Abstract

   Progressive cardiac fibrosis post myocardial infarction (MI) drives
   pathological remodeling and heart failure, yet the role of endoplasmic
   reticulum‐selective autophagy (ER‐phagy) in this process remains
   unclear. Autocrine Motility Factor Receptor (AMFR) is a recently
   identified ER‐phagy regulator, whose function under myocardial
   pathology remains poorly understood. Here, it is found that
   FAM134B‐mediated ER‐phagy activity is elevated in fibrotic mouse heart
   tissues post‐MI and in cardiac fibroblasts stimulated by TGF‐β1. AMFR
   knockout in mice aggravated cardiac fibrosis post‐MI and worsened
   cardiac function, with scRNA‐seq analysis demonstrating that AMFR‐null
   cardiac fibroblasts exhibit a myofibroblast phenotype. Simultaneously,
   AMFR overexpression in cardiac fibroblasts reduces the expression of
   profibrogenic proteins in response to TGF‐β1 stimulation. AMFR
   regulates ER‐phagy flux and turnover of FAM134B, which leads to the
   suppression of cardiac fibroblasts activation. Mechanistically, AMFR
   catalyzed K27‐linked (predominant) and K33‐linked ubiquitination of
   FAM134B and enhanced ER‐phagy flux, thereby inhibiting the
   phosphorylation of mTORC1 downstream targets such as S6K1 and 4E‐BP.
   These findings highlight the therapeutic potential of AMFR‐driven
   ER‐phagy in suppressing cardiac fibrosis post‐MI.

   Keywords: AMFR, ER‐phagy, FAM134B, fibroblast activation, mTORC1,
   myocardial infarction
     __________________________________________________________________

   By catalyzing FAM134B ubiquitination and activating ER‐phagy, AMFR
   alleviates progressive fibrosis and cardiac dysfunction by inhibiting
   the mTORC1 pathway. Consequently, these findings underscore the
   essential role of AMFR‐driven ER‐phagy in mitigating the progression of
   fibrotic responses, offering a potential therapeutic target for
   preventing heart failure post‐MI.

   graphic file with name ADVS-12-e04552-g010.jpg

1. Introduction

   Myocardial infarction (MI) is the most common and critical disease
   among cardiovascular and cerebrovascular diseases.^[ [44]^1 ^] Although
   the myocardial reperfusion therapy has significantly reduced the
   mortality of MI, a large proportion of survivors will develop heart
   failure due to adverse ventricular remodeling.^[ [45]^2 ^] Pathological
   fibrosis, characterized by excessive deposition of extracellular matrix
   (ECM), is generally recognized as central to ventricular remodeling
   post‐MI.^[ [46]^3 ^] Notably, activated fibroblasts, the main producers
   of matrix proteins, are the primary cellular effectors in cardiac
   fibrosis. Thus, elucidating the mechanisms of cardiac fibroblasts
   underlying cardiac remodeling post‐MI is crucial for developing
   targeted therapeutic interventions.

   The endoplasmic reticulum (ER) is essential for protein synthesis and
   modification in eukaryotic cells.^[ [47]^4 , [48]^5 ^] Current studies
   suggest that prolonged ER stress may lead to fibrosis through
   activation of CCAAT/enhancer‐binding homologous protein‐mediated
   apoptosis, followed by an inflammatory response and release of
   profibrotic cytokines.^[ [49]^6 ^] Endoplasmic reticulum selectively
   autophagy (ER‐phagy), a quality control mechanism that degrades damaged
   ER subdomains through the lysosomal pathway, alleviates ER stress and
   restores ER homeostasis.^[ [50]^7 ^] This selective clearance helps not
   only to eliminate abnormal ER and proteins, but also to prevent further
   cellular damage and provide cells with necessary nutrients and
   energy,^[ [51]^8 , [52]^9 ^] which may affect activation of
   fibroblasts. Recently, ER‐phagy receptors in mammals have been
   discovered successively, including FAM134B, ATL3, SEC62, CCPG1, TEX264,
   CALCOCO1, etc.^[ [53]^10 , [54]^11 , [55]^12 , [56]^13 , [57]^14 ,
   [58]^15 ^] It has been reported that ER‐phagy is favorable to diseases
   such as lung injury and kidney injury.^[ [59]^16 , [60]^17 ^] ER‐phagy
   has also been found to alleviate anthracycline‐induced cardiotoxicity
   through CCPG1‐mediated signaling.^[ [61]^18 ^] However, the role of
   ER‐phagy in the process of ventricular remodeling post‐MI and the
   mechanisms that regulate ER‐phagy have not been investigated.

   Autocrine motility factor receptor (AMFR), an E3 ubiquitin ligase
   primarily expressed on the ER membrane,^[ [62]^19 ^] is recently been
   identified as an ER‐regulator that catalyzes FAM134B ubiquitination.^[
   [63]^20 ^] It has been reported that AMFR regulates protein degradation
   and serves critical roles in biological processes such as inflammatory
   responses,^[ [64]^21 , [65]^22 ^] innate immunity,^[ [66]^23 ^]
   allergic asthma,^[ [67]^24 ^] and so on. However, existing research has
   not yet involved the regulatory role of AMFR in cardiovascular
   diseases.

   Therefore, this study aims to determine the activation status of
   ER‐phagy post‐MI and its association with fibrotic progression,
   investigate the role of AMFR in coordinating ER‐phagy and fibroblast
   activation, and elucidate the molecular mechanisms underlying
   AMFR‐mediated modulation of these processes. We found that ER‐phagy was
   activated in cardiac fibroblasts post‐myocardial infarction. AMFR
   alleviated cardiac fibrosis post‐MI through activation of ER‐phagy.
   Mechanistically, AMFR ubiquitinated FAM134B to activate ER‐phagy,
   thereby inhibiting the mTORC1 signaling pathway in cardiac fibroblasts.
   Our study identifies AMFR as an ER‐phagy regulator that suppresses
   cardiac fibrosis by promoting ubiquitination of FAM134B, and provides
   the first evidence of the crucial role of AMFR in inhibiting the mTORC1
   pathway.

2. Results

2.1. FAM134B‐Mediated ER‐phagy is Activated in Cardiac Fibroblasts Post‐MI

   To identify ER‐phagy activation post‐MI, we examined several
   established ER‐phagy receptors, such as ATL3,^[ [68]^14 ^] SEC62,^[
   [69]^11 ^] FAM134B,^[ [70]^10 ^] TEX264,^[ [71]^13 ^] CALCOCO1,^[
   [72]^15 ^] and CCPG1,^[ [73]^12 ^] as their increased transcription may
   indicate enhanced ER‐phagy activity.^[ [74]^18 ^] Among these, Fam134b,
   Ccpg1, and Sec62 were significantly up‐regulated at day 7 post‐MI, with
   Fam134b being the most prominent (Figure [75]1A). Increased
   co‐localization of FAM134B and LC3 was observed in cardiac fibroblasts
   within the infarct border zone, confirming the activation of ER‐phagy
   in fibrotic tissues after MI (Figure [76]1B).

Figure 1.

   Figure 1
   [77]Open in a new tab

   FAM134B‐mediated ER‐phagy is activated in cardiac fibroblasts post‐MI.
   A) The relative mRNA levels of Atl3, Sec62, Fam134b, Tex264, Calcoco1,
   and Ccpg1 in the mouse heart tissue sections from sham or MI mice at
   day 7 post‐MI. n = 3. B) Immunofluorescence co‐staining of LC3 (red)
   with FAM134B (green), Vimentin (white), and DAPI (blue) in the mouse
   heart tissue sections from sham or MI mice at day 7 post‐MI. C) Regimen
   of RFP‐GFP‐KDEL. D) Fluorescent imaging of RFP–GFP–KDEL transfected
   adult mouse primary cardiac fibroblasts (ACFs) isolated from WT mice
   treated with PBS or TGF‐β1(10 ng mL^−1) for 24 h. n = 4. E) Fluorescent
   imaging of RFP–GFP–KDEL transfected ACFs isolated from WT mice treated
   with si‐Ctrl or si‐FAM134B before treating with TGF‐β1(10 ng mL^−1) for
   24 h. n = 4. Data are presented as mean ± SD. The data shown in A, D,
   and E were analyzed using an unpaired t‐test. MI, myocardial
   infarction; TGF‐β1, transforming growth factor‐β1; ACFs, adult mouse
   primary cardiac fibroblasts; WT, wild‐type.

   In vitro, ER‐phagy flux was assessed using a dual‐fluorescence
   reporter, RFP‐GFP‐KDEL, as the RFP/RFP‐GFP puncta ratio reflects
   ER‐phagy flux due to quenching of GFP in autolysosomes (Figure [78]1C).
   Adult mouse primary cardiac fibroblasts (ACFs) isolated from wild‐type
   (WT) mice were treated with TGF‐β1(10 ng mL^−1) to simulate the
   activation of cardiac fibroblasts post‐MI. The results showed increased
   RFP puncta in ACFs treated with TGF‐β1, indicating enhanced ER‐phagy
   activity in cardiac fibroblasts after TGF‐β1 stimulation
   (Figure [79]1D).

   To further determine the role of FAM134B in ER‐phagy, ACFs were
   transfected with RFP‐GFP‐KDEL or mcherry‐GFP‐FAM134B, respectively,
   reflecting the total ER‐phagy flux and dedicated FAM134B‐mediated
   ER‐phagy flux. Then, cells were treated with TGF‐β1(10 ng mL^−1), and
   we found that the mCherry/mCherry‐GFP puncta ratio was close to the
   RFP/RFP‐GFP puncta ratio, indicating FAM134B being the predominant
   ER‐phagy receptor following TGF‐β1 treatment (Figure [80]S1, Supporting
   Information). Furthermore, the knockdown of FAM134B by si‐FAM134B
   abolished the increase of RFP signals induced by TGF‐β1, suggesting
   that FAM134B could be crucial in regulating TGF‐β1‐induced ER‐phagy in
   cardiac fibroblasts (Figure [81]1E). Altogether, these data elucidate
   that FAM134B‐mediated ER‐phagy is activated in fibrotic mouse heart
   tissues post‐MI and in cardiac fibroblasts stimulated by TGF‐β1.

2.2. AMFR is Up‐Regulated in Fibrotic Mice Heart Tissues After Myocardial
Infarction and in Cardiac Fibroblasts Stimulated by TGF‐β1

   Considering the strong association between AMFR and ER‐phagy,^[ [82]^20
   ^] we investigated the expression level of AMFR at different stages
   post‐MI. Mice were subjected to permanent left anterior descending
   artery (LAD) ligation or sham operation and were analyzed at 1 day, 3
   days, 7 days, 14 days, and 28 days. At the infarct and border area of
   heat, the protein levels of AMFR increased from day 3, peaked at day 7
   post‐MI, and subsequently decreased but remained higher than those in
   the sham group at 14 days post‐MI (Figure [83]2A–D). Concurrently, we
   compared the protein level of AMFR in the myocardial infarct area,
   border area, and remote area of tissues. The results showed that there
   were no significant differences in the expression levels of AMFR in the
   infarct zone and border zone tissues at all time points. The expression
   of AMFR in the infarct area was significantly higher than that in the
   remote area at day 3 post‐MI (p< 0.0001) and day 7 post‐MI (p =
   0.0007). The expression of AMFR in the border area was significantly
   higher than that in the remote area at day 1 post‐MI (p = 0.0009), day
   3 post‐MI (p = 0.0001), and day 7 post‐MI (p = 0.0002) (Figure [84]2E).
   Consistently, RT‐qPCR indicated that the mRNA level of AMFR increased
   in the infarcted area 7 days post‐MI, compared to the sham group. The
   mRNA level of AMFR in the infarcted area and border area was higher
   than that in the remote zone (Figure [85]S2, Supporting Information).
   To determine the cell types expressing AMFR during cardiac remodeling,
   we co‐stained AMFR with Vimentin, a marker of fibroblasts, and observed
   strong AMFR expression in fibroblasts in the infarct myocardium
   (Figure [86]2F). However, we observed less co‐localization of AMFR with
   the cardiomyocyte marker cTnT, while even less co‐localization of AMFR
   with the endothelial cell marker CD31 or the macrophage marker CD68
   (Figure [87]S2, Supporting Information). Single‐cell RNA‐seq
   (scRNA‐seq) provided a comprehensive view of AMFR expression across
   different cardiac cell types, and further confirmed the ubiquitous
   expression and high abundance of AMFR in cardiac fibroblasts
   (Figure [88]2G). Furthermore, we stimulated neonatal rat cardiac
   fibroblasts (NRCFs) with TGF‐β1, and Western blot analysis and RT‐qPCR
   showed a significant induction of AMFR expression (Figure [89]2H,I),
   indicating AMFR is a novel target responsive to TGF‐β1 in NRCFs. These
   data demonstrate that AMFR expression is upregulated in activated
   fibroblasts induced by myocardial infarction and is responsive to
   TGF‐β1 stimulus.

Figure 2.

   Figure 2
   [90]Open in a new tab

   AMFR is up‐regulated in fibroblasts in fibrotic heart tissues. A–D)
   Expression of AMFR was measured using Western blot analysis in the
   infarct area, border area, and remote area of WT mice at day 1, day 3,
   day 7, day 14, and day 28 post‐MI, as well as in sham control. n = 4.
   E) The XY line graph shows the differences in AMFR expression in the
   infarct area, border area, and remote area at different time points
   post myocardial infarction. n = 4. F) Immunofluorescence co‐staining of
   AMFR (red) with Vimentin (green) and DAPI (blue) in the mouse heart
   tissue sections from sham mice or MI mice at 3 days, 14 days post MI.
   G) Amfr expression in different cell types of WT mice at day 7 post‐MI.
   H) Representative Western blots and quantitative analyses of AMFR and
   α‐SMA in NRCFs treated with TGF‐β1 (10 ng mL^−1). n = 6. I) Expression
   of AMFR was measured using RT‐qPCR in NRCFs treated with TGF‐β1
   (10 ng mL^−1). n = 4. Data are presented as mean ± SD. The data shown
   in B–D was analyzed using one‐way ANOVA corrected by the post hoc
   Turkey's test. The data shown in E was analyzed using two‐way ANOVA
   corrected by the post hoc Turkey's test. The data shown in H and I were
   analyzed using an unpaired t‐test. MI, myocardial infarction; TGF‐β1,
   transforming growth factor‐β1; NRCFs, neonatal rat cardiac fibroblasts;
   WT, wild‐type.

2.3. scRNA‐seq Reveals that AMFR Deletion Promotes Fibroblast Activation and
Impairs Autophagy Activity

   To explore the mechanism of AMFR in response to acute MI, we conducted
   single‐cell RNA sequencing (scRNA‐seq) on heart tissues isolated from
   AMFR knockout and WT mice at 7 days post‐MI. After the exclusion of
   cell doublets based on unique molecular identifier counts and filtering
   based on mitochondrial genes (Table [91]S1, Supporting Information), we
   obtained high‐quality sequencing from 5218 WT and 10622 AMFR knockout
   cells. Nine main cell types were separated following principal
   component analysis (PCA) and the Louvain clustering algorithm (Figure
   [92]3A). Fibroblasts were further subdivided into 10 transcriptionally
   distinct clusters, with Clusters 2, 3, and 4 significantly increased in
   the KO group (Figure [93]3B). By analyzing the expression of key
   fibrotic marker genes, Cluster 6 was identified as quiescent resident
   fibroblasts with prominent Pdgfra expression, while Cluster 2, 3 and 4
   were identified as activated fibroblasts exhibiting high Postn
   expression (Figure [94]3C), with these identities further validated by
   hierarchical clustering and identification of the most enriched genes
   in each cluster (Figure [95]S3, Supporting Information).

Figure 3.

   Figure 3
   [96]Open in a new tab

   Single cell RNA‐seq reveals that AMFR deletion promotes cardiac
   fibroblasts activation and impairs autophagy activity. 8‐week‐old WT
   and AMFR^−/‐ male mice were subjected to MI surgery, and heart tissues
   were harvested 1 week post‐surgery for single cell RNA‐seq. A) tSNE of
   single cell transcriptomes from wild‐type (WT; 5218 cells) and AMFR
   knockout (KO; 10622 cells) mice obtained 7 days post‐MI. B) UMAP of
   single cell transcriptomes from WT and AMFR KO fibroblasts, and
   proportion of each cluster. C) The expression levels of Pdgfra and
   Postn in 10 clusters of fibroblasts. D) The distribution of 10 clusters
   of fibroblasts across pseudotime. E) The gene expression of 11 cell
   states across pseudotime and in different genotypes. F) GO enrichment
   analysis in late state(state6) and early state(state1). G) KEGG
   enrichment analysis in WT and KO Fibobalasts_2, as well as WT and KO
   Fibobalasts_4. MI, myocardial infarction; tSNE, T‐distributed
   stochastic neighbor embedding; UMAP, Uniform Manifold Approximation and
   Projection; GO, Gene ontology; KEGG, Kyoto Encyclopedia of Genes and
   Genomes.

   Pseudotemporal trajectory analysis was then performed to reveal a
   differentiation continuum from quiescent to activated states
   (Figure [97]3D; Figure [98]S3, Supporting Information). Integrating
   cell identities with the pseudotime trajectory, we observed that
   Cluster 6 with high Clec3b expression was located at the beginning of
   pseudotime, while the end of pseudotime was predominantly occupied by
   Clusters 2, 3, 4, characterized by elevated expression of Aspn, Cilp,
   Col1a1, Col1a2, Col5a2, Col8a1(Figure [99]3E). CLEC3B was found to be a
   potential target for alleviating adipose tissue fibrosis.^[ [100]^25 ^]
   Asporin is an extracellular matrix protein that regulates cardiac
   remodeling.^[ [101]^26 ^] CILP is a Novel Biomarker for Cardiac
   Fibrosis.^[ [102]^27 ^] Col1a1, Col1a2, Col5a2, and Col8a1 encode
   different types of collagen chains, which play important structural and
   functional roles in the extracellular matrix.^[ [103]^28 , [104]^29 ^]
   This suggested that the early stages of pseudotime represented a
   quiescent fibroblast phenotype, while the end indicated activated
   myofibroblasts. As fibroblasts in the KO group were particularly
   enriched in the late pseudotime clusters, Gene ontology (GO) enrichment
   analysis revealed that the late states were associated with
   extracellular matrix (ECM) organization and extracellular structure
   organization, further confirming the myofibroblast phenotype in KO
   cells (Figure [105]3F). Kyoto Encyclopedia of Genes and Genomes (KEGG)
   analysis was then performed to elucidate the underlying mechanisms.
   Cluster2 was enriched in pathways related to ECM‐receptor interaction
   and protein processing in the endoplasmic reticulum, while Cluster4 was
   enriched in pathways related to the Lysosome, phagosome, and protein
   processing in the endoplasmic reticulum (Figure [106]3G). Together,
   these findings reveal that AMFR deletion promotes cardiac fibroblast
   activation and impairs autophagy activity.

2.4. AMFR Deficiency Exacerbates Cardiac Fibrosis and Dysfunction Caused by
MI

   AMFR knockout (KO) mice were constructed using the CRISPR/Cas9
   technology (Figure [107]S4, Supporting Information). Echocardiography
   showed no noticeable differences in cardiac function between WT and
   AMFR KO mice under physiological conditions. H&E staining showed no
   significant differences in myocardial structure or arrangement between
   WT and AMFR KO mice (Figure [108]S4, Supporting Information).

   At day 1 post‐MI, no significant differences were observed between AMFR
   KO and WT mice (Figure [109]S4, Supporting Information). However, by
   day 14 post‐MI, AMFR deficiency significantly worsened cardiac
   function, as evidenced by markedly declined left ventricle ejection
   fraction (LVEF%) and left ventricle fractional shortening (LVFS%)
   (Figure [110]4A,B). Masson's trichrome staining further revealed that
   AMFR deficiency resulted in exacerbated cardiac fibrosis and increased
   collagen deposition at the infarct area compared to WT controls 14 days
   after MI (Figure [111]4C). Furthermore, AMFR deficiency significantly
   increased the MI‐induced upregulation of Col1a1, Col3a1, and α‐SMA at
   protein levels (Figure [112]4D) as well as mRNA levels
   (Figure [113]4E–G). Additionally, immunofluorescence staining showed
   that α‐SMA was significantly increased in the myocardium of AMFR KO
   mice post‐MI compared to WT littermates (Figure [114]4H). Collectively,
   these data elucidated that AMFR knockout exacerbated MI‐induced cardiac
   remodeling and dysfunction.

Figure 4.

   Figure 4
   [115]Open in a new tab

   AMFR deficiency exacerbated cardiac fibrosis and dysfunction caused by
   MI. 8‐week‐old WT and AMFR knockout male mice were subjected to Sham or
   MI surgery, and heart tissues were harvested 2 weeks post‐surgery. A,B)
   Representative serial echocardiography and quantification of
   echocardiographic parameters. n = 6. C) Representative images and
   quantitative analyses of Masson's trichrome staining. n = 4. D)
   Representative Western blots and quantitative analyses of α‐SMA,
   COL1A1, and COL3A1 in heart tissues. n = 4. E–G) The relative mRNA
   levels of Acta2, Col1a1, and Col3a1 in heart tissues. n = 4. H)
   Representative images and quantitative analyses of immunofluorescence
   staining with α‐SMA (red) and DAPI (blue). n = 4. Data are presented as
   mean ± SD. The data were analyzed using two‐way ANOVA corrected by the
   post hoc Turkey's test. MI, myocardial infarction; LVEF, left ventricle
   ejection fraction; LVFS, left ventricle fractional shortening; LVEDV,
   left ventricle end diastolic volume; LVESV, left ventricle end systolic
   volume.

2.5. AMFR Overexpression Inhibits Fibroblast Proliferation, Migration, and
Myofibroblast Differentiation

   Adenoviral vectors (Adv‐AMFR) were used to overexpress AMFR in NRCFs
   (Figure [116]S5, Supporting Information). Fibroblast proliferation and
   migration were assessed using EdU and wound healing assays,
   respectively. Compared to NRCFs with Adv‐Control, AMFR overexpression
   markedly inhibited fibroblast proliferation and migration in response
   to TGF‐β1 administration (Figure [117]5A,B). Of note, α‐SMA
   fluorescence staining indicated that AMFR overexpression significantly
   inhibited TGF‐β1‐induced fibroblast‐to‐myofibroblast
   transdifferentiation (Figure [118]5C). Western blot analysis confirmed
   that AMFR overexpression downregulated the protein levels of α‐SMA,
   COL1A1, and COL3A1 in fibroblasts after TGF‐β1 administration
   (Figure [119]5D,E). Consistently, overexpression of AMFR also reduced
   the TGF‐β1‐induced upregulation of α‐SMA, Col1a1, and Col3a1 mRNA
   levels in fibroblasts (Figure [120]5F–H). Collectively, these results
   indicated that AMFR overexpression suppressed fibroblast proliferation,
   migration, and differentiation into myofibroblasts.

Figure 5.

   Figure 5
   [121]Open in a new tab

   Overexpression of AMFR in cardiac fibroblasts inhibited proliferation,
   migration, and transdifferentiation of fibroblasts in vitro. NRCFs were
   isolated from 1 to 3‐day‐old neonatal Sprague‐Dawley rats and
   transfected with Adv‐AMFR or Adv‐Flag for 48 h before being treated
   with TGF‐β1(10 ng mL^−1) for 24 h. A) Representative image and
   quantitative analyses showing EdU (red) localization with DAPI (blue).
   n = 4. B) Representative images and quantitative analyses of wound
   healing assays. n = 4. C) Representative images and quantitative
   analyses of immunofluorescence staining with α‐SMA (red) and DAPI
   (blue). n = 4. D,E) Representative Western blots and quantitative
   analyses of α‐SMA, COL1A1, and COL3A1 in NRCFs. n = 4. F–H) The
   relative mRNA levels of Acta2, Col1a1, and Col3a1 in NRCFs. All
   normalized to HPRT. n = 3. Data are presented as mean ± SD. The data
   were analyzed using two‐way ANOVA corrected by the post hoc Turkey's
   test. TGF‐β1, transforming growth factor‐β1; NRCFs, neonatal rat
   cardiac fibroblasts.

2.6. AMFR Deficiency Impairs ER‐phagy in Cardiac Fibroblasts

   As FAM134B‐mediated ER‐phagy is activated in cardiac fibroblasts after
   TGF‐β1 stimulation (Figure [122]1), mCherry‐GFP‐FAM134B was employed to
   evaluate the role of AMFR in regulating ER‐phagy activity. After TGF‐β1
   administration, mCherry signals in AMFR knockout fibroblasts were
   reduced compared to WT fibroblasts, demonstrating impaired ER‐phagy
   (Figure [123]6A). Furthermore, Western blot analysis indicated that
   overexpression of AMFR decreased the protein level of FAM134B and ER
   marker Calnexin after TGF‐β1 administration (Figure [124]6B,C).
   Consistently, AMFR deficiency significantly increased the protein
   levels of FAM134B and Calnexin in heart tissues post‐MI
   (Figure [125]6D,E). Transmission electron microscope revealed that
   AMFR‐overexpressing fibroblasts contained massive autophagic vesicles
   and exhibited reduced ER enlargement and swelling (Figure [126]6F).
   Together, these findings suggest that AMFR plays a crucial role in
   regulating ER‐phagy in cardiac fibroblasts.

Figure 6.

   Figure 6
   [127]Open in a new tab

   AMFR deficiency causes ER‐phagy defects in cardiac fibroblasts. A) ACFs
   isolated from WT or AMFR knockout mice transfected with
   Adv‐mCherry–GFP–FAM134B before being treated with TGF‐β1(10 ng mL^−1).
   n = 4. B,C) Representative Western blots of FAM134B and Calnexin in
   NRCFs transfected with Adv‐AMFR or Adv‐Flag before being treated with
   TGF‐β1(10 ng mL^−1). n = 4. D,E) Representative Western blots of
   FAM134B and Calnexin in heart tissues of WT and AMFR knockout mice 2
   weeks post‐MI surgery. n = 4. F) Transmission electron microscopic
   imaging of NRCFs transfected with Adv‐AMFR or Adv‐Flag before being
   treated with PBS or TGF‐β1(10 ng mL^−1). Data are presented as mean ±
   SD. The data shown in A, C, and E were analyzed using an unpaired
   t‐test. MI, myocardial infarction; TGF‐β1, transforming growth
   factor‐β1; ACFs, adult mouse primary cardiac fibroblasts; NRCFs,
   neonatal rat cardiac fibroblasts.

2.7. AMFR Ubiquitinates FAM134B and Promotes ER‐phagy to Suppress Cardiac
Fibroblast Activation

   Given the role of AMFR in promoting ER‐phagy, we sought to investigate
   its underlying mechanism and its impact on cardiac fibroblast
   activation. As depicted in Figure [128]6B, overexpression of AMFR
   significantly reduced FAM134B protein levels. Subsequently, we aimed to
   determine whether this effect occurred at the transcriptional or
   post‐translational level by measuring FAM134B mRNA levels following
   AMFR overexpression. The results revealed that no significant changes
   in the transcriptional level of FAM134B were observed after AMFR
   overexpression (Figure [129]7A). AMFR overexpression decreased the
   protein level of FAM134B following cotreatment with a protein synthesis
   inhibitor, cycloheximide (CHX) (Figure [130]7B), indicating that AMFR
   may regulate the turnover of FAM134B post‐translationally.

Figure 7.

   Figure 7
   [131]Open in a new tab

   AMFR ubiquitinates FAM134B and promotes ER‐phagy to suppress cardiac
   fibroblast activation. A) The relative mRNA level of FAM134B in NRCFs
   transfected with Adv‐AMFR or Adv‐FLAG, followed by administration of
   TGF‐β1(10 ng mL^−1), measured using RT‐qPCR analysis. n = 3. B) Western
   blot analysis of FAM134B in NRCFs transfected with Adv‐AMFR or Adv‐FLAG
   and incubated with cycloheximide (CHX) for the indicated times,
   followed by administration of TGF‐β1(10 ng mL^−1). n = 3. C) NRCFs were
   transfected with Adv‐AMFR or Adv‐FLAG, then incubated with chloroquine
   (CQ) or MG132, followed by administration of TGF‐β1(10 ng mL^−1) for
   24 h. n = 3. D) NRCFs were transfected with Adv‐Flag or Adv‐AMFR for
   48 h. Immunoprecipitation with the antibody against AMFR, followed by
   immunoblot with anti‐FAM134B antibody. IgG as a negative control. n =
   3. E) NRCFs were transfected with Adv‐Flag or Adv‐AMFR for 48 h.
   Immunoprecipitation with the antibody against FAM134B, followed by
   immunoblot with anti‐AMFR, anti‐UB, and anti‐LC3 antibodies. IgG as a
   negative control. n = 3. F) HEK293T cells transduced with His‐Ub or
   His‐K6R‐UB, His‐K11R‐UB, His‐K27R‐UB, His‐K29R‐UB, His‐K33R‐UB,
   His‐K48R‐UB, His‐K63R‐UB, and MYC‐FAM134B, Flag‐AMFR vectors were
   immunoprecipitated with the antibody against MYC, and followed by
   immunoblot with anti‐UB antibody. IgG as a negative control. n = 3.
   G,H) Representative Western blots of α‐SMA in NRCFs transfected with or
   without Adv‐AMFR and siRNA‐FAM134B before being treated with
   TGF‐β1(10 ng mL^−1). n = 4. Data are presented as mean ± SD. The data
   shown in A were analyzed using an unpaired t‐test. The data shown in B,
   C, and H were analyzed using two‐way ANOVA corrected by the post hoc
   Turkey's test. TGF‐β1, transforming growth factor‐β1; CHX,
   cycloheximide; CQ, chloroquine; NRCFs, neonatal rat cardiac
   fibroblasts; UB, ubiquitin.

   Intracellular protein degradation is mainly mediated by the proteasomal
   or lysosomal pathway.^[ [132]^30 ^] Western blot analysis showed that
   overexpression of AMFR leads to the degradation of FAM134B, which can
   be reversed by the autophagy inhibitor chloroquine (CQ), but not by the
   proteasome inhibitor MG132 (Figure [133]7C), indicating that the
   degradation of FAM134B induced by AMFR is mediated through the
   autophagy pathway. Moreover, the Co‐IP assay confirmed the interaction
   between FAM134B and LC3 in NRCFs. We also examined the interaction
   between AMFR and other endoplasmic reticulum autophagy receptors in
   fibroblasts, and the results showed no significant interaction between
   AMFR and SEC62 as well as CCPG1 (Figure [134]S6, Supporting
   Information).

   As AMFR functions as an E3 ubiquitin ligase, we next investigated
   whether AMFR participates in the ubiquitination of FAM134B. We found
   that FAM134B ubiquitination was markedly increased in fibroblasts
   overexpressing AMFR (Figure [135]7D,E). As different types of
   polyubiquitin linkages mediate distinct biological functions, we
   expressed Myc‐tagged FAM134B in 293T cells together with various
   ubiquitin mutants (K6, K11, K27, K29, K33, K48, and K63), all of whose
   lysines were replaced with arginine. Notably, the K27R substitution and
   K33R substitution, rather than other point mutations, attenuated
   AMFR‐mediated ubiquitination of FAM134B (Figure [136]7F). To clarify
   which linkage (K27 or K33) predominates, HEK293T cells were transfected
   with His‐UB, His‐K27R‐UB, or His‐K33R‐UB, followed by treatment with
   chloroquine (CQ) to inhibit the degradation of autophagolysosome
   contents. Consistent with the previous results (Figure [137]7F), the
   K27R substitution and K33R substitution attenuated AMFR‐mediated
   ubiquitination of FAM134B. Chloroquine treatment increases the detected
   levels of FAM134B ubiquitination across all groups by inhibiting the
   degradation of ubiquitinated FAM134B. Notably, the ubiquitination
   levels in the K33R substitution group showed a more pronounced increase
   compared to the K27R group (Figure [138]S7, Supporting Information),
   supporting the dominance of K27‐linked ubiquitination. Importantly,
   knockdown of FAM134B by siRNA abolished AMFR‐mediated inhibition of
   α‐SMA level in cardiac fibroblasts (Figure [139]7G), providing further
   evidence that the inhibitory effects of AMFR on fibroblasts are, at
   least in part, exerted by the promotion of ER‐phagy. Collectively,
   these data prove that AMFR ubiquitinates FAM134B and targets it for
   degradation via the autophagy pathway, which is crucial for
   AMFR‐mediated suppression of cardiac fibroblast activation.

2.8. AMFR Inhibits mTORC1 Signaling in a FAM134B‐Dependent Manner

   To fully understand the molecular mechanisms underlying AMFR‐mediated
   cardiac remodeling, we next identified the pathway most significantly
   affected by AMFR. KEGG pathway enrichment analysis was used to identify
   the biological pathways involved in cardiac fibroblasts from WT and
   AMFR knockout mice after MI. The results revealed that the PI3K‐AKT
   signaling pathway was significantly enriched (Figure [140]8A). As the
   TGF‐β/SMAD pathway is a canonical pathway for the activation of
   fibroblasts, we detected the phosphorylation levels of SMAD2 and SMAD3.
   Our findings revealed that AMFR did not affect the phosphorylation
   levels of SMAD2 or SMAD3 (Figure [141]S8, Supporting Information).
   Since mTORC1 is downstream of PI3K‐AKT signaling pathway known to
   promote fibrosis,^[ [142]^31 , [143]^32 ^] we then examined
   phosphorylation of mTOR and key effectors of mTORC1, including S6K1 and
   4E‐BP1.^[ [144]^33 , [145]^34 , [146]^35 ^] The results confirmed that
   AMFR deficiency significantly increased the MI‐induced upregulation of
   p‐mTOR, p‐P70 S6K, and p‐4EBP1 protein levels (Figure [147]8B). In
   NRCFs, overexpression of AMFR inhibited the mTORC1 signaling pathway
   after TGF‐β1 administration, as evidenced by decreased p‐mTOR, p‐S6K,
   and p‐4E‐BP1 levels (Figure [148]8C,D). To verify whether the
   antifibrotic effect of AMFR is directly mediated through the mTORC1
   pathway, we employed rapamycin to inhibit the mTORC1 pathway.
   siRNA‐mediated AMFR silencing enhanced α‐SMA protein levels, but
   co‐treatment with rapamycin effectively abrogated the pro‐fibrotic
   effect induced by si‐AMFR (Figure [149]S8, Supporting Information).

Figure 8.

   Figure 8
   [150]Open in a new tab

   AMFR inhibits the mTORC1 signaling in a FAM134B‐dependent manner. A)
   KEGG analysis identified changed pathways in the cardiac fibroblasts
   isolated from WT or AMFR‐knockout mice post‐MI. B) Representative
   Western blots of p‐mTOR, p‐S6K, and p‐4E‐BP1 in heart tissues of WT and
   AMFR KO mice at 2 weeks post‐MI. n = 4. C) Representative Western blots
   of p‐mTOR, p‐S6K, and p‐4E‐BP1 in NRCFs transfected with Adv‐AMFR or
   Adv‐Flag before being treated with TGF‐β1(10 ng mL^−1). n = 4. D)
   Representative statistical result of B and C. E,F) Representative
   Western blots and statistical result of p‐4EBP1 and p‐S6K in NRCFs
   transfected with or without Adv‐AMFR and siRNA‐FAM134B before being
   treated with TGF‐β1(10 ng mL^−1). n = 4. G) Schematic diagram of the
   experimental model. 8‐week‐old WT or AMFR‐knockout male mice were
   transfected with Adv‐FAM134B or Adv‐Vector via intracardiac injection
   for one week before being subjected to Sham or MI surgery, and heart
   tissues were harvested 2 weeks post‐surgery. H) Masson's trichrome
   staining and the corresponding statistical result. n = 4. I)
   Quantification of echocardiographic parameters of ejection fraction
   (EF%) and fractional shortening (FS%). n = 6. J,K) Representative
   Western blots and statistical results of α‐SMA, COL3A1, p‐4EBP1, and
   p‐S6K in heart tissues. n = 4. Data are presented as mean ± SD. The
   data in D, F, H, I, and K were analyzed using two‐way ANOVA corrected
   by the post hoc Turkey's test. KEGG, Kyoto Encyclopedia of Genes and
   Genomes; MI, myocardial infarction; TGF‐β1, transforming growth
   factor‐β1; NRCFs, neonatal rat cardiac fibroblasts; LVEF, left
   ventricle ejection fraction; LVFS, left ventricle fractional
   shortening.

   Given the exclusive functional specificity of FAM134B in mediating
   ER‐phagy, we performed FAM134B knockdown to attenuate ER‐phagy. The
   results showed that knockdown of FAM134B by siRNA abolished
   AMFR‐mediated inhibition of p‐S6K and p‐4E‐BP1 levels in the cardiac
   fibroblasts (Figure [151]8E,F). To further explore the in vivo
   relevance, WT and AMFR KO mice were transfected with Adv‐Flag or
   Adv‐FAM134B prior to MI construction (Figure [152]8G). Transfection
   with Adv‐FAM134B in the myocardium did not affect cardiac function or
   fibrosis in sham mice at 2 weeks (Figure [153]S9, Supporting
   Information). Masson's trichrome revealed that overexpression of
   FAM134B significantly reversed the increase of fibrosis in AMFR
   knockout mice (Figure [154]8H). In MI mice, echocardiographic analysis
   showed that overexpression of FAM134B alleviated the decreased EF% and
   FS% in mice caused by AMFR deficiency (Figure [155]8I). Western blot
   analysis revealed that α‐SMA, COL3A1, p‐S6K, and p‐4E‐BP1 levels were
   significantly increased in the heart tissues of AMFR KO mice, but were
   reversed after overexpressing FAM134B (Figure [156]8J,K). Accordingly,
   these findings suggest that AMFR promotes ER‐phagy, thereby
   contributing to the inhibition of mTORC1 signaling and suppressing
   fibroblast activation.

2.9. Fibroblast‐Specific AMFR Overexpression Attenuates Cardiac Fibrosis and
Improves Cardiac Function Post‐MI

   To investigate the therapeutic potential of AMFR in cardiac remodeling,
   we established fibroblast‐specific AMFR‐overexpressing mice by
   administering rAAV9‐Tcf21‐AMFR via tail vein injection (Figure
   [157]9A). Western blot confirmed successful fibroblast‐specific AMFR
   overexpression in mice (Figure [158]9E). Echocardiographic analysis
   revealed that AMFR overexpression in fibroblasts mitigated cardiac
   dysfunction induced by MI, indicated by increased EF% and FS%
   (Figure [159]9B,C). Masson's trichrome staining showed that collagen
   deposition caused by MI was significantly reduced in the
   AMFR‐overexpressing mice hearts (Figure [160]9D). Western blot analysis
   revealed that AMFR overexpression attenuated MI‐induced increases in
   α‐SMA, COL3A1, p‐S6K and p‐4E‐BP1 levels (Figure [161]9E–H).
   Consequently, the in vivo experiments proved that fibroblast‐specific
   AMFR overexpression is a potential target to inhibit the progression of
   cardiac fibrosis post‐MI.

Figure 9.

   Figure 9
   [162]Open in a new tab

   Fibroblast‐specific AMFR overexpression attenuated cardiac fibrosis and
   improved cardiac dysfunction post‐MI. 8‐week‐old WT male mice were
   transfected with AAV9‐Tcf21‐AMFR or AAV‐Vector via tail vein injection
   one week after being subjected to Sham or MI surgery, and heart tissues
   were harvested 4 weeks post‐surgery. A) Treatment regimen. B,C)
   Representative serial echocardiography and quantification of
   echocardiographic parameters of ejection fraction (EF%) and fractional
   shortening (FS%). n = 6. D) Masson's trichrome staining and the
   corresponding statistical result of fibrosis. n = 4. E,F)
   Representative Western blots and statistical results of α‐SMA and
   COL3A1 in heart tissues. n = 4. G,H) Representative Western blots and
   statistical results of p‐S6K and p‐4E‐BP1 in heart tissues. n = 4. Data
   are presented as mean ± SD. The data were analyzed using two‐way ANOVA
   corrected by the post hoc Turkey's test. MI, myocardial infarction;
   LVEF, left ventricle ejection fraction; LVFS, left ventricle fractional
   shortening.

3. Discussion

   In this study, we have elucidated the crucial role of AMFR‐mediated
   ER‐phagy in cardiac remodeling following MI. Our findings highlight
   that AMFR targets the ubiquitination of the ER‐phagy receptor FAM134B,
   facilitating ER‐phagy and subsequent inhibition of mTORC1 signaling in
   cardiac fibroblasts. This discovery not only uncovers the role of
   ER‐phagy in regulating cardiac fibroblasts but also unveils the
   AMFR/FAM134B/mTORC1 axis as a novel mechanism during cardiac
   remodeling, providing potential therapeutic targets for the prevention
   of heart failure post‐MI (Figure [163]10 ).

Figure 10.

   Figure 10
   [164]Open in a new tab

   Schematic depiction of the regulatory mechanism in which AMFR catalyzes
   K27‐linked (predominant) and K33‐linked ubiquitination of FAM134B and
   promotes ER‐phagy activity, thereby inhibiting the mTORC1 signaling
   pathway during the progression of fibrosis post‐MI.

   Cardiac fibrosis acts as a double‐edged sword in the context of
   myocardial infarction repair. Initial fibrosis at the early stage
   post‐MI is essential for preventing ventricular rupture and maintaining
   cardiac integrity.^[ [165]^36 ^] Our previous study revealed that
   Cthrc1 deficiency attenuated fibrosis repair at the early stage
   post‐MI, leading to increased mortality attributable to cardiac
   rupture.^[ [166]^37 ^] However, unrestrained or excessive activation of
   fibroblasts can drive pathological fibrosis. Timely activation and
   suppression of pro‐fibrotic signals are critical to prevent adverse
   remodeling and cardiac dysfunction.^[ [167]^38 ^] Here, we found that
   the AMFR/FAM134B/mTORC1 axis is pivotal in regulating cardiac fibrosis
   post‐MI.

   After myocardial infarction, the level of autophagy exhibits dynamic
   changes and plays an important role in myocardial repair and
   remodeling.^[ [168]^39 ^] During both the subacute and chronic stages
   (1 and 3 weeks postinfarction, respectively), autophagy was found to be
   activated in mice subjected to MI surgery. Moreover, the activation of
   autophagy attenuates cardiac remodeling and improves cardiac
   function.^[ [169]^40 , [170]^41 ^] In this study, we discover that
   ER‐phagy is activated in fibrotic mouse‐heart tissues after myocardial
   infarction and in cardiac fibroblasts stimulated by
   TGF‐β1(Figure [171]1). ER‐phagy is the selective autophagy mediated by
   specific receptors, contributing to ER homeostasis and alleviating ER
   stress.^[ [172]^7 ^] So far, researchers have found ER‐phagy is
   associated with various human disorders such as neurological disorders,
   cancer, and infections.^[ [173]^8 , [174]^42 , [175]^43 ^] Despite
   this, reports on the role of ER‐phagy in cardiovascular diseases,
   including myocardial infarction, are still limited at present. It has
   been found to alleviate anthracycline‐induced cardiotoxicity through
   CCPG1‐mediated signaling[17]. An increasing number of studies have
   demonstrated that inhibiting endoplasmic reticulum stress or promoting
   autophagy are strategies for the prevention and treatment of
   fibrosis.^[ [176]^44 , [177]^45 ^] However, the function of the
   selective ER‐phagy has not been explored in the context of fibroblast
   activation nor in fibrosis diseases. Our investigation reveals
   significant ER‐phagy activation in both fibrotic myocardium from
   MI‐induced mouse models and TGF‐β1‐stimulated cardiac fibroblasts, with
   FAM134B showing the most pronounced upregulation among ER‐phagy
   receptors. These findings highlight the potential significance of
   ER‐phagy in regulating fibrotic remodeling post‐MI, underscoring the
   critical involvement of ER quality control pathways in MI
   pathophysiology.

   The first discovered mammalian ER‐phagy receptor, FAM134B (also known
   as RETREG1), was initially identified as an oncogene in esophageal
   squamous cell carcinoma. Over the past few decades, the potent
   biological functions of FAM134B have been gradually revealed. A large
   body of evidence has shown that its dysfunction is closely related to a
   variety of pathological processes, including neuropathy, viral
   replication, inflammation, and cancer.^[ [178]^46 ^] Building upon
   previous findings that AMFR catalyzes FAM134B ubiquitination,^[
   [179]^20 ^] our study advances this understanding by
   identifying K27‐linked (predominant) and K33‐linked ubiquitination of
   FAM134B by AMFR in cardiac fibroblasts as a critical mechanism driving
   ER‐phagy activation. Notably, prior work by D. Mookherjee et al.^[
   [180]^47 ^] demonstrated that FAM134B mediates autophagosomal
   degradation of both AMFR and OPA1, a mitochondrial fusion protein,
   thereby maintaining ER‐mitochondria homeostasis. This mechanistic
   insight at the cellular scale now calls for extended investigation into
   how such degradation pathways might have an effect at the organ level
   and contribute to pathophysiological cascades. Our findings of
   ubiquitination‐dependent regulation align with the established role of
   FAM134B in coordinating ER and mitochondrial quality control. In our
   study, we extended the role of ER‐phagy to the realm of fibroblast
   activation and confirmed that AMFR suppressed cardiac fibrosis post‐MI
   through FAM134B‐mediated ER‐phagy. These results position AMFR not
   merely as a ubiquitination enzyme but as a central orchestrator of ER
   quality control and fibroblast activation through dynamic interactions
   with FAM134B.

   Substantial evidence has verified the effects of AMFR in regulating
   invasion and metastasis of tumor cells,^[ [181]^48 ^] inflammatory
   responses,^[ [182]^21 , [183]^22 ^] innate immunity,^[ [184]^23 ^]
   allergic asthma,^[ [185]^24 ^] and differentiation of lung
   fibroblasts.^[ [186]^49 ^] Notably, previous research has highlighted
   the pro‐fibrotic role of AMFR in pulmonary contexts, where it promotes
   lung fibroblast differentiation via the CCL1‐AMFR‐SPRY1 axis,
   activating the Ras‐ERK pathway.^[ [187]^49 ^] Paradoxically, our study
   unveils its previously unrecognized cardioprotective role in
   suppressing cardiac fibrosis through a fundamentally distinct
   mechanism. Our study identifies AMFR as a critical suppressor of
   cardiac fibrosis, evidenced by exacerbated fibrotic remodeling and
   cardiac dysfunction post‐MI in global AMFR knockout mice, and further
   supported by pseudotemporal trajectory analysis revealing enhanced
   myofibroblast differentiation in AMFR deficient fibroblasts. This
   opposite function likely stems from differences in organ‐specific
   microenvironments, fibroblast types, and downstream signaling networks.
   Fibroblasts are diverse in function and transcription across and within
   organs. Despite sharing similar properties and responding to many of
   the same signal transduction pathways, tissue specific fibroblast
   functions and lineages exist to support the developmental, homeostatic,
   and repair needs of specific organs. During development, cardiac
   fibroblasts form from epicardial and endocardial epithelial cells via
   epithelial‐to‐mesenchymal transition (EMT) and
   endothelial‐to‐mesenchymal transition (EndMT), respectively. This is in
   contrast to the fibroblasts of other organs, such as the lung, that
   arise from progressive differentiation of mesenchymal progenitors.^[
   [188]^50 , [189]^51 , [190]^52 ^] Therefore, the same protein or
   molecule may exert entirely different effects in fibroblasts of
   different organs. This “paradoxical” phenomenon suggests that therapies
   targeting AMFR should be precisely regulated in conjunction with
   organ‐specific mechanisms.

   The TGF‐β/Smad signaling pathway and PI3K/AKT/mTORC1 signaling pathway
   are the classic pathways mediating tissue fibrosis.^[ [191]^53 ^] KEGG
   analysis showed that the PI3K‐AKT signaling pathway was significantly
   enriched in AMFR knockout fibroblast post‐MI. mTORC1 phosphorylates two
   key effectors of translation, S6K1 and 4E‐BP1, to conduct protein
   synthesis, which promotes cell proliferation and fibrotic response.^[
   [192]^33 , [193]^34 , [194]^35 ^] Our findings revealed that AMFR
   inhibits phosphorylation of S6K1 and 4E‐BP1 both in vitro and in vivo
   experiments rather than SMAD2 and SMAD3. Compared with other known
   fibrosis regulatory pathways (such as TGF‐β/Smad, Wnt/β‐catenin, etc.),
   mTORC1 integrates nutrient and energy signals, directly regulating
   cellular anabolism and the synthesis of extracellular matrix
   proteins.^[ [195]^54 ^] The specificity of AMFR in inhibiting the
   mTORC1 pathway highlights its intricate and unique role in modulating
   autophagy and protein synthesis. mTORC1 serves as a pivotal regulator
   throughout the autophagic cascade—suppressing autophagosome initiation
   through Atg13/ULK1 phosphorylation and TFEB inactivation.^[ [196]^55 ^]
   In turn, lysosome‐derived amino acids (leucine and arginine) and also
   other nutrients (glucose‐DHAP, cholesterol, SAM) can partially
   reactivate Rag GTPase–nutrient signaling upstream of mTORC1.^[ [197]^56
   , [198]^57 ^] This indicates that there is a mutual regulatory
   relationship between mTORC1 and autophagy. Our research has uncovered
   that AMFR promotes ER‐phagy (Figure [199]6) while simultaneously
   inhibiting the mTORC1 pathway (Figure [200]8), which is not
   contradictory to the established theory that mTORC1 inhibits autophagy.
   However, this presents a critical mechanistic dilemma: who serves as
   the upstream regulator—ER‐phagy or the mTORC1 pathway? By conducting in
   vitro FAM134B knockdown and in vivo FAM134B overexpression experiments,
   we have demonstrated that ER‐phagy is indeed upstream of the mTORC1
   pathway. This epistatic relationship definitively positions ER‐phagy
   upstream of mTORC1 in cardiac fibroblasts, establishing a novel
   regulatory axis where AMFR/FAM134B‐mediated ER‐phagy restrains
   mTORC1‐driven profibrotic signaling. Accordingly, we claimed that AMFR
   promotes the ER‐phagy activity, contributing to the inhibition of the
   mTORC1 signaling associated with fibroblast activation. For the first
   time, we have demonstrated that ER‐phagy can regulate mTORC1 signaling,
   suggesting that AMFR/FAM134B‐mediated ER‐phagy is a pivotal mechanism
   in the regulation of fibroblast activation.

   The mechanisms by which AMFR affects S6K and 4E‐BP1 through ER‐phagy
   are likely highly complex and still need further investigation. The
   mTORC1 complex comprises mTOR, Raptor, mLST8, and two inhibitory
   subunits, PRAS40 and DEPTOR. Traditionally, mTORC1 is recruited to the
   lysosomal surface and activated. This process is mediated by the Rag
   GTPase dimer, which is anchored to the lysosome through its interaction
   with the Ragulator complex.^[ [201]^58 , [202]^59 , [203]^60 , [204]^61
   ^] However, a recent study has found that mTORC1 can also be activated
   outside of the lysosome,^[ [205]^62 ^] which implies that mTORC1 may be
   affected during the early stages of ER‐phagy, rather than being
   influenced only on the lysosomal surface. Thus, ER‐phagy may inhibit
   the mTORC1 pathway by affecting the degradation of mTORC1 components,
   the localization and activation states of mTORC1 in lysosomal and
   non‐lysosomal compartments, and the regulatory effects of Rheb on
   mTORC1. However, in the process of investigating the mechanisms through
   which ER‐phagy inhibits mTORC1, our study confronts two primary
   limitations. First, there is a gap in understanding the activation
   mechanisms of non‐lysosomal mTORC1; Second, research methods, such as
   lysosomal immunoprecipitation and Immuno‐Electron microscope, serve as
   significant technical bottlenecks. These limitations prevent us from
   directly exploring the specific molecular mechanisms by which ER‐phagy
   inhibits mTORC1 now.

   Our research indicates that AMFR‐catalyzed K27‐and K33‐linked
   ubiquitination of FAM134B is a key mechanism in promoting ER‐phagy in
   fibroblasts. However, this is unlikely to be the sole mechanism. Other
   E3 ubiquitin ligases, such as TRIM13, have been shown to recruit LC3
   via self‐ubiquitination,^[ [206]^63 ^] potentially modulating ER‐phagy
   under diverse stress conditions. Additionally, non‐canonical
   ubiquitination types or alternative post‐translational modifications,
   such as phosphorylation or ufmylation,^[ [207]^64 , [208]^65 , [209]^66
   ^] might fine‐tune FAM134B activity or recruit alternative ER‐phagy
   receptors like SEC62 or CCPG1. Thus, while AMFR‐mediated ubiquitination
   is pivotal, the ER‐phagy network likely operates through redundant,
   tissue‐specific mechanisms to ensure robust stress adaptation.

   Our research identifies the AMFR/FAM134B/mTORC1 axis as a critical
   regulator of cardiac fibrosis, primarily through its role in modulating
   ER‐phagy. AMFR and FAM134B are expressed ubiquitously across tissues.
   Meanwhile, mTORC1 plays a crucial regulatory role in various organs,
   involving multiple aspects such as cell growth, proliferation,
   metabolism, and stress responses.^[ [210]^67 ^] mTORC1 has been
   reported to regulate skin wound healing, promote renal fibrosis,
   et al.^[ [211]^68 , [212]^69 ^] Thus, targeting the AMFR/FAM134B/mTORC1
   signaling axis to develop antifibrotic therapies represents a promising
   therapeutic strategy worthy of in‐depth investigation. Key research
   directions include developing small–molecule agonists or inhibitors
   targeting AMFR or the ER–phagy receptors to modulate the critical nodes
   of the axis, exploring AAV‐mediated gene therapy to regulate the
   expression of key genes such as AMFR or FAM134B, and utilizing
   biologics like antibodies targeting downstream proteins of mTORC1 to
   intervene in downstream functions. However, the field faces formidable
   challenges. The primary difficulty is that mTORC1 regulates a wide
   range of fundamental physiological processes,^[ [213]^67 ^] so drugs
   targeting this axis need to be highly selective to avoid side effects.
   It is essential to achieve tissue–and subcellular–specific drug
   delivery (e.g., targeting cardiac fibroblasts) to prevent off–target
   effects. Moreover, it is crucial to thoroughly elucidate the regulatory
   mechanisms of this axis in fibrosis of other organs and in other
   diseases (such as cancer, metabolic diseases, and neurodegenerative
   diseases) to provide a theoretical basis for precise intervention and
   to facilitate the safe clinical translation of these therapeutic
   strategies.

   There are some limitations in this study. First, the detection of
   ER‐phagy flux within tissues remains challenging, which complicates our
   understanding of its role during MI. Nevertheless, through in vivo
   immunofluorescence co‐localization experiments, we have preliminarily
   validated the role of AMFR‐mediated ER‐phagy in the context of
   myocardial infarction. Second, while our study confirmed that
   FAM134B‐mediated ER‐phagy is essential to the inhibition of cardiac
   fibroblasts, the effects of other ER‐phagy receptors, such as CCPG1 and
   Sec62, which are also increased after TGF‐β1 administration, remain
   unclear. Different ER‐phagy receptors share the same ligands (such as
   LC3/GABARAP) or upstream signaling hubs (such as the ULK1 complex).^[
   [214]^8 ^] Therefore, it cannot be ruled out that SEC62 and CCPG1 may
   compensate for the functions of FAM134B when it is downregulated.
   Future research should verify these mechanisms and their dysregulation
   in related diseases. Third, overexpression of FAM134B activated
   ER‐phagy activity in AMFR KO mice, suggesting that other ligands may
   also activate FAM134B and enhance ER‐phagy activity, which is necessary
   to clarify in the future. Fourth, our study lacks validation using
   human heart samples. Future research should conduct relevant
   experiments to enhance its clinical translational value. Lastly,
   although we demonstrated that AMFR inhibited the mTORC1 pathway, the
   specific protein targets involved in this process need further
   investigation. Additional molecular experiments and animal studies are
   required to fully elucidate the role and mechanisms of ER‐phagy in
   myocardial infarction.

4. Conclusion

   This study uncovers a novel association between myocardial fibrosis and
   AMFR/FAM134B‐mediated ER‐phagy. By catalyzing FAM134B ubiquitination
   and activating ER‐phagy, AMFR alleviates progressive fibrosis and
   cardiac dysfunction by inhibiting the mTORC1 pathway. Consequently, our
   findings underscore the essential role of AMFR‐driven ER‐phagy in
   mitigating the progression of fibrotic responses, offering a potential
   therapeutic target for preventing heart failure post‐MI.

5. Experimental Section

Experimental Animals

   Amfr conventional knockout (Amfr^−/− ) mice were generated by Cyagen
   Bioscience Inc. (Guangzhou, China), while wild‐type (WT) C57BL/6 mice
   were purchased from Beijing Vital River Laboratory Animal Technology
   Co., Ltd. (Beijing, China). All animals were housed in standard
   conditions, with 12 h light/dark cycles, constant temperature, and ad
   libitum access to food and water throughout the experiments. All animal
   experiments were approved by The Animal Care And Use Committee of
   Shanghai Jiao Tong University Affiliated Sixth People's Hospital
   (Permit Number: DWLL2025‐0055), in accordance with the Guide for the
   Care and Use of Laboratory Animals established by the US National
   Institutes of Health (NIH Publication, 8th Edition, 2011).

   Genotyping of all animals was carried out by PCR amplification of DNA
   extracted from mouse tail samples to confirm the knockout of Amfr. The
   genotyping strategies are shown in Table [215]S2 (Supporting
   Information).

Mice Model of Myocardial Infarction

   Male mice 8 weeks old were anesthetized using 1.5% isoflurane with
   oxygen at a flow rate of 0.5 L min^−1 via mechanical ventilation.
   Myocardial infarction was conducted by ligation of the left anterior
   descending (LAD) coronary artery with a 6‐0 prolene suture, while
   sham‐operated mice underwent the same procedure without LAD occlusion.
   After the indicated time, the mice were euthanized by the same
   anesthetics to obtain their heart samples for corresponding analyses.
   Subsequently, heart samples were either flash‐frozen in liquid nitrogen
   and stored at −80 °C for further RNA and protein analysis or fixed in a
   PBS solution containing 4% paraformaldehyde for histochemical analysis.

Echocardiography

   The assessment of cardiac functions was conducted utilizing the Vevo
   2100 High‐Resolution Micro‐Ultrasound System (FUJIFILM Visual Sonics
   Inc.). Mice were subjected to anesthesia via a combination of
   isoflurane (1.5%) and oxygen (0.5 L min^−1) and were placed in a supine
   position. B‐mode echocardiography was employed to guide the acquisition
   of left ventricular M‐mode images at the papillary muscle level in a
   long‐axis orientation. This approach enabled the quantification of key
   parameters, including ejection fraction (EF%) and fractional shortening
   (FS%). To ensure unbiased results, these measurements were performed by
   an investigator blinded to the experimental conditions and were
   averaged across five consecutive cardiac cycles.

Construction of Adenovirus, Adeno‐Associated Virus, and Plasmid

   Adenovirus (Adv) was packaged by OBiO (Shanghai, China).
   Adeno‐associated virus (AAV) was packaged by WZ Biosciences Inc.
   (Shandong, China). AAV‐RFP‐GFP‐KDEL and Adv‐mCherry‐GFP‐FAM134B were
   used to detect ER‐phagy flux. Adv‐AMFR and Adv‐FAM134B were used to
   manipulate the expression of AMFR and FAM134B in cultured fibroblasts
   and in mouse hearts, respectively. The heart‐specific Tcf21 promoter
   was utilized in AAV9 construction to drive AMFR overexpression
   specifically in fibroblasts. Mouse expression plasmids encoding UB or
   the corresponding UB mutant were produced and purchased from Miaoling
   (Wuhan, China).

Adenovirus and Adeno‐Associated Virus Mediated Gene Transfection In Vivo

   To establish a heart‐overexpressed FAM134B animal model, the cardiac
   adenoviral gene transfer approach was employed. The adenoviral in vivo
   delivery method was conducted as previously described.^[ [216]^70 ,
   [217]^71 ^] In brief, a 20 µL solution of adenovirus containing 1 ×
   10^9 plaque‐forming units (pfu) of either Adv‐Flag or Adv‐FAM134B was
   delivered from the apex of the left ventricle to the aortic root. A
   tourniquet was placed around the aorta and the pulmonary artery at a
   site distal to the tip of the catheter, followed by the injection of
   the solution. The tourniquet was held in place for 10 s while the heart
   pumped against a closed system and then released. To examine the
   transduction efficiency, the hearts were harvested and sectioned two
   weeks after transduction, and FAM134B expression was detected by
   Western blots. Hearts were transduced with adenovirus, followed by MI
   surgery, and were harvested 2 weeks after surgery.

   To establish a fibroblasts‐specific overexpressed FAM134B animal model,
   AAV9.Tcf21. Flag or AAV9. Tcf21.AMFR (3 × 10^11 v.g./mouse) was
   administered through tail vein injection to eight‐week‐old male
   wild‐type mice one week after the surgical procedures. The hearts were
   harvested and sectioned three weeks after transduction.

Cell Culture and Gene Transfer

   Neonatal rat ventricular fibroblasts (NRCFs) were isolated from 1 to
   3‐day‐old neonatal Sprague‐Dawley rats following established
   procedures.^[ [218]^32 ^] The cells were plated at a density of 10^6
   cells mL^−1 and cultured in Dulbecco's Modified Eagle Medium (DMEM,
   Gibco, USA) supplemented with 10% Fetal Bovine Serum (FBS, Meilunbio,
   China), and 100 units mL^−1 penicillin and streptomycin (Beyotime
   Biotechnology, China). The third generation of fibroblasts was placed
   in serum‐free medium for 30 min and subsequently transfected with
   Adv‐mCherry‐GFP‐FAM134B (MOI = 10), Adv‐Flag (MOI = 50), or
   Adv‐Flag‐AMFR (MOI = 50) for 6 h. After transfection, the medium
   containing the adenovirus was replaced with DMEM supplemented with 10%
   FBS for 24 or 48 h. To inhibit protein synthesis, NRCFs were treated
   with cycloheximide (CHX, 100 µg mL^−1) for the indicated time.
   Chloroquine (CQ, 10 µM) was used to inhibit lysosomes from degrading
   the contents of autophagosomes. and proteasome inhibitor MG132(10 µM)
   was used to inhibit proteasome activity.

   Human embryonic kidney 293T (HEK293T) cell line was used in
   co‐immunoprecipitation assays to detect protein interaction between UB
   and FAM134B.

siRNA Transfection

   The siRNA targeting FAM134B (Sense Strand: 5′‐GCUCUUUGUCCUAAGAUUATT‐3′,
   Antisense Strand: 5′‐UAAUCUUAGGACAAAGAGCTT‐3′), siRNA targeting AMFR
   (Sense Strand: 5′‐GGCAACAUCUGGUUAUCUATT‐3′, Antisense Strand:
   5′‐UAGAUAACCAGAUGUUGCCTT‐3′), and a control siRNA (Sense Strand: 5′‐
   UUCUCCGAACGUGUCACGUTT‐3′, Antisense Strand:
   5′‐ACGUGACACGUUCGGAGAATT‐3′) were obtained from OBiO (China). siRNAs
   were transfected into cells using RNAi Reagent (CALNP, China).

Single‐Cell RNA‐seq

   The heart tissues were conserved in the GEXSCOPE Tissue Preservation
   Solution (Singleron) and shipped to the Singleron lab with an ice pack.
   The specimens were washed 3 times with Hanks Balanced Salt Solution
   (HBSS, Gibco, Cat. No.14025‐076) and shredded into 1–2 mm pieces. Then
   the tissue debris was submitted to digestion with 2 ml GEXSCOPE Tissue
   Dissociation Solution (Singleron) at 37 °C for 15 min in a 15 ml
   centrifuge tube (Falcon, Cat. No.352095)with sustained agitation. Cells
   were filtered through 40‐micron sterile strainers(Falcon, Cat.
   No.352340) and centrifuged (eppendorf, 5810R) at 300 g for 5 min. Then
   the supernatant was removed, and the pellets were resuspended in 1 ml
   PBS (Hyclone, Cat. No.SA30256.01). To remove the red blood cells, which
   were frequently a significant portion of the cells produced, 2 mL RBC
   lysis buffer (Roche, Cat. No. 11 814 389 001) was added to the cell
   suspension according to the manufacturer's protocol. Centrifuge the
   cells at 500 × g for 5 min in a microfuge at 15–25 °C and resuspend in
   PBS (Hyclone, Cat. No.SA30256.01). The sample from the cell mixture was
   stained with trypan blue (Bio‐RAD, Cat. No.#1450013) and
   microscopically(Nikon, ECLIPSE Ts2) cell count was performed to make
   the concentration 1 × 105 cells mL^−1, and once the cell viability
   exceeded 80%, subsequent sample processing could be performed.

   Single‐cell suspensions were then loaded onto microfluidic devices, and
   scRNA‐seq libraries were constructed according to the Singleron
   GEXSCOPE protocol by GEXSCOPE Single‐Cell RNA Library Kit (Singleron
   Biotechnologies), which included cell lysis, mRNA trapping, labeling
   cells (barcode) and mRNA (UMI), reverse transcription of mRNA into
   cDNA, amplification, and finally fragmenting cDNA. Individual libraries
   were diluted to 4 nM and pooled for sequencing. Pools were sequenced on
   an Illumina HiSeq X with 150 bp paired end reads. Raw reads were
   processed with fastQC and fastp to remove low‐quality reads. Poly‐A
   tails and adaptor sequences were removed by cutadapt. After quality
   control, reads were mapped to the reference genome GRCm38 (ensembl
   version 99 annotation) using STAR. Gene counts and UMI counts were
   acquired by feature Counts software. Expression matrix files for
   subsequent analyses were generated based on gene counts and UMI counts.

Co‐Immunoprecipitation

   The cells were harvested and lysed following previously established
   protocols.^[ [219]^72 ^] Subsequently, the lysates were incubated
   overnight at 4 °C with either anti‐AMFR (#16675‐1‐AP, Proteintech,
   1:100) or anti‐FAM134B (#21537‐1‐AP, Proteintech, 1:100) antibodies.
   Following this, 40 µl of protein G Agarose beads (Cytiva) were
   co‐incubated with the immune complexes for 4 h at 4 °C. After
   undergoing three washes with a cold wash buffer and one wash with lysis
   buffer, the immunoprecipitations were resuspended in 40 µl of lysis
   buffer and subjected to Western blot analysis to examine the indicated
   target proteins.

Western Blot

   Protein extraction was conducted from mouse heart tissues or cultured
   cells, respectively, followed by immunoblots as previously outlined. In
   brief, Western blot analyses employed commercially available
   antibodies: anti‐AMFR ((#16675‐1‐AP, Proteintech, 1:1000), anti‐FAM134B
   (#21537‐1‐AP, Proteintech, 1:1000), anti‐CCPG1 (#500 520, Huabio,
   1:1000), anti‐SEC62(#A18589, Abclonal, 1:1000), anti‐α‐Tubulin
   (#HRP‐66031, Proteintech, 1:5000), anti‐GAPDH ((#HRP‐60004,
   Proteintech, 1:10 000), anti‐α‐SMA (# A17910, Abclonal, 1:1000),
   anti‐COL1A1 (#72 026, Cell Signaling Technology, 1:1000), anti‐COL3A1
   (#A0817, Abclonal, 1:1000), anti‐CALNEXIN (#A24433, Abclonal, 1:1000),
   anti‐LC3 (#A11280, Abclonal, 1:1000), anti‐UB (#AF1705, Beyotime,
   1:1000), anti‐FlAG (#AE005, Abclonal, 1:2000), anti‐MYC (#AE070,
   Abclonal, 1:10 000), anti‐p‐mTOR (#AP‐0115, Abclonal, 1:1000),
   anti‐mTOR (#A2445, Abclonal, 1:2000), p‐S6K (#AP0502, Abclonal,
   1:1000), S6K (#A2190, Abclonal, 1:1000), p‐4EBP1 (#AF5806, Beyotime,
   1:1000), and 4EBP1 (#AG1824, Beyotime, 1:1000), followed by incubation
   with peroxidase‐conjugated secondary antibodies. Signal detection
   utilized the enhanced chemiluminescent (ECL) system (Tanon, China), and
   densitometry scanning was performed using Image J software. α‐Tubulin
   and GAPDH were used as the loading control. Results from each
   experimental group were expressed as relative integrated intensity
   compared to the control group measured concurrently.

Real‐Time PCR

   Total RNA was isolated from the heart tissues or cultured NRCFs using
   an RNA isolation kit (#R701, Vazyme, China), and then converted using
   HiScript II Q RT SuperMix (#R223, Vazyme, China). Quantitative reverse
   transcription polymerase chain reaction (RT‐qPCR) reactions were
   performed using AceQ Universal SYBR qPCR Master Mix (#Q511, Vazyme,
   China) via LightCycler 96 Instrument (Roche, USA). HPRT was used as the
   internal reference for each sample. The primer sequences are shown in
   Table [220]S3 (Supporting Information).

Cardiac Fibroblasts Proliferation Assay

   The proliferation of NRCFs was assessed through
   5‐ethynyl‐2‐deoxyuridine (EdU) incorporation (Servicebio, G1602),
   following established procedures. The proliferative rate of NRCFs was
   quantified by normalizing the number of EdU‐positive cells to the total
   DAPI‐stained cells.

Wound Healing Assay

   For the wound healing assay, the cardiac fibroblasts transfected with
   Adv‐Flag or Adv‐AMFR were plated in 6‐well plates. Subsequently, the
   confluent monolayer cells were subjected to a linear wound area using a
   200 µL pipette tip in order to assess the migratory capacity of cardiac
   fibroblasts. The scratched plate was then gently washed to remove
   detached cells, followed by treatment with TGF‐β1 or not for 24 h.
   Images of cells were captured at the time of injury and then again 24 h
   post‐injury. The migratory capacity of fibroblasts was assessed by
   quantifying the wound healing area using ImageJ software.

Preparation of Cell Samples for Transmission Electron Microscopy

   NRCFs were treated with TGF‐β1 and adenovirus‐mediated AMFR
   overexpression. Following treatment, cells were harvested from the
   culture dish into a centrifuge tube and centrifuged using standard cell
   passage procedures. The cells were then resuspended in PBS, centrifuged
   again, and fixed overnight at 4 °C with 2.5% glutaraldehyde. After
   discarding the glutaraldehyde solution, the cells were washed with PBS
   and fixed for 1–1.5 h at 4 °C with 1% post‐fixative. Subsequent PBS
   washes were performed to remove the fixative. Dehydration was carried
   out sequentially with 50%, 70%, and 90% ethanol, followed by a 1:1
   mixture of 90% ethanol and 90% acetone, and finally 90% acetone. The
   cells were then treated three times with 100% acetone at room
   temperature for 8 min each. Resin infiltration was performed at room
   temperature with a 1:1 mixture of acetone and resin for 1 h, followed
   by a 1:2 mixture for 1 h, and finally a 1:3 mixture overnight. The
   resin was replaced after 3 h, and this step was repeated once.
   Polymerization was carried out in a 60 °C oven for 48 h. Ultrathin
   sections were prepared from relatively typical areas of cell
   ultrastructure using an ultramicrotome and stained with sodium osmium.

Immunofluorescence

   The heart tissue chips underwent dewaxing and rehydration, while cells
   were fixed using 4% paraformaldehyde. Following these steps, the washed
   heart tissue chips or cells were permeabilized with 0.1% Triton X‐100
   and then blocked with 5% normal goat serum (Servicebio, China).
   Subsequently, they were stained with anti‐Vimentin (#5741, Cell
   Signaling Technology, 1:200), anti‐cTnT (#ab8295, Abcam, 1:200),
   anti‐CD31(#ab281583, Abcam, 1:20), anti‐CD68 (#A6554, Abclonal, 1:100),
   anti‐AMFR ((#16675‐1‐AP, Proteintech, 1:200), anti‐FAM134B (#HA721752,
   Huabio, 1:100), anti‐α‐SMA (# A17910, Abclonal, 1:100), anti‐LC3
   (#ET1701‐65, Huabio, 1:500), as per the instructions. Afterward, they
   were incubated with Alexa Fluor 488, 594, or 647‐conjugated secondary
   antibodies (1:1000) (Invitrogen, USA) at room temperature for 60 min,
   followed by DAPI (Beyotime, China) staining. For imaging, five fields
   were randomly selected in each group and captured using a confocal
   microscope (NIKON, Japan).

Histological Examination

   Mouse hearts were collected and then immersion‐fixed in 4% buffered
   paraformaldehyde. Subsequently, the heart tissues were embedded in
   paraffin and cut into 5 µm sections. In brief, the heart tissue chips
   were counterstained with hematoxylin and eosin (H&E) and incubated with
   the Masson's Trichrome Stain Kit (Solarbio, China).

Statistics

   The data are presented as means ± SD. Statistical analyses were
   performed using GraphPad Prism 9.5.1. The normality of the data was
   estimated using the Shapiro–Wilk test. For comparisons between two
   groups, an unpaired t‐test was conducted. When comparing more than two
   groups, one or two‐way analysis of variance (ANOVA) was employed,
   followed by the post hoc Tukey's test, when the assumptions of normal
   distribution were met. Otherwise, the nonparametric Kruskal–Wallis test
   was employed for further analysis. In all cases, p< 0.05 was considered
   statistically significant.

Ethics Approval and Consent to Participate

   All animal experiments were approved by the Ethics Committee of
   Shanghai Jiao Tong University Affiliated Sixth People's Hospital, in
   accordance with the Guide for the Care and Use of Laboratory Animals
   established by the US National Institutes of Health (NIH Publication,
   8th Edition, 2011).

Conflict of Interest

   The authors declare that they have no competing interests.

Author Contributions

   Z.W., K.N., and W.L. contributed equally to this work. X.J. was
   involved in the conceptualization of the study. X.J. designed and
   implemented the methodology. Z.X.W., K.F.N., W.L., and X.Y.W.
   contributed to the investigations. Z.X.W., K.F.N., W.L., X.Y.W.,
   B.S.Y., T.X.L., Y.Z.C., and Y.Y.J. performed the visualization of the
   data. X.J. and Y.Y.L. acquired funding for the studies. X.J. and Y.Y.L.
   administered the project. Y.Y.L., Y.C., B.S.Y., T.X.L., Y.Z.C., and
   Y.Y.J. provided supervision for the study. X.J., Y.Y.L., Y.C., Z.X.W.,
   K.F.N., W.L., X.Y.W., B.S.Y., T.X.L., Y.Z.C., and Y.Y.J. were involved
   in the review and editing of the manuscript.

Supporting information

   Supporting Information
   [221]ADVS-12-e04552-s001.docx^ (2.1MB, docx)

   Supporting Information
   [222]ADVS-12-e04552-s002.docx^ (4.7MB, docx)

Acknowledgements