Abstract

   Metabolic reprogramming is critical in the onset of pressure
   overload–induced cardiac remodeling. Our study reveals that proline
   dehydrogenase (PRODH), the key enzyme in proline metabolism, reprograms
   cardiomyocyte metabolism to protect against cardiac remodeling. We
   induced cardiac remodeling using transverse aortic constriction (TAC)
   in both cardiac-specific PRODH knockout and overexpression mice. Our
   results indicate that PRODH expression is suppressed after TAC.
   Cardiac-specific PRODH knockout mice exhibited worsened cardiac
   dysfunction, while mice with PRODH overexpression demonstrated a
   protective effect. In addition, we simulated cardiomyocyte hypertrophy
   in vitro using neonatal rat ventricular myocytes treated with
   phenylephrine. Through RNA sequencing, metabolomics, and metabolic flux
   analysis, we elucidated that PRODH overexpression in cardiomyocytes
   redirects proline catabolism to replenish tricarboxylic acid cycle
   intermediates, enhance energy production, and restore glutathione redox
   balance. Our findings suggest PRODH as a modulator of cardiac
   bioenergetics and redox homeostasis during cardiac remodeling induced
   by pressure overload. This highlights the potential of PRODH as a
   therapeutic target for cardiac remodeling.
     __________________________________________________________________

   Promoting proline metabolism can protect cardiac remodeling by
   restoring energy and reducing oxidative stress.

INTRODUCTION

   Hypertension, the most prevalent cardiovascular disorder worldwide,
   profoundly threatens cardiac health without proper management ([44]1).
   In the early stages of elevated cardiac afterload caused by sustained
   hypertension, the heart undergoes myocardial hypertrophy, a
   pathological condition characterized by abnormal enlargement of
   cardiomyocytes. As the disease advances, compensatory cardiac
   hypertrophy is followed by disrupted myocardial structure, impeded
   blood supply, and compromised cardiac function, ultimately culminating
   in heart failure (HF) or cardiac death ([45]2–[46]4). Presently,
   therapeutic options for addressing maladaptive cardiac remodeling
   remain restricted, with a conspicuous absence of modalities capable of
   either delaying or reversing such pathological processes ([47]5).
   Therefore, comprehensive studies to uncover the pathological mechanism
   of cardiac hypertrophy and identify appropriate targets to delay or
   reverse cardiac hypertrophy might pave the way to prevent HF.

   The heart is one of the most metabolically active organs in the body.
   Before the onset of structural cardiac remodeling, the heart undergoes
   metabolic adaptation to meet heightened energy requisites during the
   cardiac hypertrophy. The metabolic remodeling of the heart can assist
   cardiomyocytes in coping with relative energy insufficiency and
   imbalanced redox homeostasis ([48]6–[49]8). Metabolic remodeling
   precedes cardiac dysfunction. Regulating the metabolic remodeling of
   cardiomyocytes can effectively prevent pathological progression in the
   early stages of remodeling ([50]9). Previous studies have demonstrated
   that reprogramming cardiomyocyte metabolism can preserve cardiac
   function during pressure overload–induced cardiac remodeling and
   prevent HF ([51]10–[52]12). These investigations primarily focused on
   regulating glycolysis and fatty acid oxidation of cardiomyocytes to
   achieve the alleviation of cardiac remodeling. In sharp contrast, there
   is relatively limited research on elucidating the role of amino acid
   metabolism in cardiac remodeling. Nevertheless, the precise metabolic
   alterations and underlying mechanisms of amino acid metabolism during
   cardiac remodeling remain largely elusive.

   Proline, a non-essential amino acid, not only serves as a fundamental
   element for protein synthesis but also participates in diverse cellular
   processes critical for cell survival and adaptation ([53]13). Emerging
   studies have elucidated its multifaceted functions, including its
   involvement in redox balance regulation and energy metabolism
   modulation ([54]14–[55]16). Hence, alterations in proline metabolism
   might have been implicated in the pathogenesis of various diseases.
   Unraveling its intricate network of interactions within cellular
   metabolism holds promise for therapeutic interventions targeting a wide
   array of pathological conditions. Proline dehydrogenase (PRODH), the
   first enzyme in proline catabolism, has been reported to be the most
   crucial enzyme in regulating proline metabolism ([56]17). PRODH
   catalyzes the conversion of proline to pyrroline-5-carboxylate, further
   converting to glutamate or ornithine ([57]18). Previous studies have
   noticed the changes of proline metabolism during myocardial remodeling
   ([58]9, [59]19). However, detailed studies are needed to investigate
   how the alternation of proline metabolism links to cardiac performance.

   In the present study, we primarily investigated the role of PRODH in
   pressure overload–induced cardiac remodeling. We aimed to elucidate the
   crucial role of PRODH in maintaining cardiac function in transverse
   aortic constriction (TAC)–induced pathological cardiac remodeling using
   a genetic approach. Furthermore, we confirmed that cardiac specific
   overexpression of PRODH exerted a protective effect on cardiac function
   and inhibited adverse myocardial remodeling. Mechanistically, we
   demonstrated that promoting proline catabolism can restore energy and
   redox balance in cardiomyocytes. These findings collectively
   demonstrate the critical regulatory role of PRODH in cardiac
   remodeling. Our results provide insights into the metabolic mechanisms
   of HF and propose a potential target for future therapeutic
   interventions.

RESULTS

The expression of PRODH was decreased in the hypertrophic hearts

   We started by assessing the relevance of PRODH in pathological cardiac
   remodeling. To achieve this, we first examined the expression levels of
   PRODH in HF and hypertrophic cardiomyopathy (HCM) by using two public
   RNA sequencing (RNA-seq) datasets: [60]GSE141910 and [61]GSE133054. We
   found that the expression of PRODH was markedly down-regulated in
   hearts of both HF and HCM ([62]Fig. 1, A and B). Following this, we
   sought to explore the changes in PRODH expression during TAC-induced
   cardiac remodeling in mice. We conducted a time course single-cell
   RNA-seq analysis on mouse hearts (0, 3 days, 1 week, and 4 weeks)
   following TAC-induced pressure overload using the public dataset
   [63]GSE95143 ([64]Fig. 1C). Consistent with our previous findings,
   PRODH expression was markedly reduced on 3 days, 1 week, and 4 weeks
   after TAC ([65]Fig. 1D). Furthermore, we established a murine model of
   cardiac hypertrophy by performing TAC surgery. After 4 weeks, we
   harvested the heart tissue and found that the expression levels of
   PRODH were substantially down-regulated at both mRNA and protein levels
   in the hearts of TAC mice compared to sham ([66]Fig. 1, E to G). In
   line with this, PRODH activity was reduced in the hearts upon TAC
   ([67]Fig. 1H). Previous studies has indicated that TAC surgery
   activates the mammalian target of rapamycin (mTOR) pathway ([68]20,
   [69]21), which is known to negatively regulate PRODH expression
   ([70]15, [71]17). To elucidate the mechanisms underlying the decreased
   expression of PRODH following TAC surgery, we collected cardiac tissue
   samples from mice treated with 4 weeks of rapamycin or saline after TAC
   surgery. We observed a notable increase in PRODH expression in the
   group treated with rapamycin, suggesting a link between mTOR activation
   and PRODH expression following TAC surgery (fig. S1, A to D). Next, to
   mimic in vitro cardiomyocyte remodeling, we treated neonatal rat
   ventricular myocytes (NRVMs) with pro-hypertrophic stimuli,
   phenylephrine (PE). Consistent with our in vivo results, both mRNA and
   protein expression were notably decreased in PE-induced hypertrophic
   cardiomyocytes ([72]Fig. 1, I to K). Overall, these data demonstrated
   that both the expression level and the activity of PRODH were reduced
   in hypertrophic cardiomyocytes.

Fig. 1. PRODH expression was decreased in hypertrophic hearts.

   [73]Fig. 1.
   [74]Open in a new tab

   (A) The mRNA expression of PRODH in heart tissues from patients with HF
   and normal cardiac myocytes (NCM) in the RNA-seq database [75]GSE133054
   (n = 7 for HF; n = 8 for NCM). (B) The mRNA expression of PRODH in
   heart tissues from patients with HCM and NCM by combining the RNA-seq
   database [76]GSE141910 and [77]GSE133054 (n = 174 for HCM; n = 36 for
   NCM). (C and D) The clustering diagram and relative PRODH expression
   after TAC acquired by single-cardiomyocyte RNA-seq of dataset
   [78]GSE95143. (E and F) Representative images and quantification of
   PRODH, β-MHC, and ANP protein levels measured by Western blotting in
   heart tissues from sham and TAC-operated mice (n = 6). GAPDH,
   glyceraldehyde-3-phosphate dehydrogenase. (G and H) Relative mRNA and
   enzyme activity of PRODH in heart tissues from sham and TAC-operated
   mice (n = 6). (I and J) Representative images and quantification of
   PRODH, β-MHC, and ANP protein levels measured by Western blotting in
   isolated NRVMs stimulated with PE (100 μM) or vehicle (Veh) for 48
   hours (n = 6). (K) Quantification of mRNA expression of PRODH, ANP, and
   BNP by reverse transcription–quantitative polymerase chain reaction
   (RT-qPCR) in isolated NRVMs stimulated with PE (100 μM) or vehicle for
   48 hours (n = 3). (Data are shown as means ± SEM. Unpaired two-tailed
   Student’s t test was conducted for the comparison.)

PRODH deficiency aggravated the adverse cardiac remodeling induced by TAC in
mice

   To investigate the role of PRODH in pressure overload–induced cardiac
   remodeling, cardiac-specific PRODH conditional knockout (cKO) mice were
   generated by intercrossing Prodh floxed mice with the α-myosin heavy
   chain promoter-driven Cre transgenic line. Subsequently, we performed
   TAC surgery on cKO mice and conducted echocardiographic analysis 4
   weeks after surgery, and their littermates were used as controls (f/f).
   Cardiac-specific PRODH knockout mice markedly exacerbated TAC-induced
   cardiac dysfunction and show reduced ejection fraction (EF) and
   fractional shortening (FS) compared to f/f mice ([79]Fig. 2, A to C),
   while no differences were observed in the survival rates between two
   strains (fig. S2A). In addition, cKO mice remarkably increased the left
   ventricular internal diameter at end-diastole (LVIDd) and left
   ventricular internal diameter at end-systole (LVIDs), suggesting left
   ventricular chamber dilation in cKO mice upon TAC ([80]Fig. 2, D and
   E). Notably, the heart function and left ventricular diameters were
   comparable between f/f and cKO mice subjected to sham operations,
   indicating that cardiac-specific knocking out PRODH did not show any
   structural or function deficiencies at baseline. As anticipated, cKO
   mice showed higher heart weight/body weight (HW/BW) and heart
   weight/tibia length (HW/TL) ratios than f/f mice ([81]Fig. 2, F and G),
   indicating aggravated cardiac remodeling. Next, we performed histology
   analysis on the mice’s hearts upon TAC or sham. Hematoxylin and eosin
   (H&E) staining showed that cKO mouse hearts were larger and had a
   greater surface area ([82]Fig. 2H). Masson’s staining showed that cKO
   mice had considerably higher levels of cardiac fibrosis than f/f mice
   after TAC ([83]Fig. 2I). Wheat germ agglutinin (WGA) staining revealed
   that the cross-sectional area of cKO mouse hearts was larger than that
   of f/f mice ([84]Fig. 2J). Dihydroethidium (DHE) staining indicated
   that PRODH deficiency exacerbated oxidative damage in cardiomyocytes
   ([85]Fig. 2K). Moreover, we isolated the adult mouse cardiac myocytes
   (AMCMs) and measured the reduced/oxidized glutathione (GSH/GSSG) ratio.
   Of note, cKO mice exhibited a more pronounced decrease in GSH/GSSG
   after TAC surgery compared to f/f mice ([86]Fig. 2L), indicating a more
   severe impairment of antioxidant capacity in cKO mice upon TAC.
   Besides, the mRNA expression levels of the hypertrophic markers atrial
   natriuretic peptide (ANP), B-type natriuretic peptide (BNP), and
   β-myosin heavy chain (β-MHC), as well as the fibrotic markers collagen
   type 1a (Col1a), collagen type 3a (Col3a), and transforming growth
   factor–β (TGFβ), were all increased in the hearts of cKO mice when
   compared with those in the f/f mice after TAC ([87]Fig. 2, M and N).
   Consistently, the protein levels of β-MHC and ANP were also
   up-regulated in the cKO hearts ([88]Fig. 2, O to R).

Fig. 2. Cardiac-specific deletion of PRODH aggravated adverse cardiac
remodeling progression under pressure overload.

   [89]Fig. 2.
   [90]Open in a new tab

   (A to E) Representative M-mode echocardiography images and
   quantification of EF and FS, LVIDd, and LVIDs among the f/f and cKO
   mice after either sham or TAC (n = 8 for sham-operated mice; n = 10 for
   TAC-operated mice). (F and G) The ratios of HW/BW and HW/TL (n = 8 for
   sham-operated mice; n = 10 for TAC-operated mice). (H to K)
   Representative images of H&E staining, Masson’s trichrome staining, WGA
   staining, and DHE and their quantitation results (n = 5 for each group
   of Masson’s trichrome staining; n = 256 cells for each group of the WGA
   staining; n = 4 for DHE staining). (L) Relative GSH/GSSG ratio of AMCMs
   isolated from the in vivo experiment (n = 4). (M and N) Relative mRNA
   expression of hypertrophic markers (ANP, BNP, and β-MHC) and fibrosis
   markers (Col1a, Col3a, and TGFβ) quantified by RT-qPCR (n = 6). (O to
   R) Representative images and quantification of PRODH, β-MHC, and ANP
   protein levels measured by Western blotting (n = 6). [Data are shown as
   means ± SEM. Two-way analysis of variance (ANOVA) was conducted for the
   comparisons.]

   Given the elevated oxidative stress levels observed in cKO mice
   compared with f/f mice after TAC, we explored whether antioxidants
   could ameliorate cardiac functional deficits in cKO mice induced by TAC
   operation. Thus, we administered N-acetylcysteine (NAC) to cKO mice and
   assessed mouse cardiac function via echocardiography 4 weeks after
   surgery (fig. S3A). The results indicated that NAC treatment
   considerably enhanced cardiac function in cKO mice after TAC, as
   evidenced by improved EF and FS (fig. S3, B and C). In addition,
   reduced LVIDd and LVIDs were observed in the cKO mice NAC-treated
   group, suggesting improved cardiac function (fig. S3, D and E).
   Consistently, cardiac hypertrophy indices, including HW/BW and HW/TL in
   cKO mice, were noticeably decreased following NAC administration (fig.
   S3, F and G). These results showed that antioxidant therapy with NAC
   partially reversed the impaired cardiac function in PRODH cKO mice.
   Collectively, cardiac PRODH deficiency accelerates adverse cardiac
   remodeling induced by pressure overload, with increased oxidative
   stress being a contributing factor.

Cardiac-specific overexpression of PRODH protected the heart from pressure
overload–induced cardiac remodeling

   To further explore the potential protective effects of cardiac-specific
   PRODH overexpression against TAC-induced cardiac remodeling, we used an
   adeno-associated virus 9 (AAV9) vector with a cardiac troponin T
   promoter to specific overexpress PRODH in cardiomyocytes. In this mouse
   model featuring cardiac-specific PRODH overexpression (AAV9-PRODH), we
   performed TAC surgery and assessed cardiac function via
   echocardiography 4 weeks later. AAV9-PRODH mice showed preserved EF and
   FS upon TAC when compared to the AAV9-Vector injected mice as a control
   ([91]Fig. 3, A to C). There were no substantive differences in the
   survival rates after TAC surgery between the AAV9-Vector mice and
   AAV9-PRODH mice (fig. S2B). In addition, while the control group
   exhibited increased LVIDd and LVIDs, indicative of ventricular
   dilation, these indicators were markedly reduced in the AAV9-PRODH mice
   ([92]Fig. 3, D and E). Consistently, it was observed that the
   AAV9-PRODH mice exhibited decreased HW/BW and HW/TL ratios compared to
   the control group, indicating a reduction in cardiac hypertrophy
   ([93]Fig. 3, F and G). Furthermore, Masson’s staining showed that PRODH
   overexpression attenuated fibrosis induced by TAC operation, as
   demonstrated by Masson’s staining ([94]Fig. 3, H and I). In addition,
   WGA staining indicated a considerable reduction in the surface area of
   cardiomyocytes in TAC mice following PRODH overexpression ([95]Fig.
   3J), accompanied by a decrease in oxidative stress damage as evidenced
   by DHE staining ([96]Fig. 3K). The GSH/GSSG ratio in isolated AMCMs
   from AAV9-PRODH mice was markedly enhanced compared to that of
   AAV9-Vector mice 4 weeks after TAC, indicating an improved antioxidant
   capacity of cardiomyocytes ([97]Fig. 3L). To elucidate how PRODH
   overexpression affected the GSH/GSSG ratio, we examined the expression
   levels of enzymes involved in GSH metabolism within AMCMs. We observed
   a marked increase in glutathione reductase (GSR) expression after TAC,
   which was even more pronounced in AAV9-PRODH mice (fig. S4, A and B).
   Meanwhile, glutamate-cysteine ligase catalytic subunit (GCLC)
   expression remained unchanged in the TAC groups following PRODH
   overexpression (fig. S4C). These findings suggested that PRODH
   overexpression enhanced the GSH/GSSG ratio by modulating GSR activity,
   thereby promoting the reduction of GSSG to GSH in cardiomyocytes.
   Subsequently, we quantified mRNA levels of hypertrophic and fibrotic
   markers. The hypertrophic markers ANP, BNP, and β-MHC alongside
   fibrotic markers Col1a, Col3a, and TGFβ were markedly elevated in the
   AAV9-Vector group after TAC but were notably suppressed in the
   AAV9-PRODH group ([98]Fig. 3, M and N). Western blotting analysis
   further substantiated that PRODH overexpression in cardiomyocytes
   reduced the protein levels of hypertrophic markers following TAC
   ([99]Fig. 3, O to R). Next, we assessed mitochondrial morphology within
   cardiac tissue sections. Mice subjected to sham surgery displayed
   well-preserved mitochondrial architecture with intact cristae.
   Conversely, mice treated with AAV9-Vector following TAC surgery
   exhibited predominantly globular and swollen mitochondria in their
   cardiac tissues. Notably, AAV9-PRODH mice maintained intact
   mitochondrial morphology after TAC surgery, suggesting a protective
   effect of PRODH overexpression on mitochondrial integrity (fig. S5A).

Fig. 3. Cardiac-specific overexpression of PRODH attenuated cardiac
remodeling under pressure overload.

   [100]Fig. 3.
   [101]Open in a new tab

   (A to E) Representative M-mode echocardiography images and
   quantification of EF and FS, LVIDd, and LVIDs among the AAV9-Vector and
   AAV9-PRODH mice after either sham or TAC (n = 7). (F and G) The ratios
   of HW/BW and HW/TL (n = 7). (H to K) Representative images of H&E
   staining, Masson’s trichrome staining, WGA staining, and DHE and their
   quantitation results (n = 5 for each group of Masson’s trichrome
   staining; n = 256 cells for each group of the WGA staining; n = 4 for
   DHE staining). (L) Relative GSH/GSSG ratio of AMCMs isolated from the
   in vivo experiment (n = 4). (M and N) Relative mRNA expression of
   hypertrophic markers (ANP, BNP, and β-MHC) and fibrosis markers (Col1a,
   Col3a, and TGF) quantified by RT-qPCR (n = 6). (O to R) Representative
   images and quantification of PRODH, β-MHC, and ANP protein levels
   measured by Western blotting (n = 6). (Data are shown as means ± SEM.
   Two-way ANOVA was conducted for the comparisons.)

   To evaluate whether cardiac-specific PRODH overexpression could reverse
   established pathological cardiac hypertrophy, we administered
   AAV9-Vector or AAV9-PRODH 2 weeks after TAC (fig. S6A). Cardiac
   function, assessed by echocardiography 6 weeks after TAC, revealed that
   cardiac-specific PRODH overexpression led to improved EF and FS (fig.
   S6, B to D), along with a reduction in LVIDd and LVIDs (fig. S6, E and
   F). These findings indicated enhanced cardiac performance and reduced
   ventricular dilation. Consistently, reduced HW/BW and HW/TL ratios in
   the AAV9-PRODH group suggested a mitigation of ongoing cardiac
   remodeling (fig. S6, G and H). Collectively, these findings strengthen
   the protective function of cardiomyocyte PRODH in cardiac remodeling.

PRODH deficiency exacerbated PE-induced cardiomyocyte hypertrophy

   Following the revelation of PRODH’s protective role against TAC-induced
   cardiac remodeling from our in vivo studies, we turned to an in vitro
   model using NRVMs. NRVMs were stimulated with PE to mimic cardiomyocyte
   hypertrophy in this study. PRODH expression was silenced in NRVMs using
   small interfering RNA (siRNA), followed by a 48-hour PE treatment to
   induce hypertrophy. Consistent with our in vivo findings,
   tetramethylrhodamine isothiocyanate (TRITC)–phalloidin staining
   revealed that PRODH deficiency notably exacerbated cardiomyocyte
   hypertrophy in vitro ([102]Fig. 4, A and B). This was in line with a
   marked increase in the mRNA levels of hypertrophic markers ANP, BNP,
   and β-MHC in PRODH deficiency in cardiomyocytes ([103]Fig. 4C).
   Similarly, the protein level of β-MHC and ANP was markedly increased
   upon PRODH loss in cardiomyocytes ([104]Fig. 4, D to G). Moreover, DHE
   staining revealed that PE treatment increased oxidative stress in
   cardiomyocytes, which was intensified by PRODH deficiency ([105]Fig. 4,
   H and I). In addition, the ratio of GSH/GSSG was diminished following
   PE induction and declined further upon PRODH deficiency ([106]Fig. 4J).
   These results collectively suggest that PRODH deficiency in
   cardiomyocytes aggravates cardiomyocyte hypertrophy and oxidative
   stress.

Fig. 4. Silencing PRODH aggravated PE-stimulated cardiomyocyte hypertrophy in
vitro.

   [107]Fig. 4.
   [108]Open in a new tab

   (A and B) Representative images and quantification of cultured NRVMs
   stained with TRITC-labeled phalloidin (red) and nuclei (blue) (n = 66
   for NC, Veh; n = 67 for si-PRODH, Veh; n = 65 for NC, PE; n = 55 for
   si-PRODH, PE). (C) Relative mRNA expression of ANP, BNP, and β-MHC
   quantified by RT-qPCR (n = 6). (D to G) Representative images and
   quantification of PRODH, β-MHC, and ANP protein levels measured by
   Western blotting (n = 6). (H and I) Representative images and
   quantification of cultured NRVMs stained with DHE (red) (n = 5). (J)
   Quantification of GSH/GSSG after silencing PRODH and/or stimulated with
   PE (n = 3). (Data are shown as means ± SEM. Two-way ANOVA was conducted
   for the comparisons.)

PRODH overexpression mitigated PE-induced cardiomyocyte hypertrophy

   Subsequently, we studied the role of PRODH overexpression in NRVMs on
   hypertrophy and oxidative damage in vitro. Therefore, an adenovirus 4
   (ADV4)–mediated PRODH overexpression adenovirus (Ad-PRODH) was added to
   NRVMs before a 48-hour PE stimulation. Notably, TRITC-phalloidin
   staining revealed that PRODH overexpression in NRVMs led to a
   noticeable reduction in the cardiomyocyte surface area after PE
   exposure ([109]Fig. 5, A and B). This reduction was paralleled by a
   marked decrease in mRNA expression of hypertrophic markers ANP, BNP,
   and β-MHC upon PE treatment ([110]Fig. 5C). Consistently, the protein
   levels of β-MHC and ANP were also reduced in PRODH-overexpressed
   cardiomyocytes ([111]Fig. 5, D to G). In alignment with these
   observations, DHE staining, the indicator of oxidative stress of NRVMs,
   showed a notably protective effect of PRODH from oxidative damage after
   PE stimulation ([112]Fig. 5, H and I). The increased GSH/GSSG ratio
   also revealed that PRODH overexpression notably up-regulated the
   antioxidant ability of cardiomyocytes ([113]Fig. 5J). In summary, these
   results reinforce the conclusion that PRODH overexpression in
   cardiomyocytes protects them against hypertrophy and oxidative stress.

Fig. 5. Overexpression of PRODH attenuated PE-stimulated cardiomyocyte
hypertrophy in vitro.

   [114]Fig. 5.
   [115]Open in a new tab

   (A and B) Representative images and quantification of cultured NRVMs
   stained with TRITC-labeled phalloidin (red) and nuclei (blue) (n = 51
   for Ad-GFP, Veh; n = 47 for Ad-PRODH, Veh; n = 46 for Ad-GFP, PE; n =
   52 for Ad-PRODH, PE). (C) Relative mRNA expression of ANP, BNP, and
   β-MHC quantified by RT-qPCR (n = 6). (D to G) The representative images
   and quantification of PRODH, β-MHC, and ANP protein levels measured by
   Western blotting (n = 6). (H and I) Representative images and
   quantification of cultured NRVMs stained with DHE (red) (n = 5). (J)
   Quantification of GSH/GSSG after silencing PRODH and/or stimulated with
   PE (n = 3). (Data are shown as means ± SEM. Two-way ANOVA was conducted
   for the comparisons.)

PRODH overexpression effectively reprogrammed cardiac metabolic shifts
induced by TAC

   To investigate the cardioprotective effects of PRODH overexpression in
   the context of pressure overload–induced cardiac remodeling, we
   simultaneously conducted RNA-seq on isolated AMCMs and cardiac tissue
   of the mice subjected to either TAC or sham surgery with
   cardiac-specific PRODH overexpression, as shown in [116]Fig. 6 and fig.
   S7, respectively, and similar observations were obtained. We first
   conducted differential gene analysis, confirming the decrease in PRODH
   levels in cardiomyocytes and cardiac tissues after TAC surgery and its
   up-regulation following PRODH overexpression ([117]Fig. 6, A and B, and
   fig. S7, A and B). The control group, treated with AAV9-Vector,
   exhibited 3768 down-regulated genes and 3867 up-regulated genes among
   AMCMs after TAC compared with the sham group. In contrast, hearts
   overexpressing PRODH showed 2956 up-regulated genes and 3339
   down-regulated genes among AMCMs following TAC compared to AAV9-Vector
   controls ([118]Fig. 6, C and D). For the results from RNA-seq of
   cardiac tissues, 3055 down-regulated genes and 2297 up-regulated genes
   were observed for the comparison between the TAC group and the sham
   group among the mice receiving AAV9-Vector. Overexpression of PRODH
   showed 3006 up-regulated genes and 2248 down-regulated genes following
   TAC compared to AAV9-Vector controls (fig. S7, C and D). A detailed
   study of these gene expression alterations revealed that PRODH
   overexpression counteracted the TAC-induced down-regulation of 1129
   genes and the up-regulation of 1345 genes among AMCMs and the
   down-regulation of 1093 genes and the up-regulation of 770 genes among
   cardiac tissues. Furthermore, we performed Kyoto Encyclopedia of Genes
   and Genomes (KEGG) enrichment analysis of these counteracted gene by
   overexpression of PRODH. The results indicated that TAC prompted
   notable alterations in cardiac metabolism, including reductions in
   valine, leucine, and isoleucine degradation; carbon metabolism;
   propanoate metabolism; and citrate cycle (TCA cycle) among AMCMs
   ([119]Fig. 6E). Notably, these metabolic perturbations were largely
   reversed by PRODH overexpression. In addition, PRODH overexpression was
   found to mitigate the enhanced expression of several key biological
   processes related to cardiac modeling, such as focal adhesion,
   extracellular matrix (ECM)–receptor interaction, and phagosome
   ([120]Fig. 6F). These results suggested that PRODH had a crucial role
   in reducing structural remodeling upon TAC. Similarly, the results from
   cardiac tissue indicated that overexpression of PRODH reversed the
   inhibited cardiac metabolism induced by TAC, including carbon
   metabolism, TCA cycle, 2-oxocaarboxylic acid metabolism, and valine,
   leucine, and isoleucine degradation (fig. S7E). It also mitigated the
   enhanced expression of several cardiomyopathic pathways after TAC, such
   as dilated cardiomyopathy, HCM, lysosome, and protein digestion and
   absorption (fig. S7F).

Fig. 6. PRODH overexpression effectively reprogrammed cardiac metabolic
shifts induced by TAC.

   [121]Fig. 6.
   [122]Open in a new tab

   (A and B) Volcano plot of two comparisons from the RNA-seq data of
   AMCMs: AAV9-Vector, TAC versus AAV9-Vector, sham and AAV9-PRODH, TAC
   versus AAV9-Vector, TAC. (C and D) Venn diagram of differentially
   expressed genes (DEGs) down-regulated by TAC surgery and up-regulated
   by PRODH overexpression and up-regulated by TAC surgery and
   down-regulated by PRODH overexpression, respectively. (E and F) KEGG
   pathway analysis of the DEGs down-regulated by TAC surgery and rescued
   by overexpression of PRODH and up-regulated by TAC surgery and
   inhibited by overexpression of PRODH, respectively. (G) GSEA of the
   dataset of the citrate cycle (TCA cycle). (H) Heatmap of the expression
   profile of proline catabolism, proline anabolism, and TCA cycle
   acquired by the RNA-seq data (n = 4 mice per group). (I to L) Relative
   proline and TCA intermediate levels of cardiac tissues after TAC
   surgery and PRODH overexpression (n = 4). (Data are shown as means ±
   SEM. Two-way ANOVA was conducted for the comparisons.)

   The replenishment of TCA intermediates in cardiomyocytes is crucial for
   maintaining cardiac function after TAC surgery. Previous studies have
   established that proline can be metabolically converted into glutamate,
   which then deaminates to α-ketoglutarate and enters the TCA cycle
   ([123]16, [124]18). Thus, we further explored the enrichment analysis
   of TCA cycle, a key metabolic process of interest to us, of two
   individual sequencing experiments. The results revealed by gene set
   enrichment analysis (GSEA) confirmed that the TCA cycle was notably
   enriched in AMCMs and cardiac tissues after PRODH overexpression
   ([125]Fig. 6G and fig. S7G). In addition, we constructed a heatmap
   based on two RNA-seq data to present the expression profiles of genes
   in proline catabolism and anabolism, alongside those in the TCA cycle
   ([126]Fig. 6H and fig. S7H). Our findings from two datasets both
   indicated a notable down-regulation of proline catabolism-related
   genes, including Prodh, pyrroline-5-carboxylate dehydrogenase (P5cdh),
   ornithine aminotransferase (Oat), and glutamate dehydrogenase 1 (Glud1)
   following TAC, which was restored upon PRODH overexpression.
   Conversely, genes in proline synthesis, such as pyrroline-5-carboxylate
   reductase (Pycr1) and pyrroline-5-carboxylate synthase (P5cs),
   displayed an up-regulation after TAC, which was reversed by PRODH
   overexpression. Furthermore, we observed that genes in the TCA cycle
   were substantially down-regulated after TAC, but their expression
   levels were substantially restored by PRODH overexpression.

   After elucidating the changes in proline metabolism and the TCA cycle
   through RNA-seq, we further conducted metabolite profiling of cardiac
   tissue after TAC via untargeted metabolomics. The results showed that
   proline levels in the TAC group among the AAV9-Vector mice were
   markedly higher than those in the sham group, accompanied by a decrease
   in levels of TCA metabolic intermediates such as succinate, malate, and
   citrate ([127]Fig. 6, I to L). Conversely, in AAV9-PRODH mice with TAC
   surgery, the proline levels were noticeably lower compared with those
   in AAV9-Vector mice, but the levels of TCA intermediates were
   prominently elevated. This result suggested that overexpression of
   PRODH not only enhanced the utilization of proline among cardiomyocytes
   but also promoted the levels of TCA cycle intermediates. Collectively,
   these findings imply that TAC suppresses proline catabolism and the
   subsequent TCA cycle, while PRODH overexpression in cardiomyocytes
   counteracts these effects and restores normal TCA cycle by reprograming
   proline metabolism.

The metabolic reprogramming induced by PRODH overexpression restored the TCA
cycle in cardiomyocyte hypertrophy

   We primarily compared the metabolic changes induced by PE stimulation
   and the metabolic effects of PRODH overexpression or knockdown on
   PE-induced cardiomyocyte hypertrophy. Orthogonal partial least squares
   discriminant analysis (OPLS-DA) revealed distinct metabolic patterns
   among cells subjected to different treatments ([128]Fig. 7, A and B,
   and fig. S8, A and B). In the PRODH overexpression experiment, PE
   stimulation modulated a subset of 33 metabolites in the
   adenovirus–green fluorescent protein (Ad-GFP) group, with 17
   up-regulated metabolites and 16 down-regulated metabolites from a pool
   of 257 analyzed metabolites, while PRODH overexpression up-regulated 20
   metabolites and down-regulated 16 metabolites compared to the
   PE-stimulated group with negative virus ([129]Fig. 7C). In the PRODH
   knockdown experiment, PE stimulation up-regulated 26 metabolites and
   down-regulated 14 metabolites out of 250 detected metabolites, whereas
   PRODH knockdown resulted in the up-regulation of 11 metabolites and
   down-regulation of 28 metabolites (fig. S8C). Volcano plots depicted
   the most profoundly altered metabolites ([130]Fig. 7, D and E, and fig.
   S8, D and E).

Fig. 7. The metabolic reprogramming induced by PRODH overexpression restored
the TCA cycle in cardiomyocyte hypertrophy.

   [131]Fig. 7.
   [132]Open in a new tab

   (A and B) OPLS-DA of metabolome data for the comparison between the
   Ad-GFP, Veh group and the Ad-GFP, PE group, and between the Ad-GFP, PE
   group and the Ad-PRODH, PE group. (C) Statistic of differently
   expressed metabolite of two comparisons: Ad-GFP, Veh versus Ad-GFP, PE
   and Ad-GFP, PE versus Ad-PRODH, PE, respectively. (D and E) Volcano
   plot of two comparisons. (F and G) Z-score plot showing the changes of
   DEMs from two comparisons. (H and I) KEGG pathway analysis of the DEMs
   of two comparisons (n = 4 for each group).

   The results of Z-score normalization intuitively illustrated the
   changes of differentially expressed metabolites (DEMs) after various
   treatments ([133]Fig. 7, F and G, and fig. S8, F and G). Both
   experiments consistently showed elevated levels of proline and TCA
   cycle intermediates such as citrate, succinate, and fumarate, along
   with increased adenosine triphosphate (ATP) levels. PE stimulation also
   decreased reduced glutathione, accompanied by elevated oxidized
   glutathione levels. Overexpression of PRODH decreased proline levels
   and further up-regulated TCA cycle intermediates and ATP levels,
   reversing the decreased glutathione and increased oxidized glutathione
   levels. Knockdown of PRODH exhibited opposite effects.

   KEGG pathway analysis further supported our observations. Enrichment
   analysis of DEMs revealed that PE stimulation up-regulated several key
   metabolic pathways in cardiomyocytes, such as central carbon metabolism
   in cancer, the glucagon signaling pathway, and the TCA cycle ([134]Fig.
   7H and fig. S8H). PRODH overexpression induced changes in pathways
   including central carbon metabolism in cancer, the GABAergic synapse,
   and the TCA cycle ([135]Fig. 7I), while PRODH knockdown mainly affected
   processes such as central carbon metabolism in cancer, protein
   digestion and absorption, aminoacyl–transfer RNA biosynthesis, and the
   TCA cycle (fig. S8I). Together, these results fully illustrated the
   metabolic differences in hypertrophic cardiomyocytes and the metabolic
   reprogramming induced by PRODH regulation. These results together
   highlight the critical role of PRODH in modulating the TCA cycle,
   energy metabolism, and glutathione homeostasis.

PRODH overexpression replenishes mitochondrial respiration and TCA cycle in
cardiomyocytes

   Next, we asked the effect of PRODH overexpression on mitochondrial
   respiration and TCA cycle restoration. To address this, we measured
   oxygen consumption rate (OCR) using Seahorse flux analysis and found
   that PRODH up-regulated oxygen consumption in cardiomyocytes,
   regardless of PE stimulation ([136]Fig. 8A). This increase in OCR was
   in agreement with the observed rise in basal OCR, ATP-linked OCR, and
   maximal OCR ([137]Fig. 8, B to D). Notably, the increase in ATP-linked
   OCR corresponded with a noticeable elevation in ATP levels in
   PRODH-overexpressed cardiomyocytes ([138]Fig. 8E).

Fig. 8. PRODH overexpression replenishes mitochondrial respiration and TCA
cycle in cardiomyocytes.

   [139]Fig. 8.
   [140]Open in a new tab

   (A) Mitochondrial respiration of NRVMs was determined by measuring OCR.
   (B to D) Basal, ATP-linked, and maximal OCRs of NRVMs measured by
   Seahorse (n = 6). (E) ATP content of NRVMs receiving different
   treatments measured by the commercial kit (n = 6). (F) Schematic of
   ^13C-labeling pattern in NRVMs. After entering the TCA cycle,
   ^13C-labeled l-proline was catabolized mostly into M5/M4 metabolites in
   the first round (round 1) of the TCA cycle, M3/M2 in round 2, M2/1 in
   round 3, and M1/0 in round 4. (G to K) Incorporation of ^13C-proline
   into the tricarboxylic acid cycle, as revealed by the ^13C fraction of
   total C of α-KG, Suc, Fum, Mal, and Cit (n = 4). α-KG, α-ketoglutarate;
   Suc, succinate; Fum, fumarate; Mal, malate; Cit, Citrate. (Data are
   shown as means ± SEM. Two-way ANOVA was conducted for the comparisons.)

   To determine whether proline catabolism via PRODH could replenish TCA
   cycle intermediates, we conducted isotope labeling experiments using
   l-proline-^13C, ^15N to trace the integration of proline into the TCA
   cycle ([141]Fig. 8F). Overexpression of PRODH in cardiomyocytes
   channeled proline into the TCA cycle, resulting in enhanced
   α-ketoglutarate (α-KG), succinate (Suc), fumarate (Fum), malate (Mal),
   and citric acid (Cit) derived from proline, independent of PE
   stimulation ([142]Fig. 8, G to K). These findings suggest that, by
   promoting the anaplerotic use of proline, PRODH overexpression
   effectively restores the TCA cycle, thereby enhancing mitochondrial
   respiration in hypertrophic cardiomyocytes.

DISCUSSION

   In this study, we mainly investigated the alterations of proline
   metabolism during pressure overload–induced cardiac remodeling and
   further determined the role of PRODH in regulating cardiac function and
   metabolic remodeling of cardiomyocytes. We observed a notable
   down-regulation of PRODH in the hearts of mice after TAC surgery,
   indicating impaired proline metabolism during pressure-induced cardiac
   remodeling. Subsequently, we found that cardiac-specific PRODH knockout
   mice showed deteriorated cardiac function after TAC, whereas PRODH
   overexpression in cardiomyocytes appeared cardioprotective. Further in
   vitro analyses revealed that PRODH overexpression in NRVMs not only
   increased TCA cycle metabolites and ATP production but also reduced
   oxidative stress by modulating glutathione synthesis and enhancing the
   GSH/GSSG ratio, thus boosting antioxidant defense in cardiomyocytes
   ([143]Fig. 9).

Fig. 9. Schematic illustration depicting the impact of PRODH overexpression
in the inhibition of cardiac remodeling.

   [144]Fig. 9.
   [145]Open in a new tab

   The wild-type mice exhibit deficiencies in energy homeostasis along
   with decreased GSH/GSSG levels following TAC surgery. Cardiomyocytes
   overexpressing PRODH facilitate the utilization of proline, promoting
   its conversion into the TCA cycle and up-regulating GSH/GSSG levels.

   In response to persistent hemodynamic stress from pressure overload,
   the heart undergoes myocardial remodeling, namely, cardiomyocyte
   hypertrophy, to adapt to the increased workload ([146]22). Enlarged
   cardiomyocytes necessitate elevated energy consumption for cell growth
   and contractile function, leading to a profound energy shortage in the
   end ([147]23). Therefore, the cardiomyocytes undergo profound changes
   to meet the demand of energy during the myocardial hypertrophy and its
   progression to HF. Cardiomyocytes primarily depend on fatty acid
   oxidation to meet the high energy demand of the heart as a biology pump
   under physiological conditions. Of note, cardiac remodeling triggers a
   metabolic shift in substrate utilization, characterized by reduced ATP
   production from fatty acids and increased dependence on glycolysis and
   ketone body oxidation for compensation of energy ([148]8). However,
   this compensation is limited because glycolysis is less efficient at
   ATP generation. The heightened glycolytic activity cannot fully
   mitigate the energy shortfall or restore cardiac function in HF.
   Therefore, modulating cardiac energy metabolism to enhance ATP
   synthesis in cardiomyocytes can remarkably preserve cardiac function.

   In addition to increased glycolysis and ketone utilization, the
   reprogramming of amino acid metabolism also exerts a pivotal role in
   redox homeostasis and cellular growth of cardiomyocytes during pressure
   overload–induced myocardial remodeling ([149]24). Sun et al. ([150]25)
   reported a notable inhibition of branched-chain amino acid (BCAA)
   catabolic gene expression in HF. The defect in BCAA catabolic enzymes
   promoted HF as the up-regulation of BCAAs directly inhibited
   mitochondrial respiration, leading to superoxide production. In
   addition, Ritterhoff et al. ([151]26) identified aspartate as a crucial
   amino acid contributing to myocardial hypertrophy. Up-regulation of
   fatty acid oxidation by deletion of acetyl-CoA-carboxylase 2 inhibited
   glucose-derived aspartate synthesis, thereby suppressing myocardial
   hypertrophy. Currently, the role of other amino acids in pathological
   myocardial remodeling remains incompletely elucidated.

   PRODH is a mitochondrial enzyme that plays a pivotal role in cellular
   homeostasis by regulating proline metabolism. It catalyzes the
   transformation of proline into glutamate, which can either be used in
   the TCA cycle for energy production or serve as a precursor for other
   amino acids ([152]18). Recent research has expanded our understanding
   of PRODH, highlighting its involvement in various cellular processes
   such as energy metabolism, apoptosis, and the oxidative stress response
   ([153]13). In our study, we observed a notable down-regulation of PRODH
   expression in cardiac tissues after TAC surgery and in hypertrophic
   cardiomyocytes stimulated by PE, with a corresponding reduction in its
   activity demonstrated in in vivo experiments. These findings suggest
   that the catabolic metabolism of proline is suppressed during cardiac
   remodeling. The underlying cause of PRODH down-regulation in
   cardiomyocytes remains unclear. However, we observed partial
   restoration of PRODH expression in mouse cardiac tissues treated with
   rapamycin after TAC surgery. Previous studies have identified mTOR as a
   inhibitory regulator of PRODH ([154]15, [155]17), and its activation
   during TAC-induced myocardial remodeling to promote compensatory
   cardiomyocyte hypertrophy is well established ([156]20, [157]21). mTOR
   senses the cellular nutrient and energy status and enhances processes
   such as glycolysis, protein, lipid, and nucleotide biosynthesis,
   thereby systematically altering cardiomyocyte synthesis and degradation
   metabolism ([158]27). Hence, we speculate that the expression of PRODH
   may be inhibited by activation of the mTOR signaling pathway during the
   cardiac remodeling.

   In the context of cardiovascular diseases, the significance of PRODH
   has been progressively recognized. Proline metabolism was initially
   identified as a differential metabolite in HF ([159]9, [160]19), yet
   research is required to elucidate the underlying mechanisms. Moreira et
   al. ([161]28) first reported a notable down-regulation of PRODH
   expression in rat myocardium following myocardial infarction. They
   found that physical exercise could reverse this down-regulation and
   concurrently enhance cardiac function. However, this study did not
   perform PRODH expression in vivo experiments to explore its protective
   effects. By using induced pluripotent stem cell–derived cardiomyocytes,
   they found that PRODH knockdown increased BNP expression, while PRODH
   overexpression up-regulated oxidative phosphorylation in
   cardiomyocytes. In addition, PRODH knockdown resulted in reduced
   mitochondrial OCR and decreased ATP synthesis, whereas PRODH
   overexpression up-regulated mitochondrial OCR and ATP production. Our
   cellular observations are aligned with their research. In addition, in
   our in vitro experiments, we confirmed that PRODH overexpression mildly
   increased mitochondrial OCR and ATP production at baseline. Meanwhile,
   PRODH overexpression markedly enhanced ATP supply by PE stimulation.
   Our RNA-seq and metabolomic studies, supported by isotopic labeling
   metabolic flux assays, confirm that PRODH promoted mitochondrial energy
   supply by augmenting the production of TCA intermediates.

   Furthermore, we substantiated the above findings by conducting
   additional metabolic profiling of in vivo experiments. We initially
   investigated the fate of proline in cardiac remodeling and found a
   remarkable up-regulation of proline levels after TAC surgery,
   consistent with the finding of down-regulation of PRODH expression
   after TAC surgery. Cardiac-specific overexpression of PRODH reversed
   the elevation of proline after TAC, indicating that PRODH
   overexpression promoted cardiac proline utilization. Subsequently, we
   also observed a notable down-regulation of TCA cycle intermediates in
   mouse cardiac tissues 4 weeks after TAC surgery, which was reversed by
   PRODH overexpression. These findings align with our RNA-seq analysis of
   primary cardiomyocytes isolated after TAC surgery, demonstrating that
   PRODH up-regulates the decreased catabolic enzyme of proline and TCA
   cycle–related enzymes during cardiac remodeling. These conclusions
   further confirm the critical regulatory role of PRODH in proline
   metabolism and the TCA cycle, suggesting a therapeutic potential for
   addressing the energy insufficiency in HF.

   In the failing heart, elevated levels of reactive oxygen species (ROS)
   lead to mitochondrial DNA damage, antioxidant depletion, and reduced
   mitochondrial ATP production, aggravating the cardiac remodeling
   eventually ([162]29, [163]30). PRODH has been previously reported to
   modulate mitochondrial function by regulating ROS and maintaining
   mitochondrial integrity ([164]31). It has been found that PRODH
   expression is up-regulated in response to oxidative stress to maintain
   cellular redox balance ([165]32). Our previous study has proven that
   PRODH exerted a protective effect against oxidative stress damage
   through the utilization of proline in myocardial infarction ([166]33).
   Specifically, dietary proline supplementation could profoundly mitigate
   oxidative stress damage and reduce myocardial apoptosis after
   myocardial infarction. Consistent with these findings, in vitro
   experiments with cardiomyocytes subjected to hypoxia further
   corroborated the same conclusion ([167]34). The findings from this
   study align with these observations above. Specifically, in PRODH cKO
   mice after TAC surgery, a marked exacerbation of oxidative stress
   damage was observed, while PRODH-overexpressed mice exhibited
   attenuated oxidative stress induced by pressure overload. Similar
   results were found in the NRVMs stimulated by PE. The administration of
   the antioxidant NAC partially improved the deteriorated cardiac
   function in cKO mice after TAC, accompanied by a decrease in heart
   weight. This finding demonstrated that the exacerbated oxidative stress
   in cKO mice is as a crucial contributor of the deterioration of cardiac
   function.

   Our results also hinted that PRODH participated in glutathione redox
   homeostasis by regulating the GSH/GSSG ratio. Overexpression of PRODH
   was found to elevate the GSH/GSSG ratio, while down-regulation of PRODH
   inhibited the GSH/GSSG ratio in both in vivo and in vitro experiments.
   These results provide insight into the role of PRODH in enhancing
   cellular antioxidant capacity. In our mechanistic investigations, we
   observed that, despite the lack of up-regulation in the expression of
   the key enzyme GCLC involved in glutathione synthesis, overexpression
   of PRODH led to an increase in the expression of GSR. This phenomenon
   suggests a potential facilitation of the conversion of GSSG to GSH
   within cardiomyocytes.

   This study illustrated the protective role of PRODH expression in
   cardiomyocytes during cardiac remodeling and elucidates its underlying
   mechanisms. However, there are still some limitations that need to be
   addressed. First, evaluations of cardiomyocyte function, especially
   mitochondrial respiratory function and isotope labeling experiments,
   were conducted in vitro. Performing an in vivo experiment would further
   solidify these findings. Second, while our experiments have
   demonstrated a correlation between the PRODH expression and mTOR
   signaling pathway, further experiments are still needed to explore
   whether there is a causal relationship, which would help explain why
   PRODH is down-regulated after TAC. Lastly, additional research
   elucidating the direct regulatory effects of PRODH on GSR expression
   would further deepen our understanding of how PRODH maintains
   glutathione homeostasis.

   In conclusion, this study reveals the reprogramming of proline
   metabolism during cardiac remodeling and its potential therapeutic
   implications. PRODH is shown to alleviate cardiac remodeling by
   reestablishing energy production through the enhanced TCA cycle and the
   improved antioxidant capacity of cardiomyocytes. These findings
   highlight the potential of targeting proline metabolic pathways as a
   treatment strategy for cardiac remodeling caused by pressure overload.

MATERIALS AND METHODS

Experimental animals and experimental protocols

   All animal experiments followed the Institutional Animal Care
   Guidelines and were approved by the Animal Experimentation Ethics
   Committee of Zhejiang University. The 8- to -10-week-old wild-type
   C57BL/6 male mice and 1- to 3-day-old neonatal Sprague-Dawley (SD) rats
   were purchased from the Vital River Laboratory Animal Technology Co.,
   Ltd. (Beijing, China). The cardiac-specific conditional PRODH knockout
   mice were generated by Cyagen Biosciences (Guangzhou, China). In brief,
   the mouse Prodh gene contains 13 exons; exon 3 to 8 of Prodh was chosen
   for targeting using a CRISPR-Cas9–mediated genome-editing approach. The
   flox mice were crossed with Myh6-iCre mice to generate the
   cardiac-specific PRODH cKO mice. The C57BL/6 mice mentioned above were
   maintained in a pathogen-free animal house with a 12-hour light/dark
   cycle and constant temperature. All animal experiments were approved by
   the Institutional Animal Care and the Use Committee of Zhejiang
   University.

   For the investigation of the relationship between mTOR and PRODH,
   wild-type mice after TAC were randomly divided into two groups. They
   were intraperitoneally injected with rapamycin at a dose of 2 mg/kg per
   day or an equivalent volume of saline. After 28 days, mouse cardiac
   tissues were obtained for relevant examinations ([168]35). To
   investigate whether NAC can restore the impaired cardiac function in
   cKO mice following TAC surgery, both f/f and cKO mice were randomly
   divided into two groups after TAC operation. They were administered
   either 500 mg/kg per day of NAC or saline orally for 28 days, followed
   by relevant experimental assessments ([169]36).

TAC model

   To build a model of pressure overload–induced cardiac hypertrophy, TAC
   surgery was performed among C57BL/6 male mice of 8 to 10 weeks old.
   Briefly, the mice were first anesthetized with 0.3% sodium
   pentobarbital (25 ml/kg, intraperitoneally). A minimally invasive
   surgical incision was conducted on the upper edge of the second rib in
   the left chest after the mouse had no toe-pinch reflex. The thoracic
   aorta was then identified with the help of a stereomicroscope. A
   27-gauge needle and 6-0 silk sutures were then used to ligate the
   thoracic aorta between the innominate and left carotid arteries. When
   the needle was removed, a region of stenosis among the thoracic aorta
   was successfully presented. Lastly, the incision was closed by 4-0 silk
   sutures. The mice were lastly placed under a heat lamp until anesthesia
   recovery. Efforts were made to minimize the suffering of the mice in
   all experiments.

Echocardiography

   Echocardiography was performed by the Vevo 1100 system (VisualSonics
   Inc., Toronto, Canada) equipped with a 30-MHz ultrasound transducer as
   previously described ([170]37). The long-axis view of M-mode images was
   captured to analyze various parameters, including EF, FS, LVIDd, and
   LVIDs.

AAV9 delivery

   For cardiac-specific exogenous expression of PRODH, recombinant AAV9
   overexpressing a full-length mouse PRODH cDNA was constructed into a
   plasmid containing a chicken troponin T promoter, obtained from Hanheng
   Biotechnology (Shanghai, China). AAV9-Vector was produced as a negative
   control. The viruses were delivered to mice with a dose of 1 × 10^11
   virus particles (vg) via tail injection before or after the TAC surgery
   according to the experimental design.

Histology analysis

   The hearts of mice after TAC surgery were harvested and immediately
   fixed in 4% paraformaldehyde (PFA) at room temperature for 24 hours.
   Subsequently, the hearts were embedded in paraffin and cut into
   5-μm-thick sections. H&E staining and Masson’s trichrome staining were
   performed to observe the gross morphology and fibrosis areas of the
   hearts following standard procedures, respectively.

Isolation of NRVMs and cell culture

   Neonatal SD rats aged 1 to 3 days old were used to isolate NRVMs.
   First, the hearts removed from the neonatal rats were cut into 1- to
   2-mm pieces and washed twice with cold phosphate-buffered saline (PBS).
   The tissue was then digested by the mixture of 0.04% Collagen II and
   0.06% trypsin 8 to 10 times for 8 min each. The cells collected from
   each digestion were temporarily kept on ice. After the digestion, the
   cells were collected by centrifugation at 1000 rpm for 5 min. The cells
   were then resuspended by red blood cell lysis buffer (Solarbio Science
   and Technology Co., Ltd., Beijing, China) to remove the red blood
   cells. The myofibroblasts were separated by the differential adhesion
   method. The myocytes were plated in gelatin-coated plates at a density
   of 5 × 10^5 cells/ml by Dulbecco’s modified Eagle’s medium (DMEM) with
   10% fetal bovine serum and bromodeoxyuridine (100 μM). After 24 hours,
   the myocytes were then cultured by the serum-free DMEM for at least 12
   hours before further experiments.

Isolation of AMCMs

   The isolation of AMCMs was conducted using the Langendorff isolated
   heart perfusion system, closely adhering to methodologies delineated in
   a prior publication ([171]38). Initially, the perfusion buffer,
   Krebs-bicarbonate solution, stopping buffer, and myocyte digestion
   buffer were prepared, which are listed in tables S1 to S4. An adult
   mouse heart was then excised and immediately placed in a 60-mm dish
   filled with perfusion buffer to ensure complete expulsion of blood. The
   aorta was identified, clamped with forceps, and mounted onto a
   Langendorff needle in a precise manner. Upon establishment of this
   setup, the perfusion was initiated with the pump activated at a flow
   rate of 3 to 4 ml/min. Following a period of 5 min, the enzyme solution
   was perfused at an adjusted flow rate of 5 to 6 ml/min. After 15 min of
   enzymatic perfusion, observations indicated that the heart had swollen
   and turned slightly pale, signifying complete digestion. The heart was
   then transferred to a fresh 60-mm dish containing 2.5 ml of the enzyme
   solution, where it was dissected into small, cotton-like pieces using
   tweezers. These fragments were subjected to pipetting 20 to 30 times
   with a wide-bore pipette before being transferred to a 15-ml centrifuge
   tube containing 2.5 ml of stopping buffer, with further pipetting
   conducted approximately 20 times. A droplet was placed on the underside
   of a dish lid for microscopic examination to estimate the percentage of
   beating cardiomyocytes among the total cells. The solution was
   subsequently supplemented with stopping buffer to reach a final volume
   of 10 ml. Centrifugation at 20g for 5 min facilitated the sedimentation
   of cardiac myocytes, leaving the cardiac fibroblasts in the
   supernatant. The sediment was collected and resuspended in
   Krebs-bicarbonate solution and then kept on ice for further
   experimental procedures.

Adenoviral infection and siRNA transfection

   For the in vitro experiments, recombinant adenoviruses expressing GFP
   or PRODH were used for exogenous expression of PRODH (Gene Pharma Co.,
   Ltd., Shanghai, China). NRVMs were infected by adenoviruses after being
   cultured by serum-free DMEM for at least 12 hours [multiplicity of
   infection (MOI) = 50]. PE treatments were conducted after 24 hours of
   transfection. Knockdown of PRODH was conducted by transfection of siRNA
   using iMax transfection reagent (Thermo Fisher Scientific, Waltham,
   USA). Briefly, NRVMs were transfected by 50 μM siRNA for 24 hours
   before PE treatment. The sequence of siRNA for PRODH was as follows:
   -GGACUAUGGUGUGGAGGAAdTdT and -UUCCUCCACACCAUAGUCCdTdT.

Immunofluorescence staining

   To determine the relative cell surface area of NRVMs, the F-actin
   cytoskeleton was stained using phalloidin-TRITC (red) (YEASEN,
   Shanghai, China). Briefly, The NRVMs were first fixed in 4% PFA for 10
   min and then permeabilized with 0.1% Triton X-100 for 5 min at room
   temperature. The cells were subsequently stained with 100 nM
   TRITC-labeled phalloidin and mounted with
   anti-fade/4′,6-diamidino-2-phenylindole (DAPI) mountant.

   For WGA staining, frozen heart sections (5 μm) were washed three times
   with PBS, then fixed with 4% PFA solution for 15 min at 37°C, and then
   stained with Alexa Fluor 594-WGA with a concentration of 10 μg/ml
   (Thermo Fisher Scientific, Waltham, USA). After washing with PBS, the
   sections were mounted with anti-fade/DAPI mountant. The cross-sectional
   areas of cardiomyocytes were measured by Image J (v1.50).

   To evaluate the ROS generated in the cells, DHE staining (YEASEN,
   Shanghai, China) was performed on the heart sections and NRVMs using a
   10 μM DHE solution at 37°C for 1 hour. The sections were then mounted
   with anti-fade/DAPI mountant after washing with PBS. The ratio of
   DHE-positive cells to the total amount of DAPI was calculated for the
   in vivo experiment, and the fluorescence intensity was measured for the
   in vitro experiment.

Electron microscopy

   Freshly dissected cardiac tissues from the left ventricle of mice were
   dissected into pieces of approximately 1 mm^2 and immediately immersed
   in a 2.5% glutaraldehyde solution for fixation, followed by overnight
   incubation at 4°C. Subsequently, they were fixed with 1% osmium
   tetroxide, dehydrated in acetone, and embedded in epoxy resin. Next,
   the samples were sectioned into 70-nm slices using an ultramicrotome,
   followed by double staining with 3% uranyl acetate and lead citrate.
   Last, the prepared samples were observed using a JEM1230 transmission
   electron microscope (JEOL, Japan).

PRODH activity assay

   The PRODH activity assay was performed using a commercial kit
   (Solarbio, Beijing, China). In brief, cardiac tissues were placed into
   an extraction solution and kept on ice, followed by centrifugation at
   1500g for 15 min at 4°C. Subsequently, the additional prepared
   extraction solution was added, and the mixture was gently agitated for
   30 min on ice. Next, the supernatant was collected by centrifugation at
   15,000g for 20 min at 4°C and incubated at 37°C for 5 min. The
   absorbance of the samples or standard solutions was then measured at a
   wavelength of 600 nm using a spectrophotometer. Last, the detected
   PRODH activity was normalized according to the samples’ mass.

GSH/GSSG ratio detection assay

   To determine the relative GSH/GSSG of the NRVMs and AMCMs, a commercial
   detection assay kit was applied according to the manufacturer’s
   instructions (Abcam, Cambridge, UK). Briefly, the supernatant of
   samples was collected after high-speed centrifugation for further
   analysis following the lysis of cells. The supernatant was mixed with
   an equal volume of assay reagent, and the mixture was incubated in the
   dark at room temperature for 1 hour before measuring the fluorescence
   intensity at 520 nm. By comparing the results with a standard curve,
   the contents of GSH and GSSG in NRVMs and AMCMs are determined, and the
   GSH/GSSG ratio is subsequently calculated.

Single-cell sequence analysis

   The single-cell RNA-seq data for this study were obtained from
   [172]GSE95140, including TAC-model mice at different stages (sham, 3
   days, 1 week, and 4 weeks). Quality control and cell annotation were
   performed according to a prior study. R package Seurat (v4.1.3) was
   used for data integration, cell filtration, normalization, clustering,
   and t-distributed stochastic neighbor embedding (t-SNE) dimensional
   reduction. Monole2 (v2.22.0) was used for cell pseudotime analysis.
   Cell progression genes were defined based on DEGs among Seurat
   clusters. Differential gene test function was used to explore gene
   dynamics during cell differentiation. Therefore, the trajectory was
   divided into five segments according to the trajectory nodes.

RNA-seq and bioinformatic analysis

   The RNA-seq was performed by Novogene Company. The total RNAs of
   cardiac tissue or AMCMs were extracted by TRIzol reagent and quantified
   using the NanoDrop spectrophotometer. The quality of RNA was then
   evaluated with a standard protocol. RNA samples with sufficient
   quantity and high quality were subjected to library preparation.
   RNA-seq libraries were created using the Illumina HiSeq Stranded mRNA
   Library Prep Kit. The prepared libraries were sequenced on an Illumina
   NovaSeq 6000 according to the manufacturer’s instructions. Raw
   sequencing data in FASTQ format were generated and subjected to quality
   control. Raw sequencing reads were preprocessed to remove low-quality
   bases, adaptors, and contaminants. The cleaned reads were aligned to
   Mus musculus using STAR with default parameters. The aligned reads were
   sorted and indexed for downstream analysis. Gene expression levels were
   quantified based on the read counts mapped to each gene. Differential
   gene expression analysis was performed using DESeq2. Genes with
   adjusted P values below a predefined threshold (FDR < 0.05) were
   considered as DEGs. Functional annotation and pathway enrichment
   analysis of DEGs were conducted using the KEGG database. Pathways
   substantially enriched with DEGs were identified using DAVID.
   Visualization of enriched pathways and related genes was performed
   using R packages.

Nontargeted metabolite profiling and bioinformatics

   In this study, nontargeted metabolite profiling was conducted by
   Panomix Biotech Company (Suzhou, China). The collection of samples,
   specifically NRVMs post-treatment, adhered to the manufacturer’s
   standard protocol. Metabolite extraction was performed using 80%
   methanol to encompass a wide spectrum of metabolites, with an optimized
   extraction process to maximize recovery. Following extraction, samples
   were subjected to vortexing and subsequent centrifugation to segregate
   the organic and aqueous phases. The extracted metabolites were then
   lyophilized for subsequent liquid chromatography–mass spectrometry
   (LC-MS) analysis. LC-MS analysis was executed using high-performance
   liquid chromatography coupled with a quadrupole orthogonal
   acceleration–time-of-flight mass spectrometer. Instrument calibration
   was carried out according to the manufacturer's recommendations.
   Chromatographic separation was achieved using a mobile phase, with
   optimization of chromatographic parameters including flow rate and
   gradient to enhance metabolite separation. Mass spectrometry analysis
   was performed in both positive and negative ionization modes, acquiring
   full-scan spectra over a predefined mass range. The raw LC-MS data
   generated during the analysis were stored in a standardized format
   established by Panomix Biotech Company. These raw data files were
   subsequently subjected to preprocessing steps, including peak
   detection, alignment, and retention time correction. Detected features
   were annotated by referencing available spectral databases to
   facilitate metabolite identification. Differential metabolite analysis
   was conducted using statistical tools, applying appropriate tests to
   identify notable differences between experimental groups. Furthermore,
   pathway enrichment analysis of the DEMs was carried out to discern
   metabolic pathways associated with these alterations. It is noteworthy
   that all data analyses described herein were conducted within the
   BioDeep cloud analysis platform.

Seahorse assay

   The OCR assessment of NRVMs was executed using an Agilent Seahorse XFe
   96 Extracellular Flux Analyzer, adhering strictly to the manufacturer’s
   guidelines. The experiment was initiated by seeding the NRVMs on XFe 96
   cell culture microplates with DMEM. This was followed by transfection
   of the cells with Ad-PRODH and subsequent induction with PE. On the day
   of the experiment, the hydration plate was meticulously rinsed twice
   with a calibration solution, followed by incubation in a 37°C culture
   incubator for 45 min. Subsequently, the NRVMs were washed twice with
   XFe 96–configured medium and incubated in a 37°C culture incubator for
   1 hour. The subsequent step involved the addition of oligomycin (1.5
   μM), carbonyl cyanide p-(trifluoromethoxy) phenylhydrazone (2.0 μM),
   and a combination of rotenone and antimycin A (0.5 μM) into the probe
   plate, which was subsequently introduced into the analyzer for
   calibration. Upon successful calibration, the incubated cells were
   transferred onto the probe plate to facilitate the final determination
   of the OCR.

ATP content assay

   The ATP content measurement was conducted using a commercial kit
   (Beyotime, Haimen, China). NRVMs were seeded in six-well plates and
   subsequently received different treatments. At the end of the
   treatment, 200 μl of lysis buffer was added to each well of the
   six-well plate. Subsequently, the lysates were centrifuged at 12,000g
   for 5 min at 4°C, and the supernatants were collected for subsequent
   assays. Next, 100 μl of ATP working solution was added into detection
   tubes incubated at room temperature for 3 to 5 min. Following this, 20
   μl of the test supernatant or standard solution was added and mixed.
   Last, the relative light units of each sample were measured using a
   luminometer. Simultaneously, the supernatant from each sample was used
   for protein quantification to normalize the ATP content to micrograms
   of protein.

l-Proline-^13C5, ^15N stable isotope tracing analysis

   The NRVMs were initially seeded on six-well plates. After 12 hours of
   serum-free culture, the cells were transfected with Ad-GFP or Ad-PRODH
   for 24 hours. Subsequently, a specialized culture medium for isotope
   experiments was prepared, consisting of serum-free, glucose-free, and
   glutamine-free DMEM; 25 mM d-glucose; 4 mM glutamine; and 2 mM
   l-proline. PE (100 μM) was added to this medium to stimulate
   cardiomyocytes for 48 hours. Upon completion of the cell treatment,
   cardiomyocytes were washed twice with a pre-cooled 0.9% sodium chloride
   solution. The cells were then extracted with −80°C pre-cooled 80%
   methanol. The collected cells were subjected to enzyme inactivation at
   −80°C overnight. Afterward, the supernatant was then dried by a vacuum
   concentrator and stored at −80°C for further LC-MS analysis. The
   protein concentration was determined by dissolving the sediment with
   200 mM NaOH. The isotope tracing analysis was performed by Lipid ALL.
   High-resolution mass spectrometry coupled with liquid chromatography
   was used to analyze ^13C metabolites. LC-MS data acquisition, raw data
   analysis, and normalization were performed using LC-MS peak intensities
   (integral peak area) to calculate the relative abundance ratio of each
   metabolite. For individual metabolites, the ^13C isotopic distribution
   was determined using the peak intensities of each ^13C isotope.

RNA isolation, cDNA synthesis, and quantitative RT-PCR

   Total RNA from NRVMs or mouse left ventricle tissue was extracted using
   the RNA quick purification kit (Esunbio Co., Ltd., Shanghai, China).
   The RNA was then reverse-transcribed for cDNA synthesis by the
   PrimeScriptTM RT Master Mix (Takara, Tokyo, Japan). Quantitative RT-PCR
   was performed to determine the relative mRNA expression of target genes
   using the Hieff UNICON qPCR SYBR Green Master Mix (YEASON, Shanghai,
   China) on the Viia 7 system (Applied Biosystems, CA, USA). All primer
   sequences of target genes are listed in table S5.

Western blotting assay

   The protein isolation from the animal tissue was performed using a
   tissue grinder. Briefly, heart tissues were lysed in the buffer with
   radioimmunoprecipitation assay (RIPA) buffer and Phenylmethylsulfonyl
   fluoride (PMSF) (100:1). The whole lysates were then collected by the
   tissue grinder. For the extraction of protein from cultured cells, the
   cells were first washed with cold PBS twice and lysed with 100 μl of
   RIPA and PMSF (100:1) for each well. After the centrifugation at 12,000
   rpm for 15 min, the supernatants were collected for the bicinchoninic
   acid assay (Beyotime, Haimen, China). Then, the protein samples were
   boiled for 5 min at 100°C. The same amounts of proteins of each sample
   were loaded onto the SDS–polyacrylamide gel electrophoresis gels (Fude
   Biological Technology, Hangzhou, China) for electrophoresis, followed
   by transferring to the 0.22-μm polyvinylidene fluoride membrane
   (Bio-Rad, CA, USA) with a constant current of 250 mA for 2 hours. The
   membrane was then blocked by the 5% nonfat milk for 1 hour at room
   temperature, followed by incubating with corresponding primary
   antibodies overnight at 4°C. The primary antibodies used in this study
   are listed as follows: ANP (27426-1-AP, Proteintech), PRODH
   (22980-1-AP, Proteintech), β-MHC (22280-1-AP, Proteintech), p-S6K
   (T389) (9234, CST), S6K (34475, CST), p-4E-BP1 (S65) (9451, CST),
   4E-BP1 (9452, CST), GCLC (12601-1-AP, Proteintech), GSR (18257-1-AP,
   Proteintech), and GAPDH (HRP-60004, Proteintech). After incubating with
   secondary anti-rabbit (70-GAM0072, MultiSciences) or anti-mouse
   (70-GAM007, MultiSciences) antibodies, the bands were detected by
   electrochemiluminescence reagents (Beyotime, Haimen, China) by the
   Amersham Imager 600 system (GE Healthcare, Buckinghamshire, England).

Statistical analysis

   All data in this study are expressed as means ± SEM. Statistical
   analysis was performed with Prism 8.0 (GraphPad Software, Inc.).
   Student’s unpaired t test was performed for the comparison between the
   two groups. One-way ANOVA followed by Bonferroni post-hoc test was
   performed by comparison among three or more groups when involving one
   factor. Two-way ANOVA with Bonferroni post-hoc test was used for the
   four-group comparison when involving two factors. P < 0.05 was
   considered as a statistical difference in this study. Statistical
   analyses were performed using GraphPad Prism (8.0).

Acknowledgments