The phospholamban R14del generates pathogenic aggregates by impairing autophagosome-lysosome fusion

Abstract

Phospholamban (PLN) plays a crucial role in regulating sarcoplasmic reticulum (SR) Ca²⁺ cycling and cardiac contractility. Mutations within the PLN gene have been detected in patients with cardiomyopathy, with the heterozygous variant c.40_42delAGA (p.R14del) of PLN being the most prevalent. Investigations into the mechanisms underlying the pathology of PLN-R14del have revealed that cardiac cells from affected patients exhibit pathological aggregates containing PLN. Herein, we performed comprehensive molecular and cellular analyses to delineate the molecular aberrations associated with the formation of these aggregates. We determined that PLN aggregates contain autophagic proteins, indicating inefficient degradation via the autophagy pathway. Our findings demonstrate that the expression of PLN-R14del results in diminished autophagic flux due to impaired fusion between autophagosomes and lysosomes. Mechanistically, this defect is linked to aberrant recruitment of key membrane fusion proteins to autophagosomes, which is mediated in part by changes in Ca²⁺ homeostasis. Collectively, these results highlight a novel function of PLN-R14del in regulating autophagy, that may contribute to the formation of pathogenic aggregates in patients with cardiomyopathy. Prospective strategies tailored to ameliorate impaired autophagy may hold promise against PLN-R14del disease.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-024-05471-1.

Introduction

Phospholamban (PLN) is a 52-amino acid transmembrane protein that regulates sarcoplasmic reticulum (SR) Ca²⁺ cycling and Ca²⁺-dependent cardiac contractility. PLN interacts with and inhibits the activity of the SR calcium ATPase SERCA2a by decreasing its affinity for Ca²⁺ [1]. Once phosphorylated at Ser¹⁶, PLN switches to a non-inhibitory state, allowing increased SR Ca²⁺ uptake by SERCA2a and cardiomyocyte relaxation [1].

To date, various PLN mutations have been identified in arrhythmogenic cardiomyopathy (ACM), dilated cardiomyopathy (DCM) and heart failure patients [2]. However, the heterozygous PLN variant c.40_42delAGA, which causes deletion of arginine at amino acid residue 14 (R14del), has attracted considerable attention and is currently under extensive investigation [3, 4]. Initially, this variant was identified in a large Greek family with DCM and heart failure that presented with arrhythmia symptoms [5]. Since the original report, the PLN c.40_42delAGA has been found in an increasing number of patients worldwide [6–9]. In the Netherlands, PLN c.40_42delAGA is the most prevalent cardiomyopathy-related mutation (> 1500 patients) and is present in ~ 12% of patients with ACM and ~ 15% of patients with DCM [10, 11].

The pathophysiological mechanisms underlying PLN-R14del have recently started to unravel based on studies using PLN-R14del animal models, patient-derived induced pluripotent stem cell cardiomyocytes (iPSC-CMs) and in vitro cell systems [12–20]. These studies revealed alterations in various pathways, including PLN structural changes, protein interaction aberrations, Ca²⁺ handling dysregulation, mitochondrial dysfunction, unfolded protein response (UPR) activation and protein aggregation, each of which contributes to disease manifestation to varying degrees (reviewed in [2]).

The formation of PLN aggregates is a characteristic feature of PLN-R14del disease and is detected primarily in the perinuclear region of ventricular cardiomyocytes in patient hearts [21, 22]. Studies in a knock-in Pln-R14del mouse model demonstrated that these aggregates can be detected in early disease stages, before the onset of functional defects, and that they can significantly alter the expression of a number of proteostasis players, including those in the ubiquitin–proteasome pathway [14]. Other studies have shown that PLN aggregates colocalize with key autophagy-related proteins, such as sequestosome-1 (p62) and microtubule-associated protein light chain 3 (LC3), in patient cells and suggest inadequate degradation by the autophagy pathway [22, 23]. Similar observations have been reported in two knock-in PLN-R14del mouse models [24, 25], but the mechanism underlying this PLN-R14del defect has not been determined.

In this study, we demonstrate that the expression of PLN-R14del in cellular systems leads to impaired autophagy and decreased autophagic flux due to reduced autophagosome-lysosome fusion. Our findings strongly indicate alterations in subcellular Ca²⁺ levels as a primary factor, disrupting the recruitment of key proteins crucial for orchestrating the necessary events leading to autophagosome–lysosome membrane fusion. The resulting aggregate formation may contribute to the increased cell death, fibrosis and cardiac dysfunction characteristic of PLN-R14del disease.

Materials and methods

Tissue culture and transfections

HEK293 cells (ECACC, Salisbury, UK) were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, Waltham, USA), as previously described [26, 27]. The generation of the GFP-PLN-WT, GFP-PLN-R14del and mCherry-PLN-R14del constructs used in this study has been previously described [20, 27]. Transient transfections in HEK293 cells were performed using Lipofectamine™ 2000 (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions.

The cardiac myoblast cell line H9c2 (ECACC) was maintained in DMEM supplemented with 10% FBS (Thermo Fisher Scientific), as previously described [20]. Transient transfection of H9c2 cells was performed using Lipofectamine™ 3000 (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. Cell differentiation and myotube formation were induced by switching to differentiation media (DMEM supplemented with 1% FBS and 10 nM retinoic acid).

Aggregate detection

Protein aggregate formation was visualized using a Proteostat^® Aggresome detection kit (Enzo Life Sciences, Inc., Lausen, Switzerland) according to the manufacturer’s instructions. For this qualitative study, H9c2 cells were transfected and differentiated for 7 days before analysis. In brief, cells were fixed with 4% formaldehyde and subsequently permeabilized with 0.5% Triton X-100 and 3 mM EDTA buffer. The samples were then incubated with Proteostat dye for 30 min at room temperature and mounted with Fluoroshield mounting medium containing DAPI (Abcam, Cambridge, UK). Imaging was performed on a Leica confocal laser scanning microscope (Leica TCS SP5 II on a DM 6000 CFS Upright Microscope) using the LAS-AF acquisition software program. Untransfected H9c2 cells treated with or without the MG-132 inhibitor (4 µM for 7 h) served as positive and negative controls, respectively.

An additional approach included examining aggregate colocalization of the p62 and LC3 proteins, as observed in human patient cells [22, 23]. For this analysis, immunofluorescence studies were performed in transfected H9c2 cells using SQSTM1/p62 (A-6) (Santa Cruz Biotechnology, Heidelberg, Germany) and MAP LC3β (G-9) (Santa Cruz Biotechnology) antibodies, as described below.

Autophagy treatments and autophagic flux measurement

Evaluation of autophagic flux was performed 24 h after transfection of HEK293 cells or 7 days postdifferentiation in H9c2 cells. The cells were treated with 25 µM chloroquine (Cell Signaling Technology, Leiden, The Netherlands) for 7 h to inhibit autophagy, while 2 h of treatment with 50 nM rapamycin (Santa Cruz Biotechnology) was used to induce autophagy. The impact of cytosolic Ca²⁺ alterations was evaluated by 4 h of treatment with 2 µM thapsigargin (Cell Signaling Technology) or 2 h treatment with 2 µM BAPTA-AM (Biotium, Fremont, USA). Following all the treatments, the cells were harvested for western blot or immunofluorescence analysis, as described below.

For autophagic flux measurements, the cells were treated with CQ as described above, and changes in LC3 localization and expression levels were evaluated before and after autophagy inhibition. The analysis included LC3 puncta quantification using the Cell Counter plugin of ImageJ [28], while LC3-II protein levels in treated versus untreated cells were determined by western blot analysis using the LC3A/B (D3U4C) XP^® antibody (Cell Signaling Technology), as described below.

RNAi-mediated knockdown

For the siRNA experiments, SERCA2 or control siRNAs (Santa Cruz Biotechnology, Heidelberg, Germany) were transfected into HEK293 or H9c2 cells using Lipofectamine RNAiMAX reagent in accordance with the manufacturer’s instructions (Invitrogen, Thermo Fisher Scientific). HEK293 cells were harvested after 48 h for western blot analysis, while H9c2 cells were differentiated for 6 days prior to immunofluorescence analysis, as described below.

Immunofluorescence analysis

Immunofluorescence studies of transfected H9c2 cells were performed after 7 days of differentiation, as previously described [20]. In brief, cells were fixed for 20 min at 25 °C with ice-cold methanol, washed three times with phosphate-buffered saline (1x PBS) and permeabilized for 30 min at 25 °C in PBS containing 0.1% Triton X-100. The cells were then washed in PBS prior to incubation with Image-iT™ FX Signal Enhancer (Thermo Fisher Scientific) for 30 min and blocking buffer (1x PBS, 1 mg/ml BSA, 10 mM NaN₃) for 1 h at 25 °C. LC3 A/B (D3U4C) XP^® (Cell Signaling Technology), LAMP1 (LY1C6), SQSTM1/p62 (A-6), MAP LC3β (G-9), Vps41 (E-10) (Santa Cruz Biotechnology), Rab7, Syntaxin 17 (D3D7H) and UVRAG (D2Q1Z) (Cell Signaling Technology) antibodies were diluted in blocking buffer and applied to the cells overnight at 4 °C. Following three washes with PBS, the cells were counterstained for 1 h at 25 °C with Alexa Fluor anti-mouse or anti-rabbit 568 and Alexa Fluor anti-rabbit or anti-mouse 633 (Invitrogen, Thermo Fisher Scientific) secondary antibodies diluted in blocking buffer. In untransfected H9c2 cells, Alexa Fluor anti-mouse or anti-rabbit 488 and Alexa Fluor anti-rabbit or anti-mouse 568 (Invitrogen, Thermo Fisher Scientific) secondary antibodies were used. After three washes with PBS, the samples were mounted with Fluoroshield mounting medium supplemented with DAPI (Abcam, Cambridge, UK).

All the samples were analyzed on a Leica confocal laser scanning microscope (Leica TCS SP5 II on a DM 6000 CFS Upright Microscope) using the LAS-AF acquisition software program. Intensity profiles were generated using Leica Application Suite X software (Leica Microsystems), while colocalization analysis was performed using the colocalization threshold plugin of ImageJ [28]. To calculate the proportion of puncta colocalized with LC3, an object-based method was used with the JACoP plugin of ImageJ [29], and automatically generated thresholds were applied, as previously described [30].

Cytosolic Ca2+ staining

Cytosolic Ca²⁺ levels were assessed using Rhod-2 AM (Biotium, Fremont, California, USA) fluorescence staining, in accordance to manufacturer’s instructions. Briefly, cells were incubated for 30 min with 1 µM Rhod-2 AM and equal volume of 20% Pluoronic F-127 (Invitrogen, Thermo Fisher Scientific) in Krebs-Ringer-glucose buffer (136 mM NaCl, 10 mM HEPES, 4.7 mM KCl, 1.25 mM MgSO₄, 1.25 mM CaCl₂, 25 mM glucose). Incubation with Rhod-2 AM was performed at room temperature in order to reduce dye compartmentalization. Following three washes with Krebs-Ringer-glucose buffer, samples were mounted with Fluoroshield mounting medium supplemented with DAPI (Abcam, Cambridge, UK) and were analyzed under the same settings on a Leica confocal laser scanning microscope (Leica TCS SP5 II on a DM 6000 CFS Upright Microscope).

Cell lysis and Western blot analysis

Following treatment of HEK293 cells with autophagy inhibitors, the samples were harvested and lysed in 1× RIPA buffer (Cell Signaling Technology) supplemented with protease inhibitors (Sigma‒Aldrich). Western blot analysis was performed using LC3A/B (D3U4C) XP^®, SQSTM1/p62, Rab7, Syntaxin 17 (D3D7H), UVRAG (D2Q1Z), SERCA2 (Cell Signaling Technology), GFP (B-2), Vps41 (E-10) (Santa Cruz Biotechnology) and GAPDH (Sigma‒Aldrich) primary antibodies and peroxidase-conjugated goat anti-rabbit (GE Healthcare Life Sciences) or anti-mouse (Sigma‒Aldrich) secondary antibodies. Immunoreactive bands were detected using Pierce ECL Plus reagents (Thermo Fisher Scientific), and protein quantification was performed using ImageJ [28].

Statistical analysis

Statistical analyses were performed in GraphPad Prism 8.0.2 software (GraphPad Software, Inc., La Jolla, CA, USA) using one-way analysis of variance (ANOVA) followed by post hoc Tukey’s test for pairwise comparisons among multiple experimental groups. In the case of two experimental group analysis, two-tailed student’s t-test statistical test was performed. A calculated P-value of < 0.05 was considered to be statistically significant. Data are presented as mean ± standard effort of the mean (SEM).

Results

PLN-R14del expression results in aggregate formation in vitro

In humanized PLN-R14del knock-in mice and cell culture models, PLN-R14del is localized at the SR and perinuclear region, similar to PLN-WT [17, 20]. We observed similar distributions of PLN-R14del in cardiac H9c2 (Fig. 1a) and HEK293 (Supplementary Fig. 1a) cells cotransfected with wild-type and mutated PLN, which were used to model the heterozygous inheritance of PLN c.40_42delAGA. In approximately 35% of the transfected cells, we noted perinuclear PLN puncta containing both the wild-type and mutated proteins, as indicated by the colocalization of the GFP and mCherry signals (Fig. 1b). Notably, the expression of the mutated PLN was crucial for puncta formation, as puncta were observed in H9c2 cells transfected with PLN-R14del but not in those transfected with PLN-WT alone (Fig. 1c). Staining with Proteostat, a reagent that fluoresces upon binding to aggregated proteins, qualitatively confirmed that the observed puncta were PLN-containing aggregates (Fig. 1c).

Fig. 1.

PLN-R14del expression leads to aggregate formation. (a) Perinuclear PLN puncta were observed in a proportion of cells cotransfected with GFP-PLN-WT and mCherry-PLN-R14del (shown by arrows). (b) Quantification of the cellular distribution (n = 75 cells from 4 experiments). (c) Proteostat staining confirms that the puncta structures observed in the presence of PLN-R14del represented protein aggregates (n = 3 experiments). Proteostat puncta colocalized with PLN-containing aggregates in all cells exhibiting aggregates (~ 35% for PLN-R14del and ~ 4% for PLN-WT). Scale bar 10 μm

Furthermore, fluorescence microscopy of H9c2 cells cotransfected with GFP-PLN-WT and mCherry-PLN-R14del that were stained with p62 or LC3 antibodies revealed colocalization of p62 and LC3 with PLN aggregates (Supplementary Fig. 1b-c). These findings collectively indicate that the in vitro expression of PLN-R14del replicates the effects observed in patient cardiomyocytes and murine disease models.

Expression of PLN-R14del leads to impaired autophagy

Since PLN aggregates colocalize with LC3 and p62 autophagy proteins in patient cells [22, 23], we hypothesized that PLN aggregate formation may result from direct interference of the autophagy pathway by PLN-R14del. To address this hypothesis, we combined immunofluorescence microscopy and immunoblot analyses to assess the subcellular distribution and expression levels of LC3 and p62 autophagosome markers in PLN-R14del- versus PLN-WT-transfected cells.

Overexpression of PLN-R14del resulted in a punctate pattern of both LC3 and p62 proteins, indicative of autophagosome formation (Fig. 2a). In contrast, transfection with PLN-WT predominantly led to diffuse cytosolic localization of LC3 and p62. Quantitative analysis of LC3 puncta, which serve as a proxy for autophagosome count [31, 32], demonstrated that the number of LC3 puncta was significantly greater in cells expressing PLN-R14del than in those expressing PLN-WT (Fig. 2b). By immunoblotting, the expression levels of endogenous LC3-II, the lipidated form of LC3 that resides on the autophagosomal membrane [31–33], and p62 were found to be elevated in cells transfected with PLN-R14del relative to those transfected with PLN-WT (Fig. 2c-d). When wild-type and mutated PLN were coexpressed, the LC3-II and p62 levels remained unchanged relative to those in the PLN-R14del group (Fig. 2e-f), suggesting that the presence of PLN-R14del itself sufficed to cause autophagy impairment. The observed alterations in the distribution pattern of LC3 and p62 autophagy-related markers, along with their protein expression level changes, provide evidence to suggest impaired autophagy by PLN-R14del. Notably, in comparison to untransfected cells, overexpression of PLN-WT itself lead to relative minor alterations in autophagy parameters (Supplementary Fig. 1d-e), most likely due to cytosolic Ca²⁺ perturbation by PLN’s inhibitory effect on SERCA2.

Fig. 2.

PLN-R14del impairs autophagy. Immunofluorescence analysis to determine the subcellular localization pattern of LC3 and p62 (a) demonstrated a punctate distribution in GFP-PLN-R14del- but not in GFP-PLN-WT-transfected H9c2 cells, and quantification (b) of LC3 puncta determined a statistically significant increase in GFP-PLN-R14del cells (n = 80 cells from 4 independent experiments; *P < 0.01 vs. PLN-WT; t test, two-tailed). Scale bar 10 μm. Western blot analysis (c) and quantification (d) of the LC3 and p62 proteins revealed significant increases in the GFP-PLN-R14del- versus GFP-PLN-WT-transfected HEK293 cells indicative of impaired autophagy (n = 4 experiments, *P < 0.05 vs. PLN-WT, two-tailed t test) (e-f) Similar changes were observed between the GFP-PLN-R14del versus GFP-PLN-WT- and mCherry-PLN-R14del-cotransfected HEK293 cells. (n = 5 experiments, *P < 0.05, **P < 0.01, one-way ANOVA). The data are expressed as the mean ± SEM

Although p62 and LC3 immunostaining patterns and levels of expression are valuable indicators of autophagy, the dynamic nature of this process is better reflected by the assessment of autophagic degradation activity, the so-called “autophagic flux”, which denotes the relative changes in LC3-II levels that occur in a lysosome-dependent manner [31–34]. To assess autophagic flux, we treated PLN-WT- or PLN-R14del-transfected cells with the lysosomal inhibitor chloroquine (CQ), which prevents autophagosome fusion to lysosomes [35]. CQ treatment resulted in a significant change in the distribution and increased formation of LC3 puncta in both PLN-WT and PLN-R14del cells, compared to those in the corresponding untreated controls (Fig. 3a). Interestingly, however, the increase in the number of LC3 puncta upon CQ treatment was significantly lower in the PLN-R14del-transfected cells than in the PLN-WT-transfected cells (Fig. 3b). Similar findings were obtained by western blot analysis, which confirmed the anticipated increase in LC3-II protein levels upon CQ treatment in both WT- and R14del-transfected cells. However, this change appeared to be significantly lower in the lysates from PLN-R14del cells than in those from PLN-WT cells (Fig. 3c-d). Indeed, quantification of autophagic flux by means of the relative change in LC3-II levels before and after CQ treatment confirmed that the autophagic degradation activity of PLN-R14del was reduced (Fig. 3e).

Fig. 3.

PLN-R14del reduces autophagic flux. Immunofluorescence analysis (a) and quantification (b) of LC3 puncta in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells treated with CQ (n = 40–50 cells from 3 independent experiments; *P < 0.05, ***P < 0.001, one-way ANOVA). Scale bar 10 μm. Western blot analysis (c) and quantification (d) of LC3 protein in transfected HEK293 cells revealed significant differences in the GFP-PLN-R14del-transfected cells compared to the GFP-PLN-WT-transfected cells following CQ treatment. (n = 5 experiments, *P < 0.05, ***P < 0.001, one-way ANOVA). (e) Calculation of the difference in LC3-II levels in the presence or absence of CQ, detected in (c-d), determined that PLN-R14del significantly suppressed autophagic flux (n = 5 experiments, *P < 0.05 vs. PLN-WT; two-tailed t test). The data are expressed as the mean ± SEM

Disruption of autophagosome-lysosome fusion by PLN-R14del

To obtain further mechanistic insight into the disruptive effect of PLN-R14del on autophagy, we performed detailed subcellular analysis of autophagosome-lysosome fusion by dual staining with antibodies against LC3 and the lysosomal marker lysosomal-associated membrane protein 1 (LAMP1). In-depth spatial distribution analysis of LC3 and LAMP1 was performed using an object-based method that quantified their colocalization by determining the proportion of LC3 puncta positive for LAMP1 relative to the total LC3 puncta [30, 36, 37]. In line with previous reports [30, 35], CQ treatment reduced LC3-LAMP1 colocalization in untransfected H9c2 cells, compared to non-treated cells, while the opposite pattern was observed upon autophagy induction with rapamycin (Supplementary Fig. 2a-c). In PLN-R14del-transfected cells, the number of LC3 puncta that were positive for LAMP1 was lower both under basal conditions and upon CQ treatment, compared to PLN-WT-expressing cells (Fig. 4a-b and Supplementary Fig. 3). In addition, the number of LAMP1 puncta that were positive for GFP was lower in GFP-PLN-R14del cells than in GFP-PLN-WT, and the colocalization of PLN-R14del with lysosomes was reduced (Fig. 4c-d). These findings suggest that the decrease in autophagic flux induced by PLN-R14del expression may be due to reduced autophagosome–lysosome fusion.

Fig. 4.

PLN-R14del suppresses autophagosome-lysosome fusion. (a) Immunofluorescence analysis of LAMP1 and LC3 in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells treated with or without (basal) CQ. (b) Quantification of LC3 puncta positive for LAMP1 relative to total LC3 puncta and (c) LAMP1 puncta positive for GFP relative to total LAMP1 following CQ treatment in H9c2. Each point in the graph represents the relative value of each cell (n = 30–45 cells from 3 independent experiments; *P < 0.05 vs. PLN-WT; two-tailed t test). (d) Quantification of Pearson’s correlation coefficient was used to determine the reduction in the colocalization of GFP-PLN-R14del and LAMP1 upon CQ treatment in H9c2 (n = 40–50 cells from 3 independent experiments; *P < 0.01 vs. PLN-WT; two-tailed t test). (e) Immunofluorescence staining of GFP-PLN-WT- and mCherry-PLN-R14del-cotransfected H9c2 cells with LAMP1 was performed to examine the distribution of aggregates (white arrow) and lysosomes. The white line in the zoomed images indicates the region used for intensity profile plot analysis, which determined that aggregates are present adjacent to lysosomes, indicating blockade of lysosome fusion. The data are expressed as the mean ± SEM. Scale bar 10 μm

To determine whether this defect relates to aggregate formation, we stained wild-type and mutated PLN cotransfected cells with LAMP1 and examined the colocalization of the aggregates with the lysosomes. This analysis revealed that aggregates were preferentially present adjacent, rather than colocalized, to lysosomes (Fig. 4e). Thus, PLN-R14del appears to impact autophagy by inhibiting the fusion of autophagosomes with lysosomes.

Alterations in Ca2+ homeostasis as the mechanism underlying impaired autophagy in PLN-R14del cells

Autophagic flux and autophagosome-lysosome fusion are known to be influenced by various factors, including cytosolic Ca²⁺ levels [38–40]. Given the Ca²⁺ cycling defects observed in PLN-R14del animal models and patient-derived iPSC-CMs [12, 13, 17, 19], we examined whether this aberration may underlie impaired autophagy. To test this hypothesis, we treated cells with thapsigargin (TG), an inhibitor of SERCA and SR Ca²⁺ uptake, which elevates cytosolic Ca²⁺ levels (Supplementary Fig. 4a-d) and blocks autophagosome-lysosome fusion [41–43].

By immunofluorescence, we first confirmed that treatment of untransfected H9c2 cells with TG reduced LC3-LAMP1 colocalization despite the increase in the number of LC3 puncta, suggesting the inhibition of autophagosome-lysosome fusion (Supplementary Fig. 2d-g). Next, we exposed PLN-WT-transfected cells to TG and observed significant redistribution of LC3 to puncta, indicating that LC3 was incorporated into autophagosomes (Fig. 5a-b and d). However, quantification of the LC3 puncta positive for LAMP1 relative to the total LC3 puncta revealed that the number of LAMP1 and LC3 double-positive puncta per cell was reduced (Fig. 5b and e), denoting the inhibition of autophagosome–lysosome fusion. A similar staining pattern was observed for the PLN-R14del-transfected cells at baseline (Fig. 5a and c). Interestingly, unlike that in PLN-WT cells, the number of LAMP1 and LC3 double-positive puncta in PLN-R14del-transfected cells was not affected by TG. These findings suggest that TG inhibits autophagosome-lysosome fusion in PLN-WT but not in PLN-R14del cells.

Fig. 5.

Alterations in cellular Ca²⁺ levels suppress autophagosome-lysosome fusion. (a) Immunofluorescence analysis of LAMP1 and LC3 in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells treated with or without thapsigargin (TG). Overlay images of LAMP1 and LC3 proteins depict alterations in LAMP1/LC3 colocalization in the TG-treated GFP-PLN-WT cell as evident by an enhanced number of red puncta compared to non-treated cells. In the case of GFP-PLN-R14del, a considerable number of red puncta are seen under both basal and TG conditions. (b) Detailed distribution analysis of LC3 and LAMP1 was also performed by intensity profile plots generated for the region indicated by the white line. The black dashed lines indicate the peak LC3 intensities. In GFP-PLN-WT-transfected H9c2 cells, the LC3 and LAMP1 peaks appeared to overlap fully (red arrows) under basal conditions, suggesting their colocalization, while TG treatment led to a moderate shift in the peak’s codistribution (black arrows), indicating reduced LC3 and LAMP1 colocalization. (c) Among the GFP-PLN-R14del-transfected H9c2 cells, the colocalization of LC3 and LAMP1 was similar between the basal and TG-treated cells and lower than in the untreated GFP-PLN-WT cells. Quantification of LC3 puncta number (d) or LC3 puncta positive for LAMP1 relative to total LC3 puncta (e) in H9c2 transfected cells with or without TG treatment revealed significant alterations predominantly in the PLN-WT cells. Each point in the graph represents the relative value of each cell (n = 40–60 cells from 3 independent experiments; *P < 0.05, ***P < 0.001, one-way ANOVA). Scale bar 10 μm. Western blot (f) and quantification (g) analyses of LC3 protein levels in transfected HEK293 cells with or without TG treatment revealed a significant increase in LC3-II following TG treatment in GFP-PLN-WT cells (n = 5 experiments; *P < 0.05, one-way ANOVA). The data are expressed as the mean ± SEM

Additional evidence supporting this conclusion was provided by immunoblot analysis of LC3-II. As shown in Fig. 2c-d, the LC3-II level was greater in the PLN-R14del-transfected cells than in the PLN-WT-transfected cells. However, upon TG treatment, we observed a statistically significant increase in LC3-II in PLN-WT but not in PLN-R14del cells, where LC3-II was similar to that in untreated cultures (Fig. 5f-g). No major effect on cell survival was observed between wild-type and mutated PLN samples (Supplementary Fig. 4e). These findings indicate that TG treatment of PLN-WT cells recapitulates the autophagic defects observed in the PLN-R14del cells that are most likely associated with elevated cytosolic Ca²⁺ levels (Supplementary Fig. 4f-g). In support to this, treatment of PLN-R14del transfected cells with the Ca²⁺-chelating agent BAPTA appeared to reduce LC3-II protein levels and LC3 puncta formation indicating reversion of these autophagy defects (Supplementary Fig. 5).

To further validate the impact of cytosolic Ca²⁺ levels on autophagy, we performed another set of experiments in which Ca²⁺ aberrations were induced by RNAi-mediated knockdown of SERCA2a. Like TG, SERCA2a siRNA resulted in a significant increase in LC3-II levels, reduced autophagic flux and caused LC3 redistribution into puncta structures (Supplementary Fig. 6), pointing to impaired autophagy.

Collectively, the aforementioned data suggest that subcellular Ca²⁺ alterations in PLN-R14del cells suppress autophagy and contribute to the inhibition of autophagosome-lysosome fusion.

Reduced autophagosomal recruitment of the Rab7 and UVRAG proteins may contribute to suppressed autophagosome–lysosome fusion

Among the various proteins participating in the autophagic process, the small GTPase Rab7 regulates autophagosome-lysosome docking and fusion [38, 44–46]. Cytosolic Ca²⁺ alterations have been reported to prevent Rab7 incorporation into autophagosomes, resulting in autophagosome-lysosome fusion blockade [30, 42, 47]. On the basis of these reports, we reasoned that Rab7 may mediate the inhibition of autophagosome-lysosome fusion induced by PLN-R14del. Indeed, treatment of untransfected H9c2 cells with TG resulted in a significant decrease in the colocalization of Rab7 with LC3, confirming that Ca²⁺ alterations led to reduced recruitment of Rab7 to autophagosomes (Supplementary Fig. 7). In PLN-R14del-transfected H9c2 cells, the colocalization and number of LC3 puncta positive for Rab7 were significantly lower than those in PLN-WT-transfected cells (Fig. 6a-d and Supplementary Fig. 8). In a parallel set of experiments, TG treatment was used to confirm the Ca²⁺-dependent nature of these defects. As shown in Fig. 6b and d, the number of LC3 and Rab7 double-positive puncta decreased in TG-treated PLN-WT cells but not in PLN-R14del cells relative to their respective untreated controls. These findings suggest that recruitment of Rab7 to autophagosomes is inhibited in PLN-R14del cells, most likely due to cytosolic Ca²⁺ alterations.

Fig. 6.

Decreased recruitment of Rab7 to autophagosomes in PLN-R14del cells. (a) Immunofluorescence analysis of Rab7 and LC3 in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells with or without TG treatment. Detailed examination of the LC3 and Rab7 distributions was performed by intensity profile plots (b) generated from the region indicated by the white line. In GFP-PLN-WT-transfected H9c2 cells, the LC3 and Rab7 peaks appeared to overlap (red arrow) under basal conditions, suggesting their colocalization; however, TG treatment caused a shift in the peak’s codistribution (black arrows), indicating reduced colocalization. In GFP-PLN-R14del-transfected H9c2 cells, the colocalization of LC3 and Rab7 (black arrows) was reduced in both basal and TG-treated cells. Quantification of LC3 and Rab7 colocalization in H9c2 transfected cells by (c) Pearson’s correlation coefficient and (d) the number of LC3 puncta positive for Rab7 relative to the total number of LC3 puncta determined a significant reduction in response to TG in GFP-PLN-WT cells but not in GFP-PLN-R14del cells (n = 65–75 cells from 3 independent experiments; *P < 0.05, ***P < 0.001, one-way ANOVA). The data are expressed as the mean ± SEM. Scale bar 10 μm

Among several Rab7 effector proteins, ultraviolet radiation resistance-associated gene (UVRAG) mediates the recruitment of Rab7 to autophagosomes. UVRAG contains a C2 Ca²⁺-dependent lipid binding domain [48–50], which suggests functional sensitivity to Ca²⁺ alterations. We therefore investigated whether the co-distribution of UVRAG with autophagosomes may be altered by PLN-R14del. We initially examined the localization of UVRAG in untransfected H9c2 cells and observed a significant decrease in its colocalization with LC3 upon TG treatment, thus confirming that cytosolic Ca²⁺ changes impaired its recruitment to autophagosomes (Fig. 7a-d). Furthermore, we observed significantly less colocalization of UVRAG with LC3 in PLN-R14del-transfected cells than in PLN-WT-transfected cells (Fig. 7e-h and Supplementary Fig. 9a). Therefore, the incorporation of Rab7 and UVRAG into autophagosomal vesicles is reduced in PLN-R14del cells, a defect that is expected to impair the process of autophagosome‒lysosome membrane fusion.

Fig. 7.

Decreased UVRAG recruitment to autophagosomes in PLN-R14del cells. (a) Evaluation of the cellular distribution of UVRAG and LC3 in untransfected H9c2 cells under basal conditions or upon TG treatment. The white lines indicate the region used for (b) intensity profile plot analysis. TG treatment induced a considerable reduction in the colocalization of LC3 and UVRAG peaks (black arrows). Quantification of LC3-UVRAG colocalization by (c) Pearson’s correlation coefficient and (d) the number of LC3 puncta positive for UVRAG relative to the total number of LC3 puncta revealed significant decreases upon TG treatment (n = 32–40 cells from 2 independent experiments; *P < 0.01 vs. Basal, t test, two-tailed). Immunofluorescence (e) and intensity plot (f) analyses of UVRAG and LC3 in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells revealed reduced colocalization of LC3 and UVRAG puncta in GFP-PLN-R14del-transfected cells compared to GFP-PLN-WT-transfected cells. Quantification of (g) LC3-UVRAG colocalization in transfected H9c2 cells by Pearson’s correlation coefficient and (h) LC3 puncta positive for UVRAG relative to total LC3 puncta revealed a significant reduction in GFP-PLN-R14del cells compared to GFP-PLN-WT cells (n = 48–55 cells from 3 independent experiments; *P < 0.05 vs. PLN-WT; two-tailed t test). The data are expressed as the mean ± SEM. Scale bar 10 μm

Increased codistribution of the membrane tethering protein VPS41 with autophagosomes suggests enhanced autophagosome–lysosome docking in PLN-R14del cells

In addition to Rab7, soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins and membrane-tethering complexes participate in the coordinated process of autophagosome-lysosome fusion [49, 51, 52]. Having shown decreased incorporation of Rab7 and UVRAG into autophagosomes, we proceeded to examine the effects of PLN-R14del versus PLN-WT on the distribution of key SNAREs and membrane-tethering proteins. Syntaxin 17 (STX17) is a SNARE protein that is recruited to autophagosomes upon autophagy and has been proposed to play an essential role in driving the fusion of autophagosomes and lysosomal membranes [49, 50, 53]. We examined the distribution of STX17 and found no difference in its colocalization with LC3 between PLN-R14del and PLN-WT cells (Supplementary Fig. 10).

In addition to this SNARE component, we also examined potential alterations in proteins of the membrane-tethering complex. In particular, we focused on members of the homotypic fusion and vacuole protein sorting (HOPS) complex, which regulates membrane docking and fusion by bringing together opposing membranes and facilitating SNARE complex formation to drive membrane fusion [49, 50, 54, 55]. The HOPS complex is composed of multiple Vps-C core (VPS) proteins. We investigated VPS41, a protein associated with both the STX17 and Rab7 effector proteins [54, 56, 57], as a representative example of a membrane-tethering constituent. Interestingly, we observed increased colocalization of VPS41 with LC3 in PLN-R14del-transfected cells compared to PLN-WT-transfected cells, indicating increased formation of autophagosomal tethers (Fig. 8a-c and Supplementary Fig. 9b). A trend toward reduced VPS41 protein levels was also found in PLN-R14del-transfected cells (Supplementary Fig. 11), likely reflecting a compensatory mechanism toward alleviating the autophagic defect. Collectively, our data suggest that PLN-R14del expression promotes autophagosome-lysosome docking but that subsequent inhibition of autophagosome-lysosome membrane fusion prevents autophagic clearance.

Fig. 8.

Increased codistribution of the membrane tethering protein VPS41 with autophagosomes in PLN-R14del cells. Immunofluorescence (a) and intensity plot (b) analyses of VPS41 and LC3 in GFP-PLN-WT- or GFP-PLN-R14del-transfected H9c2 cells revealed enhanced colocalization of LC3 and VPS41 in GFP-PLN-R14del cells compared to GFP-PLN-WT cells. Quantification analysis of (c) LC3 puncta positive for VPS41 relative to total LC3 puncta in transfected H9c2 cells revealed a significant increase in GFP-PLN-R14del cells compared to that in PLN-WT cells (n = 62–63 cells from 2 independent experiments, *P < 0.05 vs. PLN-WT, two-tailed t test). The data are expressed as the mean ± SEM. Scale bar 10 μm

Discussion

The pathogenic effects of PLN-R14del have been attributed to structural changes affecting its interaction with SERCA2a and Ca²⁺ dysregulation [2]. Among the multiple downstream molecular and cellular consequences of this defect, the formation of protein aggregates is increasingly thought to play a central role. PLN-containing aggregates are a hallmark of human PLN-R14del cardiomyocytes and have been observed in all PLN-R14del patients examined and reported thus far [22, 58]. PLN aggregates are present at early stages of the disease, well before the onset of functional deficits in Pln-R14del mice [14]. However, the mechanism leading to their formation has been elusive, with sporadic reports describing colocalization of PLN aggregates with the autophagy markers LC3 and p62 [22, 23]. We provide here the first direct evidence for a causal relationship between PLN-R14del and autophagy dysregulation. Our findings unveil that PLN-R14del reduces autophagic flux through the inhibition of autophagosome-lysosome fusion, likely through the impairment of Ca²⁺ homeostasis, and ultimately leading to aggregate formation.

To maximize the flexibility of the molecular manipulations, we modeled aggregate formation in cultured cells. Since only heterozygous PLN-R14del patients have been reported so far, we initially evaluated the effect of wild-type and mutated PLN coexpression in order to mimic the heterozygous state in patients. This resulted in aggregates containing both proteins, suggesting that PLN-R14del has a dominant effect on the PLN-WT protein. Interestingly, and in agreement with observations in PLN-R14del homozygous mice [14, 15, 24], aggregates were observed even when PLN-R14del was transfected alone, indicating that its presence in itself suffices to induce aggregate formation. In line with findings from patient tissue [22, 23], we observed that PLN aggregates in cultured cell models also contained p62 and LC3. Furthermore, we showed increased p62 and LC3-II protein levels and reduced autophagic flux in PLN-R14del transfected cells. Collectively, these data suggests that autophagy defects are a potential mechanism of impaired protein degradation and aggregate formation by PLN-R14del.

Autophagy is a complex multistep process involving autophagosome formation and maturation, followed by fusion with lysosomes and protein degradation [59–63]. Several proteins, including members of three distinct families, namely, Rab GTPases, membrane-tethering complexes and SNARE proteins, are involved in the regulation of the autophagosome-lysosome fusion step [49, 51, 52]. It is believed that during the initial stages of this process, Rab7 associates with the autophagosomal and lysosomal membranes to fuse. In turn, through a coordinated process involving several Rab7 effector proteins, tethering complexes, such as HOPS, are recruited. The HOPS tethering complex consists of at least 6 subunits, and in addition to other tether proteins, it facilitates membrane docking by bringing together the two compartments to be fused. In cooperation with Rab7, tethering complexes assist SNARE proteins in driving the actual fusion of autophagosomes and lysosomal membranes [49, 51, 52]. The critical role of autophagosome-lysosome fusion in cellular physiology is evident by the multitude of pathological conditions caused by mutations in key proteins directly participating in this process [50, 64–72].

We provide multiple lines of evidence demonstrating that the autophagic defect caused by PLN-R14del expression is due to the inhibition of autophagosome-lysosome fusion. Our findings indicate increased autophagosomal tethers (as exemplified by enhanced autophagosomal recruitment of the tether protein VPS41) and, thus, enhanced autophagosome–lysosome docking in PLN-R14del cells. Since tethering precedes membrane fusion, the inhibitory effect of PLN-R14del on autophagosome-lysosome fusion could most likely be attributed to the inhibition of the final step of membrane fusion that completes this process. In support to this, by examination of aggregate distribution in relation to lysosomes, we observed that PLN aggregates were located adjacent to, but not colocalized with lysosomes. This finding supports aggregate accumulation in autophagosomes and reflects the inability of the latter to fuse with lysosomes. Collectively, these data indicate that inhibition of autophagosome-lysosome fusion by PLN-R14del leads to defects in autophagic clearance and aggregate formation.

Recently, aggregates from both PLN-R14del patients and mice were found to contain multiple SR proteins, including SERCA2 and HRC, in addition to PLN-R14del [73]. This is reminiscent of the established process of selective clearance of ER/SR subdomains containing aberrant material [74, 75]. Consistent with this notion, aggregate-containing cardiomyocytes in PLN-R14del patients exhibit significantly reduced SR localization of PLN [22]. In PLN-R14del mouse hearts, a progressive enhancement in aggregate size was recently associated with disease development, with small speckles being observed at early stages and accumulation into large clusters containing disorganized SR occurred upon disease progression [73]. Given the transient nature of PLN-R14del expression in our cell culture systems, the aggregates observed in our study may represent these early stage aggregates that appear smaller in size. Our findings combined with these reports point to autophagy impairment and subsequent aggregate formation as a driving force for the broader disorganization of the cardiac SR that has been reported in both PLN-R14del patients and mice, ultimately leading to cell death, replacement with fibrotic tissue and cardiac dysfunction [73].

The autophagy machinery is known to be distinctly regulated by Ca²⁺, with evidence suggesting a dual impact of intracellular Ca²⁺ on autophagy, enabling either its enhancement or suppression. This occurs as changes in cytosolic Ca²⁺ concentrations can interfere with autophagy at different regulatory checkpoints [39]. For example, pharmacologically triggered increases in cytosolic Ca²⁺ levels with thapsigargin are causally linked to the inhibition of autophagosome–lysosome fusion and reduced autophagic flux [38, 39]. In animal models (iPSC-CMs, mice and zebrafish) the presence of PLN-R14del has been shown to lead to alterations in Ca²⁺ handling [5, 13, 17, 19, 76]. The initially reported super-inhibitory effect of PLN-R14del on SERCA2 function that was observed in PLN-R14del overexpression models (HEK293 cells and mouse model), has recently been challenged by some experimental evidence from in vitro studies, patient-derived iPSC-CMs and a knock-in mouse model [12, 76–80], reporting partial or complete loss of SERCA2 inhibition by PLN-R14del. However, the underlying reason for this controversy remains to be elucidated. Nevertheless, all evidence to date indicate that PLN-R14del causes Ca²⁺-handling dysfunction due to SERCA2 dysregulation. In our experimental system cytosolic Ca²⁺ alterations were induced by thapsigargin treatment and showed that thapsigargin-induced increases in cytosolic Ca²⁺ levels contribute to autophagy impairment in PLN-WT-expressing cells. Specifically, these changes triggered alterations that recapitulated the autophagic defects observed with PLN-R14del, leading to impaired autophagosome-lysosome fusion. Multiple key players in autophagosome-lysosome fusion are Ca²⁺-dependent, and their Ca²⁺-binding moiety is necessary for promoting membrane fusion, as is the case for UVRAG [48, 81]. UVRAG has been reported to regulate autophagosome-lysosome fusion by modulating Rab7 activity [50, 51]. We demonstrated that the incorporation of UVRAG and Rab7 into autophagosomal vesicles is reduced in PLN-R14del-expressing cells, providing a likely explanation for the impairment of autophagosome‒lysosome membrane fusion caused by PLN-R14del. The effect of PLN-R14del-triggered Ca²⁺ perturbations on autophagic processes, could also be mediated by mitochondrial dysfunction [39]. Indeed, mitochondrial dysfunction was recently reported in PLN-R14del models [13, 76]. Since this study did not assess mitochondrial function per se, the role of mitochondria as a contributing factor in this PLN-R14del molecular pathological mechanism remains to be determined. Interestingly, alterations in Ca²⁺ homeostasis have been implicated in autophagosome–lysosome fusion impairment in multiple different pathological conditions. For example, in familial Alzheimer’s disease, the presenilin 2 mutation T122R decreases the ER Ca²⁺ concentration, inhibiting autophagosome-lysosome fusion and consequently blocking autophagic flux [30]; meanwhile, in Parkinson’s disease, buffering cytosolic Ca²⁺ to rescue the autophagy‒lysosome pathway and prevent neuronal cell death is under investigation as a promising therapeutic avenue [82]. Future studies will be needed to explore whether such an approach may hold promise in preventing autophagy impairment in PLN-R14del disease.

While heterologous systems, such as the HEK293 and H9c2 cell lines, are valuable and widely used to gain insight into complex pathogenetic mechanisms, they can be associated with certain limitations. These include their inherent differences in physiological parameters compared to mature cardiomyocytes (e.g. Ca²⁺ dynamics and contractility), the discordant levels of transfected proteins (e.g. PLN WT or R14del), and the transient nature of protein expression (e.g. PLN-R14del), which allows assessment of tentative acute but not chronic pathophysiological effects. However, despite these limitations, the use of these heterologous expression systems has provided valuable insights into PLN-R14del pathogenetic mechanisms, opening the way for targeted future follow-up.

In conclusion, we present evidence to support the occurrence of autophagic impairment in the presence of PLN-R14del, which is most likely associated with cytosolic Ca²⁺ alterations (Fig. 9). We demonstrated that PLN-R14del cells exhibit reduced autophagic flux, which we defined as occurring due to defects in autophagosome-lysosome fusion. This may be attributed to changes in the subcellular distribution of key components of the membrane fusion machinery, including decreased recruitment of UVRAG and Rab7 to autophagosomes and increased autophagosomal tethers that cause increased membrane docking and reduced autophagosome‒lysosome membrane fusion. The resulting aggregate formation may ultimately contribute to the SR disorganization, cardiomyocyte death, and cardiac fibrosis observed in PLN-R14del disease. Future studies in patient-derived cells and PLN-R14del animal models could evaluate whether strategies designed to prevent this defect could hold promise for the management of PLN-R14del disease.

Fig. 9.

Schematic representation of the proposed hypothesis on the molecular defects associated with impaired autophagy in PLN-R14del. PLN-WT cells normally undergo autophagy, which enables normal protein homeostasis. In contrast, in PLN-R14del cells, autophagy aberrations occur, most likely due to cytosolic Ca²⁺ alterations. This causes marked changes in the subcellular distribution patterns of key components of the membrane fusion machinery, including decreased recruitment of UVRAG and Rab7 to autophagosomes and increased autophagosomal tethers. These molecular aberrations are expected to promote increased membrane docking and reduce autophagosome-lysosome membrane fusion. This defect could ultimately contribute to impaired proteostasis and increased aggregate formation

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

EV designed the study, performed the experiments, analyzed the data and prepared the manuscript. EGK and AGE provided critical input and reviewed the manuscript. DS supervised the study, acquired funding and reviewed the manuscript. All the authors read and approved the final manuscript.

Funding

This research was funded by a grant from the Leducq Foundation for Cardiovascular Research (CURE-PLaN, 18CVD01).

Data availability

Data that support the findings of this study are available within the article and the supplementary materials.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors agree to the publication of this study.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest. A.G.E is co-founder and CSO of GENOSOPHY, a spin-off company of the National and Kapodistrian University of Athens.