Abstract

   Anthracyclines are chemotherapeutic drugs used to treat solid and
   hematologic malignancies. However, life‐threatening cardiotoxicity,
   with cardiac dilation and heart failure, is a drawback. A combination
   of in vivo for single cell/nucleus RNA sequencing and in vitro
   approaches is used to elucidate the underlying mechanism. Genetic
   depletion and pharmacological blocking peptides on phosphatidylinositol
   binding clathrin assembly (PICALM) are used to evaluate the role of
   PICALM in doxorubicin‐induced cardiotoxicity in vivo. Human heart
   tissue samples are used for verification. Patients with end‐stage heart
   failure and chemotherapy‐induced cardiotoxicity have thinner cell
   membranes compared to healthy controls do. Using the
   doxorubicin‐induced cardiotoxicity mice model, it is possible to
   replicate the corresponding phenotype in patients. Cellular changes in
   doxorubicin‐induced cardiotoxicity in mice, especially in
   cardiomyocytes, are identified using single cell/nucleus RNA
   sequencing. Picalm expression is upregulated only in cardiomyocytes
   with doxorubicin‐induced cardiotoxicity. Amyloid β‐peptide production
   is also increased after doxorubicin treatment, which leads to a greater
   increase in the membrane permeability of cardiomyocytes. Genetic
   depletion and pharmacological blocking peptides on Picalm reduce the
   generation of amyloid β‐peptide. This alleviates the
   doxorubicin‐induced cardiotoxicity in vitro and in vivo. In human heart
   tissue samples of patients with chemotherapy‐induced cardiotoxicity,
   PICALM, and amyloid β‐peptide are elevated as well.

   Keywords: amyloid β‐peptide, cardiomyocytes, cardiotoxicity
     __________________________________________________________________

   Single‐cell/nucleus transcriptional profiles of cardiac cells from
   doxorubicin(DOX)‐induced cardiotoxicity show that amyloid β‐peptide is
   generated in the heart tissue of patients and mice with
   anthracycline‐induced cardiotoxicity, and its expression is regulated
   by phosphatidylinositol binding clathrin assembly (PICALM). Genetic
   depletion and pharmacological blocking peptide on PICALM in vitro and
   vivo suppress DOX‐induced amyloid β‐peptide generation in
   cardiomyocytes and necrotic cell death.

   graphic file with name ADVS-11-2401945-g006.jpg

1. Introduction

   Anthracyclines are cytotoxic antibiotics that are commonly used to
   treat adult and childhood cancers. For example, doxorubicin (DOX)
   prolongs the survival of patients with cancer.^[ [40]^1 ^] However, the
   cardiotoxicity of anthracyclines is ≈10% 3.5 months after the final
   chemotherapy.^[ [41]^1 ^] Approximately 98% of patients have
   experienced cardiotoxicity after the first year of the anthracycline
   therapy.^[ [42]^1 ^] These cardiotoxic side effects limit the
   cumulative lifetime dose that a patient can receive.^[ [43]^2 ^]
   Current therapeutic strategies for DOX‐induced cardiotoxicity are
   mostly standard anti‐heart failure and anti‐arrhythmic therapies;
   however, their effectiveness is limited. To date, specific therapies
   capable of preventing DOX‐induced cardiotoxicity, especially in heart
   transplantation are still lacking, because of the mysterious
   mechanism.^[ [44]^3 ^]

   Impaired mitochondrial quality control processes are associated with
   mitochondrial dysfunction. This is the major mechanism underlying
   DOX‐induced cardiotoxicity.^[ [45]^4 ^] Preclinical evidence
   demonstrates that the modulation of mitochondrial quality control can
   be harnessed for therapeutic benefits against DOX‐induced
   cardiotoxicity.^[ [46]^5 ^] Amyloid β‐peptide (Aβ) is on the
   mitochondrial membranes, which blocks the transport of nuclear‐encoded
   mitochondrial proteins to mitochondria, interacts with mitochondrial
   proteins, disrupts the electron‐transport chain, increases reactive
   oxygen species (ROS) production, and causes mitochondrial damage.^[
   [47]^6 ^] Patients with heart failure experience an increase in Aβ
   levels in their hearts, suggesting that exposure to Aβ peptides impairs
   the cell viability of cardiomyocytes (CMs).^[ [48]^7 ^] However, it is
   unknown whether higher levels of Aβ peptides were observed during
   DOX‐induced cardiotoxicity. PICALM is the gene encoding
   phosphatidylinositol binding clathrin assembly (PICALM) protein.^[
   [49]^8 ^] It plays essential roles in endocytosis, internalization of
   cell receptors, and intracellular trafficking.^[ [50]^9 ^] However, the
   role of PICALM in DOX‐induced cardiotoxicity remains unclear.

   Given that increased Aβ peptides in the heart involve cardiac
   dysfunction,^[ [51]^7 , [52]^10 ^] in this study, we aimed to
   investigate a possible link among the Aβ peptides, DOX‐induced
   cardiotoxicity, and the pathogenic effect of PICALM. The findings are
   that PICALM mediates the generation of Aβ peptide in CMs on DOX‐induced
   cardiotoxicity, highlighting the potential of PICALM as an effective
   intervention for cardiotoxicity during anti‐cancer therapy.

2. Results

2.1. Clinicopathologic Examination of Patients with Cardiotoxicity Caused by
Chemotherapy

   Chemotherapy‐induced cardiotoxicity is a major concern in anti‐cancer
   treatment. It may cause severe heart failure.^[ [53]^11 ^] To
   investigate the pathogenesis of chemotherapy‐induced cardiotoxicity, we
   obtained six human heart samples from the heart transplantation center
   of Fuwai Hospital, including three patients with chemotherapy‐induced
   cardiotoxicity and three healthy donors (Figure [54] 1A; Table [55]S1,
   Supporting Information). All the patients were treated with
   anthracyclines such as DOX, idarubicin, mitoxantrone, and pharmorubicin
   (Figure [56]1A). Interestingly, all the patients received reiterative
   chemotherapy. We documented the natural history, chemotherapy
   strategies, and cardiac events from the first symptoms to heart
   transplantation for each individual as shown in Figure [57]1A. Patients
   with end‐stage heart failure and chemotherapy‐induced cardiotoxicity
   showed evidence of biventricular dilatation in cardiac magnetic
   resonance (CMR, Figure [58]1B). Late gadolinium enhancement (LGE)
   helped identify an enlarged left ventricle, decreased cardiac function,
   and increased cardiac fibrosis (Figure [59]1C). Cardiac wasting,
   fibrosis, and apoptosis are hallmarks of chemotherapy‐related
   cardiotoxicity.^[ [60]^12 ^] In our study, the structure and morphology
   of CMs were destroyed in patients with chemotherapy‐induced
   cardiotoxicity (Figure [61]1D). In patients with chemotherapy‐induced
   cardiotoxicity, there was a decrease in cardiomyocyte across area, an
   increase in cardiac fibrosis (Masson: 37.17% vs 15.23%, P < 0.01), and
   an increase in apoptosis (TUNEL‐positive nuclei: 6016.00 vs 361.00, P <
   0.05) (Figure [62]1E–G). In contrast to healthy controls, patients who
   received chemotherapy exhibited cardiac ultrastructural defects,
   including disrupted myofibrillar networks and mitochondria with
   disrupted cristae (Figure [63]1H). Furthermore, the thickness of the
   myocardial intercalated disc membrane was smaller (0.015 µm) than the
   thickness of the normal disc (0.030 µm, P < 0.001)^[ [64]^13 ^] and
   wide vacuolations were detected in patients with chemotherapy
   (Figure [65]1H).

Figure 1.

   Figure 1
   [66]Open in a new tab

   Histological performance of patients with chemotherapy‐induced
   cardiotoxicity. A) Natural history of the patients, from the diagnosis
   of malignancy to heart transplantation. B) Representative CMR image of
   patient No. 1, including LGE. C) Representative graphs of macro‐views
   of patients with chemotherapy‐induced cardiotoxicity. D) Representative
   graphs of hematoxylin and eosin (HE) staining of patients with
   chemotherapy‐induced cardiotoxicity. Scale bars, 2000 µm; scale bars
   for the magnified images, 50 µm. E–G) Representative and statistical
   graphs of patients with WGA/Masson/TUNEL‐stained chemotherapy‐induced
   cardiotoxicity. Scale bars, 2000 µm; scale bars for the magnified
   WGA/Masson/TUNEL‐stained images, 200/100/50 µm. H) Representative
   electron micrographs of cardiac muscles derived from the human heart.
   Scale bars, 1 µm; scale bars for the magnified images, 0.5 µm. NC,
   normal control; CICP, chemotherapy‐induced cardiotoxicity patient. n =
   3/group. Statistical analyses were performed using an unpaired t‐test.
   ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001. Data are presented as the
   mean ± SD.

2.2. Single‐Nuclei RNA Sequencing (snRNA‐seq) and Single‐Cell RNA sequencing
(scRNA‐seq) Revealing the Cardiac Cellular Landscape of DOX‐Induced
Cardiotoxicity Mice

   To define the cardiac cellular and transcriptional landscape of
   DOX‐induced cardiotoxicity, we established a model of DOX‐treated mice
   by intraperitoneal injection of DOX (10 mg kg^−1) weekly for 4 weeks to
   mimic cardiotoxicity in human patients, following previous reports^[
   [67]^14 ^] (Figure [68] 2A). DOX‐treated mice experienced a decrease in
   left ventricular ejection fraction (LVEF: 48.26% vs 59.51%, P < 0.05)
   and a decrease in fractional shortening (FS: 22.74% vs31.19%, P <
   0.05), indicating the DOX‐induced cardiac functional deterioration,
   consistent with human clinical data (Figures [69]1B and [70]2B). Body
   weight (BW) decreased more rapidly from 8 to 10 weeks than in the first
   two weeks in DOX‐treated mice, despite a constant increase in BW in the
   normal control (NC) group during the 4 weeks post‐DOX injection. This
   suggests that DOX had a cumulative effect on the target participants
   (Figure [71]2C). Compared to the normal group, the final BW and heart
   weight (HW) at 10 weeks of the DOX‐treated mice group were lower (BW:
   19.10 vs 25.28 g, P < 0.001; HW: 83.20 vs 122.00 mg, P < 0.001;
   Figure [72]2D); The area of fibrosis increased by ≈15% in DOX‐treated
   mice compared to normal group (Figure [73]2E,F); A significant
   decreased in across area of CMs of DOX‐treated mice (Figure [74]2G).
   The DOX‐treated group showed a significant increase in TUNEL‐positive
   nuclei (2258 vs 263.6, P < 0.01), indicating cardiomyocyte injury
   (Figure [75]2H). According to phenotypic features including pathology,
   the DOX‐induced cardiotoxicity mouse model could mimic the
   corresponding clinical phenotypes of cardiotoxicity in patients.^[
   [76]^15 ^]

Figure 2.

   Figure 2
   [77]Open in a new tab

   SnRNA‐seq and scRNA‐seq reveal the cellular landscape of DOX‐induced
   cardiotoxicity. A) Flowchart of our study including the experimental
   scheme of chronic cardiotoxicity induced by DOX, scRNA‐seq, snRNA‐seq,
   and analysis for data and experiment. B) Representative images and
   statistical graphs of cardiac echocardiography of mice injected with
   phosphate‐buffered saline (PBS) or DOX. EF: ejection fraction; FS:
   fractional shortening. C) The BW from weeks 6 to 10. D) Body and heart
   weight of DOX‐treated mice group on week 10. E,F) HE and changes in
   collagen volume fraction with and without DOX administration. Scan
   bars, 500 µm; scale bars for the magnified images, 50 µm. G)
   Representative and statistical graphs of WGA staining in DOX‐treated
   mice at the tissue level. Scan bars, 50 µm; scale bars for the
   magnified images, 20 µm. H) Representative and statistical graphs of
   TUNEL staining in DOX‐treated mice at the tissue level. Scan bars,
   500 µm; scale bars for the magnified images, 50 µm. I) UMAP of 18956
   nuclei, 24365 cells, and an integrated dataset combining snRNA‐seq and
   scRNA‐seq data after quality control and data filtering. J,K) Dot plots
   generated from the integrated dataset displaying characteristic marker
   genes of each identified cell population, and pie charts showing the
   proportion of cells within the snRNA‐seq, scRNA‐seq, and integrated
   datasets. NC, normal control (PBS‐treated mice); DOX, DOX‐treated mice.
   n = 5/group. Statistical analyses were performed using an unpaired
   t‐test. ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001. Data are presented
   as the mean ± SD.

   Murine myocardial samples from the DOX‐treated group and normal
   controls were processed for snRNA‐seq (n = 5) or scRNA‐seq (n = 5)
   (Experimental Section). Single‐nucleus and single‐cell libraries were
   sequenced, aligned to the mouse reference genome, and filtered for
   quality control. Unsupervised clustering, integration, and differential
   expression analyses were performed using the uniform manifold
   approximation and projection (UMAP, Figure [78]2I). After quality
   control, nuclei samples had average gene and feature counts per cell of
   1132 and 2626, respectively, whereas cell counts were 468 and 2025,
   respectively (Table [79]S2, Supporting Information). The final
   integrated dataset consisted of 18956 nuclei and 24365 cells
   representative of 8 major cell types (Figure [80]2J). Cell identities
   were validated by expression of cell‐specific marker genes and
   transcriptional signatures (Figure [81]2K). Cell types identified in
   both snRNA‐seq and scRNA‐seq datasets included fibroblasts (FBs,
   34.12%), endothelial cells (ECs, 23.20%), CMs (17.10%), mononuclear
   phagocytes (MPs, 10.55%), neutrophils (6.89%), mural cells (4.73%), T
   cells (2.28%) and B cells (1.12%) (Figure [82]2J). A notable benefit of
   the snRNA‐seq and scRNA‐seq combination was the capacity to obtain not
   only reads from additional cell types that were not efficiently
   recovered from enzymatically digested tissues, especially CMs but also
   to detect more genes and immune cells.

2.3. Upregulated Picalm Expression in CMs Phenotype Changings During
DOX‐Induced Cardiotoxicity

   6472 CMs were re‐clustered into five clusters using UMAP for further
   analysis (Figure [83] 3A). CM1‐2 percentage was similar between
   DOX‐treated mice and normal controls (CM1: 27.4% vs 24.0%; CM2:
   23.6% vs 26.9%); CM3‐4 percentage increased in DOX‐treated mice than
   that in normal controls (CM3: 23.1% vs 16.0%; CM4: 23.0% vs 16.0%),
   while CM5 percentage decreased in DOX‐treated mice (2.7% vs 17.0%,
   Figure [84]3A; Table [85]S3, Supporting Information). Five
   cardiomyocyte clusters with differing gene expression signatures were
   identified (Figure [86]3B; Figure [87]S1, Supporting Information). CMs
   from DOX‐treated mice exhibited high Myh7, Jph2, Mypn, Ppargc1a,
   Picalm, Actc1, Maml3, and Msrb3 expression in differentially expressed
   gene (DEG) analysis (Figure [88]3C). According to the Kyoto
   Encyclopedia of Genes and Genomes (KEGG) analysis of upregulated DEG in
   DOX‐treated mice, DOX greatly affected the cytoskeleton structure and
   contraction function of CMs, transitioning to the dilated
   cardiomyopathy (DCM) phenotype, which is consistent with human clinical
   data (Figure [89]3D; Table [90]S1, Supporting Information). The KEGG
   results showed 8 significantly upregulated pathways, including DCM,
   adrenergic signaling in cardiomyocytes, cardiac muscle contraction,
   gastric acid secretion, cGMP‐PKG signaling pathway, calcium signaling
   pathway, adherens junction, and arrhythmogenic right ventricular
   cardiomyopathy (Figure [91]3D; Table [92]S1, Supporting Information).
   Gastric acid secretion, cGMP‐PKG signaling pathway, and calcium
   signaling pathway were clustered together and were associated with
   apoptosis and cell death^[ [93]^16 ^] (Figure [94]3D). The results
   suggest that DOX induces CMs apoptosis, which is consistent with the
   increased number of apoptotic cells in the DOX group (Figures [95]1G
   and [96]2G).

Figure 3.

   Figure 3
   [97]Open in a new tab

   PICALM upregulation in CMs of DOX‐treated mice groups and patients with
   chemotherapy‐induced cardiotoxicity. A) UMAP plot for sub‐cluster
   analysis of CMs and cardiomyocyte cluster proportions in PBS‐ and
   DOX‐treated mice. B) Dot plots showing the markers for each cell
   subcluster. C) Heatmap displaying the top 20 upregulated and
   downregulated genes ranked by Log2 fold‐change between DOX and NC. D) A
   heat plot of the associations between upregulated genes and pathways
   determined by KEGG analysis. Squares indicate that the corresponding
   gene belongs to the corresponding pathway. E) Heat map showing dynamic
   changes in gene expression with pseudo‐time. The x‐coordinate from left
   to right represents the time from small to large, and the y‐coordinate
   represents the gene. Each point represents the expression of a specific
   gene at a specified pseudo‐time quantity (mean value). “Cluster” refers
   to the clustering of genes, clustering genes with similar or similar
   expression patterns into one cluster. F) Monocle dimension reduction
   diagram showing the expression of Picalm. G) Survival analysis of
   patients with breast cancer with high or low expression of PICALM in
   tumor samples according to TCGA database. H) Diagram showing the
   density distribution of different cell types over pseudo‐time. I) OMIHC
   representative and statistical graphs of PICALM expression in
   myocardial cells of DOX‐treated mice at different induction times at
   the tissue level. N > 25 CMs. Scan bars, 500 µm; scale bars for the
   magnified images, 20 µm. J) OMIHC representative and statistical graphs
   of PICALM expression in the myocardial cells of patients with
   chemotherapy‐induced cardiotoxicity. Scan bars, 2000 µm; scale bars for
   the magnified images, 200 µm. NC, normal control; DOX, DOX‐treated
   mice; CICP, chemotherapy‐induced cardiotoxicity patients. Mice: n =
   5/group; human: n = 3/group. Student's t‐test was applied to analyze
   the differences between two groups. Multiple‐group comparisons were
   made by one‐way ANOVA followed by the Tukey test. ^* P < 0.05; ^** P <
   0.01; ^*** P < 0.001. Data are presented as the mean ± SD.

   To explore the temporal relationship between CMs, we calculated the
   pseudo‐time values for each cardiomyocyte cluster to predict the
   putative state of cell differentiation using Monocle 2^[ [98]^17 ^]
   (Experimental Section). Pseudo‐time trajectory analysis identified
   three CM branches and five stages (Figure [99]3E; Figure [100]S2A,B,
   Supporting Information). Normal CMs were predicted to contain two
   highly differentiated cell stages (C and E), marked by Shroom3 and Cd36
   expression (Figure [101]3E; Figure [102]S2B,C, Supporting Information).
   In contrast, DOX‐treated mouse samples displayed two highly
   differentiated cell stages (A and B), marked by Myh7, Picalm, and Msrb3
   expression (Figure [103]3E,F; Figure [104]S2C, Supporting Information).
   It showed convergence toward disease‐associated cardiomyocyte
   phenotypes in DOX‐treated mice (Figure [105]3H; Figure [106]S2D,
   Supporting Information). Pseudo‐time reconstruction analysis of CMs
   revealed an increase in Picalm expression at the end of this trajectory
   (Figure [107]3F). According to common cancer (breast cancer, cervical
   cancer, liver cancer, lung adenocarcinoma, lung squamous cell
   carcinoma, ovarian cancer, and stomach cancer) survival analysis in The
   Cancer Genome Atlas (TCGA) database, higher expression levels of PICALM
   in tumor samples indicated shorter survival in breast and stomach
   cancer (Figure [108]3G; Figure [109]S3, Supporting Information). To
   identify this phenotype, the mean expression level of Picalm in single
   cell was upregulated in CM clusters from DOX‐treated mice compared to
   normal controls (P < 0.01) (Figure [110]S4A,B, Supporting Information).
   Moreover, Picalm expression level was higher in CM3 and CM4, whose
   percentages increased after DOX treatment (Figure [111]S4A, Supporting
   Information). Consistently, DOX promoted Picalm expression, which was
   2.79 times higher than controls with a significant difference (P <
   0.0001, Figure [112]S4C, Supporting Information). Furthermore, Picalm
   expression levels in CMs marked by TNNT2 in DOX‐treated mice were 3.65
   times higher than those in controls (Figure [113]S4D, Supporting
   Information). Immunofluorescence showed Picalm expression in the
   cytoplasm, close to the membrane (Figure [114]S4E, Supporting
   Information). With DOX treatment (Figure [115]S5, Supporting
   Information), both the percentage of TUNEL‐positive nuclei and
   expression level of Picalm in CMs marked by TNNT2 increased and
   correlated positively (Figure [116]3I). This suggests that increased
   Picalm expression may cause damage to the CM during DOX treatment.
   Further investigation is necessary to ascertain whether PICALM
   expression increases in patients with chemotherapy‐induced
   cardiotoxicity. Consistently, PICALM expression levels were
   significantly increased in patients with chemotherapy‐induced
   cardiotoxicity (Figure [117]S4F, Supporting Information). The opal
   multicolor immunohistochemistry (OMIHC) results showed that PICALM
   expression in the CMs of patients with chemotherapy‐induced
   cardiotoxicity increased by approximately two times (Figure [118]3J).
   Collectively, increased PICALM expression levels in
   chemotherapy‐induced cardiotoxicity in mouse and human samples were
   confirmed. These results indicated convergence toward
   disease‐associated cardiomyocyte phenotypes in DOX‐induced
   cardiotoxicity with upregulated expression of PICALM.

   Moreover, the scRNA‐seq datasets showed that DOX did not induce the
   expression of Picalm mRNA in FBs, ECs, MPs, neutrophils, mural cells, T
   cells, Schwann cells, or B cells (Figure [119]S6, Supporting
   Information). These results imply that DOX may play certain roles
   specific to CMs.

2.4. Amyloid β (1‐40) Peptide Upregulation in CMs of DOX‐Treated Mice Groups
and Patients with Chemotherapy‐Induced Cardiotoxicity

   Gene set analysis was used to identify the pathways involved in
   DOX‐treated mice.^[ [120]^18 ^] Aβ peptide production pathways such as
   amyloid β metabolism and amyloid precursor protein biosynthesis were
   activated in DOX‐treated mice, while degradation of Aβ peptide pathways
   such as lysosome and mTOR signaling pathways were inhibited in
   DOX‐treated mice (Figure [121] 4A; Figure [122]S7A,B, Supporting
   Information). The upregulation of Picalm expression was involved in the
   generation of Aβ peptide and the downregulation of Lpin1 and Gnptab was
   involved in the degradation of Aβ peptide process (Figure [123]4B,C;
   Figure [124]S7C, Supporting Information). Collectively, Aβ generation
   pathways and related genes were upregulated in CMs of DOX‐treated mice.

Figure 4.

   Figure 4
   [125]Open in a new tab

   Amyloid β metabolic process pathway and related genes were the major
   issue in DOX‐induced cardiotoxicity mice and patients with
   chemotherapy‐induced cardiotoxicity. A) Bar plot for scGSVA analysis of
   CMs. B) Heat plot of the associations between genes and pathways. C)
   Violin plots showing the expression of Lpin1/Psap/Gnptab/App in each
   cardiomyocyte cluster. D,E) Representative and statistical graphs of
   APP expression at the myocardial cells in DOX‐induced cardiotoxicity
   group and patients. Scan bars, 50 µm; scale bars for the magnified
   images, 20 µm. F) Diagram showing the potential mechanisms of DOX
   induction in CMs. G,H) Representative and statistical graphs of Aβ40,
   PICALM expression at the CMs marked by TNNT2 in DOX‐induced
   cardiotoxicity mice, and a diagram showing the correlation of PICALM
   and Aβ40 in mice CMs. Scan bars, 1000 µm; scale bars for the magnified
   images, 200 µm. I,J) Representative and statistical graphs of Aβ40,
   PICALM expression at the CMs marked by TNNT2 in patients with
   chemotherapy‐induced cardiotoxicity, and a diagram showing the
   correlation of PICALM and Aβ40 in human CMs. Scan bars, 500 µm; scale
   bars for the magnified images, 50 µm. K) ELISA analysis of mouse serum
   Aβ40 level in DOX‐induced cardiotoxicity. L) ELISA analysis of human
   serum Aβ40 level in chemotherapy‐induced cardiotoxicity and plasma Aβ40
   levels could distinguish CICP from non‐CICP (AUC = 0.94, P = 0.03). NC,
   normal control; DOX, DOX‐treated mice; CICP, chemotherapy‐induced
   cardiotoxicity patients. Mice: n = 5/group in RNA‐seq and n = 8/group
   in immunoimaging; human: n = 3/group. Student's t‐test was applied to
   analyze the differences between two groups. Multiple‐group comparisons
   were made by one‐way ANOVA followed by the Tukey test. ^* P < 0.05; ^**
   P < 0.01; ^*** P < 0.001. Data are presented as the mean ± SD.

   Amyloid precursor protein (App) is a transmembrane precursor protein.^[
   [126]^19 ^] Its levels were higher in both the DOX‐treated mice and
   patients. It appears from the data that Aβ can be enhanced by
   anthracyclines therapy (Figure [127]4A‐E; Figure [128]S7C, Supporting
   Information). Based on the gene expression levels of Aβ metabolism, we
   inferred that Picalm induced CMs to produce Aβ during DOX treatment.
   The upregulation of Picalm facilitates clathrin‐dependent endocytosis,
   which results in early endosomal enlargement. The mTOR‐mediated
   autophagy defect and lysosome dysfunction aggravated early endosomal
   enlargement and increased Aβ production (Figure [129]4F).

   Aβ has two forms, amyloid β (1‐40, Aβ40) peptide and amyloid β (1‐42,
   Aβ42) peptide. Both of which contribute into the pathogenesis of
   Alzheimer's disease (AD) and cardiovascular disease.^[ [130]^20 ^] To
   elucidate which amyloid β‐peptide was the major form in DOX‐induced
   cardiotoxicity, we analyzed the expression of Aβ40 and Aβ42 in tissues
   and serum. The results showed that Aβ and PICALM colocalize by OMIHC
   staining in both murine models and patients. In DOX‐treated mice, the
   PICALM, Aβ40, and Aβ42 expression levels in CMs were more increased
   than the normal controls (Figure [131]4G,H; Figure [132]S7D, Supporting
   Information). The expression level of PICALM correlated positively with
   Aβ40 (R^2 = 0.9355, P < 0.0001, Figure [133]4H) and Aβ42 (R^2 = 0.7184,
   P < 0.0001, Figure [134]S7D, Supporting Information). The correlation
   coefficient between PICALM and Aβ40 (R^2 = 0.9355) was higher than that
   between PICALM and Aβ42 (R^2 = 0.7184). The data suggest that Aβ40
   rather than Aβ42 may be the main target affected by Picalm. The line
   profile showed that more than half of the signals from Aβ on the cells
   and the PICALM signals were co‐located (Figure [135]S7E, Supporting
   Information). In the patients with chemotherapy‐induced cardiotoxicity,
   Aβ40 was also highly expressed in CMs in patients, while Aβ42 was
   similar in two groups (Figure [136]4I,J; Figure [137]S7F, Supporting
   Information). It indicates that Aβ40 is the major form in patients with
   chemotherapy‐induced cardiotoxicity. In addition, the expression levels
   of Aβ40 in CMs were positively correlated with the expression levels of
   PICALM in human heart samples (R^2 = 0.9894, P < 0.0001,
   Figure [138]4J). The Aβ42 expression was also correlated positively
   with PICALM (R^2 = 0.9375, P = 0.0015, Figure [139]S7F, Supporting
   Information). Further, the concentration of Aβ40 in DOX‐treated mouse
   serum was increased than in normal mice according to the enzyme‐linked
   immunosorbent assay (ELISA) experiments (Figure [140]4K). The
   corresponding ELISA experiment was conducted with the serum obtained
   from patients with chemotherapy‐induced cardiotoxicity. The amounts of
   Aβ40 in the serum of patients with chemotherapy‐induced cardiotoxicity
   were higher than those in healthy controls (Figure [141]4L;
   Table [142]S1, Supporting Information). Moreover, the serum Aβ40
   quantity was able to discriminate between cardiotoxicity and
   non‐cardiotoxicity populations including healthy participants and DCM
   patients with an area under curve (AUC) of 0.94 (95% CI: 0.79‐1.00,
   Figure [143]4L; Figure [144]S7G, Supporting Information). Collectively,
   Picalm can regulate the generation of Aβ40 peptide in CMs during
   anthracycline therapy. It may lead to cardiotoxicity. The serum Aβ40
   peptide is a biomarker for cardiotoxicity in patients.

2.5. Picalm Regulating the Generation of Aβ Peptide in CMs

   A cell model induced by DOX was built to verify this hypothesis and
   further elucidate the mechanism (Figure [145]4F). To recapitulate the
   many features of human CM physiology, a human induced pluripotent stem
   cell‐derived cardiomyocytes (hiPSC‐CMs)‐based model of
   anthracycline‐induced cardiotoxicity was included (Figure [146]S8,
   Supporting Information). In the first step, cells were treated with DOX
   at different concentrations. We found that 1 µm DOX could reduce the
   viability of cardiomyocyte cells (hiPSC‐CMs/H9c2/HL‐1). However, this
   concentration of DOX had no significant effect on the viability of
   non‐cardiomyocyte cells such as mouse embryo fibroblasts (NIH3T3), rat
   cardiac fibroblasts (RCFs), human umbilical vein endothelial cells
   (HUVECs), and a macrophages line (Raw246.7). These data confirmed the
   effects of DOX on CMs (Figure [147] 5A; Figure [148]S9A, Supporting
   Information). Additionally, 1 µM DOX could significantly reduce the
   activity of a breast cancer cell line (MCF‐7) and a stomach cell line
   (NCI‐N87); however, not on a cervical cancer cell line (HeLa)
   (Figure [149]5A). These results are consistent with the efficacy of DOX
   in clinical practice, where DOX has been shown to be more effective
   against breast and stomach cancers than cervical cancers.^[ [150]^21 ^]
   Then, we monitored how treatment with 1 µM DOX affected cell living and
   proliferation. IncuCyte microscopic images showed that treatment with
   DOX reduced cell proliferation to the midpoint of the treatment period
   and affected the number of cells to reach the LogIC50 (Figure [151]S9B,
   Supporting Information).

Figure 5.

   Figure 5
   [152]Open in a new tab

   Si‐Picalm can alleviate DOX‐induced hiPSC‐CMs damage and Aβ peptide
   production. A) Cell viability in different groups. Cells were treated
   with the indicated concentration of DOX (0–10 µm). After incubation,
   cell viability was measured by CCK8 assay. B,C) Statistical analysis of
   vesicle area, PICALM and Aβ40 level of hiPSC‐CMs treated with 1 µM DOX
   for 0, 12, 24, and 36 h with confocal microscope (Zeiss) by using
   ImageJ plugins.^[ [153]^23 ^] D) The statistical graphs determined
   using an IncuCyte imaging system. H9c2 cells pre‐stained with 0.5 µM
   LysoTracker Red for 30 min were treated with 1 µm DOX for 24 h. E) The
   statistical graphs of LPIN1 expression determined by confocal. F) The
   WGA staining of DOX‐treated cells and statistical graphs of vesicles
   area. G) Representative images of ROS and calcium concentration of
   hiPSC‐CMs treated with DOX or Aβ40 and statistical analysis. Scan bars,
   50 µm. H) Representative images of propidium iodide (PI)‐staining of
   hiPSC‐CMs treated with DOX or Aβ40 and statistical analysis of
   apoptosis and necrotic cell percentage. Scan bars, 200 µm. I)
   Statistical analysis of Aβ40 level of hiPSC‐CMs using ELISA in the
   medium. J) The western blot of the hiPSC‐CMs groups including SI‐NC,
   SI‐PICALM, DOX+SI‐NC, DOX+SI‐PICALM, and statistical graphs of PICALM,
   APP, and Aβ40 expression. K,L) Representative images and statistical
   analysis of WGA/LPIN1 staining in the hiPSC‐CMs groups. Scan bars,
   200 µm. M) Representative images and statistical analysis of lysosome
   tracking in the hiPSC‐CMs groups. Scan bars, 50 µm. N) The cell
   viability in groups. O,P) Representative images and statistical
   analysis of TUNEL assay in the hiPSC‐CMs groups. Scan bars, 200 µm.
   DOX, DOX‐treated cells; SI‐NC, SI‐GFP; OE‐NC, overexpression vector.
   Student's t‐test was applied to analyze the differences between two
   groups. Multiple‐group comparisons were made by one‐way ANOVA followed
   by the Tukey test. ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001. Data are
   presented as the mean ± SD.

   To monitor the impact of DOX treatment on the production of PICALM and
   Aβ40 peptide, we selected four timepoints. We found that induction of
   DOX resulted in increased vesicle area, higher expression of PICALM,
   and higher expression of Aβ40 using confocal microscope
   (Figure [154]5B,C; Figure [155]S9C, Supporting Information). These
   results further confirmed that DOX could induce the expression of
   PICALM and generation of Aβ40. In terms of signaling pathways
   (Figure [156]4F), DOX promoted lysosome mass reduction (Figure [157]5D;
   Figure [158]S9D, Supporting Information),^[ [159]^22 ^] LPIN1
   expression level decreased (Figure [160]5E), and vesicle area increased
   (Figure [161]5F). The data suggests that DOX inhibits lysosome and
   autophagy pathways, resulting in vesicles secreting more Aβ40. We
   compared the effects of Aβ40 and DOX on CMs to better understand the
   impact of Aβ40 on DOX‐induced cardiotoxicity. DOX and Aβ40 led to an
   increase in ROS and calcium ions (Figure [162]5G; Figure [163]S10A,
   Supporting Information), in which free radicals can destroy the
   cytoskeleton (Figure [164]S10B, Supporting Information), cell membrane
   functions (Figure [165]S10C, Supporting Information, and organelle
   functions (Figure [166]5F; Figure [167]S10D, Supporting Information),
   and promote an increase in membrane permeability and cause cell death
   (Figure [168]5H).

   To examine the effect of PICALM on Aβ40 peptide production, we knocked
   down and overexpressed Picalm in hiPSC‐CMs. The expression of the Aβ40
   peptide was assessed using ELISA (Figure [169]S10E, Supporting
   Information). PICALM knockdown resulted in a decrease in Aβ40 levels in
   cellular supernatant, while overexpression of PICALM resulted in
   increasing in Aβ40 levels (Figure [170]5I). To test whether PICALM
   knockdown could salvage DOX‐induced cell damage, knockdowns of the
   PICALM in hiPSC‐CMs and H9c2 cells were performed and then treated with
   DOX. The PICALM and Aβ peptide expression in the DOX group were
   compared with the control group. The si‐Picalm significantly reduced
   the DOX‐induced Aβ40 peptide production, PICALM expression, and APP
   expression (Figure [171]5J; Figure [172]S10E–I, Supporting
   Information). We then analyzed the effect of DOX on the basis of
   knockout of PICALM on the lysosome and autophagy signaling pathways,
   and found that inhibition of Picalm expression could alleviate
   DOX‐induced dysfunction of the lysosome, autophagy, and organelles
   (Figure [173]5K–M). Furthermore, the CCK8 assay showed that si‐Picalm
   alleviated DOX‐induced damage (Figure [174]5N). To investigate whether
   overexpression of Picalm worsens DOX‐induced cardiotoxicity, hiPSC‐CMs
   were transfected with plasmids that overexpressed PICALM, followed by
   DOX induction (Figure [175]S10E, Supporting Information). TUNEL assay
   showed overexpression of PICALM exacerbated DOX‐induced toxicity
   (Figure [176]5O,P). Collectively, Picalm can generate Aβ peptide in
   CMs, and Aβ peptide, in turn, causes cardiotoxicity. In addition, we
   explored whether PICALM knockout affected the effectiveness of DOX in
   oncology treatment. PICALM knockout was performed in MCF‐7 and NCI‐N87,
   followed by DOX treatment in both tumor cell lines. These results
   indicate that PICALM knockout has no effect on DOX treatment
   (Figure [177]S11, Supporting Information).

2.6. Genetic Depletion of Picalm Attenuating DOX‐Induced Cardiotoxicity In
Vivo

   Conventional Picalm knock out mice were used to investigate the role of
   Picalm in the pathogenesis of DOX‐induced cardiotoxicity
   (Figure [178]S12, Supporting Information). Three groups were designed,
   including wild‐type mice (n = 5), wildtype with DOX treatment (n = 5),
   and Picalm ^+/− mice with DOX treatment (n = 6) as shown in
   Figure [179] 6A. Picalm ^+/− mice had longer survival time than the
   wild‐type mice under the same DOX treatment (P = 0.0319)
   (Figure [180]6B). Genetic depletion of Picalm rescued the cardiac
   functions of DOX‐induced cardiotoxicity, in which EF and FS in Picalm
   ^+/− mice increased significantly compared to those in DOX‐treated mice
   (Figure [181]6C).

Figure 6.

   Figure 6
   [182]Open in a new tab

   Genetic depletion of PICALM attenuated DOX‐induced cardiotoxicity in
   vivo. A) The treatment experiment design. B) Survival analysis of the
   three groups. C) Representative and statistical graphs of EF and FS
   evaluated by echocardiography of normal control, DOX induction on
   wild‐type mice and DOX induction of Picalm knockout mice. D,E)
   Representative and statistical graphs of HE and Masson staining of
   normal control, DOX induction on wild‐type mice and DOX induction of
   Picalm knockout mice. F) The IHC images of TUNEL of normal control, DOX
   induction on wild‐type mice and DOX induction of Picalm knockout mice.
   G) Representative and statistical graphs of WGA, Aβ40, PICALM
   expression at the CMs marked by TNNT2 in normal control, DOX induction
   on wild‐type mice and DOX induction of Picalm knockout mice. NC, normal
   control, n = 5; DOX, DOX‐treated mice, n = 5; DOX+KO, DOX‐treated
   Picalm knockout mice, n = 6. Student's t‐test was applied to analyze
   the differences between two groups. Multiple‐group comparisons were
   made by one‐way ANOVA followed by the Tukey test. ^* P < 0.05; ^** P <
   0.01; ^*** P < 0.001. Data are presented as the mean ± SD.

   A higher across area of CMs than that in the DOX group, and the lower
   fibrosis abundant in the heart of the Picalm ^+/− group confirmed the
   role of PICALM in DOX‐induced cardiotoxicity (Figure [183]6D,E).
   Further, apoptosis of CMs decreased in the Picalm ^+/− group
   (Figure [184]6F). Aβ40 peptide in CM expression also decreased
   (Figure [185]6G). Collectively, these data suggest that the genetic
   depletion of Picalm attenuates DOX‐induced cardiotoxicity in vivo.

2.7. Anti‐PICALM Antibody Blocking Peptides (BPs) Attenuating DOX‐Induced
Myocardial Damage and Decreasing Aβ40 Peptide Production In Vivo

   Genetic depletion of Picalm expression could attenuate DOX‐induced
   myocardial damage and Aβ peptide production in vitro and in vivo. The
   animal model has been constructed to investigate the effects of
   anti‐PICALM antibody (50 µg per mouse) by intramyocardial injection.
   This led to the development of new clinical applications (Figure [186]
   7A). Administration of the anti‐PICALM antibody rescued cardiac
   function in mice with DOX‐induced cardiotoxicity. EF and FS were
   noticeably higher in mice treated with the anti‐PICALM antibody than
   that in those treated with DOX (Figure [187]7B). Histological analysis
   revealed a greater CM area than that in the DOX group. Lower levels of
   fibrosis in the heart of the DOX anti‐PICALM antibody group confirmed
   the effectiveness of anti‐PICALM antibody therapy (Figure [188]7C,D).
   Furthermore, anti‐PICALM antibody decreased Cs apoptosis
   (Figure [189]7E,G). PICALM expression decreased in mice treated with
   the anti‐PICALM antibody (Figure [190]7F,G). Aβ40 peptide in CM
   expression also decreased after anti‐PICALM antibody treatment
   (Figure [191]7F,G). Collectively, these results suggest that
   anti‐PICALM antibodies can attenuate the cardiotoxicity caused by DOX
   and decrease the production of Aβ40 in vivo.

Figure 7.

   Figure 7
   [192]Open in a new tab

   DOX‐induced cardiotoxicity is suppressed by anti‐PICALM antibody BP. A)
   Treatment experiment design. B) Representative and statistical graphs
   of EF and FS evaluated by echocardiography in the control, DOX, and
   DOX + PICALM BP group. C,D) Representative and statistical graphs of HE
   and Masson staining of the control, DOX, and DOX + PICALM BP group. E)
   IHC images of TUNEL of the control, DOX, and DOX + PICALM BP group. F)
   Representative graphs of WGA, Aβ40, PICALM expression at the CMs marked
   by TNNT2 in the control, DOX, and DOX + PICALM BP group. G) Statistical
   graphs of TUNEL‐positive cells, and PICALM, Aβ40 expression at the CMs
   marked by TNNT2 in the control group, DOX, and DOX + PICALM BP group.
   DOX, DOX‐treated mice. Mice: n = 4/group. Student's t‐test was applied
   to analyze the differences between two groups. Multiple‐group
   comparisons were made by one‐way ANOVA followed by the Tukey test. ^*
   P < 0.05; ^** P < 0.01; ^*** P < 0.001. Data are presented as the
   mean ± SD.

3. Discussion

   Chemotherapy has been one of the most effective treatment modalities
   for cancer since its emergence in the 1940s. However, growing concerns
   have been raised regarding its cardiotoxic effects.^[ [193]^24 ^] Owing
   to limited research on its cellular mechanism, there are no efficient
   therapeutic strategies available for chemotherapy‐associated
   cardiotoxicity.^[ [194]^25 ^] Understanding the specific cellular
   regulatory processes of cardiotoxicity under diseased conditions is of
   fundamental importance for successful drug development, such as for
   developing drugs for the treatment of DOX‐induced cardiotoxicity.^[
   [195]^15 ^] In this study, we investigated the pathological
   characteristics of DOX‐induced cardiotoxicity and established a mouse
   model of chronic DOX‐induced cardiotoxicity. Cells and nuclei from the
   cardiac tissue of the mouse model were sequenced using snRNA‐seq and
   scRNA‐seq. Specific changes in the expression profiles, subpopulation
   composition, and intercellular communication in cardiac tissues of the
   mouse model were defined, especially in the CMs. By comparing gene
   expression profiles from normal participants, we showed the enrichment
   of markers of CM stress induced‐cardiotoxicity in the hearts. Based on
   this heterogeneity, we clustered the transcriptionally related cells or
   genes, each of which was likely related to different cellular
   functions. In addition, the data showed that the transcriptomic states
   of CMs in human hearts were continuous rather than discrete. Previous
   studies on cardiac tissues have reported a change at the tissue level,
   but not single cell levels. To the best of our knowledge, this is the
   first report of a single nucleus/cell study of the cardiac tissues of a
   mouse model of chronic DOX‐induced cardiotoxicity.

   Here, we report a significant stimulative effect of Picalm on
   DOX‐induced cardiotoxicity in rodent and cardiomyocytes (hiPSC‐CMs) in
   vitro and in a DOX‐induced cardiotoxicity mouse model. The data showed
   an interaction between Picalm and Aβ40 peptide, which played an
   essential role in the induction of cardiomyocyte apoptosis (TOC).
   Picalm knockout mice and iPSC‐CM experiments results indicated that
   Picalm could promote the generation of Aβ peptide, and then led to
   DOX‐induced cardiotoxicity. This intervention efficiently prevented
   DOX‐induced cardiotoxicity, highlighting the therapeutic value of
   Picalm. In addition, PICALM knockdown did not affect the therapeutic
   effects in breast and stomach cancers in vitro. TCGA data indicated
   that low PICALM expression was associated with a better prognosis.
   PICALM‐BP may be a novel treatment for anthracycline‐induced
   cardiotoxicity in the clinic and does not affect the effectiveness of
   DOX in tumor treatment. Collectively, PICALM may be a novel target for
   the treatment of anthracycline‐induced cardiotoxicity in clinical
   settings. However, drugs targeting Picalm are not currently available.
   Further investigations are required to develop new drugs for clinical
   use.

   Aβ peptide is produced through the proteolytic processing of a
   transmembrane protein, APP, by β‐ and γ‐secretases. Aβ peptide
   accumulation in the brain is proposed to be an early toxic event in the
   pathogenesis of AD, which is the most common form of dementia
   associated with plaques and tangles in the brain.^[ [196]^26 ^] Aβ
   peptide accumulation has also been found in the skeletal muscles of
   patients affected by sporadic inclusion‐body myositis and
   atherosclerotic aortas.^[ [197]^27 ^] In addition, the β‐amyloid
   peptide was observed in heart failure, which contributes to cardiac
   dysfunction via decreasing the cell viability in both ECs and CMs.^[
   [198]^7a ^] There were no reports about roles of regarding the
   β‐amyloid peptide in DOX‐induced cardiotoxicity, although it was
   reported that DOX‐treated rats significantly exhibited AD‐like brain
   injury, increased amyloid burden, enhanced neuroinflammation and
   apoptosis, and multifocal histological injury in the cerebral cortex
   with widespread vacuolations.^[ [199]^28 ^] This present study is the
   first to reveal that DOX induces amyloid generation, causing
   cardiotoxicity. Also, the relationship between Aβ and membrane
   permeability has been reported. Recent advances indicate that these
   interactions are a clear picture of the Aβ induced membrane
   disruption.^[ [200]^29 ^] The serum Aβ40 levels could be applied as a
   biomarker for discriminating cardiotoxicity patients (Figure [201]4G).
   This could serve as a diagnostic reference in the absence of an
   effective diagnostic strategy. This needs a larger cohort to confirm
   the effectiveness of serum Aβ40 levels as a specific biomarker. Taken
   together, our data suggest a possible mechanism of DOX‐induced
   cardiotoxicity. DOX upregulated the expression of Picalm, which induces
   the accumulation of Aβ to induce membrane disruption to result in the
   disturbance of cellular homeostasis. Combining our findings and
   previous studies, β‐amyloid peptide participated in the process of
   cardiovascular disease.^[ [202]^30 ^] This may indicate whether there
   was a connection between the brain and the heart, an emerging and
   essential topic that requires additional investigation.^[ [203]^31 ^]

   A limitation of our study was the lack of a mechanism by which Picalm
   expression was regulated by DOX. Further investigations on this point
   should be conducted in the future. In addition, the sample size of
   human cardiac tissues is rather small because of difficulties in
   obtaining them. All heart samples in this study were obtained from
   female patients in this study. Thus, these findings should be further
   verified in male patients.

   In summary, we performed a comprehensive cellular analysis of
   DOX‐induced cardiotoxicity. This study primarily focused on CMs. We
   identified Picalm as the hallmark gene. It was responsible for the
   generation of Aβ in CMs, which increased membrane permeability and
   induced cardiotoxicity in both human tissues and animal models. A
   potential therapy strategy for patients with end‐stage heart failure
   caused by chemotherapy could be targeting Picalm.

4. Experimental Section

Model Building and Grouping and Picalm Knockout Mice Construction

   Five‐week‐old wild‐type male C57BL/6 mice (Vital River Laboratories,
   Beijing, China) were maintained in a specific pathogen‐free facility
   and provided with a normal diet and drinking water. And Picalm‐KO mice
   (C57BL/6Smoc‐Picalm^em1Smoc, Cat. NO. NM‐KO‐200880) were purchased from
   the Shanghai Model Organisms Center, Inc. This study was approved by
   the Animal Ethics Committee of the Fuwai Hospital (FW‐2022‐0015). The
   guidelines of Directive 2010/63/EU of the European Parliament on the
   protection of animals used for scientific purposes were followed in
   this animal experiment. The animals were randomly divided into two
   groups: NC and DOX group. After 1 week of adaptive feeding, the DOX
   group was administered by weekly intraperitoneal injections of
   10 mg kg^−1 of DOX hydrochloride (Selleck, S1208) for four consecutive
   weeks. The NC group was intraperitoneally injected with PBS at the same
   time. The weight of the mice was recorded weekly. After 4 weeks of
   injection, the mice were examined using echocardiography (VisualSonics,
   Vevo2100 system).

hiPSC‐CM Culture

   Human CMs were differentiated from human induced pluripotent stem cells
   based on an established protocol using an activator and an inhibitor of
   the Wnt signaling pathway sequentially and matured using maturation
   media supplemented with L‐lactate, L‐ascorbic acid, and recombinant
   human albumin.^[ [204]^32 ^] Human CMs were characterized by
   immunostaining with TNNT2 antibodies and the percentage of TNNT2^+
   cells was counted and calculated based on the images (Figure [205]S12A,
   Supporting Information). Over 90% of differentiated cells were TNNT2^+.
   Similar results were observed using flow cytometry characterization and
   ≈97% of cells were TNNT2^+ (Figure [206]S12B, Supporting Information).
   These cells are suitable models for mimicking the response of human CMs
   to chemical or genetic perturbations. Cell differentiation was
   supported by Beijing Vercelldo Technological Lab (part of Beijing
   Wosida Xibao Jishu Company Limited).

   Cells were grown in growth media (Cor.4U Complete Culture Medium which
   contains 10% fetal bovine serum (FBS) and other essential nutrients,
   Catalog number Ax‐M‐HC250 from Ncardia) and treated in minimal media
   (BMCC Serum‐Free Culture Medium, Catalog Number Ax‐MBMCC250 from
   Ncardia, supplemented with 1% fetal bovine serum from Thermo Fisher
   Scientific) in a humidified incubator with 5% CO[2] and 37 °C.
   Cryo‐preserved hiPSC‐CMs were thawed in 75‐cm^2 flasks, cultured in
   growth media for 3 days, and reseeded into multi‐well plates. Then,
   hiPSC‐CMs were cultured in minimal media for 1.5 days before drug
   treatment and maintained in minimal media during drug treatment with
   media exchanged every other day.

Cell Culture

   H9c2 (2‐1) (ATCC CRL‐1446), non‐cardiomyocyte cells such as NIH3T3,
   RCF, HUVEC, and Raw246.7, and tumor cells line (MCF‐7 cells, NCI‐N87
   cells, and HeLa cells) were cultured in Dulbecco's Modified Eagle
   Medium (DMEM, Gibco) supplemented with 10% FBS in 5% CO[2] at 37 °C.
   HL‐1 (ATCC CRL‐12197) was cultured in Minimum Essential Medium (MEM,
   Gibco) supplemented with 10% FBS in 5% CO[2] at 37 °C. When the cell
   density reached ≈80%, the cells were digested with 0.25% trypsin every
   two to three days.

Human Heart Sample Collection

   The use of human tissue in this study was approved by the Human Ethics
   Committee of Fuwai Hospital, Chinese Academy of Medical Sciences (No.
   2021‐1465). Written informed consent was obtained from each patient.
   Human heart samples were collected from patients who underwent heart
   transplantation in the operating room. The patient who experienced
   DOX‐induced cardiotoxicity (n = 3). Healthy heart samples (n = 3) were
   obtained from brain‐dead donors with a normal circulatory supply who
   were not suitable for transplantation for technical or noncardiac
   reasons, such as BW mismatch, according to the guidelines of the China
   Transplant Services. Detailed information is provided in Table [207]S1
   (Supporting Information).

ScRNA‐seq

   Following histological examination, five hearts from each group were
   pooled to isolate single cells. The tissues were cut into small pieces
   in PBS and digested with 200 U/ml collagenase type II (Worthington,
   43J14367B) in a 37 °C water bath with mild shaking. The single‐cell
   suspension was filtered with a 40 µm cell strainer and collected by
   centrifugation at 400 g for 5 min. The supernatant was discarded, and
   the cell pellets were resuspended in 1 mL DMEM containing 10% FBS.^[
   [208]^33 ^] Isolated non‐myocytic cells from each heart were counted
   and used for single‐cell transcriptomic library preparation on the
   Chromium controller (10X Genomics). Cells were loaded into each channel
   and processed using the Chromium Single Cell 3′ v1 reagent kit.
   Following capture and lysis, the synthesis and amplification were
   conducted for 14 cycles according to the manufacturer's protocol. The
   amplified cDNA was used to construct Illumina sequencing libraries,
   each sequenced in one lane of the Illumina HiSeq X ten.

SnRNA‐seq

   Nucleus was extracted from five heart tissues in each group. After
   removing from −80 °C, the tissue was dissected in cold PBS. The
   shredded samples were then transferred to the 30 mL lysis solution and
   stirred for 20 s. The stirred liquid was homogenized in the Dounce
   tissue grinder set (pestle B) for 10 strokes. The homogenate was passed
   through 100 and 70 µm filters successively. After centrifugation at
   700 g for 10 min at 4 °C, the supernatant was discarded, and the pellet
   was resuspended using 23 mL preservation solution. The remaining 7 mL
   of preservation solution was poured into a centrifuge tube, followed by
   slow pouring of 23 mL of the resuspension into the centrifuge tube and
   centrifugation for 1 h at 13 000 rpm. After centrifugation, the
   supernatant was discarded, and the pellet was resuspended using 1%
   bovine serum albumin.^[ [209]^34 ^] The nuclei were counted using
   Trypan blue and further processed following the manufacturer's protocol
   for 10x Genomics. The snRNA‐seq libraries were prepared using the
   Chromium Next GEM Single Cell 3′ GEM, Library & Gel Bead Kit, version
   3.1, 16 reactions and sequenced using the Illumina HiSeq X ten.

Integration of Single‐Cell and Single‐Nuclei Datasets

   Scanpy v1.8.2^[ [210]^35 ^] was used for quality control,
   dimensionality reduction, and clustering using Python 3.7. For each
   sample dataset, the expression matrix was filtered using the following
   criteria: 1) cells with gene count <200 or with top 2% gene count were
   excluded; 2) cells with top 2% UMI count were excluded; 3) cells with
   mitochondrial content > 20% were excluded; 4) genes expressed in less
   than five cells were excluded. After filtering, 42 764 cells were
   retained for the downstream analysis, with on average 1274 genes and
   3243 UMIs per cell. The raw count matrix was normalized to total counts
   per cell and logarithmically transformed into normalized data matrix.
   Top 2000 variable genes were selected by setting flavor = “seurat”.
   Principle Component Analysis (PCA) was performed on the scaled variable
   gene matrix, and top 20 principal components were used for clustering
   and dimensional reduction. Cells were separated into 25 clusters by
   using Louvain algorithm and setting resolution parameter at 1.2. Cell
   clusters were visualized by using UMAP.

Data Preprocessing and Cluster Identification

   Raw reads from RNA‐seq were processed to generate gene expression
   matrixes using CeleScope
   ([211]https://github.com/singleron‐RD/CeleScope) v1.9.0 pipeline.
   Briefly, raw reads were first processed with CeleScope to remove low
   quality reads with Cutadapt v1.17 to trim poly‐A tail and adapter
   sequences.^[ [212]^36 ^] Cell barcode and UMI were extracted. After
   that, STAR v2.6.1a was used to map reads to the reference genome GRCm38
   (ensemble version 92 annotation).^[ [213]^37 ^] UMI counts and gene
   counts of each cell were acquired using featureCounts v2.0.1
   software,^[ [214]^38 ^] and used to generate expression matrix files
   for subsequent analysis.

   Cells were filtered using UMI counts of >30 000, gene counts <200 or
   >5000, and mitochondrial content >20%. After filtering, cells were
   retained for the downstream analyses. Functions from Seurat v3.1.2 for
   dimension‐reduction and clustering were used.^[ [215]^39 ^] Then
   NormalizeData and ScaleData functions were used to normalize and scale
   all gene expressions and the top 2000 variable genes with
   FindVariableFeautres function were selected for PCA analysis. Using the
   top 20 principal components, cells were separated into multiple
   clusters with FindClusters. Batch effect between samples was removed by
   Harmony.^[ [216]^40 ^] Finally, UMAP algorithm was applied to visualize
   cells in a 2D space.

DEG

   The Seurat FindMarkers function based on Wilcox likelihood‐ratio test
   with default parameters was used, and the genes expressed in >10% of
   the cells in a cluster and with an average Log (Fold Change) value
   >0.25 as DEGs were selected. For the cell type annotation of each
   cluster, the expression of canonical markers found in the DEGs with
   knowledge from literatures were combined and displayed the expression
   of markers of each cell type with heatmaps/dot plots/violin plots that
   were generated with Seurat function. Doublet cells were identified as
   expressing markers for different cell types and removed manually.

Constructing Trajectory in Pseudotime

   Cell differentiation trajectory was reconstructed with Monocle2.^[
   [217]^41 ^] Highly variable genes (HVGs) were used to sort cells in
   order of spatial‐temporal differentiation. DDRTree was used to perform
   FindVairableFeatures and dimension‐reduction. Finally, the trajectory
   was visualized by plot_cell_trajectory function. Next, CytoTRACE (a
   computational method that predicts the differentiation state of cells
   from single‐cell RNA‐sequencing data using gene Counts and Expression)
   was used to predict the differentiation potential of monocyte
   subpopulations.^[ [218]^42 ^]

Pathway Enrichment Analysis

   To investigate the potential functions of DEGs, genes were screened
   that differed significantly (LogFoldchange>0.25, P < 0.05) between
   groups or clusters. Afterward, to investigate the potential functions
   of these different expression genes, the Gene Ontology (GO) and KEGG
   analysis were used with the “clusterProfile” R packager.^[ [219]^43 ^]
   Pathways with P_adj value <0.05 were considered as significantly
   enriched. Gene Ontology gene sets including molecular function (MF),
   biological process (BP), and cellular component (CC) categories were
   used as reference.

   GSEA was performed on genes in cell clusters. For GSVA pathway
   enrichment analysis, the average gene expression of each cell type was
   used as input data using the GSVA package.^[ [220]^44 ^]

   Protein‐protein interactions (PPI) of DEGs in each cluster were
   predicted based on known interactions of genes with relevant GO terms
   in the StringDB v1.22.0.

Pathological and Electron Microscopy Analysis

   After cardiac perfusion, the hearts were removed quickly from the mice,
   wiped gently with a clean absorbent paper and cut horizontally in half.
   Half of the heart tissues were taken using paraformaldehyde (V900894,
   Merck), and the remaining tissues were stored at −80 °C.

   Hearts were dehydrated and subsequently paraffin‐embedded and then
   sectioned at 4 µm thickness for pathological analysis.
   Hematoxylin‐eosin (HE), Masson's trichrome (ab150686, Abcam), TUNEL
   (C1091, C1089, Beyotime; 11 684 795 910, Roche), Sirius Red (S8060,
   Solarbio), and WGA (1:200, W834, Thermo) staining were performed
   according to standard protocols. The slides were analyzed using an
   automatic digital slide‐scanning system (Axio Scan Z1, Zeiss). For
   electron microscopy, hearts were fixed in 4% paraformaldehyde, 1.5%
   glutaraldehyde in 0.1 mol L^−1 Cacodylate buffer. The hearts were
   subjected to immersion fixation in the same fixative and embedded in
   epoxy resin following standard protocols.

IHC

   The heart sections were repaired with antigen, blocked, and incubated
   with primary antibody overnight at 4 °C (1:500, Rabbit Anti‐PICALM
   antibody, bs‐11665R, Bioss). The secondary antibody (PV‐6000, Zsgb Bio)
   conjugated to horseradish peroxidase (HRP) was then incubated at room
   temperature for 1 h. A DAB kit (ZLI‐9019, Zsgb Bio) was used according
   to the manufacturer's instructions. Whole slides were scanned with an
   automatic digital slide‐scanning system (Axio Scan Z1, Zeiss). The IHC
   analysis was used to detect PICALM protein expression in the heart
   tissue.

OMIHC

   To analyze the correlation between Aβ and PICALM was observed in situ,
   the following protocol was followed: step 1: slide preparation and
   dewaxing; step 2: microwave treatment for epitope retrieval; step3:
   blocking; step 4: primary antibody incubation (1:500 anti‐PICALM,
   bs‐11665R, Bioss; 1:1000 anti‐TNNT2, bs‐10648R, Bioss; 1:500 anti‐
   Aβ1‐40, bs‐0106 M, Bioss; 1:500 anti‐Aβ1‐42, bs‐0107R, Bioss); step 5:
   introduction of secondary HRP conjugated secondary antibody; step 6:
   incubation of opal fluorophore; step 7: microwave treatment for
   antibody stripping; step 8: using DAPI (1:300) and WGA (1:200) for
   staining. Steps 3–7 were repeated until all targets of interest were
   detected using different Opal fluorophores (DAPI‐405, WGA‐488,
   PICALM‐540, TNNT2‐620, Aβ1‐40‐570, Aβ1‐42‐650). The slides were
   analyzed using the Vectra Quantitative 14 Pathology Imaging System
   (PerkinElmer).

ELISA

   To detect trace levels of Aβ protein in the samples, a more sensitive
   ELISA kit (E‐EL‐M3009/E‐EL‐H0542c, Elabscience) was selected to
   determine Aβ1‐40 in mice, human, and hiPSC‐CMs. Duplicate readings were
   averaged for each standard and sample, and then the average zero
   standard optical density was subtracted. A four‐parameter Logistic
   curve was plotted on the log‐log axis, with standard concentration on
   the x‐axis and OD values on the y‐axis.

Transfection of Cells with siRNAs and Plasmids

   The expression of PICALM was depleted using siRNA as previously
   described.^[ [221]^45 ^] Briefly, cells were seeded in a 6 well plate
   at a density of 150 000 cells well^−1 24 h before treatment. All
   procedures were performed in Opti‐MEM in without serum. Complexes were
   pre‐formed between the oligonucleotides and oligofectamine (Invitrogen)
   using 50 pmol oligonucleotide per 4 µl oligofectamine in a total volume
   of 200 µl Opti‐MEM. The complex was subsequently added to cells to give
   a final oligonucleotide concentration of 50 nm in a total of 1 mL
   Opti‐MEM. Cells were incubated with media, siRNA for green fluorescent
   protein (GFP) to control for off‐target effects of oligofectamine and
   oligonucleotide. 500 µl of Opti‐MEM supplemented with 12% (v/v) FBS
   were then added directly to the transfection mixture and cells were
   then incubated for 48 h and processed as required. The expression of
   PICALM was overexpressed by adding the amount of CALM‐pmCherry‐N1
   plasmid (27 691, addgene) that was twice that of normal protein
   expression during transfection.

Immunofluorescence (IF) Analysis

   To determine the localization and expression of protein in cardiac
   tissue, an IF analysis of heart section was performed. After antigen
   repair, closure, and permeability, the primary antibody (anti‐PICALM;
   anti‐TNNT; anti‐LPIN1, bs‐0759R, bioss; anti‐APP, bs‐0112 M, bioss;
   anti‐Aβ40) incubation, sections were incubated with a secondary
   antibody with Alexa Fluor Dye (Thermo Fisher Scientific). Finally, DAPI
   and WGA (L4895, Sigma, Shanghai, China) staining was used to express
   the nuclei and cell membranes.

Cell Counting Kit‐8 (CCK‐8) Assay

   Cells inoculated in a 96‐well plate were incubated with DOX at
   different time points, and then left untreated or treated with
   si‐PICALM for the next day. Then, CCK‐8 reagent (CCK8‐500, Hanbio,
   Shanghai, China) was added to the cells in the dark. After incubation
   at 37 °C for 3 h, the plate was detected at 450 nm using a microplate
   reader.

IncuCyte Imaging Analysis

   H9c2 cells pre‐stained with 5 µm carboxyfluorescein succinimidyl ester
   (CFSE, HY‐D0938, MCE)/10 µm IncuCyte CytoLight Rapid Red Reagent (4706,
   Sartorius) for 30 min were seed at ≈70% density into a 96‐well plate in
   triplicate. H9c2 cells pre‐stained with 0.5 µm LysoTracker Red
   (HY‐D1300, MCE) for 30 min were seeded at ≈70% density in a 96‐well
   plate in triplicate. The cells of the plates were treated with 1 µm DOX
   and placed into the IncuCyte imaging system (IncuCyte S3) for 24 h.

Western Blot (WB)

   RIPA lysis buffer (P0013B, Beyotime) and Cocktail (4 693 132 001,
   Roche) were added to the samples, which were then homogenized on ice.
   The supernatant was removed and denatured using a protein loading
   buffer after centrifugation. The protein was transferred to a
   nitrocellulose membrane (IB23001, Thermo Fisher) after electrophoresis,
   blocked with skim milk powder, and incubated with the following primary
   antibodies overnight: anti‐APP (ab32136, Abcam), anti‐PICALM,
   anti‐Aβ1‐40 (ab254345, Abcam), and anti‐Aβ1‐42. The next day, the
   corresponding secondary antibody was used for continuous incubation,
   and the images were collected and processed after exposure to enhanced
   chemiluminescence (ECL), GAPDH was used as an internal control.
   ECL‐plus reagent (GE Healthcare) was used to analyze the protein
   expression level.

Hochest, PI, Rhod‐4, FITC‐Phalloidin, 2,7‐Dichlorodihydrofluorescein
Diacetate (DCFH‐DA), and FITC‐Dextran/DiI ‐LDL Staining

   To visually display the vesicle size, α‐actin, living cells, dead
   cells, ROS, Ca^2+, and cell permeability, 20 µg mL^−1 WGA (L4895,
   Sigma, Shanghai, China), 20 µg mL^−1 FITC‐phalloidin(ab235137, Abcam,
   Shanghai, China), 20 µg mL^−1 Hochest 33 342 (875756‐97‐1, MERCK),
   10 µg mL^−1 PI (HY‐D0815, MCE), 5 µm DCFH‐DA (D6883, Sigma, Shanghai,
   China), 5 µM Rhod‐4 (T0404, Warbio, Nanjing, China), and 0.1 mg mL^−1
   FITC‐Dextran (HY‐128868D, MCE)/40 µg mL^−1 DiI‐LDL(20614ES76, Yeasen)
   were used to stain for 30 min in accordance with the instructions and
   then washed with PBS.

TCGA Data Acquisition and Survival Analysis

   Bulk RNA‐seq transcriptome data and clinical information were
   downloaded from TCGA database ([222]https://portal.gdc.cancer.gov/).
   Simultaneously, the clinical information and follow‐up data of the
   patients were also downloaded. Before conducting the survival analysis,
   the expression level of PICALM was extracted and used to rank the
   patients in descending order. Subsequently, the patients were
   categorized into two groups based on the median value of PICALM
   expression: a high PICALM expression group and a low PICALM expression
   group. Finally, differences in overall survival and progression‐free
   interval between high and low groups were compared using Kaplan‐Meier
   curves, with P‐values calculated via Log‐rank test, using the Survival
   package in R.

Statistical Analysis

   The images and WB results were analyzed using ImageJ software. Protein
   expression by IHC was determined using the plugin of the IHC Toolbox in
   ImageJ. GraphPad software was used for statistical analysis. All
   continuous variables are described as the mean ± SD. Student's t‐test
   was used to analyze the differences between two groups. Multiple‐group
   comparisons were performed using a by one‐way analysis of variance
   followed by the Tukey's test. P < 0.05 indicated a statistically
   significant difference. ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   M.B, X.H., X.C., and T.A. contributed equally to this work. M.B., X.H.,
   and J.S. performed conceptualization; X.C. and T.A. acquired resources;
   M.B. and X.H. wrote the original draft; H.M., Z.S., M.T., and G.Y.
   wrote, reviewed, and edited the manuscript; X.H. and J.S. acquired
   funding; and J.S. performed supervision.

Supporting information

   Supporting Information
   [223]ADVS-11-2401945-s003.docx^ (11.3MB, docx)

   Supplemental Table 1
   [224]ADVS-11-2401945-s002.xlsx^ (11.7KB, xlsx)

   Supplemental Table 2
   [225]ADVS-11-2401945-s004.xlsx^ (11.3KB, xlsx)

   Supplemental Table 3
   [226]ADVS-11-2401945-s001.xlsx^ (10.7KB, xlsx)

Acknowledgements