Abstract

   Cadmium (Cd) exposure is strongly linked to various diseases and
   dysregulation of autophagy is a pivotal mechanism in Cd toxicity.
   Targeted autophagy strategies are promising for the treatment of
   autophagy dysregulation‐related diseases, including Cd poisoning.
   However, the current understanding of autophagy mechanisms remains
   limited, hindering the development of effective strategies. Herein, a
   novel autophagy pathway, transcellular autophagy. Cd triggers this
   process in hepatocytes, facilitating the transfer of autophagosomes
   from damaged to healthy cells for degradation. Mechanistically,
   reactive oxygen species accumulation is a key driver of transcellular
   autophagy activation, while the disruption of autophagy fusion
   mechanisms triggers its activation. Notably, Cd‐induced transcellular
   autophagy relies on the tumor necrosis factor, alpha‐induced protein 2
   (TNFAIP2)‐tunneling nanotube (TNT) system. Blocking this system
   prevents autophagosome transfer and exacerbates Cd‐induced
   autophagosome overload and apoptosis. The findings offer a novel
   perspective on autophagy, and provide new insights for targeted
   autophagy strategies to treat Cd poisoning and autophagy
   dysregulation‐related diseases.

   Keywords: autophagosome transfer, cadmium, TNFAIP2, transcellular
   autophagy, tunneling nanotubes
     __________________________________________________________________

   ROS accumulation is an important upstream signal for activating
   transcellular autophagy. Incomplete autophagy and TNT formation are
   necessary conditions for activating transcellular autophagy, and the
   disruption of the autophagic fusion mechanism is a trigger for
   activating transcellular autophagy. Cd activates transcellular
   autophagy and promotes the transfer of autophagosomes from Cd‐damaged
   cells to healthy ones for fusion and degradation.

   graphic file with name ADVS-12-e02793-g006.jpg

1. Introduction

   Pollution is a leading environmental risk factor for disease and
   premature mortality worldwide.^[ [50]^1 ^] Cadmium (Cd) is a
   significant environmental pollutant that poses a serious threat to
   human and animal health. Cd accumulates in multiple organs throughout
   the food chain and is closely related to a range of diseases, including
   neurodegenerative, bone, liver, and kidney diseases, and various types
   of cancer.^[ [51]^2 , [52]^3 , [53]^4 , [54]^5 , [55]^6 ^] Notably, the
   liver accounts for ≈30% of Cd accumulation in the body, making it one
   of the organs most severely affected by Cd toxicity.^[ [56]^7 ^]
   However, an incomplete understanding of the mechanisms underlying Cd
   toxicity remains the primary hurdle in the development of targeted
   protective products.

   Autophagy is a cellular quality control mechanism that targets various
   intracellular components for degradation, and plays a crucial role in
   promoting cell survival and restoring cellular homeostasis.^[ [57]^8 ,
   [58]^9 ^] Given the significance of autophagy in disease progression,
   targeting it is considered a promising and effective treatment approach
   with immense potential.^[ [59]^10 , [60]^11 ^] Selective autophagy
   inhibitors targeting the autophagy process have been developed for
   cancer treatment.^[ [61]^12 ^] For instance, the inhibitors,
   chloroquine and hydroxychloroquine, which act at the autophagy fusion
   stage, have undergone clinical trials for cancer treatment.^[ [62]^13 ,
   [63]^14 ^] These inhibitors enhance the sensitivity of cancer cells to
   chemotherapeutic drugs, thereby potentiating their therapeutic effects.
   Notably, autophagy dysregulation is an important cytotoxic mechanism of
   Cd.^[ [64]^15 ^] Targeting autophagy also has substantial therapeutic
   potential for the prevention and treatment of Cd poisoning. Cd induces
   incomplete autophagy by blocking autophagic flux, ultimately leading to
   cell damage.^[ [65]^16 , [66]^17 , [67]^18 , [68]^19 ^] Based on this
   mechanism, several drugs, such as puerarin, taurine, quercetin,
   melatonin, paeonol, and trehalose alleviate Cd‐induced cell damage by
   restoring autophagic flux.^[ [69]^20 , [70]^21 , [71]^22 , [72]^23 ,
   [73]^24 , [74]^25 , [75]^26 ^] Therefore, targeting the regulation of
   autophagic flux is currently the main research direction of targeted
   autophagy strategies for the prevention and treatment of Cd poisoning.
   However, the current understanding of autophagy mechanisms remains
   limited, thereby hampering the further development of effective
   targeted autophagy strategies for the treatment of diseases associated
   with autophagy dysregulation.

   Multicellular organisms maintain intercellular communication through
   various mechanisms that are essential for cellular and tissue
   homeostasis.^[ [76]^27 ^] Cells located far apart can exchange signals
   and substances via secreted signaling molecules or exosomes, whereas
   adjacent cells can communicate directly through gap junctions.^[
   [77]^28 , [78]^29 , [79]^30 ^] Our previous study revealed that
   Cd‐exposed hepatocytes could induce bystander effects via the
   transmission of substances or signals between cells through gap
   junction intercellular communication.^[ [80]^31 ^] Nevertheless, gap
   junction channels are inherently limited in length, making it
   challenging to observe substances transferred within these channels
   using current experimental techniques. Tunneling nanotubes (TNTs)
   represent a novel form of direct intercellular communication capable of
   achieving long‐distance direct communication between cells.^[ [81]^32
   ^] TNTs offer a distinctive open membrane channel for intercellular
   substance exchange, with diameters spanning from tens to hundreds of nm
   and lengths extending up to 10–100 µm or longer.^[ [82]^32 , [83]^33 ^]
   TNTs enable intercellular transfer of substances, including small
   molecules, protein aggregates, various organelles, and pathogens.^[
   [84]^34 , [85]^35 , [86]^36 ^] The transfer of organelles between cells
   plays an important role in maintaining cellular homeostasis, regulating
   cellular functions, and responding to external stimuli.^[ [87]^37 ,
   [88]^38 ^] Mitochondria are the main organelles responsible for ATP
   production, and mitochondrial quality control is crucial for
   maintaining mitochondrial function and cellular homeostasis.^[ [89]^39
   ^] Recently, mitochondrial transfer, a form of intercellular signaling,
   is widely studied. Intercellular mitochondrial transfer plays an
   important role in mitochondrial quality control and maintaining tissue
   homeostasis.^[ [90]^40 ^] Our recent study reported the role of
   mitochondrial transfer in Cd‐induced nonalcoholic fatty liver
   disease.^[ [91]^41 ^] Unexpectedly, during the research process, we
   found that autophagosomes can also transfer between hepatocytes.
   Notably, intercellular autophagosome transfer occurs both between cells
   of the same type, such as human squamous cell carcinoma cells, and
   cells of different types, such as from B‐cell precursor acute
   lymphoblastic leukemia cells to mesenchymal stromal cells (MSCs).^[
   [92]^42 , [93]^43 ^] However, there are currently very limited reports
   on autophagosome transfer. Existing studies primarily focus on
   reporting the phenomenon, while the underlying mechanisms and
   functional significance of this process remain elusive. Notably, a
   recent study reported the transfer of lysosomes from healthy podocytes
   to advanced glycation end product (AGE)‐damaged podocytes, whereas
   autophagosomes were transferred from AGE‐damaged to healthy
   podocytes.^[ [94]^37 ^] This suggests that there is a potential for
   transcellular degradation of autophagosomes, which could represent a
   novel protective mechanism to alleviate the pressure caused by
   autophagosome overload.

   The specific pathways and signaling mechanisms involved in the
   transcellular degradation of autophagosomes in hepatocytes,
   particularly in the context of Cd exposure, remain unclear. In this
   study, we describe a novel autophagy pathway, transcellular autophagy.
   Our findings reveal that Cd triggers transcellular autophagy by
   disrupting autophagy fusion mechanisms, with reactive oxygen species
   (ROS) serving as a crucial upstream signal. This process involves the
   transfer of autophagosomes from Cd‐injured hepatocytes to adjacent
   healthy hepatocytes in a TNT‐dependent manner, thereby alleviating
   autophagosome accumulation in Cd‐injured hepatocytes. These findings
   reveal a unique protective mechanism that can alleviate the
   accumulation of autophagosomes mediated by incomplete autophagy and
   provide a new direction for the study of autophagy‐targeting strategies
   in disease intervention.

2. Results

2.1. Transcellular Autophagy

   To investigate transcellular autophagy, we labeled autophagosomes with
   expressing red fluorescence protein‐microtubule associated protein 1
   light chain 3 alpha [LC3] (RFP‐LC3) and observed the dynamic process of
   LC3 puncta transfer between AML12 cells using a high‐content analysis
   system. Figure [95]1A and Video [96]S1 (Supporting Information) show
   that LC3 puncta were gradually transferred from one cell to another
   through the tubular structure between cells. In the co‐culture model of
   AML12 cells labeled with RFP‐LC3 or green fluorescent protein (GFP) we
   found a distribution of RFP‐LC3 puncta in GFP‐labeled cells in the Cd
   group (Figure [97]1B), indicating that Cd can induce autophagosome
   transfer between AML12 cells. Subsequently, in the co‐culture model of
   AML12 cells labeled with RFP‐LC3 or GFP, LC3 puncta gradually
   disappeared after transferring to adjacent cells (Figure [98]1C; Video
   [99]S2, Supporting Information). This indicates that Cd can induce
   autophagosome transfer to adjacent cells for degradation. To verify
   whether autophagosomes transferred to adjacent cells could fuse with
   lysosomes in adjacent cells, we established co‐culture models of AML12
   cells labeled with RFP‐LC3 or GFP‐LAMP2 (lysosomal associated membrane
   protein 2; lysosome tags). As shown in Figure [100]1D, LC3 and LAMP2 in
   GFP‐LAMP2 labeled cells in the Cd group. This indicates that Cd can
   induce autophagosome transfer to adjacent cells and fusion with
   lysosomes in adjacent cells. These results indicate that autophagosomes
   can be transferred from one cell to a neighboring cell, where they fuse
   with lysosomes for degradation. We refer to this phenomenon as
   transcellular autophagy.

Figure 1.

   Figure 1
   [101]Open in a new tab

   Transcellular autophagy. AML12 cells labeled with RFP‐LC3 were exposed
   to 5 µm Cd and imaged every 2 min using a high‐content analysis system.
   A) Live cell imaging of the autophagosomes transfer process. Scale bar,
   10 µm. AML12 cells labeled with RFP‐LC3 were co‐cultured with AML12
   cells labeled with GFP at a ratio of 1:1 and treated with 5 µm Cd for
   6 h. B) Confocal microscopy images of LC3 puncta distribution. Scale
   bar, 10 µm. C) Live cell imaging of the degradation process of
   autophagosomes transferred to adjacent cells. Scale bar, 10 µm. AML12
   cells labeled with RFP‐LC3 were co‐cultured with AML12 cells labeled
   with GFP‐LAMP2 at a ratio of 1:1 and treated with 5 µm Cd for 6 h. D)
   Confocal microscopy images of co‐localization of LC3 with LAMP2. Scale
   bar, 10 µm.

2.2. Cd activates Transcellular Autophagy and Promotes the Transfer of
Autophagosomes from Damaged Cells to Normal Cells

   To evaluate the effect of Cd on transcellular autophagy in AML12 cells,
   we mixed RFP‐LC3‐labeled cells with GFP‐LC3‐labeled cells at a 1:1
   ratio to establish a co‐culture model (Figure [102]2A). Changes in the
   double‐positive fluorescence rate were analyzed using flow cytometry to
   reflect changes in the LC3 transfer rate, which was used to
   quantitatively assess transcellular autophagy. Figure [103]2B shows
   that the changes in the double‐positive fluorescence rate exhibited the
   same trend at different time points of Cd exposure; as the Cd exposure
   dose increased, the double‐positive fluorescence rate gradually
   increased. This indicates that Cd activates transcellular autophagy in
   AML12 cells. Based on the above results, 10 µm Cd exposure for 6 h was
   selected for subsequent mechanism research, as under this exposure
   condition, transcellular autophagy was ongoing and maintained at a
   relatively high level.

Figure 2.

   Figure 2
   [104]Open in a new tab

   Effect of Cd on transcellular autophagy in AML12 cells. A) Schematic
   diagram of the co‐culture model of AML12 cells labeled with RFP‐LC3 and
   AML12 cells labeled with GFP‐LC3. B) In the co‐culture model, different
   concentrations of Cd were exposed for different times, and the LC3
   transfer rate was analyzed using flow cytometry (n = 3). C) Schematic
   diagram of the model for evaluating the transfer of autophagosomes from
   normal to damaged cells. D) Flow cytometry analysis of the change in
   the transfer rate of LC3 from normal to damaged cells (n = 3). E)
   Schematic diagram of the model for evaluating the transfer of
   autophagosomes from damaged to normal cells. F) Flow cytometry analysis
   of the change in the transfer rate of LC3 from damaged to normal cells
   (n = 3). G) Representative confocal images of the transfer of
   autophagosomes from damaged to normal cells. Scale bar, 20 µm. Data are
   expressed as the mean ± SD. Statistical analysis was conducted using
   one‐way analysis of variance and Scheffe's F test. Compared with the
   control group, ^* p < 0.05, ^** p < 0.01.

   To investigate the directionality of autophagosome transfer between
   AML12 cells, we established two co‐culture models. In the co‐culture
   model evaluating the transfer of autophagosomes from normal to damaged
   cells (Figure [105]2C), we observed a slight increase in the
   fluorescence double‐positive rate in the Cd group compared with that in
   the control group (p< 0.05), ≈1.6 times higher than that in the control
   group (Figure [106]2D). However, when examining the transfer of
   autophagosomes from damaged cells to normal cells in the co‐culture
   model (Figure [107]2E), the fluorescence double‐positive rate in the Cd
   group significantly increased (p< 0.01), ≈3.2 times higher than that in
   the control group (Figure [108]2F), which was much higher than the
   change in the transfer rate of autophagosomes from normal to damaged
   cells. fluorescent observations are presented in Figure [109]2G,
   depicting the transfer of autophagosomes from damaged (labeled with
   RFP‐LC3) to adjacent normal cells (labeled with GFP). These findings
   suggest that the transfer of autophagosomes between cells is
   bidirectional, and during Cd‐induced transcellular autophagy,
   autophagosomes are primarily transferred from damaged to normal cells.

2.3. Transcellular Autophagy Requires TNTs

   To investigate the primary pathways of autophagosome transfer between
   AML12 cells, we examined the role of extracellular
   vesicles/endocytosis, gap junctions, and TNTs in Cd‐induced
   transcellular autophagy in AML12 cells. Using a Transwell co‐culture
   model and gap junction inhibitors (Gap 27 and 18‐β‐glycyrrhetinic acid
   [GA]), we excluded the dependence of intercellular autophagosome
   transfer on extracellular vesicles/endocytosis and gap junctions
   (Figure [110]S1, Supporting Information).

   To further investigate the role of TNTs in Cd‐induced transcellular
   autophagy, we examined the effects of Cd on TNT formation in AML12
   cells. As shown in Figure [111]3A, Cd exposure promoted the formation
   of TNT‐like structures in AML12 cells (P< 0.01). Subsequently, we
   characterized TNTs using phalloidin staining. As shown in
   Figure [112]3B,C, the intercellular connecting structures in the Cd
   group were primarily composed of F‐actin and 3D imaging revealed
   tubular structures connecting the two cells, which is consistent with
   the reported structure of TNTs.^[ [113]^32 ^] These data indicate that
   Cd promotes the formation of TNTs. Notably, transmission electron
   microscopy (TEM) revealed the presence of autophagosomes within TNTs
   (Figure [114]3D). Next, we investigated the effects of TNT inhibition
   on Cd‐induced transcellular autophagy. As shown in Figure [115]3E,F,
   inhibition of TNT formation using the TNT inhibitor, Latrunculin B
   (Lat‐B), suppressed Cd‐induced transcellular autophagy. Collectively,
   these findings suggest that Cd‐induced transcellular autophagy is
   dependent on TNTs.

Figure 3.

   Figure 3
   [116]Open in a new tab

   Cd induced transcellular autophagy requires TNTs. AML12 cells were
   exposed to various concentrations of Cd for 6 h. A) TNT‐like structures
   between cells under brightfield microscopy (TNTs were highlighted with
   black arrows). Scale bar, 100 µm (n = 6). AML12 cells was exposed to 10
   µm Cd for 6 h. B) Phalloidin staining was used to characterize TNTs
   (TNTs were highlighted with white arrows). Scale bar, 10 µm. C)
   Confocal 3D imaging of TNTs. D) TEM images of autophagosomes in TNTs.
   Scale bar, 0.5 µm. Cells were pretreated with 100 nm Lat‐B for 2 h and
   then treated with 10 µm Cd for 6 h. E) changes in the number of TNTs
   (TNTs were highlighted with black arrows). Scale bar, 100 µm (n = 6).
   F) Flow cytometry of LC3 transfer rate (n = 3). Data are expressed as
   the mean ± SD. Statistical analysis was conducted using one‐way
   analysis of variance and Scheffe's F test. Compared with the control
   group, ^* p < 0.05, ^** p < 0.01. Compared with the Cd group, # p <
   0.05, ## p < 0.01.

2.4. RNA Sequencing (RNA‐seq) of Potential Upstream Signals Regulating
Cd‐Induced Transcellular Autophagy

   To investigate the signaling pathways upstream of Cd‐induced
   transcellular autophagy, we performed RNA‐seq analysis. As shown in
   Figure [117]4A, 12008 differentially expressed genes (DEGs) were
   identified between the Control and Cd groups, of which 6815 were
   upregulated and 5193 were downregulated. KEGG enrichment analysis
   revealed that the DEGs were primarily enriched in endocytosis, the cell
   cycle, autophagy, MAPK signaling pathway, Mitophagy, AMPK signaling
   pathway, ROS, apoptosis, and other related pathways (Figure [118]4B).
   Given that autophagy is a fundamental prerequisite for cell‐to‐cell
   transmission of autophagic components and that ROS act as crucial
   upstream modulators of the aforementioned signaling pathways, this
   study focused on autophagy and ROS signaling.

Figure 4.

   Figure 4
   [119]Open in a new tab

   RNA‐seq analysis to investigate potential upstream signals regulating
   Cd‐induced transcellular autophagy. AML12 cells were treated with 10 µm
   Cd for 6 h, followed using RNA‐seq analysis (n = 6). A) Volcano plot of
   DEGs. B) KEGG pathway enrichment analysis.

2.5. Incomplete Autophagy is Required for the Induction of Transcellular
Autophagy

   First, we investigated the role of autophagy activation signals in
   Cd‐induced transcellular autophagy. We treated the cells with
   3‐Methyladenine (3‐MA; an inhibitor of autophagy) in combination with
   Cd and analyzed transcellular autophagy using flow cytometry. As
   expected, inhibition of autophagy by 3‐MA significantly reduced
   Cd‐induced transcellular autophagy (Figure [120]5A). We obtained the
   same result with AMPK pathway (important for autophagy regulation)
   inhibition (Figure [121]S2, Supporting Information). This indicated
   that autophagy activation is a prerequisite for Cd‐induced
   transcellular autophagy. Subsequently, we investigated whether
   autophagy activation could directly activate transcellular autophagy.
   Autophagy was activated using rapamycin (RAPA; an autophagy activator)
   and transcellular autophagy was analyzed using flow cytometry. Notably,
   RAPA‐induced autophagy did not directly activate transcellular
   autophagy (Figure [122]5B). We validated this finding by activating
   autophagy during starvation. Our previous study showed that the
   initiation and termination of starvation‐induced autophagy are dynamic,
   with autophagy activated after 3 h of starvation and terminated after
   6 h of starvation.^[ [123]^44 ^] However, this study showed that
   starvation for 3 or 6 h did not directly activate transcellular
   autophagy (Figure [124]5C). Given that transcellular autophagy mainly
   depends on TNTs, we evaluated the effect of starvation on the formation
   of TNTs using phalloidin staining. Notably, the results showed that
   starvation for 3 or 6 h promoted the formation of TNTs
   (Figure [125]5D). This indicates that starvation can activate autophagy
   and promote the formation of TNTs but cannot directly activate
   transcellular autophagy. Collectively, these data suggest that
   activation of autophagy is a prerequisite for Cd‐induced transcellular
   autophagy, but not a direct trigger.

Figure 5.

   Figure 5
   [126]Open in a new tab

   Incomplete autophagy is required for the induction of transcellular
   autophagy. AML12 cells were pretreated with 5 mm 3‐MA for 1 h and then
   treated with 10 µm Cd for 6 h. A) Flow cytometry analysis of LC3
   transfer rate (n = 3). AML12 cells were treated with 100 nm RAPA for
   6 h. B) Flow cytometry analysis of LC3 transfer rate (n = 3). Cells
   were treated with EBSS for 3 h or 6 h. C) Flow cytometry analysis of
   LC3 transfer rate (n = 3). D) Observation of TNTs using Phalloidin
   staining. AML12 cells were treated with EBSS and 50 nm TG for 3 h. E)
   Observation of TNT formation (TNTs were highlighted with black arrows).
   Scale bar, 100 µm. F) Flow cytometry analysis of LC3 transfer rate (n =
   3). AML12 cells were treated with 50 nm TG and 10 µm Cd for 6 h. G)
   Observation of TNT formation (TNTs were highlighted with white arrows).
   Scale bar, 10 µm. H) Flow cytometry analysis of LC3 transfer rate (n =
   3). After silencing VPS41, AML12 cells were treated with 10 µm Cd for
   6 h. I) Observation of TNT formation (TNTs were highlighted with white
   arrows). Scale bar, 10 µm. J) Flow cytometry analysis of LC3 transfer
   rate (n = 3). Data are expressed as the mean ± SD. Statistical analysis
   was performed using one‐way analysis of variance and Scheffe's F test.
   Compared with the control group, ^* p < 0.05, ^** p < 0.01. Compared
   with the Cd or NC+Cd group, ^# p < 0.05, ^## p < 0.01.

   We further investigated the effect of disruption of the autophagy
   fusion mechanism on transcellular autophagy. First, we analyzed the
   changes in TNT formation and transcellular autophagy under starvation
   and Thapsigargin (TG; an inhibitor of autophagosome‐lysosome fusion)
   treatment. As shown in Figure [127]5E,F, Earle's balanced salt solution
   (EBSS) treatment alone induced TNT formation but did not activate
   transcellular autophagy, whereas TG treatment alone did not induce TNT
   formation or activate transcellular autophagy. Combined treatment with
   EBSS and TG induced TNT formation and activated transcellular
   autophagy. This indicated that blocking the fusion of autophagosomes
   and lysosomes can activate transcellular autophagy by activating
   autophagy and inducing TNT formation. Second, we investigated the
   effect of disrupting the autophagy fusion mechanism on Cd‐induced
   transcellular autophagy using combined treatment with Cd and TG. As
   shown in Figure [128]5G,H, the inhibition of autophagosome‐lysosome
   fusion further promoted Cd‐induced TNT formation and transcellular
   autophagy. Our previous study showed that silencing vacuolar protein
   sorting 41 (VPS41) can inhibit autophagosome‐lysosome fusion, thereby
   aggravating Cd‐induced incomplete autophagy.^[ [129]^44 ^] We further
   verified the role of the disruption of the autophagy fusion mechanism
   in Cd‐induced transcellular autophagy by silencing VPS41 to inhibit
   autophagosome‐lysosome fusion. As shown in Figure [130]5I,J, silencing
   of VPS41 further promoted Cd‐induced TNT formation and transcellular
   autophagy. These data suggest that disruption of the autophagy fusion
   mechanism triggers Cd‐induced transcellular autophagy. Collectively,
   these results indicated that incomplete autophagy was required for the
   induction of transcellular autophagy.

2.6. Accumulation of ROS is a Driving Factor for Activating Transcellular
Autophagy

   To investigate the role of ROS in Cd‐induced transcellular autophagy in
   AML12 cells, we pretreated cells with 400 µm N‐acetyl‐L‐cysteine (NAC;
   a ROS scavenger) for 1 h followed by combined treatment with 5 µm Cd
   for 6 h. Changes in ROS levels and transcellular autophagy were
   analyzed using flow cytometry. As shown in Figure [131]6A,B, the
   Cd‐induced accumulation of ROS was completely inhibited by NAC, which
   also completely inhibited Cd‐induced transcellular autophagy.
   Subsequently, we analyzed changes in autophagy levels and TNT
   formation. As shown in Figure [132]6C,D, NAC inhibited Cd‐induced
   autophagy activation and TNT formation. These results suggest that ROS
   scavenging inhibits both autophagy activation and TNT formation,
   ultimately suppressing Cd‐induced transcellular autophagy. Therefore,
   ROS accumulation is an important upstream signal for Cd‐induced
   transcellular autophagy.

Figure 6.

   Figure 6
   [133]Open in a new tab

   The accumulation of ROS is a driving factor for activating
   transcellular autophagy. AML12 cells were pretreated with 400 µm NAC
   for 1 h and then treated with 10 µm Cd for 6 h. A) Flow cytometry
   analysis of LC3 transfer rate (n = 6). B) Flow cytometry analysis of
   ROS levels (n = 6). C) Confocal microscopy images of RFP‐LC3 puncta.
   Scale bar, 10 µm. D) Phalloidin staining to observe TNTs (TNTs were
   highlighted with white arrows). Scale bar, 50 µm (n = 4). AML12 cells
   were treated with 400 nm H[2]O[2] or 100 µm DMNQ for 6 h. E) Flow
   cytometry analysis of ROS levels (n = 3). F) Brightfield microscopy
   observation of TNTs (TNTs were highlighted with black arrows). Scale
   bar, 100 µm. G) RFP‐GFP‐LC3 fluorescence analysis of autophagic flux.
   Scale bar, 20 µm. H, I) Flow cytometry analysis of LC3 transfer rate (n
   = 3). Data are expressed as the mean ± SD. Statistical analysis was
   performed using one‐way analysis of variance and Scheffe's F test.
   Compared with the control group, ^* p < 0.05, ^** p < 0.01. Compared
   with the Cd group, ^# p < 0.05, ^## p < 0.01.

   To further confirm that the accumulation of ROS is an important
   upstream signal for activating transcellular autophagy, cells were
   treated with 400 nm H[2]O[2] or 100 µm 2,3‐dimethoxy‐1,4‐naphthoquinone
   (DMNQ; a ROS inducer) for 6 h, and then the levels of ROS, TNT
   formation, autophagic flux, and changes in transcellular autophagy were
   analyzed. As expected, H[2]O[2] and DMNQ induced ROS accumulation
   (Figure [134]6E), promoted TNT formation (Figure [135]6F), blocked
   autophagic flux (Figure [136]6G), and ultimately activated
   transcellular autophagy (Figure [137]6H,I). These data further confirm
   that ROS accumulation is an important upstream signal that induces
   transcellular autophagy.

2.7. Cd‐induced Transcellular Autophagy Depends on the TNFAIP2‐TNT System

   To identify the key proteins that regulate TNT formation in AML12
   cells, we first analyzed the expression of genes related to TNT
   formation using RNA‐seq. As shown in Figure [138]7A, Cd exposure led to
   the upregulation of genes, such as Tnfaip2, Rhoc, Cdc42, Myo10, and
   Ralbp1, whereas genes, such as Rhoa, Rhob, Gja1, S100a4, and Rala were
   downregulated. Lst1 showed no significant changes. After silencing
   Rhoc, Cdc42, Myo10, Ralbp1, and Lst1 using small‐interfering RNA
   (siRNA), cells were treated with 5 µm Cd for 6 h to analyze the
   formation of TNTs. As shown in Figures [139]S3 and [140]S4 (Supporting
   Information), the silencing of Rhoc, Cdc42, Myo10, Ralbp1, and Lst1 did
   not inhibit the formation of TNTs induced by Cd. We investigated the
   role of TNFAIP2 in Cd‐induced TNT formation. Western blotting (WB)
   results showed that Cd promoted the expression of the TNFAIP2 protein
   (Figure [141]7B). Phalloidin staining revealed that TNFAIP2 silencing
   inhibited Cd‐induced TNT formation (Figure [142]7C). This suggests that
   TNFAIP2 is a key protein involved in Cd‐induced TNT formation in AML12
   cells. We investigated the role of TNFAIP2 in Cd‐induced transcellular
   autophagy in AML12 cells. After silencing TNFAIP2, the cells were
   treated with 5 µm Cd for 6 h, and the level of transcellular autophagy
   was analyzed using flow cytometry. As shown in Figure [143]7D, TNFAIP2
   silencing inhibited Cd‐induced transcellular autophagy. This indicated
   that Cd‐induced transcellular autophagy depends on the TNFAIP2‐TNT
   system.

Figure 7.

   Figure 7
   [144]Open in a new tab

   Cd‐induced transcellular autophagy depends on the TNFAIP2‐TNT system.
   AML12 cells were treated with 10 µm Cd for 6 h. A) RNA‐seq analysis of
   gene expression related to TNT formation (n = 6). B) WB analysis of
   TNFAIP2 protein expression level (n = 3). After silencing TNFAIP2,
   AML12 cells were treated with 10 µm Cd for 6 h. C) Phalloidin staining
   analysis of TNT formation (TNTs were highlighted with white arrows).
   Scale bar, 20 µm. D) Flow cytometry analysis of LC3 transfer rate (n =
   3). Data are expressed as the mean ± SD. Statistical analysis was
   performed using one‐way analysis of variance and Scheffe's F test.
   Compared with the control group, ^* p < 0.05, ^** p < 0.01. Compared
   with the NC+Cd group, ^# p < 0.05, ^## p < 0.01.

2.8. Protective Effect of the TNFAIP2‐TNT System‐Mediated Transcellular
Autophagy on Cd‐Exposed AML12 Cells

   We investigated the role of the TNFAIP2‐TNT system in Cd‐induced
   autophagic flux blockage and apoptosis in AML12 cells. We first
   analyzed changes in autophagic flux by silencing TNFAIP2 or using Lat‐B
   to inhibit the formation of TNTs. As shown in Figure [145]8A, silencing
   TNFAIP2 exacerbated Cd‐induced autophagic flux blockage. Similar
   results were obtained when the formation of TNTs was inhibited using
   Lat‐B (Figure [146]8B). Subsequently, after silencing TNFAIP2, cells
   were treated with 5 µm Cd for 6 h to analyze the changes in apoptosis
   rate. As shown in Figure [147]8C, silencing TNFAIP2 exacerbated
   Cd‐induced apoptosis. These results indicate that the inhibiting of the
   TNFAIP2‐TNT system can exacerbate Cd‐induced autophagic flux blockage
   and apoptosis in AML12 cells.

Figure 8.

   Figure 8
   [148]Open in a new tab

   Protective effect of the TNFAIP2‐TNT system‐mediated transcellular
   autophagy on Cd‐exposed AML12 cells. After silencing TNFAIP2,
   RFP‐GFP‐LC3‐labeled AML12 cells were treated with 10 µm Cd for 6 h. A)
   Confocal microscopy images of RFP‐GFP‐LC3 puncta. Scale bar, 10 µm.
   RFP‐GFP‐LC3‐labeled AML12 cells were pretreated with 100 nm Lat‐B for
   2 h and then treated with 10 µm Cd for 6 h. B) Confocal microscopy
   images of RFP‐GFP‐LC3 puncta. Scale bar, 10 µm. After silencing
   TNFAIP2, AML12 cells were treated with 10 µm Cd for 6 h. C) Flow
   cytometry analysis of apoptosis rate (n = 3). RFP‐LC3‐labeled AML12
   cells (treatment with or without si‐TNFAIP2) were treated with 10 µm Cd
   for 6 h, and then co‐cultured with GFP‐LAMP2‐labeled AML12 cells
   (treatment with or without si‐TNFAIP2) for 12 h. D) Schematic diagram
   of co‐culture model. E) Confocal images of RFP‐LC3 puncta. Scale bar,
   10 µm. Data are expressed as the mean ± SD. Statistical analysis was
   performed using one‐way analysis of variance and Scheffe's F test.
   Compared with the control group, ^* p < 0.05, ^** p < 0.01. Compared
   with the NC+Cd group, ^# p < 0.05, ^## p < 0.01.

   Finally, in the co‐culture model used to evaluate the transfer of
   autophagosomes from injured to normal cells, we further verified the
   role of TNFAIP2‐TNT system‐mediated transcellular autophagy in
   Cd‐induced autophagic flux blockage in AML12 cells. The co‐culture
   model is shown in Figure [149]8D. Following the silencing of TNFAIP2 in
   RFP‐LC3‐labeled AML12 cells, the cells were exposed to 10 µm Cd for
   6 h. Subsequently, GFP‐LAMP2‐labeled AML12 cells that had been silenced
   for TNFAIP2 were inoculated and co‐cultured for 12 h. Changes in LC3
   puncta were observed using laser confocal microscopy. Figure [150]8E
   shows that in the co‐culture group without silencing TNFAIP2, the LC3
   puncta of Cd‐injured cells (labeled with RFP‐LC3) were transferred to
   adjacent normal cells (labeled with GFP‐LAMP2) and fused with
   lysosomes. However, in the co‐culture group with silenced TNFAIP2, no
   RFP‐LC3 puncta were observed in normal cells (labeled with GFP‐LAMP2),
   whereas a significant accumulation of LC3 puncta was observed in
   Cd‐injured cells (labeled with RFP‐LC3). These findings suggest that
   after inhibiting of the TNFAIP2‐TNT system, autophagosomes in
   Cd‐injured cells could not be transferred to normal cells and
   accumulated. This further demonstrates that the TNFAIP2‐TNT
   system‐mediated transcellular autophagy is a protective mechanism for
   alleviating Cd‐induced autophagic flux blockage in AML12 cells.

3. Discussion

   Autophagy is a multi‐step dynamic degradation process that maintains
   cellular homeostasis by degrading damaged organelles, aggregated
   proteins, and lipids and recycling their components.^[ [151]^9 ,
   [152]^45 ^] It is generally accepted that the complete process of
   autophagy is independently completed within the same cell. However,
   recent studies have revealed autophagosome transfer between cells,
   indicating the possibility of transcellular degradation of
   autophagosomes.^[ [153]^37 , [154]^42 , [155]^43 ^] Here, we found that
   autophagosomes can transfer from one cell to another in a TNT‐dependent
   manner and fuse with lysosomes in adjacent cells for degradation. This
   is not a simple phenomenon of autophagosome transfer, but a regulated
   and complete autophagy process called transcellular autophagy (Figure
   [156]S5, Supporting Information). We used different fluorescent markers
   to distinguish between receptor and donor cells, and successfully
   observed the complete dynamic process of transcellular autophagy for
   the first time through high‐content analysis system real‐time imaging.
   Unlike classical autophagy processes, the complete process of
   transcellular autophagy is not limited to a single cell but requires
   the transfer of autophagosomes to neighboring cells, where they
   ultimately fuse with lysosomes and undergo degradation. This new
   autophagy pathway provides cells with a new autophagosome clearance
   strategy to cope with the excessive accumulation of autophagosomes
   caused by incomplete autophagy. However, the current understanding of
   the mechanisms and functions of transcellular autophagy remains
   limited. This study focused on the directionality and pathways of
   autophagosome transfer during transcellular autophagy, as well as the
   conditions and upstream signals that trigger transcellular autophagy
   activation. We also explored the functions of transcellular autophagy.

   We explored the directionality of autophagosome transfer during
   Cd‐induced transcellular autophagy. However, the direction of organelle
   transfer remains unclear. Mitochondria are currently substantial
   interest in the study of organelle transfer. Intercellular
   mitochondrial transfer is bidirectional.^[ [157]^41 , [158]^46 ^]
   Notably, most recent studies indicate that mitochondria tend to be
   preferentially transferred from healthy to damaged cells. For example,
   in most studies using MSCs as donor cells, as a rescue mechanism, MSCs
   transfer mitochondria to damaged cells through TNTs to enhance the
   mitochondrial function and energy metabolism of damaged cells, and even
   rescue damaged cells.^[ [159]^47 , [160]^48 , [161]^49 , [162]^50 ,
   [163]^51 ^] Although autophagosome transfer has been reported, evidence
   supporting its directionality remains scarce. Our study found that
   intercellular autophagosome transfer is also bidirectional, but tends
   to transfer from Cd‐injured cells to healthy cells. Nevertheless, the
   differences in the status of autophagosomes that are transferred in
   different directions and the regulatory mechanisms governing their
   directional transfer are unclear.

   Next, we explored the autophagosome transfer pathways during Cd‐induced
   transcellular autophagy. The transfer of mitochondria between cells can
   be achieved through various ways, including TNT, extracellular
   vesicles, and gap junction channels (GJCs).^[ [164]^52 ^] Although
   reports on autophagosome transfer suggest that TNTs play a significant
   role, there is currently insufficient evidence to indicate whether
   other pathways are involved. In the present study, we ruled out the
   importance of extracellular vesicles/endocytosis and GJCs in Cd‐induced
   autophagosome transfer. Instead, we confirmed that TNT is the primary
   pathway of autophagosome transfer during Cd‐induced transcellular
   autophagy. Notably, our TEM analysis revealed the presence of
   autophagosomes within TNTs, providing intuitive and compelling evidence
   for TNT‐mediated autophagosome transfer.

   The mechanism of TNT formation is highly complex, and there is still a
   lack of comprehensive understanding of the regulatory mechanisms
   involved in its formation process. Despite this, several proteins have
   been identified in different cell types that participate in the
   regulation of TNT formation, including TNFAIP2, RHOC, MYO10, RALBP1,
   LST1, CDC42, RHOA, RHOB, S100A4, and GJA1.^[ [165]^28 , [166]^53 ,
   [167]^54 , [168]^55 ^] However, the key proteins regulating TNT
   formation may vary among different cell types, and the same protein may
   have opposite effects on TNT formation in different cell types.^[
   [169]^56 ^] This suggests the existence of cell‐specific mechanisms
   underlying TNT formation. TNFAIP2 was one of the earliest markers
   identified in of TNT formation, and its crucial role has been confirmed
   in multiple cell types.^[ [170]^57 , [171]^58 ^] In our study, among
   the validated proteins, only TNFAIP2 inhibited Cd‐induced TNT formation
   in AML12 cells. Furthermore, after silencing TNFAIP2, Cd‐induced
   autophagosome transfer was almost completely inhibited. These results
   emphasize the crucial role of the TNFAIP2‐TNT system in Cd‐induced
   transcellular autophagy.

   Next, we explored the conditions required for the activation of
   transcellular autophagy. We found that incomplete autophagy and the
   formation are prerequisites for transcellular autophagy activation.
   Previous studies have shown that starvation activates autophagy and
   induces the generation of TNTs.^[ [172]^59 , [173]^60 ^] Our findings
   revealed that starvation alone was insufficient to directly trigger
   transcellular autophagy. Starvation‐induced autophagy is complete
   autophagy with intact fusion and degradation mechanisms, allowing
   autophagosomes to be efficiently cleared through autophagic pathways.^[
   [174]^59 ^] Conversely, Cd‐induced autophagy is incomplete, with the
   disruption of autophagic fusion mechanisms, resulting in the
   accumulation of autophagosomes that cannot be cleared.^[ [175]^44 ^]
   Therefore, we speculate that the trigger of transcellular autophagy may
   be related to the disruption of autophagic fusion mechanisms. Our
   results support this hypothesis. We found that, while activating
   autophagy and inducing the formation of TNTs, blocking the fusion of
   autophagosomes and lysosomes activated transcellular autophagy. In
   summary, our findings revealed that incomplete autophagy and the
   formation of TNTs are necessary conditions for transcellular autophagy.

   After identifying the conditions necessary for transcellular autophagy
   activation, we explored the upstream signals that trigger this
   activation. ROS are upstream signals for autophagy activation and TNT
   formation, and excessive accumulation of ROS can lead to autophagic
   flux blockage.^[ [176]^61 , [177]^62 , [178]^63 ^] Autophagy, TNT
   formation, and disruption of the autophagy fusion mechanism (which can
   lead to autophagic flux blockade) are necessary conditions for
   transcellular autophagy activation. Additionally, ROS plays a pivotal
   role as an upstream signal in Cd‐induced cell injury.^[ [179]^15 ^]
   Therefore, this study focused on the role of ROS in transcellular
   autophagy activation. Our study revealed that clearing ROS accumulation
   using NAC completely inhibited Cd‐induced transcellular autophagy.
   Additionally, ROS clearance inhibited Cd‐induced autophagy activation
   and TNT formation, suggesting that ROS accumulation is a vital upstream
   signal for Cd‐induced transcellular autophagy. Notably, we discovered
   two drugs, H[2]O[2] and DMNQ, that activate transcellular autophagy in
   AML12 cells. DMNQ, a ROS inducer, can trigger ROS‐dependent cell
   death.^[ [180]^64 ^] Our findings show that inducing ROS accumulation
   through H[2]O[2] or DMNQ can promote TNT generation and induce
   autophagic flux blockade, ultimately triggering transcellular
   autophagy. This further confirms that ROS accumulation is an important
   driving factor of transcellular autophagy.

   Autophagy is a vital cellular self‐cleaning process that plays a
   crucial role in the maintenance of cellular homeostasis and survival.^[
   [181]^65 ^] Incomplete autophagy refers to a malfunctioning process in
   which accumulated autophagosomes fail to fuse with lysosomes for
   degradation. This ultimately leads to blockage of autophagic flux and
   excessive accumulation of autophagosomes, which are often detrimental
   to cell survival.^[ [182]^66 ^] A recent study revealed that healthy
   and AGE‐damaged podocytes can exchange lysosomes and autophagosomes
   through TNTs. This exchange mechanism helps to alleviate lysosomal
   dysfunction, excessive accumulation of autophagosomes, and apoptosis in
   damaged cells. However, this protective effect was eliminated by the
   tnfaip2 deletion or TNT inhibition.^[ [183]^37 ^] This suggests that
   the transfer of autophagosomes to neighboring cells may be a protective
   mechanism against excessive accumulation of autophagosomes. In this
   study, we found that silencing TNFAIP2 or inhibiting the formation of
   TNTs exacerbated Cd‐induced autophagic flux blockage and apoptosis.
   Furthermore, in co‐culture models, we found that Cd‐injured cells could
   transfer autophagosomes to normal cells for degradation, which helped
   alleviate the accumulation of autophagosomes in Cd‐injured cells.
   However, upon silencing TNFAIP2, autophagosomes in Cd‐injured cells
   failed to be transferred to normal cells, leading to their
   accumulation. These findings suggest that the TNTAIP2‐TNT
   system‐mediated transcellular autophagy is a protective mechanism that
   alleviates the excessive accumulation of autophagosomes induced by
   incomplete autophagy.

   The current study has some limitations. First, it lacks validation in
   other hepatocyte models, particularly primary hepatocytes. However,
   technical challenges in subculturing and genetic manipulation of
   primary hepatocytes restricted further verification in this study.
   Future research will explore improved methodologies for conducting such
   studies in primary hepatocytes or perform validation across different
   cell lineages. Second, the experiment was conducted under the
   traditional 2D culture conditions. In contrast, 3D spheroid cultures
   mimic the in vivo cellular microenvironment more closely. Therefore,
   future research will consider applying 3D spheroid culture techniques
   to better reflect cell–cell interactions. Third, owing to the numerous
   technical challenges in in vivo research, such as difficulties in
   multiple fluorescent labeling of live cells, accurate differentiation
   between damaged and healthy cells, and real‐time monitoring, the
   results of this study have not yet been verified in vivo. Therefore,
   future research should consider the differences in the physiological
   environments between in vivo and in vitro studies to design a more
   scientific in vivo experimental protocol.

   In summary, we revealed a novel autophagy pathway, transcellular
   autophagy, which serves as a protective mechanism for cells to mitigate
   the excessive accumulation of autophagosomes caused by incomplete
   autophagy. Transcellular autophagy has substantial research potential
   in diseases associated with autophagy dysfunction. Our study provides
   novel insights and directions for strategies targeting autophagy for
   disease intervention.

4. Experimental Section

Reagents and Antibodies

   Cd chloride (CdCl[2], 202908), dexamethasone (D4902), Hoechst33258
   (94403), Lucifer Yellow CH dilithium salt (LY, L0259), and Rhodamine B
   isothiocyanate‐Dextran (RD, R9379) were purchased from Sigma–Aldrich
   (St. Louis, MO, USA). Dulbecco's modified Eagle's Medium/Nutrient
   Mixture F‐12 (D‐MEM/F12, 12400024), fetal bovine serum (FBS, 10437028),
   and insulin‐transferrin‐selenium‐A supplement (ITS‐A, 51300044) were
   purchased from Gibco (Grand Island, NY, USA). RAPA (HY‐10219),3‐MA
   (HY‐19312), TG (HY‐13433), Lat‐B (HY‐101848), Gap 27 (HY‐P0139), GA
   (HY‐N0180), Compound C (HY‐13418A), and DMNQ (HY‐121026) were purchased
   from MedChem Express (Monmouth Junction, NJ, USA). Hoechst 33342
   (C1028), EBSS (C0214), and NAC (S0077) were purchased from Beyotime
   (Shanghai, China). The Annexin V‐fluorescein isothiocyanate
   (FITC)/propidium iodide (PI) Apoptosis Detection Kit (A211) was
   purchased from Vazyme (Nanjing, China). TRITC Phalloidin (CA1610) and
   ROS assay kits (CA1410) were purchased from Solarbio (Beijing, China).
   The Polyplus‐transfection reagent was purchased from
   Polyplus‐transfection (Illkirch, France). Anti‐LST1 (ab252839, 1:2000),
   Anti‐CDC42 antibody (ab187643, 1:10 000), and Anti‐TNFAIP2 (ab196659,
   1:1000) antibodies were purchased from Abcam (Cambridge, UK). Anti‐RHOC
   (PS02067, 1:1000), Anti‐MYO10 antibody (PS02569, 1:1000), and
   Anti‐RALBP1 ([184]T55676, 1:1000) antibodies were purchased from Abmart
   (Shanghai, China).

Cell Culture and Treatment

   The mouse hepatocyte cell line, AML12 (ATCC CRL‐2254, Manassas, VA,
   USA) was seeded in D‐MEM/F12 medium supplemented with 10% FBS, 1%
   ITS‐A, and 40 ng mL^−1 dexamethasone. Cells were incubated in a
   constant‐temperature incubator at 37 °C with 5% CO[2] and saturated
   humidity. A serum‐free medium was used for Cd exposure.

Lentivirus (LV) Infection and siRNA Transfection

   AML12 cells were seeded in a 24‐well plate and cultured to a density of
   ≈20%. Following the manufacturer's protocol, the lentivirus was
   incubated with the cells for 48–72 h. Subsequently, the cells were
   cultured in complete medium containing 4 µg mL^−1 of puromycin to
   select positive cells and establish stable cell lines. RFP‐LC3,
   stubRFP‐sensGFP‐LC3, GFP, and GFP‐LAMP2 lentiviruses (GeneChem
   Corporation, Shanghai, China) were used. For siRNA transfection, siRNA
   was mixed with the Polyplus transfection reagent according to the
   manufacturer's protocol. The mixture was then added to fresh complete
   medium without penicillin or streptomycin, and the cells were
   transfected for 24 h. Silencing efficiency was analyzed using WB. The
   siRNA sequences used in this study were purchased from RiboBio
   Corporation (Guangzhou, China) and are listed in Table [185]S1
   (Suppoerting Information).

High‐Content Analysis System

   RFP‐LC3‐ and GFP‐labeled AML12 cells were cultured separately or mixed
   in 96‐well culture plates. After reaching 70–80% confluence, the cells
   were treated with 5 µm Cd and labeled with Hoechst 33 342 to label the
   cell nucleus. Images were captured every 2 min using a high‐content
   analysis system (PerkinElmer Company, USA).

Confocal Fluorescence Analysis

   RFP‐LC3‐labeled cells were cultured either alone or mixed with GFP‐ or
   GFP‐LAMP2‐labeled cells in a 1:1 ratio on a 24‐well glass slide. The
   cells were treated according to the experimental requirements. After
   treatment, the cells were fixed with 4% paraformaldehyde, stained with
   Hoechst 33 258 to label the cell nuclei, and imaged using a laser
   confocal microscope.

Analysis of Flow Cytometry

   For the analysis of intercellular autophagosome transfer in rat,
   RFP‐LC3‐labeled cells were co‐cultured with GFP‐LC3‐labeled cells at a
   ratio of 1:1 and treated according to experimental requirements.
   Following treatment, the cells were collected and analyzed for the
   double‐positive fluorescence rate using a FACS LSRFortessa flow
   cytometer (BD Company, USA).

   For the analysis of ROS levels, the cells were treated according to the
   experimental requirements, collected through trypsin digestion, and
   incubated with the DCFH‐DA working solution at 37 °C for 20 min in the
   dark. During incubation, the solution was mixed and inverted every
   3–5 min to ensure uniform exposure. Subsequently, the residual working
   solution was washed three times with serum‐free medium to remove the
   unreacted reagents. Finally, the ROS levels in the cells were analyzed
   using flow cytometry.

   For apoptosis rate analysis, after treatment, cells were collected
   through trypsin digestion (without EDTA). Then, 5 µL of Annexin V‐FITC
   and PI were added in sequence, and the cells were incubated at
   room‐temperature for 20 min in the dark. The rate of apoptosis was
   analyzed using flow cytometry.

Cell Co‐Culture Models

   For the Transwell co‐culture model, the Transwell chamber
   (0.4 µm‐aperture) in a six‐well plate was placed upside down in a 50 mL
   beaker. RFP‐LC3‐labeled cells were then inoculated into the outer
   chamber of a Transwell plate and cultured for 12 h. Simultaneously, the
   GFP‐LC3‐labeled cells were inoculated into a six‐well plate and
   cultured for 12 h. The Transwell chamber was transferred to a six‐well
   plate and co‐cultured with GFP‐LC3‐labeled AML12 cells. After treatment
   with 10 µm Cd for 6 h, the GFP‐LC3‐labeled AML12 cells was collected
   from the six‐well plate using trypsin digestion and the fluorescence
   double‐positive rate was analyzed using flow cytometry.

   To evaluate the transfer of autophagosomes from normal to injured
   cells, GFP‐labeled cells were cultured in a six‐well plate. Following
   treatment with 10 µm Cd for 6 h, the culture medium was discarded and
   replaced with complete medium containing RFP‐LC3‐labeled cells. After
   co‐culturing for 12 h, the double‐positive fluorescence rate was
   analyzed using flow cytometry.

   To evaluate the transfer of autophagosomes from injured to normal cells
   in the co‐culture model, RFP‐LC3‐labeled cells were cultured in
   six‐well plates. Following treatment with 10 µm Cd for 6 h, the culture
   medium was discarded and replaced with complete medium containing
   GFP‐labeled cells. After co‐culturing for 12 h, the double‐positive
   fluorescence rate was analyzed using flow cytometry.

Phalloidin Staining

   After treatment, the cells were fixed with 4% formaldehyde solution at
   roommtemperature for 10 min, permeabilized with 0.5% Triton X‐100
   solution for 5 min, incubated with phalloidin working solution at
   room‐temperature in the dark for 30 min, labeled with Hoechst 333 258
   to mark the cell nucleus, and imaged using a confocal microscope (TCS
   SP8 STED, Leica, Germany) or an inverted fluorescence microscope
   (DMI3000B, Leica, Germany).

TEM Analysis

   After exposure to 10 µm Cd for 6 h, the cells were sequentially fixed
   with 2.5% glutaraldehyde and 1% osmium tetroxide. The samples were then
   dehydrated in ethanol and acetone. After immersion in acetone, resin
   embedding, and polymerization, ultrathin sections were prepared. The
   samples were then stained with uranium acetate and lead citrate before
   being photographed using TEM (HT7800 model, Hitachi High‐Tech, Tokyo,
   Japan).

Scrape Loading/Dye Transfer (SL/DT) Analysis

   The SL/DT analysis was conducted in accordance with the previous study
   to evaluate gap junction intercellular communication.^[ [186]^67 ^]
   After treatment, cell scratches were made using surgical blades,
   incubated at 37 °C for 5 min with PBS containing LY (0.5 mg mL^−1) and
   RD (2.5 mg mL^−1), and then fixed with 4% paraformaldehyde. Finally,
   the fluorescent diffusion distance from the scraped edge to the
   adjacent cells was measured using a fluorescence microscope (DMI3000B,
   Leica, Germany).

RNA‐seq Analysis

   The cells were treated with 10 µm Cd for 6 h, and total RNA was
   extracted. RNA‐seq was performed by BGI Genomics Co Ltd. (Shenzhen,
   China). The raw data obtained from sequencing were filtered using
   SOAPnukev1.5.6 to obtain clean data. HISAT2 v2.1.0 and Bowtie2 v2.3.4.3
   software were used to align the clean data against the reference genome
   sequence, while RSEM v1.3.1 software was used for gene expression
   quantification. Pheatmapv1.0.8 was used to draw a cluster heatmap of
   gene expression in different samples. DEGs were identified using a
   Q‐value ≤ 0.05 as the threshold and Phyper was applied for KEGG
   enrichment analysis of DEGs.

WB

   Cells were collected using RIPA lysis buffer supplemented with 1%
   protease inhibitors. The protein concentration of the sample was
   determined using the BCA colorimetry and lysis buffer was added to
   standardize the protein concentration. Subsequently, SDS loading buffer
   was added and denatured to obtain a protein sample. Following SDS‐PAGE,
   the proteins were transferred onto polyvinylidene fluoride membranes.
   The membranes were then sequentially incubated with the corresponding
   primary and secondary antibodies. Target proteins were visualized using
   the ECL chemiluminescence. The grayscale value was analyzed using
   ImageJ software (National Institutes of Health, United States) and the
   ratio of the grayscale value of the target protein to that of β‐actin
   was used to indicate the relative expression level of the target
   protein.

Statistical Analysis

   Data from at least three independent experiments were analyzed one‐way
   analysis of variance and Scheffe's F test using GraphPad Prism 6
   software (GraphPad Software Inc., La Jolla, CA, USA). The results are
   shown as the mean ± SD, with p > 0.05 considered not statistically
   significant and p < 0.05 or p < 0.01 considered statistically
   significant.

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [187]ADVS-12-e02793-s003.docx^ (2.2MB, docx)

   Supplemental Video 1
   [188]Download video file^ (19.8MB, mp4)

   Supplemental Video 2
   [189]Download video file^ (25.3MB, mp4)

Acknowledgements