Abstract

   Tumor‐associated macrophages (TAMs) play a crucial role in tumor
   progression within the tumor microenvironment (TME) through phenotypic
   plasticity and functional modulation. While tunneling nanotubes (TNTs)
   mediate intercellular communication, their role in shaping TAMs
   phenotypes and function remains unclear. This study explores how TNTs
   facilitate the transfer of tumor‐derived materials, particularly fatty
   acids, to TAMs, affecting macrophage polarization and function.
   Single‐cell RNA sequencing identified heterogeneous macrophage
   subpopulations in the TME. Enrichment analysis pinpointed key
   substances transferred via TNTs. Lipidomics and metabolomics analyzed
   the fatty acids involved. In vitro and in vivo experiments validated
   TNTs‐mediated material transfer, and transcriptomic analysis revealed
   the associated signaling pathways. TNTs are the primary route for
   transferring tumor‐derived fatty acids, notably arachidonic acid (AA),
   to macrophages. This transfer reprogrammed TAMs from anti‐tumor CD5L^+
   to pro‐tumor TREM2^+ phenotypes, increasing CCL18 secretion, reducing
   phagocytic activity, and impairing CD8^+ T cell proliferation.
   Mechanistically, AA activated the PI3K‐AKT pathway, driving TAMs
   polarization. These findings are confirmed in xenograft models, where
   TNTs‐induced TAMs exhibited enhanced pro‐tumor properties.
   TNTs‐mediated transfer of tumor‐derived AA reprograms TAMs via PI3K‐AKT
   activation, promoting immune suppression and tumor progression,
   highlighting TNTs and PI3K‐AKT as potential therapeutic targets in
   iCCA.

   Keywords: arachidonic acid, intrahepatic cholangiocarcinoma, PI3K‐AKT
   signaling pathway, tumor‐associated macrophages, tunneling nanotubes
     __________________________________________________________________

   Tunneling nanotubes facilitate the transfer of long‐chain fatty acids,
   primarily arachidonic acid, from tumor cells to macrophages, leading to
   lipid accumulation and activation of the PI3K‐Akt pathway. This drives
   a shift from the anti‐tumor CD5L⁺ phenotype to a pro‐tumor TREM2⁺
   phenotype. The transition reduces phagocytosis, impairs immune
   responses, and increases the release of the tumor‐promoting chemokine
   CCL18, reprogramming macrophage function and creating an
   immunosuppressive microenvironment.

   graphic file with name ADVS-12-e00148-g007.jpg

1. Introduction

   Intrahepatic cholangiocarcinoma (iCCA) is a highly aggressive
   malignancy with limited therapeutic options and dismal overall survival
   rates.^[ [38]^1 ^] The progression of iCCA is closely linked to the
   tumor microenvironment (TME), where tumor‐associated macrophages (TAMs)
   play a central role in promoting immunosuppression, angiogenesis, and
   tumor proliferation. These processes significantly enhance the
   aggressiveness of iCCA, making TAMs attractive targets for
   immunotherapeutic interventions.^[ [39]^2 ^] While tumor‐driven
   polarization of TAMs toward immunosuppressive, tumor‐supportive
   phenotypes is well established, the precise mechanisms driving this
   reprogramming.^[ [40]^3 ^] A deeper understanding of how iCCA tumor
   cells influence TAMs phenotypes could uncover novel therapeutic targets
   and advance cancer immunotherapy strategies.

   Emerging evidence highlights the importance of tumor‐derived factors in
   TAMs reprogramming.^[ [41]^4 ^] Among these, arachidonic acid (AA) and
   its derivatives have been implicated in driving TAMs toward an M2‐like,
   pro‐tumor phenotype in several cancers.^[ [42]^5 ^] Despite these
   findings, the mechanisms through which AA and other tumor‐derived
   metabolites influence macrophage polarization, particularly in iCCA,
   remain elusive. Exploring these mechanisms is crucial to decipher the
   metabolic and functional interplay between tumor cells and TAMs within
   the TME.

   Intercellular communication is fundamental to TME remodeling and tumor
   progression. While extracellular vesicles, cytokines, and chemokines
   have been extensively studied,^[ [43]^4 , [44]^6 ^] recent research has
   identified tunneling nanotubes (TNTs) as a novel mode of direct
   intercellular communication.^[ [45]^7 ^] TNTs are thin, actin‐based
   structures that facilitate the transfer of diverse cellular contents,
   including proteins, organelles, and metabolites, across distant
   cells.^[ [46]^8 ^] Notably, TNTs have been shown to enhance cancer cell
   metabolic plasticity, invasiveness, immune evasion, and resistance to
   therapy.^[ [47]^9 , [48]^10 , [49]^11 ^] However, their specific role
   in regulating macrophage polarization within the TME remains largely
   unexplored. Given the critical involvement of TAMs in tumor
   progression, elucidating how tumor cells influence TAMs function via
   TNTs may uncover novel mechanisms driving TME remodeling and tumor
   progression.

   In this study, we explored the role of TNTs in facilitating the
   transfer of tumor‐derived fatty acids, particularly AA, to TAMs within
   the iCCA microenvironment. Through integrated analyses involving
   single‐cell RNA sequencing (scRNA‐seq), lipidomics, targeted
   metabolomics, and transcriptomics, AA was identified as the predominant
   fatty acid transferred via TNTs. Mechanistically, AA was shown to
   activate the PI3K‐Akt signaling pathway, reprogramming TAMs into a
   pro‐tumorigenic TREM2⁺ phenotype marked by diminished phagocytic
   capacity and impaired support for CD8⁺ T cell proliferation. These
   findings offer novel insights into TNTs‐mediated tumor‐driven
   remodeling of the TME and highlight potential therapeutic targets for
   cancer treatment.

2. Results

2.1. scRNA‐Seq Analysis and In Vitro Experiments Reveal the Advantage of
Tumor Cells in TNT Formation

   To investigate the role of tumor cells in initiating TNTs formation, we
   performed an in‐depth analysis of the genetic alterations in epithelial
   cells within tumor tissues to uncover their potential association with
   TNTs formation. ScRNA‐seq data analysis identified 23 distinct cell
   clusters (Figure [50]S1A, Supporting Information), and cell types were
   classified based on marker gene expression (Figure [51]S1B, Supporting
   Information), successfully distinguishing cholangiocytes from
   macrophages (Figure [52]1A). Differential expression analysis further
   revealed significant differences in gene expression between tumor
   epithelial cells and adjacent normal epithelial cells (Figure [53]1B).

Figure 1.

   Figure 1
   [54]Open in a new tab

   scRNA‐seq analysis and in vitro experiments reveal tumor cell advantage
   in TNT formation. A) UMAP plot of epithelial cells and macrophages. B)
   Volcano plot showing differentially expressed genes between tumor and
   adjacent non‐tumor epithelial cells. C,D) GO and KEGG pathway
   enrichment analyses of upregulated genes in tumor cells. E,F)
   Representative images of homotypic TNTs formed between CMFDA‐labeled
   HIBEpic cells (E) and CMFDA‐HuCC‐T1 cells (F). White arrows indicate
   TNTs. G) Quantification of homotypic TNT connections per 100 cells
   (left) and percentage of TNT‐connected cells (right) between HIBEpic
   and HuCC‐T1. Data are presented as mean ± SD; n = 3. H,I) Heterotypic
   TNTs between CMFDA‐HIBEpic (H) or GFP‐HuCC‐T1 (I) and mCherry‐labeled
   macrophages. J) Quantification of heterotypic TNT connections per 100
   cells (left) and percentage of TNT‐connected cells (right) between
   HIBEpic/HuCC‐T1 and macrophages. Data are presented as mean ± SD; n =
   3.

   GO functional analysis of the upregulated genes in tumor cells
   demonstrated significant enrichment in processes such as focal
   adhesion, actin binding, extracellular matrix organization, and small
   GTPase binding (Figure [55]1C). Additionally, KEGG pathway analysis
   revealed that these genes were mainly involved in focal adhesion,
   regulation of the actin cytoskeleton, the PI3K‐AKT signaling pathway,
   and the MAPK signaling pathway (Figure [56]1D). These biological
   processes and pathways are well‐documented for their pivotal role in
   the formation of TNTs, as extensively studied in the literature.^[
   [57]^12 , [58]^13 , [59]^14 ^] These results suggest that tumor cells
   in iCCA possess a strong intrinsic capacity to initiate and facilitate
   TNTs formation, potentially through the activation of actin
   cytoskeleton‐associated pathways and signaling cascades involved in
   cellular adhesion and motility.

   Additionally, through in vitro experiments, we demonstrated that tumor
   cells exhibit a superior capacity for TNT formation compared to normal
   bile duct epithelial cells. Specifically, the percentage and number of
   homotypic TNTs formed between normal human intrahepatic biliary
   epithelial cells (HIBEpiC) were significantly lower than those observed
   among HuCC‐T1 tumor cells (Figure [60]1E–G). Likewise, heterotypic TNTs
   between HIBEpiC cells and macrophages were substantially less frequent
   compared to those between HuCC‐T1 tumor cells and macrophages
   (Figure [61]1H–J). These differences were statistically significant (p
   < 0.001), providing both quantitative data and visual evidence of TNT
   formation across these cell types. This direct observation not only
   corroborates the molecular insights derived from scRNA‐seq but also
   offers compelling functional evidence of the enhanced TNT formation
   capacity of iCCA tumor cells.

2.2. Identification of Macrophage Phenotypes and Cytokines Associated with
THEIR function

   To elucidate the mechanisms through which tumor cells reshape the TME,
   we focused on macrophages, identified as the most heterogeneous immune
   cell population in our study.^[ [62]^15 ^] ScRNA‐seq analysis
   identified five distinct macrophage subtypes (Figure [63]2A), which
   showed significant differences in their relative abundance between
   tumor tissues and adjacent normal tissues (Figure [64]2B,C).
   Specifically, the mac0 subtype was mainly enriched in normal tissues,
   whereas the mac3 subtype was significantly more abundant in tumor
   tissues. We isolated the mac0 and mac3 macrophage subtypes, excluding
   tumor cells from the mac0 group and normal cells from the mac3 group.
   The top 10 marker genes for both the mac0 and mac3 subtypes are shown
   in Figure [65]2D.

Figure 2.

   Figure 2
   [66]Open in a new tab

   Identification of macrophage subtypes and associated functional
   markers. A) UMAP plot of five macrophage subtypes. B) Distribution of
   macrophage subtypes in tumor (T) and normal (N) tissues. C) Proportions
   of macrophage subtypes in tumor versus normal tissues. D) Heatmap of
   top 10 differentially expressed genes between mac0 and mac3. E) UMAP
   visualization of TREM2 (top) and CD5L (bottom) expression across
   macrophage subtypes in tumor and normal tissues. F) The expression
   levels of TREM2 (top) and CD5L (bottom) across cell types in tumor and
   normal tissues. G,H) Quantitative analysis of TREM2 (G) and CD5L (H)
   expression in macrophages. Data are presented as mean ± SD; n = 3 (^**
   p < 0.01, ^*** p < 0.001). Scale bar: 100 µm.

   Consistent with prior studies, we observed that TREM2, a well‐known
   marker of immunosuppressive macrophages in multiple cancers^[ [67]^16
   ^] was highly expressed in the mac3 subtype within tumor tissues.
   Conversely, CD5L, previously recognized for its anti‐tumor properties
   and positive correlation with CD8⁺ T cell infiltration and M1
   macrophage activity^[ [68]^17 ^] was predominantly expressed in the
   mac0 subtype in normal tissues (Figure [69]2E). Across all analyzed
   cell types, TREM2 expression was primarily confined to tumor tissues,
   whereas CD5L was more prevalent in normal tissues (Figure [70]2F).

   To validate these findings, multiple immunofluorescence (MIF) confirmed
   that TREM2⁺ macrophages were predominantly localized within iCCA tumor
   tissues (Figure [71]2G), while CD5L⁺ macrophages were the predominant
   subtype in peritumoral tissues (Figure [72]2H). Based on these results,
   we classified mac0 as “tumor‐suppressive CD5L⁺ macrophages” and mac3 as
   “tumor‐promoting TREM2⁺ macrophages”.

   In addition, we identified CCL18, a member of the CC chemokine family
   primarily secreted by monocytes, as a potential driver of tumor
   progression. Previous studies have established its pivotal role in
   tumorigenesis across various cancers.^[ [73]^18 , [74]^19 ^] To further
   validate the pro‐tumorigenic role of CCL18, we conducted experiments by
   treating tumor cells with recombinant CCL18 (rCCL18). Using Cell
   counting kit‐8 (CCK‐8) cell proliferation and Transwell migration
   assays, we found that rCCL18 significantly enhanced the proliferative
   and migratory capabilities of tumor cells compared to the control group
   (Figure [75]S1C,D, Supporting Information). Notably, our scRNA‐seq data
   revealed that CCL18 expression was mainly enriched in macrophages and
   significantly upregulated in the TREM2⁺ macrophage subset (Figure
   [76]S1E,F, Supporting Information). This observation suggests that
   TREM2⁺ macrophages may contribute to tumor progression by secreting
   CCL18 and promoting pro‐tumorigenic effects within the TME.

2.3. The Presence of TNTs Between Tumor Cells and Macrophages

   Building on our previous findings, we observed that differentially
   expressed genes in tumor cells and macrophages were significantly
   enriched in pathways associated with TNTs formation.^[ [77]^15 ^]
   Enrichment analyses further suggested that tumor cells possess an
   intrinsic ability to initiate TNTs formation. These observations led us
   to hypothesize that TNTs might serve as a structural link facilitating
   direct communication between tumor cells and macrophages.

   To test this hypothesis, we established a co‐culture system of tumor
   cells and macrophages and employed confocal microscopy to examine TNT
   formation between these cell types (Video [78]S1, Supporting
   Information). Observations at 6 and 24 h showed no significant changes
   in the number or percentage of TNTs formed (Figure [79]3A,B). Treatment
   with CytoB, a known inhibitor of actin polymerization, substantially
   suppressed TNT formation; however, the number and percentage of TNTs
   remained consistent between the 6 and 24‐h time points
   (Figure [80]3C,D).

Figure 3.

   Figure 3
   [81]Open in a new tab

   TNT formation dynamics between tumor cells and macrophages. A) TNTs
   (white arrows) formed between mCherry‐macrophages and HuCC‐T1 cells at
   6 and 24 h. B) Quantification of TNT numbers (left) and percentage of
   TNT‐connected cells (right) at these time points. Data are mean ± SD; n
   = 3 (ns). C) Suppressed TNT formation after CytoB treatment at 6 and
   24 h. D) Quantification of TNT numbers (left) and percentage of
   TNT‐connected cells (right) at these time points after CytoB treatment
   at these time points. Data are mean ± SD; n = 3 (ns). E,F)
   Immunofluorescence staining of F‐actin (phalloidin) and tubulin in
   TNTs. White arrows indicate TNTs, yellow arrows indicate orthogonal
   views. G,H) Disruption of cytoskeletal components by CytoB,
   immunofluorescence staining of F‐actin (phalloidin) and tubulin in
   TNTs. White arrows indicate TNTs, yellow arrows indicate orthogonal
   views. I) Transfer of DIO dye from HuCC‐T1 cells to macrophages via
   TNTs (white arrows). ns: not significant.

   The formation of TNTs between tumor cells and macrophages was a
   frequent and reproducible phenomenon within the co‐culture system.
   Structural analysis confirmed that these TNTs were primarily composed
   of F‐actin (Figure [82]3E) and microtubules (Figure [83]3F), with
   Z‐axis imaging demonstrating their lack of attachment to the basal
   substrate, consistent with established TNT characteristics.^[ [84]^20
   ^] Although CytoB treatment markedly reduced the number and percentage
   of TNTs, the remaining TNTs retained the culture system. Structural
   analysis confirmed that these TNTs were primarily composed of F‐actin
   (Figure [85]3G) and microtubules (Figure [86]3H).

   To confirm the functional role of these TNTs and distinguish them from
   other cellular protrusions,^[ [87]^21 ^]we assessed their capacity for
   intercellular cargo transfer. Using the lipophilic dye DiO as a marker,
   we observed efficient transport of DiO‐labeled cargo from tumor cells
   to macrophages via TNTs (Figure [88]3I, Video [89]S2, Supporting
   Information). These findings provide compelling evidence that TNTs not
   only establish physical connections between tumor cells and macrophages
   but also mediate the transfer of cellular materials, highlighting their
   critical role in facilitating tumor‐macrophage communication.

2.4. Effects of TNTs‐Mediated Material Transfer On Macrophage Phenotypes

   In the TME, TAMs exhibit significant phenotypic plasticity, polarizing
   into either pro‐inflammatory M1‐like macrophages or pro‐tumorigenic
   M2‐like macrophages.^[ [90]^22 ^] Based on our findings, we
   hypothesized that TNTs‐mediated material transfer could drive
   macrophages toward a tumor‐promoting TREM2⁺ phenotype, a subtype
   associated with immunosuppression and tumor promotion.

   To test this hypothesis, we employed flow cytometry to analyze the
   expression of key phenotypic markers TREM2 and CD5L in macrophages
   under different experimental conditions. These conditions included M0
   (control macrophages), Coculture (macrophages co‐cultured with tumor
   cells), CytoB (TNTs formation inhibited by cytochalasin B), and
   Transwell (physical separation of macrophages and tumor cells) groups.
   Experiments were conducted using two cholangiocarcinoma cell lines,
   HuCC‐T1 and RBE. Additionally, we quantified the secretion of the
   pro‐tumorigenic chemokine CCL18 in culture supernatants using
   enzyme‐linked immunosorbent assay (ELISA).

   The results demonstrated a significant increase in TREM2⁺ macrophages
   in the Coculture group compared to other groups (Figure [91]4A–C;
   Figure [92]S2A, Supporting Information), alongside a marked reduction
   in CD5L expression (Figure [93]4D–F; Figure [94]S2B, Supporting
   Information). Furthermore, macrophages in the Coculture group exhibited
   substantially elevated levels of CCL18 secretion (Figure [95]4G,H),
   indicating a functional shift toward a tumor‐promoting phenotype. In
   contrast, TNTs inhibition (CytoB group) and physical separation
   (Transwell group) effectively suppressed these changes, supporting the
   role of TNTs‐mediated material transfer in driving macrophage
   phenotypic conversion.

Figure 4.

   Figure 4
   [96]Open in a new tab

   TNT‐mediated material transfer modulates macrophage phenotypes. A) Flow
   cytometry analysis of TREM2 expression in macrophages co‐cultured with
   RBE cells. B,C) Quantification of TREM2^+ macrophages in HuCC‐T1 (B)
   and RBE (C) co‐culture systems. D) Flow cytometry analysis of CD5L
   expression. (E‐F) Quantification of CD5L^+ macrophages in HuCC‐T1 E)
   and RBE F) systems. G,H) CCL18 secretion levels in HuCC‐T1 (G) and RBE
   (H) systems. Data are mean ± SD; n = 3 (^* p < 0.05, ^** p < 0.01, ^***
   p < 0.001).

   These findings highlight the critical role of TNTs in facilitating the
   polarization of macrophages toward the TREM2⁺ phenotype and promoting
   the secretion of CCL18, a chemokine implicated in tumor progression.
   This underscores the functional significance of TNTs‐mediated
   communication between tumor cells and macrophages in reshaping
   macrophage behavior within the TME, ultimately contributing to an
   immunosuppressive and tumor‐supportive microenvironment.

2.5. Effects of TNTs‐Mediated Material Transfer on Macrophage Function

   Tumor‐infiltrating immune cells, including CD8⁺ T cells and
   macrophages, are critical components of the tumor immune
   microenvironment (TIME). Among them, CD8⁺ T cells play a pivotal role
   in anti‐tumor immunity by mounting effective immune responses against
   tumors.^[ [97]^23 ^] However, immunosuppressive cells such as TAMs and
   myeloid‐derived suppressor cells often suppress CD8⁺ T cell function,
   thereby facilitating tumor progression in various cancers, including
   hepatocellular carcinoma.^[ [98]^24 ^] To investigate the effects of
   TNTs‐mediated material transfer on macrophage function, we evaluated
   three key parameters: macrophage phagocytic activity, their regulatory
   effects on CD8⁺ T cell proliferation, and their pro‐tumorigenic
   capacity.

   Flow cytometric analysis revealed that CytoB treatment, which inhibits
   TNTs formation, did not significantly affect macrophage phagocytic
   activity (Figure [99]S3A, Supporting Information). However, in the
   coculture groups with HuCC‐T1 and RBE cells, the phagocytic activity of
   macrophages against E. coli was significantly reduced compared to the
   control groups, as determined by flow cytometry analysis (p < 0.05,
   Figure [100]5A–C for RBE; Figure [101]S3B–D for HuCC‐T1, Supporting
   Information). This reduction was further corroborated by confocal
   microscopy through the measurement of mean fluorescence intensity (MFI)
   and Mander's coefficient (p < 0.05, Figure [102]5D–G). Moreover, the
   phagocytosis of tumor cells provided additional evidence supporting
   this observation (Figure [103]S3E–G, Supporting Information). These
   findings suggest that direct intercellular communication via TNTs
   impairs the functional phagocytic capacity of macrophages. Furthermore,
   when assessing the regulatory effects of macrophages on CD8⁺ T cell
   proliferation, we observed that proliferation was markedly suppressed
   in the Coculture groups compared to other experimental conditions (p <
   0.05, Figure [104]5H–J for RBE; Figure [105]S3H–J for HuCC‐T1,
   Supporting Information).

Figure 5.

   Figure 5
   [106]Open in a new tab

   Functional consequences of TNT‐mediated material transfer. A–C) Flow
   cytometry analysis of E. coli phagocytic activity comparing the
   Coculture group with M0, Transwell, and CytoB groups, respectively,
   along with statistical comparison. Results are shown for the RBE
   coculture system. Data are mean ± SD; n = 3. D,E) Laser confocal
   microscopy showing macrophage phagocytosis of E. coli: M0, Coculture,
   and Transwell groups (D), and DMSO and CytoB groups (E).F,G) MFI and
   Mander's coefficient analysis of phagocytosis. Data are mean ± SD; n =
   3. H–J) Proliferation of CD8^+ T cells co‐cultured with macrophages.
   Results are shown for the RBE coculture system. Data are mean ± SD; n =
   3. K) Schematic of xenograft experiments. L,M) Tumor volume
   comparisons. Data are mean ± SD; n = 6. N) H&E staining of tumors.
   (O‐P) Expression of TREM2 and CD5L in macrophages within xenograft
   tumors from the Coculture and Transwell groups O), and from the DMSO
   and CytoB groups P). Scale bar: 200 µm. Q) The molecular function of
   TREM2^+‐macrophages was significantly enriched in fatty acid binding
   (red box). R) Flowchart of co‐culturing process to observe fatty acids
   transfer. (by Figdraw) S) Transfer of fatty acids from tumor cells to
   macrophages through TNTs. (Left) Bright‐field image of hucc‐t1 cells
   labeled with BODIPY 493/503 (green) co‐cultured with mCherry‐expressing
   macrophages (red). White arrows indicate TNTs bridging tumor cells and
   macrophages. (Right) Time‐lapse fluorescence imaging (t‐axis) reveals
   dynamic lipid trafficking through TNTs. Yellow arrows highlight
   BODIPY‐labeled fatty acid cargos transferred from tumor cells to
   macrophages over time. MFI, mean fluorescence intensity Significance is
   indicated as follows: ^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001.

   To evaluate the pro‐tumorigenic capacity of macrophages, we conducted
   in vivo experiments using a nude mouse xenograft model (workflow shown
   in Figure [107]5K). Tumor volumes were significantly larger in the
   Coculture group compared to the Transwell group (Figure [108]5L), and
   in the DMSO group compared to the CytoB group (Figure [109]5M),
   highlighting the role of TNTs formation in enhancing the
   pro‐tumorigenic potential of macrophages. Hematoxylin and eosin (H&E)
   staining further confirmed the heterogeneity of the xenografts
   (Figure [110]5N).

   Immunofluorescence analysis revealed that macrophages in the Coculture
   group exhibited an increased proportion of TREM2^+ macrophages and a
   decreased proportion of CD5L^+ macrophages compared to the Transwell
   group (Figure [111]5O). Conversely, inhibition of TNTs formation by
   CytoB resulted in a lower proportion of TREM2^+ macrophages and a
   higher proportion of CD5L^+ macrophages compared to the DMSO group
   (Figure [112]5P).

   In summary, TNTs‐mediated material transfer facilitates the
   polarization of macrophages toward a TREM2⁺ pro‐tumorigenic phenotype,
   diminishes their phagocytic activity, and suppresses CD8⁺ T cell
   proliferation. This process not only reinforces the tumor‐promoting
   role of macrophages but also impairs antitumor immune responses,
   contributing significantly to tumor progression in the iCCA
   microenvironment.

2.6. Tumor Cells Transfer Fatty Acids to Macrophages via TNTs

   To identify the materials transferred from tumor cells to macrophages,
   we performed enrichment analysis on scRNA‐seq data from TREM2⁺
   macrophages and CD5L⁺ macrophages. The analysis revealed that genes
   upregulated in the TREM2⁺ macrophage subset were significantly enriched
   in pathways related to fatty acid binding and long‐chain fatty acid
   binding (Figure [113]5Q). These findings align with previous studies
   demonstrating that lipids and their metabolic derivatives are key
   regulators of macrophage polarization, pro‐inflammatory and
   anti‐inflammatory properties, as well as phagocytic function.^[ [114]^2
   ^] Based on this evidence, we hypothesized that tumor cells transfer
   fatty acids to macrophages via TNTs, resulting in lipid accumulation
   and driving their phenotypic transformation.

   To test this hypothesis, we established an experimental workflow
   (Figure [115]5R). Using fluorescent labeling and advanced imaging
   techniques, we demonstrated that fatty acids could be efficiently
   transferred from tumor cells to macrophages through TNTs
   (Figure [116]5S, Video [117]S3, Supporting Information). These findings
   provide novel mechanistic insights into how TNTs facilitate
   intercellular lipid transfer, contributing to macrophage phenotypic
   reprogramming and functional modulation within TME.

2.7. TNTs are the Primary Pathway for Tumor‐Derived Fatty Acid Transfer to
Macrophages

   To identify the primary mechanism by which tumor‐derived fatty acids
   are transferred to macrophages, we evaluated three potential pathways:
   (1) direct transfer via TNTs, (2) macrophage phagocytosis of tumor
   cells, and (3) uptake of tumor‐secreted fatty acids through macrophage
   CD36 scavenger receptors.^[ [118]^25 , [119]^26 ^]

   Using live‐cell confocal microscopy over a 13‐h observation period, no
   evidence of macrophages phagocytosing whole tumor cells was detected,
   ruling out phagocytosis as the primary pathway for fatty acid transfer
   (Video [120]S4, Supporting Information). Additionally, scRNA‐seq
   analysis showed no significant upregulation of CD36 expression in
   macrophages within tumor samples, further diminishing the likelihood of
   CD36‐mediated uptake as the primary mechanism (Figure [121]S4A,
   Supporting Information).

   Tumor‐secreted vesicles have been previously reported to transport
   fatty acids to dendritic cells and TAMs.^[ [122]^27 ^] These vesicles,
   with diameters ranging from 50 to 500 nm, can freely pass through
   Transwell membranes with 1 µm pores.^[ [123]^28 ^] To investigate the
   potential role of vesicle‐mediated fatty acid transfer, we employed a
   Transwell system. Measurements of free fatty acid concentrations in the
   upper chambers of Coculture and Transwell systems revealed no
   significant differences (p > 0.05, Figure [124]S4B,C, Supporting
   Information), indicating that the presence of the Transwell membrane
   did not impair fatty acid availability in the macrophage culture
   system.

   To evaluate the role of TNTs in fatty acid transfer, we analyzed fatty
   acid transfer rates and lipid accumulation in macrophages, with the
   experimental workflow shown in Figure [125]6A,B. Flow cytometry results
   demonstrated that fatty acid transfer rates in the Coculture group were
   significantly higher than those in the Transwell group (p < 0.05,
   Figure [126]6C–E). Moreover, treatment with the actin polymerization
   inhibitor CytoB significantly reduced fatty acid transfer rates (p <
   0.05, Figure [127]6F–H), further supporting the critical role of TNTs
   in this process. To further validate these findings, we quantified the
   percentage of acceptor macrophages receiving fatty acid using
   fluorescence microscopy with BODIPY 493/503 staining. The Coculture
   group exhibited a significantly higher percentage of macrophages with
   transferred fatty acids compared to the Transwell group (p < 0.05,
   Figure [128]6I), consistent with the flow cytometry data. Similarly,
   the percentage was higher in the DMSO control group than in the
   CytoB‐treated group (p < 0.05, Figure [129]6J), providing additional
   evidence that TNTs facilitate fatty acid transfer.

Figure 6.

   Figure 6
   [130]Open in a new tab

   TNTs facilitate tumor‐derived fatty acid transfer to macrophages. A,B)
   Schematic workflows for quantifying fatty acid transfer efficiency (A)
   and macrophage lipid accumulation (B), designed using Figdraw. C–E)
   Comparative fatty acid transfer rates between Coculture and Transwell
   systems in HuCC‐T1 (C) and RBE (D) tumor cells, with pooled statistical
   analysis (E). F–H) Pharmacological inhibition of TNT‐mediated transfer
   using CytoB versus DMSO control in HuCC‐T1 (F) and RBE (G) models,
   summarized in panel H. Data represent mean ± SD; n = 3. I,J)
   Fluorescence microscopy quantification of BODIPY 493/503^+ macrophages:
   Coculture systems exhibited enhanced fatty acid transfer compared to
   Transwell controls (I), while CytoB treatment significantly attenuated
   lipid trafficking (J) (mean ± SEM; n = 3). K–N) Confocal microscopy
   visualization of lipid accumulation: M0 macrophages (negative control),
   Coculture, and Transwell groups (K‐L), and DMSO/CytoB comparisons
   (M‐N). (mean ± SEM; n = 3) (O‐R) Flow cytometric validation of lipid
   accumulation in macrophages: HuCC‐T1 (O‐P) and RBE (Q‐R) coculture
   systems, demonstrating TNT‐dependent lipid transfer (mean ± SEM; n =
   3). Significance levels: ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.
   Scale bar, 10 µm.

   Lipid accumulation in macrophages was further evaluated using BODIPY
   493/503 staining, which specifically labels neutral lipids. In the
   Transwell group and macrophages before coculture, green fluorescent
   puncta were small and evenly distributed. However, macrophages in the
   Coculture group exhibited a marked increase in lipid droplets, observed
   as larger green fluorescent puncta (Figure [131]6K,L). Similar lipid
   accumulation patterns were observed in the DMSO group
   (Figure [132]6M,N). Flow cytometric analysis confirmed a significant
   increase in the MFI of BODIPY staining in macrophages from the
   Coculture group compared to controls, consistent across both HuCC‐T1 (p
   < 0.05, Figure [133]6O,P) and RBE cells (p < 0.05, Figure [134]6Q,R).

   In conclusion, our findings unequivocally demonstrate that TNTs serve
   as the primary pathway for the transfer of fatty acids from tumor cells
   to macrophages. The significantly higher fatty acid transfer rates and
   lipid accumulation observed in the Coculture group underscore the
   dominant role of TNTs in this process.

2.8. TNTs‐Mediated AA Transfer Drives Macrophage Phenotypic Transformation
and Functional Reprogramming

   Metabolic crosstalk between tumor cells and macrophages plays a
   critical role in shaping the immunosuppressive TME.^[ [135]^29 ^]
   M2‐like TAMs are considered the result of tumor cell reprogramming,
   with previous studies highlighting the importance of AA metabolites in
   regulating M2‐TAMs formation and promoting tumor progression.^[
   [136]^30 ^] In this study, we investigated the types of fatty acids
   transferred via TNTs and their impact on macrophage phenotype and
   function using lipidomics, oxidized lipid analysis, and targeted
   metabolic profiling.

   Lipidomics analysis revealed that upregulated lipid molecules in the
   Coculture group macrophages tended to associate with unsaturated fatty
   acid chains, whereas downregulated lipids were primarily linked to
   short‐chain or saturated fatty acids (Figure [137]7A; Table [138]S1,
   Supporting Information). Notably, significantly upregulated lipid
   species, such as PS (24:0/20:4) and PE (24:0/20:4), contained 20:4
   fatty acid chains, likely corresponding to AA. KEGG pathway analysis
   further demonstrated that these upregulated lipids were significantly
   enriched in AA and linoleic acid metabolic pathways (Figure [139]7B).
   Additionally, oxidized lipid products significantly elevated in the
   Coculture group included AA oxidation derivatives (e.g., 5‐HETE,
   17‐HETE, and 9‐HETE)^[ [140]^31 ^] and oxidation products derived from
   other unsaturated fatty acids (e.g., HEPE, HDoHE, and HETrE)^[ [141]^32
   , [142]^33 , [143]^34 ^] (Figure [144]7C; Table [145]S2, Supporting
   Information). Targeted metabolic profiling identified AA and other
   long‐chain unsaturated fatty acids, such as cis‐13,16‐docosadienoic
   acid and cis‐11‐eicosenoic acid, as significantly upregulated in
   macrophages from the Coculture group (Figure [146]7D; Table[147]S3,
   Supporting Information). Collectively, these results suggest that
   long‐chain unsaturated fatty acids, primarily AA, are the major fatty
   acids transferred via TNTs.

Figure 7.

   Figure 7
   [148]Open in a new tab

   TNT‐mediated AA transfer drives macrophage phenotypic and functional
   reprogramming. A) Lipidomics profiling reveals distinct lipid
   compositions in macrophages from Coculture (Co) versus Transwell
   (Trans) systems. B) KEGG pathway enrichment analysis of upregulated
   lipid species in Coculture macrophages. C,D) Specific elevation of
   oxidized lipids (C) and fatty acids (D), including AA, in Coculture
   macrophages. E) Time‐lapse imaging of FITC‐AA‐labeled HuCC‐T1 cells
   (green) transferring AA to mCherry^+ macrophages (red) via TNTs (white
   arrows). Yellow arrows in t‐axis panels (right) track AA transport
   dynamics. F–I) AA‐induced macrophage reprogramming: (F‐G) Flow
   cytometry quantification of TREM2 (F) and CD5L (G) expression in
   AA‐treated versus DMSO‐control macrophages. (H‐I) Immunofluorescence
   validation of TREM2 and CD5L positivity rates in DMSO and AA groups. J)
   Increased CCL18 secretion in AA‐treated macrophages. (mean ± SEM; n =
   3). K,L) Functional consequences of AA transfer: (K) Flow cytometry
   analysis of macrophage phagocytic function on E. coli in AA‐treated
   versus DMSO‐control macrophages. (L) Proliferation of CD8^+ T cells
   co‐cultured with macrophages. Data are mean ± SD; n = 3. Significance
   levels: ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001. AA, arachidonic
   acid; Co, Coculture; Trans, Transwell.

   To further validate this process, HuCC‐T1 cells were treated with
   exogenous FITC‐labeled AA and co‐cultured with macrophages.
   Observations demonstrate that FITC‐labeled AA is transported from
   HuCC‐T1 cells to macrophages via TNTs, as illustrated in
   Figure [149]7E. These findings provide compelling visual evidence of an
   intercellular lipid exchange pathway facilitated by TNTs.

   To explore the functional consequences of TNTs‐mediated AA transfer,
   exogenous AA was added to macrophages. This treatment significantly
   increased TREM2 expression and decreased CD5L expression in macrophages
   (p < 0.05, Figure [150]7F,G). Immunofluorescence analysis further
   confirmed that macrophages receiving AA exhibited a higher percentage
   of TREM2 positivity and a lower percentage of CD5L positivity compared
   to the DMSO control group, with statistically significant differences
   (Figure [151]7H,I), indicating a phenotypic polarization toward the
   TREM2⁺ phenotype. Furthermore, macrophages treated with AA exhibited
   significantly increased secretion of the pro‐tumorigenic chemokine
   CCL18 (p < 0.05, Figure [152]7G).

   Functional assays demonstrated that exogenous AA treatment
   significantly impaired macrophage phagocytic activity, as evidenced by
   reduced E. coli phagocytosis in flow cytometry analysis (p < 0.05,
   Figure [153]7K) and confirmed by confocal microscopy using Mander's
   coefficient (Figure [154]S5A, Supporting Information). Additionally, AA
   treatment compromised phagocytosis of tumor cells (Figure [155]S5B,
   Supporting Information) and markedly suppressed CD8⁺ T cell
   proliferation (p < 0.05, Figure [156]7L).

   These results indicate that AA‐driven macrophage reprogramming weakens
   their antitumor immune responses, thereby reinforcing their
   tumor‐promoting functions.

   In conclusion, our findings demonstrate that TNTs‐mediated AA transfer
   plays a pivotal role in driving macrophage polarization toward the
   TREM2⁺ phenotype and reprogramming their functional properties.

2.9. TNTs‐Mediated Material Transfer Activates the PI3K‐AKT Signaling Pathway
and Regulates Macrophage Phenotype and Function

   To explore the mechanisms by which tumor cell‐driven TNTs formation
   affects macrophages, we performed comparative transcriptome analysis on
   macrophages in the Coculture and Transwell systems (Table [157]S4,
   Supporting Information). Transcriptome analysis revealed significant
   enrichment of pathways associated with tumor growth and lipid
   metabolism, including the PI3K‐AKT signaling pathway and pathways
   related to lipid metabolism and atherosclerosis (Figure [158]8A).^[
   [159]^35 ^] GO functional analysis further indicated enrichment of
   biological processes critical for tumor initiation and progression,
   such as the positive regulation of IL‐8 production and regulation of
   intercellular adhesion (Figure [160]8B).^[ [161]^36 , [162]^37 ^] These
   findings suggest that TNTs formation significantly alters the gene
   expression profile of macrophages, particularly genes involved in tumor
   progression, immune regulation, and lipid metabolism.

Figure 8.

   Figure 8
   [163]Open in a new tab

   TNT‐mediated material transfer activates PI3K‐AKT signaling to modulate
   macrophage phenotype and function (A‐B) Pathway analysis of
   Coculture‐educated macrophages: A) KEGG enrichment of upregulated
   signaling pathways. B) Gene Ontology (GO) functional annotation of
   differentially expressed genes. (C‐H) PI3K‐AKT pathway activation: C)
   Western blot analysis of p‐PI3K and p‐AKT in M0 versus Coculture
   macrophages. D,E) Quantification of p‐PI3K/PI3K (D) and p‐AKT/AKT (E)
   ratios. Data are mean ± SD; n = 3. F–H) Reduced PI3K‐AKT activation in
   Transwell and CytoB‐treated groups compared to Coculture controls. Data
   are mean ± SD; n = 3. I–N) Phenotypic marker modulation by PI3K
   inhibition: TREM2 (I‐J), CD5L (K‐L), and CCL18 (M‐N) expression in
   Coculture macrophages treated with LY294002 (PI3K inhibitor) versus
   untreated controls, validated in HuCC‐T1 and RBE systems. Data are mean
   ± SD; n = 3. (O‐T) Functional consequences of PI3K‐AKT signaling: O,P)
   Flow cytometric quantification of macrophage phagocytic function
   against E. coli in HuCC‐T1 (O) and RBE (P) system. Q) Confocal
   microscopy analysis of macrophage phagocytic function against E. coli
   (Mander's coefficient). R) Assessment of macrophage phagocytic function
   against tumor cells. S,T) Proliferation of CD8^+ T cells in HuCC‐T1
   system (S) and RBE system (T). Data are mean ± SD; n = 3. Significance
   levels: ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

   Specifically, genes upregulated in macrophages from the Coculture group
   were significantly enriched in the PI3K‐AKT signaling pathway,
   indicating that TNTs‐mediated material transfer might activate this
   pathway (Figure [164]8A). This hypothesis was validated by western blot
   analysis, which showed a marked increase in the p‐AKT/AKT and
   p‐PI3K/PI3K ratios in macrophages after Coculture compared to baseline
   levels (Figure [165]8C–E). Furthermore, macrophages in the Coculture
   group exhibited significantly higher activation of the PI3K‐AKT pathway
   compared to the Transwell and CytoB groups (Figure [166]8F–H),
   confirming that TNTs‐mediated material transfer induces PI3K‐AKT
   signaling activation.

   To further elucidate the role of the PI3K‐AKT pathway in macrophage
   phenotypic transformation and functional reprogramming, we added the
   PI3K‐AKT pathway inhibitor LY294002 to the Coculture system. Inhibition
   of this pathway significantly decreased TREM2 expression
   (Figure [167]8I,J) and increased CD5L expression (Figure [168]8K,L).
   Additionally, the secretion of the pro‐tumorigenic chemokine CCL18 was
   markedly reduced (Figure [169]8M,N). Inhibition of the PI3K‐AKT pathway
   partially restores macrophage phagocytic capacity, as demonstrated by
   increased phagocytosis of E. coli in flow cytometry (Figure [170]8O,P)
   and confocal microscopy with Mander's coefficient (Figure [171]8Q),
   alongside enhanced phagocytosis of tumor cells (Figure [172]8R). This
   indicates that the pathway suppresses macrophage function critical for
   clearing tumor cells, and its inhibition alleviates this suppression.
   Furthermore, CD8⁺ T cell proliferation is partially recovered
   (Figure [173]8S,T), suggesting that PI3K‐AKT signaling contributes to
   immune suppression by limiting cytotoxic T cell activity, with its
   blockade enhancing anti‐tumor immunity. These results indicate that
   PI3K‐AKT signaling activation is a critical mediator of TNTs‐driven
   macrophage reprogramming.

   In conclusion, our findings demonstrate that TNTs‐mediated material
   transfer activates the PI3K‐AKT signaling pathway in macrophages,
   driving their phenotypic transformation toward the TREM2⁺
   pro‐tumorigenic phenotype and functional reprogramming. This activation
   not only modulates macrophage phenotype and function but also fosters
   an immunosuppressive microenvironment, thereby promoting tumor
   progression.

2.10. AA Activates the PI3K‐AKT Signaling Pathway and Regulates Macrophage
Phenotype and Function

   As previously described, TNTs‐mediated fatty acids transfer mainly
   involves AA and other long‐chain fatty acids, with transcriptomic
   analysis identifying PI3K‐AKT signaling pathway activation as a
   downstream effect of this transfer. To determine whether AA directly
   activates this pathway and regulates macrophage phenotypic and
   functional changes, we conducted a series of experiments.

   Western blot analysis demonstrated that treatment with exogenous AA
   significantly increased the p‐AKT/AKT and p‐PI3K/PI3K ratios in
   macrophages, indicating activation of the PI3K‐AKT signaling pathway (p
   < 0.05, Figure [174]9A–C). Conversely, adding the pathway inhibitor
   LY294002 significantly reduced these ratios, confirming that the
   pathway was effectively suppressed (p < 0.05, Figure [175]9A–C). These
   results confirm that AA can activate the PI3K‐AKT signaling pathway,
   suggesting that TNTs‐mediated AA transfer may regulate macrophage
   function through this mechanism.

Figure 9.

   Figure 9
   [176]Open in a new tab

   AA activates PI3K‐AKT signaling to modulate macrophage phenotype and
   function. A–C) PI3K‐AKT pathway activation by AA: (A) Western blot
   analysis of p‐PI3K and p‐AKT in macrophages treated with AA or AA +
   LY294002 (PI3K inhibitor). (B,C) Quantification of p‐PI3K/PI3K (B) and
   p‐AKT/AKT (C) ratios, demonstrating AA‐induced pathway activation and
   its suppression by LY294002. Data are mean ± SD; n = 3. D–F) Phenotypic
   marker regulation: Expression levels of TREM2 (D), CD5L (E), and CCL18
   (F) in macrophages treated with AA or AA+ LY294002. Data are mean ± SD;
   n = 3. (G‐I) Functional impact of AA‐mediated signaling: G) Flow
   cytometric quantification of macrophage phagocytic function against E.
   coli. H) Confocal microscopy analysis of macrophage phagocytic function
   against E. coli (Mander's coefficient). I) Assessment of macrophage
   phagocytic function against tumor cells. Data are mean ± SD; n = 3. J)
   The ability to stimulate CD8^+ T cell proliferation in macrophages
   treated with AA or AA+ LY294002. AA, arachidonic acid. Significance
   levels: ^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001.

   To validate whether AA‐induced macrophage phenotypic and functional
   changes depend on the PI3K‐AKT pathway, we co‐treated macrophages with
   exogenous AA and LY294002. Flow cytometry revealed that, compared to AA
   treatment alone, AA+ LY294002 treatment significantly reduced TREM2
   expression (Figure [177]9D), increased CD5L expression
   (Figure [178]9E), and decreased CCL18 secretion (p < 0.05,
   Figure [179]9F). These findings indicate that inhibiting the PI3K‐AKT
   pathway reverses AA‐induced macrophage polarization. Functional assays
   demonstrated that macrophages in the AA + LY294002 group exhibited
   significantly enhanced phagocytic activity. This was evidenced by
   increased phagocytosis of E. coli, as confirmed through flow cytometry
   analysis (Figure [180]9G) and laser confocal microscopy using Mander's
   coefficient (Figure [181]9H). Additionally, the group showed improved
   phagocytosis of tumor cells (Figure [182]9I). Furthermore, a
   statistically significant improvement in CD8⁺ T cell proliferation was
   observed in the AA + LY294002 group compared to the AA treatment alone
   (P < 0.05, Figure [183]9J). This suggests that blocking the PI3K‐AKT
   pathway partially restores antitumor immune functions, counteracting
   AA‐mediated functional reprogramming.

   In summary, our findings demonstrate that AA induces macrophage
   phenotypic transformation and functional reprogramming by activating
   the PI3K‐AKT signaling pathway. Inhibiting this pathway partially
   reverses these changes, offering novel insights into the metabolic
   regulation of macrophages within the TME and identifying potential
   therapeutic targets.

3. Discussion

   This study is the first to confirm that in iCCA, tumor cells transfer
   long‐chain fatty acids, primarily AA, to macrophages via TNTs. This
   process drives macrophage polarization toward a pro‐tumorigenic TREM2⁺
   phenotype while suppressing macrophage phagocytic function and reducing
   their ability to stimulate CD8⁺ T cell proliferation. Further
   investigation revealed that TNTs‐mediated AA transfer significantly
   activates the PI3K‐AKT signaling pathway, a key mechanism in macrophage
   phenotypic transformation and functional reprogramming. Inhibition of
   this pathway partially reversed the immune‐suppressive phenotype,
   restored macrophage function, and enhanced anti‐tumor immune responses.
   These findings provide novel mechanistic insights into TNTs‐mediated
   tumor‐immune cell interactions and identify TNTs and the PI3K‐AKT
   pathway as promising therapeutic targets for modulating the TME.

   While cytokines, exosomes, and microvesicles have been extensively
   studied as mediators of intercellular communication, their diffusion
   within the dynamic 3D tumor stroma often limits the specificity and
   efficiency of signaling.^[ [184]^4 , [185]^6 ^] In contrast, TNTs serve
   as direct conduits for the targeted transfer of molecules, providing a
   deterministic and efficient communication route. Previous studies have
   highlighted TNTs’ role in promoting cancer cell metabolic plasticity,
   invasiveness, and therapy resistance through the transport of RNA,
   proteins, and organelles.^[ [186]^20 ^] However, research on TNTs in
   iCCA is sparse, and this study is the first to demonstrate their role
   in directly linking tumor cells and immune cells, thereby facilitating
   the formation of an immunosuppressive TME.

   Macrophage phenotypes are shaped by their microenvironment, with M1 and
   M2 macrophages representing two extremes of activation. The
   anti‐inflammatory and pro‐tumorigenic effects of M2‐like macrophages
   are well‐documented but fail to capture the full complexity of
   macrophage activation states.^[ [187]^38 ^] In this study, we
   identified two distinct macrophage populations in iCCA: pro‐tumor
   TREM2⁺ macrophages and anti‐tumor CD5L⁺ macrophages, further clarifying
   the spectrum of macrophage phenotypic plasticity in TME remodeling.
   Consistent with prior findings in other cancers, AA and its oxidized
   derivatives were shown to promote TAMs polarization toward a
   pro‐tumoral phenotype.^[ [188]^5 ^] This study advances the field by
   demonstrating that TNTs‐mediated AA transport is a direct driver of
   macrophage functional reprogramming, emphasizing the interplay between
   lipid transfer and immune regulation.

   The PI3K‐AKT signaling pathway is a well‐established regulator of
   immune cell activation and macrophage polarization.^[ [189]^39 ^] Our
   study demonstrated that TNTs‐mediated AA transfer activates this
   pathway, inducing phenotypic transformation of macrophages toward a
   TREM2⁺ phenotype characterized by reduced phagocytic activity and
   impaired CD8⁺ T cell support. Importantly, pathway inhibition with
   LY294002 reversed these changes, restoring macrophage antitumor
   functions and reducing the secretion of pro‐tumorigenic chemokines such
   as CCL18. These findings not only confirm the central role of the
   PI3K‐AKT pathway in macrophage reprogramming but also highlight its
   potential as a therapeutic target for TME modulation.

   TNTs are capable of transporting various substances, including lipids,
   proteins, RNA, and organelles, and play an important role in regulating
   the TME.^[ [190]^20 ^] Our study reveals that AA‐dominant long‐chain
   fatty acids, transported to macrophages via TNTs, may be key factors
   driving macrophage phenotype transformation. Increasing evidence
   suggests that TAMs adapt to their metabolic needs by enhancing lipid
   uptake or storage, which is closely linked to their immune‐suppressive
   functions.^[ [191]^25 ^] It has also been shown that tumor‐derived
   long‐chain fatty acids can enter macrophages via extracellular
   vesicles, thereby providing fuel for their pro‐tumoral activity.^[
   [192]^26 ^] Our findings further confirm that fatty acids are
   transferred from tumor cells to macrophages through TNTs, representing
   one of the primary pathways for tumor‐derived lipid transfer. Although
   lipids may also be transported via vesicles, TNTs provide a more
   directed and purposeful mechanism for lipid delivery to macrophages.
   However, other molecules transported by TNTs, such as proteins,
   signaling molecules, and RNA, may work synergistically with lipids or
   independently influence macrophage function. We further validated the
   critical role of AA in macrophage phenotype transformation,
   demonstrating that exogenous AA can significantly induce macrophages to
   polarize toward a pro‐tumoral TREM2^+ phenotype, inhibit their
   phagocytic function, and reduce their immune stimulatory effects on
   CD8^+ T cells, even under conditions that do not rely on TNTs. In
   conclusion, although TNTs transport a variety of substances, AA, as a
   key lipid molecule, can independently trigger macrophage functional
   reprogramming. Future studies should explore how other components
   transported via TNTs may cooperate with AA to jointly regulate the
   dynamic changes in the TME.

   In this study, after AA is transported to macrophages via TNTs, it
   significantly activates the PI3K‐AKT signaling pathway, driving
   functional reprogramming. The PI3K‐AKT signaling pathway has been
   widely demonstrated to play a key role in immune cell metabolic
   adaptation and phenotype transformation^[ [193]^39 ^] suggesting that
   the activation of this pathway by AA may be somewhat specific. However,
   AA may not be the only acting molecule, and whether it cooperates with
   other metabolic pathways to jointly drive the activation of the
   PI3K‐AKT pathway requires further investigation.

   TNTs are not exclusive to iCCA; they are widely present in cancers such
   as breast, lung, and pancreatic cancers, primarily facilitating cancer
   cell‐to‐cell communication.^[ [194]^20 ^] This study is the first to
   highlight their role in tumor‐immune cell interactions in iCCA, raising
   the possibility that TNTs may mediate similar immunosuppressive effects
   in other cancers, such as hepatocellular carcinoma and pancreatic
   cancer. Future studies could investigate the role of TNTs in regulating
   other immune cells, including T cells and dendritic cells, providing a
   broader understanding of their impact on cancer immunotherapy.

   This study primarily relied on in vitro coculture systems and nude
   mouse xenograft models, which do not fully capture the complexity of
   the human TME. Future validation using patient‐derived organoids or
   xenograft models could provide more clinically relevant insights.
   Additionally, while this study focused on AA, other molecules
   transported via TNTs, such as proteins, nucleic acids, and
   mitochondria, may independently or synergistically influence macrophage
   function and TME remodeling. Advanced techniques, including proteomics
   and single‐cell multi‐omics, could uncover the broader repertoire of
   TNTs‐transported molecules and their roles in immune regulation.

   Moreover, the potential interactions between the PI3K‐AKT pathway and
   other regulatory pathways, such as mTOR and NF‐κB, remain unexplored.
   Investigating these interactions, as well as the effects of TME
   conditions such as hypoxia and acidic pH on TNTs formation, could yield
   additional insights into the regulation of immune cell behavior in
   tumors. Finally, constructing a TME cell communication network using
   live‐cell imaging and high‐resolution techniques could comprehensively
   reveal the dynamic interactions mediated by TNTs.

4. Conclusion

   This study establishes TNTs as a core mechanism for material transfer
   between tumor cells and macrophages in iCCA, with AA acting as a
   critical cargo molecule that activates the PI3K‐AKT signaling pathway.
   By driving TAMs polarization and functional reprogramming, TNTs
   contribute to the immunosuppressive TME and promote tumor progression.
   Targeting TNTs formation and the PI3K‐AKT pathway represents a
   promising strategy for cancer immunotherapy, with potential
   applications across multiple tumor types.

5. Experimental Section

ScRNA‐Seq Analysis

   It was obtained scRNA‐seq data for iCCA ([195]GSE138709) from the Gene
   Expression Omnibus (GEO) database
   ([196]https://www.ncbi.nlm.nih.gov/geo/). Ethical approval was
   unnecessary as per the GEO guidelines. Dimensionality reduction was
   conducted, clustering, subclustering, and mitigating batch effects.
   Cell clustering was performed with a resolution of 0.8. Marker genes
   were identified for biliary epithelial cells, malignant cells, and
   macrophages based on previous studies.^[ [197]^12 ^] Subsequently,
   epithelial cells was extracted and analyzed marker genes (|logFC|> 1,
   adj.p < 0.05), followed by GO and KEGG enrichment analysis.
   Subpopulations of macrophages were identified and further clustered
   (resolution = 0.3), with two subpopulations showing significant
   distribution differences between the tumor and normal groups. Marker
   genes were identified for the two subpopulations (|logFC| > 0.25, adj.p
   < 0.05). R packages Seurat, Harmony, org.Hs.eg.db, and ggplot2 was
   utilized for analysis and visualization.

Cell Lines

   The iCCA cell line HuCC‐T1 was obtained from Pricella (Catalog No.
   CL‐0725, China). The iCCA cell line RBE and the human monocytic cell
   line THP‐1 were sourced from the Institute of Biochemistry and Cell
   Biology, Chinese Academy of Sciences (Shanghai, China). Both cell lines
   were cultured in 1640 medium (Gibco BRL), supplemented with 10% FBS
   (Gibco) and 1% penicillin‐streptomycin at 37 °C with 5% CO[2]. HIBEpiC
   cell line was sourced from ZQXZbio (Catalog No. ZQY009, China) and
   cultured in HIBEpiC‐Im medium (ZQXZbio, Catalog No. ZMY009).

   THP‐1 cells were induced to differentiate into macrophages by treatment
   with 100 ng mL^−1 phorbol‐12‐myristate‐13‐acetate (PMA, S1819,
   Beyotime) for 24 h. Subsequently, the cells were cultured in complete
   medium for an additional 24 h to mitigate the influence of PMA,
   resulting in the generation of M0.

Patient Samples

   Tissue sections were obtained from three iCCA cases and their adjacent
   normal tissues. The diagnoses were confirmed through histological and
   clinical examinations. None of the patients had undergone preoperative
   chemotherapy, radiotherapy, or immunotherapy. Written informed consent
   was obtained from all patients prior to sample collection. Ethical
   approval was granted by the Research Ethics Committee of the Second
   Hospital of Hebei Medical University (approval letter no. 2024‐R474).

Transfection

   Lentivirus transfection was conducted following previously established
   protocols.^[ [198]^40 ^] HuCC‐T1 cells were transfected with GFP
   negative control lentivirus (Genechem, China), while THP‐1 cells were
   transfected with mCherry negative control lentivirus (Genechem, China).
   Stably transfected cell lines were obtained by selection with 1ug mL^−1
   puromycin (BL528A, biosharp) for 7 days.

Live Cell Microscopy

   Cells from various treatment groups were transferred to a Live‐Cell
   Microscope Incubation System (TOKAI HIT) and observed under a laser
   confocal microscope at 37 °C with 5% CO₂.

Cell Staining

   For F‐actin staining, cells were fixed and permeabilized with Triton
   X‐100. The cells were then stained with Phalloidin AF488 (ShareBio,
   YP0059).

   mCherry‐THP‐1 cells were induced to differentiate into mCherry‐M0 on
   gelatin‐coated coverslips and co‐cultured with logarithmically growing
   HuCC‐T1 cells. For DiO staining, HuCC‐T1 cells were labeled with DiO
   (C1993S, Beyotime), co‐cultured with mCherry‐M0 cells. Subsequently,
   cells were staining with Hoechst 33 342, and observed under a laser
   confocal microscope.

Cell Staining—Live Cell Tracing Staining

   CellTracker Green CMFDA (40721ES50) was purchased from Yeasen Biotech
   Co., Ltd. (Shanghai, China). Cells were stained with either CellTracker
   Green CMFDA at a final concentration of 1 µm for 30 min at 37 °C.
   Following staining, cells were washed twice to remove excess dye and
   subsequently cultured in fresh medium for further experiments.

   For lipid staining, cells were stained using BODIPY 493/503 (D3922,
   Invitrogen) at 37 °C for 30 min. Following staining, further
   experiments were conducted.

   For FITC‐AA staining, tumor cells were treated with FITC‐AA (R‐FT‐0020,
   Stargray) at 37 °C for 30 min. Following staining, further experiments
   were conducted.

Treatment of Different Culture Systems—Coculture System

   mCherry‐THP‐1 cells were seeded onto gelatin‐coated coverslips and
   induced to differentiate into mCherry‐M0. HuCC‐T1 or RBE cells were
   then seeded into the same culture dish at a 1:1 ratio and co‐cultured
   for 24 h.

Treatment of Different Culture Systems—Transwell Assay

   HuCC‐T1 or RBE cells were seeded into the upper chambers of a transwell
   system with 1 µm pores, while mCherry‐M0 were plated in the lower
   chambers. After 24 h of co‐culture, allowing free diffusion of
   molecules between the two chambers while preventing intercellular TNTs
   formation.

Treatment of Different Culture Systems—CytoB Treatment

   To inhibit the formation of TNTs, a known inhibitor of actin
   polymerization^[ [199]^41 ^] 350 nm CytoB was added to the co‐culture
   system (ab143482, Abcam). For the control group, an equivalent volume
   of DMSO, the solvent for CytoB, was added.

Treatment of Different Culture Systems—AA Treatment of Macrophages

   Macrophages were cultured in medium supplemented with 20 ng mL^−1 AA
   (A3611, Sigma, United States) for 24 h, designated as the AA group. For
   the control group, an equivalent volume of DMSO, the solvent for AA,
   was added. Following incubation, the cells were collected for
   subsequent experiments.

Treatment of Different Culture Systems—Cell Treatment With LY294002

   The cells were stimulated with LY294002 (10 µm) for 6 h.

Determination of Free Fatty Acids

   Free fatty acids in the medium of the cocultured system and the top of
   the transwell chamber were determined using a nonesterified free fatty
   acid assay kit (A042‐2‐1, Nanjing Jiancheng Bioengineering Institute)
   following the manufacturer's instructions.

Analysis of Fatty Acids Transfer Rate

   HuCC‐T1 or RBE cells were stained with BODIPY 493/503 before being
   seeded in the culture system at a 1:1 ratio. Cells in Coculture and
   Transwell groups, as well as in DMSO and CytoB groups, were analyzed
   using flow cytometry and fluorescence microscopy. The percentage of
   BODIPY‐positive mCherry‐expressing cells was quantified using FlowJo
   software for flow cytometry data and ImageJ for fluorescence microscopy
   images.

Analysis of Macrophage Lipid Accumulation

   Cells from different treatment groups were stained with BODIPY 493/503
   and analyzed using flow cytometry to measure the MFI of BODIPY in
   mCherry‐positive cells. Additionally, cells were co‐stained with BODIPY
   493/503 and Hoechst 33 342, and the MFI of BODIPY in mCherry‐positive
   cells was evaluated in each group using laser confocal microscopy.

Measurement of Macrophage Phenotype

   Cells from each group, both treated and untreated, were prepared as
   single‐cell suspensions and stained with anti‐TREM2 (SC‐373828, Santa
   Cruz) and anti‐CD5L (SC‐390486, Santa Cruz), respectively. Flow
   cytometry analysis was then performed to determine the proportion of
   antibody‐positive cells among the mCherry‐positive population in each
   group.

ELISA

   The CCL18 protein level in the cell culture supernatant was measured
   using the CCL18 ELISA kit (RAB0051, Sigma–Aldrich, Germany) according
   to the manufacturer's protocol.

Phagocytosis Assay

   PHrodo Green‐labeled E. coli BioParticles ([200]P35366, Thermo Fisher
   Scientific) were reconstituted in Hank's balanced salt solution with 5%
   FBS, and then added to the different treatment groups at a ratio of 1
   vial per 2 × 10^6 cells, following the manufacturer's instructions.
   After co‐incubation for 4 h, the cells were subjected to either flow
   cytometry or confocal laser scanning analysis. FlowJo or Image J
   software were utilized to assess and analyze the proportion of Green^+
   macrophages relative to the total macrophage population. Additionally,
   ImageJ was employed to calculate Mander's coefficient for
   colocalization analysis.^[ [201]^42 ^] HuCC‐T1 cells were labeled with
   CellTrace Violet ([202]C34557, Thermo Fisher) following the
   manufacturer's protocol. Macrophages from different treatment groups
   were prepared as described previously. Tumor cells and macrophages were
   co‐cultured at a 2:1 ratio for 4 h in a humidified incubator at 37 °C
   with 5% CO[2]. Phagocytosis was assessed by flow cytometry. Data were
   analysed by calculating the percentage of Violet‐positive cells among
   macrophages using FlowJo software.

Transwell Assay

   HuCC‐T1 cells were collected and resuspended in serum‐free RPMI 1640
   medium. A Corning Transwell system with 8‐µm pore inserts was assembled
   in a 6‐well plate. The lower compartment of each well was filled with
   RPMI 1640 medium supplemented with 10% FBS, while the upper compartment
   was loaded with an equivalent volume of serum‐free RPMI 1640 containing
   10 000 HuCC‐T1 cells. To induce chemotaxis, recombinant CCL18
   (4 ng mL^−1) was introduced into the lower compartment. The setup was
   incubated for 20 h at 37 °C in a 5% CO₂ environment. Post‐incubation,
   the cells were fixed using methanol and stained with crystal violet.
   The number of cells that migrated to the lower side was determined by
   counting cells in three randomly chosen microscopic fields under a
   light microscope.

CCK‐8 Counting Kit Assay

   Cell proliferation was evaluated using the CCK‐8 assay in accordance
   with the manufacturer's guidelines. HuCC‐T1 cells were plated in a
   96‐well plate at a density of 1000 cells per well. 10 µL of CCK‐8
   reagent was introduced into each well, followed by a 2‐h incubation at
   37 °C. Absorbance readings were subsequently obtained at 450 nm using a
   multi‐mode microplate reader. Data were normalized relative to the
   absorbance values of the control group for comparative analysis.

Lipidomics (LC‐MS/MS)

   Lipidomics analysis was performed using liquid chromatography‐tandem
   mass spectrometry. The experimental workflow included sample
   collection, lipid extraction, LC‐MS/MS detection, and data analysis.
   Differential compounds were selected based on the criteria of VIP>1.0,
   FC>1.5 or FC<0.667, and p‐value<0.05.

Oxidized Lipids

   Oxidized lipid species were detected using LC‐MS based on the
   high‐sensitivity SCIEX QTRAP 6500+ mass spectrometer platform. The
   threshold for selecting differential metabolites was set at a FC>1.5 or
   FC<0.667 with a p‐value<0.05. Hierarchical clustering analysis was
   performed on all differential metabolites between the comparison
   groups. The relative quantification values of the differential
   metabolites were normalized and clustered.

Targeted Metabolomics

   An ultra‐high performance liquid chromatography coupled with tandem
   mass spectrometry (UHPLC‐MS/MS) system (ExionLC AD UHPLC‐QTRAP 6500+,
   AB SCIEX Corp., Boston, MA, USA). Materials and Reagents: All 50 fatty
   acid standards were obtained from ZZ Standards Co., Ltd. (Shanghai,
   China). Isopropanol (Optima LC‐MS), acetonitrile (Optima LC‐MS), and
   formic acid (Optima LC‐MS) were purchased from Thermo Fisher Scientific
   (Fair Lawn, NJ, USA). Ultrapure water was obtained from Millipore (MA,
   USA). Metabolite Extraction: Metabolite extraction was performed using
   chromatographic and mass spectrometric methods. Differential
   metabolites were selected based on an FC threshold of > 1.2 or < 0.833,
   with a p‐value < 0.05.

Macrophage Isolation from Coculture System

   Macrophages were isolated using the EasySep Human CD11b Positive
   Selection and Depletion Kit (STEMCELL, Vancouver, Canada). Cells were
   dissociated using accutase. The cell suspension was centrifuged, and
   the pellet was resuspended in sorting buffer at a concentration of 1 ×
   10⁷ cells mL^−1. CD11b antibody mix was added, followed by magnetic
   beads. After incubation, the sample was placed in an EasySep magnet for
   3 min, and unbound tumor cells were discarded. The CD11b‐positive
   macrophages were then collected by resuspending the magnet‐bound cells
   in sorting buffer.

Proliferation Assay of T Cells

   CD8⁺ T cells were isolated from peripheral blood mononuclear cells
   (PBMCs) using the MojoSort Human CD8 Nanobeads Positive Selection Kit
   (BioLegend, 480 107). Briefly, PBMCs were separated from whole blood
   using density gradient centrifugation and washed with PBS. The cells
   were resuspended in MojoSort Buffer and incubated with 10 µL of
   MojoSort Human CD8 Nanobeads for 15 min on ice. After washing, cells
   were placed in the MojoSort magnet, and CD8⁺ T cells were isolated by
   removing the unbound fraction. The purified CD8⁺ T cells were
   resuspended in culture medium for further analysis. Macrophages and
   HuCC‐T1/RBE cells were separated using immunomagnetic bead isolation.
   The isolated CD8⁺ T cells were labeled with CellTrace Violet
   ([203]C34557, Thermo Fisher) and added to each treatment group of
   macrophages at a 10:1 ratio (T cells: macrophages). The co‐culture was
   supplemented with 10 µg mL^−1 anti‐human CD3 (AH1003110, PBM),
   2 µg mL^−1 anti‐human CD28 (AH1028110, PBM), and 50 ng mL^−1
   recombinant human IL‐2 (200‐02‐10, PeproTech). After 48 h, cells were
   collected, and T cell proliferation was assessed by CellTrace Violet
   dilution via flow cytometry.

RNA Extraction, Library Construction, and Sequencing

   Macrophages from both the Coculture and Transwell groups were
   collected, with cells from the Coculture group being separated using
   immunomagnetic bead isolation. Total RNA was extracted from three
   independent cell samples using TRIzol reagent (Aidlab, Beijing, China).
   RNA sequencing was performed by KaiTai‐Bio (Hangzhou, China). The
   integrity, concentration, and purity of the RNA samples were analyzed.
   cDNA was synthesized using the U‐mRNAseq Library Prep Kit (AT4221). The
   prepared samples were sequenced on an Illumina NovaSeq 6000 platform.
   Data cleaning involved removing reads containing adapter contamination,
   low‐quality reads, and reads with over 5% “N” bases using fastp
   (v0.21.0). Reads were aligned to the human genome (GRCh38) annotated in
   the Ensembl database (Homo_sapiens.GRCh38.101.gtf) using HISAT2
   (version 2.2.1) with default parameters. Differentially expressed genes
   were identified using the limma package, with an adjusted P value <0.05
   and logFC >1. Biological pathway analysis was conducted using the
   clusterProfiler package.

Western Blot Analysis

   Each treatment group of macrophages underwent lysis with RIPA buffer
   containing protease inhibitors. Following centrifugation, the protein
   concentration was determined using a Bradford assay, and ≈100 µg of
   protein was loaded onto 10% SDS‐PAGE gels and transferred onto PVDF
   membranes. After blocking, membranes were incubated with primary
   antibodies overnight at 4 °C, followed by fluorescence‐conjugated
   secondary antibodies (LI‐COR Biosciences). Protein bands were
   visualized using an Odyssey infrared imaging system and quantified with
   ImageJ software. The primary antibodies included anti‐PI3K (CY6951,
   1:500, Abways), anti‐p‐PI3K (CY6427, 1:500, Abcam), anti‐AKT (CY5561,
   1:500, Abcam), anti‐p‐AKT (CY5855, 1:500, Abcam), and anti‐GAPDH
   (1:5000, Abways), serving as an internal reference.

The Animal Tumor Xenograft Model

   A tumor xenograft model was established by subcutaneously injecting a
   10:1 mixture of HuCC‐T1 cells and macrophages into the backs of
   5‐week‐old BALB/c nude mice (20–25 g; HFK BIOSCIENCE, China). A total
   of 24 mice were divided into four groups: (1) Coculture group, where
   macrophages isolated from the co‐culture system were mixed with
   untreated HuCC‐T1 cells; (2) Transwell group, where macrophages
   isolated from the Transwell system were mixed with untreated HuCC‐T1
   cells; (3) DMSO group, where macrophages isolated from the DMSO group
   were mixed with untreated HuCC‐T1 cells; and (4) CytoB group, where
   macrophages isolated from the CytoB group were mixed with untreated
   HuCC‐T1 cells.  Each group received 5 × 10⁶ tumor cells in a total
   volume of 100 µL via subcutaneous injection. Tumor volume and body
   weight were measured twice a week, with tumor volume calculated using
   the formula V = (a × b^2)/2, where a and b represent the largest and
   smallest tumor diameters, respectively. Mice were sacrificed three
   weeks after tumor implantation. Throughout the experimental period, no
   significant body weight loss was observed in the mice. All animal
   experiments were approved by the Research Ethics Committee of the
   Second Hospital of Hebei Medical University (Approval No. 2024‐AE270)
   in accordance with laboratory animal welfare and ethical guidelines.

H&E Staining

   First, deparaffinize the tissue sections and rehydrate them through a
   graded ethanol series. Then, stain the cell nuclei with hematoxylin and
   the cytoplasm with eosin. Finally, dehydrate, clear, and mount the
   sections for microscopic examination.

Immunofluorescence/Multiple Immunofluorescence

   Tissue sections were sequentially treated with endogenous peroxidase
   blocking, permeabilization, and BSA blocking. The sections were then
   incubated overnight at 4 °C with primary antibodies: anti‐TREM2
   (ER1918‐04, 1:500, huabio), anti‐CD5L (17224‐1‐AP, 1:500, Proteintech),
   anti‐CD68 (PA6293, 1:500, Elabscience) and anti‐α‐tubulin (ab179484,
   1:500, abcam). Afterward, sections were incubated with HRP‐conjugated
   secondary antibodies. Signal amplification was performed using the TSA
   Kit (SB‐Y6112, ShareBio), followed by antibody elution with elution
   buffer at 37 °C. The sections were then mounted using an anti‐fade
   reagent. Imaging and data analysis were carried out using a laser
   scanning confocal microscope and ImageJ software.

Statistical Methods

   Data were analyzed using R software (version 4.2.2) with appropriate
   packages and statistical methods. ImageJ and FlowJo software were used
   for analysis. SPSS (version 26) was utilized for further statistical
   analysis, and GraphPad Prism (version 9.5.1) was employed for data
   visualization. Statistical comparisons between the two groups were
   conducted using unpaired two‐tailed heteroscedastic t‐tests. Multiple
   groups were compared using one‐way ANOVA tests with Tukey's post‐hoc
   correction for multiple comparisons. In each experiment, ≈100 cells
   were randomly selected and analyzed under a confocal microscope to
   count the number of TNTs and determine the percentage of cells
   connected by them. Experiments were performed in triplicate. Each
   experiment included at least three biological replicates. A
   significance level of p < 0.05 was considered statistically
   significant, denoted as ^* p < 0.05, ^** p < 0.01, and ^*** p < 0.001.

Abbreviations

   iCCA, intrahepatic cholangiocarcinoma; TAMs, tumor‐associated
   macrophages; TNTs, tunneling nanotubes; AA, arachidonic acid; TME,
   tumor microenvironment; scRNA‐seq, single‐cell RNA sequencing;
   HIBEpiCs, human intrahepatic biliary epithelial cells; MIF, multiple
   immunofluorescence; rCCL18, recombinant CCL18; CCK‐8, Cell counting
   kit‐8; TIME, tumor immune microenvironment; ELISA, enzyme‐linked
   immunosorbent assay; GEO, Gene Expression Omnibus; PMA,
   phorbol‐12‐myristate‐13‐acetate; CytoB, Cytochalasin B; MFI, mean
   fluorescence intensity; FC, fold change; PBMCs, peripheral blood
   mononuclear cells; H&E staining, hematoxylin and eosin staining.

Ethics Approval

   The animal study was approved by the Research Ethics Committee of the
   Second Hospital of Hebei Medical University (Approval No. 2024‐AE270).
   The study also received approval from the Research Ethics Committee of
   the same institution (Approval No. 2024‐R474) for the research
   involving human samples. Written informed consent was obtained from all
   patients prior to sample collection.

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [204]ADVS-12-e00148-s004.docx^ (7.4MB, docx)

   Supplemental Video 1
   [205]Download video file^ (31.4MB, mp4)

   Supplemental Video 2
   [206]Download video file^ (4.3MB, mp4)

   Supplemental Video 3
   [207]Download video file^ (2.4MB, mp4)

   Supplemental Video 4
   [208]Download video file^ (12.1MB, mp4)

   Supplemental Table 1
   [209]ADVS-12-e00148-s009.xls^ (76.9KB, xls)

   Supplemental Table 2
   [210]ADVS-12-e00148-s007.xls^ (2.9KB, xls)

   Supplemental Table 3
   [211]ADVS-12-e00148-s003.xlsx^ (16.6KB, xlsx)

   Supplemental Table 4
   [212]ADVS-12-e00148-s005.xls^ (2.7MB, xls)

Acknowledgements