Abstract

   Targeted delivery of glutamine metabolism inhibitors holds promise for
   cholangiocarcinoma therapy, yet effective delivery vehicles remain a
   challenge. This study reports the development of a biomimetic
   nanosystem, termed R‐CM@MSN@BC, integrating mesoporous organosilicon
   nanoparticles with reactive oxygen species‐responsive diselenide bonds
   for controlled release of the glutamine metabolism inhibitor
   bis‐2‐(5‐phenylacetamido‐1,3,4‐thiadiazol‐2‐yl) ethyl sulfide (BPTES)
   and the photosensitizer Ce6. Erythrocyte membrane coating, engineered
   with Arg‐Gly‐Asp (RGD) peptides, not only enhanced biocompatibility but
   also improved tumor targeting and tissue penetration. Upon laser
   irradiation, R‐CM@MSN@BC executed both photodynamic and
   glutamine‐metabolic therapies, inducing necroptosis in tumor cells and
   triggering significant immunogenic cell death. Time‐of‐flight mass
   cytometry analysis revealed that R‐CM@MSN@BC can remodel the
   immunosuppressive tumor microenvironment by polarizing M1‐type
   macrophages, reducing infiltration of M2‐type and CX3CR1^+ macrophages,
   and decreasing T cell exhaustion, thereby increasing the effectiveness
   of anti‐programmed cell death ligand 1 immunotherapy. This strategy
   proposed in this study presents a viable and promising approach for the
   treatment of cholangiocarcinoma.

   Keywords: cancer immunotherapy, cholangiocarcinoma, diselenide‐bond
   bridged mesoporous organosilica nanoparticles, glutamine‐metabolic
   therapy, necroptosis, tumor‐associated macrophages
     __________________________________________________________________

   The enhancement of solid tumor immunogenicity and resolution of the
   apoptosis resistance dilemma in tumors can be effectively achieved
   through the induction of necroptosis in tumor cells. R‐CM@MSN@BC
   demonstrates the capability of actively delivering photosensitizers and
   glutamine metabolism inhibitors, thereby inducing necroptosis in tumor
   cells and concurrently remodeling the immunosuppressive
   microenvironment.

   graphic file with name ADVS-11-2309203-g005.jpg

1. Introduction

   Cholangiocarcinoma (CCA), arising from the epithelial cells of the bile
   duct, presents a formidable challenge with significant heterogeneity.^[
   [50]^1 ^] The prognosis for advanced CCA is bleak, with an average
   overall survival ranging from 5–9 months due to the dearth of effective
   treatment modalities.^[ [51]^2 , [52]^3 ^] Immune checkpoint inhibitors
   (ICIs) that specifically target programmed cell death protein 1
   (PD‐1)/programmed death‐ligand 1 (PD‐L1) have exhibited substantial
   clinical efficacy across various cancers.^[ [53]^4 , [54]^5 , [55]^6 ^]
   However, this therapeutic effect is observed in less than 20% of
   patients with advanced CCA.^[ [56]^7 , [57]^8 ^] The immunosuppressive
   tumor microenvironment (TME) plays a pivotal role in hampering the
   success of ICIs in patients with CCA by impeding cytotoxic T lymphocyte
   infiltration and activation.^[ [58]^9 , [59]^10 ^] Photodynamic therapy
   (PDT) has emerged as a promising intervention for CCA due to its
   tunable light‐induced damage, non‐invasiveness, and minimal side
   effects.^[ [60]^11 , [61]^12 ^] During PDT, photosensitizers generate
   cytotoxic reactive oxygen species (ROS) upon light irradiation,
   potentially inducing immunogenic cell death (ICD) and enhancing tumor
   cell immunogenicity.^[ [62]^13 ^] However, the rapid elimination of
   PDT‐generated ROS by elevated glutathione (GSH) levels within tumor
   cells diminishes the effectiveness of ICD, necessitating additional
   therapeutic strategies to potentiate PDT and reshape the
   immunosuppressive TME.

   Tumor metabolism, particularly the reprogramming of glutamine
   utilization, significantly influences the immunosuppressive TME during
   tumor development.^[ [63]^14 , [64]^15 , [65]^16 ^] Tumor cells, in
   addition to heightened glucose demands, exhibit a dependency on
   glutamine, termed “glutamine addiction”.^[ [66]^17 , [67]^18 , [68]^19
   ^] This metabolic trait poses challenges to immune cell function and
   undermines the success of cancer immunotherapy.^[ [69]^20 , [70]^21 ^]
   Inhibition of glutamine metabolism, specifically targeting key
   metabolic enzyme glutaminase 1 (GLS1), has been shown to remodel the
   TME by inducing M1‐type macrophage polarization and increasing
   cytotoxic T‐lymphocyte (CTL) infiltration in CCA cells.^[ [71]^18 ^]
   Glutamine metabolism acts as a metabolic checkpoint that inhibits
   immune cell‐mediated antitumor responses.^[ [72]^11 , [73]^22 ^]
   Furthermore, glutamine metabolism also significantly influences the
   regulation of intracellular redox homeostasis. The inhibition of
   glutamine metabolism in tumor tissues can lead to a reduction in the
   production of vital reducing agents, such as GSH and NADPH.^[ [74]^23 ,
   [75]^24 ^] Thus, glutamine metabolism blockade could enhance treatment
   approaches that rely on oxidative damage, such as PDT, through a
   synergistic effect. The use of GLS1 inhibitors to block glutamine
   metabolism can synergistically improve PDT efficacy and alleviate the
   immunosuppressive TME. However, existing inhibitors targeting glutamine
   metabolism exhibit inadequate tumor tissue penetration, resulting in
   limited therapeutic effectiveness and substantial gastrointestinal
   adverse effects.^[ [76]^25 ^] Therefore, the development of a robust
   delivery system rooted in glutamine metabolism, with the ability to
   selectively target the CCA TME, holds significant promise for advancing
   therapeutic outcomes.

   In this study, we engineered a biomimetic multifunctional
   controlled‐release drug nanosystem, designated as R‐CM@MSN@BC, to
   synergistically integrate PDT, glutamine metabolism therapy, and
   biological imaging. Illustrated in Scheme [77] 1 , the nanosystem
   employed ROS‐responsive diselenide‐bond bridged hollow mesoporous
   organosilica nanoparticles (MSNs) as the framework, loaded with the
   photosensitizer Ce6 and the GLS1 inhibitor BPTES. Upon light
   irradiation, PDT‐induced ROS cleaved the diselenide bond, causing MSN
   skeleton disintegration and complete release of Ce6 and BPTES. The
   outer surface of the nanosystem was coated with erythrocyte membrane
   modified with Arg‐Gly‐Asp (RGD) peptide, which enhanced
   biocompatibility, targeting, and penetration capabilities toward 3D
   tumor spheroids via specific interaction between RGD peptide and
   overexpressed integrin α[ν]β[3] on CCA cell surfaces. Experimental
   findings confirmed significant ROS production by R‐CM@MSN@BC during
   therapeutic intervention, enhancing CCA immunogenicity through
   necroptosis induction in tumor cells. Importantly, time‐of‐flight mass
   cytometry (CyTOF) data indicated R‐CM@MSN@BC‐mediated glutamine
   metabolism inhibition, which reduced immunosuppressive
   macrophages—particularly M2‐type and CX3CR1^+ macrophages—and augmented
   M1‐type macrophage infiltration. Furthermore, the treatment decreased
   exhausted T cell infiltration, reversing the immunosuppressive CCA TME.
   When combined with immune checkpoint blockade therapy, R‐CM@MSN@BC
   significantly inhibited both primary and abscopal tumor growth. Thus,
   our results highlight the safety and efficacy of the “Trinity”
   biomimetic nanosystem as a comprehensive strategy for CCA treatment.

Scheme 1.

   Scheme 1
   [78]Open in a new tab

   Schematic Illustration of the Synthesis Process and Therapeutic
   Mechanism of R‐CM@MSN@BC Nanosystem. R‐CM@MSN@BC is composed of a
   diselenide‐bond‐bridged hollow mesoporous organosilica core that
   exhibits responsiveness to reactive oxygen species (ROS). Moreover, the
   biomimetic nanosystem is enveloped by an erythrocyte membrane that has
   been subjected to modifications involving RGD peptides. This
   modification facilitates the selective targeting of α[v]β[3] integrin,
   thereby facilitating the accurate administration and controlled release
   of nanomedicines. The diselenide‐bond bridged hollow mesoporous
   organosilica core is loaded with the GLS1 inhibitor BPTES and the
   photosensitizer Ce6, enabling the combination of glutamine metabolism
   therapy and photodynamic therapy (PDT). During the treatment process,
   R‐CM@MSN@BC can induce cell death in tumor cells through the
   necroptosis pathway, leading to the release of damage‐associated
   molecular patterns (DAMPs) that enhance the immunogenicity of
   cholangiocarcinoma. The combination of glutamine metabolism therapy and
   PDT holds promise in reshaping the immunosuppressive tumor
   microenvironment through reducing immunosuppressive macrophages,
   specifically M2‐type and CX3CR1^+ macrophages, while concurrently
   promoting the infiltration of M1‐type macrophages. Consequently, this
   approach has the potential to enhance the efficacy of
   cholangiocarcinoma immunotherapy.

2. Results and Discussion

2.1. GLS1 as a Potential Therapeutic Target in CCA

   Given the essential role of the glutamine metabolic pathway in tumor
   cell proliferation and the significance of GLS1 within this pathway, we
   conducted a comprehensive analysis across 33 tumor types from the The
   Cancer Genome Atlas (TCGA) database to evaluate GLS1 mRNA expression.
   Our analysis revealed a substantial increase in GLS1 expression across
   various tumors compared to corresponding normal tissues. Particularly
   noteworthy was the significantly elevated GLS1 expression in CCA
   tissues compared to normal tissues (Figure [79] 1A; Figure [80]S1A,
   Supporting Information). Confirmation of GLS1 overexpression in CCA
   tissues was achieved through real‐time quantitative reverse
   transcription polymerase chain reaction, western blot, and
   immunofluorescence staining in a set of 10 paired samples
   (Figure [81]1B–D; Figure [82]S1B, Supporting Information). Prognostic
   analysis of 36 CCA cases from the TCGA database indicated that elevated
   GLS1 expression in tumor tissues correlated with a poorer clinical
   prognosis (Figure [83]1E). Subsequent assessment of GLS1 protein levels
   in CCA cell lines (QBC‐939, RBE, HuCCT1) and human normal tissue cell
   lines (HIBEpic, MIHA) using western blot analysis demonstrated a
   significant upregulation of GLS1 protein levels in tumor cell lines
   compared to normal tissue cell lines (Figure [84]1F; Figure [85]S1C,
   Supporting Information). These findings provide strong evidence for the
   high expression of GLS1 in CCA and its association with an unfavorable
   patient prognosis. In order to ascertain the efficacy of GLS1
   inhibition in suppressing CCA cell growth and proliferation, we
   utilized CRISPR/Cas9 technology to knock down the GLS1 gene in QBC‐939
   and RBE cell lines, resulting in the generation of GLS1‐KO CCA cells
   (Figure [86]S1D, E, Supporting Information). Subsequently, the
   alterations in cell activity and proliferative potential of CCA were
   investigated following the inhibition of GLS1. The experimental
   findings indicating that either GLS1 gene knockdown or treatment with
   the GLS1 inhibitor BPTES significantly impeded the cell activity and
   proliferative potential of CCA cells. However, this effect was
   unaffected after the use of BPTES on GLS1‐KO CCA cells
   (Figure [87]1G–J; Figure [88]S1F–H, Supporting Information). Thus, the
   above experimental data suggest that BPTES‐mediated GLS1 suppression
   can inhibit the growth and proliferation of CCA cells. The suppression
   of glutamine metabolism leads to decreased production of reducing
   agents within the TME, thereby potentially improving the effectiveness
   of PDT. Consequently, the combination of glutamine metabolism therapy
   targeting GLS1 with PDT holds promise as a viable treatment approach
   for CCA.

Figure 1.

   Figure 1
   [89]Open in a new tab

   GLS1 was abundantly expressed in cholangiocarcinoma patients' tumor
   tissues and associated with a poor prognosis. A) Pan‐cancer analysis of
   GLS1 expression across cancers from TCGA and GTEx datasets.
   Abbreviations: BLCA: Bladder urothelial carcinoma; BRCA: Breast
   invasive carcinoma; CHOL: Cholangiocarcinoma; COAD: Colon
   adenocarcinoma; ESCA: Esophageal carcinoma; HNSC: Head and neck
   squamous cell carcinoma; KICH: Kidney chromophobe; KIRC: Kidney renal
   clear cell carcinoma; KIRP: Kidney renal papillary cell carcinoma;
   LIHC: Liver hepatocellular carcinoma; LUAD: Lung adenocarcinoma; LUSC:
   Lung squamous cell carcinoma; PRAD: Prostate adenocarcinoma; READ:
   Rectum adenocarcinoma; STAD: Stomach adenocarcinoma; THCA: Thyroid
   carcinoma; UCEC: Uterine corpus endometrial carcinoma. B) Relative mRNA
   expression levels of GLS1 in 10 pairs of CCA and matched non‐tumor
   tissues from our institution; n = 10 per group. C) GLS1 expression in
   CCA tissues were determined using the immunohistochemistry (IHC)
   staining. Scale bar, 100 µm. D) Western blot analysis of GLS1
   expression in CCA and matched non‐tumor tissues. N: normal liver
   tissue, T: tumor tissue; n = 3 per group. E) High expression of GLS1 in
   CCA patients predicts a poor prognosis based on TCGA dataset. F)
   Western blot and its quantification analysis of GLS1 expression in
   total cell extracts from human cholangiocarcinoma cells and human
   normal cells. G) The viability of RBE cells or GLS1‐KO RBE cells after
   different treatment. H) Flow cytometry analysis of dead cells in RBE
   cells or GLS1‐KO RBE cells after different treatments. I, J)
   Representative images and its corresponding quantitative analysis of
   the RBE cells or GLS1‐KO RBE cells after different treatment, assessed
   with colony‐forming assay. n = 3 per group n.s., no significance; *P <
   0.05; **P < 0.01; ***P < 0.001.

2.2. Fabrication, Characterization, and ROS‐Responsive Degradation of
R‐CM@MSN@BC

   The drug release kinetics of mesoporous silica are influenced by the
   dimensions of its pores, resulting in the retention of certain drugs
   within the material and hindering rapid and complete drug release.
   Additionally, the slow degradation of the mesoporous silica framework
   raises concerns about its safety in vivo, thereby restricting its
   clinical utility. Consequently, there is a pressing necessity to
   enhance traditional mesoporous silica nanomaterials with regulated drug
   delivery capabilities. A hollow mesoporous silica bridged by diselenide
   bond (MSN) was synthesized using ethyltrimethylammonium tosylate and
   triethanolamine as catalysts. The diselenide containing organosilica
   precursor bis(3‐(triethoxysilyl)propyl) diselenide (BTESe) and
   inorganic silica precursor tetraethyl orthosilicate were employed in
   the synthesis process (Figure [90]S2A,B, Supporting Information). Prior
   to synthesis, BTESe was prepared and characterized through mass
   spectrometry and ^1H nuclear magnetic resonance spectra (Figure [91]
   2A,B). Transmission electron microscopy (TEM) images revealed the
   spherical and hollow structure of MSN (Figure [92]2C). The
   Brunauer–Emmet–Teller surface area, total pore volume, and average pore
   size of the MSNs were determined to be 461.73 m^2 g^−1, 1.081
   cm^3 g^−1, and 9.36 nm, respectively (Figure [93]S2C,D, Supporting
   Information). The photosensitizer Ce6 and GLS1 inhibitor BPTES were
   loaded into the pores of the thin shell and hollow structure using
   negative pressure adsorption, resulting in successful loading into the
   MSN. Ce6 loading was confirmed by UV‐Vis‐NIR and fluorescence spectra
   of various nanoplatforms. A slight redshift observed at 654 nm to
   673 nm may be attributed to hydrophobic interactions and π–π stacking
   between the aggregated Ce6 in the core (Figure [94]2D). Loading content
   was determined by analyzing the supernatant Ce6 and BPTES using
   high‐performance liquid chromatography. The maximum loading content of
   Ce6 was approximately 7.32%, and BPTES reached 6.96%. Subsequently, MSN
   was briefly mixed with manganese permanganate (KMnO[4]) to enable the
   in situ formation of MnO[2]. As indicated in Figure [95]2E, the
   chemical state of Mn was investigated. Characteristic peaks at 653.48
   and 641.68 eV corresponded to the Mn (IV) 2p[1/2] and Mn (IV) 2p[3/2],
   indicating the predominant existence of Mn^4+ in the MnO[2] of MSN@BC.
   To investigate the ROS‐responsive properties of MSN@BC, MSN@BC was
   immersed in phosphate‐buffered saline (PBS) and subjected to different
   treatments. TEM was used to observe morphological changes and
   structural evolution, revealing clear structural breakdown and collapse
   of the majority of MSN@BC particles in a PBS solution containing 100 µM
   H[2]O[2] after a 24‐h incubation period. Furthermore, the backbone of
   MSN@BC experienced significant degradation when exposed to specific
   light wavelengths (Figure [96]2F). Notably, after 3 minutes of laser
   irradiation, MSN@BC demonstrated a drug release rate of 63.9% after 6 h
   and 73.6% after 12 h (Figure [97]2G). To confirm the production of
   singlet oxygen (^1O[2]) by MSN@BC, the generation of ^1O[2] in MSN@BC
   following irradiation with a 655 nm laser was assessed by monitoring
   the absorbance of 1,3‐diphenylisobenzofuran at 416 nm. As shown in
   Figure [98]2H, the treated MSN@BC groups exhibited significantly
   reduced absorption, indicative of ^1O[2] generation. These results
   reveal that MSN@BC exhibits good ROS‐responsive degradation behavior.

Figure 2.

   Figure 2
   [99]Open in a new tab

   Fabrication, Characterization and ROS‐responsive degradation of
   R‐CM@MSN@BC. A, B) Mass spectrometry and ^1H NMR spectra of BTESe. C)
   Representative TEM images of MSN. Scale bar, 200 nm. D) UV–vis‐NIR
   spectra of Ce6, MSN, MSN@BC, R‐CM@MSN@BC. E) XPS spectra for Mn 2p. F)
   Representative TEM images of MSN@BC after 660 nm laser irradiation
   (0.3 W cm^−2, 3 min) or incubation in H[2]O[2] (100 µM) at 12 h or
   24 h. Scale bar, 100 nm. G) The cumulative BPTES‐release profile of
   MSN@BC in H[2]O[2] (100 µM) or 660 nm laser irradiation (0.3 W cm^−2,
   3 min). H) The absorption spectra of DPBF treated with MSN@BC upon
   different times of laser irradiation (660 nm, 0.2 W cm^−2). I)
   Representative TEM images of R‐CM@MSN@BC. Scale bar, 50 nm. J)
   Representative element‐mapping images of R‐CM@MSN@BC. Scale bar, 50 nm.
   K) SDS‐PAGE analysis of empty erythrocyte membrane, MSN@BC, CM@MSN@BC
   and R‐CM@MSN@BC. L) Representative spinning disk confocal super
   resolution microscope images of R‐CM@MSN@BC. Scale bar, 100 nm.

   To enhance the biocompatibility and tumor‐targeting capabilities of
   MSN@BC, we employed a method in which the erythrocyte membrane,
   modified with RGD peptide, was enveloped onto the surface of MSN@BC,
   resulting in R‐CM@MSN@BC synthesis. R‐CM@MSN@BC exhibited a core‐shell
   structure, with the MSN core encapsulated within a thin and smooth
   membrane shell. The membrane thickness of R‐CM@MSN@BC was ≈12 nm,
   similar to the thickness of the erythrocyte membrane (Figure [100]2I;
   Figure [101]S2E, Supporting Information). Energy‐dispersive X‐ray
   spectroscopy mapping verified the presence and distribution of Si, Mn,
   Se, and P within a single R‐CM@MSN@BC entity, indicating uniform
   wrapping of the erythrocyte membrane around R‐CM@MSN@BC
   (Figure [102]2J). Subsequently, the nanosystems were subjected to
   sodium dodecyl sulfate‐polyacrylamide gel electrophoresis experiments.
   The results revealed that both R‐CM@MSN@BC and CM@MSN@BC exhibited an
   identical protein composition to that of erythrocyte membranes,
   confirming the successful encapsulation of erythrocyte membranes
   surrounding MSN (Figure [103]2K). In order to identify the presence of
   the RGD peptide modification on R‐CM@MSN@BC, the RGD peptide was
   labeled with a FITC probe and observed using a spinning disk confocal
   super‐resolution microscope. We observed that the FITC fluorescence
   enveloped the red fluorescence emitted by Ce6 in the nanosystem,
   further confirming the successful encapsulation of MSN@BC nanoparticles
   loaded with Ce6 by the erythrocyte membrane modified with FITC‐RGD
   peptide (Figure [104]2L). Fluorescence spectroscopy experiments
   provided additional evidence, as R‐CM@MSN@BC exhibited distinct
   fluorescence peaks at 525 nm and 660 nm, corresponding to FITC and Ce6
   fluorescence peaks, respectively (Figure [105]S2F, Supporting
   Information). The dynamic light scattering assay revealed that the
   diameter of R‐CM@MSN@BC measured 141.62 nm, indicating an increase of
   ≈13 nm compared to the MSN@BC particle size. This increase can be
   attributed to the successfully encapsulated erythrocyte membrane.
   Furthermore, the surface potential of the coated erythrocyte membrane
   with MSN@BC changed from −39.35 to −26.76 mV, further confirming the
   successful coating of erythrocyte membrane (Figure [106]S2G,H,
   Supporting Information). The in vivo results showed that 24 hours after
   injection of the nanosystems and 5‐minute light stimulation, a decrease
   in GLS1 protein expression was detected in mouse tumor tissues,
   suggesting that R‐CM@MSN@BC also functions as a good ROS‐responsive
   degrader in vivo (Figure [107]S2I, Supporting Information).

2.3. Biosafety and Imaging Performance of R‐CM@MSN@BC

   Effective biosafety and advanced visualization capabilities are
   essential requirements for the successful clinical implementation of
   diagnostic‐integrated nanosystems. The cytotoxicity of various
   nanosystems was evaluated in vitro using HIBEpic cells. Results from
   Figure [108] 3A and Figure [109]S3A (Supporting Information) suggest
   that when coated with an erythrocyte membrane, CM@MSN@BC and
   R‐CM@MSN@BC displayed minimal toxicity toward HIBEpic cells. R‐CM@MSN@B
   was unable to generate a sufficient amount of reactive oxygen species
   to cleave the diselenide‐bond, as the photosensitizer Ce6 was not
   present within it (Figure [110]S3B, Supporting Information).
   Consequently, the release of drugs contained within the nanoshells was
   impeded, thus limiting their cytotoxic effects. Additionally, the
   addition of R‐CM@MSN@BC to peripheral blood at a concentration of
   100 µg mL^−1 did not lead to significant hemolysis, indicating
   favorable biocompatibility (Figure [111]3B). To assess the targeting
   properties of the RGD peptide on the surface of the nanosystems for CCA
   cells, we conducted co‐incubation experiments with RBE cells and
   assessed uptake efficiency using confocal laser scanning microscopy and
   flow cytometry. R‐CM@MSN@BC exhibited superior tumor targeting
   capabilities and uptake efficiency than those exhibited by MSN@BC.
   Furthermore, this effect was observed to be attenuated in the presence
   of the integrin α[ν]β[3] inhibitor cilengitide (Figure [112]S3C–F,
   Supporting Information). The enhanced tumor targeting and uptake
   efficiency of R‐CM@MSN@BC facilitate selective eradication of tumor
   cells exposed to laser irradiation, enhancing light‐controlled
   inhibition efficacy and significantly mitigating adverse effects on
   non‐targeted cells and tissues (Figure [113]S3G, Supporting
   Information). In the context of solid tumor therapy, drug effectiveness
   relies on permeability. We accordingly assessed the tumor penetration
   characteristics of the nanosystems using a 3D tumor spheroid model
   comprising RBE cells. Different nanosystems were co‐cultured with tumor
   spheroids for 24 h, followed by imaging using a confocal high‐intensity
   imaging system at a consistent depth interval of 60 µm
   (Figure [114]3C). The spheroids incubated with R‐CM@MSN@BC exhibited
   significantly stronger red fluorescence of Ce6 in each section than
   that exhibited by the groups cultured with MSN@BC and CM@MSN@BC
   (Figure [115]3D). This indicates that the R‐CM@MSN@BC nanosystem
   exhibits heightened penetration with the RGD modification, which
   effectively enhances the targeted binding of the nanosystem to RBE
   cells that exhibit α[v]β[3] integrin overexpression on their cellular
   surfaces.

Figure 3.

   Figure 3
   [116]Open in a new tab

   Biosafety and Imaging Performance of R‐CM@MSN@BC. A) The viability of
   HIBEpic cells after co‐incubation with R‐CM@MSN@BC (5, 10, 20, 50,
   100 µg mL^−1) for 24 h; n = 3 per group. B) Hemolysis ratio and
   representative images of erythrocyte in the presence of R‐CM@MSN@BC
   with various concentrations at 6 h, 12 h and 24 h, respectively; n = 3
   per group. Scale bar, 20 µm. C) Schematic of fluorescence in various
   transverse sections with a depth interval of 60 µm of RBE
   cholangiocarcinoma cell spheroid. D) Bright‐field (BF) and fluorescence
   images of different transverse sections with an increasing depth of
   60 µm of RBE cell spheroids cultured under various conditions. Scale
   bar, 100 µm. E) Fluorescent images of the RBE tumor‐bearing Balb/c nude
   mice at different time points after injection of R‐CM@MSN@BC
   (5 mg k^−1g). F) T1‐weighted MRI of the RBE tumor‐bearing Balb/c nude
   mice at different time points after injection of R‐CM@MSN@BC
   (5 mg k^−1g).

   To scrutinize the dual imaging capability and biosafety of R‐CM@MSN@BC
   in vivo, we intravenously administered the nanosystems to mice bearing
   RBE tumors, followed by scanning using an in vivo fluorescence imaging
   system. The results highlight that the integration of RGD peptides into
   the erythrocyte membrane coating of R‐CM@MSN@BC amplifies its tumor
   targeting ability and extends its circulation time in the body.
   Consequently, a significant quantity of fluorescent signals persisted
   in the tumor tissues of mice 48 hours post‐injection of R‐CM@MSN@BC
   (Figure [117]3E, Figure [118]S3J,K, Supporting Information). The MnO[2]
   on the surface of MSN can undergo reactions with hydrogen peroxide and
   hydrogen ions within the TME, resulting in the production of manganese
   ion (Mn^2+). Mn^2+ can reduce the relaxation time of tumor tissue
   during the T1 phase of magnetic resonance imaging (MRI), thereby
   boosting the sensitivity of MRI for both tumor diagnosis and treatment
   evaluation. We investigated the potential of R‐CM@MSN@BC to function as
   an MRI enhancer. As depicted in Figure [119]3F and Figure [120]S2L
   (Supporting Information), the enhanced T1 phase post‐administration of
   R‐CM@MSN@BC displayed improved tumor visibility compared to
   pre‐enhancement imaging. Furthermore, the assessment of vital organ
   functions in various groups of mice after 3 days of treatment revealed
   no notable abnormalities in the groups treated with nanoparticles, as
   demonstrated by both biochemical indicator tests (Figure [121]S3H,
   Supporting Information) and histological examination (Figure [122]S3I,
   Supporting Information). These results affirm that the application of
   R‐CM@MSN@BC, involving the modification of erythrocyte membranes with
   RGD peptide, brings about improvements in blood circulation, tumor
   targeting, permeation, and tumor imaging capabilities.

2.4. Potent ROS Generation and Antitumor Efficiency of R‐CM@MSN@BC

   Building upon these promising experimental findings, we conducted
   additional assessments of ROS production and the antitumor efficiency
   of R‐CM@MSN@BC. As the MnO[2] in the nanosystem reacts with H[2]O[2] as
   well as H^+ in the tumor tissue to generate oxygen. Therefore, we
   tested the effect of different nanosystems on the amelioration of the
   hypoxic microenvironment in tumor tissues. We conducted western blot
   analysis to assess hypoxia‐inducible factor (HIF)−1α levels in treated
   RBE cells after different treatment. Notably, there was a significant
   reduction in HIF‐1α levels in cells treated with CM@MSN@BC and
   R‐CM@MSN@BC groups compared to the control group, providing additional
   evidence of hypoxia alleviation (Figure [123]S4A,B, Supporting
   Information). Subsequently, various nanosystems were administered via
   the tail vein to RBE tumor‐bearing Balb/C nude mice, followed by tumor
   tissue staining using a hypoxic probe (Hypoxyprobe‐1 plus kit) after
   24 h. The tumor tissues of mice injected with R‐CM@MSN@BC nanosystems
   exhibited a notable reduction in hypoxia in comparison to the control
   group, primarily due to the oxygen‐generating capabilities of
   R‐CM@MSN@BC (Figure [124] 4A). Since PDT is an oxygen‐consuming
   process, injection R‐CM@MSN@C significantly exacerbated the degree of
   hypoxia in the tumor tissues of the mice after laser irradiation.
   Interestingly, group R‐CM@MSN@BC + L significantly mitigated tumor
   tissue hypoxia compared to group R‐CM@MSN@C + L (Figure [125]4A;
   Figure [126]S4C, Supporting Information). We hypothesized that this was
   attributed to the synergistic effects of BPTES and Ce6, resulting in
   enhanced tumor cell death and reduced oxygen consumption in tumor
   tissues following light exposure. Consequently, we conducted
   investigations into the suppressive impact of R‐CM@MSN@BC on glutamine
   metabolism and its capacity to product ROS. After conducting metabolite
   analysis on mouse tumor tissue, the inclusion of BPTES in R‐CM@MSN@BC
   was observed to significantly inhibit tumor glutamine metabolism and
   the associated production of reductants (Figure [127]S4E, Supporting
   Information). Moreover, after laser irradiation, the R‐CM@MSN@BC group
   exhibited a significantly higher production of ROS than the R‐CM@MSN@C
   group. This result can be attributed partially to the inhibitory effect
   of the responsively released BPTES in R‐CM@MSN@BC on the production of
   reducing substances such as GSH, leading to a decrease in ROS
   consumption. On the other hand, the presence of MnO[2] in R‐CM@MSN@BC,
   which supplies an ample quantity of O[2] for PDT to generate ROS
   (Figure [128]4C, Figure [129]S4D, Supporting Information). The results
   of live/dead cell staining experiments demonstrated that R‐CM@MSN@BC
   exhibited enhanced cytotoxicity toward tumor cells within the same
   treatment duration (Figure [130]4B). The in vitro therapeutic efficacy
   of R‐CM@MSN@BC was evaluated against RBE cells. Due to the ability to
   efficiently generate ROS, R‐CM@MSN@BC demonstrates high efficacy in
   inducing cell death in cholangiocarcinoma cells and the therapeutic
   efficiency of R‐CM@MSN@BC was irradiance and light irradiation
   time‐dependent (Figure [131]S4F–H, Supporting Information). Treatment
   of R‐CM@MSN@BC led to significant impairments in cloning formation and
   migration abilities following laser treatment, demonstrating the most
   pronounced cell‐killing effect (Figure [132]4D,E).

Figure 4.

   Figure 4
   [133]Open in a new tab

   Robust ROS Production and Antitumor Effect of R‐CM@MSN@BC. A)
   Representative hypoxia of RBE cells of tumor slices after different
   treatments. Balb/c mice bearing RBE tumors were i.v. injection with
   indicated nanoparticles. Scale bar, 100 µm. B) Fluorescence microscope
   images of RBE cells treated with the different NPs. Cells were stained
   with Calcein‐AM and propidium iodide. Scale bar, 100 µm. C)
   Representative ROS fluorescence images of RBE cells s after different
   treatments. Scale bar, 100 µm. D, E) Representative images of the RBE
   cells after exposure to the indicated nanosystems, assessed with
   colony‐forming assay (D) and transwell assays (E). F) Schematic diagram
   of in vivo anti‐tumor experiments on RBE tumor‐bearing nude mice. G)
   Individual tumor growth kinetics of mice receiving the indicated
   treatments. H) Survival rates of RBE tumor‐bearing nude mice that
   experienced above different treatments; n = 5 per group. I) H&E and
   tunel staining images of tumor slices harvested from RBE tumor‐bearing
   nude mice after different treatments. Scale bar, 100 µm. *P < 0.05; **P
   < 0.01; ***P < 0.001.

   We further investigated the in vivo antitumor effects of R‐CM@MSN@BC.
   Subcutaneous RBE tumors were induced in immunocompromised mice, and the
   nanosystems were intravenously administered via the tail vein 7 days
   following tumor inoculation. Light was applied to the tumor site after
   24 hours (Figure [134]4F). By day 14, the average tumor mass of mice in
   groups R‐CM@MSN@B, R‐CM@MSN@C and R‐CM@MSN@BC was measured at 1.18 g,
   0.88 g, and 0.42 g, respectively. Utilizing the CDI formula,^[ [135]^26
   ^] the CDI of the R‐CM@MSN@BC group was calculated to be 0.404,
   suggesting a significant synergistic effect between the glutamine
   metabolism inhibitor BPTES and the photosensitizer Ce6 in the
   biomimetic nanosystem. At the same time, the R‐CM@MSN@BC group
   exhibited a substantial tumor suppression rate of 63.06%, whereas the
   R‐CM@MSN@C group exhibited a tumor suppression rate of only 26.32%
   (Figure [136]4G, Figure [137]S4I–K, Supporting Information). The
   survival rate of the remaining five tumor‐bearing mice was assessed,
   and after 30 days of continuous monitoring, the R‐CM@MSN@BC + L group
   exhibited a survival rate of 100% (Figure [138]4H). Tunel and
   hematoxylin and eosin (H&E) staining revealed a significant reduction
   in cell proliferation within the tumor tissue of the R‐CM@MSN@BC + L
   group (Figure [139]4I). These results were attributed to the
   synergistic effects of glutamine inhibition and photodynamic therapy,
   resulting in tumor cell death and achieving a “1+1>2” effect.

2.5. R‐CM@MSN@BC Induces Tumor Cell Death through the Necroptosis Pathway

   Nanomedicines typically impede tumor growth by inducing apoptosis in
   tumor cells.^[ [140]^27 ^] However, the development of resistance to
   apoptosis is a characteristic feature in the progression of malignant
   tumors.^[ [141]^28 ^] Necroptosis, a form of cell death known for
   eliciting more pronounced ICD effects than apoptosis, is initiated by
   substantial levels of ROS.^[ [142]^29 , [143]^30 ^] Building on
   previous experiments demonstrating that R‐CM@MSN@BC generates
   significant ROS quantities through the synergistic action of Ce6 and
   BPTES, we explored the potential of R‐CM@MSN@BC to induce necroptosis
   in CCA cells. To decipher the mechanism underlying R‐CM@MSN@BC‐induced
   tumor cell death, we conducted RNA sequencing (RNA‐seq) analysis on
   R‐CM@MSN@BC + laser‐treated RBE cells. Differential gene expression
   analysis was performed, identifying upregulated and downregulated genes
   with a 2‐fold change cutoff and a significance [adjusted P (P[adj])]
   value of <0.05 (Figure [144] 5A). Kyoto Encyclopedia of Genes and
   Genomes (KEGG) functional analysis revealed enrichment in the
   necroptosis pathway in RBE cells treated with R‐CM@MSN@BC gene
   products, suggesting potential induction of necroptosis following laser
   irradiation (Figure [145]5B). To validate RNA‐seq findings at the
   protein level, we initially confirmed the expression of
   receptor‐interacting protein kinase 3 (RIPK3, a crucial initiator and
   effector of necroptosis) in CCA cells. Results revealed a high level of
   RIPK3 expression (Figure [146]S5A, Supporting Information). Introducing
   various inhibitors targeting different death pathways to RBE cells
   treated with R‐CM@MSN@BC, our findings indicated that necrosulfonamide
   (NSA), a specific mixed lineage kinase domain‐like pseudokinase (MLKL)
   inhibitor, effectively suppressed cell death induced by R‐CM@MSN@BC
   (Figure [147]5D,E; Figure [148]S5D, [149]S6B,C, Supporting
   Information). Moreover, NSA prevented RIPK3 punctation and cell
   swelling, characteristic morphological changes associated with
   necroptosis caused by R‐CM@MSN@BC (Figure [150]5F,G; Figure [151]S5B,
   Supporting Information). In vivo experiments demonstrated a notable
   elevation in the phosphorylation of RIPK3 in tumor tissues of mice
   subjected to treatment with R‐CM@MSN@BC, as evidenced by fluorescent
   staining (Figure [152]S5I, Supporting Information). Moreover,
   necroptosis, a highly immunogenic form of cell death, resulted in the
   release of substantial quantities of damage‐associated molecular
   patterns. Membrane translocation of calreticulin and release of high
   mobility group box 1 were notably enhanced following treatment with
   R‐CM@MSN@BC in combination with laser (Figure [153]S5G, H, Supporting
   Information). These results suggest that R‐CM@MSN@BC can trigger
   subsequent ICD effects by inducing necroptosis in tumor cells.

Figure 5.

   Figure 5
   [154]Open in a new tab

   R‐CM@MSN@BC Mediates Tumor Cell Death Through the Necroptosis Pathway.
   A) The transcriptome of the RBE of Control and R‐CM@MSN@BC plus laser.
   The screen hits were selected with cut‐offs set at fold change > 2 and
   P‐value < 0.05. Red, upregulated. Blue, downregulated. B) Signaling
   pathway enrichment analysis of Kyoto Encyclopedia of Genes and Genomes
   (KEGG) in R‐CM@MSN@BC plus laser–treated RBE cells. C) The top 20 up‐
   and down‐regulated genes in RBE cells treates with R‐CM@MSN@BC plus
   laser. D) Effect of cell‐death‐suppressing compounds on R‐CM@MSN@BC
   plus laser. MLKL inhibitor NSA (10 µM), apoptosis inhibitor Z‐DEVD‐FMK
   (20 µM), pyroptosis inhibitor Ac‐FLTD‐CMK (5 µM); n = 3 per group. E)
   Western blot images of necroptosis‐related proteins after different
   treatments. The levels of indicated proteins were determined by
   immunoblotting. n = 3 independent biological repeats. F) The effects of
   R‐CM@MSN@BC plus laser on the formation of RIP3 punctation, the
   distribution of RIPK3 (green) and DAPI (blue) was detected using
   confocal microscopy. G) TEM analysis of RBE cells after different
   treatment. Scale bars, 1 µm. H) RBE cells were transfected (48 h) with
   control siRNA, or sequentially distinct DR5 siRNA, then treated with
   PBS and R‐CM@MSN@BC plus laser, respectively. The levels of indicated
   proteins were determined by immunoblotting. n = 3 independent
   biological repeats. I) A schematic model to illustrate the mechanism of
   necroptosis in tumor cells caused by R‐CM@MSN@BC plus laser (By
   Figdraw).

   To elucidate the downstream pathways through which ROS generated by
   R‐CM@MSN@BC induce necroptosis in tumor cells, we conducted a
   comparative analysis of RNA‐seq results. This revealed a significant
   increase in the expression levels of marker genes associated with
   endoplasmic reticulum stress in RBE cells following treatment with
   R‐CM@MSN@BC (Figure [155]5C). Additionally, increased protein levels of
   activating transcription factor 4, C/EBP homologous protein, and
   phosphorylation of eukaryotic translation initiation factor 2 subunit 1
   were consistently observed after treatment with R‐CM@MSN@BC in
   conjunction with laser therapy (Figure [156]5H; Figure [157]S6,
   Supporting Information). Small interfering RNA (siRNA) was employed to
   investigate the involvement of various death receptors in
   R‐CM@MSN@BC‐mediated necroptosis. Findings indicate that death receptor
   5 (DR5) deficiency, but not tumor necrosis factor receptor 1/2
   (Tnfr1/2) and Fas (also called CD95), inhibits the phosphorylation of
   receptor‐interacting protein kinase (RIPK1), RIPK3, and MLKL,
   consequently suppressing necroptosis (Figure [158]5H;
   Figure [159]S5E,F; Figure [160]S6, Supporting Information). These
   results suggest that DR5 acts as an upstream regulator of necroptosis
   in the context of R‐CM@MSN@BC combined with laser treatment
   (Figure [161]5I).

2.6. R‐CM@MSN@BC Modulates the Immunosuppressive Microenvironment of CCA

   To comprehensively examine the impact of R‐CM@MSN@BC on the TME of CCA,
   we established a subcutaneous transplanted CCA model in C57BL/6J mice
   with a normal immune response. The CCA tissue underwent treatment and
   was analyzed using CyTOF. Our metal‐labeled antibody panel was designed
   to identify primary immune cell populations infiltrating the tumor,
   including markers for cell lineage, function, and cytokine activity.
   CyTOF data revealed 11 distinct intratumoral immune cell populations,
   and the t‐distributed stochastic neighbor embedding (t‐SNE) algorithm
   was employed for data analysis (Figure [162] 6A,B). The immune
   landscape of tumors treated with R‐CM@MSN@BC plus laser (R‐CM@MSN@BC +
   L) significantly differed from both saline‐treated tumors and tumors
   treated with R‐CM@MSN@C plus laser (R‐CM@MSN@C + L), particularly in
   terms of the distribution of tumor‐associated macrophages (TAMs) and T
   cells (Figure [163]S7A, Supporting Information). The R‐CM@MSN@BC + L
   group showed a significant reduction in the M2‐type macrophage
   subpopulation (CD206^+ macrophage) and the inhibitory macrophage
   subpopulation (CX3CR1^+ macrophage) compared to the saline or
   R‐CM@MSN@C + L group, indicating a decrease in immunosuppressive
   macrophages within TAMs (Figure [164]6C,E). Additionally, R‐CM@MSN@BC +
   L treatment increased the M1‐type macrophage subpopulation (CD86^+
   macrophage), suggesting enhanced proinflammatory effects
   (Figure [165]6D). Tumor‐infiltrating T cell exhaustion is correlated
   with tumor progression and immunotherapy resistance. In contrast to
   control groups receiving saline and R‐CM@MSN@C + L, CyTOF analysis
   revealed a decrease in PD‐1 expression among intratumoral CD8^+ and
   CD4^+ T cells in mice treated with R‐CM@MSN@BC + L (Figure [166]6F,
   Figure [167]S7B,C, Supporting Information). These findings collectively
   indicate that R‐CM@MSN@BC + L effectively modifies the
   immunosuppressive TME of CCA by reducing immunosuppressive macrophages,
   specifically M2‐type macrophages and CX3CR1^+ macrophages, and
   augmenting M1‐type macrophage infiltration. Furthermore, the treatment
   leads to a decrease in exhausted T cell infiltration, attributed to the
   inhibitory impact of BPTES on glutamine metabolism within R‐CM@MSN@BC.

Figure 6.

   Figure 6
   [168]Open in a new tab

   R‐CM@MSN@BC Reshapes the Immunosuppressive Microenvironment of CCA. A)
   Heatmap depicting median signal intensity (MSI) of indicated markers
   derived from CyTOF. B) A t‐distributed stochastic neighbor embedding
   (tSNE) plot via nonlinear dimensionality reduction showing the immune
   cell clusters identified by CyTOF analysis. C to F) T‐SNE plots showing
   expression levels of (C) M2‐type macrophage, (D) M1‐type macrophage,
   (E) CX3CR1^+ macrophage and (F) PD1^+CD8^+ T cells with corresponding
   quantitative comparisons. n = 3 per group. n.s., no significance; *P <
   0.05; **P < 0.01; ***P < 0.001.

2.7. Enhanced Systemic Immune Response with R‐CM@MSN@BC in Combination with
Anti‐PD‐L1

   Immunotherapy is a commonly utilized clinical strategy for the
   treatment of tumors. At present, Anti‐PD‐L1 are the recommended
   immunotherapeutic agents for CCA; however, their therapeutic
   effectiveness is limited due to the immunosuppressive TME of CCA. Our
   previous experimental results demonstrated that R‐CM@MSN@BC enhances
   tumor immunogenicity by inducing necroptosis of tumor cells and
   reshapes the immunosuppressive TME by reprogramming glutamine
   metabolism. Thus, we evaluated the combined efficacy of R‐CM@MSN@BC and
   anti‐PD‐L1 immunotherapy (Figure [169] 7A). Due to the
   immunosuppressive TME in CCA, administering anti‐PD‐L1 solely through
   the tail vein did not significantly inhibit primary and distant tumor
   growth. However, the group treated with R‐CM@MSN@BC + L displayed
   efficacy in suppressing primary tumor growth. Interestingly, the
   combination therapy of R‐CM@MSN@BC + L and anti‐PD‐L1 exhibited a
   substantial inhibition of bilateral tumor growth. Furthermore, mice in
   the R‐CM@MSN@BC + L plus anti‐PD‐L1 combination treatment group showed
   a more favorable prognosis with regard to survival (Figure [170]7B–F,
   Figure [171]S8A–D, Supporting Information). Ki67 and H&E staining on
   primary and abscopal tumor tissues revealed a significant decrease in
   proliferating tumor cells in the R‐CM@MSN@BC + L plus anti‐PD‐L1 groups
   (Figure [172]S8F, Supporting Information). These findings suggest that
   the combination of R‐CM@MSN@BC + L and anti‐PD‐L1 exhibits a
   synergistic therapeutic effect on tumor growth inhibition, as well as a
   promising abscopal effect. To gain further insights into the mechanisms
   underlying the enhanced antitumor activity, flow cytometry was employed
   to examine T cell infiltration in both primary (Figure [173]7G,H) and
   abscopal tumors (Figure [174]7I,J). In both primary and abscopal
   tumors, R‐CM@MSN@BC administration combined with anti‐PD‐L1 resulted in
   a significant increase in the proportions of cytotoxic T‐lymphocytes
   (CD3^+CD8^+ T cells). Subsequently, enzyme‐linked immunosorbent assays
   were conducted to assess alterations in the levels of several
   proinflammatory cytokines, including IL‐12, IL‐2, interferon‐γ, and
   tumor necrosis factor‐α, in mouse sera following the different
   treatments. As indicated by our results, R‐CM@MSN@BC can stimulate
   immune cells within the TME to secrete higher levels of inflammatory
   cytokines subsequent to irradiation (Figure [175]S8G, Supporting
   Information). This study suggests that the combination therapy
   involving R‐CM@MSN@BC + L and anti‐PD‐L1 effectively stimulates a
   systemic antitumor immune response and enhances the efficacy of
   anti‐PD‐L1 in the treatment of CCA.

Figure 7.

   Figure 7
   [176]Open in a new tab

   R‐CM@MSN@BC in Combination with Anti‐PD‐L1 Effectively Enhances
   Systemic Immune Response. A) Schematic diagram of different treatments
   for inhibiting primary and abscopal tumor on C57BL/6 mice subcutaneous
   graft tumor model. B, C) Individual primary and abscopal tumor growth
   kinetics of mice receiving the indicated treatments. D, E) Primary and
   abscopal tumor weight and inhibition ratio of various treated mice. F)
   Survival rates of C57BL/6J tumor‐bearing mice that experienced above
   different treatments (n = 5). G, H) CD3^+CD8^+ T cells proportions in
   primary tumor evaluated by flow cytometry and its quantitative
   analysis. I, J) CD3^+CD4^+ T cells proportions in primary tumor
   evaluated by flow cytometry and its quantitative analysis. n = 5 per
   group. n.s., no significance; *P < 0.05; **P < 0.01; ***P < 0.001.

3. Conclusion

   In conclusion, we have successfully developed a biomimetic
   controlled‐release drug nanosystem that integrates diagnostic imaging,
   photodynamic therapy, and immunotherapy. This nanosystem demonstrates
   the capacity to generate significant ROS and inhibit GSH production
   within tumor cells upon light exposure, leading to prolonged tumor cell
   eradication. Additionally, R‐CM@MSN@BC reshapes the immunosuppressive
   TME by reprogramming glutamine metabolism. The release of ROS from
   R‐CM@MSN@BC induces ICD effects, particularly through necroptosis in
   tumor cells. Simultaneously, the GLS1 inhibitor BPTES released promotes
   M1‐type macrophage polarization and inhibits T‐cell depletion. Through
   these pathways, R‐CM@MSN@BC effectively modifies the immunosuppressive
   TME, enhancing the adaptive immune response and creating a robust
   foundation for combined anti‐PD‐L1 immune checkpoint blockade therapy.
   The combination of R‐CM@MSN@BC with anti‐PD‐L1 therapy demonstrates
   significant efficacy in impeding the proliferation of both localized
   and metastatic tumors in murine models. This introduces a novel
   approach to hinder tumor recurrence and expansion. In essence, our
   research introduces a promising nanoparticle composite, referred to as
   the “trinity” nanoparticle, which holds potential for overcoming
   existing challenges in CCA management.

4. Experimental Section

Materials

   Tetraethyl orthosilicate (TEOS),
   bis[3‐(triethoxysilyl)propyl]tetrasulfide (BTESe), triethanolamine
   (TEA), fluorescein isothiocyanate (FITC), selenium (Se) powder, sodium
   (Na) power, sodium borohydride (NaBH[4]) and ammonium nitrate
   (NH[4]NO[3]) were purchased from Xi'an Ruixi Biotechnology Company,
   China. The list of primers, sequences of the siRNAs, antibodies for
   western blot, enzyme‐linked immunosorbent assay (ELISA) kits, and
   antibodies for CyTOF was provided in Tables [177]S1 to [178]S5
   (Supporting Information), respectively.

Clinical Samples

   With the approval of CCA patients undergoing partial hepatectomy in the
   Center for Liver Transplantation and the Department of Hepatobiliary
   Surgery of Union Hospital, the paired tumor and surrounding non‐tumor
   tissues used in this study were obtained. All patients gave their
   informed consent. The research ethics committee of Union Hospital
   accepted all experimental protocols, and the same organization also
   awarded research ethics approval for this endeavor. The permission
   number of ethical approval statements was UHCT‐IEC‐SOP‐016‐03‐01.

Cell Lines and Animals

   The human bile duct epithelial cells (HIBEpic), human CCA cell lines
   (RBE, RBE, HuCCT1) and human normal liver cell lines (MIHA) were
   purchased from American Type Culture Collection (ATCC), and growed at
   37 °C with 5% CO[2] in a humidified atmosphere. The cells were grown in
   RPMI‐1640 (Gibco) supplemented with 10% FBS and 1%
   penicillin‐streptomycin at 37 °C in a humidified environment containing
   5% CO[2]. Male C57BL/6 and Balb/c nude 4–6‐week‐old mice were also
   provided by the Vital River Laboratory Animal Technology Co. Ltd. and
   Beijing HFK Bioscience Co. Ltd., both of Beijing, China (Beijing,
   China). All animal studies and feeding procedures were carried out in
   compliance with the regulations established by the Animal Care
   Committee at Tongji Medical College, HUST, Wuhan, China. The permission
   number of ethical approval statements was [2020] IACUC Number: 2910.
   This research was carried out in accordance with the Helsinki
   declaration.

Synthesis of BTESe

   Bis [3‐(triethoxysilyl) propyl] diselenide (BTESe) was synthesized
   using a previously established method.^[ [179]^31 ^] In this procedure,
   a round‐bottomed flask was filled with a solution containing 2.37 g
   (30 mmol) of selenium powder, which was then placed in an ice‐water
   bath under a nitrogen atmosphere for a duration of 10 min.
   Subsequently, a carefully measured solution of NaBH[4] (2.27 g,
   60 mmol) dissolved in 25 ml of water was added dropwise to the selenium
   suspension. The resulting reaction mixture was stirred and allowed to
   react at 0 °C until the solution became colorless and completely
   dissolved the selenium powder. Subsequently, an additional quantity of
   selenium powder (2.37 g, 30 mmol) was introduced into the
   aforementioned solution, and the flask was subjected to heating at 100
   °C for a duration of 10 min, resulting in the solution acquiring a
   reddish‐brown hue. Promptly thereafter, 12 grams of γ‐chloropropyl
   trimethoxysilane were swiftly added to the sodium diselenide stock
   solution, followed by overnight stirring. The resulting mixture was
   subjected to extraction using ethyl acetate, and subsequent rotary
   evaporation yielded an oily product. The crude product was subsequently
   purified through silica gel column chromatography, resulting in the
   formation of a dark yellow liquid product (3.32 g, ≈20% yield).

Synthesis of Diselenide‐Bond Bridged Mesoporous Organosilica Nanoparticles

   A mixture of 0.6 g of ethyltrimethylammonium tosylate (CTAT), 0.15 g of
   triethanolamine (TEA), and 40 mL of deionized water was stirred at 80
   °C for 30 min. Subsequently, a dropwise addition of a solution
   containing 4.0 g of triethoxysilane (TEOS) and 1.0 g of BTESe was made
   to the surfactant solution. The resultant mixture was stirred at 80 °C
   and 1000 rpm for an additional 4 h. The product was collected through
   centrifugation, subjected to three washes with ethanol, and
   subsequently reflux‐treated in an ethanol solution of ammonium
   aminonitrate (1% w/v) for 12 h. Ultimately, the produced
   diselenide‐bridged mesoporous organosilica nanoparticles (MSN) was
   collected, washed, and dried for subsequent experimental use.

Synthesis of MSN@BC

   Diselenide‐bond bridged mesoporous organosilica nanoparticles solution
   with the concentration of 0.6 mg mL^−1, and sonicated for well
   dispersion, subsequently, 50 mL of 1 mg mL^−1 potassium permanganate
   (KMnO[4], Sinopharm Chemical Reagent Co.) solution added slowly and
   continued to stir at room temperature for 30 min. Brown MSN were
   obtained after removing extra KMnO[4] and free MnO[2]. For the
   preparation of MSN@BC, 100 µL 10 mg mL^−1 Chlorin e6 (Ce6, Frontier
   Scientific, Inc.) dimethyl sulphoxide (DMSO, Sinopharm Chemical Reagent
   Co.) solution and 50 µL 2 mg mL^−1 BPTES DMSO solution was added into
   5 mg MSN. Stirred at room temperature for 14–16 h overnight, the
   precipitate was collected by centrifugation at 13 000 rpm and washed,
   re‐dispersed in deionized water to obtain MSN@BC. The above method was
   used to make Ce6 loaded MSN (MSN@C) and BPTES loaded MSN (MSN@B).

Preparation of Erythrocyte Membranes Modified with RGD Peptides

   In brief, 1 mL of fresh blood was withdrawn from the ophthalmic vein of
   male C57BL6/J wild‐type mice (20–22 g) through eyeball removal. This
   blood was then combined with 100 µL of EDTA anticoagulation solution.
   Erythrocytes were gathered via centrifugation at 720 g for 10 min. The
   obtained erythrocyte membranes were re‐suspended in 0.2 mm EDTA
   deionized water at a four‐fold volume to trigger membrane rupture.
   Following 60 min of hypotonic lysis, the samples were centrifuged at
   20,000 g for 20 min to obtain erythrocyte membranes. The freshly
   acquired erythrocyte membranes were stored at 4 °C and utilized within
   a 6 h timeframe. Subsequently, DSPE‐PEG2000‐RGD (FITC) was introduced,
   and the sample were mixed by ultrasonic shaking. Capitalizing on the
   lipid affinity of DSPE and erythrocyte membranes, DSPE can
   spontaneously incorporate into the erythrocyte membranes' phospholipid
   bilayer. This process efficiently and non‐destructively facilitates the
   modification of RGD on the membrane surface.

Preparation and Characterization of R‐CM@MSN@BC

   To fabricate biomimetic MSN@BC nanosystem, erythrocyte membrane
   modified with RGD peptide was mixed with the same weight of MSN@BC in
   water and sonicated for 10 min in an ice bath. The above method was
   used to make R‐CM@MSN@B and R‐CM@MSN@C. The morphology, structure, and
   surface chemistry of these nanosystem were characterized by
   transmission electron microscope (TEM, TF20, FEI company, USA), UV–vis
   spectrophotometer (TU‐1810, BJPERSEE, China), X‐Ray photoelectron
   spectroscopy (XPS, K‐Alpha, Thermo Scientific, USA), N[2] adsorption
   investigation (TriStar II 3020, Alpha technologies company, USA), and
   dynamic light scattering (DLS, NanoBrook 90plus PALS, Brookhaven, USA).
   Fluorescence spectrophotometer of the R‐CM@MSN@B, R‐CM@MSN@C, and
   R‐CM@MSN@BC were measured using a Fluoromax‐4 spectrofluorometer
   (Horiba Jobin Yvon Inc.) under 400 nm excitation. The concentration of
   Ce6 and BPTES in the supernatant was analysed by high performance
   liquid chromatography (HPLC, Agilent Technologies Inc., 1100S, USA)
   with a UV–vis detector at wavelength of 254 nm to calculate the
   encapsulation efficiency (EE) and drug loading capacity (DL) of Ce6 and
   BPTES.

   To analyze erythrocyte membranes within nanosystems, Both erythrocyte
   membranes and various nanosystems were disrupted using RIPA buffer
   supplemented with a combination of PMSF and phosphatase inhibitor. The
   entire protein content was gathered through centrifugation and
   quantified utilizing the BCA protein assay kit. Proteins derived from
   both RBC membranes and diverse nanosystems were separated using
   SDS‐polyacrylamide gels, subsequently stained, and then captured
   through photography with the application of Coomassie blue solution.
   Detection of FITC‐RGD peptides on the surface of nanosystems by a
   spinning disk confocal super resolution microscope (SpinSR, OLYMPUS,
   Japan).

Degradation Experiment and Drug Release Experiment In Vitro

   R‐CM@MSN@BC was dispersed in PBS (pH 6.5, 7.4) solution with the
   concentration of 0.2 mg mL^−1, respectively, among which pH 6.5 PBS
   solution, 200 µM H[2]O[2] was added to simulate tumor microenvironment,
   pH 6.5 PBS solution with 200 µM H[2]O[2] under laser irradiation was to
   imitate the therapeutic tumor microenvironment, and pH 7.4 PBS solution
   was to simulate in vivo. The resulting R‐CM@MSN@BC PBS solution was
   incubated separately at 37 °C in air bath shaker at 170 rpm. At the
   given time (0.5 h, 1 h, 3 h, 5 h, 7 h, 12 h, 1 d, 2 d, 3 d, and 5 d),
   2 mL of above R‐CM@MSN@BC PBS solution was centrifuged. The cumulative
   degradation content of free silicon was determined using ICP‐OES in one
   portion of the supernatant, while the content of BPTES and Ce6 was
   determined using HPLC in the other.

The Detection of Singlet Oxygen

   To measure singlet oxygen, the decline absorption of
   1,3‐diphenylisobenzofuran (DPBF) at 416 nm according to a reported
   literature. Firstly, nitrogen gas (N[2]) was blown into the PBS
   solution containing R‐CM@MSN@BC (50 µg mL^−1) for 30 min to clear away
   the O[2]. Subsequently, 250 µM H[2]O[2] was added to the R‐CM@MSN@BC
   solution above. After 30 min, 10 µg mL^−1 DPBF DMSO solution was add
   into and exposed under 655 nm laser irradiation for certain time. After
   centrifugation, the absorbance of supernatant was recorded by the
   UV–vis spectrophotometer.

Bioinformatic Analysis of RNA‐Seq Data for Human CHOL

   From the TCGA database, the RNA‐seq and clinical information of various
   cancer classifications were obtained and these data were analyzed with
   the GEPIA tool.^[ [180]^32 ^] The survival (2.44.1) packages were used
   in R (4.1.0) to compute and construct survival curves.

CRISPR/Cas9 Technique

   Single guide RNAs (sgRNAs) of Control (sgControl), GLS1 (sgGLS1‐1^#,
   sgGLS1‐2^# and sgGLS1‐3^#) were designed and purchased from Merck
   ([181]https://www.sigmaaldrich.cn/CN/zh). Then sgRNAs were cloned into
   lentiCRISPR v2 vector (#52 961, Addgene, USA). The sequences of sgRNAs
   were as flow, sgControl: CGCTTCCGCGGCCCGTTCAA; sgGLS1‐1^#:
   AAACGCGTCCGTCTCCCCG; sgGLS1‐2^#GACGCGTTTGGCAACAGCG;
   sgGLS1‐3^#AGCACGCATCCGCAGCCCG.

Cell Viability Assay

   RBE cells were seeded into a 96‐well plate at 8000 cells per well and
   cultured for 24 h. Then, the cells were treated with PBS or
   R‐CM@MSN@BC. The cells were washed three times with PBS to eliminate
   non‐internalized chemicals after a 12 h incubation in the dark, and
   200 µL of fresh culture media was added. Cells were irradiated with
   different intensities of 640 nm laser for the same time, or with the
   same intensity of 640 nm laser for different times, respectively. Cell
   viability was assessed using the CCK‐8 test (Beyotime, Shanghai, China)
   after an additional 24 h of incubation.

Immunohistochemistry

   GLS1, and Ki‐67 expression levels in tumor tissue were examined using
   immunohistochemical staining, as previously described,^[ [182]^33 ^]
   and images were taken using optical microscopy.

H&E Staining

   Images were taken using optical microscope after H&E staining to
   evaluate the histological alterations of tumors following various
   therapies, as previously described.^[ [183]^34 ^]

RNA Isolation and Quantitative Real‐Time PCR

   Trizol was employed for the extraction of total RNA from the cells, and
   the NanoDrop ND‐1000 spectrophotometer, manufactured by NanoDrop
   Technologies, Thermo, was utilized for its quantification. PrimeScript
   RT Master Mix was employed to conduct reverse transcription on 0.5 µg
   of total RNA, enabling the synthesis of the first strand cDNA. For
   quantitative real‐time PCR, a Takara SYBR Green PCR Kit was employed.
   Each sample was subjected to analysis in triplicate wells. The siRNA
   sequence and primer details can be found in Table [184]S1 and [185]S2
   (Supporting Information), respectively.

Western Blot Analysis

   The total protein content was obtained by employing a RIPA lysis
   buffer. The cytosolic protein and mitochondrial protein were extracted
   using a mitochondrial protein extraction kit and a nuclear protein
   extraction kit, respectively, following the guidelines provided by the
   manufacturer. The concentration of each sample was determined using a
   BCA protein assay kit. Equal concentrations of protein were loaded onto
   SDS‐PAGE and transferred onto PVDF membranes (Millipore, Bedford, USA),
   which were then blocked with 5% non‐fat milk for a duration of 1 h.
   Primary antibodies were incubated overnight at 4 °C. Following
   incubation with secondary antibodies conjugated to HRP, the
   immunoreactivity was detected utilizing an ECL substrate (BOSTER,
   Wuhan, China) and captured using an Image Lab imaging system. The
   comprehensive list of antibodies utilized can be found in table S3.

Flow cytometry Analysis

   According to the experimental method of previous,^[ [186]^35 ^] RBE
   cells after different treatments or tumor tissues were made into
   single‐cell suspensions. After co‐incubation with different fluorescent
   antibodies according to the manufacturer's instructions, the obtained
   cells were examined by flow cytometry (BD LSRFortessaX‐20 Special order
   product) and data were analyzed by a FlowJo software (Version 10). The
   list of antibodies were provided in Table [187]S3 (Supporting
   Information).

RNA‐Seq and Microarray Expression Data Analysis

   The RNA from RBE cells subjected to various treatments was sequenced
   using next‐generation sequencing technology by Haplox Pharmaceutical
   Technology (Shenzhen, China). The software tools Hisat2 (version
   2.2.1), htseq‐count, and edgeR were employed for mapping reads,
   calculating read counts, and analyzing differential expression,
   respectively.^[ [188]^36 , [189]^37 , [190]^38 ^] Subsequently, the
   data was normalized and analyzed using the limma package.^[ [191]^39 ^]
   Genes with a false discovery rate (FDR) less than 0.05 and an absolute
   fold change (|FC|) greater than 1.5 were identified as differentially
   expressed genes (DEGs).

Time‐Of‐Flight Mass Cytometry (CyTOF) Analysis

   Immune cells were obtained from CCA tumors utilizing the methodology
   outlined in Section Tumor Immune Microenvironment Assay using Flow
   Cytometry. These immune cells were subsequently labeled with antibodies
   containing metal tags. The antibody panel employed in this study
   comprised 42 antibodies conjugated with distinct metals (Table [192]S5,
   Supporting Information). The evaluation of conjugated target molecule
   expression was conducted by detecting the metal signals utilizing the
   CyTOF system (Helios, Fluidigm, San Francisco, CA, USA).

Transmission electron microscopy analysis

   Transmission electron microscopy analysis were performed as previously
   described to observed cell morphology after different treatments.^[
   [193]^40 , [194]^41 ^]

Nanosystems Assessment of 3D Tumor Cell Sphere Permeability

   RBE cholangiocarcinoma cells were seeded in ultra‐low adsorption
   96‐well plates at a density of 2 × 10^4 cells per well and incubated
   for 24 h. Once the cells reached a spherical morphology, equal
   concentrations of MSN@BC, CM@MSN@BC, and R‐CM@MSN@BC nanosystems were
   added individually and incubated for an additional 24 h. The cell
   nuclei were then stained using Hoechst 33342, and the resulting cell
   spheres were observed and photographed at various levels using a
   confocal high‐intensity imaging system (ImageXpress Micro Confocal,
   Molecular Devices, USA).

Internalization Efficacy Test

   The cells were randomly divided into four groups and co‐incubated with
   R‐CM@MSN@BC (10 µg mL^−1) for different time intervals (0 h, 2 h, 6 h,
   and 12 h) before being fixed with 4% paraformaldehyde for 35 min. The
   nuclei were located using Hoechst 33342 (Servicebio, Wuhan, China).
   Following three washes with PBS for five min each, confocal imaging was
   performed on the cell culture slides transferred to object slides.
   Subsequently, the ImageJ program was utilized to conduct relative
   semi‐quantitative fluorescence analysis in order to quantify the
   cellular uptake of micelles.

ROS Generation in Vitro

   Following the manufacturer's instructions, the fluorescent probe
   DCFH‐DA (Beyotime, Shanghai, China) was used to find ROS generation
   inside the cells following light exposure. On a 12‐well plate, 1 × 10^5
   RBE cells were planted. The culture medium was taken out and thoroughly
   washed with PBS twice after the cells had been attached. R‐CM@MSN@BC
   were co‐incubated with RBE cells for 6 h before DCFH‐DA was loaded into
   the cells. Cells were given two PBS washes after 20 min of incubation
   before being exposed to light at a power density of 300 mW cm^−2. After
   irradiation, fluorescence microscope and flow cytometer were used to
   measure the intensity of the fluorescence.

Hypoxia Probe Staining

   The Hypoxyprobe‐1 plus kit (Hypoxyprobe, Burlington, MA) was used to
   identify the hypoxic region in the xenografted tumor following the
   manufacturer's instructions. Briefly, nude mice bearing tumors were
   injected with 60 mg k^−1g of pimonidazole hydrochloride intravenously,
   following 640 nm irradiation (300 mW cm^−2, 3 min). The tumors were
   removed, cut into sections, and subjected to immunohistochemistry
   analysis after 90 min. In accordance with the protocol, mouse
   monoclonal antibody was used to examine frozen tissue slices. DAPI was
   used to label the cell nuclei.

Live/Dead cell staining

   In accordance with manufacturer's instructions, Calcein/PI Cell
   Viability/Cytotoxicity Assay Kit (Beyotime, Shanghai, China) were
   utilized to assess further R‐CM@MSN@BC‐mediated chemo‐PDT effects. RBE
   cells were co‐incubated with the nanoparticles for 4 h before being
   washed three times with PBS for 3 min, exposed to laser irradiation
   (640 nm, 300 mW cm^−2, 3 min), and then incubated with Calcein‐AM (2
   µM) and PI (5 µM) at 37 °C for 20 min before being photographed under a
   fluorescence microscope.

ELISA Analysis

   Following the manufacturer's instructions, the appropriate ELISA Assay
   Kits (Beyotime, Shanghai, China) were used to measure the levels of
   IFN‐γ, IL‐12, IL‐2 and TNFα in the serum of treated C57BL/6 mouse after
   various treatments. The list of elisa kit was provided in Table [195]S4
   (Supporting Information).

Immunofluorescence Staining

   Following the previous protocol,^[ [196]^34 ^] immunofluorescence
   assays were used to evaluate the level of membrane translocation of CRT
   and the extent of HMGB1 leakage in RBE cells after different
   treatments. Also was used to assess the degree of infiltration of CD4^+
   T cells, CD8^+ T cells, in tumor tissues of mice after different
   treatments.

In Vivo Imaging

   The fluorescence imaging and MRI were studied in RBE tumor‐bearing
   Balb/c nude mice. The mice were intravenously injected with the
   R‐CM@MSN@BC and then, imaged by a fluorescence imaging instrument
   (In‐Vivo FX PRO, Bruker, Germany) or an MRI scanner (4.7 T Bruker
   BioSpec, Bruker, Germany) belonged to National Center for Magnetic
   Resonance in Wuhan, Innovation Academy for Precision Measurement
   Science and Technology, CAS. The signal intensity was achieved by
   built‐in software of the imaging systems.

Metabolites Assay In Vitro

   Following the manufacturer's instructions, the intracellular
   metabolites level in RBE cells cultured in different treatment medium
   for 24 h was measured using a GSH and GSSG Assay Kit (Beyotime,
   Shanghai, China), NADP^+/NADPH Assay Kit with WST‐8 (Beyotime,
   Shanghai, China), Micro Glutamic Acid (Glu) Content Assay Kit
   (Solarbio, Beijing, China), LA Assay Kit (Solarbio, Beijing, China).

In Vivo Experiments

   In order to assess the inhibitory effect of photodynamic therapy (PDT)
   mediated by R‐CM@MSN@BC on tumors in mice, a subcutaneous injection of
   5 × 10^6 RBE cells resuspended in saline was administered to the right
   flanks of BALB/c nude mice. The mice were then randomly assigned to one
   of four groups: saline, R‐CM@MSN@BC (10 µg mL^−1), R‐CM@MSN@C
   (10 µg mL^−1) with laser irradiation (+), R‐CM@MSN@B (10 µg mL^−1) with
   laser irradiation (+), or R‐CM@MSN@BC (10 µg mL^−1) with laser
   irradiation. This randomization occurred once the average tumor volume
   reached ≈100 mm^3. Intravenous administration of all medications was
   performed, and subsequent measurements of body weight and tumor volume
   were recorded. The tumor volume was calculated by tumor length × tumor
   width × tumor width /2. Furthermore, the survival time of the BALB/c
   nude mice with tumors (n = 5 per group) was recorded daily for a period
   of 30 days. Histological changes, proliferation, and apoptosis levels
   of the tumors were assessed using H&E staining and the TUNEL assay,
   respectively, to confirm the efficacy of the R‐CM@MSN@BC treatment.
   Representative tumor tissues were also photographed.

   To verify the effect on tumor microenvironment after R‐CM@MSN@BC
   treatment, a C57BL/6 mice model was constructed with subcutaneous graft
   tumor. Briefly, an AKT/Notch intracellular domain (NICD) was first
   obtained in situ intrahepatic cholangiocarcinoma model (purchased from
   Shouzheng Pharma Biotechnology Co., Ltd.),^[ [197]^42 ^] then mice were
   executed and well‐grown tumor tissues were selected and placed in
   sterile glass dishes. tumor tissues in sterile glass dishes, and the
   tumor tissues were cut up to ≈1 mm^3. Next, the skin of the right leg
   of C57BL/6 mice was shaved off, disinfected and the skin of the right
   leg of the mice was cut, and the tumor tissue block was implanted under
   the skin of the right leg of C57BL/6 mice. Finally, the skin incision
   of the mice was sutured and again. After the subcutaneous tumor volume
   grew to about 100 mm^3, the mice were randomly divided into three
   groups (5 mice per group) and treated with saline, R‐CM@MSN@C
   (10 µg mL^−1) with laser irradiation (+), and R‐CM@MSN@BC
   (210 µg mL^−1) with laser irradiation (+), respectively. On day 14, the
   mice were executed and collected their tumor tissues and evaluated the
   immune cell infiltration in the tumor tissues by CyTOF.

   In order to assess the inhibitory effect of R‐CM@MSN@BC in combination
   with anti‐PD‐L1 treatment on distant septal tumors in mice, bilateral
   subcutaneous C57BL/6 tumor models were established. Once the tumor
   volume reached an average size of ≈100 mm^3, the tumor‐bearing mice
   were randomly divided into four groups (n = 5 per group): (1) Saline;
   (2) anti‐PD‐L1; (3) R‐CM@MSN@BC + light; and (4) R‐CM@MSN@BC + light +
   anti‐PD‐L1. The changes in volume of the bilateral tumors were
   measured, and the tumors were dissected and photographed on day 25
   after receiving the respective treatments. Ki‐67 and HE staining of
   tumor tissues were also performed to evaluate the inhibition of primary
   and distant tumors in mice by different treatment modalities.

Statistical Analysis

   All measurements were provided as mean standard deviation (SD). The
   quantitative data was evaluated using the Student's t‐test or the
   Mann‐Whitney U test. To compare the two groups statistically in paired
   samples, the Wilcoxon signed‐rank test or the paired t‐test were used.
   The count data was examined using the Pearson Chi‐Square test. The
   appropriate test was used for ranking data, either the Kruskal‐Wallis
   test or the Mann‐Whitney U test. The repeated measures ANOVA was used
   to assess the differences in tumor volume between groups. To analyze
   the statistical variances between survival curves, the log‐rank test
   was used. Prism 8.0 (GraphPad, San Diego, USA) was used to estimate the
   survival curves. The R software version 4.1.0 (Auckland, NZ) for all
   testing (*P < 0.05, **P < 0.01, ***P < 0.001) was used.

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [198]ADVS-11-2309203-s001.pdf^ (3.3MB, pdf)

Acknowledgements