Abstract

   CD47 blockade has emerged as a promising immunotherapy against liver
   cancer. However, the optimization of its antitumor effectiveness using
   efficient drug delivery systems or combinations of therapeutic agents
   remains largely incomplete. Here, patients with liver cancer
   co‐expressing CD47 and CDC7 (cell division cycle 7, a negative
   senescence‐related gene) are found to have the worst prognosis.
   Moreover, CD47 is highly expressed, and senescence is inhibited after
   the development of chemoresistance, suggesting that combination therapy
   targeting CD47 and CDC7 to inhibit CD47 and induce senescence may be a
   promising strategy for liver cancer. The efficacy of intravenously
   administered CDC7 and CD47 inhibitors is limited by low uptake and
   short circulation times. Here, inhibitors are coloaded into a
   dual‐targeted nanosystem. The sequential release of the inhibitors from
   the nanosystem under acidic conditions first induces cellular
   senescence and then promotes immune responses. In an in situ liver
   cancer mouse model and a chemotherapy‐resistant mouse model, the
   nanosystem effectively inhibited tumor growth by 90.33% and 85.15%,
   respectively. Overall, the nanosystem in this work achieved the
   sequential release of CDC7 and CD47 inhibitors in situ to trigger
   senescence and induce immunotherapy, effectively combating liver cancer
   and overcoming chemoresistance.

   Keywords: combination therapy, dual‐target, immunotherapy, senescence,
   sequential release
     __________________________________________________________________

   The sequential release of a senescence inducer (CDC7 inhibitor) and a
   CD47 inhibitor from the dual‐targeting nanosystem in situ induced
   senescence and antitumor immune responses to aganist liver cancer and
   overcome chemoresistance.

   graphic file with name ADVS-11-2308310-g004.jpg

1. Introduction

   Hepatocellular carcinoma (HCC) is one of the most common types of
   primary cancer and is the fourth leading cause of cancer‐related
   mortality worldwide.^[ [44]^1 ^] Immunotherapies based on tumor
   antigen‐targeting antibodies and immunomodulatory antibodies are
   emerging approaches that may improve HCC treatment outcomes. CD47 is an
   innate inhibitory immune checkpoint that activates the “Don't eat me”
   signal by binding to the N‐terminus of SIRP‐α (signal regulatory
   protein α), a signal regulatory protein on immune cells.^[ [45]^2 ,
   [46]^3 , [47]^4 ^] CD47 is highly expressed on HCC cells, allowing them
   to evade the immune system.^[ [48]^5 ^] Studies have shown that
   blocking the CD47/SIRP‐α signaling axis can promote the phagocytosis
   and polarization of M1 macrophages.^[ [49]^6 ^] Researchers have found
   that inhibiting CD47 can be effective in treating cancer.^[ [50]^7 ,
   [51]^8 , [52]^9 ^] The effectiveness of immunotherapy depends on the
   immune status of patients with HCC, and immunotherapy often needs to be
   combined with other treatments.^[ [53]^10 , [54]^11 ^] Combination
   treatment with CD47 monoclonal antibodies and PD‐1 monoclonal
   antibodies effectively inhibits the cancer.^[ [55]^12 ^] Combined
   targeting of CD47 and glypican‐3 is efficacious against dual
   antigen‐expressing HCC.^[ [56]^13 ^] The strategies mentioned above
   illustrate that anti‐CD47‐based combination therapies offer promise for
   the clinical treatment of HCC. However, only a limited subgroup of
   patients benefit from these treatments. Thus, more efficacious
   anti‐CD47‐based combination treatments are needed for the treatment of
   a larger proportion of HCC patients.

   Research has shown that inducing senescence in cancer cells promotes
   antitumor immune responses.^[ [57]^14 ^] Cellular senescence is usually
   a response to intrinsic or extrinsic cellular stresses such as DNA
   damage.^[ [58]^15 ^] Senescence is often considered a protective tumor
   suppressor because it prevents cancer cell proliferation and
   progression.^[ [59]^16 ^] Growth‐stunted cancer cells, however, can
   escape senescence and produce more aggressive tambours.^[ [60]^14 ^]
   The studies mentioned above indicate that we can treat HCC by inducing
   senescence. Inhibiting CDC7 prevents the start of DNA replication and
   promotes the apoptosis of cancer cells.^[ [61]^17 , [62]^18 ^]
   Moreover, CDC7 inhibition promotes HCC cell senescence,^[ [63]^19 ^]
   indicating that CDC7 inhibitors can act as senescence activators.
   Therefore, combination therapy involving both CD47 inhibitors and
   senescence activators (CDC7 inhibitors) is a promising strategy for HCC
   treatment.

   Traditional drugs are rapidly metabolized, cause systemic toxicity, and
   are insufficiently effective in the treatment of tumors.^[ [64]^20 ^]
   Moreover, for multidrug combinations, the frequency of dosing and the
   intervals between doses need to be addressed, and numerous doses are
   often needed, which imposes a certain burden on both patients and
   doctors.^[ [65]^21 , [66]^22 ^] Nanoparticles (NPs) are able to
   codeliver lipid‐soluble and water‐soluble drugs.^[ [67]^23 , [68]^24 ^]
   NPs are stable in serum, thus prolonging the drug circulation time,
   increasing the accumulation of drugs at tumor sites through enhanced
   permeability and retention (EPR) effects, reducing toxic side effects,
   and decreasing the frequency of drug administration.^[ [69]^25 ,
   [70]^26 , [71]^27 ^]

   Because the EPR effect is limited, to improve the active targeting of
   NPs to tumors, we prepared NPs modified with a CD47 antibody (AbCD47).
   Because CD47 is highly expressed on HCC,^[ [72]^5 ^] AbCD47 can target
   and block CD47 on the surface of tumor cells to disable the “Don't eat
   me” signal. However, tumor heterogeneity causes the amount of CD47 to
   vary from cell to cell,^[ [73]^28 ^] leading to targeting failure. To
   ensure adequate tumors targeting, a second target needs to be added.
   Folate receptor 1 (FOLR1), was originally identified as a folic acid
   (FA)‐binding protein,^[ [74]^29 , [75]^30 ^] has been shown to be
   barely expressed in normal cells but widely overexpressed in solid
   tumor cells, including HCC cells.^[ [76]^31 ^] These studies suggest
   that FOLR1 may serve as a specific tumor antigen for HCC cells.
   Therefore, we utilized FA as the second targeting moiety to target HCC
   cells by binding with FOLR1 to remedy the absence of target genes due
   to tumor heterogeneity.

   We constructed dual‐targeting calcium‐phosphorus NPs coloaded with CD47
   and CDC7 inhibitors using
   1,2‐distearoyl‐sn‐glycero‐3‐phosphoethanolamine‐N‐[amino(polyethylene
   glycol)−2000] (DSPE‐PEG2000‐NH[2]), which is a main component of cell
   membranes, and PEG2000, which has been approved by the FDA for use in
   humans.^[ [77]^32 ^] Calcium‐phosphorus NPs with pH‐responsive
   properties ensure the long‐term sequential release of drugs into
   tumors.^[ [78]^33 ^] Compared to conventional drug delivery, our
   nanodelivery system is expected to provide a new therapeutic option for
   liver cancer treatment due to its dual‐targeting and combined induction
   of senescence and immunomodulation. (Scheme [79]1 )

Scheme 1.

   Scheme 1
   [80]Open in a new tab

   Schematic diagram of the use of AX@NPs‐FA/AbCD47 for the treatment of
   liver cancer to induce senescence and antitumor immune responses. The
   dual‐targeting AX@NPs‐FA/AbCD47 gradually accumulated at the tumor site
   and were then taken up by HCC cells. Next, the NPs escaped the lysosome
   and then sequentially released the drugs XL413 and AZD8055 within the
   acidic tumor niche to block CDC7 and CD47, inducing apoptosis, cellular
   senescence, and antitumor immune responses for the treatment of liver
   cancer. In addition, the NPs target mitochondria and induce the
   production of ROS, thereby destroying the tumor microenvironment and
   enhancing the antitumor immune response.

2. Results

2.1. CD47 and CDC7 Overexpression Predicted Poor Prognosis in Patients with
HCC

   By screening 149 pairs of liver cancer specimens in the GEO database
   [81]GSE76297 (Figure [82]1a,b), we found that CD47 and CDC7 were highly
   expressed in cancer tissues (Figure [83]1c,d). An additional analysis
   of five pairs of liver cancer specimens ([84]GSE146049) confirmed that
   CD47 and CDC7 were highly expressed in HCC tissues (Figure [85]S1,
   Supporting Information). These results suggest that CD47 and CDC7 have
   clinical diagnostic value. To further evaluate the clinical diagnostic
   value of CD47 and CDC7, we analyzed the TCGA database and divided 365
   patients with HCC into high‐expression and low‐expression groups
   according to the fragments per kilobase million (FPKM) values of CD47
   and CDC7. Figure [86]1e,f shows that high expression levels of CD47 or
   CDC7 predicted poor prognosis in HCC patients. In addition, with the
   HCC patients with the worst prognosis (p≤0.0001) had high expression of
   both CD47 and CDC7 (Figure [87]1g). These results suggest that CD47 and
   CDC7 are promising therapeutic targets for HCC. Figure [88]1h and
   Figure [89]S2 (Supporting Information) show that the CD47 and CDC7
   expression levels in HCC patients were positively correlated,
   suggesting that these proteins may have a synergistic tumor‐promoting
   effect. The expression of major histocompatibility complex I (MHC I)
   determines the immune antitumor response.^[ [90]^34 ^] Hence, we
   analyzed the sequencing results of 149 patients, and Figure [91]S3
   (Supporting Information) shows that the expression of MHC class I
   chain‐related proteins A and B (MICA and MICB) was significantly
   upregulated in HCC patients. This result suggested that HCC tumor cells
   are sensitive to immunotherapy. To further verify the upregulated
   expression of CD47 and CDC7 in HCC, we compared the HCC cell lines
   HEP3B and HUH7 with the normal hepatocyte cell line LO2. Figure [92]1i
   shows that CD47 and CDC7 were significantly upregulated in HCC cells
   compared to normal hepatocytes. We also examined CD47 and CDC7
   expression in Hepa1‐6 in situ HCC model mice, and Figure [93]1j shows
   that CD47 and CDC7 were both more highly expressed in tumor tissue than
   in adjacent normal liver tissue. These results suggest that CD47 and
   CDC7 may be potential targets for the clinical treatment of HCC.
   Moreover, we generated a chemoresistant HCC cell line, Bel‐7402R, from
   the parental cell line Bel‐7402P. We performed RNA sequencing of
   Bel‐7402R and Bel‐7402P cells and demonstrated that cell senescence and
   DNA damage repair‐related pathways were significantly enriched after
   the development of drug resistance (Figure [94]1k; Figure [95]S4,
   Supporting Information), and CD47 was upregulated in the drug‐resistant
   cell line compared to the sensitive cell line (Figure [96]1l).
   Moreover, the expression of senescence‐associated genes (TP53 and H2AX)
   was downregulated in drug‐resistant cells (Figure [97]1l). These
   results revealed the downregulation of cellular senescence‐associated
   genes and upregulation of CD47 after the development of drug resistance
   in HCC, and CDC7 is negatively correlated with senescence, suggesting
   that the inhibition of CDC7 and CD47 may be able to overcome
   chemoresistance in HCC.

Figure 1.

   Figure 1
   [98]Open in a new tab

   CD47 and CDC7 overexpression predicted poor prognosis in patients with
   HCC. a,b) The expression of CD47 and CDC7 in 149 liver cancer patients
   (tumor tissue and adjacent normal tissue) was analyzed in [99]GSE76297.
   c,d) Statistical analysis of CD47 and CDC7 expression in HCC. e,f)
   Analysis of the TCGA database to reveal the effect of CD47 and CDC7
   expression on survival time in patients with HCC. g) Effects of CD47
   and CDC7 expression on the prognosis of patients with HCC. L indicates
   low expression, and H indicates high expression. h) Correlation
   analysis of CD47 and CDC7 in 149 liver cancer patients. i)
   Immunofluorescence analysis of CD47 and CDC7 expression in HCC cells
   (HEP3B and HUH7) and normal hepatocytes (LO2). Green represents CD47,
   red represents CDC7, and blue represents the nucleus. j)
   Immunohistochemical results of CD47 and CDC7 in Hepa1‐6 tumor‐bearing
   mice. k) Pathway enrichment analysis of Bel‐7402P and Bel‐7402R using
   GSEA software. l) Sequencing results for the differentially expressed
   genes in Bel‐7402P (1‐P, 2‐P, and 3‐P) and Bel‐7402R (1‐R, 2‐R, and
   3‐R). The data are present as the mean ± SEM, and a p‐value less than
   0.05 was considered to indicate statistical significance.

2.2. Downregulation of CD47 and CDC7 Synergistically Induced Cellular
Senescence and Phagocytosis In Vitro

   Studies have shown that CD47 signaling is mediated by the activation of
   mTOR (mammalian target of rapamycin).^[ [100]^35 , [101]^36 ^] Hence,
   we predicted that inhibiting mTOR could suppress CD47. AZD8055 is a
   mTOR inhibitor that has been investigated in clinical trials.^[
   [102]^37 , [103]^38 ^] We used computer‐simulated docking to evaluate
   the binding of the CD47 protein to AZD8055. The findings indicated that
   AZD8055 effectively docks at Lys39 (lysine 39, depicted in orange) and
   Tyr37 (tyrosine 37, depicted in blue) in CD47, by forming hydrogen
   bonds, as demonstrated in Figure [104]2a. Furthermore, treating the
   HEP3B and HUH7 cell lines with AD8055, elicited a notable reduction in
   CD47 expression, as illustrated in Figure [105]S5 (Supporting
   Information). This finding suggested that AZD8055 has the potential to
   serve as a potent small‐molecule inhibitor of CD47. XL413 is a
   selective inhibitor of CDC7. Figure [106]2b,c indicates that AZD8055
   and XL413 effectively killed HCC cells. The concentration that
   inhibited cell proliferation by 50% relative to the control (IC50) was
   calculated using GraphPad Prism 8.0 (HEP3B: AZD8055 0.5 µm, XL413 60
   µm; HUH7: AZD8055 0.4 µm, XL413 30 µm). Synergy was detected following
   the Chou‐Talalay method (combination index: CI < 1 indicates synergy)^[
   [107]^39 ^] using CompuSyn software (Ver.1.0). The synergistic index of
   HEP3B and HUH7 was less than 1 when the fraction affected (Fa) was 0.5
   (Figure [108]2d,e). Thus, we determined the synergistic ratio of the
   two drugs in HCC cells at the Fa value of 0.5 (HEP3B: 0.08 µm AZD8055 ±
   9.5 µm XL413; HUH7: 0.03 µm AZD8055 ± 2.1 µm XL413). Figure [109]2f,g
   shows that AZD8055 and XL413 downregulated CD47 and CDC7, respectively.
   We verified the synergistic effect of the two drugs with CCK8 assays,
   and the results showed that the drug combination had significant
   cytotoxicity, and the combination drug dose was much lower than the
   single drug doses (Figure [110]2h,i). Studies have indicated that XL413
   can induce cellular senescence.^[ [111]^19 ^] As shown in
   Figure [112]2j,k, XL413 promoted cellular senescence, and when combined
   with AZD8055, it also promoted cellular senescence. A certain degree of
   cellular senescence induces apoptosis,^[ [113]^40 ^] and AZD8055 can
   promote apoptosis. Figure [114]S6 (Supporting Information) shows that
   AZD8055 and XL413 promoted apoptosis in HCC cells either individually
   or in combination. The downregulation of CD47 on cells disengages the
   “Don't eat me” signal and promotes phagocytosis.^[ [115]^41 ^]
   Figure [116]2l,m and Figure [117]S7 (Supporting Information) show that
   AZD8055 significantly promoted phagocytosis; in addition, XL413
   promoted phagocytosis, which we speculate is because the drugs induced
   cellular senescence or apoptosis and macrophages perform phagocytosis
   of senescent and damaged cells (including apoptotic cells). In summary,
   we found that AZD8055 downregulated CD47 expression and synergized with
   the CDC7 inhibitor XL413 to promote cellular senescence and
   phagocytosis.

Figure 2.

   Figure 2
   [118]Open in a new tab

   Downregulation of CD47 and CDC7 synergistically induced senescence and
   phagocytosis in HCC cells. a) Left: Overview of the molecular modelling
   of AZD8055 binding to CD47 (PDB code: 2VSC); Right: The inside binding
   residues, Lys39 (orange) and Tyr37 (blue). Yellow dashed lines,
   hydrogen bonds between AZD8055 and CD47. b,c) The CCK8 assay was used
   to analyse the IC50 values of AZD8055 and XL413 against HEP3B and HUH7
   cells. d,e) CompuSyn software was used to determine the synergistic
   effect of AZD8055 and XL413 on HEP3B and HUH7 cells, and the
   Chou‐Talalay method was used (combination index: CI < 1 indicated
   synergistic effect). f,g) Immunofluorescence analysis of CD47 and CDC7
   expression in cells (HEP3B and HUH7) treated with drugs for 48 h
   (HEP3B: control, 0.5 µm AZD8055, 60 µm XL413, 0.08 µm AZD8055 ± 9.5 µm
   XL413). HUH7: control, 0.4 µm AZD8055, 30 µm XL413, 0.03 µm AZD8055 ±
   2.1 µm XL413). Green represents CD47, red represents CDC7, and blue
   represents the nucleus, n = 3, mean ± SEM, ^* p ≤0.05, ^** p ≤0.01,
   ^*** p ≤0.001. h,i) A CCK8 assay was used to determine the viability of
   HEP3B and HUH7 after incubation with certain drugs for 72 h (HEP3B:
   control, 0.5 µM AZD8055, 60 µm XL413, 0.08 µm AZD8055 + 9.5 µm XL413;
   HUH7: control, 0.4 µm AZD8055, 30 µm XL413, 0.03 µm AZD8055 + 2.1 µm
   XL413), n = 5, mean ± SEM, ^*** p ≤0.001. j,k) β‐Galactosidase staining
   was used to determine cellular senescence in HEP3B and HUH7 cells after
   treatment with drugs for 3 days (HEP3B: control, 0.5 µm AZD8055, 60 µm
   XL413, 0.08 µm AZD8055 + 9.5 µm XL413;. HUH7: control, 0.4 µm AZD8055,
   30 µm XL413, 0.03 µm AZD8055 + 2.1 µm XL413). l,m) THP‐1 cells were
   induced to become macrophages and cocultured with HCC cells with drugs
   for 1 h. (HEP3B: control, 0.5 µm AZD8055, 60 µm XL413, 0.08 µm AZD8055
   + 9.5 µm XL413; HUH7: control, 0.4 µm AZD8055, 30 µm XL413, 0.03 µm
   AZD8055 + 2.1 µm XL413) The phagocytosis of HCC cells by macrophages
   was observed. Green represents HCC cells; the bright field represents
   THP‐1 cells.

2.3. Formation and Characterization of NPs

   AbCD47 can target and block CD47 on the surface of tumor cells to
   disable the “Don't eat me” signal. However, tumor heterogeneity causes
   the amount of CD47 to vary from cell to cell,^[ [119]^28 ^] leading to
   targeting failure. To ensure tumor targeting, another target needs to
   be added. We found that FOLR1 was highly expressed in HCC tissues
   (Figure [120]S8a,b, Supporting Information), and TCGA database analysis
   indicated that high FOLR1 expression predicted poor patient prognosis
   (Figure [121]S8c, Supporting Information). FOLR1 was originally
   identified as a folic acid (FA)‐binding protein^[ [122]^29 , [123]^30
   ^] and is barely expressed in normal cells.^[ [124]^31 ^] Hence, FA was
   utilized as an auxiliary targeting molecule to improve targeting.

   Figure [125]3a illustrates the formation process of AX@NPs‐FA/AbCD47.
   Figure [126]S9a (Supporting Information) shows that the NPs were
   modified with AbCD47. The free AbCD47 moved through the gel faster than
   the AbCD47 attached to the NPs, and SDS‒PAGE indicated that the NPs
   were successfully modified with AbCD47 (Figure [127]3b). NPs‐FA/AbCD47
   were photographed by atomic force microscopy (AFM), which showed a ring
   of small particles circling the NPs, forming a “shell‐core” structure
   (Figure [128]3c). Figure [129]3d shows the adsorption of calcium and
   phosphorus on the NP surface. Figure [130]S9b–e (Supporting
   Information) shows that AX@NPs‐FA/AbCD47 had a diameter of
   122.5 ± 2.37 nm and a PDI of 0.235 ± 0.013. After incubation in 5% FBS
   solution (pH 7.4) at 37 °C for 7 days, the sizes and PDI of the NPs
   changed very little, indicating that the NPs remained stable in
   storage. At pH 7.4, these NPs were negatively charged (−10.3 mV), which
   prevents the adsorption of albumin and favors circulation in the
   blood,^[ [131]^42 ^] whereas at pH 6.0, they were positively charged
   (0.00705 mV), which facilitates uptake by tumor cells.^[ [132]^43 ^]
   The encapsulation efficiency (EE%) values for AZD8055 and XL413 were
   83.5% and 75.1%, respectively. We mimicked in vivo drug release
   conditions in vitro, and the results indicated that AZD8055 and XL413
   were released faster from NPs at pH 6.0 than at pH 7.4. Moreover, XL413
   was released from the NPs earlier than AZD8055 (Figure [133]3e–h).
   Calcein/RB@NPs‐FA/AbCD47 with an outer layer of adsorbed calcein and an
   inner layer of rhodamine (RB) were injected into mice bearing Hepa1‐6
   subcutaneous tumors via the tail vein, and small animal live imaging
   revealed changes in fluorescence with time. Little drug enrichment was
   observed at the tumor site before 0.5 h. After 6 h, the drugs
   accumulated at the tumor site, and the drug in the outer layer
   (calcein: green fluorescence) was released first, while the drug in the
   inner layer (rhodamine: red fluorescence) was released later
   (Figure [134]3i). These results confirmed that calcium and phosphorus
   were sensitive to pH changes, guaranteeing the sequential release of
   AZD8055 and XL413 in the acidic tumor niche. Moreover, the TEM results
   suggested that the NPs targeted mitochondria in the tumor cells
   (Figure [135]3j). Figure [136]S10 (Supporting Information) implies that
   the NPs do not cause haemolysis, indicating suitable biosafety. To
   explore the CD47 targeting ability of the CD47 antibody, we incubated
   free RB and RB NPs with HCC cells. Figure [137]S11 (Supporting
   Information) shows that RB@NPs‐FA/AbCD47 were taken up by HCC cells
   much more frequently than free RB, which indicated that the FA and CD47
   antibodies modifications promoted NP uptake by HCC cells. Figure
   [138]S12 (Supporting Information) shows that RB@NPs‐FA/AbCD47
   selectively targeted HCC cells but barely targeted erythrocytes.
   Escaping the degradation and clearance of the lysosome is a
   prerequisite for the effective delivery of a drug by NPs and the
   ability to exert its effects.^[ [139]^44 ^] Partial phagocytosis of NPs
   by lysosomes was observed at 1 h, and after incubation for 4 h, most of
   the NPs had escaped lysosomal degradation (Figure [140]S13, Supporting
   Information), suggesting that the NPs could effectively deliver drugs
   into cells. In summary, the NPs released drugs in a sequential manner.
   Moreover, the NPs were readily taken up by HCC cells and evaded
   lysosomal clearance.

Figure 3.

   Figure 3
   [141]Open in a new tab

   Formation and characterization of NPs. a) Synthetic procedure diagram
   of AX@NPs‐FA/AbCD47. b) Detection of AbCD47 modified NPs by SDS‒PAGE
   gel; the amount of free AbCD47 was 4 µg, and the NPs‐FA/AbCD47
   contained the same dose of AbCD47. c) Morphology of the NPs‐FA/AbCD47
   shown by AFM. d) Calcium and phosphorus in the NPs were elementally
   analyzed by field emission TEM. HAADF: high‐angle annular dark field.
   e,f) Release of AZD8055 and XL413 at pH 7.4. g,h) Release of AZD8055
   and XL413 at pH 6.0. i) Live imaging of small animals to visualize the
   release of nanodrugs (Calcein/RB@NPs‐FA/AbCD47) in animals. Red
   represents rhodamine inside the NPs; green represents calcein inside
   the shell. j) TEM images of NPs targeting mitochondria. Red arrows
   indicate NPs, and green arrows indicate mitochondria.

2.4. Mitochondrion‐Targeted AX@NPs‐FA/AbCD47 Increased Cellular Senescence
and Phagocytosis In Vitro

   One study reported that DSPE‐PEG2000‐NH[2] can participate in
   mitochondrial targeting.^[ [142]^45 ^] We treated HCC cells with
   NPs‐Cy3‐FA/AbCD47. Figure [143]4a,b shows that at 1 h, the NPs did not
   target the mitochondria; presumably, they were taken up by lysosomes
   and had not yet escaped. At 4 h, the NPs had targeted the mitochondria.
   Figure [144]4c,d shows that the nanodrugs were more effective at
   killing HCC cells. AX@NPs‐FA/AbCD47 significantly decreased the LDH
   level (Figure [145]4e,f). Mitochondria are the primary source of
   cellular reactive oxygen species (ROS),^[ [146]^46 ^] and we examined
   ROS levels in HCC cells. Figure [147]4g,h and Figure [148]S14
   (Supporting Information) show that treatment with AX@NPs‐FA/AbCD47
   elicited a large amount of intracellular ROS production. The increased
   ROS production indicates damage to the HCC cells. Our previous study
   showed that AZD8055 could act as a small molecule inhibitor of CD47,
   and that downregulation of CD47 could promote the phagocytosis of tumor
   cells by macrophages. Figure [149]4i,j and Figure [150]S15 (Supporting
   Information) show that the nanodrug significantly promoted the
   phagocytosis of HCC cells by macrophages. As XL413 promoted tumor cell
   senescence, we treated HCC cells with AX@NPs‐FA/AbCD47, and
   Figure [151]4k,l shows that AX@NPs‐FA/AbCD47 induced HCC cell
   senescence at an early stage. These results confirm that
   AX@NPs‐FA/AbCD47 can trigger HCC cell senescence and promote
   phagocytosis, thereby killing tumor cells.

Figure 4.

   Figure 4
   [152]Open in a new tab

   Mitochondrion‐targeted AX@NPs‐FA/AbCD47 increased cellular senescence
   and phagocytosis in vitro. a,b) Fluorescence micrographs of NPs
   targeting the mitochondria of HCC cells at 1 h and 4 h. Red represents
   NPs‐Cy3‐FA/AbCD47, green represents the mitochondria, and blue
   represents the nucleus. Red and green indicate that the NPs did not
   target the mitochondria, whereas red and green combined with orange
   indicate that the NPs did target the mitochondria. c,d) A cytotoxicity
   assay was used to verify the killing effect of the nanodrugs on HEP3B
   and HUH7 cells, n = 6, mean ± SEM, ^** p ≤ 0.01, ^*** p ≤0.001. e,f)
   Treatment of HCC cells with the nanodrugs and measurement of LDH levels
   released from cells, n = 3, mean ± SEM, ns means indicates no
   significant difference, ^* p ≤0.05, ^** p ≤0.01. g,h) Detection of
   intracellular ROS production by fluorescence microscopy after the
   nanodrug treatment of HCC cells. Green represents ROS; the bright field
   represents HCC cells. i,j) THP‐1 cells were induced to become
   macrophages and cocultured with HCC cells for 1 h. The phagocytosis of
   HCC cells by macrophages was observed. Green represents HCC cells;
   bright field represents THP‐1 cells. k,l) β‐Galactosidase staining to
   observe AX@NPs‐FA/AbCD47‐induced senescence in HCC cells.

2.5. AX@NPs‐FA/AbCD47 Alleviated Liver Cancer and Overcame Chemoresistance In
Vivo

   A prerequisite for NP treatment of tumors in vivo is effective delivery
   to the tumor.^[ [153]^47 ^] We subcutaneously inoculated HUH7 cells
   into nude mice and used an in vivo imaging system to evaluate the
   tumor‐targeting ability of the NPs. Figures [154]5a and [155]S16
   (Supporting Information) show that Dir@NPs‐FA/AbCD47 had the best
   targeting ability, and a large amount of fluorescence was still present
   at the tumor site at 72 h. In addition, when the subcutaneous tumor in
   the Dir@NPs‐FA/AbCD47 group metastasized to the spleen and lung, the
   NPs effectively targeted this metastasis site, while the other groups
   with metastasized tumors showed no fluorescence in the spleen and lung.
   We subcutaneously inoculated HUH7 liver cancer cells into the liver, as
   shown in Figure [156]S17a (Supporting Information). Figure [157]S17b
   (Supporting Information) shows the therapeutic period of the mice.
   Tumor growth was recorded in vivo with small animal ultrasound
   equipment, and Figure [158]5b,c revealed that the AX@NPs‐FA/AbCD47
   group had the best treatment effect, and the HUH7 tumor growth
   inhibition value reached 90.33% (Figure [159]5d). In addition, the
   growth status of the mice in the AX@NPs‐FA/AbCD47‐treated group was
   better than that of the mice in the control group (Figure [160]S18,
   Supporting Information). To further verify the effectiveness of
   AX@NPs‐FA/AbCD47, we established a Hepa1‐6 HCC in situ C57BL/6 mouse
   model, as shown in Figure [161]S17a (Supporting Information). The
   treatment period is shown in Figure [162]S17b (Supporting Information),
   and at the end of the treatment, the C57BL/6 mice were sacrificed and
   dissected. The results in Figure [163]5e suggest that treatment with
   AX@NPs‐FA/AbCD47 had the best treatment effect. In addition,
   AX@NPs‐FA/AbCD47 effectively inhibited the infiltration of HCC cells
   into normal liver tissues adjacent (<0.5 cm) to the cancerous tissues
   (Figure [164]5f). To observe the therapeutic effect of AX@NPs‐FA/AbCD47
   on drug‐resistant HCC cells, we established drug‐resistant Bel‐7402R
   subcutaneous tumors. AX@NPs‐FA/AbCD47 markedly inhibited the
   intratumoral blood supply (Figure [165]5g), and Figure [166]5h shows
   that the NPs clearly inhibited the growth of the drug‐resistant HCC
   tumors (tumor growth inhibition value: 85.15%). In addition, the CD47,
   CDC7 and mTOR levels in Bel‐7402R tumors were significantly reduced
   after treatment (Figure [167]5i). Furthermore, treatment with
   AX@NPs‐FA/AbCD47 promoted the accumulation of the senescence‐related
   protein H2AX in Bel‐7402R tumors, suggesting that the nanomedicine
   effectively induced senescence in drug‐resistant tumors
   (Figure [168]5j; Figure [169]S19, Supporting Information). The above
   results demonstrated that AX@NPs‐FA/AbCD47 were effective in treating
   HCC and drug‐resistant HCC.

Figure 5.

   Figure 5
   [170]Open in a new tab

   AX@NPs‐FA/AbCD47 alleviated liver cancer and overcame chemoresistance
   in vivo. a) An in vivo imaging system was used to detect the targeting
   of NPs (HUH7 subcutaneous tumor in nude mice). b) In situ ultrasound
   imaging results of HUH7 liver cancer in different treatment groups in
   nude mice for different periods. c) The tumor growth curve of HUH7
   liver cancer in nude mice, n = 5. d) The tumor growth inhibition value
   after treatment, n = 5, mean ± SEM, ^* p ≤0.05, ^*** p ≤0.001. e) Ex
   vivo anatomical images of different treatment groups of C57BL/6 mice
   (Hepa1‐6). f) H&E staining of liver tissue within 0.5 cm of the cancer
   after treatment of Hepa1‐6 in situ HCC. g) Intratumoral blood
   oxygenation observed by photoacoustic imaging of subcutaneous Bel‐7402
   tumors in mice after treatment and the statistical results of
   intratumoral blood oxygenation, n = 5, mean ± SEM, ^*** p ≤0.001. h)
   Photographs of tumors in mice after treatment and statistical analysis
   of tumor weight in mice, n = 5, mean ± SEM, ^*** p ≤0.001; i) Heatmap
   of CD47, CDC7, and mTOR expression in tumors after treatment with PBS
   and AX@NPs‐FA/AbCD47. j) H2AX staining in Bel‐7402R tumor‐bearing mice
   after treatment.

2.6. AX@NPs‐FA/AbCD47 Downregulated CD47 and CDC7 to Inhibit Tumor Growth and
Promote Cellular Senescence In Vivo

   We analyzed the expression levels of CD47, CDC7, and mTOR in the tumors
   of C57BL/6 mice after different treatments, and Figure [171]6a shows
   that AX@NPs‐FA/AbCD47 significantly downregulated CD47, CDC7, and mTOR
   to inhibit angiogenesis and promote apoptosis (Figure [172]6b–e). We
   examined the senescence of tumor cells, and Figure [173]6f shows that
   AX@NPs‐FA/AbCD47 significantly promoted the senescence of tumor cells.
   In addition, we performed H&E staining of the heart, spleen, lung, and
   kidney of the mice, and the results showed that the treatments had no
   toxic effects on these organs (Figure [174]S20, Supporting
   Information). These findings suggest that AX@NPs‐FA/AbCD47 promoted HCC
   cellular senescence and inhibited tumor growth by downregulating CD47
   and CDC7.

Figure 6.

   Figure 6
   [175]Open in a new tab

   AX@NPs‐FA/AbCD47 promoted cellular senescence and inhibited tumor
   growth by downregulating CD47 and CDC7. a) The expression of CD47,
   CDC7, mTOR, and CD31 in Hepa1‐6 tumors from each treatment group and
   TUNEL staining of Hepa1‐6 tumors from C57BL/6 mice in different
   treatment groups. Green represents TUNEL staining, and blue represents
   nuclei. b) Statistical analysis of CD47 expression in various treatment
   groups. c) Statistical analysis of CDC7 expression in various treatment
   groups. d) Statistical analysis of mTOR expression in various treatment
   groups; e) Statistical analysis of CD31 expression in various treatment
   groups. f) Staining for SA‐β‐Gal in Hepa1‐6 tumors from each treatment
   group. n = 5, mean ± SEM, ^* p <0.05, ^** p <0.01, ^*** p <0.001.

2.7. AX@NPs‐FA/AbCD47 Enhanced Antitumor Responses In Vivo

   We dissected the livers of C57BL/6 mice at the end of treatment,
   isolated liver cancer tissues and adjacent liver tissues (>0.5 cm), and
   extracted tumor‐infiltrating lymphocytes (TILs) and individual liver
   tissue cells (Figure [176]7a). Figure [177]7b–e shows that the
   AX@NPs‐FA/AbCD47 group exhibited more M1 macrophages
   (CD45+F4/80+CD11b+CD86+) than the other groups both in liver tissue and
   among TILs, and the number of M1 macrophages was significantly greater
   in the tumor tissue than in the adjacent normal liver tissue
   (Figure [178]7f). We examined the infiltration of mature DCs and T
   cells within TILs and found that the numbers of mature DCs, CD4+ helper
   T cells, and CD8+ cytotoxic T cells were significantly greater after
   treatment with AX@NPs‐FA/AbCD47 (Figure [179]7g–i), confirming that
   AX@NPs‐FA/AbCD47 promoted immune responses and induced immunogenic cell
   death in tumors. The above results suggest that AX@NPs‐FA/AbCD47
   promote antitumor immune responses, especially M1‐type differentiation
   of macrophages.

Figure 7.

   Figure 7
   [180]Open in a new tab

   AX@NPs‐FA/AbCD47 increased M1 macrophage differentiation and activated
   the antitumor immune response. a) Separation of tumor‐infiltrating
   lymphocytes and single liver tissue cells. b) Flow cytometry analysis
   of M1 macrophages in hepatocytes. M1 macrophages were identified as
   CD45+F4/80+CD11b+CD86+. c) Flow cytometry analysis of M1 macrophages
   among TILs. M1 macrophages were identified as CD45+F4/80+CD11b+CD86+.
   d) Differences in the percentage of M1 macrophages in adjacent normal
   hepatocytes after treatment with different drugs. e) Differences in the
   percentage of M1 macrophages in TILs after treatment with different
   drugs. f) Differences in the percentage of M1 macrophages in TILs and
   adjacent normal hepatocytes after treatment with different drugs. g)
   Statistics of infiltration of mature DCs in TILs of mice treated with
   different drugs; mature DCs were identified as CD45+CD11c+ CD80+CD86+.
   h) Statistics of infiltration of CD4+ T cells in TILs of mice treated
   with different drugs. CD4+ T cells were identified as CD45+CD3+CD4+. i)
   Statistics of the infiltration of CD8+ T cells in the TILs of mice
   treated with different drugs. CD8+ T cells were identified as
   CD45+CD3+CD8+; n = 3, mean ± SEM, ^* p <0.05, ^** p <0.01, ^***
   p <0.001.

3. Discussion and Conclusion

   Single drugs are prone to off‐target effects and ultimately lead to
   chemoresistance,^[ [181]^48 ^] while dual drugs can solve this problem
   by preventing drug insensitivity caused by single‐site mutations or the
   use of alternative pathways.^[ [182]^49 ^] Our research revealed that
   the combined inhibition of CD47 and CDC7 was effective in the treatment
   of HCC.

   To make anti‐CD47 immunotherapy clinically feasible, safety
   considerations need to be addressed.^[ [183]^50 ^] CD47 is highly
   expressed in HCC,^[ [184]^5 ^] but it is also highly expressed in
   normal erythrocytes,^[ [185]^51 ^] and anaemia can occur as a result of
   the phagocytosis induced by CD47 blockers.^[ [186]^52 ^] Currently, the
   development of CD47 blockers has not been able to overcome the
   challenge of cytotoxicity (anaemia and hemocytopenia).^[ [187]^53 ^]
   Some organizations have devoted efforts to developing SIRPα fusion
   proteins, such as TTI‐621 and TTI‐622, that bind only to CD47 on tumor
   cells but not to CD47 on the surface of erythrocytes.^[ [188]^54 ^]
   However, it is difficult to develop tumor cell‐specific fusion proteins
   because of the heterogeneity of tumor cells. Researchers have developed
   bispecific antibodies to CD47 and tumor‐specific antigens to reduce
   CD47 target‐associated toxicity by binding to other specific targets on
   the surface of tumor cells while avoiding binding to normal cell, such
   as HX009, which is a dual targeting agent that targets PD‐1 and CD47.^[
   [189]^55 ^] This method can effectively improve the targeting of
   antibody drugs, but antibody drugs are easily cleared in vivo,
   hindering in situ tumor delivery. To address this problem, we used a
   dual‐targeting nanodelivery system to deliver an anti‐CD47 drug and a
   senescence inducer. Nanotechnology was utilized to coload lipid‐soluble
   and water‐soluble drugs to minimize systemic toxicity, prolong the drug
   circulation time, and increase drug accumulation within the tumor.^[
   [190]^56 , [191]^57 ^] Calcium‐phosphorus NPs are pH‐responsive,^[
   [192]^58 ^] and our results show that this delivery system hardly
   targets erythrocytes. The specific mechanism will be investigated in
   our follow‐up studies.

   Our research revealed that the application of NPs for treating HCC
   precipitates an increase in the number of M1 macrophages among TILs and
   in the adjacent normal liver tissue. M1 macrophages, which are
   conventionally linked to anti‐neoplastic responses, secrete cytokines
   that facilitate the destruction of tumor cells and bolster additional
   immune mechanisms.^[ [193]^59 ^] The elevated numbers of M1 macrophages
   in TILs signifies an increased immune reaction against tumors.^[
   [194]^60 ^] NPs possess the capacity to directly incite the maturation
   or activation of macrophages into the M1 phenotype, which is noted for
   its antitumor properties. This effect potentially transcends the tumor
   microenvironment, as evidenced by the increase in M1 macrophages in
   normal liver tissue, indicating that NP therapy may initiate a systemic
   immune response extending beyond the localized response restricted to
   the tumor site. Such systemic activation is advantageous for addressing
   metastatic cancer cells but concurrently raises concerns regarding
   potential inflammatory responses or harm to noncancerous tissues.
   Furthermore, given the pivotal role of the liver in metabolic and
   detoxification processes, activating immune responses in normal liver
   tissue warrants cautious evaluation. The possibility of hepatotoxicity
   or compromised liver functionality, particularly with prolonged
   nanomedicine treatment or higher dosages, is a relevant risk. Although
   the upregulation of M1 macrophages through nanomedicinal intervention
   shows promise in HCC therapy, it also introduces critical
   considerations regarding systemic immune responses and potential
   adverse effects. A more comprehensive understanding of its mechanism of
   action is needed for the development of safer and more efficacious
   nanomedicine‐based treatments in future endeavors.

   In summary, CD47 and CDC7 were found to be highly expressed in patients
   with HCC, and the clinical data showed that patients with high
   expression of CD47 and CDC7 had the worst clinical prognosis,
   suggesting that simultaneously inhibiting CD47 and CDC7 is a promising
   treatment for HCC. We constructed AX@NPs‐FA/AbCD47, which are
   dual‐targeting (AbCD47 and FA) NPs that deliver both a CD47 inhibitor
   and a senescence inducer (a CDC7 inhibitor) to tumor sites. They
   sequentially release drugs to synergistically augment cellular
   senescence and immunotherapy and provide novel insight for the clinical
   treatment of HCC patients.

4. Experimental Section

Cell Culture and Establish the Drug‐Resistant Cells

   The human HCC cells (HUH7, HEP3B, Bel‐7402), human normal liver cells
   (LO2), and mouse HCC cells (Hepa1‐6) were purchased from the Cell
   Institute of Chinese Academy of Sciences, cultured in DMEM medium
   containing 10% serum and 1% double antibodies, and placed in a
   humidified 37 °C and 5% CO2 standing‐temperature incubator.
   Drug‐resistant Bel‐7402 cells were cultured by incubating them with
   ascending concentrations of 5‐fluorouracil (Selleck, USA) from 1 to 12
   µm for up to 2 months. The drug‐resistant cells were named Bel‐7402R
   and the parental cells were named Bel‐7402P.

RNA Sequencing

   Shanghai Personalbio Company was commissioned to perform Illumina
   Novaseq PE150 sequencing of Bel‐7402R and the parental cell line
   Bel‐7402P. These results were analyzed for Gene Set Enrichment Analysis
   (GSEA) pathway enrichment analysis, differentially expressed genes, and
   Kyoto Encyclopedia of Genes, and Genomes (KEGG) pathway analysis.

Preparation and Characterization of Nanoparticles

   DSPE‐PEG2000‐NH[2], DSPE‐PEG2000‐Cy3, and DSPE‐PEG2000‐FA were
   purchased from Ponsure Biotechnology (Shanghai, China).
   N‐(3‐Dimethylaminopropyl)‐N′‐ethylcarbodiimide hydrochloride (EDC), and
   N‐hydroxysuccinimide (NHS) were purchased from Sigma–Aldrich (St.
   Louis, USA). AZD8055 and XL413 were obtained from Selleck (Shanghai,
   China), and other relative reagents were obtained from Sigma–Aldrich
   (USA). AZD8055 and XL413 were encapsulated with DSPE‐PEG2000‐NH[2] or
   DSPE‐PEG2000‐FA by thin film hydration; where AZD8055 was encapsulated
   on the inner side, while XL413 was adsorbed on the nanomaterial
   surface. Subsequently, CaCl[2] solution and HBS (Hepes, Na[3]PO[4], and
   NaCl; pH 7.4) buffer were added to adsorb calcium and phosphorus onto
   the nanoparticles. AX@NPs or AX@NPs‐FA was formed by standing for
   30 min at room temperature. The surface of the NPs was then modified by
   the standard EDC/NHS‐mediated amide reaction with ultra‐LEAF purified
   anti‐mouse or anti‐human CD47 blocking antibody (AbCD47) according to
   previous reports and the previous protocols.^[ [195]^61 , [196]^62 ^]
   Briefly, AbCD47 (100 µg mL^−1), EDC (0.1 m), and NHS (0.1 m) were added
   to NPs‐FA solution (5 mg mL^−1) and gently stirred for 2 h. The
   resulting NPs‐FA/AbCD47 solution was centrifuged at 15 000 g for 15 min
   and the supernatant was discarded to remove unbound antibody. The
   NPs‐FA/AbCD47 solution was stored at 4 °C. Fourier Transform Infrared
   Spectrometer (FTIR) (NEXUS 670, Nicolet, USA) detection of the
   conjugation of AbCD47 and NPs‐FA; The 12.5% SDS‐PAGE gel (Epizyme,
   China) was prepared, and 4 ug of free AbCD47 and the same dose of
   NPs‐FA/AbCD47 were added for electrophoresis, stained with coomassie
   blue fast staining (Epizyme, China) for 15 min, and subsequently
   photographed for analysis the conjugation of AbCD47 and NPs‐FA;
   Zetasizer IV analyzer (Malvern Zetasizer Nano ZS90, Malvern, U.K.) to
   detect the particle size and potential of nanoparticles; Transmission
   electron microscope (TEM) (FEI Tecnai G2 Spirit TEM, USA) to examine
   the morphology of nanoparticles; Shapes of NPs‐FA/AbCD47 observed by
   atomic force microscopy (AFM) (Hatachi, Japan); Calcium and phosphorus
   elemental analysis of NPs was performed using field emission TEM (Talos
   F200X, USA). The high‐performance liquid chromatography (HPLC) (Agilent
   1100, USA) was used to measure the encapsulation efficiency (EE%) of
   AZD8055 and XL413 in AX@NPs‐FA/AbCD47. To carry out some other
   experiments, Dir (Biotium, USA) or rhodamine B (RB) (Beyotime, China)
   were applied to prepare Dir@NPs and Dir@NPs‐FA and Dir@NPs‐FA/AbCD47 or
   RB@NPs and RB@NPs‐FA and RB@NPs‐FA/AbCD47, respectively.

Erythrocyte Targeting Assay

   Hepa1‐6 were inoculated into 1.5 cm glass‐bottomed culture dishes
   overnight at 5000 cells well^−1. Next day, C57BL/6 mice were subjected
   to orbital blood sampling to extract erythrocytes. 1 × 10^5 mouse
   erythrocytes were added to the dishes inoculated with Hepa1‐6 in
   advance. RB@NPs‐FA/AbCD47 were added and incubated for 4 h. Photographs
   were taken with a laser confocal microscope (Olympus FV1000, Japan).

In Vitro AZD8055 and XL413 Release from Nanoparticles

   A dialysis assay was applied to confirm the release profiles of AZD8055
   and XL413 from the nanoparticles in vitro. The nanoparticles were
   dialyzed against a PBS solution containing 10 m sodium salicylate,
   adjusted to either pH 6.0 or pH 7.4. At predetermined time points,
   100 µL of the release solution was collected for subsequent analysis,
   and an equal volume of fresh dialysate was added. The analysis was
   performed with an Agilent 2000 HPLC instrument equipped with a C18
   column. The mobile phase comprised potassium dihydrogen phosphate
   (KH[2]PO[4]) at pH 4.6, acetonitrile, and methanol. The flow rate was
   maintained at 1 mL min^−1, and the injection volume was set at 10 µL.
   UV detection was performed at a wavelength of 254 nm.

Cytotoxicity Assay

   Cells were inoculated into 96‐well plates at 3000 cells/well and left
   overnight, then treated with drug and incubated for 72 h. The medium
   was aspirated off, medium containing CCK8 (Beyotime, China) was added,
   incubated for 1 h, and the OD value at 450 nm was detected with an
   enzyme marker (Thermo MultisKan FC, USA).

Western Blotting Assay

   Cells were seeded in 6‐well plates at a density of 3 × 10^5 cells per
   well and incubated overnight. Subsequently, the cells were treated with
   drugs for 48 h. Following treatment, the culture medium was discarded,
   and the cells were washed with PBS. Cell lysis was achieved by the
   addition of RIPA buffer, supplemented with both phosphatase and
   protease inhibitors (both from Beyotime, China). The cell lysates were
   then collected into tubes to facilitate protein extraction and the
   measurement of protein concentrations (Beyotime, China). Protein
   electrophoresis was conducted using 12% SDS polyacrylamide gel
   electrophoresis (PAGE) gels (Beyotime, China), after which the
   separated proteins were transferred onto polyvinylidene difluoride
   (PVDF) membranes. The membranes were subsequently blocked with 10%
   (w/v) nonfat milk powder for 1 h. Overnight incubation at 4 °C with
   primary antibodies was followed by a 1 h incubation with appropriate
   horseradish peroxidase‐conjugated secondary antibodies (mouse or
   rabbit, from Proteintech, USA) the following day. Detection was
   performed using an enhanced chemiluminescence (ECL) solution from Sigma
   (USA). Antibodies targeting CD47 and GAPDH were acquired from Abcam
   (USA).

Drug Synergy Experiment

   Cells were inoculated into 96‐well plates at 3000 cells well^−1 and
   left overnight. The next day, the drug was added for 72 h. The OD
   values were measured according to the cytotoxicity assay described
   above and the IC50 values were calculated using GraphPad Prism 8.0. The
   Chou‐Talalay method (CI <1 indicates synergism) was applied to detect
   the synergism through using the CompuSyn software (Ver.1.0).

Apoptosis Assay

   Cells were inoculated into 6‐well plates at 3×10^5 cells well^−1 left
   overnight, and treatment with drugs for 48 h the next day. Cells were
   collected and stained with Annexin V‐PI apoptosis detection kit
   (Vazyme, China) and assayed using flow cytometry (Becton Dickinson,
   USA). Data were analyzed and processed with FlowJo 10. Each experiment
   was repeated three times.

SA‐β‐Gal Staining

   SA‐β‐Gal staining was performed in 6 wells (for in vitro studies) or on
   10 µm thick frozen sections of tumor tissue (for in vivo studies) using
   the β‐Galactosidase staining kit (Beyotime, China), according to the
   manufacturer's instructions.

LDH Detection Experiment

   Cells were inoculated into 24‐well plates at 5 × 10^5 cells well^−1
   overnight and dosed for 48 h. The plates were centrifuged at 400 g for
   5 min using a multi‐well plate centrifuge. 120 µL of supernatant from
   each well was carefully aspirated into a 96‐well plate and subsequently
   assayed according to the instructions of the Lactate Dehydrogenase
   Cytotoxicity Assay Kit (Beyotime, China). The measured data were
   analyzed and processed with GraphPad Prism 8.0. Each experiment was
   repeated three times.

ROS Detection Experiment

   Cells were inoculated into 24‐well plates at 5 × 10^5 cells well^−1 and
   left overnight, and drugged for 48 h. Staining was performed according
   to the instructions of the reactive oxygen species detection kit
   (Beyotime, China), and subsequent photographic analysis was performed
   with a fluorescence microscope (Olympus Corporation, Japan).

In Vitro Phagocytosis Assay

   The assay was performed as described previously.^[ [197]^63 ^] THP‐1
   cells were inoculated into 12‐well plates at 1 × 10^5 cells well^−1,
   while Phorbol‐12‐myristate‐13‐acetate (PMA) (Sigma, USA) was added for
   48 h to induce THP‐1 into walled macrophages. After macrophages were
   treated with serum‐free culture starvation for 2 h, 1 × 10^5
   hepatocellular carcinoma cells (HEP3B and HUH7) stained in advance with
   Calcium‐AM were added and then co‐cultured in complete medium
   containing drugs for 2 h. The medium was aspirated and washed three
   times with PBS to remove free cells. Photographs were taken with a
   fluorescent microscope (Olympus Corporation, Japan). Employing the
   phagocytic index—a metric denoting the number of HCC cells ingested by
   macrophages for every 100 macrophages—the phagocytosis of HCC cells
   were quantified by macrophages following a variety of drug treatments.

Cellular Uptake Assay

   Cells were inoculated into 6‐well plates overnight at 1×10^5
   cells well^−1 and added free RB (Beyotime, China) or RB@NPs or
   RB@NPs‐FA or RB@NPs‐FA/AbCD47 for 1 and 4 h, respectively, and then two
   experiments were performed as follows: 1) fixed with 4%
   paraformaldehyde and stained with DAPI (Beyotime, China) for 10 min and
   photographed with a laser confocal microscope (Zeiss LSM510, Germany);
   2) live cells were collected and assayed using flow cytometry (Becton
   Dickinson, USA), and the data were analyzed and processed with FlowJo
   10. Each experiment was repeated three times.

Lysosomal Escape Assay

   Cells were inoculated into 1.5 cm glass bottom dishes overnight at 1 ×
   10^5 cells well^−1, and RB@NPs‐FA/AbCD47 was added for 1 and 4 h,
   respectively. The green LysoTracker probe (Beyotime, China) was added
   for 40 min. Photographs were taken with a laser confocal microscope
   (Olympus FV1000, Japan).

In Vitro Mitochondrial Targeting Assay

   Cells were inoculated into 1.5 cm glass bottom dishes overnight at
   1×10^5 cells well^−1, and NPs‐Cy3‐FA/AbCD47 was added for 1 h and 4 h,
   respectively, after which the medium was replaced with fresh medium.
   Then Mito‐Tracker Green (Beyotime, China) was added to stain for 30 min
   and Hoechst 33 342 was added for 10 min. Photographs were taken with a
   laser confocal microscope (Olympus FV1000, Japan).

Immunofluorescence Assay

   Cells were inoculated into 6‐well plates overnight at 1 × 10^5
   cells well^−1, treated with or without drugs the next day, fixed with
   4% paraformaldehyde for 10 min, broken with Triton‐100, blocked with 5%
   BSA, incubated overnight at 4 °C with specific primary antibody,
   incubated for 3 h with the corresponding fluorescent secondary
   antibody, the nuclei were stained with DAPI (Beyotime, China), and
   photographed using a laser confocal microscope (Zeiss LSM510, Germany).
   The primary antibodies used were CD47 (abcam, USA) and CDC7 (abcam,
   USA). The fluorescent secondary antibodies used were Alexa Fluor
   488‐labeled goat anti‐rabbit IgG (Beyotime, China) and Alexa Fluor
   647‐labeled goat anti‐mouse IgG (Beyotime, China).

Animal Surgeries

   All procedures in this study were performed in accordance with the
   guidelines established by the Institutional Animal Care and Use
   Committee at Renji Hospital, School of Medicine, Shanghai Jiao Tong
   University (Shanghai, China). All surgeries were performed using
   aseptic conditions.

The Tumor Targeting Abilities of NPs In Vivo

   To encapsulate Dir to form Dir@NPs, Dir@NPs‐FA, and Dir@NPs‐FA/AbCD47
   nanomaterials were used. In advance, HUH7 tumor cells were inoculated
   over the root of the right thigh of mice at the number of 2×10^6 cells,
   and when the tumors grew to 200–300 mm^3, the mice were intravenously
   administered 100 µL of Dir@NPs or Dir@NPs‐FA or Dir@NPs‐FA/AbCD47 at a
   0.5 µg/mL Dir dose. Since the lipophilic tracer Dir had a strong
   fluorescence emission in the near‐infrared region (λex/λem:
   748/780 nm), it could be used to detect the biodistribution of NPs in
   tumor‐bearing nude mice. After injection, whole‐body fluorescence
   images were obtained at 8, 24, 48, and 72 h time points using an in
   vivo imaging system (Berthold Technologies, Germany). Mice were
   sacrificed and dissected 72 h after injection, to examine the
   distribution of NP in the major organs (spleen, heart, liver, lung, and
   kidney).

Anticancer Activity In Vivo

   Three‐week‐old BALB/c nude male mice and C57BL/6 male mice were
   obtained from the Animal Technology Co., Ltd. (Beijing, China). Hepa1‐6
   and HUH7 tumor cells were inoculated at 2 × 10^6 cell volume above the
   root of the right thigh of mice, respectively. All procedures in this
   study were performed in accordance with the guidelines established by
   the Institutional Animal Care and Use Committee at Renji Hospital,
   School of Medicine, Shanghai Jiao Tong University (Shanghai, China).
   All surgeries were performed using aseptic conditions. When the tumors
   grew to ≈900 mm^3, the mice were executed and dissected, the tumors
   were cut into 1.5 mm × 1.5 mm × 1.5 mm tubercles, the mice were
   anesthetized with an inhalation anesthesia machine (isoflurane),
   sterilized, and the abdomen was opened to expose the liver, the
   tubercles were inoculated into the liver with forceps, and the mice
   were sterilized and sutured. The mice were treated from the 7th
   postoperative day. The treatment strategies for HUH7 tumor‐bearing nude
   mice (n = 5 per group): 1) PBS, 2) NPs‐FA/AbCD47, 3) AZD8055 (5 mg
   kg^−1), 4) XL413 (25 mg kg^−1), 5) AZD8055+XL413(5 and 25 mg kg^−1), 6)
   X@NPs at a XL413‐equivalent dosage of 25 mg kg^−1, 7) A@NPs at a
   AZD8055‐equivalent dosage of 5 mg kg^−1, 8) AX@NPs at a
   AZD8055‐equivalent dosage of 5 mg kg^−1 and XL413‐equivalent dosage of
   25 mg kg^−1, 9) AX@NPs‐FA at a AZD8055‐equivalent dosage of 5 mg kg^−1
   and XL413‐equivalent dosage of 25 mg kg^−1, 10) AX@NPs‐FA/AbCD47 at a
   AZD8055‐equivalent dosage of 5 mg kg^−1 and XL413‐equivalent dosage of
   25 mg kg^−1; The treatment strategies for Hepa1‐6 tumor‐bearing C57BL/6
   mice (n = 5 per group): 1) PBS, 2) NPs‐FA/AbCD47, 3) AZD8055+XL413(5
   and 25 mg kg^−1), 4) X@NPs at a XL413‐equivalent dosage of 25 mg kg^−1,
   5) A@NPs at a AZD8055‐equivalent dosage of 5 mg kg^−1, 6) AX@NPs at a
   AZD8055‐equivalent dosage of 5 mg kg^−1 and XL413‐equivalent dosage of
   25 mg kg^−1, 7) AX@NPs‐FA at a AZD8055‐equivalent dosage of 5 mg kg^−1
   and XL413‐equivalent dosage of 25 mg kg^−1, 8) AX@NPs‐FA/AbCD47 at a
   AZD8055‐equivalent dosage of 5 mg kg^−1 and XL413‐equivalent dosage of
   25 mg kg^−1. The drug was administered by tail vein injection, twice a
   week for a total of 3 weeks, and the body weight of the mice was
   recorded at each treatment. Tumor growth in nude mice was recorded
   weekly by small animal ultrasound. The tumor growth inhibition (TGI)
   value was calculated as follows: TGI = 100% × (1 – treatment
   group/control group). The mice were sacrificed, and preserve the tumors
   and vital organs (heart, liver, spleen, lung, and kidney) of C57BL/6
   mice. Vital organs were fixed with 4% paraformaldehyde and stained
   using hematoxylin‐eosin (H&E) staining. The cleaned tumors were weighed
   and fixed with 4% paraformaldehyde and then sectioned.
   Immunohistochemistry (IHC) was performed using standard methods. The
   sections were stained with TUNEL, CD31, CD47, CDC7, and mTOR
   antibodies. The CD47, CDC7, and mTOR antibodies were purchased from
   Abcam (USA), TUNEL and CD31 antibodies were purchased from Beyotime
   (China). The major organs were also fixed, sectioned, and examined by
   H&E staining.

Tissue Dissociation and Flow Cytometry

   C57BL/6 mice were executed with CO[2] asphyxiation apparatus. All tumor
   tissues and liver tissues were divided into 2 parts for subsequent flow
   cytometry and histopathology preparation. Tumors and liver tissues were
   immersed in pre‐cooled PBS, cut with scissors, wash, and then added to
   DMEM containing 2.0 mg mL^−1 Collagenase IV (Sigma, USA) and 50
   units/mL DNase I (Roche, Basel, Switzerland). The cells were digested
   for 1 h at 37 °C and filtered through 70‐µm cell filter (Biosharp,
   China) to make single‐cell suspensions. Tumor cells were added to
   percoll (GE, USA) to extract tumor‐infiltrating lymphocytes (TILs). RBC
   Lysis Buffer (BioLegend, USA) was added to the TILs and hepatocytes for
   10 min at room temperature, centrifuged and resuspended. Subsequent
   staining with FVS780 (BD, USA) at room temperature for 15 min to
   differentiate between dead and live cells; Cells were then blocked with
   an Fc Receptor Binding Inhibitor (eBioscience, USA) for 10 min at room
   temperature. Staining the cells with fluorescently labeled primary
   monoclonal antibodies for 30 min. The following antibodies were used
   for analysis of TILs and liver macrophages: anti‐CD11b‐FITC,
   anti‐F4/80‐Brilliant Violet 421, anti‐CD86‐Brilliant Violet 605,
   anti‐CD206‐Brilliant Violet 650, anti‐CD45‐PE, Fixable Viability Stain
   780 (BD, USA) stain was added to exclude dead cells; The following
   antibodies were used for analysis of TILs DC cells: anti‐CD11c‐APC,
   anti‐CD86‐Brilliant Violet 605, anti‐CD80‐FITC, anti‐CD45‐PE, Fixable
   Viability Stain 780 (BD, USA) stain was added to exclude dead cells;
   The following antibodies were used for analysis of TILs T cells:
   anti‐CD3‐Brilliant Violet 421, anti‐CD4‐Brilliant Violet 605,
   anti‐CD8‐Brilliant Violet 650, anti‐CD45‐PE, Fixable Viability Stain
   780 (BD, USA) stain was added to exclude dead cells. Acquisition was
   performed on the flow cytometry (Becton Dickinson, USA). Data analysis
   was performed using FlowJo Version 10.8.1. TILs and liver
   M1‐macrophages were identified as CD45+CD11b+F4/80+CD86+; TILs DC cells
   were identified as CD45+CD11c+CD80+CD86+; TILs CD4+T cells were
   identified as CD45+CD3+CD4+; TILs CD8+T cells were identified as
   CD45+CD3+CD8+.

In Vivo Drug Release and Mitochondrial Targeting Assay

   The Hepa1‐6 cells were injected into the above the root of the right
   thigh of three‐week‐old BALB/c nude male mice at 2 × 10^6, and when the
   tumors grew to 800 mm^3, the tail vein was administered with the
   nanodrug Calcein/RB@NPs‐FA/AbCD47 which calcein was adsorbed on the
   outer side of nanoparticles and rhodamine (RB) was wrapped on the inner
   side, and the animals were photographed at 0.5 and 6 h using a
   chemiluminescent fluorescence image analysis system (PerkinElmer, USA).
   Then, the mice were sacrificed, and preserve the tumors. Tumor tissues
   were prepared as electron microscopy samples and photographed using
   TEM.

Drug‐Resistant Liver Cancer Treatment in Mice and Photoacoustic Imaging

   Bel‐7402R cells (2 × 10^6) were inoculated above the root of the right
   thigh of three‐week‐old BALB/c nude male mice, and the therapy was
   started when they grew to ≈80 mm^3.The treatment strategies for
   Bel‐7402R tumor‐bearing nude mice (n = 5 per group): (1) PBS, (2)
   AX@NPs‐FA/AbCD47 at a AZD8055‐equivalent dosage of 5 mg kg^−1 and
   XL413‐equivalent dosage of 25 mg kg^−1. The drug was administered by
   tail vein injection, twice a week for a total of 4 weeks. Photoacoustic
   imaging of small animals (Vevo, Japan) was used to record the
   intratumoral blood oxygenation of the animals and measure the values.
   Then, the mice were sacrificed, and preserve the tumors. The cleaned
   tumors were weighed and fixed with 4% paraformaldehyde.
   Immunohistochemistry (IHC) was performed using standard methods. The
   sections were stained with CD47, CDC7, mTOR, and H2AX antibodies. The
   CD47, CDC7, and mTOR antibodies were purchased from Abcam (USA), and
   H2AX antibody was purchased from Sigma (USA).

TCGA Database Analysis

   The gene expression of CD47, CDC7, and FOLR1 were downloaded from TCGA
   database and the corresponding information of liver cancer patients,
   divided the liver cancer patients into two groups with high and low
   expression according to the expression of each gene, and analyzed the
   survival time of the two groups with GraphPad Prism 8.0 software.

TIMER Database Analysis

   The correlation between CD47 and CDC7 in HCC was analyzed by running
   the TIMER database ([198]http://timer.comp‐genomics.org/).

Molecular Docking

   The PDB file of the CD47 crystal structure (PDB code: 2VSC) was
   downloaded from Protein Bank ([199]www.pdb.org). The 3D structure of
   AZD8055 was modeled by Schrodinger, and AZD8055 was docked to CD47 via
   using the standard precision scoring function of Glide 5.5 software
   (Schrödinger, USA). The heterozygous site formed by residues Lys39 and
   Tyr37 was finally selected to define and generate the receptor grid.

Statistical Analysis

   GraphPad Prism 8.0 software was used in this study to perform
   statistical analysis of the experimental data. The analysis of data
   across two distinct groups was performed utilizing the Student's
   t‐test, when appropriate for the dataset. For comparisons involving
   multiple groups, analysis of variance (ANOVA) was used to determine the
   statistical significance of differences. Survival analyses were
   conducted using the Kaplan‐Meier method, with subsequent comparisons
   made through the log‐rank test. The data were represented as mean ±
   standard errors of the mean (SEM). Statistical significance was
   established at a p‐value threshold of ≤0.05. The symbols ^* p ≤0.05,
   ^** p ≤0.01, and ^*** p ≤0.001 indicate levels of significance, whereas
   a p‐value ≥ 0.05 was considered to indicate that the difference was not
   significant, denoted as ns. Unless otherwise stated, each experiment
   was repeated at least three times.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   K.G., J.Y.J., and Z.H.W. contribute equally to this work. K.G., J.Y.J.,
   Z.H.W., Q.W., J.H.L., Y.D., and J.T.L. conducted experiments. J.Y.J.
   conducted in situ implant surgery. K.G., J.Y.J., and Z.H.W. analyzed
   the data. J.Y. and Y.S. provided materials and assistance. K.G., Y.Z.,
   and Y.R.D wrote the paper, with input from all authors. Y.R.D
   supervised the study.

Supporting information

   Supporting Information
   [200]ADVS-11-2308310-s001.pdf^ (2.1MB, pdf)

Acknowledgements