Intravascular elimination of circulating tumor cells and cascaded embolization with multifunctional 3D tubular scaffolds†

Yijing Chen‡ ^a, Cuiwen Li‡ ^a, Jinghui Yang‡ ^b, Ming Wang ^d, Yike Wang ^a, Shibo Cheng ^a, Weihua Huang ^ac, Guohua Yuan *^b and Min Xie *^a
aCollege of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, China. E-mail: mxie@whu.edu.cn
^bSchool and Hospital of Stomatology, Wuhan University, Wuhan 430079, China
^cDepartment of Hepatobiliary and Pancreatic Surgery, Zhongnan Hospital of Wuhan University, Wuhan 430071, China
^dDepartment of Clinical Laboratory, Renmin Hospital of Wuhan University, Wuhan 430060, China

First published on 1st August 2024

---

---

1. Introduction

Metastasis, a multi-step biological process, is the main cause of cancer-related deaths in clinical settings.^1,2 In cancer, the primary tumor resembles the “root” of a plant and acts as the source of circulating tumor cells (CTCs), which migrate from the primary tumor and enter systemic blood circulation.^3,4 According to the “seed and soil” hypothesis, CTCs serve as metastasizing “seeds” that have specific affinity for certain organs and promote secondary growth, leading to metastasis.^5,6 Therefore, inhibiting the growth of “roots” and eliminating the spreading of “seeds” are essential for curing cancer.

The reasonable and efficient treatment of the primary tumor is critical in cancer therapy. Surgical resection, radiotherapy, and chemotherapy are the main clinical options; however, the side effects of these methods in patients cannot be overlooked.^7–9 In contrast, embolization is a localized approach for treating cancer by blocking the blood vessels to prevent the supply of nutrients and oxygen to the tumor. Embolization has few side effects and good anti-tumor efficacy.¹⁰ Clinically, several kinds of embolic agents including coils,¹¹ particulates,¹² gelfoam,¹³ and liquid embolic agents¹⁴ have been widely used for embolization and are suitable for the treatment of diseased or injured vasculature. However, incomplete embolization or ectopic embolization may occur occasionally, lowering the embolic efficacy.¹⁵ While novel preclinical embolic agents have been developed with improved local precision and a lower chance of recanalization for tumor embolization, it is difficult to achieve high antitumor efficacy with these agents.¹⁶ Since CTCs often flow back to the primary tumor due to the “homing effect”,^17–19 none of the existing embolization strategies can eliminate homing CTCs and embolize the primary tumor simultaneously.

The in vivo elimination of CTCs has been suggested to inhibit the metastasis of cancer.^20,21 However, the efficient removal of rare CTCs from the blood circulation system remains a significant challenge. Functionalized nanomaterials such as magnetic nanoparticles,²² anti-cancer drug-loaded liposomes,²³ micelles,^24,25 and exosomes²⁶ have been used for the specific recognition or damage of CTCs in the blood circulation system. In addition, biomedical devices (e.g., metal needles,^27,28 intravenous catheters,²⁹ and vascular scaffolds³⁰) can be implanted into the blood vessels for the in vivo capture of CTCs, which can be subsequently damaged by hyperthermia^29,30 or irreversible electroporation.³¹ Furthermore, clinically extracorporeal circuits may be employed to process liters of blood and remove CTCs.³² While these strategies are suitable for the in vivo elimination of CTCs, their potential biosafety concerns, capacity, and efficiency in CTC elimination must be considered. Therefore, innovative solutions are needed to develop more efficient strategies for eliminating CTCs.

Here, we present a cascaded embolization technique based on the implantation of a multifunctional 3D tubular scaffold in blood vessels to inhibit tumor growth and metastasis. A 3D tubular scaffold with interconnected macropores, excellent photothermal ability, and bioaffinity was prepared by embedding a gold nanotube-coated 3D scaffold (Au NT/3D scaffold) into a size-tunable poly(dimethyl siloxane) (PDMS) tube that can be easily injected and located in the blood vessels. Injecting the 3D tubular scaffold into the blood vessels upstream of the primary tumor effectively intercepted the homing CTCs and recruited platelets and immune cells^33,34 to form “micro-thrombi” for primary tumor embolization. When implanted downstream of the primary tumor, the 3D tubular scaffold captured the CTCs shed from the primary tumor in real time and was then irradiated by a laser to kill the CTCs via the photothermal effect and inhibit tumor metastasis. In vivo experiments in rabbits with auricular tumors demonstrated that the 3D tubular scaffold captured 86% of the CTCs, embolized the primary tumor to inhibit 99% of tumor growth and damaged CTCs with nearly 100% efficiency, suggesting that the scaffold might synergistically inhibit tumor metastasis.

2. Materials and methods

2.1. Fabrication of 3D tubular scaffold

A 3D scaffold was first prepared by a sacrificial template method.³⁵ Briefly, a piece of a Ni foam slice (30 mm × 7 mm × 1 mm) was sonicated in acetone, ethanol, and water for 30 min, respectively, and then dried. Subsequently, the clean Ni foam was immersed in a 2 mL centrifuge tube containing freshly prepared PDMS and centrifuged at 8000 rpm for 5 min. After that, the PDMS-stuffed Ni foam was transferred into an empty 2 mL centrifuge tube and centrifuged at 3000 rpm for 3 min. Through the two runs of centrifugation, the Ni foam coated with PDMS was obtained and then heated to 80 °C for 3 h to solidify the PDMS, which was then immersed in 7 M HNO₃ for 2 h to completely etch the Ni foam. Thus, the PDMS scaffold was prepared, then washed with ultrapure water and dried. Then, the 3D scaffold was treated with oxygen plasma for 5 min and immediately immersed into 1 mg mL⁻¹ dopamine solution (Tris–HCl buffer solution, 10 mM, pH ∼ 8.5) for 24 h. After washing with ultrapure water, the 3D scaffold was immersed in Ag NW solution. Next, the 3D scaffold was dried naturally to obtain an Ag NW/3D scaffold, which was subsequently immersed into 2 mmol L⁻¹ [Au(en)₂]Cl₃ solution at 75 °C.³⁶ Then the reacted scaffold was transferred into 0.15 M ammonium hydroxide for 30 min at 75 °C to remove the AgCl precipitate. At last, the Au NT/3D scaffold was washed with ultrapure water and dried. To protect the 3D structure, the scaffold was completely immersed in 20% gelatin solution and then cooled down at room temperature. The gelatin scaffold was cut into a certain size and plugged into a glass capillary with a certain diameter (1.0 mm or 2.0 mm). The uncrosslinked PDMS was then centrifuged (3000 rpm, 3 min) to fill into the glass capillary. After solidification at room temperature, the glass capillary was first peeled off, and the product was then immersed in 37 °C water for 2 h to remove the gelatin inside the scaffold. Finally, the 3D tubular scaffold was successfully prepared.

2.2. Functionalization of the 3D tubular scaffold and the Au NT/3D scaffold chip

The 3D tubular scaffold was immersed into 2-mercaptoacetic acid (TGA) solution (60 nmol L⁻¹) for 24 h to couple carboxyl groups. After washing with ultrapure water, the 3D tubular scaffold was activated with 10 mmol L⁻¹ EDC and 20 mmol L⁻¹ NHS for 1 h. Then, the 3D tubular scaffold was incubated with 100 μg mL⁻¹ SA for 4 h at room temperature and washed with PBS. At last, the 3D tubular scaffold was incubated with 10 μg mL⁻¹ anti-EpCAM antibody for 1 h. Besides, DyLight 488 conjugated goat anti-mouse IgG was incubated with the 3D tubular scaffold for 1 h to verify the successful modification of the anti-EpCAM antibody on the 3D tubular scaffold. The functionalization steps of the Au NT/3D scaffold chip were performed accordingly.

2.3. Hemocompatibility of 3D tubular scaffolds

Citrated human whole blood was used for all blood-related testing. The clotting test was performed according to the previously developed protocols by Gaharwar et al.³⁷ 900 μL citrated blood was mixed with 100 μL 0.1 M CaCl₂ and then 50 μL of the mixture was transferred to the wells of a 96-well plate where samples (tubular PDMS, the Au NT/3D scaffold, and the 3D tubular scaffold) in the same size were deposited at the bottom in advance. At selected time points (3, 5, 10, 15, and 20 min), each well was quickly washed with 0.9% (w/v) saline solution to halt the blood clotting and remove all soluble blood components. The blank wells with only citrated blood served as the control. Each group had three parallel experiments. Hemolysis testing was performed according to the protocols of Averyet et al.³⁸ Citrated whole blood was diluted 50× into 0.9% (w/v) saline solutions. Then, equal volumes of diluted blood and tubular PDMS, Au NT/3D scaffold, 3D tubular scaffold, saline (negative control), and DI water (positive control) were incubated at 37 °C for 2 h in a constant temperature oscillator (100 rpm, THZ-C Instrument). Samples were centrifuged (1500 rpm, 10 min), and the supernatant was transferred into the centrifuge tubes. The absorbance data were measured using a UV 2550 UV-vis spectrophotometer (Shimadzu, Japan). Percent hemolysis was calculated according to the following equation: hemolysis (%) = (A_sample − A_neg)/A_pos × 100%, where A_sample is the absorbance of samples at 545 nm, A_neg is the absorbance of the saline diluted blood, and A_pos is the absorbance of the DI water diluted blood. Each group had three parallel experiments.

2.4. Biocompatibility of 3D tubular scaffolds

Three white female New Zealand rabbits, weighing 2.0–2.5 kg, were used for the in vivo biocompatibility test. The rabbits were anesthetized with intraperitoneal injection of ketamine hydrochloride (30 mg kg⁻¹), then the 3D tubular scaffold was implanted into rabbit auricular central artery assisted by an indwelling needle which was composed of an indwelling drain and a stainless-steel needle. Briefly, the indwelling needle was implanted into the rabbit ear artery, and the stainless-steel needle was extracted. Then, the as-prepared 3D tubular scaffold (length, 1.0 cm; diameter, 1.0 mm) was placed inside indwelling drain with tweezers, and slowly pushed into the blood vessel by a flat stainless-steel needle until it completely entered the blood vessel. Finally, the indwelling needle was removed. The 3D tubular scaffold was indwelled in the auricular central artery for 28 days. During this period, blood was sampled from each rabbit once a week and used for blood routine examination and standard serum biochemical analysis.

After 4 weeks of implanting the 3D tubular scaffold in the rabbit auricular central artery, one rabbit was anesthetized to evaluate the X-ray imaging capability of the materials in vivo. Then, other rabbits of the implantation and control group were sacrificed, and their tissues of heart, liver, spleen, lungs, kidneys, and ears were obtained. The ears were obtained and fixed with 4% paraformaldehyde, then embedded and sliced for H&E, TUNEL and PCNA staining. Other organs were cleaned and lyophilized (Scientz-10N, China) to a constant weight and ground to powder. A sample of 0.1 g was digested by microwave, and then the contents of Au in the solution were analysed by ICP-MS (Agilent 8900, American). Each group was subject to three parallel experiments.

All the animal experiments were carried out strictly according to protocols (No. WP20210492) approved by the Institutional Animal Care and Use Committee (IACUC) of the Animal Experiment Centre of Wuhan University (Wuhan, China).

2.5. In vitro capture of CTCs from the mimicked circulating system

A closed loop circulating system was constructed to mimic the in vivo capture of CTCs. The system was composed of a peristaltic pump (LSP04, Longer, Baoding, China) and a rubber tube (length, 45 cm; inside diameter, 0.8 mm/1.6 mm). Before experiment, the rubber tube was preblocked with 5% BSA for 30 min. Then, 200 HepG2 cells stained with DiI were chosen as model CTCs and spiked into the mimicked circulating system, and the antibody-functionalized 3D tubular scaffold (2.0 mm in diameter) was delivered into the rubber tube (1.6 mm in diameter) to capture CTCs at a flow rate of 7 cm s⁻¹ for one circle, 0.5 h, 1 h, and 2 h, respectively. Meanwhile, 3D tubular scaffolds with diameters of 1.0 mm (with small macropores) and 2.0 mm (with big macropores) were, respectively, delivered into the rubber tube with internal diameters of 0.8 mm and 1.6 mm and sustained for 30 min to investigate their capture efficiency. After that, the 3D tubular scaffold was pulled out and observed under a fluorescence microscope to calculate capture efficiency. Besides, 100 tumor cells were spiked into PBS or whole blood to investigate the influence of lengths (0.5, 1.0, and 2.0 cm) of the 3D tubular scaffolds (2.0 mm in diameter) on cell capture capacity in the mimicked circulating system with a flow rate of 7 cm s⁻¹ for 30 min duration. Cell capture efficiency was calculated according to the following equation: cell capture efficiency (%) = the number of cells captured by a 3D tubular scaffold/the number of spiked cells in the circulation × 100%.

2.6. In vitro embolization from the mimicked circulating system

We injected a certain volume of whole blood into the closed loop circulating system established above to mimic the in vivo embolization process. The following three groups were constructed to simulate different states in vivo: a blank group (without any treatment), group I (embolized with 3D tubular scaffold), group II (embolized with a 3D tubular scaffold + 2000 HepG2 cells). To investigate the embolization performance of different groups in the closed loop circulating system, we measured the pressure change in each group using an external digital pressure gauge (YB80A, Wotian Instrument, Taicang, China) through a three-way connect to the circulating system, and then the pressure changes of each group were recorded every 30 s. Percentage of pressure changes (ΔP) was calculated according to the equation ΔP (%) = (P − P₀)/P₀ × 100%, where P is the pressure at a certain time and P₀ is the initial pressure.

2.7. Characterization of photothermal performance of 3D tubular scaffolds

To investigate the photothermal properties of the 3D tubular scaffolds, the material with a length of 1 cm was treated with an 808 nm laser at a power density of 0.2 W cm⁻² for 3 min, and the temperature change curve was recorded. The tubular PDMS and Au NT/3D scaffold in the same size were taken as control. Under the above conditions, the performance of the 3D tubular scaffold for five cycles of repetitive temperature increasing and cooling was investigated.

The 3D tubular scaffold was used to capture 10⁵ HepG2 cells spiked in 1 mL PBS for 30 min in the closed loop circulating system and was treated with an 808 nm laser for 30 s, 1 min and 2 min, respectively. Besides, 1 mm, 3 mm and 5 mm of pork slices were used to mimic the tissue which were covered above the 3D tubular scaffold and then irradiated with 808 nm laser for 3 min, 5 min and 7 min, respectively, to investigate the photothermal sensitivity of the 3D tubular scaffold. After photothermal processing, the cells on the 3D tubular scaffold were stained with Cal-AM/PI to calculate its viability.

2.8. In vivo capture and photothermal killing of tumor cells in the rabbit auricular central artery model

The in vivo capture and photothermal performance were further tested using the New Zealand rabbit model. Here, 500, 1000, and 1500 of HepG2 cells stained with DiI were, respectively, injected into the auricular central artery implanted with a 3D tubular scaffold. After 30 min, the 3D tubular scaffold was taken out to calculate the number of captured cells using fluorescence microscopy. Besides, the HepG2 cells unstained with DiI can also be identified by triple color immunostaining. Briefly, the 3D tubular scaffold was retrieved from the vessel and washed with PBS three times, then fixed with 4% paraformaldehyde for 30 min at 4 °C and permeabilized with 0.2% Triton X-100 for 30 min at 4 °C. Subsequently, 5% BSA was incubated with it for 30 min at 4 °C to reduce unspecific binding. At last, the 3D tubular scaffold was stained with Hoechst, FITC-CK, and PE-CD45 for 2 h at 4 °C. After staining, the 3D tubular scaffold was observed under a fluorescence microscope to identify captured tumor cells. Tumor cells were classified as Hoechst+/CK+/CD45−, and WBCs were defined as Hoechst+/CK−/CD45+.

The photothermal killing of tumor cells in vivo was also performed as above. Briefly, the 3D tubular scaffold was used to capture 10⁴ HepG2 cells injected into the rabbit for 30 min and then irradiated under 808 nm NIR laser (0.2 W cm⁻², 10 min). After that, the 3D tubular scaffold was retrieved out for viability assay.

The influence of the NIR laser on the living body was also performed. The 3D tubular scaffold was implanted into the rabbit auricular central artery for 4 weeks. An 808 nm laser (0.2 W cm⁻², 10 min) was applied at the implantation site of the 3D tubular scaffold. At the same time, the ear temperature changes that reflect the blood flow rate were recorded using a FLIR thermal camera before and after the exposure under the NIR laser. The blood routine examination was performed to verify the effects of cell fragmentation on the living body. The peripheral blood samples collected before irradiation were set as normal groups. And the blood samples were collected every two days four times after irradiation. At last, the ear tissues at the implantation site or downstream the implantation site were taken. All tissues were fixed by formalin and embedded in paraffin. Then, the samples were sectioned, stained with hematoxylin and eosin (H&E) and PCNA staining, and observed under a brightfield microscope.

2.9. In vivo embolization of central arteries in rabbits

Three white female New Zealand rabbits, weighing 2.0–2.5 kg, were used to evaluate the embolization performance of the 3D tubular scaffold in vivo. The rabbits were anesthetized and then the 3D tubular scaffold was implanted into the rabbit auricular central artery through an indwelling needle as described above. Subsequently, the blood perfusion in the downstream site of the 3D tubular scaffold was monitored by laser Doppler flowmetry continuously. The ear temperature changes were recorded using a FLIR thermal camera before and after 7-day implantation to verify the effect of endovascular embolization. After monitoring, the 3D tubular scaffold was taken out for SEM imaging to observe the state of cell aggregation.

2.10. In vivo evaluation of antitumor and metastasis inhibition efficacy using the 3D tubular scaffold

The anti-tumor efficacy of the 3D tubular scaffold was assessed using auricle tumor bearing rabbits. Twelve white New Zealand rabbits (female, weight 2.0–2.5 kg) were used for building tumor-bearing models by injecting 300 μL (10⁷ cells) suspension of VX2 cancer cells into the auricles of rabbits. Around 5 days, the tumors increased to a volume of approximately 1000 mm³, which was calculated using the equation V (mm³) = A × B²/2, where A and B were the largest and smallest diameters of tumors, respectively.³⁹ Then, the tumor bearing rabbit models (total n = 12) were randomly divided into four groups: a control group (without any treatment, n = 3), an up group (the 3D tubular scaffold located in the upstream of tumor blood vessel, n = 3), a down group (the 3D tubular scaffold located in the downstream of tumor blood vessel, n = 3), and an up + down group (the 3D tubular scaffold located in both the upstream and downstream of tumor blood vessel, n = 3). The tumor size and body weight of rabbits were recorded every 2 days. After 5 days of therapy, MRI was used to detect the tumors. Besides, 808 nm NIR laser (0.2 W cm⁻², 10 min) was irradiated onto the 3D tubular scaffold for CTC damage at day 15.

For further evaluation of the tumor inhibitory effect of the 3D tubular scaffold in the control and treatment groups, 1 mL of whole blood from rabbit auricular vein was collected for in vitro CTC detection using the Au NT/3D scaffold chip. After 2 weeks, one rabbit was randomly chosen from each group for sacrifice, and the primary tumor, heart, liver, spleen, lungs, and kidneys of the rabbits were taken. All tissues were fixed by formalin and embedded in paraffin. Then, the samples were sectioned, stained with hematoxylin and eosin (H&E) and observed under a brightfield microscope.

2.11. Statistical analysis

Differences between two experimental groups were determined using a two-tailed Student's t test. Graph productions were performed using software Origin 2018. All bar graphs show means ± SD. p values were *p < 0.05, **p < 0.01, and ***p ≤ 0.001; ns p > 0.05 indicates no significant difference.

3. Results and discussion

3.1. Design, fabrication, and characterization of the 3D tubular scaffold

Developing an intravascular implantable device that can cut off the nutrient supply and intercept the CTCs shed by the primary tumor would provide a novel and highly efficient strategy for cancer therapy. Toward this end, an Au NT/3D scaffold was first prepared according to our previously reported procedure.⁴⁰ Subsequently, the Au NT/3D scaffold was filled with gelatin and placed into a glass capillary. Uncrosslinked PDMS was then centrifuged into the interspace between the gelatin-filled scaffold and the glass capillary. After the solidification of PDMS, the glass capillary was peeled off, and the gelatin inside the 3D scaffold was thermally dissolved and removed to obtain a 3D tubular scaffold with a PDMS wall and a 3D internal structure (Fig. 1a). In this way, the size and length of the tubular scaffold can be simply adjusted to adapt to different types of blood vessels by changing the scaffold or the glass capillary (Fig. 1b).

Scanning electron microscopy (SEM) characterization showed that the interior 3D structure was perfectly preserved, and the interconnected macropores (110–300 μm in diameter) facilitated the efficient interception and capture of CTCs (Fig. S1, ESI†). The thickness of the PDMS wall was controlled in the range of 100–200 μm (Fig. 1c), which provided suitable mechanical flexibility for localization of the 3D tubular scaffold inside blood vessels while protecting the interior 3D structures. Fig. 1d and e demonstrate that the Au NTs were uniformly distributed on the surface of the 3D scaffold, which was then biofunctionalized with the anti-EpCAM antibody to impart the scaffold with biospecificity for recognizing CTCs (Fig. S2, ESI†).

3.2. In vitro evaluation of the 3D tubular scaffold

The in vitro evaluation of the 3D tubular scaffold was modelled in a closed-loop circulation setup (Fig. 2a), and its hemocompatibility was first tested. Hemolysis analysis showed that the 3D tubular scaffold had low hemolysis values of less than 5%, similar to those of the control groups. Thus, the scaffold can be regarded as a nonhemolytic material according to ISO 10993-4:2002 (Fig. 2b). The clotting time of the 3D tubular scaffold was tested through a coagulation assay (Fig. 2c), and clotting times comparable to those of the control groups (15 min) were obtained. Thus, the 3D tubular scaffold neither caused hemolysis nor promoted coagulation when they came into contact with blood. These results indicate that the 3D tubular scaffold exhibits excellent hemocompatibility for intravascular application.

The cell capture performance of the scaffold was then evaluated in a mimicked circulation system containing DiI-stained HepG2 cells at a velocity of 7 cm s⁻¹ (corresponding to the linear velocity of blood in human forearm vessels^28,41). The cell capture efficiency increased with circulation time (Fig. S3a, ESI†). Considering the experimental efficiency, a circulation time of 0.5 h was selected for subsequent experiments. Statistically, the 3D tubular scaffold with smaller macropores (40–250 μm) had a higher capture efficiency than the scaffold with larger macropores (110–300 μm) due to the enhanced collision frequency between the scaffold and cells (Fig. S3b, ESI†). Additionally, the cell capture efficiency was improved when the length of the 3D tubular scaffold increased from 0.5 to 2 cm (Fig. 2d). Compared with a PBS system, the complex blood components interfered with the CTC capture ability of the 3D tubular scaffold; however, nearly 40% to 78% of the measured capture efficiencies were sufficient when extremely rare CTCs existed in the blood circulation system.^28,29 The gradually captured CTCs in the 3D tubular scaffold would occlude the flow of blood which can be estimated by the dramatic increase of the fluid pressure. Comparatively, a 3D tubular scaffold without capturing CTCs slightly increased the fluid pressure because of the nonspecific absorption of blood cells. These results indicate that the 3D tubular scaffold would specifically capture CTCs in the circulating system and consequently embolize the blood vessel (Fig. S4, ESI†).

The photothermal performance of the Au NTs was confirmed (Fig. S5a, ESI†), and the temperature of the Au NT/3D scaffold under 808 nm laser irradiation was evaluated. Under irradiation at 0.35 W cm⁻² (the maximum power density used for photothermal therapy⁴²), a high temperature of 71 °C was rapidly achieved (Fig. S5b, ESI†). This temperature would likely induce empyrosis of the organism. Therefore, a lower laser power of 0.2 W cm⁻² was employed, resulting in a sharp increase in temperature to 60 °C in a few seconds (Fig. 2e, blue line). Cycling photothermal assay of the 3D tubular scaffold showed that the photothermal effects did not attenuate after five irradiation and cooling cycles (Fig. S5c, ESI†), suggesting that the 3D tubular scaffold has good reusability and stability for long-term, repetitive applications.

The photothermal damage to the HepG2 cells on the 3D tubular scaffold is shown in Fig. 2e. Of the HepG2 cells, 92% ± 2% were killed, and the cell killing efficiency reached 98% ± 2% when the exposure time was increased to 120 s (Fig. 2e, blue bar). In particular, the photothermal temperature increased to above 46.1 °C (high enough to kill cells⁴³) when even blocked by 3 mm of pork slice for 7 min laser exposure (Fig. S6, ESI†). Quantitatively, 97%, 90%, and 83% of HepG2 cells were damaged after 7 min of laser irradiation when the 3D tubular scaffolds were covered with pork slices with thicknesses of 1-, 3-, and 5-mm, respectively (Fig. 2f), demonstrating that the 3D tubular scaffold can sense laser irradiation at a certain tissue depth. Thus, the scaffold is suitable for in vivo photothermal applications.

3.3. In vivo implantation of the 3D tubular scaffold

The 3D tubular scaffold was implanted into the auricular central artery of a New Zealand rabbit (Fig. 3a) with the aid of an indwelling needle,⁴⁴ and the dark grey tubular material was clearly observed inside the vessel (Fig. 3b). Remarkably, the 3D tubular scaffold was clearly imaged at its original location by X-ray at 28 days after implantation (Fig. 3c) due to the radiopaque property of the Au NTs and the stability of the Au NT/3D scaffold. This phenomenon is consistent with the elemental analysis results showing extremely low contents of Au in various rabbit organs at 28 days after implantation (Fig. 3d; p > 0.05).

The biological safety of the 3D tubular scaffold was evaluated during the 28-day implantation period, during which the serum from the experimental rabbits and control groups was analysed once per week. Liver function was reflected by its typical biomarkers, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), while kidney function was represented by creatinine (CREA) and urea (UREA). Fig. 3e and f demonstrate that there were no significant differences (p > 0.05) in the levels of all four biomarkers between the experimental and control rabbits. Meanwhile, routine blood examination indicated that the values of white blood cells (WBCs), red blood cells (RBCs), neutrophils (NEUTs), platelets (PLTs) and hemoglobin (HGB) (Fig. 3g; p > 0.05) along with other parameters (Fig. S7, ESI†) did not deviate significantly from those of the control. These results indicate that the long-term implantation of the 3D tubular scaffold in vivo did not damage the liver or kidney function or has an obvious toxicity (hemolysis, coagulation, and inflammatory response).

During the 28-day implantation of the 3D tubular scaffold, the tissue around the implantation site did not show obvious necrosis (Fig. S8, ESI†). Hematoxylin and eosin (H&E) staining of the implantation site of the blood vessel indicated normal ear tissue formation with a complete vascular structure (Fig. 3h). In addition, the regular round shape of a blood vessel (Fig. S9a, ESI†) with a size matching that of the 3D tubular scaffold was remarkably distinguished from the irregular shape of the control (Fig. S9b, ESI†), indicating that the blood vessel was stretched and well supported by the 3D tubular scaffold. The subsequent positive proliferating cell nuclear antigen (PCNA) staining (Fig. 3i) and negative terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labelling (TUNEL) results (Fig. 3j) demonstrated that the cells around the blood vessels were in a state of proliferation rather than apoptosis. The above results all demonstrate that the 3D tubular scaffold was successfully implanted and well encapsulated by the blood vessel to achieve long-term indwelling without inducing obvious inflammation or necrosis. These properties demonstrate the capabilities and advantages of the 3D tubular scaffold as an intravascular device for in vivo applications.

3.4. In vivo enrichment and photothermal damage of CTCs

The 3D tubular scaffold was implanted in the auricular central artery of a New Zealand rabbit and indwelled in the blood vessel for 30 min to capture HepG2 cells (model CTCs) that had been previously injected in vivo (Fig. 4a). As shown in Fig. 4b, the CTCs were successfully captured and identified as Hoechst+/PE-CD45−/FITC-CK+ cells by triple-color immunostaining. Quantitatively, when 500, 1000, and 1500 HepG2 cells were injected into the rabbit, 18 ± 4, 37 ± 14, and 48 ± 7 CTCs were captured by the 3D tubular scaffold, respectively (Fig. 4c). Currently, kinds of devices or techniques, such as indwelling needles,^28,45 intravenous catheters,²⁹ intravascular aphaeretic systems,³² flexible electronic catheters,³¹etc., are developed for in vivo enrichment of CTCs. The performance of these approaches and the 3D tubular scaffold is shown in Table S1 (ESI†). It is clear that the capture efficiency of the 3D tubular scaffold is lower than those of “ZnO nanoflower-coated indwelling needles” and “electronic catheters”. These two devices have a high surface area because of the outer coating of ZnO nanoflowers or nanofibers for highly efficient cell capture; however, the coated nanomaterials may detach from the indwelling needle or catheter during injection or the indwelling process, thus causing biosafety concerns. In our strategy, the Au nanotubes are immobilized in the inner surface of the 3D tubular scaffold, and have lower chance to detach into the circulating system because of the attenuated blood flow speed. Although the the 3D tubular scaffold with a relatively lower surface area achieves less efficient cell capture efficiency, it is still comparable to that of an “indwelling needle” and superior to that of an “intravenous catheter” and an “intravascular aphaeretic system”.

The 3D tubular scaffold was used to capture CTCs and then irradiated with an 808 nm near-infrared (NIR) laser to examine its photothermal effect in vivo. Fig. 4d shows that the temperature of the laser-irradiated auricle area was gradually increased from the basal body temperature of 25.1 °C to 50.9 °C. Cal-AM/PI co-staining was used to examine the viability of the captured cells; the quantitative results indicated that nearly 100% of the cells on the laser-irradiated 3D tubular scaffold were dead, while almost all the cells on the 3D tubular scaffold that were not exposed to laser irradiation maintained good viability (Fig. 4e). The short temporal NIR laser irradiation (10 min) did not harm the surrounding tissue (both at the in situ and downstream sites of laser irradiation, Fig. S10 and S11, ESI†), but could damage the aggregated cells inside the 3D tubular scaffold and induce the recanalization of the embolized ear artery (Fig. S12, ESI†). The laser irradiation induced cell debris did not influence the normal physiological function of the living body revealed by the blood routine examination (Fig. S13, ESI†). These results demonstrate the excellent in vivo photothermal effect of the 3D tubular scaffold, resulting in efficient cell damage and the recanalization of the embolized ear artery, which can circularly filtrate the blood and capture the CTCs by the size exclusion effect and irradiate by laser for continuous CTC elimination.

3.5. Intravascular embolization by the 3D tubular scaffold

Due to the homing effect, the CTCs flow back to the primary tumor, which is one of the key factors in tumor occurrence. Therefore, the 3D tubular scaffold could be implanted into the blood vessel upstream of the primary tumor to capture homing CTCs (Fig. 5a). The blood flow rate of the 3D tubular scaffold-implanted blood vessel was consistently monitored by laser Doppler flowmetry. As shown in Fig. 5b, downstream of the 3D tubular scaffold, the blood flow decreased sharply and consistently from day 1 to day 14 and then remained at an extremely low level until day 28. The infrared thermographs also showed that the temperature downstream of the 3D tubular scaffold was lower than the temperature upstream of the scaffold and that in the control ear (Fig. 5c), indicating that the blood flow was efficiently blocked by the scaffold. The SEM images of the 3D tubular scaffold treated with the tissue-fixation process show tightly packed cell aggregates in the macropores (Fig. 5d and Fig. S14, ESI†). These phenomena all suggest that the blood could flow through the macropores of the 3D tubular scaffold to promote the in vivo enrichment of CTCs in the early stage after implantation; subsequently, the CTCs, immune cells, and platelets persistently interacted with each other and gradually formed aggregates during long-term indwelling, causing the blood flow to be blocked for tumor embolization.

3.6. In vivo application of the 3D tubular scaffold in auricle tumor-bearing rabbits

Auricle tumor-bearing rabbit models were established with xenograft tumors larger than 1000 mm³ and then implanted with 3D tubular scaffolds located upstream (up), downstream (down), or both upstream and downstream (up + down) of the primary tumor. Fig. 6a and d show that the 3D tubular scaffold dramatically inhibited tumor growth, with tumor growth inhibition (TGI) values of 92% ± 4% (up), 85% ± 4% (down), and 93% ± 3% (up + down). In contrast, the tumor volume in the control group gradually increased from approximately 1017 mm³ on day 0 to 1417 mm³ on day 9 (Fig. S15, ESI†). Magnetic resonance imaging (MRI) confirmed that the tumors in the treatment group were smaller than those in the control group (Fig. 6b), and H&E staining (Fig. 6c) of xenograft tumors had fewer cancer cells as well, demonstrating an excellent tumor inhibition effect. Here, the 3D tubular scaffold (up) exhibited better TGI performance than the 3D tubular scaffold (down) because of the tumor embolization-like effect. Meanwhile, the 3D tubular scaffold (up + down) achieved the best TGI due to the simultaneous blocking of the tumor nutrient supply (up) and inhibition of metabolite removal (down). The above differences were more obvious when the implantation period was extended to 15 days, with TGI values of 98% ± 2% (up), 93% ± 2% (down), and 99% ± 1% (up + down). The body weights of the rabbits in each group fluctuated slightly throughout the observation period without statistically significant changes (Fig. 6e), indicating that this artificial coagulation method was suitable to inhibit tumor growth.

The performance of the 3D tubular scaffold as a CTC scavenging tool was evaluated by analysing 1 mL of whole blood from the rabbits (Fig. S16, ESI†). Quantitatively, when the 3D tubular scaffold was implanted into the blood vessels of the primary tumor, the CTCs in blood decreased visibly compared with the control, with median values of 10 (up), 3 (down), 2 (up + down), and 14 (control) (Fig. 6f). The CTCs were effectively captured by the 3D tubular scaffold (down); however, only some of the CTCs homing back to the primary tumor were captured by the 3D tubular scaffold (up), which would be improved by long-term filtrating of the blood. In contrast, the 3D tubular scaffold (up + down) encircled and scavenged CTCs to the maximum extent (86%) because of its dual effects. Actually, the CTC scavenging rate could be further improved by increasing the placement amount of the 3D tubular scaffold. NIR laser irradiation was then applied to the 3D tubular scaffold to damage the aggregated CTCs and eliminate the potential for tumor metastasis (Fig. S17, ESI†). The H&E staining images of the main rabbit organs did not show obvious lesions or tissue pathological abnormalities (Fig. S18, ESI†). These results demonstrate that the 3D tubular scaffold (up) effectively inhibited tumor growth with moderate CTC scavenging ability, while the 3D tubular scaffold (down) was less efficient in inhibiting tumor growth but efficiently scavenged CTCs to greatly decrease the potential for metastasis.

Currently used cancer treatment strategies including surgical resection, chemotherapy, and radiotherapy only target the primary tumor,⁷ and the spread of CTCs can lead to cancer metastasis and recurrence.^1,2,4 Therefore, targeting only the primary tumor or CTCs may be ineffective for cancer treatment. Accordingly, we developed an innovative treatment strategy to simultaneously eliminate spreading CTCs and embolize the primary tumor. We designed an elastic and flexible 3D tubular scaffold with a size that matches that of the blood vessel for intravascular implantation. The scaffold shows excellent biosafety for long-term indwelling in vivo, and its macroporous structure filters the blood for efficient CTC capture. Synergistically, the 3D tubular scaffold acts as an artificial coagulant to efficiently inhibit tumor growth and as an intravascular CTC scavenger that accumulates CTCs for photothermal treatment, thereby reducing the potential for metastasis.

Biocompatibility and biosafety are the main concerns for the in vivo application of devices or materials. Previously, metal needles,^27,28 intravenous catheters,^29,31 and magnetic wires²² have been inserted into blood vessels for capturing CTCs. However, these materials could pierce the blood vessel, interfere with mobility, cause infection, or induce phlebitis, limiting their long-term indwelling in vivo. In contrast, while magnetic beads and other types of nanomaterials can be easily injected into the blood for CTC capture, they may promote the appearance of new metastatic sites.⁴⁶ In comparison, the 3D tubular scaffold exhibits better biosafety because of the excellent biocompatibility of PDMS⁴⁷ and the tightly bonded Au NTs during long-term indwelling.

The working mechanism of our scaffold for the in vivo capture of CTCs is revolutionary, and their CTC elimination efficiency is improved accordingly compared to that of previously reported materials. Generally, metal needles,^27,28 intravenous catheters,^29,31 and magnetic wires²² are temporally inserted in the blood vessels and have a limited surface area and a short indwelling time to interact with CTCs. In contrast, the 3D tubular scaffold, which is suitable for long-term indwelling, can be tightly encapsulated by the blood vessel, and its microporous skeleton acts like an intravascular filter to persistently capture CTCs from the blood. Theoretically, the 3D tubular scaffold could process all the blood in the organism to filter out CTCs with a high depletion rate. For short-term application, the 3D tubular scaffold can also be developed as an auxiliary tool for surgery to intercept deciduous cancer cells generated by surgical resection.

Embolization has been widely utilized for cancer treatment because of its low risk, few side effects, and high success rate.^48–50 For instance, iodized oil-based transcatheter arterial chemoembolization (TACE)⁵¹ is a good choice for the treatment of hepatocellular carcinoma (HCC) in the middle and advanced stages. The iodized oil easily induces ectopic embolization and has a short duration time for effective embolization. Besides, none of the existing embolization strategies can eliminate CTCs to improve the prognosis. In contrast, the 3D tubular scaffold is an alternative option for the TACE of HCC; the biologically captured and physically trapped homing CTCs would form permanent artificial emboli with less recanalization. Furthermore, the 3D tubular scaffold has no displacement and good localization accuracy, making it suitable for efficient embolization therapy. In the future, if the 3D tubular scaffold can be integrated with various drug release systems for chemoembolization, the antitumor efficacy would be further enhanced.

The elimination of CTCs would inhibit tumor metastasis. A conductive catheter has been developed for the intravenous capture and subsequent electroporation damage of CTCs; however, the high working potential for electroporation might cause heart attack.³¹ Here, Au NTs, a type of nanomaterial with a distinguished photothermal effect, rapidly increase the temperature of the 3D tubular scaffold for efficient cell damage and can sense laser irradiation under a certain tissue depth. Furthermore, the Au NTs have good photothermal cycling stability, making them suitable for long-term and repetitive applications. Importantly, the radiopaque property of the Au NTs^52,53 allows them to be conveniently visualized by X-ray or computed tomography without additional contrast agents, making the 3D tubular scaffold suitable for different clinical interventions.

4. Conclusions

In summary, we have developed a two-pronged strategy to eliminate CTCs and primary tumors by implanting 3D tubular scaffolds in the blood vessels upstream and downstream of the primary tumor. The implantable 3D tubular scaffold has excellent biocompatibility, high CTC capture efficiency, and a good photothermal effect. The scaffold can continuously capture escaped CTCs and achieve the cascaded embolization of the primary tumor. Quantitatively, the 3D tubular scaffold achieved 86% of the CTC scavenging rate, 99% of the tumor inhibition rate and 100% photothermal killing efficiency. Thus, the scaffold can potentially prevent the risk of tumor metastasis. The long-term implantation and photothermal processes of the 3D tubular scaffold do not have obvious side effects in blood vessels, the adjacent tissues, or major organs of the organism, demonstrating its suitability for in vivo applications. In the future, the implantable biological device can be further developed as an auxiliary surgical tool, a new chemoembolization agent, and an interventional treatment device for cancer therapy to improve the antitumor efficacy and benefit cancer patients.

Author contributions

YiJing Chen: conceptualization, data curation, formal analysis, investigation, methodology, writing – original draft, and writing – review and editing. CuiWen Li: conceptualization, data curation, formal analysis, investigation, writing – original draft, and writing – review and editing. JingHui Yang: conceptualization, data curation, and investigation. Ming Wang: conceptualization, data curation, and formal analysis. YiKe Wang: conceptualization, data curation, and resources. ShiBo Cheng: conceptualization and data curation. GuoHua Yuan: methodology, supervision, and writing – review and editing. Min Xie: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, validation, visualization, writing – original draft, and writing – review and editing. WeiHua Huang: conceptualization, data curation formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, writing – original draft, and writing – review and editing.

Data availability

Data will be made available on request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grants no. 22274120 and 21974098) and the National Key Research and Development Program of China (2022YFA1104802).

References

1. V. Plaks, C. D. Koopman and Z. Werb, Science, 2013, 341, 1186–1188 CrossRef CAS PubMed .
2. P. S. Steeg, Nat. Med., 2006, 12, 895–904 CrossRef CAS PubMed .
3. C. L. Chaffer and R. A. Weinberg, Science, 2011, 331, 1559–1564 CrossRef CAS PubMed .
4. C. Dive and G. Brady, Cell, 2017, 168, 742 CrossRef CAS PubMed .
5. S. Paget, Lancet, 1889, 133, 571–573 CrossRef .
6. I. J. Fidler and G. Poste, Lancet Oncol., 2008, 9, 808 CrossRef PubMed .
7. J. G. Hiller, N. J. Perry, G. Poulogiannis, B. Riedel and E. K. Sloan, Nat. Rev. Clin. Oncol., 2018, 15, 205–218 CrossRef PubMed .
8. M. H. Barcellos-Hoff, C. Park and E. G. Wright, Nat. Rev. Cancer, 2005, 5, 867–875 CrossRef CAS PubMed .
9. C. C. Muth, JAMA, 2017, 317, 656 CrossRef PubMed .
10. J. M. Llovet, T. De Baere, L. Kulik, P. K. Haber, T. F. Greten, T. Meyer and R. Lencioni, Nat. Rev., 2021, 18, 293–313 CAS .
11. K. Huynh, Nat. Rev. Cardiol., 2017, 14, 67 CAS .
12. X. H. Li, X. F. Ji, K. Chen, W. U. Muhammad, B. Li, J. M. Cao, L. Xiao, J. Xiao and G. Yang, Bioact. Mater., 2021, 6, 2105–2119 CAS .
13. H. P. Jander and N. A. Russinovich, Radiology, 1980, 136, 337–344 CrossRef CAS PubMed .
14. E. Cevik, H. Albadawi, Z. F. Zhang, Y. Demirlenk, D. Atar, C. Keum, J. Kim, E. Graf, S. Gunduz, S. Rehman and R. Oklu, Adv. Mater., 2024, 2402570 CrossRef CAS PubMed .
15. J. J. Hu, H. Albadawi, B. W. Chong, A. R. Deipolyi, R. A. Sheth, A. Khademhosseini and R. Oklu, Adv. Mater., 2019, 31, 1901071 CrossRef PubMed .
16. I. Altun, J. J. Hu, H. Albadawi, Z. F. Zhang, M. A. Salomao, J. L. Mayer, L. Jamal and R. Oklu, Adv. Mater., 2020, 32, 2005603 CrossRef PubMed .
17. M. Y. Kim, T. Oskarsson, S. Acharyya, D. X. Nguyen, X. H. Zhang, L. Norton and J. Massague, Cell, 2009, 139, 1315–1326 CrossRef PubMed .
18. E. Comen, L. Norton and J. Massague, Nat. Rev. Clin. Oncol., 2011, 8, 369–377 CrossRef PubMed .
19. E. Dondossola, A. S. Dobroff, S. Marchio, M. Cardo-Vila, H. Hosoya, S. K. Libutti, A. Corti, R. L. Sidman, W. Arap and R. Pasqualini, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 2223–2228 CrossRef CAS PubMed .
20. S. M. Azarin, J. Yi, R. M. Gower, B. A. Aguado, M. E. Sullivan, A. G. Goodman, E. J. Jiang, S. S. Rao, Y. Ren, S. L. Tucker, V. Backman, J. S. Jeruss and L. D. Shea, Nat. Commun., 2015, 6, 8094 CrossRef PubMed .
21. S. S. Rao, G. G. Bushnell, S. M. Azarin, G. Spicer, B. A. Aguado, J. R. Stoehr, E. J. Jiang, V. Backman, L. D. Shea and J. S. Jeruss, Cancer Res., 2016, 76, 5209–5218 CrossRef CAS PubMed .
22. O. Vermesh, A. Aalipour, T. J. Ge, Y. Saenz, Y. Guo, I. S. Alam, S. M. Park, C. N. Adelson, Y. Mitsutake, J. Vilches-Moure, E. Godoy, M. H. Bachmann, C. C. Ooi, J. K. Lyons, K. Mueller, H. Arami, A. Green, E. I. Solomon, S. X. Wang and S. S. Gambhir, Nat. Biomed. Eng., 2018, 2, 696–705 CrossRef CAS PubMed .
23. W. W. Mu, Q. H. Chu, H. Z. Yang, L. Guan, S. L. Fu, T. Gao, X. Sang, Z. P. Zhang, S. Liang, Y. J. Liu and N. Zhang, Adv. Sci., 2022, 9, 2101472 CrossRef CAS PubMed .
24. Y. J. Zhang, X. Zhu, X. L. Chen, Q. J. Chen, W. X. Zhou, Q. Guo, Y. F. Lu, C. Li, Y. Zhang, D. H. Liang, T. Sun, X. B. Wei and C. Jiang, Adv. Funct. Mater., 2019, 29, 1806620 CrossRef .
25. M. J. Mitchell, E. Wayne, K. Rana, C. B. Schaffer and M. R. King, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 930–935 CrossRef CAS PubMed .
26. K. Y. Wang, H. Ye, X. B. Zhang, X. Wang, B. Yang, C. Luo, Z. Q. Zhao, J. Zhao, Q. Lu, H. T. Zhang, Q. M. Kan, Y. J. Wang, Z. G. He and J. Sun, Biomaterials, 2020, 257, 120224 CrossRef CAS PubMed .
27. N. Saucedo-Zeni, S. Mewes, R. Niestroj, L. Gasiorowski, D. Murawa, P. Nowaczyk, T. Tomasi, E. Weber, G. Dworacki, N. G. Morgenthaler, H. Jansen, C. Propping, K. Sterzynska, W. Dyszkiewicz, M. Zabel, M. Kiechle, U. Reuning, M. Schmitt and K. Lucke, Int. J. Oncol., 2012, 41, 1241–1250 Search PubMed .
28. H. Y. Zhang, Z. Z. Jia, C. C. Wu, L. G. Zang, G. W. Yang, Z. Z. Chen and B. Tang, ACS Appl. Mater. Interfaces, 2015, 7, 20477–20484 CrossRef CAS PubMed .
29. D. Wang, C. C. Ge, W. Y. Liang, Q. H. Yang, Q. Liu, W. Ma, L. L. Shi, H. Wu, Y. H. Zhang, Z. Z. Wu, C. Y. Wei, L. D. Huang, Z. Y. Fang, L. P. Liu, S. Y. Bao and H. Zhang, Adv. Sci., 2020, 7, 2000940 CrossRef CAS PubMed .
30. M. H. Duan, X. Y. Zhu, L. L. Fan, Y. Y. He, C. Yang, R. Guo, S. Chen, X. Y. Sun and J. Liu, Adv. Mater., 2022, 34, 2205002 CrossRef CAS PubMed .
31. D. Wang, R. H. Dong, X. D. Wang and X. Y. Jiang, ACS Nano, 2022, 16, 5274–5283 CrossRef CAS PubMed .
32. T. H. Kim, Y. Wang, C. R. Oliver, D. H. Thamm, L. Cooling, C. Paoletti, K. J. Smith, S. Nagrath and D. F. Hayes, Nat. Commun., 2019, 10, 1478 CrossRef PubMed .
33. L. J. Gay and B. Felding-Habermann, Nat. Rev. Cancer, 2011, 11, 123–134 CrossRef CAS PubMed .
34. N. M. Bambace and C. E. Holmes, J. Thromb. Haemostasis, 2011, 9, 237–249 CrossRef CAS PubMed .
35. S. B. Cheng, M. Xie, J. Q. Xu, J. Wang, S. W. Lv, S. Guo, Y. Shu, M. Wang, W. G. Dong and W. H. Huang, Anal. Chem., 2016, 88, 6773–6780 CrossRef CAS PubMed .
36. Y. L. Liu, Z. H. Jin, Y. H. Liu, X. B. Hu, Y. Qin, J. Q. Xu, C. F. Fan and W. H. Huang, Angew. Chem., Int. Ed., 2016, 55, 4537–4541 CrossRef CAS PubMed .
37. A. K. Gaharwar, R. K. Avery, A. Assmann, A. Paul, G. H. Mckinley, A. Khademhosseini and B. D. Olsen, ACS Nano, 2014, 8, 9833–9842 CrossRef CAS PubMed .
38. R. K. Avery, H. Albadawi, M. Akbari, Y. S. Zhang, M. J. Duggan, D. V. Sahani, B. D. Olsen, A. Khademhosseini and R. Oklu, Sci. Transl. Med., 2016, 8, 365ra156 Search PubMed .
39. D. W. Wang, Q. R. Wu, R. Guo, C. N. Lu, M. Niu and W. Rao, Nanoscale, 2021, 13, 8817–8836 RSC .
40. Y. Qin, X. B. Hu, W. T. Fan, J. Yan, S. B. Cheng, Y. L. Liu and W. H. Huang, Adv. Sci., 2021, 8, 2003738 CrossRef CAS PubMed .
41. A. W. Stanton, B. Holroyd, J. W. Northfield, J. R. Levick and P. S. Mortimer, Vasc. Med., 1998, 3, 3–8 Search PubMed .
42. S. Veeranarayanan, M. S. Mohamed, A. C. Poulose, M. Rinya, Y. Sakamoto, T. Maekawa and D. S. Kumar, Theranostics, 2018, 8, 5231–5245 Search PubMed .
43. Y. L. Liu, K. L. Ai, J. H. Liu, M. Deng, Y. Y. He and L. H. Lu, Adv. Mater., 2013, 25, 1353–1359 CrossRef CAS PubMed .
44. S. B. Cheng, M. Wang, C. Zhang, M. M. Chen, Y. K. Wang, S. Tian, N. Zhan, W. G. Dong, M. Xie and W. H. Huang, Anal. Chem., 2020, 92, 5447–5455 CrossRef CAS PubMed .
45. J. Q. Li, X. L. Li, Y. F. Zhang, K. Jin, Y. Yuan, R. Q. Ming, Y. H. Yang and T. S. Chen, Nano Res., 2023, 16, 873–881 CrossRef CAS .
46. F. Peng, M. I. Setyawati, J. K. Tee, X. G. Ding, J. P. Wang, M. E. Nga, H. K. Ho and D. T. Leong, Nat. Nanotechnol., 2019, 14, 279–286 CrossRef CAS PubMed .
47. B. Y. Zhang, A. Korolj, B. F. L. Lai and M. Radisic, Nat. Rev. Mater., 2018, 3, 257–278 CrossRef .
48. S. P. Li, Q. Jiang, S. L. Liu, Y. L. Zhang, Y. H. Tian, C. Song, J. Wang, Y. G. Zou, G. J. Anderson, J. Y. Han, Y. Chang, Y. Liu, C. Zhang, L. Chen, G. B. Zhou, G. J. Nie, H. Yan, B. Q. Ding and Y. L. Zhao, Nat. Biotechnol., 2018, 36, 258–264 CrossRef CAS PubMed .
49. S. P. Li, Y. L. Zhang, S. H. Ho, B. Z. Li, M. F. Wang, X. Deng, N. Yang, G. Liu, Z. Lu, J. Xu, Q. Shi, J.-Y. Han, L. Zhang, Y. Wu, Y. Zhao and G. Nie, Nat. Biomed. Eng., 2020, 4, 732–742 CrossRef CAS PubMed .
50. G. Go, A. Yoo, K. T. Nguyen, M. Nan, B. A. Darmawan, S. R. Zheng, B. Kang, C.-S. Kim, D. Bang, S. Lee, K.-P. Kim, S. S. Kang, K. M. Shim, S. E. Kim, S. Bang, D.-H. Kim, J.-O. Park and E. Choi, Sci. Adv., 2022, 8, eabq8545 CrossRef CAS PubMed .
51. A. Habib, K. Desai, R. Hickey, B. Thornburg, R. Lewandowski and R. Salem, Nat. Rev. Clin. Oncol., 2015, 12, 481–489 CrossRef PubMed .
52. S. Ahn, S. Y. Jung and S. J. Lee, Molecules, 2013, 18, 5858–5890 CrossRef PubMed .
53. A. Jakhmola, N. Anton and T. F. Vandamme, Adv. Healthcare Mater., 2012, 1, 413–431 CrossRef CAS PubMed .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4tb01151a

‡ These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2024