¹⁷⁷Lu‐labeled amidoxime‐functionalized biodegradable microspheres for transarterial radioembolization of hepatocellular carcinoma

Abstract

Purpose

Transarterial radioembolization (TARE) is an important locoregional therapy for hepatocellular carcinoma (HCC). However, clinically used yttrium‐90 (⁹⁰Y) microspheres are non‐biodegradable and provide limited imaging capability, while biodegradable polymeric microspheres prepared using conventional radiolabeling strategies often exhibit poor radionuclide stability.

Experimental design

We designed a biodegradable radiometal‐chelating microsphere based on amidoxime‐functionalized polylactic acid microspheres (PLA‐g‐PAO‐Ms) engineered via electron beam‐induced graft polymerization.

Results

The PLA‐g‐PAO‐Ms exhibited a high Lu³⁺ loading capacity (33.83 mg g⁻¹) and underwent controlled biodegradation in vitro after approximately 20 d. Cell Counting Kit‐8 (CCK‐8) assays confirmed favorable cytocompatibility, with both LO2 and HepG2 cell viabilities exceeding 90% after 72 h, showing no significant difference from the control group (p > 0.05); by contrast, free Lu³⁺ (50 mg·L⁻¹) significantly reduced HepG2 viability to 51.4% (p < 0.001). In a rat orthotopic liver tumor model, the intra‐arterial administration of ¹⁷⁷Lu‐labeled PLA‐g‐PAO‐Ms (¹⁷⁷Lu‐PLA‐g‐PAO‐Ms) resulted in pronounced tumor suppression with sustained in vivo radionuclide retention for at least 14 d. Biodistribution analysis revealed selective tumor accumulation of 8.79 ± 1.74 %ID g⁻¹ , significantly higher than that in normal liver (1 ± 0.54 %ID g⁻¹ ) and other organs ( P< 0.001), whereas free ¹⁷⁷LuCl₃ showed minimal tumor uptake (0.30 ± 0.28 %ID g⁻¹ ) with predominant non‐target organ accumulation.

Conclusion

These findings suggest that PLA‐g‐PAO‐Ms provide a biodegradable microsphere system capable of stable radiometal incorporation and effective tumor suppression following intra‐arterial administration, offering a promising approach for interventional radionuclide therapy for HCC.

Scheme 1. Schematic illustration of the synthesis and therapeutic evaluation of PLA‐g‐PAO microspheres (PLA‐g‐PAO‐Ms). Monodisperse PLA microspheres (PLA‐Ms) were initially fabricated via microfluidic technology. Surface grafting of acrylonitrile monomers was subsequently achieved through radiation‐induced graft polymerization, followed by amidoximation to convert the cyano groups into amidoxime chelating ligands, yielding the target PLA‐g‐PAO‐Ms. The schematic further depicts the in vivo administration of ¹⁷⁷Lu‐labeled PLA‐g‐PAO‐Ms and their anti‐tumor efficacy assessment in a rat orthotopic hepatocellular carcinoma model.

1. INTRODUCTION

Hepatocellular carcinoma (HCC) remains a major cause of cancer‐related mortality worldwide. ¹ , ² , ³ Radical hepatectomy remains the primary curative treatment for HCC and offers the best chance for long‐term survival. ⁴ For unresectable HCC, clinical management relies on alternative therapeutic strategies, including local ablation, transarterial therapies, external beam radiotherapy (EBRT), and systemic anticancer treatments, to improve survival and disease control. ⁵

Transarterial radioembolization (TARE) is an interventional form of internal radiotherapy that is widely used in the treatment of hepatic malignancies. TARE provides localized radionuclide retention within the tumor and enables continuous internal irradiation while largely sparing the surrounding normal tissues. ⁶ In addition, owing to the relatively long tissue penetration range of β radiation, TARE can effectively deliver cytotoxic radiation to tumor cells located deep within the lesion, which may provide advantages over transarterial chemoembolization (TACE) in certain hypovascular tumors. ⁷ However, several key challenges remain in the development of TARE microsphere systems. These include efficient radionuclide loading, stable retention of the radionuclide within the tumor tissue, and physicochemical properties that ensure stable suspension and smooth delivery through microcatheters. These parameters critically influence radiation dose distribution and therapeutic efficacy. An ideal microsphere platform should exhibit controlled biodegradability. Permanent retention of non‐degradable microspheres within the hepatic vasculature may lead to long‐term complications, such as vascular stenosis or chronic inflammatory responses. Controlled degradation after the completion of radionuclide decay could facilitate microsphere clearance from liver tissue, potentially alleviating local hypoxia and reducing hypoxia‐driven tumor aggressiveness, while theoretically enabling safer retreatment with subsequent cycles of radioembolization. However, despite the increasing research interest in biodegradable radioembolization platforms, no clinically approved microsphere products with predictable and controllable degradation profiles are currently available. Achieving an optimal balance among degradability, radionuclide retention, and therapeutic safety remains a major challenge in the development of next‐generation TARE microspheres. ⁸

Polylactic acid (PLA) microspheres are widely considered promising carriers for biodegradable radioembolization because of their favorable biocompatibility and tunable degradation properties. ⁹ However, radiolabeling strategies for PLA‐based microspheres often suffer from limited labeling stability and radionuclide leakage in vitro and in vivo, which has hindered their clinical translation for TARE. ¹⁰ Holmium‐166 (¹⁶⁶Ho)‐labeled PLA microspheres represent one of the earliest biodegradable radioactive microsphere systems investigated for TARE. These microspheres are typically produced by encapsulating non‐radioactive holmium acetylacetonate (Ho(acac)₃) within poly(L‐lactic acid) (PLLA) microspheres through an emulsion–solvent evaporation process, followed by neutron activation to convert ¹⁶⁵Ho into ¹⁶⁶Ho. ¹¹ , ¹² The high‐energy neutron irradiation required for activation can induce radiolytic degradation of the PLA matrix, leading to polymer chain scission and loss of structural integrity. Such damage to the carrier material increases the risk of radionuclide leakage and unintended systemic radiation exposure. ¹³

The physical characteristics of currently used radionuclides also impose important limitations on TARE. ⁹⁰Y, the clinical standard for radioembolization, is a pure β⁻ emitter that lacks γ‐emission suitable for nuclear imaging, preventing direct visualization of microsphere distribution and accurate in vivo dosimetry using single‐photon emission computed tomography (SPECT). ¹⁶⁶Ho emits both β⁻ particles and γ‐rays, but has a short physical half‐life of 26.8 h, which restricts the duration of intratumoral irradiation and complicates clinical logistics. Lutetium‐177 (¹⁷⁷Lu) has several advantages in interventional radionuclide therapies. With a physical half‐life of 6.7 days, it allows sustained radiation delivery while maintaining practical flexibility for radiopharmaceutical preparation and administration. Its β⁻ particles (Emax = 0.497 MeV) have a tissue penetration range of approximately 2 mm, enabling localized tumor irradiation with reduced exposure to surrounding liver tissue. In addition, the concomitant emission of 208 keV γ‐rays permits real‐time nuclear imaging and quantitative dosimetry using single‑photon emission computed tomography/computed tomography (SPECT/CT), supporting image‐guided therapy and treatment monitoring. To overcome the limitations associated with conventional radiolabeling strategies, we developed a surface chelation‐mediated labeling approach based on amidoxime‐functionalized PLA microspheres. Amidoxime (AO) groups were introduced onto the microsphere surface to enable stable coordination with Lu³⁺ ions while preserving the structural integrity of the polymer matrix. Compared to neutron activation–based methods, this strategy enables efficient radiolabeling under mild conditions without radiation‐induced damage to the carrier material. The AO ligand exhibits a strong affinity for trivalent metal ions and forms stable coordination complexes with Lu³⁺. Our previous density functional theory (DFT) studies demonstrated that AO groups can generate thermodynamically stable hexadentate chelation structures with favorable enthalpy and Gibbs free energy characteristics. ¹⁴ , ¹⁵ Unlike conventional chelators, such as 1,4,7,10‐tetraazacyclododecane‐N,N′,N″,N‴‐tetraacetic acid (DOTA) or ethylenediaminetetraacetic acid (EDTA), AO‐mediated coordination can be achieved under mild conditions at room temperature, avoiding thermal damage to the microsphere matrix. ¹⁶

In this study, monodisperse PLA microspheres were fabricated using microfluidic technology. Acrylonitrile groups were introduced via electron‐beam‐induced graft polymerization and subsequently converted into amidoxime groups via amidoximation, yielding amidoxime‐functionalized polylactic acid microspheres (PLA‐g‐PAO‐Ms). Physicochemical characterization confirmed successful surface functionalization and high ¹⁷⁷Lu chelation capacity. The resulting ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms demonstrated favorable biocompatibility, stable radionuclide retention, and pronounced antitumor efficacy in hepatocellular carcinoma models (Figure 1). The γ‐emission of ¹⁷⁷Lu further enabled non‐invasive SPECT imaging of microsphere distribution in vivo, providing intrinsic theranostic capability.

FIGURE 1.

Schematic illustration of the synthesis and antitumor application of PLA‐g‐PAO‐Ms. Monodisperse poly microspheres (PLA‐Ms) were first generated by via microfluidic technology. Acrylonitrile monomers was then grafted onto the PLA‐Ms surface through electron beam‐induced graft polymerization, followed by amidoximation reaction to transform the cyano groups in the grafted chains into amidoxime chelating ligands, thus obtaining the target PLA‐g‐PAO‐Ms. The scheme also outlines the workflow of in vivo administration and antitumor therapeutic efficacy evaluation of ¹⁷⁷Lu‐radiolabeled PLA‐g‐PAO‐Ms in a rat orthotopic HCC model. Created in https://BioRender.com

2. MATERIALS AND METHODS

2.1. Materials

PLA (Aladdin, ≥99%), Acetone (Aladdin, ≥99%), methanol (Aladdin, ≥99%), N,N‐dimethylformamide (DMF, Aladdin, ≥99%), Polyvinyl alcohol (PVA, Macklin, ≥99%), Acrylonitrile (AN, Aladdin, ≥99%), Hydroxylamine hydrochloride (NH₂OH·HCl, Macklin, ≥99%) and sodium hydroxide (NaOH, Macklin, ≥99%) were used as received. ¹⁷⁷Lu chloride (¹⁷⁷LuCl₃) solution (37 MBq) was used as received. Ultrapure water was used for all experiments.

2.2. Synthesis of PLA‐g‐PAO‐Ms

Preparation of PLA microspheres (PLA‐Ms): PLA‐Ms were fabricated using microfluidic technology combined with a phase separation method, which enabled the production of PLA‐Ms with a uniform diameter, regular morphology, and excellent size controllability by tuning the flow rates of the continuous/mobile phases and the solution concentration. To prepare the continuous phase, 1 mL of PLA/acetone solution (15 mg/ mL) was mixed with 1.32 mL of methanol and vortexed for 10 min. Meanwhile, 10 mL of 0.25% (w/v) PVA aqueous solution was prepared as the mobile phase. The as‐prepared continuous and mobile phases were loaded into syringes, and the flow rate ratio of the two phases was precisely adjusted using a syringe plunger to generate uniformly sized dispersed microspheres suspended in the PVA solution. The resulting PLA‐Ms were collected by filtration, freeze‐dried, and stored for subsequent characterization and further modification.

Preparation of PLA‐g‐PAO‐Ms: Prior to graft polymerization, PLA‐Ms were pre‐irradiated in nitrogen at room temperature using an electron beam accelerator (Pioneer Electric Company GJ‐2‐II, Institute of Applied Radiation, Shanghai University) with a beam energy of 2 MeV, a maximum beam current of 10 mA, and a total irradiation dose of 20 kGy. For the graft polymerization reaction, a monomer solution consisting of 3 mL of AN, 1 mL of AA, and 6 mL of DMF was added to a flask and deoxygenated for 10 min via nitrogen purging. The pre‐irradiated PLA‐Ms were subsequently added to the flask, and the graft polymerization reaction was carried out at 57°C for 4 h. Upon completion of the reaction, the resulting grafted product was thoroughly washed with DMF to remove residual monomers and homopolymers. After vacuum freeze‐drying, the desired acrylonitrile‐functionalized polylactic acid microspheres (PLA‐g‐PAN‐Ms) were successfully prepared. For the amidoximation reaction, 10 g of NH₂OH·HCl was dissolved in 100 mL of ultrapure water, and the pH of the solution was adjusted to neutral with 5.7 g of NaOH. The as‐prepared PLA‐g‐PAN‐Ms were dispersed in the neutral NH₂OH·HCl solution, and the mixture was heated to 60°C for 12 h to fully convert the cyano groups into AO functional groups. After the amidoximation reaction was completed, the product was centrifuged at 8000 r /min, the supernatant was discarded, and the solid was washed repeatedly with ultrapure water to remove residual NH₂OH·HCl and NaOH. The final product, PLA‐g‐PAO‐Ms, was obtained after vacuum‐drying for 24 h, followed by freeze‐drying. The chemical structures, microstructures, mechanical properties, functional group densities, and distributions of the obtained materials were subsequently characterized.

Radiolabeling of PLA‐g‐PAO‐Ms with ¹⁷⁷Lu: 1 mL of PLA‐g‐PAO‐Ms dispersion (30 mg mL⁻¹) was mixed with 10 µL of ¹⁷⁷LuCl₃ solution (37 MBq), and 10 µL of 0.04 M NaOH solution was added to adjust the pH of the mixture to 6–7. The mixture was oscillated at 60°C for 30 min, followed by three cycles of washing with ultrapure water and centrifugation (8,000 rpm, 5 min each) to remove unbound ¹⁷⁷Lu.

2.3. Characterization

Fourier‐transform infrared (FT‐IR) spectra were recorded on a Fourier‐transform infrared spectrometer (Bruker GmbH, Germany) with the KBr pellet method to confirm the chemical structure of PLA‐g‐PAO‐Ms, with a scan range of 4000–500 cm⁻¹ and a resolution of 4 cm⁻¹ and 32 scans per sample.

X‐ray photoelectron spectroscopy (XPS) spectra were collected on an X‐ray photoelectron spectrometer (Kratos Ltd., UK) with a monochromatic Al Kα X‐ray source. Twenty milligrams of PLA‐Ms, PLA‐g‐PAN‐Ms, PLA‐g‐PAO‐Ms, and Lu‐PLA‐g‐PAO‐Ms were used for the measurements, with the test voltage set to 10 kV, a full‐spectrum energy range of 160 eV, and a scanning range of 0–1200 eV. The pressure of the analysis chamber was maintained at 1×10⁻⁹ Torr during the test, and the obtained XPS data were analyzed using the Thermo Avantage software.

X‐ray diffraction (XRD) patterns were recorded on an X‐ray diffractometer (Rigaku Corporation, Japan) with Cu Kα radiation as the excitation source, a scanning range of 5–80°, and a scanning rate of 5° per min to characterize the crystalline structure of PLA‐Ms, PLA‐g‐PAN‐Ms, and PLA‐g‐PAO‐Ms.

Scanning electron microscopy (SEM) images were acquired using a field‐emission scanning electron microscope (JEOL Ltd., Japan) to study the morphological characteristics of PLA‐Ms, PLA‐g‐PAN‐Ms, and PLA‐g‐PAO‐Ms. Prior to testing, the microspheres were dried in a vacuum‐drying oven at 50°C for 6 h, then mounted on sample stubs with double‐sided conductive tape and subjected to gold sputtering at a current of 20 mA for 180 s. Morphological observations were performed at an accelerating voltage of 5 kV.

Contact angles were measured using a contact angle meter (KSV Instruments Ltd., Finland) for the PLA‐Ms and PLA‐g‐PAO‐Ms pellets, which were prepared by tableting under 5 MPa pressure for 30 s. A 50 µL drop of ultrapure water (18.2 MΩ cm) was dispensed onto the sample surface, with video recording to capture images every 50 ms. Contact angles were determined from the images captured at the moment of droplet‐sample contact, and measurements were repeated five times at different positions for each sample. Thermogravimetric analysis (TGA) and derivative thermogravimetry (DTG) was performed on a TA Instruments Discovery Series TGA 550 Thermogravimetric Analyzer (TA, USA) for PLA‐Ms, PLA‐g‐PAN‐Ms, and PLA‐g‐PAO‐Ms under a high‐purity nitrogen atmosphere at a flow rate of 50 mL /min. The measurements were carried out over a temperature range of 25–800°C at a heating rate of 10°C per min. The temperature corresponding to the maximum thermal degradation rate (T_max) of each material was determined from the DTG curves, and the thermal stability of materials during modification was analyzed by comparing the TGA/DTG curves of PLA‐Ms, PLA‐g‐PAN‐Ms and PLA‐g‐PAO‐Ms.

Laser diffraction particle size distribution was determined using a Mastersizer 3000 (Malvern Panalytical Ltd., UK) for the PLA‐g‐PAO‐Ms in wet dispersion mode with ultrapure water as the dispersant. The samples were ultrasonicated for 5 min before testing for uniform dispersion, after which the median particle diameter (D50) and the proportion of particles in specific diameter ranges were obtained, and the uniformity of the microsphere size distribution was evaluated from the test results.

Dynamic light scattering (DLS) measurements were conducted on a Brookhaven NanoBook Omni Nanoparticle Size and Zeta Potential Analyzer (Brookhaven, USA) at a scattering angle of 90° to determine the hydrodynamic diameter of the PLA‐g‐PAO‐Ms under simulated physiological conditions. A 5 mL suspension of PLA‐g‐PAO‐Ms (10 mg /mL) was dispersed in phosphate‐buffered saline (PBS, pH 7.4, 0.01 mol /L) and transferred to sterilized centrifuge tubes, which were incubated in a shaking incubator at 37°C with agitation at 200 rpm for 35 d. An aliquot of the suspension was collected every 5 d; subsequently, the particle diameter was analyzed immediately and the average particle size was plotted against the degradation time to assess the microsphere stability and degradation rate.

Inductively coupled plasma mass spectrometry (ICP‐MS) measurements were performed on an Thermo Scientific iCAP TQ ICP‐MS (Thermo Scientific, USA) to determine the Lu³⁺ concentration in the supernatant of chelation reaction mixtures, with three parallel replicates for each sample and all centrifugation steps conducted at 8000 rpm for 5 min. For pH regulation experiments, a 100 mL Lu³⁺ aqueous solution (1 mg /L) was prepared, and its pH was adjusted to 5–9. Next, 0.2 mg of PLA‐g‐PAO‐Ms was added and incubated at 37°C while shaking at 150 rpm. Then, 1 mL of supernatant was collected at 10, 20, 30, and 60 min, then transferred to 2% (v/v) HNO₃ for analysis. For different initial concentration experiments, 100 mL of Lu³⁺ solution (0.1, 1, and 10 mg/ L), with the pH adjusted to 7, were used with the same chelation and testing procedure. For different solid‐liquid ratio experiments, 0.1, 0.2, and 2 mg of PLA‐g‐PAO‐Ms were added to 100 mL of Lu³⁺ solution (1 mg/ L, pH 7), and the Lu³⁺ concentration was measured at the same time points after chelation.

2.4. CCK‐8 assay

The CCK‐8 assay was conducted using a PerkinElmer multimode microplate reader (PerkinElmer Inc., USA) to evaluate the cytotoxicity of Lu‐PLA‐g‐PAO‐Ms in HepG2 and LO2 cells. HepG2 and LO2 cells in the logarithmic growth phase were digested with trypsin to prepare single‐cell suspensions, and the cell density was adjusted to 1×10⁴ cells per mL. The cell suspension was seeded into 96‐well cell culture plates at 100 µL per well, with medium‐only wells set as blank controls. The plates were pre‐incubated in a 5% CO₂ humidified incubator at 37°C for 24 h to allow complete cell adhesion. After pre‐incubation, the original medium was carefully discarded, and 100 µL of complete medium containing Lu‐PLA‐g‐PAO‐Ms at different concentrations, 500 µg/ mL LuCl₃ control solution, or blank complete medium was added to each well according to experimental groups. Each concentration was set in six replicates. After sample addition, the plate was gently tapped for mixing, and incubation was continued under the same conditions. The plates were removed at 48 and 72 h post‐incubation. CCK‐8 reagent was mixed with Dulbecco's modified Eagle medium (DMEM) basal medium at a 1:9 (v/v) ratio, and 100 µL of the mixture was added to each well under dark conditions to avoid bubble formation. After gentle mixing, the plates were returned to the incubator for 2 h of dark incubation at 37°C. Immediately after incubation, the absorbance values (OD) at 450 nm were measured using a microplate reader and raw data were recorded. Cell viability was calculated using the following formula:

$$\mathrm{Cell}\mathrm{viability}(\%)=\frac{\left(\mathrm{O}{\mathrm{D}}_t-\mathrm{O}{\mathrm{D}}_b\right)}{\left(\mathrm{O}{\mathrm{D}}_c-\mathrm{O}{\mathrm{D}}_b\right)}\times 100$$

where OD_b, OD_c, and OD_t represent the absorbance values of the blank control, negative control (blank complete medium), and experimental groups, respectively. Statistical analysis of the cell viability in each group was performed using GraphPad Prism software (version 8.0) to evaluate the effects of Lu‐PLA‐g‐PAO‐Ms at different concentrations and incubation times on the proliferation of HepG2 and LO2 cells.

2.5. In vitro radiochemical stability studies

Owing to the short half‐life of ¹⁷⁷Lu (6.6 d) and the associated radiological safety constraints, non‐radioactive natural lutetium was employed as a surrogate radionuclide to simulate the in vitro radiochemical stability of Lu‐PLA‐g‐PAO‐Ms under physiological conditions. Lu‐loaded microspheres were prepared by chelating Lu onto PLA‐g‐PAO‐Ms using the chelation procedure described above. The resulting Lu‐PLA‐g‐PAO‐Ms were thoroughly rinsed with ultrapure water to remove any unbound or loosely adsorbed Lu ions prior to stability testing.

For the stability evaluation, parallel suspensions of Lu‐PLA‐g‐PAO‐Ms (10 mg/ mL) were prepared in two different media: (i) phosphate‐buffered saline and (ii) cell culture medium supplemented with 10% (v/v) fetal bovine serum (FBS). The suspensions were transferred into sterilized centrifuge tubes and incubated in a thermostated shaking incubator at 37 ± 0.5°C with continuous horizontal agitation at 200 rpm for 10 d to mimic physiological hydrodynamic conditions. Triplicate tubes from each medium group were withdrawn from the incubator at daily intervals over a 10‐day incubation period (days 1–10). Each sample was centrifuged at 8,000 rpm for 5 min to achieve complete phase separation. The supernatant was carefully collected and filtered through a 0.22 µm polyethersulfone (PES) syringe filter to eliminate any suspended microspheres. The pelleted microspheres were subsequently washed twice with 1 mL of fresh PBS under identical centrifugation conditions to remove the residual loosely bound Lu species. The combined washes were pooled with the initial supernatant for total released Lu quantification.

For radiochemical analysis, both the pooled supernatant/washing solutions and the final recovered microspheres were subjected to inductively coupled plasma mass spectrometry (ICP‐MS) analysis. The supernatant samples were directly diluted with 2 % (v/v) HNO₃ to a final volume of 10 mL and analyzed for Lu concentration. The microsphere pellets were fully digested in concentrated nitric acid (65 %, w/w) at 80°C until complete dissolution was achieved, followed by dilution with 2 % (v/v) HNO to the same final volume. The percentage of Lu released from the microspheres at each time point was calculated using the following equation:

$$\mathrm{Lu}\mathrm{release}\left(\%\right)=\frac{C_{\mathrm{s}}\times V_{\mathrm{s}}}{C_{\mathrm{s}}\times V_{\mathrm{s}}+C_{\mathrm{m}}\times V_{\mathrm{d}}}\times 100$$

Where C_s and C_m represent the Lu concentrations determined by ICP‐MS in the supernatant and microsphere digest solutions, respectively, V_s denotes the total volume of the pooled supernatant and washings, and V_d is the final dilution volume of the digested microsphere sample. All measurements were performed in triplicate, and the results are reported as mean ± standard deviation.

2.6. Wound healing assay

A wound healing assay was performed using a Thermo Scientific EVOS M5000 inverted fluorescence microscope (Thermo Scientific, USA) to assess the effect of Lu‐PLA‐g‐PAO‐Ms on the migratory ability of HepG2 cells. HepG2 cells in the logarithmic growth phase were prepared as a single‐cell suspension at a density of 2.5×10⁵ cells per mL and seeded into 6‐well cell culture plates at 2 mL per well. The plates were cultured under standard conditions until the cells formed a confluent monolayer with a fusion rate of > 90%. Two vertical straight scratches were uniformly made on the cell monolayer in each well using a sterile 200 µL pipette tip, keeping the tip vertical to avoid uneven scratch width, and the scratch position in each well was kept consistent. After scratching, the original medium was immediately discarded and the wells were gently rinsed three times with sterile PBS to completely remove the detached cells. Under a microscope, it was confirmed that the scratch edges were clear without residual floating cells. According to the experimental groups, 2 mL of serum‐free DMEM containing Lu‐PLA‐g‐PAO‐Ms at different concentrations was added to each well. Serum‐free DMEM without microspheres was used as the blank control group. Each concentration was set in three replicates. Immediately after administration, the 6‐well plate was placed under an inverted phase‐contrast microscope and images of fixed fields of each scratch were captured at 100× magnification to record the initial scratch width at 0 h. The plate was then returned to the incubator for continued culture, and images of the same fields at the same magnification were captured again at 24 and 48 h post‐incubation to record wound healing. The ImageJ software was used to measure the scratch width at each time point, and the cell migration distance was calculated as follows: migration distance = initial scratch width × scratch width at the corresponding time point. The migration distances of the cells in each group at different time points were statistically analyzed to compare the effects of different concentrations of microspheres on the migration ability of HepG2 cells.

2.7. Calcein‐AM/PI live‐dead cell staining assay

The calcein‐AM/PI live/dead cell staining assay was performed using a Thermo Scientific EVOS M5000 inverted fluorescence microscope (Thermo Scientific, USA) to visually evaluate the cytotoxicity of Lu‐PLA‐g‐PAO‐Ms. HepG2 cells in the logarithmic growth phase were prepared as single‐cell suspensions at a cell density of 1×10⁵ cells per mL and seeded into 6‐well plates at 1 mL per well. The plates were pre‐incubated for 24 h in a 5% CO₂ humidified incubator at 37°C to allow cell adhesion. After pre‐incubation, the original medium was discarded and 1 mL of complete medium containing Lu‐PLA‐g‐PAO‐Ms at different concentrations was added to each well according to the experimental groups. Complete medium without microspheres was used as the blank control group. Each concentration was set with three replicates, and the plates were incubated for 72 h under the same standard conditions.

At the end of the incubation period, the medium from each well was carefully collected in a sterile centrifuge tube. The cells were then gently rinsed twice with sterile PBS, and the rinse solution was collected in the corresponding centrifuge tube to avoid loss of floating cells. The collected medium and rinse solution were centrifuged at 1000 r/ min for 3 min and the supernatant was discarded. According to the instructions of the live‐dead cell staining kit, 5 µL of concentrated Calcein‐AM and PI staining solution were mixed thoroughly with 10 mL of PBS to prepare the staining working solution. The working solution (1 mL) was added to each centrifuge tube to resuspend the cell pellet, and the resuspended cell suspension was transferred back to the original cell culture well. The plate was shaken gently to ensure uniform contact between cells and the staining solution, and the cells were incubated in the dark at 37°C for 15 min.

After incubation, dual‐channel images of the same field were collected under a fluorescence microscope to observe the cell survival status, morphological changes, and death to intuitively evaluate the cytotoxicity of the microspheres.

2.8. Coagulation time determination

The coagulation time was determined in 96‐well plates to evaluate the effect of Lu‐PLA‐g‐PAO‐Ms on blood coagulation. Rabbit whole blood used in the experiment was collected in tubes containing sodium citrate anticoagulant. 50 µL of PBS, 50 µL of physiological saline (9 g /L), 50 µL of CaCl₂ solution (0.1 mol/ L), and 50 µL of Lu‐PLA‐g‐PAO‐Ms physiological saline dispersions at different concentrations (50, 100, 250, 500, 1000 µg /mL) were added to consecutive wells of the 96‐well plate, respectively. 50 µL of rabbit whole blood was added to each well. At 5, 10, 20, 30, and 60 min post‐incubation, coagulation was terminated by rinsing the plates with physiological saline. The plate was repeatedly washed and the liquid was aspirated until the solution became clear (indicating that all soluble blood components had been removed). At the end of the experiment, coagulation time and blood clot formation in each well were recorded.

2.9. Hemolysis assay

Hemolysis assays were conducted using a multimode microplate reader to evaluate the haemocompatibility of Lu‐PLA‐g‐PAO‐Ms. Rabbit whole blood (1 mL) was added to a centrifuge tube and centrifuged at 1500 rpm for 5 min. The supernatant was discarded and 5–10 times the volume of physiological saline was added to the precipitate. The centrifugation and washing steps were repeated three times to prepare a red blood cell suspension (RCS). Eight 1.5 mL centrifuge tubes were prepared, and 1 mL of PBS, 1 mL of physiological saline, 1 mL of ultrapure water, and 1 mL of Lu‐PLA‐g‐PAO‐Ms physiological saline dispersions at different concentrations (50, 100, 250, 500, 1000 µg /mL) were added to the tubes, respectively. Twenty microliters of RCS was added to each centrifuge tube, and the tubes were incubated in a constant temperature incubator at 37°C for 2 h. After incubation, the supernatant was aspirated from each tube and added to six wells of a 96‐well plate (100 µL per well). The absorbance (OD) at 540 nm was measured using a microplate reader. The hemolysis rate was calculated using the following formula:

$$\mathrm{Hemolysis}\mathrm{rate}(\%)=\frac{\left(\mathrm{O}{\mathrm{D}}_q-\mathrm{O}{\mathrm{D}}_n\right)}{\left(\mathrm{O}{\mathrm{D}}_p-\mathrm{O}{\mathrm{D}}_n\right)}\times 100$$

where OD_q, OD_p, and OD_n represent the absorbance values of the Lu‐PLA‐g‐PAO‐Ms dispersion, positive control (ultrapure water), and negative control (physiological saline) groups, respectively.

2.10. Experimental animals and establishment of orthotopic N1S1 Hepatocellular carcinoma (HCC) rat model

Specific pathogen‐free (SPF) female Sprague‐Dawley (SD) rats (100–120 g) were selected as experimental animals and housed in a clean environment at the Laboratory Animal Center. Rat hepatocellular carcinoma N1S1 cells were cultured in vitro and collected when they had reached approximately 80–90% confluence (∼1×10⁷ cells per T75 flask). The collected cells were rinsed thrice with PBS, and the cell concentration was adjusted to 5×10⁷ cells per mL to prepare a cell suspension, which was stored on ice for subsequent use.

Rats were anesthetized by intraperitoneal injection of 100 µL of 3% sodium pentobarbital solution. After complete anesthesia, the rats were fixed on the operating table, and the abdominal hair was removed with depilatory cream, followed by disinfection of the abdominal skin with iodophor. A transverse incision was made below the xiphoid process using sterile ophthalmic scissors to expose the liver, and the left outer lobe of the liver was gently pulled out of the abdominal cavity using a sterile cotton swab. A 100 µL aliquot of the prepared N1S1 cell suspension was aspirated with an insulin needle and slowly injected into the intrahepatic subcapsular region of the exposed liver parenchyma. After injection, the needle hole was pressed with a cotton swab to prevent leakage of the cell suspension and the syringe was withdrawn. After confirming that there was no bleeding at the needle hole, the liver was returned to the abdominal cavity, and the surgical wound was sutured and fully disinfected before the rats were placed back into the breeding cages for routine rearing. Magnetic resonance imaging (MRI) was performed on a 9.4 T small animal MRI scanner 7 days post‐modeling to monitor tumor engraftment, and successful orthotopic HCC model establishment was confirmed by MRI results.

2.11. Grouping and hepatic artery administration of therapeutic agents

Rats with successfully engrafted orthotopic liver tumors (tumor volume approximately 150 mm³) were randomly divided into four experimental groups with six rats per group: normal saline (NS) control, ¹⁷⁷LuCl₃, blank PLA‐g‐PAO microspheres (PLA‐g‐PAO‐Ms), and ¹⁷⁷Lu‐labeled PLA‐g‐PAO microspheres (¹⁷⁷Lu‐PLA‐g‐PAO‐Ms). All therapeutic agents were administered via hepatic artery intervention, and the specific operational steps were as follows: rats were anesthetized with 3% sodium pentobarbital as described above; the abdominal cavity was opened, and the abdominal artery, hepatic artery, and gastroduodenal artery were gently separated with sterile cotton swabs and forceps; two ligators were placed around the gastroduodenal artery, with the distal end ligated, and a ligator was placed around the abdominal artery for temporary blood flow interruption; a self‐made needle was used to puncture the gastroduodenal artery at the upstream of the distal ligation site, and a catheter was inserted into the hepatic artery for agent injection.

The administration dosage for each group was strictly controlled as 100 µL per rat: the NS group was injected with 100 µL of physiological saline; the ¹⁷⁷LuCl₃ group with 100 µL of ¹⁷⁷LuCl₃ solution (11.1 MBq); the blank PLA‐g‐PAO‐Ms group with 100 µL of PLA‐g‐PAO‐Ms suspension (20 mg/ mL); and the ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group with 100 µL of ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms formulation (20 mg/ mL, 11.1 MBq). After injection, the proximal end of the gastroduodenal artery (upstream of the puncture site) was ligated, the ligator around the abdominal artery was removed to restore hepatic artery blood flow, and the abdominal cavity was sutured and disinfected.

2.12. In vivo imaging and tumor volume dynamic monitoring

Single‐photon emission computed tomography/computed tomography (SPECT/CT) imaging was performed using a nanoScan SPECT/CT scanner at 7 and 14 d post‐administration to evaluate the In vivo biological distribution and radiostability of ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms. Dynamic MRI monitoring was conducted on a 9.4 T small animal MRI scanner at 7 and 14 d post‐administration to assess the tumor growth response to different interventions. Tumor volume was calculated using the following formula: V (mm³) = (L × S²)/2, where L represents the maximum diameter of the tumor and S represents the minimum diameter of the tumor, with tumor size measured from the MRI results.

2.13. Biodistribution analysis of radionuclides

All experimental rats were euthanized 14 d after administration for biodistribution analysis. Rats from the ¹⁷⁷LuCl₃ group and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group were selected for tissue collection: heart, liver, spleen, lung, kidney, stomach, intestine, bone, muscle, skin, and tumor tissues were carefully dissected, placed in sterile plastic test tubes, and weighed immediately. The residual radioactive concentration of each tissue sample was detected using a γ‐counter, and the radioactive uptake level was calculated as the percentage of injected dose per gram of tissue (%ID /g) to quantify the in vivo distribution of ¹⁷⁷Lu.

All animal experiments were performed in compliance with standard laboratory animal care protocols and the experimental procedures were designed to ensure reproducibility with detailed operational specifications.

All experimental data were processed and plotted using Origin 2021 and GraphPad Prism 8.0 software. Statistical analysis was performed with an unpaired two‐tailed Student's t‐test for two‐group comparisons of continuous variables and Kruskal‐Wallis ANOVA with Dunn's multiple comparison test for multiple‐group comparisons. Data was expressed as mean ± standard deviation (SD) and P < 0.05 was considered statistically significant. The XPS spectral data were specifically fitted and analyzed using the Thermo Avantage software.

3. RESULTS AND DISCUSSION

3.1. Synthesis and characterization of PLA‐g‐PAO‐Ms

PLA‐g‐PAO‐Ms were generated in three steps: preparation of PLA‐Ms by microfluidic technology, surface functionalization via electron‐beam‐induced graft polymerization, and amidoximation modification (Figure 1). The microfluidic process yielded microspheres with a uniform size distribution and regular morphology. The relevant preparation parameters were systematically optimized in our previous study. Electron‐beam‐induced graft polymerization was used to introduce nitrile‐containing chains onto the surface of the PLA‐Ms under additive‐free conditions at room temperature. Subsequent amidoximation converts these nitrile groups into amidoxime ligands, yielding the target PLA‐g‐PAO‐Ms and providing coordination sites for radiometallic binding. FT‐IR confirmed the stepwise surface modification (Figure 2a). For PLA‐Ms, the characteristic peaks observed at 1082 cm⁻¹ and 1748 cm⁻¹ are assigned to the symmetric stretching vibration of the C‐O‐C ether bond and the stretching vibration of the ester carbonyl (C = O) group, respectively. ¹⁷ After irradiation‐induced grafting, a new characteristic peak emerged at 2243 cm⁻¹ in the FT‐IR spectrum of PLA‐g‐PAN‐Ms, which is attributed to the stretching vibration of the cyano (C≡N) group, confirming that acrylonitrile was successfully grafted onto the surface of the microspheres. After the amidoximation reaction, the characteristic cyano peak at 2243 cm⁻¹ in the FT‐IR spectrum of the PLA‐g‐PAO‐Ms completely disappeared. Meanwhile, new absorption peaks appeared at 3401 cm⁻¹, 1647 cm⁻¹ and 922 cm⁻¹, which correspond to the stretching vibration of the N‐H bond, the stretching vibration of the C = N bond, and the bending vibration of the N‐O bond in the amidoxime group, respectively, verifying the successful introduction of the target functional groups. ¹⁸ XPS analysis further confirmed the completion of the grafting modification and amidoximation reaction. The N 1s characteristic peak at 399.1 eV was observed in the full‐survey XPS spectra of both PLA‐g‐PAN‐Ms and PLA‐g‐PAO‐Ms, whereas no such signal was detected in the pristine PLA‐Ms, indicating that nitrogen‐containing groups were successfully introduced onto the microsphere surface via the grafting reaction. The high‐resolution N 1s spectrum of PLA‐g‐PAO‐Ms was subjected to deconvolution fitting, and three characteristic peaks were fitted with binding energies at 399.33 eV, 399.84 eV and 400.88 eV, which are assigned to C‐NH, C‐NH₂ and C = N bonds, respectively (Figure 2b and c). The high‐resolution C 1s spectrum further confirmed the presence of chemical structures, including O‐C‐O, ‐CN, C‐H, and C = C (Figure S1a, Supporting Information). XPS analysis was performed on PLA‐g‐PAO‐Ms after Lu³⁺ chelation, and the characteristic double peaks of Lu 4d_5/2 and Lu 4d_3/2 were observed at 197.48 eV and 206.43 eV, which confirmed the coordination binding between Lu³⁺ and the amidoxime groups on the microsphere surface (Figure 2d).

FIGURE 2.

Physicochemical and functional characterization of the microspheres. (a) FT‐IR spectra of PLA‐Ms, PLA‐g‐PAN‐Ms, and PLA‐g‐PAO‐Ms; (b) XPS survey spectra of PLA‐Ms, PLA‐g‐PAN‐Ms, PLA‐g‐PAO‐Ms, and Lu‐loaded PLA‐g‐PAO‐Ms (Lu‐PLA‐g‐PAO‐Ms); (c) High‐resolution N 1s XPS spectrum of Lu‐PLA‐g‐PAO‐Ms; (d) High‐resolution Lu 4d XPS spectrum of Lu‐PLA‐g‐PAO‐Ms; (e) TGA curves of PLA‐Ms, PLA‐g‐PAN‐Ms, and PLA‐g‐PAO‐Ms; (f) Particle size distribution curve of PLA‐g‐PAO‐Ms; (g) Static water contact angle profile of PLA‐Ms; (h) Static water contact angle profile of PLA‐g‐PAO‐Ms; (i) Optical photograph of the dispersibility of PLA‐Ms and PLA‐g‐PAO‐Ms in aqueous solution; (j) Particle size variation curve of PLA‐g‐PAO‐Ms during in vitro degradation; (k) Time‐dependent Lu³⁺ loading capacity of PLA‐g‐PAO‐Ms at different initial Lu³⁺ concentrations; (l) In vitro radiochemical stability of ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms.

TGA was performed to characterize the changes in the thermal stability of the microspheres during the stepwise modification (Figure 2e). For pristine PLA‐Ms, the T_max was determined to be 460°C. After irradiation‐induced graft polymerization, a distinct decrease in T_max to 330°C was observed for the obtained PLA‐g‐PAN‐Ms, corresponding to the successful grafting of acrylonitrile onto the microsphere surface. Following the amidoximation modification, the thermal stability of the final product PLA‐g‐PAO‐Ms further declined, with an onset decomposition temperature of approximately 180°C. Although stepwise surface modification reduced the overall thermal stability of the microspheres, the thermal degradation behavior of the PLA matrix backbone was not fundamentally altered and still met the thermal stability requirements for room temperature radiolabeling and subsequent interventional therapy. SEM observations showed that both the PLA‐Ms and PLA‐g‐PAO‐Ms exhibited regular spherical morphologies and smooth surface structures (Figure 3). Laser particle size analysis (wet dispersion mode) results showed that the median particle size (D50) of PLA‐g‐PAO‐Ms was 33.73±0.03 µm, with more than 90% of the microspheres presenting a particle size distribution in the range of 20–80 µm (Figure 2f), which met the particle size requirements of embolic microspheres for TARE therapy. ¹⁹ In addition, the dispersion test confirmed that the microspheres could be stably dispersed in PBS without obvious aggregation (Figure 2l). water contact angle (WCA) tests showed that The WCA of the PLA‐g‐PAO‐Ms decreased significantly from 119.1° for PLA‐Ms to 15.9° (Figure 2g and h). The remarkable reduction in the contact angle confirms that amidoximation modification can greatly improve the surface hydrophilicity of the microspheres. In vitro degradation assays for PLA‐g‐PAO‐Ms were carried out in a 10% FBS‐containing buffer system under physiological‐mimicking conditions (37°C, 100 rpm). As shown in Figure 2(j), the as‐prepared PLA‐g‐PAO‐Ms initiated the degradation process on the 20th day of incubation and achieved a mass loss rate of 72% (corresponding to 28% residual mass) on day 35. Given that the physical half‐life of ¹⁷⁷Lu is 6.7 d, its radioactivity will decay to less than 5% of the initial level after five half‐lives (approximately 33.5 d). Notably, PLA‐g‐PAO‐Ms maintained excellent structural integrity during the entire effective therapeutic irradiation window (the first 20 days after administration), which could effectively avoid premature leakage of the loaded radionuclide and its subsequent non‐specific systemic distribution, thus guaranteeing the biosafety of the carrier during the whole treatment cycle. ²⁰

FIGURE 3.

SEM images and energy‐dispersive X‐ray spectroscopy (EDS) mapping of PLA‐Ms (a), PLA‐g‐PAN‐Ms (b), and PLA‐g‐PAO‐Ms (c).

To optimize the Lu³⁺ chelation efficiency of PLA‐g‐PAO‐Ms, the effects of key parameters, including solution pH, solid‐to‐liquid ratio, reaction time, and initial Lu³⁺ concentration, on the chelation capacity were systematically investigated. In the pH range of 5 to 9, the chelation capacity increased with increasing pH, with values of 2.19, 2.28, 2.32, 3.06, and 3.60 mg/ g, respectively (Figure S1b). The significantly improved chelation efficiency at pH 8–9 can be attributed to two factors. First, the formation of a Lu(OH)₃ precipitate can be avoided at pH < 9. Second, the degree of deprotonation of the amidoxime groups increases under alkaline conditions, which enhances their chelating affinity for Lu³⁺ ions. Kinetic studies under different pH conditions showed that the chelation capacity increased with increasing reaction time and reached equilibrium within 30 min (Figure S1c). The equilibrium chelation capacities at pH 5, 7, and 9 were 2.20, 2.32, and 3.60 mg g⁻¹, respectively, with corresponding Lu³⁺ utilization efficiencies of 44.6%, 47.2%, and 73.1%, respectively. Solid‐to‐liquid ratio optimization experiments showed that the Lu³⁺ chelation capacity per unit mass of microspheres increased correspondingly with decreasing solid‐to‐liquid ratio (0.1, 0.5, and 2 mg /mL) (Figure S1d), and chelation equilibrium was reached within 30 min for all groups. The chelation capacity of the PLA‐g‐PAO‐Ms was positively correlated with the initial Lu³⁺ concentration (Figure 2k). When the Lu³⁺ concentrations were 0.1, 1, and 10 ppm, the corresponding equilibrium chelation capacities were 0.12, 3.49, and 33.83 mg/ g, respectively.

The reported specific activity of GMP‐grade ¹⁷⁷Lu ranges from approximately 740 to 2960 GBq /mg. ²¹ Based on the maximum chelation capacity (33.83 mg/ g), the theoretical ¹⁷⁷Lu radioloading capacity of 1 mg PLA‐g‐PAO‐Ms reaches 24.7–98.8 GBq, far exceeding that of clinical ⁹⁰Y glass microspheres (0.05–0.1 GBq /mg) and resin microspheres (0.002 GBq/ mg). In vitro stability studies showed ¹⁷⁷Lu retention rates above 98.9% in both PBS and 10% FBS after 10 d at 37°C (Figure 2l), approaching the >99.5% retention reported for clinical ⁹⁰Y products. ²² Benchmarked against approved TARE platforms, the D50 of 33.73 µm (20–80 µm, >90%) satisfies the embolic size window of TheraSphere (20–30 µm) and SIR‐Spheres (20–60 µm), while the PLA matrix density (∼1.25 g /mL) avoids the rapid sedimentation of high‐density TheraSphere glass (3.6 g/ mL). ²³ Unlike ⁹⁰Y—a pure β‐emitter requiring low‐resolution Bremsstrahlung SPECT (∼12–15 mm) or ultra‐low‐probability positron emission tomography (PET) for post‐treatment imaging—¹⁷⁷Lu enables direct SPECT/CT (∼8–10 mm) via inherent γ‐emissions (112.9 and 208.4 keV) ²⁴ , ²⁵ , facilitating real‐time dosimetry. Moreover, the low mean β‐energy of ¹⁷⁷Lu (0.134 MeV, tissue penetration 0.67 mm) delivers conformal irradiation for small‐to‐medium lesions while sparing normal tissue, whereas ⁹⁰Y (0.934 MeV, 2.5 mm) relies on cross‐fire effects for large tumors but increases collateral damage risk. ²³ In summary, ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms combine high radionuclide loading, excellent radiochemical stability, favorable size distribution, inherent theranostic imaging, and reduced off‐target radiation hazard, providing a qualified carrier material for subsequent biological evaluation. ²⁰ , ²⁶ , ²⁷

3.2. Cytocompatibility and hemocompatibility of Lu‐PLA‐g‐PAO‐Ms

Cytocompatibility is a core metric for evaluating the biosafety of carrier materials used in TARE. ²⁸ To systematically assess the cytocompatibility of Lu‐PLA‐g‐PAO‐Ms, HCC HepG2 cells and human normal hepatic LO2 cells were selected as in vitro models, and a comprehensive analysis was performed using a battery of assays, including the plate colony formation, CCK‐8 proliferation, scratch wound healing, and live/dead cell staining assays. The results of the plate colony formation assay showed no significant differences in the colony number and morphological characteristics of HepG2 cells between the control group and groups treated with gradient doses of Lu‐PLA‐g‐PAO‐Ms, with no statistically significant difference (P>0.05, Figure 4a and b). The colony formation ability reflects the long‐term proliferation potential and self‐renewal capacity of cells. These results indicate that the microspheres exerted no inhibitory effect on the colony formation process of HepG2 cells within the tested dose range, suggesting that the matrix material of the microspheres did not significantly interfere with cell proliferation, providing an experimental basis for subsequent in vivo safety assessment.

FIGURE 4.

Experimental results obtained under different concentrations of Lu‐PLA‐g‐PAO‐Ms: (a) Representative images of colony formation assay; (b) Quantitative statistical analysis of colony formation assay; (c) Quantitative statistical results of CCK‐8 cell viability assay; (d) Quantitative analysis of cell migration rate from scratch wound healing assay; (e) Representative images of scratch wound healing (cell migration) assay; (f) Representative images of live/dead cell viability staining assay; (g) Results of in vitro blood coagulation assay; (h) Visual images of hemolysis assay; (i) Quantitative statistical analysis of hemolysis assay.

The CCK‐8 proliferation assay further confirmed the favorable cytocompatibility of the microspheres. After LO2 cells were co‐cultured with different concentrations of Lu‐PLA‐g‐PAO‐Ms for 72 h, the cell viability of each group exceeded 90%, the viability of HepG2 cells also remained above 90% at 48 h and 72 h, with no statistically significant difference from the control group (P>0.05, Figure 4c). As a control, treatment with free Lu³⁺ (50 mg/L LuCl₃) for 72 h significantly reduced the viability of HepG2 cells to 51.4% (P<0.001). This result was consistent with the conclusion from the in vitro release assay of Lu‐PLA‐g‐PAO‐Ms that the microspheres could stably chelate Lu, indicating that Lu‐PLA‐g‐PAO‐Ms exhibited favorable in vitro cytocompatibility and provided preliminary experimental support for subsequent in vivo biosafety evaluations.

The scratch wound healing assay showed no significant difference in cell migration rates between the control group and the groups treated with different concentrations of Lu‐PLA‐g‐PAO‐Ms at 24 and 48 h (P>0.05, Figure 4d and e), suggesting that the microspheres did not affect the intrinsic migratory capacity of HepG2 cells. The live/dead staining results further supported this conclusion. After co‐culturing with different concentrations of microspheres for 48 and 72 h, the cell population was dominated by calcein‐AM‐positive (green, viable) cells, whereas PI‐positive (red, dead) cells were extremely rare (Figure 4f), with no evidence of impaired cell membrane integrity.

Multiple in vitro assays, including plate colony formation, CCK‐8 proliferation, scratch wound healing, and live/dead cell staining assays, consistently demonstrated that Lu‐PLA‐g‐PAO‐Ms exhibited favorable cytocompatibility within the tested dose range.

Given that embolic carriers are in direct contact with blood during clinical TARE procedures, hemocompatibility is a prerequisite for their translational application. ²⁹ The hemocompatibility of Lu‐PLA‐g‐PAO‐Ms was systematically evaluated by measuring the in vitro coagulation time and hemolysis rate using whole blood from New Zealand White rabbits. All animal experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of our institution. The coagulation assay showed that after the co‐incubation of microsphere suspensions at different concentrations with rabbit whole blood, there was no statistically significant difference in the coagulation time between each experimental group and the normal saline control group (P>0.05, Figure 4g), indicating that Lu‐PLA‐g‐PAO‐Ms did not activate the coagulation system. The hemolysis assay showed that after the co‐incubation of microspheres at different concentrations with rabbit erythrocytes at 37°C for 2 h, the supernatant of each group was clear and transparent (Figure 4h), and the hemolysis rate of each group was below 5% (Figure 4i), meeting the requirements of ISO 10993–4 for blood‐contacting materials. ³⁰ This favorable hemocompatibility can effectively prevent microsphere‐induced adverse reactions such as thrombosis and hemolysis, ensure successful embolization of the microspheres into the tumor microvasculature, and guarantee the safety of TARE procedures.

3.3. Radiostability and efficacy of ¹⁷⁷Lu‐PLA‐g‐PAO microspheres

Because of the excellent in vitro performance of ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms, we constructed an orthotopic N1S1 HCC rat model to validate the in vivo radiostability, tumor targeting capability, and antitumor therapeutic efficacy of ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms (Figure 5a). An orthotopic HCC rat model was established via intrahepatic subcapsular injection of N1S1 cells, and successful tumor engraftment was confirmed by MRI. Tumor‐bearing rats were randomly divided into four groups (n = 6 per group): NS control, ¹⁷⁷LuCl₃ solution, blank PLA‐g‐PAO‐Ms embolization, and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms treatment, with all interventions administered via the hepatic artery. SPECT/CT imaging was performed on day 7 and day 14 post‐treatment to evaluate the in vivo distribution and radiostability of the microspheres (Figure 5d). Representative SPECT/CT images of the ¹⁷⁷LuCl₃ group showed significant systemic spread of radioactive signals throughout the body over time, with predominant ¹⁷⁷Lu accumulation in the skeleton. Conversely, representative SPECT/CT images of the ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group showed that radioactive signals were highly restricted to the tumor lesions, with no detectable off‐target radionuclide migration observed at both day 7 and day 14 post‐administration. This indicates that ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms possess excellent in vivo radiostability and tumor targeting capability, which is attributed to the stable chelation of PLA‐g‐PAO‐Ms with ¹⁷⁷Lu, effectively hindering off‐target diffusion of radionuclides and reducing the risk of unnecessary radiation exposure to non‐target organs and tissues. MRI was performed on day 7 and day 14 post‐treatment to monitor the tumor response to the different interventions (Figure 5b). Quantitative MRI analysis showed that tumors in the NS control group and ¹⁷⁷LuCl₃ group grew continuously over the observation period. Although a mild inhibitory trend was observed in the ¹⁷⁷LuCl₃ group, the difference was not statistically significant (P>0.05). Treatment with blank PLA‐g‐PAO‐Ms only partially inhibited tumor growth compared to the NS control and ¹⁷⁷LuCl₃ groups, which was attributed to the embolization effect of PLA‐g‐PAO‐Ms that blocked the tumor blood supply, thereby slowing tumor growth. Notably, ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms yielded a more remarkable anti‐tumor effect, with significant tumor growth suppression and even partial tumor regression observed on day 14 post‐treatment, exhibiting superior therapeutic efficacy compared to the blank embolization‐only group. The tumor volumes in the blank PLA‐g‐PAO‐Ms group and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group were significantly smaller than those in the NS control group on day 14 post‐treatment (Figure 5c, P<0.001).

FIGURE 5.

Antitumor efficacy and in vivo biological behavior evaluation after hepatic artery administration in rat N1S1 orthotopic hepatocellular carcinoma models. (a) Schematic diagram of the hepatic artery administration procedure and post‐treatment evaluation workflow, outlining the establishment strategy of rat orthotopic hepatocellular carcinoma model, the experimental grouping scheme for hepatic artery administration, as well as the serial imaging evaluation and biological detection procedures after administration. Created in https://BioRender.com. (b) MRI images of tumors in the Saline group, ¹⁷⁷LuCl₃ group, PLA‐g‐PAO‐Ms group, and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group on day 0, day 7, and d 14 after administration. Red circles indicate the tumor boundaries. (c) Tumor growth curves of the above four experimental groups after administration. (d) Representative SPECT images of the ¹⁷⁷LuCl₃ group and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group on day 5 and day 10 after administration. Red circles indicate the tissue uptake sites of ¹⁷⁷Lu. (e) In vivo biodistribution profile of the ¹⁷⁷LuCl₃ group and ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group in rat N1S1 orthotopic hepatocellular carcinoma models at 14 d after hepatic artery administration.

Biodistribution analysis performed on day 14 post‐treatment further verified the in vivo SPECT/CT imaging results (Figure 5e). The biodistribution results showed that tumor tissues in the ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group exhibited the highest radioactive uptake, reaching 8.79 ± 1.74 %ID /g, with radioactive counts significantly higher than those in normal tissues including of the liver (1 ± 0.54 %ID /g), heart (0.15 ± 0.11 %ID /g), spleen (0.65 ± 0.06 %ID/ g), lung (0.38 ± 0.03 %ID /g), kidney (0.22 ± 0.002 %ID /g), and blood (0.45 ± 0.03 %ID/ g) (P<0.001). In contrast, the radioactive uptake values of other non‐target organs and tissues in the ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms group remained extremely low, indicating no severe radiation damage to the normal tissues. However, the ¹⁷⁷LuCl₃ group showed only 0.30 ± 0.28 %ID/ g of radioactive uptake in tumor tissues, with systemic distribution of radionuclides and significant accumulation in non‐target organs, especially the bone, kidney and spleen. This was mainly attributed to the rapid clearance and non‐specific protein binding of ¹⁷⁷Lu, which led to nontargeted retention in normal organs. Therefore, ¹⁷⁷Lu‐PLA‐g‐PAO‐Ms can effectively avoid off‐target radiation damage to normal tissues in vivo and significantly improve the effectiveness and biosafety of in vivo TARE applications.

A comparative analysis with previously reported ¹⁷⁷Lu‐labeled microsphere systems further highlights the novelty and competitive advantages of this work. The ¹⁷⁷Lu‐PDA‐CS‐MgO microspheres reported by Xiao et al. (50–200 µm) exhibited moderate radiochemical stability (85.15 ± 0.91% in PBS and 88.65 ± 0.20% in FBS at 14 days) and a relatively large size distribution that may restrict distal microvascular penetration. ³⁰

4. CONCLUSIONS

This study focused on critical material challenges encountered in the clinical application of TARE. Currently, clinically used radioactive microspheres generally have inherent limitations, including non‐degradability, insufficient radionuclide labeling stability, and a lack of integrated theranostic capability. To address these issues, we developed PLA‐g‐PAO‐Ms and systematically evaluated their physicochemical properties, biosafety, and in vivo antitumor efficacy as ¹⁷⁷Lu‐labeled carriersfor TARE. Our results demonstrate that the amidoxime coordination sites on the microsphere surface enable efficient and stable chelation of ¹⁷⁷Lu, with a degradation behavior that matches well with the physical half‐life of ¹⁷⁷Lu. This system maintained excellent biocompatibility while exhibiting a significant inhibitory effect on tumor growth in hepatocellular carcinoma models. These findings indicate that PLA‐g‐PAO‐Ms represent a novel material design strategy and experimental basis for the construction of biodegradable multifunctional radionuclide carriers for TARE therapy. In terms of the preparation methodology, we established a fabrication strategy combining microfluidic fabrication, electron beam irradiation‐induced grafting, and amidoximation functionalization, which addressed key technical bottlenecks in the traditional preparation process of radioactive microspheres to a certain extent. Microfluidic technology enabled precise regulation of the particle size and morphology of the microspheres. The developed PLA‐g‐PAO‐Ms have a median particle size of 33.73±0.03 µm, with approximately 90% of the microspheres distributed in the size range of 20 to 80 µm, which meets the basic requirements for embolic microspheres in TARE therapy, while simultaneously balancing catheter deliverability and tumor microvascular embolization efficiency. Compared to traditional radionuclide labeling methods that rely on reactor neutron activation, we adopted electron‐beam‐induced graft polymerization to construct functional sites on the microsphere surface. This approach allows for controllable modification of the material surface under relatively mild conditions and avoids significant damage to the polymer chain structure. This method does not require the additional introduction of chemical initiators, which reduces the risk of impurity introduction to a certain extent, while preserving the inherent physicochemical properties and degradation characteristics of the polylactic acid matrix. The material characterization results further confirmed that the amidoxime groups can be stably anchored on the microsphere surface and provide effective coordination sites, thereby offering a structural foundation for the stable chelation of radionuclides and thus improving the stability of radionuclide labeling. Biodegradability is a key feature that distinguishes this carrier from clinically available non‐degradable ⁹⁰Y microspheres, and its degradation process was rationally designed to balance both therapeutic efficacy and biosafety. In vitro degradation experiments showed that PLA‐g‐PAO‐Ms initiated obvious degradation on approximately day 20 in a simulated physiological environment, with a mass loss rate of 72% by day 35. This degradation profile matched well with the physical half‐life of ¹⁷⁷Lu. After approximately three half‐lives of ¹⁷⁷Lu (approximately day 20), its radioactivity dropped to below 12.5% of the initial value, with the treatment‐related radiation energy essentially fully released, at which point the microspheres underwent gradual degradation. This characteristic maintains the structural stability of the microspheres during the radiotherapy window, thereby reducing the risk of systemic off‐target toxicity caused by premature radionuclide release. Meanwhile, it avoids the long‐term retention of microspheres in the liver tissue, reduces the incidence of potential complications such as long‐term vascular stenosis and chronic inflammation, and reserves a certain safety margin for subsequent repeated TARE treatments that may be required. ⁹⁰ The amidoxime groups on the surface of the PLA‐g‐PAO‐Ms endowed the material with a strong radionuclide‐chelating capacity and laid a foundation for functional expansion. The experimental results showed that the maximum chelating capacity of the microspheres for Lu³⁺ couldreach 33.83 mg/ g. Based on the specific activity of clinical‐grade ¹⁷⁷Lu, its theoretical radioloading capacity could be estimated to meet the dose requirements of a single TARE treatment and exhibits high drug loading potential. Stability evaluation results demonstrated that the radionuclide retention rate of ¹⁷⁷Lu‐labeled microspheres remained above 98.9% after 10 d of incubation in physiological media. In vivo SPECT/CT imaging and a 14‐day biodistribution study both confirmed that the radionuclide could be stably retained in the tumor lesion area, with no obvious off‐target migration observed, which verified the stability of the labeling system at both in vitro and In vivo levels. Compared to the limitation of mainstream clinical ⁹⁰Y microspheres that lack real‐time imaging capability, the gamma rays emitted by ¹⁷⁷Lu enabled non‐invasive visual monitoring during the treatment process via SPECT/CT. Meanwhile, the strong coordination ability of the amidoxime groups toward a variety of transition metal radionuclides endowed the microspheres with the potential to further construct a co‐labeling system for therapeutic and imaging radionuclides, providing new possibilities for the optimization of TARE treatment strategies in different clinical scenarios.

From the perspective of clinical translation, the ¹⁷⁷Lu‐labeled PLA‐g‐PAO‐Ms developed in this study are promising candidate carriers for TARE in unresectable HCC. There is a significant unmet therapeutic need for patients with intermediate and advanced HCC. Clinically used radioactive microspheres have certain limitations in terms of their safety and scope of application owing to their non‐degradability and the potential risk of radionuclide leakage. While preserving the synergistic effect of embolization and internal irradiation in TARE technology, the biodegradable microspheres prepared in this study overcome several limitations of existing carrier materials. These microspheres exhibited excellent biocompatibility, stable radionuclide retention capacity, and a definite antitumor effect in vivo, providing an experimental foundation for subsequent preclinical evaluation in large animal models and systematic toxicological studies, as well as a new material option for further optimization of treatment strategies for unresectable HCC.

This study has certain limitations that should be addressed in subsequent studies. First, this study was limited to short‐term efficacy and safety evaluations in a rat N1S1 orthotopic hepatocellular carcinoma model and has not yet been validated in large animal models that more closely mimic the clinical application scenario. The rabbit VX2 hepatocellular carcinoma model or canine normal liver model can better simulate the conditions of clinical interventional procedures and provide an important reference for evaluating catheter deliverability, dosimetry distribution, and long‐term safety. Second, this study did not include a direct comparison with widely used clinical commercial ⁹⁰Y microspheres; thus, it lacked a systematic assessment of the differences in efficacy and safety between the two. Third, this study primarily verified the chelation and therapeutic efficacy of ¹⁷⁷Lu, whereas the labeling efficiency and In vivo performance of the amidoxime coordination structure for other common TARE radionuclides, such as ¹⁶⁶Ho and ⁹⁰Y, require further investigation. Fourth, the In vivo observation period of this study was 14 d, and systematic data on the histological response, chronic inflammation, and long‐term systemic toxicity that may occur during the complete in vivo degradation cycle of the microspheres remain unavailable. Finally, although the current in vitro radiolabeling stability data demonstrated retention rates above 98.9% over a 10‐day period in PBS and 10% FBS, more comprehensive stability evaluations in additional physiological matrixes, such as human serum, plasma, and bile, were not performed, which could be valuable to further consolidate the translational reliability of the amidoxime chelation strategy.

CONFLICT OF INTEREST STATEMENT

All authors declare no conflict of interest.

ETHICAL STATEMENT

All animal experiments were performed in accordance with the guidelines of the Dongguan People's Hospital Laboratory Animal Welfare and Ethics Committee. The study protocol was reviewed and approved by the committee.

Supporting information

Supporting Information

DATA AVAILABILITY STATEMENT

All data generated and analyzed during this study are included in this published article.