Abstract

Background

   Breast cancer remains the most common malignancy among women worldwide.
   While traditional computed tomography (CT) scans and image-guided
   radiotherapy are widely used for breast cancer diagnosis and treatment,
   their efficacy is often limited.

Methods and results

   In this study, we successfully synthesized human serum albumin
   (HSA)-based KBiF[4] nanoclusters through a simple one-pot biomimetic
   mineralization strategy. Compared to the clinical contrast agent
   iohexol, KBiF[4]@HSA significantly enhances dual-energy CT (DECT)
   imaging contrast at high keV levels, offering improved diagnostic
   accuracy for breast cancer. Furthermore, KBiF[4]@HSA exhibits a
   remarkable ability to scavenge elevated glutathione (GSH) levels and
   promote reactive oxygen species (ROS) generation. When combined with
   radiotherapy, KBiF[4]@HSA substantially increases X-ray dose deposition
   at tumor sites, leading to enhanced DNA damage and suppression of
   breast cancer progression. Importantly, KBiF[4]@HSA demonstrates
   excellent biocompatibility in vivo, with no significant tissue damage
   or inflammation observed.

Conclusions

   This study presents a novel approach for the development of
   biocompatible DECT contrast agents and radiosensitizers, offering a
   promising strategy to enhance breast cancer diagnosis and treatment.
   However, the efficacy of this approach needs to be further validated
   across diverse breast cancer subtypes to ensure its broad
   applicability, which emphasizes the necessity for continued research to
   fully translate this innovative technology into clinical practice.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-025-03530-8.

   Keywords: Bismuth, Dual-energy CT, Contrast agents, Breast cancer,
   Radiotherapy

Graphical Abstract

   [42]graphic file with name 12951_2025_3530_Figa_HTML.jpg

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12951-025-03530-8.

Introduction

   Breast cancer is becoming the most frequent malignancy among women
   worldwide, but it is curable with early-stage, nonmetastatic diagnosis
   [[43]1]. With the development of imaging diagnosis technology, earlier
   and smaller breast cancers can be diagnosed early through advanced
   imaging tools, and the five-year survival rate of breast cancer has
   improved in the past decades [[44]2, [45]3]. Radiotherapy serves as a
   noninvasive approach for eradicating breast cancer and significantly
   diminishing tumor burden. Despite its high efficacy in the majority of
   breast cancers, the nonspecific nature of irradiation can lead to
   off-target toxicity affecting adjacent tissues. To address this issue,
   image-guided radiotherapy has been developed, which integrates cancer
   imaging with radiotherapy to provide crucial functional information for
   radiotherapy planning. Dual-energy computed tomography (DECT), as a
   groundbreaking CT imaging technique, facilitates material
   differentiation and improved diagnostic capabilities compared to
   conventional single-energy CT. Furthermore, DECT provides distinctive
   qualitative and quantitative information about tissue composition,
   which is particularly beneficial in image-guided radiotherapy
   [[46]4–[47]6]. However, iodine-based CT contrast agents (CAs),
   routinely used clinically, are inadequate for DECT imaging due to their
   low K-edge energy (33.2 keV) [[48]7, [49]8]. Consequently, there is an
   urgent requirement to develop high-performance CAs to fulfill the
   demands of DECT imaging in breast cancer diagnosis and image-guided
   radiotherapy.

   Radiosensitizers have been developed to amplify the effects of
   radiation within high-grade tumors, such as triple-negative breast
   cancer (TNBC) [[50]9]. However, the therapeutic efficacies of
   radiotherapy are often impaired by the radioresistance and antioxidant
   defense system relevant to elevated glutathione (GSH) within tumors
   [[51]10]. Recently, various strategies have been developed to improve
   the effectiveness of these sensitizers. For instance, novel synthetic
   compounds, such as phenethyl isothiocyanate [[52]11] and
   epigallocatechin gallate [[53]12], have been identified to specifically
   deplete GSH and overcome radioresistance in cancer cells. In addition,
   metallic nanoparticle-based radiosensitizers, such as gold
   nanoparticles, have gained attention due to their ability to increase
   the dose of radiation absorbed by tumor tissues, thereby improving the
   overall therapeutic index of radiotherapy [[54]13]. However, these
   radiosensitizers may increase the accumulation of heavy metals in the
   normal tissues and induce potential bio-toxicity [[55]14]. More
   recently, tumor-specific radiosensitizers have been developed to
   amplify the effects of radiation within cancer cells, which helps
   mitigate the risks associated with high radiation doses by improving
   the precision of tumor targeting and minimizing exposure to adjacent
   organs at risk.

   Bismuth (Bi), as an element with a high atomic number (Z = 83) and a
   high K-edge energy (K-edge = 90.5 keV), exhibits X-ray attenuation
   capacity that possesses outstanding radiosensitizing and constant DECT
   imaging capabilities at various X-ray energies [[56]15, [57]16]. For
   instance, the CT contrast of Bi is superior to iodine at higher energy
   range (80–150 keV), making it more suitable for DECT imaging. Over the
   past decades, the rapid advancement of nanotechnology has facilitated
   the development of various Bi-based CAs, including Bi [[58]17–[59]20],
   Bi[2]S[3] [[60]21–[61]23], Bi[2]Se[3] [[62]24–[63]27], Bi[2]O[3]
   [[64]28–[65]31], BiF[3] [[66]32–[67]34], BiOI [[68]35–[69]37] Bi-DTPA
   [[70]38–[71]41], Bi[2]WO[6] [[72]42–[73]44] and other Bi-based
   nanomaterials [[74]45–[75]47]. However, the synthesis of Bi-based CAs
   presents significant challenges, such as elevated temperatures, an
   oxygen-free atmosphere, and the use of toxic organic solvents [[76]48,
   [77]49]. Furthermore, post-synthetic surface modifications are
   generally required prior to biomedical applications, potentially
   elevating systemic toxicity risks. Additionally, these Bi-based CAs
   demonstrate limitations related to effective degradation in biological
   environments, which is essential for minimizing potential toxicity and
   issues of long-term retention. Looking toward clinical translation,
   focus should be put on biocompatible and biodegradable Bi-based CAs
   that are compatible with DECT imaging and radiosensitization. Human
   serum albumin (HSA), the most abundant protein in human plasma,
   demonstrates reversible binding capabilities with both organic
   molecules and metal ions [[78]50–[79]52]. Due to their superior
   biocompatibility and non-immunogenicity, HSA-based drug formulations,
   such as Abraxane, have been clinically employed in the treatment of
   breast cancer [[80]53–[81]56]. Recently, HSA-mediated biomimetic
   mineralization of inorganic nanoparticles has garnered significant
   interest owing to milder reaction conditions and high reproducibility
   [[82]57–[83]60].

Scheme 1.

   [84]Scheme 1
   [85]Open in a new tab

   Scheme of the synthesis and degradation of HSA-based KBiF[4]
   nanoclusters for DECT imaging and GSH-scavenging radiotherapy for
   breast cancer

   In this study, we present biocompatible KBiF[4] nanoclusters using the
   HSA-mediated bioinspired mineralization strategy. Compared to iohexol,
   the synthesized KBiF[4]@HSA demonstrates a significant enhancement in
   CT value and DECT imaging contrast at 180 keV, thereby improving the
   diagnostic accuracy for breast cancer. In addition, the precise tumor
   localization achieved through DECT imaging with these nanoclusters
   could enhance image-guided radiotherapy, thereby minimizing off-target
   toxicity in adjacent tissues. Moreover, these nanoclusters could be
   rapidly degraded by intracellular GSH in tumor cells, which could
   efficiently consume radioresistance-related GSH and sensitize tumor
   cells to radiotherapy. Furthermore, the KBiF[4]@HSA could deposit more
   X-ray energy, which significantly induces DNA damage and reactive
   oxygen species (ROS) generation under radiation. These nanoclusters
   demonstrated remarkable radiosensitizing properties, effectively
   suppressing tumor growth at reduced radiation doses. Importantly, the
   biodegradable KBiF[4]@HSA exhibited no detectable tissue damage or
   inflammation, underscoring their superior biocompatibility and minimal
   in vivo toxicity. Therefore, this study investigates biocompatible and
   biodegradable Bi-based CAs for DECT imaging and radiosensitizing via an
   HSA-mediated bioinspired mineralization strategy, which shows
   considerable potential for clinical translation.

Materials and methods

Materials

   Human serum albumin (HSA), Bismuth nitrate (Bi[NO[3]][3]·5H[2]O),
   Ethylene glycol (EG, 99%, Shanghai Chemical Reagent, China), and
   Potassium fluoride (KF·2H[2]O) were purchased from Aladdin Reagent Co.,
   Ltd. (Shanghai, China). Iohexol was obtained from Beijing BEILU
   Pharmaceutical Co., Ltd, China (Beijing, China). The deionized
   distilled water (ddH[2]O) was purified by a Millipore filtration
   system.

Synthesis of KBiF[4]@HSA

   KBiF[4]@HSA was synthesized by the one-pot method. 400 mg of HSA was
   dissolved in 5 mL of deionized water under a 37 °C water bath heating,
   and 5 mL of EG was added. After stirring for 1 min, 5 mL of
   Bi[NO[3]][3]·5H[2]O (242 mg) dissolved in EG was added, and 12.5 mL of
   KF·2H[2]O (1.13 g) dissolved in EG was added with stirring for 30 min.
   The mixture was stirred for 15 min, and the light-yellow solution was
   observed to gradually change to white. It was precipitated by adding 10
   mL of deionized water and separated by centrifugation (4000 rpm,
   10 min). Subsequently, the KBiF[4]@HSA was washed twice with deionized
   water and resuspended in deionized water to obtain KBiF[4]@HSA.

Characterization

   The morphology of the prepared materials was examined using
   transmission electron microscopy (TEM, FEI Tecnai F20) and scanning
   electron microscopy (SEM, Zeiss Gemini 300). Elemental distribution was
   analyzed through elemental mapping. X-ray photoelectron spectroscopy
   (XPS, Thermo Kalpha) was performed with a Thermo Kalpha instrument.
   X-ray diffraction profiles were obtained using an X’Pert PRO MPD
   system. Fourier transform infrared spectroscopy (FT-IR) spectra were
   recorded with a Nicolet iS 10. Absorption spectra were acquired using a
   Cary 100 UV − visible spectrophotometer (Agilent Technologies) at room
   temperature. Dynamic light scattering (Nano ZS, Malvern) was used to
   record the hydrodynamic size distribution.

GSH consumption

   KBiF[4]@HSA samples at the same concentrations were added to solutions
   containing 0 mM, 1 mM, 2 mM, 5 mM, 10 mM, and 15 mM of GSH. After 2 h
   of reaction, the KBiF[4]@HSA samples treated with varying GSH
   concentrations were placed in clear glass vials for comparative
   photography and subjected to UV-Vis spectroscopy after 24 h.
   Additionally, the KBiF[4]@HSA samples were analyzed using DLS and TEM
   at 0 min, 5 min, 15 min, 30 min, and 60 min. To detect intracellular
   GSH, 4T1 cells were incubated with KBiF[4]@HSA for a period of 24 h
   using the cellular glutathione peroxidase assay kit with DTNB. Cell
   lysates were then added to the test solution. The absorbance at 412 nm
   was measured with an enzyme marker to determine the concentration of
   GSH.

Cell culture

   Rat mammary carcinoma (SHZ-88) cells, normal rat kidney (NRK) cells,
   and 4T1 cells were provided by the Chinese Academy of Sciences Cell
   Bank (Shanghai, China) and cultured in RPMI-1640 medium (Gibco, Thermo
   Fisher, USA) with 10% fetal bovine serum (Gibco, Thermo Fisher, USA)
   and 1% penicillin-streptomycin (Gibco, Thermo Fisher, USA). The fresh
   medium was changed every two days. Cells were washed with phosphate
   buffer solution (Biosharp, China), and cells were digested with 0.25%
   trypsin (Gibco, ThermoFisher, USA). The cells were incubated at 37 °C,
   5% CO[2] incubator.

Cell viability test

   To assess the safety of KBiF[4]@HSA, NRK cells, SHZ-88 cells or 4T1
   cells were seeded in 96-well plates. After 24 h, the cells were exposed
   to media containing varying concentrations of KBiF[4]@HSA and incubated
   at 37 °C with 5% CO[2] for an additional 24 h. Cell viability was then
   measured using the CCK-8 assay.

Hemolysis test

   Healthy rat blood was mixed with phosphate-buffered saline (PBS) to
   create a 2% red blood cell suspension, then treated with varying
   concentrations of KBiF[4]@HSA and incubated at 37 °C for 3 h. The blood
   was centrifuged at 5000 rpm for 10 min at 4 °C. The supernatant was
   then transferred to a 96-well plate, and its UV absorbance was measured
   at 540 nm. For the calculation of hemolysis, PBS at pH 7.4 served as
   the negative control, while 2% red blood cell lysate was the positive
   control:

   Hemolysis =(𝐴𝑠−𝐴𝑛)/(𝐴𝑝−𝐴𝑛)×100%.

   The absorbance value of the sample is denoted as As, while the
   absorbance value of the positive control is represented as Ap, and the
   absorbance of the negative control is indicated as An.

Clone formation assay

   To assess the radiosensitizing effects of KBiF[4]@HSA, 4T1 cells were
   inoculated into 6-well plates, respectively, and cultured in RPMI-1640
   medium overnight. Next, the cells were incubated with KBiF[4]@HSA
   co-incubation for a further 4 h, then irradiated with X-rays and
   cultured in normal medium. After incubation for 7 days, cells were
   fixed by paraformaldehyde (4%) and stained with crystal violet.

Detection of intracellular GSH levels

   4T1 cells in logarithmic growth phase (5 × 10^5 cells/well) were seeded
   into 6-well plates and dynamically monitored for cell adhesion under
   conditions of 37 °C and 5% CO₂. When the cell confluence reached 80%,
   complete medium containing KBiF[4]@HSA was added for co-incubation for
   6 h. Subsequently, the medium was replaced with fresh medium, and the
   experimental group was exposed to X-rays (4 Gy) while the control group
   was not. After treatment, the cells were cultured for 24 h. The cells
   were then rinsed three times with PBS solution, incubated with cell
   lysis buffer for 30 min, scraped off, and centrifuged at 12,000 rpm for
   15 min at 4 °C. The supernatant was collected as the sample to be
   tested. The protein concentration of the samples was determined using
   the BCA method. A 96-well plate was set up with blank control and
   experimental groups according to the instructions of the Cellular
   Glutathione Peroxidase Assay Kit with DTNB (Beyotime, S0057S).
   Glutathione peroxidase assay buffer, 10 mM GSH solution, 15 mM peroxide
   reagent solution, and the samples to be tested were added sequentially
   and mixed well. After incubating at room temperature for 30 min, 6.6 µL
   of DTNB solution was added to each well, mixed well, and incubated for
   10 min. The absorbance was measured at a wavelength of 412 nm using a
   microplate reader, and the glutathione peroxidase activity was
   calculated.

Intracellular ROS detection

   Intracellular ROS production was detected by
   2’,7’-dichlorodihydrofluorescein diacetate (DCFH-DA). The 4T1 cells
   were co-incubated with KBiF[4]@HSA for 12 h. After washing with PBS for
   3 times, the cells of different treatments with/without X-ray
   irradiation were incubated with a DCFH-DA probe for 15 min, washed with
   PBS and imaged with an inverted fluorescence microscope. This study
   successfully performed quantitative analysis of fluorescence intensity
   using the ImageJ software.

Two-photon digital scanned light-sheet microscopy

   4T1 cells were plated in a 35 mm petri dish and incubated for 24 h. To
   study the distribution within cells, after a 4-hour incubation with
   KBiF[4]@HSA, cells were additionally stained with mito-tracker green
   (20 nM) or lyso-tracker green (50 nM) following standard procedures.
   After washing with PBS twice, cells were observed using the two-photon
   digital scanned light-sheet microscopy.

Live/dead cell staining

   The cells that adhered well were cultured in 24-well plates with or
   without KBiF[4]@HSA for 24 h. Subsequently, cells for different
   treatments were irradiated with or without X-rays. After this period,
   the cells were incubated with calcein acetoxymethyl ester (calcein-AM)
   and propidium iodide (PI) for 30 min at 37 °C in the dark.
   Intracellular fluorescence was then observed using an inverted
   fluorescence microscope.

Animal study

   All animal studies were approved by the Ethics Committee of Xinhua
   Hospital Affiliated to Shanghai Jiao Tong University School of
   Medicine. Female Sprague-Dawley (SD) rats (4–5 weeks) were supplied by
   the Shanghai Experimental Animal Center and housed in a specific
   pathogen-free environment. Mammary carcinoma tumors in situ were
   established by injecting 2 × 10^7 SHZ-88 cells between the right
   axillary mammary fat pads of each rat. Tumor generally formed about 1
   week after cell injection. DECT imaging experiments were performed when
   the breast in situ tumor reached a diameter of about 100 mm. In
   addition, 4-week-old female BALB/c mice were used for the radiotherapy
   experiment. Firstly, 4T1 cells (1 × 10^6) were subcutaneously injected
   into the right side of each mouse, and tumor formed after ten days.
   Then, these mice were randomly divided into four groups: (1) PBS, (2)
   KBiF[4]@HSA, (3) PBS + X, (4) KBiF[4]@HSA + X. After injection of
   KBiF[4]@HSA into the tumors, the mice were exposed to X-rays. X-ray
   biological irradiators were obtained using RAD SOURCE RS 2000. After
   different treatments, measurements of the body weights and tumor
   volumes for mice were conducted every 2 days. The tumor volumes were
   calculated by the formula: V = (length × width^2)/2.

DECT imaging in vitro

   A DECT scanner (SIEMENS SOMATOM FORCE) was employed to collect HU
   values and CT images using split-filter mode at 120 kVp and
   sequential-spiral mode at 80 keV and 150 keV for various concentrations
   of iohexol and KBiF[4]@HSA. Data were analyzed to generate
   single-energy CT images and X-ray attenuation curves. The imaging
   parameters were set as follows: tube voltages of 80 kVp and 150 kVp, a
   section thickness of 0.5 mm, a field of view (FOV) of 70 mm, and a
   matrix size of 512 × 512.

DECT imaging in vivo

   DECT imaging of SD rats prior to injection was captured using a DECT
   system, followed by scanning imaging 5 min after injection within the
   in situ mammary tumors of SD rats. Prior to injection, vascular and
   cardiac imaging of SD rats was conducted using a DECT system.
   Subsequently, an indwelling needle was placed in the tail vein of the
   rat, and a whole-body DECT scan was performed immediately following the
   injection of KBiF[4]@HSA.

Histology examination

   The rats were split into two groups at random: one group was given a
   tail vein injection of PBS, while the other group received KBiF[4]@HSA.
   After a period of 7 days, major organs, including the heart, liver,
   spleen, lungs, kidneys, and tumor, were collected for hematological
   analysis, hematoxylin and eosin (H&E) staining. In addition, mice in
   radiotherapy experiment were randomized into four groups: (1) PBS, (2)
   KBiF[4]@HSA, (3) PBS + X, and (4) KBiF[4]@HSA + X. After 14 days,
   tumors were harvested and subjected to H&E staining, TUNEL staining,
   Ki67 staining, and γ-H2AX staining.

RNA sequencing methods for transcriptome analysis

   From cryopreserved murine tumor tissues, total RNA was extracted using
   TRIzol reagent (Invitrogen, CA, USA), and its purity and concentration
   were measured afterward. Using the VAHTS Universal V6 RNA-seq library
   prep kit, libraries were created and sequenced on an Illumina Novaseq
   6000 platform. DESeq2 was used for differential expression analysis,
   with significant gene expression differences defined by a Q
   value < 0.05 and a fold change > 2. R (v3.2.0) was employed to examine
   Gene Ontology (GO) terms and assess functional enrichment, covering
   biological processes, molecular functions, and cellular components.
   Transcriptome sequencing was conducted by OE Biotech Co., Ltd.
   (Shanghai, China).

Assessment of blood biochemical parameters

   The serums were obtained from whole blood after centrifugation at a
   speed of 3240 rpm for 15 min. The serum biochemical levels, including
   alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate
   aminotransferase (AST), albumin (ALB), serum total protein (TP),
   globulin (GLOB), albumin-globulin ratio (A/G), triglyceride (TG),
   creatinine (CREA), urea, calcium (Ca), potassium (K), phosphate (P),
   sodium (Na) and chloride (Cl), were assayed by an automatic biochemical
   analyzer (C501, Cobas, USA).

Statistical analysis

   Statistical analysis of the experimental data was performed using the
   one-way ANOVA, two-way ANOVA and three-way ANOVA followed by post hoc
   test. In detail, a minimum of three repetitions was conducted for each
   experiment. The levels of statistical significance for the P values
   were represented as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.005
   and ****, P < 0.0001.

Results and discussion

The synthesis and characterization of KBiF[4]@HSA

   The KBiF[4] nanoclusters were synthesized in a one-pot synthesis method
   using the HSA-mediated bioinspired mineralization strategy. First, HSA
   was dissolved in EG followed by the addition of Bi[NO[3]][3]·5H[2]O at
   37 °C. Carboxyl groups exhibit a strong attraction to Bi³⁺ ions,
   enabling HSA to firmly anchor them. After stirring for 30 min,
   KF·2H[2]O was added and reacted for 15 min to produce a milky white
   solution as KBiF[4]@HSA. After centrifugal purification, the
   KBiF[4]@HSA were characterized by TEM and SEM images, and the results
   showed that the prepared KBiF[4]@HSA is monodisperse with an average
   size of ∼100 nm (Fig. [86]1A and B). As shown in Fig. [87]1C, elemental
   mapping analysis further indicated the existence and the homogeneous
   distribution of Bi, F, K, C, N, O, and S elements in the KBiF[4]@HSA,
   suggesting the KBiF[4] nanoclusters are generated with HSA.
   Furthermore, the atomic force microscope (AFM) images of the
   KBiF[4]@HSA revealed a spherical shape with a height of 139 nm
   (Fig. [88]1D). The maximum height of KBiF[4]@HSA in the
   three-dimensional AFM image was around 140 nm, but most of the
   particles were also present in small uniform nanoparticles of 100 nm
   and below.

Fig. 1.

   [89]Fig. 1
   [90]Open in a new tab

   (A) TEM image of KBiF[4]@HSA (scale bar: 200 nm). (B) SEM image of
   KBiF[4]@HSA. (C) EDX mapping of KBiF[4]@HSA. (scale bar: 0.5 μm). (D)
   AFM images of KBiF[4]@HSA. (E) TEM images and DLS analysis of
   KBiF[4]@HSA at different timepoints during reaction (scale bar: 200 nm)

   The formation process of KBiF[4]@HSA was explored by TEM images and DLS
   measurements at reaction times of 1, 5, 10, 15, and 20 min. During the
   preparation process, it was found that the uniform and monodispersed
   KBiF[4]@HSA was created in 1 min, resulting in spherical nanoparticles
   with an average size of approximately 48 nm (Figure [91]S1). Moreover,
   the TEM images showed that KBiF[4]@HSA at 5 min are uniformly dispersed
   particles of about 70 nm, indicating the size of KBiF[4]@HSA increased
   continuously during the reaction and the formation of KBiF[4]@HSA was
   fast (Fig. [92]1E). Interestingly, a significant change in morphology
   from sphere to cluster was evidenced by the TEM image of KBiF[4]@HSA at
   10 min. Furthermore, DLS analysis showed KBiF[4]@HSA at 10 min has an
   average hydrodynamic size of 222 nm, and the aggregation may be due to
   the rapid formation and untimely dispersion of KBiF[4]@HSA. With the
   end of the reaction, the KBiF[4]@HSA dispersed again at 15 min and
   remained unchanged at 20 min. To evaluate the long-term stability of
   KBiF[4]@HSA, we conducted a 14-day colloidal stability study after
   synthesis. As consolidated in Supplementary Figure [93]S2, the hydrated
   diameters of KBiF[4]@HSA remained stable at 79.4 ± 11.61 nm (Day 1) to
   80.73 ± 21.88 nm (Day 10), with a polydispersity index (PDI) of
   0.12–0.18 across all timepoints. Additionally, the hydrated diameter
   was measured to be 75.83 ± 13.39 on day 3, 80.73 ± 21.88 on day 5, and
   76.4 ± 11.61 on day 7. This confirms excellent dispersion stability for
   at least 10 days. Therefore, KBiF[4]@HSA could be efficiently
   synthesized via a simple one-pot HSA-mediated biomimetic mineralization
   method.

   The formation of the KBiF[4]@HSA was verified through XPS, XRD and
   FT-IR spectroscopy. The presence of the primary elements Bi, F, and K
   was confirmed by XPS analysis (Fig. [94]2A). XPS survey spectrum
   exhibited characteristic peaks corresponding to F 1s, K 2p, Bi 4f, C
   1s, N 1s, and O 1s, which reveals the integration of KBiF[4] and HSA.
   In the high-resolution XPS spectrum, peaks at 164 eV and 158.65 eV
   confirmed the presence of Bi^3+ in the KBiF[4] nanoclusters
   (Fig. [95]2B), while the peak at 682.7 eV signified the presence of F
   1s (Fig. [96]2C). In addition, the peak at 291.55 eV indicated the
   presence of K 2p, and the peak at 283.75 eV confirmed the presence of C
   1s (Fig. [97]2D). Furthermore, absorbance spectra of the KBiF[4]@HSA
   demonstrated an increased absorbance within the 500–900 nm range with
   the increasing KBiF[4]@HSA concentration (Fig. [98]2E). Moreover, the
   peaks in the XRD patterns of the lyophilized KBiF[4]@HSA powder aligned
   with the characteristic peaks of the standard KBiF[4]@HSA crystal
   structure (JCPDS 48–0667), providing evidence for the successful
   synthesis of KBiF[4]@HSA (Fig. [99]2F). The XRD elemental distribution
   spectra demonstrated that the predominant elemental composition of
   KBiF[4]@HSA comprised K, Bi, and F in an atomic ratio of approximately
   1.58: 1.43: 7.64, closely aligning with the stoichiometric ratio of the
   molecular formula (Fig. [100]2G and Figure [101]S3). FT-IR spectroscopy
   confirmed the modification of HSA within KBiF[4]@HSA, indicating
   enhanced biocompatibility (Fig. [102]2H).

Fig. 2.

   [103]Fig. 2
   [104]Open in a new tab

   (A) High-resolution Bi 4f spectrum of KBiF[4]@HSA. (B) High-resolution
   F 1s spectrum of KBiF[4]@HSA. (C) High-resolution K 2p and C 1s
   spectrum of KBiF[4]@HSA. (D) XPS survey spectra of KBiF[4]@HSA. (E)
   UV − vis absorption spectra of KBiF[4]@HSA. (F) XRD pattern of
   KBiF[4]@HSA. (G) Spectra of the main elements of KBiF[4]@HSA. (H) FT-IR
   of KBiF[4]@HSA. (I) TEM of KBiF[4]@HSA incubated with GSH for 0 min,
   5 min, 15 min, 30 min, 60 min (scale bar: 200 nm). (J) DLS analysis and
   photographs of KBiF[4]@HSA with different concentrations of GSH

The GSH-responsive degradation of KBiF[4]@HSA

   Recently, studies have demonstrated that GSH concentrations in tumor
   tissue are 100–1000 times higher than those in blood and normal
   tissues, with notable intracellular GSH concentrations in tumor cells
   ranging from 1 to 15 mM [[105]49]. The GSH-responsive degradation
   behavior of KBiF[4]@HSA was investigated using TEM, DLS, and UV − vis
   absorption following incubation with GSH-containing PBS. TEM analysis
   revealed minimal morphological changes in KBiF[4]@HSA during the
   initial 15 min. However, upon extension with prolonged incubation
   beginning at 30 min, a progressive degradation of KBiF[4]@HSA was
   detected (Fig. [106]2I). To further verify the degradation process, 1
   mM of KBiF[4]@HSA was exposed to varying concentrations of GSH (0 mM, 1
   mM, 2 mM, 5 mM, 10 mM, and 15 mM) for 2 h. The results demonstrated a
   transition of KBiF[4]@HSA from a milky-white suspension to a clarified
   solution with increasing GSH concentrations. As depicted in
   Fig. [107]2J, a GSH concentration of 15 mM resulted in a reduction of
   the particle size to approximately 10 nm, which may facilitate the
   excretion of KBiF[4]@HSA in vivo. Furthermore, the increased absorption
   of KBiF[4]@HSA after incubation with GSH indicated that GSH may degrade
   KBiF[4]@HSA into bismuth ions and form multicoordinated GSH-bismuth
   chelates, as further evidenced by the displacement of absorption peaks
   (Figure [108]S4A and B). Therefore, the degradable properties may not
   only improve imaging outcomes, but also enhance the biocompatibility
   and safety profile, which is ultimately eliminated through the body’s
   metabolic processes.

DECT-guided radiotherapy in vitro

   To evaluate the biocompatibility of KBiF[4]@HSA, the viability of NRK
   cells was assessed at varying concentrations of KBiF[4]@HSA using the
   CCK-8 assay. The results demonstrated cell survival rates exceeding 80%
   following incubation with KBiF[4]@HSA, suggesting excellent
   biocompatibility in normal cells (Fig. [109]3A). Moreover, given the
   potential of nanomaterials to induce acute hemolysis upon direct
   contact with blood, the hemolytic activity of KBiF[4]@HSA was further
   evaluated. Rat erythrocytes were incubated with varying concentrations
   of KBiF[4]@HSA, and hemoglobin release was quantified by measuring
   absorbance at 540 nm. After incubation for 24 h, the positive control
   group, which utilized ddH[2]O as the erythrocyte dispersion medium,
   demonstrated significant hemolysis, as indicated by the presence of a
   red supernatant (Fig. [110]3B). In contrast, the negative control
   group, which employed PBS, along with the groups treated with different
   concentrations of KBiF[4]@HSA, showed no visible hemolysis. In
   addition, to enhance the accuracy of hemolysis quantification, the
   supernatants from all groups were subjected to centrifugation. After
   3 h, hemoglobin absorbance was measured using an enzyme marker to
   evaluate the degree of erythrocyte hemolysis. The positive controls
   exhibited nearly complete hemolysis, whereas negative controls and the
   KBiF[4]@HSA group showed no hemolysis, which indicates that KBiF[4]@HSA
   exhibits favorable in vivo safety characteristics for DECT imaging and
   radiotherapy.

Fig. 3.

   [111]Fig. 3
   [112]Open in a new tab

   (A) Cell viability of SHZ-88 cells and NRK cells after incubation with
   KBiF[4]@HSA for 24 h. (B) Hemolysis experiment of KBiF[4]@HSA. (C) Cell
   viability of 4T1 cells after incubation with KBiF[4]@HSA for 24 h. (D)
   Typical images of colony formation of 4T1 cells treated with
   radiotherapy. (E) Two-photon digital scanned light-sheet microscopy
   images of 4T1 cells with mito-tracker and lyso-tracker. (F)
   Fluorescence images of 4T1 cells stained with DCFH-DA were taken to
   observe intracellular ROS after treatments (scale bar: 200 μm). (G)
   Changes in GSH levels in 4T1 cells after PBS or KBiF[4]@HSA treatments.
   (H) Fluorescence images of 4T1 cells stained with calcein-AM/PI
   following diverse treatments. (I) Fluorescence quantitative analysis of
   calcein-AM/PI

   The potential therapeutic efficacy of KBiF[4]@HSA as a radiosensitizer
   was further investigated. The viability of 4T1 cells treated with
   KBiF[4]@HSA and radiotherapy was assessed using the CCK-8 assay. After
   24 h of co-incubation with KBiF[4]@HSA alone, the cell survival rate of
   4T1 cells was 80%. Notably, when KBiF[4]@HSA was combined with X-ray
   treatment, substantial cytotoxicity was observed (Fig. [113]3C). In
   addition, the clonal proliferation assays illustrated that KBiF[4]@HSA
   alone did not significantly impact cell colony formation
   (Fig. [114]3D). Conversely, KBiF[4]@HSA and X-rays exposure led to a
   marked reduction in cell colony formation. Moreover, as shown in
   Fig. [115]3E, the red fluorescence of squaraine dye (SQ) labeled on
   KBiF[4]@HSA appeared in cells, exhibiting mitochondria-localization
   (Pearson’s correlation coefficient, PCC: 0.73 ± 0.03) and
   lysosome-localization tendency (PCC: 0.85 ± 0.03). Furthermore,
   utilizing DCFH-DA as an intracellular ROS indicator, cells treated with
   PBS and X-rays exhibited mild fluorescence, suggesting a limited ROS
   generation by PBS + X treatment (Fig. [116]3F). In contrast, the
   KBiF[4]@HSA + X group demonstrated the most intense fluorescence,
   providing direct evidence of the radiosensitizing effects of
   KBiF[4]@HSA. Notably, cells treated with KBiF[4]@HSA alone also
   exhibited strong fluorescence, indicating ROS generation. To elucidate
   the underlying mechanism, GSH levels were subsequently measured,
   revealing a significant decrease in the KBiF[4]@HSA group
   (Fig. [117]3G). These findings indicate that KBiF[4]@HSA can
   effectively scavenge the elevated GSH and disrupt the redox balance in
   cancer cells, resulting in increased ROS levels. In addition,
   calcein-AM/PI staining revealed that the majority of cells in PBS and
   KBiF[4]@HSA groups, suggesting robust cell viability and confirming the
   safety profile of KBiF[4]@HSA (Fig. [118]3H). Notably, the
   KBiF[4]@HSA + X group had significant tumor cell death compared to
   irradiation alone. This conclusion is supported by fluorescence
   quantitative analysis (Fig. [119]3I).

In vitro DECT imaging with KBiF[4]@HSA

   The DECT imaging efficacy of KBiF[4]@HSA was evaluated, with iohexol
   clinical CAs utilized as a comparative control. The KBiF[4]@HSA and
   iohexol at different concentrations (0, 0.3125, 0.625, 1.25, 2.5, 5,
   and 10 mM) were placed in tubes and scanned by a CT imaging system. At
   the same concentration of radioactive elements, the CT signal of
   KBiF[4]@HSA was stronger than that of iohexol with increasing energy,
   especially at high keV energy imaging, where KBiF[4]@HSA maintained
   stable imaging. The monochromatic image showed that the CT values for
   both KBiF[4]@HSA and iohexol decrease as keV energy increases. However,
   the decrease in KBiF[4]@HSA was significantly slower than that of
   iohexol (Fig. [120]4A). This disparity can be attributed to the
   enhanced X-ray attenuation characteristics of bismuth elements in
   comparison to iodine elements at elevated energy levels. Monochrome
   energy images displayed across the energy range of 40–180 keV showed
   that the signal from iohexol decreased progressively with increasing
   energy. In contrast, the CT signal of KBiF[4]@HSA consistently
   exhibited greater intensity and persistence compared to iohexol across
   the entire spectrum of energy levels. This finding aligns with the
   trends observed in the Hounsfield unit (HU) curves. In comparison, the
   CT signal of KBiF[4]@HSA was persistently higher and more potent than
   iohexol across the entire energy levels. This findings align with the
   HU curves (Fig. [121]4B-F). In higher single-energy images, KBiF[4]@HSA
   shows superior X-ray attenuation properties compared to iohexol,
   indicating the advanced DECT imaging capabilities of KBiF[4]@HSA for
   higher diagnostic precision and contrast resolution.

Fig. 4.

   [122]Fig. 4
   [123]Open in a new tab

   (A) In vitro CT imaging at different keV and energy spectrum curves of
   KBiF[4]@HSA and iohexol. (B-F) In vitro CT imaging of KBiF[4]@HSA and
   iohexol at (B) 40 keV, (C) 60 keV, (D) 80 keV, (E) 150 keV, (F)
   180 keV. The data are shown as the mean ± standard deviation (SD), as
   calculated using two-way ANOVA followed by post hoc tests with
   *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001

In vivo Dect imaging with KBiF[4]@HSA

   To evaluate the DECT imaging efficacy of KBiF[4]@HSA, SHZ-88 cells were
   injected into the mammary gland of 5-week-old SD rats to establish in
   situ breast cancer tumors. Then, KBiF[4]@HSA (10 mM, 20 µL) and iohexol
   (10 mM, 20 µL) were administered via intratumoral injection. The CT
   signals of the tumor prior to injection were utilized as a control, and
   subsequent CT signals of the tumor were monitored post-injection. As
   depicted in Fig. [124]5A and B, the signal intensity within the tumor
   region exhibited significant variation in the monochrome energy image
   before and after injection. Specifically, the CT value of the tumor
   imaged before injection was approximately 56.07 HU at 80 keV and 34.06
   HU at 150 keV (Figure [125]S5A). Subsequent to the administration of
   KBiF[4]@HSA, there was a significant enhancement in the CT signal
   within the tumor region at 80 keV, with a measured CT value of 491.35
   HU, representing an approximate 9-fold increase compared to the
   pre-injection imaging. At 150 keV, the tumor region exhibited a high CT
   signal, with a recorded CT value of 368.66 HU, and this elevated signal
   intensity was maintained as the X-ray energy increased (Figure
   [126]S5B).

Fig. 5.

   [127]Fig. 5
   [128]Open in a new tab

   (A) DECT imaging of a tumor with iohexol at different radiation
   energies. (B) DECT imaging of a tumor with KBiF[4]@HSA at different
   radiation energies. (C) DECT imaging of the tumor at different
   anatomical positions with KBiF[4]@HSA and iohexol. (D) DECT imaging of
   the tumor at the effective atomic numbers (Z[eff]) with KBiF[4]@HSA.
   (E) DECT values at different keV in the tumor region with KBiF[4]@HSA.
   (The blue circle indicates the tumor location.)

   In contrast, for tumors subjected to intratumoral injection with
   iohexol, the CT value of the tumor region prior to injection was 41.29
   HU at 80 keV and 29.11 HU at 150 keV (Figure [129]S6A). After the
   injection, the CT value of the tumor area rose to about double its
   pre-injection level. Specifically, at 80 keV, the tumor region showed
   slight visualization with a measured CT value of 127.12 HU. The CT
   signal of the tumor was notably attenuated at 150 keV, with no
   observable contrast-enhanced CT signal upon further increases in X-ray
   energy, as evidenced by a measured CT value of 47.96 HU (Figure
   [130]S6B). Quantitative evaluation and statistical analysis of in vivo
   imaging signals were conducted for KBiF[4]@HSA and iohexol, further
   confirming the increased signal intensity of KBiF[4]@HSA compared to
   iohexol (Figure [131]S7-[132]8). These findings suggest that iohexol is
   effective exclusively in contrast-enhanced imaging at lower kilovoltage
   levels, with its contrast-enhancing properties diminishing
   progressively as X-ray energy levels increase. The observed phenomenon
   may stem from the inherent characteristics of iohexol. As a
   water-soluble contrast agent, its low density and rapid diffusion
   following in vivo injection prevent adequate contrast enhancement in
   DECT scans at low dosage levels. Conversely, KBiF[4]@HSA demonstrates
   elevated K-edge energy, increased density, and enhanced stability,
   along with heightened sensitivity.

   Subsequently, a comparative analysis of CT signal intensities between
   KBiF[4]@HSA and iohexol was conducted in tumor regions across various
   anatomical positions. KBiF[4]@HSA exhibited a significantly higher CT
   signal intensity than iohexol (Fig. [133]5C and D). For image
   reconstruction, the Z[eff] mode was selected using the dual-energy mode
   on the Siemens postprocessing workstation. KBiF[4]@HSA incorporates the
   high atomic number element Bi, which is capable of producing
   significant signals at the tumor site post-injection, as evidenced with
   a measured Z[eff] value of 10.13 HU (Fig. [134]5E and [135]S9). These
   findings indicate that KBiF[4]@HSA exhibits consistent X-ray
   attenuation properties, thereby establishing its reliability as a
   contrast agent for DECT imaging in the context of disease diagnosis.

   To explore the imaging capabilities of KBiF[4]@HSA, their efficacy in
   cardiac imaging was also evaluated. Firstly, vasculature imaging was
   performed on 5-week-old healthy rats via tail-vein injection, using the
   same concentration of iohexol and KBiF[4]@HSA (10 mM, 1.5 mL) for
   comparison. In rats injected with KBiF[4]@HSA, a significant increase
   in the CT signal of the abdominal aorta was observed, with an increased
   CT signal as the blood circulated to the cardiac region (Figure
   [136]S10A). Notably, upon increasing the energy of X-ray radiation, the
   cardiac CT signal intensity remained consistent across the energy range
   of 40–180 keV. In contrast, rats injected with iohexol exhibited
   diminished CT signals, lacking contrast enhancement at elevated
   radiation energies (Figure [137]S10B). Quantitative evaluation and
   statistical analysis of in vivo cardiac imaging signals were conducted
   for KBiF[4]@HSA and iohexol, further confirming the increased signal
   intensity of KBiF[4]@HSA compared to iohexol (Figure [138]S11-[139]12).
   These findings indicate that vascular contrast enhancement can be
   effectively achieved using KBiF[4]@HSA at low concentrations, whereas
   iohexol necessitates a significantly higher concentration to achieve
   visualization, which consequently elevates its toxicity within the
   body.

   Furthermore, the imaging characteristics of KBiF[4]@HSA within the
   vasculature were evaluated at multiple anatomical sites and Z[eff]
   values both pre- and post-injection. The results demonstrated a
   significant enhancement of the CT signal within the vascular system
   (Figure [140]S10C and D). This finding highlights the potential of
   KBiF[4]@HSA for imaging ventricles and major blood vessels at low
   concentrations, thereby achieving improved contrast. This capability is
   anticipated to have potential applications in the imaging of vascular
   tumors, arterial stenosis, and other related pathologies.

The efficiency of radiotherapy with KBiF[4]@HSA in vivo

   To assess in vivo radiotherapy efficiency, 4T1 cells were injected into
   the dorsal skin of mice to establish a subcutaneous tumor model.
   According to the procedures shown in Fig. [141]6A, the mice underwent
   two sessions of irradiation, with or without 6 Gy of X-ray exposure at
   the tumor sites, after intratumoral injection of either PBS or
   KBiF[4]@HSA. As shown in Fig. [142]6B, the KBiF[4]@HSA and PBS + X
   groups did not exhibit significant tumor growth inhibition compared to
   the PBS group. Significantly, tumor growth was delayed in the
   KBiF[4]@HSA + X group. In addition, no significant weight changes were
   detected across any of the treatment groups (Fig. [143]6C). At the end
   of therapy, the excellent therapeutic efficacy in the KBiF[4]@HSA + X
   group was further confirmed by the quantitative tumor weights
   (Fig. [144]6D). To elucidate the potential therapeutic mechanism of
   KBiF[4]@HSA in radiotherapy at the genetic level, a transcriptome
   analysis was performed to examine the expression of messenger RNA
   (mRNA). The analysis identified 307 differentially expressed genes
   (DEGs) when comparing the KBiF[4]@HSA group with the PBS group,
   including 240 upregulated and 67 downregulated genes. Conversely, when
   comparing the KBiF[4]@HSA group, the KBiF[4]@HSA + X group showed 288
   gene differences, including 46 upregulated and 242 downregulated genes
   (Figure [145]S13A and B). Furthermore, the heatmap analysis
   demonstrated a notable difference in mRNA expression (Fig. [146]6E and
   F). Considering the variability in RNA sequencing outcomes, we
   performed Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene
   Ontology (GO)analysis to attain a more comprehensive understanding of
   the changes in the biological functions of mRNAs and their related
   pathways (Fig. [147]6G and H). According to the KEGG enrichment
   analysis, the DEGs are chiefly enriched in the JAK-STAT signaling
   pathway, Th1 and Th2 cell differentiation, and the IL-17 signaling
   pathway.

Fig. 6.

   [148]Fig. 6
   [149]Open in a new tab

   Radiosensitizing effects of KBiF[4]@HSA in vivo. (A) Flow chart of the
   radiotherapy experiment in vivo. (B) Tumor growth curves of 4T1-bearing
   tumor mice. (C) Body weights and (D) tumor weights of mice after
   various treatments. (E) Heatmap diagram of differentially expressed
   mRNAs screened from PBS and KBiF[4]@HSA group (n = 3). (F) Heatmap
   diagram of differentially expressed mRNAs screened from KBiF[4]@HSA
   group and KBiF[4]@HSA + X group (n = 3). (G) KEGG enrichment analysis
   of the PBS group versus the KBiF[4]@HSA group (Top 20 pathways). (H)
   The results of KEGG pathway enrichment analysis of the PBS group versus
   the KBiF[4]@HSA + X group (Top 20 pathways). (I) H&E, TUNEL, Ki67,
   γ-H2AX staining images of tumors after treatments (scale bar: 100 μm).
   The data are shown as the mean ± SD, calculated using three-way ANOVA,
   followed by one-way ANOVA and post hoc tests with *p < 0.05,
   **p < 0.01, ***p < 0.001, and ****p < 0.0001

   Subsequently, H&E staining of tumor sections was conducted to further
   assess the extent of cellular damage and necrosis (Fig. [150]6I). The
   KBiF[4]@HSA + X group exhibited significant histopathological
   alterations. To detect DNA fragmentation, a hallmark of apoptosis,
   TUNEL staining was employed as both a molecular biological and
   morphological method. Enhanced green fluorescence was observed in the
   KBiF[4]@HSA + X groups compared to the PBS group, indicating increased
   apoptotic activity. Additionally, Ki67 immunofluorescence analysis was
   performed to evaluate the proliferative capacity of the tumor cells. As
   expected, mice in the KBiF[4]@HSA + X group exhibited effective
   inhibition of tumor invasion. Moreover, γ-H2AX staining was employed to
   examine DNA double-strand breaks, revealing that cancer cells treated
   with the KBiF[4]@HSA + X group exhibited significant fluorescence
   compared to other groups, confirming the greatest DNA damage for
   inhibition.

In vivo safety and metabolism assessment of KBiF[4]@HSA

   In the development of nanobiomaterials, in vivo biocompatibility and
   biodegradability are critical considerations. Following intravenous
   administration of KBiF[4]@HSA, no significant abnormal behavior or
   weight loss was observed in rats over a 7-day period, indicating an
   absence of acute toxicity. To further assess the in vivo
   biocompatibility of KBiF[4]@HSA nanoclusters, hematological analyses
   and histological examination were conducted. The liver and kidneys,
   being the primary organs for the metabolism and excretion of
   KBiF[4]@HSA, were specifically evaluated. To assess whether KBiF[4]@HSA
   impacts liver and kidney function, the blood of rats was taken for the
   important hepatic and renal function indices tests, including ALT, AST,
   ALP, TP, ALB, GLOB, Ca, K, P, Na, Cl, A/G, CREA, UREA, and TG. One day
   and seven days post-injection, the aforementioned biochemical indices
   remained within the normal range, exhibiting no significant deviations
   from standard values, thereby confirming minimal side effects on renal
   and hepatic functions (Fig. [151]7A-D). In addition to blood analyses,
   histopathological examinations were performed to assess the potential
   toxicity of KBiF[4]@HSA on major organs using H&E staining. As
   illustrated in Fig. [152]7E, no discernible tissue damage,
   inflammation, or lesions were observed in the primary organs of rats
   treated with KBiF[4]@HSA. As a result, these findings provide an
   initial validation of the biocompatibility of KBiF[4]@HSA for in vivo
   biomedical applications.

Fig. 7.

   [153]Fig. 7
   [154]Open in a new tab

   Liver function indicators: ALT, ALP (A); Ca, K, P, Na, Cl (B); ALB, TP,
   GLOB, A/G (C) with KBiF[4]@HSA pre-injection and post-injection at 1
   and 7 days. Kidney function indicators: CREA, UREA, TG (D) with
   KBiF[4]@HSA pre-injection and post-injection at 1 and 7 days. (E) H&E
   staining of tumor sections from tumor-bearing rats (scale bar: 50 μm)

Conclusion

   In summary, the Bi-based DECT CAs, KBiF[4]@HSA could be efficiently
   synthesized using a straightforward one-pot HSA-mediated biomimetic
   mineralization strategy. Compared to the clinical CAs iohexol,
   KBiF[4]@HSA significantly enhances the CT value and imaging contrast at
   high keV, thereby improving the diagnostic accuracy for breast cancer.
   Furthermore, our study illustrates that KBiF[4]@HSA effectively
   scavenges the elevated GSH and increases the ROS levels in cancer
   cells, which augments the sensitivity of radiotherapy. Meanwhile, the
   combination of KBiF[4]@HSA with radiotherapy effectively inhibits the
   progression of breast cancer. Notably, the KBiF[4]@HSA exhibits no
   detectable tissue damage or inflammation, highlighting their superior
   biocompatibility in vivo. This study introduces a novel approach for
   developing biocompatible DECT CAs and radiosensitizers, offering a
   promising strategy for improving breast cancer diagnosis and treatment.
   However, further studies are needed to validate efficacy in diverse
   breast cancer subtypes, which highlights the need for continued
   research to fully translate this technology into clinical practice.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [155]Supplementary Material 1^ (2.3MB, docx)

Author contributions

   Yuelin Huang: Writing–original draft, Conceptualization, Data curation,
   Formal analysis, Investigation, Validation, Visualization. Zhenghai Lu:
   Conceptualization, Data curation, Formal analysis, Investigation,
   Validation, Visualization. Huanhuan Liu: Formal analysis, Methodology,
   Software, Validation, Visualization. Ningxin Yang: Data curation,
   Formal analysis, Investigation, Validation; Yanhong Chen: Data
   curation, Formal analysis, Investigation, Visualization. Chunting Wang:
   Data curation, Validation, Visualization. Weixi Huang and Jingjing Hu:
   Formal analysis, Resources; Dengbin Wang: Funding acquisition, Project
   administration, Supervision. Defan Yao: Writing-original draft,
   Writing-review & editing, Project administration, Supervision, Funding
   acquisition Conceptualization, Formal analysis. All authors reviewed
   the manuscript.

Funding

   This work was supported by the National Natural Science Foundation of
   China (92159203, 82471976, 82472039, 82071870), the Explorer Program of
   the Science and Technology Commission of Shanghai Municipality
   (24TS1415300, 24TS1415200), the Construction Project of the "Discipline
   Peak-Climbing Plan" of Xinhua Hospital Affiliated to Shanghai Jiao Tong
   University School of Medicine (XKPF2024C200, XKPF2024C201,
   XKPF2024C202), Shanghai Municipal Health Youth Talent Program
   (2022YQ042), and the Shanghai Rising-Star Program (22QA1406000).

Data availability

   No datasets were generated or analysed during the current study.

Declarations

Ethical approval

   The ethical review board of the XinHua Hospital Affiliated to Shanghai
   JiaoTong University School of Medicine, China, approved the experiment
   protocol and strictly followed its guidelines (Ethical approval number:
   XHEC-F-2024-014).

Consent for publication

   All authors read and agreed to submit the manuscript.

Competing interests

   The authors declare no competing interests.

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

   Yuelin Huang and Zhenghai Lu contributed equally to this work.

Contributor Information

   Huanhuan Liu, Email: liuhuanhuan@xinhuamed.com.cn.

   Dengbin Wang, Email: wangdengbin@xinhuamed.com.cn.

   Defan Yao, Email: yaodefan@xinhuamed.com.cn.

References