Construction of in-situ self-assembled agent for NIR/PET dual-modal imaging and photodynamic therapy for hepatocellular cancer

Abstract

Hepatocellular cancer (HCC) remained a life-threatening carcinoma. Agents for HCC imaging and therapy were expected to possess different intratumoral retention time. To construct an agent with different intratumoral retention time when applied for tumor imaging or therapy remained great values. A lasialoglycoprotein receptor (ASGPR) targeted lactobionic acid derivative (LABO) was constructed for fluorescent imaging and photodynamic therapy of HCC. ¹⁸F labeled LABO (¹⁸F-LABO) was developed for PET imaging of HCC. LABO and ¹⁸F-LABO showed similar molecular structure. LABO exhibited characteristic of viscosity and concentration-induced intratumoral in-situ self-assembly to expand the intratumoral retention. LABO was non-fluorescent at free stage, but emitted NIR fluorescence and generated irradiation-induced ROS after self-assembly for fluorescent imaging and photodynamic therapy. ASGPR specificity of LABO and ¹⁸F-LABO was confirmed using HepG2 cell. Biodistribution and fluorescent imaging confirmed the different tumor retention time of LABO and ¹⁸F-LABO when used for photodynamic therapy and PET imaging. PET imaging and photodynamic therapy were performed on HepG2 tumor bearing mice, which revealed that ¹⁸F-LABO/LABO could specifically accumulated in the HepG2 tumor for tumor location/inhibition. LABO/¹⁸F-LABO with excellent HCC specificity but different intratumoral behaviors showed great values for the PET/NIR imaging and photodynamic therapy for HCC.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-024-02879-6.

Introduction

The increasing incidence of hepatocellular cancer (HCC) has attached great significance to the specific diagnosis and accurate treatment [1, 2], which have now faced with some challenges. For example, ¹⁸F-FDG PET imaging, as a gold standard for tumor diagnosis and location, showed limitations in HCC due to its low intratumoral uptake, which might result from the dephosphorylation effect of the glucose-6-phosphatase (G6Pase), a gluconeogenesis enzyme highly expressed on the well-differentiated HCC, converting glucose-6-phosphate to glucose and consequently causing reduced retention of ¹⁸F-FDG within tumor cells [3, 4]. Hence, new PET probes were supposed to be developed for HCC imaging. On the other hand, chemotherapy or targeted therapy suffered from side effects and drug resistance for HCC patients [5–7]. Photodynamic therapy is a noninvasive and promising approach for cancer treatment, with minimal systemic toxicity and spatiotemporal controllability [8, 9]. In this context, we considered that development of a integrated theranostics agent for HCC imaging and photodynamic therapy showed great significance [10]. However, agents were supposed to possess different intratumoral retention when applied to imaging and therapy. Agents with suitable tumor retention time for imaging, radionuclide labeled FAPI for example, were always deemed to show insufficient tumor retention time when used for therapy [11, 12]. When agents were applied to PET imaging, it was expected to showed relatively short tumor retention time. Tumor retention for PET agents only needed to meet the requirements for tumor imaging, and should be cleared rapidly after imaging. Overlong tumor retention after imaging not only extended the nursing time, but also caused unwanted radiation damages. However, agents for tumor therapy required longer tumor retention time to enhance the antitumor effects [13, 14]. To construct an agent that met both intratumoral retention requirements when applied to tumor imaging and therapy, we developed an asialoglycoprotein receptor (ASGPR) targeted Lactobionic acid (LA) derivative for the fluorescent/PET dual modal imaging and photodynamic therapy for HCC. ASGPR is a C-type lectin, primary expressed on the sinusoidal surface of the hepatocyte [15–17]. As a liver-specific receptor, ASGPR expression was significantly enhanced when HCC was developed. LA showed specificity to ASGPR, and LA based PET probe and antitumor agents have been reported with excellent HCC specificity and tumor inhibition effect [18–20]. Here in our work, a fluorescent LA derivative (LABO) was constructed, which exhibited characteristics of viscosity and concentration-induced intratumoral in-situ self-assembly, as shown in Fig. 1. LABO was non-fluorescent in the free state due to the free rotation of double bonds. When system viscosity was enhanced, the free rotation was inhibited, and the frame of LABO was fixed. Intermolecular π-π stacking with head-to-head model between LABO was generated, accompanying the J-aggregation, which led to the fluorescence recovery and phototoxicity of LABO. When LABO was administrated for tumor therapy, it showed high specificity to HCC due to the recognition between LA and ASGPR. After LABO was specifically carried to peritumor, its small molecule structure made it more easily to penetrate through dense extracellular matrix (ECM) and arrived at tumor, compared with those pre-nanoengineered drug. Meanwhile, high ECM and tumor cell membranes viscosity could induce the in-situ self-assembly of LABO to generate LABO based nanoparticles (LABO NPs) via J-aggregation in the tumor, thus enhancing tumor retention time [21, 22]. LABO was non-fluorescent at free statement, but showed NIR fluorescence emitting and irradiation-induced ROS generation when aggregated in the tumor for NIR imaging and photodynamic therapy of HCC. The lack of the required condition of necessary viscosity and agent concentration for self-assembly, combined with the lack of irradiation provided double security assurance for the non target organs. Meanwhile, when LABO was labeled with ¹⁸F for PET imaging via ¹⁸-¹⁹ F exchange, ¹⁸F-LABO demonstrated consistent tumor specificity and in vivo behaviors before tumor internalization due to its similar structure to LABO, and could provide detailed information including tumor localization, staging and prognosis evaluation. However, ¹⁸F-LABO showed inability for in-situ self-assembly in the tumor due to its extremely low chemical equivalent, which allowed it to be rapidly cleared from tumor and metabolic organs after imaging.

Fig. 1.

Construction of viscosity-induced intratumoral in-situ self-assembled agent for NIR/PET dual-modal imaging and photodynamic therapy for HCC. The fluorescence and irradiation-induced ROS generation of LABO was inhibited, and could be turned on in the tumor by viscosity-induced self-assembly for fluorescent imaging and photodynamic therapy. LABO was also labeled with ¹⁸F by ¹⁸F-¹⁹F exchange for PET imaging

Results and discussion

Agents for tumor imaging and therapy were supposed to possess different intratumoral behaviors. LABO demonstrating viscosity and concentration-dependant intratumoral in-situ self-assembly might meet the requirements.

Synthesis and characterizations

Synthesis of LABO and 18F-LABO

The synthesis of LABO was shown in the scheme 1. The BODIPY-3 was three-step synthesized from 2,4-dimethyl-pyrrole and 4-(hydroxymethyl)benzaldehyde. After condensation with 4-(bis(4-methoxyphenyl)amino)benzaldehyde, and glycosylation with LA, the aimed product LABO was obtained. All of the products were characterized by mass spectrum and nuclear magnetic resonance, as shown in Fig S1-S9. ¹⁸F-LABO was then synthesized by ¹⁸-¹⁹ F exchange under the catalysis of SnCl₄ as the protocols reported before [23, 24]. As a isotope labeled compounds of LABO, ¹⁸F-LABO possessed similar chemical properties (hydrophilicity and polarity e.g.) and chemical activity to LABO, which could be used to investigate the in vivo behaviors of LABO, but might also demonstrate a little differences [25, 26]. For example, the chemical reaction rate might be influenced due to the molecular weight change, which was also known as kinetic isotope effect (KIE). Meanwhile, some physical properties might also change such as vapor density, diffusion velocity and magnetic resonance behavior (¹⁸F showed no magnetic resonance signal).

Scheme 1.

Synthesis of LABO and ¹⁸F-LABO. LABO was synthesized from BODIPY derivative (BODIPY 3), and was condensed with 4-(bis(4-methoxyphenyl)amino)- benzaldehyde to induce free rotation bonds. LA was linked for ASGPR specificity. ¹⁸F-LABO was synthesized by ¹⁸-¹⁹ F exchange under the catalysis of SnCl₄

Characterizations

The optical behaviors of LABO in water/methanol mixed solution was firstly investigated, as shown in Fig. 2a. When dissolved in pure methanol, LABO showed absorption peak at 570 nm, and its absorption at 650 nm was inhibited. However, with the increased water fraction in the mixed solution, the absorption at 650 nm was increased, and reached peak when water fraction reached 70%, which might resulted from the free rotation restriction of the double bonds of LABO. Same situation was also absorbed when UV absorption spectrum was measured in glycerinum/methanol mixed system, as shown in Fig. 2b. With the increased fraction of glycerinum, the system viscosity was enhanced and the absorption at 650 nm was also increased, which arrived at peak when glycerinum fraction reached 50%. For fluorescence emitting spectrum, as showed in Fig. 2c, LABO was non-fluorescent at 720 nm when dissolved in methanol. However, with the enhanced system viscosity by increasing glycerinum fraction, the fluorescence at 720 was also enhanced. The special spectrum behaviors of LABO proved our assumptions that the in-situ aggregation of LABO was a kind of J-aggregation that might happen between conjugated molecules with poor coplanarity [10]. On the one hand, LABO possessed free rotation double bonds, which not only reduced its conjugated system and make the absorption band blue shifted, but also allowed photon-excited LABO to release the energy in a non-radiative pathway and led to the fluorescence inhibition. When viscosity enhanced, the free rotation of double bond was blocked, and the molecular rigidity was enhanced. The enlarged and fixed conjugated systems of LABO made the UV absorption band red shifted, and allowed excited electrons to return to the ground state in a radioactive pathway, thus leading to the fluorescence recovery [27]. On the other hand, the LA tail of LABO generated intermolecular steric hindrance and reduced the intermolecular coplanarity. As a result, the structure-fixed LABO was unable to form π-π stacking in a head-to-tail model (complete stacking), namely H-aggregation, by which the molecules were closely arranged, and resulted to the absorption band blue shift and fluorescence quenching. LABO tended to form π-π stacking in a head-to-head or tail-to-tail model (partial stacking), namely J-aggregation [28]. J-aggregation with strong electron coupling could cause the absorption band red shift and fluorescence enhancement, which have also been observed in our optical spectrum investigation [29].

Fig. 2.

Synthesis and characterization of LABO, LABO NPs and ¹⁸F-LABO (a: UV spectrum changes of LABO in water/methanol solution, in each titration water fraction was enhanced ~ 10%; b:UV spectrum changes of LABO in glycerinum/methanol solution, in each titration glycerinum fraction was enhanced ~ 12%; c: fluorescent spectrum changes of LABO in glycerinum/methanol solution, in each titration glycerinum fraction was enhanced ~ 7%; d: TEM imaging of LABO NPs; e: ROS induced fluorescence of DCFH-DA; f: labeling rate of ¹⁸F-LABO, g: co-injection of LABO and ¹⁸F-LABO; h: radio-HPLC analysis of ¹⁸F-LABO after formulation, i: stability of ¹⁸F-LABO in PBS and mouse serum)

To directly observe the LABO NPs, TEM imaging was performed, and LABO NPs were presented as the nanoparticles with 100–200 nm diameters, as showed in Fig. 2d.

As a BODIPY derivative, LABO NPs was considered to be phototoxic and showed potential as a photodynamic therapy agent [10, 30, 31]. Hence, the irradiation-induced reactive oxygen species (ROS) generation was monitored by ROS probe (DCFH), as shown in Fig. 2e. LABO NPs demonstrated little ROS generation when incubated with DCFH in the dark. However, When irradiation was administrated for 60 s, evident ROS generation was observed, as indicated by the enhanced fluorescence of DCF, the ROS oxidative product of DCFH. Meanwhile, the ROS outbreak was observed the the irradiation time was prolonged to 15 min, and the fluorescent intensity was enhanced by nearly 50 times, indicating the strong phototoxicity of LABO NPs.

¹⁸F labeled LABO (¹⁸F-LABO) was synthesized via ¹⁸-¹⁹ F exchange of LABO. The ¹⁸F labeling rate was 83%, as shown in Fig. 2f, on which the free ¹⁸F⁻ and ¹⁸F-LABO appeared at about 2.6 min and 5.7 on high performance liquid chromatography that equipped with radioactivity detector (radio-HPLC). LABO and ¹⁸F-LABO were co-injected into radio-HPLC to confirmed the identity of ¹⁸F-LABO, Fig. 2g. The UV peak of LABO was a little ahead of the radioactive peak of ¹⁸F-LABO (15–20 s) due to the series connection of the UV and radioactivity detector. After formulation, ¹⁸F-LABO was injected in radio-HPLC, and no chemical or radiochemical impurities was found in the radio-HPLC, Fig. 2h. Only a little unlabeled LABO appeared in radio-HPLC, and the specificity activity was 52.46 Ci/mmol. The in vitro stability of ¹⁸F-LABO was evaluated in mouse serum and PBS (pH = 7.4).¹⁸F-LABO remained more than 90% intact after 60 min incubation. After 240 min incubation (more than 2 half-life), more than 80% ¹⁸F-LABO stayed still, as shown in Fig. 2f.

Cell assays

To visualize the ASGPR specificity and intracellular ROS generation, the intracellular fluorescence colocalization and DCF fluorescent imaging were performed. 2’,7’-Dichlorodihydrofluorescein diacetate (DCFH-DA) were used as the ROS probe in the cell instead of DCFH. HepG2 cell was incubated with LABO, fluorescence-labeled ASGPR-antibody and DCFH-DA for 15 min, during which cell was also irradiated. As showed in Fig. 3a₁−3a₄, though cell was dyed by the internalized LABO and showed unsatisfied status in the bright field imaging due to the phototoxicity, color and tinting strength of BODIPY-based LABO, bright intracellular NIR fluorescence of LABO NPs (3a₂) with strong colocalization to fluorescent ASGPR-antibody (3a₃) and DCF (3a₄) was observed, not only manifested the ASGPR specificity of LABO, but also prove the intracellular ROS generation. However, when cell was pretreated with LA to block the ASGPR, the intracellular fluorescence of LABO and fluorescent antibody were significantly inhibited, resulted to the less generation of ROS and decreased DCF fluorescence, as show in Fig. 3a₅−3a₈. The cell uptake dynamic was also analyzed, and the intracellular fluorescent intensity was quantified by imageJ, as shown in Fig. 3b and c. The intracellular fluorescence of LABO NPs was enhanced (from 34.13 ± 3.65 to 58.91 ± 4.12 and 88.01 ± 9.19 as quantified by imageJ) with the prolonged incubation time from 5 min to 60 min and 120 min, indicating the continuous cell uptake and intracellular aggregation of LABO other than cell efflux like other small molecular drug. The Pearson Colocalization Coefficient was 0.91, 0.87, 0.87, 0.97 and 0.92 at 30 min, 45 min, 60 min, 90 min and 120 min after incubation, respectively. ROS generation was also improved over incubation time, the intracellular fluorescent intensity of DCF was enhanced from 26.11 ± 5.03 to 33.98 ± 5.77 and 49.61 ± 7.12 after 60 and 120 min of incubation, as also demonstrated in Fig. 3c.

Fig. 3.

Cell assays of LABO: a: fluorescent co-localization of LABO, ASGPR receptor and intracellular ROS in free state (a₁-a_4, bright filed imaging and fluorescent imaging of LABO, ASGPR antibody and DCF in experimental group cell) or ASGPR blocked state (a₅-a_8, bright filed imaging and fluorescent imaging of LABO, ASGPR antibody and DCF in preblocked cell), b/c: dynamic fluorescence change of LABO and DCF in the cell, d: phototoxicity of LABO, e: cell uptake and efflux ration of ¹⁸F-LABO

The phototoxicity of LABO was also investigated, as shown in Fig. 3d. LABO showed low cytotoxicity in the dark. Cell viability remained 89.6 ± 5.9%, 83.12 ± 7.16%, 80.77 ± 6.68% after 24, 48 and 96 h incubation with LBAO (10 µM). However, when cell was incubated with LABO under irradiation, the cytotoxicity was significantly promoted. Cell viability decreased to 23.8 ± 4.4%, 17.33 ± 6.34% and 11.22 ± 4.83% after 24, 48 and 96 h incubation with LBAO (10 µM). The IC50 was 2.12, 1.44 and 0.95 µM for 24, 48 and 96 h incubation, as calculated by GraphPad Prism. Detailed cell viability data was also presented in Table S1.

Cell recognition and specificity of LABO to ASGPR were then monitored by coincubation ¹⁸F-LABO with ASGPR-positive HepG2 cells. As the results in Fig. 3e demonstrated, ¹⁸F-LABO showed rapid binding to HepG2 cells within 30 min, and 5.72 ± 0.77% radioactivity was observed in the cell. Cell radioactivity accumulation increased with the prolonged incubation time, and reached the peak of 8.05 ± 1.33% at 90 min after incubation. Extending incubation time to 120 min showed a little help to the radioactivity accumulation (8.32 ± 1.52%). However, when cell was pretreated with LA to block the ASGPR, ¹⁸F-LABO binding was significantly inhibited (1.46 ± 0.23% at 30 min and 3.03 ± 0.46% at 90 min), indicating the ASGPR specificity of ¹⁸F-LABO.

To achieve intratumoral in-situ self-assembly, sufficient cell internalization of LABO after specific recognition and binding to ASGPR was necessary. As a results, the cell internalization and efflux of LABO after binding was evaluated by contentious incubation of ¹⁸F-LABO-bond HepG2 cell in the fresh medium. The efflux ratio was defined as the radioactivity in the supernate that deriving from the cell efflux to the radioactivity that remained in the cell. The results in Fig. 3e indicated, instead of overlong binding to the cell surfaces and being rapidly washed out, ¹⁸F-LABO was able to penetrate into cells after specific binding to ASGPR. Only 7.18 ± 3.83% and 19.83 ± 5.68% radioactivity flowed out after another 60 and 120 min incubation after binding to cell, which reflected the satisfied cell internalization and retention of LABO. We considered that the suitable lipophilicity of BODIPY derivative might play significant roles in the cell membrane penetration and internalization of ¹⁸F-LABO. Strong cell penetration capacity and high intracellular retention not only provided the concentration condition for the in-situ assembly, but also guaranteed the sufficient irradiation damages from photodynamic therapy.

In vivo evaluation

The in vivo evaluation of ¹⁸F-LABO including peripheral blood clearance and biodistribution was performed. As shown in Fig. 4A, the radioactive signal in the blood decreased dramatically from 30.77 ± 6.39%ID/g (15 s) to 5.89 ± 0.98%ID/g after 5 min post-injection (p.i). After that, the radioactivity in the blood continuously reduced over time, and could be almost cleared from blood 1 h p.i. (0.063 ± 0.012%ID/g).

Fig. 4.

In vivo evaluation of LABO and ¹⁸F-LABO (A: plasma concentration of ¹⁸F-LABO over time; B: biodistribution of ¹⁸F-LABO, C: fluorescent imaging of HepG2 tumor bearing mice, a-e: 2, 4, 6, 8, 12 h after injection of LABO, D: TEM imaging of intratumoral in-situ self-assembled nanoparticles 12 h post administration of LABO, which were pointed by the black arrows)

Meanwhile, the biodistribution studies in Fig. 4B revealed specific radioactivity in the tumor, which increased from 0.78 ± 0.23%ID/g 15 min p.i. to 2.32 ± 0.67%ID/g 90 min p.i., and showed a little decrease 120 min p.i.(1.87 ± 0.46%ID/g). Hence, the imaging time point was determined at 90 min p.i. Meanwhile, no specific uptake in the muscle (0.41 ± 0.087%ID/g 15 min p.i. and 0.11 ± 0.035%ID/g 90 min p.i.) was observed. Heart showed relatively high radioactivity accumulation at beginning (4.76 ± 1.33%ID/g 15 min p.i.) due to the blood pool and myocardium uptake, which reduced over time (0.89 ± 0.32%ID/g 90 min p.i. and 0.42 ± 0.14%ID/g 120 min p.i.). Enhanced radioactive signal was observed in the liver and kidney (6.74 ± 1.67/4.86 ± 1.54%ID/g for liver/kidney 90 min p.i.), which might resulted from the in vivo defluorination and the metabolic pathway of ¹⁸F. Signal in liver and kidney also tended to decrease (5.89 ± 1.24%ID/g and 3.96 ± 0.94%ID/g) 120 min p.i. Overlong retention in the tumor and non-target organs was unnecessary for PET imaging. ¹⁸F-LABO showed inability to self-assembly after administrated, and could be rapidly cleared. The biodistribution data was also presented in Table S2.

To confirm the different intratumoral retention of LABO during photodynamic therapy, NIR fluorescent imaging was used to monitor the tumor retention of LABO. As showed in Fig. 4c, only background fluorescence was observed in the tumor 2 h p.i. due to the insufficient LABO NPs generation in the tumor. 4 h after injection, weak fluorescence was generated in the tumor, which get brighter over time with the accumulation of LABO NPs, and reached peak 12 h p.i., making tumor clearly delineated and distinguished in the images. The radiant efficiencies were 1.18 × 10⁷, 2.43 × 10⁷, 3.66 × 10⁷, 4.23 × 10⁷ and 4.87 × 10⁷ [p/s/cm²/sr]/[µW/cm²] 2, 4, 6, 8 and 12 h p.i. respectively. The fluorescent imaging demonstrated visualized results that the generation of LABO NPs significantly enhanced the intratumoral retention of LABO when compared with ¹⁸F-LABO. Meanwhile, LABO was non-fluorescent in the non-target (heart, spleen, lung, stomach and intestine), creating a low-signal background for fluorescent imaging. High tumor-to-non target ratio (T/NT) guaranteed sufficient tumor irradiation damages and minimum unwanted non-target damages during photodynamic therapy.

To further confirm the generation and retention of LABO NPs in the tumors at different time windows, TEM imaging was performed using tumor tissues from mice that treated with LABO (12 h and 18 h after treatment). Normal liver tissue was also obtained and imaged 18 h post administrated. As the results in Fig. 4d and Fig S10a-10c indicated, intratumoral in-situ self-assembled nanoparticles were observed as black spheres in the tumors at both 12 and 18 h after injection, while barely sphere nanoparticles were observed in normal liver, as shown in Fig S10d. These results not only reflected the long intratumoral retention of LABO NPs, but gave the most intuitive evident that LABO could successfully self-assembled into nanoparticle when subjected to the high viscosity environments such as tumor microenvironment and tumor membrane. However, its aggregation in normal liver was more difficult. We considered that only in some special circumstances (e.g. mass plasma albumin absorption, abnormal accumulation in liver/kidney due to over high concentration) might induce the LBAO NPs generation. Hence, the optimal administrated dosage of LABO needed to be further explored to enhance the therapeutic effect and reduce the liver damages of LABO. However, the primary experimental evidences that obviously LABO NPs abundance difference was observed between tumors and liver have reflected that the viscosity induced in-situ self-assembly of LABO could be a potential tool to ensured the sufficient tumor damages while guaranteed non-target safety during photodynamic therapy.

In vivo PET imaging

Fluorescent imaging was fast and convenient but was limited by the scattering and tissue penetration depth [32, 33], PET imaging was then performed based on ¹⁸F-LABO as the complement, as shown in Fig. 5a (45 min p.i.) and 5b (90 min p.i). When imaged at 45 min p.i., ¹⁸F-LABO demonstrated specific uptake in the tumor, as indicated by the crosshair in Fig. 5a. Tumor could be clearly indicated and delineated by the high radioactivity accumulation, and the SUVmean was 2.67. Meanwhile, non-target organs such as heart, muscle and intestine also demonstrated ¹⁸F-LABO uptake, which reduced the tumor contrast to some extent, yet not yield significant influence on tumor observation. When imaged as 90 min p.i., the results in Fig. 5 demonstrated that ¹⁸F-LABO could specifically accumulate in the HepG2 tumor due to the ASGPR recognition, and the SUVmean was 4.47. Tumor could be clearly distinguished in the images, while non-target organs including spleen, muscle, stomach and intestine exhibited low ¹⁸F-LABO. Relatively high radioactivity in the heart (SUVmean = 2.68) was observed due to the blood pool. High radioactive signal in the bladder was observed, which might result from the metabolic pathway or in vivo defluorination. Since PET imaging based on ¹⁸F-FDG showed limitations for HCC due to the low uptake in the tumor, we considered that ¹⁸F-LABO demonstrated high ASGPR specificity and excellent HCC localization showed great potentials as a imaging tool for HCC screening and staging.

Fig. 5.

PET imaging of HepG2 tumor bearing mice 45 min (a) and 90 min (b) after injection of ¹⁸F-LABO. The axial plane was shown as the right column. The coronal/sagittal plane was shown as the upper/lower row. Tumor was indicated by the corsshair in the images

However, since PET imaging was performed based on the positron from ¹⁸F, while fluorescent imaging was performed based on the fluorescence of LABO NPs that was inhibited at the free stage, it was necessary to mention that PET imaging and fluorescent imaging might give different in vivo information. When administrated, both ¹⁸F-LABO and LABO existed as the form of free small molecule, and could subject to series of metabolic pathways such as circulation clearance, kidney retention and phagocytosis of kupffer cell in liver. In these situations, PET imaging could reflect the metabolic pathways of ¹⁸F-LABO, while fluorescent imaging might show limitations due to the fluorescence inhibition except for some special circumstances as we mentioned before (mass plasma albumin absorption, abnormal accumulation in liver/kidney due to over high concentration), which might also lead to the partial fluorescence recovery and cause the different metabolic behavior between LABO and ¹⁸F-LABO. After that, ¹⁸F-LABO and LABO could be taken in by no-specific organs such as lung, muscle and stomach, where PET imaging could also provided the in vivo information, but fluorescence of LABO was still blocked since there was no required conditions for self-assembly. As a results, PET imaging results also reflected the uptake and metabolism in non-target organs. Last but not least, when arriving at peritumor/tumor microenvironment (TME) or specifically binding to tumor surfaces, LABO was self-assembled into LABO NPs due to the high viscosity of TME and tumor. The nanoscale aggregation of LABO NPs significantly decreased tumor clearance and enhanced intratumoral retention, leading to the long time NIR fluorescence observation window and sufficient phototoxicity. For ¹⁸F-LABO, it was unable to self-assembled into ¹⁸F-LABO NPs in the TME or tumors due to its extremely low chemical equivalent. As a result, it’s tumor retention time was much shorter than than of LABO NPs, and was cleared rapidly from tumor to avoid unnecessary radiation damages. In summary, we considered that fluorescent and PET imaging could play different roles and give complementary in vivo information when investigating retention and metabolism in different organs.

Antitumor effects evaluation

The antitumor effect of LABO NPs during photodynamic therapy was evaluated on the HepG2 tumor bearing nude mice. As the results in Fig. 6a and b demonstrated, tumors growth was significantly inhibited by LABO administration combined with tumor irradiation during 12-day photodynamic therapy. Tumor volumes were reduced from 148.8 ± 24.5 mm³ to 78.5 ± 35.2 mm³ when compared with the controlling group, of which mice was administrated with saline and tumor volumes increased from 161.3 ± 16.7 mm³ to 1203.4 ± 252.9 mm³. When mice was only treated with LABO or irradiation, the antitumor effects was unconspicuous, tumor enlarged during therapy (161.12 ± 17.88 mm³ to 539.9 ± 189.1 mm³ for LABO treatment and 150.68 ± 10.35 mm³ to 412.9 ± 128.6 mm³ for irradiation treatment).

Fig. 6.

Antitumor effects of LABO: (a: Tumor photography after photodynamic therapy, tumors from up to down corresponded to the groups treated with LABO + irradiation, LABO, irradiation and saline; b: Tumor volume change during photodynamic therapy, c: Tumor ROS staining after photodynamic therapy; d: Immunohistochemical analysis of ASGPR expression on HepG2 tumors, e: Body weight monitoring during photodynamic therapy

To confirm the generation of intratumoral in-situ self-assembled nanoparticle and the photoinduced ROS, mice were then sacrificed after therapy and tumors were taken out for ROS staining. As the results in Fig. 6c demonstrated, ROS in HepG2 tumors treated with saline could be stained red (left upper). Meanwhile, ROS in the tumors treated with irradiation (right upper) or LABO (left down) was increased to some extent, but seemed not to reach the threshold of tumor tolerance to exert antitumor effects. However, when tumor was treated with LABO and irradiation, significant ROS outbreak in the tumor was observed, reflecting the phototoxicity of LABO. The immunohistochemical analysis of ASGPR was also performed, as shown in Fig. 6d, to confirmed the high expression ASGPR on the HepG2 tumor.

To confirm the biological security of LABO, mouse weight was monitored during photodynamic therapy. Mouse in all of the experimental/control groups demonstrated body weight increase, as shown in Fig. 6e. Body weight of mouse treated with saline/LABO/LABO + irradiation/irradiation was initially 14.88 ± 3.84 g 15.36 ± 4.12 g, 16.12 ± 4.38 g and 14.65 ± 2.34 g respectively, and were increased to 17.12 ± 3.78/18.33 ± 4.12 g, 16.83 ± 3.85/18.01 ± 3.05 g, 17.96 ± 4.32/19.21 ± 4.83 g and 16.02 ± 2.55/17.34 ± 3.36 g at 6/12 days post administration. Organ H&E staining of mice treated with LABO + irradiation was also performed, as shown in Fig S11. Organs of non-target including liver, kidney, heart, spleen and lung showed intact structure, confirming the biological security of LABO. Moreover, peripheral blood samples of mice treated with LABO only and LABO + irradiation were also collected, and the biochemical index, including erythrocyte count, leukocyte count, aspartate amino transferase (AST), alanine aminotransferase (ALT) and total bilirubin (T-Bil) were measured. As the results listed in Table S3 indicated, all of these blood index met the normal reference range, reflecting the low kidney/liver damages and hematotoxicity of LABO. Photodynamic therapy was considered as tumor inhibiting method with high-level security due to the controllable irradiation area, irradiation strength and irradiation time. The viscosity induced intratumoral in-situ assembly and fluorescence recover not only guaranteed effective tumor inhibition, but also added a double assurance for the non-target security.

Materials and methosd

Materials and equipment

¹⁸F was obtained from Shanghai Atomic Kexing Pharmaceutical Co., LTD. All of the reagents and solvents were purchased from Beijing InnoChem Technology Co., LTD. Cells were purchased form Shanghai Cell Bank, Chinese Academy of Sciences and cultured under the guidance. Cell culture medium, fetal bovine serum and PBS were obtained from Shanghai Aolu Biotechnology Co., LTD. Mice were obtained from Shanghai Slack Laboratory Animal Co., LTD.

Perkin-Elmer Lambda 20/2.0 UV-vis spectrophotometer and Bruker AVANCEIII 400 spectrometer were used for UV detection and NMR characterization. TEM (JEM-2010/INCA OXFORD, 200 kV) were used for TEM imaging. Agilent 1100 HPLC with a radioactivity detector was used for separation and identification. MicroPET-CT (Siemens SuperNova) and FluoroMax spectrofluorometer (HORIBA Scientific) were used for PET and fluorescent imaging.

Synthesis of LABO and 18F-LABO

150 mg POCl₃ and 10 mg DMF was dissolved in the 50 mL DCM, and was magnetically stirred for 1 h at room temperature. 300 mg 4-methoxy-N-(4-methoxyphenyl)-N-phenylaniline dissolved in 70 mL DCM was then added into the solution, and the reaction was kept reflux overnight. The reaction was then quenched and washed by Na₂CO₃ solution. After dried, DCM was removed under reduced pressure and 4-(bis(4-methoxyphenyl)amino)benzaldehyde was obtained after purification by chromatography.

BODIPY-1

The synthesis routine of LABO was listed in Scheme 1. 190 mg (2 mmol) 2,4-dimethyl-1 H-pyrrole and 540 mg 4-(hydroxymethyl)benzaldehyde were dissolved in 300 mL anhydrous dichloromethane (DCM), and drops of trifluoroacetic acid (TFA) was added. After 5 h reaction under magnetic stirring at room temperature, 500 mg 4,5-dichloro-3,6-dioxocyclohexa-1,4-diene-1,2-dicarbonitrile (DDQ) was added into the mixture, and the solution was kept stirring for another 2 h. Excess BF₃.Et₂O and Et₃N were then added, and the reaction was kept overnight. After reaction, the solution was washed by ultrapure water for 3 times and dried by anhydrous Na₂SO₄. DCM was removed under reduced pressure, and BODIPY-1 was obtained after purification by chromatography in 44% yield.

BODIPY-2

153 mg POCl₃ and 10 mg DMF was dissolved in the 50 mL DCM, and was magnetically stirred for 1 h at room temperature. 170 mg BODIPY-1 dissolved in 50 mL DCM was then added into the solution, and the reaction was kept reflux overnight. The reaction was then quenched and washed by Na₂CO₃ solution. After dried, DCM was removed under reduced pressure and BODIPY-2 was obtained after purification by chromatography in 52% yield.

BODIPY-3

In to the 50 mL acetonitrile was added 102 mg BODIPY-2, 79 mg 2-aminobenzenethiol and excess ammonium peroxodisulfate. The reaction was kept at 50 ^oC overnight under magnetic stirring. After reaction, acetonitrile was removed under reduced pressure and the mixture was redissolved in the DCM. After 3 times wash by ultrapure water, the organic layer was collected and dried. BODIPY-3 generation was monitored by LC-MS and could be directly used without purification.

BODIPY-4

BODIPY-3 solution and excess 4-(bis(4-methoxyphenyl)amino)- benzaldehyde were dissolved into 50 mL toluene, followed by the addition of small amount of piperidine and acetic acid. The mixed solution was heated to 125 ^oC for 5 h under the protection of N₂. BODIPY-4 was obtained after purification in 36% yield.

LABO

125 mg BODIPY-4, 700 mg LAB, 200 mg EDC and 20 mg DMAP was dissolved in anhydrous DMF. The mixture was heated to 100 ^oC and kept reflux overnight. After reaction, DMF was removed under reduced pressure, and the mixture was redissolved in the water. The solution was subjected to dialysis (1000 Dalton) to removed the reagents, and LABO was obtained after freeze-drying.

18F-LABO

0.1 mg LABO and 0.1 mg SnCl₄ was dissolved into anhydrous acetonitrile. 10–20 mCi dried ¹⁸F⁻ dissolved in anhydrous acetonitrile was added. The mixture was kept in the shaker for 15 min at 80 ^oC. After that, the mixture was cooled in iced bath, and was diluted with 10 mL ultrapure water. 100 µL diluted solution was injected into radio-HPLC to determine the labeling rate. Then, the diluted solution was carefully treated with C18 column to capture ¹⁸F-LABO. After 3 times wash with 3 mL ultrapure water to remove ¹⁸F⁻, acetonitrile, SnCl₄ and partial LABO, ¹⁸F-LABO was eluted by ethyl alcohol. After removing ethyl alcohol by nitrogen gas purge, ¹⁸F-LABO was formulated in the saline. 100 µL formulated saline was injected into radio-HPLC to confirm the successful synthesis of ¹⁸F-LABO and determine the chemical and radiochemical purity of ¹⁸F-LABO.

In vitro characterization of LABO and 18F-LABO

Optical behaviors of LABO

LABO was dissolved in the DMSO (2 mM) and diluted to 20 µM with the mixed solution of water and methyl alcohol (with different water fraction). 1 mL LABO solution was then taken out, and its UV-Vis spectrum and fluorescence spectrum was measured. LABO was dissolved in the DMSO (2 mM), and was then diluted to 20 µM with the mixed solution of glycerinum and methyl alcohol (with different glycerinum fraction). 1 mL LABO solution was then taken out, and its UV-Vis spectrum and fluorescence spectrum was measured. 5 µL LABO solution (glycerinum: methyl alcohol = 1: 1, v/v) was taken out for TEM imaging after negative staining.

ROS generation

2,7-dichlorodihydrofluorescein (DCFH) was used as the ROS probe to monitor the ROS generation. DCFH was added into 2 mL LABO solution (glycerinum: methyl alcohol = 1: 1, v/v) and kept the concentration at 50 µM. Then mixed solution was irradiated with a LED light (700 nm, 20 mW cm^− 2) fro different time, the fluorescent spectrum was measured and fluorescence emission at 522 nm was recorded.

18F-LABO in vitro stability

Approximately 10 µCi ¹⁸F-LABO was added into 1 mL PBS (pH = 7.4) or mouse serum, and was incubated for 4 h. At different time points after incubation, 100 µL PBS solution was directly injected into radio-HPLC to determine the radiochemistry stability. Serum was subjected with acetonitrile to precipitate proteins, and 100 µL supernatant was injected into radio-HPLC.

In vitro cell assays

Cell incubation

HepG2 cells were incubated in DMEM medium supplemented with 10% FBS and antibiotics (50 units/mL penicillin and 50 units/mL streptomycin) at 37 ^oC in a humidified atmosphere containing 5% CO₂. Cell was seeded in the 24-well cell culture plate and allowed overnight for attachment.

Cell uptake and intracellular ROS generation

After attachment, cell was pretreated with/without LA (100 µM) for 30 min. After that, the old medium was replaced with fresh medium that contained fluorescence (3,8-Diamino-5-ethyl-6-phenylphenanthridinium bromide) labeled ASGPR antibody, LABO (20 µM) and DCFH-DA (20 µM). Cell was then incubated for 15 min under irradiation. After incubation, medium was removed, and cell was washed with PBS for 3 times. Cell was then directly observed under a fluorescent microscope. For fluorescent imaging, 450 nm, 550 nm and 665 nm channel were selected for DCF, fluorescent antibody and LABO excitation, respectively.

Uptake dynamic and colocialization

After attachment, the old medium was replaced with new medium containing fluorescence labeled ASGPR antibody, LABO (20 µM) and DCFH-DA (20 µM), cells were then incubated for 60 min under irradiation. At different time points of incubation, medium was removed and cell was washed with fresh medium for 3 times. Cells were directly observed using a fluorescent microscope. 450 nm, 550 nm and 665 nm channel were selected for DCF, fluorescent antibody and LABO excitation, respectively. The fluorescent intensity was analyzed and quantified by imageJ, by which the intensity was described by the gray value.

Cytotoxicity

HepG2 cell was seeded into 96-well cell culture plate and allowed overnight for attachment. After attachment, old mediums were replaced with fresh mediums containing different fractions of LABO (0–10 µM). Cell were then incubated in the dark or under intermittent LED irradiation (every 6 h, 30 min per time) for 24, 48 and 96 h. After incubation, 10 µL CCK8 was added into each well and was incubated with cell for another 1 h. Absorbance of each at 450 nm was then measured. Cell viability was described as the ratio of absorbance in each well to that in control well where no LABO was added. Half maximal inhibitory concentrations (IC50) were then calculated using GraphPad Prism9.

Cell quantitative uptake

After attachment, cell was randomly divided into 2 groups. Cell of inhibition groups was pretreated with LAB (100 µM) to block the ASGPR. The old medium was then replaced with fresh mediums and approximately 5 µCi ¹⁸F-LABO was added into each well. Cells were then incubated for another 2 h. At different time points of incubation, medium was collected and radioactivity in the medium was measured using a γ counter. Cells were washed with PBS (pH = 7.4) for 3 times. NaOH aqueous (1 M) was added into each well to incubated with cells for 5 min. After cells lysis, NaOH solutions were collected and the radioactivity in the cells was also measured using a γ counter. Cell uptakes of ¹⁸F-LABO were described as the ratios of the radioactive counts in per million cell to the whole added dose.

Cell retention and efflux

After cell attachment, the old medium was replaced with fresh mediums. 5 µCi ¹⁸F-LABO was added into each well and incubated with cell for 2 h. After incubation, old medium was removed. Cell was washed with PBS (pH = 7.4), and 1 mL new medium was added. Cell was then incubated for another 2 h. At different time points after the second incubation, medium was removed. Cell was washed with PBS (pH = 7.4) for 3 times. NaOH (1 M) was added into each well to incubate with cells for 5 min. After cells lysis, NaOH solutions were collected and the radioactivity in cells was also measured using a γ counter. Cell retention and efflux of ¹⁸F-LABO were described as the radioactivity in the supernate that deriving from the cell efflux to the radioactivity that remained in the cell.

In vivo evaluation

Circulation clearance

Normal mice were injected with 50–100 µCi ¹⁸F-LABO via tail vein. At different time points after injection, blood was collected by capillary via tail vein and was weighted. Radioactivity in the blood was measured using a γ counter. The plasma radioactivity - time curve was depicted as the fluctuation of the radioactivity in the blood over time.

Pharmacokinetic and biodistribution studies

Normal mice were injected with 100–150 µCi ¹⁸F-LABO via tail vein. At different time points after injection, mice were sacrificed by breaking necks. Organs of interest were harvested. After weighing, the radioactivity in the organs were measured. Uptake in the organs was expressed as the percentage of the injected dose per gram tissue or organ (%ID/g).

In vivo imaging on living animals

Animal model

All of the animal experiments were performed under the guidance of the Ethics Committee of Shanghai Sixth People’s Hospital affiliated to Shanghai Jiao Tong University School of Medicine. 200 µL, 5 million HepG2 cells were subcutaneously inoculated on the right forelimb of the male nude mice (6–8 weeks). Mice were feed in normal and standard environment.

Fluorescent and TEM imaging

HepG2 tumor bearing mouse were injected with LABO (100 µL, 50 µM) via tail vein. 2, 4, 6, 8 and 12 h after injection, mice was anesthetized with 5% isoflurane delivered in 66%/33% nitrogen/oxygen, and fluorescent imaging was performed. After imaging, mice were sacrificed and tumors were taken out immediately for TEM imaging. The injection time was controlled to ensure that mice were imaged simultaneously.

HepG2 tumor bearing mouse were injected with LABO (100 µL, 50 µM) via tail vein. 18 h after imaging, mice were sacrificed. Tumors was taken out and TEM imaging were performed. For comparison, normal liver tissues were also obtained for TEM imaging.

PET imaging

For PET imaging, HepG2 tumor bearing mice were injected with 100–150 µCi ¹⁸F-LABO. 45 and 90 min after injection, mice were anesthetized with 5% isoflurane delivered in 66%/33% nitrogen/oxygen. CT scan was firstly performed for anatomical reference and attenuation correction, followed by the PET acquisition. After acquisition, PET was reconstructed with ordered subset expectation maximization algorithm (OSEM, 2iteration 3subset, 256 matrix). Mice were anesthetized with isoflurane (1–2%) throughout the period of imaging.

Photodynamic therapy using LABO

The HepG2 tumor bearing mice were divided into four groups. Mice of group A, B, C and D were treated with saline, LABO, LABO and nothing respectively (every 4 days, 3 rounds). 12 h after each administration, mice of group C and D were subjected to LED irradiation (680 nm, 20 min). 12 days after first injection with/without irradiation, mice in group B and C were anesthetized and peripheral blood samples were collected by ophthalmectomy. Peripheral blood biochemical index, including erythrocyte count, leukocyte count, aspartate amino transferase (AST), alanine aminotransferase (ALT) and total bilirubin (T-Bil) were measured to further confirm the biological safety of LABO. Mice were then sacrificed. Tumors and organs of interest were taken out for ROS/H&E staining. Mouse weight and tumor volume were monitored during photodynamic therapy.

Conclusion

We have constructed a intratumoral in-situ self-assembly agent (LABO) and its ¹⁸F labeled PET probe (¹⁸F-LABO) for the fluorescent/PET dual-modal imaging and photodynamic therapy for HCC. LABO and ¹⁸F-LABO with excellent ASGPR specificity and suitable intratumoral behaviors showed great potentials as integrated tools for the theranostics of HCC.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Xinmiao Lu, Yucheng Fu prepared Figs. 1, 2, 3 and 4. Xinmiao Lu wrote the main manuscripts. Luxinmiao, Yunyun Zhu and Chuang Xi prepared Figs. 5 and 6. Hua Pang and Quanyong Luo supervise the experiments and acquired the fund. All authors reviewed the manuscript.

Data availability

Data was available from corresponding authors under reasonable request.

Declarations

Ethical approval

The animal experiments were performed under the guidance of the Ethic Committee. of Shanghai Jiao Tong University, School of Medicine Affiliated Shanghai Sixth People’s Hospital. All of the authors have read and consent to publish the manuscript.

Competing interests

The authors declare no competing interests.