Fabrication of PBCA-MB
PBCA-MB were synthesized based on anionic-emulsion polymerization as described previously [18]. Briefly, 3 mL of n-butyl cyanoacrylate (BCA, Special Polymers, Sofia, Bulgaria) was added drop-wise to 300 mL aqueous solution containing 1% of Triton-X100 at pH 2.5. This mixture was emulsified by Ultra-Turrax T-50 basic (IKA Werke, Staufen, Germany) at 10,000 RPM for 1 h at room temperature. The resulting solution was centrifuged at 500 RPM (46 G) for 10 min and washed with 0.02% (v/v %) aqueous solution of Triton-X100 (pH = 7, Sigma-Aldrich, Munich, Germany) until the subnatant was transparent. Size and concentration of MB were measured using a Coulter counter (Multisizer 4e, Beckman, Brea, United States). To perform Coulter counter measurements, a 5 μL solution of MB was mixed with 20 mL of ISOTON® II (Beckman Coulter, Brea, United States), and triplicate readings were taken.
Functionalization of PBCA-MB
Cyclo(Arg-Gly-Asp-D-Phe-Lys) (cRGD, MedChemExpress, Monmouth Junction, USA) binds to αvβ3 integrin on the cell surface with a pretty low IC50 of 0.94 nM [19]. To achieve effective binding of cRGD-MB to αvβ3 integrin, cRGD peptides were conjugated covalently to the surface of the MB by direct coupling of the free amine group on the peptides to aminolysed ester bond of BCA backbone on the surface of the MB. To this end, 2.2 mg cRGD peptide was dissolved in 300 μL 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma-Aldrich, Munich, Germany) buffer and subsequently incubated with 109 PBCA-MB in HEPES (1 M)-Triton X-100 (0.01%) buffer (pH 7). Subsequently, 17 µL (0.05 mmol) of 0.01% Lithium Methoxide (Sigma Aldrich, Darmstadt, Germany) were added and the pH value was adjusted to 8 by adding 0.1 M NaOH (Carl Roth, Karlsruhe, Germany). The reaction was kept at room temperature for 24 h under continuous stirring at 300 RPM and then purified with HEPES/Triton X-100 buffer through centrifugation and washing steps. Cyclo(Arg-Ala-Asp-D-Phe-Lys) (cRAD, MedChemExpress, Monmouth Junction, USA) modified MB that do not bind to αvβ3 integrin, were prepared identically and used as a negative control.
High-performance liquid chromatography
The cRGD-MB solution was purified using a dialysis membrane (6-8 k MWCO, Spectrum, Massachusetts, USA) in ultrapure water to remove Triton X-100 and uncoupled peptides. The purified samples were then mixed with an equal volume of acetonitrile (ACN) to disrupt the MB to polymers prior to the characterization with high-performance liquid chromatography (HPLC). HPLC was performed using an Agilent 1260 Infinity system (Agilent technologies, Waldbronn, Germany) equipped with a diode array and multiple wavelength UV–Vis detector and a reversed phase column (Eclipse Plus C18, 3.5 μm, 4.6*150 mm). A gradient elution method was applied with eluent A (0.1% trifluoroacetic acid in ACN) increasing from 5 to 95% in 10 min (eluent B was H2O with 0.1% trifluoroacetic acid). The flow rate was set at 1 mL/min, and the injection volume was 50 µL. The UV–Vis detector operated at a wavelength of 220 nm to detect cRGD. Chromatograms were recorded and analyzed using the Agilent ChemStation software (Agilent technologies, Waldbronn, Germany).
Gel permeation chromatography
To track the aminolysis reaction, PBCA-MB were coupled with fluorescent Cy3-amine (Lumiprobe, Hannover, Germany) instead of cRGD. The Cy3-coupled MB and unmodified MB were lyophilized and further dissolved in dimethylformamide (DMF) containing 10 mM LiCl at a concentration of 5 mg/mL. The resulting solution was filtered through a 0.22 µm membrane to remove any precipitate. Gel permeation chromatography (GPC) was conducted using a PLgel 3 μm MIXED-E column (300 × 7.5 mm, Agilent technologies, Waldbronn, Germany) and PEG-standards (Agilent technologies, Waldbronn, Germany) were applied as calibration standards according to the manufacturer’s instruction. The retention time of each compound was recorded with infrared (IR) and UV–Vis detector (550 nm wavelength for Cy3 detection) at the same time. The eluent was DMF containing 10 mM LiCl, the elution rate was 0.5 mL/min, and the temperature was 55 ℃. Data analysis was carried out using a specialized Cirrus GPC software provided by the instrument manufacturer (Agilent, Waldbronn, Germany).
Cryogenic scanning electron microscopy
The morphology of standard-MB and cRGD-MB was investigated by cryogenic scanning electron microscopy (Cyro-SEM) with a Hitachi FE-SEM 4800 (Krefeld, Germany) equipped with an Alto 2500 Cryo-Gatan unit (Gatan GmbH, Munich, Germany) operating at 1 kV and 2 μA. Standard-MB and cRGD-MB were dropped onto a sample holder and frozen in liquid nitrogen. Subsequently, the sample holder was inserted into the preparation chamber for morphology observation. The MB droplets were fractured using a carbide knife and then sublimed at − 100 °C for 5 min to observe the structure and morphology of standard-MB and cRGD-MB.
Flow cytometry
Cy3 was used as a model dye and was conjugated to the PBCA-MB both using the aminolysis and the hydrolysis protocol described in previous studies [18, 20] and investigated by flow cytometry (Becton Dickinson (BD) FACS Canto II, Heidelberg, Germany). The mean fluorescence intensity (MFI) of Cy3 per MB was used to quantify the level of conjugation of the ligand. As the aminolysis protocol may result in a different MB size distribution, the variability of the MB surface coverage with dyes was investigated by correlating MB sizes with MFI.
The detailed procedure was as follows: Purified Cy3-MB (5 × 106/mL in a 0.02% triton solution) were measured at a low flow rate, and 50,000 events were recorded for each sample. Standard-MB were used as a negative control. The MFI was analyzed using flowjo-v10 (FlowJo LLC, Ashland, United States), in which standard-MB were gated as Cy3 negative.
In vitro US phantom imaging
Custom-made gelatin-based phantoms were used to analyze the echogenicity of standard-MB and cRGD-MB. The VEVO 3100 US system, incorporating a linear MX-250 transducer (FUJIFILM VisualSonics, Toronto, Ontario, Canada), was employed for US imaging. 3 × 105 MB were suspended in 4.5 mL of 2% w/v gelatin solution, and the mixture was embedded in 10% w/v gelatin solution. The transducer was fixed vertically above the phantom using a focus depth of 11 mm. US imaging was performed in non-linear contrast mode (NLC-mode) at 18 MHz frequency and 4% power. 100 frames were recorded, followed by a destructive pulse (0.5 s at 100% power) to destroy the MB in the gelatin phantom. To quantify the acoustic intensity, a region of interest (ROI) was drawn within sample-loaded gelatin and acoustic intensities were assessed before and after the US destructive pulse using the VevoLAB software version 3.2 (FUJIFILM VisualSonics, Toronto, Ontario, Canada).
Cell culture
The binding efficiency of cRGD-MB to αvβ3 integrin was investigated via Human umbilical vein endothelial cells (HUVEC). HUVEC were obtained from PromoCell (Heidelberg, Germany), maintained in Vasculife Basal Medium (Lifeline, Troisdorf, Germany) enriched with fetal bovine serum (FBS: 2% v/v), endothelial cell growth supplement, gentamycin (1% v/v), and 1% penicillin/streptomycin. For the in vitro flow chamber assay, HUVEC were grown in 1% gelatin pre-coated 35-mm petri dishes until they reached 80% confluency. 4T1 cells (murine epithelial mammary carcinoma) were used for the establishment of mouse breast tumor model. They were obtained from the American Type Culture Collection (ATCC, Manassas, Virginia, USA), and maintained in Roswell Park Memorial Institute Medium (RPMI 1640) with 10% fetal calf serum and 1% penicillin/streptomycin. Cells were passaged every three days.
In vitro flow-chamber binding studies
The specific-binding studies of cRGD-MB, cRAD-MB and standard-MB were performed in an in vitro flow chamber model as described previously [21]. Briefly, 100 μL containing 106 HUVEC/mL were added to 1% gelatin pre-coated 35-mm petri dishes. 4 h before the MB solution was added to the chamber, HUVEC were activated with 4 ng/mL of recombinant human TNF-α. The cells were then stained with Alexa Fluor 488Y-conjugated wheat germ agglutinin (WGA-AF488; Thermofisher, MA, USA) and nuclei were stained with Hoechst (Invitrogen, CA, USA) for 30 min, followed by washing 3 times with HUVEC medium. Petri dishes were placed in a flow chamber and a flow rate of 0.25 mL/min was applied using a peristaltic pump (Gilson Inc, Villiers-le-Bel, France). All MB were pre-dyed with rhodamine B (1 mg in 10 mL MB solution; Merck, Darmstadt, Germany) as described previously and were further purified by washing with HEPES/Triton X-100 buffer [9]. Subsequently, rhodamine-MB were added to the perfusion system at a concentration of 1 × 108 MB /mL and allowed to circulate for 10 min in a closed loop. After the circulation phase, the petri dishes were disconnected from the flow chambers and washed with HUVEC medium to remove unbound MB. Fluorescence microscopy (Carl Zeiss AG, Oberkochen, Germany) was used to analyze the specific binding of MB to HUVEC.
Ex vivo flow-chamber binding studies
Eight-week-old C57BL/6 J wild-type mice were used (n = 10). Mice were euthanized by cervical dislocation under anesthesia with 1.5% isoflurane (FORENE, AbbVieAG, Ludwigshafen, Germany) and subsequently their aortas were dissected. A custom-built flow chamber setup equipped with a pipette (tip diameter: 120–150 μm) was employed and the sink was filled with Ringer solution (B. Braun, Melsungen, Germany). The aortas were mounted end-to-end and the endothelium was activated by perfusion with 4 ng/mL TNF-α in Ringer solution for 4 h. The standard-MB, cRAD-MB, and cRGD-MB were introduced at a concentration of 108 MB/mL and at a flow rate of 0.25 mL/min. To image the vessel aortas in high resolution, US measurements were performed with an MX-700 US transducer (Vevo 3100, FUJIFILM VisualSonics, Amsterdam, Netherlands) at a frequency of 50 MHz, a transmit power of 1%, and a dynamic range of 60 dB. After 10 min of circulation, the MB were removed by a perfusion with Ringer solution. Bound MB were detected using the US destruction-replenishment analysis method [22]: MB were disrupted by applying 100% US power in Doppler mode at a frequency of 40 MHz with a pulse repetition frequency of 5 kHz, duration of 10 s and dynamic range of 40 dB. 100 frames were recorded in B-mode before and after the destructive US pulse and analyzed using the software MATLAB (R2022a, Natick, Massachusetts, USA). The mean ultrasound intesity per frame was determined within a ROI that covered the vessel wall. Utrasound intensity-time-curves were plotted.
In vivo molecular US imaging
All animal experiments were approved by the German State Office for Nature, Environment and Consumer Protection (LANUV) North Rhine-Westphalia. 10–12 week old female Balb/c mice (Janvier Labs, Le Genest-Saint-Isle, France) were housed on spruce granulate bedding (Lignocel, JRS, Germany) in groups of 3–5 animals in type II long individual ventilated cages (Tecniplast, Germany) under specific pathogen-free conditions. Husbandry rooms were temperature (20–24 °C) and humidity (45–65%) controlled. Water and standard pellets for laboratory mice (Sniff GmbH, Soest, Germany) were offered ad libitum. Group-housed animals were assigned individual earmarks for identification. Tumors were induced by subcutaneous injection of 4 × 104 4T1 cells in 50 μL phosphate-buffered saline (PBS) in the right hind limb. Tumor diameters were monitored by caliper measurements. After tumors had reached 5–6 mm in diameter, molecular US imaging was performed using the VEVO 3100 US system equipped with a linear MX-250 transducer. The transducer was positioned vertically over the tumor region and the focus depth was kept at the middle of the tumor. During all US measurements, mice were kept under inhalational anesthesia using 1.5% isoflurane. Imaging was performed in NLC-mode at 18 MHz and with 10% transmit power. Mice in groups of standard MB (n = 5), cRAD-MB (n = 5) and cRGD-MB (n = 5) were injected with 5 × 107 MB in 50 μL 0.9% saline solution as a bolus via a tail-vein catheter, followed by a 20 μL saline flush. During the intravenous injection of the MB solution, the signal enhancement was assessed by recording 500 frames in NLC-mode (frame rate: 10 fps). The maximum inflow of MB was described by the peak enhancement. The intravenous injection of MB was followed by an interval of 8 min, during which the MB had time to bind to their target and unbound MB were cleared from the bloodstream. Thereupon, an NLC-mode image sequence was acquired where 100 frames were recorded before and after a 1-s-long destructive pulse at 100% power. Finally, the signal deriving from bound MB was quantified by subtracting the post- from pre-destructive US mean acoustic intensity using the VevoLAB software version 3.2 (FUJIFILM VisualSonics, Amsterdam, Netherlands).
Histological analysis
After US imaging, mice were injected with fluorescein (FITC)-conjugated lectins (Vector Laboratories Inc, Burlingame, California, USA) and euthanized by cervical dislocation after 15 min. Tumors were resected, embedded in Tissue-Tek (Sakura Finetek Europe, Alphen aan den Rijn, Netherlands), snap-frozen and stored at − 80 °C until cryosectioning. Immunofluorescence stainings were performed on 8 μm-thick cryosections. The tumor sections were fixed with 80% methanol for 5 min at 4 °C followed by the addition of acetone at − 20 °C for 2 min. After fixation, sections were washed 3 times with PBS and incubated overnight with a rabbit anti-αvβ3 integrin antibody (1 µg/ µL; eBioscience, San Diego, California, USA) at a dilution of 1:100 at 4 °C. Thereafter, the sections were washed with PBS and stained using a donkey Cy3-labeled secondary anti-rabbit antibody at a dilution of 1:500 (7.5 µg/ µL; Dianova, Hamburg, Germany). The mean percentages of the αvβ3 integrin-positive areas to the total area were calculated using Image J (Carl Zeiss AG, Oberkochen, Germany).
Statistical analysis
Data are shown as mean ± standard deviation (SD). To calculate statistical differences between two groups, the unpaired t-test was used. One-way analysis of variance (ANOVA) with subsequent Tukey post-hoc testing was applied for comparing more than two groups. All statistical analyses were performed using GraphPad Prism 10 (San Diego, California, USA). A statistically significant difference was considered at *p < 0.05; **p < 0.01; ***p < 0.001, as indicated in the figure legends.
Add Comment