Synergic fabrication of pembrolizumab loaded doxorubicin incorporating microbubbles delivery for ultrasound contrast agents mediated anti-proliferation and apoptosis

Abstract This study evaluated pembrolizumab-conjugated, doxorubicin (DOX)-loaded microbubbles (PDMs) in combination with ultrasound (US) as molecular imaging agents for early diagnosis of B cell lymphomas, and as a targeted drug delivery system. Pembrolizumab, a monoclonal CD20 antibody, was attached to the surfaces of DOX-loaded microbubbles. PDM binding to B cell lymphoma cells was assessed using immunofluorescence. The cytotoxic effects of PDMs in combination with ultrasound (PDMs + US) were evaluated in vitro in CD20+ and CD20– cell lines, and its antitumor activities were assessed in Raji (CD20+) and Jurkat (CD20–) lymphoma cell-grafted mice. PDMs specifically bound to CD20+ cells in vitro and in vivo. Contrast enhancement was monitored in vivo via US. PDM peak intensities and contrast enhancement durations were higher in Raji than in Jurkat cell-grafted mice (p < 0.05). PDMs + US treatment resulted in improved antitumor effects and reduced systemic toxicity in Raji cell-grafted mice compared with other treatments (p < .05). Our results showed that PDMs + US enhanced tumor targeting, reduced systemic toxicity, and inhibited CD20+ B cell lymphoma growth in vivo. Targeted PDMs could be employed as US molecular imaging agents for early diagnosis, and are an effective targeted drug delivery system in combination with US for CD20+ B cell malignancy treatment.


Introduction
Early diagnosis is pivotal for therapeutic success in many types of cancers. Ultrasound (US) molecular imaging is a novel diagnostic approach for early detection of non-Hodgkin lymphoma (Wang et al., 2017;Heo et al., 2019;Toumia et al., 2019). Recent studies suggest that targeted microbubbles as US contrast agents (TMUCA) may serve as probes for US molecular imaging. TMUCA would improve diagnostic specificity and allow for disease monitoring in real time. TMUCAs can accumulate and remain at the tumor site for long time periods, and imaging at the molecular level can be acquired using US after TMUCA venous injection (Sanna et al., 2011;Li et al., 2018;Chen et al., 2019a,b;Zhang et al., 2020). US molecular imaging also produces quantitative data, exhibits good temporal resolution, is noninvasive, produces no ionizing radiation, and is relatively inexpensive. Over the past decade, various types of TMUCA have been applied for cell-specific targeting with US molecular imaging in vivo, specifically to assess intravascular inflammation, intravascular thrombosis, and tumor blood vessels (Chen and Sun, 2021;Zhu et al., 2021;Huang et al., 2021a,b). For early tumor diagnosis, TMUCAs were conjugated with antibodies specific for tumor cell surface antigens. Previous studies showed that the tumor neovasculature is distorted, with an imperfect basement membrane, and no smooth muscle layer. Permeability was also increased, with wall pores approximately 380-780 nm in diameter. Therefore, TMUCA diameters were adjusted to approximately 500 nm for easy passage through vascular endothelial cells and improved molecular imaging (Liu and Huang, 2011;Song et al., 2015;Wu et al., 2018;Zheng et al., 2021).
Targeted microbubbles are promising tumor-targeting drug delivery systems, although their potential utility as US contrast agents has not yet been studied. Most chemotherapy drugs currently have no targeting capabilities, and act on both diseased and non-diseased sites, leading to low therapeutic indices and severe side effects (Padmanabhan et al., 2016;Picheth et al., 2017;Guo et al., 2018;Paris et al., 2019;Wang et al., 2019). A targeted drug delivery system can increase chemotherapy drug accumulation specifically at target sites, while reducing non-target impacts (Ding et al., 2019;Zheng et al., 2020;Feng et al., 2021;Wang et al., 2021;Ma et al., 2021). Moreover, targeted microbubbles are both chemically stable and biodegradable, and exhibit prolonged circulation in the blood, with localized drug release (Lux et al., 2017;Chen et al., 2019 a,b;Szablowski et al., 2019;Brambila et al., 2020). Tumor-specific ligand-like peptides, galactose-conjugated chitosan, transferrin, folic acid, and monoclonal antibodies have been employed to target microbubbles to tumor cells for the treatment of many cancers (Chertok et al., 2016;Chen et al., 2017;Tang et al., 2018). Additionally, the combination of targeted drug-loaded microbubbles with US irradiation permeabilizes cell membranes, enhancing drug uptake by tumor cells, and selectively killing tumor cells without harming normal cells. Therefore, targeted drug-loaded microbubbles have potential use in both targeted drug delivery systems and in combination with US molecular imaging (Kheirolomoom et al., 2010;Huang et al., 2021a,b).
We hypothesized that pembrolizumab-conjugated, doxorubicin (DOX)-loaded microbubbles (PDMs) could serve as effective, biocompatible B cell lymphoma-targeting theranostic agents. The present work evaluated the specific binding potential of PDMs targeting CD20 antigen, a tetraspan membrane receptor overexpressed in B cell malignancies, in lymphoma Raji cells. We also assessed the cytotoxicity and antitumor activity of these PDMs in combination with US irradiation in vitro and in vivo. Finally, targeted US molecular imaging was explored in Raji and Jurkat cell-grafted mice.

DM preparation
PLGA microbubbles incorporating DOX were fabricated via a double US emulsion evaporation procedure. Of 0.5 g PLGA was fully dissolved in 10 mL of liquid chloroform via agitation. The PLGA solution was then combined with a 5 mg DOX solution (dissolved in 1.0 mL superpure water), and the mixture was emulsified via US for 120 min at 100 w. 1.0 mL span-80 was then added. The vial was degassed and re-perfused with nitrogen with stirring at 23,000 rpm for 5 min to obtain primary emulsified DMs. The primary emulsion was poured into cold PVA (40 mL, 5%) containing 1.0 mL tween-80, and stirred at 21,000 rpm for 30 min at room temperature for the second emulsion. The double emulsion was poured into isopropyl alcohol (40 mL, 2.5%) and mechanically agitated for 180 min at room temperature to volatilize the chloroform. The supernatant was removed after the solution was centrifuged at 4800 rpm for 5 min. The precipitate was centrifuged again at 1800 rpm for 5 min, and resuspended in superpure water. The superpure water wash was repeated several times until the supernatant become transparent. Precipites were resuspended a final time in superpure water and stored at 4 C. DMs were sterilized via cobalt 60 (60Co) irradiation (Oliveira et al., 2019;Shi et al., 2020;Lainovi c et al., 2020;Xiao et al., 2020;.

PDM preparation
Covalent bonding of the activated carboxyl groups on DM surfaces was performed using the 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) method in the presence of N-hydroxysuccinimide (NHS). Prepared DMs were resuspended in phosphate-buffered saline (PBS; pH 4.7), and EDC and NHS in an equimolar ratio were added into the suspension. The carboxyl groups were activated for 60 min at room temperature. The supernatant was removed after centrifugation, and the precipitate was resuspended in PBS. Amine-Peg2000-Biotin in MES buffer was added, and the mixture was incubated for 120 min at room temperature to obtain biotinylated DMs. Biotinylated DMs were incubated with avidin or dylight488-labeled avidin (1 mg/mL) for 10 min at room temperature. The mixture was then centrifuged three times and resuspended in PBS to remove surplus dylight488-labeled avidin/avidin. Pembrolizumab was biotinylated using the EZ-LinkTM Sulfo-LC-Biotinylatio kit according to the manufacturer's instructions, and was added to the avidin-biotin conjugated DMs and incubated for 10 min. The PDM suspension was rinsed three times and centrifuged to remove surplus biotinylated pembrolizumab.

PDM characterization
We explored PDM morphologies using SEM (Hitachi S-3400N, Tokyo, Japan) and TEM (Hitachi H-7600, Tokyo, Japan) , and determined PDM mean diameters and size distributions via dynamic light scattering (DLS) (Nanosizer-S, Malvern, London, UK). PDMs were also observed using a CLSM (Olympus, FV1000, Tokyo, Japan). Pembrolizumab coupling efficiency was determined by measuring dyLight488-labeled avidin solution and biotinylated DMs suspension absorbances with a fluorescence spectrophotometer (Jasco, FP-6500, Tokyo, Japan) at a maximum excitation wavelength of 493 nm and maximum emission wavelength of 518 nm. Pembrolizumab quantities on biotinylated DMs (binding efficiency (%)) were calculated as the ratio of the intensity of biotinylated DM to the intensity of the dyLight488-labeled avidin samples.

Assessment of DOX loading
Drug encapsulation efficiency was assessed by ultraviolet-visible spectrophotometry (Eppendorf, BioSpectrometer, Hamburg, Germany). A DOX solution standard curve was measured. Then, fresh PDMs were centrifuged and collected. The PDMs were destroyed using a 5% hydrochloric acid ethanol solution, and the mixture was centrifuged at 3000 rpm for 5 min. The optical density of the supernatant was determined at an excitation wavelength of 495 nm. Drug encapsulation efficiency was calculated using the following equation: Encapsulation efficiency ¼ Wa/Wb Â 100%, where Wa represents the total amount of drug in the PDMs, and Wb represents the total weight of DOX used in the PDM preparation (Delplace et al., 2014;Deng et al., 2015;Kim et al., 2015;Parker et al., 2016).

Drug release assay
To estimate DOX release, PDM suspensions were enclosed in dialysis bags (MWCO: 10,000 Da), which were placed in 50 mL of PBS with shaking at 100 rpm at 37 C. The suspension was then sonicated with US (power density ¼ 1.2 W/cm 2 , frequency ¼ 1 MHz, duty cycle ¼ 50%) for 60 s. At 0, 2, 4, 8, 10, 20, 30, 48, 60, and 72 h, 1 mL of dialysate was extracted and stored at 4 C for analysis. An equal volume of PBS was added to the container to insure a constant volume. The concentration of DOX in the sample was determined using an ultraviolet spectrophotometer. DOX release was depicted as a function of time. The DM suspension was assessed using the same method (Santha Moorthy et

Cell cultures
Human lymphoma B cell lines Raji (CD20þ) and Daudi (CD20þ), human lymphoma T cell line Jurkat (CD20-), and human T-acute lymphoblastic leukemia cell line CEM (CD20-), were grown in RPMI-1640 medium with 10% (v/v) fetal bovine serum (FBS, Gibco, Waltham, MA, Australian origin) and 1% penicillin-streptomycin, and incubated in a humidified atmosphere at 37 C with 5% CO 2 . For all experiments, cells growing in suspension were subcultured by centrifugation at a ratio of 1:4. After Raji and CEM cells were anchored in culture dishes with Poly L lysine solution, 50 uL targeted PDMs and non-targeted DMs were added into the dishes. Shaking was used to encourage interactions in Raji cell cultures. After 30 min at room temperature, dishes were washed twice with PBS and observed via CLSM. Five dishes were used for each experiment group. Blocking tests were performed by pre-incubating Raji cells with pembrolizumab for 30 min followed by washing to removing excessive pembrolizumab. CD20-CEM cells were employed as a control, and nonspecific uptake of PDMs by CEM cells was examined using the same methods.

DOX fluorescence intensity
Treatments were as follows: DOX, DOX þ Pembrolizumab, DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab at final DOX concentrations of 0.5 lg/mL. Raji, Daudi, Jurkat, and CEM cells were washed three times with cold PBS after 48 h treatment, centrifuged, and resuspended in 500 lL PBS. Cells treated with medium alone were used as controls. Intracellular DOX retention (red fluorescence) was examined using flow cytometry. Relative fluorescence intensity (RFI) was calculated as: FIexperiment/FIcontrol.

PDM-enhanced contrast ultrasound imaging in vivo
Cells were inoculated subcutaneously into the backs of five nude mice per cell type, with 6 Â 107 Raji cells per mouse and 2 Â 107 Jurkat per mouse. Imaging was performed using an iU22 ultrasound system (Phllips, Amsterdam, Netherland) with a 12 MHz US probe, 0.1 mechanical index, and 54% gain. Mice were anesthetized by injecting 10% hydral and fixed to entirely expose the tumor under the US probe. Nontargeted DMs were injected first for imaging studies in grafted mice. After the expurgation of non-targeted DMs, the same amount of targeted PDMs was injected. The process was monitored continuously by ultrasonography. US contrast data were quantified with PHILIPS QLab version 8.1 software . The arrival time, time to peak, peak intensity, and duration of contrast enhancement were determined. The Animal Ethics Commitment of the Southeast University approved all animal experiments.

In vivo antitumor activity
Raji and Jurkat cell-grafted mice were established additionally as described above. When lymphoma volume reached approximately 100 mm 3 , Raji and Jurkat cell-grafted mice were randomly divided into six groups (five mice per group), respectively: control group (saline), DOX, DOX þ pembrolizumab, DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab. Each mouse was treated with the appropriate formulation (

Statistical analysis
All experiments were performed in triplicate. Data were presented as means ± standard deviation and analyzed using SPSS version 16.0 software (SPSS, Chicago, IL). Comparisons were performed using Student's t-test. p < .05 was considered a significant difference.

PDM characterization
Microbubble-based targeted drug delivery has been widely investigated as an anti-tumor therapy in combination with US irradiation. Most targeted drug delivery systems exhibit high therapeutic efficacies in vitro and in vivo. However, few studies have assessed their potential roles, and the roles of microbubbles, in ultrasonic molecular imaging for diagnosis. In this study, PDMs targeted the lymphoma B cell CD20 antigen. US triggered DOX release, which was then delivered into lymphoma B cells (Figure 1). PDM morphologies and size distributions were observed via scanning electron microscopy (SEM) (Figure 2(A,B)) and transmission electron microscopy (TEM) (Figure 2(C-E)), respectively. Fluorescence imaging of PDMs revealed dense green (DyLight488-labeled avidin) and red (DOX) fluorescence with morphologies consistent with those observed via confocal laser scanning microscopy (CLSM) (Figure 2(D,E)). Targeting moiety quantities on PDM surfaces were evaluated by detecting PDM suspension fluorescence intensity after conjugation. PDM fluorescence intensity was 72.15% that of the DyLight488-labeled avidin samples (Figure 2(F)). Due to the high affinity of avidin to biotin, we presumed the same high level of adhesion of biotinylated pembrolizumab to the avidin-conjugated DMs.

Drug loading and release
PDMs as a targeted drug delivery system were evaluated via the encapsulation efficiency of DOX. DOX encapsulation efficiency in PDMs was 51.2 ± 2.05%. The release profiles of DOX from PDMs and DMs as triggered by US were also examined. The release profile was described as the percentage of cumulative released DOX as a function of time (Figure 2(G)). Total DOX released was the same for DMs þ US and PDMs þ US. The results indicated that DOX was about 50% unloaded after 5 h with sonication and about 90% unloaded after 72 h with sonication. This implied that US could promote DOX release from PDMs and DMs through cavitation.

Targeted properties of PDMs
To estimate the targeted binding capability of PDMs, the affinity of PDMs to CD20 antigen on Raji cells was determined in vitro. PDM attachment to CD20 antigen was greater than that of DMs. CLSM imaging showed large amounts of PDMs (green and red fluorescence) aggregated on Raji cell membranes, demonstrating that pembrolizumab enhanced PDM targeted binding to CD20 antigen. Few DyLight488-labeled avidin-conjugated biotinylated DMs were observed on Raji cell membranes. Competition experiments revealed that PDM targeted binding of Raji cells was reduced as CD20 antigen was blocked following pembrolizumab pre-incubation, as indicated by absence of red and green fluorescent microbubbles. Few PDMs were observed on CD20-CEM cell membranes (Figure 3).

Cell apoptosis in vitro
Raji, Daudi, Jurkat, and CEM cell apoptosis rates were detected quantitatively by flow cytometry 24, 48, and 72 h after various treatments. Raji and Daudi cell apoptosis rates were comparable following DOX, DOX þ pembrolizumab, DMs þ US, and PDMs þ US þ pembrolizumab treatment, although apoptosis was increased in all groups compared to controls. Importantly, PDMs þ US induced higher apoptosis rates than other treatments ( Figure 5(A,C), p < .05). Jurkat and CEM cell apoptosis rates were similar for DOX, DOX þ pembrolizumab, DMs þ US, and PDMs þ US þ pembrolizumab, but higher compared to controls ( Figure  5(B,D)). Additionally, PDM þ US induced higher apoptosis rates in Raji and Daudi cells as compared to Jurkat and CEM cells (p < .05). Time-dependent apoptosis rates were detected in all groups ( Figure 5). Apoptosis rate measurements were consistent with proliferation inhibition results.

Cellular uptake of DOX
Because DOX auto-fluoresces, we detected whether 48 h PDMs þ US treatment improved Raji, Daudi, Jurkat, and CEM cell DOX uptake using flow cytometry. DOX intracellular RFI for Raji and Daudi cells was similar following DOX, DOX þ pembrolizumab, DMs þ US, and PDM þ US þ pembrolizumab treatment. However, DOX intracellular RFI following PDM þ US was higher compared to all other treatments ( Figure 6(A,C), p < .01). DOX intracellular RFI was also similar following all treatments in Jurkat and CEM cells (Figure 6(B,D)). Additionally, DOX intracellular RFI following PDM þ US treatment was higher in Raji and Daudi cells as compared to Jurkat and CEM cells (p < .01). These results suggest that PDM þ US treatment increased DOX transfer into lymphoma B cells more than other treatment groups.

In vivo imaging
Arrival time, time to peak, peak intensity, and duration of contrast enhancement were compared via US imaging between non-targeted DMs and targeted PDMs in Raji and Jurkat cell-grafted mice. In Raji cell-grafted mice, there was no difference between DMs and PDMs for arrival time or time to peak, but peak intensity and duration of contrast enhancement were higher for PDMs (p < .05). In Jurkat cellgrafted mice, there was no difference between DMs and PDMs in any US measurement. Additionally, arrival times and times to peak for targeted PDMs were the same in Raji and Jurkat cell-grafted mice. However, PDM peak intensities and the durations of contrast enhancement were higher in Raji as compared to Jurkat cell-grafted mice (Figure 7, Ã p < .05). Targeted PDM (Figure 8(C,F)) and non-targeted DM ( Figure  8(B,E)) peak intensity images are shown for Raji and Jurkat cell-grafted mice.

In vivo antitumor activity
This study used a lymphoma nude mouse model to investigate the antitumor effects of PDMs þ US in vivo. PDMs þ US exhibited the strongest tumor inhibition effect in Raji-cell grafted mice. DMs þ US, and PDMs þ US þ pembrolizumabtreated mice exhibited similarly reduced Raji cell tumor growth rates compared to controls, and inhibited tumor growth more than treatment with DOX and DOX þ pembrolizumab (Figure 9(A)). DOX and DOX þ pembrolizumab only slightly inhibited tumor growth in vivo. Similarly, DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab treatment reduced grafted Jurkat cell tumor growth as compared with DOX and DOX þ pembrolizumab. Jurkat cell-grafted mouse treatment with DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab resulted in comparable growth inhibition rates (Figure 9(C)).
To assess the potential systemic toxicity of PDMs þ US in vivo, nude mouse body weights were periodically examined. Raji and Jurkat cell-grafted mice treated with DOX and DOX þ Pembrolizumab exhibited slow, continuous weight loss beginning on day 8. In contrast both Raji and Jurkat cell-grafted mouse weights increased gradually with saline, DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab treatments (Figure 9(B,D)). This suggests that DOX treatment caused severe systemic toxicity in nude mice.
TUNEL staining was used to evaluate apoptosis in Raji and Jurkat cell tumors. Sparse apoptosis (green fluorescence) in Raji and Jurkat lymphoma tissues was observed in mice treated with DOX and DOX þ pembrolizumab. Raji cell-grafted mouse tissues treated with DMs þ US, and PDMs þ US þ pembrolizumab showed moderate apoptosis, while PDMs þ US treatment induced the most apoptosis. Jurkat cell-grafted mouse tissues treated with DMs þ US, PDMs þ US, and PDMs þ US þ pembrolizumab showed moderate cell apoptosis. We observed that PDMs þ US induced greater apoptosis levels in Raji as compared to Jurkat cell-grafted mice ( Figure 10). These results confirmed that PDMs þ US could inhibit lymphoma cell growth by inducing apoptosis.

Conclusion
In conclusion, this study indicated that targeted PDMs specifically bound CD20þ B cell lymphomas. PDMs combined with US irradiation enhanced tumor targeting, reduced systemic toxicity, and inhibited B cell lymphoma cell growth in vivo.
Additionally, targeted PDMs increased peak intensity and contrast enhancement duration compared to non-targeted DMs in CD20þ B cell lymphoma-grafted mice. Our findings show that targeted PDMs could potentially be employed as US molecular imaging agents for early diagnosis, and are an effective targeted drug delivery system in combination with Figure 7. PDM arrival time, time to peak, peak intensity, and duration of contrast enhancement in Raji and Jurkat cell-grafted mice. PDM arrival times and times to peak were the same in Raji and Jurkat cell-grafted mice. PDM peak intensities and contrast enhancement durations were greater in Raji cell-grafted mice than in Jurkat cell-grafted mice. Data are represented as means ± SD (n ¼ 3). Ã p < .05, # p > .05. Figure 8. Contrast-enhanced images of targeted PDMs and non-targeted DMs at time to peak in Raji and Jurkat cell-grafted mice. Images of lymphoma before injection A and D, non-targeted DMs B and E, and targeted PDMs C and F at time to peak were acquired in Raji and Jurkat cell-grafted mice. PDM peak intensities and contrast enhancement durations were higher than those of non-targeted DMs in Raji cell-grafted mice, and were higher in Raji as compared to Jurkat cellgrafted mice.
US irradiation for the treatment of CD20þ B cell malignancies.

Disclosure statement
The authors disclose no conflicts of interest.