Scalable Fabrication of High-Payload Dendrimer-Based Nanoparticles for Targeted Atherosclerosis Therapy

Abstract

Nanoparticle-based therapeutics hold promise for the treatment of atherosclerosis, but challenges such as low drug-loading capacity and lack of scalable, controllable production hinder their clinical translation. Flash nanoprecipitation, a continuous synthesis method, offers a potential solution for scalable and reproducible nanoparticle production. In this study, we employed a custom-designed multi-inlet vortex mixer to perform cross-linking reaction-enabled flash nanoprecipitation, facilitating controlled and scalable synthesis of cross-linked polyamidoamine (PAMAM) dendrimer nanoparticles. Notably, this approach allows simultaneous nanoparticle crosslinking and drug loading in a single step. The mannose moiety enabled specific targeting of macrophages via mannose receptors, enhancing the localization of nanoparticles to atherosclerotic plaques. Atorvastatin calcium, a widely used clinical drug for atherosclerosis treatment, was selected as the model drug. This approach achieved both high production rates and high drug-loading capacities, with an output flow rate of 9.6 L/h and a nanoparticle concentration of approximately 0.4 g/L. The optimized formulation exhibited a drug-loading capacity of 37% and encapsulation efficiency of 76%. In vitro and in vivo experiments demonstrated effective macrophage and plaque targeting, leading to significant therapeutic benefits. Treatment with these nanoparticles resulted in approximately 40% inhibition of aortic root plaque progression compared to the free drug-treated group. This scalable and efficient nanoparticle platform is a promising strategy for improving the atherosclerosis treatment.

Graphical Abstract

1. INTRODUCTION

Atherosclerosis, the primary cause of heart disease and stroke, remains a leading cause of death worldwide, especially in developed countries.^{1, 2} As a chronic inflammatory disease, atherosclerosis usually develops for several years, even decades, emphasizing the critical need for early diagnosis and treatment. Current research highlights tracking macrophage recruitment in vessels offers a promising approach for the early detection of atherosclerosis.^{3, 4} In recent years, macrophage-targeted nanomedicines have gained significant attention.^5–7 Among these studies, nanoparticle therapeutics that actively target macrophages have shown potential as effective treatments for atherosclerosis. Various biomarkers, such as CCR2, IL-1R, MPO, and mannose receptors, have been explored as active targeting ligands.⁸ The mannose receptor, a carbohydrate-binding receptor, is selectively expressed by certain macrophages, dendritic cells, and nonvascular endothelial cells.^9–11 In our prior studies, mannose-functionalized polyamidoamine (PAMAM) dendrimers demonstrated efficacy by delivering both LXR and SR-A siRNA to macrophages.¹² Additionally, we developed dual-targeting nanoparticles designed to target both endothelial cells and macrophages, enabling the delivery of LOX-1 siRNA and atorvastatin.¹³ While these nanoparticles exhibit significant therapeutic potential in regressing atherosclerotic plaques, challenges such as complex design and synthesis processes hinder their scalability for industrial production and clinical translation. Furthermore, the limited drug-loading capacity of nanoparticles remains a major obstacle.¹⁴

Typically, nanoparticles that chemically bond drugs can only achieve low drug-loading capacities. High drug-loading capacity is crucial to minimize side effects from carrier materials, reduce costs, and avoid adverse effects from overdosed materials. Physical encapsulation, as opposed to chemical bonding, is a preferred strategy to enhance drug-loading capacity. Therefore, scalable methods that achieve high drug-loading capacities are urgently needed. In this work, we designed new PAMAM dendrimer-based nanoparticles for atherosclerosis therapy. Compared to traditional linear polymers, dendrimers possess unique properties that make them promising candidates for drug delivery. Their surface groups can be modified to reduce toxicity and improve biocompatibility and targeted delivery.¹⁵ Our previous studies have investigated dendrimers both as standalone delivery vehicles and as building blocks for constructing drug and gene delivery systems.^{12, 16–19}

However, achieving high drug-loading capacity via physical absorption remains challenging due to the open structure of dendrimers. This limitation can be addressed by cross-linking dendrimers into nanoparticles, which reduces specific surface area and minimizes uncontrolled drug release. To address these challenges, we developed a scalable method to produce crosslinked dendrimer-based nanoparticles using flash nanoprecipitation via a custom-designed multi-inlet vortex mixer (MIVM).^20–22 Polyethylene glycol (PEG)-mannose was also conjugated to highly positively charged PAMAM G5 dendrimers, with mannose serving as the targeting ligand (Scheme 1). This innovative approach integrates efficient drug loading with improved scalability and controlled drug release, presenting a promising strategy for atherosclerosis therapy.

Scheme 1.

Illustration of ATO-GPM NP synthesis and the desired performance in atherosclerotic lesion.

2. EXPERIMENT SECTION

2.1. Materials

DAB-core PAMAM dendrimer generation 5 was purchased from NanoSynthons (Mt Pleasant, MI). Fluorescein isothiocyanate (FITC) and atorvastatin calcium were purchased from Sigma-Aldrich (St. Louis, MO). IRDye^® 800CW NHS ester (800 CW-NHS) was from LICORbio - U.S. 3.4k Da PEG-mannose with NHS eater (NHS-PEG-mannose) and 3.4k Da PEG were purchased from Ruixi (Xi’an, China). 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester) was synthesized following the method from Li et al.²³ The ¹H-NMR result can be found in Figure S1.

2.2. Synthesis of G5-PEG-Mannose (GPM) and NIR 800CW /FITC labelled GPM

PAMAM G5 was dissolved in 0.2 mM NaHCO₃ aqueous solution with a concentration of 10 mg/mL. 3.4k Da PEG-mannose with NHS eater (NHS-PEG-mannose) was dissolved in DI water with a concentration of 11.7 mg/mL. Then, equal volumes of both solutions were mixed and stirred for 12 h at room temperature allowing the reaction between N-hydroxysuccinimide ester and amine group. The final solution was purified by dialyzing against DI water (3500 Da molecular weight cut-off). The DI water was replaced with fresh DI water thrice every 12 h. GPM can be obtained as a white powder after freeze-drying.

To get NIR 800CW /FITC labeled GPM, GPM was dissolved in DI water with a concentration of 1mg/mL. A proper amount of FITC in a molar ratio of 5:1 (FITC: G5) was dissolved in DMSO (5 mg/mL). An appropriate amount of FITC solution was added into the GPM solution and kept under stirring for 12 h at room temperature. For NIR 800CW labeled GPM, 800CW-NHS aqueous solution (1 mg/mL) was mixed with GPM solution (1mg/mL) in a molar ratio of 1: 10 (800CW-NHS: G5). The mixture is stirred at room temperature for 12 h. Both samples were purified by dialysis and freeze-dried as described before.

2.3. Synthesis of G5-PEG-Mannose Nanoparticles (GPM NPs)

As shown in Scheme 1, GPM was cross-linked via flash nanoprecipitation method. As reported in our previous work, Four inlet vortex mixers were designed and used to synthesize nanoparticles.²¹ Two inlets separately contained GPM (1 mg/mL) aqueous solution and 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester) (0.06 mg/mL) acetone solution, while the other two were used for DI water. The pumps propelled the four solutions into MIVM at a speed of 40 mL/min. The product solution was processed with a rotary evaporator to remove acetone. Final products could be obtained after dialysis and freeze drying, as described before.

2.4. Synthesis of atorvastatin calcium-loaded GPM NPs (ATO-GPM NPs)

ATO could be loaded by replacing the GPM solution with GPM and ATO mixture, which was premixed and kept under stirring for 12 h at room temperature. As described above, a similar cross-linking experiment was conducted to synthesize ATO-GPM NPs via MIVM with a GPM/ATO mixture and 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester) acetone solutions. Then, an ultra-15 centrifugal filter unit was used to remove acetone and free drug instead of rotary evaporation. The ATO-GPM NPs were obtained after freeze-drying. To determine encapsulation efficiency (EE%) and loading capacity (LC%), the formulations were subjected to LC-MS analysis. The EE% and LC% were calculated using the following formulas:

$$\text{Encapsulation}\text{efficiency}\left(\text{EE\%}\right)=\frac{\text{total}\text{amount}\text{of}\text{drug}\text{added}–\text{nonencapsulated}\text{drug}}{\text{total}\text{amount}\text{of}\text{drug}\text{added}}\times 100\%$$ (1)

$$\text{Loading}\text{capacity}\left(\text{LC\%}\right)=\frac{\text{total}\text{amount}\text{of}\text{drug}\text{in}\text{the}\text{product}}{\text{total}\text{nanoparticle}\text{mass}}\times 100\%$$ (2)

2.5. Synthesis of G5-PEG (GP), FITC labeled G5-PEG (FITC-GP), and G5-PEG nanoparticles (GP NPs)

Like the synthesis of GPM and FITC labeled GPM, NHS-PEG was used to substitute NHS-PEG-Mannose.

2.6. Material characterization

The nanoparticle size distribution was measured by using dynamic light scattering (DLS) (Malvern Panalytical Zetasizer Ultra) at 25 °C. The morphology was also characterized by TEM (JEOL JEM-1400). The conjugation between G5 and PEG-mannose and 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester) chemical structure was verified by liquid NMR (Bruker 400 MHz Avance III HD Liquid State NMR). Atorvastatin calcium concentrations were determined with LC/MS (Shimadzu 2020 Single Quadrupole LC/MS). The LC/MS method and results are provided in supporting information (Experiment section, Figure S2 and Figure S3).

2.7. Cell culture

RAW 264.7 and HUVEC cell lines were used in this study. RAW 264.7 cells were cultured with Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). HUVEC was cultured with endothelial cell growth medium with supplement Mix from PromoCell. Both cell lines were incubated at 37 °C in 5% CO₂.

2.8. Cytotoxicity assay

10,000 RAW 264.7 cells or 5,000 HUVEC cells were suspended in 100 μL corresponding culture medium and seeded in 96 well-plates for 12 h. Then, the medium was removed and replaced with fresh medium, GPM, or GPM NP medium solutions. 6 different concentrations were tested from 0.5 to 30 μg/mL. After 24 h, the cell viability was determined by CCK-8 cell proliferation method following the manufacturer’s protocol.

2.9. Cellular uptake kinetics of GPM NPs, GP NPs, ATO-GPM NPs

RAW 264.7 cells were seeded in 6-well plates at a density of 500,000 cells/well. The cells were allowed to seed and grow for 12 h. Then, the culture medium was replaced with fresh medium, FITC-labelled GP NP solution, and FITC-labelled GPM NP solution. Both GPM NP and GP NP solutions had a concentration of 10 μg/mL. After 24 hours, cells were harvested and tested with flow cytometry or a confocal microscope. For the confocal microscopy test, the cells were fixed with 4% paraformaldehyde for 15 min, followed by 0.1% Triton X-100 treatment for 15 min. Next, the cells were stained with DAPI (5 μg/mL) for 5 min. After rinsing with PBS, cells were stored at 4 °C until imaging. For the flow cytometry test, cells were rinsed with PBS three times. Then, the cells were harvested and suspended in 1 mL PBS for tests.

To test the ATO uptake, RAW 264.7 cells were seeded in 6-well plates at a density of 1,000,000 cells/well for 12h. The medium was replaced with ATO medium solution and ATO-GPM NP medium solution. ATO solution and ATO-GPM NP solution had an equal ATO concentration of 10 μg/mL. The cells were rinsed with PBS and harvested at different times: 4, 8, 12, 16, and 24 h. Samples were diluted and homogenized in 0.5 mL methanol and MQ water mixture (1: 1 in volume). The concentrations of ATO in samples were further quantified with LC-MS.

2.10. Atherosclerosis model mice establishment

The low-density lipoprotein receptor knock-out (LDLR^−/−) mice were used in this study. 4-week-old LDLR^−/− mice were randomized into four groups (n=15). After being fed with a high-fat diet for 16 weeks, the atherosclerosis model mice were obtained. Then, the feeding diet was switched to a chow diet during in vivo studies. All procedures used in the present study were conducted according to the protocol approved by the Institutional Animal Care and Ethical Committee of Missouri University of Science and Technology (Protocol No. 190–22).

2.11. Plaque targeting effect evaluation in atherosclerotic model mice

Atherosclerosis model mice were divided into 3 groups (n=3) and injected with different formulations via tail vein: PBS, NIR 800CW carboxylate, and 800CW labeled GPM NPs. The latter two solutions had an equal concentration of 800CW (16 μg/mL). After 24 h, the aortas were collected and imaged with an animal imager (AMI HTX, Spectral Instruments Imaging) with 710 nm excitation wavelength and 770 nm emission wavelength.

2.12. Pharmacodynamics study in atherosclerosis model mice

Four groups were studied in this experiment. The baseline group mice were sacrificed after the high-fat diet feeding. ATO and ATO-GPM NP groups were intravenously injected with ATO solution or ATO-GPM NP solution at an equivalent dose of 0.1 mg ATO/mouse twice a week via tail vein. The untreated group was continuously fed with a chow diet without any treatment. Aorta samples were collected after 6-week treatments. Both aortic lesions (n=10) and aortic root plaque (n=3, 3 sections for each mouse) areas were quantified. Aortic lesion samples were stained with oil red O. The aortic root sections were stained with hematoxylin and eosin (H&E). Plasma samples were also collected and tested with Amplex^™ red cholesterol assay kit for cholesterol level estimation following manufacturer instructions.

2.13. Tissue distribution and safety study in wild-type C57BL/6 mice

Three groups of 16-week-old C57BL/6 mice (n=3) were used for the biodistribution test, which was treated with different formulations via tail vein including PBS, NIR 800CW carboxylate PBS solution, and 800CW labeled GPM NP PBS solution. The latter two solutions had an equal concentration of 800CW (16 μg/mL). The main organs (heart, spleen, kidney, lung, and liver) were collected and imaged 24 h post-injection by an animal imager with a 710 nm excitation wavelength and 770 nm emission wavelength.

Two groups of 16-week-old C57BL/6 mice (n=3) were used for the safety study. PBS (100 μL/mouse) or GPM NPs (10 mg/mL, 100 μL/mouse) were injected once a day for three successive days. All the mice were sacrificed on the fourth day. The main organs (heart, spleen, kidney, lung, and liver) were collected and followed with paraffin sectioning and H&E staining. The slices were imaged with Leica ICC50W.

3. STATISTICS AND REPRODUCIBILITY

All the in vitro and in vivo experiments were performed on a minimum of three biological replicate. Data were presented as mean ±standard deviation. And a one-way analysis of variance (ANOVA) followed by Tukey’s test for multiple comparisons. T-test was applied for two group comparisons. All the statistical analyses were performed using OriginLab Origin 2021 software. And P values less than 0.05 were considered statistically significant (*P < 0.05).

4. RESULTS AND DISCUSSION

Flash nanoprecipitation method has been one efficient way to scale up nanoparticle production. As a “flash technology”, this process occurs in MIVM in a high-throughput and well-controlled way. Repeatability is much improved compared with laboratory commonly used batch methods. It shows high scalability, while the cost can be kept low compared with flow methods like microfluidics and membrane-based processes. The MIVM-based flash nanoprecipitation method would promote the clinic’s translation of nanoparticle delivery systems.^{22, 24–26} In this study, we developed ATO-GPM NP for atherosclerosis therapy via flash nanoprecipitation. A four-inlet vortex mixer is designed to conduct this synthesis method.

4.1. Preparation and characterization of nanoparticles

As shown in Figure 1a, nanoparticle size distributions of different formulations are tested with DLS at different synthesis steps. It can be found that all these samples show uniform size distributions. The sizes increase from 8 nm to 25 nm after PEG-mannose modification. The size further increases to about 190 nm after cross linking and ATO loading. Figure 1b confirms the spherical shape of GPM NP with TEM image. Meanwhile, GPM NP remains stable for up to 11 days (Figure S4). PEGylation is widely studied in relevant research, which can neutralize the highly charged surface and decrease cytotoxicity.^27–29 Generally, a positively charged nanoparticle surface can promote cell uptake because of the absorption of a negatively charged cell membrane. However, the highly positive charge can also cause severe cytotoxicity and damage tissue or organs. G5 surface was efficiently modified in this study with NHS activated PEG-Mannose, which is confirmed with ¹H-NMR. As shown in Figure S5, the peak located at 3.58 ppm corresponding to -O-CH₂-CH₂- appears in GPM spectrum. The molar ratio is calculated based on the area of peaks indicating a 10: 1 ratio between PEG-mannose and G5. Zeta potential of G5 decreases to 11.3 mV from 17.3 mV after PEGylation (Figure S6). It is further reduced to 4.1 mV, when it is modified with mannose terminated PEG.

Figure 1. Nanoparticle characterization and drug loading efficiency.

(a) DLS size distribution of G5, GPM and ATO-GPM NPs; (b)TEM image of GPM NPs; (c) DLS size distributions of prepared ATO-GPM NPs with different ATO concentrations; (d) EE% and LC% of ATO-GPM NPs synthesized with different ATO concentrations (GPM NP concentration was fixed at 1.0 mg/mL).

4.2. Drug encapsulation efficiency and loading capacity study

PAMAM G5 and PEG complex are employed as a vehicle to load ATO for drug delivery. PAMAM G5 dendrimer confers hollow structures and sufficient terminal amine groups for modifications.^30–32 However, the drug-loaded PAMAM G5 dendrimer cannot usually provide a high drug-loading capacity for its open hollow structure. Our previous studies show constructing nanoparticles or microparticles with single dendrimer molecules can be a feasible method to increase drug loading capacity and achieve a controllable drug release.^{16, 17, 33} The cross-linker can also be another anchor that can be modified to functionalize nanoparticles. In this study, 3,3′-Dithiodipropionic acid di(N-hydroxysuccinimide ester) is used as a cross-linker, which can react with GSH and promote nanoparticle intracellular degradation.

Different ATO loading concentrations are optimized for high encapsulation efficiency (EE%) and good loading capacity (LC%). As exhibited in Figure 1c, similar size distributions are acquired when the ATO concentration increases from 0.2 to 0.6 mg/mL. A size around 200 nm is observed when the ATO concentrations are 0.8 and 1.6 mg/mL. Figure 1d gives the EE% and LC% values under different synthesis conditions with a fixed GPM concentration of 1 mg/mL. EE% decreases with the increase of ATO concentrations, while LC% continuously decreases. Considering the data of EE% and LC%, a good size distribution (~200 nm), high encapsulation (76%), and high loading capacity (37%) are achieved with an ATO concentration of 0.8 mg/mL.

4.3. Low cytotoxicity of GPM NPs to RAW 264.7 and HUVEC cells

Directly modifying dendrimers with targeting ligands can be an option for active targeting nanoparticle design. As we reported before, Folic acid-modified dendrimer shows good targeting effects on head and neck cancer cells expressing high levels of folate receptors.³⁴ However, this method can not efficiently shield positive charge, especially for crosslinked nanoparticles. In this study, the PEGylated nanoparticle was employed to decrease surface charge to obtain low cytotoxicity. Macrophage (RAW 264.7) and endothelial (HUVEC) cells were selected for cytotoxicity assays in this study, as both cell types play important roles in the development of atherosclerosis. The cytotoxicity of different formulations is estimated including G5, GPM NP, and GPM NP. In a concentration range from 0.5 to 30 μg/mL, both GPM and GPM NP do not show significant toxicity with cell viabilities around 100% (Figure 2a). However, G5 exhibits significant toxicity to RAW 264.7 cells, when the concentration is higher than 10 μg/mL. At a concentration of 30 μg/mL, the cell viability is only 30%. Similar behaviors are observed from GPM and GPM NP to HUVEC cells without significant toxicity from 0.5 to 30 μg/mL. However, G5 also shows low cytotoxicity with HUVEC cells in this concentration range (Figure 2b). This difference may be caused by the different cell properties. As a kind of macrophage, RAW 264.7 usually has high activities, which leads to a high interaction with nanoparticles. In contrast, HUVEC is an endothelial cell and shows good stability which makes it endure high concentration of nanoparticles.

Figure 2. Cytotoxicity of G5, GPM and GPM NP.

(a) RAW 264.7 cell and (b) HUVEC cell viabilities were tested after incubated with G5, GPM and GPM NPs for 24 h in a concentration range from 0.5 to 30 μg/mL, (n=5).

4.4. Enhanced macrophage uptake of GPM NPs and ATO-GPM NPs

Mannose is designed as the terminal molecule for atherosclerotic plaque targeting. It has been studied and reported that mannose receptor is highly expressed on the surface of macrophages.^{9, 35–37} It is involved in a variety of pro- and anti-inflammatory responses. Generally, glucose is widely used in atherosclerosis studies. For instance, ¹⁸F-fluorodeoxyglucose has been approved by the FDA and applied for atherosclerosis PET imaging. However, the uptake of ¹⁸F-fluorodeoxyglucose can be easily affected by factors like hypoxia and increased myocardial muscle activity, which will decrease its effectiveness. Mannose can not only be taken through the glucose transporter, but it can also bond to the mannose receptors on the macrophage surface.^{11, 38} This pathology makes it more specific and compelling to target atherosclerotic plaques.

The specificity of nanoparticle to RAW 264.7 cells is studied by comparing FITC labelled GPM NP and GP NP. Cellular uptake results are analyzed by both a confocal microscope and flow cytometry. In this experiment, both kinds of nanoparticles are labeled with FITC. The cell nucleus is stained with DAPI for confocal imaging. As shown in Figure 3a, a higher fluorescence intensity from FITC (green) is obtained by the RAW 264.7 cells incubated with GPM NPs for 24 h, compared with both untreated and GP NP groups. Good colocalization is also observed, indicating successful uptake of nanoparticles (Figure S7). Flow cytometry results are consistent with confocal microscopy images (Figure 3b). A shift can be found between GP NP and GPM NP groups. The mean intensities are 1248, 9253, and 22123 for the untreated group, GP NP treated, and GPM NP treated groups, respectively.

Figure 3. In vitro analysis of GPM NP and ATO cellular uptake with RAW 264.7 cells.

(a) Fluorescent images and (b) flow cytometry analysis of RAW 264.7 cells after incubated for 24 h with different formulations: Untreated, GP NP (10 μg/mL) and GPM NP (10 μg/mL); (c) ATO drug concentrations after treated with ATO (10 μg/mL) or ATO-GPM NPs (ATO equivalent concentration 10 μg/mL), (n=3, *P<0.05).

GPM NP is used as a drug vehicle in this study, when ATO is used as a model drug. A higher cellular uptake is also verified due to an efficient targeting effect to mannose receptor. The ATO concentrations acquired at different time points are summarized in Figure 3c. No statistical concentration difference can be found after 2 h incubation with ATO and ATO-GPM NPs. Compared with ATO, ATO-GPM NP gives significantly higher drug concentrations at 4 h and 8 h, which are about 3 folds of those values from the ATO group. However, the concentrations are close to each other at 12, 16 and 24 h. Generally, small-molecule drugs can easily permeate cell membranes, whereas nanoparticles are internalized via endocytosis. Consequently, the intracellular accumulation rates differ significantly between free drugs and drug-loaded nanoparticles. Additionally, the rate of metabolism depends primarily on the intracellular drug concentration. Therefore, the observed phenomenon may be attributed to the rapid permeation and metabolism of the free drug, compared to the slower accumulation and reduced release rate of drug-loaded nanoparticles.

4.5. Enhanced targeting effects of GPM NP in atherosclerotic model mice

Atherosclerosis model mice were tested with GPM NP for plaque-targeting studies. 24 h post-injection of PBS, 800CW, and 800CW labeled GPM NP, the whole aortas were collected by heart dissection and imaged with an animal imager. All three groups are compared in a count range of 100 to 200, as shown in Figure 4a. Strong signals can be seen in the aorta roots and ascending thoracic aortas from the 800CW-GPM NP group, while only slight fluorescence was observed in the 800CW group, and no signal was from the PBS group. In all groups, no signals can be seen in descending thoracic and abdominal aorta, indicating an early stage of atherosclerosis. The signal intensities are also quantified and exhibited in Figure 4b. For the 800CW-GPM NP group, the count values are about 22000, while only about 1300 and 1500 counts can be obtained in PBS and 800CW groups, respectively. 14 times higher intensity from the 800CW-GPM NP group than those of PBS and 800CW groups, which means a sufficient specific targeting effect on macrophages.

Figure 4. Targeting effects of GPM NP in atherosclerosis model mice.

(a) Fluorescence images and (b) quantified fluorescent intensities of aortas from atherosclerotic model mice treated with 800CW-GPM NP, 800CW and PBS for 24 h, (n=3).

In this experiment, free 800CW acts like small molecular drugs usually washed out within several hours. Enhanced cycling time allows more targeting opportunities and improves the body’s drug concentration level. In vitro experiments also show an enhanced drug concentration level. This is attributed to the well-designed nanoparticle structure and GSH-triggered nanoparticle degradation.^{39, 40} The degraded formulations may be eliminated relying on hepatic clearance and renal excretion. As shown in Figures 6a and b, the high fluorescence intensity in the kidney and liver indicates an accumulation of nanoparticles in corresponding organs.

Figure 6.

Biodistribution and safety study of GPM NPs. (a) Typical fluorescent images and (b) quantified fluorescent intensities of different organs from 800CW-GPM NP, 800CW and PBS treated C57BL/6 mice, (n=3); (c) H&E histopathological staining of major organs (heart, spleen, kidney, lung and liver) after GPM NP injection for three days. Scale bar: 100 μm.

4.6. Enhanced anti-atherosclerotic efficacy of ATO-GPM NP in vivo

Furthermore, an enhanced therapy effect is observed compared with ATO. Figure 5a gives the in vivo experiment design. As shown, the diet for LDLr^−/− mice was switched to a chow diet and treated with different formulations for 6 weeks after 16-week high-fat diet feeding. Plaque areas are measured before (baseline group) and after treatments with different formulations (untreated, ATO-free drug, and ATO-GPM NP groups). Typical images are shown in Figure 5b. The plaque can be observed localizing in the aorta roots and arch parts, indicating an early stage of atherosclerosis. No visible plaques can be found after Oil red O staining in the descending aorta. The plaque areas are quantified and summarized in Figure 5c based on the images in Figure S8. It shows the plaques continuously increased after the diet switched to a chow diet from a high-fat diet. The colors of untreated, ATO, and ATO-GPM NPs group aortas exhibit deeper color than the baseline group, which may infer thicker plaques. The quantified data shows that all the others show a significant area increase compared with the baseline group. Both ATO and ATO-GPM NPs treated groups show inhibition of plaque progression by mean values. A significant plaque area decrease of 25% is observed between untreated and ATO-GPM NPs treated groups, which decreased from 10.3 to 7.7 mm². However, the differences are not significant between untreated and ATO groups or ATO and ATO-GPM NP group.

Figure 5. Assessment of anti-atherosclerotic efficacy of GPM NP in atherosclerosis model mice.

(a) Schematic of the timeline and treatment protocol for atherosclerosis model mice; (b) typical aorta images and (c) plaque area quantifications of baseline, untreated, ATO and ATO-GPM NP treated groups; (d) typical aortic root images (stained with H&E) and (e) plaque area quantifications of baseline, untreated, ATO and ATO-GPM NP treated groups (n=3, 3 sections for each mouse, *P<0.05).

Aortic root plaques are also characterized and quantified in Figure 5d and 5e. Uniformed plaque distribution is observed in untreated group. And necrotic cores are observed in all the other groups. Both untreated and ATO treated groups have an average plaque area of about 0.6 mm². An inhibition is found with ATO-GPM NP group, which has the smallest area of 0.35 mm² smaller than baseline group (0.42 mm²). About 40% shrinkage of plaque area is verified in ATO-GPM NP group comparing with untreated group. The anti-atherosclerosis efficacy is estimated from both aortic lesion and aortic root areas from longitude and latitude, giving a 3-dimensional understanding of plaque development.

Atherosclerosis continuously progresses after the diet is switched to the chow diet. It has been widely discussed that atherosclerosis is affected by multiple factors.^41–43 Diet adjustment can not effectively slow down atherosclerosis progression. The ATO-GPM NP-treated group exhibits significant lesion area inhibition compared with the untreated group, while the difference between the untreated and ATO-treated groups is insignificant. Regarding the aortic root areas, ATO-GPM NP group values are close to the baseline group but much smaller than the untreated group. Effects of ATO have been reported in recent years, including plaque stabilization, cholesterol-lowering, and anti-inflammation.^{44, 45} However, there are no differences in cholesterol levels among the untreated ATO and ATO-GPM NP groups in this study (Figure S9). A decrease is only found between the baseline and the other three groups. ATO is a 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase inhibitor. By inhibiting HMG CoA reductase, statins reduce the hepatocyte cholesterol content and increase the expression of low-density lipoprotein (LDL) receptors, responsible for LDL cholesterol uptake via receptor-mediated endocytosis.⁴⁶ However, in this study, LDLR^−/− mice are used for therapy study. And this may be the reason that no significant cholesterol level decrease is observed in treated groups. So, we may claim that the anti-atherosclerosis benefits are from the effective targeting effects and the interactions between the ATO and cells in lesions.

4.7. Biodistribution of GPM NP and biocompatibility study

Nanoparticle distribution in main organs is studied with C57BL/6 mice 24 h post injections. As shown in Figure 6a, the fluorescence of the GPM NP group is mainly concentrated in the kidney and liver after 24 h compared with the 800CW and PBS-treated groups. All the organs in the GPM NP group show significantly higher intensities than those of the PBS groups, which are about 2.8 times for the heart, 8.5 times for the spleen, 22 times for the kidney, 5.3 times for the lung, and 6.6 times for the liver. No statistical differences are found between 800CW and PBS groups (Figure 6b). Nanoparticle distribution at various time points was also monitored and is presented in Figure S10. Rapid accumulation and metabolism in the liver and kidneys were further confirmed by quantifying fluorescence changes in major organs.

GPM NP also exhibits good biocompatibility (Figure 6c). After three days of treatment with GPM NP (100 μL, 1 mg/mL) and PBS (100 μL), no apparent differences are observed from H&E-stained slices including heart, spleen, kidney, lung and liver. Furthermore, the mouse body weight changes were monitored during the treatments (Figure S11). Boy weights slightly decreased within −10% for all groups. And no toxicity of nanoparticles can be confirmed based on body weight changes. Meanwhile, the in vivo studies also show good biocompatibility of GPM NP. No significant pathological damage is found in major organs. No noticeable weight loss is observed in ATO-GPM NP treatment.

5. CONCLUSIONS

In this study, a feasible flash nanoprecipitation method is developed via a homemade MIVM device. Both crosslinking and drug loading reactions are simultaneously achieved, which is different with the conventional practice by MIVM. It is applied in the synthesis of dendrimer-based nanoparticles for drug delivery. A scalable production is achieved with a 4 g/h nanoparticle flow rate. A high drug loading capacity of 37% is achieved with the atorvastatin calcium model drug. With the help of mannose as a targeting molecule to macrophages, this well-designed nanoparticle improves the therapeutic effects of atherosclerosis compared with free drugs. The sufficient targeting effect is also verified in both in vitro and in vivo experiments. This PAMAM dendrimer-based nanoparticle exhibits great potential for scale production and a promising future in atherosclerosis therapy.

Supplementary Material

LC/MS method for atorvastatin calcium quantification, ¹H-NMR spectrum, zeta potential and stability of different formulations, single RAW 264.7 cell confocal images, Oil Red O stained aorta images, plasma cholesterol fluorescence analysis.