Cellular Interactions of Liposomes and PISA Nanoparticles during Human Blood Flow in a Microvascular Network

Abstract

A key concept in nanomedicine is encapsulating therapeutic or diagnostic agents inside nanoparticles to prolong blood circulation time and to enhance interactions with targeted cells. During circulation and depending on the selected application (e.g., cancer drug delivery or immune modulators), nanoparticles are required to possess low or high interactions with cells in human blood and blood vessels to minimize side effects or maximize delivery efficiency. However, analysis of cellular interactions in blood vessels is challenging and is not yet realized due to the diverse components of human blood and hemodynamic flow in blood vessels. Here, the first comprehensive method to analyze cellular interactions of both synthetic and commercially available nanoparticles under human blood flow conditions in a microvascular network is developed. Importantly, this method allows to unravel the complex interplay of size, charge, and type of nanoparticles on their cellular associations under the dynamic flow of human blood. This method offers a unique platform to study complex interactions of any type of nanoparticles in human blood flow conditions and serves as a useful guideline for the rational design of liposomes and polymer nanoparticles for diverse applications in nanomedicine.

An innovative methodology to characterize cellular interactions of nanomaterials under human blood flow conditions is developed. Fresh whole human blood and an artificial microvascular network are employed for the first time. This platform is easy to set up in any lab and can be applied to future studies of any type of nanoparticles.

1. Introduction

The application of nanoparticles in delivering diagnostic and therapeutic agents provides great benefits to patients including enhancing the accuracy of diagnosis, reducing side effects, and increasing therapeutic efficacy.^{[ 1} , ² , ^{3 ]} Liposomes such as Doxil are currently used in clinics to achieve such benefits in cancer treatment while other lipid nanoparticles (LNP) carrying messenger RNA (mRNA) find potential use as safe and effective vaccines.^{[ 4} , ⁵ , ^{6 ]} A relevant example is the recent approval of an mRNA LNP vaccine candidate by the U.S. Food and Drug Administration for the first clinical trial against the 2019 coronavirus (SARS‐CoV‐2).^{[ 7} , ^{8 ]} In addition, polymeric nanoparticles have demonstrated a wide variety of applications ranging from drug and gene delivery to nanocarriers for imaging contrast agents and non‐opioid drugs to treat chronic pain.^{[ 9} , ¹⁰ , ^{11 ]}

In the majority of these applications, nanoparticles first interact with human blood in blood vessels before reaching disease targets (tissues or cells).^{[ 12 ]} For example, upon administering Doxil nanoparticles, they interact with human blood during circulation, and only a small proportion of these nanoparticles can accumulate in tumors.^{[ 13} , ¹⁴ , ^{15 ]}While interactions of nanoparticles with cancer cells have been extensively studied, their interactions with healthy cells in human blood and blood vessels remain unclear as it is challenging to study.^{[ 16 ]} Understanding such important nano–bio interactions in blood vessels would allow the design of safe and effective nanoparticles for diverse applications in nanomedicine. For instance, when designing nanoparticles for cancer drug delivery, potential binding to healthy cells in human blood and blood vessels should be minimized to suppress side effects, such as immune response and cardiotoxicity.^{[ 17 ]} In contrast, to achieve safe vaccines and immunomodulators with good efficacy, the high association of nanoparticles with the antigen‐presenting cells such as B cells, monocytes/macrophages, and dendritic cells is highly beneficial while their interactions with the remaining healthy cells should be avoided.^{[ 18} , ¹⁹ , ^{20 ]}

Despite the importance to analyze the cellular interactions of nanoparticles during blood circulation, so far no study has investigated this analysis comprehensively.^{[ 21 ]} The analysis of cellular interactions during human blood flow is a long‐standing challenge due to the diverse composition of human blood and the hemodynamic flow in blood vessels. In particular, human blood is composed of plasma (containing salts, lipids, proteins, vitamins, hormones, and water) and importantly, a variety of cell types (e.g., monocytes, neutrophils, B cells, dendritic cells (DCs), T cells, and natural killer (NK) cells).^{[ 22} , ²³ , ^{24 ]} In addition, the shear rate of human blood flow varies widely as human blood vessels have a very complex geometry (not straight vessels).^{[ 25 ]} Therefore, to analyze cellular interactions of nanoparticles in blood vessels, three key requirements should be considered. First, fresh whole human blood should be used to provide a realistic biological environment, even if the cellular interaction of nanoparticles with only a specific blood component is studied for a particular application.^{[ 26} , ²⁷ , ^{28 ]} Investigating interactions with a single blood component may underestimate the complex interplay of the remaining blood components; for example, it has been well‐documented that proteins in blood plasma can form a corona on the surface of nanoparticles and completely change the way nanoparticles interact with cells.^{[ 29} , ³⁰ , ^{31 ]} In addition, blood cells and nanoparticles flow in blood vessels unevenly as they exhibit different degrees of margination to the vascular walls of blood vessels.^{[ 16 ]} As such, traditional cellular interactions in static conditions (incubating nanoparticles and blood in wells) cannot accurately model dynamic blood flow and particle margination.^{[ 32 ]} Therefore, a flow assay for analysis of nanoparticle–cell interactions in human blood vessels is required. Last but not least, an artificial microvascular network with complex geometry is needed to model human blood vessels.^{[ 33} , ³⁴ , ^{35 ]} However, taking into account all of these requirements in a single method is very challenging and has not yet been realized.

In this work, we aim to develop the first comprehensive method to analyze cellular interactions of nanoparticles under blood flow conditions in a microvascular network. Our method offers a universal and versatile platform to study such complex interactions by benefiting from the use of, for the first time, fresh whole human blood under flow conditions in a synthetic microvascular network. In particular, fluorescently labeled nanoparticles and human blood were injected into a synthetic microvascular network with a complex geometry of microchannels that are designed to mimic blood vessels. Human blood containing nanoparticles was subsequently collected and incubated with a mixture of judiciously selected fluorescently labeled antibodies. Each antibody bound to a specific cell type in the blood, allowing the identification of cellular interactions by flow cytometry (i.e., which nanoparticles bound to each cell type). Importantly, we applied our methodology to both liposomes (a clinically relevant type of nanoparticles) and polymeric nanoparticles (synthesized by reversible addition‐fragmentation chain transfer and polymerization‐induced self‐assembly (RAFT‐PISA)) with diverse potential applications. As an exemplary demonstration of the capabilities of our platform, the effect of the size and charge of nanoparticles on their interactions with six primary white blood cells (WBCs) was also evaluated for the first time.

2. Results

2.1. Nanoparticle Synthesis and Characterization

First, we identified RAFT‐PISA as a suitable method to prepare polymeric nanoparticles with three different sizes (small particles ≈40 nm (SP), medium particles ≈75 nm (MP), and large particles ≈150 nm (LP), see Figure 1 ).^{[ 36} , ³⁷ , ^{38 ]} Specifically, poly(ethylene glycol) (PEG) methyl ether acrylate and N‐hydroxyethyl acrylamide were judiciously selected as monomers to synthesize a water‐soluble, biocompatible, and end‐functional polymer by RAFT.^{[ 39 ]} This PEG‐based polymer was then used, through PISA, to form the corona (outer shell) of all three PEGylated nanoparticles (Figure 1). PEGylated nanoparticles are widely used in nanomedicine due to the excellent stabilizing and anti‐fouling properties of PEG.^{[ 40} , ⁴¹ , ^{42 ]} All nanoparticles had identical surface and three different sizes as a result of varying only the degree of polymerization (DP) (or molecular weight) of polystyrene in the core (Table 1 ). Additionally, all three particles have a carboxylic acid functional group on the surface (to provide negative charge at pH 7.4) and a far‐red fluorescent dye (Cyanine 5) covalently attached to the particle core (to provide stable fluorescent signals and prevent potential leakage in solution). These particles were carefully designed to minimize the effect of chemical components on their cellular interactions, allowing studying the exclusive effect of particle size. Subsequently, the polymeric particles were exhaustively characterized by various tools to enable quantitative comparisons between all nanomaterials.^{[ 43 ]} For particle size, we characterized the nanoparticles by dynamic light scattering (DLS), transmission electron microscopy (TEM), and nanoparticle tracking analysis (NTA or NanoSight). Results are summarized in Table 1 showing consistent sizes with all characterization methods being in good agreement. The PISA nanoparticles also exhibit high uniformity, as confirmed by TEM images in Figure 1, and low polydispersity values measured by DLS (PDI < 0.2). The zeta potential and fluorescent intensity results indicated that all particles had a relatively similar negative charge (≈−10 mV) and had been successfully labeled with the Cyanine 5 dye (Table 1 and Figure S1A, Supporting Information). In summary, three fluorescently labeled PISA nanoparticles (small, medium, and large size) were carefully designed, successfully synthesized, and thoroughly characterized for further study on the exclusive effect of particle size.

Figure 1.

Nanoparticle design and characterization. A) Top: Schematic illustration of negatively charged PISA nanoparticles with three different sizes of ≈40 nm (small particle—SP), ≈75 nm (medium particle—MP), and ≈150 nm (large particle—LP); Middle: Chemical structure of diblock copolymers of the PISA nanoparticles; Bottom: TEM images of the PISA nanoparticles. B) Top: Schematic illustration of liposomes with three different charges: anionic liposomes (AL), cationic liposomes (CL), and neutral liposomes (NL); Middle: Schematic illustration of lipids used to make liposomes with distinct charges; Bottom: Cryo‐TEM images of the liposomes. Scale bars = 200 nm.

Table 1.

Nanoparticle characterization

Nanoparticle code		SEC a)		DLS b)			TEM c)	NTA
		M n[g mol−1]	Ð	d [nm]	PdI	ζ [mV]	d [nm]	d [nm]
PISA	SP	47300	1.36	42 ± 2	0.11	−15 ± 1	39 ± 5	55 ± 1
	MP	102500	1.41	76 ± 1	0.03	−12 ± 1	73 ± 6	82 ± 1
	LP	149500	1.40	154 ± 4	0.08	−14 ± 1	142 ± 16	158 ± 2
Liposomes	AL	–	–	115 ± 4	0.12	−17 ± 2	103 ± 19	135 ± 1
	CL	–	–	115 ± 4	0.16	4 ± 1	104 ± 19	127 ± 11
	NL	–	–	128 ± 3	0.15	−3 ± 1	115 ± 19	143 ± 2

^a) SEC measurements were carried out in DMAC + 0.03 wt% of LiBr solution and using polystyrene standards for calibration

^b) DLS measurements were performed at 25 °C with reported values averaged over three measurements

^c) Diameter of the particles by TEM were calculated by FIJI.

We were also interested in expanding the scope of the nanoparticles analyzed through this new platform beyond nanoparticles synthesized in our laboratory, and therefore, also selected to study commercially available and widely used liposomes (Fluoroliposomes, Encapsula NanoSciences). We chose three fluorescently labeled liposomes of around 120 nm, as this size is similar to LNPs currently used in clinics, such as Doxil and Onpattro.^{[ 44 ]} For a more comprehensive study, liposomes with different charges (negative, neutral, and positive) were selected to represent nanoparticles used in a range of nanomedicine applications. The distinct charges of the liposomes arose from the use of three different lipids in their formulations: 1,2‐dioleoyl‐sn‐glycero‐3‐phospho‐rac‐(1‐glycerol) sodium salt (DOPG) for a negative charge, l‐α‐phosphatidylcholine for a neutral charge and 1,2‐dioleoyl‐3‐trimethyl‐ammonium‐propane chloride salt (DOTAP) for a positive charge (Figure 1). The liposome formulation also contained 1,1′‐dioctadecyl‐3,3,3′,3′‐tetramethylindodicarbocyanine 4‐chlorobenzenesulfonate salt (DiD) as a fluorescent dye in all cases. We carefully analyzed the nanoparticles by DLS, Cryo‐TEM, and NanoSight. Data in Table 1 confirmed the particle size of ≈120 nm with varying zeta potentials. The fluorescence intensity of the liposomes was measured by a fluorescence spectrophotometer (Figure S1B, Supporting Information). In short, a library of negatively charged PISA nanoparticles with three different sizes and liposomes with three different charges were synthesized or obtained, followed by extensive characterizations prior to use. All these particles were labeled with fluorescent dyes to facilitate the analysis of their cellular interactions in human blood flow.

2.2. Establishing the Blood Flow Characterization Method

To analyze interactions of nanoparticles with immune cells in human blood under hemodynamic flow conditions, we sought to develop the first method featuring the following all three requirements: (i) fresh whole human blood, (ii) flow conditions, and (iii) a synthetic microvascular network (Figure 2A). In addition, we also considered the availability of our platform to all researchers, thus allowing its widespread use. First, fresh whole human blood was taken from donors and used immediately without processing and storage. Next, we utilized a commercially available microfluidic chip (SynVivo) embedded with micro‐channels possessing the geometry of blood vessels.^{[ 45} , ⁴⁶ , ^{47 ]} The flow in these micro‐channels was designed to mimic various shear rates of blood flow.^{[ 48} , ⁴⁹ , ^{50 ]} However, the commercial microvascular network microchip is not coated with endothelial cells and importantly, to the best of our knowledge, this microfluidic chip has not been previously used to study interactions of nanoparticles with fresh human blood. Therefore, we not only needed to culture endothelial cells inside the microchip prior to use but also developed a method to employ for the first time fresh whole human blood and a synthetic microvascular network to analyze the cellular interactions during flow conditions.

Figure 2.

Blood flow characterization method. A) Top: A photo of the experimental setup. Middle: A synthetic microvascular network; Bottom: Schematic illustration of interactions between nanoparticles and different types of blood cells under flow conditions. B) Images of HUVECs growing in the microchips captured by a Nikon confocal microscopy, cells were stained with Calcein AM (cytoplasm) and Hoechst 33342 (cell nucleus). C) Nuclei of HUVECs were stained with Hoechst 33342 and focal points for imaging selected at the bottom, middle, and top views of the microchannels. Scale bars = 100 µm.

We next cultured a confluent and intact lumen of primary human umbilical vein endothelial cells (HUVECs, Lonza) inside the chip to cover all the microchannels (Figure 2B). Briefly, HUVECs at a high cell density of 2.5 × 10⁷ cells mL⁻¹ were seeded onto the microchannel surface pre‐coated with hydrogel to encourage cell adherence, and the chip then placed in an incubator at 37 °C for 1 h for cell attachment before the fresh medium flowed through the system. It is important to note that high cell density was needed to obtain a confluent and intact lumen of HUVECs and the flow of fresh medium was required for HUVECs to have similar morphology and function to those observed in vivo.^{[ 51 ]} Z‐stack images of cell nuclei captured from the top, middle, and bottom views of the channel further verified the formation of a three‐dimensionally structured lumen of HUVECs under flow conditions (Figure 2C). Finally, a dual syringe pump was used to inject nanoparticles and fresh whole human blood separately (to prevent their interactions prior to entering the microvascular network) into the synthetic microvascular network at 37 °C. From the outlet, nanoparticles and blood were collected into a tube immersed in an ice bath (to prevent further interactions, Figure 2A). It should be highlighted that all materials employed to establish this innovative blood flow method are commercially available, allowing the easy assembly of this setup in different laboratories.

2.3. Analysis of Cellular Interactions

PISA nanoparticles (≈150 nm) and liposomes (both fluorescently labeled and negatively charged) were then used in this blood flow assay. After flowing through the synthetic microvascular network at a flow rate of 5.0 µL min⁻¹, blood samples containing nanoparticles were collected in a fluorescence‐activated cell sorting (FACS) tube immersed in an ice bath. HUVECs were then detached from the microchannel surface by adding trypsin and were collected in the same FACS tube. Without the addition of trypsin, HUVECs remained attached to the microchannel surface. Fluorescently labeled antibodies were then added into the FACS tube before the samples were analyzed by flow cytometry coupled with light scattering and fluorescence detectors (see Experimental Section for more details). Using a gating strategy outlined in Figure 3 , cellular association data of nanoparticles with HUVECs and six WBCs were successfully obtained in a single experiment. In particular, forward scatter (FSC) and side scatter (SSC) plots were used to first identify peripheral blood mononuclear cells (PBMCs), granulocytes, and HUVECs due to their distinct light scattering profile (see the top left panel in Figure 3A). To differentiate B cells, monocytes, NK cells, T cells, and DCs in PBMCs, fluorescent antibodies were employed for further gating, as shown in Figure 3A. The outcome of this gating analysis was to identify the percentage of nanoparticles binding to a specific cell type. The mean fluorescent intensity data was also analyzed and reported in Figures S3 and S5 (Supporting Information). Interactions with six types of cells in the human blood could be characterized in one experiment, although our method could be easily adjusted to characterize the interactions with different cell types, cells that have become activated, or cells undergoing apoptosis, if necessary.^{[ 27 ]}

Figure 3.

Analysis of cellular interactions. A) Gating strategies to identity HUVECs and different WBC subtypes. B) Percentage of white blood cells associated with large PISA nanoparticles (LP) with (w) or without (w/o) the presence of HUVECs under static and flow conditions. C) Percentage of white blood cells associated with anionic liposomes (AL) with or without HUVEC presence under static and flow conditions. Data are presented as mean ± SD, n = 3–4 individual experiments.

For comparison purposes, we also applied this cellular association analysis for nanoparticles incubated with cells under blood flow conditions using the commercial microchip without the HUVECs. The number of nanoparticles and volume of blood injected in all experiments were kept constant at ≈3 × 10¹⁰ particles (measured by NTA) and 200 µL of fresh blood, respectively. Data in Figure 3B showed relatively similar cellular interactions between the PISA nanoparticles of ≈150 nm and all other WBCs when using the microvascular network with or without HUVECs. This finding was also true in flow and static conditions as well as for the ionic lipid nanoparticles (Figure 3C and Figure S2A, Supporting Information). We hypothesize that the presence of HUVECs in static phase does not affect the composition, shear rates, and rheology of human blood as well as the margination of nanoparticles and cells in blood vessels. In addition, the association of nanoparticles tested with HUVECs was very low (1–3% of cells positive, see Figure S2B, Supporting Information) and hence, could not significantly reduce the particle concentration in the blood. This is an important new finding as it significantly simplifies future studies without necessitating the time‐consuming coating of endothelial cells. That said, we envision that the platform using HUVECs‐coated microchips would be still required in several applications, such as cardiovascular nanomedicine in which HUVECs are the targeted cells.^{[ 52} , ⁵³ , ^{54 ]}

2.4. Effect of Flow Conditions

We next studied the effect of blood flow by comparing cellular interactions under flow against static conditions (incubating nanoparticles and blood in wells). Interestingly, we observed a profound difference in cellular association data. The overall trend was that cellular interactions of nanoparticles (for both PISA nanoparticles and liposomes) with WBCs were reduced in flow conditions, although the specific percentage of reduction varied significantly depending on the cell type, as depicted in Figure 3B,C. Specifically, B cells exhibited a high association with nanoparticles in both static and flow conditions. Proteins in human blood were found to form a corona on the surface of nanoparticles, likely contributing to their enhanced association with B cells.^{[ 23 ]} In contrast, cells exhibiting a lower association with nanoparticles (e.g., neutrophils and monocytes) had a marked decrease in cellular interaction under flow conditions (Figure 3B). We speculate that the shear force of blood flow may be strong enough to detach a proportion of nanoparticles possessing weak cellular interactions with neutrophils and monocytes but not sufficient to reduce the stronger associations between nanoparticles and B cells. As the difference in the cellular association between static and flow conditions was strongly dependent on cell type, we hypothesized that it might be also affected by the size and charge of nanoparticles.

2.5. Effect of Particle Size

PISA nanoparticles with different sizes (small ≈40 nm, medium ≈75 nm, and large ≈150 nm) were then subjected to cellular interaction studies under both flow and static conditions. Interestingly, a clear trend of gradually decreased cellular association was observed in Figure 4A. Larger PISA particles exhibited much higher cellular interaction in both flow and static conditions, especially with B cells. The effect of particle size on the cellular association may be related to distinct complemental proteins in the protein corona of small negatively charged nanoparticles, which is important for their nonspecific interaction with B cells.^{[ 23} , ^{55 ]} On the other hand, the association of T cells, NK cells, and DCs requires more specific interactions (e.g., the presence of antigens) and hence, these cells had little to no interactions with PISA nanoparticles (without antigens attached).^{[ 56 ]} When comparing associations of B cells in static and flow conditions, no marked difference was found for the large particles, whereas a significant decrease in cellular interactions was observed for the medium particles. This may be similar to our previous hypothesis that lower affinity was impacted strongly by the blood shear force. Particularly, large nanoparticles associated strongly with B cells and hence demonstrated a much lower decrease in cellular association compared to medium particles having weaker levels of interaction. This result highlights that the effect of flow on the cellular association is dependent on both the particle size and cell type.

Figure 4.

Effect of particle size of PISA nanoparticles. A) Percentage of white blood cells associated with PISA nanoparticles of three different sizes (40 nm—SP, 75 nm—MP, 150 nm—LP). Data are presented as mean ± SD, n = 3–4 individual experiments. Statistical significance of data between static and flow conditions was determined by a paired two‐tailed t‐test; * p < 0.05, ** p < 0.01. One‐way ANOVA with Tukey's pairwise comparisons post‐hoc test was used to assess statistical significance between three‐sized PISA nanoparticles; ^# p < 0.05, ^## p < 0.01, ^### p < 0.001, ^#### p < 0.0001. B) Images of B cells (lymphocytic cells) associated with medium (MP) and large (LP) PISA nanoparticles. C) Images of monocytes (phagocytic cells) associated with large PISA nanoparticles (LP). Orthogonal views of with expression of indicated markers are shown in the right and bottom images. Scale bars = 2 µm.

We next employed confocal microscopy to visualize the cellular interaction of nanoparticles with relatively high cellular associations. We studied B cells and monocytes as representative lymphocytic and phagocytic cells, respectively. Cells were isolated by using negative magnetic selections kits. Cell membranes and nuclei were labeled with different dyes before imaging. Images in orthogonal views of B cells and monocytes that associate with PISA nanoparticles are shown in Figure 4B and Figure S4 (Supporting Information). The images showed that PISA nanoparticles tend to primarily attach to the surface of B cells. For monocytes, it was more likely that all three PISA nanoparticles were internalized, which is consistent with the nature of phagocytic cells. That said, a specific hybridization internalization probe is needed to confirm the cellular uptake of these particles, and this will be the focus of a forthcoming study.^{[ 57 ]} In short, we demonstrated that the size of PISA nanoparticles has a significant effect on cellular interactions with WBCs, especially with B cells. For cancer drug delivery and imaging applications in which cellular association with WBCs should be minimized, small PISA nanoparticles seem to be promising candidates due to their low cellular interaction with WBCs. On the other hand, large PISA nanoparticles may find an application in the nonspecific targeting to B cells, which may be useful in immunomodulation studies.

2.6. Effect of Particle Charge

Motivated by the interesting results presented above, we continued to study the impact of particle charge on cellular interactions under human blood flow. Three fluorescently labeled liposomes with different charges (negative, neutral, and positive) were used, and the results are presented in Figure 5 and Figures S5 and S6 (Supporting Information). We found that all liposomes exhibited stronger interactions with B cells than other WBCs tested. This result suggests that softer nanoparticles have significantly reduced cellular uptake in immune cells, in line with previous reports.^{[ 58 ]} In particular, the hard and glassy nature of the polystyrene solid core of the PISA nanoparticles may be responsible for limiting the movement of the core.^{[ 59} , ⁶⁰ , ⁶¹ , ^{62 ]} Lipid nanoparticles composed of low molecular weight lipid and a liquid core are softer than polystyrene nanoparticles and hence have lower cellular interactions as observed in Figure 5. It is noted that liposomes still interacted with B cells depending on their charge. Similar to the case of PISA nanoparticles, B cells seem to associate with nanoparticles more than other WBCs due to their sensitivity to particle protein corona.^{[ 23 ]} Despite opposing charges, anionic liposome (AL) and cationic liposome (CL) surprisingly demonstrated higher association with B cells than neutral liposome (NL) (Figure 5A). Although the zeta potentials of the cationic and neutral liposomes show little difference, the charge effect was still statistically significant. This can be explained by the similar complemental proteins present in the protein corona of anionic and cationic liposomes, which likely dictates the interactions of B cells with liposomes and PISA nanoparticles.^{[ 23} , ^{63 ]} Overall, our data suggest that flow conditions can affect particle–cell associations depending on particle charge. The decrease in the cellular association with anionic liposomes was statistically significant (p < 0.05). Other liposomes displayed no clear trend in reduced cellular interactions due to either lower or higher association with B cells. Besides, a proportion of cationic liposomes were likely to be internalized into B cells and hence, may be less influenced by blood shear force than anionic liposomes attaching to B cell surface (Figure 5B). Similar to PISA nanoparticles, a specific hybridization internalization probe is required to further confirm the internalization as confocal microscopy images are not a quantitative technique.^{[ 57 ]} In short, these results suggest that ionic liposomes may find further application as vaccine delivery vehicles, while neutral liposomes may be suitable for cancer drug delivery.

Figure 5.

Effects of particle charge of liposomes. A) Percentage of white blood cells associated with liposomes of three different charges (anionic liposomes—AL, cationic liposomes—CL, neutral liposomes—NL). Data are presented as mean ± SD, n = 3–4 individual experiments. Statistical significance of data between static and flow conditions was determined by a paired two‐tailed t‐test; * p < 0.05. One‐way ANOVA with Tukey's pairwise comparisons post‐hoc test was used to assess statistical significance between three‐charged liposomes; ^# p < 0.05, ^## p < 0.01. B) Images of B cells (lymphocytic cells) associated with three‐charged liposomes. Orthogonal views of these cells with expression of indicated markers are shown in the right and bottom images. Scale bars = 2 µm.

3. Conclusion

We report an innovative methodology to realize the complex and challenging characterization of cellular interactions of nanoparticles under human blood flow conditions. In this method, fresh human blood and a synthetic microvascular network were employed for the first time to overcome the challenge of mimicking diverse components of human blood and the hemodynamic flow in blood vessels. Using this method, we shed new light on the complex interplay of size and charge of nanoparticles, as well as blood flow and type of blood cells, on the interactions between nanoparticles and immune cells in human blood. We show that large PISA nanoparticles exhibited higher associations to B cells, monocytes, and neutrophils than their smaller counterparts and liposomes. This also represents the first study of widely used PISA nanoparticles and clinically relevant liposomes during human blood flow in a microvascular network. Interestingly, cationic and anionic liposomes, despite having opposite charges, had relatively similar interactions with B cells but significantly higher associations than their neutral analogues. We also discovered that the blood flow conditions under the parameters examined affected weak associations more than strong interactions. Importantly, this work suggests that neutral liposomes (≈120 nm) and small anionic PISA nanoparticles (≈40 nm) would be suitable candidates for cancer drug delivery due to their minimal interactions with immune cells, while charged liposomes and large PISA nanoparticles (≈150 nm) may be useful for vaccine delivery and as immunomodulators targeting B cells. Furthermore, this much‐needed platform is easy to set up in any lab and can be applied to future studies of any type of nanoparticles with different sizes and charges, paving the way for further advances in the fundamental understanding of the nano–bio interface and the field of nanomedicine.

4. Experimental Section

Materials

Chemicals and solvents used for the synthesis of polymeric nanoparticles, including 4‐cyano‐4‐(ethylthiocarbonothioylthio) pentanoic acid (ECT), N‐hydroxyethyl acrylamide (HEAA, 97%), poly(ethylene glycol) methyl ether acrylate (PEGA, average M _n = 480), and dimethyl sulfoxide (>99.9%, anhydrous) were purchased from Sigma‐Aldrich. Cyanine 5 maleimide (Cy5) was purchased from Lumiprobe. Three different‐charged Fluoroliposomes (i.e., PG‐based negatively charged, DOTAP‐based positively charged, and neutral liposomes), which were labeled with 1,1′‐dioctadecyl‐3,3,3′,3′‐tetramethylindodicarbocyanine, 4‐chlorobenzenesulfonate salt (DiD) were purchased from Encapsula NanoSciences. All antibodies for WBC and HUVEC staining were bought from BD Biosciences, except for CD31 (BioLegend). Alexa Fluor 488 Wheat Germ Agglutinin (AF488 WGA), Calcein AM, and Hoechst 33342 were provided by ThermoFisher Scientific. Human Umbilical Vein Endothelial Cells (HUVECs) with Endothelial Growth Media (EGM) were purchased from Lonza. Synthetic vascular networks (SMN1‐D001) were obtained from SynVivo.

Synthesis of PISA Nanoparticles

Core–shell polymeric nanoparticles of three different sizes (≈40, ≈75, and ≈150 nm) were synthesized using RAFT‐PISA technique, as described previously.^{[ 13 ]} Briefly, the shell of the nanoparticles, containing poly(ethylene glycol) and carboxylic acid end groups were formed by RAFT solution polymerization. ECT was polymerized with HEAA and PEGA in DMSO for 4 h at 70 °C to obtain the macromolecular chain transfer agent (macro‐CTA). The macro‐CTA was then undergone chain extension reactions with styrene in water at 80 °C using RAFT‐mediated emulsion polymerization to form diblock copolymers. During the polymerization, an aliquot of Cy5 in styrene was added into the reaction to allow incorporation of Cy5 into the styrene core of the nanoparticles.

Nanoparticle Characterizations

Dynamic light scattering (DLS): Hydrodynamic diameter and zeta potential of nanoparticles (50 µg mL⁻¹ in PBS) were measured by DLS using a Malvern Zetasizer Nano Series with a 4 mW He–Ne ion laser (λ = 633 nm). Data measurement and analysis were performed at an angle of 173° and a temperature of 25 °C. Transmission electron microscopy (TEM): TEM images were acquired using a Tecnai F20 instrument at an operation voltage of 200 kV. Polymeric nanoparticles samples were prepared by depositing particle solutions (0.1 wt% in MilliQ water) on a Formvar‐coated copper grid (GSCu400F‐50, Proscitech), then allowing them to air‐dry overnight. NTA: The numbers of nanoparticles in 1 mL were calculated by NTA using a NanoSight NS300 (Malvern). The measurements were carried out with a nanoparticle solution of 0.1 µg mL⁻¹ in MilliQ water at room temperature with manual shutter and gain adjustments. Fluorescence measurement: Fluorescence spectra of nanoparticle solutions with an equivalent number of nanoparticles were obtained using a fluorescence spectrophotometer (Shimadzu RF‐501 PC) at an excitation wavelength of 635 nm and slit widths of 10 nm for both excitation and emission of Cy5. Cryogenic transmission electron microscopy (Cryo‐TEM) was used to examine the morphology of liposome samples. Briefly, 3.5 µL of 1 × 10⁻³ m liposome solutions in PBS were applied onto copper grids (200‐mesh) coated with holey carbon film (Quantifoil R1.2/1.3) and pre‐glow discharged in a Pelco glow discharge unit. The samples were then blotted for 3 s at a blot force of −6 in a Vitrobot plunge freezer system (FEI) to form thin sample film onto the grids, which were next vitrified by being plunged into liquid ethane at 5 °C. The vitrified samples were transferred to a Gatan 626 cryo‐holder and imaged by using a Tecnai 12 TEM at an operating voltage of 120 kV and temperatures from −170 to −175 °C. Images were captured by a Gatan Eagle high‐resolution CCD camera (4k × 4k), using the Tecnai Image Acquisition (TIA) program.

Blood Acquisition

Blood was collected from healthy human donors after obtaining informed consent in accordance with the University of Melbourne Human ethics approval 1443420 and the Australian National Health and Medical Research Council Statement on Ethical Conduct in Human Research. Fresh blood was drawn by venepuncture into sodium heparin Vacuette collection tubes (Greiner BioOne) and inverted gently at least 5 times before use.

Associations of Nanoparticles and White Blood Cells under Static Conditions

Freshly drawn blood (200 µL) was added into a 5 mL polystyrene round‐bottom tubes (“FACS” tubes, Falcon), then incubated at 37 °C for 10 min. Under static conditions, 200 µL of polymeric nanoparticles and liposomes (≈3 × 10¹⁰ particles) in PBS were added into the blood, vortexed and incubated at 37 °C for 2 min. After that, the tubes were embedded into ice for 10 min. RBCs were then lysed with Pharm Lyse buffer (BD Biosciences) at 30 × blood volume for 20 min. WBCs were spun down at 500 g for 7 min and washed twice with PBS. The cells were surface stained at 4 °C for 30 min with appropriate concentrations of antibodies against CD66b (BV421, G10F5), CD14 (APC‐H7, MϕP9), CD19 (BUV395, SJ25C1), CD3 (AF700, SP34‐2), CD56 (PE, B159), Lineage‐1 cocktail (FITC), and HLA‐DR (PerCP‐CF594, G46‐6). The cells were then washed twice with cold FACS wash buffer (0.5% bovine serum albumin (Sigma‐Aldrich) and 2 × 10⁻³ m EDTA (Ambion) in 1 × PBS) and fixed with 4% formaldehyde (Sigma‐Aldrich). Cellular associations of the particles were analyzed by flow cytometry, using a LSRFortessa (BD Biosciences). Data were analyzed by FlowJo v10.

Associations of Nanoparticles and White Blood Cells under Flow Conditions

Synthetic microvascular networks (SMN) embedded into a microfluidic device (SMN1‐D001) were used to assess nanoparticle associations with human WBCs under flow conditions. Human blood and nanoparticles were injected into the system via two different inlet ports by using 1 mL syringes connected with Tygon tubing (0.02″ID × 0.06″OD). Freshly drawn blood and nanoparticles (3 × 10¹⁰) at an equal volume of 200 µL were contemporaneously perfused through the chip at a flow rate of 5.0 µL min⁻¹ using a syringe pump (Harvard Apparatus, Holliston, MA). The nanoparticles and blood cells were allowed to interact inside the chip at 37 °C for 2 min before collecting into a FACS tube, which was placed on ice during the experiments. After that, RBCs were lysed, and WBCs were phenotyped and analyzed as described for static condition above.

Associations of Nanoparticles and White Blood Cells with Presence of Endothelial Cells

Under static conditions: HUVECs were seeded at 100 000 cells per well in 48‐well plates, then incubated overnight at 37 °C and 5% CO₂. Freshly drawn blood (200 µL) were added onto the HUVEC layer and incubated for 10 min before 200 µL of polymeric nanoparticles or liposomes (3 × 10¹⁰) in PBS were added directly into the blood, mixed and incubated in the blood at 37 °C for 2 min. At the end of the experiments, blood cells were transferred to a cold (4 °C) FACS tube while HUVECs were trypsinized and then collected into the same FACS tube. The tube was then placed on ice for 10 min. Under flow conditions: HUVECs were cultured into the synthetic microvascular networks as described previously.^{[ 13 ]} Briefly, HUVECs at 2.5 × 10⁷ cell mL⁻¹ were gently injected into the microfluidic channels, that previously coated with Matrigel (Invitro Technologies), until reaching about 90% confluence was reached. The device was then placed in an incubator (37 °C and 5% CO₂) for 1 h to allow cell attachment before EGM flowed through the chip at 0.2 µL min⁻¹ overnight using a syringe pump. An equal volume (200 µL) of human blood and nanoparticles were contemporaneously perfused through the chip with the presence of HUVECs on the vessel walls at a rate of 5.0 µL min⁻¹. After blood was collected into a FACS tube on ice at the outlet port of the chip, HUVECs were trypsinized and collected into the same tube. RBCs were then lysed, and WBCs were stained as described previously. HUVECs were phenotyped along with WBC staining with appropriate concentrations of antibodies against CD31 (BV605, WM59) and CD146 (BV711, P1H12). The cells were washed, fixed, and analyzed by flow cytometry as aforementioned.

Confocal Microscopy

The nanoparticle–blood interaction studies were carried out similar to the association assays under static conditions without the presence of HUVECs; however, the nanoparticle concentrations were increased 10 times and incubation time was prolonged to 1 h to allow particle visualization. After incubation, peripheral blood mononuclear cells (PBMCs) were purified by density gradient centrifugation by using Ficoll‐Paque PLUS (GE Healthcare), then washed twice with PBS. B cells and monocytes were then isolated from the PBMCs by negative magnetic cell selection, using EasySep Human B Cell Isolation Kit (STEMCELL Technologies) and Pan Monocyte Isolation Kit (MACS Miltenyi Biotec), respectively. The cells were stained with 5.0 µg mL⁻¹ AF488 WGA and 1.0 µg mL⁻¹ Hoechst 33342 for 15 min at RT and then fixed with 4% formaldehyde in PBS. The cells were placed into a μ‐Slide 8‐well chamber (Ibidi) and imaged on a S.P.8 confocal microscope (Leica Microsystems), using a 63 × oil 1.4 NA objective with excitation at 405, 488, and 633 nm and emission detectors set as follows: 415–485, 495–535, and 645–705 PMT. Images were acquired using LAS AF software v3.2 (Leica Microsystems) and analyzed using FIJI.

For HUVEC imaging in the chip, after reaching confluence, the cells were washed with 1 mL of PBS and stained with 5.0 µg mL⁻¹ Calcein AM and 1.0 µg mL⁻¹ Hoechst 33342 for 15 min at RT. Images of the cells were acquired on a Nikon Eclipse TiE inverted microscope with an A1R Resonant Scanning Confocal system, using a 10 × 0.30 NA Plan Fluor objective with excitation/emission of 405/450 ± 25, 488/525 ± 25, and 640/663 ± 38 for blue, green, and far‐red channels, respectively. All images were captured at 1024 × 1024 pixel resolution using a pixel dwell time of 2.4 µs per pixel.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information.

Vu M. N., Kelly H. G., Wheatley A. K., Peng S., Pilkington E. H., Veldhuis N. A., Davis T. P., Kent S. J., Truong N. P., Cellular Interactions of Liposomes and PISA Nanoparticles during Human Blood Flow in a Microvascular Network. Small 2020, 16, 2002861. 10.1002/smll.202002861