**1. Introduction**

Drug targeting has the potential to improve the therapeutic efficacy and mitigate the non-specific effects of many drugs. In the last years, several types of drug delivery vehicles have been developed, including monoclonal antibodies [1], peptides [2], proteins [3], lipoproteins [4], carbohydrates [5], and polymeric nanoparticles [6,7]. Compared with the number of studies in which nanoparticles (NPs) are used [8–12], only a small number of studies involve the use of microparticles (μPs) [13–15]. It has been reported that small sizes and positive charges favor NP intake by cells [14–16], but in some cases, larger particle sizes could be advantageous for preventing non-specific interactions

and internalization into normal non-phagocytic cells or for optimal tissue entrapment and transient retention [17]. Moreover, to target cancer cells, NPs and μPs surfaces can be modified to increase the interaction with plasma membrane-specific markers like the transferrin receptor [18], the folate receptor [19], or the human epidermal growth factor receptor 2 (HER2, also known as ERBB2) [20,21]. HER2 is a receptor tyrosine kinase which is overexpressed by some types of cancer cells and is considered a marker of poor clinical outcome in breast and ovarian cancer [22,23]. Some treatments directed to this target have already been approved and are clinically used, such as the anti-HER2 monoclonal antibody trastuzumab, alone or in combination with emtansine (T-DM1) [24], and the HER2 tyrosine kinase activity inhibitor lapatinib.

Traditionally, in vitro studies on drug carriers and drug release have been performed in static monolayer cell cultures. However, studies in microfluidic environments, mimicking the circulatory system, are currently gaining interest [12]. Compared with static cultures, microfluidic studies allow for better predictions about how a drug or a drug carrier running in a circulating flow, will interact with cells [25–27]. On the other hand, because normal and tumoral cells are intermingled in vivo, cocultures of normal and tumoral cells can better simulate tissue conditions than monocultures [28,29].

The overall objective of the present study was to evaluate the efficiency of targeting μPs to cells in physiological-like conditions (fluidic culture conditions). Two isogenic breast epithelial cell lines were used, one normal (D492) and the other overexpressing HER2 (D492HER2). Moreover, polystyrene μPs of 1 μm in diameter were biofunctionalized with a specific targeting protein, an anti-HER2 antibody, or with a non-specific secondary antibody. The specific objectives of the study were to evaluate and compare the cell internalization of these μPs in different culture conditions: monoculture versus coculture conditions, and static versus fluidic culture conditions.

### **2. Material and Methods**

### *2.1. Biofunctionalization of Polystyrene μPs*

Carboxylate polystyrene μPs of 1 μm in diameter (Polybead® Carboxylate Microspheres. Polysciences, Inc., Warrington, PA, USA) were biofunctionalized with two different targeting molecules: (1) mouse anti-c-ERBB2/c-Neu (Ab-5) clone TA-1 (Millipore, Darmstadt, Germany), herein referred to as antiH, and (2) goa<sup>t</sup> anti-mouse IgG2a secondary antibody Alexa Fluor® 647 conjugate (Life Technologies, Carlsbad, CA, USA), herein referred to as secAb. Biofunctionalization was carried out using the PolyLink Protein Coupling Kit for COOH Microspheres (Polysciences) according to the manufacturer's instructions. The size of the μPs before and after biofunctionalization was analyzed by transmission electronic microscopy (TEM) (JEOL, JEM 2011). Biofunctionalization was evaluated under a fluorescence inverted microscope (Olympus IX71, Olympus, Hamburg, Germany) and by the change in the ζ-potential.

Microscopically, biofunctionalization of μPs with secAb (μP-secAb) was evaluated directly on the basis of their far-red fluorescence emission. On the other hand, μPs biofunctionalized with antiH (μP-antiH) were incubated for 5 min with chicken anti-mouse IgG (H+L) secondary antibody Alexa Fluor® 488 conjugate (1:500. Life Technologies) before the evaluation of green fluorescence emission.

Biofunctionalized and non-biofunctionalized μPs were separately resuspended in H14 culture medium [30] and sonicated for 5 min (Fisherbrand FB15047, Fisher Scientific, Germany) to achieve a monodispersed sample. Their ζ-potential was then measured with a Zetasizer Nano ZS (Malvern Instruments, Malvern, UK).

## *2.2. Cell Lines*

Two isogenic breast epithelial cell lines, D492 and D492HER2, were used in the study. D492 is a non-tumorigenic cell line with stem cell properties that expresses low levels of the HER2 oncogene [30,31]. D492HER2 was generated by overexpressing the HER2 oncogene in D492 and is highly tumorigenic [32]. Both cell lines constitutively express green fluorescent protein (GFP).

The cells were cultured in serum-free H14 culture medium [30,31] at 37 ◦C and 5% CO2 (standard conditions). As explained below, cell culture was performed as follows: mono- or cocultures in static conditions, and cocultures in fluidic conditions.

### *2.3. Cell Cultures in Static Conditions*

For monoculture experiments, cells were seeded at a density of 60,000 cells/well in 24-well plates (μ-Plate 24-Well ibiTreat: #1.5 polymer coverslip. ibidi, Martinsried, Germany). For coculture experiments, 30,000 cells of each cell line (D492 and D492HER2) were seeded together in each well. In both cases, the cells were maintained for 24 h in standard culture conditions prior to performing any experiments.

To analyze μP internalization, μPs (μP-antiH or μP-secAb) were sonicated for 5 min, diluted (1:100), counted with a hemocytometer, and then added at a proportion of 45 μP/cell to the cell cultures and incubated for further 24 h in standard culture conditions.

### *2.4. Cell Cultures in Fluidic Conditions*

For these experiments, only cocultures were performed. Before seeding, channel slides (μ-Slides I 0.8 mm ibiTreat, ibidi) were coated with bovine collagen type I (Advanced Biomatrix, San Diego, CA, USA) to enhance cell adhesion. Then, 1.5 × 10<sup>5</sup> cells of each cell line (D492 and D492HER2) were seeded in H14 medium containing 2% penicillin/streptomycin (Biowest, Nuaillé, France) and incubated in standard culture conditions. After 24 h, the slides were connected to a microfluidic system consisting of a perfusion set (Perfusion Set Red ID 1.6 mm, ibidi) filled with fresh H14 medium containing the μPs (μP-secAb or μP-antiH) and a Fluidic Unit connected to an ibidi Pump, controlled by a pump-control software (ibidi). The microfluidic system was kept at 37 ◦C and 5% CO2. The cultures were maintained for 24 h under a unidirectional flow rate fixed at 4.32 mL/min with a shear stress of 1.50 dyn/cm<sup>2</sup> and a pressure of 7.9 mbar, as recommended by the manufacturer.

### *2.5. Evaluation of Microparticles Internalization*

After being cultured in either static or fluidic conditions, the cells were washed three times with phosphate buffer saline (PBS) at room temperature (RT), fixed for 15 min with 4% paraformaldehyde (Sigma-Aldrich, St Louis, MO, USA) in PBS, and washed again with PBS (three times). Next, the fixed cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich) in PBS for 10 min at RT, washed with PBS (three times), and blocked with 3% bovine serum albumin (BSA) (Sigma-Aldrich) in PBS for 40 min. PBS with 3% BSA was also employed to dilute the antibodies used in this work.

The cells from the monocultures were incubated with Alexa Fluor® 546 Phalloidin (1:40, Life Technologies) to label actin microfilaments and, in the case of samples with μP-antiH, also with goa<sup>t</sup> anti-mouse IgG1 secondary antibody Alexa Fluor® 647 conjugate (1:150. Life Technologies) for 1 h at RT to detect μP-antiH.

To distinguish between D492 and D492HER2 in cocultures, the cells were first incubated with rabbit anti-HER2 monoclonal antibody (1:200, Cell Signaling, Danvers, MA, USA) overnight at 4 ◦C. Then, the samples were washed three times with PBS and incubated for 2.5 h at RT with Alexa Fluor® 546 Phalloidin to label actin microfilaments to visualize the cell limit and chicken anti-rabbit IgG (H+L) Alexa Fluor® 405 conjugate secondary antibody (1:150, Life Technologies) to label HER2 in the plasma membrane; for cells incubated with μP-antiH, goa<sup>t</sup> anti-mouse IgG1 secondary antibody Alexa Fluor® 647 conjugate was also used.

Finally, the cells were washed three times with PBS and maintained at 4 ◦C in PBS until evaluation under a confocal laser scanning microscope (CLSM. Olympus, Tokyo, Japan). Orthogonal projections of z-stacks of at least 100 cells for each cell line were evaluated in each replicate. The xyz sequentially acquired images allowed for assessment of whether the particles were inside the cells or attached to their surfaces.
