Microengineered filters for efficient delivery of nanomaterials into mammalian cells

Microfabrication process

The SiN (Si₃N₄) membranes were fabricated using standard 100 mm, 480 μm thick double-sided polished silicon wafers (Fredricksburg, VA, USA). The fabrication process of the SiN microsieves has been performed in collaboration with Aquamarijn company (ZT, The Netherlands); details are shown in Fig. 1A,B. First, the Si substrate was deposited and pre-coated on both sides of wafers through low-pressure chemical vapor deposition at 850 °C. In the next step, the silicon wafers were dipped in potassium hydroxide (KOH) solution at 80 °C to etch away bulk silicon until the desired thickness for the SiN membrane was achieved. The surface of these wafers was patterned with Shipley’s positive photoresist (AZ P4330, Clariant Corp, USA) through the standard photolithography process. Next, an array of microholes with a circular pattern and size of 5 and 8 μm was etched onto the top SiN layer using reactive ion etching. This step is followed by dicing and etching the SiN wafers again in the KOH solution to remove the residuals and free the membrane. By etching away the residuals, once the membrane is freed, the SiN micropores were observed on the Si substrate. The exposed bottom side of the SiN layer was then removed by 4 h etching in the buffered oxide etch solution. Next, to enhance the strength of SiN microsieves as well as minimize the cell adhesion²⁷, a thin layer of parylene was conformally deposited over these wafers.

Schematic representation of the MFP delivery system. (A) In this platform, as the cells are forced through the microporous filters, they experience mechanical membrane deformations resulting in the generation of multiple discontinuities in the plasma membrane. These transient pores allow passive diffusion of exogenous nanomaterials across the plasma membrane. (B) Fabrication process of the SiN microsieves. First, SiN is coated on both sides of silicon (Si) wafers, followed by etching away the bulk Si using the KOH solution. Upon achieving the desired thickness, the surface of the wafers is patterned and then conformally coated with parylene to enhance the strength of the microsieves and reduce the cells attachment and prevent clogging issues. Figure (A) was generated using Biorender software (https://biorender.com/).

MFP system setup

Isopore track-etched PC membranes with 5 and 8 μm pore size and 13 mm polypropylene Swinnex filter holders were purchased from the Merck Millipore (MA, USA) and Sterlitech (WA, USA), respectively. Before the dead-end MFP, either PC or SiN membranes were mounted between two pieces of a 13 mm Swinnex filter holder, in which the top piece was connected directly to a syringe. To test the stability of the delivery setup and prime the microporous filters, 70% ethanol and sterilized DI water followed by PBS was flushed through a 3 ml syringe mounted on the syringe pump at the flow rate of 1 ml min⁻¹. Next, the cell suspension was pumped through the microporous filter membranes by a syringe pump (KDS Legato, Sigma Aldrich Inc, MO, USA), and treated cells were collected in an Eppendorf tube for further assays.

Fluid flow simulation

Additionally, to better understand the flow characteristics and fluid behavior passing through the micropores of SiN membranes, a computational fluid dynamic simulation was carried out using ANSYS Fluent 2019R3 (ANSYS Inc, Pennsylvania, USA). For simplification and lowering the computational load, only a portion of the actual MFP setup was modeled. The simulation was performed using a Laminar flow setting with a standard mesh and inlet velocity of 0.24 m s⁻¹ at the inlet corresponding to the operating flow rate of the system. Furthermore, domain size and mesh independence checks were conducted to ensure the stability of the simulation results.

Cell culture and viability assay

HeLa (cervical cancer cells, adherent) and THP1 (human acute monocytic leukemia cells, non-adherent) cell lines were purchased from ATCC (VA, USA) and cultured in DMEM and RPMI growth media, respectively, at 37 °C and 5% CO₂. The growth media was supplemented with 10% FBS (Life Technologies Inc, CA, USA) and 1% penicillin–streptomycin (Thermo Fisher Scientific Inc, MA, USA). Prior to the delivery procedure, the viability of cells was characterized using the trypan blue exclusion assay (Lonza, MD, USA). In this assay, the cells were resuspended in PBS, and one part of the cell suspension was added to one part of 0.4% trypan blue dye. Next, a drop of the final solution was added to a hemocytometer, and the total number of unstained (viable) and stained (nonviable) cells, as well as the total number of cells, were counted. Then the viability of cells was calculated as:

Delivery procedure using filter membranes

In the MFP platform, after centrifugation of cell suspension for 3 min at 500 g, the supernatant was discharged. As a proof of concept, a final concentration ranging from 5 × 10⁶– 1 × 10⁷ cells ml⁻¹ was prepared and added to the delivery buffer and passed through the 40 μm cell strainers (Corning Inc, NY, USA) to remove cell clumps from the suspension. Accordingly, DPBS, DMEM, and optiMEM were used as a delivery buffer to optimize the experimental conditions. Before loading the cell suspension into the syringes, the delivery biomolecules were added to the solution. Next, the cell suspension was forced through the microengineered porous membranes using a syringe pump (Chemyx Inc, TX, USA) at a flow rate range between 0.5 to 10 ml min⁻¹. First, we used Fluorescein isothiocyanate (FITC) dextran molecules (70 and 2000 kDa) (Sigma Aldrich, MO, USA) as delivery cargoes. Next, to prove the ability of this platform in loading large-sized cargoes, histone H2B-GFP (green fluorescent protein) plasmid DNA (Addgene plasmid # 11,680; RRID:Addgene_11680, MA, USA), kindly gifted by Geoff Wahl, was used as a delivery biomolecule to be loaded into the cells of interest²⁸. Triplicate sampling was conducted for each data point.

Cell membrane injury analysis

Next, the effect of MFP treatment on plasma membrane damage was carried out in the culture supernatant 24 h post-delivery. Accordingly, lactate dehydrogenase (LDH) assay was performed for both control and microfiltroporated samples using the pierce LDH cytotoxicity assay kit (Thermo Fisher Scientific Inc, MA, USA). The LDH activity was then calculated based on the absorption of culture supernatant at 490 and 680 nm according to the manufacturer’s recommendation using the Varioskan Lux Reader (Thermo Fisher Scientific Inc, MA, USA).

Quantitative polymerase chain reaction (qPCR) of cellular damage indicators

Immediately after the MFP treatment, total RNA was isolated from the treated and untreated cells using the PureLink ™ RNA Mini kit (Thermo Fisher Scientific Inc, MA, USA). The concentration and OD 260/280 ratio of the RNA samples were then measured using the NanoDrop ™ one/one microvolume UV–Vis spectrophotometer (Thermo Fisher Scientific Inc, MA, USA). Next, 1.0 µg of the RNA samples was used to evaluate the expression level of the DNA damage-inducible transcript 3 (DDIT3) and mitogen-activated protein kinase 14 (MAPK14) genes as indicators of cell injury and environmental stress using SuperScript ™ III platinum SYBR green one-step qPCR kit (Thermo Fisher Scientific Inc, MA, USA).

Flow cytometry and cell recovery assessment

After MFP, the cells were collected in an Eppendorf tube, centrifuged at 500 g for 3 min, and resuspended in flow cytometry buffer, which contains PBS (Thermo Fisher Scientific Inc, MA, USA) supplied with 3% FBS and 1% Pluronic F-68 (Sigma Aldrich Inc, MO, USA). Quantitative analysis of biomolecule delivery to the cell interior coined as delivery efficiency was performed using the BD FACS LSR Fortessa flow cytometer, and data analysis was carried out using BD FACSDiva™ software (Becton–Dickinson, MA, USA). To assess the delivery efficiency quantitatively, we measured the mean fluorescent intensity of individual cells using flow cytometry. To this aim, we defined a threshold of 5% as a baseline to exclude endocytosis control cells, autofluorescence, and undesired binding of 70 kDa FITC-dextran to the cell surface⁹. Then, we defined the delivery efficiency as a fraction of live cells with fluorescent signals above this threshold. The viability of cells post-treatment was characterized by Sytox blue dead cell staining (Thermofisher Ins, MA, USA) and FITC Annexin V (Biolegned, CA, USA) to exclude dead and apoptotic cells from the analysis, respectively, according to the manufacturer’s protocol. Recovery of cells post MFP process was evaluated by counting the number of viable cells in the treated and untreated sample with the same number of cells seeded in the same volume via this equation:

Microscopic assessment

The microfiltroporated cells were seeded into a 6-well plate containing the complete media for the fluorescent and confocal microscopy analysis using the Nikon A1R inverted microscope 24 h post-treatment. FITC-labeled phalloidin (Sigma Aldrich Inc, MO, USA) and Hoechst 33,342 (Thermofisher Ins, MA, USA) stains were used according to the manufacturer’s protocol to visualize cytoskeletal actin and nucleus, respectively. For scanning electron microscopy (SEM), the treated cells were further processed to visualize the pores on the cell membrane compared to the untreated ones using the Zeiss Supr 55VP scanning electron microscope.

Comparison of MFP delivery with the mainstream options

To evaluate the delivery performance of MFP using SiN microsieves, the loading efficiency and cell viability were compared with those of MFP via PC membranes, and conventional electroporation and lipofection (as representatives of benchtop methods). MFP using the PC membranes was performed in the same conditions (i.e., cell density and delivery buffer), and electroporation and lipofection were carried out using the Neon Transfection System (Thermo Fisher Scientific, MA, USA) and Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, MA, USA), respectively, according to the protocols recommended by the manufacturer.

Statistical analysis

The statistical analysis was conducted using the unpaired student T-test and one-way analysis of variance (ANOVA) method, followed by Tukey’s multiple comparison test to evaluate whether a significant difference exists between the tested groups using GraphPad Prism software 6. P-values less than or equal to 0.05 were considered statistically significant, which are displayed as *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001, and ****P-value < 0.0001.