A designer cell culture insert with a nanofibrous membrane toward engineering an epithelial tissue model validated by cellular nanomechanics

Engineered platforms for culturing cells of the skin and other epithelial tissues are useful for the regeneration and development of in vitro tissue models used in drug screening. Recapitulating the biomechanical behavior of the cells is one of the important hallmarks of successful tissue generation on these platforms. The biomechanical behavior of cells profoundly affects the physiological functions of the generated tissue. In this work, a designer nanofibrous cell culture insert (NCCI) device was developed, consisting of a free-hanging polymeric nanofibrous membrane. The free-hanging nanofibrous membrane has a well-tailored architecture, stiffness, and topography to better mimic the extracellular matrix of any soft tissue than conventional, flat tissue culture polystyrene (TCPS) surfaces. Human keratinocytes (HaCaT cells) cultured on the designer NCCIs exhibited a 3D tissue-like phenotype compared to the cells cultured on TCPS. Furthermore, the biomechanical characterization by bio-atomic force microscopy (Bio-AFM) revealed a markedly altered cellular morphology and stiffness of the cellular cytoplasm, nucleus, and cell–cell junctions. The nuclear and cytoplasmic moduli were reduced, while the stiffness of the cellular junctions was enhanced on the NCCI compared to cells on TCPS, which are indicative of the fluidic state and migratory phenotype on the NCCI. These observations were corroborated by immunostaining, which revealed enhanced cell–cell contact along with a higher expression of junction proteins and enhanced migration in a wound-healing assay. Taken together, these results underscore the role of the novel designer NCCI device as an in vitro platform for epithelial cells with several potential applications, including drug testing, disease modeling, and tissue regeneration.


Design and Fabrication of NCCI
We designed and fabricated nanofiber-based cell culture inserts by combining two scalable micro-/nano-technologies, electrospinning and 3D printing. The structural parts of cell culture inserts were fabricated by a 3D printer consisting of two components prepared from poly(lactic acid) (PLA). The diameter and the thickness of the base component (disc) are 64 mm and 0.6 mm, respectively. There are four legs with 0.5 mm in height and 1 mm in diameter connected to the base component. They serve as supports to lift the inserts above the base of the culture plates. The disc has holes of 4.5mm in diameter (Figure S1B). These structural parts of inserts can be fabricated into any dimension, depending on the resolution of the 3D printer. Thereafter, the supports of NCCI were dipped in electrolyte (0.01M KCl) solution and placed over a Whatman filter paper soiled with the electrolyte, as shown in Figure S1A. This arrangement created uniform conducting regions on the NCCI during the electrospinning process. Consequently, fibers were uniformly deposited on the eye part of the base component of the NCCI, as shown in Figure S1D, unlike the setup where nanofibers Electronic Supplementary Material (ESI) for Nanoscale Advances. This journal is © The Royal Society of Chemistry 2021 S2 were deposited without the use of electrolyte ( Figure S1C). The micropore size distribution of the conventional commercial inserts ( Figure S1E) was observed to be nearly uniform, unlike for NCCI. It was observed that nanofibers were radially aligned due to uneven stretching of nanofibers resulting from the differential spatial variation of the electric field strength near the edge of an eye of the NCCI when electrospinning was performed without the electrolyte. However, the use of the electrolyte minimizes the spatial variation in the electric field over the entire region of the eye. Thus, the uniform increase in the conductivity of the PLA substrate with the use of electrolyte results in a uniform fiber deposition, which is evident from the SEM images 1 . Due to the use of electrolyte, there might have been an increased humidity inside the electrospinning chamber. The conductivity of humid air is greater than dry air. The humid air also facilitates the uniform deposition of nanofibers.
Moreover, PCL/gelatin nanofibers being hydrophilic in nature, their nanofiber mat remained in wet condition due to capillary wicking of electrolyte and supported the layer-by-layer deposition without any decrease in conductivity of the substrate. After the uniform deposition of nanofibers of PCL/gelatin on the entire surface of the base component, they were bonded to its boundary component with partially cured PDMS glue. The PDMS was used as glue due to its inert, non-toxic nature during cell culture. Further, the PDMS glue resulted in leakproof bonding and afforded characterization of the cells by dismantling the components of the NCCI. We have systematically varied the parameters of 3D printing, electrospinning, and assembly of the components for optimal fabrication that is simple, scalable, and inexpensive.
Furthermore, we observed that ethylene oxide sterilization of the NCCIs was better than the use of ethanol.

FEM modeling and simulation for optimization of nanofiber deposition
The electrospinning of nanofibers on a substrate is dependent on its conductivity, and the pattern of deposition is primarily governed by the substrate chosen for deposition. The metallic substrates have been used conventionally to deposit the fibers, and thereafter, the fibers are peeled off for different applications. However, it is challenging to skillfully peel off the nanofibrous membrane and integrate it with other non-conducting substrates. Thus, it is imperative to develop strategies to deposit nanofibers on non-conducting polymeric substrates. Therefore, we have simulated the electric field and electric potential between a S3 needle and a substrate under a steady-state condition through a finite element method in COMSOL Multiphysics 4.2 2 . We defined the two-dimensional geometry of needle and substrate with air as a dielectric medium between them ( Figure S2A & S2B). The two cases have been investigated; one, where the substrate is a non-conducting polymeric material, and second, where a non-conducting polymeric substrate is covered with a conducting electrolyte layer.
Polymeric nanofibrous membranes were fabricated by the electrohydrodynamic drawing of polymers in an external electric field on a non-conducting substrate through an electrolyte-assisted electrospinning process. The nanofibers were deposited on a nonconducting substrate having alternate regions of conductivity. It can be observed that the strength of electric field distribution (different length of the red arrow) in the conductive and non-conductive regions were different, as shown in Figure S2C & S2D. It is due to the high dielectric constant of the non-conducting substrate that the electric field was concentrated to a greater extent in the exposed conducting part of the substrate as compared to nonconducting regions, which results in a differential electric field formation on a substrate. This method can be utilized for manipulating or patterning the electric field on a substrate in an electric field, thereby, can be used for nanofibers or nanoparticles patterning during the electrospinning process. However, the use of a thin electrolyte layer over the nonconducting substrate leads to a contiguous electric field on the substrate. This can be clearly observed by the uniform length of the red arrow near the substrate surface, irrespective of the conducting and non-conducting regions. The spatial distribution of electric field strength near the surface of the substrate (0.5 mm from the substrate) shows a sinusoidal and uniform pattern of the electric field on the non-conducting substrate without and with electrolyte, respectively ( Figure S2E and S2F). Moreover, the strength of the electric field due to the use of the electrolyte was higher as compared to one without the electrolyte. The equation describing the fluctuation of the electric field near the surface of the substrate without an electrolyte is given by (eq. 1) Where Eo is the basic electric field.

S4
It can be simulated that a charged particle (i.e., charged nanofiber/nanoparticle) will follow different velocities at different regions of the non-conducting substrate when electrospinning without an electrolyte ( Figure

FEM modeling and simulation of RVEs in NCCI
The CAD model of the free-hanging microporous membrane present in the commercial cell culture inserts, as shown in Figure S3. We have developed the NCCI with "eye" regions where the nanofibrous membrane exists as free-hanging structures. As a hanging membrane either in NCCI or commercial inserts, the membrane is always in pre-stressed condition before it is used for seeding cells. Therefore, we have used finite element modeling (FEM) to evaluate the stress distribution, available strain energy, and degree of displacement of the membrane under its own weight. We simulated the stress distribution on the representative volume element (RVE) of the hanging membrane of the commercial insert and nanofibrous membrane of NCCI under a steady-state condition using structural S5 mechanics module in COMSOL Multiphysics 4.2 3 . Under the constraint boundary and body loading condition, boundary conditions are employed during the FEM simulation. It was observed that strain energy density is comparatively higher in NCCI as compared to the commercial inserts ( Figures S3 & S4). Apart from strain energy density, the membrane in NCCI experiences higher deformation leading to a sagging structure as compared to the commercial insert substrate ( Figure S4).

Biomechanical theory of Nanoindentation
The empirical equation obtained from the force displacement data modeling ( Figure S6) has a close resemblance with the JKR theory of contact mechanics 4  Finally, there is a release of energy during the indentation process that comes from the adhesion of an indenting particle to the stretched substrate. This process drives the indentation process, and the energy liberated is given by W times the adhered area