Utilizing multiscale engineered biomaterials to examine TGF‐β‐mediated myofibroblastic differentiation

Cells integrate many mechanical and chemical cues to drive cell signalling responses. Because of the complex nature and interdependency of alterations in extracellular matrix (ECM) composition, ligand density, mechanics, and cellular responses it is difficult to tease out individual and combinatorial contributions of these various factors in driving cell behavior in homeostasis and disease. Tuning of material viscous and elastic properties, and ligand densities, in combinatorial fashions would enhance our understanding of how cells process complex signals. For example, it is known that increased ECM mechanics and transforming growth factor beta (TGF‐β) receptor (TGF‐β‐R) spacing/clustering independently drive TGF‐β signalling and associated myofibroblastic differentiation. However, it remains unknown how these inputs orthogonally contribute to cellular outcomes. Here, we describe the development of a novel material platform that combines microgel thin films with controllable viscoelastic properties and DNA origami to probe how viscoelastic properties and nanoscale spacing of TGF‐β‐Rs contribute to TGF‐β signalling and myofibroblastic differentiation. We found that highly viscous materials with non‐fixed TGF‐β‐R spacing promoted increased TGF‐β signalling and myofibroblastic differentiation. This is likely due to the ability of cells to better cluster receptors on these surfaces. These results provide insight into the contribution of substrate properties and receptor localisation on downstream signalling. Future studies allow for exploration into other receptor‐mediated processes.

epithelial cells also increases as cell contractility increases on stiffer substrates. 3,9Over-activation of TGF-β signalling due to increased substrate stiffness can contribute to the onset and progression of undesired fibrotic responses by driving myofibroblastic differentiation of fibroblasts and epithelial to mesenchymal transitions. 6,9Collectively, these prior studies highlight the importance of tissue stiffness on modulating cell behaviour.
A variety of natural and synthetic materials have been utilised to explore cellular mechanotransduction responses.Most notably, polyacrylamide (PA) gels are commonly used due to their biocompatibility, tunability to mimic various ECM stiffnesses, ease of fabrication and low cost.However, PA gels have numerous limitations.Notably, PA gels are predominantly used to alter linear elastic properties without controlling non-linear viscoelastic behaviour.Elastic behaviour refers to the ability of a material to deform in response to an applied stress and then return to this original shape in a linear manner.Viscoelastic behaviour, however, exhibit a feature known as creep, wherein the material will undergo slow deformation while subject to persistent applied stress.While the effects of elasticity on cellular behaviour have been well characterised, most biological tissues and naturally derived hydrogels, like collagen and fibrin, are viscoelastic.The role of the viscous component and the influence of cell receptor density and distribution remains understudied.
Recent studies have shown that non-linear viscoelastic properties influence cell behaviour irrespective of the substrate's elastic modulus. 10Mesenchymal stem cells seeded on hydrogels with identical elastic moduli, exhibiting either high creep or low creep, differentially influence cell adhesion, proliferation and differentiation. 11,12The high creep, more deformable hydrogels led to higher rates of adhesion and proliferation when compared to both the low creep, lower viscosity hydrogels and purely elastic hydrogels of the same stiffness.Other reports have studied myoblasts, 13 chondrocytes, 14 endothelial cells 15 and epithelial cells 16 on viscous substrates.When compared to purely elastic materials of the same stiffness, the use of non-linear viscoelastic materials resulted in increased spread area, adhesion and proliferation of these cell types.While these studies show the importance of non-linear elastic properties on influencing cellular responses, much remains to be elucidated regarding the role of viscosity in directing cell fate.Unfortunately, creating a non-linear elastic material with predictable time-dependent responses due to forces in the range of cellderived force is exceptionally difficult and usually requires trade-offs in optimsing several complicated factors.
Previously, our group has utilised microgel thin films with easily tunable viscoelastic properties to investigate the role of non-linear material properties in controlling fibroblast fate in the context of wound healing. 17By varying the amounts of intraparticle cross-linking during microgel synthesis, an inverse linear relationship exists between the polymer mobility and intraparticle cross-linking.As demonstrated previously, films constructed with highly cross-linked microgels have low loss tangents and low viscosity while films constructed with loosely cross-linked microgels have high loss tangents and high viscosity. 18We have previously characterised the film's Young's modulus and loss tangent.Low (1%) and high (7%) intraparticle cross-linked films displayed Young's modulus of 95 ± 20 and 114 ± 14 kPa and loss tangent of 1.8 ± 0.1 and 0.9 ± 0.2, respectively.Loss tangent is defined as G 00 /G 0 , with similar Young's moduli, it can be inferred that G 00 changes when intraparticle cross-linking density is changed. 17Using these films, we showed that fibroblast modes of migration were influenced by loss tangent.ROCK-mediated amoeboid migration was observed on high loss tangent films and Rac-mediated mesenchymal migration was observed on low loss tangent films.In a separate study, we created composite materials where microgel films were constructed on top of PA gels to independently control surface loss tangent and bulk modulus.We used these materials to investigate both cellular and nuclear morphology and gene expression in fibroblasts.Interestingly, the response of neonatal HDFs (HDFn) was more dependent on the underlying stiffness of the PA gels compared to the viscoelastic films. 18cause of the complex nature and interdependency of alterations in ECM composition, ligand density, mechanics and cellular responses, it is difficult to tease out individual and combinatorial contributions of these various factors in driving cell behaviour in homeostasis and disease.Therefore, the tuning of material viscous and elastic properties, along with ligand densities in combinatorial fashions would enhance our understanding of how cells process complex signals.To that end, in the studies described here, we utilise microgel thin films to control substrate mechanical properties at the cellular level and combine them with DNA origami to control receptor patterning at the sub-cellular level.DNA can self-assemble into designed nanostructures. 19,20DNA origami architectures make use of this programmable self-assembly capability to construct desired patterns with feature resolutions below 6 nm. 21DNA origami can be used to organise and display other molecules with low nanometre precision.Therefore, the combination of DNA origami with microgel thin films could allow for unparalleled control over multiscale material properties and provide insight into the synergistic contribution of substrate mechanics and receptor localisation on cell behaviour.Previous work by the LaBean group demonstrated the pre-clustering of TGF-β receptor (TGF-β-R)-binding peptide using a DNA origami decorated with streptavidin molecules for nanopatterning the biotinylated peptide. 22After exposure to low levels of TGF-β, non-transformed mouse mammary gland epithelial cells (NMuMg) seeded on these nanoassemblies displayed translocation of SMAD from the cytosol to the nucleus, indicating TGF-β activation.In comparison, cells exposed to the same level of TGF-β but without the pre-clustering nanoassembly, demonstrated no translocation of SMAD or TGF-β-R activation.That study concluded that pre-clustering of cellsurface receptors by patterned ligands on DNA origami sensitised cells to activate at lower concentrations of endogenous TGF-β. 22F-β is a key player in wound healing responses, serving as a chemoattractant for the recruitment of immune cells to the wound bed and a regulator of myofibroblastic differentiation and ECM production.Because of the known interplay between TGF-β activation/ signalling in cellular mechanotransduction responses in fibrotic conditions, we aimed to further develop and combine these two previously designed systems to interrogate how surface viscosity influences TGF-β-R clustering and downstream responses.We hypothesised that high loss tangent films would lead to an increase in TGF-β activation, due to the ability of fibroblasts to more efficiently cluster TGF-β-Rs; and that higher levels of receptor clustering would lead to a proportional increase in TGF-β signalling.In the study reported here, we use viscoelastic microgel thin films functionalised with a TGF-β-R-binding peptide (either randomly patterned or with controlled nanospacing imparted by DNA origami) to examine the effects of substrate viscoelasticity and cell surface receptor spacing on TGF-β activation.

| Material design overview
To study the complex interactions of surface viscoelasticity and receptor spacing, this study employs a rectangle DNA origami, one of the most extensively studied and characterised DNA nanostructures.
As shown in Figure 2, specific locations on the DNA origami, designated as P* anchors and O handles, are utilised to present the TGFβ-R-binding peptide in an organised, clustered, nano-patterned arrangement (P* anchors) and to bind the origami down to the gel particles (O handles).Microgel particles are conjugated with either P* or O* anchor strands, then these conjugated particles are used during thin film assembly to build the top layer of the film and bind to nonclustered peptide (using P* anchors) or DNA origami (using O* anchors).For pre-patterned substrates (Figure 1E), the O* anchor conjugated particles hybridise to the complementary O handle on the bottom of the DNA origami.On the top of the origami, the P* anchor hybridises to the complementary P-PNA-Peptide, where this peptide sequence binds TGF-β-Rs.On randomly patterned substrates (Figure 1E), the DNA origami is omitted, and the 1% or 7% crosslinked microgels are conjugated only to P* anchor.The conjugated particles then directly hybridise to the P-PNA-Peptide, displaying randomly patterned (non-clustered) peptides, which subsequently bind to TGF-β-Rs on the cultured cell surfaces.These two options, with and without origami, provide the pre-patterned and randomly patterned peptide ligand distributions, respectively.Figure 2 provides more geometric detail for the cluster of eight peptide-binding sites on each origami.The production of each individual component of this design, including microgel synthesis, film fabrication, origami production and final composite material assembly, are detailed below.

| Microgel particle synthesis and extension peptide conjugation
Microgel particles were synthesised in a precipitation-polymerisation reaction as previously described. 17,18Particles were synthesised using 1% or 7% N,N 0 -methylenebis(acrylamide) (BIS), a constant 5% acrylic acid (AAc), 94% or 88% poly(N-isopropylacrylamide) (poly-NIPAM) and 0.8 and 0.4 mM SDS, respectively.Using a total monomer concentration of 140 mM, poly-NIPAM and BIS were mixed in ultrapure water with SDS to a final volume of 100 mL.Previous research has demonstrated the ability of SDS to control microgel size distribution, where increasing SDS concentration decreases microgel size. 23The solution was quickly vortexed, filtered using a 0.22-μm Steriflip-GP polyethersulphone (PES) filter and added to a three-necked reaction vessel.The solution was equilibrated for an hour at 70 C, then 5% AAc was added to the reaction along with 1 mM ammonium persulphate (APS) to initiate the reaction.The reaction proceeded for 5.5 h, with stirring at 450 rpm and was then cooled and filtered using glass wool to remove large microgel aggregates and transferred to 1000 kDa dialysis tubing for further purification.The microgel particle suspension dialysis was conducted at volumes of 30-35 mL of microgel particles to 1000 mL of water.The dialysis was allowed to proceed for 72 h with three buffer exchanges in deionised (DI), ultrapure water.The samples were then lyophilised and stored.

| Microgel particle size characterisation
Microgel particle hydrodynamic diameter was determined using a NS300 NanoSight (Malvern), where samples were diluted at varying concentrations.Measurements were completed for a total of 3 runs at 60 s, where the equipment temperature remained at 25 C. Microgel particle dry diameter and height were measured using MFP-3D atomic force microscopy (AFM) in AC Mode.The ARROW-NCR cantilever used demonstrated a spring constant of 42 N/m and a minimum of three particles were analysed per microgel synthesis.1.These concentrations were chosen as they were found to result in approximately equivalent amounts of conjugated P or O peptide (Figure 3B).Conjugated microgels were transferred at À20 C until use for film completion.Crosslinked microgels of 1% or 7% were conjugated to O* or P* anchor.

| Conjugation of peptides to microgel particles and construction of four-layer film
Three-layer microgel thin films were created using centrifugal deposition.A glass coverslip was functionalised using a 1% solution of particles was added to coverslips in 0.9 mL of ultrapure water and centrifuged at 3700 rpm for 10 min.The microgel solution was removed and films were washed with DI water.

| Analysis of relative peptide-binding site concentrations on films
Cy5-labelled P Probe (5 0 -Cy5-GCG TTG GTG ACT GCA TAA AAA-3 0 ), which is complementary to the P* Anchor was added at a concentration of 0.1 μM for 24 h at 4 C. Validation experiments were completed to determine the respective measurement of fluorescence intensity on samples, confirming equivalent amounts of peptide-binding sites present on each sample (Figure 1E).Following incubation, the microgel films were washed three times using deionised water.The films were allowed to dry completely and then half of the film was scraped to determine background fluorescence levels and imaged using an ECHO Revolve Fluorescence Microscope at 4Â magnification.Three sections, top, middle and bottom, were captured on the fluorescent film side and the intensity density was calculated using the mean background of the scraped side of each film.At least three films were analysed per condition, capturing three images per film sample.Experiments were also performed with fluorescently labelled fibronectin (Sigma-Alrich, Fibronectin-FITC, F2733) to confirm the accessibility of TGF-β-R-binding peptides. of co-linear synthesis of an affinity peptide that binds to TGF-β-R 24 plus a PNA region with Watson-Crick complementarity to P* Anchor sequence for binding/patterning on DNA origami or directly on films (see Table 1 and Figure 1E).To decorate P-PNA-Peptide on DNA origami-film conjugates, 1 mL 100 nM P-PNA-Peptide was added and incubated at room temperature for 2 h with shaking at 100 rpm.Subsequently, excess P-PNA-Peptide stands were removed with a DNA origami folding buffer wash by shaking at 100 rpm for 5 min, with three buffer exchanges.

| Preparation of DNA origami
Following the final wash, 20 μg/mL Human Plasma Fibronectin (Thermo Fisher Scientific) was diluted in 1Â PBS, added to the microgel films and then incubated for 24 h at 4 C. Before fibroblast seeding, fibronectin was aspirated, and films were washed once with Endotoxin-Free Dulbecco's PBS (1Â, without Ca ++ or Mg ++ ).

| Culturing and seeding HDFn cells
HDFn were cultured in Dulbecco's modified Eagle's medium-high glucose, with 4500 mg/L glucose, sodium pyruvate, sodium bicarbonate, with the addition of 2 mM L-glutamine (1Â L-Glut), 10% FBS, 100 U/mL of penicillin and 100 μg/mL streptomycin.Cells were cultured until 80% confluency was reached, then cells were removed with trypsin and seeded onto fibronectin-coated films at 3249 cells/cm 2 .Upon seeding cells, a final concentration of 200 pM of active TGF-β was added to all samples.Cells were cultured on films at 37 C for 24 h.

| Immunofluorescence staining on microgel films
After 24 h in culture on films, cell culture media was aspirated and films were washed once with Endotoxin-Free Dulbecco's PBS (1Â,

| Quantification of phosphorylated SMAD3 using HDFn cell lysate
Dermal fibroblasts were seeded on control (1%, 7% Films, Glass) or experimental (1% and 7% pre-patterned or 1% and 7% randomly patterned) surfaces at 3249 cells/cm 2 (12,000 cells/well).The cells were incubated for 24 h at 37 C. Next, cells were lysed using 1Â Cell Lysis Buffer (Thermo Fisher Scientific).Samples were placed on the lysis buffer, with the cell side adhering on the solution for 4-5 min.
The 100 μL sample was pooled for three films, and three independent pooled samples were evaluated.A protein quantitation assay was performed using NanoOrange™ Protein Quantitation Kit (Thermo Fisher Scientific) to determine total protein concentration for each cell lysate and samples were then evaluated with the SMAD3 (Phospho) (pS423/ pS425) Human InstantOne™ ELISA Kit (Thermo Fisher Scientific).Cell lysates were prepared at a concentration of 0.3 mg/mL in 1Â cell lysis buffer for the SMAD3 ELISA.The positive control cell lysate was further diluted using 1Â cell lysis buffer to create a range of 40%, 20%, 10% and 5% positive control cell lysate.The total phosphorylated SMAD3 was analysed in comparison to the positive control cell lysate range.Outliers were identified and removed for 7% randomly and pre-patterned samples.A normality test completed on the data with outliers removed, using a D'Agostino and Pearson test.Data were found to be non-normal and a Kruskal-Wallis non-parametric test was completed.

| Statistical analysis
Statistical analysis was performed in GraphPad Prism.The data were analysed using a one-way ANOVA, with comparison of each subgroup mean using Tukey's test.Statistical analysis was achieved for p < 0.05.
For immunofluorescence staining, a total of three films were constructed per condition and three images were captured per film.For each individual capture, a minimum of 10 cells were analysed.Each experiment was conducted in duplicates.For the InstantOne™ ELISA, a total of nine films were constructed per condition and three films per condition were pooled for one sample.

| Characterisation of microgel particles and DNA origami
To generate microgel thin films, we first synthesised microgel particles using 1% or 7% BIS cross-linker (Figure 1A).These intraparticle crosslinking densities were chosen to regulate thin film viscoelasticity, where low intraparticle cross-linking represents high loss tangent films and more viscous substrates. 17From AFM analysis, the 1% BIS and 7% BIS particles had dry diameters of 523 ± 48 and 595 ± 27 nm (Figure 1B), respectively, when deposited on glass coverslips.Using Malvern NanoSight particle tracking analysis to determine hydrodynamic diameter, the 1% and 7% BIS particles displayed a diameter of 304 ± 12 and 324 ± 15 nm, respectively (Figure 1B).For characterisation of the DNA origami, samples are deposited on a freshly cleaved mica surface and imaged in AC mode on Asylum MFP-3D AFM.For the rectangular origami design utilised, the dimension of DNA origami is approximately 70 Â 95 nm 2 .We also measured the height of the tall rectangle DNA origami to be $2 nm, as shown in Figure 1C.DNA origami staple strands are organised in four clustered and four distal locations (Figure 2).Layer-by-layer assembly of microgels with alternating layers of PEI on clean functionalised glass coverslips was used to fabricate microgel thin films (Figure 1D).Previous studies have demonstrated the inverse relationship between intraparticle crosslinking and healing response, due to the polymer mobility. 17Here, we demonstrated a tunable material platform displaying viscoelastic properties from microgel particles to be used in conjunction with DNA origami, to control nanoscale spacing and patterning of cell surface receptors.

| Characterisation of peptide strands on microgel films using fluorescently labelled probes
To confirm the presence of equivalent levels of peptide on both randomly and pre-patterned substrates, we utilised a fluorescent P* Probe which binds to the P Handle on either the origami or fourthlayer P* anchor conjugated particles for samples without patterning (Figure 1E).Microgel films were incubated at 4 C for 24 h with

| Evaluation of cellular response in response to nanoscale patterning and material loss tangent
Following confirmation of roughly equivalent peptide densities on each surface, films were coated in 20 μg/mL fibronectin to facilitate cell attachment then human neonatal fibroblasts were plated and cultured for 24 h on the surfaces.Fibroblast protein expression and cellular morphology were evaluated after 24 h in culture to understand the relationship between material loss tangent and pre-clustered or randomly distributed cell surface receptors.After seeding human dermal fibroblasts on high or low loss tangent films randomly or pre-patterned substrates, were examined for key markers of myofibroblastic differentiation (Figure 5A).Previous research has confirmed TGF-β induces α-SMA expression. 25Dermal fibroblasts cultured on 1% BIS films, high loss tangent films, randomly patterned substrates demonstrated myofibroblastic differentiation as indicated by higher α-SMA expression (Figure 5B).Cells cultured on low loss tangent films had lower levels of α-SMA expression on randomly and pre-patterned substrates.The respective fluorescence values were: 1% prepatterned = 1.02E5 ± 6.36E4, 1% randomly patterned = 2.63E5 ± 2.00E5, 7% pre-patterned = 1.01E5 ± 6.63E4 and 7% randomly patterned = 1.38E5 ± 9.95E4.Comparing cellular outcomes on high loss tangent films, there was a two fold increase in α-SMA protein expression measured between pre-patterned to randomly patterned groups.Similarly, cellular outcomes on low loss tangent films demonstrated no significant differences in α-SMA protein expression; approximately a 1.37-fold increase between pre-patterned to randomly patterned groups.Cellular morphology was also found to be dependent on loss tangent and nanoscale patterning.On high loss tangent films, dermal fibroblasts demonstrated an elongated morphology as seen with high cell area and cell perimeter (Figure 6A,B).No statistical difference was seen between 1% pre-patterned samples and 7% pre-patterned and randomly patterned samples.In addition, the highest cell circularity was noted on high low tangent films with the presence of DNA origami (Figure 6C).These results indicate dermal fibroblasts have a more myofibroblastic phenotype on films randomly patterned and on high loss tangent films.This is presumably because cells on higher loss tangent films can more efficiently cluster TGF-β-Rs than those on lower loss tangent films when peptides are randomly patterned.Patterning of peptides 'fixes' the receptors inplace, therefore, the viscosity of the film has less of an influence.

| In vitro evaluation of phosphorylated SMAD protein expression
To further examine the role of TGF-β clustering on profibrotic responses, the expression of total phosphorylated SMAD3 was determined.First, dermal fibroblasts were seeded on high or low loss tangent films on pre-patterned or randomly patterned microgel films.The fibroblast cell lysate was prepared and the total phosphorylated SMAD3 was analysed using a Human InstantOne™ ELISA Kit.A significant increase in total pSMAD3 expression was noted on randomly patterned high loss tangent films, where there is a statistical difference with randomly and pre-patterned low loss tangent films (Figure 5C).Comparing the average mean absorbance values, randomly patterned high loss tangent films have a mean absorbance of 0.63 ± 0.59.Pre-patterned low loss tangent films have a mean absorbance of 0.18 ± 0.03.In addition, there is no statistical difference between the total pSMAD3 expression between randomly patterned high loss tangent films and 20% of total pSMAD3.

| DISCUSSION
In this study, we examined the relationship between nanoscale organisation of TGF-β-Rs and surface viscoelasticity in influencing cell adhesion, spreading, myofibroblastic differentiation and TGF-β signalling.
To do so, we created a model multi-scale material platform where microgel thin films with controllable viscoelastic properties were evidenced by higher levels of SMAD nuclear translocation compared to controls and glass samples coated with the TGF-β-R-binding peptide. 22This was hypothesised to be due to the pre-clustering of the receptors sensitising cells to lower concentrations of endogenous growth factor.In this manuscript, we further expand on these findings to determine how underlying material viscoelasticity contributes to these responses.We hypothesised that cells would be able to better cluster receptors on materials with higher levels of viscosity (high loss tangent films) compared to materials with lower levels of viscosity (i.e., low loss tangent films).We investigated this hypothesis by creating microgel thin films with high or low loss tangent with TGF-β-Rbinding peptide where the peptides were randomly patterned (i.e., non-clustered) or pre-patterned (i.e., nano-scale clustered) via DNA origami.
Cellular morphological changes in fibroblasts highlight the influence of surface viscoelasticity, or mobility of polymer chains, on spreading and adhesion (Figure 7).On high loss tangent films randomly patterned with TGF-β-R-binding peptide, there was a significant increase in α-SMA stress fibre positive cells.In addition, fibroblasts displayed cell elongation which is characteristic of increased cell contractility; previous research has noted the direct relationship between increased contractile forces and increased TGFβ activation. 26Interestingly, these results differed from recent reports of human lung fibroblasts cultured on purely elastic and viscoelastic substrates 27 where lung fibroblasts were found to display a decrease in cell spreading on both soft and stiff viscoelastic hydrogels, using fibronectin fragments to facilitate attachment.This difference likely highlights the importance of ligand specific interactions in contributing to these multiscale interactions.To confirm the activation of TGF-β, levels of total phosphorylated SMAD3 was characterised using SMAD3 (Phospho) (pS423/pS425) Human InstantOne™ ELISA Kit.A significant difference was detected in total phosphorylated SMAD3 expression between a randomly patterned high loss tangent substrate and a pre-patterned low loss tangent substrate.
These data support the hypothesis that the ability of fibroblasts to efficiently cluster cell surface receptors modulates TGF-β activation.These results provide insight into the contribution of substrate mechanics and receptor localisation on downstream signalling.These Despite the noted limitations, the results described herein, have far-reaching implications for the complex interplay of receptor clustering and tissue/substrate mechanics.Implications for the relationship between changes in viscosity and fibrotic progression have been noted in previous studies.In fibrotic, hypertrophic scars an approximate fourfold increase in surface tension measurements in comparison to healthy skin have been reported. 28In addition, previous studies have noted a decreased loss tangent values in fibrotic liver tissue compared to healthy tissue, with values of 0.03 and 0.06, respectively. 29These prior studies demonstrate how viscoelastic properties can vary during fibrosis; however, raw values are dependent on the specifics of the sample preparation and mechanical testing used. 30netheless, our biomaterial system allows us to mimic magnitude differences in healthy and fibrotic dermal tissue and will be useful for After purification, 1% and 7% BIS microgel particles were conjugated to anchor strands using N-ethyl-N 0 -(3-(dimethylamino)propyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) coupling at 50 and 20 mM, respectively.Microgel particles were added to 0.1 M 2-morpholinoethanesulphonic acid monohydrate buffer and 50 and 20 mM EDC/NHS, vortexed and incubated at room temperature for 1 h.Following incubation, peptide strands were added to microgel particles at final concentrations of: 1% P* = 40 nM, 1% O* = 20 nM, 7% P* = 200 nM and 7% O* = 150 nM, vortexed and incubated at room temperature for 2 h.At least three films were analysed per condition, measuring three locations on the film and background slide.No outliers in the normalised fluorescence were detected.A D'Agostino and Pearson test was completed on the raw data with outliers removed.Data were found to be non-normal and a Kruskal-Wallis non-parametric test was completed.The nucleotide sequences for DNA strands used are shown in Table

( 3 -
aminopropyl) trimethoxysilane in 200 Proof Ethanol, for 2 h with shaking at 50 rpm.Following functionalisation, glass coverslips were washed once with DI water and transferred to a new 12-well plate.Next, a 0.1 mg/mL microgel solution was added to the functionalised coverslips and centrifuged for 3700 rpm for 10 min.Following centrifugation, the microgel solution was removed, coverslips washed once with DI water and a 0.05 monomolar PEI solution was added.The coverslips were placed on a shaker for 30 min at 50 rpm.The PEI solution was then removed and the coverslips washed once again with DI water.This process was repeated until three microgel layers were constructed, with the final layer being microgel particles.Threelayer thin films were sterilised for 30 min in 30% ethanol solution and then incubated overnight in PBS at 4 C. Three-layer thin films were then rehydrated for 45 min in 0.05 monomolar PEI solution.The PEI solution was removed, 0.1 mL of 1 mg/mL peptide conjugated F I G U R E 1 Overview of film construction and component characterisation.(A) Microgel particles were synthesised via a precipitationpolymerisation reaction, varying the amounts of cross-linker-N,N 0 -methylenebisacrylamide (BIS) and poly(N-isopropylacrylamide) (poly-NIPAM) with a constant concentration of acrylic acid (AAc).The reaction was initiated with ammonium persulphate (APS).(B) Microgel synthesis ratios, Malvern NanoSight hydrodynamic diameter and Asylum atomic force microscopy (AFM) dry particle diameter.Representative AFM images and height traces are shown.(C) AFM height traces of DNA origami structure.(D) Overview of microgel thin films fabrication.(E) Overview of construction thin films coupled to DNA origami/peptide (pre-patterned) or TGF-β receptor-binding peptide alone (randomly patterned).

2 . 7 |
To construct tall rectangular DNA origami, ssDNA scaffold strand (M13mp18, 100 nM) was mixed with staple strands(10 equivalents)   in DNA origami folding buffer (1Â tris-acetate-EDTA (TAE) containing 12.5 mM MgCl 2 ) to a final concentration of 10 nM.The DNA origami mixture was annealed in a Thermocycler at 80-20 C over 3 h.The self-assembled DNA origami was purified using Microcon DNA Fast Flow spin filtration units to remove excess staple strands.The purified DNA origami was confirmed using AFM.Briefly, 2 μL of DNA origami T A B L E 1 DNA sequences.P* anchor 5 0 -NH2-C6-TTT TTA TGC AGT CAC CAA CGC-3 0 P probe 5 0 -Cy5-GCG TTG GTG ACT GCA TAA AAA-3 0 P-PNA-Peptide 5 0 -LTGKNFPMFHRN-GCG TTG GTG ACT GCA-3 0 O* anchor 5 0 -CGC ATT CAG GAT TCT CAA CTC GTA TTTT-C6-NH2-3 0 O handle 5 0 -(origami staple)-TAC GAG TTG AGA ATC CTG AAT GCG-3 0 F I G U R E 2 Nanoscale spacing on DNA origami.(A) Schematic showing details of the DNA origami including the pre-patterned spacing schematic for the P* (red) and O handle (green) strand sites.(B) Simplified design schematic of P* anchor strands (red) spacing on the top of the DNA origami.These sites provide eight binding points for peptide within an approximately 51 Â 42 nm patch on the origami.We have previously demonstrated that a similar pre-clustering scheme was able to sensitise cells to respond to lower levels of TGF-β. 22(C) Simplified design schematic of O handle strands (green) spacing on the bottom of the DNA origami.These sites provide binding of the origami down onto the microgel surface.was dropped on a freshly cleaved mica surface, allowing the DNA origami to absorb on the surface for 1 min.The mica surface was then washed with filtered, deionised water three times and then dried under a stream of nitrogen gas.The DNA origami was visualised using tapping mode in air with an MFP-3D AFM (Asylum Research).Adding DNA origami and P-PNA-Peptide to microgel films To immobilise DNA origami on microgel films, 2 nM DNA origami containing ssDNA O Handle stands in folding buffer (1Â TAE containing 12.5 mM MgCl 2 ) was incubated with O* Anchor-labelled microgel film at 4 C overnight.DNA origami unbound to films was removed by washing film with folding buffer by shaking at 100 rpm for 5 min and washing three times with 1Â TAE-Mg 2+ .P-PNA-Peptide was designed for ease Comparable amounts of randomly distributed or pre-patterned TGF-β receptor (TGF-β-R)-binding peptides on high or low loss tangent films.(A) Representative images of fluorescence probes binding to microgel films functionalised with randomly distributed or prepatterned TGF-β-R-binding peptides.Microgel films were constructed with equivalent amounts of respective DNA anchor strands and hybridised with a complementary fluorescence probe to elucidate peptide-binding sites.The images depicted demonstrated removing ½ of the microgel film to compare the fluorescence intensity of the surface with respective amounts of DNA anchor strands to a section without the microgel film.(B) Normalising the fluorescence intensity of 'top', 'middle' and 'bottom' scraped areas to sections with the microgel film, the corrected total fluorescence was calculated for all samples.Equivalent amounts of peptides on high or low loss tangent films were observed.w/o Ca++ and Mg++).Fibroblasts were fixed using 95% methanol and 5% glacial acetic acid and washed again with endotoxin-free Dulbecco's PBS (1Â).Samples were then blocked with 5% BSA in PBS-T (1Â PBS and 1:1000 Tween-20) for 30 min.A primary antibody solution was created using a 1:200 dilution of anti-actin, α-SMA antibody (A5228), mouse monoclonal (Sigma Aldrich) in PBS-T + 5% w/v BSA.The primary antibody solution was added to the dermal fibroblasts and placed on a shaker for 2 h at 100 rpm.After removing the primary antibody solution, an initial secondary antibody incubation of Thermo Fisher Scientific Goat anti-Mouse IgG (H + L) Cross-Adsorbed ReadyProbes™ Secondary Antibody, Alexa Fluor™ 488 and Goat anti-Rabbit IgG (H + L) Cross-Adsorbed ReadyProbes™ Secondary Antibody, Alexa Fluor™ 594 for 30 min at 50 rpm was completed.After 30 min, NucBlue™ Live ReadyP-robes™ Reagent (Hoechst 33342) was added for an additional 30 min at 50 rpm.Fibroblasts were washed once with PBS and mounted with 15 μL of Fluoromount-G mounting medium (SouthernBiotech) and then imaged using an EVOS FL Auto (Thermo Fisher Scientific) at 10Â magnification.For analysis, three films per condition and at minimum 10 cells are quantified from each image with a total of 36 images in each experiment.Experiments were conducted in duplicate.Outliers were identified and removed.A normality test was completed on the data with outliers removed, using a D'Agostino and Pearson test.Data were found to be non-normal and a Kruskal-Wallis non-parametric test was completed.

0. 1
μM of the fluorescent P* probe.Proceeding incubation, films were washed 3Â with deionised water.Microgel films were allowed to dry completely before performing a scratch on ½ of the film.Films were washed again 3Â with deionised water, dried and imaged on an ECHO Revolve Fluorescence Microscope.The mean fluorescence intensity of 'top', 'middle' and 'bottom' areas of the microgel thin films were normalised to the summative fluorescence intensity background in respective areas where film had been scraped off (Figure 3).The mean corrected total fluorescence value for each sample was as follows: 1% pre-patterned = 2.48E6 ± 2.35E6, 1% randomly patterned = 1.43E6 ± 4.29E5, 7% pre-patterned = 1.46E ± 1.38E6 and 7% randomly patterned = 1.54E6 ± 7.45E5.No statistical difference was detected between samples using a Dunn's multiple comparison test, indicating that roughly equivalent amounts of peptides were displayed on each experimental group.In addition, deposition of the DNA origami may cause clustering and variability in peptide presentation.Experiments performed with fluorescently-labelled Fibronectin and fluorescentlylabelled P* Probe confirmed the accessibility of both proteins on the surface of microgel thin films (Figure 4).

F I G U R E 4
Confirming accessibility of fluorescently labelled fibronectin and prepatterned TGF-β receptor (TGFβ-R)-binding peptides on high or low loss tangent films.Representative images of fluorescence fibronectin to microgel films functionalised to pre-patterned TGF-β-R-binding peptides indicating that both peptide and protein are accessible.Images are split into green (Fibronectin) and red (P* anchor) channels.Scale bar is 100 μm.
coupled directly to a TGF-β-R-binding peptide or DNA origami with nanoscale patterning of the same TGF-β-R-binding peptide.Previous studies by the LaBean group demonstrated that pre-clustering of cellsurface receptors by controlling nanoscale patterning of the TGF-β-Rbinding peptide via DNA origami enhanced TGF-β signalling, as F I G U R E 5 Random spacing of TGF-β receptors on high loss tangent films leads to an increase in TGF-β activation and signalling.(A) Representative images of fibroblasts cultured on 1% and 7% randomly patterned or prepatterned microgel films and stained with α-SMA (red) and DAPI (nuclei, blue).(B) Randomly patterned high loss tangent films demonstrate the highest amount of α-SMA stress fibre-positive cells.(C) Using the InstantOne™ ELISA, a statistical difference in total phosphorylated SMAD3 is noted between randomly patterned high loss tangent (1% randomly patterned) compared to randomly and pre-patterned low loss tangent films (7% randomly and pre-patterned).*p < 0.05; **p < 0.01; ****p < 0.0001.

F I G U R E 6
Microgel film loss tangent and TGF-β receptor spacing influences cell morphology.Dermal fibroblasts were seeding on fibronectin coated films for 24 h and analysed for morphological changes.A significant increase in (A) cell area and (B) cell perimeter was noted on randomly patterned high loss tangent films in comparison to pre-patterned high and low loss tangent films.Conversely, an increase in (C) cell circularity was observed on prepatterned high loss tangent films in comparison to randomly patterned high loss tangent films.****p < 0.0001.experiments were performed with the consideration that surface viscoelasticity may result in movement of the DNA origami or randomly patterned TGF-β-Rs.It should be noted that due to the differences in length scale of the microgel thin films, DNA origami and TGF-β-Rbinding peptides, there are limitations to characterisation techniques of the material platforms presented.Due to this, several characterisations were unable to be performed in the context of the present studies.However, future studies would benefit from the characterisation of TGF-β-R-binding peptide using sub-diffraction imaging and/or fluorescence resonance energy transfer microscopy and further mechanical testing using loss tangent imaging.Our prior studies characterised the loss tangent of microgel thin films alone as a function of microgel cross-linking and showed the ability to fine tune film loss tangent.We do not expect the DNA origami to greatly influence the bulk microgel thin film mechanical properties.The nanoscale DNA origami (70 Â 95 nm) is minuscule compared to the micron-scale gel particles and macro-scale thin films, and as such will only have local effects which are not expected to alter the behaviour of the whole gel.Also, the DNA origami does not introduce any new cross-links to the gel, therefore not influencing the film viscoelasticity.Another limitation of these studies was the high variability in P* binding to the 1% prepatterned substrates, which may influence interpretation of these results.
studying the interplay between receptor clustering and tissue/ substrate mechanics in a host of systems.Although these studies focused on the interplay between TGF-β and substrate mechanics, how mechanics influence receptor clustering for other growth factor receptors would also have implications in wound healing outcomes.This work also has relevance for wound healing research and applications beyond fundamental studies.For example, these multiscale films could be used as coating on wound dressings or implants to direct TGF-β responses and desired wound healing outcomes.The modular nature of this platform will enable future studies focusing on other growth factor-mediated responses, such as PDGF, VEGF and FGF which are also influential in the wound healing cascade.Notably, PDGF encourages fibroblast proliferation and subsequent myofibroblastic differentiation for collagen production.FGF functions include influencing cell proliferation and re-epithelialisation and promotion of angiogenesis.Aside from exploring other cell signalling pathways, with the programmable nature of DNA nanostructures, this platform could also be further modified in the future to optimise geometric arrangements of integrin receptors.Finally, these studies could motivate investigations in 3D microgel systems in the future to better mimic physiological ECM properties.F I G U R E 7 Summary of cell responses on pre-patterned or randomly organised TGF-β receptor (TGF-β-R)-binding peptides on low and high loss tangent films.Fibroblasts were cultured for 24 h on fibronectin-coated high or low loss tangent films.Films were coupled to either randomly attached or pre-patterned (via DNA origami) TGFβ-R-binding peptides.(A) On high loss tangent films with randomly patterned TGF-β-Rs, fibroblasts display an increase in stress fiber positive cells, an increase in cell area and perimeter correlated to a more elongated morphology; indicative of an increase in TGF-β signalling.(B) On low loss tangent films with randomly patterned TGFβ-R clustering, fibroblast display a reduction in cell spreading due to a decreased mobility of polymer chains with the increase of intraparticle cross-linking.(C,D) A significant reduction in stress fiber positive cells is noted with pre-patterned TGF-β-R clustering on both high (C) and low (D) loss tangent films compared to randomly organised peptide, with the greatest decrease being observed on low loss tangent films.