Substrates with soft rigidity (8 kPa) and normal rigidity (20 kPa) were prepared following the method mentioned in a previous research [14]. Specifically, solutions with different proportions components, including acrylamide (Sangon Biotech, China) and bis-acrylamide (Sangon Biotech, China), 10% ammonium persulfate solution and tetramethylethylenediamine (TEMED, Klamar, China) were prepared. Then 160 μL solution was dropped onto the gel slick-treated glass slide. A silanized cover glass was then placed onto the top of the liquid quickly. After polymerization, the substrate was removed from the glass slide and coated with collagen I (Corning, USA).

The rigidity of polyacrylamide (PA) substrates was detected by atomic force microscope (AFM, Bruker, Germany). The probe used for the AFM measurement was a glass ball probe with a radius of 5 μm. Three random points of each substrate were measured and the results were shown as mean ± SEM.

Human umbilical vein endothelial cells (HUVECs, Sciencell, USA) were cultured in endothelial cell culture medium (Sciencell, USA), containing penicillin/streptomycin (1%), endothelial cell growth supplement (1%) and fetal bovine serum (5%). HUVECs were culture at 37˚C and CO₂ (5%).

HUVECs were cultured on different substrates for 2 hours and imaged with a reverted microscope (Leica, German). When reached confluence, the cells were fixed with paraformaldehyde (4%). FITC-labeled phalloidin was used to stain the cytoskeleton and DAPI was used to stain the nuclei. After mounting, the samples were observed under laser scanning confocal microscope (Leica, Germany). All the processes were conducted in dark place to avoid photo-bleaching.

An RNA simple total RNA extraction kit (Tiangen Biotech, China) was used to extract total RNA following the manufacturer’s instruction. The genomic DNA was removed by gDNase treatment. After reverse transcription process, the samples were analyzed using a Talent qPCR PreMix kit (Tiangen Biotech, China) by SYBR green method. The internal control used in this study was GAPDH. All mRNA primers were provided by BioTNT (China) and the sequences were shown in Table 1.

A miRcute miR Isolation Kit (Tiangen Biotech, China) was used for extraction of microRNAs. Then the samples were reverse transcribed into complementary DNA (cDNA) using a miRcute miR First-strand cDNA synthesis kit (Tiangen Biotech, China). The cDNA was analyzed by SYBR green method using a miRcute miR qPCR Detection kit (Tiangen Biotech, China) following the instruction of the product. U6 was used a microRNA internal reference for the detection of microRNAs. The catalog numbers of U6 and miR-146a were CD201-0145 and CD201-0013, respectively. All microRNA primers were bought from Tiangen Biotech (China).

Whole cell lysates were extracted using RIPA buffer with the addition of phosphatase inhibitors and protease inhibitors (Tiangen Biotech, China). After quantification, equal amount of the samples were denaturized and loaded. Then the proteins were separated by SDS-PAGE and transferred into a PVDF membrane (Bio-Rad, USA). After blockage for 1 hour, the samples were incubated with mouse-anti-human p-JNK primary antibody (Cell signaling technology, USA) or rabbit-anti-human JNK primary antibody (Cell signaling technology, USA) at 4˚C overnight. After rinsed for 3 times, the membranes were incubated with HRP-conjugated rabbit-anti-mouse secondary antibody (Cell signaling technology, USA) or goat-anti-rabbit secondary antibody (Cell signaling technology, USA) for 1 hour at the room temperature. After washing, the samples were detected

using an enhanced chemiluminescence system (Bio-Rad, USA) and semi-quantified with Image J by comparing the intensity of the images (NIH, USA). β-tubulin was used as an internal reference in the western blot assay.

GraphPad Prism 6.0 software was applied for statistical analysis. All data represented the mean of triplicates unless otherwise indicated and error bars signified standard error of the mean (SEM). Unpaired student’s t-test was used for the comparison of two means. Significance was considered if p value < 0.05 and marked with *.

The rigidity of PA substrates can be tuned by adjusting the proportion of acrylamide and bis-acrylamide. Using AFM, the rigidity of different substrates was confirmed. As Figure 1 indicated, the slope of the curves indicated the rigidity of the matrices. The dotted line showed a more inclined slope than the solid line, indicating the higher rigidity of the 20 kPa than 8 kPa. Furthermore, young’s modulus of the substrates verified that rigidity of the fabricated substrates was consistent with our expectation. The mean rigidity was 7.37 kPa and 23.2 kPa, respectively (Table 2). Thus substrates with rigidity of 8 kPa and 20 kPa were fabricated successfully.

HUVECs were seeded onto the surface of the collagen I coated PA substrates.

After 2 hours of incubation, the ECs were imaged. As Figure 2(A) & Figure 2(B) indicate, ECs already adhered onto the PA substrates and no observable significance was found. After the ECs formed monolayers on PA substrates, the cytoskeleton was stained. However, the polygonal morphology of ECs was observed and no obvious change of morphology was acquired on both soft and normal substrates (Figure 2(C) & Figure 2(D)).

After ECs reached complete confluence on different substrates, the total RNA was extracted and analyzed. The expression of IL6, TNF-α and IL1β was dramatically increased in ECs on soft substrate compared to normal substrate (Figure 3). Concretely, the expression of IL6, TNF-α and IL1β in ECs on soft substrate was

about 1.4-fold (Figure 3(A)), 2.4-fold (Figure 3(B)) and 1.6-fold (Figure 3(C)) compared with normal substrate. Therefore, these results indicated that soft substrate induced the pro-inflammatory response of ECs.

Since the expression of inflammatory cytokines was changed by substrate rigidity, we then studied whether the inflammation-related signaling cascades were impacted by substrate rigidity. C-Jun N-terminal kinases (JNK) in mitogen-activated protein kinase (MAPK) signaling pathway were shown to regulate the expression of inflammatory cytokines, including IL6, TNF-α and IL1β [15] [16]. Thus, the activation of MAPK signaling pathway was analyzed (Figure 4). Unexpectedly, the phosphorylation of JNK in ECs on soft substrate was significantly enhanced compared to normal substrate (Figure 4(A) and Figure 4(B)), illustrating strong regulation of JNK signaling pathway by ECM rigidity.

To further explore the reason why JNK signaling pathway was involved in EC inflammatory response induced by substrate rigidity, the expression of TNF receptor associated factor 6 (TRAF6), an important regulator of JNK signaling path way, was examined. Surprisingly, the expression of TRAF6 in soft rigidity group was high than normal rigidity group (Figure 5(A)), implying activation of

JNK signaling by soft substrate. Since it has been verified that miR-146a was a direct suppressor of TRAF6 [17] [18] [19], the expression of miR-146a was determined subsequently. Consistently, miR-146a was significantly less (0.4-fold) in ECs on soft rigidity substrate compared to 20 kPa group (Figure 5(B)), indicating inhibition of miR-146a expression by soft ECM rigidity. These results implied that soft substrate promoted the phosphorylation of JNK via reducing miR-146a (targeting TRAF6) expression compared to normal substrate rigidity.