Gene Expression by Fibroblasts Seeded on Small Intestinal Submucosa and Subjected to Cyclic Stretching

Thomas W. Gilbert, Ann M. Stewart-Akers, Jennifer Sydeski, Tan D. Nguyen, Stephen F. Badylak, and Savio L-Y. Woo

Introduction
Previous research has demonstrated that porcine-derived small intestinal submucosa extracellular matrix (SIS-ECM) scaffolds can be successfully manipulated and applied for tendon and ligament repair in preclinical animal trials. In vivo experiments show that if the SIS-ECM scaffold is subjected to a site-specific mechanical environment, the healing response is improved, allowing the scaffold to promote the formation of site-specific tissue. As a result, the scaffold remodeling process reduces the incidence of scar tissue. The aim of this study is to characterize the response of fibroblasts seeded on SIS-ECM due to mechanical loading. This was accomplished in vitro by examining the changes in expression of matrix-related genes that would be predictive of the in vivo remodeling process and by correlating these changes with alterations in mechanical behavior of the SIS-ECM during stretching experiments.

Materials and Methods
For this study, a cyclic-stretching tissue culture (CSTC) system was specifically developed to independently measure in real-time the mechanical load of each scaffold. Additionally, the system applies specific displacement waveforms to each scaffold without disruption of the cell culture environment. This CSTC system consisted of eight independently operating stations, with each station containing a culture chamber, tissue clamps, a linear actuator and a load cell (Fig 1).

Figure 1. (A) Photograph of the CSTC system with the independent operating stations. (B) Representation of a single operating station, containing a linear actuator, sterile stretching chamber and a load cell to monitor load generated by the scaffold.

Porcine small intestine was harvested from market-weight pigs immediately after euthanasia. The SIS-ECM was isolated, de-cellularized in a 0.1% peracetic acid/4% ethanol solution, and rinsed in phosphate buffered saline and de-ionized water. SIS-ECM scaffolds were then rehydrated in modified DMEM and cut into fragments 4 cm in length and 0.8 cm in width. Each scaffold was seeded with 1.6 x 106 NIH 3T3 cells for 8 hours to allow for cell attachments. Subsequently, each cell-seeded scaffold was transferred to one station in the CSTC system and allowed to acclimatize to this environment for 36-40 hours before stretching. The scaffolds endured a 0.05N preload to establish a zero position and then was subjected to cyclic stretch elongated to a baseline stretch of 2.5%, 5%, 10%, or 15% at 0.1Hz, 0.3Hz, and 0.5 Hz for 20 minutes at 8-hour intervals for 3 days.

RNA was subsequently extracted from the cell/scaffold complexes using an RNeasy kit (Qiagen). RT-PCR was performed to synthesize cDNA with gene primers specific for murine Col I, Col III, SMA, TN-C, MMP-2, MMP-9, TGF-β1, TGF-β3 and GAPDH. PCR products were visualized on 2% agarose gel under UV light following DNA electrophoresis as shown in Figure 2.

Figure 2. Resulting RT-PCR bands for GAPDH, Col I, and Col II detected on ethidium bromide gels.

Finally, several cell/SIS-ECM scaffold complexes were fixed in 2% paraformaldehyde, permeabilized in 0.1% Triton X-100, washed with 0.5% BSA and incubated with Phallodin Alexa-488. The specimens were washed again with 0.5% BSA and stained with Hoechst dye and viewed with confocal microscopy.

All comparisons between relative expressions of each gene were performed using a Student’s t-test with significance set at a p value of 0.05.

Results
Actin staining revealed that fibroblasts responded to cyclic stretching by aligning in the direction of stretch, as shown in Figure 3.

Figure 3. 3D reconstruction of Hoechst and actin staining of NIH-3T3 fibroblasts seeded on SIS-ECM scaffolds and subjected to cyclic stretch at 0.1 Hz. Cells align in direction of applied stretch, as indicated by the white arrow.

Col I expression increased most significantly due to cyclic stretch and changes in expression were more dependent on frequency of stretch than on magnitude (Fig. 4A). Col III expression was primarily dependent on frequency of stretch but showed a decrease in expression with increasing stretch frequency. This magnitude of change in Col III expression was not as dramatic as that of Col I expression (Fig 4B).

Figure 4. (A) Col I and (B) Col III to GAPDH relative expression for NIH-3T3 cells seeded on SIS-ECM scaffold and subjected to different stretching regimens. NIH-3T3 cells cultured on TCP were used to determine basal levels of expression. # shows statistical difference from 0% stretch control and * indicates statistical difference between groups (p <0.05).

SMA expression by fibroblasts seeded on SIS-ECM scaffolds increased with mechanical stretch (Fig 5A). TN-C expression, however, decreased in response to cyclic stretching, but increased with increasing stretch frequency (Fig 5B).

Figure 5. (A) SMA and (B) TN-C to GAPDH relative expression for NIH-3T3 cells seeded on SIS-ECM scaffold and subjected to different stretching regimens. NIH-3T3 cells cultured on TCP were used to determine basal levels of expression. # shows statistical difference from 0% stretch control and * indicates statistical difference between groups (p <0.05).

MMP-2 expression decreased slightly with increasing stretch frequency. The only significant difference was detected for 0.3 Hz at 15% stretch, which showed a 50% decrease in expression compared to the no stretch condition (Fig 6A). MMP-9 expression increased with stretch frequency at 15% stretch, though all the results were not statistically different from the MMP-9 expression under the no stretch condition (Fig 6B).

Figure 6. (A) MMP-2 and (B) MMP-9 to GAPDH relative expression for NIH-3T3 cells seeded on SIS-ECM scaffold and subjected to different stretching regimens. NIH-3T3 cells cultured on TCP were used to determine basal levels of expression. # shows statistical difference from 0% stretch control and * indicates statistical difference between groups (p <0.05).

TGF-β1 expression showed dependency on frequency of stretch, increasing in a frequency-dependent manner (Fig 7A). TGF-β3 expression also significantly increased in a frequency-dependent manner at 10% and 15% stretch (Fig 7B).

Figure 7. (A) TGF-β1 and (B) TGF-β3 to GAPDH relative expression for NIH-3T3 cells seeded on SIS-ECM scaffold and subjected to different stretching regimens. NIH-3T3 cells cultured on TCP were used to determine basal levels of expression. # shows statistical difference from 0% stretch control and * indicates statistical difference between groups (p <0.05).

The analysis of the mechanical behavior of the fibroblasts seeded on the SIS-ECM scaffold showed that stiffness and maximum load decreased during each 24-hour period (ie. interval 1 to interval 3, interval 4 to interval 6) and subsequently increased after reapplying the preload to the specimen. As a result of this cyclic stretching, each specimen experienced a permanent creep (Fig 8).

Figure 8. Load-elongation curves for one fibroblast-seeded/SIS-ECM complex subjected to 10% cyclic stretch at 0.1 Hz.

Discussion
A novel CSTC system was developed and can be used to study the effects of cyclical mechanical loading on cells seeded in a particular scaffold. This system is useful for its ability to independently apply specific displacement waveforms to each scaffold while providing continuous measurement about loading capacity. In this study, the CSTC system provided further functionality by allowing the investigation of gene expression of matrix-related proteins by fibroblasts after cyclic stretching.

The results of this study indicate that the expression of Col I increased substantially due to cyclic stretching in a frequency-dependent manner, while the expression of Col III decreased. Previous studies show that in vitro Col I and Col III expression increase in response to mechanical stimuli. However, the SIS-ECM environment may allow for a more normal ratio of Col III to Col I and a normal distribution of collagen fibrils in healing tissue, explaining the improved mechanical properties in vivo.

Additionally, the increased expression of SMA, TN-C, and TGF-β1 suggest that the fibroblasts seeded in the SIS-ECM scaffold become more contractile with frequency of stretch, aligning in the direction of stretch. The increased expression of TGF-β3 may moderate contractile behavior, though the precise mechanisms are unknown since TGF-β3 has been shown to decrease contractility of cells. MMP-2 and MMP-9 expression did not change significantly in the presence of mechanical loading. This may be explained as MMP-2 and MMP-9 have been suggested to play a role in collagen degradation and remodeling, which may not be important in the mechanisms of in vivo scaffold degradation.

Critique
This study provides the foundation for the initial investigation into the in vivo healing response as affected by mechanical stimulation on SIS-ECM cell-seeded scaffolds. It devises a new CSTC system to specifically monitor the response of fibroblasts seeded on SIS-ECM scaffolds due to cyclic stretching. Furthermore, it comprehensively examines the expression of a variety of genes under various stretch conditions that may be predictive of the in vivo response.

However, this investigation contains several limitations in addressing its primary objectives. One limitation is the use of the NIH-3T3 cell line, an immortalized cell line that is not necessarily predictive of the in vivo healing response. While this cell line is useful for a preliminary study, it is more informative to look at primary cell lines or a variety of cell types since many cell populations are involved in the ECM remodeling process. Additionally, it may be useful to perform protein assays specific for the proteins of interest instead of simply gene expression, as many proteins may be post-translationally modified in unique ways that improve ECM remodeling. The study also may have benefited by including multiple durations of stretch, instead of one designated period. The authors do not mention that this interval most closely matches the in vivo environment, and thus, it is informative to find the optimum duration of stretch. Additionally, this study acknowledges that other stretching regimens should be applied in the future which may contribute to the remodeling response.

This study also did not incorporate a control scaffold (fibroblasts seeded on a collagen sponge or a silicone membrane) to measure the difference in gene expression due to mechanical loading on this scaffold compared to the SIS-ECM scaffold. The authors mention that the transmission of stretch within these control scaffold materials would be challenging. Therefore, it is beneficial to investigate other types of control scaffolds to have a valid standard of comparison to the experimental SIS-ECM scaffold. Finally, lengthening the duration of this study would also be helpful in applying these results to an in vivo environment since these SIS-ECM scaffolds typically degrade within 60-90 days.

As a side note, I would have appreciated it if some of the figures contained a scale bar to more easily allow the reader to follow their study. Overall, this study provides a good starting point for understanding the effects on cells seeded on SIS-ECM scaffolds due to mechanical loading, but further investigation is required.