Evaluation of two distinct placental‐derived membranes and their effect on tenocyte responses in vitro

Abstract Tendon healing is a complex, multiphase process that results in increased scar tissue formation, leading to weaker tendons. The purpose of this study was to evaluate the response of tenocytes to both hypothermically stored amniotic membrane (HSAM) and dehydrated amnion/chorion membrane (dACM). Composition and growth factor release from HSAM and dACM were evaluated using proteomics microarrays. HSAM and dACM releasate was used to assess tenocyte proliferation, migration, gene expression, extracellular matrix (ECM) protein deposition, and response to inflammation. Additionally, tenocyte–ECM interactions were evaluated. HSAM and dACM contain and release growth factors relevant to tendon healing, including insulin‐like growth factor I, platelet‐derived growth factor, and basic fibroblast growth factor. Both dACM and HSAM promoted increased tenocyte proliferation and migration; tenocytes treated with dACM proliferated more robustly, whereas treatment with HSAM resulted in higher migration. Both dACM and HSAM resulted in altered ECM gene expression; dACM grafts alone resulted in increases in collagen deposition. Furthermore, both allografts resulted in altered tenocyte responses to inflammation with reduced transforming growth factor beta levels. Additionally, dACM treatment resulted in increased expression and production of matrix metalloprotease‐1 (MMP‐1), whereas HSAM treatment resulted in decreased production of MMP‐1. Tenocytes migrated into and remodeled HSAM only. These results indicate that both grafts have properties that support tendon healing; however, the results presented here suggest that the responses to each type of graft may be different. Due to the complex environment during tendon repair, additional work is needed to evaluate these effects using in vivo models.


| INTRODUCTION
In the United States, there are approximately 32 million musculoskeletal injuries annually; of these, approximately 45% are tendon and ligament injuries (Yang, Rothrauff, & Tuan, 2013); 300,000 of the 14.4 million patients suffering a tendon or ligament injury will undergo surgical repair (Yang et al., 2013). Although surgical repair is generally successful in restoring tendon function, adequate healing may be a concern, particularly for patients with multiple comorbidities (Ackermann & Hart, 2016). In addition to adequate tendon healing, another concern is the formation of adhesions to surrounding tissue caused in part by interruption of the synovial sheath (Matthews & Richards, 1976).
Adhesion formation disrupts the normal gliding mechanics and consequently results in severe limitations in the functionality of the affected tendon. For flexor tendon repair, the occurrence of adhesion formation has been estimated at 4-10%, representing a significant number of adverse outcomes following surgery (Dy, Hernandez-Soria, Ma, Roberts, & Daluiski, 2012;Tang, 2005).
Currently, many clinicians employ the use of cellular or acellular grafts to reinforce tendon repair sites with the goal of improving healing outcomes and reducing adhesion formation. One class of grafts available is human placental tissue grafts (Riboh, Saltzman, Yanke, & Cole, 2016). Although these grafts do not provide mechanical strength for physically reinforcing the repair, they contain a multitude of factors that are potentially beneficial to repair as indicated by recent studies. In vivo transplanting amniotic cells in a sheep tendinopathy model has shown improved mechanical and structural repair and faster ECM remodeling into mature collagen (Barboni et al., 2012). Additionally, a rodent study treating tendon lesions with amniotic membrane resulted in reduced healing times, improved collagen fiber alignment, a reduction in the inflammatory response, and an increase in tenoblast proliferation (Nicodemo et al., 2017).
In limited early clinical work, patients receiving treatment with a liquid form of placental tissue matrix reported reduced pain and improved function of the treated appendage (Lullove, 2015;Warner & Lasyone, 2014;Werber, 2015). Although this existing research has demonstrated the potential for placental tissue to support tendon repair, few studies have utilized commercially available grafts.
The purpose of this study was to evaluate the effects of two commercially available placental-derived membranes on tenocyte responses in vitro. We hypothesized that placental-derived grafts would interact with tenocytes in vitro in ways that are expected to be beneficial to tendon repair. Second, we hypothesized that the processing methodologies and components of these grafts would influence how placental membranes interact with tenocytes in vitro.
The two grafts evaluated in this study included a viable amnion-only graft and a nonviable dehydrated graft consisting of both amnion and chorion membranes. The viable amnion-only graft, also referred to as a hypothermically stored amniotic membrane (HSAM), contains amniotic membrane maintained in a proprietary solution at 1-10°C in order to preserve the cellular viability and overall integrity of the membrane. Growth factors, cytokines, and mechanisms by which HSAM may support wound healing responses in vitro have been previously evaluated . Evidence from  and others (Cooke et al., 2014;Niknejad et al., 2008)

| Proteomic analysis of dACM and HSAM
To determine the concentrations of specific growth factors, 1-cm 2 samples from both HSAM (nine human tissue donors) and dACM (seven human tissue donors) were assessed utilizing a quantitative multiplex enzyme-linked immunosorbent assay (ELISA) proteomics microarray. Growth factors selected for this analysis were (a) previously identified at physiological levels within amniotic membranes and (b) identified to be relevant to tendon healing pathways. These growth factors included insulin-like growth factor I (IGF-I), acidic and basic fibroblast growth factors (aFGF and bFGF), platelet-derived growth factor-BB (PDGF-BB), and transforming growth factor beta-1 (TGF-β1; Molloy, Wang, & Murrell, 2003;Steed et al., 2008).
Growth factors released from either HSAM or dACM were evaluated by measuring the content of growth factors (

| Tenocyte proliferation assay
Cell proliferation assays were conducted with adult human tenocytes as described above.

| Tenocyte migration assays
The effect of HSAM or dACM on cell migration was assessed with a standard Boyden chamber assay as previously described in detail . Briefly, tenocytes were added to the top of the chamber at a concentration of 5,000 cells per insert and then incubated for 24 hr to allow for cell migration. Cell migration was measured in response to serum-free media (negative control), assay media + 10% fetal bovine serum (positive control), or CM from either HSAM or dACM at concentrations of 50%, 25%, and 10% (v/v). Images of the inserts were taken with an inverted microscope (Nikon Eclipse Ti, Tokyo, Japan), and migrated cells were counted.
Alternatively, tenocyte migration was evaluated using a standard scratch assay (Liang, Park, & Guan, 2007 prior to collection for reverse transcription-polymerase chain reaction.

| Gene expression and ELISAs
Tenocyte production of matrix metalloprotease-1 (MMP-1) and TGF-β1 in response to inflammatory cytokines was measured via  Table S1.

| Gene expression and ELISAs
Tenocytes were seeded at 40,000 tenocytes per well into six-well plates and cultured in assay media with or without 50% CM (dACM or HSAM) supplemented with 50 μM of ascorbic acid to allow for the deposition of collagen (Hakimi, Poulson, Thakkar, Yapp, & Carr, 2014). After 1 week of culture, the quantity of total collagen and sulfated glycosaminoglycans (sGAGs) was evaluated using the Sircol soluble collagen and Blyscan glycosaminoglycan assays, respectively (Biocolor, UK). Total ECM content was subsequently normalized to total DNA content, which was determined using a Quant-iT™ PicoGreen™ dsDNA Assay Kit (ThermoFisher, Waltham MA). Additionally, gene expression for relevant targets, including ECM-related genes, was evaluated in tenocytes cultured for 96 hr as detailed previously.

| Tenocyte culture with dACM and HSAM
To qualitatively analyze the interaction of tenocytes with HSAM and dACM grafts, 50,000 tenocytes were seeded onto dACM or HSAM (2.5 cm 2 , n = 3 from the same donor with HSAM and dACM originating from two distinct donors). Of note, grafts without tenocytes seeded were cultured concurrently and used for comparison. Grafts were then cultured in six-well plates under standard conditions and collected at 3, 7, 14, and 21 days for histological analysis.

| Histology
Postculture, grafts were fixed in a 4% solution of paraformaldehyde in phosphate-buffered saline for 24 hr. Samples were paraffin embedded, and 5-μm-thick serial sections were cut from tissue blocks. Care was taken to maintain the sectioning plane for all samples; however,

| Statistical analysis
Statistical analysis for the proteomics assays was conducted using an unpaired t test to compare protein levels between HSAM and dACM.
For all other quantitative assays, statistical analysis was conducted using a one-way analysis of variance with a post hoc Bonferroni test where p < .05 was considered significant. Unless otherwise specified, throughout this manuscript, * denotes p < .05, ** denotes p < .01, *** denotes p < .001, and **** denotes p < .0001.

| Proteomic analysis of dACM and HSAM
Proteomic microarrays confirmed physiologically relevant concentrations of growth factors related to tendon healing in both HSAM and dACM grafts (Figure 1a). Growth factors measured in this study include aFGF, bFGF, IGF-I, PDGF-BB, and TGF-β1, which are all upregulated during the natural tendon repair process (Dahlgren, Mohammed, & Nixon, 2005;Evans, 1999). Some sample-to-sample variability is known to be present in placental tissues, and the data Additionally, we hypothesized that processing of dACM may affect growth factor release from the membranes due to the compression and increased density of the ECM caused by removing the moisture from the graft. To evaluate this, the quantities of aFGF, bFGF, IGF-I, PDGF-BB, and TGF-β1 released from HSAM or dACM into assay media were quantified ( Figure 1b). Interestingly, there were no significant differences in the quantity of released growth factors measured between dACM and HSAM grafts. Taken together, these results show that although dACM grafts contain higher overall growth factor loads, the quantities of growth factors released are comparable between HSAM and dACM. In sum, both grafts contain growth factors necessary to support tendon healing.

| Tenocyte proliferation
In addition to studying specific growth factor concentrations and their release from the grafts, the effect of the released growth factors on important mechanisms of tendon healing was evaluated. Specifically, this included the effects of CM from either HSAM or dACM on tenocyte proliferation, migration, ECM deposition, and response to inflammatory stimulus. By Day 3, tenocyte proliferation was significantly upregulated in all CM groups following treatment with HSAM CM and dACM CM (p < .01) than in assay media controls (Figure 2a

| Tenocyte migration
The These results were further confirmed using a scratch assay, where a monolayer of tenocytes was scratched; repair of the scratch treated with assay media, dACM CM, or HSAM CM was quantified over 24 hr.
At 16 and 24 hr, HSAM CM resulted in a significant increase in closure than did assay media (p < .05, p < .01; Figure 3d). Additionally, at 24 hr, releasate from HSAM resulted in significantly greater percent closure than did dACM CM (p < .05). Representative bright-field images both illustrate percent closure and offer further explanation on how closure was measured (Figure 3e). Taken together with the findings of the Boyden chamber migration assay, CM from HSAM promoted more robust tenocyte migration than dACM CM. IGF-I, TGF-β, and FGF have been shown to support both proliferation and migration of tenocytes (Molloy et al., 2003).

| Collagen deposition and ECM expression
Reconstruction of the ECM following tendon injury is a critical phase during the repair process. The early production of ECM components both treatment groups compared with assay media, but these results were not statistically significant (Figure 4b). When collagen deposition was evaluated, tenocytes treated with dACM CM, but not HSAM, resulted in significantly higher levels of soluble collagen than did assay media controls (p < .05, Figure 4c).

| Tenocyte response to inflammatory cytokines
Although inflammation is a critical part of the healing process, it has been hypothesized that ineffective healing is often a result of a prolonged or excessive inflammatory phase (Thomopoulos, Parks, Rifkin, & Derwin, 2015) and that modulation of inflammation early on in the tendon repair process could result in improved outcomes (Hays et al., 2008). Therefore, we were interested in evaluating the overall response of tenocytes under inflammation and how that response changed with the addition of cytokines released from dACM or HSAM. For these studies, tenocytes were treated with either TNF-α or IL-1β for 96 hr. The rationale for using TNF-α and IL-1β in these experiments is due to their upregulation during the inflammatory response following tendon injury (Manning et al., 2014), and the concentrations used in these experiments were selected on the basis of prior studies (John et al., 2010;Tsuzaki et al., 2003). In control and TNF-α treatment groups, tenocytes treated with dACM CM had significantly decreased expression levels of TGFB1 than did assay media controls (p < .0001, Figure 5a). For HSAM groups at 1 ng/ml TNF-α, releasate from HSAM resulted in significantly decreased expression levels of TGFB1. In addition to gene expression, protein concentration was confirmed using ELISAs to measure protein production and release from the groups. With the use of ELISAs, a significant reduction in production of TGF-β1 over 96 hr using both HSAM CM and dACM CM was found (Figure 5b). In response to IL-1β stimulation, there were no significant differences in gene expression between groups under inflammatory stimulus ( Figure 5c, significantly reduced TGF-β1 in control group); however, both HSAM CM and dACM CM resulted in a reduction in TGF-β1 production in the 0.1 ng/ml IL-1β group (Figure 5d). Interestingly, HSAM CM also resulted in significant reduction in TGF-β1 production with 1 ng/ml IL-1β (Figure 5d, p < .001). Although some TGF-β1 is thought to be beneficial to tendon healing, a reduction in TGF-β1 production could be beneficial as studies have suggested that high levels of  Tenocyte response to inflammatory cytokines. Gene expression fold change of TGFB1 after stimulation with (a) TNF-α and (c) IL-1β normalized to GAPDH. Mean ± standard deviation reported; n = 12 for all groups. * denotes p < .05; • denotes p < .001; ◊ denotes p < .0001 compared with control. Protein production fold change of TGF-β1 after stimulation with (b) TNF-α and (d) IL-1β normalized to assay media. Mean ± standard deviation reported; n = 9 for all groups. ** denotes p < .01; • denotes p < .001; ◊ denotes p < .0001 compared with control. Gene expression fold change of MMP1 after stimulation with (e) TNF-α and (G) IL-1β normalized to GAPDH. Mean ± standard deviation reported; n = 12 for all groups. * denotes p < .05; • denotes p < .001; ◊ denotes p < .0001. Protein production fold change of MMP-1 after stimulation with (f) TNF-α and (h) IL-1β normalized to assay media. Mean ± standard deviation reported; n = 9 for all groups. * denotes p < .05; ** denotes p < .01; ◊ denotes p < .0001. CM, conditioned media; dACM, dehydrated amnion/chorion membrane; HSAM, hypothermically stored amniotic membrane; IL, interleukin; TGF-β1, transforming growth factor beta-1 FIGURE 6 Evaluation of tenocyte interaction with dACM and HSAM. Representative hematoxylin and eosin (H&E) images of (a) dACM and (b) HSAM seeded with tenocytes. Scale bars indicate 100 μm.
Additionally, a reduction of TGF-β1 during tendon healing has been shown to improve the range of motion of healed tendons (Xia et al., 2010) through reduced adhesion formation.
MMPs are responsible for degradation and reconstruction of the ECM. MMP-1 has a specific role in cleaving collagen fibers (Nagase, Visse, & Murphy, 2006) and plays an important role during wound healing, specifically during re-epithelialization (Gill & Parks, 2008); however, overexpression of MMPs can lead to excessive breakdown of the ECM. Additionally, production and release of MMPs are known to be stimulated by inflammatory molecules (John et al., 2010;Tsuzaki et al., 2003). After 96 hr of treatment with dACM CM and TNF-α (all concentrations), MMP1 expression and protein production were significantly upregulated compared with assay media (Figure 5e,f).
Conversely, following inflammatory stimulation, ELISAs showed significantly decreased levels of MMP-1 production (with HSAM CM) than did assay media in both 1 and 0.1 ng/ml TNF-α groups

| Tenocyte culture with dACM and HSAM
In addition to levels of growth factors, and how those released factors interact with tenocytes, the interactions between tenocytes and the ECM of these grafts were evaluated. To do this, tenocytes were seeded onto either dACM or HSAM grafts for up to 4 weeks ( Figure 6a,b). H&E staining revealed significant tenocyte attachment and migration into the HSAM graft (Figure 6b), whereas tenocytes cultured on dACM grafts attached but did not migrate into the graft (remained in a monolayer atop the graft, Figure 6a). To further evaluate how tenocytes interacted with HSAM grafts, a time course study was completed evaluating tenocytes seeded at 3, 7, 14, and 21 days on HSAM grafts ( Figure S1). These results demonstrated how tenocytes were able to migrate into and remodel the HSAM graft over time. Additional staining at 28 days also revealed that tenocyte culture on HSAM resulted in increased collagen production as determined by Masson's trichrome staining (Figure 6e,f), increases in proteoglycan production as determined by Alcian Blue (Figure 6g,h), and elastin production as determined by Verhoeff's stain (Figure 6i,j). Additionally, remodeling of HSAM and deposition of molecules important for tendon repair (collagen I/III, tenomodulin, and decorin) were further evaluated using IHC staining ( Figure S2). Images A, C, E, and G are control samples (HSAM with no tenocytes), whereas images B, D, F, and H are seeded samples (HSAM seeded with tenocytes). HSAM that was seeded with tenocytes qualitatively showed increased IHC staining for collagen III, tenomodulin, and decorin. We hypothesize the differences in tenocyte interaction between dACM and HSAM grafts is at least in part due to the preservation methods; hypothermic storage of the HSAM graft allows for maintenance and preservation of the ECM structure, which provides an open scaffold for the tenocytes to migrate through. In contrast, dehydration results in the removal of moisture, which in turn has been reported to result in a compressed ECM structure (McQuilling, Vines, Kimmerling, & Mowry, 2017). This dense structure and lack of pores, although structurally beneficial for adhesion barrier applications, likely prevents tenocytes from invading and migrating through the dACM graft.

| DISCUSSION
In this work, we have studied two different grafts to evaluate whether differences in composition and processing would result in varying responses. A summary of high-level study findings has been provided in Table 1. In this study, dACM and HSAM grafts were evaluated for total growth factor content and growth factor release from the membranes. In addition to verifying the presence and release of these factors, their ability to support key processes for tendon repair was evaluated, including proliferation and migration of tenocytes, expression and deposition of key ECM molecules, tenocyte responses to inflammatory signals, and the extent of tenocyte interactions with the ECM.
Previous studies on tendon healing have demonstrated that several key growth factors are required to promote cellular proliferation and tissue repair at the injury site (Molloy et al., 2003;Nourissat et al., 2015). These include aFGF, bFGF, IGF-I, PDGF-BB, and TGF-β1, all of which were evaluated in this study. Various studies have demonstrated that these growth factors, when provided individually following an injury, can expedite the repair process (Fukui, Katsuragawa, Sakai, Oda, & Nakamura, 1998;Kobayashi, Kurosaka, Yoshiya, & Mizuno, 1997;Kurtz, Loebig, Anderson, DeMeo, & Campbell, 1999) and potentially improve the quality of the outcome (Chan et al., 2000;Hildebrand et al., 1998;Kurtz et al., 1999;Letson & Dahners, 1994). Both dACM and HSAM grafts had physiologically relevant concentrations of all five growth factors evaluated, but dACM contained significantly more bFGF and IGF-I. Interestingly, when further evaluating the release of growth factors over time by evaluating CM, these differences were no longer discernable for any growth factors measured. We hypothesize that this is due to the processing methodology differences; HSAM is a fresh, aseptically processed graft with minimal processing steps that is stored at hypothermic temperatures, whereas dACM has been dehydrated and terminally sterilized, resulting in significant perturbations to the ECM structure.
Following assessment of important factors in both grafts, the ability of these released growth factors to support tendon repair mechanisms was evaluated in vitro. We observed that growth factors released from both HSAM and dACM triggered an increase in tenocyte proliferation; however, dACM CM significantly increased tenocyte proliferation than did HSAM conditions. When comparing HSAM with dACM, the reasons for increases in proliferation are not obvious because there are no significant differences in the levels of mediators found in CM measured in this study (Figure 1b). Although not significant, the trends for higher bFGF in dACM may be responsible for this finding. It has been reported that the combined presence of IGF-I, PDGF-BB, and bFGF has a synergistic effect on tenocyte proliferation (Costa et al., 2006), pointing to the potential value of a multifaceted treatment modality. Alternatively, this observed difference may be due to increases in other factors found within dACM or HSAM, which were not evaluated in this study.
Although proliferation of tenocytes is important, migration of tenocytes to the injury site is also essential for repair. Therefore, the ability of each graft to facilitate and enhance tenocyte migration was  (Subach & Copay, 2015).
Tenocytes cultured in CM from either dACM or HSAM exhibited blunted TGF-β1 production following TNF-α and IL-1β stimulation, a finding that can be considered beneficial to the repair process as excessive TGF-β1 is associated with scarring and adhesions (Chang, Thunder, Most, Longaker, & Lineaweaver, 2000;Farhat et al., 2015;Katzel et al., 2011;Xia et al., 2010). Previous work has demonstrated that too much TGF-β1 can result in adhesion formation, and reduction of the available TGF-β1 can improve the range of motion of the healed tendon (Xia et al., 2010).
repairs resulted in increased tensile strength at 1 and 2 weeks (Zhang et al., 2003).
One limitation of this work is the limited proteomic evaluation of these grafts. The analysis within this study consisted of five growth factors known to be relevant to tendon repair; however, these factors are only a fraction of the growth factors known to be present within placental tissues. Furthermore, in vitro work with tenocytes is limited by both the lack of complexity of cellular phenotypes found in vivo and the tendency for phenotypic drift in tenocyte cultures (Yao, Bestwick, Bestwick, Maffulli, & Aspden, 2006

| CONCLUSION
This research is the first to evaluate multiple types of commercially available placental-derived grafts as an adjunct to tendon repair.
In sum, we found that both dACM and HSAM contain and release growth factors relevant to tendon repair. Interestingly, although both grafts promoted tendon repair mechanisms, there were clear differences in how dACM and HSAM affected tenocytes in vitro. These differences are interesting and may be a result of the layers included (amnion vs. amnion/chorion), processing technique (aseptic, hypothermic storage vs. terminally sterilized, dehydrated), or some combination of these factors. Although this study illustrates several potential ways in which placental-derived membranes may support tendon repair, further in vivo work will be necessary to better understand these results.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. List of TaqMan Probes (ThermoFisher Scientific).