ARROW@TU Dublin ARROW@TU Dublin A Step Closer to Elastogenesis on Demand; Inducing Mature A Step Closer to Elastogenesis on Demand; Inducing Mature Elastic Fibre Deposition in a Natural Biomaterial Scaffold Elastic Fibre Deposition in a Natural Biomaterial Scaffold

Elastic fibres play a key role in bodily functions where fatigue resistance and elastic recovery are necessary while regulating phenotype, proliferation and migration in cells. While in vivo elastic fibres are created at a late foetal stage, a major obstacle in the development of engineered tissue is that human vascular smooth muscle cells (hVSMCs), one of the principal elastogenic cells, are unable to spontaneously promote elastogenesis in vitro . Therefore, the overall aim of this study was to activate elastogenesis in vitro by hVSMCs seeded in fibrin, collagen, glycosaminoglycan (FCG) scaffolds, following the addition of recombinant human tropoelastin. This combination of scaffold, tropoelastin and cells induced the deposition of elastin and formation of lamellar maturing elastic fibres, similar to those found in skin, blood vessels and heart valves. Furthermore, higher numbers of maturing branched elastic fibres were synthesised when a higher cell density was used and by drop- loading tropoelastin onto cell-seeded FCG scaffolds prior to adding growth medium. The addition of tropoelastin showed no effect on cell proliferation or mechanical properties of the scaffold which remained dimensionally stable throughout. With these results, we have established a natural biomaterial scaffold that can undergo controlled elastogenesis on demand, suitable for tissue engineering applications.


Introduction
Fatigue resistance and elastic recovery are critically important mechanical properties for the correct functioning of the heart, skin, lungs and other dynamic connective tissues within the human body. Elastic fibres play a key role in bodily functions where these mechanical properties are necessary [1]. Indeed, elastin is one of the most abundant proteins found in vertebrates, providing resistance to permanent tissue deformation [2] and delivering biochemical cues that regulate cell migration, preventing proliferative diseases in heart and blood vessels [3,4]. In a mature elastic fibre, approximately 90% of the volume is composed of elastin that is surrounded by a fibrillin-rich microfibrillar mesh [5]. However, the content, alignment and morphology of these mature elastic fibres is dictated by functional demands and the cells that deposit and orientate the fibrillin-rich microfibrils. For instance, elastic ligaments have approximately 50% elastin content [6] with elastic fibres that resemble rope-like networks giving joints stability and flexibility. Blood vessels have a 30% elastin content, and the elastic fibres can be observed as concentric fenestrated lamellae required for stress propagation during each cardiac cycle. Having 15% of elastin content, heart valves show a mixed architecture of elastic fibres; while the ventricularis layer of the leaflets has elastic fibres aligned radially to withstand the blood inflow [7], reticular elastic fibres are present throughout the heart valve to reduce permanent deformation [8]. Lungs on the other hand (3-7% elastin), present elastic fibres that are highly branched along the respiratory tree and facilitate alveoli expansion. Finally, skin, which contains approximately 2% of elastin [9], has elastic fibres that are thick and horizontally arranged in the dermis, whereas in the papillary dermis, these fibres are perpendicularly oriented [10]. While many biocompatible materials have been developed for tissue engineering applications, to date they are unable to fully mimic the function of native tissue, as they do not encourage the production of new elastic fibres through elastogenesis [11].
Elastogenesis is a hierarchical elastic fibre assembly process that is highly upregulated at late foetal and neonatal stage [12]. Tropoelastin, the soluble precursor of elastin, is essential for the assembly of elastic fibres [6]. In a normal elastogenesis process, cells produce tropoelastin that is chaperoned by elastin-binding proteins (EBP) to prevent premature coacervation into elastin [13]. When tropoelastin leaves the cellular membrane, it coacervates into elastin chains within the extracellular matrix (ECM). Elastin then begins to crosslink using lysyl oxidase (LOX1), copper, fibulin 4 and 5 to make insoluble elastin that is finally deposited in the microfibrillar mesh [12], allowing elastic fibres to form with a lifespan of approximately 70 years [14,15]. From the estimated 200 cell types found in the human body [16], only smooth muscle cells, fibroblasts, endothelial cells, chondroblasts and mesothelial cells are capable of assembling elastic fibres in vivo [17]. However, a major issue for researchers is that these cells are unable to spontaneously undergo elastogenesis in vitro.
Indeed, much research in recent years has focused on finding ways to induce elastogenesis in cell culture using exogenous agents [18]. In vitro, glycosaminoglycans (GAGs) such as heparin [19], hormones, growth factors [20][21][22][23] and gene expression therapy [24] can induce production of tropoelastin in cells and upregulation of microfibrillar associated genes, leading to the formation of elastic fibres in fibroblasts and smooth muscle cells (SMCs). Another way of obtaining elastic fibres in vitro is by adding a solution of recombinant human tropoelastin that can be incorporated into the ECM, thereby skipping the tropoelastin synthesis phase inside the cell [12,17,25,26]. Additionally, it has been welldocumented that similarly to natural tropoelastin, recombinant human tropoelastin can modulate the assembly of elastic fibres in the ECM, while promoting attachment of cells and regulating their proliferation [1,[27][28][29][30]. Recently, Mithieux et al. presented ways of obtaining elastic fibres in silicone and collagen constructs. This was achieved through repeated additions of recombinant synthetic human tropoelastin and conditioned growth medium, using dermal fibroblasts of different age groups [31]. As a result of these findings, there is a need to examine whether this technique can be translated into other cell-seeded biomaterial constructs. Additionally, the effect of the method and timing of recombinant human tropoelastin addition, on the deposition of elastic fibres, warrants examination.
Our research group is specialised in the fabrication of natural polymer based porous scaffolds, for several medical applications including cartilage [32,33], bone [34], nerve [35], and vascular [36] repair. We have developed a platform for soft tissue engineering that can resist cellmediated contraction using fibrin gels reinforced with a freeze-dried collagen-glycosaminoglycan matrix [37]. Although these fibrincollagen-glycosaminoglycan (FCG) scaffolds have proven to be dimensionally stable in cell culture when seeded with smooth muscle cells, the next step is to develop a repeatable method to increase and tune the content of elastic fibres when cultured with appropriate cells. This milestone will take us a step closer to meeting the unmet need of a fully functional natural biomaterial scaffold, able to undergo elastogenesis on demand, with increased elastic recoil and fatigue life. Such a biomaterial may prove suitable for a variety of applications.
Therefore, the overall aim of this study was to activate elastogenesis by human vascular smooth muscle cells (hVSMCs) seeded in the FCG scaffolds. We established the influence of the timing of tropoelastin addition on the maturing elastic fibre deposition, cell proliferation and gene expression of hVSMCs seeded on glass slides. Next, we assessed the effect of changing the method of addition of tropoelastin on the deposition and area coverage of elastin and maturing elastic fibres in FCG scaffolds seeded with hVSMCs. Finally, we studied the effect of controlled elastogenesis on the cell proliferation, mechanical and dimensional properties and elastogenesis associated gene expression in FCG scaffolds seeded with hVSMCs.

Tropoelastin solution preparation
For these experiments, recombinant synthetic human tropoelastin without hydrophilic domain 26A (SHELΔ26A) (amino acid residues 27-724 of GenBank entry AAC98394). The production process involved overexpression of tropoelastin in Escherichia coli followed by purification and this has been previously described [38,39]. The resultant lyophilised tropoelastin was dissolved overnight in phosphate buffer saline (PBS, Sigma-Aldrich), to obtain a 10 mg/ml solution.

2D elastogenesis model
To assess the optimal timing of tropoelastin addition in order to enhance the deposition of more elastic fibres, two groups were studied. Samples were treated with tropoelastin on day 0 (A0) and further samples were treated on day 10 (A10), as the rationale in this model is that a mature ECM would enhance the deposition of elastin and maturing elastic fibres. Therefore, hVSMCs were seeded in 12-well plates containing glass cover slips at a density of approximately 50,000 cells per well. 50 μl of 10 mg/ml tropoelastin was added to all the samples on their designated day. After adding growth medium, the final concentration of tropoelastin per well was 250 μg/ml. Growth medium changes were carried out on days 3 and 5 of the experiment. Samples were collected on days 1, 3 and 7 post-tropoelastin addition for cell number quantification, real-time quantitative PCR (RT-qPCR), immunofluorescence (anti-elastin) and maturing elastic fibre autofluorescence at 561 nm [31].

Elastin and elastic fibre area coverage estimation
Immunofluorescence was carried out to visually assess for the presence of elastin and maturing elastic fibres. Samples were fixated with 10% formalin for 15 min, washed twice in PBS and three times with 0.2 M Glycine (Sigma-Aldrich). Samples were incubated for 1 h at 37 • C in 5% w/v bovine serum albumin (Sigma-Aldrich), kept in the fridge overnight with 1:100 BA4 anti-elastin antibody (Abcam, Cambridge, UK), incubated with 1:500 Alexa Fluor® 488 (Abcam) for 1 h at 37 • C. Glass microscope slides and Fluoroshield with DAPI (Sigma-Aldrich) were used to mount the 2D samples while 3D samples were kept in glycine. Elastin was observed at 488 nm (green channel) and maturing elastic fibres (covalently cross-linked elastin) were visualised at a wavelength of 561 nm (red channel) using a confocal microscope, due to its autofluorescence. The following calculations were carried out using the image processing package, Fiji [40]. Relative area coverage was carried out by thresholding the elastin and maturing elastic fibres channels and quantifying the intensity of each pixel (n = 4). Maturing elastic fibre alignment was assessed by smoothing the respective channel and analysing it using the directionality function, which estimates the preferred orientation of clusters of pixels in the image using the Fourier spectrum analysis method.

Cell number quantification
To investigate the influence of tropoelastin on the proliferation of hVSMCs, a Quant-iT PicoGreen dsDNA kit (BD BioSciences, Oxford, UK) was used. This kit can detect double-stranded DNA (dsDNA) through fluorescence imaging. Samples (n = 3 for each time point) were washed in PBS and kept at − 20 • C in a solution of 0.2 M NaCO 3 and 1% Triton X (Sigma-Aldrich). Each sample was diluted 25× with 1xTE Buffer, pipetted into a 96-well plate and treated with a 200-fold PicoGreen dilution. Using a TECAN plate reader, samples were excited at 485 nm and read at 538 nm. dsDNA concentration was calculated by comparing the fluorescence levels from each sample against a standard fluorescence curve [41].

Method of tropoelastin addition to FCG scaffolds
Once the cells had been proliferating in the scaffolds for 7 days, we assessed the effect that the method of tropoelastin addition had on the formation of elastin and elastic fibres. We removed the growth medium from the wells and used two methods to add tropoelastin to the FCG scaffolds: 1) drop-loading 50 μl of 10 mg/ml tropoelastin onto the scaffolds, waiting 10 min before adding fresh growth medium and 2) adding growth medium which already contained 50 μl of 10 mg/ml of tropoelastin. In both cases, the final concentration of tropoelastin per well was 250 μg/ml. It is worth mentioning that concentration dependency studies were not carried out using our scaffolds, therefore the concentration of tropoelastin chosen for this study was based on a previous recombinant human tropoelastin publication [31]. Growth medium changes were carried out every 3 days. Seven days post initial tropoelastin addition, a second round of tropoelastin was added onto the FCG scaffolds using the two methods described above. Samples were collected on days, 3, 7 and 14 (post initial tropoelastin addition) for dsDNA quantification, RT-qPCR, immunofluorescence (anti-elastin) and autofluorescence at 561 nm, as described in previous paragraphs (n = 3 for all assays).

Quantification of elastin deposited within the FCG scaffolds
The Fastin™ Elastin Assay kit (Biocolor Ltd., Carrickfergus, UK) was used to quantify the amount of elastin deposited in FCG scaffolds. Following elastin solubilisation, 500 μl from each sample was precipitated, centrifuged and mixed with dye reagent. The resultant elastin-dye complex was centrifuged and the pellet solubilised using the dye dissociation agent. Samples were transferred to a clear 96-well plate and absorbance at 513 nm was measured using the plate reader. Observations were normalised to the corresponding blanks (0.25 M oxalic acid) and expressed in μg of elastin per 100 mg of wet scaffold.

Mechanical characterisation of FCG scaffolds
FCG scaffolds (n = 3) soaked in PBS were compressed uniaxially, perpendicular to their flat face, to 10% strain at 0.1% strain per second. The compressive modulus was obtained by calculating the slope of the stress-strain curves [47]. Dog bone-shaped FCG scaffolds (n = 3, only day 14 samples due to cell numbers involved), as per ASTM D638, were hydrated in PBS and tested using a tensile preload of 0.01 N at a tension rate of 5 mm/min until failure. Tests were carried out using a universal testing machine (Zwick/Roell) fitted with a 5 N load cell. Ultimate tensile strength, strain to failure, low strain (2-5%) and high strain (10-15%) tensile modulus were calculated from the stress-strain curves generated [48]. To assess the effect of the addition of tropoelastin on the dimensional integrity of FCG scaffolds, they (n = 3) were photographed using a digital camera, and their diameter was measured using the image processing package, Fiji [40].

Statistical analyses
Results are expressed as mean ± standard deviation. Two-way analysis of variance (ANOVA) and Bonferroni post-test were carried out to analyse dsDNA concentration, gene relative fold change, elastin quantification, elastin coverage area, maturing elastic fibre coverage area, compressive modulus and dimensional integrity. t-test and Mann-Whitney U test was carried out to assess tensile modulus results. All results were compared against untreated samples (no tropoelastin added) and P < 0.05 was defined as statistically significant.

Mature ECM enhanced elastogenesis
Samples on glass slides were treated with tropoelastin on day 0 (A0) and further samples were treated on day 10 (A10) based on the hypothesis that a higher density of cells may result in more ECM synthesis and this would enhance the deposition of elastin and maturing elastic fibres. Indeed, elastin and maturing elastic fibre deposition was visible on day 3 in A10 samples (Fig. 1E), with a steady and significant deposition of elastin (Fig. 1F, G) and maturing elastic fibres (Fig. 1J, K) on days 7 and 14. A0 samples on the other hand, showed a small amount of elastin on day 3, however no elastin or maturing elastic fibres were observed at any other time point. As expected, samples with no tropoelastin added, showed no elastin or maturing elastic fibres during the experiment (Fig. 1A-D, H).
To quantify the elastin and maturing elastic fibres observed, measurements of the relative area coverage of both were carried out using image analysis software. As expected, A10 samples had a significantly higher content of elastin and maturing elastic fibres than A0 samples (Fig. 2). On day 3, 13% of the surface area of the A10 samples was covered by elastin (p < 0.001) and by day 14, this coverage increased to almost 20% (p < 0.001). When the coverage of maturing elastic fibres was measured on day 3, there was less than 1% coverage and by day 14 the coverage was approximately 8% (p < 0.001). In addition, assessment of fibre alignment in A10 samples showed a preferred orientation of the maturing elastic fibres as indicated by the directionality curves (Fig. 2C). We have proven that a mature ECM increased significantly the deposition of elastin and maturing elastic fibres in hVSMCs seeded on glass slides.

Tropoelastin regulated cell proliferation
The effect of tropoelastin on the proliferation of hVSMCs was measured using a PicoGreen assay over a 1-week period. Untreated samples presented a higher cell proliferation rate (p < 0.001) when compared to A0 and A10 samples (Fig. 3). On day 1, A0 samples and A10 samples had a mean dsDNA concentration of 242 ± 53 ng/ml and 297 ± 36 ng/ml respectively, while the untreated samples were measured at 219 ± 46 ng/ml. Within three days of treatment, dsDNA concentration remained constant; 247 ± 0.3 ng/ml in A0 samples, 315 ± 117 ng/ml in A10 samples and 228 ± 09 ng/ml in untreated samples. Untreated samples showed a significantly higher dsDNA concentration (658 ± 117 ng/ml) than A0 samples (469 ± 27 ng/ml) and A10 samples (332 ± 13 ng/ml) after seven days. These results demonstrate that tropoelastin regulates the proliferation of hVSMCs seeded on glass slides regardless of cell density levels.
On the other hand, in A10 samples on day 1, there was downregulation of ELN (0.3 ± 0.3), LOX1 (0.5 ± 0.1), EBP (0.8 ± 0.6) and FBN1 (0.6 ± 0.3). Three days later upregulation of FBN1 (1.3 ± 2) occurred. We could see that on days 3 and 7, ELN and LOX remained downregulated. Furthermore, a relative fold change of 1.7 ± 1 in FBN1 and a 3-fold increase in EBP (p < 0.001) was observed after seven days (Fig. 4C). As there is no significant production of elastin or associated proteins (except EBP in A10 samples) by cells, we have demonstrated that elastogenesis was induced solely by adding tropoelastin on hVSMCs.

Enhanced elastin deposition in FCG scaffold groups treated with tropoelastin
Having shown that higher density of cells resulted in a higher deposition of maturing elastic fibres (Fig. 2), we used this result when designing our 3D studies where we examined the effect of the method of addition of tropoelastin. The deposition of tropoelastin in FCG scaffolds was quantified using the Fastin™ Elastin Assay kit. On day 3, drop- in the media-loaded samples. Although there was a modest 10% reduction in elastin concentration per week, this drop in elastin was not significant. In contrast, significantly higher amounts of elastin were deposited in the groups treated with tropoelastin (p < 0.001) when compared against untreated samples (Fig. 5). These results demonstrate that tropoelastin can be deposited into cell-seeded FCG scaffolds and maintained without substantial losses.

Drop-loading tropoelastin improved the surface area coverage of maturing elastic fibres in FCG scaffolds
After quantifying the deposition of tropoelastin in FCG scaffolds as a function of the method of addition, we followed with the visual assessment of the surface area coverage of elastin and maturing elastic fibres in these scaffolds. Although the amount of elastin and maturing elastic fibres looked similar in drop-loaded and media-loaded samples (Fig. 6), there was a 4-fold increase in elastin (p < 0.001) and a 2-fold increase in maturing elastic fibre surface coverage (p < 0.001) in drop-loaded samples by day 14 (p < 0.001). Three days after the addition of tropoelastin, elastin was strikingly visible in both drop-loaded (Fig. 6A) and media-loaded (Fig. 6E) samples, appearing in both as elastin globules. However, by day 7 an unexpected decrease in this elastin content was observed (Fig. 6B, F). Interestingly, by day 14, maturing elastic fibres were visible in drop-loaded and media-loaded samples (Fig. 6B, F). At each time point cells were evenly distributed and not clumped. Acellular samples (Fig. 6H) and samples free from tropoelastin (Fig. 6D), did not present elastin or maturing elastic fibres.
Relative area coverage of elastin and maturing elastic fibres was quantified on the surface of the scaffolds, being significantly higher in the drop-loaded samples by the end of the experiment. Elastin coverage significantly increased in both drop-loaded (30 ± 9%) and media-loaded (24 ± 7%) samples on day 3 (p < 0.001), as can be seen in Fig. 7A. Day 7, on the other hand, showed a drop, to approximately 2% in the elastin coverage area regardless of the method of addition. A second addition of tropoelastin increased the elastin coverage by day 14 up to 3 ± 2% in media-loaded (Fig. 7C) and 11 ± 6% (p < 0.001) in drop-loaded (Fig. 7D) samples. On day 3, the content of maturing elastic fibres in both methods of addition of tropoelastin was approximately 2% (droploaded 2.6 ± 0.8, media-loaded 2 ± 1.7%). This content of maturing elastic fibres decreased to 1.2 ± 0.7% in drop-loaded and 1.7 ± 1.4% in media-loaded samples on day 7. After the second addition of tropoelastin, drop-loaded and media-loaded samples treated with tropoelastin presented a significantly higher quantity of maturing elastic fibres (12.3 ± 4.5% and 5.9 ± 2.4% respectively) in comparison to untreated samples (0.8 ± 1%) by day 14 (Fig. 7B). Comparing addition methods, the drop-loaded samples presented a higher coverage of elastin (p < 0.001) and maturing elastic fibres (p < 0.001) than the media-loaded samples on day 14. Additionally, after assessing the maturing elastic fibre alignment both groups treated with tropoelastin presented a disordered fibre arrangement as the directionality curves were relatively flat (Fig. 7C, D), if compared with the curves from the 2D elastogenesis model (Fig. 2C). With these results we can state that drop-loading tropoelastin directly onto cell-seeded FCG scaffolds before adding cell growth medium resulted in a better coverage of elastin and maturing elastic fibres on the scaffold surface.

Consistent cell proliferation in FCG scaffolds treated with tropoelastin
Having shown that drop-loading tropoelastin resulted in a better superficial distribution of elastin and maturing elastic fibres, cell proliferation, gene expression and mechanical testing assays were carried out only on drop-loaded samples. When measured, the rate of cell proliferation proved to be constant between untreated samples and those treated with tropoelastin, across all time points (Fig. 8). Three days after the addition of tropoelastin, the average dsDNA concentration was 175 ± 10 ng/ml for the untreated samples and 213 ± 8 ng/ml for the tropoelastin-treated samples. Untreated samples had a mean dsDNA concentration of 225 ± 10 ng/ml and the tropoelastin-treated samples 192 ± 20 ng/ml by day 7. After 14 days, the concentration of dsDNA remained consistent in both untreated (198 ± 15 ng/ml) and tropoelastin-treated (186 ± 26 ng/ml) groups.

Tropoelastin-treated FCG scaffolds showed elastin gene expression
RT-qPCR analysis presented in Fig. 9 showed that there was significant ELN upregulation (1.7 ± 0.7) on day 3. The addition of tropoelastin resulted in upregulation of FBN1 (1.5 ± 0.6) on day 7. Overall, there was stable upregulation of ELN and EBP on days 7 and 14. Furthermore, LOX1 did not change at the time points studied. Therefore, similarly to the 2D studies (Fig. 4), these 3D results indicate that elastogenesis was triggered after the addition of tropoelastin.

No detrimental changes in mechanical properties and stable dimensional integrity of tropoelastin-treated FCG scaffolds
Our next step was to analyse how the mechanical properties and dimensional integrity of the FCG scaffolds were influenced by the presence of elastin and maturing elastic fibres. Based on the data shown in Table 1, fourteen days after the addition of tropoelastin, the compressive and tensile modulus, ultimate tensile strength (UTS) and strain to failure of the samples treated with tropoelastin remained comparable to those of untreated samples.
Although no significant differences were observed in the stress-strain curves (Fig. 10A, B), in terms of compressive modulus (Fig. 10C), on day 14 the modulus of treated samples was lower (0.16 ± 0.02 kPa) than untreated samples (0.19 ± 0.1 kPa). At low strain (2-5% strain), by day 14 the average tensile modulus of tropoelastin-treated samples was 1.6 ± 0.3 kPa, when compared to untreated samples (1.8 ± 0.4 kPa). At high strain (10-15% strain), the modulus of the tropoelastin-treated samples was 5.8 ± 0.1 kPa (Fig. 10D). Additionally, scaffold tensile properties were examined with representative stress-strain curves. A small increase in treated samples UTS (64.7 ± 8.3 kPa) and strain to failure (17.1 ± 1.3%) was observed when compared against untreated samples. Furthermore, the standard deviations of the tropoelastin-treated samples were smaller than the untreated group. Finally, cell-mediated contraction was studied during two weeks and only a 2% decrease in diameter was observed in the tropoelastin-treated scaffolds, which is comparable to the untreated samples (Fig. 11A, B). These assessments have confirmed that the addition of tropoelastin did not affect negatively the mechanical properties of the FCG scaffolds and in addition to this, the scaffolds were capable of resisting cell-mediated contraction.

Discussion
Encouraging and controlling the production of new elastic fibres through elastogenesis is highly desirable in regenerative medicine, as elastic fibres prevent permanent tissue deformation while regulating phenotype, proliferation and migration in cells [3,30]. Therefore, the overall aim of this study was to activate elastogenesis in human vascular smooth muscle cells (hVSMCs) seeded in fibrin collagen glycosaminoglycan (FCG) scaffolds, by adding a solution of recombinant human tropoelastin. In this manuscript, we present a repeatable method to increase and tune elastogenesis by hVSMCs seeded in FCG scaffolds. hVSMCs were chosen as they are one of the principal elastogenic cells in the human body [49] however, hVSMCs are unable to produce elastin in vitro as they age when their passage increases [2]. Although hVSMCs and recombinant human tropoelastin have been previously used together in synthetic scaffolds [50], their combined elastogenic potential has never previously been examined. Additionally, unlike previous elastogenesis experiments involving SMCs [28,51], we have used a natural biomaterial scaffold and static cell culture. Here, elastogenesis was activated after the addition of tropoelastin and furthermore, the importance of a mature ECM to obtain enhanced deposition of elastin and maturing elastic fibres in tropoelastin-treated hVSMCs was shown. Moreover, we have demonstrated that the drop-loading of tropoelastin, resulted in maturing elastic fibre surface coverage, without affecting the cell   proliferation in these scaffolds. After assessing the best day for the addition of tropoelastin on hVSMCs seeded on glass slides, we could prove that ECM maturity affects the activation of elastogenesis. Waiting 10 days before adding tropoelastin to hVSMCs (A10 samples), led to a significant increase in the content of elastin and maturing elastic fibres than hVSMCs treated with tropoelastin directly after seeding (A0 samples). Furthermore, A10 samples presented globules of elastin showing bead-like patterns on day 3 (Fig. 1E); this type of pattern has been shown by Kozel et al. to be characteristic of the first stages of elastogenesis [52]. By day 7 (Fig. 1F), the coacervation of tropoelastin was obvious and the ECM was saturated in elastin. On day 14 the maturing elastic fibre alignment resembled lamellar-like structure very similar to the ones of the tunica media in arteries [5,53], and ventricularis layer in the leaflets of aortic heart valves [54]. To prove this point, a directionality analysis was carried out (Fig. 2C), showing that the maturing elastic fibres were aligned. Although this was undoubtedly a feature of the hVSMCs used here, it was nonetheless encouraging. From these results, it can be inferred that the limiting factor in elastin deposition may be the content of ECMassociated proteins (fibulins [55,56], fibrillins [57], and microfibrillar associate proteins [58]). We speculate that in the A0 samples, as tropoelastin was added after cell seeding, the microfibrillar mesh in the ECM was not fully formed, being unable to accommodate all the coacervated elastin, resulting in the loss of elastin during each growth  medium change. This has also been observed by Kozel et al., showing that elastin is unable to transform into an elastic fibre if there is no presence of ECM [59]. Additionally, looking at the RT-qPCR results (Fig. 4) all the elastin and maturing elastic fibres present in the ECM were as a result of introducing the tropoelastin. We can assume this as the levels of elastin (ELN) and LOX1 remained downregulated, whereas Fibrillin-1 (FBN1) and EBP were upregulated. It is known that FBN1 is of one of the elastic-fibre associated proteins in the microfibrillar mesh responsible for signalling the assembly of mature elastic fibres in tissues [57]. Therefore, the upregulation of FBN1 on day 7 shown in Fig. 4 makes sense, as there was a corresponding increase in the deposition of maturing elastic fibres.
Having successfully used tropoelastin to induce elastogenesis in hVSMCs on glass slides, we translated these results into a 3D model using FCG scaffolds. After quantifying the amount of elastin deposited in drop-loaded and media-loaded samples, we observed that the deposition of elastin within the cell-seeded scaffolds occurred during the first 3 days of this study and no significant changes were observed after the second addition of tropoelastin on day 7 (Fig. 5). Additionally, while there was a reduction in elastin content during the experiment, the rate of loss was not significant. Furthermore, there were no significant differences in elastin content between drop-loaded and media-loaded samples. On the other hand, after visually assessing the scaffolds, we obtained twice the surface area coverage of maturing elastic fibres by drop-loading tropoelastin onto the cell-seeded FCG scaffolds, outperforming samples where the tropoelastin was diluted in growth medium (media-loaded). It is presumed that by drop-loading tropoelastin, the distribution of elastin and elastic fibres on the scaffold surface was higher as the tropoelastin solution was more concentrated when first in contact (for 10 min) with the hVSMCs. Initially, the concentration of tropoelastin was 10 mg/ml and when growth medium was added this reduced to 250 μg/ml in the wells. However, by day 7 there was a reduction in elastin and no visible maturing elastic fibres (Fig. 6B, F). This indicated that there was not enough ECM secreted by the hVSMCs to allow deposition of elastin on the surface of the FCG scaffolds. Interestingly, despite this, a second addition of tropoelastin resulted in an increase in elastin and maturing elastic fibre surface area coverage (Fig. 6C, G). By the end of the experiment the surface of the FCG scaffolds presented maturing branching elastic fibres along the scaffold pores, resembling the elastic fibre architecture present in lungs [10]. Fibre alignment was corroborated with a directionality analysis (Fig. 7C, D) and we saw that the maturing elastic fibres in both drop-loaded and media-loaded samples had no preferred orientation. These results imply that the hVSMCs had by this point secreted enough ECM-associated proteins to support elastin deposition and crosslinking on the surface of the FCG scaffold. To the best of our knowledge this is the first study presenting the drop-loading of tropoelastin on natural biomaterial cell-seeded scaffolds to induce and control elastogenesis in vitro.
Having shown that controlled elastogenesis was possible by droploading tropoelastin, we carried out compressive and uniaxial tensile testing on tropoelastin-treated FCG scaffolds and there were no detrimental changes in mechanical properties. We could see that the compressive modulus was lower in the samples treated with tropoelastin (Fig. 10C) and in terms of tensile properties, when collagen and elastic fibres began to straighten at low strain this resulted in a reduction of tensile modulus (Fig. 10D). Similar results were observed at high strain, presumably due to the elastic fibres transferring energy to the collagen fibres to prevent overstretching [48]. Additionally, there was also an increase in ultimate tensile strength and strain to failure (Table 1). It is also noteworthy that after two weeks in vitro, FCG scaffolds were dimensionally stable (Fig. 11B). This is an important attribute, as many biomaterial constructs need to be capable of maintaining their shape post-implantation.
Although the addition of tropoelastin resulted in an efficient way to induce the formation of maturing elastic fibres in our FCG scaffolds, we suggest that future work should focus on finding ways to further increase the overall content of elastic fibres. Indeed, the use of supplements such as microfibrillar associated protein 4 (MFAP4) [60], transforming growth factor-β1 (TGF-β1) [61,62] and LOX1 [63] are also known to induce a deposition of elastic fibres in fibroblasts and hVSMCs. Interestingly, Mithieux et al [31] obtained an increase in elastic fibre deposition using medium containing higher than usual ECM-associated proteins. Recently, the use of macromolecular crowding techniques [64] and cell co-culture with monocytes [65], have been used as alternatives to enhance the deposition of ECM-associated proteins with promising results. These techniques, in combination with dynamic conditioning, using a bioreactor, might promote gene upregulation in cells, thus improving elastic fibre deposition, ECM and cell proliferation [66][67][68] in FCG scaffolds.
This FCG scaffold used here has applications in soft tissue reconstruction, as it combines fibrin's wound-healing and remodelling properties [37], the anti-calcification properties of GAGs [69], with collagen. In this study, we have shown how tuneable amounts of elastic fibres can be obtained by adding tropoelastin to stimulate elastogenesis [70]. As it stands, the current mechanical properties of this construct makes it suitable as a platform for skin [71] and lung alveolar wall regeneration [72], and tunica media repair in blood vessels [36].

Conclusions
We have successfully demonstrated that elastogenesis can be activated in the extracellular matrix (ECM) of cell-seeded scaffolds fabricated from fibrin, collagen and glycosaminoglycan (FCG) by adding recombinant human tropoelastin. We have proven the importance of cell density and having an ECM before adding tropoelastin for the deposition of maturing elastic fibres in human vascular smooth muscle cells (hVSMCs). It was observed that this technique allows the deposition of elastin and formation of lamellar maturing elastic fibres, like the ones found in skin, blood vessels and heart valves. Improved surface coverage of maturing branched elastic fibres can be achieved by drop-loading tropoelastin onto cell-seeded FCG scaffolds prior adding growth medium. Additionally, we could see the impact of elastic fibres in terms of mechanical properties in the scaffolds. With these results, we have established a natural biomaterial scaffold that can undergo controlled elastogenesis on demand, suitable for tissue engineering applications.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.