Poly(ester-ether)s: III. assessment of cell behaviour on nano ﬁ brous sca ﬀ olds of PCL, PLLA and PDX blended with amorphous PMeDX †

The aim of this paper is to investigate the physico-chemical properties, degradation behaviour and cellular response of electrospun ﬁ bre-sca ﬀ olds of semi-crystalline PCL, PLLA and PDX blended with amorphous poly(methyl dioxanone) (PMeDX). Electrospun PCL/PMeDX and PLLA/PMeDX blend mats in varying weight ratios of the two components were fabricated and their overall performance was compared with similar composition PDX/PMeDX sca ﬀ olds. DSC analysis showed almost no change in crystallization temperature of PCL with increasing PMeDX content and TGA showed a di ﬀ erent degradation pro ﬁ le as PMeDX content increased. The appearance of two crystallization peaks for PLLA/PMeDX blends suggested stereocomplex formation. As noted from AFM images, addition of PMeDX caused a change in the width of the lamellae from 14.8 (cid:1) 2.9 nm in 100/0 mat to 32.0 (cid:1) 11.5 nm in 85/15 mat. Moreover, PCL/PMeDX blend mats show a signi ﬁ cant drop in Young's modulus for 93/7, 90/10 and 85/15 compositions compared to 100/0 and 98/2. On the other hand, no clear trend in mechanical properties was observed for espun PLLA/PMeDX mats with increasing PMeDX content. Based on these analyses, it was concluded that PCL and PMeDX were immiscible while miscible blends were obtained with PLLA and PMeDX. Initial degradation of electrospun mats over a period of 5 weeks appears to occur via a surface erosion mechanism. In vitro cell culture studies using HDFs showed that the sca ﬀ olds were bioactive and a greater density of viable cells was noted on electrospun PCL/PMeDX and PLLA/PMeDX sca ﬀ olds compared to PCL and PLLA mats respectively. HDFs in ﬁ ltrated through the entire thickness of espun 85/15 PLLA/PMeDX sca ﬀ old due to a combination of factors including morphology, porosity, surface characteristics and mechanical properties.


Introduction
Scaffolds for tissue engineering applications serve as a biomimetic extracellular matrix (ECM) and play a critical role in supporting cells. 1 They are designed to conform to a specic set of requirements which are oen conicting. 2 They should be biocompatible, biodegradable, porous and possess appropriate mechanical properties. 3Electrospinning remains a preferred method due to its low cost, high throughput, ease of operation and system control.It allows the fabrication of non-woven mats containing bres ranging from tens of microns to tens of nanometers in diameter, which can mimic both the form and function of the native ECM.Because of their mechanical properties and degradation rates that closely match those of proteins in so and hard tissues, polymers are good candidates for the development of bone and vascular scaffolds. 4Several natural and synthetic polymers have been investigated to this end.The most commonly used synthetic polymers are the aliphatic polyesters, polycaprolactone and poly(L-lactide) due to their relatively good biocompatibility and mechanical performance.However, cell affinities towards synthetic polymers are oen poor as a consequence of their hydrophobicity and lack of cell recognition sites. 5,6everal research groups have considered blending PCL and PLLA with other synthetic or natural polymers for optimized mechanical properties, degradation and bioactivity.Several studies have investigated the use of electrospun (espun) PCL/ collagen [7][8][9][10][11][12] and PCL/gelatin [13][14][15][16][17][18] for biomedical applications.Lee et al. 19 reported that compared to PCL, PCL/collagen scaffolds possessed enhanced biomechanical properties that could resist higher degrees of pressurized ow.Zhang et al. 20 found that composite PCL and gelatin scaffold had higher elongation and better exibility compared to PCL mat.Moreover, cells could not only grow and proliferate but also migrate inside the composite scaffold.Mehdinavaz Aghdam et al. 21showed that PGA increased the hydrophilicity, water uptake and mechanical properties of polycaprolactone/polyglycolic acid (PCL/PGA) nanobrous mats.Kim et al. 22 reported that the addition of PEI to PCL increased the hydrophilicity of the resulting espun mats and cell attachment on the blend mats was favoured due to the cationic nature of PEI.The miscibility of the two polymers in the blend is also an important parameter.For instance, Han et al. 23 studied espun blends of semi-crystalline PCL and amorphous poly(tetramethylene carbonate) (PTMC) and showed that the polymers were phase-separated in the bres.In a recent paper, Son et al. 24 reported on in vitro and in vivo evaluation of espun polycaprolactone/poly(methyl methacrylate) (PCL/PMMA) brous scaffolds for bone regeneration.Addition of PMMA improved the wettability of the scaffolds, while decreasing PCL content caused a decrease in the tensile strength of the espun blend mats.Higher proliferation of MG-63 cells was observed on the 7/3 PCL/PMMA scaffolds.
PLLA has been blended with natural polymers such as gelatin, 25,26 silk, 27 chitosan 28 as well as synthetic ones such as PLGA 29 and PCL. 30Shalumon et al. 28 reported that espun PLLA/ chitosan blend nanobres showed enhanced degradation and better human dermal broblasts (HDFs) cell growth compared to pure PLLA bres.Liu et al. 29 prepared PLGA/PLLA nano-brous mats and analysed their thermal, morphological and mechanical properties.Elongation was found to decrease with increasing PLLA content.In addition, Young's modulus and tensile strength of espun blend mats of PCL/PLLA decreased with increasing PCL content. 30Scaffolds supported human adipose-derived stem cells (hASCs) well.However, the 1/1 wt ratio PLLA/PCL demonstrated better properties and cellular responses in all assessments.
Blending not only inuences physico-chemical properties of the resulting espun mats but also impacts on their bioactivity and biological performance.It is well established that surface properties of scaffolds play a key role in cell-scaffold interactions, especially during the initial stage of cell-seeding.The topography created by the nanobres has a strong inuence on the biological response of cells seeded on its surface through contact guidance. 31In turn, cell morphology determines cell proliferation and inltration within a scaffold. 32Previous studies have demonstrated that cells can recognize differences in substrate stiffness such that they can tune their internal stiffness to match that of the substrate, resulting in a change in their cytoskeletal structure. 33,34This causes broblasts to adopt a more spread phenotype on stiff substrates and the cell organizes the actin cytoskeleton into stress bres. 35On the other hand, broblasts do not spread well on soer substrates and have a cortical actin cytoskeleton but no stress bres.In fact, cell size and the spreading area increase with increasing stiffness of the substrate. 36,37urthermore, Cui and Sinko 38 showed that highly crystalline and rigid PCL/PGA surfaces were more efficient in supporting broblasts growth compared to amorphous and exible ones.Yip et al. 39 showed that broblast behaviour was governed by strain on substrates soer than 20 kPa while the latter was dependent on stress for stiffer substrates (>20 kPa).In another paper by Lo et al., 40 the authors concluded on broblasts preference for stiff substrates.They showed that broblasts generate more traction force and develop a broader and atter morphology on stiff substrates than on so ones. 40In addition to mechanical properties, cells are very sensitive to surface chemistry, surface energy and surface roughness. 41In fact, surface energy plays a key role in attracting specic proteins to the surface of scaffolds which, would in turn, affect cell affinity towards these materials.Surface roughness was also found to promote cell attachment and growth on PLLA scaffolds. 41oreover, good hydrophobic-hydrophilic balance is crucial for optimized biocompatibility and cellular response. 42Indeed, recent studies have shown that cells adhere, spread and grow more easily on moderately hydrophilic substrates than on hydrophobic or very hydrophilic ones. 43n a recent study, we reported on blend lms of semi-crystalline PDX and amorphous polyDL-3-methyl-1,4-dioxan-2-one (PMeDX) and showed that low amounts of PMeDX in the blends (of the order of 15 wt%) could act as plasticizer.Mechanical tests showed overall reduced tensile properties of the blend lms.Interaction parameters from viscosity analysis and surface morphology images indicated immiscibility of the blend lms over the range of compositions studied. 44In another paper, 45 the thermal, mechanical and degradation characteristics of espun PDX/PMeDX mats were discussed.AFM images of the espun bres showed an increasing degree of morphological heterogeneity with increasing PMeDX content.Hydrolytic degradation of espun mats was found to be mainly dependent on bre diameter.Espun PDX/PMeDX nanobrous scaffolds demonstrated excellent biocompatibility as demonstrated by HDF adhesion and proliferation.
In this paper, we analyse the properties of espun mats of semi-crystalline PCL and PLLA blended with amorphous PMeDX.Our objectives are: to better apprehend the physicochemical characteristics of espun PCL/PMeDX and PLLA/ PMeDX mats; to compare the response of these two nanobrous scaffolds to HDFs behaviour with PDX/PMeDX scaffolds described previously; 45 to attempt a correlation between physico-chemical properties and biological performance of scaffolds.The miscibility of the blends, the morphology of bres, their thermal, mechanical properties and hydrolytic degradation as well as their efficacy to promote HDF cell growth and inltration will be discussed.To the best of our knowledge, this is a rst study where cell growth on PCL, PLLA and PDX scaffolds is compared under similar conditions.

Physico-chemical and mechanical characterization of electrospun blend mats
Blends of semi-crystalline homopolymers (PCL and PLLA) and amorphous PMeDX in varying weight ratios (100/0, 98/2, 93/7, 90/10 and 85/15 wt%) were espun in HFIP at a concentration of 100 mg mL À1 .Thermal behaviour, bre morphology and mechanical performance of espun PCL/PMeDX and PLLA/ PMeDX mats were analysed to get a better insight into blend miscibility and surface characteristics.In particular, the effect of increasing PMeDX content on bre diameter, pore size and porosity were investigated and compared with PDX/PMeDX bres.
Analysis of thermal properties DSC analysis: crystallinity and crystallization kinetics of electrospun mats.The melting temperature (T m ), enthalpy of melting (DH m ), crystallization temperature (T c ), enthalpy of crystallization (DH c ) of espun PCL/PMeDX and PLLA/PMeDX mats are summarized in Tables 1 and 2. The degree of crystallinity of the blends (c blend ) and that of the PCL phase in the blends (c PCL ) were calculated as reported previously. 45The enthalpy of melting for 100% crystalline PCL and PLLA were taken from the literature as 139.5 J g À1 (ref.46) and 93.7 J g À1 (ref.47) respectively.
As can be noted from Table 1, espun PCL/PMeDX blend mats crystallize in the temperature range 31.2 to 33.9 C depending on PMeDX content.The small difference in crystallization temperature suggests immiscibility of PCL and PMeDX homopolymers.It is known from the literature that when the glass transition temperature, T g of the amorphous polymer in an immiscible blend is well below the T c of the semi-crystalline polymer as in our case, the amorphous polymer doesn't affect T c due to chain mobility at that temperature. 48In contrast, the crystallization temperatures of espun PDX/PMeDX bres were found to increase from 51.0 to 80.4 C as the PMeDX content increased.
The plots of relative crystallinity against crystallization time (Fig. 1A and B) show a sigmoid shape, indicative of a fast primary crystallization during the early stage and slow secondary crystallization in a later stage.Avrami constants, K and n which are related to crystallization kinetics and mechanism respectively were determined and the results summarized in Table 3. Variation in n values indicated that PMeDX interferes with PCL nucleation and crystallite formation with an overall increase in K value, as the PMeDX content increases.The increasing value of n in espun PCL/PMeDX blend mats compared to 100/0 denotes morphological change of crystallites, as will be discussed in the next section.
A shouldering of T c is noted for all espun PLLA/PMeDX compositions in contrast to pure espun PLLA mat.This is attributed to stereocomplex formation.Indeed, the formation of stereocomplex has been reported for mixtures of PLLA and PDLLA where crystallization peaks at 110 and 130 C were attributed to PLLA homocrystallites and PLLA/PDLLA stereocrystallites. 49,50The likely formation of a stereocomplex between PLLA and PMeDX translates miscibility or partial miscibility of the two homopolymers.The Avrami K values show an overall increase in crystallization rate with increasing PMeDX content.
As can be noted in Table 2, the T m of PLLA/PMeDX blend mats are very close to that of PLLA.A slight shouldering of T m is observed for the 98/2 and 93/7 composition which could originate from the thermal history of the polymers.The presence of two distinct melting transitions at 175 C and 166 C have previously been reported for aand b-forms of PLLA. 51,52From Tables 1 and 2, it can also be noted that the enthalpy of fusion, DH m of espun PCL/PMeDX and PLLA/PMeDX mats decrease with increasing content of amorphous PMeDX.

Thermogravimetric analysis (TGA)
Thermal degradation proles of espun PCL/PMeDX and PLLA/ PMeDX mats are shown in Fig. 2 and 3 respectively.Espun PCL mat shows a two stage degradation prole with an onset degradation temperature, T onset at 374.6 C. PMeDX, on the other hand, degrades in a single step with a much lower T onset of 78 C. Espun 98/2 PCL/PMeDX follows a similar degradation prole as the 100/0 mat with no change in T onset (Table 4).This suggests that PMeDX may be partially miscible at that composition.This is also conrmed by mass loss calculations whereby PCL and PCL/PMeDX 98/2 have comparable mass loss.However, with increasing PMeDX contents (7, 10 & 15 wt%), the degradation prole changes with a rst-step degradation occurring at a lower temperature.Mass loss calculations DW 1 is equivalent to the initial wt% of PMeDX for these blend compositions (Table 5), which denotes degradation of PMeDX in the blend.This supports immiscibility of PCL and PMeDX homopolymers, in line with DSC data.
Espun PLLA mat show a T onset of 324.2 C and increasing PMeDX content led to decreased thermal stability of PLLA (Table 4) with T onset dropping from 324.2 to 306.5 C. In contrast to PCL/PMeDX blend mats, only one degradation stage was noted for all espun PLLA/PMeDX mats.This further supports the formation of PLLA/PMeDX stereocomplex as proposed in the previous section.
On the basis of these results, it can be suggested that immiscibility increases with increasing PMeDX content for PCL/PMeDX, with the 98/2 blend being partially miscible.Moreover, the formation of stereocomplex between PLLA and PMeDX is conrmed by TGA.
The stretching vibration of PLLA was found at 1631 cm À1 .For espun PLLA/PMeDX mats, only one band was observed contrary to PCL/PMeDX mats.A slight shi in carbonyl stretching (4 cm À1 ) was noted with increasing PMeDX content.
In summary, FTIR analysis suggests immiscibility of espun PCL/PMeDX mats and plasticization at low PMeDX content (2 wt%) while PLLA and PMeDX appear to be miscible based on the appearance of a single band.

AFM analysis
Surface morphology.AFM images (Fig. 5A and B) show that both espun PCL and PCL/PMeDX nanobres exhibit cylindrical morphology.Espun PCL mats had a smooth surface in contrast to blend nanobres (85/15) which showed a rougher surface, possibly due to dispersion of PMeDX within the PCL matrix.Indeed, as reported by Leclair et al., 53 deformations of the rubbery amorphous polymer occur in an immiscible semicrystalline/amorphous polymer blend if the latter is above its T g during crystallization of the semi-crystalline polymer (T g PMeDX More in-depth analysis of AFM images were carried out to investigate the nanostructure of the espun mats.Fig. 6A shows the lamellar crystal morphology within PCL spherulites crystallized at room temperature.A clear morphological change is observed with the addition of PMeDX as shown in Fig. 6B.Indeed, the presence bright domains within the lamellae can be seen in Fig. 6B.The bright regions possibly correspond to amorphous phase as reported by Gomez-Pachon et al.Moreover, addition of PMeDX caused a change in the width of the lamellae from 14.8 AE 2.9 nm in 100/0 mat to 32.0 AE 11.5 nm in 85/15 mat, suggesting that PMeDX inuences PCL crystallization as already noted from Avrami constants, K and n in the discussion on DSC analysis.This conrms that PMeDX interferes with PCL nucleation and crystallite formation. PLLA/PMeDX bres do not show clear structures as for PCL/PMeDX but exhibit corrugation-like patterns (Fig. 7) whereas some PLLA bres show a regular wavy pattern perpendicular to the direction of the bres.The lamellae appeared very thin and were arranged in a parallel manner but with changing orientation relative to the bre, unlike in the case of espun PDX bre (Fig. 8).
Fibre diameter and pore size.Table 7 summarizes bre diameters and pore sizes as determined by SEM as well as porosity values based on calculations described in the experimental section.Overall, bre diameters decrease with increasing PMeDX wt% for both espun PCL/PMeDX and PLLA/ PMeDX bres while no clear trend was observed for PDX/ PMeDX bres, data was more erratic due to protrusions on the bre surface and no clear trend was observed. 45The decrease is explained by the fact that bre diameter of espun binary blends of incompatible polymers is dependent on the internal phase morphology and solution viscosity.Indeed, if the dispersed phase in the blend has a lower viscosity, as that of PMeDX, the solution jet will be stretched and elongated more by the electric forces during the electrospinning process. 55However, the decrease in bre diameter was more pronounced for PCL/ PMeDX bres compared to PLLA/PMeDX due to its higher elasticity and stretchability.
However, as for espun PDX/PMeDX bres, there was no clear cut trend in pore size for both espun PCL/PMeDX and PLLA/ PMeDX bres.Porosity values were found to increase slightly with increasing PMeDX content in both cases which may suggest a decrease in bre packing density.
In general, the extent of whipping motion determines bre diameter.When the electrospinning jet experiences more whipping motion with crystallization most likely occurring before the jet reaches the collector, small diameter bres result with brillar structures.Large diameter bres experience less whipping and crystallization is most likely to occur aer the jet reaches the collector. 56echanical performance.Mechanical properties of espun bres depend on a combination of several factors such as bre alignment, bre diameter, bre lay-ups and interface properties of bre-bre contact (bre fusion). 57,58Tensile properties of bres are affected by their structural morphology.In fact, the strength and elastic modulus of bres is inuenced by the lamellar and amorphous fractions of chains present within bres.Elastomeric property of bres is due to the amorphous phase of the bres while dimensional stability is attributed to the crystalline phase. 59Thus, the mechanical deformation characteristic of the bre is inuenced by both the random amorphous and ordered crystalline phases in the bre. 59The internal micro/nanostructural morphology of nanobres determines their physical and mechanical properties. 60As reported in previous studies, changes in lamellae alignment and thickness result in variations of mechanical properties. 61affar et al. 62 demonstrated that elongation at break decrease with crystalline orientation.Also, Bozic et al. 63 showed that substrates with coarse lamellar structures possessed higher elongation compared to those with ne lamellar microstructures.
As reported in our previous paper, 45 the Young's modulus of espun PDX/PMeDX mats decrease with increasing PMeDX wt%.Also, in general, both strain at break and peak stress increase   with decreasing bre diameters.PCL/PMeDX blend mats show a signicant drop in Young's modulus for 93/7, 90/10 and 85/15 compositions compared to 100/0 and 98/2 (Table 8 and Fig. 9).It is likely that the formation of phase boundaries impact on mechanical performance more than crystallinity changes. 64n the other hand, the interpretation of mechanical properties for PLLA/PMeDX blends is more difficult as no clear trend is observed with increasing PMeDX content (Table 9 and Fig. 10).This could be due to antagonist effects such as drop in crystallinity which impacts negatively on mechanical properties (98/2 composition) and formation of stereocomplex which could enhance (93/7 and 90/10) but in other cases (85/15) limit mechanical performance.This can be explained by the stereocomplex formation based on thermal analysis data where the percentages of homoPLLA crystallites and PLLA/PMeDX stereocomplex crystallites may affect mechanical performance.

Hydrolytic degradation studies
Hydrolytic degradation of espun mats was carried out in PBS at 37 C for 5 weeks.We have shown recently that degradation of espun PDX/PMeDX mats occurred via surface erosion and was  found to be dependent on bre diameter of the bres. 45Larger diameter espun 98/2 PDX/PMeDX mat degraded at a faster rate than smaller diameter 85/15 mat.Similarly, espun PCL/PMeDX and PLLA/PMeDX mats appear to degrade via a surface erosion mechanism during the time period investigated as noted by the slight drop in pH (Fig. 11) and linear mass loss proles (Fig. 12).
Mass loss proles showed enhanced degradation with increasing PMeDX content which can possibly be explained by a combination of factors: decreased crystallinity in the blends, reduced bre diameters and higher porosities.Mass loss proles show that espun PCL/PMeDX and PLLA/PMeDX exhibit almost similar degradation rates.For instance, 93/7 PCL/ PMeDX and PLLA/PMeDX mat had a mass loss of about 7.

In vitro biocompatibility studies
We have previously shown that addition of PMeDX to espun PDX bres resulted in a greater density of viable human dermal broblasts compared to espun PDX mat with cells migrating up to a maximum of 45.1 AE 11.8% throughout the scaffold aer 7 days. 45This was attributed to smaller bre packing density and higher porosity.Espun PCL/PMeDX and PLLA/PMeDX mats were subjected to cell viability studies to investigate the effect of PMeDX incorporation in PCL and PLLA.Human dermal broblasts (HDFs) were cultured and seeded on the scaffolds for a period of up to 7 days.
Cell attachment and proliferation.Fig. 14 and 15 show the SEM images of HDFs cultured on espun PCL/PMeDX and PLLA/ PMeDX scaffolds aer 1 and 7 days respectively.In general, it was observed that cells spread over the mat surface.Compared to espun PCL or PLLA mats, a higher density of cells was found on PCL/PMeDX and PLLA/PMeDX scaffolds as from day 1.A change in cell morphology was noted with increasing PMeDX content from bipolar spindle (as red marked in Fig. 14) to a cobble-stone morphology with poorly organized actin laments (Table 10) which was more pronounced with PLLA/ PMeDX than PCL/PMeDX.This suggests that higher PMeDX content results in poor cell adhesion, causing the cells to adopt a cobble-stone morphology with few lopodia as can be seen in espun 85/15 PCL/PMeDX mat (Fig. 16).The cobble-stone shaped cells seem to be covered with deposited material, which is most probably ECM.Despite poor adhesion to the substrate, the balled up cells proliferated and produced matrix to create a conducive environment.As discussed in previous sections, addition of PMeDX to PLLA alters mechanical properties and crystallinity more signicantly compared to PCL and as highlighted in the introduction, cell growth is inuenced by changes in crystallinity and mechanical properties of the substrate.The presence of cobble-stone HDF morphology was detected as from 15 wt% PMeDX for PCL/PMeDX mat and as early as 2 wt% PMeDX for PLLA/PMeDX mat.A change in broblast morphology from elongated, oriented to cobble-stone was observed with the application of shear stress as reported by Braddon et al. 66 The change in HDF morphology can be explained by the change in microstructures of espun bres.Ajami-Henriquez et al. 67 reported that the presence of lamellae promotes "cell contact-guidance".In fact, they showed that other factors such as chemical composition, degree of crystallinity and surface roughness did not play a major role in determining cell preference towards a specic material.Based on the above, we deduce that the presence of lamellae in espun PCL/PMeDX bres may account for the fact that cells grown on these surfaces were mostly spindle-shaped.Furthermore, cobble-stone morphology observed on espun PLLA/PMeDX mats can possibly be related to the absence of lamellae on the mats.
In contrast, no change in morphology was observed for PDX/ PMeDX mat with 0-15 wt% PMeDX when subjected to cell growth.An increasing degree of heterogeneity and surface roughness noted with increasing amorphous PMeDX in espun PDX/PMeDX mats 45 could explain enhanced cell adhesion.It was therefore hypothesized that surface roughness in espun PDX/PMeDX was the dominant factor on broblasts proliferation compared to the decrease in crystallinity.
Cell behaviour has been reported to be dependent on three factors namely, percentage cell coverage of the surface, pore size and cell type.Indeed, as demonstrated by Salem et al., 68 broblasts displayed a co-operative pattern of cell spreading whereby pores greater than cell dimensions were bridged by group of cells using their neighbours as supports.As depicted in Table 10, the decreasing crystallinity of the blends with increasing PMeDX wt% here implies an enhanced exibility of the corresponding mats and in accordance with the introductory paragraph explains poorer cell adhesion especially in the case of PLLA/PMeDX mats where only 2 wt% PMeDX causes a drop of 10% in crystallinity.
As discussed in the Introduction, previous reports have shown that bre diameter, porosity and surface roughness inuence surface hydrophobicity which also affects cell behaviour.To have a better insight into surface hydrophobicity, contact angles (CA) were measured for a few samples and are listed in Table 11.Compared to PCL, the contact angle of a 93/7 PCL/PMeDX mat is nearly halved which translates a signicant increase in surface hydrophilicity.The same trend was previously noted for espun PDX/PMeDX where the contact angles decreased upon increasing content of PMeDX. 45Interestingly, cell adhesion appeared optimal for this 93/7 PCL/PMeDX mat, with extensive ECM secretion by day 7.This supports the fact that higher hydrophilic surfaces promote cell adhesion.On the other hand, increasing PMeDX in PLLA/PMeDX mats did not cause signicant change in surface contact angles.This could partly explain the poor adhesion of cells onto PLLA/PMeDX mats, as we highlighted in the previous paragraphs.
Cell inltration.Engineered scaffolds should not only support cell attachment and proliferation but should also allow cell inltration throughout the scaffolds so as to promote uniform tissue regeneration. 69A major challenge therefore concerns the promotion of cellular ingrowth into espun scaffolds.
Cell-seeded espun PCL/PMeDX and PLLA/PMeDX mats were cryo-sectioned and stained with 4 0 -6-diamidino-2-phenylindole (DAPI) to image cell nuclei and determine cell migration.Fig. 17 and 18 depict the uorescence microscopy images of HDFs cultured on the scaffolds.HDFs appear as bright dots as can be noted from Fig. 17.The depth of cellular inltration was quantied and results summarised in Table 12.As detailed in the Experimental section, we have privileged % migration rather than distance migration.As observed previously, espun  PDX/PMeDX mats show an inltration up to 45.1%. 45No inltration is noted for PCL/PMeDX mats independently of PMeDX composition.Inltration is observed in varying percentages for PLLA/PMeDX depending on composition.Noteworthy is the fact that a 85/15 blend shows a 100% inltration.Comparison of the different blend mats shows that cell morphology appears to be a dominating factor inuencing cell inltration.Indeed, in the case of PDX/PMeDX and PCL/PMeDX mats, HDFs adopt more of a spindle-shape whereas cells are more cobble-stone shaped in the case of espun PLLA/PMeDX mats.The pore size of the scaffold mat is also a determining factor.Indeed, literature reveals that HDFs start secreting ECM almost immediately in scaffolds with pore size less than 15 mm.
It is possible that the presence of corrugation-like structures in 85/15 PLLA/PMeDX scaffold enhanced cell inltration.Indeed, as reported previously, 70 microstructure and morphological features such as lamellar thickness or interlamellar distance were found to guide human mesenchymal stem cells (hMSCs) towards the interior of silk scaffold, producing a multilamellar hybrid construct. 70DFs penetrated the full thickness of the 85/15 PLLA/ PMeDX scaffold and were homogeneously distributed at day 7.All cells resided on the surface of the 85/15 PLLA/PMeDX scaffold at day 1.This suggests that cells observed within the scaffold at day 7 is a result of cell migration and not initial cell seeding.We can therefore conclude that espun 85/15 PLLA/ PMeDX mat had the right combination of bre diameter, pore size, porosity, mechanical and degradation properties.Such extensive cell migration is quite rare for espun materials without modication in porosity.Indeed, a number of methodologies have been proposed for improving cell inltration.For instance, electrospinning has been combined with salt leaching to produce a PCL scaffold with engineered delaminations. 71Up to 4 mm of cellular inltration was observed aer 3 weeks in culture.Another common method of increasing porosity of espun mats is via the selective removal of sacricial bres as reported in a study by Baker et al. 72 Overall, cellular inltration improved with increasing PEO content.At higher PEO contents, nearly complete inltration was observed.However, some regions remained devoid of cells and cell distribution within the scaffold was non-homogeneous.
Overall cytocompatibility results indicate that the inclusion of PMeDX into espun PCL and PLLA scaffolds improves the in vitro bioactivity of the matrix through either chemical or mechanical signalling, or a combination of the two.
In summary, in vitro cell culture studies showed that compared to espun PCL and PLLA mats, a greater density of viable cells were observed on espun PCL/PMeDX and PLLA/ PMeDX scaffolds respectively.Moreover, extensive HDF inltration was noted in espun 85/15 PLLA/PMeDX mats.

Methods
Polymerization of 3-MeDX.A typical polymerization is hereby described.A solution of Sn(Oct) 2 was prepared by dissolving 0.1 g of Sn(Oct) 2 in toluene (5 mL) in a glove box.29 mL of this solution was transferred to a quick t tube containing 3-MeDX (1.16 g).The tube was then placed in a preheated oil bath at 80 C. Aer the desired polymerization time, the reaction was quenched in liquid nitrogen and the crude sample puried by dissolving in chloroform and precipitating in petroleum ether.The product was then dried under vacuum before characterization by 1 H-NMR and 13 C-NMR.Electrospinning of PCL/PMeDX and PLLA/PMeDX blends.PCL/PMeDX and PLLA/PMeDX were blended in the following ratios: 100/0, 98/2, 93/7, 90/10, and 85/15 w/w%.Each blend was immersed in HFIP in separate vials at a concentration of 100 mg mL À1 and le on a shaker plate overnight.The polymer solutions were then loaded into a 3 mL syringe and dispensed at a constant rate using a KD Scientic syringe pump.Electrospinning parameters were optimized (rate: 3.5 mL h À1 , air-gap distance: 20 cm, voltage: +26 kV, grounded back target) to produce continuous, non-woven bres which were collected on a rotating grounded rectangular mandrel (7.5 Â 2.5 Â 0.5 cm).Aer electrospinning, scaffolds were removed from the mandrel and stored in a desiccation chamber until further analysis.
Determination of porosity.The porosity of the scaffolds was measured using a method described by Soliman et al. 74 Briey, 10 Â 10 mm 2 disks of the espun scaffolds were weighed and subsequently immersed in 70% ethanol overnight with slight mechanical agitation.This was done to allow the ethanol to penetrate into the scaffold pores.The surface of the samples was then blotted dry on a lter paper and weighed once more to determine the mass of the ethanol present within the scaffold.Measurements were made on three samples of each scaffold type.The density of ethanol, PCL and PLLA are 0.788, 1.123 and 1.23 g mL À1 respectively.The porosity (3) was calculated as: where V EtOH is the volume of the intruded ethanol and was calculated as the ratio of the observed mass change aer intrusion and r EtOH .V PCL is the volume of PCL bres and was calculated as the ratio between the dry scaffold mass before intrusion and the density of PCL (r PCL ).Non-isothermal crystallization kinetics.Non-isothermal crystallization kinetics was investigated using eqn (2) and (3) as detailed in our recent paper. 45 T 0 : onset temperature of crystallization, T: an arbitrary temperature, dH c /dT: variation of the enthalpy of crystallization as a function of temperature variation, DH c : total enthalpy of crystallization at a specic cooling rate.
t: time, T: temperature, b: constant cooling rate.Hydrolytic degradation of electrospun mats.Espun mats were placed in PBS at 37 C for ve weeks.Aer each degradation period, the samples were washed and dried under vacuum.% mass loss and % weight retention were calculated according to eqn (4) and ( 5) respectively.
In vitro biocompatibility studies.Cell culture and migration studies were carried out using human dermal broblasts.The experimental protocol was similar to the one used in our previous paper. 45To standardize the variation in scaffold thickness, cell migration was analysed as a percentage of distance travelled with respect to the thickness at that spot instead of solely distance travelled.% migration (n ¼ 30) was calculated according to eqn (6).

% migration ¼
distance of cell migration from the top thickness of scaffold at that exact point Â 100

Measurements
Differential Scanning Calorimetry (DSC) analysis was carried out using a Netzsch DSC 200 F3 Maia® thermal analyzer (Chennai, India).All PCL/PMeDX blend samples were heated from 30 to 80 C, cooled to À30 C and reheated to 80 C at 3 C min À1 .All PLLA/PMeDX blend samples were heated from 30 to 180 C, cooled to À30 C and reheated to 180 C at 3 C min À1 .Netzsch TG 209 F3 Tarsus® analyzer (Chennai, India) was used to measure and record the sample mass change with temperature over the course of the pyrolysis reaction.Thermogravimetric curves were obtained at a heating rate of 10 C min À1 between 25 C and 700 C. Nitrogen was used as an inert purge gas to displace air in the pyrolysis zone, thus avoiding unwanted oxidation of the sample.The sample mass used in this study was approximately 10 mg.SEM data were acquired on a Philips XL 30 scanning electron microscope operated at an acceleration voltage of 20 kV using the secondary electron detector.The polymer samples were sputter-coated with gold (5 nm thick) before imaging (Edwards, UK).Inner (in contact with mandrel) surface of the espun mats were imaged.To determine bre and pore size, the ImageTool 3.0 image analysis soware package was used (Shareware provided by University of Texas Health Science Center at San Antonio).The soware was calibrated using the micron scale bar of each picture.An average bre diameter was determined by measuring the diameter of 60 different bres, while an average pore size was determined by measuring the diameter of 60 different pores.Pores were identied as areas of void space bounded by bres on all sides at or near the same depth of eld, while their long and short diagonal axes were measured and averaged together to serve as their diameter. 75Intermittent contact (tapping) mode AFM imaging was done on as prepared samples on an Asylum MFP-3D atomic force microscope (Asylum Research, USA) using Olympus AC160TS cantilevers (with a resonance frequency of 300 kHz and a nominal spring constant of 40 N m À1 ) under ambient conditions.The rms amplitude of the cantilever was adjusted to 85 nm and a setpoint ratio of 0.8 was chosen.Constant amplitude images were acquired, depending on the scan size, with 512 pixels Â 512 pixels (up to 2024 pixels Â 2024 pixels); the phase shi was recorded simultaneously.The data was processed off line using the MFP-3D soware.Tensile properties of the espun mats (40 Â 10 Â 4 mm 3 ) were studied utilizing an Instron Tensile Tester 3343 (Instron, USA) at 27 C and 60% relative humidity using a crosshead speed of 10 mm min À1 , gauge length of 1 cm and 500 N load cell.Errors in Young's modulus and strain were calculated as reported in our previous paper. 45The contact angles of the modied surfaces were measured using water as a probe liquid (Milli-Q water from a Millipore Direct-Q 8 system with resistivity of 18.0MU cm À1 ) with an OCA 15plus instrument (Data Physics Instruments GmbH, Germany).Static contact angle data based on the sessile drop method were acquired immediately aer deposition of a 1 mL drop on at least three positions for each sample and are stated as arithmetic mean.A short lm sequence covering several seconds before and aer deposition of the droplet was taken.

Conclusions
Espun PCL/PMeDX and PLLA/PMeDX mats were successfully fabricated from HFIP solutions.FTIR and TGA analysis showed that immiscibility increases with increasing PMeDX content for espun PCL/PMeDX bres, with the 98/2 blend being partially miscible.AFM images revealed that espun PCL/PMeDX mats showed a rougher surface compared to espun PCL mat, possibly due to dispersion of PMeDX within the PCL matrix.Addition of PMeDX caused a change in the width of the lamellae as was noted from AFM images.Espun PLLA/PMeDX bres exhibit corrugation-like patterns.The drop in mechanical properties of espun PCL/PMeDX was explained by the likely formation of phase boundaries which impact on mechanical performance more than crystallinity changes.No clear trend was observed with increasing PMeDX content for espun PLLA/PMeDX due possibly to antagonist effects such as drop in crystallinity which impacts negatively on mechanical properties (98/2 composition) and formation of stereocomplex.Degradation of espun mats occurred via a surface erosion mechanism.In addition, biocompatibility studies conducted using HDFs showed that in general, espun PCL/PMeDX and PLLA/PMeDX scaffolds supported cell growth better than the corresponding espun homopolymer mats.Correlation of physico-chemical properties with biological performance is very complex and the nature of the polymer plays a key role.Adjustment of physico-chemical and mechanical properties to match bio-performance requires careful investigation almost on a case by case basis.

2 and 5 .
7% at week 5 and 85/15 PCL/PMeDX and PLLA/PMeDX mat had a mass loss of approximately 9.3 and 7.5% respectively at week 5.It has to be pointed out that mass loss values correspond to a relatively short degradation time.Zhao et al. 65 also observed linear mass loss proles for the hydrolytic degradation of espun PLLA and PLLA/hydroxyapatite mats.Degraded samples were analysed by SEM.Fig. 13 depicts SEM images of espun 98/2 and 85/15 PCL/PMeDX and PLLA/ PMeDX mats aer 5 weeks of degradation.Both 98/2 PCL/ PMeDX and PLLA/PMeDX mats show no change in morphology at week 5.However, the 85/15 PCL/PMeDX mat shows bre melting which is not observed in the corresponding PLLA/ PMeDX mat.This can possibly be explained by the smaller bre diameter of the 85/15 espun PCL/PMeDX mat.

Fig. 12
Fig. 12 Mass loss of (A) PCL/PMeDX and (B) PLLA/PMeDX mats as a function of hydrolysis time in PBS at 37 C.

Table 4
Variation of onset degradation temperatures of espun fibres a onset / C T 0 onset / C T max / C T onset / C T max / C a T onset : onset degradation temperature for 100/0 and 98/2 (one stage).T 0 onset : onset degradation temperature for rst stage degradation of 93/ 7, 90/10 and 85/15.

Table 5
Mass loss derived from TG degradation profiles a

Table 10
Influence of crystallinity and tensile strain on cell morphology

Table 11
Summary of static contact angle measurements on espun PCL/PMeDX and PLLA/PMeDX mats

Table 12
Summary of HDF migration in espun PCL/PMeDX and PLLA/ PMeDX mats