PROBING THE ROLE OF METHYL METHACRYLATE RELEASE FROM SPACER MATERIALS IN INDUCED MEMBRANE BONE HEALING

Abstract

In the induced membrane (IM) technique for bone reconstruction, a poly(methyl methacrylate) (PMMA) spacer is implanted to induce formation of a foreign body membrane around the defect site. Membrane development is essential for later bone grafting success, yet the mechanism by which the IM promotes bone regeneration remains unknown, as are the ways that spacer composition plays a role in the membrane’s healing potential. This study investigated the impact of leached methyl methacrylate (MMA) – the major monomeric component of PMMA – on IM development. In vitro cell culture found MMA elution did not impact endothelial cell or mesenchymal stem cell proliferation. For in vivo analysis, we advanced a streamlined rat femoral model to efficiently study the influence of spacer properties on IM characteristics. Comparison of membrane formation around polycaprolactone (PCL), MMA-eluting PCL (high dose PCL-MMA and low dose PCL-MMA), and surgical PMMA revealed robust membranes enveloped all groups after 4 weeks in vivo, with elevated expression of osteogenic bone morphogenetic protein-2 (BMP2) and angiogenic vascular endothelial growth factor (VEGF) compared with the surrounding muscle and bone tissues. Growth factor quantitation in IM tissue found no statistically significant difference between groups. New bone growth, vascularization, and CD163+ macrophage populations surrounding the polymer implants were also quantified; and blood vessel formation around high dose PCL-MMA was found to be significantly decreased compared with PCL alone. To our knowledge, these findings represent the first time that results have been obtained about the characteristics of membranes formed around PCL in the IM setting.

Introduction

The induced membrane technique for large bone defect reconstruction, also called the Masquelet technique, is one of the most successful approaches for the regeneration of segmental bone defects greater than 4 to 5 cm in length [1–4]. The method, first developed by Masquelet [5], is performed as two consecutive surgical stages. The first stage consists of extensive debridement and filling the bone defect with a poly(methyl methacrylate) (PMMA) cement spacer, which is typically left in the defect site for 4–10 weeks [2–3]. During this period, a foreign body-induced membrane develops around the surface of the PMMA cement. Then, during the second surgical stage, the cement spacer is removed but the induced membrane (IM) is left in place and an autologous bone graft is transferred to the resulting cavity.

The presence of the membrane surrounding the bone graft in the second stage creates a privileged environment for bone regeneration. Though the exact mechanisms are still not fully known, the biomembrane is speculated to provide several benefits to bone healing. As a physical envelope, the IM sequesters the transplant, preventing graft resorption and soft tissue invasion [1, 6]. Additionally, the IM is highly vascularized [7–9], aiding the reestablishment of blood vessels in the regenerating tissue. The membrane also produces growth factors such as bone morphogentic protein-2 (BMP2) and vascular endothelial growth factor (VEGF) [7–9] and contains adult mesenchymal stem cells [8].

PMMA cement is a convenient spacer material because it can be molded in situ to fill and stabilize defects of varying geometries. Surprisingly, it was shown that membranes induced by PMMA were more effective at promoting bone union than membranes induced by alternate space-holders made from titanium, despite the overall morphology of the membranes being similar [6]. The exact advantage bestowed by PMMA on the biological activity of the IM is as yet uncertain, as are which properties of the PMMA contribute to its greater success. To date, relatively few alternative spacer materials have been studied in Masquelet models, limiting our understanding of how spacer properties influence the IM regenerative potential.

Despite its increasing popularity, the induced membrane technique still has several drawbacks, not least of which is the long healing time required. The total time from placement of the PMMA until the bone is healed lasts approximately 9 months, regardless of defect size, and can take longer if complications arise requiring further surgical interventions [1, 3]. We previously reported the development of a femoral segmental bone defect model in rats to examine modifications to the traditional two stage approach [10]. Here we develop a simplified slot defect model (Figure 1) to increase throughput of investigation into the factors that impact membrane development. Briefly, in this method a slot defect is created in the rat femur, and a discoid spacer is inserted into the cavity, bringing the spacer material into contact with damaged bone, periosteal, and muscle tissues, as would occur in a larger segmental bone defect. The simplicity of the slot defect, however, avoids the need for lengthy and complex procedures involving internal or external fixation of the femur and is suitable in instances where a secondary procedure to investigate bone regeneration during the second stage of classical IM reconstruction will not be performed.

Figure 1.

Polymer specimens inserted into slot defect created in rat femur.

We speculate that in the future, an accelerated healing procedure might be possible by implanting a bioactive engineered membrane around an initial bone graft, eliminating the need for the primary spacer and the time required for membrane formation. However, since the way PMMA induces the bioactivity of the membrane remains unclear, the current study investigates one potential mechanism by which PMMA might impact IM development: the leaching of unreacted monomer units from the bulk polymer. It is well established that constructs made from PMMA release small molecules, the predominant chemical leachate being methyl methacrylate (MMA) [11–12]. MMA constitutes the main monomeric component of PMMA cements and has been associated with tissue irritation and inflammation [11, 13]. To investigate whether the release of MMA might play a role in IM development around PMMA, we characterized the membranes that formed around several types of polymer implant: polycaprolactone (PCL), PCL loaded with MMA at two concentrations (high dose PCL-MMA and low dose PCL-MMA), and a surgical PMMA bone cement.

Methods

Sample preparation

Medical grade polycaprolactone (Mn 80,000 Da; Sigma-Aldrich, USA) impregnated with methyl methacrylate (Acros Organics, Thermo Fisher, Scientific, USA) was prepared by dissolving PCL (0.50 g) in dichloromethane (DCM; 7 mL; Thermo Fisher Scientific) in a sealed vial. MMA (0.08 mL) was added to the solution and the vial was covered with a filter and allowed to evaporate to dryness in a fume hood, protected from light. Two concentrations of MMA-eluting PCL were prepared for this study: high dose PCL-MMA and low dose PCL-MMA. Levels of MMA loading were selected that fell within the range of MMA release shown by commercial PMMA products [12]. For the high dose PCL-MMA samples, the polymer solution was left to evaporate to dryness for two days until no DCM remained and was then implanted. For the low dose PCL-MMA samples, the polymer was left within the filtered vial another four days, allowing the evaporation of some of the residual MMA before being implanted. For plain PCL samples, the polymer was dissolved in DCM without the addition of MMA and the solvent evaporated to dryness. All PCL-based implants were cut into 6 mm diameter discs with one straightened edge and a thickness between 0.5–1.0 mm (see Figure 1) in a sterile culture hood.

PMMA bone cement (Simplex P, Stryker, USA) was prepared according to the package directions in a sterile culture hood; the components of the package were combined to form a slurry, which was then pressed into 0.5 mm thick sheets and allowed to fully solidify. Once set, samples were cut by scissor to the inserts seen in Figure 1. Samples were stored in sterile sealed containers protected from light until used. All PMMA samples were implanted between one and three weeks after preparation to ensure consistency between the levels of MMA released by different PMMA inserts (see Figure S-1).

Determination of MMA release kinetics

Individual polymer samples were submerged in deionized (DI) water (0.5 mL) and incubated at 37 °C, protected from light. At designated time points (day 1, 2, 7, 14, 21, & 28), the entire release media was collected and replaced with fresh DI water. The levels of MMA in the release media were quantified by gas chromatography (column: Zebron ZB-WAX plus, Phenomenex, USA; length 30 m, inner diameter 0.32 mm, film thickness 0.25 µm; carrier gas: He) with a flame ionization detector.

In vitro cell culture

To confirm that the PCL-based polymer samples would be safe for implantation, in vitro cell compatibility assays were first carried out using human umbilical vein endothelial cells (HUVEC; gift from Folkmann Lab [14]; culture medium: endothelial cell basal medium 2 + SupplementPack endothelial cell GM2, PromoCell, Germany) and human mesenchymal stem cells (hMSC; Lonza, Switzerland; culture medium: DMEM, Life Technologies, Thermo Fisher Scientific) supplemented with 10% FBS (Life technologies) and 1% penicillin-streptomycin-glutamate, (Life Technologies). To measure the viability of cells in the presence of polymer leachates, HUVEC (passage 5) and hMSC (passage 7) were cultured in a 1:1 mixture of their respective culture medium and release medium collected during quantitation of the MMA release kinetics. The cells were exposed to the release medium collected at day 1, 7, or 28 in order to observe any potential cytotoxic action by leachate from different time points in the materials’ planned residency in vivo. Cells in a 1:1 mixture of culture medium and DI water were used as a control. All cells were seeded in 96 well plates at a density of 3.1 x 10⁴ cells per cm² and allowed to attach for 4 hours, after which an equal volume of release medium or DI water was added. Cell viability was measured after 24 hours by Live/Dead staining (calcein AM/ethidium homodimer; Life Technologies) and imaging on a Zeiss fluorescent microscope. The images were processed using MATLAB to quantify the number of live (green) and dead (red) cells.

A cell proliferation assay was performed to see if leachate from the different polymer samples would influence cell growth. HUVEC (passage 5) and hMSC (passage 7) were seeded in 24 well plates at a density of 2.6 x 10³ cells per cm². Cells were allowed to attach for 20 minutes and then trans-well inserts were placed in each well with a freshly prepared PCL, high dose PCL-MMA, low dose PCL-MMA, or PMMA bone cement specimen. Cells were collected at determined time points (day 1, 4, & 7), lysed by sonication on ice in 1% Triton X-100 (Sigma-Aldrich) and quantified by PicoGreen DS DNA assay (Invitrogen, Thermo Fisher Scientific, USA).

Surgical model

Healthy 10-week-old male Sprague–Dawley (SD) rats (Charles River Laboratories, USA) used in this study were housed at the non-barrier animal facilities at Stanford University. The animal surgery protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Stanford University (APLAC-30586). All surgeries were performed under anaesthesia, and Buprenorphine Sustained-Release (1mg/kg, subcutaneous) was administered once before surgery. The operating procedure was revised based on previous publications [15–16]. A total of 24 rats were included in this study, with 6 rats randomly allocated to each group. Rats in each group received implants made from either PMMA, PCL, high-dose PCL-MMA, or low-dose PCL-MMA. Each rat was anaesthetized with 2–3% isoflurane during operation on a heated platform. After disinfection, a 2 cm incision was made through the skin from a point about 0.5 cm distal to the trochanter extending to about 0.5 cm above the femoral condyles. The fascia latae covering the thigh muscle was incised and the vastus lateralis retracted anteriorly from the intermuscular septum to expose the femur. Then a rectangular defect 6-mm in length and 1-mm in width on the lateral side of the femur was created using a fissure bur and a round bur. A polymer insert that had been wiped with ethanol was then used to fill the defect (Figure 1). After implantation of the polymer specimen, any deep tissues requiring re-approximation were repaired using resorbable sutures (4–0 Vicryl). The cutaneous layer was closed using sterile skin staples. The rats were sacrificed after 4 weeks, and samples were harvested. The femur with the polymer implant was trimmed about 5 mm from the insert on both sides and prepared for growth factor quantitation and immunohistochemistry (IHC).

Growth factor quantitation

Sections of the IM that formed between the polymer and adjacent muscle tissue were isolated during harvesting and snap frozen in liquid nitrogen, then stored at −80 °C. The tissue samples were weighed and then homogenized in protein extraction buffer by bead pulverization (Bullet blender tissue homogenizer & Navy bead lysis kit, Next Advance Inc., USA). Protein extraction buffer: 100 mM Tris (pH 7.4, Boston BioProducts, USA), 150 mM NaCl (Thermo Fisher Scientific), 1 mM EGTA (Research Products International, USA), 1 mM EDTA (Sigma-Aldrich), 1% Triton X-100, 0.5% sodium deoxycholate (Thermo Fisher Scientific); added immediately prior to use: protease inhibitor cocktail II (Abcam, UK), phosphatase inhibitor cocktail I (Abcam), 1 mM benzylsulfonyl fluoride (TCI, Japan). The homogenized IM tissue in extraction buffer was centrifuged and the supernatant was collected, aliquoted, and stored frozen at −80 °C until ready for use, as directed by ELISA manufacturer protocols. ELISA kits for VEGF (Sigma-Aldrich) and BMP2 (Abcam) were used to measure the concentrations of growth factors in the harvested membrane tissues. Growth factor expression was normalized by sample mass.

Histology and immunohistochemistry (IHC)

Femoral samples were fixed in 4% paraformaldehyde for 48 hours and the polymer inserts were carefully removed. The samples were decalcified in 10% EDTA for three weeks and then embedded in paraffin for sectioning. Tissue sections (5 µm) were cut orthogonally to the long axis of the femur and then stained either by hematoxylin and eosin (Sigma-Aldrich) or, for IHC, by primary antibody (mouse anti-rat) for VEGF (1:100 dilution), BMP2 (1:100 dilution), CD31 (1:250 dilution), and CD163 (1:250 dilution), followed by secondary antibody (goat anti-mouse; 1:100 dilution) conjugated with horseradish peroxidase (all antibodies obtained from Abcam). Secondary antibody was detected by reaction with 3,3’-diaminobenzidine (DAB, Thermo Fisher Scientific). Mayer’s Hematoxylin (Sigma-Aldrich) was used as a counterstain. Control sections stained without incubation in primary antibody can be seen in Figure S-2.

Image analysis and cell counting were performed by blinded observers. Images were taken by light microscope (BX50; Olympus, Japan) around the area where the implant material was placed in the rat femur, and image analysis was performed using Image-Pro Plus software (Media Cybernetics, USA) [17–18]. To measure the formation of new bone tissue and blood vessels, images were taken at 40x magnification and the area surrounding the polymer implants was divided into contiguous regions of interest (ROI) of 500 µm × 500 µm [19–20]. The macrophage immune response was also investigated using images captured at x400 magnification. Macrophages expressing CD163, involved in tissue repair, were counted [21–23]. The results from multiple ROI of 200 µm × 200 µm were averaged for each specimen.

Statistical analysis

Data are reported as means ± the standard deviation unless otherwise noted. All statistical analysis was performed using IBM SPSS software. Significance was defined as p < 0.05, as determined by one-way ANOVA with Tukey’s post-hoc test.

Results

MMA release kinetics

Gas chromatography was used to profile the release of MMA from medical grade PMMA and MMA-loaded PCL into a surrounding aqueous environment over the course of 28 days (Figure 2). Of the total cumulative values of MMA released over the 4 week period, both the low and high dose MMA-loaded PCL delivered more than half in the first 48 hours, during a period of burst release (57.8% and 69.7% of the cumulative release, respectively). And these respective values reached 71.1% and 83.9% of the 4 week cumulative release by the end of the first week. Simplex P PMMA displayed a more gradual burst release, with 34.5% of the cumulative MMA elution for the 28 day period occurring in the first two days and 68.4% by the end of the first week.

Figure 2.

Cumulative MMA release normalized by the mass of the polymer specimen (n=5). Values are reported as nanograms of MMA per milligram of the polymer spacer.

In vitro cell culture

To check that the PCL-MMA samples were cytocompatible prior to implantation, in vitro cell culture was carried out in the presence of the collected leachates. Live/Dead staining performed on HUVEC and hMSC cultured in media containing PCL and PCL-MMA leachates collected at 1, 7, and 28 days did not indicate any acute cytotoxicity from either the released MMA or any other potential leachate associated with the polymer processing. HUVEC showed greater than 97% viability and hMSCl showed greater than 99% viability when exposed to leachates from all time points (Figure 3A).

Figure 3.

In vitro cell culture of HUVEC and hMSC demonstrated that MMA release from the PCL-based polymers (A) did not produce cytotoxic effects (n=3) and (B) did not alter cell proliferation rates compared with PMMA (n=5). No statistical difference was found between groups (p > 0.05).

HUVEC and hMSC cultured for one week in the presence of PMMA, PCL, or PCL-MMA samples showed robust proliferation, with no statistically significant difference observed between the MMA-loaded PCL groups and the medical PMMA or plain PCL groups (Figure 3B). All cells proliferated to confluence in the available culture area.

Post-operative evaluation

Examination of gross tissue morphology showed that the bone had healed well adjacent to the embedded polymer implants. Tough fibrous membranes were observed tightly enveloping the outer surface of all polymer specimens and connecting to the periosteum on the surface of the bone. Sufficient membrane tissue was present to enable both harvesting for protein quantification (by ELISA) and histological processing. Only one rat experienced complications as a result of the slot defect model: one rat in the PCL group showed excessive bone healing at and around the defect site, with mineralized tissue fully covering the polymer sample and preventing collection of surface IM for ELISA growth factor quantitation. We believe this instance of bony overproduction may have been caused by a fracture at the defect, which stimulated excessive bone growth.

Growth factor quantitation

To determine if the levels of growth factor expression in the IM surrounding the different polymer groups varied, samples of the IM on the surface of the implant next to muscle were harvested during explantation and analyzed by ELISA. For each group, ELISA confirmed VEGF and BMP2 were present in the polymer-muscle IM tissue at detectable quantities. Although mean VEFG expression was somewhat elevated in the PCL-MMA groups compared to plain PCL, no statistically significant difference was found between the growth factor expression of the four groups (Figure 4).

Figure 4.

Growth factor expression in IM tissues harvested at the spacer-muscle interface (PMMA, Low dose PCL-MMA, High dose PCL-MMA n=6; PCL n=5). Values are reported as picograms of growth factor per milligram of IM tissue. No statistically significant difference was found between groups (p > 0.5).

Histology and immunohistochemistry

Histological analysis of the tissues surrounding the polymer inserts reveals robust membrane formation in all groups (Figure 5), both at the polymer-bone interface and at the polymer-muscle interface. At four weeks, no active inflammatory reaction was observed around the implants. Antibody staining for VEGF and BMP2 protein expression revealed that the IM tissue surrounding all polymer implants produced elevated levels of both growth factors compared with the surrounding muscle and bone tissues (Figure 6). Quantification of new blood vessels surrounding the implants revealed a statistically significant decrease in vascular growth for the high dose PCL-MMA group compared with the PCL group; while quantification of new bone growth and CD163+ macrophages found no significant difference between experimental groups, though the CD163+ macrophage population was somewhat diminished around the PCL-MMA specimens compared with PCL or PMMA implants (Figure 7).

Figure 5.

Histological staining by hematoxylin and eosin shows the tissues that developed around the implanted polymer spacers. Blue arrows indicate the induced membrane that formed at the surface of the polymer implants.

Figure 6.

Immunohistochemistry of the IM shows staining by DAB for VEGF and BMP2 in brown. Arrows point to IM tissue that developed at the polymer spacer surface – darker staining indicates a greater concentration of VEGF and BMP2 compared with adjacent tissue. Hematoxylin counterstain in purple/red.

Figure 7.

Quantification of (A,B) new bone tissue, (C,D) new blood vessels, and (E,F) CD163+ macrophages surrounding the polymer implants (n = 6). Exemplary histological pictures shown from PMMA (B,F) and PCL (D) groups. Panel in F shows macrophages captured at 400x magnification. Charts report mean values measured in the square regions of interest (ROI) lining the slot defect. ROI = 500 µm x 500 µm for new bone area and blood vessels, ROI = 200 µm x 200 µm for the macrophage count. (*) Statistical significance (p < 0.05).

Discussion

The induced membrane technique for large bone reconstruction has been practiced for several decades, yet many questions remain unanswered regarding how spacer composition influences membrane characteristics and regenerative efficacy. Previous research has shown that variation in IM morphology can be achieved by altering the spacer material [24] or by including additives in the PMMA cement [25]. Studies by McBride-Gagyi et al. investigating IM development around titanium implants observed biomembranes with comparable morphology and select growth factor expression to those surrounding PMMA [6, 26]. However, despite these similarities, PMMA was found to be more effective than Ti at promoting bone union during the second stage of Masquelet reconstruction. These results suggest that the selection of spacer material plays an important role in the IM’s regenerative potential, though which features of PMMA contribute to its greater success remain unclear. Also unclear are what membrane characteristics lead to enhanced bone regeneration and whether other polymeric materials might induce membranes with similar bioactivity to PMMA cements. Testing the IM formed around a greater variety of spacer materials would help determine which properties – both of the spacer material and of the IM tissue – are essential for IM regenerative efficacy.

In the current study, we investigated whether low levels of leached MMA monomer from PMMA cements play a role in the development of the IM. We hypothesized that the MMA released from PMMA bone cements might contribute to the specific bioactive properties of the membranes that form around PMMA implants. To test this hypothesis, we compared the membranes induced by blocks of plain PCL versus MMA-eluting PCL. Simplex P bone cement, composed of PMMA blended with polystyrene additive, was used as a positive control.

PCL is a biodegradable and biocompatible polymer that is widely used as a structural component in orthopedic tissue engineered constructs due to the good mechanical properties, processability, and osteoconductivity of the polymer and its composites [10, 27–32]. Though bioresorbable, solid PCL structures degrade slowly, with negligible polymer erosion occurring over the four-week period of our IM study, thereby avoiding potential concerns of excessive alterations to surface topography or chemistry, which could introduce additional factors impacting IM characteristics. We therefore determined PCL would provide a suitable small molecule delivery platform to study the effect of MMA elution and allow for more direct comparison with the non-degradable PMMA cement control.

Two concentrations of MMA-loaded PCL were explored: high-dose PCL-MMA and low dose PCL-MMA. Quantitation of the eluted MMA by GC (Figure 2) confirmed that both the high and low dose PCL-MMA released MMA at concentrations within the range detected for commercial PMMA cements [12]. For all three MMA-eluting groups (high dose PCL-MMA, low dose PCL-MMA, and PMMA bone cement) the greatest portion of the cumulative MMA release occurred in the first week, with steady release of low levels of the monomer detected for the remainder of the implantation period.

Cytocompatibility testing was carried out to confirm no cytotoxic effects would occur due to the MMA loading and other processing procedures, which might make the PCL-based groups unsuitable for in vivo use. Figure 3A demonstrates that cells exposed to the leachates collected from high and low dose PCL-MMA showed equivalent viability when compared with PCL alone or the control group. The materials were therefore deemed safe for in vivo implantation.

Cell growth studies carried out using HUVEC and hMSC cultured directly in the same well as the polymer samples showed robust proliferation in all groups (Figure 3B). The lack of significant differences between the growth profiles of cells exposed to the various implants suggests that leachates from PMMA, and MMA in particular, do not directly alter the growth rate of endothelial cells or MSC as a mechanism for altered membrane bioactivity.

Most previous research investigating foreign body-induced membrane formation considers subcutaneous implants [26]. However, the foreign body response is location dependent, and it cannot be inferred that membranes formed in reaction to a specific material will exhibit the same characteristics in the Masquelet milieu [33]. We have developed a novel surgical model that enables the placement of foreign spacer materials into a rat femoral slot defect to recreate the biological setting of IM bone reconstruction without the need for more complex surgeries requiring defect fixation. This enables a simplified and streamlined procedure that can be carried out with greater ease and speed at lower burden to the animal and surgeon. As with the traditional Masquelet procedure, our slot model results in direct contact between damaged bone tissue, the intramedullary cavity, and the spacer surface. The outer end of the polymer insert protrudes from the surface of the bone to interface directly with the surrounding muscle. As such, our model enables the study and comparison of the membrane that forms due to a foreign body at both the spacer-muscle interface – which represents the majority of the IM in clinical practice – as well as at the spacer-bone interface. It has been reported that IM harvested from the two interfaces have different properties [9], and so side-by-side analysis of the two regions may be of interest.

In clinic, PMMA spacers are often molded directly within the defect cavity during the first stage of IM bone reconstruction, though there are cases where pre-cured PMMA is recommended in order to avoid damage to the bone caused by heat generated during the exothermic curing reaction [2]. The implantation of PMMA cements that are not yet fully cured can lead to a large and relatively uncontrolled immediate burst release of unreacted MMA in the wound. In the present study, the surgical PMMA was prepared one to three weeks prior to implantation, which greatly decreased variance in the levels of MMA leachate observed from the PMMA group and also decreased the initial burst of MMA released from the polymer blocks (Figure S-1), though the levels of MMA released were still within the range of PMMA products used in clinic [12]. Pre-cured PMMA spacers have previously been demonstrated to promote bioactive IM formation in other in vivo models [6, 10, 26] and in the present study, the use of pre-cured PMMA implants enabled tighter control over the levels of MMA released from the individual polymer inserts, which was important for understanding the possible role of MMA concentrations on IM development. Future studies using PMMA spacers prepared in situ could offer insight into possible effects of the exothermic curing process and elevated MMA burst conditions on IM development, which the current study cannot provide. However, the narrowness of the slot defect used in the current study would make it challenging to insert implants that have not yet fully solidified.

ELISA growth factor quantitation confirmed the presence of VEGF and BMP2 in membrane tissues harvested from the graft-muscle interface (Figure 4). While no statistically significant difference was found between the groups; intriguingly, levels of VEGF were somewhat elevated in the PCL-MMA groups when compared with PCL alone, warranting further investigation. It is important to note that even when no significant quantitative difference is measured between growth factor production induced by alternate spacer materials, the membranes might still promote unequal rates of bone union [6, 26]. Further testing using a two-stage IM reconstruction model will be necessary to determine if any difference exists between the IM healing efficacy of PCL, PCL-MMA, and PMMA cement.

IHC analysis showed increased staining in the IM tissue surrounding all spacer groups at both the polymer-muscle and polymer-bone interfaces for both VEGF and BMP2 (Figure 6). Bone growth, new blood vessels, and CD163+ macrophages – which are involved in tissue repair – were also quantified around the polymer implants (Figure 7). While no statistically significant difference was found between the thickness of regenerated bone tissue or the populations of CD163+ macrophages between the groups, a significantly greater number of new blood vessels were recorded surrounding the PCL implants compared with the high dose PCL-MMA. As one of the speculative benefits of the induced membrane on bone formation is the presence of a rich vascular network, our findings suggest that the release of MMA from bone cement spacers is not responsible for this regenerative advantage. No clear qualitative differences were observed between the morphology or growth factor distribution within the membranes.

Our findings suggest that, as with PMMA, the membranes that form around the PCL and PCL-MMA spacers implanted in the Masquelet milieu express higher levels of osteogenic and angiogenic growth factors compared to adjacent tissues. To our knowledge, this work represents the first characterization of IM tissue around PCL in the induced membrane setting. PCL is widely used as a structural component in orthopedic tissue engineered constructs. The nature of the membrane induced by PCL when implanted in the Masquelet setting is therefore of interest for informing the design of future constructs as well as providing insight into how spacer composition impacts membrane properties in general. We anticipate further study of alternate spacer materials will shed light on the biological interactions of PMMA that give it enhanced efficacy in bone reconstruction.

We here propose an efficient method to examine the effects of spacer material on the IM properties. However, it is important to bear in mind that the IM characteristics which correspond to privileged bone tissue healing are not fully understood. Until future studies can determine which features of the membrane morphology produce improved bone regeneration, a two-stage study will be required to validate the surgical potential of PCL, PCL-MMA, and other polymers in comparison to PMMA spacer materials. Additionally, future investigation may well find that consideration of more membrane parameters than were included in this study – such as alternate growth factors – is required.

In conclusion, we have developed a streamlined surgical model to more efficiently investigate the variables that influence membrane induction and properties in the Masquelet setting. Though simplified, the procedure recreates the biologically relevant environment produced by more complex, time-consuming, and costly operations. Using this technique, we demonstrated that a membrane forms around PCL and PCL-MMA, producing osteogenic and angiogenic growth factors and with similar morphology to the IM that forms around PMMA cements. Different concentrations of MMA released from the polymer spacers exerted no statistically significant effect on HUVEC or hMSC proliferation in vitro, nor on the levels of VEGF and BMP2 expressed in vivo; though somewhat increased VEGF levels were observed in PCL-MMA over PCL alone, warranting further investigation. Quantitation of new blood vessels surrounding the implants revealed greater vascularization around the PCL implants in comparison to the high dose PCL-MMA material, suggesting that the release of unreacted monomer from PMMA surgical spacers is not responsible for promoting the rich vascular network found in induced membrane tissue. The robust vascularized membrane formation around PCL and PCL-based inserts indicates these materials represent interesting candidates for a two-stage IM bone reconstruction study to evaluate their healing potential.