In vivo drug release behavior and osseointegration of a doxorubicin-loaded tissue-engineered scaffold

Abstract

Bone tissue-engineered scaffolds with therapeutic effects must meet the basic requirements as to support bone healing at the defect side and to release an effect drug within the therapeutic window. Here, a rapid prototyped PCL scaffold embedded with a chitosan/nanoclay/β-tricalcium phosphate composite (DESCLAYMR) loaded with the chemotherapeutic drug doxorubicin (DESCLAYMR_DOX) is proposed as a potential multifunctional medical application for patients who undergo bone tumor resection. We showed the DESCLAYMR_DOX scaffold released DOX locally in a sustained manner in mice without significantly increasing the plasma DOX concentrations. The evaluation of osseointegration in a porcine study showed increased mineralized bone formation, unmineralized collagen fibers and significantly higher alpha Smooth Muscle Actin (α-SMA) positive areas relative to the total investigated area (TA) in defects treated solely with the DESCLAYMR scaffold than in the DESCLAYMR_DOX; and alkaline phosphatase activity, α-SMA/TA and bone formation were higher in the DESCLAYMR loaded with 100 μg per scaffold DOX (DOX_low) than with 400 μg per scaffold DOX (DOX_high). Our results suggest that the DESCLAYMR_DOX can be a viable candidate as a multifunctional medical application by delivering the chemotherapeutic agent to target remaining tumor cells and facilitate bone formation.

---

Introduction

A bone graft is often required in the reconstructive treatment of traumatic bone defects, infections, degenerative diseases, and tumor resections. Autologous bone graft is the gold standard treatment for reconstruction of bone defects. However, its use is limited because it is not readily available and it subjects the patient to significant donor site morbidity.^1–5 The development of new graft concepts may enable a multifunctional approach to skeletal reconstruction.^6–10 Poly-ε-caprolactone (PCL) scaffolds are one of the most widely researched and implemented synthetic polymers in bone tissue engineering due to their biocompatibility, biodegradability, and adequate mechanical properties.^11,12 In combined use with bioactive biomaterials such as hydroxyapatite (HAP), tricalcium phosphate, and bioactive glasses, composite PCL scaffolds offer good osteoconductivity, which PCL scaffolds on their own do not provide.^13–15

Previously, we have developed and done in vitro and in vivo studies on a rapid prototyped PCL scaffold embedded with chitosan/nanoclay/β-tricalcium phosphate (β-TCP) composite (DESCLAYMR scaffold) (Fig. 1).¹⁶ Incorporating clay with chitosan and HAP improves both the mechanical and osteogenic properties of the scaffold.^17,18 In our earlier study, we showed DESCLAYMR scaffold could serve as a bone tissue engineered scaffold;¹⁶ DESCLAYMR could co-deliver the anticancer drug doxorubicin hydrochloride (DOX) and siRNA;¹⁹ we also showed that the DESCLAYMR scaffold loaded with DOX could effectively inhibit tumor growth and prolong local drug release in mice, which suggests that DESCLAYMR scaffolds can serve as matrices for the sustained delivery of the DOX.^20,21

Fig. 1 (A) Schematic illustration of the fabrication of the 3D-printed scaffold. PCL scaffolds were printed layer by layer, as shown (left) one layer of scaffold and (right) two layers of the scaffold. (B) Photograph (above panel) and fluorescent microscopy (bottom panel) of empty DESCLAYMR scaffold (left), and DOX-loaded DESCLAYMR scaffold (right) respectively. (C) SEM image of the DESCLAYMR scaffold.

The present preclinical studies aim to further investigate the local and systemic release behavior of DOX from the DESCLAYMR scaffold in a murine model and to examine the osseointegration of DESCLAYMR loaded with and without DOX in a porcine model. In the latter study, the effects of the released anticancer drug on the bone healing process are highly relevant. This study brings a very promising medical device concept one step closer to clinical trial, and eventually patients with bone tumors would benefit from such multifunctional device: the chemotherapeutic effect of released DOX to avoid the tumor recurrence and bone reconstruction after tumor resection.

Experimental procedures

DESCLAYMR scaffold preparation

The DESCLAYMR scaffold was produced as previously described.¹⁶ Briefly, PCL (MW = 50 kDa) (Perstorp, UK) scaffolds were produced by fused deposition modeling and punched out using an Acu-Punch® (Acuderm® Inc., Florida) (Ø = 4 mm, h = 2 mm for the mice study; Ø = 8 mm, h = 10 mm for the pig study). To increase the surface hydrophilicity the scaffolds were treated with 5 M NaOH for 3 hours, washed 3 times with PBS, rinsed with sterile water, and then immersed and coated with a 1% (w/v) chitosan (Chitopharm M with a 75–85% degree of deacetylation) (Cognis, Florham Park, NJ) in 1% (v/v) acetic acid solution containing chitosan modified montmorillonite clay (Cloisite Na⁺) (Lot: 07F28GDX-008, Southern Clay Products, Inc., Moosburg, Germany) (weight ratio to chitosan is 1 : 10, clay final concentration to chitosan solution is 0.9 mg mL⁻¹), and β-TCP (Lot: TCPCH01, Berkeley Advanced Biomaterials, Inc., Berkeley, CA) (weight ratio to chitosan is 1 : 20). The scaffolds were freeze-dried, neutralized in 70% ethanol containing 0.4 M NaOH, and rinsed in phosphate-buffered saline (PBS) three times before being freeze-dried again. Scaffolds were stored at room temperature in a desiccator prior to drug loading.

Drug loading

DOX (Sigma-Aldrich, Denmark) was suspended in PBS at a concentration of 4 mg mL⁻¹ (stock). For the mice study, 15 μL of stock solution was added on the DESCLAYMR scaffold (DES_DOX contained 60 μg DOX per scaffold) and for the pig study, 200 μL of 0.5 mg mL⁻¹ and 2 mg mL⁻¹ of DOX solution, respectively, were added on the DESCLAYMR scaffolds (DOX_low and DOX_high, respectively). The scaffolds were air-dried overnight and stored at −20 °C.

Mouse model for release behavior evaluation

Six- to eight-week-old female BALB/c mice (Taconic Farms Inc., Denmark) were used for evaluation of the release behavior of DES_DOX. All animal studies, including maintenance and determination of experimental endpoints, were performed in compliance with directives of the Danish Experimental Animal Inspectorate (License no. 2012-15-2934-00364). Anesthetized with continuous 2.5% isoflurane exposure, the mice (n = 6 in each group) received either a subcutaneous implantation of DES_DOX or a subcutaneous injection of the same dose of DOX (INJ_DOX, 60 μg DOX per injection) on each side of the hind flanks. Seven to 10 μL blood samples from each animal were taken for further evaluation at 4 hours (h), 24 h, 4 days (d), 7 d and 14 d post-implantation/-injection. The experimental endpoint was reached when weight losses exceeded 10% of mouse body weight measured prior to any procedures or before skin ulcerations due to the subcutaneous use of DOX had occurred. All the mice completed the observation period and no clinical signs of complications were observed.

The mice were imaged with the Xenogen IVIS Spectrum imaging platform (PerkinElmer, MA) with continuous 2.5% isoflurane exposure at selected time points (3 h, 24 h, 4 d, 7 d, and 14 d) in order to detect DOX fluorescence, either loaded into DESCLAYMR scaffolds or injected subcutaneously. Imaging variables were maintained for comparative analysis. DOX fluorescence was captured on a CCD camera (excitation 430–465 nm/emission 540–640 nm) and analyzed after spectral unmixing. A region of interest (ROI) was manually selected according to the size of the scaffold or injection area; the area of the ROI was kept constant in the course of the experiments. Intensity was recorded as radiant efficiency ([photons per s] per [μW cm⁻²]) within a ROI.

Five microliter serum was diluted with 50 μL methanol/water mixture (110 : 100, vol : vol) and 55 μL internal standard solution (5 μg L⁻¹ daunorubicin (Sigma-Aldrich, Schnelldorf, Germany)) in methanol/water (110 : 100, vol : vol) was added. Following the addition of 165 μL acetonitrile and vortex-mixing, the sample was incubated for 10 minutes at ambient temperature and centrifuged for 10 minutes at 10 000 × g. A 220 μL supernatant aliquot was evaporated to dryness under a nitrogen flow at 40 °C and reconstituted in 100 μL 15% methanol with 0.1% acetic acid. Ten microliter sample aliquots were analyzed on a Waters Acquity UPLC system (Waters, Milford, MA, USA) coupled to a triple-quadrupole mass spectrometer (Waters Xevo TQS) equipped with an electrospray ionization source. The mass spectrometer was operated in multiple reaction monitoring mode with the subsequent transitions for DOX (m/z 544.2 > 397.2) and daunorubicin (m/z 528.2 > 321.1). The liquid chromatography methodology has been described in detail previously.²² Separate blank mouse serum samples were spiked with known amounts of DOX reference material (European Pharmacopoeia Reference Standard, Sigma-Aldrich) and used to construct a calibration curve by weighted (1/x) linear regression analysis using the DOX peak area normalized with the internal standard peak area.

Pig model for the evaluation of osseointegration

Female adolescent pigs (mean age 193 ± 11 days, mean weight 146 ± 16 kg) of breed (Danish Landrace × Yorkshire) × Duroc were used for evaluation of osseointegration of the DESCLAYMR scaffolds loaded with and without DOX. The scaffolds were tested bilaterally in the humerus of 18 pigs. In nine pigs, DOX_high were tested against empty DESCLAYMR scaffolds (Control), and in the other nine pigs, DOX_low were tested against the control. Test and control sites were randomly assigned in each animal. The experiments complied with local laws and were approved by the Danish National Authority (no. 2012-15-2934-00362).

Premedication consisted of subcutaneous injection of 0.1 mL kg⁻¹ of 125 mg zolazepam and 125 mg tiletamine (Zolotil 50 vet., Virbac, Carros, France) dissolved into 6.25 mL xylazine (20 mg mL⁻¹, Rompun vet., Bayer Health Care, Leverkusen, Germany), 1.25 mL ketaminol (100 mg mL⁻¹, MSD Animal Health, Boxmeer, Holland), and 2.5 mL butophanoltartrate (10 mg mL⁻¹, Torbugesic Vet., Scan Vet Animal Health, Fredensborg, Denmark). Anesthesia was maintained with 10 mg kg⁻¹ propofol (10 mg mL⁻¹, B. Braun Melsungen AG, Melsungen, Germany), and 25 μg kg⁻¹ fentanyl (50 μg mL⁻¹, B. Braun Melsungen AG, Melsungen, Germany). Under sterile conditions, a 3–4 cm long incision was made on the lateral proximal humerus. The muscle was incised along the muscle fibers with minimal bleeding and hemostasis was assured. The bone was stripped from muscle in an area of approximately 1.5 cm². A customized, flattened drill bit was used to drill a press-fit hole into the bone. Subsequently, the scaffolds were press-fitted into the cavity, which was sealed with a small amount of bone wax. The muscle was sutured with resorbable suture and skin was sutured with a monofilament nylon suture 2-0, disinfected with 0.5% chlorhexidine in 80% alcohol, and dressed in a sterile fashion. Postoperative analgesia was administered subcutaneously three times daily for three days (buprenorphine, 0.015 mg kg⁻¹). Antibiotics were injected daily from one day before the operation until the second postoperative day (200 000 IU benzyl penicillin and 200 mg dihydrostreptomycin per 10 kg). Professional animal keepers assessed the well-being of the animals and checked them for signs of infection on a daily basis. The five-week observation time was chosen based on studies from our lab.²³ The pigs were euthanized with an overdose of pentobarbital. The scaffolds and surrounding bone were resected en bloc for further processing and analyses. One pig was euthanized before the designated observation time due to problems with a hind limb.

The resected tissues/scaffolds were immediately fixed in 70% ethanol and stored at 4 °C. Afterwards, the specimens were dehydrated in 70% ethanol, 96% ethanol, isopropanol and cleared in xylene and embedded in methylmethacrylate for histological sectioning. Due to the difficulties of sample preparation, it was only possible to use five animals from each group for histomorphometry.

Histomorphometry

Ten micrometer slices were cut from fixed axis vertically rotated (FAVER) sections in order to minimize bias.²⁴ The slices were cut on a microtome (Polycut E, Reichert-Jung, Heidelberg, Germany) and were stained with alkaline phosphatase (ALP), Goldner's trichrome (GT), Picrosirius red, and alpha Smooth Muscle Actin (α-SMA). ALP staining was performed by incubating the sections dark in a buffer containing variamine blue B salt (94820, Fluka, Sigma-Aldrich, Denmark) and sodiumnaphthyl-phosphate (6815, Merck, Denmark) mixed in propanediol for 20 min at 4 °C. Mayer's hematoxylin was used as counterstain. GT staining and Picrosirius red staining were performed as previously reported.^20,25 For α-SMA staining, sections were rehydrated and pretreated with 1% acetic acid for 10 min at room temperature (RT), and then quenched endogenous peroxidase with 3% H₂O₂ for 20 min at RT and incubated in 10% BSA for 20 min at RT to block unspecific bindings, and afterwards incubated with α-SMA antibody (1:1600, MAB1420, R&D, Denmark), overnight at 4 °C. The sections were rinsed with PBS and incubated with biotinylated antibody (1:300, E0433, DAKO, Denmark) for 60 min at RT, and then incubated with horseradish conjugated streptavidin (1:300, P0397, DAKO, Denmark) for 60 min at RT. Sections were sensed again with PBS and H₂O the sections and finally incubated with 3-amino-9-ethylcarbazole (A6926, Sigma-Aldrich, Denmark) for 30 min at RT and rinsed with H₂O. Mayer's hematoxylin was used as counterstain. Histomorphometry was performed by two independent observers using an Olympus microscope (Olympus, Ballerup, Denmark) with Visiopharm Integrator System software (newCAST, v. 3.4.1.0, Visiopharm A/S, Horsholm, Denmark). The total investigated area (TA) was defined by the implant/bone interface and the parallel line in a distance of 2 mm from the interface (white square in Fig. 4A). A point-count technique was utilized to quantify bone formation (bone/TA), ALP positive area fractions (ALP/TA), collagen fiber formation (collagen fibers/TA), and α-SMA positive area fractions (α-SMA/TA).²⁴ Intra- and inter-observer variance was determined.

Statistical analysis

Data were analyzed with STATA v12.0 software (Stata Corporation, Lakeway Drive, TX, USA). Normal distribution was checked with QQ-plots, Shapiro–Wilk test and variance homogeneity was determined using Bartlett's test. Normally distributed data were presented as group mean ± standard errors (SEM). The data of DOX release and plasma DOX level was determined using Student's t-test (equal variance) or Wilcoxon rank-sum test (unequal variance). The data of histomorphometry was determined with one-way analysis of variance (ANOVA), and specific comparisons between defects treated with DES_DOX (DOX_low and DOX_high) and their respective controls were made with paired Student's t-test. Specific comparisons between DOX_low and DOX_high were made with unpaired, two tailed Student's t-test. A p-value of less than 0.05 was considered statistically significant.

Results and discussion

Local release of DOX in mice

DOX release in the implantation/injection area was estimated by detecting DOX fluorescence imaging in an IVIS scanner. Quantification of radiant efficiency showed that there was a burst DOX release in the local area over three hours post-DOX injection in the INJ_DOX group, followed by a fast decrease within the first 24 h (Fig. 2A). In contrast, no burst release in the local area was observed in the mice receiving the DES_DOX implant. In the first three hours, DOX fluorescence in the INJ_DOX group was significantly stronger compared with the DES_DOX group (p = 0.0018). After 24 h, the quantification of DOX fluorescence was similar in both groups, though more condensed in the DES_DOX group than in the INJ_DOX group (Fig. 2B). DOX fluorescence in the DES_DOX group was significantly stronger 4 days, 1 and 2 weeks, respectively, after implantation/injection compared with the INJ_DOX group (*p = 0.0025; p < 0.0001; p = 0.0473, respectively).

Fig. 2 DOX concentration in implanted sites estimated using an IVIS system. (A) Quantification of DOX fluorescent density in local subcutaneous area of either scaffold loaded with DOX or injection of same amount of DOX. *p < 0.05. n = 6 in each group. (B) Representative fluorescent images from different time-junctures post-treatment.

In the present study, we demonstrated that DOX release from the scaffold prolongs the presence of DOX at the treatment site to a greater extent than in the INJ_DOX group. However, this quantification is only considered as an in vivo semi-quantitative measurement, due to the limitations imposed by the profiling method. We measured DOX near the surface of the treated area using an IVIS, constituting a relative small amount of total drug amount. Ideally, the evaluation of remaining DOX in the DESCLAYMR scaffold should be done by extracting DOX from the scaffold at each investigated time-point. However, this would require a large number of animals and inter-individual variation would increase when the measurements are not performed continuously on the same animal. Nevertheless, the comparison between DES_DOX and the INJ_DOX group is feasible because the treated sites for both groups are the same. Many other studies also used local-regional treatment devices/systems for prolonged therapeutic purposes. Shi et al.²⁶ used a dual function implant to maintain osseous space and release an antibiotic to eliminate local infection in craniofacial trauma and showed that the release of colistin was above its reported minimum therapeutic concentration for 5 weeks. Poly(glycerol monostearate-co-ε-caprolactone) films were reportedly a controlled, prolonged, and low-dose delivery matrix for the potent anticancer agent 10-hydroxycamptothecin (HCPT).²⁷ These drug-loaded films showed gradual and sustained release kinetics of HCPT over 7 weeks in vitro and thereby prevented the local growth of the malignant cells in an in vivo model. Chai et al.²⁸ developed a biodegradable poly-cyclodextrin functionalized porous bioceramics, which is a bone void filler material, to achieve a high local drug concentration which may better control residual malignant cells and promote reconstruction of bone defects. It showed a prolonged release period (three days longer for cisplatin, seven days longer for gentamicin) in vitro. Castro et al.²⁹ reported a ciprofloxacin implant for treatment of bacterial bone infection and showed a prolonged release of the loaded drug, which was dependent on the erosion and disintegration of the device for up to 8 weeks in a rabbit femur. These studies either investigated the in vitro release to predict a sustained local release in vivo^26–28 or sacrificed experimental animals at each time point to determine the in vivo release profile.²⁹ In contrast to these studies, we employed the IVIS system to achieve an in vivo alive test following the same animal at progressing time points. Considering the large individual biological variance, our methods exhibit more accuracy. Furthermore, we can improve our study by prolonging the observation period and including more experimental animals to achieve a better understanding of local DOX release behavior.

Systemic release of DOX in mice

Mice receiving INJ_DOX showed significantly higher plasma DOX concentration of 68.53 ± 11.29 μg L⁻¹ 4 h post-injection compared with 21.91 ± 4.85 μg L⁻¹ DOX for mice receiving DES_DOX (*p = 0.0176, Fig. 3). The DES_DOX produced a three-fold lower peak plasma concentration compared with INJ_DOX. No significant difference was found between the two groups 24 and 96 hours post-implantation/injection of DOX. The levels of DOX after 96 hours of DES_DOX: 1.78 ± 0.40 μg L⁻¹ and INJ_DOX: 1.99 ± 0.15 μg L⁻¹ left in the circulation system was at the lower limit of quantification in both groups.

Fig. 3 Quantification of DOX concentration in mouse serum from serial time-points post-treatment by ultra high performance liquid chromatography. *p < 0.05. n = 6 in each groups.

The pharmacokinetics of DES_DOX was compared with INJ_DOX by subcutaneous administration. As shown in Fig. 3, the C_max in the plasma of mice receiving DES_DOX was significantly lower than in those receiving INJ_DOX, which indicates that DES_DOX may decrease the C_max of free DOX by gradual release of DOX from the scaffold, thus reducing the amount of the drug leaving the local treatment region. Reduced systemic drug concentration can minimize side effects while increased local-regional drug concentration improves therapeutic concentration at the target site.^30–33

It is important to consider concerns regarding the safety of using DOX, such as the amount delivered in a single- and cumulative doses and exposure period. Many studies have evaluated the systemic toxicity induced by accurate or accumulated high doses of DOX using a mouse model.^34–36 One study showed that a single dose of 10 mg kg⁻¹ injected intravenously induced cardiomyopathy in Balb C mice 30 and a cumulative DOX dose of 24 mg kg⁻¹ showed cardiotoxicity in B6C3F1 mice.³⁷ Bertazzoli et al.³⁸ have reported evidence of minimal lesions in the spleens of CD1 mice after exposure to 40 mg kg⁻¹ cumulative DOX dose. In this study, we used a single doses of 6 mg kg⁻¹ of mouse weight in both the INJ_DOX and the DES_DOX groups based on the previous study, which showed that this dose effectively inhibits tumor growth in nude mice for 2 weeks and can therefore be considered as a therapeutic dosage.²⁰ It is only logical to consider side effects when applying a therapeutic dose in the treatment. We focused on the pharmacokinetics rather than examine adverse side effects because the dosage we employed was much lower than in the literature and subcutaneous administration restricted the plasma DOX level when compared with intravenous injection employed in most of the studies.

Plasma DOX concentration was one of the most important elements in the pharmacokinetics study. Ren et al.³⁹ evaluated plasma DOX concentration using HPLC after a single intratumoral injection (dose: 20 mg kg⁻¹) and reported a rapid decrease of DOX level in plasma from 310 ng mL⁻¹ one hour post-injection. In contrast, plasma DOX levels were raised 12 hours after the injection of LipDOX and gradually decreased until the seventh day. These data were to some extent comparable to ours: the DOX were in both cases released from a compartmentalized device – either a liposomal drug carrier or a local-regional treatment scaffold instead of intravenous injection of free DOX. However, more direct comparison between Lipodox and DESCLAYMR_DOX need to be conducted in future studies.

Osseointegration evaluation in vivo

In order to investigate the effect of different concentrations of DOX on the osseointegration of the DESCLAYMR scaffold, we employed multiple staining protocols. Paired comparisons were used in each staining based on the study design, in which DOX_high and DOX_low groups were tested against their own controls.

Quantification of ALP activity

We investigated ALP activity and in-growth of the mineralized bone by means of histological ALP staining counterstained by Mayer's hematoxylin. ALP positive area was measured relative to TA. In defects treated with control, ALP/TA was significantly higher than DOX_high (*p = 0.0272 compared with DOX_high). There was no difference of ALP/TA between defects treated with DOX_low and their respective controls. ALP/TA was significantly higher in the DOX_low group compared with the DOX_high group (p = 0.0007) (Fig. 4A). Representative images of ALP staining from all three groups are shown in Fig. 4B.

Fig. 4 (A) Examples of marked area for histomorphometry evaluation and ALP positive areas estimated by histomorphometry. Data are presented as mean ± SEM. *p < 0.05. n = 5 in all groups. (B) Representative histological images of scaffold implanted in pig humerus defect. ALP positive areas appear brown and osteoblast-mediated bone formation appear condensed purple in ALP/Mayer's hematoxylin staining. Scale bar: left column = 2000 μm, right column = 200 μm.

Quantification of unmineralized collagen fibers

Picrosirius red was used to evaluate area of unmineralized collagen fibers in the investigated area. In the control, collagen fibers/TA was significantly higher than DOX_low and DOX_high respectively (*p = 0.0102 compared with the DOX_low and p = 0.0072 DOX_high groups) (Fig. 5A). There was no significant difference between scaffolds with two loading concentrations. Representative images of Picrosirius red staining from all three groups are shown in Fig. 5B.

Fig. 5 (A) Collagen fibers/TA estimated by histomorphometry. Data are presented as mean ± SEM. *p < 0.05. n = 5 in all groups. (B) Representative histological images of scaffolds implanted in pig humerus defect. Picrosirius red stained sections were imaged with polarized light. Collagen fibers appear yellow/green and bone tissue appears condensed deep orange/red in Picrosirius red staining. Scale bar: left column = 2000 μm, right column = 200 μm.

Quantification of bone formation

Area of bone formation/TA in different groups was evaluated using ALP/Mayer's hematoxylin and Picrosirius red staining. As shown in Fig. 6, bone formation/TA in control group was significantly higher than scaffold loaded with DOX (*p = 0.0152 compared with DOX_low and *p = 0.0081 compared with DOX_high). There was a significantly higher bone formation/TA in DOX_low compared with DOX_high (p < 0.0001) (Fig. 6).

Fig. 6 Bone formation evaluated using ALP/Mayer's hematoxylin and Picrosirius red staining. Data are presented as mean ± SEM. *p < 0.05. n = 5 in all groups.

Quantification of α-SMA positive areas

α-SMA is a marker for both osteoprogenitor and myofibroblast,⁴⁰ therefore a fraction of α-SMA positive areas is an important parameter to evaluate osseointegration of the DESCLAYMR scaffold. α-SMA/TA in the control group was significantly higher than the DESCLAYMR loaded with DOX (*p = 0.0218 compared with DOX_low and *p = 0.0081 compared with DOX_high) (Fig. 7A). There was a significantly higher bone formation/TA in DOX_low compared with DOX_high (*p = 0.0091). Representative images of α-SMA staining from all three groups are shown in Fig. 7B.

Fig. 7 (A) α-SMA positive areas estimated by histomorphometry. Data are presented as mean ± SEM. *p < 0.05. n = 5 in all groups. (B) Representative histological images of scaffold implanted in pig humerus defect. α-SMA positive cells appear red. Scale bar: left column = 2000 μm, right column = 200 μm.

The DOX dosages used in the evaluation of osseointegration were based on the release profile and in vitro cytotoxicity evaluation, which has proved the amount of DOX released from DOX_high group at week 5 was around 3.8 μg mL⁻¹, which could inhibit viability of human osteosarcoma cells MG-63 around 82% (Fig. S1†). Clinically, DOX is generally given with a dose of 60 mg m⁻² for every 3 weeks. It is reported that DOX reaches the peak plasma concentration at 5 μM (2.7 μg mL⁻¹).⁴¹ Therefore, DOX_high group at 5 weeks post implantation can mimic the situation of the scaffold mainly working as a chemotherapeutic tool. We could foresee that the DOX_low group possessed limited cytotoxicity after releasing DOX for 5 weeks and therefore transition gradually towards a bone repair facilitator instead of a chemotherapeutic device. We also tested the osseointegration of a drug-free construct to mimic the situation that drug was released out from the construct.

We chose to quantify the margin area of the scaffold rather than the whole area in order to focus on the bioactivity initiated from the implant/bone interface. Multiple biomarkers were applied in histology- and morphology analyses in order to fully understand the entire process of bone regeneration after being defected, including the reactive phase, the reparative phase, and the remodelling phase.

Based on previously reported studies, α-SMA was applied to identify either osteoprogenitor or myofibroblast.^40,42 Kalajzic et al.⁴⁰ indicated that the cells had the potential to differentiate into functional osteoblasts. Other studies have reported that α-SMA positive myofibroblasts can participate in fibrotic changes of various tissues following different injuries.^42,43 Taken together, the α-SMA positive cells are possibly involved in the reactive phase in bone healing and a high-level presence of these cells indicates that (1) a large number of osteogenic precursor cells might migrate from the surrounding tissue into the scaffold (Fig. 6); or (2) a fibrotic response to the implanted scaffold occurred in this area, which could be an obstacle for the formation of bone tissue. The latter possibility can decease the osseointegration of the scaffold, which is more crucial in DOX loaded ones. Therefore, further identification of myofibroblast and osteoprogenitor cells is needed.

The reparative phase includes activation and differentiation of osteoprogenitor cells, formation of randomly oriented collagen fibers, and mineralization in the collagen matrix with simultaneous vascularization. To follow this process, Picrosirius red staining imaged by polarized light was used to identify fibers and quantify collagen content.^44,45 The matrix maturation was estimated using a marker of active osteoblasts: ALP, based on the fact that proliferating osteoblasts are believed to be involved in secreting and mineralizing bone matrix.⁴⁶ We showed that unmineralized collagen fiber fractions and ALP activity were negatively associated with the amount of loaded DOX (Fig. 4 and 5). These data indicate that there was an initial induction of bone formation by active osteoblasts and formation of collagen fibers in the margin area after 5 weeks post-implantation with the presence of low amounts of DOX. Nevertheless, significantly higher ALP activity and a relatively higher content of collagen fibers in the DOX_low group compared with the DOX_high group demonstrated that the amount of DOX loaded in the scaffold influenced the osseointegration of the DESCLAYMR scaffold to some extent.

Identification of multinucleated giant cells (MGCs)

Using Goldner's trichrome staining, we observed MGCs on the edge of bone formation in all three groups (Fig. 8). The difference among these three groups was the location of these cells. The MGCs were located closer to the center of the control while they were closer to the implant/bone interface in the DOX_high. The concentration of DOX was positively related with the distance between the MGCs and the DESCLAYMR scaffold/bone interface. No clinical evidence of an inflammatory response was observed in any group at any time.

Fig. 8 MGCs revealed by Goldner's trichrome staining. The nuclei were stained as black and appear red. The cytoplasm was stained as dark blue. In the pictures of the left column, the black lines indicate the distance to the implant/bone interface. The yellow squares indicate the enlarged pictures of the right column. The MGCs are indicated using the black arrows. Scale bar: left column = 2000 μm, right column = 40 μm.

We showed that the presence of DOX affected the migration of MGCs into the DESCLAYMR scaffold. MGCs appeared only around the implant/bone interface in the DOX_high group, while MGCs were more likely to migrate into the center of an empty scaffold. To explain this observation, it would be necessary to identify cell types with this morphology. MGCs appears around the newly formed bone, and involve in the process of osteolysis could be osteoclasts and responsible for bone resorption in remodeling phase, which can be identified by examining the tartrate-resistant acid phosphatase (TRAP) activity. The presence of MGCs along with scaffold fibers can be evidence of chronic inflammation and foreign-body reaction in the reactive phase. In our study, MGCs were located in both aforementioned sites. However, no TRAP-positive cells were observed in any of the three groups (data not shown). Therefore, MGCs are considered to be responsible for chronic inflammatory reaction, which might be from the chitosan or PCL in the scaffold.^47,48 There are other bone tissue engineering scaffolds in which the appearance of inflammatory reaction has been reported. Polydioxanone sheets (PDS) were evaluated for their ability to repair experimentally created orbital wall defects in a porcine model and while they provided sufficient bridging of defects, they produced irritation of the surrounding tissues, causing a secondary foreign-body reaction.⁴⁹ PCL scaffolds with 65% porosity, used by Rohner D. et al., showed 4.5% new bone formation at three months and foreign-body giant cells were observed around the implants.⁴⁸ In their study, no clinical signs of an inflammatory reaction in experimental pigs were found which is consistent with our observation. However, due to the study design, no significant effect of DOX on the inflammatory reaction can be concluded. In addition, DOX was reported to be responsible for induction of acute inflammation⁵⁰ and initiation of cardiac inflammation.⁵¹ More investigations regarding DOX-induced inflammatory reaction in the interface area will be performed in a future study in order to further evaluate biocompatibility the DESCLAYMR scaffold loaded with DOX in vivo.

Conclusions

In the present mice and porcine studies, we demonstrated that local-regional use of the DESCLAYMR_DOX implant has the capacity to deliver and maintain a high concentration of therapeutic drug in the treatment area without increasing the risks of systemic overdose or DOX induced adverse side effects. Addition of DOX to the DESCLAYMR scaffold had a negative effect on the osseointegration of the scaffold. A high loading amount of DOX in the DESCLAYMR scaffold affected osseointegration significantly. With lower amount of DOX, the effect was less negative. So it is rational to predict that the osseointegration of the scaffold can be regained and successfully provide a substrate for supporting newly formed bone tissue when the DOX was completely released. However, the possibility of fibrotic tissue encasement, which could decrease the osseointegration capacity of the implant, can not be eliminated. Further investigation is needed. In conclusion, the DESCLAYMR_DOX implant can function as an element in the delivery of chemotherapeutic agents to target remaining tumor cells and as a potential active stimulus for bone development/healing. This proof of concept study represents an important milestone in the preclinical assessment of using DESCLAYMR_DOX for bone tumor patients following tumor resection.

Author contributions

M. S., M. C., D. Q. S. L. and C. B. conceived the idea. M. S. and M. C. designed the experiments, M. S., M. C., M. W., J. H., A. B., F. DH., J. R., J. J., H. L., H. L., and D. Q. S. L. performed experiments. M. S., and M. C. wrote the manuscript and prepared the figures. M. J., J. K., and C. B. supervised and supported the study. All authors reviewed the manuscript and have given approval to the final version of the manuscript.

Acknowledgements

We thank lab technicians Anna Bay Nielsen, Jane Pauli and Vibeke Skovhus Nielsen for experimental assistance. This study is supported by Velux Foundation, the Aarhus Spine Research Foundation and the Lundbeck Foundation Nanomedicine Center for Individualized Management of Tissue Damage and Regeneration (LUNA). This paper was accepted as Late Breaking Poster at the 61st Annual Meeting of the Orthopaedic Research Society, Las Vegas, USA, 2015.

References

1. P. C. Dell, H. Burchardt and F. P. Glowczewskie Jr, J. Bone Jt. Surg., Am. Vol., 1985, 67, 105–112.
2. D. M. Montgomery, D. D. Aronson, C. L. Lee and R. L. LaMont, J. Spinal. Disord., 1990, 3, 370–375.
3. C. Delloye, O. Cornu, V. Druez and O. Barbier, J. Bone Jt. Surg., Br. Vol., 2007, 89, 574–579.
4. R. Dimitriou, G. I. Mataliotakis, A. G. Angoules, N. K. Kanakaris and P. V. Giannoudis, Injury, 2011,(suppl. 2), S3–S15.
5. E. García-Gareta, M. J. Coathup and G. W. Blunn, Bone, 2015, 81, 112–121.
6. J. M. Lane and H. S. Sandhu, Orthop. Clin. North Am., 1987, 18, 213–225.
7. K. A. Hing, S. M. Best, K. E. Tanner, P. A. Revell and W. Bonfield, Proc. Inst. Mech. Eng., Part H, 1998, 212, 437–451.
8. T. W. Bauer and G. F. Muschler, Clin. Orthop. Relat. Res., 2000, 10–27.
9. W. G. De Long Jr, T. A. Einhorn, K. Koval, M. McKee, W. Smith, R. Sanders and T. Watson, J. Bone Jt. Surg., Am. Vol., 2007, 89, 649–658.
10. C. N. Cornell, Bull. Hosp. Jt. Dis., 2004, 62, 13–17.
11. D. W. Hutmacher and S. Cool, J. Cell. Mol. Med., 2007, 11, 654–669.
12. L. Van der Schueren, B. De Schoenmaker, Ö. I. Kalaoglu and K. De Clerck, Eur. Polym. J., 2011, 47, 1256–1263.
13. A. R. Boccaccini and J. J. Blaker, Expert Rev. Med. Devices, 2005, 2, 303–317.
14. J. Jensen, D. C. Kraft, H. Lysdahl, C. B. Foldager, M. Chen, A. A. Kristiansen, J. H. Rolfing and C. E. Bunger, Tissue Eng., Part A, 2015, 21, 729–739.
15. E. Tamjid, R. Bagheri, M. Vossoughi and A. Simchi, Mater Sci Eng C, 2011, 31, 1526–1533.
16. M. Chen, D. Q. Le, S. Hein, P. Li, J. V. Nygaard, M. Kassem, J. Kjems, F. Besenbacher and C. Bunger, Int. J. Nanomed., 2012, 7, 4285–4297.
17. K. S. Katti, D. R. Katti and R. Dash, Biomed. Mater., 2008, 3, 034122.
18. A. H. Ambre, K. S. Katti and D. R. Katti, J. Nanotechnol. Eng. Med., 2010, 1, 031013.
19. M. Chen, M. Ø. Andersen, P. Dillschneider, C.-C. Chang, S. Gao, D. Q. Le, C. Yang, S. Hein, C. Bünger and J. Kjems, RSC Adv., 2015, 5, 101718–101725.
20. M. Sun, M. Wang, M. Chen, F. Dagnaes-Hansen, D. Q. S. Le, A. Baatrup, M. R. Horsman, J. Kjems and C. E. Bünger, Acta Biomater., 2015, 18, 21–29.
21. M. Sun, C. Yang, J. Zheng, M. Wang, M. Chen, D. Q. Le, J. Kjems and C. E. Bünger, Acta Biomater., 2015, 28, 171–182.
22. L. K. Sørensen and J. B. Hasselstrøm, Anal. Methods, 2013, 5, 3185–3193.
23. J. Jensen, J. H. Rolfing, D. Q. Le, A. A. Kristiansen, J. V. Nygaard, L. B. Hokland, M. Bendtsen, M. Kassem, H. Lysdahl and C. E. Bunger, J. Biomed. Mater. Res., Part A, 2014, 102, 2993–3003.
24. H. J. Gundersen, T. F. Bendtsen, L. Korbo, N. Marcussen, A. Moller, K. Nielsen, J. R. Nyengaard, B. Pakkenberg, F. B. Sorensen and A. Vesterby, et al., APMIS, 1988, 96, 379–394.
25. R. R. Recker, Bone histomorphometry: techniques and interpretation, CRC, 1983.
26. M. Shi, J. D. Kretlow, A. Nguyen, S. Young, L. Scott Baggett, M. E. Wong, F. K. Kasper and A. G. Mikos, Biomaterials, 2010, 31, 4146–4156.
27. J. B. Wolinsky, R. Liu, J. Walpole, L. R. Chirieac, Y. L. Colson and M. W. Grinstaff, J. Controlled Release, 2010, 144, 280–287.
28. F. Chai, M. Abdelkarim, T. Laurent, N. Tabary, S. Degoutin, N. Simon, F. Peters, N. Blanchemain, B. Martel and H. F. Hildebrand, J. Biomed. Mater. Res., Part B, 2014, 102, 1130–1139.
29. C. Castro, C. Evora, M. Baro, I. Soriano and E. Sanchez, Eur. J. Pharm. Biopharm., 2005, 60, 401–406.
30. B. Baroli, J. Pharm. Sci., 2009, 98, 1317–1375.
31. V. Mourino and A. R. Boccaccini, J. R. Soc., Interface, 2010, 7, 209–227.
32. B. V. Somayaji, U. Jariwala, P. Jayachandran, K. Vidyalakshmi and R. V. Dudhani, J. Periodontol., 1998, 69, 409–413.
33. V. Mourino, J. P. Cattalini, J. A. Roether, P. Dubey, I. Roy and A. R. Boccaccini, Expert Opin. Drug Delivery, 2013, 10, 1353–1365.
34. S. K. Bjelogrlic, J. Radic, S. Radulovic, M. Jokanovic and V. Jovic, Exp. Biol. Med., 2007, 232, 1414–1424.
35. Y. Shuai, J. Guo, Y. Dong, W. Zhong, P. Xiao, T. Zhou, L. Zhang and S. Peng, Toxicol. Lett., 2011, 200, 77–87.
36. L. E. Olson, D. Bedja, S. J. Alvey, A. J. Cardounel, K. L. Gabrielson and R. H. Reeves, Cancer Res., 2003, 63, 6602–6606.
37. V. G. Desai, E. H. Herman, C. L. Moland, W. S. Branham, S. M. Lewis, K. J. Davis, N. I. George, T. Lee, S. Kerr and J. C. Fuscoe, Toxicol. Appl. Pharmacol., 2013, 266, 109–121.
38. C. Bertazzoli, O. Bellini, U. Magrini and M. G. Tosana, Cancer Treat. Rep., 1979, 63, 1877–1883.
39. S. Ren, C. Li, Y. Dai, N. Li, X. Wang, F. Tian, S. Zhou, Z. Qiu, Y. Lu, D. Zhao, X. Chen and D. Chen, J. Pharm. Pharmacol., 2014, 66, 1231–1239.
40. Z. Kalajzic, H. Li, L. P. Wang, X. Jiang, K. Lamothe, D. J. Adams, H. L. Aguila, D. W. Rowe and I. Kalajzic, Bone, 2008, 43, 501–510.
41. I. Muller, A. Jenner, G. Bruchelt, D. Niethammer and B. Halliwell, Biochem. Biophys. Res. Commun., 1997, 230, 254–257.
42. B. Hinz, S. H. Phan, V. J. Thannickal, A. Galli, M. L. Bochaton-Piallat and G. Gabbiani, Am. J. Pathol., 2007, 170, 1807–1816.
43. B. Rolfe, J. Mooney, B. Zhang, S. Jahnke, S. Le, Y. Chau, Q. Huang and H. Wang, Regener. Med. Tissue Eng., 2011, 26, 551–568.
44. J. E. Jalil, C. W. Doering, J. S. Janicki, R. Pick, W. A. Clark, C. Abrahams and K. T. Weber, J. Mol. Cell. Cardiol., 1988, 20, 1179–1187.
45. K. T. Weber, R. Pick, M. A. Silver, G. W. Moe, J. S. Janicki, I. H. Zucker and P. W. Armstrong, Circulation, 1990, 82, 1387–1401.
46. Y. Wang, T. Uemura, J. Dong, H. Kojima, J. Tanaka and T. Tateishi, Tissue Eng., 2003, 9, 1205–1214.
47. J. N. Barbosa, I. F. Amaral, A. P. Aguas and M. A. Barbosa, J. Biomed. Mater. Res., Part A, 2010, 93, 20–28.
48. D. Rohner, D. W. Hutmacher, T. K. Cheng, M. Oberholzer and B. Hammer, J. Biomed. Mater. Res., Part B, 2003, 66, 574–580.
49. H. A. Merten and H. G. Luhr, Fortschr Kiefer Gesichtschir, 1994, 39, 186–190.
50. D. V. Krysko, A. Kaczmarek, O. Krysko, L. Heyndrickx, J. Woznicki, P. Bogaert, A. Cauwels, N. Takahashi, S. Magez, C. Bachert and P. Vandenabeele, Cell Death Differ., 2011, 18, 1316–1325.
51. L. Li, G. Takemura, Y. Li, S. Miyata, M. Esaki, H. Okada, H. Kanamori, N. C. Khai, R. Maruyama, A. Ogino, S. Minatoguchi, T. Fujiwara and H. Fujiwara, Circulation, 2006, 113, 535–543.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05351c

‡ These authors contributed equally to this work.

---