In vivo bone and soft tissue response to injectable, biodegradable oligo(poly(ethylene glycol) fumarate) hydrogels

Biomaterials 24 (2003) 3201–3211

In vivo bone and soft tissue response to injectable, biodegradable oligo(poly(ethylene glycol) fumarate) hydrogels Heungsoo Shina, P. Quinten Ruhe! b, Antonios G. Mikosa, John A. Jansenb,* b

a Department of Bioengineering, Rice University, MS-142, P.O. Box 1892, Houston, TX 77251-1892, USA Department of Biomaterials, College of Dental Science, University Medical Center Nijmegen, P.O. Box 9101, Nijmegen HB 6500, The Netherlands

Received 11 December 2002; accepted 9 March 2003

Abstract This study was designed to assess in vivo bone and soft tissue behavior of novel oligo(poly(ethylene glycol) fumarate) (OPF) hydrogels using a rabbit model. In vitro degradation of the OPF hydrogels was also investigated in order to compare with in vivo characteristics. Four groups of OPF hydrogel implants were synthesized by alternation of crosslinking density, poly(ethylene glycol) (PEG) block length of OPF, and cell-binding peptide content. The in vitro degradation rate of OPF hydrogels increased with decreasing crosslinking density of hydrogels, which was characterized by measuring weight loss and swelling ratio of hydrogels and medium pH change. Examination of histological sections of the subcutaneous and cranial implants showed that an uniform thin circumferential fibrous capsule was formed around the OPF hydrogel implants. Quantitative evaluation of the tissue response revealed that no statistical difference existed in capsule quality or thickness between implant groups, implantation sites or implantation times. At 4 weeks, there was a very limited number of inflammatory and multinuclear cells at the implant–fibrous capsule interface for all implants. However, at 12 weeks, OPF hydrogels with PEG block length of number average molecular weight 6090790 showed extensive surface erosion and superficial fragmentation that was surrounded by a number of inflammatory cells, while OPF hydrogels with PEG block length of number average molecular weight 930710 elicited minimal degradation. Constant fibrous capsule layers and number of inflammatory cells were observed regardless of the incorporation of cell-binding peptide and crosslinking density of OPF hydrogels with PEG block length of number average molecular weight 930790. These results confirm that the degradation of implants can be controlled by tailoring the macromolecular structure of OPF hydrogels. Additionally, histological evaluation of implants proved that the OPF hydrogel is a promising material for biodegradable scaffolds in tissue engineering. r 2003 Elsevier Science Ltd. All rights reserved. Keywords: Oligo(poly(ethylene glycol) fumarate); In vivo bone tissue response; In vivo soft tissue response; Degradation; Tissue engineering

1. Introduction Tissue loss as a result of trauma, cancer treatment, and congenital disorders is a considerable clinical problem in reconstructive surgery. The current treatments to reconstitute lost or damaged tissue involve the use of an autograft, allograft, or a synthetic material. However, all these treatment modalities have their own limitations such as donor site morbidity, pathogen transfer, and complications due to mismatch of mechanical properties. A tissue engineering approach offers an *Corresponding author. Tel.: +31-24-3614006; fax: +31-243614657. E-mail address: [email protected] (J.A. Jansen).

alternative and involves the use of a scaffold that can provide a temporary platform for cell growth and can be degraded over time as new tissue develops [1]. Synthetic hydrogels are crosslinked polymeric networks that swell in water. The macromolecular structure of hydrogels resembles that of the extracellular matrix (ECM), which is composed of a variety of natural macromolecules including polyaminoacids and polysaccharides [2]. In addition, the versatile swelling characteristics of hydrogels allow for tissue-like elastic properties and high permeability of water [3]. Due to these characteristics, synthetic hydrogels have been explored in various tissue engineering applications, such as a substrate for targeted cell adhesion and migration [4]. Moreover, they have been utilized as a delivery

0142-9612/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0142-9612(03)00168-6

3202

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

vehicle for cells in conjunction with a bioactive molecule to activate a specific cellular function in a localized region of the body [5]. Many synthetic hydrogels have been studied for tissue engineering applications including copolymers of poly(ethylene glycol) (PEG) [6], poly(acrylic acid) derivatives [7], polypeptides [8], poly(vinyl alcohol) [9] and poly(phosphazene) [10]. Although the use of synthetic hydrogels appears to be promising, there are several limitations to be overcome for the engineering of a living tissue [2]. It is imperative that the synthetic hydrogel is biocompatible. Additionally, the degradation of the hydrogel should match the tissue formation rate. Finally, the hydrogel needs to be bioactive and capable of interacting with surrounding cells thus regulating specific cellular functions and guiding tissue development. In our laboratory, we have previously synthesized a novel linear unsaturated macromer, oligo(poly(ethylene glycol) fumarate) (OPF), which consists of two repeating units, PEG and fumaric acid that are alternatively linked to each other by ester bonds. Hydrolysis of the ester bond allows for the degradation of OPF. OPF possesses multiple double bonds that enable the formation of a crosslinked network in the presence of a watersoluble redox initiator via radical polymerization [11]. OPF hydrogels have been further investigated as an in situ crosslinkable biomimetic scaffold by covalently incorporating cell-binding peptides into the hydrogel. Modulation of cell attachment to the peptide-modified hydrogel has been accomplished by altering the macromolecular structure of hydrogels [12]. The current study was designed to assess the in vivo bone and soft tissue behavior to OPF hydrogels using a rabbit model. Four groups of hydrogel implants were synthesized with different macromolecular structures and compositions by varying the hydrogel mesh size, crosslinking density, and cell adhesive peptide content. We asked the following questions: (1) Does the PEG block length of OPF affect in vivo bone and soft tissue response to OPF hydrogels? (2) Does the crosslinking density affect the in vivo bone and soft tissue response to OPF hydrogels? (3) Does the cell adhesive peptide content affect in vivo bone and soft tissue response to bulk modified OPF hydrogels?

2. Materials and methods 2.1. Reagents PEG (nominal molecular weight 1.0 and 8.0 K), PEGdiacrylate (PEG-DA) (nominal molecular weight 575), ammonium persulfate (APS), and triethylamine (TEA) were purchased from Aldrich (Pittsburgh, PA). Fumaryl chloride was obtained from Acros (Milwaukee, WI) and

distilled prior to use. Ascorbic acid (AA) was purchased from Sigma (Saint Louis, MO). Acryloyl-PEG-Nhydroxysuccinimide (Acryloyl-PEG-NHS) (molecular weight 3.4 K) was obtained from shearwater polymers (Huntsville, AL). The model peptide, Gly-Arg-Gly-Asp (GRGD) was purchased from Bachem California (Torrance, CA). Other solvents used in the study were of reagent grade. All reagents were used without further purification unless specified. 2.2. Synthesis of oligo(PEG fumarate) Oligo(PEG fumarate) (OPF) was synthesized from PEG of nominal molecular weight of 1.0 and 8.0 K following an established procedure [11]. Briefly, PEG was dehydrated by azeotropic distillation and dissolved in anhydrous tetrahydrofuran. TEA and distilled fumaryl chloride were dropped concurrently into the dehydrated PEG solution that was placed in an ice bath. The reagents were mixed vigorously at room temperature overnight. Following solvent evaporation, the resulting product was recrystalized in ethyl acetate and precipitated in anhydrous ethyl ether. Remaining solvents were removed at 0.1 Torr for 5 h. The number average molecular weight of PEG and OPF were determined with gel permeation chromatography (GPC) based on a calibration curve from monodispersed PEG standards. The purified OPF was stored at 0
C prior to use. Acrylated-GRGD was prepared as previously described [12]. Briefly, GRGD was dissolved in a sodium bicarbonate buffer solution (pH 8.3). Acryloyl-PEGNHS was reacted with the peptide under stirring at room temperature for 2.5 h. The reaction mixture was dialyzed in distilled deionized water (DDW) for 2 days using a dialysis membrane to remove any unreacted peptide. The dialyzed polymer solution was lyophilized overnight and stored at 0
C prior to use. 2.3. Preparation of OPF hydrogel implants Four groups of OPF hydrogel implants were examined by varying the hydrogel mesh size, crosslinking density, and peptide content. The different groups of OPF hydrogel implants are presented in Table 1 and a total of 192 implants were prepared. Specifically, OPF 11, OPF 1-2, and OPF 1-3 hydrogel implants were synthesized using OPF with PEG nominal molecular weight of 1.0 K and OPF 8-1 implant was synthesized using OPF with PEG nominal molecular weight of 8.0 K. OPF hydrogels were chemically crosslinked with PEG-DA in the presence of water-soluble initiators in the following manner: OPF and PEG-DA were dissolved in DDW at 50% (w/v). The amount of added PEG-DA was determined by the ratio of number of double bonds in PEG-DA to OPF (DBR). For example,

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

3203

Table 1 Description of OPF hydrogel implants Group

Implant name

PEG nominal MW

PEG MWa

OPF MWa

DBRb

Acrylated peptide (mmol/g)

1 2 3 4

OPF OPF OPF OPF

1000 1000 1000 8000

930710 930710 930710 6090790

4470760 4470760 4470760 144307480

1 5 1 1

0 0 1.0 0

a b

1-1 1-2 1-3 8-1

The number average molecular weight of PEG and OPF was determined using GPC and a calibration curve with mono-dispersed PEG standards. DBR: ratio of double bonds present in PEG-DA to those in OPF.

0.041 ml PEG-DA were added to 0.2 g OPF 1.0 K in 0.492 ml DDW for the preparation of OPF 1-1 hydrogel implant that was crosslinked with a DBR equal to 1. The number of double bond in OPF was calculated based on the number average molecular weight of synthesized OPF. The DBR of each sample group is presented in Table 1. For OPF 1-3 hydrogel, 1 mmol/g (mmol acrylated peptide/weight of hydrogel after swelling) was added to the aqueous mixture of OPF and PEG-DA. The redox initiators, APS and AA, were added to the mixture in equal concentration of 0.025m to initiate the crosslinking reaction. Immediately, the aqueous solution was injected into a cylindrical polystyrene mold using a 10 ml syringe and crosslinked at 45
C for 30 min. A 5.8 mm in diameter and 2.1 mm in thickness mold was used for the preparation of OPF 1-1, OPF 1-2, and OPF 1-3 hydrogels. A 3.35 mm in diameter and 1.4 mm in thickness mold was used for the OPF 8-1 hydrogel to produce the same implant size following rinsing with DDW and sterilization with 70% (v/v) ethanol solution. After crosslinking, the hydrogel implants were removed from the mold and immersed in DDW for 1 week in order to remove any remaining unreacted oligomers. The dimension of the fully swollen hydrogel implants was 6.3 mm in diameter and 2.2 mm in thickness. For sterilization, the hydrogels were soaked in 70% ethanol solution (v/v) for 3 days and re-swollen in sterile phosphate buffered saline (PBS) for 3 days, changing the media twice per day under sterile conditions.

2.4. In vitro degradation of OPF hydrogels For the in vitro degradation study, OPF hydrogels were synthesized as the in vivo implants. However, following the crosslinking reaction, OPF hydrogels were dried and weighed (Wi ) instead of being rinsed with DDW. The hydrogels were then placed in PBS solution (pH 7.4) and stored in a 37
C environment on a shaker table (80 rpm). The PBS solution was changed everyday for the first week and weekly for the rest of the study. The weight loss of each specimen and change in medium pH were monitored over time. At 1, 2, 3, 4, 6, 8, and 12 weeks, specimens were removed from the media,

weighed (Ws ), and dried in a vacuum overnight, and then re-weighed (Wd ). At the same time points, pH of the PBS solution was measured and the difference from 7.4 was recorded. The percent weight loss and the swelling ratio of OPF hydrogels were calculated with the following equations: % Weight loss ¼ Swelling ratio ¼

Wi  Wd  100; Wi

Ws  Wd : Wd

The difference of the swelling ratio at each time point to that prior to degradation was then determined. 2.5. Implantation study Twelve healthy skeletally mature female New Zealand White rabbits with an average weight of 2.8–3.2 kg were used as experimental animals for subcutaneous and cranial implantation. The rabbits were divided into two groups of 6 animals depending on the implantation period, 4 weeks and 12 weeks, respectively. The implants were inserted into the parietal cranial bone and the back of the rabbits. The surgery was performed under general anaesthesia, which was maintained by the inhalation of isoflurane. The anaesthesia was induced by an intravenous injection of Hypnorms (0.315 mg/ml fentanyl citrate and 10 mg/ml fluanisone) and atropine, and maintained by a mixture of nitrous oxide, isoflurane, and oxygen through a constant volume ventilator. To reduce the perioperative infection risk, the rabbits received antibiotic prophylaxis (Baytril 2.5% (Enrofoxacin), 5–10 mg/kg). The animals were placed in a ventral position and immobilized on their abdomen for the surgery. For the insertion of the implants into the parietal cranial bone, the skull was shaved and disinfected with povidineiodine. A longitudinal incision was made down to the periosteum from the nasal bone to the occipital protuberance, and soft tissues were sharp dissected to visualize the cranial periosteum. Subsequently, a midline incision was made in the periosteum. Thereafter, the periosteum was undermined and lifted off the parietal skull. Two holes were drilled in the left parietal bone and

3204

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

two holes in the right parietal bone. The bone was drilled using a sterile hollow drill (Merck, Darmstadt, Germany) with an outer diameter of 6.3 mm. In this way, the drill made a full thickness parietal skull defect without damaging the underlying dura. The hole was prepared with low rotation speed (max. 450 rpm) and continuous internal and external cooling with physiologic saline. Following insertion of the OPF hydrogel implants, the soft tissues were closed in separate layers using resorbable vicryl 3-0 sutures. For the insertion of the subcutaneous implants, the dorsum of the rabbits was shaved, washed, and disinfected with iodine. Paravertebrally, at both sides of the spinal column, four longitudinal incisions of about 1.5 cm were made through the full thickness of the skin. Subsequently, lateral to the incisions a subcutaneous pocket was created by blunt dissection with scissors. The implants were inserted in these pockets. Finally, the wound was carefully closed intracutaneously with vicryl 3-0. Each rabbit received 4 cranial and 4 subcutaneous implants. A total of 192 OPF hydrogels were implanted into 12 rabbits. The location of hydrogel implants was varied in the following way. First of all, four implants were placed in either the carnial or dorsal surgical field in a 2  2 grid, with each implant noted. For the subsequent rabbit, the location of each implant was moved clockwise. Therefore, the animal anatomy did not affect the outcome and the four hydrogel implants were exposed to all four possible locations in the surgery field. At 4 weeks post-implantation, 6 rabbits were euthanized by an overdose of Nembutals and the implants with surrounding tissue were retrieved. The remaining 6 rabbits were euthanized similarly at 12 weeks postimplantation. Implants from 6 rabbits at 4 weeks and at 12 weeks were randomly chosen, sectioned, and analyzed for histology. During the study, the Dutch and US National Institutes of Health guidelines for the care and use of laboratory animals were observed. 2.6. Histological evaluation Immediately after retrieval, specimens were fixed in a 10% buffered formalin solution. Subsequently, the tissue blocks were dehydrated in a series of ethanol (from 10% (v/v) to 100%) and embedded in methylmethacrylate. After polymerization, non-decalcified thin (10 mm) sections were prepared in a transversal direction to the axis of the implant using a modified sawing microtome technique. Three histological sections were made for each tissue block. The sections were stained with methylene blue/basic fuchsin and examined using a light microscope (Leica BV, Rijswijk, The Netherlands). A light microscope was used for the histological evaluation. The histological evaluation consisted of a

Table 2 Histological grading scale for implants. The thickness of fibrous capsule around the implants was scored by counting the number of fibroblast layers Number of fibroblast layers

Score

1–4 5–9 10–30 >30 Defects on sectioning

4 3 2 1 Not applicable

concise description of the observed specimens and a quantitative scoring analysis of the tissue response [13]. The scoring system is presented in Table 2 and based on fibrous capsule formation as measured by the number of fibroblast layers surrounding the implants. The histological evaluation was done in six randomly determined fields along the implant–tissue interface. The quantitative measurements were performed blindly for these randomly chosen sections of each implant. The number of inflammatory cells and multinuclear cells were counted using a microscope at a magnification of  40. 2.7. Statistical analysis Single factor analysis of variance (ANOVA) and a Tukey’s HSD (Honestly Significantly Different) multiple comparison test were used to determine statistical significance of results with 95% confidence intervals (po0:05). The results are reported as means7standard deviation.

3. Results 3.1. In vitro degradation of OPF hydrogels Overall, the degradation profiles of OPF hydrogels showed a immediate weight loss during the first week of the degradation study. OPF 1-1, OPF 1-3, and OPF 8-1 hydrogels showed significantly more weight loss than OPF 1-2 hydrogel after 12 weeks (Fig. 1a). For example, the weight loss of OPF 1-1 hydrogels was 10.976.1% at 1 week, which was increased to 27.971.8% at 12 weeks. However, the weight loss of OPF 1-2 hydrogels was still 13.671.8% at 12 weeks. In Fig. 1b, the OPF 8-1 hydrogel exhibited a significant increase of swelling ratio over 12 weeks. The difference of swelling of OPF 8-1 hydrogel compared to that before degradation was 20.973.6 at 12 weeks, while the swelling change of OPF 1-1, OPF 1-2, and OPF 1-3 hydrogels ranged from 1.070.1 to 3.171.2 at 12 weeks. No significant change in the pH of the PBS solution was observed for OPF 1-2 and OPF 8-1 hydrogels

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

3205

40 35

Weight Loss (%)

30 25 20 15 10 5 0 0

5

15

10

Degradation Time (weeks)

(a)

40 35

5

25

4

20

Swelling Variation

Swelling Variation

30

15 10 5 0 -5

3 2 1 0 -1

-10 0

2

(b)

4

6

8

10

12

Degradation Time (weeks)

14

0

5

10

15

Degradation Time (weeks)

Fig. 1. (a) Percent weight loss compared to initial weight before degradation as a function of degradation time for OPF 1-1 (&), OPF 1-2 (J), OPF 1-3 (n), and OPF 8-1 (B). (b) Variation in swelling ratio compared to initial swelling before degradation as a function of degradation time for OPF 1-1 (&), OPF 1-2 (J), OPF 1-3 (n), and OPF 8-1 (B). The description of the OPF hydrogel implants is presented in Table 1. Error bars represent means 7 standard deviation for n ¼ 3:

0.6

3.2. Descriptive light microscopic evaluation

0.4

pH Variation

0.2 0 -0.2 -0.4 -0.6 -0.8 0

5

10

15

Degradation Time (weeks) Fig. 2. Variation in pH of the PBS solution with degradation time for OPF 1-1 (&), OPF 1-2 (J), OPF 1-3 (n), and OPF 1-8 (B). The description of the OPF hydrogel implants is presented in Table 1. Error bars represent means7standard deviation for n ¼ 3:

during 12 weeks of degradation study (Fig. 2). OPF 1-1 hydrogel exhibited a rapid drop of pH during the first 3 weeks. OPF 1-3 showed a moderate change in pH ranging from 0.270.1 to 0.170.0.

Examination of the histological sections of the subcutaneous implants revealed that the tissue response to all OPF hydrogel implants was fairly uniform at 4 weeks. A thin circumferential fibrous capsule was formed around the OPF hydrogel implants. The capsule contained a very limited number of inflammatory cells and typically ranged from 5 to 15 cell layers in thickness. Furthermore, mature fibroblasts and organized connective tissue appeared to surround the implants (Fig. 3a). After 12 weeks, the tissue response to OPF 1-1, 1-2, and 1-3 hydrogels was similar to that observed after 4 weeks. The fibrous capsule surrounding the implants mainly consisted of flattened fibroblasts and a few inflammatory cells. The thickness of the fibrous layer and number of inflammatory cells did not change at this prolonged implantation time. On the other hand, OPF 8-1 hydrogel implant showed evidence of degradation as seen in Fig. 3b and c, which was characterized by an extensive surface erosion of the material associated with penetration of inflammatory cells. Also, some superficial fragmentation was observed. The fragments were

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

3206

(a)

H IC FC CT

F

100 µm

(b)

H

surrounded by inflammatory cells and some macrophages. The histological sections of the cranial implants revealed similar trends as observed for the subcutaneous implants. It was noted that the implants retained their original shape and stayed in the defects during the implantation period. There was no sign of breakage or deformation of the implants. The tissue response to the OPF hydrogel implants was again relatively uniform after 4 weeks of implantation (Fig. 4). A thin layer of fibrous capsule developed around the implants, which prevented the direct contact of the implant surface with native or newly formed bone at the cranial defect perimeter (Fig. 4). There was a very limited number of inflammatory and multinuclear cells at the

IC

(a)

F 200 µm

B

H

FC

(c)

H

IC 100 µm

MN

IC 100 µm

Fig. 3. Representative histological sections of subcutaneous OPF hydrogel implants. OPF hydrogel (H) appears in all the images as unstained area. (a) OPF 1-3 hydrogel after 4 weeks of implantation, original magnification  40, (b) OPF 8-1 hydrogel after 12 weeks of implantation, original magnification  10 and (c)  40. The connective tissue (CT) exhibits a dense and organized structure regardless of the implantation period. The uniform thin fibrous tissue encapsulation (FC) that forms along the perimeter of OPF hydrogel implants is clearly visible in (a). The fibrous capsule consists of several layers of fibroblasts (F). A minimal number of inflammatory cells (IC) is observed around all types of OPF hydrogels except for OPF 8-1 hydrogel implants. OPF 8-1 hydrogel implants (b) exhibit extensive degradation with penetration of inflammatory cells. Arrows in (b) indicate fragmentation of OPF 8-1 hydrogel implants. The fragments are surrounded by inflammatory and multinuclear cells, which are clearly visible at the higher magnification image (c).

(b)

H IC FC F

100 µm

Fig. 4. Representative histological sections of cranial OPF hydrogel implants. (a) OPF 1-2 hydrogel at bone-biomaterial interface after 4 weeks of implantation, original magnification  40. (b) OPF 1-2 hydrogel at an endocranial side after 4 weeks of implantation, original magnification  40. OPF hydrogel (H) appears as unstained area. Intact bone (B) is clearly visible in the image. A thin fibrous capsule (FC) is formed, which generally prevents direct contact between OPF hydrogel implants and intact bone. Inflammatory cells (IC) and fibroblasts (F) are also seen.

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

implant–fibrous capsule interface. Despite the presence of a fibrous capsule, new bone formation was occasionally observed at the endocranial side of the implant, which was guided by the OPF hydrogel implant as shown in Fig. 5. After 12 weeks of implantation, the fibrous capsule did not change in thickness and appearance for the majority of implants except for OPF 8-1 hydrogel implants. Similarly as observed for the subcutaneous implants, the OPF 8-1 hydrogel demonstrated accumulation of inflammatory and multinuclear cells at the implant–tissue interface with infiltration of cells inside the implants (Fig. 6). 3.3. Quantitative histological analysis Fig. 7 shows the histological scoring of the fibrous capsule thickness for both subcutaneous and cranial

3207

implants. No statistical difference existed between implant groups with respect to implantation site or implantation period. The average score of the fibrous capsule for the implants ranged from 2.770.2 to 3.170.2. Fig. 8 shows the number of inflammatory and multinuclear cells that were counted along the implant perimeter for the different groups. For the cranial implants, the number of inflammatory cells was significantly increased for the OPF 8-1 hydrogel compared to the other implant groups both after 4 and 12 weeks of implantation (po0:05). Additionally, a significant increase in the number of multinuclear cells was noted between 4 and 12 weeks of implantation. For the subcutaneous implants, there was a statistical difference in the number of inflammatory and multinuclear cells for the OPF 8-1 hydrogel (po0:05) among the various OPF groups only after 12 weeks of implantation.

4. Discussion H FC

B

200 µm

Fig. 5. Histological section of OPF 1-1 hydrogel after 4 weeks of implantation. New bone (B) formation is guided along OPF hydrogel (H) surface at the endocranial side of the defect. A thin fibrous capsule (FC) surrounds the implant.

The objective of this study was to examine in vivo bone and soft tissue response to OPF hydrogels. Particularly, we evaluated the effects of the crosslinking density, the PEG block length of OPF, and the RGD cell-binding peptide content on degradation and in vivo tissue response of OPF hydrogels. In vivo response to the OPF hydrogels was assessed histologically following the implantation at two different anatomical locations in a rabbit. The weight loss and swelling ratio of OPF hydrogels over 12 weeks in PBS solution at 37
C were measured to investigate the effect of crosslinking density, PEG block length and RGD peptide content on the in vitro degradation of OPF hydrogels. The weight loss of

(a)

(b)

H

H MN

IC F IC B 200 µm

FC

100 µm

B Fig. 6. Histological sections of cranial OPF 8-1 hydrogel implants after 12 weeks of implantation; (a) original magnification  10 and (b)  40. OPF hydrogels (H) appears as unstained area. Inflammatory cells (IC) and multinuclear macrophages (MN) infiltrate into the hydrogel (H).

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

3208

Histological Scoring of Fibrous Capsule

4

3

2

1

0 (a)

OPF 1-1

OPF 1-2

OPF 1-3

OPF 8-1

OPF 1-1

OPF 1-2

OPF 1-3

OPF 8-1

Histological Scoring of Fibrous Capsule

4

3

2

1

0 (b)

Fig. 7. Quantitative scoring of tissue response to OPF hydrogel implants for the cranial defect (a) and the subcutaneous defect (b) after 4 weeks (white) and 12 weeks of implantation (black). A thin layer of fibrous capsule formation along the OPF hydrogel implants was observed. The fibrous capsule consisted of 5–15 layers of fibroblasts. Error bars represent means7standard deviation for n ¼ 6:

OPF hydrogels increased with decreasing crosslinking density without changing the PEG block length of OPF. However, the change in PEG block length from 1.0 to 8.0 K and incorporation of RGD peptide into the hydrogels did not significantly affect the weight loss. We have already shown that the crosslinking density of OPF hydrogels increased with increasing DBR [12]. Sawhney et al. demonstrated that the degradation of hydrogels was highly affected by the crosslinking density [14]. Another report showed that photo-crosslinked poly(ether-ester) networks were degraded in a manner that low crosslinking density increased degradation rate [15]. These results are consistent with our weight loss data at 12 weeks suggesting that the crosslinking density inversely affects the degradation rate of OPF hydrogels. The peptide-modified OPF hydrogel was synthesized with 1.0 mmol acrylated-GRGD peptide on the weight basis since marrow stromal osteoblasts attached and formed cytoskeletal organization on the hydrogel at this concentration in vitro [12]. The presence of peptide did not cause significant changes in degradation characteristics of the hydrogel because the amount of peptide within the crosslinked network (0.5% w/w) was

relatively small compared to the total weight of OPF and PEG-DA. The change in swelling ratio of OPF hydrogels over 12 weeks exhibited a similar trend as observed for weight loss of specimens except for the OPF 8-1 hydrogel. A significant increase in swelling for these hydrogels was observed over 12 weeks and the physical appearance of OPF 8-1 hydrogels became more jelly-like and softer than the other specimens. The swelling ratio of OPF 1-1 and OPF 1-3 hydrogels was significantly higher than that of OPF 1-2 hydrogels. The swelling ratio reflects the macromolecular structure of the crosslinked network and degradation of the polymer chains results in an increase in swelling ratios of hydrogels [16,17]. For OPF 8-1 hydrogels, the total number of double bonds and ester bonds within the crosslinked network was relatively lower compared to the hydrogels synthesized with OPF 1.0 K. Therefore, the breakdown of few ester groups in OPF 8-1 hydrogels enabled a significant increase of water uptake. The increase in swelling was limited for the hydrogels synthesized with OPF 1.0 K since there were a greater number of crosslinks per macromer chain. The degradation of the OPF hydrogels is attributed to the cleavage of ester bonds within the crosslinked network by hydrolysis. Similar hydrolytic degradations have been reported for crosslinked hydrogels as well as for non-crosslinked poly(a-hydroxy esters) [18]. The hydrolytic cleavage of ester groups results in the release of acidic degradation byproducts [18,19]. Since one of byproducts following degradation of OPF is fumaric acid [20], a decline in pH of the PBS solution was observed. The results indicate that the changes in pH corresponded to that of hydrogel weight loss. For example, OPF 1-2 hydrogels exhibited a delayed weight loss and little change in pH over 12 weeks compared to other hydrogel groups. The overall number of ester groups present in OPF 8-1 hydrogel network is one order of magnitude lower than that in the other formulations. Therefore, the change in pH for OPF 81 hydrogels was minimal although the weight loss of OPF 8-1 hydrogels was comparable with that of OPF 11, and OPF 1-3 hydrogels. The tissue response to the OPF hydrogels was characterized with a histological grading system that has been used to analyze the biomaterial–tissue response in a reproducible and comparative manner [13,21]. The histological grading of the subcutaneous implants indicated that the OPF hydrogels evoked a restricted tissue response regardless of crosslinking density, PEG block length of OPF, incorporation of peptide, and implantation period. A limited fibrous capsule formation was observed and the thickness of the fibrous tissue layer was consistent at the prolonged implantation period. Also, a low number of inflammatory cells and macrophages were seen at the implant–tissue interface

150

* 100 50 0

Number of Inflammatory Cells

(a)

(c)

*

200

OPF 1-1

OPF 1-2

OPF 1-3

OPF 8-1

250 200

*

150 100 50 0 OPF 1-1

OPF 1-2

OPF 1-3

OPF 8-1

Number of Multinuclear Cells

250

3209

25

*

20 15 10

* 5 0

(b)

Number of Multinuclear Cells

Number of Inflammatory Cells

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

OPF 1-1

OPF 1-2

OPF 1-3

25

OPF 8-1

*

20 15 10

(d)

5 0 OPF 1-1

OPF 1-2

OPF 1-3

OPF 8-1

Fig. 8. Number of inflammatory cells (a, c) and multinuclear cells (b, d) surrounding OPF hydrogel implants after 4 (white) and 12 weeks of implantation (black); (a, b) represents results from cranial OPF hydrogel implants and (c, d) represents results from subcutaneous OPF hydrogel implants. Error bars represent means7standard deviation for n ¼ 6 and the symbol ‘‘*’’ indicates a significant difference from the other three implant types at each time point (po0:05).

for all materials except for OPF 8-1 hydrogels, which showed relatively higher number of inflammatory cells at 12 weeks post-implantation. The persistent stay of inflammatory cells and macrophages indicate that the OPF hydrogels are undergoing active biodegradation [22,23]. For the subcutaneous OPF 8-1 hydrogel implants, fragmentation of hydrogels and extensive surface erosion by infiltration of inflammatory cells were observed while histological observation of OPF 11, OPF 1-2, and OPF 1-3 hydrogels showed minimal evidence of degradation. Although the hydrolytic degradation is active by the cleavage of ester groups within the crosslinked network, the tissue response in our study mainly deals with the surface interaction of the implants with surrounding tissue. This may explain that hydrogel implants prepared with OPF 1.0 K showed minimal evidence of degradation histologically despite significant in vitro weight loss at 12 weeks. In contrast, the large hydrophilic chain (longer PEG block length of OPF) in the network presumably enhanced tissue infiltration into the hydrogel implants [24]. The results also indicated that in vivo degradation products for these hydrogels elicited a mild local tissue response. For implantation of OPF hydrogels in a cranial region, we created a non-critical sized 6.3 mm-diameter defect in the rabbit skull for two reasons [25,26]. Firstly, the purpose of cranial implantation in this study was to

examine the general tissue response of OPF hydrogels that occurs at the bone–implant interface. Secondly, this allows the placement of four implants bilaterally per skull, which maximizes the comparison of experimental parameters, reduces the total number of animals, and minimizes the experimental variation between groups [27]. The tissue response to OPF hydrogel implants in the cranial defects showed a similar pattern as observed for the subcutaneous specimens. Only more inflammatory cells were observed for OPF 8-1 hydrogel implants at 4 weeks. Although a defined explanation is difficult to give, we assume that local tissue condition such as vascularization in skull bone varies compared with subcutaneous tissue and may have caused this difference. The rabbit cranial defect has been widely used as an implantation model in order to investigate bone formation using various types of tissue engineered constructs [28–30]. Although OPF hydrogels absorb large amount of biological fluids, it should be noted that the cranial hydrogel implants maintained their original shape and fit in the bone defect after 12 weeks of implantation. All implants stayed in the originally created defects without major displacement during the implantation. The RGD peptide modification appears to have limited effect on the in vivo tissue response. The fibrous capsule formation and persistence of inflammatory cells

3210

H. Shin et al. / Biomaterials 24 (2003) 3201–3211

around OPF 1-3 hydrogel implants were similar quantitatively and qualitatively as observed for other OPF hydrogels synthesized with OPF 1.0 K. We have already demonstrated that RGD peptide-modified fumarate-based hydrogels enhanced marrow stromal cell adhesion in vitro while the hydrogels without modification minimized cell adhesion [12,31]. In addition, there have been a number of reports that investigated the effect of cell-binding peptides on interaction with bone forming cells, although mostly using surface modified materials [32–34]. As opposed to in vitro experiments that can be performed under relatively well-controlled conditions of cell seeding density and concentration of media supplements, numerous interactions occur at the biomaterial–tissue interface in vivo. Several in vitro experiments demonstrated that the interaction of receptor binding domain with cells is concentration dependent [12,35]. However, the optimal concentration and spatial distribution of incorporated peptides in biomimetic materials in vivo have yet been clearly identified. In addition, the RGD model peptide sequence is ubiquitously present in ECM proteins and can be recognized by many cell types, which may limit the specific interaction between bone cells and the peptide. However, the in vivo experiments clearly indicated that OPF hydrogels allowed limited fibrous tissue encapsulation. Further studies are needed to investigate the effect of biomimetic peptides on specific tissue response.

5. Conclusions OPF hydrogels exhibited degradation in PBS at 37
C over 12 weeks of incubation, which was measured by the weight loss and swelling ratio of hydrogels and the pH change in PBS solution. The weight loss of OPF hydrogels was dependent on their crosslinking density. A significant weight loss was observed with decreasing crosslinking density. The change in the PEG block length of OPF did not influence the weight loss of hydrogels; however, the swelling ratio of crosslinked OPF network increased for the hydrogels synthesized with a longer PEG chain. A limited fibrous capsule formation was observed around the subcutaneous as well as cranial implants of which the thickness remained consistent during implantation confirming that chemically crosslinked OPF hydrogels evoked a mild tissue response. Further, histology revealed that in contrast to the in vitro degradation, only OPF hydrogels made of higher molecular weight PEG exhibited extensive fragmentation and tissue infiltration by inflammatory cells. These results showed that OPF-based hydrogels hold promise for use as a synthetic biodegradable scaffolds for tissue engineering.

Acknowledgements This work was supported by the grant from the National Institutes of Health (R01 DE13031) (AGM).

References [1] Vacanti JP, Langer R. Tissue engineering: the design and fabrication of living replacement devices for surgical reconstruction and transplantation. Lancet 1999;354(Suppl. I):32–4. [2] Lee KY, Mooney DJ. Hydrogels for tissue engineering. Chem Rev 2001;101:1869–79. [3] Nuttelman CR, Mortisen DJ, Henry SM, Anseth KS. Attachment of fibronectin to poly(vinyl alcohol) hydrogels promotes NIH3T3 cell adhesion, proliferation, and migration. J Biomed Mater Res 2001;57:217–23. [4] Hubbell JA. Biomaterials in tissue engineering. Bio/technology 1995;13:565–76. [5] Murphy WL, Mooney DJ. Controlled delivery of inductive proteins, plasmid DNA, and cells from tissue engineering matrices. J Periodon Res 1999;34:413–9. [6] Bae Y, Huh K, Kim Y, Park K. Biodegradable amphiphilic multiblock copolymers and their implications for biomedical applications. J Control Release 2000;64:3–13. [7] Meyvis T, Smedt SD, Stubbe B, Hennink W, Demeester J. On the release of proteins from degrading dextran methacrylate hydrogels and the correlation with the rheologic properties of the hydrogels. Pharm Res 2001;18:1593–9. [8] Panitch A, Yamaoka T, Fournier MJ, Mason TL, Tirrell DA. Design and biosynthesis of elastin-like artificial extracellular matrix proteins containing periodically spaced fibronectin CS5 domains. Macromolecules 1999;32:1701–3. [9] Nuttelman CR, Henry SM, Anseth KS. Synthesis and characterization of photocrosslinkable, degradable, poly(vinyl alcohol)based tissue engineering scaffolds. Biomaterials 2002;23:3617–26. [10] Laurencin CT, El-Amin SF, Ibim SE, Willoughby DA, Attawia M, Allcock HR, Ambrosio AA. A highly porous 3-dimensional polyphosphazene polymer matrix for skeletal tissue regeneration. J Biomed Mater Res 1996;30:133–8. [11] Jo S, Shin H, Fisher JP, Shung AK, Mikos AG. Synthesis and characterization of oligo(poly(ethylene glycol)fumarate) macromer. Macromolecules 2001;34:2839–44. [12] Shin H, Jo S, Mikos AG. Modulation of marrow stromal osteoblast adhesion on biomimetic oligo(poly(ethylene glycol) fumarate) hydrogels modified with Arg-Gly Asp peptides and a poly(ethylene glycol) spacer. J Biomed Mater Res 2002;61:169–79. [13] Jansen JA, Dhert WJA, Waerden JPCM, Recum AF. Semiquantitative and qualitative histologic analysis method for the evaluation of implant biocompatibility. J Invest Surg 1994;7: 123–34. [14] Sawhney AS, Pathak CP, Hubbell JA. Bioerodible hydrogels based on photopolymerized poly(ethylene glycol)-co-poly (a-hydroxy acid) diacrylate macromers. Macromolecules 1993;26: 581–7. [15] Kim B, Hrkach JS, Langer R. Biodegradable photo-crosslinked poly(ether-ester) networks for lubricious coating. Biomaterials 2000;21:259–65. [16] Hasirci V, Lewandrowski K, Gresser JD, Wise DL, Trantolo DJ. Versatility of biodegradable biopolymers: degradability and an in vivo application. J Biotech 2001;86:135–50. [17] Lee KY, Bouhadir KH, Mooney DJ. Degradation behavior of covalently cross-linked poly(aldehyde guluronate) hydrogels. Macromolecules 2000;33:97–101.

H. Shin et al. / Biomaterials 24 (2003) 3201–3211 [18] Lu L, Garcia CA. In vitro degradation of thin poly (dl-lactic-co-glycolic acid) films. J Biomed Mater Res 1999;46: 236–44. [19] He S, Timmer MD, Yaszemski MJ, Yasko AW, Engel PS, Mikos AG. Synthesis of biodegradable poly(propylene fumarate) networks with poly(propylene fumarate)-diacrylate monomers a crosslinking agents and characterization of their degradation products. Polymer 2001;42:1251–60. [20] Timmer MD, Jo S, Wang C, Ambrose CG, Mikos AG. Characterization of the crosslinked structure of fumaratebased degradable polymer networks. Macromolecules 2002;35: 4373–9. [21] Jansen JA, Ruijter JEd, Janssen PTM, Paquay YGCJ. Histological evaluation of a biodegradable polyactive/hydroxyapatite membrane. Biomaterials 1995;16:819–27. [22] Marchant R, Hiltner A, Hamlin C, Rabinovitch A, Slobodkin R, Anderson JM. In vivo biocompatibility studies, I. the cage implant system and a biodegradable hydrogel. J Biomed Mater Res 1983;17:301–25. [23] Goissis G, Junior EM, Marcantonio RAC, Lia RCC, Cancian DCJ, Carvalho WM. Biocompatibility studies of anionic collagen membranes with different degree of glurataldehyde cross-linking. Biomaterials 1999;20:27–34. [24] Bulpitt P, Aeschlimann D. New strategy for chemical modification of hyaluronic acid: preparation of functionalized derivatives and their use in the formation of novel biocompatible hydrogels. J Biomed Mater Res 1999;47:152–69. [25] Vehof JWM, Haus MTU, Ruijter JE, Spauwen PHM, Jansen JA. Bone formation in transforming growth factor beta-I-loaded titanium fiber mesh implants. Clin Oral Implants Res 2002;13: 94–102. [26] Schmitz JP, Hollinger JO. The critical size defect as an experimental model for craniomandibulofacial nonunions. Clin Orthop 1986;205:299–308.

3211

[27] Fisher JP, Vehof JWM, Dean D, Waerden JPCM, Holland TA, Mikos AG, Jansen JA. Soft and hard tissue response to photocrosslinked poly(propylene fumarate) scaffolds in a rabbit model. J Biomed Mater Res 2001;59:547–56. [28] Vehof JWM, Fisher JP, Dean D, Waerden JPCMvd, Spauwen PHM, Mikos AG, Jansen JA. Bone formation in transforming growth factor b-1 coated porous poly(propylene fumarate) scaffolds. J Biomed Mater Res 2002;60:241–51. [29] Hong L, Tabata Y, Yamamoto M, Miyamoto S, Yamada K, Hashimoto N, Ikada Y. Comparison of bone regeneration in a rabbit skull defect by recombinant human BMP-2 incorporated in biodegradable hydrogel and in solution. J Biomater Sci Polym Ed 1998;9:1001–14. [30] Robinson BP, Hollinger JO, Szachowicz EH. Calvarial bone repair with porous D,L-polylactide. Otolaryngol Head Neck Surg 1995;112:707–13. [31] Behravesh E, Zygourakis K, Mikos AG. Adhesion and migration of marrow derived osteoblasts on injectable in situ crosslinkable poly(propylene fumarate-co-ethylene glycol) based hydrogels with a covalently linked RGDS peptide. J Biomed Mater Res 2003;65A:261–71. [32] Ferris DM, Moodie GD, Dimond PM, Gioranni CWD, Ehrlich MG, Valentini RF. RGD-coated titanium implants stimulate increased bone formation in vivo. Biomaterials 1999;20:2323–31. [33] Dee KC, Anderson TT, Bizios R. Design and function of novel osteoblast-adhesive peptides for chemical modification of biomaterials. J Biomed Mater Res 1998;40:371–7. [34] Rezania A, Healy KE. Biomimetic peptide surfaces that regulate adhesion, spreading, cytoskeletal organization, and mineralization of the matrix deposited by osteoblast-like cells. Biotech Prog 1999;15:19–32. [35] Schense JC, Hubbell JA. Three-dimensional migration of neurites is mediated by adhesion site density and affinity. J Biol Chem 2000;275:6813–8.