Document [original]

Hard Tissues and Materials

Fabrication, in vitro and in vivo

studies of bilayer composite

membrane for periodontal

guided tissue regeneration

Saba Zahid

, Abdul Samad Khan

, Aqif Anwar Chaudhry

Sarah Ghafoor

, Qurat Ul Ain

, Ahtasham Raza

Muhammad Imran Rahim

, Oliver Goerke

Ihtesham Ur Rehman

and Asma Tufail

Abstract

Development of a guided occlusive biodegradable membrane with controlled morphology in order to restrict the

ingrowth of epithelial cells is still a challenge in dental tissue engineering. A bilayer membrane with a non-porous

upper layer (polyurethane) and porous lower layer (polycaprolactone and bioactive glass composite) with thermoelastic

properties to sustain surgery treatment was developed by lyophilization. Morphology, porosity, and layers attachment

were controlled by using the multi-solvent system. In vitro and in vivo biocompatibility, cell attachment, and cell pro-

liferation were analyzed by immunohistochemistry and histology. The cell proliferation rate and cell attachment results

showed good biocompatibility of both surfaces, though cell metabolic activity was better on the polycaprolactone-

bioactive glass surface. Furthermore, the cells were viable, adhered, and proliferated well on the lower porous bioactive

surface, while non-porous polyurethane surface demonstrated low cell attachment, which was deliberately designed and

a pre-requisite for guided tissue regeneration/guided bone regeneration membranes. In addition, in vivo studies per-

formed in a rat model for six weeks revealed good compatibility of membranes. Histological analysis (staining with

hematoxylin and eosin) indicated no signs of inflammation or accumulation of host immune cells. These results suggested

that the fabricated biocompatible bilayer membrane has the potential for use in periodontal tissue regeneration.

Keywords

Guided tissue regeneration, bioactive glass, bilayer membrane, in vitro cytotoxicity, animal testing, immuno-staining

Introduction

Periodontitis, one of the most progressive oral disease,

damages soft tissue and destroys tooth-supporting

bone, often resulting in the loss of a tooth.

The rate

of periodontal breakdown has significantly been

increased in the last few years. In recent years, guided

tissue regeneration/guided bone regeneration (GTR/

GBR) approaches are being used extensively for the

treatment of periodontitis. In both cases, an occlusive

periodontal membrane that acts as a barrier to prevent

down-growth of epithelial and connective tissues into

the defect and enabling periodontal regeneration is

used.

The success of GTR/GBR membranes depends

on achieving the repair of the large area through osteo-

conductivity, mechanical stability, and equilibrium

Interdisciplinary Research Centre in Biomedical Materials, COMSATS

University Islamabad, Lahore Campus, Lahore, Pakistan

Department of Restorative Dental Sciences, College of Dentistry, Imam

Abdulrahman Bin Faisal University, Dammam, Saudi Arabia

Department of Oral Biology, University of Health Sciences Lahore,

Khayaban-e-Jamia Punjab, Lahore, Pakistan

Department of Material Science and Engineering, Kroto Research

Institute, The University of Sheffield, Sheffield, United Kingdom

Fachgebiet Keramische Werkstoffe / Chair of Advanced Ceramic

Materials, Institut fu¨r Werkstoffwissenschaften und -technologien,

Technische Universit€

at Berlin, Hardenbergstr, Berlin, Germany

Corresponding author:

Asma Tufail, COMSATS University Islamabad, Lahore Campus, Lahore-

54000, Pakistan.

Email: drasmashah@cuilahore.edu.pk

Journal of Biomaterials Applications

2019, Vol. 33(7) 967–978

!The Author(s) 2018

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0885328218814986

journals.sagepub.com/home/jba

between the degradation rate of membranes and tissue

regeneration in order to restrict epithelial cells

ingrowth. Both natural (chitosan, collagen, etc.) and

synthetic (polyurethane (PU), polycaprolactone

(PCL), etc.) biopolymers are being used for the devel-

opment of GTR membranes.

3–5

PCL is a biodegrad-

able polyester with good mechanical stability and

bioactivity for endothelial cells. It degrades hydrolyti-

cally, and cells replace the degrading material by con-

tinuously infiltrating the matrix that produces collagen,

elastin, and proteoglycans.

6,7

Biodegradable PU is a

synthetic polymer with good mechanical stability and

tuneable microstructure due to flexible hard and soft

segments.

It has been used for soft tissue and bone

regeneration in vivo and showed good biocompatibility

and biodegradability for GTR membranes. The

requirement of these biopolymers differs according to

the nature of the application.

The inclusion of biocer-

amics as filler in these polymers induces osteoconduc-

tive properties and accelerates the healing process.

Among bioceramics, bioactive glasses (BGs) are

known to have osteoproductive and osteostimulative

properties. They have the ability to form an apatite-

like layer at the interface in vivo by undergoing specific

surface reactions, which leads to the development of a

strong adherent bond with the host hard/soft tissues.

Apatite layer stimulates osteoblasts (bone forming

cells) and promotes new bone formation in situ.

Furthermore, BGs have been used as filler with alginate

and found to prevent the formation of fibrous tissue at

the prosthesis–bone interface and showed in vivo new

bone formation within 7 to 14 days.

However, algi-

nate possesses low intrinsic mechanical strength.

The ability of GTR membranes to seclude epithelial

cells from the targeted area also depends on their mor-

phology and porosity,

which can be controlled by

methodology and nature of the polymers. Freeze-

drying process is relatively a new technique to fabricate

meso- and macroporous three-dimensional biomateri-

als.

This process is based on phase separation mech-

anism (solid–liquid de-mixing). The shape and

morphology of scaffolds can be influenced by modify-

ing the concentration of polymers, using a mixture of

solvents, or varying the lyophilization conditions such

as cooling rate.

In this work, bilayer GTR/GBR membranes with a

non-porous lower layer (to exclude epithelial cells

ingrowth) and porous upper layer (with the ability to

regenerate bone) were fabricated using a freeze-drying

method. In order to persuade biodegradability and fast

bone and tissue healing properties, PCL and BG were

chosen, while PU was selected due to its intrinsic

mechanical stability for handling during surgery. The

porosity of membranes was regulated by controlling

the nature and concentration of polymers, solvents,

and reaction conditions. A combination of solvents

was used to control the sublimation rate and thus

porosity. The in vitro as well as in vivo studies were

performed to fully understand the biological properties

of the membranes.

Materials and methods

Synthetic PU (PU, Z3A1, Biomer Technology Ltd.,

Runcorn, UK) and PCL (molecular weight: 80,000,

Sigma Aldrich, UK) were used for developing the

membrane. The solvents dioxane (Fischer Scientific,

England) and dimethylformamide (DMF, Sigma

Aldrich, UK) were of analytical grade, and no further

purification was required before their use.

Fabrication of bilayer GTR/GBR membranes

BG was synthesized by a base-catalyzed sol–gel tech-

nique using calcium nitrate tetrahydrate (Ca

(NO

)

.4H

O), tetraethyl orthosilicate, and diammoni-

um hydrogen phosphate [(NH

)

HPO

] as a source of

calcium oxide (CaO), silica (SiO

), and phosphorous

pentaoxide (P

), respectively, following the proce-

dure given previously.

The synthesized BG was sin-

tered at 680C. To prepare a bilayer membrane, the

upper layer was composed of the pure PU and the

lower layer consisted of analytical grade PCL with BG.

For the upper layer, 5wt.% of the PU was added to

a mixture of solvents (dioxane and DMF in a volume

ratio of 4:1) at room temperature. The lower layer of

the membrane was fabricated by dissolving 5 wt.% of

the PCL in dioxane and stirred overnight to get a clear

solution. Then, BG (40 wt.%) was added to the pre-

pared PCL solution. For fabrication of the bilayer

membrane, the PCL and the BG mixture was added

into a Petri-plate, cooled at 4C for 2 h and frozen at

20C. Then, a pre-cooled solution of the PU was

poured over the PCL-BG frozen layer and kept at

20C overnight. The bilayer mixture was freeze-

dried (CHRIST, Alpha 1–2 LDplus, UK) at 50C

for 48 h.

Physical and chemical characterization

of membranes

The physical and chemical properties of the bilayer

membrane were investigated by using different analyt-

ical techniques.

Scanning electron microscopy. The surface morphology

and porous structure were examined using scanning

electron microscopy (SEM), TESCAN Vega3 LMU.

Before analysis, the membranes were gold coated for

90 s through a sputter coater (QUORUM). The pore

968 Journal of Biomaterials Applications 33(7)

size was later quantified by using a line-drawing feature

of ImageJ software. The diameter of both small and

large pores on the images was measured manually

and then the mean pore size was determined as a

mean area (measured by software) of six straight lines

for each image.

Fourier transform infrared spectroscopy. The structural

changes due to bonding after lyophilization were

investigated using Fourier transform infrared spectros-

copy (FTIR) (Thermo Nicolet 6700, USA) in conjunc-

tion with attenuated total reflectance sampling

accessory equipped with germanium crystal. Spectra

were obtained at 4 cm

1

resolution, accumulating

256 number of scans within mid-infrared (IR) range

(4000–550 cm

1

X-ray diffraction. The phases of BG in bilayer composite

membranes were studied by X-ray diffractometer

(XPERT-PRO, PANalytical), operated at 40 kV and

40 mA using Cu Karadiation. Wide-angle X-ray dif-

fraction (XRD) patterns were made over an angle

range of 2h¼10–60at 0.02step size.

Swelling studies

The swelling ability of the membranes (1 cm length 1

cm width) was measured in phosphate-buffered saline

(PBS). Pre-weighed samples were dipped in PBS (5 mL)

at 37C temperature for periodic time intervals, i.e. 0,

15, 30, 60, 120, 140, and 160 min. After each time inter-

val, samples were taken out from PBS, and excess

water was removed by squeezing out between two

sheets of filter paper prior to weighing the samples.

The swelling percentage was measured using

the formula:

Swelling % Q

ðÞ

¼wwwd

ðÞ

=wd100 (1)

where w

and w

represent the weight of dry and wet

sample, respectively. The swelling ratio was obtained in

triplicate and expressed as mean standard devia-

tion (SD).

Cell culture

Human carcinoma cell line (MG63, ECACC general cell

collection) was maintained in Dulbecco’s modified

Eagle’s medium (Biosera, UK) added with fetal calf

serum (10%, Sigma-Aldrich, UK), l-glutamine (2 mM,

Sigma-Aldrich, UK), penicillin (100 U/mL, Sigma-

Aldrich, UK), streptomycin (1.2 mg/mL, Sigma-

Aldrich, UK), and amphotericin B (0.0625 mg/mL).

Cells were cultured in T75 tissue culture flask to 90%

confluence in an incubator (at 37C temperature under

5% CO

atmosphere, where media was changed after

every two to three days) and passaged using trypsin-

EDTA (Sigma-Aldrich, UK). The passage number was

kept between 60 and 65. Membranes were sterilized

using 70% ethanol (Sigma-Aldrich, UK) for overnight

incubation (14 h) and washed with PBS (three times,

15 min each) to eliminate any residual ethanol, followed

by dipping in cell culture medium for 1 h before plating

them on 96-well plate for cell seeding

Alamar blue assay. The cytotoxic response of membrane

was determined using Alamar blue assay on human

osteosarcoma cell line (MG63). Briefly, the sterilized

membranes were put in 96-well plate, and then MG63

cells were seeded at an initial cell density of 20,000 cells/

well. The cell viability was quantified after 24, 48, and

72 h. At each end-point, medium for cell culture was

removed and replaced with Alamar blue solution

(Sigma-Aldrich, UK working solution at the 100 mM

solution in serum-free media) for 4 h at 37C temper-

ature. Fluorescence plate reader (Bio-TEK, NorthStar

Scientific Ltd. UK) was used to measure the fluores-

cence at 570 nm. Following the measurements, the

samples were washed with PBS, and fresh serum-

containing cell culture medium was added to incubate

until next end-point. The above method was repeated

at every end-point. The experiment was done in tripli-

cates to obtain a mean value SD. The mean values of

absorbance from each scaffold were plotted over a

graph against various end-points.

Immuno-staining. The membranes were sterilized as

mentioned above and plated in 96-well plate. Human

osteosarcoma cell line MG63 was grown on either sur-

face of the membrane separately with an initial seeding

density of 20,000 cells/well. Then, these membranes

were incubated for three end-points (2, 6, and

10 days) at 37C temperature under 5% CO

atmo-

sphere. Cell culture medium was replenished every

two to three days with fresh medium. After each end-

point, washing of scaffolds was done using PBS, and

then 3.7% formaldehyde (30 min) was used for cells

fixation and 0.1% Triton X-100 for permeabilization

(for 30 min). Phalloidin-tetramethyl rhodamine B iso-

thiocyanate (TRITC) was used for immunofluorescent

staining of F-actin filaments (10 mg/mL, 2 h) and 40,6-

diamidino-2-phenylindole (DAPI) (300 nM, 15 min)

for nuclear staining. Membranes were washed with

PBS after each fluorescent probe incubation. Finally,

the membranes were left in PBS for imaging. The

images were taken by Zeiss confocal microscopy, phal-

loidin TRITC was excited at kex ¼543 nm; and emis-

sion at kem ¼565–615 nm, DAPI was excited using

coherent Chameleon pulsed IR multiphoton laser at

800 nm and emission detected at 435 to 485 nm.

Zahid et al. 969

Images were taken using Achroplan water dipping

objective (10, NA 0.3, WD 3.1) from each seeding

surface of the membrane, and the data were processed

using Zeiss LSM image browser.

Animal studies

Animal experiments were conducted according to the

instructions and permission from the animal welfare

and ethical committee of University of Health

Sciences, Lahore, Pakistan. The institutional guidelines

for the care and use of animals were followed in this

study. Animal studies were performed on wild-type

male adult rats (two months old) weighing 140 to

180 g. Rats were bred in the animal facility of

University of Health Sciences, Lahore, Pakistan.

Animals were kept in specific pathogen-free, clean,

and well-aerated individual cages with an optimum

supply of light, food, and water. Experiments were

performed independently three times on each animal

of the experimental groups and two times on each

animal of the control groups. In order to reduce the

overall sufferings and number of animals, four mem-

branes were implanted subcutaneously in a single

animal. Experimental as well as control animal

groups were further divided into two subgroups: (i)

experimental and control groups observed at 48 h

after implantations and (ii) experimental and control

groups observed at six weeks after implantation. For

implantations in experimental groups, animals were

first anesthetized with intraperitoneal injections of

xylazine hydrochloride (20 mL) and ketamine

(300 mL). Anesthetized animals were then shifted to a

clean bench, and fur from the dorsal side was removed

using a hair trimmer. Eye ointment polyfex (polymyxin

B and bacitracin; GlaxoSmithKline, UK) was then

applied on eyes to avoid dryness and infection. Four

longitudinal surgical incisions (1 cm each) were made at

different locations on the dorsal side of the animal, and

subcutaneous pouches were generated through these

incisions. Four sterilized membranes with the dimen-

sions of 1 1cm

were then inserted into these subcu-

taneously generated pouches (supporting information,

Figure S3). After implantation, surgical wounds were

closed with silk suture material (black braided silk

suture 3/0, 1

=2circle curved cutting 30 mm (care)),

and Pyodine-Iodine solution (an antiseptic used for

skin disinfection) was then applied on wound openings.

While, in the control group, four surgical pouches were

generated in each animal according to the above-

mentioned protocol; however, they were closed with

suture material without any implantations. Health

status and weight of all animals before and after the

start of experiments were regularly monitored. After

the completion of selected time intervals, animals

were peacefully euthanized by CO

asphyxiation fol-

lowed by cervical dislocation. Implanted membranes

together with tissue were then removed. The obtained

membranes were fixed in 4% paraformaldehyde and

processed for further histological examination.

Histology of tissue–implant interface. The tissues were fixed

in 4% paraformaldehyde and dehydrated with an

increasing concentration of ethanol: 70%, 85%, 95%,

and 100% for 1 h in order to avoid tissue distortion.

After dehydration, tissues were dipped in a pre-melted

wax (70C) and fixed in the form of the block for 1 h.

The blocks were placed in a refrigerator at 20C for

30 min for adequate solidification and clamped in the

carrier of the microtome and trimmed to acquire tissue

sections of 5 mm. These tissue sections were transferred

to slides subjected to hematoxylin and eosin staining,

whereby tissue samples were dipped in xylene I and

xylene II for 5 min and rehydrated by dipping in

100% (absolute ethanol), 80%, 70%, and 50% ethanol

(for 1 min each). Then, samples were dipped in deion-

ized water for 30 min and soaked in hematoxylin for

3 min. Excess stain was washed with tap water for

30 min followed by a fast dip in 0.5% acidic alcohol.

The samples were placed under running water for 1 min

and dipped in 1% ammonia solution for 1 min fol-

lowed by immersion in 50% alcohol for 1 min and

eosin for half an hour. Then, samples were soaked

again in 50%, 70% and 80% ethanol for 10 s, in

100% ethanol for 2 min and in xylene for 1 min. The

samples were dried and mounted with mixture of

Distyrene, Plasticizer (tricresyl phosphate), and

Xylene (DPX) mounting media, and images were

taken using a fluorescent microscope (VWRV

Inverted Fluorescence Microscope 89404–464, US) at

varying resolutions.

Statistical analysis. For swelling and in vitro cell studies,

statistical analysis was performed with one-way analy-

sis of variance, post hoc Tukey’s test. Furthermore,

repeated measure analysis was also performed for

swelling study only. A p-value of <0.05 was considered

significant, and tests were performed by using SPSS

version 22 (IBM, USA).

Results

Physical and chemical properties

Macro and microstructure (SEM) analysis. The objective of

this research was to develop a bilayer membrane with

the non-porous upper layer to prevent intrusion of epi-

thelial cells and porous lower layer to maintain blood

vessels and nutrient mass transfer. The pictorial images

of single PU film, single PCL-BG layer, and bilayer

970 Journal of Biomaterials Applications 33(7)

(PU/PCL-BG) membrane, respectively, are shown in

supporting information (Figure S1(a) to (c)). The thick-

ness of the bilayer membrane was about 1.85 mm.

The topography and microstructure of single and

bilayer membranes were observed by SEM (Figure 1).

Figure 1(a) to (c) shows the topography of pure PCL

layer and the PCL-BG composite layers prepared by

the freeze-drying method. A clear hierarchically

arranged pattern was observed with two different

sized pores, with the large (26–38 mm in diameter,

mean diameter 33 5) and small interconnected pores

(2–3 mm diameter), in the pure PCL (supporting infor-

mation, Figure S4 (c) and (d)). The porous structure

was symmetrical throughout the membrane. Addition

of BG in the PCL layer changed the pattern and shape

of pores. Figure 1(e) shows the asymmetrical pores

with wide openings (22–65 mm pore diameter, mean

diameter 37 22) in the PCL-BG layer (supporting

information S4(e)). The change in the pattern could

be related to the existence of a biphasic system (PCL

and BG) and density difference. Figure 1(f) shows that

the BG nanoparticles are embedded in the PCL poly-

mer. The distribution of these nanoparticles was uni-

form and homogeneous.

Microscopic image of the pure PU membrane is pre-

sented in Figure 1(g) to (i), which showed the non-

porous structure of the PU in the bilayer membrane.

Micrographs of the intersection of the two layers

(Figure 1(k) and (l)) show a good interphase or attach-

ment between the two layers, obtained by using a small

amount of the same solvent used for processing in both

the polymeric systems. No detachment was observed

throughout the surface. However, at certain points,

the lacy/fibrous structure was observed which was con-

tinuous and penetrating and hence contributes to intact

structure formation. No delamination was observed

during 28 days of soaking in PBS (not shown in results)

which also showed that bilayer structure was stable.

The membrane was flexible but not fragile, and the

interface shows that two layers intermingled with

each other. The results confirmed that a good interface

between two layers was achieved by manipulating the

number of solvents between the different polymer-

ic layers.

Fourier transform infrared spectroscopy. Chemical structur-

al properties of the PCL, the PU, and the PCL-BG

composite membranes were studied by FTIR spectros-

copy to identify the chemical structure of the backbone

of polymeric molecules and functional groups that ori-

entated themselves to the surface of the molecule. In

FTIR spectrum (Figure 2(a)) of the PCL-BG compos-

ite layer, characteristic vibrational spectral peaks of

PCL and BG were observed. Asymmetric and symmet-

ric CH

stretching vibration peaks were observed at

2944 cm

1

and 2861 cm

1

, and the peaks at 1190 cm

1

1152 cm

1

, and, 1170 cm

1

were attributed to symmet-

ric stretching of C–O–C, C–O, and C–C, respectively.

The spectral peaks at 1240 cm

1

, 1288 cm

1

, and 1714

1

were attributed to C–O–C stretching vibrations

and C–O bonds.

The characteristic vibrational spec-

tral peaks of the BG appeared in FTIR spectrum of the

PCL-BG composite membrane. A broad band with a

shoulder at 1290–890 cm

1

was specific to symmetric

and asymmetric stretching of P–O and Si–O–Si of the

BG.

18,19

A small band (appeared as shoulder) of Si–O–

NBO (non-bridging oxygen) for Si–O–M (where

M¼Ca

þ2

) bonds was observed around 960 cm

1

broad band between 840 and 760 cm

1

was assigned to

bending Si–O–Si mode. The OPCa bond appeared

at 660 cm

1

There was a slight shift in peaks due to

the interaction between the BG and the PCL in

the composite.

In the FTIR spectrum of the PU layer, the charac-

teristic band of free carbonyl (C ¼O) group appeared

as a shoulder at 1726 cm

1

and bonded C ¼Oat

1696 cm

1

. The symmetric and asymmetric stretching

vibrations of C–H with carbonyl appeared at

2845 cm

1

and 2938 cm

1

, respectively, and bending

peaks of C–H were observed at 1348 cm

1

. The stretch-

ing band of N–H was observed at 3331 cm

1

, whereas

the 1598 cm

1

peak attributed to the C ¼C vibrational

benzene ring, and 1539 cm

1

and 1224 cm

1

were the

d(N–H) þvibrational (C–N) peaks. The strong

d(N–H) þv(C–N), b(C–H) peak appeared at

1311 cm

1

X-ray diffraction. The phase development of the BG in

bilayer membrane was investigated by XRD studies.

Figure 2(b) shows the XRD pattern of the upper and

lower surface of the bilayer composite membrane.

The preparation process did not change the phases

of the BG, but there was suppression of few peaks

due to the amorphous nature of polymeric phases.

Swelling properties. The swelling ratio of the GTR/GBR

membrane was studied up to a period of 160 min

(Figure 3(a)). PCL-BG membrane showed maximum

swelling of 140% compared to the PU and bilayer

PU and PCL-BG membrane after 160 min. The bilayer

membrane absorbed the least amount of water and

showed a swelling percentage of only 77%. A variable

swelling behavior was observed for the PU membrane,

which might be due to its hydrophobic nature. The post

hoc Turkey’s test showed that there was the insignifi-

cant difference (p >0.05) among the groups except for

PU-PCL at 160 min which showed a significant differ-

ence (p <0.05) with PCL/BG and PU. Whereas repeat-

ed measure analysis showed the non-significant

difference (p >0.05) with respect to time intervals.

Zahid et al. 971

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

(j) (k)

Upper layer

Lower layer

Interface

(l)

Figure 1. (a–c) SEM images of pure PCL, (d–f) PCL-BG layer in bilayer membrane, (g) pure PU, (h and i) PU layer in bilayer

membrane, (j) bilayer membrane, and (k and l) interface of composite membranes.

972 Journal of Biomaterials Applications 33(7)

180

(a)

(b)

160

140

120

100

Swelling (%)

2.0

1.5

Fluorescence intensity (a.u)

1.0

0.5

0.0

15 30 60

Time (min)

120 140

Bilayered membrane

PU membrane

PCL-BG membrane

Bilayered membrane

TCP

24 48

Time (hours)

PCL-BG

Figure 3. Swelling studies of single layer (PU, PCL-BG) and bilayer GTR/GBR membranes (a) and metabolic activity of MG63

osteosarcoma cells (cultured for 24, 48 and 72 h) in the membranes measured using Alamar blue (b). The value represents the mean

absorbance standard deviation. PCL: polycaprolactone; BG: bioactive glass; PU: polyurethane; TCP: tissue culture plate.

2.0

(a) (b)

1.5

1.0

Intensity (a.u.)

0.5

0.0

3500 3000 2500 2000

PU layer

Pure PCL

C-O

C=O

-CH2

C-H

C-N

N-H

C-H

C=O

C=C

N-H C-H

-CH2

C-O

PO4

Si-O Si-O-Si

C-O

C-O-C

PCL + BG Layer

Wave number (cm–1)

1500 1000 10 20 30

Angle (2θ)

Upper PCL-BG Layer

Lower PU Layer

50 60

Figure 2. FTIR spectrum of PU, BG, PCL, and PCL-BG (a) and XRD pattern of lower and upper surface of bilayer membrane (b).

PCL: polycaprolactone; BG: bioactive glass; PU: polyurethane.

Zahid et al. 973

Cell studies

Alamar blue. The cell viability was evaluated by direct

cell Alamar blue assay, which also showed the cell’s

proliferation and metabolic activity as shown in

Figure 3(b). Metabolic activity of osteosarcoma cells

MG63 increased with time in all membranes, suggest-

ing an increase in cellular viability without any cyto-

toxic interference of membranes over 72 h in culture.

No statistically significant cytotoxicity was noticed in

any of the membrane (in all three end-points) when

compared with the cell in TCP (tissue culture plate,

control). The PCL-BG membrane showed significance

cell proliferation over TCP (at 72 h, p-value ¼0.0027).

Immunostaining. Osteosarcoma cell line was seeded on

either surface of the pure PU, the PCL-BG, and the

bilayer membranes. Phalloidin TRITC and DAPI

staining were used to observe cell adherence to mem-

branes. The immunostaining of cells seeded on either

surface of the PU membrane after 48 h showed consis-

tent growth of MG63 over the whole surface of the

membrane (Figure S2 supporting information).

Almost equal cell attachment was observed on either

surface of the PU membrane after initial seeding.

Figure 4 shows cell growth and attachment on the sur-

face of the PCL-BG membrane at 2, 6, and 10 days. It

was evident that the number of cells increased with the

passage of time in the PCL-BG layer. Cells attachment

and proliferation were also studied on either surface of

the bilayer membrane (Figure 5) at different end-point

(2, 6, and 10 days). The PCL-BG surface cell seeding

showed more cell attachment than the PU surface. The

number of cells on the PCL-BG was increased rapidly

at days 6 and 10, while this number almost remained

unchanged on the PU side. Figure 5 shows that the cells

were also attached inside the pores of the PCL-BG. The

more red color seen on day 2 (Figure 5) was due to

background staining (biomaterial staining), as the

material had an increased auto-fluorescence from

TRITC emission. On contrary, the bottom side (the

PU layer) was non-porous, and most of the cells died

on its surface. The PU membrane is a flat two-

dimensional surface, while the PCL-BG membrane

has pores and therefore provided more surface area

for cells to proliferate than the PU membrane. This

was evident from Figures 4 and 5 which indicate an

increased progressive growth of cells in the PCL-BG

membrane as compared to the PU membrane.

In vivo studies

In vivo studies by using animals are required prior to

the exploitation of any developed materials for clinical

application. Therefore, bilayer membranes were subcu-

taneously inserted into rats to investigate host tissue

responses. The subcutaneous tissue was selected for

implantation since it avoids complex surgical proce-

dures and represents real-time host–material interac-

tions. It allows rapid and reliable observation of

gross inflammatory host responses, such as swelling

and redness (Figure S3 supporting information).

Cell seeding on upper surface of PCL-BG membrane

Day 2

Day 6

Day 10

Cell seeding on lower surface of PCL-BG membrane

Figure 4. Cell attachment and proliferation on PCL-BG membrane observed over 10 days in culture. DAPI: 40,6-diamidino-2-

phenylindole; TRITC: tetramethyl rhodamine B isothiocyanate; PCL: polycaprolactone; BG: bioactive glass.

974 Journal of Biomaterials Applications 33(7)

Results from visual inspection did not reveal any signs

of inflammation, as no swelling or redness at the

implantation site was observed (Figure S3, supporting

information). Moreover, there were no indications of

bacterial contamination or infections, such as pus accu-

mulation at the implantation site. During the entire

study period, animals remained healthy exhibiting

normal walking, feeding, and drinking behaviors,

thereby indicating biocompatible properties of bilayer

membranes. Detailed implant–tissue interactions were

further evaluated by applying histological analysis.

After six weeks of implantation, a connective tissue

Cell seeding on lower PCL-BG surface

Day 2

Day 6

Day 10

Cell seeding on upper PU surface

Figure 5. Cell attachment and proliferation on bilayer membranes observed over 10 days in culture imaged. DAPI: 40,6-diamidino-2-

phenylindole; TRITC: tetramethyl rhodamine B isothiocyanate; PCL: polycaprolactone; BG: bioactive glass; PU: polyurethane.

(a) (b) (c) (d)

(e) (f) (g) (h)

Figure 6. Histological evaluation of implant–tissue interface (a and b ¼control group, c and d ¼bilayer membrane, e and f ¼PU, and

g and h ¼PCL-BG membranes). Membranes were subcutaneously implanted into rats and followed by 48 h and six weeks of

implantation, membranes together with adjacent tissue were fixed and subjected to hematoxylin and eosin stain. Tissue responses

towards bilayer membranes after 48 h of implantation (left side) and after six weeks of implantation (right side). i: site of implantation;

ep: epidermis; ct: connective tissue; sm: skeletal muscles; hf: hair follicle.

Zahid et al. 975

layer indicative for good material–tissue interaction

was visible around all membranes (Figure S3(d), sup-

porting information). Histological analysis of tissue

adjacent to bilayer membranes indicated no signs of

inflammation or accumulation of host immune cells

except normal foreign body responses, which was sim-

ilar to controls (Figure S2, supporting information)

(Figure 6). Overall, the absence of host immune cells,

as well as consistency in tissue adjacent to implanted

membranes, was indicative of good compatibility of

membranes for prospective clinical applications.

Discussion

The cell occlusivity is a major prerequisite for the

GTR/GBR membranes. It requires appropriate mem-

brane surface structure to prevent infiltration of epithe-

lial cells from surroundings while maintaining the

essential mass transference through the membrane. In

the present study, the morphology, adhesion, and pore

structure of the GTR/GBR membrane were controlled

by using the multi-solvent system. It has been reported

that in freeze-drying, liquid–liquid miscibility plays an

important role in porosity, densification, and morphol-

ogy of materials.

23–25

Therefore, a mixture of solvents

with good miscibility and delayed liquid–liquid phase

separation was used to fabricate a membrane with

dense non-porous structure. Binary solvent system,

comprised of DMF and dioxane, was selected for dis-

solution and fabrication of non-porous PU layer.

DMF and dioxane have relatively better miscibility

and separate very slowly (liquid–liquid phase separa-

tion) during cooling and lyophilization, subsequently

avoiding the formation of pores or cracks. Therefore,

a dense non-porous PU structure was formed after

lyophilization. Few macro-pores observed on the sur-

face of the PU layer (at low magnification) were due to

percolation of the solvent during the pouring process;

however, at high magnification, the structure

was found to be non-porous. In the lower layer

(PCL-BG), porous structure with a different pattern

(Figure 1(e)) was observed as compared to the pure

PCL layer (Figure 1(b)). This change in morphology

was due to the presence of BG nanoparticles, which

acted as filler and changed the freezing pattern of the

PCL. These pores had a diameter of 22 to 65 mm and

helped in maintaining essential mass transference

through the membrane during regeneration of bone

and tissues. The use of the same solvent (dioxane) in

both layers resulted in the formation of firm adhesion

between the two layers. The two membranes were not

detached during different physical and biological

investigations.

It has been reported that the structural morphology,

such as porosity and surface area of membranes, affects

the physical and biological properties.

In this study,

the PCL-BG layer possessed porous structure and

therefore exhibited enhanced swelling kinetics.

Swelling behavior is also a function of surface area

and therefore increased surface area due to the pres-

ence of the BG in the PCL membrane might be the

second factor for increased water absorption for the

PCL-BG membranes. The mono-layered PU mem-

brane showed poor swelling due to non-porous struc-

ture and hydrophobic nature. The bilayer membrane

showed less swelling than the single PCL-BG mem-

brane. Interconnectivity of porous structures plays an

important role in water absorption of the membranes

ultimately affecting the physical properties of the

designed materials, as the bilayer membrane showed

less swelling than the single PCL-BG membranes.

The most important criterion for new biomaterials

before their prospective clinical application is their bio-

compatibility. Therefore, synthesized membranes were

investigated for their biocompatibility and immune

responses both in vitro and in vivo. In bilayer mem-

branes, the cells were evenly spread over the surface of

membranes. The enhanced cell attachment and prolif-

eration on the lower layer (PCL-BG surface) was due

to enhanced surface area and roughness because of the

presence of the BG nanoparticles and porous structure.

The bioactive materials, specifically the BG, promote

the cellular processes and improve cell adhesion and

proliferation.

An increased cell metabolic activity

over time with a progressive increase of cell growth

and attachment through the porous structure of the

PU-BG surface of the membrane was observed. More

cell attachment was observed inside the porous struc-

ture of the PCL-BG as compared to the surface, which

is necessary for the GTR/GBR membrane.

The

number of cells on upper side (the PU surface) of the

membrane was far less than the PCL-BG surface due to

hydrophobic properties of PU, which is in accordance

to the requirement of the GTR/GBR membranes.

In the GTR/GBR membranes, the upper surface is

non-porous and hydrophobic to avoid the attachment

and down-growth of epithelial and connective tissues

into the defect, while lower surface should promote the

cells attachment and bone regeneration. In vitro cell

compatible properties of bilayer membranes were fur-

ther validated in the small animal model. All mem-

branes were implanted subcutaneously in soft tissue

to allow maximum contact between host immune sys-

tems and implanted material.

The rats were healthy

and grew normally after implantation. No inflammato-

ry responses were observed during implantation and

surgical procedures which were observed for a period

of six weeks. The population of immune cells in tissue

samples adjacent to membranes was similar to controls

after a short or long period of implantation suggesting

976 Journal of Biomaterials Applications 33(7)

good biocompatibility which was evident from favor-

able tissue response of membranes. Bacterial coloniza-

tion and subsequent biofilm formation remain a major

challenge for the in vivo performance of novel materi-

als. Interestingly, no signs of bacterial contamination,

such as accumulation of pus, could be observed near

implanted membranes. The previously reported studies

showed the optimal persistence of the GTR membranes

in vivo vary from four to several weeks.

Thus, the

controlled morphology and biocompatibility both in

in vitro and in vivo studies of the bilayer membranes

confirmed that the bioactive membranes developed in

this study have excellent potential to be used for wound

healing and tissue regeneration applications.

Conclusion

This study has established a concept design of the

bilayer GTR/GBR membrane with the upper non-

porous hydrophobic PU layer and the lower porous

hydrophobic PCL-BG layer by modified lyophilization

method. The pores or crack formation in the upper

layer was avoided using DMF and dioxane solvents

with high miscibility and low liquid–liquid phase sepa-

ration. This design prevents the penetration and

growth of epithelial cells and connective tissues into

the wound and facilitates the growth and attachment

of cells and transport of nutrition (lower surface) at the

wound site. The bilayer structure developed in this

study demonstrated excellent biocompatibility and sup-

ported the growth of cells towards lower surface. The

introduction of the BG nanoparticles into the PCL

increased the stiffness of the membrane and the

increased cell attachment and proliferation. In vivo

studies revealed that the bioactive membranes did not

cause any inflammatory response, and fast healing was

observed as compared to control, which can be con-

tributed to the bioactivity of the glass nanoparticles

impregnated within the polymeric network structure.

Authors’ Note

Asma Tufail is also affiliated with Department of Material

Science and Engineering, Kroto Research Institute, The

University of Sheffield, Sheffield, United Kingdom and

Fachgebiet Keramische Werkstoffe / Chair of Advanced

Ceramic Materials, Institut fu

¨r Werkstoffwissenschaften und

-technologien, Technische Universit€

at Berlin, Hardenbergstr,

Berlin, Germany.

Acknowledgements

We acknowledge Higher Education of Pakistan and Ministry

of Science and Technology Pakistan for providing analytical

facilities to Interdisciplinary Research Centre in Biomedical

Materials, COMSATS University Islamabad, Lahore

Campus, Pakistan. We are also grateful to Alexander von

Humboldt for facilitating us in this project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with

respect to the research, authorship, and/or publication of

this article.

Funding

The author(s) received no financial support for the research,

authorship, and/or publication of this article.

ORCID iD

Abdul Samad Khan http://orcid.org/0000-0002-2165-

800X

Asma Tufail http://orcid.org/0000-0001-9624-1842

Supplemental Material

Supplemental material for this article is available online.

References

1. Jamuna-Thevi K, Saarani NN, Abdul Kadir MR, et al.

Triple-layered PLGA/nanoapatite/lauric acid graded

composite membrane for periodontal guided bone regen-

eration. Mater Sci Eng C 2014; 43: 253–263.

2. Qasim SB, Delaine-Smith RM, Fey T, et al. Freeze

gelated porous membranes for periodontal tissue regen-

eration. Acta Biomaterialia 2015; 23: 317–328.

3. Qasim SB, Husain S, Huang Y, et al. In-vitro and in-vivo

degradation studies of freeze gelated porous chitosan

composite scaffolds for tissue engineering applications.

Polym Degrad Stab 2017; 136: 31–38.

4. Sun X, Xu C, Wu G, et al. Poly (lactic-co-glycolic acid):

applications and future prospects for periodontal tissue

regeneration. Polymers 2017; 9: 189.

5. Sheikh Z, Khan AS, Roohpour N, et al. Protein adsorp-

tion capability on polyurethane and modified-

polyurethane membrane for periodontal guided tissue

regeneration applications. Mater Sci Eng C Mater Biol

Appl 2016; 68: 267–275.

6. Wissing TB, Bonito V, Bouten CVC, et al. Biomaterial-

driven in situ cardiovascular tissue engineering—a multi-

disciplinary perspective. NPJ Regen Med 2017; 2: 18.

7. Gu

¨rbu

¨z S, Demirtas¸ TT, Yu

¨ksel E, et al. Multi-layered

functional membranes for periodontal regeneration:

Preparation and characterization. Mater Lett 2016;

178: 256–259.

8. Tsui YK and Gogolewski S. Microporous biodegradable

polyurethane membranes for tissue engineering. J Mater

Sci Mater Med 2009; 20: 1729–1741. Journal Article.

DOI: 10.1007/s10856-009-3722-4.

9. Khan AS, Aziz MS, Paul D, et al. Synthesis and in-vitro

analysis of degradative resistance of a novel bioactive

composite. J Bionanosci 2008; 2: 75–88.

Zahid et al. 977

10. Zahid S, Shah AT, Jamal A, et al. Biological behavior of

bioactive glasses and their composites. RSC Adv 2016;

6: 70197–70214.

11. Hench LL. The story of BioglassV

R.J Mater Sci Mater

Med 2006; 17: 967–978.

12. Srinivasan S, Jayasree R, Chennazhi KP, et al.

Biocompatible alginate/nano bioactive glass ceramic

composite scaffolds for periodontal tissue regeneration.

Carbohydr Polym 2012; 87: 274–283.

13. Sheikh Z, Qureshi J, Alshahrani AM, et al. Collagen

based barrier membranes for periodontal guided bone

regeneration applications. Odontology 2017; 105: 1–12.

14. Zhang H, Hussain I, Brust M, et al. Aligned two- and

three-dimensional structures by directional freezing of

polymers and nanoparticles. Nature Mater 2005;

4: 787–793.

15. Mansur HS and Costa HS. Nanostructured poly(vinyl

alcohol)/bioactive glass and poly(vinyl alcohol)/chito-

san/bioactive glass hybrid scaffolds for biomedical appli-

cations. Chem Eng J 2008; 137: 72–83.

16. Shah AT, Batool M, Chaudhry AA, et al. Effect of cal-

cium hydroxide on mechanical strength and biological

properties of bioactive glass. J Mech Behav Biomed

Mater 2016; 61: 617–626.

17. Elzein T, Nasser-Eddine M, Delaite C, et al. FTIR study

of polycaprolactone chain organization at interfaces. J

Colloid Interface Sci 2004; 273: 381–387.

18. Carta D, Knowles JC, Smith ME, et al. Synthesis and

structural characterization of P

–CaO–Na

O sol–gel

materials. J Non Cryst Solids 2007; 353: 1141–1149.

19. Fowler B. Infrared studies of apatites. I. Vibrational

assignments for calcium, strontium, and barium hydrox-

yapatites utilizing isotopic substitution. Inorg Chem 1974;

13: 194–207.

20. Shah A, Ain Q, Chaudhry A, et al. A study of the effect

of precursors on physical and biological properties of

mesoporous bioactive glass. J Mater Sci 2015;

50: 1794–1804.

21. Ahsan MR and Mortuza MG. Infrared spectra of xCaO

(1-x-z)SiO

glasses. J Non Cryst Solids 2005;

351: 2333–2340.

22. Trovati G, Sanches EA, Neto SC, et al. Characterization

of polyurethane resins by FTIR, TGA, and XRD. J Appl

Polym Sci 2010; 115: 263–268.

23. Shieh J-J and Chung TS. Effect of liquid-liquid demixing

on the membrane morphology, gas permeation, thermal

and mechanical properties of cellulose acetate hollow

fibers. J Memb Sci 1998; 140: 67–79.

24. de Groot JH, Nijenhuis AJ, Bruin P, et al. Use of porous

biodegradable polymer implants in meniscus reconstruc-

tion. 1) Preparation of porous biodegradable polyur-

ethanes for the reconstruction of meniscus lesions.

Colloid & Polymer Sci 1990; 268: 1073–1081.

25. Aubert JH and Clough RL. Low-density, microcellular

polystyrene foams. Polymer 1985; 26: 2047–2054.

26. Zhang M, Li XH, Gong YD, et al. Properties and bio-

compatibility of chitosan films modified by blending with

PEG. Biomaterials 2002; 23: 2641–2648.

27. Khan A, Wong F, McKay I, et al. Structural, mechani-

cal, and biocompatibility analyses of a novel dental

restorative nanocomposite. J Appl Polym Sci 2013;

127: 439–447.

28. Mota J, Yu N, Caridade SG, et al. Chitosan/bioactive

glass nanoparticle composite membranes for periodontal

regeneration. Acta Biomaterialia 2012; 8: 4173–4180.

29. Rahim MI, Tavares A, Evertz F, et al. Phosphate

conversion coating reduces the degradation rate and sup-

presses side effects of metallic magnesium implants in an

animal model. J Biomed Mater Res 2017; 105: 1622–1635.

30. Aurer A and Jorgie-Srdjak K. Membranes for periodon-

tal regeneration. Acta Stomatol Croat 2005; 39: 95–112.

978 Journal of Biomaterials Applications 33(7)