International Journal of Pharmaceutics 477 (2014) 218–226

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journal homepage: www.elsevier.com/locate/ijpharm

New drug-eluting lenses to be applied as bandages after

keratoprosthesis implantation
A.S. Carreira a , P. Ferreira a, *, M.P. Ribeiro b,c , T.R. Correia b , P. Coutinho b,c, I.J. Correia b ,
M.H. Gil a
a
Chemical Process Engineering and Forest Products Research Centre, Department of Chemical Engineering, University of Coimbra, Rua Sílvio Lima, Polo II,
3030-790 Coimbra, Portugal
b
CICS-UBI, Centro de Investigação em Ciências da Saúde, Faculdade de Ciências da Saúde, Universidade da Beira Interior, 6200-506 Covilhã, Portugal
c
UDI-IPG, Research Unit for Inland Development, Polytechnic Institute of Guarda, Guarda, Portugal

A R T I C L E I N F O A B S T R A C T

Article history: Corneal tissue is the most commonly transplanted tissue worldwide. This work aimed to develop a new
Received 2 June 2014 drug-eluting contact lens that may be used as a bandage after keratoprosthesis. During this work, films
Received in revised form 6 October 2014 were produced using poly(vinyl alcohol) (PVA) and chitosan (CS) crosslinked with glyoxal (GL).
Accepted 11 October 2014
Vancomycin chlorhydrate (VA) was impregnated in these systems by soaking. Attenuated total
Available online 16 October 2014
reflectance – Fourier transform infrared spectroscopy was used to confirm crosslinking. The cytotoxic
and drug release profile, hydrophilicity, thermal and biodegradation as well as swelling capacity of the
Keywords:
samples were assessed through in vitro studies. PVA and PVA/CS films were obtained by crosslinking with
Drug-eluting lens
Keratoprosthesis
GL. The films were transparent, flexible with smooth surfaces, hydrophilic and able to load and release
Poly(vinyl alcohol) vancomycin for more than 8 h. Biodegradation in artificial lachrymal fluid (ALF) with lysozyme at 37
C
Chitosan showed that mass loss was higher for the samples containing CS. Also, the samples prepared with CS
Glyoxal showed the formation of pores which were visualized by SEM. All samples revealed a biocompatible
character after 24 h in contact with cornea endothelial cells. As a general conclusion it was possible to
determine that the 70PVA/30CS film showed to combine the necessary features to prepare vancomycin-
eluting contact lenses to prevent inflammation after corneal substitution.
ã 2014 Elsevier B.V. All rights reserved.

1. Introduction
Corneal disease is the second most common cause of blindness
worldwide. Normally, this illness is caused by degenerative,
dystrophic, infectious, and inflammatory disorders. Over the last
two centuries, corneal transplantation has arisen as the principal
method for visual rehabilitation when corneal damage occurs. In
2010, 42,642 corneal transplants were performed in USA, however,
patients have to face long waiting lists, limited tissue source and
high rates of grafts rejection (Koplin et al., 2013; Tan et al., 2012;
Waldock and Cook, 2000).
Recently, different studies have been done in order to design,
produce, and fully characterize artificial corneas (keratoprostheses)
that can be used as alternatives to corneal transplantation. Refojo
(1969) studied three hydrogels for being used as artificial
membranes for corneal surgery. Dohlman et al. (1967) reported
the production of glyceryl methacrylate discs to be used as implants
* Corresponding author. Tel.: +351 239798758; fax: +351 239798703.
E-mail address: pferreira@eq.uc.pt (P. Ferreira).
for cornea replacement, Chirila et al., (1994) studied the combination
of a polymer network for being used for this biomedical application.
However, despite the intense research performed in this area, only
three materials are commonly used in surgical practice: the Boston
keratoprosthesis (Massachusetts Eye and Ear Infirmary, Boston, MA)
(Pujari et al., 2011), the AlphaCor (Addition Technology Inc., Des
Plaines, IL) (Hicks et al., 2002) and the osteo-odonto-keratopros-
thesis also known as the ‘OOKP’ (originally described by Strampelli,
1963, and then modified by Falcinelli et al., 1993 and Liu, 1998).
Furthermore, patients after being submitted to surgery usually
require the continuous use of a soft therapeutic lens (known as a
bandage contact lens). These allow the maintenance of an
adequate ocular surface hydration, improve patient comfort and
protect the eye from necrosis that can be caused by ocular surface
exposure (Wong and Yiu, 2012). Endophtalmitis is one of the
complications that may arise in the post-surgical period and may
result in a considerable loss of vision (Nouri et al., 2001).
The post-surgical therapeutics involves a topical antibiotic
prophylaxis (Durand and Dohlman, 2009), commonly using
quinolones (e.g., moxifloxacin and gatifloxacin). To further reduce
the risk of endophthalmitis, co-administration of vancomycin has

http://dx.doi.org/10.1016/j.ijpharm.2014.10.037
0378-5173/ ã 2014 Elsevier B.V. All rights reserved.
 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226 219

been also implemented to eliminate gram-positive bacteria resistant Table 1

Ratio of PVA, CS and GL in the films’ composition.
to quinolones (Durand and Dohlman, 2009). However, the topical
application of vancomycin drops has a limited efficacy due to Sample PVA CS GL
lacrimation, tear drainage and turnover (Das and Suresh, 2011). (%wt) (%wt) (%wt/wtpolymer)
Moreover, a certain amount of drug flows with tears into the PVA-GL0 100 0 0
nasolacrimal ducts and consequently may cause unwanted systemic PVA-GL20 100 0 20
PVA-GL50 100 0 50
side effects (Ali and Lehmussaari, 2006). Besides, the instillation of
70PVA/30CS-GL10 70 30 10
ophthalmic drugs as eye drops on the cornea results in rapid 50PVA/50CS-GL10 50 50 10
variation in drug concentration limiting their therapeutic efficacy. To CS-GL0 0 100 0
overcome such drawbacks, new implantable devices are required to
allow a controlled drug release over a long period of time, decreasing
drug losses and adverse effects. Additionally, the residence time of
the drug in the tear film is also increased (Ludwig, 2005). To prepare the PVA based films, the PVA solution was mixed
The patient subjected to keratoprosthesis will be using soft with GL using different ratios, as shown in Table 1. The pH of this
contact lens in a permanent basis. So the incorporation of the mixture was adjusted to 2.5, using an HCl solution. After the
desired drug in the lens composition improves patients’ compli- homogenization process, the solution was heated in a water bath at
ance, and allows drug delivery in a sustained noninvasive form. 45
C, for 2 h. Afterwards, this solution was poured into a Petri dish
Also, in the presence of contact lens, drug molecules have longer and placed in an oven at 45
C (16 h) to dry. At the end of this
residence time in the post-lens tear film leading to a higher drug period, the obtained films were removed from the Petri dishes and
flux through cornea and less drug inflow into the nasolacrimal duct immersed in water for 2 days. A film without crosslinker (PVA-GL0)
(Pattel et al., 2013). was also prepared in order to be used as a control sample in the
Prototypes of drug-eluting contact lenses, containing fluores- following characterization steps.
cein and ciprofloxacin, have been previously prepared by other PVA/CS films were prepared by using a procedure similar to the
authors (Ciolino et al., 2009). These lens were produced with a one described above. The solutions were homogenously mixed,
mixture of poly(lactic-co-glycolic acid) (PLGA) and poly(hydrox- placed in Petri dishes and then placed in the oven under the same
yethyl methacrylate) (pHEMA) and were able to perform drug conditions. The ratios between PVA/CS and GL are also described in
release in therapeutically relevant concentrations for 1 month. Table 1. Like in the case of PVA, a CS film without crosslinker (CS-
The aim of the present work was to produce drug-eluting GL0) was prepared in order to be used as a control sample during
contact lenses to be applied as bandages. PVA and CS were selected characterization.
as raw materials because of their biodegradability and biocom-
patibility as well as their ability to form films. In order to prepare 2.3. Attenuated total reflectance – Fourier transform infrared
these systems, GL was used as the crosslinker since previous spectroscopy
studies have stated its efficiency for biomedical purposes (Gupta
and Jabrail, 2006; Yang et al., 2005). PVA and PVA/CS films spectra were acquired in the range of
4000–500 cm1, using a Magma-IRTM Spectrometer 750, operating
in ATR mode (Golden Gate Single Reflection Diamond ATR). The
2. Experimental procedure
samples were previously dried in a vacuum oven at 40
C and the
data collection was performed with a 4 cm1 spectral resolution
2.1. Materials
and after 64 scans.

Acetic acid (1%), amphotericin B, L-glutamine, Eagle’s minimum

essential medium (MEM), glyoxal (40% wt.%) aqueous solution,
hydrochloric acid 0.5 M at 37% (HCl), penicillin G, poly(vinyl
alcohol) Mw 13,000–23,000 and 87–89% hydrolyzed, streptomycin,
trypsin and lysozyme (from chicken egg white) were purchased
from Sigma–Aldrich (Sintra, Portugal). Chitosan (Mw 100,000–
300,000) was obtained from Acros Organics. Fetal bovine serum
(FBS) was purchased from Biochrom AG (Berlin, Germany). The
3-[4,5-dimethylthiazol-2-yl]-5-(3-carboxy-methoxyphenyl)-2-(4-
sulfophenyl)-2H-tetrazolium, inner salt (MTS) and electron
coupling reagent phenazine methosulfate (PMS) were purchased
from Promega. T-flasks and 96-well plates were purchased from
Nunc (Denmark). Vancomycin chlorhydrate was kindly donated by
the Ophthalmology Service of the Hospital of the University of
Coimbra (Dr. Joaquim Murta).
2.2. PVA and PVA/CS films preparation
The polymeric films (PVA and PVA/CS) were prepared by
casting. A PVA solution (5 wt.%) was prepared by dissolving PVA in
distilled water. CS solution (2 wt.%) was prepared by dissolving 2 g
of CS in 100 mL of an aqueous acetic acid solution (1%). Both
solutions were stirred until they became homogeneous.
Two sets of films were prepared, one based on PVA and another
combining PVA and CS. The crosslinker agent was GL, as previously
mentioned.
2.4. Study of water uptake ability (swelling)
The swelling of different films was evaluated in ALF (pH 7.4) at
37
C and in ALF and water at room temperature. All the samples
were previously dried in a vacuum oven at 40
C and the weight of
the dried sample was determined (Wd). The dried samples were
then immersed in a constant volume of ALF. At pre-determined
intervals (t), samples were removed from the media and weighted
at different times until a maximum weight was achieved (Ws). The
swelling ratio was evaluated through Eq. (1).
 
Ws  Wd
Swellingð%Þ ¼  100 (1)
Wd
All the assays were carried out in triplicate and the results were
expressed as mean  standard deviation.
2.5. Water contact angle determination
The water contact angle (u) measurements were performed at
room temperature in an OCA 20 contact angle measurement unit
from Dataphysics. All measurements were performed on the air-
facing surfaces of the samples using the sessile drop method. Ten
measurements on different points were performed on each
sample, from which the mean static contact angle and its standard
deviation were determined.
220 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226

2.6. Thermal properties

Thermal behavior of the prepared films was evaluated by

thermogravimetric analysis (TGA). The TGA was carried out in a
TGA Q500 from Thermal Analysis at a 10
C min1 heating rate and
was performed in nitrogen atmosphere at a flow rate of 100 mL
min1.

2.7. In vitro degradation

The in vitro degradation assays were performed in ALF, at 37
C.
In order to mimic the in vivo conditions ALF was supplemented
with lysozyme (2 mg/mL). All the assays were held for a period of
4 weeks. The results were quantified gravimetrically, through the
determination of their mass loss, by using Eq. (2), where wt is the
weight of the dried sample at time t and w0 is the initial weight of
the sample. All the experiments were carried out in triplicate and
the results were expressed as mean  standard deviation.
 
w0  wt
Mass lossð%Þ ¼  100 (2)
w0
Fig. 1. Crosslinking reaction of PVA with GL, in acidic medium.
The absorbance was measured at 492 nm using a microplate reader
(Sanofi, Diagnostics Pauster). Wells containing cells in the culture
medium without materials were used as negative control. EtOH 96%
was added to wells containing cells to be used as a positive control
(Maia et al., 2009; Palmeira-de-Oliveira et al., 2010). Statistical
analysis was performed using one-way ANOVA with Dunnett post-
hoc test. Computations were performed using MYSTAT 12 statistical
package (Systat Software, a subsidiary of Cranes Software Interna-
tional Ltd.).

2.11. Drug loading and in vitro drug release

2.8. Scanning electron microscopy (SEM)
SEM technique was used to visualize the morphology of the
films at its initial stage and also after being subjected to a 4 weeks
in vitro degradation period. SEM analyses were performed by using
a Scanning Microscope JSM–5310 from Jeol. The samples were
previously dried and then placed over carbon ribbon in an
appropriated support and covered with a gold thin layer.
2.9. Proliferation of corneal endothelial cells in presence of polymeric
films
Rabbit corneal endothelial cells (CEC) were obtained as
previously described (Natu et al., 2007) and were seeded in
25 cm2 T-flasks with MEM with heat-inactivated FBS (10% v/v) and
growth factors (fibroblast growth factor (FGF), epidermal growth
factor (EGF), nerve growth factor (NGF)). T-flasks with cells were
incubated in a 5% CO2 humidified atmosphere at 37
C (Maia et al.,
2009). Then, to evaluate cell behavior in the presence of the
different materials, films were placed in 48-well plates and
sterilized by UV exposure for at least 30 min. Subsequently, CEC
were cultured at a density of 5  104 cells/well and after 24 h, cell
growth was monitored using an Olympus CX41 inverted light
microscope (Tokyo, Japan) equipped with an Olympus SP-500 UZ
digital camera.
2.10. Characterization of the cytotoxicity profile of the films
To evaluate the cytotoxicity of the polymeric films, CEC (5  104
cells/well) were cultured with MEM with 10% of inactivated FBS. At
the same time, in another plate, the culture medium was added to the
sterilized polymers, and left there for 24 h. After this period of
incubation, the culture medium used for cell growth was removed
and was replaced by 100 mL of medium that was in contact with the
polymers. Then cells were incubated at 37
C, in a 5% CO2 humidified
atmosphere, for more 24 h. Subsequently, cell viability was assessed
through the reduction of the MTS into a water-soluble brown
formazan product (n = 5) using a method previously described by
our group (Maia et al., 2009). The medium of each well was then
removed and replaced with a mixture of 100 mL of fresh culture
medium and 20 mL of MTS/PMS reagent solution. The cells were
incubated for 4 h at 37
C, under a 5% CO2 humidified atmosphere.
The previously obtained films were used to prepare VA delivery
systems. The drug was loaded into the polymeric matrices by
soaking using two aqueous solutions of VA with different
concentrations (1 mg/mL and 5 mg/mL). Briefly, the dried samples
were immersed in 1 mL of VA aqueous solutions for a period of
2 days at room temperature.
The quantification of the amount of drug released as a function
of time was performed by UV–vis spectroscopy, at 280 nm, using a
Jasco V-530 spectrophotometer. Three samples of each drug-
loaded film were transferred into a closed vial containing 5 mL of
ALF. These vials were then placed in a thermoshaker at 37
C with
continuous stirring at 100 rpm. Absorbance values were obtained
at 280 nm, at predetermined time intervals, until constant value.
Blank samples (unloaded films) were also incubated in ALF under
the same experimental conditions and absorbance of the incuba-
tion solution was also measured at the same wavelength. The value
of drug released for each film was obtained by the difference
between these two absorbance values. All the experiments were
carried out in triplicate and the results were expressed as
mean  standard deviation.
3. Results and discussion
During this work several films with different compositions
were prepared.
PVA films were obtained by crosslinking the hydroxyl groups in
PVA with the carbonyl groups of GL. This reaction led to the
Fig. 2. Crosslinking reaction of CS with GL, in acidic medium with the
establishment of imine linkages.
 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226 221

Fig. 3. ATR-FTIR spectra obtained for the PVA films (A) and PVA/CS films (B): 1 – PVA-GL0; 2 – PVA-GL20; 3 – PVA-GL50 4 – CS-GL0; 5 – 50PVA/50CS-GL10; 6 – 70PVA/30CS-
GL10.

production of a new acetal linkage as previously described by
Zhang and co-workers (Fig. 1).
The FTIR spectra of PVA and crosslinked PVA are presented in
Fig. 2A. The wide band at 3276 cm1 may be ascribed to the O—H
groups stretching, while the ones at 2920 and 2858 cm1 are
related to the stretching of C—H from alkyl groups. The band at
1710 cm1 is attributed to the stretching of C¼O and C—O from the
acetate groups remaining in PVA (after its synthesis from polyvinyl
acetate). Similarly, the band at 1244 cm1 corresponds to the
stretching of ¼C—O—C of those acetate groups.
When PVA is crosslinked with GL, its OH groups are involved in
the reaction. This fact is supported by the visible decrease in the
intensity of the band at 3276 cm1. Such results indicate that the
reaction of the PVA with GL has occurred by forming acetal
linkages between PVA hydroxyl groups and the aldehyde in acidic
conditions (Mansur et al., 2008). These conclusions are supported
by the new band that appears in the spectra of the crosslinked films
at 1043 cm1, ascribed to the C—O—C stretch of the acetal groups
(Pavia et al., 1996).
For the PVA/CS films, mixtures of both polymers were cross-
linked with GL. Since CS presents hydroxyl groups in its structure,
acetal linkages may also be formed between CS and GL. However,
amine groups on CS are also able to react with aldehyde groups of
GL in acid medium. This reaction results in the formation of imine
linkages as presented in Fig. 3 (Parida et al., 2011).
Fig. 2B presents the obtained spectra for CS and for PVA/CS
films. The spectrum of the CS film shows a band at 1545 cm1
attributed to the symmetric deformation of —NH3+ groups which
are formed by the ionization of CS primary amino groups in the
acidic medium. It is visible that the intensity of this band decreases
when PVA is crosslinked with CS. This fact is related with the
decrease of the free amino groups which react with GL forming
imine linkages. Also, a band at 1384 cm1 is visible in the spectrum
of CS and may be ascribed to the presence of carboxylic acid in the
polymers. The presence of these bands was also observed by Costa
et al. (2009). These authors prepared polymer blends based on CS
and PVA, by using glutaraldehyde (GA) as a crosslinker for skin
tissue regeneration.
The spectrum of the CS based film also presents two bands at
1151 and 897 cm1 which correspond to the saccharide structure.
However, when PVA is added to chitosan these characteristic bands
of the CS structure disappear. Moreover, the characteristic bands of
PVA begin to stand out. These include the bands at 1730 cm1
(attributed to the deformation of the carbonyl groups from the
unhydrolyzed acetate groups of PVA) and at 1235 cm1 (corre-
sponding to the stretching of ¼C—O—C of those same acetate
groups) (Pavia et al., 1996; Zhang et al., 2010).
As previously described, CS amine groups are able to react with
aldehyde groups of GL in acid medium forming imine linkages
(Parida et al., 2011). When comparing the spectrum of CS and

Fig. 4. Swelling degree in ALF at 37
C and in ALF and water at room temperature (RT).
222 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226

Table 2 properties influence the biological behavior of the polymeric

Water contact angles measured for the PVA, PVA/CS and CS films.
materials (Bahrami et al., 2003).
Samples Contact angles (
) The swelling of the films was evaluated under different
PVA-GL0 71.08  2.57 conditions in order to determine the influence of ionic strength,
PVA-GL20 70.32  1.54 pH and temperature in this parameter. Therefore, swelling experi-
PVA-GL50 63.46  1.46 ments were performed in water and ALF at room temperature (to
70PVA/30CS-GL10 83.69  2.47
mimic storage conditions) and in ALF at 37
C (to mimic biological
50PVA/50CS-GL10 87.26  1.60
CS-GL0 91.27  2.58
conditions).
Fig. 4 summarizes the results obtained for the equilibrium
swelling of the different experiments which was reached after a
CS/PVA films, it is possible to detect a new band at around period of 2 days. The results obtained for the films prepared
1630 cm1 that is not visible in the spectra of CS or PVA. This band without GL (PVA-GL0 and CS-GL0), are not shown since they were
may be attributed to the stretching of the C¼N and confirms the not able to maintain their integrity when immersed in water.
reaction between CS and GL. A similar degree of swelling was obtained for samples when
they were immersed in ALF at 37
C and at room temperature.
3.1. Study of films water uptake ability (swelling) Moreover, swelling studies showed that the water uptake
capability of the PVA films was influenced by the amount of GL
Films bulk hydrophilicity was characterized by determining the despite the tested medium, indicative of a correspondence
swelling of materials, while the surface hydrophilicity was between crosslinking degree and GL concentration. These results
evaluated by determining the water contact angle. These are related to the decrease in the mobility of the polymeric chains
when crosslinked. However, another reason is the number of free
OH groups which are available at those polymer chains. As
previously described, PVA is able to react with GL by its hydroxyl
groups forming acetal linkages. This means that the hydrophilic OH
groups are consumed in crosslinking, compromising the ability of
the polymer to form hydrogen bonds with water and therefore
decreasing its hydrophilicity. Swelling evaluation of PVA films also
showed that this value was higher when studies were performed in
water, which is related to the osmotic pressure caused by the
presence of salts in ALF. This osmotic pressure decreased the water
penetration into the polymeric matrix, reducing the swelling
degree of the samples.
By comparing the results for the PVA and the PVA/CS films in
ALF, it is visible that the films containing CS showed a decreased
swelling profile despite containing lower GL content. This decrease
in water uptake capacity is attributed to the lower hydrophilicity of
CS when compared to PVA. Moreover, CS is able to react with GL by
its NH2 groups as well as by its OH groups while PVA crosslinking
reaction is limited to its OH groups. According to previous studies,
the amine groups of chitosan are more reactive than the hydroxyls
of PVA (Costa et al., 2009). This indicates that films prepared with
CS present higher crosslinking degrees and therefore lower values
of swelling.
The swelling of CS films presented significantly higher values in
water than in ALF. This behavior may be related to the different pH
values of water (pH 6) and ALF (pH 7.4). As described in FT-IR
spectra analysis of the PVA/CS films, NH2 groups from the CS

Fig. 6. Mass loss (%) of the several films incubated in ALF, supplemented with
Fig. 5. TGA curves of (A) PVA and CS; (B) PVA films and (C) PVA/CS films. lysozyme, at 37
C.
 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226 223

(pKa = 6.5) are able to ionize and form NH3+ when pH is lower than
pKa. This provides CS chains repulsive surface charge leading to an
increase in its water sorption capacity. Bahrami et al. (2003) have
also reported that the presence of NH3+ groups in CS inhibit the
formation of hydrogen bonds between CS and PVA, and therefore
have greater disposition for binding with the water.
3.2. Water contact angle
The contact or wetting angle is usually assessed by placing a
small liquid droplet on a flat horizontal solid surface. The contact
angle is the angle which is formed by the baseline and the tangent
to the drop contour at the three-phase point. This value is specific
for any given system being determined by the interactions of the
three interfaces. In order to further characterize the surface
properties of the films, water contact angles were determined and
the results are presented in Table 2.
As shown in Table 2, the film based exclusively on CS was the
one presenting higher values of water contact angles. The
incorporation of PVA in the composition of the films led to a
decrease in those values, which means that PVA increases
wettability of the samples. This fact is due to the higher
hydrophilicity of PVA when compared to CS and has been reported
by other authors who prepared blends of PVA and CS (Zhang et al.,
2010). Moreover, crosslinking of PVA led to a decrease of the water
contact angle. This effect is attributed to the decrease of the
amount of free hydrophilic groups (OH) in PVA, which are
consumed throughout the reaction with GL. This same behavior
has been reported for the swelling profile of the samples. However,
it was observed that for the film prepared with PVA with no

Fig. 7. SEM images obtained for the several films surface: Initial structure of PVA-GL50 (A1), PVA-GL20 (B1), 70PVA/30CS-GL10 (C1) 50PVA/50CS-GL10 (D1); and after
4 weeks of incubation in ALF with lysozyme at 37
C (A2–D2).
224 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226

addition of GL (PVA-GL0), the value of the water contact angle was During the assays, it was observed that mass loss was
higher. This means that although its hydroxyl groups were not influenced by crosslinking degree and was significantly less and
involved in any chemical bonds, the film’s wettability was lower slower for PVA-GL50 than for all the other films. These results are
than the ones of the crosslinked films. Several factors may consistent with the lower mobility of the crosslinked polymeric
influence the mobility and migration of the functional groups of a chains, inhibiting the release of PVA molecules and their solubility
polymer from the bulk to the surface (as well as the other way in the aqueous medium. Moreover, incorporating PVA in the
around) (Darcovich and Kutowy, 1988). One of those factors may be materials enhanced their resistance to degradation by lysozyme.
the intra/intermolecular bonds in that polymer. In the case of PVA Lysozyme is responsible for the degradation of CS by hydrolyzing
films, strong intra/intermolecular bonds between the hydroxyl the b-(1–4) linkages between its D-glucosamine and N-acetyl-D-
groups of PVA may be established, inhibiting the formation of glucosamine residues (Verheul et al., 2009). For that reason,
hydrogen bonds with water when in contact with the surface and samples containing chitosan degraded faster. Also, the sample with
therefore increasing water contact angle. the higher amount of CS is the one presenting the higher value of
mass loss at the end of the experiment (16%).
3.3. Thermal properties The samples subjected to biodegradation were also observed by
SEM. This technique allowed the visualization of the membrane
Thermogravimetric analysis (TGA) was performed for PVA and morphology before and after incubation. As depicted in Fig. 7, all
CS as well as for all the prepared films. The obtained thermograms the films presented a very smooth surface and with no visible pores
are presented in Fig. 5A–C. at their initial state (A1, B1, C1 and D1).
The thermogram of PVA shows two weight losses (Fig. 5A). The After the incubation period (4 weeks), all the films suffered
first, at 250–350
C, accounts for approximately 70% of the total some morphological changes. Samples composed of PVA showed
weight loss. The second was registered at 400–450
C and at the an increase in their surface roughness and PVA-GL20 was the one
end of the cycle, a residue of 5% remained in the pan. This two-step with the more accentuated variation probably due to its higher
thermal degradation profile is in accordance with other studies water uptake capacity. However, only the samples prepared with
previously reported in the literature. These degradation stages CS present pore formation after degradation. These pores were
have been explained by the elimination of PVA side groups and by observed to be larger and in a higher number in the 50PVA/50CS-
the breakdown of the polymer backbone respectively (Holland and GL10 film, which is justified by the effect of the enzyme on CS. With
Hay, 2001). the increase of chitosan concentration in the films, the effect of
Two weight losses are observed in the CS TGA curve. The first, at lysozyme in their degradation becomes more pronounced.
40–100
C, is attributed to moisture evaporation while the second,
at 250–350
C, is assigned to the degradation of chitosan.
According to previous studies, pyrolysis of chitosan starts by a
random breakdown of the glycosidic bonds, followed by decom-
position to acetic, butyric and lower fatty acids (Pieróg et al., 2012).
These final products are responsible for the residue of nearly 40%
remaining after the experiment.
The TGA traces of the prepared films are shown in Fig. 5B and C.
The PVA-GL0 films present a TGA curve similar to the one of PVA in
the original state. The weight loss stages are the ones described for
PVA and occurred at the same temperature range. When PVA was
crosslinked with GL, its thermal degradation profile slightly
changes. A weight loss of approximately 5% is registered at 100–
150
C and is attributed to moisture evaporation. However, the
main variation is the disappearance of the second stage of weight
loss which is due to the loss of PVA side groups. When reacting
with GL these groups are involved in chemical bonds and are
therefore less likely to suffer thermal degradation. For this reason,
degradation of crosslinked PVA occurs in a one-step way.
The films prepared by blending PVA and CS present a two stage
thermal degradation profile. The first weight loss occurred at 200–
250
C corresponding to the thermal degradation of CS and
accounting for approximately 50% of total weight loss.
The second stage of weight loss occurred at 400–450
C and is
attributed to the degradation of PVA. The amount of residue
remaining at the end of the temperature cycles increased with
the concentration of CS in the film. For the sample 70PVA/30CS-
GL10 the percentage of residue was 25% while for the 50PVA/50CS-
GL10 this value was of 35%. This is consistent with the TGA curve of CS
and with other results reported in literature (El-Hefian et al., 2010).

3.4. In vitro biodegradation

The results of the degradation experiments are depicted in

Fig. 6. It is noticeable that mass loss occurred during all the assay Fig. 8. Microscopic photographs of CEC in the presence of different films (A);
evaluation of the cellular activity after 24 h (B). PVA-GL20; PVA-GL50, 70PVA/30CS-
period. However, the obtained mass loss profiles varied with the GL10; 50PVA/50CS-GL10; positive control (K+); negative control (K). Each result is
samples composition. the mean  standard error of the mean of at least three independent experiments.
 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226
Fig. 9. VA release profiles for samples previously immersed (A) in VA aqueous solution of 1 mg/mL and (B) in VA aqueous solution of 5 mg/mL.
3.5. Characterization of the films biocompatibility
The films cytotoxic profile was evaluated through in vitro
studies. Initially, CEC were seeded at the same initial density in the
48 well plates, with or without material. After 24 h, cell adhesion
and proliferation was monitored using an inverted light micro-
scope (Fig. 8). CEC adhered and grew in the vicinity of polymeric
films and in the negative control (Fig. 8A). In the positive control no
cell adhesion or proliferation was observed. Dead cells with their
typical spherical shape can be observed in Fig. 8A (K +).
Furthermore, to characterize the CEC viability in the presence of
the polymeric films, an MTS assay was also performed (Fig. 8B).
The MTS assay results showed a significant difference between
cells exposed to polymeric films and the positive control
(*p < 0.05) after 24 h of incubation, suggesting that the polymers
did not affect cell viability.
3.6. Drug release profiles
Drug release studies were performed in ALF, at 37
C, under
constant agitation and all the measurements were done using UV
spectroscopy. The drug released by each system was determined
using a validated calibration curve at 280 nm.
For each sample, the percentage of drug release Mt/M1 was
calculated, where Mt is the amount of drug release at a given t time
and M1 is the total amount of VA released (achieved after 24 h).
Fig. 9 shows the release pattern for the films immersed in each VA
solution (1 and 5 mg/mL), during the period of 8 h.
The observation of the drug release profiles for PVA-GL20 and
PVA-GL50 confirms that despite the concentration of the soaked
vancomycin solution, nearly 90% of the drug is released after 1 h
which is a short period of time for the pretended application. This
prompted release of VA is related with the high values of swelling
of the samples when incubated in ALF at 37
C.
The samples prepared with PVA and CS presented similar drug
release profiles with no initial burst. The ones drenched in the
1 mg/mL VA solution were able to sustain drug release through a
longer period of time and after 1 h of incubation in ALF only 30% of
VA was released. At the end of the measurements, after 8 h, nearly
the same amount of VA (between 20 and 30%) is still entrapped in
the system, which indicates that drug release would continue for a
longer period of time. Contrarily, the samples prepared by soaking
in the 5 mg/mL VA solution, presented similar results to the PVA
based systems. The higher concentration gradient induces a more
pronounced driving force and the drug is released sooner and
quicker, that is, in a less sustained manner.
225
4. Conclusion
Herein, new materials suitable to prepare drug-eluting contact
lenses were successfully produced and used to deliver vancomycin
to prevent inflammation after corneal substitution. The different
films were flexible, transparent and presented a very smooth
surface with no visible pores. Also, they possess a hydrophilic
profile being able to swell even in artificial lacrimal fluid. These
features indicate that the materials may be suitable for the
production of lenses since they would most likely be comfortable
to wear and will allow maintaining an adequate ocular surface
hydration. Moreover, vancomycin was successfully loaded onto the
films and drug release profiles indicate that sustained drug
delivery was accomplished for more than 8 h for the films
containing CS.
Biodegradation studies indicate that mass loss was more
accentuated for the samples containing CS. This fact is the result
of the action of lysozyme on CS structure and consequent
degradation. However, for the sample 70PVA/30CS, mass loss is
negligible for the first 7 days which is compatible with the use of a
contact lens that is designed for daily use and is able to deliver VA
during that time period.
Finally, all samples revealed a biocompatible character after
24 h in contact with cornea endothelial cells since cell viability was
not affected by the materials.
References
Ali, Y., Lehmussaari, K., 2006. Industrial perspective in ocular drug delivery. Adv.
Drug Deliv. Rev. 58, 1258–1268.
Bahrami, S.B., Kordestani, S.S., Mirzadeh, H., Mansoori, P., 2003. Poly(vinyl alcohol)
– chitosan blends: preparation, mechanical and physical properties. Iran.
Polym. J. 12, 139–146.
Chirila, T.V., Vijayasekaran, S., Horne, R., Chen, Y.C., Dalton, P.D., Constable, I.J.,
Crawford, G.J., 1994. Interpenetrating polymer network (IPN) as a permanent
joint between the elements of a new type of artificial cornea. J. Biomed. Mater.
Res. 28, 745–753.
Ciolino, J.B., Hoare, T.R., Iwata, N.G., Behlau, I., Dohlman, C.H., Langer, R., Kohane, D.
S., 2009. A drug-eluting contact lens. Invest. Ophthalmol. Vis. Sci. 50,
3346–3352.
Costa, E.S.J., Pereira, M.M., Mansur, H.S., 2009. Properties and biocompatibility of
chitosan films modified by blending with PVA and chemically crosslinked. J.
Mater. Sci. Mater. Med. 20, 553–561.
Darcovich, K., Kutowy, O., 1988. Surface tension considerations for membrane
casting systems. J. Appl. Polym. Sci. 35, 1769–1778.
Das, S., Suresh, P.K., 2011. Drug delivery to eye: special reference to nanoparticles.
Int. J. Drug Deliv. 2.
Dohlman, C.H., Refojo, M.F., Rose, J., 1967. Synthetic polymers in corneal surgery: I.
Glyceryl methacrylate. Arch. Ophthalmol. 77, 252.
Durand, M.L., Dohlman, C.H., 2009. Successful prevention of bacterial endoph-
thalmitis in eyes with the Boston keratoprosthesis. Cornea 28, 896–901.
El-Hefian, E.A., Nasef, M.M., Yahaya, A.H., 2010. The preparation and characteriza-
tion of chitosan/poly(vinyl alcohol) blended films. J. Chem. 7, 1212–1219.
226 A.S. Carreira et al. / International Journal of Pharmaceutics 477 (2014) 218–226
Falcinelli, G., Barogi, G., Taloni, M., Falcinelli, G., 1993. Osteoodontokeratoprosthesis:
present experience and future prospects. Refract. Corneal Surg. 9, 4.
Gupta, K., Jabrail, F.H., 2006. Glutaraldehyde and glyoxal cross-linked chitosan
microspheres for controlled delivery of centchroman. Carbohydr. Res. 341,
744–756.
Hicks, C.R., Crawford, G.J., Tan, D.T., Snibson, G.R., Sutton, G.L., Gondhowiardjo, T.D.,
Lam, D.S., Downie, N., 2002. Outcomes of implantation of an artificial cornea,
AlphaCor: effects of prior ocular herpes simplex infection. Cornea 21, 685–690.
Holland, B., Hay, J., 2001. The thermal degradation of poly(vinyl alcohol). Polymer
42, 6775–6783.
Koplin, R.S., Wu, E.I., Ritterband, D.C., Seedor, J.A., 2013. Corneal transplantation:
keratoprosthesis. The Scrub’s Bible. Springer, Berlin, pp. 113–117.
Liu, C., 1998. Indications and technique of modern osteo-odonto-keratoprosthesis
(OOKP) surgery. Eye News 5, 17–22.
Ludwig, A., 2005. The use of mucoadhesive polymers in ocular drug delivery. Adv.
Drug Deliv. Rev. 57, 1595–1639.
Maia, J., Ribeiro, M.P., Ventura, C., Carvalho, R.A., Correia, I.J., Gil, M.H., 2009. Ocular
injectable formulation assessment for oxidized dextran-based hydrogels. Acta
Biomater. 5, 1948–1955.
Mansur, H.S., Sadahira, C.M., Souza, A.N., Mansur, A.A., 2008. FTIR spectroscopy
characterization of poly(vinyl alcohol) hydrogel with different hydrolysis
degree and chemically crosslinked with glutaraldehyde. Mater. Sci. Eng. C 28,
539–548.
Natu, M.V., Sardinha, J.P., Correia, I.J., Gil, M.H., 2007. Controlled release gelatin
hydrogels and lyophilisates with potential application as ocular inserts. Biomed.
Mater. 2, 241–249.
Nouri, M., Terada, H., Alfonso, E.C., Foster, C.S., Durand, M.L., Dohlman, C.H., 2001.
Endophthalmitis after keratoprosthesis: incidence, bacterial causes, and risk
factors. Arch. Ophthalmol. 119, 484.
Palmeira-de-Oliveira, A., Ribeiro, M.P., Palmeira-de-Oliveira, R., Gaspar, C., Costa-de-
Oliveira, S., Correia, I.J., Pina Vaz, C., Martinez-de-Oliveira, J., Queiroz, J.A.,
Rodrigues, A.G., 2010. Anti-Candida activity of a chitosan hydrogel: mechanism
of action and cytotoxicity profile. Gynecol. Obstet. Invest. 70, 322–327.
Parida, U.K., Nayak, A.K., Binhani, B.K., Nayak, P., 2011. Synthesis and characteriza-
tion of chitosan–polyvinyl alcohol blended with cloisite 30B for controlled
release of the anticancer drug curcumin. J Biomater. Nanobiotechnol. 2,
414–425.
Pattel, A., Cholkar, K., Agrahari, V., Mitra, A.K., 2013. Ocular drug delivery systems: an
overview. World J. Pharmacol. 2, 47–64.
Pavia, D.L., Lampman, G.M., Kriz, G.S., 1996. Introduction to Spectroscopy. Harcourt
Brace College, San Diego, California.
Pieróg, M., Ostrowska-Czubenko, J., Gierszewska-Druzy  ska, M., 2012.
_ n
Thermal degradation of double crosslinked chtiosan membranes. PCACD 17,
67–74.
Pujari, S., Siddique, S.S., Dohlman, C.H., Chodosh, J., 2011. The Boston keratopros-
thesis type II: the Massachusetts Eye and Ear Infirmary experience. Cornea 30,
1298–1303.
Refojo, M.F., 1969. Artificial membranes for corneal surgery. J. Biomed. Mater. Res. 3,
333–347.
Strampelli, B., 1963. Osteo-odontokeratoprosthesis. Ann. Ottalmol. Clin. Ocul. 89,
1039–1044.
Tan, D.T., Dart, J.K., Holland, E.J., Kinoshita, S., 2012. Corneal transplantation. Lancet
379, 1749–1761.
Verheul, R.J., Amidi, M., van Steenbergen, M.J., van Riet, E., Jiskoot, W., Hennink, W.E.,
2009. Influence of the degree of acetylation on the enzymatic degradation and
in vitro biological properties of trimethylated chitosans. Biomaterials 30,
3129–3135.
Waldock, A., Cook, S., 2000. Corneal transplantation: how successful are we? Br. J.
Ophthalmol. 84, 813–815.
Wong, A.C., Yiu, E.P., 2012. Boston keratoprosthesis for ocular chemical injury: one-
year clinical outcome. HKJ Ophtalmol. 16, 24–29.
Yang, Q., Dou, F., Liang, B., Shen, Q., 2005. Studies of cross-linking reaction on
chitosan fiber with glyoxal. Carbohydr. Polym. 59, 205–210.
Zhang, Y., Zhu, P.C., Edgren, D., 2010. Crosslinking reaction of poly(vinyl alcohol)
with glyoxal. J. Polym. Res. 17, 725–730.