Acta Biomaterialia 97 (2019) 310–320

Contents lists available at ScienceDirect

Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actbio

Full length article

Nanofiber-reinforced decellularized amniotic membrane improves

limbal stem cell transplantation in a rabbit model of corneal epithelial
defect
Zhengbing Zhoua,b,c, Da Longd,e, Chih-Chien Hsue,f, Huanhuan Liub,g, Long Chenb,c,g,
Benjamin Slaving, Hui Line, Xiaowei Lib,c,g,i, Juyu Tanga, Samuel Yiue, Sami Tuffahah,∗,
Hai-Quan Maob,c,g,∗
a
Department of Hand and Microsurgery, Xiangya Hospital of Central South University, Changsha, Hunan Province 410008, PR China
b
Translational Tissue Engineering Center, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA
c
Institute for NanoBioTechnology, Johns Hopkins University, Baltimore, MD 21218, USA
d
Department of Ophthalmology, The First Hospital of Hunan University of Chinese Traditional Medicine, Changsha, Hunan Province 410007, PR China
e
Wilmer Eye Institute, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA
f
Department of Ophthalmology, Taipei Veterans General Hospital, National Yang-Ming University, No. 201, Shipai Rd. Sec. 2, Taipei, Taiwan
g
Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD 21218, USA
h
Department of Plastic and Reconstructive Surgery, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
i
Mary and Dick Holland Regenerative Medicine Program, Department of Neurological Sciences, University of Nebraska Medical Center, Omaha, NE 68198,
USA

a r t i c l e i n f o a b s t r a c t

Article history: Human amniotic membrane (AM) offers unique advantages as a matrix to support the transplantation of
Received 19 March 2019 limbal stem cells (LSCs) due to its inherent pro-regenerative and anti-inflammatory properties. However,
Revised 13 August 2019
the widespread use of AM in clinical treatments of ocular surface disorders is limited by its weak me-
Accepted 15 August 2019
chanical strength and fast degradation, and high cost associated with preserving freshly isolated AM. Here
Available online 19 August 2019
we constructed a composite membrane consisting of an electrospun bioabsorbable poly(ε -caprolactone)
Keywords: (PCL) nanofiber mesh to significantly improve the ultimate tensile strength, toughness, and suture re-
Amniotic membrane tention strength by 4–10-fold in comparison with decellularized AM sheet. The composite membrane
Electrospinning showed extended stability and conferred longer-lasting coverage on wounded cornea surface compared
Nanofibers with dAM. The composite membrane maintained the pro-regenerative and immunomodulatory proper-
Composite membrane ties of dAM, promoted LSC survival, retention, and organization, improved re-epithelialization of the de-
Limbal stem cell deficiency
fect area, and reduced inflammation and neovascularization. This study demonstrates the translational
Transplantation
potential of our composite membrane for stem cell-based treatment of ocular surface damage.

Statement of Significance

Human decellularized amniotic membrane (dAM) has been widely shown as a biodegradable and bioac-
tive matrix for regenerative tissue repair. However, the weak mechanical property has limited its
widespread use in the clinic. Here we constructed a composite membrane using a layer of electrospun
poly(ε -caprolactone) (PCL) nanofiber mesh to reinforce the dAM sheet through covalent interfacial bond-
ing, while retaining the unique bioactivity of dAM. In a rabbit model of limbal stem cell (LSC) deficiency
induced by alkaline burn, we demonstrated the superior property of this PCL-dAM composite membrane
for repairing damaged cornea through promoting LSC transplantation, improving re-epithelialization, and
reducing inflammation and neovascularization. This new composite membrane offers great translational
potential in supporting stem cell-based treatment of ocular surface damage.
© 2019 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

∗
Corresponding authors at: 720 Rutland Avenue, Russ 749D, Baltimore, MD
21205, USA (S. Tuffaha). 3400 N. Charles Street, Croft Hall 100, Baltimore, MD, E-mail addresses: stuffah1@jhmi.edu (S. Tuffaha), hmao@jhu.edu (H.-Q. Mao).
21218, USA (H.-Q. Mao).
https://doi.org/10.1016/j.actbio.2019.08.027
1742-7061/© 2019 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320 311

1. Introduction maintaining the pro-regenerative and immunomodulatory proper-

ties to thereby provide enhanced healing of corneal epithelial de-
The corneal epithelium depends upon limbal epithelial stem fects with reduced inflammation and vascularity in the setting of
cells (LSCs) for continual maintenance and restoration of defects LSCD.
that would otherwise accumulate and compromise this critical bar-
rier [1]. LSCs can be depleted by a number of pathologic processes
2. Materials and methods
that include chemical, thermal, infectious, and iatrogenic injury, as
well as systemic autoimmune disorders such as Stevens-Johnson
2.1. Electrospun PCL nanofiber mesh preparation and activation of
syndrome and bullous pemphigoid [2]. Conditions that significantly
carboxyl groups on fiber surface
damage the LSCs can result in an invasion of conjunctival epithe-
lium on to the corneal surface. The resulting syndrome, termed
An electrospun PCL nanofiber mesh was fabricated as previ-
limbal stem cell deficiency (LSCD), is characterized by chronic in-
ously described [17]. Briefly, the electrospinning procedure con-
flammation of the corneal surface with overlying fibrovascular tis-
sisted of dissolving a 12 wt% PCL solution (viscosity average MW,
sue and subsequent loss of vision [3].
70–90 kDa; Sigma-Aldrich) in a mixture of dichloromethane and
Contemporary treatment options for LSCD depend upon patho-
methanol (4:1, v/v). This mixture was then injected at a rate of
logical severity [2]. Conservative treatments such as supportive
2.5 mL/h via a 27-gauge needle onto a grounded and stationary
management and corneal scraping can be effective in restoring the
plate collector. There was a 12-kV positive potential between the
epithelium in the setting of mild disease [4]. Amniotic membrane
needle and ground. The use of a grounded and stationary collec-
(AM) patching has also shown promise for partial LSCD, where
tor for deposition of fibers provided random fiber meshes. Mod-
residual native LSCs are available to populate the membrane. In
ifying the collection time allowed for control of the thickness of
contrast, effective treatment of severe LSCD requires both resurfac-
the nanofiber mesh to approximately 40 μm. The PCL fiber meshes
ing of the damaged corneal surface and reconstitution of the miss-
were air-dried in a laminar hood overnight and subjected to the
ing LSC population [5]. Corneal grafts in these cases are prone to
following sequential steps: (i) treating with plasma in a radio-
failure due to chronic inflammation of the graft bed and the inabil-
frequency plasma cleaner (Harrick Plasma) at a medium dose for
ity to reconstitute LSCs, as they are not available within the grafts
10 min, (ii) incubating the treated fiber mesh with 10% acrylic acid
to begin with. The use of LSC-containing limbal rim autograft has
solution containing 0.5 mM NaIO4 , followed by exposure to ultra-
demonstrated greater efficacy; but is limited by the morbidity as-
violet light at an intensity of 30–50 mW/cm2 for 2 min in an ice
sociated with harvesting the graft from the contralateral eye [6].
bath (0–4 °C), (iii) removing un-grafted poly(acrylic acid) (PAAc)
More recently, transplantation of cultured, in vitro-expanded LSCs
from the surface of the nanofiber mesh by washing with deion-
has been explored as a means of limiting donor site morbidity
ized (DI) water, (iv) measuring the concentration of the carboxyl
[7]. However, the cell transplantation approach requires a sub-
groups on fiber mesh by the Toluidine Blue O (TBO) assay, (v)
strate to resurface the damaged corneal membrane and facilitate
activating the surface carboxylic groups by reacting the treated
engraftment.
mesh with N-hydroxysuccinimide (NHS, Sigma) and equal molar
Human decellularized amniotic membrane (dAM) has been
of N-(3-dimethylaminopropyl)-N-ethyl carbodiimide (EDC, Sigma)
used as a temporary biological wound dressing and a substrate
in 50% ethanol in a glass dish for 4 h at a molar ratio of 1:4:4 for
for LSC transplantation in LSCD treatment [8]. Numerous clinical
COOH:NHS:EDC, (vi) removing excess reagents by thorough rinse
cases and animal studies have demonstrated the efficacy of either
with 70% ethanol for three times at 5 min/wash, and (vii) drying
freshly isolated or cryopreserved dAM for this purpose [9,10]. In
the modified mesh in a laminar flow hood. The modified nanofiber
addition to its function as a cell carrier, the native growth factors
mesh was stored at −20 °C and rinsed with sterile phosphate-
and cytokines within dAM promote corneal repair and regenera-
buffered saline (PBS) before use. The amounts of residual solvents
tion when used to treat chronic corneal diseases [11]. Due to the
(dichloromethane and methanol) were analyzed by gas chromatog-
high cost associated with the preservation and delivery of fresh
raphy and found to be below 200 ppm (Impact Analytical). These
human dAM [12], decades of research efforts have focused on the
residue contents are significantly lower than the recommended
optimization of processing, with the goal of elucidating methods of
levels (600 and 2000 ppm, respectively) by the International Coun-
cost-effective storage and transportability. Currently available cry-
cil for Harmonisation of Technical Requirements for Registration of
opreserved dAM products, including AmnioGraft® and NEOX®, are
Pharmaceuticals for Human Use (ICH).
also limited by a lack of structural integrity, which hinders intraop-
erative handling and suturing, as well as a rapid degradation rate
that introduces the potential need for re-application [13–15]. 2.2. Preparation of decellularized human amniotic membrane
In order to address these limitations, we have developed a new
composite membrane consisting of an electrospun, bioabsorbable Composite membrane preparation was carried out through the
polymer fiber mesh bonded to a human dAM sheet through in- use of cryopreserved human AMs from a commercial supplier
terfacial conjugation. In vitro experiments have shown that inter- (AmnioGraft®, Bio-Tissue Inc.). The epithelial cell layer of AM
facial bonding substantially improves the tensile properties and was removed by dispase treatment. Once thawed, the sample of
toughness of dAM, making the composite membrane better able human AM was placed onto a membrane comprised of nitrocel-
to retain sutures and conform to the convex surface of the globe lulose. The epithelial cell layer was left exposed. A 2.5% (w/v)
[16]. Additional benefits observed in vivo include maintenance of solution was made through the dissolution of dispase (Millipore)
proliferation and biological activity of transplanted LSCs, as well in a DMEM/F12 medium (Life Technology). The nitrocellulose
as improved biocompatibility as seen by increased polarization membrane, along with the AM to which it was still adherent, was
of macrophages to an anti-inflammatory phenotype compared to kept for 5 h at 4 °C in a 2.5% dispase solution and subsequently
other AM formulations. triple washed using sterile PBS. An Iris spatula (Fine Science Tools)
In this study, we tested the poly(ε -carprolactone) (PCL) was utilized with the aid of a stereoscope (Nikon SMZ800) in
nanofiber-dAM composite membrane in comparison to dAM alone, order to remove the layer of epithelial cells. Lastly, the decellu-
with and without cultured LSCs, in a rabbit LSCD corneal epithe- larized AM (dAM) was triple washed with sterile PBS to ensure
lial defect model. We hypothesized that the composite membrane that it was ready for use. A DNA isolation kit (DNeasy Blood
would enhance the stability and durability of dAM alone while & Tissue Kit, QIAGEN) was employed in order to quantify the
312 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320
amount of DNA present as a measure to evaluate the effective-
ness of removal of the epithelial cells. After the decellularization
procedure, the total content of DNA in AM was 20.7 ± 13.5 ng/mg
(n = 5 separate samples), which was <5% of the level in native AM
(488.2 ± 140.7 ng/mg). These results confirm that the total level
of DNA in the dAM is below 50 ng/mg, consistent with the FDA
standard for a similar decellularized extracellular matrix products
[18]. As an additional evidence, the absence of cell nuclei on
the processed dAM was confirmed by examining the membranes
under the fluorescence microscope after treatment.
2.3. Fabrication of a composite membrane via conjugation of the
fiber mesh onto the dAM
We have previously described the fabrication of a compos-
ite membrane with conjugation of the electrospun PCL nanofiber
mesh to the dAM [16]. In brief, the dAM was placed on a sheet
of polytetrafluoroethene (PTFE). The epithelial side of the dAM
was oriented adjacently to the PTFE. Next, the surface-activated
nanofiber mesh was placed on the stromal side of the dAM, with
a second sheet of PTFE being positioned next to the nanofiber
mesh. The assembled layers were subsequently placed between
two metal plates, which applied 6 h of mechanical compression at
a temperature of 4 °C in order to maintain total thickness of the
composite at 20 μm. This technique ensured that there was suf-
ficient contact between the dAM and nanofibers. A condensation
reaction between the carboxyl groups found on the surface of the
fibers and the amino groups of the protein constituents found on
the dAM was used to complex the nanofibers to the dAM. Follow-
ing removal of the metal plates and sheets of PTFE, a freeze dryer
(Labconco Freezone 12L Cascade Freeze Dry System) was utilized
in order to lyophilize the composite membrane, which was subse-
quently stored at −20 °C until needed for further utilization.
2.4. Characterization of degradation rate
The degradation rate of the PCL-dAM composite was measured
using the method described previously in the literature [19]. In
brief, the dAM, PCL, and PCL-dAM membranes were placed in ei-
ther the 1-U pancreatin buffer or PBS as a buffer control. To de-
termine mass loss at different time points, samples were washed
with deionized (DI) water 3 times, and lyophilized; and then mass
loss was calculated accordingly for 1 week.
2.5. Rabbit LSC isolation and culture
Rabbit LSCs were isolated as previously described [16]. The
limbus of fresh New Zealand white rabbit eyeballs (purchased
from Pel-Freez Biologicals) was harvested with a scalpel under
magnification and incubated with 2.5 mg/ml dispase for 12 h at
a temperature of 4 °C. Single cells were obtained by further di-
gestion with 0.25% trypsin/EDTA for a time of 5 min. After us-
ing a trypsin inhibitor for blocking, a centrifuge was utilized at
1200 rpm for a time of 5 min in order to collect the cells. The
cells were then re-suspended in LSC culture medium contain-
ing a known keratinocyte-SFM basal medium (Invitrogen), known
keratinocyte-SFM growth supplement (Invitrogen), 1% antibiotic-
Antimycotic (Gibco), and a GFP Lentiviral Vector (ABM Inc.).
The dAM or PCL-dAM membrane was trimmed into a circu-
lar shape with a diameter of 20 mm, and then wound into a
polypropylene tube with an 18 mm diameter using a fine copper
wire to allow for the preparation of a cell culture cup. After soak-
ing for 3 h, the membranes were rinsed in PBS three times and
then soaked for an additional hour. They were then placed into
16-well plates. Corneal limbal stem cells were subsequently cul-
tivated at a density of 150 × 104 per culture cup, while the cells
were transduced with an exogenous GFP gene. After one week, the
cells covered the entire material. The cell side faced the cornea and
covered the surface of the rabbit cornea when repairing LSCD.
2.6. Alkali Burn-Induced LSCD model
Surgical and animal care procedures were conducted under
a protocol approved by the Institutional Animal Care and Use
Committee at Johns Hopkins University according to guidelines
established by National Institutes of Health and American As-
sociation for the Accreditation of Laboratory Animal Care. Alkali
burn-induced LSCD was performed as detailed in previous work
[20,21]. In brief, ketamine (100 mg/kg) and xylazine (5 mg/kg) were
injected intramuscularly into the gluteus maximus of the rabbits
in order to achieve anesthetization. This study exclusively used
the right eye of all rabbits. Ring-shaped sterile filter paper that
had been saturated with 1 N NaOH was placed onto the cornea
for 50 s. Following a rinse with 0.9% NaCl, steroids and antibiotic
eye drops (0.5% moxifloxacin and 0.1% dexamethasone) (Vigadexa,
Alcon) were applied to the rabbit eyes following an alkaline burn,
twice daily for 2 weeks. Corneal epithelial defect (CED), opacity of
the cornea, vascularization of the cornea, and conjunctivalization,
when present, were counted as signs of LSCD.
2.7. Transplantation of LSCs on PCL Fiber-dAM composite membrane
in a severe LSCD model
A total of 25 New Zealand white rabbits (8 weeks and
1.5 ± 0.2 kg) with LSCD model established in the right eye accord-
ingly to the procedure described above were grouped randomly
into 5 groups (5 rabbits per group), which received treatments
with dAM, PCL membrane, PCL-dAM composite membrane, 3 × 106
LSC-seeded dAM (LSC-dAM), and LSC-seeded PCL-dAM compos-
ite membrane (LSC-PCL-dAM), respectively, in the damaged eye at
2 weeks after alkali burn to induce LSCD. Under general anesthe-
sia, 0.2% fluorescein staining was performed to examine the repair
of the corneal surface. After the eyes were washed with saline,
a keratome was used to scrape the ocular surface scar tissue. A
graft was then placed on the ocular surface and fixed to the sclera
with a Nylon 10-O suture (Ethilon Nylon Suture, Johnson & John-
son Medical Device). An ointment combination of antibiotic and
steroid was applied to the surgically modified eye twice per day for
5 consecutive days. After two weeks, the samples were collected.
2.8. Immunohistochemistry and histology
Staining of frozen sections (8 μm) of cornea was done for rabbit
anti-CD86 (1:500, Cell Signaling) anti-KRT12 (1:200, ABIN290007,
Antibodies-online), anti-KRT13 (1:200, ABIN108618, Antibodies-
online) and anti-KRT3 (1:200, ABIN108617, Antibodies-online) as
detailed previously. In brief, the frozen sections were desiccated at
room temperature and post-fixed in chilled acetone for a time of
15 min. The sections were then washed with phosphate-buffered
saline (PBS), followed by blocking with 8% normal donkey serum in
PBS for a time of 1 h at room temperature. The samples were sub-
sequently incubated overnight at a temperature of 4 °C with one of
the primary antibodies mentioned above. Following triple wash-
ing with PBS/0.05% Tween-20, fluorescently-labeled secondary
antibody of CyTM 5 AffiniPure Donkey Anti-Mouse IgG (1:200, 715–
175-151, Jackson ImmunoResearch) and CyTM 3 AffiniPure Donkey
Anti-Goat IgG (1:200, 705–165-147, Jackson ImmunoResearch)
were incubated along with the samples at room temperature for
1 h. Following an additional three washes with PBS, the sections
were counterstained for a time of 3 min at room temperature us-
ing DAPI, a nuclear dye (1:1200; Sigma, St. Louis, MO). The slides
were mounted using an Aquamount solution and fluorescence
 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320
microscopy (Axioskop; Carl Zeiss Meditec, Thornwood, NY, USA)
was then utilized to visualize the samples. An identical exposure
time was used for each sample, and the images were analyzed
using Image-Pro Plus (Media Cybernetics, Silver Spring, MD).
2.9. RNA isolation and real-time polymerase chain reaction analysis
Real-time polymerase chain reaction (PCR) analysis was used in
order to assess the levels of mRNA of IL-6, IL-8, CCL2, and VEGF
within rabbit cornea treated with different materials. The primers
for IL-6 are GAA CAG AAA GGA GGC ACT GG (forward) and CTC
CTG AAC TTG GCC TGA AG (reverse); the primers for IL-8 are CCA
CAC CTT TCC ATC CCA AAT (forward) and CTT CTG CAC CCA CTT
TTC CTT G (reverse); the primers for CCL2 are GTG AAT TCC CCA
GTC ACC TG (forward) and TGT GTT CTT GGG TTG TGG AA (re-
verse); and the primers for VEGF are TTC AAC GTC ACC ATG CAG
AT (forward) and AAA TGC TTT CTC CGC TCT GA (reverse).
Isolation of total RNA was accomplished via RNeasy Mini Kits
(QIAGEN). First-strand cDNA synthesis reactions were carried out
though the use of a High Capacity cDNA Reverse Transcription Kit
(ThermoFisher Scientific) with an accompanying light cycler com-
ponent (Biometra TPERSONAL, German). Real-time PCR was car-
ried out through the use of the TaqMan Gene Expression Mas-
ter Mix (ThermoFisher Scientific) with a light cycler component
(StepOnePlusTM Real-Time PCR System, Applied BiosystemsTM ). PCR
cycling involved 40 cycles of amplification of the template DNA
with primer annealing occurring at a temperature of 60 °C. The
2−Ct method was then used in order to calculate the relative
expression level of each gene of interest. The amplification efficien-
cies of each primer pair were validated in order to allow for gene
expression to be quantitatively compared. Only Taqman primers
(ThermoFisher Scientific) were utilized. Every quantitative PCR was
carried out three times on three different experimental replicates.
The results were subsequently normalized to the value found using
the reference gene.
2.10. Statistical analysis
All quantitative data are expressed as a mean ± SD. Statistically
significant differences between the results of various experimental
groups were assessed via a student’s t-test. A p value of <0.05 is
considered to be indicative of statistical significance.
3. Results
3.1. Composite membrane preparation and stability
The composite membrane was prepared accordingly to our pre-
viously reported protocol [16]. Mechanical compression was ap-
plied to the tightly bonded dAM and PCL nanofiber mesh in order
to facilitate the conjugation reaction between the PCL fiber sur-
face and dAM components. The different layers that bonded to-
gether formed a composite membrane without delamination fol-
lowing lyophilization (Fig. 1). This interfacial bonding technique
not only is important to improve the integrity of the composite
membrane, but also improved the mechanical properties and sta-
bility of the membrane. The PCL-dAM composite membrane exhib-
ited about 4-fold higher ultimate tensile strength (p < 0.01), 10-fold
higher toughness (p < 0.01), and more than 10-fold higher suture
retention strength (p < 0.01) in comparison with lyophilized dAM
[16].
To determine whether this composite has the capacity to in-
crease the stability of dAM when applied to the damaged corneal
surface, we tested the degradation rate of dAM, PCL, and the PCL-
dAM composite membrane in a pancreatin buffer solution (1 U),
313
using PBS as a buffer control. From Day 1–4, dAM showed signif-
icantly more rapid degradation than PCL or the PCL-dAM compos-
ite membrane. On Day 5, the PCL only membrane maintained its
original shape and size in 1-U of pancreatin buffer, while the dAM
only membrane completely dissolved (Fig. 1). The composite slows
down the degradation rate by about 40%. These results demon-
strate that PCL bonding has the potential to extend the stability
of dAM in the presence of enzyme, thus improving its stability and
lasting coverage on a wound surface.
3.2. Alkali Burn-Induced LSCD model
We adopted an LSCD model using alkali burn. Following ex-
posure to alkali and scraping of the epithelium, typical signs of
corneal damage, including conjunctival congestion and epithelial
damage, were observed within a few hours. Inflammation involving
the anterior segment, such as soft tissue swelling of the operative
eye and white dense exudation of the anterior chamber, developed
within the first few weeks (Fig. 2). During the second week, exuda-
tion disappeared, and limbus neovascularization was indicative of
chronic inflammation. H&E staining confirmed the loss of the cen-
tral corneal epithelium and distortion of the limbal epithelium at
2 weeks post-exposure.
3.3. Assessing PCL-dAM composite membrane for corneal defect
repair in the LSCD model
3.3.1. In vivo anti-inflammation of PCL-dAM composite membrane
We demonstrated previously that the PCL-dAM composite
membrane has similar immunomodulatory effects as the dAM
and may promote the polarization of macrophages into the pro-
regenerative M2 phenotype in vitro [16]. In order to investigate the
immunomodulatory effects of PCL-dAM composite membrane in
vivo, we transplanted the dAM only membrane or PCL-dAM com-
posite membrane to the corneal epithelial defect site in a rab-
bit LSCD model. Corneal edema and inflammatory reactions grad-
ually reduced from week 1 after transplantation in both groups.
The PCL-dAM composite membrane was well-seated on the corneal
surface without degradation at week 1, while the majority of grafts
to which dAM alone had been applied demonstrated thinning and
shifting (Fig. 3). Immunohistochemistry showed a high level of
CD86-positive M1 macrophages in the control group compared
with the dAM or PCL-dAM composite membrane groups. However,
CD68, a marker of the M2 macrophage, showed negative staining
in all groups. Furthermore, real time PCR revealed lower levels of
CCL2, IL-8, and IL-6 mRNA expression in both the dAM and PCL-
dAM composite membrane groups compared with untreated con-
trols. The high expression of IL-6 mRNA was consistent with an
increased CD86 staining intensity in the control group. Reduced
VEGF mRNA expression in treated corneas indicates that both dAM
and PCL-dAM composite membrane are effective in preventing an-
giogenesis through inflammatory stimulation (Fig. 4).
3.3.2. PCL-dAM composite membrane promoting conjunctival
epithelialization of cornea in LSCD Model.
Conjunctivalisation is typically considered as a pathological pro-
cess during corneal re-epithelialization in the setting of severe
LSCD condition because neovascularization and inflammatory cell
infiltration typically accompany conjunctivalization [22]. However,
integrated conjunctival epithelialization can occur without neovas-
cularization and inflammation and contribute to corneal repair, in
which case it serves as a beneficial process to repair. In this study,
degradation of the dAM was observed in the first week follow-
ing transplantation. Conversely, the PCL and PCL-dAM exhibited
excellent stability until 4 weeks post-transplantation. Immunohis-
tochemistry demonstrated that K13-positive cells specific to the
314 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320

Fig. 1. General appearance and degradation profiles of the PCL-dAM composite membrane. (A) General appearance and (B) SEM images of dAM, PCL and the PCL-dAM
composite membranes prepared from the PCL fiber mesh and dAM. (C, D) The degradation profiles of the three samples in PBS (C) and in the presence of enzyme (1 U of
pancreatin) (D). ∗ p < 0.001 (in comparison with dAM or PCL mesh).

conjunctival epithelial cells continuously lined the corneal lesion
site in the PCL-dAM composite membrane group but formed a
much sparser dotting pattern in the dAM only group (Fig. 4). K12-
positive cells, considered a marker for basal corneal epithelium,
were not identified in either group. This finding implies that PCL-
dAM composite membrane, when used as a protective dress for
damaged cornea, can provide a stable environment for the prolif-
eration of conjunctival epithelial cells and functional conjunctival-
ization, even in the absence of LSCs.
3.4. Transplantation of LSCs with dAM vs. PCL-dAM composite
membrane for corneal defect repair in severe LSCD model
LSCs were delivered to the corneal defects in combination with
the membranes. GFP-labeled LSCs were cultured on either a dAM
or PCL-dAM composite membrane for 1 week prior to implantation.
Two weeks after implantation, conjunctival hyperemia and ocular
irritation were noticeably relieved in both groups. In the PCL-dAM
composite membrane group, the corneal epithelium gradually
recovered with improved transparency and significantly reduced
neovascularization. In contrast, the corneal epithelial defects
persisted in the dAM group and were gradually covered by the
conjunctival epithelium, while the cornea became cloudy with ex-
tensive neovascularization. GFP-labeled LSCs were identified in the
newly-formed epithelial layer in both groups received dAM + LSCs
and CM + LSCs (Fig. 5). A higher number of LSCs were found in the
group received CM + LSCs; and LSCs in this group spread over a
greater area than those identified in the group with dAM + LSCs.
The corneal epithelium is the outermost layer of the cornea.
It is composed of a single layer of basal cells and four to five
cell layers of nonkeratinized, stratified squamous epithelial cells.
Immunohistochemistry of K12 and K3, which are cornea-specific
keratins that compose the intermediate filament cytoskeleton of
corneal epithelial cells, showed linear and spotted staining, respec-
tively, in the “CM + LSCs” group. In contrast, the “dAM + LSCs” group
demonstrated predominately K12 staining and minimal K3 stain-
ing. H&E staining of the sections at week 2 following compos-
ite membrane-supported LSC transplantation revealed restoration
of corneal epithelial morphology and hierarchy. Specifically, highly
cylindrical basal cells, the upper polygonal pterygium, and super-
ficial flat cells were very similar to that of a normal corneal ep-
ithelium. Conversely, in the tissue sections from the “dAM + LSCs”
 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320 315

Fig. 2. Alkali-burn induced LSCD model results in damage to corneal epithelium. (A) Procedure of LSCD model by alkali burn. Normal: The naive rabbit eye; Day 0–1 N
NaOH: Ring-shaped sterile filter paper saturated with 1 N NaOH is placed onto the cornea for 50 s; Day 0-Rinse: the lesion site is then rinsed with 0.9% NaCl for 5 min;
Day 0-Scraping: residual corneal epithelium is scraped; Day 0-Staining: corneal damage is confirmed with fluorescein staining; Week 2-LSCD: Cornea becomes opaque as
neovascularization occurs at 2-weeks post-exposure. (B) H&E staining showing (left) normal corneal epithelial and (right) loss or distortion of limbal epithelium at 2-week
post-exposure on the LSCD model eye.

group, most epithelial cells were bulky, fusiform or polygonal, and
the connections between cells were loose. In addition, cells were
irregularly arranged, and ring-like goblet cells were observed in the
control group. The presence of basal cells is the most important ev-
idence of corneal epithelial normalization and conjunctivalization.
Comparison of basal cell count in each 40 × field for the normal,
control, dAM, PCL-dAM composite membrane, “dAM + LSCs”, and
“CM + LSCs” groups showed eyes treated with CM + LSCs had basal
cell counts closest to those of normal cornea. The cornea covered
by PCL-dAM composite membrane without LSCs exhibited better
epithelial morphological repair than dAM-covered cornea, however
there was no significant difference in basal cell count in either
group compared control (Figs. 5 and 6).
4. Discussion
Ocular burn is one of the most common forms of LSCD. Ex-
perimental alkaline burns is typically used to mimic ocular burns,
which causes severe depletion of LSCs and local inflammation,
and leads to cornea opacity and neovascularzation [23]. A grow-
ing body of work has been reported to demonstrate the promise
of LSC transplantation in treating alkaline burn-induced LSCD in
animal models [24–26]. The success of this approach, however, re-
quires a scaffold to resurface the damaged corneal epithelium and
enhance LSC engraftment. AM is well-suited for this purpose, given
its immunomodulatory and pro-regenerative properties. AMs have
many intrinsic cytokines that promote wound healing by stimulat-
ing re-epithelialization and reducing local inflammation [27]. How-
ever, the unique microenvironment of the corneal surface, as well
as shear forces from eyelid motion, make the grafted dAM more
susceptible to deformation and degradation. Consequently, clinical
LSC transplantation with dAM membranes must often be repeated
due to a lack of stable coverage [28]. Additionally, the thin and
fragile nature of AM makes it difficult to handle during the process
of implantation to the ocular surface and susceptible to tear during
suturing. Unlike approaches of chemical or enzymatic crosslinking
316 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320

Fig. 3. PCL-dAM composite membrane maintains structural integrity while reducing vascularization. PCL-dAM composite membrane (CM) remained intact without degra-
dation while dAM had almost disappeared after 1 week. Two weeks after transplantation, conjunctival hyperemia and ocular irritation had been noticeably relieved in both
groups. Compared with severe limbal neovascularization and corneal edema in control group, inflammatory reactions were reduced in all treatment groups. There was no
limbal neovascularization seen with the group received CM + LSCs. Fluorescein staining at 2-week demonstrated that the majority of corneal epithelial surface had recovered,
while higher percentages of the corneal surface stained positively in groups received dAM without LSCs, CM without LSCs, or dAM with LSCs.

that result in loss of bioactivities of the active growth factors and
cytokines found in native AM [29,30], others have reported appli-
cation of amniotic extracts in a gel form to retain more of the ac-
tive compounds [31]. However, extracting soluble AM components
will compromise the extracellular matrix structure, which has been
shown to play a role in facilitating re-epithelialization [32,33].
To address these limitations, we developed a novel compos-
ite membrane that improves the stability and toughness of AM
while maintaining its pro-regenerative and immunomodulatory
properties. The fabrication process involves interfacial bonding
of biodegradable, electrospun, surface-carboxylated PCL nanofiber
mesh to dehydrated amnion sheets. This avoids the loss of AM
biological activity that results from stratification, shedding, and
repeated chemical crosslinking used in other processing meth-
ods. Meanwhile, physical properties of the composite membrane,
including toughness and stability, were enhanced as a result of
effectively increasing the membrane thickness and slowing the
rate of degradation. During the transplantation procedure, the
 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320 317

Fig. 4. PCL-dAM composite membrane promotes conjunctival corneal epithelialization in the absence of LSCs the LSCD model. (A) Real time PCR of IL-8, IL-6, CCL2, VEGF
mRNA expression. (n = 5, ∗ p < 0.05, ∗ ∗ p < 0.01, ∗ ∗ ∗ p < 0.001, ∗ ∗ ∗ ∗ p < 0.0 0 01). (B) Immunohistochemistry showing K13-positive cells continuously lining the corneal lesion site
in PCL-dAM composite membrane (CM) group but forming a dotted pattern in dAM group. As expected, CD86-positive staining was observed more in the control group and
dAM than the PCL-dAM composite membrane group. Scale bar: 100 μm.

resilient composite membrane, with a thickness of only 20 μm,
largely improved operability over dAM alone, with the latter
demonstrating a tendency to easily tear during application to a
curved surface. In this study, the composite membrane demon-
strated stable attachment to the cornea and resistance to tearing
during suture placement and shifting from shear forces. The dAM
membrane only and dAM-supported LSC transplantation group
both have shown shifted and fully degraded membrane within one
week. In contrast, the PCL-dAM composite membrane maintained
its original shape, position and structural integrity for at least two
weeks following implantation, at which point all membranes were
explanted. We therefore speculate that the enhanced epithelial
formation in the composite membrane group in vivo was due to
the greater durability it conferred to dAM thereby providing a
more stable environment for cellular attachment, migration and
differentiation during the healing process.
In a typical rabbit LSCD model, a large number of neutrophils
infiltrated the cornea during an acute inflammatory response.
These neutrophils released proteolytic enzymes, including collage-
nase and other matrix metalloproteinases, causing progressive ly-
sis of the corneal stroma as well as the release of activated poly-
morphonuclear leukocytes, and stimulation of the expression of
inflammatory factors. The dAM membrane when used as a graft
proved capable of secreting a variety of active growth factors and
cytokines that promote the apoptosis of polymorphonuclear leuko-
cytes, while inhibiting corneal neovascularization and mitigating
the local inflammatory response [34,35].
In the present study, we used dAM membrane or PCL-dAM
composite membrane to repair corneal epithelial defects in a
rabbit model of LSCD. We observed greater epithelialization and
reduced fluorescein staining, both indicators of improved corneal
healing, in the PCL-dAM composite membrane group compared
318 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320
Fig. 5. PCL-dAM composite membrane improves LSC transplantation and the extent of corneal epithelialization in the LSCD model. Immunohistochemistry of K12 and K3
demonstrated linear and spotted staining in the PCL-dAM composite membrane (CM) supported LSC transplantation group (CM + LSCs) while K12 staining predominated over
K3 in dAM supported LSC transplantation group (dAM + LSCs). Scale bar: 100 μm.
with the untreated control and dAM group. The control group
developed the greatest degree of corneal opacity and neovascu-
larization, while the ocular surface neovascularization and opacity
were significantly decreased in the dAM and PCL-dAM composite
membrane groups. PCR demonstrated diminished IL-8, IL-6, and
VEGF mRNA in both the dAM and PCL-dAM composite membrane
repaired corneas compared with the control group. Taken together,
these findings support the hypothesis that the PCL-dAM composite
membrane maintains the pro-regenerative and anti-inflammatory
properties of the dAM components in vivo. It is worth noting
that in the absence of transplanted LSCs, PCL-dAM composite
membrane used in the setting of LSCD demonstrated a functional
pattern of conjunctivalization without neovascularization or in-
flammation, which was not observed in dAM membrane group.
Nonetheless, given that conjunctivalization has the potential to be
pathological, long term studies are needed to fully characterize the
repaired cornea structure and ensure the stability of the observed
pattern of healing.
The normal proliferation and differentiation of the epithelial
cells of the ocular surface depends on the presence of a nor-
mal epithelial basement membrane [36]. In LSCD, where there is
destruction of the epithelial basement membrane, AM provides
an ideal substitute substrate capable of supporting LSC growth,
given its tissue composition and micro-architecture that closely re-
sembles that of the ocular epithelium basement membrane [37].
In order to assess the suitability of PCL-dAM composite mem-
brane as a substrate for LSC transplantation in comparison with
dAM, GFP-labeled LSCs were seeded on the respective membranes
and tested in an LSCD corneal defect model. Both PCL-AM com-
posite and dAM membranes allowed for successful engraftment
and survival of transplanted LSCs. However, the labeled LSCs in
the group received PCL-dAM composite membrane-supported LSCs
were distributed in a more uniform band and covered a greater
surface area of the cornea in contrast to the punctuated ep-
ithelial fluorescence bands observed in the group received dAM-
supported LSCs. H&E staining revealed that the composite mem-
brane + LSCs group had better epithelial delaminated cell morphol-
ogy than the AM + LSCs group. The corneal stromal collagen con-
tent of the composite membrane group was close to that of the
normal control group, thus implying that the PCL-dAM compos-
ite membrane is more conducive to support LSC survival, reten-
tion, migration, and proliferation than dAM alone. Future studies
will focus on identifying the long-term fate of the transplanted
LSCs.
It is also important to note that there were no differences in the
number of basal cells observed in the PCL-dAM only and dAM only
groups without LSC transplantation in comparison with the un-
treated control group, despite observed improvements in collagen
organization and re-epithelialization, and inhibiting inflammation
and neovascularization. Given that the presence of basal cells is the
most important indicator of normal corneal healing, this finding
reinforces the importance of cell transplantation as an adjunct to
membrane resurfacing in the treatment of LSCD. Certainly, the im-
proved mechanical property and extended stability of the PCL-dAM
composite membrane further enhance the performance of the dAM
as a supporting matrix to harness the full potential of LSC trans-
plantation.
Other cell types such as limbal epithelial cells have been
tested as alternative to LSC transplantation. For example, the
simple limbal epithelial transplantation technique has been shown
to successfully repair damaged cornea in the case of unilateral
severe LSCD, as an alternative to conjunctival limbal autograft or
cultivated limbal epithelial transplantation [38]. Due to the rich
component this PCL-dAM composite membrane would also likely
to augment the treatment outcomes for transplantation of limbal
epithelial cells.
 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320 319

Fig. 6. PCL-dAM composite membrane group displays improved epithelial delaminated cell morphology. (A) H&E staining of the sections at week 2 following implantation
of various membranes. Scale bar: 100 μm. (B) Different layers and cell types of corneal epithelium. Scale bar: 20 μm. (C) Comparison of basal epithelial cell count in each
40 × field for the normal, control, dAM without L SCs (dAM w/o L SCs), PCL-dAM composite membrane without LSCs (CM w/o LSCs), dAM with LSCs (dAM + LSCs), and PCL-
dAM composite membranes with LSCs (CM + LSCs) (n = 5).

Although non-essential, it may be ideal that the corneal repair
membrane is optically transparent. As the PCL nanofiber mesh in
the composite membrane is not transparent, the whole membrane
is not transparent even after hydration. Nonetheless, we found that
re-epithelialization occurs within the dAM sheet that interfaces
with the corneal surface. Following integration of the LSC and re-
epithelialization and the degradation of the dAM components, the
PCL mesh can be easily removed, so as not to obscure vision af-
ter healing. A transparent or translucent version of the compos-
ite membrane can be produced by using a transparent nanofiber
sheet, which is currently ongoing.
5. Conclusions
The PCL-dAM composite membrane maintains and augments
the beneficial pro-regenerative and immunomodulatory prop-
erties of dAM, with enhanced ability to support the survival,
retention, migration and organization of LSCs, thus promoting
the re-epithelialization and reducing inflammation and neovascu-
larization in the repair site of a rabbit LSCD model. In addition
to preserving the biological activity of dAM, the addition of the
PCL nanofibers stabilizes and toughens the dAM to provide more
durable and longer-lasting coverage of the defect site during the
healing process. The PCL-dAM composite membrane therefore
provides a superior alternative to cryopreserved dAM with greater
clinical utility and practicality.
Declaration of Competing Interest
The authors declare no competing financial interest.
Acknowledgements
This work was partially supported by Johns Hopkins University
School of Engineering and a grant from Wuhan Kangchuang Tech-
nology Co. Ltd. to Johns Hopkins University.
References
[1] J. Vazirani, D. Nair, S.S. Shanbhag, S. Wurity, A. Ranjan, V. Sangwan, Limbal
stem cell deficiency—demography and underlying causes, Am. J. Ophthalmol.
(2018) 99–103.
[2] M. Pauklin, K.P. Steuhl, D. Meller, Characterization of the corneal surface in
limbal stem cell deficiency and after transplantation of cultivated limbal ep-
ithelium, Ophthalmology 116 (6) (2009) 1048–1056.
[3] Q. Le, J. Xu, S.X. Deng, The diagnosis of limbal stem cell deficiency, Ocul. Surf.
16 (1) (2018) 58–69.
[4] A. Solomon, P. Ellies, D.F. Anderson, A. Touhami, M. Grueterich, E.M. Espana,
Long-term outcome of keratolimbal allograft with or without penetrating ker-
atoplasty for total limbal stem cell deficiency, Ophthalmology 109 (6) (2002)
1159–1166.
320 Z. Zhou, D. Long and C.-C. Hsu et al. / Acta Biomaterialia 97 (2019) 310–320
[5] M. Grueterich, E.M. Espana, A. Touhami, S.E. Ti, S.C. Tseng, Phenotypic study
of a case with successful transplantation of ex vivo expanded human limbal
epithelium for unilateral total limbal stem cell deficiency, Ophthalmology 109
(8) (2002) 1547–1552.
[6] M.S. Santos, J.A. Gomes, L. Hofling-Lima, Survival analysis of conjunctival lim-
bal grafts and amniotic membrane transplantation in eyes with total limbal
stem cell deficiency, Am. J. Ophthalmol. 140 (2) (2006) 223.e1 9.
[7] G. Amescua, M. Atallah, N. Nikpoor, A. Galor, V.L. Perez, Modified simple limbal
epithelial transplantation using cryopreserved amniotic membrane for unilat-
eral limbal stem cell deficiency, Am. J. Ophthalmol. 158 (3) (2014) 469–475.e2.
[8] S.C. Tseng, P. Prabhasawat, K. Barton, Amniotic membrane transplantation with
or without limbal allografts for corneal surface reconstruction in patients with
limbal stem cell deficiency, Arch. Ophthalmol. 116 (4) (1998) 431–441.
[9] A.L. Sabater, V.L. Perez, Amniotic membrane use for management of corneal
limbal stem cell deficiency, Curr. Opin. Ophthalmol. 28 (4) (2017) 363–369.
[10] Q. Le, S.X. Deng, The application of human amniotic membrane in the surgical
management of limbal stem cell deficiency, Ocul. Surf. 17 (2) (2019) 221–229.
[11] P. Prabhasawat, N. Tesavibul, W. Komolsuradej, Single and multilayer amniotic
membrane transplantation for persistent corneal epithelial defect with and
without stromal thinning and perforation, Br. J. Ophthalmol. 85 (12) (2001)
1455–1463.
[12] K. Jirsova, G.L. Jones, Amniotic membrane in ophthalmology: properties, prepa-
ration, storage and indications for grafting—a review, Cell Tissue Bank. 18 (2)
(2017) 193–204.
[13] C.L. Allen, G. Clare, E.A. Stewart, M.J. Branch, O.D. Mcintosh, M. Dadhwal,
A. Hopkinson, Augmented dried versus cryopreserved amniotic membrane as
an ocular surface dressing, PLoS ONE 8 (10) (2013) e78441.
[14] C.E. Uhlig, C. Frings, N. Rohloff, C. Harmsenaasman, R. Schmitz, L. Kiesel,
A.F. Alex, Long-term efficacy of glycerine-processed amniotic membrane trans-
plantation in patients with corneal ulcer, Acta Ophthalmol. 93 (6) (2015)
e481–e487.
[15] Y. Zhao, L. Ma, Systematic review and meta-analysis on transplantation of ex
vivo cultivated limbal epithelial stem cell on amniotic membrane in limbal
stem cell deficiency, Cornea 34 (2015) 592–600.
[16] H. Liu, Z. Zhou, H. Lin, J. Wu, B. Ginn, J.S. Choi, H. Mao, Synthetic nanofiber-re-
inforced amniotic membrane via interfacial bonding, ACS Appl. Mater. Inter-
faces 10 (17) (2018) 14559–14569.
[17] S.H. Lim, X.Y. Liu, H. Song, K.J. Yarema, H. Mao, The effect of nanofiber-guided
cell alignment on the preferential differentiation of neural stem cells, Bioma-
terials 31 (34) (2010) 9031–9039.
[18] P.M. Crapo, T.W. Gilbert, S.F. Badylak, An overview of tissue and whole organ
decellularization processes, Biomaterials 32 (2011) 3233–3243.
[19] D.Y. Chau, S.V. Brown, M.L. Mather, V. Hutter, N.L. Tint, H.S. Dua, F.R. Rose,
A.M. Ghaemmaghami, Tissue transglutaminase (TG-2) modified amniotic
membrane: a novel scaffold for biomedical applications, Biomed. Mater. 7
(2012) 045011.
[20] M.S. Conners, F. Urbano, C. Vafeas, R. Stoltz, M.W. Dunn, M.L. Schwartzman,
Alkali burn-induced synthesis of inflammatory eicosanoids in rabbit corneal
epithelium, Invest. Ophthalmol. Vis. Sci. 38 (10) (1997) 1963–1971.
[21] R. Arora, D. Mehta, V. Jain, Amniotic membrane transplantation in acute chem-
ical burns, Eye 19 (3) (2004) 273–278.
[22] A. Miri, T. Alomar, M. Nubile, M. Alaqaba, M. Lanzini, U. Fares, H.S. Dua, In
vivo confocal microscopic findings in patients with limbal stem cell deficiency,
Br. J. Ophthalmol. 96 (4) (2012) 523–529.
[23] F.L. Gimeno, V. Lavigne, S. Gatto, J.O. Croxatto, L. Correa, J.E. Gallo, Advances in
corneal stem-cell transplantation in rabbits with severe ocular alkali burns, J.
Cataract Refract. Surg. 33 (11) (2007) 1958–1965.
[24] N. Zakaria, T. Possemiers, S.N. Dhubhghaill, I. Leysen, J.J. Rozema, C. Koppen,
M. Tassignon, Results of a phase I/II clinical trial: standardized, non-xenogenic,
cultivated limbal stem cell transplantation, J. Transl. Med. 12 (1) (2014) 58 58.
[25] A. Trounson, C. Mcdonald, Stem cell therapies in clinical trials: progress and
challenges, Cell Stem Cell 17 (1) (2015) 11–22.
[26] S. Ti, M. Grueterich, E.M. Espana, A. Touhami, D.F. Anderson, S.C. Tseng, Corre-
lation of long term phenotypic and clinical outcomes following limbal epithe-
lial transplantation cultivated on amniotic membrane in rabbits, Br. J. Ophthal-
mol. 88 (3) (2004) 422–427.
[27] T. Miyai, R. Hara, R. Nejima, K. Miyata, T. Yonemura, S. Amano, Limbal allo-
graft, amniotic membrane transplantation, and intraoperative mitomycin c for
recurrent pterygium, Ophthalmology 112 (7) (2005) 1263–1267.
[28] J. Liu, H. Sheha, Y. Fu, L. Liang, S.C. Tseng, Update on amniotic membrane trans-
plantation, Expert Rev. Ophthalmol. 5 (5) (2010) 645–661.
[29] J. Lai, Y. Li, Functional assessment of cross-linked porous gelatin hydrogels for
bioengineered cell sheet carriers, Biomacromolecules 11 (5) (2010) 1387–1397.
[30] J.Y. Lai, P.R. Wang, L.J. Luo, S.T. Chen, Stabilization of collagen nanofibers
with l-lysine improves the ability of carbodiimide cross-linked amniotic mem-
branes to preserve limbal epithelial progenitor cells, Int. J. Nanomed. 9 (2014)
5117–5130.
[31] Y.K. Choi, F.U. Din, D.W. Kim, Y.I. Kim, J.O. Kim, S.K. Ku, J.C. Ra, J.W. Huh, J.I. Lee,
D.H. Sohn, C.S. Yong, H.G. Choi, Amniotic membrane extract-loaded double-lay-
ered wound dressing: evaluation of gel properties and wound healing, Drug
Dev. Ind. Pharm. 40 (2014) 852–859.
[32] C. Malhotra, A.K. Jain, Human amniotic membrane transplantation: different
modalities of its use in ophthalmology, World J. Transpl. 4 (2) (2014) 111–121.
[33] D. Bauer, M. Hennig, S. Wasmuth, H. Baehler, M. Busch, K.P. Steuhl, S. Thanos,
A. Heiligenhaus, Amniotic membrane induces peroxisome proliferator-acti-
vated receptor-gamma positive alternatively activated macrophages, Invest.
Ophthalmol. Vis. Sci. 53 (2) (2012) 799–810.
[34] D.F. Anderson, P. Ellies, R.T. Pires, S.C. Tseng, Amniotic membrane transplan-
tation for partial limbal stem cell deficiency, Br. J. Ophthalmol. 85 (5) (2001)
567–575.
[35] M. Ebrahimi, E. Taghi-Abadi, H. Baharvand, Limbal stem cells in review, J. Oph-
thalmic Vision Res. 4 (1) (2009) 40–58.
[36] A. Joseph, A. Powellrichards, V.A. Shanmuganathan, H.S. Dua, Epithelial cell
characteristics of cultured human limbal explants, Br. J. Ophthalmol. 88 (3)
(2004) 393–398.
[37] M.A. Meallet, E.M. Espana, M. Grueterich, S.E. Ti, E. Goto, S.C. Tseng, Amniotic
membrane transplantation with conjunctival limbal autograft for total limbal
stem cell deficiency, Ophthalmology 110 (8) (2003) 1585–1592.
[38] S. Basu, S.P. Sureka, S.S. Shanbhag, A.R. Kethiri, V. Singh, V.S. Sangwan, Sim-
ple limbal epithelial transplantation: long-term clinical outcomes in 125 cases
of unilateral chronic ocular surface burns, Ophthalmology 123 (5) (2016)
10 0 0–1010.