Prevention of in vivo lung tumor growth by prolonged local delivery of hydroxycamptothecin using poly(ester-carbonate)-collagen composites

Abstract

Local tumor recurrence has a major impact on long-term patient survival following the surgical treatment of most cancers, and this is especially true with lung cancer. Consequently, methods to deliver chemotherapeutics locally at a lung tumor resection margin would be beneficial since: 1) systemic treatment approaches are ineffective or highly toxic; 2) the incidence of local recurrence does not warrant universal treatment of all patients with a highly morbid systemic therapy; and 3) surgical resection of recurrent disease is not an option and alternative rescue therapies are generally unsuccessful. To begin to meet this clinical need, we have prepared poly(glycerol monostearate-co-epsilon-caprolactone) films as a controlled, prolonged, and low dose delivery matrix for the potent anticancer agent 10-hydroxycamptothecin (HCPT). These drug-loaded films were applied to a collagen-based scaffold clinically indicated for the mechanical buttressing of lung tissue following surgical resection, resulting in a flexible composite that can be secured to the tissue that releases HCPT over seven weeks and thereby prevents the local growth and establishment of Lewis lung carcinoma tumors in vivo (a freedom of local tumor growth of 86%). In comparison, all animals treated with a larger intravenous dose of HCPT or unloaded composites developed rapid local tumors.

Full text

Prevention of in vivo lung tumor growth by prolonged local delivery of
hydroxycamptothecin using poly(ester-carbonate)-collagen composites
Jesse B. Wolinsky
a,b
, Rong Liu
c
, Joe Walpole
c
, Lucian R. Chirieac
d
, Yolonda L. Colson
c
, Mark W. Grinstaff
a,b,
⁎
a
Department of Biomedical Engineering, Boston University, Boston, MA 02215, United States
b
Department of Chemistry, Boston University, Boston, MA 02215, United States
c
Division of Thoracic Surgery, Department of Surgery, Brigham and Women's Hospital, Boston, MA 02115, United States
d
Department of Pathology, Brigham and Women's Hospital, Boston, MA 02115, United States
abstractarticle info
Article history:
Received 7 January 2010
Accepted 15 February 2010
Available online 22 February 2010
Keywords:
Prevention
Lung tumors
Recurrence
Films
Local delivery
Local tumorrecurrence has a major impacton long-term patient survival followingthe surgical treatmentof most
cancers, and this is especiallytrue with lung cancer. Consequently, methods to deliver chemotherapeutics locally
at a lung tumor resection margin would be beneficial since: 1) systemic treatment approaches are ineffective or
highly toxic;2) the incidence of localrecurrence does not warrant universaltreatment of all patients with a highly
morbid systemic therapy; and 3) surgical resection of recurrent disease is not an option and alternative rescue
therapies are generally unsuccessful. To begin to meet this clinical need, we have prepared poly(glycerol
monostearate-co-ε-caprolactone) films as a controlled, prolonged, and low dose delivery matrix for the potent
anticancer agent 10-hydroxycamptothecin (HCPT). These drug-loaded films were applied to a collagen-based
scaffold clinically indicated for the mechanical buttressing of lung tissue following surgicalresection, resulting in
aflexible composite that can be secured to the tissue that releases HCPT over seven weeks and thereby prevents
the local growth and establishment of Lewis lung carcinoma tumors in vivo (a freedom of local tumor growth of
86%). In comparison, all animals treated with a larger intravenous dose of HCPT or unloaded composites
developed rapid local tumors.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Polymers have enabled increasingly sophisticated therapies for the
diagnoses and treatment of cancer over the last two decades. Drugs,
contrast agents, and targeting moieties have been covalently-bound or
entrapped by a polymer in the form of prodrugs, micelles, particles, or
bulk material in an attempt to both increase solubility and localize
delivery to tumors via systemic targeting or local delivery [1–3].The
treatment of choice for localized solid tumors is surgery, where the
feasibility of removing all of the cancerous tissue is balanced against the
resultant morbidity to the patient. Local recurrence is of significant
concern following primary treatment for many cancers including head
and neck [4],breast[5], lung [6],colon[7],rectal[8], and pancreatic [9]
malignancies. Locally recurrent tumors are initiated by residual cancer
cells remaining at or near the resection margins or site of initial
treatment. Given that microscopic disease can remain,surgical resection
is often used in conjunction with radiation and/or chemotherapy to
improve local control. While radiation and chemotherapy are common-
ly utilized as adjuvant therapies for more advanced primary cancers,
preventative therapy is administered only in selected cases to decrease
recurrence risk, due to theoften severe side effectsassociated with these
treatments and the inability to accurately predict in which patients the
benefits of treatment would warrant the additional morbidity.
Consequently, a drug delivery platform that locally delivers thera-
peutic doses of drug directly to the site at highest risk for recurrent
disease, prevents recurrence at resection margins, and diminishes
significantly the systemic toxicity associated with intravenous chemo-
therapy and external radiation, would offer significant advantages over
other approaches. Loco-regional delivery is particularly beneficial in
situations where: 1) therapeutic levels of chemotherapy are not
achievable due to poor aqueous solubility, non-ideal pharmacokinetics
or biodistribution; 2) systemic treatment approaches are ineffective or
highly toxic; 3) the incidence of local recurrence does not warrant
universal treatment of all patients with a highly morbid systemic
therapy; or 4) surgical resectionof recurrent disease is not an option and
alternative rescue therapies are generally unsuccessful. All of these
examples are characteristic of the clinical practice of lung cancer surgery.
Our interest is to utilize a tunable drug-eluting polymeric delivery
system to prevent recurrence of malignant disease in lung cancer
patients, as limited pulmonary reserve makes this patient population
particularly susceptible to local tumor recurrence. To this aim, we
have developed a tunable polymeric drug delivery platform. Herein
we describe poly(glycerol monostearate-co-ε-caprolactone) films
loaded with the potent anticancer agent 10-hydroxycamptothecin
Journal of Controlled Release 144 (2010) 280–287
⁎Corresponding author. Department of Biomedical Engineering, Boston University,
Boston, MA 02215, United States.
E-mail address: mgrin@bu.edu (M.W. Grinstaff).
0168-3659/$ –see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconrel.2010.02.022
Contents lists available at ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel

(HCPT), an analogue of irinotecan which has been clinically utilized in
the treatment of various lung cancer subtypes [10,11]. These films can
be applied to a collagen-based scaffold clinically indicated for the
mechanical buttressing of lung tissue following surgical resection,
resulting in a flexible composite that can be secured to the tissue along
the resection margins. We demonstrate that these films are capable of
providing effective sustained release of HCPT over seven weeks and
thereby prevent the local growth and establishment of malignant cells
in an in vivo model of microscopic disease.
2. Materials and methods
2.1. Materials
All solvents were dried and freshly distilled before use or were
purchased from Sigma (Toluene). CH
2
Cl
2
was distilled under N
2
from
calcium hydride. Stannous 2-ethylhexanoate, ε-caprolactone, stearic
acid, N,N′-dicyclohexylcarbodiimide, and 4-(dimethylamino)pyridine
were purchased from Aldrich. Palladium on carbon was purchased
from Acros. 10-hydroxycamptothecin was purchased from Sigma.
5-benzyloxy-1,3-dioxan-2-one was prepared as previously reported
[12]. All reactions were performed under nitrogen atmosphere unless
otherwise noted. NMR spectra were recorded on a Varian INOVA spec-
trometer (
1
H at 400 MHz). Chemical shifts were referenced to residual
solvent peaks (CHCl
3
peak at 7.24 ppm). DCM= dichloromethane,
DCC = N,N′-dicyc lohexylcarbodiimid e, DM AP =4 -(dimethylamino)
pyridine, HCPT= 10-hydroxycamptothecin, PGC-OH= poly(glycerol-
co-ε-caprolactone), PGC-C18 = poly(glycerol monostearate-co-ε-
caprolactone), Pd/C= 10% palladium on activated carbon, and PBS=
phosphate buffered solution.
2.2. Synthesis of poly(glycerol-co-ε-caprolactone)
Poly(glycerol-co-ε-caprolactone), or PGC-OH, was prepared as
reported previously [12].Briefly, the ε-caprolactone and 5-benzyloxy-
1,3-dioxan-2-one monomers were mixed at a 4:1 molar ratio in a
schlenk flask and subsequently evacuated and flushed with N
2
three
times. Sn(oct)
2
was used (M/I=500) to catalyze the ring-opening
polymerization of the co-monomers at 140 °C for 48 h, and the resulting
copolymer was isolated by precipitation in cold methanol (92% yield).
The benzyl-protecting groups were removed via palladium-catalyzed
hydrogenolysis (99% yield).
1
H NMR (CDCl
3
)1.30–1.40 (m, 2H, CH
2
),
1.57–1.68 (m, 4H, CH
2
CH
2
), 2.25–2.28 (m, 2H, CH
2
), 3.82 (broad p, 1H,
OCH), 4.03–4.06 (t, 2H, OCH
2
), and 4.10–4.21 (m, 4H, OCH
2
CH).
2.3. Synthesis of poly(glycerol stearic acid-co-ε-caprolactone)
Stearic acid (0.78 mmol), poly(glycerol-co-ε-caprolactone) (2.6 mmol),
dicyclohexylcarbodiimide (0.63 mmol), and 4-(dimethylamino)pyridine
(0.26 mmol) were dissolved in CH
2
Cl
2
. The solution was stirred at
room temperature for 18 h. The DCU was filtered and the solvent
evaporated. The product, PGC-C18, was dissolved in CH
2
Cl
2
and
precipitated in cold methanol. The solvent was decanted and subse-
quently dried by evaporation (83% yield). Addition of the stearic acid
side chain was determined by the presence of the methylene group at
the end of the alkyl chain, with a corresponding peak in the
1
HNMR
spectrum at 0.82 (m, 3H, CH
3
).
2.4. Preparation of films on glass
Films (5 mg polymer; 100 μg HCPT) were cast from polymer
solutions onto glass cover-slips using a microsyringe. First, 10-
hydroxycamptothecin (HCPT) was added to CH
2
Cl
2
(0.1% w/v) in a
glass vial and homogenized for 60 min in a sonication bath to break
apart aggregates. Polymer (5% w/v) was then dissolvedinto the solution
by vortexing for 1 min. The solution was slowly added to the glass
surface and left to evaporate for 24 h, followed by further drying under
reduced pressure for another 24 h.
2.5. Preparation of films on scaffolds
Polymer films (3 mg polymer; 60 μg HCPT) were adhered to
collagen-based scaffolds for in vivo implantation studies. Peri-Strips
Dry® with Veritas® Collagen Matrix (Synovis LT) collagen strips were
chosen as a viable scaffolding material based on their wide use as
surgical staple-line reinforcement materials and demonstrated bio-
compatibility in human patients. HCPT-loaded films were adhered to
the collagen scaffolds using a modified method as was used for the
glass cover-slips. Following the application and complete drying of a
base layer of unloaded polymer, an additional layer of HCPT-loaded
polymer was added, evaporated, and dried under reduced pressure.
2.6. Size exclusion chromatography
Molecularweight determinations were performed via size exclusion
chromatography using THF as the eluent on a Polymer Laboratories
PLgel 3 µm MIXED-E column (3 µm bead size) and a Rainin HPLC
system (temp= 25 °C; flow rate=1.0 mL/min). Polystyrene standards
(Polysciences, Inc.) were used for calibration. These data confirm that
the post-modification of PGC-OH with stearic acid did not result in
polymer degradation (PGC-OH: M
n
=8600, M
w
/M
n
=1.96; PGC-C18:
M
n
=12,300,M
w
/M
n
=1.85).
2.7. Scanning electron microscopy
Samples for scanning electron microscopy were prepared by
mounting the films on an aluminum sample stub and then sputter-
coating with a 5 nm layer of gold–palladium alloy. Samples were then
imaged on a Zeiss SUPRA 40VP field emission scanning electron
microscope using an accelerating voltage of 2 kV. Cross-sectional images
were obtained from films that were fractured after submersion in liquid
nitrogen and subsequently clamped perpendicular to the base.
2.8. HCPT release kinetics from films
The release kinetics from HCPT-loaded films (n=4) were assessed
in PBS buffer at 37 °C. Specifically, films comprised of 5 mg polymer
(PGC-OH or PGC-C18) and 100 µg HCPT were cast onto glass cover-slips
and submerged in 50 mL of PBS. At specific time points, an aliquot of
release media was removed and the concentration of HCPT was
measured by fluorescence spectroscopy (λ
Ex
=382 nm and
λ
Em
=550 nm, Photon Technology International QM-4/2005 spectro-
fluorimeter). The release medium was changed at regular intervals to
maintain the pH and ensure sink conditions. A calibration curve was
constructed from nine samples of known HCPT concentrations with
linearity observed from 10 nM to 5 µM (R
2
=0.999) from which the
drug concentration of each aliquot was determined. After the
completion of release, the HCPT remaining in each film was quantified.
Films were dissolved in CH
2
Cl
2
to release encapsulated drug, the solvent
evaporated, and PBS (150 mL) was added under rigorous stirring. The
fluorescence spectrum was recorded and remaining drug measured.
2.9. Stability of HCPT
The ratio of active lactone to inactive carboxylate HCPT was
quantified by collecting aliquots of release media before significant
conversion to the carboxylate form could occur. Samples were analyzed
by HPLC using an adapted method reported previously [13]. AnHP 1090
HPLC system was used with a fluorescence detector (λ
ex
=380 nm,
λ
em
=540 nm), with a Phenomenex Prodigy 5 ODS reverse-phase
column (150× 4.6 mm, 5 μm). The mobile phase was composed of 25%
acetonitrile and 75% ammonium acetate buffer adjusted to a pH of 6.4
281J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

and delivered at 0.8 mL/min. Calibration curves were constructed for
both lactone (rt=3.7 min) and carboxylate (rt= 2.8 min) forms with
sensitivities of 0.5 ng/mL and 2 ng/mL respectively. At designated time
points, drug-loaded films were transferred to fresh PBS (pH=6.4,
adjusted with acetic acid) for 60 min at 37 °C and aliquots were
removed andimmediately analyzed. HCPT converted from 100% lactone
form to 87% lactoneat pH = 6.4 over60 min for the control experiments.
The stability study was performed until the amount of drug released
became lower than our limit of detection. Long-term release kinetics
were also monitored for 58 days and confirmed the release profile
observed from the fluorescence experiment. Cumulative release was
measured by removing aliquots from the release media and replaced
with fresh media at regular intervals to maintain sink conditions.
2.10. In vitro cytotoxicity assays
In vitro experiments were performed with Lewis lung carcinoma
(LLC) cells received as a kind gift from Dr. Judah Folkman's laboratory
at Children's Hospital, Boston and cultured with medium containing
10% FBS and 1% PS. Cells were plated at a density of 30,000 cells/well
in 24-well Transwell plates (Corning Inc.) and grown overnight. Films
were sterilized overnight under UV light and placed on tissue culture
treated polyester membrane inserts (0.4 μm pore size) before being
incubated with cultured cells and adding serum-containing culture
media. Films were co-incubated with the cells for 5-day intervals
before being transferred to new wells with freshly-plated cells to
assess continued drug release and cytotoxicity in the setting of new
media and freshly-plated tumor cells. To avoid reaching the HCPT
solubility limit (and thereby to decrease the rate and duration of drug
release), shorter intervals were used over the first 5 days to simulate
the anticipated drug gradient present in living tissues. Cell viability
was quantified using a standard MTT cell proliferation assay.
2.11. Animal and tumor models
All animal studies were approved by the Institutional Animal Care
and Use Committee of Dana-Farber Cancer Institute, and all animal
care was performed humanely and in accordance with the designated
and approved standards. Female C57BL/6J mice at six to eight weeks
of age were obtained from Jackson Laboratories (Bar Harbor, Maine).
Mice were anesthetized with intraperitoneal injection of ketamine
(120 mg/kg) and xylazine (8 mg/kg). A 0.8 cm wide subcutaneous
pocket was prepared by blunt dissection under sterile conditions
between the shoulders on the dorsum of the mice. Ultraviolet-
sterilized polymer films dried on pericardial strips (1.0 × 0.8 cm) with
or without HCPT (60 µg per film) were inserted into the subcutaneous
pockets. The incision was closed with 5–0 polypropylene sutures.
Mice receiving sham surgery without film implantation served as
control for wound healing studies (n= 3). Mice receiving an
intravenous injection (i.v.) of 200 µg HCPT served as a second control
group for administration of HCPT (dissolved in DMSO and diluted
with PBS (1:5 v/v)). The HCPT solution consisted of 100% lactone form
as measured by HPLC and was injected immediately upon formula-
tion. The body weights of mice were measured with a digital balance
biweekly. White blood cell counts from tail vein blood were
performed at day 10 after surgery.
Wound healing, including the integrity of skin and any signs of
wound infection, was assessed daily without knowledge of the type of
the films implanted. Implanted films and the surrounding tissues
were sampled for histological evaluation 10 days after surgery and
fixed in 10% formalin. Formalin-fixed, paraffin-embedded pathology
specimens from each mouse were cut into 5-μm thick serial sections
and hemotoxylin and eosin (H&E) staining was performed. Micro-
scopic examination of the H&E sections was assessed independently
by a pathologist (L. Chirieac), without knowledge of the type of
implant or drug delivery regimen. Each specimen was evaluated for
tissue integrity, stage of infection, the magnitude and type of local
inflammatory changes, reactive fibrosis, scar and granulation tissue
formation.
In the tumor establishment study, incisions were allowed to heal
for two days before LLC cell injection to avoid leakage of the cells. LLC
cells (750,000), suspended in 0.1 mL of PBS, were injected subcuta-
neously on top of the implanted film via a 27-gauge needle. Tumor
growth and its local relationship to films were monitored by palpation
and a digital caliper biweekly. Mice were euthanized if tumors
reached 2 cm in size. Rapid tumor growth in control mice and
morbidity resulting from distal metastases required sacrifice of
multiple animals 2 weeks following LLC-injection, and thus equiva-
lent comparison of tumor growth was performed within 14 days of
tumor inoculation for all mice in the study.
2.12. Statistical analyses
Analysis of variation (ANOVA) was used to compare body weights
and percent change of body weight between experimental and
control animals; the post-hoc Tucky's multiple comparison test was
performed if the overall difference was significant (Pb0.05). The WBC
counts were compared by Student's t-test. The incidence of tumor
growth overtopping films in mice that received either HCPT-loaded
or unloaded films was compared with Fisher's exact test. All sta-
tistics were performed by GraphPad Prism 4.0 software (La Jolla, CA).
Pb0.05 is considered statistically significant. All data are presented as
mean±s.e.m.
3. Results
3.1. Film formation, characterization, and drug release studies
HCPT-loaded poly(glycerol-co-ε-caprolactone) (PGC-OH) and
poly(glycerol stearic acid-co-ε-caprolactone) (PGC-C18) films were
prepared and evaluated. PGC-OH was synthesized as reported
previously[12]. PGC-C18 was synthesized by conjugating stearic acid
to PCG-OH to yield a significantly more hydrophobic copolymer
(contact angles of PGC-OH and PGC-C18 films were 87° and 125°,
respectively). The drug was pre-mixed in the solvent and homoge-
nized in an ultra-sonication bath to yield a colloidal solution of drug
particulates, followed by the addition of polymer and film casting. In
this procedure, small insolubilized drug aggregates were embedded
into the polymer matrix as the volatile solvent evaporated. The PGC-
OH films fractured over the first few days when exposed to PBS at
37 °C. Conversely, the PGC-C18 formed uniform films with improved
mechanical integrity. The scanning electron micrograph of a repre-
sentative PGC-C18 film is shown in Fig. 1. The film has a uniform
thickness (∼40 µm) with smooth surfaces and a homogenous non-
porous microstructure is evident via a cross-sectional view.
Drug-loaded PGC-C18 films demonstrated gradual and sustained
release kinetics with no significant initial burst of HCPT as compared to
unmodifiedPGC-OH (Fig. 2) or other well known polymers such as PLGA
[14]. Specifically, PGC-OH released 25% of its load over two days and
continued to elute drug for approximately three weeks. In contrast,
HCPT-loaded PGC-C18 films released in a much more controlled
manner, initially releasing about 2% loading per day for the first two
weeks, decreasing to approximately 1% per day for the following three
weeks, before diminishing to 0.5% release per day, with continued
release detected for at least 7 weeks. The films were dissolved following
release and 100% mass balance was observed for HCPT in both groups.
The films were also cast onto pericardial substrates for mechanical
support; pericardial scaffolds are currently used as reinforcement
materials for several types of surgeries including bariatric and
cardiothoracic procedures. The release kinetics of loaded PGC-C18
films from pericardial substrates were similar to films cast on glass
supports (Fig. 3). The lactone–carboxylate conversion of HCPT was
282 J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

investigated for loaded PGC-C18 films. The drug was loaded into the
polymer in its 100% lactone (active) form and at regular intervals, the
films were transferred to fresh release media (pH =6.4) for 60 min, the
media collected for stability analysis, and the films transferred to fresh
media forlong-term release.In this manner,the lactone stabilityof HCPT
remaining in the films over time can be determined. The conversion
from 100% lactone to carboxylate in these conditions over 60 min
yielded approximately 87% lactone form in the control studies. The
conversion to the carboxylate form was greatly reduced usingour films,
stabilizing to about 84% lactone form over 49 days of release (Fig. 4). For
comparison, the t
1/2
of HCPT in PBS at 37 °C is reported as 22 min with
∼80% conversion of the lactone to the carboxylate occurring at
equilibrium [15].
3.2. Cytotoxicity studies
To determine the anti-proliferative efficacy and duration of
effective drug release for this polymeric film on a lung cancer cell
line, HCPT-loaded PGC-C18 (2% drug w/w) films were exposed to
cultured Lewis Lung Carcinoma (LLC) cells for 5-day intervals up to
50 days. Based on the release kinetic experiments described above, a
minimum HCPT concentration of 1 µM accumulates at the end of each
5-day interval, well-above the measured HCPT IC
50
of LLC cells
(b100 ng/mL). The films exhibited significant cytotoxicity for each
5-day interval exposure during the 50 day experiment (Fig. 5),
demonstrating that the films are capable of releasing a sustained
dose of therapeutic drug for at least 7 weeks. In contrast, unloaded
PGC-C18 films were not cytotoxic for the duration of the experiment.
3.3. Evaluation of composites in a lung cancer animal model
Based on the positive results from the in vitro study, a similar
HCPT-loading was used for these in vivo tests. Poly(glycerol stearic
acid-co-ε-caprolactone) was selected over PGC-OH as a delivery
matrix due to its preferable film stability and release properties.
HCPT-loaded (2% wt/wt; 60 μg of HCPT) PGC-C18 films adhered to
pericardial substrates were implanted subcutaneously into C57BL/6
mice. Two days were allowed for the healing of the incision before
Lewis lung carcinoma cells were injected on top of the films to
approximate microscopic malignant disease. Local tumor growth was
defined as the development of subcutaneous tumor nodules overtop
the implanted films. The experimental mice that received the polymer
films loaded with HCPT (∼0.4 μg/mm
2
) exhibited a freedom of local
tumor growth of 86% (6/7 mice). In contrast, all but one mouse from
the animal group treated with unloaded polymer films developed
significant tumor growth within 14 days, for a freedom of local tumor
Fig. 1. Top. Chemical structures of poly(glycerol-co-ε-caprolactone) (PGC-OH) and poly(glycerol monostearate-co-ε-caprolactone) (PGC-C18). Bottom. Scanning electron micrographs of
10-hydroxycamptothecin-loaded poly(glycerol monostearate-co-ε-caprolactone) films (left: cross-section; right:surface).
Fig. 2. Release profiles of 10-hydroxycamptothecin from poly(glycerol-co-ε-caprolactone)
films (n=3). PGC-OH: poly(glycerol-co-ε-caprolactone), PGC-C18: poly(glycerol mono-
stearate-co-ε-caprolactone).
283J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

growth of less than 12% (1/9 mice) (Fig. 6). Additionally, all of the
mice treated with a clinically-relevant intravenous dose of HCPT
(200 µg, approximately ∼10 mg/kg) developed rapid local tumor
growth. For several mice, tumor cells gained access to more distant
regional areas as a consequence of the tumor cell injection procedure
resulting in the delayed development of small regional tumors outside
the periphery of the HCPT-loaded polymer films, also highlighting the
local nature of HCPT release from the films.
Notably, this result indicates that this delivery strategy is local, and
does spare healthy surrounding tissues from high levels of drug,
thereby minimizing chemotherapy-induced toxicity. Consistent with
local delivery, there was no clinical evidence of impaired healing in
mice treated with unloaded or HCPT-loaded films. The minimal
inflammatory reaction and fibrosis was consistent with normal
wound healing and was indistinguishable between mice receiving
HCPT-loaded and unloaded films, and mice undergoing sham surgery
(Fig. 7). There was no evidence of persistent local inflammation,
infection, or subcutaneous fluid accumulation upon inspection in any
group. Histological analysis of tissues at the surgical site demonstrat-
ed only mild inflammatory changes, fibrosis and granulation tissue
formation consistent with the acute phase of normal, post-surgical
wound healing process. Mice receiving either unloaded or HCPT-
loaded films did not experience any significant decrease in body
weight over time. White blood cell counts were comparable between
mice implanted with unloaded or HCPT-loaded films (9.53 ±
1.26×10
6
/mL, 7.00±0.23 × 10
6
/mL, respectively; PN0.05; Fig. 8).
4. Discussion
Although the field of polymeric delivery devices for cancer
therapies is undergoing rapid growth, there has been a limited focus
on polymer implants for prevention of local tumor recurrence.
Our approach utilized a novel copolymer composed of glycerol and
ε-caprolactone biocompatible building blocks whose properties can
be tuned via functionalizable side chain group modification to yield a
delivery system with prolonged release kinetics that can be applied to
pericardial strips already approved for suture line reinforcement in
lung surgery [12]. The resulting composites are thin and flexible (see
Fig. 3), and easily administered along the surface or resection margins
of soft tissue via the application of a standard surgical stapler
commonly utilized in the resection of localized tumors. Specifically,
poly(glycerol-co-ε-caprolactone) (PGC-OH) was chosen as a delivery
matrix because the copolymer can be easily modified to affect a range
of properties (i.e., crystallinity, permeability, hydrophobicity, etc.)
known to alter drug release from bulk polymer [12,16]. A second
Fig. 3. Left. Release of HCPT-loaded poly(glycerol monostearate-co-ε-caprolactone) films off glass or pericardial substrates (n= 3). Right. Photograph of a flexible PGC-C18 film
adhered to pericardium while submerged in phosphate buffered solution after 1 day.
Fig. 4. Stability of the lactone form of 10-hydroxycamptothecin remaining in poly
(glycerol monostearate-co-ε-caprolactone) over time (n= 4). Media collected after
60 min of release at pH = 6.4, dashed line represents conversion of control HCPT in
same conditions.
Fig. 5. Proliferation of LLC cells after exposure to HCPT-loaded poly(glycerol
monostearate-co-ε-caprolactone) films (n= 3). Fresh cells were exposed to films for
multi-day intervals over the release life of the films. For example, after 45 days of
release in vitro, drug-loaded films were exposed to a new plate of fresh cells for five
days demonstrated significant cytotoxicity compared to unloaded films. Percent
proliferation is normalized to positive serum control.
284 J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

polymer comprised of poly(glycerol monostearate-co-ε-caprolactone)
(PGC-C18), a stearic acid-modified analogue of PGC-OH, was also
synthesized for these studies (Fig. 1). By doing so, we were able to
investigate two copolymer analogues with identical back-bone struc-
tures yet featuring disparate side chains, ultimately leading to
substantially different physical properties. Fatty acid side chains were
incorporated into this polymer in order to increase the hydrophobicity
of the polymer matrix and allow the comparison of drug release
kinetics from polymer analogues with markedly divergent hydropho-
bicities. An in-depth evaluation of a variety of other lipophilic PGC-OH
analogues in addition to PGC-C18 was performed, and the results of
these experiments will be reported elsewhere.
10-Hydroxycamptothecin (HCPT) was selected as a prototype
drug because it is a known chemotherapeutic agent that has
demonstrated strong anti-tumor activity but whose practical use
has been limited due to poor aqueous solubility and instability in
aqueous solution. Furthermore, like the FDA-approved camptothecin
Fig. 6. Freedom from tumor growth overtop films implanted subcutaneously is sig-
nificantly higher for HCPT-loaded films compared to unloaded (P= 0.0022), HCPT–i.v.
(P=0.0005) and sham surgery (P=0.0005) (symbols used for clarity). For the sham
surgery and i.v. control, growth occurred at the site of tumor cell injection. Some mice
were euthanized before day 14 due to excessive tumor mass. *Mice with loaded films
eventually develop tumor distant from the film, confirming that drug release is local.
HCPT content: loaded film (60 µg), i.v. injection (200 µg).
Fig. 7. Normal wound healing after polymer film implantation. A. Sham surgery (× 40); B. unloaded films (×40); C. loaded film (× 40); D. sham surgery (× 200); E. unloaded films
(×200); F. loaded film (× 200).
Fig. 8. White blood cell count of mice implanted with unloaded or HCPT-loaded films
ten days following treatment (PN0.05).
285J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

analogues irinotecan and topotecan, systemic administration of HCPT
is associated with significant morbidity [17]. Thus, the entrapment of
HCPT in the copolymer films was expected to circumvent the
solubility obstacle and reduce side effects by locally delivering the
anticancer agent directly to the site of microscopic malignant disease.
Additionally, the camptothecins have a pH-sensitive lactone ring that
opens to become inactive at a physiological pH (t
1/2
=22 min) [18].In
agreement with other polymer systems, the polymer-embedded drug
was mostly preserved in its active lactone form [18,19]. As PGC-18
does not degrade appreciably over at least six months, the
preservation of the lactone form of HCPT was likely due to its
protection from the aqueous environment inside the polymer matrix.
10-Hydroxycamptothecin has been formulated into polymer
nanoparticles [25–27], microparticles [28],andmicellarstruc-
tures [29,30] and demonstrated some inhibition of primary tumor
growth following systemic injection but efficacy against local tumor
recurrence has yet to be demonstrated with these controlled release
formulations. PEGylated niosomes were modified with transferrin for
receptor-mediated delivery of HCPT and shown to inhibit tumor
growth (79% inhibition) significantly more than non-transferrin
modified niosomes (46%) and intravenous HCPT injection (29%) in a
sarcoma murine model as reported by Hong et al. [31]. Similarly,
Zhang et al. described HCPT-loaded poly(caprolactone-co-lactide)-b-
PEG-b-poly(caprolactone-co-lactide) nanoparticles that demonstrat-
ed inhibition of tumor growth rate by 80% in a sarcoma murine model
compared to that of the systemic injection of HCPT (67%) [25].
Perhaps more importantly, polymeric delivery of camptothecin drugs
has been shown to stabilize camptothecin molecules in their active
lactone form while the drug is embedded in its polymer delivery
vehicle. Shenderova et al. reported that encapsulation of HCPT within
poly(lactide-co-glycolide) microspheres stabilizes the camptothecin
in its lactone form over at least two months of drug release in vitro,
attributed most likely to the acidic environment present within the
slowly-degrading microsphere [28]. Hatefiet al. observed a similar
result with biodegradable injectable ε-caprolactone oligomers over a
release period of 16 weeks. The minimal conversion to the carboxyl-
ate form was attributed to the impedance of water into the
hydrophobic oligomer, delaying polymer degradation and subsequent
creation of an acidic microenvironment [19].
For our application, lung cancer recurrence represents a significant
clinical challenge in patients who commonly have limited pulmonary
reserve. Surgery for lung cancer patients with poor lung function due
to emphysema or other underlying lung diseases is often limited to
non-anatomic wedge resections of the involved parenchyma rather
than the traditional lobectomy which involves removal of significantly
more lung tissue. Unfortunately, when compared to lobectomy,
smaller wedge resections are associated with a significantly higher
incidence of microscopically positive surgical resection margins and
double the local recurrence rate to ∼16% [20]. Given the limited
treatment options available for recurrent disease, and, the fact that
the majority of these patients will die as a result of their disease,
interventions to prevent recurrence are critically important to this
patient population [21,22]. Therefore, a lung cancer model was chosen
to establish proof of concept for the described HCPT-loaded PGC-C18
film delivery system in vivo.
Today, there are only a handful of delivery devices reported for the
prevention of local growth of early or residual disease and only one of
these technologies has been successfully translated towards clinical
use. Paclitaxel-loaded thermosensitive chitosan-based hydrogels,
developed by Leroux et al., were implanted four days after tumor
cell inoculation and demonstrated a significant decrease, but not
complete inhibition, of tumor growth. Matsuda et al. created a
polyurethane-based pouch which was sutured subcutaneously in
tumor-bearing mice and loaded with gemcitabine three days after
tumor inoculation [23]. Four of six mice supporting loaded devices
had no observable tumor mass during the 30 day observation period,
but the remaining two mice developed tumors at a rate comparable to
the control mice. Cross-linked chitosan hydrogels have also been
loaded with a radioisotope to deliver localized radiotherapy for
prevention of tumor recurrence in breast cancer [24]. Implants loaded
with
131
I-norcholesterol were co-implanted with 4T1 metastatic
mammary mouse tumor cells to simulate microscopic residual disease
and tumor growth was prevented in 69% of the mice. The most
established local delivery device to date is used in the treatment of
malignant glioma, an aggressive brain cancer that often recurs near
the resection margins of the primary tumor. Commercially manufac-
tured by MGI Pharma (now Eisai) under the brand name Gliadel®, the
device delivers the chemotherapeutic carmustine over several days
from a rigid biodegradable polyanhydride wafer placed near the
resection margins, that degrades within three weeks of implantation.
Patients receiving these wafers, which are placed in the resection
cavity after tumor debulking, exhibited an increase in mean survival
from 20 weeks to 28 weeks. Although the wafer had a modest impact
on the survival of treated patients, patients reported a markedly
higher quality of life compared to those treated by conventional
systemic chemotherapy [32–34]. These previous results inspire the
development of additional technologies directed towards improving
clinical efficacy and expanding local treatment to other types of
tissues.
An important design requirement for this application is the
sustained release of drug over several weeks where microscopic
disease is most likely to reside. Continuous administration of systemic
chemotherapy is precluded by dose-limiting toxicity, while post-
operative treatment is avoided due to concerns of wound healing
impairment and increased risk of infection. These complications are
circumvented given that prolonged local delivery requires a substan-
tially smaller amount of drug, thereby minimizing systemic complica-
tions. Moreover, recurrent solid tumors will develop at a positive
margin following resection with a much more rapid growth rate
compared to the initial primary tumor [35].As10–15% of cells in a
solid tumor are expected to be in the DNA synthesis phase, only those
cells would experience increased sensitivity to anti-neoplastic agents
at a given time point [36–38]. For these reasons, we expect that the
continuous exposure of a chemotherapeutic drug to tumor cells over
multiple cell cycles would be significantly more effective than a bolus
or short term administration of drug.
The controlled and sustained drug delivery from poly(glycerol
monostearate-co-ε-caprolactone) films described in this study offers
several unique advantages over other local delivery vehicles. First,
complete inhibition of local tumor establishment was achieved at the
surface of the HCPT-loaded implanted films in all but one animal (a
dramatic result when compared to prior studies) whereas a standard
intravenous dose of HCPT three times greater than the dose delivered
by a film did not affect tumor growth. Second, the copolymer
implemented here can be chemically modified for tunable drug
release kinetics, as exemplified by the unmodified (PGC-OH) and
lipophilic (PGC-C18) derivative films compared in this study. For
example, the burst release from hydrophilic PGC-OH is likely due to
the dissolution of HCPT deposited at the surface of the PGC-OH film
and this effect may be exacerbated as the film fractures, thus
increasing the surface area of the film and facilitating rapid release
over several days. The hydrophobic PGC-C18 demonstrated no such
burst. Third, highly localized delivery was observed, since local tumor
growth was prevented but the surrounding tissue was unaffected by
HCPT, displaying only mild local inflammation associated with normal
wound healing. This observation indicates that only treated surfaces
receive significant concentrations of chemotherapy, thus enabling
normal healing and preserving surrounding healthy tissue. Finally, the
HCPT-polymer platform is assembled via facile processing methods,
allowing for the preparation of flexible films with custom shapes and
surface areas. The results suggest that a polymeric film-based delivery
system can be locally administered to prevent tumor cell proliferation
286 J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287

and recurrence following limited surgical resections and other
primary therapies where surgical margins are more likely to exhibit
small foci of microscopic residual tumor. We are developing animal
models to assess the efficacy of these films to prevent recurrent tumor
growth in vivo following surgical resection. The encouraging results
obtained in the current study warrant further exploration of drug-
loaded film-based implants for the delivery of anticancer agents to
regions of potential microscopic disease with the ultimate goal of
preventing tumor recurrence following surgical resection.
Acknowledgements
This work was supported in part by a grant from the Wallace H.
Coulter Foundation (MWG & YLC) and the George H.A. Clowes, Jr., MD,
FACS Memorial Research Career Development Award (YLC) through
the American College of Surgeons. The authors wish to thank Boston
University and Brigham and Women's Hospital for their support, and
the Animal Resources Facility at Dana-Farber Cancer Institute for their
excellent animal care.
References
[1] W.M. Saltzman, L.K. Fung, Polymeric implants for cancer chemotherapy, Adv.
Drug Deliv. Rev. 26 (1997) 209–230.
[2] T.M. Allen, P.R. Cullis, Drug delivery systems: entering the mainstream, Science
303 (2004) 1818–1822.
[3] M.A. Moses, H. Brem, R. Langer, Advancing the field of drug delivery: taking aim at
cancer, Cancer Cell 4 (2003) 337–341.
[4] S.J. Wong, M. Machtay, Y. Li, Locally recurrent, previously irradiated head and neck
cancer: concurrent re-irradiation and chemotherapy, or chemotherapy alone?
J. Clin. Oncol. 24 (2006) 2653–2658.
[5] M. Clemons, S. Danson, T. Hamilton, P. Goss, Locoregionally recurrent breast
cancer: incidence, risk factors and survival, Cancer. Treat. Rev. 27 (2001) 67–82.
[6] A. El-Sherif, H.C. Fernando, R. Santos, B. Pettiford, J.D. Luketich, J.M. Close, R.J.
Landreneau, Margin and local recurrence after sublobar resection of non-small
cell lung cancer, Ann. Surg. Oncol. 14 (2007) 2400–2405.
[7] T.E.Read,M.G.Mutch,B.W.Chang,M.S.McNevin,J.W.Fleshman,E.H.Birnbaum,
R.D. Fry, P.F. Caushaj, I.J. Kodner, Locoregional recurrenceand survival after curative
resection of adenocarcinoma of the colon, J. Am. Coll. Surg. 195 (2002) 33–40.
[8] Y. Moriya, Treatment strategy for locally recurrent rectal cancer, Jpn. J. Clin. Oncol.
36 (2006) 127–131.
[9] E.A. Newman, D.M. Simeone, M.W. Mulholland, Adjuvant treatment strategies for
pancreatic cancer, J. Gastrointest. Surg. 10 (2006) 916–926.
[10] F. Oshita, H. Saito, K. Yamada, K. Noda, Phase II study of paclitaxel and irinotecan
chemotherapy in patients with advanced nonsmall cell lung cancer, Am. J. Clin.
Oncol. 30 (2007) 358–360.
[11] M. Kawahara, Irinotecan in the treatment of small cell lung cancer: a review of
patient safety considerations, Expert Opin. Drug Safety 5 (2006) 303–312.
[12] J.B. Wolinsky, W.C. Ray III, Y.L. Colson, M.W. Grinstaff, Poly(carbonate-ester)s based
on units of 6-hydroxyhexanoic acid and glycerol, Macromolecules 40 (2007)
7065–7068.
[13] D.F. Chollet, L. Goumaz, A. Renard, G. Montay, L. Vernillet, V. Arnera, D.J. Mazzo,
Simultaneous determination of the lactone and carb oxylate forms of the
camptothecin derivative CPT-11 and its metabolite SN-38 in plasma by high-
performance liquid chromatography, J. Chromatogr., B, Biomed. Sci. Appl. 718
(1998) 163–175.
[14] B. Ertl, P. Platzer, M. Wirth, F. Gabor, Poly(D,L-lactic-co-glycolic acid) microspheres
for sustaineddelivery and stabilizationof camptothecin,J. Control. Release 61 (1999)
305–317.
[15] T.G. Burke, Z. Mi, E thyl sub stitu tion at the 7 posit ion ext ends th e half-l ife of
10-hydroxycamptothecin in the presence of human serum albumin, J. Med.
Chem. 36 (1993) 2580–2582.
[16] C.G. Pitt, M.M. Gratzl, A.R. Jeffcoat, R. Zweidinger, A. Schindler, Sustained drug
delivery systems II: factors affecting release rates from poly(epsilon-caprolactone)
and related biodegradable polyesters, J. Pharm. Sci. 68 (1979) 1534–1538.
[17] Y.H. Ping, H.C. Lee, J.Y. Lee, P.H. Wu, L.K. Ho, C.W. Chi, M.F. Lu, J.J. Wang, Anticancer
effects of low-dose 10-hydroxycamptothecin in human colon cancer, Oncol. Rep.
15 (2006) 1273–1279.
[18] A.Shenderova, T.G.Burke,S.P.Schwendeman,Stabilizationof 10-hydroxycamptothecin
in poly(lactide-co-glycolide) microsphere delivery vehicles, Pharm. Res. 14 (1997)
1406–1414.
[19] A. Hatefi, D. Knight, B. Amsden, A biodegradable injectable thermoplastic for
localized camptothecin delivery, J. Pharm. Sci. 93 (2004) 1195–1204.
[20] R.J. Landreneau, D.J. Sugarbaker, M.J. Mack, S.R. Hazelrigg, J.D. Luketich, L.
Fetterman, M.J. Liptay, S. Bartley, T.M. Boley, R.J. Keenan, P.F. Ferson, R.J. Weyant,
K.S. Naunheim, Wedge resection versus lobectomy for stage I (T1 N0 M0) non-
small-cell lung cancer, J. Thorac. Cardiovasc. Surg. 113 (1997) 691–698 discussion
698–700.
[21] F. Detterbeck, M.P. Rivera, M.A. Socinski, J. Rosenman, Diagnosis and Treatment of
Lung Cancer: An Evidence-Based Guide for the Pr acticing Physician, W. B.
Saunders Company, Philadelphia, 2001.
[22] N. Martini, M.S. Bains, M.E. Burt, M.F. Zakowski, P. McCormack, V.W. Rusch, R.J.
Ginsberg, Incidence of local recurrence and second primary tumors in resected
stage I lung cancer, J. Thorac. Cardiovasc. Surg. 109 (1995) 120–129.
[23] T. Manabe, H. Okino, R. Maeyama, K. Mizumoto, M. Tanaka, T. Matsuda, New
infusion device for trans-tissue, sustained local delivery of anticancer agent to
surgically resected tissue: potential use for suppression of local recurrence of
pancreatic cancer, J. Biomed. Mater. Res. B. Appl. Biomater. 73 (2005) 203–207.
[24] A.K. Azab, J. Kleinstern, V. Doviner, B. Orkin, M. Srebnik, A. Nissan, A. Rubinstein,
Prevention of tumor recurrence and distant metastasis formation in a breast
cancer mouse model by biodegradable implant of 131I-norcholesterol, J. Control.
Release 123 (2007) 116–122.
[25] L. Zhang, Y. Hu, X. Jiang, C. Yang, W. Lu, Y.H. Yang, Camptothecin derivative-loaded
poly(caprolactone-co-lactide)-b-PEG-b-poly(caprolactone-co-lactide) nanoparticles
and their biodistribution in mice, J. Control. Release 96 (2004) 135–148.
[26] L. Zhang, M. Yang, Q. Wang, Y. Li, R. Guo, X. Jiang, C. Yang, B. Liu, 10-
Hydroxycamptothecin loaded nanoparticles: preparation and antitumor activity
in mice, J. Control. Release 119 (2007) 153–162.
[27] J. Wang, R. Wang, L.B. Li, Preparation and properties of hydroxycamptothecin-
loaded nanoparticles made of amphiphilic copolymer and normal polymer,
J. Colloid Interface Sci. 336 (2009) 808–813.
[28] A. Shenderova, T.G. Burke, S.P. Schwendeman, The acidic microclimate in poly
(lactide-co-glycolide) microspheres stabilizes camptothecins,Pharm. Res. 16 (1999)
241–248.
[29] C. Zhang, Y. Ding, L.L. Yu, Q. Ping, Polymeric micelle systems of hydroxycamp-
tothecin based on amphiphilic N-alkyl-N-trimethyl chitosan derivatives, Colloids
Surf., B Biointerfaces 55 (2007) 192–199.
[30] X. Yang, L. Li, Y. Wang, Y. Tan, Preparation, pharmacokinetics and tissue
distribution of micelles made of reverse thermo-responsive polymers, Int. J.
Pharm. 370 (2009) 210–215.
[31]M.Hong,S.Zhu,Y.Jiang,G.Tang,Y.Pei,Efficient tumor targeting of
hydroxycamptothecin loaded PEGylated niosomes modified with transferrin,
J. Control. Release 133 (2009) 96–102.
[32] H. Brem, M.S. Mahaley Jr., N.A. Vick, K.L. Black, S.C. Schold Jr., P.C. Burger, A.H.
Friedman,I.S. Ciric,T.W. Eller, J.W.Cozzens, etal., Interstitialchemotherapywith drug
polymer implants for the treatment of recurrent gliomas, J. Neurosurg. 74 (1991)
441–446.
[33] H. Brem, S. Piantadosi, P.C. Burger, M. Walker, R. Selker,N.A. Vick, K. Black, M.Sisti, S.
Brem, G. Mohr, et al., Placebo-controlled trial of safety and efficacy of intraoperative
controlled delivery by biodegradable polymers of chemotherapy for recurrent
gliomas.The Polymer-brainTumor TreatmentGroup, Lancet345 (1995) 1008–1012.
[34] H. Brem, Polymers to treat brain tumours, Biomaterials 11 (1990) 699–701.
[35] S.E. Shackney, G.W. McCormack, G.J. Cuchural Jr., Growth rate patterns of solid
tumors and their relation to responsiveness to therapy: an analytical review, Ann.
Intern. Med. 89 (1978) 107–121.
[36] V.T. DeVita, T.S. Lawrence, S.A. Rosenberg, DeVita, Hellman, and Rosenberg's
Cancer: Principles and Practice of Oncology, Lippincott Williams & Wilkins, 2008.
[37] W.H. Wolberg, F.J. Ansfield, The relation of thymidine labeling index in human
tumors in vitro to the effectiveness of 5-fluorouracil chemotherapy, Cancer Res.
31 (1971) 448–450.
[38] J.S. Meyer, Cell kinetic measurements of human tumors, Human Pathol. 13 (1982)
874–877.
287J.B. Wolinsky et al. / Journal of Controlled Release 144 (2010) 280–287