Content uploaded by Sebastian Kalamajski
Author content
All content in this area was uploaded by Sebastian Kalamajski on Nov 19, 2015
Content may be subject to copyright.
The FASEB Journal •Research Communication
Vaccination against the extra domain-B of fibronectin
as a novel tumor therapy
Elisabeth J. M. Huijbers,* Maria Ringvall,* Julia Femel,* Sebastian Kalamajski,*
,1
Agneta Lukinius,
†
Magnus Åbrink,* Lars Hellman,
‡
and Anna-Karin Olsson*
,2
*Department of Medical Biochemistry and Microbiology and
‡
Department of Cell and Molecular
Biology, Biomedical Center, and
†
Department of Genetics and Pathology, Rudbeck Laboratory,
Uppsala University, Uppsala, Sweden
ABSTRACT Monoclonal antibody-based therapies have
made an important contribution to current treatment strate-
gies for cancer and autoimmune disease. However, the cost
for these new drugs puts a significant strain on the health-
care economy, resulting in limited availability for patients.
Therapeutic vaccination, defined as induction of immu-
nity against a disease-related self-molecule, is therefore
an attractive alternative. To analyze the potential of
such an approach, we have developed a vaccine against
the extra domain-B (ED-B) of fibronectin. This 91-aa
domain, inserted by alternative splicing, is expressed
during vasculogenesis in the embryo, but essentially
undetectable under normal conditions in the adult.
However, ED-B is highly expressed around angiogenic
vasculature, such as in tumorigenesis. Here, we demon-
strate that it is possible to break self-tolerance and
induce a strong antibody response against ED-B by
vaccination. Nineteen of 20 vaccinated mice responded
with production of anti-ED-B antibodies and displayed
a 70% reduction in tumor size compared to those
lacking anti-ED-B antibodies. Analysis of the tumor
tissue revealed that immunization against ED-B induced
several changes, consistent with an attack by the im-
mune system. These data show that tumor vascular
antigens are highly interesting candidates for develop-
ment of therapeutic vaccines targeting solid tumors.—
Huijbers, E. J. M., Ringvall, M., Femel, J., Kalamajski,
S., Lukinius, A., Åbrink, M., Hellman, L., Olsson, A.-K.
Vaccination against the extra domain-B of fibronectin
as a novel tumor therapy. FASEB J. 24, 4535–4544
(2010). www.fasebj.org
Key Words: therapeutic 䡠immunization 䡠neovascularization
䡠extracellular matrix 䡠angiogenesis
Vaccination programs have served to virtually erad-
icate disabling and life-threatening diseases such as
polio, diphtheria, and smallpox. The possibility of
using vaccination as a treatment strategy also for can-
cer, allergy, and other inflammatory conditions, has for
the past years been the focus of intense research.
However, no vaccination that targets a self-molecule is
yet in clinical use. There are several reasons why the
development of vaccines for noninfectious disease is
more difficult. First, the antigen to be targeted is a
self-molecule. In contrast to vaccination against foreign
substances, such as virus or bacterial antigens, there is a
requirement to break self-tolerance of the immune
system toward the self-antigen (1). Vaccination against
tumor cell antigens is further complicated by the fact
that tumor cells have developed mechanisms to evade
recognition by the immune system, for example, by
decreased presentation of endogenous peptides on
major histocompatibility complex (MHC) (2). We have
carefully addressed these issues with the aim of increas-
ing the chances of clinical development of a cancer
vaccine. The immunization technique that we use is
highly efficient in breaking self-tolerance compared to
traditional methods, since it does not depend on MHC
presentation of the self-antigen (3, 4). To circumvent
the immune escape by the tumor cells, we have focused
on a tumor vascular antigen as a vaccine target, instead
of a tumor cell antigen. Tumor vessels are significantly
different from normal vessels, which can be exploited
for development of new cancer therapies (5). Not only
are the tumor vessels less functional, they also express
or overexpress distinct molecules, which can be used
for targeted cancer therapies. One well-characterized
marker in this category is the extra domain-B of fi-
bronectin (ED-B). ED-B is a 91-aa domain inserted into
fibronectin by alternative splicing (6, 7) (Fig. 1A). The
ED-B-containing splice variant of fibronectin is highly
expressed by developing blood vessels during embryo-
genesis. In the adult, ED-B is absent under normal
conditions and only expressed in situations involving
neovascularization, such as wound healing and tumor
growth (8, 9). In addition, ED-B is not only produced
by the vasculature, but it can also be expressed by the
tumor cells themselves in certain situations (10). Tar-
geting of ED-B by administration of antibodies coupled
1
Present address: Department of Orthopaedic Surgery and
Genetics, Children’s Hospital and Harvard Medical School,
Boston, MA 02115, USA.
2
Correspondence: Department of Medical Biochemistry
and Microbiology, Uppsala University, Biomedical Center,
Husargatan 3, Box 582, SE-751 23 Uppsala, Sweden. E-mail:
anna-karin.olsson@imbim.uu.se
doi: 10.1096/fj.10-163022
45350892-6638/10/0024-4535 © FASEB
to radioactive or cytotoxic agents has been developed
by the research groups of Luciano Zardi and Dario
Neri, and these strategies have given promising results
in mouse models of cancer (11–14) and also in a
clinical study (15). Instead of using ED-B as a target for
the delivery of cytotoxic compounds to the tumor
tissue, we have developed a vaccine that induces pro-
duction of antibodies directed against ED-B, which
elicits an immune response against the tumor tissue
where ED-B is expressed. A major advantage of this
approach is that the manufacturing cost will only be a
fraction of that required for repeated and long-term
administration of a GMP-produced monoclonal anti-
body.
MATERIALS AND METHODS
Expression vectors for recombinant proteins
TRX-EDB
The region encoding the ED-B domain and a N-terminal
His-tag (6xhistidine) (309 bp) was inserted in frame, down-
stream the bacterial TRX sequence (354 bp), in a pET-21a
vector (Novagen; EMD Chemicals, Gibbstown, NJ, USA). The
resulting expression vector was named pET21a TRX-EDB.
GST-EDB
The TRX sequence in the pET21a TRX-EDB vector was
replaced with a PCR-amplified GST sequence (678 bp). The
resulting vector construct was named pET21a GST-EDB.
Expression and purification of recombinant fusion-proteins
The vectors pET21a TRX-EDB and pET21a GST-EDB were
transformed into Escherichia coli Rosetta gami (DE3) (Nova-
gen; EMD Chemicals) for expression of the fusion proteins.
Rosetta gami (DE3) is an E. coli strain that is optimized for
production of eukaryotic proteins. Overnight cultures were
diluted 1:10 and grown until OD
600
0.5. Protein expression
was induced with 1 mM isopropyl -D-1-thiogalactopyrano-
side (IPTG) (VWR International AB, Stockholm, Sweden).
TRX-EDB protein expression was induced at 22°C for 16 h,
and expression of GST-EDB was induced at 37°C for 4 h. The
lower temperature was used to increase the soluble fraction of
TRX-EDB.
Fusion proteins were released by sonication of the bacteria
on ice for 4 ⫻30 s (repeated 3 times). Bacterial debris was
pelleted by centrifugation at 2500 gand the supernatant was
collected. Thirty milliliters of bacterial supernatant (originat-
ing from a 1.2-L bacterial culture) was mixed with 1 ml
Ni-NTA agarose slurry (Qiagen, Hilden, Germany) and incu-
bated “end-over-end” at 4°C for 3 h. The Ni-NTA-agarose was
pelleted by centrifugation, washed 5 times with PBS pH 7.0/1
M NaCl/0.1% Tween-20, transferred to a column with a glass
filter (Sartorius AG, Go¨ttingen, Germany) and washed again.
The protein was eluted with 100 mM imidazole (Sigma-
Aldrich GmbH, Steinheim, Germany) in 20 mM Tris (pH
8.0)/0.1 M NaCl in 500-l fractions. Purification of the
GST-EDB protein was performed as above, with the exception
that the Ni-NTA-agarose column with bound GST-EDB was
treated with 50 mM imidazole before elution, to reduce
background. Protein-containing fractions were pooled and
dialyzed against PBS (pH 7.0) [Spectra/Por CE (cellulose
ester) membrane, 6- to 8-kDa M
W
cutoff; Spectrum Medical
Industries, Los Angeles, CA, USA]. Final protein concentra-
tion was estimated by comparison with a BSA fraction V
(Roche Diagnostics, Mannheim, Germany) standard on SDS-
PAGE and by a protein quantification assay (BCA Protein
Assay; Pierce, Rockford, IL, USA). Purified fractions of both
TRX-EDB and GST-EDB protein were analyzed by mass-
spectrometry to confirm their identity.
Cell culture
T241 fibrosarcoma cells (American Type Culture Collection,
Manassas, VA, USA) were cultured in DMEM ⫹Glutamax
(31966; Invitrogen AB, Lidingo¨, Sweden) supplemented with
10% FCS (ECS 0180L; Euroclone Ltd, Devon, UK).
Animal studies
Animal work was approved by the Uppsala University board of
animal experimentation (reg. no. C207/7) and thus per-
formed according to the UK Coordinating Committee on
Cancer Research (UKCCCR) guidelines for the welfare of
animals in experimental neoplasia (16). The mice were
anesthesized with isoflurane (Forene; Abbott Scandinavia,
Solna, Sweden) (induction 4.5%; maintenance 2.5–3%) dur-
ing all manipulations. Four- to 8-wk-old female C57BL/6 mice
(Taconic, Lille Skensved, Denmark), were immunized in the
groin with 100 l of an emulsion containing 100 g recom-
binant TRX-EDB protein in PBS, mixed 50:50 with Freund’s
complete adjuvant (FCA) (263810; Difco Laboratories, De-
troit, MI, USA) atd1oftheexperiment. Booster injections
containing the same amount of recombinant protein but in
Freund’s incomplete adjuvant (FIA) (F5506; Sigma-Aldrich)
were given in the opposite groin at d 14 and 28. Control
animals received PBS in FCA or FIA. Ten mice/group were
included in both experiments. At d 35, all mice were inocu-
lated subcutaneously in the left flank with 0.5 ⫻10
6
T241
fibrosarcoma cells in a total volume of 100 l in PBS. The
tumors were allowed to grow for 21 d (first experiment) or
26 d (second experiment), until the maximum allowed size
was reached. One mouse was lost in the first experiment and
2 mice in the second experiment, due to ruptured tumors. At
the last day of the experiment, blood samples were taken
from all mice before euthanization. Tumors were removed
and measured with a caliper before cryopreservation in
isopentane/dry ice. Tumor volume was calculated using the
formula width
2
⫻length ⫻/6. The frozen tissue was stored
at ⫺80°C until further processing.
ELISA
For analysis of total anti-ED-B antibody levels, blood samples
were left to coagulate overnight at 4°C and centrifuged twice
at 13,000 rpm for 5 min in a microcentrifuge. Supernatants
(serum) were collected and stored at ⫺20°C until use. ELISA
plates (Costar MaxiSorp surface; Corning Incorporated,
Corning, NY, USA) were coated with GST-EDB (10 g/ml
protein in PBS) and blocked with centrifuged horse serum
(Life Technologies; Biocult, Glasgow, UK). Plates were incu-
bated with serum from control or TRX-EDB vaccinated mice
(diluted 1:25 in horse serum), followed by biotinylated goat
anti-mouse IgG (H⫹L) (BA-9200; Vector Laboratories, Burl-
ington, ON, Canada) diluted 1:500 and streptavidin-horserad-
ish peroxidase (SA-HRP) (SA-5004; Vector Laboratories)
diluted 1:500. All incubations were performed at 37°C. HRP
activity was detected by incubation with TMB substrate
4536 Vol. 24 November 2010 HUIJBERS ET AL.The FASEB Journal 䡠www.fasebj.org
(T8665; Sigma-Aldrich). The absorbance was measured at
405 nm.
For analysis of the levels of anti-ED-B antibodies of differ-
ent subclasses serum from control and TRX-EDB vaccinated
mice without tumors, was analyzed by ELISA. The assay was
performed as described above with the exceptions that bio-
tinylated goat anti-mouse antibodies specific against IgG
subclasses 1, 2a, 2b, and 3 (1070-08, 1080-08, 1090-08, 1110-
08; Southern Biotech, Birmingham, AL, USA) were used as
secondary reagent (diluted 1:2500 in PBS) and that the
absorbance was read at 650 nm.
Analysis of tumor necrosis
Estimation of the necrotic area fraction (%) was performed
on tumor sections stained with Mayer’s hematoxylin and
eosin (HistoLab products AB, Go¨teborg, Sweden) and CD31,
using an ocular grid. Necrosis was defined as avascular areas
lacking nuclear structures. Four tumors of relatively equal size
from control and TRX-EDB vaccinated mice from experi-
ment 1 and 2, respectively (in total n⫽8/group), were
selected for the analysis, which was performed in a masked
fashion. Pictures of necrotic areas were taken in a Nikon
Eclipse 90i microscope with NIS Elements 3.06 software
(Nikon Instruments B.V. Europe, Amstelveen, The Nether-
lands), with the ⫻40 objective.
IHC of glioma tissue using serum from control and
TRX-EDB immunized mice
Cryosections of normal brain tissue and grade III glioma from
Gtv-a Arf
⫺/⫺
mice (17) were used for IHC with serum from
control or TRX-EDB immunized mice. The transgenic Gtv-a
Arf
⫺/⫺
glioma model expresses the avian leukemia virus
receptor tv-a in glial progenitor cells. Injection of the avian
virus RCAS, engineered to express platelet-derived growth
factor B (PDGFB), in the brain of newborn Gtv-a mice
induces development of glioma, with a similar histology to
humans. The tissue used for the current IHC analysis was
derived from RCAS-PDGFB virus injected Gtv-a mice that
were also deficient for the tumor suppressor gene Arf. The
sections were blocked with FCS, incubated with mouse-serum
(diluted 1:10) from control or TRX-EDB-immunized animals,
and treated with 1% hydrogen peroxide (Merck KGaA,
Darmstadt, Germany). Detection of primary antibody was
performed with biotinylated anti-mouse IgG (BA-9200; Vector
Laboratories) diluted 1:500, followed by incubation with
SA-HRP diluted 1:500 and AEC substrate (SK-4200; Vector
Laboratories). Counterstaining was performed with Mayer’s
hematoxylin (HistoLab). For negative controls, the primary
antibody was omitted.
Electron microscopy
T241 fibrosarcoma tumor tissue from 4 control- and 4 TRX-
EDB-immunized mice were fixed in 0.15 M sodium cacody-
late-buffered 2.5% glutaraldehyde and postfixed in 0.1 M
s-collidine-buffered 2% osmium tetroxide, dehydrated, and
embedded in epoxy resin. Ultrathin sections were analyzed by
electron microscope (Philips CM-10; FEJ; Philips AB, Stock-
holm, Sweden). Approximately 15–20 vessels from each indi-
vidual were analyzed.
Fibrinogen staining and quantification
T241 fibrosarcoma cryosections from control- and TRX-EDB-
immunized mice (n⫽4/group) were blocked in 3% BSA for
1 h at room temperature (RT) and double stained overnight
at 4°C with purified rat anti-mouse CD31 (PECAM-1) mono-
clonal antibody (cat. no. 553370; BD Pharmingen, BD Bio-
sciences, Stockholm, Sweden) diluted 1:1000 and polyclonal
rabbit anti-fibrinogen (cat. no. A0080; DakoCytomation,
Glostrup, Denmark) diluted 1:500 in blocking solution. De-
tection antibodies used were Alexa555 donkey anti-rabbit
(cat. no. A31572; Invitrogen) and Alexa488 rabbit anti-rat IgG
(H⫹L) (cat. no. A21210; Invitrogen), at a concentration of 2
g/ml in blocking solution. Incubation time was 45 min at
RT. Nuclei were visualized with Hoechst 33342 (VWR Inter-
national) at 1 g/ml. Sections were washed, mounted with
Fluoromount-G (Southern Biotech), and stored at 4°C until
use. For negative controls, the primary antibody was omitted.
For quantification of the fibrinogen-stained area, three pic-
tures of each tumor section (n⫽12 areas) were taken at
random in a Nikon Eclipse 90i microscope with NIS Elements
3.06 software, with the ⫻20 objective and exposure time 400
ms. Analysis of the pictures was performed with Image J 1.42
software (National Institutes of Health, Bethesda, MD, USA).
Neutrophil staining and quantification
T241 fibrosarcoma cryosections from control- and TRX-EDB-
immunized mice (n⫽4/group) were blocked with FCS (ECS
0180L; Euroclone, Devon, UK) for1hatRTandstained with
purified rat anti-mouse Ly6G and Ly-6C (Gr-1) monoclonal
antibody (cat. no. 553123; BD Pharmingen, BD Biosciences,
Stockholm, Sweden) at a concentration of 0.5 g/ml in
blocking solution overnight at 4°C. Detection of primary
antibody binding sites was done with an Alexa555 goat anti-rat
IgG (H⫹L) (Invitrogen) at a concentration of 2 g/ml in
blocking solution for 45 min at RT. Nuclei were visualized
with Hoechst 33342 (VWR International) at 1 g/ml. Sec-
tions were washed, mounted with Fluoromount-G (Southern
Biotech), and stored at 4°C until use. For negative controls,
the primary antibody was omitted. For quantification of the
Gr1-stained area, 3 pictures of each tumor section (n⫽12
areas) were taken at random in a Nikon Eclipse 90i micro-
scope with NIS Elements 3.06 software, with the ⫻20 objective
and exposure time 300 ms. Analysis of the pictures was
performed with the Image J 1.42 software.
Wound-healing assay
Animal work was approved by the Uppsala University board of
animal experimentation (reg. no. C63/9). Immunocompe-
tent C57BL/6 mice were immunized as described previously.
One week after the second boost, mice were anesthesized with
isoflurane (Forene; Abbott Scandinavia) (induction 4.5%;
maintenance 2.5–3%) and a full-thickness wound, 8 mm in
diameter, was placed on the back using a disposable biopsy
punch (Pfm Medical, Carlsbad, CA, USA). Buprenorphine sc
(Temgesic; Schering-Plough AB, Stockholm, Sweden) 0.05–
0.1 mg/kg was given 2⫻/dforthefirst2daspain medica-
tion. In addition, Bimotrim (trimethoprim: sulfadoxine 1:5;
CEVA Vetpharma, Lund, Sweden) 0.2% (w/v) was adminis-
tered to the drinking water during the first 5 d of the
experiment. Wounds were protected with Cavilon no-sting
barrier spray (3M, Sollentuna, Sweden). Mice were observed
every day for the first7doftheexperiment. Pictures of the
wounds were taken on d 0, 7, and 14 after wounding with a
Nikon D70 camera. At d 20, the experiment was terminated,
and all animals were sacrificed, and blood was drawn for
analysis of anti-ED-B antibody titers with ELISA.
Arthritis analysis
Control- and EDB-immunized mice were examined for signs
of arthritis, both at the macroscopic and the cellular level by
4537TARGETING TUMOR VESSELS BY THERAPEUTIC VACCINATION
a person with experience from mouse models of rheumatoid
arthritis.
Macroscopic analysis
The examined mice (n⫽6/group) had anti-ED-B antibodies
in their circulation for 8 mo. Signs of arthritis were analyzed
as described previously (18) in a masked fashion.
Cellular analysis
Control- and TRX-EDB-immunized mice were sacrificed ⬃2
mo after immunization; front and hind paws were incubated
in 10% buffered formalin solution, extensively washed in tap
water, and transferred to Parengy solution (BIE & Berntsen
A/S, Roedovre, Denmark) for decalcification. The solution
was exchanged to fresh Parengy solution after 30 h, and the
tissue was further incubated for 4 d, washed in tap water for
24 h, dehydrated, and paraffin embedded. The paws were
sectioned sagitally, and stained with hematoxylin and eosin.
Sections of 3 animals in each group were examined in a
masked analysis for histopathological changes, such as cellu-
lar infiltration and cartilage destruction.
Statistical analysis
For statistical analysis a Mann-Whitney Utest was performed,
and values of P⬍0.05 were considered significant. The
Mann-Whitney Utest was used because of the small sample
size (n⫽10) and because normal distribution could not be
assumed.
RESULTS
The basis of the vaccination technique is to produce
the self-antigen of choice as a fusion protein with a
foreign (i.e., nonself) part, derived, for example,
from bacteria (1). The mechanism by which autore-
active B cells are activated to become antibody-
producing plasma cells is described in Fig. 1B. The
polyclonal anti-self antibodies bind to the target pro-
tein, and the resulting immune complexes will be
cleared by phagocytosis if the target protein is soluble.
When the antigen is membrane- or matrix-bound (as in
the case of ED-B), “frustrated phagocytosis” will com-
mence, since the immune complexes cannot be phago-
cytosed. This results in recruitment of monocytes/
macrophages, neutrophils, and natural killer cells,
which leads to tissue damage where the antigen is
expressed.
Two essentially pure and soluble recombinant fusion
proteins containing ED-B as the self-antigen part were
produced by expression in bacteria (see Fig. 1Cfor a
schematic illustration of the expression vector). ED-B
was either fused to the E. coli-derived protein thiore-
doxin (TRX) or glutathione-S-transferase (GST; de-
rived from the parasite worm Schistozoma japonicum),
generating the fusion proteins TRX-EDB and GST-EDB
(Fig. 1D). C57BL/6 mice were immunized with TRX-
EDB or vehicle in Freund’s adjuvant. After 5 wk,
including 2 booster immunizations, all mice received
subcutaneous injections of T241 fibrosarcoma cells.
Before the experiment was initiated, expression of
ED-B in this particular tumor type was confirmed using
a synthetic anti-ED-B antibody, kindly provided by Prof.
Dario Neri (ETH Zurich, Zurich, Switzerland; data not
shown). After 21 d, blood was collected from all mice,
tumors were removed, measured with a caliper and
cryopreserved. The tumor volume in TRX-EDB vacci-
Figure 1. Mechanism for breaking self-tolerance (adapted from ref. 1). A) Illustration of a fibronectin monomer with the
alternatively spliced ED-B indicated. B)1) Antigen presenting cells (APC) internalize the fusion protein (TRX-EDB) and
present nonself (TRX) and self (ED-B) peptides via MHC class II. 2) Nonself peptides are recognized by the T-cell receptor
(TCR) on T-helper cells (TH) that become activated. Self-peptides are not recognized, since autoreactive T cells are deleted in
the thymus during development. 3) Autoreactive B cells are, however, present in the circulation and recognize the fusion
protein’s self-part (ED-B) via their B-cell receptor (BCR). These autoreactive B-cells take up the fusion protein and present
peptides from both the self- and the nonself part on MHC class II. TH cells previously activated by the foreign part of the fusion
protein now activate these autoreactive B cells, since they present the same foreign peptides. 4) Autoreactive B-cells undergo
clonal expansion and produce anti-ED-B antibodies. C) Schematic illustration of the pET21a bacterial expression vector
encoding TRX-EDB under the control of the IPTG-inducible T7lac promoter. His-tag is included for the purpose of purification.
D) Schematic representation of the fusion proteins TRX-EDB (20 kDa) and GST-EDB (33 kDa) and their appearance on
reducing SDS-PAGE after purification.
4538 Vol. 24 November 2010 HUIJBERS ET AL.The FASEB Journal 䡠www.fasebj.org
nated mice was significantly reduced compared to the
control group (Fig. 2A; 1st experiment). An attenuated
tumor growth in TRX-EDB vaccinated mice was con-
firmed in a second independent experiment (Fig. 2A;
2nd experiment). Antibodies against ED-B were present
in the TRX-EDB-immunized mice in both experi-
ments (Fig. 3A). Interestingly, the only mouse that
for unknown reasons did not generate any significant
amount of anti-ED-B antibodies above background
(Fig. 3A, 2nd experiment; pound sign) developed the
largest tumor in this group (Fig. 2A, 2nd experiment;
pound sign). The average reduction of tumor volume
in mice immunized against ED-B, compared to those
lacking anti-ED-B antibodies, was 68.5% (1st experi-
ment, 71%; 2nd experiment, 66%; Fig. 2B). Histolog-
ical analysis of hematoxylin-and-eosin-stained tumor
tissue from both experiments showed larger necrotic
areas in the TRX-EDB group compared to controls
(Fig. 2C). The pictures in Fig. 2Dare examples of a
viable and a necrotic area of the tumor tissue.
To analyze the antibody response against ED-B in
more detail, we determined which IgG subclasses that
Figure 2. Reduced tumor size in TRX-EDB-vaccinated mice.
A) TRX-EDB (EDB)-vaccinated mice displayed a significantly
reduced tumor size compared to the control group (Ctrl) in 2
independent experiments: 1st experiment, P⫽0.0007; 2nd
experiment, P⫽0.0188. Each dot represents the tumor
volume of 1 individual. One animal with no or very low
amounts of anti-ED-B antibodies is denoted by pound sign (#;
see Fig. 3A, 2nd experiment). B) Mice from the 1st and 2nd
experiment grouped according to the presence of anti-ED-B
antibodies (EDB ab
⫺
and EDB ab
⫹
;P⬍0.0001). C) Necrotic
area fraction (%) in 8 tumors from control and TRX-EDB-vaccinated mice, respectively, from the 1st and 2nd experiment
(P⫽0.0148). D) Photos illustrating a viable (left) and a necrotic (right) tumor area. Viable tissue has clearly distinguishable
nuclei and is vascularized, as illustrated by the CD31-positive blood vessel (red). Necrotic tissue contains no clear structures.
Scale bar ⫽50 m.
Figure 3. Functional anti-ED-B antibodies are produced in TRX-EDB
vaccinated mice. A) Serum from control (Ctrl) and TRX-EDB (EDB)-
vaccinated mice was analyzed for the presence of anti-ED-B antibodies
using ELISA. Each bar represents an individual animal. Animal lacking
anti-ED-B antibodies is indicated by pound sign (#). B) Serum from
control and TRX-EDB-vaccinated mice without tumors was analyzed for
anti-ED-B IgG subclasses; Ig1, IgG2a, IgG2b, and IgG3. C) Immunohis-
tochemical staining of normal brain tissue (top panels) and grade III
experimental glioma from mouse (bottom panels) using serum from
control (left panels) and TRX-EDB (right panels)-immunized animals.
Scale bar ⫽20 m.
4539TARGETING TUMOR VESSELS BY THERAPEUTIC VACCINATION
were generated in immunized mice (Fig. 3B). As ex-
pected, the major subclass was found to be IgG1,
characteristic of a TH2-response, meaning an antibody-
mediated immune response.
To confirm that the anti-ED-B antibodies in immu-
nized mice were functional and recognized native ED-B
in tissue, we performed immunohistochemical staining
of murine grade III glioma tissue (17), using serum
from control or TRX-EDB vaccinated mice. Gliomas are
known to express high levels of ED-B around the tumor
vasculature (19), in contrast to normal brain tissue,
where ED-B is undetectable. As expected, no staining
was detected when normal brain tissue was incubated
with serum from control or TRX-EDB vaccinated mice
(Fig. 3C, top panels). However, when glioma tissue was
incubated with serum from TRX-EDB vaccinated mice,
a typical ED-B expression pattern was visualized along
the vasculature, which was not seen with serum from
the control animals (Fig. 3C, bottom panels). This
clearly demonstrates that functional anti-ED-B antibod-
ies, recognizing native ED-B in tissue, are produced in
mice immunized with TRX-EDB.
Stereological analysis of tumor vascularization (20)
using CD31 as a marker, showed no obvious differences
between the control and ED-B-vaccinated group (data
not shown). However, electron microscope analysis of
the tumor tissue revealed an altered morphology of the
vasculature in TRX-EDB immunized mice (Fig. 4). The
vessel wall displayed invaginations and protrusions
(Fig. 4B2, arrow) not seen in the control group (Fig.
4A1, arrow). Furthermore, transport vesicles present
in endothelium from controls (Fig. 4A1, arrowheads)
were less frequent in TRX-EDB immunized mice.
Instead, a large number of free ribosomes was seen in
tumor endothelial cells from mice carrying anti-ED-B
antibodies, indicating an acute cellular response
(Fig. 4B3, arrowheads). In addition, macrophages
actively trying to phagocyte the endothelium were
seen in tumors from mice vaccinated against ED-B.
These macrophages displayed a dilated endoplasmic
reticulum (Fig. 4B4, asterisk), indicating production
of proteolytic enzymes for lysosomal degradation.
Collectively, these observations are consistent with an
immunological response against the tumor blood
vessels in TRX-EDB-immunized mice and indicate a
decreased functionality. To address vessel function,
Figure 4. Altered vessel morphology and fibrinogen extravasation in tumors from TRX-EDB-vaccinated mice. Representative electron
micrographs of blood vessels from ultrathin sections of tumors from control (Ctrl; A) and TRX-EDB (EDB; B) vaccinated mice. Panels
A1 and B2–B4 are magnifications of the corresponding area in Aand B.A1) Arrowheads indicate transport vesicles; asterisk indicates
pericyte. B2) Arrow indicates invaginations and protrusions of the vessel wall, not seen in the control group (compare with A1; arrow).
B3) Arrowhead indicates free ribosomes. B4) Asterisk indicates dilated endoplasmic reticulum. C) Top panels: representative
immunohistochemical stainings for CD31 (green) and fibrinogen (red) in tumor tissue from control and TRX-EDB-vaccinated
animals. Bottom panel: tumors from animals with anti-ED-B antibodies showed a significantly increased area positive for fibrinogen
staining (P⫽0.0351). Data are derived from n⫽4 tumors/group; each dot represents 1 quantified field, 3 fields/tumor. Similar
symbols are used to denote individual tumors in the 2 groups. Scale bars ⫽2m(A,B); 200 m(C).
4540 Vol. 24 November 2010 HUIJBERS ET AL.The FASEB Journal 䡠www.fasebj.org
we performed immunohistochemical staining for fi-
brinogen in tumor tissue derived from control- or
EDB-immunized mice, as a measurement of vascular
leakage. The stainings revealed significantly in-
creased amounts of extravasated fibrinogen in the
tumor tissue derived from mice with anti-ED-B anti-
bodies (Fig. 4C).
To address whether an ongoing immune response in
the tumor tissue could be detected at a cellular level, we
performed immunohistochemical staining for neutro-
phils using the marker Gr1. The area with Gr1-positive
staining was quantified using an image analysis program.
A significantly increased amount of infiltrating neutro-
phils was detected in TRX-EDB-immunized mice (Fig. 5).
As previously mentioned, ED-B is expressed during
neovascularization in wound healing. From a safety
perspective, it is, therefore, important to address
whether wound healing is affected by the presence of
anti-ED-B antibodies. A wound-healing assay was, there-
fore, performed in control- and EDB-immunized mice.
A full-thickness wound of ⬃1 cm in diameter was
placed on the back of the mice using a biopsy punch.
We followed the mice closely during the healing period
but could not detect any differences between the
groups. After 14 d, the wounds of all mice were
completely healed (Fig. 6A). As can be seen in Fig. 6B,
all TRX-EDB-immunized mice in the wound healing
assay carried anti-ED-B antibodies.
In addition to situations with neovascularization,
there are also some reports on ED-B expression in
cartilage (21, 22). One concern was, therefore, whether
TRX-EDB immunized mice could develop arthritis. We
have carefully analyzed the TRX-EDB immunized mice,
both at a macroscopic and cellular level, as described in
Materials and Methods, but failed to detect any signs of
arthritis (data not shown).
DISCUSSION
Vaccination against ED-B induced high levels of anti-
ED-B antibodies, which was previously believed to be
very difficult, or even impossible, due to the high grade
of conservation of this protein domain (23, 24). The
amino acid sequence of ED-B is, in fact, identical in
mouse, rat, rabbit dog, monkey, and human (5). Our
Figure 5. Increased neutrophil infiltration in tumors from TRX-EDB-immunized mice. Left and center panels: representative
immunohistochemical stainings for neutrophils (red), using the marker Gr1, in tumor tissue from control (Ctrl) and TRX-EDB
(EDB)-vaccinated animals. Right panel: Gr1-positive area (%) was significantly increased in tumors from TRX-EDB-vaccinated
mice (P⫽0.0061). Data are derived from n⫽4 tumors/group; each dot represents 1 quantified field, 3 fields/tumor. Similar
symbols are used to denote individual tumors in the two groups. Scale bars ⫽200 m.
Figure 6. No impairment of wound healing in TRX-EDB-
vaccinated mice. A) Full-thickness wounds of control (Ctrl) and
TRX-EDB (EDB) vaccinated animals (n⫽5/group) are depicted
at d 0, 7, and 14 after wounding. No difference in wound healing
between the two groups was observed. B) Anti-ED-B antibodies
were present in serum of all 5 TRX-EDB-immunized mice
included in the wound healing assay, but not in the control
mice. Bars represent individual animals.
4541TARGETING TUMOR VESSELS BY THERAPEUTIC VACCINATION
data now show that immunization against ED-B is,
indeed, possible and therefore also isolation of monoclo-
nal anti-ED-B antibodies using hybridoma technology.
The anti-ED-B antibodies generated by the vaccine
had a protective effect against tumor growth; ⬃70%
reduction in tumor volume was achieved by vaccination
against this single tumor vascular antigen. Targeting a
combination of 2 or more such antigens may prove
even more effective with respect to suppression of
tumor growth. In this context, it is important to remem-
ber that the antibodies that are generated by immuni-
zation are polyclonal and therefore significantly more
efficient in immune complex formation and activation
of the immune system than a monoclonal antibody.
Electron microscope analysis, in combination with im-
munohistochemical staining for fibrinogen, revealed
an aberrant morphology and increased leakage of
tumor vessels in TRX-EDB immunized mice compared
to controls. Moreover, an increased infiltration of neu-
trophils in tumors from mice with anti-ED-B antibodies
support the conclusion of an ED-B targeted attack by
the immune system.
An important safety factor is that this type of vacci-
nation is reversible and not life-long, in contrast to
immunization against certain foreign antigens (4, 25,
26). By “reversible” we mean 2 things: first, the antibody
titers decrease to background levels at a faster rate
when you immunize against a self-molecule than is the
case for immunization against a foreign antigen; and
second, a new injection of the fusion protein TRX-EDB
is required to increase the antibody titers again. Using
the same therapeutic vaccination technique as in the
present study, others (4, 25) demonstrated the revers-
ibility of the immune response against IgE and IL-13.
After 3–4 mo, the majority of the therapeutic antibod-
ies were gone and after ⬃11 mo, the antibody titers
were reduced by 89–99% (4). Ongoing studies in our
laboratory indicate that the same kinetics apply to
antibody production against ED-B (unpublished re-
sults). The main reason for this observation is that the
antigen is a self-molecule. When an autoreactive B cell
meets its antigen (for instance, ED-B) in the absence of
T-cell help (which is provided by the TRX-part of the
fusion protein TRX-EDB in the vaccine), the result is
inactivation of the B cell (anergy). This mechanism is
part of the peripheral tolerance of the immune system,
which is aimed to reduce the risk of autoimmune
responses. If instead a B cell has specificity for a nonself
antigen, the result will be activation of the B cell on
antigen encounter, since at the same time, it will
receive help from T cells directed against the same
nonself molecule. Even though repeated immuniza-
tions with the fusion protein may be required, the
reversibility is an advantage in case adverse effects are
observed. A relevant comparison is the time it takes to
clear an injected monoclonal antibody (MAb) from the
circulation. MAbs have a half-life of ⬃3 wk, which
means that it takes ⬃3 mo to clear 90% of the injected
antibody.
Using the immunization strategy with fusion proteins
presented in this work, we consider the risk for auto-
immune disease as very low for several reasons. It is
generally very difficult to break self-tolerance. Although
our approach works well for this purpose, 2 require-
ments have to be fulfilled at the same time: the pres-
ence of the fusion protein TRX-EDB, and concomitant
administration of an adjuvant. The presence of TRX-
EDB alone, without additional immunostimulatory
compounds, will not induce production of anti-ED-B
antibodies. Moreover, as described above, the presence
of endogenous ED-B will suppress, not enhance, activa-
tion of autoreactive B cells with specificity for ED-B.
Another safety issue that has been raised is whether
the described immunization strategy against ED-B
could potentially induce immune complex glomerulo-
nephritis. Immune complexes will, however, only de-
posit in the kidney if the targeted antigen is soluble. In
our case, the antigen is matrix bound. Moreover,
vaccination induces a slow increase in antibody titers,
and macrophages, therefore, have the capacity to clear
immune complexes as they form, even for a soluble
antigen. This is in sharp contrast to injection of hun-
dreds of milligrams of a monoclonal antibody at one
single occasion, which is usually the case when treating
patients with monoclonal antibodies.
In a recent report by Haller et al. (27), the re-
searchers clearly show that immunization against the
endothelial tip cell protein delta-like ligand 4
(DLL4) delays growth of mammary carcinomas in
mouse. In this study, a DNA vaccine approach was
used where the foreign epitopes were provided by
human DLL4. However, there are additional safety
issues to consider when targeting a self-molecule with
a DNA vaccine. There is always a risk for integration
of the vaccine DNA into the host genome. As long as
a foreign antigen, like a virus, is the target, the risk of
host integration may be acceptable from a clinical
perspective. But when it comes to self-molecules, it is
crucial that the immune response is reversible, espe-
cially if the expression pattern of the target antigen is
not completely known.
There are also previous studies reporting therapeutic
effects on tumor growth in mice following vaccination
against angiogenesis-related molecules, such as VEGF
and VEGFR2. However, the chances of clinical approval
of a vaccine against VEGFR2, present on all blood
vessels in the body, are minor. The important distinc-
tion between those vaccines and the ED-B vaccine we
present here is that immunological targeting of ED-B is
clinically feasible due to its restricted expression pat-
tern. The first clinical study targeting ED-B was recently
reported (15). In this study, Hodgkin lymphoma pa-
tients received an anti-ED-B antibody conjugated to
131
I
(
131
I-L19SIP). The treatment induced a sustained par-
tial remission in two out of three relapsed patients (15),
indicating that targeting of ED-B is interesting for
future development as a cancer treatment. This conclu-
sion is also supported by the well-documented tumor-
specific expression pattern of ED-B in a large number
of distinct tumor types, such as head and neck tumors
4542 Vol. 24 November 2010 HUIJBERS ET AL.The FASEB Journal 䡠www.fasebj.org
(28), high-grade astrocytomas (24), colorectal carcino-
mas (29), breast cancer (30), as well as lung and gastric
carcinoma and malignant melanoma (7).
The effect of the EDB vaccine on established tumors
is highly relevant and remains to be addressed. There
is, however, a major obstacle connected to this ap-
proach and that is that we need a slow-growing tumor
model in an immunocompetent mouse strain for this
purpose. The majority of subcutaneous tumor models
in immunocompetent mice are fast growing. From
when the tumor has become established and is palpa-
ble, it may take 2 to 3 wk or less until the mouse has to
be euthanized, either due to its health condition or
because the tumor has reached the maximum allowed
size according to ethical regulations. Antibody produc-
tion is not instantaneous after immunization, but re-
quires a few weeks (and repeated injection) to develop.
It is, therefore, not possible to perform this type of
experiment in the fast growing subcutaneous models
available. There are plenty of slow-growing tumor mod-
els in immunocompromised mice (SCID, Nude), but
for obvious reasons these mouse strains cannot be used
when studying effects of vaccines that require both B
and T cells. The situation in humans is of course very
different, where a tumor can develop over years.
We could not detect any adverse effects on wound
healing in mice carrying anti-ED-B antibodies, despite
expression of ED-B in this situation. The reason for this
is not known, but one possible explanation could be
that the expression of ED-B in a wound is transient
compared to the situation in tumors. Another possibil-
ity is that the quality of the vessels is different in tumors
and during wound healing (pathological vs. physiolog-
ical angiogenesis). Tumor vessels are constantly leaky,
while increased leakiness is a transient phenomenon in
vessels undergoing physiological angiogenesis. The an-
tibodies need to cross the endothelium to reach ED-B
in the extracellular matrix and targeting of ED-B may,
therefore, be more efficient in the tumor.
As mentioned previously, expression of ED-B in carti-
lage has also been reported. Unlike other connective
tissue, cartilage is not vascularized. The source of the
fibronectin splice variants is the chondrocytes themselves.
One possible explanation for the finding that cartilage
does not seem to be attacked by the immune system in
TRX-EDB-immunized mice, is that the lack of blood
vessels in cartilage may reduce the accessibility of the
anti-ED-B antibodies for their target.
Breaking self-tolerance is a central issue in the field
of therapeutic vaccines. To transfer the ED-B vaccine to
the clinic, it must be based on a less toxic, but equally
potent, adjuvant than Freund’s. On the basis of a large
screen, we have recently identified alternative adju-
vants, which are as efficient as Freund’s in evoking a
self-antigen response, but are biodegradable and non-
toxic (31). This finding will significantly aid clinical
development of therapeutic vaccines.
In summary, our data show that therapeutic vaccines
targeting tumor vascular antigens, single molecules or
combinations, are highly interesting for clinical devel-
opment and could provide potent and cost-efficient
new treatment strategies for cancer.
The authors thank Dr. Lene Uhrbom (Department of
Genetics and Pathology, Uppsala University) for providing
tissue sections of normal mouse brain and glioma and Prof.
Dario Neri (Institute of Pharmaceutical Sciences, ETH Zu-
rich, Zurich, Switzerland) for providing SIP anti-ED-B anti-
bodies. The authors also thank Prof. Wilhelm Jahnen-Dech-
ent and Dr. Daniela Dreymu¨ller (Institute for Biomedical
Engineering, RWTH Aachen, Aachen, Germany) for gener-
ously teaching us the wound-healing model. Financial sup-
port was provided by The Swedish Research Council, The
Swedish Cancer Society, The Swedish Society of Medicine,
Jeansson⬘s Foundation, and the Åke Wiberg Foundation.
REFERENCES
1. Hellman, L. (2008) Therapeutic vaccines against IgE-mediated
allergies. Expert Rev. Vaccines 7, 193–208
2. Ridolfi, L., Petrini, M., Fiammenghi, L., Riccobon, A., and
Ridolfi, R. (2009) Human embryo immune escape mechanisms
rediscovered by the tumor. Immunobiology 214, 61–76
3. Hellman, L. (1994) Profound reduction in allergen sensitivity
following treatment with a novel allergy vaccine. Eur. J. Immunol.
24, 415–420
4. Vernersson, M., Ledin, A., Johansson, J., and Hellman, L.
(2002) Generation of therapeutic antibody responses against
IgE through vaccination. FASEB J. 16, 875–877
5. Neri, D., and Bicknell, R. (2005) Tumour vascular targeting.
Nat. Rev. Cancer 5, 436–446
6. Zardi, L., Carnemolla, B., Siri, A., Petersen, T. E., Paolella, G.,
Sebastio, G., and Baralle, F. E. (1987) Transformed human cells
produce a new fibronectin isoform by preferential alternative
splicing of a previously unobserved exon. EMBO J. 6, 2337–2342
7. Carnemolla, B., Balza, E., Siri, A., Zardi, L., Nicotra, M. R.,
Bigotti, A., and Natali, P. G. (1989) A tumor-associated fibronec-
tin isoform generated by alternative splicing of messenger RNA
precursors. J. Cell Biol. 108, 1139–1148
8. Castellani, P., Viale, G., Dorcaratto, A., Nicolo, G., Kaczmarek,
J., Querze, G., and Zardi, L. (1994) The fibronectin isoform
containing the ED-B oncofetal domain: a marker of angiogen-
esis. Int. J. Cancer 59, 612–618
9. Astrof, S., and Hynes, R. O. (2009) Fibronectins in vascular
morphogenesis. Angiogenesis 12, 165–175
10. Midulla, M., Verma, R., Pignatelli, M., Ritter, M. A., Courtenay-
Luck, N. S., and George, A. J. (2000) Source of oncofetal
ED-B-containing fibronectin: implications of production by
both tumor and endothelial cells. Cancer Res. 60, 164–169
11. Tarli, L., Balza, E., Viti, F., Borsi, L., Castellani, P., Berndorff, D.,
Dinkelborg, L., Neri, D., and Zardi, L. (1999) A high-affinity
human antibody that targets tumoral blood vessels. Blood 94,
192–198
12. Demartis, S., Tarli, L., Borsi, L., Zardi, L., and Neri, D. (2001)
Selective targeting of tumour neovasculature by a radiohaloge-
nated human antibody fragment specific for the ED-B domain
of fibronectin. Eur. J. Nucl. Med. 28, 534–539
13. Halin, C., Rondini, S., Nilsson, F., Berndt, A., Kosmehl, H.,
Zardi, L., and Neri, D. (2002) Enhancement of the antitumor
activity of interleukin-12 by targeted delivery to neovasculature.
Nat. Biotechnol. 20, 264–269
14. Schliemann, C., Palumbo, A., Zuberbuhler, K., Villa, A., Kaspar,
M., Trachsel, E., Klapper, W., Menssen, H. D., and Neri, D.
(2009) Complete eradication of human B-cell lymphoma xeno-
grafts using rituximab in combination with the immunocytokine
L19-IL2. Blood 113, 2275–2283
15. Sauer, S., Erba, P. A., Petrini, M., Menrad, A., Giovannoni, L.,
Grana, C., Hirsch, B., Zardi, L., Paganelli, G., Mariani, G.,
Neri, D., Durkop, H., and Menssen, H. D. (2009) Expression
of the oncofetal ED-B-containing fibronectin isoform in
hematologic tumors enables ED-B-targeted 131I-L19SIP ra-
4543TARGETING TUMOR VESSELS BY THERAPEUTIC VACCINATION
dioimmunotherapy in Hodgkin lymphoma patients. Blood
113, 2265–2274
16. Workman, P., Balmain, A., Hickman, J. A., McNally, N. J., Rohas,
A. M., Mitchison, N. A., Pierrepoint, C. G., Raymond, R.,
Rowlatt, C., Stephens, T. C., and et al. (1988) UKCCCR guide-
lines for the welfare of animals in experimental neoplasia. Lab.
Anim. 22, 195–201
17. Tchougounova, E., Jiang, Y., Brasater, D., Lindberg, N.,
Kastemar, M., Asplund, A., Westermark, B., and Uhrbom, L.
(2009) Sox5 can suppress platelet-derived growth factor
B-induced glioma development in Ink4a-deficient mice
through induction of acute cellular senescence. Oncogene 28,
1537–1548
18. Magnusson, S. E., Pejler, G., Kleinau, S., and Abrink, M. (2009)
Mast cell chymase contributes to the antibody response and the
severity of autoimmune arthritis. FASEB J. 23, 875–882
19. Santimaria, M., Moscatelli, G., Viale, G. L., Giovannoni, L., Neri,
G., Viti, F., Leprini, A., Borsi, L., Castellani, P., Zardi, L., Neri,
D., and Riva, P. (2003) Immunoscintigraphic detection of the
ED-B domain of fibronectin, a marker of angiogenesis, in
patients with cancer. Clin. Cancer Res. 9, 571–579
20. Gundersen, H. J., Bendtsen, T. F., Korbo, L., Marcussen, N.,
Moller, A., Nielsen, K., Nyengaard, J. R., Pakkenberg, B.,
Sorensen, F. B., Vesterby, A., and West, M. J. (1988) Some new,
simple and efficient stereological methods and their use in
pathological research and diagnosis. Apmis 96, 379–394
21. Burton-Wurster, N., Lust, G., and Macleod, J. N. (1997) Carti-
lage fibronectin isoforms: in search of functions for a special
population of matrix glycoproteins. Matrix. Biol. 15, 441–454
22. Peters, J. H., Chen, G. E., and Hynes, R. O. (1996) Fibronectin
isoform distribution in the mouse. II. Differential distribution of
the alternatively spliced EIIIB, EIIIA, and V segments in the
adult mouse. Cell. Adhes. Commun. 4, 127–148
23. Carnemolla, B., Neri, D., Castellani, P., Leprini, A., Neri, G.,
Pini, A., Winter, G., and Zardi, L. (1996) Phage antibodies with
pan-species recognition of the oncofoetal angiogenesis marker
fibronectin ED-B domain. Int. J. Cancer 68, 397–405
24. Castellani, P., Borsi, L., Carnemolla, B., Biro, A., Dorcaratto, A.,
Viale, G. L., Neri, D., and Zardi, L. (2002) Differentiation
between high- and low-grade astrocytoma using a human recom-
binant antibody to the extra domain-B of fibronectin. Am. J.
Pathol. 161, 1695–1700
25. Ma, Y., HayGlass, K. T., Becker, A. B., Fan, Y., Yang, X., Basu, S.,
Srinivasan, G., Simons, F. E., Halayko, A. J., and Peng, Z. (2007)
Novel recombinant interleukin-13 peptide-based vaccine re-
duces airway allergic inflammatory responses in mice. Am. J.
Respir. Crit. Care. Med. 176, 439–445
26. Le Buanec, H., Paturance, S., Couillin, I., Schnyder-Candrian,
S., Larcier, P., Ryffel, B., Bizzini, B., Bensussan, A., Burny, A.,
Gallo, R., Zagury, D., and Peltre, G. (2007) Control of allergic
reactions in mice by an active anti-murine IL-4 immunization.
Vaccine 25, 7206–7216
27. Haller, B. K., Brave, A., Wallgard, E., Roswall, P., Sunkari, V. G.,
Mattson, U., Hallengard, D., Catrina, S. B., Hellstrom, M., and
Pietras, K. (2010) Therapeutic efficacy of a DNA vaccine target-
ing the endothelial tip cell antigen delta-like ligand 4 in
mammary carcinoma. [E-pub ahead of print] Oncogene doi:
10.1038/onc.2010.176
28. Birchler, M. T., Milisavlijevic, D., Pfaltz, M., Neri, D., Odermatt,
B., Schmid, S., and Stoeckli, S. J. (2003) Expression of the extra
domain B of fibronectin, a marker of angiogenesis, in head and
neck tumors. Laryngoscope 113, 1231–1237
29. Hauptmann, S., Zardi, L., Siri, A., Carnemolla, B., Borsi, L.,
Castellucci, M., Klosterhalfen, B., Hartung, P., Weis, J., Stocker,
G., Haubeck, H.-D., and Kirkpatrick, C. J. (1995) Extracellular
matrix proteins in colorectal carcinomas. Expression of tenascin
and fibronectin isoforms. Lab. Invest. 73, 172–182
30. Kaczmarek, J., Castellani, P., Nicolo, G., Spina, B., Allemanni,
G., and Zardi, L. (1994) Distribution of oncofetal fibronectin
isoforms in normal, hyperplastic and neoplastic human breast
tissues. Int. J. Cancer 59, 11–16
31. Ringvall, M., Huijbers, E. J., Ahooghalandari, P., Alekseeva, L.,
Andronova, T., Olsson, A. K., and Hellman, L. (2009) Identifi-
cation of potent biodegradable adjuvants that efficiently break
self-tolerance–a key issue in the development of therapeutic
vaccines. Vaccine 28, 48–52
Received for publication May 11, 2010.
Accepted for publication July 1, 2010.
4544 Vol. 24 November 2010 HUIJBERS ET AL.The FASEB Journal 䡠www.fasebj.org