The pro-inflammatory peptide LL-37 promotes
ovarian tumor progression through recruitment
of multipotent mesenchymal stromal cells
Seth B. Coffelta, Frank C. Marinib, Keri Watsonb, Kevin J. Zwezdarykc, Jennifer L. Dembinskib, Heather L. LaMarcad,
Suzanne L. Tomchuckc, Kerstin Honer zu Bentrupc, Elizabeth S. Dankac, Sarah L. Henklec, and Aline B. Scandurroc,1
aTumor Targeting Group, University of Sheffield School of Medicine, Sheffield, United Kingdom;bDepartment of Stem Cell Transplant and Cellular Therapy,
University of Texas M. D. Anderson Cancer Center, Houston, TX; andcDepartment of Microbiology and Immunology, Tulane University, New Orleans, LA;
anddDepartment of Molecular & Cellular Biology, Baylor College of Medicine, Houston, TX
Communicated by Darwin J. Prockop, Tulane University, New Orleans, LA, January 9, 2009 (received for review April 22, 2008)
Bone marrow-derived mesenchymal stem cells or multipotent
mesenchymal stromal cells (MSCs) have been shown to engraft
into the stroma of several tumor types, where they contribute to
tumor progression and metastasis. However, the chemotactic sig-
nals mediating MSC migration to tumors remain poorly under-
the C-terminal peptide of human cationic antimicrobial protein 18,
stimulates the migration of various cell types and is overexpressed
in ovarian, breast, and lung cancers. Although there is evidence to
support a pro-tumorigenic role for LL-37, the function of the
peptide in tumors remains unclear. Here, we demonstrate that
neutralization of LL-37 in vivo significantly reduces the engraft-
ment of MSCs into ovarian tumor xenografts, resulting in inhibi-
tion of tumor growth as well as disruption of the fibrovascular
network. Migration and invasion experiments conducted in vitro
indicated that the LL-37-mediated migration of MSCs to tumors
likely occurs through formyl peptide receptor like-1. To assess the
response of MSCs to the LL-37-rich tumor microenvironment,
conditioned medium from LL-37-treated MSCs was assessed and
found to contain increased levels of several cytokines and pro-
angiogenic factors compared with controls, including IL-1 receptor
Similarly, Matrigel mixed with LL-37, MSCs, or the combination of
the two resulted in a significant number of vascular channels in
nude mice. These data indicate that LL-37 facilitates ovarian tumor
progression through recruitment of progenitor cell populations to
serve as pro-angiogenic factor-expressing tumor stromal cells.
FPRL1 ? hCAP-18 ? mesenchymal stem cell ? ovarian cancer ?
(leucine, leucine-37) is up-regulated in these malignancies (1–3).
LL-37 was originally identified as a component of host defense
peptides released by innate immune cells to combat microor-
ganisms (4–6). However, recent investigations have revealed
more complex and diverse functions of the peptide (7–9). LL-37
is synthesized as the 37-aa C terminus of human cationic
antimicrobial protein 18 (hCAP-18) and maintained in an
inactive state until release by enzymatic cleavage (4, 5, 10–12).
Expression and secretion of LL-37 is elevated at sites of inflam-
mation and wound healing, where the peptide functions as a
proliferative signal and pro-angiogenic factor (7–9). The peptide
also acts as a potent chemoattractant for various immune cells
through activation of formyl peptide receptor like-1 (FPRL1)
(13, 14). In contrast to LL-37’s established functions in host
defense and tissue damage, the role of the peptide in the tumor
microenvironment and the advantage given to tumor cells by its
overexpression is not entirely clear.
The heterogeneous population of progenitor cells known as
multipotent stromal cells or mesenchymal stem cells (MSC) has
istological examination of ovarian, breast, and lung tumors
has shown that the pro-inflammatory peptide LL-37
been shown to engraft within tumor microenvironments, where
they incorporate into the stroma of solid tumors as tumor-
associated fibroblasts or pericyte-like cells and potentiate tumor
progression through the release of paracrine signals (15–25).
Kaposi sarcoma seems to be the only exception to this phenom-
enon, as MSCs inhibit the growth of this tumor type (23).
However, the tropism of MSCs for Kaposi sarcoma is main-
tained, suggesting that the factors responsible for MSC recruit-
ment to tumors are commonly secreted by multiple tumors of
different tissue origin. In fact, a number of soluble factors have
been implicated in MSC migration to tumors, including many of
the same inflammatory mediators up-regulated in injured and
inflamed tissues (17, 18, 20, 26–28). Therefore, given the similar
expression pattern of LL-37 in tumors, damaged tissue, and
inflammation, where MSCs are prominent, as well as the ability
of the peptide to stimulate chemotaxis of various cell types, we
the tumor microenvironment to support cancer progression.
LL-37 Promotes Migration and Invasion of MSC in Vitro.AsLL-37has
been shown to activate migration through the FPRL1 receptor
in various cell types, several donor pools of MSCs were examined
for expression of FPRL1 (13). Flow cytometry analyses con-
firmed the expression of FPRL1 on all MSC donor pools,
corroborating results from other laboratories (Fig. 1A) (29).
We extended our previous findings to determine the optimal
dosage of the peptide using in vitro chemotaxis assays (26). As
shown in Fig. 1B, LL-37 induced the migration of MSCs in a
dose-dependent manner, and the peptide performed as well as
EGF, an established chemotactic factor for these cells (17).
MSCs were pretreated with pertussis toxin (Ptx), a G?iinhibitor,
before activation with LL-37 or EGF to prevent FPRL1 signal-
ing. Ptx treatment followed by LL-37 stimulation resulted in a
significant inhibition of MSC migration, whereas no significant
difference was observed between EGF-stimulated, Ptx-treated
cells and EGF-stimulated cells alone (Fig. 1C). LL-37 and EGF
were also preincubated with a neutralizing LL-37 antibody and,
as expected, the neutralizing antibody (?LL-37) abolished LL-
37’s chemotactic effects on MSCs but had no effect on EGF-
stimulated cells (Fig. 1C). No decrease in MSC migration was
observed in wells with a control IgG antibody (data not shown).
MSC invasion through Matrigel-coated inserts was also sig-
Author contributions: S.B.C., F.C.M., K.J.Z., and A.B.S. designed research; S.B.C., F.C.M.,
K.W., K.J.Z., J.L.D., H.L.L., S.L.T., K.H.z.B., E.S.D., S.L.H., and A.B.S. performed research;
S.B.C., F.C.M., K.J.Z., H.L.L., S.L.T., and A.B.S. analyzed data; and S.B.C., F.C.M., and A.B.S.
wrote the paper.
The authors declare no conflict of interest.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
March 10, 2009 ?
vol. 106 ?
no. 10 www.pnas.org?cgi?doi?10.1073?pnas.0900244106
nificantly enhanced by LL-37 stimulation (Fig. 1D). Pretreat-
ment of MSC with Ptx significantly attenuated LL-37’s ability to
promote invasion (Fig. 1E). EGF-stimulated cells were slightly
affected by Ptx, but this was not significant. The anti-LL-37
antibody (?LL-37) significantly blocked LL-37 from binding to
MSC surface receptors, as MSCs in this experimental group did
not invade as effectively (Fig. 1E). By contrast, the anti-LL-37
antibody did not affect EGF stimulation of MSC invasion. An
IgG control antibody did not interfere with the ability of LL-37
or EGF to induce MSC invasion (data not shown). Taken
together, these data suggest that LL-37 induces MSC trafficking
through a G?i-coupled receptor, such as FPRL1.
As further validation of FPRL1 involvement in LL-37-
mediated responses, MSCs were assessed for activation of sig-
naling molecules downstream of this receptor. Western blot
analysis of MSC lysates showed that ERK-1 and -2 are robustly
phosphorylated beginning 10 min after LL-37 treatment and
maintained over 60 min (Fig. 1 F and G). However, MSCs
pretreated with Ptx before stimulation by LL-37 reduced ERK-
1/2 activation, providing further evidence in support of notion
that LL-37 stimulates MSCs through FPRL1.
The growth kinetics of MSCs were then analyzed after treat-
of serum. After a 72-hour time course, LL-37 failed to stimulate
the proliferation of MSCs (data not shown).
To assess the importance of LL-37 in recruitment of MSCs to
tumors, an in vivo migration assay was used. Human tumor
xenografts were established in SCID mice by injection of
OVCAR-3 ovarian cancer cells, whose expression of hCAP-18/
LL-37 was previously reported (1). After 3.5 weeks, mice were
randomly divided into 2 groups. One group (n ? 11) was given
an IgG control antibody and the other group (n ? 14) was given
an anti-LL-37 antibody to neutralize the peptide’s effect before
injection of firefly luciferase (ffLUC)-labeled MSCs. Biolumi-
nescence images were taken throughout the experiment to
monitor engraftment of MSCs into the developing ovarian
tumors. Neutralization of tumor-derived LL-37 in anti-LL-37
antibody-treated animals (?LL-37) abrogated the number of
engrafted MSCs into tumors compared with IgG-treated ani-
mals (Fig. 2A). Bioluminescence from engrafted MSCs was
quantified, and a significant difference was observed between
IgG-treated animals and anti-LL-37-treated animals for each
day an image was taken (Fig. 2B).
At the end of the experiment, tumors were analyzed via
immunohistochemical staining to confirm the reduced MSC
engraftment into tumors of anti-LL-37-treated animals. MSCs
were identified with an anti-ffLUC antibody, and their engraft-
ment into the tumor stroma as fibroblast-like cells was readily
observed (Fig. 2 C and D; white arrows). The difference in
overall MSC engraftment between IgG- and anti-LL-37-treated
was evident in these representative sections, validating biolumi-
nescence imaging. In addition to ffLUC, tumor sections were
co-stained for expression of LL-37. Tumor cells of IgG-treated
mice expressed measurable levels of hCAP-18/LL-37, whereas
expression of the peptide was dramatically reduced in tumors of
anti-LL-37-treated mice (Fig. 2C). Surprisingly, co-localization
of ffLUC and LL-37 was noted in tumor-infiltrating MSCs and,
when compared with tumor cells, MSCs expressed considerably
noted in perivascular regions neighboring vessels, indicative of
pericyte-like differentiation of these cells (Fig. 2F).
11 tumor-bearing animals that received the IgG control anti-
body, 10 animals had visible tumors upon sacrifice. By contrast,
5 of the 14 animals given the anti-LL-37 antibody treatment
regimen had no detectable tumor mass. Tumors from both
groups were weighed after surgical removal and those from
IgG-treated animals were significantly larger than those from
anti-LL-37-treated animals (98.43 ? 13.51 mg vs. 46.39 ? 12.65
were stained for the proliferation marker Ki-67. A dramatic
difference was again noted between tumors from IgG- and
anti-LL-37-treated animals. Tumor nuclei of IgG-treated mice
were almost all positive for Ki-67, in contrast to the few positive
nuclei in tumors from LL-37 antibody-treated mice (Fig. 3B).
Quantification of Ki-67?nuclei in both treatment groups re-
vealed a statistically significant difference (Fig. 3C). Notably, the
stromal components of tumors from anti-LL-37-treated mice,
including fibroblasts and endothelial cells, were absent in a large
majority of areas (Fig. 3B; yellow arrows). Necrotic regions were
in IgG-treated tumors. As shown in Fig. 3D, these necrotic areas
stained positively for LL-37, suggesting that the peptide is
- - -
* ** **
- - -
0 10 30 60 min LL-37 treatment M 10
Ptx - - - - +
min LL-37 treatment
coupled receptor. (A) FPRL1 expression on 3 different donor pools of MSCs
analyzed by flow cytometry. (B) Graphic representation of MSC migration
stimulated as indicated in a modified Boyden chamber. EGF and PMA were
used at 10 ng/mL. (C) MSC migration after pretreatment of cells with 100
ng/mL pertussis toxin (Ptx), or preincubation of LL-37 and EGF with an anti-
LL-37 neutralizing antibody (?LL-37). (D) Invasion of MSCs through Matrigel-
coated inserts following stimulation as indicated. (E) MSC invasion after
pretreatment of cells with Ptx or preincubation of LL-37 and EGF with ?LL-37
antibody.*, P ? 0.05;**, P ? 0.01. (F) Lysates from LL-37-treated MSCs
analyzed for ERK phosphorylation by Western blot. MSCs in the far right lane
molecular weight marker. (G) Quantification of Western blot band intensity
by densitometry (n ? 3), plotted as a bar graph.
LL-37 mediates MSC migration and invasion through a G protein-
Coffelt et al.PNAS ?
March 10, 2009 ?
vol. 106 ?
no. 10 ?
sequestered in the debris. Taken together, these data strongly
implicate a pro-tumorigenic role for tumor-derived LL-37
through its recruitment of progenitor cell populations capable of
differentiating into supportive stromal cells.
MSC Exposure to LL-37 Promotes Secretion of Angiogenic and Inflam-
matory Molecules. It is well established that MSCs produce many
trophic factors with pro-tumorigenic functions (19, 26, 30).
Evidence from the in vivo experiments described here indicated
that MSCs also produce LL-37. We tested conditioned medium
from several MSC donor pools growing in culture by ELISA and
found that these cells readily secrete the peptide (Fig. 4A).
Next, we set out to determine how tumor-infiltrating MSCs
would react to the LL-37-rich microenvironment of ovarian
tumors. We have previously reported that LL-37 enhances the
secretion of IL-1?, IL-6, IL-8, IL-10, and TNF-? from MSC
while diminishing the secretion of IL-12 (p70) (26). To identify
additional MSC-derived cytokines and growth factors that may
be regulated by ovarian tumor-derived LL-37 and expand our
previous findings, conditioned medium from various LL-37-
After 48 h of LL-37 treatment, MSCs were stimulated to release
significantly more of the following cytokines compared with
untreated cells: IL-1 receptor antagonist, IL-6, IL-10, CCL5
(regulated upon activation, normal T cell expressed and se-
creted; RANTES), and VEGF (Fig. 4B).
Conditioned medium was also analyzed for the presence and
activation of matrix metalloproteinases (MMPs) by zymography
assays. Untreated MSCs secreted large amounts of the MMP-2
MMP-2 pro-form, regardless of treatment (Fig. 4C Upper).
However, enzymatic activity of the active form of MMP-2 was
increased after treatment of MSCs with LL-37, EGF, and
phorbol myristate acetate (PMA; Fig. 4 C and D Lower). MMP-9
activity was undetectable after any treatment, and no expression
was noted in casein gels, indicating that MSCs do not secrete
measurable levels of stromelysins such as MMP-3 (data not
We tested whether conditioned medium from LL-37-treated
MSCs could increase endothelial cell tubule formation in vitro.
Serum-starved human umbilical vein endothelial cells
(HUVECs) were seeded onto growth factor-reduced Matrigel in
the presence of MSC-conditioned medium. As shown in Fig. 4E,
all 3 donor pools of LL-37-treated MSC-conditioned medium
stimulated HUVECs to form tubules at a faster rate than
medium from untreated MSCs. HUVECs exposed to medium
from LL-37-treated MSCs began to migrate and organize into
tube-like structures after only 2 hours. These data not only
confirmed that LL-37-treated MSC-conditioned medium con-
LL-37 ffLUC DAPI
LL-37 ffLUC DAPI
cantly reduces engraftment of
MSCs into ovarian tumors. Human
ovarian tumor xenografts were es-
tablished i.p. in SCID mice. Mice
were treated with IgG or anti-LL-37
antibodies (?LL-37) twice a week
for 4 weeks. ffLUC-labeled MSCs
were injected 4 times at weekly in-
tervals 1 day after the first weekly
injection of antibody, then visual-
ized by bioluminescence in live
mice. (A) Representative images of
MSC engraftment into ovarian tu-
mors 7 days after each injection of
MSC. (B) Quantification of lumines-
cence units emanating from tumor-
engrafted MSCs. Values are mean ?
(red) in ovarian tumor sections of
IgG- and ?LL-37-treated mice. MSCs
were identified using an anti-ffLUC
antibody (green) and are indicated
by white arrows. Nuclei were de-
tected with DAPI. Sections are mag-
antibodies followed by hematoxylin counterstain. (F) Example of LL-37-expressing MSCs in perivascular areas. (Scale bar, 50 ?m, D–F.)
Inhibition of LL-37 signifi-
control Ki-67 LL-37
of LL-37. (A) Graphic representation of tumor weights from IgG- (n ? 10) and
?LL-37-treated (n ? 9) animals obtained after surgical removal. Values are
with hematoxylin counterstain. Arrows indicate mouse stroma in human
xenograft tumors. (Scale bar, 50 ?m.) (C) Graphic representation of the
average number of Ki-67?nuclei per high-powered field. Values are mean ?
an ?LL-37-treated mouse. (Scale bar, 100 ?m.)
Growth of ovarian tumor xenografts is diminished by neutralization
www.pnas.org?cgi?doi?10.1073?pnas.0900244106Coffelt et al.
tains increased amounts of pro-angiogenic molecules, they also
confirmed that the pro-angiogenic milieu is functional and has
LL-37 Enhances the Pro-Angiogenic Activity of MSC. Because MSCs
secreted larger amounts of pro-angiogenic factors in response to
LL-37, we hypothesized that tumor-associated MSCs are stim-
ulated in the same manner by a microenvironment abundant in
LL-37. To this end, serum-starved MSCs were mixed into
Matrigel with and without growth factors, and then injected into
the flanks of nude mice. LL-37, as well as the combination of
basic FGF2 and VEGF, were added to Matrigel without cells as
controls. Matrigel plugs containing growth factors, but no cells,
induced a significant number of vascular channels compared
with plugs without growth factors (Fig. 5 A and B). Surprisingly,
the stimulation of angiogenesis by LL-37 exceeded that of the
FGF2/VEGF combination at the concentrations used here.
MSCs, on their own, induced a similar response as FGF2/VEGF.
The numbers of vascular channels in Matrigel plugs with MSC
and LL-37 were significantly greater than those containing
MSCs alone or MSCs and FGF2/VEGF. However, there was no
synergistic or additive effect on angiogenesis in plugs containing
MSCs and growth factors, as we had anticipated. MSCs with
either LL-37 or FGF2/VEGF did not significantly increase the
number of vascular channels compared with plugs without cells.
as well as murine endothelial cells.
As shown in Fig. 5C, by using Ki-67 to identify human cells,
MSCs were observed around endothelial cells, but did not
incorporate into the vasculature, suggesting pericyte-like differ-
entiation and corroborating our previous reports (18, 27). We
investigated LL-37’s effect on the pericyte-like differentiation of
MSCs by tubule formation assay. Serum-starved MSCs were
seeded onto Matrigel and treated with LL-37 or FGF2 or left
untreated. LL-37 stimulated MSCs to form organized, capillary-
like structures in the same manner as FGF2 treatment, whereas
MSCs without growth factor influence remained as single-cell
entities, indicating that LL-37 may be involved in differentiation
of MSCs into pericytes in tumors (Fig. 5D).
Emerging evidence indicates that inflammatory molecules play
a pivotal role in tumor progression; however, the function of
many of these tumor-derived inflammatory mediators remains
poorly understood (31). Herein, we demonstrate that LL-37
promotes ovarian tumor progression through recruitment and
engraftment of MSCs into tumors, where these cells provide
pro-angiogenic and immunomodulatory factors to support tu-
mor growth and progression.
This study directly tested the importance of a tumor-derived
MSC chemotactic factor by using an in vivo assay. Previous
reports from our laboratory and others suggest that additional
chemokines, growth factors, and danger signals may be involved
in MSC migration to tumors (17, 18, 20, 26, 27). Thus, it is
unlikely that LL-37 is acting alone in the recruitment of MSCs
to the tumor microenvironment given the vast production of
chemotactic factors by tumor and stromal cells. In support of this
notion, neutralization of LL-37 in tumor-bearing animals did not
completely block MSC engraftment. Additionally, few chemo-
kine receptors have been implicated in this process. Our recent
data suggest that CCR2—the receptor for CCL2 (also called
monocyte chemotactic protein 1)—is involved, as incubation of
murine MSCs with anti-CCR2 blocking antibodies decreases
migration of these cells (18). The data presented here indicate
that LL-37 mediates MSC migration through FPRL1.
maintained a fibroblast appearance or differentiated into pericyte-
like cells—observations consistent with our previous reports and
those of others (15–20, 23, 32–34). However, the inhibition of MSC
engraftment into tumors resulted in disorganization of the fibro-
vascular network, strongly suggesting a critical role for LL-37 in
recruitment of progenitor populations that serve as carcinoma-
associated fibroblasts and pericytes. Our data suggest that MSCs,
factors that initiate angiogenesis and/or differentiate into blood
vessel-supporting cells. Although evidence of this pro-angiogenic
levels, such as IL-10 and CCL5, in response to LL-37 was not
an innate ability to modulate the function of various immune cell
Donor 1Donor 2Donor 3
untreated MSC LL-37-treated MSC
and inflammatory mediators after LL-37 stimulation.
(A) The concentration of LL-37 in conditioned medium
taken from unstimulated MSCs in culture. (B) Serum-
then conditioned medium was analyzed by Luminex-
based cytokine arrays. Values are mean ? SE. *, P ?
0.05, **, P ? 0.01. (C) Analysis of MSC-conditioned
medium after treatment for 48 h as indicated by gel-
atin zymography. The representative image depicts
the electrophorectic pro-MMP-2 (72 kDa) and active
MMP-2 (62 kDa). MMP-9 (92 kDa) was not undetect-
able. (D) Quantification of zymography by densitom-
etry. Intensity of the lower band (active MMP-2, 62
kDa) is plotted as a bar graph (n ? 3). (E) Conditioned
medium from untreated and LL-37-treated MSCs was
cently labeled cells were monitored were for forma-
tion of capillary-like tubules. Photographs are repre-
sentative of HUVECs after 2 h incubation.
MSCs secrete increased levels of angiogenic
Coffelt et al.PNAS ?
March 10, 2009 ?
vol. 106 ?
no. 10 ?
populations via these factors (35, 36). Another effect of LL-37 on
MSCs was the elevated secretion of CCL5. MSCs have been shown
to promote breast cancer metastases through release of this cyto-
kine, raising the possibility that LL-37 may be the instigator in this
Unexpectedly, staining for LL-37 in ovarian tumor xenograft
sections proved a better identifier of MSCs than ffLUC. As
noted in Fig. 2, LL-37 expression was much higher in MSCs than
ovarian cancer cells, indicating that both tumor cells and MSCs
contribute the peptide to the cytokine milieu of the tumor. This
increase in total LL-37 levels may then lead to recruitment of
more MSCs perpetuating the progression of the tumor. Perivas-
cular MSCs may also influence endothelial cells through secre-
tion of LL-37, as the peptide’s effects on endothelial cells has
already been established through FPRL1 activation (7).
The results presented here and in previous reports suggest the
following sequence of events summarized in Fig. 6: (i) genetic
alterations in ovarian surface epithelial cells or other tissues
elevates expression of LL-37, a peptide whose expression is low
or absent in normal cells (1); (ii) LL-37 feeds back on the tumor
epithelium, stimulating epithelial cell proliferation (1–3); (iii)
LL-37 activates the MSC population to migrate into the tumor
mass and enhances MSC secretion of pro-angiogenic factors; (iv)
MSCs provide additional LL-37 to the tumor microenvironment;
(v) LL-37 as well as other tumor- and MSC-derived molecules
induce angiogenesis as the tumor expands; (vi) MSC produce
immunosuppressive factors, likely in response to the LL-37-rich
microenvironment, that function to dampen anti-tumor immu-
nity; and (vii) cytokines such as CCL5 released by LL-37-
stimulated MSCs enable an invasive phenotype from tumor cells.
The overall consequence of LL-37’s actions through its recruit-
ment of MSCs is advancement of tumor progression.
Materials and Methods
More detailed methods are presented in SI Materials and Methods.
Cell Culture. Human MSC were obtained from Tulane University’s Center for
Gene Therapy (New Orleans, LA) and Lonza/Cambrex (Walkersville, MD). The
cells were characterized by flow cytometry and various differentiation assays,
and the cells were propagated as described (26, 27). MSC were used at
passages no greater than 5. OVCAR-3 ovarian cancer cells were cultured as
described in ref. 1.
Alexa-488-conjugated goat anti-rabbit secondary antibodies.
Boyden Chamber Migration Assay. Chemoattractants with and without an
anti-LL-37 neutralizing antibody was added to the lower compartment of a
48-well modified Boyden chamber. Serum-starved MSC were added to the
upper chamber. Where indicated, MSC were pretreated with 100 ng/mL
pertussis toxin (Ptx).
Invasion Assay. The assay was performed in a similar manner to migration
assays using inserts coated with growth factor-reduced Matrigel.
Western Blot Analysis. LL-37-treated MSC lysates were electrophoresed and
transferred to nitrocellulose membranes. Quantification of band intensity
was performed using National Institutes of Health ImageJ software.
In Vivo Migration Assay. Female SCID/CB17 mice were injected i.p. with
OVCAR-3 ovarian cancer cells and tumors were allowed to establish for 3.5
weeks. After this time, half the mice were treated with 50 ?g of nonspecific
twice a week for the duration of the experiment. MSC were infected with 500
viral particles per cell of Ad-ffLUC-RDG, 24 h before injection. D-Luciferin was
used to detect MSC, 7 d after MSC injections. Bioluminescent images were
acquired from anesthetized mice with the IVIS-Xenogen system (Caliper Life
of interest using Living Image software.
empty FGF2/VEGF LL-37
avg. number of vascular
untreated LL-37 FGF2
combination of FGF and VEGFA was added to cold Matrigel with or without
MSCs and injected into nude mice (n ? 6). The absence of growth factors and
cells served as negative control. After 7 to 10 days, Matrigel plugs were
surgically removed, fixed, sectioned, and stained by H&E. Representative
images of vascular channels are shown. (Scale bar, 100 ?m.) (B) The average
3 high-powered fields of view then graphically represented. Values are
mean ? SE. *, P ? 0.05, **, P ? 0.01, ***, P ? 0.001. (C) Example of MSCs in
perivascular areas identified by Ki-67 staining. (D) Fluorescently labeled,
serum-starved MSCs were seeded onto Matrigel in the presence of 5 ?g/mL
LL-37 or 10 ng/mL FGF2 and allowed to incubate overnight. Formation of
tubules, indicative of their pericyte-like differentiation, was captured by
microscopy at ?200 the next day.
LL-37 enhances the pro-angiogenic activity of MSCs. (A) LL-37 or the
progression (see Discussion).
Schematic illustration of effects of LL-37 and MSCs on ovarian tumor
www.pnas.org?cgi?doi?10.1073?pnas.0900244106Coffelt et al.
Histology and IHC. Tumors were fixed in formalin solution and embedded in Download full-text
paraffin. Sections were stained with hematoxylin and eosin. Immunofluores-
cence staining for LL-37 and firefly luciferase (ff-LUC) was performed using
Alexa-488- and Alexa-568-conjugated antibodies, respectively. IHC was per-
formed as previously described using Dako’s Animal Research Kit (1). Images
were evaluated using a Zeiss Axioplan 2 fluorescence microscope and Intelli-
gent Innovations software (SlideBook version 4).
Analysis of MSC-Secreted Soluble Factors. MSC-conditioned medium was
analyzed by a luminex-based assay. Zymography assays were performed as
before (1). Quantification of band intensity was performed using ImageJ
Tubule Formation Assay. HUVECs were resuspended in MSC-conditioned me-
dium and seeded onto Matrigel then fluorescently labeled. For MSC differ-
entiation on Matrigel, cells were treated with LL-37 or FGF2.
Matrigel Plug Assay. Female BALB/c nude mice (n ? 14) were used as
described in ref. 27. Matrigel was loaded with LL-37 (final volume ? 5
?g/mL), FGF2/VEGF (final volume ? 10 and 25 ng/mL, respectively), and
2.5 ? 105MSC when appropriate. Vascular channels were identified from
H&E stained sections and the average number from 3 separate 400? fields
was calculated for each plug.
Keuls post hoc test was used for P values.
ACKNOWLEDGMENTS. We thank Dr. Jeff Rosen and his team for their accom-
modation and Dr. Cindy Morris for the HUVECs. This work was supported in
part by National Institutes of Health grants 1P20RR20152–01 (to A.B.S.) and
CA-1094551, CA-116199, and CA49639 (to F.C.M.); Susan G. Komen Breast
Cancer Foundation grant BCTR0504372 (to K.W., J.L.D., and F.C.M.), and the
W. M. Keck Foundation.
and its derivative LL-37 increases ovarian cancer cell proliferation and invasion. Int J
2. Heilborn JD, et al. (2005) Antimicrobial protein hCAP18/LL-37 is highly expressed in
breast cancer and is a putative growth factor for epithelial cells. Int J Cancer 114:713–
3. von Haussen J, et al. (2008) The host defence peptide LL-37/hCAP-18 is a growth factor
for lung cancer cells. Lung Cancer 59:12–23.
and expressed in bone marrow and testis. Proc Natl Acad Sci USA 92:195–199.
5. Larrick JW, et al. (1995) Human CAP18: A novel antimicrobial lipopolysaccharide-
binding protein. Infect Immun 63:1291–1297.
peptides. Immunotechnology 1:65–72.
7. Koczulla R, et al. (2003) An angiogenic role for the human peptide antibiotic LL-37/
hCAP-18. J Clin Invest 111:1665–1672.
8. Shaykhiev R, et al. (2005) Human endogenous antibiotic LL-37 stimulates airway
epithelial cell proliferation and wound closure. Am J Physiol 289:L842–L848.
9. Heilborn JD, et al. (2003) The cathelicidin anti-microbial peptide LL-37 is involved in
re-epithelialization of human skin wounds and is lacking in chronic ulcer epithelium.
J Invest Dermatol 120:379–389.
10. Cowland JB, Johnsen, AH, Borregaard, N (1995) hCAP-18, a cathelin/pro-bactenecin-
like protein of human neutrophil specific granules. FEBS Lett 368:173–176.
11. Sorensen OE, et al. (2001) Human cathelicidin, hCAP-18, is processed to the antimicro-
bial peptide LL-37 by extracellular cleavage with proteinase 3. Blood 97:3951–3959.
A novel mechanism of generating antimicrobial peptides in vagina. J Biol Chem
13. Yang D, et al. (2000) LL-37, the neutrophil granule- and epithelial cell-derived cathe-
licidin, utilizes formyl peptide receptor-like 1 (FPRL1) as a receptor to chemoattract
human peripheral blood neutrophils, monocytes, and T cells. J Exp Med 192:1069–
alpha-defensins are expressed by specific lymphocyte and monocyte populations.
and targeted-delivery vehicles for anticancer agents. J Natl Cancer Inst 96:1593–1603.
16. Studeny M, et al. (2002) Bone marrow-derived mesenchymal stem cells as vehicles for
intereron-beta delivery into tumors. Cancer Res 62:3603–3608.
17. Nakamizo A, et al. (2005) Human bone marrow-derived mesenchymal stem cells in the
treatment of gliomas. Cancer Res 65:3307–3318.
18. Klopp AH, et al. (2007) Tumor irradiation increases the recruitment of circulating
mesenchymal stem cells into the tumor microenvironment. Cancer Res 67:11687–
19. Karnoub AE, et al. (2007) Mesenchymal stem cells within tumour stroma promote
breast cancer metastasis. Nature 449:557–563.
20. Dwyer RM, et al. (2007) Monocyte chemotactic protein-1 secreted by primary breast
21. Komarova S, Kawakami Y, Stoff-Khalili MA, Curiel DT, Pereboeva L (2006) Mesenchy-
mal progenitor cells as cellular vehicles for delivery of oncolytic adenoviruses. Mol
Cancer Ther 5:755–766.
effects in a model of Kaposi’s sarcoma. J Exp Med 203:1235–1247.
24. Ramasamy R, et al. (2007) Mesenchymal stem cells inhibit proliferation and apoptosis
of tumor cells: Impact on in vivo tumor growth. Leukemia 21:304–310.
25. Djouad F, et al. (2003)Immunosuppressive effect of mesenchymal stem cells favors
tumor growth in allogeneic animals. Blood 102:3837–3844.
26. Tomchuck SL, et al. (2008) Toll-like receptors on human mesenchymal stem cells drive
their migration and immunomodulating responses. Stem Cells 26:99–107.
27. Zwezdaryk KJ, et al. (2007) Erythropoietin, a hypoxia-regulated factor, elicits a pro-
angiogenic program in human mesenchymal stem cells. Exp Hematol 35:640–652.
28. Spaeth E, Klopp A, Dembinski J, Andreeff M, Marini F (2008) Inflammation and tumor
29. Viswanathan A, Painter RG, Lanson NA, Jr, Wang, G (2007) Functional expression of
N-formyl peptide receptors in human bone marrow-derived mesenchymal stem cells.
Stem Cells 25:1263–1269.
30. Uccelli A, Moretta, L, Pistoia, V (2008) Mesenchymal stem cells in health and disease.
Nat Rev Immunol, in press.
31. Coffelt SB, Scandurro, AB (2008) Tumors sound the alarmin(s). Cancer Res 68:6482–
facilitate engineering of long-lasting functional vasculature. Blood 111:4551–4558.
33. Covas DT, et al. (2008) Multipotent mesenchymal stromal cells obtained from diverse
human tissues share functional properties and gene-expression profile with CD146?
perivascular cells and fibroblasts. Exp Hematol 36:642–654.
34. Mishra PJ, et al. (2008) Carcinoma-associated fibroblast-like differentiation of human
mesenchymal stem cells. Cancer Res 68:4331–4339.
35. Beyth S, et al. (2005) Human mesenchymal stem cells alter antigen-presenting cell
maturation and induce T-cell unresponsiveness. Blood 105:2214–2219.
36. Ren G, et al. (2008) Mesenchymal stem cell-mediated immunosuppression occurs via
concerted action of chemokines and nitric oxide. Cell Stem Cell 2:141–150.
37. Yotnda P, et al. (2004) Comparison of the efficiency of transduction of leukemic cells
by fiber-modified adenoviruses. Hum Gene Ther 15:1229–1242.
Coffelt et al.PNAS ?
March 10, 2009 ?
vol. 106 ?
no. 10 ?