Vascular normalizing doses of antiangiogenic
treatment reprogram the immunosuppressive tumor
microenvironment and enhance immunotherapy
Yuhui Huanga, Jianping Yuanb, Elda Righib, Walid S. Kamouna, Marek Ancukiewicza, Jean Nezivarb,
Michael Santosuossob, John D. Martina, Margaret R. Martina, Fabrizio Vianellob, Pierre Leblancb, Lance L. Munna,
Peigen Huanga, Dan G. Dudaa, Dai Fukumuraa, Rakesh K. Jaina,1, and Mark C. Poznanskyb
aEdwin L. Steele Laboratory of Tumor Biology, Department of Radiation Oncology andbVaccine and Immunotherapy Center, Infectious Diseases Medicine,
Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114
Contributed by Rakesh K. Jain, September 8, 2012 (sent for review July 16, 2012)
The recent approval of a prostate cancer vaccine has renewed hope
for anticancer immunotherapies. However, the immunosuppressive
tumor microenvironment may limit the effectiveness of current
immunotherapies. Antiangiogenic agents have the potential to
modulate the tumor microenvironment and improve immunother-
vessels and paradoxically may compromise various therapies. Here,
we demonstrate that targeting tumor vasculature with lower
vascular-normalizing doses, but not high antivascular/antiangio-
genic doses, of an anti-VEGF receptor 2 (VEGFR2) antibody results
in a more homogeneous distribution of functional tumor vessels.
tumor-associated macrophages from an immune inhibitory M2-like
facilitating CD4+and CD8+T-cell tumor infiltration. Based on this
mechanism, scheduling lower-dose anti-VEGFR2 therapy with T-cell
activation induced by a whole cancer cell vaccine therapy enhanced
anticancer efficacy in a CD8+T-cell–dependent manner in both im-
mune-tolerant and immunogenic murine breast cancer models.
These findings indicate that vascular-normalizing lower doses of
anti-VEGFR2 antibody can reprogram the tumor microenvironment
away from immunosuppression toward potentiation of cancer vac-
cine therapies. Given that the combinations of high doses of bevaci-
zumab with chemotherapy have not improved overall survival of
breast cancer patients, our study suggests a strategy to use antian-
giogenic agents in breast cancer more effectively with active immu-
notherapy and potentially other anticancer therapies.
vascular normalization|hypoxia|myeloid-derived suppressor cell|
tumor tissue vaccine
istration approved the first cancer vaccine (Provenge) to treat
advanced prostate cancer in 2010 (1). However, tumor control by
vaccination alone is minimal, and survival benefits remain modest.
Thus, new approaches that improve the clinical benefits of anti-
cancer vaccines are urgently needed.
The induction of high numbers of tumor-specific cytotoxic T
Unfortunately, the presence of a high number of tumor antigen-
specific cytotoxic T cells in peripheral immune organs often is not
associated with clinical benefit (2, 3). Thus, other factors are likely
involved in this poor clinical outcome. Among these, the tumor
microenvironment is now considered a key player (4–8).
Emerging data indicate that abnormal tumor vasculature, result-
ingfromthe prevalence of pro- versus antiangiogenic signals,fosters
an immunosuppressive tumor microenvironment that enables the
tumor to evade host immunosurveillance (4, 6, 9). Proangiogenic
factors not only suppress the function of various immune cells (10)
but also diminish leukocyte–endothelial interactions and hinder the
Clinical studies consistently support the view that malignant tumors
fter decades of research on harnessing the power of the im-
mune system to fight cancer, the US Food and Drug Admin-
are nonpermissive to T effector cell accumulation (3, 7, 12). In
suppressors, such as tumor-associated macrophages (TAMs), my-
eloid-derived suppressor cells (MDSCs), and regulatory T cells
(Tregs) (13–15). Importantly, TAM and Treg accumulation within
tumors correlates with poor prognosis (12–14, 16). Within breast
cancer lesions, TAMs represent the dominant myeloid cell pop-
associated with the hypoxic tumor microenvironment (14, 17, 18).
These immune-evasion mechanisms could be altered by anti-
angiogenic treatment. Rationally scheduled antiangiogenic treat-
ment can transiently normalize tumor vessels, improve vessel
perfusion, decrease hypoxia, and enhance cytotoxic therapies (4,
19–21). In genetic studies, vascular normalization by deletion of
Rgs5 increased T-cell infiltration into tumors and substantially
preclinical studies have suggested that antiangiogenic therapy
could increase tumor-infiltrating T cells (22–25). However, no
antiangiogenic agent has been shown to improve breast cancer
vaccine therapy in a clinically relevant model of immune-tolerant
breast cancer (23). Here, we evaluate the effects of treatment with
different doses of an anti-VEGF receptor 2 (VEGFR2) antibody
(DC101) and establish a combinational regimen that synchronizes
T-cell activation with breast cancer vascular normalization. In
models of both immune-tolerant and immunogenic breast cancer,
we show that lower doses, but not high dose, of DC101 can re-
program the immunosuppressive tumor microenvironment in
a manner that augments anticancer vaccine therapy.
Lower Doses of Anti-VEGFR2 Antibody Treatment Enhance Vaccine
Therapy in a Model of MCaP0008 Breast Cancer. To test the dose-
dependent effect of antiangiogenic treatment on cancer vaccine
therapy in a clinically relevant breast cancer model, we vaccinated
C-pretreated MCaP0008 cancer cell vaccine, following different
doses of DC101 treatment (Fig. 1A). Although vaccination stimu-
Author contributions: Y.H., J.Y., E.R., M.S., F.V., D.G.D., D.F., R.K.J., and M.C.P. designed
research; Y.H., J.Y., E.R., J.N., M.S., and P.L. performed research; Y.H., W.S.K., J.D.M.,
M.R.M., L.L.M., P.H., D.G.D., D.F., R.K.J., and M.C.P. contributed new reagents/analytic
tools; Y.H., J.Y., E.R., W.S.K., M.A., M.S., J.D.M., F.V., D.G.D., D.F., R.K.J., and M.C.P. ana-
lyzed data; and Y.H., D.G.D., D.F., R.K.J., and M.C.P. wrote the paper.
Conflict of interest statement: R.K.J. received research grants from Dyax, MedImmune,
and Roche; consultant fees from Dyax, Enlight, Noxxon, and SynDevRx; owns equity in
Enlight, SynDevRx, and XTuit; and serves on the Board of Directors of XTuit and Boards of
Trustees of H&Q Healthcare Investors and H&Q Life Sciences Investors. M.C.P. serves as
a scientific adviser to Evaxion-Biotech and owns equity in Celtaxsys. No reagents or fund-
ing from these companies was used in this study. Therefore, there is no significant finan-
cial or other competing interest in the work.
Freely available online through the PNAS open access option.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| October 23, 2012
| vol. 109
| no. 43
lated IFN-γ production in splenic CD8+T cells (Fig. S1), vaccina-
tion alone did not inhibit tumor growth, suggesting the immune
tolerance of MCaP0008 breast cancer (Fig. 1B). Consistent with
previous preclinical studies (26), monotherapy with a standard
antivascular/antiangiogenic high-dose DC101 (40 mg/kg, hereafter
referred as “full-dose”), but not lower-dose DC101 (10 mg/kg,
hereafter referred to as “quarter-dose”), significantly delayed tu-
mor growth (Fig. 1B). However, the combination of full-dose
tumor growth compared with full-dose DC101 monotherapy. Re-
markably, tumor growth was inhibited significantly by quarter-dose
DC101 combined with vaccine treatment as compared with quar-
ter-dose DC101 alone (P = 0.038) (Fig. 1B). In vivo depletion of
CD8+T cells abrogated the improvement seen with the combina-
tion of quarter-dose DC101 and vaccine treatment, indicating that
T-cell–dependent manner (Fig. 1C). Similarly, the combination of
half-dose DC101 (20 mg/kg) and vaccine therapy significantly re-
duced tumor volume as compared with half-dose DC101 mono-
therapy (P = 0.031) (Fig. 1D). These data show that lower-, but not
high-dose, anti-VEGFR2 antibody treatment enhances the anti-
cancer efficacy of a vaccine therapy in a model of immune-tolerant
Lower-Dose Anti-VEGFR2 Antibody Treatment Normalizes Breast
Tumor Vasculature and Improves Overall Tissue Perfusion. Abnor-
mal tumor vasculature and the immunosuppressive tumor micro-
environment are two major barriers for cancer vaccine therapy (6,
8, 9, 16). We hypothesized that lower-dose antiangiogenic treat-
ment, in contrast to a high dose, alleviates the immunosuppressive
tumor microenvironment via vascular normalization. Here, we
adapted a previously established procedure to label tumor areas
proximal to perfused vessels with Hoechst 33342. Thus, Hoechst
fluorescence positivity is inversely related to hypoxia status (27).
We analyzed perfusion in entire cross-sections of tumor tissue
(Fig. 2A). Quantitative analysis revealed that a significantly higher
and quarter-dose, but not full-dose, DC101 as compared with IgG
control (Fig. 2B). Consistently, quarter-dose DC101 treatment
significantly reduced the hypoxic area as compared with both IgG
control and full-dose DC101 treatment (Fig. 2C and Fig. S2C). In
addition, we found a remarkable change in the distribution of
with low-dose DC101. In IgG control tumors, perfused vessels
were distributed unevenly throughout the tumor. Some areas of
tumor tissue were well perfused, but other areas had little or no
DC101 treatments, perfused vessels were distributed more evenly
throughout the tumor section. [The variances are significantly
different in IgG vs. quarter-dose (P = 0.027) and IgG vs. half-dose
(P = 0.026), exact Wilcoxon test; areas under relative operating
characteristic (ROC) curves are 0.76 and 0.77, respectively (Fig. 2
A and D).] On the other hand, there was no significant improve-
ment in functional vessel distribution with full-dose DC101 treat-
ment (Fig. 2 A and D). These data indicate that only lower-dose
DC101 treatment results in a more homogeneous distribution of
functional blood vessels in tumors.
To dissectfurther the mechanismsofimprovedtumorperfusion
during lower-dose DC101 treatment, we evaluated pericyte cov-
and tumor vessel perfusion by labeling functional vessels with i.v.-
injected FITC-lectin. Consistently, half-dose DC101 treatment
significantly increased NG2-positive pericyte coverage (Fig. 2E)
and the proportion of functional vessels labeled by FITC-lectin
in comparison with IgG control (Fig. 2F). Together, these data
demonstrate that lower-dose DC101 treatment normalizes breast
tumor vasculature and improves tissue distribution of functional
blood vessels in breast cancers.
Lower-Dose Anti-VEGFR2 Antibody Treatment Polarizes the Number,
Distribution, and Phenotype of Tumor-Infiltrating Myeloid Cells from
Immunosuppressive to Immunostimulatory. In breast cancer, the
abundant TAMs and MDSCs suppress anticancer immunity and
promote tumor progression (5, 8, 14, 16). Considering the positive
effects of lower-dose DC101 on vaccine therapy, we determined
whether lower-dose DC101 treatment could modulate the tumor
infiltration, distribution, and phenotype of immune-suppressive
myeloid cells. We found that lower-dose, but not full-dose, DC101
treatment significantly reduced the fraction of CD45+CD11b+
Gr1+F4/80−cells (CD11b+Gr1+or MDSC) in total viable cells in
MCaP0008 breast cancer tissues (Fig. 3A and Fig. S3A). Concur-
rently, lower-dose DC101 treatment significantly increased the
percentage of CD45+CD11b+Gr1−F4/80+TAMs in total viable
cells as compared with IgG control or full-dose DC101 treatment
(Fig. 3A and Fig. S3A). To differentiate TAMs proximal and distal
to blood vessels, we injected Hoechst 33342 dye i.v. before tissue
harvest. TAMs were separated by flow cytometry as Hoechst
33342-positive TAMs (Ho+TAMs) proximal to perfused vessels
vessels. We found that full-dose DC101 treatment significantly
decreased the percentage of Ho+TAMs among total TAMs in
cine therapy in a model of MCaP0008 breast cancer.
(A) Treatment protocol. Seven days after implanta-
tion of an MCaP0008 breast tumor, mice were di-
vided randomly into two groups and injected i.p.
with 5 × 106CD45−, mitomycin C-treated MCaP0008
tumor tissue cells or with an equal volume of PBS at
four time points. These mice subsequently were
treated with four doses of DC101 [10 or 40 mg/kg
body weight (bw)] at 3-d intervals or with IgG (40 mg/
kg bw) 1 d before the fourth vaccination. Mice in the
in vivo CD8 depletion study were treated with anti-
CD8a or 2A3 (isotype rat IgG2a, 200 μg per mouse) on
days −1 (1 d before the first vaccination), 1, 7, and 13.
(B) Tumor growth curves. Tumor size was measured
every 3 d starting at day 7 after the first vaccination
(the first day of DC101 treatment). *P < 0.05, PBS/
DC101-10 vs. vaccine/DC101-10. n = 10 mice per
group. (C) Depletion of CD8 T-cells abrogated the
improvement of quarter-dose DC101 treatment on
vaccine therapy. **P < 0.01,vaccine/DC101-10 vs.
vaccine/DC101-10/anti-CD8. The IgG2a group had six mice; all other groups had 10 mice. (D) Tumor growth curves. MCaP0008 tumor-bearing mice were treated
with rat IgG, half-dose DC101, vaccine, or a combination as described in A. Tumor size was measured at 3-d intervals. *P < 0.05. n = 8–11 mice per group. Data are
mean ± SEM.
Lower-dose DC101 treatment enhances vac-
| www.pnas.org/cgi/doi/10.1073/pnas.1215397109Huang et al.
dose DC101 treatments (Fig. 3B and Fig. S3B).
Next, we tested the effect of DC101 treatment on a spontaneous
MMTV-PyVT breast cancer, a widely used mouse model of breast
(28). We orthotopically transplanted spontaneous MMTV-PyVT
breast tumors into syngeneic FVB mice and treated these mice
bearing first-generation (F1) isografts with IgG or DC101. We
confirmed that full-dose DC101 treatment significantly decreased
Ho+TAMs in MMTV-PyVT breast tumors compared with IgG
control and quarter-dose DC101 treatment (Fig. 3C and Fig. S3C).
Together, our data suggest that lower-dose DC101 treatment
decreases CD11b+Gr1+cells, whereas high-dose DC101 treatment
decreases a subset of TAMs proximal to perfused tumor vessels in
murine breast cancer models.
The abnormaltumor vasculature creates a highlyheterogeneous
and hypoxic tumor microenvironment (4) that can convert TAMs
hypothesized that TAMs in the area distal to perfused vessels have
more M2-like features than TAMs proximal to perfused vessels.
Indeed, flow-sorted Ho−TAMs had expressed higher levels of
M2-like genes, including Arg1, CSF-1, TGF-β, and MMP9, as com-
pared with Ho+TAMs (Fig. S4A). Next, we examined whether
lower-dose DC101 treatment polarizes TAMs to an immunosup-
portive M1-like phenotype to facilitate a vaccine therapy. Indeed,
lower-dose DC101 treatments (10 and 20 mg/kg) up-regulated
before tumor harvest on day 11 after DC101 treatment. Perfusion images of whole tumor tissue were taken by multispectral confocal microscopy. (A) Rep-
resentative perfusion images of whole tumor tissue treated with (Left to Right) IgG, DC101-10, DC101-20, and DC101-40. Green, Sytox staining; red, Hoechst
33342 staining. (Scale bars, 1,000 μm.) (B) The fractions of Hoechst 33342-positive area in whole tumor area. n = 10–14 mice per group. *P < 0.05. (C) The
fractions of pimonidazole-positive area in total viable areas. n = 10 miceper group. *P < 0.05, **P < 0.01. (D) A distribution histogram of the Ho33342-positive
areas. Tumor areas were subdivided based on a 700-μm grid, the percentage of Hoechst 33342-positive area in each grid was quantified, and their fractions in
the total grid were calculated for each tumor. A distribution histogram for each group was plotted. The x-axis shows the percentage of Ho33342-positive area
in each grid (0.49 mm2). The y-axis shows the percentage of Ho33342-positive area in the total grids analyzed. n = 10–14 mice per group. (E) Quantification of
pericyte coverage (fraction of area covered) in DC101- and IgG-treated groups (20 mg/kg). CD31-positive endothelial cells are stained red; NG2-positive
pericytes are stained green. Confocal images were taken within randomly selected fields excluding the tumor periphery (four to six fields per tumor, six to
eight tumors per group). A 20× objective was used for imaging. (Scale bars, 100 μm.) (F) Quantification of tumor vessel perfusion (fraction of area perfused) in
DC101- and IgG-treated groups (20 mg/kg). CD31-positive endothelial cells are stained red; FITC-lectin–perfused vessels are stained green. Data are shown as
mean ± SEM. **P < 0.01.
Lower-dose DC101 treatment normalizes breast tumor vasculature. When MCaP0008 tumors reached 4–5 mm in diameter, mice were treated with
whereas high-dose DC101 treatment decreases the proportion of
TAMs proximal to perfused tumor vessels. When tumors reached
of CD11b+Gr1+and TAM cells in total viable cells in MCaP0008
tumors. (B) Full-dose DC101 treatment decreased the proportion of
Ho+TAMs in total TAMs in MCaP0008 tumors. (C) Full-dose DC101
treatment decreased the proportion of Ho+TAMs in total TAMs in
Lower-dose DC101 treatment decreases CD11b+Gr1+cells,
Huang et al. PNAS
| October 23, 2012
| vol. 109
| no. 43
M1-like gene transcription compared with full-dose DC101 (Fig.
4A). We separated Ho+TAMs and Ho−TAMs and analyzed their
significantly elevated the levels of M1-type genes (iNOS, IL-12a,
IL-β, CXCL-9, and CXCL-11) and reduced the expression of
with IgG control (Fig. 4B). Consistently, quarter-dose DC101
treatment substantially elevated iNOS and IL-12a and decreased
Arg1 compared with IgG control and full-dose DC101 treatment in
MMTV-PyVT tumors (Fig. S4B). Ho−TAMs displayed mixed
M1/M2-like phenotypes after half-dose DC101 treatment: In
MCaP0008 tumors they had elevated levels of both typical M1
compared with IgG control (Fig. 4C). However, full-dose DC101
distribution of tumor-infiltrating myeloid cells and polarizes TAM
from an immunosuppressive to an immunostimulatory phenotype.
Lower-Dose Anti-VEGFR2 Antibody Treatment Promotes T-Cell Tumor
Infiltration More Potently than Full-Dose Treatment. Improvement
phenotype can facilitate T-cell tumor infiltration (9, 14, 17). Thus,
we evaluated the effect of various doses of DC101 treatment on
T-cell infiltration into tumors. In MCaP0008 tumors, flow cyto-
metric analysis revealed significantly increased tumor-infiltrating
CD4+and CD8+T cells in all DC101-treated groups (10–40 mg/
the increase in T-cell tumor infiltration appeared inversely corre-
lated with DC101 doses. Quarter-dose DC101 treatment signifi-
cantly increased the percentage of tumor-infiltrating CD8+T cells
the proportion of Hoechst 33342+CD8+T cells was higher after
quarter-dose DC101 treatment than after either IgG control or
full-dose DC101 treatment (Fig. 5B). To validate this finding fur-
ther, we treated F1 isografts of spontaneous MMTV-PyVT breast
IgG control. In addition, we tested a spontaneous autochthonous
breast cancer in syngeneic C3H mice by transplanting it ortho-
topically in C3H mice. Consistently, half-dose DC101 treatment
increased tumor-infiltrating CD8+T cells as compared with full-
dose DC101 treatment or IgG control (Fig. S6). Together, these
data demonstrate that lower-dose DC101 is more effective than
high-dose treatment in facilitating CD8+T-cell tumor infiltration
in breast cancers.
Lower-Dose Anti-VEGFR2 Antibody Combined with Vaccine Prolongs
Survival in a Model of MMTV-PyVT Breast Cancer. Next,wetestedthe
therapy in a model of immunogenic MMTV-PyVT breast cancer.
Whole cancer tissue cell vaccination alone induced regression of
MMTV-PyVT breast cancer, supporting its immunogenic char-
acter (Fig. 6A). Consistent with the inhibition of tumor growth,
up-regulated IFN-γ expression in CD8 T cells, and reduced the
proportion of tumor-infiltrating CD4+CD25+FoxP3+Tregs within
the CD4+T-cell population (Fig. 6 B and C and Fig. S7). These
data demonstrated that the whole cancer tissue cell vaccine in-
treatment appeared to stabilize tumor growth, whereas full-dose
DC101 induced tumor regression (Fig. 6A). Together, these data
indicated that orthotopically transplanted MMTV-PyVT breast
cancers are immunogenic and very sensitive to DC101 treatment.
with vaccine moderately enhanced the inhibition of tumor growth
during the treatment period, as compared with vaccine alone (Fig.
prolongation of survival as compared with vaccine monotherapy
40 mg/kg bw) or IgG (40 mg/kg bw) and were perfused with Hoechst33342 as
described in Fig. 2. Gene transcription in different TAM populations was ana-
lyzed by quantitative real-time PCR (Table S1). (A) In MCaP0008 tumors lower-
dose (10 and 20 mg/kg bw) DC101 treatments up-regulated typical M1-like
gene expression in TAMs as compared with both IgG and full-dose (40 mg/kg
bw) DC101 treatment. TAMs were enriched by CD11b-microbead and sepa-
rated by flow sorting. *P < 0.05, **P < 0.01. (B) Half-dose DC101 treatment
elevated expression of M1-like genes and down-regulated expression of M2-
after half-dose DC101 treatment. Data are shown as mean ± SEM. TAMs from
8–10 tumors were pooled as three samples in each group. (B and C) Horizontal
dash: the value of 1. Vertical dash: separates the genes as M1-type (left side)
and M2-type (right side).
breast cancer parenchyma. MCaP0008 (A and B) and MMTV-PyVT (C and D)
tumor-bearing mice were treated with DC101 or rat IgG as described in Figs. 2
and 3. (A) Percentage of CD4+and CD8+T cells in total viable cells. n = 10 mice
per group. In 7AAD−CD45+tumor-infiltrating immune cells, lymphoid cells
were gated according to the side scatter (SSC) and forward scatter (FSC) and
were analyzed for expression of CD4 and CD8a by flow cytometry. *P < 0.05,
**P < 0.01. (B) Quarter-dose DC101 treatment increased the proportion of
Hoechst 33342-positive CD8+(Ho+CD8+) T cells in total CD8+T cells in
MCaP0008 tumors. n = 10 mice per group. *P < 0.05, **P < 0.01. (C) Repre-
sentative flow figures of tumor-infiltrating CD4+and CD8+T cells in spon-
taneous MMTV-PyVT breast tumors. Numbers show the percentages of CD4+
and CD8+T cells in total viable cells. (D) The percentage of tumor-infiltrating
CD4+and CD8+T cells in total viable cells. n = 5 mice per group. Data are
shown as mean ± SEM. *P < 0.05.
Lower-dose DC101 treatment promotes the infiltration of T cells into
| www.pnas.org/cgi/doi/10.1073/pnas.1215397109Huang et al.
vs. vaccine/DC101-40; log-rank test)(Fig. 6D). Depletion of CD8+
T cells in vivo significantly shortened the survival, demonstrating
that the improvement in survival is CD8+T-cell–dependent in
the vaccine plus quarter-dose DC101 group (Fig. 6D and Fig. S8).
Collectively, these data suggest that lower-dose anti-VEGFR2
whencombined witha vaccine therapy ina model of immunogenic
breast cancer sensitive to both vaccine and antiangiogenic thera-
pies and that CD8+T cells mediate this effect.
Malignant tumors escape from host immune surveillance through
multiple mechanisms (5, 7–9). Of these, abnormal tumor vascula-
ture and numerous immune-inhibitory factors producedby tumor-
infiltrating myeloid cells are critical in establishing an immuno-
active cancer immunotherapy.Ourstudydemonstratesthat lower-
dose antiangiogenic treatment normalizes tumor vasculature,
polarizes TAMsto reduceimmune-regulatorysignals, and thereby
creates an immune-supportive microenvironment to recruit and
activate CD8+T cells. Through this mechanism, lower-dose anti-
angiogenic treatment enhancestheanticancerefficacy ofa vaccine
therapy. In contrast, high-dose antiangiogenic/antivascular treat-
ment has smaller or adverse effects on the tumor immune micro-
environment (Fig. 7).
Antiangiogenic treatment is an established therapy in several
solid cancers (29). Although bevacizumab alone does not improve
clinical outcome, the combination of bevacizumab with chemo-
therapy prolongs the survival of patients with advanced non-small
the clinical benefits are in the order of months, and resistance
eventually develops (29). The role of antiangiogenic therapy in
breast cancer remains controversial. Addition of bevacizumab to
chemotherapy improved progression-free survival in patients with
advanced breast cancer butresulted in no overallsurvival benefit in
three large randomized phase III trials (32–34). These disappoint-
ing results raise questions about how antiangiogenic therapy might
[bevacizumab, 10 or 15 mg/kg body weight (bw)] to treat breast
cancers (32, 33). It is likely that high-dose antiangiogenic therapy
prunes tumor vessels excessively, rather than normalizing them,
and thus decreases the delivery of chemotherapeutics (31, 33,
35). This excessive pruning also may exacerbate, rather than re-
verse, the immunosuppressive tumor microenvironment and thus
may compromise the efficacy of active cancer immunotherapy. In
this study, we showed that lower-dose DC101 treatment did not
change vessel density in tumors significantly (Fig. S2) but normal-
ized tumor vessels by improving overall vessel perfusion and cre-
ating a homogeneous distribution of perfused vessels throughout
the tumor. As a result more T cells can be delivered into tumors
(9, 36). Improved vessel perfusion and decreased hypoxia could
polarize TAMs to an M1-like phenotype, and, in turn, elevated
CXCL9 expression in M1-like TAMs further promotes T-cell tu-
mor infiltration (14). In addition, low-dose DC101 treatment in-
creased TAMs and decreased MDSCs in MCaP0008 tumors. This
effect might be caused by the promotion of differentiation of
MDSCs toward M1-like TAMs by low-dose DC101 (16). Fur-
thermore, lower-dose antiangiogenic therapy is likely to reduce
toxicity as well as drug resistance. Therefore, our study suggests
that appropriate lower-dose antiangiogenic therapy could be an
effective strategy to reengineer the tumor microenvironment for
active immunotherapies in a clinical setting.
Different breast cancers have different immunogenicity and
respond differently to a cancer vaccine therapy. In immunogenic
MMTV-PyVT breast cancers, vaccination alone dramatically in-
creased CD8+T cells in tumor tissue and led to tumor regression.
However, in immune-tolerant MCaP0008 breast cancers, vacci-
nation alone did not increase CD8+T-cell tumor infiltration and
did not inhibit tumor growth, even though vaccination activated
CD8+T cells in the spleen. When we compare the intratumor
vessels in these cancer models, the vasculature in MMTV-PyVT
to maximize anticancer efficacy in the MMTV-PyVT tumor model. When
orthotopically transplanted MMTV-PyVT tumors reached 3 mm in diameter,
mice received vaccine, DC101, IgG, or anti-CD8 antibody treatments as de-
scribed in Fig 1A. (A) Tumor growth curves. Tumor size was measured at 3-d
intervals beginning on day 7 after the first vaccination (the first day of
DC101 treatment). The vaccine/DC101-10/anti-CD8 group had 10 mice; all
other groups had 11 mice. (B) The percentages of tumor-infiltrating CD4+
CD25+Foxp3+Tregs in the total CD4+T-cell population. n = 8 mice per group.
*P < 0.05, **P < 0.01. (C) Vaccination up-regulated IFN-γ production in
MMTV-PyVT tumors. Mitomycin C-treated whole breast tumor tissue cells
were used for vaccination. Tumor-infiltrating CD8+T cells were isolated by
anti-CD8 microbeads and then were stimulated by coculturing with mito-
mycin C-treated whole tumor cells in vitro. IFNγ+CD8+T cells were analyzed
by flow cytometry. n = 8 mice per group. *P < 0.05. (D) The combination of
quarter-dose DC101 treatment and vaccine improved survival significantly
compared with vaccine monotherapy. Mice were euthanized when tumors
reached 1,300 mm3. P < 0.0001, vaccine/DC101-10 vs. vaccine/DC101-10/anti-
CD8; P < 0.05, vaccine/IgG vs. vaccine/DC101-10 (log-rank test). The vaccine/
DC101-10/anti-CD8 group had seven mice; all other groups had 11 mice. Data
are shown as mean ± SEM. *P < 0.05.
Quarter-dose DC101 treatment combined with vaccine is sufficient
reprograms the tumor microenvironment from immunosuppressive to immu-
nosupportive and potentiates cancer vaccine therapy. Abnormal tumor vas-
culature creates a hypoxic tumor microenvironment, which impedes the
infiltration of T effector cells into the tumor and polarizes TAMs to the im-
mune-inhibitory M2-like phenotype that suppresses the function of T effector
cells. Lower-dose antiangiogenic treatment normalizes the tumor vasculature
and generates a homogeneous distribution of perfused tumor vessels, which
promotes the infiltration of T effector cells, redirects TAMs to an immune
stimulatory M1-like phenotype, and thereby substantially improves the anti-
cancer efficacy of a cancer vaccine therapy. Conversely, high-dose anti-
angiogenic treatment prunes tumor vessels, increases hypoxia, fails to induce
TAM M1-like polarization, and restricts the infiltration of T effector cells into
tumor parenchyma, resulting in impaired cancer vaccine therapy.
A schematic model showing that lower-dose DC101 treatment
Huang et al.PNAS
| October 23, 2012
| vol. 109
| no. 43
cancer appears to be more functional than that in MCaP0008 Download full-text
cancer in term of pericyte coverage and vessel perfusion (Fig. 2 E
and F and Fig. S9). These factors may be favorable for the in-
filtration of activated CD8+T cells and elicit a relatively immu-
nosupportive tumor microenvironment.
Vaccine therapy usually requires time to generate, activate, and
boost a host immune response against a tumor. Vaccination of
MMTV-PyVT cancer cells pretreated with mitomycin C appeared
to induce tumor regression 10 d after vaccination. Our previous
reports indicated that antiangiogenic treatment could normalize
tumor vessels as early as 2 d posttreatment (4, 19). Therefore, the
schedule of combination treatment (Fig. 1A) was designed to
synchronize vascular normalization and T-cell activation. A pre-
vious study suggested that antiangiogenic therapy preceding vac-
cine therapy had a better anticancer effect than vaccine therapy
followed by antiangiogenic treatment (37). Thus, a comparison of
the efficacy of different combination schedules might yield even
better treatment regimens in the future.
In summary, our data provide preclinical evidence that lower-
dose antiangiogenic therapy can normalize breast tumor vascu-
lature, reprogram the tumor microenvironment from immuno-
suppressive to immunosupportive, and enhance a cancer vaccine
therapy (Fig. 7). This finding may have implications beyond ac-
tive immunotherapy in breast cancer, where the use of low doses
of bevacizumab with chemotherapy ultimately may contribute to
improvements of overall survival (33, 34).
Materials and Methods
Fragments of spontaneous murine mammary carcinoma from MMTV-PyVT
mice or MCaP0008 tumor fragments (38) were implanted orthotopically in
the mammary fat pad of syngeneic immunocompetent FVB mice. When
tumors reached 4–5 mm in diameter, mice were divided into appropriate
groups and received four doses by i.p. injection of either control rat IgG or
DC101 (10, 20, or 40 mg/kg bw) administered at 3-d intervals. For vaccination
experiments, breast tumor-bearing mice were divided randomly into ap-
propriate groups and received four i.p. injections of 5 × 106mitomycin
C-treated, CD45−breast tumor tissue cells or an equal volume of PBS, ad-
ministered at 2- to 3-d intervals (Fig. 1A). CD8 T cells were depleted using 200
μg anti-CD8a monoclonal antibody. On the day before the last vaccination,
both vaccination and control groups were divided randomly into several
groups and were treated with different doses of DC101 or IgG. Experimental
procedures are explained in detail in SI Materials and Methods.
ACKNOWLEDGMENTS. We thank ImClone/Lilly for their gift of DC101;
Glenn Dranoff for advice on cancer vaccine preparation; Sylvie Roberge,
Julia Kahn, Carolyn Smith, Ned Kirkpatrick, Ramone Williams, Madzia
Kowalski, Rachel Ingraham, and Amy Yang for technical assistance; and
Shom Goel, Vikash Chauhan, Eleanor Ager, and Sergey Kozin for their
helpful scientific input. This work was supported by National Institutes of
Health Grants R01-CA115767 and R01-CA126642 (to R.K.J.), R01-CA096915
(to D.F.), and R21-CA139168 and R01-CA159258 (to D.G.D.); a National
Cancer Institutes Federal Share grant (to M.C.P. and J.Y.); Department of
Defense (DoD) Breast Cancer Innovator Award W81XWH-10-1-0016 (to
R.K.J.); DoD Research Fellowship W81XWH-11-1-0619 (to Y.H.). This work
was also supported by the Marsha Rivkin Foundation (E.R. and M.C.P.),
Friends of the Vaccine and Immunotherapy Center, and the Frank Lynch
Jr. Cancer Research Fund (J.Y. and M.C.P.).
1. Madan RA, Gulley JL, Fojo T, Dahut WL (2010) Therapeutic cancer vaccines in prostate
cancer: The paradox of improved survival without changes in time to progression.
2. Rosenberg SA, et al. (2005) Tumor progression can occur despite the induction of very
high levels of self/tumor antigen-specific CD8+ T cells in patients with melanoma. J
3. Offringa R (2006) Cancer. Cancer immunotherapy is more than a numbers game.
4. Jain RK (2005) Normalization of tumor vasculature: An emerging concept in anti-
angiogenic therapy. Science 307:58–62.
5. Schmid MC, Varner JA (2010) Myeloid cells in the tumor microenvironment: Modu-
lation of tumor angiogenesis and tumor inflammation. J Oncol 2010:201026.
6. DeNardo DG, et al. (2011) Leukocyte complexity predicts breast cancer survival and
functionally regulates response to chemotherapy. Cancer Discov 1(1):54–67.
7. Quezada SA, Peggs KS, Simpson TR, Allison JP (2011) Shifting the equilibrium in
cancer immunoediting: From tumor tolerance to eradication. Immunol Rev 241:
8. Schlom J (2012) Therapeutic cancer vaccines: Current status and moving forward.
J Natl Cancer Inst 104:599–613.
9. Hamzah J, et al. (2008) Vascular normalization in Rgs5-deficient tumours promotes
immune destruction. Nature 453:410–414.
10. Huang Y, et al. (2007) Distinct roles of VEGFR-1 and VEGFR-2 in the aberrant hema-
topoiesis associated with elevated levels of VEGF. Blood 110:624–631.
11. Wu NZ, Klitzman B, Dodge R, Dewhirst MW (1992) Diminished leukocyte-endothe-
lium interaction in tumor microvessels. Cancer Res 52:4265–4268.
12. Jochems C, Schlom J (2011) Tumor-infiltrating immune cells and prognosis: The po-
tential link between conventional cancer therapy and immunity. Exp Biol Med
13. Liyanage UK, et al. (2002) Prevalence of regulatory T cells is increased in peripheral
blood and tumor microenvironment of patients with pancreas or breast adenocar-
cinoma. J Immunol 169:2756–2761.
14. Biswas SK, Mantovani A (2010) Macrophage plasticity and interaction with lympho-
cyte subsets: Cancer as a paradigm. Nat Immunol 11:889–896.
15. Schmid MC, et al. (2011) Receptor tyrosine kinases and TLR/IL1Rs unexpectedly acti-
vate myeloid cell PI3kγ, a single convergent point promoting tumor inflammation
and progression. Cancer Cell 19:715–727.
16. Corzo CA, et al. (2010) HIF-1α regulates function and differentiation of myeloid-de-
rived suppressor cells in the tumor microenvironment. J Exp Med 207:2439–2453.
17. Rolny C, et al. (2011) HRG inhibits tumor growth and metastasis by inducing macro-
phage polarization and vessel normalization through downregulation of PlGF. Cancer
18. Huang Y, Snuderl M, Jain RK (2011) Polarization of tumor-associated macrophages: A
novel strategy for vascular normalization and antitumor immunity. Cancer Cell 19:
19. Winkler F, et al. (2004) Kinetics of vascular normalization by VEGFR2 blockade gov-
erns brain tumor response to radiation: Role of oxygenation, angiopoietin-1, and
matrix metalloproteinases. Cancer Cell 6:553–563.
20. Chauhan VP, et al. (2012) Normalization of tumour blood vessels improves the de-
livery of nanomedicines in a size-dependent manner. Nat Nanotechnol 7:383–388.
21. Sorensen AG, et al. (2012) Increased survival of glioblastoma patients who respond to
antiangiogenic therapy with elevated blood perfusion. Cancer Res 72:402–407.
22. Li B, et al. (2006) Vascular endothelial growth factor blockade reduces intratumoral
regulatory T cells and enhances the efficacy of a GM-CSF-secreting cancer immuno-
therapy. Clin Cancer Res 12:6808–6816.
23. Manning EA, et al. (2007) A vascular endothelial growth factor receptor-2 inhibitor
enhances antitumor immunity through an immune-based mechanism. Clin Cancer Res
24. Shrimali RK, et al. (2010) Antiangiogenic agents can increase lymphocyte infiltration
into tumor and enhance the effectiveness of adoptive immunotherapy of cancer.
Cancer Res 70:6171–6180.
25. Righi E, et al. (2011) CXCL12/CXCR4 blockade induces multimodal antitumor effects
that prolong survival in an immunocompetent mouse model of ovarian cancer.
Cancer Res 71:5522–5534.
26. Prewett M, et al. (1999) Antivascular endothelial growth factor receptor (fetal liver
kinase 1) monoclonal antibody inhibits tumor angiogenesis and growth of several
mouse and human tumors. Cancer Res 59:5209–5218.
27. Olive PL, Chaplin DJ, Durand RE (1985) Pharmacokinetics, binding and distribution of
Hoechst 33342 in spheroids and murine tumours. Br J Cancer 52:739–746.
28. Lin EY, et al. (2003) Progression to malignancy in the polyoma middle T oncoprotein
mouse breast cancer model provides a reliable model for human diseases. Am J Pathol
29. Carmeliet P, Jain RK (2011) Molecular mechanisms and clinical applications of an-
giogenesis. Nature 473:298–307.
30. Hurwitz H, et al. (2004) Bevacizumab plus irinotecan, fluorouracil, and leucovorin for
metastatic colorectal cancer. N Engl J Med 350:2335–2342.
31. Sandler A, et al. (2006) Paclitaxel-carboplatin alone or with bevacizumab for non-
small-cell lung cancer. N Engl J Med 355:2542–2550.
32. Miller K, et al. (2007) Paclitaxel plus bevacizumab versus paclitaxel alone for meta-
static breast cancer. N Engl J Med 357:2666–2676.
33. Robert NJ, et al. (2011) RIBBON-1: Randomized, double-blind, placebo-controlled,
phase III trial of chemotherapy with or without bevacizumab for first-line treatment
of human epidermal growth factor receptor 2-negative, locally recurrent or meta-
static breast cancer. J Clin Oncol 29:1252–1260.
34. Miles DW, et al. (2010) Phase III study of bevacizumab plus docetaxel compared with
placebo plus docetaxel for the first-line treatment of human epidermal growth factor
receptor 2-negative metastatic breast cancer. J Clin Oncol 28:3239–3247.
35. Van Cutsem E, Lambrechts D, Prenen H, Jain RK, Carmeliet P (2011) Lessons from the
adjuvant bevacizumab trial on colon cancer: What next? J Clin Oncol 29:1–4.
36. Johansson A, Hamzah J, Payne CJ, Ganss R (2012) Tumor-targeted TNFα stabilizes
tumor vessels and enhances active immunotherapy. Proc Natl Acad Sci USA 109:
37. Farsaci B, Higgins JP, Hodge JW (2012) Consequence of dose scheduling of sunitinib
on host immune response elements and vaccine combination therapy. Int J Cancer
38. Huang P, Duda DG, Jain RK, Fukumura D (2008) Histopathologic findings and es-
tablishment of novel tumor lines from spontaneous tumors in FVB/N mice. Comp Med
| www.pnas.org/cgi/doi/10.1073/pnas.1215397109 Huang et al.