692 | Nature | Vol 577 | 30 January 2020
Article
the phenotypes of T cells that infiltrate the brain (Fig. 3e, Extended
Data Fig. 9a). This result, combined with the increase in CD45+BFP+
cells in the deep cervical lymph nodes (Extended Data Fig. 7i), suggests
that the increased priming and subsequent infiltration of T cells is a
result of increased antigen drainage. Mice that were treated with the
VEGFC mRNA construct also showed a stable number of TCF1+ CD8
T cells—a population that is important for eliciting antitumour effector
responses^26 ,^27 (Extended Data Fig. 9c-f )—in the three compartments
(brain, meninges and deep cervical lymph nodes), along with increased
FOXP3+ CD4 T cells (Extended Data Fig. 9f ). Notably, a higher propor-
tion of FOXP3+ cells expressed T-bet, which is a critical regulator of the
differentiation of T helper 1 cells (Extended Data Fig. 9f ) and is associ-
ated with increased levels of IFNγ in regulatory (CD4+CD25+) T cells
and reduced suppression of CD4 effector cells^28. These data indicate
that there is a shift towards an antitumour immune response in mice
that are treated with VEGF-C. Consistent with this, CD8 T cells in the
brain of VEGF-C treated mice were also capable of producing multiple
cytokines after ex vivo stimulation (Extended Data Fig. 9e–i)—a positive
prognostic factor for immunotherapy^29. All of these changes seemed to
occur independently of a direct effect on T cells, as no VEGFR-3 expres-
sion was detectable in CD4 and CD8 T cells (Extended Data Fig. 7n), and
VEGF-C did not affect the proliferation of T cells in vitro (Extended Data
Fig. 7p). We did not observe expression of VEGFR-3 in other immune
compartments, and stimulation of bone-marrow-derived dendritic
cells with VEGF-C induced no change in the expression levels of MHC
II or costimulatory molecules (Extended Data Fig. 7o). Together, these
results show that VEGF-C provides a potent antitumour environment,
which—in association with increased lymphatic drainage and antigen
presentation—promotes multifunctional and durable T cell immunity
against glioblastoma.VEGF-C restores CNS-restricted T cell priming
In contrast to what is seen in patients with glioblastoma, reports have
shown that combined therapy with nivolumab and ipilimumab had
intracranial efficacy, concordant with extracranial activity, in patients
with melanoma that had metastasized to the brain^30. Consistent with
clinical observations, mice with both a flank and an intracranial tumour
responded better to immunotherapy than did those with just an intrac-
ranial melanoma^31. To examine whether VEGFC mRNA is effective in
treating other (non-glioblastoma) types of cancer in the CNS, we used
the melanoma cell lines YUMMER1.7 and B16 (Extended Data Fig. 10a).
Mice bearing only intracranial YUMMER1.7 tumours showed signifi-
cant survival benefits when treated with VEGFC mRNA and checkpoint
inhibitor therapy (Fig. 4a). By contrast, mice with both intracranial and
flank tumours benefited from checkpoint inhibitor therapy regard-
less of VEGFC mRNA treatment (Fig. 4b). In fact, the survival benefits
to mice with only intracranial tumours that received combination
therapy were similar to mice with both intracranial and flank tumours
that were treated with checkpoint inhibitor therapy alone (Fig. 4c). In
addition, ligation of the deep cervical lymph nodes only affected mice
with intracranial tumours, not mice with both intracranial and flank(^0051015202530)
5.0 × 105
1.0 × 106
1.5 × 106
Days
RLU
VEGFC mRNA + isotype
GFP mRNA + anti-PD-1
GFP mRNA + isotype
VEGFC mRNA + anti-PD-1
(^0020406080100)
20
40
60
80
100
Days
Survival (%)
VEGFCGFPGFP mRNA + isotype mRNA + anti-PD-1 mRNA + isotype
VEGFC mRNA + anti-PD-1
a
GFP
mRNA
VEGFC
mRNA
0
1
2
3
4
Lymph node
Te tramer
-positive
cells (%)
GFP
mRNA
VEGFC
mRNA
0
2
4
6
8
10
Tumour
**
b
c
d
e
(^1079111315)
5
106
Days
No. of CD3
- cells
Brain
Luc-mRNA
VEGFC mRNA
3.31 7.55
1.54 3.65
PE–Cy7 CD45
APC–Tetramer MHC I
GFP mRNA
Tumour
VEGFC mRNA
Tumour
GFP mRNA
dCLN
VEGFC mRNA
dCLN
GL261-Luc
0 7
Intracranial tumour
50,000 cells
VEGFC mRNA/GFP mRNA
intracisternal injection
9 11
anti-PD-antibodies (200 μg) × 3
Monitor
survival
–1 0 4 8 12 16 X
200 μg per mouse at day –1, 0 and every 4 days
100
Rechallenge contralateral
hemisphere
anti-CD4 anti-CD8 Isotype antibodies
Day
Day
Te tramer
-positive
cells (%)
Fig. 3 | Therapeutic delivery of VEGF-C potentiates checkpoint inhibitor
therapy by enhancing T cell priming and recruitment. a, Schematic of
treatment plans. b, Mice inoculated with 50,000 GL261-Luc cells were treated
with VEGFC mRNA or GFP mRNA constructs (day 7) and with either anti-PD-1
(RMP1-14) antibodies or isotype control antibodies (days 7, 9 and 11), and
monitored for survival (top; Kaplan–Meier curve; GFP mRNA + isotype, n = 1 1;
VEGFC mRNA + isotype, n = 9; GFP mRNA + anti-PD-1, n = 8; VEGFC mR NA +
anti-PD-1, n = 10) and tumour burden (bottom; GFP mRNA + isotype, n = 7;
VEGFC mRNA + isotype, n = 5; GFP mRNA + anti-PD-1, n = 4; VEGFC mR NA + anti-
PD -1, n = 5). Data are pooled from 2 independent experiments. *P < 0.0001.
RLU, relative luminescence units. c, d, Mice were inoculated with 50,000
GL261-Luc cells and treated with GFP mRNA or VEGFC mRNA at day 7. Seven
days after mRNA treatment, the deep cervical lymph nodes (dCLN) and
tumour-bearing brain hemispheres were collected to detect tetramer-positive
CD8 T cells. c, Concatenated f luorescence-activated cell sorting (FACS) plots of
CD45+CD3+CD8+CD44+ T cells in tumour-bearing brain and deep cervical lymph
nodes after treatment with GFP mRNA or VEGFC mRNA. d, Quantification of
deep cervical lymph nodes (left; circle, ipsilateral; square, contralateral
(GFP mRNA, n = 6; VEGFC mRNA, n = 12) or tumour-infiltrating tetramer-positive
CD8 T cells (right) (GFP mRNA, n = 3; VEGFC mRNA, n = 5). The experiment was
repeated independently with similar results. P = 0.007 (lymph nodes);
**P = 0.002 (tumour). e, Mice were inoculated with 50,000 GL261-Luc cells and
treated with Luc mRNA or VEGFC mRNA at day 7. Tumour-inoculated brain
hemisphere tissue was collected and analysed using FACS, and the number of
CD3-positive cells was assessed (n = 3; 3 mice were pooled for each sample).
Data are mean ± s.d. P values were calculated by two-tailed unpaired Student’s
t-test or two-sided log-rank Mantel–Cox test.