Letters

https://doi.org/10.1038/s41563-018-0028-2

A facile approach to enhance antigen response for

personalized cancer vaccination
Aileen Weiwei Li1,2, Miguel C. Sobral1,2, Soumya Badrinath3, Youngjin Choi4, Amanda Graveline2,
Alexander G. Stafford2, James C. Weaver2, Maxence O. Dellacherie1,2, Ting-Yu Shih1,2, Omar A. Ali2,
Jaeyun Kim4,5,6, Kai W. Wucherpfennig3 and David J. Mooney1,2*

Existing strategies to enhance peptide immunogenicity for
cancer vaccination generally require direct peptide altera-
tion, which, beyond practical issues, may impact peptide pre-
sentation and result in vaccine variability. Here, we report a
simple adsorption approach using polyethyleneimine (PEI)
in a mesoporous silica microrod (MSR) vaccine to enhance
antigen immunogenicity. The MSR–PEI vaccine significantly
enhanced host dendritic cell activation and T-cell response
over the existing MSR vaccine and bolus vaccine formulations.
Impressively, a single injection of the MSR–PEI vaccine using
an E7 peptide completely eradicated large, established TC-1
tumours in about 80% of mice and generated immunological
memory. When immunized with a pool of B16F10 or CT26 neo-
antigens, the MSR–PEI vaccine eradicated established lung
metastases, controlled tumour growth and synergized with
anti-CTLA4 therapy. Our findings from three independent
tumour models suggest that the MSR-PEI vaccine approach
may serve as a facile and powerful multi-antigen platform to
enable robust personalized cancer vaccination.
Cancer vaccines targeting multiple tumour-specific antigens
can elicit broad immune responses and decrease tumour escape1,2,
and recent advances enable identification of tumour-specific muta-
tions (‘neoantigens’)3,4. Neoantigens are attractive vaccine targets
as they are not expressed in healthy tissues and are predicted to
have strong major histocompatibility complex (MHC)-binding
affinity5. Recent clinical data have shown that neoantigen vaccines
could generate T cells that specifically target heterogeneous tumour
clones6. However, neoantigen peptides exhibit rapid clearance and
low immunogenicity, which limits optimal presentation by antigen-
presenting cells to initiate strong T-cell responses7. Macro- and
nano-engineering strategies have been designed to overcome these
challenges8–11, but many approaches require chemical modification
or physical emulsification of the peptides, potentially altering their
presentation capacity. Moreover, since neoantigen vaccines typi-
cally require many peptides12, modification of individual peptides
is cumbersome for clinical translation and is likely to result in high
batch-to-batch variability.
We propose a facile strategy to enhance antigen immunogenicity
using polyethyleneimine (PEI) combined with a mesoporous silica
microrod (MSR) vaccine. The MSR vaccine can be injected using
standard needles, was shown to effectively concentrate and acti-
vate large populations of host antigen-presenting cells and induced
more potent humoral responses and prophylactic tumour protec-
tion than traditional vaccine formulations13. Moreover, the MSR
surface could potentially be modified to induce stronger responses.
Recent studies have shown that complexes based on PEI, a widely
used cationic polymer14,15, can stimulate pro-inflammatory cytokine
production16,17, and induce potent humoral responses when com-
plexed with glycoproteins18. Here, we explore the application of PEI
to co-present an antigen in a facile, layered adsorption manner in
the MSR vaccine.
MSRs were adsorbed with PEI (MSR–PEI) by simply mixing
with a PEI solution for 15 minutes; subsequently, an antigen pool
was directly adsorbed onto MSR–PEI particles (Fig. 1a). Both rela-
tive molecular mass 60,000 (60K) branched PEI (B60K) and 25K
linear PEI (L25K) absorbed to MSR with high efficiency (Fig. 1b),
with an incorporation capacity of about 20 μ​g PEI per mg MSR.
Over 90% of B60K and L25K PEI polymers were adsorbed after
1 min of mixing (Fig. 1c). Zeta potential measurements confirmed
PEI incorporation (Fig. 1d). MSR–PEIs maintained the intrinsic
mesopores of MSRs (Supplementary Fig. 1a), pore structure and
bulk particle structure (Supplementary Fig. 1b–d), with reduced
surface area and pore volume as expected (Supplementary Table 1).
MSRs and MSR–PEIs showed high incorporation efficiency for
net positive and neutral example murine (Fig. 1e) and human
(Fig. 1f) peptides, but MSR–PEI enhanced the incorporation of net
negative peptides.
The underlying adjuvant effect of PEI19,20 on bone-marrow-
derived dendritic cells (BMDCs) was next examined. BMDCs take up
free PEI, reaching maximum uptake at 24 h (Supplementary Fig. 2a),
and showed a significant increase in CD86 and MHC-II expression
(Fig. 1f) and tumour necrosis factor α​ (TNF-α​) (Fig. 1g) and inter-
leukin-6 (IL-6) (Fig. 1h) production in a PEI dose-dependent man-
ner. BMDCs stimulated with MSR–PEI also showed significantly
increased CD86 expression (Supplementary Fig. 2b) and TNF-α​
production (Supplementary Fig. 2c). MSR–PEI also triggered the
increased production of IL-1β​, a key cytokine produced in response
to Nlrp3 inflammasome activation21 (Supplementary Fig. 3a). This
was probably a result of lysosomal rupture upon MSR–PEI uptake
(Supplementary Fig. 3b), leading to the release of phagosomal con-
tents into the cytosolic compartment. Interestingly, as TNF-α​ and
IL-6 have been shown to be Nlrp3-independent cytokines22, it is
possible that MSR–PEI particles can stimulate multiple damage-
associated molecular pattern (DAMP) receptors. The impact of PEI

1
John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA. 2Wyss Institute for Biologically Inspired
Engineering, Harvard University, Boston, MA, USA. 3Department of Cancer Immunology and Virology, Dana-Farber Cancer Institute, Boston, MA, USA.
4
School of Chemical Engineering, Sungkyunkwan University, Suwon, Republic of Korea. 5Department of Health Sciences and Technology, Samsung
Advanced Institute for Health Science & Technology (SAIHST), Sungkyunkwan University, Suwon, Republic of Korea. 6Biomedical Institute for Convergence
at SKKU (BICS), Sungkyunkwan University, Suwon, Republic of Korea. *e-mail: mooneyd@seas.harvard.edu

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Letters NATuRE MATERIALS
a b 100 c 100

Loading efficiency (%)

80

Loading efficiency (%)

Antigen
PEI pool 60
50
37 °C 37 °C
40
15 min 3h
Bare MSRs MSR-PEI B60K
B60K
L25K 20
L25K

0 0
5 10 20 30 5 10 20 30 0 5 10 15
PEI (μg)/MSR (mg) Time (min)

d 20 e f MSR B60K
L25K
MSR B60K 4 6941 MSR 7
B60K CT29-M19
MSR

Charge at neutral pH
10

Charge at neutral pH
Zeta potential (mV)

CT26-M03 2 7412 3

B16-M27 0.2 2
0 6943
CT26-M90 0.1 P = 0.003
6783 –2

**
P = 0.006 –3
–10 B16-M30

**
P = 0.0005
P = 0.004 6942 –5

***
CT26-M20 –5

**
–20
0 1 5 10 20 0 50 100 0 50 100
PEI (μg)/MSR (mg) Adsorption efficiency (%) Adsorption efficiency (%)

g h i j
50 600 250

SIINFEKL-H2Kb+ CD11c+ (%)

P < 0.0001 30 P < 0.0001
P < 0.0001
P < 0.0001****
P = 0.0008
**** ****
40 200 *** P = 0.0036
CD86+ MHC-II+ (%)

**** P = 0.0003
TNF-α (pg ml–1)

**
IL-6 (pg ml–1)

400 P = 0.0004 20
30 150
*** *** P = 0.02
20 100 10 *
200
10 50
0
0 0 0

S
VA

μg EI
I
PE
PB

10 g P
O
PBS 1 7 1 7 PBS 1 7 1 7 PBS 1 7 1 7

μ
5
L25 B60 L25 B60 L25 B60
OVA+PEI

Fig. 1 | PEI can be rapidly incorporated onto MSRs and leads to murine and human DC activation. a, Schematic representations of PEI and subsequent
antigen adsorption onto bare MSRs. b, Incorporation efficiency of various doses of soluble B60K and L25K PEI into MSRs (n =​ 3). c, Incorporation kinetics
of soluble B60K and L25K PEI into MSRs (representative data, repeated at least three times). d, Zeta potential of MSR–PEI particles using various doses
of soluble B60K and L25K PEI (n =​ 3). e,f, Incorporation efficiency of murine (e) and human (f) neoantigen peptides (colour-codedaccording to the net
charge at neutral pH) onto bare MSR or MSR–PEI particles using B60K PEI (n =​ 3, two-tailed t-test). g, Flow cytometry analysis of CD86 and MHC-II
expression on murine BMDCs after 24 h of stimulation with 1 μ​g or 7 μ​g of soluble PEI or PBS (n =​ 4, compared with PBS by one-way analysis of variance
(ANOVA)). h,i, Enzyme-linked immunosorbent assay (ELISA) analysis of TNF-α​ (h) and IL-6 (i) concentration in murine BMDC supernatant after 24 h of
stimulation with 1 μ​g or 7 μ​g of soluble B60K and L25K PEI or PBS (n =​ 4, compared with PBS by one-way ANOVA). j, Flow cytometry analysis of SIINFEKL-
presenting murine BMDCs after stimulation with PBS, OVA, and OVA with 5 μ​g or 10 μ​g of soluble B60K PEI (n =​ 3, compared with OVA by one-way
ANOVA). Data depict mean ±​ s.d.

on antigen presentation was next examined by pulsing BMDCs with (GM-CSF) to recruit host DCs, the TLR-9 agonist CpG-ODN and
ovalbumin (OVA) either alone or together with B60K PEI for 2 days. a model antigen OVA as previously described13 with the MSR–PEI
B60K PEI led to an approximately 10- to 20-fold increase in anti- vaccine incorporating these agents (Fig. 2a). In vitro, GM-CSF
gen cross-presentation when compared with OVA alone (Fig. 1i). and PEI were released in a sustained manner from the MSR–
Finally, as the MSRs and PEI solutions were endotoxin free PEI vaccine, while CpG was rapidly released due to its negative
(Supplementary Table 2), and BMDC activation was unaffected by charge (Supplementary Fig. 5). Mice were immunized with the
TLR4 neutralization (Supplementary Fig. 4), the activation of DCs MSR vaccine (V) or the MSR–PEI vaccine (VP). In both vaccines,
was probably not a result of material contamination. Together, these CpG-ODN was incorporated, as TLR-9 activation can efficiently
data suggest that PEI can be efficiently and easily incorporated induce a type 1 and CD8 killer T-cell (CTL) response23. The total
into MSRs, that a broad range of neoantigen peptides can be sub- numbers of cells recruited to the vaccine on day 3 were compa-
sequently incorporated via adsorption and that PEI and MSR–PEI rable between V and VP immunized mice (Fig. 2b). However,
particles can enhance DC activation and cross-presentation. VP showed a significant enrichment of CD11c+ CD86+ activated
The ability of an MSR–PEI vaccine to improve DC activation DCs (Fig. 2c), CD11c+ CCR7+ LN homing DCs (Fig. 2d) and
in vivo was then analysed by comparing the MSR vaccine incor- CD11c+ SIINFEKL-H2Kb+ antigen cross-presenting DCs (Fig.
porating granulocyte–macrophage colony-stimulating factor 2e). The vaccine draining lymph node (dLN) total cellularities

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATuRE MATERIALS Letters
a V VP b P = 0.63 c d
ns P = 0.002
1 × 105 1 × 105
MSR: CpG MSR: CpG ** **
1.5 × 106 P = 0.0096
8 × 104 8 × 104

Total number of cells

CD11c+CCR7+
CD11c+CD86+
MSR-PEI: OVA

activated DCs
MSR: OVA

activated DC
1.0 × 106 6 × 104 6 × 104
MSR: GM-CSF MSR: GM-CSF
4 × 104 4 × 104
5.0 × 105
2 × 104 2 × 104
Lyophilize Lyophilize
0.0 0 0
Mix and inject Mix and inject V VP V VP V VP

e f P < 0.0001 g P < 0.0001 h

P < 0.0001

Number of CD11c+ MHC-II+ DC

3 × 107 **** 8 × 105 6 × 103 P = 0.036
CD11c+ SIINFEKL-MHC-I+ DCs

Number of CD11c+ CD86+ DC

Number of DQ-OVA+ CD11c+

N **** 1 × 106 N
Total number of cells in dLN

N **** N
4 × 104 V V *
* 8 × 105
V
5
V
P = 0.02 VP VP 6 × 10 VP VP
3 × 104 2 × 107 4 × 103
P < 0.0001 6 × 105 P = 0.0001 P < 0.0001
2 × 104
**** 4 × 105

4 × 105 *** *****

1 × 107 2 × 103
5
1 × 104 5 2 × 10
2 × 10

0 0 0 0 0
V VP Day 3 Day 5 Day 3 Day 5 Day 3 Day 5 Day 3 Day 5

i j P = 0.02 k P = 0.007
l P = 0.01
4 × 105 1 × 106 5 × 103
* *
5
** 3
CD11c+ CD86+ DC
4 × 10
CD11c+ CD86+ DC

8 × 10

DQ-OVA+ CD11c+
Trans VP Cis VP 3 × 105

Number of
Number of

MSR-PEI: CpG
Number of

MSR: CpG 6 × 105 3 × 103

2 × 105
MSR: OVA MSR-PEI: OVA 4 × 105 2 × 103
5
1 × 10
2 × 105 1 × 103
MSR: GM-CSF MSR: GM-CSF
0 0 0
Trans Cis Trans Cis Trans Cis
VP VP VP

Fig. 2 | MSR–PEI vaccine enhances DC activation and trafficking in situ. a, Schematic representations of the MSR vaccine (V), boxed in black, and
MSR–PEI vaccine (VP), boxed in red. b, Total cell number at the vaccine site explanted on day 3 after immunization with V or VP using B60K PEI (n =​ 4,
two-tailed t-test). c–e, Total number of CD11c+ CD86+ activated DCs (n =​ 4, two-tailed t-test) (c), CD11c+ CCR7+ LN homing DCs (n =​ 4, two-tailed
t-test) (d) and SIINFEKL-presenting DCs (n =​ 4, two-tailed t-test) (e) recruited to the vaccine site on day 3 after immunization with V or VP using B60K
PEI. f–h, Total number of cells (n =​ 4 for day 3 and n =​ 5 for day 5, two-way ANOVA) (f), of CD11c+ CD86+ or CD11c+ MHC-II+ activated DCs (n =​ 4 for
day 3 and n =​ 5 for day 5, two-way ANOVA) (g) and OVA+ DCs (n =​ 4 for day 3 and n =​ 5 for day 5, two-way ANOVA) (h) in the dLN on days 3 and 5
after immunization with V or VP using B60K PEI or left unimmunized (N). i, Schematic representations of the MSR–PEI trans vaccine (trans VP) and the
MSR–PEI cis vaccine (cis VP). j, Total number of CD11c+ CD86+ activated DCs at the vaccine site on day 3 after immunization with the trans VP vaccine
or the cis VP vaccine (n =​ 5, two-tailed t-test). k,l, Total number of CD11c+ CD86+ activated DCs (k) and CD11c+ OVA+ DCs (l) in the dLN on day 5 after
immunization with the trans VP vaccine or the cis VP vaccine (n =​ 5, two-tailed t-test). Data depict mean ±​ s.d.

between V and VP immunized mice were similar on day 3, but the MSR–PEI vaccine enhanced host DC activation, antigen pre-
VP immunized mice showed strikingly higher dLN cellularity on sentation and trafficking to secondary lymphoid organs.
day 5 (Fig. 2f). Additionally, VP elicited a strong enrichment in The ability of the MSR–PEI vaccine to induce a CTL response
activated DCs (Fig. 2g and Supplementary Fig. 6a) and antigen- with the model antigen OVA was next examined. Circulating
presenting DCs (Fig. 2h), but not in the macrophage populations peripheral blood mononuclear cells (PBMCs) were analysed 7 days
(Supplementary Fig. 6b,c) in the dLN. Next, the impact of a direct after immunization with V or VP vaccines. Mice immunized with
association between antigen and PEI was evaluated. While main- VP generated about twice as many circulating IFN-γ​+ (Fig. 3a) and
taining a constant dose of GM-CSF, CpG, OVA and PEI, two types tetramer+ (Fig. 3b) CTLs when compared with V. Furthermore,
of VP vaccine were tested: (1) a trans form (trans VP) that com- effector T cells outnumbered regulatory T cells by about 15-fold at
bined CpG adsorbed onto MSR–PEIs, OVA adsorbed onto bare the VP vaccine site, almost three times higher than at the V vaccine
MSRs and GM-CSF adsorbed onto also bare MSRs, or (2) a cis site (Fig. 3c). Increasing the PEI dose in VP to above 40 μ​g per vac-
form (cis VP) that combined OVA adsorbed onto MSR–PEIs, CpG cine led to a decrease in CTL responses (Fig. 3d). Similarly to VP
adsorbed on to bare MSRs and GM-CSF adsorbed onto also bare using B60K PEI, the VP formulation using L25K PEI also showed
MSRs (Fig. 2i). The incorporation efficiencies of GM-CSF, CpG an enhanced CTL response when compared with V (Fig. 3e). No
and OVA were comparable (Supplementary Table 3). The cis VP significant differences were observed when varying PEI structure
vaccine showed twice as many activated DCs in the scaffold (Fig. and molecular weight (Supplementary Fig. 7a). Finally, and consis-
2j) and three times as many activated and antigen-presenting DCs tent with our previous finding on host DC activity, a cis VP vaccine
in the dLN (Fig. 2k,l) when compared with the trans VP vaccine. showed a significantly higher CTL response when compared with
The myeloid cell population in the scaffold was not impacted a trans VP vaccine and the MSR vaccine (Fig. 3f). To exclude the
(Supplementary Fig. 6d). Collectively, these data demonstrate that possibility that a higher dose of PEI in the trans VP vaccine may

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Letters NATuRE MATERIALS

a 5 P = 0.03 b 5 P = 0.037
* *

IFNy+ within CD3+CD8+ (%)

N V VP

Tet+ within CD3+CD8+ (%)

4 4
105 0.00% 105 1.78% 105 2.39%
4 4
10 10 104 3 3
103 103 103
IFNγ-APC

IFNγ-APC

IFNγ-APC
2 2
102 102 102
0 0 0
1 1
0 102 103 104 105 0 102 103 104 105 0 102 103 104 105
CD8a-FITC CD8a-FITC CD8a-FITC
0 0
N V VP N V VP

c d P = 0.01 e f * (P = 0.017)
8
* P = 0.03 # (P = 0.015)
20 IFNy+ within CD3+CD8+ (%)
8 *
P = 0.003
6 6

IFNy+ within CD3+CD8+ (%)

IFNy+ within CD3+CD8+ (%)
**
Scaffold Teff /Treg ratio

15 6

4 4
10 4

2 2
5 2

0 0 0 0
V VP N 10 20 40 80 N V VP L25 N V Trans Cis

VP (μg PEI) VP

Fig. 3 | MSR–PEI vaccine enhances CD8 cytotoxic T-cell response against OVA. a, Percentage of IFN-γ​+ CD8+ T cells isolated from peripheral blood on
day 7 after immunization with V or VP using B60K PEI or left unimmunized, and stimulated with SIINFEKL (three primary fluorescence-activated cell
sorting (FACS) plots on the left, quantifications from the FACS plots on the right) (n =​ 5, one-way ANOVA). b, Percentage of SIINFEKL-tetramer+ CD8+
T cells isolated from peripheral blood on day 7 after immunization with V or VP using B60K PEI or left unimmunized (N) (n =​ 5 for VP, n =​ 4 for N and
V, one-way ANOVA). c, Ratio of CD8+ effector T cells (Teff) to CD4+ Foxp3+ regulatory T cells (Treg) at the MSR vaccine site on day 11 after immunization
with V or VP using B60K PEI (n =​ 5, two-tailed t-test). d, Percentage of IFN-γ​+ CD8+ T cells isolated from peripheral blood on day 7 after immunization
with VP containing various doses of B60K PEI or left unimmunized (N) (n =​ 4, one-way ANOVA). e, Percentage of IFN-γ​+ CD8+ T cells isolated from
peripheral blood on day 7 after immunization with V or the MSR-PEI vaccine using L25K PEI (VP L25) or left unimmunized (N) (n =​ 4, one-way ANOVA).
f, Percentage of IFN-γ​+ CD8+ T cells isolated from peripheral blood on day 7 after immunization with V, the MSR-PEI trans vaccine (trans, VP) using B60K
PEI or the MSR-PEI cis vaccine (cis, VP) using B60K PEI, or left unimmunized (N), and stimulated with SIINFEKL (n =​ 9, * between cis VP and trans VP, #
between cis VP and V by one-way ANOVA). Data depict mean ±​ s.d.

improve the response, two increasing doses of PEI in the trans inoculation, were completely tumour free after the rechallenge
VP vaccine were evaluated, and the CTL response decreased with (Fig. 4e). In a separate experiment, VP was evaluated against a tra-
increasing PEI (Supplementary Fig. 7b). ditionally formulated bolus vaccine. Consistent with previously
To test whether the MSR–PEI vaccine may induce anti-tumour reported findings, the bolus vaccine was partially effective and
immunity against tumour-specific peptides, a synthetic long pep- induced complete regression in only about 20% of the animals9,26.
tide derived from the E7 oncoprotein of human papilloma virus In comparison, VP induced faster regression and about 80% of
(HPV) was used as the antigen. Beyond cervical cancer, HPV is treated animals again showed complete regression (Supplementary
associated with 30–60% of oropharyngeal, vaginal and head-and- Fig. 8a,b). The anti-tumour effect of VP was antigen specific
neck cancers24–26. Consistent with our findings using OVA antigen, (Fig. 4f) and mediated by CD8+ T cells (Fig. 4g, h). Importantly, the
VP induced about two to three times higher E7-specific IFN-γ​+ therapeutic regression accomplished with just one immunization
(Fig. 4a) and tetramer+ (Fig. 4b) circulating CTLs when compared with VP was stronger than multiple treatments of conventionally
with V. Mice immunized with VP showed significantly elevated adjuvanted E7 vaccines, and messenger RNA and nanoparticle-
blood TNF-α​levels (Fig. 4c). Next, mice were inoculated with based technologies in similar tumour studies9,27–29. To further dem-
E7-expressing TC-1 carcinoma subcutaneously and treated with onstrate that a direct association between the antigen and PEI in
a single injection of V or VP when tumour areas reached about the vaccine is important, animals bearing established subcutaneous
25 mm. VP triggered rapid and complete regression in a majority TC-1 tumours were treated with trans VP (E7 VP trans) and cis
of the animals, whereas V induced only partial regression (Fig. 4d). VP (E7 VP cis). Mice treated with cis VP showed superior tumour
Importantly, VP led to complete regression of even tumours that regression (Supplementary Fig. 8c), indicating that co-presentation
reached large dimensions (~1 ×​ 1 cm2). VP also enhanced overall of the antigens and PEI in the vaccine is probably a crucial compo-
survival by about twofold when compared with V. Impressively, nent to generating anti-tumour immunity.
about 80% of animals treated with VP survived long term, beyond Finally, the MSR–PEI vaccine’s ability to serve as a facile and
150 days (Fig. 4e). To test immunological memory, the animals effective platform for multiple neoantigens to induce tumour control
that survived from the first inoculation were rechallenged with the was tested. First, VP demonstrated superior CTL response against
same E7-expressing TC-1 carcinoma after 6 months. All mice that B16F10 neoantigens in both tumour-free and tumour-bearing
were treated with either V or VP, and survived the first tumour mice when compared with V (Supplementary Figs. 9, 10). Next,

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATuRE MATERIALS Letters
a b c
P = 0.06
*P = 0.03 *P = 0.02
2.5 4 60 **

IFNy+ of CD3+CD8+ (%)

*P = 0.03

Tet+ of CD3+CD8+ (%)

P = 0.0495 P = 0.2
2.0
3 ns

TNF-α (pg ml–1)

1.5 40
2
1.0
0.5 1 20

0.0 0
N V μg
μg N V μg
μg 0
520 520 N V VP
VP (μg PEI) VP (μg PEI)

f 100
d TC-1 V e Rechallenge ***
300 P = 0.0001
N 100
P = 0.03

Percentage survival
V 0 8
Tumour area (mm2)

Percentage survival
VP
Day
P value * Naive
200 (V and VP) N SIINFEKL VP
18 0.03 V 50 E7 VP
50 VP
20 0.0006
100 22 0.003
*
*** 24 0.01
** 26 0.02
***
0 * 0 0
28 0.03
0 10 20 30 40 0 100 200 300 0 10 20 30
30 0.03
Time Time Time
(days after inoculation) (days after inoculation) (days after inoculation)

g h Naive isotype
Isotype mab a-CD8a mab 300
5 5 VP isotype
10 10
Tumour area (mm2) Naive a-CD8
7.09% 0.040% VP a-CD8
104 104 200
P value
103 103 Day (VP a-CD8
and VP iso)
CD8a - PE

CD8a-PE

12.2% 100
102 102 20.4% **** 15 0.02
0 0
**** 17 <0.0001
* **** ****
0 102 103 104 105 0 102 103 104 105 19 <0.0001
0
CD4 – Pac blue CD4–Pac blue 0 5
10 15 20 25 21 <0.0001
Time 23 <0.0001
(days after inoculation)

Fig. 4 | MSR–PEI vaccine enhances CD8 cytotoxic T-cell response against E7 and regresses established tumours. a,b, Percentage of IFN-γ​+ CD8+ T cells
in response to RAHYNIVTF stimulation (a) and percentage of tetramer+ CD8+ T cells (b) in peripheral blood on day 7 after immunization with the MSR
E7 vaccine (V) or the MSR–PEI E7 vaccine (VP) using 5 μ​g or 20 μ​g of B60K PEI, or left unimmunized (N) (n =​ 4, one-way ANOVA). c, ELISA analysis of
TNF-α​level in serum 24 h after vaccination with V or the MSR-PEI (B60K) vaccine (VP), or left unimmunized (N) (n =​ 4, compared with N by one-way
ANOVA). d,e, Tumour growth (d) and overall survival (e) of mice bearing established E7-expressing TC-1 tumours (allowed to develop for 8 days) and
treated with V or VP using L25K PEI, or left untreated (N), and subsequently rechallenged with TC-1 cells 6 months after the first inoculation (n =​ 10,
compared with V by two-way ANOVA for d and by log-rank test for e). f, Overall survival of mice bearing established E7-expressing TC-1 tumours and
treated with the MSR–PEI vaccine containing E7 (E7 VP) or the MSR–PEI vaccine containing SIINFEKL (SIINFEKL VP), or left untreated (Naive) (n =​ 8,
compared with SIINFEKL VP by log-rank test). g, Flow cytometry analysis of blood T cells 3 days after treatment with a-CD8a monoclonal antibody (mab)
or an isotype monoclonal antibody (representative data, repeated three times). h, Tumour growth of mice bearing established E7-expressing TC-1 tumours
and treated with the MSR–PEI vaccine with either a-CD8a monoclonal antibody or an isotype monoclonal antibody (n =​ 8, compared with VP a-CD8 by
two-way ANOVA). In a–c data depict mean ±​ s.d. and in d,h data depict mean ±​ s.e.m.

VP was evaluated in the highly aggressive and immune-suppressive with repeated immunizations of RNA-based neoantigen vaccines5.
B16F10 and CT26 models using combinations of neoantigen A standard prime and boost of VP improved subcutaneous B16F10
peptides5. As effector tumour-infiltrating lymphocytes (TILs) are growth control (Fig. 5c) and induced temporary regression in a
needed to achieve tumour clearance7, the TIL population was iso- subset of the animals (Fig. 5d). Checkpoint blockade therapies are
lated, directly stained and analysed on day 15 after inoculation, by effective in various types of cancer, but the response rate depends
which point some tumours in the untreated and V-treated animals on a pre-existing immunity30. In this aggressive model, anti-CTLA4
had developed to the maximum allowed size. Impressively, tumours therapy alone did not confer significant tumour growth control.
in VP-treated animals showed stronger enrichment in the IFN-γ​+, However, one injection of VP in combination with anti-CTLA4
TNF-α​+ and granzyme B+ TILs (Fig. 5a, Supplementary Fig. 11). In treatment significantly augmented the response (Fig. 5e).
contrast, V did not generate enhanced TIL response when compared This study describes a potentially transformative, simple and
with untreated animals. Lung metastases of B16F10 (Fig. 5b) and modular strategy to enhance antigen immunogenicity and drive
CT26 (Supplementary Fig. 12) tumours were efficiently eradicated effective anti-tumour immunity. This approach can effectively drive
by one injection of VP containing a combination of B16 or CT26 immune responses against libraries of cancer-specific mutations
neoantigens, respectively. Notably, these responses achieved with and synergize with other immunotherapies. The vaccine is assem-
a single immunization of VP were comparable to those achieved bled in less than 3 h by simple mixing of all components, and can

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Letters NATuRE MATERIALS
a b e
T Vax
(IV) (M27, M30) Lung T Vax
P = 0.0052
(SC) (M27, M30, M47, M48)
1,500 ** a-CTLA4
0 1 16

(per 500,000 cells)

CD44+ TNF-α+
T Vax 1,000 N VP
(SC) (M27, M30) TIL 0 5
P = 0.6598
ns N
500 300
0 5 15

Tumour area (mm2)

0 200
N V VP
250
P = 0.022 P = 0.0096
100
800 * *

Lung metastasis (no.)

30 200
CD44+ granzyme B+
(per 500,000 cells)

600 P = 0.467 (per 500,000 cells) 0

CD44+ IFNy+

150
ns 20 0 5 10 15 20
400 P = 0.776 100 a-CTLA4
ns 300
10 P = 3.6 × 10–7 ns, P = 0.07
200

Tumour area (mm2)

50
****
0 0 200
0
N V VP N V VP N VP
100
c T Vax Vax d
(SC) (M27, M30) (M27, M30)
200 0
0 5 10 15 20
0 3 13
Tumour volume change (%)

150 VP + a-CTLA4
500 500 300
N VP*, P = 0.02 *, P = 0.046

Tumour area (mm2)

Tumour area (mm2)

400 400 ***, P = 0.0008

100
200
300 300 P = 0.0004
50 ***
200 200
100
100 100 0
0 0 0
0 10 20 30 0 10 20 30 –50 0 5 10 15 20
Time (days post inoculation) N VP Time (days post inoculation)

Fig. 5 | MSR–PEI vaccine enhances melanoma TIL effector function and induces tumour control and synergy with anti-CTLA4 therapy using combined
B16 neoantigens. a, Number of CD44+ IFN-γ​+, CD44+ TNF-α​+ and CD44+ granzyme B+ TILs per 500,000 tumour cells on day 15 after inoculation. Mice
bearing established B16F10 tumours (allowed to develop for 5 days) and treated with the MSR vaccine (V) or the MSR–PEI vaccine (VP) using L25K PEI
and 50 μ​g of the B16 neoantigens, or left untreated (N) (n =​ 5 for VP, n =​ 3 for N and V, one-way ANOVA). b, Number of lung metastases formed on day 16
after inoculation in mice that received IV inoculation of B16F10 melanoma cells (allowed to develop for 1 day) and were treated with VP using L25K PEI and
50 μ​g of the B16 neoantigens, or left untreated (N). Primary representative photographs of excised lungs are shown (n =​ 6, two-tailed t-test).
c, Tumour growth in mice bearing established B16F10 tumours (allowed to develop for 3 days) and treated with two injections of VP using L25K PEI and
50 μ​g of the B16 neoantigens on days 3 and 13, or left untreated (N) (n =​ 8, two-tailed t-test). d, Tumour volume change between days 13 and 17 after
tumour inoculation (n =​ 8, two-tailed t-test). e, Tumour growth of mice bearing established B16F10 tumours (inoculated with 1 ×​ 105 cells) and treated
with anti-CTLA4 antibody (a-CTLA4), anti-CTLA4 antibody in combination with the MSR–PEI vaccine (VP +​ a-CTLA4) using L25K PEI and 50 μ​g of the
B16 neoantigens on day 5, or left untreated (N) (n =​ 8, *significant difference between VP +​ a-CTLA4 and a-CTLA4, ***significant difference between
VP +​ a-CTLA4, ns between a-CTLA4 and N by one-way ANOVA). In a,b,d data depict mean ±​ s.d., in c,e data depict individual tumour growth.

be stored in a lyophilized form before or after antigens are added. 3. Hacohen, N., Fritsch, E. F., Carter, T. A., Lander, E. S. & Wu, C. J.
Overall, this approach may have high value as a therapeutic multi- Getting personal with neoantigen-based therapeutic cancer vaccines.
antigen platform to enable robust personalized vaccination. Cancer Immunol. Res. 1, 11–15 (2013).
4. Linnemann, C. et al. High-throughput epitope discovery reveals frequent
Methods recognition of neo-antigens by CD4+​T cells in human melanoma. Nat. Med.
21, 81–85 (2015).
Methods, including statements of data availability and any asso- 5. Kreiter, S. et al. Mutant MHC class II epitopes drive therapeutic immune
ciated accession codes and references, are available at https://doi. responses to cancer. Nature 520, 692–696 (2015).
org/10.1038/s41563-018-0028-2. 6. Ott, P. A. et al. An immunogenic personal neoantigen vaccine for patients
with melanoma. Nature 547, 217–221 (2017).
Received: 26 April 2017; Accepted: 19 January 2018; 7. van der Burg, S. H., Arens, R., Ossendorp, F., van Hall, T. & Melief, C. J.
Vaccines for established cancer: overcoming the challenges posed by immune
Published: xx xx xxxx
evasion. Nat. Rev. Cancer 16, 219–233 (2016).
8. Moon, J. J. et al. Interbilayer-crosslinked multilamellar vesicles as synthetic
References vaccines for potent humoral and cellular immune responses. Nat. Mater. 10,
1. Gubin, M. M., Artyomov, M. N., Mardis, E. R. & Schreiber, R. D. Tumor 243–251 (2011).
neoantigens: building a framework for personalized cancer immunotherapy. 9. Liu, H. et al. Structure-based programming of lymph-node targeting in
J. Clin. Invest. 125, 3413–3421 (2015). molecular vaccines. Nature 507, 519–522 (2014).
2. Schumacher, T. N. & Schreiber, R. D. Neoantigens in cancer immunotherapy. 10. Cho, N. H. et al. A multifunctional core–shell nanoparticle for dendritic
Science 348, 69–74 (2015). cell-based cancer immunotherapy. Nat. Nanotech. 6, 675–682 (2011).

Nature Materials | www.nature.com/naturematerials

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATuRE MATERIALS
11. Cho, H. J. et al. Immunostimulatory DNA-based vaccines induce cytotoxic
lymphocyte activity by a T-helper cell-independent mechanism.
Nat. Biotechnol. 18, 509–514 (2000).
12. Bobisse, S., Foukas, P. G., Coukos, G. & Harari, A. Neoantigen-based cancer
immunotherapy. Ann. Transl. Med. 4, 262–270 (2016).
13. Kim, J. et al. Injectable, spontaneously assembling, inorganic scaffolds
modulate immune cells in vivo and increase vaccine efficacy. Nat. Biotechnol.
33, 64–72 (2015).
14. Song, W. J., Du, J. Z., Sun, T. M., Zhang, P. Z. & Wang, J. Gold nanoparticles
capped with polyethyleneimine for enhanced siRNA delivery. Small 6,
239–246 (2010).
15. Sakai, S., Yamada, Y., Yamaguchi, T., Ciach, T. & Kawakami, K. Surface
immobilization of poly(ethyleneimine) and plasmid DNA on electrospun
poly(L-lactic acid) fibrous mats using a layer-by-layer approach for gene
delivery. J. Biomed. Mater. Res. A 88, 281–287 (2009).
16. Oh, Y.-K. et al. Enhanced adjuvanticity of interleukin-2 plasmid
DNA administered in polyethylenimine complexes. Vaccine 21,
2837–2843 (2003).
17. Mulens-Arias, V., Rojas, J. M., Pérez-Yagüe, S., Morales, M. P. & Barber, D. F.
Polyethylenimine-coated SPIONs trigger macrophage activation through
TLR-4 signaling and ROS production and modulate podosome dynamics.
Biomaterials 52, 494–506 (2015).
18. Wegmann, F. et al. Polyethyleneimine is a potent mucosal adjuvant for viral
glycoprotein antigens. Nat. Biotechnol. 30, 883–888 (2012).
19. Sheppard, N. C. et al. Polyethyleneimine is a potent systemic adjuvant for
glycoprotein antigens. Int. Immunol. 26, 531–538 (2014).
20. He, W. et al. Re-polarizing myeloid-derived suppressor cells (MDSCs) with
cationic polymers for cancer immunotherapy. Sci. Rep. 6 (2016).
21. Hornung, V. et al. Silica crystals and aluminum salts activate the NALP3
inflammasome through phagosomal destabilization. Nat. Immunol. 9,
847–856 (2008).
22. Eisenbarth, S. C., Colegio, O. R., O’Connor, W., Sutterwala, F. S. & Flavell, R.
A. Crucial role for the Nalp3 inflammasome in the immunostimulatory
properties of aluminium adjuvants. Nature 453, 1122–1126 (2008).
23. Vollmer, J. & Krieg, A. M. Immunotherapeutic applications of CpG
oligodeoxynucleotide TLR9 agonists. Adv. Drug Deliv. Rev. 61,
195–204 (2009).
24. Bartkowiak, T. et al. Unique potential of 4-1BB agonist antibody to promote
durable regression of HPV+​tumors when combined with an E6/E7 peptide
vaccine. Proc. Natl Acad. Sci. USA 112, E5290–E5299 (2015).
25. Chaturvedi, A. K. Beyond cervical cancer: burden of other HPV-related
cancers among men and women. J. Adolesc. Health 46, S20–S26 (2010).
Nature Materials | www.nature.com/naturematerials
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Letters
26. Welters, M. J. et al. Vaccination during myeloid cell depletion by cancer
chemotherapy fosters robust T cell responses. Sci. Transl. Med. 8, 334ra352 (2016).
27. Zwaveling, S. et al. Established human papillomavirus type 16-expressing
tumors are effectively eradicated following vaccination with long peptides.
J. Immunol. 169, 350–358 (2002).
28. van der Sluis, T. C. et al. Therapeutic peptide vaccine-induced CD8 T cells
strongly modulate intratumoral macrophages required for tumor regression.
Cancer Immunol. Res. 3, 1042–1051 (2015).
29. Kranz, L. M. et al. Systemic RNA delivery to dendritic cells exploits antiviral
defence for cancer immunotherapy. Nature (2016).
30. Kvistborg, P. et al. Anti-CTLA-4 therapy broadens the melanoma-reactive
CD8+​T cell response. Sci. Transl. Med. 6, 254ra128 (2014).
Acknowledgements
The authors would like to thank C. J. Wu (Dana Farber Cancer Institute) for providing
the human neoantigen peptides. We are also grateful to G. Dranoff, C. S. Verbeke,
G. J. Xu and A. S. Cheung for their helpful discussions and feedback on the manuscript.
This work was supported by the National Institutes of Health (NIH) R01EB015498
and R01EB023287, the Melanoma Research Alliance Foundation, the National Science
Foundation (NSF) Graduate Research Fellowship Program (AWL) and the Wyss Institute
for Biologically Inspired Engineering.
Author contributions
A.W.L. and D.J.M. conceived the study, designed the experiments and wrote the
manuscript, A.W.L., M.C.S., S.B., Y.C., A.G., A.G.S., J.C.W., M.O.D. and T-Y.S. carried
out the experiments. S.B. and K.W.W. designed and carried out the TIL experiments. Y.C.
and J.K. designed and carried out the TEM and MSR pore analysis experiments. O.A.A.
contributed to the study design.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41563-018-0028-2.
Reprints and permissions information is available at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to D.J.M.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Letters
Methods
Materials, cell line and peptides. Branched 60K PEI (Sigma 181978), branched
2K PEI (Polysciences 06089), linear 25K PEI (Polysciences 23966) and linear 2K
PEI (Polysciences 24313) were diluted to 5.5 mg ml−1 in PBS, pH adjusted to 7.4
and stored away from direct light. Endotoxin-free OVA was used throughout the
study (Invivogen, vac-pova). OVA was diluted to 5 mg ml−1 in endotoxin-free water.
All antibodies were used according to the manufacturer’s protocols. SIINFEKL
and E749-57 tetramers were obtained from the Emory NIH Tetramer Core Facility.
TC-1 cells were generated in the laboratory of T.-C. Wu (Johns Hopkins University,
Baltimore, MD, USA, tested to be mycoplasma free) and maintained in RPMI
supplemented with 10% fetal bovine serum (FBS), 1% penicillin–streptomycin
and 50 μ​g  ml−1 G41824. The TC-1 cells were tested to be mycoplasma-free and not
further authenticated. B16F10 (ATCC) and CT26 cells (ATCC) were authenticated
and tested to be mycoplasma-free by ATCC maintained in DMEM supplemented
with 10% FBS and 1% penicillin–streptomycin according to protocols from
ATCC. Peptides used in this study are listed in Supplementary Table 4. All murine
peptides used in this study were synthesized to at least 95% purity from Peptide
2.0. All human neoantigen peptides were provided by Dr C. J. Wu (Dana Farber
Cancer Institute, Harvard Medical School, Boston, MA, USA). Peptides were first
dissolved in DMSO to 50 mg ml−1 and further diluted in endotoxin-free water. Flow
cytometry antibodies used in this study are listed in Supplementary Table 5.
Vaccine formulations. MSRs were synthesized as described previously13. All
vaccine components were incorporated after the MSRs were synthesized.
To formulate the MSR–PEI or cis MSR–PEI vaccine, 2 mg of the MSR
(50 mg ml−1 working solution in PBS) was adsorbed with 100 μ​g murine class B
CpG-ODN (sequence TCCATGACGTTCCTGACGTT, IDT, 10 mg ml−1 working
solution in dH2O) for 6–8 h at room temperature and subsequently lyophilized.
The MSR (2 mg; 50 mg ml−1 working solution in PBS) was adsorbed with various
doses (indicated in figure legends) of linear or branched PEI (in 100 μ​l PBS,
pH 7.4) at 37 °C for 15 min (MSR–PEI particles), and various doses of antigen
(typically 100 μ​g OVA or 50 μ​g peptides) were adsorbed onto MSR–PEI particles
for 1–3 h at 37 °C under shaking and subsequently lyophilized. Separately, 1 mg
of the MSRs (50 mg ml−1 working solution in PBS) was loaded with 1 μ​g murine
GM-CSF (Peprotech, 1 mg ml−1 working solution) for 1 h at 37 °C under shaking.
The MSR components (5 mg total) were combined and resuspended in cold PBS
(150 μ​l per vaccine) prior to immunization.
To formulate the trans MSR–PEI vaccines, 2 mg of the MSR (50 mg ml−1
working solution in PBS) was adsorbed with various doses of antigens (typically
100 μ​g OVA or 50 μ​g peptides) for 6–8 h at room temperature and subsequently
lyophilized. 2 mg of the MSR (50 mg ml−1 working solution in PBS) was adsorbed
with various doses (indicated in figure legends) of linear or branched PEI (in 100 μ​l
PBS, pH 7.4) at 37 °C for 15 min (MSR–PEI particles), and adsorbed with 100 μ​g
murine class B CpG-ODN (sequence TCCATGACGTTCCTGACGTT, IDT,
10 mg ml−1 working solution in dH2O) onto MSR–PEI particles for 1 h at 37 °C
under shaking. Separately, 1 mg of the MSRs (50 mg ml−1 working solution in PBS)
was loaded with 1 μ​g murine GM-CSF (Peprotech) for 1 h at 37 °C under shaking.
The MSR components (5 mg total) were combined and resuspended in cold PBS
(150 μ​l per vaccine) prior to immunization.
To prepare the MSR vaccines, 4 mg of the MSR (50 mg ml−1 working
solution in PBS) was adsorbed with 100 μ​g murine class B CpG-ODN
(sequence TCCATGACGTTCCTGACGTT, IDT, 10 mg ml−1 working solution in
dH2O) and the antigen (typically 100 μ​g OVA or 50 μ​g peptides) for 6–8 h at
room temperature under shaking, and subsequently lyophilized. Separately, 1 mg
of the MSRs (50 mg per ml working solution in PBS) were loaded with 1 μ​g
murine GM-CSF (Peprotech) for 1 h at 37 °C under shaking. The MSR
components were combined and resuspended in cold PBS (150 μ​l per vaccine)
prior to immunization.
To prepare the bolus vaccine, 100 μ​g murine class B CpG-ODN (sequence
TCCATGACGTTCCTGACGTT, IDT, 10 mg ml−1 working solution in dH2O), 1 μ​g
murine GM-CSF and the antigen (typically 50 μ​g peptides) were combined in
150 μ​l cold PBS.
Vaccine components were tested to be endotoxin free (Charles River
Laboratories) (Supplementary Table 2).
Animals. C57BL/6J and Balb/c mice were purchased from Jackson Laboratories
and housed at Harvard University. Female mice between 6 weeks and 10 weeks old
at the start of the experiments were used. All studies were carried outin accordance
with institutional guidelines approved by Harvard University’s Institutional Animal
Care and Use Committee (IACUC)31.
Immunization. MSR vaccines were resuspended in 150 μ​l of cold PBS immediately
before injection. They were injected, via an 18 G needle, subcutaneously in the
intrascapular region with the mouse under brief isoflurane anaesthesia.
Pore measurement. MSRs and MSR–PEI particles were first lyophilized. Nitrogen
sorption isotherms of the particles were measured at 77 K using the ASAP
2000 system (Micromeritics). To determine the surface area of the particles,
the Brunauer–Emmett–Teller method was used. To determine the pore size
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATuRE MATERIALS
distributions, the Barret–Joyner–Hallenda model from the adsorption branch of
the isotherm was used31.
Transmission electron microscopy (TEM). To prepare the samples for TEM
imaging, MSR and MSR–PEI particles were first diluted in 200-proof ethanol
(Sigma-Aldrich) to about 0.1 mg ml−1 and briefly sonicated. An aliquot of 5–20 μ​l
of the diluted particles was dripped onto formvar/carbon TEM grids (Ted Pella)
and dried at room temperature. The particles were imaged using the JEOL 2100
transmission electron microscope31.
Endotoxin measurement. Samples were analysed without dilution using EndoSafe
cartridge kits (Charles River Laboratories, catalogue no. PTS55005F). Samples were
assayed according to manufacturer protocols. Briefly, 25 μ​l of sample was added to
samples and control wells and the assay was run on an EndoSafe multi-cartridge
system (Charles River Laboratories).
Incorporation efficiency assays. To determine the incorporation efficiency of
peptides onto the MSRs, 50 μ​g of the indicated peptides was adsorbed onto 2 mg of
MSRs or MSR–PEI particles overnight. The MSRs were centrifuged at 2,000g for
10 min to separate the unbound peptides from the bound peptides. Supernatant
containing the unbound peptides was collected and the concentration of the
peptides was quantified using micro-BCA (ThermoScientific). Quantification of
peptide incorporation was confirmed using HPLC liquid chromatography–mass
spectrometry (LC-MS) (Supplementary Fig. 13a). Incorporation efficiency was
calculated as the total added to MSR minus the unbound portion. To ensure
the accuracy of this approach, the bound peptides were dissociated from the
MSRs with 40% Tween-20 under vigorous shaking for 4 h at room temperature.
Subsequently, the concentrations of the peptides in the bound portion and the
unbound portion were measured using LC-MS, and it was confirmed that nearly
100% of the peptides initially introduced to the particles were accounted for
(Supplementary Fig. 13b).
To determine the incorporation efficiency of PEI onto the MSRs, PEI at
the indicated doses was adsorbed onto MSR for 15 min at 37 °C. Subsequently,
the MSRs were centrifuged at 2000g for 10 min to separate the bound PEI from
the unbound PEI. Supernatant was collected and the concentration of PEI was
measured using fluoraldehyde (Sigma-Aldrich). Incorporation efficiency was
calculated as the total minus the unbound portion.
In vitro release studies. A 2 mg portion of MSRs was loaded with either 2 μ​g
of GM-CSF, 100 μ​g of CpG or 10 μ​g of rhodamine–PEI (L25K). The individual
vaccine components were then resuspended in 1 ml of release medium, which
is composed of RPMI (Sigma-Aldrich) supplemented with 1% penicillin–
streptomycin and 10% heat-inactivated FBS (Sigma-Aldrich), in low-binding
Eppendorf tubes. The release study was then carried out at 37 °C under gentle
shaking. Periodically, supernatant containing the released component was
collected by pelleting the particles at 2000g for 10 min. The release medium was
subsequently replaced. For analysis, GM-CSF content was determined using ELISA
(R&D). CpG content was detected using OliGreen (ThermoFisher)13.
BMDC differentiation, cytokine production and antigen presentation assays.
DCs were differentiated from bone marrow cells isolated from C57Bl/6J mice
(Jackson Laboratories). Bone marrow was collected from the femur and tibia
bones, and single-cell suspensions were cultured in RPMI (Sigma-Aldrich)
supplemented with 10% heat-inactivated FBS (Sigma-Aldrich), 1% penicillin–
streptomycin, 50 µ​M β​-mercaptoethanol and 20 ng ml−1 murine GM-CSF
(Peprotech). Non-adherent and loosely adherent cells between days 7 and 10
of differentiation were collected and used for the studies. Differentiation was
confirmed by staining the cells with anti-mouse CD11c and CD11b. To assess
the uptake kinetics of PEI by BMDCs, 0.5 ×​  106 ml−1 immature BMDCs were
plated onto non-tissue-culture-treated 12-well dishes and stimulated with 7 μ​g
of rhodamine-labelled L25K PEI for various lengths of time (0 h, 2 h, 6 h, 24 h,
72 h). BMDCs were harvested, resuspended in FACS buffer and the rhodamine
signal analsed using flow cytometry (BD Biosciences). To assess the BMDCs’
activity after PEI stimulation, 1 ×​  106 ml−1 immature BMDCs were plated onto
tissue-culture-treated 12-well dishes and stimulated with various doses of L25K
or B60K PEI or MSR–PEI for 24 h, in either the presence or absence of 20 ng ml−1
lipopolysaccharide. BMDCs were harvested, resuspended in FACS buffer and
stained with 7-AAD (eBioscience), anti-CD11c, anti-CD86 and anti-MHCII for
15 min on ice. BMDCs were then washed in FACS buffer and analysed using flow
cytometry. Separately, supernatant was collected and analysed for IL-1β​,
IL-6, TNF-α​and IFN-γ​using ELISA according to the manufacturer’s protocols
(eBioscience). To assess BMDC cross-presentation, 0.5 ×​  106 ml−1 BMDCs were
pulsed with 20 μ​g OVA (Invivogen) either alone or in the presence of 5 μ​g or 10 μ​g
B60K PEI for 48 h. BMDCs were then harvested and stained using 7-AAD
(Biolegend), anti-CD11c and anti-SIINFEKL/H-2Kb for 15 min on ice, washed and
analysed using flow cytometry (LSRFortessa, BD)32.
Lysosomal rupture in BMDCs. Immature BMDCs (1 ×​  106 ml−1) were plated
onto tissue-culture-treated 12-well dishes. BMDCs were stimulated with 40 μ​g of
Nature Materials | www.nature.com/naturematerials
NATuRE MATERIALS
MSRs, 40 μ​g of MSR–PEI particles (12.5 μ​g of L25K-PEI per mg MSRs) or 2 μ​g free
PEI (L25K) for 24 h. Subsequently, BMDCs were incubated with acridine orange
(ThermoFisher, diluted 1:20,000) for 4 h. Acridine orange signal was assessed
using flow cytometry at 488 nm (excitation) and 650–690 nm (emission)
(LSRFortessa, BD)21.
TLR-4 neutralization. 1 ×​  106 ml−1 immature BMDCs were plated onto tissue-
culture-treated 12-well dishes and treated with TLR4 (CD284)/MD2 monoclonal
antibody (Biolegend, clone MTS510) for 1 h at 37 °C. Subsequently, the BMDCs
were stimulated with 40 μ​g of MSR or MSR–PEI particles (12.5 μ​g PEI per mg
MSRs) for 24 h. BMDCs were collected, stained with anti-CD11c, anti-CD86 and
7AAD, and analysed using flow cytometry (LSRFortessa, BD).
Therapeutic tumour studies. Cells were used at passage number 5 or earlier and
passaged at least twice prior to inoculation. In the subcutaneous therapeutic TC-1
model, female C57BL/6 mice were injected with a single-cell suspension of 2 ×​  105
TC-1 cells in 100 μ​l cold PBS in the back of the neck on day 0. In the subcutaneous
therapeutic B16F10 model, female C57BL/6 mice were injected with a single-cell
suspension of 1 ×​  105 B16F10 cells in 100 μ​l cold PBS in the back of the neck on
day 0. Mice were immunized on indicated days, and tumour area was monitored
using a calliper: longest surface length (a) and width (b), and the tumour size was
expressed as area (ab). Mice were killed when the longest side reached 2 cm or
according to IACUC standards. In the therapeutic B16F10 or CT26 lung metastasis
model, mice were injected intravenously with 2 ×​  105 B16F10 or CT26 cells in
100 μ​l cold PBS on day 0. Lungs were excised on day 16 and fixed in Fekete’s
solution. Lung metastasis nodules were then manually enumerated. For anti-
CTLA4 therapies, 100 μ​g per mouse of anti-mouse CTLA-4 antibodies (BioXcell,
clone 9D9) were administered intraperitoneally on days 1, 3, 6 and 9 after
tumour inoculation.
CD8+ T-cell depletion. CD8+ T cells were depleted by injecting 100 μ​g of
anti-CD8a monoclonal antibody (BioXCell, clone 2.43) intraperitoneally twice
weekly, starting the day before vaccination33. Depletion was confirmed using
flow cytometry.
Scaffold explant and analysis. Scaffold tissues were surgically removed from the
animals after euthanasia. They were first processed using mechanical disruption,
and subsequently digested in 250 U ml−1 collagenase IV (Worthington) in RPMI
for 30 min. at 37 °C under gentle shaking. Cell and tissue suspensions were filtered
through a 40 μ​m cell strainer to separate single cells from large MSR particles.
The cells and small remaining MSR particles were pelleted, washed three times
and counted manually using a haemocytometer. Dead cells were visualized using
trypan blue (ThermoFisher). Subsequently, cells were stained with anti-mouse
CD11c, CD86, CCR7 and SIINFEKL/H-2Kb monoclonal antibodies in FACS
buffer (PBS supplemented with 1% BSA and 0.1% sodium azide) for 15 min on
ice. Dead cells and MSR particles were stained and excluded from analysis using
7-AAD (eBioscience) or a fixable viability dye (eBioscience). Finally, cells were
washed three times in FACS buffer and analysed using flow cytometry
(LSRFortessa, BD)13.
Lymph node analysis. Vaccine dLNs were surgically removed from the animals.
They were then processed through mechanical disruption and digested for 30 min
at 37 °C in RPMI containing 0.5 mg ml−1 collagenase 4 and 0.1 mg ml−1 DNAse.
Cells were then filtered through a 40 μ​m cell and washed in cold PBS. Single-cell
suspensions were enumerated (Beckman-Coulter) and stained with an Alexa Fluor
780 dead exclusion dye (eBioscience), anti-mouse CD11c, CD11b, CD86, MHC-II
and F4/80 for 15 min on ice. DQ-OVA was visualized in the FITC channel. The
cells were then washed and analysed using flow cytometry (LSRFortessa, BD).
Tetramer analysis and peptide restimulation of circulating PBMCs. About
100 μ​l of peripheral blood was collected into heparin-coated tubes (BD) and kept
on ice. Red blood cells (RBCs) were removed by adding 1 ml RBC lysis buffer
(BioLegend) for 1 min at room temperature; lysis was quenched by adding 9 ml
cold PBS directly. Cells were pelleted and washed in FACS staining buffer. To carry
out tetramer staining, blood cells were incubated with 50 μ​l of 7 μ​g  ml−1 Alexa
Fluor 647 conjugated peptide-MHC tetramers (NIH Tetramer Core) for 20 min
at 37 °C. Subsequently, cells were pelleted and stained with 7-AAD, anti-mouse
CD3e and CD8a for 15 mins on ice. Finally, cells were washed three times in FACS
staining buffer and analysed using flow cytometry (LSRFortessa). To carry out
peptide restimulation, blood cells were pulsed with 2 μ​g  ml−1 of various MHC-I
restricted peptides for 1–1.5 h at 37 °C in T-cell medium (RPMI supplemented with
Nature Materials | www.nature.com/naturematerials
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Letters
110 mg l−1 sodium pyruvate, 5 mM HEPES, 50 μ​M β​-mercaptoethanol, 10% heat-
inactivated FBS and 1% penicillin–streptomycin). Subsequently, cytokine secretion
was stopped by directly adding Golgi Plug (Biolegend) to the culture. Cells were
cultured for another 4 h at 37 °C. Cells were pelleted and stained with Alexa Fluor
780 dead exclusion dye, anti-mouse CD3e, CD8a and CD4 for 15 min on ice. Cells
were then fixed and permeabilized (eBioscience 88-8824-00), and stained with
anti-mouse IFN-γ​for 30 min at 4 °C. Finally, cells were washed and analysed using
flow cytometry (LSRFortessa, BD).
Peptide restimulation of splenic CD8+ T cells. To isolate splenic DCs, naïve
spleens were retrieved and digested in RPMI supplemented with 1 mg ml−1
collagenase 4 (Worthington) and 0.1 mg ml−1 DNase (New England Biolabs)
for 30 min at 37 °C. RBCs were lysed using RBC lysis buffer (Biolegend). Dead
cells were removed (Miltenyi Biotec), and splenic DCs were obtained using a
CD11c+ magnetic sorting kit (Miltenyi Biotec). DCs were pulsed with 5 μ​g  ml−1
of M27, M30 or a control peptide. Separately, spleens from vaccinated mice were
retrieved (8 days after vaccination). Splenocytes were isolated from the spleens by
physically grinding the spleens on a cell filtration device in RPMI supplemented
with 2% FBS. RBCs were lysed and CD8+ or CD4+ T cells were then obtained
using a negative CD8+ or CD4+ T-cell selection kit (Miltenyi Biotec). The sorted
T cells were labelled with CFSE (ThermoFisher). T cells and DCs were cultured
together at a 10:1 (T:DC) ratio for 3 days at 37 °C to allow for antigen-specific T-cell
proliferation. On day 3, cytokine secretion was stopped by incubating with Golgi
Plug (Biolegend) for 4 h at 37 °C. Subsequently, cells were collected and stained
with Alexa Fluor 780 dead exclusion dye, anti-mouse CD3e, CD8a and CD4 for
15 min on ice before fixing and permeabilizing (eBioscience 88-8824-00). Cells
were then stained with anti-mouse IFN-γ​for 30 min at 4 °C and analysed using
flow cytometry (LSRFortessa, BD)5,34.
TIL analysis. Tumours were explanted and cut into small pieces in 5 ml of RPM.
The samples were then digested in 15 ml RPMI supplemented with 2% FCS,
50 U ml−1 collagenase IV (Invitrogen), 100 μ​g  ml−1 hyaluronidase and 20 U ml−1
DNase (Roche) for 2 h at 37 °C using gentleMACS dissociators (Miltenyi Biotech).
Suspensions were filtered through a 40 µ​M strainer and washed three times with
PBS. Lymphocyte population was enriched through gradient centrifugation at
50g for 5 min, repeated three times. Supernatant was collected and stained with a
Zombie-UV dead exclusion dye, anti-mouse CD45, CD3e, CD4, CD8a, CD62L and
CD44 for 15 min on ice before fixing and permeabilizing (eBioscience 00-5523-00).
Cells were then stained with anti-mouse IFN-γ​, TNF-α​and granzyme B for 30 min
at 4 °C and analysed using flow cytometry (LSRFortessa, BD)35.
Statistical analysis. All values in the present study are expressed as mean ±​  s.d.
unless otherwise indicated in the figure legends. Statistical analysis was performed
using GraphPad Prism. The significance between two groups was analysed by a
two-tailed Student t-test. Sample variance was tested using the F-test. For multiple
comparisons, a one-way or a two-way ANOVA test was used. In all cases, a P value
of less than 0.05 was considered significant. Details for statistical analyses for each
comparison are reported in Supplementary Table 6.
Life Sciences Reporting Summary. Further information on experimental design is
available in the Life Sciences Reporting Summary.
Data availability. The datasets generated during and/or analysed during the
current study are available from the corresponding author on reasonable request.
References
31. Li, W. A. et al. The effect of surface modification of mesoporous silica
micro-rod scaffold on immune cell activation and infiltration. Biomaterials
83, 249–256 (2016).
32. Kim, J. et al. Effect of pore structure of macroporous poly (lactide-co-
glycolide) scaffolds on the in vivo enrichment of dendritic cells.
ACS Appl. Mater. Interf. 6, 8505–8512 (2014).
33. Moynihan, K. D. et al. Eradication of large established tumors in mice by
combination immunotherapy that engages innate and adaptive immune
responses. Nat. Med. 22, 1402–1410 (2016).
34. Höpken, U. E. et al. The ratio between dendritic cells and T cells determines
the outcome of their encounter: proliferation versus deletion. Eur. J. Immunol.
35, 2851–2863 (2005).
35. Zhou, P. et al. In vivo discovery of immunotherapy targets in the tumor
microenvironment. Nature 506, 52 (2014).
 nature research | life sciences reporting summary
Corresponding author(s): David Mooney
Initial submission Revised version Final submission

Life Sciences Reporting Summary

Nature Research wishes to improve the reproducibility of the work that we publish. This form is intended for publication with all accepted life
science papers and provides structure for consistency and transparency in reporting. Every life science submission will use this form; some list
items might not apply to an individual manuscript, but all fields must be completed for clarity.
For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research
policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

` Experimental design
1. Sample size
Describe how sample size was determined. Sample sizes were determined to allow the statistical significance of differences of
50% or greater, and according to similar studies conducted in the field. The specific
sample size required depended on the experiment.
2. Data exclusions
Describe any data exclusions. No samples were excluded
3. Replication
Describe whether the experimental findings were Results from each experiment were reproduced in the lab at least twice
reliably reproduced.
4. Randomization
Describe how samples/organisms/participants were Animals were randomly sorted to various experimental or control groups.
allocated into experimental groups.
5. Blinding
Describe whether the investigators were blinded to No formal blinding was used, but the various steps in key studies were performed
group allocation during data collection and/or analysis. by different investigators, who were each independent of the steps performed by
the other investigator.
Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

6. Statistical parameters
For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or in the
Methods section if additional space is needed).

n/a Confirmed

The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.)
A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same
sample was measured repeatedly
A statement indicating how many times each experiment was replicated
The statistical test(s) used and whether they are one- or two-sided (note: only common tests should be described solely by name; more
complex techniques should be described in the Methods section)
A description of any assumptions or corrections, such as an adjustment for multiple comparisons
The test results (e.g. P values) given as exact values whenever possible and with confidence intervals noted
June 2017

A clear description of statistics including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range)
Clearly defined error bars

See the web collection on statistics for biologists for further resources and guidance.

1
` Software

nature research | life sciences reporting summary

Policy information about availability of computer code
7. Software
Describe the software used to analyze the data in this Statistical analysis was performed using GraphPad Prism.
study.
For manuscripts utilizing custom algorithms or software that are central to the paper but not yet described in the published literature, software must be made
available to editors and reviewers upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). Nature Methods guidance for
providing algorithms and software for publication provides further information on this topic.

` Materials and reagents

Policy information about availability of materials
8. Materials availability
Indicate whether there are restrictions on availability of All materials used are readily available from the authors or common commercial
unique materials or if these materials are only available sources (provided in the manuscript).
for distribution by a for-profit company.
9. Antibodies
Describe the antibodies used and how they were validated List of antibodies used, including manufacturing source and clone, were provided
for use in the system under study (i.e. assay and species). in the manuscript.

10. Eukaryotic cell lines

a. State the source of each eukaryotic cell line used. B16F10 - ATCC; TC-1 - Johns Hopkins University, CT26 - ATCC

b. Describe the method of cell line authentication used. ATCC Cell lines were authenticated using STR profiling by ATCC, no further
authentication was performed

c. Report whether the cell lines were tested for TC-1 cells were tested free for mycoplasma contamination
mycoplasma contamination.

d. If any of the cell lines used are listed in the database N/A
of commonly misidentified cell lines maintained by
ICLAC, provide a scientific rationale for their use.

` Animals and human research participants

Policy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines
11. Description of research animals
Provide details on animals and/or animal-derived C57BL/6J or Balb/c mice were obtained from Jackson Laboratory. Female, 6-10
materials used in the study. weeks old at the start of experiment were used. All animal studies were performed
in accordance with NIH guidelines, under approval of Harvard University’s
Institutional Animal Care and Use Committee.

Policy information about studies involving human research participants

12. Description of human research participants
Describe the covariate-relevant population N/A
characteristics of the human research participants.
June 2017

2
 nature research | flow cytometry reporting summary
Corresponding author(s): David Mooney
Initial submission Revised version Final submission

Flow Cytometry Reporting Summary

Form fields will expand as needed. Please do not leave fields blank.

` Data presentation
For all flow cytometry data, confirm that:
1. The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
2. The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of
identical markers).
3. All plots are contour plots with outliers or pseudocolor plots.
4. A numerical value for number of cells or percentage (with statistics) is provided.

` Methodological details
5. Describe the sample preparation. For scaffold samples, the scaffolds were processed through mechanical
disruption and digested for 30 min. at 37 C in 250 U/ml Collagenase IV in
RPMI. The resulting cell suspension was then filtered through a 40μm cell
strainer to isolate the cells from the larger sized MSRs. The cells and small
remaining MSR particles were pelleted and washed with cold PBS.
For LN samples, they were processed through mechanical disruption and
digested for 30 min. at 37 °C in RPMI containing 0.5mg/ml Collagenase 4
and 0.1 mg/ml DNAse. Cells were then filtered through a 40μm cell and
washed in cold PBS.
Tumor sample processing is detailed in the methods section.
For all samples, cells were first stained with a dead exclusion dye,
followed by antibodies against surface antigens. In some experiments, cells
were subsequently fixed, permeablized and stained for intracellular
antigens.

6. Identify the instrument used for data collection. LSRFortessa, BD

7. Describe the software used to collect and analyze Flowjo 7.6 and Flowjo 10.4
the flow cytometry data.

8. Describe the abundance of the relevant cell No sorting was performed

populations within post-sort fractions.

9. Describe the gating strategy used. In general, cells were first gated on FSC/SSC. Singlet cells were gated using
FSC-H and FSC-A. Dead cells were then excluded and further surface and
intracellular antigen gating was performed on the live cell population.
Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
June 2017