Accepted Manuscript

Dendritic cell-derived exosomes elicit tumor regression in autochthonous hep-atocellular carcinoma mouse models

Zhen Lu, Bingfeng Zuo, Renwei Jing, Xianjun Gao, Quan Rao, Zhili Liu, HanQi, Hongxing Guo, HaiFang Yin

PII: S0168-8278(17)32055-XDOI: http://dx.doi.org/10.1016/j.jhep.2017.05.019Reference: JHEPAT 6544

To appear in: Journal of Hepatology

Received Date: 7 December 2016Revised Date: 12 May 2017Accepted Date: 15 May 2017

Please cite this article as: Lu, Z., Zuo, B., Jing, R., Gao, X., Rao, Q., Liu, Z., Qi, H., Guo, H., Yin, H., Dendriticcell-derived exosomes elicit tumor regression in autochthonous hepatocellular carcinoma mouse models, Journalof Hepatology (2017), doi: http://dx.doi.org/10.1016/j.jhep.2017.05.019

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting proof before it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Dendritic cell-derived exosomes elicit tumor regression in autochthonous

hepatocellular carcinoma mouse models

Zhen Lu1§, Bingfeng Zuo1§, Renwei Jing1, Xianjun Gao1, Quan Rao1, 2, Zhili Liu1,

Han Qi1, Hongxing Guo

2, HaiFang Yin

1*

1 Department of Cell Biology and Research Centre of Basic Medical Science, Key

Laboratory of Immune Microenvironment and Disease (Ministry of Education),

Tianjin Medical University, Qixiangtai Road, Heping District, Tianjin, 300070, China

2 Third Central Clinical College, Tianjin Medical University, Jintang Road, Hedong

District, Tianjin, 300170, China

Correspondence

HaiFang Yin

Email: haifangyin@tmu.edu.cn

Tel: +86 (0)22 83336537

Fax: +86 (0)22 83336537

Address: Tianjin Medical University, Qixiangtai Road, Heping District, Tianjin,

300070, China

§ These authors contributed equally as joint first authors.

Key words: Exosome; hepatocellular carcinoma; immunotherapy; alpha fetoprotein

2

Electronic Word count:

5512

Number of figures and tables:

8 figures and 0 tables

Financial Support: This research was supported by National Natural Science

Foundation of China (no. 81672124, 81273420, 81501531 and 81671528), Key

Program of National Natural Science Foundation of Tianjin (no. 14JCZDJC36000),

Key Program of Tianjin Municipal Health Bureau (no. 2013-GG-05), the Key

Laboratory of Immune Microenvironment and Disease (Tianjin Medical University),

Ministry of Education, and Tianjin Municipal 13th

five-year plan (Tianjin Medical

University Talent Project).

Conflicts of Interest

No potential conflicts of interest were disclosed.

Author Contributions

H.Y. designed the study. Z. L., B. Z., R. J., X.G., Q. R., Z. L., H. Q. and H. G.

performed the experiments. Z.L. and H.Y. analyzed the data. H.Y. supervised the

study and wrote the manuscript. All authors contributed to the revision of the

manuscript and approved the final version.

3

Abstract

Background & Aims: Dendritic cell (DC)-derived exosomes (DEXs) form a new

class of vaccines for cancer immunotherapy. However their potency in hepatocellular

carcinoma (HCC), a life-threatening malignancy with limited treatment options in the

clinic and responds poorly to immunotherapy, remains to be investigated.

Methods: Exosomes derived from α-fetoprotein (AFP) - expressing DCs (DEXAFP)

were investigated in three different HCC mouse models systemically and tumor

growth and microenvironment were monitored.

Results: DEXAFP elicited strong antigen-specific immune responses and resulted in

significant tumor growth retardation and prolonged survival rates in mice with ectopic,

orthotopic and carcinogen-induced HCC tumors that displayed antigenic and

pathological heterogeneity. The tumor microenvironment was improved in

DEXAFP-treated HCC mice, demonstrated by significantly more γ-interferon (IFN-γ)

-expressing CD8+ T lymphocytes, elevated levels of IFN-γ and interleukin-2 (IL-2),

and fewer CD25+Foxp3

+ regulatory T (Treg) cells and decreased levels of

interleukin-10 (IL-10) and transforming growth factor-β (TGF-β) in tumor sites. Lack

of efficacy in athymic nude mice and CD8+ T cell-depleted mice showed that T cells

contribute to DEXAFP-mediated antitumor function. Dynamic examination of the

antitumor efficacy and the immune microenvironment in DEXAFP-treated orthotopic

HCC mice at different time-points revealed a positive correlation between tumor

suppression and immune microenvironment.

Conclusions: Our findings provide evidence that AFP-enriched DEXs can trigger

potent antigen-specific antitumor immune responses and reshape the tumor

4

microenvironment in HCC mice and thus provide a cell-free vaccine option for HCC

immunotherapy.

Electronic word count:

230

Lay Summary

Dendritic cell (DC)-derived exosomes (DEXs) form a new class of vaccines for

cancer immunotherapy. However their potency in hepatocellular carcinoma (HCC)

remains unknown. Here, we investigated exosomes from HCC antigen-expressing

DCs in three different HCC mouse models and proved their feasibility and capability

of treating HCC, and thus provide a cell-free vaccine for HCC immunotherapy.

5

Introduction

Hepatocellular carcinoma (HCC) presents as one of the most lethal malignancies

worldwide due to its aggressive nature, high mortality and low response rates to

treatments in the clinic [1]. A multitude of therapeutic approaches are under intensive

investigation. Among them, chemotherapeutic and radiotherapeutic interventions have

been extensively used, however the survival benefit is limited [2]. Resection surgery

is effective for early-stage HCC patients, but is only amenable to a subpopulation of

patients and HCC frequently recurs. Dendritic cell (DC)-based immunotherapy is

promising, but is hampered by a costly and tedious preparation protocol and relies on

the availability of patient-derived DCs due to its short shelf-life [3-6].

Exosomes are small extracellular vesicles with sizes of 30-100 nm (in diameter)

secreted by a wide variety of cells [7, 8]. Exosomes from DCs (DEXs) bear abundant

signature proteins from their parental cells, such as major histocompatibility complex

class I (MHC-I), class II (MHC-II) and co-stimulatory molecules, and thus have been

employed as a cell-free alternative option to DC vaccines for cancer immunotherapy

[9]. Zitvogel et al demonstrated that exosomes derived from tumor antigenic

peptide-pulsed DCs could elicit strong immune responses and tumor suppression in

mastocytoma and mammary carcinoma mice [10], showing that DEXs can replace

DCs as effective vaccines. This concept was adopted in other tumor models [11, 12].

Importantly, recent clinical trials on advanced melanoma patients with DEXs have

demonstrated encouraging results, highlighting DEXs’ applicability [13]. Unlike DCs,

6

DEXs are stable vesicles with long shelf-lives when frozen and can be

tailor-manufactured from genetically modified cell lines.

Here we examine the feasibility of exosomes derived from α-fetoprotein

(AFP)-expressing DCs (DEXAFP), recovered by high speed ultracentrifugation and

0.22 µm diafiltration, to stimulate antigen-specific immune response in HCC models.

AFP is a fetal liver protein and highly elevated in 50-80% of HCC patients and thus

has been extensively used as a HCC antigen for diagnostics and therapeutics [14]. Our

study demonstrates that DEXAFP could induce strong antigen-specific immune

responses and tumor retardation in ectopic and orthotopic HCC mice, particularly in

diethylnitrosamine (DENA)-induced autochthonous HCC mice. The tumor immune

microenvironment was reshaped from immuno-inhibitory to immuno-stimulatory

after repeated administration of DEXAFP in orthotopic HCC mice based on the

cytokine profiles, and γ-interferon (IFN-γ)-expressing CD8+ cytotoxic T lymphocytes

(CTLs) contributed substantially to the effect. Our finding provides evidence for the

first time that exosomes from HCC antigen-modified DCs could be used as cell-free

vaccines for HCC and thus opens a new avenue for HCC immunotherapy. More

importantly, the long shelf-life and availability of large amounts of DEXs affords

practical advantages over DCs in clinical deployment of DC-based vaccines.

Materials and Methods

Mice

7

C57BL/6 wild-type (H-2b) and BALB/C nude mice (H-2

d) (6-8 weeks old, no gender

preference) were used in all experiments (5 or 10 mice were used in each group for

ectopic or orthotopic studies, respectively, and the experiments were repeated for 3

times unless otherwise specified). All the animal experiments were carried out in the

animal unit, Tianjin Medical University (Tianjin, China) according to procedures

authorized and specifically approved by the institutional ethical committee (Permit

Number: SYXK 2009–0001).

Statistical analysis

All data are reported as mean values ± SEM. Statistical differences between treatment

and control groups were evaluated by SigmaStat (Version 3.5; Systat Software,

London, UK), with significance set at p<0.05. Both parametric and non-parametric

analyses were applied, in which the Mann-Whitney Rank Sum Test (Mann-Whitney U

test) was used for samples on a non-normal distribution whereas a two-tailed t test

was performed for samples with a normal distribution, respectively. Sample size was

determined by PASS (Power Analysis and Sample Size) (Version 11, NCSS, LLC, UT,

US). The power analysis with our pre-specified sample size (n=5) in orthotopic HCC

mice showed an associated power of 0.921 when the significance set was p<0.05.

Further detailed methods are reported in the Supplementary Material.

Results

Enrichment of AFP and immune molecules in DEXAFP

8

To test the ability of DEXs to elicit antitumor immunity in HCC, we infected mouse

DC cell line (DC2.4; H-2b, MHC haplotype) with a lentivirus expressing murine AFP

gene (DCAFP). Similar levels of AFP protein expression were observed in DCs serially

infected 3 or 4 times, therefore DCs infected 3 times were used subsequently

(Supplementary Fig. 1A). MHC-I, co-stimulatory molecules CD80 and CD86 were

up-regulated in DCAFP, though not for MHC-II and intercellular cell adhesion

molecule (ICAM) (Supplementary Fig. 1B), suggesting that AFP expression promoted

DC maturation. Unsurprisingly, compared to DCs, DCAFP showed stronger capacity to

prime and elicit T cell expansion based on CFSE assay (Supplementary Fig. 1C) and

triggered higher levels of IFN-γ and IL-2 secretion (Supplementary Fig. 1D), resulting

in greater antigen-specific cytolysis in Hepa1-6 cells (H-2b) (Supplementary Fig.

1E-F). Similarly, DCAFP could also significantly inhibit tumor growth and prolong

survival rate in ectopic HCC mice therapeutically after tumor establishment

(Supplementary Fig. 2A-B) and prophylactically before tumor challenge

(Supplementary Fig. 2C-D), indicating that AFP confers DC antigen-specific

immunity.

Exosomes derived from DCAFP (DEXAFP) expressed AFP and exosomal marker Alix

but not mitochondrial cytochrome C (Fig. 1A), with a characteristic saucer-cup shape

under transmission electron microscopy (Fig. 1B). The yield of DEXAFP was 1.6-1.8

µg of protein per million cells in 24 h, similar to unmodified DCs, suggesting that

AFP expression did not affect exosome production. Strikingly, MHC-I, II and

9

co-stimulatory molecules were up-regulated on DEXAFP compared to DEX (Fig. 1C),

indicating that AFP expression promotes the enrichment of immune molecules on

DEXs.

DEXAFP promotes antigen-specific immune responses in vitro

To examine if DEXAFP can trigger T cell-mediated antigen-specific cytolysis against

Hepa1-6 cells (H-2b), we intravenously administered 40 µg DEXAFP, DEX derived

from DCs infected with control lentivirus (DEXVEC) or DEX into MHC-matched

C57BL6 mice (H-2b) weekly for 3 weeks. Splenic T cells were harvested from

immunized mice 3 days after the last injection and re-stimulated with 40 µg DEXAFP

or DEX for 72 h. Significantly increased levels of IFN-γ, a functional parameter of T

cell immune response [15], and IL-2 were secreted from T cells derived from

DEXAFP-immunized mice after re-stimulation with DEXAFP, compared to all other

groups (Fig. 2A). DEXAFP also increased antigen-specific cytolysis of Hepa1-6 cells

(60.8+3.6%), compared to DEXVEC (26.3+2.9%), DEX (21.6+10.2%) or

un-stimulated T controls (21+2.6%) (Fig.2B). To demonstrate the antigen-specificity

elicited by DEXAFP, DEXAFP and DEX derived from green fluorescence protein

(GFP)-expressing DCs (DEXGFP) were administered into C57BL6 mice under the

identical conditions and splenic T cells were harvested (Supplementary Fig. 3A). As

expected, restimulation with DEXAFP or AFP212, an immunodominant epitope for

HCC identified previously [16], significantly increased levels of IFN-γ and IL-2

secretion in DEXAFP-immunized T cells, but not in DEXGFP-immunized T cells

10

(Supplementary Fig. 3B). Consistently, significantly higher cytolysis rate against

Hepa1-6 cells at the effector T : target cell (E: T) ratio of 10:1 was only observed in

DEXAFP–immunized T cells after restimulation with DEXAFP or AFP212 compared to

other groups (Supplementary Fig. 3C). Furthermore, cross-protective cytolysis were

not seen with allogeneic H22 (H-2d), a murine HCC cell line lacking AFP expression,

pancreatic cancer (Panc02, H-2b) and Lewis lung cancer cells (LLC1, H-2

b) after

co-incubation with DEXAFP-immunized T cells re-stimulated with DEXAFP (Fig. 2C).

These findings support the conclusion that DEXAFP can induce an AFP-specific

activation of T cells in vivo and that the activated T cells specifically target

AFP-expressing cells.

Levels of IFN-γ and IL-2 secreted by DEXAFP-treated T cells also increased

concordantly with escalating amounts of DEXAFP stimulation (Supplementary Fig.

4A). Higher E: T ratios showed higher cytolysis rates against Hepa1-6

(Supplementary Fig. 4B).

DEXAFP induces effective tumor suppression in ectopic HCC mice

To investigate if DEXAFP can trigger tumor rejection in vivo, we injected DEXAFP (40

µg) intravenously into ectopic MHC-matched tumor-bearing C57BL6 mice (H-2b)

(0.44+0.05 cm longitudinal diameter) weekly for 3 weeks [17]. Significantly slower

tumor growth was detected in DEXAFP-treated mice compared to DEX and control

groups on day 13 (P<0.001), 16 (P=0.011), 19 (P=0.011), 22 (P=0.002), 25, 28 and 31

11

(P<0.001, n=15) after inoculation (Fig. 3A). Importantly, prolonged survival rate was

achieved in 100% DEXAFP-treated mice at 57 days after tumor challenge, whereas no

mice survived in DEX or PBS treatment groups (Fig. 3B), indicating that DEXAFP is

potent at eliciting tumor-specific antitumor immunity in ectopic HCC mice.

Different administration routes might affect the immunotherapeutic outcome, thus we

compared intravenous and subcutaneous injections for DEXAFP in ectopic HCC mice

under an identical dosing condition (40 µg/mouse/week for 3 weeks). Significant

retardation of tumor growth was observed with both intravenous and subcutaneous

injections (Fig. 3C). Interestingly, a slightly stronger immune response with

prolonged survival rate was observed in 100% DEXAFP-treated mice with intravenous

injection at 62 days after tumor challenge, whereas 80% of the DEXAFP-treated mice

with subcutaneous injection survived (Fig. 3D). As intravenous delivery appears to be

slightly more beneficial in prolonging life-span, we chose intravenous injection for

subsequent studies. To examine whether a similar antigen-specific antitumor effect to

DCAFP can be achieved with bone marrow-derived DCs (BMDC), we infected

BMDCs with the same AFP-expressing lentivirus (BMDCAFP) and harvested

exosomes from BMDCAFP (BMDEXAFP) (Supplementary Fig. 5A). Unsurprisingly, a

similar antitumor effect to DEXAFP was observed in BMDEXAFP-treated ectopic HCC

mice under identical dosing conditions (Supplementary Fig. 5B). Further examination

on the antigen-specific effect of BMDEXAFP confirmed that BMDEXAFP elicit an

AFP-specific immune response in HCC mice (Supplementary Fig. 5C-D)

12

DEXAFP suppresses tumor growth and improves tumor microenvironment in

orthotopic HCC mice

It is well-recognized that HCC is notoriously difficult to treat as the tumor promotes

an immunotolerogenic environment that precludes an anti-HCC immune reaction [18].

To better mimic the tumor microenvironment, C57BL6 mice bearing

day-7-established orthotopic HCC were used. DEXAFP (150 µg) were administered

intravenously into orthotopic HCC mice weekly for 3 weeks. Tumor growth was

significantly reduced in DEXAFP-treated mice at 26 days after inoculation, with

average tumor sizes of 0.59+0.07, 1.13+0.14 or 1.32+0.16 cm for DEXAFP, DEX or

PBS groups respectively (Fig. 4A-B). DEXAFP treatment also significantly prolonged

the survival (Fig. 4C). Notably, the same effect was also established with DEXAFP in

day-5-established orthotopic HCC mice under a slightly different dosing regimen (150

µg, 4 intravenous injections every 4 days) (Supplementary Fig. 6), showing that

DEXAFP can elicit a strong antitumor immune response in orthotopic HCC mice.

Importantly, repeated administration of DEXAFP resulted in significant improvement in

orthotopic HCC microenvironment, demonstrated by significantly more IFN-γ and

IL-2 (Fig. 4D) and significantly less immuno-inhibitory cytokines such as

interleukin-10 (IL-10) and transforming growth factor-β (TGF-β) (Fig. 4E). These

results suggest that DEXAFP treatment altered the tumor milieu from

immuno-inhibitory to immuno-stimulatory, which may improve the prognosis of HCC

patients [19].

13

T cell activation contributes to the potency of DEXAFP

Given the striking antitumor effect of DEXAFP on immunocompetent mice, we next

examined whether DEXAFP can mount antitumor immunity in immunodeficient

BALB/c nude mice (H-2d), a model for T cell deficiency due to the lack of thymus

[20]. The same dosing regimen used in immunocompetent mice was applied in

immunodeficient nude mice (BALB/C) with day-5-established ectopic tumors. Tumors

were significantly retarded in immunocompetent C57BL6 mice treated with DEXAFP

but not for immunodeficient nude mice (Fig. 5A), suggesting that T cell activation is

responsible for DEXAFP’s efficacy. Consistently, DEXAFP significantly prolonged the

lifespan of treated HCC mice with 100% survival rate at 51 days after tumor

challenge whereas no mice survived in DEXAFP-, DEX- and PBS-treated

immunodeficient and PBS-treated immunocompetent mice (Fig. 5B).

To further confirm the direct involvement of T cells in the potency of DEXAFP, we

measured the number of effector T cells in splenic lymphocytes from DEXAFP-treated

orthotopic mice and the FACS results indicated a significant increase in the number of

IFN-γ-expressing CD3+ T cells compared to DEX or PBS treatments (Fig. 5C),

supporting the conclusion that T cells play an important role in DEXAFP-mediated

antitumor effect. To investigate whether effector T cells were actively recruited to

primary tumor sites and thus contributed to the antitumor effect, we examined levels

of T cell infiltration in tumor tissues. Immunohistochemical staining of CD3+ T cells

14

and CD4+FoxP3

+ regulatory T (Treg) cells in tumor tissues revealed more CD3

+ T

lymphocytes and fewer FoxP3+ Treg cells in tumors from DEXAFP-treated mice

compared to DEX or PBS groups (Fig. 5D). Consistently, significant increases in the

number of IFN-γ-expressing CD3+CD8

+ and CD3

+CD4

+ T cells were detected in

tumor tissues treated with DEXAFP compared to DEX or PBS groups demonstrated by

FACS (Fig. 5E), showing that T cell activation and infiltration contribute substantially

to the functionality of DEXAFP. To further elucidate the role of CD8+ and CD4+ T cells

in DEXAFP-mediated anti-tumor immunity, we depleted CD8+ or CD4+ or both T cells

in day-7-established ectopic HCC mice after repeated administration of antibodies

(Supplementary Fig. 7). Strikingly, the DEXAFP-mediated antitumor effect was

completely diminished when CD8+ or both CD8

+ and CD4

+ T cells were depleted,

whereas only a marginal effect observed in mice when CD4+ T cells were depleted

(Fig. 5F), indicating that CD8+ T cells are largely responsible for the antitumor

immunity of DEXAFP.

DEXAFP-mediated antitumor effect correlates with the improved immune

microenvironment in orthotopic HCC mice

To verify the correlation between DEXAFP-mediated antitumor effect and the

improved immune microenvironment in day-7-established orthotopic HCC mice, we

monitored the change of tumor growth and effector T cells in orthotopic HCC mice at

different time-points after treatment (Fig. 6A). Significant tumor retardation was

detected in HCC mice treated by DEXAFP 19 days after tumor challenge compared to

15

DEX or PBS groups (Fig. 6B), and the difference became more pronounced at day 26

after the third DEXAFP injection (Fig. 6B-C), demonstrating that repeated

administration of DEXAFP can consistently suppress tumor growth. Significant

increases in the number of CD8+ CTLs (Fig. 6D) and IFN-γ-expressing CD3

+ effector

T cells (Supplementary Fig. 8A) in serum were observed after 3 DEXAFP injections,

compared to first or second treatment and the corresponding DEX or PBS groups,

suggesting a cumulative immune effect. Furthermore, significantly elevated serum

levels of IFN-γ and IL-2 (Fig. 6E) and decreased levels of IL-10 and TGF-β (Fig. 6F)

suggest that DEXAFP treatment reshapes the HCC immune microenvironment. There

was also a significant increase in serum IL-12, a cytokine important for promoting

Th0 toward Th1 cell, from DEXAFP-treated HCC mice compared to DEX or PBS

groups at different time-points (Supplementary Fig. 8B), strengthening the importance

of T cell activation for the functionality of DEXAFP. These findings demonstrated a

correlation between DEXAFP-mediated antitumor effect and the improved HCC

immune microenvironment in orthotopic HCC mice.

DEXAFP elicits effective tumor suppression and alters tumor milieu in

DENA-induced autochthonous HCC mice

To further investigate the potency of DEXAFP in a more clinically relevant setting, we

established a DENA-induced autochthonous HCC mouse model, a model commonly

used for closely mimicking clinical manifestations in HCC patients [16]. Tumor

nodules and pathological heterogeneity were detected 7 months after one

16

intraperitoneal (i.p.) injection of 50 mg/kg DENA into 15-day old neonatal mice

(H-2b/k) (Supplementary Fig. 9A). As expected, AFP and glypican 3 (GPC3), two

well-characterized HCC antigens in the clinic [21, 22], were expressed in tumor

nodules derived from DENA-induced autochthonous HCC mice (Supplementary Fig.

9B), underlying the similarity of DENA-induced HCC mice to patients’ HCC.

DENA-induced autochthonous HCC mice had an immunosuppressive milieu with

decreased IFN-γ and increased IL-10 and TGF-β in serum compared with normal

controls (Supplementary Fig. 9C).

To test the antitumor efficacy of DEXAFP in DENA-induced autochthonous HCC mice,

we applied an identical dosing regimen to that used in orthotopic transplantation HCC

mice (Fig. 7A). Strikingly, a significant tumor suppression was achieved in

DEXAFP-treated mice compared to DEX or PBS groups, reflected by significantly

fewer tumor nodules (29.4+1.69, 21.2+2.08 or 7+0.89 for DEXAFP, DEX or PBS

groups, respectively), and decreased ratio of liver to body weight (Fig. 7B). Maximal

tumor nodule volumes of DEXAFP-treated mice were significantly lower than DEX

and PBS groups (Fig. 7C). Consistently, significantly more IFN-γ-expressing CD8+

CTLs and fewer CD25+Foxp3+ Treg cells were detected in serum from

DEXAFP-treated mice compared to DEX and PBS groups (Fig. 7D), suggesting that

DEXAFP accomplishes this function through T cell activation. As expected, IFN-γ and

IL-2 were significantly elevated (Fig. 7E), whereas IL-10 and TGF-β were

significantly decreased in serum from DEXAFP -treated mice (Fig. 7F), demonstrating

17

that DEXAFP is able to alter the HCC immune microenvironment in diverse HCC

models.

Importantly, the tumor milieu was reshaped with more CD8+ CTLs and a significantly

greater CD8+ to CD4

+ T cell ratio, a prognostic indicator for other tumors [23], in

tumor tissues from DEXAFP-treated mice compared to control groups (Fig. 8A).

CD25+CD4+ Treg cells were also significantly fewer in tumor tissues from

DEXAFP-treated mice compared to control groups (Fig. 8B). The results were

corroborated with significantly higher levels of IFN-γ and IL-2 (Fig. 8C) and

significantly lower levels of IL-10 and TGF-β (Fig. 8D). Consistently, more CD3+ T

cells and fewer Foxp3+ Treg cells infiltrated into tumor tissues from DEXAFP-treated

mice compared to control groups (Fig. 8E-F), further demonstrating that DEXAFP

treatment is capable of promoting active recruitment of activated T cells to tumor sites

in autochthonous HCC mice.

Discussion

DEX is promising as a cell-free cancer vaccine candidate due to a long shelf-life when

frozen, scalability and genetic modifiability [24], thus its performance in poorly

immunogenic HCC warrants investigation. Here we demonstrate that exosomes from

HCC antigen-expressing DCs (DEXAFP) could induce an effective antigen-specific

immune response in ectopic or orthotopic HCC mice. The observation that DEXAFP

treatments improved the tumor immune microenvironment of DENA-induced

18

autochthonous tumor, demonstrated by increases in immuno-stimulatory cytokines

and infiltrating CD8+ CTLs, reductions in immuno-inhibitory cytokines and Treg cells,

indicated the potential of DEXs as cell-free replacements for DC vaccines in HCC.

This is the first study, to our knowledge, to investigate if antigen-modified DEXs can

elicit antigen-specific antitumor immunity in a HCC model. Our study provides

evidence that DEX can be a cell-free vaccine for HCC immunotherapy and thus

provide impetus to test if DEX can effectively control HCC progression in patients.

Although significant tumor growth inhibition was achieved with DEXAFP in vivo, no

complete tumor eradication was observed in our study, underlining the tenacity of

HCC at overcoming the immune onslaught and the importance of targeting HCC

simultaneously with as many therapeutic options rather than relying on a “cure-all”.

The dosing regimens and delivery routes are also likely to impact the

immunotherapeutic effect of DEXAFP, thus further optimization before human clinical

trials is warranted. Similar antitumor effects were achieved when we compared

exosomes from AFP-expressing DC2.4 cells and bone marrow-derived primary DCs.

However the low yield of bone marrow-derived DCs limits the practicality of using

bone marrow DC-derived exosomes for in vivo studies; therefore we used a DC2.4

cells throughout. For future clinical use, human induced pluripotent stem cell-derived

DCs will likely be source cells for obtaining large quantities of antigen-modified

exosomes.

19

Based on our previous study, lower survival rate was obtained in orthotopic HCC

mice compared to ectopic ones [6], therefore we used much higher doses of DEXAFP

in orthotopic studies (150 µg) in our current study. Surprisingly, much higher survival

rate was achieved in orthotopic HCC mice than ecotopic ones after treatment,

suggesting that the dose of DEXAFP can be significantly reduced in future studies.

Strikingly, a consistent antitumor effect of DEXAFP was observed in mice with

autochthonous and transplanted HCC, suggesting that DEXAFP works in spite of tumor

heterogeneity. Previous studies had identified the molecular mechanisms involved in

the functionality of DEX in other tumor models as the direct or indirect activation of

T cells [9, 24]. Our study revealed that DEXAFP could activate functional T cells,

particularly CD8+

CTLs, to slow down the progression of HCC. There is growing

evidence showing that DEX activates natural killer (NK) cells [13, 25]. Significantly

more CD3-NK1.1+ NK cells (Supplementary Fig. 10A) and functional TNF-α+ IFN-γ+

NK cells (Supplementary Fig. 10B) were found in tumor tissues and serum from

DEXAFP-treated mice compared to control groups, suggesting that DEXAFP also

operate through this mechanism in HCC. Nevertheless, more detailed mechanistic

studies are warranted for elucidating the cellular pathway of DEXAFP.

The selection of antigens is critical for HCC immunotherapy as no single antigen is

ubiquitously expressed in HCC cells. Here we chose AFP for our proof-of-concept

study as AFP is one of the most common HCC antigens accounting for 50-80% HCC

patients and also was shown to be an effective HCC antigen for HCC immunotherapy

20

previously [14]. Although it was shown previously that HLA-restricted peptides failed

to trigger immune response and there is a risk of immune escape with a single antigen

[26, 27], AFP-modified DEX elicited a strong immune response in three different

HCC mice in our study. We speculate that this effect may be attributed to the use of

full-length AFP protein, which can be processed into multiple peptides. The strong

antitumor immunity elicited by DEXAFP in DENA-induced HCC models suggests that

DEXAFP may be successful in patient populations with heterogeneous HCC.

Conclusions

Taken together, our study demonstrates for the first time that antigen-modified DEX

can elicit a strong antigen-specific immune response and tumor suppression in diverse

HCC mouse models, demonstrating its potential as a cell-free vaccine for HCC

immunotherapy.

Acknowledgements

The authors thank Dr Yiqi Seow (Biomedical Sciences Institutes, A*STAR, Singapore)

for critical review of the manuscript.

References

[1] El-Serag HB. Epidemiology of viral hepatitis and hepatocellular carcinoma. Gastroenterology

2012;142:1264-1273 e1261.

[2] Lin S, Hoffmann K, Schemmer P. Treatment of hepatocellular carcinoma: a systematic review.

Liver cancer 2012;1:144-158.

[3] Lee WC, Wang HC, Hung CF, Huang PF, Lia CR, Chen MF. Vaccination of advanced

hepatocellular carcinoma patients with tumor lysate-pulsed dendritic cells: a clinical trial. Journal of

21

immunotherapy 2005;28:496-504.

[4] Chuma M, Terashita K, Sakamoto N. New molecularly targeted therapies against advanced

hepatocellular carcinoma: From molecular pathogenesis to clinical trials and future directions.

Hepatology research : the official journal of the Japan Society of Hepatology 2014.

[5] Palmer DH, Midgley RS, Mirza N, Torr EE, Ahmed F, Steele JC, et al. A phase II study of

adoptive immunotherapy using dendritic cells pulsed with tumor lysate in patients with hepatocellular

carcinoma. Hepatology 2009;49:124-132.

[6] Rao Q, Zuo B, Lu Z, Gao X, You A, Wu C, et al. Tumor-derived exosomes elicit tumor

suppression in murine hepatocellular carcinoma models and human in vitro. Hepatology 2016.

[7] Gould SJ, Raposo G. As we wait: coping with an imperfect nomenclature for extracellular vesicles.

Journal of extracellular vesicles 2013;2.

[8] Lotvall J, Hill AF, Hochberg F, Buzas EI, Di Vizio D, Gardiner C, et al. Minimal experimental

requirements for definition of extracellular vesicles and their functions: a position statement from the

International Society for Extracellular Vesicles. Journal of extracellular vesicles 2014;3:26913.

[9] Pitt JM, Charrier M, Viaud S, Andre F, Besse B, Chaput N, et al. Dendritic cell-derived exosomes

as immunotherapies in the fight against cancer. Journal of immunology 2014;193:1006-1011.

[10] Zitvogel L, Regnault A, Lozier A, Wolfers J, Flament C, Tenza D, et al. Eradication of established

murine tumors using a novel cell-free vaccine: dendritic cell-derived exosomes. Nature medicine

1998;4:594-600.

[11] Andre F, Chaput N, Schartz NE, Flament C, Aubert N, Bernard J, et al. Exosomes as potent

cell-free peptide-based vaccine. I. Dendritic cell-derived exosomes transfer functional MHC class

I/peptide complexes to dendritic cells. Journal of immunology 2004;172:2126-2136.

[12] Morse MA, Garst J, Osada T, Khan S, Hobeika A, Clay TM, et al. A phase I study of dexosome

immunotherapy in patients with advanced non-small cell lung cancer. Journal of translational medicine

2005;3:9.

[13] Escudier B, Dorval T, Chaput N, Andre F, Caby MP, Novault S, et al. Vaccination of metastatic

melanoma patients with autologous dendritic cell (DC) derived-exosomes: results of thefirst phase I

clinical trial. Journal of translational medicine 2005;3:10.

[14] Cany J, Barteau B, Tran L, Gauttier V, Archambeaud I, Couty JP, et al. AFP-specific

immunotherapy impairs growth of autochthonous hepatocellular carcinoma in mice. Journal of

hepatology 2011;54:115-121.

[15] Bradley LM, Dalton DK, Croft M. A direct role for IFN-gamma in regulation of Th1 cell

development. Journal of immunology 1996;157:1350-1358.

[16] Hong Y, Peng Y, Guo ZS, Guevara-Patino J, Pang J, Butterfield LH, et al. Epitope-optimized

alpha-fetoprotein genetic vaccines prevent carcinogen-induced murine autochthonous hepatocellular

carcinoma. Hepatology 2014;59:1448-1458.

[17] Qazi KR, Gehrmann U, Domange Jordo E, Karlsson MC, Gabrielsson S. Antigen-loaded

exosomes alone induce Th1-type memory through a B-cell-dependent mechanism. Blood

2009;113:2673-2683.

[18] Pardee AD, Butterfield LH. Immunotherapy of hepatocellular carcinoma: Unique challenges and

clinical opportunities. Oncoimmunology 2012;1:48-55.

[19] Chew V, Toh HC, Abastado JP. Immune microenvironment in tumor progression: characteristics

and challenges for therapy. Journal of oncology 2012;2012:608406.

[20] Cordier AC, Haumont SM. Development of thymus, parathyroids, and ultimo-branchial bodies in

22

NMRI and nude mice. The American journal of anatomy 1980;157:227-263.

[21] Murugavel KG, Mathews S, Jayanthi V, Shankar EM, Hari R, Surendran R, et al.

Alpha-fetoprotein as a tumor marker in hepatocellular carcinoma: investigations in south Indian

subjects with hepatotropic virus and aflatoxin etiologies. International journal of infectious diseases :

IJID : official publication of the International Society for Infectious Diseases 2008;12:e71-76.

[22] Shirakawa H, Kuronuma T, Nishimura Y, Hasebe T, Nakano M, Gotohda N, et al. Glypican-3 is a

useful diagnostic marker for a component of hepatocellular carcinoma in human liver cancer.

International journal of oncology 2009;34:649-656.

[23] Liu Y, Peng Y, Mi M, Guevara-Patino J, Munn DH, Fu N, et al. Lentivector immunization

stimulates potent CD8 T cell responses against melanoma self-antigen tyrosinase-related protein 1 and

generates antitumor immunity in mice. Journal of immunology 2009;182:5960-5969.

[24] Viaud S, Thery C, Ploix S, Tursz T, Lapierre V, Lantz O, et al. Dendritic cell-derived exosomes

for cancer immunotherapy: what's next? Cancer research 2010;70:1281-1285.

[25] Munich S, Sobo-Vujanovic A, Buchser WJ, Beer-Stolz D, Vujanovic NL. Dendritic cell exosomes

directly kill tumor cells and activate natural killer cells via TNF superfamily ligands. Oncoimmunology

2012;1:1074-1083.

[26] Makarova-Rusher OV, Medina-Echeverz J, Duffy AG, Greten TF. The yin and yang of evasion

and immune activation in HCC. Journal of hepatology 2015;62:1420-1429.

[27] Butterfield LH, Ribas A, Dissette VB, Lee Y, Yang JQ, De la Rocha P, et al. A phase I/II trial

testing immunization of hepatocellular carcinoma patients with dendritic cells pulsed with four

alpha-fetoprotein peptides. Clinical cancer research : an official journal of the American Association for

Cancer Research 2006;12:2817-2825.

23

Figure Legends

Figure 1. Characterization of exosomes derived from AFP-expressing DCs

(DEXAFP). (A) Western blot analysis for detecting the expression of exosomal

markers and cellular protein in DEXAFP. Total protein (25 µg) was loaded for all the

samples. (B) Transmission electron microscopy (TEM) image for DEXAFP (scale bar

= 100 nm). (C) Flow cytometry for analyzing levels of surface proteins in DEXAFP or

DEX (n=9; triplicates each time and repeated for 3 times).

Figure 2. Investigation of antigen-specific immune responses of DEXAFP in vitro.

(A) Measurement of IFN-γ and IL-2 in splenic T cells harvested from immunized

C57BL6 mice after re-stimulation with 40 µg DEXAFP or DEX. (B) Cytolytic

activities of effector T cells against Hepa1-6 cells after re-stimulation with DEX or

DEXAFP. Significant differences were detected between DEXAFP and other groups. (C)

Cytolysis comparison between different tumor cells. Significant differences were

achieved between Hepa1-6 and other groups. All experiments contained 3 mice per

group and repeated 3 times (n=9) and statistical significance determined with

two-tailed t test (**P<0.001; ***P<0.0001).

Figure 3. Examination of delivery routes on the antitumor effect of DEXAFP in

MHC-matched ectopic C57BL/6 tumor-bearing mice (H-2b). Subcutaneous

tumor-bearing C57BL6 mice were treated with PBS(�), DEX (�) or DEXAFP (▲) (40

µg/mouse/week for 3 weeks) intravenously (i.v.) or subcutaneously (sc). (A)

Measurement of tumor volume at different time-points after inoculation. Significant

24

differences were obtained between DEXAFP and other groups (n=15, 5 mice per group

and repeated for 3 times; two-tailed t test, *P<0.05;**P<0.01). (B) Survival rate of

treated subcutaneous tumor-bearing mice. (C) Analysis of tumor volume from

DEXAFP-treated tumor-bearing mice. (D) Survival rate of DEXAFP-treated

tumor-bearing mice.

Figure 4. Systemic evaluation of DEXAFP in orthotopic MHC-matched HCC mice.

Treatments (150µg/mouse/week for 3 weeks) were applied to orthotopic HCC mice

with PBS (�), DEX (�) or DEXAFP (▲), respectively. (A) Representative images for

treated mice. (B) Measurement of tumor volumes on day 26 after implantation.

Significant tumor suppression was yielded between DEXAFP and other groups (n=30,

10 mice per group and repeated for 3 times). (C) Survival rate of treated mice. (D)

Measurement of IFN-γ and IL-2 in serum from treated mice on day 26 after

implantation. Significant differences were found between DEXAFP and other groups

(n=15, 5 mice per group and repeated for 3 times). (E) Measurement of TGF-β and

IL-10 in serum from treated mice with ELISA. Significant differences were found

between DEXAFP and other groups (n=15). Significance was determined with

two-tailed t test (*P<0.05; **P<0.01).

Figure 5. Examination of the involvement of T cells and improvement in tumor

microenvironment in DEXAFP–treated orthotopic HCC mice. (A) Measurement of

tumor volume in treated ectopic immunodeficient or immunocompetent

25

tumor-bearing mice. Tumor-bearing nude mice were treated intravenously with

DEXAFP (▲) or DEX (�) (40 µg/mouse/week for 3 weeks). A significant difference

was detected between DEXAFP-treated immunodeficient and immunocompetent mice

(n=15). (B) Survival rate of treated tumor-bearing mice. A significant difference was

observed between DEXAFP–treated immunodeficient and immunocompetent mice. (C)

Analysis of IFN-γ-expressing CD3+ T lymphocytes in spleens from treated orthotopic

tumor-bearing mice (n=15). Significant increases in IFN-γ-expressing CD3+ T

lymphocytes were detected between DEXAFP and other groups. (D)

Immunohistochemistry of CD3+ and FoxP3

+ Treg cells in tumor sections (scale bar =

100 µm). (E) Analysis of IFN-γ-expressing CD3+CD4

+ and CD3

+CD8

+ T

lymphocytes in tumor tissues from treated orthotopic tumor-bearing mice. Significant

increases in the number of IFN-γ-expressing CD3+CD4+ and CD3+CD8+ T

lymphocytes were detected between DEXAFP and other groups. (F) Measurement of

tumor volumes in DEXAFP-treated ectopic tumor-bearing mice after CD8+ and CD4

+ T

cells depletion. Depl represents depletion. A significant difference was detected

between CD8+ depletion and no depletion groups. Significance was determined with

two-tailed t test (*P<0.05; **P<0.01; ***P<0.0001).

Figure 6. Dynamic examination on DEXAFP-mediated antitumor effect and

immune improvement in orthotopic HCC mice. (A) Schematic diagram for the

dosing regimen of DEXAFP in orthotopic HCC mice. (B) Representative images of

tumor nodules. (C) Measurement of tumor volume in treated orthotopic

26

tumor-bearing mice. Significant tumor suppression was yielded between DEXAFP and

other groups (n=15, 5 mice per group and repeated for 3 times). (D) Flow cytometry

analysis of CD8+ CTLs in serum from treated orthotopic tumor-bearing mice.

Significant increases in the number of CD8+ CTLs were detected between DEXAFP

and other groups. Compared to mice treated with DEXAFP once or twice, significant

increases in the number of CD8+ CTLs were detected for the third time. (E)

Measurement of IL-2 and IFN-γ in serum from treated orthotopic tumor-bearing mice.

Significant increases in levels of IL-2 and IFN-γ were detected between DEXAFP and

other groups. (F) Measurement of IL-10 and TGF-β in serum from treated orthotopic

tumor-bearing mice. A significant reduction in levels of IL-10 and TGF-β was

detected between DEXAFP and other groups. Significance was determined with

two-tailed t test (*P<0.05; **P<0.01).

Figure 7. Systemic investigation on the potency of DEXAFP in DENA-induced

autochthonous HCC mice. Treatments (150 µg/mouse/week for 3 weeks) were

applied to autochthonous HCC mice with PBS (�), DEX (�) or DEXAFP (▲),

respectively. (A) Schematic diagram for the dosing regimen of DEXAFP. (B-C)

Measurement of tumor nodules, the ratio of liver to body weight and tumor volume of

maximal tumor nodules in autochthonous HCC mice on week 40 after DENA

induction. Significant decreases were yielded between DEXAFP and other groups

(n=15, 5 mice per group and repeated for 3 times). (D) Flow cytometry analysis of

CD8+IFN-γ+ and CD25+Foxp3+ T lymphocytes in serum from treated autochthonous

27

HCC mice. A significant increase in the number of CD8+IFN-γ

+ and decrease in the

number of CD25+Foxp3+ T lymphocytes was detected between DEXAFP and other

groups. (E) Measurement of IFN-γ and IL-2 in serum from treated tumor-bearing

mice on week 40 after induction. Significant differences were detected between

DEXAFP and other groups. (F) Measurement of immunosuppressive cytokines

including TGF-β and IL-10 in serum from treated mice with ELISA. Significant

differences were detected between DEXAFP and other groups. Significance was

determined with two-tailed t test (*P<0.05; **P<0.01).

Figure 8. DEXAFP improved tumor microenvironment in DENA-induced

autochthonous HCC mice. (A) Analysis of CD8+CD3

+ and CD4

+CD3

+ T

lymphocytes in tumor tissues from treated tumor-bearing mice. A significant increase

in the number of CD8+ CTLs was detected in DEXAFP-treated mice compared to other

groups (n=15, 5 mice per group and repeated for 3 times). (B) Flow cytometry

analysis of CD25+CD4

+ Treg cells in tumor tissues from treated tumor-bearing mice.

A significant decrease in the number of CD25+CD4

+ Treg cells was detected in

DEXAFP-treated mice compared to other groups. (C) Measurement of IFN-γ and IL-2

in tumor tissues from treated tumor-bearing mice on week 40 after induction.

Significant differences were observed between DEXAFP and other groups. (D)

Measurement of TGF-β and IL-10 in tumor tissues from treated mice with ELISA.

Significant differences were detected between DEXAFP and other groups. (E)

Immunohistochemistry of CD3+ and FoxP3+ Treg cells in tumor sections from treated

28

autochthonous HCC mice (scale bar = 100 µm). (F) Quantitative analysis of FoxP3+

Treg cells in tumor tissues from treated autochthonous HCC mice. Significance was

determined with two-tailed t test (*P<0.05; **P<0.01).

29

30

31

32

33

34

35

36

37

38

Highlights

� DEXAFP elicits competent antigen-specific immune responses in hepatocellular

carcinoma mice.

� DEXAFP reshapes tumor immune microenviornment in hepatocellular carcinoma

mice.

� Significant tumor suppression correlates to improved immune microenvironment.

� CD8+ T cells are largely responsible for the functionality of DEXAFP.