A live attenuated Salmonella Typhimurium oral T cell vaccine against PD-L1

protects 100% of animals from a leukemia challenge

Sébastien Wieckowski1, Heiko Smetak2, Marco Springer3, Iris Kobl3, Amine A. Berkane4, Ming Wei4, Albrecht Meichle3,

Klaus M. Breiner1, Philipp Beckhove2, Marc Mansour1, Matthias Schroff1, Heinz Lubenau3

1VAXIMM AG, Basel, Switzerland; 2Regensburg Center for Interventional Immunology (RCI), Regensburg, Germany; 3VAXIMM GmbH, Mannheim, Germany; 4CellVax S.A.S., Romainville, France.

Background

VA M

VAXIMM's oral T-cell vaccine platform is based on the approved, live attenuatedSalmonella Typhi strain Ty21a vaccine, which has been administered in millions ofindividuals for prophylactic vaccination against typhoid fever. This strain has beenthoroughly studied, is safe and well tolerated. The bacteria are modified to deliveran eukaryotic expression plasmid, which encodes the genetic information of aspecific target antigen1.

Figure 2. (A) Schematic representation of VXM10 oral T-cell vaccine, and (B) domains of the murine PD-L1 protein encoded in VXM10 (orange) and VXM10a (brown) vaccines.

VXM01 lead vaccine encodes vascular endothelium growth factor receptor 2(VEGFR2) in order to evoke an immune response specifically directed against thetumor vasculature. It is currently in clinical development as a treatment forvarious solid cancer types. The murine analogue of VXM01 has shown consistentanti-angiogenic activity in different tumor models in several animal studies2. Anincrease in tumor immune cell infiltration was recently shown. A proposedmechanism of action of VXM01 is described in Figure 1.

Figure 1. Intra-lymphatic delivery of Salmonella Typhi strain Ty21a T-cell vaccines via the oral route leading to target-specific T-cell activation.

The current study summarizesthe immunogenicity andpreclinical anti-cancer efficacy forthe Salmonella TyphimuriumSL7207 murine vaccines VXM10and VXM10a (Figure 2A),transformed with eukaryoticexpression plasmids encoding thefull-length murine programmeddeath-ligand 1 (PD-L1) proteinand a truncated form of PD-L1,respectively (Figure 2B).Indeed, the deletion of the signalpeptide (SP) prevents the properlocalization of the native PD-L1protein to the cell surface. Theempty vector, i.e. withoutplasmid, was used as negativecontrol throughout the study.

Antibody response

Figure 7. (A) Experimental design, and (B) anti-PD-L1 antibody response in sera collected 79 days after the final vaccination. The green dashed line represents the cut-off value (for 95% confidence). Soluble recombinant murine PD-L1 was used for immunization with CFA/IFA in the positive control group (blue).

The systemic antibody response was evaluated by ELISA in the serum ofanimals vaccinated with either VXM10 or VXM10a, 79 days after the finalvaccination (Figure 7A). Anti-PD-L1 antibodies were detected in a few animalsvaccinated with VXM10 and VXM10a, and the response was more pronounced inthe VXM10a/high-dose group, with 50% of the animals (3 out of 6) showingsignal-to-background ratio above the cut-off value (Figure 7B).

Antitumor efficacy

We evaluated the prophylactic anti-cancer activity of VXM10 and VXM10a in theFBL-3 disseminated model of leukemia expressing PD-L13 (Figure 3A). Emptyvector, VXM10 and VXM10a were given by oral gavage at ca. 108 CFU and 1010

CFU, on days 1, 3, 5 and 7 as a prime vaccination, and on days 14 and 22 asboosts (Figure 3B). C57BL/6 mice (n=6 per group) then received 5×106 viableFBL-3 cells by intraperitoneal injection on day 20. All surviving animals were re-challenged with 5×106 viable FBL-3 cells by intraperitoneal injection on day 100.

1. Darji A. et al., Cell 1997; 91:765. 2. Niethammer AG. et al., Nature Medicine 2002; 8:1369.3. Yamazaki T. et al., Journal of Immunology 2002; 169:5538.

References

Contact and Information

Heinz Lubenau, Ph.D.

VAXIMM GmbH

Chief Operating Officer

Office: +49 621 8359 687 10

Fax: +49 621 8359 687 99

heinz.lubenau@vaximm.com

www.vaximm.com

MAFINEX-Technologiezentrum

Julius-Hatry-Straße 168163 Mannheim

Germany

Poster No. B057 presented during the Cancer Vaccines and Targets session at the Third CRI-CIMT-EATI-AACR International Cancer lmmunotherapy Conference on September 8th 2017 in Mainz/Frankfurt, Germany.

Figure 3. (A) Expression of PD-L1, but not PD-L2, by FBL-3 cell line3, as measured by flow cytometry (left inset) and RT-PCR (right), and (B) experimental design and treatment schedulein the prophylactic and re-challenge experiment.

Figure 4. (A) Evolution of the mean bodyweight, and (B) overall survival in the indicatedtreatment groups and doses. The blue arrows represent the time points of leukemiachallenge. Treatment-naive animals (yellow curves) were used as a control for the FBL-3 rechallenge and received the leukemia cells only day 100.

We finally evaluated the therapeutic efficacy ofVXM10 and VXM10a in the FBL-3 model. C57BL/6mice (n=8 per group) received 5×106 viable FBL-3cells by intraperitoneal injection on day 0. Emptyvector, VXM10 and VXM10a were thenadministered by oral gavage at a dose of 109 CFUon days 1, 3, 5 and 7 as a prime vaccination, andon days 14 and 21 as boosts (Figure 5).

Therapeutic vaccination with VXM10 and VXM10a was well tolerated (Figure6A), and induced full leukemia control, with 100% (8 out of 8) of survivinganimals 94 days after leukemia challenge (P<0.0001). In contrast, treatmentwith the empty vector control did not show any anti-cancer effect (Figure 6B).

Conclusions

▪ Prophylactic and therapeutic vaccinations with VXM10 and

VXM10a induced a strong and sustained anti-cancer activity in

the FBL-3 model of leukemia.

▪ This study provides evidence that VAXIMM’s oral T-cell

vaccination platform can be used to stimulate anti-tumor

immunity against antigens of the immune checkpoint regulatory

protein PD-L1.

▪ These data paved the way for advancing the clinical

development of VXM10, in particular in leukemia.

Figure 5. Experimental design of the therapeutic study.

Figure 6. (A) Evolution of the bodyweight in each individual animal, and (B) overall survivalin all treatment groups, in the therapeutic setting. The blue arrow represents the time point of FBL-3 challenge.

D1,3,5,7 D14 D21

FBL-3 challenge

D0

D94

Prophylactic vaccination with VXM10 and VXM10a was highly tolerated, as nodeterioration in general status nor significant body weight loss were observedduring the treatment (Figure 4A). It also generated a rapid and sustained anti-leukemia effect with 100% (6 out of 6) of surviving animals 80 days after leukemiachallenge (P=0.0005) in the highest dose groups. In contrast, vaccination with theempty vector control did not show any anti-cancer activity, as the median survivalreached 41 days, and 0% (0 out of 6) of cancer regression was observed (Figure4B). Importantly, 100% of surviving mice in the high dose groups resisted re-challenge with FBL-3 cells for at least 100 days (P=0.0002), demonstrating thatvaccination with VXM10 and VXM10a generated a potent memory T cell responseagainst the leukemia (Figure 4B).

A B

0 40 80 120 160 200

0

20

40

Days after initiation of treatment

Mean

% b

od

yw

eig

ht

ch

an

ge o

ver

baselin

e

empty vector

VXM10 108 CFU

VXM10 1010 CFU

VXM10a 108 CFU

VXM10a 1010 CFU

untreated (traitment-naive)

0 40 80 120 160 200

0

50

100

Days after tumor challenge

Perc

en

t su

rviv

al

empty vector

VXM10 108 CFU

VXM10 1010 CFU

VXM10a 108 CFU

VXM10a 1010 CFU

untreated(treatment-naive)

***

*

A B

0 20 40 60 80 100

0

10

20

30

Days after tumor challenge

Perc

en

t b

od

yw

eig

ht

ch

an

ge o

ver

baselin

e

empty vector

VXM10

VXM10a

0 20 40 60 80 100

0

50

100

Days after tumor challenge

Perc

en

t su

rviv

al

empty vector

VXM10

VXM10a

****

Oral administration of a suspension

containing attenuated Salmonella bacteria

carrying plasmids encoding for target antigens.

Bacteria pass through M cells into

Peyer’s patches and are taken up

by macrophages.

Bacteria die inside

macrophages and release

plasmids. Plasmids enter

the nucleus and the encoded

antigen is expressed.

Bacterial infection induces

macrophages to undergo

apoptosis (cell death).

Apoptotic vesicles

contain antigen.

Apoptotic vesicles are phagocytized

by dendritic cells, processed, and

antigen epitopes are presented on

the surface via MHC Class I.

Antigen-specific CD8+ T-cells

are activated.

CD8+ T-cells circulate through the

body and bind to target cells

expressing the specific

antigen, initiating

cell death.

Salmonella bacterium

Plasmid

Small intestine

Small intestinallumen

Macrophage

M cell

T-cell

Salmonella bacteriacarrying plasmids

Peyer’spatch

Phagocytosis

of bacteria

Bacteria

carrying

plasmids

Antigen

Plasmids in nucleusMacrophage

Apoptotic

macrophageAntigen

Apoptotic

vesicle

Dendritic cell

phagocytizing vesicle

Antigen

epitope

MHC

class I

Antigen-specific

CD8+ T-cells

Activated dendritic cell

Blood

vessel

CD8+ T-cell

Target (tumor) cell

MHC class Imounted antigen

7

6

54321

D1,3,5,7 D14 D21

FBL-3

challenge

D20

D205

FBL-3

rechallenge

D100

A B

Flow cytometry RT-PCR

PD

-L1

PD

-L2

PD-L2

PD-L1

β-actin

A B

D1, 3, 5, 7 D14 D21

D100

serum

OD (signal:background ratio)

FBL-3

challenge

D20

cut-off

0.5 1.0 1.5 2.0 2.5 3.0

immun. sPD-L1

negative control

VXM10a 1010

CFU

VXM10a 108

CFU

VXM10 1010

CFU

VXM10 108

CFU

A

B murine PD-L1

VXM10

VXM10a

SP ECD TM ICD

1–18 240–260

29026123919

Salmonella TyphimuriumSL7207 carrier

Eukaryotic expressionplasmid encodingmurine PD-L1 protein