RES EARCH

CANCER IMMUNOTHERAPY of CLDN6 and the corresponding amino acid

sequences of these claudins is 81, 85, and 98%,
An RNA vaccine drives expansion and efficacy of respectively, bearing the risk of cross-reactivity
and off-target toxicity of the CAR. We found
claudin-CAR-T cells against solid tumors that the CLDN6-transfected target cells were
killed but not those transfected with the rel-
Katharina Reinhard1*, Benjamin Rengstl1*, Petra Oehm1*, Kristina Michel1, Arne Billmeier1, ated claudins, demonstrating precise targeting
Nina Hayduk1, Oliver Klein1, Kathrin Kuna1, Yasmina Ouchan1, Stefan Wöll1, Elmar Christ1, by CLDN6-CAR-T cells (Fig. 1E).
David Weber2, Martin Suchan2, Thomas Bukur2, Matthias Birtel1, Veronika Jahndel1, Karolina Mroz1, To measure cognate immune activation, we
Kathleen Hobohm1, Lena Kranz1, Mustafa Diken2, Klaus Kühlcke1, Özlem Türeci1†, Ugur Sahin1,2,3†‡ cocultured CLDN6-CAR-T cells with human
tumor cell lines. We found interferon-g (IFN-g)
Chimeric antigen receptor (CAR)–T cells have shown efficacy in patients with B cell malignancies. secretion and up-regulation of T cell activation
Yet, their application for solid tumors has challenges that include limited cancer-specific targets markers upon coculture with CLDN6pos tar-
and nonpersistence of adoptively transferred CAR-T cells. Here, we introduce the developmentally gets but not CLDN6neg cells (Fig. 1F). CLDN6-
regulated tight junction protein claudin 6 (CLDN6) as a CAR target in solid tumors and a strategy to CAR-T cells were able to efficiently clear
overcome inefficient CAR-T cell stimulation in vivo. We demonstrate that a nanoparticulate RNA CLDN6pos PA-1 ovarian carcinoma spheroids
vaccine, designed for body-wide delivery of the CAR antigen into lymphoid compartments, stimulates and to kill repetitively upon rechallenge (Fig.
adoptively transferred CAR-T cells. Presentation of the natively folded target on resident 1G). Deletion of CLDN6 by CRISPR-Cas9–
antigen-presenting cells promotes cognate and selective expansion of CAR-T cells. Improved mediated genetic knockout (Fig. 1G, top) com-
engraftment of CAR-T cells and regression of large tumors in difficult-to-treat mouse models was pletely abrogated CAR-T cell recognition of

Downloaded from http://science.sciencemag.org/ on March 31, 2021

achieved at subtherapeutic CAR-T cell doses. PA-1, further confirming high potency and
target-specificity of CLDN6-CAR-T cells.

A
Next, we studied in vivo antitumor activity
doptive cell therapy (ACT) with geneti- titative real-time–polymerase chain reaction of human CLDN6-CAR-T cells in mice xeno-
cally engineered T lymphocytes express- (qRT-PCR), significant CLDN6 transcript ex- grafted subcutaneously with a human tumor
ing chimeric antigen receptors (CARs) pression was ruled out (Fig. 1A and fig. S2B). cell line. Of note, the mouse is not a suitable
has been clinically successful in patients In addition, CLDN6 protein was not detectable species for studying toxicity of this CAR be-
with B cell malignancies (1, 2). However, in any of the adult human normal tissue types cause the binding affinity of CLDN6-CAR to
in patients with solid tumors, the efficacy of (>40 tested) assessed by IHC staining (Fig. 1B). the mouse CLDN6 ortholog is 15-fold lower
CAR-T cell therapy is challenging and much In line with previous studies (9, 10), high than to human CLDN6 and, whereas human
less effective (3). One key hurdle is the limited CLDN6 transcript levels were frequent in CLDN6 is strictly confined to the embryonic
number of cell-surface targets with high cancer- various human solid cancers such as testicular, stage, murine CLDN6 is expressed in some
specific expression to allow for efficient tumor ovarian, uterine, and lung adenocarcinoma postembryonic somatic tissues. Immunode-
eradication and low risk of off-tumor on-target (Fig. 1A and fig. S2, A to C). IHC staining ficient NOD-scid IL2Rgnull (NSG) mice with
toxicity (4–6). We and others have recently re- showed membrane expression of CLDN6 pro- large ovarian OV90 tumors (mean volume
ported cancer-associated expression of claudin 6 teins in these human cancers that was high 168 mm3) underwent ACT with a single dose
(CLDN6), a tetraspanin membrane protein and homogeneous in many of the tested spec- of human CLDN6-CAR-T cells or control cells.
that is involved in tight junction formation imens (fig. S2C). These findings indicate ex- Notably, all CLDN6-CAR-T cell–treated mice
(7). To evaluate the suitability of CLDN6 as a quisitely tight and complete silencing of CLDN6 experienced complete tumor regression within
target for CAR-T cell therapy, we profiled its in normal human tissues and suggest that 2 weeks, compared with control group mice
expression in a comprehensive set of human CLDN6 is a strictly oncofetal cell-surface anti- with tumors that progressed rapidly (Fig. 1H).
and mouse tissues. In mice, CLDN6 has been gen with an ideal expression profile for CAR-T Circulating CLDN6-CAR-T cells were detectable
reported to be developmentally regulated (8). cell targeting (11). in cured mice for the full observation period of
By immunohistochemical (IHC) staining, we We designed a second-generation CLDN6- up to 25 days after ACT (fig. S3).
found CLDN6 to be broadly expressed in fetal CAR with a 4-1BB costimulatory domain. For Engraftment and persistence of transferred
organs but prenatally down-regulated, result- the receptor domain, we engineered a single- CAR-T cells are known to be critical for their
ing in lack of expression in most organs of chain variable fragment (scFv) with exquisite clinical effect (12–14). In hematological malig-
adult mice (fig. S1A). In humans, CLDN6 tran- specificity and high binding affinity to CLDN6 nancies, CAR-T cells are directed against lin-
script levels were high in fetal tissues derived in the nanomolar range (Fig. 1C). First, we char- eage antigens of B cells and encounter their
from stomach, pancreas, lung, and kidney but acterized CLDN6-CAR–engineered human T cells targets on the host’s normal and malignant
undetectable in the corresponding adult tissue in vitro. CLDN6neg human COLO-699N lung B cells. These act as antigen-presenting cells
samples (fig. S1B). In more than 160 noncan- carcinoma cells were transfected with increasing (APCs) that provide strong proliferation signals
cerous healthy human samples from more amounts of CLDN6 RNA and assessed for killing and promote persistence of CAR-T cells (13, 14).
than 50 adult tissue types analyzed by quan- by CAR-T cells (Fig. 1D). We observed highly sen- However, in the solid-tumor setting, the fre-
sitive recognition and lysis of CLDN6-transfected quency of CAR-T cells typically declines rapid-
1
Biopharmaceutical New Technologies (BioNTech) Corporation,
target cells by the CLDN6-CAR, even at the ly (15–17) owing to the impaired accessibility of
BioNTech Cell & Gene Therapies GmbH, BioNTech Innovative lowest target expression level. CAR-T cells to tumor cells within solid lesions
Manufacturing Services GmbH, An der Goldgrube 12, 55131 In a similar experimental setting, we eval- and the absence of proliferation signals when
Mainz, Germany. 2TRON–Translational Oncology at the
uated the CLDN6-CAR for cross-recognition of CAR-T cells encounter the target in an immu-
University Medical Center of Johannes Gutenberg University
gGmbH, Freiligrathstr. 12, 55131 Mainz, Germany. 3Helmholtz CLDN3, CLDN4, and CLDN9, the most close- nosuppressive tumor microenvironment. We
Institute for Translational Oncology Mainz, HI-TRON Mainz, ly related claudin family members that, in hypothesized that expression of the CAR tar-
Obere Zahlbacher Str. 63, 55131 Mainz, Germany. contrast to CLDN6, are expressed in toxicity- get in its native conformation on the surface of
*These authors contributed equally to this work. †These authors
contributed equally to this work. relevant normal tissues. The homology be- professional APCs in lymphoid tissues would
‡Corresponding author. Email: [email protected] tween the CAR-targeted first extracellular loop render it accessible for cognate CAR-T cell

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 1 of 8

RES EARCH | R E P O R T

A Metastasis
B a b c
1,000
Primary tumor
Relative expression

Healthy tissue
100
d e f

10
g h i

adrenal gland
blood vessel
brain
cartilage
colon
dorsal root ganglion
epididymis
esophagus
eye
fallopian tube

lung
lymph node
nerve
omentum

pancreas
PBMC
pituitary gland
placenta
prostate
rectum
salivary gland
skeletal muscle
skin
small intestine
spinal cord
spleen
stomach
synovial membrane
testis
thrombocytes
thymus
thyroid
tongue
tonsil
trachea
umbilical cord
uterus
breast

heart
kidney
bone marrow

oropharynx
ovary
bladder

gall bladder

liver

ureter
Ovarian cancer
j k l
C

CLDN6- m n o
scFv D E CLDN3 CLDN4
80

CD8 hinge 60

Lysis [%]
0
CLDN6 RNA [µg]

p q r
10 40

Downloaded from http://science.sciencemag.org/ on March 31, 2021

4-1BB 1 CLDN6 CLDN9
0.1 20
0.01
0
CD3ζ 0.001 s t u

CLDN3
CLDN4
CLDN6
CLDN9
0 50 100
CLDN6 Lysis [%] CLDN

G PA-1 PA-1-CLDN6-/-
v w x
CLDN6-CAR-T + PA-1
CLDN6-CAR-T + PA-1-CLDN6-/-
Spheroid eGFP signal

1.2 x107
non-transd. T + PA-1
0.9 x107 CLDN6 non-transd. T + PA-1-CLDN6-/- CA1 CA2 CA3

0.6 x107

0.3 x107
x200
0
0 24 48 72 96 120 144 168 192 216 240 264 288 F CLDN6/ CLDN9+ cells [%] 100
Time [h]
CLDN6
αCLDN6 CLDN9
H isotype ctrl CD4+ CD8+ 1,500 50
Tumor volume [mm3]

CLDN6-CAR-T
ACT

OV90 tumor cells 16.3 18.4 1,000

0
25
CLDN6 500
20.0 23.5 CLDN6-CAR-T
20 non-transd. T
IFNγ [ng/mL]

CLDN6-CAR-T
GFP-transd. T GFP-transd. T
GFP

15
0
-10 -5 0 5 10 15 20 25
10
CAR Days after ACT
5

0
Fig. 1. The oncofetal antigen CLDN6 is a target for CAR-T cell therapy. (A and B) qRT-PCR expression of CLDN6
60
transcripts (A) and IHC analysis of protein (B) in human tissues [(a) adrenal gland, (b) fallopian tube, (c) kidney,
Activated CD4/ CD8 [%]

(d) liver, (e) thyroid, (f) prostate, (g) esophagus, (h) stomach, (i) colon, (j) cerebrum, (k) cerebellum, (l) spinal cord, CD4+OX40+
40 CD8+4-1BB+
(m) thymus, (n) spleen, (o) bone marrow, (p) pancreas, (q) skin, (r) bladder, (s) placenta, (t) heart muscle, (u)
striated muscle, (v) testis, (w) ovary, (x) lung, (CA1) testicular cancer, (CA2) ovarian cancer, and (CA3) lung cancer].
20
(C) Design of CLDN6 CAR. (D) COLO-699-N cells (no endogenous claudin expression) electroporated with increasing
amounts of CLDN6 RNA, as analyzed by flow cytometry (left), and dependency of lysis by human CLDN6-CAR-T
0
cells [right; effector:target cell (E:T) = 20:1, mean + SD of technical triplicates] on the level of CLDN6 surface
NEC8

NCI-N87_CLDN18.2
MCF7
23132-87
LCLC-103H

COLO-699-N
HEK-293

CPC-N
JAR

MDA-MB-231
PA-1-CLDN6-/-
NIH-OVCAR-3
PA-1

SK-OV-3

expression. (E) Surface expression of highly homologous claudins on COLO-699-N cells electroporated with
CLDN RNAs assessed by flow cytometry (left; control: isotype antibody) and analysis of cross-recognition and lysis by
cocultured CLDN6-CAR-T cells (right; E:T = 7:1, mean + SD of technical triplicates. (F) Human tumor cell lines
analyzed by flow cytometry for CLDN6 and CLDN9 surface expression (top) were cocultured with CLDN6-CAR or
nontransduced T cells (E:T = 10:1). IFN-g secretion (middle; mean + SD of technical duplicates) and expression of
activation markers OX40 on CD4+ and 4-1BB on CD8+ T cells after coculture (bottom), as assessed by flow cytometry, are shown. (G) Serial killing of CLDN6pos and
CLDN6−/− PA-1 tumor spheroids cocultured with either CLDN6-CAR or nontransduced (non-transd.) T cells (E:T = 10:1), as measured by enhanced GFP (eGFP) real-time
imaging (mean of technical triplicates). (H) NSG mice bearing subcutaneous CLDN6pos OV90 xenografts were treated with human T cells transduced with CLDN6-CAR or
eGFP. Tumor and T cell characteristics (left and middle) and tumor growth kinetics in individual mice (right) were analyzed. ctrl., control.

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 2 of 8

RES EARCH | R E P O R T

stimulation in an optimal immune-activating vaccine (referred to hereafter as CARVac), we DCs and macrophages but not on lymphocytes
environment. conducted a series of experiments. First, we (Fig. 2C and fig. S4A), confirming in vivo de-
Recently, we introduced intravenously ad- tested if CLDN6 can be natively displayed on livery of the CAR antigen exclusively to APCs.
ministered liposomal antigen-encoding RNA dendritic cells (DCs) to stimulate CLDN6-CAR-T APCs were activated and underwent matura-
(RNA-LPX) to stimulate tumor-associated cells in vitro. We measured concentration- tion (fig. S4B), and strong activation of nat-
T cells in the natural repertoire of cancer dependent surface expression of CLDN6 on ural killer (NK), B, and T cells was detected in
patients (18). This nanoparticulate vaccine DCs treated with different amounts of CLDN6- the spleen and lymph nodes of RNA-LPX–
delivers antigen to APCs in the spleen, lymph encoding RNA-LPX (herein CLDN6-LPX) injected mice (fig. S4C).
nodes, and bone marrow and concomitantly (Fig. 2A, top). The resulting expression of Next, naïve C57BL/6 mice were engrafted
initiates a Toll-like receptor–dependent type I CLDN6 on DCs induced stimulation, cytokine with CLDN6-CAR-T cells labeled with a cell pro-
IFN–driven immune-stimulatory program, pro- secretion, and proliferation of co-cultured liferation dye and vaccinated with CLDN6- or
moting priming and strong expansion of CLDN6-CAR-T cells in a dose-dependent man- control-LPX. Spleen and lymph nodes from all
antigen-specific T cells. ner (Fig. 2B, top). When BALB/c mice were major body regions resected from CLDN6-
To test whether this approach could be injected intravenously with CLDN6-LPX, CLDN6 LPX–vaccinated, but not control-treated mice,
adapted to act as a CAR-T cell–amplifying RNA surface expression was detected on splenic displayed significantly increased proportions

A Anionic mRNA mRNA-Lipoplex DC Transfected DC B

- - = RNA-LPX 104 IFNγ TNFα IL-2 100 CD4 CD8

Proliferating T cells [%]

-

Cytokine [pg/mL]
- - 103

Downloaded from http://science.sciencemag.org/ on March 31, 2021

+
Cationic Liposomes 102 50
+ 1
+ 10
+

nd

nd
nd
+ 100 0
+

0
100
10
1
0.1
0
100
10
1
0.1
0
100
10
1
0.1

0
100
10
1
0.1
0
100
10
1
0.1
CLDN6-LPX [µg/mL]
0 100 10 1 0.1 CLDN6-LPX [µg/mL] CLDN6-LPX [µg/mL]

104 IFNγ TNFα IL-2 CD4 CD8

100

Proliferating T cells [%]

Cytokine [pg/mL]

103

102 50
CLDN6
CLDN18.2-LPX [µg/mL] 101

0 100 10 1 0.1
nd

nd
100 0

0
100
10
1
0.1
0
100
10
1
0.1
0
100
10
1
0.1
0
100
10
1
0.1
0
100
10
1
0.1
CLDN18.2-LPX [µg/mL] CLDN18.2-LPX [µg/mL]

CLDN18.2

C D CPD450+ 20 µg
6
Fig. 2. Activation of CAR-T cells by RNA-LPX– RNA-LPX i.v. injection in vivo transfected 1.5 x10 CAR-T RNA-LPX
splenocytes 1°
mediated display of the CAR target on C57BL/6
dendritic cells is antigen specific and dose- -4 0 48 hours
dependent. (A) RNA-LPX were generated
organ collection
by mixing anionic mRNA with cationic liposomes. 100
Proliferating CAR-T cells [%]

Surface expression of CLDN6 (top) and 16 90

saline 80 CLDN6-LPX
CLDN18.2 (bottom) on DCs pulsed with RNA-LPX ****
CLDN6+ cells [%]

CLDN6-LPX 70 ctrl-LPX
12
encoding the respective CLDN assessed by 60
ctrl-LPX
flow cytometry is shown. (B) Cytokine secretion 50
8
of CAR-T cells analyzed by a multiplex assay after 25 ***
20
24 hours of coculture of claudin-expressing 4 15 **
DCs with carboxyfluorescein succinimidyl ester 10 *
(CFSE)–labeled CLDN6-CAR (top left) or 0 5
*
c dim
0+
4+

8+

0
s

CLDN18.2-CAR-T cells (bottom left). Proliferation

C

C
/8
D

11
8+

8-

LN een

LN )

ce ft)

N
C

F4

g. ght

+ +
of CD4 and CD8 CAR-transduced T cells was
.L

.L
D

(le
D

rv
i

ax
C

sp

( r

analyzed by flow cytometry after 5 days (right,

g.

in
in

top and bottom). Means + SD of technical

triplicates are indicated; nd indicates not detected. (C) Surface expression of CLDN6 on splenic immune cell populations of BALB/c mice analyzed by flow cytometry
24 hours after a single intravenous (i.v.) injection of 25 mg RNA-LPX encoding either CLDN6 or an irrelevant control (mean + SEM of biological duplicates). See fig. S4A for
histograms. (D) CAR-T cell proliferation in secondary lymphoid tissues resected 48 hours after intravenous administration of RNA-LPX (CLDN18.2 as control). Data indicate
mean ± SEM of biological replicates (n = 5 mice per group). LN, lymph node; ing. LN, inguinal LN; cerv. LN, cervical LN; ax. LN, axillary LN. P values were determined by unpaired
Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 3 of 8

RES EARCH | R E P O R T

A TBI
+ B
6 CLDN6-LPX Transferred Endogenous
10 CAR-T
1° (Thy1.1+) (Thy1.2+)
C57BL/6 50
BrdCrHsd-TyrC 0 8 Days post ACT
40

CD8+ KLRG1+
of parent [%]
d8
baseline 30

RNA-LPX 0 20 10 5 2.5 1.25 0.625 µg Expansion index d11 20

102
10

compared to baseline
d9

xfold expansion
0
100
d11 101

CD8+ CD62L+
80

of parent [%]
60

d14 40
100 20

0
20
10
5
2.5
1.25
0.625
0
d21

0
20
10
5
2.5
1.25
0.625
0
20
10
5
2.5
1.25
0.625
CLDN6 RNA [µg]

min max CLDN6 RNA [µg]

C TBI 20 µg
+ CLDN6-LPX

Downloaded from http://science.sciencemag.org/ on March 31, 2021

3 7 / saline 103 CAR-T + LPX
10 -10 CAR-T

Thy1.2
1° 2°
C57BL/6
BrdCrHsd-TyrC 0 1 8 Days post ACT 1° 2° treatment 0.01 1.00
107 CAR-T + saline 107
transferred CAR-T cells
107 106 105 103 Thy1.1 (transferred)

Thy1.1+ of lymphocytes [%]

Total flux [p/s]
d1
baseline 2
**
106 CAR-T + saline
1° treatment saline CLDN6-LPX 106
1.5
d4 1
105 CAR-T + saline
2° treatment saline CLDN6-LPX 103 CAR-T + LPX 0.5
d11 105 0
0 5 10 15 post 1st post 2nd
min max Days post ACT LPX (d4) LPX (d11)

D TBI
+
1.5 x106/ 20 µg
Time point a Time point b
7.5 x106 CAR-T RNA-LPX
1° 2° B16-hCLDN6 B16-WT B16-hCLDN6 B16-WT
Specific lysis of target cells [%]
Specific lysis of target cells [%]

C57BL/6 **
0 7 14 Days post ACT 80 80
1° 2° RNA-LPX
60 * 60 CLDN6-LPX
** ** ctrl-LPX
Total flux [p/s]

107 CLDN6-LPX ***

ctrl-LPX 40 40
* *
20 20
106
a b
0 0
5 10 15 20 25 1 3 7 10 30 1 3 7 10 30 1 3 7 10 1 3 7 10
Days post ACT Effector : target ratio (x :1) Effector : target ratio (x :1)

E TBI
+ 20 µg
106 CAR-T RNA-LPX
Prior After Time point c
1° 2° 3° 4° 5°
C57BL/6 vacc. (a) 3° RNA-LPX (b) 8 *
***
total CD8+T cells [%]

BrdCrHsd-TyrC 0 8 15 22 49 86 Days post ACT

0.73 16.89 CD127+
1° 2° 3° 4° 5° RNA-LPX CLDN6- 6
CD62Lneg
CAR-T of

LPX
108 KLRG1neg
4
CD127+
CAR
Total flux [p/s]

1.14 0.49 2 CD62L+

ctrl-
107 LPX KLRG1neg
CLDN6-LPX 0
ctrl-LPX
X

GFP
LD l-LP

LP

106
6-
r
N
ct
C

a b c
0 10 20 30 40 50 60 70 80 90
Days post ACT
Fig. 3. CARVac promotes efficient in vivo expansion, superior functionality, and memory formation of CAR-T cells. (continued on next page)

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 4 of 8

RES EARCH | R E P O R T

(A and B) Impact of dose level of intravenously administered target-antigen shown. (D) Ex vivo cytotoxic activity of low-dose CAR-T cells from CLDN6-LPX–
encoding RNA-LPX on expansion of CAR-T cells in vivo. Luc-expressing Thy1.1+ vaccinated mice (1.5 × 106 CAR-T + CLDN6-LPX) compared with high-dose
CLDN6-CAR-T cells (106 per animal) were transferred into lymphodepleted CAR-T cells sorted from control-vaccinated mice (7.5 × 106 CAR-T + CLDN18.2-
Thy1.2+ C57BL/6-albino mice (n = 5 mice per group). Eight days later, mice LPX) 3 days (time point a) and 7 days (time point b) after second vaccination
were injected intravenously with 40 mg RNA-LPX in total containing the (n = 5 mice per treatment group per time point). Sorted, pooled CAR-T cells
indicated titrated doses of CLDN6-LPX. Kinetics of CAR-T cell expansion per treatment group were cocultured for 20 hours in the presence of human
[(A), left] by bioluminescence imaging (BLI) and the expansion index of CAR-T CLDN6-transduced B16 mouse melanoma cells or wild-type (WT) control
cells [(A), right] and the frequencies of KLRG1- and CD62L-expressing endogenous at indicated E:T ratios (mean ± SD of technical triplicates). (E) Luc-eGFP–
(Thy1.2+) and transferred (Thy1.1+) CD8+ T cells in peripheral blood (B) 11 days expressing Thy1.1+ CLDN6-CAR-T cells transferred into lymphodepleted Thy1.2+
after ACT by flow cytometry (mean + SEM) are shown. d, day. (C) Impact of C57BL/6-albino mice (n = 2 or 3 mice per group) followed by repetitive
repetitive intravenous dosing of target-antigen encoding RNA-LPX on expansion vaccination with RNA-LPX (OvaI as control). CAR-T cell kinetics by BLI (left;
of CAR-T cells in vivo. BLI kinetics of different dose levels of Thy1.1+ Luc-expressing mean ± SEM of treatment groups) are shown. Frequency of eGFP+ CAR-T cell
CLDN6-CAR-T cells transferred into lymphodepleted Thy1.2+ C57BL/6-albino subsets in the peripheral blood in pretreatment samples (time point a, day 7
mice are shown. Mice in the lowest CAR-T cell dose group (103 cells) were after ACT) and after third RNA-LPX treatment (time point b, day 26 after ACT)
vaccinated twice with 20 mg CLDN6-LPX (n = 6 mice), whereas all other groups (middle) are shown. Frequency of memory CAR-T cells in the CD8+ T cell
received saline (n = 4 mice per group). Representative imaging (left) and population 31 days after fourth treatment (right; time point c, day 80 after
mean ± SEM of treatment groups (middle) are shown. Thy1.1+ subsets in the ACT) is shown. P values were determined by paired (C) and unpaired [(D) and
peripheral blood of individual mice determined by flow cytometry (right) are (E)] two-tailed Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001.

Downloaded from http://science.sciencemag.org/ on March 31, 2021

of proliferating CLDN6-CAR-T cells, suggest- a single intravenous dose of CLDN6-LPX in- cells isolated from unvaccinated mice, they
ing body-wide functional expression of the duced a profound expansion of circulating produced higher levels of IFN-g (fig. S8B) and
CAR antigen within lymphoid compartments CLDN6-CAR-T cells (Fig. 3A and fig. S6). The exerted significantly higher and strictly antigen-
(Fig. 2D). expansion correlated with the CLDN6-LPX dose dependent cytolytic activity upon ex vivo co-
To assess the broader applicability of this level and was substantial at even the lowest culture with CLDN6pos tumor cells (Fig. 3D).
approach, we selected CLDN18.2, which is a dose of 0.625 mg CLDN6 RNA. Quantitative and The low-dose CAR-T cell groups benefited
distantly related cancer-associated member of phenotypic analysis of peripheral blood T cells more from repetitive RNA-LPX treatment, as
the claudin family. CLDN18.2 is expressed in in treated animals confirmed increased fre- indicated by increased expansion. In vivo ex-
various high–medical need tumors, such as gas- quencies of Thy1.1+ CAR-T cells exhibiting pansion in the high-dose CAR-T cell groups
troesophageal and pancreatic cancers (19–21). an activated phenotype (KLRG1hi, CD62Llow), stagnated after reaching high levels, suggest-
Both in human and mice, its expression in whereas endogenous T cells were not affected ing a saturation threshold that was presum-
normal tissues is restricted to tight junctions at any dose after RNA-LPX treatment (Fig. 3B). ably due to T cells competing for homeostatic
of differentiated cells of the gastric mucosa, The CAR-T cell numbers peaked 3 to 4 days gc-cytokines and niches (Fig. 3C, middle, and
in which it is shielded. Only upon cancer- after RNA-LPX vaccination followed by a de- fig. S7).
associated perturbation of the tight junction cline, mimicking the dynamics of a physio- To assess the impact of repetitive RNA-LPX
architecture does the CLDN18.2 antibody- logical response of antigen-specific T cells vaccination on long-term persistence of CAR-T
binding epitope become exposed (21). Mono- to stimulation, with an initial expansion and cells, CLDN6-CAR-T cell–engrafted mice received
clonal antibodies (22, 23) and CAR-T cells (24) subsequent retraction phase (Fig. 3A and three weekly doses of RNA-LPX followed by
against CLDN18.2 are being evaluated now in fig. S6A). two further RNA-LPX administrations with
clinical studies. We engineered a CLDN18.2- In another experiment, groups of mice re- longer treatment-free intervals (4 and 4.5 weeks).
CAR by substituting the CLDN6-specific scFv ceived different dose levels of CLDN6-CAR-T The first CLDN6-LPX exposure rapidly ampli-
with an anti-CLDN18.2 scFv that exhibits spe- cells, starting as low as 103 cells per mouse, fied CAR-T cells by more than two orders of
cific binding with similar affinity to both human and either were left untreated or received a magnitude, and subsequent weekly treatments
and mouse CLDN18.2 (22). CLDN18.2-CAR-T CLDN6-LPX regimen shortly after ACT. In maintained CAR-T cells at a high level, result-
cells were shown to exert similar functional mice that did not receive CLDN6-LPX, prim- ing in a frequency of more than 15% of total
features as observed for the CLDN6-CAR, in- ary CAR-T cell engraftment (as quantified by peripheral blood lymphocytes (Fig. 3E, left and
cluding strictly antigen-specific activation and bioluminescence) correlated linearly with the middle). For the treatment group in which
killing of tumor cells in vitro (fig. S5A) and number of adoptively transferred cells and re- CLDN6-LPX treatment–free intervals were ex-
complete rejection of advanced CLDN18.2pos mained stable or slowly declined over time tended to up to 35 days, the blood CAR-T cell
tumors in vivo (fig. S5B). CLDN18.2-CAR-T (Fig. 3C, left and middle, and fig. S7, A and B). frequency declined. CAR-T cell numbers did
cells cocultured with CLDN18.2-LPX–treated Notably, in mice treated using the CARVac not drop to the baseline level of engraftment
DCs showed cognate activation and prolifera- strategy, CAR-T cells were expanded irrespec- but rather stabilized at a 10-fold higher fre-
tion (Fig. 2, A and B, bottom). tive of the starting dose. CLDN6-LPX medi- quency. After each treatment-free interval,
Next, we studied the in vivo performance of ated expansion of only 103 CAR-T cells resulted CLDN6-CAR-T cells could be robustly reex-
the CARVac strategy in a series of mouse ex- in detectable frequencies in peripheral blood panded by CLDN6-LPX, indicating memory
periments. Thy1.2+ C57BL/6 mice underwent (Fig. 3C, right). Almost the entire adoptively formation of CAR-T cells. Enrichment of CAR-T
total body irradiation (TBI) for lymphodeple- transferred CAR-T cell population underwent cells with an effector memory (CD127+, CD62Lneg,
tion and were then engrafted with congenic activation and proliferation by RNA-LPX, as KLRG1neg) and a central memory (CD127+, CD62L+,
Thy1.1+ CLDN6-CAR-T cells coexpressing lucif- indicated by transient up-regulation of Ki67 KLRG1neg) phenotype was confirmed by flow
erase (Luc) and green fluorescent protein (GFP) on the majority of transferred T cells (fig. S8A). cytometry (Fig. 3E, right).
and subsequently vaccinated with CLDN6-LPX. The RNA-LPX expanded CLDN6-CAR-T cells Cytokine release syndrome as a clinical man-
In vivo bioluminescence imaging revealed that were fully functional. As compared with CAR-T ifestation of excessive and prolonged secretion

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 5 of 8

RES EARCH | R E P O R T

A LL/2-LLc1-
hCLDN6 TBI
1 x105 CLDN6-CAR-T/ 20 µg B CT26-
mCLDN18.2 TBI
4 x105 CLDN18.2-CAR-T/ 20 µg
RNA-LPX
5 x106 non-transd. T RNA-LPX 3 x106 ctrl-CAR-T
1° 1°
C57BL/6 BALB/c
-19 -7 0 1 Days post ACT -26 -7 0 1 Days post ACT

CLDN6-CAR-T + CLDN6-LPX CLDN18.2-CAR-T + CLDN18.2-LPX

CLDN6-CAR-T + ctrl-LPX CLDN18.2-CAR-T + ctrl-LPX
non-transd. T + CLDN6-LPX ctrl-CAR-T + CLDN18.2-LPX

1° RNA-LPX 1° RNA-LPX
1,500 1,500
ACT

ACT
TBI

TBI
ns 100 100

Tumor volume [mm3]

Tumor volume [mm3]

1,250 1,250

**
1,000 1,000

*
**
****

***
750

****
750

****
50 50
****

***
500 500

250 250

*
*
0 0 0 0
-15 -10 -5 0 5 10 15 20 0 5 10 15 20 25 -15 -10 -5 0 5 10 15 20 25 30 35 10 15 20 25 30 35 40
Days post ACT Days post ACT Days post ACT Days post ACT

105 CLDN6-CAR-T/
C CLDN6-CAR-T D OV90
20 µg
RNA-LPX
1° 2° RNA-LPX 107 non-transd. T

Downloaded from http://science.sciencemag.org/ on March 31, 2021

1° 2° 3°
Total flux [p/s]

106 NSG
-30 0 3 10 17 Days post ACT
CLDN6-LPX
CD45+ CD4+ CD8+

105 CLDN6-CAR-T 107 non-transd. T

CLDN18.2-LPX **** ns
CLDN6-CAR-T + CLDN6-LPX 3.0 67 31
**** 1.5 0.5
CLDN6-CAR-T + ctrl-LPX
105 CLDN6-
non-transd. T + CLDN6-LPX LPX
0 5 10 15
Days post ACT 1° 2° 3° RNA-LPX 45 54
1,200 14.6
ACT

6.1 40.4
CLDN6-
CLDN18.2-CAR-T
Tumor volume [mm3]

1,000 LPX
800
Total flux [p/s]

60 35
106 CLDN18.2-LPX 4.6
600 ctrl- 1.9 8.6
CLDN6-LPX LPX
FSC

CD4

GFP
400

200 CD45 CD8 CAR

105
0 5 10 15 0
Days post ACT -15 -10 -5 0 5 10 15 20 25 30
Days post ACT

E Low-dose CAR-T followed by expansion with RNA-LPX

High-dose CAR-T without vaccination
Fig. 4. Subtherapeutic CAR-T cell doses are efficacious against solid tumors after
1° 2° 3° 4° CARVac Treatment
RNA-LPX vaccination. (A and B) Mice with large established tumors were lymphodepleted
and treated with syngeneic nontransduced or CLDN-CAR–redirected mouse T cells Dose
CAR-T blood counts

Limiting
followed by single intravenous administration of CLDN or control RNA-LPX. Tumor Toxicity
growth (left; mean ± SEM) and survival (right) were determined. CLDN6-CAR was tested
in C57BL/6 mice bearing LL/2-LLc1 tumors transduced with human CLDN6 (n = 9 or Optimal
Therapeutic
10 mice per group; tumor size at start of treatment was 209 mm3) (A), and CLDN18.2-CAR Window
was tested in BALB/c mice bearing mouse CLDN18.2-transduced CT26 (n = 9 mice per
group; tumor size at start of treatment was 78 mm3) (B). (C) Human Luc-expressing Insufficient
CLDN-specific CAR-T cells in naïve NSG mice vaccinated twice with CLDN-LPX. CAR-T cell Therapeutic
Activity
expansion was analyzed by measuring the splenic BLI signal (mean ± SEM of 2 or 3 mice
Time / Persistence
per group). (D) NSG mice with OV90 xenograft tumors (tumor size at start of treatment
was 60 mm3) were treated with a subtherapeutic dose of human CLDN6-CAR (105 cells
per animal) or nontransduced T cells followed by three weekly repetitions of RNA-LPX coding for CLDN6 or a control. Tumor growth curves (left; mean ± SEM of 9 or
10 mice per group) and representative CAR-T cell frequencies after third RNA-LPX treatment in peripheral blood, as assessed by flow cytometry (right), are shown.
(E) Maintaining frequency of circulating CAR-T cells within a therapeutic window by CARVac. P values were determined by two-way analysis of variance with Tukey’s
multiple-comparisons test [(A), left; (B), left; and (D), left]. The time period from ACT until >50% of mice in the control group were euthanized was used for calculation.
Survival benefit was determined with the log-rank test [(A), right; and (B), right]. ns indicates not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

of proinflammatory cytokines in the expansion strategy, we analyzed IFN-g, interleukin-6 and transient increase of IFN-g, no relevant
phase is the most prominent severe adverse (IL-6), and tumor necrosis factor–a (TNFa) increases of the tested proinflammatory cyto-
event of CAR-T cells against B cell markers (25). serum concentrations in gently preconditioned kines were observed (fig. S9A), and treated
To explore the possibility of increased systemic CLDN6-CAR-T cell–engrafted mice after expo- mice were of normal appearance, displaying
cytokine release in conjunction with the CARVac sure to CLDN6-LPX. Except for an early mild regular weight gain over time (fig. S9B).

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 6 of 8

RES EARCH | R E P O R T

Because repeated application of RNA-LPX CAR-T cells in conjunction with CLDN18.2-LPX conformational epitopes, indicating that CARVac
and strong expansion of cytotoxic T cell effec- in the NSG mice xenograft model (fig. S13). is a universally applicable approach.
tors might bear the risk of depletion of APCs in Our study has established two key findings. Our data establish the feasibility and safety
the lymphoid tissues, we analyzed the spleens First, our data support CLDN6 as an oncofetal of single as well as repetitive administration
of treated mice because this is the organ with cell-surface antigen that appears suitable for of CARVac for tunable expansion of engineered
the highest RNA-LPX exposure. Spleens ex- CAR-T cell targeting. In humans, the CLDN6 T cells. RNA-LPX–stimulated CAR-T cells ap-
posed to single or repetitive doses of RNA-LPX gene is strictly silenced in healthy adult tissues pear superior to nonstimulated versions with
did not display any overt pathological altera- but aberrantly activated in various solid tumors regard to cytokine response and cytolytic ac-
tions in spleen architecture or in appearance of high medical need, resulting in expression tivity upon antigen recognition. They form
of red and white pulp (fig. S10, A and B). Flow of high protein levels. This, together with the memory T cells and persist at higher frequen-
cytometry of the cellular composition of spleen feasibility of engineering a CLDN6-directed cies. CARVac not only improves the engraft-
at different time points after repetitive RNA- CAR of high sensitivity, precise specificity, and ment of transferred CAR-T cells but also enables
LPX treatment showed mild and transient re- strong potency against this surface mole- therapeutic tumor control at lower CAR-T cell
ductions of CD11c+ DC and F4/80+ macrophage cule, proposes it as an ideal target for CAR-T doses (Fig. 4E).
populations and no quantitative changes in cell therapy of solid cancers. Tumors without The expansion, retraction, and restimulation
T, B, and NK cell populations (fig. S10C). No homogeneous CLDN6 expression bear the risk kinetics of CAR-T cells mediated by RNA-LPX
changes were noted in the cellular distribu- of outgrowth of antigen-loss variants. However, mimics the physiological process T cells un-
tion of APC subsets in spleen tissue sections activated CLDN6-CAR-T cells are strongly IFN-g– dergo upon antigen-specific priming and boost-
from corresponding time points (fig. S10D). secreting effectors, and hence their antitumor ing. Given that the magnitude of CAR-T cell
Finally, we studied the impact of RNA-LPX activity is thought to drive inflammatory re- expansion depends on RNA-LPX dose, control

Downloaded from http://science.sciencemag.org/ on March 31, 2021

on the therapeutic efficacy of CAR-T cells in modeling of the suppressive tumor microenvi- of the levels of circulating CAR-T cells and
tumor-bearing mice. Lymphodepleted C57BL/ ronment and release of endogenous tumor titration of CAR-T cell frequencies can be
6 mice with large CLDN6pos LL/2-LLc1 Lewis antigens, which together promote antigen- achieved within an optimal therapeutic window.
lung tumors (mean tumor volume 209 mm3) spread and counteract the rapid outgrowth of In addition to lack of suitable targets and
underwent ACT with a subtherapeutic dose antigen-loss variants (26). fast decline of CAR-T cells in the circulation,
of mouse CLDN6-CAR-T cells followed by a We also present the CARVac strategy as an other barriers for efficacy of CAR-T cells in
single injection of CLDN6-LPX or control. approach to improve the antitumor efficacy of human solid cancer exist, including tumor
Tumor control by CLDN6-CAR-T cells alone CAR-T cells. The CAR antigen is displayed in antigen heterogeneity, impaired T cell traffick-
was incomplete, and tumor growth was only its native conformation on the surface of APCs ing and extravasation to tumor sites, exhaustion,
delayed. By contrast, 6 of 10 mice receiving residing in lymphoid compartments, which is and an immunosuppressive microenvironment.
CAR-T cells together with CLDN6-LPX vacci- the ideal setting for costimulation and potent Maintaining optimally stimulated CAR-T cells
nation showed complete rejection of large expansion of T cells. Of note, it is likely that the within a therapeutic window may provide a
tumors, with a significantly higher median same APCs concurrently process and present good foundation for overcoming those con-
survival (Fig. 4A). We reproduced these find- CLDN6 on major histocompatibility complex straints as well.
ings in BALB/c mice with CLDN18.2pos CT26 molecules, which may result in priming and
colon carcinomas (mean tumor volume 78 mm3) activation of endogenous CLDN6-specific T cells REFERENCES AND NOTES
for CLDN18.2 CAR-T cells in conjunction with (18). Recently, different approaches have been 1. S. S. Neelapu et al., N. Engl. J. Med. 377, 2531–2544
a single administration of CLDN18.2-LPX, fur- explored for antigen-specific expansion of CAR-T (2017).
ther supporting the applicability of improving cells (27–31). The CARVac approach presented 2. S. L. Maude et al., N. Engl. J. Med. 378, 439–448 (2018).
3. I. Scarfò, M. V. Maus, J. Immunother. Cancer 5, 28 (2017).
the antitumor effect of CAR-T cells with the here combines various distinctive features. One 4. R. A. Morgan et al., Mol. Ther. 18, 843–851 (2010).
CARVac (Fig. 4B). of the advantages of CARVac is that single- 5. C. H. Lamers et al., Mol. Ther. 21, 904–912 (2013).
To explore the possibility of CARVac for stranded RNA as a natural Toll-like receptor 6. S. A. Richman et al., Cancer Immunol. Res. 6, 36–46
(2018).
human CAR-T cells, we used the CLDN6pos ligand combines delivery of the antigen and
7. K. Turksen, T. C. Troy, Dev. Dyn. 222, 292–300 (2001).
OV90 xenograft tumor model in NSG mice. adjuvanticity in one molecule. Importantly, 8. K. Turksen, T. C. Troy, J. Cell Sci. 117, 2435–2447 (2004).
In pilot experiments, we confirmed that NSG the approach does not require a reengineer- 9. T. Ushiku, A. Shinozaki-Ushiku, D. Maeda, S. Morita,
mice are capable of splenic uptake of RNA ing of the CAR scaffold or adaptation of T cell M. Fukayama, Histopathology 61, 1043–1056 (2012).
10. P. Micke et al., Int. J. Cancer 135, 2206–2214 (2014).
(Fig. 4C and fig. S11A) and of promoting spe- transduction protocols nor does it depend on 11. C. H. June, R. S. O’Connor, O. U. Kawalekar, S. Ghassemi,
cific expansion of human CAR-T cells upon the cumbersome identification and character- M. C. Milone, Science 359, 1361–1365 (2018).
repetitive RNA-LPX administration (fig. S11B). ization of peptide ligands as vaccine mimo- 12. S. L. Maude et al., N. Engl. J. Med. 371, 1507–1517 (2014).
13. M. Kalos et al., Sci. Transl. Med. 3, 95ra73 (2011).
NSG mice bearing advanced CLDN6pos OV90 topes. Nanoparticulate RNA-LPX is fast and 14. D. L. Porter et al., Sci. Transl. Med. 7, 303ra139 (2015).
tumors received a subtherapeutic dose of 1 × inexpensive to produce for any protein-based 15. T. Gargett et al., Mol. Ther. 24, 1135–1149 (2016).
105 CLDN6-CAR+ T cells (fig. S12) followed by antigen. In ongoing clinical trials, the RNA- 16. K. Feng et al., Sci. China Life Sci. 59, 468–479 (2016).
17. D. M. O’Rourke et al., Sci. Transl. Med. 9, eaaa0984 (2017).
repetitive CLDN6-LPX or control treatment LPX vaccine platform is being used for inducing 18. L. M. Kranz et al., Nature 534, 396–401 (2016).
(Fig. 4D). The advanced tumors were com- CD4+ and CD8+ T cell responses against a spec- 19. S. Wöll et al., Int. J. Cancer 134, 731–739 (2014).
pletely rejected in CLDN6-LPX–treated mice, trum of different tumor antigens (NCT02410733, 20. C. Rohde et al., Jpn. J. Clin. Oncol. 49, 870–876 (2019).
21. U. Sahin et al., Clin. Cancer Res. 14, 7624–7634 (2008).
whereas they rapidly progressed in the con- NCT02316457, and NCT03815058), with early 22. U. Sahin et al., Eur. J. Cancer 100, 17–26 (2018).
trol group engrafted with the same CAR-T clinical data supporting the lymphoid target- 23. O. Türeci et al., Ann. Oncol. 30, 1487–1495 (2019).
cell dose (Fig. 4D, left). Effective tumor con- ing and execution of the intended mode of 24. H. Jiang et al., J. Natl. Cancer Inst. 111, 409–418 (2019).
25. J. N. Brudno, J. N. Kochenderfer, Blood Rev. 34, 45–55 (2019).
trol correlated with a high frequency of CAR-T action in humans (18, 32). As exemplified for
26. J. H. Sampson et al., Clin. Cancer Res. 20, 972–984 (2014).
cells in the peripheral blood, proving their ef- CLDN6, CLDN18.2, and also CD19 (fig. S14), a 27. C. Berger et al., Cancer Immunol. Res. 3, 206–216 (2015).
ficient in vivo expansion and improved per- matched RNA vaccine can be immediately gen- 28. C. Y. Slaney et al., Clin. Cancer Res. 23, 2478–2490 (2017).
sistence upon CLDN6-LPX vaccination (Fig. erated and manufactured in good manufactur- 29. M. Tanaka et al., Clin. Cancer Res. 23, 3499–3509 (2017).
30. L. Ma et al., Science 365, 162–168 (2019).
4D, right). As with CLDN6-CAR-T cells, these ing practice (GMP) grade for essentially any 31. X. Wang et al., Clin. Cancer Res. 21, 2993–3002 (2015).
findings were reproduced for human CLDN18.2- existing CAR, including those directed against 32. S. Pektor et al., EJNMMI Res. 8, 80 (2018).

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 7 of 8

RES EARCH | R E P O R T

ACKN OW LEDG MEN TS manuscript. Competing interests: K.R., B.R., P.O., K.Mi., A.B., courtesy of Astellas Pharma GmbH. Material requests should be
We thank B. Jesionek, N. Brüne, C. Stofft, and S. Krapp for N.H., O.K., K.K., Y.O., S.W., E.C., M.B., K.Mr., K.H., L.K., K.K., Ö.T., directed to Astellas Pharma GmbH.
technical support and A. Goß for project management. Funding: and U.S. are employees at BioNTech SE (Mainz, Germany). M.D.
This work was supported by the CI3 Cutting Edge Cluster works as a consultant for BioNTech SE (Mainz, Germany). U.S.,
for Individualized Immune Intervention and funded by the German Ö.T., K.R., B.R., P.O., K.Mi., and K.Mr. are inventors on patents and SUPPLEMENTARY MATERIALS
Federal Ministry of Education and Research (BMBF). Author patent applications, which cover parts of this article. K.R., P.O.,
science.sciencemag.org/content/367/6476/446/suppl/DC1
contributions: U.S. was responsible for the conception and O.K., L.K., V.J., M.D., K.K., Ö.T., and U.S. are stockowners. Ö.T. and
Materials and Methods
experimental strategy of the study. Design and analysis U.S. are management board members of BioNTech SE (Mainz,
Figs. S1 to S14
of the experiments were done by K.R., B.R., P.O., S.W., and E.C. Germany). All other authors declare no competing interests. Data
References (33–37)
K.Mi., A.B., N.H., O.K., K.K., Y.O., K.H., and M.S. performed and materials availability: The results shown here are based in
experiments and acquired the data. D.W. and T.B. performed part on data generated by the TCGA Research Network (www. View/request a protocol for this paper from Bio-protocol.
analysis of RNA sequencing datasets. M.B. and K.Mr. established cancer.gov/tcga; dbGap accession: phs000178) and the Genotype-
in vitro assays. L.K. and M.D. established the RNA-LPX technology. Tissue Expression (GTEx) project (https://gtexportal.org; dbGap 2 July 2019; accepted 18 December 2019
V.J. and K.K. coordinated the GMP manufacturing. K.R., B.R., accession: phs000424.v4.p1). Claudin-specific antibodies IMAB027 Published online 2 January 2020
P.O., Ö.T., and U.S. interpreted the data and drafted the and IMAB362 and their respective anti-idiotype antibodies are 10.1126/science.aay5967

Downloaded from http://science.sciencemag.org/ on March 31, 2021

Reinhard et al., Science 367, 446–453 (2020) 24 January 2020 8 of 8

An RNA vaccine drives expansion and efficacy of claudin-CAR-T cells against solid tumors
Katharina Reinhard, Benjamin Rengstl, Petra Oehm, Kristina Michel, Arne Billmeier, Nina Hayduk, Oliver Klein, Kathrin Kuna,
Yasmina Ouchan, Stefan Wöll, Elmar Christ, David Weber, Martin Suchan, Thomas Bukur, Matthias Birtel, Veronika Jahndel,
Karolina Mroz, Kathleen Hobohm, Lena Kranz, Mustafa Diken, Klaus Kühlcke, Özlem Türeci and Ugur Sahin

Science 367 (6476), 446-453.

DOI: 10.1126/science.aay5967originally published online January 2, 2020

A one-two, CAR-T cell punch

Chimeric antigen receptor (CAR)−T cells have been clinically effective in killing certain hematological
malignancies, but achieving long-term patient responses for solid tumors remains a challenge. Reinhard et al. describe a

Downloaded from http://science.sciencemag.org/ on March 31, 2021

two-part ''CARVac'' strategy to overcome poor CAR-T cell stimulation and responses in vivo. They introduce the tight
junction protein claudin 6 (CLDN6) as a new CAR-T cell target and designed a nanoparticulate RNA vaccine encoding a
chimeric receptor directed toward CLDN6. This lipoplex RNA vaccine promotes CLDN6 expression on the surface of
dendritic cells, which in turn stimulates and enhances the efficacy of CLDN6-CAR-T cells for improved tumor therapy.
Science, this issue p. 446

ARTICLE TOOLS http://science.sciencemag.org/content/367/6476/446

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2019/12/30/science.aay5967.DC1
MATERIALS

RELATED http://stm.sciencemag.org/content/scitransmed/11/511/eaaw9414.full
CONTENT
http://stm.sciencemag.org/content/scitransmed/11/499/eaau5907.full
http://stm.sciencemag.org/content/scitransmed/11/485/eaau7746.full
http://stm.sciencemag.org/content/scitransmed/10/430/eaao2731.full
REFERENCES This article cites 37 articles, 16 of which you can access for free
http://science.sciencemag.org/content/367/6476/446#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS.
Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of
Science. No claim to original U.S. Government Works