NEWS & VIEWS doi:10.

1038/nature21506

IMMUNO LO GY Tcells has been used to target the CCR5 gene

using zinc-finger nuclease enzymes4,5, and

T-cell tweaks to RNA-based nuclease enzymes have integrated

CAR sequences into the CCR5 locus6.
The direct-insertion strategy for CAR

target tumours
sequences chosen by Eyquem and co-workers
used CRISPRCas9 gene-editing tools. The
authors introduced RNA encoding the DNA-
cleaving nuclease enzyme Cas9 into Tcells
Immune cells known as T cells can destroy tumour cells, but their clinical use along with an RNA guide sequence that
requires complex preparation and the cells can lose effectiveness over time. A targets the TRAC sequence. The Cas9 pro-
new approach might improve the efficiency of T-cell therapy. tein used the RNA guide sequence to create a
targeted double-stranded break in the TRAC
DNA sequence. A viral vector with a CAR
M A R C E L A V. M A U S binds a specific antigen. Tcells with TCRs that sequence flanked by sequences homologous
have a high affinity for antigens expressed to the TRAC sequence was introduced into

T
he T cells of the immune system by the bodys own cells (self antigens) are the Tcells. The double-stranded break in the
can kill cells that express a molecule destroyed to avoid autoimmunity. Attacks on TRAC sequence was repaired using the viral
known as an antigen, if the antigen is tumours by Tcells are usually weak because vector sequence as a template, resulting in the
specifically recognized by a T-cell receptor tumours express self antigens. Most current original TRAC sequences being replaced by
protein (TCR). Tcells can be removed from approaches to making CAR Tcells insert the the introduced CAR sequences (Fig. 1).
the body and genetically modified in vitro to CAR-encoding gene into T cells without dis- In genetic engineering approaches, viral
insert the sequence encoding an engineered rupting the resident TCR gene. This means promoter sequences can be used to drive gene
TCR known as a chimaeric antigen receptor that the patients own T cells have to be used expression, which can result in high expres-
(CAR). A CAR can be designed to recognize an to avoid the risk of rejection or graft-versus- sion. However, expression often drops off
antigen expressed in a cancer cell, such as the host disease when the Tcells are transplanted over time. Constitutively active promoters
CD19 antigen. These genetically engineered back into the body. However, a recent study has offer another way to drive gene expression.
cells, called CARTcells, are returned to the explored disrupting the resident TCR -chain Both types of promoter approach have been
patients body, and this approach has success- constant region (TRAC) sequence when cre- shown to drive constitutive CAR signalling in
fully treated some cancers1. In a paper online ating CAR T cells, thereby disabling the TCR Tcells7,8, and this can affect the ability of CAR
in Nature, Eyquem et al.2 have developed and enabling a patient to be treated with CAR Tcells to control tumours. Constitutive T-cell
an approach to integrating DNA sequences T cells made from another persons cells3. activation might cause T-cell exhaustion, a
encoding CARs into Tcells that offers several To create a consistent way to introduce CAR dysfunctional T-cell state characterized by the
advantages over the strategies currently used. sequences into Tcells, Eyquem and colleagues inability to exert antitumour effects and the
During T-cell development, T cells decided to integrate a CAR sequence directly expression of some inhibitory receptor pro-
rearrange gene sequences that encode the TCR into the TRAC. A direct-insertion strategy to teins such as PD-1 (ref.9). Constitutive sig-
so that each Tcell expresses a unique TCR that genetically modify DNA sequences in human nalling might also result in toxicity because of
unusually high release of immune-signalling
molecules such as cytokines.
a b c To test how the pattern of CAR expression
CAR affects the effectiveness of the T-cell response,
T cell
TCR
Eyquem and colleagues tested four promoter
sequences that drive gene expression. The
promoters tested were of different strengths,
and the authors also tested promoterless
TCR CRISPRCas9 CARs integrated into two different gene loci.
Their results demonstrate that integration of
CAR
Inhibitory a promoterless CAR into the TRAC sequence
receptor
results in a more consistent baseline level of
CAR expression, more regulated expression
of the CAR at both the transcriptional and
protein level, and more effective antitumour
Figure 1 | Making CAR T cells. a, T cells can destroy tumour cells if a T-cell receptor protein (TCR) responses in mouse models of leukaemia. This
binds and specifically recognizes an antigen molecule (not shown) that is expressed by the tumour cell. was probably because CAR expression driven
Human T cells can be modified in vitro to insert a sequence into the genome at a random location that by the natural TCR promoter avoids constitu-
encodes a genetically engineered TCR called a chimaeric antigen receptor (CAR). The CAR is designed tive CAR signalling and the CAR expression
to recognize a tumour-cell antigen, and these cells, known as CAR Tcells, are being used in the clinic.
is coupled to T-cell activation. It remains to
If the CAR is expressed constitutively, the cells often eventually express inhibitory receptors that make
the cells enter a dysfunctional state called exhaustion. b, Eyquem et al.2 have developed a way to make be determined whether this CAR-expression
CAR Tcells that uses the gene-editing tool CRISPRCas9 to replace the cells TCR-encoding gene with approach results in any improvement in the
a CAR-encoding sequence flanked by a small region of TCR sequences. c, These CRISPRCas9-edited toxicity profile of CAR Tcells, which remains
Tcells express only one type of TCR, the CAR, and provide better antitumour responses than CAR Tcells a substantial challenge in the clinic.
generated using conventional techniques. In more than 500 patient-years of follow-up

| NAT U R E | 1

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 RESEARCH NEWS & VIEWS

from genetically modified T-cell treatment,
the random integration of viral vectors has not
been implicated as the cause of any cancers10.
However, the risk is present, and the long-term
follow-up required in these studies is expen-
sive; perhaps both could be eliminated if there
were sufficient confidence in the rarity of off-
target effects in the method used. If Eyquem
and co-workers approach is used, CRISPR
Cas9 off-target gene edits in CAR Tcells are
likely to occur only at defined regions of the
genome, and the risks from such edits should
be easier to assess, given that the technologies
for identifying the frequency and location of
such off-target edits are improving rapidly11.
The use of CAR Tcells has resulted in some
successful clinical outcomes, especially in
CD19-expressing B-cell acute leukaemias1.
However, there is room for improvement in
the treatment of some CD19-expressing lym-
phoma cancers and of solid tumours, for which
CAR T-cell therapy has not been very effec-
tive. Treatment of some of these conditions
might be hindered by T-cell exhaustion. Other
researchers have combined CAR Tcells with
gene editing to remove exhaustion-associated
inhibitory pathway proteins12. In the future, it
might be interesting to compare how different
combinations of gene knockouts (such as tar-
geting T-cell-exhaustion pathways) and CAR
integration might affect T-cell function. Such
analysis could enhance our understanding of
T-cell biology in a way that also illuminates
therapeutic uses.
Eyquem and colleagues strategy provides
three important improvements for T-cell-
based therapies. First, CAR Tcells generated
using their method made more-effective anti-
tumour responses than CAR Tcells generated
using current standard techniques. Second,
the targeted nature of their CAR integration
might prove safer than random integration,
which carries the potential risk of generat-
ing a harmful mutation. Third, this approach
might enable off-the-shelf CAR Tcells to be
made that need not come from a patients own
Tcells. This would enable easier and cheaper
manufacture of CAR Tcells, an outcome that
would be particularly helpful for the treatment
of severely immunocompromised patients.
Marcela V. Maus is in the Department of
Medicine, Harvard Medical School, and
the Massachusetts General Hospital Cancer
Center, Boston, Massachusetts 02114, USA.
e-mail: mvmaus@mgh.harvard.edu
1. Maus, M. V. & June, C. H. Clin. Cancer Res. 22,
18751884 (2016).
2. Eyquem, J. et al. Nature http://dx.doi.org/10.1038/
nature21405 (2017).
3. Qasim, W. et al. Sci. Transl. Med. http://dx.doi.
org/10.1126/scitranslmed.aaj2013 (2017).
4. Lombardo, A. et al. Nature Methods 8, 861869
(2011).
5. Wang, J. et al. Nucleic Acids Res. 44, e30 (2016).
6. Sather, B. D. et al. Sci. Transl. Med. 7, 307ra156
(2015).
7. Frigault, M. J. et al. Cancer Immunol. Res. 4,
356367 (2015).
8. Long, A. H. et al. Nature Med. 6, 581590 (2015).
9. Wherry, E. J. & Kurachi, M. Nature Rev. Immunol. 15,
486499 (2015).
10. Scholler, J. et al. Sci. Transl. Med. 4, 132ra53 (2012).
11. Tsai, S. Q. & Joung, J. K. Nature Rev. Genet. 17,
300312 (2016).
12. Ren, J. et al. Clin. Cancer Res. http://dx.doi.
org/10.1158/1078-0432.CCR-16-1300 (2017).

2 | NAT U R E |

2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.