Perspective

CAR immune cells: design principles,

resistance and the next generation

https://doi.org/10.1038/s41586-023-05707-3 Louai Labanieh1,2,3 & Crystal L. Mackall2,3,4,5 ✉

Received: 5 March 2022

Accepted: 4 January 2023 The remarkable clinical activity of chimeric antigen receptor (CAR) therapies in B cell
Published online: 22 February 2023 and plasma cell malignancies has validated the use of this therapeutic class for liquid
cancers, but resistance and limited access remain as barriers to broader application.
Check for updates
Here we review the immunobiology and design principles of current prototype CARs
and present emerging platforms that are anticipated to drive future clinical advances.
The field is witnessing a rapid expansion of next-generation CAR immune cell
technologies designed to enhance efficacy, safety and access. Substantial progress
has been made in augmenting immune cell fitness, activating endogenous immunity,
arming cells to resist suppression via the tumour microenvironment and developing
approaches to modulate antigen density thresholds. Increasingly sophisticated
multispecific, logic-gated and regulatable CARs display the potential to overcome
resistance and increase safety. Early signs of progress with stealth, virus-free and
in vivo gene delivery platforms provide potential paths for reduced costs and
increased access of cell therapies in the future. The continuing clinical success of
CAR T cells in liquid cancers is driving the development of increasingly sophisticated
immune cell therapies that are poised to translate to treatments for solid cancers and
non-malignant diseases in the coming years.

CARs are synthetic modular proteins that redirect immune cell reactiv- responses have not been demonstrated (Table 1). Autologous cell manu-
ity toward a target of interest. This versatile platform has demonstrated facturing is labour-intensive and expensive and commercial scaling is
substantial clinical effects in the treatment of B cell and plasma cell not yet adequate to meet clinical needs. This Perspective synthesizes
malignancies, and the potential to expand its application is driving current understanding of the immunobiology of CAR T cells, emphasiz-
rapid technological developments and large investments from aca- ing resistance mechanisms in cancer, design principles and emerging
demia and the biopharmaceutical industry. Six CAR T cell products approaches to enhance efficacy. We focus primarily on developing CAR
have been approved by the US Food and Drug Administration (FDA) T cells for cancer treatment, but many of the principles are relevant to
for 12 indications, including large B cell lymphoma1–4 (LBCL), B cell other immune cell therapies for cancer and to nascent efforts to develop
acute lymphoblastic leukaemia5–7 (B-ALL), mantle cell lymphoma8 and cell therapies for non-malignant diseases. Owing to space constraints,
follicular lymphoma9. In pivotal trials, CD19-CAR therapy outperformed we focus primarily on the most recent literature and on the emerging
standard of care (SOC) as second line therapy for LBCL10,11, and was efforts to enhance efficacy, and refer the reader to recent authorita-
highly effective as a first line therapy12, paving the way for its applica- tive reports for additional information on CAR-related toxicities30 and
tion in earlier-stage disease. The generalizability of the CAR platform clinical outcomes31.
beyond CD19 targeting is now established, with two BCMA-CAR T cell
therapies (BCMA is also known as TNF receptor superfamily mem-
ber 17) having been approved by the FDA for treatment of multiple CAR T immunobiology and mechanisms of resistance
myeloma13,14, and high response rates with CD22-CARs in B-ALL15,16 Sustained broad-based advances by many groups focused on develop-
and LBCL17, CD30-CARs in Hodgkin lymphoma18, CD7-CARs in T cell ing immune cell therapies for cancer have been essential for the success
acute lymphoblastic leukaemia19–22 (T-ALL), CD20-CARs in LBCL23, and of CAR T therapy (Fig. 1). CARs were invented by Eshhar and colleagues
GPRC5D-CARs in multiple myeloma24 (Table 1). Standardized toxicity with the goal of harnessing the expansion, killing and persistence of
grading and management has resulted in low treatment-related mor- natural T cells while overcoming major histocompatibility complex
tality with current commercial CAR T cells1–11. (MHC) restriction of the T cell receptor (TCR), to enable broader thera-
Despite this progress, many challenges remain. Fewer than 50% of peutic applicability32,33. After iterative optimization by many groups34,35,
patients treated with commercial CAR T cells for B cell malignancies a receptor incorporating a scFv as the antigen-binding domain, a hinge/
experience durable disease control1–7. CAR T cells have shown signs transmembrane domain, TCRζ and a CD28 or 4-1BB costimulatory
of activity in solid tumours25–29, but high rates of consistent durable endodomain emerged as the CAR prototype (Fig. 2). This architecture

1
Department of Bioengineering, Stanford University, Stanford, CA, USA. 2Center for Cancer Cell Therapy, Stanford Cancer Institute, Stanford University, Stanford, CA, USA. 3Parker Institute for
Cancer Immunotherapy, San Francisco, CA, USA. 4Division of Hematology, Oncology, Stem Cell Transplantation and Regenerative Medicine, Department of Pediatrics, Stanford University,
Stanford, CA, USA. 5Division of Blood and Marrow Transplantation and Cell Therapy, Department of Medicine, Stanford University, Stanford, CA, USA. ✉e-mail: [email protected]

Nature | Vol 614 | 23 February 2023 | 635

Perspective
Table 1 | Targets of CAR T cell therapies with clinical evidence of efficacy

Target Disease Response ratea Survival Comments Date of FDA Refs.

approval
Haematological malignancies
CD19 B-ALL CR or CRi: 81%, EFS: 50% Tis-cel approved for R/R B-ALL Aug 2017 5

OS: 76% at 12 months (≤25 yr of age)

CD19 LBCL CR: 58%, PFS: 44% at 12 months Axi-cel approved as 3rd line treatment Oct 2017 3

OS: 52% at 18 months for LBCL (>18 yr of age)

CD19 LBCL CR: 40% RFS: 65% Tis-cel approved as 3rd line treatment May 2018 4

OS: 49% at 12 months for LBCL (>18 yr of age)

CD19 MCL CR: 67% PFS: 61% Brex-cel approved for R/R MCL July 2020 8

OS: 83% at 12 months (>18 yr of age)

CD19 FL CR: 74% PFS: 65% Axi-cel approved as 3rd line treatment Mar 2021 222

OS: 87% at 18 months for R/R FL (>18 yr of age)

CD19 LBCL CR: 53% PFS: 44% Liso-cel approved for 3rd line LBCL Feb 2021 1

OS: 58% at 12 months (>18 yr of age)

BCMA MM CR: 33% Median PFS: 8.8 months Ide-cel approved for 5th line treatment Mar 2021 13

OS: 78% at 12 months for MM (>18 yr of age)

CD19 B-ALL CR: 56% RFS: 58% at 6 months Brex-cel approved for R/R B-ALL Oct 2021 223

OS: 71% at 12 months (>18 yr of age)

BCMA MM sCR: 67% PFS: 77% Cilta-cel approved for 5th line MM Feb 2022 14

OS: 89% at 12months (>18 yr of age)

CD19 FL CR: 69% PFS: 67% at 12 months Tis-cel approved for 3rd line treatment May 2022 9

of FL (>18 yr of age)
CD19 LBCL (Axi-cel vs SOC) (Axi-cel vs SOC) Axi-cel approved as 2nd line treatment April 2022 10

CR: 65% vs 32% EFS: 41% vs 16% for LBCL (>18 yr of age)
OS: 61% vs 52% at 24 months
CD19 LBCL (Liso-cel vs SOC) (Liso-cel vs SOC) Liso-cel approved as 2nd line treatment June 2022 11

CR: 66% vs 39% EFS: 45% vs 24% for LBCL (>18 yr of age)
OS: 79% vs 64% at 12 months
CD19 LBCL CR: 78% PFS: 75% Front line therapy for high-risk LBCL 12

OS: 91% at 12 months

CD22 B-ALL CR: 70% Median RFS: 6 months CD19-CAR T cell therapy had failed in 15,16

Median OS: 13.4 months 88% of these patients

CD22 LBCL ORR: 86% Median PFS: not reached CD19-CAR T cell therapy had failed in 17,224

CR: 67% 95% of these patients

CD30 HL CR: 59% PFS: 36% Greater CD30 CAR T persistence and 18

OS: 94% at 12 months higher PFS with fludarabine-based LD

CD7 T-ALL CR: 90% Not available Allogeneic donor-derived CD7-CAR 20

T cells; GVHD grade 1–2 in 60% of

patients
CD7 T-ALL or TLBL CR: 7/8 Not available Autologous CD7-CAR T cells rendered 19

fratricide-resistant using a CD7 PEBL

CD38 AML CR or CRi: 4/6 50% relapse rate at 6 months Allo-HSCT refractory patient 225

population; no off-target effects on

monocytes or lymphocytes
κ light chain NHL, CLL, or CR: 2/9 Not available No or limited pre-treatment LD. One CR 226

MM sustained for at least 3 yr

CD20 LBCL CR: 54.5% PFS 41.7% at 24 months All patients had prior rituximab; longest 23

CR at least 57 months
Solid tumours
GD2 NB CR: 27% of patients with Median OS: 31 months 1st generation CAR expressed by 227

active disease EBV-reactive T cells; one patient had

sustained CR for at least 60 months
GD2 DMG 9/10 patients with Not available Initial IV infusion followed by multiple 28,228

radiographic or clinical ICV infusions; one patient had >95%

benefit reduction in tumour volume
HER2 Sarcomas CR: 27% Not available No on-target, off-tumour toxicity 229

of HER2-CARs; patient with RMS

metastatic to bone marrow had a CR for
>12 months
IL-13Rα2 GBM CR: 1/1 Not available CR sustained for 7.5 months with 27

16 locoregional administrations over

220 days
Continued

636 | Nature | Vol 614 | 23 February 2023

Target Disease Response ratea Survival Comments Date of FDA Refs.
approval
EGFR BTC CR: 6% Median PFS: 4 months One out of 17 patients achieved a CR 230

for at least 22 months. Manageable

mucosal toxicities.
Mesothelin MPD 11% complete metabolic OS: 83% at 1 yr Regionally delivered intrapleural CAR 98

response by PET Median OS: 23.9 months T cell administration plus PD-1 blockade
Claudin-18.2 GC or PC ORR: 48.6% Median PFS: 3.7 months 83% of patients showed tumour 29

Disease control rate: 73.0% OS: 81% at 6 months regression; 11% showed reversible grade
3/4 gastrointestinal toxicities
PSMA MCRPC 5/13 patients had >30% Median PFS: 4.4 months PSMA CAR T cells expressing a 172

reduction in PSA Median OS: 15.9 months dominant-negative TGFβRII. 5 out of 13

patients with high-grade CRS, one fatal
Non-malignant disease
CD19 SLE 1/1 clinical remission Not available Rapid decrease of autoantibodies from 213

>5,000 U ml−1 to 4 U ml−1
a
If fewer than ten patients were treated, absolute response numbers are provided as a fraction; otherwise, they are provided as the percentage response rate.
Allo-HSCT, allogeneic haematopoietic stem cell transplant; AML, acute myeloid leukaemia; BTC, biliary tract cancer; CLL, chronic lymphocyte leukaemia; CR, complete response; CRi,
complete response with incomplete haematologic recovery; CRS, cytokine release syndrome; DMG, diffuse midline glioma; EBV, Epstein–Barr virus; EFS, event-free survival; FL, follicular
lymphoma; GBM, glioblastoma; GC, gastric cancer; GVHD, graft-versus-host disease; HL, Hodgkin lymphoma; ICV, intracerebroventricular; IV, intravenous; LD, lymphodepletion; MCL, mantle
cell lymphoma; MCRPC, metastatic castration-resistant prostate cancer; MM, multiple myeloma; MPD, malignant pleural disease; NB, neuroblastoma; NHL, non-Hodgkin lymphoma; ORR,
overall response rate; OS, overall survival; PC, pancreatic cancer; PEBL, protein expression blocker; PET, positron emission tomography; PFS, progression-free survival; PSA, prostate-specific
antigen; PSMA, prostate-specific membrane antigen; RFS, relapse-free survival; RMS, rhabdomyosarcoma; R/R, relapsed or refractory; sCR, stringent complete response; SLE, systemic lupus
erythematosus; TLBL, T cell lymphoblastic lymphoma. Axi-cel, axicabtagene ciloleucel; brex-cel, brexucabtagene autoleucel; cilta-cel, ciltacabtagene autoleucel; ide-cel, idecabtagene
vicleucel; liso-cel, lisocabtagene maraleucel; tis-cel, tisagenlecleucel.

is utilized by five out of the six FDA-approved agents, with the sixth the higher affinity of single-chain variable fragments (scFvs) com-
incorporating the same architecture with two nanobody heavy chains pared with TCRs and the generally higher density of CAR expression
(Vhh) as the antigen-binding domains14. Antigen engagement of the pro- compared with TCR–CD3 complexes36, TCRs induce full activation
totype 1.5–2.2 kilobase (kb) receptor largely replicates antigen specific in response to less than 100 peptides per antigen presenting cell39,40,
activation and killing mediated by the TCR–CD3 complex in natural whereas CARs require more than 1,000 target molecules per target
T cells; however, significant distinctions exist between the biology cell15,41–45. The basis for the difference includes diminished proximal
of CAR T cells and natural T cells that provide opportunities and chal- kinase recruitment by CARs38,44,46,47, a less developed immune syn-
lenges for application of these therapeutic agents, as discussed below. apse46, reduced engagement of co-receptors and greater induction
of negative downstream regulators36—in part related to tonic signal-
Resistance due to antigen modulation ling, in which CAR aggregation, often driven by the scFv, induces
A major distinction between CAR and TCR signalling is that CARs antigen-independent activation48. Modifications to the design of CAR
require higher antigen density for full T cell activation36–38. Despite prototypes can tune the antigen density threshold to some extent, with

Biopharma Tis-cel approved by FDA for

invests in CAR children and young adults
T cell therapies with R/R B-ALL—the 1st cell FDA approval of:
therapy approved for cancer • axi-cel for R/R
10-yr follow up of treatment and 1st gene follicular lymphoma
2nd generation Safety and patients with HIV therapy approved in the USA • liso-cel for 3rd line
CD3/CD28 CAR prototype antitumour activity following CD4-TCRζ LBCL
scFv linked to beads induce with CD28 co- of 1st generation CAR (1st generation Axi-cel • ide-cel for 5th line
TCRα or TCRβ 1st generation potent ex vivo stimulation GD2-CAR T cells prototype) reveals approved by multiple myeloma
activates T cell CAR prototype expansion of shows enhanced shown in safety and cell FDA for 3rd line • brex-cel for R/R
hybridoma32 described33 human T cells232 potency34,233 neuroblastoma236 persistence77 LBCL B-ALL in adults

1991 2004 2008 2012 2018 2020 2022

1989 1993 1998 2001–2002 2009 2010–2015 2017 2021

CD8–TCRζ chimeric Lymphopenia elevates CD19.BB.z-CAR Lymphodepletion Tis-cel Brex-cel FDA approval of:
receptor activates homeostatic cytokines shows potent activity plus CD19.28z-CAR approved approved by • cilta-cel for 5th line
Jurkat cells231 and enhances efficacy in preclinical or CD19.BB.z-CAR by FDA for FDA for R/R multiple myeloma
of adoptive T cell leukaemia models35 induce potent 3rd line mantle cell • axi-cel and liso-cel for
transfer234,235 sustained benefit in LBCL lymphoma 2nd line LBCL
patients with • tis-cel for R/R
CD19.BB.z-CAR shows refractory B cell follicular lymphoma
enhanced anti-leukemic lymphomas, B-ALL
efficacy and prolonged and CLL58,238–241 GD2-CAR mediates
survival in mouse benefit in patients with
xenograft models of diffuse midline glioma28
primary human B-ALL237.
Significant activity of
claudin-18-CAR
in gastrointestinal
carcinomas29

Concept phase Prototype iteration Clinical proof of concept Commercialization

Fig. 1 | Timeline of key milestones in CAR T cell development.A timeline of developments in CAR T cell therapies231–241.

Nature | Vol 614 | 23 February 2023 | 637

Perspective
Linker
scFv
VH VL
• Affinity or antigen density
threshold242
• VH–VL pairing, tonic signalling FW
and potency (linker)78
• Epitope proximity52
• Tonic signalling (framework)48,243 CDR Hinge or spacer
• Immunogenicity140,244 • Optimal distance for
synapse formation242
Transmembrane • Dimerization or
• Homo- or oligomerization103
hetero-dimerization245 • Antigen density threshold
• Association with endogenous Cell and potency44,49
proteins246 membrane
• CAR expression level247
• Antigen density threshold and
potency44,49
Costimulation
• Persistence237,248
• Memory formation75
• Metabolic requirements249
• Antigen density threshold
and potency44,49,75
CD3ζ
• Turning down signal to avoid
dysfunction79
• Turning up signal for increased
potency44
• Resting T cells by CD3ζ         
inhibition121

Fig. 2 | CAR structure–function relationships. Prototype CARs comprise a Small modifications in CAR structure can have profound effects on CAR T cell
target-binding domain such as a scFv, a hinge or spacer domain that projects function, including modulation of the antigen density threshold for activation,
the binder away from the cell surface and provides conformational flexibility, persistence, potency, tonic signalling, CAR expression level and the propensity
a transmembrane domain that anchors the receptor in the cell surface, and for dimerization237,242–249. CDR, complementarity-determining region; FW,
costimulatory and CD3ζ signalling domains that provide activation signals. framework region.

features that modulate signal strength, scFv affinity, CAR expression shedding69. Downregulation of some targets is amenable to therapeutic
density, hinge/transmembrane architecture and synapse spacing, intervention with small molecules, such as CD22 upregulation by bry-
having significant effects44,49–52 (Fig. 2). Because greater signal strength ostatin70, CD70 upregulation by azacitidine71 and BCMA upregulation
lowers the antigen density threshold, features that enhance T cell fit- by γ-secretase inhibitors, which inhibit antigen shedding69,70.
ness independently of CAR design also reduce the CAR antigen density
threshold44,53. These insights are foundational for developing safe and Resistance due to inadequate T cell function
effective CAR T cell therapies, since toxicity and efficacy are intimately A second major cause of CAR T cell resistance is related to inadequate
related to the expression characteristics of the targeted antigen. CARs T cell potency, persistence, functional persistence and/or dysfunction,
targeting molecules that are absent from vital tissues, such as B cell and is typically associated with disease recurrence in the absence of
lineage antigens, should be engineered for activation at low antigen antigen modulation. Dysfunction often results from T cell exhaustion,
density to diminish the risk of low antigen recurrence, whereas CARs characterized by global transcriptional and epigenetic reprogramming
targeting molecules that are highly expressed in cancer but with low that converges on terminal differentiation53,72. T cells in the apher-
expression in vital tissues should be engineered with higher antigen esis and/or manufactured CAR T cell product sometimes manifest
density thresholds to exploit a therapeutic window on the basis of exhaustion73,74, and high tumour burdens induce exhaustion following
differential antigen density49. adoptive transfer. CAR-intrinsic factors also contribute to exhaustion,
Antigen modulation is a major cause of CAR T cell resistance in B cell with the costimulatory domain having a major role. CD28-costimulated
malignancies, and is likely to pose an even greater challenge in solid CARs manifest more rapid and greater expansion, secrete more inflam-
tumours, where most targetable antigens show significant heteroge- matory cytokines, and show limited persistence owing to T cell exhaus-
neity54–56. In B-ALL in children and young adults, approximately 50% of tion when compared with 4-1BB and first-generation CARs, which in
relapses are associated with CD19 loss5,57,58, and in LBCL, approximately some cases, may persist for years5,58,75–77. The pro-exhaustion effect of
30% of relapses are CD19-negative and an additional 30% express CD19 the CD28 costimulatory domain is magnified in CARs with tonic sig-
at levels below the antigen density threshold for commercial CARs45,59. nalling48,53. The effect of tonic signalling depends on the magnitude of
The role of antigen modulation in resistance to BCMA-CAR T therapy the signal and is context-dependent, with some CARs demonstrating
in multiple myeloma is less well defined. Baseline BCMA expression enhanced function and persistence in the presence of tonic signalling78.
levels are heterogeneous among patients and have not been associ- Tuning down the signalling strength of CD28-based CARs through
ated with clinical responses60–62. BCMA loss—associated with genetic mutations in CD3ζ or CD28 domains can attenuate their propensity for
mutation and deletion—is a rare cause of resistance13,63,64 (less than 5% of exhaustion and improve persistence79,80, as can mutations that interfere
cases), however, antigen modulation is observed following BCMA-CAR with downregulation and ubiquitination of 4-1BB-costimulated CARs81.
treatment61,62. Antigen density can be modulated through a variety Of interest, a recent long-term follow-up study demonstrated that
of mechanisms, including genetic mutation63,65,66, alternative RNA long-lived CARs were CD4-positive, raising the prospect that this sub-
splicing65, cell lineage switching67, epigenetic and/or posttranscrip- set may be less susceptible to exhaustion and thereby exhibit greater
tional mechanisms15, trogocytosis50, hyperglycosylation68 and antigen persistence76,82.

638 | Nature | Vol 614 | 23 February 2023

 Box 1

Next-generation CAR enhancements

Enhancements in development that are in clinical testing are listed in bold; those at proof-of-concept stage are underlined.
Multispecificity and/or logic gating
OR gates: bivalent45,100,104,105, bicistronic103, two vectors101,102, co-infusion250, higher order251
NOT gates: PD-1179,180, CTLA-4179, TIGIT180, LIR-1181
IF–THEN gates: synNotch173,174
AND gates: split CAR252, LINK CAR176
TME-localized: masked CAR253, hypoxia sensing254, conditional granzyme255, EGFR BiTE114
Fitness enhancements
Gene overexpression: c-Jun53, BATF117, PGC1A (also known as PPARGC1A)256
Gene knockout: PD1 (also known as PDCD1)123, TGFBR2142, HPK1 (also known as MAP4K1)132, NR4A2126, RASA2127,131, TET2115, DNMT3A116,
TOX257, CBLB127, CD5127, SOCS1127, TCEB2 (also known as ELOB)127, REGNASE-1 (also known as ZC3H12A)128, DHX37129, PTPN2130, IKZF2124, TLE4124,
ID372, SOX472, ACAT1258, adenosine A2A receptor133, diacylglycerol kinase259, LAG3260, GM-CSF (also known as CSF2)261, mediator kinase
module125
Small molecule: dasatinib28,121, AKT inhibitor119, ibrutinib120
Suicide gene: EGFRt153, CD20 epitope262, HER2t263, HSV-tk152, iCasp9138,151
Regulatable platforms
AP1903-inducible costimulation264, antibody-coupled265, fluorescein-CAR266, switchable CAR T cells267, antigen receptor complex
(ARC) T cells268, SNIP162, SUPRA177, co-LOCKR178, drug-regulated degrons121,157,161, drug-induced dimerization156,157, drug-activated binders159,
CAR disruption158, protease-cleavable CARs162,163, Tet-inducible155, PROTACs160, ultrasound269, light270
Armouring
Dominant negative: TGFβR171,172, PD-1143, FAS145
Checkpoint: PD-1-Fc, anti-PD-1 scFv146
Cytokines: IL-12109,110,271, IL-18111,272,273, NFAT-induced IL-12274, IL-15 and IL-21275
Switch receptors: PD-1–CD28276, IL-4R–IL-2Rβ277, IL-4R–IL-7R147, FAS–41BB148, GM–IL-1885
Other: FAP-CAR278, heparanase overexpression279, catalase overexpression280, solHVEM281, PKA disruptor282
Engaging the endogenous immune system
Chemokine overexpression: CXCR5283, CCR4284, CCL1999, CCR2285, CX3CR1286
Other: CD40L overexpression112, RN7SL1 extracellular vesicles108, FLT3L overexpression113
Expansion and persistence
IL-15: secretion138,196, sushi domain287, tethered288
IL-7: secretion99, mutant constitutive136, tethered289
Other: JAK–STAT CAR137, ortho IL-2290, IL-2–IL-9 chimera291, chimeric costimulatory receptor107,292
Alternative signalling
TRuC185, Ab–TCR184, STAR183, HIT182
Stealth or fratricide resistant
Knockout: TRAC150,190,262, B2M293, CIITA191, CD52190, CD7134, CD5135
Overexpression: HLA-E192, CD47193
Endoplasmic reticulum retention: CD7 PEBL19, CD3 PEBL294

The clinical effect of shorter persistence of CD28- versus signalling has been demonstrated to be required for productive CAR
4-1BB-costimulated CARs varies by disease. In LBCL, tumour eradi- T cell adhesion and cytotoxicity in solid but not liquid tumours91. CAR
cation occurs rapidly, and CD28 and 4-1BB costimulated CAR T cells T cell potency is also limited by immunosuppressive molecules (TGFβ,
demonstrate similar efficacies1,3,4,83. By contrast, in B-ALL, persis- IL-10, IL-6 and checkpoint molecules) in the tumour microenviron-
tence of CAR T cells beyond 6 months is associated with increased ment (TME), and work is underway to combine CAR T cell therapies
rates of relapse—thus, CD28 costimulated CAR T cells are less effec- with immunomodulators designed to activate immunity within the
tive unless patients receive a post-CAR bone marrow transplant to TME and/or to arm immune cells to resist specific immunosuppressive
consolidate remission57,84. In multiple myeloma, functional persis- mediators (Box 1).
tence of anti-BCMA-CAR T cells is associated with a longer duration of
response13. It remains unclear whether CD28 or 4-1BB costimulation is Impaired trafficking and locoregional delivery
preferred for solid tumours, where both strong signalling strength and Impaired trafficking to the tumour site may also limit CAR T cell effi-
persistence are desirable. Prototype CARs incorporating both CD28 cacy, especially in solid tumours. In preclinical models of central nerv-
and 4-1BB costimulatory domains have not demonstrated superiority, ous system tumours, intratumoral or intracerebroventricular (ICV)
leading investigators to integrate novel85,86 or synthetic costimula- administration has improved therapeutic benefit, with an approxi-
tory domains87 with the goal of endowing maximal signalling power mate tenfold lower regional dose being required to achieve the same
alongside durable persistence. Pooled CAR screening has been used efficacy as intravenous administration92–94. Several clinical trials have
to identify optimal CAR signalling domains and designs and elucidate demonstrated the safety of locoregional delivery of CAR T cells into
CAR design principles87–89. Genome-wide CRISPR screens have identi- the central nervous system27,28,95 and in a patient with glioblastoma
fied the CD2–CD58 axis as a mediator of T cell potency90 and IFNγR multiforme, ICV delivery of IL-13Rα2 CAR T cells induced a complete

Nature | Vol 614 | 23 February 2023 | 639

Perspective
Prototype Bivalent SynNotch LINK CAR SNIP CAR Stealth CAR Transcriptional TRuC HIT Tethered
CAR CAR reprogramming cytokine
Priming Cytotoxic Antigen Antigen
Antigen Antigen antigen antigen A B
A B HLA-E
scFv
scFv CD47 Vα Cα Linker Cytokine
Vβ VL VH
Cβ Cα Cβ
Hinge Cleavage
TM site
δ γ
Costim LAT SLP-76 ε ε δ γ
TF ζζ ε ε
c-Jun ζζ
CD3ζ Cytokine
overexpression
CAR NS3 receptor
expression T cell protease signalling
activation

mRNA
Protease
inhibitor Growth and
B2M T2
TET2 survival
ON
TRAC NR4A
TF response Gene knockouts Gene knockouts
element

Conventional OR gate IF–THEN gate AND gate Drug-regulated Allogeneic Improved Alternative Alternative Persistence
CARs fitness signalling signalling

Fig. 3 | Next-generation platforms. Investigators have leveraged numerous leukocyte antigen (HLA)-E and CD47 overexpression shield stealth CAR T cells
bioengineering strategies to develop advanced CAR platforms to improve the from natural killer cell- and macrophage-mediated rejection. Deletion of NR4A
safety and efficacy of immune cell therapies. Multispecific (bivalent) CARs, and TET2 genes or ectopic overexpression of c-Jun results in transcriptional
which operate as an OR gate, may be able to overcome obstacles related to rewiring that renders T cells exhaustion resistant. By fusing a scFv domain
tumour heterogeneity and antigen loss, whereas combinatorial antigen to TCR subunit, TRuC and HIT receptors redirect the specificity of the
sensing systems such as SynNotch or LINK CAR increase the specificity of endogenous receptor and exhibit increased antigen sensitivity compared
CAR T cells by requiring two antigens for activity. Safety switches such as with CARs, while secreting lower levels of cytokine. Integration of a cytokine
drug-regulated and adapter CARs could mitigate CAR-induced toxicities and gene in the CAR vector provides a growth and survival signal to improve the
enhance efficacy by tuning signalling. Knockout of TRAC, B2M and CIITA genes persistence of T cells. Costim, costimulatory domain; TF, transcription factor;
ablates TCR, MHC class I and MHC class II expression, respectively, and human TM, transmembrane domain.

response, whereas intratumoral administration was not effective27. mediated diminished potency toward CD22 and tumour cell variants
In a study of patients with diffuse midline gliomas, ICV delivery of exhibiting low or no surface expression of CD19 emerged45. Cilta-cel,
GD2-CAR T cells induced antitumour effects and clinical responses, a BCMA-CAR recently approved by the FDA, incorporates two tandem
and repeated dosing was associated with sustained benefit, raising Vhh binders in one receptor, which binds two different epitopes on
the prospect that delivery to the central nervous system may abrogate BCMA. Clinical results with cilta-cel demonstrate a sCR of 83% and 55%
immune sensitization, which has probably limited the effectiveness PFS at 27 months, the highest reported so far using CARs for multiple
of multidose intravenous CAR T cell regimens28,96,97. In patients with myeloma14,106 (Table 1). In summary, clinical data with multispecific
lung cancer involving the pleura, regional delivery of mesothelin-CAR CARs is nascent but demonstrates safety and promise for improved
T cells in combination with PD-1 blockade mediated stable disease efficacy by diminishing antigen escape.
and metabolic responses98. Cell-intrinsic strategies to improve T cell Novel receptors designed to lower the antigen density threshold
homing to and persistence in the TME, such as secretion of IL-7 and are also being developed. Katsarou and colleagues have expressed a
CCL1999, are also being explored. chimeric costimulatory receptor (CCR), which lacks a CD3ζ domain,
in trans with a prototype CAR, and reported that CCR engagement
activated the prototype CAR at very low antigen density, preventing
The next generation of CAR T cells antigen low escape in preclinical models107. Induction of antitumour
The various next-generation platforms being used to overcome tumour responses toward non-CAR T cell antigens—as reported following CAR
resistance mechanisms, augment immune cell fitness, improve specific- T cell therapy in a patient with rhabdomyosarcoma—could diminish
ity, tune CAR signalling, enhance safety, and increase antigen sensitivity  resistance due to antigen modulation54. Several approaches are under
are discussed in this section. development to augment innate and adaptive immunity (Box 1), includ-
ing CAR-mediated delivery of the immunostimulatory RNA RN7SL1108,
Platforms to diminish antigen escape coexpression of ligands or cytokines that reshape the TME such as
Bispecific CAR targeting may be achieved by administration of a IL-12109,110, IL-18111, CD40L112 or Flt3L113, engineering CAR T cells to secrete
mixed cell product, bicistronic expression of two receptors, two scFvs bispecific T cell engagers (BiTEs), taking advantage of CAR T cell accu-
incorporated into a single receptor100, or co-transduction of multiple mulation within the tumour site and avoiding systemic toxicity of the
CARs, with each approach presenting opportunities and challenges. BiTE114, or using non-traditional immune cells that may mediate more
Co-infusion is financially, labour- and cell-intensive and co-infusion and potent endogenous antitumour activity.
co-transduction generate heterogeneous products, risking the emer-
gence of a subpopulation that dominates the pool of cells after infu- Enhancing T cell potency
sion101,102. Bicistronic vectors may result in reduced protein expression, Extensive work is underway to enhance immune cell fitness (Fig. 3 and
and in one clinical trial, a bicistronic construct demonstrated limited Box 1). Significant effort is focused on epigenetic modulation, in part on
persistence103. Several trials with bispecific receptors targeting CD19 the basis of an exceptional responder in a clinical trial of CD19-CAR for
plus CD20 or CD22 have been reported45,104,105, and in one, the receptor CLL—in which lentiviral integration disrupted the TET2 gene, a mediator

640 | Nature | Vol 614 | 23 February 2023

of DNA methylation, resulting in substantial clonal T cell proliferation Kill switches such as iCasp9151, HSV tyrosine kinase152 (HSV-tk) and
and a sustained antitumour response115. Similarly, knockout of the epitope tags153 enable the depletion of engineered cells in the event of
DNMT3A gene enhances the antitumour activity of CAR T cells in pre- severe toxicity, and a transgene-free safety switch that renders T cells
clinical models116. Overexpression of transcription factors to prevent auxotrophic for uridine has been developed154. Regulatable platforms
exhaustion has also shown promise, including overexpression of the can serve as reversible safety switches and also tune CAR signalling,
AP-1 factor JUN, which enhances T cell expansion and persistence, thereby enhancing T cell potency by providing rest periods that prevent
diminishes terminal differentiation and lowers the antigen density T cell exhaustion121. Numerous regulatable platforms have been devel-
threshold, presumably owing to increased signal strength53. Similarly, oped using drug-sensitive promoters155, induced dimerization156,157,
overexpression of BATF transcription factors has been reported to disruption of split CARs158, drug-dependent activation of binders159,
enhance T cell potency117. Manufacturing strategies are being devel- proteolysis-targeting chimeras160 (PROTACs), chemically-dependent
oped to optimize CAR T cell phenotype towards stem-like and central degron domains121,157,161 and drug-regulated CAR proteolysis162,163.
memory subsets, including shorter culture duration118, inhibition of These systems represent significant advances in synthetic biology,
PI3K–mTOR–AKT119, BTK120 or tyrosine kinase121 signalling, and culture but remain challenged by leaky activity in the OFF state that risks tox-
in memory-promoting cytokines122. icity, diminished CAR expression or activity in the ON state and the
CRISPR-mediated gene editing was first applied clinically in the set- use immunosuppressive drugs as regulators121,155–161. Labanieh et al.
ting of adoptive T cell therapy, in which PD-1 was deleted from cells recently developed a protease-regulated grazoprevir-induced ‘drug ON’
engineered to express NY-ESO-1, a cancer-specific TCR transgene123. platform, signal neutralization by an inhibitable protease (SNIP), which
The engineered cells did not demonstrate enhanced persistence or shows no leaky activity and full functional capacity162 (Fig. 3). Similar
potency, but the study demonstrated the feasibility and safety of the to synNotch164, SNIP demonstrates superior antitumour efficacy com-
approach, and accelerated efforts to apply gene editing technologies pared with conventional CAR T cells owing to reduced exhaustion, and
to enhance immune cell therapies. Several genes have been identified in an on-target off-tumour ROR1 toxicity model, decreased grazoprevir
as candidates for editing to enhance T cell fitness72,124–133 (Box 1), and dosing tuned SNIP CARs to open a therapeutic window in which healthy
CRISPR-mediated disruption of T cell markers such as CD7 and CD5 has tissue was spared but ROR1-expressing tumour cells were killed162.
enabled CAR T cell therapies for T cell malignancies, while avoiding CAR Similarly, Hernandez-Lopez et al. iterated the synNotch platform to
T cell lysis134,135 (termed ‘fratricide’). We anticipate increasing clinical target very highly expressed tumour antigens while avoiding lower
trial activity incorporating gene-edited immune cells into adoptive levels of the antigens on normal tissues165. Thus, regulatable CARs show
immune cell therapy platforms to enhance their potency, expand the promise for enhancing efficacy and diminishing toxicity.
landscape of targetable antigens and avoid immune sensitization.
To enhance persistence, some investigators have sought to inte- Enhancing specificity through Boolean logic
grate cytokine signals into the CAR receptor or express cytokines in B cell and plasma cell malignancies are especially suited to CAR T cell
trans136,137, including a clinical trial in which CAR-expressing natural therapy owing to the high, homogenous expression of lineage antigens
killer cells transgenically expressing IL-15 demonstrated prolonged that are co-expressed predominantly on B cells and plasma cells, the
persistence138. Immune rejection may also limit CAR T cell persistence depletion of which is tolerable. However, a recent case report showed
as anti-CAR immune responses—often targeting mouse, humanized or the development of parkinsonism in a patient after BCMA-CAR T cell
fully human scFvs—can be measured in many patients29,97,139,140. Consist- therapy, with postmortem analysis revealing expression of BCMA on
ently, clinical experience demonstrates the limited utility of second subsets of neurons and astrocytes in the patient’s basal ganglia166. In
and subsequent intravenous CAR T cell doses, which can be improved another study, single-cell RNA-sequencing analysis showed the expres-
using enhanced lymphodepleting regimens96. These findings raise the sion of CD19 on brain mural cells, raising the prospect that on-target
prospect that stealth platforms—which are currently being developed killing may be responsible for neurotoxicity after CD19-CAR T cell
to enable off-the-shelf allogeneic products (discussed in ‘Platforms to therapy. These results highlight the challenge of identifying targets
enhance access and efficacy’)—could enhance CAR T cell efficacy by that are not expressed on vital tissue.
enhancing persistence or enabling multiple CAR T dosing regimens. So far, the paucity of tumour-specific surface targets on solid tumours
Diverse efforts are underway to address the suppressive TME (Box 1), has limited the application of the CAR prototype to solid tumours, with
including genetic ablation or expression of dominant-negative TBGβ141,142, unacceptable off-tumour, on-target toxicity having been observed in
PD-1143,144 or Fas receptors145, and engineering CAR T cells to secrete trials of CARs targeting CAIX167 and CEACAM5168. However, several clini-
checkpoint-blocking scFvs146. Some investigators have engineered switch cal trials of CAR T cells and other potent antibody-directed therapies
receptors, fusion proteins that convert a suppressive signal within the have demonstrated good safety profiles in solid tumours (Table 1). The
TME to an activating signal in the CAR T cells147,148. Whether tonic acti- high CAR antigen density threshold is likely to explain the safe target-
vating signals induced by such receptors result in long-term CAR T cell ing of some antigens with known expression on vital tissues—such as
enhancement or predispose them to exhaustion and terminal differen- GD2, which is expressed at low levels on neural tissues169,170. A recent
tiation remains to be determined. Biomaterials-based approaches for trial demonstrated promising clinical activity of claudin-18.2-CARs was
enhancing the expansion and persistence are also being explored149. associated with significant but non-dose-limiting toxicity, potentially
explained by antigen expression restricted to differentiated epithe-
CAR tuning and regulatable platforms lial cells buried in gastric mucosa that may be less accessible to CAR
Substantial efforts are underway to enhance safety and potency by tun- T cells29. Identifying additional molecules with sufficient differential
ing or dampening CAR signalling to diminish toxicity and exhaustion. expression levels for safe targeting, such as oncofetal cell-surface tar-
This concept was first proposed by Eyquem and colleagues, who used gets is essential for expanding the reach of CAR T cells beyond B cell and
CRISPR to knock-in CAR receptors into the TRAC locus and observed plasma cell malignancies. However, the safety of specific targets will
improved potency and diminished exhaustion due to antigen-induced need to be continually reassessed as potency and persistence enhance-
CAR downregulation mediated by endogenous TRAC regulatory ments are deployed, as in recent studies with a PSMA-targeted CAR
elements150. Weber and colleagues extended this principle using syn- integrating a dominant-negative TGFβ receptor that was associated
thetic biology or small molecules to transiently cease CAR signalling, with lethal toxicity171,172.
which enhanced CAR T cell potency when used during manufacturing Next-generation receptors incorporating logic gates could allow
and improved antitumour effects when applied in vivo after adoptive better discrimination between tumour and healthy tissue through
transfer121. combinatorial antigen sensing, and expand the repertoire of potential

Nature | Vol 614 | 23 February 2023 | 641

Perspective
antigens (Fig. 3). Roybal et al. developed synNotch, an IF–THEN circuit current interest, especially for therapies targeting rare indications,
incorporating a synthetic notch receptor against antigen A, which such as paediatric cancers189.
upon engagement, triggers the transcription of a conventional CAR Allogeneic CAR T cells manufactured from healthy ‘super donors’
against antigen B173,174. The synNotch system has not been tested clini- could improve potency by avoiding preexisting T cell dysfunction and
cally, but in preclinical models it prevented on-target, off-tumour decrease the cost and logistical challenges of manufacturing, thereby
toxicity when tumours and susceptible vital tissues are not colocal- enhancing access. However, allogeneic T cell therapies must overcome
ized175. Tousley et al. developed an AND gate platform called LINK, the risk of GVHD mediated by the TCR and rejection of the transferred
which utilizes the proximal TCR signalling molecules LAT and SLP76, cells by the host immune system. Gene editing of the endogenous TCR
each fused to a membrane-bound scFv specific for a unique antigen176. eliminates the risk of GVHD190, but endowing stealth properties to avoid
Engagement of both antigens colocalizes LAT and SLP76, leading to immune rejection remains a significant challenge, since CD8+, CD4+,
T cell activation. In an on-target, off-tumour ROR1 toxicity model, natural killer and macrophage cells can reject allogeneic cells and each
LINK CAR T cells cured mice of tumours expressing both antigens are regulated by distinct axes, necessitating multiple enhancements
without ROR1-mediated toxicity, whereas mice treated with synNotch (Box 1). Knockout of β2-microglobulin can eliminate HLA class I surface
T cells succumbed to toxicity175. Other approaches for combinatorial expression, but paradoxically increases the risk of rejection by natural
antigen targeting that are under development include SUPRA177 and killer cells. Additional strategies for inducing allogeneic tolerance
co-LOCKR178, which redirect CAR T cell specificity through protein include knockout of the CIITA gene to ablate MHC class II expression191,
switches. Although combinatorial antigen sensing could expand and overexpression of HLA-E192 and CD47193 to ameliorate natural killer
the landscape of targetable tumour antigens, the increased risk of cell- and macrophage-mediated cell rejection.
tumour escape owing to loss of either antigen is a potential concern. Many allogeneic approaches use CRISPR–Cas9, and the risks of
An alternative approach to enhance specificity is to use an AND NOT CRISPR-based mutagenic events could be magnified when producing
gate, in which a prototype activating CAR is expressed in trans with hundreds or thousands of allogeneic products with a singular manu-
an inhibitory CAR (iCAR) targeting an antigen that is expressed on facturing process. Alternative platforms, such as base editing or prime
healthy tissue but not on tumour tissue179–181. Limited engineering editing may emerge as preferred alternatives to nuclease-based genome
with NOT gates has been undertaken so far and these applications editing since they probably involve lower risk owing to an absence of
have not been tested clinically. double strand DNA breaks194. CRISPR–Cas systems targeting RNA could
also provide opportunities for multiplexed gene knockdowns with
TCR-like CARs greater specificity and efficiency compared with RNA-mediated inter-
With the goal of targeting antigens that are expressed at low levels, ference. Although allogeneic donor-derived cells containing multiple
HLA-independent TCRs (HIT) are designed with the variable domain gene edits could provide significant advantages, these technologies
of the endogenous TCR being altered to target scFvs by gene editing are nascent and their toxicity profiles remain unknown. Some groups
the endogenous TRAC locus182. When CD80 and 4-1BBL are provided have attempted to prevent immune rejection by augmenting immune
in trans, CD19-directed HITs display superior antigen sensitivity com- suppression of the host using conventional chemotherapy or immu-
pared with prototype CD19-CARs (Fig. 3). Synthetic TCR and antigen nosuppressive antibodies for which the targets are edited from the
receptors (STARs) have a similar design but are not knocked in to the CAR T cells190. Early response rates with this approach are promising,
TRAC locus; thus, the endogenous TCR specificity is retained183. Other but long-term safety and efficacy have not been demonstrated and
approaches for redirecting TCR specificity include the antibody–TCR concerns remain regarding infectious risks associated with intensive
(AbTCR) platform184, which replaces the variable domains of TCRγδ immune-depleting regimens190.
with a Fab fragment and TCR fusion constructs185 (TRuC), which fuse
an scFv to a CD3 subunit. A recent comparison of TCR-like chimeric Alternative immune cells
receptors showed that STAR and HIT receptors reproduce TCR antigen Several non-T immune cells, including natural killer cells, invariant
sensitivity, whereas TruCs do not186. One potential drawback of CAR natural killer T (iNKT) cells, γδ T cells and macrophages exhibit innate
T cells compared with native T cells is the inability to target intracellular antitumour activity and do not induce GVHD, raising the prospect that
antigens, since most aberrant proteins that drive cancer are intracel- they could provide an off-the-shelf source of cells with reduced toxic-
lular. Yarmarkovich et al. overcame this by developing a prototype ity, enhanced tumour trafficking and/or target antigen-negative vari-
CAR with specificity for peptides presented by MHC187 (pMHC). Using ants through innate tumour recognition. However, allogeneic innate
scFv binders specific for a PHOX2B peptide–MHC overexpressed in immune cells remain susceptible to rejection, raising concerns regard-
neuroblastoma, they targeted pMHCs across several HLA allotypes. This ing the durability of their effects if they are not engineered for stealth.
strategy could greatly expand the landscape of CAR targets, including Cord blood-derived allogeneic natural killer cells incorporating ectopi-
key oncogenic drivers. cally expressed IL-15 have shown promise in a phase I trial for NHL and
CLL138. iNKT-CAR cells mediated activity in mouse models, in part by
cross-priming host CD8 cells towards tumour antigens195, and their
Platforms to enhance access and efficacy safety and feasibility in a phase I trial for neuroblastoma have been
Diverse approaches are under exploration to increase access of cell demonstrated196. γδ T cells engineered to express a CD20-CAR have
therapies, diminish the high manufacturing costs, create stealth also shown impressive activity in early studies197. Expressing CARs
immune cells resistant to rejection, and leverage the unique proper- in macrophages requires substantial adaptations of vectors and sig-
ties of alternative immune cells. nalling domains, but antitumour effects associated with augmented
phagocytosis, modification of the TME and recruitment of T cells198,199
Distributed manufacturing and allogeneic products have been demonstrated in preclinical models198, with CD3ζ-based
Engineering advances have yielded automated closed-system manufac- CARs demonstrating equivalent phagocytic activity as Fcγ-based CARs
turing, which is providing opportunities for point-of-care manufactur- Efforts are also underway to create induced pluripotent stem (iPS)
ing to diminish the costs, delays and logistical challenges associated cell-derived CAR T cells200, natural killer cells201 and macrophages202.
with the centralized manufacturing models that are the industry stand- The differentiation of iPS cells to natural killer cells has been particularly
ard. A recent multicentre trial demonstrated the safety and efficacy successful, and clinical testing of these off-the-shelf therapies is cur-
of cells manufactured at the point of care188. Defining the regulatory rently in progress203, whereas iPS cell differentiation to fully functional
requirements for point-of-care manufacturing is an area of significant T cells has been more challenging204. Given their nearly inexhaustible

642 | Nature | Vol 614 | 23 February 2023

expansion potential, iPS cell-derived products could enable mass pro- homeostasis in models of liver fibrosis220. CARs targeting fibroblast
duction of a homogenous cell product integrating numerous enhance- activation protein (FAP) have improved cardiac function in a mouse
ments to endow stealth properties, safety switches and potency, and model of cardiac fibrosis221 and in vivo generation of FAP-CARs using
the long-term safety and efficacy results of these emerging platforms CD5-directed lipid nanoparticles loaded with mRNA also demon-
are thus eagerly anticipated. strated benefit211. In this model, the non-integrating nature of mRNA
ensured that CAR expression was transient, thereby mitigating the
Next-generation gene delivery risk of toxicity associated with widespread elimination of activated
Viral vector-based gene delivery has been the gold standard in the field, fibroblasts.
but vector production and qualification is costly and time consuming.
Viral-free platforms for gene delivery are under development, with
CRISPR-based gene delivery in human T cells demonstrating proof Outlook
of principle, although DNA templates are toxic to T cells and the effi- Adoptive immune cell therapy is established as a transformative thera-
ciency of this approach remains lower than with viral vectors205. Clinical peutic modality. The past decade has witnessed significant progress in
feasibility has been demonstrated with a CD19-CAR site-specifically understanding the biology of prototype CAR T cells, identifying antigen
delivered to the PD-1 locus inducing a high CR rate in NHL, although the modulation and T cell dysfunction as major resistance mechanisms
manufacturing process did not meet dose requirements for a relatively and highlighting the logistical challenges of delivering cell therapies
high proportion of patients206. Modifications to DNA templates and to all patients who could benefit. Modifications to prototype CARs can
small-molecule inhibitor cocktails are improving knock-in efficiencies augment their potency, but increasingly investigators are designing
and cell yields207. Transposon-based gene delivery has also been used, next-generation platforms to create advanced cellular therapies that
although malignant transformation of CAR-engineered T cells was incorporate a diverse array of enhancements. The fields of immunol-
reported in two patients associated with high-copy-number integra- ogy, synthetic biology, genetic engineering and cell manufacturing
tion using a Piggybac transposon platform208. In vivo gene delivery are synergizing to create smarter, safer and more accessible cellular
is another emerging approach that could improve accessibility and therapies that are poised for increased efficacy and access, diminished
diminish cost. In this approach, DNA or RNA is delivered systemically risk and cost, and broader utility, for the treatment of cancer as well as
using viral vectors209 or nanoparticles210 that preferentially target non-malignant diseases.
and transduce immune populations in vivo. Immunogenicity could
prohibit repeat administration of viral vectors owing to the induc- 1. Abramson, J. S. et al. Lisocabtagene maraleucel for patients with relapsed or refractory
large B-cell lymphomas (TRANSCEND NHL 001): a multicentre seamless design study.
tion of neutralizing antibodies. Stable expression of a CD19-CAR has
Lancet 396, 839–852 (2020).
been demonstrated using lipid nanoparticles targeting CD3 in mice210 2. Nastoupil, L. J. et al. Standard-of-care axicabtagene ciloleucel for relapsed or refractory
and T cell-targeted lipid nanoparticles incorporating optimized RNA large B-cell lymphoma: results from the US lymphoma CAR T consortium. J. Clin. Oncol.
38, 3119–3128 (2020).
diminished cardiac fibrosis in a mouse model211.
3. Neelapu, S. S. et al. Axicabtagene ciloleucel CAR T-cell therapy in refractory large B-cell
lymphoma. N. Engl. J. Med. 377, 2531–2544 (2017).
4. Schuster, S. J. et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-cell
CAR therapy for non-malignant diseases lymphoma. N. Engl. J. Med. 380, 45–56 (2019).
5. Maude, S. L., Laetsch, T. W., Buchner, J. & Grupp, S. A. Tisagenlecleucel in children and
The CAR T platform has been optimized for cancer treatment, but young adults with B-cell lymphoblastic leukemia. N. Engl. J. Med. 378, 439–448 (2018).
the design principles and expansive synthetic biology toolbox used 6. Pasquini, M. C. et al. Real-world evidence of tisagenlecleucel for pediatric acute
lymphoblastic leukemia and non-Hodgkin lymphoma. Blood Adv. 4, 5414–5424 (2020).
for CAR T cells are providing opportunities to extend this therapeu-
7. Schultz, L. M. et al. Disease burden affects outcomes in pediatric and young adult B-cell
tic approach to non-malignant diseases, including autoimmunity, lymphoblastic leukemia after commercial tisagenlecleucel: a pediatric real-world
senescence, fibrosis and infectious diseases. In preclinical studies, chimeric antigen receptor consortium report. J. Clin. Oncol. 40, 945–955 (2022).
8. Wang, M. et al. KTE-X19 CAR T-cell therapy in relapsed or refractory mantle-cell lymphoma.
CD19-CAR T cells have demonstrated beneficial effects in systemic N. Engl. J. Med. 382, 1331–1342 (2020).
lupus erythematosus212, and a case study reported sustained activity 9. Fowler, N. H. et al. Tisagenlecleucel in adult relapsed or refractory follicular lymphoma:
of CD19-CAR therapy in a patient with refractory lupus nephritis213 . the phase 2 ELARA trial. Nat. Med. 28, 325–332 (2022).
10. Locke, F. L. et al. Axicabtagene ciloleucel as second-line therapy for large B-cell lymphoma.
Chimeric autoantibody receptors (CAARs) are prototype CARs that N. Engl. J. Med. 386, 640–654 (2021).
incorporate a scFv targeting the idiotype of an autoreactive B cell clone This pivotal trial demonstrated CD19-CAR therapy outperforming autologous stem-cell
or use autoantigens as the recognition domain. In preclinical studies, transplant for second line therapy of LBCL.
11. Kamdar, M. et al. Lisocabtagene maraleucel versus standard of care with salvage
CAARs mediated therapeutic effects against pemphigus vulgaris, and chemotherapy followed by autologous stem cell transplantation as second-line
clinical testing is underway. Adoptive transfer of T regulatory (Treg) treatment in patients with relapsed or refractory large B-cell lymphoma (TRANSFORM):
cells, which mediate suppression rather than cytotoxicity, is an alter- results from an interim analysi. Lancet 399, 2294–2308 (2022).
12. Neelapu, S. S. et al. Axicabtagene ciloleucel as first-line therapy in high-risk large B-cell
native approach for treating autoimmunity. Non-engineered Treg cells lymphoma: the phase 2 ZUMA-12 trial. Nat. Med. 28, 735–742 (2022).
have demonstrated activity in mouse models of GVHD, allograft trans- This study demonstrated that axi-cel is highly effective in the first-line therapy setting
plantation, type 1 diabetes, systemic lupus erythematosus and multiple for high-risk LBCL.
13. Munshi, N. C. et al. Idecabtagene vicleucel in relapsed and refractory multiple myeloma.
sclerosis, and early clinical data demonstrate feasibility of manufactur- N. Engl. J. Med. 384, 705–716 (2021).
ing and a good safety profile214. Compared with non-engineered cells, 14. Berdeja, J. G. et al. Ciltacabtagene autoleucel, a B-cell maturation antigen-directed
Treg cells expressing a CAR targeting antigens expressed on the diseased chimeric antigen receptor T-cell therapy in patients with relapsed or refractory multiple
myeloma (CARTITUDE-1): a phase 1b/2 open-label study. Lancet 398, 314–324 (2021).
tissues show enhanced specificity and potency215,216. Recent data have This study showed impressive clinical activity of the BCMA-directed therapy
demonstrated that inadvertent expansion of CAR Treg cells limits the ciltacabtagene autoleucel.
efficacy of commercial CAR T cells, providing proof-of-concept for 15. Fry, T. J. et al. CD22-targeted CAR T cells induce remission in B-ALL that is naive or resistant
to CD19-targeted CAR immunotherapy. Nat. Med. 24, 20–28 (2018).
the utility of CAR-engineered Treg cells217,218. Approaches are underway 16. Shah, N. N. et al. CD4/CD8 T-cell selection affects chimeric antigen receptor (CAR) T-cell
to engineer FOXP3 expression to enforce lineage stability and incor- potency and toxicity: updated results from a phase I anti-CD22 CAR T-cell trial. J. Clin. Oncol.
porate safety switches to diminish risk214. Recent promising preclinical 38, 1938–1950 (2020).
17. Baird, J. H. et al. CD22-directed CAR T-cell therapy induces complete remissions in CD19-
data were generated in haemophilic mice, in which Treg cells expressing directed CAR–refractory large B-cell lymphoma. Blood 137, 2321–2325 (2021).
a factor VIII-targeted CAR and FOXP3 prevented the development 18. Ramos, C. A. et al. Anti-CD30 CAR-T cell therapy in relapsed and refractory Hodgkin
of neutralizing anti-factor VIII antibodies219. Senolytic CAR T cells lymphoma. J. Clin. Oncol. 38, 3794–3804 (2020).
19. Zhang, M. et al. Autologous nanobody-derived fratricide-resistant CD7-CAR T-cell therapy
targeting urokinase-type plasminogen activator receptor have been for patients with relapsed and refractory T-cell acute lymphoblastic leukemia/lymphoma.
demonstrated to target senescent cells in vitro and restore tissue Clin. Cancer Res. 28, 2830–2843 (2022).

Nature | Vol 614 | 23 February 2023 | 643

Perspective
20. Pan, J. et al. Donor-derived CD7 chimeric antigen receptor T cells for T-cell acute 53. Lynn, R. C. et al. c-Jun overexpression in CAR T cells induces exhaustion resistance. Nature
lymphoblastic leukemia: first-in-human, phase I trial. J. Clin. Oncol. 39, 3340–3351 (2021). 576, 293–300 (2019).
21. Lu, P. et al. Naturally selected CD7 CAR-T therapy without genetic manipulations for This study showed that overexpression of c-Jun improves the fitness of CAR-T cells.
T-ALL/LBL: first-in-human phase 1 clinical trial. Blood 140, 321–334 (2022). 54. Hegde, M. et al. Tumor response and endogenous immune reactivity after administration
22. Yang, J. et al. High effectiveness and safety of anti-CD7 CAR T-cell therapy in treating of HER2 CAR T cells in a child with metastatic rhabdomyosarcoma. Nat. Commun. 11,
relapsed or refractory (R/R) T-cell acute lymphoblastic leukemia (T-ALL). Blood 138, 473 3549 (2020).
(2021). 55. Krenciute, G. et al. Transgenic expression of IL15 improves antiglioma activity of IL13Rα2-
23. Zhang, W. Y. et al. Long-term safety and efficacy of CART-20 cells in patients with refractory CAR T cells but results in antigen loss variants. Cancer Immunol. Res. 5, 571–581 (2017).
or relapsed B-cell non-Hodgkin lymphoma: 5-years follow-up results of the phase I and IIa 56. O’Rourke, D. M. et al. A single dose of peripherally infused EGFRvIII-directed CAR T cells
trials. Signal Transduct. Target. Ther. 2, 17054 (2017). mediates antigen loss and induces adaptive resistance in patients with recurrent
24. Mailankody, S. et al. Phase I first-in-class trial of MCARH109, a G protein coupled receptor glioblastoma. Sci. Transl. Med. 9, eaaa0984 (2017).
class C group 5 member D (GPRC5D) targeted CAR T cell therapy in patients with relapsed 57. Gardner, R. A. et al. Intent-to-treat leukemia remission by CD19 CAR T cells of defined
or refractory multiple myeloma. Blood 138, 827 (2021). formulation and dose in children and young adults. Blood 129, 3322–3331 (2017).
25. D’angelo, S. P. et al. Antitumor activity associated with prolonged persistence of adoptively 58. Lee, D. W. et al. T cells expressing CD19 chimeric antigen receptors for acute lymphoblastic
transferred NY-ESO-1c259T cells in synovial sarcoma. Cancer Discov. 8, 944–957 (2018). leukaemia in children and young adults: a phase 1 dose-escalation trial. Lancet 385, 517–528
26. Robbins, P. F. et al. Tumor regression in patients with metastatic synovial cell sarcoma and (2015).
melanoma using genetically engineered lymphocytes reactive with NY-ESO-1. J. Clin. 59. Plaks, V. et al. CD19 target evasion as a mechanism of relapse in large B-cell lymphoma
Oncol. 29, 917–924 (2011). treated with axicabtagene ciloleucel. Blood 138, 1081–1085 (2021).
27. Brown, C. E. et al. Regression of glioblastoma after chimeric antigen receptor T-cell therapy. 60. Raje, N. et al. Anti-BCMA CAR T-cell therapy bb2121 in relapsed or refractory multiple
N. Engl. J. Med. 375, 2561–2569 (2016). myeloma. N. Engl. J. Med. 380, 1726–1737 (2019).
This case report showed that IL-13Rα2-targeting CAR-T cells administered via 61. Cohen, A. D. et al. B cell maturation antigen–specific CAR T cells are clinically active in
intracavitary and intraventricular delivery routes were safe and effective in a patient multiple myeloma. J. Clin. Invest. 129, 2210–2221 (2019).
with glioblastoma. 62. Brudno, J. N. et al. T cells genetically modified to express an anti-B-cell maturation antigen
28. Majzner, R. G. et al. GD2-CAR T cell therapy for H3K27M-mutated diffuse midline gliomas. chimeric antigen receptor cause remissions of poor-prognosis relapsed multiple myeloma.
Nature 603, 934–941 (2022). J. Clin. Oncol. 36, 2267–2280 (2018).
This study demonstrated clinical activity of GD2 CAR-T cells in diffuse midline gliomas. 63. Da Vià, M. C. et al. Homozygous BCMA gene deletion in response to anti-BCMA CAR
29. Qi, C. et al. Claudin18.2-specific CAR T cells in gastrointestinal cancers: phase 1 trial T cells in a patient with multiple myeloma. Nat. Med. 27, 616–619 (2021).
interim results. Nat. Med. 28, 1189–1198 (2022). 64. Samur, M. K. et al. Biallelic loss of BCMA as a resistance mechanism to CAR T cell therapy
30. Morris, E. C., Neelapu, S. S., Giavridis, T. & Sadelain, M. Cytokine release syndrome and in a patient with multiple myeloma. Nat. Commun. 12, 868 (2021).
associated neurotoxicity in cancer immunotherapy. Nat. Rev. Immunol. 22, 85–96 (2022). 65. Sotillo, E. et al. Convergence of acquired mutations and alternative splicing of CD19
31. Majzner, R. G. & Mackall, C. L. Clinical lessons learned from the first leg of the CAR T cell enables resistance to CART-19 immunotherapy. Cancer Discov. 5, 1282–1295 (2015).
journey. Nat. Med. 25, 1341–1355 (2019). 66. Orlando, E. J. et al. Genetic mechanisms of target antigen loss in CAR19 therapy of acute
32. Gross, G., Waks, T. & Eshhar, Z. Expression of immunoglobulin-T-cell receptor chimeric lymphoblastic leukemia. Nat. Med. 24, 1504–1506 (2018).
molecules as functional receptors with antibody-type specificity. Proc. Natl Acad. Sci. 67. Jacoby, E. et al. CD19 CAR immune pressure induces B-precursor acute lymphoblastic
USA 86, 10024–10028 (1989). leukaemia lineage switch exposing inherent leukaemic plasticity. Nat. Commun. 7, 12320
33. Eshhar, Z., Waks, T., Gross, G. & Schindler, D. G. Specific activation and targeting of (2016).
cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding 68. Heard, A. et al. Antigen glycosylation regulates efficacy of CAR T cells targeting CD19.
domains and the γ or ζ subunits of the immunoglobulin and T-cell receptors. Proc. Natl Nat. Commun. 13, 3367 (2022).
Acad. Sci. USA 90, 720–724 (1993). 69. Pont, M. J. et al. γ-Secretase inhibition increases efficacy of BCMA-specific chimeric antigen
34. Maher, J., Brentjens, R. J., Gunset, G., Rivière, I. & Sadelain, M. Human T-lymphocyte receptor T cells in multiple myeloma. Blood 134, 1585–1597 (2019).
cytotoxicity and proliferation directed by a single chimeric TCRζ/CD28 receptor. Nat. 70. Ramakrishna, S. et al. Modulation of target antigen density improves CAR T-cell functionality
Biotechnol. 20, 70–75 (2002). and persistence. Clin. Cancer Res. 25, 5329–5341 (2019).
35. Imai, C. et al. Chimeric receptors with 4-1BB signaling capacity provoke potent cytotoxicity 71. Leick, M. B. et al. Non-cleavable hinge enhances avidity and expansion of CAR-T cells for
against acute lymphoblastic leukemia. Leukemia 18, 676–684 (2004). acute myeloid leukemia. Cancer Cell 40, 494–508.e5 (2022).
36. Harris, D. T. et al. Comparison of T cell activities mediated by human TCRs and CARs that 72. Good, C. R. et al. An NK-like CAR Tcell transition in CAR Tcell dysfunction. Cell 184,
use the same recognition domains. J. Immunol. 200, 1088–1100 (2018). 6081–6100.e26 (2021).
37. Anikeeva, N. et al. Efficient killing of tumor cells by CAR-T cells requires greater number This study used continuous antigen exposure to identify gene signatures associated
of engaged CARs than TCRs. J. Biol. Chem. 297, 101033 (2021). with CAR dysfunction.
38. Gudipati, V. et al. Inefficient CAR-proximal signaling blunts antigen sensitivity. Nat. Immunol. 73. Finney, O. C. et al. CD19 CAR T cell product and disease attributes predict leukemia
21, 848–856 (2020). remission durability. J. Clin. Invest. 129, 2123–2132 (2019).
39. Sykulev, Y., Joo, M., Vturina, I., Tsomides, T. J. & Eisen, H. N. Evidence that a single peptide– 74. Fraietta, J. A. et al. Determinants of response and resistance to CD19 chimeric antigen receptor
MHC complex on a target cell can elicit a cytolytic T cell response. Immunity 4, 565–571 (CAR) T cell therapy of chronic lymphocytic leukemia. Nat. Med. 24, 563–571 (2018).
(1996). This study identified features of CAR-T immunobiology by comparing responding and
40. Irvine, D. J., Purbhoo, M. A., Krogsgaard, M. & Davis, M. M. Direct observation of ligand non-responding patients with CLL treated with CD19-CAR T cells.
recognition by T cells. Nature 419, 845–849 (2002). 75. Salter, A. I. et al. Phosphoproteomic analysis of chimeric antigen receptor signaling
41. Watanabe, K. et al. Target antigen density governs the efficacy of anti-CD20–CD28–CD3 ζ reveals kinetic and quantitative differences that affect cell function. Sci. Signal 11,
chimeric antigen receptor-modified effector CD8+ T cells. J. Immunol. 194, 911–920 eaat6753 (2018).
(2015). 76. Melenhorst, J. J. et al. Decade-long leukaemia remissions with persistence of CD4+ CAR
42. Walker, A. J. et al. Tumor antigen and receptor densities regulate efficacy of a chimeric T cells. Nature 602, 503–509 (2022).
antigen receptor targeting anaplastic lymphoma kinase. Mol. Ther. 25, 2189–2201 (2017). This study identified a subset of CD4+ CD19 CAR-T cells that exhibited extended
43. Majzner, R. G. et al. CAR T cells targeting B7-H3, a pan-cancer antigen, demonstrate persistence in CLL patients.
potent preclinical activity against pediatric solid tumors and brain tumors. Clin. Cancer 77. Scholler, J. et al. Decade-long safety and function of retroviral-modified chimeric antigen
Res. 25, 2560–2574 (2019). receptor T cells. Sci. Transl. Med. 4, 132ra53 (2012).
44. Majzner, R. G. et al. Tuning the antigen density requirement for car T-cell activity. Cancer 78. Singh, N. et al. Antigen-independent activation enhances the efficacy of 4-1BB-costimulated
Discov. 10, 702–723 (2020). CD22 CAR T cells. Nat. Med. 27, 842–850 (2021).
This study demonstrated that modifications to the CAR prototype structure modulate 79. Feucht, J. et al. Calibration of CAR activation potential directs alternative T cell fates and
antigen density threshold. therapeutic potency. Nat. Med. 25, 82–88 (2019).
45. Spiegel, J. Y. et al. CAR T cells with dual targeting of CD19 and CD22 in adult patients with 80. Guedan, S. et al. Single residue in CD28-costimulated CAR-T cells limits long-term
recurrent or refractory B cell malignancies: a phase 1 trial. Nat. Med. 27, 1419–1431 (2021). persistence and antitumor durability. J. Clin. Invest. 130, 3087–3097 (2020).
46. Dong, R. et al. Rewired signaling network in T cells expressing the chimeric antigen 81. Li, W. et al. Chimeric antigen receptor designed to prevent ubiquitination and
receptor (CAR). EMBO J. 39, e104730 (2020). downregulation showed durable antitumor efficacy. Immunity 53, 456–470.e6 (2020).
47. Salter, A. I. et al. Comparative analysis of TCR and CAR signaling informs CAR designs 82. Yang, Y. et al. TCR engagement negatively affects CD8 but not CD4 CAR T cell expansion
with superior antigen sensitivity and in vivo function. Sci. Signal. 14, eabe2606 (2021). and leukemic clearance. Sci. Transl. Med. 9, eaag1209 (2017).
This study demonstrated diminished recruitment of proximal signaling molecules in 83. Frank, M. J. et al. Monitoring of circulating tumor DNA improves early relapse detection
CARs compared to TCRs. after axicabtagene ciloleucel infusion in large B-cell lymphoma: results of a prospective
48. Long, A. H. et al. 4-1BB costimulation ameliorates T cell exhaustion induced by tonic multi-institutional trial. J. Clin. Oncol. 39, 3034–3043 (2021).
signaling of chimeric antigen receptors. Nat. Med. 21, 581–90 (2015). 84. Shah, N. N. et al. Long-term follow-up of CD19-CAR T-cell therapy in children and young
49. Heitzeneder, S. et al. GPC2-CAR Tcells tuned for low antigen density mediate potent adults with B-ALL. J. Clin. Oncol. 39, 1650–1659 (2021).
activity against neuroblastoma without toxicity. Cancer Cell 40, 53–69.e9 (2022). 85. Lange, S. et al. A chimeric GM-CSF/IL18 receptor to sustain CAR T-cell function. Cancer
50. Hamieh, M. et al. CAR T cell trogocytosis and cooperative killing regulate tumour antigen Discov. 11, 1661–1671 (2021).
escape. Nature 568, 112–116 (2019). 86. Mata, M. et al. Inducible activation of myD88 and CD40 in CAR T cells results in controllable
This study showed that CAR-T trogocytosis could lower the antigen density of target and potent antitumor activity in preclinical solid tumor models. Cancer Discov. 7, 1306–1319
tumor cells and provide an alternative mechanism of resistance to therapy. (2017).
51. James, S. E. et al. Antigen sensitivity of CD22-specific chimeric TCR is modulated by 87. Daniels, K. G. et al. Decoding CAR T cell phenotype using combinatorial signaling motif
target epitope distance from the cell membrane. J. Immunol. 180, 7028–7038 (2008). libraries and machine learning. Science 378, eabq0225 (2022).
52. Haso, W. et al. Anti-CD22-chimeric antigen receptors targeting B-cell precursor acute 88. Goodman, D. B. et al. Pooled screening of CAR T cells identifies diverse immune signaling
lymphoblastic leukemia. Blood 121, 1165–1171 (2013). domains for next-generation immunotherapies. Sci. Transl. Med. 14, eabm1463 (2022).

644 | Nature | Vol 614 | 23 February 2023

89. Gordon, K. S. et al. Screening for CD19-specific chimaeric antigen receptors with 123. Stadtmauer, E. A. et al. CRISPR-engineered T cells in patients with refractory cancer.
enhanced signalling via a barcoded library of intracellular domains. Nat. Biomed. Eng. 6, Science 367, eaba7365 (2020).
855–866 (2022). 124. Wang, D. et al. Crispr screening of CAR T cells and cancer stem cells reveals critical
90. Yan, X. et al. CD58 loss in tumor cells confers functional impairment of CAR T cells. Blood dependencies for cell-based therapies. Cancer Discov. 11, 1192–1211 (2021).
Adv. 6, 5844–5856 (2022). 125. Freitas, K. A. et al. Enhanced T cell effector activity by targeting the Mediator kinase
91. Larson, R. C. et al. CAR T cell killing requires the IFNγR pathway in solid but not liquid module. Science 378, eabn5647 (2022).
tumours. Nature 604, 563–570 (2022). 126. Chen, J. et al. NR4A transcription factors limit CAR T cell function in solid tumours. Nature
This study showed that the IFNγR pathway was required for CAR-T activity against solid 567, 530–534 (2019).
but not liquid tumors. This study demonstrated enhanced T cell fitness by deletion of exhaustion-associated
92. Donovan, L. K. et al. Locoregional delivery of CAR T cells to the cerebrospinal fluid for transcription factors.
treatment of metastatic medulloblastoma and ependymoma. Nat. Med. 26, 720–731 127. Shifrut, E. et al. Genome-wide CRISPR screens in primary human T cells reveal key
(2020). regulators of immune function. Cell 175, 1958–1971.e15 (2018).
93. Theruvath, J. et al. Locoregionally administered B7-H3-targeted CAR T cells for treatment 128. Wei, J. et al. Targeting REGNASE-1 programs long-lived effector T cells for cancer therapy.
of atypical teratoid/rhabdoid tumors. Nat. Med. 26, 712–719 (2020). Nature 576, 471–476 (2019).
94. Priceman, S. J. et al. Regional delivery of chimeric antigen receptor-engineered T cells 129. Dong, M. B. et al. Systematic immunotherapy target discovery using genome-scale in vivo
effectively targets HER2+ breast cancer metastasis to the brain. Clin. Cancer Res. 24, CRISPR screens in CD8 T cells. Cell 178, 1189–1204.e23 (2019).
95–105 (2018). 130. LaFleur, M. W. et al. PTPN2 regulates the generation of exhausted CD8+ T cell subpopulations
95. Vitanza, N. A. et al. Locoregional infusion of HER2-specific CAR T cells in children and and restrains tumor immunity. Nat. Immunol. 20, 1335–1347 (2019).
young adults with recurrent or refractory CNS tumors: an interim analysis. Nat. Med. 27, 131. Carnevale, J. et al. RASA2 ablation in T cells boosts antigen sensitivity and long-term
1544–1552 (2021). function. Nature 609, 174–182 (2022).
96. Gauthier, J. et al. Factors associated with outcomes after a second CD19-targeted CAR 132. Si, J. et al. Hematopoietic progenitor kinase 1 (HPK1) mediates T cell dysfunction and is a
T-cell infusion for refractory B-cell malignancies. Blood 137, 323–335 (2021). druggable target for T cell-based immunotherapies. Cancer Cell 38, 551–566.e11 (2020).
97. Turtle, C. J. et al. CD19 CAR− T cells of defined CD4+:CD8+ composition in adult B cell ALL 133. Beavis, P. A. et al. Targeting the adenosine 2A receptor enhances chimeric antigen
patients. J. Clin. Invest. 1, 2123–2138 (2016). receptor T cell efficacy. J. Clin. Invest. 127, 929–941 (2017).
98. Adusumilli, P. S. et al. A phase I trial of regional mesothelin-targeted CAR T-cell therapy in 134. Gomes-Silva, D. et al. CD7-edited T cells expressing a CD7-specific CAR for the therapy of
patients with malignant pleural disease, in combination with the anti–PD-1 agent T-cell malignancies. Blood 130, 285–296 (2017).
pembrolizumab. Gynecol. Oncol. 11, 2748–2763 (2020). 135. Chun, I. et al. CRISPR–Cas9 knock out of CD5 enhances the anti-tumor activity of
99. Adachi, K. et al. IL-7 and CCL19 expression in CAR-T cells improves immune cell infiltration chimeric antigen receptor T cells. Blood 136, 51–52 (2020).
and CAR-T cell survival in the tumor. Nat. Biotechnol. 36, 346–351 (2018). 136. Shum, T. et al. Constitutive signaling from an engineered IL7 receptor promotes durable
100. Hegde, M. et al. Tandem CAR T cells targeting HER2 and IL13Rα2 mitigate tumor antigen tumor elimination by tumor-redirected T cells. Cancer Discov. 7, 1238–1247 (2017).
escape. J. Clin. Invest. 126, 3036–3052 (2016). 137. Kagoya, Y. et al. A novel chimeric antigen receptor containing a JAK–STAT signaling
101. Annesley, C. et al. SCRI-CAR19x22v2 T cell product demonstrates bispecific activity in domain mediates superior antitumor effects. Nat. Med. 24, 352–359 (2018).
B-ALL. Blood 138, 470 (2021). 138. Liu, E. et al. Use of CAR-transduced natural killer cells in CD19-positive lymphoid tumors.
102. Gardner, R. et al. Early clinical experience of CD19 × CD22 dual specific CAR T cells for N. Engl. J. Med. 382, 545–553 (2020).
enhanced anti-leukemic targeting of acute lymphoblastic leukemia. Blood 132, 278 This study showed that HLA-mismatched, cord blood-derived anti-CD19-CAR natural
(2018). killer cells were safe and effective against CD19-positive cancers.
103. Cordoba, S. et al. CAR T cells with dual targeting of CD19 and CD22 in pediatric and 139. Hege, K. M. et al. Safety, tumor trafficking and immunogenicity of chimeric antigen
young adult patients with relapsed or refractory B cell acute lymphoblastic leukemia: a receptor (CAR)-T cells specific for TAG-72 in colorectal cancer. J. Immunother. Cancer 5,
phase 1 trial. Nat. Med. 27, 1797–1805 (2021). 22 (2017).
104. Shah, N. N. et al. Bispecific anti-CD20, anti-CD19 CAR T cells for relapsed B cell 140. Brudno, J. N. et al. Safety and feasibility of anti-CD19 CAR T cells with fully human binding
malignancies: a phase 1 dose escalation and expansion trial. Nat. Med. 26, 1569–1575 domains in patients with B-cell lymphoma. Nat. Med. 26, 270–280 (2020).
(2020). 141. Shaim, H. et al. Targeting the αv integrin/TGF-β axis improves natural killer cell function
105. Zhang, Y. et al. Long-term activity of tandem CD19/CD20 CAR therapy in refractory/ against glioblastoma stem cells. J. Clin. Invest. 131, e142116 (2021).
relapsed B-cell lymphoma: a single-arm, phase 1–2 trial. Leukemia 36, 189–196 (2022). 142. Tang, N. et al. TGF-β inhibition via CRISPR promotes the long-term efficacy of CAR T cells
106. Martin, T. et al. Ciltacabtagene autoleucel, an anti–B-cell maturation antigen chimeric against solid tumors. JCI Insight 5, e133977 (2020).
antigen receptor T-Cell therapy, for relapsed/refractory multiple myeloma: CARTITUDE-1 143. Cherkassky, L. et al. Human CAR T cells with cell-intrinsic PD-1 checkpoint blockade resist
2-year follow-up. J. Clin. Oncol. https://doi.org/10.1200/JCO.22.00842 (2022). tumor-mediated inhibition. J. Clin. Invest. 126, 3130–3144 (2016).
107. Katsarou, A. et al. Combining a CAR and a chimeric costimulatory receptor enhances 144. Ren, J. et al. A versatile system for rapid multiplex genome-edited CAR T cell generation.
T cell sensitivity to low antigen density and promotes persistence. Sci. Transl. Med. 13, Oncotarget 8, 17002–17011 (2017).
1962 (2021). 145. Yamamoto, T. N. et al. T cells genetically engineered to overcome death signaling
108. Johnson, L. R. et al. The immunostimulatory RNA RN7SL1 enables CAR-T cells to enhance enhance adoptive cancer immunotherapy. J. Clin. Invest. 129, 1551–1565 (2019).
autonomous and endogenous immune function. Cell 184, 4981–4995.e14 (2021). 146. Rafiq, S. et al. Targeted delivery of a PD-1-blocking scFV by CAR-T cells enhances
109. Agliardi, G. et al. Intratumoral IL-12 delivery empowers CAR-T cell immunotherapy in a anti-tumor efficacy in vivo. Nat. Biotechnol. 36, 847–858 (2018).
pre-clinical model of glioblastoma. Nat. Commun. 12, 444 (2021). 147. Mohammed, S. et al. Improving chimeric antigen receptor-modified T cell function by
110. Alsaieedi, A., Holler, A., Velica, P., Bendle, G. & Stauss, H. J. Safety and efficacy of reversing the immunosuppressive tumor microenvironment of pancreatic cancer. Mol.
Tet-regulated IL-12 expression in cancer-specific T cells. Oncoimmunology 8, 1542917 Ther. 25, 249–258 (2017).
(2018). 148. Oda, S. K. et al. A Fas-4-1BB fusion protein converts a death to a pro-survival signal and
111. Avanzi, M. P. et al. Engineered tumor-targeted T cells mediate enhanced anti-tumor enhances T cell therapy. J. Exp. Med. 217, e20191166 (2020).
efficacy both directly and through activation of the endogenous immune system. Cell 149. Liao, Z. et al. Leveraging biomaterials for enhancing T cell immunotherapy. J. Control.
Rep. 23, 2130–2141 (2018). Release 344, 272–288 (2022).
112. Curran, K. J. et al. Enhancing antitumor efficacy of chimeric antigen receptor T cells 150. Eyquem, J. et al. Targeting a CAR to the TRAC locus with CRISPR/Cas9 enhances tumour
through constitutive CD40L expression. Mol. Ther. 23, 769–778 (2015). rejection. Nature 543, 113–117 (2017).
113. Lai, J. et al. Adoptive cellular therapy with T cells expressing the dendritic cell growth This study demonstrated that genetic engineering of a CAR into the TRAC locus
factor Flt3L drives epitope spreading and antitumor immunity. Nat. Immunol. 21, 914–926 improved T cell fitness compared to viral overexpression.
(2020). 151. Straathof, K. C. et al. An inducible caspase 9 safety switch for T-cell therapy. Blood 105,
114. Choi, B. D. et al. CAR-T cells secreting BiTEs circumvent antigen escape without detectable 4247–4254 (2005).
toxicity. Nat. Biotechnol. 37, 1049–1058 (2019). 152. Bordignon, C. et al. Transfer of the HSV-tk gene into donor peripheral blood lymphocytes
115. Fraietta, J. A. et al. Disruption of TET2 promotes the therapeutic efficacy of CD19-targeted for in vivo modulation of donor anti tumor immunity after allogeneic bone marrow
T cells. Nature 558, 307–312 (2018). transplantation. Hum. Gene Ther. 6, 813–819 (1995).
This study showed that random integration of the CAR into the TET2 locus disrupted the 153. Wang, X. et al. A transgene-encoded cell surface polypeptide for selection, in vivo
gene and resulted in improved CAR-T function and persistence in a patient with CLL. tracking, and ablation of engineered cells. Blood 118, 1255–1263 (2011).
116. Prinzing, B. et al. Deleting DNMT3A in CAR T cells prevents exhaustion and enhances 154. Wiebking, V. et al. Metabolic engineering generates a transgene-free safety switch for cell
antitumor activity. Sci. Transl. Med. 13, eabh0272 (2021). therapy. Nat. Biotechnol. 38, 1441–1450 (2020).
117. Seo, H. et al. BATF and IRF4 cooperate to counter exhaustion in tumor-infiltrating CAR 155. Sakemura, R. et al. A Tet-On inducible system for controlling CD19-chimeric antigen
T cells. Nat. Immunol. 22, 983–995 (2021). receptor expression upon drug administration. Cancer Immunol. Res. 4, 658–668 (2016).
118. Ghassemi, S. et al. Reducing ex vivo culture improves the antileukemic activity of chimeric 156. Wu, C.-Y., Roybal, K. T., Puchner, E. M., Onuffer, J. & Lim, W. A. Remote control of
antigen receptor (CAR) T cells. Cancer Immunol. Res. 6, 1100–1109 (2018). therapeutic T cells through a small molecule-gated chimeric receptor. Science 350,
119. Zheng, W. et al. PI3K orchestration of the in vivo persistence of chimeric antigen receptor- aab4077 (2015).
modified T cells. Leukemia 32, 1157–1167 (2018). 157. Jan, M. et al. Reversible ON- and OFF-switch chimeric antigen receptors controlled by
120. Good, Z. et al. Proliferation tracing with single-cell mass cytometry optimizes generation lenalidomide. Sci. Transl. Med. 13, eabb6295 (2021).
of stem cell memory-like T cells. Nat. Biotechnol. 37, 259–266 (2019). 158. Giordano-Attianese, G. et al. A computationally designed chimeric antigen receptor provides
121. Weber, E. W. et al. Transient rest restores functionality in exhausted CAR-T cells through a small-molecule safety switch for T-cell therapy. Nat. Biotechnol. 38, 426–432 (2020).
epigenetic remodeling. Science 372, eaba1786 (2021). 159. Zajc, C. U. et al. A conformation-specific ON-switch for controlling CAR T cells with an
This study showed that CAR-T functionality could be restored and features of orally available drug. Proc. Natl Acad. Sci. USA 117, 14926–14935 (2020).
exhaustion could be diminished by providing CAR-T cells with periods of rest. 160. Lee, S. M. et al. A chemical switch system to modulate chimeric antigen receptor T cell
122. Xu, Y. et al. Closely related T-memory stem cells correlate with in vivo expansion of CAR. activity through proteolysis-targeting chimaera technology. ACS Synth. Biol. 9, 987–992
CD19-T cells and are preserved by IL-7 and IL-15. Blood 123, 3750–3759 (2014). (2020).

Nature | Vol 614 | 23 February 2023 | 645

Perspective
161. Richman, S. A. et al. Ligand-induced degradation of a CAR permits reversible remote 191. Kagoya, Y. et al. Genetic ablation of HLA class I, class II, and the T-cell receptor enables
control of CAR T cell activity in vitro and in vivo. Mol. Ther. 28, 1932 (2020). allogeneic T cells to be used for adoptive T-cell therapy. Cancer Immunol. Res. 8,
162. Labanieh, L. et al. Enhanced safety and efficacy of protease-regulated CAR-T cell 926–936 (2020).
receptors. Cell 185, 1745–1763.e22 (2022). 192. Wang, B. et al. Generation of hypoimmunogenic T cells from genetically engineered
The SNIP drug-regulated CAR system manifested a full dynamic range of activity allogeneic human induced pluripotent stem cells. Nat. Biomed. Eng. 5, 429–440 (2021).
devoid of leakiness, and in a ROR1 toxicity model could be tuned to selectively target 193. Hu, X. et al. Engineered hypoimmune allogeneic CAR T cells exhibit innate and adaptive
tumour cells expressing high levels of ROR1 while sparing healthy cells expressing low immune evasion even after sensitization in humanized mice and retain potent anti-tumor
levels of ROR1. activity. Blood 138, 1690 (2021).
163. Li, H.-S. et al. High-performance multiplex drug-gated CAR circuits. Cancer Cell 40, 194. Anzalone, A. V., Koblan, L. W. & Liu, D. R. Genome editing with CRISPR–Cas nucleases,
1294–1305 (2022). base editors, transposases and prime editors. Nat. Biotechnol. 38, 824–844 (2020).
164. Hyrenius-Wittsten, A. et al. SynNotch CAR circuits enhance solid tumor recognition and 195. Simonetta, F. et al. Allogeneic CAR invariant natural killer T cells exert potent antitumor
promote persistent antitumor activity in mouse models. Sci. Transl. Med. 13, eabd8836 effects through host CD8 T-cell cross-priming. Clin. Cancer Res. 27, 6054–6064 (2021).
(2021). 196. Heczey, A. et al. Anti-GD2 CAR-NKT cells in patients with relapsed or refractory
This study demonstrated that SynNotch-inducible CAR-T cells have lower tonic neuroblastoma: an interim analysis. Nat. Med. 26, 1686–1690 (2020).
signalling, greater stemness and enhanced potency compared with conventional 197. Neelapu, S. S. et al. A phase 1 study of ADI-001: anti-CD20 CAR-engineered allogeneic
CAR-T cells. gamma delta (γδ) T cells in adults with B-cell malignancies. J. Clin. Oncol. 40, 7509
165. Hernandez-Lopez, R. A. et al. T cell circuits that sense antigen density with an ultrasensitive (2022).
threshold. Science 371, 1166–1171 (2021). 198. Klichinsky, M. et al. Human chimeric antigen receptor macrophages for cancer
166. Van Oekelen, O. et al. Neurocognitive and hypokinetic movement disorder with features immunotherapy. Nat. Biotechnol. 38, 947–953 (2020).
of parkinsonism after BCMA-targeting CAR-T cell therapy. Nat. Med. 27, 2099–2103 (2021). The authors adapted vectors for efficient viral gene delivery of a CAR to macrophages.
167. Lamers, C. H. J. et al. Treatment of metastatic renal cell carcinoma with CAIX CAR- 199. Morrissey, M. A. et al. Chimeric antigen receptors that trigger phagocytosis. eLife 7,
engineered T cells: Clinical evaluation and management of on-target toxicity. Mol. Ther. e36688 (2018).
21, 904–912 (2013). 200. Themeli, M. et al. Generation of tumor-targeted human T lymphocytes from induced
168. Thistlethwaite, F. C. et al. The clinical efficacy of first-generation carcinoembryonic pluripotent stem cells for cancer therapy. Nat. Biotechnol. 31, 928–933 (2013).
antigen (CEACAM5)-specific CAR T cells is limited by poor persistence and transient 201. Zhu, H. et al. Metabolic reprograming via deletion of CISH in human iPSC-derived NK
pre-conditioning-dependent respiratory toxicity. Cancer Immunol. Immunother. 66, cells promotes in vivo persistence and enhances anti-tumor activity. Cell Stem Cell 27,
1425–1436 (2017). 224–237.e6 (2020).
169. Straathof, K. et al. Antitumor activity without on-target off-tumor toxicity of GD2-chimeric 202. Zhang, L. et al. Pluripotent stem cell-derived CAR-macrophage cells with antigen-
antigen receptor T cells in patients with neuroblastoma. Sci. Transl. Med. 12, eabd6169 dependent anti-cancer cell functions. J. Hematol. Oncol. 13, 153 (2020).
(2020). 203. Bachanova, V. et al. Safety and efficacy of FT596, a first-in-class, multi-antigen targeted,
170. Mount, C. W. et al. Potent antitumor efficacy of anti-GD2 CAR T cells in H3-K27M+ diffuse off-the-shelf, iPSC-derived CD19 CAR NK cell therapy in relapsed/refractory B-cell
midline gliomas letter. Nat. Med. 24, 572–579 (2018). lymphoma. Blood 138, 823 (2021).
171. Kloss, C. C. et al. Dominant-negative TGF-β receptor enhances PSMA-targeted human 204. Wang, Z. et al. 3D-organoid culture supports differentiation of human CAR+ iPSCs into
CAR T cell proliferation and augments prostate cancer eradication. Mol. Ther. 26, highly functional CAR Tcells. Cell Stem Cell 29, 515–527.e8 (2022).
1855–1866 (2018). 205. Roth, T. L. et al. Reprogramming human T cell function and specificity with non-viral
172. Narayan, V. et al. PSMA-targeting TGFβ-insensitive armored CAR T cells in metastatic genome targeting. Nature 559, 405–409 (2018).
castration-resistant prostate cancer: a phase 1 trial. Nat. Med. 28, 724–734 (2022). The authors developed a non-viral approach to engineer T cells with genomic
173. Roybal, K. T. et al. Precision tumor recognition by T cells with combinatorial precision.
antigen-sensing circuits. Cell 164, 770–779 (2016). 206. Zhang, J. et al. Non-viral, specifically targeted CAR-T cells achieve high safety and
174. Zhu, I. et al. Modular design of synthetic receptors for programmed gene regulation in efficacy in B-NHL. Nature 609, 369–374 (2022).
cell therapies. Cell 185, 1431–1443.e16 (2022). 207. Shy, B. R. et al. High-yield genome engineering in primary cells using a hybrid ssDNA
175. Srivastava, S. et al. Logic-gated ROR1 chimeric antigen receptor expression rescues repair template and small-molecule cocktails. Nat. Biotechnol. https://doi.org/10.1038/
T cell-mediated toxicity to normal tissues and enables selective tumor targeting. Cancer S41587-022-01418-8 (2022).
Cell 35, 489–503.e8 (2019). 208. Micklethwaite, K. P. et al. Investigation of product-derived lymphoma following infusion
176. Tousley, A. M. et al. Coopting T cell proximal signaling molecules enables Boolean logic-gated of piggyBac-modified CD19 chimeric antigen receptor T cells. Blood 138, 1391–1405
CAR T cell control. Preprint at bioRxiv https://doi.org/10.1101/2022.06.17.496457 (2022). (2021).
LINK CAR utilizes proximal signalling molecules fused to membrane-bound scFvs to 209. Weidner, T. et al. Genetic in vivo engineering of human T lymphocytes in mouse models.
create an AND-gated CAR system. Nat. Protoc. 16, 3210–3240 (2021).
177. Cho, J. H., Collins, J. J. & Wong, W. W. Universal chimeric antigen receptors for multiplexed 210. Smith, T. T. et al. In situ programming of leukaemia-specific T cells using synthetic DNA
and logical control of T cell responses. Cell 173, 1426–1438.e11 (2018). nanocarriers. Nat. Nanotechnol. 12, 1–21 (2017).
178. Lajoie, M. J. et al. Designed protein logic to target cells with precise combinations of 211. Rurik, J. G. et al. CAR T cells produced in vivo to treat cardiac injury. Science 375, 91–96
surface antigens. Science 369, 1637–1643 (2020). (2022).
179. Fedorov, V. D., Themeli, M. & Sadelain, M. PD-1- and CTLA-4-based inhibitory chimeric This study utilized CD5-targeting nanoparticles for in vivo gene engineering of T cells
antigen receptors (iCARs) divert off-target immunotherapy responses. Sci. Transl. Med. 5, to treat cardiac fibrosis.
215ra172 (2013). 212. Kansal, R. et al. Sustained B cell depletion by CD19-targeted CAR T cells is a highly
180. Richards, R. M. et al. NOT-gated CD93 CAR T cells effectively target AML with minimized effective treatment for murine lupus. Sci. Transl. Med. 11, eaav1648 (2019).
endothelial cross-reactivity. Blood Cancer Discov. 2, 648–665 (2021). 213. Mougiakakos, D. et al. CD19-targeted CAR T cells in refractory systemic lupus erythematosus.
181. Tokatlian, T. et al. Mesothelin-specific CAR-T cell therapy that incorporates an HLA-gated N. Engl. J. Med. 385, 567–569 (2021).
safety mechanism selectively kills tumor cells. J. Immunother. Cancer 10, e003826 This case report showed a strong clinical benefit for treatment of systemic lupus
(2022). erythematosus with CD19-CAR T cells.
182. Mansilla-Soto, J. et al. HLA-independent T cell receptors for targeting tumors with low 214. Raffin, C., Vo, L. T. & Bluestone, J. A. Treg cell-based therapies: challenges and perspectives.
antigen density. Nat. Med. 28, 345–352 (2022). Nat. Rev. Immunol. 20, 158–172 (2020).
This study showed that T cells expressing HIT receptors targeted to the TRAC 215. MacDonald, K. G. et al. Alloantigen-specific regulatory T cells generated with a chimeric
locus are able to recognize and kill tumour cells with low antigen density. antigen receptor. J. Clin. Invest. 126, 1413–1424 (2016).
183. Liu, Y. et al. Chimeric STAR receptors using TCR machinery mediate robust responses 216. Noyan, F. et al. Prevention of allograft rejection by use of regulatory T cells with an MHC-
against solid tumors. Sci. Transl. Med. 13, eabb5191 (2021). specific chimeric antigen receptor. Am. J. Transplant. 17, 917–930 (2017).
184. Xu, Y. et al. A novel antibody–TCR (AbTCR) platform combines Fab-based antigen 217. Good, Z. et al. Post-infusion CAR Treg cells identify patients resistant to CD19-CAR therapy.
recognition with gamma/delta-TCR signaling to facilitate T-cell cytotoxicity with low Nat. Med. 28, 1860–1871 (2022).
cytokine release. Cell Discov. 4, 62 (2018). 218. Haradhvala, N. J. et al. Distinct cellular dynamics associated with response to CAR-T
185. Baeuerle, P. A. et al. Synthetic TRuC receptors engaging the complete T cell receptor for therapy for refractory B cell lymphoma. Nat. Med. 28, 1848–1859 (2022).
potent anti-tumor response. Nat. Commun. 10, 2087 (2019). 219. Fu, R. Y. et al. CD4+ T cells engineered with FVIII-CAR and murine Foxp3 suppress anti-factor
186. Burton, J. et al. Inefficient exploitation of accessory receptors reduces the sensitivity of VIII immune responses in hemophilia a mice. Cell. Immunol. 358, 104216 (2020).
chimeric antigen receptors. Preprint at bioRxiv https://doi.org/10.1101/2021.10.26.465853 220. Amor, C. et al. Senolytic CAR T cells reverse senescence-associated pathologies. Nature
(2022). 583, 127–132 (2020).
187. Yarmarkovich, M. et al. Cross-HLA targeting of intracellular oncoproteins with peptide- This study showed that CAR T cells directed against the urokinase-type plasminogen
centric CARs. Nature 599, 477–484 (2021). activator receptor could be used as senolytic agents.
The authors developed a peptide-centric CAR targeting a peptide derived from an 221. Aghajanian, H. et al. Targeting cardiac fibrosis with engineered T cells. Nature 573,
intracellular protein that is overexpressed in neuroblastoma. 430–433 (2019).
188. Maschan, M. et al. Multiple site place-of-care manufactured anti-CD19 CAR-T cells induce 222. Jacobson, C. A. et al. Axicabtagene ciloleucel in relapsed or refractory indolent non-
high remission rates in B-cell malignancy patients. Nat. Commun. 12, 7200 (2021). Hodgkin lymphoma (ZUMA-5): a single-arm, multicentre, phase 2 trial. Lancet Oncol. 23,
189. Pearson, A. D. et al. Paediatric Strategy Forum for medicinal product development of 91–103 (2022).
chimeric antigen receptor T-cells in children and adolescents with cancer. Eur. J. Cancer 223. Shah, B. D. et al. KTE-X19 for relapsed or refractory adult B-cell acute lymphoblastic
160, 112–133 (2022). leukaemia: phase 2 results of the single-arm, open-label, multicentre ZUMA-3 study.
190. Benjamin, R. et al. Genome-edited, donor-derived allogeneic anti-CD19 chimeric antigen Lancet 398, 491–502 (2021).
receptor T cells in paediatric and adult B-cell acute lymphoblastic leukaemia: results of 224. Frank, M. J. et al. CD22-CAR T-cell therapy mediates high durable remission rates in adults
two phase 1 studies. Lancet 396, 1885–1894 (2020). with large B-cell lymphoma who have relapsed after CD19-CAR T-cell therapy. Blood 138,
This study showed the feasibility of using allogeneic, genome-edited CAR T cells. 741 (2021).

646 | Nature | Vol 614 | 23 February 2023

225. Cui, Q. et al. CD38-directed CAR-T cell therapy: a novel immunotherapy strategy for 259. Jung, I. Y. et al. CRISPR/Cas9-mediated knockout of DGK improves antitumor activities of
relapsed acute myeloid leukemia after allogeneic hematopoietic stem cell transplantation. human T cells. Cancer Res. 78, 4692–4703 (2018).
J. Hematol. Oncol. 14, 82 (2021). 260. Zhang, Y. et al. CRISPR–Cas9 mediated LAG-3 disruption in CAR-T cells. Front. Med. 11,
226. Ramos, C. A. et al. Clinical responses with T lymphocytes targeting malignancy- 554–562 (2017).
associated κ light chains. J. Clin. Invest. 126, 2588–2596 (2016). 261. Sterner, R. M. et al. GM-CSF inhibition reduces cytokine release syndrome and
227. Louis, C. U. et al. Antitumor activity and long-term fate of chimeric antigen receptor- neuroinflammation but enhances CAR-T cell function in xenografts. Blood 133, 697–709
positive T cells in patients with neuroblastoma. Blood 118, 6050–6056 (2011). (2019).
228. Majzner, R. G. et al. Major tumor regressions in H3K27M-mutated diffuse midline glioma 262. Qasim, W. et al. Molecular remission of infant B-ALL after infusion of universal TALEN
(DMG) following sequential intravenous (IV) and intracerebroventricular (ICV) delivery of gene-edited CAR T cells. Sci. Transl. Med. 9, eaaj2013 (2017).
GD2-CAR T cells. Cancer Res. 82, CT001 (2022). 263. Johnson, A. J. et al. Rationally designed transgene-encoded cell-surface polypeptide tag
229. Navai, S. A. et al. Administration of HER2-CAR T cells after lymphodepletion safely for multiplexed programming of CAR T-cell synthetic outputs. Cancer Immunol. Res. 9,
improves T cell expansion and induces clinical responses in patients with advanced 1047–1060 (2021).
sarcomas. Cancer Res. 79, LB147 (2019). 264. Becerra, C. R. et al. Ligand-inducible, prostate stem cell antigen (PSCA)-directed GoCAR-T
230. Guo, Y. et al. Phase I study of chimeric antigen receptor–modified T cells in patients with cells in advanced solid tumors: preliminary results from a dose escalation. J. Clin. Oncol.
EGFR-positive advanced biliary tract cancers. Clin. Cancer Res. 24, 1277–1286 (2018). 37, 282–283 (2019).
231. Irving, B. A. & Weiss, A. The cytoplasmic domain of the T cell receptor zeta chain is 265. Munoz, J. et al. A phase 1 study of ACTR087 in combination with rituximab, in subjects
sufficient to couple to receptor-associated signal transduction pathways. Cell 64, with relapsed or refractory CD20-positive B-cell lymphoma. Blood 134, 244 (2019).
891–901 (1991). 266. Ma, J. S. Y. et al. Versatile strategy for controlling the specificity and activity of engineered
232. Levine, B. L. et al. Large-scale production of CD4+ T cells from HIV-1-infected donors after T cells. Proc. Natl Acad. Sci. USA 113, E450–E458 (2016).
CD3/CD28 costimulation. J. Hematotherapy Stem Cell Res. 7, 437–448 (1998). 267. Nikolaenko, L. et al. First in human study of an on/off switchable CAR-T cell platform
233. Haynes, N. M. et al. Single-chain antigen recognition receptors that costimulate potent targeting CD19 for B cell malignancies (CLBR001 + SWI019). Blood 138, 2822 (2021).
rejection of established experimental tumors. Blood 100, 3155–3163 (2002). 268. Frigault, M. J. et al. Phase 1 study of CART-ddBCMA in relapsed or refractory multiple
234. Fry, T. J. et al. A potential role for interleukin-7 in T-cell homeostasis. Blood 97, 2983–2990 myeloma. J. Clin. Oncol. 40, 8003–8003 (2022).
(2001). 269. Wu, Y. et al. Control of the activity of CAR-T cells within tumours via focused ultrasound.
235. Dudley, M. E. et al. Cancer regression and autoimmunity in patients after clonal repopulation Nat. Biomed. Eng. 5, 1336–1347 (2021).
with antitumor lymphocytes. Science 298, 850–854 (2002). 270. Allen, M. E. et al. An AND-gated drug and photoactivatable Cre- loxP system for
236. Pule, M. A. et al. Virus-specific T cells engineered to coexpress tumor-specific receptors: spatiotemporal control in cell-based therapeutics. ACS Synth. Biol. 8, 2359–2371
persistence and antitumor activity in individuals with neuroblastoma. Nat. Med. 14, (2019).
1264–1270 (2008). 271. O’Cearbhaill, R. E. et al. A phase I clinical trial of autologous chimeric antigen receptor
237. Milone, M. C. et al. Chimeric receptors containing CD137 signal transduction domains (CAR) T cells genetically engineered to secrete IL-12 and to target the MUC16ecto antigen
mediate enhanced survival of T cells and increased antileukemic efficacy in vivo. Mol. Ther. in patients (pts) with MUC16ecto+ recurrent high-grade serous ovarian cancer (HGSOC).
17, 1453–1464 (2009). Gynecol. Oncol. 159, 42 (2020).
238. Kochenderfer, J. N. et al. Eradication of B-lineage cells and regression of lymphoma in a 272. Hu, B. et al. Augmentation of antitumor immunity by human and mouse CAR T cells
patient treated with autologous T cells genetically engineered to recognize CD19. Blood secreting IL-18. Cell Rep. 20, 3025–3033 (2017).
116, 4099–4102 (2010). 273. Svoboda, J. et al. Interleukin-18 secreting autologous anti-CD19 CAR T-cells (huCART19-IL18)
239. Brentjens, R. J. et al. Safety and persistence of adoptively transferred autologous CD19- in patients with non-Hodgkin lymphomas relapsed or refractory to prior CAR T-cell
targeted T cells in patients with relapsed or chemotherapy refractory B-cell leukemias. therapy. Blood 140, 4612–4614 (2022).
Blood 118, 4817–4828 (2011). 274. Zhang, L. et al. Improving adoptive T cell therapy by targeting and controlling IL-12
240. Savoldo, B. et al. CD28 costimulation improves expansion and persistence of chimeric expression to the tumor environment. Mol. Ther. 19, 751–759 (2011).
antigen receptor-modified T cells in lymphoma patients. J. Clin. Invest. 121, 1822–1826 275. Batra, S. A. et al. Glypican-3-specific CAR T cells coexpressing IL15 and IL21 have superior
(2011). expansion and antitumor activity against hepatocellular carcinoma. Cancer Immunol.
241. Porter, D. L., Levine, B. L., Kalos, M., Bagg, A. & June, C. H. Chimeric antigen receptor- Res. 8, 309–320 (2020).
modified T cells in chronic lymphoid leukemia. N. Engl. J. Med. 365, 725–33 (2011). 276. Liu, X. et al. A chimeric switch-receptor targeting PD1 augments the efficacy of
242. Hudecek, M. et al. Receptor affinity and extracellular domain modifications affect tumor second-generation CAR T cells in advanced solid tumors. Cancer Res. 76, 1578–1590
recognition by ROR1-specific chimeric antigen receptor T cells. Clin. Cancer Res. 19, (2016).
3153–3164 (2013). 277. Papa, S. et al. A phase I trial of T4 CAR T-cell immunotherapy in head and neck squamous
243. Landoni, E. et al. Modifications to the framework regions eliminate chimeric antigen cancer (HNSCC). J. Clin. Oncol. 36, 3046–3046 (2018).
receptor tonic signaling. Cancer Immunol. Res. 9, 441–453 (2021). 278. Kakarla, S. et al. Antitumor effects of chimeric receptor engineered human T cells directed
244. Awasthi, R. et al. Tisagenlecleucel cellular kinetics, dose, and immunogenicity in relation to tumor stroma. Mol. Ther. 21, 1611–1620 (2013).
to clinical factors in relapsed/refractory DLBCL. Blood Adv. 4, 560–572 (2020). 279. Caruana, I. et al. Heparanase promotes tumor infiltration and antitumor activity of
245. Leddon, S. A. et al. The CD28 transmembrane domain contains an essential dimerization CAR-redirected T lymphocytes. Nat. Med. 21, 524–9 (2015).
motif. Front. Immunol. 11, 1519 (2020). 280. Ligtenberg, M. A. et al. Coexpressed catalase protects chimeric antigen receptor-
246. Bridgeman, J. S. et al. The optimal antigen response of chimeric antigen receptors redirected T cells as well as bystander cells from oxidative stress-induced loss of
harboring the CD3 transmembrane domain is dependent upon incorporation of the antitumor activity. J. Immunol. 196, 759–66 (2016).
receptor into the endogenous TCR/CD3 complex. J. Immunol. 184, 6938–6949 (2010). 281. Boice, M. et al. Loss of the HVEM tumor suppressor in lymphoma and restoration by
247. Fujiwara, K. et al. Hinge and transmembrane domains of chimeric antigen receptor regulate modified CAR-T Cells. Cell 167, 405–418.e13 (2016).
receptor expression and signaling threshold. Cells 9, 1182 (2020). 282. Newick, K. et al. Augmentation of CAR T-cell trafficking and antitumor efficacy by
248. Brentjens, R. J. et al. Eradication of systemic B-cell tumors by genetically targeted blocking protein kinase a localization. Cancer Immunol. Res. 4, 541–551 (2016).
human T lymphocytes co-stimulated by CD80 and interleukin-15. Nat. Med. 9, 279–86 283. Li, G. et al. CXCR5 guides migration and tumor eradication of anti-EGFR chimeric antigen
(2003). receptor T cells. Mol. Ther. 22, 507–517 (2021).
249. Kawalekar, O. U. et al. Distinct signaling of coreceptors regulates specific metabolism 284. Di Stasi, A. et al. T lymphocytes coexpressing CCR4 and a chimeric antigen receptor
pathways and impacts memory development in CAR T cells. Immunity 44, 380–390 targeting CD30 have improved homing and antitumor activity in a Hodgkin tumor model.
(2016). Blood 113, 6392–6402 (2009).
250. Yan, Z. et al. A combination of humanised anti-CD19 and anti-BCMA CAR T cells in patients 285. Craddock, J. A. et al. Enhanced tumor trafficking of GD2 chimeric antigen receptor
with relapsed or refractory multiple myeloma: a single-arm, phase 2 trial. Lancet Haematol. T cells by expression of the chemokine receptor CCR2b. J. Immunother. 33, 780–788
6, e521–e529 (2019). (2010).
251. Schneider, D. et al. Trispecific CD19–CD20–CD22-targeting duoCAR-T cells eliminate 286. Siddiqui, I., Erreni, M., van Brakel, M., Debets, R. & Allavena, P. Enhanced recruitment
antigen-heterogeneous B cell tumors in preclinical models. Sci. Transl. Med. 13, of genetically modified CX3CR1-positive human T cells into Fractalkine/CX3CL1
eabc6401 (2021). expressing tumors: importance of the chemokine gradient. J. Immunother. Cancer 4, 21
252. Kloss, C. C., Condomines, M., Cartellieri, M., Bachmann, M. & Sadelain, M. Combinatorial (2016).
antigen recognition with balanced signaling promotes selective tumor eradication by 287. Feng, J. et al. Treatment of aggressive T cell lymphoblastic lymphoma/leukemia using
engineered T cells. Nat. Biotechnol. 31, 71–75 (2013). anti-CD5 CAR T cells. Stem Cell Rev. Rep. 17, 652–661 (2021).
253. Han, X. et al. Masked chimeric antigen receptor for tumor-specific activation. Mol. Ther. 288. Hurton, L. V. et al. Tethered IL-15 augments antitumor activity and promotes a stem-cell
25, 274–284 (2017). memory subset in tumor-specific T cells. Proc. Natl Acad. Sci. USA 113, E7788–E7797
254. Kosti, P. et al. Hypoxia-sensing CAR T cells provide safety and efficacy in treating solid (2016).
tumors. Cell Rep. Med. 2, 100227 (2021). 289. Hunter, M. R. et al. Chimeric γc cytokine receptors confer cytokine independent
255. Ho, P., Ede, C. & Chen, Y. Y. Modularly constructed synthetic granzyme B molecule engraftment of human T lymphocytes. Mol. Immunol. 56, 1–11 (2013).
enables interrogation of intracellular proteases for targeted cytotoxicity. ACS Synth. Biol. 290. Zhang, Q. et al. A human orthogonal IL-2 and IL-2Rβ system enhances CAR T cell expansion
6, 1484–1495 (2017). and antitumor activity in a murine model of leukemia. Sci. Transl. Med. 13, eabg6986
256. Scharping, N. E. et al. The tumor microenvironment represses T cell mitochondrial (2021).
biogenesis to drive intratumoral T cell metabolic insufficiency and dysfunction. Immunity 291. Kalbasi, A. et al. Potentiating adoptive cell therapy using synthetic IL-9 receptors. Nature
45, 374–388 (2016). 607, 360–365 (2022).
257. Seo, H. et al. TOX and TOX2 transcription factors cooperate with NR4A transcription 292. Krause, A. et al. Antigen-dependent CD28 signaling selectively enhances survival and
factors to impose CD8+ T cell exhaustion. Proc. Natl Acad. Sci. USA 116, 12410–12415 proliferation in genetically modified activated human primary T lymphocytes. J. Exp.
(2019). Med. 188, 619–626 (1998).
258. Yang, W. et al. Potentiating the antitumour response of CD8+ T cells by modulating 293. Ren, J. et al. Multiplex genome editing to generate universal CAR T cells resistant to PD1
cholesterol metabolism. Nature 531, 651–655 (2016). inhibition. Clin. Cancer Res. 23, 2255–2266 (2017).

Nature | Vol 614 | 23 February 2023 | 647

Perspective
294. Kamiya, T., Wong, D., Png, Y. T. & Campana, D. A novel method to generate T-cell Squibb. L.L. is a cofounder of, consults for, and holds equity in CARGO Therapeutics. L.L. is a
receptor-deficient chimeric antigen receptor T cells. Blood Adv. 2, 517–528 (2018). consultant for and holds equity in Lyell Immunopharma.

Additional information
Acknowledgements C.L.M. and L.L are members of the Parker Institute for Cancer
Correspondence and requests for materials should be addressed to Crystal L. Mackall.
Immunotherapy, which supports the Stanford University Cancer Immunotherapy Program.
Peer review information Nature thanks Justin Eyquem, Kristen Hege, Hans Stauss and Jie Sun
This work was supported by the St Baldrick’s Foundation Empowering Pediatric
for their contribution to the peer review of this work.
Immunotherapies for Childhood Cancer (EPICC) Team and NCI 5P30CA124435 (C.L.M.).
Reprints and permissions information is available at http://www.nature.com/reprints.
L.L. was supported by a Siebel Scholars Fellowship, Stanford Graduate Fellowship,
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
National Science Foundation Graduate Research Fellowship (DGE-1656518), and Discovery
published maps and institutional affiliations.
Innovation Award philanthropically supported through the Biomedical Innovation Initiative.

Author contributions L.L. and C.L.M. conceptualized, wrote and edited the manuscript. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this
article under a publishing agreement with the author(s) or other rightsholder(s); author
Competing interests L.L. and C.L.M. are inventors on several patents related to CAR T cell self-archiving of the accepted manuscript version of this article is solely governed by the
therapies. C.L.M. is a cofounder of Lyell Immunopharma, CARGO Therapeutics and Link Cell terms of such publishing agreement and applicable law.
Therapies, which are developing CAR-based therapies, and consults for Lyell, Syncopation,
Link, Apricity, Nektar, Immatics, Ensoma, Mammoth, Glaxo Smith Kline and Bristol Myers © Springer Nature Limited 2023, corrected publication 2023

648 | Nature | Vol 614 | 23 February 2023