Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator

Article abstract of DOI:10.1016/j.chembiol.2020.10.004

Chimeric antigen receptor (CAR)-T-based therapeutics are a breakthrough in cancer treatment; however, they are hampered by constitutive activation, which leads to worrisome side effects. Engineering CAR-T cells to be as tightly controllable as possible remains a topic of ongoing investigation. Here, we report a photoswitchable approach that uses a mediator for the at-will regulation of CAR-T cells. This mediator carries dual folate and fluorescein isothiocyanate moieties tethered by an ortho-nitrobenzyl ester photocleavable linker. CAR-T cells were shown to be highly cytotoxic to targeted cells only in the presence of the mediator and acted in a dose-dependent manner. The toxicity of CAR-T cells can be rapidly terminated by cleavage of the mediator, and the effects of CAR-T cells can be activated again by resupplementation with the mediator without compromising tumor therapy. The approach described here provides a direction for enhancing the controllability of CAR-T cells and can likely be applied in other immunotherapies.CAR-T is a powerful technology for cancer therapy, but is largely limited by inherent controllability issues. Zhang et al. developed an accurate controllable approach based on the bond-cleavage chemistry combined with universal anti-FITC CAR-T cells, allowing the regulation of CAR-T cells in a switchable manner.

Full text of DOI:10.1016/j.chembiol.2020.10.004

Article
Photoswitchable CAR-T Cell Function In Vitro and
In Vivo via a Cleavable Mediator
Graphical Abstract
Authors
Bo Zhang, Yan Wang,
Shenlong Huang, ..., Demin Zhou,
Lingjun Li, Xuan Zhang
Correspondence
deminzhou@bjmu.edu.cn (D.Z.),
lingjunlee@aliyun.com (L.L.),
zxpumch2003@sina.com (X.Z.)
In Brief
CAR-T is a powerful technology for
cancer therapy, but is largely limited by
inherent controllability issues. Zhang
et al. developed an accurate controllable
approach based on the bond-cleavage
chemistry combined with universal anti-
FITC CAR-T cells, allowing the regulation
of CAR-T cells in a switchable manner.
Highlights
d
d
d
d
Folate-O-FITC switch CAR-T cells to an activated state in a
dose-dependent manner
CAR-T-induced cytotoxicity can be rapidly terminated by the
cleavage of folate-O-FITC
CAR-T can be turned on/off on demand by supplementation/
cleavage of folate-O-FITC
Such switchable regulated manner would not compromise
antitumor effects of CAR-T cells
Zhang et al., 2021, Cell Chemical Biology 28, 1–10
January 21, 2021 ª 2020 Elsevier Ltd.
https://doi.org/10.1016/j.chembiol.2020.10.004
ll

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Photoswitchable CAR-T Cell Function
In Vitro and In Vivo via a Cleavable Mediator
Bo Zhang,1,2,3,6 Yan Wang,2,6 Shenlong Huang,4,6 Jiaqi Sun,2 Min Wang,3,5 Wenxiao Ma,2 Yanbo You,4 Ling Wu,2 Jin Hu,1,3
4,
Wei Song,1 Xudong Liu,1 Shengjie Li,1 Hua Chen,3,5 Guisheng Zhang,4 Lihe Zhang,2 Demin Zhou,2, Lingjun Li,
*
*
and Xuan Zhang3,5,7,
*
1Department of Medical Research Center, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union
Medical College, Beijing 100730, China
2State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University, Beijing 100191, China
3Clinical Immunology Center & Epigenetics Center, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing
100730, China
4Henan Key Laboratory of Organic Functional Molecule and Drug Innovation, School of Chemistry and Chemical Engineering Henan Normal
University, Xinxiang, Henan 453007, China
5Department of Rheumatology and Clinical Immunology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and
Peking Union Medical College, Beijing 100730, China
6These authors contributed equally
7Lead Contact
*Correspondence: deminzhou@bjmu.edu.cn (D.Z.), lingjunlee@aliyun.com (L.L.), zxpumch2003@sina.com (X.Z.)
https://doi.org/10.1016/j.chembiol.2020.10.004
SUMMARY
Chimeric antigen receptor (CAR)-T-based therapeutics are a breakthrough in cancer treatment; however,
they are hampered by constitutive activation, which leads to worrisome side effects. Engineering CAR-T cells
to be as tightly controllable as possible remains a topic of ongoing investigation. Here, we report a photo-
switchable approach that uses a mediator for the at-will regulation of CAR-T cells. This mediator carries
dual folate and fluorescein isothiocyanate moieties tethered by an ortho-nitrobenzyl ester photocleavable
linker. CAR-T cells were shown to be highly cytotoxic to targeted cells only in the presence of the mediator
and acted in a dose-dependent manner. The toxicity of CAR-T cells can be rapidly terminated by cleavage of
the mediator, and the effects of CAR-T cells can be activated again by resupplementation with the mediator
without compromising tumor therapy. The approach described here provides a direction for enhancing the
controllability of CAR-T cells and can likely be applied in other immunotherapies.
INTRODUCTION
on-target, on-tumor reactions against heavy tumor loads
(Acharya, et al., 2019; Brudno and Kochenderfer, 2016). Ironi-
Immunotherapy has emerged as a potential approach for the cally, CRS is considered an ‘‘on-target’’ effect, in that its pres-
treatment of cancer, with chimeric antigen receptor (CAR)-T cells ence demonstrates that active T cells are at work in the body
already in use in the clinic (Grupp, et al., 2013). The premise of (Brudno and Kochenderfer, 2016).
CAR-T cells is the genetic modification of T cells to express a
AlthoughhowCAR-Tcellsworkisunderstood,translationofthis
CAR for the recognition of tumor surface antigens specific to knowledge to the management of serious CRS is ongoing.
an individual patient. After infusion into a patient, CAR-T cells Numerous approaches have been implemented to address this
become activated upon contact with their targeted antigen and problem, such as neutralizing inflammatory cytokines by antago-
proceed to proliferate and destroy cancer cells. CAR-T cells nist proteins (Giavridis, et al., 2018; Kotch, et al., 2019; Norelli,
become cytotoxic through several mechanisms, including et al., 2018), inducing CAR-T cell death by incorporating suicide
extensive cell proliferation, increasing the degree to which the genes (Di Stasi, et al., 2011; Straathof, et al., 2005), and applying
cells are toxic to other living cells and, more importantly, causing bifunctional mediators to bridge CAR-T cells and target cells
the increased secretion of cellular factors, such as cytokines, in- (Cao, et al., 2016; Kim, et al., 2015; Tamada, et al., 2012; Urban-
terleukins, and growth factors (Barrett, et al., 2014; Grupp, et al., ska, et al., 2012). Despite the promise of these approaches,
2013). In the human body, a population of CAR-T cells expands achieving immediate and complete control of the toxicity of
by 100- to 10,000-fold in response to their antigen, making the CAR-T cells is still challenging considering the time-consuming
magnitude of their toxicity unpredictable and difficult to manage. procedures required for bispecific adaptors based on metabolism
These adverse reactions, including cytokine release syndrome (Lee, et al., 2019); furthermore, antibodies can only be locked
(CRS), are worrisome and sometimes fatal due to either on- to known cytokines, which may not be sufficiently comprehen-
target, off-tumor reactions against healthy tissue or excessive sive. As recent chemical developments have focused on
Cell Chemical Biology 28, 1–10, January 21, 2021 ª 2020 Elsevier Ltd.
1

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 1. Schematic Representation of a Photoswitchable CAR-T Cell System Based on a Cleavable Mediator
(A) Photoswitchable CAR-T cell system based on a cleavable mediator.
(B) Structure of the photocleavable mediator folate-O-FITC in which the FITC and folate moieties are tethered together via a photosensitive ortho-nitro-
benzyl ester.
‘‘bond-cleavage’’ reactions with excellent characteristics in the neko, et al., 2011; Menard-Moyon, et al., 2015; Sarfaty, et al.,
spatial and temporal regulation of biological functions (Li and 2018; Tan, et al., 2009; Tso, et al., 2016) (Figure S1A). In brief,
Chen, 2016; Oliveira, et al., 2020; Tomlin, et al., 2018; Wang, the ortho-nitrobenzyl ester bearing a terminal alkyne and free
et al., 2019), it is thus becoming possible to further enhance the amine group was obtained by propargylation, ester exchange,
controllability of CAR-T cells in a potentially switchable manner.
and ammonolysis from a commercial ortho-nitrobenzylic
As a proof of concept, here, we report a photoswitchable alcohol. Coupling of the ortho-nitrobenzylic ester to folate was
approach using the well-known mediator folate-fluorescein conducted with dicyclohexylcarbo-diimide as the condensing
isothiocyanate (FITC) (Figure 1A) (Kim, et al., 2015), but its reagent, and conjugation to FITC was then completed using
conventional tether was replaced with a photocleavable linker the CuAAC reaction. The structure and purity of folate-O-FITC
(Figure 1B) (Klan, et al., 2013) to render CAR-T cells as control- were verified by MALDI-TOF-MS and nuclear magnetic reso-
lable as possible. This system includes three key steps: (1) acti- nance (Figure S1B), and their photosensitivity was confirmed
vation of inert CAR-T cells through a mediator, which carries by UV365 irradiation.
folate and fluorescein isothiocyanate moieties tethered by an
To create a FITC-specific CAR, the anti-FITC scFv clone
ortho-nitrobenzyl linker (folate-O-FITC), for binding CAR-T cells 4M5.3 (Midelfort, et al., 2004; Vaughan, et al., 1996) was subcl-
and target cells, respectively; (2) termination of the toxic effects oned into a lentiviral vector (Figure S2A) and then transduced
of CAR-T cells via photoirradiation that rapidly breaks down the into human T cells as described previously (Kim, et al., 2015;
linker of folate-O-FITC; and (3) reactivation of CAR-T cells by Ma, et al., 2016). The transduction efficiency was ~90%, as
supplementation with the mediator as needed.
measured by flow cytometry (Figure S2B). Proper scFv folding
We find that folate-O-FITC has the ability to switch inert CAR-T on the T cell surface was verified by staining CAR-T cells with
cells to an activated state, and the persistence of the mediator anti-mouse IgG F(ab’)2-APC and the mediator (FITC) simulta-
forces CAR-T cells to remain in the active state. With exposure neously. We observed that nearly all T cells expressing the
to light pulses at nondetrimental wavelengths, CAR-T cells are scFv bound FITC (~50%), which was a significantly higher level
switched off following breakdown of the mediator and can be of binding than that observed in nontransduced T cells (Fig-
switched on again by resupplementation with the mediator. ure S2C), indicating that the scFv expressed on the cellular sur-
The versatile approach described here may provide a direction face was solvent exposed and accessible to the FITC ligand.
for enhancing the controllability of CAR-T cells and can likely
be applied in other immunotherapies.
Activation of Anti-FITC CAR-T Cells by Folate-O-FITC
In Vitro
RESULTS
Next, we evaluated the ability of folate-O-FITC to direct inert anti-
FITC CAR-T cells to bind to folate receptor (FR)-positive KB cells
using a cytotoxicity assay with FR-negative A549 cells as a
negative control (Kim, et al., 2015). We found that either at an ef-
Generation and Characterization of Folate-O-FITC
Mediator and Anti-FITC CAR-T Cells
We first synthesized the folate-O-FITC mediator based on a fector:target cell (E:T) ratio of 3:1 or 10:1, folate-O-FITC demon-
modular assembly (Chang, et al., 2016; Jung, et al., 2010; Ka- strated high potency in directing anti-FITC CAR-T cells to kill KB
2
Cell Chemical Biology 28, 1–10, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 2. Characterization of Folate-O-
FITC-Mediated Activation of Anti-FITC
CAR-T Cells In Vitro
(A) Dose-dependent activation of CAR-T cells by
folate-O-FITC. The cytotoxicity of CAR-T cells
induced by the mediator were measured by incu-
bation of anti-FITC CAR-T cells and FR-positive
KB cells at a ratio of either 3:1 or 10:1, and the
amount of LDH released into the culture medium
detected. The KB cells cultured with non-
transduced
T cells and the FR-negative A549
cultured group were applied as negative control.
The percentage cytotoxicity was defined as the
percentage of (experimental – effector sponta-
neous – target spontaneous)/(target maximum –
target spontaneous), where ‘‘effector’’ means
CAR-T cells or nontransduced control T cells, and
‘‘target’’ means KB or A549 cells, respectively. The
data are presented as the mean SEM derived
from triplicate samples (n = 3).
(B) Quantification of cytokine (IFN-g and IL-2)
levels in the cultures described in (A) by ELISA.
Data are presented as the mean SEM derived
from triplicate samples (n = 3).
(C) Characterization of folate-O-FITC-mediated
activation of CAR-T cells by flow cytometry anal-
ysis of CD25 and CD69 expression on CAR-T cells.
CAR-T cells were cocultured with KB cells either in
the presence or absence of the mediator for 24 h
with nontransduced T cells as a control.
(D) Morphological features of folate-O-FITC-
dependent CAR-T cell aggregates. All images
were acquired from the center field of each well
and are shown at 1003 magnification. For (C) and
(D), representative results from one of three ex-
periments are shown.
cells over 24 h (half maximal effective concentration = 109.5
to an activated state, and persistence of the mediator in the me-
28.86 or 93.2 12.65 pM for E:T = 3:1 or 10:1, respectively, Fig- dium forces CAR-T cells to remain in this activated state.
ure 2A). In contrast, the mediator failed to induce the lysis of KB
cells by nontransduced T cells. No significant cytotoxicity was Inactivation of CAR-T Cells by Photoirradiation and
observed with A549 cells and either anti-FITC CAR-T cells or Reactivation by Supplementation with Folate-O-FITC
nontransduced T cells, demonstrating the high selectivity of To explore whether the activated anti-FITC CAR-T cells could be
folate-O-FITC for the activation of anti-FITC CAR-T cells. silenced by photoirradiation, we first tested the capacity of irra-
Consistent with these results, increased secretion of cellular fac- diation to promote the detachment of CAR-T cells from KB cells.
tors was detected during this 24-h assay, and the levels of the As described above, anti-CD3-APC-labeled CAR-T cells and
cytokines, interferon g (IFN-g) and interleukin-2 (IL-2), were posi- anti-FR-PE-labeled KB cells were cocultured at a 1:1 ratio for
tively correlated with the concentration of folate-O-FITC only in 1 h at 4ꢀC. The percentage of PE/APC double-positive cells
the case of FR-positive cells (Figure 2B). The binding affinity was significantly greater in the presence of the mediator, as
and cytotoxicity of the mediator were also compared with those shown in Figure 3A, than in the absence of the mediator. We
of its parental folate-FITC; nonsignificant differences were then irradiated half of the experimental cells at 365 nm, with
observed, indicating that the installed nitrobenzyl linker has little the remaining cells serving as a negative control. We found
effect on these functions (Figures S3A and S3B).
that, upon irradiation for 5 min, the percentage of double-posi-
The activation of CAR-T cells by the mediator was further tive cells sharply decreased to the level in the nonirradiated
tested by monitoring the activation markers and, as anticipated, group, suggesting that the ortho-nitrobenzyl linker within the
folate-O-FITC significantly increased the CD25 and CD69 mediator was effectively broken. A parallel experiment by detec-
expression levels of anti-FITC CAR-T cells compared with those tion of photoirradiated Jurkat or KB cells labeled with corre-
of nontransduced T cells only in the presence of FR-positive cells sponding fluorescent antibody was also performed, and no
(Figure 2C). In addition, the selective formation of aggregates, obvious decrease of fluorescent intensity was found even
indicative of crosslinking between T cells and target cells, was when the duration of photoirradiation was up to 15 min, suggest-
visible only in cocultures of anti-FITC CAR-T and KB cells with ing that the decrease of the double-positive cell was indeed
the mediator but not in any of the controls (Figure 2D). Clearly, ascribed to the detachment of Jurkat and KB cells (Figure S4A).
supplementation with the mediator switches inert CAR-T cells Notably, analysis of cell clusters through fluorescence and
Cell Chemical Biology 28, 1–10, January 21, 2021
3

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 3. Characterization of Photoirradia-
tion-Induced Detachment of CAR-T/KB
Cell Clusters through Fluorescence Imaging
(A) Folate-O-FITC-mediated attachment and pho-
toswitchable detachment of CAR-T cells and KB
cells analyzed by flow cytometry in triplicate ex-
periments. The data are presented as the mean
SEM of triplicate experiments (n = 3), and the p
values shown were determined by two-tailed, un-
paired Student’s t tests; ***p % 0.001.
(B) Representative confocal fluorescence images
of cell clusters with PE and APC double-staining
obtained by flow cytometry, as shown in (A). All
images were acquired with
a 403 objective,
analyzed by IDEAS software, and illustrated in a
channel series. The channels used for imaging
were as follows: bright-field (BF, channel 1), PE
(anti-FR-PE for KB cells, channel 2), and APC (anti-
CD3-APC for CAR-T cells, channel 3). Scale
bar, 10 mm.
ation, and mediator resupplementation.
Therefore, we concluded that CAR-T
cells were able to shuttle between an
bright-field imaging showed that the cell clusters in the presence activated and inert state by folate-O-FITC acting as a switch
of the mediator were mostly composed of both KB and CAR-T and a UV365 pulse acting as a regulator.
cells at a ratio of 1:1 or 1:2, whereas the nonspecific self-adhe-
sion of CAR-T cells and KB cells was predominant in the nonirra- Spatial and Temporal Photoswitching of CAR-T Cells in
diated groups (Figure 3B). Optimization of the irradiation time FR-Positive Xenograft Tumor Models
indicated an increased degree of detachment with increasing Given the feasibility of this system in vitro, we next explored the
irradiation time up to 10 min for CAR-T cells and up to 20 min possibility of the spatial and temporal photoswitching of CAR-T
for KB cells (Figure S4B). No significant decrease in cell viability cells in vivo. We first examined the turn-on effect of folate-O-
was observed for either cell line, even with irradiation for as long FITC using a diffuse SKOV3 tumor model (Song, et al., 2011).
as 30 min (Figure S4C).
We found that SKOV3 is an FR-positive cell line that showed a
Upon validating the feasibility of the mediator-induced weaker cytotoxic response to CAR-T cells than KB cells (Figures
detachment of CAR-T cells and target cells, we then designed 5A and 5B), but could be easily transplanted into mice via intra-
a time course experiment (Figure 4A) to monitor the mediator- venous injection. As anticipated, mice treated with high doses of
induced in vitro cytotoxicity and photoirradiation-mediated the mediator exhibited significantly inhibited tumor proliferation
regulation. The coculture of CAR-T cells with KB cells in the compared with mice treated with a low dose of the mediator
presence of folate-O-FITC for 4 h led to marginal but detect- and mice not treated with the mediator (Figure 5C), indicating
able cytotoxicity, accompanied by mild IL-2 and IFN-g release that the in vivo activation of CAR-T cells was successfully initi-
(Figure 4B). Then, some of the cultured cells were given a pulse ated by the mediator in a dose-dependent manner.
of photoirradiation that led to an immediate attenuation of
We next investigated the effects of the switchable regulation of
cytotoxicity and a decrease in the release of IL-2 and IFN-g CAR-T cells in vivo. To this end, we first evaluated the pharma-
compared with the nonirradiated group. The nonirradiated cokinetics of the mediator after subcutaneous or intratumoral in-
group showed a continuous increase in cytotoxicity and IL-2 jection in healthy or KB tumor-bearing mice, respectively. The
and IFN-g release, which plateaued overnight (Figure 4B). Sub- mediator was present in the circulation with pharmacokinetic
sequently, the resting CAR-T cells become reactivated upon properties, i.e., a half-life (T1/2) of ~3 h and a mean retention
resupplementation with folate-O-FITC, as manifested by the time of ~4 h, comparable with those of its parental folate-FITC
significant release of IL-2 and IFN-g. Impressively, the cytotox- (Figures S5A and S5B), and it remained in the local tumor tissue
icity of the CAR-T cells even exceeded that of the nonirradi- for at least 12 h after a single intratumoral injection (Figure S5C).
ated cells 4 h after resupplementation, accompanied by a Moreover, the accessibility of percutaneous UV irradiation to the
synchronous increase in IL-2 release that surpassed that in tumor burden was verified by ex vivo flow cytometry analyses
the nonirradiated group (Figure 4B). Concomitant with this (Figure S5D). These results suggest the feasibility and availability
progress was the formation, dispersion, and reformation of of evaluating the mediator in a tumor-bearing mouse model
cellular aggregates (Figure 4C), and phenotypical changes in vivo.
serving as indicators of CAR-T cell activation, deactivation,
We then intratumorally injected KB tumor-bearing SCID/beige
and reactivation (Figure 4D), respectively. These observations mice with an adequate number of CAR-T cells on day 0, followed
correlated well with folate-O-FITC supplementation, UV irradi- by alternate-day dosing with the mediator. Inactivation was
4
Cell Chemical Biology 28, 1–10, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 4. Time Course of CAR-T Cell Activa-
tion by Folate-O-FITC, Inactivation by Pho-
toirradiation, and Reactivation by Folate-O-
FITC Supplementation
(A) Flow chart of the in vitro analysis. CAR-T cells
and KB cells were cocultured with the mediator at
0 h, and UV irradiation and reactivation were con-
ducted at 4 and 21 h, respectively. The UVꢁ group
received the mediator at 0 h and no further treat-
ment beyond that time; the UV+ group received the
mediator at 0 h, UV irradiation at 4 h, and no further
treatment beyond that time; and the UV+/M+
group received the mediator at 0 h, UV irradiation
at 4 h, and another dose of the mediator at 21 h.
(B) Quantification of LDH and cytokine (IFN-g and
IL-2) levels over time as described in (A). Data are
presented as the mean SEM derived from tripli-
cate samples (n = 3). The p values shown were
determined by two-tailed, unpaired Student’s
t tests, *p % 0.05, **p % 0.01, ***p % 0.001, ****p %
0.0001.
(C) Microscopic images of CAR-T cells, including
after activation by folate-O-FITC, inactivation by
photoirradiation, and reactivation by folate-O-FITC
supplementation, as described in (A and B). All
images were acquired from the center field of each
well and are shown at 1003 magnification.
(D) Characterization of folate-O-FITC-mediated
switchable regulation of CAR-T cells by flow cy-
tometry analysis of CD25 and CD69 expression.
CAR-T cells were cocultured with KB cells in the
presence of the mediator and subsequently
stained with anti-mouse-IgG F(ab’)2 antibody fol-
lowed by anti-CD25 and anti-CD69 antibodies at
the indicated times, as described in (C) The results
represent findings from three independent exper-
iments.
achieved by UV photoirradiation 6 h after each mediator injection toxicity of CAR-T cells can be regulated by transient cleavage
on days 6, 8, and 10, and changes in both body weight and tumor of the mediator in a switchable manner without compromising
growth were monitored for 2 weeks (Figure 6A). According to our antitumor activities.
results, the treatment of KB tumor-bearing mice with an
adequate dose of CAR-T cells plus uninterrupted alternate-day DISCUSSION
doses of the mediator promoted rapid and continuous weight
loss, suggesting the occurrence of cytokine-induced toxicity, CAR-T cell-based therapeutic approaches have improved in
and UV irradiation rapidly reversed such weight loss compared recent years but are still associated with significant safety and
with those in the uninterrupted adapter-treated group. The controllability issues. Numerous approaches have been imple-
continuing body weight loss was observed again upon resupple- mented to address these issues, such as neutralizing inflamma-
mentation with the mediator (Figure 6B), implying the reactiva- tory cytokines by antibodies, administering immunosuppressive
tion of CAR-T cells. These weight loss changes were correlated drugs, and inducing CAR-T cell death via the suicide gene
well with changes in cytokine production, as demonstrated by (Acharya, et al., 2019; Brentjens, et al., 2013; Lee, et al., 2014;
the related changes in the IFN-g and IL-2 levels (Figure 6C). Straathof, et al., 2005). Because all of these methods require
Notably, with the controllable cytokine secretion and rapid the suppression of CAR-T activity, it is thus necessary to balance
decrement in body weight loss, the integral antitumor effects the benefits of suppressing CRS and inhibiting CAR-T cells in
were not compromised by this temporary switchable regulation clinical applications. As the application of antagonist proteins,
(Figure 6D). These results are consistent with those of a parallel such as tocilizumab and anakinra, for blocking IL-6R and IL-1R
bioluminescence imaging (BLI) assay, in which photoirradiation (Giavridis, et al., 2018; Kotch, et al., 2019; Norelli, et al., 2018),
treatment (Mediator+/UV+) led to obvious tumor growth relapse respectively, can only be locked to known cytokines, the effect
compared with mediator-only treatment (Mediator+/UVꢁ) on may not be sufficiently comprehensive to suppress CRS. Aiming
day 10, while CAR-T cell reactivation and tumor inhibition were to develop a feasible strategy to further enhance the controlla-
achieved again by reinjection of the mediator (Figures 6E–6G). bility and safety of CAR-T cell technology, as a proof of concept,
Collectively, these results suggest that the cytokine-induced we report an approach using a photocleavable mediator and a
Cell Chemical Biology 28, 1–10, January 21, 2021
5

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 5. Folate-O-FITC-Mediated Activa-
tion of CAR-T Cells in Diffuse SKOV3 Tumor
Xenografts
(A) Comparison of folate receptor expression on
A549, KB, and SKOV3 cells using anti-FR-PE an-
tibodies (top) and the mediator (bottom), respec-
tively. Cells were first incubated with 10 nM
mediator and then stained with anti-FITC-APC
antibodies. The results shown represent the find-
ings from three experiments.
(B) In vitro comparison of anti-FITC CAR-T cell
responses with KB and SKOV3 cells. Anti-FITC
CAR-T cells were cocultured for 24 h with KB or
SKOV3 cells in the presence of 1 nM mediator at an
E:T ratio of 20:1. Cells cultured without the medi-
ator were used as a negative control. Cytotoxicity
(top), IL-2 (middle), and IFN-g (bottom) levels were
measured and calculated as percentages. Data
are presented as the mean SEM derived from
triplicate samples (n = 3).
(C) Time course of folate-O-FITC-mediated acti-
vation of CAR-T cells in diffuse SKOV3 tumor xe-
nografts. Xenografts were monitored and quanti-
tatively analyzed as regions of interest by BLI. NOG
mice inoculated with luciferase-expressing SKOV3
cells were infused with 1 3 107 CAR-T cells; 7 days
later, folate-O-FITC treatment, performed every
other day, was executed six times at the indicated
doses. Mice treated with high doses of the medi-
ator achieved significant tumor growth inhibition.
Data are presented as the mean SEM derived
from three mice per group (n = 3), the p values
shown were determined by one-way ANOVA
(Dunnett’s multiple comparison test compared
with mediator (ꢁ) treated group). *p % 0.05, **p %
0.01. The experiment was repeated three times
with similar results, and one of three representative
results is shown.
nondetrimental laser pulse that allows the regulation of CAR-T convert to a resting state immediately upon exposure to a
cells in a switchable manner at will, limiting the possible adverse photoirradiation pulse because significantly lower levels of
effects without comprising the antitumor effects. Folate-FITC, cytokines, such as IL-2 and IFN-g, lasted for some time
with its conventional linker having been replaced with a UV-sen- compared with those in the nonirradiated group. The persistent
sitive ortho-nitrobenzyl ester group, was adopted here because cytokines released into medium may also contribute to the
of its well-studied and successful application in the clinic (Lu, release of LDH.
et al., 2009; Lu, et al., 2006; van Dam, et al., 2011).
Whether this approach had the capacity to switchable regula-
We found that CAR-T cells were indeed highly cytotoxic to FR- tion of CAR-T cells was evaluated using a xenograft model
positive cells only in the presence of the mediator and exhibited in vivo. KB cells were used because they express high levels of
cytotoxic effects in a dose-dependent manner both in vitro and FR and can be easily implanted as solid tumors. According to
in vivo, verifying the specificity of the mediator for CAR-T cell our results, a body weight loss of ~10% was observed in tu-
activation. Furthermore, the cytotoxicity and cytokines release mor-bearing mice within 10 days of a single injection of an
of CAR-T cells were rapidly inhibited upon photoirradiation, adequate dose of CAR-T cells plus uninterrupted alternate-day
and reactivation of inactive CAR-T cells can be achieved by re- doses of the mediator, suggesting the occurrence of CRS (Lee,
supplementation of mediator reversibly, addressing the inherent et al., 2019; Ma, et al., 2016). The potent decrement in cyto-
issues regarding the rapid and reversible control of CAR-T cells. kine-induced toxicity was rapidly promoted by simple UV irradi-
Nevertheless, as in vitro assessment demonstrated in Figure 4, ation, which broke the link between CAR-T cells and target cells,
we found that the effects of UV-mediated detachment on cyto- and CAR-T cell reactivation was achieved by supplementation
toxicity of CAR-T cells are compromised compared with those with the mediator, enabling switchable regulation of the cyto-
of cytokines release of CAR-T cells. The continuous persistence kine-induced toxicity of CAR-T cells. Although this demon-
of cytotoxicity reflected by lactate dehydrogenase (LDH) might strated that the antitumor activity of CAR-T cells has not been
be explained by several factors, as LDH release might continue significantly affected by such switchable regulation, the contrast
for a while, depending on renewal of the damaged plasma mem- of the tumor suppression effect of the CAR-T cells between "with
brane, even when CAR-T cells are inactivated by photoirradia- mediator" and "without mediator" groups in Figure 6D is, to a
tion. In addition, activated CAR-T cells might not be able to lesser degree than that of previously reports using other
6
Cell Chemical Biology 28, 1–10, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Figure 6. Verification of Photoswitchable CAR-T Cells in KB Tumor Xenograft Models
(A–F) Characterization of switchable regulation of CAR-T cell-mediated cytotoxicity in vivo. (A) Flow chart of the exploration of the in vivo activation and inac-
tivation behavior of folate-O-FITC and photoirradiation, respectively. The KB tumor-bearing SCID/beige mice were intratumorally injected with 1.5 3 107 anti-
FITC CAR-T cells on day 0, followed by alternate-day dosing with the mediator (500 nmol/kg). Changes in body weight and tumor growth were monitored
alternately over 2 weeks. (B) Changes in body weight after administration of CAR-T cells in the absence (Mꢁ) or presence of the mediator (M+/UVꢁ). Inactivation
was achieved by UV photoirradiation 6 h after each mediator injection on days 6, 8, and 10 (M+/UV+), and reactivation was achieved by replenishing of mediator
4 h after each photoirradiation, respectively, as demonstrated in (A). (C) Analysis of IFN-g and IL-2 secretion by CAR-T cells in plasma by ELISA on day 8. (D) The
tumor growth after administration of CAR-T and mediator. For (B)–(D), data are presented as the mean SEM derived from five mice per group (n = 5), and the p
values shown were determined by two-tailed, unpaired Student’s t tests. Representative data from one of three experiments are shown and similar results were
observed; (E) bioluminescence imaging and quantitative analyses of the region of interest (F) in mice inoculated with luciferase-transduced KB cells.
(G) Measurement of body weight in V at the end of the study period (day 18). For (F) and (G), data are presented as the mean SEM of three mice per group (n = 3),
and p values shown were determined by two-tailed, unpaired Student’s t tests; *p % 0.05, **p % 0.01, ***p % 0.001.
FR-positive cells, such as MDA-MB-231 cells (Lee, et al., 2019), of KB cells are beneficial to the rapid assessment of the in vivo
and all groups of mice eventually succumbed to excessive tumor behavior of CAR-T cells, but also inevitably compromised their
burden (data not show). The rapid growth and easy implantation antitumor effects.
Cell Chemical Biology 28, 1–10, January 21, 2021
7

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
The use of bispecific adaptors to control CAR-T cell-derived switch. With the use of a light pulse, the mediator-based on-
toxicity has also been studied by other groups, and promising re- and-off switching of CAR-T cells at will may limit side effects
sults have been generated (Kim, et al., 2015; Lee, et al., 2019; without compromising antitumor effects, providing an attractive
Ma, et al., 2016). Achieving rapid and flexible control of the immunotherapeutic approach for overcoming some of the safety
toxicity of CAR-T cells, however, is still challenging due to the hurdles associated with current CAR-T cell-based therapies.
time-consuming procedures required for these adaptors, which Furthermore, with replacement of the folate moiety in the medi-
generally include elaborate concentration and dosing schedules ator with other ligands with a specific cell or tissue tropism, the
based on the metabolism and pharmacokinetics of the mediator photoswitchable CAR-T cells reported here may become a
(Lee, et al., 2019). The switchable CAR-T cells described here, universal system that can be redirected against theoretically
which functioned intermittently under irradiation 6 h after each any antigen on the tumor surface. We believe that such a
application of the mediator and were reactivated by resupple- photoswitchable CAR-T cell system provides a direction for
mentation with the mediator 4 h after each irradiation, as controllable immunotherapy and helps to increase the safety of
described in Figure 6A, created a 44-h ‘‘active’’ duration and a CAR-T cells for future clinical applications.
4-h ‘‘inactive’’ pulse within the 2-day interval; the effect in terms
of overall antitumor activities was comparable with that of unin-
terrupted alternate-day dosing but was more beneficial in terms
SIGNIFICANCE
of tuning the toxicity of the CAR-T cells on demand. Given the
high efficiency of UV-mediated cleavage, the active and inactive
CAR-T is a powerful technology for cancer therapy, but is
significantly limited by the safety issues because of the
durations can be adjusted at will, regardless of the concentration
and half-life of the mediator, which is a benefit of this approach.
inability to flexibly control and terminate their activations.
To this end, we developed an accurate and controllable
Nevertheless, cutting back on the dosage of the mediator but
increasing the dosing frequency during the same time interval
approach based on a cleavable on-off mediator, where the
bond-cleavage chemistry, i.e., photocleavable reaction,
may have a similar effect but is beyond the scope of this study.
The versatile approach described here provides an appealing
was combined with universal anti-FITC CAR-T cells. The
activation of CAR-T cells strictly depended on the presence
direction for clinical applications of photoswitchable CAR-T
cells. There are, however, some limitations. One limitation is
of the mediator and acted in a dose-dependent manner;
also, the CAR-T-induced cytotoxicity can be rapidly termi-
the wavelength of light: ~365 nm used in this study to perform
photoswitching. The best wavelengths to trigger mediator cleav-
nated by the cleavage of the mediator, and the effects of
CAR-T cells can be reactivated by the supplementation of
age are in the UV wavelength region, which is not medically use-
ful due to its poor tissue penetration (less than a few millimeters)
the mediator without compromising therapeutic effects.
This switchable approach was shown to be highly effective
and its ability to induce DNA damage and mutations (Li and
Chen, 2016). A series of photocleavable groups have emerged
in the regulation of CAR-T cells both in vitro and in vivo.
We believe that such a photoswitchable CAR-T system
over the last decade to regulate bioactive molecules, and red
and infrared light within wavelengths of ~650–950 nm may be
would pave a path for controllable immunotherapy and
help to increase the safety of CAR-T for future clinical
optional choices because of less harm to living cells and the sig-
nificant possibility for deeper tissue penetration (Papageorgiou,
applications.
et al., 2000; Shanmugam, et al., 2016; Stolik, et al., 2000). Opti-
mizations of the switchable parts, including but not limited to
STAR+METHODS
photocleavable groups, remain a topic of ongoing investigation,
and become crucial to ensure this approach for clinical transfor-
Detailed methods are provided in the online version of this paper
mation. A variety of bioorthogonal cleavage reactions, including
and include the following:
light-, chemical-, and enzyme-triggered reactions, have recently
emerged (Li and Chen, 2016). The UV-photosensitive ortho-ni-
trobenzyl ester was selected as a linker in this study due to its
high cleavability and rapid cleavage under biocompatible photo-
irradiation at 365 nm (Klan, et al., 2013; Li and Chen, 2016); these
properties afford immediate and efficient detachment at the
living cell level as needed. Given that the cleavage of mediators
is not damaging to CAR-T cells, our results indicate that the con-
trol of CAR-T cells using a cleavable mediator is effective for
reversing CRS without inhibiting the efficacy of CAR-T cell treat-
ment, enabling the regulation of CAR-T cells in a switchable
manner.
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITY
B Lead Contact
B Materials Availability
B Data and Code Availability
d EXPERIMENTAL MODELS AND SUBJECT DETAILS
B Animals
B Mammalian Cell Culture
d METHOD DETAILS
B Detailed Synthetic Procedures to Folate-O-FITC (Com-
pound 10)
In conclusion, we report a platform that provides an instanta-
neous and recyclable methodology to enable the regulation of
CAR-T cell activities in a flexible, on-and-off switchable manner.
The spatial and temporal control of the toxicity of CAR-T cells un-
der physiological conditions indicates the excellent biocompat-
ibility of using such a triple-functional mediator as a molecular
B Compound 1 1,4-Diazidobutane
B Compound 2 4-azidobutan-1-amine
B Compound 3 5-(3-(4-azidobutyl)thioureido)-2-(3,6-di-
hydroxy-9H- xanthen-9-yl) benzoic acid
B Compound
phenyl) ethan-1-one
4
1-(3-methoxy-4-(prop-2-yn-1-yloxy)
8
Cell Chemical Biology 28, 1–10, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
B Compound
yloxy) phenyl) ethan-1-one
B Compound 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-
yloxy) phenyl) ethan-1-ol
B Compound 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-
yloxy) phenyl) ethyl (4-nitro phenyl) carbonate
B Compound 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-
5
1-(5-methoxy-2-nitro-4-(prop-2-yn-1-
Cao, Y., Rodgers, D.T., Du, J., Ahmad, I., Hampton, E.N., Ma, J.S., Mazagova,
M., Choi, S.H., Yun, H.Y., Xiao, H., et al. (2016). Design of switchable chimeric
antigen receptor T cells targeting breast cancer. Angew. Chem. Int. Ed. 55,
7520–7524.
6
Chang, T.C., Adak, A.K., Lin, T.W., Li, P.J., Chen, Y.J., Lai, C.H., Liang, C.F.,
Chen, Y.J., and Lin, C.C. (2016). A photo-cleavable biotin affinity tag for the
facile release of a photo-crosslinked carbohydrate-binding protein. Bioorg.
Med. Chem. 24, 1216–1224.
7
8
yloxy) phenyl) ethyl (2-amino ethyl) carbamate
B Virus Production and Generation of CAR-Ts
B Evaluation of Transduction Efficiency and Function of
CAR-Ts by Flow Cytometry Analysis
van Dam, G.M., Themelis, G., Crane, L.M., Harlaar, N.J., Pleijhuis, R.G.,
Kelder, W., Sarantopoulos, A., de Jong, J.S., Arts, H.J., van der Zee, A.G.,
et al. (2011). Intraoperative tumor-specific fluorescence imaging in ovarian
cancer by folate receptor-alpha targeting: first in-human results. Nat. Med.
17, 1315–1319.
B Cytotoxicity and Cytokine Release Assay
B Photocleavage Experiments at the Cellular Level
B In Vivo Regulated Efficacy Assays
Giavridis, T., van der Stegen, S.J.C., Eyquem, J., Hamieh, M., Piersigilli, A., and
Sadelain, M. (2018). CAR T cell-induced cytokine release syndrome is medi-
ated by macrophages and abated by IL-1 blockade. Nat. Med. 24, 731–738.
d QUANTIFICATION AND STATISTICAL ANALYSIS
Grupp, S.A., Kalos, M., Barrett, D., Aplenc, R., Porter, D.L., Rheingold, S.R.,
Teachey, D.T., Chew, A., Hauck, B., Wright, J.F., et al. (2013). Chimeric antigen
receptor-modified T cells for acute lymphoid leukemia. New Engl. J. Med. 368,
1509–1518.
SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.
chembiol.2020.10.004.
Jung, M.E., Ouk, S., Yoo, D., Sawyers, C.L., Chen, C., Tran, C., and
Wongvipat, J. (2010). Structure-activity relationship for thiohydantoin
androgen receptor antagonists for castration-resistant prostate cancer
(CRPC). J. Med. Chem. 53, 2779–2796.
ACKNOWLEDGMENTS
Kaneko, S., Nakayama, H., Yoshino, Y., Fushimi, D., Yamaguchi, K., Horiike,
Y., and Nakanishi, J. (2011). Photocontrol of cell adhesion on amino-bearing
surfaces by reversible conjugation of poly(ethylene glycol) via a photocleav-
able linker. Phys. Chem. Chem. Phys. 13, 4051–4059.
We thank Mr. Sen Dou for his help in generating graphics. This work was sup-
ported by grants from the National Natural Science Foundation of China
(81788101, 81630044, 81803419, 91753202, 21778015 and 81821004), the
National Key Research and Development Program of China (2019ZX09739),
Chinese Academy of Medical Science Innovation Fund for Medical Sciences
(CIFMS 2017-12M-1-008，2016-12M-1-003), and the PUMC Youth Fund
(2017350001). All authors declare no competing financial interests. Additional
data can be found in the supplementary materials.
Kim, M.S., Ma, J.S., Yun, H., Cao, Y., Kim, J.Y., Chi, V., Wang, D., Woods, A.,
Sherwood, L., Caballero, D., et al. (2015). Redirection of genetically engi-
neered CAR-T cells using bifunctional small molecules. J. Am. Chem. Soc.
137, 2832–2835.
Klan, P., Solomek, T., Bochet, C.G., Blanc, A., Givens, R., Rubina, M., Popik,
V., Kostikov, A., and Wirz, J. (2013). Photoremovable protecting groups in
chemistry and biology: reaction mechanisms and efficacy. Chem. Rev. 113,
119–191.
AUTHOR CONTRIBUTIONS
B.Z., D.Z., and L.L. designed the study. B.Z., Y.W., S.H., Y.Y., and J.S. per-
formed the experiments and analyzed the data. J.H., W.L., H.C., and W.S.
helped to polish the language. W.M., S.L., and X.L. donated blood and contrib-
uted to the collection and assembly of data, data analysis, and interpretation.
M.W., G.Z., and L.Z. provided comments. B.Z., X.Z., and D.Z. wrote the manu-
script. All authors reviewed the manuscript.
Kotch, C., Barrett, D., and Teachey, D.T. (2019). Tocilizumab for the treatment
of chimeric antigen receptor T cell-induced cytokine release syndrome. Expert
Rev. Clin. Immunol. 15, 813–822.
Lee, D.W., Gardner, R., Porter, D.L., Louis, C.U., Ahmed, N., Jensen, M.,
Grupp, S.A., and Mackall, C.L. (2014). Current concepts in the diagnosis and
management of cytokine release syndrome. Blood 124, 188–195.
DECLARATION OF INTERESTS
Lee, Y.G., Chu, H., Lu, Y., Leamon, C.P., Srinivasarao, M., Putt, K.S., and
Low, P.S. (2019). Regulation of CAR T cell-mediated cytokine release syn-
drome-like toxicity using low molecular weight adapters. Nat. Commun.
10, 2681.
The authors declare no competing interests.
Received: August 7, 2020
Revised: September 7, 2020
Accepted: October 7, 2020
Published: October 27, 2020
Li, J., and Chen, P.R. (2016). Development and application of bond cleavage
reactions in bioorthogonal chemistry. Nat. Chem. Biol. 12, 129–137.
Lu, Y., Xu, L.C., Parker, N., Westrick, E., Reddy, J.A., Vetzel, M., Low, P.S., and
Leamon, C.P. (2006). Preclinical pharmacokinetics, tissue distribution, and
antitumor activity of a folate-hapten conjugate-targeted immunotherapy in
hapten-immunized mice. Mol. Cancer Ther. 5, 3258–3267.
REFERENCES
Acharya, U.H., Dhawale, T., Yun, S., Jacobson, C.A., Chavez, J.C., Ramos,
J.D., Appelbaum, J., and Maloney, D.G. (2019). Management of cytokine
release syndrome and neurotoxicity in chimeric antigen receptor (CAR) T cell
therapy. Expert Rev. Hematol. 12, 195–205.
Lu, Y., Klein, P.J., Westrick, E., Xu, L.C., Santhapuram, H.K., Bloomfield, A.,
Howard, S.J., Vlahov, I.R., Ellis, P.R., Low, P.S., et al. (2009). Strategy to pre-
vent drug-related hypersensitivity in folate-targeted hapten immunotherapy of
cancer. AAPS J. 11, 628–638.
Barrett, D.M., Singh, N., Porter, D.L., Grupp, S.A., and June, C.H. (2014).
Chimeric antigen receptor therapy for cancer. Annu. Rev. Med. 65, 333–347.
Ma, J.S., Kim, J.Y., Kazane, S.A., Choi, S.H., Yun, H.Y., Kim, M.S., Rodgers,
D.T., Pugh, H.M., Singer, O., Sun, S.B., et al. (2016). Versatile strategy for con-
trolling the specificity and activity of engineered T cells. Proc. Natl. Acad. Sci.
U S A 113, E450–E458.
Brentjens, R.J., Davila, M.L., Riviere, I., Park, J., Wang, X., Cowell, L.G.,
Bartido, S., Stefanski, J., Taylor, C., Olszewska, M., et al. (2013). CD19-tar-
geted T cells rapidly induce molecular remissions in adults with chemo-
therapy-refractory acute lymphoblastic leukemia. Sci. Transl. Med. 5,
177ra138.
Menard-Moyon, C., Ali-Boucetta, H., Fabbro, C., Chaloin, O., Kostarelos, K.,
and Bianco, A. (2015). Controlled chemical derivatisation of carbon nanotubes
with imaging, targeting, and therapeutic capabilities. Chemistry 21,
14886–14892.
Brudno, J.N., and Kochenderfer, J.N. (2016). Toxicities of chimeric antigen re-
ceptor T cells: recognition and management. Blood 127, 3321–3330.
Cell Chemical Biology 28, 1–10, January 21, 2021
9

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Midelfort, K.S., Hernandez, H.H., Lippow, S.M., Tidor, B., Drennan, C.L., and
Wittrup, K.D. (2004). Substantial energetic improvement with minimal struc-
tural perturbation in a high affinity mutant antibody. J. Mol. Biol. 343, 685–701.
Stolik, S., Delgado, J.A., Perez, A., and Anasagasti, L. (2000). Measurement of
the penetration depths of red and near infrared light in human "ex vivo" tissues.
J. Photochem. Photobiol. B Biol. 57, 90–93.
Straathof, K.C., Pule, M.A., Yotnda, P., Dotti, G., Vanin, E.F., Brenner, M.K.,
Heslop, H.E., Spencer, D.M., and Rooney, C.M. (2005). An inducible caspase
9 safety switch for T-cell therapy. Blood 105, 4247–4254.
Norelli, M., Camisa, B., Barbiera, G., Falcone, L., Purevdorj, A., Genua, M.,
Sanvito, F., Ponzoni, M., Doglioni, C., Cristofori, P., et al. (2018). Monocyte-
derived IL-1 and IL-6 are differentially required for cytokine-release syndrome
and neurotoxicity due to CAR T cells. Nat. Med. 24, 739–748.
Tamada, K., Geng, D., Sakoda, Y., Bansal, N., Srivastava, R., Li, Z., and Davila,
E. (2012). Redirecting gene-modified T cells toward various cancer types using
tagged antibodies. Clin. Cancer Res. 18, 6436–6445.
Oliveira, B.L., Stenton, B.J., Unnikrishnan, V.B., de Almeida, C.R., Conde, J.,
Negrao, M., Schneider, F.S.S., Cordeiro, C., Ferreira, M.G., Caramori, G.F.,
et al. (2020). Platinum-triggered bond-cleavage of pentynoyl amide and
N-propargyl handles for drug-activation. J. Am. Chem. Soc. 142,
10869–10880.
Tan, L.P., Wu, H., Yang, P.Y., Kalesh, K.A., Zhang, X., Hu, M., Srinivasan, R.,
and Yao, S.Q. (2009). High-throughput discovery of Mycobacterium tubercu-
losis protein tyrosine phosphatase B (MptpB) inhibitors using click chemistry.
Org. Lett. 11, 5102–5105.
Papageorgiou, P., Katsambas, A., and Chu, A. (2000). Phototherapy with blue
(415 nm) and red (660 nm) light in the treatment of acne vulgaris. Br. J.
Dermatol. 142, 973–978.
Tomlin, F.M., Gordon, C.G., Han, Y., Wu, T.S., Sletten, E.M., and Bertozzi, C.R.
(2018). Site-specific incorporation of quadricyclane into a protein and photo-
cleavage of the quadricyclane ligation adduct. Bioorg. Med. Chem. 26,
5280–5290.
Sarfaty, M., Leshno, M., Gordon, N., Moore, A., Neiman, V., Rosenbaum, E.,
and Goldstein, D.A. (2018). Cost effectiveness of nivolumab in advanced renal
cell carcinoma. Eur. Urol. 73, 628–634.
Tso, K.K., Leung, K.K., Liu, H.W., and Lo, K.K. (2016). Photoactivatable cyto-
toxic agents derived from mitochondria-targeting luminescent iridium(III)
poly(ethylene glycol) complexes modified with a nitrobenzyl linkage. Chem.
Commun. 52, 4557–4560.
Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image to
ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675.
Urbanska, K., Lanitis, E., Poussin, M., Lynn, R.C., Gavin, B.P., Kelderman, S.,
Yu, J., Scholler, N., and Powell, D.J., Jr. (2012). A universal strategy for adop-
tive immunotherapy of cancer through use of a novel T-cell antigen receptor.
Cancer Res. 72, 1844–1852.
Shanmugam, S., Xu, J., and Boyer, C. (2016). Light-regulated polymerization
under near-infrared/far-red irradiation catalyzed by bacteriochlorophyll a.
Angew. Chem. Int. Ed. 55, 1036–1040.
Song, D.G., Ye, Q., Carpenito, C., Poussin, M., Wang, L.P., Ji, C., Figini, M.,
June, C.H., Coukos, G., and Powell, D.J., Jr. (2011). In vivo persistence, tumor
localization, and antitumor activity of CAR-engineered T cells is enhanced by
costimulatory signaling through CD137 (4-1BB). Cancer Res. 71, 4617–4627.
Vaughan, T.J., Williams, A.J., Pritchard, K., Osbourn, J.K., Pope, A.R.,
Earnshaw, J.C., McCafferty, J., Hodits, R.A., Wilton, J., and Johnson, K.S.
(1996). Human antibodies with sub-nanomolar affinities isolated from a large
non-immunized phage display library. Nat. Biotechnol. 14, 309–314.
Di Stasi, A., Tey, S.K., Dotti, G., Fujita, Y., Kennedy-Nasser, A., Martinez, C.,
Straathof, K., Liu, E., Durett, A.G., Grilley, B., et al. (2011). Inducible apoptosis
as a safety switch for adoptive cell therapy. N. Engl. J. Med. 365, 1673–1683.
Wang, J., Liu, Y., Liu, Y., Zheng, S., Wang, X., Zhao, J., Yang, F., Zhang, G.,
Wang, C., and Chen, P.R. (2019). Time-resolved protein activation by proximal
decaging in living systems. Nature 569, 509–513.
10 Cell Chemical Biology 28, 1–10, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
STAR+METHODS
KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Antibodies
Anti-Mouse IgG, F(ab’)2 specific, Alexa Fluor 647
Folic acid Antibody (FA1) [Alexa Fluorꢀ 647]
Jackson Immuno Research
Novus Biologicals
Cat# 115-605-006; RRID: AB_2338903
Cat#
NB100-72975AF647; RRID: AB_2876876
Anti-FITC antibody (PE)
Abcam
Cat# 25539; RRID: AB_732396
Cat# 310938; RRID: AB_2562307
Cat# 908304; RRID: AB_2629795
Cat# 302606; RRID: AB_314276
Cat# 300412; RRID: AB_314066
Cat# 20376-HP; RRID: AB_2876877
Brilliant Violet 605 anti-human CD69
PE anti-FOLR1 (Folate binding protein)
PE anti-human CD25
Biolegend
Biolegend
Biolegend
APC anti-human CD3
Biolegend
Monoclonal Anti-FITC IgG-HRP Conjugate
Biological Samples
Alpha Diagnostic International
Human PBMC, health donor
Peking Union Medical College
Hospital
N/A
Chemicals, Peptides, and Recombinant Proteins
Aim-V medium
Thermo
Cat# 12055091
Cat# 27016-021
Cat# 41090036
Cat# 11344D
RIMI 1640 Medium, no folic acid
MEM, GlutaMAX Supplement
Dynabeads Untouched Human T Cells
Dynabeads Human T-expander CD3/CD28
Human IL-2 Recombinant Protein
Lipofectamine 3000 Transfection Reagent
Penicillin/streptomycin
Thermo
Thermo
Thermo
Thermo
Cat# 11141D
Thermo
Cat# BMS334
Thermo
Cat# L3000015
Cat# 15140163
Cat# 16140071
Cat# 17-1440-02
Cat# 17501; CAS 7240-37-1
CAS 910661-33-5
Thermo
Fetal Bovine Serum
Thermo
Ficoll-Paque PLUS
GE Healthcare
AAT Bioquest
(Kim, et al., 2015)
Macklin
7-Aminoactinomycin D
Compound Folate-FITC
4’-Hydroxy-3’-methoxyacetophenone
Cat# H830242;
CAS 498-02-2
Propargyl bromide
Macklin
Aladdin
Aladdin
Aladdin
Macklin
Macklin
Macklin
This paper
Cat# P815876;
CAS 106-96-7
Sodium borohydride
4-Nitrophenyl Chloroformate
Ethylenediamine
Cat# S108355;
CAS 16940-66-2
Cat# N159046;
CAS 7693-46-1
Cat# E112643;
CAS 107-15-3
N-Hydroxysuccinimide
N,N0-Dicyclohexylcarbodiimide
Folic acid
Cat# N811124;
CAS 6066-82-6
Cat# N806920;
CAS 538-75-0
Cat# F809516;
CAS 59-30-3
Compound Folate-O-FITC, detailed synthesis
described in Supplemental Information
N/A
Critical Commercial Assays
CytoTox 96 ꢀ Non-Radioactive
Cytotoxicity Assay
Promega
Dakewei
Cat# G1780
Human IL-2 ELISA Kit
Cat# 1110202
(Continued on next page)
Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021 e1

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Continued
REAGENT or RESOURCE
Human IFN-g ELISA Kit
Folic Acid (FA) ELISA kit
Oligonucleotides
4m5.3 anti-FITC scFv
4-1BB
SOURCE
IDENTIFIER
Dakewei
Cat# 1110002
Cat# 0365-0B9
Alpha Diagnostic International
(Midelfort, et al., 2004)
Genebank
N/A
U03397.1
J04132.1
CD3 z
Genebank
Software and Algorithms
Prism 6.0
Graphpad
https://www.graphpad.com/scientificsoftware/prism/
https://www.perkinelmer.com/category/ chemdraw
https://www.beckman.ae/flow-cytometry
https://imagej.nih.gov/ij
ChemDraw Professional 18.0
CytExpert 2.3
PerkinElmer
Beckman Coulter
(Schneider et al., 2012)
ImageJ
RESOURCE AVAILABILITY
Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Xuan
Zhang, MD (zxpumch@sina.com);
Materials Availability
This study did not generate any unique biological reagents. Synthesis of all chemicals is detailed within the manuscript.
Data and Code Availability
All data generated or analyzed during this study are included in this published article and its Supplementary Information files. The raw
datasets supporting the current study have not been deposited in a public repository but are available from the corresponding author
on request.
EXPERIMENTAL MODELS AND SUBJECT DETAILS
Animals
Six- to eight-week-old female SCID/beige or NOG mice (Vital River Laboratory Animal Technology Co., Ltd. Beijing) were maintained
at Peking University Health Science Center in compliance with Peking University Institutional Animal Care and all studies reported
here were approved by the Ethics Review Board of PUMC Hospital, Chinese Academy of Medical Science (CAMS). Mice were
co-housed at 3-5 mice per cage and maintained on folic acid-deficient diet under pathogen-free conditions on a 12-h light cycle
with access to food and water ad libitum.
Mammalian Cell Culture
All cells were maintained in a humidified incubator at 37C and 5% CO2. KB cells were kindly provided by Prof. Xinjing Tang (Peking
University) and maintained in RPMI 1640 medium (without folic acid) supplemented with 10% heat-inactivated fetal bovine serum
(FBS) plus 1% antibiotics. The virus-producing cell line HEK293FT was purchased from Life Technologies (USA) and maintained
in Dulbecco’s modified Eagle’s medium containing 10% FBS, 1% antibiotics, 0.1 mM MEM/nonessential amino acids, 1 mM sodium
pyruvate, and 500 mg/mL geneticin. Human CD3+ T cells were isolated from healthy volunteer donors by Ficoll-Pacque (Dakewei
Biotech Co., Ltd.) and then subjected to negative selection (Life Technologies). All specimens were collected according to a protocol
approved by the Institutional Review Board of our hospital. Isolated human T cells were subsequently cultured in Aim-V medium sup-
plemented with 10% FBS and 300 IU/mL IL-2 (Life Technologies). Jurkat, A549 and SKOV3 cells were purchased from American
Type Culture Collection (ATCC, Manassas, VA, USA) and cultured according to ATCC recommendations.
METHOD DETAILS
Detailed Synthetic Procedures to Folate-O-FITC (Compound 10)
The synthesis of the designed mediator molecule was based on a modular assembly as shown in Figure S1. Briefly, the ortho-nitro-
benzyl ester bearing a terminal alkyne and free amine group was obtained by propargylation, ester exchange, and ammonolysis from
a commercial ortho-nitrobenzylic alcohol. Coupling of the ortho-nitrobenzylic ester to folate was conducted with DCC as the
1
condensing reagent, and conjugation to FITC was then completed using the CuAAC reaction. H NMR spectra were recorded on
e2 Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
400 or 600 MHz using the residual solvent signal TMS as internal standard. Chemical shifts were reported in parts per million and
coupling constants quoted in Hz. Matrix Assisted Laser Desorption/Ionization time-of-flight mass spectrometry with 2,5- Dihydrox-
ybenzoic acid (DHB) as the matrix. MS were performed with Bruker BIFLEX III and Bruker Daltonics Autoflex speed MALDITOF/TOF.
1H NMR, 13C NMR data were recorded with an Avance 400/DPX (Bruker) and Avance III HD600 (Bruker). The more details about syn-
thesis are described below:
Compound 1 1,4-Diazidobutane
To a solution of 1,4-dibromobutane (2.38 mL, 20.0 mmol) in DMF (20 mL) was added a solution of NaN3 (2.73 g, 42.0 mmol, 2.1 eq.)
in H2O (10 mL). After stirring overnight at 20ꢀC, the reaction mixture was cooled down to room temperature and diluted with brine
(40 mL) and hexanes (40 mL). The organic layer was separated and the aqueous layer was extracted with hexane (2340 mL). The
combined organic layers were dried over MgSO4, filtered and concentrated in vacuo to give title compound 1 as clear oil (2.73 g,
19.6 mmol, 98%). 1H NMR (400 MHz, CDCl3) d 3.30 (t, J = 6.8 Hz, 2H), 2.74 (t, J = 6.7 Hz, 2H), 1.70-1.60 (m, 2H), 1.53
(dt, J = 14.2, 7.0 Hz, 2H). Analytical data are in agreement with literature (Jung, et al., 2010).
Compound 2 4-azidobutan-1-amine
1,4-Diazidobutane (2.72 g, 19.4 mmol) was dissolved in a mixture of Et2O (12 mL), EtOAc (12 mL) and aqueous HCl (1 M, 40 mL).
Triphenylphosphine (5.09 g, 19.4 mmol, 1.00 eq.) was divided into several parts and added within 1 h at 0ꢀC. The mixture was warmed
to 21ꢀC and shirred for an addition 20 hours. The aqueous layer was separated and washed with Et2O (2 3 40 mL). The pH of the
aqueous layer was adjusted to pH 13 with aqueous NaOH (2 M). After extraction with DCM (33 60 mL), the combined organic layers
were dried over MgSO4, filtered and concentrated in vacuo to give title compound 2 as clear oil (1.70 g, 77%). 1H NMR (400 MHz,
CDCl3) d 3.27 (t, J = 6.8 Hz, 2H), 2.71 (t, J = 6.7 Hz, 2H), 1.68-1.57 (m, 2H), 1.50 (dt, J = 14.2, 7.0 Hz, 2H). Analytical data are in agree-
ment with literature (Jung, et al., 2010).
Compound 3 5-(3-(4-azidobutyl)thioureido)-2-(3,6-dihydroxy-9H- xanthen-9-yl) benzoic acid
To a solution of 4-azidobutan-1-amine (34 mg, 0.3 mmol) in DMSO (5mL) were added FITC (100 mg, 0.25 mmol) and DIPEA (65 mg,
0.5 mmol). After stirring for 6 hours. After the reaction was completed, the solvent was removed under vacuum, which also removed
part of the excess 4-azidobutan-1-amine. The compound was purified by flash chromatography. The product was obtained as an
1
orange solid in 40% yield. H NMR (400 MHz, Methanol-d4 ) d 7.99 (s, 1H), 7.71-7.68 (m, 2H), 7.19 (d, J = 8.0 Hz, 1H), 7.12-7.10
(m, 2H), 6.64-6.60 (m, 4H), 3.69-3.65 (m, 2H), 3.37 (t, J = 8.0 Hz, 2H), 1.79-1.66 (m, 4H). 13C NMR (151 MHz, Methanol-d4)
d 182.3, 172.5, 157.4, 141.4, 137.5, 131.8, 130.6, 129.2, 119.2, 114.2, 103.7, 55.5, 52.0, 27.1. HRMS: m/z C25H21N5NaO5S
[M+Na]+ 526.1156, Found 526.1160.
Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021 e3

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Compound 4 1-(3-methoxy-4-(prop-2-yn-1-yloxy) phenyl) ethan-1-one
To a solution of 4-Methoxy-3-hydroxyacetophenone (3.32 g, 20.0 mmol) in DMF (20 mL) was added potassium carbonate (2.76 g,
20.0 mmol) and propargyl bromide (2.38 g, 20.0 mmol). The reaction mixture was stirred at room temperature for 2h. The reaction
mixture was extracted with DCM (33100 mL). The organic layer was dried over MgSO4, filtered and concentrated in vacuo to
1
give tile compound 4 as a light brown solid (4.0 g, 98%). Compound 4 H NMR (600 MHz, CDCl3) d 7.57 (d, J = 8.4 Hz, 1H), 7.55
(s, 1H), 7.05 (d, J = 8.2 Hz, 1H), 4.84 (s, 2H), 3.93 (s, 3H), 2.57 (s, 3H). Analytical data are in agreement with literature (Tan, et al., 2009).
Compound 5 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-yloxy) phenyl) ethan-1-one
To a solution of 70% nitric acid (47 mL) and acetic anhydride (17 mL) was added dropwise a solution of compound 4 (2.17 g,
8.20 mmol) dissolved in acetic anhydride (7 mL) at 0ꢀC. The resultant yellow mixture was stirred for 3 h at room temperature.
Cold deionized water (300 mL) was added to the mixture, which was then stirred for 2 h at 0ꢀC. The yellow precipitate was collected
by suction filtration and purified by column chromatography on silica gel. Compound 5 was obtained as a yellow solid (1.45 g, 58%).
Compound 5 1H NMR (400 MHz, CDCl3) d 7.80 (s, 1H), 6.78 (s, 1H), 4.86 (s, 2H), 3.98 (s, 3H), 2.61 (s, 1H), 2.51 (s, 3H). Analytical data
are in agreement with literature (Tso, et al., 2016).
Compound 6 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-yloxy) phenyl) ethan-1-ol
The solid compound 5 (1.2 g, 4.9 mmol) was dissolved in THF/MeOH (20mL, 1/2, v/v) and cooled at 0ꢀC for 5 min. Sodium boro-
hydride (101 mg, 2.66 mmol) was added to the solution. The mixture was transferred to room temperature for 3 h. The reaction was
quenched by deionized water (5 mL) and the pale-yellow crude product was extracted with CH2Cl2 (33100 mL). The organic layers
were concentrated by rotary evaporation and purified by column chromatography on silica gel. Compound 6 was obtained as a yel-
low solid (1.11 g, 90%). Compound 6 1H NMR (400 MHz, CDCl3) d 7.73 (s, 1H), 7.33 (s, 1H), 5.56 (q, J = 6.0 Hz, 1H), 4.80 (d, J = 2.4 Hz,
2H), 3.99 (s, 3H), 2.57 (t, J = 2.4 Hz, 1H), 1.55 (d, J = 6.4 Hz, 3H). Analytical data are in agreement with literature (Tso, et al., 2016).
e4 Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
Compound 7 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-yloxy) phenyl) ethyl (4-nitro phenyl) carbonate
To a solution of compound 6 (1.25 g, 5 mmol) in CH3CN (5 ml) was added 4-nitrophenyl chloroformate (1.2 g, 6 mmol) and pyridine
(5 mL) at 0ꢀC. The mixture was stirred at room temperature for 6 h. The reaction was extracted with CH2Cl2 (33100 mL). The organic
layers were concentrated by rotary evaporation and purified by column chromatography on silica gel. compound 7 was obtained as a
light yellow solid (1.35 g, 65%). Compound 7 1H NMR (400 MHz, CDCl3) d 8.24 (d, J = 9.2 Hz, 2H), 7.78 (s, 1H), 7.34 (d, J = 9.2 Hz, 2H),
7.15 (s, 1H), 6.54 (q, J = 6.4 Hz, 1H), 4.83 (d, J = 2.0 Hz, 2H), 4.02 (s, 3H), 2.60 (s, 1H), 1.77 (d, J = 6.4 Hz, 3H). 13C NMR (151 MHz,
CDCl3) d 155.32, 154.40, 151.46, 146.08, 145.43, 139.68, 132.40, 126.16, 125.32, 121.68, 115.67, 110.30, 108.32, 73.73, 57.02,
56.62, 22.01. HRMS: m/z C19H17N2O9 [M+H]+ 417.0929, Found 417.0931.
Compound 8 1-(5-methoxy-2-nitro-4-(prop-2-yn-1-yloxy) phenyl) ethyl (2-amino ethyl) carbamate
To a solution of ethylenediamine (180 mg, 3 mmol) and DIPEA (623 mL, 4 mmol) in DCM (10 mL) were added dropwise a solution of
Compound 7 (832 mg, 2 mmol) in DCM (5 mL) at room temperature. After stirring at room temperature for overnight, the reaction
mixture was concentrated to dryness in vacuo. The residue was purified using silica gel chromatography. compound 8 was obtained
as a yellow solid (566.7 mg, 84%). Compound 8 1H NMR (400 MHz, CDCl3) d 7.75 (s, 1H), 7.04 (s, 1H), 6.38 (q, J = 8.0 Hz, 1H), 5.23 (s,
1H), 4.80 (d, J = 4.0 Hz, 2H), 3.96 (s, 3H), 3.23-3.15 (m, 2H), 2.81 (t, J = 8.0 Hz, 2H), 2.57 (t, J = 4.0 Hz, 1H), 1.60 (d, J = 4.0 H, 3H). 13C
NMR (151 MHz, CDCl3) d 155.67, 154.17, 145.42, 139.43, 135.13, 110.25, 108.46, 68.76, 56.97, 56.47, 53.49, 43.04, 41.38, 22.23.
HRMS: m/z C15H20N3O6 [M+H]+ 338.1347, Found 338.1352.
Compound 10 5-(3-(4-(4-((4-(1-(4-(((2-amino-4-oxo-3,4-dihydropteridin-6-yl) methyl)amino)phenyl)-3-carboxy-1,6,11-trioxo-12-
oxa-2,7,10-triazatetradecan-13-yl)-2-methoxy-5nitrophenoxy)methyl)-1H-1,2,3-triazol-1-yl)butyl)thioureido)-2-(6-hydroxy-3-oxo-
3H-xanthen-9-yl) benzoic acid
A solution of folic acid (302 mg, 0.68 mmol,.), NHS (86 mg, 0.75 eq.), and dicyclohexylcarbo-diimide (DCC) (154 mg, 0.75 eq.) in
anhydrous DMSO (16 mL) was stirred for 4 h in the dark. The reaction mixture was filtered over a PTFE (0.45 mm) membrane to remove
dicyclohexylurea. Compound 8 (229 mg 0.68 mmol) were added to the filtrate. The reaction mixture was stirred for 2 days in the dark.
The clear solution was then concentrated by partial removal of the solvent under reduced pressure at 40ꢀC. The solution was poured
into a cold solution of acetone/diethyl ether (30:70% v/v ratio) resulting in precipitation of the product. The obtained precipitate was
washed with 50 mL of acetone/diethyl ether solution (30:70% v/v ratio) and diethyl ether (3350 mL) to remove traces of unreacted
reagents and DIPEA-NHS salt. The product was dried in vacuum over P2O5 overnight (310 mg, 60%). The compound 9 was direct
used for the following click reactions without further purifications. To a solution of compound 9 (343 mg, 0.68 mmol) in DMSO (10 mL)
was added compound 3 (343 mg, 0.68 mmol), CuSO4$5H2O (17 mg, 0.068 mmol) and Sodium L-Ascorbate (54 mg, 0.272 mmol)
Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021 e5

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
under nitrogen atmosphere. The reaction was stirred at room temperature for 2 days. The solution was poured into a cold solution of
acetone/diethyl ether (30:70% v/v ratio) resulting in precipitation of the product. The obtained precipitate was washed with DCM
(50 mL) and water (50 mL). The product was purified by HPLC and freeze-drying to give compound 10 as a yellow solid
1
(395.4 mg, 41%). H NMR (600 MHz, DMSO) d 8.63 (s, 1H), 8.38-8.08 (m, 2H), 7.99-7.76 (m, 2H), 7.75-7.52 (m, 2H), 7.42-7.28 (d,
J = 6.0 Hz, 1H), 7.28-7.13 (m, 1H), 7.11-7.04 (m, 1H), 6.98-6.81 (s, 2H), 6.71-6.52 (m, 5H), 6.18-5.98 (s, 1H), 5.32-5.13 (m, 2H),
4.62-4.30 (m, 4H), 3.97-3.68 (m, 3H), 3.04-2.91 (m, 5H), 2.16-2.07 (s, 1H), 2.06-1.69 (m, 4H), 1.52-1.46 (s, 6H). MALDI-TOF-MS:
m/z C59H57N15NaO16S [M+Na]+1286.37, Found 1286.85; C59H57KN15O16S [M+K]+ 1302.35, Found 1302.83.
Virus Production and Generation of CAR-Ts
A gene cassette containing the 4M5.3 anti-FITC scFv (Ma, et al., 2016; Midelfort, et al., 2004; Vaughan, et al., 1996), the CD8a hinge
and transmembrane region, and the cytoplasmic domains of 4-1BB and CD3z was synthesized by Icartab Co., Ltd. (SuZhou, China)
and cloned into a lentivirus vector. Lentivirus production and transduction of human T cells were performed as previously described
(Kim, et al., 2015; Ma, et al., 2016). Briefly, HEK293FT cells were transfected with the anti-FITC CAR, pspax2, and pMD2.g plasmids
using Lipofectamine 3000 (Life Technologies). The medium was changed 6 h after transfection, and the viral supernatant was har-
vested 48 h after transfection. Viral particles were concentrated 30-fold by ultracentrifugation for 2 h at 25,000 rpm with a Beckman
Ti70 rotor (Beckman Coulter) and frozen at ꢁ80ꢀC until ready for use. Human T cells were purified from peripheral blood mononuclear
cells using a Dynabead Untouched Human T cells kit (Life Technologies) and further activated for 24 h with CD3/CD28-coated mag-
netic beads (Life Technologies) before infection. Concentrated lentivirus was then applied to activated human T cells (~50 mL/106
human T cells in 24-well culture plates) in the presence of 10 mg/mL polybrene and 50 IU/mL IL-2 and centrifuged at 1000 3 g for
2 h at 32ꢀC. The next day, viral supernatants were replaced with fresh medium containing 300 IU/mL IL-2, and the transduced
T cells were maintained at 0.5 3 106 cells/mL in complete growth medium, which was replenished every 4 days.
Evaluation of Transduction Efficiency and Function of CAR-Ts by Flow Cytometry Analysis
The transduction efficiency and function of CAR-Ts were both verified by flow cytometry, as previously described (Kim, et al., 2015).
Briefly, to assess anti-FITC CAR expression, CAR-Ts were incubated with APC-conjugated anti-mouse IgG F(ab’)2 antibodies for 1 h
at 4ꢀC and then analyzed using a CytoFLEX flow cytometer (Beckman Coulter, Inc.). To determine the availability of the folate-based
mediator, CAR-Ts were stained with the mediator and APC-conjugated anti-mouse IgG F(ab’)2 antibodies simultaneously, and the
binding capacity of FITC was measured based on the percentage of cell clusters positive for both FITC and APC. The affinity of KB
cells for the mediator was verified by incubation of KB cells with 10 nM mediator for 1 h at room temperature, followed by staining with
an anti-FITC antibody conjugated to a PE dye (Abcam, Cambridge, UK) for more than 45 min. After being washed twice, the cells were
analyzed using a CytoFLEX flow cytometer.
Cytotoxicity and Cytokine Release Assay
KB cells were mixed with anti-FITC CAR-Ts at an E:T ratio of 10:1 in 200 mL of folate-deficient medium with different concentrations of
mediator for 24 – 48 h. Cytolytic activity was determined using a lactate dehydrogenase-based (LDH) method with a CytoTox 96
Nonradioactive Cytotoxicity Assay kit (Promega, Madison, WI, USA) following the manufacturer’s protocol. The percent cytotoxicity
was calculated as follows: % cytotoxicity = ((experimental – effector spontaneous – target spontaneous) / (target maximum – target
spontaneous)) 3 100. The remaining culture supernatants were harvested and assayed for the presence of IFN-g and IL-2 using an
enzyme-linked immunosorbent assay kit (Dakewei Biotech Co., Ltd.), according to the manufacturer’s instructions.
Photocleavage Experiments at the Cellular Level
For cleavage assays, KB cells were incubated with 100 nM mediator for 1 h at room temperature and then washed twice with wash
buffer (PBS + 3% bovine serum albumin) to remove excess mediator. Cells were resuspended in 200 mL PBS at 1 3 106 cells/mL and
then subjected to UV irradiation at 365 nm approximately 5 cm above the cell surface using a UV lamp (15 W 3 16, UV lamp power:
2.76 mW/cm2; Panchum, photochemical reactor PR-2000) for 20 min at room temperature. A control sample was not subjected to
photoirradiation. After UV irradiation, the cleaved part of the mediator was removed from reactions by a series of three washes. The
cleavage of CAR-Ts was similar to that of KB cells, except that the CAR-Ts were first stained with the mediator and then stained with
an anti-folate antibody (Novus Biologicals) conjugated to an Alexa Fluor 647 dye. The cleavage efficiency of the mediator was deter-
mined by the attenuation of the fluorescent signal as determined by flow cytometry analysis. Additionally, the viability of cells after UV
irradiation was confirmed by flow cytometry analysis using both the 7-AAD and trypan blue methods.
For immunofluorescent imaging, CAR-Ts and KB cells were stained with anti-CD3-APC and anti-FR-PE antibodies, respectively,
and cocultured in the presence or absence of the mediator at a ratio of 1:1 for 1 h at 4ꢀC. After three washes with PBS, some of the
mediator-treated cells were subjected to UV irradiation (365 nm) for 20 min and then washed adequately to remove the cleaved medi-
ator. The remaining mediator-treated cells were used as a non-photoirradiated control and were subjected to the same washing pro-
cess as the photoirradiation group. Analyses were performed using Amnis IDEAS software to evaluate the binding and separation of
CAR-Ts and KB cells mediated by the mediator. All images were acquired using a 403 objective (ImageStream x, Amnis), analyzed
using IDEAS software (Amnis), and illustrated with a channel series, including Bright field mode (BF, channel 1). The fluorescence
channels used for imaging KB and CAR-T cells were channel 2 (PE, anti-FR-PE) and channel 3 (APC, anti-mouse-IgG F(ab’)2-APC),
respectively.
e6 Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021

Please cite this article in press as: Zhang et al., Photoswitchable CAR-T Cell Function In Vitro and In Vivo via a Cleavable Mediator, Cell Chemical
Biology (2020), https://doi.org/10.1016/j.chembiol.2020.10.004
ll
Article
In Vivo Regulated Efficacy Assays
To evaluate the effects of the regulation of CAR-T cells in vivo, a KB tumour-bearing SCID/beige mouse model was applied as
previously described with slight changes (Lee, et al., 2019). Briefly, immunodeficient SCID/beige mice (6-8 weeks of age, Vital River
Laboratory Animal Technology Co., Ltd., Beijing) were implanted subcutaneously with 1 3 107 KB cells and received an intratumoural
injection of 1.5~2 3 107 anti-FITC CAR-T cells when the tumour reached ~100 mm3 in size on day 0. Alternate-day dosing with the
mediator was applied starting on this day, and inactivation was achieved by UV photoirradiation 6 h after each mediator injection on
days 6, 8 and 10 ( ~5 cm from the skin for 5 min using a 3 W flashlight). Mouse blood was also collected to measure the levels of IL-2
and IFN-g using ELISA (BioLegend, CA), and CAR-T cell-induced toxicity was monitored by measuring body weight loss. The tumour
dimensions were measured with callipers, and the tumour volume was calculated using the following formula: volume = (length 3
width2)/2, in which the length was the greatest longitudinal diameter, and the width was greatest transverse diameter. For the diffuse
tumour model, female NOD.Cg-PrkdcscidIl2rgtm1Sug/JicCrl mice (NOG mice; 6–8 weeks of age) were intravenously inoculated with
2 3 106 SKOV3 cells transfected with luciferase, and engraftment was confirmed by BLI 1 week later. The next day, CAR-T cells
were infused intravenously (1 3 107 cells/per mouse) and treated with the indicated amounts of the mediator. In parallel, mice in con-
trol groups were injected with only CAR-T cells. Tumour growth was monitored weekly by BLI.
QUANTIFICATION AND STATISTICAL ANALYSIS
All statistical calculations were performed using GraphPad Prism 6. Accordingly, data sets were subjected to one-way ANOVA (Dun-
nett’s multiple comparison) and the two-tailed unpaired Student’s t-test as indicated in the figures. A p values less than 0.05 were
considered significant. *p % 0.05, **p % 0.01, ***p % 0.001, ****p % 0.0001. Sample sizes are indicated in the respective figure
legend.
Cell Chemical Biology 28, 1–10.e1–e7, January 21, 2021 e7