NKG2D-CAR T cells eliminate senescent cells in aged mice and nonhuman primates

Abstract

Cellular senescence, characterized by stable cell cycle arrest, plays an important role in aging and age-associated pathologies. Eliminating senescent cells rejuvenates aged tissues and ameliorates age-associated diseases. Here, we identified that natural killer group 2 member D ligands (NKG2DLs) are up-regulated in senescent cells in vitro, regardless of stimuli that induced cellular senescence, and in various tissues of aged mice and nonhuman primates in vivo. Accordingly, we developed and demonstrated that chimeric antigen receptor (CAR) T cells targeting human NKG2DLs selectively and effectively diminish human cells undergoing senescence induced by oncogenic stress, replicative stress, DNA damage, or P16INK4a overexpression in vitro. Targeting senescent cells with mouse NKG2D-CAR T cells alleviated multiple aging-associated pathologies and improved physical performance in both irradiated and aged mice. Autologous T cells armed with the human NKG2D CAR effectively delete naturally occurring senescent cells in aged nonhuman primates without any observed adverse effects. Our findings establish that NKG2D-CAR T cells could serve as potent and selective senolytic agents for aging and age-associated diseases driven by senescence.

Full text

AGING
NKG2D-CAR T cells eliminate senescent cells in aged
mice and nonhuman primates
Dong Yang
1
†, Bin Sun
1
†, Shirong Li
1
, Wenwen Wei
1
, Xiuyun Liu
2
, Xiaoyue Cui
2
, Xianning Zhang
2
,
Nan Liu
1
, Lanzhen Yan
1
, Yibin Deng
3
*, Xudong Zhao
1,2
*
Cellular senescence, characterized by stable cell cycle arrest, plays an important role in aging and age-associated
pathologies. Eliminating senescent cells rejuvenates aged tissues and ameliorates age-associated diseases.
Here, we identied that natural killer group 2 member D ligands (NKG2DLs) are up-regulated in senescent
cells in vitro, regardless of stimuli that induced cellular senescence, and in various tissues of aged mice and
nonhuman primates in vivo. Accordingly, we developed and demonstrated that chimeric antigen receptor
(CAR) T cells targeting human NKG2DLs selectively and eectively diminish human cells undergoing senescence
induced by oncogenic stress, replicative stress, DNA damage, or P16
INK4a
overexpression in vitro. Targeting sen-
escent cells with mouse NKG2D-CAR T cells alleviated multiple aging-associated pathologies and improved
physical performance in both irradiated and aged mice. Autologous T cells armed with the human NKG2D
CAR eectively delete naturally occurring senescent cells in aged nonhuman primates without any observed
adverse eects. Our ndings establish that NKG2D-CAR T cells could serve as potent and selective senolytic
agents for aging and age-associated diseases driven by senescence.
Copyright © 2023 The
Authors, some
rights reserved;
exclusive licensee
American Association
for the Advancement
of Science. No claim
to original U.S.
Government Works
INTRODUCTION
Although senescence plays a beneficial role in tumor suppression,
wound healing, and possibly embryonic development, aberrant ac-
cumulation of senescent cells in tissues contributes to age-related
chronic diseases and degenerative disorders (1–3). To combat this
issue, senolytics have been developed, primarily using chemical
drugs that target senescent cells. These senolytics have been demon-
strated to rejuvenate aged tissues and prolong the life span of mice
(4–9). However, because of the heterogeneity of senescent cells (10),
it remains necessary to develop more senolytics.
Immunotherapy that specifically targets cell surface antigens has
displayed promising results in treating diseases, such as cancer, that
escape endogenous immune surveillance (11–15). Using this ap-
proach, a potential immunotherapeutic strategy for destroying sen-
escent cells is through vaccination. For instance, a vaccine targeting
glycoprotein nonmetastatic melanoma protein B (GPNMB) have
been shown to remove senescent vascular endothelial cells and leu-
kocytes in mice with atherosclerosis (16). A vaccine targeting
CD153 has also been found to deplete senescent T cells from
high-fat diet–induced obese mice (17). A recent study used seno-
lytic chimeric antigen receptor (CAR) T cells targeting senescent
cells expressing urokinase-type plasminogen activator receptor
(uPAR). The uPAR-CAR T cell treatment extended the survival of
mice carrying lung adenocarcinomas by selectively eliminating
mitogen-activated protein kinase kinase (MEK) and cyclin-
dependent kinase 4/6 (CDK4/6) inhibitor treatment–induced sen-
escent tumor cells. The CAR T cells also alleviated liver fibrosis in
mice by removing carbon tetrachloride- or diet-induced senescent
cells expressing uPAR (15).
Despite the fact that natural killer group 2 member D ligands
(NKG2DLs) are highly expressed in senescent cells and not in
normal cells, the senescent cells that express NKG2DLs escape en-
dogenous natural killer (NK) cell–mediated innate immune clear-
ance (18). Thus, we conducted a study to determine whether
adoptive transfer of NKG2D-CAR T cells could selectively and effi-
ciently eliminate NKG2DL-expressing senescent cells and explore
their senolytic roles in targeting naturally occurring senescent
cells expressing NKG2DLs, potentially delaying or blocking aging
phenotypes in vivo.
Our studies demonstrated that NKG2D-CAR T cells could safely
be used as senolytic agents by selectively and effectively depleting in
vivo senescent cells expressing NKG2DLs in aged mice and nonhu-
man primates. Furthermore, the use of murine NKG2D-CAR T
cells in two independent mouse models of aging reversed senes-
cence-associated phenotypes, highlighting their potential in treat-
ing aging and age-related diseases.
RESULTS
NKG2D-CAR T cells selectively kill cells undergoing
senescence induced by dierent stresses
To assess the impact of NKG2D-CAR T cells on different types of
cellular senescence, including replicative exhaustion, DNA damage
induced by etoposide treatment, and oncogene-induced senescence
triggered by Kras stress, we established cellular senescence models
using the human lung fibroblast cell line IMR90 (4,19,20). Senes-
cence was confirmed by senescence-associated β-galactosidase (SA-
β-Gal) staining and the detection of the senescence marker P16
INK4a
mRNA expression in IMR90 cells (fig. S1). The mRNA and protein
expression of different NKG2DLs, such as major histocompatibility
1
Department of Targeting Therapy and Immunology and Laboratory of Animal
Tumor Models, Cancer Center and State KeyLaboratory of Biotherapy and National
Clinical Research Center for Geriatrics and Frontiers Science Center for Disease-
related Molecular Network, West China Hospital, Sichuan University, Chengdu,
Sichuan 610041, China.
2
Key Laboratory of Animal Models and Human Disease
Mechanisms of Chinese Academy of Sciences, Kunming Institute of Zoology,
Chinese Academy of Sciences, Kunming,, Yunnan 650223, China.
3
Department
of Urology, Masonic Cancer Center, University of Minnesota Medical School, Min-
neapolis, MN 55455, USA.
*Corresponding author. Email: zhaoxudong@wchscu.cn (X.Z.);
dengx103@umn.edu (Y.D.)
†These authors contributed equally to this work.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 1 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

complex (MHC) class I polypeptide–related sequence A and B
(MICA and MICB) and UL16 binding protein 1 (ULBP1) to
ULBP3, detected by quantitative real-time polymerase chain reac-
tion (qRT-PCR) and by flow cytometry, respectively, was enhanced
in various stress-induced cellular senescence in IMR90 cells (Fig. 1A
and fig. S2). These results are consistent with previous studies (21,
22) that NKG2DLs could serve as a cell surface marker for senescent
cells. Thus, we proposed the possibility that NKG2D-CAR T cells
may effectively target senescent cells expressing NKG2DLs
through their specific natural receptor, NKG2D.
A lentivirus-mediated human NKG2D (hNKG2D) CAR com-
prising the NKG2D extracellular domain (NKG2D-ECD), CD137
(4-1BB), and CD3ζ signaling domains was constructed and trans-
duced into human T cells. A vector lacking the NKG2D extracellular
domain (mock) was used as a negative control (fig. S3). Upon co-
culturing with NKG2D-CAR T cells, senescent cells induced by
various stresses were efficiently eliminated in a dose-dependent
manner (Fig. 1B). Correspondingly, enzyme-linked immunosor-
bent assays (ELISAs) of the coculture medium showed significant
increases of tumor necrosis factor–α (TNF-α) and interferon-γ
(IFN-γ) production upon challenge with NKG2D-CAR T cells (P
< 0.001; Fig. 1C), indicating the activation of NKG2D-CAR T
cells. The secretion of granzyme B and perforin, crucial effector
molecules mediating activated T cell functions, was also signifi-
cantly increased in the coculture medium (P< 0.05; Fig. 1D).
These results were also reproduced using the HEL1 and WI38
human lung fibroblast cell lines (fig. S4).
P16
INK4a
, a tumor suppressor protein, is a key contributor to per-
manent cell cycle arrest during aging, and accordingly, clearance of
P16
INK4a
-positive cells in mice alleviated age-related diseases and
increased the healthy life span (5,23). However, recent studies re-
ported that P16
INK4a
-induced senescence in human cells or tissues
Fig. 1. hNKG2D-CAR T cells eliminate human cells undergoing senescence induced by dierent stresses. (A) Flow cytometric analysis of NKG2DL (MICA/MICB and
ULBP1 to ULBP3) expression on IMR90 cells undergoing senescence (SEN) induced by etoposide (IMR90-Et, 50 μM), overexpression of oncogene Kras (IMR90-Kras), or
replicative exhaustion (IMR90-Rep). Cells treated with DMSO, cells expressing empty lentiviral vector, or cells at a population doubling level (PDL) of 25 were defined as
controls (CON), respectively. The percentage of positive cells compared with the controls is shown in each histogram. All NKG2DL-specific antibodies were allophyco-
cyanin (APC)–conjugated. The results are representative of N= 2 independent experiments, each performed in triplicate.(B) Cytotoxicity of the mock- and hNKG2D-CAR T
cells to the indicated cells at an E:T ratio of 1:2, 1:1, and 2:1 after 8 hours of coculture. The results are representative of N= 2 independent experiments, each performed in
triplicate. (Cand D) The concentrations of cytokines TNF-α and IFN-γ (C) and the effector molecules granzyme B and perforin (D) released by mock- and hNKG2D-CAR T
cells were quantified. The supernatant was collected and measured by ELISAs after 8 hours of coculture with the indicated cell lines at an E:T ratio of 1:1. The results are
representative of N= 2 independent experiments, each performed in triplicate. Data are presented as the mean ± SD; data were analyzed by two-way ANOVA with
multiple comparisons (B) or one-way ANOVA with multiple comparisons (C and D): **P≤0.01 and ***P≤0.001.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 2 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

does not express sufficient NKG2DLs, allowing the cells to escape
endogenous NK cell–NKG2D–mediated immune surveillance (24).
To investigate whether P16
INK4a
-induced senescence increases
NKG2DL expression and whether this induced expression in senes-
cent human cells is sufficient to be targeted by NKG2D-CAR T cells,
we established three human lung fibroblast cell lines stably express-
ing doxycycline-induced P16
INK4a
and a marker, enhanced green
fluorescent protein (EGFP) (fig. S5A). The induction of P16
INK4a
overexpression resulted in cellular senescence in more than 90%
of the cells, as measured by SA-β-Gal staining in the various
human lung fibroblast cell lines (fig. S5B), identical to the effects
induced by other stressors (fig. S1A). The expression of NKG2DL
mRNA and protein was induced in senescent human lung fibroblast
cells expressing P16
INK4a
, as analyzed by qRT-PCR and flow cytom-
etry analysis (Fig. 2A and fig. S6, A and B). The individual ligands of
NKG2D, namely, MICA, MICB, and ULBP1 to ULBP3, induced by
P16
INK4a
, as analyzed through flow cytometry, are generally lower in
comparison with those induced by DNA damage drug etoposide
and oncogene Kras. However, the expression of ULBP2 is enhanced
in senescent human lung fibroblast cells induced by P16
INK4a
(Figs.
1A and 2B). These findings suggest that different stressors may elicit
diverse NKG2D ligands. Because NKG2D can recognize all types of
ligands expressed on the cell surface (25), we used a recombinant
hNKG2D Fc chimera protein to detect total NKG2DLs. We ob-
served that overexpression of P16
INK4a
led to 42.7 to 62.1% of
human lung fibroblasts expressing NKG2DLs (fig. S6B). Consistent
with the increased expression of NKG2DLs, NKG2D-CAR T cells
efficiently eliminated senescent human cells expressing P16
INK4a
(Fig. 2B and fig. S6C) and produced increased concentrations of cy-
tokines (TNF-α and IFN-γ) and effector molecules (granzyme B
and perforin) (P< 0.01; Fig. 2, C and D). These data together
support that NKG2D-CAR T cells effectively eliminate senescent
human cells expressing NKG2DLs upon various aging-associated
stress challenges in vitro.
Fig. 2. hNKG2D-CAR T cells deplete senescent human cells expressing NKG2DLs induced by exogenous expression of P16
INK4a
.(A) Flow cytometric analysis of
NKG2DL (MICA/MICB and ULBP1 to ULBP3) expression on senescent IMR90 and HEL1 cells (human lung fibroblasts) with P16
INK4a
expression induced by doxycycline
(+DOX) compared to the corresponding control cells (−DOX). The percentage of positive cells is shown in each histogram. All NKG2DL-specific antibodies were APC-
conjugated. The results are representative of N= 2 independent experiments. (B) Cytotoxicity of mock- and hNKG2D-CAR T cells to the IMR90 and HEL1 cells treated with
or without doxycycline at an E:T ratio of 1:2, 1:1, and 2:1 after 8 hours of coculture. The results are representative of N= 2 independent experiments, each in triplicate. (C
and D) The concentrations of cytokines TNF-α and IFN-γ (C) and effector molecules granzyme B and perforin (D) were quantified in the coculture supernatant. Mock- or
hNKG2D-CAR T cells were cocultured without (−) DOX or with (+) DOX–treated human lung fibroblasts (IMR90 and HEL1) at an E:T ratio of 1:1 for 8 hours, and the
supernatant was collected and measured by ELISAs. The results are representative of N= 2 independent experiments, each in triplicate. Data are presented as the
mean ± SD; data were analyzed by two-way ANOVA with multiple comparisons (B) or one-way ANOVA with multiple comparisons (C and D): ***P≤0.001.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 3 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Elimination of senescent cells by mNKG2D-CAR T cells
delays aging-associated phenotypes in irradiated mice
To investigate the effects of NKG2D-CAR T cells on age-related pa-
thologies in mice, we developed a retroviral vector containing the
extracellular domain of mouse NKG2D (mNKG2D), CD137, and
CD3ζ signaling domains to generate mNKG2D-CAR T cells
through retrovirus-mediated transduction of mNKG2D-CAR into
mouse spleen T cells (fig. S7A).
Initially, to determine the capacity of mNKG2D-CAR T cells to
kill senescent cells in vitro, we treated mouse embryonic fibroblasts
(MEFs) from C57BL/6 mice with etoposide (50 μM for 36 hours) or
radiation [10 J/m
2
ultraviolet-C (UV-C)] to induce cellular senes-
cence, as evidenced by elevated SA-β-Gal staining (fig. S7B) and in-
creased mRNA expression of senescence markers, including
P16
INK4a
,P21, and Pai-1 in senescent MEFs (fig. S7C). mRNA ex-
pression of mouse NKG2DLs, such as Rae-1 and Mult-1, was signif-
icantly enhanced in drug- or irradiation-induced senescent cells (P
< 0.001; fig. S7D), implying that senescent MEFs expressing
mNKG2DLs could be selectively targeted by mNKG2D-CAR T
cells. Subsequently, we cocultured mNKG2D-CAR T cells with
control or senescent MEFs to evaluate the cytotoxicity of CAR T
cells. The results demonstrated that mNKG2D-CAR T cells effi-
ciently destroyed senescent MEFs, but not control normal MEFs,
in a dose-dependent manner (fig. S7, E and F). Accordingly, in-
creased mRNA expression of the cytokines Ifng,Tnfα, Il2, and Il6
was observed after mNKG2D-CAR T cell coculture (fig. S7G).
These findings indicate that mNKG2D-CAR T cells can selectively
and efficiently deplete senescent MEFs expressing NKG2DLs
in vitro.
In our study, we observed that NKG2D-CAR T cells also had a
moderate impact on the lysis of uninduced cells. This effect was ac-
companied by increased production of TNF-α, IFN-γ, granzyme B,
and perforin, in comparison with mock-CAR T cells (Figs. 1 and 2
and fig. S7E). These effects might be attributed to a basal degree of
cell culture–induced stresses (CCISs) in primary or nonimmortal-
ized cells, resulting in elevated NKG2DLs and the onset of cellular
senescence (26). Consequently, NKG2D-CAR T cells could poten-
tially delete the baseline of the NKG2DL-expressing cells induced
by CCIS, even in the absence of exogenous senescence inducers.
To validate this possibility, we also examined cultured primary
MEF cells that underwent senescence at different passages (27).
We found a modest degree of senescence and up-regulated
NKG2DL expression at passage 7 (fig. S8, A and B). Corresponding-
ly, mNKG2D-CAR T cells showed modest cytotoxicity at passage 7,
whereas no cytotoxicity was observed at passages 0 and 3, where no
senescent cells were present (fig. S8, C and D).
To evaluate the in vivo effects of mNKG2D-CAR T cells on sen-
escent cells during aging, we used a mouse aging model accelerated
by x-ray irradiation (4 Gy), which shares many age-related features
found in humans (28). Three months after radiation, the hair of the
mice turned white, and the proportion of white hairs increased over
time (fig. S9A). Increased mRNA expression of P16
INK4a
was de-
tected in multiple tissues at 3 and 11 months after irradiation (fig.
S9B). Expression of the NKG2DLs, including Rae-1,Mult-1, and
H60b, was also up-regulated in tissues from irradiated mice and
generally displayed identical trends to P16
INK4a
(fig. S9, C to E).
At 11 months after irradiation, 1 × 10
6
mNKG2D-CAR T cells en-
gineered from inbred donors were infused through the tail vein.
One month after cell infusion, mRNA expression of NKG2DLs,
including Rae-1,Mult-1,orH60b, was reduced in all the assessed
tissues [adipose, skeletal muscle (SkM), heart, lung, liver, and
kidney] from NKG2D-CAR T cell–treated mice (Fig. 3A and fig.
S10A). These data suggest that NKG2D-CAR T cells efficiently
eliminate cells expressing NKG2DLs in various tissues in vivo.
To assess the efficacy of NKG2D-CAR T cells in depleting sen-
escent cells in vivo, we chose senescence-sensitive inguinal adipose
tissue (IAT) from x-ray irradiation–induced aging mice. The whole-
mount SA-β-Gal staining of adipose tissues in mice treated with
mNKG2D-CAR T cells showed a weaker blue compared with
mice treated with mock-CAR T cells (Fig. 3B), indicating the effec-
tive removal of senescent cells from adipose tissues by mNKG2D-
CAR T cells. In addition, mRNA expression of senescence markers
P16
INK4a
and P21, as well as senescence-associated secretory phe-
notype (SASP)–related cytokines, such as Il6,Pal1,Mmp3, or
Igfbp2, were reduced in all tested tissues in mNKG2D-CAR T
cell–treated mice (Fig. 3C and fig. S10B). We observed no signifi-
cant body weight loss in the CAR T cell–treated mice (P> 0.05;
fig. S10C). These results support that mNKG2D-CAR T cells effec-
tively eliminate senescent cells in vivo without any observable
adverse effects in an accelerated aging mouse model.
Given that age-dependent loss of adipose tissue is a widely used
marker inversely correlated with aging in rodents (5,29,30), we ex-
amined IAT histology, adipocyte size, and abdominal adipose tissue
volume to assess the aging-associated alterations in mice. The mean
area of adipocytes from the mNKG2D-CAR T cell–treated mice was
significantly higher than that of adipocytes from the mock-CAR T
cell–treated mice (P< 0.05; Fig. 3D). Correspondingly, IAT weights
from mNKG2D-CAR T cell–treated mice increased (Fig. 3E). We
then used microcomputed tomography (micro-CT) to evaluate
the effects of mNKG2D-CAR T cells on adipose tissue volume
and age-associated osteoporosis in vivo. Consistent with our find-
ings in IAT, mNKG2D-CAR T cell treatment resulted in a signifi-
cant increase in subcutaneous adipose tissue (SAT) (P< 0.05;
Fig. 3F). Although mNKG2D-CAR T cell–treated mice had no sig-
nificant (P= 0.081) increases in bone density (Fig. 3G), the
mNKG2D-CAR T cells significantly attenuated age-related trabec-
ular bone loss in cancellous bone of the distal femur (P< 0.05;
Fig. 3H). This phenotype may be caused by an increase in osteo-
blasts and a decrease in osteoclasts after the clearance of senescent
cells from the bone microenvironment (9,31). We also observed a
significant increase in gastrocnemius myofiber size after 6 months
of treatment with mNKG2D-CAR T cells (P< 0.05; Fig. 3I), suggest-
ing an improvement in muscle aging and sarcopenia in aged mice
and attenuation of age-related muscle atrophy upon mNKG2D-
CAR T cell challenge. Considering physical functions with im-
proved skeletal and muscle tissues (Fig. 3, H and I), the mice
treated with mNKG2D-CAR T cells exhibited significantly better
performance regarding maximal treadmill endurance, grip strength,
and walking speed than mice treated with mock-CAR T cells (P<
0.05; Fig. 3, J to L). These findings suggest that mNKG2D-CAR T
cells have the potential to be an effective therapeutic intervention
for age-related diseases.
Elimination of senescent cells by mNKG2D-CAR T cells
delays aging-associated phenotypes in naturally aged mice
Our findings support that mNKG2D-CAR T cells could slow down
irradiation-induced accelerated aging. To explore whether these
CAR T cells could also be effective in naturally aged mice, we
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 4 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Fig. 3. mNKG2D-CAR T cells deplete senescent cells expressing NKG2DLs and alleviate aging-associated pathologies in irradiation-accelerated aged mice. (A)
The mRNA expression of Rae-1,Mult-1, and H60b in IATand SkM tissue samples was detected by qRT-PCR analysis. Tissues were collected from mice treated with mock- or
mNKG2D-CAR T cells 1 month after cell infusion. Each column on the graph represents an individual mouse (n= 5). The results are representative of N= 2 independent
experiments. Each gene was analyzed in triplicate. (B) Representative images of IAT stained with SA-β-Gal (left) and quantification of the average staining intensity using
ImageJ software (right). IAT was collected from the mice treated with mock- or mNKG2D-CAR T cells 1 month after cell infusion. Each IAT sample was collected from an
individual mouse (n= 5). (C) The mRNA expression of senescence markers (P16
INK4a
and P21) and SASP-related cytokines (Il-6,Pal-1,Mmp3, and Igfbp2) was detected by
qRT-PCR analysis of IAT and SkM tissue. The tissues were collected from mice treated with mock- or mNKG2D-CAR T cells 1 month after cell infusion. Each column
represents an individual mouse (n= 5). The results are representative of N= 2 independent experiments. Each gene was analyzed in triplicate. (D) Left: Representative
images of H&E staining of IAT samples collected at different time points fromthe mice treated with mock- or NKG2D-CAR T cells. Right: Quantificationof relative adipocyte
areas in the IAT samples. The data were normalized to those of the mock group. Each dot represents an individual mouse (n= 5). Scale bar, 50 μm. (E) Weight of IAT
samples collected from mice treated with mock- or NKG2D-CAR T cells 1 month after cell infusion. The data are normalized to the mouse body weight. (F) Left: Micro-CT
images of SAT of mice treated with mock- or mNKG2D-CAR T cells at 6 months after cell infusion. Green, subcutaneous adipose tissue (SAT); red, bone. Right: Quantifi-
cation of the adipose tissue of mice based on the micro-CT images from the left panel. The data were normalized to the mouse body weight. Each dot represents an
individual mouse (n= 5). (G) The bone density of the femur was quantified and calculated withcaliper analysis software on the basis of the micro-CT image from (F). Each
dot represents an individual mouse (n= 5). (H) Left: Representative micro-CT images of the distal femurs of the mice treated with mock- or mNKG2D-CAR T cells at 6
months after cell infusion. Right: The relative bone volume fraction [bone volume (BV)/total volume (TV)] of cancellous bone in the distal femur was quantified from the
left panel and calculated with caliper analysis software. Each dot represents an individual mouse (n= 5). (I) Top: Representative images of H&E staining of gastrocnemius
collected at different time points from the mice treated with mock- or NKG2D-CAR T cells at indicated time point. Scale bar, 50 μm. Bottom: Relative cell areas in the
gastrocnemius were quantified from the top panel. The data were normalized to those of the mock group. Each dot represents an individual mouse (n= 5). (Jto L)
Treadmill endurance (J), grip strength (K), and maximal walking speed (L) of the irradiation-induced aging mice treated with mock- or mNKG2D-CAR T cells were recorded
at 6 months after cell infusion. Each dot represents an individual mouse (n= 5). Data are presented as the mean ± SD; data were analyzed by unpairedtwo-tailed Student’s
ttests (B and E to L) or Mann-Whitney Utests (D and I): *P≤0.05; **P≤0.01; ***P≤0.001; ns, not significant.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 5 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

infused mNKG2D-CAR T cells into 24-month-old mice through
the tail vein. Although only 1 × 10
6
NKG2D CAR T cells were
infused, CAR T cells expanded for 4 weeks and still could be detect-
able at 15 weeks after CAR T cell infusion (Fig. 4A). The infused
CAR T cell expansion resulted in a slight rise in serum interleu-
kin-6 (IL-6) and IFN-γ 14 days after treatment, but they returned
to baseline by day 28 (Fig. 4B). Hypothermia, which was caused
by high-dose NKG2D-CAR T cells in the previous study (32,33),
was not observed in our experiments (fig. S11A). The body
weight loss in the aged mice was alleviated after 1 month of CAR
T treatment (Fig. 4C). The NKG2D-CAR T cells did not produce
Fig. 4. mNKG2D-CAR T cells deplete senescent cells expressing NKG2DLs in naturally aged mice. (A) The persistence of mNKG2D-CAR T cells was detected by qRT-
PCR analysis of the copy number of mNKG2D CAR in blood samples collected from the naturally aged mice treated with mock- or NKG2D-CAR T cells at indicated time
points. The results are representative of n= 3 mice per group from two independent experiments. (B) The concentration of cytokines IL-6 and IFN-γ in blood samples
collected from the naturally aged mice was evaluated by ELISAs. Each dot represents an individual mouse (n= 5). The results are representative of N= 2 independent
experiments. (C) The weight change of the naturally aged mice treated with mock- or NKG2D-CAR T cells was recorded. n= 5 mice per group. (D) The mRNA expression of
mouse NKG2DLs (Rae-1,Mult-1, and H60b) was detected by qRT-PCR analysis of the indicated tissues collected from the mice treated with mock- or mNKG2D-CAR T cells
at day 21 after cell infusion. Each row represents an individual mouse (n= 5). The results are representative of N= 2 independent experiments. Each gene was analyzed in
triplicate. (E) Total NKG2DL expression in liver and lung tissues collected from the mice treated with mock- or mNKG2D-CAR T cells at day 21 after cell infusion was
detected by flow cytometric analysis using a recombinant mouse NKG2D Fc chimera protein. Each dot represents an individual mouse (n= 5). (F) The mRNA expression
of senescence markers was detected by qRT-PCR analysis of the indicated tissues. Each row represents an individual mouse (n= 5). The results are representative of N= 2
independent experiments. Each gene was analyzed in triplicate. (G) Representative images of IAT with SA-β-Gal staining (left) and quantification of the average staining
intensity using ImageJ software (right). IAT was collected from the mice treated with mock- or mNKG2D-CAR T cells at day 21 after cell infusion. Each IAT sample was
obtained from an individual mouse (n= 5). (H) Full scans of SA-β-Gal staining in the liver and lung collected from the naturallyaged mice treated with mock- or mNKG2D-
CAR T cells. Scale bar, 500 μm. (I) Representative images of Flag (red, C terminus–tagged CAR), CD3 (green, T cells), DAPI (blue, DNA), and SA-β-Gal staining (blue green,
senescence) for liver and lung samples collected from the mice treated with mock- or mNKG2D-CAR T cells at day 21 after cell infusion. Scale bar, 10 μm. Right: Quan-
tification of the immunofluorescence. Data are presentedas the mean ± SD; data were analyzed by unpaired two-tailed Student’sttests (B, E, and G), two-way ANOVAwith
multiple comparisons (C), or Mann-Whitney Utests (I): *P≤0.05, **P≤0.01, and ***P≤0.001.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 6 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

any significant changes in the hematological parameters (P> 0.05;
fig. S11B).
To determine whether NKG2D-CAR T cells can reduce senes-
cent cell abundance in various tissues in naturallyaged mice, we an-
alyzed the IAT, SkM, liver, and lung in the mice treated with mock-
or NKG2D-CAR T cells for 21 days. The NKG2D-CAR T cell treat-
ment resulted in reduction in the mRNA expression of NKG2D
ligands (Fig. 4D), a marked reduction of NKG2DL-positive cells
in liver and lung confirmed by flow cytometry (P< 0.001;
Fig. 4E), and a decrease in the expression of the senescence
markers (Fig. 4F). This reduction was accompanied by a decrease
in SA-β-Gal staining intensity (Fig. 4, G and H). NKG2D-CAR T
cells were detected around senescent cells after 21 days upon CAR
T cell infusion in liver and lung (Fig. 4I), indicating that these CAR
T cells effectively infiltrate senescent tissues and potentially delete
senescent cells in naturally aged mice. Moreover, the infiltration of
CAR T cells was also observed in the IAT samples collected from the
CAR T cell–treated aged mice. In contrast, only a small number of
wild-type T cells and no Flag-positive cells were detected in the IAT
in control mice (fig. S11C). These results together support the effec-
tiveness of mNKG2D-CAR T cells in targeting and eliminating sen-
escent cells in naturally aged mice.
We then proceeded to investigate whether the administration of
NKG2D-CAR T cells could alleviate age-related pathologies 6
months after treatment. Blood biochemical analysis indicated im-
provements in liver-, kidney-, heart-, and metabolism-related bio-
chemical markers (Fig. 5A). The mean area of adipocytes from the
CAR T cell–treated mice was significantly larger compared with
Fig. 5. mNKG2D-CAR T cells alleviate aging-associated pathologies and improve physical function in naturally aged mice. (A) Quantification of metabolic
enzymes and substrates reflecting liver, heart, and kidney function. Blood biochemical analysis was performed to determine the concentrations of aspartate aminotrans-
ferase (AST), alanine aminotransferase (ALT), total protein (TP), albumin (ALB), globulin (GLB), cholesterol (CHOL), triglyceride (TG), glucose (GLU), blood urea nitrogen
(BUN), creatinine (CRE), and creatine kinase (CK) in the naturally aged mice treated with mock- or mNKG2D-CAR T cells. The data were normalized to those of the young
group. Each dot represents an individual mouse (n= 5). (B) Left: Representative images of H&E staining of IAT samples collected from aged mice treated with mock- or
NKG2D-CAR T cells. Quantification of relative adipocyte areas in the IAT samples from the left panel are shownon the right. Right: The data were normalized to those of the
mock group. Young mice were used as control. Each dot represents an individual mouse (n= 6). Scale bar, 50 μm. (C) Micro-CT images of the SAT of the mice treated with
mock- or mNKG2D-CAR T cells at 6 months after cell infusion (left). Green, SAT; red, bone. Right: Quantification of the adipose tissue of mice based on the micro-CT images
from the left panel. The data were normalized to the mouse body weight. Each dot represents an individual mouse (n= 5). Twelve-month-old (young) mice were used as
positive controls. (D) Left: Representative micro-CT images of the distal femurs of the mice treated with mock- or mNKG2D-CAR T cells at 6 months after cell infusion. The
relative bone volume fraction (BV/TV) of cancellous bone in the distal femur was quantified from the left panel and calculated with caliper analysis software (right). Each
dot represents an individual mouse (n= 5). (Eto H) Treadmill endurance (E), maximal walkingspeed (F), grip strength (G), and hanging endurance (H) of the naturally aged
mice treated with mock- or mNKG2D-CAR Tcells were recorded at 6 months after cell infusion. Each dot represents an individual mouse (n= 5). Data are presented as the
mean ± SD; data were analyzed by Mann-Whitney Utests (A) or unpaired two-tailed Student’sttests (B to H): *P≤0.05, **P≤0.01, and ***P≤0.001.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 7 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

mock-CAR T cell–treated mice (P< 0.01; Fig. 5B), and micro-CT
revealed corresponding increases in SAT (Fig. 5C). The results of
SA-β-Gal staining revealed a decreasing trend in the abundance
of senescent cells in IAT samples obtained from aged mice 6
months after NKG2D-CAR T cell treatment. However, the observed
decrease in senescent cell abundance was not statistically significant
(P= 0.2388; fig. S12). Consistent with the results observed in irra-
diated mice, mNKG2D-CAR T cell treatment resulted in an increase
in trabecular bone in naturally aged mice (Fig. 5D). Although there
was not an increase in muscle fiber after treatment with mNKG2D-
CAR T cells (fig. S13), the mice showed significant improvements in
physical functions (P< 0.05; Fig. 5, E to H). Collectively, these
results support the idea that NKG2D-CAR T cells can serve as a
potent senolytic therapy by targeting senescence-driven aging and
age-associated pathologies in vivo.
Autologous T cells armed with hNKG2D-CAR eciently
eliminate senescent cells in naturally aging nonhuman
primates without observable systemic toxicity
Rhesus and cynomolgus macaques, two species of nonhuman pri-
mates, are more closely related to humans and recapitulate the
salient features of the human aging process and development of
age-associated diseases (34). The extracellular domain of NKG2D
in these macaques is almost identical, where only two amino
acids are different from hNKG2D (fig. S14). This almost-identical
extracellular domain of NKG2D indicates that hNKG2D-CAR T
cells might recognize and eliminate senescent cells expressing
NKG2DLs in nonhuman primates. To test this possibility in vitro,
we treated primary T cells from a rhesus macaque with the DNA
damage drug etoposide to induce cellular senescence, which was
confirmed by flow cytometric analysis of SA-β-Gal staining up to
70% (fig. S15A). Concomitantly, the protein expression of
NKG2DLs was highly up-regulated, as measured by flow cytometric
analysis using a recombinant hNKG2D Fc chimera protein (fig.
S15B). These senescent T cells expressing NKG2DLs were
Fig. 6. Autologous T cells armed with hNKG2D-CAR eliminate senescent cells in naturallyaging nonhuman primates. (A) The persistence of hNKG2D-CAR T cells in
nonhuman primates treated with a dose of 1 × 10
6
cells/kg was determined by analyzing the copy number of hNKG2D-CAR using qRT-PCR in blood samples collected
from identified individuals: 98106, 00065, 00085, 00102, and 01102. The data shown represent results from N= 2 independent experiments, each performed in triplicate.
(B) Concentrations of IFN-γ and IL-6 in blood samples collected from the nonhuman primates treated with hNKG2D-CAR T cells were evaluated at the indicated times
using MesoScale Discovery V-Plex assay kits. (C) The fold changes in mRNA expression of senescence markers in SAT, skeletal muscle (SkM), liver, and kidney collected
from the nonhuman primates before and 2 months after hNKG2D-CAR T cell treatmentwere determined by qRT-PCR analysis. The data shown represent results from N= 2
independent experiments, each performed in triplicate. (D) Representative images of SAT with SA-β-Gal staining. SAT was collected from the nonhuman primates before
and 2 months after hNKG2D-CAR T cell treatment and stainedwith SA-β-Gal. The average staining intensity values were calculated using ImageJ software. (Eand F) Body
weight changes (E) and temperature changes (F) in nonhuman primates treated with NKG2D-CAR T cells weremonitored and recorded atthe indicated time points. Data
are presented as the mean ± SD; data were analyzed by unpaired two-tailed Student’sttests: *P≤0.05.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 8 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

efficiently lysed in a dose-dependent manner when incubated with
autogenous T cells armed with hNKG2D-CAR (fig. S15, C and D).
Furthermore, this lysis was accompanied by a significant increase in
the concentrations of the cytokines IFN-γ and IL-6 (P< 0.001; fig.
S15E). These results from the cellular model support the conserva-
tion between the human and macaque NKG2DL-NKG2D axis and
suggest that hNKG2D-CAR T might eliminate NKG2DL-express-
ing senescent cells in aged macaques.
To investigate whether hNKG2D-CAR T cells diminish naturally
occurring senescent cells in different tissues in vivo, we carried out
the experiments using three rhesus macaques and two cynomolgus
macaques at 19 to 22 years of age, approximately corresponding to
60 years in humans (35). Autologous T cells armed with hNKG2D-
CAR were generated with a transduction efficiency of 18 to 45% (fig.
S16) and administered to primates at a dose of 1 × 10
6
cells/kg to
evaluate the safety and capacity of the hNKG2D-CAR T cells. RT-
PCR quantification of lentiviral DNA in the blood showed that the
CAR T cells were mildly expanded in vivo, peaking at 18 days after
CAR T cell autotransplantation (Fig. 6A). IFN-γ and IL-6 were gen-
erally maintained at low concentrations in the blood after
hNKG2D-CAR T cell treatment (Fig. 6B). Although IL-6 in one
of nonhuman primates (00085) was elevated (about 22 pg/ml) at
day 21, it was still below the threshold known to trigger cytokine
release syndrome (36), and subsequently, IL-6 was restored to a
low concentration. SAT, SkM, liver, and kidney were used to evalu-
ate the mRNA expression of senescence markers at 2 months after
CAR T cell treatment (Fig. 6C). P16
INK4a
expression was reduced in
all the tested samples, and the expression of P14,P21, and SASP
genes, including LGFBP2,IL6,MMP3, and CXCL1, was decreased
in most of the samples (Fig. 6C). Given that adipose tissue is highly
susceptible to senescence with aging (5,37), we used adipose tissue
to assess senescence status affected by NKG2D-CAR T cells in vivo.
Adipose tissues from all five aged nonhuman primates showed
intense blue staining for SA-β-Gal activity before NKG2D-CAR T
cell therapy. In contrast, the SA-β-Gal–positive cells were reduced in
SAT 2 months after NKG2D-CAR T cell infusion, as reflected in the
reduction in blue staining of all five adipose tissues (Fig. 6D). These
data demonstrate that hNKG2D-CAR T cells efficiently deplete sen-
escent cells in different tissues of aged nonhuman primates.
During the initial 2 weeks, there was an apparent reduction in
food intake (fig. S17). However, this may not be attributed to
CAR T therapy because the downward trend started before treat-
ment. Accordingly, the nonhuman primates except one (00065) ex-
perienced temporary weight loss but subsequently recovered
(Fig. 6E). None of the five nonhuman primates showed aberrant
changes in temperature during the treatment period (Fig. 6F). No
abnormal changes were observed in routine blood tests after
NKG2D-CAR T cell treatment (fig. S18A). Blood samples were
also analyzed for metabolic enzymes and substrates to assess liver,
heart, and kidney functions. Although slight changes in some pa-
rameters, such as alanine aminotransferase (ALT) and aspartate
aminotransferase (AST), were observed, they were all within
normal ranges (fig. S18B). In conclusion, our data support that
hNKG2D-CAR T cells function as a senolytic agent to effectively
deplete the in vivo senescent cells without causing any severe side
effects in aged nonhuman primates.
DISCUSSION
Experimental studies and clinical evidence have established a solid
link among cellular senescence, senescent cell accumulation, and
the secretion of SASP components with various pathologies, such
as obesity, diabetes, fibrosis, cardiovascular disease, cancer, and
other age- and chronic disease–related organ dysfunction (38–40).
Thus, clearance of senescent cells may ameliorate age- and chronic
disease–associated disorders (3,38,41). Clinical trials have thus far
focused on safety, target engagement, and efficacy of chemical-
based senolytic agents, such as a cocktail of the Src kinase inhibitor
dasatinib and the flavonoid quercetin and, subsequently, BCL-2
family inhibitors, which clear senescent cells (42,43). However,
these combined senolytic drugs predominantly inhibit anti-apopto-
tic pathways, raising the potential concern of off-target effects in
normal tissues (28,44–46).
Accumulating evidence suggests that improvements of existing
senolytics, particularly in selectivity, are possible through immuno-
therapy. Vaccines targeting GPNMB or CD153 have been shown to
reduce senescent cells expressing GPNMB or CD153 and alleviate
multiple aging-associated pathologies (16,17). Senolytic uPAR-
CAR T cells provide selectivity by deleting uPAR-expressing senes-
cent lung adenocarcinoma cells induced by MEK and CDK4/6 in-
hibitors and senescent cells in liver fibrosis induced by carbon
tetrachloride or a high-fat diet (15). Therefore, the development
of immunotherapies that specifically target senescent cells express-
ing cell surface proteins could provide a selective and effective ap-
proach for the treatment of aging and age-related pathologies.
Ligands for NKG2D are expressed in cancer cells and are typi-
cally absent in normal cells (47), making them promising CAR
targets. NKG2D-CAR T cells have been specifically targeted to
cancer cells expressing NKG2DLs in clinical trials. Unfortunately,
the limited clinical efficacy may be due to the shedding of
NKG2DLs by cancer cells or limited function of CAR T cells
under the complex tumor microenvironment (48,49). Although
it was previously reported that senescent cells escape from endoge-
nous NKG2D-mediated NK cell immunosurveillance during aging
(18), our results demonstrate that senescent cells expressing
NKG2DLs in vivo can be selectively and effectively eliminated by
the adoptive transfer T cells armed with NKG2D-CARs. Recent
studies have found higher expression of NKG2DLs (ULBP2 and
MICA) in senescent fibroblasts present in the skin of older individ-
uals in comparison with younger individuals (50). Our studies in-
dicate that senescent cells expressing NKG2DLs in aged humans
could potentially be targeted using NKG2D-CAR T cells.
Currently, NKG2D-based CAR T cells have been tested in
various types of cancer in clinical trials and have shown a favorable
safety and tolerability profile in humans (51,52). No dose-limiting
toxicities, cytokine release syndrome, or autoimmune reactions
were observed until the dose was increased to 3 × 10
8
CAR T cells
per injection every 2 weeks for three administrations, all of which
resolved with standard treatment (52–54). In our study, NKG2D-
CAR T cells efficiently eliminated senescent cells and ameliorated
age-associated functional decline of organs in both irradiation-
induced and naturally aged mice without observable side effects.
When treated with a dose of 1 × 10
6
/kg, the nonhuman primates
only experienced transient weight loss but recovered thereafter
without any additional treatment. Together, our cell-based studies
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 9 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

and animal experiments strongly support that engineered NKG2D-
CAR T cells function as selective, effective, and safe senolytics.
Although NKG2D-CAR T cells have shown a favorable safety
profile in cancer clinical trials, it is essential to thoroughly investi-
gate their safety and potential toxicities in older individuals through
clinical studies. NKG2DLs are highly expressed in senescent cells
from aged individuals (50), and our studies demonstrated that
NKG2D-CAR T cells do not exhibit overt toxicity in aged mice
and nonhuman primates. These findings suggest that the adoptive
transfer of engineered T cells armed with NKG2D-CAR, specifically
targeting naturally occurring senescent cells expressing NKG2DLs,
could have an impact on the field of geriatric medicine.
Although CAR T cell therapy has shown promising outcomes in
clinical trials, it is important to acknowledge the limitations of its
application in humans. NKG2D-CAR T cells are designed to iden-
tify and eradicate cells expressing NKG2DLs, including senescent
cells. However, it is worth noting that NKG2DLs can be induced
on cells during instances of inflammation or specific bacterial and
viral infections (55). Consequently, NKG2D-CAR T cells may rec-
ognize and eliminate these cells, potentially exacerbating tissue
damage and inflammation, leading to unintended adverse effects
in clinical settings. Therefore, it is crucial to thoroughly evaluate
the potential risks associated with tissue damage when considering
the use of NKG2D-CAR T cells in older individuals or in individ-
uals with inflammatory and infectious diseases. In addition, patients
may require repeated administrations of the therapy because of the
short-lived effects of CAR T cell therapy. Despite observing a
decline in the abundance of senescent cells in IAT samples obtained
from aged mice 6 months after NKG2D-CAR T cell treatment, the
absence of statistical significance suggests the possibility of senes-
cent cells potentially reappearing within or around the 6-month
time frame. However, there is encouraging evidence indicating
that the persistence of CD4
+
CAR T cells in patients treated with
CD19-CAR T cells has resulted in decade-long leukemia remissions
(56). This finding implies that the clonal makeup of CAR T cells can
be stabilized in vivo over the long term, providing valuable insights
into the therapy’s potential long-term efficacy.
MATERIALS AND METHODS
Study design
The goals of this study were to investigate whether NKG2DLs are
up-regulated in senescent cells induced by various aging-related
stresses and to determine the efficacy of NKG2D-CAR T cells in
eliminating senescent cells expressing NKG2DLs and leading to re-
duction in aging-related pathologies and improvement in physical
performance in aged animal models. The study objectives were ac-
complished by (i) detecting mRNA and protein expression of
NKG2DLs in senescent cells induced by different types of stress
in vitro, as well as in various tissues of aged mice and nonhuman
primates; (ii) generating NKG2D-CAR T cells and determining
their effectiveness in eliminating senescent cells in mice experienc-
ing either irradiation-induced or naturally occurring aging and aged
nonhuman primates; and (iii) evaluating aging-associated patholo-
gies, physical performance, and potential toxicity in the aged animal
models after treatment with NKG2D-CAR T cells. For all experi-
ments, mice were randomly assigned to either the control group
or NKG2D-CAR T cell treatment group. In all in vivo experiments,
no animals were excluded from the analysis. Sample sizes for the
experimental groups were determined on the basis of previous
reports and calculations. All pathology analyses were performed
blindly by a pathologist. Detailed information regarding the
number of replicates, the statistical test used, and the corresponding
Pvalues are provided in the figure legends.
Cell culture
The human lung fibroblast cell lines IMR90 (catalog number CCL-
186) and WI38 (catalog number CCL-75) were purchased from the
American Type Culture Collection (ATCC), and the HEL1 cell line
was purchased from the China Infrastructure of Cell Line Resources
(catalog number KCB 86024). Cells were authenticated by short
tandem repeat profiling and tested negative for mycoplasma con-
tamination. Cells were cultured in Eagle’s minimum essential
medium (Thermo Fisher Scientific) containing 10% fetal bovine
serum (Thermo Fisher Scientific), penicillin (100 U/ml), and strep-
tomycin (100 mg/ml) (Thermo Fisher Scientific). Human embry-
onic kidney 293T (catalog number CRL-3216) virus packaging
cells were purchased from ATCC and cultured in Dulbecco’s mod-
ified Eagle’s medium (DMEM; Thermo Fisher Scientific) with 10%
fetal bovine serum. MEFs were isolated from E13.5 C57BL/6 mouse
embryos and cultured in DMEM with 10% fetal bovine serum, 2
mM glutamine, penicillin (100 U/ml), and streptomycin (100 mg/
ml). All the cells were cultured in a humidified incubator with 5%
CO
2
at 37°C.
NKG2D-CAR design and construction
The sequence of hNKG2D-CAR elements, comprising the extracel-
lular domain of hNKG2D, CD137, and CD3ζ signaling domains,
was described previously (57). The CAR elements were cloned
into the lentiviral expression vector pTomo (catalog number
26291, Addgene) between the Xba I and Sal I restriction sites.
The mNKG2D CAR elements, comprising the extracellular
domain of mouse NKG2D, mouse CD137, and mouse CD3ζ signal-
ing domains, were subcloned into the retroviral vector murine stem
cell virus (MSCV)–internal ribosomal entry site (IRES)–GFP
(catalog number 20672, Addgene) between the Eco RI and Sal I re-
striction sites. NKG2D-CAR vectors lacking the extracellular
domain of NKG2D were defined as mock vectors.
CAR T cell preparation and expansion
Human T cells were isolated from the peripheral blood of healthy
donors using RosetteSep Human T Cell Enrichment Cocktail
(15021, STEMCELL Technologies) and activated with CD3/CD28
Dynabeads (11131D, Life Technologies). The protocol was ap-
proved by the Institutional Review Board at Kunming Institute of
Zoology, Chinese Academy of Sciences (approved ID: SMKX-
2019022). The healthy blood donors provided written informed
consent in compliance with the international ethical standards for
biomedical research involving human participants. Macaque T cells
were isolated from peripheral blood using the EasySep Non-Human
Primate T Cell Isolation Kit (19581, STEMCELL Technologies) and
activated with CD2/CD3/CD28 Dynabeads (130-092-919, Miltenyi
Biotec). Mouse T cells were isolated from single-cell suspensions of
splenocytes prepared by mechanical disaggregation using the
EasySep Mouse T Cell Isolation Kit (19851, STEMCELL Technolo-
gies) and activated with CD3/CD28 Dynabeads (11452D, Life Tech-
nologies). All experiments were performed in strict accordancewith
the manufacturer’s instructions.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 10 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Lentivirus expressing human mock or NKG2D-CAR was pre-
pared as described previously (57). Retrovirus expressing mouse
mock- or NKG2D-CAR was produced using pCL-Eco packaging
vector (catalog number 12371, Addgene) and concentrated by poly-
ethylene glycol (Abcam) according to a method described previous-
ly (58). For virus infection, T cells were suspended at 1 × 10
6
cells/ml
in advanced RPMI 1640 medium (Thermo Fisher Scientific) con-
taining 10% fetal bovine serum, 2 mM L-glutamine, and IL-2 (200
U/ml). Forty-eight hours later, the virus was added to the cultures at
a multiplicity of infection (MOI) of 50 in the presence of 1% Lenti-
Boost (Sirion Biotech GmbH). For viral infection of macaque T
cells, an additional 5 μg/ml of cyclosporin A (Sigma-Aldrich) was
added. T cells infected with the mock vector–or NKG2D-CAR–ex-
pressing virus were defined as mock- or NKG2D-CAR T cells, re-
spectively. The cells were cultured at 37°C in a humidified incubator
with 5% CO
2
and transferred to fresh medium every other day. The
cells were counted every other day, and the cell density was main-
tained at 1 × 10
6
cells/ml. For in vitro experiments, the cells were
used 3 days after virus transduction. For in vivo experiments, the
cells were used 7 days after virus transduction.
Senescent cell models
For etoposide-induced senescence, cultured human lung fibroblasts
(IMR90, WI38, and HEL) or MEFs at about 60% confluency were
exposed to 50 μM etoposide for 36 hours. Dimethyl sulfoxide
(DMSO) was used as a vehicle control. For Kras-induced senes-
cence, human lung fibroblasts at about 30% confluency were infect-
ed with a lentivirus produced from pTomo-KrasG12D-EGFP at an
MOI of 50 in the presence of polybrene (8 μg/ml). Lentiviral parti-
cles with the empty vector were used as a control. For P16
INK4a
over-
expression–induced senescence, human lung fibroblasts at about
30% confluency were infected with a lentivirus produced from
pTomo-Teton-P16
INK4a
-T2A-EGFP at an MOI of 50 in the pres-
ence of polybrene (8 μg/ml) and then induced with doxycycline
(1 μg/ml). DMSO or the vehicle of doxycycline was used as a
control. For UV irradiation–induced senescence, MEFs at 60 to
70% confluency were exposed to UV radiation (10 J/m
2
). UV-free
MEFs were used as controls. The cells treated as mentioned above
were cultured in fresh medium for another 8 days. For replicative
exhaustion–induced senescence, human lung fibroblasts at popula-
tion doubling level 25 (PDL25) were cultured until growth ceased
(PDL50 to PDL59). Human lung fibroblast cells at PDL25 were
used as controls. All senescent cell models were verified by SA-β-
Gal staining and then used for experiments.
SA-β-Gal staining
For adipose tissue, frozen tissue section, or adherent cells, SA-β-Gal
activity was evaluated using the Senescence β-Galactosidase Stain-
ing Kit (Cell Signaling Technology). The average staining intensity
values of the whole-mount SA-β-Gal staining of adipose tissues
were calculated using ImageJ software. For cells in suspension,
SA-β-Gal activity was determined using the CellEvent Senescence
Green Flow Cytometry Assay Kit (Thermo Fisher Scientific). All
procedures were carried out in strict accordance with the manufac-
turers’instructions.
Immunouorescence staining
Fixed frozen sections (10 μm) were permeabilized in phosphate-
buffered saline (PBS) containing 0.3% Triton X-100 for 30 min.
Subsequently, the sections were blocked with 10% goat serum at
room temperature for 1 hour. For the double staining of Flag and
CD3, the tyramide signal amplification (TSA) Plus Fluorescence
System (AKOYA) was used according to the manufacturer’s proto-
col. Briefly, the sections were incubated with the primary rabbit
monoclonal antibody anti-Flag (1:500; 14793, Cell Signaling Tech-
nology) at 4°C overnight, followed by incubation with peroxidase-
conjugated anti-rabbit secondary antibody (1:5000; Sigma-Aldrich)
for 2 hours at room temperature. Fluorescent signals were generated
using TSA Plus Cyanine 3. Next, the sections were subjected to mi-
crowave heat treatment and subsequently incubated with the
primary rabbit monoclonal antibody anti-CD3 (1:500; 78588, Cell
Signaling Technology). The same anti-rabbit secondary antibody
was used as described above. Fluorescent signals were generated
using TSA Plus Fluorescein. Last, the sections were stained with
40,6-diamidino-2-phenylindole dihydrochloride (DAPI; Sigma-
Aldrich), mounted using antifade reagent (Abcam), and examined
under a Nikon ECLIPSE Ni-E microscope (Nikon).
Flow cytometry
Cells were harvested, washed twice with 1× PBS, and resuspended in
cold PBS containing 5% bovine serum albumin and 1% sodium
azide (1 × 10
6
cells/ml). Subsequently, labeled primary antibodies
were added to the cell suspension according to the manufacturer’s
instructions, and the cells were incubated for 1 hour at 4°C in the
dark. Immediately after incubation, the cells were washed three
times with ice-cold PBS, and flow cytometry was performed using
a BD LSRFortessa cell analyzer (BD Biosciences). Data were ana-
lyzed using the FlowJo analysis software package (Tree Star Inc.).
The following primary antibodies were used: MICA (1:100;
FAB1300A, R&D Systems), MICB (1:100; FAB1599A, R&D
Systems), ULBP1 (1:100; FAB1380A, R&D Systems), ULBP2/
ULBP5/ULBP6 (1:100; FAB1298A, R&D Systems), ULBP3 (1:100;
FAB1517A, R&D Systems), NKG2D (1:100; 320808, BioLegend),
hNKG2D-Fc chimera (10 μg/ml; 1299-NK-050, R&D Systems),
mouse NKG2D-Fc chimera (10 μg/ml; 139-NK-050, R&D
Systems), CD4 (1:100; 317416, BioLegend), and CD8α (1:100;
301021, BioLegend). The details for the antibodies can be found
in table S1.
Coculture assays
Senescent and corresponding control cells were seeded in 96-well
flat-bottom culture plates in triplicate with a density of 5000 cells
per well in RPMI 1640 with 10% fetal bovine serum, 2 mM gluta-
mine, penicillin (100 U/ml), and streptomycin (100 mg/ml). After
24 hours, effector cells were added to each well at an effector:target
(E:T) ratio of 0.5:1, 1:1, or 2:1. After 8 hours of coculture, the culture
supernatant was collected for measurement of cytokine concentra-
tions. The concentrations of human TNF-α, IFN-γ, granzyme B,
and perforin were measured using ELISA kits (BD Biosciences) ac-
cording to the manufacturer’s instructions. The expression of
mouse cytokines in effector cells isolated from the coculture
system was detected by qRT-PCR analysis. For cytotoxicity assays
in adherent target cells, suspended cells were removed by gentle as-
piration, and the adherent cells were stained with DAPI (1 μg/ml;
Sigma-Aldrich) for 10 min after fixation and counted with the MD
ImageXpress Micro XL High-Content Analysis System (Molecular
Devices). Cytotoxicity assays in suspended target cells were per-
formed using the Cell-Mediated Cytotoxicity Fluorometric Assay
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 11 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Kit (BioVision) according to the manufacturer’s instructions. The
percentage of target cell lysis in wells cocultured with T cells was
calculated according to the following equation: cell lysis (%) =
[(negative control treatment)/negative control] × 100. Negative
control wells had target cells only.
RNA isolation and qRT-PCR analysis
Total RNA was isolated from cells or tissues using TRIzol reagent
(Sigma-Aldrich), and DNA was removed using the TURBO DNA-
free Kit (Thermo Fisher Scientific). Total RNA was verified by spec-
trophotometry (NanoDrop One, Thermo Fisher Scientific) and
agarose gel electrophoresis. RNA was reverse-transcribed with
random primers, a process that was carried out using the RevertAid
First-Strand Complementary DNA Synthesis Kit (Thermo Fisher
Scientific). qRT-PCR was performed in triplicate using SYBR
Select Master Mix (Applied Biosystems) on a QuantStudio 3
system (Applied Biosystems). All experiments were performed in
strict accordance with the manufacturers’instructions. Relative ex-
pression was calculated using the 2
−ΔΔCT
method with normaliza-
tion against 18Sribosomal RNA expression. The DNA sequences
for the primers can be found in table S2.
Experimental animals
All mice were kept in specific pathogen–free conditions at a cons-
tant temperature (20° to 25°C) and humidity (50 to 60%). Naturally
aged C57BL/6 mice were purchased from Jiangsu Wukong Biotech-
nology Co. Ltd. For the irradiation-accelerated mouse model of
aging, wild-type C57BL/6 mice (8 to 10 weeks old) were exposed
to 4 Gy of x-ray radiation using the RS 2000 Biological Research Ir-
radiator (RadSource) operating at 160 kV and 25 mA at a dose rate
of 1.346 Gy/min. The mice were raised for another 11 months and
used for follow-up experiments. Irradiated or naturally aged (24
months old) mice were treated with mock- or mNKG2D-CAR T
cells (1 × 10
6
cells per mouse) through tail vein injection. About 1
month after treatment, some of the mice were euthanized by cervi-
cal dislocation, and tissues were collected to detect changes in the
expression of senescence markers and NKG2DLs. Six months after
treatment, the remaining mice were assessed to measure changes in
age-related phenotypes and physical function. All experiments with
mice were approved by the animal ethics committee of the Kunming
Institute of Zoology, Chinese Academy of Sciences.
Three rhesus macaques and two cynomolgus macaques used in
this study were obtained from the Kunming Primate Research
Center at the Kunming Institute of Zoology, Chinese Academy of
Sciences and housed at the Association for Assessment and Accred-
itation of Laboratory Animal Care (AAALAC)–accredited facility of
Primate Research Center of Kunming Institute of Zoology. The
rhesus macaque identification numbers were 98106 (female, 22
years old), 00065 (male, 20 years old), and 00085 (male, 20 years
old). The cynomolgus macaque identification numbers were
00102 (female, 20 years old) and 01102 (female, 19 years old).
The monkeys were treated with autologous T cells armed with
hNKG2D-CAR (1 × 10
6
cells/kg) through intravenous injection.
Before and after treatment, peripheral blood counts were performed
using a veterinary hematology analyzer (Hemavet 950, Drew Scien-
tific Inc.), and blood biochemical analyses were performed using the
Dimension EXL 200 Integrated Chemistry System (Siemens). Cyto-
kine profiling was carried out using MesoScale Discovery V-Plex
assay kits (MesoScale Discovery). SAT and SkM were collected
surgically, and liver and kidney tissues were collected by ultra-
sound-guided percutaneous puncture before and 2 months after
NKG2D-CAR T cell treatment. All surgeries were performed
under aseptic conditions with animals under general anesthesia
by a combination of ketamine [10 mg/kg, intramuscularly (i.m.)],
sodium pentobarbital (20 mg/kg, i.m.), and atropine (0.05 mg/kg,
i.m.). All protocols with monkeys were approved by the animal
ethics committee of the Kunming Institute of Zoology, Chinese
Academy of Sciences (approval ID: IACUC19006). The experiments
were performed in accordance with approved guidelines.
Micro-CT imaging
Micro-CT imaging was carried out on the mice treated with mock-
or NKG2D-CAR T cells with a Quantum GX micro-CT imaging
system (PerkinElmer). Before analysis, the mice were continuously
anesthetized with 3% isoflurane/oxygen and were naturally prone
on the micro-CT bed. The scanning conditions for adipose tissue
were as follows: voltage, 80 kV; current, 100 μA; voxel size, 90
μm; time, 4 min. The scanning conditions for femur were as
follows: voltage, 80 kV; current, 100 μA; voxel size, 50 μm; time,
14 min. The scanning results were reconstructed in three dimen-
sions with Caliper analysis software (PerkinElmer).
Physical function analysis
Maximal walking speed was assessed using an accelerating rotarod
system (SA102, Jiangsu SANS Biological Technology Co. Ltd.) as
previously described (6). Briefly, mice were allowed to acclimate
to the rotarod system for three consecutive days. The maximal
walking speed of the mice was tested for three consecutive days be-
ginning on the fourth day. The speed and time were recorded when
the mouse fell from the balance bar. The results were averaged from
the 3-day test data and are presented as the mean ± SD.
A treadmill test was performed on a motor-driven treadmill
(ZH-PT, Huaibei Zhenghua Bio Equipment Co. Ltd.) to evaluate
the endurance capacity. The treadmill was set at a 5° incline, and
the mice were encouraged to run using 0.5-mA electrical stimula-
tion. The mice were allowed to acclimate to the treadmill for three
consecutive days at 10 m/min for 5 min. After the mice were allowed
to rest for 1 day, they were run on the treadmill at an initial speed of
5 m/min. The speed was increased from 5 to 13 m/min and then
maintained until the mice were unable to keep running. The
running time and running distance were recorded.
The grip strength of all mouse limbs was measured using a grip
strength meter (SA417, Jiangsu SANS Biological Technology Co.
Ltd.). The meter was set to the peak mode. The experiment was re-
peated seven times. After exclusion of the maximum and minimum
data points, the remaining five points were used to calculate the
mean and SD of grip strength.
H&E staining
Adipose or muscle tissues were fixed in a 4% paraformaldehyde sol-
ution, dehydrated using an ethanol gradient, cleared with xylene,
embedded in paraffin, and sectioned into 4-μm-thick sections.
These sections were then stained with hematoxylin and eosin
(H&E) using a Beyotime H&E Staining kit (C0105). After staining,
the sections were examined and recorded under a Nikon ECLIPSE
Ni-E microscope (Nikon). Cell area measurements were quantita-
tively evaluated using ImageJ software.
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 12 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Statistical analysis
Raw data for experiments with sample size n< 20 are presented in
data file S1. All statistical analyses were performed using GraphPad
Prism statistical software. Data are presented as the mean ± SD. The
normality of the data was assessed using the Shapiro-Wilk test. For
normally distributed data (P> 0.05), statistical differences between
two groups were determined using an unpaired two-tailed Student’s
ttest, with Welch’s correction applied where appropriate. Statistical
differences among three or more groups were analyzed by one- or
two-way analysis of variance (ANOVA) followed by Tukey’s multi-
ple comparisons post hoc test. Nonparametric tests, specifically the
Mann-Whitney Utest or Kruskal-Wallis test with Dunn’s correc-
tion for multiple comparisons, were used if the data were not nor-
mally distributed (P≤0.05). Statistical significance was defined as
*
P≤0.05,
**
P≤0.01, and
***
P≤0.001.
Supplementary Materials
This PDF le includes:
Figs. S1 to S18
Tables S1 and S2
Other Supplementary Material for this
manuscript includes the following:
Data file S1
MDAR Reproducibility Checklist
REFERENCES AND NOTES
1. J. M. van Deursen, The role of senescent cells in ageing. Nature 509, 439–446 (2014).
2. Y. Deng, S. S. Chan, S. Chang, Telomere dysfunction and tumour suppression: The sen-
escence connection. Nat. Rev. Cancer 8, 450–458 (2008).
3. S. He, N. E. Sharpless, Senescence in health and disease. Cell 169, 1000–1011 (2017).
4. M. J. Schafer, T. A. White, K. Iijima, A. J. Haak, G. Ligresti, E. J. Atkinson, A. L. Oberg, J. Birch,
H. Salmonowicz, Y. Zhu, D. L.Mazula, R. W. Brooks, H. Fuhrmann-Stroissnigg, T. Pirtskhalava,
Y. S. Prakash, T. Tchkonia, P. D. Robbins, M. C. Aubry, J. F. Passos, J. L. Kirkland,
D. J. Tschumperlin, H. Kita, N. K. LeBrasseur, Cellular senescence mediates fibrotic pul-
monary disease. Nat. Commun. 8, 14532 (2017).
5. D. J. Baker, T. Wijshake, T. Tchkonia, N. K. LeBrasseur, B. G. Childs, B. van de Sluis,
J. L. Kirkland, J. M. van Deursen, Clearance of p16
Ink4a
-positive senescent cells delays
ageing-associated disorders. Nature 479, 232–236 (2011).
6. Y. Cai, H. Zhou, Y. Zhu, Q. Sun, Y. Ji, A. Xue, Y. Wang, W. Chen, X. Yu, L. Wang, H. Chen, C. Li,
T. Luo, H. Deng, Elimination of senescent cells by β-galactosidase-targeted prodrug at-
tenuates inflammation and restores physical function in aged mice. Cell Res. 30,
574–589 (2020).
7. M. Xu, T. Pirtskhalava, J. N. Farr, B. M. Weigand, A. K. Palmer, M. M. Weivoda, C. L. Inman,
M. B. Ogrodnik, C. M. Hachfeld, D. G. Fraser, J. L. Onken, K. O. Johnson, G. C. Verzosa,
L. G. P. Langhi, M. Weigl, N. Giorgadze, N. K. LeBrasseur, J. D. Miller, D. Jurk, R. J. Singh,
D. B. Allison, K. Ejima, G. B. Hubbard, Y. Ikeno, H. Cubro, V. D. Garovic, X. Hou, S. J. Weroha,
P. D. Robbins, L. J. Niedernhofer, S. Khosla, T. Tchkonia, J. L. Kirkland, Senolytics improve
physical function and increase lifespan in old age. Nat. Med. 24, 1246–1256 (2018).
8. B. G. Childs, H. Li, J. M. van Deursen, Senescent cells: A therapeutic target for cardiovascular
disease. J. Clin. Invest. 128, 1217–1228 (2018).
9. J. N. Farr, M. Xu, M. M. Weivoda, D. G. Monroe, D. G. Fraser, J. L. Onken, B. A. Negley,
J. G. Sfeir, M. B. Ogrodnik, C. M. Hachfeld, N. K. LeBrasseur, M. T. Drake, R. J. Pignolo,
T. Pirtskhalava, T. Tchkonia, M. J. Oursler, J. L. Kirkland, S. Khosla, Targeting cellular sen-
escence prevents age-related bone loss in mice. Nat. Med. 23, 1072–1079 (2017).
10. N. S. Gasek, G. A. Kuchel, J. L. Kirkland, M. Xu, Strategies for targeting senescent cells in
human disease. Nat. Aging 1, 870–879 (2021).
11. A. D. Waldman, J. M. Fritz, M. J. Lenardo, A guide to cancer immunotherapy: From T cell
basic science to clinical practice. Nat. Rev. Immunol. 20, 651–668 (2020).
12. J. H. Park, I. Riviere, M. Gonen, X. Wang, B. Senechal, K. J. Curran, C. Sauter, Y. Wang,
B. Santomasso, E. Mead, M. Roshal, P. Maslak, M. Davila, R. J. Brentjens, M. Sadelain, Long-
term follow-up of CD19 CAR therapy in acute lymphoblasticleukemia. N. Engl. J. Med. 378,
449–459 (2018).
13. M. Yarmarkovich, Q. F. Marshall, J. M. Warrington, R. Premaratne, A. Farrel, D. Groff, W. Li,
M. di Marco, E. Runbeck, H. Truong, J. S. Toor, S. Tripathi, S. Nguyen, H. Shen, T. Noel,
N. L. Church, A. Weiner, N. Kendsersky, D. Martinez, R. Weisberg, M. Christie, L. Eisenlohr,
K. R. Bosse, D. S. Dimitrov, S. Stevanovic, N. G. Sgourakis, B. R. Kiefel, J. M. Maris, Cross-HLA
targeting of intracellular oncoproteins with peptide-centric CARs. Nature 599,
477–484 (2021).
14. H. Aghajanian, T. Kimura, J. G. Rurik, A. S. Hancock, M. S. Leibowitz, L. Li, J. Scholler,
J. Monslow, A. Lo, W. Han, T. Wang, K. Bedi, M. P. Morley, R. A. Linares Saldana, N. A. Bolar,
K. McDaid, C. A. Assenmacher, C. L. Smith, D. Wirth, C. H. June, K. B. Margulies, R. Jain,
E. Pure, S. M. Albelda, J. A. Epstein, Targeting cardiac fibrosis with engineered T cells. Nature
573, 430–433 (2019).
15. C. Amor, J. Feucht, J. Leibold, Y.J. Ho, C. Zhu, D. Alonso-Curbelo, J. Mansilla-Soto, J. A. Boyer,
X. Li, T. Giavridis, A. Kulick, S. Houlihan, E. Peerschke, S. L. Friedman, V. Ponomarev,
A. Piersigilli, M. Sadelain, S. W. Lowe, Senolytic CAR T cells reverse senescence-associated
pathologies. Nature 583, 127–132 (2020).
16. M. Suda, I. Shimizu, G. Katsuumi, Y. Yoshida, Y. Hayashi, R. Ikegami, N. Matsumoto,
Y. Yoshida, R. Mikawa, A. Katayama, J. Wada, M. Seki, Y. Suzuki, A. Iwama, H. Nakagami,
A. Nagasawa, R. Morishita, M. Sugimoto, S. Okuda, M. Tsuchida, K. Ozaki, M. Nakanishi-
Matsui, T. Minamino, Senolytic vaccination improves normal and pathological age-related
phenotypes and increases lifespan in progeroid mice. Nat. Aging 1, 1117–1126 (2021).
17. S. Yoshida, H. Nakagami, H. Hayashi, Y. Ikeda, J. Sun, A. Tenma, H. Tomioka, T. Kawano,
M. Shimamura, R. Morishita, H. Rakugi, The CD153 vaccine is a senotherapeutic option for
preventing the accumulation of senescent T cells in mice. Nat. Commun. 11, 2482 (2020).
18. B. I. Pereira, O. P. Devine, M. Vukmanovic-Stejic, E. S. Chambers, P. Subramanian, N. Patel,
A. Virasami, N. J. Sebire, V. Kinsler, A. Valdovinos, C. J. LeSaux, J. F. Passos, A. Antoniou,
M. H. A. Rustin, J. Campisi, A. N. Akbar, Senescent cells evade immune clearance via HLA-E-
mediated NK and CD8
+
T cell inhibition. Nat. Commun. 10, 2387 (2019).
19. N. Tasdemir, A. Banito, J. S. Roe, D. Alonso-Curbelo, M. Camiolo, D. F. Tschaharganeh,
C. H. Huang, O. Aksoy, J. E. Bolden, C. C. Chen, M. Fennell, V. Thapar, A. Chicas, C. R. Vakoc,
S. W. Lowe, BRD4 connects enhancer remodeling to senescence immune surveillance.
Cancer Discov. 6, 612–629 (2016).
20. P. Liu, F. Li, J. Lin, T. Fukumoto, T. Nacarelli, X. Hao, A. V. Kossenkov, M. C. Simon, R. Zhang,
m
6
A-independent genome-wide METTL3 and METTL14 redistribution drives the senes-
cence-associated secretory phenotype. Nat. Cell Biol. 23, 355–365 (2021).
21. A. Iannello, T. W. Thompson, M. Ardolino, S. W. Lowe, D. H. Raulet, p53-dependent che-
mokine production by senescent tumor cells supports NKG2D-dependent tumor elimi-
nation by natural killer cells. J. Exp. Med. 210, 2057–2069 (2013).
22. V. Krizhanovsky, M. Yon, R. A. Dickins, S. Hearn, J. Simon, C. Miething, H. Yee, L. Zender,
S. W. Lowe, Senescence of activated stellate cells limits liver fibrosis. Cell 134,
657–667 (2008).
23. D. J. Baker, B. G. Childs, M. Durik, M. E. Wijers, C. J. Sieben, J. Zhong, R. A. Saltness,
K. B. Jeganathan, G. C. Verzosa, A. Pezeshki, K. Khazaie, J. D. Miller, J. M. van Deursen,
Naturally occurring p16
Ink4a
-positive cells shorten healthy lifespan. Nature 530,
184–189 (2016).
24. D. P. Munoz, S. M. Yannone, A. Daemen,Y. Sun, F. Vakar-Lopez, M. Kawahara, A. M. Freund,
F. Rodier, J. D. Wu, P. Y. Desprez, D. H. Raulet, P. S. Nelson, L. J. van ’t Veer, J. Campisi,
J. P. Coppe, Targetable mechanisms driving immunoevasion of persistent senescent cells
link chemotherapy-resistant cancer to aging. JCI Insight 5, e124716 (2019).
25. J. Fan, J. Shi, Y. Zhang, J. Liu, C. An, H. Zhu, P. Wu, W. Hu, R. Qin, D. Yao, X. Shou, Y. Xu,
Z. Tong, X. Wen, J. Xu, J. Zhang, W. Fang, J. Lou, W. Yin, W. Chen, NKG2D discriminates
diverse ligands through selectively mechano-regulated ligand conformational changes.
EMBO J. 41, e107739 (2022).
26. C. J. Sherr, R. A. DePinho, Cellular senescence: Mitotic clock or culture shock? Cell 102,
407–410 (2000).
27. S. Parrinello, E. Samper, A. Krtolica, J. Goldstein, S. Melov, J. Campisi, Oxygen sensitivity
severely limits the replicative lifespan of murine fibroblasts. Nat. Cell Biol. 5,
741–747 (2003).
28. B. G. Childs, M. Gluscevic, D. J. Baker, R. M. Laberge, D. Marquess, J. Dananberg, J. M. van
Deursen, Senescent cells: An emerging target for diseases of ageing. Nat. Rev. Drug Discov.
16, 718–735 (2017).
29. M. Xu, A. K. Palmer, H. Ding, M. M. Weivoda, T. Pirtskhalava, T. A. White, A. Sepe,
K. O. Johnson, M. B. Stout, N. Giorgadze, M. D. Jensen, N. K. LeBrasseur, T. Tchkonia,
J. L. Kirkland, Targeting senescent cells enhances adipogenesis and metabolic function in
old age. eLife 4, e12997 (2015).
30. M. Xu, T. Tchkonia, H. Ding, M. Ogrodnik, E. R. Lubbers, T. Pirtskhalava, T. A. White,
K. O. Johnson, M. B. Stout, V. Mezera, N. Giorgadze, M. D. Jensen, N. K. LeBrasseur,
J. L. Kirkland, JAK inhibition alleviates the cellular senescence-associated secretory phe-
notype and frailty in old age. Proc. Natl. Acad. Sci. U.S.A. 112, E6301–E6310 (2015).
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 13 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

31. Z. Yao, B. Murali, Q. Ren, X. Luo, D. V. Faget,T. Cole, B. Ricci, D. Thotala, J. Monahan, J. M. van
Deursen, D. Baker, R. Faccio, J. K. Schwarz, S. A. Stewart, Therapy-induced senescence
drives bone loss. Cancer Res. 80, 1171–1182 (2020).
32. M. L. Sentman, J. M. Murad, W. J. Cook, M. R. Wu, J. Reder, S. H. Baumeister, G. Dranoff,
M. W. Fanger, C. L. Sentman, Mechanisms of acute toxicity in NKG2D chimeric antigen
receptor T cell-treated mice. J. Immunol. 197, 4674–4685 (2016).
33. H. VanSeggelen, J. A. Hammill, A. Dvorkin-Gheva, D. G. Tantalo, J. M. Kwiecien,
G. F. Denisova, B. Rabinovich, Y. Wan, J. L. Bramson, T cells engineered with chimeric
antigen receptors targeting NKG2D ligands display lethal toxicity in mice. Mol. Ther. 23,
1600–1610 (2015).
34. R. J. Colman, Non-human primates as a model for aging. Biochim. Biophys. Acta Mol. Basis
Dis. 1864, 2733–2741 (2018).
35. S. J. Park, O. Gavrilova,A. L. Brown, J. E. Soto, S. Bremner, J. Kim, X. H. Xu, S. T.Yang, J. H. Um,
L. G. Koch, S. L. Britton, R. L. Lieber, A. Philp, K. Baar, S. G. Kohama, E. D. Abel, M. K. Kim,
J. H. Chung, DNA-PK promotes the mitochondrial, metabolic, and physical decline that
occurs during aging. Cell Metab. 25, 1135–1146.e7 (2017).
36. M. L. Davila, I. Riviere, X. Wang, S. Bartido, J. Park, K. Curran, S. S. Chung, J. Stefanski,
O. Borquez-Ojeda, M. Olszewska, J. Qu, T. Wasielewska, Q. He, M. Fink, H. Shinglot,
M. Youssif, M. Satter, Y. Wang, J. Hosey, H. Quintanilla, E. Halton, Y. Bernal, D. C. Bouhassira,
M. E. Arcila, M. Gonen, G. J. Roboz, P. Maslak, D. Douer, M. G. Frattini, S. Giralt, M. Sadelain,
R. Brentjens, Efficacy and toxicity management of 19-28z CAR T cell therapy in B cell acute
lymphoblastic leukemia. Sci. Transl. Med. 6, 224ra225 (2014).
37. D. J. Baker, C. Perez-Terzic, F. Jin, K. S. Pitel, N. J. Niederlander, K. Jeganathan, S. Yamada,
S. Reyes, L. Rowe, H. J. Hiddinga, N. L. Eberhardt, A. Terzic, J. M. van Deursen, Opposing
roles for p16
Ink4a
and p19
Arf
in senescence and ageing caused by BubR1 insufficiency. Nat.
Cell Biol. 10, 825–836 (2008).
38. R. Di Micco, V. Krizhanovsky, D. Baker,F. d’Adda di Fagagna, Cellular senescence in ageing:
From mechanisms to therapeutic opportunities. Nat. Rev. Mol. Cell Biol. 22, 75–95 (2021).
39. T. Tchkonia, Y. Zhu, J. van Deursen, J. Campisi, J. L. Kirkland, Cellular senescence and the
senescent secretory phenotype: Therapeutic opportunities. J. Clin. Invest. 123,
966–972 (2013).
40. C. D. Wiley, R. Sharma, S. S. Davis,J. A. Lopez-Dominguez, K. P. Mitchell, S. Wiley, F. Alimirah,
D. E. Kim, T. Payne, A. Rosko, E. Aimontche, S. M. Deshpande, F. Neri, C. Kuehnemann,
M. Demaria, A. Ramanathan, J. Campisi, Oxylipin biosynthesis reinforces cellular senes-
cence and allows detection of senolysis. Cell Metab. 33, 1124–1136.e5 (2021).
41. A. Hernandez-Segura, J. Nehme, M. Demaria, Hallmarks of cellular senescence. Trends Cell
Biol. 28, 436–453 (2018).
42. E. O. Wissler Gerdes, A. Misra, J. M. E. Netto, T. Tchkonia, J. L. Kirkland, Strategies for late
phase preclinical and early clinical trials of senolytics. Mech. Ageing Dev. 200,
111591 (2021).
43. L. J. Hickson, L. G. P. Langhi Prata, S. A. Bobart, T. K. Evans, N. Giorgadze, S. K. Hashmi,
S. M. Herrmann, M. D. Jensen, Q. Jia, K. L. Jordan, T. A. Kellogg, S. Khosla, D. M. Koerber,
A. B. Lagnado, D. K. Lawson, N. K. LeBrasseur, L. O. Lerman, K. M. McDonald, T. J. McKenzie,
J. F. Passos, R. J. Pignolo, T. Pirtskhalava, I. M. Saadiq, K. K. Schaefer, S. C. Textor,
S. G. Victorelli, T. L. Volkman, A. Xue, M. A. Wentworth, E. O. Wissler Gerdes, Y. Zhu,
T. Tchkonia, J. L. Kirkland, Senolytics decrease senescent cells in humans: Preliminary
report from a clinical trial of dasatinib plus quercetin in individuals with diabetic kidney
disease. EBioMedicine 47, 446–456 (2019).
44. D. Montani, E. Bergot, S. Gunther, L. Savale, A. Bergeron, A. Bourdin, H. Bouvaist, M. Canuet,
C. Pison, M. Macro, P. Poubeau, B. Girerd, D. Natali, C. Guignabert, F. Perros,
D. S. O’Callaghan, X. Jais, P. Tubert-Bitter, G. Zalcman, O. Sitbon, G. Simonneau,
M. Humbert, Pulmonary arterial hypertension in patients treated by dasatinib. Circulation
125, 2128–2137 (2012).
45. M. Borghesan, W. M. H. Hoogaars, M. Varela-Eirin, N. Talma, M. Demaria, A senescence-
centric view of aging: Implications for longevity and disease. Trends Cell Biol. 30,
777–791 (2020).
46. S. Cang, C. Iragavarapu, J. Savooji, Y. Song, D. Liu, ABT-199 (venetoclax) and BCL-2 inhib-
itors in clinical development. J. Hematol. Oncol. 8, 129 (2015).
47. B. Sun, D. Yang, H. Dai, X. Liu, R. Jia, X. Cui, W. Li, C. Cai, J. Xu, X. Zhao, Eradication of he-
patocellular carcinoma by NKG2D-based CAR-T cells. Cancer Immunol. Res. 7,
1813–1823 (2019).
48. A. Zingoni, R. Molfetta, C. Fionda, A. Soriani, R. Paolini, M. Cippitelli, C. Cerboni, A. Santoni,
NKG2D and its ligands: “One for all, all for one”.Front. Immunol. 9, 476 (2018).
49. L. L. Lanier, NKG2D receptor and its ligands in host defense. Cancer Immunol. Res. 3,
575–582 (2015).
50. T. Hasegawa, T. Oka, H. G. Son, V. S. Oliver-Garcia, M. Azin, T. M. Eisenhaure, D. J. Lieb,
N. Hacohen, S. Demehri, Cytotoxic CD4
+
T cells eliminate senescent cells by targeting
cytomegalovirus antigen. Cell 186, 1417–1431.e20 (2023).
51. S. Curio, G. Jonsson, S. Marinovic, A summary of current NKG2D-based CAR clinical trials.
Immunother. Adv. 1, ltab018 (2021).
52. S. H. Baumeister, J. Murad, L. Werner, H. Daley, H. Trebeden-Negre, J. K. Gicobi,
A. Schmucker, J. Reder, C. L. Sentman, D. E. Gilham, F. F. Lehmann, I. Galinsky, H. DiPietro,
K. Cummings, N. C. Munshi, R. M. Stone, D. S. Neuberg, R. Soiffer, G. Dranoff, J. Ritz,
S. Nikiforow, Phase I trial of autologous CAR T cells targeting NKG2D ligands in patients
with AML/MDS and multiple myeloma. Cancer Immunol. Res. 7, 100–112 (2019).
53. D. A. Sallman, J. Brayer, E. M. Sagatys, C. Lonez, E. Breman, S. Agaugue, B. Verma,
D. E. Gilham, F. F. Lehmann, M. L. Davila, NKG2D-based chimeric antigen receptor therapy
induced remission in a relapsed/refractory acute myeloid leukemia patient. Haematologica
103, e424–e426 (2018).
54. D. A. Sallman, J. B. Brayer, X. Poire, V. Havelange, A. Awada, P. Lewalle, K. Odunsi, E. S. Wang,
C. Lonez, T. Lequertier, E. Alcantar-Orozco, N. Braun, A. Flament, I. Moors, T. Kerre, Results
from the completed dose-escalation of the hematological arm of the phase I think study
evaluating multiple infusions of NKG2D-based CAR T-cells as standalone therapy in
relapse/refractory acute myeloid leukemia and myelodysplastic syndrome patients. Blood
134, 3826 (2019).
55. S. Hosomi, J. Grootjans, M. Tschurtschenthaler, N. Krupka, J. D. Matute, M. B. Flak,
E. Martinez-Naves, M. G. del Moral, J. N. Glickman, M. Ohira, L. L. Lanier, A. Kaser,
R. Blumberg, Intestinal epithelial cell endoplasmic reticulum stress promotes MULT1 up-
regulation and NKG2D-mediated inflammation. J. Exp. Med. 214, 2985–2997 (2017).
56. J. J. Melenhorst, G. M. Chen, M. Wang, D. L. Porter, C. Chen, M. A. Collins, P. Gao,
S. Bandyopadhyay, H. Sun, Z. Zhao, S. Lundh, I. Pruteanu-Malinici, C. L. Nobles, S. Maji,
N. V. Frey,S. I. Gill, A. W. Loren, L. Tian, I. Kulikovskaya, M. Gupta, D. E. Ambrose, M. M. Davis,
J. A. Fraietta, J. L. Brogdon, R. M. Young, A. Chew, B. L. Levine, D. L. Siegel, C. Alanio,
E. J. Wherry, F. D. Bushman, S. F. Lacey, K. Tan, C. H. June, Decade-long leukaemia remis-
sions with persistence of CD4
+
CAR T cells. Nature 602, 503–509 (2022).
57. D. Yang, B. Sun, H. Dai, W. Li, L. Shi, P. Zhang, S. Li, X. Zhao, T cells expressing NKG2D
chimeric antigen receptors efficiently eliminate glioblastoma and cancer stem cells.
J. Immunother. Cancer 7, 171 (2019).
58. T. Zhang, B. A. Lemoi, C. L. Sentman, Chimeric NK-receptor-bearing T cells mediate anti-
tumor immunotherapy. Blood 106, 1544–1551 (2005).
Acknowledgments: We would like to thank Z. Hu from the Kunming Primate Research Center,
Kunming Institute of Zoology, Chinese Academy of Sciences for technical support in
nonhuman primate surgeries. We also acknowledge the Core Facility of West China Hospital (L.
Chai, Y. Li, X. Xu, L. Wu, and L. Bai) for technical assistance. Funding: Thiswork wassupported by
National Natural Science Foundation of China grant U1702289 and 82172701 (to X. Zhao);
National Key R&D Program of China grant 2018YFC200043 (to X. Zhao); 1.3.5 Project for
Disciplines of Excellence, West China Hospital, Sichuan University grantZYYC20002 (to X. Zhao);
and Department of Science and Technology of Sichuan Province grant 2020JDRC0019 (to D.Y.).
Author contributions: X. Zhao and Y.D. designed the study. D.Y., B.S., and S.L. designed and
constructed the human NKG2D-CAR vectors and prepared the lentivirus. X.C., X.L., and X. Zhang
generated senescent models with human cells. D.Y., X.C., and S.L. analyzed the cytotoxicity of
hNKG2D-CAR T cells on human senescent cells. D.Y., B.S., and X.L. detected the safety and
effects of hNKG2D-CAR T cells in monkeys. D.Y. designed and constructed mNKG2D-CAR
vectors. D.Y., B.S., S.L., and W.W. prepared the retrovirus, generated senescent MEFs, and
analyzed the cytotoxicity of mNKG2D-CAR T cells on senescent MEFs. D.Y., B.S., W.W., N.L., and
L.Y. detected the effects of the mNKG2D-CART cells in mice. X. Zhao, D.Y., and Y.D. analyzed the
results and wrote the manuscript. X. Zhao was responsible for study supervision. Competing
interests: X. Zhao, D.Y., and S.L. are inventors on a Chinese patent (Application of NKG2D CAR-
immune cells in anti-aging, patent no. ZL202011482080.0) and a pending U.S. patent
(Application of NKG2D CAR-immune cells in anti-aging, application no.18/255,808) related to
this work. The other authors declare that theyhave no competing interests. Data and materials
availability: All data associated with this study are present in the paper or the Supplementary
Materials. Cell lines and plasmids can be provided by X. Zhao through a completed material
transfer agreement. Requests for the materials should be submitted to X. Zhao (zhaoxudong@
wchscu.cn). Information about all commercially available reagents is provided in Materials and
Methods and table S1.
Submitted 26 May 2022
Resubmitted 17 May 2023
Accepted 10 July 2023
Published 16 August 2023
10.1126/scitranslmed.add1951
SCIENCE TRANSLATIONAL MEDICINE |RESEARCH ARTICLE
Yang et al.,Sci. Transl. Med. 15, eadd1951 (2023) 16 August 2023 14 of 14
Downloaded from https://www.science.org at Sichuan University on August 16, 2023

Use of this article is subject to the Terms of service
Science Translational Medicine (ISSN ) is published by the American Association for the Advancement of Science. 1200 New York Avenue
NW, Washington, DC 20005. The title Science Translational Medicine is a registered trademark of AAAS.
Copyright © 2023 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works
NKG2D-CAR T cells eliminate senescent cells in aged mice and nonhuman
primates
Dong Yang, Bin Sun, Shirong Li, Wenwen Wei, Xiuyun Liu, Xiaoyue Cui, Xianning Zhang, Nan Liu, Lanzhen Yan, Yibin
Deng, and Xudong Zhao
Sci. Transl. Med., 15 (709), eadd1951.
DOI: 10.1126/scitranslmed.add1951
View the article online
https://www.science.org/doi/10.1126/scitranslmed.add1951
Permissions
https://www.science.org/help/reprints-and-permissions
Downloaded from https://www.science.org at Sichuan University on August 16, 2023