nature biotechnology
Article https://doi.org/10.1038/s41587-023-02060-8
IL-10-expressing CAR T cells resist dysfunction and mediate durable clearance of solid tumors and metastases
Received: 1 April 2022
Accepted: 8 November 2023
Published online: 02 January 2024
<>] Check for updates
Yang Zhao©1,2,8, Jiangqing Chen3,8, Massimo Andreatta©4'5, Bing Feng1,2 Yu-Qing Xie©1, Mathias Wenes4, Yi Wang1'2, Min Gao©1, Xiaomeng Hu1, Pedro Romero©4, Santiago Carmona©4'5, Jie Sun©3, Yugang Guo©1'2'6'7 & Li Tang©12
The success of chimeric antigen receptor (CAR) T cell therapy in treating several hematopoietic malignancies has been difficultto replicate in solid tumors, in part because of T cell exhaustion and eventually dysfunction. To counter T cell dysfunction in the tumor microenvironment, we metabolically armored CAR T cells by engineering them to secrete interleukin-10 (IL-10). We show that IL-10 CAR T cells preserve intact mitochondrial structure and function in the tumor microenvironment and increase oxidative phosphorylation in a mitochondrial pyruvate carrier-dependent manner. IL-10 secretion promoted proliferation and effector function of CAR T cells, leading to complete regression of established solid tumors and metastatic cancers across several cancer types in syngeneic and xenograft mouse models, including colon cancer, breast cancer, melanoma and pancreatic cancer. IL-10 CAR T cells also induced stem cell-like memory responses in lymphoid organs that imparted durable protection against tumor rechallenge. Our results establish a generalizable approach to counter CAR T cell dysfunction through metabolic armoring, leading to solid tumor eradication and long-lasting immune protection.
Chimeric antigen receptor (CAR) T cell therapy has achieved remarkable clinical success in treating B cell malignancies. However, CAR T cells have so far shown limited efficacy against solid tumors, and long-term disease control remains rare. One of the major factors that hinders the efficacy of CARTcell therapy isTcell dysfunction in tumors due to T cell exhaustion1. Exhausted T cells are characterized as a distinct population with completeor partial loss of proliferative capacity
and effector function, upregulation of multiple immune inhibitory receptors and transcriptional and epigenetic alterations2-5. Metabolic fitness is essential to sustain T cell survival and function. Emerging evidence suggests that metabolic alteration and deficiency of intra-tumoral T cells drive T cell exhaustion68. Mitochondrial dysfunction imposed by mitochondrial depolarization and oxidative stress was shown to reinforce phenotypic and epigenetic exhaustion programs
'Institute of Bioengineering, Ecole Polytechnique Federate de Lausanne, Lausanne, Switzerland, institute of Materials Science & Engineering, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland, department of Cell Biology and Bone Marrow Transplantation Center of the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China. "Department of Oncology, University of Lausanne, Lausanne, Switzerland. 5Swiss Institute of Bioinformatics, Lausanne, Switzerland. 6Present address: Institute of Drug Metabolism and Pharmaceutical Analysis, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, China. 'National Key Laboratory of Advanced Drug Delivery and Release Systems, Zhejiang University, Hangzhou, China. "These authors contributed equally: Yang Zhao, Jiangqing Chen. Ele-mail: sunj4@zju.edu.cn; yugang.guo@zju.edu.cn; li.tang@epfl.ch
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
in tumor-infiltrating T cells7,8. In addition, impairing mitochondrial oxidative phosphorylation (OXPHOS) suppresses T cell proliferation and promotes exhaustion9. Strategies sustaining mitochondrial fitness and respiration, such as antioxidant treatment, could restore the proliferation and effector functions of exhausted T cells7,8. CAR T cells cultured in medium supplemented with cytokines, such asengi-neeredinterIeukin-2 (IL-2) or IL-15, exhibited improved metabolic fitness and enhanced antitumor immunity in vivo when transferred to treat tumor-bearing mice10,11. However, it remains challenging to modulate T cell metabolism in vivo and rejuvenate exhausted Tcells in the tumor microenvironment (TME).
We recently reported that in vivo administration of an IL-10-Fc fusion protein reprograms intratumoral T cell metabolism toward OXPHOS and increases mitochondrial respiratory capacity, leading to potently enhanced expansion and effector function of terminally exhausted CD8+tumor-infiltrating lymphocytes (TILs)12. This metabolic reprogramming strategy could therefore be exploited to engineer metabolically armored CARTcell therapy to better treat solid tumors. Here, we show that CART cells engineered to express IL-10 prevented the functional impairment associated with Tcell exhaustion in the TME, leading to enhanced proliferative capacity and effector function. Further analyses revealed that IL-10 expression improved mitochondrial fitness and increased OXPHOS in CARTcells in a mitochondrial pyruvate carrier (MPC)-dependent manner. Infusion of mouse and human IL-10-expressing CAR T cells eradicated established solid tumors in multiple syngeneic and xenograft mouse models, respectively, and induced stem cell-like memory (Tscm cell) responses to control tumor recurrence. The metabolically armored CARTcell is a promising therapeutic strategy to counter T cell exhaustion-associated dysfunction and induce potent and long-term antitumor immunity.
Results
IL-10-expressing CAR T cells counter dysfunction in tumors
Adoptive transfer of CARTcells against human epidermal growth factor receptor 2 (HER2) failed to control established solid tumors partially due to functional impairment of tumor-infiltrating CAR T cells, which exhibited reduced cytotoxicity and higher expression levels of inhibitory receptors than the transferred CAR T cells found in spleen (Extended Data Fig. la-e). To counter T cell dysfunction in the TME, we produced a mouse IL-10-expressing HER2 CAR T cell (IL-10 HER2 CAR T) using a tandem construct encoding the second-generation anti-HER2 CAR and mouse IL-10 linked with a cleavable 2A peptide sequence (Fig. la). IL-10 HER2 CAR T cells expressed approximately the same levels of CAR as HER2 CART cells (Extended Data Fig. If) but secreted a high level of IL-10 in the culture (Extended Data Fig. lg). The secreted IL-10 showed negligible impact on the phenotypes and viability of CAR T cells during the ex vivo preparation phase before infusion (Extended Data Fig. lh-q).
We next intravenously (i.v.) transferred IL-10 HER2 CAR T cells (3 x io6) or HER2 CAR T cells (3 x io6) in the presence or absence of i.v.-administered free mouse IL-10 to treat established subcutaneous
Fig. 11 IL-10-expressing C AR T cells counter dysfunction in tumors.
a, Schematic depicting the HER2 CAR and IL-10 HER2 CAR constructs; scFv, single-chain variable fragment; TM, transmembrane domain; LTR, long terminal repeat, b-1, C57BL/6 mice were inoculated s.c. with MC38-HER2 mouse colon adenocarcinoma cells (1 x io6) and sublethallylymphodepleted by irradiation on day 5 and received i.v. adoptive transfer of IL-10 HER2 CAR T cells (3 x io6) or HER2 CAR T cells (3 x io6) in the presence or absence of i.v.-administered IL-10 (1 ug) on day 6 (n = 5 mice; data a re pooled from two independent experiments). On day 14, mice were killed, and the indicated tissues were processed and analyzed by flow cytometry, b, Experimental timeline, c, Counts of viable HER2 CAR T cells in tumors, d, Frequencies of Ki67* HER2 CAR T cells in tumors, e-h, Counts of viable IL-2* (e), PD-1*TIM3* (f), granzyme B* (g) and polyfunctional (h) HER2 CAR T cells in tumors, i, Representative flow cytometry plots and average mean
(s.c.) HER2-expressing MC38 (MC38-HER2) mouse colon adenocarcinoma tumors in a mouse model (Fig. lb). Compared to conventional CART cells, ILIO HER2 CART cellsexhibited an -5.3-fold increase in cell counts, augmented Ki67 expression and increased expansion of the population that secretes IL-2 in the tumor, suggesting that IL-10 expression markedly enhanced CARTcell proliferation (Fig. Ice). Consistent with previous reports12, the PD l+TIM-3+ terminally exhausted subset was notably expanded, whereas the counts of progenitor exhausted Tcells (TCF 1+PD l+TIM-3_) remained unchanged (Fig. If and Extended Data Fig. 2a). IL-10 secretion enhanced cytotoxicity and polyfunc-tionality of HER2 CAR T cells, including both terminally and progenitor exhausted subsets (Fig. lg,h and Extended Data Fig. 2b,c). IL-10 HER2 CARTcells also exhibited reduced expression levels of PD-1, suggesting alleviated exhaustion (Fig. li). By contrast, secreted IL-10 showed negligible effects on other immunecells or tumor-infiltrating endogenous Tcells, including regulatory T (Treg) cells (Extended Data Fig. 2d-g). HER2 CAR T cells with the highest CAR expression density showed decreased cytokine production and increased PD-1 and TIM-3 expression (Fig. lj-l and Extended Data Fig. 2h,i), suggestingthat strong antigen stimulation through CAR may driveexhaustion of intratumoral CARTcells13 M. The extent of functional restoration of ILIO HER2 CAR T cells by IL-10 secretion seemed dependent on CAR signaling; the reinvigoration of exhausted CARTcells was in general more prominent when there was a higher CAR density. Of note, we found that delivery of IL-10 through CARTcell secretion was superior to i.v. administration in reinvigorating exhausted CAR Tcells (Fig. lc-l), likely because secreted IL-10 was enriched in tumors and predominantly acted on tumor-infiltrating CART cells.
To extend this strategy to human CAR T (hCAR T) cells, we prepared IL-10-expressing CD19-targeted T cell antigen receptor-a constant (77?/lC)-deficient hCARTcells (CD19 hCAR; Extended Data Fig. 3a,b). ILIO CD19 hCAR T cells secreted human IL-10 in culture and showed enhanced tumor-lytic potential against CD19-expressing PANC1 human pancreatic tumor cells (PANC1-CD19) and Raji lymphoma cells in vitro (Extended Data Fig. 3c-e). After multiple rounds of CAR stimulation by CD19-expressing NIH/3T3 or PANC1-CD19 cells in long-term culture, ILIO CD19 hCAR T cells exhibited increased expansion compared to conventional CD19 hCAR T cells (Extended Data Fig. 3f,g). Similar to mouse CAR T cells, ILIO CD19 hCAR T cells that were transferred (i.v.) to immune-deficient NODsritf gamma (NSG) mice bearing PANC1-CD19 tumors showed enhanced production of granzyme B and interferony (IFNy) compared to conventional CD19 hCAR T cells (Fig. lm-o). Together, incorporation of IL-10 secretion countered CARTcell dysfunction in tumors by promoting their proliferation and effector function.
IL-10-expressing CAR T cells sustain mitochondrial fitness
Impaired mitochondrial fitness has been shown to reinforce T cell exhaustion8,15. We next examined the mitochondrial mass and membrane potential of CART cells by staining them with MitoTracker Green (MG) and MitoTracker Deep Red (MDR), respectively. Consistent with prior reports of TILs8, we found that CAR T cells with dysfunctional
fluorescence intensity (MFI) showing PD-1 expression levels on HER2 CAR T cells in tumors, j-1, CAR T cells in tumors were classified into several subpopulations based on gating of CAR density (j). Shown are curves of the MFI of IFNy (k) and PD-1 (1) as a function of CAR density, m-o, NSG mice were inoculated s.c. with PANC1-CD19 cells (5 x io6) and received i.v. adoptive transfer of CD19 hCAR T cells or IL-10 CD19 hCAR T cells (1 x io6; n = 5 mice). On day 38, mice were killed, and the indicated tissues were processed and analyzed by flow cytometry, m, Experimental timeline. n,o, Frequencies of granzyme B* (n) and IFNy* (o) hCAR T cells in tumors. All data represent the mean ± s.e.m. and were analyzed by two-tailed Student's f-test (n and o) or one- or two-way analysis of variance (ANOVA) with Tukey's multiple-comparisons tests (c-i, kand 1). Data are representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
mitochondria (defined as a low ratio of MDR to MG; MDR/MG'0) were and expressed much higher levels of PD-1 and TIM-3 than the MDR/ notably enriched in tumors compared to in the spleen (Fig. 2a,b). The MGhl population, showing an exhaustion phenotype (Extended Data MDR/MG10 subpopulation was substantially lessactive in degranulation   Fig. 4a). Further, analysis of mitochondrial ultrastructure by electron
HER2 CAR construct
IL-10 HER2 CAR construct
Inoculation of MC38-HER2 tumors (s.c.) in C57BL/6
Kill and analyze
14 Time(d)
Lymphodepletion j
HER2 CAR T cells + IL-10 (i.V.) /IL-10 HER2 CAR T cells (i.V.)
1—	o	2	x103 -|
	E		
<		1.5	x103 -
u	öi		
	E	1	x103 -
CO	o		
uno	CL	5	x102 -
U	eel		0 -
P= 0.9996 -1
4
P = 0.0066 -1
cd o
* E 60-'S 3
oj c
o)"" 40-co J2
C 0
m o on. 2 \-0 q;
°- < o-
P = 0.2557 P = 0.0225 -II-1
a. o 600
< E
400
=! E
0 S 200
01 CL
c J2
o
0
P = 0.9996 P= 0.0069
	o	
CO	E	
		1,000 -i
Ú.		
i—	mg	800 -
Q	0	600 -
Q_	a	
		400 -
nts	o	200 -
	I—	0 -
o u	CAR	
<
m E
O =3
E ~
400
300
£ £ 200 2 o
O) Q- 100
4- 01
o =:
0
o
i P= 0.0002 P= 0.9957 J
«!> ^ ,«!>
o z.
2 E
P = 0.0239 -1
8
I IFNy*
I IFNY*granzyrne B* I IFNy'TNFa* IFNY*granzyrne B*TNFct*
I HER2 CAR T
I HER2 CAR T+ IL-10
I IL-10 HER2 CAR T
0        104 10b
PD-1 (APC/Cy7)
4x104 3x104 2x104 1 x104 0
P= 0.9726
P= 0.0117
- o8§ rfta A
,4> <A <ß>
ty -V
CAR density of CAR T cells
HER2 CAR (PE/Cy7)-
2x104
■ HER2 CART
■ HER2 CAR T +IL-10
■ IL-10 HER2 CART
P = 0.0744
P< 0.0001
5x104      1x10°     1.5x10° 2x10°
CAR density
■B- HER2 CAR T ■B- HER2 CAR T +IL-10 IL-10 HER2 CAR T
♦
P = 0.0027
5x104     1x10°    1.5x10° 2x10° CAR density
Inoculation of PANC1-CD19 tumors (s.c.) in NSG mice
27 Í
Kill and analyze
38 Day
CD19 hCAR T/IL-10 CD19 hCAR T (i.V.)
!S £ 100
£ E
CO Zl
o <
60 40 20 0
A
o o
0 _^
— E 60 o 3
0 C
si 40
C 0
0 o 20
2 I-
0 q;
-< 0
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
microscopy (EM) indicated that CARTcells in tumors possessed mitochondria with an enlarged balloon-like shape and loosely structured cristae, whereasspIenicCARTcellsexhibited tubular mitochondria with tight cristae (Fig. 2c). In addition, mitochondria in tumor-infiltrating CAR T cells showed reduced crista numbers per mitochondrion and decreased length of cristae per mitochondrial area compared to those in the spleen (Fig. 2d,e). These data reveal the accumulation of dysfunctional mitochondria in exhausted CARTcells in tumors.
IL-10 expression sustained mitochondria fitness in tumor-infiltrating CAR T cells, with substantially reduced frequencies of dysfunctional mitochondria in ILTOHER2 CARTcells (5.5%) compared to HER2 CAR T cells alone (22.9%) and HER2 CAR T cells combined with exogenous IL-10 (23.5%; Fig. 2f,g). IL-10 expression also increased the ratio of MDR to MG in IL-10 HER2 CART cells (Fig. 2h). EM imaging analysisof mitochondrial ultrastructure provided additional evidence of enrichment of mitochondria with a tubular shape, well-structured cristae,and increased number and length of cristae in tumor-infiltrating IL-10 HER2CARTcellscompared to theconventionalHER2 CARTcells (Fig. 2i-l and Extended Data Fig. 4b). HER2 CART cells with increased CARexpression density exhibited a higher frequency of dysfunctional mitochondria, which was considerably reduced in IL-10 HER2 CAR Tcells (Fig. 2m and Extended Data Fig. 4c). Similarly, IL-10 CD19 hCAR T cells exhibited decreased frequency of cells with dysfunctional mitochondria and an increased ratio of MDR to MG in tumors compared to CD19 hCART cells (Fig. 2n,o). These results together suggest that antigen stimulation may drive mitochondrial dysfunction in CAR Tcells, whereas ILTO expression could sustain mitochondrial fitness of CARTcells, especially those experiencing strong antigen stimulation.
IL-10 expression promotes OXPHOS of CART cells
We next assessed whether ILTO expression could reprogram CARTcell metabolism. To mimic persistent tumor antigen stimulation in the TME, wecoculturedMC38-HER2 cells and HER2 CARTcells with or without IL-10 expression. After antigen stimulation, IL-10 HER2 CART cells exhibited markedly elevated basal and maximal oxygen consumption rates (OCRs) compared to HER2 CART cells alone (Fig. 3a,b), whereas extracellular acidification rate (ECAR) remained almost unchanged (Extended Data Fig. 4d,e). Therefore, IL-10 expression increased the ratios of OCR to ECAR and reprogrammed CART cell metabolism (Fig. 3c). Addition of exogenous IL-10 also increased the basal OCR of HER2 CAR T cells (Fig. 3b). Notably, in the absence of antigen stimulation, neither exogenous nor secreted IL-10 had any effects on the OCR of CAR T cells (Extended Data Fig. 4f). In addition, metabolomic analysis confirmed that intratumoral IL-10 HER2 CARTcells exhibited a distinct metabolomic profile compared to conventional HER2 CAR T cells, with markedly increased levels of the pyruvate production intermediate phosphoenolpyruvate and the tricarboxylic acid cycle intermediate succinate (Extended Data Fig. 4g-j and Supplementary Table 1). These results indicate that IL-10 expression reprograms CAR Tcell metabolism toward OXPHOS in an antigen-dependent manner.
Fig. 21 IL-10 expression sustains the mitochondrial fitness of CART cells.
a-m, The experimental setting was the same as described in Fig. lb. CAR T cells were subjected to mitochondrial phenotype analysis by flow cytometry (n = 5 mice; data were pooled from two independent experiments) or sorted for EM analysis (n = 15 biologically independent samples). Mitochondrial mass and membrane potential of CAR T cells were examined by staining with MG and MDR, respectively, a, Representative flow cytometry plots showing MDR/MGhi and MDR/MG'° (defined as dysfunctional mitochondria) subpopulations of HER2 CAR T cells in the spleen and tumors. The numbers indicate the frequency (%). b, Frequencies of HER2 CAR T cells with dysfunctional mitochondria in the spleen and tumors, c, Representative EM images of sorted splenic and intratumoral HER2 CAR T cells. d,e, Quantification of crista numbers per mitochondrion (d) and total crista length per mitochondrial area (e) in splenic and intratumoral HER2 CAR T cells, f, Representative flow cytometry plots showing dysfunctional
We recently showed that ILTO-Fc promotes OXPHOS in terminally exhausted CD8+TILs through MPC12. Consistently, by blocking pyruvate transportation using an MPC inhibitor UK5099, but not by inhibiting fatty acid oxidation with etomoxir, we could completely abrogate the increase of OXPHOS in IL-10 HER2 CAR T cells (Fig. 3d). In the cocul-ture with MC38-HER2 cells, IL-10 HER2 CAR T cells exhibited greatly enhanced proliferation, polyfunctionality and killing efficiency of target cells compared to HER2 CAR T cells (Fig. 3e-g), which was consistent with the in vivo results described earlier. Inhibition of MPC-dependent OXPHOS substantially diminished the enhancement of proliferation and polyfunctionality of IL-10 HER2 CAR T cells (Extended Data Fig. 4k,l), suggesting that elevated mitochondrial OXPHOS is necessary for reinvigoration of exhausted CAR T cells. Furthermore, we used Mpci-knockout (Mpcl-KO) T cells to generate IL-10 HER2 CAR T cells, which exhibited increased frequencies of accumulated dysfunctional mitochondria and reduced ratios of OCR to ECAR compared to IL-10 HER2 CAR T cells generated from wild-type (WT) T cells (Fig. 3h-j). Moreover, Mpcl-KO IL-10 HER2 CARTcells failed to induce improved proliferation, polyfunctionality or tumor cell-killingcapacity compared to WT IL-10 HER2 CAR T cells (Fig. 3k-m). Collectively, we show that IL-10 expression promotes OXPHOS metabolism of CAR T cells in an MPC-dependent manner, which isessential for rejuvenatingexhausted CAR Tcells with enhanced proliferative capacity and effector function.
IL-10 expression alters transcription of CART cells
To investigate the impact of IL-10 expression on CART cell fate in the TME at the transcriptional level, we performed a single-cell RNA-sequencing (scRNA-seq) analysisof freshly sorted HER2 CARTor IL-10 HER2 CAR T cells from MC38-HER2 tumors (Extended Data Fig. 5a). After unsupervised clustering, four distinct clusters were identified based on their gene expression profiles (Fig. 4a and Extended Data Fig. 5b). Notably, cluster 1 was dominantly enriched in IL-10 HER2 CAR T cells, whereas HER2 CAR T cells were mostly found in cluster 0 (Fig. 4b,c). Projection of these scRNA-seq data into a reference atlas of TILs using ProjecTILs16 revealed that the majority of CAR T cells were mapped to the CD8+ terminally exhausted T cell subtype (Extended Data Fig. 5c). However, compared to cluster 0, cluster 1 displayed higher expression of cytotoxic and effector molecules, including Czmb, Gzmc, Prfl, Gzmf'and Ifng, as well as transcription factors Junjunb and Fos (Fig. 4d,e and Extended Data Fig. 5d). Gene set enrichment analysis (GSEA) between these two clusters indicated that cluster 1 was strongly associated with mitochondrial OXPHOS (Fig. 4f and Extended Data Fig. 5d). Consistent with our previous observation that IL-10-Fc enhanced mitochondrial respiration via pyruvate, we found that cluster 1 showed an enriched pyruvate metabolism pathway compared to cluster 0 (Fig. 4f). In addition, cluster 2, a cluster associated with high expression levels of cell proliferation and cycling genes (Fig. 4d and Extended Data Fig. 5d), was also enriched in IL-10 HER2 CARTcells (Fig. 4a-c), suggesting higher proliferativecapacity of IL-10 HER2 CAR T cells. These results align with our previous observations and suggest
mitochondria-enriched CAR T cells in each treatment group. The numbers indicate the frequency (%). Data are representative of two independent experiments, g, Frequencies of CAR T cells with dysfunctional mitochondria, h, The ratioof MDR/MG in tumor-infiltrating CAR T cells, i, Representative EM images of sorted intratumoral CAR T cells, j-l, Quantification of mitochondrion number per cell (j), crista numbers per mitochondrion (k) and total crista length per mitochondrial area (I) in sorted intratumoral CAR T cells as shown in i. m, Frequencies of CAR T cells with dysfunctional mitochondria in the subpopulation of different CAR densities. n,o, The experimental setting was the same as described in Fig. lm (n = 5 mice), n, Frequencies of hCAR T cells with dysfunctional mitochondria, o, The ratio of MDR/MG in tumor-infiltrating hCAR T cells from the indicated treatment groups. All data represent the mean ± s.e.m. and were analyzed by two-tailed Student's f-tests (b, d, e and j-o) or one-way ANOVA with a Tukey's multiple-comparisons test (g and h).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
that IL-10 expression rejuvenates exhausted CART cells, sustaining a cell state with enhanced cytotoxicity, effector function and prolifera-tivecapacity, which is engaged with metabolic reprogramming toward mitochondrial OXPHOS.
IL-10 CAR T cells eradicate established solid tumors
The discovery that IL-10 HER2 CAR T cells resist dysfunction motivated us to assess their efficacy against solid tumors. In a therapeutic setting of preestablished MC38-HER2 tumors in mice with
Spleen
S o E ceo
O   =3 JZ
□
i
Spleen
Tumor
		10 -i	P = 0.0037
	c		
	o		O □
iber	ndri	8 -	cm o
E	o	6 -	<D □
	si		
c	o		
co	o	4 -	o m
rist	E	2 -	o um ĽE
U	per	0 -	
2 -b'
°3>
o ■
HER2 CAR T
HER2 CAR T + IL-10
IL-10 HER2 CAR T
105 106
105 106
Soč c c o o =3 jz "'S o
P - 0.9604 P < 0.0001
oo o§8o
■v
dl
N*-
•v
HER2 CAR T
IL-10 HER2 CAR T
c/)		
bei		50 -i
E		
		40 -
c	_.	
c	0	30 -
o	O	
■a	o	20 -
c	a	
ho		10 -
o		
o		0 -
i		
1 a.
90 A44
4ŕ
U) —
c co
20
r~ 15 E
3. 10 E
3 5
o
p<	0.0001
	
-	A
	
(	> *
- o §8	
Q. TJ ■
o HER2 CAR T
ä IL-10 HER2 CART
n
Low Medium CAR density
High
Soč ceo
O'S o 10
0
3.4- ■
Íl
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Mitochondrial respiration (+MC38-HER2)
■e- HER2 CAR T
■B- HER2 CAR T + IL-10
* IL-10 HER2 CAR T
;<!>
20     40 60 Time (min)
	500
c	400
E	
"o	300
pm	200
	
q;	100
U	
O	0
o HER2 CAR
□ HER2 CAR T+IL-10
& IL-10 HER2 CART
P = 0.0256 /
P= 0.0207
P= 0.9993
JÜ1
nn
Basal
'0.3125 J A
l#l
Maximal
	Ü -6 -
<	
u	
lu	4 -
q?	
U	
O	2 -
	0 -
o HER2 CAR T
□ HER2 CAR T+IL-10
a IL-10 HER2 CART
P = 0.0372
i-1 r
P= 0.0048
P = 0.6348
□DO
Basal
P = 0.0375 P = 0.4462 J
■ f
Maximal
□ HER2 CAR T
4 IL-10 HER2 CART
IL-10 HER2 CAR T + oligomycin ♦ IL-10 HER2 CAR T + UK5099 o IL-10 HER2 CAR T+ETO
01	2.5	*104 -|
		
C	2	*104-
O		
O	1.5	*104 -
"o		
o	1	x104 -
1—		
AR	5	x103 -
O		0 -
P< 0.0001   P< 0.0001
4? *
WT
HER2 CAR T
Mpc7 KO
IL-10 HER2 CAR T
HER2 CAR T
IL-10 HER2 CAR T
30.5
104 105 106      0       104  105 106 o
104 105 106 o
104  105 106
MG-
o HER2 CAR T
A IL-10 HER2 CAR T
P< 0.0001
P= 0.2977
r i
Mpc7 KO
IL-10HER2 CART versus HER2 CAR T
IL-10 HER2 CART versus HER2 CAR T
WT Mpc7KO
Mpc7 KO
8. 2 5-
U) o
Fig. 31 IL-10 expression enhances OXPHOS in CAR T cells in an MPC-dependent manner, a-c, IL-10 HER2 CAR T cells or HER2 CAR T cells in the presence or absence of IL-10 were cocultured with MC38-HER2 cells at an effector-to-target (E:T) ratio of 5:1 for 18 h. CAR T cells were then isolated for a Seahorse assay (n = 3 biologically independent samples), a, Real-time analysis of OCR. b, Average basal and maximal OCR. c, Ratios of basal and maximal OCR to ECAR. d, Average basal OCR of IL-10 HER2 CAR T cells in coculture with MC38-HER2 cells (E:T = 5:1) in the presence of different inhibitors (n = 3 biologically independent samples); ETO, etomoxir. e-g, IL-10 HER2 CAR T cells or HER2 CAR T cells in the absence or presence of mouse IL-10 (145 ng mr1) were cocultured with MC38-HER2 cells at an E:T ratio of 0.5:1 for 48 h (n = 3 biologically independent samples). Shown are viable CAR T cell counts (e), frequencies of granzyme B*IFNy*TNF(x* polyfunctional CAR T cells (f) and percent lysis of MC38-HER2 cells (g); Ctrl T, untransduced T cells, h-m, HER2 CAR T cells and IL-10 HER2 CAR
IL-10 HER2 CART versus HER2 CAR T
WT Mpc7KO
0 HER2 CAR T
1 IL-10 HER2 CART
^ £. 100 -O
co o
„ a.
C l±j
P = 0.2369 I-1
WT
Mpc7 KO
T cells generated from WT orMpcl-KO T cells were cocultured with MC38-HER2 cells (E:T = 0.5:1) for 48 h. CAR T cells were analyzed by flow cytometry or isolated for a Seahorse assay (n = 4 biologically independent samples), h, Representative flow cytometry plots showing a dysfunctional mitochondria-enriched subpopulation of CAR T cells in each group. The numbers indicate the frequency (%). i, Frequencies of CAR T cells with dysfunctional mitochondria, j, Relative ratios of basal OCR to ECAR of WT and Mpcl-KO CAR T cells (IL-10 HER2 CAR T cells versus HER2 CAR T cells), k, Relative CAR T cell counts of WT and Mpcl-KO CAR T cells (IL-10 HER2 CAR T cells versus HER2 CAR T cells). 1, Relative frequencies of polyfunctional WT and Mpcl-KO CAR T cells (IL-10 HER2 CAR T cells versus HER2 CAR T cells), m, Percent lysis of MC38-HER2 cells. All data represent the mean ± s.e.m. and were analyzed by two-tailed Student's f-tests (i-m) or one- or two-way ANOVA with a Tukey's multiple-comparisons test (b-g). Data are representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Unsupervised clustering
0
UMAP1
HER2 CAR T IL-10 HER2 CAR T
0 5
UMAP2
I HER2 CAR T
I IL-10 HER2 CAR T
Gzmfa
Gzmc
Slfn2
Dgatl Fmb •
Cc(5
i «
itgbj !
PrfJ Gzmb
*Ctsd
•if9-' Jf"S
I **Jun ."Gzma
SppJ
,V*' *~Fos
^ ~"J(jnb "«V;d69 HspaJa Mra5 ,
Cc(3 Gzme
Gzmg
log2 (fold change)
Hallmark oxidative phosphorylation (ES = 0.53, FDR q = 2.4 x 10~5)
KEGG oxidative phosphorylation (ES = 0.41, FDR q = 0.063)
KEGG pyruvate metabolism (ES = 0.55, FDR q = 0.068)
Fig. 41 IL-10 expression modulates CART cell transcription for enhanced cytotoxicity, proliferation and mitochondrial OXPHOS. C57BL/6 mice were inoculated with MC38-HER2 tumor cells (1 x 106, s.c.) and sublethally lymphodepleted by irradiation on day 5 and received i.v. adoptive transfer of HER2 CAR T cells (3 x 106) or IL-10 HER2 CAR T cells (3 x 106) on day 6. On day 18, tumor-infiltrating HER2 CAR T cells or IL-10 HER2 CAR T cells were sorted for scRNA-seq analysis (scRNA-seq samples were pooled from five mice for the HER2 CAR T cell group and six mice for the IL-10 HER2 CAR T cell group).
12,000 16,000 Cluster 0
a, Uniform manifold approximation and projection (UMAP) of unsupervised cell clusters, b, Distribution of HER2 CAR T cells and IL-10 HER2 CAR T cells over the UMAP space, c, Cluster composition for HER2 CAR T cells or IL-10 HER2 CAR T cells, d, Single-cell expression of key marker genes over the UMAP space, e, Volcano plot showing key differentially expressed genes between clusters 1 and cluster 0. Genes with Pvalues <10~5 and with log2 (fold change) values of >0.5 are highlighted in red. f, GSEA between clusters 1 and 0 of selected signatures from the mSigDB database; ES, enrichment score; FDR, false discovery rate.
lymphodepletion preconditioning, adoptive transfer of ILIO HER2 CAR Tcells (3 x 106) induced complete tumor regression and durablecures in 90% of treated mice (Fig. 5a-c and Extended Data Fig. 6a). By contrast, adoptive transfer of HER2 CART cells alone or combined with single or multiple i.v. administrationsof native IL-10 transiently controlled tumor growth but failed to induce any durable tumor regression. In fact, treatment with ILIO HER2 CAR T cells was equivalent to the combination of HER2 CAR Tcells and ILTO-Fc (peritumoral injection) and outperformed the combination of HER2 CAR T cells and ILTO-Fc (i.v.) in therapeutic efficacy (Extended Data Fig. 6b-e). No body weight loss or overt elevation in serum cytokines was observed after ILIO HER2 CAR T cell treatment, suggesting a favorable safety profile (Extended Data Fig. 6f,g). We next
extended this strategy to the CAR specific to the melanoma-associated antigen tyrosinase-related protein-1 (TRPT) and prepared ILIO TRPT CAR Tcells (Extended Data Fig. 7a-c). ILIO TRP-1 CART cells exhibited enhanced killing efficiency of TRPT-expressing B16F10 melanoma cells in vitro compared to conventional TRPT CART cells (Extended Data Fig. 7d). To treat the poorly immunogenic and highly aggressive mouse B16F10 melanoma tumor, we transferred IL-10 TRP-1 CARTcells (3 x 106) through i.v. administration, which led to 60% tumorclearanceandcura-tive responses in the s.c. tumor model without body weight loss (Fig. 5a,d,eand Extended Data Fig. 7e,f). Furthermore, when applied totreatan orthotopic B16F10 melanoma model, IL-10 TRPT CARTcells induced complete responses in about 40% of treated mice (Extended Data Fig. 7g-j).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
We next evaluated whether IL-10-expressing CAR T cells could control metastatic cancer. We prepared epidermal growth factor receptor variant III (EGFRvlll)-targeting CAR T cells with IL-10 expression (Extended Data Fig. 8a-c). Consistent with the finding in HER2 CAR T cells with a 4-1BB endodomain, IL-10 expression also enhanced the basal OCRofCD28-incorporated IL-10 EGFRvIII CARTcells when cocul-tured with target cells (Extended Data Fig. 8d). IL-10 EGFRvIII CART cells exhibited enhanced in vitro killing efficiency of 4T1 tumor cells that were transduced to express EGFRvIII and luciferase (4Tl-EGFRvIII-Luc) compared to EGFRvIII CARTcells (Extended Data Fig. 8e). To establish a metastasis-like tumor model, BALB/c mice received i.v. injection of 4Tl-EGFRvIII-Luc cells. Mice that developed metastatic nodules in the lungs were sublethally lymphodepleted and then received theadoptive transfer of EGFRvIII CARTcells aloneor supported by IL-10 infusion or IL-10 EGFRvIII CAR T cells (Fig. 5a). Tumor growth was monitored by bioluminescence imaging (Fig. 5f). Notably, IL-10 EGFRvIII CARTcells exhibited superior antimetastasis activity and led to durable cures in 100% of treated mice with no body weight loss, whereas EGFRvIII CAR T cells with or without exogenous IL-10 infusion showed only modest tumor burden control (Fig. 5f-h and Extended Data Fig. 8f). The long-lastingsuppression of metastases was likely due to the persistence of IL-10 EGFRvIII CARTcells in the blood (Extended Data Fig. 8g).
We further extended this therapeutic strategy to hCAR Tcells. In immune-deficient mice bearing established s.c. Raji or PANC1-CD19 tumors, transfer of conventional CD19 hCAR T cells (1 x 106) resulted in minimum or transient tumor growth inhibition, respectively, but ultimately failed to control tumor progression (Fig. Si—1 and Extended Data Fig. 9a,b). By contrast,all mice treated with IL-10-expressingCD19 hCAR T cells exhibited complete tumor regression in both models with no recurrence (Fig. Si—1 and Extended Data Fig. 9a,b), suggesting that IL-10-expressinghCARTcells have superior antitumor capacities against solid tumors in xenograft models. In addition, weestablished an orthotopic human tumor model of pancreatic ductal adenocarcinoma (PDAC) with PANC1-CD19-Luc tumor cells inoculated in the mouse pancreas (Extended Data Fig. 9c-e). IL-10 CD19 hCAR Tcells (1 x 10s) eliminated orthotopic PDAC tumors much more effectively than conventional CD19 hCAR T cells (1 x 10s), leading to complete responses in 100% of treated mice (Extended Data Fig. 9c-e). Notably, IL-10 CD19 hCAR T cells demonstrated markedly higher expansion after hCAR T cell infusion until at least day 50 (Extended Data Fig. 9f,g). Collectively, these results suggest that IL-10-expressing CAR T cells represent a potent immunotherapy against multiple solid tumors.
IL-10 expression induces stem cell-like memory
To investigate whether IL-10-expressing CARTcells developed antitumor immune memory, we rechallenged the surviving mice 3 months after adoptive CAR T cell transfer (Extended Data Fig. 10a). All long-term survivors treated with IL-10 HER2 CAR T cells, IL-10 TRP-1 CAR T cells or IL-10 EGFRvIII CART cells rejected the second challenge of the original tumor cells (Fig. 6a-c). Similarly, in the human PDAC tumor model, after orthotopic tumor cell rechallenge, all survivors from the
Fig. 51 Infusion of IL-10 CART cells mediates solid tumor clearance in multiple syngeneic and xenograft models, a-c, C57BL/6 mice were inoculated (s.c.) with MC38-HER2 colon cancer cells (3 x 10s) and lymphodepleted and received IL-10 HER2CART cells (3 x 106, i.v.) or HER2 CARTcells (3 x 106, i.v.) alone or with IL-10 as a single dose (1 \ig i.v. on day 6) or multiple doses (3.5 \ig i.v. on days 6,8,10,12 and 14; n = 10 mice for PBS, HER2 CAR T cells and IL-10 HER2 CAR T cells and n = 5 mice for other groups), a, Experimental timeline, b, Average tumor growth curves, c, Survival curves. d,e, Similar to as shown in a. C57BL/6 mice were inoculated (s.c.) with B16F10 melanoma cells (3 x 105) and lymphodepleted and received IL-10 TRP-1 CAR T cells (3 x 106, i.v.) or TRP-1 CAR T cells (3 x 106, i.v.) alone or with IL-10 (1 \ig, i.v.) on day 6 (n = 5 mice), d, Average tumor growth curves, e, Survival curves, f-h, Similar to as shown in a. BALB/c mice were inoculated (i.v.) with 4Tl-EGFRvIII-Luc breast cancer cells (5 x 10+) and
treatment group that received IL-10 CD19 hCART cells rapidly rejected the second challenge (Extended Data Fig. 9f,h,i). This robust immune memory response motivated us to examine the memory phenotypes of IL-10-expressing CARTcells in lymphoid tissues and the circulation (Fig. 6d). Flow cytometry analysis of CAR T cells (initially transferred CAR T cells were all CD44hi) 12 d after treatment showed that IL-10 HER2 CARTcellsenriched for a population with a Tscm cell phenotype (defined as CD62LhiCD44'° and Sca-1+CD122+)1720 in the spleen and peripheral blood (Fig. 6e,f and Extended Data Fig. 10b,c). IL-10 HER2 CART cells in the spleen showed -3.2-fold higher frequencies of CD62LhiCD44'° T cells than HER2 CART cells alone, among which the majority (-71.2%) were Sca-1+CD122+ Tscm cells (Fig. 6e,f). In addition, IL-10 HER2 CAR Tcells exhibited substantially increased expression levelsofSca-1 compared to HER2 CAR T cells alone or HER2 CAR T cells plus exogenous IL-10 among all CD62LhiCD44'° CAR T cells (Fig. 6g). Similarly, IL-10 EGFRvIII CAR T cells promoted a higher frequency of Tscm cells in the spleen and bone marrow than EGFRvIII CARTcells (Extended Data Fig. lOe-i). This finding was further confirmed by the observation that IL-10 HER2 CARTcells werecomposed of -3.7- and 2.6-fold higher proportions of IL-7Ra+KLRGr long-lived memory precursor T cells21 than HER2 CARTcells in the spleen and blood, respectively (Fig. 6h and Extended Data Fig. lOd). GSEA confirmed that splenic IL-10 HER2 CAR T cells exhibited a clear transcriptomic state of'Sternness/Memory' (identified by mouse Sell, Tcf7,Lefl,H7r and CcrT) and were highly enriched for Tscm cell-associated genes, including/d2, Ly6a, Cxcr3, Eomes and Bcl2 (Fig. 6i,j and Extended Data Fig. 10j)22. Additionally, we observed that, compared to CD19 hCAR T cells, treatment of IL-10 CD19 hCAR T cells appeared enriched with Tscm cells, inducing 22.4-and 31.9-fold higher counts of CD8+ CAR T cells with a Tscm cell phenotype (defined by CD45RA+CD27+CD62L+CD95+CCR7+) in the spleen and bone marrow, respectively (Extended Data Fig. lOk-m). Furthermore, scRNA-seq analysisconfirmed that IL-10 CD19 hCART cells in the spleen had a markedly higher fraction of cells with the gene signature 'Sternness/Memory' (identified by human TCF7, CCR7, LEF1 and SELL) and expressed higher levels of Tscm cell-associated genes, including IL7R and CXCR3, than conventional CD19 hCAR Tcells (Fig. 6k-n and Extended Data Fig. lOn). Together, these results provideclear evidence that IL-10 signaling may induce the formation of mouseand human Tscm CAR T cells that contribute to long-term antitumor immunity.
Discussion
Metabolic modulation is a promising strategy to counter T cell dysfunction for improved CARTcell immunotherapy. Among different strategies, increasing mitochondrial fitness and OXPHOS may be an effective intervention to improve persistence and function of CAR T cells and other adoptive T cell therapies. In the design of second-generation CARTcells, 4-1BB co-stimulation has been reported to enhance mitochondrial biogenesis and OXPHOS and ameliorate T cell exhaustion and therefore induces better therapeutic outcomes than the CD28 co-stimulatory domain23,24. Moreover, preconditioning T cells with IL15 (ref. 11), coexpressing IL15 (ref. 25) or deletingZC3«22/l26 could
lymphodepleted and received IL-10 EGFRvIII CAR T cells (3 x 106, i.v.) or EGFRvIII CAR T cells (3 x lo6, i.v.) alone or with IL-10 (1 \ig, i.v.) on day 6 (n = 5 mice), f, Bioluminescent imaging, g, Individual radiance (p s_1 cnT2 sr1). h, Survival curves, i-l, NSG mice were inoculated s.c. with Raji lymphoma cells (1 x lo6) or PANC1-CD19 human pancreatic tumor cells (2 x lo6) and received CD19 hCAR T cells or IL-10 CD19 hCAR T cells (1 x lo6, i.v.) on day 8 (n = 5 mice for the Raji model and n = 9 mice for the PANC1-CD19 model), i, Experimental timeline. j,k, Shown are average tumor growth curves (j) and survival curves (k) of the Raji model. I, Tumor growth curves of the PANC1-CD19 model. Indicated are the numbers of tumor-free mice per total number of mice in the group (c, e, h, kand I). Data are representative of two independent experiments (d-h). All data represent the mean ± s.e.m. and were analyzed by two-way ANOVA with a Tukey's multiple-comparisons test (b, d, g, j and I) or log-rank test (c, e, h and k).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
improve antitumor immunity ofT cells through enhanced mitochondrial fitness and OXPHOS. Here, we developed metabolically armored IL-10-secreting CAR T cells, which exhibited enhanced proliferation and effector function by sustaining mitochondrial fitnessand promoting OXPHOS, leading to complete remission in multiple syngeneic and xenograft solid tumor models. Moreover, ILIO secretion in CAR
T cells induced Tscm cell responses in the peripheral blood and spleen, which bestowed durable protection in treated mice against tumor rechallenge.
Although enhancement of pyruvate-dependent OXPHOS was critical for countering CARTcell dysfunction, OXPHOS activities that rely on other substrates, in particular, fatty acids, can also preserve
Inoculation of MC38-HER2 (s.c.)/ B16F10 tumors (s.c.) in C57BL/6 or 4T1-EGFRvlll-Luc tumors (i.V.) in BALB/c mice
Soluble IL-10 (i.V.)
.....       Single dose 1     1     i    I     1  or multiple doses
5 6 8 10 12 14 Time(d) *.*
Lymphodepletion  ^ |
HER2 CAR T/TRP-1 CAR T/EGFRvlll CAR T (i.V.)
b -♦- PBS
■O- HER2 CAR T
t> HER2 CAR T + IL-10 (1 |jg) ♦ HER2 CAR T + IL-10 (3.5 lig x 5) IL-10 HER2 CART
160 -
0 10 20 30 40
Time after inoculation (d)
C —
PBS
HER2 CAR T
HER2 CAR T+ IL-10 (1 lig)
HER2 CAR T+ IL-10 (3.5 |jg x 5) ■
IL-10 HER2 CART
0        20       40       60 Í Time after inoculation (d)
-O- PBS
<■- TRP-1 CAR T
<*- IL-10 TRP-1 CAR T
0 10 20 30 40
Time after inoculation (d)
PBS
TRP-1 CART IL-10 TRP-1 CAR
0        20       40       60 Í Time after inoculation (d)
PBS
EGFRvlll CART
EGFRvlll CAR T+ IL-10        IL-10 EGFRvlll CART
Day 6
Day 9
Day 12
Day 15
Day 18
Day 24
Day 75
Inoculation of PANC1-CD19 /Raji tumor cells (s.c.) in NSG mice
Raji
+■ PBS
■D- CD19 hCAR T
■*■ IL-10 CD19 hCAR T
V 0
8 Time (d)
CD19 hCART/ IL-10 CD19 hCAR T (i.V.)
Raji
— PBS
— CD19 hCAR T i
— IL-10 CD19hCART '
0 20 40 60
Time after inoculation (d)
0 20 40 60 8< Time after inoculation (d)
PBS
EGFRvlll CART
EGFRvlll CAR T +IL-10 (1 |jg)
IL-10 EGFRvlll CART
0 20 40 60
Time after inoculation (d)
— PBS
— EGFRvlll CART
— EGFRvlll CAR T+IL-10 (1 lig)
— IL-10 EGFRvlll CAR T
J n
60 40 20 0
0
0 20 40 60
Time after inoculation (d)
PANC1-CD19 (s.c.)
-♦- PBS
* CD19 hCAR T
IL-10 CD19hCART
0 20 40 60
Time after inoculation (d)
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
60 40 20 0
MC38-HER2
■ Naive Cured ' (IL-10 HER2 CAR T) 1
A/A
0 20 40 60
Time after rechallenge (d)
60 40 20 0
B16F10
■ Naive
Cured I P < 0.0001
(IL-10 TRP-1 CAR T) '
_3/3
4T1-EGFRvlll-Luc
0 20 40 60
Time after rechallenge (d)
■ Naive Cured
(IL-10 EGFRvlll CART)
60 40 20 0
5/5
0 20 40 60
Time after rechallenge (d)
d       Inoculation of MC38-HER2 tumors (s.c.) in C57BL/6 mice
\
IL-10 (i.V.)
Kill and analyze
\
Lymphodepletion *•*     ♦        HER2 CAR T/
I   IL-10 HER2 CAR T (i.V.)
Time (d)
^r ^r
8 80
ČN 60
CO Q
U 40
20
Spleen
P < 0.0001 -1
P < 0.0001
P = 0.5555
55 j
i
Spleen
Stemness/Memory (Sell, Left, \l7r, Tcf7, Ccr7)
Spleen
< 0.0001
UMAP1
UMAP1
Inoculation of PANC1-CD19 tumors   Kill and (s.c.) in NSG mice analyze
0      8     39 Time (d) <
CD19 hCART/ IL-10 CD19 hCAR T (i.V.)
2.5 0
-2.5 -5.0
I CD19 hCAR T
I IL-10CD19hCART
-5 0 UMAP1
Stemness/Memory (TCF7, CCR7, LEFT, SELL)
Stemness/Memory
0.3 0.2 0.2
ľ o,
• CD19 hCART
• IL-10CD19hCART
proliferation and function of tumor-infiltrating T cells2728. Therefore, how specific metabolic pathways determine T cell differentiation remains unclear. Recently, it has been reported that CAR T cells engineered to express supportingcytokines29~31,checkpointblockadeanti-body32 or c-Jun33 or targetinga CAR to the 77?/lClocus34 resist exhaustion or dysfunction. Yet, whether and how these strategies alter T cell metabolism remains to be determined. Compared to most existing armored CARTcells, IL-10 expression in CARTcells enhanced proliferation and
function of terminally exhausted CD8+Tcells(PDrTIM-3+), a subset of exhausted T cells that lack responses to other stimulatory cytokines8 or checkpoint blockades35 but are direct killers of cancer cells owing to their toxicity1235. Therefore, IL-10 expression complements existing strategies for developing dysfunction-resistant CAR Tcell therapy.
We previously showed that multiple local administrations of the half-life-extended ILTO-Fc fusion protein are highly effective in reinvig-oratingintratumoral exhausted Tcells and greatly enhance therapeutic
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Fig. 61 IL-10-expressing CAR T cells confer a memory response against tumor rechallenge and exhibit a stem cell-like memory phenotype in lymphoid organs. Survivors after treatment with IL-10 HER2 CAR T cells, IL-10 TRP-1 CAR T cells and IL-10 EGFRvIII CAR T cells were rechallenged s.c. with MC38-HER2 (1 x IO6), B16F10 (1 x IO5) and 4Tl-EGFRvIII-Luc (5 x IO5) cells, respectively, on day 90 after primary tumor inoculation. Naive WT mice (n = 5 mice) received an equal number of tumor cells as controls, a-c, Shown are survival curves of the rechallenged MC38-HER2 (a), B16F10 (b) and 4Tl-EGFRvIII-Luc (c) tumor models. Indicated are the numbers of tumor-free mice per total mice in each group, d-j, C57BL/6 mice were inoculated with MC38-HER2 cells (1 x IO6, s.c.) and lymphodepleted and received IL-10 HER2 CAR T cells (3 x io6, i.v.) or HER2 CAR T cells (3 x io6, i.v.) alone or with IL-10 (1 ug, i.v.) on day 6. Mice were killed on day 18 for flow cytometry analysis of CAR T cells in the spleen and blood (n = 5 mice) and for scRNA-seq analysis of splenic IL-10 HER2 CAR T cells (samples were pooled from five mice), d, Experimental timeline. e,f, Average frequencies
efficacy in combination with adoptive cell therapies, including CAR T cell therapy. Efficient delivery and accumulation of IL-10 in the TME seems crucial because half-life extension through the fusion protein design, local injection and multiple dosages are all required. In this report, we achieved highly efficient intratumoral delivery of IL-10 by engineering CAR T cells with IL-10 expression. CAR T cells secreting a high level of IL-10 likely had the capacity to maintain an effective local concentration of IL-10 in the TME for invigoration of tumor-infiltrating CARTcells. By contrast, i.v.-administered exogenous recombinant IL-10 (without Fc fusion) of an equivalent dose showed minimal effects on intratumoral Tcells in vivo likely due to its short half-life and low level oftumoraccumulation.Consequently,a single doseofIL-10-expressing CAR T cells through systemic administration led to equivalent therapeutic efficacy as observed for the combination therapy of CAR Tcells in tandem with multiple intratumoral injections of IL-10-Fc12. Therefore, using IL-10-expressing CAR T cells as a monotherapy may be a promising future clinical strategy, particularly when intratumoral injection is not feasible.
We found that the heterogeneity of CARexpression density leads to sharp differences in T cell phenotypes, function and extent of responses to secreted IL-10. In general, CAR T cells with higher CAR density expressed higher levels of inhibitory receptors (PD-1 and TIM-3), displayed reduced effector function and responded more prominently to IL-10. These results suggest that strong antigen stimulation signals likely drive T cell exhaustion314; CAR density may therefore be an important parameter to control in the development of CART cells with well-balanced tumoricidal activities and resistance toward antigen stimulation-induced exhaustion and dysfunction13. IL-10 was most effective toward the highly exhausted subset with the highest CAR density for metabolic reprogrammingandreinvigoration but showed fewer effects in the scenario with low strength of antigen stimulation. These results were in line with the in vitro observation that exogenous or secreted IL-10 failed toenhance OXPHOS in CART cells without antigen stimulation. It has been recently reported that IL-10 receptor-STAT3 signaling controls the balance between PD-lhl exhausted CD8+ T cells and functional PD-rntTCF 1+CD8+Tcells by limiting excessive activation of CD8+Tcells36. In a microenvironment of low tumor antigen burden, such as the peripheral blood or healthy tissues, secreted IL-10 likely has fewer effects in expandingor activating CARTcells, thus inducing fewer risks of systemic toxicities. This intrinsic safeguard mechanism could be an additional advantage for IL-10-expressing CAR T cells for potential clinical application.
The IL-10 receptor-STAT3 pathway has been reported to be important for T cell memory formation. IL-10 deficiency leads to a markedly reduced frequency of KLRGlloIL-7Rhi memory precursor CD8+ T cells in acute infection37. Furthermore, IL-10 produced from either Treg cells or CDllc+dendritic cells during the resolution of infection promotes the development of KLRGll0CD127hi and CD62LhiKLR-Gl'° central memory CD8+ T cells38. Here, we observed that constitutive
of CD62LhiCD4410 cells among total splenic CAR T cells (e) and Sca-1*CD122* cells among CD62Ll"CD44'° CAR T cells (f). g, Representative flow cytometry plots and Sca-1 MFI among CD62LhiCD44'° splenic CAR T cells, h, Frequencies of IL-7Ra*KLRGr cells among total splenic CAR T cells, i, UCell gene signature scores of IL-10 HER2 CAR T cells for the gene set 'Sternness/Memory', j, Single-cell expression of key marker genes over UMAP maps, k-n, NSG mice were inoculated with PANC1-CD19 cells (2 x io6, s.c.) and received i.v. adoptive transfer of CD19 hCAR T cells or IL-10 CD19 hCAR T cells (1 x IO6, i.v.) on day 8. On day 39, mice were killed for scRNA-seq analysis of splenic hCAR T cells (samples were pooled from five mice), k, Experimental timeline. 1, Distribution of hCAR T cells over UMAP maps, m, UCell gene signature scores of hCAR T cells for the gene set 'Sternness/ Memory', n, Frequencies of hCAR T cells with the 'Sternness/Memory' signature among total hCAR T cells. All data represent the mean ± s.e.m. and were analyzed by one-way ANOVA with a Tukey's multiple-comparisons test (e-h) or log-rank test (a-c).
expression of IL-10 by CAR T cells induced the loss of CD44 expression, and the majority of IL-10-secreting CARTcells acquired a stem cell-like memory phenotype (CD44loCD62LhiSca-l+CD122+) during tumor clearance. Tscm cells possess superior survival capacity, robust proliferative potential after antigen reexposure and the ability to give rise to all memory and effector T cell subsets18-20. Promoting sternness is essential for durable protection against relapse with CAR Tcell therapy. Additionally, this finding suggests that controlling the strength and persistence of IL-10 signaling may fine-tune CAR T cell differentiation states. Metabolically, memory T cells typically have higher mitochondrial mass and respiration spare capacity3940, which is in agreement with our observation that IL-10-expressing CART cells have increased mitochondrial activity compared to conventional CARTcells.
Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; detailsof author contri-butionsandcompetinginterests;and statements of data and codeavail-ability are available at https://doi.org/10.1038/s41587-023-02060-8.
References
1. Hou, A. J., Chen, L. C. & Chen, Y. Y. Navigating CAR-T cells through the solid-tumour microenvironment. Nat. Rev. Drug Discov. 20, 531-550 (2021).
2. Blank, C. U. et a I. Defining 'Tcell exhaustion'. Nat. Rev. Immunol. 19, 665-674(2019).
3. Wherry, E. J. & Kurachi, M. Molecular and cellular insights into Tcell exhaustion. Nat. Rev. Immunol. 15,486-499 (2015).
4. Wherry, E. J. etal. Molecular signature of CD8* Tcell exhaustion during chronic viral infection. Immunity 27, 670-684 (2007).
5. Sen, D. R. etal. The epigenetic landscape of Tcell exhaustion. Science 354,1165-1169 (2016).
6. DePeaux, K. & Delgoffe, G. M. Metabolic barriers to cancer immunotherapy. Nat. Rev. Immunol. 21,785-797 (2021).
7. Scharping, N. E.etal. Mitochondrial stress induced by continuous stimulation under hypoxia rapidly drives Tcell exhaustion.
Nat. Immunol. 22,205-215 (2021).
8. Yu, Y.-R. et al. Disturbed mitochondrial dynamics in CD8*TILs reinforce T cell exhaustion. Nat. Immunol. 21,1540-1551 (2020).
9. Vardhana, S. A. et al. Impaired mitochondrial oxidative phosphorylation limits the self-renewal of T cells exposed to persistent antigen. Nat. Immunol. 21,1022-1033 (2020).
10. Mo, F. et al. An engineered IL-2 partial agonist promotes CD8* T cell sternness. Nature 597, 544-548 (2021).
11. Alizadeh, D. etal. IL15 enhances CAR-T cell antitumor activity by reducing mTORCI activity and preserving their stem cell memory phenotype. Cancer Immunol. Res. 7,759-772 (2019).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
12. Guo, Y. et a I. Metabolic reprogramming of terminally exhausted CD8* Tcells by IL-10 enhances anti-tumor immunity.
Nat. Immunol. 22,746-756 (2021).
13. Scarfo, I. & Maus, M. V. Unraveling the signaling balance of activation and exhaustion of CAR Tcells. Cancer Cell 37,143-144 (2020).
14. Mueller, S. N. & Rafi, A. High antigen levels are the cause of T cell exhaustion during chronic viral infection. Proc. Natl Acad. Sci. USA 106, 8623-8628 (2009).
15. Liu, X. & Peng, G. Mitochondria orchestrate T cell fate and function. Nat. Immunol. 22,276-278 (2021).
16. Andreatta, M.etal. Interpretation of T cell states from single-cell transcriptomics data using reference atlases. Nat. Commun. 12, 2965 (2021).
17. Gattinoni, L. et al. Wnt signaling arrests effector T cell differentiation and generates CD8* memory stem cells. Nat. Med. 15, 808-813 (2009).
18. Flynn, J. K. & Gorry, P. R. Stem memory T cells (Tscm)—their role in cancer and HIV immunotherapies. Clin. Transl. Immunol. 3, e20 (2014).
19. Buck, M. D. etal. Mitochondrial dynamics controls T cell fate through metabolic programming. Cell 166, 63-76 (2016).
20. Gattinoni, L. et al. A human memory T cell subset with stem cell-like properties. Nat. Med. 17,1290-1297 (2011).
21. Herndler-Brandstetter, D.etal. KLRG1* effector CD8* Tcells lose KLRG1, differentiate into all memory T cell lineages, and convey enhanced protective immunity. Immunity 48,716-729.e8 (2018).
22. Andreatta, M. & Carmona, S. J. UCell: robust and scalable single-cell gene signature scoring. Comput. Struct. Biotechnol. J. 19,3796-3798 (2021).
23. Kawalekar, O. U. et al. Distinct signaling of coreceptors regulates specific metabolism pathways and impacts memory development in CAR Tcells. Immunity 44,380-390 (2016).
24. Long, A. H. etal. 4-1BBcostimulation amelioratesTcell exhaustion induced by tonic signaling of chimeric antigen receptors. Nat. Med. 21, 581-590 (2015).
25. Hurton, L. V. et a I. Tethered IL-15 augments antitumor activity and promotes a stem-cell memory subset in tumor-specific Tcells. Proc. Natl Acad. Sci. USA 113, E7788-E7797 (2016).
26. Wei, J. et al. Targeting REGNASE-1 programs long-lived effector T cells for cancer therapy. Nature 576,471-476 (2019).
27. Zhang, Y. et al. Enhancing CD8* T Cell Fatty Acid Catabolism within a Metabolically Challenging Tumor Microenvironment increases the Efficacy of Melanoma Immunotherapy. Cancer Cell 32,377-391.e9(2017).
28. Chowdhury, P. S., Chamoto, K., Kumar, A. & Honjo, T. PPAR-induced fatty acid oxidation in T cells increases the number
of tumor-reactive CD8* Tcells and facilitates anti-PD-1 therapy. Cancer Immunol. Res. 6,1375-1387 (2018).
29. Ma, X. et al. lnterleukin-23 engineering improves CAR T cell function in solid tumors. Nat. Biotechnol. 38,448-459 (2020).
30. Adachi, K. et al. IL-7and CCL19 expression in CAR-T cells improves immune cell infiltration and CAR-T cell survival in the tumor.
Nat. Biotechnol. 36,346-351 (2018).
31. Avanzi, M. P. et al. Engineered tumor-targeted T cells mediate enhanced anti-tumor efficacy both directly and through activation of the endogenous immune system. Cell Rep. 23, 2130-2141 (2018).
32. Rafiq, S. etal. Targeted delivery of a PD-1-blocking scFV by CAR-T cells enhances anti-tumor efficacy in vivo. Nat. Biotechnol. 36, 847-858(2018).
33. Lynn, R. C. etal. c-Jun overexpression in CARTcells induces exhaustion resistance. Nature 576,293-300 (2019).
34. Eyquem, J. et al. Targeting a CAR to the TRAC locus with CRISPR/ Cas9 enhances tumour rejection. Nature 543,113-117 (2017).
35. Miller, B. C. et al. Subsets of exhausted CD8*Tcells differentially mediate tumor control and respond to checkpoint blockade. Nat. Immunol. 20,326-336 (2019).
36. Hanna, B. S. et al. lnterleukin-10 receptor signaling promotes the maintenance of a PD-1 intTCF-rCD8*Tcell population that sustains anti-tumor immunity. Immunity 54,2825-2841.e10 (2021).
37. Cui, W., Liu, Y, Weinstein, J. S., Craft, J. & Kaech, S. M. An interleukin-21-interleukin-10-STAT3 pathway is critical for functional maturation of memory CD8*Tcells. Immunity 35, 792-805 (2011).
38. Laidlaw, B. J. et al. Production of IL-10 by CD4* regulatory Tcells during the resolution of infection promotes the maturation of memory CD8*Tcells. Nat. Immunol. 16, 871-879 (2015).
39. Bordon, Y. IL-15 provides breathing space for memory. Nat. Rev. Immunol. 12,77 (2012).
40. Li, W. & Zhang, L. Rewiring mitochondrial metabolism for CD8* Tcell memory formation and effective cancer immunotherapy. Front. Immunol. 11,1834 (2020).
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to Springer Nature America, Inc. 2024
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Methods
Ethics statement
Experimentsand handling of micein syngeneic tumormodels werecon-ducted under federal, stateand local guidelines with approval from the Swiss authorities (Canton of Vaud, animal protocol IDs 3206 and 3533) and performed in accordance with the guidelines from CPG of EPFL and the animal facility of University of Lausanne. Human peripheral blood mononuclear cells were received from healthy volunteers with written informed consent, and the protocol was approved by the Ethics Committeeof Zhejiang University School of Medicine (no. 2020-003). All NSG mouse experiments in xenograft models were approved by the Institutional Animal Care and Use Committee of Zhejiang University (animal protocol no. 20220178).
Mice
Five- to 6-week-old female C57BL/6 (C57BL/6J) mice and BALB/c (BALB/cByJ) mice were purchased from Charles River Laboratories. Six- to 12-week-old N0D-scid/It2rg"u" (NSG) mice were purchased from GemPharmatech. T cell antigen receptor-transgenic OT-I mice (C57BL/6-Tg(TcraTcrb)1100Mjb/J) wereoriginally purchased from Jackson Laboratory and maintained in the Ecole Polytechnique Federale de Lausanne(EPFL) Center of PhenoGenomics(CPG)animalfacility./Wpcln/fl mice were crossed to Cd4cre mice on an OT-I background to generate Mpcl-KO OT-I mice. Experimental procedures in mouse studies were approved by the Swiss authorities (Canton of Vaud, animal protocol IDs 3206 and 3533) and Institutional Animal Care and Use Committee of Zhejiang University (animal protocol no. 20220178) and were performed in accordance with the guidelines from the CPG of EPFL, the animal facility of University of Lausanne and Zhejiang University.
Cells and tumor models
HER2-transduced MC38 mouse colon cancer cells (MC38-HER2), CD19-expressing PANC1 epithelioid cells (PANC1-CD19) and CD19-expressing NIH/3T3 cells (NIH/3T3-CD19) were generated according to previous publications41,42. B16F10 melanoma cells, Phoenix-Eco cells and Raji cells were originally acquired from ATCC, and 4Tl-EGFRvIII-Luc mouse breast cancer cells were provided by D.J. Irvine (Massachusetts Institute of Technology). All mouse tumor cells and PANC1-CD19 and NIH/3T3-CD19 cells were cultured in complete DMEM (DMEM (Gibco/Thermo Fisher Scientific) supplemented with 10% (vol/vol) fetal bovine serum (FBS; Gibco/Thermo Fisher Scientific) and 1% (vol/vol) penicillin/streptomycin (Gibco/Thermo Fisher Scientific)). Raji cells were maintained in RPMIT640 supplemented with 10% (vol/vol) FBS (Gibco/Thermo Fisher Scientific) and 1% (vol/vol) penicillin/streptomycin (Gibco/Thermo Fisher Scientific). MC38-HER2 (3 x 10s) and B16F10 tumor cells (3 x 10s) were implanted s.c. into the right flanks of C57BL/6 WT mice to establish the syngeneic s.c. tumor models. B16F10 tumor cells (2 x 10s) werealso implanted intradermally into the right flanks of C57BL/6 WT mice to establish the syngeneic orthotopic melanoma model. 4T-EGFRvlll-Luc tumor cells (5 x io4) were injected i.v. into BALB/c WT mice. PANC1-CD19 tumor cells (2 x io6) and Raji tumor cells (1 x 106) were inoculated s.c. into the right flanksof NSG mice to establish the xenograft tumor models. PANCTCD19-Luc tumor cells (2 x 106) wereorthotopically implanted into the tails of the pancreas of NSG mice to establish an orthotopic PDAC model.
Construction of mouse and human CARs
CAR constructs targeting HER2 were generated as previously described41. Briefly, the complete HER2 CAR sequence was composed of a mouse CD8 signal peptide, antigen-specific scFv (4D5 scFv), mouse CD8a hinge and transmembrane domain, 4TBB co-stimulatory domain and CD3Z, intracellular domain. Constructs for CARs targeting TRPT and EGFRvIII were provided by D.J. Irvine. Thecomplete CAR sequence was composed of a mouse CD8 signal peptide, antigen-specific scFv (TA99 scFv for monospecific TRP-1 CAR and 139scFv for EGFRvIII CAR),
mouse CD8a hinge and transmembrane domain, CD28co-stimulatory domain and CD3(;intracellular domain. To facilitate CAR detection by flow cytometry, a c-Myc tag was inserted between the scFv and CD8a hinge for the EGFRvIII CAR and TRP-1 CAR. The CD19 hCAR comprised an scFv specific for human CD19 preceded by a CD8a leader peptide and followed by CD28 hinge-transmembrane-intracellular regions and a CD3Z, intracellular domain, as described in previous publications34. Full-length mouse 1110 (accession number NP 034678.1) and human IL10 (accession number NM 000572.2) were amplified by PCR from cDNAclonescustom purchased from Twist Bioscienceand Gene-script, respectively. IL-10 CAR constructs were generated by fusing2A self-cleaving peptide and//20genefragments into the CAR containing viral vector (pMSGV for HER2 CAR, pMSCV for TRP-1 CAR and EGFRvIII CAR and pAAV for CD19 hCAR).
Retrovirus production
Retrovirus production was performed using the calcium phosphate method following the manufacturer's protocol (Clonetech). Briefly, Phoenix cells were seeded in a 10-cm dish and cultured for 18 h before transfection and replenished with 10 ml of prewarmed medium without disturbing the cells before transduction. For each transfection, 14 pg of plasmid (8.5 pg of CAR plasmid plus 5.5 pg of pCL-Eco packaging plasmid) was added to 628 pi of ultrapure water, followed by the addition of 72 pi of a CaCI2 solution (2 M; Sigma-Aldrich). HEPES buffered saline (700 pi x2;Sigma-Aldrich) was then added in a dropwise manner with gentle vortexing. After a 20-min incubation at 25 °C, the transfection mixture was gently added to Phoenix cells and replenished with 10 ml of prewarmed medium 16 h later. After 36-72 h of incubation, virus-containing supernatant was collected and passed through a 0.45-pm filter (Merck Millipore) to removecell debris. Virus-containing supernatant was then aliquoted and stored at -80 °C.
Preparation of mouse CART cells
Spleens from WT OT-I mice were disintegrated mechanically and filtered through a 70-pm strainer (Fisher Scientific). Red blood cells (RBCs) were lysed with ACK lysis buffer (2 ml per spleen; Gibco/Thermo Fisher Scientific) for 5 min at 25 °C. The splenocytes were washed once with cold complete RPMI medium (RPMI-1640 (Gibco), FBS (10% (vol/vol); Gibco/Thermo Fisher Scientific), HEPES (1% (vol/vol); Gibco/Thermo Fisher Scientific), penicillin/streptomycin (1% (vol/ vol); Gibco/Thermo Fisher Scientific), sodium pyruvate (1% (vol/vol); Gibco/Thermo Fisher Scientific) and 2-mercaptoethanol (0.1% (vol/ vol); Gibco/Thermo Fisher Scientific)) and resuspended at a cell density of 2 x io6 cells per ml in complete RPMI medium supplemented with IL-2 (10 ng ml"1; PeproTech) and IL-7 (10 ng ml"1; PeproTech). For WT T cell activation, six-well plates were precoated with 1 ml of anti-CD3 (1 pg ml"1; 17A2, BioXCell) and anti-CD28 (5 pg ml"1; 37.51, BioXCell) per well at 4 °C for 18 h. OT-I T cells were activated with OVA2S7_264 peptide (1 pM;GenScript). Tcells wereenriched by using Ficoll-Paque PLUS (GE Healthcare) and seeded onto precoated six-well plates at 3 x 106 cells per well in complete RPMI medium (3 ml per well). Cells were cultured at 37 °C for 48 h without disturbance. Six-well plates were coated with protamine (10 pg ml"1; Sigma-Aldrich) for 24 h before transduction. On day 2, protamine-coated plates were blocked with FBS (0.5% (vol/ vol)) containing PBS for 30 min before use. Virus supernatant was added into each well of the blocked plates (1 ml per well), and plates were centrifuged at 2,000gfor 2 h at 32 °C. Activated Tcells were collected, enriched again by using Ficoll-Paque PLUS and resuspended at 2 x io6 cells per ml in complete RPMI medium supplemented with IL-2 (10 ng ml"1). The cell suspension described above was added to virus-enriched plates (1 ml per well) and mixed by gentle shaking; spin transfection was performed at 500gfor 30 min at 32 °C. Plates were then transferred to an incubator and maintained overnight. A second transduction was conducted on day 3, and Tcells were collected and added into virus-containing plates after centrifugation. After overnight
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
incubation,Tcells were collected and expanded in fresh complete RPMI medium supplemented with IL-2 (10 ng mr1). On day5, the transduction efficiency was determined by flow cytometry analyses. For HER2 CAR T cells, CAR expression was detected by using a biotinylated human HER2/ERBB2 protein (ACROBiosystems) and the secondary staining reagentStreptavidin-PE/Cyanine7(BioLegend).TRP-lCARor EGFRvlll CAR expression was evaluated by surface staining of the c-Myc tag using anti-Myc (9B11, Cell Signaling). Untransduced T cells activated and cultured in parallel with virus-free medium were used as a control. CART cells were used on day 6 for both in vitro and in vivo experiments.
Generation of hCAR T cells
Peripheral blood was obtained from healthy volunteers. All blood samples were handled following the required ethical and safety procedures. Peripheral blood mononuclear cells were isolated by density gradient centrifugation, and T cells were purified using a Pan T Cell Isolation kit (Miltenyi Biotec). T cells were then stimulated with CD3/ CD28 T cell Activator Dynabeads (Thermo Fisher Scientific) and cultured in X-VIVO15 serum-free medium (Lonza) supplemented with FBS (10% (vol/vol); HyClone), penicillin/ streptomycin (1% (vol/vol); HyClone), human IL-7(5 ng ml"1; Novoprotein) and human IL-15 (5 ng ml" Novoprotein). Human T cells were stimulated for 48 h and debeaded for gene targeting experiments. Gene targeting experiments were performed as described previously34. In short, modified guide RNA (C*A*G*GGUUCUGGAUAUCUGU; * represents 2'0-methyl 3' phospho-rothioate) targeting the first exon of the constant chain of the TRACgene was synthesized by Genescript. Cas9 protein was purified fromEscheri-chia coll strain Rosetta by using a heparin column and a Superdex 200 increase column on an AKTAPure system (GE Healthcare), as described previously43. Adeno-associated viruses containing the CAR gene were produced by Vigene Biosciences. Ribonucleoprotein complexes were produced by complexing two components: guide RNAandCas9 protein. Stimulated human Tcells were incubated with ribonucleoprotein complexes for 15 min at room temperature, and adeno-associated viruses were added (multiplicity of infection = lx 10s) within 20 min after elec-troporation. After a 16-h incubation, cells were continually cultured, and medium was changed every 2 d. On day 7 after electroporation, cells were collected for flow cytometry analysis to determine the CAR knock-in efficiency, and supernatant was collected to determine human IL-10 concentration with a Cytometric Bead Array kit (BD Biosciences).
Flow cytometry analyses
For surface marker staining, cells were collected into U-bottom,96-well plates (Thermo Fisher Scientific), blocked with anti-mouse CD16/32 (BioLegend) and incubated with the indicated antibodies at 4 °C for 20 min, followed by live/dead staining with DAPI (Sigma-Aldrich). Cells were then washed and resuspended with PBS containing bovine serum albumin (BSA; 0.2% (wt/vol); Sigma-Aldrich) for flow cytometry analyses. For mitochondrial staining, cells were stained with MG (100 nfvl; Thermo Fisher Scientific) and MDR (10 nfvl; Thermo Fisher Scientific) for 20 min at 37 °C for mitochondrial mass and membrane potential assessments, respectively. Afterward, cells were further stained for surface markers and DAPI for flow cytometry analyses. For intracellular cytokine staining, cells were first stimulated with Cell Stimulation Cocktail (Invitrogen/Thermo Fisher Scientific) at 37 °C for 5 h. After stimulation, cells were stained for surface markers and with Zombie Aqua Fixable Dye (BioLegend) and fixed and permeabi-lized with a Cytofix/Cytoperm Fixation/Permeabilization Solution kit (BD Biosciences). Intracellular staining with the indicated antibodies was performed following the manufacturer's protocol. Cells were detected using an Attune NxT Flow Cytometer with Attune NxT Software v.3 (Invitrogen/Thermo Fisher Scientific) or an LSRFortessa with FACSDiva Software (BD Biosciences). Analyses were performed using Flowjo 10.8.1 (Tree Star). Gate margins were determined by isotype controls and fluorescence minus one controls.
Antibodies and reagents for flow cytometry
The following antibodies or staining reagents were purchased from BioLegend: CD16/32 (93,101302), CD45.2 (104,109814), CD8(3 (YTS256.7.7,126606), CD4 (RM4-5,100526), NK1.1 (PK136,108740), F4/80 (BM8,123108), CD3e (17A2,100306), CDllc (N418,117348), I A/I E (MHC-II,M5/114.15.2,107643),Siglec-F(S17007L, 155508), CD80 (16-10A1,104734), CD86 (GL-1,105006), CDllb (Ml/70,101228), Ki67 (16A8,652424), granzyme B (GB11,515403), IFNy (XMG1.2,505826), TNFa (MP6-XT22,506308), IL-2 (JES6-5H4,503822), Gr-1 (RB6-8C5, 202519), PD-1 (29F.1A12,135216), TIM-3 (RMT3-23,119706), CD107a (H4A3,328608), CD44 (IM7,103028), CD62L (MEL-14,104432), CD122 (5H4,105906), Seal (D7,108106), KLRG1 (2F1/KLRG1,138410), IL-7Ra (SB/199,121111), human CD95 (DX2,305606), Streptavidin-PE/Cya-nine7 (405206), human granzyme B (QA16A02,372208) and human IFNy (B2,552887). The following antibodies or staining reagents were purchased from BD Biosciences: human CD4 (SK3,563550) and human CD8 (SKI, 557834). The followingantibodies or staining reagents were purchased from Invitrogen: human CD62L (DREG56,48-0629-42), human CD45RA (H100, 69-0458-42), human CD27 (0323,56-0279-42) and CCR7 (3D12,61-1979-42). Biotinylated human HER2/ERBB2 protein (HE2-H822R) was purchased from ACROBiosystems. Myc tag antibody (9B11,3739) was obtained from Cell Signaling Technology. Goat anti-mouse IgG was purchased fromjackson ImmunoResearch. MitoTracker Green FM and MitoTracker Deep Red FM were obtained from Thermo Fisher Scientific.
Cytotoxicity assay of hCAR T cells
On day 7 after electroporation, CD19 hCAR T cells and target PANC1-CD19 (5 x io4) or Raji (5 x io4) cells were cocultured in complete RPMI-1640 medium at the indicated E:T ratios using black, 96-well flat plates (Thermo Fisher Scientific). Target cells alone were plated at the same cell density to determine the maximal luciferase expression (relative light units (RLU)). After 18 h of coculture, luciferin (1.5 mg ml"1, 100 ul per well; GoldBio) was added to each well. Luminescence was detected in a plate reader. Percent lysis was determined by using the following equation: (1 - RLUsamp|e/RLUmax) x 100%.
Antitumor therapy and rechallenge experiments
C57BL/6 or BALB/c WT mice bearing established tumors were sublethal^ lymphodepleted by total body irradiation (4 Gy) on day 5. On day 6, mice received i.v. adoptive transfer of CARTcells (HER2 CARTcells, TRP-1 CAR T cells or EGFRvlll CAR T cells, 3 x 10" CAR+ T cells) in the presence or absence of a single doseof i.v.-administered IL-10 (1 ng) on day 6, multiple doses of IL-10 (3.5 ng, equivalent to 20 ng of IL-10-Fc12) or multiple doses of IL-10-Fc (20 ng five times) through i.v. or peritu-moral administration on days 6,8,10,12 and 14; IL-10 CARTcells (IL-10 HER2 CARTcells, IL-10 TRP-1 CARTcellsor IL-10 EGFRvlll CARTcells, 3 x 106 CARTcells); untransduced Tcells (equivalent total number of cells infused) or PBS. For the MC38-HER2 or B16F10 tumor model in C57BL/6 mice, tumor area and body weight were measured every other day. For the s.c. PANC1-CD19 tumor model and the Raji tumor model, NSG mice received i.v. adoptive transfer of CD19 hCAR T cells (1 x 106 CAR+ T cells) or IL-10 CD19 hCART cells (1 x 106 CAR+ T cells) on day 8 after tumor inoculation. For the orthotopic PANC1-CD19-Luc tumor model, NSG mice received i.v. adoptive transfer of CD19 hCAR T cells (1 x 105 CAR+ T cells) or IL-10 CD19 hCAR T cells (1 x 105 CAR+ T cells) on day 4 after tumor inoculation. For the s.c. tumor model, tumor area was calculated by the formula area = length x width from caliper measurements of two orthogonal diameters. Forthe4Tl-EGFRvIII-Luc metastatic tumor model and orthotopic PDAC model, mice were anesthetized and intraperitoneally injected with bioluminescent substrate D-luciferin potassium salt (100 ul; GoldBio) prediluted at 30 mg ml"1 in lx PBS. Ten minutes after injection, mice were subjected to luminescent imaging using a Xenogen IVIS fluorescence/bioluminescence imaging system or Photon Imager (Biospace Lab) for tumor growth
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
monitoring. In tumor cell rechallenge experiments in the syngeneic model, MC38-HER2 (1 x 10"), B16F10 (1 x 105) and 4Tl-EGFRvlll-Luc (5 x io5) cells were reinoculated s.c. into the left flanks of mice that survived from treatment groups 3 months after adoptive CAR T cell transfer. Naive WT mice were s.c. inoculated with the same number of tumor cells as the control group. In the tumor cell rechallenge experiments in the xenograft model, PANC1-CD19-Luc (2 x io6) cells were orthotopically inoculated into the pancreas tail of mice that survived from treatment groups 46 d after adoptive CAR T cell transfer. Naive NSG mice were orthotopically inoculated with the same number of tumor cells as the control group. Survival of rechallenged mice was monitored for at least another 60 d. Mice were killed when body weight loss was beyond 15% of baseline weight, tumor area reached 150 mm2 (mouse syngeneic models) or the maximum diameter reached 20 mm (NSG mouse models) or if any signsof discomfort were detected by the investigators or as recommended by the veterinarian who monitored the mice every other day.
Analyses of tumor-infiltrating immune cells and serum cytokines
C57BL/6 mice were inoculated s.c. with MC38-HER2 tumor cells (1 x 106) and sublethally lymphodepleted by irradiation on day 5 and received i.v. adoptive transfer of HER2 CAR T cells (3 x 106 CAR+ T cells) in the presence or absence of i.v.-administered IL-10 (1 \ig), IL-10 HER2 CAR T cells (3 x io6 CAR+ T cells) or PBS control on day 6 after tumor inoculation. On day 14, mice were killed, and tumors and sera were collected. NSG mice were inoculated s.c. with PANC1-CD19 cells (5 x io6) and received i.v. adoptive cell transfer of CD19 hCART cells or IL-10 CD19 hCAR T cells (1 x io6). On day 27, mice were killed, and tumors were collected. Collected tumors were weighed, mechanically minced and digested in RPMI-1640 medium supplemented with collagenase type IV (1 mg ml"1; Gibco/Thermo Fisher Scientific), dispase II (100 \ig mL Sigma-Aldrich), hyaluronidase (100 \ig ml"1; Sigma-Aldrich) and DNase I (100 ug ml"1; Sigma-Aldrich) at 37 °C for 60 min. RBC lysis was performed on the digested tumor samples with ACK lysing buffer. TILs were then enriched by Percoll (GE Healthcare) density gradient centrifugation, resuspended in PBS with 0.2% BSA (wt/vol), stained with the indicated antibodies and analyzed by flow cytometry. Serum was collected from C57BL/6 mice and analyzed by enzyme-linked immunosorbent assay (ELISA) using the ELISA MAX Standard Set for mouse IL-1(3, IL-2, IL-6 and IFNy (BioLegend) to determine the serum concentrations of the indicated cytokines.
Analyses of CAR T cells from spleen, blood and bone marrow
For the MC38-HER2 model, C57BL/6 mice were inoculated s.c. with MC38-HER2 tumor cells (1 x io6) and sublethally lymphodepleted by irradiation on day 5 and received i.v. adoptive transfer of HER2 CAR T cells (3 x io6 CAR+ T cells) in the presence or absence of i.v.-administered IL-10 (1 ug), IL-10 HER2 CAR T cells (3 x 10" CAR+ T cells) or PBS control on day 6 after tumor inoculation. On day 18, spleens and blood were collected for phenotypic analyses of CAR T cells. Forthe4Tl-EGFRvIII-Luc model, BALB/c mice were inoculated i.v. with 4Tl-EGFRvlll-Luc tumor cells (5 x io4) and sublethally lymphodepleted by irradiation on day 5 and received i.v. adoptive transfer of EGFRvIII CARTcells (3 x 106 CAR+Tcells) in the presenceor absence of i.v.-administered IL-10 (1 ug), IL-10 EGFRvIII CARTcells (3 x 106 CAR+ T cells) or PBS control on day 6 after tumor inoculation. On days 6, 9,12,15 and 18, mice were bled to monitor CAR T cell counts in the blood. For the PANC1-CD19 model, NSG mice were inoculated s.c. with PANC1-CD19 cells (2 x 106) and received i.v. adoptivecell transfers of CD19 hCAR T cells (1 x 106 CAR+ T cells) or IL-10 CD19 hCAR T cells (1 x io6 CAR+ T cells) on day 8. On day 63, mice were killed for phenotypic analyses of CAR T cells in the spleen and bone marrow. For the PANC1-CD19-Luc model, NSG mice wereorthotopically implanted with PANC1-CD19-Luc cells (2 x 106) and received i.v. adoptive cell transfers
of CD19 hCAR T cells (5 x 10s CAR+ T cells) or IL-10 CD19 hCAR T cells (5 x 105 CAR+ T cells) on day 4. On days 18,28 and 50, blood samples were collected and subsequently analyzed by flow cytometry to assess the expansion of CARTcells in the blood. Spleens were dissected from the surrounding tissues, ground and filtered through a 70-um strainer (Fisher Scientific). RBC lysis was performed on the spleen samples with ACK lysing buffer (2 ml per spleen; Gibco/Thermo Fisher Scientific) and resuspended in PBS with BSA (0.2% (wt/vol)). Blood samples were resuspended in PBS with ethylenediaminetetraacetic acid (2 mM; Inv-itrogen). RBCs were lysed with ACK lysis buffer, debris was removed by Percoll (GE Healthcare), and cells were resuspended in PBS with 0.2% BSA (wt/vol). Bone marrow was collected from freshly isolated femurs and tibiae. Bones were crushed in 5 ml of PBS with ethylenediaminetetraacetic acid (2 mM; Sigma-Aldrich) and filtered with strainers. Remaining RBCs were lysed with ACK buffer. Collected cells from the samples described above were stained with the indicated antibodies and analyzed by flow cytometry.
Cell sorting
Splenic IL-10 HER2 CAR T cells, tumor-infiltrating HER2 CAR T cells or IL-10 HER2 CAR T cells from the MC38-HER2 tumor model were enriched by density gradient centrifugation against Percoll (GE Healthcare) and stained with surface markers and DAPI (Sigma-Aldrich), followed by sorting with an Aria II sorter (BD Biosciences) at the EPFL Flow Cytometry Core Facility. Splenic CD19 hCARTcells or IL-10 CD19 hCAR T cells from the PANC1-CD19 tumor model were enriched by using a human CD45 isolation MACS kit (Miltenyi Biotec) and stained with surface markers and DAPI (Sigma-Aldrich), followed by sorting with an Aria II sorter at the FACS facility at Zhejiang University.
Electron microscopy
Sorted HER2 CAR T cells or IL-10 HER2 CAR T cells were fixed for 1 h in 2.5% glutaraldehyde (vol/vol; Sigma-Aldrich) and 1% formaldehyde (vol/vol; Sigma-Aldrich) in sodium cacodylate buffer (0.12 M; Sigma-Aldrich). Cell sections were prepared by the EPFL Biological Electron Microscopy Facility. Ultrathin sections (50 to 60 nm) were obtained using a Leica EM UC7 ultramicrotome. Sections were examined with a Talos L120C TEM for Life Sciences (Thermo Fisher Scientific) at an accelerating voltage of 80 keV. Mitochondrial parameters were measured by using Fiji imaging software (ImageJ).
In vitro coculture of CAR T cells and tumor cells
MC38-HER2, B16F10 or 4Tl-EGFRvIII-Luc tumor cells (1 x 10s cells per well) were seeded in 12-well plates (Thermo Fisher Scientific) in complete DMEM and incubated overnight. After aspiration of tumor culture medium, HER2 CAR T cells, TRP-1 CAR T cells or EGFRvIII CAR T cells on day 6, as described above, were suspended in complete RPMI supplemented with IL-2 (10 ng ml"1) and added to the tumor cell culture at an E:T ratio of 0.5:1 in the presence or absence of IL-10 (145 ng ml"1). After another 2-d coculture, cells in the plates were collected, and flow cytometry analyses were conducted to determine the tumor cell viability and the proliferation and phenotypes of CAR T cells. IL-10 concentration in theculture supernatants was determined by an ELISA MAX Standard Set Mouse IL-10 kit (BioLegend) after 3 d of coculture.
In vitro restimulation of hCAR T cells with tumor cells
Seven days after electroporation, CAR T cells were cocultured with irradiated NIH/3T3-CD19 or PANC1-CD19 cells for weekly restimulation. Typically, NIH/3T3-CD19 or PANC1-CD19 cells (1.25 x 10s) were plated on 24-well tissue culture plates 12 h before the addition of CAR T cells (5 x io5) in X-VIVO15 serum-free medium (Lonza) supplemented with 10% (vol/vol) FBS,1% (vol/vol) penicillin/streptomycin,5 ng ml"1 human IL-7 and 5 ng ml"1 human IL-15. Every week, the total number of cells were counted using Countstar Rigel S2 (Alit Biotech). CAR expression was detected by flow cytometry.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Seahorse assay
MC38-HER2 tumor cells (1 x io6 cells per flask) were seeded in T25 flasks (Thermo Fisher Scientific) in complete DMEM and incubated overnight. After aspiration of tumor cell culture medium, HER2 CAR T cells (WT, OT-I or Mpc-KO OT-I) cells or IL-10 HER2 CAR T cells (WT, OT-I or Mpc-KO OT-I) in complete RPMI medium supplemented with IL-2 (10 ng ml"1) were added to the tumor cell culture at an E:T ratio of 5:1. In some experiments, the medium was supplemented with IL-10 (145 ng ml"1) or various inhibitors as indicated. After 16 h of coculture, CARTcellsintheflaskswerecollectedand isolated byFicoll-PaquePLUS for the Seahorse assay. Seahorse assays were performed to measure OCR and ECAR of CAR T cells. CAR T cells (3 x 105 cells per well) with different treatment conditions were seeded in a Seahorse culture plate (Seahorse Bioscience) in a non-C02 incubator at 37 °C for 40 min. OCR and ECAR were measured with a Seahorse XFe96 Analyzer (Seahorse Bioscience) following the manufacturer's instructions. During a Seahorse assay, cells were treated with oligomycin (1 uM; Sigma-Aldrich), carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (2 \lM; Sigma-Aldrich), rotenone (0.5 uM; Sigma-Aldrich),antimycin A (0.5 uM; Sigma-Aldrich),glucose (10 mM;Sigma-Aldrich) and 2-deoxy-D-glucose (50 mM; Sigma-Aldrich). Each condition was performed with three to five replicates in a singleexperiment. Basal or maximal OCR and ECAR were calculated according to previous reports44.
Metabolic inhibitor treatments
MC38-HER2 cells (1 x 10s cells per well) were seeded on 12-well plates (Thermo Fisher Scientific) and incubated overnight. HER2 CARTcells or IL-10 HER2 CAR T cells on day 6, as described above, were cocultured with preseeded MC38-HER2 cells at an E:T ratio of 0.5:1 in complete RPMI medium containing IL-2 (10 ng ml"1) and the indicated inhibitors (oligomysin, 1 uM; UK5099 (Sigma-Aldrich), 100 uM; etomoxir (Sigma-Aldrich), 200 uM) for 2 d. Absolute counts and phenotypes of CAR Tcells were determined by flow cytometry analyses.
Metabolomic analyses
Mice bearing MC38-HER2 tumors were sublethally lymphodepleted by total body irradiation (4 Gy) on day 5, followed by i.v. adoptive transfer of HER2 CAR T cells or IL-10 HER2 CAR T cells on day 6. On day 14, mice were killed, and CAR T cells were isolated from tumors as described above. Metabolomic analyses were performed by the University of Lausanne Metabolomics Unit. Cell lysates were preex-tracted and homogenized by the addition of 200 \i\ of methanol in a Cryolys Precellys 24 sample homogenizer (2 x 20 s at 10,000 r.p.m.; Bertin Technologies) with ceramic beads. Homogenized extracts were centrifuged for 15 min at 4,000gat 4 °C (Hermle). The resulting supernatant was collected and evaporated to dryness in a vacuum concentrator (LabConco). Dried sample extracts were resuspended in methanokwater (4:1 (vol/vol)) according to the total protein content. The protein pellets were evaporated and lysed in Tris-HCI buffer (20 mM, pH 7.5) with guanidine hydrochloride (4 M), NaCI (150 mM), Na2EDTA (1 mM), EGTA (1 mM), Triton (1%), sodium pyrophosphate (2.5 mM), [^glycerophosphate (1 mM), Na3V04 (1 mM) and leupeptin (1 ng ml"1) using a Cryolys Precellys 24 sample homogenizer (2 x 20 s at 10,000 r.p.m.; Bertin Technologies) with ceramic beads. A BCA Protein Assay kit (Thermo Scientific) was used to measure total protein concentration (adsorption at 562 nm; Hidex). Extracted samples were analyzed by hydrophilic interaction liquid chromatography coupled to tandem mass spectrometry in both positive and negative ionization modes using a 6495 triple quadrupole system interfaced with a 1290 UHPLC system (Agilent Technologies). Pooled quality control (QC) samples were analyzed periodically throughout the overall analytical run toassess the quality of the data,correct the signal intensity driftand remove peaks with poor reproducibility. In addition, a series of diluted QC samples was prepared by dilution with methanol: 100% QC,50% QC, 25% QC, 12.5% QC and 6.25% QC. Metabolites were then selected also
considering the linear response on the diluted QC series. Raw liquid chromatography-tandem mass spectrometry data were processed using Agilent Quantitative analysis software (version B.07.00; Mass-Hunter Agilent Technologies). Relative quantification of metabolites was based on extracted ion chromatogram areas for the monitored multiple reaction monitoring transitions. The obtained tables (containing peak areas of detected metabolites across all samples) were exported to R software (http://cran.r-project.org/). Signal intensity drift correction and noise filtering were done within MRM PROBS software. The preprocessed data with peak areas were imported into Metaboanalyst 5.0 for further data analysis.
scRNA-seq and bioinformatics analyses
MC38-HER2-bearing mice were sublethally lymphodepleted by total body irradiation (4 Gy) on day 5, followed by i.v. adoptive transfer of HER2 CAR T cells or IL-10 HER2 CAR T cells on day 6. On day 18, mice were killed, and CAR T cells were isolated from tumors or spleens as described above. NSG mice were inoculated (s.c.) with PANC1-CD19 cells (2 x io6) and received i.v. adoptive cell transfer of CD19 hCAR T cells or IL-10 CD19 hCAR T cells (1 x 106) on day 8. On day 39, mice were killed, and CARTcells were isolated from spleens as described above. Sorted HER2 CAR T cells from tumors were then pelleted for 5 min at 400gand washed once in PBS with BSA (0.04% (wt/vol)). Afterward, HER2 CARTcells or IL-10 HER2 CARTcells were suspended in 100 ul of Cell Multiplexing Oligo (LMO, lOx Genomics). LMO barcode labeling was performed for 10 min at room temperature,and cells were washed twice with PBS, resuspended in PBS with BSA(0.04% (wt/vol)), filtered and pooled. Labeled cells were subjected to single-cell encapsulation using a Chromium Single Cell Instrument and reagents. Sorted CD19 hCAR T cells or IL-10 CD19 hCAR T cells from spleens were subjected to single-cell encapsulation using a Chromium Single Cell Instrument and reagents separately. A Chromium Next GEM Chip G was loaded with the appropriate number of cells, and the sequencing were libraries prepared using lOx Genomics reagents accordingto the manufacturer's instructions and passed QC. Briefly, an emulsion encapsulating single cells into droplets with reagents and gel beads containing a unique molecular identifier (UMI), reverse transcription reagents and cell barcoding oligonucleotides was generated. cDNAs were obtained and amplified after droplets broke. For the 5' Gene Expression library, the cDNA was fragmented, ligated to a sequencingadaptor and PCRampli-fied. The generated 5'Gene Expression libraries were sequenced using an Illumina HiSeq 4000 with a sequencing depth of115,000 paired-end reads per cell for HER2 CAR T cells from tumors and a sequencing depth of46,000 paired-end reads per cell for CD19hCARTcells from spleens. The fastq files were generated and demultiplexed by cellrangermkfastq from lOx Genomics (version 3.0.2), and primary data analysis was performed with Cell Ranger (version 2.2.0) using a custom reference package based on the mmlO or GRCh38 reference genome and GEN-CODE gene models. For IL-10 HER2 CAR T cells from the spleen, cells were subjected to single-cell partitioning usinga microchip, and mRNA capture and sequencing libraries were prepared using GEXSCOPE Single Cell RNA Library Kit Cell V2 reagents (Singleron Biotechnologies) according to the manufacturer's instructions and passed QC. Briefly, single partitioning was performed by loading singlecells into a microchip and loading barcode beads containing a UMI. Cells were lysed, and mRNA capturing beads were retrieved. cDNAs were obtained via reverse transcription, amplified and purified. For the 5' Gene Expression libra ry, the cDNA was fragmented, ligated toa sequencingadaptor and PCR amplified. The generated 5' Gene Expression libraries were sequenced using an Illumina HiSeq 4000 with a sequencing depth of 50,000 paired-end reads per cell. The fastq files were generated and demultiplexed by CeleScope rna from Singleron (version 3.0.1), and primary data analysis was performed with CeleScope (version 1.10.0) using a custom reference package based on reference genome Mus musculus ensembl 92.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
To demultiplex samples of HER2 CARTcellsfrom tumors based on LMO barcodes, we applied the HTODemux function implemented in Seurat v4.0.5 (ref. 45) using default settings. QC was applied to remove outlier cells with less than 500 or over 50,000 UMI counts, less than 500 or over 6,000 genes and abnormal ribosomal or mitochondrial gene content (over 40% and 7.5%, respectively). To remove potential contaminants or cell doublets, we applied the scGate package vl.0.0 (ref. 46) using a default T cell filter. Dimensionality reduction was performed by calculating ten principal components from 500 variable genes, further reducing dimensionality by UMAP for two-dimensional visualization. Unsupervised clustering was performed on the principal component reduction using the FindNeighbors and FindClusters functions implemented in Seurat with parameters k.param = 30 and resolution = 0.35. Differentially expressed genes between clusters were determined using the FindMarkers function of Seurat, which applies a Wilcoxon test to determine significance. Differentially expressed genes were visualized using the EnhancedVolcano R package (https://github. com/kevinblighe/EnhancedVolcano) with a log2 (fold change) cutoff of 0.5 and P value of 10s. Gene sets for relevant biological processes and pathways were obtained from the database mSigDB47, and signature scores for these gene sets were calculated using the UCell package22. GSEA between clusters was performed on average gene expression by cluster usingtheRpackageclusterProfiler^and relevant signatures from mSigDB48. Projection of HER2 CARTcells or IL-10 HER2 CARTcells into a reference atlas of TILs was performed using the ProjecTILs method with default parameters.
Statistical analysis
Statistical analysis was performed using GraphPad Prism 9 (GraphPad software) except scRNA-seq data, which were analyzed with R (described above). Metabolomics data (Extended Data Fig. 5h) were imported into MetaboAnalyst 5.0 for further analysis. Data are presented as mean ± s.e.m. unless otherwise indicated. Comparisons of two groups were performed by using two-tailed unpaired Student's t-tests unless otherwise indicated. Comparisons of multiple groups were performed by using a one- or two-way ANOVA with a Tukey's multiple-comparisons test. Survival data were analyzed using log-rank tests. No statistically significant differences were considered when P values were larger than 0.05.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All data generated and supporting the findings of this study are available within the paper. scRNA-seq data for the mouse CART cells from this study have been deposited in the Gene Expression Omnibus under accession number GSE245517. scRNA-seq data forthe hCARTcells are available in the CNGB Sequence Archive of the China National Gen-eBank Database under accession number CNP0003547. Source data are provided with this paper. Additional information and materials will be made available upon reasonable request.
Code availability
The Rcode required to reproduce the scRNA-seq data shown in Figs. 4 and 6 and Extended Data Figs. 5 and 10 is available at the following repository: https://github.com/carmonalab/LiTang IL10 CART.
References
41. Tschumi, B.O.etal. CART cells are prone to Fas-and DR5-mediated cell death. J. Immunother. Cancer 6,71 (2018).
42. Brentjens, R. J. et a I. Eradication of systemic B-cell tumors by genetically targeted human T lymphocytes co-stimulated by CD80 and interleukin-15. Nat. Med. 9,279-286 (2003).
43. Jing, R. et al. Cas9-cleavage sequences in size-reduced plasmids enhance nonviral genome targeting of CARs in primary human Tcells. Small Methods5,2100071 (2021).
44. Michelet, X. et al. Metabolic reprogramming of natural killer cells in obesity limits antitumor responses. Nat. Immunol. 19, 1330-1340 (2018).
45. Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184,3573-3587.e29 (2021).
46. Andreatta, M., Berenstein, A. J. & Carmona, S. J. scGate: marker-based purification of cell types from heterogeneous single-cell RNA-seq datasets. Bioinformatics 38,2642-2644 (2022).
47. Liberzon, A. et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics 27,1739-1740 (2011).
48. Yu, G., Wang, L.-G, Han, Y. & He, Q.-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS16,284-287 (2012).
Acknowledgements
We thank D. J. Irvine (Massachusetts Institute of Technology) for providing TRP-1 CAR and EGFRvIII CAR plasmids and 4T1-EGFRvlll-Luc cells, J. Auwerx (EPFL) for providing access to a Seahorse XFe96 Analyzer, and J. Shi and Y. Wang (Zhejiang University) for technical support for the orthotopic PDAC model. We acknowledge the EPFL CPG, Flow Cytometry Core Facility, Gene Expression Core Facility and Biological Electron Microscopy Facility and the University of Lausanne Center of Metabolomics Platform for technical assistance. We thank Y. Li, Y. Huang and C. Guo from the Core Facilities, Zhejiang University School of Medicine, for their technical support. This work was supported, in part, by the Swiss National Science Foundation (315230J73243, 316030J89686, 315230_204202, IZLCZ0_206035 and CRSII5_205930), Swiss Cancer Research Foundation (KFS-4600-08-2018), Kristian Gerhard Jebsen Foundation, the European Research Council under the ERC grant agreement MechanoIMM (805337), Anna Fuller Fund grant, Xtalpi, Inc., and EPFL. J.S. was funded by the National Natural Science Foundation of China grants 82161138028 and 31971324, Zhejiang Provincial Natural Science Foundation grant LR20H160003 and National Key R&D Program of China grant 2021YFA0909900. S.C. was supported by the Swiss National Science Foundation Ambizione grant 180010. P.R. was supported, in part, by grants from the Swiss National Science Foundation (310030J82735) and the Swiss Cancer League (KFS-4404-02-18). M.G. was supported by the Chinese Scholarship Council (201808320453). X.H. has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sktodowska-Curie grant agreement 754354 and was supported by the China Scholarship Council (201700260266).
Author contributions
Y.Z., J.C., J.S., Y.G. and LT. conceived the study. Y.Z., J.C., J.S., Y.G and L.T. designed the experiments. Y.Z., J.C., M.A., B.F., Y.-Q.X., M.W., Y.W., M.G., X.H., P.R. and S.C. performed the experiments. Y.Z., J.C., J.S., Y.G. and L.T. analyzed the data and wrote the paper. All authors edited the paper.
Competing interests
Y.G., L.T. and Y.Z. are inventors on patents related to the technology described in this paper. L.T. and Y.G. are cofounders, share-holders and advisors for Leman Biotech. The interests of L.T. were reviewed and managed by EPFL. The other authors declare no competing interests.
Additional information
Extended data is available for this paper at https://doi.org/10.1038/s41587-023-02060-8.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41587-023-02060-8.
Correspondence and requests for materials should be addressed to Jie Sun, Yugang Guoor Li Tang.
Peer review information Nature B/otechnoiogythanksthe anonymous reviewers for their contribution to the peer review of this work.
Reprints and permissions information is available at www.nature.com/reprints.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Inoculation of MC38-HER2 tumors (s.c.) in C57BL/6
0 6
t
HER2 CAR-T (i.v.)
day
- PBS
- Ctrl T
- HER2 CAR-T
PBS Ctrl T
HER2 CAR-T
' 100 80 60 40 20 0
		0/5
L_		
		1
Time post innoculation (day)
3 10 20 30 40 Time post inoculation (day)
z 1 801
<D >•£ 60-u) z rr ro u_ <
40-
^ c/"c/
6*10 4*10 2x104-0
CtrIT
HER2 CAR-T IL-10 HER2 CAR-T
Extended Data Fig. 11 In vitro characterizations of IL-10 CAR-T cells prior to infusion, a-e, C57BL/6 mice were inoculated subcutaneously (s.c.) with MC38-HER2 colon cancer cells (3 x 10s) and received intravenously (i.v.) adoptive cell transfer of HER2 CAR-T cells (3 x lo6), untransduced T cells (Ctrl T, equivalent cell number infused), or phosphate buffered saline (PBS) on day 6. Mice were either monitored for tumor growth (n = 5 mice) or sacrificed on day 14 for the analysis of CAR-T cells (n = 10 mice), a, Experimental timeline, b-e, Shown are individual tumor growth curves (b), survival curves (c), frequencies of granzyme B*IFNY*TNFa* polyfunctional CAR-T cells in spleen and tumor (d), and MFI of PD-1 and TIM-3 (e). Indicated are numbers of tumor-free mice/total mice in each group (c). f, HER2 CAR or IL-10 HER2 CAR constructs were introduced via retroviral vectors. CAR expression levels were assessed by flow cytometry. Data are one representative often independent experiments, g, HER2 CAR-T or IL-10
1.5*103n		
	P= 0.8941	
1x103-	Ogo AAg	
5x102-		
0-		
HER2 CAR-T cells were co-cultured with MC38-HER2 tumor cells for 3 days (n = 3 biologically independent samples)). The culture supernatants were assessed for IL-10 concentration using an enzyme-linked immunosorbent assay (ELISA). h-q, CAR-T cells prior to infusion were examined for the expression levels of indicated markers by flow cytometry (n = 9 biologically independent samples for IL-7Ra, and n = 10 biologically independent samples for others). Shown are average MFI of CD44 (h), CD62L (i), Sca-1 (J), CD122 (k), IL-7Rct (1), KLRG-1 (m), PD-1 (n), TIM-3 (o), and CD69 (p) expression in HER2 CAR-T and IL-10 HER2 CAR-T cells, q, Average frequencies of viable CAR-T cells. All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (d, e, g-q), or one- or two-way ANOVA with Tukey's multiple-comparisons test (b) or log-rank test (c). ns, not significant (P > 0.05). Data are one representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
TCF-1*PD-1*TIM-3" CD8* CAR-T
J 40-|> 30 2 20 10-0-
o
TCF-1*PD-1*TIM-3" CD8*CAR-T
o 40 z
j- 30
o E
TCF-rPD-1*TIM-3* CD8* CAR-T P= 0.0103
			O	50n
	» & a		Li_	
		+		40-
O	a	Z li-	CG	30-
1		'S	:yme	20-
			ra	10-
_,_			O	n-
a a
Nature killer cells P= 0.0961
Dendritic cells P= 0.3095
Eosinophils
s 5x102
§ 4x102
|> 3x102
2 2x102 c
O Ü
	P= 0.0720
1 p	P= 0.2053 = 0.9523
	
	□ a
Tumor-associated macrophages. P= 0.3360
P= 0.6730
J 1.5*103 E 1*103
(A
§ 5»102
o
O
0
„ o --š-- nfe hps   g°88 n?B
■fr
Gr-1*CD11b* Myeloid-Derived Suppressor Cells
I 2:
tn 1; c
3 5
:103n
103-103-103-102-
0 —
c/- „* cf
oO- -5
Inoculation of MC38-HER2 tumors (s.c.) in C57BL/6
6 t
Sacrifice & analyses
I
14 day
HER2 CAR-T/IL-10 HER2 CAR-T (i.V.)
s 2x103-11.5x103-
D)
I 1x103-
1 5x102-0-
Endogenous T cells
P=0.1518 P= 0.8147 P=0.37
8*5 ÜT
150-100-50 0
Endogenous Tregs
P= 0.4257 P= 0.9486 P= 0.6012
□ □
f
HER2 CAR-T IL-10 HER2 CAR-T
HER2 CAR-T IL-10 HER2 CAR-T
0       5x104    1x105  1.5«106 2x105 CAR density
Extended Data Fig. 21 See next page for caption.
5x10"   1x105  1.5x10= 2x CAR density
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Extended Data Fig. 2 IL-10 expression shows negligible effects on other immune cells or endogenous T cells in tumor, a-c, The experimental setting was the same as described in Fig. lb (n = 5 mice), a, Counts of TCF-1*PD-rTIM3XD8* HER2 CAR-T cells in tumors. b,c, Frequencies of IFNY*Granzyme B*TNFa* polyfunctional cells among PD-l*TCF-l*TIlv13~ (b) and PDT*TCFT~TIM3* (c) CD8* CAR-T cells, d, The experimental setting was the same as described in Fig. lb (n = 10 mice). Counts of indicated immune cell subsets in the MC38-HER2 tumors from each treatment group, e-g, C57BL/6 mice were inoculated s.c. with MC38-HER2 colon cancer cells (1 x 106) and received i.v. adoptive transfer
of IL-10 HER2 C ART cells (3 x 106) or HER2 C ART cells (3 x 106) on day 6 (n = 5 mice), e, Experimental timeline, f, Counts of endogenous T cells, g, Counts of Foxp3*CD25*CD4* endogenous Tregs. h,i, The experimental setting is described in Fig. lb. CAR-T cells in tumors were classified into several subpopulations based on the gating of CAR density (Fig. lj). Shown are response curves of MFI of TNFa (h) and TIM-3 (i) as a function of CAR density (n = 4 mice). Indicated are Pvalues. All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (a-c, h, i), or one-way ANOVA with Tukey's multiple-comparisons test (d, f, g). Data are one representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
CD19 hCAR construct
P2A
IL-10 CD19 hCAR construct
P2A
CD28   CD3(   PA     TRAC RHA
T2A
CtrIT
CD19 hCAR-T IL-10 CD19hCAR-T
1.5*10'
1x10'
5*10:
0        7       14 21 Days post 3T3-CD19 stimulation
0.0001
AAA
AAA AaA AAA
0 J-£gp-,-
2:1 1:1 1:2 1:4 1 E:T ratio
1:1 1:2 1:4 1:81:16 E:T ratio
0        7       14 21 Days post PANC1-CD19 stimulation
Extended Data Fig. 31 Preparation and in vitro characterizations of IL-10-secreting CD19 hCAR-T cells, a, Schematic depicting constructs of CD19-targeted second-generation CD28-based human CAR (CD19 hCAR) and human IL-10-secreting CD19 hCAR (IL-10 CD19 hCAR). b, The expression levels of CD19 hCAR were analyzed by flow cytometry. Ctrl T, untransduced control human T cells, c, The culture supernatants were examined for the concentration of hIL-10 on day 8 after hCAR-T cell preparation (n = 12 biologically independent samples), d, e, CD19 hCAR-T and IL-10 CD19 hCAR-T cells were cocultured with PANC1-CD19
cells or Raji cells at different E:T ratios for 48 h (n = 3 biologically independent samples). Shown are the percentage of lysis of PANC1-CD19 cell (d) or Raji cell (e). f,g, The proliferation of CD19 hCAR and IL-10 CD19 hCAR-T cells after multiple weekly stimulation (indicated byarrows; n = 3) with irradiated NIH/3T3-CD19 (f) or PANC1-CD19 (g) cells (n = 3 biologically independent samples). All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test. Data are one representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Extended Data Fig. 41 IL-10 expression improved mitochondrial fitness and increased the levels of pyruvate production intermediates and TCA cycle intermediates in CAR-T cells. The experimental setting is described in Fig. 2a. a, MFI of PD-1, TIM-3, and CD107a in MDR/MGhi and MDR/MG1" CAR-T cells in tumor (n = 10 mice), b, Quantification of roundness index of mitochondria in CAR-T cells as shown in Fig. 2i (n = 15 biologically independent samples), c, Representative contour plot showing CAR-T cells in tumor were classified into three subpopulations based on CAR density, d-f, The experimental setting is described in Fig. 3a-c (n = 3 biologically independent samples), d, Real-time analysis of EC AR. e, Average basal and maximal EC AR. f, Average basal and maximal OCR of HER2 CAR-T and IL-10 HER2 CAR-T cells in the presence or absence of IL-10 without antigen stimulation, g-j, C57BL/6 mice were inoculated s.c. with MC38-HER2 colon cancer cells (1 x 106), lymphodepleted, and received IL-10 HER2 CAR-T cells or HER2 CAR-T cells (3 x 106, i.v.) on day 6. On day 14,
tumor-infiltrating CAR-T cells were sorted for metabolomics analysis (n = 3 biologically independent samples, each sample is pooled from n = 5 mice), g, Experimental timeline, h, Clustered heatmap of significantly altered metabolites (P< 0.05) in HER2 CAR-T and IL-10 HER2 CAR-T cells. The color blocks indicate relative metabolite levels, with high expression in red and low expression in green, i j, Fold changes of phosphoenolpyruvate (i) and succinate (j) (relative to HER2 CAR-T cells). k,l, HER2 CAR-T or IL-10 HER2 CAR-T cells were cocultured with MC38-HER2 cells in the presence of indicated inhibitors (n = 3 biologically independent samples). Shown are relative CAR-T cell counts (k) and relative frequencies of granzyme B*IFNy*TNFa* polyfunctional CAR-T cells (I) (IL-10 HER2 CAR-T vs. HER2 CAR-T). Data are one representative of two independent experiments. All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (a, b, i, j) or by one- or two-way ANOVA with Tukey's multiple-comparisons test (e, f, k, I).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Inoculation of MC38-HER2 Sorted CAR-T cells tumors (s.c.) in C57BL/6      for scRNA-seq
0
6 18 day
Lymphodepletion *s4 A       HER2 CAR-T/
1   IL-10 HER2 CAR-T (i.v
(i.V.)
Top2a Birc5 Mki67 Mcm3 Mcm7
Normalized expression
2 1
< 0
HER2 CAR-T
0
UMAP 1
< 0
IL-10 HER2 CAR-T
0 4 UMAP 1
m HER2 CAR-T
■ IL-10 HER2CAR-T
Cell state
CD8+ Terminally-exhausted T CD8+ Precursor-exhausted T CD8+ Effector-memory T CD8+ Early-activation T
Qj CD8* Naive-like T CD4* Naive-like T
| CD4* Follicular-helper T CD4*Type 1 T helper (Th1)
| CD4* T cells Regulatory T (Treg)
Cytotoxicity Cell Cycling HALLMARK
(Prf1, Gzma, Gzmb, Gzmc, Gzmf)       (Tirosh et al. Science 2016) OXIDATIVE_PHOSPHORYLATION
I
	
	a.
0.75	
0.50	
0.25	
0.00	
	cn
	0_
0.75	MA
0.50	
0.25	
0.00	
I
	
	
0.25	<
0.20	UM
0.15	
0.10	
0.05	
	cn
	n
	<
0.2	s
	
0.1	
0.0	
UMAP_1 UMAP_1
Extended Data Fig. 51 IL-10-expressing HER2 CAR-T cells exhibit an expanded subpopulation of CD8* terminally exhausted T cells with upregulated gene expression encoding cell cycling, cytotoxicity, and OXPHOS. a, The
experimental setting and timeline is described in Fig. 4. b, Heatmap representing the expression levels of representative marker genes in each cluster, c, Projection of IL-10 HER2 CAR-T cells and HER2 CAR-T cells into a reference atlas of tumor-infiltrating lymphocytes. Reference atlas states are indicated
I
0.35 0.30 0.25 0.20 0.15
as colored cells. CD8* terminally-exhausted T cell cluster was characterized as high expression of granzymes, multiple inhibitory receptors (Pdcdl, Ctla4, Lag3, Tigit, Haucr2/TM-3, etc.) and Tox16. Shown are cell subtype compositions for each group. Black contour lines represent the density of projected cells, d, Gene signature scores for each group of samples (HER2 CAR-T and IL-10 HER2 CAR-T cells) for the indicated gene sets.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
PBS
HER2 CAR-T
HER2 CAR-T + IL-10 i.V. (1 ug) 200n
0 10 20 30 40 Time post innoculation (day)
HER2 CAR-T + IL-10 i.v. (3.5 (jg*5)
200
0 10 20 30 40 Time post innoculation (day)
IL-10 HER2 CAR-T
0 10 20 30 40 Time post innoculation (day)
	200n
E	150-
m	
ra	100-
mor	50-
zi	
	0-
0 10 20 30 40 Time post innoculation (day)
Inoculation of MC38-HER2 (s.c.) in C57BL/6
I
IL-10-Fc (i.v.) orlL-IO-Fc(p.t)
II   I   I I
0
Lymphodepletion
5
6
t
10  12   14 day
0 10 20 30 40 Time post innoculation (day)
HER2 CAR-T (i.V.)
HER2 CAR-T + IL-10-Fc i.V. (20 ugx5)
HER2 CAR-T+ IL-10-Fc p.t. (20 ugx5)
HER2 CAR-T + IL-10-Fc i.v. (20 ugx5)
HER2 CAR-T + IL-10-Fc p.t. (20 ugx5)
	160n
E	
E	120-
03 cd	sn
CD	ou
mor	40-
□	
	0-
0 10 20 30 40 Time post inoculation (day)
-100 80 60 40 20 0
2/5
HER2 CAR-T + IL-10-Fc i.V. (20 ugx5)
HER2 CAR-T + IL-10-fc p.t. (20 ugx5)
0 20 40 60 80 Time post inoculation (day)
0 10 20 30 40 Time post innoculation (day)
Time post innoculation (day)
	
	150-,
weigl	120-
>. -o	90:
_Q	801
>	40-0-
tr	i
10 15 20 25 Time post innoculation (day)
PBS
HER2 CAR-T
IL-10 HER2 CAR-T
HER2 CAR-T + IL-10 i.v. (1 ug)
HER2 CAR T + IL-10 i.v. (3.5 ugx5)
HER2 CAR-T + IL-10-Fc i.V. (20 ug*5)
HER2 CAR-T + IL-10-Fc p.t. (20 ugx5)
= 0.6344
I 20
0 j-^ťp-h^pwapiHnrjD—
Ay j>. <xl
Extended Data Fig. 61 In vivo assessment of therapeutic efficacy and toxicity of IL-10 HER2 CAR-T cells in the MC38-HER2 tumor model, a, Experimental setting is described in Fig. 5a. Shown are individual tumor growth curves. Indicated are numbers of tumor-free mice among the total number of mice in the group (a), b-e, The experimental setting was the same as described in Fig. 5a except another two treatment groups were added: HER2 CAR-T cells (3 x 106) in combination with IL-10 - Fcu (20 iig x 5, equivalent to native IL-10 at the dose of 3.5 \ig x 5) through i.v. or peritumoral (p.t.) administration (n = 5 mice). Shown are experimental timeline (b), average tumor growth curves (c), survival curves
(d), and individual tumor growth curves (e). Indicated are numbers of tumor-free mice among the total number of mice in the group (d, e). f, Body weight of treated mice in both Fig. 5b and Extended Data Fig. 6c (n = 10 mice for the PBS, HER2 CAR-T, and IL-10 HER2 CAR-T groups, and n = 5 mice for the rest of the groups), g, The experimental setting was the same as described in Fig. lb (n = 5 mice). Serum was collected and examined for the concentration of IL-lß, IL-6, IL-2, and IFNy by ELISA. All data represent the mean ± s.e.m. and are analyzed by one-way ANOVA with Tukey's multiple-comparisons test.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
TRP-1 CAR construct
c-Myc   S  CD8TM     CD28   CD3? 5 3'LTR
0 10 20 30 40 Time post innoculation (day)
0 10 20 30 40 Time post innoculation (day)
0 10 20 30 40 Time post innoculation (day)
0 10 20 30 40 Time post innoculation (day)
Inoculation of B16F10 tumors (i.d.) in C57BL/6 mice
PBS -a- TRP-1 CAR-T * IL-10 TRP-1 CAR-T
- PBS
— TRP-1 CAR-T IL-10 TRP-1 CAR-
T]p< 0.0001
Lymphodepletion
5
6
t
day
TRP-1 CAR-T/ IL-10 TRP-1 CAR-T (i.v.)
	150n
E	
E	
"tc	100-
o	
b	50-
E	
	
1-	0:
	0
		
-		3/7
----		
B16F10 (orthotopic)
5 10 15 20 Time post inoculation (day)
0 20 40 60 80 Time post inoculation (day)
TRP-1 CAR-T
IL-10 TRP-1 CAR-T
0 10        20        30 0      10     20     30 40
Time post innoculation (day) Time post innoculation (day)
Extended Data Fig. 71 Preparation and efficacy assessment of TRP-1 CAR-T cells in the subcutaneous and orthotopic B16F10 tumor models, a, Schematic depicting constructs of TRP-l-directed second-generation CD28-based CAR (TRP-1 CAR) and murine IL-10-secreting TRP-1 CAR (IL-10 TRP-1 CAR), b, The expression levels of TRP-1 CAR were examined by staining the c-Myc tag. Ctrl T, untransduced T cells as control, c, IL-10 TRP-1 CAR-T or TRP-1 CAR-T cells were co-cultured with B16F10 cells for 3 days (n = 3 biologically independent samples). The culture supernatants were examined for the concentration of IL-10 by ELISA. d, IL-10 TRP-1 CAR-T or TRP-1 CAR-T cells in the absence or presence of IL-10 (145 ng/ml) were cocultured with B16F10 cells at the E:T ratio of 0.5:1 for 48 h (n = 3 biologically independent samples). The percentage of tumor cell lysis was analyzed by flow cytometry. e,f, Experimental setting is described in Fig. 5d
3/7 cured
0 10 20 30 40 50 Time post innoculation (day)
(n = 5 mice). Shown are individual tumor growth curves (e) and relative body weight (f). Indicated are numbers of tumor-free mice among the total number of mice in the group (e). g-j, C57BL/6 mice were inoculated intradermals (i.d.) with B16F10 cells (2 x 10s) to establish an orthotopic melanoma model. Mice were then lymphodepleted, and received i.v. adoptive transfer of IL-10 TRP-1 CAR-T cells (3 x lo6), or TRP-1 CAR-T cells (3 x lo6) on day 6 (n = 7 mice), g, Experimental timeline, h-j, Shown are average tumor growth curves (h), mouse survival curves (i), and individual tumor growth curves (j) of the orthotopic B16F10 melanoma model. Indicated are numbers of tumor-free mice among the total number of mice in the group (i, j). All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (c), or by one- or two-way ANOVA with Tukey's multiple-comparisons test (d, h), or by log-rank test (i).
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
EGFRvIII CAR construct
EGFRvlll-scFv
IL-10 EGFRvIII CAR construct
3~I
EGFRvlll-scFv
CD8TM     CD28    CD3? J 3'LTR
CtrIT
EGFRvIII CAR-T IL-10 EGFRvIII CAR-T
c-Myc (PE)
♦ PBS
* EGFRvIII CAR-T
o EGFRvIII CART+IL-10 (1 ug) -a- IL-10 EGFRvIII CAR-T
•g 300
0
1 200 DC
o "o
0
\ \
P< 0.0001 I   P = 0.0158 P= 0.0021
I
EGFRvIII CAR-T
EGFRvIII CAR-T + IL-10 (1 ugh
IL-10 EGFRvIII CAR-T
(A		
		4x104-i
o		
R-T	lood	3*104-
<		
O	3.	2x104-
O	o	
(A	o	1x104-
		
Co		0-
6 9 12 15 18 21 24 Time post innoculation (day)
Extended Data Fig. 81 Preparation and efficacy assessment of EGFRvIII CAR-T cells in the4Tl-EGFRvIII-Luc tumor model, a, Schematic depicting constructs of EGFRvIII-directed second-generation CD28-based CAR (EGFRvIII CAR), and murine IL-10-secreting EGFRvIII CAR (IL-10 EGFRvIII CAR), b, The expression levels of EGFRvIII CAR were examined by staining the c-Myc tag. Ctrl T, untransduced T cells as control, c, EGFRvIII CAR-T or IL-10 EGFRvIII CAR-T cells were co-cultured with 4Tl-EGFRvIIIl-Luc cells for 3 days (n = 5 biologically independent samples for EGFRvIII CAR-T, n = 10 biologically independent samples for EGFRvIII CAR-T). The culture supernatants were examined for the concentration of IL-10 by ELISA. d, IL-10 EGFRvIII CAR-T or EGFRvIII CAR-T cells in the presence or absence of IL-10 were cocultured with 4Tl-EGFRvIII-Luc cells at the E:T ratio of 5:1 for 18 h (n = 4 biologically independent samples). CAR-T cells were then isolated for a seahorse assay. Shown are average basal OCR. e, IL-10 EGFRvIII CAR-T or EGFRvIII CAR-T cells in the absence or presence of IL-10
0       5      10      15 20 Time post inoculation (day)
(145 ng/ml) were cocultured with 4Tl-EGFRvIIIl-Luc cells at the E:T ratio of 0.5:1 for 48 h (n = 4 biologically independent samples). The percentage of tumor cell lysis was analyzed by flow cytometry, f, Experimental setting is described in Fig. 5 f (n = 5 mice). Shown are relative body weight, g, BALB/c mice were i.v. inoculated with 4Tl-EGFRvIII-Luc cancer cells (5 x io4), lymphodepleted, and received i.v. adoptive transfer of IL-10 EGFRvIII CAR-T cells (3 x 106), or EGFRvIII CAR-T cells (3 x io6) in the presence or absence of i.v. administered IL-10 (1 \ig) on day 6 (n = 5 mice). On day 6,9,12,15, and 18, the blood samples were collected and analyzed by flow cytometry. Shown are counts of viable EGFRvIII CAR-T cells in the peripheral blood of individual mouse. All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (c, d), or by one- or two-way ANOVA with Tukey's multiple-comparisons test (e, g). Data are one representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Raji (s.c.)
CD19hCAR-T
IL-10CD19hCAR-T
5/5 cured
Time post innoculation (day)
PANC1-CD19(S.C.)
PBS
0 20 40
Time post innoculation (day)
CD19 h CAR-T
0 20        40 60
Time post innoculation (day)
IL-10 CD19hCAR-T
rT200-| E
■E 150
100-50 0
0      20     40 60 Time post innoculation (day)
Inoculation of PANC1-CD19-Luc tumor cells in the pancreas of NSG mice
0 20 40 60 80 Time post innoculation (day)
0 20 40 60 80 Time post innoculation (day)
CD19 hCAR-T
IL-10 CD19 hCAR-T
day
CD19 hCAR-T/IL-10 CD19 hCAR-T (i.v.)
PANC1-CD19-Luc (orthotopic)
-»■ PBS
» CD19hCAR-T
IL-10CD19hCAR-T
u		2.5x105n
TO	-í	2x10=-
ra	E	1.5x10s-
□		1x10=-
	"o.	5x104-
<		0-
3     11     18    25 32 Time post inoculation (day)
Inoculation of PANC1-CD19-Luc tumor cells in the pancreas of NSG mice
[IIIMlIflUtífl
[fflfflimnii] iin minii
::iifflffliira]
Rechallenge survivors with PANC1-CD19-Luc tumor cells (orthotopically)
*
0 4 I
CD19hCAR-T/ IL-10 CD19hCAR-T (i.v.)
18        28 50
t        f I
Analyses of CAR-T cells in blood
day
1 250
o
it 200
< 8 150 u ji
-5 3.100
(A
č 50
CD19hCAR-T IL-10 CD19hCAR-T
P= 0.0012
P = 0.0105
Cured [CD19 hCAR-T]  Cured [IL-10 CD19 hCAR-T]
D4
OQQo	1	g f * i				i
OOpo	f	e	i OP	mm		
El.SL 1 il			II,	Ul	m	
18 28 50
Time post inoculation (day)
PANC1-CD19-Luc (orthotopic)
Naive
- Cured [CD19 hCAR-T]
- Cured [IL-10 CD19 hCAR-T]
¥ i"5
™E 104
I 103.
8 102
.1 10' 1 10»
0 2 4 6 8 10 Time post rechallenge (day)
Extended Data Fig. 91 See next page for caption.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Extended Data Fig. 91 Efficacy assessment of hC AR-T cells in the subcutaneous Raji and PANC1-CD19 xenograft tumor models, and the orthotopic PANC1-CD19-Luc xenograft tumor model, a, b, Experimental setting is described in Fig. 5i. a, b, Shown are individual tumor growth curves of the subcutaneous Raji model (a) and PANC1-CD19 model (b). Indicated are numbers of tumor-free mice among the total number of mice in the group (a, b). c-e, PANC1-CD19-Luc human epithelioid carcinoma cells (2 x in6) were implanted in the tail of the pancreas of NSG mice to establish an orthotopic human PDAC model. Mice were then received i.v. adoptive cell transfer of CD19 hCAR-T or IL-10 CD19 hCAR-T cells (1 x io5) on day 4 (n = 5 mice), c, Experimental timeline, d, Longitudinal bioluminescent imaging was performed to monitor orthotopic tumor growth, e, Average radiance (p/s/cm2/sr) of mice of different treatment groups, f-i, PANC1-CD19-Luc human epithelioid carcinoma cells (2 x io6) were orthotopically implanted in the tail of the pancreas of NSG mice. Mice were
then received i.v. adoptive cell transfer of CD19 hCAR-T or IL-10 CD19 hCAR-T cells (5 x io5) on day 4 (n = 5 mice). On day 18,28 and 50, the blood samples were collected and analyzed by flow cytometry. Survivors from treatment groups of CD19 hCAR-T and IL-10 CD19 hCAR-T were rechallenged orthotopically with PANC1-CD19-Luc (2 x io6) cells on day 50 post primary inoculation. Naive NSG mice were inoculated with the same number of tumor cells as controls, f, Experimental timeline, g, Shown are counts of viable hCAR-T cells in peripheral blood on day 18,28 and 50. n = 10 mice pooled from two independent experiments, h, Longitudinal bioluminescent imaging was performed to monitor tumor growth post rechallenge (n = 5 mice), i, Individual radiance (p/s/cm2/sr) showing the rechallenged tumor growth of survivors from indicated treatment groups (n = 5 mice). All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (g, i), or by Mann-Whitney U test (e). Data are one representative of two independent experiments.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Inoculation of MC38-HER2/B16F10 Rechallenge survivors with tumors (s.c.) in C57BL/6 MC38-HER2/B16F10/ or 4T1-EGFRvlll-Luc tumors 4T1-EGFRvll-Luc (i.v.) in BALB/c mice tumor cells (s.c.)
5 6
90 day
Lymphodepletion m£ ^■
IL-10 HER2/TRP-1/EGFRVIII CAR-T (i.v.)
Inoculation of 4T1-EGFRvlll-Luc
tumors (i.v.) in BALB/c        Sacrifice & analyses
_t-i-►
0 5      6        18 day
Lymphodepletion Y
EGFRvlll CAR-T/IL-10 EGFRvlll CAR-T (i.v.)
Blood P< 0.0001
J?
I 8 too
♦   é 80
s <t
5 P 60
o 5
t- □ 40
o
j 20
Blood
P < 0.0001 P < 0.0001 P =0.1433
I33 |
Blood
P < 0.0001 |    P< 0.0001 P = 0.9574
60 40-20
On V Q   40- S " □
. >    .<>-' -x1-     , ~ ,x," tg"
«3-  „?• <£•
«3- >v
Bone marrow
Spleen
Inoculation of PANC1-CD19   Sacrifice & tumors (s.c.) in NSG mice analyses
-»
0       8       63 day I
CD19hCAR-T/ IL-10 CD19hCAR-T (i.v.)
4x10: J 3x102
Q.
S) 2*10: E
S. 1-102-
F	= 0.0132
	A £
	A
—8-°-o-.-	
< ä
O S
co aj
Q E
P 0)
2.5x106-| 2x106
1.5*106 1*106-5*105 0
Bone marrow P = 0.0020
T
o
Extended Data Fig. 101 See next page for caption.
Nature Biotechnology
Article
https://doi.org/10.1038/s41587-023-02060-8
Extended Data Fig. 101 IL-10 expression promotes sternness in mouse and human CAR-T cells, a, The experimental timeline for Fig. 6a-c. b-d, The experimental setting is described in Fig. 6d (n = 5 mice). Shown are frequencies of CD62LhiCD4410 cells among total CAR-T cells (b) and Sca-1*CD122* cells among CD62LhiCD4410 CAR-T cells (c) in blood, d, Shown are the frequencies of IL-7Ra*KLRG-r among total CAR-T cells in blood, e-i, BALB/c mice were inoculated with 4Tl-EGFRvIII-Luc (5 x 10+, i.v.), sublethally lymphodepleted by irradiation on day 5, and received i.v. adoptive transfer of IL-10 EGFRvIII CAR-T cells (3 x io6), or EGFRvIII CAR-T cells (3 x io6) on day 6 (n = 5 mice). On day 18, mice were killed for phenotype analyses of CAR-T cells in spleen and bone marrow by flow cytometry, e, Experimental timeline, f, h, Average frequencies of CD62LhiCD44'° cells among total CAR-T cells in spleen (f) and bone marrow (h). g, i, Average frequencies of Sca-1*CD122* cells among CD62LhiCD44l0CAR-T cells in spleen (g) and bone
marrow (i). j, The experimental setting is described in Fig. 6d. Shown are single-cell expression of key marker genes over the UMAP representation of the map. k-m, NSG mice were inoculated (s.c.) with PANC1-CD19 cells (2 x io6) and received i.v. adoptive cell transfers of CD19 hCAR-T cells (1 x IO6) or IL-10 CD19 hCAR-T cells (1 x io6) on day 8 (n = 4 mice). On day 63, mice were killed for phenotype analyses of CAR-T cells in spleen and bone marrow by flow cytometry, k, Experimental timeline. 1, m, Shown are counts of CD45RA*CD27*CD62L*CD95*CCR7t human CD8* CAR-T cells in spleen (1) and bone marrow (m). n, The experimental setting is described in Fig. 6k. Shown are single-cell expression of key marker genes over the UMAP representation of the map. All data represent the mean ± s.e.m. and are analyzed by two-tailed Student's t-test (f, g, h, i, 1, m), one-way ANOVA with Tukey's multiple-comparisons test (b-d).
Nature Biotechnology
nature portfolio
Reporting Summary
Corresponding author(s):   Li Tang, Yugang Guo, Jie Sun
Last updated by author(s): Oct 9,2023
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a
□
□
□
□
□
□
Confirmed
^1 The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
^1 A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
K-pi The statistical test(s) used AND whether they are one-or two-sided
^ Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested M A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
□
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates {e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)
For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable.
] For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings ~] For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes ] Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.
Software and code_
Policy information about availability of computer code
Data collection
Flow cytometry data collection was performed with Attune NxT Software version 3 (Thermal Fischer Scientific) or LSRFortessa with FACSDiva Software version 8.0.1 (BD Biosciences). Transmission electron microscopy data collection was performed using Velox imaging software (Thermal Fischer Scientific). In vivo bioluminescence data collection was performed using Living Image Software (PerkinElmer). Seahorse data was collected using Seahorse Wave Desktop Software (Agilent). Metabolomics data was collected using the Agilent Quantitative analysis software (version B.07.00, MassHunter Agilent technologies). csRNA sequencing data was collected on a HiSeq4000 (lllumina).
Data analysis
Flow cytometry data were analyzed using FlowJo 10.8.1 (Tree Star, Oregon, USA). Statistical analysis was performed using GraphPad Prism 9 (GraphPad software, Inc, La Jolla, CA, USA). Relative quantification of metabolites was based on EIC (Extracted Ion Chromatogram) areas for the monitored MRM transitions. Signal intensity drift correction and noise filtering was done within the MRM PROBS software. The preprocessed data with peak areas were imported into Metaboanalyst 5.0 for further data analysis. csRNA sequencing raw data of mouse intratumoral HER2 CAR-T cells and human splenic CD19 CAR-T cells were demultiplexed by cellranger mkfastq from lOx Genomics (version 3.0.2) and primary data analysis performed with Cell Ranger (version 2.2.0) using a custom reference package based on mouse reference genome (mmlO or GRCh38) and GENCODE gene models. csRNA sequencing raw data of mouse splenic HER2 CAR-T cells CeleScope rna from Singleron (version 3.0.1) and primary data analysis was performed with CeleScope (version 1.10.0) using a custom reference package based on reference genome (Mus_musculus_ensembl_92). Unsupervised clustering was performed on the PC reduction using the FindNeighbors and FindClusters functions implemented in Seurat with parameters k.param=30 and resolution=0.35. Differentially expressed genes between clusters were determined using the FindMarkers function of Seurat, which applies a Wilcoxon test to determine significance. Differentially expressed genes were visualized using the EnhancedVolcano R package (https://github.com/kevinblighe/EnhancedVolcano) with cutoff on Iog2(fold-change)=0.5 and p-value=10-5. Gene sets for relevant biological processes and pathways were obtained from the database mSigDB, and signature scores for these gene sets were calculated using the UCell package. Gene set enrichment analysis between clusters was performed on average gene expression by cluster using the R package clusterProfiler and relevant signatures from mSigDB.40 and relevant signatures from mSigDB. Projection of HER2 CAR-T or IL-10 HER2 CAR-T cells into a reference atlas of TILs was performed using the ProjecTILs I methodwith default parameters.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
Data_
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy
All data generated and supporting the findings of this study are available within the paper. The scRNA-seq data for murine CAR-T cells from this study have been deposited in the GEO under accession number 245517. The scRNA-seq data for the human CAR-T cells are available in the CNGB Sequence Archive (CNSA) of China National GeneBank Database with accession number CNP0003547. The R code required to reproduce the scRNA-seq data shown in Fig. 4, Fig. 6, Extended Data Fig. 5 Extended Data Fig.10 is alailable at the following repository: http://github.com/carmonalab/LiTang_IL10_CART. The reference genme for primary data analysis of HER2 CAR-T cells in tumors was mmlO, GRCh38for CD19 hCAR-T cells, and ensembl_92 for HER2 CAR-T cells in spleen. Source Data are provided with the online version of the paper. Additional information and materials will be made available upon reasonable request.
Field-specific reporting_
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. £3 Life sciences ] Behavioural & social sciences      ] Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summarv-flat.pdf
Life sciences study design
All studies must disclose on these points even when the disclosure is negative Sample size
Group sizes for in vivo validation experiments were selected empirically based on previous results of the intragroup variation of tumor growth upon similar treatments. Similarly, group sizes in vitro were selected on the basis of prior knowledge of variation. Ref: Nat Immunol [2021Jun;22(6):746.756
Data exclusions
Rout outlier tests were run with default parameters (Q= 1%) in Prism on all mouse experimental data due to inherent variability within the model system.
Replication        [All presented results were repeatable. Replicates were used in all experiments as noted in figure legend or methods.
Randomization    Age and sex-matched animals were used for each experiment. Mice were randomized prior to treatment. In the in vitro experiments, samples with same pretreatment conditions were randomly assigned to a treatment group with a pipette.
Blinding
No blinding was performed due to requirements for cage labeling and staffing needs.
Reporting for specific materials, systems and methods_
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
2
Materials & experimental systems
Methods
n/a	Involved in the study	n/a	Involved in the study
□	£3 Antibodies		] ChlP-seq
□	|^| Eukaryoticcelllines	□	|^| Flow cytometry
El	~2 Palaeontology and archaeology		~2 MRI-based neuroimaging
□	|^| Animals and other organisms		
□	|^| Human research participants		
	~2 Clinical data		
	] Dual use research of concern		
Antibodies
Antibodies used
Validation
The following antibodies or staining reagents were purchased from BioLegend: CD16/32 (93,101302), CD45.2 (104, 109814), CD8(3 (YTS256.7.7, 126606), CD4 (RM4-5, 100526), NK1.1 (PK136, 108740), F4/80 (BM8,123108), CD3e (17A2, 100306), CDllc (N418, 117348), l-A/l-E (MHC-II, M5/114.15.2, 107643), Siglec-F (S17007L, 155508), CD80 (16-10A1, 104734), CD86 (GL-1, 105006), CDllb (Ml/70,101228), Ki67 (16A8, 652424), Granzyme B (GB11, 515403), IFNy (XMG1.2, 505826), TNFa (MP6-XT22, 506308), IL-2 (JES6-5H4, 503822), Gr-1 (RB6-8C5, 202519), PD-1 (29F.1A12, 135216), TIM-3 (RMT3-23, 119706). CD107a (H4A3, 328608), CD44 (IM7, 103028), CD62L (MEL-14,104432), CD122 (5H4,105906), Sca-1 (D7,108106), KLRG1 (2F1/KLRG1,138410), IL-7Ra (SB/199, 121111), human CD95 (DX2, 305606), Streptavidin-PE/Cyanine7 (405206), human Granzyme B (QA16A02, 372208), human IFN-y (B2, 552887). The following antibodies or staining reagents were purchased from BD Biosciences: human CD4 (SK3, 563550), human CD8 (SKI, 557834). The following antibodies or staining reagents were purchased from Invitrogen: human CD62L(DREG56, 48-0629-42), human CD45RA (H100, 69-0458-42), human CD27 (0323, 56-0279-42), CCR7 (3D12, 61-1979-42). Biotinylated Human Her2/ErbB2 Protein (HE2-H822R) was purchased from ACROBiosystems. Myc-tag antibody (9B11, 3739) was obtained from Cell Signaling Technology. Goat anti-mouse IgG was purchased from Jackson ImmunoResearch (Polyclonal, 115-606-003), In VivoMAb anti-mouse CD3 (BE0002,17A2) and In VivoMab anti-mouse CD28 (BE0015-1, 37.51) were purchased from BioXcell.
Validated by manufacturer as indicated on the websites. In Vivo anti-mouse CD3 and CD28 were used for stimulate the proliferation and cytokine production of murine T cells. The other antibodies described above are applied to Flowcytometry in this study, and thier validation can be found on the manufacturer website. Antibodies with human reactivity are labeled as such, remaining antibodies without labels have mouse reactivity.
Eukaryotic cell lines
Policy information about cell lines Cell line source(s)
[hER2 transduced MC38 mouse colon cancer cells (MC38-HER2) were provided by Prof. Pedro Romero (University of Lausanne). B16F10 melanoma cells, Phoenix-Eco cell sand Raji cells were originally obtained from the American Type Culture Collection. 4Tl-EGFRvlll-Luc mouse breast cancer cells were provided by Prof. Darrell J. Irvine (Massachusetts Institute of Technology). NIH/3T3-CD19 cells, PANC1-CD19 cells and PANC1-CD19-Luc cells were generated by Prof. Jie Sun (Zhejiang I University).
Authentication
None of the cell lines were authenticated in these studies. In all related studies, cell lines with low passage number were used.
Mycoplasma contamination      f All cell lines were confirmed mycoplasma negative.
Commonly misidentified lines (See ICLAC register)
No commonly misidentified cell lines were used.
Animals and other organisms
Policy information about studies involving animals: ARRIVE guidelines recommended for reporting animal research Laboratory animals
Five to six-week old female C57BL/6 (C57BL/6J) mice, BALB/c (BALB/cByJ) mice were purchased from Charles River Laboratories (Lyon, France). Six to twelve-week old NOD/SCID/IL-2Ry null mice were purchased from GemPharmatech (Nanjing, China). T-cell receptor (TCR)-transgenic OT-I mice (C57BL/6-Tg(TcraTcrb)1100Mjb/J) were originally purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained in Ecole Polytechnique Federale de Lausanne (EPFL)-Center of PhenoGenomics (CPG) animal facility. Mpclfl/fl mice were crossed to Cd4cre mice on an OT-I background to generate MPC1 knock out (MPC1-KO) OT-I mice. Mpcl fl/fl breeder mice and OT-I breeder mice range in age from six to 40 weeks. Six to twelve-week OT-I T mice and MPC1-KO OT-I mice were used for T cell isolation. Mice were housed under a controlled 12-hour light and 12-hour dark cycle, with tempratures maintained between 22-24 degree and humidity levels between 40-60% to ensure a comfortable environment.
Wild animals
Study did not involve wild animals.
Field-collected samples [Study did not involve field-collected samples Ethics oversight
Experimental procedures in syngeneic mouse studies were approved by the Swiss authorities (Canton of Vaud, animal protocol ID 3206 and 3533) and NSG mouse studies were approved by Institutional Animal Care and Use Committee of Zhejiang University
(Hangzhou, China, animal protocol NO.2021914). Animal studies were performed in accordance with the guidelines from CPG of EPFL, the animal facility of University of Lausanne, and Zhejiang University. Mice were euthanized when body weight loss was beyond 15% of baseline weight, tumor area reached 150 mm2 (syngeneic mouse models) or the maximum diameter reached 20 mm (NSG mouse models), or any signs of discomfort were detected by the investigators or as recommended by the veterinarian who monitored the mice every other day.
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Human research participants
Policy information about studies involving human research participants
Population characteristics
Recruitment
Ethics oversight
Whole blood samples from healthy volunteers (female/male, 19-32 years old) were collected by a doctor of the First I Affiliated Hospital of Zhejiang University School of Medicine
Volunteers recruitment occurred through campus public media. Healthy volunteers aged 18-35 with written informed consent were recruited and the protocol was approved by the Ethics Committee of Zhejiang University School of Medicine (Hangzhou, China, NO.2020-067). Donors were compensated according to Institute policy of Zhejiang University School of Medicine. Other than ensuring good health, no other bias were identifed. While variability among different donors may affect batch-to batch variability, it impact on comparing treatment groups within the same batch of T cells is minor.
^Ethics Committee of Zhejiang University School of Medicine (Hangzhou, China, NO.2020-067).
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Flow Cytometry
Plots Confirm that:
^ The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
^The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
^ All plots are contour plots with outliers or pseudocolor plots.
^ A numerical value for number of cells or percentage (with statistics) is provided.
Methodology Sample preparation
Instrument
Software
Tumors were dissected from the surrounding tissues, weighed, mechanically minced, and stirred at 1000 rpm in RPMI-1640 medium with Collagenase Type IV (1 mg/mL, Gibco / Thermo Fisher Scientific), Dispase II (100 ng/mL, Sigma-Aldrich, St. Louis, Missouri, USA), Hyalurondase (100 ng/mL, Sigma-Aldrich), and DNase I (100 ng/mL, Sigma-Aldrich) at 37 "Cfor 60 min for digestion. Red blood cell lysis was performed on the tumor digestion samples with ACK Lysing Buffer (Gibco /Thermo Fisher Scientific), and then tumor-infiltrating leukocytes were then enriched by density gradient centrifugation against Percoll (GE healthcare) and resuspended in PBS with bovine serum albumin (0.2%, wt/v, Sigma-Aldrich). Spleen were dissected from the surrounding tissues, grinned, and filtered with strainers. Red blood cell lysis was performed on the spleen samples with ACK Lysing Buffer and then resuspended in PBS with bovine serum albumin (0.2%, wt/v, Sigma-Aldrich). Blood samples were collected from abdominal aorta, and resuspend in PBS with 2mM EDTA. Red blood cell was removed by ACK lysis buffer(Gibco/Thermo Fisher Scientific), then red blood cell debris were removed by density gradient centrifugation against Percoll (GE healthcare), and then resuspended in PBS with bovine serum albumin (0.2%, wt/v, Sigma-Aldrich). Bone marrow I was harvested from freshly isolated femurs and tibiae. Bones were crushed in 5 ml of PBS with ethylenediaminetetraacetic acid (2mM, Sigma-Aldrich) and filtered with strainers. Remaining RBCs were lysed with ACK buffer.
Data were collected using Attune NxT Flow Cytometer (Invitrogen / Thermal Fischer Scientific) or LSRFortessa Flow Cytometer (BD Biosciences). Cell sorting was performed with FACS Aria II (BD Biosciences).
Flow cytometry data were analyzed using FlowJo 10.8.1 (Tree Star, Oregon, USA). Flow cytometry data collection was performed using Attune NxT Software version 3 (Invitrogen /Thermal Fischer Scientific) or LSRFortessa with FACSDiva Software (BD Biosciences).
Cell population abundance        Purity was determined by flow cytometry for CAR positive T cells (> 95 %).
Gating strategy
We used standard gating strategies: gating on the typical lymphocyte population based on FSC-SSC signals, doublet exclusion based on FSC-H and FSC-A comparison, Live/Dead discrimination based on DAPI or fixable Aqua dye signals. Cell populations were identified based on the expression markers listed below. CAR-T cells: CD45+/CD3+/CAR+; natural killer cells: CD45+/ CD3-/NK1.1+; dendritic cells: CD45+/Gr-l-/CDllb+/CDllc+/MHCII+; Tumor-associated macrophages: CD45+/Gr-l-/CDllb+/ F4/80+/MHCII+; Gr-l+CDllb+ Myeloid-Derived Suppressor Cells: CD45+/Gr-l+/CDllb+; Eosinophils: CD45+/CDllb+/Siglec-F +. All the representative FACS gating strategies were described in Supplemental Fig. 1.
] Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
4