The growing world of CAR T cell trials: a systematic review

Abstract

In recent years, cancer treatment involving adoptive cell therapy with chimeric antigen receptor (CAR)-modified patient’s immune cells has attracted growing interest. Using gene transfer techniques, the patient’s T cells are modified ex vivo with a CAR which redirects the T cells toward the cancer cells through an antibody-derived binding domain. The T cells are activated by the CAR primary signaling and costimulatory domains. Such “second generation” CAR T cells induced complete remission of B cell malignancies in the long-term. In this fast-moving field with a growing number of engineered T cell products, we list about 100 currently ongoing trials here that involve CAR T cells targeting hematopoietic malignancies and solid cancer. Major challenges in the further development of the therapy are briefly discussed.

Introduction

Over the past few years, adoptive T cell therapy aimed at pre-defined targets has become one of the most promising strategies in cancer immunotherapy, recognized as a “breakthrough therapy” by the US Food and Drug Administration (FDA). T cells were genetically equipped with a chimeric antigen receptor (CAR), which is a composite membrane receptor molecule and provides both targeting specificity and T cell activation (Fig. 1). The CAR targets the T cell through an antibody-derived binding domain in the extracellular moiety, and T cell activation occurs via the intracellular moiety signaling domains when the target is encountered. The CAR design has several advantages over the naturally occurring T cell receptor for antigen (TCR). These include the modular composition of the receptor molecule, redirection toward a broad variety of targets beyond classical MHC-bound peptides, and the combination of signaling domains for sustained T cell activation. Once the CAR engages its cognate target, CAR T cells initiate their immune response with the release of pro-inflammatory cytokines and the cytolytic elimination of the target cell [1]. Activated T cells undergo multiple rounds of amplification and cytolytic attacks, turning them into “serial killers”. To sustain the immune response in the long-term, the primary TCR signal through CD3ξ was combined with a costimulatory signal, mostly CD28 or 4-1BB (CD137), within the same CAR [2, 3] in order to orchestrate a distinct pattern of effector functions. Such “second generation” CARs have proven an important advancement in the development of clinically efficacious T cell therapy. The combination of primary and costimulatory signals sustains distinct T cell functions; the ideal combination of signals seems to be different for the individual T cell subsets and is currently being explored in trials. The 4th generation of CARs contains a CAR inducible expression cassette for a transgene, which may be IL-12 or any other “payload”, thereby making CAR T cells to a targeted factory which delivers a defined transgenic protein to the targeted tissue [4].

Fig. 1.

Families of CARs. a The prototype CAR consists in the extracellular domain of a single chain fragment of variable region (scFv) antibody and a spacer, typically the IgG1 CH2CH3, and in the intracellular domain of the CD3ξ signaling chain providing the primary signal (1st generation CAR). The CAR may harbor one costimulatory or two costimulatory domains in addition to the CD3ξ chain (2nd generation CAR, 3rd generation CAR, respectively). Other signaling chains and various combinations are also described. TRUCK T cells (4th generation) are CAR T cells with CAR inducible release of a “payload” which is a transgenic product, e.g., a cytokine such as IL-12 [4]. b The inhibitory CAR (iCAR) harbors a suppressing instead of an activating signal domain. c Split CAR systems consist of two CARs in the same T cell; the CARs recognize different antigens by their individual scFvs, one CAR providing the CD3ξ signal, the other CAR the costimulatory signal required for full T cell activation (costimulatory CAR) or an suppressor signal to suppress T cell activity (inhibitory CAR). d The switch receptor harbors the PD-1 extracellular domain, or any other receptor for an inhibitory ligand, which is linked to the CD28 costimulatory intracellular domain, thereby switching engagement of an inhibitory ligand into an activating signal

In recent years, procedures have been developed for ex vivo engineering of patient’s T cells with a CAR in good manufacturing practice (GMP) quality and for amplification of the engineered cells to produce clinically relevant numbers for use in adoptive cell therapy. Retro- and lentiviral gene transfer techniques are commonly applied, and DNA and RNA transfer are also used in some trials. Supplementation with IL-7, IL-15 and IL-21 favors the ex vivo expansion of T cells with a less mature phenotype and with improved persistence, along with more potent cytokine release and cytolytic activities. All these activities are thought to be crucial for an efficacious anti-tumor response. In this context, the “pre-conditioning” of the patient’s immune system through non-myeloablative chemotherapy prior to adoptive cell therapy proved to be advantageous in sustaining engraftment and persistence of the transferred CAR T cells. These basic procedures were efficacious in pilot trials, and a number of recent trials are exploring modifications of the basic principle for the treatment of other cancer entities.

CAR T cell trials

As of the beginning of 2016, we are aware of about 100 CAR T cell trials registered at clinicaltrials.gov (Table 1). Most trials have been held in the USA, a growing number is being initiated in Asia and only a few are being carried out in Europe (Table 2). A majority of early-phase trials have been and are currently being performed to treat B cell malignancies, with only a minority of trials targeting solid cancer. Objective regression was achieved in patients with acute lymphoblastic leukemia (ALL), chronic lymphocytic leukemia (CLL) and various other types of B cell lymphoma upon application of CAR T cells which are redirected against CD19. CD19 is a well-established target in the immunotherapy of B cell malignancies due to its restricted expression by mature B cell lineage cells, but not by other hematopoietic cells or non-hematopoietic tissues. Compared with conventional therapies, including advanced radio- or chemotherapy, CAR T cell trials targeting CD19 exhibited a favorable and lasting clinical outcome. For instance, pediatric and adult patients with ALL experienced a complete remission rate of about 90 % and sustained remission for up to 2 years [5]; 63 % of patients with CLL experienced overall responses and 19 % complete remission [6].

Table 1.

Clinical trials in adoptive cell therapy using second- and third-generation CAR T cells

Target antigen	Disease	CAR	Gene transfer	T cell origin	Infused dose	Preconditioning	Number of patients	Response	PI	Center	Identifier	References
BCMA	Myeloma	4-1BB-CD3ξ	NA	Autologous	NA, split dose				Cohen	Abramson Cancer Center	NCT02546167
BCMA	Myeloma	CD28-CD3ξ		Autologous	0.3–15 × 106 CAR T cells/kg, escalating doses	CTX, FLU			Kochenderfer	NCI	NCT02215967
CD19	FL	CD28-CD3ξ	RV	Autologous	108 CAR T cells day 1, 3 × 108 CAR T cells day 2	CTX, FLU, IL-2	1	1 × PR	Rosenberg	NCI	NCT00924326	[35]
CD19	DLBCL, FL	CD3ξ; CD28-CD3ξ	RV	Autologous	0.2, 1, 2 × 108 CAR T cells/m2	None	6	2 × SD, 4 × NR	Savoldo	BCM		[7]
CD19	CLL	CD28-CD3ξ	RV	Autologous	1.2–3 × 107 CAR T cells/kg, 0.4–1 × 107 CAR T cells/kg, split dose over 2 days	3 CLL: none 5 CLL: CTX	8	4 × NR, 1 × PR, 2 × SD	Park	MSKCC	NCT00466531	[14]
CD19	CLL	CD28-CD3ξ	RV	Autologous	0.3–3 × 107 CAR T cells/kg	CTX, FLU	4 (8 in total)	1 × CR, 1 × SD, 2 × PR	Rosenberg	NCI	NCT00924326	[15]
CD19	CLL	CD28-CD3ξ	RV	Autologous	1–4 × 106 CAR T cells/kg	CTX, FLU	4 (15 in total)	3 × CR, 1 × PR	Rosenberg	NCI	NCT00924326	[36]
CD19	CLL	CD28-CD3ξ	RV	Allogeneic, donor derived	1.5, 4.5, 12 × 107 T cells/m2	None	4	1 × PR, 1 × SD	Ramos	BCM	NCT00840853	[8]
CD19	Leukemia	CD28-CD3ξ	RV	Allogeneic, donor derived	0.4–7.8 × 106 CAR T cells/kg	None	10	1 × SD, 2 × NR, 1 × CR	Kochenderfer	NCI	NCT01087294	[37]
CD19	CLL			Autologous	NA, split dose	CTX, FLU			Hosing	MDACC	NCT01653717
CD19	CLL, SLL	4-1BB-CD3ξ	NA	Autologous	1–5 × 107 CAR T cells, 1–5 × 108 CAR T cells				Frey	Abramson Cancer Center	NCT01747486
CD19	ALL	4-1BB-CD3ξ	LV	Autologous	0.14–1.2 × 107 CAR T cells/kg	1: None, 1: ETO-CTX	2	2 × CR	Grupp	UPenn	NCT01626495	[25]
CD19	ALL	4-1BB-CD3ξ	LV	Autologous	0.76–20.6 × 106 CAR T cells/kg, split dose	3: None, 27: physician’s choice salvage chemotherapy	30	27 × CR, 3 × NR	Grupp	UPenn	NCT01626495	[5]
CD19	ALL	CD28-CD3ξ	RV	Autologous	0.3, 1, 3 × 107 CAR T cells/kg, split dose over 2 days	CTX	2 enrolled, 1 treated	1 × CR	Park	MSKCC	NCT01044069	[14]
CD19	ALL	CD28-CD3ξ	RV	Autologous	1.5–3 × 106 CAR T cells/kg, split dose day 1 and 2	CTX	5	5 × CR	Park	MSKCC	NCT01044069	[38]
CD19	ALL	CD28-CD3ξ	RV	Autologous	3 × 106 CAR T cells/kg, 1/3 dose day 1, 2/3 dose day 2	Physician’s choice salvage chemotherapy and CTX	16	14 × CR, 2 × NR	Park	MSKCC	NCT01044069	[27]
CD19	ALL	CD28-CD3ξ	RV	Autologous	1, 3 × 106 CAR T cells/kg	CTX, FLU	21 enrolled	14 × CR, 3 × SD, 3 × PD	Lee	NCI	NCT01593696	[17]
CD19	ALL	CD28-CD3ξ	RV	Allogeneic, donor derived	1.5, 4.5, 12 × 107 T cells/m2	None	4	1 × CR, 1 × NR	Ramos	BCM	NCT00840853	[8]
CD19	ALL	4-1BB-CD3ξ	LV	Allogeneic, donor derived	NA, split dose	NA			Porter	Abramson Cancer Center	NCT01551043
CD19	ALL	CD28-CD137-CD3ξ	LV	Autologous	NA, split dose	NA			Deng, Hu	Affiliated Hospital to Academy of Military Medical Sciences	NCT02186860
CD19	ALL pediatric	NA	LV	Autologous	NA, split dose	CTX			Gardner	Seattle Children’s Hospital	NCT016832794
CD19	ALL	CD28-CD3ξ	LV	Autologous	NA	Yes			Khaled	COH	NCT02146924
CD19	ALL	CD28-CD3ξ	NA	Autologous	1, 3 × 106 CAR T cells/kg, split dose	CTX			Curran	MSKCC	NCT01860937
CD19	ALL	4-1BB-CD3ξ	LV	Autologous	1–5 × 108 CAR T cells	NA			Frey	Abramson Cancer Center	NCT02030847
CD19	Leukemia	CD28-CD3ξ	RV	Autologous	NA	CTX			Park	MSKCC	NCT01416974
CD19	Leukemia/lymphoma	4-1BB-CD3ξ	LV	Autologous	0.146 −16 × 106 CAR T cells/kg, split dose over 3 days	BEN ± RTX; PEN/CTX	30	27 (30) × CR	Frey	Abramson Cancer Center	NCT01029366	[5, 11, 12]
CD19	Leukemia/lymphoma	CD28-CD3ξ versus CD3ξ	RV	Autologous	0.2, 1, 2 × 108 CAR T cells/m2	None	14	14 × no sustained responses	Ramos	BCM	NCT00586391	[31]
CD19	Leukemia/lymphoma	CD3ξ (EBV-CTLs), CD28-CD3ξ (T cells)	RV	Autologous; syngeneic	0.2, 1, 2 × 108 CAR T cells/m2	None			Ramos	BCM	NCT00709033	[31]
CD19	Leukemia	4-1BB-CD3ξ	LV	Allogeneic, donor derived	NA				Gardner	Seattle Children’s Hospital	NCT02028455
CD19	Leukemia/lymphoma	4-1BB-CD3ξ	RV	Autologous; allogeneic	NA, split dose				Han	Chinese PLA General Hospital	NCT01864889
CD19	NHL	4-1BB-CD3ξ	NA	Autologous	1–5 × 108 CAR T cells				Schuster	Abramson Cancer Center	NCT02030834
CD19	NHL	CD28-CD3ξ	RV	Autologous	3 × 105 CAR T cells/kg, 1/3 dose day 1, 2/3 dose day 2	CTX; BEN			Ozawa	Jichi Medical University	NCT02134262
CD19	NHL	CD27-CD3ξ	LV	Autologous					Zhu	Peking University, University of Florida	NCT02247609
CD19	NHL	CD28-CD3ξ	LV	Autologous	TCM versus TN/MEM				Popplewell	COH	NCT02051257
CD19	B-cell NHL	CD28-CD3ξ	LV	Autologous	TCM				Popplewell	COH	NCT01815749
CD19	CLL, NHL, ALL	4-1BB-CD3ξ	LV	Autologous	NA				Maloney	FHCRC	NCT01865617
CD19	B-cell NHL, ALL, or CLL	CD28-CD3ξ	RV	Autologous; allogeneic	0.1, 0.5, 1 × 106 CAR T cells/kg, 0.5, 1, 5 × 106 CAR T cells/kg				Ramos	BCM	NCT02050347
CD19	MCL	4-1BB-CD3ξ	RV	Autologous; allogeneic	NA, split dose				Wang	Chinese PLA General Hospital	NCT02081937
CD19	Stratum 1: NHL stratum 2: CLL, PLL, SLL	CD28-CD3ξ	LV	Autologous	TCM	CTX; BEN; FLU, CTX; ETO, CTX; CTX, ETO			Siddiqi	COH	NCT02153580
CD19	B-cell NHL, ALL, CLL	CD28-CD3ξ, CD28-4-1BB-CD3ξ	RV	Autologous; allogeneic	1, 5, 20 × 106 CAR T cells/m2	CTX			Ramos	BCM	NCT01853631
CD19	HL	4-1BB-CD3ξ	RNA EP	Autologous	NA				Svoboda	Abramson Cancer Center	NCT02277522
CD19	ALL, CLL, NHL	NA	NA	NA	NA				Qian	Southwest Hospital	NCT02349698
CD19	leukemia/lymphoma	CD28-CD3ξ	NA	Allogeneic, donor derived	EBV-CTLs, escalating doses				Curran	MSKCC	NCT01430390
CD19	DLBCL, PMBCL, TFL	NA	NA	Autologous	2 × 106 CAR T cells/kg	CTX, FLU			Go	Kite Pharma	NCT02348216
CD19	Leukemia/lymphoma	CD28-4-1BB-CD3ξ	RV	Autologous	NA	Yes			Enblad, Hagberg	Uppsala University	NCT02132624
CD19	NHL	CD28-CD3ξ	NA	Autologous	0.5, 1, 2 × 107 CAR T cells/kg	Carmustine, ETO, Cytarabine, Melphalan			Sauter	MSKCC	NCT01840566
CD19	B-lymphoid malignancies post-HSCT	CD28-CD3ξ	DNA EP	Autologous	NA, split dose	Carmustine, ETO, Cytarabine, Melphalan			Kebriaei	MDACC	NCT00968760
CD19	B-lineage lymphoid malignancies post-UCBT	CD28-CD3ξ	DNA EP	Allogeneic, donor derived	NA	NA			Shpall	MDACC	NCT01362452
CD19	Leukemia/lymphoma	NA	LV	Autologous	CD8+ TCM	NA			Popplewell	COH	NCT01318317
CD19	Leukemia/lymphoma	NA	NA	Allogeneic	CD8+ CMV-CTLs, CD8+ EBV-CTLs	NA			Turtle	FHCRC	NCT01475058
CD19	myeloma	4-1BB-CD3ξ	LV	Autologous	1–5 × 107 CAR T cells	Melphalan	10	1 × CR, 3 × PR, 4 × PD	Stadtmauer	Abramson Cancer Center	NCT02135406	[39]
CD19	Leukemia/lymphoma	NA	NA	NA	NA	NA			Liang	Shanghai Tongji Hospital, Tongji University School of Medicine	NCT02537977
CD19	Leukemia/lymphoma	NA	NA	Autologous	NA	FLU, CTX			Wang	Shenzhen Second People’s Hospital	NCT02456350
CD19	Leukemia/lymphoma	NA	NA	Autologous	0.5, 1, 3 × 106 CAR T cells/kg	FLU, CTX			Qian	Second Military Medical University	NCT02644655
CD19	Leukemia/lymphoma	NA	NA	Autologous	NA, escalating doses	Yes			Gangyi	Beijing Doing Biomedical Co.; First Hospital of Jilin University	NCT02546739
CD19	Lymphoma	CD28-CD3ξ	RV	Autologous	NA, split dose	Yes			Zhu	Xinqiao Hospital of Chongqing	NCT02652910
CD20	MCL, B-NHL	CD28-4-1BB- CD3ξ	DNA EP	Autologous	Infusion with 108, 109, 3.3 × 109 CAR T cells/m2	CTX, Aldesleukin	4 enrolled, 3 treated	2 × SD, 1 × PR	Till	FHCRC	NCT00621452	[40]
CD20	Lymphoma	4-1BB-FcεRIγ	LV	Autologous	NA				Han	Chinese PLA General Hospital	NCT01735604
CD22	ALL, FL, NHL, LCL	4-1BB-CD3ξ	LV	Autologous	0.3, 1, 3, 10 × 106 T cells/kg	CTX, FLU			Fry	NCI	NCT02315612
CD30	HL, NHL	CD28-CD3ξ	NA	Autologous	0.2, 1, 2 × 108 T cells/m2	NA			Ramos	UNC Lineberger CCC	NCT01316146
CD30	Cutaneous T Cell lymphoma, transformed Mycosis fungoides	CD28Δ-CD3ξ	RV	Autologous	1, 3, 10 × 107 CAR T cells, local application	NA			Mauch	University of Cologne	NCT01645293
CD30	HL, NHL	NA	NA	Autologous	escalating doses	CTX, FLU			Han	Chinese PLA General Hospital	NCT02259556
CD30	HL, NHL	CD28-CD3ξ	NA	Autologous	2, 5, 10 × 108 CAR EBV-CTLs/m2	NA			Heslop	BCM	NCT01192464
CD30	Lymphoma	NA	LV	NA	NA				Zhu	Peking University	NCT02274584
CD33	AML	4-1BB-CD3ξ	LV	Autologous	1.12 × 109 T cells, escalating doses		1	Marked decrease of blasts in bone marrow (2. wk), gradual increase until florid disease progression (9. wk), death (13. wk)	Han	Chinese PLA General Hospital	NCT01864902	[41]
CD123	AML	CD28-CD3ξ	LV	Autologous; allogeneic	NA	CTX, FLU, ETO			Budde	COH	NCT02159495
CD133	Leukemia/lymphoma	CD3ξ; 4-1BB-CD3ξ	NA	Autologous	NA, escalating doses				Weidong	Chinese PLA General Hospital	NCT02541370
CD138	Myeloma	4-1BB-CD3ξ	RV	Autologous; allogeneic	NA				Han	Chinese PLA General Hospital	NCT01886976
LeY	AML	CD28-CD3ξ	RV	Autologous	5–13  ×  108 T cells	FLU	4	3 × clinical responses, 1 × NR		Peter MacCallum Cancer Center		[42]
LeY	Myeloma, AML, myelodysplastic syndrome	CD28-CD3ξ	RV	Autologous	NA	FLU			Prince	Peter MacCallum Cancer Center	NCT01716364
NKG2D-L	Myeloma, AML, myelodysplastic syndrome	NA	NA	NA	1, 3, 10, 30 × 106 CAR T cells	None			Nikiforow	Celdara Medical, LLC	NCT02203825
Igk	Leukemia/lymphoma, myeloma	CD28-CD3ξ	RV	Autologous	0.2, 1, 2 × 108 CAR T cells/m2	CTX			Ramos	BCM	NCT00881920
ROR1	CLL, SLL	NA	NA	Autologous	Starting at 105 CAR T cells/kg, escalating doses	FLU, CTX, RTX; BEN, RTX; FLU, BEN, RTX			Wierda	MDACC	NCT02194374
CEA	Carcinoma	CD28-CD3ξ	RV	Autologous	NA				Junghans	Roger Williams MC	NCT01723306
CEA	Colorectal cancer	CD28-CD3ξ	RV	Autologous	0.1, 1, 10 × 1010 T cells				Junghans	Roger Williams MC	NCT00673322
CEA	Breast cancer	CD28-CD3ξ	RV	Autologous	0.1, 1, 10 × 1010 T cells				Junghans	Roger Williams MC	NCT00673829
CEA	Liver metastases	CD28-CD3ξ	RV	Autologous	3 infusions in 6 weeks into the hepatic artery, intra-patient dose escalating: 0.1, 1, 10, 30 × 109 CAR T cells				Katz	Roger Williams MC	NCT01373047	[30]
CEA	Carcinoma	NA	NA	Autologous	NA				Quian	Southwest Hospital	NCT02349724
CEA	Carcinoma	NA	NA	NA	NA				Quian	Southwest Hospital	NCT02349724
cMet	Breast cancer	NA	RNA EP	Autologous	NA				Tchou	Abramson Cancer Center	NCT01837602
ErbB2	Carcinoma	NA		Autologous	NA, split dose				Wang	Chinese PLA General Hospital	NCT01935843
ErbB2	Carcinoma	CD28-CD3ξ	RV	Autologous	0.01, 0.03, 1, 3, 10, 30, 100 × 106 CAR T cells/m2				Gottschalk	BCM	NCT00889954
ErbB2	Glioblastoma	CD28-CD3ξ	RV	Autologous	1, 3, 10, 30, 100 × 106 CMV-CTLs/m2				Ahmed	BCM	NCT01109095
ErbB2	Sarcoma	CD28-CD3ξ	RV	Autologous	0.01, 0.03, 1, 3, 10, 30, 100 × 106 CAR T cells/m2	CTX, FLU			Ahmed	BCM	NCT00902044
ErbB2	HNSCC	CD28-CD3ξ	RV	Autologous	107–109 CD4+ T cells, escalating doses	None			Maher	KCL	NCT01818323	[43]
ErbB2	Carcinoma	NA	NA	Autologous	NA	CTX, FLU, Mesna			Rosenberg	NCI	NCT00924287
ErbB2	Breast cancer	CD28-CD3ξ	RV	Autologous	NA	Yes			Niu	Fuda Cancer Hospital	NCT02547961
EGFRvIII	Glioblastoma	NA	LV	Autologous	NA				O’Rourke, Chang	Abramson Cancer Center	NCT02209376
EGFRvIII	Glioma	CD28-4-1BB- CD3ξ	RV	Autologous	NA	CTX, FLU, Aldesleukin			Rosenberg	NCI	NCT01454596	[44]
EGFR	Glioma	NA	NA	Autologous	NA				Li	RenJi Hospital	NCT02331693
EGFR	Carcinoma	4-1BB-CD3ξ	LV	Autologous	NA, split dose				Wang	Chinese PLA General Hospital	NCT01869166
EphA2	Glioma	NA	NA	NA	NA	None			Niu	Fuda Cancer Hospital	NCT02575261
FAP	Mesothelioma	CD28-CD3ξ	RV	Autologous	1 × 106 CD8+ CAR T cells				Stupp	University Hospital Zurich	NCT01722149
FR-α	Ovarian cancer		NA	NA					Rosenberg	NCI	NCT00019136
GD2	Sarcoma, melanoma	OX40-CD3ξ	NA	Autologous	0.1, 1, 3, 10 × 106 CAR T cells/kg	CTX			Kaplan	NCI	NCT02107963
GD2	Neuroblastoma	OX40-CD3ξ	RV	Autologous	1.5, 2 × 108 CAR T cells	Pembrolizumab, CTX, FLU			Heczey	BCM	NCT01822652
GD2	Neuroblastoma	NA	RV	Allogeneic, donor derived	NA				Myers	Children’s Mercy Hospital Kansas City	NCT01460901
GD2	Neuroblastoma	CD28-OX40-CD3ξ	RV	Autologous	0.3, 1, 3, 10 × 107 CAR NK T cells	CTX, FLU			Heczey	BCM	NCT02439788
GD2	Sarcoma	CD28-OX40- CD3ξ	NA	Autologous	0.1, 1, 10 × 107 CAR-VZV-CTLs/m2				Wang	BCM	NCT01953900
IL-13Ra2	Glioma	4-1BB-CD3ξ	LV	Autologous	NA				Badie	COH	NCT02208362
L1-CAM	Neuroblastoma, ganglioneuroblastoma	4-1BB-CD3ξ; CD28-4-1BB-CD3ξ	LV	Autologous	0.5, 1, 5, 10 × 106 T cells/kg	Yes			Park	Seattle Children’s Hospital	NCT02311621
Mesothelin	Carcinoma	4-1BB-CD3ξ	LV	Autologous	1, 3, 10, 30 × 107 CAR T cells/m2				Haas	Abramson Cancer Center	NCT02159716
Mesothelin	Metastatic cancer	NA	RV	Autologous	NA	CTX, FLU, Aldesleukin			Rosenberg	NCI	NCT01583686
Mesothelin	Metastatic pancreatic ductal adenocarcinoma (PDA)	4-1BB-CD3ξ	RNA EP	Autologous	1–3 × 108 CAR T cells/m2 (3 times/wk, 3 wks)				Beatty	Abramson Cancer Center	NCT01897415
Mesothelin	Malignant pleural mesothelioma	4-1BB-CD3ξ	RNA EP	Autologous	108 CAR T cells, 3 times and 109 CAR T cells, 3 times		4	1 × transient PR, 1 × PD	Haas	Abramson Cancer Center	NCT01355965	[23, 45]
Mesothelin	Carcinoma	4-1BB-CD3ξ	NA	Autologous	NA, escalating doses, split dose				Weidong	Chinese PLA General Hospital	NCT02580747
Mesothelin	Pancreas carcinoma	4-1BB-CD3ξ	LV	Autologous	1–3 × 107 CAR T cells/m2, 1-3 × 108 CAR T cells/m2, split dose, two separate infusions (Meso CAR T cells and CD19 CAR T cells)	CTX			Ko	UPenn	NCT02465983
MUC1	Carcinoma	CD28-4-1BB-CD3ξ (SM3scFv, IL-12; pSM3scFv)	LV	Autologous	5 × 105 CAR T cells per tumor lesion, intratumoral injection	None	1	pSM3-CAR treated tumor lesion showed necrosis	Yang	PersonGen Biomedicine	NCT02587689	[46]
MUC1	Carcinoma, glioma	NA	NA	NA	NA				Yang	PersonGen Biomedicine, Anhui General Hospital of Armed Police Forces	NCT02617134
PSMA	Prostate cancer	CD28-CD3ξ	RV	Autologous	1, 3, 10 × 107 CAR T cells/kg	CTX			Slovin	MSKCC	NCT01140373
PSMA	Prostate cancer	NA	RV	Autologous	1, 10 × 1010 T cells				Junghans	Roger Williams MC	NCT00664196
PSMA	Prostate cancer	NA	RV	Autologous					Junghans	Roger Williams MC	NCT01929239
VEGFR-II	Metastatic cancer, melanoma, renal cancer	NA	RV	Autologous	1 × 106–3 × 1010 T cells, escalating doses	CTX, FLU, Aldesleukin			Rosenberg	NCI	NCT01218867

Trials using second or third generation CAR T cells in the treatment of malignant diseases and registered at http://clinicaltrials.gov as of May 2016 are listed

ALL acute lymphoblastic leukemia, BCM Baylor College of Medicine, BEN bendamustine, CLL chronic lymphocytic leukemia, COH City of Hope Medical Center, CR complete response, CTX cyclophosphamide, DLBCL diffuse large B cell lymphoma, EP electroporation, ETO etoposide, FHCRC Fred Hutchinson Cancer Research Center, FL follicular lymphoma, FLU fludarabine, HL Hodgkin’s lymphoma, LV lentiviral, MC Medical Center, MCL mantle cell lymphoma, MDACC MD Anderson Cancer Center, MSKCC Memorial Sloan Kettering Cancer Center, NA not available, NCI National Cancer Institute, NHL non-Hodgkin lymphoma, NR non-responder, PEN pentostatin, PLL prolymphocytic leukemia, PMBCL primary mediastinal B cell lymphoma, PR partial response, RTX rituximab, RV retroviral, SD stable disease, SLL small lymphocytic lymphoma, TCM central memory T cell, TN/MEM naïve memory T cell, UCBT umbilical cord blood transplantation, UPenn University of Pennsylvania, wk week

Table 2.

World distribution of clinical trials with second- and third-generation CAR T cells

USA and Canada	74
PR China	27
Europe	4
Japan	1
Australia	1

The clinical benefit of CAR T cell therapy in these and other trials is well documented. However, the success rate varies broadly between the trial centers and the pathological type of the disease. For instance, 2 out of 6 lymphoma patients achieved stable disease for some months before progression [7], while half the patients (4 out of 8) experienced complete and partial remission in another trial [8]. However, a direct comparison of the outcome is difficult to make due to a number of technical and design variations, including the CAR composition, production and amplification of CAR T cells, patient pre-conditioning and cytokine support, CAR T cell dose and other variations [9]. Lymphodepletion is also a relevant variable in both autologous and allogeneic cell therapy (Table 3). A recent meta-analysis of CD19 CAR T cell trials confirmed lymphodepletion and CAR T cell dose as key factors for a favorable prognosis, while IL-2 co-administration is not recommended [10]. The situation is even more complex when targeting solid cancer lesions. This is due to the fact that, in contrast to CD19, most targeted antigens are likewise expressed by healthy tissues, although often at lower levels. Designed to test for safety, the phase I trials explored the possibility of ruling out the “on-target off-tumor” auto-reactivity of CAR T cells toward healthy tissue which in case of targeting CD19 produced a lasting depletion of healthy B cells. First, phase I trials targeting B cell leukemia/lymphoma have been successfully completed and are currently entering phase II development by global pharmaceutical companies.

Table 3.

Patient pre-conditioning by non-myeloablative lymphodepletion impacts clinical outcome of CAR T cell therapy

CAR T cells	Pre-conditioning		No pre-conditioning
Allogeneic			n = 8	CR: 2 (25 %) PR: 1 SD: 2 NR: 3
Autologous	n = 75	CR: 53 (70.7 %) PR: 7 SD: 7 NR: 8	n = 24	CR: 1 (4.2 %) PR: 14 SD: 2 NR: 7

Hematologic malignancies

The first clinical investigations at the University of Pennsylvania with second generation CAR T cells achieved two complete responses in the treatment of three patients with refractory advanced CLL using anti-CD19 CAR T cells [11, 12]; a recent update by the same group confirmed an overall response rate of 57 % [13]. T cells engineered with a “second generation” CAR with combined 4-1BB-CD3ξ signaling underwent extensive amplification upon administration to the patients, eliminated high tumor burdens and persisted for at least 3 years, with retention of anti-tumor activity. With respect to the eliminated tumor mass, it was calculated that one CAR T cell is capable of killing as many as 1000 leukemic cells. At the same time, similar clinical results were obtained by the Memorial Sloan Kettering Cancer Institute [14] and National Cancer Institute [15] using anti-CD19 CAR T cells with combined CD28-CD3ξ signaling. In all trials, the anti-tumor effect correlated with the persistence of CAR T cells in the peripheral blood of the individual patients. Up to now, over 40 trials are targeting CD19 to treat hematologic malignancies (Table 1), including non-Hodgkin’s lymphoma (NHL), CLL and ALL, with other trials targeting CD20 (NCT00621452), CD22 (NCT02315612), Igκ (NCT00881920) and the B cell maturation protein BCMA (NCT02215967).

Larger trials on the treatment of adult and pediatric ALL revealed even more impressive response rates. Complete clinical responses of 90 % were reported in 30 relapsed or refractory ALL patients with ongoing responses over 4 years [5]. Likewise, other trials also obtained complete response rates of 88 and 70 % [16, 17]. A number of patients had previously undergone allogeneic hematopoietic stem cell transplantation, and the T cells from the transplant donor were engineered with the respective CAR. The kinetics of ALL cell clearance in those patients was more rapid compared to that for CLL patients. In addition to an increased sensitivity toward CAR T cells, some further differences in the disease itself may also contribute to the effect, including different tumor burdens, accessibility to immune cells and a more homogeneous patient population treated thus far. In the long-term, most patients treated with 4-1BB-CD3ξ CAR T cells did not receive further treatment, whereas patients treated with CD28-CD3ξ CAR T cells frequently underwent subsequent allogeneic stem cell transplantation. The difference may be biased by the clinical decision to prefer watchful waiting over consolidation of the results, a decision that is based partly on the prolonged persistence of 4-1BB-CD3ξ CAR T cells of up to 2 years, compared to about 30 days for CD28-CD3ξ CAR T cells.

Solid cancer

The application of the CAR T cell strategy to non-hematopoietic cancer requires the consideration of additional factors. These include disease status and tumor burden, immune repression in the tumor tissue and CAR T cell infiltration, as well as the recruitment and activation of other pro-inflammatory and repressor immune cells. Some early trials are currently investigating the safety and potential side effects of CAR T cell-targeting of solid cancer lesions after systemic application, while other trials apply CAR T cells through local administration. Not all effective variables for therapeutic efficacy have as yet been identified; among these, the choice of target antigen is a major issue. Redistributed expression of an antigen throughout the entire membrane of cancer cells, while polarized expression in glandular or intestinal epithelia, makes such an antigen targetable by CAR T cells. These targets include prostate-specific membrane antigen (PSMA), carcinoembryonic antigen (CEA) and mucin 1 (Muc1), all currently being investigated in clinical trials. Two trials using anti-CEA CAR T cells are currently recruiting (NCT00673829, NCT01723306); one trial was terminated due to treatment-related adverse effects (NCT01212887).

CAR T cell therapy is a tightrope walk

Although adoptive CAR T cell therapy achieved spectacular efficacy in the treatment of leukemia/lymphoma, this treatment is often associated with significant toxicities which need to be taken into account, in particular, when entering advanced phases of clinical evaluation.

1. Most toxicities in CAR T cell therapy are due to the engagement of the cognate antigen on healthy cells. Such “on-target off-tumor” targeting may harbor life-threatening risks, in particular, when the antigen is expressed by essential tissues such as lung, heart or liver. In the case of anti-CD19 CAR T cells for the treatment of B cell leukemia/lymphoma, lasting B cell aplasia and, consequently, hypo-gammaglobulinemia are consistently observed, which is considered a biomarker for anti-CD19 CAR T cell function. The situation can be clinically managed by immunoglobulin substitution, and patients generally did not develop opportunistic infections. To allow reconstitution of the B cell compartment, CAR T cell elimination after complete tumor eradication would be ideal which can be achieved by various means as described below.

   The severity of “on-target off-tumor” toxicity became obvious after the death of a patient during the treatment of metastatic colon cancer (NCT00924287), when CAR T cells targeted HER2/neu (ErbB2), which is highly expressed by the carcinoma cells but also by healthy tissue, although at lower levels, causing acute toxicity toward cardiopulmonary epithelia [18]. Other trials are currently exploring anti-HER2 CAR T cells in the treatment of brain tumors and various carcinomas (NCT00889954, NCT00902044, NCT01818323, NCT01935843). Despite the risk of toxicity, anti-HER2 CAR T cells can be safe when using a less potent signaling CAR and applying a more cautious dose-escalation regimen [19].

   In cases of toxicity, CAR T cells can be eliminated by the activity of co-expressed inducible suicide genes (herpes simplex virus thymidine kinase, HSV-tk; inducible caspase 9, iCasp9), by the administration of a clinically approved depleting antibody targeting an epitope within the CAR, or a co-expressed non-functional surface protein, all strategies that are being explored in trials. For instance, HSV-tk, which phosphorylates the guanosine analog ganciclovir which initiates the arrest of DNA replication, is used in trials targeting PSMA in prostate cancer (NCT01140373), CD19 in NHL (NCT00182650) and L1-CAM in neuroblastoma (NCT00006480). Inducible caspase 9, which encodes a modified truncated caspase 9 and initiates an apoptotic cascade upon drug-initiated dimerization, is being explored in trials targeting the disialoganglioside GD2 in neuroblastoma (NCT01822652), sarcoma (NCT01953900), other solid cancers (NCT02107963) and in trials targeting CD19 for the treatment of NHL (NCT02247609). The truncated EGFR is co-expressed with the CAR in the same T cell and targeted by the anti-EGFR therapeutic antibody cetuximab in various trials (NCT01815749, NCT02051257).

   The level of antigen expression is clearly a critical factor for “off-tumor” auto-immunity. Moreover, the CAR T cell strategy demands a target which is exclusively expressed by cancer cells, ideally required for their survival and harboring mutations that are large enough to produce new epitopes that can be specifically recognized by the CAR [20]. In the absence of such an ideal target, however, target antigens are preferred that are either co-expressed by non-essential tissues or topologically sequestered from T cells. In the latter case, epithelial antigens with polarized membrane distribution in healthy cells, while uniformly distributed on cancer cells, like CEA, provide some tumor selectivity based on altered topology.

   Several strategies are currently being developed to furthermore increase tumor selectivity. These strategies include the co-expression of inhibitory CARs (iCARs) with PD-1 or CTLA-4 intracellular domains and targeting a surface antigen on healthy tissues which is not present on cancer cells, by an inhibitory CAR thereby providing a dominant inhibitory signal when engaging healthy cells (Fig. 1) [21]. Alternatively, the CAR T cell activation may depend on the CAR engagement of antigen pairs, where only binding to both antigens by both CARs drives full T cell activation [22]. Clinical exploration of the latter strategy remains challenging since it requires the adjustment of CARs to the individual antigen levels in tumors and healthy tissues.

   In this situation, T cells which are transiently modified by a CAR, thereby limiting the persistence and function of CAR T cells, were proposed. RNA-modified CAR T cells require repetitive CAR T cell infusions as applied in a trial that targeted mesothelioma (NCT01355965) [23]. Other trials are using RNA-modified CAR T cells against pancreatic cancer (NCT01897415), breast cancer (NCT01837602) and Hodgkin’s lymphoma (NCT02277522).

2. CAR T cell activation is accompanied by extensive release of toxic levels of pro-inflammatory cytokines which can cause cytokine release syndrome (CRS), characterized by nausea, fever, hypotension, vascular leakage and life-threatening multiple organ failure. This syndrome is closely associated with systemic macrophage activation syndrome, resembling hematophagocytic lymphohistiocytosis. CRS is correlated with clinical efficacy. The pathomechanism seems to be based on the extensive activation of CAR T cells, with high IFN-γ and TNF-α release and the activation of monocytes or macrophages, which release IL-6, which in turn produces a toxic effect on various organs such as kidney, liver and brain. Major risks for CRS are therefore thought to be a high tumor burden and the dose and potency of applied T cells. However, patients with low tumor loads also suffered from CRS. The severity of CRS requires intensive care treatment, and the death of two cancer patients in 2014 has demonstrated that CRS constitutes a severe limitation of CAR T cell therapy [15, 24, 25]. Interference with the IL-6 pathway through application of the neutralizing anti-IL-6 antibody tocilizumab successfully reduced the symptoms without eliminating CAR T cells [26]. In order to clinically monitor CRS and to grade the severity of CRS, a clear treatment algorithm has recently been established based on tumor burden, age, comorbidities and other factors [27, 28].

   Acute and reversible neurotoxicity after CD19 CAR T cell application were observed in about 40 % of patients during CRS, with 20 % of patients experiencing late-onset manifestations with aphasia, hallucinations, and delirium. Adult patients and children equally suffered from CRS [29]. While the exact pathomechanism remains unclear and only partly overlaps with CRS, there is some evidence that CAR T cells may infiltrate the brain parenchyma, induce diffuse encephalopathy and release IL-6, which is not neutralized by i.v. application of tocilizumab.

3. Alternative routes for the delivery of T cells beyond i.v. infusion are also being explored to minimize the risk of “off-tumor” toxicities. In particular, the CAR T cells were administered in a localized fashion by intratumoral (CEA, NCT01373047; HER2, NCT01818323), intracerebral (EGFRvIII, NCT01454596) or intrapleural (mesothelin, NCT01355965) injections. Individual solid cancer metastases were treated by locally introducing the CAR T cells into the tumor lesion. For instance, endoscopic delivery of CAR T cells to CEA⁺ metastases in the liver (NCT01373047) resulted in shrinkage of metastases without systemic side effects [30]. Likewise, mesothelin-specific CAR T cells were introduced into the pleura of mesothelioma patients [23]. Local application of the CAR T cells is assumed to have some advantages, including the prompt antigen-induced activation to sustain robust T cell amplification, persistence in the tumor lesion and the low dose of CAR T cells required to induce efficacy when compared to systemic infusion. At present, some clinical trials have incorporated such tools for minimizing potential systemic toxicities.

4. Once accumulated at the tumor site, the CAR T cell anti-tumor response is repressed by various means, for instance, by preventing escape from the vasculature and penetration into the tumor tissue, repression by repressor cells and/or soluble factors, or by nutrient deprivation. Moreover, CAR T cells express immune repressive receptors upon activation, including programmed cell death-1 (PD-1) and cytotoxic T lymphocyte-associated antigen-4 (CTLA-4). Pre-clinical models have demonstrated strategies to counteract the PD-1 or CTLA-4-mediated CAR T cell repression, while the effect of the co-administered anti-CTLA-4 antibody, ipilimumab, is currently being investigated in a pilot trial (NCT00586391) [31].

Future perspectives

While CAR T cell therapy for refractory leukemia/lymphoma is becoming more and more established in clinical practice, CAR T cell therapy for solid cancer is still in its infancy and requires extensive clinical investigation [32]. Moreover, the comparison of trial results is complicated by a number of differences in CAR T cell formulation, trial protocols, pre-conditioning of patients and other factors. The current situation demands a more rigorous standardization; some major obstacles need attention in the near future.

1. Target selection: Most targeted antigens are tumor-associated but not tumor-selective as tumor-associated antigens are also expressed by healthy tissues, making them potential CAR T cell targets. In extensive pre-clinical studies and sophisticated mouse models, researchers are trying to define the risk of CAR-mediated immunity against healthy tissues. However, the limitations of such models demand a responsible design of early-phase trials. It more and more becomes obvious that CAR T cells can be safe when targeting with less affinity the appropriate target and with less signaling capacities. Translation to clinical application still demands a cautious dose-escalation regimen and a rigorous clinical monitoring. In this respect, trials are aiming at minimizing the “off-tumor” risk while increasing CAR T cell efficacy toward cancer cells. As a consequence, engineered T cells, which recognize target antigen signatures, instead of an individual antigen, indicative for cancer cells by split receptor systems (Fig. 1), are experimentally explored. Two CARs are co-expressed in T cells complementing their activating signals when engaging both antigens, thereby increasing cancer selectivity and likely improving therapeutic safety. In a further development, a CAR with extracellular PD-1 domain, which binds to PD-L1 and PD-L2, provides CD28 co-stimulation as the required second signal in an immune suppressive tumor environment with PD-1 ligand expression [33]. Such split CARs can also be used to provide inhibitory signals to the T cell when targeting an antigen on healthy tissues.

2. CAR design and T cell formulation: Different CAR designs with respect to the extracellular spacer, transmembrane domain and costimulatory moieties are used; most modifications are merely empiric or were found functional in the particular context, others still need systemic evaluation. So far, the binding affinity, the targeted epitope and the extracellular length of the CAR were identified to be crucial for optimal T cell activation. Some optimization applies for the ex vivo amplification of engineered T cells to clinically relevant numbers, which is performed in the presence of IL-2, now preferentially in the presence of IL-7 and IL-15. In addition, the T cell subset undergoing CAR modification substantially impacts clinical outcome. Standardization of the manufacturing processes will help to obtain cell products which allow comparing the clinical performance in different trials on a more robust basis. Currently available clinical data suggest that CAR T cells with 4-1BB-ξ signaling domain inducing a central memory phenotype persist longer than T cells with CD28-ξ CAR inducing an effector memory phenotype. Moreover, a recent analysis points to differences in the metabolic reprogramming by 4-1BB-ξ versus CD28-ξ CARs to central versus effector T cells, respectively [34]. However, other T cell subsets and T cells in a more advanced stage of maturation need other costimulatory signals or combinations thereof to persist in the long-term. Until now, autologous and allogeneic CAR T cells were applied with different outcome; however, there are too many differences in the trial and CAR design which do not allow a well-founded trial comparison. Future efforts should be focused on a series of comparative trials with the aim to identify the optimal design.

3. Pre-conditioning: CAR T cell amplification after application is clearly required for an anti-tumor response; non-myeloablative lymphodepletion of patients prior T cell therapy is crucial in this context (Table 3). The additional impact of pre-conditioning on sensitizing the tumor milieu by depleting suppressor cells or inducing the release of antigen needs to be evaluated; however, the pre-treatment protocols and their efficacies substantially differ and need further standardization. In addition, the optimal dose of IL-2 substitution during and shortly after T cell therapy in order to sustain T cell amplification and the use of alternative cytokines such as IL-7 and/or IL-15 need to be clinically explored.

4. Control of side effects: The cytokine release and the vascular leakage syndromes are severe, life-threatening side effects which seem to be associated with anti-tumor activity of CAR T cells but need intensive care attention. With the establishment of a CRS screening protocol and IL-6 neutralization [27–29], first steps toward a more standardized clinical management of those side effects of the CAR T cell therapy are being made.

5. Hematopoietic stem cell transplantation: When CAR T cell therapy has induced clinical remission, a number of patients were transplanted with allogeneic hematopoietic stem cells. While the anti-leukemia efficacy of CAR T cells is clearly shown, the capacity to control leukemia in the long-term without stem cell transplantation needs to be established in forthcoming trials.

Abbreviations

ALL

Acute lymphocytic leukemia

CEA

Carcinoembryonic antigen

CLL

Chronic lymphocytic leukemia

CRS

Cytokine release syndrome

NHL

Non-Hodgkin’s lymphoma

PSMA

Prostate-specific membrane antigen

tk

Thymidine kinase

Compliance with ethical standards

Conflict of interest

Hinrich Abken serves on the scientific advisory board of Miltenyi Biotec. The other authors declare no conflict of interest.