Multipurposing CARs: Same engine, different vehicles

Abstract

T cells genetically engineered to recognize and eliminate tumor cells through synthetic chimeric antigen receptors (CARs) have demonstrated remarkable clinical efficacy against B cell leukemia over the past decade. This therapy is a form of highly personalized medicine that involves genetically modifying a patient’s T cells to recognize and kill cancer cells. With the FDA approval of 5 CAR T cell products, this approach has been validated as a powerful new drug in the therapeutic armamentarium against cancer. Researchers are now studying how to expand this technology beyond its use in conventional polyclonal αβ T cells to address limitations to the current therapy in cancer and applications beyond it. Considering the specific characteristics of immune cell from diverse lineages, several preclinical and clinical studies are under way to assess the advantages of CAR-redirected function in these cells and apply the lessons learned from CAR T cell therapy in cancer to other diseases.

Graphical abstract

Chimeric antigen receptors (CARs) expressed in conventional polyclonal αβ T cells have demonstrated remarkable clinical efficacy against B cell leukemia. Current research is exploring ways to expand this technology to alternative immune cells in order to address limitations to current therapy in cancer and applications beyond it.

Introduction

Chimeric antigen receptor (CAR) T cells have achieved remarkable success in the treatment of blood cancers1, 2, 3, 4, 5, 6 and are a fast growing sector of immuno-oncology. This technology involves the genetic modification of autologous αβ T cells to express synthetic receptors against antigen present on tumor cells. These receptors enable modified T cells to recognize and produce an immunogenic response against cancer. To date, five CAR T cell therapies have been approved by the U.S. Food and Drug Administration (FDA) to treat B cell malignancies, four targeted toward CD19 (tisagenlecleucel, axicabtagene ciloleucel, brexucabtagene autoleucel, and lisocabtagene maraleucel) and one directed against B cell maturation agent (BCMA) (idecabtagene vicleucel).^4,7, 8, 9, 10 These achievements have sparked new hope for the application of CARs to a wider range of cancers and other diseases. To this end, there are currently more than 500 clinical trials investigating the therapeutic efficacy of CAR T cells.¹¹

Despite some clinical success, many challenges remain that limit the therapeutic applicability and efficacy of CAR T cells against a wider range of diseases. In cancer, obstacles to effective therapy include the reduced durability and tumor homing of CAR T cells, tumor antigen loss, the immunosuppressive tumor microenvironment (TME), and the time and cost associated with manufacturing autologous T cell products.12, 13, 14, 15 Moreover, serious toxicities associated with CAR T cell therapies, such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), have raised major concerns.

Investigators are dedicating tremendous effort worldwide to overcome these issues. A number of clinical and preclinical research are under way to replace the bespoke manufacturing needed with autologous, patient-derived CAR T cells products with donor-derived allogeneic CAR T products. However, the possibility of life-threatening graft-versus-host disease (GvHD) with an allogeneic cell therapy product has motivated researchers to use the CAR molecule in immune cells other than “conventional” polyclonal αβ T cells. These include “nonconventional” T cells, natural killer (NK) cells, and macrophages. In this review we discuss recent developments in CAR-based cellular immunotherapy using immune cell types aside from conventional T cells.

CARs in nonconventional T cells

Natural killer T cells

Type 1 natural killer T (NKT) cells are a small and distinct subset of αβ T cells that are CD1d restricted and express an invariant T cell receptor.16, 17, 18, 19 NKT cells mainly recognize self and foreign lipid antigens.^16,20 Invariant natural killer T (iNKT) cells are a major subset of NKT cells and express the unique invariant Vα24Jα18 TCRα chain in humans.²¹ iNKT cells control the adaptive immune response by coordinating diverse immune cells such as CD4+/CD8+ T cells, NK cells, dendritic cells, and B cells.22, 23, 24, 25, 26

The immune regulatory potential and the rapid cytokine production of iNKT cells suggest that these cells may be harnessed for redirection with CAR. To this end, Poels et al.²⁷ expressed CD38 or B cell maturation antigen-directed CARs in Vα24-iNKT cells and demonstrated effective elimination of multiple myeloma (MM) cells. Additionally, Xu et al.²⁸ demonstrated potent anti-tumor activity by GD2-directed CAR NKT cells against neuroblastoma cells. In addition, autologous GD2-directed CAR NKT cells co-modified to secrete interleukin (IL)-15 demonstrated superior anti-tumor function, cellular persistence, and tumor infiltration compared with GD2-CAR-NKT cells without IL-15 against neuroblastoma. Interim reporting from a phase I clinical trial using these anti-GD2 CAR NKT cells (NCT03294954) demonstrated that the cells underwent in vivo expansion and localization to the tumor in patients.²⁹ Another allogeneic CAR NKT cell (ANCHOR) clinical trial (NCT03774654) is ongoing to study anti-CD19 CAR NKT cells modified to secrete IL-15 in patients with lymphoma.³⁰

Although the CAR NKT cell strategy is in its early stages, the inherent pro-inflammatory characteristics of NKT cells, such as robust cytokine production and rapid cytotoxicity,²⁹ lend these cells great potential in adoptive cell therapy.

Cytokine-induced killer cells

Cytokine-induced killer (CIK) cells are a unique subset of cytotoxic T lymphocytes that demonstrate both T cell and NK cell phenotype and are capable of tumor killing in a major histocompatibility complex (MHC)-unrestricted manner.31, 32, 33 CIK cells are manufactured by ex vivo activation of peripheral blood mononuclear cells (PBMCs) or cord blood cells with interferon (IFN)-γ, anti-CD3 monoclonal antibody, and interleukin-2 for 2–3 weeks. These cells express NK cell markers in addition to T cell markers.³⁴ Increased expression of NKG2D and DAP10 in CIK cells plays a major role in killing tumor cells.³⁵ Importantly, the overproduction of IFN-γ by CIK cells results in reduced GvHD in mice.^36,37

Because of the multivariate immunotherapeutic role of CIK cells, incorporation of CAR into this cell type has demonstrated clinical success.³⁸ In a recent phase I and II clinical trial (NCT03389035), allogeneic anti-CD19 CAR-expressing CIK cells (CARCIK-CD19), engineered using the non-viral Sleeping Beauty (SB) transposon, demonstrated complete response (CR) or CR with incomplete blood count recovery (CRi) at day 28 in 6 of 7 patients with B cell acute lymphoblastic leukemia (B-ALL).^38,39 Five of 6 patients who achieved CR in this study were minimal residual disease (MRD) negative, and no patients developed GvHD or neurotoxicity at the highest dose tested. Robust expansion of CARCIK-CD19 was observed in the majority of patients, and the CAR-expressing cells were detectable 10 months post-infusion. This research group also developed CD33-targeting CAR CIK cells using the SB transposon system for the treatment of acute myeloid leukemia (AML), which were significantly effective against AML and chemotherapy-resistant AML in a preclinical setting.⁴⁰ In another study, Merker et al.⁴¹ engineered ERBB2-CAR CIK cells to treat high-risk rhabdomyosarcoma. This preclinical study demonstrated that ERBB2-CAR CIK cells eradicated RH30 tumors in mice but failed to clear tumor burden in relapsed disease. Finally, second-generation CAR CIK cells specific to chondroitin sulfate proteoglycan 4 (CSPG4) was developed by Leuci et al.⁴² to treat soft tissue sarcomas (STS) and demonstrated strong anti-tumor activity in STS spheroid and xenograft mouse models compared with unmodified CIK cells.

Taken together, the promising safety profile in patients in an allogeneic setting, the capacity for MHC-unrestricted cytotoxicity, and a unique NK cell-like phenotype make CAR CIK cells a viable alternative approach to target cancer.

γδ T cells

γδ T cells are a distinct subgroup of T cells based on the variable chains in the T cell receptor (TCR) protein complex.⁴³ Although only 1%–5% of circulating T cells are γδ T cells, these cells are abundant at epithelial surfaces.⁴⁴ γδ T cells demonstrate functional characteristics of both the innate and adoptive immune systems through expression of cytotoxic receptors such as NKG2D, NKp30, and/or NKp44 and Toll-like receptors (TLRs).45, 46, 47, 48, 49, 50 γδ T cells can kill cancer cells in a major histocompatibility complex-unrestricted manner, a function that distinguishes these cells from αβ T cells.⁵¹ Moreover, upon activation, γδ T cells can function as professional antigen-presenting cells to activate CD8 αβ T cells by cross-presenting tumor antigen.⁴⁴

Given these characteristics, adoptive transfer of CAR-modified γδ T cells can be a viable potential therapeutic option for some cancer patients. A recent study demonstrated that γδ T cells expressing a GD2-targeting second-generation CAR improved the tumor killing and expansion capacity of unmodified γδ T cells, without impairing the cells’ ability for tumor-specific migration.⁴⁴ Furthermore, a short-term (4 h) in vitro cytotoxicity assay in this study demonstrated that tumor lysis by GD2-specific CAR γδ T cells was equivalent to GD2-specific CAR αβ T cells. A separate study demonstrated that CD19-targeting CAR γδ T cells were effective in clearing disease from the bone marrow in a mice model of leukemia.⁵² Importantly, this study also showed that unlike conventional CD19-targeting CAR T cells, CAR γδ T cells were also capable of killing leukemia cells in a CD19-independent manner. This approach is noteworthy, as CD19 antigen loss is a major problem for patients who relapse after treatment with conventional CAR T cells.53, 54, 55 Thus, γδ CAR T cells may be able to overcome antigen escape. Additionally, as γδ CAR work in an MHC-independent manner and do not cause GvHD, these cells can be used in allogeneic setting.

Induced pluripotent stem cell-derived T cells

Many different strategies are under investigation for the development of “off-the-shelf” CAR T cell therapy. The tedious manufacturing process and high cost of autologous CAR T cells may exclude some patients from receiving this therapy. The use of induced pluripotent stem cells (iPSCs) as a starting source for CAR T cells is an attractive approach. Production and expansion of reprogrammed iPSC from antigen-specific cytotoxic T cell clones have demonstrated anti-tumor function in preclinical models.56, 57, 58, 59 Other approaches produce T lymphocytes from iPSCs in feeder cell or serum-free culture conditions.60, 61, 62 For example, iPSC clones can be generated by retroviral transduction of peripheral blood T lymphocytes with reprogramming factors KLF4, SOX2, OCT-4, and C-MYC.⁵⁹ These transduced cells are then seeded onto mouse embryonic fibroblast (MEF) feeder cells in T cell media. T-iPSCs are generated after further culturing in embryonic stem cell (ESC) medium for 22–25 days. The resulting T-iPSCs were then lentivirally transduced to express a CD19-directed second-generation CAR (called 19-28zT-iPSC) and expanded on Delta like 1-expressing OP9 (OP9-DL1) feeder cells. The 19-28zT-iPSCs showed activation of T cell markers such as CD25 and CD69, secretion of cytokine markers such as IL-2, tumor necrosis factor (TNF)-α and interferon-γ, and reactivity against CD19 antigen, indicating successful generation of iPSC-derived CAR T cellular therapy. 19-28zT-iPSC lysed tumor cells in vitro and demonstrated anti-tumor function in vivo.

Despite the successful generation of CAR T iPSCs from PBMCs, the use of feeder cells at multiple stages has raised safety concerns for contamination in the final product. Thus, a feeder-free iPSC differentiation system can be a promising approach for cellular immunotherapies. A recent study by Iriguchi et al.⁶⁰ outlined a feeder-free iPSC system to generate induced T (iT) cells. This approach involves inducing hematopoietic progenitor cells (HPCs) to form embryoid bodies, differentiation of HPCs into CD4+CD8αβ+ cells, differentiation of CD4+CD8αβ+ cells to CD8αβ+ single-positive cells, and finally expansion of CD8αβ+ single-positive cells. The feeder-free iPSC T cells are modified to express anti-CD19 CAR (iCART). These cells express low levels of CD45RA and CCR7 but increased levels of CD45RO and CD62L. iCART cells had no expression of PD1 and TIM3 exhaustion markers but did expressed LAG3 and exerted improved anti-tumor activity in vitro and in vivo against NALM-6 tumor-bearing mice compared with unmodified iT cells.

Regulatory T cells

Given the promising results in targeting hematological malignancies using CARs, efforts have been made to use CARs to redirect T cells for other purposes. Regulatory T cells (Tregs) are a subset of CD4+ T cells characterized by expression of the transcription factor FOXP3 and the immunophenotype CD25+, CD127^low. Tregs promote immune tolerance through suppression of multiple immune effectors, including cytotoxic T cells. Adoptive transfer of Tregs for the treatment of multiple medical conditions is an active area of investigation.^63,64 Several groups have explored the use of CARs to target Tregs to specific immune cell populations in preclinical models of graft-versus-host disease, organ transplantation, and multiple autoimmune diseases.

CARs targeted to major histocompatibility complex class I molecules have been shown by multiple groups to suppress the activity of alloreactive cytotoxic T cells without eliminating MHC I-expressing cells.65, 66, 67 Tregs expressing CARs on the basis of antibodies against the common MHC allele HLA-A2 (frequently referred to as A2-CAR) were shown to promote immune tolerance in multiple preclinical models of GvHD and tissue transplantation.65, 66, 67, 68, 69, 70 Forced overexpression of FOXP3 to maintain the Treg phenotype in A2-CAR Tregs has also been tested in preclinical GvHD systems.⁷¹ Furthermore, co-expressing IL-10 by A2-CAR Tregs enhanced the suppressive capacity of these cells.⁷² In a more flexible and modular approach, CAR Tregs were designed to bind monoclonal antibodies to MHC I molecules, and this system promoted allograft tolerance in multiple preclinical tissue transplantation settings.⁷³ Immune cells besides MHC I-expressing T cells can also be targeted by CAR Tregs; for example, expression of CD19-targeted CARs in Tregs suppressed differentiation and IgG antibody production by CD19+ B cells, thereby reducing GvHD⁷⁴ in one model system.

In addition to GvHD and tissue transplantation, the use of CAR Tregs has also been explored in autoimmune diseases. Some of the earliest work in the field demonstrated that CAR Tregs are able to reduce tissue inflammation in multiple preclinical models of autoimmune colitis75, 76, 77 and that this reduction in inflammation reduced subsequent colorectal tumor burden.⁷⁷ Various CAR Treg approaches have also demonstrated activity in preclinical models of hemophilia A (HA), a clotting disorder characterized by deficiency of clotting factor VIII (FVIII), which is often associated with formation of antibodies that inhibit therapeutic recombinant FVIII. CAR Tregs targeting FVIII suppressed both B and T cell responses to therapeutic FVIII in HA models,⁷⁸ similar to what was seen using a transgenic T cell receptor.⁷⁹ T cells modified to express both FOXP3 and a CAR targeted to FVIII and sorted for a Treg immunophenotype inhibited proliferation of FVIII-specific effector T cells and improved clotting activity in a murine model of HA.⁸⁰ Tregs expressing a CAR incorporating an FVIII antigen (termed a B cell-targeting antibody receptor [BAR]) were shown to inhibit B cells selectively, reducing levels of FVIII-specific antibodies.^81,82 Preclinical activity of CAR Tregs has been showing in other autoimmune diseases, including reactive airway disease/asthma,⁸³ rheumatoid arthritis,⁸⁴ vitiligo,⁸⁵ and multiple sclerosis.⁸⁶ Tregs expressing CARs targeting insulin and pancreatic endocrine antigens have been designed to address type 1 diabetes. However, although these efforts have demonstrated proof of concept, they were ultimately unable to prevent diabetes in preclinical models.^87,88

Importantly, several of these studies have demonstrated that the CAR molecule provides antigen-specific localization and activation of the CAR Tregs, thus providing an advantage of increased specificity compared with unmodified polyclonal Tregs. Additionally, compared with cytotoxic T cells, Tregs can suppress lymphocytes targeted to other antigens in the local microenvironment (often referred to as “bystander suppression”), without damaging normal cells that express the target antigen. This creates a localized zone of reduced inflammation and immune tolerance even when multiple or unknown antigens are targeted by endogenous effector cells and antibodies.

From a translational perspective, human CAR Tregs have been produced,^89,90 and methods to generate CAR Tregs have been described.^91,92 Multiple groups have explored the optimal costimulatory domain for CAR Tregs. These efforts have generally suggested that, in contrast to some observations with antineoplastic cytotoxic CAR T cells, CD28-based co-stimulation is superior to other costimulatory systems (e.g., 4-1BB) for CAR Treg activity.^93,94 Similar to efforts in conventional CAR αβ T cells (reviewed by Rafiq et al.⁹⁵), technologies to develop modular CAR Tregs for increased utility have been demonstrated.^73,96

CAR NK cells

Natural killer cells are granular lymphocytes that are CD3− and express surface markers CD56+ and NKp46+. Although considered to be a part of innate immunity,^97,98 NK cells are functionally parallel to CD8⁺ cytotoxic T cells and kill target cells through similar cytotoxic mechanisms. However, these cells lack a somatically rearranged and unique antigen-specific TCR.99, 100, 101 NK cells are able to recognize and kill infected cells, allogeneic cells, and tumor cells without prior sensitization in a non-HLA (human leukocyte antigen)-restricted manner. NK cells are able to exhibit potent anti-tumor immunity through multiple mechanisms.¹⁰² Activated NK cells are capable of eliminating tumor cells through direct cell cytotoxicity with perforin and granzymes and/or production of pro-inflammatory cytokines.¹⁰³ In addition, cytokines and chemokines such as IFN-γ and granulocyte macrophage colony-stimulating factor (GM-CSF) produced by NK cells can modulate the immune responses of other cells.^104,105

Because of several limitations with autologous αβ T cell therapy, NK cells represent a promising cell type for the application of CAR technology. First, allogeneic NK cells do not need complete HLA matching and cause little or no graft-versus-host disease compared with allogeneic T cells. This opens the possibility to produce “off-the-shelf” allogeneic cell therapy products that can be prepared in advance and are readily available for multiple patients on demand.106, 107, 108 In fact, preclinical data in mouse models suggest that NK cells may prevent GvHD in some cases by inhibiting activated alloreactive T cells while retaining beneficial graft-versus-tumor effects.¹⁰⁸ On the contrary, CAR T cells can cause GvHD even when HLA matched, as minor mismatches may still be reactive.¹⁰⁸ Additionally, the levels of cytokines release by NK cell infusions are moderate¹⁰⁹ compared with the levels seen in life-threatening cases of CRS caused by CAR T cells.¹⁰⁶ NK cells have a limited lifespan in circulation and do not produce memory cells except against some viruses. Thus, a primary NK cell-based CAR approach may not need to be modified to include a suicide gene as a safety switch to reduce the CRS and other harmful side effects of long-lasting CAR T cells such as on-target, off-tumor toxicities.^14,110,111 However, suicide switches may still be necessary in CAR NK cells that are further modified to enhance in vivo expansion (133). A further potential advantage of using CAR NK cells is that NK cells can be derived from umbilical cord blood (UCB) or cell lines such as NK-92, both of which have the potential for use as “off-the-shelf” therapeutic products.¹⁰⁶

NK cell source

PBMCs, UCB, CD34+ hematopoietic progenitor cells, NK cell lines (such as NK-92), and induced pluripotent stem cells can be used to manufacture clinical-grade NK cells. Because of the unlimited proliferative capability of NK-92 cells in vitro and consistent phenotypic and functional features in the presence of interleukin-2, these cells are a major source of CAR-NK cells.112, 113, 114, 115 NK-92 cells provide distinct advantages for use in manufacturing “off-the-shelf” CAR-NK products for clinical use. These include shorter manufacturing time and lower cost. NK-92 cells were initially derived from a non-Hodgkin lymphoma patient, and thus being a tumor-derived cell line, have inherent drawbacks such as the potential risk for tumorigenicity.¹¹⁶ NK-92 cells lack expression of FcγRIII and thus cannot mediate antibody-dependent cell-mediated cytotoxicity. Furthermore, these cells require lethal irradiation before clinical use, which diminishes in vivo expansion potential. These characteristics prevent NK-92 cells from being an ideal cell source for most CAR-NK cell therapy approaches.^116,117

Another important source of primary NK cells is human PBMCs, which have been used in numerous clinical trials (see Table 1). Several methods have been used to easily isolate NK cells from healthy donors. The most commonly used method is depletion of CD3 cells followed by selection of CD56 cells that are cultured with IL-2. Good Manufacturing Practices (GMP)-grade NK cells can also be prepared using lymphokine-activated killer (LAK) cell culture media with IL-15 for 28 days.118, 119, 120 Activated PBMC-derived CAR NK cells expressing a wide range of activating receptors and can be administered without irradiation, allowing in vivo expansion. PBMC-derived NK (PB NK) cells, up to 90% of which are CD56^dimCD16+, typically exhibit a mature phenotype, with increased cytotoxicity and reduced proliferative capacity.¹²¹

Table 1.

Current clinical trials of CAR engineered immune cells, excluding conventional αβ T cells

CAR immune cells	ClinicalTrials.gov identifier	Target	Disease	Starting material	CAR construct	Sponsor	Purpose
NK cells	NCT03692767	CD22	refractory B cell lymphoma	NA	NA	Allife Medical Science and Technology Co., Ltd.	to investigate the safety and efficacy of anti-CD22 CAR NK
	NCT03690310	CD19	refractory B cell lymphoma	NA	NA	Allife Medical Science and Technology Co., Ltd.	to assess the safety and efficacy of anti-CD19 CAR NK
	NCT03692637	Mesothelin	epithelial ovarian cancer	NA	NA	Allife Medical Science and Technology Co., Ltd.	to investigate the safety and efficacy of anti-mesothelin CAR NK
	NCT03692663	PSMA	Castration-resistant prostate cancer	NA	NA	Allife Medical Science and Technology Co., Ltd.	to evaluate the safety and feasibility of anti-PSMA CAR NK cells
	NCT03824964	CD19/CD22	refractory B cell lymphoma	NA	NA	Allife Medical Science and Technology Co., Ltd.	to study on the safety and efficacy of anti-CD19/CD22 CAR NK cells
	NCT03940833	BCMA	R/R multiple myeloma	NK-92 cells	NA	Asclepius Technology Company Group (Suzhou) Co., Ltd.	to assess the safety and feasibility followed by BCMA CAR-NK cells
	NCT03941457	ROBO1	pancreatic cancer	NK-92 cells	NA	Asclepius Technology Company Group (Suzhou) Co., Ltd.	to assess safety profile followed by administration of BiCAR-NK cells
	NCT03940820	ROBO1	solid tumor	NK-92 cells	NA	Asclepius Technology Company Group (Suzhou) Co., Ltd.	to assess the safety and effectiveness of ROBO1 CAR-NK cells to treat solid tumors
	NCT04324996	viral S protein and host NKG2DL	COVID-19	UCB cells	NA	Chongqing Public Health Medical Center	to assess the safety, tolerability, and efficacy of NKG2D-ACE2 CAR-NK cells
	NCT04245722	CD19	B cell lymphoma, CLL	iPS cells	CAR.19-NKG2D-2B4-CD3ζ-IL15RF-hnCD16	Fate Therapeutics	to assess the incidence, nature and safety of FT596 (anti-CD19 CAR NK) monotherapy and with combination of anti-CD20 antibody
	NCT03383978	HER2	GBM	NK-92 cells	anti-HER2 NK-92/5.28.z	Johann Wolfgang Goethe University Hospital	to determine safety, tolerability and maximum tolerated dose (MTD) or maximum feasible dose (MFD) followed by NK-92/5.28z.
	NCT04847466	PDL-1	gastroesophageal junction (GEJ) cancers, advanced HNSCC	NK-92 cells	NA	National Cancer Institute (NCI)	to determine the clinical response rate (CR + PR) with irradiated PD-L1 CAR-NK cells in combination with N-803 plus pembrolizumab
	NCT01974479	CD19	B-ALL	NK cells	anti-CD19-BB-zeta	National University Health System, Singapore	to determine the minimal disease residual (MRD) monitoring followed by NK cell infusion
	NCT02944162	CD33	AML	NK-92 cells	anti-CD33,28,BB and zeta	PersonGen BioTherapeutics (Suzhou) Co., Ltd.	to determine safety profile followed by administration of anti-CD33 CAR-NK cells
	NCT02742727	CD7	lymphoma, leukemia	NK-92 cells	anti-CD7,28,BB and zeta	PersonGen BioTherapeutics (Suzhou) Co., Ltd.	to determine safety profile followed by administration of the anti-CD7 CAR-pNK cells
	NCT02839954	MUC1	solid tumor	NK-92 cells	anti-MUC1-pNK	PersonGen BioTherapeutics (Suzhou) Co., Ltd.	to determine safety profile followed by administration of the anti-MUC1 CAR-pNK cells
	NCT02892695	CD19	lymphoma, leukemia	NK-92 cells	anti-CD19,28,BB and zeta	PersonGen BioTherapeutics (Suzhou) Co., Ltd.	to determine the adverse events attributed to the administration of the anti-CD19 CAR-NK cells
	NCT04887012	CD19	B cell non-Hodgkin lymphoma	NK cells	NA	Second Affiliated Hospital, School of Medicine, Zhejiang University	to study the safety and effectiveness of HLA haploidentical CAR-NK cells
	NCT00995137	CD19	B-ALL	NK cells	anti-CD19-BB-zeta	St. Jude Children's Research Hospital	to determine the maximum tolerated dose of genetically modified NK cells
	NCT03415100	NKG2DL	metastatic solid tumor	NK cells	NA	The Third Affiliated Hospital of Guangzhou Medical University	to determine the anti-tumor response followed by CAR-NK cell infusions
	NCT04796675	CD19	ALL CLL NHL	CB cells	NA	Wuhan Union Hospital, China	to evaluate the primary safety and efficacy of CAR-NK-CD19
	NCT04639739	CD19	NHL	NA	NA	Xinqiao Hospital of Chongqing	to assess safety and efficacy of anti-CD19 CAR NK
	NCT03656705	NA	solid tumor	NK cells	NA	Xinxiang medical university	to evaluate the safety and feasibility of CAR-NK cell treatment
	NCT03056339	CD19	relapsed and refractory B lymphoid malignancies	UCB	CD19-CD28-zeta-2A-iCasp9-IL15	M.D. Anderson Cancer Center	to assess the optimal dose, toxicity, and efficacy of CD19-CD28-zeta-2A-iCasp9-IL15
NKT cells	NCT03774654	CD19	refractory B cell lymphoma, relapsed adult ALL relapsed CLL, relapsed non-Hodgkin lymphoma	NKT cells	NA	Baylor College of Medicine	to determine the dose limiting toxicity (DLT) rate
	NCT04814004	CD19	relapsed, refractory, or high-risk B cell malignancies	iNKT	hCD19.IL15.CAR-iNKT	Kai Lin Xu; Jun Nian Zheng and Xuzhou Medical University	to evaluate the safety and feasibility of IL-15-expressing anti-CD19 CAR iNKT (hCD19.IL15.CAR-iNKT) cells
	NCT03294954	GD2	neuroblastoma	NKT cells	CAR.GD2-CD28-CD3ζ-IL15	Baylor College of Medicine	to determine the maximum tolerated dose (mTD)
CAR γδ T cells	NCT04702841	CD7	T-ALL	γδ T cells	NA	PersonGen BioTherapeutics (Suzhou) Co., Ltd.	to evaluate the efficacy of CAR - γδ T cells
	NCT04107142	NKG2DL	colorectal cancer, triple-negative breast cancer, sarcoma, nasopharyngeal carcinoma, prostate cancer, gastric cancer	γδ T cells	NA	CytoMed Therapeutics Pte Ltd	to evaluate the safety of CTM-N2D and the feasibility to produce CTM-N2D for three target dose levels
	NCT04735471	CD20	B cell malignancies	γδ T cells	NA	Adicet Bio, Inc	to assess the safety and efficacy study of ADI-001 anti-CD20 CAR-engineered allogeneic gamma delta T cells in adults with B cell malignancies, in monotherapy and combination with IL-2
CAR Tregs	NCT04817774	HLA-A2	kidney transplantation	Treg cells	NA	Sangamo Therapeutics	safety and tolerability of TX200-TR101
CAR macrophages	NCT04660929	HER2	HER2-positive carcinoma	macrophages	NA	Carisma Therapeutics Inc	assess the safety and tolerability of CT-0508 and to assess the feasibility of manufacturing CT-0508
CAR CIK cells	NCT03389035	CD19	B-ALL	CIK cells	CAR.19-CD28-OX40- CD3ζ	Fondazione Matilde Tettamanti Menotti De Marchi Onlus	feasibility and safety of a single dose of transposon-manipulated allogeneic CARCIK-CD19
iPSC-derived CAR T cells	NCT04629729	CD19	B cell lymphoma, CLL, B cell ALL	iPS cells	NA	Fate Therapeutics	to assess the incidence, nature, and safety of FT819 (truncated TCR anti-CD19 CAR T) monotherapy and with combination with IL-2

NK cells can also be isolated from UCB or the placenta. Although about 10¹⁰ NK cells, with about 92.37% viability, can be obtained from a 21 day culture, these numbers are still limited for clinical use. Compared with PBMC-derived NK cells, UCB NK cells have an immature phenotype, with decreased expression of certain adhesion molecules such as CD2, CD11a, CD18, CD62L, CD16, KIRs, perforin, and granzyme B, and higher expression of inhibitory molecules such as NKG2A.¹²² This ultimately leads to lower tumor cell lysis potential against K562 cells.¹²³

NK cells derived from induced pluripotent stem cells have become an attractive source of CAR NK cells because of the unlimited proliferative capacity of these cells¹²⁴ with improved anti-tumor activity.¹²⁵ Compared with differentiated NK cells, iPSCs can be more efficiently engineered to stably express CARs. Li et al. demonstrated that TrypLe-adapted CAR-expressing iPSCs can be cultured for 11 days in APEL culture media containing human stem cell factor (SCF), human vascular endothelial growth factor (VEGF), and recombinant human bone morphogenetic protein 4 (BMP4) differentiated into hematopoietic progenitor cells. CAR NK cells are then allowed to further differentiate with IL-3, IL-15, IL-7, SCF, and FLT3L in culture medium.^115,125 This study demonstrated that NK cells derived from iPSCs (NK-CAR-iPSC-NK) significantly inhibited tumor growth and prolonged survival as compared with CAR T cells.

NK CAR designs

Considering the distinct properties of NK cells, several novel CAR constructs have been designed especially for signaling within these cells. The novel CARs impart specificity on NK cells and demonstrate improved cytotoxicity and cytokine production.^126,127

CAR constructs contain four major structural elements: a target-specific binding domain, a hinge region, a transmembrane domain, and an intracellular signaling domain.¹⁴ Manipulating each of these components can significantly change the cellular behaviors and tumor-killing properties of the transduced cell. Thus, CAR NK cell function can greatly depend the CAR design. For instance, Imai et al.¹²⁸ reported that a second-generation CD19-targeting scFv CAR with 4-1BB costimulatory and CD3ζ activation domain expressed in primary NK-92 cells demonstrated increased activation, interferon production, and markedly enhanced CD19-targeted killing of acute lymphoblastic leukemia (ALL). CD28 costimulatory domains have also been studied in CAR NK cells¹²⁹ against HER2, EGFR or EGFR variant III, and CS1.130, 131, 132 Although 4-1BB/CD28-containing CARs are engineered for use in T cells, these constructs can exert anti-tumor activities when used in NK cells. Considering the need for NK-specific co-stimulation, domains such as DAP10, DAP12, and 2B4 are under investigation.133, 134, 135 NK cells with 2B4 costimulatory domains were reported to demonstrate enhanced cytotoxicity against CD5⁺ malignant T-ALL cells with rapid proliferation, increased cytokine production, and decreased apoptosis compared with transduced NK cells bearing the conventional 4-1BB containing CAR.¹³⁶ In another study, CARs containing the DAP12 signaling domain demonstrated more efficient tumor lysis than receptors with the CD3ζ signaling domain.¹³⁷ These results suggest that NK cell-specific activation signaling may affect CAR performance.¹³⁶ An NK CAR construct optimization study was performed by Li et al.¹²⁵ testing ten different anti-mesothelin CARs in NK-92 cells in cytotoxicity assays. Among the ten CAR structures studied, only the three that contained NKG2D transmembrane and 2B4 costimulatory domains exhibited enhanced expression of CD107a and increased in anti-tumor activity when stimulated by mesothelin^high targets. The recruitment of DAP10 can be a possible mechanism for improved performance by these NKG2D TM and 2B4 under the condition of mesothelin^high A1847 stimulation.¹¹⁵

In a recent study, CAR NK (S309-CAR-NK) cells were generated against the highly conserved epitope of spike (S) glycoproteins present on different variant of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viruses using an S309 scFv domain. This preclinical study demonstrated specific killing of S protein expressing A549 cells representing pseudo-SARS-CoV-2 virus¹³⁸ and highlighted the potential for the use of CAR NK cells against infectious diseases.

Clinical trials with CAR NK cells

Despite the safety and promising clinical efficacy of unmodified allogeneic NK cells and the potential to address some of the key challenges associated with CAR T cells, the clinical translation of CAR NK cells is still in early stages of development.¹³⁹

Several of the clinical trials reported thus far demonstrate promising activity against different types of cancer. For instance, early results from the first large-scale trial (NCT03056339) using CAR-NK cells derived from umbilical and cord blood (CB) demonstrated no increase in the levels of inflammatory cytokines and thus no CRS in patients with CD19+ CLL and B cell lymphoma.¹⁴⁰ Of the 11 patients who were treated on this trial, 8 (73%) had objective responses, 7 (4 with lymphoma and 3 with CLL) had complete remission, and 1 had remission of the Richter’s transformation component but had persistent CLL. Although this study demonstrated promising preliminary results, most patients had short-term responses, and 5 of 8 patients received post-remission therapy.

As of August 2021, there are 25 studies registered at ClinicalTrials.gov evaluating the safety and efficacy of CAR-NK cells in cancer patients, including two additional completed phase I/II trials (Table 1). Among them, the safety data results from one of the trials (NCT00995137) are not available yet. The other phase I safety trial (NCT02944162) targeting CD33 in relapsed and refractory AML demonstrated no major adverse effect when dosed up to 5 × 10⁹ cells per patient. However, this trial did not demonstrate any obvious clinical efficacy.¹⁴¹

CARs in other immune cells

Multipotent mesenchymal stem cells

Multipotent mesenchymal stem cells (MSCs), also known as mesenchymal stromal cells, are a heterogeneous group of progenitor cells. MSCs were first isolated from the bone marrow stroma in late 1960s and are capable of forming colony-forming unit fibroblast (CFU-fibroblast).142, 143, 144 MSCs have self-renewal capacity and can differentiate into several other cell types, including adipocytes, osteoblasts, and chondrocytes.144, 145, 146, 147 Aside from the immune regulatory role of MSCs, the tumor-promoting phenotype of these cells has also been a focus of research.147, 148, 149, 150, 151, 152

Researchers hypothesize that the crosstalk between the immune suppressing effect of MSCs and tumor cells contributes to tumor growth and progression.^151,153 Because of the high tumor-homing capacity of MSCs, including to metastatic lesions and less immunogenic milieu, these cells have become an attractive approach as targeted cellular therapy. A particularly innovative approach to infiltrating solid tumors is the modification of stromal cells themselves, rather than T cells. Multiple groups have investigated the modification of MSCs to express truncated CAR molecules that recognize tumor-associated antigens in order to enhance the innate ability of MSCs to infiltrate solid tumors. This approach was shown to enhance MSC tumor infiltration into glioma xenografts¹⁵⁴ and ovarian xenografts¹⁵⁵ and facilitated the delivery of the apoptosis-promoting molecule TRAIL to tumor cell in vitro.¹⁵⁶

Additionally, Hombach et al.¹⁵⁷ engineered MSCs to release IL-7 and IL-12 in the TME to create favorable environment for CAR T cells. The engineered MSCs demonstrated improved activation, proliferation, and cytokine production of carcinoembryonic antigen-targeting CAR T cells. It is well studied that IL-12 and IL-7 not only promote the cytotoxicity of CAR T cells but also modify the tumor microenvironment and enhance T cell infiltration and ultimately exert improved anti-tumor function.158, 159, 160, 161, 162, 163, 164 Thus, secretion of IL-7 and IL-12 by MSCs was an innovative approach to reduce solid cancer progression. Alternatively, Aliperta et al.¹⁶⁵ demonstrated that expression of bispecific antibodies (bsAbs) by SCP-1 mesenchymal stromal cells redirected T cells against a CD33-expressing leukemia model.

Macrophages

Tumor-associated macrophages (TAMs) infiltrate solid tumor microenvironments and play role toward immune suppression and tumor progression.166, 167, 168 TAMs can make up to 50% of the total solid tumor mass.¹⁶⁹ The ability to infiltrate tumors and the high prevalence of macrophages in tumor microenvironment make macrophages an attractive cell type for CAR expression and have motivated researchers to assess CAR-directed tumor cell phagocytosis by macrophages.

Morrissey et al.¹⁷⁰ demonstrated that bone marrow-derived macrophages (BMDMs) engineered to express anti-CD19 CAR with a CD8 transmembrane domain and Megf10 and FcRɣ cytosolic domains (CAR-Ps) triggered antigen-dependent phagocytosis. CAR-P demonstrated enhanced engulfment of different sized antigen-coated beads and significantly reduced Raji cells numbers. Similarly, Klichinsky et al.¹⁷¹ developed CD19-directed CAR macrophages (CAR-Ms) containing a CD3ζ intracellular domain that demonstrated phagocytosis of tumor cells in an antigen-specific manner. Solid tumor targeting anti-mesothelin or anti-HER2 CAR macrophages eradicated antigen-positive cancer cells in vitro and in preclinical mouse models.

Macrophage Toll-like receptor-chimeric antigen receptors (MOTO-CARs) are engineered to add Toll/interleukin-1 receptor (TIR) signaling domain to polarized macrophages toward a pro-inflammatory M1 phenotype.¹⁷² A recent conference abstract demonstrated that these cells express high levels of CD14, CD80, and CD206 and low levels of CD163. As M1 macrophages have tumor suppressor activity, MOTO-CARs exhibited target-specific killing of solid tumor. Finally, a recent study of induced pluripotent stem cell-derived, CAR-expressing macrophage cells (CAR-iMac) showed similar enhanced antigen-induced tumor phagocytic properties.¹⁷³

Apart from targeting cancer, CAR macrophages have been tested against SARS-CoV-2 viruses. A recent study by Fu et al.¹⁷⁴ tested a series of CAR macrophages to target various recognition domains of the S protein present in SARS-CoV-2. This study demonstrated that MER tyrosine kinase (MERTK)-expressing CAR macrophages (CAR_MERTK) are capable of reducing viral load without upregulating of pro-inflammatory cytokines except IFN-γ, IL-6, and IL-8 in an in vitro setting.

Current limitations and next steps

Despite the remarkable success of conventional CAR αβ T cells in certain disease settings, expression of CAR in different immune cells is being explored to address the limitations of current therapy. These include the generation of off-the-shelf/allogeneic products such as NK cells, CIK cells, or γδ T cells to avoid manufacturing failures or using immune cells such as Tregs and macrophages with unique characteristics to address hurdles in cancer and other diseases.

However, similar to CAR T cells, other CAR-expressing cell therapies still face obstacles to clinical efficacy, such as loss of targeted antigen, tumor heterogeneity, and the hostile TME. Therefore, strategies should be considered to maximize the efficacy of CAR-based immune cell therapy in the future. Unlike T cells, many of these alternative cells can mediate cytotoxic activity against cancer cells in both CAR-dependent and CAR-independent manner. Therefore, a CAR can be developed to induce a moderate activating signal. For example, a CAR without costimulatory domains in NK cells could preferentially use the natural killing mechanism of NK cells and weakly use CAR-mediated direct cytotoxicity. One study reported that the use of a 4-1BB costimulatory domain caused excessive activation-induced cell death and reduced numeric expansion of NK cells compared with respective CD28-based constructs,²⁸ indicating the careful consideration should be taken while designing CAR constructs in alternative immune cells. These cells could even be engineered to express a non-signaling CAR that does not induce a direct killing signal but promotes homing and adhesion to targets, allowing cytotoxicity triggered by natural killing mechanisms, with no or minimal on-target, off-tumor effects. In this scenario, a wider range of antigens could be used as targets, including antigens such as HER2, EGFR, and mesothelin that are preferentially expressed by tumor cells but also maintain low levels in some healthy cells.

Additionally, using advanced gene engineering technologies, immune cells can be developed to co-express other molecules, including cytokines, antibodies, protease, and others, that are able to promote immune cell proliferation, trafficking, and penetration into tumor. Engineered CAR cells that can express molecules to reprogram the tumor microenvironment, switching an immunosuppressive environment to an immuno-activating one, can encourage infiltration of endogenous immune cells to facilitate anti-tumor immune responses independent of the CAR-mediated cytotoxicity.¹⁷⁵

Finally, using these approaches to expand the use of CARs to diseases beyond cancer is promising and highlights the vast potential of this technology. Several preclinical CAR α/β T cell therapy studies are under active investigation to treat diseases such as cardiac fibrosis,¹⁷⁶ HIV,^177,178 aging,¹⁷⁹ and type 1 diabetes.¹⁸⁰ The use of other cellular platforms with CAR technology may widen the number of diseases that can be treated using this approach (Figure 1).

Figure 1.

Advantages and disadvantages of CAR engineered immune cells, excluding conventional αβ T cells

NK, natural killer; Treg, regulatory T cell; CIK, cytokine-induced killer; TCR, T cell receptor; HLA, human leukocyte antigen; CRS, cytokine release syndrome; GvHD, graft-versus-host disease. Figure created using BioRender.com