CAR-modified T-cell therapy for cancer: an updated review

Abstract

The use of chimeric antigen receptor (CAR)-modified T cells is a promising approach for cancer immunotherapy. These genetically modified receptors contain an antigen-binding moiety, a hinge region, a transmembrane domain, and an intracellular costimulatory domain resulting in T-cell activation subsequent to antigen binding. Optimal tumor removal through CAR-modified T cells requires suitable target antigen selection, co-stimulatory signaling domain, and the ability of CAR T cells to traffic, persist, and retain antitumor function after adoptive transfer. There are several elements which can improve antitumor function of CAR T cells, including signaling, conditioning chemotherapy and irradiation, tumor burden of the disease, T-cell phenotype, and supplementary cytokine usage.

This review outlines four generations of CAR. The pre-clinical and clinical studies showed that this technique has a great potential for treatment of solid and hematological malignancies. The main purpose of the current review is to focus on the pre-clinical and clinical developments of CAR-based immunotherapy.

Keywords::

• cancer
• chimeric antigen receptor
• clinical studies
• immunotherapy
• T cells

Introduction

Immunotherapy can be considered as a promising approach for the treatment of various diseases such as cancer, autoimmune disorder, and allergic and hypersensitivity disorders (Burks et al. 2013, Perosa et al. 2005, Tomihara et al. 2013, Zou et al. 2007). Cancer immunotherapy is generally classified into two main classes, including active and passive methods. Some interventions which promote the immune system of the patient, for example, by vaccination or adjuvant therapy, actively promote antitumor effector mechanisms to improve cancer elimination. On the other hand, administration of specific monoclonal antibodies (mAbs) against different tumor antigens and the adoptive transfer of genetically-modified specific T cells are currently the most rapidly developing approaches for targeted therapy of cancer (Ahmed and Cheung 2013, Brayer and Pinilla-Ibarz 2013, Fenoglio et al. 2013, Kvistborg et al. 2012, Pranchevicius and Vieira 2013, Hosseini et al. 2015, Kazemi et al. 2015).

CD8⁺ T lymphocytes play a critical role in tumor eradication (Kvistborg et al. 2012). The main routes for T-cell therapy are described as follows: tumor-infiltrating lymphocyte (TIL) therapy, engineered TCR, and chimeric antigen receptor (CAR) (Hu et al. 2013, Jena et al. 2013, Kvistborg et al. 2012).

Recent efforts have been made to develop the recognition specificity of T cells, which redirects T cells to specific tumor-associated antigens (TAA). This technique uses both specific T and B cells by assembling an antigen-binding moiety, collected with an activating immune receptor (Hudecek et al. 2010).

The first chimeric TCR was created by joining a single-chain variable fragment (scFv) directly to the C region of either one of the α or β TCR chains, in 1989 (Gross et al. 1989).

CARs are proteins expressed on the surface of T and NK cells which contain extracellular binding domains, a hinge region which mediates the linkage of extracellular to transmembrane domains, a transmembrane domain, and an intracellular signaling domain (Chu et al. 2014, Till et al. 2012).

The use of CARs is associated with several advantages, such as: CAR protein recognizes TAA in a major histocompatibility complex (MHC)-independent manner and therefore, this feature enables the T cells to overcome the tumor^'s capacity to escape immune recognition by downregulation of MHC molecules on the cell surface (Sadelain et al. 2003, Schreiber et al. 2011, Seliger 2008). The use of CARs can be applicable to a broad range of patients irrespective of MHC type, because tumor targeting is not restricted to MHC molecules. CAR proteins are able to recognize any cell surface antigen including proteins, carbohydrates, and glycolipids. These receptors enable genetically modified T cells to react to a broad range of molecules in comparison with the native TCR (Hombach et al. 1997, Kailayangiri et al. 2012, Kochenderfer et al. 2009). The T-cell engineering for expression of CAR protein can redirect the specificity for the most of the T-cell populations, including CD4, CD8, naive, and memory or T-cell effectors. This is essentially important since naive and antigen-experienced T cells have different functional capabilities which can make them more or less effective for adoptive cell therapy (Jameson and Masopust 2009). CAR modification can involve one or more T- cell-activating signals which will be associated with the higher potential of T cells to proliferate, persist, and lyse target cells (Pulè et al. 2005). Finally, the ability to produce a large number of tumor-specific T cells in a moderately short period of time makes this technique more attractive to use in the clinical background (Hollyman et al. 2009, June 2007).

CAR-modified T-cell production

To produce CAR-modified T cells, CD4⁺ and CD8⁺ T lymphocytes are separated from the patient's blood through leukapheresis. Isolated T cells are then transfected with a viral vector such as lentivirus, containing genes for CAR, directed against the tumor target antigen. The virus binds to the T-cell membrane and fuses with the cell membrane, and reverse transcription, DNA integration, expression of CAR genes, and insertion of CAR protein into the cell membrane of T cells then occur.

In the next step, the genetically modified T cells are cultured with a sufficient amount of IL-2 in the laboratory. To reduce the level of white blood cells, the patient receives lymphodepleting conditioning chemotherapy. Subsequently, autologous CAR T cells are injected, and patients monitor for the treatment response and for the function of modified T cells (Jacobson and Ritz 2011, Porter et al. 2011).

CAR specificity

Antigen-targeting through CAR is a critical point in CAR design. Generally, the recognition of a target by CAR includes the usage of scFv that has been assembled from monoclonal antibodies (Bridgeman et al. 2010, Sadelain et al. 2003). Eradication of tumor by CAR requires recognition of antigen that only expresses on malignant cells and is pivotal for tumor cell survival. Since this antigen is critical for tumor survival, it cannot escape from immune response (Goldberger et al. 2008). Unfortunately, the expression of most antigens is not limited to tumor cells, resulting in undesirable effects on the normal tissue. CAR-modification of T cells can only recognize antigens expressed on the target cell surface (Stewart-Jones et al. 2009, Tassev et al. 2012). Moreover, tumor targeting by scFv is similar to antigen–antibody interaction. Various factors, including the affinity of binding domains, structure of epitopes, and amount of antigen which expresses on the tumor cells can affect the response of the modified T cells to target antigen (Hudecek et al. 2013).

Several studies have addressed various anchored ligands to the extracellular domain which contain scFv against certain molecules like CD19, CD20, HER2, FITC, etc., on tumor cell surface, or some agents like avidin that are designed, at the N-terminal of CAR, to recognize soluble biotin targets which are conjugated with monoclonal antibodies (Brentjens et al. 2011, Pinthus et al. 2003, Tamada et al. 2012, Urbanska et al. 2012, Wang et al. 2004).

Until now, the genetically engineered T cells have successfully been developed against various tumor antigens such as CD19 and CD20 for the treatment of B-cell malignancies (Brentjens et al. 2011, Wang et al. 2004), ERRB2 (estrogen-related receptor beta type 2) for the treatment of prostate and breast cancer (Pinthus et al. 2003, Teng et al. 2004), Carbonic anhydrase IX (CAIX) for the treatment of renal cell carcinoma (Lamers et al. 2006), prostate-specific membrane antigen (PSMA) for the treatment of prostate cancer (Gong et al. 1999, Maher et al. 2002), Lewis Y for the treatment of lung cancer (Mezzanzanica et al. 1997, Westwood et al. 2005), carcinoembryonic antigen for the treatment of colon cancer (Darcy et al. 1998, Haynes et al. 2002), TAG72 for the treatment of breast, colorectal, pancreatic cancer, etc. (Sharifzadeh et al. 2013), folate-binding protein and MUC16 for the treatment of ovarian cancer (Chekmasova et al. 2010, Hwu et al. 1993), and the diasialoganglioside GD2 for the treatment of neuroblastoma (Rossig et al. 2002). Studies have shown various CARs containing receptor polypeptide ligands linked to intracellular domains such as an integrin-binding peptide (anti-avb6), a polypeptide against vascular endothelial growth factor receptor (anti-VEGFR2), heregulin (anti-Her3/4 receptor), or interleukin (IL)-13 mutein (anti-IL13 receptor-a) (Kahlon et al. 2004, Muniappan et al. 2000, Niederman et al. 2002, Pameijer et al. 2007).

Recently, two general CAR systems have been described which include anti-biotin–CAR and an anti-fluorescein isothiocyanate (anti-FITC)–CAR. In the anti-biotin–CAR system, the extracellular domain of CAR attaches to avidin, which can bind to soluble biotin-conjugated therapeutic molecules (monoclonal antibodies, nucleic acid-based aptamers, or ligands). In other CAR systems that have been developed by Tamada et al., the binding region anchored to an anti-FITC scFv recognized tumor cells via binding to the FITC monoclonal antibody. Patients received these conjugated therapeutic molecules before infusion of CAR T cells, which labeled the tumor target. This technique allows the recognition of various tumor antigens which restrict the capacity of tumors to escape targeting by downregulating a single target antigen. Furthermore, genetically modified T cells can be useful for several types of tumors with only one CAR (Tamada et al. 2012, Urbanska et al. 2012). In another experiment, Ahmed et al. introduced a dual-specific CAR T cell—TanCAR—which contains two scFvs associated with a flexible linker on the surface of T cells against HER2 and CD19. This approach could enhance the anti-tumor function of CAR T cells, in vivo (Grada et al. 2013). One disadvantage of the CAR strategies using scFvs derived from murine antibodies is the appearance of both antibody- and cell- mediated responses against these foreign molecules. In order to avoid this adverse effect, we can use the humanized scFv domains or human antibodies to produce CARs (Hombach et al. 2000, Lamers et al. 1995) (Table I).

Table I. Targets for CAR-modified T cells.

Target	Malignancies	Selected references
CD19	B cell	Brentjens et al. (2011), Brentjens et al. (2007), Hollyman et al. (2009), Jena et al. (2013), Kochenderfer et al. (2014), Kowolik et al. (2006), Lee et al. (2011), Savoldo et al. (2011)
CD20	B cell	Till et al. (2008), Till et al. (2012), Wang et al. (2004)
k light chain	B cell	Vera et al. (2006)
CD30	Hodgkin's and non-Hodgkin's lymphomas	Di Stasi et al. (2009)
ROR1	B cell	Hudecek et al. (2010)
GD2	Neuroblastoma, Ewing's sarcoma	Kailayangiri et al. (2012), Louis et al. (2011), Pule et al. (2008)
CD171	Neuroblastoma	Park et al. (2007)
Folate receptor-α	Ovarian	Hwu et al. (1993)
VEGFR	Several	Niederman et al. (2002)
CEA	Several	Haynes et al. (2002), Hombach et al. (2000), Katz et al. (2015)
ERbB2	Prostate, breast	Pinthus et al. (2003), Teng et al. (2004)
ErbB3/4	Several	Muniappan et al. (2000)
CAIX	Renal cell carcinoma	Lamers et al. (2006), Lamers et al. (2011)
Lewis Y	Lung cancer	Mezzanzanica et al. (1997), Westwood et al. (2005)
PSMA	Prostate	Gong et al. (1999), Maher et al. (2002)
MART1	Melanoma	Rosenberg and Dudley (2009)
MAGE-A3	Melanoma	Morgan et al. (2013)
NKG2D ligands	Ovarian	Spear et al. (2012)
Mesothelin	Several	Carpenito et al. (2009)
IL-13 receptor α2	Several	Kahlon et al. (2004)
MUC16	Ovarian	Chekmasova et al. (2010), Koneru et al. (2015)
αvβ6 integrin	Several	Pameijer et al. (2007)
TAG72	Several	Sharifzadeh et al. (2013)
Universal	All	Niederman et al. (2002), Pameijer et al. (2007)

CAR T-cell Signaling

The optimal activation of T cells requires at least two distinct signals including interaction between the TCR–CD3 complexes with peptide presented by MHC molecules (signal 1), and recruitment of CD28 on the T-cell surface with co -stimulatory molecules CD80 (B7.1) or CD86 (B7.2) on the antigen-presenting cells such as dendritic cells (signal 2). Signal 1 and signal 2 are necessary for the optimal activation, effector function, proliferation, and survival of T cells. Delivery of signal 1 without signal 2 leads to the anergy and apoptosis of T cells and consequently leads to their hyporesponsiveness (Chen 2004, Chen and Flies 2013, Croft 2005, Ladygina et al. 2013, Näslund et al. 2013). The first generation of CARs contain a scFv joined to the intracellular signaling domain of CD3ζ (Eshhar et al. 1993) (Figure 1). The CD3ζ-derivedsignal provides the required ‘signal 1’ resulting in T-cell activation, modest IL-2 secretion, target cell lysis, and in vivo antitumor function (Eshhar et al. 1993, Haynes et al. 2001, Heuser et al. 2003). The activation of the first-generation CARs with only signal 1 results in T-cell anergy, unfavorable cytokine secretion, weak proliferation, and apoptosis of CAR T cells (Grossi et al. 1992, Lenschow et al. 1996). The second-generation CAR T cells are composed of scFv linked to co-stimulatory domains such as CD28 and CD3ζ which were generated to improve CAR activation signals and increase the proliferation and cytokine secretion (Figure 1). Interaction between CD28 and B7 family molecules such as B7.1 and B7.2, which are located on the surface of target cells, provides a second activation signal that enhances ‘signal 1’ from the TCR–CD3 complex (Bretscher 1999, Chambers and Allison 1997, Smith-Garvin et al. 2009). Delivery of signal 1 and signal 2 via CAR molecules in the second-generation CARs leads to T-cell proliferation, lysis of target cells, IL-2 secretion, and expression of some proteins such as anti-apoptotic protein Bcl-xL (Chambers and Allison 1997). Studies have shown that the second-generation CAR T cells against PSMA are much more effective than the first-generation in IL-2 production and proliferation (Maher et al. 2002).

Figure 1. Schematic structure of CAR-modified T cells in cancer immunotherapy.

Another study indicated that the treatment of established tumor using the second-generation CAR which contains CD28 increased the eradication efficacy in comparison with the first-generation CAR for the CD19 antigen (Brentjens et al. 2007). The use of CD28-containing second-generation CARs was also associated with their increased persistence (Savoldo et al. 2011). Generally, the enhanced cytokine production will increase the cell proliferation and upregulation of anti-apoptotic proteins such as Bcl-xL in the second- generation CARs containing CD28, in comparison with the first-generation CARs (Beecham et al. 2000, Haynes et al. 2002, Hombach et al. 2001a, 2001b, Kalos et al. 2011, Kowolik et al. 2006, Vera et al. 2006). Other B7 family members such as inducible co-stimulation (ICOS) and tumor necrosis factor receptor consisting of CD137–4-1BB or CD134–OX-40 can be replaced with CD28-containing second-generation CARs (Finney et al. 2004).

To improve CAR design and enhance the activation of the second-generation CARs, the ‘third-generation’ CARs were developed containing 3 co-stimulatory domains (Figure 1). The third-generation CARs provide signal 1, signal 2, and a supplementary co-stimulatory signal to boost T-cell activation signals, and include CD28, 4-1BB, and CD3ζ. Several studies have demonstrated the improvement of T-cell survival, cytokine production, and anti-tumor potential for genetically modified T cells that express the third-generation CAR (Carpenito et al. 2009, Chang et al. 2007, Wang et al. 2007, Zhao et al. 2009, Zhong et al. 2010). In another study, Zhong et al. (Zhong et al. 2010) have reported that the use of PMSA targeting CAR T cells was associated with the improved in vivo T-cell survival, enhanced cytokine production, improved PI3K/AKT pathway activation, enhanced tumor elimination, reduced T-cell apoptosis, and enhanced Bcl-xL expression. In recent studies, scientists have developed T cells redirected for universal cytokine-mediated killing or TRUCKs (fourth-generation CARs) (Figure 1). This type of CAR is able to secrete a pro-inflammatory cytokine such as IL-12 into the tumor microenvironment (Chmielewski et al. 2011, Pegram et al. 2012). Interaction between the fourth-generation CARs and their target leads to the activation of T cells and secretion of IL-12 by the CAR T cells in the target tissue. The accumulation of IL-12 is an important strategy for attraction of innate immune cells, mainly NK cells and macrophages, in a local tumor lesion (Kerkar et al. 2011, Kerkar et al. 2010). IL-12 is a heterodimeric protein which is physiologically secreted by macrophages and dendritic cells in response to intracellular pathogens (Trinchieri et al. 2003). Secretion of IL-12 by the fourth-generation CARs, leads to the infiltration of NK cells and macrophages to the tumor site, which consequently improve tumor eradication (Pritchard et al. 2005, Zhang et al. 2011). The optimal tumor eradication with CARs requires incorporation of different signaling domains which leads to complete activation of T cells to promote the secretion of certain cytokines.

CAR T-cell Trafficking

In order to achieve effective tumor eradication, genetically modified T cells should traffic to the tumor site(s) where CAR T cells will be able to target tumor cells. In fact, poor T-cell homing to tumor sites is the main reason for the reduced efficiency of adoptive immunotherapy based on CAR-engineered T cells (Parente-Pereira et al. 2011). Migration of genetically-modified T cells to the tumor site has been demonstrated through a dual bioluminescent imaging of CAR T cells in a murine tumor model (Santos et al. 2009). Accumulation of CAR T cells at the most site(s) of tumor, and persistence of CAR T cells, are the major principles for effective tumor eradication (Parente-Pereira et al. 2011). Studies have shown that genetically-modified T cells which express the second-generation CAR will accumulate in the disease sites (Brentjens et al. 2011, Kalos et al. 2011, Savoldo et al. 2011). Trafficking of CAR T cells to several sites including liver, lymph nodes, and bone marrow have been demonstrated by Brentjens et al. through the second-generation CAR against CD19 (Brentjens et al. 2011). In another investigation, Savoldo et al. (Savoldo et al. 2011) has reported the homing of genetically-modified T cells to a skin lesion after adoptive cell therapy in patients with non-Hodgkin's lymphoma (NHL). The second-generation CAR containing CD28 has shown effective trafficking to the tumor site compared to the first-generation CAR against CD19 (Savoldo et al. 2011). Overproduction of several molecules such as integrins and chemokines was associated with T-cell trafficking to tumor sites. Chemokines and their receptors play a critical role in the homing of CAR T cells to tumor sites (Di Stasi et al. 2009). Kershaw et al. (Kershaw et al. 2002) have demonstrated that the expression of CXCR2 and CCR4 in CAR-modified T cells will enhance T-cell trafficking to the disease site(s).

Persistence of CAR T cells

Poor survival of the infused T cells limits the efficiency of adoptive T-cell immunotherapy. Genetically modified T cells need to survive for a sufficient time period to achieve complete and successful tumor elimination. Robbins et al. have indicated that the efficacy of anti-cancer response is correlated with the persistence of infused T cells (Robbins et al. 2004). Brocker et al. have also reported poor survival of the first-generation CAR T cells containing CD3ε against lymphoma (Brocker 2000). In another study, Lamers et al. demonstrated poor persistence of the first-generation CAR T cells against carbonic anhydrase IX (CAIX) in metastatic renal cell carcinoma (Lamers et al. 2006). In order to overcome the limitations of the first-generation CARs, the second- and third-generation CARs have been developed to increase persistence of genetically modified T cells. Activation of CAR T cells with two or more co-stimulatory domains can enhance the persistence of infused T cells (Carpenito et al. 2009, Wang et al. 2007).

Moreover, lymphodepleting chemotherapy can also improve the persistence of CAR T cells through the modification of the host environment prior to the infusion of CAR T cells (Muranski et al. 2006, Rosenberg and Dudley 2009). Conditioning chemotherapy can remove suppressive cells such as regulatory T (Treg) cells and myeloid cells and enhance the persistence of CAR T cells.

Conditioning chemotherapy and low-dose irradiation can improve the survival of adoptively transferred T cells through a number of mechanisms, as demonstrated in preclinical models (Curiel et al. 2004, Durrant and Metzger 2010, Gattinoni et al. 2005). Brentjens et al. reported their observations for 10 patients with chronic lymphocytic leukemia (CLL) or acute lymphoblastic leukemia (ALL). In the first group, patients had been treated with the second- generation CAR containing CD28 anti-CD19 without conditioning chemotherapy, and showed undetectable persistence. Controversially, observations in the next group of patients that received prior lymphodepleting conditioning chemotherapy (cyclophosphamide), showed that genetically modified T cells can persist for 6 weeks after adoptive transfer in the bone marrow (Brentjens et al. 2011).

A chemotherapy regimen can reduce the patient's disease burden and enhance the persistence of the infused T cells. Correlation between patient's tumor burden and persistence of CAR T cells was demonstrated by Park et al. in patients with metastatic neuroblastoma (Park et al. 2007).

One approach to improve the persistence of genetically modified T cells can be the administration of IL-2. An enhanced persistence of the first-generation CAR anti- CD20 was also observed in patients with indolent NHL who were administered low dosages of subcutaneous IL-2. Moreover, production of some cytokines such as IL-2 and IL-15, as well as the overexpression of their receptors, improved the persistence of CAR T cells (Kershaw et al. 2006, Till et al. 2008).

Another approach to enhance in vivo T-cell persistence involves the application of virus-specific T cells which are genetically modified to express CAR. Murphy et al. have demonstrated the increased persistence of CAR-specific influenza virus T cells in mice with breast cancer (Murphy et al. 2007). In another experiment, improved persistence of CAR-specific Epstein Barr virus (EBV) T cells against GD2 was observed in patients with neuroblastoma (Pule et al. 2008). Increased persistence of CAR T cells in the circulation for up to 6 weeks was also observed in patients that were treated with EBV-specific T cells engineered with a CAR recognizing the diasialoganglioside GD2, representing an antigen expressed by neuroblastoma cells (Louis et al. 2011).

In conclusion, conditioning chemotherapy, cytokine supplementation, the patient's tumor burden, T-cell phenotype, co-stimulatory signaling, and possibly the use of viral-specific T cells are several factors which are able to improve the persistence of infused T cells.

CAR T cells and immunosuppressive tumor microenvironment

The maintenance of cytolytic function is an important element for successful tumor eradication by CAR-genetically modified T cells subsequent to an adoptive transfer. In addition, several mechanisms may inhibit the function of adoptively transferred CAR T cells. These mechanisms include suppressor cells such as Tregs cells, myeloid-derived suppressor cells (MDSCs), and immunosuppressive molecules/cytokines (e.g., TGF-β) (Rabinovich et al. 2007, Schreiber et al. 2011). Moreover, some genetic modifications mentioned above can also help genetically modified T cells to resist immunosuppressive factors in the tumor microenvironment. For instance, the incorporation of co-stimulatory molecules such as CD28 into CARs can assist T cells in being more resistant to Treg cells and TGFβ (Koehler et al. 2007, Lee et al. 2011, Loskog et al. 2006). Overexpression of AKT can also increase the resistance of CAR (Sun et al. 2010). A promising approach to inhibit the effects of TGFβ is expression of a decoy form of the TGFβ receptor, which can be achieved through genetic strategies in CAR T-cell design (Bollard et al. 2002, Gorelik and Flavell 2002). Activation of T cells can be suppressed by programmed cell death protein 1 (PD1) which is expressed on the surface of T cells. Interaction between PD1 and their ligands that are expressed in surface of cancerous cells can suppress the activation of CAR T cells in the tumor microenvironment. PD1-blocking by small interfering RNA (siRNA) or mAbs is a promising approach to reduce this form of immunosuppression (Borkner et al. 2010, Topalian et al. 2012). Expression of IL-12 by genetically modified CAR T cells is another strategy to enhance the resistance and may also change the tumor microenvironment. Genetic engineering of T cells to express IL12 can also enhance antitumor function and will improve the capacity to resist the immunosuppression exerted by Treg cells (Kerkar et al. 2010, Spear et al. 2012). Secretion of IL-12 by T cells leads to the formation of immunosupportive myeloid cells in tumors (Chmielewski et al. 2011, Pegram et al. 2012). Genetically modified T cells expressing a chimeric NKG2D were found to improve cytokine secretion, leading to a change in myeloid cells from an immunosuppressive to an immunostimulatory phenotype (Kerkar et al. 2011). Lee et al. have demonstrated that conditioning chemotherapy can also reduce the number of Tregs and improve the antitumor function of T cells (Lee et al. 2011). Other immunosuppressive factors such as galectins and amino acid metabolites may also suppress the function of CAR T cells, and future approaches that aim to block these factors may generate more effective tumor-specific T cells.

Toxicity of CAR T cells

Unwanted toxicity is a major problem that can limit CAR T-cell-based immunotherapy. Genetically modified T cells have to recognize antigens that only express on the surface of tumors. Morgan et al. treated a colon cancer patient with the third-generation CAR against ERBB2. The patient that received lymphocyte-depleting chemotherapy died 5 days after infusion of CAR T cells. Interaction between genetically modified CAR T cells and ERBB2 led to the release of pro-inflammatory cytokines. Hence, pulmonary toxicity, multi-organ failure, and death of the patient were observed after infusion of T cells (Morgan et al. 2010). In another clinical study, 11 renal-cell carcinoma patients were treated with the first-generation CAR-engineered T cells recognizing the CAIX, representing an antigen expressed by renal cells. Targeting CAIX in bile duct epithelial cells led to serious liver toxicity in 5 patients after the infusion of CAR T cells (Lamers et al. 2006, Lamers et al. 2011). In another clinical trial, neurotoxicity was reported in 2 of 8 patients treated with genetically modified T cells to express a MAGE-A3-specific CAR. It was concluded that neurotoxicity was the result of cross-reactivity between MAGE-12, which is expressed on some neurons, and MAGE-A3 that is only expressed on tumors (Morgan et al. 2013). These studies demonstrated the need for tumor-targeting antigens that are only expressed on cancer cells. Some important characteristics of the fourth-generation of CAR T cells have been summarized in Table II.

Table II. Different characteristics of the four generations of CAR T cells.

	First-generation	Second-generation	Third-generation	Fourth-generation
Intracellular domain	CD3ζ	CD3ζ/CD28 or ICOS, CD137, CD134	CD3ζ/CD28/4-1BB	CD3ζ/CD28 or ICOS, CD137, CD134
Signaling	Poor proliferation	Increased proliferation	Increased proliferation	Increased proliferation
Trafficking	Pelvic mass	Liver, lymph nodes, peripheral blood, and bone marrow	Pulmonary infiltrate	–
Persistence	Poor survival	Increased persistence	Increased persistence	Increased persistence
Function	Poor cytotoxic activity	Increased cytotoxic function	Increased cytotoxic function	Increased cytotoxic function
Toxicity	Liver toxicity	B-cell aplasia, decreased plasma cell number, and hypogammaglobulinemia	Patient died 5 days after infusion	–
Cytokine secretion	IL-2	IL-2	IL-2	IL-2/IL-12

Clinical trials with different generations of CAR T cells

Several studies have evaluated the antitumor effectiveness of the first-generation CAR-modified T cells. No significant results were found in the early clinical trials using the first-generation CAR T cells. Furthermore, there are certain differences between these clinical trials, including the targeted antigen selection, gene transfer method, the technique of ex vivo expansion, injection of exogenous IL-2, cell dosage, and the use of conditioning therapy.

Lamers and colleagues (Lamers et al. 2006, Lamers et al. 2011) described their experience on 11 patients with metastatic cell renal carcinoma treated with the first-generation anti-CAIX CAR which contained a CD3ζ endodomain with or without subcutaneous IL-2. Liver toxicity was observed after infusion of CAR T cells in 5 out of 8 patients. This side-effect is due to the low-level expression of CAIX on bile duct epithelial cells which are targeted through CAR T cells. Moreover, no liver toxicity was observed in 3 patients who received anti-CAIX mAb (to block CAIX on normal liver tissue) prior to the infusion of CAR T cells. The persistence of infused T cells was limited for all patients, and no objective clinical responses were observed. Importantly, humoral and cellular anti-CAR responses were developed in all patients, which can describe the limited persistence of infused T cells.

Kershaw et al. (Kershaw et al. 2006) reported their studies on 14 patients with ovarian cancer treated with CAR-modified T cell anti-α-folate receptor in the presence and absence of subcutaneous IL-2. Infused T cells had limited persistence, and no objective clinical responses were observed.

Park et al. (Park et al. 2007) tried to treat 6 neuroblastoma patients with CAR-modified T cells against L1-cell adhesion molecule (L1-CAM; CD171); however, no objective clinical responses were observed. However, increased T-cell survival was observed in a patient with a low disease burden in comparison with the patients with larger tumor burdens.

Till et al. (Till et al. 2008) treated 7 patients with relapsed or refractory indolent NHL with T-cell targeting CD20. In the first group, no objective clinical responses were observed, and the persistence of infused CAR T cells was also limited. However, patients in the second group who received subcutaneous IL-2, the survival of modified CAR T cells was improved significantly. One patient in the second group demonstrated a limited remission, and one patient showed a reduction in the metabolic activity of tumor.

Moreover, Jensen et al. (Jensen et al. 2010) investigated the efficacy of CAR-modified T cells targeting CD19 and CD20 in the treatment of four patients with recurrent NHL. Two patients with B-cell lymphoma were treated with modified T cells expressing a CD20-specific CAR. Treatment of two follicular lymphoma patients with CAR-modified T cells expressing a CD19-specific CAR and low dosage of subcutaneous IL-2 showed the limited persistence of CAR T cells, which was associated with no clinical responses.

The initial clinical trial based on the application of the second-generation CAR-modified T cells developed in 2010, also focused on B-cell malignancy.

Kochenderfer et al. (Kochenderfer et al. 2010) described their experience regarding the 1 patient with NHL treated with a CD28-containing second-generation CAR anti-CD19 antigen. The patient received lymphodepleting conditioning chemotherapy that included cyclophosphamide and fludarabine prior to the infusion of CAR T cells and IL-2. A partial remission of the patient's lymphoma was also observed for up to 32 weeks. The persistence of CAR T cells was shown by quantitative polymerase chain reaction (Q-PCR) in the peripheral blood for 27 weeks, and no significant toxicity was found.

Savoldo et al. (Savoldo et al. 2011) tried to treat 6 relapsed or refractory NHL patients with genetically modified anti-CD19CAR T cells. Their patients were treated with a mixture of engineered T cells included CD3ζ-containing first- generation, and CD28-, CD3ζ-containing second-generation CAR anti-CD19, without lymphodepletion. They showed that the second-generation CARs had higher persistence, expansion, and traffic in comparison to the first-generation CARs.

Brentjens and colleagues (Brentjens et al. 2011) treated 9 leukemia patients with CD28-containing second-generation CAR T cells against CD19 antigen. In this study, 8 CLL patients and 1 ALL patient were treated with CAR T cells with and without conditioning chemotherapy. Enhanced persistence of the modified CAR T cells was observed in 6 patients who received lymphodepleting conditioning chemotherapy with cyclophosphamide before T-cell infusion. One patient died 2 days after modified CAR T-cell infusion. Out of 4 patients with bulky CLL, 3 demonstrated a response to treatment with conditioning chemotherapy and cyclophosphamide followed by infusion of modified CAR T cells. The one ALL patient treated with CAR T cells showed a B-cell aplasia. Modified CAR T cells showed efficient traffic to the sites of disease and retained ex vivo antitumor function after adoptive transfer.

Kalos et al. (Kalos et al. 2011) described the adoptive transfer of 4-1BB-containing second-generation CAR-modified T cells directed against CD19 antigen in 3 patients with CLL who received conditioning chemotherapy prior to the T-cell infusion. Modified T cells showed traffic to site of the disease and were detected at high levels for up to 6 months post-infusion in peripheral blood and bone marrow. Responses to treatment and tumor lysis syndrome were observed in 3 patients (2 complete remissions and 1 partial remission) which correlated with increased CAR T-cell number and pro-inflammatory cytokine levels.

Kochenderfer et al. (Kochenderfer et al. 2012) treated 8 advanced B-cell malignancy patients with second-generation CAR directed against CD19 with low dosage of IL-2. Six treated patients (6 of 7 evaluable) had a clinical remission. Trafficking of modified T cells to bone marrow and complete remission was demonstrated in few patients.

Saar Gill et al. (Gill et al. 2014) have introduced CD123, a transmembrane chain of IL-3R, as a worthy target for AML-directed CAR T-cell therapy. As it is over-expressed in most of the AML cells in vivo, eradication of primary AML with the help of human CD123-CD28-CD3ζ T cells (CAR T123) in immunodeficient mice suggested that CAR T123 is one of the best choices for AML therapy, in this study.

Kochenderfer and colleagues (Kochenderfer et al. 2014) evaluated the efficacy of CD28- and CD3ζ-containing second-generation CAR-engineered T cell anti-CD19 in 15 patients with advanced B-cell malignancies following conditioning chemotherapy. In this study, 9 patients with diffuse large B-cell lymphoma (DLBCL) and 6 patients with indolent B-cell malignancies were treated. Of the 7 patients with DLBCL, 4 achieved complete remissions (CRs), 2 achieved partial remissions (PRs), and 1 had stable disease (SD) after T-cell therapy. Of the 6 evaluable patients with indolent B-cell malignancies, 4 achieved CR and 2 achieved PR. One patient died after infusion of CAR T cells, and one patient was withdrawn to follow-up. Several manifestations including hypotension, fever, delirium, and other neurologic toxicities were observed in some patients after treatment; these toxicities were eliminated within 3 weeks after T-cell infusion.

Katz et al. (Katz et al. 2015) treated 6 patients with liver metastases (LM) with second-generation of CAR-T cell (CD28 and CD3ζ) anti-CEA with and without IL-2 injection. Of the 6 evaluable patients, 5 died due to disease progression. One patient had stable disease at 23 months after treatment. Elevated serum IFN-γ levels and improved CEA responses were demonstrated in second-generation groups who received systemic IL-2.

Morgan et al. treated 1 colon cancer patient with CD28- and 4-1BB-containing third-generation CAR targeting ERBB2 with lymphocyte-depleting chemotherapy including cyclophosphamide and fludarabine (Morgan et al. 2013). The patient died 5 days after the infusion of modified T cells. Tumor-targeting T cells recognized ERBB2 on the lung epithelium, which ultimately leads to death of the patient.

Till et al. reported their experience regarding 4 patients with mantle cell or follicular non-Hodgkin's lymphoma. In their investigation, patients were treated with CD28- and 4-1BB-containing third-generation CAR targeting CD20. Two patients remained progression-free for 12 and 24 months. An objective partial remission was found in the third patient 12 months after infusions (Till et al. 2012).

Recently, fourth-generation CARs have been introduced. Some studies have revealed the anti-tumor function of CAR-modified T cells in vitro and in several animal models.

Koneru et al. (Koneru et al. 2015) have designed the fourth-generation of CAR (4H11-28z/IL-12) against MUC-16^ecto antigen that overexpresses on the surface of ovarian tumor cells and is the residue of MUC-16 after CA-125 cleavage. This study indicates that co-expression of IL-12 and 4H11-28z CAR T cells’ increased proliferation and IFN-γ secretion when compared with 4H11-28z CAR T cells in the SKOV3 cell line. The antitumor efficacy of IL-12-secreting 4H11-28z CAR T cells is also examined in SCID-Beige mice with human ovarian cancer xenografts, through enhanced survival, prolonged persistence of T cells, higher systemic IFN-γ level, and in turn, complete eradication of tumors.

Although these experiments have shown the potential of T cells engineered with the fourth-generation CARs for the treatment of patients, further investigations in various clinical studies (see Table III) (Chmielewski et al. 2011, Kerkar et al. 2011, Kerkar et al. 2010, Pegram et al. 2012) are needed.

Table III. Clinical trials with different types of CAR-modified T cells.

Author	Number of patients	Tumor target	Antigen	Lymphodepletion	Cytokine IL-2	CAR generation	Effect
Lamers et al.	7	Metastatic cell renal carcinoma	CAIX	No	Yes/No	First	No
Kershaw et al.	14	Ovarian cancer	α-folate receptor	No	Yes/No	First	No
Park et al.	6	Neuroblastoma	L1-CAM (CD171)	Yes	No	First	No
Till et al.	7	NHL	CD20	Yes	Yes/No	First	2 CR, 1 PR, 4 SD
Jensen et al.	4	NHL	CD19/CD20	No	Yes/No	First	No
Kochenderfer et al.	1	NHL	CD19	Yes	Yes	Second	PR
Savoldo et al.	6	NHL	CD19	No	No	First and second	Improves expansion and persistence of 2nd generation as compared with 1st generation
Brentjens et al.	9	CLL and ALL	CD19	Yes/No	No	Second	NOR,PD,SD
Kalos et al.	3	CLL	CD19	Yes	No	Second	2 CR, 1 PR
Kochenderfer et al.	8	Advanced B-cell malignancy	CD19	Yes	Yes	Second	6 PR, 1 CR, 1 SD 1 DOD
Kochenderfer	15	Advanced B-cell malignancy	CD19	Yes	No	Second	4 PR, 1 SD 8 CR
Katz	6	Liver metastases	CEA	No	Yes/No	Second	5 DOD, 1 SD
Morgan et al.	1	Colon cancer	ERBB2	Yes	No	Third	DOD
Till et al.	4	NHL	CD20	Yes	Yes	Third	1 PR, 2 SD 1 CR

CAIX: Carbonic anhydrase-IX; L1-CAM: L1-cell adhesion molecule; NHL: Non-Hodgkin's lymphoma; ERBB2: Estrogen-related receptor beta type 2; CLL: Chronic lymphocytic leukemia; ALL: Acute lymphocytic leukemia;

CR: Complete response; DOD: Died of disease; PD: Progressive disease; PR: Partial response; SD: stable disease; NOR: No objective response; PD: Progressive disease.

Conclusion& Future perspective

CD4⁺and CD8⁺ T lymphocytes play an important role in the treatment of cancer. CD8⁺ CTLs are able to eradicate tumor cells efficiently, whereas CD4⁺ T lymphocytes can improve the capacity of antigen-presenting cells (APC) and antitumor function of CTLs through the enhanced secretion of some cytokines. CAR-based immunotherapy was developed for use both in cellular and humoral arms of the immune system for effective tumor eradication. The first generation of CARs includes a binding domain which specifically targets a tumor antigen, and a lymphocyte-activating intracellular domain. Clinical trials utilizing the first-generation CAR-engineered T cells have failed to exhibit significant clinical benefits. The next generations of CARs were developed to improve the persistence and anti-tumor function of the first generation of CARs. Genetically modified T cells which express CARs are capable of eradicating malignant cells in vitro, and the infusion of engineered T cells results in remission of tumors in various animal models. Recent clinical trials which are based on the infusion of CAR-engineered T cells into tumor patients have indicated that this approach is feasible and safe. Despite some promising results which were derived from clinical studies, some side-effects were reported which can limit the efficacy of CAR T cells. Additionally, unwanted toxicity is a major problem which was observed in several clinical studies. Another problem in immunotherapeutic approaches that are based on CAR-modified T cells is their restricted availability of proper surface tumor-associated antigen (TAAs). Finally, the poor persistence of infused CAR-genetically modified T cells results in restricted antitumor responses. Second, the third and fourth generations of CAR-engrafted T cells were also developed to overcome this obstacle. These generations of CARs exhibited increased proliferation rates and increased cytokine secretion, and are more persistent in the immunosuppressive mechanism of the tumor microenvironment. It is expected that in the near future, many efforts will focus particularly on the development of the third and fourth generations of CARs to acquire more successful results for the treatment of cancer.

Disclosures of financial & competing interests

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript.