Redirected T cells in cancer therapy

Abstract

Genetic engineering of T lymphocytes is an appealing strategy to confer and enhance new antitumor specificities to generate effective anticancer cell products for adoptive immunotherapy. The two main approaches are based either on transgenic tumor-antigen specific T cell receptors (TCR) or chimeric antigen receptors (CAR). Initial clinical trials reported important results against selected diseases, along with relevant warnings. Ongoing research challenges are directed toward a widespread application of this approach enhancing the range of possible target antigens, antitumor activity and safety, but also addressing logistic issues regarding cost/effectiveness, up-scaled/automated production and compliance with regulations.

Keywords:

• adoptive immunotherapy
• CAR
• engineered T cells
• TCR

1. Introduction

The concept of genetically redirected adoptive cell therapy (ACT) is based on the impressive progresses of molecular technologies that allow to engineer T cells with transgenic, clinical graded, tumor-specific receptors.[1]

The main form of ACT has been developed in the setting of metastatic melanoma, with ex vivo expansion and clinical re-infusion of tumor infiltrating lymphocytes (TILs). This strategy provided important proof of concept along with relevant initial clinical results; important issues have, however, prevented its application to other solid tumors and the procedure remained confined in very few centers worldwide.[2] Generation of TILs is in fact a very personalized approach, requiring availability of surgically removed tumor tissue carrying relevant infiltrating lymphocytes, dedicated Good Manufacturing Procedures–compliant facilities, and laborious ex vivo expansion processing. The recently impressive clinical successes with immune-checkpoint programmed death (PD-1) and cytotoxic T lymphocyte antigen 4 (CTLA-4) inhibitors base their activity on the re-activation of endogenous TILs, which helped to overcome mechanisms of a tumor-induced adaptive tolerance.[3]

In both these strategies, the key is to potentiate or unleash a pre-existing antitumor immune response. Genetic engineering of T lymphocytes exploits a different, and in a way complementary, approach as it may confer new tumor specificity to conventional circulating lymphocytes, generating an artificial immune response that would be lacking otherwise in spontaneous conditions. The two main strategies to redirect T lymphocytes are based on the genetic transferring of tumor-specific T-cell receptors (TCRs) or chimeric antigen receptors (CARs).[4]

1.1 TCR engineered T cells

In this approach, T lymphocytes are transduced with transgenes encoding for α and β TCR chains specific for a given tumor-associated antigen (TAA). This approach produced important initial clinical data, recently confirmed and updated, against myeloma but also solid tumors like melanoma and synovial sarcomas.[5,6] In this case, the target was a HLAA2 restricted NY-ESO1 antigen, that is currently one of the most promising and safer target for TCR engineered ACT. Besides the case of virus-induced tumors, the family of non-mutated cancer testis antigens currently represents the principal and more tested targets for TCR-transfer strategies. An interesting future perspective will be to explore the generation of artificial TCRs against mutated neoantigens that were recently described as the main targets of the immune response unleashed by immune-checkpoint modulators.[4] Side effects, potentially severe, have to be considered with TCR transfer approach, either due to cross-recognition of non-tumoral targets, generation of unexpected antigen specificities or autoimmune reactions. The most striking example came from a trial where artificial anti-MAGE3-TCR, directed against melanoma and myelomas, induced two lethal events following abnormal recognition of cardiac protein Titin. Of note are also cases of severe neurotoxicity consequent to the attack of cross-reactive antigens in the brain.[7] One of the most reliable approach currently explored to reduce these kinds of risks is the introduction of an inducible suicide gene, capable to switch off engineered lymphocytes in case of undesired effects.[8,9] It is interesting to note, however, that most of the severe adverse reactions described in the trials occurred very rapidly questioning the complete efficacy of an inducible suicide system. The new encoded α and β TCR chains may give rise to heterodimers with endogenous TCR chains, determining a loss of efficiency or generation of new unplanned specificities. Many research efforts are exploring creative and effective strategies to obviate this problem, like the possibility to knock down endogenous TCR genes, modify the protein structure to prevent mixed-dimerization or generate single-chain TCRs.[10–13] Researchers are also dedicated to enhance the affinity of transgenic TCRs, as the naturally occurring T cells are usually characterized by low-affinity recognition of self-antigens. The issue may be different for mutated tumor neoantigens where a spontaneous high-affinity TCR recognition may spontaneously exist and justified by the “non-self” nature of the target.[4] Artificial maturation of TCR affinity is, however, a delicate challenge as it may eventually impair the function of T cells and increase the risk of toxicities.[14,15]

TCR engineered lymphocytes have the potential to recognize both intracellular and extracellular antigens, broadening the pathway of potential eligible tumor targets. As a possible downside, their activity always requires human leukocyte antigen (HLA) presentation of TAA and it is consequently restricted to specific HLA haplotypes, potentially limiting the number of patients that might benefit from these approaches.

1.2 CAR engineered T cells

The second main strategy to redirect T lymphocytes is based on their engineering with tumor-specific CARs. CARs are made by an extracellular antibody-derived single chain variable fragment (scFv), capable of recognizing TAA, fused into a chimeric receptor with TCR-derived signaling domains. Interaction of scFV with a given TAA induces physiologic T-cell activation and consequent tumor killing.[1] Differently from TCRs, CARs do not need HLA-mediated presentation of their targets, with implications for a broaden benefit of this strategy to all patients regardless of their HLA haplotype. On the other side, the spectrum of possible targets is limited to extracellular antigens, excluding, for example, cancer testis family TAA or neoantigens derived by cancer genome instability. Three main generations of CARs can be distinguished, where the last two are endowed with progressively enhanced activity, survival and persistence due to the addition or association of co-stimulatory signals (e.g. CD28; CD137; OX40).[1] While initial clinical studies with first-generation CARs were disappointing, the improved co-stimulatory signaling of second generation brought impressive clinical results in the field of hematologic B-cell malignancies including chronic lymphocytic leukemia and acute lymphocytic leukemia.[1,16–18] Reported responses rates, including complete remissions, were around or higher than 50% and were shown to correlate with in vivo persistence and proliferation of CAR-T cells. These positive data have not been replicated so far against solid tumors (e.g. renal, colorectal and ovarian carcinomas) even if encouraging data derived from initial studies against neuroblastoma.[19] Along with the importance of in vivo persistence and necessity of appropriate co-stimulatory signaling, conditioning lymphodepleting regimens seem to be a key factor for effective ACT against solid tumors. As for TCR-transfer strategies, toxicity is an open issue even for CAR-redirected T lymphocytes. Cytokine release syndrome is the most common side effect, present in the majority of patients who respond to CAR-T cells. Its intensity may be varied up to very severe or even fatal events and research efforts are directed into definition of predictive biomarkers that could anticipate the clinical diagnosis. Great hope is hold by research of cytokine-blockade agents that could effectively contrast cytokine release syndromes (e.g. anti-IL6 receptor antibody or agents blocking IL1-cascade).[1,4] Even with CAR-T cells, important toxicities may be derived by undesired recognition of the target antigen in healthy tissues. B-cell aplasia is common following treatment with CARs, anti-CD19 or anti-CD20 antigens, expressed also on normal B cells; liver toxicity was reported with CAR-T lymphocytes directed against carbonic anhydrase IX for renal cancer and a fatal event was registered with anti-human pidermal rowth factor receptor 2 (HER2) CAR lymphocytes against colorectal cancer, likely due to HER2 expression in lungs and cytokine storm.[1,4,19]

2. Expert opinion

The present and next future of genetically redirected ACT will be hopefully characterized by a trend toward its widespread application as cancer therapy, progressively increasing the number of targetable disease settings. In this direction the next years will see researchers facing important scientific but also logistic issues. Scientific challenges will be to enhance the activity, applicability and safety of ACT, while logistic issues will regard cost-effectiveness, up-scaled/automated production of cell products and compliance with regulations.

2.1 Efficacy and safety issues

It emerged from initial clinical trials that in vivo persistence and proliferation of engineered lymphocytes is a key element for their efficacy. Potential improvements might then derive from a planned upfront selection of different T-cell subsets to be engineered and adoptively infused, preserving naïve and memory compartments that could sustain a prolonged in vivo response. Similarly, a different approach under investigation includes engineering lymphocytes other than αβ T cells, for example, γδ T lymphocytes or even natural killer (NK) cells. Possible advantages would be to confer TAA specificity to lymphocytes already capable of innate antitumor activity plus overcoming the risk of TCR heterodimerization with endogenous chains. A complementary emerging approach is that of engineering T cells with activating innate immune receptors, like NKG2D, derived from γδ or NK lymphocytes. Promising preclinical reports demonstrated that a chimeric receptor, with NKG2D fused to TCR-derived signaling domains, can mediate T-cell activation and cytotoxicity upon engagement with its ligands on tumor cells.[20] Main ligands of NKG2D are stress-inducible molecules (e.g. MIC A/B or ULBPs) that are widely expressed on several tumor histotypes. Data from more studies in the next future will allow better understanding of potentialities and possible toxicities of this approach.

The limiting issue of immunogenicity of transgene receptors and tolerance versus self-antigens may be contrasted by generation of fully human or humanized CARs and obtainment of TCRs from transgenic mice or HLA mismatched donors. While these alternative sources would broaden the available repertoire and affinity of antitumor receptors, researches are ongoing to limit the consequent enhanced risk for off-target toxicities. More general inflammatory and autoimmune-mediated effects may also be derived by effective ACT, similarly to what observed with immune-checkpoint inhibitory antibodies. This type of effects may become more evident as ACT will hopefully widen its clinic applications, including patients with underlying or clinically silent autoimmune problems.

An appealing line of research that might enhance the spectrum of targetable TAA in the next future might be engineering TCRs specific for tumor neoantigens, recently uncovered as final targets of immune responses unleashed by immune-checkpoint inhibitors.

Actually the potential synergism with immune-checkpoint modulators is itself an appealing perspective, supported by preclinical models and with initial clinical studies ongoing. It may be realized by current monoclonal antibodies on the market or similar effects may be searched by further engineering ACT products with genome editing techniques that could knock down inhibitory molecules like PD-1 or CTLA-4.

Ongoing and future improvements in the efficacy and safety of redirected ACT will require a parallel development of reliable predictive preclinical models, capable of exploring new biotechnology strategies but also display unexpected toxicities before translation into clinical studies.

2.2 Logistic issues

Besides scientific challenges, the widespread application of genetically redirected ACT faces logistic issues involving the manufacturing processes, economic sustainability and regulatory elements.

Most of research efforts described above tend toward realization of universal ACT products, with off-shelf features that might help the widespread application of these strategies. The manufacturing processes will need to be up-scaled and hopefully automated; the recent interest of several pharmaceutical companies in the field of engineered ACT may provide the requested energy and tools in this direction, probably selecting upfront those approaches with a demonstrated efficacy that may justify the economic effort. Costs are in fact a major and delicate issue, in need of great attention and compromise by administrative, academic and regulatory organs involved. Clinical application of genetically redirected ACT may follow the path shown by current immune-checkpoint modulators, a breakthrough in the treatment of several diseases but also sustainability challenges for our health systems.

Declaration for interest

The author declares that no competing financial interests exist.

Financial and competing interests disclosure

D. Sangiolo has received funding from Associazione Italiana per la Ricerca sul Cancro (AIRC) via grants MFAG 2015 N. 15731; FPRC ONLUS 5 per mille Ministero della Salute 2012; Ministry of Health GR-2011-02349197; University of Torino Fondi Ricerca Locale 2013. He has no other 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. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Acknowledgements

The author was supported in part by grants from “Associazione Italiana Ricerca sul Cancro” (AIRC) MFAG 2014 N.15731, Ricerca Finalizzata-Giovani Ricercatori (GR-2011-02349197), Fondo Ricerca Locale 2013, Università degli Studi di Torino, FPRC-ONLUS 5 per mille Ministero della Salute anno 2012.