by Silvia Hnatova
Gene-edited cell therapies showed potential as a treatment for many diseases, especially for blood-based cancer patients who do not respond to any other treatment. Gene-edited T cells reprogrammed to express CAR transgene (CAR T cells) can serve as the ‘living drugs’ for leukaemia and lymphoma (Bailey and Maus, 2019).
CAR T cells are reprogrammed to target tumour cells expressing a specific antigen (Gross et al., 1989). CAR T therapy emerged as a potent anticancer therapy achieving up to 94% remission rate in clinical trials (Seimetz et al., 2019). CAR T cells can survive for a long time upon being engrafted in the patient and can be trafficked to all body tissues.
In the present, CAR T cells are being reprogrammed to express small molecules and drugs that can serve as an addition to the existing therapies. In this article, we will review the manufacturing challenges for gene-edited cell therapies, considerations for allogeneic CAR-T cells, industry developments, best practices and opportunities for innovation.
CAR-T cell therapies, although promising, are recent and have not yet been standardised. Efficacy and reproducibility of the therapies remains to be determined and manufacturing costs can be high.
All manufacturing steps for CAR-T cell therapies include isolation of patient’s white blood cells, activation of T cells and transduction with CAR transgene. The T cells are then expanded and following quality control, they are ready for injection into the patient. The patient has to undergo a chemotherapy that depletes lymphocytes prior to the injection (Vormittag et al., 2018).
Such approach was adopted by Novartis in the development of Kymriah. Kymriah (tisagenlecleucel) is composed of autologous T cells that are reprogrammed for expressing a CAR that recognizes CD19+ tumour cells using a lentiviral vector. The manufacturing showed ~9% failure rate and the main concern was to address replication competent lentivirus presence during FDA review. This was addressed using improved vector design, achieving a limited copy number in a cell (Seimetz et al., 2019). Kymriah underwent several clinical trials in order to establish the manufacturing failure rate, efficacy, safety and remission rate. In a multi-centre, phase II study determining efficacy and safety of Kymriah in adults with relapsed or refractory diffuse large B-cell lymphoma, 90% of patients survived beyond 1 year, out of which 76% showed a response, with 43% demonstrating a complete response (NCT02445248).
A similar strategy was implemented by KITE Pharma’s Yescarta for non-Hodgkin lymphoma. Yescarta (axicabtagene ciloleucel). Yescarta, similar to Kymriah, is formed of autologous T cells reprogrammed by a retroviral vector to express a CAR directed against CD19+ cells. The main manufacturing challenge was a demonstrated efficacy, that was addressed using Phase I, II and III clinical studies. In ZUMA-1, an open-label multi-centre Phase I/II study for B-cell non-Hodgkin lymphoma, the overall remission rate based on 111 patients was 72% and it took on average 0.9 months for the patients to respond to the treatment (NCT02348216).
Currently, the autologous CAR-T cell manufacturing is labour intensive, hard to scale, costly and prone to failure. The manufacturing costs can be significant in order to generate sufficient number of cells required (1.2 x 106 to 6 x 108 CAR-positive viable t cells). The costs of Kymriah vary from $475,000 for leukaemia and $373,000 for lymphoma per treatment. The cost of Yescarta is estimated at $378,000 per patient. The high costs arise from achieving a therapeutically relevant number of anti-tumour T cells during reprogramming. T-cell activation is pivotal for ensuring therapeutic efficiency and the purification steps can result in significant cell loss. Retroviral and lentiviral vectors for gene delivery are extremely expensive, due to intense manufacturing labour and requirements for cleanliness and safety. Stable packaging lines could help to reduce the manufacturing costs (Roddie et al., 2019).
In order to improve the manufacturing of CAR T cell therapies, we need to improve the robustness and scalability of these approaches, and to simplify and standardize the manufacturing processes (Kaiser et al., 2015). Use of enclosed systems prevents cross-contamination and has previously been implemented by Tumaini et al. (2013). The input material needs to be standardized. Some authors use a subset of T cells for gene editing (van Loenen et al., 2014). Targeted use of naïve, central memory or memory T cells has shown advantages over undefined cell populations (Gattinoni et al., 2011). Reagents and cell culture media need to be standardized and serum free. Vera et al. (2011) optimized the process of T cell culturing by using gas-permeable culture devices, reducing complexity and costs of manufacture.
Last, leukaemia patients often display lower number or under-functioning of T cells (Graham et al., 2018) creating a further obstacle for use of autologous CAR-T cells.
Allogeneic CAR-T cells therapy, using reprogrammed T cells from healthy donors in multiple patients, is economically more viable than the labour-intense autologous CAR-T cell therapy. Allogeneic CAR-T therapy could become a ‘universal’ CAR-T therapy. So far, only a handful of allogeneic blood therapies have been approved, using umbilical cord blood cells (Seimetz et al., 2019). In a clinical study assessing the efficacy of allogeneic CAR T cells, eight out of twenty patients achieved remission (Brudno et al., 2016). The most successful remission was achieved in patients with acute lymphoblastic leukaemia, followed by chronic lymphocytic leukaemia and lymphoma.
The most significant challenge in developing allogeneic products in the graft rejection by the host, resulting in graft versus host disease (GVHD). GVHD results as a consequence of a mismatch via the T cell αβ receptor (TCRαβ) and the results can be fatal. Although haematopoietic stem cell transplant donors are HLA matched, this does not fully prevent GVHD.
The manufacturing strategies for preventing GVHD could include gene editing in order to prevent GVHD. TCRαβ could be knocked out in order to prevent recognition. This could be done via CRISPR-Cas9 or any other gene editing techniques (Graham et al., 2018; Singh et al., 2018). Georgiadis et al. (2018) used CRISPR-Cas9 in order to achieve TCRαβ cleavage, resulting in potent anti-leukemic effects in immunodeficient mice. Eyquem et al. (2017) directed the CAR targeting to the TCRαβ locus, achieving both CAR expression and preventing GVHD in one step.
An alternative to using CAR T cells could also be engineered natural killer cells (NK) or gamma-delta (γδ) T cells. γδ T cells are a subset of T cells expressing NK receptor NKG2D, specific to tumour cells (Patel et al., 2019). Most studies carried out previously were in the preclinical stage and currently, with limited studies at clinical stage. γδ T cells were expanded in vivo by Bennouna et al. (2008) and infused into ten patients with metastatic renal cell carcinoma, showing toleration by the patients. Immunotherapy for patients with hepatocellular carcinoma was trialled (NCT00562666) in a Phase I study that was terminated. γδ T cells remain a powerful alternative to CAR T cell therapies, but safety and efficacy of a therapy using γδ T cells remains unknown and yet to be determined (Lawand et al., 2017; Nussbaumer and Koslowski, 2019).
Manufacturing considerations in using γδ T cells would have to address these, although γδ T cells are less likely to cause GVHD and display a reduced toxicity (Fisher and Anderson, 2018).
Cell therapies using NK cells are also promising, due to their reduced risk of triggering GVHD and transitory activity in vivo (Patel et al., 2019). Both autologous and allogeneic approaches are being developed for NK cell therapies, and the tolerance of autologous NK cells was good in digestive cancer patients, without adverse effects, but also without any clinical response (Sakamoto et al., 2015). Potential for clinical therapy is currently hindered by lack of knowledge about the specificity and efficacy of the treatment using NK cells (Bachanova and Miller, 2015). Phase II clinical trial using allogeneic NK cells for treatment of leukaemia, lymphoma and myelodysplastic syndrome is underway (NCT02727803).
The biggest challenges in CAR T cell therapies remain variability and barriers to scalability. Current industry developments aim to address these challenges, via improving the manufacturing process and the production line. BioSpherix use Xvivo modular laminar flow system in order to enable each step of the manufacturing process to be carried out in separate physical space. CliniMACS Prodigy employ device-based manufacturing in order to reduce the number of manual handling steps. In contrast to the linear flow, dysfunction of one unit in device-based manufacturing does not impact the rest of the products and could be rapidly exchanged.
Scalability of autologous CAR T cell therapies can be overcome by the rapid advancements of allogeneic CAR T products. The main barrier towards improving allogeneic CAR T cells is the potential non-specificity of CRISPR-Cas9 and other gene editing techniques, creating potential for off-target effects. The main focus therefore will be on reduction of off-target gene cleavage. Another attractive possibility could be creating a bank of CAR T cell lines with different HLA types (Graham et al., 2018). CAR T cells have previously been generated using induced pluripotent stem cells (iPSCs) and demonstrated anti-tumour activity in a xenograft model (Themeli et al., 2013).
The product process of CAR T cell therapies is strongly regulated by the authorities, including the European regulation 1394/2007 and the Guidance for Human Somatic Cell Therapy and Gene Therapy by the FDA. A massive quality control (QC) pipeline is employed in CAR T cell therapy manufacturing process, in contrast to standard pharmaceuticals. This is to prevent contamination and cross-contamination. QC is integral to maintaining safety and efficacy in CAR T cell therapies, including ensuring activated T cells that express the target CAR domain.
The leukapheresis process has to optimize for collection of T cells. Standard leukapheresis takes place before lymphodepleting chemotherapy is initialized if absolute T cell count is below 200-300 in a patient. Leukapheresis is generally safe and the efficacy of the procedure could be enhanced using automated cell-washing devices, e.g. Haemonetics or ClinicMACS (Levine, 2015). Anticoagulants (citrate) are using during the process to prevent clumping. Recently, gradient centrifugation was implemented for the filtering step removing red blood cells and platelets (CT01886976, NCT01864902; Vormittag et al., 2017).
Optimal T cell activation can be ensured using stimulation via anti-CD3/anti-CD28 monoclonal antibodies, enabling selection of activated T cells (Levine et al., 1997). Magnetic beads coated with anti-CD3/CD28 antibodies are used for this process, allowing efficient separation using strong electromagnets (Kalos et al., 2011). Anti-CD3/CD28 antibody selection of activated T cells remains the dominant method for T cell activation, used in conjunction with paramagnetic beads. Such approach is being implement by 626 out of 952 available products, including Yescarta and Kymriah (Vormittag et al., 2017).
Gene delivery using lentiviruses or retroviruses is labour and cost demanding but achieves 4-70% transduction efficiency. Viruses remain the preferred method of choice in gene delivery, used in 919 out of 978 available products (Vormittag et al., 2017).
Expansion of activated CAR T cells can be achieved using T-flasks or rocking motion bioreactors.
Allogeneic CAR T cell production includes steps to ensure HLA compatibility. First clinical trials testing the safety of gene-edited CAR T cells are being conducted, evaluating the safety and efficacy of such methods (Mehta and Rezvani, 2018; Exley et al., 2000).
The future of CAR T cells therapies lies in deploying automation and technological innovations in order to improve cost-effectiveness, scalability and to reduce variability, whilst conforming with all necessary QC steps.
The main opportunities for automation in CAR T cell therapy development lie in using automated cell washing and selection devices, like CliniMACS Plus for targeted cell harvesting, selection and activation of T cells (Miltenyi Biotec). Closed or semi-closed systems aid the manufacturing processes whilst preventing cross-contamination, like those adapted by Tumaini et al. (2013).
Main challenges in technological innovation will be in addressing the gene editing safety and efficacy of allogeneic CAR T cells, that would facilitate a streamlined manufacturing process. Optimized use of gene editing techniques, CRISPR-Cas9 or zinc finger nucleases, would help to prevent GVHD whilst facilitating the use of allogeneic CAR T cells, by enhancing their function and ensuring immune compatibility (Rafiq et al., 2020). No gene editing technique can ensure 100% TCRαβ knockout and a safe level of contamination by TCRαβ-positive allogeneic donor CAR T cells has to be established.
iPSC-derived T cells could overcome this challenge, by ensuring that the MHC restriction is determined by host MHC molecules, reducing GVHD risk (Rafiq et al., 2020). Such technique has previously been trialled with promising results (Themeli et al., 2013; Van Der Stegen et al., 2017; Clarke et al., 2018).
Enhancing the development of defined allogeneic CAR T cells would enable the development of a universal CAR T cell therapy, further lowering down manufacturing costs and speeding up automation.
Future development of gene editing techniques not relying on lentiviral or retroviral constructs would help to lower down costs and make the treatment more accessible and affordable (Rafiq et al., 2020).
CAR T cell activity could also be improved using drugs. This approach is being explored in a single-centre Phase I/II clinical trial using pembrolizumab following anti-CD19 CAR T cell therapy in order to reactivate the exhausted CAR T cells (NCT02650999). This followed a case study of a patient from ZUMA-1 with refractory diffuse large B-cell lymphoma, who did not respond to CAR T cell therapy and experienced multiple adverse effects.
Follow-up of the patient revealed rapid programmed cell death, agreeing with retrospective immunohistochemical analyses showing PD-L1 expression pre-treatment (Hill et al., 2019). Following administration of nivolumab, anti-PD-L1 agent, the patient experienced tumour regression and rapid recovery, including increase in cytokines related to proliferation, immune modulation, inflammation and immune effector function. This case study led the ground of the on-going clinical trial NCT02650999, assessing the safety and efficacy of checkpoint inhibitors in CAR T cell re-activation following exhaustion.
In summary, both autologous and allogeneic CAR T cell therapies are undergoing intense development and innovation in order to streamline manufacturing processes, reduce costs and establish safety and efficacy. Allogeneic CAR T cell therapies have the potential to become a future universal cancer therapy, enabling re-using the CAR T cells for different patients and so lowering the costs of CAR T therapy.
