T cell receptor-engineered T cells to treat solid tumors: T cell processing toward optimal T cell fitness.

Therapy with autologous T cells that have been gene-engineered to express chimeric antigen receptors (CAR) or T cell receptors (TCR) provides a feasible and broadly applicable treatment for cancer patients. In a clinical study in advanced renal cell carcinoma (RCC) patients with CAR T cells specific for carbonic anhydrase IX (CAIX), we observed toxicities that (most likely) indicated in vivo function of CAR T cells as well as low T cell persistence and clinical response rates. The latter observations were confirmed by later clinical trials in other solid tumor types and other gene-modified T cells. To improve the efficacy of T cell therapy, we have redefined in vitro conditions to generate T cells with young phenotype, a key correlate with clinical outcome. For their impact on gene-modified T cell phenotype and function, we have tested various anti-CD3/CD28 mAb-based T cell activation and expansion conditions as well as several cytokines prior to and/or after gene transfer using two different receptors: CAIX CAR and MAGE-C2(ALK)/HLA-A2 TCR. In a total set of 16 healthy donors, we observed that T cell activation with soluble anti-CD3/CD28 mAbs in the presence of both IL15 and IL21 prior to TCR gene transfer resulted in enhanced proportions of gene-modified T cells with a preferred in vitro phenotype and better function. T cells generated according to these processing methods demonstrated enhanced binding of pMHC, and an enhanced proportion of CD8+, CD27+, CD62L+, CD45RA+T cells. These new conditions will be translated into a GMP protocol in preparation of a clinical adoptive therapy trial to treat patients with MAGE-C2-positive tumors.

Long-term stability of T-cell activation and transduction components critical to the processing of clinical batches of gene-engineered T cells.

BACKGROUND AIMS: The generation of gene-modified T cells for clinical adoptive T-cell therapy is challenged by the potential instability and concomitant high financial costs of critical T-cell activation and transduction components. As part of a clinical trial to treat patients with metastatic renal cell cancer with autologous T cells engineered with a chimeric antigen receptor (CAR) recognizing carboxy-anhydrase-IX (CAIX), we evaluated functional stability of the retroviral vector, T-cell activation agent Orthoclone OKT3 (Janssen-Cilag, Beerse, Belgium) monoclonal antibody (mAb) and the transduction promoting agent RetroNectin (Takara, Otsu, Japan). METHODS: Carboxy-anhydrase-IX chimeric antigen receptor retrovirus-containing culture supernatants (RTVsups) were generated from two packaging cell lines, Phoenix-Ampho (BioReliance, Sterling, UK) and PG13, and stored at -80°C over 10 years and 14 years. For Orthoclone OKT3 and RetroNectin, aliquots for single use were prepared and stored at -80°C. Transduction efficiencies of both batches of RTVsups were analyzed using the same lots of cryopreserved donor peripheral blood mononuclear cells, Orthoclone OKT3 and RetroNectin over time. RESULTS: We revisit here an earlier report on the long-term functional stability of the RTVsup, observed to be 9 years, and demonstrate that this stability is at least 14 years. Also, we now demonstrate that Orthoclone OKT3 and RetroNectin are functionally stable for periods of at least 6 years and 10 years. CONCLUSIONS: High-cost critical components for adoptive T-cell therapy can be preserved for ≥10 years when prepared in aliquots for single use and stored at -80°C. These findings may significantly facilitate, and decrease the financial risks of, clinical application of gene-modified T cells in multicenter studies.

Treatment of metastatic renal cell carcinoma with CAIX CAR-engineered T cells: clinical evaluation and management of on-target toxicity.

Autologous T cells genetically modified to express a chimeric antibody receptor (CAR) against carboxy-anhydrase-IX (CAIX) were administered to 12 patients with CAIX-expressing metastatic renal cell carcinoma (RCC). Patients were treated in three cohorts with a maximum of 10 infusions of a total of 0.2 to 2.1 × 10(9) CAR T cells. CTC grade 2-4 liver enzyme disturbances occurred at the lowest CAR T cell doses, necessitating cessation of treatment in four out of eight patients in cohorts 1 and 2. Examination of liver biopsies revealed CAIX expression on bile duct epithelium with infiltration of T cells, including CAR T cells. Subsequently four patients were pre-treated with CAIX monoclonal antibody (mAb) G250 to prevent CAR-specific toxicity and showed no liver toxicities and indications for enhanced peripheral T cell persistence. No clinical responses were recorded. This report shows that CAIX-targeting CAR T cells exerted antigen-specific effects in vivo and induced liver toxicity at the lowest dose of 0.2 × 10(9) T cells applied, illustrating the potency of receptor-modified T cells. We provide in-patient proof that the observed "on-target" toxicity is antigen-directed and can be prevented by blocking antigenic sites in off-tumor organs and allowing higher T cell doses.

Immune responses to transgene and retroviral vector in patients treated with ex vivo-engineered T cells.

Adoptive transfer of immune effector cells that are gene modified by retroviral transduction to express tumor-specific receptors constitutes an attractive approach to treat cancer. In patients with metastatic renal cell carcinoma, we performed a study with autologous T cells genetically retargeted with a chimeric antibody receptor (CAR) directed toward carbonic anhydrase IX (CAIX), an antigen highly expressed in renal cell carcinoma. In the majority of patients, we observed distinct humoral and/or cellular anti-CAIX-CAR T-cell immune responses in combination with a limited peripheral persistence of transferred CAIX-CAR T cells in the majority of patients. Humoral immune responses were anti-idiotypic in nature and neutralized CAIX-CAR-mediated T-cell function. Cellular anti-CAIX-CAR immune responses were directed to the complementarity-determining and framework regions of the CAR variable domains. In addition, 2 patients developed immunity directed against presumed retroviral vector epitopes. Here, we document the novel feature that therapeutic cells, which were ex vivo engineered by means of transduction with a minimal γ-retroviral vector, do express immunogenic vector-encoded epitopes, which might compromise persistence of these cells. These observations may constitute a critical concern for clinical ex vivo γ-retroviral gene transduction in general and CAR-retargeted T-cell therapy in particular, and underscore the need to attenuate the immunogenicity of both transgene and vector.

Gibbon ape leukemia virus poorly replicates in primary human T lymphocytes: implications for safety testing of primary human T lymphocytes transduced with GALV-pseudotyped vectors.

The Food and Drug Administration/Center for Biologics Evaluation and Research has defined that for retroviral gene therapy, the vector-producing cell, the vector preparation, and the ex vivo gene-transduced cells have to be tested for absence of replication-competent retrovirus (RCR) if the transduced cells are cultured for >4 days. We assessed the sensitivity of the "extended PG4(S+L-) assay" to detect gibbon ape leukemia virus (GALV) RCR, and applied this assay to measure GALV RCR spread in retrovirally transduced T cells. To this end, T cells were expanded for 12 days after transduction with a GALV-envelope pseudotyped retroviral vector expressing single chain variable fragment (anticarbonic anhydrase IX) in presence or absence of GALV RCR. Results showed that: (1) the "extended PG4(S+L-) assay" detects 1 focus-forming unit (ffu) GALV RCR and thus is applicable and sufficiently sensitive to screen human T-cell cultures for absence of infectious GALV RCR; (2) although GALV RCR infect human T cells, it very poorly replicate in T cells; (3) GALV RCR, when present at low levels immediately upon gene transduction (ie, 100 ffu/20x10 T cells in 100 mL), did not spread during a 12-day T-cell culture at clinical scale. Our observation that GALV RCR poorly spreads in primary human T-cell cultures questions the relevance of testing T-cell transductants for RCR on top of testing the vector-producing cells and the clinical vector batch for RCR and warrants evaluation of the current policy for safety testing of ex vivo retrovirally transduced T lymphocytes for GALV RCR.