IND-Enabling Studies for a Clinical Trial to Genetically Program a Persistent Cancer-Targeted Immune System

Adoptive transfer of T cells genetically modified to express tumor-specific T-cell receptors (TCRs) has shown remarkable antitumor efficacy in several clinical trials (1–7). However, durable long-term clinical responses have been more challenging to maintain (4, 6), whereas engraftment and persistence of the modified T cells have been associated with better antitumor responses (4, 5, 8).

To improve the persistence of the TCR-expressing T cells, we propose to use genetically modified hematopoietic stem cells (HSCs) as a source for constant endogenous renewal of TCR-engineered T cells. The use of genetically modified HSCs was first shown to be efficacious in primary immunodeficiencies, where the transplantation of genetically modified HSCs demonstrated long-term correction of the disease (9–11). Regarding its use in cancer immunotherapy, it has been previously shown in murine and humanized murine models that transplantation of HSCs engineered to express a tumor-specific TCR results in the output of functional T cells with the defined specificity after proper thymic selection. The newly generated T cells showed antitumor activity and were able to differentiate into memory T cells after antigen stimulation (12–18). The endogenously produced TCR-engineered T cells demonstrated allelic exclusion, a process in which the TCR transgenic chains inhibit the expression of the endogenous TCR chains, thereby avoiding TCR mispairing and consequently reducing the potential toxicity of the therapy (19).

To test the approach of TCR genetic engineering of HSCs in the clinic, we designed a clinical trial of double cell therapy coadministering HSCs and T cells both genetically modified to express an NY-ESO-1 TCR. We propose to use 2 TCR-engineered cell therapies because we anticipate that the TCR-engineered HSCs will endogenously differentiate into fully active mature T cells with a long delay in their appearance in the periphery, as new T cells reaching peripheral circulation must undergo a thymic selection process which takes 1 to 3 months (20, 21). In this clinical trial, TCR-engineered mature lymphocytes and TCR-engineered HSCs will be coadministered to patients with NY-ESO-1–positive advanced cancers after a myelo- and lymphodepleting conditioning regimen. We hypothesize that the TCR-engineered mature lymphocytes will expand in vivo and provide the first wave of transient antitumor activity. This will serve as a bridge until the second wave of TCR-transgenic cells arising from the bone marrow–engrafted gene-modified HSCs has gone through the T-lymphocyte maturation process and partially repopulated peripheral tissues (Supplementary Fig. S1).

We propose to use the NY-ESO-1 TCR for this clinical trial because the toxicities observed in adoptive T-cell therapy clinical trials using this TCR were not dose-limiting and unrelated to the TCR recognizing its cognate antigen, which in adult life is presented by the cancer cells and absent in most normal somatic cells (3, 5, 6). NY-ESO-1 is a cancer-testis antigen expressed at high homogeneous levels in synovial sarcoma, with lower frequency of expression in melanoma and multiple myeloma, and at even lower levels in other cancers (22–24). The NY-ESO-1 TCR used in this work recognizes the SLLMWITQC NY-ESO-1 epitope in the context of HLA-A*0201 and contains a 2 amino acid substitution in the third complementary determining region that increases its affinity for peptide–MHC complexes (25).

Herein, we report on preclinical studies performed to demonstrate the safety and feasibility of the approach. These studies led to the investigational new drug (IND) application approval for a new clinical trial that is being conducted at UCLA (NCT03240861). Based on the previous experience with adoptive T-cell transfer using the NY-ESO-1 TCR, the major safety concerns were focused on the transplantation with lentivirally modified HSCs and the untested effects of the coadministration of the HSCs and T cells both expressing the same NY-ESO-1 TCR. The preclinical studies performed to justify the safety of our approach adressed (i) the ability of the genetically modified and ex vivo–cultured lineage-negative (Lin⁻) cells to differentiate into all blood lineages; (ii) the effect of the coadministration of genetically modified Lin⁻ cells and T cells on the engraftment and proliferation of the T cells and the engraftment and differentiation of the HSCs; (iii) the immunogenicity of the NY-ESO-1 TCR and the sr39TK transgenes caused by their expression in the various hematopoietic cells; (iv) the potential genotoxicity induced by the lentiviral vector integration in the Lin⁻ cells; and (v) the ability of the sr39TK suicide gene to ablate transduced stem cell progeny in case of a serious adverse event. To test the feasibility of the clinical trial, we validated the HSC cell product manufacturing process and demonstrated the stability and functionality of the final product. The data presented on the article are focused on the approach-specific studies. Detailed information on the viral vector manufacturing, lot release criteria, and optimization of the gene-modified cell manufacturing can be found in the Supplementary Data Section.

All animal experiments were performed under the UCLA Animal Research Committee protocol #2013-095 that was previously approved by the Institutional Animal Care and Use Committee. Lin⁻ cells were isolated from the bone marrow of HLA-A2/K^b transgenic mice (26), prestimulated with cytokines, transduced with previously validated lentiviral vectors LV-NY-ESO-1 TCR/sr39TK, LV-NY-ESO-1 TCR, or LV-empty, and cultured in vitro. For the coadministration experiments, T cells were isolated from spleens from HLA-A2/K^b transgenic mice, activated with CD3/28 and IL2 stimulation, transduced with the RV-NY-ESO-1 TCR, and expanded in vitro. The methods and reagents to manufacture and validate both cell products are described in the Supplementary Materials. Transduced Lin⁻ cells alone or together with transduced T cells were administered systemically to 8- to 12-week-old HLA-A2/K^b mice that had received total body irradiation (TBI) with 900 cGy the day prior to cell administration. A detailed description of the cell testing performed prior to administration and the cell doses can be found in the Supplementary Material and Methods. Five days or 3 months after the bone marrow transplantation (BMT), mice were euthanized, and hematology, serum chemistry, complete histopathology survey, lentivirus and retrovirus vector copy number (VCN) in the blood, spleen, and bone marrow, and TCR expression and phenotype analysis in the bone marrow and spleen were performed. A detailed explanation of these procedures can be found in the Supplementary Materials and Methods.

Lin⁻ cells from C57BL/6J (Jackson Laboratories) mice were transduced with the LV-NY-ESO-1 TCR/sr39TK vector or the retrovirus SF91-eGFP-WPRE, expanded, and seeded at limiting dilution. The growth of transformed clones was measured. This procedure is detailed in the Supplementary Materials and Methods.

The CD34⁺ cell population was enriched using the CliniMACSCD34 reagent system, prestimulated overnight, and transduced with LV-NYESO TCR/sr39TK vector supernatant at a final multiplicity of infection (MOI) of 50 (2 transduction cycles of 25 MOI each). After 18 ± 6 hours from the first transduction, cells were harvested, formulated in cryopreservation solution, and cryopreserved using a controlled-rate freezer. The cell product was tested for sterility, endotoxin levels, mycoplasma contamination, clonogenic potential, and transduction efficiency. Long-term stability of the cryopreserved cell product was evaluated 30 ± 7, 90 ± 7, and 180 ± 10 days after cryopreservation. The post-thaw stability of the cell product was tested over a 48-hour period. See Supplementary Materials for further details.

It is described in the Supplementary Materials.

The lentiviral vector used in this clinical trial, LV-NY-ESO-1 TCR/sr39TK (RRL-MSCV-optNYESO-optsr39TK-WPRE, Fig. 1A), expresses the alpha and beta chains of the NY-ESO-1 TCR and the sr39TK suicide/PET reporter gene linked by a 2A self-cleavage peptide, and has been previously described in preclinical studies (17). The expression of the transgenes is driven by the murine stem cell virus (MSCV) promoter derived from the MSCV retrovirus long terminal repeat (LTR). Both genes had been codon-optimized to improve expression in mammalian cells. This vector was produced at Good Manufacturing Practice (GMP) grade at the Indiana University Vector Production Facility (IU VPF). To have enough vector for all of the IND-enabling studies and product manufacture optimization, a first manufacture of a 20-liter preparation without full lot release testing was labeled as GMP-comparable lot and was used for the majority of the preclinical studies. A second GMP production of 60 liters that was tested and met all of the lot release criteria (Supplementary Tables S1 and S2) was used for preclinical studies of cell product manufacturing and functionality, and will be used for clinical product manufacturing. To characterize the functionality of the lentiviral vector, we measured (i) the expression of the NY-ESO-1 TCR on the cell surface by dextramer staining in transduced human PBMCs (Fig. 1B) and (ii) the sr39TK suicide gene function in transduced human CD34⁺ peripheral blood stem cells (PBSC). Following treatment of human CD34⁺ PBSC with increasing doses of ganciclovir in vitro, the number of NY-ESO-1 TCR-positive cells (V_β13⁺ cells) decreased, whereas the number of TCR-negative cells remained constant (Fig. 1C).

To genetically engineer the peripheral T cells for short-term antitumor activity while awaiting the emergence of TCR transgenic T cells from the transplanted HSCs, a preparation of 18 liters of GMP-grade retroviral vector expressing the NY-ESO-1 TCR chains, RV-NY-ESO-1 TCR (MSGV1-A2ab-1G4A-Ly3H10), was made at the IU VPF. The vector met all of the release criteria (Supplementary Table S3). This retroviral vector is a splicing-optimized MSCV-derived vector expressing the alpha and the beta chains of the NY-ESO-1 TCR linked by a P2A self-cleavage peptide (a kind gift from Dr. Steven A. Rosenberg and Dr. Paul Robbins, Fig. 1D). For the studies in mice presented in this article, we pseudotyped the clinical vector with an ecotropic envelope, and a 3-liter GMP-comparable batch was produced at the IU VPF. For the GMP-comparable vector, only partial lot release testing was performed (Supplementary Table S4). To demonstrate the functionality of the ecotropic RV-NY-ESO-1 TCR retroviral vector, we transduced murine T cells from HLA-A2/K^b mice and measured the expression of extracellular and surface NY-ESO-1 TCR staining for the beta chain of the TCR (V_β13 staining, Fig. 1E).

To formally test if the coadministration of both cell products is safe, it was necessary to set up all of the conditions required to conduct a complete toxicology study under Good Laboratory Practice (GLP) conditions in compliance with 21 CFR Part 58. As GLP studies require exclusively dedicated facilities, a new laboratory space was set up with certified and calibrated equipment, with continuous automated monitoring of all the equipment and a tracking system in place to assure that the raw materials, test articles, and specimens were adequately preserved. One hundred twelve standard operating procedures (SOPs) were generated that captured all of the intended procedures for the study, including the specifics of the scientific protocols, facility and equipment management, personnel training, and GLP procedures. All of the personnel involved in the study were qualified to perform their assignments, and curricula vitae, job descriptions, and training records were maintained to document their proficiency. The study required 32 personnel with job descriptions previously defined and recorded (see organizational chart in Supplementary Fig. S2). All personnel were required to undergo documented training to conduct GLP studies and to perform the procedures for which they were responsible. In addition, all personnel were required to demonstrate proficiency using the pertinent SOPs according to their job descriptions. Prior to study initiation, the identity, purity, and potency of the test articles were verified, acceptance criteria for the manufactured cells were established, and the study protocol was approved by the study director, the test facility management, and the sponsor. This documentation was included in the IND later submitted to the FDA. Before the beginning of the study, a master schedule was generated and distributed to all personnel involved in the study. Sterility testing, blood and serum analysis, and histopathology analysis were outsourced to GLP-compliant contract research organizations. The study included a full histopathologic analysis performed by an American College of Veterinarian Pathologist board-certified veterinary pathologist. All raw data were signed by the study personnel performing the activity, reviewed by the study director or designee, and archived. Any deviation from the SOPs, study protocol, or GLP regulations was summarized in a deviation report, and its impact on the study data addressed by the study director and archived with the study documentation. All of the data generated in the study were included in a final study report that was submitted to the FDA in support of the IND application. All of the study raw data, specimens, and test articles will be archived for 5 years after submission of the IND to the FDA. Finally, an independent Quality Assurance Unit audited all of the experimental portions of the study, the facilities, the SOPs, the study records, and the GLP report and generated a Quality Assurance statement that was included in the final study report.

Fully immunocompetent HLA-A2/K^b transgenic mice express a chimeric MHC class I complex that includes the human HLA-A2.1 α1 and α2 domains, allowing their cells to present the same epitopes as HLA-A2.1 subjects, and maintain the murine α3 domain, permitting murine CD8 coreceptor engagement (26). To manufacture the T-cell products, T cells from HLA-A2/K^b mice were activated, transduced with the RV-NY-ESO-1 TCR vector, and expanded ex vivo. For the HSC product, Lin⁻ progenitor stem cells were purified from the bone marrow of HLA-A2/K^b mice, prestimulated with cytokines, and transduced with the LV-NY-ESO-1 TCR/sr39TK vector. Both cell types were coadministered by tail vein injection to myelodepleted HLA-A2/K^b mice that had previously received a lethal dose of TBI (900 cGy). We performed 12 pilot studies to establish the best conditions for BMT in this model. The main conclusions from these pilot studies were: (i) TBI at 900 cGy is myelodepleting and lethal without BMT (Supplementary Fig. S3A); (ii) specific animal care procedures had an important role, in particular having long acclimatization periods in the animal facility before treatment, preventive antibiotic treatment, and supplementary hydration and nutritional gel diet during the post-irradiation recovery; (iii) culture of Lin⁻ cells for a short time, 48 hours, was important for HSCs engraftment, progeny persistence, and survival (Supplementary Fig. S3B); (iv) doses from 0.5 to 2 × 10⁶ short culture (48 hours) Lin⁻ cells showed no differences in engraftment and survival (Supplementary Fig. S3C); and (v) increasing the retrovirus VCN in T cells was not toxic (Supplementary Fig. S3D).

The formal GLP-compliant toxicity study design is described in detail in the study protocol (Supplementary Materials Section 4). This study included a total of 110 mice divided into 5 groups (Supplementary Table S5). The number of mice had been prospectively defined by power calculation based on the assumptions made after interpreting the pilot studies (see study protocol in the Supplementary Materials Section 4). TCR-transduced Lin⁻ cells and TCR-transduced T cells (cohort E) were coadministered to myelodepleted HLA-A2/K^b mice and compared with control groups of mice receiving mock-transduced Lin⁻ cells and mock-transduced T cells (cohort B), transduced Lin⁻ cells and mock-transduced T cells (cohort C), mock-transduced Lin⁻ cells and transduced T cells (cohort D), or untreated control mice (cohort A). Different cohorts of mice were euthanized at 5 days or 3 months after BMT to allow analyses of toxicity without interfering with the overall survival endpoint (Supplementary Table S5). Prior to the study start, we established acceptance criteria for the cell manufacture and planned to not proceed with any batches of cells that did not meet the criteria. The criteria included cell purity, viability, VCN, mycoplasma, sterility, endotoxin, cell mix homogeneity, and stability during administration (Supplementary Table S6).

After BMT and T-cell administration, mice were followed for 3 months. No differences in overall survival were observed among cohorts (log-rank test P value 0.48, Fig. 2A). A statistically significant decrease in total body weight was observed in all cohorts receiving TBI and BMT compared with the untreated mice, but it was independent of the expression of the NY-ESO-1 TCR and sr39TK transgenes (Fig. 2B; Supplementary Fig. S4).

Overall, the clinical observations, gross pathology, and histopathology from the 10 animals that died or were euthanized before the end of the study did not identify toxicities that could be attributed to a specific treatment leading to increased mortality of a specific cohort. Lack of engraftment together with intestine necrosis was observed in the 3 mice that died within the first 2 weeks after BMT: one in cohort C (transduced Lin⁻ cells and mock T cells), one in cohort D (mock Lin⁻ cells and transduced T cells), and one in cohort E (transduced Lin⁻ cells and transduced T cells). Necrotic intestines and cecal torsions associated with irradiation sickness were also observed in 2 additional mice from cohort C (transduced Lin⁻ cells and mock T cells) that died at days 64 and 69 after BMT. Urethral obstruction and urinary bladder distension were observed in 1 mouse from cohort C that died at day 52 after BMT (transduced Lin⁻ cells and mock T cells) and one from cohort D that died at day 21 after BMT (mock Lin⁻ cells and transduced T cells). In addition, 1 mouse from cohort A (untreated) died at day 71 presenting an atypical enlargement of the heart. No significant findings were identified in 1 mouse from cohort B (mock Lin⁻ cells and mock T cells) that died at day 35 and 1 from cohort E (transduced Lin⁻ cells and transduced T cells) that died at day 85. We performed a full necropsy for all of the mice euthanized at the study end, 3 months after treatment. The necropsy included organ weights (Fig. 2C) and histopathology of all protocol-specified collected tissues (Supplementary Table S7). A marked reduction in testes weight was observed in all cohorts receiving TBI, consistent with germ cell ablation identified by histopathology analysis. No other consistent microscopic findings associated with a treatment or with irradiation were identified.

Complete blood cell counts at day 5 after BMT showed that mice receiving TBI were lymphodepeleted and presented with statistically significant decreases in red blood cell and platelet counts (Fig. 2D and E; Supplementary Fig. S4C). TBI caused a sharp reduction in splenocyte and bone marrow cell counts, with a decrease in spleen weight that was recovered 3 months after BMT (Supplementary Fig. S5). At 3 months after BMT, mice from all cohorts had reconstituted all blood cell compartments. All white blood cell lineages had recovered completely (Fig. 2D and E), and red blood cell parameters (hemoglobin and hematocrit) were slightly increased in irradiated mice, whereas platelets were not completely recovered relative to untreated mice (Fig. 2D; Supplementary Fig. S4). Serum chemistry analysis at 3 months after BMT showed a slight decrease in albumin levels in the coadministration group (cohort E). Liver function tests and creatinine kinase activities were highly variable and in general increased in all cohorts receiving BMT (Fig. 2F; Supplementary Fig. S6). However, there was no evidence of hepatocyte or muscle damage in tissue sections or the clinical observations. Altogether, these data demonstrate that the coadministration of Lin⁻ cells and T cells expressing an NY-ESO-1 TCR is safe and does not increase toxicity beyond those related to the BMT procedure.

To assess the engraftment and persistence of the T cells and Lin⁻ cells expressing the NY-ESO-1 TCR, we quantified the VCN using droplet digital PCR with primers specific for the inserted retroviral and lentiviral vectors. We demonstrated the presence of transduced T cells at 5 days and 3 months after treatment in blood, spleen, and bone marrow cells with no statistical difference between the groups receiving mock-transduced Lin⁻ cells or NY-ESO-1/sr39TK-transduced Lin⁻ cells together with the transduced T cells (Fig. 3A and B). Similarly, we quantified the engraftment of the transduced Lin⁻ cells and the persistence of their progeny 5 days and 3 months after treatment in blood, spleen, and bone marrow. No statistically significant differences were identified among the cohorts receiving transduced Lin⁻ cells together with mock-transduced T cells or with NY-ESO-1 TCR-transduced T cells (Fig. 3C and D).

Finally, to study the differentiation of the Lin⁻ cells, we performed an extensive phenotype characterization of the bone marrow cells and splenocytes 3 months after BMT. In the bone marrow, there were no differences in the frequencies of HSCs and precursor populations, Lin⁻, LSK (Lin⁻, ScaI⁺, cKit⁺), and HSCs (LSK, CD150⁺ CD48⁻), among the cohorts receiving BMT, although the frequency of Lin⁻ and HSCs populations was decreased compared with untreated mice (Fig. 4A; Supplementary Fig. S7). Intracellular expression of the NY-ESO-1 TCR by clonotypic V_β13 staining was observed in mice receiving BMT with TCR-transduced Lin⁻ cells independently of the T cells received (Fig. 4B; Supplementary Fig. S7). Interestingly, there were no differences in the frequencies of the Lin⁻, LSK, and HSCs cells between the total bone marrow and the V_β13⁺ population (Fig. 4C; Supplementary Fig. S7). Similarly, phenotypic characterization in the spleen showed no differences in NKT cells, CD4⁺ T cells, CD8⁺ T cells, B cells, granulocytes, neutrophils, and monocytes/macrophages among all cohorts receiving BMT, although there was a decrease in the NKT cell population and an increase in the CD4⁺ T-cell population when compared with the untreated mice (Fig. 4D; Supplementary Fig. S7). Intracellular NY-ESO-1 TCR was expressed in the cohorts receiving transduced Lin⁻ cells independent of the T cells received (Fig. 4E; Supplementary Fig. S7). The NY-ESO-1 TCR-positive cells displayed an increase in CD8⁺ T cells, NKT, and macrophages and a decrease in CD4⁺ T cells compared with the total splenocyte population (Fig. 4F; Supplementary Fig. S7).

To test if the expression of the NY-ESO-1 TCR and the sr39TK in the hematopoietic cell progeny after BMT in myelodepleted hosts is immunogenic, we performed BMT experiments with Lin⁻ cells transduced with the lentiviral vectors expressing the NY-ESO-1 TCR and the sr39TK, the NY-ESO-1 TCR alone, or an empty vector using the same animal model described above. We demonstrated that after the BMT, engraftment was observed in mice from all cohorts with no difference in survival (Supplementary Fig. S8A). The peripheral blood cell populations were reconstituted 3 months after BMT, and no significant differences were observed among cohorts in the white cell and red blood cell populations (Fig. 5A; Supplementary Fig. S8B). As seen in the coadministration experiment, at 3 months, the platelets had not completely recovered to the level of untreated control mice (Fig. 5A). To test the potential immunogenicity of the transgenes, we assessed the engraftment and persistence of modified Lin⁻ cells progeny by VCN quantification using primers designed within the lentivirus packaging signal to be able to detect the 3 lentiviruses. If the transgenes were immunogenic, then we would anticipate that cells expressing NY-ESO-1 TCR and sr39TK would not persist long term. As the engraftment and persistence of the transduced stem cell progeny in the bone marrow, spleen, and blood were not affected by the expression of the transgenes (Fig. 5B), we concluded that expression of the NY-ESO-1 TCR and sr39TK was not immunogenic 3 months after autologous BMT in this immunocompetent mouse model.

To assess the potential genotoxicity of the lentiviral vector in Lin⁻ cells, we performed an in vitro immortalization (IVIM) experiment (27). Briefly, Lin⁻ stem cells from C57BL/6J mice were purified and either mock-transduced, transduced with the retroviral vector SF91-eGFP-WPRE (28), or transduced with the LV-NY-ESO-1 TCR/sr39TK lentiviral vector. The retrovirus SF91-eGFP-WPRE is known to be oncogenic and to integrate preferentially close to cancer-related genes (27, 29, 30), thereby serving as a positive control for the IVIM experiment. The Lin⁻ cells were cultured in bulk for 2 weeks and then seeded in limiting dilution and cultured for an additional 2 weeks. At the end of the expansion, the replating frequency/VCN ratio was measured. Transformation occurred in 8 of the 20 independent assays performed with the cells transduced with the positive control vector, SF91-eGFP-WPRE. In contrast, no growth was observed in the mock-transduced Lin⁻ cells (n = 3) or the Lin⁻ cells transduced with LV-NY-ESO-1 TCR/sr39TK (n = 17; Fig. 5C). These data demonstrated lack of genotoxicity of the vector proposed for use in our clinical trial by the IVIM assay.

For the chemistry, manufacturing, and controls (CMC) section of the IND, we performed 5 manufacturing validation runs together with their lot release testing. Three of the validation runs were performed using the GMP-compliant clinical-grade vectors. We purchased G-CSF–mobilized peripheral blood leukapheresis products obtained from healthy human donors and enriched them for CD34⁺ PBSC using a clinical-grade magnetic sorting system. Analysis of 5 enrichment procedures demonstrated mean CD34⁺ fraction purity of 96% and viability of 92%. Following enrichment, the cells were stimulated for approximately 18 hours in X-Vivo 15 media supplemented with rhSCF, rhTPO, rhFLT3L, and rhIL-3 and transduced with the lentiviral vector encoding NY-ESO-1 TCR and sr39TK for an additional 18 hours. Following the transduction, the PBSCs were harvested and cryopreserved; average cell viability was 95%. An aliquot evaluated for bacterial, fungal, and mycoplasmal contamination using both organism-specific cultures and multiple histological stains demonstrated no infections. Endotoxin levels were ≤ 5 EU/kg of body weight. We assessed the transduction efficiency of PBSC with the LV-NY-ESO-1 TCR/sr39TK vector by VCN with an average of 0.3. The percentage of cells expressing NY-ESO-1 TCR (by V_β13 staining) was 24%, and the percentage of colony-forming units (CFUs) positive for Psi (Ψ) sequence was 33%. These results show the feasibility of the manufacturing procedure and were used to establish the lot release acceptance criteria for the clinical trial batches (see Supplementary Table S8 summarizing the cell product specifications for all of the qualification runs and Supplementary Materials Section 5 containing the CMC section of the approved IND).

To study the effect of gene modification on PBSC functionality, we compared the clonogenic potential of PBSC product with that of freshly isolated peripheral blood CD34⁺ cells and untransduced CD34⁺ cells cultured under the same conditions. The clonogenic potential of peripheral blood CD34⁺ cells was similar between the groups as established by percentage of CFU per total cells plated. The average percentage of CFUs was 39 ± 11%, 48 ± 6% and 51 ± 11% (n = 5, P = 0.1), respectively, in the unstimulated, untransduced, and transduced PBSC final product. Finally, we showed that the transduced and in vitro–cultured stem cells were able to differentiate in vitro in artificial thymic organoids into functional T cells expressing the NY-ESO TCR (Supplementary Fig. S9), as previously described by Seet and colleagues (31).

To measure the stability of the TCR-transduced HSC product, we confirmed that the total cell viability, CD34 recovery, and clonogenic potential as well as the CFU subtype distribution were not affected after 1, 3, and 6 months after cryopreservation (Fig. 6A and B). See Supplementary Table S9 and Supplementary Materials Section 6 for the detailed description of the stability evaluation for the LV-NYESO TCR/sr39TK PBSC product. In addition, we evaluated the post-thaw stability of the PBSC product after 3, 6, 24, and 48 hours. The cells were split and stored at either ambient room temperature (RT) or 4°C to 8°C. Based on the results of CD34⁺ cell viability and clonogenic potential, we concluded that the PBSC product should be stable for up to 6 hours after thawing. Although no significant changes in total cell viability were observed (Fig. 6C), the number of viable CD34⁺ cells in the cell product was significantly decreased in both storage conditions after 24 hours (Fig. 6D). The percent CFU per total cells plated significantly decreased after 24 hours in both storage conditions, from 46% to 13% (RT) and 29% (4°C −8°C; Fig. 6E). The average recovery of CD34⁺ cells after 24 hours in storage at RT decreased to 81%, whereas no significant change in recovered CD34⁺ cells was observed while stored at 4°C to 8°C (Fig. 6F).

The IND-enabling studies presented herein were performed based on the written requirements from FDA reviewers discussed at a pre-IND meeting. The safety concerns regarding the proposed clinical trial were mainly focused on the effect of ex vivo modification of HSCs with lentiviral vectors on genotoxicity and stem cell differentiation and the effect of the coadministration with TCR-transduced T cells. Our studies demonstrate that neither of these concerns was evident in preclinical models. Accordingly, the proposed clinical trial has subsequently been approved by the FDA to proceed to patient accrual.

We have previously shown that human CD34⁺ cells transduced with the same lentiviral vector and ex vivo–cultured were able to engraft in myelodepleted neonatal NSG HLA-A2.1 mice and differentiate into B cells and functional T cells (17). In these experiments, we have also demonstrated that the modified HSCs and their progeny could be tracked using PET/CT scan and could be completely ablated following ganciclovir treatment (17). We also have demonstrated that human CD34⁺ cells transduced with the LV-NY-ESO-1 TCR vector were able to differentiate in vitro in artificial thymic organoids to functional T cells with tumor antigen-specific cytotoxicity and near-complete lack of endogenous TCR V_β expression due to allelic exclusion (31). In the current studies using immunocompetent mouse models, we have established that coadministration of genetically modified Lin⁻ cells and T cells does not affect the engraftment and persistence of T cells and the engraftment, differentiation, and progeny persistence of the Lin⁻ cells. Based on these new results, we conclude that the expression of NY-ESO-1 TCR and sr39TK do not alter the differentiation of the Lin⁻ cells or the persistence of their progeny, supporting the lack of immunogenicity of these transgenes after BMT in myelodepleted mice. Overall, these data indicate that coadministration of TCR-engineered T cells and TCR-engineered HSCs represents a viable means of obtaining cells capable of mounting both an immediate and a prolonged tumor-targeted immune response.

A concern regarding the HSCT with lentivirally transduced HSCs is the potential genotoxicity of the vector caused by its integration close to cancer-related genes leading to clonal expansion, a process termed insertional mutagenesis. This concern arises from previous clinical trials using gamma-retroviral vectors that preferentially integrate near oncogenes (LMO2, MDS1-EVI1, PRDM16, SETBP1, CCND2 among others) and deregulate their expression due to the strong viral promoters located in their LTRs; this phenomenon led to the development of leukemia and myelodysplasia in the first gene therapy clinical trials using gene-modified HSCs (32–35). More recently, clinical trials of HSCT with HSCs genetically modified with self-inactivating (SIN) lentiviral vectors have proven the safety of this approach, showing high polyclonality and no expansion of clones with integrations near cancer-related genes previously associated with insertional mutagenesis (36–40). In the current clinical trial, we plan to use a SIN-lentiviral vector with an MSCV promoter derived from the MPSV murine gamma retrovirus (41) to drive the expression of the transgenes. There is no prior clinical experience with this lentiviral vector. However, clinical trials using SIN-lentiviral vectors with the closely related MND promoter derived from the MPSV murine gamma-retrovirus (42) to drive the expression of the ABCD1 gene have shown safety and lack of clonal expansion of the modified cells (43, 44). Interestingly, Biffi and colleagues compared the integration profile of different lentiviruses in human HSCs engrafted in immunodeficient mice and showed that integration clustered in megabase-wide chromosomal regions of high lentiviral vector integration density demonstrating the bias toward specific genomic regions rather than by oncogenic selection (37). In line with these data, we demonstrated that the transduction of murine Lin⁻ cells with our vector does not lead to transformation in vitro in contrast with the gamma-retroviral vector used as a positive control.

T cells are a self-renewing population, but we and others have previously shown that their persistence is limited and associated to increased antitumor responses (4, 5, 8). HSC gene therapy holds great promise for immunotherapy due to the long-term persistence of the modified cells. Similar to the clinical trial proposed in this article, the use of TCR genetically engineered HSCs has been proposed for HIV-1 therapy (45, 46) in which feasibility has been shown in preclinical models. In line with this, cancer therapy and HIV-1 therapy approaches using chimeric antigen receptors (47–49) and invariant natural killer T (iNKT) cell receptors (50) have been proposed in preclinical studies. These studies demonstrate the ability of HSCs to differentiate into functional T cells or iNKT cells expressing the selected antigen-specific receptor and show antitumor activity or functional antiviral response to limit HIV-1 replication. The coadministration of T cells together with HSCs is key to allow the HSCs to engraft and generate new functional T cells and at the same time provide a therapeutic effect, especially in patients with advanced disease.

B. Bolon is an employee of Flagship Biosciences. R.P. Hollis is an employee of Curative Therapeutics. No potential conflicts of interest were disclosed by the other authors.

Conception and design: C. Puig-Saus, A. Garcia-Diaz, S. Sandoval, E.H. Gschweng, L. Yang, D.B. Kohn, P.J. Kaplan-Lefko, B. Berent-Maoz, A. Ribas

Development of methodology: C. Puig-Saus, G. Parisi, A. Garcia-Diaz, P.E. Krystofinski, S. Sandoval, R. Zhang, A.S. Champhekar, J. McCabe, A. Vega-Crespo, G.M. Crooks, R.P. Hollis, B. Campo-Fernandez, L. Yang, D. Baltimore, B. Comin-Anduix, B. Berent-Maoz, A. Ribas

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Puig-Saus, A. Garcia-Diaz, P.E. Krystofinski, S. Sandoval, R. Zhang, A.S. Champhekar, J. McCabe, G.C. Cheung-Lau, N.A. Truong, A. Vega-Crespo, M.D.S. Komenan, J. Pang, M.H. Macabali, J.D. Saco, J.L. Goodwin, B. Bolon, C.S. Seet, A. Montel-Hagen, D. Bischof, K. Cornetta, A. Ribas

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Puig-Saus, G. Parisi, A. Garcia-Diaz, P.E. Krystofinski, S. Sandoval, A.S. Champhekar, A. Vega-Crespo, B. Bolon, C.S. Seet, A. Montel-Hagen, B. Campo-Fernandez, B. Comin-Anduix, X. Wang, B. Berent-Maoz, A. Ribas

Writing, review, and/or revision of the manuscript: C. Puig-Saus, G. Parisi, P.E. Krystofinski, S. Sandoval, A. Vega-Crespo, B. Bolon, D. Bischof, K. Cornetta, A. Nguyen, D.B. Kohn, X. Wang, P.J. Kaplan-Lefko, B. Berent-Maoz, A. Ribas

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Puig-Saus, A. Garcia-Diaz, P.E. Krystofinski, R. Zhang, N.A. Truong, A. Vega-Crespo, M.D.S. Komenan, J. Pang, J.L. Goodwin, B. Campo-Fernandez, C. Adelson, A. Nguyen, O.N. Witte, P. Cabrera, P.J. Kaplan-Lefko, A. Ribas

Study supervision: C. Puig-Saus, G. Parisi, P.E. Krystofinski, J. McCabe, O.N. Witte, P.J. Kaplan-Lefko, B. Berent-Maoz, A. Ribas

Other (optimization of vector production, and generation of the vector used in this study): D. Bischof

Other (regulatory and GLP support): A. Nguyen

This work was funded by the California Institute for Regenerative Medicine (CIRM; A. Ribas), NCI grants P01 CA168585 and R35 CA197633 (A. Ribas), the PhaseOne Foundation (A. Ribas), the Caltech/UCLA Joint Center for Translational Medicine (D. Baltimore, O.N. Witte, and A. Ribas), the Ressler Family Fund, the Grimaldi Family Fund, the Samuels Family Fund, and the Garcia-Corsini Family Fund (A. Ribas). A. Ribas and O.N. Witte are members of the Parker Institute for Cancer Immunotherapy.

We would like to acknowledge the support of the CIRM, the UCLA Broad Stem Cell Research Center, and the CIRM Alpha Stem Cell Clinic at UCLA. The National Gene Vector Biorepository (funded by a P40 grant from the NHLBI 9P40HL116242, K. Cornetta) provided cells and plasmids available in their repository and archive the test articles and specimens generated in the GLP studies. Flow cytometry was performed in the UCLA Jonsson Comprehensive Cancer Center (JCCC) and Center for AIDS Research Flow Cytometry Core Facility, which is supported by NIH awards CA-16042 and AI-28697, and the JCCC, the UCLA AIDS Institute, and the David Geffen School of Medicine at UCLA. We thank Dr. Stephen T. Smale for his role as Test Facility Management in the GLP study, Dr. Richard C. Koya and Dr. Thinle Chodon for their support in the early stages of the project, all Division of Laboratory Animal Medicine (DLAM) personnel for their support with the animal work and GLP implementation, and Justin Tran for his help with the logistics of the project.

The NY-ESO TCR used in this study is covered by patents including US 8,143,376 and is used under license from Adaptimmune Limited. As of September 7, 2017, GlaxoSmithKline plc (LSE:GSK) (NYSE:GSK) has exercised its option under a collaboration and license agreement signed in 2014 to exclusively license the right to research, develop, and commercialize Adaptimmune's NY-ESO SPEAR T-cell therapy program.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.