Bispecific anti-CD20, anti-CD19 CAR T cells for relapsed B cell malignancies: a phase 1 dose escalation and expansion trial.

Chimeric antigen receptor (CAR) T cells targeting CD19 are a breakthrough treatment for relapsed, refractory B cell malignancies1-5. Despite impressive outcomes, relapse with CD19- disease remains a challenge. We address this limitation through a first-in-human trial of bispecific anti-CD20, anti-CD19 (LV20.19) CAR T cells for relapsed, refractory B cell malignancies. Adult patients with B cell non-Hodgkin lymphoma or chronic lymphocytic leukemia were treated on a phase 1 dose escalation and expansion trial (NCT03019055) to evaluate the safety of 4-1BB-CD3ζ LV20.19 CAR T cells and the feasibility of on-site manufacturing using the CliniMACS Prodigy system. CAR T cell doses ranged from 2.5 × 105-2.5 × 106 cells per kg. Cell manufacturing was set at 14 d with the goal of infusing non-cryopreserved LV20.19 CAR T cells. The target dose of LV20.19 CAR T cells was met in all CAR-naive patients, and 22 patients received LV20.19 CAR T cells on protocol. In the absence of dose-limiting toxicity, a dose of 2.5 × 106 cells per kg was chosen for expansion. Grade 3-4 cytokine release syndrome occurred in one (5%) patient, and grade 3-4 neurotoxicity occurred in three (14%) patients. Eighteen (82%) patients achieved an overall response at day 28, 14 (64%) had a complete response, and 4 (18%) had a partial response. The overall response rate to the dose of 2.5 × 106 cells per kg with non-cryopreserved infusion (n = 12) was 100% (complete response, 92%; partial response, 8%). Notably, loss of the CD19 antigen was not seen in patients who relapsed or experienced treatment failure. In conclusion, on-site manufacturing and infusion of non-cryopreserved LV20.19 CAR T cells were feasible and therapeutically safe, showing low toxicity and high efficacy. Bispecific CARs may improve clinical responses by mitigating target antigen downregulation as a mechanism of relapse.

In vivo ligand-inducible regulation of gene expression in a gutless adenoviral vector system.

Transcriptional regulation that is rapid, reversible, and repeatedly inducible would greatly enhance the safety and efficacy of many gene therapy strategies. We developed a chimeric ligand-inducible regulation system based on the human estrogen receptor. This system has two components, the responsive promoter driving expression of the transgene of interest, and the ligand-inducible chimeric transcription factor. The transcription factor is composed of a novel DNA binding domain and a modified estrogen receptor ligand-binding domain. A point mutation in the ligand-binding domain significantly reduces estrogen binding while allowing binding of the estrogen antagonist, tamoxifen. We used a gutless adenoviral vector system and incorporated both components into two separate vectors. A single gutless vector encoding both system components was also generated. The tamoxifen-mediated induciblity of transgene expression of the gutless vector system was compared in vitro and in vivo with the analogous components incorporated into early generation, E1/E2a/E3-deficient adenoviral vectors. In normal mice, both the gutless vector and early generation systems displayed inducibility in the presence of tamoxifen. Importantly, the gutless vector system was inducible to extremely high levels, at least four times over a 2-month period. In contrast, the early generation vector system was inducible only once. Furthermore, the early generation system displayed significant toxicity, as evidenced by extremely high liver enzyme levels, abnormal liver pathology, and rapid loss of vector DNA from the liver, while the gutless vector system displayed minimal toxicity. These data directly demonstrate the improved in vivo function of the tamoxifen-inducible transcriptional regulation system in the context of the gutless adenoviral vectors.

Intraocular expression of endostatin reduces VEGF-induced retinal vascular permeability, neovascularization, and retinal detachment.

Endostatin, a proteolytic fragment of collagen XVIII, is an endogenous inhibitor of tumor angiogenesis that also inhibits choroidal neovascularization. In this study, we assessed the effects of increased intraocular expression of endostatin on vascular endothelial growth factor (VEGF)-induced changes in the retina. After subretinal injection of a pair of gutless adenoviral vectors (AGV) designed to provide tamoxifen-inducible expression of endostatin, diffuse endostatin immunoreactivity was induced thoroughout the retina by administration of tamoxifen. Induction of endostatin in double transgenic mice with doxycycline-induced expression of VEGF in the retina resulted in significant suppression of leakage of intravascular [3H]mannitol into the retina. The ability of endostatin to reduce VEGF-induced retinal vascular permeability was confirmed by using [3H]mannitol leakage and two other parameters, fluorescein leakage and retinal thickness, after subretinal injection of a bovine immunodeficiency lentiviral vector coding for endostatin (BIV-vectored endostatin, or BIVendostatin). Subretinal injection of BIVendostatin resulted in more discrete, less intense staining for endostatin in the retina than that seen with the inducible AGV system, which suggested lower levels and allowed visualization of sites where endostatin was concentrated. Endostatin staining outlined retinal blood vessels, which suggested endostatin binding to a component of vessel walls. More prolonged or higher level expression of VEGF in the retina resulted in neovascularization and retinal detachment, both of which were also significantly reduced by BIVendostatin. These data suggest that endostatin may be an endogenous inhibitor of vasopermeability as well as neovascularization. In patients with diabetic retinopathy, endostatin gene transfer may provide a way to decrease the risk of three causes of visual loss: macular edema, neovascularization, and retinal detachment.

Optimization of the generation and propagation of gutless adenoviral vectors.

Adenoviral vectors devoid of all viral coding regions are referred to by many names, including gutless vectors. Gutless vectors display reduced toxicity and immunogenicity, increased duration of transgene expression, and increased coding capacity compared to early generation vectors, which contain the majority of the viral backbone genes. However, the production of gutless vectors at a scale and purity suitable for clinical use has limited the utility of this technology. In this work we describe the optimization of the production of gutless vectors. We constructed an improved helper virus and generated an alternative gutless vector producer cell line, PERC6-Cre. We demonstrated increased gutless vector yields, minimal helper virus contamination, and no replication-competent adenovirus contamination using the optimized system. Furthermore, the PERC6-Cre cells were adapted to serum-free suspension culture and high-titer gutless vector preparations were produced using bioreactor technology, suggesting the feasibility of gutless vector scale-up for clinical use. Finally, we observed that helper virus lacking a packaging signal could be packaged at a low frequency, revealing an inherent limitation to the differential packaging strategy for gutless vector propagation.