Human cancer-targeted immunity via transgenic hematopoietic stem cell progeny

Abstract

Adoptive transfer of genetically engineered T cells expressing a tumor-antigen-specific transgenic T cell receptor (TCR) can result in clinical responses in a variety of malignancies. However, these responses are frequently short-lived, and patients typically relapse within several months. This phenomenon is largely due to poor persistence of the transgenic T cells, as well as a progressive loss of their functionality and terminal differentiation in vivo. This underscores the need for cell therapy approaches able to sustain the initial antitumor efficacy and lead to long-term antitumor efficacy. Herein, we report the use of tandem cell therapies involving autologous T cells and hematopoietic stem cells engineered to express the NY-ESO-1 TCR for the treatment of solid tumors in a first-in-human phase I clinical trial (NCT03240861). This therapy is shown to be safe, feasible, and leads to initial tumor regression activity. T cell progeny from the HSC progenitors is shown to provide circulating transgenic NY-ESO-1 TCR-T cells, which display tumor-antigen-specific antitumor functionality, without any evidence of anergy or exhaustion. These results demonstrate the utility of transgenic HSCs to generate a self-renewing source of tumor-specific cellular immunotherapy in human participants. Clinicaltrials.gov: NCT NCT03240861.

Fig. 1. Overview of manufacturing protocol, and clinical trial procedures.

Peripheral blood stem cells (PBSCs) were isolated from patients via leukapheresis following bone marrow mobilization with filgrastim and plerixafor. CD34+ cells are isolated via magnetic bead separation using the CliniMACS platform, and were then transduced with the lentiviral vector encoding the NY-ESO-1 TCR and sr39TK reporter/suicide gene (enabling non-invasive visualization of the transgenic HSC progeny in vivo within the bone marrow niche, as well as serve as a safety ablation feature), and were then cryopreserved. Following final lot release criteria for the gene-modified PBSCs, patients were admitted following unmobilized leukapheresis to collect peripheral blood mononuclear cells (PBMCs), which were then transduced with a retroviral vector encoding the NY-ESO-1 TCR while the patient received conditioning chemotherapy with busulfan and fludarabine in parallel, in order to selectively myeloablate and lymphodeplete the patient, respectively. Gene-modified PBSCs and PBMCs were administered to the patient on day 0 and day +1, respectively, and patients subsequently received 7 days of IL-2 for transgenic T-cell expansion in vivo. Created in BioRender. Nowicki, T. (2025) https://BioRender.com/4bcin2w.

Fig. 2. Retroviral and lentiviral-transduced cell engraftment kinetics in patient NYSCT-01.

A Peripheral NY-ESO-1 TCR+ expansion and contraction over time in vivo. B Vector copy number (VCN) kinetics in peripheral circulation. While there was a predominance of retrovirus (RV)-derived product early, there was predominantly lentiviral (LV)-derived product coinciding with the day +29 [18 F]-FHBG PET (C), followed by a loss of LV-derived product coinciding with the day +120 PET scan (D). Data in 2B represent two experimental replicates, ± SD.

Fig. 3. Retroviral and lentiviral-transduced cell engraftment kinetics in patient NYSCT-05.

A Peripheral TCR+ expansion and contraction over time in vivo. B Vector copy number (VCN) kinetics in peripheral circulation. While there was a predominance of retrovirus (RV)-derived product early, there was no significant detectable lentiviral (LV) progeny at any point, consistent with the non-engraftment observed which was observed at the day +35 [18 F]-FHBG PET scan (C). Data in 3B represent two experimental replicates, ± SD.

Fig. 4. Retroviral/lentiviral engraftment kinetics in patient NYSCT-03.

A Peripheral TCR+ expansion and contraction over time in vivo. B Vector copy number (VCN) kinetics in peripheral circulation. While there was a predominance of retrovirus (RV)-derived product early, there was predominantly lentiviral (LV)-derived product coinciding with the day +25 [18 F]-FHBG PET (C). There was no data beyond day +43, as the patient had passed away, and there was insufficient sample at day +53 for analysis. Data in 4B represent two experimental replicates, ± SD.

Fig. 5. Single nuclei RNA sequencing (snRNAseq) of PBMC from patient NYSCT-03 at day 43.

A Overview of ddPCR VCN kinetics at timepoint selected for analysis, which yielded 4166 nuclei after doublet removal and QC (as shown in Fig. 4B). Of the PBMC population, 1954 nuclei were lentiviral-positive (pLenti + , 46.9%) and 2212 nuclei were lentiviral-negative (pLenti-, 53.1%) (B), while only 10 nuclei were retroviral-positive (pretro + ) and 4156 were retroviral-negative (pretro-) (C). Both pLenti+ and pLenti- cells expressed all immune cell types from Azimuth References at comparable levels in RNA and ATAC UMAPs (D), and all immune cell type states at comparable levels; myeloid cells clustered together, B cells clustered separately, and T cells and NK cells clustered together in RNA and ATAC UMAPs (E). NYESO1 TCRB lentivirus encodes the same amino acids as T cell receptor beta (TRBC1) gene, but is codon optimized. T cells represented 185 nuclei (≈4.5%) of all nuclei, which were equally represented by pLenti+ and pLenti- nuclei (47% and 53%, respectively). Of pLenti+ Tcells, 37% co-expressed genomic TRBC1/2 + codon optimized lentiviral NY-ESO-1 TCR (pRRL-lenti), while 63% exclusively expressed codon optimized pRRL-lenti TCR (F). Of the T cells, 32 nuclei co-expressed the genomic TRBC1/2 + the pLenti TCR, while 98 and 55 nuclei had mutually exclusive expression of either genomic TRBC1/2 or codon optimized pRRL-lenti TCR, respectively (G). T cell CD4:CD8 ratio was 3:1, and the proportion of nuclei belonging to CD4 T cell, CD8 T cell, or other mature T cells was comparable for nuclei expressing exclusively TRBC1/2 (blue) or nuclei expressing codon optimized pRRL-lenti TCR (H). Of all pLenti+ nuclei (1954), 7% co-expressed genomic TRBC1/2 (131), while of all pLenti+ T cells (87 nuclei), 37% co-expressed genomic TRBC1/2 (32 nuclei) (I). Heatmap for expression of pLenti TCR, TRBC2, and TRBC1 per nucleus for pLenti+ T cells revealed allelic exclusion in the majority of nuclei. 45% of the CD4 T cell clusters/state co-expressed the NYESO1 pLenti TCR and genomic TRBC1/2, while 15.4% of the CD8 T-cell cluster/states co-expressed NYESO1 and genomic TRBC2 but not TRBC1, remaining nuclei exclusively expressed the codon optimized NYESO1 pLenti TCR (J).