Bipotent Lymphoid Progenitors Independently Maintain Long-Term Genetically Engineered T and NK Cell Production in Humans

Historically, survival and activity of individual human hematopoietic progenitor subtypes have been studied exclusively via transplantation in permissive mouse models. Despite being a relevant investigational tool, such animal models do not recapitulate the human hematopoietic milieu. Specifically, the production of human T and NK cells in these mice is severely impaired therefore, as of today, it has been impossible to measure with enough confidence the in vivo dynamics of human T/NK lymphoid progenitors. Hematopoietic stem cell gene therapy (GT) using integrating viral vectors has opened a unique opportunity to trace the fate of transplanted cells for the first time directly in vivo in humans by means of integration sites (IS) clonal tracking. Our mathematical modelling of IS data in clinical GT supported the existence of a population of long-term lymphoid progenitors (LtLP) capable of surviving independently from hematopoietic stem cells (HSC). However, to date, no experimental setting has been available to validate such statistical prediction neither in the mouse nor in humans.

We here report the first formal evidence, directly in vivo in humans, that a population of bipotent lymphoid progenitors is capable of surviving and maintaining de novo T/NK production for at least 15 years after loss of transplanted HSC. In 5 SCID-X1 patients treated with gammaretroviral vector HSC-GT we observed that genetically engineered myeloid and B cells have stayed constantly below detection limits starting from 6 months after GT. On the contrary T and NK cells remained vector positive for up to 19 years (average vector copies/cell equal 2.1 and 2.3 respectively). We initially thought that such data would be consistent with permanent loss of engineered HSC and survival of vector-marked long-living circulating mature lymphocytes. However, strikingly and unexpectedly, by a comprehensive immunophenotyping of T cells overtime, we detected vector-positive short-living CD45RA+CD62L+CD95- naïve T cells (Tn) in the circulation of these patients. We first obtained functional validation of the identity of such Tn cells by IFN-y productionassays and we confirmed positive thymic activity by measurement of TREC content. High throughput sequencing of 1,193 T-cell receptor (TCR) rearrangements in FACS-sorted T cell subtypes confirmed that these engineered Tn cells had normal TCR diversity and that new rearrangements were detectable overtime. Similarly, Vbeta repertoire measured by spectratyping displayed normal profiles in all patients. These results suggested that a de novo T-cells production is maintained in these patients by a putative LtLP population in absence of supply by gene-corrected HSC. Using Tn data as surrogate markers of LtLP clonal dynamics and composition, we collected and analysed 12,756 unique IS from 5 populations including Tn cells in a window of 10.1 to 14.9 years after the observed loss of transplanted HSC. Clonal diversity was stable in all T subtypes and IS sharing across subpopulations and timepoints was high and consistent with active in vivo output by LtLP and differentiation of Tn into memory/effector T cells. Analyzing nature and recapture probability of IS we observed that IS in LtLP are significantly enriched in genes involved in lymphocyte survival/activation and we could estimate that T cell production is currently sustained by 2,092-6,056 individual engineered LtLP clones. Moreover, by studying 651 IS collected from CD3-CD56+ purified NK cells we observed that up to 41.2% of them were shared with independently analyzed Tn suggesting that the LtLP active in these patients have a dual T/NK restricted potential. Lastly, we detected and tracked overtime 52 clones bearing IS in MECOM, CCND2 and LMO2 including the IS originally associated with vector-induced leukemia in one patient. Our data show absence of any sustained clonal expansions within the engineered LtLP population for up 19 years after GT, the longest follow up available to date in clinical HSC-GT.

In conclusion, our data provide the first formal evidence in vivo in human that a de novo production of genetically engineered T and NK cells can be physiologically maintained by a population of LtLP surviving up to 15 years after loss of transplanted HSC (manuscript submitted). Identification and exploitation of such human LtLP population might be crucial for the development of next generation GT and cancer immunotherapy approaches.