The Clonal Inventory of Gene Corrected Hematopoiesis in Three Successful Clinical Gene Therapy Trials

To enhance the safety and efficacy of future gene therapy trials using integrating vector systems it is necessary to analyse the clonality of the genetically modified cell pool. The comparative analyses of integration site distribution and cellular gene expression will further reveal causal mechanisms of in vivo clone selection. We followed the repopulation clonality of 21 patients which participated in 3 successful clinical gene therapy trials via linear amplification mediated PCR (LAM PCR). We identified the integration sites (IS) of pre and post transplantation samples by Sanger sequencing and accomplished RNA analyses. The comparative results from all trials showed that vector integration is favoured in gene coding regions, in particular transcriptional start sites. In both X-SCID trials significantly more post transplantation IS were located in or in the vicinity of genes encoding proteins with kinase or transferase activity. In pre transplantation samples no uniform gene class was overrepresented. In both trials we detected common insertion sites mainly post transplantation and the effect was more pronounced in the trial where 4 patients developed vector induced leukemia. Notably, we detected no significant differences regarding the IS distribution in leukemic versus non leukemic patients. The gene corrected repopulation of patient 1 and 2 of the X-CGD trial was polyclonal until 542 and 777 days after transplantation, respectively. 5 months after therapy dominant clones appeared. In patient 1, between 616 and 820 days post transplantation (post mortem time point) the number of participating clones and the contribution of a dominant clone decreased while the contribution of another dominant clone increased. In both patients the integrated vector induced the upregulation of the genes MDS1/EVI1, PRDM16 or SETBP1 and thus led to the in vivo expansion of affected cell clones. From these trials we sequenced >2000 unique IS by Sanger sequencing and several thousand via pyrosequencing (datamining is ongoing). Our data show that the integration site distribution was non random, that the integrated vector influenced the cellular gene expression which caused subtle to massive changes in the repopulation clonality and that it will soon be possible to define the clonal inventory of patients using next generation sequencing technologies.