Identification and validation strategy for hematopoietic stem cell regulatory genes

In this study we used large data sets of genomic γ-retroviral (γRV) integration sites (IS) from a clinical gene therapy trial to identify HSC regulatory genes and developed a pooled competitive overexpression strategy for functional validation in vivo. We hypothesized that common integration sites (CIS) indicate actively transcribed genes in human HSCs or genetic loci, which are deregulated by viral enhancer and promotor elements nearby transcription start (TSS) sites, leading to clonal expansion. Systematic analysis of the IS repertoire of 10 Wiskott-Aldrich syndrome patients revealed 130,699 unique IS in the vicinity of 12,437 genes. Applying stringent selection criteria based on the cluster size and distance of IS to TSS, we identified 1,545 gene associated CIS. Approximately 25% of the top ranked CIS are located in the vicinity of published HSC regulators, such as MECOM, LMO2 or HMGA2. Moreover, pathway analysis using the CIS tagged gene signature revealed significant enrichment for hematological system development and function as well as immune response pathways, thus strongly validating our selection strategy. To study the influence of all 20 top ranked candidate genes in parallel, we generated an inducible and barcoded lentiviral overexpression library, consisting of 2 unique barcodes (BCs) per gene and 10 barcoded GFP-only controls. Nested amplification of BCs and indexing of 96 samples/lane (HiSeq 2500) was established and tested intensively in a “one BC/sample” experiment, showing averagely 1.37E6 reads per sample with only 11 off target reads per BC (0.0008%±2.8E-6 [SEM]). Pooled transduction of LSKRosa26-rtTA cells and Doxycycline administration 4 weeks after transplantation showed long-term GFP expression in 1° and 2° recipients (94.4%±0.4 engraftment; 4.2%±1.1 GFP+ [SEM]). Serial bleedings and multiplexed HT-sequencing of amplified BCs from gDNA of B-, T-, and myeloid cells revealed a complete representation of all BCs in all mice (n=23) before and after Dox administration with 15,643±656 (SEM) read counts/BC at week 4 and 46,222±3276 (SEM) read counts/BC at week 16. Our pooled competitive overexpression strategy allows us to study 20 genes in parallel facilitating the discovery of novel key players in hematopoiesis.