Using barcode vectors for neutral genetic marking to study clonal dynamics of hematopoietic reconstitution

Abstract

Haematopoietic stem cell transplantation (SCT) is the only curative therapy option for a variety of malignant and non-malignant blood diseases. The contribution and importance of various cell populations to hematopoietic reconstitution have been studied extensively, but little is known about clonal dynamics within these populations. To study these dynamics, stable, inheritable marking of donor cells is required, which can be provided by integrating retroviral vectors. Unfortunately, stable genome insertion is associated with the risk of insertional mutagenesis potentially leading to malignant transformation of affected cells as documented in several gene therapy studies. Improved retroviral vector design has greatly reduced, but not removed the likelihood of insertional mutagenesis in the last decade. In parallel, development of genetic barcoding techniques has opened the possibility to analyse clonal composition and dynamics in greater detail than before. We reasoned that genetic barcoding of hematopoietic cells with state-of-the art retroviral vector system should facilitate high-resolution analysis of neutral hematopoietic reconstitution, unaffected by the marking procedure itself. Within this thesis, I wanted to evaluate the influence of the vector type and their internal promoters on clonal dynamics of hematopoietic reconstitution after SCT. Based thereon, the vector construct best suited to study neutral reconstitution should be determined. To achieve this task, I took advantage of a genetic barcoding system with colour-coding capabilities. Alpha- and lentiviral vector constructs equipped with either a strong, intermediate or no promoter upstream of a fluorescence protein (FP) were barcoded and used to independently transduce lineage-negative cells of donor mice. I studied four groups with different competitive in-vivo setups by transplanting grafts, containing up to three different vector constructs, into lethally irradiated recipients. Genomic DNA was extracted from peripheral blood (PB) samples taken monthly, while selected time points were additionally analysed by flow cytometry (FC). Samples from PB, spleen, bone marrow as well as flow cytometrically sorted subsets of T cells,  cells and granulocytes were collected eight to twelve months after SCT. Barcode analysis via next-generation sequencing (NGS) created a dataset with temporal dynamics from successive PB samples, while the different populations sampled at the final analysis time point show the spatial distribution of clones. FC analysis of chimaerism and FP expression confirmed stable long-term engraftment of the marked populations detectable over the whole observation period. The number of barcoded cells significantly contributing to haematopoiesis declined over time in most animals, although I observed big variability between individual mice. Temporal and spatial analyses of clonal dynamics in the animals showed a diverse picture from monoclonal to polyclonal situations. In most samples, approximately 15 clones per construct contributed to more than 75% of the marked fraction. Nearly all clones measurably contributing to haematopoiesis at final analysis were already present six weeks after transplantation, and only their frequencies changed over time. Dominant or prominent clones representing a big fraction of the marked haematopoiesis were detectable for all vector constructs tested without any bias towards any of the constructs used. Finally, the obtained dataset was used to calculate the amount of cells contributing to haematopoiesis. These calculations indicate that around 350 cells actively supply the blood production six weeks after transplantation. Eight to twelve months after transplantation, this number decreases to around 260 cells. In the transplantation setup investigated here, lenti- and alpharetroviral vector constructs equipped with different promoters showed comparable clonal dynamics and/or trends in all analyses. There might be some negative characteristics for the construct with the strong viral promoter, but the collected data is not sufficient for a final assessment. Thus, in principle all constructs appeared suitable for investigating undisturbed reconstitution of the haematopoietic system after transplantation. On the other hand, I also observed dominant, or at least prominent, clones marked by vector constructs without any promoters indicating that intrinsic cell features might promote clonal dominance. In conclusion, this work demonstrates the feasibility to mark and track several distinct cell populations in parallel in vivo within single animals with a barcoding system. This allows competitive setups, where effects of various parameters (e.g. promoters), can be compared based on the colour-coded barcode backbones, while the individual barcodes provide further information about the clonal dynamics within a population.