DNA Barcoding in Nonhuman Primates Reveals Important Limitations in Retrovirus Integration Site Analysis

In vivo tracking of retrovirus-tagged blood stem and progenitor cells is used to study hematopoiesis. Two techniques are used most frequently: sequencing the locus of retrovirus insertion, termed integration site analysis, or retrovirus DNA barcode sequencing. Of these, integration site analysis is currently the only available technique for monitoring clonal pools in patients treated with retrovirus-modified blood cells. A key question is how these two techniques compare in their ability to detect and quantify clonal contributions. In this study, we assessed both methods simultaneously in a clinically relevant nonhuman primate model of autologous, myeloablative transplantation. Our data demonstrate that both methods track abundant clones; however, DNA barcode sequencing is at least 5-fold more efficient than integration site analysis. Using computational simulation to identify the sources of low efficiency, we identify sampling depth as the major factor. We show that the sampling required for integration site analysis to achieve minimal coverage of the true clonal pool is likely prohibitive, especially in cases of low gene-modified cell engraftment. We also show that early subsampling of different blood cell lineages adds value to clone tracking information in terms of safety and hematopoietic biology. Our analysis demonstrates DNA barcode sequencing as a useful guide to maximize integration site analysis interpretation in gene therapy patients.