Estimated comparative integration hotspots identify different behaviors of retroviral gene transfer vectors

Abstract

Integration of retroviral vectors in the human genome follows non random patterns that favor insertional deregulation of gene expression and may cause risks of insertional mutagenesis when used in clinical gene therapy. Understanding how viral vectors integrate into the human genome is a key issue in predicting these risks. We provide a new statistical method to compare retroviral integration patterns. We identified the positions where vectors derived from the Human Immunodeficiency Virus (HIV) and the Moloney Murine Leukemia Virus (MLV) show different integration behaviors in human hematopoietic progenitor cells. Non-parametric density estimation was used to identify candidate comparative hotspots, which were then tested and ranked. We found 100 significative comparative hotspots, distributed throughout the chromosomes. HIV hotspots were wider and contained more genes than MLV ones. A Gene Ontology analysis of HIV targets showed enrichment of genes involved in antigen processing and presentation, reflecting the high HIV integration frequency observed at the MHC locus on chromosome 6. Four histone modifications/variants had a different mean density in comparative hotspots (H2AZ, H3K4me1, H3K4me3, H3K9me1), while gene expression within the comparative hotspots did not differ from background. These findings suggest the existence of epigenetic or nuclear three-dimensional topology contexts guiding retroviral integration to specific chromosome areas.

Figure 1. Integration densities of HIV and MLV in CD34⁺ cells, for chromosome 6.

We analyzed each strand separately: the upper half is the + strand and the lower the - strand. In blue the estimated variability band at level 0.99 for HIV integrations (n = 1629), in red for MLV (n = 1815). Candidate comparative hotspots are plotted in the two central x-axes, the color indicating which of the two vectors had stronger integration intensity (HIV: blue; MLV: red). In the other four x-axes, each tick represents one integration site, with the same color code. Because of resolution, many ticks fall on the same point and cannot be distinguished.

Figure 2. Two typical situations in comparative hotspots.

In panel A (left side) the bands don't overlap in plus and minus strands, suggesting the presence of two candidate comparative hotspots (hotspots ID: hiv_36 and hiv_40, see supplementary material, Table S1). Differences in integration densities in one versus the other strand may reflect a preferential integration orientation at that particular locus. In panel B (right side) the bands don't overlap in the plus strand (upper panel) whereas on the minus strand they do (lower panel), suggesting only one candidate (hotspot ID: mlv_43, see supplementary material, Table S1). These examples are taken from chromosome 6.

Figure 3. Gene density of HIV and MLV comparative hotspots.

Histogram of the number of target genes per hotspot normalized by hotspot length in HIV (blue bars, upper panel) and MLV (red bars, lower panel).

Figure 4. Robustness plots.

Panel A: chromosome 6, arm p; strand -. Panel B: chromosome 6, arm q; strand -. Here we can observe how hotspots would change in length and location if we were to use different smoothing parameters. Most importantly, we see that the hotspots identified at level 1, corresponding to our choice of the smoothing parameters, persist at slightly larger and smaller values, confirming their validity. At smaller levels of smoothing many spurious hotspots appear, of very short length. There is no support from the data for these, as they either disappear for more smoothing or they merge into larger and more robust segments. Large smoothing either impairs the creation of hotspots (as bands tend to become large and flat) or they deliver very large hotspots, which are difficult to interpret biologically.

Figure 5. Blind regions.

These plots illustrate the presence of blind regions, which are scattered over the genome and usually short, but occasionally also of appreciable length (panel A: mlv_53, chromosome 7, arm p strand -; panel B: mlv_147, chromosome 22 arm q, strand +). Not all the candidate comparative hotspots that we identified were clearly distinguishable, see for example panel A.