Airway gene-addition therapy could offer lasting benefits for people with CF, particularly those that do not have access to highly effective CFTR modulator therapies. Amongst all the potential gene transfer vehicles, LV vectors have been demonstrated to have characteristics that make them particularly appealing, one of which is their low immunogenicity which makes redosing possible. The ability to increase levels of initial gene expression, or boost levels when they wane over time will likely be vital for any vector delivery system. Although LV redosing has been examined in the past [14, 22], neither the use of the VSV-G pseudotype nor the use of LPC conditioning in a repeat-dosing scenario have been thoroughly examined. The primary aim of this study was to determine the optimal dosing schedule of a VSV-G pseudotyped LV vector delivered into mouse lungs.

Our first hypothesis was that multiple closely-spaced early doses would increase initial lung gene expression levels, resulting in better long term expression when compared to a single dose. This was based on Griesenbach et al. who found evidence for the effectiveness of early multiple doses, reporting that five or ten daily doses of a F/HN-SIV-Lux vector both significantly increased nasal and lung luciferase expression compared to a single dose [14], and Sinn et al. who reported an increase in nasal luciferase expression following seven doses over a 1-week period [22]. In our study, there was no evidence from the in vivo flux assessments that any dosing schedule produced significantly higher mean expression than the dose group at any imaging time point. However, in the ex vivo BLI assessment of lungs removed at the end of the study, the estimated mean flux for the group was higher than our standard dose group, providing some evidence that multiple early doses are advantageous. Interestingly, there was not a strong benefit from more than two doses. The period between those two doses appeared to be important, because the group had the largest within-group reduction in flux (36.9×) at 1-month compared to the first imaging time point, while the group had the smallest reduction (2.39×). Whether this timing makes a difference in a clinical context remains unknown.

We speculate that there could be an unintended consequence of using multiple LPC deliveries. It is feasible that some transgene-expressing cells are removed by subsequent LPC conditioning doses, which means that multiple LPC administrations might in fact be detrimental. In addition, regardless of the dosing regimen chosen, expression levels in all groups waned by at least one log by three months, potentially due to the turnover of terminally differentiated transduced cells [29]. Together this data suggests that further studies are needed to examine whether the benefit from multiple gene vector deliveries is outweighed by multiple LPC deliveries.

It is important to note that there was higher mortality in the and groups, due to the impact of the injectable anaesthetic and the greater challenges associated with LPC and LV delivery to the mouse lung compared to the nose. This likely limits the number of doses that can be delivered to mouse lungs at the early time points. An inhalable anaesthetic such as isoflurane could reduce this mortality, but its use in conjunction with airway conditioning can be complicated due to the one-hour interval between LPC conditioning and LV delivery required for optimal gene transduction. Finally, although BLI could not detect increased levels of luciferase expression over the 12-month period, the levels of flux we achieved in the lung in this study were orders of magnitude higher than those previously reported in the mouse nose and lung following multiple doses [14, 22]. Our choice of vector pseudotype and promoter could be responsible for these differences.

Our second hypothesis was that repeat-dosing over longer periods would help maintain gene expression over time, but our results suggest that using dosing periods of longer than seven days (e.g. and ) was detrimental to expression levels, compared to the standard dose group when measured in vivo and ex vivo. Our data was also in contrast to Sinn et al. who reported a dose-dependent increase in nasal luciferase expression following three, five, or seven 1× weekly doses [22]. Nonetheless, taken together, our in vivo and ex vivo data suggest that the group produces the highest levels of flux of all the dosing groups. The immunohistochemical analysis demonstrated the presence of GFP-positive cells in the conducting airways, which supports the data from our previous LV-LacZ reporter gene studies [10, 26].

In this study, we performed a sensitivity analysis to ascertain how to handle the BLI data from the animals that had no detectable flux after dosing with LV-Luc. This analysis was necessary because some animals had no detectable flux at a particular time point, but the flux at subsequent time points had returned to previous levels. This could have been due to the non-uniform localisation of D-Luciferin in the lung (see below), which may have confounded the findings. One approach was leaving these values as being at the limit of detection (100) but because this value was two to three logs lower than the other values this skewed the data sets. An alternative approach involved treating all BLI values at the detection limit as missing values, but this meant that animals that really had no luciferase expression were inappropriately excluded. As a balance between these two options, we chose to only set values as missing if they returned to above the detection limit.

The strength of this study is that it was a well-powered longitudinal assessment of a range of LV-vector lung dosing strategies over a long period of time (12 months). However, the study had some limitations. Although our chosen statistical analysis approach dealt with the flux values at the limit of detection, their presence could have been due to physiological variability in the D-luciferin nasal dosing approach. There may have been mismatches between the locations within the lung that were targeted by the LPC conditioning and vector delivered via an endotracheal tube, and the delivery of D-luciferin via nasal sniffing. However, due to welfare concerns it was not possible to intubate each animal at every BLI time point for D-luciferin delivery. Other groups have delivered D-luciferin via i.p. injection [22], but in our experience this produces much lower flux values. We also speculate that the sensitivity of the IVIS system for detecting small increases in flux from the repeat dosing schedules is low. In future studies, it would be advantageous to assess whether circulating antibodies (and other immune/cytokine responses) correlate with the decline in expression levels, particularly for the 5 × 1 m dosing schedule that had the largest reduction in gene expression at 12 months. In addition, the effect of the presence of the GFP transgene in the LV-Luc vector is unknown, and it is also possible that immune responses may be different if a higher titre vector is used. Finally, while our results and those of Sinn et al. and Griesenbach et al. are biologically and mechanistically interesting reporter gene studies, further development and testing with therapeutic CFTR gene vectors is required [14, 22].

The conclusions from our study are that repeat-dosing a VSV-G pseudotyped LV vector to the murine lung is feasible, but that longer repeat-dosing intervals are detrimental to expression levels compared to a dose. There was also some evidence to suggest that the short interval group produced higher ex vivo flux than our standard dose group. However, further detailed examination, potentially including the use of small non-human primate repeat-dosing feasibility studies, to model the responses of the human lung and immune system more accurately, is necessary for clinical development.