The purpose of this study was to assess the utility of the AAV-TF-RhEpo-2.3w vector, at two doses (1.5×1011 vg and 1.5×1012 vg/parotid gland), to direct clinically adequate, tightly controlled expression of RhEpo in the serum of rhesus macaques. As shown in vitro, no expression of RhEpo resulted unless transduced cells were treated with rapamycin (). Five animals completed the in vivo study, but rapamycin-regulation of the vector was unable to achieve significant elevations in serum levels of RhEpo following parotid gland transduction. However, in the single animal that developed pneumonia and became anemic (see below), rapamycin administration was followed by a large increase (~3.7-fold) in the serum RhEpo level detected.
The overall weak production of RhEpo in healthy macaques was unanticipated, as we previously showed that a conventional, unregulated, AAV2 vector (rAAV2RhEPO), used at 20% of the high dose employed herein, readily transduced macaque parotid glands and led to significant increases in serum RhEpo levels (Voutetakis et al, 2007a
). Accordingly, the present results are unlikely to be related to use of AAV2 vectors per se in this tissue. However, there are two major genomic differences between the rAAV2RhEPO vector used previously and AAV-TF-RhEpo-2.3w that theoretically could have contributed to the disparity in RhEpo production observed: (i) the promoter utilized to drive RhEpo expression and (ii) the actual size of the genome. While both should have been considered more when determining the vector doses used herein, it is likely that the first reason was most significant. rAAV2RhEPO uses a promoter from Rous sarcoma virus that directs high levels of sustained transgene expression in salivary glands (Voutetakis et al, 2004
; Zheng and Baum, 2005
). Conversely, AAV-TF-RhEpo-2.3w uses a relatively weak promoter, derived from interleukin-2. Secondly, the genome of rAAV2RhEPO is ~2.3kb in size, while that of AAV-TF-RhEpo-2.3w is ~4.7kb. The latter is roughly comparable to the genome size in wild type AAV2, albeit tightly packed with exogenous components. A similar AAV2 vector (rAAV-TF2.3-hEPO, expressing human Epo; Wang et al, 2006
) worked well after delivery to murine submandibular glands. The dose used in that study was ~3.3×1011
vg/kg, i.e., a dose roughly comparable to the high dose used here (~2.5 ×1011
vg/kg). Thus, it does not seem that a >4kb, rapamycin-regulated AAV2 vector cannot effectively transduce salivary glands per se. Additionally, an AAV2 vector that was virtually identical to the AAV-TF-RhEpo-2.3w vector used here led to high serum RhEpo levels following transduction of non-human primate liver at doses similar to those used herein (Rivera et al, 2004
). While it is possible that the use of a higher vector dose would lead to clinically adequate serum RhEpo levels from transduced healthy macaque parotid glands, we conclude that the macaque parotid is apparently not as good a target for the AAV-TF-RhEpo-2.3w vector as the liver.
The present study, however, demonstrates that an AAV2 vector, delivered directly to a parotid gland at a significantly higher dose (~1.5×1012 vg) than used in previous salivary gland studies, does not by itself result in any untoward effects. The five macaques that completed the study had essentially normal CBC and serum chemistry values over the course of the study, with no consistently abnormal values observed. In addition, none of these animals exhibited any purulence in their saliva. All displayed normal patterns of food consumption and were able to gain weight. Extrapolating the highest dose used herein (~2.5 ×1011 vg/kg) to dosing in humans, our results suggest that administration of an AAV2 vector at a dose of up to ~1.5 ×1013 vg would be well tolerated in a single parotid gland.
Despite the general safety of the AAV-TF-RhEpo-2.3w vector, during this study there was a severe adverse event in one macaque: animal #6851 developed pneumonia in his left lung that eventually required euthanasia. Although extensive post-mortem evaluations were performed, the etiology of the pneumonia could not be determined. There was no evidence of aspiration pneumonia grossly, or in the lung tissue sections examined histologically, but the possibility of aspiration inciting the abscess, e.g., as a result of collecting whole saliva on anesthetized animals, cannot be ruled out. This was the first adverse event that we have seen following the use of a viral vector (either Ad5 or AAV2) in macaque parotid glands (Voutetakis et al, 2007a
, Voutetakis et al, 2009
). In our previously reported studies we employed a total of 20 macaques. Thus, when including the six animals studied herein, only #6851 of 26 total macaques treated experienced a study-related adverse event of any type. Animal #6851 may just have responded idiosyncratically. In this regard it is interesting to note that in a previously reported study using quite high levels of AAV vectors encoding RhEpo administered intramuscularly to eight cynomologous macaques, two developed a severe autoimmune anemia, while the other six were polycythemic (Gao et al, 2004
). It is also important to recognize that all macaques had been administered rapamycin twice prior to day 37 when #6851 first showed signs of the adverse event (i.e., on days 14 [albeit in an inappropriate diluent] and 28). It is therefore possible that the pneumonia that developed in this animal was in part related to transient immunosuppressive effects of rapamycin.
Additionally, the QPCR biodistribution study performed with selected tissues from #6851 represents the first time that we have found a significant level of a parotid gland-administered AAV2 vector outside the targeted gland in a large animal model (macaques and miniature pigs; Voutetakis et al, 2007a
; Hai et al, 2009
). At necropsy two tissues had high and roughly similar levels of vector copies present: the targeted right parotid gland and the liver. A general confounder to understanding these results is that the necropsy was conducted under standard, not GLP (good laboratory practice) conditions, i.e., separate instruments were not used for each tissue and the obtained tissues were not removed in sequence, from least likely to have vector present to most likely presence (i.e., targeted parotid gland; see Voutetakis et al, 2007a
). Accordingly, the possibility of some inadvertent tissue contamination cannot be unequivocally excluded.
However, finding a fairly high vector level in the liver strongly suggests that AAV-TF-RhEpo-2.3w had entered the bloodstream in this animal, as liver is the primary target for an AAV2 vector delivered intravascularly (e.g., Xiao et al, 1998
; Nakai et al, 2000
). The most likely reason for that occurrence would be by trauma while delivering the vector, e.g., damaging the cannulated duct. Since the targeted right parotid gland itself appeared normal on histopathological and gross examination, it seems there was no significant inflammatory reaction in the gland from which vector could have been lost and entered the bloodstream.
In summary, despite the occurrence of a severe adverse event in one animal, the present study has shown that administration of high doses of AAV-TF-RhEpo-2.3w to single parotid glands of healthy macaques was generally safe. This vector, however, was unable to direct the expression of clinically significant levels of serum RhEpo, in a rapamycin-dependent manner, except in the one sick animal.