In order to support the clinical development of the MV1-F4 vaccine, we have compared the toxicology, biodistribution and shedding profiles of the HIV-1 candidate vaccine MV1-F4 to those of the parental Schwarz vaccine strain, using a cynomolgus macaque model. In addition, the results may serve to increase our knowledge of the toxicity or in vivo tropism of MV vectors or vaccine strains, which has been the focus of few preclinical studies to date (e.g., de Vries et al. 2010
; Lemon et al. 2011
; Myers et al. 2007
; Peng et al. 2003
Our data show that no shedding of infectious virus was observed for either of the two vaccines. Furthermore, no toxic effect in relation to the MV vaccination was found with these vaccines. One or three IM injections of a full human dose of MV1-F4 or of the reference vaccine were well tolerated and induced no clinical symptoms of local or systemic reactogenicity. For both vaccines, virus replication was predominantly observed in secondary lymphoid organs (including spleens, Peyer’s patches and all major lymph nodes) and to a lesser extent in epithelium-rich tissues (including intestine, larynx, trachea and urinary bladder) and liver, which were thus designated as potential target organs for intrinsic toxicity of the vaccines. However, no gross or histopathological effects indicative of toxicity were observed in these target organs, or in any other organs. Mean spleen weights were increased after three doses of either vaccine, which corresponded in some animals with enlargements of the white pulp, due to generation or growth of germinal centers. This effect likely resulted from immune activation of the secondary lymphoid organs by the vaccines, which has also been observed in other preclinical studies (Sheets et al. 2006
; Speijers et al. 1988
). It is less likely that local virus replication in these organs has caused the increased spleen weights, as no clear relationship between increased spleen weights, enlargements of the white pulp and the presence of viral RNA in the spleens could be established. All other observations were considered to be unrelated to treatment, as they occurred incidentally, without correlation between genders, or at similar frequencies in control and immunized animals.
The occasional reductions in food consumption had no impact on body weights and were not considered to be indicative of systemic toxicity. For DNA plasmid vaccines, it has been observed that food consumption data and body weights did not correlate with toxicity or other findings (Sheets et al. 2006
). A similar conclusion could potentially be drawn for recombinant live-attenuated vaccines such as MV1-F4.
The only clinical pathological finding for MV1-F4 treatment groups was a slight but statistically significant decrease of the aPTT (but not of the PT) for males on day 56. However, shortening of the aPTT has generally little clinical relevance and may be associated with suboptimal sample collection or processing (Adcock et al. 1998
; Awad et al. 2006
). Therefore this is not considered indicative of coagulation abnormalities of toxicological relevance. We also observed that on day 85, the mean CK activity for Rouvax-treated males was significantly higher than for control males. Elevated CK values are used as a marker of injury, heart attack, severe muscle breakdown, muscular dystrophy or acute renal failure (Lott and Abbott 1986
). None of these conditions were observed for the two males involved, and urinalysis showed no treatment-related effect for either vaccine. In addition, no elevated CK activity was found in these animals on previous time points. Possibly these elevations reflect the effects of IM-delivered sedation for ophthalmology on day 84. This effect is commonly seen in preclinical studies (Stewart et al. 2008
) and is therefore not considered to be an adverse reaction to treatment.
Cell tropism for vaccine and wild-type MV strains is largely determined by virus entry (Tatsuo et al. 2000a
). Wild-type MV, which use predominantly SLAM/CD150 expressed on activated immune cells (Tatsuo et al. 2000b
), is known to mainly replicate in lymphoid organs and epithelial tissues such as the skin, respiratory tract, intestine and urinary bladder (Griffin 2007
; Takeda 2008
). Recent studies suggest that after aerosol infection, wild-type MV initially targets macrophages and dendritic cells in alveolar tissues, which is followed by replication in regional lymph nodes and systemic spreading to lymphoid organs (Lemon et al. 2011
). How MV infects CD150-negative epithelial cells remains largely unclear, although putative epithelial receptors have recently been described (Watanabe et al. 2010
). As MV vaccine strains use the ubiquitous CD46 receptor in addition to the receptors used by wild-type MV (Dorig et al. 1993
), the vaccine strains were expected to exhibit a wider tropism than that documented previously for wild-type MV (Griffin 2007
; Lemon et al. 2011
; Takeda 2008
). However, in macaques infected intratracheally or by aerosol with wild-type or attenuated (Edmonston tag) recombinant MV, only the pathogenic MV caused significant viremia and widespread distribution. The attenuated MV had a restricted systemic spread and was rarely detected in lymphoid tissues (de Vries et al. 2010
), in contrast to observations in the current study. Our data seem to be more aligned with studies in mice immunized intraperitoneally or intravenously with an attenuated oncolytic (Edmonston) recombinant strain, in which infected cells were mainly detected in secondary lymphoid organs, liver and lungs (Myers et al. 2007
; Peng et al. 2003
). These disparities between studies may be explained by differences in administration routes, MV strains or methods of MV detection.
The RT-qPCR results for tissues and biological fluids confirm a peak of replication around 10 days after the first dose, supporting other studies in monkeys with wild-type or attenuated MV strains (Auwaerter et al. 1999
; Pan et al. 2005
; Permar et al. 2003
). This resembles the clinical situation, as vaccine-induced peaks are known to occur a few days before those induced by natural infection taking place between 11 and 14 days after exposure (Strebel et al. 2004
). We observed that viral clearance for both vaccines was not fully completed at 28 days after the third dose, since viral RNA was still detectable in lymphoid and epithelium-rich tissues. This is consistent with an earlier report showing that MV RNA remained detectable in monkeys for 4–5 months (in this case, in PBMCs), even though clearance of viremia occurred within 14 or 29 days (Pan et al. 2005
An additional study objective was to assess the potential of virus shedding of the MV1-F4 vaccine (in parallel to that of the reference vaccine). In natural infection, MV can be isolated from urine up to 10 days after rash onset (Gresser and Katz 1960
), and viral RNA can be detected in PBMC and nasopharingeal specimens up to 100 days after rash onset (Riddell et al. 2007
). The few known studies of shedding by attenuated MV vectors report detection of (Edmonston) viral RNA in buccal swabs from monkeys (after intravenous administration; Myers et al. 2007
), but no detection in human urine or saliva (after intraperitoneal administration; Galanis et al. 2010
). Consistent with reports of detection of viral RNA from attenuated MV vaccine strains in human urine up to 14 days after vaccination (Rota et al. 1995
), we detected MV1-F4-derived RNA in urine at day 11. Infectious viral particles from MV vaccines have rarely been found in human throat or nasopharyngeal secretions, and there is only one reported case of isolation of (Schwarz) MV vaccine virus from throat swabs (Morfin et al. 2002
). While we detected viral RNA from both vaccines at the peak of viremia in a few samples such as throat and nasopharingal swabs, none of these contained infectious virus at a LOD of 10 CCID50
/well (with 95 % probability). Therefore, transmission of MV1-F4 virus between persons is considered unlikely to occur, and has also never been documented previously for any of the current measles vaccines (WHO 2009
The immunogenicity of the MV1-F4 candidate vaccine was not investigated in detail in the current study, as our study was primarely designed to assess the toxicology, biodistribution and shedding profiles of the candidate vaccine. However, immune responses directed against the F4 transgene were characterized in a separate study using the same animal model (MV-seronegative cynomolgus macaques). In the latter study, the MV1-F4 candidate vaccine was shown to induce F4-specific CD4+ and CD8+ T cell responses, as well as antibody responses to F4 (manuscript in preparation).
In conclusion, while viral dissemination was observed in various organs, no shedding of infectious viral particles was noted, and no toxic effect in relation to the MV vaccination was found following one or three IM injections with a clinically relevant dose of MV1-F4, with the same outcomes for the (Schwarz) measles comparator vaccine. Moreover, either vaccine virus replicated predominantly in secondary lymphoid organs and, to a lesser extent, in epithelium-rich tissues. Thus, as expected, introduction of the F4 transgene did not change the toxicological profile, shedding capacity or tropism of the parental strain.