Recent data obtained with the live-attenuated tetravalent dengue CYD-TDV vaccine showed higher protective efficacy against dengue virus type 4 (DENV-4) than against DENV-2. In contrast, results from previous studies in nonhuman primates predicted comparable high levels of protection against each serotype. Maximum viral loads achieved in macaques by subcutaneous inoculation of DENV are generally much lower than those observed in naturally dengue virus-infected humans. This may contribute to an overestimation of vaccine efficacy. Using more-stringent DENV infection conditions consisting of the intravenous inoculation of 10⁷ 50% cell culture infectious doses (CCID₅₀) in CYD-TDV-vaccinated macaques, complete protection (i.e., undetectable viral RNA) was achieved in all 6 monkeys challenged with DENV-4 and in 6/18 of those challenged with DENV-2, including transiently positive animals. All other infected macaques (12/18) developed sustained DENV-2 RNAemia (defined as detection of viral RNA in serum samples) although 1 to 3 log₁₀ units below the levels achieved in control animals. Similar results were obtained with macaques immunized with either CYD-TDV or monovalent (MV) CYD-2. This suggests that partial protection against DENV-2 was mediated mainly by CYD-2 and not by the other CYDs. Postchallenge induction of strong anamnestic responses, suggesting efficient vaccine priming, likely contributed to the reduction of DENV-2 RNAemia. Finally, an inverse correlation between DENV RNA titers postchallenge and vaccine-induced homotypic neutralizing antibody titers prechallenge was found, emphasizing the key role of these antibodies in controlling DENV infection. Collectively, these data show better agreement with reported data on CYD-TDV clinical vaccine efficacy against DENV-2 and DENV-4. Despite inherent limitations of the nonhuman primate model, these results reinforce its value in assessing the efficacy of dengue vaccines.

IMPORTANCE The nonhuman primate (NHP) model is the most widely recognized tool for assessing the protective activity of dengue vaccine candidates, based on the prevention of postinfection DENV viremia. However, its use has been questioned after the recent CYD vaccine phase III trials, in which moderate protective efficacy against DENV-2 was reported, despite full protection against DENV-2 viremia previously being demonstrated in CYD-vaccinated monkeys. Using a reverse translational approach, we show here that the NHP model can be improved to achieve DENV-2 protection levels that show better agreement with clinical efficacy. With this new model, we demonstrate that the injection of the CYD-2 component of the vaccine, in either a monovalent or a tetravalent formulation, is able to reduce DENV-2 viremia in all immunized animals, and we provide clear statistical evidence that DENV-2-neutralizing antibodies are able to reduce viremia in a dose-dependent manner.