Unique challenges faced by the clinical evaluation of dengue vaccines

Keywords::

• Aedes aegypti
• dengue fever
• dengue hemorrhagic fever and shock syndrome
• dengue virus
• flavivirus
• Phase I–IV clinical trials
• vaccines

The spread of dengue virus via its Aedes mosquito vector throughout most of the tropics and subtropics has led to a worldwide explosion of clinical dengue. Infection by any one of the four dengue virus (DV) serotypes can produce a broad spectrum of illness, including asymptomatic infection, mild febrile illness, classic dengue fever, and the lethal dengue hemorrhagic fever (DHF) and shock syndrome. During the past 50 years, dengue has evolved into one of the world’s major infectious diseases [1]. Over 3.5 billion people are at risk of dengue in over 100 countries, and more than 50–100 million infections occur each year. Dengue is now a leading cause of morbidity in American and European travelers and military personnel, rivaling or exceeding malaria in many countries [2]. No effective antiviral treatment for dengue infection exits. Prevention has relied upon mosquito control, which has been largely ineffective. For all these reasons, policy makers have identified the critical need for safe and effective dengue vaccines [3,4].

Unfortunately, vaccine development efforts have been hampered by two unique features of dengue biology and epidemiology; namely, the existence of multiple DV serotypes and the tendency for a second, heterotypic DV infection to provoke severe or life-threatening illness after a priming infection. Some of these vaccine impediments will be discussed in this article. Comments are limited to live-attenuated tetravalent dengue vaccine (TDV) candidates, because they have entered field trials and some immunogenicity and safety data are available [5]. Subunit and whole killed DV vaccines are, generally speaking, in preclinical development.

Need for a tetravalent dengue vaccine

There is a general consensus among dengue and public health experts that a multivalent dengue vaccine consisting of the four dengue serotypes will be required [6]. This requirement has complicated and slowed vaccine development. Taking into consideration its background, the need for a TDV is supported by three key epidemiological and immunological facts. First, from a limited series of dengue challenge experiments in volunteers carried out in 1944 by Albert Sabin, it appears that a single serotype can provide short-term heterologous protection limited to 3–4 months against the other three serotypes, but homologous immunity can persist for years [7]. Second, only a tetravalent, rather than a monovalent or bivalent vaccine, would reliably protect DV-exposed persons in any locale, because one or several DV serotypes can circulate simultaneously and serotypes can change unpredictably from one season to the next. Third, DV infections are more severe in individuals who have acquired dengue antibodies actively from a previous heterologous DV infection and are apparently mediated by nonneutralizing heterospecific antibodies that are residual from that earlier infection [8,9]. Antibody dependent enhancement (ADE) has provided an explanatory hypothesis whereby pre-existing, cross-reactive DV antibodies facilitate DV entry and replication within target cells, thereby increasing virus burden [10,11]. A massive cellular release of virus and soluble DV protein triggers complement activation, which together with inflammatory cytokines, synergize locally to trigger the vascular leak observed in severe dengue [12]. The secondary infection hypothesis and ADE suggests that dengue vaccines must induce protective neutralizing antibodies to all four serotypes simultaneously rather than sequentially to avoid enhancement of natural dengue illness in incompletely vaccinated individuals. A tetravalent vaccine that would confer only suboptimal immunity to multiple serotypes, or one that allows a fall-off in neutralizing antibody, leaving in its wake residual, non-neutralizing enhancing antibody, could theoretically prime an individual for DHF upon natural DV exposure. Although ADE is a theoretical concern, to date no evidence exists that DV vaccines actually promulgate clinically recognizable ADE. On the contrary, there is one limited, but reassuring study, in which 140 Thai children followed for 3–8 years after immunization with an attenuated tetravalent vaccine experienced multiple subsequent dengue infections without apparent clinical or serologic enhancement [13].

An in vitro quantitative ADE assay did not uncover any apparent risk of in vivo sensitization in this Thai cohort of vaccinated children [14]. In addition, the risk of acquiring a third or fourth DV infection is considerably reduced after two infections [15], which suggests that, in theory, neutralizing immunity to only two or three serotypes may be sufficient to protect against the remaining serotypes, based on long-lived, cross-reacting heterologous immune priming [16]. However, it should be noted that ADE and clinical enhancement of DV is complex and incompletely understood. There is a growing consensus that in addition to ADE, age, gender, DV strain virulence, sequence of infecting DV serotypes, interval between infections, genetic propensity of the host, and CD4⁺ and CD8⁺ mononuclear cell responses all modulate the development of severe dengue and DHF. Clearly, we still have a lot to learn.

Viral interference in TDV formulations

The generation of a balanced immune response with multi-typic viral vaccines has historically been difficult. Immunological interference, intrinsic differences in viral fitness, and/or the suppression or enhancement of replication of one virus by another viral component of the vaccine has markedly lengthened the time required to construct a balanced formulation [17,18]. The generation of robust, balanced antibody responses to all four DV serotypes following vaccination with three different live-attenuated vaccine candidates has proven no exception [19–24]. Currently, no attenuated vaccine candidate has conferred robust immunity to all four DV serotypes with a single dose; two or three doses have been necessary. We now realize that timing between doses must allow 4–6 months for the cross-reactive antibody to wane so that vaccine replication and the resulting antiviral responses are adequate to all serotypes after the second or third vaccination. Vaccinees living in dengue-endemic areas during these 4–6 month vaccination intervals, or persons who receive only the first dose of a multidose schedule, must be followed carefully to assure that incompletely protected volunteers are not predisposed to develop more severe dengue after natural infection. To this point, an epidemiological study of two Cuban epidemics of sequential dengue-1 and dengue-2 infections revealed that the epidemic with the 20-year interval between serial infections was associated with more severe illness than the epidemic with the 4-year interval [25]. Although this is only one study in one locale, the results suggest that many years of follow-up may be required to assure the safety of vaccine-induced immunity. Nevertheless, I suspect that waning immunity induced by attenuated vaccine strains will be qualitatively different than the waning immunity induced by fully virulent DV; however, this hypothesis requires formal testing in Phase II, III and IV post-licensing trials [26]. Because the antibody response to TDV immunization is more robust in the setting of pre-existing DV, Japanese encephalitis virus (JEV) or yellow fever virus (YFV) immunity [27], I would predict that interference between DV vaccine components would be diminished in populations with prevalent immunologic priming caused by wild flavivirus infections or by JEV or YFV vaccination. Again, TDV field trials must be designed to confirm this postulate.

Serological assays & correlates of TDV efficacy

An unavoidable limitation of DV serological assays is that a specific antibody response is generated against the infecting serotype as well as cross-reactive, transient responses against other DV serotypes. The inability of the 50% plaque reduction neutralization test (PRNT₅₀) to discriminate these cross-reactive antibodies from infection-derived, protective antibodies complicates the interpretation of immunogenicity data obtained in TDV trials. Because the protective antibody titer for any of the DV serotype infections is unknown, the characterization of the optimal immune response to vaccination therefore remains to be defined. Whereas a PRNT₅₀ of 1:10 or less has been accepted as a surrogate of efficacy for the JEV and YFV flavivirus vaccines [28], protective titers for DV may be much higher and may vary among serotypes [29]. In the instance where cross-reactive and specific responses cannot be distinguished, the optimal immune response may simply be defined as one that is broad and of a high titer. Large field efficacy trials of TDV candidates will be needed to correlate protection with the preinfection antibody titer against the subsequent infecting DV serotype. Although it may be difficult to obtain adequate numbers of correctly timed serological specimens to permit such a correlation, successful attempts to do so would greatly simplify the testing and approval of second-generation vaccines [26].

Additional TDV trial concerns

Other issues include, first, the need to conduct trials in both Asia and the Americas owing to different dengue epidemiology, DV strain variation and clinical outcomes. Second, the annual variability in the proportion of DV serotypes circulating in study areas, which complicates adequate sample size calculations needed to achieve desirable statistical power to detect vaccine efficacy against any one DV serotype. Third, measurement of possible enhancing or detrimental effects of prior flavivirus exposure on TDV performance, which would likely require large post-licensure trials. Fourth, the calculation of the vaccine coverage rate necessary to interrupt mosquito transmission of DV in order to achieve herd immunity. And finally, building an infrastructure of national regulatory authorities, ethical research committees, pharmaceutical companies, academic institutions and field sites with sufficient laboratory capacity and clinical trial experience to carry such a large effort forward [5,26].

This article highlights but a sample of the more pressing and unresolved virology, vaccinology and public health questions surrounding live-attenuated TDV trials. Despite real and theoretical concerns, TDVs have entered clinical trials for immediate and long-term safety and efficacy in Latin America and Southeast Asia. The pipeline of dengue vaccines looks promising for the first time in 60 years.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.