Research Article: Licensed Dengue Vaccine: Public Health Conundrum and Scientific Challenge

Date Published: October 05, 2016

Publisher: The American Society of Tropical Medicine and Hygiene

Author(s): Scott B. Halstead.


A tetravalent live attenuated vaccine composed of chimeras of yellow fever 17D and the four dengue viruses (chimeric yellow fever dengue [CYD]) manufactured by Sanofi Pasteur has completed phase III clinical testing in over 35,000 children 2–16 years of age. The vaccine was recently licensed in four countries. During the first 2 years of observation, CYD vaccine efficacy ranged between 30% and 79% in 10 different countries with an overall efficacy of 56.8%. During year 3, there was an overall efficacy against hospitalization of 16.7%, but a relative risk of hospitalization of 1.6 among children younger than 9 years and 4.95 in children 5 years of age and younger. Vaccination of seronegative children resulted in universal broad dengue neutralizing antibody responses, but poor protection against breakthrough dengue cases. Unless proven otherwise, such breakthrough cases in vaccinated subjects should be regarded as vaccine antibody-enhanced (ADE). The provenance of these cases can be studied serologically using original antigenic sin immune responses in convalescent sera. In conventional dengue vaccine efficacy clinical trials, persons vaccinated as seronegatives may be hospitalized with breakthrough ADE infections, whereas in the placebo group, dengue infection of monotypic immunes results in hospitalization. Vaccine efficacy trial design must identify dengue disease etiology by separately measuring efficacy in seronegatives and seropositives. The reason(s) why CYD vaccine failed to raise protective dengue virus immunity are unknown. To achieve a safe and protective dengue vaccine, careful studies of monotypic CYD vaccines in humans should precede field trials of tetravalent formulations.

Partial Text

A six-decade-long effort to develop a dengue vaccine culminated in December 2015, with the licensing of a tetravalent live attenuated yellow fever chimeric dengue vaccine in Brazil, El Salvador, Mexico, and the Philippines.1–4 This vaccine, Dengvaxia, is a mixture of chimeric yellow fever and dengue viruses (DENV) 1, 2, 3, and 4. Each component was developed by inserting the structural genes for the premembrane and envelope proteins of each of the four DENVs into the genes of the capsid and nonstructural proteins of yellow fever 17 D vaccine virus.5 After a decade of preclinical development at OraVax (Cambridge, MA), then Acambis Inc., in 2005, the company and dengue vaccine were acquired by Sanofi Pasteur, who managed further development and clinical testing. For the purposes of this presentation, this vaccine is designated chimeric yellow fever dengue (CYD).

Sanofi Pasteur conducted extensive phase III efficacy trials of their CYD vaccine, now in their third–fourth year, involving over 35,000 children, 2–16 years of age, resident in 10 dengue-endemic countries.6 During the first 2 years of observation, CYD vaccine efficacy against mild-severe dengue disease ranged between 30 and 79 in 10 different countries with an overall efficacy of 56.8%.7,8 During year 3, there was an overall efficacy against hospitalization of 16.7% (65 hospitalizations in vaccinees, 39 in placebo group), but a relative risk of hospitalization of 1.6 among children younger than 9 years and 4.95 in children 5 years of age and younger.6 Vaccination of seronegative children resulted in universal broad neutralizing antibody responses but poor protection against breakthrough dengue cases.7,8 During year 3, clinical observations on vaccinated children and placebo controls showed the vaccine to be asymmetrically protective and enhancing, that is, some age groups were protected, whereas in others, disease accompanying breakthrough dengue infections was increased. A review of published data suggests that “all or nearly all” hospitalizations of vaccinated children over the 3-year postvaccination period may have occurred in children who were susceptible when vaccinated, and are attributed to vaccine ADE.9

CYD developers have provided hypotheses to explain the third-year clinical responses observed during breakthrough DENV infections in vaccinated children.10 Summaries of these hypotheses receive comment here:

The empiric nature of dengue vaccine development and the requirement for individual testing of each monovalent element, including, critically, demonstration of protection using a human challenge model has received prior discussion.24 Since that review, important new understanding has been gained concerning the structure and function of monovalent DENV neutralizing antibodies.25 But, does tetravalent CYD vaccination raise these “protective” monovalent antibodies? Evidence of a protective role of T-cells in human DENV infections has been mounting.26 T-cells directed at nonstructural protein epitopes similar to those observed in individuals infected with wild-type DENV have been detected in recipients of the live attenuated tetravalent dengue vaccine produced by the National Institutes of Health.27 Might T-cell immunity after administration of CYD vaccine be blunted because yellow fever not DENV nonstructural proteins are presented? Finally, must an effective dengue vaccine contain dengue NS1? An analogy has been observed between the function in humans of bacterial lipopolysaccharides (LPSs) and that of DENV NS1.28 Both compounds interact with toll-like receptor 4 (TLR4) on the surface of monocytes, macrophages, and endothelial cells to induce the release of a range of cytokines and chemokines. These same mediators have been identified in the blood of patients with dengue hemorrhagic fever (DHF)/dengue shock syndrome (DSS). NS1-mediated cytokine release was inhibited by the TLR4 antagonist LPS-Rhodobacter sphaeroides suggesting an avenue for therapeutic intervention. Crucially, this same observation has been confirmed in an in vivo model. DENV 2 NS1 inoculated intravenously at physiologically relevant concentrations in sublethal DENV 2-infected interferon-α/β receptor −/−C57BL/6 mice produced lethal vascular permeability.29 In vitro, NS1 when added to cultured human endothelial cells resulted in endothelial permeability and disruption of endothelial cell monolayer integrity. These observations suggest that DSS may be a viral protein toxicosis. It was further shown that vaccination of mice with DENV 2 NS1 protected against endothelial leakage and death due to lethal DENV 2 challenge. Immunization with DENV 1, 3, and 4 NS1 proteins partially protected against heterologous DENV 2 challenge. The successful prevention of death in mice due to DENV by immunizing with NS1 was established long ago and repeated many times.30,31 Is it possible that the CYD fails to protect seronegatives because it does not contain DENV NS1 antigens?

The CYD tetravalent vaccine is closely followed in clinical testing by two additional live attenuated tetravalent dengue vaccines, each containing chimeric viruses; the Takeda vaccine contains chimeras of DENV-1, 3, and 4 on a DENV-2 backbone,35 whereas the tetravalent National Institute of Allergy and Infectious Diseases vaccine contains a DENV-2/4 chimera.36 To assure that these chimeras are fully protective, the efficacy trial design must be changed.

Clearly, any dengue infection occurring in a subject receiving a dengue vaccine is evidence of vaccine failure. Clinically, however, the mild disease that accompanies primary dengue infections in children may exactly mimic dengue disease occurring in the presence of antibodies raised by the vaccine. Vaccination may raise mixtures of protective and enhancing antibodies, and with the passage of time, this balance may shift toward enhancement. Categorizing mild disease as “vaccine-enhanced” requires evidence of a statistically significant increased rate of such disease among vaccinated compared with appropriate placebo controls. Because hospitalized disease, DHF, or severe dengue seldom accompany primary dengue infections in children, when such cases occur among vaccinated individuals, they should be recognized as serious adverse events, that is, vaccine-enhanced dengue disease.

On the basis of the described protection of seropositives and reduction of severe disease in the CYD phase III clinical trials, it is expected that one or more doses of vaccine will efficiently protect the high-risk group—monotypic DENV-immunes—from acquiring disease when exposed to DENV.6 The duration of this protective immunity is unknown. Neither is it known whether vaccination of monotypic immunes prevents subsequent DENV infection and viremia. Today worldwide, DENV infections of susceptible and monotypic-immune adults contribute to the pool of virus in circulation. According to Sanofi, CYD will not be given to children under the age of 9 years. Children in this age group should sustain community DENV force of infection at fairly high rates well into the future even in the face of massive vaccination of individuals 9 years of age and older. High rates of immunization of individuals 9 years of age and older can be expected to reduce illness burden substantially, but only among monotypic immunes. In Mexico, where a large percent of vaccinated population is seronegative (∼47%), overall protective efficacy was low (31.3%).40 The low efficacy must be the result of high rates of vaccine-enhanced DENV disease. The challenge confronting public health officials is how to immunize those who will benefit from vaccination but shield those who are at risk to vaccine-acquired enhanced DENV disease. To do this, it has been suggested that the vaccine be given only to dengue immunes.9 Manufacturers, regulators, and public health authorities must grapple with this question to find an equitable, affordable, and ethical solution.