New England Journal of Medicine
November 19, 2015 Vol. 373 No. 21
http://www.nejm.org/toc/nejm/medical-journal

.
Editorials
Vaccine-Resistant Malaria
C.V. Plowe
[Free Full Text]
To ensure efficacy against wild poliovirus, Jonas Salk methodically classified circulating polio strains before choosing three to use in his inactivated vaccine.1 Several other successful vaccines against viruses and bacteria have likewise included immunogen variants that were selected through careful assessments of pathogen genetic diversity and strain-specific protective immunity.

In contrast, virtually all vaccines against the malaria parasite Plasmodium falciparum, including RTS,S/AS01, have been designed with the use of genetic sequences that are derived from a single, well-characterized reference strain thought to be of West African origin — 3D7. In light of the extreme diversity of malaria parasites, could this “freezer epidemiology” approach to vaccine development be one reason it has been so hard to make an effective malaria vaccine?

Phase 2 field trials of two malaria vaccines targeting highly polymorphic blood-stage antigens have shown strain-specific efficacy — that is, parasites that are genetically divergent from the strain used in the vaccine escaped its effect.2,3 However, field trials of RTS,S, which is based on the pre-erythrocytic P. falciparum circumsporozoite protein, showed no clear evidence of strain-specific efficacy.4,5 Circumsporozoite protein has polymorphism within two major T-cell epitopes, and although there is some variation in the immunodominant central repeat region, for decades it has been hoped that antibodies targeting a single dominant epitope based on the tetrapeptide repeat NANP would provide strain-transcending immunity. A molecular epidemiologic study showed no evidence of naturally acquired strain-specific immunity to different variants of circumsporozoite protein in African children,6 which added to the hope that RTS,S might not be threatened by “vaccine-resistant malaria.”7

These previous studies each examined malaria parasites in a few hundred children at a single African site and were constrained by the challenges of sequencing the long central repeat region of the circumsporozoite protein. Neafsey et al.8 now describe in the Journal how they used next-generation sequencing to measure strain-specific efficacy in a phase 3 trial of RTS,S/AS01 involving 15,000 children at 11 study sites across Africa. Parasite DNA was extracted from several thousand dried blood spots collected from children who were randomly assigned to receive the malaria vaccine or a control vaccine, and the efficacy of the vaccine against infection and clinical malaria with circumsporozoite protein identical to, or different from, the vaccine strain 3D7 was measured by means of a sieve analysis.9

The approach is straightforward: if vaccine-induced immunity is acting like a sieve that filters out parasites that are identical to the vaccine strain at immunologically important polymorphic amino acid positions, then less vaccine-type malaria should occur among vaccinated children, as compared with the control group. The large sample size was needed — vaccine-type parasites were in the minority at all study sites. Several predetermined analyses examining different variant components of circumsporozoite protein showed that, at least in older children, RTS,S/AS01 was indeed better at protecting against malaria caused by vaccine-type parasites. There was no such allele-specific efficacy in young infants, a group in which RTS,S efficacy was lower and in which there may have been different immune responses to both vaccination and malaria infection.

The effect of parasite genetic diversity on efficacy in older children was modest. Various measures of efficacy were approximately 10 to 15 percentage points lower when they were evaluated against non–vaccine-type parasites. This is a much smaller blow to efficacy than the more than 40-point deficit seen with a blood-stage vaccine.3

If we had a vaccine with 80% or 90% efficacy, we might be willing to accept a loss of this magnitude. But RTS,S/AS01 starts off with about 50 to 60% efficacy at best; the efficacy is lower in younger children, and it wanes over time. As this first malaria vaccine moves toward licensure, the results of this study should give pause to those considering whether, where, and when to deploy it. If RTS,S/AS01 is introduced into wide use, over time the loss of efficacy could be more profound than that seen during just a year of follow-up among children who are exposed to a large surrounding population of malaria parasites that are not under selection pressure from vaccine-induced immunity. The most prudent course may be to use RTS,S/AS01 in the specific populations that will benefit most, while redoubling efforts to improve this pioneering vaccine on the basis of new understanding provided by this study of the strain-specific basis of its partial efficacy.

The news is not all bad. The modest loss of efficacy against divergent parasites implies that RTS,S/AS01 offers some degree of cross-protection — vaccine escape is not complete. A multivalent version of RTS,S with carefully chosen circumsporozoite protein variants, possibly combined with additional antigens, might offer broader protection.10 The large set of molecular epidemiologic data on the prevalence of circumsporozoite protein variants across Africa that was generated for this study should provide the evidence needed to select a combination of strains for a more broadly efficacious, next-generation malaria vaccine.