Vaccines have prevented more disease than any other modality, except clean water. This point was made in the first edition of Vaccines,1 the textbook that is now in its sixth edition2 and has grown in size and associated content. In this brief article, it is impossible to discuss all vaccines, so I will discuss vaccines against only four viruses: rubella, rabies, rotavirus and cytomegalovirus (CMV). I was involved in the development of all of these. These four vaccines cover the spectrum of public health, in as much as rubella vaccine is in the process of eradicating the disease, rabies vaccine could do so if it were more generally available, rotavirus vaccine is just now being applied and CMV vaccine is under development.
Interest in rubella began in the 1940s with the observation by an Australian ophthalmologist, Norman Gregg, that infection in pregnant women led to congenital abnormalities in their infants, including cataracts, deafness, congenital cardiac malformations and mental retardation.3 The problem of congenital rubella syndrome (CRS) was soon recognized in other parts of the world,4 but the need for a vaccine became obvious when a large epidemic of rubella began in Europe in 1962 and spread to the USA.5 In the USA, thousands of babies were damaged and rubella became a feared complication of pregnancy.6
A number of groups went to work to develop a rubella vaccine, including, notably, one group at the US Food and Drug Administration and another at GlaxoSmithKline. At the Wistar Institute in Philadelphia, PA, USA, the use of diploid fibroblast cell strains developed from human fetuses had just begun,7 and it was soon realized that those cells were free of contaminating agents, in contrast to cells obtained from free-living monkeys and other animals. Accordingly, we set about to attenuate rubella virus in WI-38, a human diploid cell. A rubella strain was isolated from one of the fetuses sent to my laboratory after an abortion requested by the parents because of maternal rubella. The virus was isolated directly in WI-38 and adapted to a level suitable for good replication by serial passage in those cells.8,9 After a limited number of passages at 37 °C, the incubation temperature was lowered to 30 °C.10 This had to be carried out because we had learned from work with polioviruses that adaptation to growth at low temperatures selected for attenuated mutants. Indeed, clinical trials conducted at various passage levels confirmed increasing attenuation and by the twenty-fifth passage, it appeared that a good level of attenuation had been obtained without damaging immunogenicity.11
At this stage, trials were carried out in larger numbers of subjects under conditions of natural exposure to rubella. The results were gratifying in that seroconversion occurred in ≥ 95% of those vaccinated and, when they were exposed to rubella, protection was seen in about 95% (Table 1).12,13,17 The vaccine strain, called RA 27/3, was first made in Europe by Wellcome Labs and the Institut Mérieux, but later was adopted by Merck (West Point, PA), GlaxoSmithKline (Rixensart, Belgium) and the Serum Institute of India (Pune, India). It is usually incorporated with measles and mumps strains to be given in the triple vaccine, measles, mumps and rubella (MMR).2
|Reference||Population studied||Number of vaccines exposed||Protective efficacy|
|Beasley et al. 196912||Primary schools, Taiwan||198||99.5%|
|Furukawa et al. 197013||Boys’ school, Japan||24||100%|
|de Valk and Rebiere 199814||Primary school, France||119||95%|
|Greaves et al. 198315||High school children, USA||> 600||90%|
|Davis et al. 197116||Institution USA||22||100%|
Reactions to RA27/3 are unusual in children and most reactions after MMR are due to the measles component. In adult women, about 25% will have arthralgia or arthritis, but such phenomena are transient and do not result in long-term sequelae.18,19
Two initial strategies in the use of rubella vaccine were suboptimal. In the USA, vaccination was launched in infants, whereas in the UK the target was prepubertal girls. The first strategy failed because pregnant women could still be infected by their male partners or other adults carrying the rubella virus and the second strategy failed because girls not vaccinated were still exposed to virus circulating in children.20,21 Moreover, the infant vaccination strategy decreased circulation of the virus sufficiently to allow older girls to reach child-bearing age without immunity, when they could be exposed to other infected adults. This could lead to a paradoxical increase in CRS, at least temporarily until immune cohorts became adults.22 Therefore, it became clear that combined vaccination strategies were necessary, involving both vaccination in infancy and vaccination of older girls. Indeed, both boys and girls are vaccinated in infancy to reduce circulation of virus and to provide later immunity in women. Prepubertal girls are targeted to give them immunity later in life and women at the end of a pregnancy are vaccinated to give them immunity in a subsequent pregnancy. Some countries, particularly in Latin America, have also vaccinated adult men in order to protect their women, with the expected result that the rubella virus disappears.23,24
The application of rubella vaccination has resulted in elimination of endogenous infection from North America, Latin America and indeed from the entire Western hemisphere. In Europe, Scandinavia is now free of the rubella virus and the UK has enjoyed periods of little or no CRS. Serious outbreaks of rubella in the Pacific Islands have caused the Western Pacific Region to install routine rubella vaccination with the aim of eliminating the virus. Many countries in Asia have begun to vaccinate against rubella and even in Africa some countries have instituted rubella vaccination through the use of the measles and rubella or MMR vaccinations. As there is no animal reservoir of rubella virus and there is no reason why eradication of the virus cannot be accomplished; therefore, we hope that CRS will become a disease of the past.25–27
Rabies has been called the incurable wound, and in fact few people infected with the rabies virus survive. Those who do have usually been exposed to strains from bats, which may be less virulent than strains from other animals.28 However, in the great majority of rabies cases in humans, the biting animal is a dog, and 50 000 deaths are reported annually, almost all in those who did not receive post-exposure prophylaxis.29–31 The reasons why the vaccine was not given are mainly the cost and complexity of the regimen, failure to appreciate the risk of rabies and lack of availability of the vaccine. In addition, exposure to rabies often occurs in children, who may fail to report animal bites.
The earliest rabies vaccines were made by injection of virus into the central nervous system of laboratory animals and then collecting the infected brain material for injection.28 The brain tissue could cause severe autoimmune reactions and the low titre of the virus necessitated as many as 21 injections.32,33 Adaptation of the virus to duck embryos also resulted in low-titre material that required many injections.34 The advent of cell culture allowed for growth of virus to high titres, as identified by Tad Wiktor in Hilary Koprowski’s laboratory.33 The best cell appeared to be the same human diploid cells chosen for rubella virus attenuation.7 I became associated with Wiktor and Koprowski in the effort to show that immunization with rabies virus grown in cell culture and then inactivated was safe and could induce high titres of protective antibody against the virus and protect against exposure.35,36
Volunteers were obtained from laboratory personnel (including ourselves), veterinary students and others potentially exposed to the rabies virus. The results were phenomenal in that reactions to the vaccine were rare (later traced to the human albumin used as a stabilizer) and the neutralizing antibody titres against the rabies virus were high even after only three injections.37 Indeed, a series of three injections is recommended for pre-exposure immunization. For greater certainty, four or five injections are recommended for those bitten by a rabid animal, together with rabies antiserum.38,39 The standard pre-exposure regimen, given intramuscularly or intradermally, is three doses at 0, 7 and 21–28 days. There are many post-exposure regimens used throughout the world, but a minimum of four doses should be administered intramuscularly, for example at 0, 3, 7 and 14 days after an exposure. Alternatively, intradermal regimens often consist of 4–8 injections at the first visit, followed by boosters at days 7, 28 and 90. If re-exposure occurs, two doses at 0 and 3 days are recommended.40
Human diploid cells are relatively difficult to grow in large numbers and the cost of a rabies vaccine made in those cells is high. Therefore, the virus was adapted to other cell types, notably African green monkey kidney Vero cells and chick embryo cells.41,42 Vero cells are a stable cell line, easy to grow in fermenters, and most of the rabies vaccines used today are probably made in those cells. Cell culture rabies vaccines are now made by several manufacturers and nerve tissue vaccine is gradually disappearing. An additional value of the cell culture products is that they can be delivered via the intradermal as well as the intramuscular route. Intradermal vaccination allows for a five- to tenfold reduction in the quantity of vaccine required and, thus, a considerable reduction in costs of rabies prophylaxis.43–45
The remaining problem in the prevention of rabies is distribution of the vaccine so that it is available everywhere and extension of pre-exposure vaccination in countries where rabies exposure is common. Years ago we showed that children could be easily vaccinated against rabies if the vaccine is included in the routine immunizations during infancy.46 Although antibodies will decline after pre-exposure immunization, they will persist in many individuals, and in others a single booster at the time of rabies exposure will almost certainly result in an anamnestic response, although two doses are recommended in that situation.47
Attention must also be paid to the vectors of rabies for humans: dogs, cats, other animals that come in contact with humans and bats.48 Little can be done with respect to bats, but dogs and cats should be routinely vaccinated against rabies.49 In areas of the world where domestic animal rabies is controlled, the problem of rabies in free-living species such as foxes, raccoons and skunks is important.50 This problem has been solved in principle by the development of vectors such as poxviruses into which the gene for rabies glycoprotein has been inserted.51 When animals eat baits containing these vectors they become infected with them. The infection generates antibodies against both the vector and the rabies protein generated by the inserted gene. As a result, the wild animal rabies reservoir is diminished and fewer humans come in contact with the rabies virus. This strategy has been highly successful in both the USA and Europe.52,53
Gastroenteritis is a major killer of infants. Of course, bacterial agents such as Shigella, Salmonella and Escherichia coli cause many infections, but studies carried out in the 1970s showed that the major viral cause was a group of viruses called rotaviruses, so named because of their wheel-like appearance on electron microscopy.54 An estimated 450 000 infants die each year from rotavirus infection.55 The rotaviruses contain double-stranded RNA, with 11 separate segments each coding for a protein.56 It also became apparent that rotaviruses infected and caused diarrhoea in many species of mammals, but that each species was infected by a different group, so that there were mouse rotaviruses, bovine rotaviruses, monkey rotaviruses, etc.57
Efforts to identify protective mechanisms have not been entirely successful, although serum IgA responses to vaccination correlate well with efficacy and probably reflect IgA responses in the intestine.58 However, IgG antibodies in the serum can also be protective.59
A group of workers at the US National Institutes of Health led by Albert Kapikian used the fact that the animal rotaviruses are species specific and that the RNA segments permitted reassortment of viral genes to create a vaccine.60 The vaccine was based on a rhesus rotavirus, which was reassorted with each of the four most common serotypes of human rotaviruses by coinfection in cell culture with selection of those viruses that contained 10 RNA segments from the murine virus and one from the human virus. The latter coded for a human protein called VP7, which induced neutralizing antibodies.61 The final vaccine was a quadrivalent vaccine that, when given orally, could induce antibodies to the VP7 of human G serotypes 1–4. Unfortunately, after licensure it became apparent that the rhesus vaccine caused intussusceptions (intestinal invagination) in about 1 in 10 000 infants and it was taken off the market.62 This was a debacle for vaccination.
Nevertheless, two groups continued efforts seeking to develop a better and safer rotavirus vaccine. One of them, in my laboratory, had previously worked with a bovine rotavirus strain that was clearly attenuated in humans, but which was insufficiently protective.63
Accordingly, we created new reassortants, this time a pentavalent mixture, as shown in Figure 1. Four reassortants were similar to the previous vaccine in that they contained 10 RNA segments from the animal virus and one segment from a human virus coding for the VP7 of G1, G2, G3 or G4. In addition, a fifth reassortant was created, containing an RNA segment coding for a protein called VP4, also capable of inducing neutralizing antibodies.63–68 The particular VP4, called P1A,8 is the most common in human rotaviruses. The idea was to induce a neutralizing response against all of the common human strains. Although the result was successful, it turned out later that heterotypic immunity is also induced and thus there is also protection against serotypes not in the vaccine.66
The other group, located at the Cincinnati Children’s Hospital, successfully attenuated a single strain of human rotavirus by serial passage in cell culture.69 Both new vaccines were eventually licensed in the USA and Europe. The pentavalent vaccine was introduced into routine vaccination in the USA in 2006, with three oral doses recommended at 2, 4 and 6 months of age. The effect on rotavirus disease was immediate, with a sharp reduction in laboratory diagnoses of infection and in admissions to hospital caused by rotavirus gastroenteritis.59,70 There was also evidence of a herd effect: fewer cases in unvaccinated infants.71 The monovalent vaccine also showed high efficacy when tested in Europe and Latin America.72–74
However, when the two vaccines were tested in developing countries, the efficacy was much lower. Studies conducted in Asia and Africa gave 50–80% prevention of severe rotavirus diarrhoea.75–77 Investigations are being carried out to determine the reason for the differences in efficacy between developed and developing countries. Possible explanations for the lower efficacy in developing countries include neutralization by antibodies in breast milk, neutralization by higher maternal antibodies, interference by other viruses and changes in the small intestine caused by those prior infections.59 Nevertheless, from a public health point of view, because of the high incidence of serious and fatal gastroenteritis in tropical countries, even a reduction of 50% results in a major beneficial effect of rotavirus vaccination and various strategies are being explored to improve efficacy.
Although the two new rotavirus vaccines are clearly saving countless lives, the issue of possible causation of intussusceptions has not gone away. The data are mixed, in that some studies have failed to show a relationship between vaccination and intussusceptions whereas others have done so.59,78 Because intussusceptions occurs with a background rate of approximately 1 in 3000 infants, it has been difficult to discern if vaccination increases the rate.79 However, it appears that an attributable rate could be between 1 in 50 000 and 1 in 100 000 of those who have been vaccinated, but because gastroenteritis kills so many infants, both the World Health Organization and the Centers for Disease Control and Prevention have decided that the risk–benefit calculation is in favour of vaccination.80
In the 1970s, after it became apparent that the rubella vaccination would eventually eradicate congenital rubella, the importance of congenital CMV (cCMV) infection became evident. Indeed, it has since been confirmed that cCMV occurs in between 0.5% and 2% of all pregnancies81,82 and, thus, that it is a major cause of birth defects, including severe central nervous system malformations and, primarily, deafness.83,84 It is truly the ‘changeling demon’ as a child apparently normal at birth may later develop problems that impinge on normal development.85 In addition, the era of solid organ and haematogenous stem cell transplantation revealed the importance of CMV infection and disease, which has been called the ‘troll of transplantation’.86 Aside from fever, serious pneumonia and gastrointestinal disease, CMV increases the risk of solid organ and stem cell transplant rejection.87,88
Natural infection gives considerable protection against fetal disease;56 thus, the first efforts to develop a CMV vaccine were based on live, attenuated strains. A British team briefly tested the AD-169 laboratory strain and we isolated and attenuated a strain called Towne from a patient with cCMV.89,90 The Towne strain was passaged in human diploid cells until it developed a deletion that restricted replication and latency in vivo. At 125 passages, it was tested both in normal adults and in recipients of kidney transplants.91,92 Vaccination of previously seronegative recipients with a single dose reduced serious CMV disease and transplant infection by approximately 80% but did not significantly prevent infection.93,94 On the other hand, despite induction of antibodies and cellular immune responses, the vaccine virus did not become latent in the vaccinees.95
The next step was to test the Towne vaccine in seronegative mothers exposed to children in day-care centres, where CMV excretion is common and transmission is rife. Unfortunately, vaccination did not prevent women from becoming infected from their infants;96 therefore, it appeared that Towne is overattenuated. In order to increase the immunogenicity, recombinants were made with a low-passage isolate called Toledo.97 These recombinants are currently being tested in seronegative subjects to see if the immune response can be improved.
Meanwhile, other avenues were pursued to develop a vaccine. The neutralizing antibody response to CMV is induced by two antigens on the surface of the virus: a glycoprotein called gB98 and a pentameric complex called gH/gL/UL128/UL130/UL131.99–101 Notably, the pentameric complex was missing on the Towne virus.102 Many attempts are being made to induce antibodies to one or more of the proteins in the pentameric complex because antibodies to the complex prevent entry of CMV into epithelial and endothelial cells. However, the gB glycoprotein can be easily produced in Chinese hamster ovary cells and, therefore, a number of trials have been carried out with that protein combined with an adjuvant.103,104 Antibodies are induced at reasonably high levels after three doses of gB.105
The gB protein combined with an oil in water adjuvant, MF-59, was tested in a placebo-controlled trial conducted in young women who had a high rate of acquisition of CMV.106 The results were positive in that over the first 18 months of the study, 60% protection was shown, but efficacy waned afterwards to 50% at 42 months. This was associated with a peak antibody titre from 7 to 18 months that faded quickly thereafter. More striking was a study of the same vaccine in recipients of kidney and liver transplants conducted in the UK.107 Both seronegative and seropositive recipients of organs showed 90% protection against CMV viraemia compared with placebo recipients, indicating that the antibodies could prevent both primary infection and reinfection.
Another group has combined gB with AS01, an adjuvant containing a toll-like receptor agonist, and showed that antibodies persist much longer than after vaccination with gB/MF59 (Professor Arnaud Marchant, University of Louvain, 2013, personal communication). However, no efficacy data are yet available for this approach.
DNA plasmids have long been known to induce immune responses against the proteins coded by the DNA. Protection of haematogenous stem cell recipients against CMV viraemia and the need for antiviral treatment was demonstrated in a placebo-controlled clinical trial.108,109 The proteins generated were gB and pp65, the latter being a tegument protein of the virus that is the principal inducer of cytotoxic T cells against CMV. Although responses to gB were not high, the responses to pp65 were high and it makes sense that cytotoxic T lymphocytes would restrict reactivation of CMV from latency in the haematogenous stem cell recipients. Thus, it appears that we already have the tools to prevent CMV disease in both solid organ and haematogenous stem cell recipients.
The situation for prevention of congenital CMV infection is less clear. The results with gB in young women suggest that protection is feasible if antibodies to gB can be maintained at high levels, but uncertainty exists as to the need for antibodies to the pentameric complex. In as much as most of the neutralizing antibody after natural infection is directed against the complex,101 many believe this should be induced by a vaccine and many groups are working towards this end. Aside from simply adding one or more of the proteins in the complex to a vaccine, strategies to induce antibodies to those proteins include replication-defective CMV, virus-like particles and DNA plasmids.
How would a CMV vaccine be used to prevent cCMV? There are several possible targets, depending on the properties of a vaccine. The most obvious is the immunization of seronegative girls before they enter the child-bearing years. As seropositive girls also have a risk of reinfection, albeit with lesser consequences to the fetus, ideally a vaccine would boost their immunity.81,110 Of course, vaccines could be given to older women contemplating pregnancy. From a public health viewpoint, an interesting approach would be vaccination of infants before they come in contact with other children, as they are the chief source of infection to their mothers. Such vaccination could reduce circulation of CMV (Table 2).
In any case, the prospects that a vaccine against CMV will become available are good, in as much as efficacy has already been shown for protection of seronegative women and transplant recipients. Fortunately, multiple companies and biotechnology companies have development programmes and we can hope that one day pregnancy will be even safer.111
I have reviewed four viral vaccines whose development I have been deeply involved in. The common theme is that one must identify the type of immune response that is necessary for prevention at an early stage and then seek to induce that response using the safest and most effective method. Fortunately, modern vaccinology contains numerous approaches to manipulate viral proteins and, therefore, the principal obstacle is to convince people that a vaccine is required.