Protective immunity following vaccination: How is it defined?

Introduction

The goal of vaccination is to induce protective immunity against disease. For infectious disease, vaccination works not only through direct protection of the vaccinee, but also through herd immunity by limiting disease spread to susceptible individuals. Interestingly, despite the clear medical significance of vaccination there are no specific guidelines for how protective immunity is defined. This is a challenging task since a correlate of immunity must be defined individually for each vaccine and will vary depending on the characteristics of the disease that is being studied. Here, we describe the immunological correlates (or the lack thereof) for five vaccines and ask the question of how correlates of protective immunity for new vaccines may be determined in the future.

Smallpox

The smallpox vaccine was first described by Edward Jenner in 1798¹, leading to the ultimate eradication of natural smallpox in 1977. At the time of its development, the fields of virology and immunology did not exist and so a visual marker of successful vaccination was used as a surrogate to an immunological correlate of immunity. Unlike injected vaccines that are administered subcutaneously or intramuscularly, the original smallpox vaccine was delivered by scarification, a term indicating that the vaccine inoculum is introduced by scratching the skin surface. As the virus replicates, the vaccination site typically evolves through macular, papular, and vesicular stages of development and the vesicular or pustular stage is often referred to as a “Jennerian vesicle” or a vaccine “take”. Evidence of the formation of a vesicular lesion has been used for over 200 years as evidence of protective immunity against smallpox. Even today, with our advances in cutting-edge immunological techniques, the Jennerian vesicle remains the only universally accepted correlate of successful vaccination.² New smallpox vaccines, such as modified vaccinia virus Ankura (MVA), are injected instead of being scratched on the skin surface and therefore do not induce vesicle formation, the only correlate of protection that has been accepted by the medical community. This has caused somewhat of a dilemma because for the first time since its introduction by Edward Jenner, there is a need to identify a new correlate of protective immunity following smallpox vaccination. One suggestion for defining protective immunity was to show protection against vesicle formation following revaccination with the traditional smallpox vaccine.³ This was based on the assumption that immune/protected individuals would not form a Jennerian vesicle upon cutaneous challenge. However, this approach has been disputed since approximately 85% of vaccinees will still demonstrate vesicle formation within just a year after successful vaccination.⁴ Since it is established that protection against smallpox is maintained for many years after vaccination⁵, this indicates that protection against direct cutaneous challenge is not an accurate measure of systemic protective immunity.

Perhaps a better approach would be to use immunological techniques to determine correlates of protection instead of visual examination of vesicle formation at the vaccination site. In this regard, small independent epidemiologic studies examining pre-existing neutralizing antibody titers against vaccinia have shown that neutralizing titers of 1:20 or 1:32 are indicative of protective immunity.^6,7 Neutralizing antibody is highly protective against orthopoxviruses in animal models⁵ and has been shown to be both necessary and sufficient for protection of non-human primates against monkeypox, the most virulent orthopoxvirus related to smallpox.⁸ Despite these results and both direct and indirect evidence for the role of neutralizing antibodies in mediating protective immunity in humans following smallpox infection⁵, there has been no consensus in terms of the requirements for protective immunity against smallpox. There is no animal reservoir for smallpox and although experimental animal models of smallpox infection are useful for studying specific aspects of pathogenesis, they do not demonstrate the full spectrum of human disease manifestations with physiological doses of virus. For this reason, and the knowledge that smallpox is now extinct in nature, it is unlikely that a defined immunological correlate of protective immunity will be accepted at any time in the near future.

Tetanus

Tetanus is a rare, but life-threatening disease caused by the tetanus toxin produced following infection with the bacterium, Clostridium tetani. Since the disease is toxin-mediated, immunity is achieved through antibody-mediated neutralization. Accordingly, the World Health Organization (WHO) and the Centers for Disease Control and Prevention (CDC) suggest that a serum antibody level of 0.01 IU/ml will provide a minimum level of protective immunity against the disease. The putative protective level of 0.01 IU/ml was first suggested in 1937⁹ and was based primarily on challenge studies performed in guinea pigs⁹ and horses in the 1930s.^10–12 Although these represent excellent animal models for tetanus, guinea pigs and horses are only distantly related to humans and this may have delayed acceptance of this immunological correlate had it not been for direct evidence in humans. There is one remarkable study published in 1942 in which two German researchers, with pre-existing antibody titers of only 0.01 IU/ml and 0.007 IU/mL, purposefully injected themselves with 2 to 3 lethal doses of tetanus toxin in order to demonstrate that their pre-existing humoral immunity was indeed protective.¹³ Neither subject died, indicating that 0.01 IU/ml was sufficient for protection against experimental inoculation with normally lethal doses of tetanus toxin and 0.01 IU/ml was later established by the WHO as the immunological correlate of immunity. Studies published 30 years after the direct human challenge studies¹³ provide comparable results¹⁴; 54/64 (84%) of tetanus patients had antitoxin levels below 0.01 IU/ml and although 10 patients had titers of >0.01 IU/mL at the time of clinical analysis, all of the deaths occurred in patients with antitoxin levels of 0.002 IU/mL or less. These observations provide further support for the continued use of 0.01 IU/ml as a correlate of protection against lethal tetanus.

Yellow Fever

Yellow fever (YF) is a mosquito-borne viral disease responsible for an estimated 200,000 infections annually in the tropical regions of South America and Africa. A live attenuated vaccine strain of the virus (strain 17D and its derivatives) was developed in the 1930s, with virus-specific antibody responses considered key in achieving immunity against wild type infection.¹⁵ The package insert of the U.S. vaccine, YF-VAX®, states that a log₁₀ neutralizing index (LNI) of 0.7 is indicative of protective immunity and recent clinical trials use an LNI of >0.7 as a cutoff for seroconversion and a correlate of protective immunity.^16,17 What many researchers may not realize is that the 0.7 LNI correlate of protective immunity was not initially based on human subjects but is instead based on a study performed in non-human primates. In 1974, Mason et al. performed a series of experiments in rhesus macaques immunized with graded doses of the yellow fever vaccine and then challenged 20 weeks later with a lethal dose of the virulent Asibi strain of yellow fever.¹⁸ Of the monkeys tested, 94% (51/54) of survivors demonstrated an LNI ≥0.7, whereas 91% (10/11) of fatal infections occurred in animals that exhibited an LNI of <0.7. Primates represent the natural animal reservoir for yellow fever and often demonstrate the full spectrum of human disease and mortality under experimental conditions - and this may explain why there is wide acceptance of 0.7 LNI as a correlate of immunity for yellow fever.

Measles

Measles is a highly infectious viral disease with clinical presentation ranging from relatively mild illness to death.¹⁹ Although representing a vaccine-preventable disease, measles accounts for up to 10% of worldwide mortality in children less than 5 years of age.²⁰ Live attenuated measles vaccines were developed in the 1950s, and licensed for use in the U.S. in 1963. Following large scale introduction of the vaccine, there has been striking success in reducing the occurrence of disease.¹⁹ Humoral immunity has long been known to play a vital role in protection against measles, as demonstrated by immunotherapy with convalescent serum²¹ and maternal antibodies providing protective immunity in infants exposed to measles infection.²² Currently, anti-measles antibody levels of greater than 200 mIU/ml are considered protective.²³ Similar to smallpox, there is no animal reservoir for measles and studies in experimental animal models were not used to define the correlates for protective immunity. Instead, an “experiment of nature” that occurred in 1985 revealed an immunological correlate of protection that is still used today. The level of 200 mIU/ml as a correlate of immunity against measles is derived from a study comparing pre-existing levels of neutralizing antibody to disease outcome following a measles outbreak in a dormitory setting shortly following a campus blood drive.²⁴ Samples from the blood drive provided the critical data on pre-existing antibody titers that the study authors were then able to correlate with protection following the outbreak. They found that clinical measles occurred in 88% (8/9) of subjects with a neutralizing titer of ≤120, whereas fully symptomatic measles was not observed in any (0/71) subjects whose neutralizing titer exceeded 220 (≈200 mIU/ml). Upon further examination, 64% of subjects with pre-existing anti-measles antibody titers ranging from 216–874 mIU/ml, showed large spikes in their levels of neutralizing antibodies (an average of 42-fold), indicating that they had been infected with measles, but spared from most overt clinical symptoms. In contrast, none of the tested subjects with a pre-existing titer of ≥1052 demonstrated a spike in antiviral antibodies or showed signs of illness. This suggests that individuals with ≥200 mIU/ml of anti-measles antibodies will not necessarily be protected against infection (this may require very high antibody titers of ≥1000 mIU/ml), but they will likely be protected against overt disease and are less likely to spread the virus compared to patients with clinically apparent cases of measles.

Chickenpox

Chickenpox is caused by infection with varicella zoster virus (VZV). This disease is usually self-limiting and benign but can lead to severe complications in young infants, pregnant women and immune-compromised individuals. A live attenuated vaccine against chickenpox (Varivax®) was approved by the FDA in 1995 and the efficacy of this vaccine was determined by comparing the rate of varicella incidence in vaccinated versus unvaccinated children with confirmed exposures. Initially, vaccine efficacy was calculated at 86%²⁵, but later studies showed that immunity wanes over time.²⁶ introduction of routine vaccination against chickenpox has led to a dramatic decrease in VZV-related hospitalizations and deaths but the immunological correlate of immunity is open to debate. Similar to smallpox and measles, human VZV does not naturally infect other animals. Simian varicella virus (SVV) represents a closely related herpesvirus and provides an excellent model in non-human primates for studying varicella pathogenesis.²⁷ However, this model has been underutilized – especially in terms of determining the underlying correlates of vaccine-mediated protection and this may be partly due to the inability of VZV to cause disease in primate species other than humans. Nevertheless, vaccination with VZV can protect primates against SVV infection, indicating that this model might be useful in determining immunological correlates of protection against these pathogens. Epidemiological studies have shown that the risk for breakthrough varicella in vaccinated individuals is inversely related to antibody titers²⁸ with a titer of >5 gpELISA units/ml at 6 weeks post vaccination associated with 95.5% vaccine efficacy compared to 83.5% when the titer is <5 gpELISA units/ml.²⁹ It is unclear if VZV-specific antibody is playing a direct role in vaccine-mediated protection or if it is simply a surrogate marker of vaccine-elicited T cell responses that accompany seroconversion. For instance, although the percentage of vaccinees with >5 gpELISA units/ml is stable from 2 to 9 years post-primary vaccination (91–95%, respectively)³⁰, the rates of breakthrough varicella changes significantly from 1.6 cases/1000 person-years at 1 year after vaccination to 9.0 cases/1000 person-years at 5 years, and 58.2 cases/1000 person-years at 9 years post-vaccination.²⁶ If 5 gpELISA units/ml is an accurate and specific measure of protective immunity, then one would have expected a significant decrease in gpELISA scores during the timeframe that varicella breakthrough was steadily increasing. Instead, the discrepancy between long-term maintenance of antibody ELISA titers and varicella breakthrough indicates that a new or perhaps more complex immunological correlate is warranted. This may require improved standardized techniques for measuring cellular immunity and a better understanding of the combined roles of humoral and cellular immune responses in vaccine-induced protection against VZV.

Conclusions

Correlates of protective immunity differ greatly, and can range from the visual observation of a pustule at the vaccination site (e.g., smallpox vaccination), to the development of a defined serological titer of antigen-specific antibody (e.g., tetanus, measles, and yellow fever). These correlates may derive strictly from clinical observation following exposure, analysis in appropriate animal studies (when available), or some combination of both approaches. Future challenges in defining vaccine-mediated correlates of protective immunity will likely come from complex viruses such as VZV as well as other pathogens that may require a combination of both cellular and humoral immunity in order to provide effective host defense. Under these circumstances, new correlates of immunity will need to be defined and validated. Until such correlates are demonstrated, the long-term efficacy for new vaccines, and the potential susceptibility of vaccinated populations, may in some cases remain uncertain.