A Guide to Vaccinology: From Basic Principles to New Developments PDF

REvIEWS A guide to vaccinology: from basic principles to new developments Andrew J. Pollard 1,2 ✉ and Else M. Bijker1,2 Abstract | Immunization...

REvIEWS A guide to vaccinology: from basic principles to new developments Andrew J. Pollard 1,2 ✉ and Else M. Bijker1,2 Abstract | Immunization is a cornerstone of public health policy and is demonstrably highly cost-effective when used to protect child health. Although it could be argued that immunology has not thus far contributed much to vaccine development, in that most of the vaccines we use today were developed and tested empirically, it is clear that there are major challenges ahead to develop new vaccines for difficult-to-target pathogens, for which we urgently need a better understanding of protective immunity. Moreover, recognition of the huge potential and challenges for vaccines to control disease outbreaks and protect the older population, together with the availability of an array of new technologies, make it the perfect time for immunologists to be involved in designing the next generation of powerful immunogens. This Review provides an introductory overview of vaccines, immunization and related issues and thereby aims to inform a broad scientific audience about the underlying immunological concepts. Antigens Vaccines have transformed public health, particularly on infectious diseases to provide insight into the key Parts of the pathogen (such as since national programmes for immunization first issues facing immunologists today. We also provide proteins or polysaccharides) became properly established and coordinated in the some perspectives on current and future challenges that are recognized by the 1960s. In countries with high vaccine programme cov- in continuing to protect the world’s population from immune system and can be erage, many of the diseases that were previously respon- common pathogens and emerging infectious threats. used to induce an immune response by vaccination. sible for the majority of childhood deaths have essentially Communicating effectively about the science of vacci- disappeared1 (Fig. 1). The World Health Organization nation to a sceptical public is a challenge for all those Protection (WHO) estimates that 2–3 million lives are saved each engaged in vaccine immunobiology but is urgently The state in which an individual year by current immunization programmes, contributing needed to realign the dialogue and ensure public health8. does not develop disease after being exposed to a pathogen. to the marked reduction in mortality of children less than This can only be achieved by being transparent about 5 years of age globally from 93 deaths per 1,000 live births what we know and do not know, and by considering the in 1990 to 39 deaths per 1,000 live births in 2018 (ref.2). strategies to overcome our existing knowledge gaps. Vaccines exploit the extraordinary ability of the highly evolved human immune system to respond to, What is in a vaccine? and remember, encounters with pathogen antigens. A vaccine is a biological product that can be used to However, for much of history, vaccines have been devel- safely induce an immune response that confers protection oped through empirical research without the involve- against infection and/or disease on subsequent exposure ment of immunologists. There is a great need today for to a pathogen. To achieve this, the vaccine must contain improved understanding of the immunological basis antigens that are either derived from the pathogen or for vaccination to develop vaccines for hard-to-target produced synthetically to represent components of the pathogens (such as Mycobacterium tuberculosis, the bac- pathogen. The essential component of most vaccines 1 Oxford Vaccine Group, terium that causes tuberculosis (TB))3 and antigenically is one or more protein antigens that induce immune Department of Paediatrics, variable pathogens (such as HIV)4, to control outbreaks responses that provide protection. However, polysac- University of Oxford, that threaten global health security (such as COVID-19 charide antigens can also induce protective immune Oxford, UK. or Ebola)5,6 and to work out how to revive immune responses and are the basis of vaccines that have been 2 NIHR Oxford Biomedical responses in the ageing immune system7 to protect developed to prevent several bacterial infections, such Research Centre, Oxford the growing population of older adults from infectious as pneumonia and meningitis caused by Streptococcus University Hospitals Trust, Oxford, UK. diseases. pneumoniae, since the late 1980s9. Protection conferred ✉e-mail: andrew.pollard@ In this Review, which is primarily aimed at a broad by a vaccine is measured in clinical trials that relate paediatrics.ox.ac.uk scientific audience, we provide a guide to the history immune responses to the vaccine antigen to clinical end https://doi.org/10.1038/ (Box 1), development, immunological basis and remark- points (such as prevention of infection, a reduction in s41577-020-00479-7 able impact of vaccines and immunization programmes disease severity or a decreased rate of hospitalization). NaTure RevIeWS | ImmUnology volume 21 | February 2021 | 83 Reviews a Diphtheria b Capsular group C meningococcus Introduction of 1,800 80,000 vaccination (1940) 1,600 Introduction of Number of cases 70,000 1,400 vaccination (1999) Notifications 60,000 1,200 50,000 1,000 40,000 800 30,000 600 20,000 400 10,000 200 0 0 1914 1924 1934 1944 1954 1964 1974 1984 1994 2003 20 9/2 9 20 0/2 0 20 1/2 1 20 2/2 2 20 3/2 3 20 4/2 4 20 5/2 5 20 6/2 6 20 7/2 7 20 8/2 8 20 9/2 9 20 0/2 0 20 1/2 1 /2 2 3 9 99 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 0 00 1 01 1 01 12 01 01 Year 19 8/1 9 19 Year c Polio d Haemophilus influenzae type B 1,000 Introduction of 900 vaccination (1992) 800 Laboratory reports 7,000 700 Introduction of 600 6,000 vaccination (1956) Notifications 5,000 500 4,000 400 3,000 300 2,000 200 1,000 100 0 0 1912 1922 1932 1942 1952 1962 1972 1982 1992 2006 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 05 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 19 Year Year e Measles f Pertussis 900,000 800,000 Introduction of 200,000 700,000 vaccination (1968) 180,000 Notifications 160,000 Notifications 600,000 Introduction of 140,000 500,000 vaccination (1950s) 120,000 400,000 100,000 300,000 80,000 60,000 200,000 40,000 100,000 20,000 0 0 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013 2016 Year Year Fig. 1 | The impact of vaccination on selected diseases in the UK. The introduction of vaccination against infectious diseases such as diphtheria (part a), capsular group C meningococcus (part b), polio (part c), Haemophilus influenzae type B (part d), measles (part e) and pertussis (part f) led to a marked decrease in their incidence. Of note, the increase in reports of H. influenzae type B in 2001 led to a catch-up vaccination campaign, after which the incidence reduced. For pertussis, a decline in vaccine coverage led to an increase in cases in the late 1970s and 1980s, but disease incidence reduced again after vaccine coverage increased. Adapted with permission from the Green Book, information for public health professionals on immunisation, Public Health England, contains public sector information licensed under the Open Government Licence v3.0. Finding an immune response that correlates with pro- The distinction between live and non-live vaccines tection can accelerate the development of and access to is important. The former may have the potential to repli- new vaccines10 (Box 2). cate in an uncontrolled manner in immuno­compromised Vaccines are generally classified as live or non-live individuals (for example, children with some pri- Attenuated (sometimes loosely referred to as ‘inactivated’) to distin- mary immunodeficiencies, or individuals with HIV A reduction in the virulence guish those vaccines that contain attenuated replicating infection or those receiving immunosuppressive drugs), of a pathogen (through either deliberate or natural changes strains of the relevant pathogenic organism from those leading to some restrictions to their use11. By contrast, in virulence genes). that contain only components of a pathogen or killed non-live vaccines pose no risk to immunocompromised whole organisms (Fig. 2). In addition to the ‘traditional’ individuals (although they may not confer protection in Virus-like particles live and non-live vaccines, several other platforms have those with B cell or combined immunodeficiency, as Particles constructed of viral proteins that structurally mimic been developed over the past few decades, including viral explained in more detail later). the native virus but lack the vectors, nucleic acid-based RNA and DNA vaccines, and Live vaccines are developed so that, in an immuno- viral genome. virus-like particles (discussed in more detail later). competent host, they replicate sufficiently to produce a 84 | February 2021 | volume 21 www.nature.com/nri  Reviews Adjuvant strong immune response, but not so much as to cause the portfolio of adjuvants is steadily expanding, with An agent used in a vaccine to significant disease manifestations (for example, the liposome-based adjuvants and oil-in-water emulsions enhance the immune response vaccines for measles, mumps, rubella and rotavirus, being licensed in the past few decades14. The mech- against the antigen. oral polio vaccine, the Mycobacterium bovis bacillus anism of action of aluminium salts (alum), although Danger signals Calmette–Guérin (BCG) vaccine for TB and live atten- extensively used as an adjuvant for more than 80 years, Molecules that stimulate a uated influenza vaccine). There is a trade-off between remains incompletely understood15, but there is increas- more robust immune response enough replication of the vaccine pathogen to induce a ing evidence that immune responses and protection can together with an antigen. strong immune response and sufficient attenuation of be enhanced by the addition of newer adjuvants that pro- Endogenous mediators that are the pathogen to avoid symptomatic disease. For this rea- vide danger signals to the innate immune system. Examples released in response to infection or injury and that interact with son, some safe, live attenuated vaccines require multiple of these novel adjuvants are the oil-in-water emulsion pattern recognition receptors doses and induce relatively short-lived immunity (for MF59, which is used in some influenza vaccines16; AS01, such as Toll-like receptors to example, the live attenuated typhoid vaccine, Ty21a)12, which is used in one of the shingles vaccines and the activate innate immune cells and other live attenuated vaccines may induce some mild licensed malaria vaccine17; and AS04, which is used in a such as dendritic cells. disease (for example, about 5% of children will develop vaccine against human papillomavirus (HPV)18. Innate immune system a rash and up to 15% fever after measles vaccination)13. Vaccines contain other components that function as The evolutionarily primitive The antigenic component of non-live vaccines can be preservatives, emulsifiers (such as polysorbate 80) or part of the immune system killed whole organisms (for example, whole-cell pertus- stabilizers (for example, gelatine or sorbitol). Various that detects foreign antigens sis vaccine and inactivated polio vaccine), purified pro- products used in the manufacture of vaccines could the- in a non-specific manner. teins from the organism (for example, acellular pertussis oretically also be carried over to the final product and AS01 vaccine), recombinant proteins (for example, hepatitis are included as potential trace components of a vaccine, A liposome-based adjuvant B virus (HBV) vaccine) or polysaccharides (for exam- including antibiotics, egg or yeast proteins, latex, for- containing 3-O-desacyl- ple, the pneumococcal vaccine against S. pneumoniae) maldehyde and/or gluteraldehyde and acidity regulators 4′-monophosphoryl lipid A (Fig. 2). Toxoid vaccines (for example, for tetanus and (such as potassium or sodium salts). Except in the case of and the saponin QS-21. AS01 triggers the innate immune diphtheria) are formaldehyde-inactivated protein toxins allergy to any of these components, there is no evidence system immediately after that have been purified from the pathogen. of risk to human health from these trace components of vaccination, resulting in an Non-live vaccines are often combined with an adjuvant some vaccines19,20. enhanced adaptive immune to improve their ability to induce an immune response response. (immunogenicity). There are only a few adjuvants Vaccines induce antibodies that are used routinely in licensed vaccines. However, The adaptive immune response is mediated by B cells that produce antibodies (humoral immunity) and Box 1 | A brief history of vaccination by T cells (cellular immunity). All vaccines in rou- tine use, except BCG (which is believed to induce Epidemics of smallpox swept across Europe in the seventeenth and eighteenth centuries, accounting for as much as 29% of the death rate of children in London137. Initial efforts to T cell responses that prevent severe disease and innate control the disease led to the practice of variolation, which was introduced to England by immune responses that may inhibit infection; see later), Lady Mary Wortley Montagu in 1722, having been used in the Far East since the mid-1500s are thought to mainly confer protection through the (see Nature Milestones in Vaccines). In variolation, material from the scabs of smallpox induction of antibodies (Fig. 3). There is considerable lesions was scratched into the skin in an attempt to provide protection against the supportive evidence that various types of functional anti- disease. Variolation did seem to induce protection, reducing the attack rate during body are important in vaccine-induced protection, and epidemics, but sadly some of those who were variolated developed the disease and this evidence comes from three main sources: immu- sometimes even died. It was in this context that Edward Jenner wrote ‘An Inquiry into the nodeficiency states, studies of passive protection and Causes and Effects of the Variole Vaccinae…’ in 1798. His demonstration, undertaken by immunological data. scratching material from cowpox lesions taken from the hands of a milkmaid, Sarah Nelms, into the skin of an 8-year-old boy, James Phipps, who he subsequently challenged with smallpox, provided early evidence that vaccination could work. Jenner’s contribution Immunodeficiency states. Individuals with some to medicine was thus not the technique of inoculation but his startling observation that known immunological defects in antibodies or associ- milkmaids who had had mild cowpox infections did not contract smallpox, and the ated immune components are particularly susceptible serendipitous assumption that material from cowpox lesions might immunize against to infection with certain pathogens, which can provide smallpox. Furthermore, Jenner brilliantly predicted that vaccination could lead to the insight into the characteristics of the antibodies that eradication of smallpox; in 1980, the World Health Assembly declared the world free of are required for protection from that particular path- naturally occurring smallpox. ogen. For example, individuals with deficiencies in the Almost 100 years after Jenner, the work of Louis Pasteur on rabies vaccine in the 1880s complement system are particularly susceptible to menin- heralded the beginning of a frenetic period of development of new vaccines, so that gococcal disease caused by infection with Neisseria men- by the middle of the twentieth century, vaccines for many different diseases (such ingitidis21 because control of this infection depends on as diphtheria, pertussis and typhoid) had been developed as inactivated pathogen products or toxoid vaccines. However, it was the coordination of immunization as complement-mediated killing of bacteria, whereby com- a major public health tool from the 1950s onwards that led to the introduction of plement is directed to the bacterial surface by IgG anti- comprehensive vaccine programmes and their remarkable impact on child health bodies. Pneumococcal disease is particularly common that we enjoy today. In 1974, the World Health Organization launched the Expanded in individuals with reduced splenic function22 (which Programme on Immunization and a goal was set in 1977 to reach every child in the may be congenital, resulting from trauma or associated world with vaccines for diphtheria, pertussis, tetanus, poliomyelitis, measles and with conditions such as sickle cell disease); S. pneumo- tuberculosis by 1990. Unfortunately, that goal has still not been reached; although niae bacteria that have been opsonized with antibody and global coverage of 3 doses of the diphtheria–tetanus–pertussis vaccine has risen to complement are normally removed from the blood by more than 85%, there are still more than 19 million children who did not receive basic phagocytes in the spleen, which are no longer present vaccinations in 2019 (ref.105). in individuals with hyposplenism. Antibody-deficient NaTure RevIeWS | ImmUnology volume 21 | February 2021 | 85 Reviews Box 2 | Correlates of protection surface polysaccharides of invasive bacteria such as meningococci (N. meningitidis)30 and pneumococci The identification of correlates of protection is helpful in vaccine development as they (S. pneumoniae)31, provide considerable protection can be used to compare products and to predict whether the use of an efficacious against these diseases. It is now known that these vac- vaccine in a new population (for example, a different age group, medical background or cines do not induce T cell responses, as polysaccha- geographical location) is likely to provide the same protection as that observed in the rides are T cell-independent antigens, and thus they must original setting. There is considerable confusion in the literature about the definition of a correlate of protection. For the purposes of this discussion, it is useful to separate out mediate their protection through antibody-dependent two distinct meanings. A mechanistic correlate of protection is the specific functional mechanisms. Protein–polysaccharide conjugate vaccines immune mechanism that is believed to confer protection. For example, antitoxin contain the same polysaccharides from the bacterial sur- antibodies, which are induced by the tetanus toxoid vaccine, confer protection directly face, but in this case they are chemically conjugated to by neutralizing the activity of the toxin. A non-mechanistic correlate of protection does a protein carrier (mostly tetanus toxoid, or diphtheria not in itself provide the protective function but has a statistical relationship with the toxoid or a mutant protein derived from it, known as mechanism of protection. An example of a non-mechanistic correlate of protection CRM197)32–34. The T cells induced by the vaccine recog- is total IgG antibody levels against pneumococci. These IgG antibodies contain the nize the protein carrier (a T cell-dependent antigen) and mechanistic correlate (thought to be a subset of opsonophagocytic antibodies) but these T cells provide help to the B cells that recognize the the mechanism of protection is not being directly measured. Correlates of protection can polysaccharide, but no T cells are induced that recognize be measured in clinical trials if there are post-vaccination sera available from individuals who do or do not develop disease, although large-scale serum collection from participants the polysaccharide and, thus, only antibody is involved is rarely undertaken in phase III clinical efficacy trials. An alternative approach is to in the excellent protection induced by these vaccines35. estimate the correlates of protection by extrapolating from sero-epidemiological Furthermore, human challenge studies offer the opportu- studies in a vaccinated population and relating the data to disease incidence in the nity to efficiently assess correlates of protection (Box 2) population. Human challenge studies have also been used to determine correlates of under controlled circumstances36, and they have been protection, although the dose of challenge bacterium or virus and the experimental used to demonstrate the role of antibodies in protection conditions may not relate closely to natural infection, which can limit the utility of these against malaria37 and typhoid38. observations. Vaccines need T cell help individuals are susceptible to varicella zoster virus Although most of the evidence points to antibod- (which causes chickenpox) and other viral infections, ies being the key mediators of sterilizing immunity but, once infected, they can control the disease in the induced by vaccination, most vaccines also induce T cell AS04 An adjuvant consisting of same way as an immunocompetent individual, so long responses. The role of T cells in protection is poorly aluminium salt and the as they have a normal T cell response23. characterized, except for their role in providing help for Toll-like receptor agonist B cell development and antibody production in lymph monophosphoryl lipid A. Passive protection. It has been clearly established nodes. From studies of individuals with inherited or Complement system that intramuscular or intravenous infusion of exoge- acquired immunodeficiency, it is clear that whereas anti- A network of proteins that nous antibodies can provide protection against some body deficiency increases susceptibility to acquisition of form an important part of the infections. The most obvious example is that of passive infection, T cell deficiency results in failure to control a immune response by enhancing transfer of maternal antibodies across the placenta, pathogen after infection. For example, T cell deficiency the opsonization of pathogens, which provides newborn infants with protection against results in uncontrolled and fatal varicella zoster virus cell lysis and inflammation. a wide variety of pathogens, at least for a few months infection, whereas individuals with antibody deficiency Opsonized after birth. Maternal vaccination with pertussis24, teta- readily develop infection but recover in the same way as A state of a pathogen in which nus25 and influenza26 vaccines harnesses this important immunocompetent individuals. The relative suppression antibodies or complement protective adaptation to reduce the risk of disease soon of T cell responses that occurs at the end of pregnancy factors are bound to its surface. after birth and clearly demonstrates the role of antibod- increases the severity of infection with influenza and Opsonophagocytic ies in protection against these diseases. Vaccination of varicella zoster viruses39. antibodies pregnant women against group B streptococci27 and res- Although evidence for the involvement of T cells Antibodies that bind to a piratory syncytial virus (RSV)28 has not yet been shown in vaccine-induced protection is limited, this is likely pathogen, which subsequently to be effective at preventing neonatal or infant infection, owing, in part, to difficulties in accessing T cells to can be eliminated by phagocytosis. but it has the potential to reduce the burden of disease in study as only the blood is easily accessible, whereas the youngest infants. Other examples include the use of many T cells are resident in tissues such as lymph nodes. T cell-independent antigens specific neutralizing antibodies purified from immune Furthermore, we do not yet fully understand which types Antigens against which B cells donors to prevent the transmission of various viruses, of T cell should be measured. Traditionally, T cells have can mount an antibody response without T cell help. including varicella zoster virus, HBV and measles virus29. been categorized as either cytotoxic (killer) T cells or Individuals with inherited antibody deficiency are with- helper T cells. Subtypes of T helper cells (TH cells) can T cell-dependent antigen out defence against serious viral and bacterial infections, be distinguished by their profiles of cytokine production. An antigen for which T cell help but regular administration of serum antibodies from an T helper 1 (TH1) cells and TH2 cells are mainly important is required in order for B cells to immunocompetent donor can provide almost entirely for establishing cellular immunity and humoral immu- mount an antibody response. normal immune protection for the antibody-deficient nity, respectively, although TH1 cells are also associated Human challenge studies individual. with generation of the IgG antibody subclasses IgG1 and Studies in which volunteers are IgG3. Other TH cell subtypes include TH17 cells (which deliberately infected with Immunological data. Increasing knowledge of are important for immunity at mucosal surfaces such as a pathogen, in a carefully conducted study, to evaluate immuno­logy provides insights into the mechanisms the gut and lung) and T follicular helper cells (located the biology of infection and the of protection mediated by vaccines. For example, in secondary lymphoid organs, which are important efficacy of drugs and vaccines. polysaccharide vaccines, which are made from the for the generation of high-affinity antibodies (Fig. 3)). 86 | February 2021 | volume 21 www.nature.com/nri  Reviews Licensed vaccines Type of vaccine using this technology First introduced Measles, mumps, rubella, Live attenuated yellow fever, influenza, oral (weakened or polio, typhoid, Japanese 1798 (smallpox) inactivated) encephalitis, rotavirus, BCG, varicella zoster Whole-cell pertussis, Killed whole polio, influenza, organism Japanese encephalitis, 1896 (typhoid) hepatitis A, rabies Toxoid Diphtheria, tetanus 1923 (diphtheria) Pertussis, influenza, Subunit (purified protein, hepatitis B, meningococcal, recombinant protein, 1970 (anthrax) pneumococcal, typhoid, polysaccharide, peptide) hepatitis A Virus-like Human papillomavirus 1986 (hepatitis B) particle Outer Pathogen 1987 membrane antigen Gram-negative bacterial outer Group B meningococcal (group B vesicle meningococcal) membrane Polysaccharide Protein–polysaccharide Haemophilus influenzae 1987 (H. influenzae conjugate type B, pneumococcal, type b) meningococcal, typhoid Carrier protein Viral vector Pathogen gene Viral Ebola 2019 (Ebola) vectored Viral vector genes RNA Nucleic acid DNA vaccine SARS-CoV-2 2020 (SARS-CoV-2) Lipid coat Pathogen gene Bacterial vectored Experimental – Bacterial vector Antigen- Pathogen presenting antigen Experimental – cell MHC Fig. 2 | Different types of vaccine. Schematic representation of different types of vaccine against pathogens; the text indicates against which pathogens certain vaccines are licensed and when each type of vaccine was first introduced. BCG, Mycobacterium bovis bacillus Calmette–Guérin. Studies show that sterilizing immunity against car- cytotoxic T cells are required to control and clear riage of S. pneumoniae in mice can be achieved by the established infection. transfer of T cells from donor mice exposed to S. pneu- moniae40, which indicates that further investigation Features of vaccine-induced protection of T cell-mediated immunity is warranted to better Vaccines have been developed over the past two centuries under­stand the nature of T cell responses that could be to provide direct protection of the immunized individ- harnessed to improve protective immunity. ual through the B cell-dependent and T cell-dependent Although somewhat simplistic, the evidence there- mechanisms described above. As our immunological fore indicates that antibodies have the major role in understanding of vaccines has developed, it has become prevention of infection (supported by TH cells), whereas apparent that this protection is largely manifested NaTure RevIeWS | ImmUnology volume 21 | February 2021 | 87 Reviews Vaccine Skin Muscle Vaccine antigen Adjuvant (containing danger signals) Dendritic Memory B cell cell proliferation PRR MHC class II Maturation of the Soluble antibody response vaccine Peptide of antigen vaccine antigen CD4+ T cell BCR Activation and trafficking to T cell help Proliferation draining lymph node Plasma cell differentiation and antibody production TCR B cell MHC T cell class II help CD8+ effector T cell Bone marrow MHC class I CD8+ CD8+ T cell memory T cell Long-lived plasma cell Fig. 3 | The generation of an immune response to a vaccine. The immune response following immunization with a conventional protein antigen. The vaccine is injected into muscle and the protein antigen is taken up by dendritic cells, which are activated through pattern recognition receptors (PRRs) by danger signals in the adjuvant, and then trafficked to the draining lymph node. Here, the presentation of peptides of the vaccine protein antigen by MHC molecules on the dendritic cell activates T cells through their T cell receptor (TCR). In combination with signalling (by soluble antigen) through the B cell receptor (BCR), the T cells drive B cell development in the lymph node. Here, the T cell-dependent B cell development results in maturation of the antibody response to increase antibody affinity and induce different antibody isotypes. The production of short-lived plasma cells, which actively secrete antibodies specific for the vaccine protein, produces a rapid rise in serum antibody levels over the next 2 weeks. Memory B cells are also produced, which mediate immune memory. Long-lived plasma cells that can continue to produce antibodies for decades travel to reside in bone marrow niches. CD8+ memory T cells can proliferate rapidly when they encounter a pathogen, and CD8+ effector T cells are important for the elimination of infected cells. through the production of antibody. Another important to future infections with different pathogens, so called feature of vaccine-induced protection is the induction non-specific effects, perhaps by stimulating prolonged Immune memory of immune memory. Vaccines are usually developed to changes in the activation state of the innate immune The capacity of the immune prevent clinical manifestations of infection. However, system. system to respond quicker some vaccines, in addition to preventing the disease, and more effectively when a may also protect against asymptomatic infection or col- Immune memory. In encountering a pathogen, the pathogen is encountered again after an initial exposure that onization, thereby reducing the acquisition of a patho- immune system of an individual who has been vaccinated induced antigen-specific B cells gen and thus its onward transmission, establishing herd against that specific pathogen is able to more rapidly and and T cells. immunity. Indeed, the induction of herd immunity is more robustly mount a protective immune response. perhaps the most important characteristic of immuni- Immune memory has been shown to be sufficient for Incubation period The period from acquisition of zation programmes, with each dose of vaccine protect- protection against pathogens when the incubation period a pathogen to the development ing many more individuals than the vaccine recipient. is long enough for a new immune response to develop of symptomatic disease. Some vaccines may also drive changes in responsiveness (Fig. 4a). For example, in the case of HBV, which has an 88 | February 2021 | volume 21 www.nature.com/nri  Reviews a Primary immune Secondary immune Fig. 4 | Immune memory is an important feature of response response vaccine-induced protection. Antibody levels in the circulation wane after primary vaccination, often to a level below that required for protection. Whether immune Antibody titre memory can protect against a future pathogen encounter Protective depends on the incubation time of the infection, the antibody quality of the memory response and the level of antibodies level induced by memory B cells. a | The memory response may be sufficient to protect against disease if there is a long Vaccine incubation period between pathogen exposure and the antigen Exposure to Long incubation period onset of symptoms to allow for the 3–4 days required for exposure pathogen (for example, hepatitis B memory B cells to generate antibody titres above the virus): antibody levels protective threshold. b | The memory response may not above protective be sufficient to protect against disease if the pathogen threshold has a short incubation period and there is rapid onset of symptoms before antibody levels have reached the b Primary immune Secondary immune protective threshold. c | In some cases, antibody levels after response response primary vaccination remain above the protective threshold and can provide lifelong immunity. Antibody titre Protective administered. For example, the virus-like particles used antibody in the HPV vaccine induce antibody responses that can level persist for decades, whereas relatively short-term anti- body responses are induced by pertussis vaccines; and Vaccine the inactivated measles vaccine induces shorter-lived antigen Exposure to exposure antibody responses than the live attenuated measles pathogen vaccine. Short incubation period So, for infections that are manifest soon after acqui- (for example, Haemophilus influenzae type B): insufficient sition of the pathogen, the memory response may be time to raise protective insufficient to control these infections and sustained antibody levels immunity for individual protection through vaccina- c Primary immune tion can be difficult to achieve. One solution to this response is the provision of booster doses of vaccine through childhood (as is the case, for example, for diphtheria, tetanus, pertussis and polio vaccines), in an attempt to Antibody titre sustain antibody levels above the protective threshold. Protective It is known that provision of five or six doses of teta- antibody nus45 or diphtheria46 vaccine in childhood provides life- level long protection, and so booster doses of these vaccines Lifelong protective antibody throughout adult life are not routine in most countries Vaccine levels after vaccination that can achieve high coverage with multiple childhood antigen (for example, yellow fever) exposure doses. Given that, for some infections, the main burden is in young children, continued boosting after the second year of life is not undertaken (for example, the invasive incubation period of 6 weeks to 6 months, a vaccinated bacterial infections including Hib and capsular group B individual is usually protected following vaccination meningococci). even if exposure to the virus occurs some time after The exception is the pertussis vaccine, where the vaccination and the levels of vaccine-induced antibody focus of vaccine programmes is the prevention of dis- have already waned41. Conversely, it is thought that ease in infancy; this is achieved both by direct vacci- immune memory may not be sufficient for protection nation of infants as well as by the vaccination of other against rapidly invasive bacterial infections that can age groups, including adolescents and pregnant women cause severe disease within hours or days following in some programmes, to reduce transmission to infants acquisition of the pathogen42 (Fig. 4b). For example, and provide protection by antibody transfer across the there is evidence in the case of both Haemophilus influ- placenta. Notably, in high-income settings, many coun- enzae type B (Hib) and capsular group C meningococcal tries (starting in the 1990s) have switched to using the infection that individuals with vaccine-induced immune acellular pertussis vaccine, which is less reactogenic than memory can still develop disease once their antibody (and therefore was thought to be preferable to) the older levels have waned, despite mounting robust, although whole-cell pertussis vaccine that is still used in most not rapid enough, memory responses43,44. The waning of low-income countries. It is now apparent that acellular Booster doses antibody levels varies depending on the age of the vac- pertussis vaccine induces a shorter duration of protec- Repeat administration of a vaccine after an initial priming cine recipient (being very rapid in infants as a result of tion against clinical pertussis and may be less effective dose, given in order to enhance the lack of bone marrow niches for B cell survival), the against bacterial transmission than is the whole-cell the immune response. nature of the antigen and the number of booster doses pertussis vaccine47. Many high-income countries have NaTure RevIeWS | ImmUnology volume 21 | February 2021 | 89 Reviews Vaccine coverage below Vaccine coverage above No vaccination threshold for herd protection threshold for herd protection Infection passes from individuals with disease Infection can still pass to susceptible Infection cannot spread in the population to susceptible individuals and spreads individuals and spread throughout the and susceptible individuals are indirectly throughout the population population except to those who are vaccinated protected by vaccinated individuals Diseased Susceptible Vaccinated Fig. 5 | Herd immunity is an important feature of vaccine-induced protection. The concept of herd immunity for a highly contagious disease such as measles. Susceptible individuals include those who have not yet been immunized (for example, being too young), those who cannot be immunized (for example, as a result of immunodeficiency), those for whom the vaccine did not induce immunity, those for whom initial vaccine-induced immunity has waned and those who refused immunization. observed a rise in pertussis cases since the introduc- pathogen, potentially putting them at increased risk of tion of the acellular vaccine, a phenomenon that is not infection or more severe disease. Strategies to overcome observed in low-income nations using the whole-cell this include the use of adjuvants that stimulate innate vaccine48. immune responses, which can induce sufficiently By contrast, lifelong protection seems to be the rule cross-reactive B cells and T cells that recognize differ- following a single dose with some of the live attenuated ent strains of the same pathogen, or the inclusion of as viral vaccines, such as yellow fever vaccine49 (Fig. 4c), many strains in a vaccine as possible, the latter approach although it is apparent that protection is incomplete obviously being limited by the potential of new strains to with others. In the case of varicella zoster and measles– emerge in the future54. mumps vaccines, some breakthrough cases are described during disease outbreaks among those individuals who Herd immunity. Although direct protection of individu- have previously been vaccinated, although it is unclear als through vaccination has been the focus of most vac- whether this represents a group in whom immunity has cine development and is crucial to demonstrate for the waned (and who therefore needed booster vaccination) licensure of new vaccines, it has become apparent that or a group for whom the initial vaccine did not induce a key additional component of vaccine-induced protec- a successful immune response. Breakthrough cases are tion is herd immunity, or more correctly ‘herd protection’ less likely in those individuals who have had two doses (Fig. 5). Vaccines cannot protect every individual in a of measles–mumps–rubella vaccine50 or varicella zos- population directly, as some individuals are not vacci- ter vaccine51, and cases that do occur are usually mild, nated for various reasons and others do not mount an which indicates that there is some lasting immunity to immune response despite vaccination. Fortunately, how- the pathogen. ever, if enough individuals in a population are vaccinated, An illustration of the complexity of immune mem- and if vaccination prevents not only the development of ory and the importance of understanding its underlying disease but also infection itself (discussed in more detail immunological mechanisms in order to improve vac- below), transmission of the pathogen can be interrupted cination strategies is provided by the concept of ‘origi- and the incidence of disease can fall further than would nal antigenic sin’. This phenomenon describes how the be expected, as a result of the indirect protection of immune system fails to generate an immune response individuals who would otherwise be susceptible. against a strain of a pathogen if the host was previously For highly transmissible pathogens, such as those exposed to a closely related strain, and this has been causing measles or pertussis, around 95% of the popu- demonstrated in several infections, including dengue52 lation must be vaccinated to prevent disease outbreaks, and influenza53. This might have important implica- but for less transmissible organisms a lower percentage tions for vaccine development if only a single pathogen of vaccine coverage may be sufficient to have a substan- strain or pathogen antigen is included in a vaccine, as tial impact on disease (for example, for polio, rubella, vaccine recipients might then have impaired immune mumps or diphtheria, vaccine coverage can be ≤86%). responses if later exposed to different strains of the same For influenza, the threshold for herd immunity is highly 90 | February 2021 | volume 21 www.nature.com/nri  Reviews variable from season to season and is also confounded positive interferon-γ release assay response to M. tuber- by the variability in vaccine effectiveness each year55. culosis if they had previously been BCG vaccinated64. Modest vaccine coverage, of 30–40%, is likely to have The lack of a T cell response in previously vaccinated an impact on seasonal influenza epidemics, but ≥80% individuals indicates that the BCG vaccine induces an coverage is likely to be optimal56. Interestingly, there innate immune response that results in ‘early clearance’ might be a downside to very high rates of vaccination, of the bacteria and prevents infection that induces an as the absence of pathogen transmission in that case adaptive immune response. It will be hugely valuable will prevent natural boosting of vaccinated individuals for future vaccine development to better understand the and could lead to waning immunity if booster doses of induction of such protective innate immune responses vaccine are not used. so that they might be reproduced for other pathogens. Apart from tetanus vaccine, all other vaccines In the case of the current pandemic of the virus in the routine immunization schedule induce some SARS-CoV-2, a vaccine that prevents severe disease and degree of herd immunity (Fig. 5), which substantially disease-driven hospitalization could have a substantial enhances population protection beyond that which public health impact. However, a vaccine that could also could be achieved by vaccination of the individual only. block acquisition of the virus, and thus prevent both Tetanus is a toxin-mediated disease acquired through asymptomatic and mild infection, would have much infection of breaks in the skin contaminated with the larger impact by reducing transmission in the community toxin-producing bacteria Clostridium tetani from and potentially establishing herd immunity. the environment — so, vaccination of the community with the tetanus toxoid will not prevent an unvaccinated Non-specific effects. Several lines of evidence indicate that individual acquiring the infection if they are exposed. As immunization with some vaccines perturbs the immune an example of the success of herd immunity, vaccination system in such a way that there are general changes in of children and young adults (up to 19 years of age) with immune responsiveness that can increase protection capsular group C meningococcal vaccine in a mass cam- against unrelated pathogens65. This phenomenon has paign in 1999 resulted in almost complete elimination been best described in humans in relation to BCG and of disease from the UK in adults as well as children57. measles vaccines, with several studies showing marked Currently, the strategy for control of capsular groups A, reductions in all-cause mortality when these vaccines C, W and Y meningococci in the UK is vaccination of are administered to young children that are far beyond adolescents, as they are mainly responsible for transmis- the expected impact from the reduction in deaths attrib- sion and vaccine-mediated protection of this age group uted to TB or measles, respectively66. These non-specific leads to community protection through herd immu- effects may be particularly important in high-mortality nity58. The HPV vaccine was originally introduced to settings, but not all studies have identified the phenom- control HPV-induced cervical cancer, with vaccination enon. Although several immunological mechanisms programmes directed exclusively at girls, but it was sub- have been proposed, the most plausible of which is that sequently found to also provide protection against HPV epigenetic changes can occur in innate immune cells as infection in heterosexual boys through herd immu- a result of vaccination, there are no definitive studies in nity, which led to a marked reduction in the total HPV humans that link immunological changes after immuni- burden in the population59,60. zation with important clinical end points, and it remains unclear how current immunization schedules might Prevention of infection versus disease. Whether vac- be adapted to improve population protection through cines prevent infection or, rather, the development of non-specific effects. Of great interest in the debate, recent disease after infection with a pathogen is often difficult studies have indicated that measles disease casts a pro- to establish, but improved understanding of this dis- longed ‘shadow’ over the immune system, with deple- tinction could have important implications for vaccine tion of existing immune memory, such that children who design. BCG vaccination can be used as an example to have had the disease have an increased risk of death from illustrate this point, as there is some evidence for the pre- other causes over the next few years67,68. In this situation, vention of both disease and infection. BCG vaccination measles vaccination reduces mortality from measles prevents severe disease manifestations such as tubercu- as well as the unconnected diseases that would have lous meningitis and miliary TB in children61 and animal occurred during the ‘shadow’, resulting in a benefit that studies have shown that BCG vaccination reduces the seems to be non-specific but actually relates directly to spread of M. tuberculosis bacteria in the blood, medi- the prevention of measles disease and its consequences. Interferon-γ release assay ated by T cell immunity62, thereby clearly showing that This illustrates a limitation of vaccine study protocols: An assay in which blood is vaccination has protective effects against the develop- as these are usually designed to find pathogen-specific stimulated with Mycobacterium ment of disease after infection. However, there is also effects, the possibility of important non-specific effects tuberculosis antigens, after good evidence that BCG vaccination reduces the risk of cannot be assessed. which levels of interferon-γ (produced by specific memory infection. In a TB outbreak at a school in the UK, 29% T cells if these are present) are of previously BCG-vaccinated children had a memory Factors affecting vaccine protection measured. T cell response to infection, as indicated by a positive The level of protection afforded by vaccination is interferon-γ release assay, as compared with 47% of the affected by many genetic and environmental factors, Epigenetic changes Changes in the expression of unvaccinated children63. A similar effect was seen when including age, maternal antibody levels, prior antigen genes that do not result from studying Indonesian household members of patients exposure, vaccine schedule and vaccine dose. Although changes in DNA sequence. with TB, who had a 45% reduced chance of developing a most of these factors cannot be readily modified, age of NaTure RevIeWS | ImmUnology volume 21 | February 2021 | 91 Reviews vaccination and schedule of vaccination are important of childhood meningitis and bacterial pneumonia, and and key factors in planning immunization programmes. the development of the conjugate vaccine technology in The vaccine dose is established during early clinical the 1980s has transformed global child health9. development, based on optimal safety and immunogenic- Immune responses are also poor in the older popu­ ity. However, for some populations, such as older adults, lation and most of the vaccines used in older adults a higher dose might be beneficial, as has been shown for offer limited protection or a limited duration of protec- the influenza vaccine69,70. Moreover, intradermal vaccina- tion, particularly among those older than 75 years of tion has been shown to be immunogenic at much lower age. The decline in immune function with age (known (fractional) doses than intramuscular vaccination for as immunosenescence) has been well documented79 but, influenza, rabies and HBV vaccines71. despite the burden of infection in this age group and the increasing size of the population, has not received Age of vaccination. The highest burden of and mortality sufficient attention so far amongst immunologists from infectious disease occur in the first 5 years of life, and vaccinologists. Interestingly, some have raised the with the youngest infants being most affected. For this hypothesis that chronic infection with cytomegalovirus reason, immunization programmes have largely focused (CMV) might have a role in immunosenescence through on this age group where there is the greatest benefit from unfavourable effects on the immune system, including vaccine-induced protection. Although this makes sense clonal expansion of CMV-specific T cell populations, from an epidemiological perspective, it is somewhat known as ‘memory inflation’, and reduced diversity of inconvenient from an immunological perspective as the naive T cells80,81. induction of strong immune responses in the first year of In high-income countries, many older adults receive life is challenging. Indeed, vaccination of older children influenza, pneumococcal and varicella zoster vaccines, and adults would induce stronger immune responses, although data showing substantial benefits of these but would be of little value if those who would have ben- vaccines in past few decades in the oldest adults (more efited from vaccination have already succumbed to the than 75 years of age) are lacking. However, emerging disease. data following the recent development and deployment It is not fully understood why immune responses of new-generation, high-dose or adjuvanted influenza to vaccines are not as robust in early infancy as they vaccines82 and an adjuvanted glycoprotein varicella zos- are in older children. One factor, which is increasingly ter vaccine83 suggest that the provision of additional sig- well documented, is interference from maternal anti- nals to the immune system by certain adjuvants (such as body72 — acquired in utero through the placenta — AS01 and MF59) can overcome immunosenescence. It is which might reduce antigen availability, reduce viral now necessary to understand how and why, and to use replication (in the case of live viral vaccines such this knowledge to expand options for vaccine-induced as measles 73) or perhaps regulate B cell responses. protection at the extremes of life. However, there is also evidence that there is a physio- logical age-dependent increase in antibody responses in Schedule of vaccination. For most vaccines that are used infancy72. Furthermore, bone marrow niches to support in the first year of life, 3–4 doses are administered by B cells are limited in infancy, which might explain the 12 months of age. Conventionally, in human vaccinol­ very short-lived immune responses that are documented ogy, ‘priming’ doses are all those administered at less in the first year of life74. For example, after immuniza- than 6 months of age and the ‘booster’ dose is given at tion with 2 doses of the capsular group C meningococcal 9–12 months of age. So, for example, the standard WHO vaccine in infancy, only 41% of infants still had protec- schedule for diphtheria–tetanus–pertussis-containing tive levels of antibody by the time of the booster dose, vaccines (which was introduced in 1974 as part of the administered 7 months later75. Expanded Programme on Immunization84) consists of In the case of T cell-independent antigens — in other 3 priming doses at 6, 10 and 14 weeks of age with no words, plain polysaccharides from Hib, typhoid-causing booster. This schedule was selected to provide early bacteria, meningococci and pneumococci — animal data protection before levels of maternal antibody had indicate that antibody responses depend on development waned (maternal antibody has a half-life of around of the marginal zone of the spleen, which is required for 30–40 days85, so very little protection is afforded to the maturation of marginal zone B cells, and this does not infants from the mother beyond 8–12 weeks of age) and