Abstract
When an individual is exposed to a viral pathogen for the first time, the adaptive immune system is naive and cannot prevent virus replication. The consequence may be severe disease. At the same time, the host may rapidly generate a pathogen-specific immune response that will prevent disease if the virus is encountered again. Parvovirus B19 provides one such example. Children with sickle cell disease can experience life-threatening transient aplastic crisis when first exposed to parvovirus B19, but an effective immune response confers lifelong protection. We briefly examine the induction and benefits of virus-induced immunity. We focus on three human viruses for which there are no licensed vaccines (respiratory syncytial virus, human immunodeficiency virus type 1, and parvovirus B19) and consider how virus-induced immunity may inform successful vaccine design.
Introduction
In the mid 20
RSV Infection and Immunity
RSV is a negative-strand RNA paramyxovirus that causes severe and potentially lethal lower respiratory tract infections (LRTI) in humans. It is estimated that these infections are responsible for ∼3 million hospitalizations worldwide each year and ∼60,000 hospital deaths among children under the age of 5 years. When community deaths are considered, the estimated number of fatalities per year rises to ≥100,000. Infants are most susceptible to disease at ∼2 months of age when maternal antibodies have waned, but endogenous B cell responses have not yet been stimulated (28,62,111,148).
Upon RSV infection, humans mount immune responses comprising RSV-specific T cells and antibodies, including IgM, IgG, and IgA. The RSV membrane proteins F and G are prominent targets of virus-specific B cells and neutralizing antibodies. Both systemic and mucosal immune responses are observed. Debates continue as to which cell type, antigen specificity, and/or function confers the best protection (7,93,96,99,103,107,122,139,150,163,165,169,170).
The protective capacity of B cell-mediated virus-induced immunity is demonstrated by the protection afforded to infants by maternal antibodies measured in cord blood and neonatal serum (48,117). In addition, when RSV hyperimmune sera are passively transferred to naive infants as prophylaxis, recipients experience reductions in rates of LRTI, frequencies of hospitalizations and intensive care unit admissions, and lengths of hospital stays (63,64).
The immune response induced by an RSV infection is not sterilizing; it does not prevent reinfection and might therefore be deemed “suboptimal” (66). Nonetheless, the most serious illnesses and hospitalizations occur in infants and young children and it is only when immune competence is compromised, as in the elderly who experience waning immunity, that the threat of serious RSV disease resurfaces (51,65,171). Vaccines designed to induce immune responses similar to responses induced by RSV infections, but mounted before a virus exposure occurs, could save an extraordinary number of human lives each year.
RSV Vaccine Development
The RSV research field has developed dozens of vaccine products (60,73,113,137,140), but has experienced numerous setbacks, most notably in the 1960s when a formalin-inactivated RSV vaccine was tested and vaccine recipients suffered worse disease than controls upon subsequent RSV exposures (32,54,84). This inactivated vaccine did not elicit strong neutralizing antibodies, a situation that rendered children susceptible to LRTI and immunopathology (108). Based on lessons learned from the clinical study and from virus-induced immunity, today's RSV vaccine candidates are more often judged by the induction of functional antibodies (60). Among many current vaccine candidates are (1) a baculovirus-expressed RSV F vaccine (ResVax), which was tested in a phase III study and reduced RSV LRTI hospitalizations among infants born to vaccinated mothers (although results fell short of expectations) (59,78), (2) a stabilized prefusion RSV F protein that significantly boosted neutralizing antibodies in a phase I study of adults (38,104), and (3) an RSV F-expressing Sendai virus-based vector that induced neutralizing antibodies in research animals and is currently in a phase I clinical trial (79,140,141,178).
HIV-1 Infection and Immunity
HIV-1 is a positive-strand RNA lentivirus. A lipid membrane, acquired by virus budding from the mammalian cell surface, is studded with the virus-encoded envelope (Env) protein. Despite the discovery of HIV-1 decades ago and an extraordinary research effort, there remains no licensed vaccine (147). According to UNAIDS, there had been a total of 32 million deaths worldwide due to AIDS-related illnesses by the end of 2018. In 2018 there were ∼37.9 million people living with HIV-1 worldwide; 1.7 million people were newly infected and only 62% of infected people were accessing treatment (unaids.org, accessed 2019).
Due to the impressive successes of antiretroviral drugs, individuals who can access and tolerate treatments usually live relatively normal lives. However, for those individuals who cannot access or adhere to treatments, HIV-1 infections predictably lead to immunodeficiencies, opportunistic infections, and death. The number of annual deaths has been reduced by >55% from its peak in 2004, but there were nonetheless 770,000 deaths due to AIDS-related illnesses in 2018 (unaids.org, accessed 2019).
Researchers have used HIV-1, SIV (simian immunodeficiency virus), and SHIV (a chimera of HIV-1 and SIV) to study lentivirus disease control and disease progression. B cell and T cell responses are rapidly induced by each of these infections. The virus-specific antibodies neutralize virus and also exhibit functions including antibody-dependent cell-mediated cytotoxicity, antibody-dependent cell-mediated virus inhibition, and phagocytosis (33,72,134,173).
The adaptive immune response cannot clear an established virus infection (105), but affords some protection against superinfections from an exogenous source (30,35,37,41,49,128,134,135,145,146,160,161,164). Moreover, the passive transfer of antibodies from infected individuals can protect against virus infections in a naive individual (52,101,114,125,138,149).
The extent of protection afforded by HIV-1, SIV, or SHIV infections is variable and depends on coevolution of endogenous virus and the immune response (35,41,129,132,172). When an individual is first infected, the founder virus displays little diversity in sequence or structure and, therefore, induces a focused B cell and T cell response that can recognize the founder, but fails to recognize virus variants (172). In privileged sites, the virus will then mutate and escape contemporaneous B cell and T cell activities. New viruses trigger new B cell and T cell effectors with diverse antibodies and T cell receptors (selected from a vast repertoire of receptor specificities) (42,100,129,132,172). In addition, B cells will undergo somatic mutations and if/when a cell expresses an antibody with improved affinity, avidity, and/or breadth (e.g., a broadly neutralizing antibody, bNab), it will expand (22,74). With each cycle of virus escape and lymphocyte activation, immune breadth improves (129,172). Eventually, the composite of B cells and T cells may recognize a plethora of diverse viral antigens and may prevent infections from an exogenous source.
In sum, the immunodeficiency virus infections serve as vaccines by providing, at least to some degree, protective immunity against repeat virus exposures. Ronen et al. showed, for example, that HIV-1 infected women in Kenya were significantly less likely than uninfected controls to be infected by exogenous HIV-1 (135). However, because the immune system cannot clear endogenous virus from privileged sites, untreated HIV-1 will ultimately deplete T cells and hamper or reverse immunity. The difficulty of generating and sustaining a protective immune response in the context of an ongoing infection with an immunodeficiency virus highlights the dire need for a successful prophylactic vaccine that induces a diverse immune response before, not after, a virus exposure occurs.
HIV-1 Vaccine Development
HIV-1 vaccine development has continued for decades. Researchers have studied dozens of unique vectors, including vectors from Semliki Forest virus, Venezuelan equine encephalitis virus, adenovirus, adeno-associated virus, modified vaccinia virus Ankara, canarypox/fowlpox, vesicular stomatitis virus, plants, bacteria, and yeast (1,14,43,67,88,98,155,174,175). External and internal HIV-1 antigens have been studied (2,29,46,102,109), each with a variety of antigen modifications. Sequences have encoded consensus, mosaic, and/or scrambled HIV-1 proteins. One concern associated with scrambled sequences is that viral protein three-dimensional structure is lost. When B cell or T cell antigenic epitopes are taken out of context, they may no longer match those of the virus (26,29,40,56,74,76,91,95,144,154,156,157,159).
Several phase III clinical trials have been performed that were not successful (e.g., with two gp120 Env proteins or with an adenovirus 5 vector expressing HIV-1 internal proteins) (36,55,133). One study (RV144) indicated partial protection by a vaccine. This study included >16,000 study participants and used a heterologous prime-boost strategy. The priming vaccine comprised a canarypox vector expressing HIV-1 Env, Gag, and protease proteins, and the booster vaccine was composed of two purified HIV-1 Env proteins in an alum adjuvant (12,130). In a modified intent-to-treat analysis, data suggested that there were ∼30% fewer HIV-1 infections in vaccinated individuals compared to controls. Several additional clinical trials are in progress (55).
Might scientists design a safe HIV-1 vaccine to match the protection conferred by virus infection? To capture the diverse HIV-1 protein structures that induce protection, researchers study viruses collected longitudinally from infected persons (100,129,132,172). These viruses may be grouped using antibody-antigen cartography (151) to define proteins with unique antigenic determinants. Then, mutually-exclusive “immunotypes” may be formulated into vaccine cocktails (21,26,75,76,80,157,179), delivered either as a mixture or in succession. The cocktail vaccine approach has repeatedly yielded successful licensed vaccines in other fields [e.g., rotavirus (34), pneumococcus (16,119), and polio (18 –20,82,112)].
In the HIV-1 vaccine field, the cocktail vaccine approach is gaining momentum. Two different goals are fueled. One goal is to induce successive somatic mutations in a B cell clone to generate a rare bNab (74). The second goal is to harness the extraordinary receptor diversity present among mammalian B cells and T cells (11,42,61,71,92), and thereby recruit an army of diverse lymphocytes to tackle HIV-1 (22,129). Perhaps the cocktail vaccine approach will ultimately advance through clinical trials, satisfy both goals, and prove successful in the HIV-1 field.
Parvovirus B19 Infection and Immunity
Parvovirus B19 is a small, nonenveloped, single strand DNA virus. The minor capsid protein (VP1) and major capsid protein (VP2) assemble with a ratio of ∼1:20 to form an icosahedron coat. VP1 and VP2 share amino acid sequence at the C-terminus. The VP1 N-terminus contains a unique region (VP1u) that is crucial for virion binding to its target cell (126). Parvovirus B19 is thought to be transmitted primarily through the respiratory route, but the virus can also infect by an individual's direct exposure to contaminated blood. The virus specifically infects erythroid progenitor cells by binding globoside (P antigen) and causes transient cessation of erythropoiesis (25,126).
Outbreaks of parvovirus infections occur frequently in schools and daycare centers and can cause symptoms in >40% of infected children (23,58,126,162). Typical presentation is as erythema infectiosum (fifth disease), first with a facial (“slapped cheek”) rash and later progressing to the development of erythematous lesions on the torso and limbs. Parvovirus B19 also infects adults with several manifestations. Vulnerable populations include pregnant women and adults who are immunocompromised. When an expectant mother is infected (47,90), massive edema can occur in the fetus (hydrops fetalis) followed by intrauterine fetal death or death of the newborn. In immunocompromised individuals, parvovirus B19 may persist in the bone marrow as a chronic infection, causing pure red cell aplasia (PRCA). Additional serious outcomes that may occur following parvovirus B19 infections include arthritis, myocarditis, encephalitis, meningitis, and neuropathy (8,17,24,110,126).
Perhaps the most predictable consequences of parvovirus B19 infections are among patients with hemolytic disorders, such as sickle cell disease (SCD). When parvovirus B19 infects and destroys erythroid progenitor cell targets, there is insufficient erythropoietic reserve to meet the high demand for erythrocytes in patients with chronic hemolytic anemia due to the short lifespan of red blood cells. Children suffer transient aplastic crisis due to aregenerative stress as anemia worsens and hemoglobin levels drop, often requiring hospitalization and transfusion. Other complications of an acute parvovirus B19 infection in patients with SCD include strokes [which are in some cases silent (118)] with permanent neurologic deficits, glomerulonephritis, cardiac dysfunction, and sometimes death (87,127,153).
The immune response to parvovirus B19 is robust, even in patients with SCD (68). Parvovirus-specific B cell and T cell responses (53,85,115,167) are predictably observed just days after infection. Parvovirus-specific IgM appears first followed by IgG. Multiple IgG subclasses, as well as IgE and IgA antibody isotypes, are observed (126). Antibodies are detected both in sera and secretions (breast milk and saliva) (131,152). Common antibody targets include structural (VP1 and VP2) and nonstructural (e.g., NS1) proteins (50,69,81,89,166,168). VP1u is an important target of neutralizing antibodies (5,6,57,83,136,142,143,176). The efficacy of antibody-mediated protection against parvovirus B19 is demonstrated by the transfer of human intravenous immunoglobulin to patients as an effective treatment for virus-associated chronic anemia due to PRCA (94,106,177).
The B cell and T cell immune responses elicited by primary infection usually provide lifelong protection from symptomatic disease [although there have been rare reports of infections in immunocompetent seropositive individuals (4,68,86)]. Such observations encourage the development of a parvovirus B19 vaccine informed by virus-induced protection.
Parvovirus B19 Vaccine Development
As for RSV and HIV-1, parvovirus B19 vaccine research has progressed for decades. In 1991, Kajigaya et al. described the production of virus-like particles (VLPs) matching the structure of the parvovirus B19 capsid (83). When parvovirus coat proteins VP1 and VP2 or VP2 alone were overexpressed in insect cells, VLPs would spontaneously assemble. These VLPs containing VP1 and VP2 retained important virus epitopes that could induce virus-specific neutralizing antibody responses typical of those observed in a virus-exposed individual. Importantly, the neutralizing antibody response correlated with VP1 content in the empty capsids.
One strategy for clinical vaccine manufacture was to coinfect insect cells with two baculoviruses that, respectively, expressed VP1 and VP2. This allowed enrichment of VP1 in VLP compared to native virions. Resultant VLPs were then formulated with MF59 adjuvant for testing in adult clinical trials (9,10). Study participants received up to three intramuscular injections with vaccine, after which virus-specific binding and neutralizing antibodies were induced. Usually responses were retained for many months. Unfortunately, in one phase I study, vaccination was associated with unexpected cutaneous eruptions at the injection site among study volunteers, and consequently, the research was temporarily halted. Ultimately, the clinical hold was lifted, but there was insufficient commercial momentum for the vaccine program to progress (15).
Several explanations were given for the unexpected cutaneous reactogenicity. One suggestion was that the reaction was directed toward baculovirus antigens contaminating the vaccine. A second suggestion was that the viral phospholipase A2 (PLA2) activity of VP1 was the cause (15).
An improved parvovirus B19 candidate vaccine was then developed. This vaccine was produced in Saccharomyces cerevisiae. Cells were transfected with a bicistronic plasmid expressing both VP1 (with a mutation that inactivated PLA2 activity) and VP2. The vector design supported better control of the VP1 and VP2 content in VLPs and a VP1:VP2 ratio of ∼1:1 that enhanced presentation of the VP1u neutralizing determinant. The vaccine was successfully tested in wild-type mice (31) and then in a mouse model for SCD (121). In both models, rapid and robust neutralizing antibody responses were observed. Results encourage further development of this attractive parvovirus B19 vaccine candidate.
Beyond the Adaptive Immune Response
The discussion above is focused on adaptive immunity, but additional factors will influence immune success.
Adjuvants
When a vaccine is replication competent, it can be administered alone, whereas protein-based and subunit vaccines often benefit from formulation with adjuvant. The alum adjuvant has been used in many of the licensed vaccines described above, whereas newer adjuvants have now been licensed for clinical use. These include MF59 and AS01, a liposome-based adjuvant containing MPL and QS-21 (44,45,97).
Innate immune cells
Apart from B cells and T cells, many other cell types play a significant role in pathogen clearance. For example, epithelial cells in skin and mucosal passages provide barrier function and serve as a first line of defense against invading pathogens. Innate cells such as macrophages and dendritic cells also counter viruses by presenting antigens to the adaptive immune system, clearing pathogens, and signaling other cells with cytokines/chemokines. Natural killer cells may now be categorized as both innate and adaptive; they do not carry classical V-D-J-C gene-encoded receptors, but they can nonetheless exhibit memory in response to a second pathogen exposure (27,116,120).
Logistical Challenges to Vaccine Development
Advancement of vaccine products from preclinical development to licensure is a difficult course, which can hamper development of promising vaccine candidates. Phase III clinical trials often require recruitment of thousands of healthy participants, a costly requirement that cannot be met by many vaccine developers. Public antivaccine sentiments now add to vaccine development difficulties (13). Because vaccines do not cure disease, their value is not always recognized. In fact, when vaccines and consequent herd immunity are successful, populations do not experience disease consequences and are not reminded of vaccine benefit. Today's negative sentiments toward vaccines are in marked contrast to the public's pro-vaccine efforts of the 1950s that underpinned Salk's polio vaccine success (82).
The Way Forward
Vaccine design may be assisted by attention to basic immune concepts, revealed by previous vaccine successes and by natural virus infections (39,77). Concepts include: Natural virus infections of the healthy host often induce a degree of protective immunity. Successful vaccines “look like” viruses and induce cross-reactive, protective immunity without disease. Vaccines are best administered before a natural virus exposure occurs. Safe vaccines can be produced with a variety of strategies, including attenuation of a human virus, isolation of a related virus from a nonhuman species (the Jennerian approach), production of VLPs, or production of individual viral antigens. A vaccine that mimics the pathogen's coat proteins/carbohydrates in three-dimensional structure, not just sequence, can induce a potent virus-specific immune response. Amino acid sequences, when taken out of context, do not necessarily mimic the pathogen's B cell or T cell epitopes. Diverse pathogens can be countered by cocktail vaccines that combine serotypes/immunotypes.
We propose that application of these basic principles, in combination with improved public education programs, will hasten the development and licensing of new vaccines.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
Writing was supported by NIH NCI P30CA21765 and ALSAC.
