Abstract

1. Measles in Developed Countries: the Need for New Knowledge

Despite the existence of an effective measles vaccine, 266,701 measles cases were reported worldwide in 2014 with more than 146,000 measles-related deaths reported in 2013 (the majority of fatal measles cases occurring in Africa and Asia) [1]. In recent years, there has been a resurgence of measles cases in the United States and across Europe [2]. From 2010 to 2014, the European region reported 135,600 measles cases, with 26,436 and 14,059 cases in 2013 and 2014, respectively; large outbreaks were recorded (mainly among unvaccinated and individuals with unknown vaccine status) in France, Spain, Italy, Germany and Romania [2–6]. In 2014 alone, the U.S. reported 668 measles cases across 27 states—the highest number of annual cases since the U.S. measles elimination declaration in 2000 [2]. During January – April 2015, a total of 159 measles cases (of which 18% had received measles vaccine) were reported to the U.S. Centers for Disease Control and Prevention [5, 6]. The vast majority of measles cases are due to failure in administering or receiving the vaccine [6]. However, in countries with high measles vaccine coverage, outbreaks have revealed measles vaccine failure among individuals previously vaccinated with two doses of measles-containing vaccine [2, 3, 5, 7–10].

Given the ongoing public health threat of measles, it is critical to understand the development and determinants of measles vaccine immunogenicity – both those that drive initial protective responses and those that lead to vaccine failure. In this review, we examine measurements of measles-specific humoral immunity, vaccine correlates of protection, and factors associated with variability in measles-specific humoral immunity, with a focus on immunogenetics. We discuss how new “OMICS” technologies, systems biology and vaccinomics approaches to studying vaccine responses can be applied to explain the variations in immune responses to the measles vaccine. These new developments, in addition to available datasets for other vaccines at a human systems level, offer an exciting opportunity to search for evidence of common immune responses, pathways and signatures among various infectious diseases following immunization. Furthermore, these current technological advances may indeed serve to better identify specific biomarkers of vaccine immunogenicity, and/or any potential adverse reactions presented in response to one or several group(s) of vaccines.

2. Variation in Measles Vaccine Responses: General Principles

Primary vaccine failure arises when a vaccinated individual does not develop a protective immune response after immunization. Secondary failure (waning immunity) occurs when an individual develops a protective immune response after vaccination (based on the established correlates of protection), but the vaccination fails to protect the vaccinated individual from subsequent infection upon exposure. The current measles vaccines available in the U.S. contain the Edmonston-Enders-based Moraten measles strain in combination with other viruses: measles-mumps-rubella (MMR), or measles-mumps-rubella-varicella (MMRV). Other Edmonston-based strains used worldwide, with similar immunogenicity and safety profiles, include the Schwarz (produced in Brazil and Europe); the Edmonston-Zagreb (the most frequently used vaccine in the WHO immunization programs, India, Croatia, Switzerland); and the AIK-C strain (used in Japan). The non-Edmonston-based vaccines are derived independently and include the CAM-70 (produced and used in Japan and Indonesia); the Leningrad-16 (produced and used in Russia); the Changchun-47 and the Shanghai-191 strains (produced and used in China) [11]. It was anticipated that a two-dose MMR vaccination program would lead to substantial reductions in measles morbidity and measles elimination (Box 1); however, various studies have approximated that 2–10% of individuals vaccinated with two MMR doses may not develop or sustain protective measles humoral immunity, allowing a gradual accumulation of individuals susceptible to infection and subsequently, the occurrence of viral outbreaks [2–4, 6–10, 12, 13].

The correlate of protection for measles is based on measles-specific humoral immunity; namely, an antibody response. The current “gold standard” is based on quantification of neutralizing antibodies against the viral hemagglutinin (H) and fusion (F) surface glycoproteins by the plaque reduction neutralization test (PRN), or its high-throughput version, the PRMN assay [12, 14, 15, 16]. Experimental evidence (serum depletion studies of H- and F-specific antibodies) suggests that H- (to a greater extent) and F-specific antibodies contribute to virus neutralization and protection [17, 18]. Thus, neutralizing antibodies are the correlate of protection which assesses disease susceptibility, with measles antibody titers ≥120 mIU/ml (or PRN/PRMN titer of 120) rendering protection against the disease, and titers ≥1,000 mIU/ml rendering protection against both infection and disease [16]. The importance of cellular immunity to vaccine-induced protection is not completely understood, though data suggest that seronegative vaccinated children may still have protection against measles, thus supporting an involvement of cellular immunity [16, 19].

In a study of 763 healthy children living in a community with no circulating natural infection (and no immune response boosting from wild type virus exposures), only 91% demonstrated protective humoral immunity, 7.4 years (median) after vaccination with two medical record-documented doses of MMR vaccine [12]. This study also demonstrated inter-individual variations in neutralizing antibody titers to measles (GMT of 832 mIU/mL [95% CIs: 776; 891]) and no correlations of antibody titer with cellular immunity measures [12, 20]. Additional information was gathered from studies demonstrating the variability and persistence of measles-specific antibodies after two measles vaccine doses, as summarized in Table 1.

3. HLA Influence on Antibody Variations to Measles Vaccination

There is a growing interest in applying novel vaccinomics approaches to vaccine studies. Vaccinomics is generally defined as “the integration of a systems biology approach with immunogenetics, immunogenomics, immune profiling and functional studies in order to understand and predict vaccine-induced immune responses” [21–23].

It is believed that host genetic factors, including both major histocompatibility human leukocyte antigen genes (HLA) and non-HLA genes, as well as other variables (Key Figure, Fig. 1) modulate immunity following measles vaccination. A study conducted on twins revealed a strong genetic influence on the variance in measles vaccine-specific humoral immune responses (heritability 88.5%, p<0.0001) [24, 25], and we have estimated (previously described methodology [26]) that single nucleotide polymorphisms (SNPs), together with HLA alleles, explain ~30% of the inter-individual variability in measles-specific antibody titers [3]. A large body of evidence has now demonstrated that several markers in HLA genes (highly polymorphic gene region on human chromosome 6p21.31) are associated with the heterogeneity of antibody responses to measles vaccination [27–31]. A replication study using two independent cohorts (346 and 388 healthy individuals in cohorts 1 and 2, respectively) demonstrated that specific class I HLA alleles (B*35:03 and B*57:01) and class II alleles (DQA1*02:01, DQB1*03:03, DQB1*06:02, and DRB1*07:01) were associated with significant differences in antibody titers following two doses of the measles-containing vaccine [31]. Furthermore, the HLA-B7 supertype that shares common peptide binding motifs between various class IB alleles was associated with approximately 18% higher measles vaccine-induced antibodies in both cohorts [31]. There is indication that low antibody responsiveness after a first (single) dose of measles-containing vaccine is at least partially influenced by class II HLA alleles, such as DRB1*03 (p=0.004) and DQA1*02:01 (p=0.05) [27]. Conversely, very high levels of measles-specific antibodies following a first vaccine dose were found to be associated with HLA allelic variants, such as B*7 (p=0.05), DQA1*01:04 (p=0.02), and DPA1*02:02 (p=0.04) [28]. However, the second dose of measles vaccination had an extinguishing effect on these HLA allele associations, suggesting that an additional dose may overcome HLA restriction [32].

Open in a separate window

Fig. 1

Naïve individuals (white color, no previous exposure to measles vaccine) are vaccinated with one or two doses of MMR vaccine. An array of non-genetic (upper box panel) and genetic factors (lower box panel) perturbate the development and/or persistence of measles-specific humoral immunity, leading to a spectrum (different shades of gray) of antibody titers in vaccinated individuals.

HLA haplotypes associated with measles-specific low antibody response included class I A*24-Cw*03-B*15 (p=0.04), and class II DRB1*07-DQB1*02-DPB1*02 (p=0.05) and DRB1*07-DQB1*03-DPB1*04 (p=0.001) haplotypes [31, 33]. In contrast, the class II haplotype DRB1*15/16-DQB1*06-DPB1*04 (p=0.02) was associated with high measles virus antibody titers after two doses of the measles vaccine [31, 33]. These population-based studies provided evidence that HLA genes and haplotypes, as well as immunologically relevant non-HLA class III genes (i.e., TNF, LTA, and others) [34], are important determinants in predicting an individual’s humoral response to measles vaccination.

The immune response to major measles virus structural proteins is controlled by T lymphocytes that recognize measles virus-derived endogenous and exogenous antigens in connection with specific HLA class I and class II molecules, respectively. Mass spectrometry studies have allowed the isolation and identification of naturally processed as well as HLA-presented measles virus-derived peptides from HLA class I alleles (A*02:01, B*27:05), and the class II (DRB1*03:01) allele [35, 36], associated with measles vaccine non-responses [27]. By isolating HLA-restricted peptides, novel promiscuous multi-peptide HLA supertype vaccines administered with specific directed adjuvants can be developed and used with greater efficacy [21]. These immunogenic measles virus peptides are being utilized in the design of novel measles vaccine candidates to overcome HLA polymorphic restrictions [37].

4. Non-HLA and Multigenic Influences on Antibody Variations to Measles Vaccination

Recent candidate gene association studies have shown that genetic variation in non-HLA genes is linked to variations in humoral immune responses after measles vaccination. Polymorphisms in genes involved in viral recognition, viral binding/entry (SLAM, CD46) and pathogen-associated molecular patterns (PAMPs) sensing (TLRs, CD209/DC-SIGN); as well as cytokine/cytokine receptor genes (IL2, IL10, IL12B, IL4RA, IL12RB1, IL7R, IL6, TNFA); antiviral genes (TRIM5, DDX58/RIG-I, OAS1, ADAR, MX2, OAS3, VISA); vitamin A and vitamin D receptor genes (RARA, RARB, RARG RXRA), and a growing set of other innate immunity-related or signaling genes, have been associated with a variation in measles-specific neutralizing antibodies or IgG titers after vaccination [3, 21, 22, 31, 37–52].

Host antiviral defenses and immune response activation after live viral vaccination and/or infection requires efficient recognition (related to viral cell entry and sensing) to trigger innate and adaptive immunity. Evidence from recent reports points to an association of genetic polymorphisms in the measles virus cell entry receptors CD46 and SLAM (and in particular CD46) with variations in humoral immunity after vaccination, as well as with adverse events following administration of the measles virus-containing vaccine [38, 40, 42, 49, 51, 53, 54] (Table 2). For example, CD46 is an ubiquitously expressed type I membrane cofactor protein and a complement regulatory factor acting as a cellular-entry receptor for measles virus vaccine strains. A candidate gene association study of 339 schoolchildren (12–18 years of age, 94% Caucasians) receiving two doses of the measles-mumps-rubella (MMR, Merck) vaccine demonstrated a significant association (p=0.01) of an intronic SNP in the CD46 gene (rs2724384) with measles vaccine-specific IgG antibody titers [40]. This result was also replicated in two subsequent independent studies [42, 49] (Fig. 1, Table 2). Increased representation of the minor G allele for this CD46 SNP was significantly associated (p = 0.0007) with a two-fold decrease in neutralizing antibody titers in a cohort of 745 individuals (80.3 % Caucasians; median age at enrollment = 15 years), following two doses of MMR vaccine [42]. Concomitantly, CD46 SNP rs2724384 was also associated with variations in IFN-α (p = 0.006), IL-6 (p = 0.027) and TNF-α (p = 0.0006) secretion [42]. This SNP was also significantly associated with neutralizing antibody responses in a multivariate/multigenic assessment model (p=0.0006) [38]. In a cohort of 137 Australian children (12 to 14 months old) following a primary dose of measles vaccination, the CD46 polymorphism was significantly associated with measles virus-specific IgG levels (p=0.018) [49]. Notably, a recently published state-of-the-art genome-wide association-replication study in a Danish population revealed a high correlation between the CD46 (1q32.2) genetic region and the intronic SNP rs2724384 with MMR-related febrile seizures (p=1.7E-10), where the risk allele [A] for febrile seizures is related to higher antibody titers [54]. Furthermore, data from the 1000 Genomes Project, based on genome and transcriptome sequencing and analysis of lymphoblastoid cell lines of 462 subjects, revealed an association of the CD46 SNP rs2724384 [A] with increased CD46 gene expression [55], and provided evidence of the functional consequences of this and/or other tagged causal CD46 SNPs. A coding synonymous genetic variant (rs164288, Thr210Thr) in the SLAM gene has also been associated with measles vaccine-induced antibody levels [40] (Table 2). Although this genetic association was not successfully replicated in two independent studies [42, 51], the same SLAM polymorphism was found to be associated with a measles-specific IFNγ ELISPOT response after measles vaccination [42]. The genetic influence of the newly discovered measles virus receptor PVRL4 (nectin-4) [56, 57] on antibody responses has yet to be elucidated. In summary, several genetic association and replication-validation studies implicate the CD46 genetic region in the regulation of humoral immunity after measles vaccination, and in susceptibility to adverse events (febrile seizures) after vaccination. Further fine mapping and functional studies are warranted to identify the causal genetic risk factors (variants), and to elucidate the downstream immunologic consequences of measles vaccination.

Given the profound importance of cytokines and their receptors in innate and adaptive immunity, these genes have been the focus of several measles vaccine immunogenetic studies, as reviewed in Haralambieva, et al. [3]. For example, a significant association of the functional genetic variant rs3212227 (TaqI polymorphism; located in the 3′UTR region of the IL12B gene) with variations in measles vaccine-induced IgG antibody titers (p =0.01) has been reported [39] (Fig. 1; Table 2). This genetic association has been subsequently validated in a larger genetic association study of 994 common cytokine and cytokine receptor SNPs in 764 subjects (p=0.037) [44]. Interleukin-12 (IL-12) is a proinflammatory cytokine with a key role in Th1 differentiation, IFNγ secretion, and induction of Th1 immune responses. It can thus modulate the Th1/Th2 cytokine balance and indirectly have an effect on antibody responses. The replicated 3′UTR TaqI polymorphism (rs3212227) can potentially influence transcriptional activity, mRNA stability and protein abundance, and was previously reported to correlate with IL-12 protein secretion, gene expression and antibody titers following HBsAg immunization [58, 59]; therefore, it is a promising target for future functional studies.

The discovery of gene polymorphisms having a potential impact on measles vaccine-induced humoral immunity offers a unique opportunity to assess high-priority candidates in vaccination strategies. Focused fine-mapping and functional studies could subsequently aim to validate the significance and biological relevance of given polymorphisms in measles vaccination. This research may also increase our understanding of the mechanisms through which such polymorphisms affect gene function and, ultimately, the responses and ensuing immunological memory to a measles vaccine. As such, functional characterization studies of measles-specific gene associations are being performed, and are reviewed elsewhere [3].

Genetic data from candidate gene panels has provided the identification of SNP combinations that conjointly contribute to variability in antibody responses using a multigenic analysis approach [3, 38]. This approach identified genetic variants in the CD46 (rs2724384), innate pattern recognition receptor genes (DDX58, TLR2, TLR4, TLR5, TLR7), antiviral effector and signaling genes (e.g., TRIM5, VISA), and cytokine/cytokine receptor genes (IFNA1, IL4, IL6, IL8RA, IL2RB) that cooperatively contribute to the observed variability in measles-specific antibody responses in a multigenic fashion [3, 38] (Fig. 1; Table 2). It is important to note that while most of these findings are specific to the measles component of the MMR vaccine (the immune outcomes in these analyses were measles-specific IgG or neutralizing antibodies), some of the genetic associations were described together with the antibody responses against the rubella component of the vaccine (e.g., DDX58) [60].

At this point, due to the limited number of consistent findings, it is difficult to integrate the current knowledge into a more comprehensive view of the mechanisms leading to hyper- and hypo-responsiveness to the measles vaccine. However, it is obvious that initial infection stages (infection of susceptible cells and activation of innate immunity) are crucial for the genesis of an adaptive immune response to measles vaccination. Protective immunity against measles begins with (and is greatly influenced by) the infection of susceptible cells through measles-specific cellular surface receptors (e.g., CD46) as well as innate receptor mediated-sensing of pathogen-associated molecular patterns, such as through Toll-like receptors and RIG-I (DDX58). Together, innate receptors trigger various intracellular signaling cascades and IFN responses. Secretion of type I interferons (e.g., IFNα), pro-inflammatory cytokines (e.g., IL-6) and chemokines, initiate innate antiviral responses (e.g., VISA, TRIM5), induce HLA expression, differentiation and activation of antigen-presenting cells, and augment cellular chemotaxis and inflammation. It is well known that these events trigger and amplify measles-specific adaptive immunity.

A next step in vaccinology is to uncover the yet unknown genetic (or other) factors influencing measles vaccine-induced immunity. This could be accomplished in part, by performing larger genome-wide association/replication studies (such as those performed for other viral vaccines, [61–63]) to interrogate genome associations amongst millions of SNPs and specific measles vaccine-induced phenotypes of interest (many studies are currently in progress). While such studies pose analytic challenges, they also present a unique opportunity for comprehensive gene set- and pathway-based assessments of genetic associations [3].

Sets of identified genetic markers and data serve as a starting point for future genetic and functional studies. Functional information will further enhance our understanding of the correlation between expression and function of the genetic variants likely to regulate inter-individual immune response variations in antibody responses following measles vaccination. Ultimately, such studies may delineate biomarkers of protective immunity or risk (adverse events or low/non-response to vaccines). This may assist the development of new or improved measles vaccines.

5. New Technologies and Approaches for Studying Immune Responses to Vaccines

As we move further away from empirical vaccine development and toward more rational and directed approaches, several critical elements are necessary for success. First, we must have a detailed understanding of the pathogenesis of the measles virus. Second, we need to define better correlates of protection—correlates that go beyond measuring antibody titers. Third, it is essential to understand what drives a vaccine response, a vaccine non-response, and even adverse events following vaccination. Fourth, it is important to develop a complete picture of vaccine-induced immunity. This picture must accurately cover both innate and adaptive immunity, as well as capture as many complex interactions between immune function-related components as possible [64].

Over the last decade, there have been multiple advances in immunology, virology, molecular biology, bioinformatics, and related research fields that have provided new tools and technologies aimed at enhancing our understanding of measles vaccine-induced immunity and directly address the aforementioned points. Although this review focuses on specific areas of measles vaccine immunity, these approaches can be, and in many cases have been, applied to other viral vaccines as well.

We have seen vast improvements in our ability to simultaneously capture high dimensional data in biology. The “OMIC” technologies are capable of measuring gene expression, protein levels, epigenetic events, and metabolic processes at a comprehensive level [65]. Thus, it is now possible to simultaneously examine changes in essentially every gene, protein, or biological product in response to a given stimulus. Indeed, sequencing experiments can now provide large gene sequence datasets, identify different isoforms, quantify and evaluate differential mRNA transcript usage, measure miRNA species, and assess DNA methylation status across the entire genome.

These advances in data acquisition have, of necessity, spurred the development of sophisticated bioinformatics algorithms capable of handling and evaluating immense datasets [66]. Many of these tools and databases are online, providing the research community with publicly available large dataset information and easily accessible tools for data mining capabilities [67]. Epitope-prediction algorithms have moved beyond simply defining potential immunogenic T and B cell epitopes, to complex sequence analysis and structural vaccinology-modeling tools that identify putative epitopes, model 3D structures of antibody-antigen interfaces, evaluate energy and binding requirements, or optimize antigen backbones, side chains and antibody interaction surfaces [68, 69]. Such approaches have not been used for measles, but have been successfully applied to the development of broadly neutralizing antibodies to other viruses such as influenza and RSV, among others [70, 71]. These online tools cover a range of applications including viral sequencing data, comparative analyses, RNA folding, miRNA and siRNA functional studies, protein-protein interactions, structural analyses, and host-pathogen interactions.

Many advanced “OMICS” technologies, combined with cutting-edge bioinformatics tools, have made it possible to study vaccine responses from a systems-level perspective; however, systems-level findings are still lacking for measles-specific vaccination and immunity [64] (Box 2). Other techniques such as flow cytometry, are readily amenable to single-cell analysis, and high-dimensional “OMICS” technologies are also being coupled to single cell analysis. Mass cytometry, or CyTOF (cytometry time of flight), utilizes mass spectrometry to dramatically enhance the number of distinguishable markers compared to conventional flow cytometry. CyTOF technology has allowed complex phenotyping of lymphocyte subsets, [72] and detailed T cell epitope mapping [73]. Others have combined flow cytometry with single-cell PCR, or next-generation sequencing technologies in order to study single cell transcriptomes [74]. In terms of immune response profiling and vaccine development, single-cell transcriptomics has provided insights into the control of dendritic cell heterogeneity and function, and can be used to assess transcriptomic changes in specific cell populations relevant to the development and persistence of protective immune responses to measles vaccines [75, 76].

Understanding the role of genetic variation in measles vaccine immunity has led to the identification of genetic polymorphisms associated with a measles vaccine response or non-response. Identifying the specific activation or suppression “genotype” of viral receptors (e.g., CD46), would potentially allow the development of new vaccine candidates – with the use of adjuvants (and/or different moieties), it might be possible to induce “correction” of an underlying genetic defect responsible for an immunosuppression phenotype. One can also consider (with regard to current knowledge) taking advantage of the RIG-I (DDX58) pathway and/or TLR agonists to boost innate immunity and elicit optimal protection (neutralizing antibody levels and cellular immunity) after measles immunization. The aforementioned immunogenetics, vaccinomics and systems vaccinology published reports [23, 65, 77–81], provide examples of how such studies might lead to new vaccine development (Box 2).

There are reports that the route of vaccine administration has a significant impact on the resulting immune response, with many studies focused on the mobilization of skin-resident APCs in order to enhance antigen presentation and subsequent adaptive immunity [82]. Therefore, there has been considerable interest in evaluating different routes of administration. A recent clinical trial found that seroconversion rates were 10% lower for the aerosolized measles vaccine than for the conventional injected vaccine [83, 84]. However, the measles vaccine administered via microneedle patch has shown promise in non-human primate studies [85]. The evaluation of immune responses has also been improved considerably by the introduction of new assays and more sophisticated animal models, to name some of the various novel vaccine approaches (Box 3). Collectively, these new techniques provide advantageous tools, allowing investigators to study the mechanisms of measles vaccine immune responses with far greater power and precision than previously possible.

To date, the nature of inter-individual variability in humoral immune responses to live attenuated measles vaccine, and the mechanisms behind it, remain only partially explained, with much yet to discover. Future research avenues and scientific questions are further summarized in this review (see Box 4 and Outstanding Questions).

6. Limitations of genetic association analyses in vaccine research

A major issue with genetic association studies is being able to adjust for all confounding variables and to control false positive associations by using stringent p-values. For agnostic GWAS, a common practice is to use a p-value threshold of 5×10⁻⁸[86]. The choice of critical p-values should consider the number of statistical tests performed, but also, the prior probability/evidence that the association between the SNP/HLA allele and the outcome is real [87]. Performing analyses that appropriately account for the large number of tests in candidate gene studies requires a Bonferroni correction (or other multiple testing approaches) to correct for false discovery rate (FDR) for the number of tested SNPs or HLA alleles, which was not performed in some of the reported findings in this review. With that in mind, the current review summarizes and interprets consistent and/or replicated findings, with the understanding that false positives are less likely to demonstrate consistent associations. In addition, due to the high dimensionality of many new technologies applicable to vaccines (e.g., mRNA-Seq, Mass spectrometry, etc.), FDRs need to be computed to assess statistical significance, as is standard practice in the field [88]. Furthermore, filtering strategies and analytical methods can be designed for data reduction and for incorporating a priori knowledge from outside sources. This would allow a knowledge-driven analytical approach to minimize false discoveries [64, 66].

7. Concluding Remarks

Recent genetic association and replication studies implicate HLA alleles, CD46 and other genetic variants of innate pattern recognition receptors, cytokine/cytokine receptors, as well as antiviral effector and signaling genes in the regulation of humoral immunity (Fig. 1), and in the susceptibility to febrile seizures following measles vaccination. Future genome-wide, systems biology and functional studies are warranted to further expand our knowledge of the genetic influences on measles vaccine-induced antibody responses. Technological and conceptual developments coupled with new genetic data now offer unprecedented opportunities to characterize the mechanisms of measles vaccine-induced immune responses from a systems-level perspective. However, important scientific questions must still be addressed (Outstanding Questions). The bigger questions are whether the aforementioned advancements and breakthroughs in the field of vaccinology enhance our ability to manipulate specific targets to obtain optimal immune responses and immunological memory. Also, can they lead to the development of prediction algorithms when assessing immune response outcomes after measles vaccination? Can these also be applied to personalized vaccination approaches, and to the evaluation of novel measles (or other pathogen) vaccines to benefit public health?

Acknowledgments

Footnotes

References