Systems Approaches to Dissecting Immunity

Abstract

Systems biology is the comprehensive and quantitative analysis of the interactions between all of the components of biological systems over time. Cells of the innate immune system are the first line of defense against invading pathogens and orchestrate the ensuing adaptive response, which is critical to the establishment of long-term protective immunity. Innate immunity is well suited for systems analysis, because the relevant cells can be isolated in various functional states and many of their interactions can be reconstituted in a biologically meaningful manner. Application of the tools of systems biology to the innate immune system will enable comprehensive analysis of the complex interactions that maintain the fine balance between host defense and inflammatory disease. In this review, we discuss innate immunity in the context of the systems biology concepts, emergence, robustness, and modularity. We also describe recent efforts to apply these approaches to enable rational vaccine design and accelerate the pace of clinical vaccine trials.

1 Systems Biology

For the purpose of this review, we define systems biology as the comprehensive, quantitative, and temporal analysis of the manner in which all of the components of a biological system interact. Systems biology is a holistic rather than reductionist approach to deciphering complexity and understanding emergent properties. This approach requires the capture and integration of measurements from as many hierarchical levels of information as possible. These can include DNA sequences, RNA and protein measurements, protein–protein and protein–DNA interactions, biomodules, signaling and gene regulatory networks, cells, organs, individuals, populations, and ecologies. Raw measurements are then imported and annotated into comprehensive databases, many of which are accessible online to the scientific community. Both detailed graphical visualizations and mathematical modeling are used to integrate the vast quantities of individual data points into molecular networks that underlie the biology of the system. These models suggest specific hypotheses that are tested experimentally by selective molecular perturbations thereby tying the phenotypic features of the system directly to the behavior of protein and gene regulatory networks. Repeated cycles of iteration refine the model; ultimately, these models will explain the systems or emergent properties of the biological system of interest. Once a model is sufficiently accurate and detailed, it will allow biologists to accomplish two tasks never possible before: (1) predict the behavior of the system given any perturbation and (2) redesign or perturb the gene regulatory networks to create completely new emergent properties. This latter possibility lies at the heart of preventive medicine. Thus, systems biology is hypothesis driven, global, quantitative, iterative, integrative, and dynamic.

Maximizing the potential of systems approaches requires an interdisciplinary team of investigators that is also capable of developing the novel technologies and computational tools. In this model, biology dictates what new technology and computational tools should be developed. These tools often open new frontiers in biology that go well beyond the original question, driving an iterative cycle of development and discovery. Thus, biology drives technology and computation, and, in turn, technology and computation revolutionize biology. Biological systems, as opposed to engineered man-made systems, are not the result of a rational design process, but rather the result of a random evolutionary process that selects only for function. For this reason, reverse-engineering approaches predicated on the assumption of a rational underlying design will often fail to unravel a biological system.

1.1 Basic Concepts Crucial in Understanding Complex Biological Systems: Emergence, Robustness, and Modularity

1.1.1 Emergence

Complex systems display ‘emergent properties’ that are not present in their individual parts and cannot be predicted even with a full understanding of the parts alone. The arch is an example of an emergent property that arises from simple constituents. A comprehensive analysis of the physical properties of rocks will not predict that they give rise to an arch when assembled in a specific context. Life is emergent and not inherent in the individual components of an organism. Simply mixing DNA, RNA, proteins, carbohydrates, and lipids does not generate a biological system: life is a consequence of the specific organization of these components and interactions between them. A systems approach is, therefore, necessary to understand how the emergent properties of living organisms are derived from their individual components.

1.1.2 Robustness

Biological systems tend to maintain phenotypic stability despite diverse perturbations from the environment, stochastic events, and genetic variation. Robustness often arises as an emergent property through positive and negative feedback loops and other forms of regulatory control that constrain gene outputs at the transcriptional, translational, or post-translational levels. These feedback mechanisms insulate the system from environmental fluctuations. Robustness is also achieved through redundancy of pathways that perform the same biological function.

1.1.3 Modularity

A network module can be defined as a set of nodes that interact strongly and perform common function. Modularity can contribute to robustness by confining damage to independent parts, preventing the spread of damage to the entire network. Modularity can also contribute to evolution of the system, where adaptation can be achieved by rewiring connections between modules rather than reconstituting the modules themselves.

1.2 A Systems Biology Approach to Studying Immunity

The complex interactions within the innate immune system that result in effective host defense under normal conditions and inflammatory disease when perturbed can only be dissected in a comprehensive way by systems biology approaches. Immunology is particularly well suited for such analysis, as the cells can be isolated in various functional states and many aspects of the immune response can be reconstituted in a biologically meaningful manner. In this review, we highlight by example three aspects of the immune response that are particularly well suited to analysis by systems-level approaches. First, we describe the complex interactions between the multitude of phagocytic and pattern-recognition receptors that initiate the immune response to invading microbes. Next, we discuss a transcriptional regulatory circuit that tunes inflammatory responses in macrophages and discriminates transient from persistent stimulation. Finally, we describe how the tools of systems biology can be used to gain an understanding of the molecular and cellular interactions that govern the response to vaccination. We argue that systems approaches offer a route to accelerating the pace of efficacy trials by identifying correlates of protection.

2 Innate Immune Receptors

The recognition, phagocytosis, and presentation of pathogens by macrophages represent emergent properties that arise from the concerted action of a number of receptors and signaling pathways. Specific pathogen-derived molecules are detected by chemotactic receptors on the macrophage, leading to alterations in the cytoskeleton that culminate in directed movement. The macrophage then uses pattern-recognition receptors (PRRs), which include the Toll-like receptors (TLRs), the NOD-like receptors (NLRs), and the RIG-I-like receptors (RLRs), to identify the nature of the pathogen by recognizing specific pathogen-associated molecular patterns (PAMPs). Phagocytic receptors, such as the Fc receptor, the complement receptor, and DECTIN, bind the particle and activate signaling pathways that lead to its internalization (Underhill and Ozinsky 2002). Upon internalization, the pathogen is degraded, and pathogen-derived antigens are presented to cells of the adaptive immune system; this process of antigen presentation constitutes the mechanism by which the innate immune system instructs adaptive immunity.

It is not possible to predict the complex behavior underlying chemotaxis, phagocytosis, and antigen presentation by having a complete understanding of each individual receptor and its cognate signaling pathway in isolation. Systems biology approaches will enable an understanding of how the crosstalk between these pathways results in the emergent properties that give rise to these functional responses.

2.1 Crosstalk Between Phagocytic Receptors and PRRs

It has long been known that phagocytosis can be uncoupled from the induction of an inflammatory response (Aderem et al. 1984, 1985). For example, phagocytosis of latex beads is not accompanied by the production of arachidonic acid metabolites unless the macrophages are primed with bacterial lipopolysaccharide (LPS), in which case a synergistic response is observed (Aderem et al. 1986). Similar synergy also occurs for Fc receptor and zymosan-induced phagocytosis but not for complement-induced phagocytosis, which will not induce arachidonic acid metabolite release even with LPS priming (Aderem et al. 1986). These interactions are even more subtle when considering the internalization of bacteria. When macrophages internalize Gram-negative bacteria, tumor necrosis factor (TNF) is only produced in the presence of TLR4. By contrast, TLR2 is required for TNF production during phagocytosis of Gram-positive bacteria (Underhill and Ozinsky 2002).

Phagocytosis of fungal zymosan provides an example of how phagocytic and PRR pathways can function as interlocking pieces in their regulation of the macrophage response (reviewed in Goodridge and Underhill 2008). Zymosan is recognized by both TLR2 and DECTIN. TLR2 signaling induces inflammatory cytokines through the MyD88 pathway and activation of NF-κB but does not induce reactive oxygen species (ROS), phagocytosis, and only weak arachidonic acid release. DECTIN, which recognizes β-glucan, activates Syk kinase, induces zymosan phagocytosis, ROS induction, and weak arachidonic acid release. When both TLR2 and DECTIN are activated, inflammatory cytokine induction, ROS production, and arachidonic acid metabolism are all synergistically enhanced.

Interactions between PRR signaling and phagocytic pathways extend beyond internalization and inflammation. TLR signaling has been implicated in the enhanced maturation of phagosomes (Blander and Medzhitov 2004). More importantly, the presence of TLR ligands within a dendritic cell phagosome markedly enhances the MHC class II-mediated presentation of antigens within that phagosome (Blander and Medzhitov 2006). Thus, the entire set of functional macrophage responses to pathogens are shaped and modulated by complex interactions between PRR, phagocytic, and other pathways.

2.2 Crosstalk Between PRRs

Macrophages are not confronted with purified PAMPs in nature. Rather, they interact with complete pathogens that present a cocktail of agonists to the numerous PRRs they express (Underhill and Ozinsky 2002; Trinchieri and Sher 2007). These combinations of PAMPs enable the innate immune cell to carry out ‘multiparameter analysis’, which permits far greater accuracy in the determination of the threat. For example, if TLR4, TLR5 and the NLR IL-1β-converting enzyme protease-activating factor (IPAF) are simultaneously activated, the cell can compute that it has encountered a Gram-negative flagellated bacterium that contains a type III secretion system (Miao et al. 2006). Activation of TLR4 and TLR5 culminates in NF-κB-dependent inflammatory gene expression while detection of flagellin by IPAF recruits (Geddes et al. 2001; Poyet et al. 2001) and activates the caspase-1 inflammasome (Masumoto et al. 2003) which processes IL-1β and IL-18 for secretion (Dinarello 1998).

Dual sensing of flagellin by TLR5 and IPAF suggests a complex, two-step process for regulating the response to invading bacteria. When a macrophage encounters a Salmonella bacterium, TLR5 is initially stimulated by flagellin (in addition to activation of TLR4 by LPS). This signal induces, among others, the mRNAs encoding IL-1β and IL-18 and their precursor proteins. Once the bacterium is in the phagosome, flagellin is injected into the cytoplasm via the type III secretion system, and IPAF is subsequently activated. Conceptually, TLR signaling in the absence of NLRs may constitute a ‘yellow alert’, indicating that microbes have penetrated the physical barrier of the epithelial layer. The inflammasome NLRs, when activated in conjunction with the TLRs, may then trigger a ‘red alert’, alerting the immune system to the presence of pathogens which harbor more threatening virulence factors such as the type III secretion system (Miao et al. 2007). Signaling by TLRs alone or by NLRs alone does not initiate the red alert, and thus the red alert emerges from the convergent activation of the two pathways. IL-1β is not known to be capable of activating the inflammasome itself, and thus paracrine IL-1β signaling can propagate the yellow alert but not the red alert, which is reserved for the infected macrophage. A similar distinction between the reserved red alert status of the infected cell and the yellow alert status for neighboring cells activated by paracrine cytokine signaling has been postulated for viral nucleic acid detection (Stetson and Medzhitov 2006): cytotoxic lymphocytes and natural killer cells must be able to distinguish between virus-infected cells that should be targeted for apoptosis and cells that have been activated into an antiviral state by paracrine type I IFN signaling.

The system is even more complex than described due to crosstalk arising from simultaneous engagement of multiple TLRs and other receptor families. Viral RNA is recognized by at least five PRRs [TLR3, TLR7, TLR8, melanoma differentiation-associated gene 5 (MDA5), and RIG-I], and it is interesting to speculate on how convergent detection can lead to synergistic, virus-specific responses. Results from the Akira laboratory (Kumar et al. 2008) suggest that the adjuvant effects of the double-stranded RNA (dsRNA) analog polyinosinic–polycytidylic acid (polyI:C) arise from cooperative activation of TLR and cytoplasmic RLR pathways. Thus, pathogen recognition by the innate immune system is perhaps best considered as a process in which activation of several PRR pathways in combination gives rise to an emergent, pathogen-specific response that seeks to neutralize the threat, alert neighboring cells to the presence of microbes, and initiate an appropriate adaptive immune response.

2.3 Robustness and Modularity in Innate Immunity

While combinatorial PAMP detection by PRR pathways allows macrophages to accurately determine threat levels posed by invading pathogens, it also illustrates two additional key properties of the innate immune system: robustness and modularity.

To provide protection, the innate immune system must be robust: pathogens must be detected and the immune system alerted, even as evolution favors development of pathogen strategies to evade detection. The large number of PAMPs that may be detected by macrophage PRRs thus constitutes a robust, ‘fail-safe’ detection system: if a particular PRR fails to detect a pathogen, or if a pathogen evolves a strategy to evade a particular PRR, it nevertheless will be detected by all of the relevant remaining PRRs expressed by the cell. This level of robustness is revealed by gene-targeting studies, in which specific PRR knockouts or knockdowns fail to exhibit phenotypes. For example, TLR3, which detects viral dsRNA, when ablated does not result in universally enhanced susceptibility to viral infection (Edelmann et al. 2004), presumably because signaling by other viral RNA detectors (RIG-I, MDA5, and TLR7) is sufficient for protection. Similarly, we have demonstrated that inflammasome activation in response to Listeria monocytogenes involves detection by three or more cytoplasmic receptors: IPAF, NALP3, and at least one other NLR utilizing the adapter ASC (apoptosis-associated speck-like protein containing a C-terminal caspase recruitment domain) (Warren et al. 2008).

Modularity in the PRR pathways is typified by the modularity in the structures of the PRRs themselves. In the TLR family, for example, a less conserved N-terminalleucine-rich repeat (LRR) domain is coupled to a more highly conserved C-terminal Toll/Interleukin-1 receptor (TIR) domain (Roach 2005) by a single transmembrane domain. The LRR domains are so variable that they cannot be aligned over large evolutionary distances; alignment can only be accomplished using the TIR domains. The TIR domain couples the TLR to the restricted set of adapters [the linker adapters, MyD88 adapter-like protein (MAL) and translocating chain-associating membrane protein (TRAM), and the major signaling adapters, MyD88 and TRIF], whereas the LRR domain is responsible for PAMP recognition. Evolution of LRRs has resulted in an extraordinary diversity in ligands detected by the TLRs, giving rise to six major TLR families in vertebrates (Roach 2005). Structural studies of TLR–ligand complexes have revealed diversity in LRR ligand binding mechanisms (reviewed in (Jin and Lee 2008)). While TLR2/TLR1 heterodimer binding of Pam₃CSK₄ is achieved by hydrophobic interactions at the boundary between central and C-terminal domains (Jin et al. 2007), TLR3 dimers bind dsRNA at two regions near the N-terminal and C-terminal ends (Liu et al. 2008).

Robustness in the innate immune system emerges not only from the modularity of the PRRs and the pathways they activate but also from the feedback architectures of the pathways themselves. Type I IFN induction by cytoplasmic viral sensors in fibroblasts is an example of a positive feedback loop which results in robust induction of an antiviral state (reviewed in Honda et al. 2006). Cytoplasmic detection of viral RNA by the RLRs RIG-I or MDA5 results in type I IFN induction by activated IFN regulatory factor-3 (IRF3) and IRF7 transcription factors. The type I IFN then feeds back on the cells in an autocrine manner to induce IRF7 to high levels. IRF7 then induces additional type I IFN species and increases the expression of the sensors RIG-I and MDA5 themselves, which presumably renders the cell more sensitive to viral RNA. On the other hand, precise control and robustness to intracellular noise is partly achieved in PRR pathways by negative feedback loops. For example, TLRs induce the expression of many genes that negatively regulate the TLR pathways (reviewed in Liew et al. 2005). In particular, the ubiquitin-editing protein A20 (Tnfaip3) is both induced by and is a negative regulator of TLR, RLR, and NLR pathways (Boone et al. 2004; Wang 2004; Saitoh et al. 2005; Lin et al. 2006; Hitotsumatsu et al. 2008), acting directly on key adapter molecules such as tumor necrosis factor receptor-associated factor 6 (TRAF6), TRIF, and receptor-interacting protein 2 (RIP2). The second example of this type of regulation is illustrated by the network containing the transcription factors NF-κB (Rel), ATF3, and C/EBPδ. (See Fig. 1).

Fig. 1

Regulation of cytokine production in macrophages by the NF-κB, ATF3, and C/EBPδ circuit. Stimulation of TLR4 activates NF-κB, which initiates transcription of a number of cytokines. ATF3 is also activated and represses transcription of a subset of these cytokines. NF-κB and ATF3 also act similarly to modulate transcription of Cebpd which amplifies the transcription of a subset of cytokines as well as itself. Cytokines were classified into three categories based on genome-wide localization analysis of NF-κB, ATF3, and C/EBPδ and the classification confirmed by measuring transcriptional responses in ATF3^−/− and Cebpd^−/− macrophages. (Adapted from (Alon 2007))

3 Network Analysis of Innate Immune Responses

3.1 A Network that Enables Innate Immune Cells to Discriminate Between Transient and Persistent Activation

It is well established that transcriptional programs are propagated by sequential cascades of transcription factors (Bolouri and Davidson 2003; Smith et al. 2007). We have shown that stimulation of macrophages with LPS induced the transcription of multiple clusters of transcription factors within 3 h. We used a combination of mathematical modeling and biological experiments to predict and confirm the existence of a transcriptional network involved in TLR4 activation. The power of the approach lies in its ability to rapidly identify complex interactions between transcription factors and to define the functional emergent properties of the system, which in turn suggest the molecular underpinnings of the biological response. Analysis of the transcription factors activated immediately by LPS predicted the existence of many networks involved in the TLR4 response.

One of these networks contained the transcription factors NF-κB (Rel), ATF3, and C/EBPδ (Fig. 1). High-density temporal measurements of LPS-induced binding of these transcription factors to the Il6 promoter, combined with gene-deletion studies, enabled us to construct a model of a regulatory circuit that participates in the transcription of this cytokine-encoding gene. In this model, TLR4 stimulates translocation of NF-κB to the nucleus, where it activates weak transcription of Il6. Concomitant with that, NF-κB induces C/EBPδ, which then binds to the Il6 promoter and acts together with NF-κB to stimulate maximum transcription of Il6. At a later time point, ATF3 attenuates transcription of Cebpd and Il6. ATF3 recruits histone deacetylase 1 to the Il6 promoter in an LPS- dependent way. The ATF3-associated histone deacetylase 1 then deacetylates histones, resulting in the closure of chromatin and inhibition of Il6 transcription. It is known that C/EBPδ binds to and recruits the histone acetylase CBP to its target promoters, leading to more histone acetylation and chromatin opening. It is, therefore, likely that epigenetic chromatin remodeling contributes to this network.

The relationship between NF-κB and C/EBPδ suggests coherent feed-forward type I regulation (Alon 2007). This type of regulation has been suggested to protect biological systems from unwanted responses to fluctuating inputs (Alon 2007). The inflammatory response is like a double-edged sword, and it is therefore critical that inflammatory cells be modulate their response appropriately. The coherent feed-forward type I regulatory circuit described above could in principle enable immune cells to distinguish transient stimuli from more dangerous persistent activation. We used a combination of motif-scanning, microarray, and ChIP-on-chip analysis to identify many LPS-induced targets of C/EBPδ. These genes showed differences in transcriptional responsiveness to persistent and transient LPS-dependent stimulation of macrophages in vitro, and many have ascribed functions in host defenses against bacterial infection. Consistent with our in vitro studies, Cebpd-null mice were able to resist transient infection with a low dose of E. coli H9049 but were highly susceptible to persistent infection with a higher dose.

In summary, we have used the tools of systems biology to show that TLR4-induced inflammatory responses are regulated by the integration of transcriptional ‘on’ and ‘off’ switches with ‘amplifiers’ and ‘attenuators’. In addition, we have demonstrated a mechanism by which the macrophages are able to discriminate between transient and persistent activation. Collectively, these regulatory elements may facilitate the maintenance of effective host defense and the prevention of inflammatory disease.

3.2 Unraveling Complexity in Innate Immune Signaling

Genetic analysis of the mouse, whether through targeted gene deletions studies, chemical- or radiation-induced mutations, or spontaneous mutations has been one of the most powerful tools for unraveling immune responses. Although knockout mice are generally produced based on a hypothesized function of the targeted gene in a particular context, many genes that were originally identified by their role in other aspects of mouse biology have subsequently been shown to impact immune responses. Furthermore, large-scale phenotypic screening of mutagenized mice can uncover unpredicted components of immune regulatory pathways. In both of these cases, considerable effort is required to determine the mechanism by which these genes impact immunity.

Systems biology approaches organize information into sets of interacting networks that can serve to contextualize the role of a gene. A reference library of networks, such as those that we defined for NF-κB, ATF3, and C/EBPδ (Fig. 1), which are generated in a highly standardized manner, can be used as a comparator to identify signaling pathways that are functionally associated with mutated genes of interest. For example, the responses of macrophages carrying a mutation or gene deletion to a panel of immune stimuli can be compared with a compendium of responses from wild-type macrophages and macrophages lacking known components of TLR-induced signaling and gene regulatory networks. By identifying overlapping patterns in the responses, a testable hypothesis for the role of the gene in the immune response, and even its likely interaction partners, can be identified. These networks are generated using thousands of data points (e.g., entire transcriptomes) making it far less likely that such an overlap occurs by chance.

We have used this approach to link the cpdm mutation in SHARPIN to pathways known to regulate TLR responses (Zak et al. 2011). We identified SHARPIN as a potential regulator of macrophage responses in the course of a systems-level transcriptional and epigenomic analysis of combinatorial TLR pathway activation. To evaluate the role of SHARPIN in innate immunity, we measured TLR responses in macrophages derived from cpdm mice, which bear a null mutation in the Sharpin gene (Seymour et al. 2007). IL-12p40 production was markedly impaired in response to nearly all TLR ligands evaluated, including Pam₃CSK₄ (TLR2), LPS (TLR4), CpG-B (TLR9), and R848 (TLR7). The cpdm mutation also strongly attenuated macrophage production of IL-12p40 in response to infection with Listeria monocytogenes, which signals through TLR2, TLR5, and various Nod-like receptor family members (Zenewicz and Shen 2007; Warren et al. 2010; Leber et al. 2008).

Transcriptome analysis of wild-type macrophages identified 400 genes induced threefold or more by a stimulation with Pam₃CSK₄. SHARPIN deficiency arising from the cpdm mutation resulted in threefold impaired induction of 87 of these genes, including many pro-inflammatory cytokines. To identify the transcription factors that mediate the effect of SHARPIN on macrophage responses, we performed promoter analysis. We used PAINT (Vadigepalli et al. 2003) to scan the proximal promoter sequences of all 400 Pam₃CSK₄-regulated genes, and we then applied the gene set enrichment analysis (GSEA) algorithm (Subramanian et al. 2005) to determine which transcription factors were associated with impaired Pam₃CSK₄ responses. The only transcription factor binding sites that were over-represented in the promoters of SHARPIN-dependent genes relative to the overall set of 400 Pam₃CSK₄-induced genes were NF-κB and AP-1. This result suggests that SHARPIN may be required for maximal NF-κB and AP-1 activation in response to TLR2 stimulation in macrophages.

We analyzed the link between SHARPIN, NF-κB, and AP-1 in greater depth by integrating the SHARPIN-dependent gene set defined above with our database of transcriptome responses in mutant macrophages. The set of 87 SHARPIN-dependent genes overlapped significantly with genes regulated by the panr2hypomorphic mutation in NEMO (Siggs 2010) a central node in the TLR2/NF-κB pathway. The extraordinarily strong association between the effects of these mutants suggested that SHARPIN might interact with NEMO. This interaction was confirmed by biochemical analysis and was abrogated by the panr2 mutation.

In addition to pinpointing the location of SHARPIN in the TLR2/MyD88/NF-κB signaling cascade, this approach also revealed a previously unknown branch point in the pathway that controls a subset of the response. Although similar, the effects of SHARPIN-deficiency on macrophage responses were weaker than those of the panr2 mutation. Some pro-inflammatory cytokine induction remain in SHARPIN-deficient macrophages that is not observed in panr2 macrophages suggesting that the panr2 mutation was also able to impair a SHARPIN-independent pathway. Recently, it was shown that a paralog of SHARPIN, RBCK1/HOIL-1L (Lim et al. 2001), interacts with NEMO as part of the LUBAC complex (Tokunaga et al. 2009), and therefore might mediate the SHARPIN-independent pathway. This hypothesis was reinforced by the observation that the panr2 mutation ablates the RBCK1–NEMO interaction. Furthermore, it has recently been shown that SHARPIN and RBCK1 are present in distinct LUBAC complexes that are both capable of polyubiquitinating NEMO (Gerlach 2011; Ikeda 2011; Tokunaga 2011). Comparison of signaling defects induced by SHARPIN deficiency and by the panr2 mutation suggested a model in which the MyD88 pathway bifurcates at NEMO. In this model, summarized in Fig. 2, maximal induction of many pro-inflammatory cytokines requires SHARPIN while the activation of a significant number of downstream genes occurs independently. Combined with transcriptional network analysis, this also suggests previously unappreciated specificity in NF-κB and AP-1 activities since these molecules are effectors for both arms of the pathway.

Fig. 2

SHARPIN is an essential adaptor distal to the branch point defined by the panr2 mutation in NEMO. a The signaling responses most strongly impaired by SHARPIN deficiency and NEMO L153P (panr2) are the phosphorylation of p105 and ERK, suggesting that p105 IκB activity and TPL2 sequestration are dominant regulators of Toll-like receptor 2 (TLR2)-induced proinflammatory cytokine expression. The greater deficiency in signaling and pro-inflammatory cytokine induction observed in panr2 compared with cpdm macrophages may result from SHARPIN-independent interactions between NEMO and the SHARPIN paralog and the linear ubiquitin chain assembly complex constituent RBCK1, which are also abrogated by NEMO L153P. b TLR2-induced IκBα degradation, phosphorylation of p38 and JNK, and Nfkbia gene induction were unimpaired in cpdm macrophages and panr2 mutant macrophages, implying the existence ofa branch of NEMO-dependent I-kappa-B kinase (IKK) and MAPK activity that proceeds independently of SHARPIN and NEMO residue L153. (Adapted from Zak et al. 2011)

4 Systems Vaccinology

To date, vaccines have been created by “trial and error”; vaccines are generated from related pathogens, attenuated pathogens, or pathogen components. Systems biology will enable rational vaccine design.

The response of an individual to vaccination depends on a multitude of interacting genetic, molecular, and environmental factors spanning numerous temporal and spatial scales. Systems biology provides a powerful toolset for deciphering complex biological networks and has been applied extensively to identify and contextualize novel regulators of the innate immune response (Amit et al. 2009; Zak 2011; Litvak et al. 2009; Ramsey et al. 2008; Gilchrist 2006; Suzuki 2009). The application of this approach to explore the innate-adaptive interface in the context of vaccination has already yielded new insights into the mechanisms of action of the ‘gold standard’ yellow fever vaccine YF-17D (Querec 2009; Gaucher 2008) and the seasonal influenza vaccine (Nakaya 2011). Furthermore, systems analysis of vaccination promises to generate useful biomarkers for protection and to identify mechanisms of immunogenicity that will guide rational vaccine design. As these topics have already been discussed in numerous reviews (Rappuoli and Aderem 2011; Pulendran et al. 2010; Zak and Aderem 2009; Shapira and Hacohen 2011; Gardy 2009), we will instead provide a high level perspective on systems vaccinology analysis that, by necessity, involves large number of model systems, each providing unique opportunities for discovery despite numerous practical constraints.

Systems vaccinology can be divided into five essential steps: measurements of the innate (1) and adaptive responses (2) to vaccination, determination of vaccine efficacy (3), systems-level data integration leading to the identification of biomarkers and mechanistic insights (4), and perturbation of the vaccine response in an appropriate experimental system (5). In the paragraphs below, we follow one cycle through the iterative systems vaccinology process, defining the inherent constraints and opportunities at each step.

4.1 An Iterative, Multistep Approach

• Step I The starting point of the approach is the comprehensive analysis of the innate immune response to vaccination. A wide range of technologies is employed to make these measurements including transcriptomics, high-throughput serum analyte profiling, and proteomics. Transcriptome analysis is the most reliable and robust and is the predominant technique employed in systems vaccinology. In humans, vaccine-induced innate responses are most often measured indirectly by profiling readily accessible blood-derived cell populations (Querec 2009; Gaucher 2008; Nakaya 2011; Bosinger 2009; Palermo 2011).

  Innate response measurements made by profiling whole blood or blood cell subsets, although indirect, are nevertheless highly informative. This analysis probes multiple aspects of the response, all of which occur in parallel. These include the subset of cells that respond directly to the vaccine, cells responding to inflammatory mediators induced by the vaccine, and changes in the composition and activation states in circulating cells.

• Step II The next step in the approach is to measure vaccine-induced adaptive immune responses (immunogenicity). In contrast to innate responses, measurements of immunogenicity can be directly obtained from cells accessible in the blood or mucosa. These include antibody responses and antigen-specific CD4⁺ and CD8⁺ T cell responses (Hersperger 2011). Importantly, these measurements of adaptive immune function can be easily quantified at multiple time points to define the peak and memory responses as well as the impact of the initial vaccine ‘primes’ and subsequent ‘boosts’.

• Step III The third step of the approach is the measurement of vaccine efficacy. In some cases, such as malaria, efficacy can be measured directly in challenge studies. Alternatively for infections such as HIV where challenge studies are impossible, efficacy can be determined through measurements of vaccine-reduced acquisition rates, post-infection viral loads, transmission, or other aspects of the infection.

• Step IV The full compendium of measurements are then computationally integrated in a systems-level analysis to derive mechanistic insights and biomarkers. When direct measurements of vaccine-induced innate immune responses are available, it is possible to make computationally guided predictions about the causal regulatory networks controlling the vaccine-induced responses. We, and others, have employed these approaches to derive and validate novel regulatory networks controlling Toll-like receptor activated networks in innate immune cells (Amit et al. 2009; Zak 2011; Litvak et al. 2009; Ramsey et al. 2008; Gilchrist 2006). These approaches can be readily extended to predict regulatory networks controlling responses to vaccines, which are likely to activate several innate immune pathways in parallel (Querec 2006; Lindsay 2010; Delaloye 2009). When vaccine-induced innate immune responses (direct or indirect) and immunogenicity or efficacy measurements are available from the same animal or volunteer, it becomes possible to computationally identify predictive signatures of protection (Fig. 3). Currently, the most powerful application of systems vaccinology is the identification of these immunogenicity and efficacy signatures. In the best case, robust predictive signatures illuminate novel mechanistic insights; in the worst case these signatures serve as valuable biomarkers (Querec 2009; Nakaya 2011; Brooks 2008; Zou 2005).

  Fig. 3

  

  Network of gene expression signatures associated with CD4 ⁺ responses and SIV titers in macaques. The network represents innate immune signatures, measured days after primary vaccination, which predict enhanced SIV Gag-specific CD4⁺ T cell responses, and reduced SIV load after challenge, measured several months later. Each circle represents a gene expressed in PBMCs of macaques 6 days after vaccination. Blue edges represent gene pairs associated with enhanced CD4⁺ response; red edges represent gene pairs associated with immediate protection after challenge. (Adapted from Rappuoli and Aderem 2011)

• Step V While biomarkers achieve practical utility once they are validated in additional cohorts, the power of the mechanistic insights obtained in the first round of analysis is only realized when they are used to design and execute appropriate systems-level perturbations in an experimental model. The perturbations most directly related to molecular signatures are overexpression or knockdown (in vitro) or genetic ablation (murine in vivo) of the relevant genes. Although in vitro systems are the most easily perturbed, they are also the least appropriate for evaluating vaccine immunogenicity and efficacy. The murine genetic ablation validation strategy was recently applied in a study that identified immunogenicity signatures for the seasonal influenza vaccine in humans (Nakaya 2011). Small molecule agonists and inhibitors, specific for the networks implicated by the predictive signatures, can also be used as perturbations. Combinations of this class of drugs and vaccines have been explored in model systems (Araki 2009; Tan 2011) and may ultimately identify pharmacologic agents that can be combined with vaccines to improve efficacy.

4.2 Accelerating Efficacy Trials

During the 30 years since the discovery of HIV only four efficacy trials have been performed, an average of one trial every 8 years. Two of them have shown that anti-gp120 antibodies alone do not work; one has shown that T cells alone do not work; and one has shown that a prime-boost regime involving B and T cells may work. Altogether, only three hypotheses (elicitation of anti-gp120 antibodies, activation of T cells, and simultaneous elicitation of B and T cell responses) have been tested. Similarly, although challenge models are possible for malaria and numerous vaccines have been tested in phase I studies, all of these trials tested two hypotheses: peptide-based vaccines and RTS, S-based vaccines. To date, no efficacy trials have been performed for a new preventive vaccine against tuberculosis.

Accelerated clinical development can be achieved by improving the design of trials to test several hypotheses in parallel, incorporating systems biology to derive mechanistic insights and biomarkers, and employing a flexible strategy to expand the arms of the trial that are most promising (Freidlin and Simon 2005; Campbell 2009). For instance, several prime/boost strategies could be initiated concurrently in a large phase II study where subsets of the enrollees are monitored by systems biology approaches to test both safety and immune responses. This approach would identify vaccines that elicit qualitatively similar immune responses and permit rapid discrimination of different vaccine platforms and exploration of diverse approaches. Information collected during the early phases of the trial could be used to expand the most promising arms of the trial in order to gain sufficient statistical power to show the efficacy required for vaccine registration. Although this approach may require larger budgets during the initial phases, over the entire course of vaccine development, it will save money and time and will increase the probability of success. Several studies have demonstrated the ability to use early vaccine-response signatures to predict later immune responses (Querec 2009; Gaucher 2008; Nakaya 2011) and therefore, in principle, be used to make early decisions regarding the course of a clinical trial.