Alpha and beta type 1 interferon signaling: passage for diverse biologic outcomes

Summary

Type I interferon (IFN-I) elicits a complex cascade of events in response to microbial infection. Here, we review recent developments illuminating the large number of IFN-I species and describing their unique biologic functions.

Introduction

Type I interferon (IFN-I) is a sophisticated early defense mechanism positioned to combat microbial infections and assist in initiation of the adaptive immune response. By evolutionary selection, multiple IFN-I species evolved that signal through the same IFN-I receptor (IFNAR). The 14 type I alpha IFNs (IFN-α) in mice and 13 in humans, and one beta (IFN-β) have suggested that these molecules are retained because of unique biologic functions. Yet, the function of these various IFN-I species, their differing physical/chemical profiles, receptor affinities, and downstream gene activation profiles is largely unclear. Recently, data has emerged examining the role of various IFN-I species, illuminating both separate and overlapping roles for the members of the IFN-I family specific to the interacting host and infecting microbe.

IFN-I during microbial infection

IFN-I is associated with host-beneficial, anti-microbial activities, however, mounting evidence indicates the effects of this group of cytokines is more complex. As host-beneficial effects have been well-reviewed elsewhere, this review will concentrate on recent data describing the potentially detrimental effects of IFN-I to highlight the increasingly complex role that IFN-I plays and how its nuances are controlled through a single receptor. Recent studies in mice documented IFN-I mediated suppression of anti-microbial immune responses. Although these findings need to be borne out in human studies, observations of elevated IFN-I signatures in several chronic infections suggest that the role of IFN-I is at least equally complex in humans.

Early observations regarding the complexity of IFN-I were made in infections with intracellular bacteria. During infection with Listeria monocytogenes, mice lacking functional lymphocytes or with lymphocytes in which the IFNAR receptor is absent experience lower bacterial loads, highlighting the dependence of burden on IFN-I signaling directly on lymphocytes (Auerbuch et al., 2004; Carrero et al., 2006). IFN responses during infection with Mycobacterium leprae (leprosy) correlated with pathogenicity as shown by an inverse association between IFN-β and IFN-γ. IFN-β and its downstream genes, including interleukin-10 (IL-10), are expressed in progressive lepromatous lesions while IFN-γ and its downstream antimicrobial genes were expressed in self-healing tuberculoid lesions. Further, the IFN-γ-induced response, though able to mediate antimicrobial activity against M. leprae in vitro, was inhibited by IFN-β and IL-10, suggesting that differential production of IFNs contributes to protection versus pathogenesis and clinical outcome (Teles et al., 2013). Active infection with M. tuberculosis, causative agent of tuberculosis, is associated with high IFN-I signaling profile and excessive IFN-I signaling exacerbates tuberculosis (Mayer-Barber et al., 2014). The disadvantage to the host incurred by excessive IFN-I may be due to IFN-I mediated suppression of IFN-γ signaling which alters the cytokine profile to one that suppresses the immune response and promotes cellular conditions that support bacterial persistence.

Similar to infection with M. tuberculosis, profiling of persistent human viral infections with HIV, hepatitis B virus (HBV), and hepatitis C viruses (HCV), also indicate elevated IFN-signaling (Bolen et al., 2013; Doyle et al., 2015). The hallmarks of HIV, HBV, and HCV, and persistent viral infection in mice with lymphocytic choriomeningitis virus (LCMV) include generation of negative immune regulating (NIR) molecules that suppress antiviral CD4 and CD8 T cell responses resulting in decreased T cell function (T cell exhaustion). For several persistent viral infections, disordered secondary lymphoid structure and infection of key cell types (including dendritic cells and fibroblastic reticular cells) likely further interfere with activation of antiviral T cells via disruption of T cell migration and DC/T cell interactions, which is exacerbated by hyperactivation of T, B, and NK cells, and proinflammatory cytokines. Because IFN-I signaling is upstream of a number of inflammatory genes, IFN-signaling may play a deciding role in generating the hyperimmune environment accompanying viral infections and several autoimmune diseases.

Definitive data for the role of individual IFN-I species in generating harmful or beneficial environments was revealed during examination of LCMV infection. LCMV Armstrong clone 53b (ARM) and Cl-13 differ by three amino acids yet have profoundly different biologies resulting in an acute infection that is cleared within 8–12 days or a persistent infection characterized by CD4 and CD8 T cell hyporesponsiveness/exhaustion (Ng et al., 2011; Sullivan et al., 2011). During Cl-13 infection, mice significantly elevate IFN-α and IFN-β at 16–24 hours post-infection compared to ARM-infected mice (Figure 1B) (Teijaro et al., 2013; Wilson et al., 2013). Early clearance was associated with enhancement of antiviral CD8 T cells and upon marked enhancement of antiviral CD4 T cells, agreeing with previous observations that IFN-I suppressed CD4 T cells (Osokine et al., 2014). The improvements were dependent upon two factors: (1) IFNAR blockade suppressed IL-10 and PD-L1 levels prior to onset of T cell exhaustion, supporting the hypothesis that high levels of IFN-I signaling promote T cell exhaustion; (2) Blockade of IFNAR signaling protected secondary lymphoid structure and T cell migration which becomes severely distorted during infection (Teijaro et al., 2013; Wilson et al., 2013), and, thus, interfering with T cell migration and DC/T cell interactions.

Figure 1.

Biochemical, functional and sequence comparison of IFN-β and IFN-α2. (A) Structure of the heterodimeric IFN-I receptor. IFNAR-1 (green) with the SD4 modeled from pdb 3WCY and IFNAR-2 (magenta) (pdb 3SE3). (B) IFN-I species including IFN-α2 (blue) and IFN-β bind in the same groove using a conserved geometry to elicit ligand-specific AV and AP activity profiles. (Upper panel) Differences in serum IFN-α and IFN-β production following infection with LCMV Cl-13 or ARM. (Middle panel) IFN-β has a 20- to 30-fold higher affinity for either IFNAR-1 or IFNAR-2 compared to IFN-α2. (Bottom panel) The high affinity properties of IFN-β result in a 40- to 60-fold improvement in AP (cell proliferation assay) potency relative to IFN-α2. IFN-β exhibits a modest 2-fold enhancement in AV (inhibition of vesicular stomatitis virus cytopathogenicity) potency over IFN-α2, reflecting the "maxed out" AV response of the IFN-I system. (C–G) Sequence conservation of IFN-β and IFN-α2 interaction networks with IFNAR. (C) IFN-β and IFN-α2 have a low sequence identity of 32%. Coloring the structure of IFN-α2 (YNS) from pdb 3SE3 reveals sequence conservation between IFN-β and IFN-α2 in the core and outer facial residues important for receptor binding (Thomas et al., 2011). (D) The low affinity IFNAR-1 binding site. Conserved amino acids belonging to different receptor interaction-networks are shown as sticks. (E) IFN-α2/IFNAR-1 two-dimensional interaction map. Residues are depicted as nodes in the interaction map. IFNAR-1 residues are shown as green rectangles with circles indicating the hotspot residues. IFN-α2 residues are shown as ovals and are colored by sequence conservation. Side chain interactions are indicated by lines, interactions between side chains and backbones are depicted as arrows directed towards the backbone. Hydrophobic and van der Waals interactions are shown as solid lines, aromatic interactions as dotted lines, and electrostatic or hydrogen-bond interactions as dashed lines. (F) Close-up view of the high affinity IFNAR-2 binding site. (G) IFN-α2/IFNAR-2 two-dimensional interaction map as described in (E) except IFNAR-2 residues are colored in magenta.

Differing roles for IFN-I species

The presence of a large number of IFN-I subspecies that signal through a common receptor make it likely that they were evolutionarily retained for selective functions. While a degree of redundancy is advantageous, the conserved nature of these subspecies suggests necessity beyond redundancy. Indeed, of the IFN-I subspecies, only IFN-β is known to induce NIRs IL-10 and PD-L1 (Ng et al., 2012; Saraiva and O'Garra, 2010; Sharpe et al., 2007) which suppress T cell responses contributing to microbial clearance. IFN-I induction during microbial infection depends upon intracellular and extracellular innate sensing pathways that recognize pathogen-associated molecular patterns. Detection of microbial components through these pathways initiates the production of IFN- I and is essential to determining whether a microbial infection will be cleared acutely or establish long-term infection. In the case of LCMV, the virus triggers the cytoplasmic RNA sensing pathways in hematopoietic cells via RIG-I and MDA-5 triggering IRF3 to induce IFN-I signaling (Sullivan et al., 2015). Although both IFN-α and IFN-β are elicited by Cl-13 infection, early clearance of LCMV persistent infection initiated by blocking IFNAR is primarily due to loss of IFN-β signaling. Blockade of IFN-β results in early viral clearance similar to that observed by blocking the IFNAR despite the presence of high levels of IFN-α; blockade of at least six forms of IFN-α (A, 1, 4, 5, 11, and 13) did not accelerate viral clearance (Ng et al., 2015). However, IFN-α was key to controlling viral spread, as only IFN-α blockade altered early viral dissemination (Ng et al., 2015), indicating that the roles played by IFN-β as compared to IFN-α in controlling viral infection are different.

The observed dichotomy between IFN-α and IFN-β biologic activity observed in animal models is likely reflected in human disorders. For example, the finding in the LCMV model that blockade of IFN-α, but not IFN-β, was effective for control of early viral dissemination provides a possible explanation for the partial effectiveness of IFN-α but not IFN-β therapy in lowering HCV RNA. Correspondingly, the effectiveness of IFN-β therapy in treatment of multiple sclerosis likely represents IFN-β immunosuppressive activity that reduces the T cell response responsible for severity and progression of this autoimmune disease. The separate roles of IFN-α and IFN-β are also supported by studies with other infection models and indicate the importance of host genes and timing of IFN-I signaling. For example, LCMV Cl-13 inoculation into certain strains of immunocompetent mice (C57Bl/6, Balb/CDJ, C3H, or SWR/J) lead to the persistent infection phenotype. In contrast, the same infection in NZB, SJL, or FVB/N mice induces death by 7 days post-challenge (Baccala et al., 2014; Schnell et al., 2012). The two distinct phenotypes are mediated by IFN-I but, in contrast, to the persistent infection, acute death is primarily due to IFN-α and not IFN-β (Baccala et al., 2014).

Factors such as timing, signaling magnitude, and source of the individual IFN-I subspecies play an important role in coordinating the outcome of IFN-I signaling. Control of SIV is associated with downregulation of IFN-I responses while persistently elevated IFN-I responses are associated with pathogenic SIV infection (Harris et al., 2010). However, comparison of blockade vs. administration of IFN-I resulted in similar outcomes of increased SIV reservoir size and accelerated CD4 T-cell depletion and clinical progression. Early blockade of IFNAR in SIV-infected rhesus macaques decreased T-cell activation but also antiviral gene expression, while administration of IFN-α2a, although initially slowing viral dissemination, eventually induced IFN-I desensitization and decreased antiviral gene expression (Sandler et al., 2014). Consequently, timing of IFN-induced innate responses likely affects overall disease course. Certainly, the timing of early events is key as early innate signals and IFN-I induction are necessary in determining disease outcome (Kazar et al., 1971; Sullivan et al., 2015). The cellular source(s) of IFN-I define timing and localization of IFN-I expression. During MCMV infection, pDC-derived IFN-I is key to control of viral spread (Krug et al., 2004; Schneider et al., 2008) while early release by stromal cells likely coordinates control of the initial inoculum and initiates further IFN-I production, demonstrating importance of sequence of events in defining the functional role of IFN-I. Timing may also involve activity of different IFN-I subtypes. Each IFN-I subtype may have overlapping but unique windows during which the subtype’s specific signaling best generate a beneficial effect. Signaling magnitude is also a factor as antibody blockade of IFNAR-1 produces accelerated viral clearance (Teijaro et al., 2013; Wilson et al., 2013) while deletion or mutation of the gene results in the inability to clear the virus in mice and increased susceptibility to infection in humans. Such complexities indicate that there is an ideal range of IFN-I signaling below which signaling is not sufficient to generate immunity and above which the overwhelming signal is detrimental to host immunity.

The mechanism behind the differential signaling profiles of IFN-I species involves complex structure-activity relationships where differences in the relative binding affinities of the different species to each IFNAR receptor are translated into distinct functional outcomes (Piehler et al., 2012). The IFNAR receptor is heterodimeric, consisting of two subunits, IFNAR-1 and IFNAR-2 (Figure 1). IFNAR-2 is the primary binding protein and binds IFN-I molecules with high affinity; IFNAR-1 subsequently binds the molecule with low affinity. The binding event triggers phosphorylation of the cytoplasmic portion of IFNAR-1 and consequent signaling through the JAK/STAT pathway via STAT1 and STAT2 (Schreiber and Piehler, 2015). The ability to sensitively convey the extracellular binding parameters into distinct functional outcomes is accomplished through a combination of effects on receptor internalization, and negative and positive intracellular regulators and feedback loops (Piehler et al., 2012; Schreiber and Piehler, 2015).

The root of the functional discrimination of IFN-I ligand types is through their interaction chemistries with IFNAR-1 and IFNAR-2. All of the IFNs contain a set of common ‘hotspot’ residues for receptor binding that serve as “anchor point” and ensure that all IFN bind to the IFNAR (Figure 1D–G) (Piehler et al., 2012). Interspersed within the anchor points are specificity determining residues that differ between the IFNs and serve to modulate affinity for the IFNAR (Figure 1D–G), thereby imparting functional specificity. IFN-I major subtypes α and β display different signaling activities in vitro, likely due to their unique binding affinities for each of the receptors and divergence in their respective complex lifetimes (Jaks et al., 2007; Lavoie et al., 2011; Piehler et al., 2012; Thomas et al., 2011). IFN-β has the highest binding affinity among the IFN-I family (Figure 1A,B) (Jaks et al., 2007; Lavoie et al., 2011). The binding interfaces of IFN-β are likely to be very similar to that of other IFN-Is in that the binding hotspots are conserved, which will manifest as identical receptor docking topologies. However, peripheral to the hotspots, it is clear that IFN-β possesses amino acids that impart a superior binding chemistry to IFNAR and resulting in higher affinity.

Conclusion

IFN-I signaling, determined by the physical-chemical properties of interactions between IFNAR and various IFN-I species, serves as a gateway to activation of selected genes generating different biologic phenotypes. The dichotomy uncovered between α and β IFN responses during various pathogenic infections suggest there are unique molecular and signaling pathways controlled by IFN-I that await discovery. Clearly, further careful mapping of α and β activity and receptor binding coupled with development of decoys and therapeutic molecules is likely to open new avenues for successfully understanding pathway usage and mechanism by uncovering specific activation signals, roles and activities for various IFN-I subspecies. Correlation of screening studies that illuminate the activities of interferon stimulated genes on pathogens (Liu et al., 2012; Schoggins et al., 2014; Schoggins et al., 2011) with study of the types and timing of the IFN-I species and the genes they activate during infection will lead to a better understanding of the IFN-I mediated chess game between the pathogen and host and generate strategies to manipulate the IFN-I pathway for biomedical benefit.