Journal of Leukocyte Biology eBioscience full spectrum cell analysis
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published online as doi:10.1189/jlb.0603252 on October 2, 2003

Published online before print October 2, 2003
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
jlb.0603252v1
75/2/163    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Schroder, K.
Right arrow Articles by Hume, D. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Schroder, K.
Right arrow Articles by Hume, D. A.
(Journal of Leukocyte Biology. 2004;75:163-189.)
© 2004 by Society for Leukocyte Biology

Interferon-{gamma}: an overview of signals, mechanisms and functions

Kate Schroder*,{dagger}, Paul J. Hertzog{dagger},{ddagger}, Timothy Ravasi*,{dagger} and David A. Hume*,{dagger},1

* Institute for Molecular Bioscience, University of Queensland, St. Lucia, Brisbane, Australia;
{dagger} CRC for Chronic Inflammatory Diseases, Parkville, Victoria, Australia; and
{ddagger} Molecular Genetics Group, Institute of Reproduction and Development, Monash Medical Centre, Monash University, Clayton, Victoria, Australia

1 Correspondence: Institute for Molecular Bioscience, University of Queensland, St. Lucia, Brisbane 4072, Australia. E-mail: D.Hume{at}imb.uq.edu.au


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
Interferon-{gamma} (IFN-{gamma}) coordinates a diverse array of cellular programs through transcriptional regulation of immunologically relevant genes. This article reviews the current understanding of IFN-{gamma} ligand, receptor, signal transduction, and cellular effects with a focus on macrophage responses and to a lesser extent, responses from other cell types that influence macrophage function during infection. The current model for IFN-{gamma} signal transduction is discussed, as well as signal regulation and factors conferring signal specificity. Cellular effects of IFN-{gamma} are described, including up-regulation of pathogen recognition, antigen processing and presentation, the antiviral state, inhibition of cellular proliferation and effects on apoptosis, activation of microbicidal effector functions, immunomodulation, and leukocyte trafficking. In addition, integration of signaling and response with other cytokines and pathogen-associated molecular patterns, such as tumor necrosis factor-{alpha}, interleukin-4, type I IFNs, and lipopolysaccharide are discussed.

Key Words: macrophage • cytokine • lipopolysaccharide • Toll-like receptor • inflammation


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
This review focuses on the interplay between interferon-{gamma} (IFN-{gamma}) and macrophages in inflammation and acquired immunity during infection. Macrophages are extremely versatile cells involved in a number of complex functions in disease and health. A pathogen encounters macrophages soon after host entry, and the result of these encounters is fundamental to the host’s ability to mount an effective immune response. Macrophage "activation", broadly defined as "acquisition of competence to execute a complex function" [1 ], is among the first actions to occur in innate immunity to potential pathogens. In most situations, macrophages are activated to acquire microbicidal effector functions and secrete proinflammatory cytokines, resulting in inflammation and recruitment of immune cells and subsequent elimination of the microbe by phagocytosis or release of toxic metabolites.

Macrophages respond to a range of different cell products during the innate and acquired immune response. Of these, IFN-{gamma} (originally called macrophage-activating factor) is among the most important. Macrophage stimulation with IFN-{gamma} induces direct antimicrobial and antitumor mechanisms as well as up-regulating antigen processing and presentation pathways. IFN-{gamma} orchestrates leukocyte attraction and directs growth, maturation, and differentiation of many cell types [2 3 4 ], in addition to enhancing natural killer (NK) cell activity [5 ] and regulating B cell functions such as immunoglobulin (Ig) production and class switching [4 , 6 ]. This review provides a brief overview of IFN-{gamma} biology with respect to the well-characterized responses that alter macrophage function during infectious challenge.


   THE IFNs
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
The IFNs were originally discovered as agents that interfere with viral replication [7 ]. Initially, they were classified by the secreting cell type but are now classified into type I and type II according to receptor specificity and sequence homology. The type I IFNs are comprised of multiple IFN-{alpha} subtypes (14–20, depending on species), IFN-ß, IFN-{omega}, and IFN-{tau}, all of which are structurally related and bind to a common heterodimeric receptor (IFNAR, comprised of IFNAR1 and IFNAR2 chains). Although type I IFNs are secreted at low levels by almost all cell types, hematopoietic cells are the major producers of IFN-{alpha} and IFN-{omega}, whereas fibroblasts are a major cellular source of IFN-ß [8 ]. IFN-ß is also produced by macrophages under appropriate stimulus (discussed later). Viral infection is the classic stimulus for IFN-{alpha} and IFN-ß expression [8 , 9 ]. Secretion of IFN-{tau} has only been reported in ruminants [10 ].

IFN-{gamma} is the sole type II IFN. It is structurally unrelated to type I IFNs, binds to a different receptor, and is encoded by a separate chromosomal locus. Initially, it was believed that CD4+ T helper cell type 1 (Th1) lymphocytes, CD8+ cytotoxic lymphocytes, and NK cells exclusively produced IFN-{gamma} [8 , 11 ]. However, there is now evidence that other cells, such as B cells, NKT cells, and professional antigen-presenting cells (APCs) secrete IFN-{gamma} (reviewed in refs. [5 , 12 13 14 15 16 ]). IFN-{gamma} production by professional APCs [monocyte/macrophage, dendritic cells (DCs)] acting locally may be important in cell self-activation and activation of nearby cells [12 , 13 ]. IFN-{gamma} secretion by NK cells and possibly professional APCs is likely to be important in early host defense against infection, whereas T lymphocytes become the major source of IFN-{gamma} in the adaptive immune response [12 , 17 ].

IFN-{gamma} production is controlled by cytokines secreted by APCs, most notably interleukin (IL)-12 and IL-18. These cytokines serve as a bridge to link infection with IFN-{gamma} production in the innate immune response [18 19 20 21 22 23 24 ]. Macrophage recognition of many pathogens induces secretion of IL-12 and chemokines [e.g., macrophage-inflammatory protein-1{alpha} (MIP-1{alpha}); ref. 25 ]. These chemokines attract NK cells to the site of inflammation, and IL-12 promotes IFN-{gamma} synthesis in these cells [25 , 26 ]. In macrophages, NK and T cells, the combination of IL-12 and IL-18 stimulation further increases IFN-{gamma} production [20 , 23 , 24 , 27 , 28 ].

Negative regulators of IFN-{gamma} production include IL-4, IL-10, transforming growth factor-ß, and glucocorticoids [17 , 21 , 27 28 29 ]. Given the complexity of IFN-{gamma} regulation, it is not surprising that inbred mouse strains vary in their ability to secrete this cytokine; for example, T lymphocytes of C57BL/6 and C3H mice secrete significantly higher amounts of IFN-{gamma} compared with the T lymphocytes of BALB/c and B10.D2 mice. Increased IFN-{gamma} production in these strains is associated with greater resistance to bacteria and viruses [30 31 32 ]. Many excellent reviews on the regulation of IFN-{gamma} production have been published recently and the reader is referred to these for further information [11 12 13 ].


   THE IFN-{gamma} RECEPTOR
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
Functional IFN-{gamma} receptor (IFNGR) is comprised of two ligand-binding IFNGR1 chains associated with two signal-transducing IFNGR2 chains and associated signaling machinery. IFNGR1 and IFNGR2 chains belong to the class II cytokine receptor family, a class of receptors that bind ligand in the small angle of a V formed by the two Ig-like folds that constitute the extracellular domain. It is likely that the IFNGR chains and other family members (type I IFN receptor and tissue factor) evolved from primitive adhesive molecules [33 , 34 ].

The IFNGR2 chain is generally the limiting factor in IFN-{gamma} responsiveness, as the IFNGR1 chain is usually in surplus [8 , 35 ]. The IFNGR2 chain is constitutively expressed, but its expression level may be tightly regulated according to the state of cellular differentiation or activation [8 ]. For example, some CD4+ Th1 populations have very low levels of cell-surface expression of the IFNGR2 chain and thus low expression of active IFNGR, leading to a functional blockade of some aspects of IFN-{gamma} signaling [36 , 37 ]. As the growth-inhibitory effects of IFN-{gamma} are blocked, T cells expressing low levels of the IFNGR2 chain continue to proliferate during IFN-{gamma} treatment. Conversely, IFN-{gamma} exposure to CD4+ Th2 populations displaying high levels of IFNGR2 inhibits proliferation and may induce the apoptotic program [35 , 38 , 39 ]. This mechanism may aid in the Th2-to-Th1 phenotype switch apparent when CD4+ cells are treated with IFN-{gamma}: the Th2 population decreases as a result of growth-inhibitory and proapoptotic effects of IFN-{gamma}, and the Th1 population continues to proliferate as a result of blockade of IFN-{gamma} function. As a consequence, IFN-{gamma} signaling causes IFNGR2 down-regulation and switching to a Th1 phenotype as a T cell population but not at a single-cell level [37 ].

Both IFNGR chains lack intrinsic kinase/phosphatase activity and so must associate with signaling machinery for signal transduction. The IFNGR1 intracellular domain contains binding motifs for the Janus tyrosine kinase (Jak)1 and the latent cytosolic factor, signal transducer and activator of transcription (Stat)1. In humans, the Jak1-binding motif LPKS is a membrane-proximal sequence located at residues 266–269 [40 41 42 ]. The human Stat1-binding site YDKPH is positioned at residues 440–444 [43 ]. This motif contains an essential Y440 phosphorylation site that is phosphorylated during signal transduction to allow Stat1 recruitment to the receptor [42 43 44 ]. Residues 441DKPH444 are responsible for the binding specificity of IFNGR1 and Stat1 [41 , 43 , 44 ]. The Jak1- and Stat1-binding motifs are required for receptor phosphorylation, signal transduction, and induction of biological response. Receptor-ligand internalization is mediated by an isoleucine-leucine sequence at residues 270 and 271 of the mature IFNGR1 chain in humans. Mutant receptors with deletion or substitution of this sequence cannot internalize ligand, although signal transduction is not affected [42 , 45 ].

The intracellular region of IFNGR2 contains a noncontiguous binding motif for recruitment of Jak2 kinase for participation in signal transduction. In humans, these sites are 263PPSIP267 and 270IEEYL274 [46 , 47 ]. The IFNGR2 chain is not tyrosine phosphorylated during signal transduction [46 ]. Cross-linking experiments with labeled human IFN-{gamma} demonstrated that IFN-{gamma} only associates with IFNGR2 when the IFNGR1 chain is present, indicating that the primary interaction between the IFNGR2 chain and the IFN-{gamma}:IFNGR1 complex is found between IFNGR1 and IFNGR2, although it is likely that IFNGR2 also interacts weakly with the ligand [46 , 48 ].

IFN-{gamma}:IFNGR1 and IFNGR1:IFNGR2 interactions are species-specific. A number of studies with chimeric receptors in which the intracellular, transmembrane, or extracellular domains were swapped between the human and murine IFNGR chains showed that species specificity in interactions of the ligand:receptor complex is restricted to the receptor extracellular domains [49 50 51 52 53 ].


   IFN-{gamma} SYSTEM DYSFUNCTION
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
IFN-{gamma}-/- and IFNGR1-/- mice showed no overt developmental defects, and their immune system appeared to develop normally [54 ]. However, these mice show deficiencies in natural resistance to bacterial, parasitic, and viral infections such as vaccinia virus, Theiler’s murine encephalomyelitis virus, Leishmania major, Toxoplasma gondii, Listeria monocytogenes, and several poorly virulent mycobacteria species [54 55 56 57 58 59 ]. Some viruses (e.g., vaccinia virus, Theiler’s murine encephalomyelitis virus, and lymphocytic choriomeningitis virus) appear to require both type I and II IFN pathways, whereas viruses such as Semliki forest virus and vesicular stomatitis virus require a predominantly type I IFN response for efficient clearing [60 ]. These observations have prompted the suggestion that the two IFN systems may have evolved to complement each other in overlapping but nonredundant activities to defend against a broad spectrum of pathogens [60 , 61 ]. Historically, the type I IFNs were considered to be primarily antiviral agents with limited immunomodulatory activity, whereas IFN-{gamma} was considered to be primarily an immunomodulator with limited antiviral activity [8 ]. It is now apparent that both types of IFN possess antiviral and immunomodulatory activities to some degree [62 ]. The antiviral activity of both types of IFNs is often important in the early stages of viral infection, but their immunomodulatory activities become important in later stages of infection when the adaptive immune response becomes critical [60 ].

Patients with inactivating mutations of the human IFNGR1 or IFNGR2 chains show clinical presentation similar to the mouse models. Human loss-of-function mutations in the IFNGR1 or IFNGR2 chain are closely associated with severe susceptibility to poorly virulent mycobacteria, often presenting as early onset bacillus Calmette-Guerin infection and fatality in childhood [63 64 65 66 67 68 69 70 71 ]. Partial deficiency of either chain gives a milder phenotype and a much better prognosis [63 64 65 ]. Loss of functional IFNGR1 appears to be associated only with increased susceptibility to infection with some viruses and intracellular bacteria; increased susceptibility to other common bacterial and fungi pathogens have not been reported [72 ].

In addition to recurrent infection, infants with deficient production of IFN-{gamma} exhibited decreased neutrophil mobility and NK cell activity, highlighting the importance of IFN-{gamma} in the inflammatory response and immunoregulation [73 ]. It is interesting that natural IFN-{gamma} polymorphisms have been correlated with increased longevity [74 ]. It has been proposed that a slightly dampened inflammatory status caused by an IFN-{gamma} polymorphism, while not enough to significantly impact on the individual’s ability to clear infection, may prevent or defer inflammation-related diseases such as cardiovascular disease, neurodegeneration, osteoarthritis, osteoporosis, and diabetes [74 ].

Aside from functions in host defense, IFN-{gamma} may also contribute to autoimmune pathology. Although IFN-{gamma} production was shown to be disease-limiting in autoimmune models such as murine experimental allergic encephalomyelitis (EAE) [75 , 76 ], it may contribute to autoimmune nephritis [77 , 78 ]. In humans, IFN-{gamma} is implicated in pathology of diseases such as systemic lupus erythematosus [79 , 80 ], multiple sclerosis [81 ], and insulin-dependent diabetes mellitus [82 83 84 ].

IFN-{gamma} function is significant in tumor surveillance. IFNGR1 knockout (KO) mice, cells with dominant-negative IFNGR1 mutations, and cells treated with IFN-{gamma}-neutralizing antibodies display compromised tumor rejection [85 86 87 ]. IFN-{gamma} protects against tumor development and directs the immunogenic phenotype of tumors that arise in an immunocompetent host (the "cancer immunoediting" concept; reviewed in refs. [88 , 89 ]).


   SIGNAL TRANSDUCTION
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
IFN-{gamma} primarily signals through the Jak-Stat pathway, a pathway used by over 50 cytokines, growth factors, and hormones to affect gene regulation [90 ]. Jak-Stat signaling involves sequential receptor recruitment and activation of members of the Janus family of kinases (Jaks: Jaks 1–3 and Tyk2) and the Stats (Stats 1–6, including Stat5a and Stat5b) to control transcription of target genes via specific response elements. As this signaling mechanism is a recurring theme amongst members of the cytokine receptor superfamily, IFN-{gamma}-induced Jak-Stat signaling has emerged as the global paradigm for class II cytokine receptor signal transduction.

The IFNGR1 and IFNGR2 subunits of the IFN-{gamma} receptor were thought not to strongly associate with each other in the absence of ligand [13 , 40 , 47 , 91 , 92 ], but new techniques allowing the study of receptor chain interactions in intact cells have shown the receptor complex is assembled before ligand binding [93 ]. Upon ligand binding, the intracellular domains of the receptor chains open out to allow association of downstream signaling components. Biologically active IFN-{gamma} is a noncovalent homodimer formed by the self-association of two mature polypeptides in an antiparallel orientation [94 , 95 ] and thus binds IFNGR1 in 2:2 binding stoichiometry [47 , 95 96 97 ]. Ligand binding induces Jak2 autophosphorylation and activation, which allows Jak1 transphosphorylation by Jak2 (Fig. 1 ) [98 ]. The activated Jak1 phosphorylates functionally critical tyrosines on residue 440 of each IFNGR1 chain to form two adjacent docking sites for the SH2 domains of latent Stat1 [41 , 99 100 101 ]. The receptor-recruited Stat1 pair is phosphorylated near the C terminus at Y701, probably by Jak2 [98 ]. Phosphorylation induces dissociation of a Stat1 homodimer (also known as {gamma}-IFN activation factor) from the receptor [100 ]. The four critical tyrosines (contained by Jak1, Jak2, IFNGR1, and Stat1) are phosphorylated within 1 min of IFN-{gamma} treatment [41 , 99 ].



View larger version (22K):
[in this window]
[in a new window]
  		Figure 1. The current paradigm for IFN-{gamma} signal transduction. Ligand binding causes a conformational change in the IFN-{gamma}R (IFNGR1, yellow; IFNGR2, green), such that the inactive Jak2 kinase undergoes autophosphorylation and activation, which in turn allows Jak1 transphosphorylation by Jak2. The activated Jak1 phosphorylates functionally critical tyrosines on residue 440 of each IFNGR1 chain to form two adjacent docking sites for the Src homology (SH)2 domains of latent Stat1. The receptor-recruited Stat1 pair is phosphorylated near the C terminus at Y701. Phosphorylation induces dissociation of a Stat1 homodimer from the receptor. To a lesser extent, IFN-{gamma} signaling also produces Stat1:Stat1:IFN regulatory factor (IRF)-9 and Stat1:Stat2:IRF-9 [IFN-stimulated gene factor 3 (ISGF3)] complexes. Stat1 homodimers travel to the nucleus and bind to promoter IFN-{gamma}-activation site (GAS) elements to initiate/suppress transcription of IFN-{gamma}-regulated genes. Many of IFN-{gamma}-regulated genes are in fact transcription factors (e.g., IRF-1), which are activated by IFN-{gamma} and are able to drive regulation of the next wave of transcription (e.g., induction of IFN-ß). Stat1:Stat1:IRF-9 heterodimers, ISGF3, and IRF-1 are able to bind to IFN-stimulated response element (ISRE) promoter regions in target genes to regulate transcription. IRF-1 is also able to promote transcription of Stat1 through an unusual ISRE site (IRF-E/GAS/IRF-E). Signaling molecules activated by IFN-{gamma} are depicted in red text. ICAM-1, Intercellular adhesion molecule-1; MIG, monokine induced by IFN-{gamma}; iNOS, inducible nitric oxide synthase.

 
The dissociated Stat1 homodimer enters the nucleus and binds to promoter elements to initiate or suppress transcription of IFN-{gamma}-regulated genes (reviewed in refs. [102 103 104 ]). Stat1 homodimers bind DNA at GAS elements of consensus sequence TTCN(2-4)GAA [105 ]. The key function of Stat1 in mediating IFN-{gamma} signal transduction is indicated by the phenotype of Stat1-/- mice, which largely phenocopy IFNGR1-/- mice during IFN-{gamma} stimulation [106 ].

The first wave of IFN-{gamma}-induced transcription occurs within 15–30 min of IFN-{gamma} treatment [107 ]. Many of the induced genes are in fact transcription factors (for example, IRF-1), which are activated by IFN-{gamma} and are able to further drive regulation of the next wave of transcription. Transcription of many IFN-{gamma}-responsive genes is controlled by a GAS element or an ISRE [108 ].


   Stat1
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
The crystal structure for tyrosine phosphorylated Stat1 bound to DNA has been determined [109 ]. The structure verified that the SH2 domain necessary for dimerization is also involved in DNA binding. The C-terminal transactivation domain contains the Y701 phosphorylation site and also a S727 phosphorylation site, both of which are functionally important for efficient signaling [110 111 112 113 ]. All Stat proteins possess a SH2 domain that when tyrosine phosphorylated, mediates homodimerization or heterodimerization with other Stat molecules [114 , 115 ]. Although the simplistic model described above suggests that IFN-{gamma} signaling produces active Stat1 homodimers alone, other active complexes such as Stat1 heterodimers (e.g., Stat1:Stat2) and heterotrimers (e.g., Stat1:Stat1:IRF-9, Stat1:Stat2:IRF-9) form during signaling [102 , 116 117 118 ]. Stat2 is the only Stat that does not contain a DNA-binding domain [119 ]. The Stat1:Stat2:IRF-9 complex (known as ISGF3) was thought to be the typical type I IFN transcription factor. It is now evident that type I IFN signals primarily through ISGF3 but also uses Stat1 homodimers and conversely, type II IFN signals primarily through Stat1 homodimers but also uses ISGF3 to activate transcription [116 ]. This may explain many of the overlapping effects of the two types of IFNs.

Phosphorylation of Stat1 at S727 is essential for maximal ability to activate transcription of target genes [111 112 113 , 120 121 122 ]. A number of different stimuli induce Stat1 serine phosphorylation, including types I and II IFN, lipopolysaccharide (LPS), IL-2, IL-12, tumor necrosis factor {alpha} (TNF-{alpha}), and platelet-derived growth factor [110 111 112 , 123 , 124 ]. This may be a mechanism whereby S727 serves as an avenue for modulation of IFN-{gamma} signaling by independent extracellular cues. For example, LPS signaling increases Stat1 S727 phosphorylation independently of Y701 phosphorylation in macrophages, thereby augmenting cellular responses to IFN-{gamma} [110 ].

The ability of Stat1 to activate or repress gene transcription depends on the presence of other transcription factors binding to the promoter element and Stat1 interaction with other factors. Stat1 activation is necessary but not sufficient for transcription of a number of genes [125 , 126 ]. Likewise, factors such as oncostatin M, which participate in Jak-Stat signaling to produce activated Stat1, do not induce transcription of IFN-{gamma}-inducible genes [127 ]. Stat1 interaction with transcription factors such as IRF-9, upstream stimulatory factor-1, specificity protein-1, and heat shock factor-1 have been reported (reviewed in ref. [119 ]) and are likely to influence specificity of DNA binding and transactivator ability. Phosphorylation of S727 is necessary for interaction with some proteins, such as BRCA1 and MCM-5, and this interaction potentiates Stat1-mediated transcription [128 , 129 ]. In some promoters, transcription is maximal when more than one GAS-binding transcriptional activator binds to tandem GAS sites in the target gene promoter [130 , 131 ].

In the absence of IFN-induced tyrosine phosphorylation, Stat1 drives constitutive transcription of genes such as caspases [132 , 133 ]. The Stat1 KO mouse showed increased lymphocyte survival and proliferation, which was attributed to insufficient caspase levels to carry out the apoptotic program [133 ]. Stat1 also constitutively regulates expression of the low molecular protein 2 (LMP2) gene [134 ]. A complex of unphosphorylated Stat1 and IRF-1 may mediate this.

The mechanism of Stat1 entry into the nucleus is still controversial, but the involvement of importin-{alpha}-1 (NPI-1) is implied [135 ]. Two major models have been suggested, one being the classic nuclear importin mechanism and the other involving a requirement for intracellular IFN-{gamma}. Nuclear entry of Stat1 is apparent at 15 min and almost complete after 30 min exposure to IFN-{gamma} [136 ]. Stat1 nuclear translocation does not appear to be a result of liberation from a cytoplasmic anchor or transport by the cellular cytoskeleton or microtubules [136 ].

Experiments with green fluorescent protein-tagged Stat1 implied that it travels through the cytoplasm to the nucleus by random walk movement [136 ]. A model was proposed whereby Stat1 dimerization allows interaction with importin NPI-1, which then mediates translocation through the nuclear pore [135 ]. The nuclear membrane forms an efficient barrier to inactivated Stat1, and entry requires a gain of nuclear localization sequence (NLS) function [136 137 138 ]. To date, no "classic" NLS (single stretch of basic residues or bipartite motif; ref. [139 ]) has been identified for Stat1, indicating an as-yet unidentified, novel NLS is unmasked during Stat1 homodimerization, or the NLS of a Stat1 homodimer-binding partner mediates nuclear translocation.

The second model for Stat1 nuclear entry involves a requirement for intracellular IFN-{gamma}. IFN{gamma} enters the cell using a mechanism that is unclear at present. Nuclear accumulation of IFN-{gamma} is apparent shortly after cellular exposure to IFN-{gamma} [140 , 141 ]. Nuclear translocation is mediated by a C-terminal sequence very similar to the Simian virus-40 T antigen NLS, 126RKRKRSR132 in mice and 128KTGKRKR134 in humans [139 , 142 143 144 145 ]. This sequence is a functional NLS and is essential for the full biological activity of IFN-{gamma} [144 145 146 147 148 ]. A deletion mutant of the putative NLS in human IFN-{gamma} is still able to bind with high affinity to the IFNGR1 chain, but induction of biological response is absent [149 ]. The addition of a heterologous NLS is able to rescue biological responsiveness in this system. Another study tracked the cellular localization of components of the IFNGR complex following ligand exposure [150 ]. IFN-{gamma}, Stat1, and IFNGR1 were found to be quickly internalized, colocalized, and accumulated in the nucleus, while the majority of the IFNGR2 subunit remained at the cell surface throughout signaling. Stat1 nuclear translocation is mediated by NPI-1 [135 ], and immunoprecipitation of Stat1 following IFN-{gamma} treatment showed that IFN-{gamma}, IFNGR1, Stat1, and NPI-1 form a complex at the nuclear pore during translocation [142 , 143 ]. Subramaniam et al. [90 ] proposed a model suggesting a function for intracellular ligand in Stat1 nuclear localization. In this model, extracellular ligand binds to the IFNGR1 chain, causing endocytosis of the complex and recycling of the IFNGR2 chain to the cell surface. The intracellular domain of IFNGR1 associates with activated Stat1 and intracellular IFN-{gamma}, and the IFN{gamma}:IFNGR1:Stat1 complex is actively transported through the nuclear pore via IFN-{gamma}–NLS interaction with the NPI-1 importin. Activated Stat1 may then bind to promoter elements to activate transcription.

Functional Stat1 is crucial to host response to infection. KO of Stat1 renders mice extremely sensitive to infection by viral and microbial pathogens (e.g., vesicular stomatitis viruses, mouse herpes virus, L. monocytogenes) [106 , 151 ]. This susceptibility has been attributed to defective immune system function as a result of the combined loss of types I and II IFN responses [151 ]. NO production in Stat1-/- macrophages primed with types I or II IFN and treated with LPS is greatly impaired compared with the wild type (WT), resulting in reduced macrophage microbicidal function [106 ]. Impaired macrophage microbicidal function may contribute to the increased susceptibility to infection apparent in Stat1-/- mice.

Stat1 KO mice also demonstrated the existence of Stat1-independent mechanisms in IFN-{gamma} signaling. IFN-{gamma} may exert pro- and antiproliferative signals through Stat1-independent and Stat1-dependent pathways, respectively [152 153 154 ]. Although IFN-{gamma} treatment of lymphocytes and fibroblasts inhibited cell growth, the same treatment in Stat1-/- cells promoted proliferation and decreased apoptotic signals [133 , 153 , 154 ]. Microarray analysis revealed that many genes, including the growth-promoting c-myc and c-jun genes, were induced in response to IFN-{gamma} in Stat1-/- cells [152 ]. This signaling was not dependent on the Stat1-docking site of IFNGR1. Although much more susceptible than their WT counterparts, Stat1-/- mice are significantly more resistant to viral infection than IFNGR1/IFNAR1 double-KO mice, indicating the existence of IFN-dependent, Stat1-independent, antiviral mechanisms [155 ].

A loss-of-function, heterozygous mutation in the Stat1 gene was recently described in humans, resulting in a dominant-negative effect over the WT allele for Stat1 activation [156 ]. Although these patients demonstrated increased susceptibility to bacterial infection, viral infections took the normal clinical course. This suggests that IFN-{gamma}-induced, physiologically significant, Stat1-independent control of antiviral responses also occurs in humans.

Mechanisms of Stat1-independent IFN-{gamma} signaling are currently unclear. One possibility is the existence of an unknown IFN-{gamma} receptor or IFN-{gamma} receptor complex linked to an alternative signaling pathway. This hypothesis was suggested when it was found that anti-IFN-{gamma} antibodies affected the clinical course of EAE in IFNGR1 KO mice [75 ].


   IRF-1
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
CELLULAR EFFECTS OF IFN-{gamma}
 CONCLUDING REMARKS
 REFERENCES
 
The IRF gene family is intimately involved in mediating the type I and II IFN signal cascades. Members of the IRF family share similar structure, including an amino-terminal DNA-binding domain containing multiple tryptophan residues and a carboxy-terminal transcriptional activation or repression domain (reviewed in ref. [157 ]). IRF-1, IRF-2, and IRF-9 (p48, ISGF3-{gamma}) all participate in IFN-{gamma} signaling. Basal IRF-1 expression has functions in constitutive gene expression [134 ], but Stat1 and nuclear factor (NF)-{kappa}B interaction with promoter elements dramatically increases IRF-1 transcription [158 159 160 ]. IRF-1 expression is up-regulated in response to types I and II IFN, virus, or cytokines. IRF-1 transcriptional regulatory activity is regulated independently in response to IFN-{gamma}, virus, and dsRNA [161 ]. The physiological relevance of IRF-1 function in transcriptional control of IFN-{gamma}-regulated genes is highlighted by IRF-1 KO mice, in which many genes are submaximally induced by IFN-{gamma} when compared with their WT littermates [162 , 163 ]. IRF-1 functions are not limited to IFN-{gamma} signaling; IFN-{gamma}-independent activities in viral/bacterial infection, lymphocyte development, regulation of cell cycle, growth inhibition, apoptosis and tumor suppression have been reported [107 , 123 , 135 , 159 , 164 165 166 167 168 ].

IRF-1 drives inducible expression of many target genes through interaction with the IRF-E site, consensus G(A)AAAG/CT/CGAAAG/CT/C [102 , 164 ]. This specificity overlaps with the ISRE consensus site A/GNGAAANNGAAACT, recognized by ISGF3, which is induced by type I IFN and to a lesser extent, type II IFN [102 ]. In this way, IRF-1 is able to induce a subset of the full spectrum of IFN-inducible genes. IRF-2 also binds to the IRF-E site and largely functions to repress transcription of IRF-1-inducible genes through its transcriptional repressor domain [102 , 159 ].

Negative regulation of IFN-{gamma} signaling
Stat1 activation is inhibited within 1 h of IFN-{gamma} treatment, despite the continued presence of extracellular IFN-{gamma}, and so mechanisms must exist to control the extent of ligand stimulation of IFN-{gamma} signaling (Fig. 2 ) [117 , 169 ]. These mechanisms involve every level of the pathway.



View larger version (29K):
[in this window]
[in a new window]
  		Figure 2. Negative regulation of IFN-{gamma} signaling. A number of factors limit the extent and duration of IFN-{gamma} signal transduction. Receptor IFNGR2 (green) expression is tightly regulated in some cell types, whereas IFNGR1 (yellow) surface concentration may be regulated by the extent of recycling following receptor:ligand internalization. Protein tyrosine phosphatases (PTPs), such as SH2-containing tyrosine phosphatase-2 (Shp2), dephosphorylate Jak1, Jak2, and IFNGR1, and suppressors of cytokine signaling-1 (SOCS-1) and SOCS-3 interfere with Jak activity to turn off signaling after ligand binding. Stat1 activation is down-regulated by Stat1 dephosphorylation in the nucleus. IRF-2 antagonizes transcriptional activation of many (IRF-1-inducible) genes containing ISRE or IRF-E promoter elements by competing for binding sites without promoting gene expression. Signaling molecules activated by IFN-{gamma} are depicted in red text.

 
Following signal transduction, the IFN-{gamma}:IFNGR1 complex internalizes and enters the endosomal pathway, where the complex dissociates [170 ]. In many cell types, the IFNGR1 chain eventually recycles to the cell surface in its uncoupled, dephosphorylated form, and the ligand is degraded [45 , 171 , 172 ]. In some cell types, IFN-{gamma} signaling may induce degradation of the internalized receptor, thereby down-regulating IFNGR1 surface expression [45 ]. This may be a mechanism by which these cells can prevent overstimulation of the pathway through cellular desensitization. Such a mechanism may also allow differential regulation of IFN-{gamma}-induced responses in different cell types [37 ].

One of the most inducible targets of IFN-{gamma} is a specific feedback inhibitor, SOCS-1, which associates with Jak1/2, interfering with tyrosine kinase activity and inhibiting downstream IFN-{gamma} signaling [173 174 175 176 177 ]. SOCS-1 may also promote degradation of signaling machinery by binding to and targeting signaling molecules for the ubiquitin-proteasome pathway [178 ]. Overexpression of SOCS-1 in cells in vitro or in transgenic mice causes loss of responsiveness to IFN-{gamma} [176 , 179 ]. Conversely, SOCS-1-/- mice are hyper-responsive to microbial infection, exhibit enhanced macrophage cytocidal activity, and die of IFN-{gamma}-dependent, multisystem inflammatory tissue destruction in the absence of infection [180 , 181 ]. This indicates that IFN-{gamma} is normally present and performs important functions in vivo, even in the absence of infection. Another SOCS protein, SOCS-3, is also induced by IFN-{gamma} and negatively regulates IFN-{gamma} signaling, although perhaps less effectively than SOCS-1 [182 ].

Aside from SOCS protein function, IFN-{gamma} signal down-regulation can occur by receptor and Jak dephosphorylation by the PTPs Shp1 and Shp2 [183 184 185 186 187 ]. Shp2 null cells exhibit elevated tyrosine phosphorylation and DNA-binding ability of Stat1 and show enhanced suppression of cell growth with IFN-{gamma} treatment [184 ]. Shp1 KO mice (motheaten mice, me/me) die from a condition caused by dysregulation of a large number of cytokines, possibly including IFN-{gamma} [188 ].

Stat1 phosphorylation may also be controlled by nuclear dephosphorylation. Stat1 homodimers enter the nucleus and are actively retained by a mechanism dependent on the kinase function of Jak1 [189 ]. McBride et al. [137 ] proposed a model in which nuclear export is mediated by a leucine-rich nuclear export signal (NES) at residues 197–205, which is hidden when Stat1 is bound to DNA. Stat1 dephosphorylation by a nuclear protein tyrosine phosphate (PTP) allows the release of DNA and subsequent recognition of the NES by chromosome region maintenance/exportin 1 (CRM1), an export receptor for NES-containing proteins [190 , 191 ]. CRM1 then mediates Stat1 translocation through the nuclear pore in a Ran-GTPase-dependent manner. This model is supported by work from other groups, demonstrating Stat1 tyrosine dephosphorylation in the nucleus and subsequent recycling of inactive Stat1 to the cytoplasm, and is consistent with the crystal structure of Stat1 bound to DNA [137 , 189 , 192 , 193 ].

Signaling specificity
The many cytokines that use Jak-Stat signaling direct transcription through specific but not necessarily unique combinations of Jaks and Stats. This raises the question of how signal specificity is maintained when many cytokines use the same signaling machinery.

Evidence for a devoted role of Jaks in signal transduction of specific cytokines is controversial. Mutant cell lines lacking Jak1 or Jak2 have demonstrated deficiency in types I and II IFN signaling (Jak1) or IFN-{gamma} signaling (Jak2) [194 , 195 ]. These deficiencies are reflected in the Jak1 and Jak2 KO mice, suggesting that these Jaks are functionally specific to their cytokine systems, and other Jaks cannot substitute for them in vivo. Jak1 gene KO mice exhibit a number of abnormalities, all attributable to signaling deficits in three cytokine receptor families: type II cytokine receptors (i.e., types I and II IFN and IL-10 receptors); all cytokines that use the {gamma}c subunit for signaling, thereby promoting lymphopoiesis (e.g., IL-2, IL-4, IL-7 receptors); and gp130 receptor family members (e.g., IL-6, IL-11, leukemia inhibitory factor receptors) [196 ]. These results showed that in vivo, the role of Jak1 is essential and nonredundant in signal transduction and induction of biological response for a specific range of cytokine receptors. Jak2 gene KO in mice is embryonic lethal as a result of deficiencies in erythropoiesis [197 , 198 ]. Other than suppression of IFN-{gamma} signaling, the cytokine response deficit of Jak2 KO mice is largely nonoverlapping compared with Jak1 KO mice [197 , 198 ]. This indicates that other Jaks cannot functionally substitute for Jak2 in vivo, although Jak3 can substitute for Jak2 in vitro to produce Stat1 homodimers and activate transcription of class I major histomcompatibility complex (MHC) [199 ].

Although specific Stats appear to be dedicated to specific subsets of cytokine signaling, the same Stat or combination of Stats can be activated in different cytokine systems to give completely different biological effects (e.g., IFN-{gamma} and oncostatin M as described earlier). Possible mechanisms giving rise to specificity of response with respect to Stats include: different thresholds/timeframes of Stat activation may be required for specific biological responses, and specific Stat complexes induced by cytokines may give different biological effects. It is widely agreed that signal specificity is largely conferred by the particular Stats activated in cytokine signaling [106 , 151 , 200 201 202 203 204 205 ]. This is supported by the Stat1 KO mouse, which displays loss of only IFN signaling, despite multiple reports of Stat1 activation in response to a range of other cytokines in vitro [106 , 151 ]. Similarly, KO mice of the other Stat proteins showed a loss of biological responsiveness to a single cytokine, indicating that Stat activation may be specific for a single cytokine in vivo if not in vitro [169 ]. In addition to the Stats themselves, other factors (e.g., intracellular IFN-{gamma} as described above, or cell-specific transcription factors), which associate with Stats may contribute to the specificity of the signal.

Cross-talk between IFN-{alpha}/ß and IFN-{gamma} pathways
The IFN-{gamma} and IFN-{alpha}/ß signal pathways cross-talk at multiple levels. The signal pathways and target genes used by types I and II IFN are partially overlapping, which gives the opportunity for cross-talk to synergize or antagonize particular functions within the cell. This cross-talk is biologically relevant, as cells in vivo are not stimulated with one cytokine in isolation, rather a cytokine cocktail instructs gene expression through the integration of multiple signaling pathways.

Although IFN-{gamma} primarily signals through Stat1 homodimers in the initial signal cascade, additional signaling molecules are activated. As noted above, the archetypal type I IFN signaling molecule ISGF3 is activated by IFN-{gamma}, thus providing a mechanism for cross-talk between the types I and II IFN pathways [116 ]. ISGF3 is able to induce expression of type I IFN, thereby further amplifying the response of IFN-{alpha}/ß-induced genes [206 , 207 ]. Conversely, type I IFN can elicit classic type II IFN signaling molecules such as active Stat1 homodimers, which are able to bind to GAS sites to activate transcription of target genes [17 , 208 ].

Takaoka et al. [209 ] published data indicating that IFN-{gamma} signaling in mouse embryonic fibroblasts requires a constitutive subthreshold IFN-{alpha}/ß signal. They have suggested that IFN-{alpha}/ß may promote the interaction of the IFNGR2 and IFNAR1 chains in the caveolar membrane domains of the cell membrane. In this model, low level IFN-{alpha}/ß signal is necessary for maintaining its receptor in the phosphorylated form, thus providing a functional aid for efficient assembly of IFN-{gamma}-activated Stat1 homodimers.


   CELLULAR EFFECTS OF IFN-{gamma}
TOP
 ABSTRACT
 INTRODUCTION
 THE IFNs
 THE IFN-{gamma} RECEPTOR
 IFN-{gamma} SYSTEM DYSFUNCTION
 SIGNAL TRANSDUCTION
 Stat1
 IRF-1
 CELLULAR EFFECTS OF IFN-{gamma}
CONCLUDING REMARKS
 REFERENCES
 
Class I antigen presentation pathway
Types I and II IFN up-regulate multiple functions within the class I antigen presentation pathway to increase the quantity and diversity of peptides presented on the cell surface in the context of class I MHC. Up-regulation of cell-surface class I MHC by IFN-{gamma} (Table 1 ) is important for host response to intracellular pathogens, as it increases the potential for cytotoxic T cell recognition of foreign peptides and thus promotes the induction of cell-mediated immunity.


View this table:
[in this window]
[in a new window]
  		Table 1. IFN-{gamma}-Regulated Genes and Their Function in Producing IFN-{gamma} Effects

 
IFN-{gamma} stimulation induces a replacement of the constitutive proteasome subunits with "immunoproteasome" subunits. In unstimulated cells, the proteasome enzymatic subunits are ß1, ß2, and ß5, encoded outside of the MHC locus. IFN-{gamma} induces expression of new subunits, LMP2, MECL-1, and LMP7, which competitively replace ß1, ß2, and ß5 subunits of the proteasome, respectively [210 211 212 213 214 215 216 ]. In this way, new species of proteasome are formed, consisting of LMP2:MECL-1:LMP7 (the immunoproteasome), LMP2:MECL-1:ß5, or ß1:ß2:LMP7, as well as low levels of the ß1:ß2:ß5 constitutive proteasome [216 ]. Inducible proteasome replacement is thought to be a mechanism by which IFN-{gamma} can increase the quantity, quality, and repertoire of peptides for class I MHC loading. The quantity is increased, as overall expression levels of proteasome are increased. The cleavage specificity of the immunoproteasome may allow production of peptides better able to bind class I MHC and thereby increase efficiency in the system. Peptide diversity is thought to increase as a result of differences in cleavage specificities between different species of proteasome. As a whole, this serves to increase levels and diversity of epitopes presented for CD8+ T cell recognition in the context of class I MHC and thus increase immune surveillance (reviewed in ref. [216 ]). This mechanism may have evolved to ensure LMP2/LMP7/MECL-1-dependent epitopes are only produced in sites of inflammation and thus avoid autoimmunity without compromising appropriate T cell stimulation [216 ].

The efficiency of peptide generation is further increased with the IFN-{gamma}-induced PA28, which is composed of PA28{alpha} and PA29ß subunits and associates with the proteasome and alters proteasome proteolytic cleavage preference. It is thought to gear antigen processing toward more efficient generation of TAP- and class I MHC-compatible peptides to increase overall efficiency of class I MHC peptide delivery [217 , 218 ].

The IFN-{gamma}-inducible TAP transporter is vital in peptide transport from the cytosol to the ER lumen [307 ]. TAP transiently associates with class I MHC to aid in efficient peptide loading [307 , 308 ]. Genes encoding the TAP subunits are located within the class II MHC locus and are coordinately expressed with MHC class I [220 , 221 ].

The class I MHC complex is composed of a heavy chain (consisting of {alpha}1, {alpha}2, and {alpha}3 domains) and a light chain (ß2microglobulin) and is up-regulated by IFN-{gamma} stimulation [4 , 222 , 224 225 226 227 228 229 230 231 232 ]. Chaperones such as tapasin and GP96 implicated in aiding in the efficient assembly of peptide:MHC I complexes are also up-regulated by IFN-{gamma} [4 , 309 , 310 ].

Class II antigen presentation pathway
Of the IFNs, IFN-{gamma} alone can efficiently up-regulate the class II antigen presenting pathway and thus promote peptide-specific activation of CD4+ T cells [4 , 234 ]. IFN-{gamma} treatment further up-regulates class II MHC molecules in cells constitutively expressing class II MHC, such as B cells, DCs, and cells of the monocyte-macrophage lineage (professional APCs) [234 ]. IFN-{gamma} is also able to induce class II MHC expression in cells that do not constitutively express these genes (nonprofessional APCs) [76 ]. IFN-{gamma} up-regulates the quantity of peptide:MHC II complexes on the cell surface by promoting expression of several key molecules (Table 1) : the Ii chain and the components of the class II MHC complex [234 235 236 237 238 239 ]; cathep sins B, H, and L, lysosomal proteases implicated in production of antigenic peptides for class II MHC loading [240 , 241 ]; and DM, a regulator of peptide accessibility to the peptide-binding cleft of class II MHC [235 , 237 ].

The CIITA mediates coordinate transcriptional control over these genes. Targeted disruption of CIITA in mice causes complete loss of constitutive or inducible display of class II MHC molecules [311 ]. Likewise, loss of constitutive/inducible class II MHC display is found in humans with CIITA mutation (Bare Lymphocyte Syndrome) [312 313 314 ]. As CIITA is the limiting factor in a complex that directs transcriptional regulation, it acts as a switch for rapid up-regulation of class II MHC-related genes by IFN-{gamma} [315 ].

IFN-{gamma} and development of Th1 response
IFN-{gamma} is a major product of Th1 cells and further skews the immune response toward a Th1 phenotype. IFN-{gamma} achieves this by promoting characteristic Th1 effector mechanisms: innate cell-mediated immunity (via activation of NK cell effector functions), specific cytotoxic immunity (via T cell:APC interactions), and macrophage activation (discussed below) [4 ]. IFN-{gamma}-induced, specific cytotoxic immunity is promoted by direct and indirect mechanisms. IFN-{gamma} promotes specific cytotoxic immunity by indirect mechanisms, such as growth inhibition of Th2 populations and up-regulation of antigen processing, presentation, and APC costimulatory molecules, thereby increasing CD4+ differentiation. IFN-{gamma} also influences naïve CD4+ cell differentiation toward a Th1 phenotype more directly. The phenotype adopted by a naive T cell during T cell activation is heavily influenced by the cytokine milieu present at the time of T cell receptor engagement. IFN-{gamma} and IL-12 are the prototypic cytokines directing Th1 differentiation during the primary response to antigen, and IL-4 directs differentiation of Th2 populations. IFN-{gamma} induces IL-12 production in phagocytes [296 ] and inhibits IL-4 secretion by Th2 populations [39 ], which may further drive Th1 differentiation in vivo.

IL-12 and IFN-{gamma} coordinate the link between pathogen recognition by innate immune cells and the induction of specific immunity, by mediating a positive feedback loop to amplify the Th1 response [4 ]. LPS and other pathogen-associated molecular patterns directly trigger IL-12 production upon recognition by macrophages, DCs, and neutrophils [316 317 318 319 ], which in turn induces IFN-{gamma} secretion in antigen-stimulated, naive CD4+ T cells and NK cells [320 , 321 ]. IL-12-induced IFN-{gamma} participates in positive feedback by further promoting IL-12 production in macrophages [296 , 297]. This amplification may be important in initiation or stabilization of the Th1 response (reviewed in refs. [4 , 322 ]). IL-18 is highly synergistic with IL-12 for IFN-{gamma} production and has important functions in the in vivo Th response [23 , 24 ].

Secretion of large amounts of IFN-{gamma} or IL-4 is the defining feature of Th1 or Th2 cells, respectively [323 ]. These cytokines function to directly promote cell-mediated immunity (IFN-{gamma}) or humoral immunity (IL-4) and to reciprocally antagonize each other’s actions, further polarizing the response (reviewed in ref. [108 ]). For example, IL-4 inhibits IFN-{gamma}-dependent activation of macrophage-effector functions [324 , 325 ]. Conversely, IFN-{gamma} antagonizes IL-4-dependent induction of the IgE receptor, Fc{varepsilon}RII [326 ].

The IFN-induced antiviral state
The IFN system regulates innate and adaptive immunity to viral infection. Viral invasion directly triggers induction of type I IFNs, through a mechanism involving IRF-3 and IRF-7 (reviewed in ref. [327 ]). Although both types of IFN are crucial in the immediate cellular response to viral infection, the immunomodulatory activities of IFN-{gamma} (discussed elsewhere) become important later in the response in coordinating the immune response and establishing an antiviral state for longer term control [54 , 60 , 328 , 329 ].

IFN-{gamma}-induced antiviral mechanisms include induction key antiviral enzymes (Table 1) , most notably PKR, which is a serine/threonine kinase greatly induced by types I and II IFN stimulation [242 , 330 ]. PKR is inactive in its constitutive form and requires an activating signal for autophosphorylation. dsRNA, a necessary intermediate in replication of RNA viruses, is the best characterized activator of PKR, although other agents such as heparin are able to activate PKR [331 ]. Association of PKR with dsRNA is likely to cause a conformational change that unmasks the catalytic domain responsible for PKR autophosphorylation [242 , 332 ]. PKR is then activated for dsRNA-independent phosphorylation of specific cellular substrates [333 ]. One of these substrates is the eIF-2{alpha} subunit, a rate-limiting factor in the normal cellular translational machinery. Phosphorylation by PKR prevents recycling of eIF-2–GTP from its guanosine 5'-diphosphate-bound form, thereby inhibiting viral and cellular protein synthesis [242 ]. PKR is implicated in numerous other functions, including activation of NF-{kappa}B [243 , 334 ], TNF-{alpha} transcript splice regulation [244 ], induction of fas-mediated apoptosis [245 , 246 ], and regulation of Stat1 activity [154 , 247 ].

It is likely that IFNs induce many as-yet undiscovered antiviral proteins, as the enzymes above and described in Table 1 do not account for all IFN-induced antiviral effects [335 , 336 ]. Also, the host of factors mediating the profound growth-inhibitory and proapoptotic effects of IFN-{gamma} would have no small part in limiting viral replication within the cell and spread to other cells.

Cell cycle, growth, and apoptosis
Macrophages are produced in large amounts by the bone marrow, and many die shortly after production by a process of programmed cell death ("apoptosis") [337 ]. After release from the bone marrow, macrophages proliferate, differentiate, acquire specialized functions (e.g., microbicidal activity), or die by apoptosis depending on the presence or absence of extracellular cues. Many reports indicate that IFN-{gamma} arrests the macrophage cell cycle and provides a survival signal, and others suggest that it serves as a proapoptotic signal. Although recent advances in cell cycle control and apoptosis have greatly increased our understanding in these functions in isolation, the inter-relationship between these intimately related processes is still somewhat confusing. For this reason, the effects of IFN-{gamma} on cellular proliferation and apoptosis are described here separately.

One of the most easily observed effects of IFN-{gamma} is cell growth inhibition. Types I and II IFN are able to protect against pathogen-induced apoptosis and suppress colony stimulating factor type 1 (CSF-1)-dependent growth of bone marrow-derived macrophages (BMM) [252 ]. Granulocyte macrophage-CSF, which is usually released by T cells at the same time as IFN-{gamma}, is able to protect against the growth-inhibitory effects of IFN-{gamma} on macrophages, and this may be physiologically important in vivo [338 ]. The IFNs most commonly arrest the cell cycle at the G1/S checkpoint, although blockages of other cell cycle stages have been reported [339 340 341 ]. It has been suggested that IFN-induced G1 arrest may actually reflect exit from the cell cycle into G0 [342 ].

IFNs inhibit proliferation primarily by increasing protein levels of Ink4 and Cip/Kip CKIs [252 253 254 255 , 343 , 344 ]. IFN-{gamma} transcriptionally induces p21 and p27 CKIs (Table 1) [256 , 345 , 346 ]. p21 and p27 inhibit the activity of cyclin E:CDK2 and cyclin D:CDK4 complexes, respectively, thereby arresting the cell cycle at the G1/S boundary. Decreased cyclin:CDK activity during G1 results in hypophosphorylation of the tumor suppressor Rb. In its hypophosphorylated state, Rb sequesters the E2F family of transcription factors from activation of genes (e.g., c-myc) required for cell cycle progression from G1 to S phase [292 ]. c-myc controls G1/S transition by activating cyclin:CDK complexes and inducing transcription of genes required for S phase (reviewed in ref. [263 ]). Expression of c-myc is induced rapidly by a number of growth factors and cytokines and is down-regulated by IFN-{gamma} [154 , 264 ]. IFN-{gamma} influences c-myc expression through multiple pathways, including suppression of Rb phosphorylation, resulting in decreased E2F activity as well as Rb-independent pathways, which may be a result of PKR action [251 ]. In its active form, c-myc is associated with max, and the myc:max heterodimer activates transcription of genes required for cell cycle progression [265 ]. The level of active myc is negatively regulated by mad1, which sequesters the max coactivator away from myc and suppresses transcription of myc-inducible genes [259 ]. In this way, the mad1-to-c-myc ratio determines whether a cell will proliferate (as a result of high levels of myc:max) or arrest in G1 (as a result of high levels of mad1:max) [259 , 260 ]. In addition to decreasing the abundance of myc, IFN-{gamma} increases mad1 levels, thereby further antagonizing myc activity and inhibiting CSF-1-dependent proliferation in BMM [347 ].

Apoptosis is a process in which the cell directs its own demise by activating intrinsic suicide machinery (reviewed in refs. [337 , 348 ]). In macrophages, apoptosis may be desirable to prevent colonization by intracellular pathogens. Conversely, many pathogens actually induce host-cell death through secretion of macrophage-specific toxins. Induction of apoptosis by signals such as DNA damage requires the IRF-1 tumor-suppressor gene [165 , 266 , 267 ]. Levels of IRF-1 may be a deciding factor in whether IFN-{gamma} induces or protects from apoptosis on treated cells [35 , 349 , 350 ]. It is proposed that IFN-{gamma} treatment of cells with high levels of functional IFNGR very rapidly activates Stat1, thereby producing high levels of IRF-1 that are able to induce apoptosis. In contrast, IFN-{gamma} treatment of cells with low levels of functional IFNGR may activate Stat1 more slowly, thereby producing lower levels of IRF-1 that are not sufficient to induce apoptosis [35 ]. Experiments in which overexpression of functional IFNGR on normally low-level IFNGR-expressing cells changed the IFN-{gamma} response of these cells from an antiapoptotic/proliferative phenotype to a proapoptotic phenotype are consistent with this hypothesis [35 ]. This may also explain why myeloid cells are more sensitive to the proapoptotic actions of IFN-{gamma} than other cells such as T cells, as myeloid cells express relatively high numbers of functional IFNGR on the cell surface [35 ].

Many of the proapoptotic effects of IRF-1 are mediated by the IRF-1-induced caspase 1 (IL-1ß-converting enzyme) [