The regulation of inflammation by interferons and their STATs.

Interferons (IFN) are subdivided into type I IFN (IFN-I, here synonymous with IFN-α/β), type II (IFN-γ) and type III IFN (IFN-III/IFN-λ) that reprogram nuclear gene expression through STATs 1 and 2 by forming STAT1 dimers (mainly IFN-γ) or the ISGF3 complex, a STAT1-STAT2-IRF9 heterotrimer (IFN-I and IFN-III). Dominant IFN activities in the immune system are to protect cells from viral replication and to activate macrophages for enhanced effector function. However, the impact of IFN and their STATs on the immune system stretches far beyond these activities and includes the control of inflammation. The goal of this review is to give an overview of the different facets of the inflammatory process that show regulatory input by IFN/STAT.

Inflammation constitutes an essential part of the innate immune response to pathogens or the release of self molecules acting as endogenous danger signals. Exposure of peripheral tissue to pathogen- or danger-associated molecular patterns (PAMPs and DAMPs, respectively) stimulates the release of proinflammatory mediators by tissue-resident cells that activate the endothelium of blood vessels and initiate a chain of events that ends with the transmigration of blood leukocytes and their penetration of the infected or otherwise irritated tissue. The strength and persistence of the proinflammatory stimulus decides whether systemic responses such as the mobilization of bone marrow leukopoiesis or the liver acute phase response ensues. The potentially harmful consequences of the inflammatory response need to be tightly controlled. Otherwise inflamed tissue may be irreversibly damaged as a consequence of lytic enzyme release or through oxidative stress. Moreover, an overshooting systemic response may cause a generalized shock syndrome., Consequently the outcome of the innate response to infection is determined by the balance between microbicidal effects and the damage inflicted to the host organism by the inflammation-induced loss of cell, tissue or organ function.

A large number of cytokines and chemokines that regulate the generation, trafficking and effector activity of leukocytes forming the inflammatory cell infiltrate control inflammation. In some cases the predominant effect of these cytokines is clearly proinflammatory, as in the case of TNF, or predominantly anti-inflammatory as in the case of TGF-β or IL-10. In other cases cytokines may act to support or suppress inflammation, depending on context. Interferons (IFN) are frequent contributors to the inflammatory cytokine stew. According to structural similarities, IFN and IL-10 families are grouped as class II cytokines.– Based on their evolution, structure and interaction with distinct receptor complexes IFN are subdivided into three distinct types. With together around 20 members the mammalian type I IFN (IFN-I) includes more than 10 IFN-α and usually a single IFN-β. Most likely all tissues and cell types produce IFN-I when exposed to appropriate pathogen or danger-associated molecular patterns. This contrasts the type II IFN, IFN-γ, which is produced predominantly by various T cell and NK cell populations. Type III IFN is comprised of three family members called IFN-λ1−3, or, synonymously, IL-29, IL-28A and IL-28B. The conditions and molecular mechanism controlling their synthesis are most likely similar, although not identical,, to those of IFN-I and both differ strongly from the regulation of IFN-γ synthesis by T and NK cells. Unlike IFN-I and IFN-γ, IFN-III receptors show highly restricted tissue distribution and appear to be expressed mainly on epithelia and, in humans but not in mice, on hepatocytes.

Receptors of all IFN types belong to the class II of cytokine receptors and share the attribute of employing JAK-STAT signal transduction for nuclear signaling.,, In keeping with highly similar biological properties of IFN-I and IFN-III, their receptor complexes (IFNAR and IFNλR, respectively), although composed, respectively, of IFNΑR1 and IFNΑR2 chains and IL28R/IL-10R2 chains, associate with JAK1 and Tyk2 kinases to phosphorylate and activate STAT1 as well as STAT2. This causes the formation of a STAT1/STAT2 heterodimer that associates with a third subunit, IRF9, to form the transcriptional complex ISGF3. By contrast the IFN-γ receptor (IFNGR), composed of IFNGR1 and IFNGR2 chains, uses JAK1 and JAK2 kinases and strongly favors association with STAT1 over all other STATs. Consistently, transcriptional responses to IFN-γ are dominated by the activity of the STAT1 homodimer. Although STAT3, STAT4 and STAT5 are activated by IFN in some cell types, all available evidence suggests that STATs 1 and 2 are the main mediators of cellular and organismic IFN biology. Hence the terms “STAT” and particularly “IFN/STAT” are used in this review to indicate STATs 1 and 2, unless there is explicit reference to a different member of the STAT family.

To study the impact of individual IFN types in animal models of inflammatory disease, mice deficient for the IFNΑR1 chain or for either the Ifng or Ifngr1 gene are used. IL28R−/− mice have been generated, yet publications describing their application in experimental models of inflammatory disease are scarce. STAT1−/− mice are used to eliminate the impact of all IFN. Although not subject of this review, STAT1 is also activated by receptors for IL-27 and IL-35, which may influence the contribution of T cells, which would be particularly relevant for IL-35-mediated Treg differentiation. Inflammatory responses have not been widely studied in STAT2−/− mice. In theory this should equal a composite deficiency for IFN-I and IFN-III responses. However, in some tissues the lack of STAT2 causes a significant drop of STAT1 levels, thus producing at least partial absence of IFN-γ responsiveness. Thus, phenotypic properties of STAT2−/− mice are more difficult to interpret.

IFN/STAT upregulate the immunocompetence of many cell populations, which is particularly evident in the case of macrophage activation. This function corresponds to a local or systemic increase in IFN production. Particularly, type I IFN fulfill an important role as inducers of tonic IFNΑR signaling in the steady-state. Low levels of synthesis provide sufficient stimulus to maintain transcription of inflammation/immunity-relevant genes and to keep cells in a state of alertness., Recent studies show that constitutive IFN-I synthesis can be stimulated by commensal bacteria and their manipulation of host cell chromatin. In interpreting data from IFN/STAT-deficient mice this constitutive IFN activity must be taken into account.

Impact of IFN and STAT1 on the Generation, Differentiation and Death of Hematopoietic Cells

Persistent inflammation stimulates bone marrow hematopoiesis and thus produces an increase in blood leukocyte numbers. Mice deficient for IFN receptors have no major hematopoietic abnormalities, thus IFN are not major regulators of steady-state hematopoiesis. It appears clear, however, that they feed information from ongoing infection/inflammation to hematopoietic stem cells (HSC) of the bone marrow. Treatment of mice with the proinflammatory, IFN-inducing agent poly-IC causes the mobilization of dormant hematopoietic stem cells (HSC) and this response is abrogated in mice lacking either the IFN-I receptor or STAT1. Exhaustion of the stem cell niche in this situation is prevented by the interferon regulatory factor-2 (IRF2), a negative regulator of IFN and IFN-induced genes. In HSC IRF2 abolishes IFN-α-mediated suppression of genes limiting excessive cycling and maintaining the self-renewing potential of HSC. In a murine model of mycobacterial disease HSC mobilization required IFN-γ and STAT1. IFNAR deficiency had a minor effect in this situation despite the ability of Mycobacteria to stimulate IFN-I synthesis. Together these studies suggest that STAT1 target genes push HSC from a dormant state into the cell cycle, shifting the balance toward enhanced hematopoietic differentiation. Contrasting mouse HSC, the growth of human CD34+ hematopoietic cells, which represent uncommitted progenitors, is inhibited by IFN-α. This response requires the p38MAPK pathway and may be independent of STAT signaling. A recent study suggests the importance of the IFN-γ pathway in regulating the differentiation of inflammatory dendritic cells in situ. Intraperitoneal infection of mice with Toxoplasma gondii causes the replacement of resident peritoneal leukocytes with blood-borne monocytes that differentiate in situ in both macrophages and inflammatory DC. These events required NK cell-derived IFN-γ. The study did not include STAT1−/− mice.

During an infection cells may become sensitive to a death-enhancing effect of IFN-I. Studies from our lab demonstrate enhanced death of macrophages infected with intracellular L. monocytogenes due to the activity of a pathway requiring IFN-I, STAT1 and NO production. Possibly linked to this, death mechanisms requiring Rip3 kinase or caspase 1/11 activation referred to as necroptosis or pyroptosis, respectively, are enhanced by IFN-I., In the latter situation upregulation of the inflammasome subunit Aim2 by STAT-signaling provides an explanation for enhanced bacteria-induced macrophage death. In virally infected mice IFN-I-mediated neutrophil depletion may occur. This provides a potential explanation for the increased sensitivity to bacterial superinfection in the wake of viral disease. It should be noted however that reduced neutrophil counts upon virus infection have also been explained by IFN-I-mediated alterations in chemokine production., During infection with the intracellular bacterium Listeria monocytogenes massive death of splenic T cells occurs and this is almost completely reversed upon conditional deletion of either the IFNΑR or STAT1., Removal of apoptotic lymphocytes is thought to stimulate IL-10 synthesis and the suppression of a protective inflammatory response. Treatment of mice with TNF increased the fraction of apoptotic enterocytes and hepatocytes. The apoptotic response was reduced in mice deficient for the IFN-I receptor. Together these reports support the view that IFN/STAT are important regulators of cell survival during innate, proinflammatory immune responses. However, favorable or adverse consequences of abrogating their regulatory input are observed, depending on the proinflammatory stimulus and environment.

Regulation of Cell Trafficking: the Impact of IFN/STAT on Chemokine Expression

The recruitment of leukocytes to sites of infection or to sites exposed to non-infectious inflammatory agents, and the subsequent formation of inflammatory infiltrates, is controlled by the CCL and CXCL families of chemokines A large number of studies document the ability of IFN-I, IFN-γ or both, to regulate chemokine synthesis (Table 1). Among the chemokines consistently reported to show IFN regulation are the CCR2 ligand CCL2, CCL5, which binds to multiple receptors, and the CXCR3 ligands CXCL9, CXCL10 and CXCL11. Not all studies investigating the regulation of chemokines by IFN establish a clear link to STAT1 dimer or ISGF3 activity. The interpretation of gene targeting is complicated by the fact that STAT1-deficient mice reveal differences in chemokine production, but do not allow to distinguish between direct action of STAT1 on chemokine promoters and more indirect effects. That said, several chemokine promoters such as those regulating CCL2/MCP1, CCL5/RANTES, CXCL9 or CXCL10 contain binding sites for STAT1 dimers and/or ISGF3, thus fulfilling this important criterion of direct target genes.– The CCL5/RANTES promoter associates with members of the interferon regulatory factor family through a proximal promoter element, its regulation by IFN-γ thus involves a STAT1-IRF1 pathway. The vast majority of investigated chemokine promoters respond to signals from classical proinflammatory pathways, most prominently the NFκB pathway. Functional cooperation between IFN and NFκB- activating cytokines, or, more directly, between STATs, IRFs and NFκB is frequently observed.,,– Moreover, the tissue-specific transcription factor Pu.1 was found to allow for the IFN-γ-induced expression of CXCL9 in myeloid cells, but not other cell types. During infection with L. monocytogenes CCL2/MCP1 was shown to be under control of the MyD88 pathway early after infection and predominantly under IFN-I control at a later stage. This regulatory switch may indicate cooperativity of NFκB and STAT pathways by sequential deployment.

Table 1. Regulation of chemokine synthesis by IFN and STATs 1/2

Cell/animal	Stimulus/disease	IFN type involved	STAT involved	Chemokine regulated	References
Mouse macrophages	IFN-γ or IFN-γ/TNF	IFN-γ	STAT1	CCL2/MCP1 ↑	37, 45 and 128
				CXCL9/MIG ↑
				CXCL10/IP10 ↑
				CCL12/MCP5 ↑
	IFN-γ/PamCys	IFN-γ	nd	CCL2/MCP1 ↑	52
	LPS	nd	STAT1	CXCL10/IP10 ↑	45 and 129
				CCL12/MCP5 ↑
	LPS/IFN-γ	IFN-γ	STAT1	CXCL1/KC/GROα ↓	45
				CXCL2/MIP2/GROβ ↓
				CCL2/MCP1 ↓
				CCL4/MIP1β ↓
	TNF	IFN-β	STAT1	CCL5/RANTES ↑	11
				CXCL9/MIG ↑
				CXCL10/IP10 ↑
	Listeria monocytogenes	IFN-I	nd	CCL2/MCP1 ↑	42
Mouse splenocytes	IFN-α	IFN-α	STAT1	CCL2/MCP1 ↓	46
				CCL5/RANTES ↓
				CCL7/MCP3 ↓
Human monocytes	IFN-γ	IFN-γ	STAT1	Inhibition of migration in response to CCL2	130
	TNF	IFN-β	STAT1	CCL5/RANTES ↑	11
				CXCL9/MIG ↑
				CXCL10/IP10 ↑
	M1 polarization (IFN-γ/LPS)	IFN-γ	nd	CCL5/RANTES ↑	44 and 131
				CCL15/MIP1δ ↑
				CCL20/MIP3α ↑
				CXCL9/MIG ↑
				CXCL10/IP10 ↑
Human monocyte-derived DC	Sendai virus	IFN-I	ISGF3	CCL19/MIP3β ↑	40
	Salmonella Typhimurium
Human PBMC	IFN-λ	IFN-λ	nd	CXCL9/MIG ↑	132
				CXCL10/IP10 ↑
				CXCL11/I-TAC ↑
Mouse primary cortical neurons	IFN-α	IFN-α	STAT1	CXCL10/IP10 ↑	133
Human astrocytes	IFN-γ/IL-1β	IFN-γ	nd	CXCL11/I-TAC ↑	36
	IFN-β/IL-1β	IFN-β	ISGF3 (?)	CCL5/RANTES ↑	32
Mouse microglia	IFN-γ	IFN-γ	STAT1	CXCL9/MIG ↑	41
Human plasmacytoid DC (MS patient)	IFN-β/TLR9 ligand	IFN-β	nd	CCL3/MIP1α ↓	48
				CCL4/MIP1β ↓
				CCL5/RANTES ↓
Human T cells (MS patient)	IFN-β	IFN-β	nd	CCL3/MIP1α ↓	47
				CCL5/RANTES ↓
				CCR5 ↓
Mouse	MCMV infection	IFN-I	nd	CCL2/MCP1 ↑	49
				CCL3/MIP1α ↑	50
				CCL7/MCP3 ↑	51
				CCL12 ↑
Mouse liver	MCMV infection	IFN-γ	nd	CXCL9/MIG ↑	49
Mouse brain	LCMV infection	IFN-γ	nd	CCL2/MCP1 ↑	55
				CCL3/MIP1α ↑
				CCL5/RANTES ↑
				CXCL10/IP10 ↑
Mouse cornea	HSV-1	IFN-α	nd	CCL2/MCP1 ↑	34
Mouse	Vaccinia virus	IFN-I, IFN-γ	nd	CXCL9/MIG ↑	134
	West Nile virus	IFN-I	nd	CCL2/MCP1 ↑	135
				CCL5/RANTES ↑
				CXCL10/IP10 ↑
	Influenza virus	IFN-I	nd	CXCL1/KC ↓	23
	Streptococcus pneumoniae			CXCL2/MIP2 ↓
Mouse peritoneal cavity	Listeria monocytogenes	IFN-I	nd	CCL2/MCP1 ↑	42
		nd	STAT1 (macrophages)	CCL3/MIP1α ↓	26
			STAT1 (DC)	CCL2/MCP1 ↑
				CCL5/RANTES ↑
			STAT1 (T cells)	CCL2/MCP1 ↑
				CCL5/RANTES ↑
Mouse	Francisella tularensis	IFN-γ	nd	CCL2/MCP1 ↑	52
	Polymicrobial peritonitis (CASP)	IFN-I	nd	CCL2 ↓	78
	Polymicrobial peritonitis (CLP)	IFN-I	nd	CXCL10 ↑	53
	Trypanosoma cruzi	IFN-γ	nd	CXCL9/MIG ↑	136
				CXCL10/IP10 ↑
	Paracoccidioides brasiliensis	IFN-γ	nd	CCL5/RANTES ↑	137
				CXCL1/KC/GROα ↓
				CCL3/MIP1α ↓
	Candida albicans	IFN-I	nd	CCL2/MCP1 ↑	71
				CXCL1/KC/GROα ↑
Mouse thyroid gland	IFN-γ transgene (thyroid gland)	IFN-γ	nd	CXCL9/MIG ↑	56
				CXCL10/IP10 ↑
				CCL4/MIP1β ↑
				CCL5/RANTES ↑
				CXCL11/I-TAC ↑
Mouse mast cells	Imiquimod	IFN-I	nd	CCL2/MCP1 ↑	59
Mouse	TNF	IFN-I	nd	CXCL9/MIG ↑	27
				CXCL10/IP10 ↑
				CXCL11/I-TAC ↑
	Allergic asthma	IFN-γ	STAT1	CXCL9/MIG ↑	58
				CXCL10/IP10 ↑
	ConA-induced hepatitis	IFN-γ	STAT1	CXCL5/ENA-78 ↑	138
				CXCL9/MIG ↑
				CXCL10/IP10 ↑
				CXCL11/I-TAC ↑
Mouse (astrocytes)	EAE	IFN-γ	nd	CXCL10/IP10 ↑	57
Human (MS patient)	IFN-β	IFN-β	nd	CCL2/MCP1 ↑	139
				CXCL10/IP10 ↑

nd, not determined

Regulation by IFN is an attribute of chemokines associated with the IFN-γ/LPS-induced M1 polarization of macrophages whereas M2 polarization is associated with the production of a distinct set of chemokines., Therefore, besides activation of nonpolarized macrophages, the induction of macrophage polarization is one of several ways by which IFN/STAT can influence the chemokine spectrum synthesized during inflammation. In addition, IFN/STAT may change the abundance and composition of chemokine-producing cell populations. Active suppression of chemokine synthesis by IFN-I or IFN-γ in activated macrophages, splenocytes or pDC has also been reported, and may contribute to the successful treatment of MS patients with IFN-β., Chemokines not usually upregulated by IFN/STAT such as CXCL1 and CXCL2 can be suppressed, whereas others show situation-dependent upregulation or suppression (Table 1). The mechanisms underlying the inhibition of chemokine synthesis or the factors determining the balance between induced synthesis and inhibition are not known.

Both chemokines recruiting predominantly myeloid cells (neutrophils, inflammatory macrophages and DC; e.g., the CCR2 ligand CCL2, the CCR1/5 ligand CCL3, the CCR1/3/5 ligand CCL5, and the CCR1/2/3 ligand CCL7) and chemokines recruiting predominantly lymphoid cells (NK cells, effector T cells; the CXCR3 ligands CXCL9, CXCL10 and CXCL11) are controlled by IFN/STAT (Table 1). IFN/STAT-regulated chemokine synthesis thus exerts profound effects on the mobilization, tissue infiltration and activation of inflammatory cell populations. This in turn is thought to alter immune responses of IFNΑR, IFNGR, or STAT1-deficient mice and explain in part why such animals survive better or worse when infected or treated with inflammatory agents. While many studies document changes of cell recruitment in absence of IFN/STAT responses (most animal studies listed in Table 1, e.g., refs. 26, 34, 42 and 49–52) this alone does not support the conclusion of a causal relationship to the outcome of an immune response. For example, we noted significantly altered chemokine production in Listeria-infected mice lacking STAT1 in DC, but this had no impact on the survival of such animals. Comparison of IFN/STAT deficiency with mice lacking the regulated chemokine allows an estimation of the contribution of that chemokine to the immune response under study,,, but not of the contribution of its regulation by IFN/STAT. More convincingly the impact of chemokine regulation can be determined by rescuing the effects of IFN/STAT deficiency by chemokine injection. This approach has not been widely used.

Studies linking regulation of chemokines by IFN/STAT with the establishment of immunity suggest this can benefit or weaken host immunity. For example, infections with MCMV or HSV-1 viruses, or the intracellular bacteria L. monocytogenes and F. tularensis are accompanied by IFN/STAT-mediated upregulation of CCR2 ligands, particularly CCL2. These mediate the recruitment of inflammatory monocytes and neutrophils to increase resistance to infection. In addition to inflammatory monocyte recruitment by CCR2 ligands, MCMV infection triggers a protective cascade of IFN-α-upregulated CCL3, NK cell recruitment, IFN-γ synthesis and IFN-γ-induction of CXCL9 production. In accordance with the beneficial effect on MCMV infection, IFN-α/STAT1-induced chemokines increased resistance against corneal infection with HSV-1, an infection protocol resulting in viral spread to the brain stem., Contrasting these examples, intracranial infection with LCMV was worsened by IFN-γ-mediated recruitment of inflammatory cells, shown by the protective effects of a dominant negative IFN-γR expressed in macrophage-lineage cells. In addition to infection models with a single pathogen, chemokine synthesis was shown to underlie the reduced resistance of mice previously infected with influenza virus to secondary infection with S. pneumoniae. Shahangian and colleagues demonstrated that the inhibition of the CXCR2 ligands CXCL1 and CXCL2 by virus-induced IFN-I lead to a drop in neutrophil infiltrates and a corresponding inhibition of bacterial clearance.

Aside from infection, chemokine regulation by IFN/STAT has been studied using non-infectious inducers of inflammation. For example, expressing an IFN-γ transgene in the thyroid gland caused increased expression of CCL4, CCL5, CXCL9, CXCL10 and CXCL11 and a mononuclear infiltrate in the thyroid gland. In the brain of mice suffering from experimental autoimmune encephalitis (EAE), IFN-γ-mediated protection correlated with the expression of CXCL10/IP10 in astrocytes. Induction of hepatitis with the lectin Concanavalin A (ConA) in IFN-γ or STAT1-deficient mice resulted in decreased inflammatory infiltrates that correlated with reduced production of CXCL family chemokines in hepatic cells (Table 1). In a mouse asthma model IFN-γ and STAT1 contributed to allergic inflammation through enhanced production of CXCL9 and CXCL10. The synthetic TLR7 agonist Imiquimod is an effective treatment against human skin cancer. In a mouse melanoma model the drug was shown to stimulate mast cells for TLR7-dependent IFN-I synthesis. Subsequent IFNAR signaling caused CCL2 production and the recruitment of a tumoricidal plasmacytoid dendritic cell (pDC) infiltrate.

While synthesis of chemokines is an important factor for the recruitment of inflammatory infiltrates, additional regulation can be provided through the expression of their receptors and by cellular adhesion molecules that mediate leukocyte transmigration. Alterations of chemokine receptor expression contribute to the distinct migratory properties of naïve and effector lymphocytes, DC before and after maturation, or of differently polarized macrophages. While the role of IFN/STAT in these processes has not been widely studied, preliminary evidence for their participation has been reported.,, We expect future studies to confirm that all steps controlling leukocyte migration and tissue invasion are under surveillance of IFN/STAT.

Other Mediators of Inflammation Regulated by IFN/STAT

The inflammatory environment is shaped by the products of tissue resident cells as well as cells belonging to the inflammatory infiltrate. IFN-γ/STAT1 are well-established regulators of M1 polarization and classical activation of macrophages, thus controlling the synthesis of cytokines, nitric oxide (NO), reactive oxygen intermediates (ROI) and enzymes required for tissue remodelling.,,, Conditional STAT1 gene deletion in mice convincingly demonstrated the importance of macrophage activation in IFN-γ-dependent protective immunity and inflammation against L. monocytogenes infection. Apart from macrophages the important role of inflammatory dendritic cells has been widely recognized, cells that may differentiate in situ from inflammatory monocytes. One population of inflammatory DC are Tip-DC, characterized by production of large quantities of NO and TNF during L. monocytogenes infection. NO is an important contributor to inflammation owing to its properties as a microbicidal agent as well as a signaling molecule and regulator of cell death., NO synthesis is catalyzed by inducible nitric oxide synthase (Nos2 or iNOS). Whereas its regulation in Tip-DC has not been studied, reports performed in macrophages show that the Nos2 gene is synergistically activated by NFκB and STAT pathways. Interaction of a STAT1 dimer activates the Nos2 promoter when IFN-γ and PAMPS are present. By contrast PAMPS like LPS, or pathogens such as Listeria monocytogenes stimulate Nos2 transcription through an IFN-I intermediate and ISGF3 activation. ISGF3 and NFκB cooperate in the assembly of a transcription initiation complex with ISGF3 holding responsible for the recruitment of RNA polymerase II and NFκB for promoter binding of the kinases phosphorylating the carboxyterminal domain of RNA polymerase II. It will be of interest to determine in how far this mode of cooperation between STAT and NFκB pathways is paradigmatic for the regulation of proinflammatory genes and whether it extends to cell types other than macrophages, such as Tip-DC.

A novel, but as yet fairly unexplored activity of IFN/STAT is their regulation of IL1-β production by inflammasomes. IFN-I via STAT1 exert suppressive activity on NLRP1 and NLRP3 inflammasomes that induce caspase 1 to process the IL1-β precursor in response to a large variety of intracellular PAMPs. Two mechanisms have been proposed to explain the reduced IL1-β processing in IFN-I-stimulated cells. On the one hand STAT1 target gene products directly repress these inflammasomes. On the other hand the IFN-I/STAT1 pathway increases IL-10 synthesis, IL-10-mediated STAT3 activation and the suppression of IL1-β precursor synthesis by activated STAT3. A study addressing M. tuberculosis infection reported an intriguing connection between IFN/STAT and NO synthesis. Upregulation of NO production by IFN-γ caused nitrosylation and inactivation of the NLRP3 inflammasome, reduced IL1-β synthesis and a concomitant decrease in inflammatory pathology. Since IFN-I, like IFN-γ, is an inducer of Nos2 synthesis and NO production, the same mechanism may underlie the suppression of inflammasome-mediated IL1-β production by IFN-I. IFN-I-dependent suppression of IL1-β synthesis correlated with increased susceptibility to Candida albicans infection., It is also thought to contribute to the benefits of IFN-β for the anti-inflammatory treatment of MS, as monocytes from treated patients secrete significantly less IL1-β69.

IFN/STAT in Infection-Associated Systemic Inflammation and Sepsis

Treatment of mice with LPS causes a septic shock syndrome, and, ultimately, death. STAT1 enhances the systemic inflammation resulting from LPS administration through the recruitment and activation of macrophages and additional inflammatory leukocytes. Consistently, IFN-γ plays a well-documented role in the pathogenesis of the endotoxin shock., STAT1−/− mice survive moderate LPS quantities better than wildtype mice, but succumb to relatively high doses of LPS, that are survived by mice lacking a functional IFN-β gene., The reasons underlying this STAT1-independent contribution of IFN-β are not completely understood. One potential explanation is provided by the surprising finding that late-stage induction of IFN-induced genes after viral infection can occur in a STAT2-dependent, STAT1-independent manner. Possibly this STAT1-independent phase of the IFN response accelerates the septic shock syndrome. IFNΑR or IFN-β-deficient mice also show a remarkable degree of resistance when the septic shock syndrome is evoked by injection of TNF. Huys and colleagues attribute resistance to a combination of protection from cell death, reduced synthesis of proinflammatory cytokines, and deregulated chemokine production. The prominent contribution of IFN-I to systemic, pathogen-induced inflammation has sparked the idea of IFN-I neutralization as a means of anti-septic therapy.

A more physiologic way of studying sepsis is to injure the intestine by surgical procedures. Two commonly used methods are cecal ligation and puncture (CLP) or colon ascendens stent peritonitis (CASP). Surprisingly the two procedures produced opposing effects of IFN-I on resistance to the resulting polymicrobial peritonitis. CASP resulted in increased survival of Ifnar1−/− mice and a corresponding increase in CCL2 secretion, neutrophil infiltration and ROI production. By contrast, survival of Ifnar1−/− mice in comparison to wildtype controls was decreased following CLP. In this case IFNAR deficiency abolished IFN-I-mediated upregulation of CXCL10 and the concomitant increase of neutrophil phagocytotic activity. Counterintuitive to the report by Kelly-Scumpia, STAT1−/− mice showed increased resistance to CLP despite decreased production of IFN-α and CXCL10. At present the factors determining the difference between STAT1−/− and Ifnar1−/− in the CLP model are elusive.

To explain the discrepant role of IFN-I in the CASP and CLP models the authors suggest that the intensity of the inflammatory response may decide between adverse or protective effects of IFN-I. If correct, this implies that IFN-I are neither “good” nor “bad” regulators of inflammation, but that their protective or adverse character varies with more or less pronounced inflammatory environments. In line with this notion our recent findings show that the impact of IFN-I on L. monocytogenes infection varies with the route of infection. Whereas IFN-I worsen the outcome of i.p. infection, they protect after infection through the gastrointestinal tract. This is correlated with different kinetics and intensity of the proinflammatory response (ref. 28 and our unpublished results). In correspondence with the detrimental effects of type I IFN after CASP or intraperitoneal infection with L. monocytogenes, adverse effects of IFN-I were reported for a mouse model of C. albicans sepsis. Here, the protective effect of Ifnar1 gene deletion was explained by the lack of IFN-I-mediated upregulation of chemokines recruiting and activating neutrophils and inflammatory monocytes for increased kidney destruction.

IFN/STAT in Intestinal Inflammation

The incidence of intestinal inflammation in humans has strongly risen over past decades. According to current knowledge it results from a disturbed interplay between the intestinal mucosa and gut microbiota. The most prominent type of intestinal inflammation is inflammatory bowel disease (IBD), consisting of two major sub-pathologies named Crohn disease (CD) and ulcerative colitis (UC). The relevance of JAK-STAT signaling for IBD was recently confirmed by a report that the JAK inhibitor tofacitinib improved the condition of patients suffering from UC. In detail the JAK-dependent pathway(s) targeted by the inhibitor remain to be clarified.

The first direct evidence for a role of STAT1 in inflammatory bowel disease came from a study on human patients developing pouchitis, which is a major inflammatory complication after proctocolectomy in UC and familial adenomatous polyposis. In this study, increased expression and activation of STAT1 was observed in biopsies from inflamed mucosa compared with normal mucosa. The same increase was shown in samples of ulcerative colitis and CD patients. The cell types showing increased STAT1 (activation) were identified as infiltrating monocytes and neutrophils. Determination of STAT1 levels in lamina propria mononuclear cells (LPMC) and T cells of CD and UC patients by flow cytometry demonstrated a trend toward increased total STAT1 in LPMC during CD, whereas in UC a similar trend was observed for phosphorylated STAT1. Confirming this result, a microarray on biopsies showed STAT1 expression to be increased in CD, but not in UC and non-IBD infectious colitis. Biopsies from non-inflamed ileal pouches from a cohort of UC patients demonstrated a significant increase of phosphorylated STAT1 over biopsies of familial adenomatous polyposis patients and controls. A more recent study on patients with UC and CD showed STAT1 mRNA slightly but not significantly increased, while reporting significantly increased expression of STAT1-induced genes encoding IRF1, Socs1, IP10, IL-12/23 p40 and T-bet in active CD, and of IP10 and IL-12/23 p40 in active UC.

Prominent animal models of intestinal inflammatory disease are dextran sodium sulfate (DSS)-induced colitis, trinitrobenzene sulfonic acid (TNBS)-induced colitis, or the transfer of CD45RBhigh T cells into Rag-deficient mice. Surprisingly, data on STAT1-deficient mice in these models are scarce. DSS treatment of STAT1−/− mice on a 129S6/SvEv background suggested decreased tissue damage and hyaluronan deposition compared with wild-type controls, suggesting a contribution of STAT1 to colitis. In our own experiments STAT1 deficiency in a different genetic background slightly protected concerning crypt damage and amount of tissue involved in inflammation caused by DSS. There are no data on STAT1 deficient mice in other models of colitis, however recent papers using the colitis model of CD45RBhigh T cell transfer into Rag-deficient mice suggest that the balance of STAT1 and STAT3 in the intestine is crucial for the equilibrium of Treg/TH17/TH1 levels, and if disturbed, can lead to increased or decreased pathology., A small-molecule compound that triggers the tyrosine phosphorylation of Src homology 2-containing protein tyrosine phosphatase 2 (SHP-2) ameliorated TNBS colitis. The mechanism of this amelioration was shown to be interaction of tyrosine phosphorylated SHP-2 with cytosolic STAT1, preventing the recruitment of STAT1 to the IFNGR.

Reperfusion injury is a type of tissue damage caused when blood supply returns to tissue after a period of ischemia, which can occur in clinical settings. STAT1-deficient mice showed increased survival to ischemia/reperfusion of the small intestine. Their intestines were protected from gross histomorphological tissue destruction and neutrophil infiltration. In a study on celiac disease (gluten-sensitive enteropathy), STAT1 was found to be activated to higher levels than in controls, and in explant cultures of biopsies, gliadin induced the activation of STAT1.

An infectious cause of intestinal inflammation whose containment in the intestinal tract and subsequent clearance was shown to be entirely dependent on STAT1 signaling is Norwalk virus. After oral infection, wild-type animals clear murine Norwalk virus from the gastrointestinal tract within one day, while STAT1-deficient animals cannot contain it, develop pathology in several organs and die at a rate of over 70%. Rotavirus, another diarrhea-causing virus especially dangerous for newborns, was shown to be shed 100-fold more from STAT1-deficient animals than wild-types; however, the knockout animals could still clear the virus via the adaptive immune system. A later study demonstrated that the STAT1-dependent pathway for rotavirus containment was IFN-λ signaling in epithelial cells of the intestine.

Taken together the data suggest that STAT1 upregulation and/or activation are hallmarks of colonic inflammation. However, the role of STAT1 as a colitogenic driver is modest and varies depending on etiology.

Compared with STAT1, a larger data set is available describing the role of cytokines signaling via STAT1 for the development of IBD/colitis. Already in 1996, it was shown that IFN-γ is increased in human CD. In clinical trials with fontolizumab, an IFN-γ blocking antibody, the clinical response was weak, but reduced inflammation was noted based on decreases in C-reactive protein levels., T cell transfer-induced colitis in mice was prevented by the administration of an anti-IFN-γ antibody for up to 12 weeks. Systemic administration of an anti-IFN-γ antibody produced a rather mild, yet significant beneficial effect on the development of DSS-induced colitis., Consistently, IFN-γ-deficient animals were protected as well. The effect of IFN-γ deficiency was subsequently attributed to homeostatic functions of IFN-γ signaling in enterocyte proliferation and apoptosis through serine- threonine protein kinase AKT-β-catenin and Wingless-Int-β-catenin signaling pathways. Contrasting the DSS or T cell transfer models, TNBS-induced colitis remained unaffected by antibody-mediated neutralization of IFN-γ activity. Moreover, TNBS administration had equal effects on IFN-γ-receptor-deficient mice and wild type controls., Surprisingly, knockout mice for the ligand, IFN-γ, were slightly protected from mortality upon high TNBS doses however at intermediate doses the disease rather changed from a TH1 type to a TH2 type pathology in the absence of IFN-γ.

Mice deficient for IL-10 spontaneously develop colitis. Interestingly, colitis in these animals was exacerbated upon IFN-γ deficiency. This anti-inflammatory role of IFN-γ was explained through inhibition of proinflammatory IL-23 in colonic Cd11b+ cells.

Data on IFN-I in colitis have recently been reviewed elsewhere. In brief, IFN-I produced by intestinal dendritic cells upon TLR9 stimulation or given systemically are thought to protect against acute DSS colitis. Our own studies suggest that the benefit of IFN-I varies with the severity of colonic inflammation. Protective effects are seen only at high DSS concentrations and a correspondingly strong inflammation (our unpublished results). As with models of bacterial infection and sepsis, the impact of IFN on inflammation again varies with the intensity of the inflammatory process. Clinical data from IFN-I treated IBD patients appear to confirm this notion by showing highly variable responses. Contrasting bacterial sepsis however, IFN-I protect better with increasing severity of DSS-induced colitis. This may be explained by a dual role of IFN-I signaling in the protection of the gut epithelium on the one hand and the promotion of tissue-destructive inflammation on the other.

Two very recent papers utilizing the T cell transfer model of colitis showed that IFN-I signaling is important in this model, too. In the colitogenic CD45RBhigh cells it leads to CD69 induction which decreases their efficacy. If suppressive CD45RBlow T cells are transferred along with the CD45RBhigh effector T cells, IFN-I signaling is important for the maintenance of Foxp3 expression and thereby their disease suppressing potential. IFN-I signaling as a regulatory mechanism for suppressor T cell activity adds yet another component to the multitude of pro- and antiinflammatory mechanisms under IFN/STAT control that requires future attention.

IFN/STAT in the Regulation of Autoimmunity-Related Inflammation

This topic has been subject to many previous reviews,, and will be covered very briefly. Autoimmune syndromes such as multiple sclerosis, rheumatoid arthritis or the systemic lupus erythematosus (SLE) are characterized by local or systemic inflammatory episodes. The participation of IFN in some of these autoinflammatory syndromes was recognized with the finding that patients suffering from SLE have elevated plasma IFN-I levels and that their blood leukocytes display gene signatures with a prominent fraction of IFN-induced genes., Recognition of self nucleic acids from necrotic cells by endosomal toll-like receptors is widely accepted as a mechanism inducing IFN-I synthesis by plasmacytoid dendritic cells (pDC) present in inflamed organs. It is enhanced by the presence of autoantibodies to nucleic acids (reviewed in ref. 116). Consistently, IFN-I accelerate the SLE syndrome of lupus-prone mouse strains and Ifnar1 gene deletion is protective in such animals. Anti-IFN-I therapy is considered as a promising therapeutic strategy in humans. STAT1 is activated in cells from lupus patients and although there is no genetic evidence for its requirement, the STAT1 target gene signature provides a strong indication of its contribution to SLE-associated inflammation. The role of IFN/STAT is further emphasized by the demonstration that IFN-I as well as STAT1 and IRF9 enhance plasma cell differentiation and autoantibody production,, suggesting activity of the ISGF3 complex in the differentiation and autoantibody production by SLE patient B cells. An IFN-I contribution sharing similarities with that in SLE patients was shown for human psoriasis. Skin lesions display gene signatures resulting from IFN-I production by infiltrating pDC. Association of the cationic antimicrobial peptide LL37 with self nucleic acids is thought to facilitate their transport and association with endosomal TLR.

The proinflammatory activities of IFN-I in SLE, psoriasis and other autoinflammatory syndromes contrast with the anti-inflammatory properties of IFN-I in at least some neurodegenerative disorders. Most prominently, patients afflicted with multiple sclerosis (MS) benefit from IFN-β therapy. An animal model recapitulating some of the properties of MS is experimental autoimmune encephalitis (EAE). In experimental animals inflammatory neurodegeneration is caused by immunization with myelin basic protein. The disease shows a strong involvement of TH1 cells, as mice deficient for the TH1 fate-determining transcription factor T-bet are highly resistant. Conditional ablation of the IFNΑR on myeloid cells strongly increased EAE pathology, demonstrating that the IFN-I response of myeloid cells exerts a protective effect.

Both IFN-γ−/− and STAT1−/− mice are highly susceptible to EAE., This suggests that IFN-γ, like IFN-I, protects against EAE and that STAT1 is involved in signaling an anti-inflammatory response to both IFN types. The lack of STAT1 did not alter suppressor T cell abundance or function, but TH1 cells were prominently produced. In the study evaluating the effects of IFN-I, Prinz et al. (2008) did not find a change in the composition of helper T cell populations. In contrast, another study of EAE reported IFN-I-mediated suppression of proinflammatory TH17 cells. The potential of STAT1 signaling to alter TH differentiation is further emphasized by a recent study of human STAT1 mutations. Unlike all other STAT1 mutations found in the human population, which cause loss-of-function-associated immune defects, the 12 patients described in this paper suffer from chronic mucocutaneous candidiasis (CMCD) resulting from a STAT1 gain-of-function. The mutations are localized to the N-terminal coiled coil domain and lead to a stronger activation of STAT1 by IFN, but also by cytokines usually activating other STATs, such as IL-6. This causes a suppression of TH17 activity, thus depriving affected subjects from a major effector system against fungal pathogens such as C. albicans.

Collectively the studies addressing the role of IFN/STAT in autoimmune-related inflammation emphasize their profound regulatory input into complex inflammatory scenarios.

Concluding Remarks

We have summarized considerable but by no means all evidence documenting that IFN/STAT exert control over important aspects of inflammation reaching from leukocyte migration and tissue invasion to their activation and effector functions (Fig. 1). Beyond the innate response, inflammation promoted by TH subsets or its suppression by Treg is under IFN/STAT control, although there is still little understanding of the importance of this for inflammatory disease or the positive impact of inflammation on the clearance of infection. Particularly the IFN-I/STAT system is linked to a number of autoinflammatory syndromes and can act both as a driver or suppressor of inflammation-related tissue damage. Exploring the factors determining this yin-yang character as well as fathoming the therapeutic potential of IFN-I inhibition at the potential cost of losing antiviral immunity appears of utmost importance.

Figure 1. The control of Inflammation by interferons and their STATs. Arrows indicate whether a particular event is increased by IFN, decreased or whether either can occur in different inflammatory diseases. The IFN-γ receptor operates by activating a STAT1 dimer, whereas both the type I IFN (IFN-α/β) and IFN-λ receptors operate through the ISGF3 (STAT1/STAT2/IRF9) complex.