In search of a function for human type III IFNs: insights from inherited and acquired deficits

Abstract

The essential and redundant functions of human type I and II IFNs have been delineated over the last three decades by studies of patients with inborn errors of immunity or their autoimmune phenocopies, but much less is known about type III IFNs. Patients with cells that do not respond to type III IFNs due to inherited IL10RB deficiency display no overt viral disease, and their inflammatory disease phenotypes can be explained by defective signaling via another IL10RB-dependent pathways. Moreover, patients with inherited deficiencies of ISGF-3 (STAT1, STAT2, IRF9) present viral diseases also seen in patients with inherited deficiencies of the type I IFN receptor (IFNAR1/2). Finally, patients with auto-antibodies neutralizing type III IFNs have no obvious predisposition to viral disease. Current findings thus suggest that type III IFNs are largely redundant in humans. The essential functions of human type III IFNs, particularly in antiviral defenses, remain to be discovered.

Introduction

Almost 70 years after the discovery of type I interferons (IFNs) in 1957 by Isaacs and Lindenmann ^{[1, 2]}, and type II IFN in 1965 by Wheelock ^[3], type III IFNs (IFN-λ1, IFN-λ2 and IFN-λ3 in 2003 and IFN-λ4 in 2013) became the latest type of this fascinating family of secreted molecules named on the basis of their antiviral function to be discovered ^[4–6]. Human IFN-λ1, IFN-λ2, and IFN-λ3 were previously known as IL-29, IL-28A and IL-28B, respectively, perhaps due to the low level of similarity of their sequences to those of both IL-10 and type I IFNs (~11–13% amino-acid identity for IL-10, 15–19% amino-acid identity for IFN-α) ^[5]. The receptor for type III IFNs was identified at the same time as IFN-λ1–3 ^{[4, 7]}. The amino-acid sequences of human IFN-λ1, IFN-λ2, and IFN-λ3 are 81~96% identical. IFN-λ4 is most closely related to IFN-λ3, but these two molecules display only ~30% amino-acid sequence identity ^{[8, 9]}. Consistently, IFN-λ1, IFN-λ2, and IFN-λ3 have similar antiviral functions via the induction of interferon stimulated genes (ISGs), with IFN-λ1 the most potent ^[5]. IFN-λ4 also has more rapid antiviral activity than the other three IFN-λs, and several studies have suggested this rapidity may be important for the downregulation of type I IFN responses ^[9]. Interestingly, the type III IFN receptor has one subunit, IL10RB, in common with the receptor for IL-10 family cytokines, whereas the other subunit, IFNLR1, is specific to the receptor for type III IFNs ^[10]. In humans, the genomic structure of the genes encoding the three types of IFNs and their receptors further attests to their connection with the IL-10 cytokine family (Figure 1). The genes encoding type I and III IFNs form two separate clusters on chromosomes 9 and 19, whereas the gene encoding the single type II IFN, IFN-γ, clusters with those encoding IL-26 and IL-22 on chromosome 12. The genes for the two subunits of the type III IFN receptor are located close to the genes for the type I and II IFN and IL-10 family cytokine receptors on chromosome 21 (Figure 1) ^[11]. The receptors for type I, II, and III IFNs, respectively IFNAR1/2, IFNGR1/2, and IFNLR1/IL10RB, are members of the class II cytokine receptors ^[12].

Figure 1. Genomic structure of the genes for IFNs, IL-10 family cytokines and their receptors.

Diagram of the genomic loci encoding type I (red), II (purple), and III (blue) IFNs, IL-10 family cytokines (green), and their receptors. Chr: chromosome.

Like type I IFNs, human type III IFNs are thought to have antiviral activity. Both type I and III IFNs are induced by viral infection at least through Toll-like receptors (TLRs), retinoic acid-inducible gene 1 (RIG-1) -like receptors (RLRs), and cGAS-STING ^[13]. Both are monomers that bind to their receptor to activate the ISGF-3 complex (STAT1-STAT2-IRF9) via JAK-STAT pathways ^[10]. Both trigger the expression of largely overlapping ISGs (Figure 2) ^[10]. And both have antiviral activity in vitro ^{[10, 14]}. What differentiates type I IFNs from type III IFNs is their site of action, determined essentially by the expression patterns of their receptors. Type I IFN receptors are widely expressed, whereas type III IFN receptors are restricted to anatomic barriers, including the respiratory epithelium, the fenestrated endothelium of the liver, the gut epithelium, the skin epithelium, and the endothelial junctions of the blood-brain barrier, as well as hematopoietic cells including B lymphocytes, plasmacytoid dendritic cells (pDCs), monocytes, and neutrophils ^{[10, 14–16]}. However, studies of inborn errors and their phenocopies performed since 2003 have shown that human type I IFNs are essential for protective immunity to certain viruses but surprisingly redundant against many more viruses ^[17], whereas the essential and nonredundant functions of human type III IFNs remain unknown. This is an important issue, as studies of inborn errors of type II IFN and its autoimmune phenocopy have shown that type II IFN is much more a macrophage-activating factor than an antiviral IFN ^[18]. We provide here an overview of what is currently known about human type III IFNs, focusing in particular on what is known about their nonredundant functions. Mouse type III IFNs have been reviewed elsewhere ^{[10, 13, 14]}.

Figure 2. Signaling pathways for type I IFN, type III IFN, and IL-10 cytokines.

Signaling pathways of type I IFNs (IFN-α, -β, -ω, -ε, and -κ), type III IFN (IFN-λ), and IL-10 family cytokines (IL-10, -22, and -26).

Human type III IFNs are related to type I IFNs through their induction pathways

During viral infections, type III IFNs can be induced by microbial sensors, including at least TLRs, RLRs, and cGAS-STING ^[19]. The activation of appropriate sensors triggers a signaling cascade involving the recruitment of adapter proteins, such as MyD88 and TRIF for the TLRs, and MAVS for the RLRs ^[20–22]. The adapter proteins activate a series of downstream proteins and transcription factors, including IRF3, IRF7, and NF-кB. Activated IRF3/7 undergo homodimerization and translocation to the nucleus, where they elicit the expression of IFN genes, subsequently inducing the autocrine production and secretion of both type I and III IFNs (Figure 2) ^[23]. Some microbial products or cellular by-products preferentially induce type III IFNs in specific cell types. For example, cell surface TLRs preferentially induce type III IFNs ^[24], whereas peroxisomal MAVS signaling preferentially induce type III IFNs in intestinal epithelial cells ^[21]. However, most viral infections trigger the production of both type I and III IFNs, as multiple sensors are activated by different products. Infected cells in vitro produce similar amounts of type I and III IFNs, with some cells producing larger amounts of type III IFNs than of type I IFNs ^[25–29]. Many, if not all, cells, including tissue-resident cells and leukocytes, can produce type I and III IFNs, both constitutively in basal conditions and following stimulation with various synthetic agonists mimicking viral infection-related products or with the viruses themselves ^{[10, 14, 26–29]}. In addition, the different subtypes of type I IFNs are produced preferentially by different cell types. For example, IFN-β was initially identified as the ‘fibroblastic’ IFN as it is the only type I IFN produced by human fibroblasts and hPSC-derived neurons, the 13 IFN-α are produced by leukocytes, IFN-ε in the reproductive tract ^[30], and IFN-κ is produced by keratinocytes ^[31]. By contrast, type III IFNs are produced by all tissue-resident cells and leukocytes tested ^{[25, 28]}. The production of type III IFNs by such a wide range of cells raised an intriguing question as to whether these IFNs might have far-reaching biological effects throughout the human body, albeit by localized actions given the more restricted pattern of expression of their receptor.

Type III IFNs are related to IL-10 cytokines through their receptors

The type III IFN receptor is a heterodimer of IFNLR1 and IL10RB ^{[4, 5]}. The latter is also a subunit of receptors for IL-10 (with IL10RA), IL-22 (with IL22RA1), and IL-26 (with IL20RA, which is also the receptor for IL-19, together with IL20RB) ^[32]. IL-10 family cytokine receptors display low levels of sequence similarity, but their structure is conserved ^[33–35]. The ligand/receptor interaction is a two-step process in which the cytokines first bind to their specific high-affinity chain (IFNLR1, IL10RA, IL22RA1, IL20RA). This binding leads to a change in conformation, promoting binding of the common, low-affinity chain, IL10RB ^[36]. The ligand-receptor interaction triggers the activation of JAK1 and TYK2, constitutively associated with IFNLR1 and IL10RB, respectively, leading to the heterodimerization of STAT1/STAT2, which, together with IRF9, form the ISGF3 transcription complex ^[4]. Unlike type I IFNs, type III IFNs can also phosphorylate JAK2 in addition to JAK1 and TYK2, and this phosphorylation appears to trigger a different downstream signaling pathway ^[21]. The specificity of the biological activities of the various cytokines of the IL-10 family is determined by the restricted expression of the high-affinity binding chain, the low-affinity chain, IL10RB, being much more widely expressed ^[37]. The restricted expression of IFNRL1 renders IFN-λ promising antiviral drug, as fewer side effects would be expected than with type I IFNs, which have wide-ranging effects in most, if not all cell types, due to the universally expressed type I IFN receptors ^[10].

Human type III IFNs are related to type I IFNs through their antiviral functions

Type III IFNs are produced by a wide range of cells, but their activity is limited to epithelial and endothelial cells expressing IFNLR1. Furthermore, type I IFNs, especially IFN-β, are much more potent than type III IFNs in terms of their activity in vitro ^[38–40]. Type I IFNs therefore have a stronger and broader impact than type III IFNs. Despite signaling through a unique heterodimeric receptor complex consisting of IFNGR1 (IL28Rα) and IL10RB, type III IFNs activate the ISGF3 complex through a JAK-STAT pathway remarkably similar to that used by type I IFNs, driving the expression of overlapping ISGs ^[10]. The TLR3-dependent basal levels of IFN-β and IFN-λ present before viral infection may provide protection against the first wave of viral invasion in infected tissues ^[28]. The type I and III IFN signaling pathways differ in terms of kinetics, bandwidth, and the responding cells ^[10]. Type I and III IFNs induce similar sets of ISGs in human epithelial cells, but ISG expression is generally weaker and occurs later after treatment with type III IFNs than with type I IFNs ^{[38, 39]}, although it is unclear whether these differences are related to differences in the IFN concentrations tested. Despite their lower potency, type III IFNs can induce well-known antiviral ISGs that are also induced by type I IFNs. Treatment with type III IFNs mainly IFN-λ1 or IFN-λ3 inhibits the replication of at least dengue virus 2, sindbis virus, vesicular stomatitis virus, encephalomyocarditis virus, influenza A virus, and coronaviruses, including SARS-CoV-2 in some human epithelial cells ^{[4, 5, 26, 41, 42]}. All leukocytes respond to type I IFNs, which induce systemic antiviral and inflammatory responses that are thought to be characteristic. However, recent studies have shown that human peripheral monocytes, pDCs, and B cells might also respond to type III IFNs ^{[16, 43]}. Nevertheless, the biological functions of these molecules in humans remain to be established ^{[16, 43]}. Except for these peripheral hematopoientic cells, the type III IFN receptors are mostly restricted to endothelial and epithelial barriers, suggesting that type III IFNs may serve as a frontline defense limiting viral infection locally without triggering a more potent systemic type I IFN response ^[10].

Population genetics of the four type III IFNs and their two receptor chains

Population genetic studies can suggest genes that are essential (deleterious alleles under negative/purifying selection), redundant (deleterious alleles under relaxed natural selection), or adaptable (advantageous alleles under positive selection), even without knowledge of the biological functions concerned ^[44]. Manry et al. showed in a population genetic study of IFNs that the single type II IFN, IFN-γ, is under strong negative/purifying selection, whereas the type of selection acting on 17 type I IFN loci varied from locus to locus. No negative selection was detected for type III IFNs and their receptors (Table 1) ^[45]. This study also found no evidence of strong negative/purifying selection on the four subunits of the type I and II IFN receptors (IFNAR1, IFNAR2, IFNGR1, and IFNGR2). More recently, Rapaport et al. developed a scoring system, the CoNeS score, for evaluating the strength of negative selection on human genes through comparisons with known inborn errors of immunity (IEIs) ^[46]. They found evidence for strong negative selection on IFNG, but not on IFNL1–3 (Table 1) ^[46]. However, neither study took into account the fact that vast majority of missense variants are untested and probably functionally neutral, regardless of in silico predictions ^[47–50]. Scores based exclusively on predicted loss-of-expression variants might provide a simple alternative to this problem. For example, pLI score, although not available for most type I IFNs, indicates that type II IFN and its receptor genes are under strong negative selection, whereas type III IFNs and their receptor genes are under relaxed negative selection. Nevertheless, homozygous carriers of predicted loss-of-expression variants are present in the gnomAD database for IFNL4, but not for IFNL1–3 or their receptors, consistent with the hypothesis that IFNL1–3 are under moderate negative/purifying selection (Table 1).

Table 1.

Population genetics of IFNs

Type of IFN	IFN	pLI	CoNes	Natural selection [45]	High MAF pLOF variants (at least one homozygous individual in gnomAD v4.0.0)
Type I	IFNA1		−0.145448
	IFNA2		1.18740137
	IFNA4		1.56400038
	IFNA5		1.14489871		p.Gln115Ter (0.005510)
	IFNA6		0.6863197	Negative	p.Gln85Ter (0.00008116)
	IFNA7		1.72208985		p.Arg56Ter (0.0002733)
					p.Asn34Ter (0.000007437)
	IFNA8		0.38405775	Negative
	IFNA10		2.97343741		p.Cys20Ter (0.2134)
					p.Gln125Ter (0.001563)
					p.Gln85Ter (0.0005082)
	IFNA13		−0.0225397	Negative
	IFNA14		1.33816144	Negative
	IFNA16		1.74857971
	IFNA17		1.03545511		p.His57GlnfsTer2 (0.03362)
	IFNA21		2.20734529		p.Glu126Ter (0.0002546)
	IFNB1		1.28478561		p.Arg186Ter (0.001260)
	IFNE	0.014672	0.42632874		p.Gln71Ter (0.02178)
					p.Met84IlefsTer53 (0.00004226)
	IFNK	9.0615E-10	1.32160607		p.Trp13PhefsTer4 (0.02996)
					p.Gln70Ter (0.00006382)
					p.Ala86ProfsTer31 (0.000002052)
	IFNW1		0.72036958
	IFNAR1	3.2501E-12	0.411163583
	IFNAR2	8.5332E-06	1.503264334		p.Tyr322Ter (0.001360)*
Type II	IFNG	0.47156	−0.775619	Negative
	IFNGR1	0.015503	−0.063432717
	IFNGR2	0.95281	−0.330452454		p.Ter61LysfsTer7 (0.00004247)*
Type III	IFNL1	0.00028469	0.74760998	Positive in Eurasians
	IFNL2	4.9903E-08	0.96851769	Positive in Asians	p.Arg188Ter (0.0003674)
	IFNL3	2.4233E-08	2.61624542	Positive
	IFNL4				p.Gly23AlafsTer158 (0.3176)
					p.Ser151Ter (0.0001770)
					c.443-2A>G (0.0004238)
					c.151+1G>A (0.000005684)
	IFNLR1	0.062532	0.117759003
	IL10RB	2.5103E-06	0.462201582		p.Gln321GlyfsTer60 (0.00005887)
					p.Gln321HisfsTer7 (0.00005893)
					c.50-2A>T (0.00001314)

^*pLOF only in minor transcripts that are unlikely to be protein-coding, while intronic in MANE transcripts.

A human inherited disorder of type III IFNs: IL10RB deficiency

The study of inborn errors of immunity can determine the contribution of cells and/or proteins to protective immunity ^{[51, 52]}. No patients with IEIs of IFNLR1 or IFNL1–3 have ever been reported, but many other IEIs can disrupt type III IFN production or activity, in addition to their effects on other cytokines, potentially contributing to the phenotypes of the affected patients (Table 2). Autosomal recessive IL10RB deficiency (complete or partial) is the only such IEI in which type I IFN immunity is intact, make it the only known IEI capable of revealing the antiviral functions of type III IFNs independently of type I IFNs. IL-10, IL-22, and IL-26 signaling are impaired in patients with this deficiency. The major phenotype of IL10RB deficiency is early-onset inflammatory bowel disease (EOIBD), as in patients with autosomal recessive IL-10 or IL-10RA deficiency ^[53]. Bone marrow transplantation leads to clinical improvement in IL10RB-deficient patients, but without the rescue of type III IFN immunity in nonhematopoietic cells, suggesting that the principal clinical phenotypes of IL10RB deficiency result from defective IL-10 signaling in leukocytes ^[53]. Intriguingly, patients with IL10RB deficiency do not generally seem to be susceptible to viral infections either before or after HSCT, with the exception of two siblings carrying biallelic mutations (tryptophan to glycine at amino acid position 100, Trp100Gly) who suffered from fulminant viral hepatitis (FVH) ^[54]. However, this FVH was not caused by defective type III IFN signaling, as this signaling was intact in these patients whose IL-10RB deficiency is partial, as opposed to complete. Instead, it was probably due to defective IL-10 signaling leading to excessive IFN-γ activity ^[54]. The studies of human patients with inherited IL-10RB deficiency performed to date suggest that type III IFNs may be largely redundant.

Table 2.

IEIs disrupting type III IFN responses

IEI	Type III IFN	Type I IFN	Other cytokine pathways disrupted	Phenotype	References
IL10RB	Defective	Intact	IL-10, IL22, IL-26	IBD, fulminant HAV hepatitis	[54, 137]
JAK1	Defective	Defective	IFN-γ, IL-2, IL-4, IL-7, IL-9, IL-15, IL-21, IL-27, IL-6 family cytokines, IL-10 family cytokines	Mycobacterial infections	[57]
TYK2	Defective	Defective	IL-12, IL-23	Viral (including cutaneous Molluscum contagiosum, HSV, VZV; PIV3 pneumonia, EBV associated B cell lymphoma), fungal, and mycobacterial infections	[61, 62, 138–141]
STAT1	Defective	Defective	IFN-γ	Viral (including HSE, fulminant EBV infection, CMV, MMR encephalitis, HHV6, disseminated chicken pox) and mycobacterial infections	[56, 57, 60, 142–147]
STAT2	Defective	Defective		LAV infection (MMR), cutaneous HSV, VZV, EBV, enterovirus, adenovirus, and influenza pneumonia	[148–150]
IRF9	Defective	Defective		Influenza and COVID-19 pneumonia, biliary perforation after MMR	[151, 152]
TLR3	Defective	Defective		HSE, Enterovirus encephalitis, influenza pneumonia, COVID-19 pneumonia	[28, 40, 153, 154]
UNC93B1	Defective	Defective		HSE and COVID-19 pneumonia	[49, 155, 156]
TBK1	Defective	Defective		HSE and COVID-19 pneumonia	[156, 157]
TICAM1	Defective	Defective		HSE and COVID-19 pneumonia	[156, 158]
TRAF3	Defective	Defective		HSE	[159]
IRF3	Defective	Defective		HSE and COVID-19 pneumonia	[156, 160]
IRF7	Defective	Defective		Influenza, RSV, adenovirus, and COVID-19 pneumonia	[26, 156, 161]
TLR7	Defective	Defective		COVID-19 pneumonia	[49]
MDA5	Defective	Defective		Enterovirus encephalitis, Rhinovirus pneumonia	[29, 50, 162]
IRAK4/MYD88	Defective	Defective		Bacterial infections and COVID-19 pneumonia	[163]
NEMO	Defective	Defective	IFN-γ	Viral and mycobacterial infections	[164]
APCED/APS-1	Defective	Defective	IL-17	Fungal infections and COVID-19 pneumonia	[84, 92, 165, 166]

Human inherited disorders of type I and III IFNs

Many IEIs disrupt both type I and III IFN signaling, with or without signaling defects for other cytokines (Table 2). These IEIs include IRF7, STAT2 and IRF9 deficiencies (defective type I IFN signaling) ^[55], STAT1 deficiency (defective type I, II, and III IFN signaling) ^[56], JAK1 deficiency (defective type I and II IFN, IL-2, IL-4, IL-7, IL-9, IL-15, IL-21, IL-27, IL-6 family and IL-10 family signaling) ^{[57, 58]}, TYK2 deficiency (defective type I IFN, IL-12, and IL-23 signaling) ^[59], deficiencies of TLR3 and its downstream pathway (defective type I and III IFN induction) ^[28], deficiencies of TLR7 and its downstream pathway (defective type I and III IFN induction) ^[49], and MDA5 deficiency (defective type I and type III IFN induction) ^[29]. However, the response to type III IFNs has not been rigorously tested for all these IEIs and the cells affected may differ. For example, for complete STAT1 deficiency, there is no response to type III IFNs in EBV-transformed B cells, SV40-transformed fibroblasts transduced with IFNLR1, and human pluripotent stem cell-derived lung epithelium cells ^[60]. By contrast, for complete TYK2 deficiency, SV40-transformed fibroblasts have very weak responses, whereas EBV-B cells respond normally ^{[54, 61]}. The response to IFN-λ1 in HAP1-TYK2^KO cells is normal, suggesting that type III IFN signaling may be TYK2-independent, at least in some cell types ^[62]. Globally, the set of viral phenotypes seen in patients with these IEIs is apparently indistinguishable from that seen in patients with inherited IFNAR1 or IFNAR2 deficiency, in which type I IFN immunity is specifically disrupted. Studies of these IEIs do not, therefore, point to essential roles of human type III IFN against particular viral infections. Admittedly, as with IL-10RB deficiency, only a few patients have been diagnosed with these IEIs. It is, therefore, important to diagnose more patients, to define the range of viruses that can be controlled in the absence of type III IFNs more precisely, perhaps with the identification of certain viruses that cannot be controlled in these conditions.

Common variants of type III IFNs associated with HCV susceptibility

In 2009, genome-wide association studies (GWAS) connected single-nucleotide polymorphisms (SNPs) upstream from IFNL3, rs12979860 and rs8099917 with the clinical course of HCV infection, including spontaneous clearance, and HCV clearance induced by peginterferon IFN-λ plus ribavirin bitherapy or peginterferon IFN-λ plus ribavirin plus first-generation direct-acting antiviral (DAA) tritherapy ^[63–76]. In 2013, further investigations of these associations led to the discovery of a dinucleotide variant, rs368234815 (TT or ΔG, comprised of rs11322783 and rs74597329 which are in full linkage disequilibrium), with the unfavorable rs368234815-ΔG allele creating a new interferon gene (IFNL4) between IFNL3 and IFNL2 (Figure 1) ^{[77, 78]}. The expression of IFN-λ4 is regulated by this common variant and associated with a less favorable outcome of HCV infection ^{[42, 77–81]}. The amino-acid sequence of IFN-λ4 is 30% identical to that of IFN-λ3 and these two molecules have antiviral properties similar to but weaker and faster than those of other type III IFNs in vitro ^{[42, 77–81]}. The IFNL4-encoding allele is the major allele in African (~70%) populations, but is a minor allele in European (~30%) and East Asian (~10%) populations ^[81]. IFN- λ3 residues under positive selection, such as the arginine in position 70, have been validated in epidemiological studies as playing a key role in the spontaneous clearance of HCV ^[45]. There are also other nonsynonymous common variants on the haplotypes carrying IFNL4 that might contribute to phenotypes studied. The biochemically favorable (due to antiviral activity) but clinically unfavorable functions of IFNL4 is a paradox that has yet to be solved. Several hypotheses have been proposed, including a role for IFN-λ4 in downregulating type I IFN signaling ^[82], poor induction by viral infection due to the weakness of the IFNL4 promotor, which has no IRF binding site ^[83], and toxicity due to the accumulation of IFN-λ4 in the ER ^[79]. This paradox adds another layer of complexity to the question of the possible redundancy of type III IFNs in humans.

Auto-Abs against type III IFNs do not underlie a clear viral phenotype

Autoantibodies (auto-Abs) against type III IFNs were first discovered in patients with autoimmunity, including autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED, also known as autoimmune polyendocrinopathy syndrome type 1, APS-I) ^{[84, 85]}, thymoma ^[86], and systemic lupus erythematosus (SLE) ^[87]. Interest in auto-Abs against type I IFNs has increased since 2020, due to their presence in ~15% of patients with critical COVID-19 pneumonia, ~20% of patients dying from COVID-19 ^[88–119], ~5% of patients with severe influenza pneumonia ^[120], three in eight cases of severe infections due to live attenuated yellow fever vaccine ^[121], cases of herpes zoster infection (particularly in patients with COVID-19) ^{[107, 122]}, and 40% of cases of West Nile virus encephalitis ^[123]. Auto-Abs neutralizing type I IFNs are surprisingly common in the general population, being found in <1% in individuals under the age of 70 years but more than 6% of those over the age of 70 years ^[89]. However, the contribution of auto-Abs against type III IFNs has been little studied in the context of viral or other conditions. We recently investigated the prevalence of auto-Abs against type III IFNs in the general population and in COVID-19 patients, with a luciferase-based immunoprecipitation assay using an IFN-λ1–3 mixture as the antigen ^[124]. About 0.6% of the 1,500 SARS-CoV-2-naïve individuals tested had auto-Abs neutralizing all three IFN-λ subtypes tested, a prevalence similar to that for auto-Abs neutralizing type I IFNs ^[89]. Most of the individuals with these antibodies were women over the age of 70 years, but one 13-year-old girl also tested positive. However, no evidence was found of enrichment in auto-Abs neutralizing any of the three type III IFNs tested in patients with severe COVID-19 pneumonia (3.0%) relative to those with mild COVID-19 (3.8%). Auto-Abs neutralizing all three IFN-λ subtypes tested were found in five patients with severe COVID-19 (0.5%) and in none of the patients with mild COVID-19 tested. However, four of the five patients with these auto-Abs also had auto-Abs neutralizing type I IFNs. It is unknown whether auto-Abs against type III IFNs neutralize the antiviral effect of IFN-λ in lung epithelial cells following infection with SARS-CoV-2. The prevalence and role of auto-Abs against type III IFNs in other viral diseases have not been investigated.

Treatment of viral infections with peginterferon IFN-λ

Peginterferon lambda-1a (PegIFN-λ1a) was initially developed for the treatment of chronic hepatitis C virus (HCV) infections and clinical trials of this drug alone or in combination with antiviral drugs have shown it to be superior to type I IFN treatment ^{[125, 126]}. Its favorable safety profile encouraged its inclusion in several trials of possible treatments for SARS-CoV-2 infection. The results of two phase II studies characterizing the effect of PegIFN-λ1a on SARS-CoV-2 viral load have been published ^{[127, 128]}. One of these trials, a randomized, single-blind placebo-controlled trial including 120 unvaccinated participants receiving subcutaneous injections of 180 μg PegIFN-λ1a within 72 hours of diagnosis, found no significant clinical improvement or shortening of the period of viral shedding in COVID-19 outpatients ^[128]. The other randomized, double-blind controlled trial, on 60 unvaccinated outpatients, found that PegIFN-λ1a accelerated viral clearance, as shown by analyses performed seven days after the subcutaneous injection of 180 μg PegIFN-λ1a in patients with a high baseline viral load (>10⁶ copies of RNA per ml) ^[127]. A recent large, phase III, randomized, placebo-controlled study recruited almost 2000 outpatients with COVID-19, 83% of whom were vaccinated ^[129]. The incidence of hospitalization or emergency department visits was significantly lower among those who received a single dose of PegIFN-λ1a than among those who received placebo. PegIFN- λ1a appears to be a promising drug in terms of its antiviral effects with excellent safety profile. However, the prevalence of auto-Abs against IFN-λ, especially for IFN-λ1, is a particular concern for the development of this molecule as a drug, analogous to the problem represented by anti-type I IFN antibodies in the context of treatment with IFN-α or IFN-β.

Conclusions and perspectives

The studies performed to date suggest that human type III IFNs may be redundant in host defense. However, the search continues for patients with inherited or acquired deficiencies of type III IFNs and a clinical phenotype, particularly for viral infection phenotypes. Indeed, this apparent redundancy is surprising in light of evolution, as type III IFNs have been subject to significant positive selection during the course of human evolution ^[45]. Over a longer evolutionary time scale, type III IFNs appeared early in vertebrate evolution. They have been found in all tetrapods tested, including mammals, birds, reptiles, and amphibians ^{[4–6, 130–132]}. Most type III IFN genes retain an exon-intron structure, like other IL-10 family cytokine genes, but some amphibians and mammals have intron-less IFN-λs genes more closely resembling the genes encoding type I IFNs, possibly due to the retrotransposition of intron-containing type III IFNs ^{[132, 133]}, suggesting a need for the rapid translation of both type I and III IFNs. Type III IFNs have not been detected in fishes, but an IFNLR1 gene has been identified in many fish genomes, and sharks and humans display a high degree of synteny ^{[134, 135]}. There is also other evidence suggesting that, like type I IFNs, type III IFNs may have originated in the earliest jawed vertebrates. IFNR-like genes for both type I and type III IFNs have been predicted in silico in the genomes of jawless fish, the most primitive vertebrates, which separated from other vertebrates 500 million years ago ^[136] (Feng, et al., in prep). Thus, both type I and type III IFNs may have a very early common ancestor, in the last common ancestor of all vertebrates. Studying IFNs in vertebrate groups that diverged early in evolution may provide us with insight into the co-evolution of virus and host. Together with the search for inborn errors of type III IFN immunity and their autoimmune phenocopies in human patients, particularly those suffering from viral diseases affecting the corresponding tissues, and in the general population, this should help to clarify the redundant and essential roles of the mysterious type III IFNs.

Abbreviations

IFN

Interferon

ISGF-3

Interferon-stimulated gene factor 3

ISGs

interferon-stimulated genes

TLRs

Toll-like receptors

RIG-1

retinoic acid-inducible gene 1

RLRs

retinoic acid-like receptors

IEI

inborn error of immunity

JAK-STAT

janus kinase-signal transducers and activators of transcription

pDC

plasmacytoid dendritic cell

DC

dendritic cell

EOIBD

early-onset inflammatory bowel disease

HSCT

haematopoietic stem cell transplantation

FVH

fulminant viral hepatitis

EBV

Epstein–Barr virus

SV40

polyomavirus simian virus 40

GWAS

genome-wide association studies

SNPs

single-nucleotide polymorphisms

DAA

direct-acting antiviral

HCV

hepatitis C virus

ER

Endoplasmic reticulum

IRF

interferon regulatory factor

APECED, also known as autoimmune polyendocrinopathy syndrome type 1, APS-I

autoimmune polyendocrinopathy candidiasis ectodermal dystrophy

SLE

systemic lupus erythematosus

COVID-19

coronavirus disease 2019

PegIFN-λ1a

Peginterferon lambda-1a