The role of interferon in the thymus

Abstract

Interferons (IFNs) are a family of proteins that are generated in response to viral infection and induce an antiviral response in many cell types. The COVID-19 pandemic revealed that patients with inborn errors of type I IFN immunity were more prone to severe infections, but also found that many patients with severe COVID-19 had anti-IFN autoantibodies (AAB) that led to acquired defects in type I IFN immunity. These findings revealed the previously unappreciated finding that central immune tolerance to IFN is essential to immune health. Further evidence has also highlighted the importance of IFN within the thymus and its impact on T cell development. This review will highlight what is known of IFNs role in T cell development, T cell central tolerance, and the impact of IFN on the thymus.

Interferon production in the thymus

T cell central tolerance occurs within the thymus where developing thymocytes expressing newly generated T cell receptors (TCRs) are selected for potential functionality and deleted if they possess high self-reactivity [1]. While IFN is primarily thought of as a proinflammatory cytokine produced transiently following pathogen exposure, recent evidence has revealed IFNs are continuously expressed in the thymus [2,3]. IFNs can be categorized into three families: type I, type II and type III. In mouse, type I IFNs are composed of 13 IFN-a isoforms, IFN-b, IFN-ζ, IFN-e, and IFN-k. IFN-g is the single type II IFN, and IFN-l2 and 3 are in the type III IFN family [4]. Type I and III IFNs will be the focus of this review. Type I IFNs signal through the type I IFN receptor (IFNaR, IFNAR1/ IFNAR2 heterodimer), while type III IFNs bind to the type III IFN receptor (IFNlR, IFNLR1/IL-10R2 heterodimer). Type I and III IFN binding to their respective receptors results in the transcription of similar interferon stimulated genes (ISGs), though some differences in the induction of proinflammatory genes have been identified [5]. Characterization of the differences between type I and III IFN are still ongoing, but evidence suggests specialized and unique roles for each IFN type.

Within the thymus, both type I and III IFNs are predominantly produced by medullary thymic epithelial cells (mTECs) [2,3]. mTECs are unique because they are one of the few cell types that can express the gene encoding Autoimmune regulator (AIRE) [6]. Models show that AIRE promotes the expression of tissue restricted antigens (TRAs) in the thymus, thereby inducing clonal deletion or Foxp3+ regulatory T (Treg) cell induction in thymocytes with TCRs specific for TRAs [6–9]. Along with promoting TRA expression, AIRE also controls the maturation and differentiation of mTECs. LTβR, CD40 and RANK signaling activates the alternative NF-κB pathway to drive medullary thymic epithelial precursor cells (AIRE-) to mature into MHCII^hi AIRE+ mTEC (mTEC^hi) [10,11]. mTEC^hi cells will then further mature into post-AIRE mTEC, as identified by the reduction of MHCII and AIRE expression. Even though post-AIRE mTEC represent a terminal differentiation step of mTEC development [12], these cells are likely important in selection of developing thymocytes as a source of self-antigen, as recent works have shown the post-AIRE mTEC are composed of a heterogenous group of cells that mimic the gene signatures diverse extra-thymic cell types[13]. AIRE KO mice showed strong reductions in Ifnb and modest reduction in Ifnl2/3 gene expression in mTEC, suggesting IFN production is AIRE regulated [2,3]. However, as AIRE is important for mTEC^hi maturation, it is unclear if IFN production is an AIRE-driven TRA or is dependent on mTEC maturation. Re-analysis of mTEC lineage tracing transcriptomics data showed the highest expression of type I and III IFNs in mTEC^hi, with reduced expression in post-AIRE mTEC [10]. It has been hypothesized that inflammatory mediators, such as IFNs and TLR stimuli, are the exclusive product of post-AIRE cells [12]. Such hypotheses are interesting in the context of recent work demonstrating a portion of mTEC^hi and post-AIRE cells acquire the transcriptional profile and characteristics of peripheral cell types and can all produce IFNs [13]. Future studies will be required to precisely define the cells that produce IFNs in the thymus and identify what regulates IFN production by these cells.

Interferon impact on T cell selection

A growing body of work has demonstrated that thymic IFN has striking impacts on both thymocytes and antigen presenting cells. IFNaR is broadly expressed by cells in the thymus; both on developing thymocytes and antigen presenting cells (APCs) (Figure 1). Meanwhile, IFNlR is restricted to thymic epithelial cells and select APCs, including B cells and thymic DC. In thymocytes, thymic type I IFNs are involved in the late-stage maturation of single positive thymocytes though cell intrinsic effects of IFNaR signaling [14]. While thymocytes rely on type I IFN signaling, it was recently shown that type III IFN is required for thymic B cell licensing and optimal Treg cell selection, even though the B cells expressed IFNaR [15]. Similarly, mature XCR1+ dendritic cells (DC), a subset of APCs essential for central tolerance, strongly express ISGs in the thymus, but are largely unaffected by deficiency of IFNaR alone [16]. It will be critical to determine the impact of IFNlR and type III IFN on thymic XCR1+ maturation and self-antigen presentation. Lastly, type III IFN also controls MHC class I expression in cortical thymic epithelial cells (cTECs) and mTECs, suggesting it may impact positive as well as and negative selection [3]. Thus, IFN functions are important for generating a healthy T cell repertoire.

Figure 1. Thymic interferon drives thymocyte and thymic antigen presenting cell maturation.

Type I and III interferon (IFN) are produced at steady state by MHCII^hi AIRE+ medullary thymic epithelial cells (mTEC^hi) in a partially AIRE-dependent manner. Interferons (IFNs) have differential impact on cells within the thymus, likely due to receptor expression or IFN localization. The type I IFN receptor (IFNaR) is expressed by thymocytes, dendritic cells (DC) and B cells, and the type III IFN receptor (IFNlR) is expressed on DC and B cells [14–16]. CD4 single positive (SP) thymocytes and type III IFN drive thymic B cell licensing/activation (top panel), inducing isotype class switch recombination and generate a B cell that can select Foxp3+ regulatory T (Treg) cells. While IFNaR is expressed on thymic B cells, type I IFN does not impact B cell licensing or interferon stimulated gene (ISG) expression. Semi-mature thymocytes require type I IFN signaling for maintaining ISG expression associated with thymocyte maturation, with no intrinsic impact of type III IFN due to lack of IFNlR expression (right panel). XCR1+ DC maturation demonstrates high expression by ISGs but is not greatly impacted by loss of IFNaR (left panel). The role of IFNlR in DC maturation is still unknown.

There is emerging evidence the thymus is also required to maintain T cell tolerance against IFN itself. During the COVID-19 pandemic, approximately 20% of COVID-19 deaths could be attributed to anti-IFN autoantibodies (AAB) [17,18]. These autoantibodies were present in individuals prior to SARS-CoV-2 infection and could block type I IFN signaling [18]. The prevalence of anti-IFN AABs increases with age and were found in approximately 4% of healthy individuals over the age of 70 years old [18]. Consequently, immune tolerance to type I IFNs were more likely to be lost in older individuals, which was correlated with severe COVID-19 in these patients. Further investigation into patient populations with high prevalence of anti-IFN AAB has revealed that AIRE-mediated immune tolerance in the thymus is likely required to prevent anti-IFN AAB [19].Patients with autoimmune polyendocrinopathy ectodermal dystrophy (APECED), also known as autoimmune polyglandular syndrome type 1 (APS-1) have inherited mutations in AIRE that lead to reduced expression/function of the protein, resulting in changes to mTEC maturation, TRA expression and alterations in clonal deletion and Treg cell selection [20–22]. These patients also have a remarkably high prevalence of class switched, high affinity anti-IFN AAB [23,24] indicative of an adaptive immune response against type I IFNs. Identification of other patients with a high prevalence of anti-IFN AB further supports the idea that the thymus and mTECs are required for immune tolerance to IFN. This includes patients with defects in the alternative NF-KB pathway (NIK, RELB, specific NFKB2 deficiencies) [25,26], loss-of-function (LOF) NEMO mutations [17], LOF FOXP3 mutations [27] or hypomorphic RAG1/RAG2 mutations [28]. Defects in the alternative NF-kB pathway result in altered mTEC development and maturation [26,29–31], while male patients with LOF FOXP3 mutations have dysfunctional Treg cells and present with a condition known as immune dysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX)[27]. Patients with hypomorphic RAG mutations are hypothesized to have altered mTECs, as mature single positive thymocytes deliver CD40L and RANKL signals to mTEC for maturation and AIRE induction [28]. Thus, mTEC maturation with AIRE induction and Treg cell generation is required for immune tolerance to IFN [26] (Figure 2). Tolerance is not lost to all type I and III IFNs, but instead anti-IFN AAB are often found against IFN-a and IFN-w and less commonly identified against IFN-b,-k, or -e, which is puzzling as all IFNs are produced by mTEC [10,18]. The propensity to generate anti-IFN AAB against certain IFNs is not yet understood, but likely involves the process of how immune tolerance to IFNs is generated. This process is still undefined. The simplest possibility is that mTEC directly present IFN-derived peptides on MHCII induce central tolerance. In this situation, one would predict IFN-l, -b and -e expression would be the most likely target of autoimmunity, as these are most highly expressed IFN transcripts in mTEC and would be most impacted by AIRE deficiency/mTEC dysregulation[32]. However, this is not seen in patients, as antibodies against IFN-l, -b and -e are rare compared to antibodies against IFNs with lower transcript expression, such as IFN-a2. Instead, mTEC mediated IFN-tolerance is likely more complicated and studies will be required to understand the complex mechanism of autoimmune reactions against IFN.

Figure 2. Key players in the generation of central tolerance to interferons.

CD4 single positive (SP) thymocytes recognizing self-pMHCII complexes on MHCII^lo medullary thymic epithelial cells (mTEC^lo) precursors deliver signals to drive alternative NF-kB signaling (NIK, NFKB2, RELB) to induce AIRE and drive mTEC differentiation to MHCII^hi mTEC (mTEC^hi). mTEC^hi produce IFN and select for Foxp3+ regulatory (Treg) cells that can prevent B cell activation and generation of autoantibody responses against type I interferons (IFNs). Patients with loss-of-function mutations in the factors important in this process (highlighted in red) have a high incidence of autoantibodies against type I IFNs [20]. As both type I and type III IFNs are produced by mTEC and only type I IFN is targeted in autoimmunity, it is likely that loss of IFN expression in the thymus is not the only determinant controlling anti-IFN autoantibody production. We hypothesize chronic inflammation and type I IFN signaling in the peripheral tissues drives interferon stimulated genes (ISGs) and type I IFN production, which primes the anti-IFN immune response typically prevented by central tolerance.

Interferonopathies and the thymus

The importance of regulating type I IFN production and signaling can be appreciated in patients with inborn errors of type I IFN and interferonopathies (disorders of dysregulated IFN production and signaling). Patients with inborn errors of type I IFN are at risk for pathogen mediated pathology [19], while patients with interferonopathies show a range of symptoms associated with autoinflammation that include skin and CNS disease, lupus and developmental delay [33]. While the thymus has not been widely studied in patients with inborn errors of type I IFN or interferonopathies, dysregulation of IFN signaling in the thymus has been associated with myasthenia gravis (MG) and may represent a subtype of interferonopathy [34]. MG is a heterogeneous autoimmune disease mediated by autoantibodies targeting proteins of the neuromuscular junction, in particular the acetylcholine receptor (a-AChR) [34]. MG can be separated into those with early onset MG (EOMG), late onset MG (LOMG) and MG with thymoma (MGT). Thymoma is a neoplasm of the thymic epithelium defined by histological subtypes based on the location of neoplastic cells in the thymic cortex and/or medulla. In both EOMG and MGT, type I and III IFN have been shown to drive a-AChR expression in mTEC and are hypothesized to be important in initiating the process of generating anti-a-AChR antibodies [35,36]. Increased IFN production in the thymus of EOMG patients is associated with B cell infiltrates, ectopic germinal center formation and increased ISG expression, while increased type I IFN found in MGT is associated with a loss of AIRE expressing mTEC due to the underlying thymoma [35,37]. Interestingly, over half of MGT and thymoma patients have circulating anti-IFN AABs but these are not found in EOMG patients [38]. Thus, while IFNs may be over-produced in both MG subtypes, there are likely distinct pathophysiological processes underlying the diseases. In EOMG, thymic B cell activation and generation of neuromuscular autoantibodies are associated with increased IFN signals in the thymus. Recent work has found thymic B cells express Ifnlr1 and rely on type III IFN for thymic licensing and activation, but it is not known if excessive type III IFN alters thymic B cell activation and drives germinal center formation [15]. As there is loss of AIRE+ mTEC in MGT due to thymoma, it is not surprising that anti-IFN AAB would be formed in these cases. However, there is also an increase in thymic IFNs and ectopic B cell activation often found adjacent to thymoma, suggesting either the neoplastic cells or other cell types are producing IFN in this disease and may be driving anti-a-AChR immune responses [39]. Future studies will be required to discriminate the underlying mechanisms driving MGT and EOMG to understand the role IFNs play in each disease.

Tonic interferon impact on T cell reactivity

Type I IFN is known to impact T cell activation and differentiation during infection, but recent evidence has found that a subset of naïve CD8SP, CD4SP and Foxp3+ regulatory T cells express an interferon stimulated gene signature at steady state [40–44]. It is not yet clear if this reflects IFN signaling in the thymus or the periphery or both (type I and type III IFNs are produced in intestinal sites at steady state [2,45]). Nonetheless, the fact that IFN is impacting the transcriptome of the cells suggests it may also be impacting their function. In Treg cells, type I IFN signaling optimizes antiviral and antitumor immunity during immune responses, and may play a role in thymic Treg cell development [46,47]. Transcriptional studies of developing thymocytes have identified a subset of Treg cells that express ISGs, suggesting a potential role of type I IFN in a subset of developing Treg cells [42,43]. However, more study is needed to determine how intrinsic type I IFN signaling impacts Treg cell selection and development. In conventional thymocytes, naïve CD4SP thymocytes lacking Ifnar have lower levels of Stat1 and Ifr7 [14]. Furthermore, a subset of CD8SP thymocytes and naïve CD8 T cells expresses Ly6C and are dependent on Ifnar [41,44]. This Ly6C+ CD8 T cell subset showed increased sensitivity to low-affinity ligands and unique effector and memory differentiation when compared to Ly6C-counterparts [41,44]. Thus, evidence suggests constitutively expressed IFN impacts CD8 T cell function and may keep T cells primed to rapidly respond to viral infection.

Interferon and thymic atrophy

The thymus can be directly infected by a host of pathogens, including Mycobacterium tuberculosis, Toxoplasma gondii, HIV and LCMV [48–50]. Infection with these pathogens is associated with loss of thymic cellularity, also known as thymic atrophy [48]. Thymic atrophy has multiple causes but can be mediated by type I IFNs [49–51]. However, this is associated with infection derived IFNs, as those produced at steady state have not been associated with thymic atrophy. In a chronic LCMV infection, LCMV-specific CD8 T cells directly recognize infected dendritic cells located in the thymic medulla and mediate deletion of thymic epithelial cells and CD4+CD8+ double positive (DP) thymocytes [49]. In this model, both TCR:pMHCI interactions and type I IFN signaling in LCMV-specific CD8 T cells are required for thymic atrophy. Recovery of thymic cellularity in chronic LCMV coincides with CD8+ T cell exhaustion and can be prevented with PD-1 blockade [49]. While there is an overall loss of developing thymocytes in type I IFN mediated atrophy in LCMV, there is an increased survival of self-reactive T cells that would normally undergo clonal deletion. These decreases in clonal deletion are potentially due to loss of thymic epithelial cells, but other work has demonstrated IFN-b may directly impact mTEC differentiation and maturation by terminating RANK signaling, thereby altering the size of the thymocyte selection niche [52]. Type I IFN subtype and concentration may modulate this effect as chronic LCMV produces different IFN subtypes than acute LCMV, which is not associated with long lasting thymic atrophy [51]. Thus, infection induced IFN produced during chronic infections can impact the thymus and mechanisms of T cell central tolerance. It is interesting to speculate if IFN is produced in the thymus during chronic LCMV infection and if this acutely produced IFN has different impacts than steady state produced IFN due to the differences in IFN subtype, concentration, or timing during thymocyte developmental stage.

Conclusions

Thymic IFN plays crucial roles in the development of a healthy T cell repertoire. IFN impacts thymocyte selection and maturation through T cell intrinsic signaling and extrinsically through activation of APC. Although some of the effects of steady-state thymic IFN have been elucidated, recent evidence suggests that the development and maturation of mTEC are required for immune tolerance to IFN. The same mechanisms that promote IFN production in mTEC are likely involved in the development of tolerance to IFN and ISG-encoded proteins. To understand the influence of IFNs on thymocyte development, future studies will be needed to dissect the distinct functions played by type I and III IFNs in both the maturation of thymic antigen-presenting cells, and their direct effects on thymocytes. Proteomic studies are needed to characterize how IFN changes the immunopeptidome displayed in the thymus. The identification of IFN/ISG-derived CD4 T cell epitopes and the development of mouse models will also be helpful for detailed tolerance mechanistic studies. Animal models will likely be required to understand the mechanism by which mTEC promote tolerance to IFN, however, to date, mouse models of mTEC disfunction have not been reported to develop autoantibodies to IFNs. Lastly, it will be important to investigate how loss of tolerance to IFN impacts the severity of viral or bacterial infections other than COVID-19.