|Home | About | Journals | Submit | Contact Us | Français|
The tumor necrosis factor superfamily (TNFSF) and its corresponding receptor superfamily (TNFRSF) form communication pathways required for developmental, homeostatic, and stimulus-responsive processes in vivo. Although this receptor–ligand system operates between many different cell types and organ systems, many of these proteins play specific roles in immune system function. The TNFSF and TNFRSF proteins lymphotoxins, LIGHT (homologous to lymphotoxins, exhibits inducible expression, and competes with HSV glycoprotein D for herpes virus entry mediator [HVEM], a receptor expressed by T lymphocytes), lymphotoxin-β receptor (LT-βR), and HVEM are used by embryonic and adult innate lymphocytes to promote the development and homeostasis of lymphoid organs. Lymphotoxin-expressing innate-acting B cells construct microenvironments in lymphoid organs that restrict pathogen spread and initiate interferon defenses. Recent results illustrate how the communication networks formed among these cytokines and the coreceptors B and T lymphocyte attenuator (BTLA) and CD160 both inhibit and activate innate lymphoid cells (ILCs), innate γδ T cells, and natural killer (NK) cells. Understanding the role of TNFSF/TNFRSF and interacting proteins in innate cells will likely reveal avenues for future therapeutics for human disease.
The tumor necrosis factor superfamily (TNFSF) and its corresponding receptor superfamily (TNFRSF) mediate developmental, homeostatic, and stimulus-responsive processes in many organ systems (Locksley et al. 2001; Wiens and Glenney 2011). The ligands and receptors in the tumor necrosis factor (TNF) superfamilies form communication pathways between many different cell types (Ware 2005). The focus of this review is on the lymphotoxin and LIGHT (TNFSF-14)-related subset of TNFSF, their interacting receptors, and in their roles as effectors and regulators of innate immunity and inflammation. In particular, we focus on recent results that illustrate the key roles of these cytokines in the ability of both conventional innate lymphoid cells, such as natural killer (NK) cells, and unconventional innate-acting cells, such as B lymphocytes and γδ T cells, to initiate and attenuate inflammation.
TNF is well known as a critical factor in eliciting rapid inflammatory events acting through two distinct receptors, TNFR1 and TNFR2 (for review, see Walczak 2011). In general, ligation of these receptors results in activation of caspases, E3:ubiquitin ligases, or both. Death domain containing receptors, such as TNFR1, recruit caspase 8, whereas lymphotoxin-β receptor (LT-βR) forms an E3 ligase liberating the nuclear factor κ-light-chain-enhancer of activated B-cell (NF-κB)-inducing serine kinase (NIK) from ubiquitination and degradation (Sanjo et al. 2010). LT-βR signaling plays a key role in lymphoid organogenesis and homeostasis of lymphoid tissue microarchitecture. Herpes virus entry mediator (HVEM), TNFRSF14 acts as a molecular switch between proinflammatory and inhibitory signaling by serving as both ligand and receptor for multiple ligands (or coreceptors), creating a network for cellular communication (Kaye 2008; Cai and Freeman 2009; Murphy and Murphy 2010; Ware and Šedý 2011). The cellular ligands for HVEM come from two distinct families: the TNF-related cytokines LIGHT (TNFSF14) and lymphotoxin-α (LT-α) (Mauri et al. 1998), and the Ig superfamily members B and T lymphocyte attenuator (BTLA) (Šedý et al. 2005) and CD160 (Cai et al. 2008). The cross-utilization of ligands by HVEM, the LT-β receptor, and the two receptors for TNF (Fig. 1) create a network of signaling systems that together form a broader network of pathways regulating inflammation, and innate and adaptive immune responses (Ware 2005; Šedý et al. 2008; Ware and Šedý 2011). The critical issue currently being addressed is interpreting these molecular pathways with physiological processes, particularly in the context of host defense. The recent examples we provide in this review hopefully help address how to interpret inflammatory models to aid in designing new approaches to alter real-world disease processes in patients.
Peripheral lymphoid organs form during embryonic development requiring lymphoid tissue inducer cells (LTi) that engage stromal organizing cells. Although often thought of as the critical sites for generating adaptive immune responses, the spleen, lymph nodes, and Peyer’s patches also serve to centralize innate defenses, particularly the initial type 1 interferon (IFN-I) response to virus infections (see below). LT-αβ–LT-βR signaling is the key pathway in the formation of lymph nodes and Peyer’s patches. The LT-αβ that activates stromal cell LT-βR to initiate lymph node (LN) development is expressed in LTi cells, a highly specialized type of hematopoietic cell (Mebius et al. 1997; Cupedo et al. 2009). Embryonic LTi cells originate in the fetal liver and respond to CXCL13 to migrate to the nascent lymphoid tissues (van de Pavert et al. 2009). LTi lack LT-αβ before entry into newly formed LN, but express surface LT-α1β2 on activation through receptor activator of NF-κB (RANK) (TNFRSF11a) by its ligand (Vondenhoff et al. 2009). Following the expression of LT-α1β2, lymphoid tissue formation initiates. Expression of LT-α1β2 by LTi is also induced and sustained by IL-7 (Yoshida et al. 2002), whereas stromal cell-produced chemokines recruit additional LTi in a feed-forward loop. Successful completion of the lymphoid developmental program in rodents is largely concluded during the first week of neonatal life (Nolte et al. 2003).
Although unique during embryonic life, in the adult, LTi cells constitute a subset of a recently appreciated family of lymphocytes that are collectively known as innate lymphoid cells (ILCs) (Fig. 2). ILCs have been categorized into three main groups depending on their functional capacity. Thus, there are group 1 ILCs that secrete mainly IFN-γ (ILC1), group 2 ILCs that specialize in the production of type 2 cytokines, such as IL-5 and IL-13 (ILC2), and group 3 ILCs, which primarily produce IL-17 and IL-22 (ILC3) (Spits et al. 2013). LTi cells have been classified as ILC3 because they express IL-22 within the embryo and share a common progenitor that depends on constitutive expression of the transcription factor retinoid-related orphan receptor γt (RORγt) (Sawa et al. 2010; Vonarbourg et al. 2010). ILCs play very important roles in many inflammatory and infectious responses (reviewed elsewhere, see Bernink et al. 2013). Interestingly, it appears that LT-αβ and LT-α are also important for the function of ILC3 by regulating the IL-22-dependent clearance of intestinal bacterial infections (Wang et al. 2010) and inducing immunoglobulin A (IgA) production (Kruglov et al. 2013).
In addition to LT-αβ, LTi, and ILC (particularly ILC1 and ILC3) express high levels of an array of TNFSF ligands including but not limited to TNF-α, LIGHT (TNFSF14), OX40-ligand (TNFSF4), CD30-ligand (TNFSF8), RANK (TNFSF11a), RANK-ligand (TNFSF11), and receptors including HVEM, TNFR2 (TNFRSF1B), and death receptor (DR) 3 (TNFRSF25) (Kim et al. 2003, 2006b). With the exception of CD30-ligand and the provision of survival signals to NK cells during viral infection (Bekiaris et al. 2009), the individual contribution of each receptor or ligand in ILC-driven innate immune responses has not been formally addressed. However, given their shared signaling properties with LT-αβ, TNF, and LIGHT may contribute compensatory mechanisms for ILC function during immune development and infection.
ILC are not the only lymphocytes that have been linked with TNFR-driven innate immunity. There are at least two subpopulations of γδ T cells that have lost T-cell-receptor responsiveness, yet display bona fide innate functionality (Wencker et al. 2014). These two γδ T-cell subsets can be broadly identified by the expression of the TNFRSF CD27 as CD27+IFN-γ+ and CD27−IL-17+ (Ribot et al. 2009). Embryonic thymic signaling from CD27 and LT-βR in the stromal cells have been shown to regulate the balanced development of IFN-γ- versus IL-17-producing γδ T cells (Silva-Santos et al. 2005; Ribot et al. 2009), whereas LT-βR is critical for the full maturation and activation of CD27−IL-17+ γδ T cells (Powolny-Budnicka et al. 2011). It is likely that the LT-βR requirement is indirect because LT-βR expression is restricted to stromal or cells of myeloid lineage (Murphy et al. 1998). Furthermore, CD30 has also been shown to provide survival and activation signals to γδ T cells at mucosal sites (Sun et al. 2013).
In humans, we recently reported the presence of a CD4+CD3− innate-like T cells that are characterized by high levels of constitutive TNF and inducible surface LT-αβ that are prevalent in the blood of rheumatoid arthritis patients (Bekiaris et al. 2013b). The functional role of TNF and LT-αβ in CD4+CD3− innate T cells has not been elucidated.
Although not required for the formation of the spleen, the LT-αβ–LT-βR pathway provides key signals that drive the maturation of the spleen and its microarchitecture. Maintenance of mature splenic organization and homeostasis is initially dependent on embryonic LTi cells, whereas B lymphocytes provide homeostatic signals at later times (Cyster 2003). In contrast, innate lymphocytes continue to provide key homeostatic signals via the LT-αβ–LT-βR pathway in lymph nodes. The LT-αβ–LT-βR system plays multiple roles in forming and maintaining the microarchitecture providing key differentiation signals to stromal cells (Cyster 2003), subsets of dendritic cells (DC) (CD11chiCD8−) (Kabashima et al. 2005; De Trez et al. 2008), follicular dendritic cells (Endres et al. 1999), and marginal sinus macrophages (Kraal and Mebius 2006). In addition to regulating antiviral responses via type I IFN, engagement of the LT-βR on the embryonic lymphoid tissue stromal cells initiates a signaling cascade that promotes the differentiation of these cells into a specialized population known as stromal organizer cells (White et al. 2007). Stromal organizing cells then express the homeostatic chemokines CXCL13, CCL19, and CCL21 (Dejardin et al. 2002) that will attract B and T lymphocytes after birth and organize them into their distinct T-cell zones and B follicles the regions found in mature lymph nodes. In addition, stromal organizing cells express the cytokine interleukin (IL)-7, which is the major survival factor for naïve lymphocytes within the developing and mature lymphoid organ (Link et al. 2007; Meier et al. 2007). Recent work has shown that within lymph nodes the specific progenitor to stromal organizing cell is a preadipocyte, which differentiates using an LT-βR-dependent mechanism (Benezech et al. 2012).
The sinuses in the spleen and lymph nodes provide the essential filtering mechanism that capture blood- and lymph-borne pathogens (Junt et al. 2008). Within these structures, the sinus-lining macrophages in the spleen and lymph nodes are critically important for capturing pathogens. Their differentiation requires B-cell expression of LT-αβ to produce the initial wave of type 1 interferons (IFN-I including both IFN-β and IFN-α). Studies of two distinct viruses in their natural hosts, mouse cytomegalovirus (CMV) a β-herpesvirus with a large DNA genome (Schneider et al. 2008) and vesicular stomatitis virus (VSV), a Rhabdovirus with a small RNA genome (Moseman et al. 2012) revealed the innate role of B cells in providing LT-αβ that drives the initial IFN-I response. Mice deficient in Lt-β, specifically in B cells fail to produce IFN-I in response to CMV or VSV infection. The “innateness” of B cells in IFN-I-mediated defense against VSV was clearly defined using the DHLMP2A mice (Casola et al. 2004). Activating signals from the herpesvirus LMP2A gene retain B cells and have normal lymphoid tissue architecture, yet are devoid of surface or secreted antibodies.
The initial IFN-I defense to CMV occurs rapidly, peaking within 8–12 h after infection and accounts for >80% of the IFN-I in the circulation. Mouse CMV infects the CCL21 expressing LT-βR-dependent stromal cells in the splenic marginal sinus and T-cell zone (Benedict et al. 2006). IFN-I production requires active signaling by the LT-βR within these infected cells (Banks et al. 2005). IFN-I induction measured as IFN-β or IFN-α mRNA increases by several orders of magnitude in LT-βR+ stromal cells paralleling viral gene expression, yet is independent of TLR signaling (Schneider et al. 2008). In contrast, LT-βR signaling in DC produces IFN-β at levels reflecting homeostasis (Summers deLuca et al. 2011). Thus, communication between innate B cells and stromal cells initiates and amplifies the earliest IFN-I response to mouse CMV infection (Fig. 3) (Schneider et al. 2008).
Infection with VSV differs from CMV in that macrophages are the primary target following subcutaneous infection and macrophages provide the source of IFN-I. Innate B cells via LT-αβ–LT-βR pathway differentiate CD169+ subcapsular macrophages into a permissive state for VSV replication. In macrophages, the permissive state of these cells is controlled by Usp18, which encodes an ISG-15 deconjugating enzyme that also restricts IFN-I receptor signaling, destabilizing the effects of IFN-I (Kim et al. 2006a, 2008). The Usp18 gene is needed to mount an effective adaptive (antibody) response to VSV (Honke et al. 2012).
The consequences of infection are severe when the LT-IFN axis fails to function. In the absence of LT-βR signaling, IFN-I production and VSV replication in lymph nodes is blocked; however, neurons innervating the nodes become a virus target leading to spread of virus to the central nervous system with ensuing paralytic disease. In contrast, CMV replicates in the absence of IFN-I or LT-βR signaling with disastrous consequences for lymphoid tissues, causing a massive apoptosis of T and B cells in the absence of direct infection (splenic necrosis). IFN-I is a key factor in promoting T- and B-cell viability during antigen challenge (Marrack et al. 1999) indicating that the LT-IFN axis promotes adaptive defenses.
The conceptual outcome of these results suggests that the LT-IFN axis creates a restricted microenvironment containing sentinel pathogen permissive cells located in the sinuses of lymphoid organs as the first line of innate defense to viral pathogens (Fig. 3) (Khanna and Lefrancois 2012; Ware and Benedict 2012). The LT-βR system controls expression of CXCL13 and CCL21 that position CD169+ macrophages within lymph node and splenic microenvironment. The membrane-anchored position of the LT-αβ complex indicates that only those macrophages in cell contact with innate B cells will differentiate into a permissive state, and, consequently, creating a restricted microenvironment for pathogen replication. Usp18 is an example of a gene that induces unresponsiveness to IFN-I signaling, thus restricting virus replication in the sinus-lining macrophages. It is not clear whether this mechanism operates in stromal cells. This IFN-I unresponsive state allows for high virus replication and, thus, antigen production in a restricted microenvironment. High antigen concentration would promote efficient B-cell recognition for transport to follicular regions to generate high affinity antibody responses. Additionally, it is predicted that IFN-I resistant mutations would not be selected during the initial rounds of virus replication. Additional pathogens show strong dependence on sinus-lining macrophages and the LT-IFN axis suggesting a broader host defense strategy. It is not surprising then that some successful pathogens have evolved strategies to evade these pathways (Šedý et al. 2008).
HVEM is one of the most evolutionary conserved TNFRSF members with orthologs expressed in many species including humans and lamprey (Guo et al. 2009). Diverse binding modalities of HVEM with LIGHT, LT-α3, and the immunoglobulin (Ig) superfamily members BTLA and CD160 in trans and in cis have opened up the TNFR signaling pathways to further regulation and unique immunological properties. HVEM was originally identified as an entry receptor for herpes simplex virus and is a widely expressed in hematopoietic cells and in epithelia cells, and binds the inducible TNF family members LT-α3 and LIGHT (Murphy et al. 2006; Ware and Šedý 2011).
BTLA is an inhibitory receptor that binds HVEM at a distinct surface from TNF ligands through its Ig domain, resulting in bidirectional signaling in cells expressing both proteins (Šedý et al. 2005; Cheung et al. 2009; Murphy and Murphy 2010). BTLA was originally identified as a Th1-specific transcript, but is expressed in many cells in the immune system including B cells, αβ and γδ T cells, NK cells, DC, and macrophages (Watanabe et al. 2003; Murphy and Murphy 2010; Bekiaris et al. 2013a; Šedý et al. 2013). HVEM activates BTLA resulting in phosphorylation of its cytoplasmic domain and recruitment of the Src homology-2 (SH2) containing phosphatase (SHP)-1 and 2 to its carboxy-terminal motif (Watanabe et al. 2003; Šedý et al. 2005). The hematopoietic and epithelial cell–specific SHP-1 protein functions predominantly to dephosphorylate tyrosine within activated kinases or adaptor proteins, thus limiting signal transduction pathways (Van Vactor et al. 1998; Pao et al. 2007). In contrast, a major function of the more ubiquitously expressed SHP-2 protein is to activate RAS/mitogen-activated protein kinase (MAPK) signaling, although SHP-2 has been associated with activation of several inhibitory receptors (Tiganis and Bennett 2007). Direct targets of BTLA-associated SHP-1 and -2 phosphatase activity are thought to include the CD3ζ chain in T cells and Syk in B cells (Wu et al. 2007; Vendel et al. 2009), but little is known about BTLA signaling in other cell subsets. A peptide derived from an additional tyrosine-containing motif within the BTLA cytoplasmic tail could recruit growth factor receptor bound-2 (Grb2) and the p85 regulatory subunit of phosphoinositide-3-kinase (PI3K), although a function of these associations has not been ascribed in vivo (Gavrieli and Murphy 2006). HVEM and LIGHT are each expressed on a variety of cell types including T cells, DC, macrophages, and NK cells, and activation of HVEM by LIGHT or BTLA leads to NF-κB-driven gene activation and inflammatory signaling (Murphy et al. 2006).
HVEM was also identified as a ligand for CD160, a glycosylphosphoinosiol (GPI)-linked receptor expressed at low levels in many cells and showing greater expression in NK cells (Cai et al. 2008; Le Bouteiller et al. 2011). CD160 has been shown to interact MHC-I proteins in mouse and human, including HLA-C (Le Bouteiller et al. 2002; Maeda et al. 2005). HLA-C binds CD160 in NK cells to activate cytolysis, and the expression of IFN-γ, TNF-α, and IL-6 (Le Bouteiller et al. 2002; Barakonyi et al. 2004). CD160 may function as a costimulatory receptor in the absence of CD28 on T cells to activate PI3K signaling (Nikolova et al. 2002; Rabot et al. 2007). The mechanism of GPI-linked CD160 signaling is unclear, although it may transmit signals with cell surface–associated CD2 in NK cells (Rabot et al. 2006). Additionally, in CD4+ T cells, CD160 was shown to inhibit T-cell activation in response to HVEM ligation (Cai et al. 2008). In contrast, activated NK cells can express an alternate spliced form of CD160 that encodes a transmembrane domain and cytoplasmic tail. This form of CD160 costimulates protein kinase B (PKB; AKT) and extracellular signal-regulated kinase (ERK) signaling in response to HVEM activation.
NK cells mediate clearance of infected cells largely through the secretion of inflammatory cytokines, such as TNF-α or IFN-γ, or through direct lysis of infected cells (Vivier et al. 2011). Many pathogens that establish chronic or latent infections use a variety of immunoregulatory mechanisms to evade initial clearance by innate effector cells, such as NK cells (Lanier 2008). The role of HVEM as a receptor for HSV raised the possibility that viruses may alter the activity of HVEM or its ligands during infection as an immunoregulatory mechanism. We previously showed that human CMV expresses a viral mimic of HVEM (orf UL144) that binds and activates BTLA to inhibit T-cell proliferation (Benedict et al. 1999; Cheung et al. 2005). We and others found that the orf UL144 protein could not bind LT-α3 or LIGHT, likely because of its limited domain structure (Benedict et al. 1999; Poole et al. 2006). However, we also found that the orf UL144 protein could not bind CD160, which binds an overlapping site on HVEM with BTLA (Kojima et al. 2011; Šedý et al. 2013). HVEM binding to CD160 enhanced NK-cell activation in response to CMV and costimulated NK cytolytic function and cytokine release, whereas NK lines expressing high levels of BTLA had reduced cytolytic function.
Together, these data showed that BTLA and CD160 counter-regulate NK-cell activation in response to HVEM binding (Fig. 4). Expression of BTLA following NKG2D stimulation likely serves to limit CD160 costimulatory signals as a means to shut down activation (Šedý et al. 2013). The expression of a BTLA-binding protein in CMV that avoids CD160 activation may have evolved to promote infection by selectively engaging inhibitory pathways and tip the balance in favor of reduced NK-cell activation. As an additional measure against NK killing, CMV expresses a second protein, orf UL141, which binds and down-regulates expression of two apoptotic TNF receptors (DR4 and 5) in infected cells (Smith et al. 2013).
HVEM costimulation of NK cells is greatest in donors that had not been previously infected with human CMV. In CMV seropositive individuals, an NKG2C+ population of NK cells expands that is suggested to acquire memory cell properties and, thus, may be less reliant on costimulatory signals (Lopez-Verges et al. 2011; Foley et al. 2012a,b; Muntasell et al. 2013). We did not detect differences in CD160 expression between NKG2C+ or NKG2C− cells. Thus, the response of NKG2C− cells to HVEM may reflect a requirement for CD160 signaling to initiate transcriptional programs already active in NKG2C+ cells. Activating NKG2C and inhibitory NKG2A receptors bind to human leukocyte antigen (HLA)-E, which presents leader peptides from MHC class I. In accordance with the missing-self hypothesis of NK-cell activation, the absence of NKG2A inhibitory signaling through SHP-1 allows for NK activation. The role for lower affinity interactions between NKG2C and HLA-E-peptide complexes is less clear, although it has been proposed that a stable complex of HLA-E with a CMV nonamer peptide from the orf UL40 protein may activate NKG2C (Heatley et al. 2013; Muntasell et al. 2013). Intriguingly, HLA-E polymorphisms have been associated with the development of psoriasis lesions and decreased frequencies of NKG2C+ NK cells (Batista et al. 2013; Patel et al. 2013; Zeng et al. 2013). Future studies will be required to clarify how these receptors are involved in the development of disease.
We and others recently observed the inhibitory effect of BTLA in innate γδ T cells (Bekiaris et al. 2013a; Gertner-Dardenne et al. 2013). Mice deficient in BTLA had significantly increased IL-17-producing RORγt+CD27− γδ T cells that are implicated in autoimmune diseases (Sutton et al. 2012; Ribot and Silva-Santos 2013). Additionally, we found that BTLA deficiency conferred a competitive advantage to γδ T cells in mixed bone marrow chimera. γδ T-cell homeostatic expansion was previously shown to be dependent on common γ chain signaling through IL-7 and IL-15 (Baccala et al. 2005). BTLA-deficient γδ T cells also show increased IL-17 and TNF-α production in response to IL-7. Together these data indicate that BTLA may directly regulate cytokine signals in innate cells. SHP-1 has been shown to regulate signaling of several cytokines, including those using the common γ chain, whereas Janus kinase (JAK) and signal transducer and activator of transcription (STAT) proteins have been described as substrates for both SHP-1 and SHP-2 (Rakesh and Agrawal 2005; Pao et al. 2007; Tiganis and Bennett 2007). Recently, SHP-1 was reported to limit TH17 development through inhibition of IL-21 signaling, and T-cell homeostasis through inhibition of IL-4 (Mauldin et al. 2012; Johnson et al. 2013). Notably, it is unclear how SHP phosphatases are recruited to cytokine receptors, or whether BTLA participates in recruitment to receptors. We observed induction of BTLA expression in response to IL-7 signaling, providing a mechanism through which cytokine signals can be regulated (Fig. 5). This cytokine signaling circuit was also present in ILC3, which also gained a competitive advantage by BTLA deficiency in mixed bone marrow reconstitution experiments, and which also induced BTLA expression. Thus, in ILC and γδ T cells, IL-7 is regulated by negative feedback through BTLA.
Using the Aldara/Imiquimod (IMQ) cream–induced psoriasis model (van der Fits et al. 2009), we further show the necessity of BTLA to suppress skin inflammation by regulating γδ T-cell expansion and cytokine secretion. IMQ cream–induced psoriasis is strongly dependent on skin resident nonlymphoid cells (Cai et al. 2011) and is largely TLR7-dependent, although it may also induce TLR7-independent inflammation (Pantelyushin et al. 2012; Walter et al. 2013). Whether BTLA or other inhibitory receptors act very early to suppress TLR signaling in the skin is currently unknown. Interestingly, whereas IMQ can induce IFN-αβ in a TLR7/MyD88-dependent mechanism (Hemmi et al. 2002), it was recently shown that signaling downstream from the type I IFN receptor is not required for the development of psoriasis or the activation of IL-17-producing γδ T cells (Walter et al. 2013; Wohn et al. 2013). Because of the importance of BTLA in inhibiting CD27−IL-17+ γδ T-cell responses and the key role of innate IL-17 during bacterial infection, it would make evolutionary sense to maintain low levels of BTLA at steady state. In this regard, we show that RORγt repressed BTLA transcription through direct binding to the Btla promoter (Bekiaris et al. 2013a).
In humans, proliferation of γδ T cells expressing Vγ9- and Vδ2-encoding T-cell receptors was reduced by HVEM binding to BTLA, and enhanced by antibodies or decoy receptors that blocked the HVEM-BTLA interaction (Gertner-Dardenne et al. 2013). Vγ9Vδ2+ T cells are the predominant population of γδ T cells circulating within human blood, and have potent cytolytic activity against a wide variety of tumors and infected cells (Caccamo et al. 2010; Kalyan and Kabelitz 2013). Although the cognate receptor for Vγ9 Vδ2+ T cells remains unclear, they are activated by nonpeptide phosphorylated isoprenoid pathway metabolites whose presentation is facilitated by stress-induced proteins and CD277 (Harly et al. 2012). Recently, much attention has been focused on evaluating the effectiveness of phosphoantigens or activated Vγ9Vδ2+ T cells as anticancer therapies (reviewed in Caccamo et al. 2010). Blocking the BTLA-HVEM interaction has been proposed as a further measure to enhance antitumor responses of Vγ9Vδ2+ T cells (Lopez 2013).
The role of TNFRSF in balancing DC homeostasis is well documented. In this regard, LT-βR is necessary for the proliferation of lymphoid tissue resident CD8− DC (Kabashima et al. 2005). In contrast, although mice deficient in LT-βR have reduced numbers of CD8− DCs, their numbers in BTLA- or HVEM-deficient animals show the opposite phenotype, with significantly increased CD8− DC (De Trez et al. 2008), suggesting that ITIM-dependent signals suppress the impact of the LT-βR pathway to balance DC homeostasis. Similarly, recent data have shown that along with Notch2, LT-βR instructs the normal development of tissue resident and migratory intestinal DC subsets (Satpathy et al. 2013). Moreover, BTLA expression in DC and macrophages is necessary to suppress LPS-induced TNF and IL-12 production, and prevent endotoxin shock (Kobayashi et al. 2013). The importance of BTLA in DC biology has been recently reemphasized by the discovery that it is one of the most highly expressed genes in mouse and human CD103+CD11b− DC (Watchmaker et al. 2014).
A wide variety of tumors express HVEM and, thus, have the potential to inhibit the proliferation of Vγ9Vδ2+ T cells. Interestingly, several groups have reported that a high percentage of adult onset and pediatric follicular lymphoma and diffuse large B-cell lymphoma contain deletions in the gene-encoding HVEM (TNFRSF14), and are associated with worse prognosis (Cheung et al. 2010; Launay et al. 2012; Lohr et al. 2012; Bjordahl et al. 2013; Martin-Guerrero et al. 2013). We previously argued that TNFRSF14 deletions may be acquired in more aggressive tumors as an adaptation to prevent NK cell costimulation through CD160 (Šedý et al. 2013). CD160 is also present in human γδ T cells, and identifies the recently characterized lytic ILC1 population, although it is not clear how this receptor signals in these cells (Maiza et al. 1993; Fuchs et al. 2013). Another possible selective pressure for follicular lymphoma to delete TNFRSF14 is to prevent interactions with BTLA-expressing TFH cells (Chtanova et al. 2004; Nurieva et al. 2008; M’Hidi et al. 2009). Follicular lymphomas containing a higher content of T cells with TFH markers are associated with increased survival (Byers et al. 2008; Carreras et al. 2009; Pangault et al. 2010). Further studies will be required to determine the relative contributions of innate and adaptive compartments in shaping the tumor microenvironment and the development of more aggressive tumors.
BTLA has also been shown to control inflammation induced by a number of infections in mice including malaria and listeria (Lepenies et al. 2007; Sun et al. 2009). In humans, increased expression of BTLA in lymphocytes is associated with the presence of chronic infections, such as cytomegalovirus and hepatitis B (Serriari et al. 2010; Cai et al. 2013). In contrast, decreased BTLA expression in lymphocytes is observed during infection with human immunodeficiency virus, possibly owing to regulation by type I interferon (Xu et al. 2009; Zhang et al. 2011; Boliar et al. 2012; Larsson et al. 2013). The HVEM-BTLA-CD160 signaling complex is also critical for the innate immune properties of intestinal intraepithelial lymphocytes. Thus, following bacterial infection CD160 in intraepithelial lymphocytes engages HVEM on the epithelium to increase levels of the IL-22 receptor and initiate a STAT3-dependent pathway necessary for clearing the infection (Shui et al. 2012). Whether these are CD160-expressing ILC1 (see above) remains to be determined.
The prevention of autoimmune recognition is thought to be controlled by lymphocyte inhibitory receptors, such as CTLA4, PD-1, BTLA, and their ligands (Watanabe and Nakajima 2012). BTLA has been shown to be important in regulating autoimmunity in several disease models, such as experimental autoimmune encephalitis, autoimmune cardiomyopathy, dermatitis, and airway hypersensitivity (Watanabe et al. 2003; Tao et al. 2005; Deppong et al. 2006; Bekiaris et al. 2013a). In a model of CD4+ T-cell-driven colitis, HVEM expression in recipient animals lacking T or B cells was shown to prevent the development of disease (Steinberg et al. 2008). A SNP in the human BTLA locus is associated with rheumatoid arthritis (RA) (Table 1) (Lin et al. 2006; Oki et al. 2011). SNPs in TNFRSF14 have also been associated with RA as well as multiple sclerosis, ulcerative colitis, and celiac disease, possibly owing to the role of HVEM as a focal point of network interactions (Raychaudhuri et al. 2008; Dubois et al. 2010; Perdigones et al. 2010; Anderson et al. 2011; Blanco-Kelly et al. 2011; Jostins et al. 2012; Herraez et al. 2013).
The association of innate cells in diseased tissues and their role in promoting inflammation in autoimmune disease is well established. Blockade of TNF has proven effective in clinical settings of inflammation (rheumatoid arthritis psoriasis and inflammatory bowel diseases), although significant subsets of patients are refractory to treatment with TNF inhibitors, and in some diseases (e.g., multiple sclerosis), TNF inhibitors are contraindicated. One must consider whether other pathways are active in these pathologies, and indeed whether these pathologies may arise from alterations in inhibitory signaling pathways that may, at least in part, be because of defects in innate cell function.
The authors thank Lisa Marie Bellovich for figures and editing and the National Institutes of Health (AI-033068, AI48073, and CA164679) and Jean Perkins Family Foundation for support.
Editor: Ruslan M. Medzhitov
Additional Perspectives on Innate Immunity and Inflammation available at www.cshperspectives.org