Journal list menu
A Guide to IL-1 family cytokines in adjuvanticity
Abstract
Growing awareness of the multiplicity of roles for the IL-1 family in immune regulation has prompted research exploring these cytokines in the context of vaccine-induced immunity. While tightly regulated, cytokines of the IL-1 family are normally released in response to cellular stress and in combination with other danger-/damage-associated molecular patterns (DAMPs), triggering potent local and systemic immune responses. In the context of infection or autoimmunity, engagement of IL-1 family receptors links robust innate responses to adaptive immunity. Clinical and experimental evidence has revealed that many vaccine adjuvants induce the release of one or multiple IL-1 family cytokines. The coordinated release of IL-1 family members in response to adjuvant-induced damage or cell death may be a determining factor in the transition from local inflammation to the induction of an adaptive response. Here, we analyse the effects of IL-1 family cytokines on innate and adaptive immunity with a particular emphasis on activation of antigen-presenting cells and induction of T cell-mediated immunity, and we address in detail the contribution of these cytokines to the modes of action of vaccine adjuvants including those currently approved for human use.
Abbreviations
-
- APCs
-
- antigen-presenting cells
-
- cDCs
-
- conventional dendritic cells
-
- CMI
-
- cell-mediated immunity
-
- CTLs
-
- cytotoxic T lymphocytes
-
- DAMPs
-
- danger-/damage associated molecular patterns
-
- EAE
-
- experimental autoimmune encephalomyelitis
-
- GM-CSF
-
- granulocyte-macrophage colony-stimulating factor
-
- HMGB-1
-
- high mobility group box 1
-
- IFN
-
- interferon
-
- IFN-I
-
- type I interferon
-
- IL
-
- interleukin
-
- LCs
-
- Langerhans cells
-
- LPS
-
- lipopolysaccharide
-
- MAMPs
-
- microbe-associated molecular patterns
-
- MAPK
-
- mitogen-activated protein kinase
-
- MHC
-
- major histocompatibility complex
-
- MOA
-
- mode of action
-
- moDCs
-
- monocyte-derived dendritic cells
-
- NETs
-
- neutrophil extracellular traps
-
- NF‑κB
-
- nuclear factor‑κB
-
- NK
-
- natural killer
-
- PAD4
-
- peptidylarginine deiminase 4
-
- PBMCs
-
- peripheral blood mononuclear cells
-
- pDCs
-
- plasmacytoid dendritic cells
-
- RA
-
- receptor antagonist
-
- s.c.
-
- subcutaneous
-
- Tc
-
- T cytotoxic
-
- TCR
-
- T-cell receptor
-
- Th
-
- T helper
-
- TIR
-
- Toll-IL-1-receptor
-
- TLR
-
- Toll-like receptor
-
- WT
-
- wild-type
Introduction
Vaccine adjuvants: current status and future challenges
Vaccination is one of the most successful and cost-effective medical interventions, having eradicated or nearly eradicated diseases such as smallpox, rubella and poliomyelitis. Traditional vaccines consist of attenuated or inactivated pathogens that provide antigen arrays combined with microbe-associated molecular patterns (MAMPs) that act as natural enhancers of immunity. Recombinant antigen production led to a shift towards well-defined subunit formulations which are less reactogenic but also less immunogenic [1]. Adjuvants became essential components of subunit formulations to enhance immunogenicity and boost subunit-vaccine efficacy.
Adjuvants influence the type, quality and breadth of the immune response. For 70 years, aluminium salts remained the only approved adjuvant for human use. Although the term alum strictly corresponds to potassium aluminum sulfate, the literature has adopted it to generally refer to most aluminium-containing adjuvants including aluminum hydroxide and aluminum phosphate. Hence in this review term alum generally refers to alum-containing adjuvants. Alum drives T helper 2 (Th2) and potent humoral responses sufficient to tackle pathogens with a stable antigen repertoire and their secreted toxins by promoting the generation of neutralising antibodies. However, protection against diseases such as tuberculosis, malaria, HIV, dengue and cancer also requires potent cell-mediated immunity (CMI) including Th1, Th17 and cytotoxic T lymphocytes (CTLs) [2, 3]. Since the 1990s, few new adjuvants have been licensed for human use. Among the newly licensed adjuvants, only AS01E has been shown to elicit protective Th1 cell-mediated immunity [4]. Three other adjuvant candidates in clinical development IC31, CAF01 and GLA-SE also induce strong T-cell responses in humans [3, 5-7]. Until now, adjuvants have been empirically developed, but as inadequate adjuvant selection may lead to vaccine failure even with a promising antigen, it is important that we understand the adjuvants’ modes of action (MOA) to select the most suitable candidate according to the type of response required [8]. Below we address the roles of IL-1 family cytokines in innate and adaptive immunity and their effects on activation of antigen-presenting cells and induction of T cell-mediated immunity. We present the evidence that IL-1 family members play key roles in adjuvant MOA and discuss how selective induction of IL-1 family cytokines by adjuvants could be harnessed to induce and direct cell-mediated immunity and improve vaccine efficacy.
An overview of the IL-1 family members: agonists, antagonists and receptors
The interleukin 1 (IL-1) cytokine family members are central mediators of immune responses and inflammation, regulating many processes in health and disease, with broad effects on immune and nonimmune cells [9]. IL-1 was the only identified member of the family until the cloning and expression of IL-1 DNA in 1984 revealed that it actually consisted of two proteins with overlapping effects: IL-1α and IL-1β [10]. Currently, the IL-1 family comprises 11 ligands including seven proinflammatory cytokines: IL-1α, IL-1β, IL-18, IL-33, IL-36α, IL-36β and IL-36γ; three receptor antagonists (RA) which inhibit their biologic activities: IL-1RA (which blocks IL-1 receptor signalling) and IL-36RA and IL-38 (both antagonists for IL-36R). IL-37 is the only known anti-inflammatory cytokine of the IL-1 family [11, 12] (Table 1).
Common name | Alternative name | Subfamily | Receptor | Coreceptor | Antagonist/inhibitory receptor | Functions | Known secretion signal peptide | Proteolytic activation/inactivation | Protease/s | Released upon cell death | Expression |
---|---|---|---|---|---|---|---|---|---|---|---|
IL-1α | IL-1F1 | IL-1 | IL-1R1 | IL-1RAcP | IL-1RA, IL-1R2 | Proinflammatory | No | Active membrane-bound form and soluble precursor further activated by proteases [16] | Calpain, chymase, NE, GzmB [16] | Yes | Keratinocytes, thymus, hepatocytes, endothelial cells, fibroblasts, epithelial cells, monocytes, macrophages, DCs, B lymphocytes [12] |
IL-1β | IL-1F2 | IL-1 | IL-1R1 | IL-1RAcP | IL-1RA, IL-1R2 | Proinflammatory | No | Produced as inactive pro-IL-1β, activated upon cleavage [16] |
Casp-1 (inflammasome activation). Noncanonical casp-8; also NE, PR-3, CathG, chymase [16] |
Yes | Primarily produced by monocytes, macrophages, and DCs (microglia). Low production by B cells and NK cells [12] |
IL-18 | IL-1F4 | IL-18 | IL-18Rα | IL-18Rβ | IL-18BP | Proinflammatory | No | Produced as inactive pro-IL-18, activated upon cleavage. | Casp-1 (inflammasome activation). GzmB, chymase. | Yes | Constitutively expressed in blood monocytes, intestinal epithelial cells (human). Also expressed in murine macrophages, DCs, endothelial cells, intestinal epithelial cells, and keratinocytes under steady state [155] |
IL-33 | IL-1F11 | IL-1 | ST2L (IL-1RL1) | IL-1RAcP | Soluble ST2 (sST2) | Proinflammatory | No | May be activated or inactivated by cleavage [16] |
Activated by: CathG, chymase, NE. Inactivated by: Casp-3 and 7, poor substrate for casp-1 [156, 157] |
Yes | Epithelial cells at barriers, i.e. skin, lung, intestine; adipocytes; endothelial cells; tumour cells [39] |
IL-36α | IL-1F6 | IL-36 | IL-36R (IL-1Rrp2) | IL-1RAcP | IL-36RA | Proinflammatory | No | Cleavage at 9 amino acids N-terminal to a conserved A-X-Asp (K6) for optimal activity [134] | Unknown | Unknown | Monocytes, T/B cells, spleen, bone marrow, tonsils, lymph nodes, skin [134] |
IL-36β | IL-1F8 | IL-36 | IL-36R (IL-1Rrp2) | IL-1RAcP | IL-36RA | Proinflammatory | No | Cleavage at 9 amino acids N-terminal to a conserved A-X-Asp (R5) for optimal activity [134] | Unknown | Unknown | Monocytes, T/B cells, bone marrow, tonsils, heart, lung, testis, colon, neurons, glial cells [134] |
IL-36γ | IL-1F9 | IL-36 | IL-36R (IL-1Rrp2) | IL-1RAcP | IL-36RA | Proinflammatory | No | Cleavage at 9 amino acids N-terminal to a conserved A-X-Asp (S18) for optimal activity [134] | Unknown | Release noted in apoptotic keratinocytes [142] | Predominantly secreted by epithelial cells; also monocytes and THP1, macrophages and T cells; IL-36γ production induced in keratinocytes after TNF-α or PMA treatment [134, 158] |
IL-1RA | IL-1F3 | N/A | IL-1R1 | – | N/A | Anti-inflammatory | Yes | No | N/A | Yes | Usually produced by the same cells that express IL-1α or IL-1β, i.e. monocytes, macrophages, DCS, neutrophils and epithelial cells [104] |
IL-36RA | IL-1F5 | IL-36 | IL-36R (IL-1Rrp2) | Blocks IL-1RAcP recruitment | N/A | Anti-inflammatory | No | Removal of its N-terminal methionine [134] | Neutrophil elastase [159] | Unknown | Constitutively expressed in keratinocytes; present in monocytes, B cells, DCs, skin, brain, kidney, heart [134] |
IL-37 | IL-1F7 | IL-18 | IL-18Rα | SIGIRR (TIR8, IL-1R8) [160] | Unknown | Anti-inflammatory | No | Active precursor and cleaved forms. Casp-1 cleavage may promote nuclear translocation and anti-inflammatory activitya | Casp-1b, poor substrate for casp-4 [161] | No | Absent in mice. Expressed in human colonic plasma cells, tonsils, colonic lamina propria and stromal cells and skin cells [161], DCs, mononuclear cells and monocytes [162] |
IL-38 | IL-1F10 | IL-36 | IL-36R (IL-1Rrp2) | Blocks IL-1RAcP recruitment | Unknown | Anti-inflammatory | No | Unknown | Unknown | One study reporting release during apoptosis [163] | mRNA expressed in heart, placenta, fetal liver, skin, spleen, thymus and tonsil. Protein found in skin and proliferating B cells in the tonsils [164] |
- N/A, not applicable; IL, interleukin; RA, receptor antagonist; DCs, dendritic cells; NK, natural killer; NE, neutrophil elastase; GzmB, Granzyme B; Casp-, caspase; CathG, cathepsin G; PR-3, proteinase-3; PMA, phorbol-mirystate acetate; TNF-α, tumour necrosis factor alpha; mRNA, messenger ribonucleic acid.
- a Studies performed using recombinant human IL-37 in murine macrophages; results of this study must be carefully interpreted [165].
- b Cleavage of IL-37 by endogenous caspase-1 has not yet been definitely demonstrated as mutating the proposed caspase-1 cleavage site only marginally affects LPS-induced cleavage of IL-37. This may suggest that other proteases might be involved in processing of IL-37 [165].
Ten IL-1 family receptors (IL-1R) have been identified and named chronologically on their discovery as IL-1R1, IL-1R2, etc. Each cytokine binds to and signals through its cognate receptor, triggering recruitment of an accessory chain to form a heterotrimeric complex containing three immunoglobulin (Ig)-like regions. Receptor activation induces phosphorylation of Toll-IL-1-receptor (TIR) domains, activation of nuclear factor‑κB (NF‑κB) and mitogen-activated protein kinase (MAPK) (reviewed in [13, 14]).
Interestingly, the lack of a canonical secretion signal sequence is a common feature of most proinflammatory members of the IL-1 family cytokines and their release has been linked to cell death (Table 1) [15]. Many IL-1 family cytokines are produced as procytokines and require protease cleavage to gain full agonist properties (Table 1). IL-1β and IL-18, can be activated by multiprotein complexes known as inflammasomes which act as molecular platforms for procaspase-1 activation and processing of the procytokines into mature active forms (Fig. 1) [8]. Pro-IL-1β and pro-IL-18 can also be activated by neutrophil and mast cell proteases, granzymes and some apoptotic caspases (Table 1) [16]. IL-1α, IL-33 and IL-1RA can exert their effects without being cleaved, although processing by neutrophil proteases and granzymes from cytotoxic lymphocytes or natural killer (NK) cells greatly amplifies the proinflammatory activity of IL-1α and IL-33 [16].

Receptor activation is tightly regulated by receptor antagonists (i.e. IL-1RA and IL-36RA), inhibitory receptors (i.e. IL-1R8, also known as SIGIRR or TIR8) and decoy receptors (IL-1R2 and IL-18-binding protein, IL-18BP) that prevent intracellular signalling and inhibit the cytokines’ actions to avoid pathologic inflammation [13, 15]. The biochemistry of IL-1 family cytokine processing is beyond the scope of this review and has been addressed elsewhere [9, 13, 17].
Modulation of innate and adaptive immunity by IL-1 family cytokines
Cellular immunity is key to protection against several pathogens and in immunity to cancer. T cells typically require three signals from antigen-presenting cells (APCs) for activation: 1) engagement of the T-cell receptor (TCR) by complexes of antigen and major histocompatibility complex (MHC) class I or II displayed on the APC surface; 2) costimulation by the APC via molecules including CD80, CD86 and CD40 which promote prosurvival signals in the T cell; and 3) secretion of polarising cytokines including many members of the IL-1 family, which trigger differentiation of naïve T cells into different T helper (Th) or T cytotoxic (Tc) subsets (Fig. 2) [18, 19]. Adjuvants can influence these three signals, shaping T-cell responses.

Influence of the IL-1 family of cytokines on antigen-presenting cells
IL-1α and IL-1β are potent activators of DC subsets including conventional (cDCs) and Langerhans cells (LCs) which are highly phagocytic cells populating the epidermis and the first line of defence against skin pathogens. LCs can differentiate into DCs that polarise naïve CD4+ T cells into Th2 [20], Th17 [21] or IL-22 secreting T cells [22] and prime and cross-prime CD8+ T cells [23]. IL-1α enhances maturation of human monocyte-derived DCs (moDCs), increasing their ability to stimulate allogeneic CD4+ T-cell responses and enhancing their secretion of interferon (IFN)-γ and IL-13 [24]. Similarly, IL-1β enhances CD40L-mediated activation of human DCs, augmenting their ability to drive IFN-γ secretion in T cells [25] and to produce IL-12 [26]. IL-1α activates LCs in vitro when combined with granulocyte-macrophage colony-stimulating factor (GM-CSF) [27], whereas intradermal administration of IL-1β promotes LC migration into the lymph nodes [28], enhances MHC-II expression and naïve T-cell proliferation more efficiently than IL-1α [29]. It must be noted that since IL-1α and IL-1β use the same receptor and trigger identical signalling pathways, differences in the responses evoked by these cytokines are largely a consequence of the context in which they are released rather than a ligand-specific effect. It is also worth considering that while comparing IL-1α and IL-1β potency by injecting the recombinant proteins is a valid approach, these experiments may not mimic the in vitro or in vivo situations where IL-1α is mostly presented as a membrane-bound protein and do not always contemplate the effect of proteolytic activation of IL-1α.
IL-18 is expressed by human LCs and certain DC subsets in mice, displaying autocrine functions [30]. When injected intradermally, IL-18 induces migration of epidermal LCs into the draining lymph nodes, with an activity comparable to that of IL-1β and tumour necrosis factor (TNF)-α [31]. IL-18 stimulates migration of cDCs and plasmacytoid (pDCs) to the lymph nodes. Plasmacytoid DCs (pDCs) are specialised type I interferon (IFN-I)-producing cells and are potent mediators of antiviral immune responses. Besides, IL-18 synergises with IL-12 for Th1-polarisation of murine and human cells [30, 32, 33]. Although both IL-1β and IL-18 synergise with IL-12 for Th1 polarisation [34], it has been proposed that only IL-18 is required for IFN-γ production in vivo [35, 36].
IL-33 is associated with allergic inflammation and Th2 responses. DCs are sensitive to IL-33 stimulation as they express the membrane-bound IL-33 receptor ST2L (also known as IL-1RL1) and produce IL-33 which has autocrine [37, 38] and paracrine actions on immune and nonimmune cells (i.e. mast cells, basophils, innate lymphocytes, NK cells, eosinophils and macrophages [39]). In vitro, IL-33 induces IL-6 secretion and maturation of DCs which can stimulate Th2 responses [37, 40, 41]. In the airways, IL-33 promotes homing and activation of DCs, sensitising to allergic inflammation [40], whereas in the gut, IL-33 promotes tolerogenic DCs and expansion of regulatory T cells [42]. IL-33 also acts on macrophages, inducing M2 polarisation and upregulation of IL-13 in T cells, contributing to Th2 polarisation in vivo [43].
IL-36α, IL-36β and IL-36γ regulate skin inflammation in mice and humans [44-46]. Dermal macrophages, CD1a+ DCs and LCs express high levels of IL36R messenger RNA. Exposure to IL-36β and IL-1β stimulates similar cytokine secretion in LCs but IL-1β is a more potent activator of dermal macrophages than IL-36 [47], suggesting a nonredundant role for IL-1 family members in stimulation of skin APCs subsets. IL-36γ induces IL-18 and IL-12p70 secretion in moDCs which triggers IFN-γ production in T cells in vitro suggesting that IL-36γ can skew T-cell polarisation towards Th1/Tc1 responses. Human pDCs express IL-36R constitutively and at higher levels compared to myeloid DCs [48]. IL-36α and IL-36β can also induce maturation of human moDCs imprinting a Th1 polarising phenotype [49]. IL-36α can synergise with IFN-γ to induce upregulation of Toll-like receptor 4 (TLR4), CD14 and CD11c, facilitating phagocytosis of complement opsonised particles [49]. Murine bone marrow-derived DCs also express IL-36R and respond to IL-36α, IL-36β and IL-36γ by secreting IL-1β, IL-6, IL-12, TNF-α and upregulating maturation markers [50].
IL-37 is the only anti-inflammatory cytokine of the family and also modulates DC activation. IL-37 is absent in mice and its functions in humans have not been fully resolved. As IL-37 isoforms bind to IL-18Rα expressed in mice, a transgenic mouse model of human IL-37b isoform expression (IL-37tg) has been used to explore IL-37 function, showing that overexpression of IL-37 impairs splenic DC activation by lipopolysaccharide (LPS) [51]. IL-37tg mice produced less inflammatory cytokines and were protected against LPS challenge compared to wild-type (WT) mice. Silencing of IL-37 in human peripheral blood mononuclear cells (PBMCs) induced overproduction of IL-1β, IL-6 and TNF-α, whereas IL-37 impaired secretion of IL-1α, IL-1β, TNF-α and IL-8 in human monocytic and epithelial cell lines in the steady state and after stimulation with LPS [51]. Whether IL-37 can be induced by adjuvants remains unknown, but if excessively produced it could potentially have detrimental effects on vaccine priming of local innate responses.
Altogether these studies highlight the importance of IL-1 family members as regulators of APC-induced T-cell responses and potentially in vaccine adjuvant-induced CMI.
The IL-1 family cytokines as regulators of T-cell differentiation and function
The functions of the IL-1 family cytokines in T-cell biology go beyond lineage commitment of naïve T cells. Several members regulate activation and maintenance of differentiated effector and memory T cells. Below we discuss the main roles of IL-1 family cytokines in T-cell differentiation and lineage maintenance.
IL-1 was the first family member linked to T-cell differentiation, particularly Th17 cells. IL-1β synergises with cytokines including IL-6, IL-23, TGF-β and IL-21 to promote differentiation of naive CD4+ T cells into Th17 cells in mice and humans in vitro [52-54], and also in vivo in models of experimental autoimmune encephalomyelitis (EAE), vaccination or colitis [53, 55, 56]. IL-1 can directly act on naive and memory Th1, Th17 and Th2 cells regulating their expansion and survival upon antigen stimulation and enhancing B cell help as reflected by the higher antibody titres observed in mice immunised with ovalbumin and IL-1α or IL-β [57, 58]. It must be noted that some studies failed to demonstrate IL-1R1 expression on Th1 cells [59] despite others showing that Th1 cells respond to IL-1 [57]. Recent studies revealed an unexpected plasticity of Th17 cells which can differentiate into a “nonconventional” subset of Th1 cells upon prolonged exposure to IL-12. This Th17-derived Th1 subset expresses IL-1R1 and responds to IL-1 driving enhanced inflammatory responses in humans and mice [60].
IL-1β directly acts on antigen-specific CD8+ T cells to enhance their expansion, survival, migration and effector function, including expression of granzyme B and IFN-γ secretion [61, 62].
IL-18 seems to be more exclusively linked to the generation of Th1 cell responses. In vitro, IL-18 plus IL-12 promote Th1 differentiation [63]. However, most studies suggest that IL-18 has a more prominent role in CD4+ phenotype stabilisation rather than initial lineage commitment [64]. In vivo, IL-18 is required for induction of optimal CD8+ antiviral responses against influenza in an IL-12-independent manner [65]. Increasing evidence suggests that memory T cells carry out several antigen-independent effector functions and IL-18 has been implicated in the modulation of noncognate memory T-cell functions. During Salmonella infection in mice, DCs sense bacteria via the NLRC4 inflammasome, triggering IL-1β and IL-18 secretion. In this context, IL-18 (but not IL-1β) induces IFN-γ secretion from noncognate CD8+ memory T cells which are important for protection [66]. Similarly, monocyte-derived IL-18 plays a role in early activation of noncognate CD8+ T cells and protection against Listeria monocytogenes [67]. These results reveal an important role for IL-18 in reactivation of memory T cells even in the absence of cognate antigen that could be extremely relevant in vaccine-induced responses.
IL-33 directly influences almost all subsets of Th cells and certain Tc subsets [68]. IL-33 has been mainly linked to Th2 responses as it amplifies production of IL-4, IL-5 and IL-13 from polarised Th2 cells [41, 69] and also acts as a chemoattractant for Th2 cells [70]. While most evidence suggests that ST2L is absent in Th1 cells, it has been suggested that ST2L is transiently expressed in activated Th1 cells in a T-bet/STAT4-dependent fashion, and that ST2L deficiency in virus-specific Th1 cells impairs their function and expansion [71]. Tc1 lymphocytes express ST2L in a T-bet-dependent fashion and respond to IL-33 when combined with IL-12 in vitro by secreting IFN-γ [68]. In line with this, IL-33 contributes to antiviral defences during RNA and DNA viral infections in mice [72]. IL-33 proved to be an effective adjuvant in DNA vaccines encoding tuberculosis or human papillomavirus antigens and IL-33 isoforms. In this setting, coadministration of plasmids encoding IL-33 enhanced antigen-specific IFN-γ responses, polyfunctional CD4+ and CD8+ T cells and immunological memory [73, 74].
In models of colitis, IL-33 induced secretion of IL-9 in T cells [75] and was also shown to promote Treg expansion, Foxp3 and ST2L expression [76]. ST2L signalling in Tregs also ameliorates EAE by increasing Treg frequency which reduces IL-17 and IFN-γ production and pathology [42, 77, 78]. IL-33 also induces Treg infiltration and function in a model of skeletal muscle injury, promoting tissue repair [79]. Hence, although IL-33 can boost Tc1/Th1 responses, IL-33 can also boost Treg function, Th9 and Th2 responses highlighting that context-specific roles of IL-33 should be taken into account for rational adjuvant and vaccine design.
The IL-36 subfamily recently emerged as a regulator of T-cell function and differentiation as IL-36 family members were shown to stimulate secretion of IFN-γ, IL-4, and to a lesser extent IL-17 in cultured splenocytes and activated CD4+ T cells. Recombinant IL-36β acts as an adjuvant, stimulating Th1 responses in vivo [50]. IL-36R is predominantly expressed on naïve CD4+ T cells in mice which also constitutively express IL-36β. All three IL-36 cytokines stimulate proliferation and IL-2 secretion in naïve CD4+ T cells. When combined with IL-12 in vitro, IL-36β induces polarisation of naïve T cells into Th1 cells. Using a model of intravenous challenge with live Bacillus Calmette–Guerin (BCG), Vigne et al. showed that IL-36 cytokines are also required for optimal T-cell responses in vivo, as splenocytes from IL-36R KO mice had impaired secretion of IFN-γ, IL-6 and TNF-α upon ex vivo restimulation [80]. Nevertheless, impaired IL-36R signalling did not affect susceptibility to BCG infection [80, 81].
IL-36β or IL-36γ synergise with IL-12 or IL-2 to induce TCR-independent IFN-γ synthesis in CD8+ T cells [82], whereas IL-36γ can inhibit Tregs while promoting Th9 responses and intestinal inflammation [83]. Induction of IL-36 cytokines during vaccination could be a strategy for promoting Th1 responses.
IL-1 family cytokines and adjuvant modes of action
The diverse effects of IL-1 family members on APC function, T-cell differentiation and induction of cell-mediated immunity during infections or autoimmunity highlights the potential of exploiting these cytokines in the development of improved vaccines. Below we examine what is currently known about the involvement of IL-1 family members in the MOA of clinically approved adjuvants and other promising candidates for induction of CMI.
Adjuvants in licensed vaccines or in vaccines at advanced clinical stages
Alum-based adjuvants
Alum has been the most widely used adjuvant in human vaccines since the 1920s; nevertheless, its MOA remains incompletely understood. Alum drives a Th2-biased immune response accompanied by antibody responses, mainly of the IgG1 type in mice. Several mechanisms have been proposed to contribute to alum's MOA. An initial report by Li et al. put the IL-1 family at centre stage after showing that alum induced IL-1β and IL-18 secretion in a NLRP3-dependent fashion in vitro [84]. Soon, some suggested that NLRP3 and IL-1β were required for alum's adjuvanticity in vivo [85, 86]; however, despite several reports showing that alum induced NLRP3 activation and IL-1β secretion in vitro, the role for IL-1 or NLRP3 in alum's adjuvanticity in vivo has not been confirmed [87-90].
Alum injection induces the release of endogenous danger signals or alarmins from resident or recruited cells that die after encountering alum crystals. These alarmins, including uric acid, ATP, HMGB-1 (high mobility group box 1) protein, host DNA and intracellular cytokines including IL-1α and IL-33 have been proposed to mediate alum's adjuvanticity [8]. DNA released from dying cells was shown to be a major contributor to alum-induced adaptive immunity [91]; however, the cellular source of that DNA remained unidentified. Intraperitoneal alum injection induces neutrophil infiltration [92-94] in an IL-1-dependent and NLRP3-independent fashion, whereas IL-18 is not required for innate cell recruitment after intraperitoneal vaccination with alum [89]. As neutrophil extracellular traps (NETs) are a source of chromatin and other alarmins, it was hypothesised that these could contribute to alum's adjuvanticity [95]. A recent study by Stephen et al. [96] showed that mice deficient in peptidylarginine deiminase 4 (PAD4), the enzyme that catalyses histone hypercitrullination and chromatin decondensation during NET formation [97], had reduced CD4 T cells, B cells and IgG1 titres after alum injection, suggesting that NETs can contribute to the adjuvanticity of alum [96]. However, we found that despite the impaired neutrophil recruitment seen in IL-1R1 deficient mice, both humoral and CD8+ T-cell responses were fully intact after intraperitoneal alum injection [89]. In line with this, other studies have also reported intact humoral and/or cellular responses in mice depleted of neutrophils (using monoclonal antibodies) either after intraperitoneal [93] or intramuscular [98] vaccination with alum. A previous study by Munks et al. showed that several cell types form extracellular traps in response to alum injection which contribute to local formation of nodules composed by chromatin, histones and fibrin. Fibrin deficient mice with impaired nodule formation and reduced local accumulation of extracellular traps exhibited comparable cellular and adaptive responses to wild-types following injection of alum-containing vaccines [91]. Hence, despite the documented contribution of DNA to alum adjuvanticity [91], the role of local neutrophil recruitment and formation of extracellular traps remains controversial. As different vaccination routes were used in these studies, it is possible that the presence of NETs differentially impacts tissue-resident cells and cell recruitment, which may determine the type of local response elicited by the adjuvant.
It is worth noting that IL-1α, IL-1β and IL-18 induce IgG and IgE responses when administered intranasally with an antigen [99]. Moreover, instillation of alum in the lungs induces necrotic death of alveolar macrophages which results in IL-1α release and formation of inducible bronchus-associated lymphoid tissue in the lungs, Th2 responses and IgE production [100]. Thus, it is possible that the IL-1 family cytokines regulate alum-induced responses in a tissue-dependent manner which is dictated by the susceptibility of resident cells to alum-induced necrosis [95]. Alum-containing human vaccines are administered intramuscularly but many mechanistic studies have used alternative routes, hence more research will be required to assess this context-dependent contribution of IL-1 family cytokines to alum-induced humoral and cellular immunity.
A recent study explored the role of IL-33 in alum's adjuvant effect. Elevated IL-33 concentrations were detected locally after mice were injected with alum intraperitoneally. Lack of IL-33 or impaired IL-33 signalling reduced local cytokine and chemokine secretion (IL-5, IL-13, MCP-1, G-CSF) but IL-33 was not required for alum-induced antibody responses [101]. In conclusion, although IL-1 and IL-33 contribute to alum-induced innate responses, they seem to be dispensable for alum-induced IgG responses.
In line with the aforementioned studies, antigen-specific humoral responses were intact in MyD88 KO mice after intraperitoneal vaccination with alum-adsorbed antigens [102, 103]. As MyD88 mediates signal transduction upon engagement of the receptors for IL-1, IL-18, IL-33 and IL-36 (reviewed in [104]), it seems unlikely that these IL-1 family members play a prominent role as positive regulators of alum-induced antibody responses. The alum-containing adjuvant system AS04 incorporates the TLR4 agonist monophosphoryl lipid A (MPLA), a nontoxic derivative of LPS [105]. AS04 enhances antigen-specific T-cell responses compared to alum alone [106]. Intramuscular injection of this adjuvant transiently induces local NF-κB activation, cytokine production, antigen uptake and migration of DCs and monocytes to draining lymph nodes [106]. Although the specific role of IL-1 in the AS04 MOA has not yet been assessed, it is worth noting that contrary to LPS, MPLA poorly stimulates IL-1β secretion, does not induce NLRP3 expression and has been reported to suppresses activation NLRP3 [107]. Hence, it is unlikely that IL-1β plays an important role in AS04 MOA.
Squalene-containing oil-in-water emulsions – MF59, GLA-SE, AS03
MF59 was the first emulsion adjuvant licensed for human use in influenza vaccines (Table 2) after incomplete Freund's adjuvant (IFA) used in polio and influenza vaccines was withdrawn due to severe local reactogenicity [108]. MF59 is a well-tolerated squalene-based emulsion that enhances antibody production and Th1 responses in humans [109] and mixed Th1/Th2 responses in mice [110]. MF59 activates tissue-resident cells, i.e. macrophages and muscle cells, inducing cytokine and chemokine secretion and local innate infiltrates. MF59 triggers ATP release which acts as a danger signal and boosts adjuvanticity [110]. ATP can engage the purinergic P2X7 receptor and activate the NLRP3 inflammasome, inducing release of IL-1β and IL-18 [111, 112]. Nevertheless, NLRP3 is dispensable for MF59 adjuvanticity [113, 114]. It was suggested that the inflammasome adaptor protein ASC, which acts as a bridge between the inflammasome sensor (e.g. NLRP3, NLRC4, AIM2, among others) and caspase 1 (Fig. 1), was needed for MF59-induced antibody responses and cytokine secretion, implying that a different inflammasome could be required for MF59 adjuvanticity [114]. However, it must be noted that this result was not reproduced in a different Pycard−/− (ASC KO) mouse strain [115] and may have been linked to a documented deficiency in ASC KO mice which affects Dock2, a central regulator of the cytoskeleton. Dock2 deficiency impairs lymphocyte migration, pDC function and antigen uptake [116]. Although more research is required to confirm these observations, the results suggest that inflammasome-derived IL-1 and IL-18 or ASC may not be required for MF59 adjuvanticity. Given the important role of purinergic receptor signalling in MF59 adjuvanticity, it is worth noting that exposure of bronchial epithelial cells to rhinoviruses or allergens induces an increase in extracellular ATP, activating purinergic receptor signalling which is linked to IL-33 release [117]. It would be interesting to investigate if MF59-triggered ATP release and P2X7 signalling in the skeletal muscle also results in the release of IL-33 and whether the release of IL-33 could play a role in MF59 adjuvanticity.
Name | Type | Components | IL-1 family cytokines and adjuvant MOA | Vaccines | Administration route | Approval stage | Approval reference |
---|---|---|---|---|---|---|---|
Alum | Mineral salt | Aluminium oxyhydroxide or aluminium hydroxyphosphate | NLRP3-independent IL-1 and IL-33 promote innate recruitment. Not required for adjuvanticity [87-90, 101] | Various | Intramuscular | Licensed (worldwide) | Various |
AS04 | Mineral salt + TLR4 agonist | Alum-adsorbed MPLA | Not fully addressed; NLRP3-derived IL-1 seems dispensable [107] |
Cervarix (HPV bivalent types 16 and 18) Fendrix (HBV) |
Intramuscular |
Licensed EU, USA. Withdrawn USA 2016. Europe |
https://www.medscape.com/viewarticle/870853 [166] |
MF59 | Oil-in-water | Squalene, polysorbate 80, sorbitan trioleate, citrate buffer | NLRP3-derived IL-1 and IL-18 not required [113, 114] | Influenza (many: Fluad, Aflunov, Focetria, Celtura) | Intramuscular | Licensed (EU and or worldwide excluding USA) | [108] |
AS03 | Oil-in-water | Squalene, α-tocopherol, polysorbate 80, phosphate-buffered saline | ND | Influenza (many: Prepandrix, Pumarix, Pandemrix, Arepanrix) | Intramuscular | Licensed (EU and or worldwide excluding USA) | [108] |
AS02 | Oil-in-water | AS03 + MPLA and QS-21 | NLRP3 shown to negatively regulate QS-21 adjuvanticity [122]. IL-1 family cytokine role in AS02 MOA remains unknown. |
HBVAXPRO (HBV) FMP2.1/AS02A and RTS,S/AS02A (malaria) Mtb72F/AS02A (Tuberculosis) NefTat/gp120/AS02A (HIV) |
Intramuscular |
Phase III Phase II Phase II Phase I |
NCT00291954/NCT00480116 NCT00460525/NCT00197054 NCT00146744 NCT00027365/NCT00117429 |
AF03 | Oil-in-water | Squalene, polyoxyethylene cetyl-stearylether, sorbitan oleate, mannitol, phosphate-buffered saline | ND | Humenza (influenza) | Intramuscular | Approved by European Medicines Agency in 2010; withdrawn in 2011 | [108] |
GLA-SE | Oil-in-water | Glucopyranosyl lipid A in a stable squalene emulsion | NLRP3-independent secretion of IL-18 required for IFN-γ and CD8+ T cells [118] |
ID93 + GLA-SE (tuberculosis) rSm14/GLA-SE (Schistosomiasis) LEP-F1 + GLA-SE (leprosy) PanBlok (Pandemic Influenza) |
Intramuscular |
Phase IIa (tuberculosis) Phase II (Schistosomiasis) Phase I (Leprosy) Phase I (influenza) |
NCT02465216 NCT03041766 NCT03302897 NCT01612000 |
Montanide ISA-51 | Water-in-oil | Squalene, mannide monooleate | ND | CIMAvax-EGF (lung cancer, therapeutic) | Intramuscular/subcutaneous | Licensed Cuba | [108] |
AS01 adjuvants | Liposomes | Cholesterol, QS-21 and MPLA | ND |
Mosquirix (malaria, RTS,S/AS01) and Shingrix (Hz/su, shingles) |
Intramuscular |
Mosquirix: approved by European Medicines Agency Shingrix: Phase III |
NCT02690207, NCT02735915 |
CAF01 | Liposomes (cationic) | Two-component liposomal suspension with DDA and TDB | IL-1R signalling required for IL-17 and IFN-γ production but not for antibodies [126] | Ag85B-ESAT-6:CAF01 | Intramuscular | Phase I | NCT00922363 |
IC31 | Cationic particles | 11-mer antibacterial peptide KLKL(5)KLK and synthetic immunostimulatory oligodeoxynucleotide | ND |
H1/H56: IC31 SSI, Valneva, Aeras and H4:IC31 Sanofi Pasteur, SSI, Aeras (both tuberculosis) |
Intramuscular | Phase IIa | http://www.aeras.org/img/uploads/attachments/Tait_Stakeholder_102016_-_2.pdf |
ISCOMATRIX | Self-assembled nanoparticles | Quillaja saponins, cholesterol and phospholipids at a molar ratio ~ 1 : 1 : 1 | NLRP3, NLRC4-independent IL-18 required for IFN-γ and IgG2c [124] |
NY-ESO-1:ISCOMATRIX (melanoma) V950:ISCOMATRIX (Alzheimer) Octavalent HPV:ISCOMATRIX V180:ISCOMATRIX (dengue) |
Intramuscular |
Phase II Phase I Phase I Phase I |
NCT00518206 NCT00464334 NCT00851643 NCT01477580 |
- MPLA, 3-O-desacyl-4ʹ-monophosphoryl lipid A; TLR, Toll-like receptor; AS, adjuvant system; AF, adjuvant formulation; EGF, epidermal growth factor; QS, Quillaja saponaria, HPV, human papilloma virus; TDB, α,α’-trehalose dibehenate; DDA, N,N’-dimethyl-N,N’-dioctadecylammonium bromide; HCS, Hepatitis C virus; HBV, Hepatitis B virus; ND, not determined.
The adjuvant GLA-SE consists of the synthetic TLR4 agonist glucopyranosyl lipid A in a stable squalene-in-water emulsion [118]. GLA-SE is in Phase I clinical trials for vaccines for influenza, leishmaniasis, schistosomiasis, malaria and tuberculosis, showing robust enhancement of Th1 and balanced IgG1/IgG2 responses in humans [108]. Induction of Th1 responses by GLA-SE requires MyD88, TRIF, T-bet and IL-12 signalling in addition to IFN-I supporting IFN-γ secretion by memory CD8+ T cells and NK cells [119, 120]. Regarding involvement of IL-1 family cytokines in its MOA, the injection of GLA-SE into the footpad of mice induced IL-18 in the local draining lymph nodes, stimulating noncognate CD8+ T cells with a memory phenotype which secreted IFN-γ and enhanced the activation marker CD69 [118]. The same study showed that GLA-SE also promotes migration of IFN-γ+ neutrophils carrying the vaccine antigen into the lymph nodes of mice where they colocalise with T cells. NLRP3 is dispensable for GLA-SE adjuvanticity, and while IL-18R1 and caspase1/11 are not required for cell recruitment, they regulate the expression of IFN-γ [118]. These results underscore the role of IL-18 in the GLA-SE MOA and highlight the contribution of neutrophils in IL-18-induced IFN-γ production after vaccination with antigen and GLA-SE. It remains to be determined if neutrophil-derived IFN-γ is required for GLA-SE adjuvant effect and how inflammasomes, other than NLRP3, contribute to IL-18 activation and GLA-SE adjuvanticity.
Regarding the role of IL-1 family cytokines in the adjuvanticity of AS03, very little is known. AS03 injection induces early and transient expression of inflammatory cytokines including IL-1β in the muscle [121] but the role of IL-1β in the adjuvanticity of AS03 has not yet been explored.
QuilA and QS-21-containing adjuvants: AS02, AS01, ISCOM/ISCOMATRIX
Saponins, derived from the soapbark tree Quillaja saponaria are amphipathic triterpene glycosides with intrinsic adjuvanticity. QuilA®, a mixture of Quillaja derived saponins has been used in veterinary vaccines and induces potent humoral and cellular responses. Saponins can be quite toxic as they are haemolytic. QS-21, a fraction of QuilA® exhibits robust adjuvanticity and low toxicity when formulated in cholesterol-containing liposomes [122]. Saponin-containing adjuvants including AS01 (cholesterol-containing liposomes, QS-21 and MPLA) and AS02 (Squalene oil-in-water emulsion AS03 plus MPLA and QS-21) have been licensed in human vaccines or are at advanced clinical development stages (see Table 2). The ISCOMATRIX™/Matrix-M™ adjuvants are self-assembling 40 nm particles containing QuilA, cholesterol and phospholipids in phase I and phase II trials (Table 2).
QuilA induces IL-1β release in human PBMCs and murine macrophages (requiring NLRP3 in the latter) [86]. The adjuvant QS-21 promotes antigen-specific antibody responses, Th1 and cytotoxic CD8+ T cells in primates and mice; however, CD8+ responses are normally weak in humans vaccinated with the QS-21 containing adjuvant AS01 and a subunit antigens [123]. In mice, QS-21 stimulates innate immunity by enhancing NK cell function and inflammatory cytokine release and induces NLRP3-dependent release of IL-1β and IL-18 in murine DCs and macrophages in vitro [122]. Surprisingly, NLRP3 was reported to negatively regulate QS-21-induced humoral and cellular responses in vivo as NLRP3-deficient mice immunised intramuscularly with HIV-1 gp120 and QS-21 had significantly higher Th1 and Th2 responses and increased IgG1 and IgG2c titres compared to WT controls [122]. The molecular basis of this finding remains unknown and the role of IL-1 in QS-21 adjuvanticity in vivo remains to be assessed.
The saponin-containing adjuvant ISCOMATRIX™ (IMX) induces humoral responses and shows an unparalleled ability to induce Ag-specific CD8+ T cells in humans compared to other saponin-containing adjuvants such as AS01 or AS02 [124]. In mice, IMX promotes early activation of NK cells and IFN-γ production and induces NLRP3-dependent IL-1β and IL-18 secretion in DCs in vitro. However, only IL-18 is required in vivo for IMX-induced IFN-γ production, NK cell activation, priming of antigen-specific CD8+ T cells and antibody class switching to IgG2c, which is independent of NLRP3 or NLRC4. Interestingly, ASC-deficient mice had a partial impairment in NK cell activation and IFN-γ production [124]. Based on these results, the authors concluded that inflammasome-dependent and independent sources of IL-18 may contribute to IMX adjuvant effects, but as mentioned before, this conclusion must be interpreted with care as ASC KO mice may also have a defect in Dock2. In any case, the requirement of IL-18 for IMX adjuvanticity seems clear although inflammasomes may be dispensable for its activation. IMX induces cell death [124] which may contribute to release of intracellular pro-IL-18. As pro-IL-18 can be activated by proteases such as granzyme B produced by activated NK cells, inflammasomes may not be essential for its activation.
Overall, the current evidence suggests that Quillaja-derived saponins presented in the form of small nanoparticles can induce release of IL-18 which is essential for the adjuvant activity of ISCOMATRIX adjuvants. Regarding the role of the licensed human adjuvants AS01 and AS02 that contain QS-21, the role of NLRP3, IL-1 and other IL-1 family cytokines remains to be determined.
Cationic particulates: CAF01 and IC31
The cationic adjuvants CAF01 (Statens Serum Institut) and IC31 (Valneva Technologies) are two promising candidates for tuberculosis vaccines in clinical phase I and phase IIa trials, respectively. Both consist of cationic particulates. CAF01 is composed of cationic liposomes containing N,N’-dimethyl-N,N’-dioctadecylammonium bromide (DDB) and the immunostimulatory α,α’-trehalose dibehenate (TDB), a synthetic analogue of the mycobacterial cord factor trehalose-6,6-dimycolate (TDM) a Mincle agonist [3, 125]. IC31 contains an immunostimulatory oligodeoxynucleotide ODN1a (a TLR9 agonist) and an 11-mer antibacterial peptide (KLK) which spontaneously forms particles when combined with antigens [125]. Both adjuvants are potent stimulators of T-cell responses, with CAF01 triggering Th1/Th17 responses, while IC31 favours Th1 responses, although generally these two adjuvants induce lower antibody titres when compared with MF59 or alum in mice immunised subcutaneously [125].
The role of the IL-1 family cytokines in the MOA of CAF01 and IC31 has not been explored in detail. Indeed to date, there are no reports assessing the involvement of IL-1 family cytokines in IC31 adjuvanticity and only one report has evaluated the contribution of one the components of CAF01, TDB on the induction of some IL-1 family members [126]. The C-type lectin receptor Mincle was previously identified as the receptor for TDM and the synthetic analogue TDB included in CAF01, which triggers FcRc-Syk-Card9 signalling, APC activation and adjuvanticity [127]. Although MyD88 was not required for stimulation of APCs by CAF01 in vitro, MyD88 KO mice showed impaired antigen-specific IFN-γ and IL-17 production as well as lower IgG2a titres after vaccination with CAF01 and the tuberculosis antigen H1. The MyD88 requirement was not linked to TLR2, TLR3, TLR4, TLR9 or TLR7 signalling as TLR2/3/4/7 quadruple knockouts and TLR9 KO responded normally to immunisation with TDB. As TDB induces the expression and secretion of IL-1β in macrophages in vitro and MyD88 is required for IL-1, IL-18 and IL-33 signalling, the authors analysed the contribution of these IL-1 family cytokines in adjuvanticity of CAF01. Interestingly, IL-1R1-signalling deficiency or administration of an IL-1 receptor antagonist abrogated IL-17 and IFN-γ production in vaccinated mice but did not affect antibody production. IL-18 and ST2 were not required for adjuvant activity although less inflammation was observed at the injection site in IL-18 KO mice injected with CAF01 compared to WT controls [126]. The defect seen in MyD88-deficient mice was probably linked to impaired IL-1R signalling, which is necessary for CAF01 adjuvanticity.
IL-1 family cytokines and chitosan
The cationic polysaccharide chitosan derived from partial deacetylation of chitin (a major component of crustacean and arthropod exoskeletons) has been approved for human use in cosmetics, food supplements, bandages and other biomedical applications and holds great adjuvant potential. We and others showed that chitosan induces inflammasome-dependent IL-1β secretion in DCs and macrophages in vitro [128-130]. When combined with the TLR9 agonist CpG, chitosan induced potent NLRP3-dependent Th1/Th17 responses in vivo. In addition, NLRP3- and IL-1R1-deficient mice show reduced antigen-specific IgG2b and IgG2c titres following vaccination with chitosan-CpG [129]. We have recently found that chitosan induces mitochondrial stress leading to release of DNA into the cytosol, triggering the DNA sensing cGAS-STING pathway. This leads to IFN-I secretion and DC maturation, stimulating Th1 responses, IFN-γ production, antibody secretion and class switching [130]. Although IFN-I have been shown to suppress NLRP3 inflammasome and IL-1β activation by caspase-1 via a mechanism involving STAT-1 [131], we showed that NLRP3 is required for chitosan-induced Th1 response as this response was abrogated in NLRP3 KO mice. It is worth noting that antibody responses were intact in NLRP3 KO mice [130]. This highlights the need to evaluate both the humoral and cellular responses when determining the MOA and signalling pathways involved in the adjuvanticity of different formulations.
Conclusion
Growing evidence implicates members of the IL-1 cytokine family in the differentiation and function of Th1, Th17 and cytotoxic T cell responses, yet this effect seems largely context-dependent. For example, while IL-1 enhances the survival, expansion and effector function of antigen-specific CD8+ T cells in vivo during L. monocytogenes or vaccinia virus infection, or in a tumour model [61, 62], IL-1 can promote Th2 responses when administered intranasally [132]. Similarly, IL-18 can support both Th1 or Th2 responses depending on the availability of IL-12 [133]. IL-33 has been mainly linked to amplification of Th2 responses [41, 69] particularly in the airways [40], but in the gut, IL-33 promotes regulatory T cells [42], whereas when IL-33 is combined with IL-12, Tc1 lymphocytes expressing ST2L respond by secreting IFN-γ [68]. IL-33 also contributes to antiviral immunity in vivo [72] and enhances antigen-specific IFN-γ responses, polyfunctional CD4+ and CD8+ T cells and immunological memory when used as an adjuvant in DNA vaccines [73, 74]. The IL-36 subfamily has also emerged as a prominent driver of Th1 responses and IFN-γ being most prominent in the skin [134]. In this regard, it is important to note that the injection route plays a critical role in determining which IL-1 family members are being released in response to an adjuvant as the expression patterns of IL-1 family members vary depending the tissue and recruited cell types. At the same time, the expression pattern of cytokine receptors will influence the downstream response. Furthermore, as most IL-1 family members are subject to processing by proteases [16], the proteases released by resident or recruited cells at the injection site may also play a key role in determining the activation or inactivation of different IL-1 family cytokines [97].
Clinical and experimental evidence indicates that most adjuvants cause local irritation and some degree of cytotoxicity stimulating DAMP release which may act as endogenous immune enhancers for adjacent viable cells at the injection site (Fig. 3). The DAMPs released in response to adjuvants are structurally diverse, ranging from extracellular ATP and uric acid to chromatin or mitochondrial DNA [91, 109, 110, 122, 130]. While these molecules may be immunostimulatory in certain contexts, research suggests that the release of preformed DAMPs per se may not always be sufficient to elicit an effective immune response. This was illustrated by studies where cells undergoing accidental necrosis after freeze-thawing cycles or boiling were unable to activate DCs in vitro or to elicit protective immunity against tumours in vivo [117, 135, 136]. In contrast, preactivation of cells prior to induction of cell death by γ-irradiation resulted in activation of DCs and promoted T-cell proliferation in vitro, an effect not seen when the same cells were killed without prior activation [137]. This implies that the expression and/or activation of canonical DAMPs must be initiated and precisely timed, before cell death is induced to trigger effective adaptive responses. As elegantly presented in a recent review [97], the IL-1 family members have all the characteristics of canonical DAMPs which can be induced when danger or damage is detected and released by dying cells. IL-1 family cytokines are tightly regulated, signal though very specific receptors that share signalling components with innate receptors (namely TLRs), and like TLRs, activation of the IL-1 family receptors links robust innate responses to adaptive immunity in the context of infection and autoimmunity [97]. Therefore, the coordinated release of IL-1 family members with an array of preformed DAMPs in response to adjuvant-induced cell death may be a key factor acting as a tipping-point in the transition from inflammation to the induction of an adaptive response (Fig. 3).

Adjuvants induce cell death by different mechanisms: alum activates lysosomal-mediated necrosis in macrophages [138, 139] and netosis in neutrophils and mast cells [93, 96]; ISCOMATRIX also triggers lysosomal destabilisation and necrosis in macrophages [124], whereas MF59 elicits apoptotic-like death in macrophages [109] and saponins and surfactant-containing adjuvants induce caspase-independent cytolysis [122, 140, 141]. As most IL-1 family cytokines lack conventional secretion signals, their release is largely linked to some form of cell death (Table 1) [15]. The type of cell death differentially regulates the release and activity of IL-1 family cytokines and while necrotic cell death promotes the release of IL-33 [39], activation of the proapoptotic caspases 3 and 7 inactivates IL-33 [16]. On the other hand, IL-36γ released from keratinocytes requires a cell death mechanism that involves activation of caspase-1 which induces IL36G gene expression and caspases 3/7 for IL-36γ secretion [142]. While inflammasome and caspase 1/11-mediated pyroptosis is possibly the best characterised mechanism linking cell death with maturation and release of IL-1β and IL-18 [143], the inflammasome does not seem to play a major role during adjuvant-induced release of IL-1 or IL-18 in vivo suggesting that other mechanisms may be more relevant. Recently, necroptosis and components of the necroptosome have been shown to regulate IL-1α and IL-1β secretion in vitro and during inflammatory autoimmune conditions [144, 145]; however, it remains to be determined if adjuvants can induce necroptosis and if this may impact adjuvanticity. Hence, the contribution of the IL-1 family members to adjuvanticity may be directly linked to the susceptibility of resident cells to adjuvant-induced cell death and studying how cell death triggered by adjuvants regulates IL-1 family members in vivo may contribute to unveiling the adjuvants MOA.
Most vaccines are administered into the skeletal muscle, a “nonclassical” immune-privileged tissue that despite not being separated by the typical anatomical barriers encountered in sites such as the brain and the eye, presents few immune cells under homeostasis and supports immunomodulatory responses by recruiting alternatively activated macrophages and Tregs in response to sterile inflammation, which favour regeneration and limit autoimmunity [137]. However, many studies addressing the adjuvants MOA utilise other immunisation routes that may not be representative of the real environment the adjuvant will encounter when injected in patients.
It is interesting to note that IL-1 family cytokines act as strong adjuvants when administered mucosally. For example, intranasal administration of IL-18 with influenza hemagglutinin (HA) induced the secretion of Th1 and Th2 cytokines, increased serum IgG levels and mucosal IgA [146], while intranasal immunisation of mice with IL-1α and a HIV antigen induced CTL responses and antigen-specific IFN-γ secretion [147]. Therefore, the role of IL-1 family cytokines in the MOA of mucosal adjuvants deserves more attention, as the induction of these cytokines could be the key to development of much needed mucosal adjuvants.
The intricate immune mechanisms triggered by administration of an adjuvant and the complex biology of the IL-1 family cytokines highlight the importance of studying adjuvant-induced responses in an appropriate context, one that can more closely reflect the route of vaccination in a clinical setting.
Acknowledgments
We sincerely thank Dr Michael Carty for his critical reading of the review and valuable discussions. This work has been supported by the Irish Research Council GOIPD-2017-1257 (NMW) and Science Foundation Ireland Investigator Award 12/1A/1421 (ECL). We apologise to the authors whose work could not be cited due to space limitations.