Targeting interleukin-17 in chronic inflammatory disease: A clinical perspective

IL-17A, the founding member of the IL-17 family, was first discovered in mice in 1993 (Rouvier et al., 1993) and 2 yr later in humans (Yao et al., 1995b). The IL-17 family is evolutionary conserved within vertebrates and composes six structurally related members in mammals, IL-17A, IL-17B, IL-17C, IL-17D, IL-17E, and IL-17F (Malagoli, 2016). All cytokines bind to their receptors as disulfide-linked homodimers, with the exception of IL-17A and IL-17F, which can also form heterodimers (Wright et al., 2007; Fig. 1).

Five IL-17 receptor subunits have been identified (IL-17RA, IL-17RB, IL-17RC, IL-17RD, and IL-17RE). All are single-pass transmembrane receptors with two extracellular fibronectin II–like domains and a cytoplasmatic “SEFIR” motif for activation of downstream signaling pathways (McGeachy et al., 2019). The receptor for IL-17A and IL-I7F homodimers and IL-17A/F heterodimers consists of a heterodimeric receptor complex composed of IL-17RA and IL-17RC subunits. IL-17RA can also pair with IL-17RE or IL-17RB, which then bind IL-17C and IL-17E, respectively. IL-17B has been shown to engage with IL-17RB, but a coreceptor is still missing, as well as the cognate receptors for IL-17D and a cytokine that binds to IL-17RD.

IL-17R signal transduction has been best studied for IL-17RA/IL-17RC complexes after IL-17A ligation (Gu et al., 2013). Receptor engagement leads to a conformational change, which allows the adaptor protein Act1 to form a homotypic interaction between the receptor’s “SEFIR” domain and its own. The E3 ligase activity of Act1 initiates the canonical pathway via ubiquitylation of TNF receptor-associated factor 6 (TRAF6), recruitment of TAK1 and subsequent activation of the canonical NF-κB and MAPK pathways including ERK, p38, and JNK, as well as the CCAAT-enhancer-binding proteins pathway (Monin and Gaffen, 2018). An alternative pathway depends on the phosphorylation of Act1 by inducible IκB kinase, which leads to the assembly of TRAF2 and TRAF5 with mRNA stabilizing factor human antigen R, thereby increasing stability of multiple mRNAs (Amatya et al., 2017). IL-17A and F lie in close proximity in the genome on chromosome 1 in mice and chromosome 6 in humans, share the highest structural identity among all IL-17 family members, and have been shown to have largely overlapping biological activities (Akimzhanov et al., 2007; Monin and Gaffen, 2018). Hereafter, we focus on IL-17A/F and their role in chronic inflammatory disorders and use IL-17 as shorthand.

Mosmann et al. (1986) were the first to propose a bifurcation in CD4⁺ Th cells. Th1 cells were defined as mediators of “cellular immunity” directed against intracellular pathogens and were characterized mainly by their production of IL-2 and IFN-γ, but also other cytokines such as GM-CSF and TNF. Th2 cells were described as IL-4, IL-5, IL-13, and IL-10–secreting cells involved in “humoral immunity” directed against extracellular pathogens (reviewed in Raphael et al., 2015). The discovery of IL-17 initially did not influence the Th1/Th2 paradigm prevailing in the field at this time even though numerous reports indicated that it could be involved in several inflammatory disorders (Kotake et al., 1999; Kostulas et al., 1999; Antonysamy et al., 1999; Chabaud et al., 1998, 1999). Langrish et al. (2005) found that among many other mediators, IL-17A is produced by Th cells upon engagement of their IL-23R. In vitro, IL-17 secretion could be elicited from naive CD4 T cells upon polarization with TGF-β and IL-6 (Veldhoen et al., 2006). In humans, IL-6 and IL-1β and IL-21 were later also identified as promoters of the IL-17 expression in Th cells (reviewed in Sallusto et al., 2012). IL-17 expression is driven by the transcription factor RORγt (via a STAT3-dependent mechanism; Ivanov et al., 2006). The term Th17 cell was coined, and for some time Th17 polarization and their role in immunity became the core interest of the immunology community. The discovery and characterization of Th17 cells are reviewed in more detail elsewhere (Korn et al., 2009; McGeachy and Cua, 2008). Over the years, it crystallized that the commonly called “IL-23/IL-17 axis” is critically involved in the regulation of tissue inflammation, in the context of both host defense and chronic inflammatory disorders.

In addition to Th17 cells, innate cells located at epithelial and mucosal barriers such as CD27⁻ γδ T cells, NK(T) cells, and innate lymphoid cells (ILCs) emerged as potent IL-17 producers (Happel et al., 2003; Ferretti et al., 2003; Zheng et al., 2008; Fig. 2). Their constitutive IL-23R and RORγt expression allows them to respond quickly (within hours) to injury or pathogenic triggers (reviewed in Cua and Tato, 2010). Also, other leukocytes such as CD8 T cells and B cells and even Paneth cells have been reported as cellular sources of IL-17 (He et al., 2006; Schlegel et al., 2013; Vazquez-Tello et al., 2012; Takahashi et al., 2008; Fig. 2). Whether myeloid cells such as neutrophils or macrophages produce IL-17 is still controversially discussed (see Fig. 2).

IL-17 is a pleiotropic cytokine that acts mainly on cells of mesenchymal origin such as endothelial cells, epithelial cells, and fibroblasts but has also effects on immune cells including dendritic cells and macrophages (reviewed inXu and Cao, 2010; Fig. 2). IL-17 executes key functions to prevent pathogen invasion by the secretion of proinflammatory cytokines, chemokines, matrix metalloproteinases, growth factors, and antimicrobial peptides (reviewed in Onishi and Gaffen, 2010). Hereby IL-17 often acts in synergy with other mediators such as TNF and IL-22 (Shen et al., 2005;Albanesi et al., 1999). Conversely, excessive IL-17 production contributes via similar mechanisms to pathogenic inflammation.

Most of our understanding of IL-17 and its role in inflammatory disorders was gained through various preclinical models of disease and has prompted the development of new drugs targeting IL-17. Several drugs were designed to block the IL-17 pathway either selectively or in combination with other inflammatory cytokines, or cytokines involved in the maintenance of IL-17 producing cells such as IL-23. Table 1 summarizes the therapeutic agents that are already approved or in clinical development. In the following section, we will discuss preclinical experimental findings about the role of IL-17 in prototypic chronic inflammatory disorders and compare the results to data from clinical trials.

Psoriasis is a chronic immune-mediated skin disease, characterized by the proliferation of keratinocytes and infiltration of T cells (Nikaein et al., 1991). In xenograft models of psoriasis, nonlesional skin from psoriasis patients is transplanted onto immunocompromised mice (Boehncke et al., 1994; Boyman et al., 2004). The model uncovered a pathogenic role of IL-23 (Tonel et al., 2010), IFN-α (Nestle et al., 2005), and TNF (Boyman et al., 2004) in psoriasis and paved the path for the successful introduction of TNF inhibitors as therapeutic options for psoriasis vulgaris (Leonardi et al., 2003). In 2009,van der Fits et al. (2009) introduced the Aldara (5% Imiquimod)-model of psoriasis, which very closely mimics human psoriasiform lesions. We and others then found that psoriasiform inflammation greatly depends on the IL-23/IL-17 axis (Pantelyushin et al., 2012; Cai et al., 2011). Skin of Aldara-treated mice expressed high levels of IL-17A/F (Pantelyushin et al., 2012; Cai et al., 2011) and IL-22 (Pantelyushin et al., 2012). In line with this, it was reported that IL-17A, IL-22, and IL-23 are up-regulated in the serum of patients with active psoriasis, and increased levels of IL-17A, F, and C were found in psoriatic skin lesions (Johansen et al., 2009; Fotiadou et al., 2015). Surprisingly, γδ T cells and not Th17 cells were found to be the primary source of IL-17A/F in the Aldara-mouse model, and their ablation led to disease protection (Cai et al., 2011; Pantelyushin et al., 2012). Similarly, increased numbers of IL-17 producing γδ T cells were also found in the skin of psoriasis patients (Laggner et al., 2011). There is some debate as to which cells in developing lesions initially deliver IL-17 as CD8⁺ IL-17 producing cells were also identified in psoriatic lesions (Ortega et al., 2009). The disease-promoting role of IL-17 in psoriasiform inflammation was shown to be mainly mediated via the activation of epithelial cells and keratinocytes and their increased secretion of antimicrobial peptides, cytokines, and chemokines mobilizing the recruitment of neutrophils (reviewed in Perera et al., 2012).

After the identification of the IL-23/IL-17 axis as a key factor in the pathogenesis of psoriasis, Ustekinumab, the first p40 inhibitor targeting both IL-12 and IL-23, was introduced to the clinic with success (Papp et al., 2008). However, we found that only IL-23 is disease-promoting in psoriasiform inflammation, whereas IL-12 has a protective role (Kulig et al., 2016), suggesting that collateral neutralization of IL-12 instead of IL-23 alone limited the effectiveness of these drugs.

Recently, the first three selective IL-23 inhibitors targeting the IL-23p19 subunit were licensed for treatment of plaque psoriasis after successful completion of several phase III clinical trials: Tildrakizumab, Risankizumab, and Guselkumab (Reich et al., 2017; Gordon et al., 2018; Blauvelt et al., 2017b). IL-23p19 inhibitors were more efficient in improving skin lesions than TNF inhibitors (Reich et al., 2017) and in line with findings from the Aldara-model of psoriasiform inflammation, superior compared with IL-12/IL-23 inhibitor Ustekinumab (Gordon et al., 2018). The selective IL-23 inhibitor Guselkumab was also effective in psoriasis patients where combined IL-12/IL-23 inhibition had failed (Langley et al., 2018), emphasizing that sparing IL-12 is necessary for resolution of skin inflammation in some psoriasis patients.

Between 2015 and 2017, three IL-17 pathway inhibitors reached the market. Secukinumab and Ixekizumab are monoclonal antibodies specific for IL-17A and thus inhibit both IL-17A homodimers as well as IL-17A/F heterodimers (Liu et al., 2016), while IL-17F homodimers remain unaffected. In contrast, the monoclonal antibody Brodalumab targets IL-17RA, and thus blocks the activity of IL-17A, IL-17F, IL-17C, and IL-17E. All three biologicals are approved by the Food and Drug Administration and European Medicine Agency for the treatment of moderate to severe plaque psoriasis. Several pivotal phase III studies demonstrated rapid and sustained skin improvement in the treated patients compared with placebo (Langley et al., 2014;Gordon et al., 2016; Lebwohl et al., 2015). Depending on the design of the studies, up to 42% of patients who received one of the IL-17 pathway inhibitors achieved disease-free skin (Papp et al., 2016). Clinical responses are paralleled by decreased IL-17A, IL-17F, and IL-22 expression and downmodulation of leukocyte infiltration-promoting chemoattractants in the skin (Kolbinger et al., 2017).

IL-17 inhibitors showed superior efficacy in skin clearance than previously approved psoriasis therapy with IL-12/IL-23 inhibitor Ustekinumab (Blauvelt et al., 2017c; Paul et al., 2019; Lebwohl et al., 2015) or TNF inhibitor Etanercept (Griffiths et al., 2015; Langley et al., 2014). In addition, Ixekizumab demonstrated efficacy in Etanercept nonresponders (Blauvelt et al., 2017a) and Brodalumab in Ustekinumab nonresponders (Langley et al., 2014). Head-to-head trials comparing the efficacy between the available IL-17 and IL-23p19 inhibitors have not yet been performed, but a recent meta-analysis of 77 randomized controlled trials failed to detect a difference between the drug classes, as Brodalumab, Ixekizumab, Guselkumab, and Risankizumab were equally efficient to achieve skin improvement (Sawyer et al., 2019).

Besides its presence in diseased skin, IL-17 is also observed in healthy skin and has a nonredundant role in tissue protection and immunity against fungi (Naik et al., 2015, 2017; Linehan et al., 2018; Okada et al., 2016; Sparber et al., 2019). In spite of this, the safety profile of IL-17 pathway inhibitors is favorable over long-term treatment with a consistent spectrum of adverse events across all three therapies (Deodhar et al., 2019; Langley et al., 2019; Lebwohl et al., 2015). Upper respiratory tract infections occur more frequently, followed by a higher susceptibility to mucocutaneous Candida spp. infections (Saunte et al., 2017).

We speculate that IL-17 production in cells associated with homeostasis in the skin is largely independent of IL-23R signaling, suggesting that the blockade of IL-23 would trigger fewer adverse effects compared with the general inhibition of IL-17, which may serve a function in homeostatic barrier maintenance and the control of the commensal flora. Indeed, the risk of Candida infections was lower when targeting IL-23p19 alone compared with studies with IL-17 inhibitors (Blauvelt et al., 2018). Whether IL-23 may be required for homeostatic IL-22 remains unclear.

IBDs, which comprise of Crohn’s disease and ulcerative colitis, are inflammatory conditions of the colon and small intestine (Abraham and Cho, 2009). Single nucleotide polymorphisms in IL23R are associated with the susceptibility to IBD, and both increased IL-23 and IL-17 levels were found in colon biopsies of IBD patients, especially during active disease (Fujino et al., 2003; Duerr et al., 2006; Hölttä et al., 2008).

Also in mouse models of IBD, IL-23 was clearly associated with colitis, and as for psoriasis and experimental autoimmune encephalomyelitis (EAE), an animal model for multiple sclerosis (MS), IL-23 and not IL-12 is critically driving the disease (Buonocore et al., 2010; Ahern et al., 2010;Yen et al., 2006; Hue et al., 2006; Elson et al., 2007). In contrast, targeting IL-17 or its receptors led to controversial results in different animal models of colitis. Already in 2004, Ogawa et al. (2004) showed that neutralization of IL-17 leads to increased disease severity in the dextrane sulfate sodium (DSS) model. Several studies confirmed this result in the DSS as well as in other models of IBD (O’Connor et al., 2009; Song et al., 2015; Lee et al., 2015;Maxwell et al., 2015) and are in conflict with the few reports showing a disease promoting role of IL-17 (Feng et al., 2011; Leppkes et al., 2009; Yen et al., 2006). Despite the description of a protective role of IL-17 in different murine models of colitis, IL-17A (Hueber et al., 2012) and IL-17RA inhibitors (Targan et al., 2016) were tested for therapeutic efficacy in patients with Crohn’s disease. The treatment failed and the studies had to be terminated due to exacerbation of colitis. In fact, it was already proposed in the early 2000s that IL-17 has an important role in the maintenance of barrier properties of epithelial tissues in vitro and later also in vivo (Kinugasa et al., 2000; Lee et al., 2015). Two independent reports using different colitis models demonstrated that IL-17A increases the barrier integrity in epithelial cells and demonstrated that tissue-protective IL-17 is produced by γδ T cells in the absence of IL-23 (even though they do express IL-23R; Lee et al., 2015; Maxwell et al., 2015).

Interestingly, new-onset IBD is a rare adverse event in psoriasis patients treated with IL-17 inhibitors (Schreiber et al., 2019), and mucosal inflammation is also not part of the disease spectrum in patients with inborn errors of the IL-17 pathway (Okada et al., 2016). However, during an established mucosal inflammation like in Crohn’s disease, the beneficial function of IL-17 on the integrity of the epithelial barrier might prevail (Maxwell et al., 2015).

While IL-17 inhibition resulted in alarming disease exacerbation, IL-23 inhibition has revealed promising results in clinical trials for Crohn’s disease (Feagan et al., 2017, 2018; Sands et al., 2017). These clinical data suggest that selective targeting of IL-23 allows the IL-23–independent IL-17 production in the gut needed to keep the epithelial barrier intact, while IL-23–driven disease-promoting IL-17–producing cells decrease, mirroring results from previous animal studies (Lee et al., 2015).

Given the critical role of IL-23 in the maintenance of disease-promoting IL-17–producing cells, selective IL-23 class inhibitors hold the promise to interfere especially with the development of these pathogenic IL-17–producing cells while sparing the IL-23–independent IL-17–producing cells involved in host defense againstCandida and maintenance of the gut barrier. Several reports indicated that the protective effect of IL-17 could be mediated via the control of the commensal microflora (Tang et al., 2018; Song et al., 2015; Kumar et al., 2016), which were already earlier shown to induce IL-17 in the colon (Ivanov et al., 2009). Interestingly, disrupted IL-17R signaling in gut epithelial cells and the following dysbiosis was also connected to a higher predisposition to neuroinflammation, indicating that homeostatic IL-17R signaling in the gut is vital for a normal gut microflora, which in turn can control the autoimmune potential of lymphocytes (Kumar et al., 2016). Furthermore, it was shown that only deletion of IL-17F but not IL-17A prevents disease in both DSS and T cell–induced colitis models and that IL-17F promotes colitis by inhibiting the colonization of the gut by regulatory T cell–promoting commensals (Tang et al., 2018). These findings indicate that specific IL-17F inhibitors also could be a potential option for the treatment of Crohn’s disease.

Fig. 3 summarizes the divergent roles of IL-23 and IL-17 inhibition in Crohn’s disease and psoriasis.

The success of IL-17 inhibition in psoriasis encouraged the initiation of clinical trials in other chronic inflammatory diseases with elevated IL-17 expression. Serum IL-17 and IL-23 levels are increased in children with atopic dermatitis and correlated with disease severity (Leonardi et al., 2015). IL-17 and IL-23 inhibitors are currently in phase II trials; however, two trails with the combined IL-12/IL-23 inhibitor Ustekinumab failed to show superiority compared with placebo (Khattri et al., 2017; Saeki et al., 2017), suggesting redundancy of the IL-23/IL-17 axis in atopic dermatitis. Several case reports reported successful outcomes of patients suffering from Hidradenitis suppurativa (Jørgensen et al., 2018) and Pityriasis rubra pilaris (Wain et al., 2018) with IL-17 inhibitors, indicating shared pathomechanisms of these inflammatory skin diseases with psoriasis despite different clinical presentation. IL-17 pathway inhibition did not show a benefit over placebo for the treatment of asthma in multiple trials (Busse et al., 2013; see Table 1), suggesting that the IL-23/IL-17 axis is not a major pathological driver in asthmatic airway inflammation.

Overexpression of IL-17 in mice leads to inflammation of epithelial barriers as well as granulocytosis (Croxford et al., 2014; Haak et al., 2009). However, no inflammation-related symptoms of other, nonbarrier tissues are induced (Haak et al., 2009). This argues against the notion that the dysregulation of IL-17 would elicit immunopathology in nonbarrier tissues. The exception seems to be psoriasis arthritis (PsA), where inflammation of the skin usually precedes joint pathology and has a very different disease etiology than rheumatoid arthritis (RA). Nevertheless, the role of IL-17 in inflammatory disorders of nonbarrier tissues was extensively studied using preclinical models, and IL-17 inhibitors were tested in the clinics.

PsA is the most common psoriasis-associated disorder, found in up to 30% of psoriasis patients (Ritchlin et al., 2017; Villani et al., 2015) and is among others associated to polymorphisms in IL23R (Bowes et al., 2015). Increased numbers of IL-17 producing ILCs as well as CD4 and CD8 T cells are found in the synovial fluid of PsA patients (Benham et al., 2013; Leijten et al., 2015;Menon et al., 2014). In the joints, IL-17 seems to act on osteoblasts inducing osteoclast differentiation and thereby increases bone resorption (Sato et al., 2006). Until now, no specific murine model for has PsA existed. However, in some models for psoriasis, the development of arthritis was observed and demonstrated that changes in the skin can be sufficient to induce pathological changes in the joints (Zenz et al., 2005; Yamamoto et al., 2015; Croxford et al., 2014). In line with this, most drugs that are effective to treat psoriatic skin lesions also inhibit arthritis in PsA patients, indicating a shared pathophysiology (McInnes et al., 2013, 2015; Mease et al., 2015). However, it is still unclear how inflammatory processes in the skin, which often precede PsA, lead to arthritis (Eder et al., 2016).

Soon after the success in treating plaque psoriasis, Secukinumab and Ixekizumab received the Food and Drug Administration/European Medicine Agency approval for treatment of PsA (McInnes et al., 2015; Mease et al., 2017); Secukinumab has extended the approved indication spectrum even for ankylosing spondylitis (Baeten et al., 2015). Both IL-17A inhibitors have shown clinical efficacy and long-term improvements, even in patients unresponsive to first-line therapy with TNF inhibitors (Coates et al., 2018; Nash et al., 2017). However, compared with the high rates of complete skin clearance in plaque psoriasis, effects on PsA are relatively modest, possibly due to a less pronounced IL-17 signature in affected joints compared with skin (Belasco et al., 2015). The safety profile in the setting of both diseases is comparable to IL-17 pathway inhibition in plaque psoriasis (Mease et al., 2019; Deodhar et al., 2019).

IL-23 inhibition is effective in phase II clinical studies in PsA (Deodhar et al., 2018), and phase III studies are ongoing (NCT03552276), while there was no evidence of clinically meaningful improvements in ankylosing spondylitis (Baeten et al., 2018), although both diseases respond well to IL-17 blockade. A hypothesis for the dichotomy of clinical efficacy is that in contrast to PsA, IL-17–producing cells become IL-23–independent in the chronic phase of joint inflammation in ankylosing spondylitis as shown in an animal model (van Tok et al., 2018). Counterintuitive to the proposed synergistic action of TNF and IL-17A in PsA, a recent study failed to show superiority of dual neutralization of TNF and IL-17A with a bi-specific antibody over TNF inhibition alone (Mease et al., 2018a).

Early after the discovery of IL-17, several reports showed an association between IL-17 and RA, which is a systemic autoimmune disorder characterized by the chronic inflammation of several organs including the synovia of joints (Scott et al., 2010). Those reports demonstrated the presence of IL-17 and IL-17R in synovial fluids and tissue from RA patients, and synovial IL-17 expression was identified as being predictive for joint damage progression (Ziolkowska et al., 2000; Honorati et al., 2001; Kirkham et al., 2006;Chabaud et al., 1999, 1998). The main producers of IL-17 in synovial tissues of RA patients were identified as CD4⁺ memory T cells (Kotake et al., 1999). IL-17 was shown to potently induce cytokines such as TNF, IL-1β, and GM-CSF from different joint cells as synovial fibroblasts, chondrocytes, and macrophages (Yao et al., 1995a, 1995b; Fossiez et al., 1996; Chabaud et al., 1998; Jovanovic et al., 1998) and was associated to bone and cartilage degradation (Chabaud et al., 2001; Cai et al., 2001; Chabaud and Miossec, 2001; Kim et al., 2015; Osta et al., 2015).

The disease-promoting role of IL-17 in RA was also described in mice. In wild-type mice, inflammatory arthritis can be induced by injecting recombinant IL-17 into the intra-articular space of knee joints (Chabaud et al., 2001; Cai et al., 2001). With rare exceptions (Doodes et al., 2008), the large majority of reports on IL-17 using different mouse models of RA showed that ablation of IL-17 reduced disease severity (Lubberts et al., 2001; Nakae et al., 2003;Katayama et al., 2013; Wu et al., 2010; Hirota et al., 2007; Duarte et al., 2010; Chabaud et al., 2001). The ablation of IL-17 was associated with reduced systemic levels of IL-6 as well as IL-1β and RANKL (receptor activator of NF-κB ligand)–producing cells in the synovium (Lubberts et al., 2004). Also, the transdifferentiation of Foxp3 regulatory T cells into IL-17–producing cells was associated to arthritis (Komatsu et al., 2014). Furthermore, it was proposed that changes in the microbiome and IL-17 in the gut are early events in RA (Wu et al., 2010), supporting the mucosal origins hypothesis suggesting that the disease-promoting immunity in RA is initiated at mucosal sites (leading to the circulation of auto-antibodies before clinical symptoms are observed) and only later spreads to the synovial joints generating symptomatic inflammation (reviewed in Holers et al., 2018).

Despite the extensive rationale from preclinical studies in mice and human to target IL-17 in RA, results from clinical studies fell far short of expectations. Although IL-17A inhibition was superior to placebo and improved symptoms in TNF nonresponders (Blanco et al., 2017; Genovese et al., 2014), response rates were still lower compared with an established second-line biological CTLA4-Ig, Abatacept (Blanco et al., 2017). As these studies provided evidence that IL-17A is a central pathogenic driver only in a small subgroup of RA patients and other available substance classes are more effective, clinical development was terminated after phase III (NCT01377012).

Many of the preclinical findings on IL-17 in RA were made using the collagen-induced arthritis (CIA) model where DBA-1 mice are immunized against type 2 collagen and immune response is boosted with CFA (Courtenay et al., 1980). Despite its valuable contributions to a better understanding of RA, in the CIA model, γδ T cells are the main producers of IL-17, while these cells were hardly detected in joints of RA patients (Ito et al., 2009). We speculate that the CFA used in the CIA model induces the excessive IL-17 expression, thus inflating the role of IL-17 in the model compared with human RA.

Furthermore, it was recently shown that the transition of disease initiation to symptomatic arthritis could be mediated by IL-23 rather than IL-17. Pfeifle et al. (2017) found that IL-23–activated Th17 cells promoted the production of pro-inflammatory antibodies by B cells via IL-22 and IL-21, but not IL-17. However, at least in established RA, a recent study with a selective IL-23 inhibitor showed no improvement in disease symptoms (Smolen et al., 2017).

As for other inflammatory disorders, dysregulated cytokines play an important role in MS. In 2002, it was shown that, as for psoriasis, IL-23 and not IL-12 is disease-promoting in the EAE model of MS (Cua et al., 2003; Becher et al., 2002). IL-23–deficient mice were shown to be resistant to EAE and had reduced levels of IL-17, while the transfer of IL-17–producing T cells led to disease exacerbation (Langrish et al., 2005). The presence of IL-17–producing cells in active brain lesions of MS patients (Tzartos et al., 2008) and increased IL-17 levels in cerebrospinal fluid especially during clinical exacerbation (Matusevicius et al., 1999), supported the idea that IL-17 might be implicated in MS pathogenesis.

Surprisingly, blockade of IL-17A or F in EAE only led to a disease reduction in SJL/J mice, but had only a modest effect in C57/BL6 mice (Park et al., 2005; Hofstetter et al., 2005; Komiyama et al., 2006; Langrish et al., 2005; Haak et al., 2009).McGeachy et al. (2007) reported that IL-17–producing Th cells are only pathogenic when they are driven by IL-23, but not by IL-6 and TGFβ. In addition, IL-23 decreased the IL-10 expression by IL-17–producing cells, promoting their encephalitogenic capacity. In support of this, we found that IL-23 authorizes T cells to invade the central nervous system (CNS) and to polarize to Th17 cells (Gyülvészi et al., 2009). Furthermore, several reports suggested the existence of nonpathogenic Th17 cells (Yang et al., 2009; Esplugues et al., 2011). All these findings demonstrate that while IL-23 is essential for the development of EAE, IL-17 is largely redundant for disease development, and IL-23 seems to induce other pathogenic mediators. Indeed, we and others reported that IL-23 converts nonpathogenic into pathogenic T cells by the induction of GM-CSF (Codarri et al., 2011; El-Behi et al., 2011). Within the CNS of EAE mice, GM-CSF is produced almost exclusively by neuro-antigen–specific Th cells (Komuczki et al., 2019), and systemic dysregulation of GM-CSF leads to severe and fatal CNS inflammation dominated by the invasion into the CNS of monocytes (Spath et al., 2017). Lee et al. (2012) found that T-bet, which was already associated to the expression of IL-23R in Th17 cells (Gocke et al., 2007) and protection from EAE (Panitch et al., 1987), is driving the pathogenic functions of Th17 cells independently of their IL-17 production (Lee et al., 2012). In line with this, clinically stable MS patients had higher expression of IL-10 in Th17 cells, and single-cell clones from patients with MS showed enhanced production of IL-17, GM-CSF, or IFN-γ, while in healthy controls, IL-10 dominated (Cao et al., 2015; Hu et al., 2017).

In summary, targeting of IL-17 does not efficiently inhibit disease formation in mice, whereas targeting IL-23 selectively blocks disease-promoting Th cells, independently of IL-17. The role of IL-17 in MS remains controversial. Although IL-17 pathway inhibition slowed down the formation of new Gd-enhancing T1 brain lesions in a phase II study, the trial was terminated because the cumulative number of lesions was not significantly different compared with placebo at the primary endpoint (Havrdová et al., 2016). Although according to preclinical data the new selective IL-23 inhibitors offer a promising strategy for MS, clinical trials have not been initiated yet. Combined IL-12/IL-23 inhibition failed to show clinical efficacy in patients (Segal et al., 2008; Vollmer et al., 2011), possibly by a negative effect of the concomitant IL-12 pathway inhibition.

There are ongoing trials exploring the possibility of IL-17 inhibitors in primary biliary cirrhosis, giant cell arteritis, and Sjögren’s syndrome (Table 1). The rational for such trials came from observations that IL-17–producing cells were present in the inflammatory infiltrates of the respective affected organs of patients (Katt et al., 2013; Deng et al., 2010; Zhang et al., 2018). A clinical trial investigating IL-17 inhibition in diabetes mellitus type 1 patients was terminated during phase II (Table 1).

It is becoming increasingly clear that dysregulated cytokine networks are fundamentally involved in chronic inflammatory diseases and that targeting cytokines therapeutically has great value in treating patients suffering from these disorders. Of note, excessive production of single cytokines in transgenic mice actually led to tissue-specific inflammation (Keffer et al., 1991; Haak et al., 2009; Spath et al., 2017). IL-17 in particular is connected with barrier function, and dysregulation of IL-17 is closely linked to psoriasiform skin inflammation. The most likely explanation for the skin tropism of IL-17–mediated inflammation is the fact that the receptor complex for IL-17 is highly abundant in the skin compared with other tissues. The normal steady-state function is likely the control of the commensal microflora, and as a quasi-innate cytokine, IL-17 can rapidly recruit other leukocytes in response to dysbiosis or pathogen invasion through barriers. So, neutralization of the IL-17/IL-23 axis has been shown to be beneficial in treating psoriasis, while in Crohn's disease only the targeting of IL-23 shows clinical benefit, possible because the steady-state function of IL-17 in gut barrier function is too fragile for external intervention. Nevertheless, the impact of the IL-23/IL-17 axis on the microbiome is abundantly clear. The nature of the commensal flora in the steady-state has in turn a dramatic effect on immunity in general. We thus propose that the conflicting data regarding the role of IL-17 in preclinical models of inflammation (outside of barrier tissues) is likely the result of the altered gut immune axis observed in IL-17–deficient mice. The clinical success of targeting the IL-17/IL-23 axis in chronic inflammation of the body lining but not of diseases of internal organs (e.g., RA, diabetes mellitus type 1, MS, etc.), was largely predicted in the preclinical studies, which provides further support for integrative and comprehensive bench to bedside research.

This work was supported by grants from the Swiss National Science Foundation (310030_146130, 316030_150768, and 310030_170320 to B. Becher), the European Union FP7 Project ATECT (Advanced T-cell Engineered for Cancer Therapy; to B. Becher), the European Union H2020 Project iPC (Individualized Paediatric Cure), and the University Research Priority Project Translational Cancer Research (to B. Becher).

The authors declare no competing financial interests.

Author contributions: S. Unger, P. Zwicky, and B. Becher wrote the manuscript.