Identification of small‐molecule elastase inhibitors as antagonists of IL‐36 cytokine activation

IL‐1 family cytokines act as apical initiators of inflammation in many settings and can promote the production of a battery of inflammatory cytokines, chemokines and other inflammatory mediators in diverse cell types. IL‐36α, IL‐36β and IL‐36γ, which belong to the extended IL‐1 family, have been implicated as key initiators of skin inflammation in psoriasis. IL‐36γ is highly upregulated in lesional skin from psoriatic individuals, and heritable mutations in the natural IL‐36 receptor antagonist result in a severe form of psoriasis. IL‐36 family cytokines are initially expressed as inactive precursors that require proteolytic processing for activation. The neutrophil granule‐derived protease elastase proteolytically processes and activates IL‐36α and IL‐36γ, increasing their biological activity ~ 500‐fold, and also robustly activates IL‐1α and IL‐33 through limited proteolytic processing. Consequently, inhibitors of elastase activity may have potential as anti‐inflammatory agents through antagonizing the activation of multiple IL‐1 family cytokines. Using in silico screening approaches, we have identified small‐molecule inhibitors of elastase that can antagonize activation of IL‐36γ by the latter protease. The compounds reported herein may have utility as lead compounds for the development of inhibitors of elastase‐mediated activation of IL‐36 and other IL‐1 family cytokines in inflammatory conditions, such as psoriasis.

IL-1 family cytokines act as apical initiators of inflammation in many settings and can promote the production of a battery of inflammatory cytokines, chemokines and other inflammatory mediators in diverse cell types. IL-36a, IL-36b and IL-36c, which belong to the extended IL-1 family, have been implicated as key initiators of skin inflammation in psoriasis. IL-36c is highly upregulated in lesional skin from psoriatic individuals, and heritable mutations in the natural IL-36 receptor antagonist result in a severe form of psoriasis. IL-36 family cytokines are initially expressed as inactive precursors that require proteolytic processing for activation. The neutrophil granule-derived protease elastase proteolytically processes and activates IL-36a and IL-36c, increasing their biological activity~500-fold, and also robustly activates IL-1a and IL-33 through limited proteolytic processing. Consequently, inhibitors of elastase activity may have potential as anti-inflammatory agents through antagonizing the activation of multiple IL-1 family cytokines. Using in silico screening approaches, we have identified small-molecule inhibitors of elastase that can antagonize activation of IL-36c by the latter protease. The compounds reported herein may have utility as lead compounds for the development of inhibitors of elastase-mediated activation of IL-36 and other IL-1 family cytokines in inflammatory conditions, such as psoriasis.
Individuals that carry loss-of-function mutations in the IL-36 receptor antagonist (IL-36RA) display a severe form of psoriasis, called generalized pustular psoriasis [16][17][18][19][20]. Furthermore, expression of all three IL-36 family members has been found to be dramatically elevated (100-fold) in skin biopsies from individuals with the most common form of psoriasis, psoriasis vulgaris, compared with nonlesional skin from the same individuals, or nonaffected controls [13,14]. This suggests that deregulated IL-36 cytokine signalling is sufficient to initiate psoriatic-type skin inflammation. Moreover, deregulated expression of IL-36a in the mouse leads to a psoriasis-like condition that is exacerbated with the skin irritant, phorbol acetate [12]. In addition, transplantation of human psoriatic lesions onto immunodeficient (SCID) mice produces a psoriasis-like condition that is ameliorated through IL-36 receptor neutralization [13].
As is common within the extended IL-1 cytokine family, IL-36a, IL-36b and IL-36c are all expressed as leaderless cytokines that lack biological activity unless cleaved within their N termini [11,22,23]. Thus, limited proteolytic processing of IL-36 family cytokines to remove up to 15 Nterminal amino acids dramatically increases their proinflammatory activity by over 500-fold [11,22,24,25]. We have recently reported that the neutrophil-derived proteases cathepsin G (CatG) and elastase are potent IL-36processing proteases, with elastase playing a major role in the processing and maturation of IL-36a and IL-36c [11,25]. Furthermore, IL-1a and IL-33 are also proteolytically processed and activated by elastase [reviewed in Ref. 26]. Because deregulated IL-36 cytokine activation appears to be instrumental in the initiation of inflammation in the skin barrier, inhibitors of IL-36 activation and/or downstream activity are likely to have potential for the treatment of inflammatory skin conditions. Neutrophil infiltrates are common in psoriasis, and these cells are the major source of elastase [27][28][29]. Thus, small-molecule inhibitors of neutrophil elastase (NE) may have significant potential as inhibitors of IL-36 activation in psoriasis.
Here, we have used in silico screening approaches to identify small-molecule inhibitors of elastase, followed by functional testing of candidates. We show that these inhibitors are capable of antagonizing elastase-mediated processing and activation of IL-36c, suggesting that these compounds may be useful leads for the generation of therapeutic modulators of IL-36 cytokine activity in inflammatory conditions.

Elastase processes and activates IL-36c
Similar to other members of the extended IL-1 family, such as IL-1b and IL-18 [26], IL-36 cytokines possess minimal pro-inflammatory activity as fulllength proteins and require N-terminal processing for activation [11,22]. However, as we have recently reported [11,25], IL-36c is robustly activated upon incubation with elastase ( Fig. 1), a protease that is released in large quantities into the extracellular space from the cytoplasmic granules of activated neutrophils. Elastase processes IL-36c at Val15, liberating a new N terminus that unleashes the pro-inflammatory activity of this cytokine, possibly through provoking a conformational change in the latter or through eliminating steric interference within the receptor-binding domain of this cytokine [11]. Of note, previous studies have shown that IL-36c is dramatically upregulated at the mRNA and protein levels in lesional skin from psoriasis patients, compared with unaffected skin from the same individuals, or from control subjects [13,14,24,30]. Coupled with observations that loss-of-function mutations in the natural IL-36 receptor antagonist promote a highly severe form of psoriasis [16][17][18][19][20], this suggests that IL-36c may be an important driver of the inflammation seen in this condition.
Migration of circulating neutrophils into peripheral tissues is a major amplifier of inflammation and is commonly seen in psoriatic lesions. Neutrophilderived proteases such as elastase and CatG, although generally thought of as antimicrobial enzymes, are also potent instigators of inflammation [4,11,25,26], most likely through processing and activation of IL-1 family cytokines such as IL-36a and IL-36c. Therefore, inhibitors of NE may have therapeutic potential as anti-inflammatory agents through antagonizing processing and activation of multiple IL-1 family cytokines.

Identification of candidate elastase inhibitors using an in silico screening approach
To identify novel small-molecule inhibitors of elastase, we performed in silico screening of the elastase crystal structure with an in-house compound library comprised of over 100 000 unique molecular entities (Saint Petersburg Technical University). Using molecular dynamics simulation, compounds were docked in multiple poses into the substrate-binding pocket of elastase, as illustrated by the examples presented in Fig. 2. Using this approach, and guided by known inhibitors of elastase such as dihydropyrimidine (DHPI) ( Fig. 2A,B) and Alvelestat/AZD9668 (Fig. 2C,D), we identified a small molecule, designated LCB016, which fit the binding parameters expected of a candidate inhibitor (Fig. 2E,F).

LCB016 and derivatives thereof exhibit elastaseinhibitory activity
LCB016 has an azolactone structure (Fig. 3A); therefore, a series of azolactone analogues were synthesized (LCB001-LCB165; Table 1), which were then assessed for their ability to antagonize elastase activity, initially using a synthetic substrate (AAPV-AMC) hydrolysis assay. As Fig. 3B demonstrates, LCB016 and several of its derivatives were found to inhibit NE activity as assessed by the ability of these compounds to antagonize hydrolysis of the synthetic elastase substrate peptide Suc-AAPV-AMC. Substitution of the bromine in LCB016 by hydrogen (LCB108, LCB109, LCB111) led to a significant decrease in calculated binding energy to elastase (data not shown), and consequently to diminished inhibitor potency (Fig. 3B). Substitution of the ethoxy group to smaller methoxy in LCB016 also led to a decrease in binding energy (Table 1 and Fig. 3B). Furthermore, the presence of the benzoxy group was found to be essential to the binding (data not shown).
Based on these preliminary screening results, we performed titrations of LCB016 and several of its derivatives (LCB091, LCB092, LCB095, LCB113, LCB161, LCB163 and LCB164) against purified elastase (Fig. 4). Once again, we observed reproducible elastase-inhibitory activity of LCB016 and several of its derivatives, with LCB092 and LCB113 exhibiting greater potency than LCB016 (Fig. 4). We also performed time-course analyses of LCB016 and the most promising derivatives over a range of drug concentrations, and as Fig. 5A illustrates, LCB092, LCB113 and LCB116 exhibited the greatest potency towards elastase.

Specificity of candidate elastase inhibitors towards other inflammatory proteases
There is approximately 56% structural similarity between NE and proteinase-3 (PR3), and these proteases exhibit almost identical substrate preferences [11]. Thus, LCB016 and its derivatives also exhibited some inhibitory activity towards PR3, as expected (Fig. 5B). However, no cross-inhibitory effect was observed towards the other major neutrophil protease, CatG (Fig. 6A). Furthermore, caspase-1, which plays a key inflammatory role through processing and activation of IL-1b and IL-18, was also unaffected by the candidate elastase inhibitors (Fig. 6B). Collectively, these data suggest that the compounds identified herein exhibit good specificity towards NE.

Small-molecule elastase inhibitors can antagonize IL-36c activation and cytokine production downstream
Based on the synthetic substrate hydrolysis data, six compounds (LCB016, LCB090, LCB092, LCB113, LCB115 and LCB116) were selected for further testing in the more physiologically relevant context of elastase-mediated IL-36c processing. Here, we incubated recombinant full-length IL-36c in the presence or absence of elastase, either alone or in combination with the candidate elastase inhibitors (Fig. 7A). The products of the latter reactions were then incubated with HeLa IL-36R cells, which respond to active (i.e. elastase-processed) IL-36c by secreting a range of proinflammatory cytokines and chemokines such as IL-6, IL-8 and CXCL1 (Fig. 7A) [11,23]. As Fig. 7B illustrates, while LCB016 and some of its derivatives clearly suppressed elastase-mediated activation of IL-36c, as indicated by the inhibition of IL-6 synthesis from HeLa IL-36R cells, LCB092 and LCB113 exhibited greater potency in this regard, in agreement with the synthetic substrate hydrolysis assays (Fig. 5A).
To explore whether the inhibitors identified herein could also inhibit IL-36c processing and activation by elastase released from activated neutrophils, we activated purified human neutrophils via treatment with phorbol 12-myristate 13-acetate (PMA; Fig. 8A). The latter treatment triggered the release of multiple neutrophil granule proteases, including elastase, PR3 and CatG, into the extracellular space as expected (Fig. 8B). Using activated neutrophil supernatants (SNs) as a source of elastase to process and activate IL-36c, we observed that the lead compounds (LCB016, LCB092 and LCB113) robustly suppressed activation of the latter by neutrophil-derived proteases (Fig. 8C). As we have reported previously [11], IL-36b is predominantly activated by CatG, and in accordance with this, we found that our elastaseinhibitory compounds had minimal effects on IL-36b activation by neutrophil degranulate preparations. However, inhibition of CatG activity, using a commercial inhibitor of the latter, robustly suppressed activation of IL-36b by activated neutrophil degranulate preparations (Fig. 8D).
In summary, we have identified a number of compounds that are capable of acting as inhibitors of elastase and may have utility as lead compounds designed to suppress inflammation through blocking the actions of this protease against cytokines such as IL-36c.

Discussion
Here, we report the identification of several small molecules that exhibit significant inhibitory activity against NE and can antagonize the processing and activation of IL-36c by the latter. Because elastase has also been implicated in the processing and activation of IL-1a, IL-1b, IL-33 and other cytokines [26], the compounds reported herein are also likely to suppress the activation of multiple members of the extended IL-1 family. Further experiments will explore the latter possibility and will also explore the activity of these compounds in animal models of psoriasis and other inflammatory diseases.
Neutrophil infiltration is a hallmark of a number of skin-related inflammatory diseases. In particular, psoriatic plaques are heavily infiltrated with neutrophils, dendritic cells, macrophages and T cells [27,28,31]. Neutrophils are well-known first responder cells of the innate immune system and play a central role in the early stages of infection or tissue damage [32][33][34]. Although it is widely appreciated that release of neutrophil proteases (through degranulation or neutrophil extracellular trap [NET] formation) can exert profound antimicrobial effects during infection, the latter proteases can also cause extensive tissue damage and exacerbate inflammation [35][36][37]. Psoriasis is often preceded by tissue damage or abrasion, called the Koebner reaction [38], which is most likely due to Thiophene is substituted on bioisosteric furan ring LCB164 Thiophene is substituted on benzene ring LCB165 Azolactone ring is substituted with sulfur consequence of tissue trauma, may play an important initiating role in psoriasis, especially in individuals lacking endogenous buffers of IL-36 activity, such as deficiency in the IL-36R antagonist [16][17][18][19][20]. Because previous studies have also implicated NE in the activation of IL-1a [39], IL-1b [40], IL-18 [41] and IL-33 [42] and IL-36 receptor antagonist [43], our observations also suggest that this protease may serve as a therapeutic target for suppressing the activation of multiple IL-1 family cytokines in inflammatory diseases.
Psoriasis is associated with a pro-inflammatory signature with elevated levels of tumour necrosis factor (TNF), IL-17C and IL-36 among other cytokines. At present, multiple biotherapeutics, directed against cytokines such as TNF, IL-17, IL-12/IL-23 and IL-36, have been approved or are in development for the treatment of psoriasis, as well as other inflammatory conditions. However, although these cytokine-neutralizing approaches are highly effective, they also suffer from several drawbacks including high cost, a necessity for systemic delivery (which can increase vulnerability to opportunistic infections) and their single-molecule specificity. The need for systemic delivery of cytokine-neutralizing biotherapeutics is particularly problematic due to the neutralization of cytokine throughout the body and not just where it is overproduced. A more desirable approach would be to neutralize cytokine only in the affected lesions. Thus, small molecules, such as the compounds reported herein, have the advantage that they are less costly to produce, can be applied directly to affected areas of skin and may simultaneously suppress activation of multiple cytokines that are processed and activated by elastase [11,29].

Cell culture
HeLa cells were cultured in RPMI media (Gibco (BioSciences), Dublin, Ireland), supplemented with 5% FBS. HeLa.IL-36R cell lines were generated by transfection with pCXN2.empty or pCXN2.IL-1Rrp2 (IL-36R) plasmids followed by selection using G-418 antibiotic (Sigma). IL-36Roverexpressing clones were expanded from a single cell. Clones were selected by demonstration of acquired optimal responsiveness to active forms of IL-36 via ELISA. All cells were cultured at 37°C in a humidified atmosphere with 5% CO 2 .

Expression and purification of recombinant proteins
Full-length IL-36b and IL-36c were generated by cloning the human coding sequence in frame with the polyhistidine tag sequence in the bacterial expression vector pET45b. Protein was expressed by addition of 600 lM IPTG to exponentially growing cultures of BL21 strain Escherichia coli followed by incubation for 3 h at 37°C. Bacteria were lysed by sonication and polyhistidine-tagged proteins were captured using nickel/NTA agarose (Qiagen (UK), Manchester, UK), followed by elution into PBS, pH 7.2, in the presence of 100 mM imidazole. Recombinant polyhistidine-tagged caspase-1 and caspase-3 were expressed and purified as described previously [44].
In silico screening and molecular modelling Ligand 3D structures were prepared with LigPrep suite in OPLS_2005 force field. The Schrodinger package was used for molecular modelling studies. Glide was used for SM virtual screening (VS) and docking studies; the ligands from the in-house SM library were treated as flexible. The standard precision protocol was used for VS. The positions of the identified azolactone hit compounds were specified within extra-precision protocol. Azolactone derivatives were redocked as Z-isomers corresponding XRD data. The X-ray human neutrophil elastase (HNE) structure (PDB ID 1HNE) was used for in silico screening; the water and cocrystallized molecules were removed from the protein model. Hydrogens were added using PROPKA program pH 7.4 and their positions optimized. The 25 9 25 9 25 A 3 box (with 12 A diameter ligand mid-point), centred on the coordinates of 57 His residue (active site), was used for grid map calculations; the 94 Tyr, 54 Ser, 195 Ser and 214 Ser hydroxyl groups were allowed to rotate during docking runs. For the validation of the docking protocol, the DHPI ligand was blindly docked to the HNE structure (PDB ID 3Q77) and the poses of docked ligand and cocrystallized with HNE were aligned (Fig. 2) showing good theory-to-experiment correspondence.