A Truncated Form of HpARI Stabilizes IL-33, Amplifying Responses to the Cytokine

The murine intestinal nematode Heligmosomoides polygyrus releases the H. polygyrus Alarmin Release Inhibitor (HpARI) - a protein which binds to IL-33 and to DNA, effectively tethering the cytokine in the nucleus of necrotic cells. Previous work showed that a non-natural truncation consisting of the first 2 domains of HpARI (HpARI_CCP1/2) retains binding to both DNA and IL-33, and inhibited IL-33 release in vivo. Here, we show that the affinity of HpARI_CCP1/2 for IL-33 is significantly lower than that of the full-length protein, and that HpARI_CCP1/2 lacks the ability to prevent interaction of IL-33 with its receptor. When HpARI_CCP1/2 was applied in vivo it potently amplified IL-33-dependent immune responses to Alternaria alternata allergen, Nippostrongylus brasiliensis infection and recombinant IL-33 injection, in direct contrast to the IL-33-suppressive effects of full-length HpARI. Mechanistically, we found that HpARI_CCP1/2 is able to bind to and stabilize IL-33, preventing its degradation and maintaining the cytokine in its active form. This study highlights the importance of IL-33 inactivation, the potential for IL-33 stabilization in vivo, and describes a new tool for IL-33 research.


INTRODUCTION
Heligmosomoides polygyrus is a parasitic nematode that infects the intestines of mice. It has a fecal/oral lifecycle, with infective L3 larvae being ingested, and then rapidly penetrating the epithelium of the proximal duodenum. There, the larvae develop to L4 stage and emerge as adults into the intestinal lumen at around day 10 of infection (1,2). The transit of the parasite through the intestinal wall is likely to cause epithelial damage and cell death, resulting in the release of alarmins such as IL-33 from stromal cells or mast cells (3), in turn inducing an anti-parasite type 2 immune response (4). In order to negate this response, and allow persistence of the parasite in the host, H. polygyrus secretes multiple immunomodulatory factors, including Hp-TGM, a protein mimic of host TGF-β (5), and microRNA-containing extracellular vesicles (6) which modulate transcription of multiple host genes, including suppression of Suppression of Tumorigenicity 2 (ST2), the IL-33 receptor. Furthermore, our recent work shows that H. polygyrus secretes HpBARI, a protein which binds and blocks ST2 (7). We previously showed that the parasite also secretes the H. polygyrus Alarmin Release Inhibitor (HpARI), which blocks IL-33 responses (8).
IL-33 is an alarmin cytokine constitutively produced by epithelial cells. It is stored preformed in the nucleus and released on necrotic cell death, due to mechanical, protease-mediated or chemical damage to the epithelium (9). On necrotic cell death, proteases from the cell cytoplasm, or those secreted by recruited mast cells, neutrophils or those in allergens can then cleave the cytokine between the N-terminus chromatinbinding domain and the C-terminus receptor binding domain, potently increasing the activity of the cytokine (10)(11)(12). The IL-33 receptor-binding domain contains four free cysteine residues, which upon release from the reducing nuclear environment into the oxidizing extracellular environment rapidly form disulphide bonds, changing the cytokine's conformation, rendering it unable to bind to its receptor and effectively inactivating it (13). Proteases can also further degrade IL-33 to smaller, inactive forms (12). Thus, the active form of IL-33 has only a very short half-life, and by 1 h after release the vast majority of IL-33 is inactive or degraded.
HpARI binds to the active reduced form of IL-33 and to genomic DNA. This dual binding tethers IL-33 within the nucleus of necrotic cells, preventing its release, and inhibiting interaction of IL-33 with ST2. The HpARI protein consists of 3 Complement Control Protein domains (CCP1-3), and our previous data showed that HpARI binds IL-33 through the CCP2 domain, while DNA-binding was mediated by the CCP1 domain (8).
Here, we further characterize the functions of the CCP domains of HpARI, finding that CCP3 stabilizes the interaction between HpARI and IL-33, increasing its affinity and being required for blockade of IL-33-ST2 interactions. Furthermore, we show that HpARI_CCP1/2 (the HpARI truncation lacking CCP3) is able to stabilize IL-33, increasing its half-life and amplifying its effects.

Protein Expression and Purification
Constructs encoding HpARI, HpARI_CCP1/2 and HpARI_CCP2/3 (all with C-terminus myc and 6-His tags) were cloned into the pSecTAG2A expression vector as previously described (8). Purified plasmids were transfected into Expi293F TM cells, and supernatants collected 5 days later. Expi293F TM cells were maintained, and transfections carried out using the Expi293 Expression System according to manufacturer's instructions (ThermoFisher Scientific). Expressed protein in supernatants were purified over a HisTrap excel column (GE Healthcare) and eluted in 500 mM imidazole. Eluted protein was then dialysed to PBS, and repurified on a HiTRAP chelating HP column (GE Healthcare) charged with 0.1 M NiSO 4 . Elution was performed using an imidazole gradient and fractions positive for the protein of interest were pooled, dialysed to PBS and filter-sterilized. Protein concentration was measured at A280 nM (Nanodrop, ThermoFisher Scientific), using calculated extinction coefficient.
SPR single-cycle kinetic titration binding experiments were performed at 25 • C. Three-fold dilution series of mIL-33 (2.47 nM to 200 nM), were injected over the sensor surface, in 10 mM NaH 2 PO 4 , pH 7.5; 150 mM NaCl; 50 µM EDTA; 0.05% surfactant P20, at 30 ml.min −1 for 30 s followed by a final 600 s dissociation phase. The on-(k + ) and off-rate (k − ) constants and the equilibrium dissociation constants were calculated from the double referenced sensorgrams by global fitting of a 1:1 binding model, with mass transport considerations, using analysis software (v2.02) provided with the Biacore T200 instrument.

Animals
BALB/cAnNCrl and C57BL/6JCrl mice were purchased from Charles River, UK. Heterozygous IL-13eGFP +/GFP mice (15) were bred in-house. All mice were accommodated and procedures performed under UK Home Office licenses with institutional oversight performed by qualified veterinarians.

Nippostrongylus brasiliensis Infection
The life cycle of N. brasiliensis was maintained in Sprague-Dawley rats as previously described (17), and infective L3 larvae were prepared from 1 to 3 week rat fecal cultures. C57BL/6 mice were subcutaneously infected with 400 L3 N. brasiliensis larvae, and culled 3 or 6 days later.

CMT-64 Cell Line
CMT-64 cells (ECACC 10032301) were maintained by serial passage in "complete" RPMI [RPMI 1640 medium containing 10% fetal bovine serum, 2 mM L-glutamine, 100 U/ml Penicillin and 100 µg/ml Streptomycin (ThermoFisher Scientific)] at 37 • C, 5% CO 2 . Cells were seeded into 24-or 96-well plates for Triton-X100 or freeze-thaw treatment, respectively. Cells were grown to 100% confluency prior to 2 washes with PBS. For Triton-X100 treatment, cells were then washed into RPMI 1640 containing 0.1% BSA with or without 0.1% Triton-X100, and incubated at 37 • C as indicated, prior to collection of supernatants and measurement of IL-33 by ELISA and western blot. For freeze-thaw assays, cells were then washed into complete RPMI containing 10 µg/ml of HpARI or HpARI_CCP1/2, frozen on dry ice for at least 1 h, then thawed and incubated at 37 • C as indicated, prior to collection of supernatants and application to bone marrow cell cultures.

Bone Marrow Cell Culture
Single cell suspensions of bone marrow cells were prepared from C57BL/6 mice, by flushing tibias and femurs with RPMI 1640 medium using a 21 g needle. Cells were resuspended in red blood cell lysis buffer (Sigma) for 5 min at room temperature, prior to resuspension in medium and passing through a 70 µm cell strainer. Cells were cultured in round-bottom 96-well-plates in a final 200 µl volume, containing 0.5 × 10 6 cells/well. IL-2 and IL-7 were added at 10 ng/ml final concentration, with 50 µl of CMT-64 freeze-thaw supernatant. Cells were then cultured at 37 • C, 5% CO 2 , for 5 days, prior to assessment of responses by cytokine ELISA and flow cytometry.

Statistical Analysis
All data was analyzed using Prism (Graphpad Software Inc.). One-way ANOVA with Dunnet's multiple comparisons posttest was used to compare multiple independent groups, while two-way ANOVA and Tukey's multiple comparison's post-test was used to compare multiple timepoints or concentrations between independent groups. Where necessary, data was logtransformed to give a normal distribution and to equalize variances. * * * * p < 0.0001, * * * p < 0.001, * * p < 0.01, * p < 0.05, N.S. = Not significant (p > 0.05).

HpARI_CCP1/2 Increases Responses to IL-33
We previously showed that HpARI_CCP1/2 was capable of suppressing the release of IL-33 in vivo, 15 min after Alternaria alternata administration (8). To assess whether HpARI_CCP1/2 could replicate the inhibition of IL-33-dependent responses seen with full-length HpARI, we administered HpARI or HpARI_CCP1/2 together with Alternaria allergen and OVA protein and assessed type 2 immune responses after OVA challenge 2 weeks later (Figure 2A). While HpARI suppressed allergic reactivity in this model (as shown previously (8)), HpARI_CCP1/2 had the opposite effect, increasing BAL and lung eosinophil, and lung ILC2 and ICOS + ST2 + Th2 cell numbers (18) (Figure 2A and Supplementary Figure 1).
When the innate Alternaria-induced immune response was assessed 24 h after initial administration of the allergen to naïve mice, we found that although HpARI_CCP1/2 did not change the eosinophil response compared to Alternaria alone, HpARI_CCP1/2 increased BAL neutrophil numbers. At this timepoint, no ILC2 proliferation has yet occurred, as previously described (19), so total lung ILC2 cell numbers were similar in all groups (data not shown). However, allergen-activated ILC2s showed strong upregulation of CD25 expression, as described previously during activation of ILC2s in this model (20), which was further increased by HpARI_CCP1/2 ( Figure 2B).
To exclude the possibility that HpARI_CCP1/2 is interfering with the Alternaria allergen directly, exacerbating the response to it, we used a second model of IL-33-dependent responses (21)(22)(23), infecting mice with Nippostrongylus brasiliensis and administering HpARI or HpARI_CCP1/2 to the lungs during the first 3 days of infection. During N. brasiliensis infection, L3 larvae migrate through the lung at days 1-4, enter the intestines as L4 larvae and develop to adults at days 4-10 post-infection (21). Mice were culled at days 3 and 6 post-infection, when parasites were present in the lung and gut, respectively, and the type 2 immune response in the lung was assessed at both timepoints. Again, HpARI suppressed type 2 immune responses as shown previously (8), while HpARI_CCP1/2 increased BAL eosinophilia, IL-5 and IL-13 production ( Figure 2C). Neither HpARI nor HpARI_CCP1/2 had any effect on BAL neutrophilia at these timepoints (data not shown), implying that neutrophil recruitment in N. brasiliensis is not IL-33 dependent. Similarly, in Strongyloides venezuelensis lung-stage infection, neutrophil recruitment is IL-33-independent (24).
Finally, we utilized a model of recombinant IL-33 intraperitoneal injection, which induces a mast cell-dependent neutrophilia (25,26), in contrast to the ILC2-dependent, largely eosinophilic response seen on IL-33 release in the lung. Again, here we found that while HpARI suppressed IL-33 induced neutrophilia, HpARI_CCP1/2 exacerbated it ( Figure 2D).
In conclusion, HpARI_CCP1/2 amplifies IL-33-dependent responses in vivo. We hypothesized that this activity was due to stabilization of the cytokine, increasing its effective half-life. To test this hypothesis, we developed an in vitro model of IL-33 release and IL-33 responses.

HpARI_CCP1/2 Maintains IL-33 in Its Active Form
The CMT-64 cell line constitutively produces IL-33, which is released on cellular necrosis (12). Confluent CMT-64 cells were washed into PBS+0.1% BSA, and necrosis induced by addition of 0.1% Triton-X100, in the presence or absence of HpARI Frontiers in Immunology | www.frontiersin.org or HpARI_CCP1/2. Over a 24 h timecourse following Triton-X100 addition, we assessed IL-33 release by ELISA and western blot. IL-33 ELISA showed that Triton-X100 caused rapid IL-33 release, with high concentrations of the cytokine detected in culture supernatants within 15 min of addition of the detergent in control wells. IL-33 levels then gradually decreased at later timepoints, presumably as the protein was degraded (Figure 3A) (12). HpARI addition ablated the IL-33 signal seen in the ELISA, as shown in our previous study (8): as well as retarding the release of the cytokine, HpARI binding also out-competes the ELISA antibodies, abolishing detection of IL-33. HpARI_CCP1/2 did not abolish detection of IL-33 in the ELISA, but did reduce the IL-33 signal at early timepoints. Moreover, in the presence of HpARI_CCP1/2, IL-33 accumulated over the timecourse and maintained high levels at later timepoints.
In contrast, when IL-33 in the same samples was assessed by western blot, a very strong signal was seen at all timepoints at a size consistent with full-length IL-33 protein (∼30 kDa), while a weaker signal was seen at around 18-20 kDa, consistent with processed mature IL-33 ( Figure 3B and Supplementary Figures 2A,B). While a strong full-length IL-33 band was seen across all timepoints and treatments, the density of the mature bands were dynamically altered by the presence of each treatment. In control wells, mature IL-33 was present early after Triton-X100 treatment and was degraded at later timepoints. In contrast, in the presence of HpARI_CCP1/2, the mature form was present at lower intensities than in control wells at early timepoints, but accumulated over the timecourse and was strongest at 24 h post Triton-X100 treatment, reflecting ELISA data ( Figure 3A). HpARI treatment had a similar effect to HpARI_CCP1/2 when IL-33 was assessed by western blot. Quantification of band intensities by densitometry reflected this increase of mature IL-33 signal in the presence of HpARI or HpARI_CCP1/2 (Supplementary Figure 2C). The difference in IL-33 signal strength between ELISA and western blot in the presence of HpARI was seen in a previous study (8), and is thought to be due to interference with antibody binding to the endogenous IL-33-HpARI complex in ELISA, but in a denaturing western blot proteins from this complex are dissociated and available for antibody detection. Together, this data suggests that binding of IL-33 by HpARI or HpARI_CCP1/2 stabilizes the mature cytokine, protecting it from degradation.
To assess the activity of the cytokine released, we induced necrosis of CMT-64 cells via freeze-thaw treatment. This treatment could be carried out in complete culture medium (without toxic additives such as Triton-X100), allowing downstream assessment of cellular responses to the released cytokine. On thaw, necrotic CMT-64 cells were incubated for up to 48 h at 37 • C, and IL-33 levels in supernatants assessed by ELISA. Similarly to Triton-X100-mediated necrosis, we found high levels of IL-33 released rapidly after freeze-thaw necrosis, which gradually decreased over the 48 h timecourse in control wells, while IL-33 levels increased over the timecourse in the presence of HpARI_CCP1/2 ( Figure 3C). These supernatants were applied to total bone marrow cells from IL-13eGFP +/GFP reporter mice (15) cultured in the presence of IL-2 and IL-7 (to support ILC2 differentiation), and cytokine responses were assessed 5 days later. As shown in Figure 3D, control freezethaw CMT-64 supernatants could only induce bone marrow cell IL-5 and IL-13 production at early timepoints post-thaw, implying that after ∼6 h post-thaw, all IL-33 present in the culture medium was inactive. This response appeared IL-33dependent as HpARI entirely inhibited IL-5 and IL-13 release. In contrast, supernatants from cells freeze-thawed in the presence of HpARI_CCP1/2 were able to maintain high levels of IL-5 and IL-13 stimulation (∼10-fold higher than the peak production seen in control wells) and this stimulation was maintained even when supernatants had been incubated for 48 h post-thaw. To specifically assess the ILC2 response within these total bone marrow cell cultures, we used flow cytometry for IL-13eGFP reporter or CD25 expression on ICOS + lineage − CD45 + ILC2s to confirm that these cells were activated by supernatants from medium of freeze-thaw control wells at early (45 min post-thaw), but not late (48 h post-thaw) timepoints, while wells containing HpARI_CCP1/2 remained highly activated throughout the timecourse (Figure 3E and Supplementary Figure 3).

DISCUSSION
HpARI blocks IL-33 responses and is secreted by H. polygyrus, as part of a suite of immunomodulatory effector molecules which act to prevent immune-mediated ejection of the parasite (27). HpARI acts by binding to IL-33 through the HpARI CCP2 domain and to genomic DNA in necrotic cells through the HpARI CCP1 domain, tethering the cytokine within the necrotic cell nucleus and preventing its release (8). Here, we further characterize these interactions, showing that a synthetic, nonnatural construct lacking the CCP3 domain (HpARI_CCP1/2) binds IL-33 with an approximately 10-fold lower affinity than the full-length HpARI protein, and lacks the blocking activity of HpARI against IL-33-ST2 interactions. Furthermore, HpARI_CCP1/2 had the surprising effect of stabilizing and amplifying IL-33 responses in vitro and in vivo.
As opposed to HpARI_CCP1/2, HpARI_CCP2/3 showed high affinity binding to IL-33, and prevented ligation of ST2 by IL-33, replicating the IL-33-blocking effects of full-length HpARI. In a previous study (8), we showed that HpARI_CCP2/3 lacked the DNA-binding activity of full-length HpARI and HpARI_CCP1/2, implying that this activity is mediated by the CCP1 domain. We previously also showed that HpARI_CCP2/3 increased, rather than decreased IL-33 levels in the bronchoalveolar lavage of mice 15 min after Alternaria allergen treatment. Our work here supports the hypothesis that this increase in IL-33 is due to HpARI_CCP2/3 preventing the rapid uptake and degradation of bound IL-33 by ST2-expressing immune cells (13,28,29), while lacking the DNA-binding (and hence tethering function) of HpARI or HpARI_CCP1/2. Thus, all IL-33 released is retained in the bronchoalveolar lavage, leading to increased IL-33 levels compared to controls.
IL-33 is known to mediate parasite expulsion in a type-2 dependent-manner (22). The HpARI_CCP1/2 truncated protein maintains the activity of IL-33, potentially amplifying its antiparasitic effects. It is worthwhile emphasizing that this truncated construct is not a protein naturally secreted by the parasite, but rather a synthetic product with an unexpected activity.  As the IL-33 pathway is strongly implicated in human asthma, HpARI, with its unique mechanism of action and strong binding to IL-33, is a potential therapeutic agent. IL-33 is a potently inflammatory cytokine which is kept tightly regulated. Once released, IL-33 undergoes rapid oxidation and degradation, confining its effects to a short time after release (12,13). Addition of HpARI or HpARI_CCP1/2 prevented degradation of the cytokine and maintained it in its active form, possibly due to steric hinderance of proteases. As HpARI also blocked the interaction of IL-33 with its receptor there was no cellular response to IL-33 in the presence of HpARI, while HpARI_CCP1/2, which lacks this IL-33-ST2 blocking activity, was unable to inhibit responses to IL-33. Furthermore, most surprisingly, HpARI_CCP1/2 was able to maintain the effects of IL-33 over a long timecourse, potently exacerbating IL-33dependent responses in vivo and in vitro.
The effects of HpARI_CCP1/2 may not be confined to extending the half-life of IL-33 by preventing its degradation, but may prevent the much more rapid oxidation of the cytokine. Partial oxidation of IL-33 occurs in vivo within 15 min of release (13), therefore the activity of released IL-33 in vivo may be less than that of fully active IL-33. Indeed, when a purified wild-type or an oxidationresistant mutant of human IL-33 were tested in vitro, the mutant form of IL-33 was found to be 30-fold more potent than WT IL-33 (13). In this study, we were not able to measure the difference between reduced and oxidized IL-33, therefore we cannot make definitive statements about this activity of HpARI_CCP1/2. However, inhibition of IL-33 inactivation, either through prevention of oxidation or proteolytic degradation, could be a potent method for amplifying IL-33-dependent responses.
HpARI_CCP1/2 could also be a useful tool for IL-33 research. Modulating IL-33 responses by using HpARI and HpARI_CCP1/2 in parallel allows assessment of the role of IL-33 in a system in the absence of potentially confounding effects of recombinant cytokine administration or genetic manipulation. In addition, the strategy of IL-33 stabilization by HpARI_CCP1/2 may be able to be replicated using a monoclonal antibody-based therapy, with low-affinity or non-blocking antibodies potentially able to amplify IL-33 responses. As anti-IL-33 treatments enter clinical trials (37), this is an important consideration, as suboptimal antibodies could result in amplification rather than suppression of IL-33 responses.
This study sheds further light on the mechanism of binding of HpARI to IL-33, the function of the domains of HpARI, and the effects of IL-33 degradation and inactivation. Further structural characterization of HpARI-IL-33 binding will be useful in characterizing this interaction and could allow guided design of more effective IL-33-blocking or IL-33-amplifying therapeutic agents.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/Supplementary Material.

ACKNOWLEDGMENTS
Flow cytometry data was generated with support from the QMRI Flow Cytometry and cell sorting facility, University of Edinburgh.