Efficacy of Chemokine Receptor Inhibition in Treating IL-36α–Induced Psoriasiform Inflammation

Several forms of psoriasis are associated with higher-than-normal levels of IL-36 cytokines in the skin (1–4). One rare type of psoriasis, generalized pustular psoriasis (GPP), appears to result directly from loss-of-function mutations in IL36RN, a gene encoding an important negative regulator of the IL-36 receptor (5, 6). GPP is a rare disease for which there is a dearth of clinical research and no universally accepted, evidence-based guidelines for its treatment and management (7). New therapies are needed because biologic therapies effective in more common forms of psoriasis are often ineffective in GPP (7, 8), and treatments that directly target GPP have not yet been developed (9–11).

We have recently reported an orally bioavailable small-molecule chemokine receptor antagonist effective in the imiquimod (IMQ)–induced murine model of psoriasis (12). Such chemokine-focused therapies are designed to block the migration of inflammatory leukocytes into tissues from the peripheral blood, thus preventing them from participating in and amplifying autoimmune or other inflammatory reactions. Chemokine ligands of CXCR2 and CCR6 are found within several types of psoriasis, including GPP (2, 13).

As an approach to identify novel therapies for IL-36–related skin diseases, we used a mouse psoriasis model in which activated IL-36α is injected directly into skin (2, 14). We hypothesized that blocking CCR6 and CXCR2 would be beneficial in treating IL-36α–induced skin inflammation. In this study, we have assessed the effectiveness of CCX624, a small-molecule CCR6/CXCR2 antagonist, in alleviating skin inflammation induced by direct injection of activated IL-36α into murine skin.

CCX624 was discovered and synthesized at ChemoCentryx (U.S. Patent Application no. 15/353,889; Mountain View, CA). We have previously published using CCX2553 as a CCR6/CXCR2 inhibitor (12). CCX2553 is from an earlier generation of compounds, and continued chemistry has led to the development of the more advanced compound CCX624. The chemistry effort through structure activity relationship focused on improving the potency and pharmacokinetic properties of the previous generation of compounds. CCX624 was specifically chosen for its improved potency on the murine CCR6 and CXCR2 receptors (Supplemental Table IA) as well as its improved clearance in rodents. These attributes allow for better and more reliable target coverage when used to dose mice.

For in vitro determination of chemokine receptor inhibitor activity, CCX624 was dissolved in DMSO to generate a stock concentration of 10 mM that was further diluted in chemotaxis migration buffer (HBSS, 1% BSA, and 1% HEPES) to create a 10-point inhibitor gradient ranging from 100 to 0.01 nM. Migration assay was then run in 100% mouse serum in response to rmCCL20 for CCR6 activity or rmCXCL1 for CXCR2 (mCCR6–transfected BaF3 and mouse bone marrow cells) activity using ChemoTx plates (Neuro Probe, Gaithersburg, MD) and assessed by fluorescence staining of migrated cells with CyQUANT (Thermo Fisher Scientific).

BALB/c mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and housed at ChemoCentryx animal facility in accordance with guidelines described in the Guide and Use of Laboratory Animals of the National Research Council. All studies were approved by the ChemoCentryx Institutional Animal Care and Use Committee.

For in vivo dosing, mice were injected s.c. with CCX624 in 1% HPMC once daily at 90 mg/kg, unless stated otherwise, starting on day 0 of a given study. This 90 mg/kg in vivo dosage was chosen based on in vitro potency in mice. The minimum antagonist concentration (at trough) to fully cover a gradient-sensing chemoattractant receptor has been determined to correspond with the IC₉₀ concentration (15). The IC₉₀ of CCX624 in 100% serum is 45 ng/ml for CCR6 inhibition and 145 ng/ml for CXCR2 inhibition. The actual trough concentrations achieved in vivo for CCX624 at 30, 60, and 90 mg/kg are shown in Supplemental Table IB.

The right ear of 8-wk-old BALB/c mice received four daily intradermal injections of recombinant truncated murine IL-36α Arg8-His160 (20 μl of a 150 μg/ml solution per mouse, totaling 3 μg/mouse/d formulated in PBS; BioLegend). Intradermal injections of PBS were given to the left ear as a control. Ear thickness was measured prior to start of study and each day throughout the study using a digital micrometer.

The IMQ model of psoriasiform dermatitis was performed as previously described (12, 16). Briefly, 5% IMQ (Fougera Pharmaceuticals) was applied daily to the shaved and depilated backs of BALB/c mice for up to 4 d. Control mice were treated with an application of Vaseline. Erythema, desquamation, and skin thickening were scored independently on a scale from 0 to 4, in which 0 = no disease, 1 = slight disease, 2 = moderate disease, 3 = marked disease, and 4 = very marked disease. These measurements each contributed to the total psoriasis activity and severity score (scale 0–12), which served as measurement of inflammation (12, 16). Skin from the backs of mice was excised at the end of the study and processed for flow cytometry analysis.

Ears or excised skin samples were digested in collagenase A and 1 U/ml DNAse I with agitation for 30 min at 37°C. Cells were then dislodged from skin, filtered through a 70-μM sieve, washed, and resuspended in FACS buffer (PBS with 10% FBS) for analysis.

Directly conjugated mAbs were from R&D Systems (Minneapolis, MN), BioLegend (San Diego, CA), or eBioscience/Thermo Fisher Scientific (San Diego, CA). CD45.2 (104) in Alexa Fluor 488, Ly6C (HK1.4) in PE, CD90.2 (30-H12) in PerCP-Cy5.5, TCRVγ4 (UC3-10A6) in PE-Cy7, TCRβ (H57-597) in APC, CD4 (GK1.5) in APC/eFluor 780, TCRγδ (GL3) in BV-421, LIVE/DEAD Fixable Aqua Stain (Molecular Probes, Eugene, OR), CD8α (53-6.7) in BV-605, CD11c (N418) in BV650, Ly6G (1A4) in BV711, and CD11b (M1/70) in BV785. Flow cytometry data were acquired with a Fortessa (BD Biosciences) cytometer and analyzed using FlowJo v10.2.

Flow cytometry of leukocytes extracted from inflamed skin shown in Fig. 2 and Supplemental Fig. 2 involved a panel of unconjugated monoclonal cells (or their isotype-matched controls), followed by an appropriate APC-conjugated polyclonal (FAB)₂ (Jackson ImmunoResearch Laboratories, West Grove, PA). These unconjugated reagents included anti-CCR6 (140706) from R&D Systems and anti-CXCR2 (SA044G4) from BioLegend, all rat IgG2as. Unconjugated mAbs against lineage markers CD103, CD11c, FLT3, CD205, and F4/80 were all obtained from BioLegend. After staining with unconjugated monoclonal and APC polyclonal, the cells were blocked with 10% mouse, 10% hamster, and 10% rat sera (Jackson ImmunoResearch Laboratories) prior to staining with direct conjugates.

Mouse anti–IL-17RA (R&D Systems) was dosed as a positive control at 200–500 μg/mouse, i.p., every day. The 200 μg/ml/mouse/d dosage was found to be well within the plateau of maximal activity for this treatment (Fig. 3A).

Ear tissue collected after 4 d of intradermal IL-36α treatments was homogenized in cold PBS containing protease inhibitors (Roche) then centrifuged to remove debris. The soluble fraction was assayed using a multiplex assay kit according to the manufacturer’s instructions (R&D Systems) and read on a MAGPIX (Luminex) analyzer. The concentration of tissue chemokines was normalized against the total protein levels measured for each sample using the standard Bradford assay.

Ear tissue was initially fixed in paraformaldehyde for paraffin embedding. Samples were processed by standard procedures. Epidermal thickness was measured as the average of seven measurements made along the center third (the area containing the lesion) of the length of the H&E-stained ear sections using Photoshop CS4 software (Adobe).

Statistical significance was determined by Mann–Whitney calculation (unless specified otherwise) using GraphPad Prism 6.0 software.

We compared the inflammatory cell populations obtained from the inflamed skin of two murine psoriasiform dermatitis models: the well-established IMQ model (12, 16), in which the TLR7/8 agonist IMQ is applied daily to the surface of depilated skin; and a newer model introduced by Foster et al. (2), in which truncated, activated mouse IL-36α is administered by intradermal injections (14).

After 4 d of treatment, flow cytometry showed inflammatory cell infiltrates isolated from skin (gated as shown in Supplemental Fig. 1A) differ appreciably between the two models. As demonstrated in previous studies involving both the C57BL/6 and BALB/c strains (17–20), IMQ treatment generated a large neutrophil population (Fig. 1A, 1C). In contrast, activated IL-36α generated nearly equal numbers of T cells, neutrophils, and Ly6C^hi/Ly6G⁻ myeloid cells (Fig. 1B, 1C). The T cell populations from each model were also highly divergent from each other (Fig. 1D); IMQ-treated skin (red bars) had a prominent γδT17 (Vγ4⁺) population, as previously observed (12, 21), whereas T cells from IL-36α–injected skin (blue bars) consisted almost entirely of CD4⁺ conventional αβ T cells.

The large population of myeloid cells (in similar abundance to neutrophils and CD4 T cells) observed by flow cytometry in IL-36α–treated skin (Fig. 1B, 1C) supports previous immunohistological findings (2). We found these myeloid cells to express a combination of markers characteristic of both myeloid and dendritic cells (DCs), including CD103, CD11c, and F4/80. They did not express Flt3, a marker diagnostic of classical DCs, nor did they express CD205 (DEC205) (Supplemental Fig. 1B). Based on this immunophenotype and the location of these cells within actively inflamed skin, we believe these cells to be monocyte-derived inflammatory DCs (iDCs) (22) and will refer to them as such for the remainder of the report.

We observed that chemokine ligands for CCL20 and CXCR2 were significantly increased in skin by treatment with IL-36α (Fig. 2A), in agreement with previous work (12, 16). We used flow cytometry to examine CCR6 and CXCR2 expression on each of the three leukocyte subtypes comprising most of the CD45⁺ infiltrate that accumulated in response to IL-36α treatment (Fig. 2B). Of these three, only neutrophils expressed CXCR2 levels greater than isotype control staining (Fig. 2B, left column, middle row). A minority of the iDCs and a majority of CD4 T cells expressed CCR6 (Fig. 2B, right column).

The prominence of CCR6 and CXCR2 ligand in the inflamed skin coupled with the expression of these two receptors by the IL-36α–induced subsets prompted us to assess a CCR6/CXCR2 antagonist for efficacy in preventing these cells from accumulating within the skin (and thereby alleviating IL-36α–induced skin inflammation). CCX624 is a small molecule with a m.w. of∼440 and an IC₅₀ of 10 nM on mouse CCR6 and 20 nM on mouse CXCR2, as assessed by in vitro inhibition of chemotaxis in 100% serum (see Supplemental Table IA) and is further optimized from a previously reported compound, CCX2553, that was efficacious in the IMQ model (12). Consistent with our previous report using CCX2553, CCX624 is effective at alleviating IMQ-induced skin inflammation and reducing the accumulation of γδT17 cells in the inflamed site (Supplemental Fig. 2) (12).

We measured ear thickness after four daily intradermal injections of PBS alone (control) or activated IL-36α in PBS (Fig. 3). The mean thickness of PBS-treated ears was∼0.2 mm, half as thick as the IL-36α–treated ears (∼0.5 mm, Fig. 3A). CCX624 was s.c. dosed once daily (on the back of the mouse, distal from the IL-36–treated ear) and achieved dose-dependent decreases in IL-36–induced ear thickening (Fig. 3A). A time course revealed that the effects of CCX624 were appreciable after the second day of treatment (Fig. 3B; Supplemental Fig. 3A).

We next compared the effectiveness of CCX624 with that of an anti–IL-17RA mAb (mice were given a dosage of 200 or 500 μg/mouse i.p. once daily). Anti–IL-17RA treatment was effective at reducing IL-36–induced ear swelling (Fig. 3A) but was significantly less effective than the 90 mg/kg dosage of CCX624. The separation in effectiveness between a saturating dose of anti–IL-17RA mAb and dosage of 90 mg/kg CCX624 became evident after the second day of treatment (Fig. 3B; also, see Supplemental Fig. 3B for direct comparison of anti–IL-17RA mAb treatment with the isotype-matched control for this mAb at 500 μg/mouse/d).

We next examined sections of mouse ears taken after four daily injections of activated IL-36α (or control) for measurement of epidermal thickness (Fig. 4). In agreement with the direct measurements of ear swelling shown in Fig. 3, epidermal thickening was evident in the IL-36α–treated ear versus the PBS-treated ear (Fig. 4A). Administration of CCX624 or anti–IL-17RA significantly reduced IL-36α–mediated epidermal thickness (Fig. 4B; please note that the biological effects of anti–IL-17RA were shown to be saturating at 200 μg/mouse/d dosage in Fig. 3A, which was the dosage chosen for Figs. 4 and 5).

We next examined the effects of CCX624 on the inflammatory cell subsets accumulating within IL-36–treated skin by flow cytometry (Fig. 5). Immune cells were isolated from the skin after four daily ear injections of IL-36α and treatment with CCX624 or anti–IL-17RA mAb. We found that CCX624 significantly reduced the accumulation of CD4⁺ T cells, neutrophils, and inflammatory iDCs. In contrast, anti–IL-17RA did not reduce the accumulation of CD4⁺ T cells, but had effects similar to CCX624 on neutrophils and iDCs.

Psoriasis is composed of multiple subtypes, including pustular psoriasis, a rare skin disorder further subdivided into generalized and localized forms (7, 10, 11). The generalized form (GPP) is associated with significant morbidity and, in some cases, mortality (23). GPP is characterized by a widespread eruption of pustules and erythematous plaques. In the acute variant, patients usually appear systemically ill because the sudden eruption of pustules is accompanied by pain and fever. Life-threatening leukocytosis, electrolyte abnormalities, hypoalbuminemia, and elevated liver enzymes can also occur in the acute variant (7).

Because of the rarity of this disease, data on treatment outcomes consist primarily of retrospective studies, case reports, and expert opinions (7, 9). Current treatments involve supportive care for those patients who are systemically ill, followed by medical treatments to control the skin disease (9). Chronic, slowly progressing disease is typically managed by oral retinoids or methotrexate. Acute disease is treated with cyclosporine or one of the anti–TNF-α biologics. None of these treatments are highly effective, and all have side effects (9). There are no U.S. Food and Drug Administration–approved treatments for GPP. Thus, there is an unmet clinical need for the identification of new targets and novel therapeutics for GPP management.

Loss-of-function mutations in the IL36RN gene are common in GPP patients, especially those without concomitant symptoms of plaque psoriasis (5). The IL36RN gene encodes a protein known as the IL-36R antagonist (IL-36RA). In healthy individuals, IL-36RA competes with activated IL-36α, β, and/or γ cytokines for binding to the IL-36R, and these anti-inflammatory properties help maintain homeostasis. Aberrant structure and function of IL-36RA engenders dysregulated secretion of inflammatory cytokines and chemokines (5). In this study, we report that an optimized small-molecule antagonist targeting CCR6/CXCR2 was more effective than anti–IL-17 therapy in reducing skin thickness and leukocyte infiltration after direct intradermal injection of activated IL-36α into mouse skin.

Small-molecule antagonists targeting CXCR2, including SCH57123 (24) and AZD5069 (25), have safely advanced to human clinical trials without causing increased prevalence of immune related adverse events, suggesting that CXCR2 antagonists can be safely dosed in humans. Although small-molecule CCR6 antagonists have not yet been dosed in human subjects, the CCR6 gene has been ablated in mice by three independent groups (26–28), and none reported increased susceptibility to bacterial infections in these strains. Thus, we see no clinical or preclinical rationale to preclude molecules that antagonize pharmaceutical targets such as CXCR2/CCR6 from further therapeutic exploration and for careful advancement to intervention in psoriasiform diseases.

To begin to understand the mechanisms driving skin inflammation resulting from a dysregulated IL-36 cytokine axis, we examined the populations of cells that accumulate within IL-36α–treated skin. In addition to CD4⁺ T cell and neutrophil populations, we detected large numbers of myeloid cells expressing markers characteristic of iDCs. Because these cells did not express FLT3/CD135, we believe them to be derived from monocytes rather than classical DCs. Significantly, a similar, although less well-characterized, myeloid cell population resembling these iDCs has also been detected in the skin of patients with GPP (7), and it has been previously demonstrated that DC derived from human blood monocytes express high levels of IL-36R (2).

We used the IL36α-injection model to compare the effectiveness of a specific CCR6/CXCR2 antagonist with that of an anti–IL-17RA mAb. Biologics against the IL-17 axis are often used as a second-line treatment for GPP in the clinic (29). We found the 90 mg/kg dosage of CCX624 to be significantly more effective at reducing ear swelling than a saturating dosage (200 or 500 μg/mouse/d) of the anti–IL-17RA Ab (Fig. 3).

CCX624 is an optimized small-molecule antagonist of CCR6 and CXCR2, and it appears that both of these activities contribute to its efficacy in the IL-36α injection model. Anti–IL-17RA mAb, even at the very high dosage of 500 μg/mouse/d was less efficacious than CCX624 in reducing disease activity. CCX624 significantly reduced the numbers of CD4⁺ αβ T cells, neutrophils, and iDCs in the IL-36–treated skin, whereas the anti–IL-17RA mAb reduced only the neutrophils and iDCs. Taken together, we interpret these results to indicate that the CCR6 antagonist activity of CCX624 on CCR6⁺ T cells and iDCs acts in concert with the CXCR2 antagonist activity on neutrophils to reduce immune cell infiltration and disease activity in this model. CCX624 thus represents a novel, and mechanistically sound, therapeutic agent for diseases involving a dysregulated IL-36 cytokine axis.

In summary, we have shown that the inflammation resulting from activated IL-36α skin injections involves neutrophil, iDC, and CD4 T cell accumulation, similar to what is seen in GPP patients. Selective inhibition of CCR6 and CXCR2 by CCX624 reduced all three of these inflammatory cell types and alleviated skin inflammation. Blockade of the IL-17 axis reversed neutrophil and iDC, but not CD4 T cell, accumulation in skin. CCX624 is a more effective therapeutic agent than the saturating concentrations of an anti–IL-17RA mAb assessed in this IL-36α–induced model of psoriasis. These findings suggest that CCR6/CXCR2 antagonism may constitute a novel target class and a mechanistically distinct therapeutic approach to treating dysregulation of the IL-36 cytokine axis (as in GPP) by specifically acting upon the inflammatory cells that likely mediate the disease.

All authors are full-time employees of ChemoCentryx, Inc.