Modulating immune toxicity in ulcerative colitis using a small molecule RAB27A inhibitor

Abstract

Background

Recent evidence has shown that ulcerative colitis patients with increased neutrophil biomarkers are linked to complicated disease and refractory response. In the pathology of colitis, neutrophils are recruited to the colon and yield acute damage via reactive oxygen species (ROS), degranulation, and neutrophil extracellular traps (NETosis). RAB27A is a regulator of intracellular membrane trafficking and has been shown to play a role in neutrophil activation by enabling the increased release of proteolytic granules and ROS. Studies have shown significant upregulation of RAB27A expression in biopsies of moderate and severe ulcerative colitis patients, while preclinical studies genetically deleting RAB27A in ulcerative colitis models have shown improved outcomes.

Methods

To evaluate the impact of RAB27A inhibition in ulcerative colitis and lower activated neutrophil toxicity in the disease, we developed IMYX-3, a first-in-class small molecule oral inhibitor of RAB27A activity and applied it to the study of diseased neutrophils and models.

Results

IMYX-3 lowers neutrophil ROS, degranulation, NETosis, and cytokine secretion. Preclinical efficacy was demonstrated in dextran sulfate sodium and Il10^−/− mouse models, in which IMYX-3 treatment led to improved colon length, reduced systemic inflammation, lower levels of activated neutrophil function, and comparable disease outcomes compared with standard-of-care control mice. In vitro toxicity profiling as well as pharmacokinetic and toxicology studies in rodents and canines showed a favorable drug profile, with no significant adverse events observed at doses up to 30 times the predicted therapeutic range.

Conclusion

IMYX-3 effectively inhibits RAB27A activation in neutrophils, yielding improved outcomes in preclinical disease models with no major safety concerns.

Graphical abstract

Graphical Abstract.

Key Messages.

• Neutrophils are found in high numbers in the diseased colon of patients with ulcerative colitis, yet their exact role remains unclear.

• There are no therapeutic modalities known to directly modify the activity of neutrophils in ulcerative colitis.

• IMYX-3 is a first-in-class RAB27A inhibitor that reduces neutrophil-mediated damage by decreasing ROS, degranulation, and NET formation.

• Preclinical studies in DSS and IL10^−/− mice demonstrate improved disease outcomes including longer colon length, reduced systemic inflammation, histological remission, and lower levels of activated neutrophil function.

• IMYX-3 also shows favorable in vitro and in vivo safety profiles, supporting potential therapeutic application in ulcerative colitis patients.

Introduction

Ulcerative colitis (UC) is a chronic, idiopathic inflammatory disease that affects the colon. The disease is characterized by chronic progressive mucosal inflammation, starting in the rectum and variably extending to proximal segments of the colon. Ulcerative colitis tissue from patients is infiltrated by numerous immune cell types, with neutrophils and natural killer cells being a significant presence in the inflamed colon. The aim of current therapy is to reduce inflammatory damage to maintain clinical and endoscopic remission to prevent debilitating symptoms. Key treatments for the advanced disease include tumor necrosis factor α inhibitors like infliximab (Remicade) and adalimumab (Humira); vedolizumab (Entyvio), an integrin inhibitor; Xeljanz, a JAK inhibitor; and ustekinumab (Stelara), which targets interleukin (IL)-12 and IL-23, all major immune pathway modulators. However, despite major advances in inflammatory bowel disease (IBD) therapeutics, advanced biologics and development-stage drugs are only effective in ∼25% of patients who have failed conventional therapies.^1,2

A possible cause of the refractory patient population is that most drugs used for UC have a mechanism of action which directly modifies the T cell response and only modulate activated neutrophil toxicity through intermediate regulators. Developing novel therapeutics for UC with the capacity to modulate neutrophil activation has been supported by several preclinical and clinical evaluations. In animal models of IBD, neutrophil depletion using antineutrophil antibodies ameliorates colitis induced by dextran sulfate sodium (DSS) or trinitrobenzene sulfonic acid.^3,4 Specifically, neutrophils are found in high number in intestinal inflammation biopsies (particularly in crypt abscesses described as focal active colitis), and the extent of neutrophil infiltration correlates with the severity of the disease as part of the scoring system of UC severity.^5,6 Moreover, patients with high levels of neutrophil markers such as fecal calprotectin and fecal myeloperoxidase (MPO) have a significantly higher probability of developing complicated IBD or becoming refractory to drug treatment.⁷ Last, granulocyte/monocyte apheresis (GMA), which removes activated neutrophils from the bloodstream, is used as an effective first-line therapy in Japan and some areas of Europe for treatment of UC.^3,8

RAB27A belongs to the Ras superfamily of monomeric G proteins and localizes to distinct cellular compartments to regulate specific steps of intracellular membrane trafficking and secretion.^9–11 Notably, the protein is expressed and is significantly upregulated in the intestinal epithelial cells during UC.¹² Beyond the basic mucosa, RAB27A is also highly expressed in immune infiltrates found in patient samples of colitis.^11,13 RAB27A activity is crucial for toxic activity of difficult-to-target immune cell populations, such as neutrophils.^14,15 A recent study showed that genetic deletion of RAB27A improves disease outcomes in preclinical colitis models both in vitro and in vivo.¹⁶ Mouse models of RAB27A deficiency and humans lacking RAB27A (GS2) activity are viable and show no gastrointestinal symptoms, with specific side effects (skin discoloration and leukopenia) likely due to developmental defects that would not occur, and have not been found, with our drug inhibition approach.

Previous attempts to pharmacologically modulate RAB27A have yielded structures shown to down-regulate neutrophil toxicity. Specifically, Nexinhib20 was identified as a small molecule capable of inhibiting RAB27A association with JFC1 via a high throughput screening assay.¹⁷ The molecule was then found to effectively down-regulate neutrophil activation in vitro. However, in vivo studies did not show a robust effect even at high concentrations of the drug administered intraperitoneally, which could be due to the lack of solubility. Lack of solubility is a central issue, given the advantages and preference of developing orally deliverable molecules for patients with chronic disease. Moreover, the molecule contains a nitro group, which has been shown to cause DNA damage, hepatotoxicity, and disrupted proteins in humans.¹⁸

To create a compound capable of inhibiting RAB27A in vivo for neutrophil modulation in disease, we developed a soluble and structurally safe RAB27A inhibitor based on literature knowledge of structures yielding inhibition activity. In this study, we present IMYX-3, which is a soluble and safe RAB27A inhibitor. First, we characterized the molecule using in vitro pharmacology assays, and then, second, showed its effect on neutrophil activation in vitro. Given the possible central role of neutrophils in UC, we evaluated the effect of IMYX-3 in two independent UC disease models. Last, we conducted toxicology assays on rodents and canines showing positive safety. Taken together, these results present an orally available small molecule developmental candidate for RAB27A inhibition in UC.

Methods

Animals

Eight-week-old C57BL/6 mice were purchased from Harlan. UC was induced in healthy 8-week-old male and female C57BL/6 mice by administering 3% DSS (MPBIO; 36 000-50 000 molecular weight MP Grade) in their drinking water for 5 consecutive days. Subsequently, the mice were provided with DSS-free water for 4 to 5 days. Animals were treated with IMYX-3 either from day 0 to day 8 or after DSS treatment, starting on day 4 until day 8. In all mouse experiments, IMYX-3 was prepared in phosphate-buffered saline (PBS) and dosed to animals by gavage at a volume of 100 µL. Animals were weighed daily, and colon length was measured at sacrifice.

Germ-free 129S6/SvEv background wild-type and Il-10–deficient (Il-10^−/−) mice were obtained from the National Gnotobiotic Rodent Resource Center at the University of North Carolina at Chapel Hill and transplanted with previously validated, mouse-adapted pooled human IBD patient fecal microbiota to induce chronic experimental colitis, as previously described.¹⁹ All animal experiments were conducted under approved Institutional Animal Care and Use Committee protocols. Serial measurement of fecal lipocalin 2 (fLcn2) was performed with a commercial ELISA kit (DuoSet; R&D Systems) as previously described. Formalin-fixed, paraffin-embedded colon specimens were prepared and stained for hematoxylin and eosin. These samples were then scored blinded by a trained histopathologist to quantify histologic tissue inflammation on a scale of 0 to 4 for 4 colonic tissue segments (cecum, proximal colon, distal colon, and rectum), as previously described.²⁰ Total histology score is the summation of all 4 segments. Neutrophil staining was performed with a monoclonal MPO antibody (rabbit) followed by horseradish peroxidase. MPO staining area was quantified with ImageJ2 (2.14.0/1.54F, National Institutes of Health).

Rat and canine experiments were performed at Wuxi AppTec under approved Institutional Animal Care and Use Committee protocols. IMYX-3 was prepared in PBS and dosed daily for 2 weeks by trained technicians using gavage appropriate for each species.

Neutrophil function assays

Neutrophil purification from mice and humans was conducted as previously described.²¹ All activation studies used 100 nM PMA (Sigma-Aldrich). Reactive oxygen species (ROS) production was evaluated by plating 10⁴ isolated normal density neutrophils in Hanks’ Balanced Salt Solution (HBSS) without phenol red into a white 96-well flat-bottom plate. Subsequently, each well received 20 μL of a Luminol solution at a concentration of 500 μM in HBSS, supplemented with 40 U/mL of horseradish peroxidase. Following this, the cells were exposed to 100 nM PMA. Chemiluminescence kinetics were observed for a duration of 60 minutes using the InfiniteF200Pro system from TECAN. Degranulation was quantified via flow cytometry on circulating neutrophils or isolated from the colon, utilizing specific Ly6G markers for neutrophils and CD11b markers present on the granules. Neutrophil apoptosis was measured by quantifying propidium iodide (1:100) staining in the CD66b⁺ population of cells. Mean fluorescence intensity was considered during analysis. Isolated leukocytes or neutrophils were incubated with Fc blocking reagent (MACS; Miltenyi Biotec) in PBS containing 0.5% bovine serum albumin, followed by incubation with the respective antibodies mentioned previously. The samples were then processed for flow cytometric analysis using either a MACSQuant analyzer (Miltenyi Biotec) or LSRFortessa cell analyzer (BD Biosciences). The data were subsequently analyzed using dedicated software packages FlowJo V10 (BD Biosciences).

Neutrophil bacteria killing assay

Escherichia coli (ATCC 25922 Seattle 1946 strain) were grown ON in LB tubes at 37 °C with shaking and erythromycin (stock 250 mg/mL ethanol, working 1:1000). Bacteria were diluted 1:5000 in a 50 mL tube and quantified based on optical density. E coli was then incubated with human neutrophils (100 000) with dilutions of neutrophil to bacteria of 1:10 in a final volume of 200 µL HBSS with and without IMYX-3 at increasing concentrations. A total of 50 µL of 2.5% serum solution in HBSS (final 0.4% serum) was added to activate opsonization, and samples were incubated at 37 °C for 4 hours. Following incubation, 10 µL of the 200 µL were plated 1:10 on regular LB plates ON. Colonies on the plates were counted after 16 hours to determine neutrophil killing.

Tissue histology

Tissue immunohistochemistry was performed on 5 μm sections from the FFPE tissue samples. Hematoxylin and eosin staining was performed using an automated stainer (Leica Biosystems). The following antibodies were used to stain the colon tissue: MPO staining (Abcam; ab300650 Rat) at 1:500 dilution and CitH3 staining (Abcam; ab281584 Rabbit) at 1:100 dilution.

ELISA and cytokine panel analysis

The quantification of IL-6 in serum was performed utilizing an ELISA kit (BioLegend) according to the manufacturer’s instructions. Mouse fLcn2 was extracted from fecal matter directly from the animal by homogenization in PBS with 0.1% Tween 20, then protein levels quantified using ELISA (DY1857; R&D Systems), as previously described.²² Human MPO was measured using an ELISA kit (Abcam) according to the manufacturer’s instructions for evaluating protein in cell culture supernatant. Cytokine panel screening was performed using a RayBiotech array (#QAH-CYT-1-2) for 20 human cytokines. Cell culture supernatant was used to quantify neutrophil cytokines secretion with and without induction by PMA.

RNA sequencing

HL-60 cells (ATCC) were cultured with 1% dimethyl sulfoxide for 7 days to induce differentiation and then stimulated with PMA (100 nM) together with 5 µM of IMYX-3 for 6 hours. RNA was extracted immediately using a Qiagen RNeasy extraction kit and then stored at −80 °C. RNA integrity and quantity were assessed on an Agilent Bioanalyzer 2100 using the RNA 6000 Nano assay chip. Samples with an RNA Integrity Number ≥8.0 were considered acceptable for library preparation. Library preparation and sequencing were performed by Azenta Life Sciences. Ribosomal RNA and globin RNA depletion was performed prior to library preparation to enrich for messenger RNA and low-abundance transcripts. Sequencing libraries were prepared using a stranded total RNA library preparation protocol and sequenced on an Illumina platform using a paired-end 2 × 150 bp configuration, generating a minimum of 20 million reads per sample. Raw sequencing reads were assessed for quality using FastQC, and adapter sequences and low-quality bases were trimmed using Trimmomatic (v0.39). Trimmed reads were aligned to the human reference genome (GRCh38/hg38, Ensembl annotation release 109) using STAR (v2.7.x) with default two-pass alignment parameters. Aligned reads were quantified at the gene level using HTSeq-count (or featureCounts) in stranded mode against the Ensembl GTF annotation file. Analysis was performed using Cytoscape (BiNGO) followed by analysis with Microsoft Excel. All included genes had a P value <.001.

Peptide mass fingerprinting

Reactions were conducted in a buffer containing 20 mM Tris (pH 8.0), 150 mM NaCl, 5 mM MgCl₂, and 0.5 mM TCEP.RAB27A (Construct code: FJE1, Lot: PC20737) was used at a final concentration of 35 µM. The compound was added at 712 µM, corresponding to a 20-fold molar excess over the protein, with a final dimethyl sulfoxide concentration of 2%. Reactions were incubated for 16 hours at 4 °C. Samples underwent proteolytic in-solution digestion followed by liquid chromatography–tandem mass spectrometry (MS/MS) analysis. Three separate digestions were performed using trypsin/LysC, chymotrypsin, and pepsin to ensure comprehensive sequence coverage. MS analysis was performed using an Agilent 6546 LC/Q-TOF system with an ACQUITY UPLC CSH C18 column. Data acquisition was conducted in auto MS/MS mode. Peptide mass fingerprinting results were matched against the RAB27A construct amino acid sequence using MassHunter BioConfirm. The compound was included as a variable amino acid modifier to assess conjugation.

In silico modeling

The protein structure used for docking was taken from PDB with the code 7 OPR. Cys188 in this structure was mutated to serine. For docking, this serine was mutated back to cysteine manually and the structure was then prepared and minimized using the Protein Preparation Workflow from the Schrödinger Software Suite (2021-04). All covalent docking jobs were performed using CovDock module from the Schrödinger Software Suite. Docking mode was set to Pose Prediction (Through) and default parameters were used.

The 50-ns unbiased MD simulations were performed with explicit solvent at 300K using the AMBER20 suite of programs. The atomic charge of ligands was generated by the GAFF2 with the AM1 bcc charges. The AMBER ff14SB force field was used for protein. The periodic cubic TIP3P water box was used as the solvent, and the minimum distance from the molecule to the boundaries of the water box was set to 12 Å. The LEaP module in AmberTools21 was used to parameterization every system. The parameterized system was first energy-minimized in total 2000 steps, including 1200 steps of the steepest descent method and 800 steps of conjugate gradient algorithm. The Langevin dynamics was used for heating the system, the system was gently heated to 300 K in 100 ps with a 1-ps-1 collision frequency. All heavy atoms were restrained by 20 kcal/(mol·Å2) harmonic potential during temperature-rise process. The nonbonded cutoff of all MD simulations was set to 8 Å.

Bioanalytical methods

An aliquot of 50 µL unknown sample (diluted blood and plasma), calibration standard, and controls were added to 1.5 mL tube. A 0.25 mL of internal standard mixture in 2M HCl was added, and the mixture was vortexed, and then incubated at 90 °C for 2 hours. After 2 hours, the mixture was neutralized by adding 0.25 mL of 2M NaOH. The mixture was vortexed and centrifuged for 15 min at 18 000 g, 4 °C. After centrifugation, 200 μL of each sample was transferred to well containing 600 μL of stop solution (5% TCA in H₂O) followed by mixing. Then, the mixture was vortexed for 1 min and centrifuged for 10 min at 12 000 g, 4 °C. An aliquot of 50 µL supernatant from each sample was transferred to 96-well plate and centrifuged for 5 min at 3220 g, 4 °C. Then, the supernatant was directly injected for liquid chromatography–MS/MS analysis.

Statistical analysis

Statistical analysis was performed with GraphPad Prism 10 (GraphPad Software). Significant differences (P < .05 and P < .01) reflect P values from 2-tailed t test.

Ethics statement

Animal experiments were performed in compliance with all relevant ethical regulations for animal testing and research as described. DSS studies received ethical approval from the Hebrew University Institutional Animal Care and Use Committee (MD-24-17499-5). IL10 knockout studies received ethical approval by the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee (18-266.0-B). The human fecal sample for transplantation to mice was collected in compliance with all relevant ethical regulations for work with human participants, and with informed consent from the participant under the approved University of North Carolina Institutional Review Board protocol #17-1528. The use of human neutrophils was approved for use by Hadassah Hospital Institutional Review Board (0615-16-HMO). Rat and canine experiments were performed at Wuxi AppTec under AAALAC International guidelines with study approval by internal IRB (069-002-CTX and 069-0006-CTX).

Results

To modulate the effect of activated neutrophils in disease, we created a soluble small molecule inhibitor of RAB27A. Following our medicinal chemistry program and lead compound selection, we performed liquid chromatography–MS intact mass analysis to validate the interaction between the inhibitor, IMYX-3, and purified RAB27A protein. We identified the sites of interaction using peptide mass fingerprinting at Cys123 and Cys188 in the RAB27A binding pocket. These findings were validated using in silico modeling to visualize the blocking interaction occurring in the protein binding pocket (Figure 1).

Figure 1.

In silico modeling of IMYX-3 Rab27A interaction. Computer modeling based on x-ray crystallography shows the molecule’s interactions.

We then assessed the functional capacity of IMYX-3 to inhibit RAB27A activity in immune cells by evaluating the blockade of PMA-activated human neutrophil ROS in response to treatment with the molecule. Using isolated human neutrophils activated by PMA (100 nM), we determined the half maximal inhibitory concentration of IMYX-3 in vitro at 0.36 µM (Figure 2A). A decrease in degranulation activity was also seen by evaluating cytokine secretion levels in human neutrophils, in which a dose-dependent response in secretion of multiple factors was demonstrated (Figure 2B; Supplementary Data 1). NETosis was also decreased by IMYX-3 (Figure 2C), yet apoptosis was not induced by the compound in primary neutrophil assays (Figure 2D). Finally, to gauge potential changes in human neutrophil inflammatory profiles, we activated differentiated HL-60 cells (a neutrophil-like immortalized cell line) with PMA (10 nM) and IMYX-3 (5 µM) and performed RNA sequencing comparing the PMA activated expression vs IMYX-3–treated PMA sample expression (Figure S1). IMYX-3–treated HL-60 cells demonstrated downregulation of transcription of neutrophil inflammatory proteins. Although HL-60 cells are neutrophil-like cells and not true primary neutrophils, these data provide a strong indication for the potency of IMYX-3 in regulating neutrophil-mediated inflammatory responses.

Figure 2.

IMYX-3 inhibits RAB27A mediated immune toxicity in vitro. (A) Reactive oxygen species (ROS) activity stimulated by PMA (100 nM) is blocked at an half maximal inhibitory concentration of 0.36 µM human (n = 3) when cultured with 10 000 cells, and degranulation is also significantly decreased, (B) as measured by a cytokine secretion panel in human neutrophils. (C) NETosis is inhibited by IMYX-3 (1 µM) when tracked by Incucyte over 5 hours following PMA induction (100 nM). (D) Primary white blood cells were cultured with increasing concentrations of IMYX-3 for 3 hours and then costained for propidium iodide (PI) and CD66B⁺ to evaluate induction of apoptosis by fluorescence-activated cell sorting. AU, arbitrary units; HDF, high-density fraction; IL, interleukin; MCP-1, monocyte chemoattractant protein-1; MIP, macrophage inflammatory protein; MPO, myeloperoxidase; PBS, phosphate-buffered saline; VEGF, vascular endothelial growth factor.

Following confirmation of target binding and activity, we assessed the in vitro toxicity and metabolism of the molecule using a variety of standard assays available for compound profiling (Table 1). The molecule was evaluated for AMES mutagenicity, CYP inhibition, liver microsome assays, multispecies in vitro metabolites (MET-ID), and 44-panel safety. No significant concerns were found during the toxicity profiling studies (Table 1). Prior to initiation of in vivo studies, we prepared a salt formulation of IMYX-3 using HCl and increased synthesis to a batch of 500 g, demonstrating the ability to scale up molecule synthesis.

Table 1.

In vitro toxicity and metabolism profiling of IMYX-3.

Assay	Results
AMES mutagenicity	No mutagenicity
CYP binding (IC50 uM)	IC50 > 50 µM for CYP1A2, CYP2C9, and CYP2C19 and ∼5 µM for CYP2D6 and CYP3A-M
Liver microsome human and mouse (T1/2 min)	45.9 and 34.2
In vitro metabolites (Met-ID)	<2% Glutathione conjugated drug metabolite
44 panel safety	43 receptors under 50% (1 at 66% hERG)
Multispecies hepatocyte stability (T1/2 min)	36.7 (Mouse), 40.7 (Rat), 54.7 (Dog), 37 (Monkey), 39.6 (Mini-Pig) 43.3 (Human)

Six assays were used to assess the potential for IMYX-3 to cause toxicity in vivo, based on standard pharma development tools. No significant toxicity issues were identified during the profiling. Multispecies MET-ID found similar profiles between mouse, rat, dog and human.

To determine the effect of IMYX-3 on experimental colitis in vivo, we used the DSS-induced model of intestinal damage and assessed dose response to IMYX-3 using male and female C57BL/6 mice treated with 3% DSS. We first treated animals with DSS for 5 days in drinking water to induce experimental colitis and dosed in parallel with IMYX-3 via gavage. DSS was discontinued after 5 days, and IMYX-3 oral dosing was continued until day 8. We identified a dose-responsive decrease in inflammation of the drug, yielding significantly improved colon length; lower systemic neutrophil toxicity, as measured by Ly6G⁺CD11b⁺ (day 4); and decreased colon ROS levels (day 5) (Figures 3A–C). Treated animals also showed lower levels of colonic neutrophil NETosis, as identified by costaining intestinal sections with MPO and CitH3 (Figure 3D). To determine whether the effect of the drug was systemic or local in the colon, we administered IMYX-3 rectally to the animals using the same DSS protocol, with dosages adjusted for biodistribution reaching the colon when the drug is administered orally (Figure S2). No difference was seen in colon length when the drug was delivered rectally, suggesting that the effect of IMYX-3 occurs in the systemic environment and not specifically in the colon.

Figure 3.

IMYX-3 coadministration in the dextran sulfate sodium (DSS) mouse model of inflammatory bowel disease via oral administration. In parallel to disease induction with DSS 3% for 5 days, mice were treated orally with IMYX-3 at ascending doses to day 8 and showed improved (A) colon length, (B) systemic neutrophil CD11b (day 5), and (C) reactive oxygen species (ROS) levels in the colon (day 4). (D) Colon neutrophil NETosis was also found to decrease in animals treated with 5 mg/kg IMYX-3 based on %CitH3⁺MPO⁺ cells. *P < .05; **P < .01; ***P < .001; ****P < .0001. MFI, mean fluorescence intensity; ns, not significant.

To better reflect colon regeneration in a disease setting, mice were first administered DSS for 5 days followed by IMYX-3 orally at 0.5 mg/kg and 5 mg/kg for 4 days. IMYX-3 was compared with the drug ABX-464 (40 mg/kg) as a positive control, given ABX-464’s oral dosing, previously reported therapeutic activity in the DSS model,²³ and positive performance in clinical studies for IBD.¹ We found that IMYX-3 at both concentrations yielded a significantly improved colon length, animal weight, and IL-6 level, showing superiority of IMYX-3 to ABX-464 in all categories (analysis of variance) (Figures 4A–C).

Figure 4.

Oral IMYX-3 treatment following dextran sulfate sodium (DSS) induction. (A) After disease induction with DSS 3% for 5 days, mice were treated orally with IMYX-3 at 0.5 mg/kg and 5 mg/kg, and compared with positive control ABX-464 (40 mg/kg), showing significant improvement and superiority in (A) colon length, (B) weight, and (C) interleukin (IL)-6 levels. *P < .05; **P < .01. MC, methylcellulose; ns, not significant; PBS, phosphate-buffered saline.

Next, we evaluated the effect of IMYX-3 in the Il10^−/− mouse model of chronic T cell–mediated colitis, which is considered the gold standard for predicting Il10-independent drug response in UC. Eight-week-old male and female Il10^−/− mice were transplanted with mouse-adapted human IBD patient fecal microbiota originating from 3 pooled human IBD patients.¹⁹ These animals were compared with wild-type (Il10^+/+) mice on the same background (germ-free 129S6/SvEv) that were transplanted with the same mouse-adapted human IBD fecal microbiota. Three days after fecal transplantation, mice were treated by oral gavage with either 0.25 mg/kg or 0.5 mg/kg of IMYX-3 daily for 5 days/week. fLcn2, a key biomarker of IBD disease severity, was measured weekly using ELISA, and mice were harvested on day 31 (Figure 5A). Significant improvement in fLcn2 (ng/g feces) was found in both IMYX-3 treatment groups, in addition to a significant improvement in the colonic histology score in both IMYX-3 treatment groups (Figure 5B), with treated mice developing less severe disease as shown by representative images (Figure 5C). We then quantified the presence of neutrophils in the mouse colon by staining for MPO, showing a significant upregulation of neutrophils in Il10^−/− animals and a significant downregulation of neutrophils in the diseased animals treated with IMYX-3 (Figures S3A and B).

Figure 5.

IMYX-3 assessment of prevention efficacy in the IL10 knockout (KO) model of inflammatory bowel disease. (A) Fecal lipocalin (fLcn) was measured weekly to determine difference of the clinical biomarker showing significantly lower amounts of fLcn after IMYX-3 treatment. (B) Areas of the colon were scored for the effect of IMYX-3 treatment on disease burden and summarized in a total score. (C) Representative histologic images of different rectum areas are shown, including treatment response. **P < .01; ***P < .001. IL, interleukin; PBS, phosphate-buffered saline; WT, wild-type.

To further validate the therapeutic efficacy of the drug in the Il10^−/− mouse model, we repeated the experiment with a longer disease induction period, higher IMYX-3 dose, and comparison with positive control anti-IL12p40 monoclonal antibody, a Food and Drug Administration–approved IBD biologic therapy that neutralizes the common subunit of IL-12 and IL-23 cytokines. Eleven days after Il10^−/− mice received mouse-adapted human IBD fecal microbiota transplantation, 1 group of animals was injected with anti-IL12p40, while 2 additional groups were treated daily 5 days/week with IMYX-3 at 0.5 mg/kg and 5 mg/kg until day 32 (Figure 6). All treated groups showed a significant decrease in fLcn2 levels from preinduction (day 11) baseline, when compared with untreated, diseased animals that had increased fLcn2 levels (Figure 6A; Supplementary Data 2). Additionally, the drug showed a trend toward dose response in histologic response, particularly demonstrating improvement in the distal colon and rectum (Figures 6B and C). Last, costaining with MPO and CitH3 also showed a significant decrease of NETosis in the IMYX-3–treated Il10^−/− model (Figure 6D).

Figure 6.

IMYX-3 dose response compared with positive control in a IL10 knockout (KO) treatment model of inflammatory bowel disease. After 11 days of induction, mice were treated with a-IL12p40 and IMYX-3 at 0.5 mg/kg and 5 mg/kg showing (A) a significant decrease in fecal lipocalin (fLcn) and (B) improvement in total histology, (C) specifically improving the distal colon and rectum comparable with the positive control. (D) Colon neutrophil NETosis was also found to decrease in animals treated with 5 mg/kg IMYX-3 based on %CitH3⁺MPO⁺ cells. *P < .05.

Next, we examined the pharmacokinetics of the drug in rodent models. A bioanalytical method was developed to determine the quantity of IMYX-3 in blood. The drug was administered both intravenously and orally to 3 rats per group, and blood was withdrawn at 8 time points up to 48 hours. Oral T1/2 was found to average 10.4 hours, suggesting a once daily dosing regimen for clinical evaluation. We next evaluated toxicity of the drug in rodent and canine models using both maximum tolerated dose and dose range finding toxicity experimental designs. The single maximum tolerated dose was measured at 400 mg/kg in rats and dogs. In rat dose range finding experiments (14 days orally), up to 200 mg/kg dosing yielded no toxicity in biochemistry, complete blood count, immune profiling, animal weight, organ weight, or organ histopathology as assessed by an expert blinded animal pathologist. The drug was then evaluated in a canine model for dose range finding (beagle dogs, 14 days) and showed no adverse events at 20 mg/kg and emesis at 100 mg/kg, suggesting a nontoxic dose between 20 mg/kg and 100 mg/kg.

We then tested the possibility that IMYX-3 may affect anti-microbial killing by white blood cells. GFP + E coli was cultured for 3 hours with human blood (10:1 [bacteria:blood]) in the presence and absence of increasing concentrations of IMYX-3. Samples were plated and cultured aerobically overnight at 37 °C. The killing capacity was evaluated based on the number of bacterial colonies, with no decrease in killing shown when white blood cells were cultured with 5, 10, and 50 µM of drug (Figure 7).

Figure 7.

IMYX-3 does not affect white blood cell (WBC) antibacterial cytotoxicity. GFP⁺Escherichia coli was cultured with human blood (10:1) for 3 hours in the presence of escalating quantities of IMYX-3 (5, 10, and 50 µM). The culture was then left overnight on plates, and bacterial killing was tested compared with nontreated cultures. No difference was found in WBC killing in the presence of high quantities of IMYX-3.

Finally, we performed an initial evaluation of the pharmacokinetics-to-pharmacodynamics relationship of IMYX-3 in BALB/c mice to determine basic pharmacokinetic properties and biodistribution of the drug. At the previously identified efficacious dose of 2 mg/kg in the DSS model, we administered the drug for 5 days orally to mice (Figure 8A). We found the maximum concentration (C_max) to be ∼0.5 µM at 4 hours in the blood. Notably the drug showed a similar C_max in the colon, further demonstrating that IMYX-3 is efficacious both in the blood and the colon and can be used for IBD therapy (Figure 8B). The drug was not present in brain, confirming that it does not cross the blood-brain barrier.

Figure 8.

Pharmacokinetic (PK)-to-pharmacodynamic modeling and biodistribution of IMYX-3. Healthy BALB/c mice were dosed orally daily with 2 mg/kg of IMYX-3. On day 5, one group of animals was sacrificed prior to dosing followed by additional groups on a time course of 0.5, 1, 2, 4, 6, and 24 hours. (A) At each time point, samples of blood, liver, brain, and intestine were prepared for bioanalytical evaluation in liquid chromatography–mass spectrometry and (B) PK parameters were quantified with specific IMYX-3 concentrations for blood, colon, liver, and brain plotted over time.

Discussion

Effective and convenient treatments for UC remain a major unmet need in drug development, with 50% to 75% of the standard drug refractory patient population lacking effective treatment options. For example, drugs like ABX-464 (IL-22 modulation) and RVT-3101 (anti-TL1A inhibitor) yield a maximum of 26.2% and 34% clinical remission rates, respectively, in biologic therapy–experienced treatment groups.^1,24 Many patients undergo colectomy or experience poor quality of life for extended periods. Most drugs available for UC focus on direct T cell modulation affecting other immune populations downstream. While this is generally effective for many patients, patients with high neutrophil burdens have been found to become refractory to treatment and experience worse disease. The presence of highly activated neutrophils in the colon is undoubtedly a cause of significant damage to the epithelium and likely yields significant cytokine secretion, causing chemoattraction of additional activated immune populations.

To this end, we developed IMYX-3 to modulate the toxic activation of neutrophils in UC. In our studies we show that IMYX-3 binds RAB27A and effectively lowers the activation of neutrophil activity, including degranulation, ROS, and NETosis, while not increasing neutrophil apoptosis. The efficacy of the molecule was confirmed using 2 separate in vivo models of experimental colitis (DSS and Il10^−/− mice with human IBD fecal transplantation) using both prevention and treatment protocol and then shown to be safe with maximum tolerated dose and dose range finding testing. These studies confirm previous preclinical reports demonstrating that neutrophil activation can be detrimental in UC and that RAB27A modulation can improve outcomes in animal models of UC.^25,26

Our results also support the clinical practice showing that GMA is an effective treatment for UC, in which studies show between 25% and 60% remission rates with the procedure.^1,8,27 Notably, there is variability within these results, with some studies not showing significant improvement in patient populations undergoing GMA.²⁸ This report notes that the disparate results could be caused by heterogeneity among the patient population or different trial protocols and populations. Importantly, in the study showing lack of efficacy for GMA in UC,²⁸ there was significant improvement in patients who had a modified Riley score of 7 indicating severe acute inflammation, erosion, and ulceration. Clinical remission and response were seen in 24% of sham patients and 54% of GMA-treated patients. These findings are indicative of a high neutrophil burden in the colon demonstrated by crypt abscesses, suggesting that neutrophil biomarkers can be effectively used to identify patients with high levels of neutrophil activation and work toward modulating them in clinical studies with therapeutic agents. Fecal calprotectin, fecal MPO, and neutrophil-to-lymphocyte ratio have all been shown to identify patient populations with uniquely high neutrophil activity and can be used as a companion diagnostic for IMYX-3 in clinical studies. Moreover, no GMA studies found safety or tolerability issues leading to adverse events, confirming that downregulation of overall neutrophil number by ∼65% during GMA did not yield negative consequences for the patient population. Notably, no increase in infections was noted. This suggests that therapeutics that downregulate neutrophil activity, while not ablating them, may be safe for clinical practice.

Adverse events may also be possible for a RAB27A inhibitor, as RAB27A is expressed in many secretory cell types. Indeed, mouse models of RAB27A deficiency and humans lacking RAB27A activity (Griscelli syndrome type 2) have been shown to have leukopenia and partial albinism due to the failure of melanocytes to effectively distribute melanin in the cytoplasm. In GS2, it is notable that despite the broad expression of RAB27A across several cell types (eg, small intestine and pancreas) there is no deleterious phenotype in secretory cells except leukocytes and melanocytes, likely due to the redundancy of other RAB GTPases such as RAB27B and its ability to compensate for the lack of RAB27A. In our studies, we did not note any alterations to animal color or changes in white blood cells levels in either rodent or canine even at higher concentrations of the drug (up to 100 mg/kg for 2 weeks). This is likely due to the partial inhibition of IMYX-3 for the RAB27A across daily oral dosing. This parallels the activity of many therapeutics in which the complete loss of function of their target could yield extreme toxicity and embryonic lethality, yet the drugs are well tolerated. This includes VEGF inhibitors (bevacizumab), mTOR inhibitors (everolimus), JAK inhibitors (Rinvoq), proteasome inhibitors (Velcade), and others. Alternatively, it has been shown that there is variability among the phenotypes of GS2 patients, depending on the location of their RAB27A mutations with some patients not showing albinism.²⁹ This may suggest that IMYX-3 binds to RAB27A at a site that does not inhibit melanocyte function, similar to the human patients. Further studies will be required to examine and map this possibility.

While several attempts are currently underway to therapeutically modulate neutrophil activity in disease, IMYX-3 maintains a unique positioning by broadly modulating neutrophil activity while not completely inhibiting its function or altering cell number. In our pharmacokinetic/pharmacodynamic modeling, we found that the efficacious dosing of the molecule yields a 0.5 µM concentration in the blood, which was found to only partially downregulate ROS and degranulation, preserving cellular activity to combat infection. Previous attempts to drug the neutrophil yielded neutropenia (as seen with CXCR2 inhibitors such as danirixin and navarixin) or were not sufficiently efficacious (as shown by elastase inhibitors such as alvelestat and sivelestat). Intriguingly, brensocatib, a cathepsin C inhibitor, has recently shown efficacy in clinical trials on bronchiectasis.³⁰ Cathepsin C acts in the bone marrow and cleaves proenzymes of elastase, cathepsin G, and proteinase 3.³⁰ The upstream inhibition leaves these enzymes inactive when they are secreted from the mature neutrophil and lowers damage caused to the lungs during active disease and exacerbations. This suggests that broader modulation of toxic neutrophil function in disease may yield improved outcomes, further supporting our approach. Additionally, studies on brensocatib did not show any significant increase in adverse events, again showing that modulating neutrophil function is not inherently toxic. Other molecules currently in development for neutrophil manipulation include ADS051, which modulates neutrophil trafficking, CIT-013, which targets neutrophil extracellular traps by binding to citrullinated histones H2A and H4, and NTR-L301 inhibiting enzymes related to NETosis. However, these drugs do not affect ROS activity or degranulation, which are key neutrophil functions in numerous inflammatory diseases.

IMYX-3’s functional modulation is not limited to neutrophils and has effects on additional inflammatory cells. Particularly in our in vitro studies, we found that it can downregulate secretion of inflammatory cytokines from activated peripheral blood mononuclear cells. This broad effect is likely additive to the anti-inflammatory activities of the drug found in our UC modeling because both T-cells and monocytes are active in UC. Here too, we found only partial inhibition of cytokine secretion at the expected C_max of the drug in our efficacy experiments. Notably, RAB27A expression is significantly higher in the neutrophils than in monocytes or T cells, suggesting that its inhibition will be more substantial in neutrophils than other cell types. Importantly, our biodistribution studies show that the drug also effectively reaches the colon, with a similar C_max between blood and colon at 4 hours demonstrating that the drug effect is likely both on the systemic activated neutrophils and on colonic neutrophils. Functionally, this can be seen in our experiment, showing lower ROS levels in the diseased whole colon after treatment with the drug.

First-in-human clinical evaluation will be needed to better define the ability of IMYX-3 to functionally modulate neutrophil activity and yield clinical remission in UC. A first set of single ascending and multiple ascending dose studies will be conducted to determine pharmacokinetics and safety. Additionally, these studies should include evaluation of neutrophil activity markers to determine the extent of neutrophil modulation at specific drug concentrations. Following these studies, optimal dosing ranges will be selected and patients with UC will be evaluated for pharmacokinetics, safety and disease modulation by Mayo score. Because UC patients have activated neutrophil populations, it will be possible to determine the extent of the effect of IMYX-3 in lowering deleterious functions. Selection of UC patients will also include patient criteria based on activated neutrophil markers (eg, fecal calprotectin, fecal MPO, and neutrophil-to-leukocyte ratio) for a precision medicine approach focusing on candidates that will most benefit from IMYX-3 (as seen in the GMA studies). Beyond these initial clinical studies, future clinical evaluation may include both naive and refractory to treatment patient populations and may also involve coadministration with existing drugs that block orthogonal pathways to further improve efficacy.

The current study shows for the first time the ability to modulate neutrophil toxicity via RAB27A inhibition using a novel and developed pharmacological agent in UC models. Additional studies on human UC-derived immunocompetent organoids may also be beneficial in understanding human disease dynamics. Notably, neutrophil activity has been implicated as an important factor in additional inflammatory diseases, such as Crohn’s disease, chronic obstructive pulmonary disease, lupus (systemic lupus erythematosus), rheumatoid arthritis, and hidradenitis suppurativa.^31,32 Additional evaluation of the drug in these preclinical disease models will yield a deeper understanding of the possible broad disease efficacy.

Author contributions

M.N. carried out most of the experimental work, collected and analyzed the data, and drafted the initial figures and methods. D.B., N.I., Y.M., J.W.H., S.G., and R.B.S. performed additional experiments within the text. J.A., D.C.D., C.H.E., and Z.G.F. contributed to experimental design. S.S. and Z.G. conceived and designed the study, supervised all experimental planning and data interpretation, secured funding, and revised the manuscript with contributions from R.B.S. All authors discussed the results and approved the final version of the manuscript.

Supplementary material

Supplementary data is available at Inflammatory Bowel Diseases online.

Funding

R.B.S. received grant support from ImmunyX, National Institutes of Health (NIH) P30DK034987 and P40OD010995, and the Crohn’s and Colitis Foundation. S.G. received support from National Institutes of Health (NIH) grants T32DK007737 and P30DK034987. ImmunyX received support from the Crohn’s and Colitis Foundation IBD Ventures (Award 1372889) and an National Institutes of Health (NIH) Small Business Innovation Research (SBIR) Phase II Award (1R44DK145292-01).

Conflicts of interest

M.N., N.I., Y.M., D.C.D., J.A., C.H.E., Z.G.F., S.S. and Z.G. are employees of ImmunyX. R.B.S. has received grant support from Immunyx. All other authors disclose no conflicts.

Data availability

The data underlying this article are available in the article and in its online supplemental material.