Bacteria were lyophilized in the presence of a cryoprotectant to stabilize ADS024. Lyophilized ADS024 material (lot #210421) was used in all in vitro experiments. Lyophilized ADS024 material (lot #230920) was used in the animal experiment. A cryoprotectant-containing control filtrate without bacteria (lot #180122) was used as the negative control. The bacterial stock was dissolved to 6×10⁹ colony forming units (cfu)/mL in RPMI1640 (pH 8) and incubated at 37°C for 30 min before use (Xie et al., 2022a). The bacterial suspensions were filtered with 0.22 μm filters to produce sterile filtrates (FS), which were further serially diluted to 1X, 0.1X, 0.01X, and 0.0001X in PBS (pH 7.4). 1 μL of the sterile filtrates were added to cells and tissue cultures (1 mL) to final titers equivalents of 6 × 10⁶, 6 × 10⁵, 6 × 10⁴, and 6 × 10³ cfu/mL. The bacterial filtrates were used within 1 h of resuspension to ensure freshness. Sterile filtration of ADS024 keeps the metabolites, proteins, or extracellular vesicles secreted by the live bacteria during the 30-min rehydration period while avoiding any solvent that might denature secreted membrane vesicles or proteins. The sterile filtration also prevented bacterial contamination of the cell culture incubator.

For molecular weight cut-off (MWCO) testing, the ADS024 filtrate was filtered by Vivaspin 2 MWCO columns (100 kDa: 45-001-570; 50 kDa: 45-001-569; 30 kDa: 45-001-568; 10 kDa: 45-001-567; 5 kDa:45-001-566; 3 kDa:45-001-565) from Cytiva (Xie et al., 2022a).

Serum samples from UC patients were prospectively collected from 2012 to 2015 (Tran et al., 2017). UCLA Institutional Review Board (IRB 12-001499) approved this study. UCLA Pathology obtained written informed consent from all subjects, so UCLA IRB waived a separate informed consent.

Sera were pooled from 6 UC patients to prepare serum exosomes (UCSE) by total exosome isolation reagent (#4478360, ThermoFisher; Wang et al., 2020). The protein concentrations in the UCSE were determined by bicinchoninic acid (BCA) protein assay (#23225, ThermoFisher).

Fresh colonic tissues from UC patients with moderately inflamed histology from UCLA Surgical Pathology were obtained between 2021 and 2022. UCLA IRB (12-001499) approved the study. The fresh human colonic explants in 3 × 3 mm per piece were incubated in serum-free RPMI1640 media with or without 100 μg/mL UCSE for 30 min. Control and ADS024 filtrates were subsequently added and incubated for 24 h.

The baseline characteristics of intestinal tissues and serum samples are shown in Table 1.

Human primary colonic epithelial cells (HPEC; H6047, Cell Biologics) were cultured in complete human epithelial medium (H6621, Cell Biologics) containing 10% fetal bovine serum and 1% penicillin–streptomycin to 80% confluence and then switched to serum-free media overnight (Xie et al., 2022a,b). Cells were subsequently pretreated with 1 μL/mL control and ADS024 sterile filtrates for 30 min, followed by 100 μg/mL UCSE and further incubated at 37°C for 24 h.

To quantify apoptosis, serum-starved HPECs in white-wall clear-bottom 96-well plates (165306, ThermoFisher) were treated with 100 μg/mL UCSE, followed by control sterile filtrate and ADS024 sterile filtrates at various dilutions and MWCOs. After adding UCSE and sterile filtrates, the cells were added with Promega RealTime-Glo Annexin V apoptosis assay reagents (JA1011, Promega) in a 1:1,000 ratio. After 24 h, the luminescence (apoptosis) signal was measured with a BioTek Synergy H1 plate reader (Xie et al., 2022a).

To identify apoptosis-related proteins, the serum-starved HPECs in 6-well plates were pretreated with sterile filtrates for 30 min, followed by 100 μg/mL UCSE for 24 h. The cell lysates (300 μg protein/group) were collected for Proteome Profiler Human Apoptosis Array Kit (ARY009, R&D Systems) and Human Apoptosis Signaling Array C1 (AAH-APOSIG-1-4, RayBiotech). All needed reagents were included in the kits. The signals were captured and analyzed using a Bio-Rad ChemiDoc Imaging system and Bio-Rad Image Lab software (Xie et al., 2022a).

A fresh blood sample from a UC patient were obtained through UCLA Pathology and isolated PBMCs with SepMate-15 (#85415, Stemcell Technologies). A vial of UC-PBMCs (#70051) was purchased from Stemcell Technologies. The baseline characteristics of UC-PBMC are shown in Table 1. The UC-PBMC was cultured in RPMI1640 medium with 10% fetal bovine serum and 1% penicillin–streptomycin. Secreted human interleukin 8 (IL-8) was detected by DuoSet ELISA kit (DY208) from R&D Systems.

The human colonic epithelial T84 cell monolayer was cultured on Transwell inserts (#3470, Corning) at 2 × 10⁵ cells/insert density and transepithelial electrical resistance (TEER) was measured with a dual-electrode connected to an epithelial volt/ohm meter (World Precision Instruments). When the measured TEER reached 2,000 Ω, 10 ng/mL tumor necrosis factor-alpha (TNFα) and 10 ng/mL interferon-gamma (IFNγ) were added to the basolateral side and ADS024 sterile filtrate to both sides. TEER was measured at 0 and 4–6 h. The average value was recorded as the measured TEER using the calculation: TEER (Ω·cm²) = (measured TEER–blank control TEER) × growth area (Xie et al., 2023).

For tight junction protein measurement, the serum-starved T84 cells were incubated with 10 ng/mL TNFα, 10 ng/mL IFNγ, and sterile filtrates for 24 h. Then, the cells were lyzed with radioimmunoprecipitation assay (RIPA) buffer (#89900, ThermoFisher) containing 1X protease inhibitor cocktail (#78429, ThermoFisher). Tight junction protein 1/TJP1 (MBS2605490), occludin (MBS2704294), and claudin 1 (MBS2704401) in the lysates were detected by ELISA kits from MyBioSource. Protein concentrations in cell lysates were determined by a bicinchoninic acid (BCA) assay (#23225, ThermoFisher).

Six to eight-week-old male C57BL/6 mice (Charles River Laboratories) with an average starting body weight (± standard deviation) of 21.0 ± 1.76 g were used. The animal rooms had HEPA air filtration at a temperature of 70 ± 5°F, 50% ± 20% relative humidity, 12–15 air changes per hour, and 12/12-h light/dark cycle. The facility staff swept and mopped the floor daily with a commercial detergent. They disinfected walls and cage racks with diluted bleach solution monthly. All workers always disinfected all surfaces and materials introduced into the hood with a commercial disinfectant.

Mice were acclimatized for 3 days and then randomized into four groups at the start of the study: one group of six mice, one group of 15 mice, and two groups of 12 mice each. Each mouse was identified by an ear punch corresponding to an individual number. The mice were housed in cages with sterile Bed-o’Cobs® or equivalent bedding, Labdiet 5,053 sterile rodent diet, and reverse osmosis-purified water ad libitum. The facility washed cages, tops, and water bottles with a commercial detergent, followed by air drying. Each cage card shows information to identify the study, dose, animal number, and treatment group.

Colitis in mice was induced by 3% dextran sulfate (DSS; MP Biomedicals, Cat #0260110) in drinking water from day 0 to day 5. Control naïve mice did not receive DSS. Mice were dosed by oral gavage of 5 × 10⁸ CFU ADS024 in phosphate-buffered saline (PBS) twice daily starting on Day 6 for 14 days. Negative control mice were dosed with PBS (vehicle). All interventions were performed during the light cycle.

Mice were monitored daily for weight loss, diarrhea, and blood in the stool. On day 19, all mice were sacrificed for histology assessment of the colon. Colonic cytokines and myeloperoxidase (MPO) levels were measured by ELISA. Distal colon samples from 44 mice (88 pieces), fixed in 10% neutral buffered formalin, were shipped to Inotiv Boulder (formerly HistoTox Labs) for hematoxylin and eosin (H&E) staining and evaluation by a board-certified veterinary pathologist using light microscopy. Each sample was trimmed into three transverse sections and embedded together in a single block (six pieces of tissue each with proximal toward the label and distal away from the label). Three slides were sectioned from each block at 5 μm and stained with hematoxylin and eosin (H&E). Lesions were scored according to severity 0–5 (0 = not present/normal, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked, 5 = severe). Scored features were added together for each sample to obtain a sum colitis score (Range: 0–25). For cytokine and MPO measurements, a piece of the colon from each animal was rinsed and snap-frozen. The colon tissue was homogenized in PBS plus protease inhibitor. The resulting homogenate was centrifuged to obtain a clear supernatant containing the total tissue protein. This colon tissue supernatant was used to determine the concentrations of interleukin 1beta (IL-1β), interleukin 6 (IL-6), interleukin 10 (IL-10), and tumor necrosis factor-alpha (TNF-ɑ) via multiplex (Luminex). Colon MPO concentrations were also determined in the same groups by ELISA. The mouse experiment was conducted at Biomodels, Waltham, MA, USA.

All experiments were repeated to ensure reproducibility. Two data groups were compared with unpaired Student’s t-tests and multiple data groups with ordinary one-way ANOVAs using GraphPad Prism. Results were expressed as mean +/− standard deviation (SD) or standard error of means (SEM). Significant p values in each figure are indicated.

Additional unpublished data from the study may be shared. Please contact HK or Adiso Therapeutics (https://adisotx.com/).

Our previous report demonstrated that ADS024 sterile filtrate prevented mucosal injury in C. difficile toxin B-exposed fresh human colonic explants (Xie et al., 2022a). Here, this study explored the direct protective effects of ADS024 sterile filtrate in fresh colonic explants from UC patients. The colonic mucosa of UC patients and the ileal mucosa of CD patients are characteristic of disrupted mucosal structures with numerous infiltrating immune cells (Figures 1A,B).

Compared to the control group, ADS024 filtrate treatment significantly reduced mRNA expression of pro-apoptotic Bcl-2 Associated X-protein (BAX) and pro-inflammatory tumor necrosis factor (TNF) in the fresh colonic explants from UC patients (Figure 1C), suggesting potential protective effects in UC. The 24-h treatment with ADS024 filtrate could not reverse the preexisting histological mucosal injury in the UC colonic explants (data not shown). ADS024 filtrate treatment did not impact the mRNA expression of BAX and TNF in the fresh CD patient-derived ileal explants (Figure 1D), suggesting a UC-specific protective effect.

Exposure to UCSE induced apoptosis in HPECs (Figure 2A), simulating a UC environment in vitro. Compared to the control, the ADS024 filtrate at 1X prevented UCSE-mediated apoptosis in HPEC. Further dilutions of ADS024 lost their anti-apoptotic effect (Figure 2A).

The molecular weight of the potential anti-apoptotic agent(s) in the ADS024 filtrate was estimated by filtering the ADS024 filtrates with MWCO columns. ADS024 filtrates at MWCO at 100 kDa and below did not exert anti-apoptotic effects in the UCSE-treated HPECs (Figure 2B). Therefore, the anti-apoptotic agent(s) in ADS024 is likely to be at least 100 kDa.

A protein array was utilized to detect 35 anti-/pro-apoptotic proteins in UCSE-treated HPECs quantitatively. UCSE increased activated pro-apoptotic cleaved caspase 3 protein expression level (Figure 3A). Another protein array was utilized to detect 19 apoptosis-signaling proteins in UCSE-treated HPECs. Consistent with our previous C. difficile toxin B study (Xie et al., 2022a), ADS024 sterile filtrate at 1X reduced the protein level of cleaved caspase 3 in UCSE-treated HPECs (Figure 3B).

The anti-apoptotic effect of ADS024 sterile filtrate against UCSE was attenuated by a caspase 3 activator (PAC1; Figure 3C). This finding suggested that ADS024 filtrate exerts an anti-apoptotic effect against UCSE by inhibiting caspase 3 cleavage in human colonic epithelial cells.

Disruption of epithelial barrier function leads to gut bacteria and proinflammatory substances penetrating the mucosal layer and triggering inflammation in UC. Consistent with another study (Fischer et al., 2013), TNFα and IFNγ caused epithelial barrier dysfunction with reduced transepithelial electrical resistance (TEER; Figure 4A). A lowered TEER indicated more electrical current flowing through the leaky epithelial layer.

Adding ADS024 filtrate did not affect baseline TEER (Figure 4A). However, ADS024 filtrate at 1X significantly diminished the reduction of TEER in TNFα- and IFNγ-treated T84 cells, indicating a partial restoration of epithelial barrier function (Figure 4A). The molecular weight of the potential cytoprotective agent(s) in the ADS024 filtrate was estimated by MWCO column filtration. Only filtrates of 100 kDa or above exhibited the cytoprotective effects (Figure 4A).

Three common tight junction-related proteins were quantified by ELISA. TNFα and IFNγ treatment significantly reduced TJP1 expression in T84 cells, and expression was partially restored by exposure to ADS024 sterile filtrate at 1X (Figure 4B). Neither TNFα and IFNγ nor ADS024 sterile filtrate affected claudin 1 and occludin protein expression in T84 cells (Figures 4C,D).

IL-8 is a chemokine to induce chemotaxis in immune cells (Bickel, 1993). UC patient-derived PBMCs (UC-PBMC) and HPECs secrete high levels of proinflammatory IL-8. Lipopolysaccharide (LPS), a bacterial cell wall component, stimulated IL-8 secretion in UC-PBMCs and HPEC (Figures 4E,F). ADS024 filtrate dramatically increased IL-8 secretion when alone and in the presence of LPS treatment (Figures 4E,F). Thus, ADS024 filtrate does not possess direct anti-IL-8 effects in immune and epithelial cells.

The protective effects of ADS024 were further evaluated by inducing UC-like colitis in mice with DSS, followed by 14 days of oral live ADS024 treatment and monitoring until day 20 (Figure 5A). DSS colitis caused the most severe weight loss in mice, significantly minimized by live ADS024 treatment (Figure 5B). The colitis was comprehensively assessed with a disease activity index (DAI) that evaluates weight loss, diarrhea, and bloody stool (Figures 5C–F). DSS colitis caused weight loss, diarrhea, and bloody stool, resulting in a significant increase in DAI, which was reduced by oral live ADS024 treatment (Figures 5C–F). The cryoprotectant alone did not affect the DSS-mediated weight loss, diarrhea, bloody stool, and DAI (Figures 5C–F).

Compared to water-treated control, DSS treatment caused severe colitis exhibiting the expected histologic lesions of subacute inflammation, mucosal necrosis/gland loss, erosions, submucosal edema, and epithelial hyperplasia with infiltration of neutrophils, lymphocytes, plasma cells, and macrophages into the mucosa or submucosa. These observations are characteristic of the loss of colonic mucosal structure and immune cell infiltration, indicating the loss of barrier integrity (Figure 6A). Oral live ADS024 treatment partially improved mucosal structure and reduced infiltrating immune cells (Figure 6A).

DSS treatment increased colonic levels of several representative pro-inflammatory markers, including IL-6, IL-1β, TNFα, and myeloperoxidase (MPO) in mice (Figure 6B). Of note, oral live ADS024 treatment significantly reduced colonic IL-6 and TNFα, but not IL-1β, and MPO, levels in the DSS-treated mice (Figure 6B). Therefore, oral live ADS024 treatment reduced DSS-mediated weight loss, clinical disease activity, colonic mucosal injury, and select colonic pro-inflammatory cytokine expression in mice.