Leaky gut model of the human intestinal mucosa for testing siRNA-based nanomedicine targeting JAK1

Complex in vitro models of human immune cells and intestinal mucosa may have a translation-assisting role in the assessment of anti-inflammatory compounds. Chronic inflammation of the gastrointestinal tract is a hallmark of inflammatory bowel diseases (IBD). In both IBD entities, Crohn's disease and ulcerative colitis, impaired immune cell activation and dysfunctional epithelial barrier are the common pathophysiology. Current therapeutic approaches are targeting single immune modulator molecules to stop disease progression and reduce adverse effects. Such molecular targets can be difficult to assess in experimental animal models of colitis, due to the disease complexity and species differences. Previously, a co-culture model based on human epithelial cells and monocytes arranged in a physiological microenvironment was used to mimic inflamed mucosa for toxicological and permeability studies. The leaky gut model described here, a co-culture of Caco-2, THP-1 and MUTZ-3 cells, was used to mimic IBD-related pathophysiology and for combined investigations of permeability and target engagement of two Janus kinase (JAK) inhibitors, tofacitinib (TOFA) and a JAK1-targeting siRNA nanomedicine. The co-culture just before reaching confluency of the epithelium was used to mimic the compromised intestinal barrier. Delivery efficacy and target engagement against JAK1 was quantified via downstream analysis of STAT1 protein phosphorylation after IFN-γ stimulation. Compared to a tight barrier, the leaky gut model showed 92 ± 5% confluence, a barrier function below 200 Ω*cm2, and enhanced immune response to bacteria-derived lipopolysaccharides. By confocal microscopy we observed an increased accumulation of siJAK1-nanoparticles within the sub-confluent regions leading to uptake into immune cells near the epithelium. A concentration-dependent downregulation of JAK/STAT pathway was observed for siJAK1-nanoparticles (10 ± 12% to 16 ± 12%), whereas TOFA inhibition was 86 ± 2%, compared to untreated cells. By mimicking the status of severely damaged epithelium, like in IBD, the leaky gut model holds promise as a human in vitro system to evaluate the efficacy of anti-inflammatory drugs and nanomedicines.

SI Figure 1 Surface area measurements for Caco-2 confluence calculation. Examples of confocal images with subconfluent regions within the leaky gut model in top view (xy-dimension) and overlaid fluorescence channels. Co-culture was counterstained for cell nuclei (DAPI; blue) and cytoskeleton (AF488 phalloidin; green). Caco-2 confluency of the leaky gut model was calculated after 6/7 days of cultivation by ImageJ (version 1.52, NIH, Bethesda, USA). After ´free hand´ selection of Caco-2 cell-free areas (yellow line), respective regions were measured using ´Analyze > Measure´ and the resulting area divided by the total area of the image (represented as percentage of confluency).

SI Figure 2 Representative graphs obtained from automated CASY Cell counter with cell type-specific size distributions.
Representative graphs of cell size distribution for Caco-2, MUTZ-3, THP-1, and dTHP-1 cells (A). Cell number and percentage viability was determined by different size ranges pre-set for each individual cell line in order to differentiate between cell debris (< 8 µm; I), dead or dying cells (IIa), and viable cell population (IIIa). Representative graphs of resulting single-cell suspension after dissociation of tight barrier and leaky gut model (B). Appropriate size ranges were determined to separate for dendriticlike cells (MUTZ-3) with a range of 8-13 µm (IIb) and the other two cell types, epithelial (Caco-2) and macrophage-like cells (dTHP-1), with similar size ranges between 13-50 µm (IIIb).

SI Figure 4 Confocal microscopic images of tight barrier model after nanoplex incubation using fluorescently-labeled nanocarrier (DiI-labeled LNPs) and siRNA (AF647-labeled siJAK1).
Tight barrier model in absence of nanoplexes (control) was used to set up the optimal fluorescence detection of respective fluorophores of the nanocarrier (DiI-LNPs) and cargo (AF647-siJAK1) without generating unspecific background signal (A). Nanoplexes with N/P ratio of 8 (45 µg LNPs and 135 nM siJAK1; N/P 8 Low) (B), and control treatments were incubated twice (apically) for 6 h including blank LNPs (90 µg per well; LNPs High) (C), and naked siJAK1 (270 nM; siJAK1 High) (D). Afterwards, cells were fixed and stained for cytoskeleton (AF488-phalloidin; green) and cell nuclei (DAPI; blue). Recorded z-stacks of around 200 µm depth (axis unit = µm) are represented in top (xy-) or side (xz/yz-dimension) view with overlaid fluorescence signal (merge) or split channels for DiI-LNPs (red) and AF647-siJAK1 (cyan). Basal site of epithelial cell layer (attached to the collagen matrix) is marked by a dashed line to identify the localization of nanocarrier and/or its cargo.

All events
All in siJAK1 High

DiI-labeled LNPs (PE-A)
SI Figure 7 Flow cytometric analysis of STAT1 phosphorylation (pY701) using anti-phospho-STAT1 antibodies and assay protocol adaptation for JAK/STAT pathway activation and downregulation. STAT1 activation was tested in different mono-cultures (Caco-2, dTHP-1 and MUTZ-3) stimulated with IFN-g for 1 h ranging from 3 to 25 ng/well (A). Cells were pre-treated twice with TOFA (0.5 to 5 µM) for 6 h on two subsequent days and afterwards stimulated with 25 ng/well IFN-g (50 ng/mL) for 1 h (B). Phosphorylation profile of pSTAT1 of leaky gut model pre-treated twice on day 6/7 with 5 µM TOFA for 6 h applied to both compartments (C) and of tight barrier model pre-treated on day 11/12 with 5 µM TOFA for 6 h applied either to the apical (AP) or basolateral (BL) compartment (D). After pre-treatment, both co-cultures were stimulated with 50 ng/mL IFN-g for 1 h and subsequently dissociated for phospho-specific STAT1 staining and FACS analysis. Phosphorylation of STAT1 is represented as histogram overlay with AF647 emission collected at APC instrument setting (APC-A) and summarized in both percentage of pSTAT1-positive cells (%) and mean fluorescence intensity (MFI). pSTAT1-positive cells are compared to unstimulated (control; grey), untreated cells (red), and TOFA pre-treatments applied to the apical (AP; yellow), basolateral (BL; violet), or both compartments (orange). Data of are means ± SD, for mono-cultures (n=3; N=1), tight model (n=9; N=3), and leaky model (n=15; N=5).  Figure 10 Confocal microscopic images of co-culture in absence of Caco-2 epithelial cell layer. Co-culture model without Caco-2 cells on top was used to verify the penetration of nanocarrier through the collagen matrix. Fluorescently-labeled nanocarrier (DiI-labeled LNPs; 90 µg/well) complexed with siRNA (AF647-siJAK1; 450 nM) were applied apically and incubated for 6 hours (A). Uptake of nanocarriers by immune cells within the collagen matrix (top view). Images of 3D cells are projections of 20-30 µm z-sections (scale bar = 50µm). (B). In comparison, diffusion of carboxylated polystyrene nanoparticles (FluoSpheres ® ; FS) of 100 nm size (50 µg/well) after 6 h of incubation (C). Cells are stained for cytoskeleton (AF488-phalloidin; green) and cell nuclei (DAPI; blue). Recorded z-stacks of around 350 µm depth (axis unit = µm) are represented in top (xy-) or side (xz/yz-dimension) view with overlaid (merged) fluorescence signal or in split channels for DiI-LNPs and FS (red) and AF647-siJAK1 (cyan). Figure 11 Confocal microscopic images of tight barrier model comprising a confluent Caco-2 cell layer on top and uptake study of LNPs. Nanocarriers were applied apically (90 µg/well) and incubated for either 4 or 24 h in a healthy (A) and LPS-inflamed co-culture model (B). For all conditions, co-localization of nanocarriers within epithelial cells and barrier function decrease upon LPS-inflammation do not increase particle penetration and access to the underlying immune cells, even for longer incubation time. Recorded z-stacks of around 250 µm depth (axis unit = µm) are represented in top (xy-) or side (xz/yz-dimension) view in split fluorescence channels for cytoskeleton (AF488-phalloidin; green), cell nuclei (DAPI; blue), and nanocarrier (DiI-labeled LNPs; red). Basal site of epithelial cell layer (attached to the collagen matrix) is marked by a dashed line to identify the localization of nanocarrier.