Non-canonical Inflammasome-Mediated IL-1β Production by Primary Endometrial Epithelial and Stromal Fibroblast Cells Is NLRP3 and Caspase-4 Dependent

Abstract

Inflammation of the post-partum uterus is a normal physiological event, crucial for tissue involution and repair. However, in the bovine, some cows fail to resolve this inflammation, resulting in endometritis, which compromises fertility. Earlier work has identified upregulated expression of the potent inflammatory cytokine IL-1β early post-partum, in cows which subsequently develop endometritis. As a result, targeting IL-1β expression holds potential as a novel treatment for this disease, yet the regulatory mechanisms contributing to IL-1β expression in the bovine endometrium remain unknown. To investigate this, endometrial tissue samples were obtained 7 and 21 days post-partum (DPP) from cows that were diagnosed with endometritis at 21 DPP and cows that experienced a physiological level of inflammation throughout involution. IL-1β was measured by qPCR, ELISA, and immunohistochemistry. Seven DPP, endometrial IL-1β protein levels were significantly higher in animals that proceeded to develop endometritis at 21 DPP. IL-1β production could be detected in luminal and glandular epithelium, in underlying stromal fibroblasts as well as infiltrating immune cells. To investigate the mechanisms regulating IL-1β expression, primary endometrial epithelial cells, stromal fibroblasts and PBMCs were stimulated with LPS and the inflammasome activator nigericin. Stromal fibroblast cells were particularly potent producers of IL-1β. Basolateral LPS stimulation of polarized epithelial cells induced IL1B mRNA and a previously undescribed IL-1β protein isoform, with preferential protein secretion into the apical compartment. Key inflammasome components [nod-like receptor protein 3 (NLRP3), nima-related kinase-7 (NEK7), apoptosis speck like protein containing a CARD (ASC), and gasdermin-D] were expressed by endometrial cells following stimulation. Endometrial cell stimulation in the presence of NLRP3 receptor (MCC950) and pan-caspase (Z-VAD-FMK) inhibitors blocked IL-1β production, demonstrating its dependence on the NLRP3 inflammasome and on caspase activity. Furthermore, caspase-4 specific siRNA prevented IL-1β production, confirming that inflammasome activation in endometrial cells is caspase-4 but not caspase-1 dependent, as shown in other species. Identifying the tissue- and species-specificity of inflammasome assembly and activation has critical relevance for our understanding of inflammation and suggests new therapeutic targets to enhance the resolution of inflammatory pathologies including endometritis in cattle.

Keywords: IL-1β; NLRP3 inflammasome; endometritis; epithelial cells; inflammation; stromal fibroblasts.

Figure 1

IL-1β protein levels and NLRP3 receptor mRNA expression are elevated in endometritic animals. Protein and RNA were extracted from endometrial samples obtained by cytobrush at 7 and 21 DPP from healthy cows and cows with cytological endometritis alone or cytological endometritis with PVD. (A). Protein levels were quantified via BCA assay and IL-1β levels quantified by a bovine specific ELISA kit. Data is presented as mean cytokine expression (pg) per mg of total protein + SEM (n = 3–7). ^*P < 0.05 was calculated using a Mann Whitney test in GraphPad Prism 7 software. (B–E) NLRP3, IFI16, NLRC4, and NLRP1 inflammasome receptor expression were quantified via qPCR on mRNA extracted and reverse transcribed to cDNA. Data are presented as mean mRNA expression + SEM relative to the reference gene H3F3A (n = 3–7).

Figure 2

IL-1β protein is primarily localized to the epithelial cells and stromal fibroblasts in the post-partum endometrium. Endometrial biopsies were collected from healthy or endometritic cows at either (A) 7 DPP or (B) 21 DPP. Cows were classified as healthy or as having cytological endometritis based on immune cell infiltration at 21 DPP. Biopsies were paraffin embedded and subsequently sectioned and stained for IL-1β or CD45. Nuclei were counterstained with haemotoxylin. Tissue sections were visualized using the Aperio ScanScope Imager under 20x magnification. Scale bar indicates 100 μm. Lum., uterine lumen; L.E., luminal epithelium; G.E., glandular epithelium; S.F., stromal fibroblasts.

Figure 3

IL-1β production is induced in endometrial cells in response to LPS and the inflammasome activating signal nigericin. Expression of IL-1β was examined in primary endometrial epithelial, stromal fibroblasts, and peripheral blood mononuclear cells in response to stimulation with LPS and the inflammasome activator nigericin over a time course of 24 h. Expression of IL1B mRNA (A–C) and IL-1β protein production (D–F) in PBMCs, epithelial cells and stromal fibroblasts following stimulation with 2 μg/ml LPS over a time-course of 24 h and subsequent stimulation with 10 μM nigericin for 1 h. Induction of mRNA was examined by qPCR and protein production was examined by a bovine specific IL-1β ELISA in cultured supernatants (n = 5). Pro-IL-1β protein expression was examined in epithelial cells and stromal fibroblasts by western blotting (G,H). For qPCR analysis, results are presented as mean fold change (+SEM) relative to the reference gene H3F3A. For ELISA analysis results are presented as mean cytokine concentrations (+SEM) in cell supernatants. ^*p < 0.05, ^**p < 0.01 were calculated using a Mann Whitney test in GraphPad Prism 7, comparing untreated control to a timepoint stimulation.

Figure 4

IL-1β induction in polarized epithelial cells is mediated via the basolateral membrane. Expression of IL1B mRNA (A) and IL-1β protein secretion (B,C) by polarized endometrial epithelial following basolateral or apical stimulation with LPS (2 μg/ml) and nigericin (10 μM) over a time course of 24 h (n = 5). For qPCR analysis, results are presented as mean fold change (+SEM) relative to the reference gene H3F3A. All values were normalized to the untreated control within the same time point. For ELISA analysis results are presented as mean cytokine concentrations (+SEM) in cell supernatants. ^**p < 0.05 was calculated using a Mann Whitney test in GraphPad Prism 7, comparing untreated control to a timepoint stimulation.

Figure 5

Inflammasome components are expressed in endometrial cells in response to inflammatory stimuli. Expression of components of the NLRP3 inflammasome complex were examined in PBMCs and primary endometrial epithelial cells and stromal fibroblasts following stimulation with LPS and the inflammasome activator nigericin over a time course of 24 h. NLRP3 (A–D), NEK-7 (E–H), ASC (I–L), gasdermin-D (M–P), and caspase-4 (Q–T) mRNA expression was examined across PBMCs, 2D cultured epithelial cells, polarized epithelial cells and stromal fibroblasts (n = 5). Results are presented as mean fold change (+SEM) relative to the reference gene H3F3A. All values were normalized to the untreated control within the same time point. ^*p < 0.05, ^**p < 0.01 were calculated using a Mann Whitney test in GraphPad Prism 7, comparing untreated control to a timepoint stimulation.

Figure 6

IL-1β production in endometrial cells is dependent on the NLRP3 inflammasome complex. IL-1β secreted from polarized endometrial epithelial cells and stromal fibroblasts was examined in the presence of inhibitors targeting specific components of the NLRP3 inflammasome complex. Cells were either treated with increasing concentrations of (A,B) the NLRP3 receptor inhibitor MCC950 (1–100 nM) before LPS priming or (C,D) the pan-caspase inhibitor Z-VAD-FMK (1–100 ng/ml). Treatment with DMSO was included as a vehicle control. Supernatants were subsequently harvested and IL-1β measured using a bovine specific ELISA (n = 5). Results are mean cytokine concentrations (+SEM). ^*p < 0.05, ^**p < 0.01 were calculated using a Mann Whitney test in GraphPad Prism 7, comparing the LPS and nigericin stimulation to individual inhibitor treatments.

Figure 7

Endometrial cells display differential expression of inflammatory caspases. Levels of caspase-1 and caspase-4 expression was examined in PBMCs, endometrial epithelial cells and stromal fibroblasts (n = 5). (A,B) mRNA expression of the inflammatory caspases caspase-1 and caspase-4 was examined in PBMCs and primary endometrial epithelial cells and stromal fibroblasts via qPCR. Expression levels are relative to the reference gene H3F3A. Expression of active levels of caspase-1/caspase-4 in was examined by flow cytometry using the FAM-FLICA® caspase assay following stimulation with the inflammasome activators LPS (2 μg/ml) for 6 h and subsequent stimulation with nigericin for 1 h. Gating strategies for PBMCs (C), epithelial cells (D), and stromal fibroblasts (E) are shown. (F) Expression levels of caspase-1/caspase-4 in PBMCs, epithelial cells and stromal fibroblasts.

Figure 8

Inflammasome activation of IL-1β is mediated through caspase-4 in endometrial cells. IL-1β secreted from polarized endometrial epithelial cells and stromal fibroblasts was examined in the presence of a commercial inhibitor or siRNA targeting caspase-4 (n = 5). (A,B) Cells were treated with increasing concentrations of the caspase-4 inhibitor Z-LEVD-FMK (0.02–2 μM) before LPS priming. Supernatants were subsequently harvested and IL-1β measured using a bovine specific ELISA. (C,D) Caspase-4 mRNA expression levels in epithelial and stromal cells was examined by qPCR and (E,F) IL-1β measured by ELISA following treatment with siRNA directed against caspase-4, scrambled siRNA, a vehicle control or left untreated for 48 h followed by stimulation with LPS and nigericin for 6 h. mRNA expression levels are relative to the reference gene H3F3A. ELISA results are mean cytokine concentrations (+SEM). ^*p < 0.05, were calculated using a Mann Whitney test in GraphPad Prism 7, comparing the LPS and nigericin stimulation to individual inhibitor treatments.

Figure 9

Proposed mechanism of the switch to pathological inflammation in endometritis. A physiological inflammatory response in the post-partum endometrium is characterized by a (mostly) intact epithelium, inflammasome activation in the epithelial cells with the majority of the subsequent release of the pro-inflammatory cytokine IL-1β directed into the uterine lumen. A pathological inflammatory immune response is characterized by a disrupted epithelium resulting in exposure of the stromal fibroblasts. Inflammasome activation subsequently occurs in the exposed stromal fibroblasts resulting in production of IL-1β and its release into both the lumen and surrounding endometrial tissue, propagating the pathological inflammation.