Hypoxia ameliorates intestinal inflammation through NLRP3/mTOR downregulation and autophagy activation

1Crohn’s disease (CD) and ulcerative colitis (UC), the main clinical phenotypes of inflammatory bowel diseases (IBD), are complex diseases arising from the interplay between genetic and environmental factors that influence mucosal homeostasis and trigger an inappropriate immune response,. In recent years, environmental hypoxia has been increasingly recognized as an important environmental factor influencing the development of IBD,. Because of the integral role of the epithelium in maintaining intestinal homeostasis, intestinal epithelial cells (IECs) have evolved many molecular mechanisms to cope with fluctuations in blood flow, oxygen availability and metabolism, resulting in tissue hypoxia, particularly during the severe shifts associated with active inflammation in IBD,. Adaptive transcriptional responses to oxygen deprivation are mediated mainly through the hypoxia inducible factor (HIF) complexes, which are tightly regulated through the action of three prolylhydroxylases and one asparagine hydroxylase. In steady state conditions, hydroxylases suppress the activity of HIF through the hydroxylation of Pro402 and Pro564 of HIF-1α, leading to ubiquitination and proteasomal degradation. A further hydroxylation of Asn803 prevents HIF transcriptional activity. Hypoxia, blocks these processes, allowing HIF-1α to accumulate and induce the expression of HIF target genes. Both HIF-1α and HIF-2α are induced in the inflamed mucosa from IBD patients and mouse models of colitis,, and promote epithelial barrier function and the resolution of inflammation in various mouse models of disease,,. Hypoxia activates further innate immune protective mechanisms such as autophagy. Thus, HIF-1α induces the expression of several autophagosome markers, including LC3, autophagy related gene (ATG)5 and beclin-1, as well as the conversion of LC3-I to LC3-II.

2Functional polymorphisms in gene loci, including autophagy proteins such as ATG16L1, IRGM, and nucleotide-binding oligomerization domain-containing protein (NOD)2 are associated with an increased risk of CD, suggesting defective autophagy in the pathogenesis of IBD,. Autophagy ensures energetic homeostasis in cells by providing nutrients during metabolic stress and also serves a protective role by removing damaged components from the cytosol. Autophagic adaptor proteins, such as p62 (also known as SQSTM1/sequestosome 1), deliver target substrates for degradation to the autophagosome via binding to the microtubule-associated protein 1 light chain 3 (LC3). LC3 is linked to phosphatidylethanolamine, yielding a product known as LC3-II, which remains associated with autophagosomes until its destruction. Thus, the amount of membrane-bound LC3-II provides a good index of autophagy induction. Environmental stress induces autophagy through the inhibition of mammalian target of rapamycin (mTOR), a serine/threonine protein kinase that belongs to the phosphatidylinositol 3-kinase-related kinase family,. Hypoxia inhibits mTOR signaling through HIF-1-dependent and -independent mechanisms, as well as through proteins that prevent the binding of mTOR with the mTOR upstream activator Ras homolog enriched in brain (RHEB). Despite recent efforts, the molecular mechanisms underlying the induction of autophagy in response to hypoxia remain incompletely understood.

3Hypoxia has also been associated with the activation of NOD-like receptor, pyrin domain-containing (NLRP)3. NLRP3 is best known as a component of the inflammasome, a multiprotein signaling complex responsible for the activation of caspase-1 and subsequent maturation of pro-interleukin (IL)-1β and pro-IL-18 that can be activated by a wide range of stimuli, like adenosine triphosphate (ATP), reactive oxygen species (ROS), and inorganic materials,. NLRP3 and the inflammasome pathways play a decisive role in maintaining intestinal homeostasis, and single-nucleotide polymorphisms in the NLRP3 gene have recently been associated with the development of CD. Accumulating evidence indicates that NLRP3 and autophagy pathways are linked by reciprocal regulation, and recent studies have shown autophagy-dependent degradation of NLRP3,.

4Our results identify NLRP3 as a binding partner of mTOR, and show that environmental hypoxia ameliorates intestinal inflammation through the reduction of NLRP3 and mTOR binding, and the subsequent activation of autophagy.

5To study the effects of hypoxia on colonic inflammatory pathways, healthy subjects and patients with IBD in clinical remission were subjected to hypoxic conditions, resembling an altitude of 4000 m above sea level (m.a.s.l.) for 3 h. Transcriptional analysis of distal colon biopsies from patients with CD, but not UC, evidenced increased levels of NLRP3 and tumor necrosis factor (TNF)α, but not IL-1β or IL-6 mRNA prior to hypoxia (Fig. 1a–d). Hypoxia significantly reduced the mRNA expression of TNFα (Fig. 1a), IL-6 (Fig. 1c) and NLRP3 (Fig. 1d) immediately after hypoxia, and this reduction was still evident 1 week later. Interestingly, hypoxia downregulated inflammatory gene expression concomitantly with the induction of autophagy-associated p62 (Fig. 1e), but not LC3 (Fig. 1f). Although histology samples did not show any differences in the inflammatory status of patients with IBD prior to or after hypoxia, immunostaining for p62 of distal colon sections showed a trend towards reduced protein levels of p62 1 week after hypoxia in HV and patients with CD, indicating a hypoxia-mediated increase in autophagy flux (Fig. 1g, arrowheads).

6To further study the influence of hypoxia in inflammation, and the role of NLRP3 in hypoxia-mediated effects, WT and Nlrp3 −/− mice were administered with 2% dextran sulfate sodium (DSS) for 7 days in drinking water and subjected to normoxia or hypoxia for 18 h. DSS-treated WT and Nlrp3 −/− mice presented increased mRNA expression of proinflammatory mediators, including Tnfα, Il-6, Il-1β, and Nlrp3 (Fig. 2a–d). Hypoxia significantly reduced Tnfα, Il-6, Il-1β, and Nlrp3 mRNA levels in WT mice, but not Nlrp3 −/− mice (Fig. 2a–d), supporting a role for NLRP3 in hypoxia-mediated inhibitory effects. Hypoxia also induced the expression of the autophagy proteins p62 and Lc3 in both WT and Nlrp3 −/− mice (Fig. 2f and g). Hypoxic conditions were confirmed through the induction of the hypoxia marker vascular endothelial growth factor (Vegf) (Fig. 2e). Under normoxia, DSS-treated mice presented increased levels of phosphorylated NF-κB subunit p65/RelA concomitantly with p62 accumulation (Fig. 2h and Supplementary Fig. 1A). In WT mice, hypoxia reversed DSS-induced p65/RelA phosphorylation and NLRP3 expression, and restored autophagy as evidenced by the reduction of mTOR phosphorylation and degradation of p62, as well as the increase of the ratio LC3-II/LC3-I and LC3 degradation (Fig. 2h and Supplementary Fig. 1B–D). Interestingly, Nlrp3 −/− mice showed reduced levels of phosphorylated mTOR when compared to WT mice as well as impaired responses to hypoxia evidenced by the lack of marked changes in the ratio LC3-II/LC3-I and LC3 degradation (Fig. 2h and Supplementary Fig. 1C, D), suggesting a role for NLRP3 in the control of autophagy.

7Next, we sought to confirm the anti-inflammatory effects of hypoxia in the Il-10 −/− mouse model of colitis. WT, Il-10 −/−, Nlrp3 −/−, and Il-10 −/− Nlrp3 −/− double knockout mice were subjected to normoxia or hypoxia for 18 h prior to colon biopsy collection. Under normoxic conditions, Il-10 −/− mice presented increased mRNA levels of Tnfα, Il-6, Il-1β and Nlrp3, which were significantly reduced under hypoxia (Figs. 3a–d), together with the induction of p62 expression (Fig. 3f). Although not significant, our results also show a clear trend towards an increase in the mRNA levels of Lc3 under hypoxia (Fig. 3g). Under normoxia, Il-10 −/− mice also presented an increased accumulation of phosphorylated p65/RelA, phosphorylated mTOR, p62 and LC3 compared to Il-10 −/− NLRP3 −/− mice (Fig. 3h and Supplementary Fig. 2A, C–E), suggesting a blockage of autophagy promoted by NLRP3 concomitantly with the induction of the NF-κB signaling pathway. Conversely, and in line with our results in the DSS mouse model, hypoxia reduced NLRP3 expression and p65/RelA phosphorylation, and restored autophagy as evidenced by the reduction of phosphorylated mTOR, the increase of the ratio LC3-II/LC3-I and subsequent degradation of LC3 and p62 (Fig. 3h and Supplementary Fig. 2A–E).

8Since our results in patients with CD and the DSS and Il-10 −/− mouse models of colitis suggest a protective effect of hypoxia in the colonic mucosa linked to the reduction of NLRP3 and the activation of autophagy, we sought to elucidate the molecular mechanisms governing the regulation of these pathways in IECs. For this purpose, we subjected the human IEC line HT-29 to hypoxia in the presence and absence of lipopolysaccharide (LPS). HIF-1α protein accumulation peaked after 6 h of hypoxia (Fig. 4a and Supplementary Fig. 3A). In line with our results in mice, constitutive phosphorylation of NF-κB subunit p65/RelA and the NF-κB upstream inhibitor IκBα was reversed under hypoxia in a time-dependent manner (Fig. 4a and Supplementary Fig. 3B, C). Hypoxia also activated autophagy, as evidenced by the reduction of the phosphorylation/activation of mTOR, the conversion of LC3-I into LC3-II, and the subsequent degradation of p62 and LC3 protein (Fig. 4a and Supplementary Fig. 3E–G). Interestingly, hypoxia also reduced constitutive and LPS-induced protein expression of NLRP3 (Fig. 4a and Supplementary Fig. 3D). In accordance with our in vivo results, hypoxia triggered p62 mRNA expression (Fig. 4b). Likewise, hypoxia induced a transient expression of NLRP3 mRNA after 6 and 18 h, which returned to basal levels after 24 h of hypoxia (Fig. 4c), indicating that NLRP3 protein expression was not due to a reduction in the levels of NLRP3 mRNA. Protein levels of NLRP3 were also reduced in the presence of the proteasome inhibitor MG132, revealing that hypoxia induces NLRP3 clearance through a proteasome-independent mechanism (Fig. 4e and Supplementary Fig. 3H). Interestingly, hypoxia reversed the accumulation of ubiquitinated proteins in the presence of the proteasome inhibitor, pointing to the activation of alternative protein degradation mechanisms under hypoxia such as autophagy (Fig. 4e).

9Following our results on hypoxia-mediated regulation of p62 and NLRP3 expression, we sought to elucidate the molecular mechanisms governing the transcriptional regulation of both genes. Sequence analysis of the p62 promoter identified putative binding sites for HIF-1α and p65/RelA at −234 and +42, respectively (Fig. 4f). One putative binding site for p65/RelA was also found at −468 and one binding site for HIF-1α at −150 in the NLRP3 promoter (Fig. 4g). To investigate the binding activity of both transcription factors to the p62 and NLRP3 promoters in HT-29 cells subjected to hypoxia (0.2 % O2, 6 and 24 h), chromatin immunoprecipitation (ChIP) analysis was performed. In line with the stabilization of HIF-1α after 6 h of hypoxia, ChIP analysis shows increased biding of HIF-1α to the promoter of p62 after 6 h of hypoxia compared to normoxia, suggesting a role for HIF-1α in the induction of p62 expression observed under low oxygen level conditions (Fig. 4f). In accordance with the reduction in p65/RelA phosphorylation under hypoxia, binding of NF-κB to the p62 promoter was not induced at low oxygen levels (Fig. 4g). In contrast, and in line with the constitutive expression of NLRP3 in unstimulated cells, we observed constitutive binding of NF-κB to the NLRP3 promoter. Of note, HIF-1α was not found in the promoter of NLRP3 despite the accumulation of HIF-1α observed during hypoxia (Fig. 4g).

10Biosynthesis and accumulation of lysosomes in HT-29 cells under hypoxia was monitored using LysoTracker Yellow-HCK-123, a fluorescence probe that accumulates in acidic organelles such as lysosomes. Hypoxia induced the accumulation of lysosomes, further confirming the induction of autophagy at low oxygen levels. Active mitochondria monitored using MitoTracker Deep Red FM were not affected after 24 h of hypoxia (Fig. 5a). In order to clarify the causality between hypoxia-mediated autophagy and the reduction of inflammatory processes, HT-29 cells were subjected to hypoxia in the presence of chloroquine, a late-stage autophagy inhibitor that prevents the fusion of autophagosome with lysosome. As expected, autophagy was induced 24 h after hypoxia as demonstrated by a significant increase in staining for LC3 (green) in the presence of chloroquine (Fig. 5b). Hypoxia-induced p62 degradation was blocked in the presence of chloroquine, as well as the reduction of phosphorylated p65/RelA (Fig. 5c and Supplementary Fig. 4A, B). These results point to the fact that hypoxia-induced autophagy is crucial for the downregulation of proinflammatory signaling. Conversely, hypoxia still reduced the protein expression of NLRP3 in the presence of chloroquine (Fig. 5c and Supplementary Fig. 4C), indicating that NLRP3 degradation is not autophagy-mediated.

11Following our results showing that blockage of autophagy in Il-10 −/−mice was NLRP3-dependent, and that hypoxia reduces NLRP3 expression concomitantly with the activation of autophagy in the DSS and Il-10 −/−mouse models of colitis, as well as cultured IECs, we sought to determine the role of NLRP3 in the regulation of autophagy. HT-29 cells transfected with NLRP3-specific siRNAs displayed a reduction in mTOR phosphorylation and an increased basal autophagy as evidenced by the degradation of the autophagy proteins p62 and LC3 (Fig. 6a and Supplementary Fig. 5A–D), suggesting that NLRP3 plays a crucial role in the constitutive blockage of autophagy through the promotion of mTOR phosphorylation. Moreover, co-immunoprecipitation (Co-IP) experiments identified NLRP3 as a direct binding partner of mTOR and showed that this direct interaction is reversed under hypoxia in IECs (Fig. 6b and Supplementary Fig. 5E, F).

12To study the role of HIF-1α stabilization in hypoxia-mediated anti-inflammatory effects, WT mice were administered with 2% DSS for 7 days in drinking water and treated concomitantly with the hydroxylase inhibitor DMOG. DMOG activates the HIF pathway by mimicking hypoxia through the inhibition of hydroxylase activity, leading to stabilization and transactivation of HIF-1α. DSS-treated mice receiving DMOG presented a less severe colitis compared to mice receiving vehicle as evidenced by a significant reduction in body weight (Fig. 7a) and colonoscopy results (Fig. 7c and d). DMOG-treated mice also showed a clear trend towards less shortening of the colon (Fig. 7b). Hematoxyline and eosin (H&E)-stained sections of distal colon tissue taken from animals treated with DMOG presented a significant lower histological score compared to phosphate-buffered saline (PBS)-treated mice, and showed intact crypts in large areas without extensive infiltration or thickening of the mucosa (Fig. 7d and e). DMOG reduced DSS-induced Tnfα, Il-6, Il-1β, and Nlrp3 mRNA expression (Figs. 7g–j), and induced the expression of the autophagy protein p62 (Fig. 7l). These mice also presented a tendency towards an increase of Lc3 (Fig. 7m).At protein level, DMOG-treated mice showed reduced NF-κB phosphorylation and NLRP3 expression, as well as enhanced autophagy evidenced by the increase of the ratio LC3-II/LC3-I and degradation of p62, when compared with vehicle-treated mice (Fig. 7n and Supplementary Fig. 6A–E).

13Our results show that environmental hypoxia reduces TNFα, IL-6, and NLRP3 expression concomitantly with the induction of p62 mRNA expression and p62 protein clearance in the colon of patients with CD, suggesting an increase in autophagy turnover. In contrast, biopsies from patients with UC did not display changes in the expression of inflammatory mediators or p62. It is increasingly recognized that CD and UC constitute distinct molecular entities. The discrepancy between patients with CD and UC might reflect the different molecular mechanisms underlying both pathologies. Interestingly, recent studies have shown different HIF expression patterns in intestinal biopsies from patients with UC and CD, suggesting divergent responses to hypoxia. The heterogeneity in disease and treatments might also contribute to the contrasting molecular responses. In line with the results obtained in patients with CD, prolonged hypoxia downregulated NF-κB signaling and the expression of Tnfα, Il-6, and Il-1β in the DSS and Il-10 −/−mouse models of colitis. In HT-29 cells, hypoxia triggered transient accumulation of HIF-1α and resulted in the reduction of basal and LPS-induced NF-κB and IκBα phosphorylation. These results are in agreement with recent reports showing that hypoxia inhibits LPS-induced TNFα and NF-κB activation in colon epithelial cancer cells, and confirm previous studies showing protective effects of hypoxia on mucosal inflammation,,. Conversely, several studies suggest that reduced oxygen levels induce the expression of TNFα and IL-6 in monocytes and directly activate the NF-κB pathway by blocking hydroxylation of factors in that pathway,.This discrepancy, however, might reflect that responses to hypoxia differ in different cell types. Recent studies indicate that aircraft travel-associated low oxygen partial pressure heightens the risk of flares in IBD patients,, and altitude-associated hypoxia increases the expression of inflammatory cytokines in the duodenum of healthy individuals. Nonetheless, these effects can be due to confounding factors such as stress, acute exercise, or local changes in circulation,. On the other hand, there is mounting evidence that hypoxia promotes resolution of inflammation in several models of disease,,. In transgenic mice, loss of HIF-1α expression in the epithelium resulted in a more severe colitis, while constitutively active HIF-1 was protective,,.

14In line with recent studies linking deficient autophagy with IBD,,, DSS-treated WT mice and Il-10 −/− mice presented accumulation of phosphorylated/activated mTOR, the main inhibitor of autophagy, and the autophagy protein p62, together with increased NF-κB phosphorylation and inflammatory gene expression. Hypoxia restored autophagy clearance of p62 reducing mTOR phosphorylation and proinflammatory gene expressionand signaling inthese mice. Hypoxia also upregulated p62 mRNA expression, pointing to a compensatory mechanism for autophagosomal p62 protein degradation. In HT-29 cells, hypoxia ameliorated inflammatory processes together with the inactivation of mTOR, and induced degradation of p62 and LC3, confirming that hypoxia triggers autophagy through the regulation the catalytic activity of mTOR. This is in good agreement with studies showing that mTOR represses autophagy by phosphorylating, and thereby downregulating the autophagy-associated proteins unc-51-like kinase 1 and ATG13,. The blockage of autophagy at late stage with chloroquine reverted hypoxia-mediated inhibition of NF-κB phosphorylation, confirming a causal link between hypoxia-mediated autophagy and the downregulation of inflammation. These results point to an effect upstream of the NF-κB signaling cascade. In this regard, several studies show that autophagy inhibits NF-κB activation through the suppression of the upstream regulatory factor IκB kinase (IKK), and the recruitment of phosphorylated IKK to the autophagosome compartment. Hypoxia-mediated inhibition of NF-κB nuclear translocation is also linked to reduced IKKβ phosphorylation in colon epithelial cells. The two main nuclear factors involved in the regulation of mucosal responses to oxygen deprivation, NF-κB, and HIF-1α, appear to be intimately linked at several mechanistic levels. In line with the significant increase of p62 mRNA expression observed under hypoxia, ChIP analysis in HT-29 cells showed an increased recruitment of HIF-1α, but not NF-κB, to the promoter of p62, indicating a role for HIF-1α in the expression of autophagy proteins. Conversely, NF-κB, but not HIF-1α, was bound to the NLRP3 promoter under normoxic and hypoxic conditions. The opposing binding of these nuclear factors to p62 and NLRP3 promoters suggest a converse transcriptional regulation of both genes, and appears to align with their antagonistic function. After 24 h of hypoxia, we found lysosomal accumulation in HT-29 cells, but could not detect any effects on active mitochondria. Since lysosomal formation precedes mitochondrial degradation, longer hypoxia periods might be needed to elicit mitochondria damage and clearance.

15Il-10 −/− Nlrp3 −/− mice did not present blockage of autophagy under normoxia when compared to Il-10 −/− mice, suggesting a role for NLRP3 in the regulation of autophagy. Interestingly, hypoxia downregulated NLRP3 concomitantly with the induction of autophagy in the DSS and Il-10 −/− mouse models of colitis, suggesting a role for NLRP3 in hypoxia-induced autophagy. These results were confirmed in HT-29 cells, where hypoxia triggered NLRP3 clearance in a proteasome-independent manner. In contrast to reports showing that autophagy regulates NLRP3 inflammasome function through the autophagosomal degradation of NLRP3, hypoxia-mediated NLRP3 clearance appears to be autophagy-independent, pointing to differential regulatory mechanisms for NLRP3 inflammasome-dependent and -independent effects. siRNA-mediated silencing of NLRP3 in HT-29 cells further promoted autophagy clearance of p62 and LC3, confirming a role for NLRP3 in the inhibition of autophagy and reinforcing the idea that NLRP3 degradation precedes and triggers autophagy activation. NLRP3 silencing also reduced the levels of phosphorylated mTOR, indicating that NLRP3 regulates autophagy upstream of the mTOR signaling pathway. Furthermore, Co-IP experiments identified NLRP3 as a novel binding partner of mTOR. Our results point to that hypoxia-mediated NLRP3 clearance reduces the physical interaction between NLRP3 and mTOR, suggesting that the interaction of both factors is required for mTOR function and subsequent repression of autophagy under normoxia. Inflammasome-independent effects for NLRP3 have been increasingly recognized, and recent research has revealed functions for NLRP3 that do not require the assembly of the inflammasome complex, including the upstream regulation of transforming growth factorβ/Smad signaling,. It is well stablished that the catalitic activity of mTOR is tightly regulated by its physical interaction with mTOR complex-binding partners. For example, Raptor regulates mTOR activity by orchestrating the assembly of the mTOR complex and recruiting substrates for mTOR,. Hypoxia has been proposed to inhibit mTOR through the regulation of proteins that interfere with its binding with RHEB. Thus, PML protein binds mTOR during hypoxia and inactivate it through the secuestration in nuclear bodies. Likewise, the hypoxia-inducible factor BNIP3 inhibits mTOR activation by interacting with RHEB. Our results suggest that hypoxia-mediated NLRP3 degradation not only counteracts mTOR activation, but might also limit NLRP3 inflammasome-dependent responses trigered by danger signals such as ATP. Notably, hypoxia has been recently discovered to contribute to the control of inflammation, promoting the phosphohydrolysis of extracellular ATP, thereby preventing purinergic signaling.

16The finding that hypoxia might promote the resolution of inflammation has sparked the interest in the therapeutic use of hydroxylase inhibitors to stabilize HIF for the treatment of IBD. Thus, the hydroxilase inhibitors AKB-4924 and DMOG are protective in the DSS and trinitrobenzene sulfonid acid mouse models of colitis,. Our results indicate that DMOG ameliorates DSS colitis and significantly reduces inflammatory gene expression and signaling. Interestingly, HIF-1α stabilization triggered clearance of NLRP3 and p62, and the conversion of LC3-I to LC3-II. These results confirm our findings in the DSS and Il-10 −/− mouse models of colitis, and highlight the important contribution of HIF-1α to hypoxia-mediated autophagy. Recent research has underscored the importance of HIF signaling in controlling inflammatory and innate host defense responses. For example, bacteria- and LPS-mediated induction of HIF-1α has been demonstrated through an oxygen-independent mechanism in myloid and colonic epithelial cells–. Consequently, manipulating HIF levels has been considered in the context of cancer therapy, infectious diseases, and chronic inflammatory conditions such as rheumatoid arthritis and IBD.

17In summary, our results identify NLRP3 as a binding partner of mTOR and support a novel role for NLRP3 in the regulation of hypoxia-mediated anti-inflammatory effects through NLRP3 clearance, and subsequent autophagy activation and reduction of inflammatory processes in IECs (Fig. 8). Hypoxia-mediated NLRP3 degradation not only prevents stimulatory effects on mTOR signaling but might also limit NLRP3 inflammasome-dependent responses.

18Healthy subjects (n = 10), patients with CD (n = 11) and patients with UC (n = 9) in stable remission, were subjected to hypoxic conditions resembling an altitude of 4000 m.a.s.l. for 3 h using a hypobaric pressure chamber at the Swiss aeromedical center in Dubendorf, Switzerland (ClinicalTrials.gov number, NCT02849821). Clinical activity in patients with CD and patients with UC was assessed using the Harvey Bradshaw Index and the Partial Mayo Score, respectively. The participant characteristics are shown in Table 1. Distal colon biopsies were taken the day before entering the hypobaric chamber, immediately after hypoxia, and 1 week after collection of the first biopsy at the Division of Gastroenterology and Hepatology of the University Hospital Zurich. This study was approved by the Ethics Committee of the Canton of Zurich (KEK-ZH Nr. 2013-0284) and all participants signed an informed consent.

19Intestinal specimens from the distal third of the colon were embedded in paraffin and cut in 5 µm sections with a microtome. Immunostaining for p62 was performed on Leica Bond Max instruments using Refine HRP-Kits (Leica Biosystems, Newcastle, UK). Paraffin-embedded tissue was pretreated for 20 min at 100 °C with citrate-based buffer at pH 6.0 (Leica Biosystems) prior to the incubation with p62 monoclonal antibody (Cat. No. GP62-C; PROGEN Biotechnik, Heidelberg, Germany) at a 1:750 dilution. Rabbit anti-Guinea pig (Cat. No. P0141; Dako, Glostrup, Denmark) was used as a secondary antibody at a 1:100 dilution. Immunohistochemistry score was determined by two independent, blinded investigators who assigned score values according to the prevalence of p62 staining (<10% positive cells = 0, 10–25% = 1, 25–50% = 2, 50–75% = 3, and 75–100% = 4).

20Female WT (C57BL/6J) and Nlrp3 −/− mice between 8–10 weeks of age were exposed to 2% DSS (MP Biomedicals) in drinking water ad libitum for 7days. Intestinal inflammation was induced in WT, Il-10 −/−, Nlrp3 −/−, and Il-10 −/− Nlrp3 −/− mice by oral administration of 1.5% DSS in drinking water ad libitum for 5 days,. Mice were separated into hypoxia and normoxia groups. The hypoxia group was exposed to hypoxia (8% O2) in a normobaric hypoxic chamber (Coy Laboratory Products, Grass Lake, MI), whereas the normoxia group (21% O2) was used as control. After 18 h of hypoxia, mice were killed for colonic biopsy collection. WT mice were exposed to 2% DSS in drinking water ad libitum for 5 days and were co-administered with 8 mg DMOG (Frontier Scientific, Logan, UT) dissolved in PBS, injected intraperitoneally every second day. Control mice received PBS. After a recovery period, mice were killed at day 9 and distal colon biopsies were taken. Ethical approval for all animal experiments was obtained by the University of Zurich from the Veterinary Office of the Canton of Zurich, Switzerland.

21Animals were anesthetized intraperitoneally with a mixture of ketamine 90–120 mg/kg body weight (Vétoquinol) and xylazine 8 mg/kg body weight (Bayer), and colonoscopy was performed using the Coloview miniendoscopic system (Karl Storz, Tuttlingen, Germany). Animals were killed and 1 cm of tissue was removed from the distal third of the colon and fixed in 4% formalin overnight. H&E-stained sections of the paraffin-embedded tissue were used for histological analysis. Histological score was calculated as the sum of epithelial damage and infiltration score, and was determined by two independent, blinded investigators.

22Colonic biopsies were disrupted using the GentleMACS Dissociator (Miltenyi Biotech, Gladbach, Germany). Total RNA from human and mice biopsies, and cells was extracted using RNeasy Plus Kit (Qiagen, Hombrechtikon, Switzerland) according to manufacturer’s instructions. For reverse transcription, the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) was used following the manufacturer’s instructions. RT-qPCR was performed using the TaqMan system 7900HT (Applied Biosystems), under the following cycling conditions: 20 s at 95 °C, followed by 40 cycles of 95 °C for 3 s, and 60 °C for 30 s with the TaqMan Fast Universal Mastermix. TaqMan Gene Expression Assays (all from Applied Biosystems) used in this study were human TNF (Hs00174128_m1), IL1B (Hs01555410_m1), IL6 (Hs00174131_m1), NLRP3 (Hs00918082_m1), p62 (SQSTM1) (Hs01061917_g1), LC3 (MAP1LC3A) (Hs01076567_g1) and β-actin Vic TAMRA (4310881E), and mouse Tnf (Mm99999068_m1), Il1b (Mm00434228_m1), Il6 (Mm00446190_m1), Nlrp3 (Mm00840904_m1), Vegfa (Mm00437306_m1), p62 (Sqstm1) (Mm00448091_m1), Lc3 (Map1lc3a) (Mm00458724_m1) and β-actin VIC TAMRA (4352341E). Relative mRNA expression was calculated by the comparative ∆∆Ct method using β-actin as housekeeping gene.

23Total protein from biopsies and cells was collected in M-PER lysis buffer (Thermo Fisher Scientific, Reinach, Switzerland). Phospho-p65/RelA (Ser536, 93H1) (Cat. No. 3033), p65/RelA (D14E12) (Cat. No. 8242), phospho-mTOR (Ser2448) (Cat. No. 2971), mTOR (7C10) (Cat. No. 2983), p62 (Cat. No. 5114), phospho-IκBα (Ser32, 14D4) (Cat. No. 2859), IκBα (Cat. No. 9242), ubiquitin (Cat. No. 3933) (Cell Signaling Technology, ZA Leiden, The Netherlands), anti-mouse NLRP3 (Cat. No. AG-20B-0014-C100; AdipoGen, Liestal, Switzerland), anti-mouse HIF-1α (EP1215Y) (Cat. No. ab51608; Abcam, Cambridge, UK), anti-human HIF-1α (Cat. No. NB100-479; Novus Biologicals, Littletown, CO), anti human NLRP3 (Cat. No. ALX-804-819-C100; Enzo Life Sciences, Lausen, Switzerland), LC3 (Cat. No. L8919) and β-actin (Cat. No. A5441) (Sigma-Aldrich, St Louis, MO) antibodies were used at 1:1000 dilution for western blot. Co-IP was performed overnight at 4 °C using the human NLRP3 antibody at 1:100 dilution. Immunocomplexes were collected with protein G sepharose beads (GE Healthcare, Glattburgg, Switzerland) for 1 h at 4 °C prior to Western blotting. Densitometry of bands was measured using ImageJ software.

24HT-29 cells were obtained from the German Collection of Cells and Microorganisms (DSMZ, Braunschweig, Germany), and cultured in DMEM medium supplemented with 10% FCS (VWR, Dietikon, Switzerland). In some experiments, HT-29 cells were treated with 10 µg/ml LPS (Invitrogen, Carlsbad, CA). Late-stage autophagy was inhibited with 40 µM chloroquine (Sigma-Aldrich) for 1 h prior to hypoxia. Proteasomal activity was inhibited with 20 µM MG132 (Sigma-Aldrich) for 1 h prior to 24 h hypoxia. Cells were exposed to hypoxia (0.2% O2) in a hypoxia workstation incubator (In vivo 400, Baker Ruskinn, Bridgend, UK). Cells maintained in normoxia (21% O2) for the same time-period were used as controls.

25After 24 h under normoxia or hypoxia, HT-29 cells were stained with LysoTracker Yellow-HCK-123 and MitoTracker Deep Red FM (Thermo Fisher Scientific) according to manufacturer’s instructions. For LC3 staining, HT-29 cells subjected to normoxia or hypoxia for 24 h were fixed with 4% paraformaldehyde for 20 min and then permeabilized in 100% methanol. After blocking with 3% bovine serum albumin, cells were incubated with LC3 antibody (Cat. No. PM036; MBLI, Woburn, MA) at 1:200 dilution overnight at 4 °C. Cells were then incubated with Alexa 488 conjugated secondary rabbit antibody (Invitrogen) for 1 h and DAPI for 5 min before mounting with antifade medium (Dako). Cells were analysed by a Leica SP5 laser scanning confocal microscope (Leica Microsystems, Wetzlar, Germany). Fluorescence images of the same sample were acquired and processed using Leica confocal software (LAS-AF Lite; Leica Microsystems). Quantification of LC3/DAPI was performed by two independent, blinded investigators using ImageJ software in at least tenareas of interest pooled from four independent experiments.

26HT-29 cells were transfected with 10 pmol of three sets of siRNAs for NLRP3 (Life Technologies) using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions.

27ChIP was performed using the ChIP-IT Express Enzymatic kit (Active Motif, Carlsbad, CA) according to the manufacturer’s instructions. Immunoprecipitation of 25 µg of DNA was carried out overnight at 4 °C using anti-HIF-1α (Cat. No. NB100-479; Novus biologicals, Littletown, CO) or anti-p65/RelA (D14E12) (Cat. No. 8242; Cell Signaling Technology) antibodies at a concentration of 1 μg/ml. DNA isolated from an aliquot of the total nuclear extract was used as a loading control for the PCR (input control). PCR was performed with total DNA and immunoprecipitated DNA using the following promoter-specific primers, p62 promoter-binding site for HIF-1α: 5′-ATCCCGATCGGCTCCCG-3′ (forward), 5′-ACTCCCACTCCAGCCTTACC-3′ (reverse) (70 bp), and NF-κB: 5′-ACTGGGCGGTTGAAGCC-3′ (forward), 5′-TCCAGAGGCCCAGAGGAAG-3′ (reverse) (87 bp), NLRP3 promoter-binding site for HIF-1α: 5′-TTCTTCTGGCGTCTCCAAGG-3′ (forward), 5′-TTGCACACGTAGAGCCCAAA-3′ (reverse) (81 bp), and NF-κB: 5′-CCAGTATGGAACCGAGACACG-3′ (forward), 5′-CAGAGCAGAGGCAGAGAGACT-3′ (reverse) (97 bp). The PCR products (10 µl) were subjected to electrophoresis on a 2% agarose gel.

28Statistical analysis was performed using Student t test, or one-way ANOVA followed by Tukey’s post-test. The results are expressed as mean + standard error of the mean (s.e.m.) and significance was set as p < 0.05.

29The authors declare that all the data supporting the findings of this study are available in the manuscript or are available from the corresponding author upon reasonable request.