Hugh Thomas

NAFLD: A critical role for the NLRP3 inflammasome in NASH

NLRP3 inflammasome blockade reduces liver inflammation and fibrosis in experimental NASH in mice. J. Hepatol. http://dx.doi.org/10.1016/ j.jhep.2017.01.022 (2017) Inhibition of the NLRP3 inflammasome in a mouse model of NASH reduces hepatic inflammation and liver fibrosis, according to new research published in the Journal of Hepatology. Excess hepatic levels of lipotoxic lipids, such as ceramides, diacylglycerols and free fatty acids, have been suggested as key molecules promoting progression of NAFLD to NASH. In the past few years, other lipid species have also been investigated. “We and several other groups have found evidence to support free cholesterol as the lipotoxin,” explains author Geoffrey Farrell. “In the atherosclerosis literature, it has been shown that cholesterol crystals ... provoke activation of the NLRP3 inflammasome.” Hepatic NLRP3 inflammasome activation, which promotes caspase-1-dependent IL-1β production, occurs in patients with NASH, and evidence from knockout mouse models suggests activation of the inflammasome is important in NAFLD progression. To probe the role of the NLRP3 inflammasome in NASH, Farrell and colleagues treated an overnutrition mouse model of NASH with a selective small molecule NLRP3 inhibitor, MCC950, or a control solution for 8 weeks. Mice treated with the control solution had upregulated levels of hepatic NLRP3, IL-1β and active caspase 1, in addition to NASH pathology and cholesterol crystal accumulation in the liver. By contrast, MCC950 prevented increased expression of these proteins and reduced levels of plasma inflammatory markers. Despite no treatment effect on body weight, hepatic steatosis extent and cholesterol crystal levels, NASH severity was reduced in treated mice, primarily as a result of reduced histologically assessed liver inflammation and hepatocyte ballooning. Liver fibrosis progression over the 8 weeks was also lessened in mice treated with MCC950. In vitro, Kupffer cells and macrophages were activated by cholesterol crystals, releasing IL-1β. Treatment with MCC950 inhibited this activation and associated neutrophil chemotaxis, suggesting that the treatment suppresses cholesterol-crystalmediated immune activation. “By using an intervention in whole animals that specifically blocks NLRP3, these results are the most powerful to date that indicate a key role for NLRP3 activation in the liver for NASH pathogenesis,” Farrell concludes. “We now plan to study more rigorously, using a longer time course, the extent to which NLRP3 blockade (using MCC950 as our tool compound) can reverse liver fibrosis in established NASH in our mouse model.” Hugh Thomas N A F L D

Liver: Regulating profibrotic transcription

NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY www.nature.com/nrgastro JQ1 treatment in mice with established CCl4‐induced liver fibrosis reduced the expression of activated HSC markers and seemed to resolve fibrosis, as assessed by histological scoring. “Currently, there are limited therapeutic options for liver fibrosis. Our studies show that small molecule epigenetic modulators hold significant therapeutic promise in the treatment of liver fibrosis,” explains corresponding author Michael Downes. Hugh Thomas Profibrotic genes BRD4 Acetylated lysine

Diabetes: Circulating factors implicated in diabetic enteropathy.

and a substantial decrease in epithelialcell proliferation. A subsequent screen of circulating serum factors showed reduced levels of IGF-1 and increased levels of IGFBP-3 in patients with T1DM compared with healthy individuals. Crypts isolated from patients with T1DM and cultured to form organoids failed to grow, unlike those cultured from healthy people. The investigators showed that addition of IGF-1 to the culture medium rescued organoid growth, but addition of IGFBP-3 negated this improvement. IGFBP-3 was found to promote apoptosis by directly binding the IGFBP-3 receptor (IGFBP-3R, also known as TMEM219). DIABETES

Pancreatic cancer: CYP3A5 contributes to PDAC chemoresistance

mechanisms underlying resistance to chemotherapeutics in patients with pancreatic ductal adenocarcinoma (PDAC). Increased expression of CYP3A5, which metabolizes the cytoskeleton-targetting drug paclitaxel and the tyrosine-kinase inhibitor erlotinib, contributes to tumour-cell-autonomous and acquired drug resistance in different PDAC subtypes. PDAC is an aggressive cancer with a poor prognosis. Current chemotherapy drugs produce limited survival benefit, in part because PDAC tumours are intrinsically resistant or acquire resistance to treatment. “Although some progress has been made in recent years, the underlying mechanisms of treatment resistance in PDAC remain poorly understood,” says author Martin Sprick. “Our first goal was to develop improved model systems which closer resemble patient tumours.” Using patient-derived PDAC tumour samples xenografted on to the pancreata of immunodeficient mice, the researchers first generated primary PDAC cell lines. Whereas only two of the three tumour subtypes (exocrine-like, quasi-mesenchymal (QM) and classical) found in patient tumour samples were represented in previously available PDAC cell lines, the primary cell lines derived by the authors contained all PDAC subtypes. To assess whether PDAC subtypes differ in their sensitivity to chemotherapy agents, the authors treated the patient-derived PDAC cell lines with paclitaxel or erlotinib. The exocrine-like tumour subtype was markedly more resistant to these drugs than other subtypes — subsequent analysis revealed increased expression of CYP-family enzymes involved in drug metabolism in exocrine-like tumour cells. “Although expression of CYP-family enzymes has been described for various tumours, to our knowledge no one has demonstrated that tumours can acquire this mechanism to locally inactive drugs,” explains author Elisa Noll. CYP3A5 was highly expressed exclusively in exocrine-like PDAC tumour cells; knockdown of this enzyme rendered these cells sensitive to erlotinib and paclitaxel. Furthermore, mice with established exocrine-like tumours in which CYP3A5 expression was ablated were more sensitive to erlotinib and paclitaxel than mice with tumours expressing CYP3A5. The effectiveness of smallmolecule drugs in PDAC is also limited by acquired resistance. Notably, CYP3A5 was upregulated by treatment with paclitaxel or erlotinib in xenograft tumours that were previously CYP3A5-negative. The researchers also showed that acquired resistance to erlotinib and paclitaxel in classical-subtype and QM-subtype cell lines was associated with induced CYP3A5 expression. Knockdown of CYP3A5 in these cell lines restored drug sensitivity. “Developing approaches to inhibit the CYP3A5 enyzmatic cascade could pave the way to overcoming drug resistance in a subset of patients with PDAC,” concludes author Andreas Trumpp. Hugh Thomas PA N C R E AT I C C A N C E R

Tuft cells revealed as norovirus target

Tuft cells are the targets of norovirus infection in mice, according to a new study published in Science. The findings have important implications for understanding norovirus pathogenesis and how interactions with other gut organisms alter the disease. Globally, human noroviruses are the leading cause of acute viral gastroenteritis, an important disease burden responsible for around 200,000 deaths a year. There are no vaccines or antiviral therapies for human noroviruses, in part because understanding of the mechanisms of norovirus pathogenesis is limited. In a new study, Craig Wilen and colleagues sought to shed light on how mouse noroviruses interact with the host and cause disease. “We use mouse norovirus as a model for human norovirus because mouse norovirus can replicate in cell culture and has a reverse genetics system,” says Wilen. As the cellular tropism of norovirus in mice or humans was unclear, the researchers first used bone marrow transplants in a mouse model of norovirus infection to show that the target cell was radiationresistant, and therefore an epithelial cell. I N F E C T I O U S D I S E A S E

NAFLD: Loss of CD4 + T cells in HCC

fatty acids in MYC-ON mice, only levels of linoleic acid were increased by MCD diet. Isolated mouse and human CD4+ T cells showed increased apoptosis, mediated by upregulated mitochondrial production of reactive oxygen species and superoxide, when cultured with linoleic but not palmitic acid. Finally, MYC-ON mice fed MCD and given the antioxidant N-acetylcysteine had markedly fewer microscopic tumours than untreated MCD-fed MYC-ON mice, an effect lost when N-acetylcysteine-treated mice were injected with anti-CD4 antibody. In future work, Greten and colleagues plan to explore how linoleic acid induces mitochondrial dysfunction in CD4+ T cells, and study the effect of NAFLD on anti-tumour immunity in a therapeutic setting. Hugh Thomas

Intestinal tract: Gut endothelial cells — another line of defence

characterize the gut–vascular barrier (GVB), showing that endothelial cells control the passage of antigens and commensal gut microbiota from the intestine into the bloodstream. The intestinal barrier prevents the passage of gut microbiota into the circulation yet allows nutrients to be absorbed. Bacteria that do manage to cross the intestinal epithelium are prevented from entering the bloodstream, despite the close proximity of blood vessels to epithelial cells. “We wondered why bacteria would not easily enter the blood vessels and reach the liver ... thus, we hypothesized that the endothelial cells of the blood vessels might generate a barrier [to bacteria],” explains corresponding author Maria Rescigno. Rescigno and colleagues determined the permeability of the GVB in mice by intravenously injecting fluorescently labelled particles of varying molecular weights. The researchers showed that the GVB of healthy mice were impermeable to 70 kDa particles, indicating that bacteria, which are larger, might be physically excluded from the circulation. In common with other vascular barriers, endothelial cells in the GVB expressed a number of tight junction and adherens junction proteins. To determine if the ability to cross the epithelial barrier is enough for bacteria to cross the GVB, the investigators constructed a strain of Escherichia coli expressing a recombinant protein that enabled it to penetrate the gut epithelium. This bacteria was excluded from the circulation of infected mice despite crossing the epithelial barrier, suggesting that intestinal endothelial cells constitute a distinct vascular barrier against gut microorganisms. Lastly, the researchers showed that the GVB is present in healthy human intestine by staining for junctional protein expression on intestinal endothelial cells. In intestinal tissue from patients with coeliac disease and no observable epithelial barrier defect, a marker of endothelial permeability was upregulated and was associated with increased liver damage, defined by raised serum alanine aminotransferase levels. “We now want to focus on [endothelial barrier function in] ... diseases that are related to the gut–liver axis and that might be triggered by mucosal barrier defects,” concludes Rescigno.

Gut microbiota: Growing up together [mdash] gut microbiota assembly and IgA

assembly of the gut microbiota in humans over the first 2 years of life, as well as the associated changing gut mucosal IgA responses. Development and maturation of the human gut microbiota is a complex process that begins at birth and continues over the next few years. Despite the known importance of mucosal IgA in regulating gut barrier function, little is known about how gut mucosal IgA responses evolve during early life. “We designed the current study to determine how IgA responses to the microbiota normally develop in healthy infants and children, as well as how they are impacted by human genetics, environmental factors, and diet history,” explains author Jeffrey Gordon. Using a cohort of 40 healthy twin pairs, the investigators characterized the bacterial composition of 1,670 faecal samples taken monthly throughout the first 2–3 years of life by 16S ribosomal RNA (rRNA) sequencing. They identified a series of bacterial taxa defining a programme of gut microbial community assembly; this programme was shared across twin pairs, and varied by family membership and whether the child was fed breast or formula milk. Next, the researchers investigated how IgA responses co-developed with the gut microbiota. For each stool sample, fluorescenceactivated cell sorting employing anti-IgA labelling was used to separate microbiota that had elicited an IgA response. This technique generated IgA and IgA gut microbiota fractions for each individual at each time point; 16S rRNA sequencing was then used to determine composition and abundance of present bacterial species. This analysis identified specific microbiota members that were consistently targeted, or not targeted, by IgA responses. These mucosal IgA responses, although varying between twin pairs early in life, converged to a common pattern of targeting by the second postnatal year. Lastly, the researchers transplanted faecal samples obtained at 6 and 18 months of age from two twin pairs into germ-free mice. “We find that age-associated IgA responses observed in the human population are largely recapitulated in these mice,” explains first author Joseph Planer. The authors next plan to assess how IgA responses to microbiota assembly vary depending on environmental influences. “Our group is performing these analyses now,” says Planer.

Liver cancer: IGF2 — an epigenetic oncodriver in HCC

ME ME Overexpression of insulin-like growth factor 2 (IGF2), mediated by epigenetic mechanisms, drives tumorigenesis in some hepatocellular carcinomas (HCCs), according to new research. HCC is a substantial cause of global mortality. The only approved systemic treatment for patients with late-stage HCC is sorafenib, and research efforts in the past decade have sought to uncover genetic alterations associated with the disease, with a view to defining new druggable targets. However, less is known about the epigenetic landscape of HCC. Expression of IGF2, which is paternally imprinted, is known to be markedly upregulated in some HCCs, and epigenetic mechanisms could be responsible for these changes. To define the role and mechanisms of IGF2 dysregulation in HCC, Josep Llovet and colleagues analysed 228 tumour samples from patients with the disease, as well as 168 paired non-tumour adjacent cirrhotic tissue samples and 10 normal liver samples from patients without HCC. Samples from tumours showed increased IGF2 mRNA levels compared with non-tumour samples; importantly, tumours with high IGF2 expression (>20-fold increase versus normal liver samples) had greater hypomethylation of promoters typically only active in fetal tissue than samples with lower IGF2 expression (<20-fold increase). In human HCC cell lines with low IGF2 expression and high fetal promoter methylation, forced demethylation of these promoters resulted in IGF2 overexpression. Using genetically engineered mouse models of HCC, the researchers found that Igf2 overexpression accelerated tumour progression and decreased survival. However, liver-specific overexpression did not promote tumour initiation, suggesting Igf2 is not a transforming oncogene in HCC. Given the overexpression of IGF2 in human HCCs and evidence of its tumorigenic action, Llovet and colleagues assessed whether a monoclonal antibody (BI 836845, also known as xentuzumab) against IGF1 and IGF2 might have therapeutic potential. Mice with xenograft HCC tumours were given an inactive control, sorafenib, xentuzumab or a combination of both active compounds. Compared with standalone sorafenib or control compound treatment, xentuzumab alone or in combination with sorafenib reduced tumour growth and prolonged survival. “There are two direct implications of this work,” concludes Llovet. “First, to test xentuzumab in clinical proof-of-concept trials in HCC. Second, to explore the ideal biomarker to identify the subpopulation who might benefit from this drug.” Hugh Thomas L I V E R C A N C E R

NASH: NASH and TLR9

Laura Marshall /N PG New research shows that TLR9, an endosomal pattern-recognition receptor required for NASH development in mice, is activated by mitochondrial DNA (mtDNA) released from hepatocytes. NAFLD is a leading cause of chronic liver disease in the developed world. NAFLD can progress to NASH in some individuals, although the mechanisms underlying this progression remain unclear. “We have shown that TLR9 is required to elicit various types of liver injury in mice, including NASH, yet TLR9 is ubiquitously expressed. This made us consider if the difference between people who are obese with or without NASH is the presence of the ligand for TLR9,” explains corresponding author Wajahat Mehal. Because mtDNA is released by hepatocytes after liver injury, the researchers measured plasma mtDNA concentrations in patients who were obese, with or without NASH. Patients with NASH had higher levels of plasma mtDNA than patients without NASH. In addition, plasma from patients with NASH produced stronger activation of a TLR9 reporter cell line than plasma from patients without NASH. Using a mouse model of NASH induced by high-fat diet feeding, the authors next investigated whether loss of TLR9 specifically in lysozyme-producing cells, which include liver-resident macrophages, could reduce liver inflammation and steatosis. Mice lacking Tlr9 in these cells had less severe NASH and lower expression of pro-inflammatory genes than wild-type mice when both groups were fed a high-fat diet. Administration of a TLR7–TLR9 antagonist to wild-type mice with established NASH reduced the hepatic expression of inflammatory cytokines and markedly reduced NASH severity. The authors plan to investigate the mechanisms of mtDNA release from hepatocytes in future work. “MtDNA from hepatocytes might be also responsible for activation of macrophages via TLR9 in other places such as adipose tissue or atheroma plaques,” remarks Mehal. Hugh Thomas R E S E A R C H H I G H L I G H T S