Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract Non-alcoholic fatty liver disease (NAFLD) is the most common liver disease in industrialized countries and is increasing in prevalence. The pathomechanisms, however, are poorly understood. This study assessed the unexpected role of the Hedgehog pathway in adult liver lipid metabolism. Using transgenic mice with conditional hepatocyte-specific deletion of Smoothened in adult mice, we showed that hepatocellular inhibition of Hedgehog signaling leads to steatosis by altering the abundance of the transcription factors GLI1 and GLI3. This steatotic 'Gli-code' caused the modulation of a complex network of lipogenic transcription factors and enzymes, including SREBP1 and PNPLA3, as demonstrated by microarray analysis and siRNA experiments and could be confirmed in other steatotic mouse models as well as in steatotic human livers. Conversely, activation of the Hedgehog pathway reversed the "Gli-code" and mitigated hepatic steatosis. Collectively, our results reveal that dysfunctions in the Hedgehog pathway play an important role in hepatic steatosis and beyond. https://doi.org/10.7554/eLife.13308.001 eLife digest The liver is one of the main organs responsible for processing everything that mammals eat and drink. Nutrients absorbed by the gut like sugars and lipids (fats) are processed by the liver and are stored or distributed to provide energy to other organs. Sometimes these metabolic processes become unbalanced. This can lead to lipids accumulating in the liver – a process known as steatosis, which is a feature of human non-alcoholic fatty liver disease. In organs like the liver, cells are instructed how to behave via signaling pathways. A protein outside the cell signals to specific proteins inside, which switch on a set of target genes. One such pathway is the Hedgehog pathway, which primarily regulates tissue regeneration and the development of embryos. A component of this pathway is the Smoothened gene, which indirectly switches on proteins called GLI factors that regulate metabolic genes, including those involved in lipid metabolism. The Hedgehog pathway has been found to control the metabolism of lipids in fat tissue but it is not known whether it is important for lipid metabolism in the liver. Matz-Soja et al. investigated this possible role of the Hedgehog pathway in the liver using mice with a Smoothened gene that could be deleted specifically in that organ. This deletion disrupted Hedgehog signaling and led to lipids accumulating in the liver and eventually to steatosis. These changes were associated with an increase in the amounts and activityof several enzymes (and the proteins that regulate these enzymes) that help to synthesize lipids. Steatosis was also associated with low amounts of two of the three GLI factors; indeed, this seems to be key for triggering problems with lipid metabolism. Human livers with steatosis showed the same changes in levels of the GLI factors. Increasing the amount of GLI factors in liver cells taken from mice with steatosis reduced the accumulation of lipids and brought lipid metabolism back to its normal balance. A focus of future studies will be to understand how the Hedgehog signaling pathway interacts with other signaling pathways known to regulate liver lipid metabolism, such as insulin signaling. This knowledge will help clinicians to design new treatments for lipid-associated diseases like non-alcoholic fatty liver disease. https://doi.org/10.7554/eLife.13308.002 Introduction As the most common liver disease in the western countries and a disease with an increasing prevalence, non-alcoholic fatty liver disease (NAFLD) is the subject of many investigations and studies (Ahmed et al., 2015; Dongiovanni et al., 2015; Zhang and Lu, 2015) to identify unknown risk factors and new treatment strategies. Hepatic steatosis is the hallmark feature of NAFLD (Enomoto et al., 2015) and has the potential to develop into more severe steatohepatitis (NASH), which can progress to fibrosis, cirrhosis and cancer. Steatosis occurs when the rate of fatty acid delivery exceeds the rate of fatty acid removal (oxidation and export). A comprehensive knowledge of the hepatic lipid metabolism and its control mechanisms is crucial for preventing and treating liver steatosis. To date, these control mechanisms, e.g. the factors governing hepatocyte heterogeneity in lipid metabolism, remain largely unknown and have not yet received adequate attention in the discussion of NAFLD (Postic and Girard, 2008). Metabolic zonation of the liver is of considerable importance for the optimal integration and regulation of the plethora of different hepatic functions and metabolic homeostasis (Gebhardt, 1992; Gebhardt and Matz-Soja, 2014). Recently, Wnt/beta-catenin signaling was recognized as a master regulator of the zonal distribution of nitrogen metabolism in the adult liver (Benhamouche et al., 2006; Gebhardt and Hovhannisyan, 2010; Monga, 2015) and it influences the balance between the anabolic and catabolic functions of glucose metabolism (Chafey et al., 2009). The Hedgehog (Hh) signaling cascade is another important pathway that determines embryonic patterning, cell growth and tissue repair (Omenetti et al., 2011; Gu and Xie, 2015), and often acts in close crosstalk with Wnt/beta-catenin signaling (Toku et al., 2011). Similar to Wnt/beta-catenin signaling, the Hh pathway can act in canonical and non-canonical manners, as explicitly described by Teperino et al. (Teperino et al., 2014). To activate canonical Hh signaling in mammals, the ligands Sonic, Indian and Desert Hedgehog (SHH, IHH and DHH) interact with the surface receptors Patched 1 and 2 (PTCH1 and PTCH2) which remove their inhibition of the co-receptor Smoothened (SMO). Active SMO triggers the activation of the GLI (Glioma-associated oncogene) transcription factors (TFs) (GLI1, 2 and 3) by preventing the conversion of GLI2 and GLI3 into transcriptional repressors (Ruiz i Altaba, 1999; Infante et al., 2015). According to Ruiz I Altaba et al., the 'Gli-code', i.e. the combinatorial and cooperative function of the repressing and activating forms of all GLI factors, forms the basis of the integration of Hh signals (as well as of multiple other morphogens and cytokines) in embryogenesis and carcinogenesis (Ruiz i Altaba, 1999; Ruiz i Altaba et al., 2003; 2007). To date, Hh signaling has been found to play a considerable role in various scenarios of adult liver regeneration from progenitor cells (Sicklick et al., 2006) and in the regulation of the compensatory outgrowth of progenitors and myofibroblasts (Jung et al., 2010). Regarding liver metabolism, we recently showed that the regulation of the IGF-axis (insulin-like growth factor) in mature hepatocytes is controlled by Hh signaling (Matz-Soja et al., 2014). Moreover, we were also able to show that the TFs GLI1, GLI2 and GLI3 form a unique self-stabilizing network in hepatocytes and regulate several metabolic genes, including lipid associated factors (Schmidt-Heck et al., 2015). Therefore, in this study, we have focused on the alterations in lipid metabolism as a consequence of hepatic Hh signaling modulations in vivo and in vitro. To address this aim, a mouse model with a conditional hepatocyte-specific deletion of Smo in adult mice was established, in order to avoid interference with developmental effects of Hh signaling. For the in vitro investigations, we used RNA interference (RNAi) technology to knock down several genes of the Hh pathway in cultured hepatocytes (Böttger et al., 2015; Schmidt-Heck et al., 2015). The results of our investigations clearly show that the Hh pathway is a strong regulator of lipid metabolism in the adult liver. Furthermore, we show that impaired Hh signaling leads to increased expression of lipogenic TFs and enzymes with different zonal preference which finally results in steatosis. Conversely, we demonstrate that slight activation of the pathway by Sufu knockdown, small molecule agonists or GLI overexpression can mitigate lipid accumulation in steatotic livers. Results Breeding of transgenic mice As depicted in Figure 1A–B, triple transgenic mice which have a conditional hepatocyte-specific ablation of Smo in response to doxycycline (abbreviated SLC mice) were generated. Thus, the Smotm2Amc/J mice (Jackson Laboratories), which possess loxP sites flanking exon 1 of the Smo gene (Long et al., 2001), were crossed with the LC-1/rTALAP-1 mice which are working with a tetracycline-controlled transcriptional activation of the Cre recombinase protein (provided by Hermann Bujard) (Schonig et al., 2002). In the LC-1/rTALAP-1 mice the synthetic transactivator variant (rtetR) of the tet-repressor present in rTALAP-1 mice is driven by the LAP-promoter (PLAP). In the presence of doxyxycline, rtetR binds to an array of seven tet operator sequences (tetO7) activating transcription of the Cre recombinase gene (tet-on-system) (Figure 1A). The offspring were genotyped by PCR for the Smo wildtype (Smo WT) and Smo floxed (Smo flx) alleles, the doxycycline responsible element (rtetR) and the Cre recombinase (Primer: Supplementary file 1A). Figure 1 with 1 supplement see all Download asset Open asset Strategy for conditional and hepatocyte-specific deletion of Smo. (A) Scheme of the tet-on system in the LC-1/rTALAP-1 mice. (B) Structure of Smo with loxP sites. (C) PCR, for Cre recombinase, yielded a 400 bp fragment in SLC-KO mice only. (D) PCR from liver tissue during treatment with doxycycline, yielding a 600 bp amplicon for SLC-WT alleles and a 350 bp amplicon for the recombinant Smo alleles in the SLC-KO mice. (E) Immunohistochemical staining of Cre recombinase in liver sections of the SLC-WT and the SLC-KO mice. Bar: 100 µm. (F) qRT-PCR of Smo in different tissues and isolated hepatocytes of the SLC-WT (n=6–10) and the SLC-KO (n=6–10) mice. Source files of all data used for the quantitative analysis are available in the Figure 1—source data 1. https://doi.org/10.7554/eLife.13308.003 Figure 1—source data 1 Source data of qRT-PCR of Smo in different tissues and isolated hepatocytes of the SLC-WT and the SLC-KO mice (Figure 1F). https://doi.org/10.7554/eLife.13308.004 Download elife-13308-fig1-data1-v1.docx At the age of 8 weeks (to avoid the hormonal complications of adolescence), hepatocyte-specific ablation of Smo was induced by adding 2 mg/ml doxycycline hydrochloride (Sigma, Germany) to the drinking water for 10 days to promote the expression of the Cre recombinase (Figure 1C). During this period, the Smo rec. (recombinant) primer yielded a 350 bp fragment, indicating the deletion of the floxed domain (Figure 1D) (Primer: Supplementary file 1A). After 10 days, nearly all hepatocytes were positive for Cre recombinase protein (Figure 1E). After this treatment, the mice were maintained under normal conditions until 12 weeks of age. After sacrifice, the specificity of the Smo deletion was measured via qRT-PCR, indicating that there was a significant decrease in Smo expression in the liver material and isolated hepatocytes (Figure 1F). As already shown in our previous article, no adverse side effect of the doxycycline treatment could be observed on body weight (Matz-Soja et al., 2014). Likewise, the expression of important genes of lipid metabolism in hepatocytes was not affected by doxycycline (Figure 1—figure supplement 1). Phenotypic features of Smo knockout mice When maintained without doxycycline, the transgenic SLC mice developed without a phenotype. The deletion of Smo at 8 weeks of age resulted in pronounced liver steatosis within 5 weeks (Figure 2A). The H&E and fat red staining clearly showed lipid droplet accumulations, which are most prominent in the midzone to periportal zone, but occasionally encompassed the entire parenchyma (Figure 2A). Quantification of the fat red staining revealed a 7-fold increase in the SLC-KO mice compared with the WT controls (Figure 2B). Because the insulin and glucose levels changed only marginally (Table 1), we conclude that insulin resistance does not contribute to the steatosis in SLC-KO mice at least during the first five weeks after doxycycline administration. This assumption is confirmed by the fact that we could not detect any significant changes in hepatic expression for insulin receptor (Insr), insulin receptor substrate 1 and 2 (Irs1/2) (Figure 2—figure supplement 1). With the exception of the steatosis and lower liver/body weight ratio compared with the SLC-WT mice (Figure 2C), there were no other indications of overt liver damage in the SLC-KO mice. Accordingly, the serum activities of ASAT (aspartate aminotransferase), ALAT (alanine aminotransferase) and GLDH (glutamate dehydrogenase) were not different from those of the SLC-WT mice (Table 1). The plasma triglyceride levels were significantly increased in the VLDL (very low density lipoprotein) fraction, while no changes were found in LDL (low density lipoprotein) and HDL (high density lipoprotein) fractions (Figure 2—figure supplement 2). These findings indicate that steatosis is not caused by an impairment of triglyceride secretion from the hepatocytes. Figure 2 with 3 supplements see all Download asset Open asset Liver phenotype of the SLC mice. (A) The H&E and fat red staining of liver sections and hepatocytes showed strong steatosis in the male SLC-KO mice compared to the SLC-WT mice (bars: 200 µm, 100 µm and 50 µm) (pc: pericentral, pc: periportal). (B) Quantification of the fat red-stained liver sections from the male SLC-WT (n=10) and SLC-KO (n=7) mice. (C) Comparison of the liver/body ratio. Source files of all data used for the quantitative analysis are available in the Figure 2—source data 1. https://doi.org/10.7554/eLife.13308.006 Figure 2—source data 1 Source data of quantification of the fat red-stained liver sections from the male SLC-WT and SLC-KO mice (Figure 2B) and comparison of the liver/body ratio (Figure 2C). https://doi.org/10.7554/eLife.13308.007 Download elife-13308-fig2-data1-v1.docx Table 1 Serum concentrations of glucose, insulin and enzyme activities of ALAT, ASAT and GLDH in SLC mice. https://doi.org/10.7554/eLife.13308.011 parameterSLC-WTnSLC-KOninsulin (pmol/l)80.06 ± 9.9813110.88 ± 31.317glucose (mmol/dl)10.16 ± 1.8057.68 ± 1.026ALAT (µkat/l)1.02 ± 0.3250.81 ± 0.035ASAT (µkat/l)2.95 ± 0.9052.50 ± 0.415GLDH (µkat/l)0.42 ± 0.1350.28 ± 0.055 In addition to these results in vivo, the inhibition of Hh signaling in cultured mouse and human hepatocytes using the SMO inhibitor cyclopamine (Hovhannisyan et al., 2009) also resulted in marked steatosis within 48 to 72 hr (Figure 2—figure supplement 3). Alterations in downstream signaling of the Hh Pathway due to the deletion of Smo To study the signaling cascade after the deletion of Smo, we analyzed the alterations of Hh-related genes in isolated hepatocytes. Regarding Hh ligands, we could observe a down-regulation of Ihh and Shh, consistent with a decrease in pathway activity due to the deletion of Smo (Figure 3A). It is worth noting that Ihh is the most abundantly expressed ligand in hepatocytes, whereas Shh is hard to detect and Dhh was not measureable. Regarding the receptors, we did not observe significant changes in the Ptch1, Ptch2 and Hhip (Hedgehog-interacting protein) transcripts, which was also true for Fu (Fused) and Sufu (Suppressor of fused) (Figure 3B,C). However, the Gli1 and Gli3 mRNAs were significantly decreased in the SLC-KO mice, while the Gli2 mRNA remained unchanged (Figure 3D). To visualize the amount and the distribution of the GLI1 and GLI3 protein in the liver parenchyma of the SLC mice, we performed immunohistological stainings of GLI1 and GLI3 as well. (Figure 3E,F). The results clearly demonstrate that GLI1 and GLI3, are well detectable in hepatocyte nuclei in SLC-WT mice, but are absent in nuclei of SLC-KO hepatocytes (white arrows). In SLC-KO livers, these TFs are only present in non-parenchymal cells, e.g. bile duct epithelial cells (Figure 3E,F, yellow arrowheads). These results were confirmed by analyses of the GLI3 protein content by western blotting in isolated hepatocytes from SLC-WT and SLC-KO mice. The results clearly show that the amount of GLI3/A (full length activator protein) was significantly reduced in SLC-KO hepatocytes (Figure 3—figure supplement 1A,B). In order to find out whether these distinct alterations of the GLI factors are characteristic for steatotic livers, we measured the expression signature of the Gli TFs in isolated hepatocytes from melanocortin-4-receptor-deficient mice (MC4R) and Lepob/ob mice, which are characterized by massive steatosis as a result of over-nutrition (Sandrock et al., 2009; Itoh et al., 2011; Trak-Smayra et al., 2011). In both mouse models, the expression of Gli1 and Gli3 was obviously reduced, whereas Gli2 expression either did not change (MC4R-KO mice) or was increased (ob/ob mice) (Figure 3—figure supplement 2A,B). Furthermore we could detect the same transcriptional changes of the Gli factors in human patients with clinical relevant steatosis compared to non-steatotic patients (Figure 3—figure supplement 2C). Figure 3 with 2 supplements see all Download asset Open asset Expression of genes and proteins related to Hh signaling in SLC mice. (A–D) qRT-PCR data from isolated hepatocytes from the male SLC-WT (n=6–10) and the SLC-KO (n=6–10) mice illustrating the expression of (A) the ligands Ihh, Shh and Dhh (n.d.: not detectable); (B) the ligand binding molecules Ptch1, Ptch2 and Hhip1; (C) the downstream genes Fu and Sufu and (D) the TFs Gli1, Gli2 and Gli3 of the Hh signaling pathway. (E–F) Immunohistochemistry of liver sections from male SLC-WT and SLC-KO mice of (E) GLI1 and (F) GLI3. Labeled hepatocyte nuclei for both Gli factors are seen only in WT, but not in KO animals (white arrows). Staining in non-parenchymal cells, e. g. bile ductular cells (yellow arrowheads) is not affected by the knockout. Scale bars: 200 μm; 100 μm and 50 μm. Source files of all data used for the quantitative analysis are available in the Figure 3—source data 1. https://doi.org/10.7554/eLife.13308.012 Figure 3—source data 1 Source data of the expression of genes related to Hh signaling in SLC mice (Figure 3A–D). https://doi.org/10.7554/eLife.13308.013 Download elife-13308-fig3-data1-v1.docx Gene set enrichment analysis of global gene expression To get an impression of the global changes in gene expression in response to the deletion of Smo, microarray studies were performed. In order to account for possible inter-individual variations of gene expression, total RNA was prepared from freshly isolated hepatocytes of four pairs of SLC-KO and SLC-WT mice. At a cut-off level of 1.5-fold, 179 genes were up-regulated in Chip Arrays and 106 genes were down-regulated. Gene set enrichment analysis (GSEA) using ClueGO revealed highly significant changes in a number of metabolic functions, particularly those involved in lipid metabolism (Figure 4A) (Figure 4—source data 1—data). For instance, the GO term ‘lipid metabolic process’ showed pronounced up-regulation of 30 genes (p-value of 7.70E-11) many of which (e.g. Ppara, Pparg, Srebf1, Aacs, Elovl6 and Fas) were verified by qRT-PCR. In addition, the GSEA revealed that many regulated genes in the SLC-KO mice (e.g. Cd36, Avpr1a, Ttc23, Lifr andFabp2) belong to the top 50 ones described to be mostly correlated with elevated hepatic triglyceride level among 100 unique inbred mouse strains (Hui et al., 2015). The GO term ‘metabolic process’ showed pronounced up-regulation of 113 genes reaching a p-value of 1.24E-9 indicating that genes from other metabolic processes also responded to the Smo knockout. The GO terms ‘organic acid metabolic process’ and ‘steroid biosynthetic process’ showed up-regulation of several genes with p-values of 1.61E-6 and 6.09E-5, respectively. Furthermore, genes belonging to the GO term ‘response to hormone stimulus’ such as Lipin1 and Ramp2 were found to be up-regulated. Among down-regulated genes those involved in ‘activation of protein kinase activity’ and ‘microtubule-based processes’ prevailed with p-values of 1.59E-3 and 1.62E-3, respectively. Figure 4 Download asset Open asset Gen and protein expression of hepatic TFs involved in lipid metabolism in SLC mice. (A) Volcano blot visualizing differentially expressed genes in male SLC-KO mice detected by Affymetrix microarray analysis (n=4). All colored dots (blue and magenta) indicate an expression fold change equal or higher than 1.5; magenta: central genes of lipid metabolism; blue: other regulated genes. (B) qRT-PCR of Chrebp, Srebf1, Srebf2, Ppara, Pparb/d and Pparg from hepatocytes of male SLC-WT (n=6–13) and SLC-KO (n=6–13) mice. (C–D) Immunohistochemistry in liver sections from male SLC-WT and SLC-KO mice. (C) SREBP1 is strongly induced and shows a higher incidence of nuclear staining in pericentral hepatocytes of SLC-KO mice. (D) PPARG shows a much higher incidence in hepatocyte nuclei and a slight cytoplasmic increase in pericentral hepatocytes of SLC-KO mice. Scale bars: 200 μm and 100 μm. Source files of all data used for the quantitative analysis are available in the Figure 4—source data 2. https://doi.org/10.7554/eLife.13308.016 Figure 4—source data 1 Gene set enrichment analysis of isolated hepatocytes from SLC-WT and SLC-KO mice. https://doi.org/10.7554/eLife.13308.017 Download elife-13308-fig4-data1-v1.docx Figure 4—source data 2 Source data of gene expression of hepatic TFs involved in lipid metabolism in SLC mice (Figure 4B). https://doi.org/10.7554/eLife.13308.018 Download elife-13308-fig4-data2-v1.docx Smo deficiency up-regulates key lipogenic transcription factors and enzymes QRT-PCR was used to confirm major results of the microarray analysis and to detect additional regulated genes with higher sensitivity and accuracy. As shown in Figure 4B, the significant up-regulation of several TFs involved in the regulation of lipid and carbohydrate metabolism was detected including Chrebp (Carbohydrate-responsive element-binding protein), Srebf1, Srebf2 (Sterol regulatory element binding transcription factor 1/2), Ppara and Pparg (Peroxisome proliferator activated receptor alpha/gamma) in SLC-KO mice. To confirm this data on protein level, immunohistochemical staining was performed for the expression of SREBP1 and PPARG in liver sections of SLC-WT and SLC-KO mice. The SREBP1 protein showed a slightly pericentral and rare nuclear preference in liver parenchyma from SLC-WT mice and was very strongly enhanced in the pericentral zone of livers from SLC-KO mice. In particular, the frequency of nuclear staining in hepatocytes was much higher in these mice compared to control mice (Figure 4C). Likewise, PPARG protein was much more frequent in hepatocyte nuclei of SLC-KO mice than in SLC-WT mice. Also with PPARG heterogeneous distribution of the protein was obvious (Figure 4D). Additionally, the gene expression of the anti-adipogenic TF Gata4 (Patankar et al., 2012) was significantly reduced, whereas Gata6 remained unchanged (Figure 5A). The Nfyb (Nuclear transcription factor-Y beta) mRNA was also markedly reduced, while the Nfyg (Nuclear transcription factor-Y gamma) and Lxra (Liver X receptor alpha) mRNAs remained unaffected. The expression of Foxa1 and Foxa2 (Forkhead box A1/2), which are known to mediate the effects of insulin on lipid metabolism (Wolfrum et al., 2004), did not change (Figure 5B). Interestingly, the expression of Nr1d2 (Rev-ErbA beta) was strongly down-regulated, while that of Nr1d1 (Rev-ErbA alpha) remained unchanged (Figure 5B). Figure 5 Download asset Open asset Expression of the hepatic TFs and enzymes involved in lipid metabolism in SLC mice. (A–D) qRT-PCR data from hepatocytes from the male SLC-WT (n=6–13) and SLC-KO (n=6–13) mice. (A) Gata4, Gata6, Nfyb, Nfyg and Lxra. (B) Foxa1, Foxa2, Nr1d1 and Nr1d2. (C) Acaca, Acacb, Fasn, Gpam, Elovl6 and Elovl3. (D) Aacs, Hmgcr, Lss and Pnpla3. (E) The PPI (protein-protein interaction) network obtained from the STRINGv10 database using steatosis- and Hh-signaling-related genes as the query. The colored lines indicate co-expression (black), experimental data (purple), database scan (blue) and published scientific abstracts (green). Source files of all data used for the quantitative analysis are available in the Figure 5—source data 1. https://doi.org/10.7554/eLife.13308.019 Figure 5—source data 1 Source data of expression of the hepatic TFs and enzymes involved in lipid metabolism in SLC mice (Figure 5A–D). https://doi.org/10.7554/eLife.13308.020 Download elife-13308-fig5-data1-v1.docx Of the key lipogenic enzymes, we found increased expression of those involved in the biosynthesis of fatty acids and triglycerides, including Acaca (Acetyl-Coenzyme A carboxylase alpha), Fasn (fatty acid synthase), Elovl6 (Elongation of long chain fatty acids) and Gpam (Glycerol-3-phosphate acyltransferase) (Figure 5C). In contrast, the expression of Acacb was not changed, and Elovl3 expression was significantly decreased (Figure 5C). The transcripts for enzymes involved in cholesterol biosynthesis, such as Aacs (acetoacetyl-CoA synthetase), Hmgcr (3-hydroxy-3-methylglutaryl-Coenzyme A reductase), and Lss (Lanosterol synthase), were also increased (Figure 5D). These findings suggest a coordinated response of genes favoring fatty acid, triglyceride and cholesterol biosynthesis. Intriguingly, the newly discovered NAFLD-associated gene Pnpla3 (Adiponutrin) (Chow et al., 2014; Smagris et al., 2015) was also dramatically increased (Figure 5D). Using this data, the results of our microarray analysis, and published databases, we created a hypothetical protein-protein interaction network (PPI) of the studied steatosis- and Hh-signaling-related genes with the STRINGv10 software (Figure 5E). This network provides an overview of the studied proteins and confirms the observed diverse and highly complex connections of the TFs and enzymes related to lipid metabolism and Hh signaling. PNPLA3 expression is connected to SREBF1 activity, and the PPAR family is linked to the FOXA and GLI TFs. The lipogenic enzymes (e.g. FASN, ACACA/B, ELOVL6, and GPAM) are closely connected to the TF SREBF1 and to each other as part of the fatty acid and triglyceride biosynthetic pathways. Coordinated response of enzymes and metabolic pathways Immunohistochemical analyses of FASN were performed to confirm the differences in expression at the protein level. The stronger staining not only confirmed the up-regulation of FASN in the SLC-KO mice, but, surprisingly, revealed a shift from the known pericentral (Gebhardt, 1992; Postic and Girard, 2008) to the periportal localization in the SLC-WT mice, matching the preferential site of lipid accumulation (Figure 6A). These results suggest that in addition to Wnt/beta-catenin signaling (Benhamouche et al., 2006; Gebhardt and Hovhannisyan, 2010), Hh signaling is required to maintain the normal zonation of the liver parenchyma, at least with respect to lipid metabolism. Furthermore, we measured the activity of several lipid metabolism pathways in cultured hepatocytes. Fatty acid biosynthesis, as determined by the incorporation of [14C]-labelled acetate, was almost doubled in the hepatocytes from the SLC-KO mice (Figure 6B). However, cholesterol biosynthesis from labelled acetate was not changed in vitro (Figure 6B), corresponding to the unchanged serum cholesterol levels in vivo (Figure 1—figure supplement 1A). Interestingly, the hepatocytes from the SLC-KO mice showed a significantly higher rate of fatty acid synthesis from [14C(U)]-glucose in the presence of 10 mM glucose and 0.1 µM insulin (Figure 6C), suggesting increased channeling of the high concentrations of glucose into fatty acid biosynthesis. The fact that neither the glycogen content (Figure 6D), nor basal or stimulated glycolysis as determined by the conversion of [14C(U)]-glucose to lactate (Figure 6E) were altered in the hepatocytes from the re-fed SLC-KO mice, suggests that there is no shortage of potential fuel for glycolytic acetyl-CoA formation. Figure 6 with 1 supplement see all Download asset Open asset Expression of genes and proteins involved in lipid and mitochondrial energy metabolism in SLC mice. (A) Immunohistochemistry of FASN in liver sections of male SLC-WT and SLC-KO mice (pp: periportal, pc: pericentral). (B–H) Measurements in freshly isolated hepatocytes from male SLC-WT and SLC-KO mice (B) Determination of fatty acid and cholesterol biosynthesis. (C) Conversion of [14C(U)]-glucose to fatty acids. (D) Glycogen content. (E) Determination of glycolysis. (F) Electron microscopy of liver tissue. (G) ATP content. (H–I) qRT-PCR data from the male SLC-WT (n=6) and the SLC-KO (n=6) mice: (H) Acox1, Cpt1a, Cpt2 and Acadvl, (I) Slc25a1, Slc25a2 and Slc25a20. (J) Serum concentrations of ketone bodies. Eight to ten SLC mice were used in each of the experiments depicted in (B, C,
As part of the neuronal cytoskeleton, neurofilaments are involved in maintaining cellular integrity. In the setting of ischemic stroke, the affection of the neurofilament network is considered to mediate the transition towards long-lasting tissue damage. Although peripheral levels of distinct neurofilament subunits are shown to correlate with the clinically observed severity of cerebral ischemia, neurofilaments have so far not been considered for neuroprotective approaches. Therefore, the present study systematically addresses ischemia-induced alterations of the neurofilament light (NF-L), medium (NF-M), and heavy (NF-H) subunits as well as of α-internexin (INA). For this purpose, we applied a multi-parametric approach including immunofluorescence labeling, western blotting, qRT-PCR and electron microscopy. Analyses comprised ischemia-affected tissue from three stroke models of middle cerebral artery occlusion (MCAO), including approaches of filament-based MCAO in mice, thromboembolic MCAO in rats, and electrosurgical MCAO in sheep, as well as human autoptic stroke tissue. As indicated by altered immunosignals, impairment of neurofilament subunits was consistently observed throughout the applied stroke models and in human tissue. Thereby, altered NF-L immunoreactivity was also found to reach penumbral areas, while protein analysis revealed consistent reductions for NF-L and INA in the ischemia-affected neocortex in mice. At the mRNA level, the ischemic neocortex and striatum exhibited reduced expressions of NF-L- and NF-H-associated genes, whereas an upregulation for Ina appeared in the striatum. Further, multiple fluorescence labeling of neurofilament proteins revealed spheroid and bead-like structural alterations in human and rodent tissue, correlating with a cellular edema and lost cytoskeletal order at the ultrastructural level. Thus, the consistent ischemia-induced affection of neurofilament subunits in animals and human tissue, as well as the involvement of potentially salvageable tissue qualify neurofilaments as promising targets for neuroprotective strategies. During ischemia formation, such approaches may focus on the maintenance of neurofilament integrity, and appear applicable as co-treatment to modern recanalizing strategies.
Experimental stroke studies yielded insights into single reactions of the neurovascular unit (NVU) and associated extracellular matrix (ECM). However, the extent of simultaneous processes caused by ischemia and their underlying transcriptional changes are still poorly understood. Strictly following the NVU and ECM concept, this study explored transcriptional responses of cellular and non-cellular components as well as their morphological characteristics following ischemia. Mice were subjected to 4 or 24 h of unilateral middle cerebral artery occlusion. In the neocortex and the striatum, cytoskeletal and glial elements as well as blood-brain barrier and ECM components were analyzed using real-time PCR. Western blot analyses allowed characterization of protein levels and multiple immunofluorescence labeling enabled morphological assessment. Out of 37 genes analyzed, the majority exhibited decreased mRNA levels in ischemic areas, while changes occurred as early as 4 h after ischemia. Down-regulated mRNA levels were predominantly localized in the neocortex, such as the structural elements α-catenin 2, N-cadherin, β-catenin 1, and βIII-tubulin, consistently decreasing 4 and 24 h after ischemia. However, a few genes, e.g., claudin-5 and Pcam1, exhibited increased mRNA levels after ischemia. For several components such as βIII-tubulin, N-cadherin, and β-catenin 1, matching transcriptional and immunofluorescence signals were obtained, whereas a few markers including neurofilaments exhibited opposite directions. In conclusion, the variety in gene regulation emphasizes the complexity of interactions within the ischemia-affected NVU and ECM. These data might help to focus future research on a set of highly sensitive elements, which might prospectively facilitate neuroprotective strategies beyond the traditional single target perspective.
Ischemic stroke not only affects neurons, but also glial and vascular elements. The development of novel neuroprotective strategies thus requires an improved pathophysiological understanding of ischemia-affected cell types that comprise the 'neurovascular unit' (NVU). To explore spatiotemporal alterations of oligodendrocytes, astrocytes and neurons after experimental ischemic stroke, we applied a permanent middle cerebral artery occlusion model in mice for 4 and 24 h. Using fluorescence microscopy, the oligodendrocyte marker 2',3'-cyclic nucleotide phosphodiesterase (CNP), the neuronal neurofilament light chain (NF-L) and the astroglial aquaporin-4 (AQP4) were analyzed in regional relation to one another. Immunofluorescence intensities of CNP and NF-L were simultaneously increased in the ischemic neocortex and striatum. AQP4 immunoreactivity was decreased in the ischemic striatum, which represents the initial and potentially strongest affected site of infarction. The more distant ischemic neocortex and infarct border zones exhibited areas with alternately increased or decreased AQP4 immunoreactivity, leading to an increase of fluorescence intensity in total. Further, deformed CNP-immunopositive processes were found around axonal spheroids, indicating a combined affection of oligodendrocytes and neurons due to ischemia. Importantly, altered AQP4 immunosignals were not limited to the ischemic core, but were also detectable in penumbral areas. This applies for CNP and NF-L also, since altered immunosignals of all three markers coincided regionally at both time points. In conclusion, the present study provides evidence for a simultaneous affection of oligodendrocytes, astrocytes and neurons after experimental focal cerebral ischemia. Consequently, CNP, AQP4 and NF-L immunofluorescence alterations can be utilized to identify ischemia-affected tissue. The simultaneity of the described alterations further strengthens the concept of interdependent NVU components and distinguishes NF-L, CNP and AQP4 as highly ischemia-sensitive elements. Consequently, future therapeutic approaches might influence stroke evolution via strategies simultaneously addressing both neuronal and glial functions.
The Hedgehog signalling-driven Gli transcription factors in hepatocytes form a regulatory network identified by a fuzzy-logic modelling approach. The network explains dynamic features important for hepatocyte function and fate.
Beim hepatozellulären Karzinom (HCC), handelt es sich um einen hochmalignen Tumor aus entarteten Hepatozyten, der mit mehr als 1 Mio. Neuerkrankungen pro Jahr zu einem der häufigsten Krebsarten weltweit zählt. Als Ursache gelten chronische Hepatitis Infektionen, Leberzirrhose und das Einwirken von Kofaktoren, wie Alkoholabusus oder eine Aflatoxin-Exposition. Genetische Veränderungen in morphogenen Signalkaskaden, wie dem Wnt/β-Catenin und dem Hedgehog Signalweg, treten häufig bei Patienten mit einem HCC auf. So wird bei 30% der Erkrankten eine erhöhte CyclinD Expression bzw. eine Veränderungen in der β-Catenin Transkription diagnostiziert. Ebenfalls einen Anstieg in seiner Aktivität zeigt der Hedgehog Signalweg. Mehrere Studien an verschiedenen Karzinomarten haben gezeigt, dass es unterschiedliche Crosstalks und Feedback-Loops zwischen diesen beiden Signalkaskaden gibt. Auch in primären murinen Hepatozyten lassen sich diverse Verknüpfungen aufzeigen. Wir stellten uns die Frage, ob Hepatozyten sowie Hepatoblastom- und Hepatomzellen einen Crosstalk zwischen Hedgehog und Wnt/β-Catenin Signaling aufweisen.
Hintergrund: Die morphogenen Signalwege Wnt/-Catenin und Hh (Hedgehog) spielen in der Embryogenese, Organogenese und Kanzerogenese eine wichtige Rolle. Neuere Arbeiten konnten zeigen, dass beide Signalwege in adulten Geweben ebenfalls wichtige Funktionen des Metabolismus steuern und sich dabei gegenseitig beeinflussen.
As part of the extracellular matrix (ECM), perineuronal nets (PNs) are polyanionic, chondroitin sulfate proteoglycan (CSPG)-rich coatings of certain neurons, known to be affected in various neural diseases. Although these structures are considered as important parts of the neurovascular unit (NVU), their role during evolution of acute ischemic stroke and subsequent tissue damage is poorly understood and only a few preclinical studies analyzed PNs after acute ischemic stroke. By employing three models of experimental focal cerebral ischemia, this study was focused on histopathological alterations of PNs and concomitant vascular, glial and neuronal changes according to the NVU concept. We analyzed brain tissues obtained 1 day after ischemia onset from: (a) mice after filament-based permanent middle cerebral artery occlusion (pMCAO); (b) rats subjected to thromboembolic MACO; and (c) sheep at 14 days after electrosurgically induced focal cerebral ischemia. Multiple fluorescence labeling was applied to explore simultaneous alterations of NVU and ECM. Serial mouse sections labeled with the net marker Wisteria floribunda agglutinin (WFA) displayed largely decomposed and nearly erased PNs in infarcted neocortical areas that were demarcated by up-regulated immunoreactivity for vascular collagen IV (Coll IV). Subsequent semi-quantitative analyses in mice confirmed significantly decreased WFA-staining along the ischemic border zone and a relative decrease in the directly ischemia-affected neocortex. Triple fluorescence labeling throughout the three animal models revealed up-regulated Coll IV and decomposed PNs accompanied by activated astroglia and altered immunoreactivity for parvalbumin, a calcium-binding protein in fast-firing GABAergic neurons which are predominantly surrounded by neocortical PNs. Furthermore, ischemic neocortical areas in rodents simultaneously displayed less intense staining of WFA, aggrecan, the net components neurocan, versican and the cartilage link protein (CRTL) as well as markers in net-bearing neurons such as the potassium channel subunit Kv3.1b and neuronal nuclei (NeuN). In summary, theconsistent observations based on three different stroke models confirmed that PNs are highly sensitive constituents of the NVU along with impaired associated GABAergic neurons. These results suggest that PNs could be promising targets of future stroke treatment, and further studies should address their reorganization and plasticity in both stabilizing the acute stroke as well as supportive effects during the chronic phase of stroke.
