Palmitate and Fructose Interact to Induce Human Hepatocytes to Produce Pro-Fibrotic Transcriptional Responses in Hepatic Stellate Cells Exposed to Conditioned Media
Ignazio S. Pirasa Glenn S. Gerhardb Johanna K. DiStefanoa
aTranslational Genomics Research Institute, Phoenix, AZ, USA, bLewis Katz School of Medicine at Temple University, Philadelphia, PA, USA
Key Words
Hepatocyte • Fructose • Saturated fat • Transcriptomics • Hepatic steatosis • Hepatic stellate cell • Long noncoding RNA • Nonalcoholic fatty liver disease
Abstract
Background/Aims: Excessive consumption of dietary fat and sugar is associated with an elevated risk of nonalcoholic fatty liver disease (NAFLD). Hepatocytes exposed to saturated fat or sugar exert effects on nearby hepatic stellate cells (HSCs); however, the mechanisms by which this occurs are poorly understood. We sought to determine whether paracrine effects of hepatocytes exposed to palmitate and fructose produced profibrotic transcriptional responses in HSCs. Methods: We performed expression profiling of mRNA and lncRNA from HSCs treated with conditioned media (CM) from human hepatocytes treated with palmitate (P), fructose (F), or both (PF). Results: In HSCs exposed to CM from palmitate-treated hepatocytes, we identified 374 mRNAs and 607 lncRNAs showing significant differential expression (log2 foldchange ≥ |1|; FDR ≤0.05) compared to control cells. In HSCs exposed to CM from PF-treated hepatocytes, the number of differentially expressed genes was much higher (1198 mRNAs and 3348 lncRNAs); however, CM from fructose-treated hepatocytes elicited no significant changes in gene expression. Pathway analysis of differentially expressed genes showed enrichment for hepatic fibrosis and hepatic stellate cell activation in P- (FDR =1.30E-04) and PF-(FDR =9.24E-06) groups. We observed 71 lncRNA/nearby mRNA pairs showing differential expression under PF conditions. There were 90 mRNAs and 264 lncRNAs strongly correlated between the PF group and differentially expressed transcripts from a comparison of activated and quiescent HSCs, suggesting that some of the transcriptomic changes occurring in response to PF overlap with HSC activation. Conclusion: The results reported here have implications for dietary modifications in the prevention and treatment of NAFLD.
Nonalcoholic fatty liver disease (NAFLD) arises from excessive triacylglycerol deposition in hepatocytes and encompasses a histological spectrum with simple steatosis at one end and nonalcoholic steatohepatitis (NASH), often accompanied by fibrosis, at the other [1, 2]. Hepatic fat accumulation results from increased rates of de novo lipogenesis and hepatic fatty acid uptake or reduced levels of fatty acid oxidation and distribution of lipids from the liver to the circulation [3]. In the United States, NASH is the major cause of chronic liver disease and is poised to become the most common indication for liver transplantation [4].
Type 2 diabetes (T2D) and obesity increase the risk of developing NAFLD [5]. The close association among T2D, obesity, and NAFLD has likely antecedents in higher levels of overall daily energy intake and increased consumption of processed foods, which are major sources of sugar and saturated fat in the modern diet [6]. Indeed, high intake of saturated fat and cholesterol has been associated with hepatic steatosis [7-12]. In addition, dietary fructose has been associated with depletion of adenosine triphosphate, hepatic de novo lipogenesis, post-prandial hypertriglyceridemia [13, 14], and the development of NAFLD in the absence of traditional risk factors [15, 16]. Dietary fructose consumption has also been linked with severe fibrosis in NASH patients [17]. A number of different diet-induced animal models have been generated to investigate the effects of specific macronutrients on liver metabolism and NAFLD pathogenesis [18]. Many studies suggest a synergistic effect of high fat and fructose feeding on biological and histological parameters of NAFLD [19].
A number of studies have investigated changes in intercellular metabolism resulting from exposure to these nutrients in primary or immortalized hepatocytes and hepatic stellate cells (HSCs) [20-28]. HSCs are quiescent under physiological conditions, but are activated to a myofibroblastic state in response to injurious stimuli, whereupon they begin to secrete cytokines and components of the extracellular matrix (ECM). Under chronic conditions of hepatic inflammation, ECM components accumulate and eventually lead to hepatic scarring [29]. Conditioned media from primary human hepatocytes treated with palmitate was able to activate HSCs, decrease HSC apoptosis, and elicit changes in expression of pro-inflammatory and pro-fibrogenic genes [27]. Similarly, conditioned media from HepG2 cells treated with palmitate significantly increased alpha-SMA expression in immortalized human HSCs (LX-2 cells) [20]. Treatment of LX-2 cells with exosomes from palmitate-treated Huh7 cells resulted in significant changes in expression of genes related to fibrosis [23]. HepG2 cells treated with a combination of fructose, glucose, and fatty acids (palmitate and oleate) showed increased levels of triacylglycerols, total cholesterol, and inflammatory cytokines, as well as upregulated expression of genes involved in carbohydrate and lipid metabolism [22], although fructose alone did not affect expression of lipogenic genes [28]. In mice, however, fructose-feeding induced expression of lipogenic genes, while inhibiting expression of genes involved in fatty acid oxidation [19]. In humans, fructose consumption for ten weeks led to increased hepatic de novo lipogenesis and 23-hour postprandial hypertriglyceridemia [14], suggesting that fructose may contribute to the development of NAFLD via the circulating lipid pool [8]. These data are consistent with recent evidence indicating that transcriptomic changes accompanying the early development of NAFLD occur predominantly in hepatocytes [24]. To date, however, the impact of palmitate- and fructose-treated human hepatocytes on transcriptomic changes in HSCs has not yet been characterized.
We hypothesized that fructose may amplify the effects of saturated fat by contributing to the biosynthesis of intracellular palmitate, leading to the enhanced expression of profibrotic genes. In this study, we compared gene expression patterns in HSCs treated with conditioned media from primary human hepatocytes exposed to palmitate, fructose, or a combination of the two. We also compared a subset of RNAs showing differential expression in HSCs [30] treated with conditioned media from palmitate + fructose-treated hepatocytes with transcripts that were dysregulated in HSCs undergoing myofibroblastic transactivation, suggesting that these dietary macronutrients may contribute to fibrosis through HSC activation.
Materials and Methods
Cell treatments
Primary human hepatocytes (Thermo Fisher Scientific; Waltham, MA) were cultured in 500 μL William's E Medium supplemented with primary Hepatocyte Maintenance Supplement on collagen-coated 6-well plates. Culture medium was replaced the first day after thawing. LX-2 cells (Merck Millipore; Billerica, MA), an immortalized hepatic stellate cell line [31], were cultured in T-75 flasks containing 12 ml Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 2% fetal bovine serum (FBS). Culture medium was replaced the first day after seeding, and then every 72 hours. Cell line authentication was performed using short tandem repeat (STR) profiling (Cell Line Genetics; Madison, Wi), which confirmed the presence of a single cell line and alleles matching the known DNA fingerprint [32].
PHH were treated with 1 mM palmitate (P), 10 mM fructose (F), or a combination of 1 mM palmitate + 10 mM fructose (PF) for 48 hours. This concentration of palmitate was selected to approximate supraphysiological levels and promote triacylglycerol storage and lipotoxity [33, 34]. We assayed a range of concentrations for fructose treatment (5-25 mM), and found that 10 mM fructose optimally induced expression of lipogenic genes, similar to other findings [28]. A schematic overview of the experimental design is shown in Fig. 1A. Palmitate (Sigma-Aldrich; St. Louis, MO) was conjugated to bovine serum albumin (BSA; Omega Scientific; Tarzana, CA) at a final concentration of 1 mM palmitate and 1% BSA by heating at 37°C for 30 minutes prior to cell treatments [35]. Fructose (Sigma-Aldrich) was dissolved at a concentration of 300 mM in sterile water. Cells were serum-starved overnight, and then treated with either 1% BSA or 1 mM palmitate, in the presence or absence of 10 mM fructose for 24 and 48 hours. Oil Red O staining was performed to confirm uptake of palmitate (Supplementary Fig. S1 – for all supplementary material see www.cellphysiolbiochem.com) [36].
LX-2 cells were seeded at 1 x 104 cell/well on 24-well plates coated with 85 mL of (1mg/mL) Matrigel Growth Factor Reduced (GFR) basement membrane matrix (Corning Inc; Corning, NY) and cultured for 72 hours at 37˚C to induce a state resembling biological quiescence [37]. Media from treated PHH was collected and centrifuged twice at 1700 x g for 15 minutes to remove cell debris and aggregates. Cells were then treated with a cocktail containing 250 μL conditioned media from the different hepatocyte treatment groups and 250 μL LX-2 growth media and cultured for 48 hours.
Array hybridization and data analysis
Total RNA was extracted from three biological replicates for each treatment group using the RNeasy Mini Kit (Qiagen). All RNA preparations were diluted to 5 μg/μl. Samples were amplified and transcribed into fluorescent cRNA using a random priming method (Arraystar Flash RNA Labeling Kit [Arraystar; Rockville, MD]). Labeled cRNAs were purified using the RNeasy mini Kit (Qiagen) and the concentration and specific activity of the labeled cRNAs were measured using the NanoDrop ND-1000. One microgram of each labeled cRNA was fragmented and heated according to the manufacturer’s protocol, and then added to the microarray slides. Slides were incubated for 17 hours at 65°C in an Agilent Hybridization Oven. The microarray used in this study was the Human LncRNA Microarray V4.0 (Arraystar), which allows detection of 40,173 lncRNAs and 20,730 protein-coding mRNAs. After washing and fixing, arrays were scanned using the Agilent DNA Microarray Scanner. Array images were analyzed using Agilent Feature Extraction software (version 11.0.1.1). Low quality mRNAs and lncRNAs were removed from further analysis, including those with at least three out of 15 samples with flags in “Present” (P) or “Marginal” (M) according to GeneSpring GX v12.1. Raw signal intensities for informative features were background-corrected and quantile-normalized using the R-package limma [38]. Quality controls (QC) were conducted using the R-package arrayQualityMetrics [39], inspecting heatmaps of inter-array expression distances, boxplots of expression values, and MAplots, which includes the log-intensity ratios (M-values) versus log-intensity averages (A-values). Differential expression analysis between treatment groups was conducted using a linear model as implemented in limma. P-values were adjusted for multiple testing using the False Discovery Rate method (FDR) [40]. Transcripts were considered to be differentially expressed if the FDR was <0.05 and the log2 fold change was ≥|1|.
Pathway analysis
To interpret the biological significance of the microarray data, we uploaded the list of differentially expressed genes containing gene identifiers, FDR values, and log2FC into the Ingenuity Pathway Analysis (IPA) software (Qiagen). The “core analysis” function was applied to identify canonical pathways, upstream regulators, and gene networks relevant to the differentially expressed transcripts. The significance of the overrepresentation of differentially expressed genes in a pathway was assessed using a right-tailed Fisher’s exact test and adjusting the p-values with the Benjamini-Hochberg method.
Co-expression analysis of lncRNAs and associated mRNAs
We mapped mRNAs located within 10 kb of differentially expressed lncRNAs, retaining the mRNAs/lncRNAs pairs showing statistically significant evidence for differential expression. We measured the relationship in expression between lncRNAs and associated mRNAs using Pearson’s correlation and post hoc adjustment using the Benjamini-Hochberg method. Furthermore, we summarized the different classes of lncRNA/mRNA relationship (Bidirectional, Exon sense overlapping, Intron sense-overlapping, Intronic antisense, and Natural antisense) by concordant or discordant log2 fold change direction.
Comparison with differentially expressed genes in activated hepatic stellate cells
We previously characterized mRNA and lncRNA expression changes that occurred during myofibroblastic activation of hepatic stellate cells using the same array platform [30]. We selected the differentially expressed genes from that study using the same cutoff applied in the current study (log2 FC ≥|1| and FDR < 0.05). Then, we intersected by probe ID the obtained list of genes with the differentially expressed genes from the palmitate + fructose treatment. The relationship between the two experiments was conducted computing the Pearson’s correlation coefficient between the log2 fold changes.
Results
Changes in RNA expression in LX-2 cells in response to palmitate and fructose
We first analyzed the pooled data obtained from the microarray experiments. Following removal of RNAs that did not meet quality control measures, 88.9% of the RNAs on the array were detected in HSCs, corresponding to 19,837 mRNAs and 32,471 lncRNAs. The mean normalized intensity was greater in mRNAs (log2 normalized intensity = 9.592) compared to lncRNAs (log2 normalized intensity =6.617), consistent with previous findings [30]. Sixty-eight percent of the detected lncRNAs were intergenic; the remainder of the lncRNAs had a bidirectional (5%), exon sense-overlapping (2%), intron sense-overlapping (3%), intronic (12%), or natural antisense (10%) orientation (Supplementary Fig. S2).
To establish baseline levels of potential residual media effects from palmitate (P) and fructose (F) on RNA expression in LX-2 cells, we compared mRNA and lncRNA levels in HSCs exposed to conditioned media from untreated hepatocytes with those from cells treated with palmitate and fructose (PF)-containing media from the hepatocyte-free control group (Fig. 1A). We identified 360 (162 upregulated and 198 downregulated) and 1207 (290 upregulated and 917 downregulated) differentially expressed mRNAs and lncRNAs, respectively, with log2 fold-change ≥|1| and FDR <0.05 (Supplementary Fig. S3). To account for potential residual effects of palmitate and fructose not taken up by hepatocytes, we filtered all subsequent comparisons between treatment groups by this gene set. Filtering was conducted by probe ID to account for the presence of different transcript variants for the same gene.
We first assessed the effects of conditioned media from P-treated hepatocytes on transcriptomic changes in HSCs (Fig. 1B). We observed 374 differentially expressed mRNAs (176 upregulated and 198 downregulated) showing a log2fold change ≥|1| and FDR <0.05 in LX-2 cells treated with conditioned media from hepatocytes exposed to palmitate compared to the control treatment (CT) cells (Supplementary Table S1). Transcripts showing the strongest evidence for differential expression (based on adjusted p-value) are shown in Table 1. Of the most dysregulated genes in Table 1, only FNBP1L has been linked with NAFLD in humans [24], although PGF [41], SLC15B [42, 43], and WWP2 [44] have been associated with biological attributes of NAFLD in mouse and in vitro models.
We also observed 607 dysregulated lncRNAs: 459 upregulated and 148 downregulated molecules (Fig. 1B and Supplementary Table S2), compared to control conditions. Table 2 lists the lncRNAs showing the strongest evidence for differential expression between the two groups. Of note, the hypoxia upregulated 1 gene (HYOU1), associated with lncRNA RP11-110I1.6, is a known anti-apoptotic protein that plays a prominent role in hepatocyte survival [45].
In HSCs treated with conditioned media from hepatocytes exposed to 10 mM fructose, no statistically significant evidence (FDR < 0.05) for differential mRNA or lncRNA expression was observed, indicating that at this concentration, fructose alone does not produce major transcriptional changes in our model (Fig. 1C). Despite the lack of significant changes in gene expression in the fructose group, treating HSCs with conditioned media from hepatocytes exposed to PF led to significant changes in expression for 1198 mRNAs, including 557 upregulated and 641 downregulated genes (Fig. 1D and Supplementary Table S3). Those transcripts showing the strongest evidence for differential expression are shown in Table 2. Not surprisingly, we observed substantial overlap between the differentially expressed genes shown in Tables 1 and 3 (P and PF-treated cells, respectively). Seven mRNAs (MYO3D, CREBF, DPP9-AS1, ZMAT4, MAT2A, OR6M1, and FBO40) were unique to the gene list from the combined PF-treated cells. We also observed differential expression of 3348 lncRNAs (1450 upregulated and 1898 downregulated) showing a log2 fold change ≥|1| and FDR <0.05 (Supplementary Table S4). LncRNAs showing the strongest evidence for differential abundance are shown in Table 4. In contrast to the overlap in differentially expressed mRNAs between the P and PF groups, only two lncRNAs, XLOC_006774 and G028271, were shared between the gene lists in Tables 2 and 4.
Given the overlap between mRNAs showing the strongest evidence for differential expression in Tables 1 and 3, we next sought to directly compare the datasets from the P and PF treatments (Fig. 2). In this comparison, we observed an overlap of 326 differentially expressed mRNAs between datasets (Supplementary Table S5). Of the shared differentially expressed genes, 276 showed higher variation in the PF group compared to the P group. Comparison of the lncRNAs from the P and PF treatment groups showed an overlap of 573 differentially expressed transcripts between the two conditions, with 522 showing higher expression in the PF group relative to the P group. We then compared the PF group with the P and CT groups combined. Using this approach, we identified 92 upregulated and 221 downregulated mRNAs and 480 upregulated and 1204 downregulated lncRNAs (Fig. 3).
Enriched biological pathway analysis of differentially expressed genes
We used functional enrichment analysis to identify biological attributes associated with the differentially expressed genes in individual group comparisons. The most significant canonical pathways involved in the different treatment conditions are shown in Table 5. Of note, in the comparison between HSCs exposed to conditioned media from untreated hepatocytes with HSCs treated with PF-containing media without hepatocytes, we observed enrichment in several biological pathways relevant to NAFLD, including Hepatic Fibrosis/Hepatic Stellate Cell Activation, LXR/RXR activation, TREM1 Signaling, Inhibition of Matrix Metalloprotease, and Hepatic Fibrosis Signaling Pathway, suggesting that palmitate and fructose exert direct, i.e., not mediated by hepatocytes, effects on key fibrogenic pathways in HSCs. The most significant canonical pathway in the palmitate versus control treatment group was Fatty Acid Activation, while the most significant pathway in the PF group was Hepatic Fibrosis/Hepatic Stellate Cell Activation. Although the Hepatic Fibrosis/Hepatic Stellate Cell Activation pathway was enriched across all comparisons, the PF treatment had the highest ratio of genes (12.9%) and the strongest significance (FDR = 9.24E-06) compared to the other groups.
Functional enrichment analysis of the 326 differentially expressed genes shared between the P and PF treatment groups identified Fatty Acid Activation as the top canonical pathway (P=1.89E-05). Interestingly, pathway analysis for the 48 and 872 differentially expressed genes unique to the P and PF treatment groups, respectively, identified Hepatic Fibrosis/Hepatic Stellate Cell Activation as the top pathway for both gene sets suggesting that palmitate exposure drives the expression of this subset of genes.
Co-expression analysis of differentially expressed lncRNAs and mRNAs
We performed co-expression analysis of differentially expressed pairs of lncRNAs and nearby mRNAs to investigate the presence of potential regulatory mechanisms (Table 6). For the P versus CT comparison, we identified ten significant lncRNA/mRNA pairs. For the comparison of PF versus CT, we detected 71 significant mRNA/lncRNAs pairs (Supplementary Table S6); the ten most significant pairs are listed in Table 6. For the comparison of PF versus P + CT, we detected nine significant mRNA/lncRNAs pairs. The distribution of the transcriptional relationships in the PF dataset included 12 bidirectional, 1 exon sense overlapping, 7 intron sense overlapping, 27 intronic antisense, and 24 natural antisense lncRNAs (Supplementary Fig. S4); 56% of differentially expressed lncRNA/nearby mRNA pairs in this comparison showed expression changes in opposing directions.
Differentially expressed RNAs in the PF treatment group are correlated with dysregulated genes in activated hepatic stellate cells
We previously characterized mRNA and lncRNA expression changes that occurred during myofibroblastic activation of HSCs [30]. Because the most significant pathway in the PF group was Hepatic Fibrosis/Hepatic Stellate Cell Activation, we compared this dataset with the RNAs that we found to be dysregulated between quiescent and activated HSCs (Fig. 4). We identified 675 probes that were differentially expressed in both analyses (log2FC ≥|1|; FDR <0.05), corresponding to 301 mRNAs and 374 lncRNAs. A total of 480 probes showed a concordant log2FC, and of these, 343 upregulated transcripts were located in a cluster with highly correlated features (R = 0.745; P< 2.2E-16). Pathway analysis of these 343 genes identified a number of overlapping canonical pathways, many of which are relevant to hepatic stellate cell function and activation (Fig. 5).
Discussion
We attempted to model the potential interaction between hepatocytes and HSC in response to common macronutrients. Although hepatocytes are the most abundant cell type in the liver, comprising approximately 80% of the total mass, the liver is comprised of multiple cell types, including HSCs, Kupffer cells, and sinusoidal endothelial cells [46]. In the presence of chronic excessive fat accumulation, hepatocytes generate toxic lipid metabolites, which contribute to ballooning degeneration, cell injury, and cell death. Hepatocytes also participate in the initiation and progression of fibrosis through complex processes involving nearby cells, particularly HSCs [47], which are triggered by hepatocyte injury [48]. In response to hepatic injury, quiescent HSCs transdifferentiate to an activated myofibroblastic state and begin to secrete cytokines and other molecules [49]. We found that conditioned media from hepatocytes treated with palmitate and fructose elicit changes in transcriptional programs associated with hepatic fibrosis in hepatic stellate cells. We also observed a cluster of 343 differentially expressed genes in cells treated with palmitate + fructose that were strongly correlated with genes upregulated during hepatic stellate cell activation, suggesting that paracrine factors released by hepatocytes in response to saturated fat and sugar may stimulate hepatic stellate cell activation and subsequent fibrogenesis in the liver.
Our results are consistent with previous in vitro studies that have investigated changes in cellular metabolism resulting from exposure to palmitate, fructose, or a combination of the two in primary or immortalized hepatocytes and HSCs. Conditioned media from HepG2 cells treated with palmitate significantly increased alpha-SMA expression in LX-2 cells, although LX-2 cells directly treated with palmitate did not show the same response, suggesting that a secondary metabolite from palmitate-treated HepG2 cells was responsible for HSC activation [20]. Windemuller et al. [26] showed that increasing the fructose to glucose ratio in Huh7 cells corresponded with elevated triacylglycerol and cholesterol synthesis, but this effect was not observed with increasing concentrations of glucose. In contrast, HepG2 cells treated with fructose did not show altered expression of lipogenic genes [22], although a combination of fructose, glucose, and fatty acids (palmitate and oleate) not only increased levels of triacylglycerols, total cholesterol, and inflammatory cytokines, but also upregulated expression of genes involved in carbohydrate and lipid metabolism [28]. These data are consistent with recent evidence indicating that transcriptomic changes accompanying the early development of NAFLD occur predominantly in hepatocytes [24]. The results reported here add to these findings by characterizing the impact of palmitate- and fructose-treated human hepatocytes on transcriptomic changes in HSCs.
Interestingly, our results did not show statistically significant changes in either mRNA or lncRNA expression in HSCs treated with conditioned media from hepatocytes exposed to fructose alone. This could be due to the concentration of fructose used in this study, 10 mM, which may be too low to induce metabolic effects in hepatocytes that lead to the secretion of paracrine factors. Alternatively, treatment with fructose alone may not be sufficient to cause metabolic changes in liver cells. Some studies have shown that dietary fat produces stronger hepatic effects than added sugars [7, 8]. However, because fructose is known to stimulate de novo lipogenesis, and one of the major initial products of de novo lipogenesis is palmitate [7, 50, 51], it would be expected that fructose would exert metabolic effects similar to palmitate. We postulate that fructose, in the presence of palmitate, may amplify effects on gene expression by augmenting endogenous palmitate production in hepatocytes.
The roles of dietary saturated fat and added sugars in the pathogenesis of NAFLD is well recognized. Hepatic fat fraction has been associated with high intake of energy, total fat, and saturated fat in NAFLD [7] and NASH patients [52]. Saturated fat was also positively correlated with hepatic lipid content (β=0.45; p=0.03) and intrahepatic lipid (β=1.16; p=0.03), following adjustments for total energy intake, age, and BMI [7]. Saturated fat overfeeding had significantly stronger impacts on liver fat, plasma ALT levels, and atherogenic lipid levels compared to unsaturated fat [9, 12], which actually yielded a slight decrease in liver fat, improved lipid profiles, and no change in ALT [12]. Sugar overfeeding likewise led to a 33% increase in IHTG, specifically through the stimulation of de novo lipogenesis [9]. Hepatic de novo lipogenesis, lipid metabolism, and the 23-hour postprandial triacylglycerol AUC increased with fructose, but not glucose, consumption [14]. Rats fed a high fructose diet (60%) for 16 weeks developed both hepatic steatosis and dyslipidemia [8]. Combined, these findings provide evidence that dietary saturated fat and fructose specifically lead to changes in metabolic profiles relevant to the pathogenesis of NAFLD.
A novel aspect of the present work was a focus on differential expression of lncRNAs. We observed a significant overrepresentation of differentially expressed lncRNAs in the PF vs CT and PF vs P + CT groups (Chi-square P < 0.0001), but not in the P vs CT group (p = 0.341), suggesting that expression of these transcripts is more sensitive to paracrine effects from hepatocytes. In addition, co-expression analysis identified a number of lncRNA/mRNA pairs that were both differentially expressed suggesting the presence of potential regulatory mechanisms between co-expressed lncRNAs and nearby mRNA transcripts. Compared with protein-coding genes, lncRNAs show stronger tissue-specific patterns of expression [53], and these transcripts may associate with other co-expressed RNAs to produce similar phenotypic effects [54]. Finally, in the comparison with data from a transcriptomic analysis of activated HSCs [30], we observed overlap of 306 lncRNAs showing a log2FC concordance, including 258 lncRNAs that were highly correlated between the two datasets. To date, the majority of lncRNAs are unannotated, so the significance of these findings remains to be determined. However, because lncRNAs are known to influence expression of genes located in proximity (cis-acting) or elsewhere (trans-acting) through interactions with DNA, RNA, and protein [55], it is possible that some of the changes in expression of protein-coding genes are a direct result of regulatory lncRNAs.
The main limitation of this study is that the factors released by hepatocytes in response to palmitate and fructose that result in RNA expression changes in HSCs remain uncharacterized. A number of studies have sought to identify mechanisms by which hepatocytes may communicate a pro-fibrotic message to HSCs. For example, HepG2 cells treated with palmitate were found to release sphingosine 1-phosphate, which increased expression of fibrogenic markers in LX-2 cells [20]. Another study showed that extracellular vesicles (EVs) were released from palmitate-treated mouse hepatocytes and subsequently internalized by both primary mouse HSCs and human LX-2 cells, resulting in increased expression of markers of HSC activation, migration, and proliferation [25]. In a similar study, exosomes from palmitate-treated Huh7 cells were shown to significantly alter expression of genes related to fibrosis in LX-2 cells [23]. Results from these two studies implicate a novel pathway by which lipotoxicity in hepatocytes may trigger fibrogenesis in HSCs.
We also recognize that palmitate may have cytotoxic effects on the liver and can result in cell injury and death [56]. Palmitate is known to activate hepatic stellate cells through mechanisms involving inflammasomes and hedgehog signaling [57]. In our studies, we did not observe increased cell death among hepatocytes treated with palmitate. Further, treatment of LX-2 cells with conditioned media from the palmitate + fructose no-cells control group did not increase expression of markers of activation, nor did morphological changes associated with transactivation to a myofibroblastic phenotype occur.
Conclusion
The results obtained in the current study demonstrate that conditioned media from hepatocytes treated with palmitate elicit changes in transcriptional programs associated with hepatic fibrosis in hepatic stellate cells, and these changes are amplified in the presence of fructose. The results have implications for dietary modifications in the prevention and treatment of NAFLD. Future investigations, including functional characterization of dysregulated transcripts and lncRNA-mRNA co-expressed networks, identification of factors released by hepatocytes in response to palmitate and fructose, and exploration of dietary interventions in appropriate animal models of NAFLD, will be important to extend these findings.
Acknowledgements
The authors acknowledge the technical support of Ms. Amanda Hanson.
GSG: reviewing and editing the manuscript.
ISP: data analysis; reviewing and editing the manuscript.
JKD: Project conception; experimental design; data analysis, writing and editing manuscript.
Funding
This paper was financially supported by the National Institutes of Health (NIH), Grant No. DK120890 and No. DK107735.
References
|
1
Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ:
Nonalcoholic fatty liver disease: a spectrum of clinical and pathological
severity. Gastroenterology 1999;116:1413-1419. |
|
|
|
|
|
2 Rafiq N, Bai C, Fang Y, Srishord M, McCullough
A, Gramlich T, et al.: Long-term follow-up of patients with nonalcoholic
fatty liver. Clin Gastroenterol Hepatol 2009;7:234-238. |
|
|
|
|
|
3 Ipsen DH, Lykkesfeldt J, Tveden-Nyborg P:
Molecular mechanisms of hepatic lipid accumulation in non-alcoholic fatty
liver disease. Cell Mol Life Sci 2018;75:3313-3327. |
|
|
|
|
|
4 Charlton MR, Burns JM, Pedersen RA, Watt KD,
Heimbach JK, Dierkhising RA: Frequency and outcomes of liver transplantation
for nonalcoholic steatohepatitis in the United States. Gastroenterology
2011;141:1249-1253. |
|
|
|
|
|
5 Younossi ZM: Non-alcoholic fatty liver disease
- A global public health perspective. J Hepatol 2019;70:531-544. |
|
|
|
|
|
6 Forouhi NG, Krauss RM, Taubes G, Willett W:
Dietary fat and cardiometabolic health: evidence, controversies, and
consensus for guidance. BMJ 2018;361:k2139. |
|
|
|
|
|
7 Cheng Y, Zhang K, Chen Y, Li Y, Li Y, Fu K, et
al.: Associations between dietary nutrient intakes and hepatic lipid contents
in NAFLD patiens quantified by 1H-MRS and Dual-Echo MRI. Nutrients
2016;8:527-545. |
|
|
|
|
|
8 Jensen VS, Hvid H, Damgaard J, Nygaard H,
Ingvorsen C, Wulff EM, et al.: Dietary fat stimulates development of NAFLD
more potently than dietary fructose in Sprague-Dawley rats. Diabetol Metab
Syndr 2018;10:4. |
|
|
|
|
|
9 Luukkonen PK, Sadevirta S, Zhou Y, Kayser B,
Ali A, Ahonen L, et al.: Saturated Fat Is More Metabolically Harmful for the
Human Liver Than Unsaturated Fat or Simple Sugars. Diabetes Care
2018;41:1732-1739. |
|
|
|
|
|
10 Parry SA, Hodson L: Influence of dietary
macronutrients on liver fat accumulation and metabolism. J Investig Med
2017;65:1102-1115. |
|
|
|
|
|
11 Puri P, Baillie RA, Wiest MM, Mirshahi F,
Choudhury J, Cheung O, et al.: A lipidomic analysis of nonalcoholic fatty
liver disease. Hepatology 2007;46:1081-1090. |
|
|
|
|
|
12 Rosqvist F, Kullberg J, Stahlman M, Cedernaes
J, Heurling K, Johansson HE, et al.: Overeating saturated fat promotes fatty
liver and ceramides compared to polyunsaturated fat: a randomized trial. J
Clin Endocrinol Metab 2019;104:6207-6219. |
|
|
|
|
|
13 Jensen T, Abdelmalek MF, Sullivan S, Nadeau
KJ, Green M, Roncal C, et al.: Fructose and sugar: A major mediator of
non-alcoholic fatty liver disease. J Hepatol 2018;68:1063-1075. |
|
|
|
|
|
14 Stanhope KL, Schwarz JM, Keim NL, Griffen SC,
Bremer AA, Graham JL, et al.: Consuming fructose-sweetened, not
glucose-sweetened, beverages increases visceral adiposity and lipids and
decreases insulin sensitivity in overweight/obese humans. J Clin Invest
2009;119:1322-1334. |
|
|
|
|
|
15 Abid A, Taha O, Nseir W, Farah R, Grosovski
M, Assy N: Soft drink consumption is associated with fatty liver disease
independent of metabolic syndrome. J Hepatol 2009;51:918-924. |
|
|
|
|
|
16 Assy N, Nasser G, Kamayse I, Nseir W,
Beniashvili Z, Djibre A, et al.: Soft drink consumption linked with fatty
liver in the absence of traditional risk factors. Can J Gastroenterol
2008;22:811-816. |
|
|
|
|
|
17 Abdelmalek MF, Suzuki A, Guy C, Unalp-Arida
A, Colvin R, Johnson RJ, et al.: Increased fructose consumption is associated
with fibrosis severity in patients with nonalcoholic fatty liver disease.
Hepatology 2010;51:1961-1971. |
|
|
|
|
|
18 Stephenson K, Kennedy L, Hargrove L,
Demieville J, Thomson J, Alpini G, et al.: Updates on Dietary Models of
Nonalcoholic Fatty Liver Disease: Current Studies and Insights. Gene Expr
2018;18:5-17. |
|
|
|
|
|
19 DiStefano JK: Fructose-mediated effects on
gene expression and epigenetic mechanisms associated with NAFLD pathogenesis.
Cell Mol Life Sci 2019;77:2079-2090. |
|
|
|
|
|
20 Al Fadel F, Fayyaz S, Japtok L, Kleuser B:
Involvement of Sphingosine 1-Phosphate in Palmitate-Induced Non-Alcoholic
Fatty Liver Disease. Cell Physiol Biochem 2016;40:1637-1645. |
|
|
|
|
|
21 Hetherington AM, Sawyez CG, Zilberman E,
Stoianov AM, Robson DL, Borradaile NM: Differential Lipotoxic Effects of
Palmitate and Oleate in Activated Human Hepatic Stellate Cells and Epithelial
Hepatoma Cells. Cell Physiol Biochem 2016;39:1648-1662. |
|
|
|
|
|
22 Hirahatake KM, Meissen JK, Fiehn O, Adams SH:
Comparative effects of fructose and glucose on lipogenic gene expression and
intermediary metabolism in HepG2 liver cells. PLoS One 2011;6:e26583. |
|
|
|
|
|
23 Lee YS, Kim SY, Ko E, Lee JH, Yi HS, Yoo YJ,
et al.: Exosomes derived from palmitic acid-treated hepatocytes induce
fibrotic activation of hepatic stellate cells. Sci Rep 2017;7:3710. |
|
|
|
|
|
24 Li L, Liu H, Hu X, Huang Y, Wang Y, He Y, et
al.: Identification of key genes in nonalcoholic fatty liver disease
progression based on bioinformatics analysis. Mol Med Rep 2018;17:7708-7720. |
|
|
|
|
|
25 Povero D, Panera N, Eguchi A, Johnson CD,
Papouchado BG, de Araujo Horcel L, et al.: Lipid-induced hepatocyte-derived
extracellular vesicles regulate hepatic stellate cell via microRNAs targeting
PPAR-gamma. Cell Mol Gastroenterol Hepatol 2015;1:646-663 e644. |
|
|
|
|
|
26 Windemuller F, Xu J, Rabinowitz SS, Hussain
MM, Schwarz SM: Lipogenesis in Huh7 cells is promoted by increasing the
fructose: Glucose molar ratio. World J Hepatol 2016;8:838-843. |
|
|
|
|
|
27 Wobser H, Dorn C, Weiss TS, Amann T,
Bollheimer C, Buttner R, et al.: Lipid accumulation in hepatocytes induces
fibrogenic activation of hepatic stellate cells. Cell Res 2009;19:996-1005. |
|
|
|
|
|
28 Zhao L, Guo X, Wang O, Zhang H, Wang Y, Zhou
F, et al.: Fructose and glucose combined with free fatty acids induce
metabolic disorders in HepG2 cell: A new model to study the impacts of
high-fructose/sucrose and high-fat diets in vitro. Mol Nutr Food Res
2016;60:909-921. |
|
|
|
|
|
29 Friedman SL: Hepatic stellate cells: protean,
multifunctional, and enigmatic cells of the liver. Physiol Rev
2008;88:125-172. |
|
|
|
|
|
30 Gerhard GS, Davis B, Wu X, Hanson A,
Wilhelmsen D, Piras IS, et al.: Differentially expressed mRNAs and lncRNAs
shared between activated human hepatic stellate cells and nash fibrosis.
Biochem Biophys Rep 2020;22:100753. |
|
|
|
|
|
31 Xu L, Hui AY, Albanis E, Arthur MJ, O'Byrne
SM, Blaner WS, et al.: Human hepatic stellate cell lines, LX-1 and LX-2: new
tools for analysis of hepatic fibrosis. Gut 2005;54:142-151. |
|
|
|
|
|
32 Weiskirchen R, Weimer J, Meurer SK, Kron A,
Seipel B, Vater I, et al.: Genetic characteristics of the human hepatic
stellate cell line LX-2. PLoS One 2013;8:e75692. |
|
|
|
|
|
33 Chu X, Jin Q, Chen H, Wood GC, Petrick A,
Strodel W, et al.: CCL20 is up-regulated in non-alcoholic fatty liver disease
fibrosis and is produced by hepatic stellate cells in response to fatty acid
loading. J Transl Med 2018;16:108. |
|
|
|
|
|
34 Lee JY, Cho HK, Kwon YH: Palmitate induces
insulin resistance without significant intracellular triglyceride
accumulation in HepG2 cells. Metabolism 2010;59:927-934. |
|
|
|
|
|
35 Zhao NQ, Li XY, Wang L, Feng ZL, Li XF, Wen
YF, et al.: Palmitate induces fat accumulation by activating
C/EBPbeta-mediated G0S2 expression in HepG2 cells. World J Gastroenterol
2017;23:7705-7715. |
|
|
|
|
|
36 Yao HR, Liu J, Plumeri D, Cao YB, He T, Lin L, et al.: Lipotoxicity in HepG2 cells triggered by free fatty acids. Am J Transl Res 2011;3:284-291. |
|
|
|
|
|
37 Gerhard GS, Hanson A, Wilhelmsen D, Piras IS,
Still CD, Chu X, et al.: AEBP1 expression increases with severity of fibrosis
in NASH and is regulated by glucose, palmitate, and miR-372-3p. PLoS One
2019;14:e0219764. |
|
|
|
|
|
38 Ritchie ME, Phipson B, Wu D, Hu Y, Law CW,
Shi W, et al.: limma powers differential expression analyses for
RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47. |
|
|
|
|
|
39 Kauffmann A, Gentleman R, Huber W:
arrayQualityMetrics--a bioconductor package for quality assessment of
microarray data. Bioinformatics 2009;25:415-416. |
|
|
|
|
|
40 Benjamini Y, Hochberg Y: Controlling the
False Discovery Rate: a Practical and Powerful Approach to Multiple Testing.
J R Statist Soc B 1995;57:289-300. |
|
|
|
|
|
41 Li X, Jin Q, Yao Q, Zhou Y, Zou Y, Li Z, et
al.: Placental Growth Factor Contributes to Liver Inflammation, Angiogenesis,
Fibrosis in Mice by Promoting Hepatic Macrophage Recruitment and Activation.
Front Immunol 2017;8:801. |
|
|
|
|
|
42 Aguayo-Orozco A, Bois FY, Brunak S, Taboureau
O: Analysis of Time-Series Gene Expression Data to Explore Mechanisms of
Chemical-Induced Hepatic Steatosis Toxicity. Front Genet 2018;9:396. |
|
|
|
|
|
43 Ijssennagger N, Janssen AWF, Milona A, Ramos
Pittol JM, Hollman DAA, Mokry M, et al.: Gene expression profiling in human
precision cut liver slices in response to the FXR agonist obeticholic acid. J
Hepatol 2016;64:1158-1166. |
|
|
|
|
|
44 Ge MX, Liu HT, Zhang N, Niu WX, Lu ZN, Bao
YY, et al.: Costunolide represses hepatic fibrosis through WW
domain-containing protein 2-mediated Notch3 degradation. Br J Pharmacol
2020;177:372-387. |
|
|
|
|
|
45 Haimerl F, Erhardt A, Sass G, Tiegs G:
Down-regulation of the de-ubiquitinating enzyme ubiquitin-specific protease 2
contributes to tumor necrosis factor-alpha-induced hepatocyte survival. J
Biol Chem 2009;284:495-504. |
|
|
|
|
|
46 Gao B, Jeong WI, Tian Z: Liver: An organ with
predominant innate immunity. Hepatology 2008;47:729-736. |
|
|
|
|
|
47 Magee N, Zou A, Zhang Y: Pathogenesis of
Nonalcoholic Steatohepatitis: Interactions between Liver Parenchymal and
Nonparenchymal Cells. Biomed Res Int 2016;2016:5170402. |
|
|
|
|
|
48 Tu T, Calabro SR, Lee A, Maczurek AE,
Budzinska MA, Warner FJ, et al.: Hepatocytes in liver injury: Victim,
bystander, or accomplice in progressive fibrosis? J Gastroenterol Hepatol
2015;30:1696-1704. |
|
|
|
|
|
49 Tsuchida T, Friedman SL: Mechanisms of
hepatic stellate cell activation. Nat Rev Gastroenterol Hepatol
2017;14:397-411. |
|
|
|
|
|
50 Jacobs S, Jager S, Jansen E, Peter A, Stefan
N, Boeing H, et al.: Associations of Erythrocyte Fatty Acids in the De Novo
Lipogenesis Pathway with Proxies of Liver Fat Accumulation in the EPIC-Potsdam
Study. PLoS One 2015;10:e0127368. |
|
|
|
|
|
51 Yilmaz M, Claiborn KC, Hotamisligil GS: De
Novo Lipogenesis Products and Endogenous Lipokines. Diabetes
2016;65:1800-1807. |
|
|
|
|
|
52 Musso G, Gambino R, De Michieli F, Cassader
M, Rizzetto M, Durazzo M, et al.: Dietary habits and their relations to
insulin resistance and postprandial lipemia in nonalcoholic steatohepatitis.
Hepatology 2003;37:909-916. |
|
|
|
|
|
53 Kapusta A, Kronenberg Z, Lynch VJ, Zhuo X,
Ramsay L, Bourque G, et al.: Transposable elements are major contributors to
the origin, diversification, and regulation of vertebrate long noncoding
RNAs. PLoS Genet 2013;9:e1003470. |
|
|
|
|
|
54 Cabili MN, Trapnell C, Goff L, Koziol M,
Tazon-Vega B, Regev A, et al.: Integrative annotation of human large
intergenic noncoding RNAs reveals global properties and specific subclasses.
Genes Dev 2011;25:1915-1927. |
|
|
|
|
|
55 Chen LL: Linking Long Noncoding RNA
Localization and Function. Trends Biochem Sci 2016;41:761-772. |
|
|
|
|
|
56 Arain SQ, Talpur FN, Channa NA, Ali MS,
Afridi HI: Serum lipid profile as a marker of liver impairment in hepatitis B
Cirrhosis patients. Lipids Health Dis 2017;16:51. |
|
|
|
|
|
57 Duan NN, Liu XJ, Wu J: Palmitic acid elicits
hepatic stellate cell activation through inflammasomes and hedgehog
signaling. Life Sci 2017;176:42-53. |
|