Non-alcoholic fatty liver disease (NAFLD) includes a wide compass of liver pathologies, ranging from simple steatosis, usually a mild, benign and non-progressive condition, to non-alcoholic steatohepatitis (NASH), which may progress to liver cirrhosis and ultimately hepatocellular carcinoma (HCC). With NAFLD affecting both children and adults alike, it is postulated to emerge as the leading cause of end-stage liver diseases in the coming years [1
]. Several cellular and molecular events conspire and collaborate to transform simple steatosis to NASH to HCC. However, the underlying precise mechanisms of disease pathogenesis and NAFLD progression have just begun to be understood. Some of the newly emerging concepts include iron overload, inflammation, dysregulated fat metabolism, oxidative stress, gut microbiota and angiogenesis [2
Angiogenesis or new blood vessel formation is a crucial aspect of inflammation and a critical step in tissue damage, healing, and vascular remodeling. Changes in liver vascular architecture have been linked to the progression of fibrosis, cirrhosis and HCC in chronic liver diseases (CLD) [3
]. Almost all experimental and clinical conditions of CLD, including NASH, have been associated with an over-expression of pro-angiogenic cytokines and related receptors. Although various studies have reported an upregulation of angiogenic factors, particularly VEGF in NASH, the underlying mechanisms that regulate angiogenesis, inflammation and fibrogenesis in NASH pathology remain obscure [5
In the current study, we explored the role of Runt-related transcription factor 1 (RUNX1
) in the pathogenesis of NASH. RUNX1
, also known as acute myeloid leukemia 1 (AML1
), is a powerful and pivotal regulator of hematopoiesis and angiogenesis [8
]. A defect in RUNX1
is associated with impairment in angiogenesis accompanied by the absence of hematopoietic stem cells [10
]. Given the significance of RUNX1
in angiogenesis and its rarely identified role in NASH, we investigated the expression and function of RUNX1 in NASH pathology by addressing its emergence in endothelial cells (ECs).
2. Materials and Methods
2.1. Study Subjects and Collection of Samples
Human liver tissues were histologically examined for patients without NAFLD (n
= 33), patients with simple liver steatosis (n
= 46) and patients with NASH (n
= 43) as described earlier [11
] (for tissue characteristics see Supplementary Table S1
). A subset of these samples was used for a mRNA microarray analysis: patients without NAFLD (n
= 7), patients with simple liver steatosis (n
= 7), and with non-alcoholic steatohepatitis (NASH) (n
= 7). The experimental procedures were performed according to the guidelines of the charitable state-controlled foundation HTCR (Human Tissue and Cell Research, Regensburg, Germany), with written informed consent from patients. The study in Germany and the consent form were approved by the local ethical committee of the University of Regensburg (ethics statement 12-101-0048, University of Regensburg, Germany). Additionally, immunohistochemistry (IHC) studies were conducted on liver biopsies collected from NASH patients’ samples (n
= 16) and control liver tissues (n =10) collected in ILBS, New Delhi (for patient characteristics see Supplementary Table S2
). The study performed in India was duly approved by the Human ethics committee of ILBS, New Delhi (ethics approval F25/5/64/ILBS/AC2014/1484). All experiments involving human tissues and cells were carried out in accordance to The Code of Ethics of the World Medical Association (Declaration of Helsinki).
2.2. Differential Gene Expression Studies and qRT-PCRs
About 17 differentially expressed genes (DEGs) obtained from a microarray experiment and associated with gene ontology (GO) term angiogenesis were selected for further Taqman quantitative real time-PCR (qRT-PCR) validation studies (Supplementary Table S3A
) using a larger cohort of NAFLD liver tissue samples (Supplementary Table S1
). For in vitro assays, SYBR Green PCR master mix (Applied Biosystems, Foster City, CA, USA) based qRT-PCR studies were done (Supplementary Table S3B
2.3. Immunohistochemistry Analysis
Samples of human liver tissues were fixed and stained as per standard protocols. IHC scoring was done on a scale of 1–4 by counting RUNX1 positive cells per field. Details of the protocols and antibodies used are given in the supplementary material and Supplementary Table S4
2.4. Culture of Endothelial Cells with Conditioned Medium from Hepatoma Cells Treated with Palmitic Acid
Huh7 cells or mouse primary hepatocytes were treated with 200 µM palmitic acid-BSA (PA) for 48 h according to previously published methods [13
]. BSA treated cells served as controls. To investigate the effect of steatotic liver cells on gene expression, human umbilical vein endothelial cells (HUVECs) or LSECs (mouse) were incubated with conditioned medium (CM) from BSA/PA treated Huh7 cells or primary hepatocytes, respectively, for 24 h and then assayed for gene expression. For validation of VEGF in the induction of RUNX1
gene expression, studies were also conducted by adding VEGF blocking antibody in HUVECs along with CM from BSA/PA treated Huh7 cells.
2.5. Induction of RUNX1 Expression in HUVECs
To study induction of RUNX1 expression in HUVECs, HUVECs were treated with or without 10 ng/mL VEGF (Himedia Laboratories, Mumbai, India) or TGF-β for 24 h. After 24 h, cells were trypsinized for analysis of RUNX1 gene expression in unstimulated and stimulated cells.
2.6. RUNX1 Inhibition and Overexpression in HUVECs
To elucidate the role of RUNX1 in inflammation and angiogenesis, steatotic or activated HUVECs were transfected with 50 nmol/mL × 106 cells of either negative control (NC) siRNA (Catalogue no# AM4635) or pre-designed RUNX1-specific siRNA (siRNA ID: s229352) (Thermofisher Scientific, Waltham, MA, USA) using lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Forty-eight hours after transfection, cells were analyzed by qRT-PCR to confirm the knockdown of RUNX1. For overexpression studies, HUVECs were transfected with 2μg/L × 106 cells of control plasmid (pControl, empty vector) or RUNX1 plasmid (pcDNA3.1+ /C-(K)-DYK vector with RUNX1b, Gene Script # OHu26354), referred to as pRUNX1, in the absence or presence of 10 ng/mL VEGF. For transfection, lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) was used according to the manufacturer’s instructions. pcDNA3-EGFP plasmid vector (kind gift from Dr. Vijay) was used as the control of transfection efficiency and expression in all the transfection experiments. Forty-eight hours after plasmid transfection, cells were analyzed by qRT-PCR. HUVECs with loss of RUNX1 or gain of RUNX1 expression were assayed for gene expression of adhesion molecules, angiogenic markers by qRT-PCR, flow cytometry and angiogenic functions by matrigel assays. CCL2 levels were assayed by ELISA in HUVECs under different conditions.
2.7. Flow Cytometry Analysis
The labeled antibodies used for flow cytometry are given in Supplementary Table S4
. After antibody incubation in PBS for 45 min at 4 °C, the cells were fixed with paraformaldehyde in PBS. Multicolor flow cytometry was performed using FACS Verse (BD Biosciences, San Jose, CA, USA) and minimum of 1 million events using live cells were acquired. Analysis of flow cytometry data was performed using Flow-Jo v10 software (BD Biosciences, San Jose, CA, USA). Unstained cells without any antibody were used as negative controls.
2.8. Statistical Analysis
PCR data obtained from patient samples were evaluated for normality distribution by a Shapiro–Wilk test. Statistical differences between the two groups were analyzed by a two-tailed Mann–Whitney U Test or a Student's unpaired t-test (in vitro and flow cytometry experiments) and statistical differences between several groups (data from human samples) by a Kruskal–Wallis Test (SPSS Statistics program, IBM, Leibniz Rechenzentrum, München, Germany). A value of p < 0.05 was regarded as significant. The Pearson correlation (r) was calculated using the IBM SPSS Statistics program. Each experiment was performed in at least triplicates and data were expressed as means ± SD (standard deviation).
In the current study, we report increased expression of RUNX1 in liver NPCs, presumably among others in ECs of NASH livers. We describe a novel angiogenic and inflammatory role of RUNX1 in NASH pathogenesis. Signals such as VEGF from PA-treated hepatocytes induce/increase the expression of RUNX1 in ECs, resulting in an enhanced expression of angiogenic and adhesion molecules in these cells, potentially augmenting increased leucocyte migration and adhesion.
Oxidative stress and inflammation-driven pathological angiogenesis is an important mechanism in the progression of NAFLD from steatosis to NASH to cirrhosis and subsequently HCC [17
]. Studies have documented both increased and decreased hepatic expression of the angiogenic genes, such as VEGFA and their cognate receptors in patients with NAFLD when compared to control tissues [5
]. In our study subjects, we did not observe a significant difference in the gene expression of known angiogenic factors such as VEGFA, VEGFR1
. However, we found a group of angiogenesis associated genes to be differentially expressed in NASH. Among the transcription factors known to control these DEGs, we became particularly interested in RUNX1, which was significantly positively correlated with the histopathological features of NASH. PPARγ
was another transcription factor, which was a part of our validated DEGs and showed a good correlation with NAS, steatosis and inflammation degree of the patients and the role of PPARγ in NAFLD is also well established [19
RUNX1 is a salient factor that is known to control diversification between hematopoietic and endothelial cell lineages [20
]. Gain and loss of function in RUNX1 has been correlated with cancer progression and metastasis, most notably in acute myeloid leukemia [22
]. The role of RUNX1 in enhancing TLR4-mediated inflammation has been demonstrated previously [23
]. In our study, there was an increased expression of RUNX1 in liver endothelial cells of NASH patients that significantly correlated with severity of disease, hypothesizing a pathogenic role of endothelial-specific expression of RUNX1 in NASH. Our findings are in concordance with those of Lam et al. who have also identified RUNX1
as a gene upregulated in CD31-positive vascular ECs obtained from human proliferative diabetic retinopathy fibrovascular membranes [24
]. Although RUNX1 was not present in parenchymal liver cells, hepatocytes and cholangiocytes, we did find faint RUNX1
mRNA expression in Huh7 cells, which is in concordance with the observation of a low RUNX1 nuclear expression in other hepatoma cells such as human Hep3B and mouse AML12 [25
]. Furthermore, RUNX1 was shown to regulate TIMP1 (Tissue Inhibitor of Metalloproteinase 1) expression in hepatic stellate cells (HSCs) and to play a role in activating HSCs in a mouse NASH model [14
]. This may be because in earlier studies, they solely analyzed RUNX1
in the HSCs and in our study, we focused on endothelial-specific expression of RUNX1 and its potential role in NAFLD pathogenesis. However, it is highly possible that an increased RUNX1 expression in HSCs or other NPCs such as Kupffer cells also contributes to NASH pathogenesis via distinct mechanisms.
Performing in vitro cultures, we observed an increased expression of RUNX1
in HUVECs which were treated with CM from PA treated hepatoma cells, indicating that non-saturated fatty acids and/or high fat conditions, which are associated with oxidative stress and generation of reactive oxygen species (ROS), may be inducing RUNX1
expression in HUVECs. Interestingly, a study has reported an increase in mRNA expression of RUNX1
in liver tissue of a NAFLD guinea pig model, suggesting a regulatory role of RUNX1
for organic cation transporter N1 (OCTN1) [27
]. OCTN1 specifically transports ergothioneine, a natural radical scavenger and therefore, augments its anti-oxidative and anti-inflammatory properties. Previous studies in model organisms have shown that high glucose levels act as a trigger for RUNX1
expression via ROS–mediated upregulation of hypoxia-inducible factor 1 [24
]. Therefore, RUNX1
expression seems to be triggered by cellular oxidative stress, dietary factors and with regard to hepatic steatosis, this is mediated by VEGF and TGF-β, which were shown to be released from hepatoma cells after PA treatment. The activation of RUNX1
by TGF-β and other transcription factors such as SMAD1
has also been well reported in previous studies [28
Overexpression of RUNX1 in ECs resulted in substantial increase in the expression of adhesion molecules VCAM1 and PECAM1 and also that of CCL2, independent of additional treatment with VEGF. Based on this finding, we propose a novel role of RUNX1 in the potential recruitment of inflammatory cells in NASH, because VEGF is known to aggravate endothelial cell chemokine production in vitro and in vivo and functions in the recruitment of monocytes and T cells [16
]. Hence, both VEGF and RUNX1 may act together as crucial angiogenic and pro-inflammatory inducers in NASH. We confirmed the angiogenic properties of RUNX1 in ECs and show that RUNX1 may be an important downstream effector of VEGF in mediating endothelial angiogenesis. However, angiogenic factors other than VEGF may also be involved in RUNX1 regulation and need further characterization.
RUNX1 attenuation also led to a decrease in the expression of known RUNX1 target genes, including CCL2, PI3KCA, PRKCE
. Furthermore, a decrease in angiogenic activity of ECs while RUNX1
mRNA expression is silenced validated its role as an angiogenic modulator. The contribution of RUNX1
towards angiogenesis in steatosis may also be deduced by the fact that many of its target genes, including eNOS
, are known to be involved in various aspects of angiogenesis, including EC proliferation, sprouting and vascular stabilization during hepatic steatosis and inflammation in high fat diet mice [30
]. PRKCE is already known to play a critical role in mediating fat-induced hepatic insulin resistance through the buildup of diacylglycerol in NAFLD [32
]. CCL2 is mainly secreted by infiltrating inflammatory monocytes and HSCs in an injured liver and is known to provide pro-angiogenic signals during chronic liver injury [35
]. In our study, CCL2 significantly correlated with the severity of human NASH and CCL2 release from ECs was significantly altered by RUNX1 expression, adding novel insight into how RUNX1/CCL2 mediates infiltration of inflammatory cells and angiogenesis in NASH. In addition, the expression of VEGFR1
was found to be attenuated in HUVECs after RUNX1
mRNA silencing, implicating that VEGFR1
may be a vital factor downstream of RUNX1-mediated angiogenesis and disease progression in NAFLD [36
]. However, whether VEGFR1
is a direct target of RUNX1 remains to be determined.