Targeting Extracellular RNA Mitigates Hepatic Lipotoxicity and Liver Injury in NASH

1. Introduction

Non-alcoholic fatty liver disease (NAFLD) refers to a spectrum of liver conditions affecting people who drink little to no alcohol [1]. This ranges from benign hepatic fat deposition (steatosis) to a much severe condition known as non-alcoholic steatohepatitis (NASH) [2]. NASH is a clinically alarming stage of NAFLD which results in liver injury, fibrosis, cirrhosis, and may eventually culminate into hepatocellular cancer (HCC) [2]. Unfortunately, we have no approved drug therapy for NASH and its progression may become irreversible in some individuals.

Hepatic injury due to lipotoxicity associated with NASH, is triggered by lipids which include saturated free fatty acids such as palmitic acid (PA) [3]. Upon entry into the hepatocytes, PA induces a number of cellular derangements including activation of stress kinases, oxidative stress, mitochondrial dysfunction, [4] and eventually, cell lysis [5,6]. The death of hepatocytes caused by lipotoxicity results in the release of a variety of damage-associated molecular patterns (DAMPs) such as DNA fragments, histones, ATP, uric acid, and cholesterol crystals, that act as stress signals to activate DAMP receptors, such as toll-like receptors (TLRs), purinoreceptors (P2X and P2Y) and C-type lectin domain (CLEC)12A, thereby initiating sterile inflammation and exacerbating tissue injury in NASH [7,8,9,10,11,12]. Extracellular RNA (eRNA) also belongs to this heterogeneous group of DAMPs and includes several types of RNAs present in the extracellular environment, including microRNA (miRNA), transfer RNA (tRNA), small interfering RNA (siRNA), and long noncoding RNA (lncRNA) [13]. eRNA released as a result of tissue injury may play a role in differentiation, chromatin modification, and inflammation, as well as tissue injury and repair [13,14,15,16,17,18,19,20,21]. eRNA binds to the Toll-like receptor (TLR) family receptors to induce apoptosis and inflammation in diseases such as ischemia-reperfusion and sepsis injury in cardiomyocytes [22,23,24]. Consistently, the administration of Ribonuclease1 (RNase 1), which degrades eRNA, has shown efficacy in decreasing disease severity in several preclinical disease models [22,23,24]. However, it is unknown whether eRNA plays a role in hepatic cell injury and pro-inflammatory cytokine production in NASH and whether RNase 1 administration can act as an interventional strategy to reduce NASH progression.

Here in this study, we demonstrate that eRNA released from injured/dead hepatocytes upon lipotoxic insult, increases hepatic injury along with pro-inflammatory cytokine production. Furthermore, the degradation of eRNA by RNase 1 treatment attenuates lipotoxicity both in vitro and in a mouse model of NASH.

2. Materials and Methods

2.1. Cell culture and treatment

HepG2 and AML12 hepatic cells were procured from American Type Culture Collection (ATCC, Manassas, VA, USA). HepG2 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Mouse hepatocytes, AML12 cells (CRL-2254) were maintained in DMEM-F12 1:1 containing 10% FBS, 1x ITS (Thermo Fisher Scientific, 41400), 10 nM dexamethasone and 1x penicillin/streptomycin. Both cell lines were maintained at 37°C in presence of 5% CO2. At 80% confluence, lipotoxicity was induced into the cells using 0.75 mM PA (saturated fatty acid), dissolved in ethanol; 0.5% BSA was used as a carrier for PA. Cells in the vehicle group were given only ethanol in 0.5% BSA containing media. For the administration of RNase 1, cells were incubated with 2.8U/mL RNase 1, one hour prior to the addition of PA and the treatment was continued upto 24 hours to maintain a sustained release of RNase 1 (Invitrogen, #12091021) in the culture media.

2.2. MTT assay

HepG2 and AML12 cells (1x10⁴) seeded into 96 well plates, were treated with PA with and without RNase 1. Vehicle control cells were given only ethanol in 0.5% BSA containing media. After 24 hours of treatment, medium was supplemented with MTT reagent (5mg/ml) for 4 hours at 37°C. DMSO (100ul) was then added to dissolve the purple formazan crystals and the plate was shaken for 10 min. Absorbance was recorded at 570nm using a multi-well spectrophotometer (ELISA reader).

MTT assay was also performed in HepG2 cells treated with PA in presence and absence of 10uM of TLR3/dsRNA complex inhibitor (Merck, #614310). In this case, cells in the vehicle group received both DMSO and ethanol in 0.5% BSA containing media as the TLR3/dsRNA complex inhibitor was dissolved in DMSO.

2.3. Analysis of total eRNA levels

Total eRNA levels in the conditioned media from vehicle control, PA treated, and PA+RNase 1 treated HepG2 cells were analyzed using the eRNA quantitation kit (Promega, Cat No: E3310 QuantiFluor® RNA System), according to the manufacturer’s protocol, using a multimode reader.

2.4. Oil Red O Staining

Lipid droplets in HepG2 cells were assayed by Oil Red O Stain. 96 well plates containing HepG2 cells were washed twice with phosphate-buffered saline. Cells were fixed with 10% paraformaldehyde for 30 min. After this, cells were washed with distilled water and 60% isopropanol was added for 5 minutes. Cells were then incubated in Oil Red O stain for 20 minutes, followed by washing with distilled water to remove the excess stain. Next cells were incubated with Hematoxylin for 1 minute. Oil Red O stain was extracted with 100% isopropanol for 5 minutes and absorbance was recorded at 492 nm using a multi-well spectrophotometer (ELISA reader).

2.5. Animal experiments

HFMCD NASH Model: 6-8-weeks-old male C57BL/6N mice were fed with a high-fat methionine-choline-deficient (HFMCD) diet consisting of 60 kcal% fat and 0.1% methionine by weight (A06071302; Research Diets, New Brunswick NJ), littermate mice fed with normal chow diet (NCD) were used as control mice. Mice were divided into three groups: NCD, HFMCD diet alone for 4 weeks, and HFMCD with RNase 1 (50ug/kg). RNase1 was injected intra-peritoneally, starting every alternate day after 2 weeks feeding on HFMCD diet, and continued for the next 2 weeks. All the animals were sacrificed after 4 weeks, and serum and liver tissues were collected. All animal procedures were carried out in accordance with the institutional guidelines for animal research at SGPGIMS.

2.6. RNA isolation and qRT-PCR

Total RNA from mice liver tissues and HepG2 cells was isolated using Trizol reagent (#343909). qRT-PCR was performed using the QuantiTect SYBR Green PCR Kit (Qiagen, 204141) according to manufacturer’s instructions. The GAPDH gene was used for normalization. KiCqStart SYBR Green Primers were purchased from Sigma-Aldrich, United States. Primer IDs were, Mouse Tnf-alpha (M_Tnf_1),Mouse IL6(M_IL6_1), Mouse IL1 beta (M_IL 1 beta _1), Mouse CCL20 (M_CCL20_1), Mouse Ccl2 (M_Ccl2_1), Mouse Ccl3 (M_Ccl3_1), Mouse Cxcl10 (M_Cxcl10_1), Mouse GAPDH (M_ Gapdh_1), Human Tnf- alpha (H_Tnf_1 ), Human IL-6 (H_IL6_1) Human CCL3(H_CCL3L3_1), Human CXCL10 (H_CXCL10_1), Human CCL20 (H_CCL20_1) and Human GAPDH (H_Gapdh_1).

2.7. Western blotting

HepG2 cells and mice liver tissue samples were lysed using RIPA buffer (Sigma, #R0278) mixed with phosphatase and protease inhibitor cocktails. Fifty microgram of protein was subjected to sodium dodecyl sulfate—polyacrylamide gel electrophoresis and electro-transferred onto a nitrocellulose membrane as per the manufacturer’s guidelines (Bio-Rad Laboratories, United States). Image acquisition was done using ChemiDoc (Syngene Model No: -G: BOX CHEMIXRQ). Densitometry analysis was performed using ImageJ software (NIH, Bethesda, MD, United States). Antibodies used were anti-GAPDH (Cell Signaling Technology, # 5174), anti- F4/80 (Cell Signaling Technology, # 70076), anti-Cleaved PARP (Cell Signaling Technology,#5625), anti-Phospho-c-JUN (Cell Signaling Technology,#3270), anti-Phospho-p-38MAPK (Cell Signaling Technology,#4511), anti-Phospho-SAPK/JNK(Cell Signaling Technology,#4668), anti-JNK ( Cell Signaling Technology,#9252 ), anti-c-JUN(Cell Signaling Technology,#9165), anti-p-38MAPK(Cell Signaling Technology,#8690), anti-LC3B(Cell Signaling Technology,#2775) and anti-P62(Cell Signaling Technology,#39749).

2.8. H&E staining

Tissues from the liver were fixed in situ after opening the abdomen, immersed in 4% paraformaldehyde in PBS, and processed into paraffin blocks. Sections were sliced at 5-micron thickness and mounted on slides to perform H&E staining. Images were analyzed using the ImageJ software, NIH. The non-alcoholic fatty liver disease activity scoring (NAS) system was used to quantify steatosis (0–3), lobular inflammation (0–3), and hepatocellular ballooning (0–2). Liver biopsies scored NAS ≥5 were classified as definitive NASH [25].

2.9. Serum ALT test

Alanine aminotransferase (ALT) levels in the serum were assessed using the kit (Abcam calorimetric assay kit #ab105134) as per the manufacturer’s protocol.

2.10. Liver TG analysis

Liver triglyceride (TG) levels were assessed using the kit (Cayman, Triglyceride Colorimetric assay kit # Cat No. 10010303) as per the manufacturer’s protocol.

2.11. Confocal microscopy

Immunofluorescence experiments were performed on treated HepG2 cells in chambered slides and in paraffin-embedded sections of mouse liver. In brief, formalin-fixed cells and tissues were permeabilized with 0.1% Triton X-100 (SIGMA-ALDRICH, X100) in PBS for 5-10 min and blocked with 3% BSA-PBS for 30 min, at room temperature. Cells and tissues were incubated with the primary antibody (1:200 in 3% BSA-PBS) overnight at 4°C followed by fluorochrome labelled secondary antibodies (Molecular probes) and cell imaging was performed using an LSM710 Carl Zeiss (Carl Zeiss Microscopy GmbH, Germany) confocal microscopy. The primary antibody used for HepG2 cells and liver sections was anti-NFkB (Cell Signaling Technology, #8242). Colocalization studies were performed using ImageJ software with JACoP plugin.

2.12. Mitochondrial ROS detection

Mitochondrial ROS was detected in vehicle control and treated HepG2 cells using 5 µM MitoSOX™ Red (catalog number #M36008), in which cells were incubated for 20-30 min at 37°C. After washing with PBS, reading was taken in a fluorimeter at an excitation wavelength of 510 nm and an emission wavelength of 580 nm.

2.13. Mitochondrial membrane potential (MMP)

Mitochondrial membrane potential (MMP) was assessed by JC-1 staining. HepG2 cells seeded into 96 well black bottom plates were treated with PA with and without RNase 1 for 24 hours. Cells were washed with PBS and incubated with 2 µM JC-1 dye (Invitrogen, catalog number #T3168) at 37 °C for 30 min. Fluorescence was measured by a fluorimeter (Synergy HTX multimode microplate reader, BioTek) at an excitation wavelength of 488 nm and an emission wavelength of 530 nm (monomer) and 590nm (aggregates). Loss of MMP was calculated from the ratio of emission wavelength at 590nm/530nm.

2.14. Statistics

One-way ANOVA with Tukey's post hoc test was used to compare among groups. Statistical analysis was performed using the Graph-Pad Prism software, version 5.0. *p < 0.05, **p < 0.01, and ***p< 0.001 were considered statistically significant.

3. Results

3.1. RNase 1 attenuates PA induced cellular injury in HepG2 cells

Lipotoxicity by PA is known to induce apoptosis and cellular stress in hepatic cells, and mimics hepatocyte damage in NASH [26]. To determine if PA induced hepatocyte injury leads to the release of eRNA, we assessed the eRNA content in the conditioned media of PA treated HepG2 cells and observed a significant increase in the eRNA levels. Furthermore, the co-administration of RNase 1 significantly reduced the released eRNA levels upon PA treatment, when compared with PA treatment alone (Figure 1A). Concurrently, RNase 1 treatment also significantly rescued the cellular viability of HepG2 upon lipotoxic insult (Figure 1B). These results were further supported by the decrease in cleaved PARP levels (apoptosis marker) in RNase 1 and PA treated cells, compared to increased apoptosis, as observed in PA alone treated cells (Figure 1C,D). Additionally, PA induced hepatic steatosis was also reduced by RNase 1 administration (Supplementary Figure 1). Cellular stress kinases such as p-JNK and p-p38MAPK are key mediators of PA induced cellular stress [27]. In line with these observations, we also assessed the effect of RNase 1 treatment on PA induced JNK and p38MAPK activation. Our results showed that, indeed, similar to its effect on cellular viability, the targeted degradation of eRNA by RNase 1 significantly reduced the PA induced activation of JNK, its downstream target c-jun, and p38MAPK (Figure 1E-J). Furthermore, Rnase 1 treatment also rescued PA induced autophagy block which is a major mediator of lipotoxicity [28] (Supplementary Figure 2). To verify if the observed effects of RNase 1 are specifically due to its effects on eRNA, we also used a pharmacological inhibitor of TLR3, a receptor of eRNA and observed a similar rescue effect on PA induced cell death (Figure 1K). This small molecule inhibitor specifically inhibits the TLR3/dsRNA complex formation and has been shown to prevent eRNA signaling [29]. Concurrently, the administration of RNase 1 alone to HepG2 cells had no effect on cell viability (Supplementary Figure 3A). These sets of results, thus, rule out any possible off target effect of RNase 1. We further validated the effect of RNase 1 on another hepatocyte cell line (AML-12) and observed similar effects as observed in HepG2 cells (Supplementary Figure 3B).

PA is also known to increase mitochondrial dysfunction via its effect on mitochondrial reactive oxygen species (ROS) generation and loss of MMP [4]. Therefore, to investigate the effect of RNase 1 on PA induced mitochondrial damage, we first assessed the levels of mitochondrial ROS and observed significant protection conferred by RNase 1 to PA induced mitochondrial ROS generation (Figure 2A). Similarly, in agreement with its effect on mitochondrial ROS, RNase 1 prevented PA induced reduction in MMP, in HepG2 cells, as measured by JC-1 staining (Figure 2B).

Therefore, collectively these data indicate a direct involvement of eRNA in PA induced hepatic injury in vitro and the abrogation of these effects by RNase 1 treatment.

3.2. RNase 1 attenuates PA induced pro-inflammatory signalling in HepG2 cells

PA induced proinflammatory and pro-apoptotic signalling is linked to hepatic inflammation and cell death in NASH [30], and is associated with NF-κB activation [31]. Intriguingly, eRNA released from injured cells have been previously shown to induce TLR receptors and stimulate NF-κB dependent cytokine and chemokine synthesis [13]. Therefore, we speculated if PA induced cytokine production is partly mediated by eRNA signalling. To test this, we treated HepG2 cells with PA alone or PA along with RNase 1 and assessed the activation of NF-κB along with pro-inflammatory cytokine expression. Our results definitively showed that RNase 1 treatment almost completely inhibited PA induced NF-κB nuclear translocation, which measures its activation (Figure 3A,B). In concurrence with NF-κB nuclear translocation, RNase 1 also prevented the PA induced activation of NF-κB transcriptional targets IL-6 and TNFα in presence of PA (Figure 3C). As TNFα is a major cell death inducer in hepatic cells [30], its inhibition may also explain the observed increase in the viability upon RNase 1 administration as described earlier (Figure 1B). Similarly, the PA induced expression of several other NF-κB induced chemokines such as CCL3, CCL20 and CXCL10, were repressed by RNase 1 co-treatment (Figure 3D). Therefore, these data suggest an anti-inflammatory action of RNase 1 in hepatic cells treated with PA.

3.3. RNase 1 administration mitigates NASH induced liver injury in mouse

To determine if eRNA is also involved in NASH pathogenesis in vivo, we used C57BL/6N mice and fed them HFMCD diet supplemented with 60 kcal% fat and 0.1% methionine by weight. These animal models rapidly developed severe NASH phenotype owing to the lipotoxic action of stored hepatic fat when compared to mice fed with control diet [32,33,34]. We administered RNase 1 in animals fed with HFMCD diet and followed its effects on the NASH phenotype. RNase 1 (50ug/kg), i.p. injections starting every alternate day after 2 weeks of HFMCD feeding were continued for the next 2 weeks. Mice fed either NCD or HFMCD diet for 4 weeks were taken for comparison. All the animals were sacrificed after 4 weeks.

Histological evaluation of the livers of mice fed with NC, HFMCD diet and HFMCD diet with RNase 1 intervention, showed a marked increase in hepatic steatosis in the HFMCD fed group, and its reduction upon RNase 1 co-administration (Figure 4A). These results were further corroborated by hepatic TG measurement in the different experimental groups (Figure 4B). HFMCD diet consumption damages hepatic tissue, which is reflected by the elevated levels of serum ALT in HFMCD mice, consistent with histological findings (Figure 4C). However, RNase 1 administration significantly reduced the levels of ALT in mice fed with HFMCD diet (Figure 4C). Additionally, cellular stress markers such as p-JNK and p-p38MAPK which were upregulated in HFMCD fed mouse livers were also reduced by RNase 1 administration (Supplementary Figure 4). Therefore, these results indicate a protective effect of RNase 1 on lipid induced liver injury in vivo.

3.4. RNase 1 administration reduces NASH induced liver inflammation

The HFMCD mouse model mimics several aspects of human NASH pertaining to inflammation and immune cell infiltration [35]. Therefore, to validate our in vitro findings, we looked at the effect of RNase 1 on NASH induced NF-κB activation and pro-inflammatory cytokine expression. Our results demonstrated that similar to the effect of RNase 1 observed in HepG2 cells, in vivo blocking of eRNA action by RNase 1 significantly prevented hepatic NF-κB nuclear translocation (Figure 5A,B) and the expression of pro-inflammatory genes (Figure 5C,D & Supplementary Figure 5) in the liver of HFMCD diet fed mice. Additionally, the mRNA expression of chemokines was also similarly altered by RNase 1 administration in the liver of HFMCD diet fed mice (Figure 5E & Supplementary Figure 5). Furthermore, we also observed an inhibitory effect of RNase 1 administration on intrahepatic macrophage infiltration in HFMCD diet fed mouse liver as assessed by hepatic F4/80 levels (Figure 5F,G). Furthermore, the outcome of steatosis, inflammation, and hepatic ballooning, were assessed histologically by providing the NAFLD activity score, which was significantly decreased upon RNase 1 treatment (Figure 5H). Collectively, these results demonstrated an anti-lipotoxic action of RNase 1 by attenuating NASH induced tissue damage and inflammation.

4. Discussion

As a disease, NAFLD displays a spectrum of disease stages ranging from benign steatosis to NASH [2], which is further presented as steatohepatitis, fibrosis and cirrhosis [1]. Unfortunately, despite its increasing prevalence across the globe, we still do not have an effective treatment strategy for NASH, other than life-style modifications [36]. At the molecular level, a tipping point during progression from benign steatosis to NASH, is the increase in the incidence of hepatocyte injury which is associated by necrotic/apoptotic death of hepatocytes, release of pro-inflammatory cytokines, and infiltration of macrophages leading to inflammation [37]. Therefore, understanding these events that link hepatocyte injury to inflammation is the key to preventing the transition and progression of steatosis to NASH.

One of the key concepts that help to explain the large spectrum of NAFLD seen in patients is the process called “Lipotoxicity” [37]. Indeed, when a liver cell experiences an increased flux of lipids, it attempts to store it into inert lipid droplets, however, in case the lipid load overruns the adaptive capacity of hepatocytes, the fatty acids inside the cells lead to the generation of reactive lipid species which trigger cell death by several distinct molecular mechanisms, encompassing the process of lipotoxicity [38]. Lipotoxic injury results in the release of DAMPs by hepatocytes, also known as danger signals that trigger the activation of sterile (i.e. in the absence of infection) inflammatory pathways, that when perpetuated over time, result in chronic injury and an abnormal wound healing response with fibrosis [39,40].

Basic and clinical research have provided evidence that the eRNA released by injured cells is an important player in the crosstalk between immunity and tissue injury in several diseases [13]. In fact, (patho-) physiological functions of eRNA are associated with, and in many cases, causally related to conditions such as arterial and venous thrombosis, atherosclerosis, ischemia/reperfusion injury, or tumor progression, especially in association with the elevated inflammatory status of these diseases [16,19,41,42,43,44]. eRNA can be liberated from the cells in a free form or bound to proteins or phospholipids as well as in association with extracellular vesicles (EVs) or apoptotic bodies [13]. Once released from an injured cell, eRNA binds to the adjacent cells via surface-bound Toll-like receptors (TLRs) and activates the pro-inflammatory NF-kB pathway, leading to the synthesis and secretion of pro-inflammatory cytokines like TNF-α and IL-6 [13]. Therefore, eRNA may serve as a catalyst to increase inflammatory environment and worsening of disease pathology. However, the adverse effects of eRNA are countered by several extracellular RNase, which proteolytically degrade eRNA [13]. In this regard, the exogenously administered RNase 1 bears considerable potential as new therapeutic agent, based on its tissue-protective functions that may translate into anti-inflammatory properties in different pathological situations [13].

As NASH represents a milieu wherein injured hepatocytes and cytokine production co-exist, it may be possible that eRNA drives NASH progression, however, this possibility has still not been tested so far. In the results presented here, we provide for the first-time, the direct evidence of eRNA acting as a paracrine pro-inflammatory mediator of lipotoxicity in hepatic cells. Our results demonstrate that inhibiting the biological activity of eRNA by RNase 1, prevents PA induced cellular stress in HepG2 cells as well as the expression of pro-inflammatory cytokines. Additionally, these data were confirmed in an animal model of NASH which showed that RNase 1 administration significantly reduced liver injury and inflammation associated with diet- induced NASH. Our study is in agreement with a previous report showing the possible involvement of TLR3 in human NASH, [45] which further strengthens our proposition of eRNA mediated inflammation in NASH.

Further studies are needed to profile eRNA across stratified NAFLD patients to assess the differences associated with the progression of the disease. Additionally, further experiments will be required to engineer stable RNase 1 with minimum side effects which can enter human clinical trials for NASH. In line with this, RNase therapy is already in clinical trials for cancer treatment [46]. Furthermore, based on our in vitro results, the use of specific TLR3 inhibitors should also be tested in preclinical models of NASH.

In summary, our results put forward eRNA released by injured hepatocytes under a lipotoxic environment constituting a feed forward loop, wherein hepatocyte injury is intertwined with the maintenance of a pro-inflammatory environment. Therefore, targeting eRNA by RNase 1 or specific TLR3/dsRNA inhibitors may be a new therapeutic strategy to reduce the pathogenesis of NASH (Figure 6).

5. Conclusions

Our results provide an insight into the pathological role of eRNA released by injured hepatocytes and put forth a proof-of-concept for limiting eRNA action to blunt NASH progression.