Loss of MicroRNA-21 Protects Against Acetaminophen-Induced Hepatotoxicity in Mice

Abstract

Acetaminophen (APAP) induced acute liver failure (ALF) is recognized as the most common cause of ALF in Western societies. APAP-induced ALF is characterized by coagulopathy, hepatic encephalopathy, multi-organ failure, and death. MicroRNAs are small, non-coding RNAs that regulate gene expression at the post-transcriptional level. MicroRNA-21 (miR-21) is dynamically expressed in the liver and is involved in the pathophysiology of both acute and chronic liver injury models. We hypothesize that miR-21genetic ablation attenuates hepatotoxicity following acetaminophen intoxication. Eight-week old miR-21knockout (miR21KO) or wild-type (WT) C57BL/6N male mice were injected with acetaminophen (APAP, 300 mg/kg BW) or saline. Mice were sacrificed 6 or 24 hours post-injection. MiR21KO mice presented attenuation of liver enzymes ALT, AST, LDH compared with WT mice 24 hours post-APAP treatment. Moreover, miR21KO mice had decreased hepatic DNA fragmentation and necrosis than WT mice after 24 hours of APAP treatment. APAP-treated miR21KO mice showed increased levels of cell cycle regulators CYCLIN D1 and PCNA, increased autophagy markers expression (Map1LC3a, Sqstm1) and protein (LC3AB II/I, p62), and an attenuation of the APAP-induced hypofibrinolytic state via (PAI-1) compared with WT mice 24 post-APAP treatment. MiR-21 inhibition could be a novel therapeutic approach to mitigate APAP-induced hepatotoxicity and enhance survival during the regenerative phase, particularly to alter regeneration, autophagy, and fibrinolysis. Specifically, miR-21 inhibition could be particularly useful when APAP intoxication is detected at its late stages and the only available therapy is minimally effective.

INTRODUCTION

Acute liver failure (ALF) is characterized by severe and sudden loss of hepatocellular function in patients with previously normal function. The American Association for the Study of Liver Diseases includes evidence of coagulation abnormality and mental alteration in its definition of ALF and lists drug-induced liver injury as one of the most prominent causes (Polson and Lee 2005). Specifically, acetaminophen (N-acetyl-p-aminophenol, paracetamol, APAP) intoxication accounts for approximately 50% of all ALF related cases (Lee 2017). APAP is both a widely used over-the-counter analgesic and antipyretic as well as a prescribed medication used in combination with opioids for pain management. Both by intentional and by non-intentional overdose of APAP, the drug induced hepatotoxicity and liver failure can lead to the need of a liver transplant in approximately 20% of cases (Lee 2017). Although clinical stages of toxicity present in a predictable timeline, as the response to APAP has been well-studied as a hepatotoxin, these symptoms can often vary among individuals (Yoon et al. 2016). Because of its widespread use, APAP is considered to be the most used analgesic in the United States, the compound is often available in many of the most commonly used over-the-counter drugs for both adults and children (Herndon and Dankenbring 2014).

The process of APAP hepatotoxicity in mice begins with formation of N-acetyl-p-benzoquinoneimine (NAPQI) when stores of glutathione (GSH) are limited. Most commonly, phase II enzymes metabolize APAP which may be glucuronidated and sulfonated into a non-toxic form and excreted in urine. In the liver approximately 10% of the drug is metabolized by P450 enzymes, specifically CYP2E1, to become a more toxic form of the compound NAPQI (Yoon et al. 2016). If stores of GSH are not available to further conjugate NAPQI, it is able to form covalent bonds with cellular proteins and lead to adduct formation, inflammation, and ultimately hepatic centrilobular necrosis (Ramachandran and Jaeschke 2017). Current therapies are limited in number and in effectiveness, and exploring novel options is an area of actively ongoing research (Jaeschke 2019). The only recognized clinical antidote (Rumack and Bateman 2012), N-acetylcysteine (NAC)- a promoter of glutathione synthesis, has a therapeutic efficacy that is time-dependent but may only be useful within 8 hours of overdose (Morrison et al. 2019). Investigating effective secondary strategies for therapies that can be delivered at later times following APAP overdose would be beneficial to a population at high-risk for liver transplant following APAP intoxication.

MicroRNAs (miRNAs) are small non-coding RNAs that regulate and silence protein expression. The influence of miRNAs in repression of protein expression is due to their ability to pair with specific mRNA targets, which differ between different miRNAs (Guo et al. 2010). Additionally, these miRNAs are able to destabilize and degrade mRNA (Guo et al. 2010) as well as recruit silencing proteins to specific mRNA targets (Huntzinger and Izaurralde 2011). Various microRNAs have been identified for their potential to serve as biomarkers of APAP-induced hepatotoxicity prior to elevation of liver enzymes, as they are readily detectable in circulating blood levels (Thulin et al. 2014; Yang et al. 2015). MicroRNA-21 (miR-21) has been identified for exhibiting dynamic biological functions in multiple organs including proliferation, fibrosis, apoptosis (Krichevsky and Gabriely 2009). MiR-21 has been shown to regulate expression of target genes involved in a multitude of liver related diseases (Zhang et al. 2020) including hepatocellular cancer (Tomimaru et al. 2012) and non-alcohol fatty liver disease (Loyer et al. 2016). Patients with spontaneous recovery following acute liver failure have been shown to have higher serum concentrations of circulating miR-21 (John et al. 2014). Genetic ablation or the use of miR-21 antisense oligonucleotides have been shown to improve liver inflammation, steatosis, and fibrosis in other acute and chronic liver pathology models (Du et al. 2020; Rodrigues et al. 2017). Several downstream targets of miR-21 have been identified in the liver, although the role of miR-21 in liver injury following APAP intoxication has yet to be explored. We aim to elucidate the role of miR-21, and the molecular mechanisms involved in its effect in APAP mediated hepatotoxicity. We hypothesize that miR-21 genetic ablation will attenuate APAP-induced liver injury and may provide a novel therapy for acute liver injury following APAP.

MATERIALS AND METHODS

Materials

Acetaminophen (4-acetamidophenol) was obtained from Acros Organics (Geel, Belgium). The drug was dissolved in warm saline for i.p. injection to mice.

Animals

Generation of the miR-21 knock-out (miR21KO) mice were previously reported (Lu et al. 2011). Transgenic mice were backcrossed for more than 10 generations in a C57BL/6N (Charles River Laboratories, Wilmington, MA) background. Genotype was assessed by DNA extraction and endpoint PCR using specific primers (Lu et al. 2011). Wild-type (WT) littermates were used as controls. Mice were housed 12:12 light dark cycle under temperature-controlled facility conditions. Mice were maintained on regular chow diet (Teklad 2018, Envigo). All experimental protocols were performed in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animal 8^th edition (2011) and reviewed and approved by the Institutional Animal Care and Use Committee of the University of Mississippi Medical Center.

Experimental design

Eight-week old miR21KO or WT C57BL/6N male mice were used in the study. The animals were subdivided into two groups to be injected with APAP (300 mg/kg BW, i.p.) or saline (10 mL/kg, i.p.) following overnight fasting (12–15 hours). Food was provided after APAP injection and water was not restricted at any time. Mice were euthanized 6 or 24 hours post-injection. The dose of APAP was selected to achieve subtoxic effect and human clinical relevance (McGill and Jaeschke 2019) based on other reports.

Preparation of blood and tissue samples

At the end of the treatment period, mice were euthanized by left ventricle exsanguination and organ removal under isoflurane anesthesia. Liver sections were weighed, flash frozen in liquid nitrogen, and stored at −80°C for further analysis or fixed in 10% neutral buffered formalin and paraffin-embedded (FFPE) for histology, immunohistochemistry, and TUNEL analysis. Blood samples were incubated at 24C for 30 min, centrifuged (1.5xg for 10 min at 4°C) and serum was stored at −80°C until measurements were performed.

Liver Enzymes

Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST) and Lactate Dehydrogenase (LDH) were determined in the serum using an automated chemical analyzer (Vet Axcel, Alfa Wassermann Diagnostic Technologies).

APAP-Cysteine Protein Adducts

APAP-Cysteine protein adducts were measured using high pressure liquid chromatography with electrochemical detection in frozen liver sections as previously described (McGill et al. 2013) at the Analytical Core Laboratory (University of Kansas Medical Center, Kansas City, KS). Values below the limit of detection (LOD) are expressed as LOD divided the square root of two.

GSH and GSSG

Hepatic GSH and GSSG were measured with a modified Tietze method in liver homogenates at the University of Kansas as previously described (McGill and Jaeschke 2015) at the Analytical Core Laboratory (University of Kansas Medical Center, Kansas City, KS). Briefly, frozen tissue was homogenized in 3% sulfosalicylic acid and trapped with 10 mM N-ethylmaleimide. Samples were diluted in 0.01 N HCl, centrifuged, and diluted with 100 mM potassium phosphate buffer. The samples were then assayed using dithionitrobenzoic acid. All data are expressed as GSH-equivalents.

Histology

Liver sections were used for histological analysis. Following excision, sections were fixed overnight in 10% neutral buffered formalin and transferred to 70% ethanol for processing. Tissues were embedded in paraffin, cut into 5 micrometer sections, and followed by hematoxylin and eosin staining or left unstained for mounting with paraffin. Whole section images were acquired with an Olympus DP80 camera mounted on an Olympus BX63 motorized microscope. Images were used for necrosis quantification with ImageJ software (Fiji).

Terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL)

FFPE liver sections were deparafinized and rehydrated for in situ apoptosis detection using a TUNEL detection kit (Trevigen Apoptotic Cell Systems, Gaithersburg, MD). Briefly, following two washes in PBS, samples were pretreated with Proteinase K solution for 20 minutes at room temperature. Labeled dUTP nucleotides were incorporated onto DNA fragments in 37C humidity chamber for 60 minutes, and then incubated with biotinylated anti-BrdU antibody. Samples were then incubated with a DAB substrate solution (Vector Labs, Burlinghame, CA) for up to 10 minutes and followed by methyl green counterstain. Images were acquired with an Olympus DP80 camera mounted on an Olympus BX63 motorized microscope. Apoptosis was quantified with ImageJ (Fiji) and expressed as percentage of total area.

RNA and RT-qPCR

Total RNA was extracted with TRI-Reagent (Cat No: TR118, Molecular Research Center Inc.) following recommended protocol by the manufacturer. RNA samples were resuspended in nuclease-free water and underwent routine DNAse-free treatment following manufacturer’s protocol (Cat No: AM1907, TURBO DNA-free, Ambion). RNA concentrations were quantified using UV absorbance on a Nanodrop spectrophotometer (Nanodrop technologies). Reverse transcription was carried out using Maxima First Strand cDNA Synthesis kit (Cat No: K1641, Thermofisher). Gene expression was performed using hydrolysis probes technology with Fast Advanced Master Mix (Cat No: 4444556, Applied Biosystems) and TaqMan assays (Table 1) or with DNA intercalanting dies using Titanium Taq DNA polymerase (Cat No: 639208, Clontech) and primers listed in Table 2 as previously described (Ball et al. 2017) using a QuantStudio3 cycler (Appled Biosystems). Relative gene expression was normalized to the geometric mean of housekeeping genes (B2M, ActB, Gapdh), standardized to vehicle-treated WT mice (=1), and expressed in arbitrary units (AU).

Table 1.

qPCR Taqman assays

Gene Name	Gene Symbol	Assay ID
BCL2 Apoptosis Regulator	Bcl2	Mm00477631_m1
Beclin 1	Becn1	Mm01265461_m1
BTG Anti-Proliferation Factor 2	Btg2	Mm00476162_m1
CyclinD1	Ccnd1	Mm00432359_m1
Cytochrome P450 Family 2 Subfamily E Member 1	Cyp2e1	Mm00491127_m1
Heat Shock Protein Family A Member 1A	Hspa1a	Mm01159846_s1
Heme oxygenase 1	Hmox1	Mm00516005_m1
Interferon Gamma	Ifng	Mm01168134_m1
Interleukin-1β	Il1b	Mm00434228_m1
Interleukin-6	Il6	Mm00446190_m1
Kruppel Like Factor 6	Klf6	Mm00516184_m1
Microtubule Associated Protein 1 Light Chain 3 Alpha	Map1lc3a	Mm00458724_m1
Microtubule Associated Protein 1 Light Chain 3 Beta	Map1lc3b	Mm00782868_sH
Parkin RBR E3 ubiquitin protein ligase	Prkn (Park2)	Mm01323528_m1
Phosphatase and Tensin Homolog	Pten	Mm00477208_m1
Programmed Cell Death 4	Pdcd4	Mm01266062_m1
PTEN induced Kinase 1	Pink1	Mm00550827_m1
Sequestosome1	Sqstm1	Mm00448091_m1
Serpin Family E Member 1	Serpine1	Mm00435858_m1
Sprouty RTK Signaling Antagonist 1	Spry1	Mm01285700_m1
Sprouty RTK Signaling Antagonist 2	Spry2	Mm00442344_m1
Tumor Necrosis Factor	Tnf	Mm00443258_m1
Wnt Family Member 4	Wnt4	Mm01194003_m1
Wnt Family Member 5A	Wnt5a	Mm00437347_m1
Actin Beta	Actb	Mm02619580_g1
Beta-2 Microglobulin	B2m	Mm00437762_m1
Glyceraldehyde-3-Phosphate Dehydrogenase	Gapdh	Mm99999915_g1

Table 2.

qPCR Primer sets

Gene	Gene ID	Sequence	Annealing Temperature
SMAD family member 7	Smad7	F: GGGCTTTCAGATTCCCAACTT R: AGGGCTCTTGGACACAGTAGA	60
Transforming growth factor beta 1	TGFb1	F: TGCGCTTGCAGAGATTAAAA R: CTGCCGTACAACTCCAGTGA	60
Beta-2-microglobulin	B2m	F: GCTCGGTGACCCTGGTCTTT R: TGTTCGGCTTCCCATTCTCC	65

miRNA PCR

MiR-21 expression was performed in liver samples using RNA extracted and DNAse treated as described above. Reverse transcription was performed with 10 ng total RNA using miRCURY LNA RT Kit (Cat No. 339340, Qiagen); following this, diluted cDNA samples were run with miRCURY LNA SYBR green Master Mix (Cat No: 339347, Qiagen) using hsa-miR-21–5p (YP00204230), hsa-miR-122–5p (YP00205664), or RNU5G (YP00203908) specific primers on a QuantStudio3 cycler (Applied Biosystems). Relative gene expression was normalized to t RNU5G, standardized to vehicle-treated WT mice (=1), and expressed in AU.

Western blot

Liver samples were homogenized in radioimmunoprecipitation assay buffer containing Halt protease and phosphatase inhibitor cocktail (Cat No:78444, Thermo Scientific). Total protein was quantified with the bicinchoninic acid protein assay kit (Cat No:23225, Thermo Scientific). Protein were separated by SDS-PAGE using TGX Stain-Free Gels following UV-activation and transferred to PVDF membranes. Membranes were blocked with 5% nonfat dry milk or 5% bovine serum albumin in Tris-buffered saline containing 0.1% Tween 20 for 1h at room temperature and then incubated overnight at 4°C with the following primary antibodies at the indicated dilutions: Parkin (Prk8; 1:1000; Cell Signaling Technology 4211s); Beclin (Beclin1; 10,000, Abcam ab-207612); LC3 (LC3A/B; Cell Signaling Technologies 12741; 1:10,000); PINK1 (Pink1; 1:10,000, abcam 23707); plasminogen activator inhibitor type 1 (PAI-1; 1:30,000; Abcam ab-182973), phosphatase and tensin homolog (PTEN; 1:30,000; Abcam ab-170941), programmed cell death 4 (PDCD4; 1:30,000; Abcam ab-8448), BCL2 (1:3,000; Abcam ab-692), CYCLIND1 (1:3000, Abcam ab-16663), SQSTM1/p62 (1:3000, CST 23214), PCNA (1:3000, CST#13110), JNK (1:3000, CST #9252), phospho-JNK(Thr183/Tyr185) (1:3000, CST #4668).

Membranes were probed with horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse IgG secondary antibodies (Jackson Immuno Research 111–035-003 and 115–035-003) for 1 h at room temperature. Chemiluminescence detection was performed on a ChemiDoc MP imaging system (Bio-Rad) using SuperSignal West Pico PLUS kit (Cat No: 34578, Thermo Scientific) as substrate. For normalization, membranes were stripped with Restore Western Blot Stripping Buffer (Cat No:21059, Thermo Scientific), incubated overnight at 4°C with anti-GAPDH antibody (1:1,000,000; Cell Signaling Technology 5174), and processed as described above. All quantification analyses were carried out using Image Lab software (Bio-Rad). Data was standardized to vehicle-treated WT mice (=1) and expressed in AU.

Statistics

All data are presented as mean ± SEM. For normally distributed data, a Student’s t-test was performed for comparison of two groups or a two-way analysis of variance (ANOVA) for unequal sample sizes followed by Newman-Keuls post hoc contrasts for all. Assumptions on normality, independence, and equal variance of residuals were checked using Shapiro Walk test, Dubin Watson Test, and Levene test. All gene expression and protein level data is expressed as a fold-change relative to WT vehicle of each time point. Statistical differences were considered significant if p<0.05. For normally distributed data, a Student’s t-test was performed for comparison of two groups with Welch’s correction for unequal variances. Assumptions on normality and equal variance were checked using Shapiro-Wilk and F tests. All analyses were performed in GraphPad Prism 8 software version 8.4.3 (GraphPad Software Inc, La Jolla, CA) or R Statistical Software (R Foundation for Statistical Computing Vienna, Austria).

RESULTS

MicroRNA-21 genetic ablation attenuates APAP-induced biochemical serum marker elevation and APAP-induced necrosis

During the process of hepatic necrosis, the marked rise in the release of cellular contents can be determined by elevation of serum biochemical aminotransferases alanine aminotransferase (ALT) and aspartate transaminase (AST) in addition to lactate dehydrogenase (LDH). All APAP treatment groups showed significant increases of ALT, AST, and LDH compared to control groups. APAP-treated WT mice showed significantly higher serum levels of the hepatic injury marker ALT (20,077±2,084 U/L vs. 14,160±1619 U/L, Fig. 1A) than their miR21KO counterparts after 6 hours. However, no significant differences were observed between genotypes in AST or LDH 6 hours post-APAP overdose (Fig. 1B,C). While no significant difference was observed between genotypes in 24 hour treated animals for ALT (WT: 11,125±1,151 U/L vs. miR21KO: 9,158±1,404 U/L, Fig. 1A), both serum AST (WT: 12,867±2,045 U/L vs. miR21KO: 8,275±1,168, n=12, Fig. 1B) and LDH (WT: 32,850±3791 U/L vs. mir21KO: 20,558±4543 U/L, n=12, Fig. 1C) were significantly reduced in miR21KO animals 24 hours post-APAP. Taken together, these results suggest that liver injury was decreased in miR21KO mice following APAP treatment compared with WT mice.

Figure 1. Effect of microRNA-21 ablation on APAP-induced liver enzymes and necrosis.

Serum (A) Alanine aminotransferase (ALT, U/L) (B) Aspartate aminotransferase (AST, U/L) (C) Lactate dehydrogenase (LDH, U/L) were measured with a chemical analyzer (n=10–12/APAP groups). (D) Left: Quantification of necrotic regions in H&E-stained liver sections (n=11–12/APAP groups) Right: Representative Images. (E) Left: Quantification of DNA fragmentation regions by TUNEL assay (n=8/group) Right: Representative Images. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA (A, B, C) or Student’s t-test (D, E) for each time point. *: p <0.05.

Compared with their WT counterparts, miR21KO mice presented with attenuated liver necrosis represented as relatively lower percentage of necrotic over total areas (6 hr: 54.34±3.4 vs. 47.41±4.6 %, p=0.16, n=11/grp; 24h: 58.84±5.39 vs. 38.95±7.02 %, p=0.034, n=12/grp, Fig. 1D) post-APAP treatment. Liver injury has been shown to peak around 12 hours post APAP administration (Bhushan et al. 2014). In our miR21KO mice, liver injury improves some time between 6 and 24 hours post APAP overdose. Additionally, Terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL) staining allows for the visualization of DNA breaks by biotin-labeled nucleotides attached by the enzyme terminal deoxynucleotide transferase to the 3’-hydroxy terminus of the broken strand. Although traditionally used to measure final-stage apoptotic events, cytosolic staining by TUNEL in APAP hepatotoxicity indicates DNA strand breaks owing to mitochondrial damage and does not reflect apoptosis-induced DNA fragmentation, in part supported by the lack of increased caspase activity (Cover et al. 2005; Jaeschke et al. 2018). Inability to detect apoptosis following APAP is generally believed to be the cause of ATP depletion along with extensive mitochondrial damage (Chao et al. 2018). The latter effect was demonstrated as decreased hepatic DNA fragmentation in miR21KO mice compared with WT mice at 24 hours (58.26±4.5 vs. 42.60±4.9 %, p<0.03, n=8, Fig 1E). There was no difference observed between strains in 6 hour treated mice in concordance with necrosis analysis.

Absence of differences in liver protein adduct formation at 6 and 24 hours but increased recovery of GSH in miR21 KO mice

Once glutathione (GSH) stores have been depleted after APAP overdose, the most toxic metabolite of APAP, NAPQI, is capable of forming protein adducts in the liver which results in vast hepatocellular necrosis. GSH and its oxidized form glutathione disulfide (GSSG) are a redox buffering and detoxification system. The increased ratio of GSSG/GSH has been shown to be representative of mitochondrial oxidative stress. When 70% of GSH has been depleted from liver stores (Mitchell et al. 1973), NAPQI begins to bind to cellular proteins to form cysteine adducts of cellular proteins (Streeter et al. 1984). Expectantly, APAP was shown to decrease total GSH (GSH+GSSG μmol/g liver) relative to WT control and increase the ratio of GSSG/GSH after 6 hours; conversely, total GSH was unchanged compared to controls following 24 hours (Fig. 2A, B). APAP dramatically increased APAP-Cysteine adducts 6 hours after overdose, but miR-21 genetic ablation did not modify APAP-Cysteine adducts at either 6 or 24 hour post APAP treatment (Fig. 2C). However, miR21KO-APAP mice showed a borderline significant trend in increased recovery of GSH in their livers (0.49±0.07 vs. 0.73±0.09 μmol/mL, p=0.069, Fig. 2A) compared with WT-APAP mice 24 hours after APAP overdose, though this advantage did not relay into decreased APAP-Cysteine adducts in miR21KO mice compared with WT mice. In APAP toxicity, mitochondrial proteins are often the target of adduct formation leading to increased mitochondrial oxidative stress. From these results, we cannot conclude that the difference in an improved prognosis for miR21KO mice following APAP is due to differences in formation of protein adducts or increased GSH.

Figure 2. APAP-induced changes to glutathione and protein adduct formation.

(A) Total Glutathione in liver homogenates quantified by glutathione (GSH) and its oxidized form glutathione disulfide (GSSG) (GSH+GSSG μmol/g liver, n=3–9/group). (B) GSSG/GSH, expressed as a percent, n=3–9/group. (C) APAP-Cysteine protein adducts (nmol/mg liver protein, n=3–7/group). (D) mRNA fold-change expression of Cytochrome 450 Family 2 Subfamily E Member 1(Cyp2e1) relative to WT Vehicle group for each time point, 6 or 24 hours standardized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase (N = 5–12/group). Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

Additionally, the expression of one of the enzyme that contributes to production of NAPQI, cytochrome P450 2E1 (Cyp2e1), was decreased in all groups post-APAP after 6 hours (WT: 0.42; miR21KO: 0.63-fold) or 24 hours (WT: 0.13; miR21KO: 0.16-fold, Fig. 2D) relative to WT vehicles. At 6 hours there was a significant increase in Cyp2e1 expression in miR21KO-APAP mice compared with their WT-APAP counterparts. Decreased expression of Cyp2e1 at both 6 and 24 hours are consistent with changes to adduct formation between groups at these times.

Modulation of APAP-induced expression changes in inflammatory and fibrotic markers

Following 6 hour of APAP administration, there were significant decreases in inflammation markers tumor necrosis factor-α (Tnf, WT: 0.49-fold, Fig. 3A) and interferon gamma (Ifng, WT: 0.37 and miR21KO: 0.38-fold) (Fig. 3B) compared with WT vehicles. After 24 hours of APAP administration, APAP increased the expression of Tnf (WT: 9.19 and miR21KO: 7.69-fold, Fig.3A), Il-1β (WT: 6.67 and miR21KO: 6.44-fold, Fig. 3C), and Smad7 (WT: 4.12 and miR21KO: 3.49-fold, Fig. 3D) compared with vehicle controls. Surprisingly, Tgfb1 mRNA expression was only upregulated in WT-APAP (3.63-fold, Fig. 3E) treated mice compared with their vehicle controls and not in miR21KO-APAP mice (2.70-fold, Fig. 3E) at 24 hours. There was a borderline significant interaction term for Smad7 and a significant interaction term for Tgfb1 at 24 hours, indicating the importance of miR-21 in pro-fibrotic processes (Fig. 3E). We did not find significant differences between groups for interlukin-6 expression; however, at both time points treatment had a significant effect by two-way ANOVA (Fig. 3F). At 6 hours after APAP, there were large increases in mRNA expression of cytoprotective Heme oxygenase 1 (Hmox1, WT: 75.49; miR21KO: 83.69-fold, Fig. 3G), which was similar at 24 hours although to a much lower magnitude.

Figure 3. Inflammation and fibrotic markers expression altered by APAP and miR-21 genetic ablation.

(A) Tumor necrosis factor-α (Tnf), (B) interferon gamma (Ifng), (C) interlukin-1β (Il-1b), (D) SMAD family member 7 (Smad7), (E) Transforming growth factor beta 1 (Tgfb1), (F) interlukin-6 (Il-6), (G) Heme oxygenase 1 (Hmox1) mRNA expression expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Gene expression was normalized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase. Jun-amino terminal kinase (JNK) protein expression and phosphorylation status was analyzed by Western-blot as shown in (I) pJNK/JNK and (J) Total JNK. GAPDH was used for normalization and data expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

There were no significant differences between APAP-treated groups in the stress-related activation by phosphorylation to the Jun-amino terminal kinase pathway (JNK) (Fig. 3I); however, APAP treatment had a significant effect on JNK phosphorylation only at 6 hours by two-way ANOVA (Fig. 3I). Neither treatment nor genotype had an effect on levels of total JNK protein (Fig. 3J).

MicroRNA-21 genetic ablation increased markers of hepatic cell regeneration after APAP overdose

In animals recovering from APAP toxicity, molecular markers of cell proliferation can indicate an animal’s hastened ability to regenerate the liver. Cyclin D1 has been suggested as a marker of cell proliferation due to its critical function of cell cycle progression. At 6 hours, only WT-APAP mice showed a decreased Cyclin D1 mRNA (Ccnd1, 0.43-fold, Fig. 4A) and protein (0.16-fold, Fig. 4B) compared with their vehicle counterparts while this decrease was not significant in miR21KO mice. Conversely following 24 hours of APAP, both WT and miR21KO mice showed a significant increase in Cyclin D1 mRNA expression compared with their controls (WT:3.37 vs. 1.00-fold; miR21KO: 4.94 vs. 1.46-fold, Fig. 4A). At the protein level, APAP miR21KO had a greater increase than treated WT mice (7.68 vs. 4.10-fold, respectively, Fig. 4B) and there was a significant effect of strain and of treatment, as well as a borderline significant trend for the interaction between the two factors (F(1,12)=4.05, p=0.067). The decreased expression of mRNA CyclinD1 suggests active inhibition of the cell cycle, as evidenced in WT mice, while at 24 hours the increase in Cyclin D1 expression in miR21KO mice suggests enhanced liver regeneration. At both times, miR21KO mice showed an attenuated response to injury at 6 hours and an enhanced recovery following 24 hours. Based on this finding, we studied proliferating cell nuclear antigen (PCNA) expression, a protein associated with DNA replication. The expression of PCNA followed a similar trend as CYCLIND1 whereby APAP decreased its levels following 6 hours for only WT-APAP-treated animals but this marker was significantly upregulated in 24 hour miR21KO-APAP treated mice (4.87 vs. 1.0-fold, Fig. 4C) compared with WT-treated animals. Similarly to CYCLIN D1 protein at 24 hour, by two-way ANOVA strain, treatment, and interaction term effects were significant. Together our results indicate an enhanced regenerative response in miR21KO mice following 24 hours of APAP-induced injury.

Figure 4. Improved molecular markers of liver regeneration in miR21KO mice following APAP overdose.

WT or miR21KO animals were treated with acetaminophen (APAP) or saline (Veh) for 6 or 24 hours. Cyclin D1 (Ccnd1) (A), BTG anti-Proliferation Factor 2 (Btg2) (D), Heat shock protein family A member 1A (Hspa1a) (E), Kruppel Like Factor 6 (Klf6) (F), Wnt4 (G), and Wnt5a (H) mRNA was quantified by RT-qPCR and standardized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase (N = 5–12/group). CYCLIN-D1 (B) and PCNA (C) protein was quantified by Western-blot and normalized to GAPDH (N = 4/group). Data are expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

The cell cycle inhibitor BTG anti-Proliferation Factor 2 (Btg2) was significantly increased to a similar extent in both APAP treated groups at 6 hours (WT: 57.24 and miR21KO: 61.3-fold, Fig. 4D). Only the 24 hour APAP treated WT mice had a significant increase in Btg2 (10.93-fold) versus vehicle, while there was an attenuation of this increase in the miR21KO mice. Heat shock protein family A member 1A (Hspa1a) gene is important for liver regeneration following partial hepatectomy (Wolf et al. 2014) and serves a primary role in responding to improperly folded proteins after cell stress. After 6 hours of APAP treatment, Hspa1a mRNA expression increased by a large magnitude in both APAP treated groups (WT:408.8-fold; miR21KO:271.9-fold), resulting in a significant effect of the treatment term by two-way ANOVA (Fig. 4E). Hspa1a expression was significantly increased (54.23-fold, Fig. 4E) in 24 hour APAP-treated miR21KO mice compared to vehicles, while the increase in WT mice was not significant. The transcription activator Kruppel Like Factor 6 (Klf6) was similarly upregulated by APAP treatment at both timepoints (Fig. 4F). Taken together, these results suggest miR21KO mice show an enhanced response to APAP induced anti-proliferative effects and enhanced management of cell stress during the liver regeneration period.

Lastly, Wnt4 and Wnt5a expression are enhanced during liver regeneration, which is considered to occur between 12 and 24 hours post-APAP (Bhushan et al. 2014). In the current study, we found that only APAP-treated miR21KO mice have significant increases in Wnt4 expression 24 hours (1.67-fold, Fig. 4G) compared with vehicles. Wnt5a was non-significantly increased in both APAP treated groups compared to vehicles 24 hours post-APAP (Fig. 4H). Together, these results indicate the potential for improved functionality of Wnt/Catenin signaling, which is necessary for regeneration following injury.

MicroRNA-21 targets are altered by APAP

As a number of different canonical targets of miR-21 have been recognized, we analyzed their expression at both the mRNA and protein levels. Bcl2, a member of the pro-apoptotic B-cell lymphoma 2 (Bcl-2) family, was found to be significantly increased at the mRNA level of WT and miR21KO to a similar extent (WT: 1.42 vs. miR21KO: 1.35-fold, Fig. 5A) following 24 hours of APAP albeit unchanged after 6 hours. Bcl2 overexpression has been shown to significantly aggravate necrosis in a BAX specific manner 24 hours after APAP administration, supporting its pro-oxidant role (Adams et al. 2001). However, BCL2 protein was increased by APAP in WT mice 6 hours after APAP significantly more than in APAP treated miR21KO animals (WT: 9.29 vs. miR21KO: 5.59-fold, Fig. 5B). Only at the 24 hour timepoint did strain have a borderline significant trend of an effect (F(1,12)=3.28, p=0.095), and at this time there was again a significant attenuation in the APAP-induced BCL2 protein increase in miR21KO animals (WT: 5.39 vs. miR21KO: 2.08-fold, Fig. 5B).

Figure 5. Alterations of microRNA-21 targets expression with APAP and miR-21 genetic ablation.

B-cell lymphoma 2, Bcl-2 mRNA (A) and protein (B), Phosphatase and tensin homolog (Pten) mRNA (C) and protein (D), and programmed cell death 4 (Pdcd4) mRNA (E) and protein (F) expression expressed as fold-change. mRNA was quantified by RT-qPCR and standardized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase (N = 5–12/group). Protein was quantified by Western-blot and normalized to GAPDH (N = 4/group). Expression of miR-21 (G) and miR-122 (H) are represented as fold-change and normalized to expression of small nuclear RNA U5G (RNU5G). Data are expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

Decreased expression was both phosphatase and AKT signaling pathway inhibitor Phosphatase and tensin homolog (Pten, Fig C,D) and programmed cell death 4 (Pdcd4, Fig. 5 E,F) was observed in both 6 and 24 hours post-APAP. No difference between APAP treated groups was observed at either time point, although both APAP groups had decreased Pten downregulated at 24 hours. Interestingly, at 6 hours there was an attenuation in the APAP-induced decrease in mRNA expression of both Pten (WT:0.72 vs. miR21KO: 0.91-fold, Fig 5C) and Pdcd4 (WT:0.43 vs. miR21KO: 0.79-fold, Fig 5E) in miR21KO mice compared with WT APAP-treated mice. Contrary to the observed Pdcd4 mRNA downregulation, Pdcd4 protein showed a similar upregulation in both APAP-treated groups (WT: 2.12; miR21KO: 2.67-fold) at 6 hours (Fig. 5E,F). Similarly at 24 hours, there was a significant effect of strain and of treatment, as well as a borderline significant trend for the interaction between the two factors in PDCD4 protein expression (F(1,12)=4.09, p=0.066).

Lastly, liver microRNA-21 expression was shown to be biphasic following APAP overdose. Liver miR-21 expression was quantified in all animals in the study; lack of miR-21 expression in all miR21KO mice was confirmed at both time points (Fig. 5G). Interestingly, miR-21 was significantly decreased (0.47-fold) in WT mice after 6 hours of APAP administration but increased (2.31-fold) after 24 hours compared with their vehicle counterparts (Fig. 5G). Additionally microRNA-122, a liver-specific microRNA which constitutes approximately 70% of all liver miRNAs (Jopling 2012), was decreased in both APAP treated groups at 6 hours (WT-APAP: 0.45 vs. WT-Vehicle: 1.00-fold; miR21KO-APAP: 0.29 vs. miR21KO-Vehicle: 0.79-fold) (Fig. 5H). APAP-mediated miR-122 decrease was not observed for either APAP treated group 24 hours after APAP administration.

Specific elevations in the hepatic autophagic response in MicroRNA-21 KO mice after APAP treatment

The Light-Chain 3 mRNA marker Map1lc3a was upregulated in APAP treated miR21KO mice (1.44-fold, Fig. 6A) compared to its vehicle group at 6 hours; this upregulation was consistent between APAP treated groups at 24 hours (WT: 1.25-fold; miR21KO: 1.47-fold, Fig. 6A), resulting in a significant effect of both strain and treatment. APAP upregulated mRNA expression of another Light-Chain isoform autophagy marker Map1lc3b similarly between APAP groups at 6 hours and 24 hours (Fig. 6B). Consistent with these findings at the mRNA level, we observed a significant increase in Light Chain-3 isoform AB-II protein relative to GAPDH in APAP treated mice (Fig. 6C) at 6 and 24 hours, and a significant increase in the LC3AB II/I protein ratio at 6 hours (WT: 3.33-fold; miR21KO: 4.21, Fig. 6D) in APAP mice. Comparing APAP groups at 24 hours, the LC3AB II/I protein ratio was increased in only the miR21KO mice (WT: 1.05 vs. miR21KO: 2.36-fold, Fig. 6D) and all effects including strain, treatment, and interaction terms were significant by two-way ANOVA. This result is indicative of increased autophagy in miR21KO animals. Beclin-1 mRNA expression was only increased in miR21KO mice compared with the WT treated group at 24 hours (1.23 vs. 1.02-fold, Fig. 6E), although this did not result in any changes in protein levels (Fig. 6F). There were no effects in mRNA or protein levels of Beclin-1 across strain or APAP treatment at 6 hours.

Figure 6. Effect of miR-21 genetic ablation in autophagic response following APAP overdose.

mRNA expression of markers involved in autophagy (A) Microtubule Associated Protein 1 Light Chain 3 Alpha (Map1lc3a), (B) Microtubule Associated Protein 1 Light Chain 3 Beta (Map1lc3b), (E) Beclin-1 (Becn1), (G) p62 (Sequestosome1), (I) PTEN induced Kinase (Pink1), (K) Parkin RBR E3 ubiquitin protein ligase (Parkin) were normalized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase. Protein markers (C) Light Chain 3 isoform AB II, (D) Light Chain 3 isoform AB II/ Light Chain 3 isoform AB I (F), BECLIN (H), p62 (J), PINK1, and (L) PARKIN was quantified by Western-blot (n=3–5/group). All protein markers except (D) were normalized to GAPDH. Data are expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

Another marker of autophagy, Sequestosome-1 or p62, a multifunctional cargo protein was increased with APAP-treatment similarly between strains at both 6 and 24 hour timepoints (Fig. 6G). However, at 24 hours of treatment, only the protein expression p62 was significantly elevated in miR21KO mice (18.98-fold, Fig. 6H) compared with vehicle treatment although there was no significant effect of strain by ANOVA. The ubiquitin kinase Pink1, which is involved in mitophagy, was significantly reduced in all APAP treated groups for both time points (Fig. 6I) while, surprisingly, protein levels were not changed (Fig. 6J). Lastly the marker Parkin, which is specifically recruited to promote mitophagy, was downregulated with APAP after both 6 and 24 hours (Fig. 6K). However, Parkin protein showed the opposite response whereby it was increased in WT APAP treated groups at both 6 (21.61-fold) and 24 (13.28-fold) hours compared to their vehicle controls (Fig. 6L). Parkin protein increase was attenuated in miR21KO mice for both time points (miR21KO 6hr: 11.0-fold; 24hr: 4.76-fold). Together, following 24hours of APAP miR21KO show a greater modulation and enhancement of the autophagic response.

Attenuation of APAP-induced increase of PAI-1 mRNA expression and protein in MicroRNA-21 KO mice

Another mechanism by which miR-21 could be mediate additional injury in WT mice is by alteration of a coagulation response. The plasminogen activator system primarily functions to regulate fibrinolytic activity. The plasminogen activator inhibitor 1 (PAI-1) acts to inhibit uPA and tPA, preventing plasminogen to plasmin, and ultimately blocks the breakdown of fibrin matrices following a clot. Following 6 hours of APAP, both miR21KO and WT mice showed increased Serpine1, the gene encoding for PAI-1, mRNA expression relative to WT vehicle (Fig. 7A). At 24 hours after APAP overdose, an even greater increase in mRNA expression was observed in both treated groups (Fig. 7A). Comparing both APAP-treated groups, the miR21KO mice had an attenuated increase in mRNA expression of PAI-1 compared to WT mice (Fig. 7A, WT: 727.46-fold vs. miR21KO: 368.9-fold). Consistent with mRNA expression, at the protein level, only the WT APAP-treated mice had significantly increased levels of PAI-1 6 hours (WT: 44.55-fold) and 24 hours (79.05-fold) post-APAP compared with their vehicle counterparts (Fig. 7B). Additionally, there was a significant effect of treatment, and borderline significant trends for the effects of strain (F(1,12)=4.23, p=0.062) and interaction terms (F(1,12)=4.06, p=0.067) at 24 hours. Compared with WT mice, miR21KO mice showed a drastic attenuation in their response to modulation in the coagulation response, specifically at 24 hours following APAP toxicity.

Figure 7. Attenuation of APAP-induced PAI-1 increase with miR-21 genetic ablation.

WT or miR21KO animals were treated with acetaminophen (APAP) or saline (Veh) for 6 hours or 24 hours. PAI-1 (Serpin Family E Member 1) mRNA (A) and protein (B) were quantified by RT-qPCR and Western blot. Relative expression of mRNA was standardized to the geometric mean of housekeeping genes (GMHK) β-Actin, β2-Microglobulin, and Glyceraldehyde-3-Phosphate Dehydrogenase (N = 5–12/group) and protein was normalized to GAPDH (N = 3–5/group). Data are expressed as fold-change relative to WT Vehicle group for each time point, 6 or 24 hours. Data are expressed as mean ± SEM. Data were analyzed by two-way ANOVA followed by Newman-Keuls multiple comparisons test. *p < 0.05.

DISCUSSION

The liver has a remarkably high regenerative capacity; in mice, regeneration of liver mass can be observed between 5–7days following partial hepatectomy (Michalopoulos 2007). The APAP-induced hepatotoxicity can be considered in different stages: the initiation and progression liver injury and the compensatory regeneration and injury regression (Bhushan and Apte 2019). Timely liver regeneration is important for recovery, and dysregulation of this response by inhibition of the cell cycle is shown to result in greater degree of sustained injury and delayed recovery following APAP (Bhushan et al. 2014). The regenerative phase occurs between 12 and 24 hours following APAP administration (Bhushan et al. 2014). Although regeneration can be proportionate to initial injury, in the current study, we did not see a difference between the initial injury at 6 hours in APAP treated WT or miR21KO mice. MiR-21 expression did not alter several initial mechanisms controlling APAP-induced necrosis progression, including those involved in inflammation, adduct formation, and mitochondrial oxidative stress, since we generally did not observe significant differences between the WT and miR21KO groups at 6 hours. Given that drug-induced progression of injury can lead to a reduced ability to regenerate, and that we did not observe a difference in initial injury, we can conclude that miR-21 may play a role in delaying regeneration during the proliferative stage of recovery. In our study, hepatotoxicity progressed injury after 6 hours which was then followed by enhanced regeneration, cessation of necrosis progress, and sustained liver repair in miR21KO mice after 24 hours. The findings of this study indicate that miR-21 genetic ablation may be protective against APAP-induced liver injury by 1) enhancing the ability of the liver to regenerate by promoting the cell cycle entry and proliferation, 2) hastening the induction of autophagy, 3) modulating the coagulation pathway to balance the hypofibrinolytic state.

The present findings indicate that several mechanisms that are involved in liver injury repair response following 24 hours of APAP treatment (regenerative phase) were enhanced in miR21KO mice. Wnt/βCatenin signaling is important following regeneration of liver injury (Apte et al. 2009) and plays a specialized role in metabolic zonation of the liver (Torre et al. 2011b). CyclinD is a pro-proliferative target of β-catenin signaling in hepatocytes (Torre et al. 2011a), and Catenin has been shown to bind to the CyclinD promoter following APAP (Bhushan et al. 2014). Our data show that following 24 hours of APAP dosing, Wnt4 mRNA was increased in APAP-treated miR21KO but not WT-treated mice (Fig. 4G). Promotion of the Wnt/βcatenin signaling is a possible mechanism by which regeneration of Cyclin D1 was shown to increase nearly 8-fold in 24 hour APAP-treated miR21KO mice (Fig. 4B). There is extensive crosstalk between miRNAs and Wnt signaling pathways (Peng et al. 2017), including miR-21. Interestingly, miR-21 has been shown both to be upregulated by β-catenin via STAT3 in glioma cells (Han et al. 2012) and to display control over Wnt signaling via the negative regulator Dickkopf2 (DKK2) in oral cancer (Kawakita et al. 2014). Potential inhibition of cell cycle entry in WT mice could delay the initiation of regeneration, which was also evidenced by their lack of change in PCNA after 24 hours, a time at which miR21KO showed an approximate 5-fold increase in PCNA protein (Fig. 4C). Proliferating Cell Nuclear Antigen (PCNA) regulates the sliding clamp mechanism to ensure that DNA polymerase can function during replication and also helps repackage the nucleosome of newly synthesized DNA by attracting histone chaperones, and increased expression indicates DNA replication. At 24 hours post-APAP at the dose used in our studies, most cells are in the G1 phase of the cell cycle, while many have progressed to the S phase (Bhushan et al. 2014). Surprisingly, PCNA protein was not upregulated 24 hours post-APAP in WT mice in contrast to previous reports (Bajt et al. 2008; James et al. 2003). APAP-induced PCNA protein expression is dose-dependent, with APAP high doses (600 mg/kg) delaying and reducing PCNA expression compared with APAP low doses (300 mg/kg) in C57BL/6J mice (Bhushan et al. 2014). There is C57BL/6 substrain differential sensitivity to APAP hepatotoxicity (Duan et al. 2016). The C57BL/6N substrain used in our studies has been reported to be more susceptible to APAP hepatotoxicity than the widely used C57BL/6J substrain (Duan et al. 2016). We speculate that the low dose APAP (300 mg/kg) in mice with the C57BL/6N genetic background causes a delay and attenuation in PCNA protein expression similar to a high dose APAP (600 mg/kg) in C57BL/6J mice. PCNA protein induction delay and attenuation in C57BL/6N may have precluded us from observing an increase in PCNA as we only pursued our studies for up to 24 hours. Longer follow-up studies (48–72 hours) may allow testing of this hypothesis. Another mechanism by which miR-21 may exhibit control over regeneration is in the context of mRNA expression of TGF-β. There was a significant effect of the interaction of strain and treatment on Tgf-β mRNA expression at 24 hours (Fig. 3E). TGF-β inhibition has been shown to benefit liver regeneration in mice treated with APAP by reduced senescence and increased regeneration (Bird et al. 2018). The expression of miR-21 may influence the regenerative capacity following APAP by this potential mechanism.

Evidence of a proproliferative role of miR-21 has been shown in cancers, as human solid tumors are shown to overexpress miR-21 (Fu et al. 2011; Volinia et al. 2006). In WT mice, hepatic expression of miR-21 was decreased following 6 hours of APAP treatment, but after 24 hours was increased relative to vehicle controls (Fig. 5G). Our findings are consistent with other studies showing that miR-21 expression is upregulated during the proliferative phase following partial hepatectomy (Marquez et al. 2010). Our findings suggest that ablation of miR-21 is protective because increased regenerative markers and decreased necrosis was observed following 24 hours post-APAP, yet these findings stand in opposition to oncogenic nature of miR-21 reported in different tumors (Krichevsky and Gabriely 2009). One other study has investigated time course changes to circulating miR-21 expression compared to vehicles in an APAP model but they did not find significant changes (Park et al. 2016). However, that study was performed in rats, which unlike mice used in this study, has been shown to be a poor model of APAP-induced liver injury due to the differences in mechanisms and severity of liver injury than those of humans (Jaeschke et al. 2014). In human serum, a rise in specific microRNAs, including miR-21–5p, has been identified as indicative of APAP poisoning that is clinically distinguishable from microRNA profiles observed in other types of liver injury (Carreiro et al. 2019; Ward et al. 2014). The role of miR-21 in regeneration following partial hepatectomy has been investigated (Marquez et al. 2010); however, our study is the first to examine the role of miR-21 mediated liver regeneration following APAP overdose. In several hepatotoxin-induced liver regeneration models, including APAP, the exposure itself causes functional changes that specifically impair regenerative functions (Abu Rmilah et al. 2019). For instance, miR-21 expression was upregulated during liver regeneration following partial hepatectomy with chronic ethanol consumption, yet these animals still showed inhibited cell proliferation (Dippold et al. 2012). Upon further examination, inhibition of miR-21 using a locked nucleic acid (LNA)-DNA antisense oligonucleotide yielded improvement in liver regeneration (Juskeviciute et al. 2016). These findings stand in contention of the pro-proliferative nature overwhelming the literature on miR-21 in tumors, but is consistent with our 24 hour APAP treated WT mice. In this way, the proliferative role of miR-21 may be tissue-specific and it may have anti-proliferative roles in non-cancerous cells. Our WT mice showed an approximately 3-fold increase in expression of miR-21 compared with non-treated WT mice at 24 hours post-APAP. Considering recovery at this time point, these results indicate that miR-21 may function to inhibit regeneration rather than to enhance proliferation. Lastly, the upregulation of miR-21 in WT mice observed at 24 hours could also be mediated by a pro-fibrotic feedback loop from APAP-induced increase in TGF-β expression (Fig. 3E). TGF-β has been shown to positively regulate miR-21 (Davis et al. 2008; Li et al. 2013) and may act in a positive feedback to enhance miR-21 expression in our APAP WT mice. In the current study, during the recovery stage following liver injury, WT mice showed decreased expression of proliferative markers, thus the initiation of regeneration is probably delayed by mechanisms controlled by miR-21. Together our results suggest that genetic ablation of miR-21 did not alter mechanisms involved in improvement of initial injury progression compared to WT mice, given that injury and inflammatory pathways were not reduced at 6 hours. Genetic ablation of miR-21, however, also did not impair regenerative capacity, unlike WT mice, and regeneration mechanisms were even enhanced following a greater time post-APAP.

Autophagy, namely macroautophagy and chaperone-mediated autophagy, is the fundamental process of protein degradation by which normal cell function is maintained following initiation of autophagosome formation and lysosomal fusion. The regulation of the autophagic network is complex and involves specific signaling regulatory pathways, including mTOR and AMPK, in order to ultimately alter cellular energy homoeostasis by ridding the body of damaged or misfolded proteins and by providing additional nutrients by cellular breakdown (Kim et al. 2011; van Oosten-Hawle et al. 2016). Autophagy has been implicated in normal cell processes including cell repair (Kaushik and Cuervo 2006), differentiation (Clarke and Simon 2018), and cell defense (Deretic 2011). Inherent in acetaminophen toxicity, the autophagic pathway has been implicated as a major regulator of sustained necrosis and survival and this process serves in a protective role. Specifically, selective autophagy by the removal of damaged mitochondria has been shown to benefit the repair process following APAP intoxication (Chao et al. 2018). In the current study, PDCD4 was increased by APAP at 6 hours and decreased following 24 hours post-APAP (Fig. 5D). Both LC3-II and p62, proteins that are associated with autophagy initiation and degradation, and are involved in the mechanism of Pdcd4 degradation in cells (Manirujjaman et al. 2020) which was consistent with the inverse relationship with Pdcd4 at 24 hours (Fig 5 D;G). Specifically, p62 makes up the sequestosome and works as a cargo protein in conjunction with LC3-II to allow for a selected authophagy in response to APAP. Induction of the p62/keap1/Nrf2 antioxidant pathway following APAP serves in a protective role; our results indicate that miR-21 may prevent p62-specific upregulation after 24 hours of APAP to interfere with selected autophagy aiding the antioxidative response (Shen et al. 2018).

The phosphastase and tensin homolog (PTEN)-induced kinase 1 (PINK1)-Parkin pathway is a well-characterized selective autophagy pathway for removal of damaged mitochondria (Williams and Ding 2018) which has been shown to be required following APAP treatment (Wang et al. 2019). Interestingly, we did not find significant changes to PTEN or PINK1 between our miR21KO and WT mice despite being recognized as a miR-21 target. In the healthy mouse liver, it has been shown that miR-21 specifically reduces binding to polysome-associated target mRNAs which can affect the silencing activity (Androsavich et al. 2012). Parkin, the E3 ubiquitin ligase recruited by PINK1, was significantly upregulated in APAP-treated WT mice at both 6 and 24 hours compared with APAP-treated miR21KO mice (Fig. 6L). In a Parkin knockout model of APAP, levels of LC3-II were increased and the number of autophagasomes were comparable to WT mice (Williams et al. 2015). This data suggests that autophagic processes in APAP can function independent of Parkin. Considering there were no basal level changes between vehicle groups in Parkin levels, our results suggest miR21KO mice had increased autophagy following APAP overdose without the same considerable increase in Parkin as their WT counterparts. Importantly, ubiquitination can be coordinated by other E3 ligases to induce mitophagy other than Parkin in APAP hepatotoxicity (Chao et al. 2018). There is also a potential that miR-21 elevations in WT APAP treated mice involve improved compensatory mitochondrial-specific authophagy, which lead to increased Parkin. Lastly, in addition to its crucial involvement in liver regeneration in a non-toxicant induced model, heat shock proteins may crucially influence the cellular response of misfolded proteins for autophagic response following APAP treatment. We show that heat shock protein (Hspa1a) mRNA expression was significantly increased in miR21KO mice at 24 hours (Fig. 4E) in addition to modulation of some autophagy-specific processes. Together these results indicate that miR-21 may retard the autophagic breakdown of damaged organelles and misfolded protein response, which may impair the liver’s regenerative ability following APAP overdose.

We are the first to report the potential influence of miR-21 to the fibrinolytic pathway following APAP overdose. Plasminogen Activator Inhibitor 1 (PAI-1) can modulate pathways involved in regeneration in addition to its more typical role in fibrinolysis regulation. Although it inhibits both urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA) to block fibrinolysis, PAI-1 has also been shown to increase in response to various types of stress (Beier and Arteel 2012). Hepatocyte Growth Factor (HGF) has important roles in injury related regeneration in the liver (Fukushima et al. 2018). Pro-hepatocyte growth factor must be proteolytically activated in order to have a biological effect, and urokinase-type plasminogen activator (uPA) has been shown to be involved in this activation (Fukushima et al. 2018), which is regulated by PAI-1. In APAP treated mice there is an increase in PAI-1, which is known to result in a hypofibrinolytic state. The blood coagulation cascade controls the formation and lysis of clots following APAP overdose. In the current study, we show an attenuation in the APAP-mediated increase in PAI-1 in miR21KO mice (Fig 7 A,B). Previous reports have shown that PAI-1 is involved in the regulation of miR-21 in muscle (Ardite et al. 2012). We did not measure fibrin deposition or thrombin in the liver in this study; however, studies on APAP-induced liver injury have shown both of these to be increased (Ganey et al. 2007; Pant et al. 2018). PAI-1 plays a critical role in this fibrinolytic balance as evidenced by increased APAP-induced liver injury following APAP in PAI-1 deficient mouse models (Pant et al. 2018). Our results show that miR-21 can directly modulate the elevated levels of both PAI-1 mRNA and protein, and its ablation significantly aids in limiting the extent of this increase to ameliorate the hypofibrinolytic state.

In summary, the current research indicates three pathways that may be beneficially modulated by downregulation of miR-21 in APAP overdose including cell regeneration, autophagy, and coagulation homeostasis. Although miR-21 genetic ablation did not seem to alter the mechanisms involved in the improvement of initial injury progression, our study suggests that it may enhance the regenerative capacity following a greater time post-APAP. However, longer follow-up studies are needed to prove further that miR-21 genetic ablation increases liver regenerative capacity post-APAP. Antisense oligonucleotide (ASO)-based therapies to target miR-21 are currently in clinical trials (Huang et al. 2020). Despite the potential for unintended off-target effects, the use of this therapy in APAP has been explored in animal models for other miRNAs (Zhang et al. 2021). Given the results of our study, miR-21 specific ASOs may be a useful therapeutic option to mitigate late-stage APAP hepatoxicity.

ABBREVIATIONS

ALF

acute liver failure

ALT

alanine aminotransferase

APAP

Acetaminophen

AU

arbitrary units

AST

aspartate transaminase

Bcl-2

B-cell lymphoma 2

Ccnd1

Cyclin D1

Cyp2e1

cytochrome P450 2E1

GSH

glutathione

GSSG

glutathione disulfide

Hspa1a

Heat shock protein family A member 1A

JNK

Jun-amino terminal kinase pathway

Klf6

activator Kruppel Like Factor 6

LDH

lactate dehydrogenase

miRNAs

MicroRNAs

miR-21

MicroRNA-21

miR21KO

miR-21knockout

NAC

N-acetylcysteine

NAPQI

N-acetyl-p-benzoquinoneimine

PAI-1

Plasminogen Activator Inhibitor 1

PCNA

proliferating cell nuclear antigen

Pdcd4

programmed cell death 4

PINK1

phosphastase and tensin homolog (PTEN)-induced kinase 1

Pten

Phosphatase and tensin homolog

Tnf

tumor necrosis factor-α

tPA

tissue-type plasminogen activator

TUNEL

Terminal deoxynucleotidyl transferase dUTP nick end labeling assay

uPA

urokinase-type plasminogen activator

WT

wild-type

DATA AVAILABILITY

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.