Parts of the study were funded by the German Research Foundation (Cluster of Excellence REBIRTH; EXC 62/1 awarded to Tobias Cantz and SFB 738 awarded to Michael Ott) and the German Ministry of Education and Research (ITCF-01GU0618, awarded to Toni Cathomen. Amar Deep Sharma receives grant support from HILF, Hannover Medical School.
Potential conflict of interest: Nothing to report.
Death receptor-mediated apoptosis of hepatocytes contributes to hepatitis and fulminant liver failure. MicroRNAs (miRNAs), 19-25 nucleotide-long noncoding RNAs, have been implicated in the posttranscriptional regulation of the various apoptotic pathways. Here we report that global loss of miRNAs in hepatic cells leads to increased cell death in a model of FAS/CD95 receptor-induced apoptosis. miRNA profiling of murine liver identified 11 conserved miRNAs, which were up-regulated in response to FAS-induced fulminant liver failure. We show that ectopic expression of miR-221, one of the highly up-regulated miRNAs in response to apoptosis, protects primary hepatocytes and hepatoma cells from apoptosis. Importantly, in vivo overexpression of miR-221 by adeno-associated virus serotype 8 (AAV8) delays FAS-induced fulminant liver failure in mice. We additionally demonstrate that miR-221 regulates hepatic expression of p53 up-regulated modulator of apoptosis (Puma), a well-known proapoptotic member of the Bcl2 protein family. Conclusion: We identified miR-221 as a potent posttranscriptional regulator of FAS-induced apoptosis. miR-221 may serve as a potential therapeutic target for the treatment of hepatitis and liver failure. (HEPATOLOGY 2011;)
Hepatocytes are highly sensitive to death receptor-mediated apoptosis.1, 2 The extrinsic apoptotic pathways in hepatocytes involve receptors such as FAS, tumor necrosis factor (TNF), and TNF-related apoptosis inducing ligand (TRAIL).3, 4 FAS receptors and downstream apoptotic events have been implicated in hepatitis including hepatitis B and hepatitis C virus infection, fulminant liver failure, nonalcoholic fatty liver disease, and hepatocellular carcinoma (HCC).4 A number of pro- and antiapoptotic proteins including caspases mediate hepatocyte apoptosis, all of which are regulated at the transcriptional and/or translational level.4 Among the posttranscriptional regulators, microRNAs (miRNAs) are new players, which inhibit protein translation.5-7 One of the first reports demonstrating the involvement of miRNAs in apoptosis came from studies using the model organism Drosophila melanogaster, in which two miRNAs, miR-14 and Bantam, were reported to control apoptosis.8, 9 A number of reports describe a role for miRNAs in hepatic apoptosis.10-12 However, their direct involvement in apoptosis of primary hepatocytes during hepatitis and fulminant liver failure has not yet been elucidated in detail.
In the current study we aimed to evaluate the role of miRNAs in apoptosis during fulminant liver failure in mice. Our results indicate that miRNAs are important regulators of apoptosis. Furthermore, overexpression of miR-221 protects hepatocytes from apoptosis and delays fulminant liver failure in mice.
AAV8, adeno-associated virus serotype-8; DGCR8, DiGeorge syndrome critical region gene 8; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; miRNA, microRNA; PTEN, phosphatase and tensin homolog; Puma, p53 up-regulated modulator of apoptosis; shRNA, short hairpin RNA; UTR, untranslated region.
Materials and Methods
Animal experiments were performed according to the guidelines of Hannover Medical School, Germany. BALB/c mice were purchased from Charles River Laboratories (Germany).
In Vitro Knockdown of DiGeorge Syndrome Critical Region Gene 8 (DGCR8 and DROSHA).
Hepa 1-6 mouse hepatoma cells (American Tissue Culture Collection, ATCC, Manassas, VA) were cultured in Dulbecco's modified Eagle's medium (DMEM) (PAA Laboratories) supplemented with 10% fetal bovine serum (FBS) (PAA), L-glutamine (PAA), and penicillin/streptomycin (PAA). Cells were transduced with retrovirus expressing short hairpin RNA (shRNA) against DGCR8 and DROSHA (Addgene plasmid) as described.13 After transduction, shRNA-expressing cells were selected in medium supplemented with puromycin (Invitrogen) at a concentration of 1 μg/mL. Loss of DGCR8 and DROSHA was confirmed by western blots.
Primary mouse hepatocytes were isolated by two-step collagenase (Roche) perfusion followed by Percoll (Sigma) density gradient centrifugation as described.14 Purified mouse hepatocytes were cultured in Primaria dishes (BD Labware) in the presence of hepatocyte basal medium supplemented with hepatocyte single quotes (Lonza). Targefect hepatocyte reagent (Targetingsystems) for plasmid transfection and Targefect F2 reagent (Targetingsystems) were used for small interfering RNA (siRNA) transfection into primary hepatocytes.
FAS and TNF-α-Induced Apoptosis.
For in vitro apoptosis induction, a final concentration of 0.5 μg/mL Jo2 antibody was added to the hepatocyte culture medium. In vitro apoptosis by TNF-α was induced as described.15In vivo apoptosis was induced by intraperitoneal injection of 0.4 μg/g body weight Jo2 antibody (BD Pharmingen) in 8 to 10-week-old BALB/c mice. Mice were sacrificed at indicated timepoints. Liver tissues were harvested and immediately snap-frozen in liquid nitrogen and fixed in 4% paraformaldehyde (Sigma).
Serum Parameter Analysis.
Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured as described.16
Western blots were performed as described.17 DGCR8 antibody (Abcam, dilution 1:250), p53 up-regulated modulator of apoptosis (PUMA) (Abcam, 1:1000), p27 (BD Pharmingen, 1:500), phosphatase and tensin homolog (PTEN) (Cell Signaling 1:1000), FAS (Santa Cruz, 1:250), and Tubulin (Sigma, 1:1,000) were used.
Liver tissues were fixed in 4% paraformaldehyde for 4 hours at 4°C, washed in phosphate-buffered saline (PBS), and embedded in OCT to prepare frozen blocks. The 7-μm sections were cut and air-dried for 20 minutes before staining. A terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay (Millipore) was performed according to the manufacturer's guidelines.
miRNA profiling was performed and analyzed by FeBIT (Heidelberg, Germany). Briefly, total RNA was isolated from liver tissue using the miRNeasy kit (Qiagen). Following on-column DNase treatment, total RNA quality was determined by Nanodrop (NanoDrop Technologies) and Bioanalyzer (Agilent). For each array, 250 ng RNA was suspended in FeBIT's proprietary miRNA Hybridization Buffer (25 μL per array). Hybridization was performed for 16 hours at 42°C using the GeniomRT-Analyzer. Data analyses and presentation (in Table 1) were performed as described.18
Table 1. MicroRNA Profiling of Apoptotic Liver
Differentially Regulated MicroRNAs in Response to Jo2-Induced Apoptosis in Mouse Liver
6 Hours After Jo2 Injection
12 Hours After Jo2 Injection
*Denotes the star miR-135a, generated from passage strand.
One μg and 10 ng total RNA was used for first-strand complementary DNA (cDNA) synthesis for gene expression analysis and miRNA expression, respectively. The Taqman miRNA RT kit (for miR-cDNA synthesis), Taqman Universal Real Time PCR kit (for miRNA quantitative reverse transcription [qRT]-PCR), and SYBR green PCR master mix (for gene expression analysis) were purchased from Applied Biosystems. Gene expression was normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Data were analyzed according to the delta-delta Ct method. The primer sequences were: Puma forward: 5′-CTGTA TCCTGCAGCCTTTGC-3′, Puma reverse: 5′-ACGGG CGACTCTAAGTGCT-3′, GAPDH forward: 5′-ATG GCCTTCCGTGTTCCT-3′, and GAPDH reverse: 5′-CGGCACGTCAGATCCA-3′.
AAV8 vectors were prepared as described.19 293T cells, at 50%-60% confluency, were transfected with two plasmids by the calcium phosphate method. The plasmid pDP8.ape (PlasmidFactory, Germany) was used to provide necessary genes for AAV8 production such as rep/cap, E2A, E3, and E4 genes. The minimal transthyretin (Ttr) promoter20 was kindly provided by Dr. Weidong Xiao (Temple University, PA). To generate pD.AAV.Ttr.Cre plasmid Ttr Promoter was cloned into pD.cmvsHAnlsCre, which was previously digested with Kpn1 and BspE1 to remove CMV promoter. For constructing pD.AAV.Ttr.miR-221 plasmid miR-221 was PCR-amplified from mouse genomic DNA before cloning into pD.AAV.Ttr.Cre plasmid using forward primer 5′-CAGGCTGAACAT CCAGGTCT-3′ and reverse primer 5′-TGGCTCCTA GAAAAGTTGACTC-3′. Then 72 hours after transfection with pDP8.ape and transgene plasmids pD.AAV.Ttr.Cre or pD.AAV.Ttr.miR-221 providing Cre recombinase or miR-221, cells containing virus were harvested and AAV8 was purified. The titer was determined by qRT-PCR using primers spanning the region of the Ttr promoter. Ttr Forward primer 5′-TCAGCTT GGCAGGGATCAG-3′ and Ttr reverse primer 5′-GAC GGCTTCTCCTGGTGAAG-3′.
Detection of Cell Viability, Caspase-3/7 Activity, and Annexin V Staining.
Primary hepatocytes were grown on Primaria dishes in Hepatocyte Basal Medium (HBM, Lonza). WST assay (Roche) for cell viability and caspase-3/7 activity assay (Promega) for apoptosis were performed according to the manufacturer's instructions. APC-conjugated Annexin V (eBioscience) staining was performed according to the manufacturer's protocol. Propidium iodide (PI) was added just before data collection at FACSCalibur.
Luciferase Reporter Assay.
Mouse 3′ untranslated region (UTR) was amplified from genomic DNA using forward primer 5′-GAGTCCGCTAGCGTGCC TACACCCGCCCGGGG and reverse primer 5′-GAT GTAGTCGACCACTGTTCAATCTGATTT-3′. Six hours after seeding, hepatocytes were transfected with miRNA mimics or inhibitors (Dharmacon) followed by transfection of miR-glo-PUMA UTR plasmid or control plasmid at 18 hours after seeding cells. Twelve hours after reporter construct transfection, cells were lysed and Dual-Glo luciferase assay system (Promega) was used to detect the firefly luciferase activity on a luminometer (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity. Position 2, 3, and 4 of seed sequences were mutated in Puma 3′ UTR and mutated plasmid was created using the QuickChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies). Data are shown as percentage activity by setting the control to 100%.
In Vitro Knockdown of PUMA, PTEN, and BMF.
The 1 × 106 primary hepatocytes were seeded in each well of 6-well Primaria dishes. siRNA against PUMA, PTEN, and BMF were purchased from Qiagen. Hepatocytes were transfected with siRNA 1 and 2 using Targefect F2 (Targeting systems). Then 48 hours after transfection hepatocytes were trypsinized and cell lysate was prepared for western blot.
Cotransfection of Target Protectors and miR-221 Mimic.
miRNA target protectors for PUMA, PTEN, and BMF (Qiagen) were cotransfected with miR-221 mimic using the Targefect F2. Then 24 hours after cotransfection apoptosis was induced by supplementing the culture with Jo2.
Significance was determined with a two-tailed Student's t test. P < 0.05 was considered significant. A log-rank test was used to compare two Kaplan-Meier survival curves to obtain significance in Fig. 4A.
To investigate the role of miRNAs in apoptosis, we first generated a global loss of miRNAs in Hepa 1-6 mouse hepatoma cells by knockdown of DGCR8, an essential component of the microprocessor complex for miRNA biogenesis.21 For stable knockdown we generated a retroviral vector, which expresses shRNA against DGCR8. After transduction and selection in the presence of puromycin, we observed an efficient loss of DGCR8 protein in Hepa 1-6 (Fig. 1A) (henceforth these cells will be referred as shDGCR8 cells). As a result of DGCR8 deficiency, levels of miR-122, the most abundant miRNA in hepatocytes, were significantly decreased in these cells (Fig. 1B). In addition, miR-21 and miR-221 were down-regulated in shDGCR8 cells (Fig. 1B), confirming that the loss of DGCR8 resulted in a reduction of miRNA expression in these cells.
We next sought to determine the effect of global loss of miRNAs on apoptosis. We treated normal Hepa 1-6 and shDGCR8 cells with a FAS-agonist antibody (anti-CD95, clone Jo2) to induce apoptosis. Jo2 antibody causes rapid apoptosis in hepatocytes in vitro as well as in vivo and induces fulminant liver failure in mice.22, 23 At 12 and 24 hours after Jo2 treatment, cell viability was measured by WST assay. We found that loss of DGCR8 and thus global loss of miRNAs in shDGCR8 cells sensitizes them to FAS-induced cell death (Fig. 1C). WST assay determines the number of viable cells by measuring the activity of mitochondrial dehydrogenase. In order to confirm whether lower cell viability in shDGCR8 cells was indeed a result of increased apoptosis, and not merely due to reduced proliferation, we measured caspase-3/7 activity, which is an indicator of apoptosis. Increased caspase-3/7 activity confirmed that lower cell viability in shDGCR8 cells was indeed due to increased apoptosis (Fig. 1D). Furthermore, staining with Annexin V, another marker for detection of apoptosis also showed a higher number of Annexin V-positive shDGCR8 cells by FACS analysis (Fig. 1E). Cells in early apoptosis (Annexin V-positive but PI-negative) as well as in late apoptosis (Annexin V-positive and PI-positive) contributed to the high number of apoptosis in shDGCR8 cells.
Next we sought to determine whether another model of global miRNA inhibition also leads to increased FAS-induced apoptosis in Hepa 1-6 cells. We therefore knocked down DROSHA, another component of the microprocessor complex, in Hepa 1-6 cells, which resulted in reduction of miRNA levels (Supporting Fig. S1a,b). Basal level of apoptosis in DROSHA or DGCR8 knockdown cells was similar to control cells (Supporting Fig. S1c). After induction of apoptosis by FAS we found that DROSHA knockdown, similar to DGCR8 knockdown, also leads to increased apoptosis in Hepa 1-6 cells (Supporting Fig. S1d,e). Thus, global loss of miRNAs in hepatoma cells sensitizes them to FAS-induced apoptosis in vitro.
To investigate the significance of miRNAs in fulminant hepatic failure, we injected a lethal dose of Jo2 antibody in BALB/c mice intraperitoneally. We administered 0.4 μg/g body weight of Jo2 antibody, a dose which has previously been reported to cause 100% mortality in mice due to acute apoptotic cell death.24 First, we documented the hepatic damage by analyzing serum ALT and AST. We found markedly elevated levels of ALT and AST after Jo2 injection, indicating severe liver injury at 6 hours and 12 hours (Supporting Fig. S2a). TUNEL staining of liver sections showed moderate and extensive apoptosis at 6 hours and 12 hours, respectively (Supporting Fig. S2b). On the basis of ALT, AST levels, and TUNEL staining we selected liver samples for miRNA expression profiling from the 0-hour timepoint as control livers, 6-hour timepoint for early apoptosis, and 12-hour timepoint for advanced stage apoptosis beyond which mice start to die. miRNA microarrays enabled us to detect the expression of 600 miRNAs in the liver samples (miRBASE 13.0). We found that 5 and 32 miRNAs were significantly differentially regulated at 6 hours and 12 hours, respectively, after FAS-induced apoptosis in the liver (Table 1). We validated the differentially regulated miRNAs by qRT-PCR and found that most miRNAs showed the same expression pattern as in our miRNA profiling (Supporting Fig. S2c).
For functional analyses we selected 11 significantly deregulated miRNAs that were conserved between mouse and human (Fig. 2A). To analyze direct effects of miRNAs on apoptosis we aimed to transfect primary hepatocytes with miRNA mimics and miRIDIAN inhibitors for gain and loss of miRNA function experiments, respectively. Using liposome complexed reagents, up to 80% of primary mouse hepatocytes were successfully transfected (Supporting Fig. S2d). First, we transfected hepatocytes with mimics of each of the conserved and differentially regulated miRNAs. After confirming the overexpression of miRNAs (data not shown), we investigated their effect on FAS-induced apoptosis in hepatocytes. By cell counting, we found that miR-221 inhibited Jo2-induced cell death most prominently in comparison to other miRNAs (Fig. 2A). We therefore focused on miR-221 for further analyses. WST assay also revealed that miR-221 overexpression followed by Jo2-treatment led to an increase in cell survival, whereas inhibition of miR-221 decreased cell survival (Fig. 2B). In order to determine whether increased hepatocyte survival was due to inhibition of apoptosis we measured the activity of caspase-3/7. We found decreased caspase-3/7 activity in hepatocytes transfected with miR-221 mimic, whereas increased caspase-3/7 activity was seen in hepatocytes transfected with miR-221 inhibitor (Fig. 2C). Together, cell viability assay and caspase-3/7 assay provide evidence that miR-221 protects cultured hepatocytes from Jo2-induced apoptosis.
We then addressed the question whether overexpression of miR-221 can rescue the observed high sensitivity of shDGCR8 cells to FAS-induced apoptosis. To this end, we transfected shDGCR8 cells with miR-221 mimic followed by FAS-induced apoptosis. By Annexin V staining and caspase-3/7 assay we found that shDGCR8 cells transfected with mimics of miR-221 had reduced apoptosis (Fig. 2D,E). Therefore, miR-221 can partially rescue cells from FAS-induced apoptosis and hence partially compensate for global loss of miRNAs in shDGCR8 cells.
Next we sought to investigate whether overexpression of miR-221 can protect mice from FAS-induced fulminant liver failure. To overexpress miRNAs in vivo, we used AAV8. AAV8 has been shown to transduce up to 100% hepatocytes when injected intravenously in mice.25 Moreover, a successful miRNA delivery by AAV8 and regression of tumors has recently been shown in a mouse model of hepatocellular carcinoma.26 We therefore prepared AAV8-Ttr-miR-221 vector expressing miR-221 under the control of a hepatocyte-specific transthyretin (Ttr) promoter and an AAV8-Ttr-Cre (control) vector expressing Cre recombinase (Fig. 3A). In vivo hepatocyte transduction efficiency was assessed by injecting 1 × 1011 viral particles of AAV8-Ttr-Cre vectors into the tail vein of ROSA26 reporter mice,27 which contains a floxed stop codon upstream of the β-galactosidase reporter gene. Efficient hepatocyte transduction was confirmed by X-gal staining (Fig. 3B). We then injected 1 × 1011 viral particles of either AAV8-Ttr-Cre or AAV8-Ttr-miR-221 vector intravenously into BALB/c mice. AAV8-injected mice showed normal histology (Fig. 3B) and normal levels of transaminases (data not shown). Four days after AAV8 injection, we detected 8-fold higher expression of miR-221 in purified hepatocytes from mice injected with AAV8-Ttr-miR-221 compared to control hepatocytes from AAV8-Ttr-Cre-injected mice (Fig. 3C). Consistent overexpression of miR-221 could also be detected up to 2 weeks after AAV8 injection (data not shown).
Four days after AAV8 injection in mice, we administered Jo2 antibody intraperitoneally and observed the effect of miR-221 overexpression on Jo2-induced apoptosis. We found delayed death due to fulminant liver failure in mice overexpressing miR-221 compared to control mice (Fig. 4A). To confirm that the observed survival effect of miR-221 was indeed due to inhibition of apoptosis in the liver, four mice in both groups were sacrificed at 9 hours after Jo2 injection. We found that miR-221 overexpressing mice had reduced pathological signs of liver injury (Fig. 4B). Consistent with liver morphology, miR-221 overexpressing mice had decreased levels of serum transaminases and less hepatocyte apoptosis as detected by lower caspase-3/7 activity (Fig. 4C,D) and a reduced number of TUNEL-positive nuclei (Fig. 4E) than their respective controls. Thus, miR-221 overexpression in mouse liver delays FAS-induced fulminant liver failure by inhibiting hepatocyte apoptosis.
Next, we investigated the mechanism of protection of hepatocytes from apoptosis by miR-221. To find protein targets of miR-221 we used PICTAR, Target Scan, and miRANDA algorithms. In silico analyses identified PUMA, a proapoptotic protein, as a target for miR-221. If Puma mRNA is a true target of miR-221, expression of PUMA protein levels should change in the liver at 12 hours after Jo2-injection in mice, because miR-221 expression increases at this timepoint (Table 1; Supporting Fig. S2c). We therefore determined the level of PUMA at 12 hours after Jo2 injection in primary hepatocytes. We found that PUMA protein expression in hepatocytes was dramatically reduced at 12 hours after Jo2 injection compared to 0 hours control hepatocyte lysates (Fig. 5A), albeit its transiently elevated levels at earlier timepoints (Supporting Fig. S3a). Decrease in PUMA protein levels at 12 hours suggests regulation of PUMA at the posttranscriptional level. However, this decrease may simply be due to transcriptional down-regulation. We therefore determined Puma mRNA levels after Jo2-induced apoptosis by qRT-PCR. We found that mRNA levels of Puma did not decrease; we rather observed a 1.8-fold increase in Puma mRNA expression levels in isolated primary hepatocytes at 12 hours after Jo2-injection (Fig. 5B). Thus, decreased protein levels of PUMA at 12 hours but not those of its mRNA strongly suggest posttranscriptional regulation of PUMA in hepatocytes.
To further investigate whether Puma is posttranscriptionally regulated by miR-221 we cloned the 3′ UTR of Puma downstream of a firefly luciferase gene in a miRGLO vector (henceforth, this construct will be referred as miR-GLO-PUMA). Luciferase reporter assay using the miRGLO vector was used to demonstrate the direct binding of a mature miRNA with 3′ UTR of mRNA. If Puma mRNA is a real target of miR-221, then luciferase activity should decrease or increase after overexpression or inhibition of miR-221, respectively. We transfected miR-GLO-PUMA into mouse primary hepatocytes in which miR-221 was either overexpressed or inhibited. Indeed, luciferase reporter assay showed that hepatocytes transfected with mimics of miR-221 produced lower luciferase activities, whereas hepatocytes transfected with inhibitor of miR-221 showed high luciferase activity, thus confirming the binding of this miRNA with the Puma 3′ UTR (Supporting Fig. S3b,c). Furthermore, overexpression and inhibition of miR-221 decreased and increased PUMA protein levels 48 hours after transfection, respectively (Fig. 5C,D). Thus, luciferase reporter assay and altered protein expression after transfection of mimics in hepatocytes confirms that Puma mRNA is a target of miR-221.
Finally, we investigated whether loss of PUMA in primary hepatocytes mimics the antiapoptotic effects seen by miR-221 overexpression. We used siRNA to knockdown PUMA in primary mouse hepatocytes (Supporting Fig. S3d). After confirming the efficient knockdown, apoptosis was induced. By WST and caspase-3/7 activity assays, we found that loss of PUMA protected hepatocytes from Jo2-induced apoptosis (Fig. 5E,F), a similar effect, which was observed after miR-221 overexpression. Thus, PUMA contributes to the observed antiapoptotic effect of miR-221. Furthermore, in accordance with our in vitro findings, we found decreased levels of PUMA in AAV8-Ttr-miR-221-injected mice (Fig. 6). PTEN and p27, previously known targets of miR-221, were also found to be down-regulated in AAV8-Ttr-miR-221-injected mice. FAS receptor levels were unchanged in miR-221-overexpressing mice, which further suggests that the miR-221-mediated antiapoptotic effect involves molecules other than FAS receptor.
To address the extent to which the antiapoptotic function of miR-221 relies on PUMA, we examined two other important miR-221 targets; PTEN12 and Bmf.28 We used siRNA to knockdown PTEN and BMF in primary hepatocytes (Supporting Fig. S4a,b). WST assay and caspase-3/7 activity assay showed that cells transfected with Pten siRNA but not with Bmf siRNA were protected against FAS-induced apoptosis (Fig. 7A,B). Furthermore, transfection of miR-221 mimic in hepatocytes treated with Puma siRNA or Pten siRNA increased the protection against FAS compared to hepatocytes treated with Puma siRNA or Pten siRNA alone (Fig. 7A,B). Thus, PUMA and PTEN show antiapoptotic behavior in primary hepatocytes in response to FAS-induced apoptosis. Moreover miR-221 further protects hepatocytes, treated with Puma and Pten siRNA, from FAS-induced apoptosis.
To further investigate the mechanism of the antiapoptotic effect of miR-221 in primary hepatocytes, we transfected primary hepatocytes with miR-221 and target protectors of PUMA, PTEN, or BMF. miRNA target protectors are oligo-nucleotides which bind to the miRNA target site in 3′ UTRs of target mRNA and therefore do not allow miRNA mediated repression of a gene. Forty-eight hours after transfection with miRNA target protectors of Puma, Bmf, and Pten in primary hepatocytes, we found that miRNA-target protectors are able to relieve the repression of PUMA, BMF and PTEN (Supporting Fig. S4c,d,e). Next, to assess the contribution of PUMA, PTEN, and BMF to apoptosis, we cotransfected miRNA-target protectors and miR-221 mimic in primary hepatocytes before induction of apoptosis. WST assay and caspase-3/7 activity assay showed that cells cotransfected with Puma or Pten target protectors and miR-221 mimic are able to increase apoptosis slightly, although significantly compared to apoptosis detected in cells cotransfected with control target protectors and miR-221 mimics (Fig. 7C,D). Importantly, cells transfected with Puma, Pten target protectors and miR-221 mimic still showed significantly less apoptosis, as detected by lower caspase-3/7 activity than with the control target protector alone (Fig. 7D). This indicates that other targets may contribute to the observed antiapoptotic effect of miR-221.
Finally, we investigated whether miR-221 also affects TNF-α-induced apoptosis. After miR-221 transfection, we treated hepatocytes with TNF-α (50 ng/mL or 25 ng/mL). At a higher dose of TNF-α (50 ng/mL) the protective effect of miR-221 was not detected (data not shown). However, 24 hours after apoptosis induction by a lower dose of TNF-α (25 ng/mL), WST assay (Supporting Fig. S4f) and caspase-3/7 activity assay (Supporting Fig. S4g) showed a moderate but significant antiapoptotic effect of miR-221.
Together, our findings suggest that miRNAs are differentially regulated during fulminant liver failure. We demonstrate that of the deregulated miRNAs, miR-221 protects mouse hepatocytes from apoptosis in vitro and in vivo. We found that levels of PUMA protein decrease in hepatocytes in contrast to its mRNA levels. Indeed, we show that miR-221 binds to 3′ UTR of Puma mRNA and regulates its protein expression in mouse hepatocytes. In accordance with our findings, Puma regulation by miR-221 has been suggested very recently in glioblastoma cells.29 Our findings of Puma regulation by miR-221 in hepatocytes are important, as it has been shown that regulation of an apoptotic pathway gene and, hence, of cell death by a miRNA is a cell type-specific phenomenon. For example, miR-21 serves as antiapoptotic miRNA in glioblastoma30 and in MCF-7 cells,31 whereas, surprisingly, the same miRNA in HeLa cells functions as a prosurvival miRNA and has no effect on cell survival in A549 human lung cancer cells.32
Overexpression of miR-221 leads to delayed progression of fulminant liver failure, in part by targeting the proapoptotic PUMA protein. However, we do not rule out the involvement of other miR-221 targets, which may have contributed to the observed antiapoptotic effect in mice. Consistent with previous findings, we also observed decreased levels of p27 and PTEN protein. Loss of these proteins has been shown to increase hepatocyte proliferation, which may also render protection to hepatocytes from apoptosis, as proliferating hepatocytes have been shown to resist the lethal effect of Jo2.24 We found that PTEN, another target of miR-221, also acts as an antiapoptotic protein in primary hepatocytes in response to FAS-induced apoptosis.
The two findings, miR-221 up-regulation in wild type mouse liver after FAS-induced apoptosis and delayed fulminant liver failure after ectopic expression of miR-221, led us to hypothesize that miR-221 mediated Puma regulation may be one of the decisive factors for progression or abrogation of fulminant liver failure. Indeed, in vitro PUMA knockdown experiments in primary hepatocytes showing delayed apoptosis confirmed this hypothesis. Similarly, Puma−/− mice have been reported previously to develop less apoptosis after TNF-α-induced apoptosis.33 Consistent with previous findings, miR-221 was also found to be antiapoptotic when primary hepatocytes were subjected to a lower dose of TNF-α-induced cell death. However, a study of FAS-induced apoptosis in Puma−/− knockout mouse liver would further elucidate the role of miR-221 in apoptosis. Thus, up-regulation of miR-221 is a cellular protective mechanism in response to an apoptotic stimulus by reducing the expression of proapoptotic proteins such as PUMA. However, previously known targets of miR-221 in fulminant liver failure such as p27,34 PTEN,12 TIMP3,12 and DDIT411 may also play an important role during fulminant liver failure. Previously, miR-221 has been reported to be antiapoptotic in liver cancer cells.12 However, for the first time, to our knowledge, we provide evidence for the antiapoptotic role of miR-221 in primary hepatocytes during fulminant liver failure. A similar study previously identified miR-491_5p as an up-regulated miRNA after acute liver injury. Overexpression of miR-491_5p leads to increased sensitivity of human hepatoma cells in response to TNF-α-induced apoptosis.35 Therefore, in addition to miR-221, other miRNAs may play essential roles during acute liver injury.
In conclusion, we provide evidence that miR-221 plays an important antiapoptotic role during fulminant liver failure, at least in part, by regulating Puma at the posttranscriptional level in hepatocytes. Our study is the first demonstration of in vivo modulation of fulminant liver failure by delivering an miRNA by AAV8. Thus, miR-221 may be considered one of the therapeutic targets for intervention of fulminant liver failure in the future.
The authors thank Andreas Kispert, Institute for Molecular Biology, Hannover Medical School for providing Rosa26 Cre reporter mice. We thank Heike Bantel and Arndt Vogel for critical discussion of the data. We thank Julia Norden and Jutta Lamlé for expert help with mouse experiments and hepatocyte isolation. We thank Asha Balakrishnan (UCSF), Deepa Subramanyam (UCSF), and Manvendra Kumar Singh (UPEN) for critical reading of the article.