Cluster of Excellence REBIRTH, Hannover Medical School, Hannover, Germany
Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Hannover, Germany
Ph.D., Stem Cell Biology, Cluster of Excellence REBIRTH, Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Hans Borst Zentrum J-11-02-6530, OE 8881, Carl-Neuberg-Str-1, D-30625 Hannover, Germany
Potential conflict of interest: Nothing to report.
Supported by grants from the German Research Foundation (EXC 62/1, SFB 738, SH640/1-1) and HiLF from Hannover Medical School. S.M. and B.R. receive stipends from the Hannover Biomedical Research School (HBRS).
The tightly controlled replication of hepatocytes in liver regeneration and uncontrolled proliferation of tumor cells in hepatocellular carcinoma (HCC) are often modulated by common regulatory pathways. Several microRNAs (miRNAs) are involved in HCC progression by modulating posttranscriptional expression of multiple target genes. miR-221, which is frequently up-regulated in HCCs, delays fulminant liver failure in mice by inhibiting apoptosis, indicating a pleiotropic role of miR-221 in hepatocytes. Here, we hypothesize that modulation of miR-221 targets in primary hepatocytes enhances proliferation, providing novel clues for enhanced liver regeneration. We demonstrate that miR-221 enhances proliferation of in vitro cultivated primary hepatocytes. Furthermore, applying two-thirds partial hepatectomy as a surgically induced liver regeneration model we show that adeno-associated virus-mediated overexpression of miR-221 in the mouse liver also accelerates hepatocyte proliferation in vivo. miR-221 overexpression leads to rapid S-phase entry of hepatocytes during liver regeneration. In addition to the known targets p27 and p57, we identify Aryl hydrocarbon nuclear translocator (Arnt) messenger RNA (mRNA) as a novel target of miR-221, which contributes to the pro-proliferative activity of miR-221. Conclusion: miR-221 overexpression accelerates hepatocyte proliferation. Pharmacological intervention targeting miR-221 may thus be therapeutically beneficial in liver failure by preventing apoptosis and by inducing liver regeneration. (HEPATOLOGY 2013;)
MicroRNAs (miRNAs) have emerged as important posttranscriptional regulators in normal liver physiology and in liver diseases.1-4 In recent years, various groups have identified deregulated miRNAs in HCC by miRNA profiling and quantitative real-time polymerase chain reaction (PCR) analyses.5-18 The functional roles of many of these deregulated miRNAs during HCC progression remain to be elucidated in detail. Among the deregulated miRNAs, miR-221 has been identified to be upregulated in human HCC independently by many laboratories.5, 7, 9, 12, 13, 15, 16 Inhibition of apoptosis may not only contribute to the development of HCC but also render hepatoma cells resistant to therapeutic agents such as tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL).19 miR-221-mediated TRAIL resistance is due to targeting of phosphatase and tensin homolog (PTEN) and tissue inhibitor of metalloproteinases 3 (TIMP3).19 In addition to inhibiting apoptosis, miR-221 has been suggested to modulate the proliferation of HCC cells by regulating the expression of cyclin-dependent kinase inhibitor 1B (CDKN1B or p27) and CDKN1C (p57).20, 21 However, it remains unclear whether miR-221 up-regulation is a cause or consequence of extensive proliferation of primary hepatocytes. A recent publication by Callegari et al.18 suggests that constant overexpression of miR-221 induces HCCs in a subgroup (50%) of mice after a period of 9 months. We previously have shown that miR-221 overexpression inhibits apoptosis of hepatocytes and delays fulminant liver failure.22 However, the direct effect of miR-221 overexpression on proliferation of primary hepatocytes and in a liver regeneration model remains to be investigated.
Liver regeneration involves a well-orchestrated interplay of growth factors and cytokines, which themselves are regulated at various levels. Significant miRNA expression changes have been reported during two-thirds partial hepatectomy (2/3 PH), which represents a well-established in vivo model for liver regeneration.2, 23-25 For example, recent reports demonstrating profound effects of miR-21 and miR-33 knockdown on liver regeneration in mice shed light on the requirement of optimum miRNA expression levels for hepatocyte proliferation.26, 27 Furthermore, these results also indicate potential oncogenic risks that may be associated with altered miRNA expression levels in hepatocytes. Therefore, modulation of miRNAs during 2/3 PH is a reliable model to investigate whether a particular miRNA possesses the capacity to modulate proliferation of primary hepatocytes in vivo.
In our present study we aimed to investigate the regulatory role of miR-221 in hepatocyte proliferation in vitro and during liver regeneration in a murine 2/3 PH model. Our results indicate that overexpression of miR-221 promotes hepatocyte proliferation. Furthermore, we demonstrate that miR-221 regulates the expression of Arnt, which contributes to the pro-proliferative activity of miR-221.
Animal experiments were performed according to the guidelines of the Hannover Medical School, Germany. BALB/c mice were purchased from Charles River Laboratories (Germany).
Hepatocyte Transfection and Proliferation Assay.
Primary hepatocytes were isolated from mouse liver as described.22 For all in vitro transfection experiments we used Percoll density gradient purified hepatocytes to achieve high transfection efficiency. 1.25 × 106 cells per well of a 6-well Primaria plate (BD Bioscience) or 10,000 primary hepatocytes per well of a collagen-coated 96-well plate (BD) were seeded. Twelve hours after seeding hepatocytes were transfected with miR-221 mimic or control mimic using the Targefect reagent in the presence of virofect enhancer (Targetingsystems). Transfected hepatocytes were supplemented with 30 ng/mL epidermal growth factor (EGF) to induce hepatocyte proliferation. Seventy-two hours after transfection hepatocytes were fixed with 4% paraformaldehyde (PFA) and subjected to immunofluorescence staining for 5-bromo-2′-deoxyuridine (BrdU) (Abcam) and Ki67 (Labvision). For the quantitative colorimetric BrdU proliferation assay (Roche), BrdU was added 12 hours before fixation of hepatocytes with 4% PFA.
The 2/3 PH was performed as described.2 Five mice per timepoint (0, 18, 36, 72, and 96 hours) were used for 2/3 PH. All partial hepatectomies were performed at early hours in the morning.
For immunocytochemistry, hepatocytes were fixed in 4% PFA for 15 minutes at room temperature. For BrdU, but not for Ki67 staining, cells were incubated with 2N HCL for 30 minutes before primary antibody incubation. For immunohistochemistry (IHC), tissues were fixed in 10% formalin before embedding them into paraffin blocks. Four μm paraffin sections were prepared for Ki67, proliferating cell nuclear antigen (PCNA), and BrdU staining. We used BrdU staining kit (Invitrogen) and PCNA staining kit (Invitrogen) for IHC. For Ki67 staining we used the antibody at a dilution of 1:100. Ki67 staining was performed similar to PCNA staining protocol (Invitrogen). Terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick end labeling (TUNEL) assay was used to stain apoptotic cells according to the manufacturer's guidelines.
These methods were performed as described in detail.22, 28 Briefly, antibodies for ARNT (Cell Signaling), cyclins D1, E1, A, B1 (Santa Cruz), and tubulin (Sigma) immunoblotting were used at 1:500 and 1:1,000 dilutions, respectively.
Luciferase Reporter Assay.
The 3′ untranslated region (UTR) of Arnt messenger RNA (mRNA) was amplified by PCR from genomic DNA using following primers; forward primer: GGAAAGTTTAAAC TGTACAGTTATTACAGGGCTTCTGA and reverse primer: GGAAATCTAGA TTTGCTACATGTCATTATTTTTGC. The amplicon was cloned into a miRGLO vector (Promega). The first three nucleotides of the miR-221 seed sequence in the 3′ UTR of Arnt was mutated using the QuikChange Lightning Kit Site-Directed Mutagenesis Kit (Agilent Technologies). Luciferase reporter assay was performed as described.22
Transfection of Target Protectors.
miRNA target protectors for Arnt were purchased from Qiagen. Hepatocytes were cotransfected with miR-221 mimic and Arnt Target Protectors or control target protectors using the Targefect F2 with enhancer virofect (Targetingsystems). Twelve hours after cotransfection, medium was supplemented with 30 ng/mL EGF to induce proliferation.
Two μg of pCMV-SPORT6-ARNT plasmid (BioCat), containing Arnt cDNA, was cotransfected with miR-221 mimic in primary hepatocytes using Targefect F2 and virofect enhancer. Overexpression of ARNT was determined by western blot.
Significance was determined with the two-tailed Student's t test. P < 0.05 was considered significant.
To investigate the role of miR-221 on hepatocyte proliferation we transfected freshly isolated and purified primary hepatocytes with 25 nM of miR-221 mimic or control mimic. Hepatocyte proliferation was induced by supplementing the medium with EGF. BrdU and Ki67 immunofluorescence stainings were used to determine the hepatocyte proliferation. At 72 hours after transfection, we found that primary hepatocytes transfected with miR-221 mimic had higher numbers of BrdU- and Ki67-positive nuclei than hepatocytes transfected with control mimic (Fig. 1A). Thus, BrdU and Ki67 immunofluorescence staining indicated the enhanced proliferation of primary hepatocytes in the presence of miR-221 mimic. To further confirm the enhanced hepatocyte proliferation we performed a quantitative BrdU proliferation assay. At first, we tested whether the quantitative BrdU assay can be used for primary hepatocytes in the presence or absence of EGF. Consistent with previous reports, we found that BrdU incorporation was higher in hepatocytes cultured in the presence of EGF (Supporting Fig. S1). Next, we determined BrdU incorporation in primary hepatocytes in the presence of 25, 50, 100, and 150 nM of miR-221 mimic (Fig. 1B). We found increased BrdU incorporation in primary hepatocytes transfected with 25, 50, 100, and 150 nM of miR-221 mimic, an observation also detected by Ki67 and BrdU immunofluorescence staining. At 150 nM, we observed cell toxicity in hepatocytes transfected with miR-221 mimic. This toxicity may be due to off-target effects because it was also observed in hepatocytes transfected with control mimic. Based on these results we chose 100 nM (high dose) or 25 nM (low dose) of miR-221 mimic for further in vitro experiments. Together, our in vitro proliferation assay and immunofluorescence staining suggest that miR-221 promotes hepatocyte proliferation.
We next addressed the question whether overexpression of miR-221 affects hepatocyte proliferation in vivo. To answer this question we overexpressed miR-221 in the mouse liver by an AAV8 (AAV8-Ttr-miR-221) vector, which expresses miR-221 under the transcriptional control of transthyretin (TTR) promoter. This gene transfer system ensures hepatocyte specific transduction by taking advantage of hepatic tropism of AAV8 and the hepatic promoter activity of TTR, as described.22 1 × 1011 viral particles of AAV8-Ttr-miR-221 or AAV8-Ttr-Cre (control) were separately injected into BALB/c mice by way of the tail-vein (Fig. 2A). Seven days after AAV8 injection, we performed 2/3 PH to induce hepatocyte proliferation in vivo. To confirm overexpression of miR-221 in the liver of mice injected with AAV8-Ttr-miR-221 we analyzed the expression of miR-221 by qRT-PCR. We found that mice injected with AAV8-Ttr-miR-221 indeed expressed high levels of miR-221 (Fig. 2B), which is consistent with results from our previous study.22 In control mice we found that miR-221 level increased slightly after 2/3 PH (Fig. 2B). Importantly, miR-221 levels continued to remain higher after PH in livers of mice injected with AAV8-Ttr-miR-221 than in the livers of control mice. Next, in vivo hepatocyte proliferation was analyzed by immunohistochemistry for Ki67, BrdU, and PCNA antigens (Fig. 2C). Initially, we choose the timepoint of 36 hours after 2/3 PH, which coincides with the peak of DNA synthesis in mice. We found that miR-221 overexpressing mice exhibited higher numbers of Ki67, BrdU, and PCNA-positive nuclei in hepatocytes compared to their respective controls at this timepoint (Fig. 2C,D). Higher numbers of BrdU and PCNA-stained positive cells indicate that miR-221 overexpressing hepatocytes have entered more rapidly into S-phase than the control hepatocytes. Although the peak of DNA synthesis in mouse liver occurs at 36 hours after 2/3 PH, proliferation can be detected at later timepoints as well.29-32 We therefore analyzed proliferation at 72 and 96 hours after 2/3 PH. Similar to 36 hours, we found a significant increase in proliferation at 72 hours after 2/3 PH (Fig. 2D). In contrast, we found no significant difference in proliferation at 96 hours in liver of mice injected with AAV8-Ttr-miR-221 compared to liver of mice injected with AAV8-Ttr-Cre (Fig. 2D). Notably, we did not find any spontaneous proliferation of hepatocytes in miR-221 overexpressing mice at 0 hours after 2/3 PH (Fig. S2). In addition to similar proliferative activity at 96 hours after 2/3 PH, the liver weight/body weight ratio was not different in miR-221 overexpressing mice compared to controls, indicating a controlled termination of liver regeneration after 1 week (Fig. S3A,B). Thus, our in vivo proliferation analyses suggest that AAV-8-mediated miR-221 overexpression accelerates liver regeneration in mice mainly at an early phase after 2/3 PH but does not induce exceeding growth of the regenerating liver in later stages.
Cyclins are important regulators of the cell cycle. To gain further insights into the cell cycle of proliferating hepatocytes after 2/3 PH we analyzed the expression of various cyclins such as cyclin D1, cyclin E1, cyclin A2, and cyclin B1 by qRT-PCR (Fig. 3A-D). We found that miR-221 overexpressing mice have higher expression of cyclin D1, E1, and A2 at 36 hours after 2/3 PH. Similar to Ki67, BrdU, and PCNA staining, we found increased expression of cyclin D1, E1, A2, and also B1 at 72 hours after 2/3 PH. Furthermore, we did not find any difference in the expression levels of cyclin D1, E1, A2, and B1 at 96 hours in liver of mice injected with AAV8-Ttr-miR-221 than in the liver of mice injected with AAV8-Ttr-Cre. In addition, higher protein levels of these cyclins were found in miR-221 overexpressing mice at 36 and 72 hours after 2/3 PH (Fig. 3E). Thus, Ki67, BrdU, and PCNA staining as well as expression analyses of a panel of cyclins revealed that miR-221 overexpression in mouse liver accelerates liver regeneration.
To elucidate the mechanism of accelerated hepatocyte proliferation during liver regeneration in miR-221 overexpressing mice, we carried out in silico analyses to identify a novel target of miR-221. Our in silico analyses (TargetScan followed by DAVID functional annotation analyses) predicted the 3′ UTR of Aryl hydrocarbon receptor nuclear translocator (Arnt) mRNA as a novel putative target of miR-221 (Fig. 4A). Importantly, miR-221 binding site was conserved in Arnt mRNA of human, mouse, dog, and chicken. ARNT (also known as hypoxia inducible factor 1 beta) is a member of the basic helix loop helix / Per-ARNT-SIM (bHLH/PAS) superfamily. It is expressed in a variety of cells including hepatocytes. Loss of Arnt has been associated with increased HCC recurrence and lower survival in patients.33 Furthermore, loss of ARNT has been shown to affect proliferation of human hepatoma cells.33 We therefore sought to investigate whether Arnt mRNA is regulated by miR-221. To this end, we transfected 100 nM miR-221 mimic in primary hepatocytes and analyzed the expression of ARNT protein levels by western blot and Arnt mRNA by qRT-PCR. We found that Arnt mRNA and protein expression decreased in hepatocytes, which were transfected with the miR-221 mimic (Fig. 4B,D). Conversely, inhibition of miR-221 by transfection of 100 nM miR-221 inhibitor in primary hepatocytes led to increase in ARNT protein and Arnt mRNA levels (Fig. 4C,E). To further investigate whether miR-221 regulates expression of Arnt at the posttranscription level, we cloned 3′ UTR of Arnt downstream of firefly luciferase gene in miR-GLO vector (miR-GLO-ARNT). Hepatocytes were transfected with miR-221 mimic or inhibitor followed by transfection with miR-GLO-ARNT vector. Luciferase reporter assay showed that hepatocytes transfected with miR-221 mimic had lower luciferase activity compared to hepatocytes transfected with control mimic (Fig. 4F). Likewise, hepatocytes transfected with inhibitor of miR-221 showed higher luciferase activity than the hepatocytes transfected with control inhibitor (Fig. 4G). This set of data was reproduced, using with a lower dose (25 nM) of miR-221 mimic or inhibitor, which also resulted in modulation of ARNT expression, which further confirmed the posttranscriptional regulation of Arnt by miR-221 (Fig. S4). Furthermore, we sought to investigate whether Arnt expression changes after 2/3 PH. By qRT-PCR analysis we found that Arnt expression decreases after 2/3 PH (Fig. 4H), whereas miR-221 level increases after 2/3 PH (Fig. 2B), indicating an in vivo regulation of Arnt by miR-221. Thus, our western blot analyses, luciferase reporter assays, and Arnt expression analyses after 2/3 PH established the 3′ UTR of Arnt as a target of miR-221.
Next, we investigated whether loss of ARNT in primary hepatocytes mimic the pro-proliferative effect of miR-221. To answer this question we used small interfering RNAs (siRNAs) against Arnt mRNA. As expected, the transfection of Arnt siRNA led to efficient knockdown of ARNT as detected by western blot (Fig. 5A). After confirming the loss of Arnt we analyzed the effect of ARNT loss on the proliferation of primary hepatocytes by Ki67 immunofluorescence staining. We found that hepatocytes transfected with 25 nM Arnt siRNA had higher numbers of Ki67-positive nuclei than hepatocytes transfected with control mimic (Fig. 5B,C). Similarly, quantitative BrdU proliferation assay also showed a 1.3- and 1.8-fold higher number of BrdU-positive nuclei in hepatocytes transfected with 25 nM and 50 nM Arnt siRNA, respectively (Fig. 5D). Thus, similar to the overexpression of miR-221, the loss of ARNT promotes the proliferation of primary hepatocytes.
Next, to investigate whether cell cycle regulators are affected by knockdown of Arnt mRNA, we determined the protein levels of the cdk inhibitors, p21 and p27. We found that levels of p27, but not of p21, were reduced in hepatocytes transfected with Arnt siRNA (Fig. 5E). Therefore, reduced levels of p27 may also contribute to enhanced proliferation of ARNT knockdown cells.
We then addressed the question to what degree does miR-221-mediated enhanced proliferation depend on ARNT. To address this, we cotransfected primary hepatocytes with miR-221 mimic and Arnt target protectors. Arnt target protectors would specifically inhibit the binding of miR-221 with 3′ UTR of Arnt without affecting the interaction of miR-221 with other targets. At first, we confirmed the target protector-mediated effect on the 3′ UTR of Arnt by western blot. We found derepression of ARNT in hepatocytes transfected with Arnt target protector compared to hepatocytes transfected with control target protectors (Fig. 6A). We then analyzed the proliferation of primary hepatocytes in the presence of Arnt target protector by quantitative BrdU assay. We found that hepatocytes cotransfected with miR-221 mimic and Arnt target protectors continue to promote proliferation; however, to lesser but significant levels compared to hepatocytes cotransfected with miR-221 mimic and control target protectors (Fig. 6B). Experiments with a low dose of miR-221 mimic (25 nM) also showed a partial reduction in miR-221-mediated enhanced proliferation in the presence of Arnt target protectors (Fig. S5a). Therefore, our target protector experiments revealed that loss of ARNT contributes to the miR-221-mediated increased proliferation of hepatocytes. In addition to target protectors, in a separate experiment we overexpressed Arnt and miR-221 by cotransfecting purified primary hepatocytes with a pCMV-SPORT6-ARNT plasmid and the miR-221 mimic. By western blotting, we first confirmed the ARNT overexpression in transfected hepatocytes before analyzing whether overexpression of ARNT can inhibit miR-221-induced proliferation (Fig. S5b). We found that, similar to target protectors, overexpression of Arnt diminished miR-221 induced hepatocyte proliferation (Fig. S5c).
We next examined whether miR-221 levels changed during EGF induced proliferation. By qRT-PCR analyses we found that miR-221 levels increased slightly at 72 hours after EGF stimulation (Fig. S6a). We then sought to investigate the effect of inhibition of miR-221 on proliferation of primary hepatocytes in the presence of EGF. To inhibit miR-221 we transfected the miR-221 inhibitor in purified primary hepatocytes. BrdU and Ki67 staining showed no significant difference in proliferation (Fig. S6b). Furthermore, quantitative BrdU proliferation assay also showed no difference in proliferation of primary hepatocytes transfected with 100 nM or 25 nM miR-221 inhibitor (Fig. S6c).
Because miR-221 acts as an antiapoptotic miRNA in hepatocytes we investigated the effect of miR-221 on apoptosis during EGF-induced proliferation in vitro and in the 2/3 PH model. We analyzed apoptosis by TUNEL assay. We found only rare apoptotic cells during EGF-induced hepatocyte proliferation or in the regenerating liver after 2/3 PH in the presence or absence of miR-221 overexpression (Figs. S7a, S8). However, in response to apoptosis inducing Jo2 antibody (Fas agonist), a protective role of miR-221 on hepatocyte protection was observed (Fig. S7b,c).
MiR-221 has been proposed as a potential target for therapeutic intervention in HCC and fulminant liver failure.19, 21, 22 The antiapoptotic role of miR-221 is well established; however, its role in the proliferation of primary hepatocytes has not been elucidated. In the present study we investigated the precise role of miR-221 in hepatocyte proliferation in vitro and in vivo. Primary hepatocytes rarely proliferate in vitro; however, in the presence of growth factors such as EGF or HGF hepatocyte proliferation can be induced.34-36 Our data show that miR-221 overexpression in primary hepatocytes increases proliferation in the presence of EGF. Consistent with these findings, miR-221 overexpression by AAV8 vector in vivo accelerated liver regeneration by rapid S-phase entry of hepatocytes in a 2/3 PH model.
We demonstrate that the observed enhanced proliferation of primary hepatocytes in the presence of miR-221 is, at least in part, due to posttranscriptional regulation of Arnt (Fig. 6C). Our data indicate that miR-221 binds to the 3′ UTR of Arnt mRNA and thereby regulates its expression at the posttranscriptional level. ARNT, a member of the PAS superfamily, is an essential molecule for hypoxia and xenobiotic responses. Recently, a function of ARNT in the proliferation of hepatic cells has been reported.33, 37 Specifically, the loss of ARNT has recently been shown to increase the proliferation of human hepatoma cells.33 Here we show that siRNA-mediated knockdown of Arnt mRNA leads to enhanced proliferation of primary hepatocytes, suggesting that ARNT contributes to miR-221-mediated increased proliferation. Undoubtedly, other targets of miR-221 such as p27 and p57 may contribute to accelerated hepatocyte proliferation in response to miR-221 overexpression. In fact, p27 null mice were shown to have accelerated DNA replication after partial hepatectomy.38 Furthermore, it is important to note that miR-221 up-regulation,5, 7, 9, 12, 13, 15, 16 p27 down-regulation,17, 39 p57 down-regulation,40 and reduced levels of ARNT33 have commonly been associated with advanced tumor stage, lower survival, and higher recurrence rates in human HCC. In addition to previously known targets of miR-221 such as p27 and p57, we show that miR-221 enhances proliferation of hepatocytes by regulating the expression of Arnt at the posttranscriptional level. The role of ARNT in hypoxia is well known; however, miR-221-mediated enhanced proliferation by way of ARNT is most likely independent of hypoxia, as 2/3 PH is a model that does not have hypoxia at early hours after 70% PH.41 However, to further elucidate the effect of loss of ARNT on liver regeneration, a future study of 2/3 PH in a hepatocyte-specific Arnt knockout mouse model would be required.
Notably, the effect of loss of ARNT on hepatocyte proliferation was not as pronounced as was observed in the case of miR-221 overexpression. Similarly, derepression of Arnt mRNA in the presence of specific target protectors reduced the miR-221-mediated pro-proliferation effect only partially. ARNT has been shown to positively regulate the expression of cdk inhibitors such as p27 and p21.42, 43 Our findings revealed that ARNT loss decreases the expression of p27 but not p21 in primary hepatocytes treated with Arnt siRNA. This suggests that ARNT loss enhances proliferation, at least in part, by reducing p27 levels. The level of p27 decreases after 2/3 PH44, 45 and its degradation by ubiquitin-proteasome pathway is important for G1 to S phase progression.44 Our data indicate that reduced levels of ARNT may contribute to decreased levels of p27 and may therefore accelerate hepatocyte proliferation during liver regeneration in vivo. Similar to our findings, sustained activation of aryl hydrocarbon receptor (a heterodimer partner of ARNT) by an agonist causes attenuation of liver regeneration by blocking cell cycle progression at the G1/S checkpoint of the cell cycle.46 Taken together, our findings indicate that miR-221 regulates ARNT, which, in turn, regulates p27 levels and thereby affects hepatocyte proliferation (Fig. 6C).
In contrast, loss of miR-221 did not significantly affect proliferation of primary hepatocytes. This inability to block proliferation by knocking down miR-221 may be related to the modest endogenous levels of miR-221 in primary hepatocytes. Considering studies that aim to knockdown miR-221 in HCC by use of antisense oligonucleotides as a therapeutic tool, the unperturbed cell cycle of normal hepatocyte after knockdown of miR-221 could even be an advantage.
In addition to miR-221, other miRNAs such as miR-21 and miR-33 have recently been proposed to modulate hepatocyte proliferation during liver regeneration.26, 27 Inhibition of miR-21 impaired liver regeneration, whereas inhibition of miR-33 improves liver regeneration. However, it remains to be investigated whether in vivo modulation of miR-33 and miR-21 can also significantly accelerate proliferation of primary hepatocytes in vitro and during liver regeneration. Interestingly, modulation of all three miRNAs (-221, -33, and -21) seems to affect the G1/S checkpoint of hepatocyte cell cycle. In fact, the G1/S checkpoint is very often disrupted in HCC.47 Furthermore, it is important to note that long-term stable overexpression of miR-221 in transgenic mice leads to HCC formation in 50% of mice at the age of 9 months.18 Therefore, miR-221 may act as a double-edged sword because its pro-proliferative function may be harnessed for treatment of fulminant liver failure but long-term stable overexpression in the liver may increase the risk of HCC. Hence, careful strategies, which provide a constant balance between proliferation and nontumorigenicity, need to be developed. Strategies involving transient overexpression by AAV or oligonucleotide delivery would therefore be the preferred approaches for overexpression of potential oncogenic miRNAs such as miR-221 or miR-21. In addition to our study, AAV-mediated delivery technology has been applied by other groups to be effective in the treatment of liver diseases in mice.48
In summary, our results reveal that miR-221 overexpression accelerates hepatocyte proliferation in vitro and in vivo. Our findings show that miR-221 has pro-proliferative properties in primary hepatocytes in addition to the previously shown antiapoptotic function. Our results indicate that modulation of miR-221 may represent new therapeutic approaches to treat liver failure caused by impaired hepatocyte proliferation.
We thank Jessica Wenzl for help during AAV preparations and Christian Viering as well as Merlin Witte for routine technical assistance.