The role of miR-21 in liver regeneration miRCURY LNA™ microRNA Inhibitors
Dr. Holger Willenbring
, Associate professor and Raymond Ng
, PhD student, from Eli and Edythe Broad Center of Regeneration Medicine and Stem Cell Research Department of Surgery, Division of Transplantation at the University of California San Francisco study many aspects of liver regeneration. They have been using Exiqon array services and miRCURY LNA™ microRNA in vivo
inhibitors to study the role of miRNAs in this process. Their results show in molecular detail how miR-21 stimulates rapid proliferation of hepatocytes in the regenerating liver by stimulating Akt1-mTORC1 signaling. For more details see the recent paper in Journal of Clinical Investigation.
1. Please tell us about the main focus of your research and how you got involved with miRNA research?
Initially, we were curious about whether miRNAs play a role in liver regeneration. So we generated mice that lacked all miRNAs in hepatocytes because of knockout of DGCR8 (an essential component of a microprocessor complex containing Drosha that releases pre-miRs from pri-miRs transcripts in the nucleus). When we performed 2/3 partial hepatectomy on these mice we saw that their hepatocytes had a cell cycle delay in the early phase of liver regeneration, that is, at the G1-to-S phase transition. Prompted by this observation we decided to search for the miRNAs involved in S phase entry in regenerating hepatocytes.
2. You submitted samples for miRNA array profiling in Exiqon Services. Can you tell us about the objective of the experiments and the results and why the study initially surprisingly failed to give any results?
The objective of the miRNA array profiling was to identify miRNAs that are differentially expressed during liver regeneration and thus may regulate hepatocyte proliferation. For this purpose, we sampled livers of mice at different time points after 2/3 PH. To our surprise the initial study which included 3 biological replicates at each time point suggested no or very modest changes to the miRNA profile during liver regeneration. However, in a second study including sufficient numbers of biological replicates (5), we identified miR-21 as a miRNA rapidly induced by 2/3 PH with good statistical significance.
3. What was your next step?
Next we wanted to know what miR-21 does in regenerating hepatocytes. We decided to specifically antagonize the surge in miR-21 expression occurring after 2/3 PH because miR-21 is expressed highly in normal hepatocytes and we reasoned that fully depleting it by genetic mutation may disrupt homeostasis and interfere with our analyses. So we worked with Niels Frandsen from Exiqon to generate a miR-21 inhibitor for intravenous injection into mice.
4. How did you determine the right dose regime?
We tail-vein-injected mice with different doses of the miR-21-ASO and analyzed their livers at different time points after injection to determine the levels of hepatic miR-21 inhibition. This allowed us to determine that a single injection of miR-21-ASO acts to inhibit miR-21 and de-represses its target gene Btg2 in the liver within just 6 hours and that, after 2/3 PH, the effect persists for 48-72 hours.
5. How did you ascertain that you were indeed achieving inhibition of miR-21?
We checked miR-21 levels using real-time PCR and also performed qRT-PCR and Western Blots of several established miR-21 target genes such at PTEN, Spry1 and Btg2.
6. How did you decide on when to administer the inhibitor and when to look for effects of miR-21 inhibition?
We knew from staining of proliferation markers such as Ki67 and PCNA that there was a cell cycle delay from 18-48 hours after 2/3 PH in hepatocytes lacking all miRNAs. Our data also showed that miR-21 expression increases rapidly 6 hours after 2/3 PH and comes back down by 36 hours after 2/3 PH. Therefore, we reasoned that we had to inhibit the surge of miR-21 before it reaches its peak at 18-24 hours after 2/3 PH. Since we had established that inhibition kicks in around 6 hours after intravenous injection we tried injecting the miR-21-ASO at 6 hours and 10 hours after 2/3 PH. We used a dose of 25 mg/kg miR-21-ASO because it afforded repression of the miR-21 surge whilst maintaining baseline levels. We determined whether miR-21 was effectively inhibited by checking levels of miR-21 as well as its target gene Btg2. We determined the effects of miR-21 knockdown on hepatocyte proliferation by analyzing livers at 2 different time points, 18 hours and 36 hours after 2/3 PH, for Ki67 expression, which begins in late G1 phase, and expression of PCNA and BrdU labeling, both S phase markers.
7. Although the surge in miR-21 expression is not essential to liver regeneration is does play an important role at an early stage of the process – tell us what you found.
Consistent with the phenotype observed with the DGCR8 mutant we found that inhibition of miR-21 impaired the transition of hepatocytes to the S phase of the cell cycle. This effect was observed 18-36h after 2/3h PH but after 72h hepatocytes injected with miR-21-ASO had similar levels of phosphorylated histone H3. However, we only observed this effect when the inhibitor was administered 6 hours after 2/3 PH. Cell cycle progression was normal when we injected the inhibitor 10 hours after 2/3PH. We conclude that the surge in miR-21 activity is required during a brief window of time (12-16 hours after 2/3 PH) for rapid G1-to-S phase transition. At the molecular level we observed that miR-21 inhibition suppressed cyclin D1 levels, which could be the cause of the cell cycle arrest. In fact, injection of a cyclin D1 inhibitor after 2/3 PH also led to a delay in G1-to-S phase progression of regenerating hepatocytes.
8. The effect of the miR-21 inhibitor was quite subtle. Correct timing of administration and collection of samples was crucial. If you had injected the oligo only at 10hr post 2/3PH or taken samples only after 72hr you would not have observed any effect at all?
Yes, we did not observe any differences between the control and mice injected with miR-21-ASO at 10 hours after 2/3 PH. And since the effects of the miR-21-ASO were compensated for by increased cyclin E1 and A2 expression, the hepatocytes were able to overcome the cell cycle entry delay and catch up by M phase which is roughly 72 hours after 2/3 PH.
9. So the question was how does miR-21 activity affect Ccnd1 expression? How did you solve this question?
We observed that miR-21 suppression led to decreased cyclin D1 protein levels but mRNA levels remained the same in miR-21-ASO-injected mice. Therefore, we checked for cyclin D1 degradation by measuring levels of G3K3β, the main enzyme responsible for degrading cyclin D1. The enzyme levels were similar in all the samples suggesting that degradation of cyclin D1 wasn’t the cause of decreased cyclin D1 in the miR-21-ASO livers. Thus, we decided to perform polysome analysis to check if cyclin D1 was being translated differently. Indeed, we found that cyclin D1 mRNA was decreased in the polysomal fractions confirming that it was being translated at a lower rate when miR-21 was inhibited by miR-21-ASO. Furthermore, we discovered that increased miR-21 expression facilitates cyclin D1 translation in the early phase of liver regeneration by stimulating Akt1/mTORC1 signaling and thus eIF-4F-mediated translation initiation of Ccnd1 mRNA. This effect of miR-21 is mediated by suppression of Rhob.
10. How did you home in on Rhob as a miR-21 target?
We had identified a cohort of 63 target genes for miR-21 in the liver, which we narrowed down to 6 by performing gene ontology analysis. We selected for genes which were negative regulators of the cell cycle. Rhob stuck out because it had the highest prediction score and due to it’s previously reported ability to inhibit Akt1/mTORC1 signaling. Furthermore, the levels of Rhob were inversely correlated with those of miR-21. We confirmed that Rhob was a direct miR-21 target gene and showed that Rhob knockdown could restore normal cyclin D1 levels in miR-21-ASO-treated cells. This finding showed that miR-21 was acting through Rhob to regulate cyclin D1 translation.
11. What are the main conclusions of your published work that are of general significance outside the field of liver regeneration?
Our findings reveal that miR-21 accelerates cyclin D1 translation in hepatocytes, which enables rapid liver regeneration. This is achieved by stimulation of Akt1/mTORC1 signaling by direct targeting of Rhob. Because miR-21 is an oncomiR, it is possible that this mechanism plays a role in accelerating translation of not only cyclin D1 but also other mRNAs subject to cap-dependent translation in tumors. It would also be interesting to examine other tissues with a high proliferation rate (e.g., small intestine and hair follicle) to see if a similar mechanism is used to promote cyclin D1 expression.
12. Where is this project taking you now?
Rapid progression of hepatocytes through G1 and into S phase is critical for survival from acute liver injury. We are currently exploring whether miR-21 mimics can be used to accelerate liver regeneration.
13. Do you have any advice for researchers thinking about performing miRNA inhibition studies in vivo ?
We think it is important to first determine the strength and duration of inhibition of your targeted miRNA. Since we work in the liver (an organ that take up oligonucleotides very efficiently), we had the advantage of not having to worry about problems with tissue targeting. Other researchers may have to overcome that hurdle. Always use functional assays (derepression of miRNA targets) when assessing whether miRNA inhibition has been achieved – miRNA PCR assays are prone to give false positive results and cannot stand alone. Remember to include a mismatched control inhibitor –reviewers are likely to ask for that.