#MicroTwJC: Dicing Up Virus Genomes Part 3: The Controversy

In this weeks Microbiology Twitter Journal Club we are going to be discussing two papers published a month ago in Science. These papers purport to provide evidence for RNA Silencing as an antiviral mechanism in mammals.

For those of you who don't know what RNA silencing is, or have no idea about these papers, feel free to check out my first two posts for this weeks MicroTwJC.

Basically an enzyme known as DICER spots double stranded RNA within cells, and chops it up into short 22 base pair sequences.  These sequences are then grasped by the RNA Silencing complex. The RNA Silencing complex uses these short sequences to recognise full length RNA floating within the cell. Once the RNA Silencing complex recognises these full length RNAs, they break them down. Through this, they can "silence" the expression of genes.

For a long time, its been known that many viruses use RNA to make their genomes, and when they reproduce their genomes they can form double stranded RNA, which is a prime target for DICER and the RNA Silencing Complex.

Can organisms use RNA Silencing to fight against viruses ?

The answer to that question up till this point has been a Yes, with an important caveat. There has been definite evidence for RNA silencing playing a role in resisting viruses in plants and in invertebrates.... but not mammals. 

Whilst mammals are known to have active RNA silencing systems, the evidence for them playing a role in combating viruses rather than just regulating cellular RNA has been hard to discover.

Interfering Interferon

The main problem with proving that RNA Silencing  is an antiviral response in mammalian cells is that mammals have evolved a much stronger ways of responding to viral double stranded RNA.

For a start, there are Pathogen recognition receptor system, which acts like an early warning system for a cell when its being attacked by a pathogen. These receptors can detect specific bacterial or viral components, and send out an alarm to the rest of the cell so that it can take appropriate action.

Toll-like Receptor 3 is perhaps one of the most important ones for recognising viruses, as it can specifically recognise double stranded RNA from viruses. It is usually found in Dendritic cells, which are immune cells that tend to mop up debris from damaged areas of the human body, so if a virus is causing damage, it's likely the'll be around to detect it.

Once they do recognise it, they produce Interferon, which warns cells that viruses may be in the area. The cells then upregulate production of anti-RNA effectors, such as RNA activated Protein Kinase and Latent Ribonucleases. These slow down protein production in response to double stranded RNA and break down any double stranded RNA in the cell. In extreme cases, they can push the cell self destruct button (known as Apoptosis) and kill the cell, taking the virus with it.

These antiviral measures tend to swamp any effects that small interfering RNA's may exert. Remember, the main effect of RNA interference is a decrease in the number of viral genomes floating around a cell, a number that would already be reduced by Ribonucleases.

Viruses do have some defences against this response. If they protect their dsRNA from attack, they can avoid being detected by TLR3 and RNA-activated Protein Kinases, and being broken down by ribonucleases. In a funny/frustrating coincidence, the proteins that enable this defence also enable protection against DICER and the RNA-silencing complex.

This makes it very difficult to obtain unequivocal evidence that RNA silencing actually occurs in mammalian cells.

What evidence needs to be gained ?

In 2006, Bryan Cullen threw down the gauntlet, and came up with three predictions of what you would see if the RNA silencing complex genuinely broke down viral RNAs. I will quote directly from the article.

• "viral infection should result in the production of siRNA of viral origin"
• "inhibition of the RNAi response should enhance virus replication"
• "as an adaptive response to that antiviral mechanism, many viruses should have evolved gene products that specifically inhibit RNAi" 

Do these papers show that the RNA Silencing complex can break down viral RNA into siRNA's ? 

Both of these papers fulfil the first prediction. They detected siRNA fragments of a viral origin in mouse embryonic stem cells, in hamster kidney cells, and during infection of full grown mice.

Is there evidence that inhibition of the RNAi response enhances viral replication ?

Let's take a look at the kinds of evidence researchers present to prove that inhibition on the RNAi response improves replication.

There are a number of ways which the researchers investigated this. The first was to look at what happens when we remove the activity of viral suppressors of RNA interference.

Li et al focussed their study on Nodamura virus with or without expression of B2. In one study they infected the B2 deficient mutant into cells that were engineered to produce RNA interference suppressors, either B2 or VP35. When these cells were infected with wild type viruses, they tended to accumulate more viral RNA than when the wild type virus infected normal cells. The over expression of these RNA suppressors appeared to enhance replication.

At least it would if the researchers had used the mysterious science of statistics to analyse this data. There could be a trend there, but we can't be sure.

Both of them however demonstrate that the removal of B2 reduces the abundance of the RNA genome within a cell.

But as I noted before, B2 can potentially protect against all of the other antiviral mechanisms within a cell. So this evidence alone isn't necessarily enough. They needed to disrupt the DICER system, and show that removing that improves the survival of viral RNA's.

Neither study showed this in their published papers, although Maillard et al came the closest, as they actually produced a DICER knockout model. In their paper, they showed that without DICER, virus derived siRNAs didn't show up.

However, if you dig into the supplementary section of the paper, and do some hunting, you will find a really interesting figure. In the DICER knockout, we find that there is a higher accumulation of VP3 RNA, indicating that knocking out DICER improves viral survival.

This seems to strongly suggest that in embryonic stem cells at least, the ECM virus infection is impaired by RNA silencing.

Do these papers show evidence that the virus response evolved to specifically hamper RNA interference in mammals?

This is the hardest of the predictions to prove, because as I mentioned earlier, you can argue that any protein that protects RNA could hamper both the TLR response and the RNA interference response.

It's not enough that viruses have evolved genes like B2 and VP35, you need to prove that these genes actually have a functional effect that is not related to inhibiting the interferon response.

In embryonic stem cells, the interferon response doesn't generally happen. In fact, the researchers showed that when the interferon response rolled into action, RNA interference ceased to be important in infection.

Things get a lot weirder when we get to the mouse experiments, where researchers took samples of RNA from mice during infection and used that to check whether there were any differences in the expression of immune genes. All of that data was stuffed into a supplementary table that I did not have the energy to get through yesterday. Essentially, the proteins that regulate interferon are expressed in different patterns, and there is a slight up-regulation in the production of interferon alpha, but the relevance is debatable

My main problem with this data is that yet again we don't get any idea of the variance of the data, which is very important. 

Even more importantly, we don't get any idea of when in the infection these mice are being sampled. If the infection has resolved in one group of mice and not the other, then it is effectively the same as comparing infected mice with uninfected mice. Any differences due to the infection of a mutated virus could have completely been resolved by the time point they are examining. The inflammatory changes could simply be due to the infection resolving. The researchers needed to communicate what day of infection they were examining these genes on, and also provide the reader with an idea of the viral abundance at this point.

All I can say for this is that it looks like RNA interference is important for embryonic stem cells, and not unconditionally proven for the rest of the other models.

Is RNA interference a relevant antiviral response for whole organisms ?

These studies have demonstrated that in embryonic stem cells, RNA interference does suppress the activity of the ECM virus and the Nodamura virus. I will concede that there is enough evidence for that. 

I can also say that based on the evidence presented, viral RNA's from the Nodamura virus and the ECMV get processed into short RNAs during the course of infections, and that B2 can halt this process.

The main problem comes from that damn interfering interferon response. The changes in viral genome abundance could alternatively be explained as a result of B2 flipping off the interferon response.

Whilst Li et al tried to account for it, they haven't presented their data in enough detail for me to conclusively say that B2 had no effect on the interferon response. Key details are missing, like how long after infection did they collect their samples? Where exactly from the mouse's anatomy did those samples come from ? 

Conclusions

These papers have enough in them to convince me that RNA interference of viruses can occur in mammalian cells, and that they could be particularly important in embryonic stem cells. This is an evolutionarily ancient antiviral response, so it make sense that it could be important for the most basal cells we know of. But just as the Interferon response superseded these responses evolutionarily, they superseded them developmentally during the lifespan of an organism.

Verdict

This is a really difficult one for me. These papers have provided quite compelling evidence that RNA interference does affect viruses in certain situations, but not enough evidence to fully support RNA interference as a key part of the mammalian immune response. 

These papers have pushed the goal posts of this debate towards debating whether siRNA's are important in viral infection rather than whether they actually exist.

But the controversy is alive and kicking.

Join in the discussion this Tuesday at 8pm GMT. Follow the #microtwjc hashtag and be a part of Microbiology Twitter Journal Club !

Cullen B., Cherry S. & tenOever B. (2013). Is RNA Interference a Physiologically Relevant Innate Antiviral Immune Response in Mammals?, Cell Host & Microbe, 14 (4) 374-378. DOI: 10.1016/j.chom.2013.09.011

Cullen B.R. (2006). Is RNA interference involved in intrinsic antiviral immunity in mammals?, Nature Immunology, 7 (6) 563-567. DOI: 10.1038/ni1352