Home sweet home


Picking the perfect place to live is important, especially if you’re a retrovirus.

Unlike viruses such as the common cold, which infect cells and completely destroy them in the process, HIV and other retroviruses convert their genetic information into DNA and insert it into the genome of their host – a human immune cell known as a T cell.

Some infected cells use this viral DNA to make many copies of the virus, which go and infect new hosts. But others go into a resting state: the viral DNA still lurks within the cell’s genome but it doesn’t produce new viruses. This is known as a latent infection. These latent, hidden viruses then get reactivated at a later date, starting up an active infection that can develop into full-blown AIDS.

Up to 10 in every million immune cells in an HIV patient may harbour a latent infection. Current HIV treatments only work on active viruses, so researchers have been designing drugs that can ‘evict’ the hidden viruses from their DNA home and make them vulnerable to therapy. In theory, HIV can settle into any location within the genome, and it should be equally easy to remove it from any place. But none of these treatments can shift all the latent viruses, so HIV infection is still impossible to fully cure.

Now Guillaume Filion and his team at the CRG have developed a way of tracking down the exact location of latent HIV in the genome, revealing important clues about why some are harder to reactivate than others.

“We have known since the 1930s that some genes are more active than others, depending on their location in the genome,” Filion explains. “So it makes sense that the same would be true for HIV – that it would be easier to integrate or reactivate from certain locations compared with others.”

When it comes to deciding where to live, location is everything. If you’re moving to an exciting, vibrant city you’d probably choose an area with bustling cafes, bars and shops, rather than a dead zone where everything is closed all the time. And, as Filion and his team have found, HIV has the same idea when it moves into human cells.

Cracking the barcode

One major problem with studying latent HIV infection is that a single cell can have multiple viruses embedded in many places within the genome. This makes it hard to identify the location of each specific virus and work out which ones are reactivated or remain hidden after treatment.

The trick to Filion’s new method lies in a genetic ‘barcode’ – a unique sequence of DNA that the researchers embed in the genetic code of individual virus particles. Once the viruses have inserted themselves into the genome of the host cell (in this case human immune cells grown in the lab) the researchers then extract the cellular DNA and sequence it to find the locations of the latent viruses. And because each virus has its own unique barcode, it’s easy to see where each one has embedded. The barcode can also be used to see whether a particular virus has become active and is being ‘read’, in a similar way to the host cell’s regular genes.

“We have been able to use our technique, which we call B-HIVE, to make a map of where HIV likes to go,” Filion says. “It is very biased and prefers active regions of the genome – it wants to go where the action is!”

Even amongst the active regions of the genome, the researchers found that some particular areas in the genome are much more desirable than others. For example, some ‘addresses’ are 100 times more likely to be home to a virus than another, although why they are quite so popular remains a mystery.

As might be expected, the exact location of HIV in the genome affected whether the virus was likely to become active again or not. Viruses that settled down near the ‘control switches’ responsible for turning on genes were much more active than those in other places. Even so, as Filion points out, there is still a lot of unexplained variability.

“We assumed that the insertion site in the genome would be the main determinant of HIV activity, but we have seen the virus go into the same place in different cells and then behave differently. So there is something else going on that we don’t yet understand.”

Moving on

As well as spotting patterns in the places that HIV likes to live in the genome, Filion and his colleagues also noticed key differences in how viruses in specific locations responded to two reactivation drugs, vorinostat (VOR) and phytohaemagglutinin (PHA), which each have different modes of action. Intriguingly, they found that viruses in one set of genomic ‘addresses’ were more likely to respond to VOR, while those that had inserted in other locations were preferentially reactivated with PHA.

This suggests that any treatment aimed at treating latent infections might need to combine a cocktail of several different drugs with separate actions in order to flush out all the dormant viruses. And as genetic techniques improve, it might be possible in the future to work out the best combination of reactivation drugs based the different types of viral locations within an individual patient’s immune cells, ensuring that all the latent viruses are evicted from every part of the genome.

For Filion and his team, the next step is to see whether they can apply their barcoding technique to label latent HIV DNA within immune cells in animals or even taken directly from infected patients, rather than cells grown in the lab. There are hopes that new gene editing technology, known as CRISPR, might make this possible, although there’s a lot more work still to be done.

“The endgame here is not that there will be one drug that can cure HIV, and it may never be possible to reactivate and get rid of every single virus in the body,” says Filion. “More importantly, our work helps us to understand the complex relationship between the virus and the host genome – each one is a very small needle in a very big haystack, but we can now find exactly where they are hiding.”

Reference work

Chen HC, Martinez JP, Zorita E, Meyerhans A, Filion GJ.

“Position effects influence HIV latency reversal.”

Nat Struct Mol Biol, 24(1):47-54. doi: 10.1038/nsmb.3328. Epub 2016 Nov 21.