Debunking myths on genetics and DNA

Thursday, December 22, 2011

Guest post: How to camouflage a virus and why it's important

Last week I covered a couple of recent gene therapy studies and discussed the different types of vectors used in order to make these therapies more efficient. One of the obstacles that hinders the efficiency of gene therapy is the immune system: if the patient has previously developed immunity against the viral vector, the virus will be quickly cleared out of the system without being able to deliver the genes. Therefore, the question is: how can we prevent the immune system from attacking the viral vector?

One of my regular readers here on the blog, antisocialbutterflie is an expert on "capsid recognition," and kindly offered to discuss the topic. The capsid is the outer shell of the virus and the idea is to disguise it so the antibodies won't recognize it. Without further ado, I'm going to let my guest take over the post.

The monumental advances in genetics over the last six decades have improved our understanding of diseases and where they originate. Now armed with this knowledge there are many disorders that we could fix if we had a way to deliver the right DNA. In many ways gene therapy is like the Cold War. We have nuclear warheads but it isn’t going to do any good if we don’t have an effective delivery system. A delivery system must first be able to hit the target but it must also allow the device to do its job and do it covertly enough that it doesn’t call up the defenses of its enemy too quickly. In the case of gene therapy you can’t just inject someone with naked DNA and expect something to happen. Thankfully nature helped us out by providing a fabulous rocket powered, laser guided missile system in the form of viruses.

You typically think of viruses in terms of disease but there are plenty of viruses that coexist without any the deleterious effects. The virus of choice for many gene therapy trials is the adeno-associated virus or AAV. AAV is a single stranded DNA virus indigenous to primates from the family parvoviridae, which may sound vaguely familiar if you’ve ever had a pet. This is the same family of viruses that give us canine parvo and B19 in humans, but in the case of AAV it doesn’t really do much of anything. In fact AAV can’t even replicate unless its host is also infected with a helper virus like adenovirus or herpes. They are also mildly immunogenic so they can travel (mostly) under the immune system’s radar without eliciting a strong response.

The virus only has two genes that can make seven proteins when you factor in alternate start sites and splice variants. Most gene therapy vectors utilize the cap gene and their resultant 3 VP proteins to create a viral shell for the therapeutic gene they want to deliver. Sixty of these proteins interlock together to make an icosahedron. Basically the pieces, when assembled, look like 20-sided dice with a therapeutic gene stuffed inside. As you might imagine the core fold of the protein has to be fairly specific to fit together but there are loopy bits on the exterior of the capsid called “variable regions” that give each variant its unique properties. These variable regions act like velcro to grab onto cell receptors. In multicellular organisms different tissues express different receptors so you can use these regions to target a specific cell type. Unfortunately it is also these variable regions that immune cells target for neutralizing antibody production. Since these viruses target humans, antibodies already exist to the naturally occurring variants, which is why many of the gene therapy vectors have to be tweaked to be effective.

There are several ways of engineering a good gene therapy vector that can escape the immune system while still targeting the tissues that you want. One strategy is to mix and match. This can be accomplished by mixing up whole capsid proteins from several serotypes or cutting out bits and pieces of different protein sequences and pasting them together to make a single new one. There is also directed evolution where you introduce errors in the gene to create a new protein sequence and then select for the mutations that confer the appropriate tissue specificity while still escaping the neutralizing antibodies. Another strategy involves adding a peptide to the capsid that you know binds a specific receptor as a way to target it. You can even create protective coats for your virus made from things like PEGs or lipids to allow it to evade the immune system in the same way enveloped viruses do. All of these alterations run the risk of blocking capsid assembly, changing tissue specificity, and/or reducing infectivity so it’s a bit of a crap shoot as to whether your efforts are all for naught. For every successfully engineered vector there are hundreds that didn’t work.

As mentioned in a previous post, there is also the issue of gene size. There is a finite amount of space inside that d20 and some of it has to be taken up by the inverted terminal repeat sequence that packages the DNA. If you have a bigger gene you need a bigger virus like adenovirus, HSV or HIV. These bigger viruses are made up of several capsid proteins that have to be accounted for and provoke a bigger immune response making the can of worms even larger.

I hope this enlightens you a bit to the field of gene therapy vector design and its challenges. Thanks to Elena for having me.

That was not only enlightening, but also exceptionally clear -- thank you! It was fascinating to learn that one of the ways this is accomplished is by introducing artificial mutations in order to trick the immune system... Another lesson learned from viruses themselves, as that is exactly how HIV escapes our immune defenses.

Antisocialbutterflie is an X-ray crystallographer with a PhD in biomedical science. She is currently a postdoc and dreaming of a day when she can step away from the bench, preferably before she sets fire to her FPLC. In addition to the random blog comment she occasionally writes fiction when time permits.

Pulicherla, N., & Asokan, A. (2011). Peptide affinity reagents for AAV capsid recognition and purification Gene Therapy, 18 (10), 1020-1024 DOI: 10.1038/gt.2011.46


  1. Thanks Antisocialbutterflie (and Elena).

    If an artificial mutation allows a gene-therapy virus to escape the immune system and complete its task can the same viral vector be used against the same patient again or will the immune system develop a response? Or, does it vary from patient to patient--some develop a response fairly quickly and a new mutation has to be used and some don't?

  2. Thanks for your question, Steve. I'll let Antisocialbutterflie pitch in. :) What I've seen with HIV is that if the mutation escapes the immune system, then the virus is pretty much free to do what it needs to do. It's also true that the immune system keeps trying to get to it, though, so probably if you keep the virus in there long enough without mutating it again you _might_ start seeing a response...

  3. Hi Steve, good questions.

    Like Elena mentioned with HIV an escape mutant stays an escape mutant because it has specifically evolved for that purpose. With AAV there is a very real risk of the patient developing immunity to the vector. Usually this is circumvented by larger subsequent doses but the hope is that an array of different capsid proteins with the same tissue specificity can be developed like a gene therapy buffet to address this very issue. In a perfect situation the gene sequence will integrate into the host genome and it occur in a cell with longevity so one dose will be enough, but as we know the world is rarely perfect.

    To address your second question, you are correct, the degree of immune response is very patient specific. Originally they thought that AAV was entirely non-immunogenic which led to some unfortunate repercussions once human clinical trials started. Most people are exposed to AAVs in early childhood, but even then the effective dose can vary widely. That's why you usually read about one person who was successfully treated by gene therapy when there were a dozen in the study. It's really hard to get statistical significance or a standard working protocol which is why many of these studies stall in clinical trials.

    I hope this answers your question.


Comments are moderated. Comments with spam links will be deleted and never published. So, if your intention is to leave a comment just to post a bogus link, please spare your time and mine. To all others: thank you for leaving a comment, I will respond as soon as possible.