In a setback for a novel approach to curing HIV, researchers involved in a technique that uses enzymes to remove viral genes from the DNA of infected cells have found that HIV rapidly develops resistance to the guide molecules that target the correct part of the DNA sequence. The resistant viruses that develop can in some cases replicate even faster than viruses not exposed to the gene therapy (though they are still susceptible to conventional antiretroviral (ARV) drugs).
Furthermore, the researchers suggest, the way the gene therapy works may actually promote the development of resistance, as it actively creates small mutations at the site where it bisects the cellular DNA. Resistance arises rapidly – within 8-10 days of the therapy first starting to work.
This does not mean the whole gene-splicing approach is doomed to failure, but it does imply that the gene-degrading enzyme would need to be attached to a variety of different gene probes, designed to attach to a number of different points on the viral DNA hidden within the human DNA inside infected cells.
The gene technology involves transporting a DNA-degrading enzyme called CRISPR or cas9 into the heart of the nucleus of human cells. The cas9 enzyme, which was originally found within bacteria as a natural defence against viruses, is attached to a single-strand length of ‘guide RNA’ (sgRNA) that guides the cas9 to the particular piece of rogue DNA that needs to be removed.
The concept is not dissimilar to the versatile gene therapy called short-interfering RNA (siRNA), which is being investigated for a number of diseases including chronic hepatitis B. But whereas siRNA targets and degrades the RNA messenger and component molecules that act as the replication machinery of viruses within the main part of the cell (cytoplasm), sgRNA/cas9 targets integrated DNA, the ‘master template’ for viral manufacture that retroviruses like HIV insert into a cell’s core genetic instructions, and which exists in the nucleus, not the surrounding cytoplasm.
In this lab-dish study the researchers infected T-cells with three different sgRNA/cas9 gene probes designed to target different sections of integrated HIV DNA.
One, called T4, attached to the gag/pol area of the HIV genome, which includes HIV’s ‘copying machinery’; a second, T10, attached to the env/rev area, which includes the instructions for making the viral envelope. In these two cases the sgRNA attached itself at a single specific point in the viral DNA and cleaved it in two, allowing cas9 to degrade the frayed ends.
An intracellular molecular repair mechanism called NHEJ (non-homologous end joining) eventually repairs the frayed ends of DNA, but this is prone to copying mistakes and introduces changes into the DNA chain, some of which may confer resistance.
The third gene probe, called LTR-B, was similar to the one used in a study we reported on at the end of March. This cuts the DNA at two points, which are located at or near the sequences called LTRs (long terminal repeats) that represent the end points of the HIV genetic material. It thus performs a complete removal of all HIV material from the cell.
One implication of LTR-B resistance is that if the cell remains able to produce HIV, it can't be because there is some remaining defective DNA that is nonetheless capable of producing infectious virus. It must be because the HIV DNA has mutated into a form that is resistant to attachment by the sgRNA in the first place.
Results and implications
Production of viral particles was compared between cells treated with sgRNA/cas9 and cells treated with cas9 alone. The researchers found that, as expected, viral reproduction was initially severely impaired in cells infected with the sgRNA/cas9 gene probe. Overall, peak viral production levels were 83% lower in cells treated with T4, by 95% in those treated with T10, and by about 98% in the cells treated with LTR-B.
Viral production started about five days after infection in control cells but was delayed by about four days in cells treated with LTR-B and by about ten days in cells treated with T4 or T10. There was, however, significant viral production in the end. In the LTR-B cells, viral levels at the peak pf production – which was also delayed by four days – was still 55% lower than in control cells. But in T10-treated cells it was exactly the same as in control cells – though delayed by eight days – and was actually about 20% higher in T4 cells. This suggests that viral resistance happens rapidly.
In samples of virus, 74% of T4, 70% of T10 and 72% of LTR-B viruses had had their DNA altered in the way expected by the three different genetic probes. In the other 16%, 20% and 18%, there were mutations of some sort, some of which conferred resistance.
The researchers performed the experiment again on cells infected with HIV viruses taken from the peak of viral production in the previous experiment, which were all expected to be resistant. The resistant T4-treated viruses again produced 20% more virus than control cells and in the case of T10-treated cells, they produced the same amount of virus as control cells, but actually reached the peak of viral production six days earlier. This suggests that HIV that had become resistant to T4 or T10 was at least as reproductively fit as control virus, or fitter.
In the case of the LTR-B treated with resistant virus, the peak of viral production was also reached about six days earlier than control virus but viral production levels remained about 30% lower, suggesting that LTR-B resistant virus might be paying a small reproductive price for its resistance.
In T4-resistant viruses, there was a predominant single-point mutation (a mutation with just one genetic ‘letter’ changed). This represented 81% of resistant viruses, and another single-point mutation represented 13% of them. In T10-resistant virus, while one single-point mutation represented 38% of resistant viruses, the other resistant viruses had in general more complex 3- or 4-point changes.
The resistance in the LTR-B resistant viruses was more unusual. T4 and T10 remove only small parts of the viral genome; if it is then repaired inaccurately, this can create resistant strains. In this case, it is the human NHEJ cellular machinery, that rejoins the frayed DNA inaccurately, that is the core cause of resistance. But LTR-B should remove the entire viral genome, so the rejoined viral DNA should not have the chance to even start producing mutated viruses. Where was the resistance coming from?
The researchers found that a small proportion of the T4 and T10-resistant viruses, but all the LTR-B resistant viruses, did not have substitutions of one genetic base for another like most resistant viruses, but had whole sections either removed or inserted into the genetic code – so-called ‘indels’.
What this suggested was that it was the sgRNA/cas9 itself that was causing the resistance in these cases. The cas9 was deranging the DNA at the cleavage site in such a way that it was producing resistant viruses itself.
In this case, resistance was not being caused by viral turnover in the presence of low levels of drug, that exert a selective evolutionary pressure – it was being caused directly by the drug itself. In short, sgRNA/cas9 was acting as a mutagen, a direct driver of viral mutation.
Some unconventional antiviral drugs work by being mutagens, such as ribavirin, but in this case they work on the viral genome. siRNAs can also introduce mutations directly into cytoplasmic RNA. But this is the first time a proposed HIV treatment has been seen directly contributing to the production of resistant mutations within the proviral DNA at the heart of cells.
While the production of resistance would be brought down to a minimum by the use of multiple sgRNAs, as the researchers suggest, this study shows that unexpected setbacks may lie in wait for researchers on the way to a cure for HIV and that the novel gene-editing and other techniques involved could pose risks of their own.
Wang Z et al. CRISPR/Cas9-derived mutations both inhibit HIV-1 replication and accelerate viral escape. Cell Reports, 15, pages 1-9. DOI: http://dx.doi.org/10.1016/j.celrep.2016.03.042. April 2016.