Snipping the HIV Genome Out of Latently Infected Cells

One of the more futuristic-sounding ideas for curing HIV infection involves trying to remove the genome of the virus from the genome of the cells into which it has integrated. On paper, the idea is very appealing, but there are a multitude of challenges associated with trying to identify integrated HIV DNA (referred to as proviral DNA or provirus) and then excise it from the DNA of an infected cell without causing untoward effects. In 2014, the research group of Kamel Khalili at Temple University in Philadelphia drew extensive news coverage when they reported some success in laboratory experiments; this work and the media response were covered at the time on TAG's media monitor page. Khalili and colleagues have now published a new paper and again have generated considerable press (in broad terms, TAG's previous commentary on interpreting the research and associated stories remains relevant). In addition to Khalili's new findings, another paper has been published recently that describes a slightly different approach toward excising HIV proviral DNA, also reporting encouraging results but similarly limited to the laboratory setting.

The paper from Khalili's group is published in the open access journal Scientific Reports. The researchers employed CRISPR/Cas9, a DNA editing approach derived from bacteria, and targeted it to relatively conserved regions of the HIV genome, successfully excising proviruses from some (but not all) infected CD4 cells in laboratory cultures. Delivery of the HIV-targeted CRISPR/Cas9 using a lentivirus vector also had a protective effect on uninfected CD4 T cells, which the researchers suggest was likely due to editing of pre-integrated HIV DNA. Additionally, the technology significantly reduced HIV replication (as measured by p24 protein production) in CD4 cells sampled from HIV-positive individuals; this activity appeared to be mediated both by provirus excision and the induction of crippling mutations in proviruses that were not eliminated.

The second paper, by Janet Karpinski and colleagues from the Technische Universität, Dresden, also prompted some news articles. These researchers modified an enzyme from bacteria called a Cre recombinase so that it targeted a region of the HIV genome that is conserved in approximately 90% of virus isolates from subtypes A, B and C. The resulting enzyme, dubbed broad-range recombinase 1 (Brec1), was then tested for activity in laboratory cultures and humanized mice. As in the CRISPR/Cas9 study, delivery was via a lentivirus vector. The researchers report that Brec1 reduced virus production by infected CD4 cells sampled from HIV-positive individuals, and depleted proviruses from these cells.

Additional experiments were conducted in which infected cells from HIV-positive donors were transduced with Brec1 or a control vector and then transferred into humanized mice; similar to the in vitro results, viral loads progressively declined to undetectable levels in recipients of the Brec1-transduced HIV-infected cells, but not in the controls. The researchers also transduced human stem cells with Brec1 or a control and used these cells to generate a humanized immune system in immunodeficient mice. The mice were subsequently challenged with HIV, and recipients of the stem cells transduced with Brec1 displayed viral load declines and evidence of depletion of HIV proviruses, in contrast to the controls.

There are a number of overarching issues relating to these gene-editing technologies that are addressed by both groups of researchers in their papers. Chief among them is safety: manipulation of the genome carries the risk of altering genes in ways that might lead to cancer or other problematic alterations in gene function. Extensive analyses were conducted to look for evidence of potentially dangerous off-target effects in cells exposed to CRISPR/Cas9 or Brec1, and the researchers report that the approaches appeared safe, while noting that these in vitro assessments have limitations. Karpinski and colleagues explain that the presence of two HIV proviruses integrated in a single cell could be a concern, because rather than excising a single provirus and stitching the genome back together at the points where the provirus had been located, the gene-editing approaches could potentially excise all the genes between the locations of the two separate proviruses. A number of reasons are offered as to why the risk of this occurrence is likely to be low, but it nevertheless needs to be borne in mind.

Beyond safety, another major challenge facing researchers trying to develop these gene editors into therapies is delivery to the cells where they are needed. While Khalili and colleagues have issued press releases promoting their work in fairly glowing terms (leading to the extensive news coverage), they have very little to say in their paper about how it might be delivered in vivo, simply stating: "improved delivery of CRISPR/Cas9 will be required to target the majority of circulating T-cells." They do not mention that, because Cas9 is a bacterial protein, the human immune system is likely to recognize it as foreign and generate immune responses against it (this problem has already been described in mice).

Karpinski and colleagues offer more discussion regarding the problem of delivering their Brec1 approach, suggesting it could be used to genetically modify stem cells, which would then be transferred into HIV-positive people in the hopes of generating HIV-resistant CD4 T cells (similar to trials that are being conducted by Calimmune and Sangamo BioSciences using gene-editing approaches that knock out the CCR5 co-receptor). They also mention the possibility of using adeno-associated virus (AAV) vectors to target delivery of Brec1 to central memory CD4 T cells, where latent HIV most commonly resides. Again, however, the risk that the bacteria-derived Brec1 enzyme might provoke an immune response is not discussed.

HIV's notorious genetic instability, leading to the presence of many virus variants in HIV-positive individuals, also presents a significant hurdle for excision approaches. The DNA-cutting enzymes are guided to their target by recognition of specific HIV sequences and thus could be stymied by sequence variations. Khalili and colleagues acknowledge this issue, and suggest that it would likely need to be addressed by "analysis of the HIV-1 quasi-species harbored by patients' CD4+ T-cells and design of suitable, i.e. personalized CRISPRs" -- a requirement that could clearly have implications for turning the idea into a practical therapy.

Overall, these are exciting technologies that have understandably generated a lot of interest regarding their potential application to eliminating HIV from latently infected cells. But, without wanting to be overly naysaying, the media coverage has probably not conveyed how serious the challenges are when it comes to translating the promising laboratory findings into therapies that could feasibly be administered to people. The road to human trials is likely to be long, and there may turn out to be obstacles that are insurmountable. On a more optimistic note, there is widespread interest in developing gene-editing technologies to treat a vast range of different diseases, so many scientists are currently engaged in findings ways to make them more amenable to clinical use (a recent open access review in the journal Molecular Therapy offers an overview, albeit a fairly technical one, of the technologies now under investigation).

UPDATE 4/4/2016: Last Friday, The Daily Telegraph published a horrendously inaccurate article claiming that this research may lead to an HIV cure "within three years" -- see response published today by Ben Ryan for POZ Magazine and TAG's media monitor page for additional information.

Also, since writing this post I've learned that Kamel Khalili has founded a company, Excision BioTherapeutics, which is partnering with Temple University with the aim of developing and commercializing the approach.

UPDATE 4/6/2016: Thankfully_The Daily Telegraph_ has now edited the headline and first paragraph of their article to remove the mistaken claim that the approach may cure HIV infection "within three years" or "a few years."

Richard Jefferys is the coordinator of the Michael Palm HIV Basic Science, Vaccines & Prevention Project Weblog at the Treatment Action Group (TAG). The original blog post may be viewed here.