There’s a rather large breakthrough in CRISPR gene editing – yeah, another one – that has downstream implications for drug discovery. A group at the Salk Institute reports in Cell that they’ve found a way to do both loss- and gain-of-function without double-stranded DNA breaking. That’s quite different from the various “classic” forms of CRISPR, and for all its utility, there’s always been concern about leaving a trail of DNA damage during the process (especially in human therapeutic use).
So I need to back up, because this really isn’t “gene editing” any more; the sequences haven’t changed. What it is, is transcriptional editing. The technique uses shorter guide RNAs (which keep Cas9 from actually cutting the DNA) along with a set of transcriptional activators to zero in on a gene of interest and start it reading off. This was accomplished by cramming the guide RNA sequence(s) and transcriptional machinery carefully into a single AAV viral vector (previous systems tended to be too large to accomplish this). When this is introduced into a Cas9-expressing mouse, voila – transcriptional effects where you want them. OK, here comes a blast of detail:
Previous attempts to do this sort of epigenetic transcriptional-level targeting had not worked very well in adult animals (or at all), particularly gain-of-transcription experiments. But this one looks pretty impressive. The group shows effects in several model systems: induction of mouse follistatin (Fst) increased muscle mass, as it should, induction of either IL-10 or Klotho ameliorated kidney damage (as they should), activating Pdx1 in the liver turned enough cells into insulin-producing ones to partially reverse a Type I diabetes model, and expression of either Klotho or utrophin rescued a known mouse model (Mdx) for Duchenne muscular dystrophy (both of which had previously been demonstrated by other groups by other means). Finally, they turned around and demonstrated similar effects when the Cas9 was also delivered by a viral vector (instead of being wired into the mice), and this (importantly) could also be done with the nuclease-dead version of the enzyme. Another interesting finding (with the Pdx experiment) was that histone marks that generally show up around activated genes (a methylation, H3K4me3, and an acetylation, H3K27ac) both increased on treatment, showing that this method really does seem to recapitulate the endogenous process, rather than hijacking things and dragging them over.
OK, that’s a lot of CRISPRing in one paragraph, and it’s even money or worse whether anyone who’s not already conversant with the field has even made it down this far. But the take-home is that this looks like a very useful system to manipulate animal model genes at the very least, and it also has immediate implications for people trying to extend CRISPR to human therapy. Here’s how the authors wind it up:
In summary, the in vivo TGA system described here can be used to transcriptionally activate endogenous genes (either single genes or combinations of genes), including extremely large genes. This system can be used to express genes to compensate for disease-associated genetic mutations, or to overexpress long non-coding RNAs as well as GC-rich genes to reveal their biological functions, a difficult feat until now (La Russa and Qi, 2015, Vora et al., 2016). Finally, combined loss- and gain-of-function manipulations can be applied to rapidly establish epistatic relationships between genes in vivo. In the future, improvements in the specificity of the system by designing tissue-specific promoters, and by using AAVs with specific tropism, will enhance and widen the applicability of the system. Thus, the versatile and efficient in vivo CRISPR/Cas9-mediated gene activation system introduced here holds great promise, both as a tool for in vivo biomedical research and as a targeted epigenetic approach for treating a wide range of human diseases.
That last line is the sort of thing many papers promise, but I think that in this case, it has a good chance of being true. This may be the most powerful gene-manipulation toolkit that has yet been described, and you can expect to see a lot of work on it in the coming months as other groups give it a shakedown. It will also be quite interesting to see what the Salk’s strategy is for positioning this for human therapy – will they do licensing/codevelopment deals with existing companies, or start one of their own? Because someone’s going to go after human therapy with this technique, one way or another.