CRISPR is usually thought of as a laboratory tool to edit DNA in order to fix genetic defects or enhance certain traits—but the mechanism originally evolved in bacteria as a way to fend off viruses called bacteriophages. Now scientists have found a way to adapt this ability to fight viruses in human cells.

In a recent study, Catherine Freije, Cameron Myhrvold and Pardis Sabeti at the Broad Institute of the Massachusetts Institute of Technology and Harvard University, and their colleagues programmed a CRISPR-related enzyme to target three different single-stranded RNA viruses in human embryonic kidney cells (as well as human lung cancer cells and dog kidney cells) grown in vitro and chop them up, rendering them largely unable to infect additional cells. If further experiments show this process works in living animals, it could eventually lead to new antiviral therapies for diseases such as Ebola or Zika in humans.

Viruses come in many forms, including DNA and RNA, double-stranded and single-stranded. About two thirds of the ones that infect humans are RNA viruses, and many have no approved treatment. Existing therapies often use a small molecule that interferes with viral replication—but this approach does not work for newly emerging viruses or ones that are evolving rapidly.

“CRISPR” refers to a series of DNA sequences in bacterial genomes that were left behind from previous bacteriophage infections. When the bacteria encounter these pathogens again, enzymes called CRISPR-associated (Cas) proteins recognize and bind to these sequences in the virus and destroy them. In recent years, researchers (including study co-author Feng Zhang) have reengineered one such enzyme, called Cas9, to cut and paste DNA in human cells. The enzyme binds to a short genetic tag called a guide RNA, which directs the enzyme to a particular part of the genome to make cuts. Previous studies have used Cas9 to prevent replication of double-stranded DNA viruses or of single-stranded RNA viruses that produce DNA in an intermediate step during replication. Only a small fraction of RNA viruses that infect humans produce such DNA intermediates—but another CRISPR enzyme, called Cas13, can be programmed to cleave single-stranded RNA viruses.

“The nice thing about CRISPR systems and systems like Cas13 is that their initial purpose in bacteria was to defend against viral infection of bacteria, and so we sort of wanted to bring Cas13 back to its original function—and apply this to mammalian viruses in mammalian cells,” says Freije, who is a doctoral student in virology at Harvard. “Because CRISPR systems rely on guide RNAs to specifically guide the CRISPR protein to a target, we saw this as a great opportunity to use it as a programmable antiviral.”

Freije and her colleagues programmed Cas13 to target three different viruses: lymphocytic choriomeningitis virus (LCMV), influenza A virus (IAV) and vesicular stomatitis virus (VSV). LCMV is an RNA virus that mostly infects mice—but it is in the same family as the virus that causes Lassa fever, which is found in West Africa and is much more dangerous to study in the lab. IAV is a flu virus; although some antiviral medications for flu already exist, such viruses evolve rapidly, so there is a need for better options. Finally, VSV is a model for many other single-stranded RNA viruses.

To determine how effective Cas13 was at destroying the viruses, the researchers also used it as a diagnostic tool to see how much viral RNA was being released from infected cells. They saw a twofold to 44-fold reduction in RNA, depending on which virus they were looking at and the time point. They also looked at how well the released RNA was able to go on and infect new cells. In most cases, they saw a 100-fold reduction in infectivity—and in some cases, more than 300-fold—according to Freije. The findings were published online on October 10 in Molecular Cell.

“The results are very impressive,” says Chen Liang, a professor at the Lady Davis Institute at Jewish General Hospital and the department of microbiology and immunology at McGill University in Montreal, who was not involved in the study. His own laboratory has used the Cas9 enzyme to deactivate DNA viruses. The concept is very similar, but Cas13 has a few advantages, he says. For one, Cas13 can be used to target one virus using several guide RNAs, making it difficult for the virus to “escape.” Secondly, the new study also used Cas13 to detect how much viral RNA was left over to infect cells. The amount of viral knockdown the group achieved is “very significant,” Liang says. “If you can target and inactivate all three [of these] viruses, in principle, you can inactivate any virus.”

As with any approach, there are limitations. One is the question of how to deliver the Cas13 to target a virus in a living person, Liang notes, and the researchers have not yet done any animal studies. Another is the fact that viruses will eventually develop resistance. But Cas13 has an advantage here: when Cas9 cuts viral DNA, mammalian cells repair it and can cause mutations that make the virus more resistant. Yet with Cas13, these cells do not have the mechanism to repair the RNA and introduce errors that would help the virus escape being destroyed. Even if a virus does evolve resistance, or if a new virus is encountered, the method could be quickly adapted. 

“One of the things that’s most exciting about this approach is the programmability,” says Myhrvold, a postdoctoral fellow at Harvard. “Once you figure out how to do this well for one virus it’s not that hard to design sequences against another virus—or another one. Furthermore, if the virus changes its own sequence—as viruses are known to do, just during an outbreak or in response to therapy—you can very easily update the CRISPR RNA sequence and keep up with the virus.”

Freije agrees. “We are definitely excited about future prospects of optimizing the system and trying it out in mouse models,” she says. Beyond therapeutics, the team hopes to understand more about how viruses operate—how they replicate and what parts of their genomes are most important. Using approaches like this, “you can really start to get a better picture of what parts of these viruses are and, most importantly, what really makes them tick.”