First clinical trial of human gene editing

Get ready, world, scientists are going to use CRISPR/Cas9 on human patients for the first time, extracting a population of cells, modifying their genomes, amplifying them in tissue culture, and then injecting the modified cells back into the human host. It’s being done in China, where the ethical constraints are a bit more loose, which isn’t always good…but in this case, it sounds like a good, safe (as safe as experimental therapies can be) approach.

They intend to use CRISPR/Cas9 to knock out the PD-1 gene in immune system cells. PD-1 is a cell surface molecule on T-cells that inhibits the cells, and acts as a constraint on immune system activity. The “PD” is short for “Programmed Death”, and what it does is compel the cells to commit suicide when stimulated — so if immune system cells get a bit overzealous and go on a rampage attacking healthy cells, they can be switched off. The immune system has multiple checkpoints to prevent it from going rogue, and this procedure will remove one of them. By knocking out the PD-1 gene, the scientists are creating particularly unrestrained cells that they hope will do a more effective job killing cancer cells, because cancer cells are known to use the signaling mechanisms that tell the immune system to die.

Are there drawbacks and risks? There are always drawbacks and risks. This technique is a variation on an existing pharmaceutical approach, which uses drugs that inhibit PD-1 in cancer patients, so we know a bit about its effects — it’s just that taking out the whole gene with CRISPR/Cas9 is a dramatically thorough way of demolishing the molecule. But we do have some drugs, like Nivolumab and Pembrolizumab, that target PD-1 already and are in clinical trials. We’ve also experimentally knocked out the gene in mice.

So, we have an idea of what could go wrong, and in the immortal words of Dr House, it’s lupus. Or lupus-like effects. By jacking up the immune system and removing one restraint on its activity, you can get complex system-wide problems, which Dr House will tell you are pretty hard to treat, but at least they’re not as severe as terminal cancer. They are also editing a terminal cell type — it’s not going to proliferate — so eventually, we hope after they’ve killed cancer cells, the injected cells will die of natural causes and the effect will fade away.

This is not a treatment that affects the germ cell line, so these patients, if they survive, will not be passing on an edited gene to their offspring. It’s also got to be a rather expensive therapy that has to be customized for each new patient, so it’s not going to be routine. It is a first step into the exciting world of genetically modifying humans, though.


  1. JustaTech says

    This is very exciting! There are some immunotherapies on the market or in clinical trials that do patient-specific cells, so that’s a mostly solved problem. I think this will have an interesting impact on the new CAR-T cell therapies (the chimeric antigen receptor T-cell therapies which require cloning out T-cells) if it is successful.

    But given how complex (insane) the immune system is, and how it can still be kind of unpredictable (which mostly means we have a lot to figure out still) I do think there is a lot of risk to this trial. I profoundly hope all the patients survive the therapy (even if they don’t survive their disease) but I think it’s pretty likely someone will die.

  2. imaginggeek says

    PZ, there is an error in your post. Although T cells do represent a terminally differentiated cell type (e.g. they are one of the termini within the lymphopoiesis cascade), they are absolutely capable of further cell division. Indeed, cell division is a key aspect of how T cells function – only about 1 in every hundred thousand or so T cells will recognise antigens from a particular pathogen. As such, following antigen recognition, the first things T cells do is undergo massive proliferation to build their numbers up to levels where they can engage in a meaningful immune response. Activated T cells are among the fastest dividing cells in our bodies, replicating an estimated 2-3 times per day. Likewise, T cells can also differentiate into long lived (e.g. years) memory cells, meaning that once formed, they can stick around for a long time.

    This is one reason why we’ve been slow to test therapies like this – other genetically-altered T cell therapies (e.g. chimeric receptor T cells/CARs) have been tested in clinical trial, and while successful, the ability of the cells to proliferate and their long life spans have led to pretty severe long-term issues in patients treated with them. For example, one of the first CAR clinical trials targeted a B cell malignancy, and while successful, treated patients lost all antibody production, with this immune deficit persisting as of a 5 year follow-up. Autoimmunity has been observed in patients treated with PD-1 blocking antibodies, and as such this remains a real concern with T cells depleted of PD-1. There has been discussion of including in future CARs (and presumably in CRISPR approaches such as these) inducible terminator genes to rid the patient of the cells once their job is done, but AFAIK no practical approach to this has yet been demonstrated.

    That’s not to say that this approach is doomed to failure – the choice of somewhat controllable autoimmunity versus death by cancer is an easy one to make. But we’ve got a long ways to go before these treatments will be ready for the front lines.