Cells edited using CRISPR–Cas9 injected into a person for the first time | Hacker News Reader

Cells edited using CRISPR–Cas9 injected into a person for the first time

nature.com

201 pointssnake11752 comments

So this title is a bit misleading; something like, "cells edited with CRISPR injected into a person for the first time" would be better. While CRISPR is promising for topological treatments, that's not what happened here.

The team took white blood cells out of a patient's body, used CRISPR to knock out a gene which suppresses immune response, then injected those same cells back into the patient, hoping that they would attack the cancer without that inhibiting gene.

If this were a live CRISPR treatment in a human, it would probably make more sense to just knock out the activated oncogene(s) in the patient's cancer cells and/or repair the deactivated tumor suppressant gene(s).

> it would probably make more sense to just knock out the activated oncogene(s) in the patient's cancer

I've been wondering about that for the past couple of years, since I heard of CRISPR basically. Who is working on this aspect most seriously right now? Is there any interesting published progress? I understand it will take a while to make real progress, but mostly I'm just curious as an outsider to the industry.

I don't think anyone is; it's still too far out there. I'm sure that there's plenty of research building towards it, but I couldn't name any names. I'm also an outsider to the industry, but CRISPR is actually a simple protocol that doesn't require much advanced equipment, and I've been looking into using it for some home science projects.

The problem as I understand it is that right now, CRISPR requires some manual steps in the lab to make it work well. You can design the protein and guide RNA and spacers until you're blue in the face, but if you can't find a specific enough cutting site on the genome, or one that is close enough to your target, or something like that, you'll wind up with a lack of specificity in that cut, which can lead to unwanted mutations. And since the CRISPR system was originally a sort of bacterial immune system, it's really geared towards 'knocking out' specific genes by (I think, but I'm really shaky on this part) introducing a bunch of extra mutations when the cut it makes in the DNA gets repaired. But there are two kinds of repair mechanisms, and apparently there's a way to encourage the more accurate one, although this is really an area I need to read more about.

Anyways, the technique has also been used to introduce entirely new genes, but I think that involves using modified CRISPR proteins called 'nickase's to make single-stranded cuts in the DNA, rather than double-stranded ones. You introduce plasmid DNA with your genes to match up with the cleavage sites, and ideally it gets taken up.

Whatever the approach, you still aren't going to get 100% expression or transfection, and stable transfection either requires invasive techniques like biolistics (shooting DNA through membranes on accelerated nanoparticles) or delivery by viruses, which has a lot of potential, comparatively minimal side effects, and can even be targeted somewhat to certain types of cells. But I think the FDA is nervous about, especially in humans.

Again though, I'm a layman here too, so someone please correct me where I'm wrong :)

Project idea: The death cap mushroom is really tasty according to people who are now dead. Disable the poison.

Maybe the poison is what makes it tasty?

you would do that chemically, after collecting the mushroom/while cooking it, because that's easy.

not for the death cap

The poison is a bicyclic polypeptide that shuts down ribosomes in your liver. Destroying or removing the poison is as difficult as destroying or removing all the protein. (like dealing with mad cow prions) It's just not going to happen unless you like your mushrooms charred pure black all the way through. Well, that or dusty white ash.

Just half a mushroom will kill an adult human. This isn't something to take chances with.

Right, my protocol would have been: fully homogenize the mushroom, treat it with an antibody known to disable all the toxins, then cook and eat. I wasn't really proposing anybody does this.

> CRISPR requires some manual steps in the lab to make it work well. You can design the protein and guide RNA and spacers until you're blue in the face, but if you can't find a specific enough cutting site on the genome, or one that is close enough to your target, or something like that, you'll wind up with a lack of specificity in that cut, which can lead to unwanted mutations.

I am not a biologist, but nevertheless I found out something interesting today regarding the problem of designing good guide RNA via machine learning.

The work that I learnt about today addresses 2 key questions via machine learning approaches: off-target effects of CRISPR (http://biorxiv.org/content/early/2016/10/05/078253) and optimization of the guide RNA sequence to maximize on-target activity and minimize off target effects (http://www.nature.com/nbt/journal/v34/n2/full/nbt.3437.html).

I learnt of this stuff from one of the authors who works at MSR New England (http://nicolofusi.com/).

Also of possible interest is their project page https://www.microsoft.com/en-us/research/project/crispr/, where they are planning to release a cloud-based service for end-to-end guide design by incorporating the approaches developed in the above work.

My sister, despite not even having complete her masters yet, is the leading CRISPR-Cas9 person in my country, and was sent to MIT to improve her technique.

You are completely correct, in fact one thing that was notable about it was how much work intensive it is, she was forced to miss lots of family events because she had to go in the lab work on the cells, it requires lots and lots of manual labor.

That said, the stuff IS awesome, after you figure how to get it right, it can get right a lot. (but she did this with animals, where she CAN attempt trial and error until she figures the right reference gene sequence to use to guide the proteins, I dunno how that would work when you are racing against time to save a live patient for example, specially if you want to edit his heart or some other area that is hard to access).

By the way, one of the motivations of MIT new variations of CRISPR (that don't use Cas9) is that Cas9 technique require big needles if you want to work in a living organism, the new ones would allow you to use much smaller needles, that are more practical to stab into a patient... and it is very forward thinking, because people didn't even tried stabbing CRISPR needles into someone yet, it is just that MIT guys realized it would be impractical with Cas9 and decided to invent another technique before actually trying it.

Startups working on this (with credible founders, i.e. inventors of CRISPR) include Editas Medicine, Caribous Biosciences, and CRISPRx. Look out for Doudna, Qi, Charpentier, Elledge, and Church. Also see J. Listgarten for computation approaches to improving CRISPR/CAS9.

This isn't ready for in vivo use yet due to off-target effects (CAS9 is a very permissive protein that often binds to the wrong part of the DNA).

In the US, Carl June's group at U Penn is working on editing adoptive, engineered CAR T cells for cancer therapy, going into clinical trials. They're probably bummed to have been scooped (first to do an adoptive CRISPR engineered T cell transfer), though their work is certainly more clinically relevant than Lu's.

The headline seems fine to me. Your complaint seems to be that you think "CRISPR gene-editing" should only refer to editing of cells while they are still inside a live body. But I don't see why that specific type of therapy should claim full ownership of the term "CRISPR gene-editing". Grammatically, any therapy that uses CRISPR modified genes would qualify.

Watch the low-quality comments for an illustration of how to misread that headline as "We edited a human! I want wings and a giant pulsating brain!" /s

The term you're looking for is in vivo.

This study was in vitro (and then in vivo!)

I would agree, though, that most people would have thought "in vivo" from the title alone, since the prospect of and actual practicality of doing it in vivo is what really differentiates CRISPR from the tech that came before it.

ANy time a person says about human medical processes "more sense to just..." it means you don't understand the complexity of the problem.

We've already seen previous methods of attempting to "knock out" or "Repair" genes with gene therapy; it's an entirely non-trivial problem for a large number of different reasons.

Here's a Kurzgesagt video about CRISPR for anyone that's curious: https://www.youtube.com/watch?v=jAhjPd4uNFY

I was reading Oryx and Crake and I was thinking "There's no way this could happen... right?" then Trump said he basically wants to get rid of the FDA and EPA and now there's talk of a biomedical duel in human gene-editing?

Oh boy.

I can't wait to get a rakunk, myself.

With CRISPR, it's even better right; live gene editing. You can be a rakunk, rather than just own one!

On a side note I read that book too & loved it.

>"The researchers removed immune cells from the recipient’s blood and then disabled a gene in them using CRISPR–Cas9, which combines a DNA-cutting enzyme with a molecular guide that can be programmed to tell the enzyme precisely where to cut. The disabled gene codes for the protein PD-1, which normally puts the brakes on a cell’s immune response: cancers take advantage of that function to proliferate.

Lu’s team then cultured the edited cells, increasing their number, and injected them back into the patient, who has metastatic non-small-cell lung cancer. The hope is that, without PD-1, the edited cells will attack and defeat the cancer."

Or maybe it was this:

>"The researchers removed immune cells from the recipient’s blood and then selectively killed most cells containing a certain sequence using CRISPR–Cas9, which combines a DNA-cutting enzyme with a molecular guide that can be programmed to tell the enzyme precisely where to cut. The targeted gene codes for the protein PD-1, which normally puts the brakes on a cell’s immune response: cancers take advantage of that function to proliferate.

Lu’s team then cultured the surviving cells, increasing their number, and injected them back into the patient, who has metastatic non-small-cell lung cancer. The hope is that, without PD-1, the selected-for cell population will attack and defeat the cancer."

Since there is no paper (only press release) we can't say much more about which explanation is most plausible in this case.

Luckily the article linked to the clinical trials.gov record for this trial [1] which suggests that the first approach (editing vs selection) is being used.

1. https://clinicaltrials.gov/ct2/show/NCT02793856?term=crispr&...

That seems to be the favorite interpretation. We need to see what data they present that favors it (ie % survival of the treated cells, number of initial cells, % mutants at that location detected in control cells). Also, this is kind of weird, because they say it is non-randomized:

  Study Design: 	
  Allocation: Non-Randomized
  Endpoint Classification: Safety Study
  Intervention Model: Parallel Assignment
  Masking: Open Label
  Primary Purpose: Treatment

But later it says:

  Progression free survival - PFS [ Time Frame: From date of randomization until 
  the date of first documented progression or date of death from any cause, whichever
  came first, assessed up to average 10 months ] [ Designated as safety issue: No ]

  Overall Survival - OS [ Time Frame: The time from randomization to death from any
  cause, assessed up to 2 years ] [ Designated as safety issue: No ]

So is the treatment randomized or not? The info on that site may not be reliable... maybe it makes sense somehow though.

I would be surprised if this safety trial were randomized with respect to the CRISPR manipulation. I agree with your assessment of needing more data to believe that the gene editing manipulation was successful.

I suspect the randomization relates to the allocation to the experimental arms as described on the reporting page:

    This is a dose-escalation study of ex-vivo knocked-out, 
    expanded, and selected PD-1 knockout-T cells from 
    autologous origin. Patients are assigned to 1 of 3  
    treatment groups to determine the maximal tolerant dose

Nice catch, why do you think a safety trial should not use randomized allocation of the treatment (but it does make sense for dose) though?

It's certainly the former; CRISPR is used to mess with a gene, not kill a population of cells. This isn't what the trial is looking to validate.

>"CRISPR is used to mess with a gene, not kill a population of cells."

Look it up, in every paper where they report on that aspect, the vast majority of the cells die. Also, they detect small numbers of mutants in the control (non-"modified") cells.

I haven't read a ton of CRISPR/CAS papers, but I have read the seminal ones, and didn't see that. Mutations in unmodified cells are completely expected, of course, but not mutations in line with where CRISPR is being targeted—I'd hardly believe that to be the case. It looks like the cell perforation (getting CRISPR/CAS into the cells) can certainly affect cell viability, but those numbers have been steadily, rapidly improving. Quick search shows that in 2015, a study found 80% viability post perforation and CRISPR/CAS application (http://advances.sciencemag.org/content/1/7/e1500454.full).

Looking at that paper, the 80% viability doesn't refer to after CRISPR/Cas9 had time to do the selection, it refers to passage through their chip and puromycin selection (which just indicates the plasmid got in the cell):

"Plasmids encoding Cas9 and sgRNA targeting phosphatase and tensin homolog (Pten) (fig. S6A) were delivered into MCF7 cells, followed by culture for 48 hours and puromycin selection. More than 80% of the cells survived the selection process, indicating the high delivery efficiency of our method.

[...]

Similar to Pten knockout, more than 80% of 53BP1 knockout cells survived the selection process."

You have to look at figure 5B, where the caption reads:

"Cells (5 × 10^4) from (A) were seeded in 60-mm dishes in complete medium and cultured for 7 days. Cells were trypsinized and collected for cell count in a Countess II FL Automated Cell Counter (Life Technologies) daily for 7 days."

So they started with 5 x 10^4 cells but if you look at the first day after being in culture, it is much less than that. I can't tell how low from the chart which shows it as basically indistinguishable from zero. They do not seem to report those values anywhere.

Also, you can look at figure S5 and see that they saw GFP get knocked out in >2% of control cells (Cas9 only). They don't do a similar quantification for the Pten, but if they had, I am pretty sure some low percentage of cells would be shown as already knocked out for Pten.

>"knocked out in >2% of control cells"

That should be 1%.

This makes me feel that i.e. looking for bone marrow donors may become obsolete in some cases, if we know the gene sequence responsible for the disease. Instead of looking for donors, it may actually be possible to repair bone marrow cells and inject them back, fixing the problem. Am I missing something here?

Nope, not missing anything. It should also be possible to use it it to reprogram mature cells (such as skin cells) into homeopoetic stem cells in-mass, and simply introduce them into the appropriate site (eg bone marrow, for your example)

Wait, I thought that CRISPR was a matter of gene editing but that stem versus other types of cells was a matter of gene activation? Wouldn't that require different techniques?

Good! I can't wait for the day I can get a shot to knock out my debilitating 'seasonal' allergies permanently, at the gene level, instead of having to rely on potions and powders of steroids, anti-inflammatories, and other fun chemicals.

> The hope is that, without PD-1, the edited cells will attack and defeat the cancer.

It seems that we already have good drugs which inhibit interaction between PD-1 and PD-L1 is there really enough benefit from this?

The effectiveness of drugs like pembro varies with cancer type and other factors. Here's a recent study where patients exhibited only a 56% response rate: http://www.nejm.org/doi/full/10.1056/NEJMoa1603702

The field is extremely promising and it's worth exploring fully, especially considering that even good responders can suffer relapses due to subclonal evolution.

Those drugs cost $100K+ per year, and have toxicity.

I can't imagine that personalized gene editing therapy will ever be cheaper than the current checkpoint blockade agents.

It will be interesting to see if there are differences in side effect profiles in the PD-1 knockout T-cells vs our current biologicals, however.

I took the original comment to be asking why bother taking out PD-1 from the T-cells. The answer is that it obviates the need for checkpoint inhibitors, which you would otherwise need to administer together with the CAR T-cells for best efficacy. Maybe they meant something different.

Checkpoint blockade alone is a great development, but having treated patients with these drugs, I can tell you it is far from a cure for 80% of people.

Agree with the other comment that CRISPR is unlikely to cost less.

Also, there is toxicity associated with gene modification. Take a look at the patient deaths with CAR-T therapy.

I think the best (or worst?) part is that CRISPR can be used with a 'gene drive' it keeps the changes active on on-going it's not a one shot thing.

GATTACA?

The "gene drive" - the quine of the genetic world.

Can't CRISPR achieve stable transfection through pathways like AAV, though? If that's the case, can't the cell lines carry on the changes without any need for on-going therapy?

Yup. You'd target sex cells instead, but that's only a slight modification.

GATTACA just relied on gene analysis (including among a number of blastocycts) but no gene editing. Though I was very annoyed at that movie for implying that people who get diseases just have insufficient willpower.

> We want to give your child the best possible start. Believe me, we have enough imperfection built in already. Your child doesn't need any more additional burdens. Keep in mind, this child is still you. Simply, the best, of you. You could conceive naturally a thousand times and never get such a result.

I always thought the editing was implied even if it wasn't explicitly called out.

Jerome was different in that he was not edited.

You're right that the text doesn't call it out but I was assuming PGSD[1] which would have the same practical effects and would be less controversial.

[1]https://en.wikipedia.org/wiki/In_vitro_fertilisation#Preimpl...

I have a feeling this won't stop at medical use cases

Most of technology is neither good or evil, and has usages in both civilian and military space. That is how the things are with technological progress. Machine learning may be used to find corelation between gene mutations and cancer, but it may also be used to put missile into unsuspecting enemy.

Roman road system allowed for flow of goods and people within empire on scale never before seen in acient world, but also for their legions to strike empire's enemies.

That's true. But I'm not talking just about military use.

I'm talking about germline gene editing, beyond gene therapy.

It will have consequences that span multiple generations, and are self-replicating. It will completely change the fabric of society. We are going to be changing what we are as life forms, without really understanding life in the first place.

http://www.nature.com/news/don-t-edit-the-human-germ-line-1....

Isn't this the back story to I am legend?

A biotech arms race with the US?

That's laughable after this election.

https://en.wikipedia.org/wiki/CRISPR -> Cures Disease

Eventually...

https://en.wikipedia.org/wiki/CRISPR -> https://en.wikipedia.org/wiki/Gattaca

And...

https://en.wikipedia.org/wiki/CRISPR -> https://en.wikipedia.org/wiki/Gremlins_2:_The_New_Batch

I hope we cure a lot of diseases before we get into the bad side of genmods.

Reply on news.ycombinator.com