Curious about zink finger nuclesases / gene therapy

[Mega]

Hi,

I read about zink finger nucleases, and that they may be used in living organisms to alter their genome (after birth, as an adult) . 

I was wondering how that was posssible, how do you get the enzyme in every cell of the body? Is it just injected in the blood stream and then gets distributed automatically? 

Or do you inject it into the organ needed? But how can it penetrate the cell wall? Is it so small that it fits through the celll membranes? 

[Cathal Garvey]

I studied Zinc Finger Nucleases long enough to know that they are a dead
end. Seriously, six-feet-under dead end. They're already being replaced
by TALENs because TALENs aren't as Patent-encumbered.

Essentially, about half of the known zinc finger families are patented
to the hilt, and the patent troll that owns them charges in the 10,000s
for licenses. The remaining zinc fingers aren't sufficient for reliable
assembly; it turns out that, despite their glowing reputation for
"modular DNA targeting proteins", they are actually very hard to get
working correctly. You can add a zinc finger for "TAG" to a zinc finger
for "GAC" and have it totally fail to bind to TAGGAC, and instead bind
to TTAGWGAC, or something like that.

I haven't studied TALENs enough to know for sure yet, but I gather they
are somewhat more predictable. Being less patented is ALWAYS a boon, and
it seems that it's paying off.

As to how they work; when using them in-vitro, you can use them as
proteins are normally used; protein + buffer + DNA etc.
When used for gene therapy, it's a totally different situation; you
don't deliver the protein, you deliver a DNA agent that encodes the
protein. So, you design a plasmid containing your ZFN/TALEN, plus the
DNA you want to replace the chromosomal target with (it must have
significant homology to the chromosomal target to encourage crossover
after enzyme-cleavage), and you deliver that through electroporation, or
chemical treatment, or viruses.

The TALEN/ZFN gets transcribed/translated and, if you've got the
appropriate nuclear targeting peptides attached, gets sent back into the
nucleus, where it cuts the target site. Then, homology-directed DNA
repair (homologous recombination) leads to the target site getting
replaced with your alternative sequence in the plasmid at some efficiency.

If your replacement is designed not to contain the target site, this is
one-way and reasonably high efficiency, provided that the DNA-cleaving
TALEN/ZFN works as intended and gets into the nucleus. You have to
provide lots of plasmid to ensure there's plenty of template lying
around when the break occurs in the target strand.

Efficiency per targeted cell is good, but don't expect to transform
whole organisms, it will not happen. Efficiencies of transduction in
animals is very low for naked DNA, often less than 10%. When using
viruses, you can get higher efficiencies, but at the cost of high
specificity (only certain cell types get transformed) and potential
immune overreactions. Methods for non-naked-DNA, non-viral DNA delivery
include electroporation, high pressure delivery, and liposomal delivery.

Options for Viral delivery which aren't insanely risky to attempt on
humans include: Adeno-Associated-Virus (NOT Adenovirus). And that's it.
The others are highly risky in terms of immune response, and if you use
them as they are most often used, they also carry significant cancer
risks due to random-ish, gene-preference integration. AAV is remarkable
as it has such low immune stimulation effects, and it has a target site
in the human genome that seems to have very low risks of any cancerous
side-effects (provided it's unoccupied by wild AAVs).. but it has really
low capacity for DNA, so it's hard to deliver anything useful.

Naked DNA has low efficiency of transformation, but at least it's A)
Easy and B) More controllable; integration is only likely where you
direct it to occur, and immune reaction to naked DNA is generally mild,
particularly if you produce it via PCR or DAM/DCM negative strains of
E.coli, so it's not methylated in ways that the immune system treats as
suspicious.

--
www.indiebiotech.com
twitter.com/onetruecathal
joindiaspora.com/u/cathalgarvey
PGP Public Key: http://bit.ly/CathalGKey

[Andreas Sturm]

Thanks, that makes it more complicated :D

Well, what I was thinking that you just deliver the protein and it will  work. (Just like doing a restriction digest in vivo)
That sounds even more difficult, when you have to get plasmids into cells some way.

If you want a gene knock-out, that will be easier to archieve than HR, I think?

say  we want to delete those 2 base-pairs, we just need two ZFN (TALENs?),
ATAGATAGATAGAG ||| AG ||| AGTAGTAGAT   The AG is cut out, and it's unlikely (?)  that it is reinserted, is it?

--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

[Cathal Garvey]

The size issue you raise is pretty significant, I'll grant you. When
there was any chance of me working with interesting things like
ZFNs/TALEs, it was with the possibility of using Adeno-Associated Virus,
and you can easily fit ZFNs into an AAV vector, but probably not a TALE.

So, when delivery size is a constraint, ZFNs still have a significant
advantage. A zinc finger binding 3/4 nucleotides can be as little as
20-30 amino acids long, barely longer than the loose consensus of a
"peptide". TALEs seem, on cursory inspection, to be significantly larger.

But, when size isn't an issue (cost of synthesis falling, vector insert
size not a constraint), I think the effort and wo/man-hours saved by not
"optimising" makes TALEs far more attractive.

But then, I'm lazy. ;)

On 27/09/12 10:23, jhwangc wrote:
> Although a lot of the field is moving to working on TALEN's zinc fingers
> haven't yet been completely replaced.
>
> Although zinc finger proteins do have context-dependent effects like you
> mentioned (you can't string together any two zinc fingers into an array)
> there are currently two different assembly methods for zinc finger arrays
> which work around this (OPEN and CoDA) by building arrays using zinc
> fingers that are known to work well in combination together or optimizing
> the zinc finger array at every step of construction. These processes are
> pretty long and cumbersome to just stringing together TALE's (TALENs are
> the the TALE protein + a nuclease domain) though which don't have any known
> context dependent effects.
>
> BUT the size of the TALE protein arrays is a concern for a lot of
> researchers because for a zinc finger array, each individual zinc finger
> binds 3 base pairs of DNA but for a TALE array, each TALE only binds one
> base pair. As a result, a zinc finger array only needs 6 zinc finger
> proteins to bind to an 18 base pair sequence while a TALE array would
> need18 proteins to bind to an 18 bp sequence.
>
> The field is definitely moving toward TALE's but zinc fingers will still be
> around for at least a little longer until more research on TALE's is
> done/the kinks get ironed out.

[Anselm Levskaya]

Speaking as someone who used to do genome editing -

ZFNs were always simply too hard to evolve/select into a working
reagent for targeting a new DNA string. Only dedicated labs were
working on them, and even then progress was slow. TALEs already seem
to be easy enough for a determined individual in a lab to build a new
reagent with. Plus, there has only been minimal work optimizing this
new family.

Although size is an issue in AAV viral vectors, not being to get a
protein that targets your string of choice is a much bigger issue!
There just aren't that many sites you can target with ZFNs.

Besides, one of the major application of these proteins is editing the
germline (via NHEJ deletions or HR swaps), in which case the size of
the genetic payload doesn't matter much (since microinjection is often
used here). The first successful demonstration ever of targeted HR
editing of the zebrafish happened very recently using TALENs.

-a

[Mega [Andreas Stuermer]]

While thinking about thene therapy once again, one question arised:

Usually, the Adeno-Associated visus integrates into a specific locus in chromosome 19. But the integration site does not seem to be very sequence specific.


Now, what I wonder: Once you had a AAV-gene therapy, you cannot have a second one? Or will the second therapy using the same vector integrate randomly? Or some other place in chromosome 19?


I know that's not diy bio, but a lot of smart people here :D

[Koeng]

How about using CRISPR for genome engineering? CRISPR can actually be used to interfere with a genome instead of breaking it (catalytically inactivated), so hypothetically if you could get it into immune cells you could prevent the mRNA production of a provirus such as HIV/AIDS...

-Koeng

[Josh Tycko]

AAV integrates at a relatively low efficiency compared to other viral vectors, and the majority of your gene expression comes from episomal non-integrated DNA. This is generally considered to be an advantage (lower chance of integration in oncogenes) but does mean that the delivered gene will not be replicated and expressed in daughter cells if your target cell is dividing.

You're right, in the majority of cases, AAV cannot be administered twice but it's actually an issue of immune response and not integration. After the first administration, the body generates AAV capsid-specific antibodies which will neutralize the second administration before the gene is effectively delivered. But, there are different serotypes (capsids) of AAV, so you could administer AAV8 and then still be able to administer AAV9 the next time.

AAV naturally have a packing capacity of 4.7kb, which is potentially enough for a small promoter, TALE, and some nucleases. George church had a paper this summer involving AAV delivery of TALE fusions. AAV may not be so good for CRISPR though, but I guess we'll see what Editas Medicine comes up with.

Also you should definitely check out the work from Carl June delivering zinc finger nucleases to T-cells (ex vivo) to remove the CCR5 receptor, to create an immune system invisible to HIV. They're looking at using RNA electroporation instead of viral vectors because the short term transient expression of the nuclease is completely sufficient for genome engineering, whereas AAV is great when you need stable expression.

Lots of fun stuff to think about when choosing a vector!

And since it's my first post I'll add - I am a huuuge fan of this listserv.

[Cathal Garvey]

Well, not sure what this has to do with Zinc Finger Nucleases, aside
from the fact that some AAV therapies have embedded ZFNs as their "payload".

As to AAV itself, yes; you'd expect that a cell pre-infected with an AAV
at its target locus would be immune to re-"infection" with the same
locus-targeting AAV. That's not to say the payload mightn't have a
chance to get expressed though, it doesn't have to be integrated for
that to happen... it just means it'll be ephemeral and won't outlast a
cell division or two.

This mightn't be the big issue it seems at first. Your payload can do a
lot of things in one move without requiring integration; it could help
trigger homologous recombination to repair one faulty gene with its
partner allele, in the case of a heterozygote. It could trigger outright
deletion of a gene, if you can manage that with a payload as small as
AAVs will carry. It could kill the cell altogether, which sidesteps a
need for integration or, er, caution.

One of the challenges of AAVs though is the small size of the payload
they can carry compared to other viruses. This is more than made up for
by the fact that AAV gets a "free pass" from the immune system, of
course. But it may mean you simply can't do certain things with AAV.

It's worth looking into the integrative mechanism behind AAV: I'd be
surprised if there hasn't been work altering its target-site specificity.

And, as Koeng suggests, CRISPR could be used to provide an alternate
spcificity.. if memory serves, CRISPR is a pretty small system? So it
may fit on an AAV like a ZFN, with a little room for additional payload.

[Ravasz]

For my PhD I've been designing plasmids which replicate like a chromosome and can be maintained extrachromosomally (even replicated and delivered to the daughter cells upon proliferation) in the human cell. This completely avoids any integration issues, but it has only been tested in cell cultures.

Of course this method does not allow replacing genes, only adding new ones, but still has a lot of potential IMO.

[Cathal Garvey (Phone)]

If the plasmid encodes a replacement and a CRISPR/TALEN it could certainly replace endogenous genes, but you'd have to watch put for copy number then as you could end up with plasmid encoded plus two chromosomal! Also, prolonged expression of DNA modifying enzymes would, I feel, be a cancer risk as they never have 100% specificity..

--
Sent from my Android device with K-9 Mail. Please excuse my brevity.

[Mega [Andreas Stuermer]]

For my PhD I've been designing plasmids which replicate like a chromosome and can be maintained extrachromosomally (even replicated and delivered to the daughter cells upon proliferation) in the human cell. This completely avoids any integration issues, but it has only been tested in cell cultures.

That sounds awesome. Just make sure the artificial chromosomes won't get into germ cells, else the enhanced human isn't compatible to "old" / "original" humans anymore and interbreed won't be possible I assume.
 

Of course this method does not allow replacing genes, only adding new ones, but still has a lot of potential IMO.

RNA silencing? Then you can knock out the gene. And add the same gene with another codon usage / or another gene from yeast with the same function. 

As to AAV itself, yes; you'd expect that a cell pre-infected with an AAV
at its target locus would be immune to re-"infection" with the same
locus-targeting AAV.

But as it is not really seqence dependend, there may be a chance it integrates close to the  iother integrated virus DNA?

In case it were sequence specific you could include the targeted sequence in the virus, so it integrates a new integration site :D 

[Cathal Garvey (Phone)]

Aside re: compatibility at germ line: not a concern. At least, not a fertility concern. Chromosomal abnormalities are bad because of gene dosage: too many or too few *important* genes due to too many/too few chromosomes. If you had a million copies of a large chromosome that encoded nothing important, the worst you'd suffer would be phosphate deficiency (because of all that wasted DNA).

So a plasmid encoding a "plugin" wouldn't stop you breeding with others lacking the plasmid, but it's an open question as to how inheritance would work: probably depends on the plasmid's segregation system and how it behaves during meiosis.

One area where compatibility *would* be an issue is two people with different 'plugins' on the same plasmid backbone. Interference would cause plasmid loss of one or the other eventually, but how soon after fertilisation? My guess is somewhere in blastocyst stage, with plasmid assortment being effectively random if the systems are identical (copy number, stability etc), so the baby would be chimeric for both traits.

Early enough, and you'd get organ-level chimerism. So your kidneys are green fluorescent, your skin red. Late enough and it'd be intra-organ tissue-level or cell-lineage level. Total chimerism.

Because you're dealing with a random system, likely you'd get a spectrum of chimerism: some completely homogenous organs, some totally blended organs.

What effect this would have depends on the chimeric genes. For curing cystic fibrosis you only need to "cure" a critical percentage of cells to clear the lungs of excess ions, but for something like a myopathy or motor neuron disease you probably need to cure as many as possible: every lost cell is permanent damage.

Speculation is fun!