0:00
So it's Ben Wilkes and Kevin Jackson from Waters.
0:06
So Ben will cover the liquid chromatography and mass spectrometry, right.
0:14
And the Kevin will talk about spectroscopy, terminal analysis and chromatography.
0:24
Thank you for joining our double act.
0:26
Hopefully it'll be more Morecambe and Wise than the Chuckle Brothers.
0:30
So I'd like to talk about 3 aspects here and Kevin will talk about the light scattering aspects.
0:40
So in terms of oligonucleotide and RNA analysis, things I'd like to cover are RNA sequencing, CQA monitoring and formulation efficiency, particularly regarding LNPs.
0:54
So firstly, RNA sequencing Waters have been working for quite some time in the RNA sequencing space and the first thing I'd like to present is round digestion and enzymatic digestion for mapping here.
1:11
So commonly the digestion enzyme of choice is RNase T1.
1:18
For oligonucleotides, the problem that we often see with RNase T1, and RNase 4 is unspecific digestion fragments coming out at the end.
1:30
So we should see short overlapping digestion fragments that can be isobaric and they can just be the same fragments.
1:38
So Waters have developed 2 new digestion enzymes here.
1:41
Firstly MC1 and Cusativin.
1:45
These have different specificities.
1:47
MC1 has five prime uridine specificity and Cusativin 3 prime cytidine.
1:54
What we enable with these two enzymes is more larger fragments, so we have less overlapping fragments.
2:03
We have larger fragments that are easier to be uniquely identifiable by LC-MS.
2:13
We've combined this with our hardware and software technologies, particularly the LC-MS systems, the BioAccord and the Xevo G3 to develop a workflow for RNA mapping here end to end with the digestion enzyme LC-MS analysis of the digestive fragments.
2:32
In parallel, we have software to generate in silico digestion products that we can feed into the analysis software to review against the digestion fragments that we see experimentally.
2:48
This can be done in MAP Sequence, which is a specifically designed software.
2:52
There's not a redesigned peptide mapping software.
2:55
It's specifically been generated for oligonucleotide mapping and sequencing.
3:01
We can then feed this data into some other software to generate our sequence coverage.
3:08
And with this workflow, particularly these new enzymes, we've seen some really excellent coverage, up to 100% sequence coverage for 100mers and above.
3:16
So we're getting some really good data out of this and this is a bit of a busy slide, but just to show you some digestion fragments that we see from RNase T1 and RNase 4.
3:30
These are some of the short overlapping or non-overlapping sequences that we see.
3:34
And using MC1 and Cusativin, we generate a lot more fragments, bigger fragments and unique fragments, enabling confident identification of your RNA sequence.
3:49
The next topic that I'd like to fly through is CQA mapping, particularly 5 prime cap and 3 prime Poly(A) using a similar type of digestion strategy.
4:00
Here we can digest our sample and even capture in case of five prime cap to look at what our CQAs are using LC-MS technology and our software intact mass to acquire, process and report the software.
4:22
So to look at 5 prime capping, if we have our desired structure on the right here, we can have a number of impurities along the way to generate our five prime cap.
4:32
And what we're trying to look for is modification on a single nucleotide, which at the intact level, if you're looking at thousands of base pairs sequence, then it's quite challenging to see.
4:48
So using the digestion strategies, the enzymes, the chemistries and the hardware and software that Waters have generated, we can digest the product, separate it chromatographically and analyse using mass spec to generate.
5:03
What we see here is this large peak as our intended 5 prime cap and the uncapped precursors down here.
5:09
And we can identify these by LC-MS, the software when you input your sequence and modifications, we can generate a simple tabulated view of the percentage purity of your intended 5 prime cap product.
5:24
So we can see very quickly with automatic processing, we've got 97.3% intended 5 prime cap here with 2.7% impurities.
5:41
And to look at the Poly(A) tail as well, so if we do a digestion with MC1 or Cusativin, we generate a map sequence like this.
5:49
This large peak at the end is the Poly(A) tail and looking at this with mass spec we see a very heterogeneous signal here.
5:58
Interpretation of that can be quite difficult but using the intact mass software and its deconvolution algorithms, we can deconvolute this into a really sensible and pun intended easily digestible result here where we can see the mass difference of 329 Daltons, which is the repeating adenosine units.
6:21
There is an application note on this and if you come to our booth, I do have the link to that if we don't scan it now.
6:30
And last but not least from the LC-MS side is lipid nanoparticle characterization, raw material characterization and monitoring as well.
6:41
So we have two workflows for LNP characterization monitoring.
6:45
Firstly, an impurity identification workflow.
6:49
We can acquire LC-MS MSE data, apply our discovery tools to identify these unknown impurities and feed this report and the impurities we identify back into a monitoring and quantification method.
7:03
Here we're using the same type of approach.
7:07
We can do a targeted screening and consultation of the LNP product as well as the raw materials and impurities from that and generate report.
7:17
This is all under GMP software as well validatable.
7:22
So from a single injection we can do impurity identification and monitoring and quantitation of targeted products.
7:35
I would like to pass over to Kevin to talk about the light scattering solutions we have now.
7:47
Thank you for that.
7:47
I'm about to throw a load of ideas at you.
7:50
All the stuff we do will not fit into a 10 minute talk, so please bear with me if there isn't anything that rings a bell with you.
7:58
Please come and see us on the booth.
7:59
It's #35 I think it is, anyway.
8:03
We've jumped straight in.
8:04
This is one of our normal chromatograms.
8:07
There's three different materials here.
8:08
What we have is the line that's going across the non-chromatogram is the molecular weight calculated by multi-angle light scattering on our chromatogram.
8:19
What you have here is the thyroglobulin, which is just a reference protein and you can see the molecular weight of that from 10 million down to what about a million.
8:28
And when it eludes, you can see it eludes in the middle of where my RNAs mRNAs elute.
8:34
So they elute at different times.
8:36
The light scattering doesn't care when it eludes, it still gives you the molar mass at that point.
8:41
So you know, you don't have to calibrate this with masses.
8:44
It's an absolute calibration.
8:46
So you're actually getting the right result out from your answers.
8:50
That's the very basic bit we do with light scattering.
8:54
On the other side, we can also measure the hydrodynamic radius on online at the same time.
8:58
And that shows actually the hydrodynamic radius of those materials are at the points where they elute fairly similar.
9:05
Therefore that's why they're eluding at that point.
9:12
For gene therapies, we kind of started in this with AAVs and we have a technique which uses the multi angle light scattering to give three critical quality attributes.
9:25
4 AAVs which can be done in an automated way, an online with an autosampler, et cetera, in about 20 to 30 minutes.
9:48
This equation up here shows how we're doing the calculation.
9:50
So the amount of scattered lights proportion to the mass times the concentration and this which is a number we put in related to your samples.
9:57
The important bit here is this concentration because we can use different concentration sources to see different things within your molecule.
10:07
That means for, it means with the light scattering detector, we can do particle concentration.
10:15
So we get our chromatogram and each point in the chromatogram we can measure a particle concentration out.
10:21
If you put different amounts in, we can still go down to a pretty low limit.
10:27
So we need around about 2 to 5 times 10 to the 10 particles per mil injected into the system.
10:32
It's not necessarily what comes out.
10:37
The empty full ratio is the bit I was talking about the different detectors.
10:40
So if you have a full AAV, it's got the DNA inside and it's got the capsid on the outside.
10:46
By using two different detectors, either differential refractive index and UV or two different UV wavelengths, we can see different parts of that construct.
10:54
We can then separate that out and use the light scattering to work out the molecular weight of each component within the molecule.
11:01
And just in case anybody's in the wrong thing, you're doing antibodies.
11:04
We can also do it with antibodies for the drug, antibody ratio, carbohydrate contents, glycosylation, all this other stuff.
11:10
But for AAVs again.
11:12
So here we have, we can work out the total molecular weight, we can work out the molecular weight of the capsid, and we can work out the molecular weight of what's loaded inside.
11:21
Of course, if it's empty, the DNA becomes very low molecular weight.
11:25
There's none in there, but the capsid remains the same molecular weight.
11:29
You can also see on this slide that where we have a slight overlap, there's a little tiny amount of dimer in here.
11:36
We can see that coming into the system.
11:39
We can also measure this in both the dimer and the monomer content.
11:44
So if you do have a significant amount of oligomerization within the molecule, it may be preferentially taking your load into that molecule.
11:52
And that's what's that can be something you can assess in your production.
11:58
Of course, being light scattering, we can also do aggregation.
12:01
So you can see on here we've got again the molar mass and you can see aggregate down here, two different aggregates down here and then molecular weights associated.
12:09
And we get the particle concentration on every slice within the chromatogram, whether it's empty or full.
12:15
And that can all be measured.
12:19
Just in case you didn't believe me, there's lots of publications out there now using the technique.
12:23
This is one of them.
12:26
As a side you'll get a copy of the slide.
12:28
If you want to look it up later, please do.
12:31
It's just too busy to go into.
12:32
But it does compare very well to other techniques.
12:36
So a summary of what we're doing with this.
12:38
It's a SEC-MALS, so size exclusion chromatography followed by multi angle light scattering.
12:44
It's in the native state, so we're not changing it.
12:46
You can run it in your formulation buffer or whatever buffer you feel is suitable.
12:50
We can get things like the molecular weight.
12:52
We can get the Rh if you put a DLS detector on the side of it as well and it's going to be automated.
12:59
It's also done under compliance.
13:00
Of course, the limitations, there's always limitations to the thing.
13:06
We're not resolving the difference between empty and full.
13:09
They are overlapping each other because they didn't.
13:11
They're basically the same size.
13:13
So anything which is a partial capsid looks like not fully loaded.
13:18
So that is the one limitation we have on the system.
13:21
Our sensitivity is down to between 2 and 5 times 10 to the 10 particles per mil loaded onto the column.
13:28
And the one assumption we're making in here is that we're assume what you put onto the column comes out because we want to give you a particles per mil.
13:35
We'll measure the number of particles in what came out.
13:38
If not all of it came out, your chromatography might not be perfect, then that is one assumption made in there.
13:47
We can also apply exactly the same thing to LNPs.
13:50
So even larger constructs, even more complicated constructs, we can get all of these different answers out.
13:58
Again, I'm going to be very quick in here, but under SEC you have some interaction effects.
14:05
You also have obviously the possibility of sheer degradation or taking the any ligaments in there and sharing them apart.
14:14
LNPs are basically unstable materials that we can deform and change, so very difficult to analyse.
14:23
If you want to do a separation, we have another technique called FFF fill flow fractionation, which will allow you to much more gently separate these materials and there's much less problems with these interactions.
14:35
That's really all I can talk to in that time.
14:37
It's a longer technique, so I really have race through this.
14:42
This is good.
14:43
Last of all of these things I'm talking about, we can also do most of these in line.
14:49
So we can actually put them inside your production system and give real time feedback into the process control.
14:56
So if you want to know what's going on, and that was the question I was going to ask the previous speaker, have you got any in line control?
15:02
But you did mention VATs, that was nice.
15:05
We can do this sort of thing in there.
15:07
There's also outline controls which we can be done in sort of 10 to 15 minutes.
15:12
But the inline control is really where we're heading a lot with a lot of these techniques.
15:16
Anything with a fluid system we can take a line out of, we can get that sort of information.
15:21
We can give you real time feedback.
15:22
So if something starts to go wrong, we can stop your process before you ruin your batch.
15:26
It will never replace QC.
15:28
You'll still have to do all your release testing because that's compliance, but it will mean your batches will be more reliable and less likely to fail afterwards.
15:37
I think that was the last bit.
15:39
So I've really gone through that quickly.
15:43
Thank you very much.
