0:00
Called variance and you can see here separations typically last.
0:03
This is a 5 kilobase pDNA, typically lasts about 14 minutes.
0:09
Again, it's you're basically mimicking agarose gels.
0:13
It's a gel based separation and electrophoresis conditions.
0:17
We've actually looked at different types and different sizes of plasmas.
0:20
You'll find actually that the plasma size varies depending on what you want to produce.
0:26
Typically for some work I think was presented a couple of years ago by Pfizer and a poster goes up to 12 kb.
0:33
So you can you and what you tend to do then is you look at a change in the gel constitution to deal with the larger plasmids.
0:43
Typically you dilute it with TBE as a way to enable you to use the classical gel with the larger plasmids.
0:50
And we have several manufacturers now are using this type of approach to quality control their plasmids in production CDMOs as well as innovative companies.
1:00
And this real approach is really reproducible and it just shows a sort of a DNA ladder to give you just an idea the reproducibility, the separation you want to get by CE, which makes it ideal technique then to use in the quality control of your building block of your mRNA vaccines and therapeutics.
1:20
So really that sort of where we start and plasmids, you know, companies like Plasmid Factory, Cobra Biologics, which is now I think part of Charles River, these companies have been using CE to assay the plasmids for well over 10-15 years.
1:36
But what's next important is understanding the purity of your, of really your active part of your drug, which is your mRNA itself.
1:44
Now I would love to run intact mRNA into my spec.
1:47
That's simply not possible at the moment because you can't separate it for a start.
1:51
And it's very big.
1:53
So again, you're looking now towards a CE based technique to separate size based impurities and to give you, we'll call it genome integrity to give you an idea of the impurities present in your mRNA part of your drug.
2:12
And again, at Sciex, we've been working with our users in the industry to devise new ways of doing this.
2:20
And because of the demand, you know, over the last couple of years, because of the sort of the generation of the mRNA vaccines to treat COVID, we actually looked at devising a kit specifically now in a gel based kit to look at mRNA.
2:34
And here you can see an example of an mRNA being analysed by from an LNP or lipid nanoparticle.
2:42
From this assay you can pick up both lower molecular weight and higher molecular weight impurities.
2:49
And this sort of gives you sort of an integrity assay for your RNA.
2:53
We actually now put this on a multi capillary system.
2:56
So where historically you would run 11 sample every 15 to 20 minutes, if you look at the runtime now you can run 8 samples in that same amount of time.
3:07
So it's now starting to be used even in process development in this particular respect to assay their production of mRNA vaccines.
3:22
And again, how we do this is when you're dealing with LNPs or lipid nanoparticles, you have to of course get released the mRNA first of all.
3:30
So we treat it with surfactants like Triton X and extract it and heat it.
3:34
So when you generate your LNP or lipid nanoparticle, you have to release, of course your mRNA.
3:38
And the simplest way to do it is to disrupt the lipids.
3:40
And we do that by adding surfactants.
3:42
So it's quite a simple step to release the RNA and then we just profile the RNA as shown here using gel electrophoresis again.
3:52
Now if you want to get an accurate sizing of the RNA, what's really important is you have to use a similar type of reference standard.
4:01
So we found that if you want to do sizing of mRNA, then use an mRNA reference standard, OK.
4:07
If you want to do single stranded DNA, which is of course in AAV biotherapeutics, then you have to use a single stranded DNA reference standard because that gives you a far more accurate sizing of your product.
4:21
OK.
4:22
If you don't, if you use a different, the wrong type of standard, when you do this, you'll get a lot, a bit of a mass shift or a sort of a size difference, which is just because you're using the wrong reference standard.
4:37
And then this gives an example of how we use this.
4:39
And this, I think this work was done in collaboration with a company in Norway called SINTEF.
4:44
And so Jeremy helped us produce some of these results.
4:47
And these were based on MC3 LNPs, MC3 is basically the ionizable lipid, which is used to build a nanoparticle.
4:55
And you can see that the corrected peak area, the main product is very reproducible.
5:01
And we're getting, you know, RSDs of low percentage share of reproducibility for the corrected peak area and also migration time.
5:10
So again, it's a very reproducible technique to assay your active part of your therapeutic.
5:17
And again, just looking at here, you can see the detection of minor products and main products and the error expected on the size.
5:25
So we're only seeing a couple of sort of basis difference in size from what we're seeing in CE compared to the literature value.
5:36
And again, we're using this case an RNA ladder to calculate the sizing of the mRNA product just to give you an idea of the linearity.
5:45
This is important to actually quantitate your impurities and this sort of assay.
5:50
In this case we are using laser induced fluorescence as a detection technique.
5:54
So we would use things like SYBR Green or SYBR Gold as a labelling dye.
5:58
Here you can get down to two to three orders of linearity and ULO ultra low limited detection of 1.6 x 10 to the -3 mix per mil of the RNA in solution.
6:12
OK.
6:16
So that just looks at basically the whole R messenger RNA.
6:21
Look at size based impurities.
6:23
If you want to start to go down to more in depth characterisation, then you have to think about digesting similar to proteomics.
6:32
You digest your RNA, you cap it to stop digesting all of it.
6:36
What you have to do here in when you're handling messenger RNA to digester, to get down to look at the microheterogeneity is typically you sort of cap the part you want to keep.
6:46
You digest away the others using RNAs and then you release it and look at what's left.
6:52
And you can do this by the three prime end or the five prime end.
6:55
And this is when you start using mass spectrometry.
6:58
And this is a workflow now, which has been out for a couple of years.
7:02
So let's just show you an example of five prime 5 prime capping process.
7:07
This is important when you manufacture your mRNA so you don't cap it properly, then the mRNA will degrade.
7:12
All right.
7:13
So capping really protects the mRNA from degradation.
7:16
So it's an important part in the process of generating your drug effectively your active part of your drug.
7:22
So the capping process is really a way you protect the RNA from digestion.
7:28
Say you just do this sort of N methyl transferases using enzymes and you go from a G cap to an O cap and then eventually add the five prime ends.
7:38
OK.
7:39
So this is a chemical process.
7:41
And what you want to be able to do is understand the different levels of the different parts of the process in your final product to understand how effective the capping process has happened.
7:49
And you can do this basically using TOF MS Here we're dealing quite large chunks of RNA, so it could be 40 or 50 base pairs.
7:59
So we tend to use TOF MS in negative mode because of course you're dealing with acidic monitors.
8:06
And you can then do the reconstruction of the data and look to see if you've got, you can see the bottom here, the G cap and the cap presence on the species you pulled out of the RNA.
8:20
And again, you could with this sort of size of RNA now because we've chopped it up, we're no longer dealing with 1000 base pairs.
8:27
We're dealing with less than 100 base pairs.
8:29
Now you can actually run it through an LC.
8:32
So it's effectively converts it to an oligonucleotide, and you do an oligonucleotide analysis.
8:37
Now it's been in the industry by mass spec for a fair number of years.
8:40
And you can do this either using some of the new hillock based methods or use in ion pair reagents like HFBA and stuff like that to help separate your oligonucleotides.
8:50
And you'll also see here from the on the far left hand side, you see the uncapped and capped actually are separated chromatographically as well, which helps aid identification of these two products in your sample.
9:05
And using MS, you can dig down and you can say identify the uncapped end capped.
9:10
And actually in this case, the sodium, potassium and adducts you see will help with identification.
9:17
So it's in this case, it's quite useful to have these adducts present because it helps with the ID and the sort of the byproducts and the sort of mis-capping of your RNA.
9:29
So from that point now as thus gives you an idea of how to dig down to sort of the micro heterogeneity of the RNA, I want to now look at the delivery vehicles themselves.
9:40
Now, as with COVID vaccine, there's two possible ways of delivering mRNA into a patient.
9:45
One is using viral vectors and the other one's using lipid nanoparticles.
9:50
Of course, I've been involved in a in biopharmaceutical and pharmaceutical analysis now for over 25 years.
9:55
And as you get, as they become more and more modern drugs, it becomes more and more challenging.
9:59
So no longer do you have to just look at the payload, you have to look at the delivery vehicle as well.
10:04
OK.
10:04
So in this case we would, we're now looking at the two different types of delivery vehicle.
10:10
And the first one, this is more of a CGT or cell and gene therapy rather than an adenovirus used in COVID.
10:15
It's an AAV type viral vector as a delivery vehicle.
10:20
People like using AAV because it's be very targeted.
10:24
The serotype will help you target the cell line you want the cells you want to go to.
10:29
And it's relatively safe compared to other virus types.
10:32
So that's why people tend to use it.
10:34
And it's, I know a lot of the AAV current AAV cell gene therapies in production are things to treat disease like eye disease, etcetera.
10:44
And but again, you can use two different techniques in this case.
10:46
Again, if you look at more intact particles and intact proteins which build up the particles, then again we would tend to use CE in this respects.
10:55
And what you're doing here is profiling the intact proteins, OK.
11:00
And again, you have to disrupt the particle.
11:02
So you just reduce the particle, you add in DDT and you heat it to release the viral proteins and then you profile, you generate a profile of the viral proteins.
11:11
In AAV, typically the VP3 dominates and you want in a 10:1:1 ratio.
11:17
When the ratio charges shift, then the properties, the particle starts to change and it's, it becomes less useful to deliver its payload into the target cells in the body.
11:26
So this is actually an important assay for the quality control AAV particles.
11:31
And again, looking at this, you can see several different serotypes which you can put for the CE system and it can be applied to say we've applied to several different serotypes again, because now you're looking at low concentrations of your active viral particles.
11:47
Again, we use lift based detection techniques.
11:50
OK, So we actually label the proteins in this case with a chromium based dye, which a Pyrylium based dye to increase our sensitivity.
11:58
And this really jumps your sensitivity up by sometimes up to 500-fold, which allows you then to start thinking about applying this type of technique to process development, not just to QC of your final products.
12:11
And again, this just shows you the reproducibility you can get for a for an assay.
12:14
So we've labelled the chromium dye the P5O3 and you can see that again, the assay is really reproducible and you can put this on the multi capillary system as well.
12:22
So you can again, eight times your throughput running on that system, which allows you then to send to move it to more of a process control type assay.
12:30
And again, the linearity is very good and the sensitivity we can get down to is E to the 9 gene copies per mil.
12:38
That's the way you measure viral vector concentration.
12:44
But again, if you want to go down to the, I see micro heterogeneity, right down to the structural changes of the proteins, which can also affect the ability of the viral particle to deliver its payload.
12:54
Now you shift again to mass spec.
12:56
And similar to the RNA we would use, as mentioned in the previous talk, you would use peptide mapping as a way to pull out some structural chains of the proteins.
13:05
Things like deamidation, charge variance, which will affect the structure of the viral particle and the sort of classical workflows you see here.
13:13
I would say I've been used by academics for 20-30 years, right?
13:17
It's the classical reduce denature reduce alkylate and tryptic digest your proteins present in your viral particle sample and just simply peptide map it or you can even do intact analysis.
13:30
OK and do sort of like top down type sequencing approach.
13:36
Again, we've used this to look at and there's several papers that to look at some variance present on these viral particles.
13:44
Actually there's a lot of variance around the cap at the end point.
13:47
So we get C terminal acetylation, which is commonly seen in these.
13:50
So that's quite easy to pick up by peptide mapping.
13:53
And again, these actual viral particles, viral proteins are very similar in structures.
13:59
So mass spec really helps to dig down any changes and you can see that we can even use intact mass for this case.
14:08
This is used just some visuals taken from a biologist explorer software, which we're now using to process some of this data that's based on gene data software package.
14:18
But it helps us to visually show the data quite well and highlight presence of this case, presence of say variants of the viral proteins in the sample.
14:30
But in using this even at an intact level, as I mentioned, you can buy C terminal acetylation is quite common.
14:36
And also because now you're dealing not with the monoclonal antibodies, you're dealing with more biologically produced proteins.
14:41
So now you get phosphorylations as also a modification which happens on these proteins.
14:46
And again, you can pick up phosphorylation and variants on the intacts’ size.
14:52
And these days using things like EAT or Electro activity dissociation, you can actually now start to think about sequencing this.
14:59
We haven't done this.
15:00
We're planning to do it.
15:01
It's so now you because that's only was released a couple of months ago, about a month ago.
15:05
We're now planning to put these through and do some sort of top down sequencing of some of these proteins to see if we can pull this, the actual position of these modifications out from the intact mass.
15:14
Just me removing one step in the sample prep.
15:19
But if you can't do it from an intact level, then the way to do it classically as I mentioned is using bottom up approach or tryptic digestion.
15:28
And here you can see sort of the classical map you get when you do tryptic digestion, your coverage really with common.
15:35
Again, we use a Q TOF here we use our Zenotof.
15:38
It's really quite good.
15:39
So you'd end up getting 95% coverage of your sequence and it allows you also to using alternative fragmentation techniques to pull up positions of variants.
15:51
And this is includes, to be honest, isobaric variants like I mentioned in the previous talk, Asp, isoAsp can be pulled out using EAD because that will actually give you C and Z fragments.
16:02
So it allows you to give diagnostic ions for isoAsp and Asp if that variant occurs in the sequence of the viral proteins.
16:11
And again, you can see some of the data we pulled out from the peptide map.
16:15
A good one here is the position of the phosphate on the serine.
16:22
And again, using some of the alternate fragmentation techniques that allows you to accurate position where this modification happens on the protein.
16:30
So that's looking at how to look at the AAV delivery vector.
16:34
Yeah, I've got time.
16:35
The last one really is looking at LNPs.
16:39
I was at a conference about a year ago.
16:42
It was in Switzerland where I was listening to Melissa Moore, who's the chief scientific and officer Moderna and she mentioned that there was a problem, there was a paper in Nature about a problem with the ionizable lipids where they could have an impurity which deactivated your RNA.
16:57
Last thing you want, right?
16:59
Ionizable lipid are very important integral part of lipid nanoparticle.
17:02
It helps to stabilise the particle, and it sort of surrounds the RNA.
17:07
And you can imagine if that ingredient of the LNP is deactivating the mRNA, it's really a problem.
17:14
OK.
17:15
And so we actually work with Precision NanoSystems to develop an assay to put pick up impurities of LNPs.
17:22
NC 3 is a common ionizable lipid.
17:25
You see from an ionizable lipid, what the issue here is that it has a tertiary amine which is quite chemically active, right?
17:32
And that could be a point of oxidation from an oxide.
17:35
And that's the bit which deactivates the mRNA, OK, that can bind onto mRNA covalently.
17:40
And that means your mRNA then it's deactivated.
17:43
Just one point of binding stops the message being transferred and it stops the RNA molecule working.
17:48
So it's really important to understand if that's present, you also have this say carboxylic acid and also unsaturation on the alkyl chains as well.
18:00
These are all points of potential modification of the ionizable lipid.
18:05
And when we took this sample from the manufacturer, they knew there was impurities present in it and we could pick them up.
18:12
So we picked up three different impurity peaks.
18:14
Again, this is using LC/MS as an assay technique and we want to understand it was one of these the N oxide and well, straight away you can pick up probably an oxidation by an addition of 16.
18:28
So it's quite easy to pick up an oxidation from the MC 3 just by looking at the parent mass, right.
18:33
But the issue really is more about, you know, where is that oxidation on the molecule.
18:41
One of the benefits you'll see is that we tend to use two different fragmentations next just to dig into the details and EAD or electroactive dissociation which is in our ZenoTOF system actually gives us more information down the alkyl region.
18:54
So we would we've used that technique now from majority of the analysis.
18:59
So again, this is one point of oxidation we found by looking at that alkyl region, we found that you get an epoxide ring formation and basically a change in this in the saturation or the unsaturation of one of those double bonds it oxidises.
19:16
So that was one point of oxidation.
19:21
OK.
19:22
And we only saw that the actual, the head group hadn't changed because you got this diagnostic fragment 132 and the where we saw the oxidation, you can see these diagnostic fragments coming around about 550 odd, 570 odd that helped us position the oxidation, of the alkyl chain.
19:41
And you can see these ions being present in the EAD spectrum.
19:48
So where did the N-oxidation happen?
19:51
Well, we then look to see if we could find the diagnostic ions, the N-oxidation if the N-oxidation happened, of course, you add oxygen now to the N group, the N to the sort of the nitrogen at the end, the tertiary amine.
20:04
And this will again change the mass from 132 to 148, the addition of 16.
20:10
And we actually saw that in one of the impurities.
20:12
So we localised this position and we saw the fragment times of 61, 148.
20:18
So it then highlighted that the oxidation had happened at the tertiary amine.
20:24
So in this particular case, it did have a percentage of that impurity present.
20:29
But what about there's another peak we found towards the end?
20:32
What was that?
20:33
So and what we found actually by looking at the third impurity was actually that was a as a reduction of sort of a change in the double bond nature.
20:43
So you lost one of the double bonds, OK, so you got reduced.
20:47
So in this case, that was a change in the alkyl chain length.
20:51
And by using EAD, we can actually position which double bond had changed.
20:56
This is one of the beauties of EAD and actually allows you to position double bonds, the point of unsaturation of the alkyl chain.
21:03
And we can tell, we could tell from the Spectra that basically this bond here, I don't have a pointer, but this point here had changed.
21:11
So that had been reduced, and we lost that double bond.
21:16
So going back to look at the sort of the relative impurities we could pick up using mass spec with a ZenoTOF system, we got down to picking up really low abundant impurities down to .1% in the LNP particle.
21:28
The PPM error was as expected around about one PPM or better error.
21:33
So we could really confirm the identity of the impurities.
21:37
And this is again where you could use mass spec to assay the delivery vehicles.
21:40
This is a really good example where we're now applying mass spec to pick up impurities which could adversely affect your drug in LNP particles.
21:50
At that point then I think I'm out of time.
21:52
So I'd like to hope like your attention today and if you have any questions, I'll be around for the rest of the conference.
21:57
So please find me.
21:58
Thank you very much.
