0:34
So my company is Bachem.
0:38
Since Peter Grogg was founding the company, he started with some minor acid derivatives and since then we were always focusing on the peptides, mostly synthetic peptides and solid phase peptide synthesis.
0:51
And because of the analogies, we then also went on into preparing oligonucleotides.
0:59
And during this over 50 years, we gained a lot of knowledge, and this also helps to tackle this dramatically increasing market with just one direction for quality, for limits, for guidance, for everything, also for speed.
1:18
And this is also true for the analytic part, of course, where we need to put all our know-how in to cope up with all this.
1:27
And of course, there are also interesting things which drive innovation that we have to look at, for example, into processes like aggregation.
1:38
For me it was completely new field and so I will start with an introduction.
1:45
Even so, Michael already did quite a lot which is relevant for this topic as well and also showed a lot of about the peptides which are the main driver in this field.
1:56
And then later on, I go on to what we are doing analytically to manage the downstream process and then later on what we have as a routine analytic in the end, if we talk about the quality material.
2:15
So peptides, we're always in the in between small molecules and the biologics and there was a huge gap in the guidance.
2:25
So it was filled somehow.
2:28
And nowadays in the last years, they're coming up more and more and they're even more in draft.
2:33
And it showed us that we are on the right way.
2:37
And in this example you see Glucagon and I think everybody knows that this is quite aggregation prone.
2:44
So we're in the heart of this topic.
2:47
Peptides can aggregate.
2:48
So we have to take care of that. And what we have as a peptide, we have the patient safety, we have to ensure as it's a drug substance.
3:02
And if we have aggregation, we can reduce the amount of active compound in the drug product, so reducing the efficiency, but we can also have an immunogenicity issue and then reaching from the depletion in the human body up to anaphylactic shock.
3:22
So there's serious issues with this and we have to get over this by having a good material quality to at least ensure that's not coming from the peptide itself initiating instability in the drug product solution.
3:40
And of course for our own manufacturing purpose, we need to ensure that we have readily filterable solutions that we do not have any G links somewhere in between reaching the final product.
3:58
The peptides made quite an evolution if we are talking about this diabetic market and a lot of improvements were implemented, but mostly they are addressing the efficiency and half life in the human body.
4:13
So under the physiological conditions and our customers of course they are looking for the drug product stabilities, but the physical stability becomes better if you're talking about our production environment, if you add fatty acids or other things to it, then you can have a higher physical instability in the end.
4:43
So looking a bit into the aggregation process, so it's complex, it's an example, there are others.
4:49
This goes into fibrillation reaction and in the beginning, you have the lag face.
4:54
There you have reversible processes going on some of them but non covalently and so you are still fine.
5:04
Then once the growth started to you have monomers adding to it and self-propagation is taking place or you have a strong growth of your aggregate and then again other variables are coming with the second order nucleation, fragmentation, and elongation so that you have quite a complex creation.
5:26
If you think in theoretical mathematical terms that all influences the aggregation pathway.
5:35
And the point you don't want to reach is a nucleation.
5:47
So usually after nucleus or smaller aggregates and you refer to them as seeds and if they are in the material, they cannot only in the process but later on also induce aggregation.
6:00
So without intense stress on the material.
6:06
So therefore it's important to develop the process so that you avoid aggregation and that you get a high quality material.
6:16
Our process, as I said before is mostly solid phase peptide synthesis and we are starting on the left with the reactor, doing the building of the peptide chain, coming to acid cleavage precipitation.
6:31
This is referred to as the upstream process.
6:33
And then we are going into the purification of the peptides, which is then the downstream process.
6:38
So we have column purification, buffer exchange, we do home organisations filtrations and then finally and sometimes in between lyophilisation.
6:49
So quite a lot of steps and the advantage of the downstream is you can purify your material, and you can also purify it from high molecular weight aggregates.
7:00
But of course you have to see more and more to the end that you do not run into aggregation issues or generate seeds while getting to an end.
7:11
So the criticality and the impact of the aggregation becomes higher to the end of our process.
7:18
And the aggregation therefore becomes one part of the puzzle we need to address when we do a downstream process development.
7:27
And for this we have to look at all the different manufacturing steps we have in order to look at those condition and optimise them also towards the physical stability.
7:40
Usually, traditionally mostly the chemical purity was in the focus of the downstream process.
7:46
So addressing the removal of reagents and side products and by holding time studies and so on, looking that the chemical degradation is not taking place.
8:00
So if you look into the instrumentations, we find a lot of different mechanical stress we are applying by steering, pumping, it might foam sometimes, we're freezing the solutions down and of course we have all the material surfaces.
8:18
We also find surfaces like air to solvent surface or the solvent to ice surfaces and they all can trigger aggregation.
8:27
And then we have the solution characteristics, which can be unlikely, but it's also our option to tune those in order to get stable solutions.
8:41
Then maybe it's a question when should we do so?
8:44
So sometimes we have peptides, they do not have any tendency to aggregate.
8:49
So you can be more relaxed.
8:52
But from some basic structure, you already can guess that they will aggregate from what is known in the literature.
8:58
And then it should be done very early as is part of the puzzle.
9:02
The aggregation or let's call it the physical instability of the peptide has to be also addressed during the early process development because otherwise you're redoing it later on if you're going into upscale or further on with the project.
9:22
Yeah, that's just introduction into the environment where we are working this, and now we're coming to the part with the analytics how we tackling it at Bachem.
9:34
So there's a lot of analytics out and they're coming out more and more.
9:39
For me, it's very important that the analytic we are using is meaningful.
9:43
So I need to know my instrument and what they can do and what the limitations are very well, because I want to have finally something be said about my material and not discussions about the method.
10:01
And they don't always need to be complicated to be interesting and good.
10:06
So we are talking about a kinetic process and if I have a visual detection, I maybe have a very high limit, but if I have one seed, I do hard to see it anywhere.
10:20
But if I have some time it will come into my range to observe this also with a visual inspection.
10:27
And then we come to the advantages of this method.
10:30
So it's generic, I can use any buffer system for my visual inspection.
10:37
I can run samples in parallel, different batches, different stress conditions.
10:43
And therefore it's a very versatile method which I can use.
10:51
Of course there are also instrument helping you with this.
10:53
So we don't need to record this all visually.
10:56
But if we can do a multi variable screening, we are quite fast to look through all our environment.
11:06
And this can be either because the instrument has one sample and screening certain conditions throughout the samples, or we use instruments with different chambers that we can apply different conditions to the same sample like with Crystal16 screening instrument or the ARGEN.
11:26
And then also great are plate readers as you can do quite a lot of measurements in parallel.
11:33
So measuring in triplicate positive and negative samples and of course different conditions that you choose before in the pretreatment of the samples, which one I like most is light scattering instrument, because it has a quite nice range.
11:54
I already see my monomers in there and I also see the growing aggregate quite a long time.
12:01
It gives you quality aspects of the material.
12:05
And of course, I said before, you can measure quite a lot in parallel.
12:12
The basic principle is you're looking at the aggregates with a laser and detecting the scattered light by the branch movement the particles are doing.
12:24
I can calculate this back on the hydrodynamic radius.
12:28
And another advantage here, I can see different species in one measurement, so I do not just have any average values.
12:37
I do have multi component system and see the growing aggregates.
12:42
While I still can detect my monomeric unit.
12:47
It's also something which works generic.
12:51
So it's not the fluorescence or anything which disturbs.
12:56
As long as you do not have two huge components in the buffer system, you are very versatile in your samples you can look at. And another advantage is that the radius of the particle is going into the intensity by a power of 6, giving you a good sensitivity on high molecular weight structures.
13:21
And if you go into the instrument, probably with every instrument you find different things you can read out.
13:27
So here of course you can have average values, but you can also look into each single spectrum on different days and see whether we have just a peak broadening or another species is developing.
13:42
Then with this instrument you can do different tests like R2 screening.
13:48
So you have basically different concentration of your drug product and getting out whether they are generally having attractive or repulsive interactions.
13:59
And last but not least with all the data, the instrument already does some statistics.
14:04
So if the auto correlation function, which is required to calculate the radius is not working, we will take this out of the calculation, and you can see how many were taken out.
14:18
So you can judge about whether your results are meaningful or if it's valuable to discuss about it or not.
14:27
Because if you end up in discussions with a lot of values you’ve generated, you lose a lot of time and it's not beneficial to go on.
14:37
So on the other hand, you can use it if you know what you have and know your system very well.
14:42
You can use also these statistics to say, okay, this sample was bad.
14:46
For example, if we observe G links, usually we don't have a good autocorrelation.
14:50
So we see on this data that this that conditions were not good at all.
14:58
And then we come to the design of experiment.
15:02
This brings everything together.
15:04
So you have all the variables you can screen, you have the instruments giving you read out, and then you can look for everything you want yourself.
15:13
And the nice thing about the readout is you don't need to concentrate on one single technique or one readout of a single technique.
15:21
You can make up your own scale which is meaningful to you and combine this together to end up with your visualisation of your robust operating window directing your synthesis or your downstream process in the right direction.
15:42
Then we come to the analytic part.
15:46
So this is a qualitative aspect, how we can look at stability in solution.
15:53
And now we are coming how we look at materials if we want to see is it a good quality, is it a bad quality, will it be a sufficient quality.
16:02
And also here the kinetic for me it's quite important because the seed concentration usually which you're heading for is unknown.
16:16
Of course you can look for high molecular weight species then mostly you're good and fine with the SEC.
16:23
But if you want to know what critical seeds or what amount of critical high molecular weight species you have, you need to push them, to trigger them, and initiate the aggregation process.
16:38
If you do so or if you just wait it can take quite a while until the kinetic starts.
16:48
We also see in triplicate that you can have outliers and as I showed before in the beginning of the lag phase, we use some processes ongoing or reversible, but in statistic events they start to aggregate and the same is true for later on if the aggregation is one started going on heavily.
17:12
There are also different terms in your equation which can bring you a variability.
17:21
So we try to reduce the lag phase to push the aggregation further on to get over the lag phase and ideally then to measure when you're already in the aggregation process.
17:41
Of course, at that point, SEC maybe can be applied because you already started to differentiate with the material.
17:50
It would be versatile as it can handle active solution as well as organic solutions.
17:58
You can validate it.
17:59
It's said to be sensitive and yeah, doing chromatography is one of our major analytic methods, so easy to implement.
18:09
But with LOQ of 0.1%, you’re usually in the range where you look at chemical impurities, but you do not know whether your seeds are at that level.
18:20
Maybe you have less, maybe you can't see it.
18:23
Whether this level you have is based on the critical seeds.
18:27
So there's a lot of uncertainty.
18:30
And if you're at once running aggregation, it's even hard to distinguish your chromatography from a column that went bad because then you get a flat chromatogram, reduced intensity as the aggregates will be in front of the column and slowly eluting over this.
18:53
So we are back at kinetics.
18:54
Here's again an example where we have run 2 batches in the SEC without seeing any significant differences.
19:04
But then going to something as simple as the filtration, it's also done on a chromatography instrument with two pressure sensors, one before and one after the filter.
19:16
And then the solution is eluted over the filter to see an increasing back pressure in the case that the aggregates if they're present starting to clog the filter and then leading to the increase of the pressure.
19:32
But still it's on a chromatography, so we have to run it one by one.
19:36
And if I want to fix as many variables in my system as I as needed, then in the morning I have to prepare all my samples each after 15 minutes to then at lunchtime take them.
19:53
And each 15 minutes I start a chromatography to have everywhere the same time for the pretreatment and get it timely onto this system.
20:04
Of course, this makes nothing I'm heading for if I'm talking about a routine assay.
20:12
It's working, but nicer if you go on to the well plates again and you can run everything in parallel.
20:20
For example in this case thioflavin does a nice job.
20:29
This is well established.
20:31
It's also mentioned from the authorities and the and the documents that this can be applied, and it's also claimed to be possible to validate. And with a good pretreatment you also get rid of the long lag phase.
20:54
Anyway, I think jumping over the lag phase and then do not look too long on the aggregation process gives you the best results as it fixes all variables which you do not want to look at and concentrate more on the initial seed concentration, which is the initiator of this slope.
21:17
And if I talk about fixing variables, I mean, if you set up your system, you have to look at everything what's there.
21:25
So how do I prepare?
21:27
And do I have the same vessels?
21:29
Do they steer all the same?
21:30
It's the temperature everywhere the same?
21:33
But also in theoretical recreation, for example, the monomer concentration should be high enough to have the physical chemistry torque constant.
21:44
And if you fix all these variables, then the increase of the slope should be directly dependent on the seed concentration.
21:56
And this makes you a test finally where you can read out the seed concentration of an unknown material.
22:02
You have something, you know, you run it in parallel and then finally you get out the value what the current material is supposed to be.
22:15
Yeah, this was my little journey through my aggregation environment in peptide manufacturing.
22:23
And each peptide is individual.
22:27
We have peptides with four minor assets.
22:29
They already have physical instabilities.
22:32
If you go break there, for sure we'll get problems.
22:35
So you have to individually look at those, but it's important to look at the downstream process to optimise so that you deplete the high molecular weight impurities and avoid inducing aggregation or bringing new seeds into the material.
22:53
And therefore you need good analytic techniques.
22:56
You must be expert of what you're using.
22:58
I don't think that you need all of them which are available, but you have to look for yourself what gives me meaningful values.
23:07
I like multi screening techniques as I can have very comparable results reducing by this way.
23:15
Also again variability in the systems.
23:19
And with this, I want to end my talk being happy to speak to you and always being supported from a small team, from R&D and QC if we are going to investigate those.
23:35
Thanks.
