Latest / The BioHub - by Avetix / Episode 68 – Stefan Lutz – Codexis

Transcript

Episode 68 – Stefan Lutz – Codexis · The BioHub - by Avetix

0:05
Welcome back to another episode of the Bio Hub.
0:08
Today I'm joined by Doctor Stefan Lutz, the SVP of Research
0:13
at Codexis, a leader in enzyme engineering and the force behind
0:17
some of the most exciting advancements in RNA
0:19
manufacturing. Stefan's journey spans academia
0:22
and industry, from chairing the Chemistry department at Emory
0:26
University to now driving innovation at Codexis.
0:30
He and his team are helping solve some of biotech's biggest
0:32
challenges, from improving RNA synthesis to using AI to unlock
0:37
new frontiers in protein protein engineering.
0:41
This one's all about the evolution of RNA, the role of AI
0:45
in discovery, and what it takes to lead with a growth mindset in
0:49
modern science. Pleasure to have you, Stefan.
0:54
Thanks Max for the introduction. Great to be here.
0:56
Now it's a pleasure really is. Let's just start with your
1:00
story, Stefan. So first of all, what brought
1:04
you from academia into biotech? And then in particular, what
1:08
drew you towards Codexis? Yeah.
1:12
Well, I think probably the, the, the my favorite quote that
1:18
captures the motivation behind that transition from academia to
1:22
to industry is, is division without execution is
1:26
hallucination. Now, don't get me wrong, I love
1:30
the creative freedom, the opportunities to pursue ideas
1:34
that that academia had offered. But I had also seen too often
1:40
that these ideas and, and those creative freedoms ended up with
1:44
a simple publication or patent application, but no more.
1:49
So I really joined Cadaxes because of because I wanted to
1:52
see more impact that science can have to apply scientific
1:57
innovation to solve real world problems.
2:01
Yeah, and sometimes you get what you ask for.
2:04
Just a few months after I joined, one of our engineered
2:07
enzymes was needed to manufacture PEX Lowit during
2:11
COVID-19. So in six months we had to go
2:14
from what was on a test in a test tube on the bench to
2:18
producing 10s of metric tons of enzyme, helping to support these
2:22
lifestyle, the production of these life saving medicines.
2:26
It was one of those moments where you kind of realize, you
2:28
know, where things come together, where signs, purpose
2:32
and impact kind of align. That experiment didn't just
2:36
validate my decision to leave academia, it really reminded me
2:40
of why I became a scientist in the 1st place.
2:45
Now you might wonder, is this an isolated incident?
2:48
Is this something that you can do every now and then?
2:52
But it turns out two years ago, Codaxis made a strategic
2:57
decision to turn the organization's expertise and
3:00
experience towards addressing a major unmet need in
3:04
manufacturing modified RNA oligonuclear ties, which are the
3:08
foundation of RNA interference or RNA I therapeutics.
3:12
Now in that short period of time, we designed, built and are
3:17
now scaling an enzyme catalyzed manufacturing solution for these
3:21
oligonuclear type therapeutics. And in the process we've
3:25
revolutionized the field introducing A simpler, cleaner
3:28
and more sustainable process. So what my transition to
3:33
industry has brought is the opportunity to so to deliver
3:38
real solutions with real impacts that benefit real people.
3:44
That's amazing. How did you find that switch
3:47
from academia into biotech? You know, it was, it was, it's
3:54
definitely a different mindset. You know, here the economic
3:58
realities are are front and centre.
4:03
It is not just about having great ideas, creative ideas,
4:07
it's also translating them into economically viable product,
4:12
which is definitely a brings its own challenges.
4:18
Yeah, I can imagine talking about RNA as well.
4:23
RNA therapeutics have what exploded in recent years?
4:28
How have you seen the field evolve and what's driving in
4:31
demand for for better RNA? Yeah, you're, you're certainly
4:36
right there. I both firmly believe that RNA
4:39
oligonucleotides as a therapeutic entity are a super
4:45
exciting and emerging modality. But I also believe that today
4:51
what we're seeing today is only the tip of the iceberg.
4:55
You know over the last decade we've seen roughly two dozen
4:58
approved oligonuclear type therapeutics hitting the market,
5:03
but there are several 100 new assets in preclinical and the
5:06
clinical pipeline, many targeting large indications with
5:11
millions of patients globally benefiting from these drugs that
5:15
are working their way through the clinic.
5:17
You know it, there's actually estimates that put the demand
5:22
for oligonucleotide product at 30 metric tons in as soon as the
5:27
early 2000 thirties. Now you know, what is it about
5:32
these RNA oligonucleotides, what makes them such tremendously
5:36
promising therapeutic modalities?
5:39
It is really their ability to very effectively address
5:43
previously underserved and undruggable diseases.
5:47
You know, they allow us to modulate the production of
5:50
specific proteins by targeting for degradation the
5:54
corresponding message or RNA. Now a good example of of the
5:59
promise and the future potential of these drugs, in my opinion,
6:04
is an asset to treat neurodegenerative diseases such
6:08
as Alzheimer's. You know, clinical studies have
6:11
shown that these RNA oligo or RNA I therapeutics can intercept
6:18
the messenger RNA, which encodes for the plaque forming Tau
6:22
protein in your brain. And so by preventing the protein
6:26
production, we can prevent aggregation and thereby prevent
6:30
the neural deterioration in the 1st place.
6:33
So tremendous promise. So where are we going next?
6:37
Maybe worthwhile highlighting 2 directions here.
6:41
The first one is to enhance the potency of the existing
6:44
therapeutics and then also to explore new modes of action
6:49
beyond what just mRNA degradation that I've mentioned.
6:55
On the former, you know, how do we increase the potency of these
6:58
therapeutics? There is a lot of ongoing
7:01
studies that enhance our understanding of the mechanism
7:05
of these drugs. So from the moment they are
7:08
administered to a patient finding their way journey
7:12
through the into the cell and to their messenger RNA targets,
7:18
there's a lot of biology and understanding each step of their
7:22
journey in more detail is empowering the design of more
7:26
potent drugs. That is, for example, simply
7:30
getting the drug where it's most needed to the specific tissue
7:35
kind of including address tags that allow for efficient
7:39
delivery. Then there is also the the
7:43
opportunity to tweak to optimize the specificity of these drugs
7:48
once they reach their target by introducing new modified
7:51
nucleotide building blocks and eliminating off target effect.
7:57
Now beyond the targeting of mRNA, as I already indicated, a
8:01
whole new world of biology has opened up where we are learning
8:06
around micro RNA, opportunities of micro RNA and its broader
8:11
regulatory functions in the cell.
8:13
There's also interesting ideas of utilizing these
8:16
oligonucleotides as shuttles to pull cellular entities into
8:22
lysosomal degradation. So that there are many new
8:25
directions that the field is only now discovering that I
8:30
believe will in the foreseeable future have the potential to
8:34
translate into promising new drugs for patients to treat
8:40
terrible diseases. Yeah.
8:44
I mean, I think what you guys are doing here like amazing, but
8:49
you're right, like RNA as a whole in the industry is it's
8:53
crazy how it's blown up in the developments you've seen.
8:56
If we if we look at Codex in particular, what what are the
8:59
biggest hurdles for you and how are you also working to overcome
9:04
them? Yeah.
9:07
Well, you know, I think the single biggest hurdle is what's
9:12
the point of developing to discovering all these exciting
9:17
potential therapeutic assets, these RNA oligonucleotides, if
9:22
they can't reach the patient because we can't manufacture
9:26
them at scale. Now there's a real technical
9:30
challenge. I've mentioned, you know, the
9:32
the 30 metric tons of oligonucleotides that are needed
9:36
in the next 5 to 10 years. The existing chemical process is
9:43
is a working solution, but in order to make even one metric
9:48
ton of oligonucleotide with this today's technology, you need to
9:53
run 100 to 200 batches to make enough product.
9:59
There is real limitations in the scalability of the chemical
10:03
approach. Also the harsh chemical
10:06
conditions that the chemistry represents impacts negatively
10:11
the purity of these oligonucleotide products.
10:15
In addition, this entire process runs in organic solvents,
10:20
actually vast quantities of organic solvents, flammable
10:23
toxic reagents. And so there's a real concern
10:28
over the sustainability of this process.
10:31
Last but not least, because of the the circumstances,
10:36
manufacturing with the chemical method requires a major
10:40
infrastructure investment and a dedicated production facility.
10:45
All these points can be addressed with our new, I
10:50
believe game changing enzyme powered manufacturing solution.
10:54
We call it the Ecosythesis platform and it makes the
10:58
process scalable. It is a process that runs in
11:01
water, and because of the latter, the milder reaction
11:05
conditions also offer a benefit to the purity of the Oracle
11:09
product. On top of all that, because the
11:13
process is much more simple, it also eliminates the highly
11:18
specialized equipment, which offers significant savings to a
11:21
manufacturer, to a producer and the infrastructure investment
11:25
that is needed. So how do we?
11:29
How do we? What's behind the ecosynthesis
11:32
platform? Probably a good question to
11:36
start is to ask yourself who is the world's foremost nucleic
11:40
acid chemist. I would argue it is nature.
11:45
You know, our cells utilize enzymes in the most elegant way
11:49
to repeatedly and reliably synthesize nucleic acid.
11:54
Now the question we have to ask ourselves is, is there a way to
11:59
harness that capability? Turns out we're not the first
12:03
people that asked that question. In fact, this idea actually
12:09
predates the chemical methods that I've just referred to.
12:12
Earlier. Research back in the 1950s
12:16
actually explored the use of enzymes to synthesize
12:20
oligonucleotides, DNA and RNA. But it was truly an idea ahead
12:25
of its time. The native enzymes that these
12:28
researchers had access to simply could not perform under what is
12:33
industrially relevant production conditions.
12:37
Now here we are 7080 years later.
12:42
Today we have enzyme engineering tools such as could access as
12:46
code of all the technology to effectively and efficiently turn
12:50
enzymes into high performance bio catalysts.
12:53
We still start with enzymes in nature, but we're putting them
12:57
through this multi dimensional optimization process.
13:00
We're training the enzyme to work with the substrates that we
13:05
need to incorporate into these therapeutic agents.
13:08
We make sure the enzyme can work at high substrate loadings.
13:12
We make the enzyme versatile so it can synthesize many different
13:16
oligonoptate types with high efficiency.
13:20
And last but not least, we build a stable enzyme that can be
13:24
manufactured at scale and that actually is robustly performing
13:28
under process conditions. We then take these enzymes,
13:32
incorporate them into the process or into the processes,
13:36
optimizing each component and the conditions so they work hand
13:40
in hand to create a bio catalytic toolbox that allows us
13:45
to go from individual building blocks from the nucleosides to a
13:48
complete oligonucleotide therapeutic asset.
13:54
It is with that approach that we can effectively address today's
13:58
hurdles, the quality, the scalability and the
14:01
sustainability concerns that I've mentioned earlier.
14:06
Enzymes are catalysts. We only need them in small
14:09
amounts, so that allows us to fundamentally change how we
14:14
operate the process, keeping the oligonucleotide in solution
14:18
instead of immobilizing it on a solid matrix, which makes a
14:22
significant impact or benefit to the scalability of the process.
14:27
Given that nucleotides and the enzymes all work in water, we
14:32
can all but eliminate organic solvent meat.
14:36
These aqueous conditions also turn out to really help with the
14:40
quality of the product. The cleanliness, the the
14:43
impurities are significantly reduced as we've been able to
14:47
demonstrate in our studies and that translates into higher
14:51
product yield. It reduces the need for
14:54
purification, which further benefits yields and is also a
14:58
significant cost driver. Now, just as a bonus, these
15:02
enzymatic approaches also offer really fundamentally new
15:05
capabilities. So for example, it offers a
15:09
really elegant solution for handling the chorality of
15:12
oligonucleotides. You know, the fact that these
15:14
drug molecules can be left or right-handed, which makes a
15:19
significant difference in biology, but is something that
15:22
chemistry cannot effectively address.
15:26
Our enzymatic approach offers a straightforward, logical and
15:29
scalable solution for that. So where are we with this new
15:33
technology? You know, are we ready to
15:35
manufacture 30 metric tons today?
15:38
Of course, the answer is not now or not yet.
15:43
At the same time, we have gone beyond a science project.
15:47
Earlier this year, we opened a pilot plant, what we call the
15:50
Eco Innovation Lab, which allows us to synthesize up to 100 gram
15:55
scale of oligonucleotide. This is an important milestone
16:01
because it generates the process and the product data that are
16:05
critically important to demonstrate the robustness and
16:09
the reproducibility of our manufacturing solution.
16:13
It also forms the foundation to allow us to build productive
16:18
predictive models for larger scale production.
16:23
It's amazing. It's amazing.
16:25
You're talking about like scaling up in like the potential
16:30
ways of getting there. Everyone's also talking about
16:35
what AI machine learning, which I suppose could be a really
16:38
vital part in those kind of problems.
16:41
And just to get your broad take about AI and machine learning
16:45
and potentially how you would use those tools at Codex is?
16:49
Yeah. Thanks for raising raising that
16:52
topic. There's really two things I want
16:55
to say about AI machine learning.
16:59
I actually personally prefer the term predictive analytics sounds
17:02
a little less blosterous. Predictive analytics is
17:07
tremendously powerful and I firmly believe that it offers
17:11
invaluable guidance towards faster and more a more efficient
17:17
or more accurate delivery of products and services.
17:21
However, I think we need to also control our enthusiasm a little
17:27
bit and remember that predictive analytics is not a magic wand.
17:32
Today, hype regularly outpaces the reality and that actually
17:37
hurts the scientific endeavour. We need facts, not fantasies.
17:43
Let me, maybe I'll take this apart a little bit more.
17:46
Machine learning is a proven, highly effective tool to
17:50
accelerate the engineering of complex systems, including
17:53
enzymes. I do believe Codaxis can speak
17:57
with some authority on the subject because back in 2007 we
18:02
actually published one of the very early tools in the field of
18:06
enzyme engineering applying what is multi component analysis to
18:10
deconvolute the structure activity relationships in
18:13
proteins. You know, the success of that
18:17
tool enabled us to really guide scientists at the bench to
18:23
accelerate to the successful engineering of enzymes.
18:28
Now the technology has vastly improved.
18:31
You know, we have new automation, we have new
18:33
miniaturization, we have advances in the architecture of
18:36
these algorithms. But what hasn't changed is the
18:41
fundamental that you gain, well, you you get what you train on.
18:47
We need quality data and ideally as much as possible of it.
18:53
Now I want you to hold that Paul for a second quality and
18:57
quantity. Well, can we have both the
19:03
quantity of data in enzyme engineering TPA, Lot of people
19:08
revert to public databases where scientists in academia and
19:14
industry report the biochemical data for enzymes often operating
19:20
with natural substrates on the variable conditions and often
19:25
times with incomplete details on the reaction conditions, which
19:29
matter a great deal if one wants to play deploy predictive
19:33
analytics. Now we have the quantity but we
19:37
are lacking the quality. On the flip side, we have a few
19:43
publications that really deal with deploying these enzymes
19:48
under industrially relevant conditions, but those studies
19:52
are few and far in between. If one also remembers that these
19:59
algorithms, that these predictive analytics are
20:02
generally very effective in exploring but not exploiting.
20:09
They are actually effective in interpolating but not in
20:13
extrapolating. It really highlights the the the
20:16
conclusion that. With today's data sets, we do
20:21
not have the information, the quality and the quantity of
20:26
information that allows us to build global predictive
20:30
frameworks that a lot of people are pursuing and advertising
20:36
instead. From my perspective, machine
20:39
learning is there. It has a future, but it needs to
20:42
be deployed where it makes sense, where it has impact.
20:47
Maybe it's a smaller, it's a more more defined problem.
20:53
That's where the technology can shine.
20:56
But in my opinion, we should shy away from these global broad
21:01
brush stroke statements that portray machine learning as as
21:06
the solution to all our problems.
21:10
So what you're saying is there's a time and place for machine
21:14
learning, and people are ultimately trying to fit squares
21:20
through circles to make sure that, you know, like machine
21:23
learning could work in certain scenarios.
21:25
And really, it's not just the answer to everything.
21:28
Absolutely, I think yeah, use it where it makes sense.
21:33
Don't use it as a magic wand. OK, yeah, get you.
21:38
That was that was really nice to hear.
21:40
To be fair, complete change of topic.
21:43
Now you've mentored a whole lot of scientists over the years.
21:49
What What advice do you have for young scientists and engineers
21:52
looking to grow their careers in biotech?
21:56
Well, first and foremost, Congrats.
21:58
I think you've made a great career choice.
22:02
I think biotech offers tremendous growth potential and
22:08
and lots of opportunities to have real impact across many
22:13
disciplines, you know, be it healthcare, be it agriculture,
22:17
be it industrial production, you know, think biofuel,
22:20
biomaterials. There are many people out there
22:24
talking about the future. I believe if you work in
22:28
biotech, you can help build it. Think of biotech as a relatively
22:34
young field. So entering you need to be
22:39
prepared to deal with a lot of volatility.
22:42
There's a lot of change, but also with that change, the
22:47
opportunity to express your creativity and really be part of
22:53
of innovation. It's an extremely fast moving
22:57
and highly interdisciplinary field.
22:59
So be prepared and I hopefully you enjoy this continuous
23:04
learning, not just in your field of expertise, but also in
23:07
adjacent areas. As a scientist, you work with
23:11
business, with regulatory, with process scientists.
23:15
And that team effort, that cross disciplinary thinking is really
23:20
a way of of building impactful products.
23:25
Last but not least, as I've just hinted, biotech is a team sport.
23:30
I think you need to practice, you need to strengthen
23:33
communication but also your listening skills because it is
23:37
essential to work effectively. I love that.
23:42
I mean, that's awesome. I mean, there's a lot of young
23:46
people in biotech coming into the field in a what is difficult
23:49
market. So that you know, that kind of
23:53
advice goes a long, long way at the moment.
23:57
Yeah, changes. Changes is definitely a dominant
24:03
element working in this industry, again with its
24:07
uncertainties but also with the opportunities it brings.
24:11
Yeah. And then like my next question
24:14
is about growth mindset, which has become a a massive theme in
24:19
today's market. What, when you talk about growth
24:23
mindset, what does it mean to you as a leader in science and
24:26
and how do you foster that at Codexis?
24:33
I think growth mindset to me it's, it's really captured in
24:36
that that little try, fail, learn and repeat.
24:41
I think that that nicely concisely captures my thinking
24:45
there. Innovation depends on on that,
24:51
on that logic, on that cycle. You know, innovation requires
24:57
risk taking. Now you want to take calculated
25:00
risk to test your ideas, but the culture within an organization
25:06
that embraces the growth mindset that allows for that risk taking
25:10
is essential. I think part that our society is
25:15
struggle to this day is to embrace failure as an option.
25:21
I think we cannot emphasize it enough that a failed experiment
25:27
can teach you as much as a successful one, you know?
25:30
The trick is, of course, that you ask when you set up your
25:34
experiments, when you pursue your ideas, that you ask the
25:37
most relevant question so that if you fail, you fail quickly
25:42
and not going down too many rabbit holes.
25:46
Last but not least, there's an element of learning, integrating
25:51
the insights from your your successes and your failures.
25:57
Maybe folk, you know, noodling a little bit more on the the
26:02
learning part operating or having a a position in science
26:07
is a commitment for lifelong learning, to keep an open mind,
26:11
to seek new knowledge to to think outside the box.
26:17
It's, it's kind of hard to get precise numbers, but I've
26:21
actually heard estimates that the years that it takes for half
26:26
of what you've learnt to be proven inaccurate or obsolete
26:32
can be counted on one hand in fast moving fields such as
26:36
biotech. So what you've learned 5-10
26:39
years ago, a lot of it is today actually overhauled.
26:44
And so the need to stay on top to, to constantly educate, to
26:51
open your mind to new ideas is absolutely critical.
26:56
The good news is, of course, you don't have to do this alone.
27:00
In working in an academic setting, working in an
27:03
industrial environment, you have your colleagues, your peers with
27:06
whom you can, through communication, through idea
27:09
sharing, drive forward to, to stay on top of your field.
27:14
And in my experience, it's never been too difficult to find
27:18
people with different opinions in the sciences.
27:22
So it's it's always a great idea to discuss your own view and
27:27
kind of compare your assumptions against other perspectives,
27:32
which helps you to hopefully solidify to refine your
27:36
perspective. Having said all that, I think
27:41
it's important to value and to respect experience, but at the
27:46
same time, it should not be taken as the truth and nothing
27:52
but the truth. So I actually do remember a
27:55
story when I was a young assistant professor where I
27:58
thought I had a pretty cool idea of how one could engineer
28:03
proteins in a novel way. I've discussed this idea with
28:07
one of my highly regarded senior colleagues.
28:11
He's also a personal friend and he listened patiently and then
28:17
told me this is never going to work, why waste your time?
28:22
I believed in that idea. I went ahead, executed it, and
28:27
guess what? I think in retrospect, I can say
28:30
it made a huge difference. In fact, I firmly believe it
28:33
contributed to my successful tenure in my academic career.
28:38
So the take away message is believe in yourself.
28:41
Believe in your own creative mind and you know being a little
28:45
rebel every now and then is not a bad thing.
28:49
Yeah, I completely agree with that.
28:50
I think what you said earlier really hit home as well, because
28:54
in my field we often talk about growth mindset and the
28:58
importance of it. You know, leave your ego at the
29:01
door, have a constant willingness to leave, to learn
29:04
all the time. But you brought up a massive
29:08
point, which is in biotech, like it's even more critical, right?
29:11
Like you say, things are constantly evolving.
29:14
It's not even a want to just have that growth mindset.
29:18
It's an absolute need to have it, which, yeah, which is so, so
29:24
true. Yeah.
29:27
And I wanted to move on to the final question as well, Stefan,
29:30
which is a question we ask every guest that comes in the podcast.
29:34
So if you could have dinner with any scientist or innovator, past
29:39
or present, who would it be and why?
29:43
Oh a great question. Honestly, many people come to
29:52
mind, but if I could choose one kind of conversation that has
29:59
always inspired me that that has always been special to me is the
30:06
opportunity to sit down with young, with the future
30:10
generation of of innovators, students, postdoctoral fellows.
30:16
So I find them keen and often times less biased observers that
30:22
that exhibit a great deal of unfiltered curiosity.
30:26
And you know, I think there is something to be said about
30:30
diversity of thought, a fresh perspective.
30:34
So I would probably pick having dinner with a group of students
30:38
and post docs, talk about the signs, talk about their ideas,
30:42
share some of my experience, not to not to discourage them to
30:46
pursue their own, but simply to engage in a intellectually
30:51
stimulating conversation. Amazing.
30:54
That's such a good note to finish it on.
30:56
Thank you so much for coming on Stefan.
30:58
It was an absolute pleasure. Yeah, thank you for having me.
31:02
And please everyone like and subscribe to the Bio Hub and and
31:05
thanks again, Stefan.