Modulating Multispecific Manufacturability With AP Biosciences’ JJ You, Ph.D.

In this episode of “Better Biopharma,” host Tyler Menichiello is joined by Jhong-Jhe You, Ph.D., VP of antibody discovery at AP Biosciences. You explains how molecular complexity and aspects of antibody design can influence a product’s manufacturability and developability.

Keep up with everything biomanufacturing at Bioprocess Online: https://www.bioprocessonline.com/

Follow Tyler Menichiello on LinkedIn: https://www.linkedin.com/in/tmenichiello/

Subscribe to our newsletters: https://www.bioprocessonline.com/user/edit/subscriptions

Tell us how we can improve. Take our 60-second survey: https://docs.google.com/forms/d/e/1FAIpQLSfZGJluCmEhm1MBQvcm7_gtBEXuTx5v42LJ1mx9bCUKo8fwfg/viewform?usp=header

Transcript

SPEAKER_01 0:16

Hello and welcome back to Better Biopharma, the official podcast of Bioprocess Online. I'm your host, Tyler Manichello, and on this episode, I'm joined by Dr. Zhang Jaio, AP Bioscience's Vice President of Antibody Discovery, to talk about the ways in which antibody design can affect product developability and manufacturability. Dr. Yo, thank you so much for joining me today.

SPEAKER_00 0:36

Yeah. Hi, everyone. Great to be here, Taylor. And thank you and the Bioprocess Online teams for the invitation. I would like to share our some experience at the AP Bio about the job development or manufacturing problem.

SPEAKER_01 0:53

Thank you so much. I think a great start would be to give us a bit of background about yourself and AP Biosciences and specifically uh AP Biosciences pipeline of multi-specifics, bi-specifics, and multi-specific antibodies.

SPEAKER_00 1:06

Yeah, sure. Uh actually, my okay, you can everyone can call me my short name. It's about JJs, and my scientific roots are actually in neurobiology. But after my PhD, I met a pretty big people into the antibody engineering during my postal training. And about two years later, I joined AB Bio as a research scientist. And actually, I was the company's first employee. And since then, I have grown alongside the company that it has expanded. And today my role has uh broadened significantly in addition to antibody discovery, functional catalyzation, and validation. I'm also responsible for overseeing the GMP manufacturing through the CDMO partnership for the clinical stage material. On top of that, I manage the key preclinical development activity, including uh toxicology studies and other IND enabling work. Overall, my roles are now spend an entire range from the early functional validation through to the clinical entry. That's uh ensuring that each program is well integrated and executed, ready and aligned for the development success. This is a whole about me. And for the AP Bio, uh actually, AP Bio was founded around 2013 and is based, just like Terry mentioned, uh, is based at Taipei Bioinnovation Park in Taiwan. We are a research-driven biotech company, and we don't operate a large-scale manufacturing ourselves. Instead, we focus on the drug design and play phone development. And our core focus is on developing novel uh antibody therapeutics, mainly in uh eye disease and uh immune oncology, and more recently, we have also expanded to the autoimmune disease. And around uh after the 30 years of development, we are uh expected to list on Taiwan's OTC market on June 3rd. So by the time this is aired, we may already be a public company in Taiwan, actually. So our business model is quite flexible because the drug development is a long and capital-intensive process. Our current business model is uh centered on our licensing and the co-development partnership. This approach helps reduce the risk for investors while also accelerating the progression of our pipeline. And uh today we have a complete two preclinical outlicensing deals, and we continue to advance several internal biospecific antibody programs into the clinical stage. So, this is a background about our AB bioscience. And so far, our current pipeline is entirely built around multiple function and bi-specific antibody program, starting with our most advanced assay, like IBI302, which is a bifunctional antibody designs for the well-m D, also called edge-related muscular degeneration and diabetes muscular edema. Shown and is a DME. This all licensed program is being developed by our partner, Innovat Biologists now. In the FAST 3 data released this March, is uh exciting a primary efficacy endpoint was successfully meet, and we expect regulatory filing activity in China to follow shortly. And another one's program is APE505, which we have a license to test the biopharma in China, is currently being evaluated in two ongoing phase two clinical trial, including one in chorectrical cancer. This is representing our solid tumor portfolio. In addition to this partner essay, we are also advancing the three interim programs based on a novel T cell engaged platform. The core concept is to use the bi-specific antibodies to reiterate and activate the T cell, bring them into the close proximity with the tumor cell to trigger a targeted immune response against the tumor or cancer. To address the different tumor indications, we combine this platform with the different tumor specific target, which has led to the three distant programs, such as AP203, 402, 601. These are currently invest one clinical trial in Taiwan or Australia. Overall, this whole portfolio reflects not only our successful partnership, but also the strength of our internal innovation in Gene, from validated later stage assay to our next TCO engagement platform. So, uh one thing I would like to emphasize is also my product, that one, uh, with the exception of IBI 302, which was rationalized, rationally designed by the recombining functional domain from the target receptor. All of our other products were discovered from our own fully human antibody library, OMIMAP. This means that all intellectual property and antibody sequence are fully owned by the AP BIO. So the OMIMAP, a server of the critical innovation source, uh consistently providing a unique functional antibody that strengths our next generation antibody development. Yeah. It's also wrong about uh our company, myself, and the whole Pine 9. Yeah.

SPEAKER_01 7:40

Thank you so much for the comprehensive uh overview, JJ, and background. I appreciate that. And can preemptive congratulations on the uh the IPO in in Taiwan as well as uh the good positive clinical readout that you alluded to um in your LEAP program. That's awesome. Uh yeah, thank you again. And so I really want to dive into kind of how your how your insights in discovery and and molecule design have downstream effects on product development and and the scaling of that manufacturing and all the considerations therein, right? And so to start off, I'm curious if you can kind of tell me a little bit about the key elements as you see it of biospecific or multi-specific antibody design and how that kind of might maybe conflict with traditional antibody design. And um and yeah, we'll start there.

SPEAKER_00 8:35

Yeah, sure. Um actually uh from the design perspective, uh, actually the biospecific and the multi-specific antibody really comes down to three key pillars or elements. First is the biology, second is the drug format, and the third one is the developability. And the first, uh just like I mentioned, is the biology. The target pair for the bi-specific or multi-specific, the target pair has to make the biological sense together. In other words, you need a clear rationale for why these two pathways or cell types should be engaged at the same time and what kind of functional synergy you expected to achieve. In many complex diseases, especially on college and immunomedial disorder, a single target is often not sufficient. So the design always starts from that biological hypothesis. So the second is a drug formate. Once the biologic is defined, the next question is how to physically build that functional into a single molecule. Here the binding geometry and the valency become just a very important as affinity. It's not only about whether each arm binds, but how the molecule we have in space and whether it can bring the rice cell or targets into strong connection. Champion uh strategy, a seismetrics, and the FC design all play the critical role in inspiration, purification, and overall manufacturing ability. For example, in our TQ platform, we use the symmetrics IgG fused to SCAV format, which is designed to efficiently bridge the immune infector cell and the tumor cell. This is optimizing the geometry for the functional activity rather than just the binding strengths. Importantly, the intended route of the administration and the dose in the regime also feedback into the format selection. And the last one is uh developability, which is something many in the early stage technology-driven biotech startup tend to underestimate. Uh, even if a molecule looks uh excellent in vitro, it still has to be stable, manufactured manufacturable, and have acceptable pharmacokinetics and the safety in vivo. So we spend a lot of effort optimizing the drug properties like the aggregation risk, uh, expression, and health life at the early stage to make sure it can actually become a real drug. So there's uh uh three things, three elements we we will consider when we design the multi-specific. And also uh manufacturing will be considered in a very early stage.

SPEAKER_01 12:05

Thank you, JJ. And yeah, like you're touching on, and you noted earlier in our briefing call as well, that many teams tend to ignore manufacturability or manufacturing difficulty until clinical efficacy is proven. And I'm curious to hear from you at what point in antibody engineering should teams begin thinking about manufacturability? How early should that take place? And is there such a thing as thinking about manufacturability too early in development?

SPEAKER_00 12:31

Yeah. Uh actually, I would say the manufacturing ability should be considered from the candidate's concept stage. Yeah. Uh not after you have seen the clinical efficacy. Uh, in more the traditional development models, manufacturing was often the trigger's uh downstream step. First, uh you validate the biology, and only then do you ask whether the molecule can be produced at a scale. But for bispecific or multi-specific antibody, that approach is too risky. If you wait too late, you often end up trying to rescue a molecule that was never really designed for manufacturing ability in the first place. And at that point, you have already invested heavily in the biology, only to discover issues like poor expression, instability, aggregation, or complex purification that are very difficult and expensive to fix. That's why, in our view, the manufacturing ability needs to be considered in the design process from the day one, even at the earliest screening stage. We are already looking at the key developability parameters, such as the expression of behavior, stability, aggregation tendency, and purification feasibility. Good teams don't run the biologist, uh biology and the developability sequentially. They run them in parallel because the reality is the cost of redesigning a molecule late in development is extremely high, both in time and resource. So early antibody or protein sequence optimization is not just a technical problem or step. It's a fundamentally determining whether a promising biological idea can actually become a scalable therapeutics. If you get the right upfront, everything uh downstream becomes uh far more predictable and uh efficient.

SPEAKER_01 14:56

Yeah. Yeah, thank you, JJ. I think it's it's often I don't I'm not gonna say it's lost on people, but it is it is um you you encounter this paradigm where great science and promising biology conflicts with the reality of this is an industry that requires products being taken to market. And if that product is inherently not manufacturable, or at least not manufacturable at scale, it doesn't matter how good the biology is, right? If you can't if you can't develop it and bring it to market. And I think that's it's worth impressing upon early stage companies that might be in this you know early phase of design as you're as you're alluding to. And we're gonna get a little bit more into it, but I'm curious um if you could specifically note how design decisions improve manufacturing outcomes, like what kind of decisions can affect what kind of manufacturing um levers or our outputs or just aspects of manufacturing.

SPEAKER_00 15:56

Yeah. Okay. Actually, our design decisions have a very direct and over uh underestimated impact on the manufacturing outcome. Uh at the fundamental level, what you define uh at the protein sequence and the format stage essentially determines how the molecule will behave through the product throughout the production. Even the uh related small choice, such as a framework selection, domains orientation, linker designs, or the specific amino acid optimization can meaningfully influence the expiration level, stability, aggregation risk, and overall product quality. So in general, simpler and more symmetric formats tend to be more robust. They improve the chain pairing, increase the title or ear, and simplify the purification. FC domain and the linker design in particular can have a major impact on both the expression and the stability, while also influencing how cleanly the molecule can be processed downstream. Actually, in many cases, uh avoiding the unnecessary structural complexity is itself a design advantage because it's reduced the risk of heterogeneity and manufacturing the failure model. So this format decision also reflects directly in uh into the CMC performance, affecting uh uh process consistency, scale up reliability, and even uh analytical simplicity. So the impact is not just uh biochemical, it's uh uh operational and industrial. A good example is our T cube, like the IgG fuse to SCLV design. It is uh built on uh symmetric IgG scale for with uh an SDL fused to the C terminal of the heavy chain, which is uh old-fashioned design, but is well suited for bridging the two different cell types from a functional standard standpoint. From a manufacturing perspective, this format is also highly practical. The intake IgG backbone allows us to leverage the established protein A-based purification workflow without introducing an entirely new downstream system. At the same time, the seismetrix format reduces the heavy light chain, mispairing a common challenge in more complex biospecific format, which in turn improves the ear and the product homogeneity. Overall, the better design molecules are not just biologically more effective, they are also significant, easier to develop, manufacture, and scale. In practical, good design directly reduces the downstream the CMC burden and increase the probability that a candidate can successfully become a job. Again, the final format still has to be aligned with the mechanism of action of the therapeutic you are trying to achieve. Thank you, JJ.

SPEAKER_01 19:43

Um, I'm gonna pose a bit of a hypothetical. I'm curious to hear your thoughts on it. Um, or maybe just looking back at your experience, have you ever found yourself in conflict or at a at a crossroads where you could maintain a certain aspect of a molecule or antibody's structure and design that might be complex, but complex in favor of its efficacy or mechanism of action? Um at the at the sacrifice of manufacturability and scale up. Or, you know, when you find yourself at a crossroads, what as a as a company and as an individual, what do you what do you put more weight in? What's more important to retain, like complex structure or ease of production? And and how do you kind of walk me through the way you make these kind of decisions as they come up in design?

SPEAKER_00 20:34

Yeah, it's kind of the balance. Yeah. Actually, uh when when we design by specific prototype, uh, a lot of the key optimization work is actually done very early. Uh before we even think about uh scale up, the goal is to de-risk manufacturing ability at the molecule design stage rather than trying to fix the issue later in the process development. So one of the first things we look at is the domain selection or interface stability. For example, uh taking uh AP505 as a reference, it's a VGF trapping domain, it's engineered from the VGF receptor domain through the recombinant designs. Because this domain is highly glycosylate and naturally very hydrophilic. So it contributes significantly to the overall stability of the molecule. From a manufacturing ability perspective, this translates into the excellent stability and allows us to achieve a relatively high formulation concentration over 40 mic per male, which very helpful.

SPEAKER_01 22:00

uh product development and clinical dosing convenience especially for the multi-specific program so um when we design our drug we always uh focus on what kind of mechanism we should uh focus on and uh what synergy in fact we would like to achieve so i think the potency could be the very important for us and the the second one we try to achieve the same potency through the engineering to uh overcome some uh developabilities problem so let's uh how we uh uh design or how we um implement in my in our uh designing the strategy got it thank you thank you um and yeah i'm curious too like i'm sure indication plays a role right whether you know for example um ibi 302 for eye diseases requires high concentrations uh for intravitrial injection um but that of course higher doses impose a a different production demand than uh a systemical a systemically injected antibody right and so i'm curious how how do things like indication and dosage uh affect your design decisions across your your portfolio and also you know going back to the your your point about balance how do you strike a balance between therapeutic novelty and long-term developability a very clear way to illustrate is the let's say to compare the IDC and the systemic oncology because the indication itself fundamentally uh reshaps what good design mean for a biologic molecule uh at the most the best basic level the indication determines the dose level dosing frequency root of the administration require concentration and the pharmacokinetic expectation of the molecule these uh parameters are not uh downstream the cost trend they are actually the upstream design input that should guide format decision from the very beginning take the intra bacteria delivered for the well and D, such as uh you mentioned IBI302 as an example, because the drug is injected directly into the eyeball, so the injection value is extremely small, which immediately force the molecule into very high local concentration environment that creates a completely different set of development pressure compared with the systemic administration.

SPEAKER_00 25:13

So here the viscosity subbility stability become the dominant constraint even if the biologist is strong, the molecule must remain stable at a high concentration without aggregation or the precipitation under the formulation stress. IBS02 is a good example our partner innovative biologist has done very strong clinical development work and importantly they have also built additional intellectual property around a high concentration formulation strategy. This is not just a formulation detail it's a strongly protect the product design space for the intraver administration and reflect the tidal integration between the molecule design formulation engineering and the IP strategy. It shows how early alignment across this dimension can matter strength the overall development package. In contrast systemic antibody used in oncology or immunologies affect a very different set of priorities. High concentration formulation is often less critical but instead of you prioritize the health life, systemic exposure, tumor penetration and target mediates trigger behavior. In this case FC engineering recycling through the FCR intention and the PK optimization often become the more central extreme formulation density or viscosity management. So at the end the indication is the downstream consideration is define the therapeutic context and the context directly determine what kind of molecule you should build both at the sequence level and the overall structure and the formal level.

SPEAKER_01 27:25

Thank you JJ Yeah I mean what you're saying makes sense right like higher concentrations would would correct me if I'm wrong but higher concentration for a something like the eye versus uh would that concentration be different from something that's systemically injected for example and yeah as a result that that just has that has downstream effects on um the way that these molecules behave when they're in in solution right and so I wanted to talk a bit about because you know when you whenever multi-specifics bi-specifics come up the issue I always hear about the challenge downstream at least is like you said aggregation precipitation um things just act funny when you get a lot of them together and you start purifying and also dosing affects affects purity and such but when you're designing a molecular prototype of a bi-specific for example what are the most common design level optimizations that you use to preemptively address these common manufacturing challenges such as aggregation or solubility challenges that can impact scaling of manufacturing yeah actually it's our strategy uh here I just want to give you some example um the first uh just like I mentioned earlier the the first thing we look into the domain selection just like I mentioned earlier the second one is the key area is the linker designs and the domain arrangement.

SPEAKER_00 28:58

In uh by specific format the spatial relationship between the binding domain is critical. A good example comes from our TCube uh CD137 SCLV best program where we systemically evaluate different linkers designs and SCIV orientation we found that this design choice can have a very large impact on developability especially just like we mentioned like the stability in fact the difference between the best and the worst configuration can reach up around three fourths in stability, simplified based on how the SDL is oriented and how the linker is designed or engineered. So these are not the mirror optimization they often determine whether a molecule is manufactured. Another important aspect is charge distribution and surface engineering as the teams we routinely perform the structure simulation or prediction analysis on the drug prototype, including the mapping of the electrostatic charge and hydrophobic distribution, the goal is to avoid a large surface patch of either strong positive or negative charge, as well as expose the hydrophobic region, since these are common driver of the self association and aggregation, when we identify this risk area we introduce the rational amino acid substitution to smooth out this localized patch and improve the overall molecule behavior. So the overall aggregation and stability are not just downstream manufacturing concern. They are built at the design stable if you optimize the prototype properly from start you can significantly improve the scalability robot and overall GMP manufacturing ability thank you JJ that's great. Before we get to our closing question the homework question of the show I just wanted to ask if there was anything else you wanted to tell our audience any lessons that you've learned along the way that you want to impart um any any stories about success or failure in in antibody design yeah I think uh actually I think uh in our development process I think uh first uh you you still have to consider all everything um manufacturing ability is still very important for uh when you develop a new drug when you uh consider any multi-specific or multi-functional antibodies you have to put this uh manufacturing ability into your mind first and then also you have to consider how to achieve to have the synergy in fact within a single molecule so if you decide at a very early stage and consider the manufacturing ability at an early stage uh I think your drug will have a big chance to be the success actually and uh one more thing I would like to share about uh this uh industrial field um the key issue is to actually when we develop a new antibody drug actually uh we we don't we don't worry about the the the uh we don't have a sufficient or uh brilliant scientist or advanced technology. The real the bottleneck is that industrial still operating under an invisible cost trend around maybe 10% clinical success rate ceiling that reflects how efficiently we translate early innovation into the real patient benefit. The key issue is that too many progress still optimized biologists first and treated developability and manufacturing ability as a second concern, secondary concern. This sequential mindset often lead to the late stage failure that could have been anticipated much easier sorry much earlier what we need instead is much tighter integration between the discovery, CMC translational science and the clinical strategy. So let the decisions at a very uh every stage are aligned from the start at the same time we should pay more attention to the molecule format that are both innovative and practical not just the biologically powerful but also feasible to manufacture formulated and deliver. AI and the computational tools can significantly accelerate this process but only when they are running a strong experimental validation and systemic integration. So going forward the focus should be on solving the real biological bottleneck, reducing the unnecessary complexity and the designing with the clinical in mind rather than the publication. This also includes using the data not only for discovery but to better understand and improve clinical trial design itself. So the bottom line is the the industry the advanced when we stop treating the manufacturing and the clinical feasibility as a downstream the cost train and start treating them as the integral part of what makes a molecule truly innovative yeah I hope that this uh kind of experience sharing can help you everyone to move forward. Yeah absolutely and so it's correct me if I'm wrong and tell me or I guess I'll just ask in your opinion uh JJ would you say pharmacology and manufacturability or development developability should be equal as far as you're considering antibody design they shouldn't be one shouldn't be weighted more than the other pharmacology and manufacturability should be kind of considered equally yeah that's right and uh when you design your molecule you have to consider this uh manufacturing ability at the beginning uh otherwise you will spend a lot of resource or money uh to overcome the developability when you get some clinical uh uh pro concept it's it's uh it's uh occupy it will suck in you your resource when you uh consider it's too late yeah it's a great point JJ thank you and uh last but not not least sorry thank you JJ and uh last but not least I'm gonna ask you the hallmark question of the show that I ask every guest and that is in your opinion big or small specific or broad how do you JJ think we can better biopharma yeah I think uh the the maybe just like I mentioned uh actually uh we we face uh some uh automate actually just like I mentioned as just 10% the clinical uh success rate the ceiling uh let's uh reflect uh uh how efficiently we translate early innovation into the real patient so I think um just we discussed earlier the manufacturing ability and the molecule the innovation should be considered at the same time so this will speed up our uh development process and also reduce our resource allocation uh so I think this uh when we consider these two key parameters when you design this job I think it will be good for every even a researcher or every biotech company yeah I agree uh JJ thank you so much for joining me on this episode of Better Biopharma the official podcast of Bioprocess Online and thank you out there for listening or tuning in wherever you get your podcasts we'll see you next time