The challenges in solid tumour targeting with bispecific ADCs

Czech biotechnology company, SOTIO, recently started preclinical development of its first two bispecific antibody drug conjugates (ADCs), which it says are “just the beginning” of what it has in store for its next-generation ADC pipeline. DDW’s Megan Thomas speaks with SOTIO’s Chief Scientific Officer, Martin Steegmaier, about the company’s pipeline. MT: Bispecific ADCs are a promising frontier in drug discovery. What specific challenges in solid tumour targeting is SOTIO aiming to overcome with its approach? MS: Unlike standard ADCs targeted to a single tumour antigen, bispecific ADCs can be targeted to two antigens simultaneously. This dual-targeting format could improve the precision of ADC delivery — ensuring it gets to the cancerous tissue and not other, normal tissues — and could also help ADCs overcome the challenge of tumour heterogeneity, which is a key challenge in treating some of the solid tumours SOTIO is focused on. By targeting two distinct antigens, bispecific ADCs can more effectively target and kill a wider range of cancer cells within a heterogeneous tumour. Furthermore, targeting two distinct tumour antigens bares the potential to prevent or to overcome potential emerging resistances. If one target were to be downregulated by the tumour (or its associated molecular mechanism of internalisation), the cytotoxic payload could still be delivered to the tumour cell via the second internalising tumour antigen. MT: What considerations go into selecting and validating the two targets for your bispecific ADCs, and how do you ensure they synergise rather than interfere? MS: Selecting two targets for a bispecific ADC approach may sound straight forward but this is substantially more complex than the selection of a tumour antigen for a mono-targeting ADC approach. At SOTIO, we invest great thought and lots of resources in this key innovation step. The below mentioned questions are just some of the ones that need to be tackled in this process: It starts with the question which tumour indication we would like to address and whether for this tumour indication (or indications) there are at least to some extent validated internalising tumour targets. Equally important is the question whether there is underlying biology that hints that the potential target pairs physically interact or at least are colocalising on the surface of the tumour cell and whether the two target molecules are abundantly expressed on the surface of the tumour cells but ideally not or to a substantially lesser extent on the surface of any normal or vital healthy tissue to avoid added toxicities. There are the more mechanistical and technical questions. Things lake spatial consideration such as proximity and subcellular co-localisation of targeted epitopes and whether such pairs of targets (or monomers of one target) actually do get efficiently internalised and then end up in the lysosome to release the payload. Needless to say that all these questions will need to be answered by detailed expression and localisation analysis on tumour and healthy tissue via multiple approaches including mRNA and protein expression (proteomics, immunohistochemistry), often guided by big data and AI approaches. Equally important are mechanistic and proof-of-mechanism investigations in vitro (2D and 3D systems) and in vivo. MT: From a discovery and development standpoint, how do bispecific ADCs complicate or accelerate the preclinical pipeline? MS: Due to the increased complexity, timelines from project start to clinical candidate selected tend to be longer. In the case of a bispecific ADC, two binders that match or fit together in terms of affinities and epitopes need to be identified or engineered. There is another layer of complexity that needs to be tackled. The number of molecules that need to be tested and profiled multiplies. Instead of testing molecules on low or high target expression cells, all of a sudden one needs to look at all permutations of target 1 and 2 high and low. Needless to say that not only the discovery phase becomes more complex. Complexity is also added to the development aspects. Often one needs to develop two companion diagnostic tools instead of one. What is the expression level cut-off for target 1 and 2 to select or include patients in a trial? This complexity in most approached extends the timelines to the clinical candidate selected milestone and possibly even beyond, but if done right it will deliver highly differentiated drug candidates. MT: Can you describe the technology platform or engineering strategy SOTIO is using to construct its bispecific ADCs, and how it supports scalability and manufacturability? MS: We are utilising multiple technology platforms to construct our bispecific molecules, depending on the antibody binders we use or have available. The most straightforward technology is the so called common light chain approach combined with KiH (knob-into-hole) technology to ensure correct heavy chain pairing but in case our selected binders do not carry matching light chains, we rely on other, more sophisticated approaches. But in every case, we are ensuring good physicochemical properties of the molecules and thereby ensure good manufacturability and scalability of the production process. MT: Heterogeneity of tumour antigen expression is a major hurdle in oncology. How do your designs help mitigate this, and what data supports that promise so far? MS: You are probably familiar with the terms “AND” or “OR” gate approaches. In addition, we are (for some projects) also applying a so called “AND/OR” gate approach. This combines the best of both worlds. This allows us to combine a relatively clean target (but maybe more heterogeneously expressed target), with a somewhat “dirtier”, more broadly expressed target. By combining this more heterogeneously expressed but clean target with a somewhat “dirty” but highly internalising target, one can overcome the heterogeneous expression for this clean target and, if the design of the molecule in terms of affinities and binding valency is right, one still has a highly internalising and clean (safe) molecule. In preclinical tumourmodels we could indeed show improved activity on tumours that display a heterogenious antigen expression. We are about to start safety studies in non-human primates to demonstrate that the more “promiscuous” nature of the targeting antibody that allows us to overcome the tumour heterogeneity is nevertheless clean and well tolerated. MT: What are the major learnings from the development of your first two bispecific ADC candidates, and how will they shape the next generation in your pipeline? MS: We experienced that a simple so called 1+1 molecular design of the bispecific targeting antibody often is not sufficient to translate into an improved and differentiated profile of the drug in comparson to already existing monotargeting approaches. The majority of the bispecific ADC projects today are such 1+1 approaches but tumour and target biology is more complicated or complex and more sophisticated designs are needed. For all our bispecific projects we are nowadays not only testing antibody pairs in the standard 1+1 format, but we also include a format screen including 2+2 or asymmetric bispecific formats such as 2+1. This ensures that we can indeed address such often more complex target biologies and distributions. MT: For drug discovery researchers working on ADCs or multispecific formats, what will differentiate clinical success from failure? MS: Many of the bispecific ADC approaches are built on simply combining two “available” binders for at first sight obvious target combinations in a 1+1 format, but biology is much more complex than that. Just look at how many 1+1 bispecific ADCs targeting EGFR and HER2 or similar combinations are seen in the drug development pipelines. We have to go beyond such first generation bispecific approaches! This is of course technically much more challenging and consequently more lengthy and resource intensive but but putting in more thought and molecule optimisation cycles will in the long term pay off ensuring a really diffenentiated molecule and not only one that only caries the label “bispecific” but is otherwise not truely differentiated and improved versus a monotargeting approach. MT: Looking ahead, how does SOTIO envision the role of bispecific ADCs within the broader oncology therapeutic landscape? Will they complement or compete with existing modalities like cell therapies or immune checkpoint inhibitors? MS: In contrast to cell therapies, bispecific ADCs can be developed as “off-the shelf” precision medicines. This will allow them to not only get approval in late line settings but we eventually will see them moving into earlier lines of therapies offering differentiated treatment options for many more patients. Considering the logistical complexity of (at least autologous) cell therapy approaches and the associated high cost of such a treatment, (in contrast) a bispecific ADC approach for me can develop into a stand alone regiment even in earlier line settings. Combining this with CART approaches are less likely to be seen. Mechanistically, this makes lesser sense. In contrast, combinations with CPIs are indeed to be considered for the potential biological synergies such combinations can and will create. Biography Martin Steegmaier has held leadership roles at Boehringer Ingelheim, Roche, and MorphoSys, focusing on oncology drug discovery and antibody-based therapeutics. In 2022, he became Chief Scientific Officer at SOTIO. He earned a PhD in biochemistry from the University of Basel and an MBA from Edinburgh Business School.