BioMap Announces Positive Preclinical Data for ImmuBot Pipeline and Launches Modular Immunotherapy Toolkit

Combining cutting-edge AI and biotechnology, BioMap has built xTrimo, a trillion-parameter foundational large model for life sciences centered on protein language. Based on this, BioMap has also developed the AI Generative Protein Design Platform (AIGP). By modeling biological evolution and decoding the laws of life, BioMap has achieved breakthrough results in models across a series of fields including disease target discovery, innovative antibody drug design, and the design of short peptides and enzymes.

In 2023, BioMap and Sanofi reached an AI model development commercial cooperation deal worth over 1 billion US dollars. For the first time, the milestone was set as model development rather than drug research progress, becoming a new case of the unique MaaS (Model as a Service) business model in the large model era.

And what are the actual effects of these AI models, and what breakthroughs can they bring to drug development? After two years of refinement, BioMap's innovative drug development portfolio, ImmuBot (Immune Robot), has also delivered a exemplary答卷, with the preclinical data from the first batch of pipelines meeting expectations and steadily advancing toward the clinical stage. In addition, BioMap has developed its own patented plug-and-play modular immune cell engager multi-antibody platform, which enables precise regulation and generalizable combinations of various immune cells and multiple tumor targets, providing more possibilities for immunotherapy.

In the past decade, immunotherapy has thoroughly activated key markets such as oncology and autoimmune diseases. However, the pursuit of more precise, more efficient, and balanced broad-spectrum and targeted precision medicine has never stopped.When the world's largest foundational model for life sciences enters the pipeline and platform validation stage, will the plug-and-play, modular, and prefabricated treatment methods driven by large models become the future of "tailor-made" precision medicine?VCBeat brought these questions to interview Dr. Jiang Zhaoshi, Senior Vice President of BioMap.

Dr. Zhao Shi Jiang, Senior Vice President of BioMap

Dr. Jiang Zhaoshi has 20 years of experience in bioinformatics, drug target discovery, and drug design. Before joining BioMap, he served as a member of the drug R&D leadership team, Executive Director, and Head of the Bioinformatics and Data Science Department at Gilead Sciences. The team he led supported nearly 100 clinical trial studies and contributed to the clinical approval of nearly ten new drugs, including well-known medications such as Jyseleca (autoimmune), Vemlidy (hepatitis B), Biktarvy (HIV), and Remdesivir (antiviral). In addition to his expertise in clinical R&D, he has also made significant contributions to genomic scientific research, publishing over 40 high-quality scientific papers, five of which were featured in Nature. Prior to joining Gilead, as a senior scientist at Genentech, he led several large-scale cancer genomics research initiatives. He earned his Ph.D. in Genomics from the University of Washington under the mentorship of Dr. Evan Eichler, a member of the National Academy of Sciences and a renowned geneticist. Previously, he graduated from the Clinical Medicine program at Zhejiang University and worked for many years as a clinical pathologist at Peking Union Medical College Hospital.

VCBeat: Since its debut in September 2022, what internal iterations and new advancements has the ImmuBot immunobot undergone?

Jiang Zhaoshi:As an AI-centric enterprise, BioMap starts by collecting unprecedented, rich datasets, downloading, cleaning, and integrating almost all compliant and publicly available multi-omics big data related to human immunology, thereby creating the world's largest multimodal pre-trained model in life sciences.

Based on a high-throughput dry-wet closed-loop bio-computing engine to analyze the immune system, we utilize large models to precisely predict and decode complex immune patterns and disease mechanisms. We anticipate developing a new generation of protein or antibody-based drugs to achieve precision treatment for currently hard-to-address immune-related diseases. In September 2022, we first publicly disclosed BioMap's breakthrough innovative drug combination, ImmuBot (Immune Robot).

To date, we are pleased to share with you that, based on the company's proprietary technology, we have successfully completed the design, optimization, construction, and preclinical validation of BioMap's ImmuBot innovative drug asset portfolio. This portfolio includes the following functional characteristics:

Based on the aforementioned drug components and platforms, the company has developed multiple bispecific antibody pipelines targeting liver cancer, gastrointestinal solid tumors, and autoimmune diseases, particularly for patients who are resistant to current treatments or show poor response to first-line immunotherapies.Some pipelines have reached the preclinical stage, and both in vitro and animal experiments have demonstrated that our design has excellent efficacy and safety, achieving the expected design goals. We plan to further advance to CMC and clinical development.

VCBeat: In the current era where AI innovation is breaking through biotechnology barriers, how does the ImmuBot immune drug platform pave the way for precision in immunotherapy?

Jiang Zhaoshi:In order to address the pain points in immunotherapy and related drug development mentioned above, and further enhance the precision of immunotherapy, our ImmBot drug platform is expected to achieve the following goals:

First, a refined and controllable full set of immunotherapy tools has been developed.

First, it is refined and controllable.Take our first core bispecific antibody pipeline as an example, this is aCD3 T Cell Engager Bispecific Antibody with Conditional Activation Function, A unique molecular design ensures that this drug remains in a "standby" state in non-tumor tissues and is only activated by specific proteases in the tumor microenvironment, achieving the goal of targeted killing of tumor cells. This condition-triggered mechanism is expected to significantly reduce the on-target/off-tumor toxicity risk of CD3 antibodies and expand the therapeutic window of the drug. In comparative experiments, we have observed activation efficiency and therapeutic windows several times higher than competing drugs in the industry, along with very positive tumor suppression effects and high safety in animal studies.

Next is the full set of immunotherapy tools:In addition to conditional activation, we also have various functional regulatory elements such as logic gates and enhancers, which enable our immunotherapeutic drugs to possess powerful precise targeting and precise activation capabilities. In addition to CD3 T cells, we have also developed immune cell engagers for natural killer (NK) cells, gd T cells, and other types of immune cells, enhancing their killing efficacy through various regulatory elements. These innate immune cell engagers, due to their relatively high safety, are expected to demonstrate great potential in future combination therapies for cancer.

Second, establish a library of highly specific tumor-targeting warheads.

As an AI platform company in the life sciences field, BioMap possesses powerful database resources, target discovery algorithm capabilities, and experimental validation loop capabilities. By leveraging hundreds of billions of public immune databases and hundreds of millions of unique proprietary experimental data sets, we have established a top-tier immune target discovery platform in the industry and, based on this, identified a series of globally pioneering unique novel targets.

Many existing drug targets also have low expression in normal tissues, posing a risk of off-tumor toxicity. Therefore, by combining methods such as differential target expression and subcellular localization prediction, we have discovered new targets that traditional methods could not find. Additionally, for hard-to-drug targets with toxicity risks, we have implemented designs like logic gates to further enhance their safety. We believe these targets and target combinations will shine as the warheads of next-generation immune conjugate bispecific antibodies or ADC drugs.

VCBeat: What pain points in immunotherapy can the ImmuBot immune drug platform and innovative protein drugs address?

Jiang Zhaoshi:I think immunotherapy is the most crucial direction for treating complex human diseases. Take cancer as an example: why does cancer have certain characteristics associated with old age? Simply put, the immune system is a surveillance system that plays a strong role in monitoring tumor cells that undergo malignant transformation. In younger bodies, the probability of tumor cell mutations is lower, and the immune system is relatively stronger, which means it is highly likely to eliminate tumor cells. As aging occurs, immune system function declines, exposure time to carcinogens lengthens, and the likelihood of mutations gradually increases. With these opposing trends, cancer becomes relatively more common in older populations.

Therefore, from a therapeutic perspective, activating and enhancing the body's own immune mechanisms to kill, regulate, and treat tumors is the most natural and efficient approach. This is also why immunotherapy can achieve long-term tumor control with excellent efficacy in some patients.

Over the past 20 years, a significant advancement in the field of cancer treatment has been the invention of immune checkpoint inhibitors, which addressed the previous reliance on treatments such as radiotherapy, chemotherapy, ADC, and targeted therapy alone.However, the current efficacy rate of first-line immune checkpoint inhibitors still needs improvement, with only an average of 20%-30% of patients responding. Some patients experience drug resistance and recurrence after treatment. Additionally, the first generation of T cell-based immunotherapies has severe toxic side effects, which also impacts the applicability of such drugs. In summary, there is still significant room for expansion in cancer immunotherapy.

Where do the difficulties of these problems lie? At their root, it's because every individual is different. By studying the characteristics of the tumor microenvironment through multi-omics, especially single-cell sequencing and spatial transcriptomics research, we will discoverThe tissue microenvironment of tumor patients exhibits strong heterogeneity.This heterogeneity is manifested as: different infiltration levels of various immune cells, the benign or malignant degree of tumors, and the diversity of mesenchymal stem cell components associated with tumors. These differences can all influence the clinical efficacy of immunotherapy.

Therefore, future cancer treatment needs personalized design to formulate a treatment plan that fits the patient's characteristics and addresses the current challenge of inconsistent efficacy in immunotherapy. Designing products based on patient diversity requires the product itself to have sufficient flexibility —

The design of ImmuBot fully embodies this concept: modular, plug-and-play, and flexible product components, ranging from tumor-related targets, to different immune conjugates, to functional regulatory elements, which can bring technical possibilities for future personalized designs targeting different diseases, different patients, and different clinical stages.

VCBeat: Personalization is one of the main trends in precision medicine currently, but it also comes with highly controversial issues, such as industrialization, scalability, accessibility, and cost. How do you view this controversy?

Jiang Zhaoshi:The personalization of precision medicine will certainly involve challenges such as industrialization and large-scale production. However, personalization does not mean the ideal state of designing a product for each patient. Our philosophy is to find a balance between personalized design and large-scale production.

Traditional treatment plans that use the same medication and dosage for all patients are clearly unreasonable. Since each patient has a different tissue microenvironment, it is not surprising that clinical efficacy and toxic side effects vary widely. However, the challenges of personalization are twofold: First, the challenge of clinical development. It is not easy to strike a balance between personalized design and providing sufficient evidence to demonstrate its efficacy and safety during the regulatory review process. Second, the cost of production will be high. The contradiction between personalized design and large-scale production of products is not easily reconciled.

This is also why BioMap emphasizes the concept of plug-and-play today: we believe that,Not every patient needs to have a module and a product designed for them. Take a certain type of tumor as an example: by setting a specific biomarker as an identifying feature, patients can be divided into various possible subgroups. Within these patient subgroups, we can design one or more combination or sequential treatment plans, which are expected to achieve the optimal balance between personalized treatment and scalable production.

VCBeat: Therefore, what differentiated advantages do the characteristics of prefabrication, flexible assembly, plug-and-play, and modularity bring to BioMap's ImmuBot immune drug discovery platform?

Jiang Zhaoshi:Our immune cell engager multi-antibody platform has achieved a highly flexible plug-and-play capability at the antibody design level, meaningIn the development of "immune robot" drug molecules with different characteristics, various functional components can be flexibly allocated and rapidly assembled according to requirements (such as targeting different tumors and related antigens) to ensure that the drug molecules achieve optimal performance with low toxicity and high efficiency, while reducing the cost and cycle of development.

The various functional components we have designed includeThe possibility of reuse brings dozens or even hundreds of times improvement to the overall efficiency of drug development, minimizing the challenges that new products may bring in downstream large-scale production.

Most importantly, based on this technology, we willPossibleDevelop more "tailor-made" refined drugs for patient groups of various subtypes.Moreover, it is also conducive to the development of drugs with greater market demand and clinical urgency.

VCBeat: As both a platform and a drug (pipeline), how does BioMap envision the commercialization model for ImmuBot?

Jiang Zhaoshi:As for BioMap, the main business model will still focus on providing AI models and platforms in the life sciences field as external services. ImmuBot, as an important application scenario and demonstration project in our biopharmaceutical development field, we hope to further advance these high-value drug components and pipelines into clinical trials.

Based on immunomodulatory components and highly specific tumor targets, combined with an in-depth exploration of unmet clinical needs, we have independently initiated or collaboratively developed multiple bispecific antibody drugs. These drugs are gradually completing LEAD confirmation and in vivo validation, and their Best-in-class potential is beginning to show. We are highly looking forward to advancing these excellent molecules and drug pipelines into the clinical stage rapidly with the support of our partners and investors, ultimately benefiting patients.

At the same time, the gradual validation of these drug pipelines also means that BioMap's entire suite of immune cell engager platforms, immune modulation component platforms, and target combination discovery platforms have been firmly established. Moving forward, we warmly welcome interested partners to join hands with us in exploring the infinite possibilities of precision immunotherapy.

VCBeat: You just mentioned designing one or more immunotherapy combination regimens based on different populations or biomarker specificities. What is your perspective on combination immunotherapy?

Jiang Zhaoshi:In the past, I worked on HIV therapy development at Gilead. Initially considered a fatal disease, AIDS was later addressed with the introduction of the cocktail therapy, which involves the combined use of three or more drugs to reduce viral resistance or escape under single-drug treatment, maximally suppress viral replication, and restore compromised immune function, thereby extending patient survival. Nowadays, in long-term HIV treatment spanning 10 to 20 years, drug resistance rarely emerges, and patients can take a single pill daily to achieve high-quality, long-term survival. This has essentially transformed a once fatal condition into a manageable chronic illness.

This successful case also has a demonstration effect on tumor treatment. Today, a single therapy, whether it is chemotherapy, ADC, or immune checkpoint inhibitors, is difficult to completely and thoroughly eliminate solid tumors.In the foreseeable future, effective combination therapy or sequential treatment regimens are expected to offer cancer patients a long-term (20-30 years), high-quality life cycle.If we can achieve this goal, in a fundamental sense, it is already a functional cure for cancer. This is also the fundamental reason why we are building a plug-and-play, modular full set of immunotherapy toolkits today, and further developing multiple immunotherapy pipelines.

VCBeat: What are the empowerment directions for AI and large language models in combination therapies or multi-mechanism strategies?

Jiang Zhaoshi:The first direction, the discovery of innovative target combinations.When the efficacy of a single target is insufficient, AI algorithms and data science can help determine which new targets patients should choose, what marker expression levels, and what target combinations will achieve the best therapeutic effect and highest safety.

The second direction, AI therapy may help in the design and component selection of complex individual combination therapies.The complexity of selecting combination therapies goes far beyond today's simple experience-based drug combination schemes. For instance, factors such as the expression levels of biomarkers potentially having a weak correlation with prognosis, and the combination of multiple immune characteristics and features of the tumor microenvironment influencing the efficacy of combination therapies, all need to be incorporated into the design considerations for combination therapy.

The third direction, how immunotherapy targets tumor tissues, is essentially about identifying the differential characteristics compared to normal tissues.For example, the design of ImmuBot incorporates components with conditional triggers, such as changes in the enzymatic profile and pH levels within the tumor microenvironment. Through sufficient conditional triggering, ImmuBot can further increase the therapeutic window and enhance safety and efficacy.

VCBeat: From the broader perspective of biological computing and AI computing, what are the significant implications of exploring ImmuBot immune robots?

Jiang Zhaoshi:Since its establishment, BioMap has been a platform enterprise that hopes to empower life sciences through AI and data science.An excellent AI enterprise can never be a castle in the air or too sophisticated for practical use. After developing a highly advanced model, it is essential to have a grounded and applicable practice for validation.If the platform has strong predictive capabilities, then the success rate of future molecular design, product effectiveness, and R&D cycles should all show significant improvements to directly reflect the value of the AI platform.

Conversely,In drug development practice, the large amount of real-world data collected also provides more targeted data support for further optimizing the predictive capabilities of large language models in drug design. It can help the model achieve finer tuning, thereby enhancing its predictive power.Therefore, the two actually have a good interactive effect.

In the past, AI was relatively mature in small molecule drug design, but protein drug design was a significant gap. Protein 3D structure prediction, protein affinity prediction, and others have always been major technical challenges on AI computing platforms. However, today, a series of breakthroughs brought by high-performance computing and cutting-edge AI algorithms increasingly surpass the limitations of traditional experimental methods.

We have seen cases where targets previously considered "difficult to drug" or "undruggable" are gradually being overcome with the help of computational methods.In the next five to ten years, with the assistance of computers and AI, innovative protein-based drug products will continue to emerge, completely transforming the landscape of this field and bringing good news to patients.