e14068 Background: Although comprehensive genomic profiling has revealed the molecular features of glioblastoma (GBM), most identified genomic alterations fail to translate into effective therapeutic targets to date, underscoring a critical gap between molecular characterization and druggable GBM biology. Tumor cell surface antigens play a crucial role in novel cancer therapies such as CAR-T and antibody-drug conjugates. The lack of clinically actionable, tumor-selective cell-surface biomarkers remains a fundamental obstacle to such novel approaches for GBM. Alpaca-derived variable domain of heavy chain of heavy chain (VHH) antibodies possess a simple structure, are easily digitized through sequencing, and enable AI-driven biomarker discovery. Methods: We established Inverse BioMarker Exploration Technology (IBMET) using VHH, a fundamentally different biomarker discovery approach that treats antibodies as structural sensors rather than affinity reagents. A large-scale alpaca-derived VHH repertoire was generated through immunization using multiple GBM cell lines, followed by phage display and deep sequencing. Instead of selecting antibodies against predefined targets, IBMET applies enrichment-based statistical analysis across millions of VHHs to identify convergent binding patterns indicative of shared tumor-specific structural epitopes. Candidate VHHs were validated by immunohistochemistry and immunofluorescence using extensive panels of GBM and normal human tissues. Antigen identity was resolved by cross-linking immunoprecipitation and LC–MS/MS. Clinical relevance was assessed in an independent GBM cohort (n = 20). Results: IBMET uncovered multiple previously inaccessible structural biomarker candidates in GBM. The lead clone, VHH19, selectively recognized a novel antigen protein not previously implicated in GBM biology. VHH19 demonstrated robust and highly tumor-specific staining in GBM tissues, with negligible reactivity across normal organs. Notably, VHH19 positivity was observed in over 20% of GBM cases, defining a clinically relevant tumor subset unified by a shared structural epitope rather than genomic alteration or expression level. These findings reveal a previously unrecognized layer of tumor stratification based on conformational antigen states. Conclusions: IBMET represents a paradigm shift in biomarker discovery by enabling direct, high-throughput identification of structurally defined tumor-selective targets independent of prior molecular assumptions. By bridging antibody-level structural recognition with translational validation, this approach opens a new route to actionable targets in GBM, with immediate implications for antibody–drug conjugates, radioligand therapies, and companion diagnostics. Four GBM-selective VHH antibodies identified through IBMET will be released for research use in June 2026, facilitating rapid downstream translation.

Abstract Background: In cancers with substantial unmet medical need (UMN), the limited availability of reliable and tumor-selective surface biomarkers continues to hinder therapeutic development. This challenge is particularly evident in triple-negative breast cancer (TNBC) and pancreatic ductal adenocarcinoma (PDAC), where currently available targets often show insufficient specificity or restricted applicability across patient subgroups. Methods: We established the Inverse Biomarker Exploration Technology (IBMET), a systematic framework designed to detect pathological structural alterations on tumor cells. IBMET utilizes a large alpaca-derived VHH antibody repertoire as highly sensitive structural probes. Alpacas were immunized with multiple tumor cell lines, and the resulting VHH library was analyzed using phage display and next-generation sequencing. Candidate antibodies were selected based on statistical enrichment and evaluated by IHC/IF across an extensive panel of tumor and normal tissues. Antigen identification was performed by cross-linking immunoprecipitation, SDS-PAGE, and LC-MS/MS. Clinical relevance was examined using a breast cancer biopsy cohort (n = 106). Results: IBMET identified several structural biomarker candidates relevant to TNBC and PDAC. The lead antibody, VHH89, selectively recognized a previously uncharacterized low-molecular-weight isoform of ALCAM (approximately 70 kDa; designated ALCAM70). VHH89 showed lesion-selective staining in tumor tissues with minimal reactivity in normal organs, indicating a high degree of tumor specificity. In clinical breast cancer specimens, VHH89 demonstrated positivity in more than 20% of TNBC cases. In addition, subsets of pancreatic cancer and cholangiocarcinoma specimens also showed VHH89 positivity, suggesting that this isoform may represent a structurally altered antigen present across multiple tumor types. Conclusion: IBMET offers a reproducible and scalable approach for identifying structurally defined biomarkers that may not be detectable using genomic or transcriptomic analyses alone. By systematically excluding antibodies that react with normal tissues and enriching for tumor-associated conformational epitopes, IBMET broadens the spectrum of actionable targets for aggressive cancers. These findings support the potential integration of IBMET-derived antibodies into the development of next-generation antibody-drug conjugates, radioligand therapies, and companion diagnostics. To facilitate broader validation, COGNANO will make eight PDAC-selective VHH antibodies available for research use at AACR 2026. Citation Format: Akihiro Imura, Ryota Maeda, Hiroyuki Yamazaki, Tsuyoshi Inoue, Sadako Aakashi-Tanaka, Hiroko Tsukada. Ultra-sensitive structural biomarker discovery for TNBC and pancreatic cancer using a deep VHH repertoire (IBMET): A translational platform for tissue-agnostic target identification [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6414.

Background Conventional antibody discovery approaches that do not account for enrichment-driven biases, such as epitope immunogenicity, PCR amplification bias, or protein expression efficiency, may result in under-representation of rare yet functionally relevant clones, necessitating labor-intensive in vitro screening to identify agonistic antibodies among a large number of dominant clones. Thus, efficient screening methods for agonistic antibodies are urgently needed. OX40 is a promising target for cancer immunotherapy due to its role in enhancing T-cell activation and survival. However, effective anti-OX40 agonistic antibodies have not yet been developed. Methods We developed a novel screening strategy that involves the selection of nanobody clone pools enriched by biopanning against gp34-engaged and non-engaged OX40-expressing cells, next-generation sequencing, and computational clustering and subtraction analysis to identify clones recognizing the ligand-receptor interface. Representative nanobody clones underwent in vitro validation, including epitope mapping, binding affinity measurements, and functional assessments. Furthermore, we engineered the selected nanobody to enhance its in vivo efficacy. We also performed structural analysis of the nanobody-OX40 complex. Results Our epitope-directed approach efficiently identified nanobody clones recognizing functionally relevant epitopes distinct from dominant immunogenic regions. Notably, clone Nb479 demonstrated robust agonistic activity, closely mimicking the natural ligand gp34 with extensive OX40-binding interactions. Trimerization of Nb479 facilitated potent OX40 activation without the need for a cross-linking scaffold. Conjugation of the Nb479 trimer with an anti-serum albumin nanobody exhibited significantly improved pharmacokinetics in vivo and enhanced antitumor activity in a mouse model treated with CD19 chimeric antigen receptor T cells. Conclusion This study presents an innovative epitope-directed approach that greatly accelerates the discovery of functionally potent agonistic nanobodies by effectively circumventing enrichment-driven epitope bias. Our approach and engineered multivalent anti-OX40 nanobody offer a powerful platform to advance immunotherapeutic strategies for cancer treatment.

Antibodies are crucial proteins produced by the immune system to eliminate harmful foreign substances and have become pivotal therapeutic agents for treating human diseases. To accelerate the discovery of antibody therapeutics, there is growing interest in constructing language models using antibody sequences. However, the applicability of pre-trained language models for antibody discovery has not been thoroughly evaluated due to the scarcity of labeled datasets. To overcome these limitations, we introduce AVIDa-SARS-CoV-2, a dataset featuring the antigen-variable domain of heavy chain of heavy chain antibody (VHH) interactions obtained from two alpacas immunized with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike proteins. AVIDa-SARS-CoV-2 includes binary labels indicating the binding or non-binding of diverse VHH sequences to 12 SARS-CoV-2 mutants, such as the Delta and Omicron variants. Furthermore, we release VHHCorpus-2M, a pre-training dataset for antibody language models, containing over two million VHH sequences. We report benchmark results for predicting SARS-CoV-2-VHH binding using VHHBERT pre-trained on VHHCorpus-2M and existing general protein and antibody-specific pre-trained language models. These results confirm that AVIDa-SARS-CoV-2 provides valuable benchmarks for evaluating the representation capabilities of antibody language models for binding prediction, thereby facilitating the development of AI-driven antibody discovery. The datasets are available at https://datasets.cognanous.com.

Abstract Currently, candidate molecules such as MUC1, Mesothelin, CEACAM6, and Claudin18.2 have been proposed as immunotherapy targets in pancreatic cancer. These molecules are significantly overexpressed in cancer cells, allowing differentiation from normal cells, and are being targeted for drug development using immunotherapy modalities like antibody drug conjugate (ADC), T cell bispecific (TCB), and CAR-T. Therefore, if more specific target molecules exist, it would be possible to develop even higher quality drug. However, do such target molecules exist? We conducted long-term immunization of alpacas, which are single-domain antibody producers, using more than 10 types of human pancreatic cancer cell lines. Over time, we constructed a large library of VHH antibodies recognizing the cell surface. Based on this, we developed an algorithm to cluster millions of antibody amino acid sequences in silico and identify characteristic clusters for each pancreatic cancer cell line. The Inverse Biomarker Exploring Technology (IBMET) is a bio-AI fusion technology that comprehensively converts biologically generated single-domain antibodies into a digital library to discover new biomarker molecules by using features correlated with epitopes as indicators. We validated antibodies recognizing new markers in patient tissues and ultimately identified several biomarkers. Among these, we here report the most distinctive biomarker, VHH89T antigen. The antigen annotation revealed by MSpec analysis identified it as CD166, known as ALCAM, a type I membrane protein. ALCAM is a membrane molecule widely expressed in the body, including immune cells. However, VHH89T specifically recognized only the complex formed by a small subset of ALCAM with its partner molecule. Furthermore, the partner molecule of the ALCAM complex was identified as membrane protein X and it was confirmed that the multimeric complex was rapidly internalized into the cell upon antibody binding. Consequently, we formulated VHH89T into an ADC bound to SN-38 and tested its cytotoxic function in cultured cells, confirming very favorable efficacy. Also efficacy has been confirmed in a mouse xenograft model. The novelty and advantages of IBMET are described as The ability to identify abnormalities (features) in post-translational modifications or complex formation, which are not detectable by genome or mRNA expression analysis, by converting antibodies that recognize target molecule structures into amino acid sequence information. Immediate formulation into ADCs and other drugs upon biomarker discovery, as antibodies are already available. Instant response to necessary companion diagnostics in parallel with drug formulation. This enables the provision of a super-rapid and highly probable next-generation antibody drug development platform. With IBMET, we have successfully identified similar marker molecules in not only pancreatic cancer but also cholangiocarcinoma and TNBC, and are currently aiming to pipeline multiple targets. Citation Format: Akihiro Imura, Kazutaka Araki, Tsuyoshi Inoue, Ryota Maeda, Hiroyuki Yamazaki. Inverse Biomarker Exploring Technology (IBMET) Mathematical Identification of Cancer-Specific Epitopes and Novel Targets (Biomarkers) from Big Data of Single-Domain Antibodies Recognizing Higher Structures [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research; 2024 Sep 15-18; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl_2):Abstract nr A071.

The continuous emergence of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) variants associated with the adaptive evolution of the virus is prolonging the global coronavirus disease 2019 (COVID-19) pandemic. The modification of neutralizing antibodies based on structural information is expected to be a useful approach to rapidly combat emerging variants. A dimerized variable domain of heavy chain of heavy chain antibody (VHH) P17 that has highly potent neutralizing activity against SARS-CoV-2 has been reported but the mode of interaction with the epitope remains unclear. Here, we report the X-ray crystal structure of the complex of monomerized P17 bound to the SARS-CoV-2 receptor binding domain (RBD) and investigated the binding activity of P17 toward various variants of concern (VOCs) using kinetics measurements. The structure revealed details of the binding interface and showed that P17 had an appropriate linker length to have an avidity effect and recognize a wide range of RBD orientations. Furthermore, we identified mutations in known VOCs that decrease the binding affinity of P17 and proposed methods for the acquisition of affinity toward the Omicron RBD because Omicron is currently the most predominant VOC. This study provides information for the rational design of effective VHHs for emerging VOCs.

Breast cancer can be classified into several types according to the expression patterns of human epidermal growth factor receptor 2 (Her2), oestrogen receptor (ER), and progesterone receptor (PgR) proteins. The prognosis of patients with tumors showing low Her2 expression and no ER and PgR expression—categorized as triple-negative breast cancer (TNBC)—is worst among these groups. Due to the lack of specific antibodies for TNBC, curative treatments for TNBC remain limited. Antibodies targeting TNBC have potential as diagnostic and therapeutic tools. Here, we generate a panel of nanobodies targeting TNBC cell lines by immunizing alpacas and subsequently panning the resulting phage libraries with TNBC cell lines. We show that several clones exclusively stain Her2-negative cells in tissues of breast cancer patients, and a few clones stain both Her2-positive and Her2-negative regions in these tissues. These clones can be applied to patient-specific therapies using drug-conjugated antibodies, radiolabelled antibodies, chimaera antigen receptor T cells, or drug delivery components, as well as to TNBC diagnosis.

Antibodies have become an important class of therapeutic agents to treat human diseases. To accelerate therapeutic antibody discovery, computational methods, especially machine learning, have attracted considerable interest for predicting specific interactions between antibody candidates and target antigens such as viruses and bacteria. However, the publicly available datasets in existing works have notable limitations, such as small sizes and the lack of non-binding samples and exact amino acid sequences. To overcome these limitations, we have developed AVIDa-hIL6, a large-scale dataset for predicting antigen-antibody interactions in the variable domain of heavy chain of heavy chain antibodies (VHHs), produced from an alpaca immunized with the human interleukin-6 (IL-6) protein, as antigens. By leveraging the simple structure of VHHs, which facilitates identification of full-length amino acid sequences by DNA sequencing technology, AVIDa-hIL6 contains 573,891 antigen-VHH pairs with amino acid sequences. All the antigen-VHH pairs have reliable labels for binding or non-binding, as generated by a novel labeling method. Furthermore, via introduction of artificial mutations, AVIDa-hIL6 contains 30 different mutants in addition to wild-type IL-6 protein. This characteristic provides opportunities to develop machine learning models for predicting changes in antibody binding by antigen mutations. We report experimental benchmark results on AVIDa-hIL6 by using machine learning models. The results indicate that the existing models have potential, but further research is needed to generalize them to predict effective antibodies against unknown mutants. The dataset is available at https://avida-hil6.cognanous.com.

We are amid the historic coronavirus infectious disease 2019 (COVID-19) pandemic. Imbalances in the accessibility of vaccines, medicines, and diagnostics among countries, regions, and populations, and those in war crises, have been problematic. Nanobodies are small, stable, customizable, and inexpensive to produce. Herein, we present a panel of nanobodies that can detect the spike proteins of five SARS-CoV-2 variants of concern (VOCs) including Omicron. Here we show via ELISA, lateral flow, kinetic, flow cytometric, microscopy, and Western blotting assays that our nanobodies can quantify the spike variants. This panel of nanobodies broadly neutralizes viral infection caused by pseudotyped and authentic SARS-CoV-2 VOCs. Structural analyses show that the P86 clone targets epitopes that are conserved yet unclassified on the receptor-binding domain (RBD) and contacts the N-terminal domain (NTD). Human antibodies rarely access both regions; consequently, the clone buries hidden crevasses of SARS-CoV-2 spike proteins that go undetected by conventional antibodies.

Abstract SARS-CoV-2 Omicron variants are highly resistant to vaccine-induced immunity and human monoclonal antibodies. Here, we demonstrate that a novel nanobody TP86 potently neutralized both BA.1 and BA.2 Omicron variants, and that the TP17 and TP86 nanobody cocktail broadly neutralized in vitro all VOCs as well as D614G. Furthermore, this cocktail showed therapeutic efficacy in vivo on VOCs using human ACE2 transgenic mice.

Background SARS-CoV-2 Omicron variants are highly resistant to vaccine-induced immunity and human monoclonal antibodies. Methods We previously reported that two nanobodies, P17 and P86, potently neutralize SARS-CoV-2 VOCs. In this study, we modified these nanobodies into trimers, called TP17 and TP86 and tested their neutralization activities against Omicron BA.1 and subvariant BA.2 using pseudovirus assays. Next, we used TP17 and TP86 nanobody cocktail to treat ACE2 transgenic mice infected with lethal dose of SARS-CoV-2 strains, original, Delta and Omicron BA.1. Results Here, we demonstrate that a novel nanobody TP86 potently neutralizes both BA.1 and BA.2 Omicron variants, and that the TP17 and TP86 nanobody cocktail broadly neutralizes in vitro all VOCs as well as original strain. Furthermore, intratracheal administration of this nanobody cocktail suppresses weight loss and prolongs survival of human ACE2 transgenic mice infected with SARS-CoV-2 strains, original, Delta and Omicron BA.1. Conclusions Intratracheal trimerized nanobody cocktail administration suppresses weight loss and prolongs survival of SARS-CoV-2 infected mice.

We are in the midst of the historic coronavirus infectious disease 2019 (COVID-19) pandemic caused by severe respiratory syndrome coronavirus 2 (SARS-CoV-2). Although countless efforts to control the pandemic have been attempted—most successfully, vaccination1–3—imbalances in accessibility to vaccines, medicines, and diagnostics among countries, regions, and populations have been problematic. Camelid variable regions of heavy chain-only antibodies (VHHs or nanobodies)4 have unique modalities: they are smaller, more stable, easier to customize, and, importantly, less expensive to produce than conventional antibodies5, 6. We present the sequences of nine alpaca nanobodies that detect the spike proteins of four SARS-CoV-2 variants of concern (VOCs)—namely, the alpha, beta, gamma, and delta variants. We show that they can quantify or detect spike variants via ELISA and lateral flow, kinetic, flow cytometric, microscopy, and Western blotting assays7. The panel of nanobodies broadly neutralized viral infection by pseudotyped SARS-CoV-2 VOCs. Structural analyses showed that a P86 clone targeted epitopes that were conserved yet unclassified on the receptor-binding domain (RBD) and located inside the N-terminal domain (NTD). Human antibodies have hardly accessed both regions; consequently, the clone buries hidden crevasses of SARS-CoV-2 spike proteins undetected by conventional antibodies and maintains activity against spike proteins carrying escape mutations.

The posttranslational modification of histones is crucial in spermatogenesis, as in other tissues; however, during spermiogenesis, histones are replaced with protamines, which are critical for the tight packaging of the DNA in sperm cells. Protamines are also posttranslationally modified by phosphorylation and dephosphorylation, which prompted our investigation of the underlying mechanisms and biological consequences of their regulation. On the basis of a screen that implicated the heat shock protein Hspa4l in spermatogenesis, we generated mice deficient in Hspa4l (Hspa4l-null mice), which showed male infertility and the malformation of sperm heads. These phenotypes are similar to those of Ppp1cc-deficient mice, and we found that the amount of a testis- and sperm-specific isoform of the Ppp1cc phosphatase (Ppp1cc2) in the chromatin-binding fraction was substantially less in Hspa4l-null spermatozoa than that in those of wild-type mice. We further showed that Ppp1cc2 was a substrate of the chaperones Hsc70 and Hsp70 and that Hspa4l enhanced the release of Ppp1cc2 from these complexes, enabling the freed Ppp1cc2 to localize to chromatin. Pull-down and in vitro phosphatase assays suggested the dephosphorylation of protamine 2 at serine 56 (Prm2 Ser56) by Ppp1cc2. To confirm the biological importance of Prm2 Ser56 dephosphorylation, we mutated Ser56 to alanine in Prm2 (Prm2 S56A). Introduction of this mutation to Hspa4l-null mice (Hspa4l -/-; Prm2 S56A/S56A) restored the malformation of sperm heads and the infertility of Hspa4l -/- mice. The dephosphorylation signal to eliminate phosphate was crucial, and these results unveiled the mechanism and biological relevance of the dephosphorylation of Prm2 for sperm maturation in vivo.

In this paper, we propose a high voltage brushless AC starter that contributes to improved fuel efficiency and a reduction in the cost of the one-motor two-clutch hybrid system, which we call a 1MG2CL system. We have named it the HV starter, and it is composed of an AC motor, inverter and pinion with a shift mechanism. One of the issues with the 1MG2CL system is the high electrical energy when starting an ICE as it switches over from EV drive to HEV drive. While the ICE is starting, the main motor has to crank the ICE via the clutch; the clutch slips to absorb the main motor power, so the main motor has to output a high power to overcome the loss. Therefore, to contribute to reducing the electrical power by eliminating clutch slip losses, we developed an HV starter as a dedicated ICE starting device. Thanks to the reduction in electrical power, the HV starter is able to improve fuel efficiency and reduce system costs. However, the major issue is the cranking noise generated by the pinion mechanism of the starter. We have developed a noise control strategy to solve this issue. The concept is to reduce the dependency of the pinion operation on the starter. We aim to utilize the ICE inertia energy in the later cranking process instead of starter cranking. Using the HV starter, which is a high power and controllable AC motor, we have managed to control the cranking noise to the target level, which is equal to the road noise during EV driving.

Model predictive instantaneous-current control (MPIC), which was proposed in our earlier works, enables us to achieve better instantaneous current control using mathematical models of an inverter and permanent magnet synchronous motors (PMSMs). However, dead-time to avoid short breakdown in the inverter is not usually considered in a general inverter model. Such an unmodeled part in the inverter model prevents accurate prediction of current evolution in motor systems based on the model predictive control. Therefore, in this paper, we analyze current response resulting from the dead-time in the MPIC, and propose a refined inverter model considering the dead-time so that control performance is improved. The effectiveness of the proposed method is verified through simulation and experiments. (c) 2013 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

The direct torque control (DTC) of ac motors leads to a faster torque response with a small number of switching operations in inverters when compared with a conventional approach such as vector control. DTC exhibits a hybrid nature in the sense that the system is composed of continuous variables of torque and flux and involves discrete switching in the inverter. The output of the inverter is limited to the finite discrete values at each instance of sampling. Model predictive control (MPC) is applied to the system so that an optimal switching sequence is derived subject to the given constraints. The proposed MPC-based approach reduces the torque ripple with a small number of switching operations. The effectiveness of the proposed MPDTC approach is verified through simulations and experiments by comparing the proposed approach with conventional DTC.