2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - 393-396
This session explores how artificial intelligence, advanced molecular technologies, and innovative model systems are transforming hematologic diagnostics and precision medicine. Presentations span AI-enabled digital pathology and CBC-based immune profiling, rapid genomic testing in AML, computational prediction of cancer immunogenicity, and emerging platforms for therapeutic response testing—highlighting the evolution from traditional morphology and molecular assays toward integrated, data-driven diagnostics.
Technology and Procedures, Assays, Molecular Testing, Bioinformatics, Emerging Technologies, Treatment Considerations, Drug Development, Artificial Intelligence (AI), Machine Learning, Biological Processes, Immune Mechanism, Diseases, Lymphoid Malignancies, Lymphoid Leukemias, Myeloid Malignancies, Acute Myeloid Malignancies, CLL, Research, Fundamental Science, Translational Research, Study Population, Human
PAPER

142 Structure-guided cyclic peptide decoy design for interception of anti-AAV8 neutralizing antibodies

Bingqi Xu, PhD, Xiaolei Pei2, Dandan Yu3, Lei Zhang, MD4
1 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Tianjin, China
2 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Tianjin, China
3 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Tianjin, China
4 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Tianjin, China
Background:
Adeno-associated virus (AAV) vectors represent the cornerstone platform for in vivo gene therapy of inherited monogenic disorders. However, the high prevalence of pre-existing neutralizing antibodies (NAbs) in human populations and the mandatory induction of high-titer NAbs following initial administration severely impair vector transduction efficiency, presenting a major barrier to patient eligibility and vector re-administration. Current clinical modalities primarily rely on non-specific systemic immunosuppression or plasmapheresis, which suffer from off-target toxicity, transient efficacy, and a lack of epitope selectivity. Compared to flexible linear peptides, cyclic peptides offer superior thermodynamic and structural advantages. Through rigid conformational pre-organization, cyclic peptides restrict main-chain flexibility, minimizing entropic loss upon paratope binding while accurately mimicking the complex topological loops of capsid surface epitopes. Furthermore, the absence of free termini endows cyclic peptides with exceptional resistance against serum proteolytic degradation. Thus, this study established a structure-guided computational framework to design high-affinity cyclic peptide decoys for competitive interception of anti-AAV8 NAbs, offering a novel modality to overcome humoral immune barriers in gene therapy.
Methods:
We developed an integrated biophysical scanning and generative AI pipeline for structural epitope mapping and cyclic peptide decoy discovery. Using high-resolution AAV8 capsid structures, solvent-accessible surface area analysis was applied to systematically partition the exposed capsid surface into 177 spatial patches. AI-assisted structural generation, combined with backbone cyclization constraints, was deployed to reconstruct conformational mimetics of flexible variable regions (VRs). Spatial network clustering algorithms were then utilized to group correlated patches into distinct epitope communities, facilitating the rational design of broad-spectrum “pan-decoys” and domain-targeted “cluster-specific decoys.” Candidate peptides were systematically evaluated for binding kinetics, serum metabolic stability, and functional NAb neutralization rescue efficiency in cell-based AAV8 transduction assays under NAb challenge.
Results:
Topological analysis mapped key epitope communities onto canonical antigenic hot spots, including VR-IV, VR-V, VR-VIII, VR-IX, and the HI-loop, validating the precision of our computational model. The resulting library yielded pan-decoys with broad surface-masking potential across multiple patches, alongside cluster-specific decoys tailored to resolve polyclonal antibody heterogeneity. Biophysical and functional evaluations demonstrated that, benefiting from enhanced binding affinity and serum stability conferred by structural cyclization, lead cyclic decoys achieved submicromolar binding affinities and prolonged serum stability, significantly outperforming their linear counterparts. Under NAb challenge, lead cyclic decoys effectively sequestered anti-AAV8 antibodies and significantly rescued vector transduction efficiency in a dose-dependent manner.
Conclusion:
We established a robust “scanning–clustering–mimicking” platform for designing structure-guided cyclic peptide decoys. By leveraging the biophysical advantages of cyclic peptides—including high affinity, precise structural matching, minimal entropic penalty, and superior metabolic stability—this approach achieves both broad-spectrum interception and epitope-specific targeting of anti-AAV8 NAbs. This work provides fundamental insights into polyclonal antibody recognition patterns and delivers promising candidates to enable vector re-administration in clinical gene therapy.
Bingqi Xu, PhD: No relevant conflicts of interest to declare.