ASGCT 2026: Process Development, Technology Transfer, Clinical Manufacturing, and Process Characterization of AFTX-201, an Investigational New Medicine for the Treatment of BAG3-Associated Dilated Cardiomyopathy

Process Development, Technology Transfer, Clinical Manufacturing, and Process Characterization of AFTX-201, an Investigational New Medicine for the Treatment of BAG3-Associated Dilated Cardiomyopathy Matt Edwards, Vice President Process Science 29th Annual Meeting of the American Society of Gene and Cell Therapy, Boston, Massachusetts May 13th, 2026 2 Disclosures ▪ I am an employee of Affinia Therapeutics BAG3 DCM is a devastating disease 3 DeGroot, S. (2024, August 16). Facebook; accessed October 2, 2024. Dominguez et al., JACC, 2018 • Current treatments for symptoms only • 22% require a heart transplant • ~76K patients in developed markets • Present with heart failure phenotype High unmet need AFTX-201 has the potential to restore lost function in patients affected by BAG3 DCM ATC-0187 Capsid (Engineered AAV9) + Muscle-Specific Promoter + BAG3 cDNA 4 Process flow diagram for AFTX-201 Use of Affinia’s proprietary plasmid design resulted in high AAV yields The formulation for AFTX-201 resulted in product stability for at least 12 months The purification process yielded a high-quality vector suitable for clinical use Harvest yields for AFTX-201 exceeding 6e15 vg/L are achievable using Affinia’s plasmid system 5 Tech Transfer Run (50L) Demonstration Run (50L) GMP Run (50L) Harvest Yield (vg/L) 4e15 4.48e15 6.07e15 Total vg per 50L run (clarified lysate) 2e17 2.24e17 2.9e17 Total process yield 21.5% 24.7% 43% High process yields and a low clinical dose enables ample manufacturing capacity at the 50L scale, making supply of thousands of doses readily achievable We believe the manufacturing process is readily scalable to meet demand, if needed AFTX-201 process generates high-quality, clinical-ready vector 6 Tech Transfer Run (50L) Demonstration Run (50L) GMP Run (50L) Capsid Purity (% VP1/2/3) 98.3 100 100 hcDNA (ng per 1e13) 15.3 33 30 HCP (ng/mL) BLOQ BLOQ BLOQ Residual plasmid DNA (cp/mL) 4e11 5.75e11 3.04e11 Residual E1A DNA (cp/mL) BLOQ BLOQ BLOQ AUC (% Full / Partial / Empty) 95 / 3 / 1 95 / 2 / 2 96 / 2 / 2 Aggregation (% Monomer) 90.7 96.6 99.5 In vitro Potency (% Reference Standard) 187.9 86.0 92.3 Replication competent AAV BLOQ BLOQ BLOQ The AFTX-201 formulation has demonstrated stability through one year of storage at −80°C. Stability studies are ongoing, with an anticipated drug product shelf life of up to three years. 7 Establishing a representative scale-down model for the AFTX-201 process to support process validation Clinical Process Scale-Down Process 8 Establishing a late-stage representative scale-down model for the AFTX-201 process Identify a scale-down bioreactor model that replicates 50 -L performance while remaining within midstream operating parameters Develop a scale-down CsCl gradient model that reproduces manufacturing scale performance These processes can be scaled linearly 9 1 x 25 cm2 2 x 1 mL 1 x 0.01 m2 4 x 25 cm2 1 x 10 mL 5 x 0.01 m2 8.9 mL 2 x 10 mL 1 x 0.1 m2 1 x 340 cm2 4 x 10 mL 2 x 0.1 m2 39 mL 1 x 1020 cm2 1 x 75 mL 4 x 0.1 m2 Multiple Tubes 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Primary Clarification Secondary Clarification TFF1 UC Liters of Lysate Harvested Scaledown Operating Ranges 1L Scale: Allows efficient representative processing to DP of upstream experiments. 4.4L Scale: Provides material for testing multiple downstream conditions / more precise downstream parameter evaluation Clarification filter sizing is the primary driver of midstream and bioreactor scaling 10 AFTX-201 upstream scales efficiently across reactor sizes 50L 1L 2L 3L 4.5L 0 25 50 75 100 Reactor Scale Producvity Relative to 50L All the reactor scales performed similarly with a slight decrease observed compared to 50L scale Continued development work is underway to increase small scale yields to match 50L scale 11 Critical parameters for a representative CsCl UC scale-down model ▪ To properly scale UC, the following factors must be considered: – TFF2 Load material concentration (analogous to tube loading) – K-factor (separation efficiency) – Spin time (must be adjusted if K -factors differ) ▪ Workflow – Identify smaller tubes (and a compatible rotor) – Calculate the r-min, r-max, and k-factor for the tube-rotor combination – Calculate the equivalent spin time to the at -scale process – Assess % full capsids after band pulling from each tube size Image modified from Beckman Website 12 Equivalent banding pattern observed across tube sizes Consistently pulling only the full band is more difficult in 4.2mL tube; so that size won’t be used for scale-down purposes 39mL Tube 13.5mL Tube 8.9mL Tube 4.2mL Tube 0 20 40 60 80 100 % Full Capsids 13 AFTX-201 scale-down process generates similar quality vector to clinical process GMP Run Scale Down 1 Scale Down 2 Scale Down 3 Capsid Purity (% VP1/2/3) 100 95.8 94.6 95.3 hcDNA (ng per 1e13) 30 18.4 33.1 41.1 HCP (ng/mL) BLOQ BLOQ BLOQ BLOQ Residual plasmid DNA (cp/mL) 3.04e11 6.83e11 7.98e11 9.65e11 Mass Photometry (% Full / Partial / Empty / HMW) 96 / 2 / 2 95.3 / 3.2 / 2.6 94.5 / 1.5 / 4 93.6 / 3.1 / 3 Aggregation (% Monomer) 99.5 99.74 99.61 99.73 In vitro Potency (% Reference Standard) 92.3% 94.1 72.4 80.3 14 Conclusions AFTX-201 clinical manufacturing ▪ A scalable, high-yield AAV manufacturing process was successfully developed and transferred to Forge Biologics for GMP production of AFTX-201 ▪ First GMP run achieved exceptional harvest titers (> 6e15 vg/L), exceeding typical industry benchmarks ▪ The 50L scale can enable supply for thousands of patients, with scalability available as demand grows Scale-down model development ▪ Representative scale-down models have been established for both bioreactor production and CsCl ultracentrifugation ▪ Bench-scale models show strong concordance with the 50L process for performance and key quality attributes ▪ These predictive models enable efficient process characterization and support continued clinical advancement 15 Acknowledgments ▪ Process Science Team – Matt Bennett – Paul Freeman – Ramin Kamran Sami – Rong Cong ▪ Analytical Science Team – Shahrzad Parker – Jordan Shufro – Hannah Czeladko – Lauren Sargent ▪ Rob May Our CDMO Partner: affiniatx.com Expanding the reach of gene therapies 43 Foundry Ave, Suite 120, Waltham, MA 02453 Confidential

The release as fetched from its publisher. affiniatx.com ↗