ASGCT 2025: Development of a Flexible High Yielding, High Performing Process for Manufacturing of AFTX-201, a Novel Investigational AAV Gene Therapy for Treatment of BAG3 Dilated Cardiomyopathy

Confidential Confidential Development of a Flexible High Yielding, High Performing Process for Manufacturing of AFTX-201, a Novel Investigational AAV Gene Therapy for Treatment of BAG3 Dilated Cardiomyopathy 2025 ASGCT Matt Edwards, VP Process Science Confidential BAG3 DCM is a devastating disease US 34,000EUR 37,000 Other developed markets, 5,000 No. of patients with BAG3 DCM in developed markets 1) DeGroot, S. (2024, August 16). Facebook; accessed October 2, 2024. 2) Sentero Pharma market research & forecast Disease biology and unmet medical need • Monogenic disease with haploinsufficiency • Diagnosed at a mean age of 37 years • 64% are in NYHA Class II – IV at presentation • Current treatments for symptomatic improvement • 22% with DCM require a heart transplant A significant market for gene therapy 76K patients in developed markets Increasing genotyping and diagnosis Forecasted peak annual WW sales >$2B2 2 Confidential 3 AFTX-201 reversed disease pathology in the BAG3 cKO mouse model: increased EF and improved dilated cardiomyopathy Affinia solution: AFTX-201 ▪ Restore BAG3 protein to normal level with a full-length, fully-human BAG3 transgene using Affinia’s novel cardiotropic capsid ▪ Reverse disease pathology: increase contractility, reduce dilation, increase ejection fraction and exercise capacity BAG3 protein Cardiac contractility Dilated heart Patient exercise capacity WT Vehicle AFTX-201 10 20 30 25 20 27 Fractional Shortening % ✱✱✱ ns ✱✱ BAG3 cKO Fractional shorteningEjection fraction WT Vehicle AFTX-201 30 40 50 60 50 41 54 Ejection Fraction % BAG3 cKO ✱✱✱ ns ✱✱ BAG3 cKO mouse model closely mimics the structural, functional, and molecular defects observed in patients affected by BAG3 DCM Confidential Affinia’s novel cardiotropic capsids are engineered to address the limitations of conventional capsids in heart disease 4 1) Internal analysis based on public information from AAV9-based clinical programs Left: PR-0025 cynos (n=3-4), CAG.GFP 3e13 vg/kg IV, day 28; heart LV % cardiomyocytes GFP+ Right: iPSC derived cardiomyocytes, 96-well assay, 50K cells per well, 72h incubation, in duplicate >30% AAV dose ≥ 1e14 vg/kg IV <30% AAV dose ≤ 6e13 vg/kg IV Dose not tolerated Dose tolerated % cardiomyocytes transduced In clinical trials, conventional capsids have been unable to transduce >30% cardiomyocyte at tolerated doses1 Affinia’s novel cardiotropic capsids have demonstrated superiority vs. conventional capsids in NHP and in human cardiomyocytes Target capsid profile AAV9 ATC-187 0 50 100 % GFP+ (NHP heart) Head-to-head capsid performance in NHP heart at 3e13 vg/kg IV 10 8 10 9 10 10 10 11 1000 10000 vg per well Avg GFP per cell 5000 2500 AAV9 ATC-187 Head-to-head capsid performance in human cardiomyocytes Confidential 5 Process development for AFTX-201 Goal was to develop a robust process for a novel capsid to support a Phase 1/2 clinical trial for BAG3 DCM ▪ High-Yielding – Conduct manufacturability assessment during capsid selection to ensure good productivity of novel capsid – Enable lower manufacturing costs ▪ High performing – Achieve low residuals and high % full capsids ▪ Stability – Conduct manufacturability assessment to ensure stability of novel capsids ▪ Flexibility – Demonstrate strong performance across multiple payloads and capsids AFTX-201 ATC-0187 AAV9 with peptide insertion Confidential 6 Process flow diagram for AFTX-201 Utilize Affinia's novel plasmid design to increase yields Explore new depth filters to increase capacity Explore multiple Affinity resins to ensure high performance Using CsCl gradient to ensure >90% full capsids Confidential 7 Use of Affinia plasmid design led to yields of >3e15 vg/L 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 0 10 20 30 40 50 60 70 Ratio Packaging 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 0 1×10 12 2×10 12 3×10 12 4×10 12 Ratio Harvest yield (vg/mL) We chose condition with highest yield to move forward since we were utilizing CsCl gradient for polishing Affinia plasmid design Condition Condition % full capsids Confidential 8 Harvest RC filters were optimal process choice – Higher loading capacity – Cheaper at scale – Reduced HCP and hcDNA levels Harvest RC filter better choice than Millipore DF + + Millipore plus STIC + Harvest RC plus STIC HCP Breakthrough hcDNA Breakthrough Note: Millipore filter was stopped due to rising pressure 0 20 40 60 80 100 0 2×10 6 4×10 6 6×10 6 Filter Loading (L/m2) HCP (ng/E+13vg) Harvest RC Millipore 0 50 100 150 200 1×10 0 1×10 1 1×10 2 1×10 3 1×10 4 1×10 5 1×10 6 Filter Loading (L/m2) hcDNA (pg/E+13vg) Millipore Harvest RC Confidential 9 AviPure Affinity resin were the optimal process choice – Similar binding capacity – Similar yield (>85%) – Reduced G3BP contaminant – Easily cleaned and reusable Avipure resin better choice than Poros AAV9 0 5×10 14 1×10 15 1.5×10 15 0 20 40 60 80 100 120 vp/mL loading % Breakthrough AviPure PorosAAV9 Purified with Poros AAV9 Purified with AviPure AAV9 G3BP contaminant Confidential 10 AFTX Lot 2 Forge Lot 2 Potency assay for AFTX-201 developed to help support process development Protein extraction Protein expression- ELISA Cell plating Viral serial dilution RS TA Infection Incubation at 37 ◦C Confidential 11 Process for AFTX-201 was successfully tech transferred to Forge Biologics AFTX (Hyperforma) Forge (Pall) 0 1×10 15 2×10 15 3×10 15 4×10 15 5×10 15 vg/L at harvest AFTX Lot 1 AFTX Lot 2 Forge Lot 1 Forge Lot 2 hcDNA 5.17e4 pg / 1e13 Pending 9.13e3 pg / 1e13 2.9e4 pg / 1e13 HCP BLOQ BLOQ BLOQ BLOQ % Full / Partial / Empty by AUC 95 / 4 / 1 94 / 4 / 2 95 / 3 / 1 96 / 2 / 2 Aggregation 99.3 97.6 90.73 96.56 rcAAV BLOQ BLOQ BLOQ BLOQ Relative potency ND Passed ND Passed Confidential 12 Conclusions ▪ Developed a robust process for a novel capsid that achieves high yield and quality in support of Phase 1/2 clinical trials ▪ Attain yields of >4e15 vg/L at scale with the AFTX-201 process ▪ Ensure excellent critical quality attributes (CQAs) ▪ Reduce manufacturing costs 50L Bioreactor ~2e17 total vg Drug Product ~5e16 total vg ~60 doses 8e14 total vg (80kg @1e13vg/kg) ~30K/dose Confidential 13 Acknowledgments ▪ Process Science Team – Esther Aribilola – Matt Bennett – Patrick Bresnahan – Rong Cong – Paul Freeman – Ramin Kamran Sami – Mike White ▪ Rob May ▪ Analytical Science Team – Hannah Czeladko – Spencer Fielder – Graham Lilley – Shahrzad Parker – Lauren Sargent – Jordan Shufro ▪ Upstream Team – Danielle Sexton – Aimee Doiron – Caitlin Austin ▪ Downstream Team – Corben Davis – Nicole Hoefer – Mreedul Koirala – Benjamini Stillions ▪ Potency Assay Team – Devon Blake – Julie Selich-Anderson

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