# ASGCT 2026 Recap — Georg Feichtinger **Source:** `ASGCT recap Georg.pptx` **Extracted:** 2026-08-06 19:12 **Slides:** 68 --- ## Slide 01 **Text:** 1423 ASGCT 2026 Parvotec relevant talks and posters --- ## Slide 02 **Text:** I. Chronic pain gene therapy ASGCT 2026 - Abstracts, talk slides and posters --- ## Slide 03 **Text:** 3 Relevant oral presentations Chronic pain gene therapy **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 04 **Text:** 4 SereNeuro Oral abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 05 **Text:** 5 SereNeuro (key slides attached) Oral abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 06 **Text:** 6 Exgenesis Bio (key slides attached) Oral abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 07 **Text:** 7 Sanofi (no poster pdfs available) Poster abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 08 **Text:** 8 Encoded (poster attached) Poster abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 09 **Text:** 9 Sangamo (poster attached) Poster abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 10 **Text:** 10 Sangamo (poster attached) Poster abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 11 **Text:** 11 Borea Therapeutics (poster attached) Poster abstracts **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 12 **Text:** Abstract 487 AAV-Mediated RNA Interference Targeting Nav1.7 Provides Durable Relief of Hyperalgesia in Preclinical Models of Chronic Pain Exgenesis Bio Inc, China (CJ Song) --- ## Slide 13 **Text:** 13 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] AAV-Mediated RNA Interference Targeting © Na,1.7 Provides Durable Relief of Hyperalgesia in Preclinical Models of Chronic Pain CJ Song A 3 : Ms, Exegenesis Bio Ve. “y (h » May 15%" 2026 hye | Ly | ee --- ## Slide 14 **Text:** 14 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] _» Workflow for EXG117 Development © EXEGENESIS BIO [| * >1,000 candidates designed using different In Silico Design algorithms and screened in silico a sgegassgsgens * >100 miRNA mimics screened in dual luciferase screening seeeselecsese system and HEK293 cells stably overexpressing hNa, 1.7 a Yity o® Spessipsisess * Screening of AAV capsid, miRNA scaffold, promoter, poly iC go’ Aetc. \ = \ * Efficacy: pain measurement in rat chronic pain model Rog yor? ) * Safety: behavior, histopathological assessment and off- ge SS =< targeting evaluation Or * Distribution, miR expression and Na, 1.7 knockdown My or a * Behavior, histopathological assessment and off-targeting 3 wl A\ evaluation Regt Le Confidential --- ## Slide 15 **Text:** 15 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] -» miRNA Mimics In Vitro Screening a) : . EXEGENESIS BIO In Vitro miRNA Screening 1.0 rs co DSRS o “miRacle” in vitro Screening Platform 5 iosener F : 2 os ong t pe Sl = Th * Invitro dual luciferase system was § ii aft used to screen miRNA mimics. 3 : ate Ss 00 a eke . 3 i * Candidates for human or rat Na,1.7 2 S have been tested and ranked based -0.5 on their knockdown efficiency. Human Rat Nav1.7 Nav1.7 Confidential --- ## Slide 16 **Text:** 16 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] EXG117 Alleviated Pain Hypersensitivity in Rat CFA Inflammatory Pain Model ®) DW EXEGENESIS BIO S wy) CFA Model Allodynia Mechanical Allodynia Heat & | 7 (von Frey) P (Hargreaves test) BS ae & 0.01 2 = <0.00 > 15 BS Dose er a ote iS + e000 | 3 | Im 5 | : £ i 2 ty eal} 5 3x 10" vg/rat G fe ay = +e Fal (1.x 10" vg/human) Edie 3 10+ § . 1x 10” vg/rat = 2 | be (3 x 10% vg/human) 2 . & | i 4x 10" vg/rat Ba ‘i =. (1.2 x 10! vg/human) = * = if | | ROA c a ; aL 10 (¥/-1) BL 9-1) a Days Post CFA Modeling Days Post CFA Modeling Mechanical Pain Alleviation Thermal Pain Alleviation EXG117 sign antly alleviated me anical and thermal allodynia in rat CFA inflammatorv pain model Confidential --- ## Slide 17 **Text:** 17 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] _» EXG117 Alleviated Pain Hypersensitivity in Rat SNI Neuropathic Pain Model eo he qPCR Na,1.7 mRNA Allodynia Mechanical SIN 20 EXEGENESIS BIC 1 ») = 64 Sham*saiine (Von Frey) SNI Model i ie Ne . 33 ro 05. § _ peont = ——————w 5 ». fee & + a ar Sr oy Nav1.7 expression increased in SNI rats ahermallPainiAllauiation Br Dose moore CER Alouynia Cold 3s sy | #2 sNiesaine a g (Cold Pain Test) > 1x 10" vg/rat Pete Goce J § °] ~ +t shamesatine = yeas Eo] = 62 sniesatine | 3 2 > G-3 SNIVEXG117, p<0.01 «00 ze > S gt — ROA se. ° 5? nt & Bodyweight Gain Improvement Mechanical Pain Alleviation EXGIAveienificantlyalleviatedimechanicallandanhenmallallodyniainlateniimedel cmaental --- ## Slide 18 **Text:** 18 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] EXG117 Alleviated Pain Hypersensitivity in Rat CIPN Model Chemotherapy-induced peripheral neuropathy (CIPN) EXEGENESIS BIO i FI ~_modeling, 1 — —=— ney WEY Ge Paclitaxel (PTX) PTX Ame/ke/dose; Administration «gg aman) angie) anaha every other day; 4 SSN doses in total Nay 1.7 mRNA in Rat DRG-L Allodynia Mechanical Allodynia Cold (von Frey) Cold Pain Test 3 «18 . EMD ams cayenne so px 7 sc4 as ) a = 1 = |2eGeeer qe | + escmnvexctt7 ter vanima 216 a | 7 CSP uERCHE tet von Ba = 2 3 ] B44 2 SE. 5 t 3 ae PS . “ Fs \ 4 E 10 B® in 3 10 \5 in = 08 7 - . Saline CIPN (PTX) BL Vani Wks WT Wko ° Swit viko whe wee Weeks Post EXGH17 Administration Woks Poot EXG117 Adriaan --- ## Slide 19 **Text:** 19 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] _» EXG117 Repressed Na, 1.7 by Over 80% in Rat DRG © 15 pesnoot D 4x108 se <0.001 2 tT B I 38 om past, 00. 1 Ean 5 Fy 23 5 2 E "Eg 0.010 if 3 5 2 3 Sos . aD & = 0.005 — E &s 2 : > 4x104 r = 0.000 0.0 Vehicle EXG117 2 Vehicle EXG117 Vehicle EXG117 EXG117 genome was highly - her} distributed in DRG miRNA was robustly expressed Na,1.7 mRNA was significantly repressed High biodistribution and miRNA expression achieved significant NaV1.7 knockdown in rats confidential --- ## Slide 20 **Text:** 20 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] -» EXG117 Was Well-Tolerated in Rats e < * No body weight difference in 4x10*? vg/rat EXG117-treated animals compared to execenesis aio naive controls. * No behavior abnormalities in EXG117-treated animals * No histopathological findings in DRG or spinal cord after IT administration of EXG117 i Vehicl EXG117 Body weight — = £ : 1 exe117 HD (4e12 vg/animal) | te oe Sets uth + Naive Control y Se DRG ac Se z Xx sl fs . ma . a Body Weight Comparable to Control No Histopathological adverse findings --- ## Slide 21 **Text:** 21 Abstract 487 – Exgenesis Bio Inc **OCR from screenshots:** [Screenshot OCR] EXG117 Demonstrated Strong Efficiency and Safety in NHPs © 2 yy» © Significant EXG117 miRNA expression, with a dose-dependent manner. EXEGENESIS BIO y * Na,1.7 mRNA was remarkably repressed in EXG117-treated animals. e EXG117 EXG117 miR Expression Nav1.7 mRNA Knockdown ctrl 3e13 vg 5el3v 5 oO = ais = Iie Bese: ea: Y.. sme: os. eo ee but absent in CG, > i oe i tee @ ss Ei ee. ensuring efficacy = Ae bee toee se ld (A p| Fa and safety. é <0 et A : ay | Effective Na,1.7 Knockdown in DRG No Na,1.7 Knockdown in Cardiac Ganglia Ontidential --- ## Slide 24 **Text:** Abstract 537 Human Sensory Neuron Profiling Enables KCNQ2 Gene Therapy for Neuropathic Pain SereNeuro, USA (Tea Soon Park) --- ## Slide 25 **Text:** 25 **OCR from screenshots:** [Screenshot OCR] 1. Human Genetics 2. iPSC Sensory Neurons 3. Pain Therapy ee 2 @ @ - ni p => ii i 2 4| \ @ ) People with Identify gene Undifferentiated iPSC-derived Target protective Lt pain conditions variants that ipsc sensory neurons genes / pathways Develop gene increase or (human pain therapies to reduce pain neurons in a dish) reduce pain Goal: Goal: Goal: Find natural “protective” Model how genes affect Create long-lasting therapies or pain-modulating genes human pain neuron activity that quiet overactive pain neurons Discover genes Test in human Translate to therapies 0) that influence pain => KK pain neurons (iPSC) > ay for chronic pain SereNeuro | SN102-a gene therapy for chronic pain 2 --- ## Slide 26 **Text:** 26 **OCR from screenshots:** [Screenshot OCR] TRIGEMINAL ux, | Ta¥estineuropsthie pain NEURALGIA » BN a cove —e J) leckel's cave acasiti | Ae | zs 4s 42 ‘SereNeuro | SN102-a gene therapy for chronic pain 3 --- ## Slide 27 **Text:** 27 **OCR from screenshots:** [Screenshot OCR] GFP tagged SCNYA and TRPV1 enable to enrich nociceptors A B (CG ui Ul I Ul H9 (Neg CTRL) TRPV1::GFP SgRNA Ly MRGPRX1 Noam 2 ee ‘ Pe tof +Cas9 nucle i g * oo >, ] S105. sa 3| & z Z a ——— (earmPARGRR PCR Pore FoRam SY) ge Fee > A S| s0RNA 3 7 oo | & SCN9A San AP 284 —ea_ aa — Stop 4Gns9 nuclesse EGFP EGFP vruady D a SCN9A::GFP MRGPRX1::GFP < —® TRPV1::GFP. F 240 TRPV1 Cae) Y, 3 —® MRGPRX1::GFP_ PAM oe a =< | ee E |e = g Stop €2s0 nuctoase % Fea * a ‘Gaacotch G 3 & 34 é Wa 33.7 m5, Aeleee 0 2 50 75 100 E » EGFP yp Cg Days of differentiation ax Sis Replate >= Gf F G ral ry-commit Neural Mature sensory - x praZies eee ee ee -e __e ___e-. -e 4 3° Pa) we p10) 0 2 4 6 8 12 14 18 22 70+ Day = igo S60. # ig 8 | Sed Geltrex PO/FN/LMN 3 ie ie we B® KSB Brainphys Brainphys x rs j Sa ra LDN-193189 LDN-193189 FGF2 NGF. NGF 8 e \ * obpgan cg 3 mol $B-431542 SB-431542 Y-27632. GDNF GDNF ° y y wig f Ha DAPT BONF AA Nez »5B ry; KRY ww 4 aaa NT-3 cAMP —— € € iy mt ee: cAMP --- ## Slide 28 **Text:** 28 **OCR from screenshots:** [Screenshot OCR] Resilience to Pain: A Peripheral Component Identified Using JNeu roscli Induced Pluripotent Stem Cells and Dynamic Clamp Ceol ee ‘Malgorzata A. Mis,!* ©Yang Yang,'°* Brian S. Tanaka,!? Carolina Gomis-Perez,'? Shujun Liu,'? 2019 ‘Fadia Dib-Hajj,:* Talia Adi,'? Rolando Garcia-Milian,’ Betsy R. Schulman, '? ©Sulayman D. Dib-Hajj,'* and ©Stephen G. Waxman’? 5 Erythromelalgia Ly Erythromelalgia (No KCNQ2 variant) ‘Same Family (With KCNQ2 variant) 1. KCNQ2 BASE EDITING ( - 7 . 7 Se ee |. blots peer ecb. § 4 we z J ; ee Base editing upregulates | Jo Ps . et AL Neuronal excitability - ae sone! Boo + Pain signals fire easily + Pain signals are quieter + More pain } + Less pain | pee i @ KCNQ2 variant may act as a natural “brake” on pain nerves, reducing nerve overactivity and lowering pain. ee _ * _ ‘SereNeuro | SN102-a gene therapy for chronic pain 6 --- ## Slide 29 **Text:** 29 **OCR from screenshots:** [Screenshot OCR] Hargreaves Von Frey PAMWT gi El are : : ai} GL Bart . : : ie I ._a> Se. &, .&@ > °. = C . } 4 d = (fa i=l PES em 2wks v A Se ® 4d lala l ys aye aS = in A ; “ | | | & i, A ° : e fel ACLT Install nociception- | Assessment of knee fe 150 surgery resilient osteoarthritis pain & 7 oe = mutation g =I 10) Ca =! eg {el iy BE] 53 wo. ey ea) ie fd EI a [3 "eS “Poe “SSS & & & & oe ra oe EEE Ss ° Rs —— SereNeuro | SN102-a gene therapy for chronic pain 8 --- ## Slide 30 **Text:** 30 **OCR from screenshots:** [Screenshot OCR] SN102 Base Editing in the spare nerve injury (SNI) Mouse Model 2 F F | \ \ , = -——— Nontargeting = A Ht am fe EL Te GG en 4 % 80: st snes ance och ry ‘ 5 | A | | 0 _/romecks \ , 3 mics SE” Ey Eh aE x S 5 <> & ; . 4 x x 2 _/r0meeks ©, & & & s s FS ee | ts i FX SF F ‘ Re « RS ase “tise 9 SN-KENQ2BE Electophysiology g Fs ° Intrathecal _Intraplantar BAMS-22 (Hind Paw) B Cervical Thoracic - o - H mitczhteaniton 15 250: oo * KCNQ2 BE $2004 | | B 100 _ (2 = A a E10. z Lumbar, Sacral 8 Y £ & 50. 3 \ a 3” $5 es $ \ i 3 2 50 =e 25 $ z + 2 E ° B ° 2 & co PS gO St © J © J 2 OO oP Pf PPPs oe 3 oe < eo PN Cia o = & < = < * < = WE) SereNeuro | SN102 -a gene therapy for chronic pain s 9 --- ## Slide 31 **Text:** 31 **OCR from screenshots:** [Screenshot OCR] In vitro In vivo behavioral test characterization Observe pain modulation LODMT efficacy test post effect on OA model o0@ 7 =a Target base-editing __._ | Differentiation args caee® L 5 gene iB. @: . &. ee Intra K ik6 i E See- SCNQA, TRPV1 A <7 a eee \s Ss: ‘=r hPSC-sensory AAV-PHPS = iPS neurons ca") . + iPSC-nociceptor provide tools to study pain gene therapy Spared Nerve Injury (SNI) pain model Trigeminal Neuralgia murine model * Proof of concept for SN102 efficacy = Sa () A A os) o Tigeminal * Cross-model validation XP Juan Reza, Myat Soe, Ricky Thu, Pua Dhanota, Steven Tan, Annie Tanenhaus, Archana Bell, Sirka Play, Suresh Pod, Stephani Taglatela. Encoded Therapeutics, South San Francisco, Calforia, USA INTRODUCTION NociPro Preserves Potency Comparable to Benchmark Promoters NociPro Drives DRG-Restricted Expression, Eliminates CNS Exposure, and Selectively Targets Nociceptors in Mice ** Chronic pain is driven by aberrant activity in nociceptive sensory neurons, and Na,1.7 "= NociPro (-3.3 kb in size) was derived from regulatory elements in the SCN10A genome '* Wild type C57BLIGN mice were dosed at PND1 with AAV9-NociPro-miRNAMS°¥*, (SCN9A) knockdown is a genetically validated approach for pain treatment (Figure 1)° locus and designed to hamess the nociceplor specificity of Nay1.8 (Figure 2A) ‘AAVS-hSyn1-miRINA=", or vehicle via sICV delivery. NociPro achieved: Smet nt Near ces, 0 CRMC ie ye Monin bag oe eS ‘+ NociPro-miRNA'S°" demonstrated dose-dependent knockdown of SCN9A in human = Comparable expression to hSynt in the DRG of sICV-dosed mice (Figure 3A) A rer aarers Ce eS ory nouns (Figure 28) ~ Notable exresson in cox Figure 34) 2\ « ee a tne + honiPots a pomotar dud om ragdsoy sloments wn he human SCNAOA weus, |", NOBPraetoved up 060% SCNDA knockdown, comparable to ubiutous (U8) and pane Robust target engagement in DRG esue, matching te potency of hSyn (Figure 38) fie son Bow Nr Pom oe eam AS oal ee ease ‘neuronal (hSyn) promoter contro (Figure 28) «Expression was envched in pain colovant neurons in ORG } j HI ce — te ~ Preferential expression in small-diameter nociceptors (Figure 4A) $3" ow GL ERS A. scntostoow @—— ye, ~ Potent and selective expression within Peripherin+ nociceptors, with Hi « u ‘Using a fixed, previously validated SCN9A-targeting miRNAPSN* NociPro-driven waco — — ‘minimal expression in NF200+ mechanosensory neurons (Figure 4B and 4C) i 33° bolenay ancl epaciicity were beachanerked agsinet wall-eeiebliehed promoters in — Distinct from broad neuronal expression profile of hSyn1 Som ae a ae Ee lm Sint Mouse in vivo models P Cat rowe en?) * — Human iPSC-derved neurons = a. awe Bo le = c a seas = [Sirsa Bs = + ) sera Z on z - — wes, OY UY i = i j i : NociPro Selectivity Conserved in Human Neuronal Models NociPro Shows Equivalent Potency to U6 in NHPs, with Improved Tolerability CONCLUSIONS * On-target activity and off-target sparing were confirmed in human iPSC-derived neurons *-‘NociPro-treated animals showed DRG-estricted expression with ‘+ NociPro demonstrated comparable target engagement to U6 with improved safety NociPro delivers selective, potent, and translatable nociceptor targeting with an improved therapeutic = Robust (63%) SCNOA knockdown in PSC-dorved sensory naurons (Figure SA) ‘minimal or no expression in SC, brain, or peripheral organs (Figure 6A) ~ Significant Nay1.7 protein knockdown, comparable to U6 (Figure 68) index, supporting its use A en gene tierapy. approaches ta chrorsc paln = No detectable activity in iPSC-derived cortical neurons? (Figure SB) ‘ap gewwey aay Improved safety profile relative to U6 candidate, with no adverse findings (Figure 6C) | & NociPro enables potent and selective transgene expression in DRG nociceptors, with activity comparable to ~ Maintained potney while improving specity v= UShMSyat ¥-) VE oe benchmark promoters hSynt and US ‘se Leer So © [setetyParameter [Nociro Tus * Itdemonstrates consistent DRG-restricted activity across species, with no detectable expression in non-target CNS . ‘s a) “econ — ae earn * It shows strong cell-type specificity in human sensory neurons, with enriched expression in nociceptors and minimal = = 1 Pa | * In NHPs, NociPro achieves robust Na, .7 protein reduction at clinically relevant doses with improved tolerability |. Sateen | Site [ceceman | iimmeemnanre | | 7 eels tn on tang ty at ea se i i | I i Lo oss = NociPro is well tolerated with no adverse findings observed fe i 4 owes \ _— i ee i i i ¢ | ov i | eax | “oma | See | = Advancement of AAVS-NociPro-miRNA'®°N* toward IND-enabling pharmacology and toxicology studies ov ese (adverse) * Potential for further optimization (e.g. size reduction for AAV payload compatibility) or use with diverse payloads “aie “ate Yypttiertag mam a # é Su Seether Sarre Sieces eiaremrcen momar Date eninge ms we ramet get aponeeeton bmtonea OG memo menecegeemet anata mates! Seiatatate arama Prasat ipsanucta sats ee arte’ meeeicu cect ae {dAPCR, normalized to nontergeting seramble (SCRM) contol. N=3 biclogcal replicates. Err bars are SEM. inca the it of quentaton ofthe respective assays. (B) Reaive total Ney 7 protein concentration for lumbosacral DRG regions. Nav 7 protein levels were sssessed by MSD sossy. Each be represents Sandan snore moa, eV semis: hssorbronnrcer TE ge mgngemars TSS reseipton sat ns WT wld ype ACKNOMLEDGEMENTS These ses nos --- ## Slide 34 **Text:** Abstract 1309 ST-503 Nonclinical Safety Studies Evaluating Zinc Finger Repressors Regulating the Expression of the Nav1.7 Gene for Treatment of Small Fiber Neuropathy Sangamo Therapeutics, USA --- ## Slide 35 **Text:** 35 **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 36 **Text:** Abstract 1033 A Modular Platform for Chemical Capsid Functionalization of AAV Vectors, Enabling Receptor-Guided Targeting and Programmable Payload Release for Pain Gene Therapy Borea Therapeutics, Italy --- ## Slide 37 **Text:** 37 **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 38 **Text:** II. Capsid engineering ASGCT 2026 - Abstracts, talk slides and posters --- ## Slide 39 **Text:** 39 Elements of our engine present at ASGCT Parvotec Capsid Discovery Engine Everyone R-Scan for Motoneurons (poster) ? **OCR from screenshots:** [Group Screenshot OCR] AAVI-X1 AAV9-X1.1 & K,=6+2uM K)=19+5 uM e PKD2-MBP_ e ——— OO ——— | c &. & @ é | FI 1 — a AAV 2 \ B10 substitution ———— =} Bio-AAVO Fab ge \ 10 aa: 22 £ | | \ > . SA @ as) Non-functional AAV particle + ITR maintained > Truncation not detected by ITR-based qPCR or ddPCR * Overestimation of functional titer [Screenshot OCR] * Introduction of mismatches between the passenger and guide strand TG UG TG UG Transcription Bulge remains, weakening RNA hairpin GT GU UG TG Non-canonical G:U base pairing stabilizes RNA hairpin Both options weaken the DNA hairpin and improve read-through during AAV genome replication [Screenshot OCR] Untreated Treated 7 ; =r VG ae. (=) oe mi405/miR871 mute Nett AAV aN Anns —> & a & sink > 1 ae * Smeaton —p TS Stable cell lines for potency assay D> i oe, Potency assay readout [Screenshot OCR] . . of, miR8714°2 improves % full and potency % te Truncated — © Full 2 100 as _ (o) 5 80 80 Ww ns 2 60 Z 6 F 40 & 40 5 20 20 E xo ° o v — Ny @ Ss ow os 2 x ao Yr § FS wr 3 & & & © & € & [Screenshot OCR] Reduction of miR-truncations Improved product purity Increase in potency Strong RNA hairpin —) Weak DNA hairpin --- ## Slide 41 **Text:** 41 Relevant oral presentations Capsid engineering **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 42 **Text:** 42 Relevant oral presentations Capsid engineering **OCR from screenshots:** [Screenshot OCR failed: cannot find loader for this WMF file] [Screenshot OCR failed: cannot find loader for this WMF file] --- ## Slide 43 *(no extractable content)* --- ## Slide 44 **Text:** 44 **OCR from screenshots:** [Screenshot OCR] In vitro and in vivo screening of AAVé clade variants 2) DNA extraction OHSU » » vilbt viib2 DNA > wi pLibo of ot 4, on, a extraction S 5 a J é sls wr 87 = vLibt = ald >. Sz/ 5 V a ‘ a 3 - 2» Human ane) #4 —— 3 \ aire> A Renal vein(RV) | & “7 privet [P islets 4 & So a 3 Renal pelvis (RP) | § pl cs PIGS eA blood etc. S © SF wwe ras S| 3 £ 3 G:) Mouse IV Ss ve FLO os S 2 >, 2 xD Pa o oy = vLib0.1 @, oO) .< > > <2 2 E y 1, = plibt vLibt & 2 aft &, 3 a) ef oD Brod > a> & He 3 2,9°7|2 sp sts > g Barcode library = (V E ° » 2 2 Pancreatic duct | @ —) & v <2, % \V pb) 5 Sp efe SY = E> >> a> E papred|= E —_—— blood etc: --- ## Slide 45 **Text:** 45 **OCR from screenshots:** [Screenshot OCR] Reprogramming AAVé capsids by known BBB-penetrant 2) peptides OHSU Loop VIII Peptide insertion [| _ReceptorFamily [SCS Peptide | LY6A | LY6C1 | CA-IV Reference A CS | gpos | + | Nonnenmacher M, et al. Molecular Therapy Methods & | opis | + | | | pHp.c2| [| + | | PHP.C3 [ [+ [J Ravindra kumar S, et al. Nat Methods. 2020 May;17(5) AAVF | | + | pes [| | Clinical Development, 2020; 20, 366-378 =| Chuapoco, M.R., et al. Nat. Nanotechnol. 18, 1241— . Cap-Mac] N/A | N/A | N/A |Chuspoco. MR cay Met Manoieennas 18,126 | --- ## Slide 46 *(no extractable content)* --- ## Slide 47 **Text:** 47 **OCR from screenshots:** [Screenshot OCR] The DB1 Family of AAV5-Based Capsids Is Engineered for Potent Brain Targeting in Primates The AAVS capsid contains a We delivered this massive library liver-tropic sialic acid directly to NHPs to uncover AAV binding loop capsids with novel tissue tropism : We removed and replaced - this loop with a library of Pe Beg ORE >1 billion sequence variants ae ee Rei. ‘<2 § eco W a NHP Brain DNA - ® ee; ie} Ea 7p) --- ## Slide 48 **Text:** 48 **OCR from screenshots:** [Screenshot OCR] How Does DB1 Cross the BBB? Co-IP LCMS for Receptor Identification + We looked for DB1 novel receptors by identifying proteins in macaque brain lysate that bind to the capsid + Mass-Spec analysis identified dozens of candidates, which we followed up on with a functional screen in vitro 4 rN * 4 4 Vix’ = 7 5 fa. > > > +e > FI + i ie « ‘ys Clarify Bind \*/ Wash es Isolate |t, tems ont _ . } jpn ot + Gas) =H) & 7p) Shape Therapeutics Inc. 2026 - Proprietary --- ## Slide 49 *(no extractable content)* --- ## Slide 50 **Text:** 50 **OCR from screenshots:** [Screenshot OCR] ’ oo oe AAVnerGene’s Solution: ATHENA Platform eee ATHENA-I ATHENA-II ATHENA-III AAVI. AAV2— AAV3B O99) eH4) € Zé AAVX AAVy — AAVz ees @25 ~cuomn €884 AAVm = AAVn AAV J ; * Known barcoded capsid library * Random peptide insertion library * Rational DNA shuffling library * Evaluate different capsids * Evolve novel capsids * Create novel AAVs * >1000 AAV capsids * High library complexity (> 1e9) * Combine the advantages of * 1 capsid = 3 barcodes * Any capsid, any VRs, any peptides selected capsids --- ## Slide 51 **Text:** 51 **OCR from screenshots:** [Screenshot OCR] o@ ATHENA: A Platform for Iterative AAV Evolution ©%® sss $< A AY NELGene s 1 were Yo cases I bey, ys") 82 9 Any y Cross-Species as tag AAV Templates y 10 Four Rounds of + Multi-rounds selection _ ATHENA-II 8 $3 + Machine Learning & . S «FE | MMII. cosa tncapass 3 Fvolution 3 ge Models with ATHENAI 00 2 82 og + Individual Testing fling S!9r oleran NA Tissue Tropism AAVs 3e8 (Starting with Best AAV Templates from ATHENA-| or -III) --- ## Slide 52 *(no extractable content)* --- ## Slide 53 **Text:** 53 **OCR from screenshots:** [Screenshot OCR] Rei ining viral deli starting with AAV Holistic datasets Al-Optimised Rapid data generation across . . tropism, pot id Full capsid mutation a) ©) imtoune-ovasio for every High impact mutations across designed variant the entire capsid surface, £) g ) giving us a complete picture f. Lir’s integrated 7, lab-in-the-loop ae Targeted Biological __ ee Endpoints nAAVigator® Al , A full-stack viral design engine enabelling rapid iteration and precision --- ## Slide 54 **Text:** 54 **OCR from screenshots:** [Screenshot OCR] iS nAAVigator® G La x ae Viral Design Sas S E Targeted i ASD) Engine a PLM development with Endpoints 7 z in vivo datasets Targeted library design mee Het. os 2 ¥ tox > Variant enrichment Ph gy ae & & Serene PLM training crenata . - High-throughput invivo Tissue processing Model fitting and engineered library biodistribution and NGS library redesign experiments --- ## Slide 55 **Text:** 55 **OCR from screenshots:** [Screenshot OCR] [..-) DNA Struct. VAS ue Models Models a Library Generators | Library Scoring Mm Quality Control Score-based Filtering AAV Variant Sampling Library --- ## Slide 56 *(no extractable content)* --- ## Slide 57 **Text:** 57 **OCR from screenshots:** [Screenshot OCR] iii WHITELAB Ag GENOMICS Ga [ 168: A multimodal, end-to-end rational design platform breaks native AAV tropism constraints Dylan SERILLON, PharmD, PhD Global Director of Drug Design 05.13.2026 --- ## Slide 58 **Text:** 58 **OCR from screenshots:** [Screenshot OCR] Whitelab Genomics’ Rational Guided Design Approach Oh WAITELDS About Whitelab Genomics Rational Guided Approach e Al specialist in genomic medicine, aw ee. making treatments more accessible and SY "eee safer for patients a In Silico Selection Peptide Design Vector Design Of high-potential targets Optimized for receptor Optimized for viability, identified via ML binding targeting and safety e Develops a proprietary Al platform combining computational and Impact structural biology with data science v Predicted Mechanism of Action Vv Improved results robustness (10K to 100K max. variants per library) v Reduced number of in vivo screening iterations e TheAl enables rapid design of optimized vectors, promoters, and payloads for = Proprietary patented platform: a unique rational approach combining machine learning, molecular design, and in vivo validation. specific cell and tissue targeting. 2026 - Property of WhiteLab Genomics. Any cstrbuton, reproduction or use without prior authorization i strictly prohibited UNLEASH THE POTENTIAL OF GENOMIC MEDICINE USING AL 3 --- ## Slide 59 **Text:** 59 **OCR from screenshots:** [Screenshot OCR] . . . . . . . ‘i WHITELAB Rational Guided Design Relies on Five Experimentally Validated Steps ‘" come 01 02 03 04 05 © aN Target Receptor Peptide Capsid Viral Functional Discovery Design Engineering Fitness Validation : TAT NO Sake, fe tT beg ett ee % bs ee 4 BPRORPLEEE EE Al-driven identification Al-driven design of Rational capsid design Evaluation of variant In vitro and in vivo of novel tissue targets targeting peptides with modelling strategy producibility specificity confirmation An Al-powered rational guided design platform - from biological target to optimized capsid in a single platform. a se --- ## Slide 60 **Text:** 60 **OCR from screenshots:** [Screenshot OCR] The Challenge of AAV Engineering a. WHITELAB e 1argetin EX 9 Conventional approaches suppose a systematic trade-off: optimizing one axis compromises the other two. Targeting Targeting vs vs Efficacy Production Be 8 a 63°38 By re-thinking these trade-offs, WLG 4 platform jointly optimizes all three 3 Efficacy dimensions simultaneously -) vs ey Production & SP or Ki % es --- ## Slide 61 **Text:** 61 **OCR from screenshots:** [Screenshot OCR] : . . Wy WHITELAB Step 1 - Target Receptor Discovery - A Rigorous Selection Funnel eo Human Candidate Receptors Kidney (multi-omic analysis) 30) Proximal tubule-specific receptors identified with confirmed human- NHP conservation FILTER 1 Cell-Type Specificity CNS Blood-brain barrier-crossing targets with neuronal cell-type selectivity across species FILTER 2 Cross-Species Conservation Eye Retinal cell-specific receptors ¥ Immuno - Oncology Validated Targets Tumor-selective receptor targeting a . Every target earned its place - studied for Human The pipeline already discovered new relevance, specificity and cross-species translatability, targets, experimentally validated not only expression. 2026 - Property of WhiteLab Genomics. Any distribution, reproduction or use without prior authorization is strictly prohibited. UNLEASH THE POTENTIAL OF GENOMIC MEDICINE USING Al --- ## Slide 62 **Text:** 62 **OCR from screenshots:** [Screenshot OCR] . 5 A A {fi WHITELAB Our Approach: Al-Driven Screening of peptides Oh WHITER Generative Virtual Iterative inchouse Reptige Ribena edvisualizatton Screening Docking SAR* 0 pacisters © Binders © Non-Binders ~1M+ candidates Structural Affinity % filtering optimization AC a 3 a & 4M+ Proprietary peptide database oo . * = £ yp ’ 5-15 a.a. Small peptides cS + Fj . a " , a . + Virtual screening trained on receptor-peptide interactions Vs es fy Y + Virtual docking pre-filters for structural feasibility e * Iterative SAR* loop between predictions and SPR validation ; . " A unified Al pipeline - molecular coverage, with nw : : predicted 3D structure, & biophysical relevance Proprietary peptides - binders vs. non-binders receptor-trained, project-agnostic “Structure Activity Relationship 2026 - Property of WhiteLab Genomics. Any distribution reproduction or use without prior authorization is strictly prohibited UNLEASH THE POTENTIAL OF GENOMIC MEDICINE USING Al --- ## Slide 63 **Text:** 63 **OCR from screenshots:** [Screenshot OCR] A : : . A ow A ‘i WHITELAB Step 3 & 4 - Capsid Engineering - Multi-Trait Optimization eo Production performances (10K-100K size libraries) Peptide Insertion Capsid Backbone Receptor targeting Stability - Assembly - Fitness > 90% Predicted capsid variants are produced > 80% Designed capsid variants with NGS RPM > 5 Al Multi-Trait Optimization . , Targeting * Efficacy x Production - simultaneously > 10% _ Variants with fitness > parental capsid Receptor Specificity § Capsid transduction | Fitness| sid Aen pes i. ~ Py sa v All three axes optimized jointly - no trade-off ors | fe Aras nee Pre + fis x Sa » WLG engineered rAAV . ,.| 4 Powe, || a capsids are SOK FRR aa 74] ibs = No more trade-offs - the capsid variants are significantly different ~~ ‘ =e “ as >i a 5 ae ti tabili a ducibilit fromthe WT parental...) 6 10 engineered for targeting, stal ility, and producibility capsid ae as be thy in a single optimization loop. 0.75 % Mes ¥, ¥ tas” 5 Ney rt Fr -1.00 oe 5 “100 075 -050 -025 000 025 050 075 100 em Spee rere is 0 Stee ks Casein recht ta everett cue dene cose tantra teniy chalet (ea. EMSH THE POTENTIAL OF GENDAEC MEENCRE Uma Al --- ## Slide 64 **Text:** 64 **OCR from screenshots:** [Screenshot OCR] A A H ‘i, WHITELAB Step 5 - Evaluation of the Mechanism of Action us sone Productivity Performances - ~10K Variants Library In vitro MoA WLG AAV2-derived library Oo Library Titer 1.8E13 vg/mL (1.5 mL) Oo VETO 84 % of variants robustly produced Receptor expressing cells Receptor - KO cells fails Nash l NGS reads per million (RPM) > 5 . ” 87 % of variants with good fitness + Binding selectivity Variant Fitness 7 a Finess > Value from negative control + Transduction selectivity Cross-packaging Not detected Composite Score - 5 Selection Criteria Applied Transduction Robustness Safety — ge Xo, +k eae aaa --- ## Slide 65 **Text:** 65 **OCR from screenshots:** [Screenshot OCR] i . i ; R . ‘i, WHITELAB In vivo Results: Brain Transduction & Liver Detargeting ee Bon Key Findings Project 2 - CNS Delivery - AAV9 BBB i 1) Robust BBB crossing and CNS . To ahead a CNS Benchmark transduction - 600 candidates successfully . a iS a penetrate the brain in vivo ° Uc 7 2) Top variants combine high CNS c . Ps transduction with complete liver 4 Brain Brain & Liver detargeting B , targeting only targeting = a 3} SPR _> in vivo predictivity confirmed: SPR 2 a selectivity translates to functional delivery 2 “7 oo AAV “7 P| a e 1 a Liver A “Ze ~ 600 rAAV with BBB penetration per Pe targeting (°° °7 ~15,000 tested after a single round of design Uc - -! (~4%) 0 = +—_______ 0 1 2 3 4 5 log10(liver + 1) 2026 - Property of WhiteL ab Genomics, Any distribution, reproduction or use without prior authorization is strictly prohibited UNLEASH THE POTENTIAL OF GENOMIC MEDICINE USING Al --- ## Slide 66 **Text:** 66 **OCR from screenshots:** [Screenshot OCR] H ‘i, WHITELAB Conclusions un sons VO C1) Rational Guided Design Improves Capsid Engineering A biology-first, Al-accelerated pipeline: from target identification to optimized capsid in a single, structured workflow {2 ) Rigorous Multi-Omic Target Discovery Receptor selection based on tissue specificity, low off-target expression, and cross-species translatability C3) Al-Driven Peptide Design Validated by SPR Generative models combined with biophysical validation deliver high-confidence binders before any in vivo experiment (4) Simultaneous Multi-Trait Capsid Optimization Tropism, fitness, and producibility engineered together: no trade-offs between targeting and manufacturability Cs) In vitro Validation Confirms Mechanism of Action ~30 validated hits from 10,000 candidates with a single screening iteration, outperforming conventional approaches eo In vivo Selectivity and Liver Detargeting Demonstrated Cross-program reproducibility across distinct receptors, libraries, and organ panels in preclinical models Rational guided design delivers validated, tissue-selective AAV capsids: from target to candidate in a single platform 2026 - Property of WhiteLab Genomics. Any distribution, reproduction or use without prior authorization is strictly prohibited. UNLEASH THE POTENTIAL OF GENOMIC MEDICINE USING Al --- ## Slide 67 **Text:** 67 **OCR from screenshots:** [Screenshot OCR] ; wi WHITELAB Protein & Vector Computational Translational External GENOMICS Engineering Biology Science Partners Ne GENOV om. eS Tomee 6076 CMAN ayo a eee [el ey are la eye el ats ey rer a =) mc S = K oss = é, . y Oa -S \ | S S Ck sal Ete “ate f \y ee 4 77 », an . we oh! ". tm Nec yy OS BG 8 olla ts [om Leading Academic and Industrial Partners Awards and Affiliations i] os at. Spine... SANOFI CD cytiva Que a 4 bivlead r image io r “a Universe r . Le SGENETHON ARTICO = TaR@el . aN Lo) Biotech sifted, _ CONTAC in OFFICES 4 mm, . i gx.c 2 Bayer Co.Lak 3900-1, Place Ville Marie — Ute TaDOX:¢ : = 8 Rue Jean Antoine de Baif 75013. 238 Main St, Cambridge, MA Montréal Québec H3B4M7, Soe a a Paris, France 02142 Canada i 20c5 JerpberivieriWritelen Genomics. Any dlatribution, reprodctionar Use; hou prior li Eaaalltam] - I or cenowic meoicie UIE --- ## Slide 68 **Text:** 68 ---