Preclinical and clinical investigation of VES001, an oral small molecule sortilin receptor inhibitor for the treatment of frontotemporal dementia: a first-in-human randomized placebo-controlled phase 1 trial Authors: Louise Klem, Daniel Dumas, Imogen J Swift, Sofie L Frandsen, Ditte Nedergaard Mikkelsen, Laerke Fensbaek, Yalçin Yavuz, Okke Merton, Jesminne Castricum, Simone Prichardt, Graham M Wynne, Anders Riisager, Nina Schultz, Paul B Little, Ove Pedersen, Eva Thijssen, Anders Nykjaer, Philip Kremer, Mads Kjolby Published: 2026 ABSTRACT AbstractGRN haploinsufficiency reduces progranulin and causes frontotemporal dementia (FTD-GRN). VES001, an oral small-molecule sortilin inhibitor, increases progranulin without downregulating sortilin. Preclinically, VES001 bound sortilin selectively and elevated progranulin in plasma, cerebrospinal fluid (CSF), and hippocampal interstitial fluid of male rats. In a randomized, double-blind, placebo-controlled first-in-human phase 1 A study, 79 healthy adults (63/37% male/female; mean age 28 years) received VES001 or placebo as single ascending doses (10–990 mg) or multiple ascending doses for 7 days (180 mg once or twice daily, or 450 mg twice daily). Safety and tolerability, the primary endpoints, were met: no deaths, serious adverse events or treatment-related discontinuations; headache and fatigue were mild and not dose-related. Pharmacokinetics were dose-proportional, with blood–brain barrier penetration. Prespecified pharmacodynamic endpoints were met as single dosing increased plasma progranulin, and twice-daily dosing produced greater, sustained increases versus placebo in plasma (peak 49% change from baseline) and CSF (18% change from baseline). Although cohorts were small, dosing lasted 7 days, and participants were healthy, these findings establish pharmacological proof of mechanism for oral sortilin inhibition and support further development of VES001 as a potential disease-modifying therapy for FTD-GRN. EU CTR 2023-505847-38-00; ClinicalTrials.gov NCT06226064.Subject terms: Dementia, Drug development, Dementia Introduction Frontotemporal dementia (FTD) is a progressive neurodegenerative disease marked by atrophy of the frontal and temporal lobes and is the second most common cause of early-onset dementia in individuals under 65 years, after Alzheimer’s disease1. Epidemiological studies indicate that FTD accounts for approximately 10% of dementia cases in individuals under 65 years of age, with a point prevalence of 15–22 per 100,000 in the 45–64-year age group2–5. Clinically, FTD manifests as progressive behavioral and/or language dysfunction, profoundly impairing quality of life1. Current treatments provide only symptomatic relief and do not target the underlying molecular pathology, underscoring an urgent need for disease-modifying therapies. A strong genetic component underlies FTD, with 30–50% of FTD cases being familial4,6,7. Approximately 80% of hereditary cases are linked to mutations in 3 genes: chromosome 9 open reading frame 72, microtubule-associated protein tau, and the progranulin (PGRN) encoding gene (GRN)4,6,7. GRN loss-of-function mutations are found in 5–20% of familial FTD cases and 1–5% of sporadic FTD cases5. These heterozygous mutations result in an approximately 50% reduction in PGRN levels and lead to the development of FTD with complete penetrance (FTD-GRN)2,8. Therapeutic strategies that restore PGRN to physiological levels are therefore considered a rational and potentially disease-modifying approach for FTD-GRN9–12. PGRN elevation consistently suppresses inflammation and attenuates neurodegeneration across both PGRN-deficient and non-PGRN-deficient models8,13–24. PGRN interacts with several binding partners, including sortilin, a type I transmembrane receptor in neurons, as well as prosaposin, which can facilitate binding and internalization via prosaposin-specific receptors, such as LRP125. Sortilin and LRP1 mediate endocytosis of PGRN and route it to the lysosomal compartment, where it is proteolytically processed by cathepsin L and related cysteine proteases into a series of stable ~6 kDa granulin peptides (GRNs)25–27. While sortilin-mediated uptake was initially characterized as a primary clearance mechanism for PGRN, an extensive and growing body of evidence now supports the view that PGRN -and particularly its GRN peptide products, possibly acting as multi-GRN complexes - plays active and critical roles within the lysosome. Specifically, PGRN loss-of-function impairs lysosomal acidification and dysregulates cathepsin D maturation28, disrupts prosaposin processing leading to reduced glucocerebrosidase (GCase) activity29, alters levels of the anionic endolysosomal phospholipid bis(monoacylglycero)phosphate (BMP)30, and has been linked to dysregulation of legumain and other lysosomal hydrolases31. This alternative conceptual framework - in which PGRN serves as a precursor whose GRN peptide cleavage products regulate lysosomal homeostasis - is now supported by substantial preclinical data and is relevant to the interpretation of PGRN-targeted therapeutic strategies. Sortilin inhibition increases PGRN levels, and the anti-inflammatory and neurotrophic effects of PGRN are independent of sortilin23,32–35. Importantly, sortilin itself also exerts PGRN-independent pro-inflammatory effects and facilitates apoptosis through the cell surface complex with the P75 neurotrophin receptor (P75NTR)36–40. Different strategies are being developed to elevate PGRN levels in vivo, including sortilin inhibition, delivery of modified or recombinant PGRN, and gene therapies designed to enhance GRN expression. Among these, sortilin inhibition using a small molecule binding to the neurotensin interaction site 1 (NTIS1) represents a testable approach. It may both restore PGRN levels and directly counteract sortilin-mediated inflammatory and apoptotic signaling, without interfering with other important functions of sortilin, such as Trk receptor transport41. This contrasts with other sortilin-targeting strategies, as an antibody (latozinemab) is known to downregulate sortilin and may impact functions of the receptor41,42. Here, we report on the pre-clinical development of VES001 and the subsequent randomized and placebo-controlled first-in-human single and multiple ascending dose studies in healthy volunteers, describing pharmacokinetics (PK) and the pharmacodynamic (PD) changes in PGRN levels. VES001 is an orally delivered small molecule that binds to the NTIS1 and inhibits the protein-protein interaction between sortilin and PGRN. Furthermore, VES001 does not impact sortilin levels and has a limited impact on other functions of sortilin that do not involve the NTIS1. Importantly, it is also effective at crossing the blood-brain barrier, resulting in a favorable CNS exposure profile. Preclinical development experiments included evaluating in vitro binding to sortilin and other members of the receptor family39, and an in vitro sortilin internalization/downregulation assay. Further preclinical evaluation included a rat microdialysis experiment, assessing PK/PD in plasma, cerebrospinal fluid (CSF), and hippocampal interstitial fluid (Hpc ISF). These preclinical studies led to the phase 1 A clinical trial, designed to assess safety, tolerability, PK, and PD of both single and multiple ascending dose administrations in healthy volunteers. Results Preclinical: VES001 selectively inhibits human sortilin without receptor downregulation in vitro Grating-coupled interferometry analyses confirmed that VES001 binds human sortilin with high affinity (mean ± standard deviation (s.d.), Kon = 1.55*106 ± 4.88*105 M−1 s−1; Koff = 8.35*10−2 ± 2.13*10−2 s⁻¹; KD = 55.1 ± 4.98 nM). VES001 also binds rat sortilin with similar high affinity (mean ± standard deviation (s.d.), Kon = 1.19*106 ± 5.05*105 M−1 s−1; Koff = 6.46*10−2 ± 3.44*10−3 s⁻¹; KD = 60.6 ± 15.4 nM). No measurable binding to other members of the Vps10p family was detected (Supplementary Fig. 1)39. Furthermore, VES001 showed no relevant inhibition or stimulation across a SafetyScreen 87 pharmacology panel and a panel of 8 cytochrome P450 enzymes. In flow-cytometry assays, VES001 potently inhibited the interaction between human PGRN and sortilin in a HEK293-FL cell line, with an IC₅₀ (mean ± s.d.) of 4.424 ± 4.498 µM and a Kᵢ of 2.121 ± 2.106 µM. The mean and s.d. for the unbound fraction in human plasma was 10.3 ± 0.6% (recovery mean 92.3%). Notably, the tested concentrations of VES001 (3–9000 ng mL⁻¹) had no detectable effect on cell-surface or total sortilin protein levels in COLO792 cells, whereas a latozinemab biosimilar reduced both by 50–60% (n = 6) (Fig. 1A, B). These findings demonstrate that VES001 acts through functional competitive inhibition rather than receptor down-regulation. Overview of VES001 pre-clinical data. Results from the flow cytometry sortilin downregulation assay detecting changes in cell surface (A) and total (B) sortilin protein levels, compared to a sortilin-targeting antibody (latozinemab biosimilar). C–F Results from a male rat microdialysis PK/PD experiment. C Overview of the treatment and sampling. D Plasma, CSF, and HPC ISF compound levels (ng/mL). E Absolute changes in PGRN from baseline (set to 0%) in CSF and HPC ISF of the left Y-axis and in plasma on the right Y-axis. F Percentage changes in plasma, CSF, and HPC ISF PGRN from baseline (set to 0%). All data are reported as the mean and standard errors of the mean (SEM). n is listed directly in each graph. Source data are provided as a Source Data file. Created in BioRender. Bio, V. (2026) https://BioRender.com/g742mp3. Preclinical: VES001 penetrates the BBB and elevates PGRN in vivo Following repeated oral administration in male rats (n = 7), VES001 was detected in Hpc ISF, CSF, and plasma with stable PK profiles (Fig. 1C, D). Peak mean concentrations were 815 ng mL⁻¹ at 2 h in Hpc ISF, 2175 ng mL⁻¹ at 4 h in CSF, and 29571 ng mL⁻¹ at 4 h in plasma, with Hpc ISF and CSF levels approximately 10-fold lower than plasma, when not adjusting for protein binding. VES001 produced clear, time-dependent increases in PGRN in all compartments, separating from baseline by 1 h (plasma), 3 h (CSF), and 5 h (Hpc ISF) after the first dose (Fig. 1E, F). Maximal mean responses occurred at 9 h in Hpc ISF (6.12 ng mL−¹; 144% increase), 7 h in CSF (23.91 ng mL⁻¹; 93% increase), and 10 h in plasma (981.67 ng mL⁻¹; 114% increase). Vehicle-treated male rats showed no significant change (n = 5). These data confirm BBB penetration and dose-dependent elevation of PGRN following systemic VES001 administration in a pre-clinical model. Clinical: participant disposition and baseline characteristics A total of 48 healthy adults were enrolled in the first-in-human single-ascending-dose (SAD) study and completed dosing and follow-up (Fig. 2A). In the multiple-ascending-dose (MAD) study, 31 participants completed dosing and follow-up (Fig. 2B). One participant in the MAD cohort 2 (2 × 180 mg) discontinued PK sampling for personal reasons unrelated to study conduct and was replaced. Baseline demographics were comparable across cohorts (Table 1). The study design is shown in Fig. 3. CONSORT diagrams of the single-ascending-dose (SAD) and multiple-ascending-dose (MAD) studies. CONSORT diagrams showing participant flow through the SAD (A) and MAD (B) phases of the study. All participants completed dosing and safety follow-up. † In cohort 2, one participant missed one intermediate assessment day between the final dose and the last follow-up visit due to personal reasons unrelated to study conduct or VES001. Their PD and safety data (except for the missed assessments) were analyzed, but their PK data were unusable due to technical issues. This participant was replaced, resulting in nine participants receiving VES001 in this cohort. ‡ In cohort 3, one participant declined the post-dose lumbar puncture; all other PK, PD, and safety assessments were performed and included in the analyses. Created in BioRender. Bio, V. (2026) https://BioRender.com/9rkef6i. Table 1Demography of phase 1A participantsSAD cohort (n = 48)MAD cohort (n = 31)Sex, n (%)Female18 (37.5%)11 (35.5%)Male30 (62.5%)20 (64.5%)Race, n (%)Black or African American0 (0%)1 (3.2%)Mixed2 (4.2%)1 (3.2%)White46 (95.8%)29 (93.5%)Age, yearsMean ± s.d.28.2 ± 9.928.2 ± 8.7Median2425Min–Max19–5419–50BMI (kg m−2)Mean ± s.d.23.3 ± 2.6123.1 ± 2.78Median23.122.1Min–Max18.8–30.418.0–29.3BMI body mass index, MAD multiple ascending dose, SAD single ascending dose, s.d. standard deviation. Demography of phase 1A participants BMI body mass index, MAD multiple ascending dose, SAD single ascending dose, s.d. standard deviation. Schematic of study design. A Overview of the single ascending dose (SAD) and multiple ascending dose (MAD) study cohorts, including the number of participants (N) per treatment group and placebo. B Overview of the study process, including the screening, randomization (R), dosing, follow-ups, and washout for both the SAD and MAD studies. C Sampling times and timeline illustration of the plasma (open circle) and cerebrospinal fluid (CSF) samples (cross) drawn for the SAD and MAD studies. † In cohort 2, one participant missed one intermediate assessment day between the final dose and the last follow-up visit due to personal reasons unrelated to study conduct or VES001. Their PD and safety data (except for the missed assessments) were analyzed, but their PK data were unusable due to technical issues. This participant was replaced, resulting in 9 participants receiving VES001 in this cohort. ‡ In cohort 3, one participant declined the post-dose lumbar puncture; all other PK, PD, and safety assessments were performed and included in the analyses. Created in BioRender. Bio, V. (2026) https://BioRender.com/ite8d6r. Clinical: safety and tolerability In this study, VES001 was generally safe and well-tolerated over a short treatment period in healthy volunteers, with no deaths, serious adverse events, or discontinuations due to treatment-emergent adverse events. Adverse events were reported in 51.9% of SAD and 80.6% of MAD participants, most commonly mild headache and fatigue (Supplementary Tables 1 and 4). In the SAD study, most events were unrelated to treatment (23/28 events); in the MAD study, approximately 1/3 were unrelated (27/93 events). The incidence of headache (VES001: 22.2–62.5% vs. placebo 16.7%) and fatigue (VES001: 22.2–50.0% vs. placebo 16.7%) showed no dose-related trend. Clinical laboratory values, electrocardiograms (ECGs), and vital signs revealed no clinically meaningful differences between VES001 and placebo. Overall, these findings are directionally reassuring and support acceptable short-term tolerability across the tested dose range; however, the small study population and limited exposure duration preclude assessment of rare or delayed toxicities. Clinical: single-ascending-dose (SAD) study, pharmacokinetics Single oral doses of VES001 (10–990 mg) produced dose-proportional increases in plasma exposure (Fig. 4A and Supplementary Table 2). Higher doses remained for longer durations above the GCI KD value, representing 50% sortilin receptor inhibition, with over 24 h after 810 mg and 990 mg, ~12 h after 270 mg, ~6 h for 90 mg, ~4 h for 30 mg, while the 10 mg dose did not reach the KD. The mean terminal half-life ranged from 7.2 to 9.5 h for doses ≥30 mg, and individual concentration-time curves showed a biphasic decline (Supplementary Fig. 3). Under fed conditions, the median Tmax was 4 h compared with 1.5 h in the fasted state, whereas the overall exposure (AUC) was unchanged (Fig. 4B). VES001 was measurable in CSF 7 h post-dose in the 810 and 990 mg cohorts, yielding mean ± s.d. CSF/plasma ratios of 0.17 ± 0.05 and 0.16 ± 0.03, respectively (Fig. 4E). These data indicate dose-linear pharmacokinetics and blood–brain barrier penetration following single-dose administration. Ph1A VES001 SAD plasma and CSF PK/PD. Plasma was sampled repeatedly throughout the experiment, while CSF was only sampled at –17 h (pre-dose) and at 7 h (post-dose). A Plasma total compound levels (ng/mL) for the fasted cohorts. B Plasma total compound levels (ng/mL) and areas under the curve (AUC) (ng*h/mL) for the fed and fasted cohorts receiving 90 mg VES001. C Absolute changes in plasma PGRN (ng/mL) for the fasted cohorts from the pre-dose baseline. D Percentage change in plasma PGRN from the pre-dose baseline (set to 0%) for the fasted cohorts. E Compound levels in CSF (ng/mL) for the fasted cohorts. F Absolute change in CSF PGRN (ng/mL) from the pre-dose baseline for the fasted cohorts. G Percentage change in CSF PGRN from baseline (set to 0%) for the fasted cohorts. The graphs depict the mean and the standard errors of the mean (SEM). A black line in plasma PK graphs denotes the GCI KD value, representing 50% sortilin receptor inhibition. A black line in plasma %CFB graphs denotes the baseline. n = 6 (4M and 2F; 5M and 1F; 2M and 4F; 5M and 1F; 3M and 3F; 3M and 3F, respectively) for each VES001 cohort, n = 12 for the plasma placebo cohort (7M and 5F), and n = 4 (2M and 2F) for the CSF placebo cohort. Source data are provided as a Source Data file. Clinical: single-ascending-dose (SAD) study, pharmacodynamics Single oral doses of VES001 produced dose-dependent but variable increases in plasma PGRN concentrations (Fig. 4C, D and Supplementary Table 3). No significant differences from placebo were observed at doses up to 270 mg, whereas the 810 mg and 990 mg cohorts exhibited statistically significant elevations based on nominal p values relative to baseline and placebo (810 mg: 19.5% [95% CI 5.6–35.4%], p < 0.01; 990 mg: 18.8% [95% CI 4.4–35.2%], p = 0.010). Increases relative to baseline were evident within the first dosing day and persisted over 48 h, with peaks obtained at 48 h for 810 mg (35%) and at 24 h for 990 mg (27%). Mixed-model analysis identified effects consistent with contributions of baseline, treatment, and time (all p < 0.0001–0.01) with a significant treatment × time interaction (p < 0.0001), consistent with dose- and time-dependent modulation of PGRN. CSF PGRN concentrations showed no clear differences between VES001 and placebo in the 810 mg and 990 mg cohorts (Fig. 4F, G). In this compartment, baseline levels were the only significant predictor of post-dose concentrations (p < 0.0001). These data suggest that single dosing increases plasma PGRN at higher doses but is insufficient to drive robust CSF changes. Multiple dosing is required for stronger, more durable pharmacodynamic effects. Clinical: multiple-ascending-dose (MAD) study, pharmacokinetics Repeated oral dosing for 7 days with VES001 (1 × 180 mg, 2 × 180 mg, or 2 × 450 mg) produced predictable, dose-proportional plasma pharmacokinetics (Fig. 5A and Supplementary Table 5). The pre-dose values for the twice daily dosing groups remained above the GCI KD value, suggesting continuous >50% inhibition of the sortilin receptor, while the once daily dosing group was only above the KD for the first 12 h post-dose. Mean steady-state concentrations were achieved by Day 3 for the 1 × 180 mg cohort and by Day 2 for the 2 × 180 mg and 2 × 450 mg cohorts, with accumulation ratios of 1.0, 1.26, and 1.19, respectively, consistent with minimal accumulation. The concentration-time profiles had similar shapes on Days 1 and 7 across regimens (Fig. 5A). Ph1A VES001 MAD plasma and CSF PK/PD. Plasma was sampled repeatedly throughout the experiment, while CSF was only sampled at –18 h (pre-dose) and at 150.2 h (post-dose). A Plasma total compound levels (ng/mL). B Absolute changes in plasma PGRN (ng/mL). C Percentage change in plasma PGRN from the pre-dose baseline (set to 0%). D Compound levels in CSF (ng/mL). E Absolute change in CSF PGRN (ng/mL) from the pre-dose baseline. F Percentage change in CSF PGRN from the pre-dose baseline (set to 0%). The graphs depict the mean and the standard errors of the mean (SEM). A black line in plasma PK graphs denotes the GCI KD value, representing 50% sortilin receptor inhibition. A black line in plasma %CFB graphs denotes the baseline. In all graphs, the placebo cohort has n = 6 (n = 2M and n = 4F) and the 1 × 180 mg cohort has n = 8 (n = 4M and n = 4F). The 2 × 180 mg cohort has n = 8 (n = 6M and n = 2F) for PK and n = 9 for PD. The 2 × 450 mg cohort has n = 8 (n = 7M and n = 1F) for plasma, and n = 7 for CSF. Source data are provided as a Source Data file. At 150 h after the first dose, mean ± s.d. CSF/plasma ratios were 0.11 ± 0.06, 0.19 ± 0.05, and 0.17 ± 0.09 for the 1 × 180, 2 × 180, and 2 × 450 regimens, respectively (Fig. 5D). Semi-quantitative profiling identified 13 plasma and 15 urinary metabolites; together, the metabolites accounted for <6.6% of total drug-related exposure. Together, these findings confirm dose-proportional exposure, CNS penetration, and limited metabolite formation with repeated dosing. Clinical: multiple-ascending-dose (MAD) study, pharmacodynamics Consistent with the expectation from the SAD study, repeated twice-daily dosing of VES001 induced sustained, dose-dependent increases in PGRN in both plasma and CSF (Fig. 5B, C, E, F and Supplementary Table 6). In plasma, PGRN was increased in the 2 × 180 mg and 2 × 450 mg cohorts relative to placebo (2 × 180 mg: 30.2% [95% CI 15.9–46.2%], p < 0.0001; 2 × 450 mg: 29.9% [95% CI 16.4–45.0%], p < 0.0001), whereas the 1 × 180 mg regimen produced no significant change. Mixed-model analysis identified effects of baseline, treatment, and time, and a treatment × time interaction (all p < 0.0001), indicating a consistent pharmacodynamic effect across repeated dosing. Plasma PGRN concentrations in the 2 × 180 mg and 2 × 450 mg cohorts increased early during the first dosing day (starting 6 h post-dose) and continued to rise with repeated administration, remaining elevated for up to 48 h after the final dosing day (Day 7). The highest percentage increases were reached by Day 7 (2 × 180 mg: 25%; 2 × 450 mg: 49%), indicating that pharmacodynamic effects may not yet have reached steady state (Fig. 5C, D and Supplementary Table 6). In CSF, PGRN concentrations showed consistent increases in the 2 × 180 mg and 2 × 450 mg cohorts compared with placebo (2 × 180 mg: 17.2% [95% CI 7.0–28.5%], p < 0.01; 2 × 450 mg: 17.7% [95% CI 7.4–29.0%], p < 0.01), whereas the 1 × 180 mg regimen showed no significant effect. The magnitude and persistence of these CSF effects (2 × 180 mg: 15%; 2 × 450 mg; 18%) exceeded those seen in the SAD study, demonstrating that multiple dosing is required for a more substantial pharmacodynamic response (Fig. 5E, F and Supplementary Table 6). Discussion The rationale for PGRN normalization in FTD-GRN is supported by converging genetic and biomarker evidence. GRN-haploinsufficiency causes FTD-GRN, and reduced PGRN is directly linked to neuroinflammation, neurodegeneration, and impaired lysosome function8,17,43. Drugs that reliably restore intra- and extracellular PGRN levels could therefore be a promising approach to disease modification. Several therapeutic strategies aimed at restoring PGRN levels are under clinical evaluation for FTD-GRN. These include AAV-mediated GRN gene-replacement approaches, such as PR006/LY3884961, PBFT02, and AVB-101; the anti-sortilin monoclonal antibody latozinemab; and recombinant PGRN replacement using blood–brain-barrier transport technology, as exemplified by DNL593. Latozinemab is administered intravenously and increases PGRN levels by targeting sortilin-mediated clearance, although the Phase 3 INFRONT-3 trial did not meet its primary clinical endpoint despite evidence of PGRN elevation. AAV-based approaches are intended to provide durable GRN expression after a single CNS-directed administration, but involve invasive delivery and are not readily reversible or dose-titratable. DNL593 is an intravenously administered recombinant PGRN fused to a transferrin-receptor-binding transport vehicle to enhance CNS exposure. VES001 differs from these biologic and gene-replacement approaches as an orally available small molecule that competitively blocks the PGRN–sortilin interaction at the neurotensin/NTIS1-binding pocket, thereby increasing extracellular PGRN without reducing sortilin levels. These differences relate to modality, route of administration, reversibility, dose titratability, and mechanism of action. This first-in-human study demonstrates that VES001, a novel and selective small molecule competitive inhibitor of sortilin, is safe and well-tolerated in healthy adults and produces dose-dependent increases in extracellular PGRN levels in plasma and CSF. These effects on PGRN levels are consistent with preclinical findings showing potent, specific binding of VES001 to human sortilin and inhibition of the PGRN-sortilin interaction. No non-specific activity was detected against other tested targets, which included a wide panel of targets and members of the Vps10p family39. Notably, VES001 does not alter cell surface or total sortilin protein levels, while antibody-based approaches are known to cause receptor downregulation42. Preclinical studies in wild-type cells and rodents—selected to match the physiology of healthy subjects—confirmed effective BBB penetration for the compound, resulting in downstream readouts confirming target engagement, including increased Hpc ISF PGRN concentrations measured by in vivo microdialysis. In the latter study, PGRN levels increased over time in all three compartments with steady levels of VES001. Together, these data provide a strong mechanistic foundation for targeting sortilin in FTD-GRN, a condition characterized by genetically driven PGRN deficiency. In both the SAD and MAD studies, VES001 was well tolerated at all tested doses, with no serious adverse events or dose-limiting toxicities. The most common treatment-emergent adverse events (headache and fatigue) were mild and not dose-dependent. In the SAD study, a dose-dependent increase in plasma PGRN was observed, although the magnitude of response varied across cohorts, possibly reflecting inter-individual differences in anthropometrics and smaller group sizes relative to the MAD study. The dose-dependent effects aligned with long durations above the GCI KD value for the two highest doses, with moderate durations for the two middle doses, and with limited durations for the two lower doses. The findings in the SAD study suggest that repeated dosing would produce larger and more sustained PGRN elevations, a prediction confirmed in the MAD study. Specifically, twice daily dosing with 270 mg or higher was predicted to sustain 50% sortilin receptor inhibition. In the MAD study, twice-daily administration sustained 50% sortilin receptor inhibition and caused greater and more sustained PGRN elevations in plasma and CSF than once-daily dosing. Interestingly, increases in PGRN levels continued to rise in magnitude even after PK had reached steady state. This temporal dissociation between pharmacokinetics and pharmacodynamics suggests a delayed and progressive biological response, potentially mediated by saturation of buffering mechanisms such as prosaposin-PGRN complex formation44,45. The observation that PGRN continued to rise despite stable VES001 concentrations parallels preclinical findings, indicating that full pharmacodynamic expression may require sustained receptor inhibition. Contextualization of these effects within the Genetic FTD Initiative (GENFI), which quantified PGRN using the same assay, provides a translational benchmark46,47. When superimposing the VES001-induced increases from the MAD study onto the reduced baseline levels observed in asymptomatic GRN-carriers, the projected concentrations approach those measured in non-carrier individuals46. We speculate that this supports the hypothesis that VES001 could normalize PGRN in genetically affected populations, a key therapeutic objective. Furthermore, previous studies of sortilin inhibition in Grn-haploinsufficient animals and asymptomatic human carriers have demonstrated amplified PGRN responses relative to controls, suggesting that patient populations may exhibit enhanced pharmacodynamic sensitivity42. Several limitations warrant consideration. The analyses were exploratory, and no adjustments for multiple comparisons were performed, increasing the risk of type I error. Consequently, reported p values should be interpreted as nominal and hypothesis-generating. Additionally, the sample sizes, although sufficient for evaluating safety and pharmacology, limit the precision of estimated effects and the ability to detect rarer adverse effects or subtle dose-response relationships. The 7-day treatment duration precludes conclusions about long-term safety or efficacy. In addition, differences in sortilin expression, baseline PGRN levels, or compensatory mechanisms between healthy volunteers and patients may influence the pharmacodynamic response. These aspects should be addressed in upcoming patient trials that combine extended dosing with patient-specific biomarker monitoring. In conclusion, VES001 is a promising first-in-class therapeutic candidate for FTD-GRN. Its favorable safety profile, BBB penetration, and ability to elevate PGRN in humans, together with the mechanistic concordance between in vitro, in vivo, and clinical data, provide a strong rationale for advancing to studies in genetically defined patient populations. Future trials will determine whether sustained PGRN normalization translates into disease modification. Methods The following sections describe in full the methods and materials used for supporting preclinical experiments and for the first-in-human Phase 1 clinical trial. Preclinical: grating-coupled interferometry (GCI) Sortilin The direct in vitro binding properties of VES001 to recombinant human sortilin (Bio-Techne Ltd, 3154-ST-050) or rat sortilin immobilized on a 4PCH chip (Creoptix AG) were determined by conducting GCI experiments using the Creoptix WAVEdelta equipment (Malvern Panalytical). The GCI assay involved three steps: (1) chip conditioning and immobilization—the preparation of the 4PCH GCI chip prior to sortilin immobilization and the capturing of the progranulin upon the surface of a flow cell, (2) analyte screening—flowing VES001 over the immobilized sortilin upon the surface and a reference flow cell, and (3) data analysis—double referenced binding data (following blank and reference subtraction) fitted with a 1:1 kinetic model to estimate the affinity and kinetics of the measured interaction. Each of these steps, and the production of the recombinant rat sortilin, is detailed below. Chip conditioning and immobilization The 4PCH chip was conditioned across all flow cells using 0.2X concentration running buffer with the following composition: 1X HBS-N at pH 7.4 (Cytiva, BR-1006-70), 3.4 mM EDTA (Merck Life Science UK, E7889-100ML), and 1% DMSO (Sigma-Aldrich, D8418), using the pre-set conditioning wizard (WAVEcontrol software) injecting 0.1 M borate (Xantec Bioanalytics, B BELU-50) and 1 M NaCl (pH 9), followed by three start-up injections of 0.2X running buffer. All buffers described were filtered using a 0.2 µm filter (Nalgene, 10300461) and degassed for at least 15 min prior to use. A flow cell temperature of 25 °C was used throughout the experiment. A buffer exchange to a 1X running buffer was performed during the immobilization procedure. An initial injection of a 1:1 solution of EDC and NHS (Cytiva, BR100050) was performed across all four flow cells to activate the surface for amine coupling of ligands (sortilin). The assay was designed such that each run included four separate flow cells, with one flow cell left blank to be used as a reference channel. Recombinant human sortilin was prepared ahead of the assay by reconstitution at a concentration of 100 µg/mL in sterile PBS and 50 µL aliquots flash frozen in 0.5 mL Eppendorf tubes on dry ice before storage at –80 °C. For the assay, protein aliquots were thawed quickly by hand and centrifuged at 13,300 rpm for 10 min. A 15 µg/mL protein solution was then made in pH 5.0 acetate (Cytiva, BR100351) and injected once for 20 min over non-blank flow cells. A final 7-min passivation injection of 50 mM Tris (Merck Life Science UK, 93352-1KG) was used across all flow cells. A flow rate of 10 µL/min was used for all conditioning and immobilization cycles. Analyte screening The assay was designed to measure the binding of compound VES001 to all immobilized proteins, using the Creoptix WAVE ‘RAPIDKinetics wizard’. A single batch of VES001 (NIDE002-058-1) was tested in 24 runs (N = 24). For each run, VES001 was independently weighed twice, and appropriate volumes of DMSO were added to make a final stock concentration of 10 mM. A compound was used throughout the experiment as a positive control to quantify surface activity, based on in-house experience. Compounds were screened at 1 µM using the in-built ‘intermediate binders’ settings: 100 µL/min, 45 s baseline, 25 s association, 300 s dissociation, with blanks every 5th sample and a DMSO correction (running buffer containing 1.5% DMSO) performed at the start and end of the experiment as well as every 20 cycles. An acquisition rate of 10 Hz and a flow rate of 100 µL/min were used throughout the experiment. Both the control and VES001 were screened at a final assay concentration of 1 µM and a final DMSO concentration of 1%. The compounds were diluted in DMSO from 100 or 10 mM stocks (for the control and VES001, respectively) to 100 µM (100× final assay concentration), then diluted 1:100 in running buffer, which contained no DMSO, to establish a final assay concentration of compound of 1 µM and a final DMSO concentration of 1%. Compounds and DMSO were mixed by plate shaking at 1000 rpm for 60 s using a Bioshake instrument. The reagents were added to the assay plate, and the experiment was run. Data analysis All measurements made were double referenced (blank and reference flow-cell subtracted) before a 1:1 kinetic model was fit to the measured response. These steps were performed using the ‘Kinetic Analysis’ tool of the GCI ‘WAVEcontrol’ software (version 4.5.13) with data fitted using a standard 1:1 kinetic BioModel. The software was used to calculate the association rate constant (Kon), dissociation rate constant (Koff), and the equilibrium dissociation constant (KD) for each run, and the means and standard deviations (s.d.) were calculated across runs in Excel using the AVERAGE and STDEV.P functions. Production of recombinant rat sortilin for GCI binding experiments Recombinant rat sortilin was produced by Peak Proteins Ltd (Macclesfield, UK). A construct encoding amino acids 1–754 of rat sortilin (UniProt O54861), corresponding to the extracellular domain (ECD), was codon-optimized for mammalian cell expression and subcloned into a proprietary vector containing a C-terminal hexahistidine (6×His) tag. The construct was co-transfected transiently with recombinant human furin in HEK293-6E suspension cells (National Research Council Canada) to facilitate propeptide removal at the furin cleavage site. Following a 5–7-day culture period, conditioned medium was harvested, and recombinant rat sortilin was purified by nickel affinity chromatography (Ni-NTA) followed by size exclusion chromatography. Protein purity was assessed by SDS-PAGE, identity confirmed by mass spectrometry, and concentration determined by absorbance at 280 nm corrected for the molar extinction coefficient. Other Vps10p family members The direct in vitro binding properties of VES001 to recombinant human SorCS1 (Bio-Techne Ltd, 3457-SR-050), SorCS2 (Bio-Techne Ltd, 4238-SR-050), SorCS3 (Bio-Techne Ltd, 3326-SR-050), and SorLA (Bio-Techne Ltd, 9880-LA-050) were conducted as described above for sortilin, but with these Vps10p family members immobilized instead. The setups were validated by testing the binding of ligands to the receptors prior to testing VES001. VES001 did not show any binding to SorCS1, SorCS2, SorCS3, or SorLA, as shown with some examples in Supplementary Fig. 1. Preclinical: SafetyScreen87 panel and CYP panel The off-target binding properties of VES001 were tested in the commercially available Eurofins’ SafetyScreen87 panel-FR at 10 µM in vitro. The target classes included G-protein-coupled receptors, ion channels, enzymes, and transporters. The full overview of the 87 targets is available on the Eurofins website48. Moreover, 10 µM VES001 was also analyzed with the commercially available Eurofins’ Cytochrome P450 inhibition panel (overview available on Eurofins website), comprised of eight Cytochrome P450 isozymes39. Compound binding was calculated as a percentage inhibition of the binding of a ligand specific for each target. The compound enzyme inhibition effect was calculated as the percentage inhibition of control enzyme activity. In vitro absorption was evaluated as the compound transporter inhibition effect, calculated as the percentage inhibition of the vehicle control activity. Results showing an inhibition or stimulation higher than 50% would represent significant effects of the test compound. To assess in vitro metabolism, as previously, compound binding was calculated as a percentage inhibition of the binding of a ligand specific for each target, and results showing an inhibition higher than 50% would represent significant effects of the test compound. Preclinical: PGRN–sortilin competition binding by flow cytometry The effect of VES001 on PGRN–sortilin binding was assessed in an in vitro flow cytometry displacement assay (MACSQuant X, Miltenyi Biotec) using the HEK293-FL cell line (HEK293 over-expressing full-length sortilin), from which VES001 IC50 and Ki values were determined. The assay protocol and data analysis are detailed below. Assay protocol The HEK293-FL stable cell line was maintained in T175 cell culture flasks using culture media (DMEM + 10% FBS + 1% penicillin/streptomycin (100 units penicillin, 100 µg/mL streptomycin final concentration) + 500 µg/mL Zeocin). When cells reached 80% confluency, they were either passaged for maintenance of the cell line or harvested for use in screening assays. After harvesting, cells were counted using trypan blue and the Luna II cell counter. Cells were then passed through a cell strainer and added to all wells of the 384-well assay plate using the Certus liquid handler. Control or test compounds were then dispensed into the assay plate, with test compounds dispensed in duplicate as 8-point half-log concentration-response curves, with a top concentration of 30 µM, and dimethyl sulfoxide (DMSO) normalized to 0.3%. The plate was then sealed, briefly vortexed, and pulsed using the centrifuge before incubating on ice for 1 h. PGRN was custom-made with an N-terminal Avi-tag and biotinylated using BirA co-expression during biosynthesis, to ensure biotinylation in the Avi-tag region only (6HIS-Avi-PGRN). This PGRN was prepared at 120 nM (4× concentrated) in assay buffer, with 10 µL then added to each well of the 384-well assay plate, followed by incubation on ice for a further 1.5 h. Streptavidin-AF647 was prepared at 1200 nM (4× concentrated) in assay buffer, with 10 µL then added to each well, followed by a 0.5 h incubation. The MACSQuant X (Miltenyi Biotec) was turned on to acquisition mode, and the lasers were allowed to warm to 37 °C. The machine was calibrated using calibration beads prior to sample acquisition. Following run completion, ‘.mqd’ data files were copied from the machine to the project folder for analysis. Data analysis Screening data were analyzed using FlowJo, and the gating strategy is shown in Supplementary Fig. 2. The median fluorescence intensity (MFI) statistic was applied to the single cell gated population on the PGRN-AF647 channel, with the data for each individual sample then exported using the table editor to Excel for analysis. Raw MFI data were normalized to the control compounds and loaded into Prism or Dotmatics for analysis using a 4-parameter curve fit (variable slope model) and IC50 values reported. Inhibitory constant (Ki) values were then determined using the Cheng-Prusoff equation: Ki = IC50/[1 + (L/Kd)]. For the calculation, L (ligand concentration) = 30 nM and Kd = 28.8 nM (determined during assay validation). Z′-factor and assay window plate statistics were calculated using the high and low controls and used alongside a VES001-similar control IC50/Ki for passing screening plates. Pass criteria comprised Z′ > 0.5, assay window >2, and positive control IC50 within 3-fold of its expected value. Preclinical: sortilin downregulation assay by flow cytometry The impact of VES001 on the cell surface and total sortilin protein levels was evaluated in an in vitro flow cytometry assay (MACSQuant X) using COLO792 cells. Permeabilized cells were used for the total sortilin levels, while non-permeabilized cells informed on the cell surface levels. Prior to permeabilization, the cells were treated for 24 h with a wide concentration range of VES001 (3, 9, 30, 90, 300, 900, 3000, and 9000 ng/mL). VES001 was tested alongside a sortilin-targeting antibody known to downregulate sortilin (1500 ng/mL), as a positive control, DMSO, and a rabbit IgG isotype control. Cell culturing COLO792 cells (Merck, 93052616-1VL) were cultured in Roswell Park Memorial Institute (RPMI) medium (Gibco, 11875093) supplemented with 10% fetal bovine serum (FBS) (SLS, F9665-500 mL). On the day of the assay, cells were counted and seeded at a density of 50,000 cells per well in two 48-well plates, with a total volume of 500 µL per well. Plates were incubated at 37 °C with 5% CO2 for approximately 4 h. Compound treatment All compound dilutions were done in culture media (RPMI + 10% FBS). The sortilin-targeting antibody latozinemab biosimilar (HY-P99688-1mg) was added approximately 4 h after plating. To dispense the compound, a 1 in 100 stock solution was prepared, and 6.4 µL was added to all high control wells. Fixation and permeabilization Media was removed using a multichannel pipette, and plates were rinsed with 500 µL of DPBS. After aspirating the DPBS, 200 µL of IC fixation buffer (Invitrogen, 00-8222-49) was added to all wells and incubated at room temperature for 30 min. Fixation buffer was then removed, and plates were rinsed twice with 500 µL DPBS. Permeabilization buffer (200 µL) (Invitrogen, 00-8333-56) was added to plate 2 only and incubated for 15 min at room temperature. All wells were then rinsed twice with 500 µL of DPBS. Staining Cells were stained with SORT1-CoraLite647 antibody (Proteintech, CL647-12369) at 2.5× the recommended concentration (1 µg per 1 × 106 cells; stock: 1 mg/mL; supplier recommendation: 0.4 µg per 1 × 106 cells). Given that the cells were seeded 24 h prior, a maximum of 100,000 cells per well were assumed at the time of staining. Therefore, 0.1 µg of antibody was used per well in 200 µL volume. For staining, 20 mL of SORT1-CoraLite647 solution was prepared by adding 10 µL of antibody stock to 20 mL of DPBS + 0.5% BSA. The rabbit IgG isotype control (stock: 0.2 mg/mL) (Invitrogen, 51-4616-82) was prepared similarly in a total volume of 5 mL (12.5 µL in 5 mL DPBS + 0.5% BSA). Plates were incubated at room temperature, protected from light (wrapped in aluminum foil), for 1 h. After incubation, cells were rinsed twice with DPBS. Cell detachment TrypLE (200 µL) (Gibco, 12604021) was added to all wells. Plates were incubated at 37 °C, 5% CO2, and monitored under a microscope until cell detachment was observed (typically within 20–30 min). Once detachment occurred, TrypLE was inactivated by adding 100 µL of culture media containing FBS to every well. Cells were transferred to 96-well V-bottom plates according to the plate map using a multichannel pipette. Plates were centrifuged at 300 × g for 5 min. Using a Viaflow instrument, 150 µL of supernatant was carefully aspirated, leaving approximately 150 µL per well for flow cytometry analysis. Plates were stored at 4 °C, wrapped in foil, until the flow cytometry experiment. Analysis The flow cytometer (MACSQuant X) was calibrated prior to the assay using calibration beads. Forward scatter (FSC) and side scatter (SSC) gain parameters were manually adjusted using 20 µL of the sample. The assay was run using a gain of 450 for the APC channel. All raw data were first analyzed using FlowJo software. Median fluorescence intensity (MFI) values and cell counts were exported as CSV files. Two batches of VES001 were each analyzed in three runs, giving six runs in total with varied n for each group within each run. Within each run, the data were normalized with the isotype control set to 0% and the DMSO control set to 100%. The normalized data were then pooled across runs (N = 6), and the arithmetic means and standard deviations were plotted using GraphPad Prism (version 10.1.1). Preclinical: rapid equilibrium dialysis (RED) on human plasma In vitro protein binding of VES001 in human plasma of mixed sex (BioIVT, lHMN646547) was determined using the Rapid Equilibrium Dialysis (RED) method and LC-MS/MS equipment. Three technical replicates (N = 3) of 5 µM VES001 in mesylate salt form (batch NIDE002-058-1) and 1 µM of the positive control compound, propranolol, were incubated in human plasma for 4 h at 37 °C. After incubation, aliquots of donor side matrices were collected and diluted with an equal volume of the blank receiver side matrix. Aliquots of the receiver side matrices were diluted with an equal volume of blank donor side matrix. All samples were protein precipitated by the addition of a two-fold volume of internal standard solution. The donor side volume was 200 µL, the receiver side volume was 350 µL, and the receiver side solution was 150 mM PBS at pH 7.4. The stock solution was 100-fold in DMSO. After centrifugation for 10 min at 13,200 rpm with an M-20 swinging bucket rotor, the supernatants were analyzed with LC-MS/MS to obtain the unbound fraction of test substance in plasma. The recovery sample was prepared by pooling the remaining amounts of spiked donor side matrices (replicates) together and preparing these for analysis in the same manner as donor side samples. The sample was prepared for analysis immediately after starting the RED experiment. The RED analysis was conducted using the following HPLC and MS equipment: Waters ACQUITY PREMIER UPLC, Waters TQ Absolute triple quadrupole MS, and the Waters ACQUITY BEH C18 column (50 × 2.1 mm, particle size 1.7 µm) with a guard filter column. The polarity was ESI+, with a desolvation flow of 900 L/h, cone gas of 200 L/h, capillary voltage of 1.2 kV, desolvation temperature of 650 °C, and source temperature of 150 °C. The first 0.5 min of the run was directed into waste by using a divert valve to decrease the ion source contamination by early-eluting matrix constituents. The run included a gradient elution of 0.1% formic acid in water and acetonitrile with an injection volume of 2 µL, a constant flow of 0.5 mL/min, and a column oven temperature of 40 °C. The gradient elution ran with a ratio of 0.1% formic acid in water to acetonitrile of 98:2 at 0.00 and 0.5 min, 35:65 at 2.5 min, 2:98 at 2.7 and 3.5 min, and finally 98:2 at 4.5 min. The VES001 MRM reactions were 306 > 117, 306 > 147, and 306 > 160 at collision energies of 28, 14, and 10, respectively. The unbound fraction and recovery were calculated from the analyte peak areas (A) and the matrix volumes (V) as follows, and are reported as arithmetic means (AVERAGE) and standard deviations (STDEV): Preclinical: in vivo microdialysis study on VES001 in male rats Animals and ethics The microdialysis study was conducted in thirteen fasted male Sprague-Dawley rats (290–375 g; Charles River Laboratories, Germany), which were divided into two treatment groups to receive either vehicle or VES001. The experiments were conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 2011) and were in accordance with European Union directive 2010/63 and the Dutch law. The experiments were carried out under license number AVD23600202011187, issued by the national committee for licensing of animal experiments (Centrale Commissie Dierproeven) and were approved by the Animal Care and Use Committee (Instantie voor Dierenwelzijn) of Charles River Discovery Groningen. Following arrival, animals were housed in groups of up to 5 in polycarbonate cages (56 × 33 × 20 cm) with a wire mesh top in a temperature (22 ± 2 °C) and humidity (55 ± 15%) controlled environment on a 12-h light cycle (07.00–19.00 h). After surgery, animals were housed individually (cages 30 × 30 × 30 cm). Standard diet (R/M-M A04 mod., Ssniff, Germany) and domestic quality mains water were available ad libitum. Stereotaxic surgery Sampling probes were positioned by stereotaxic surgery (Kopf Instruments, USA). All rats were implanted with guides and push-pull microdialysis probes with a 4 mm exposed polyethersulfone (PES 200) membrane (CRL Groningen, the Netherlands) targeting the left hippocampus to measure PGRN concentrations. In animals assigned to receive VES001, guides and MetaQuant microdialysis probes with a 4 mm exposed ethylene vinyl alcohol (EC20) membrane (CRL Groningen, the Netherlands) were additionally placed in the right hippocampus to measure VES001 concentrations. Coordinates for the tip of the probes were: AP = –5.3 mm (to bregma), lateral ±4.8 mm (to midline), ventral –8.0 mm (to dura). The incisor bar was set at −3.3 mm. All coordinates were based on ‘The rat brain in stereotaxic coordinates’ by Paxinos and Watson (2009). The probes were attached to the skull with stainless steel screws and dental cement. The rats receiving VES001 were also surgically equipped for cerebrospinal fluid (CSF) and blood sampling in the same surgical procedure. A stainless-steel cannula (CRL Groningen, the Netherlands) was positioned in the cisterna magna to allow for CSF sampling, and an indwelling cannula (35–42 mm, CRL Groningen, the Netherlands) was positioned in the jugular vein and exteriorized through an incision on top of the skull to allow for blood sampling. The end of each cannula was fixed in position with dental cement and attached to the skull with stainless steel screws. Before surgery, carprofen (5 mg/kg, s.c.) was administered for analgesia during surgery and the post-surgical recovery period. During surgery, animals were anesthetized using isoflurane (2.5–3.0% and 800 mL/min O2). A mixture of bupivacaine and epinephrine was used for local analgesia of the incision site and periosteum of the skull. Formulation and dosing VES001 was formulated in 20% sulfobutylether-β-cyclodextrin (SBE-β-CD) in deionized water at a concentration of 20 mg/mL and administered by oral gavage at a dose of 100 mg/kg in a volume of 5 mL/kg. All formulations were prepared fresh, and the doses were adjusted for the salt correction factors. This pharmacokinetic (PK)/pharmacodynamic (PD) experiment commenced 1 day following surgery. All rats received oral doses at 0, 3, and 6 h relative to the first administration. A total of 6 rats received 0 mg/kg (vehicle), and 7 rats were administered 100 mg/kg of VES001 formulated in 20% SBE-β-CD. Food was withdrawn 2 h before the first dose (–2 h) and returned after the third oral administration at 6 h. Microdialysis and sample collection The probes were perfused with artificial CSF (aCSF: containing 147 mM NaCl, 3.0 mM KCl, 1.2 mM CaCl2, and 1.2 mM MgCl2) at a flow rate of 0.5 µL/min (push-pull probes) or with a slow flow of aCSF at a flow rate of 0.12 µL/min and a carrier flow of ultrapurified H2O at a flow rate of 0.8 µL/min (MetaQuant probes). After 2 h of prestabilization, microdialysis samples were collected in 1-h intervals, and each sample was set to the end of the sampling period. For example, three basal samples were collected from –3 to 0 h, corresponding to samples at –2, –1, and 0 h. Microdialysate samples were collected continuously for 12 h after the first compound administration. Blood and CSF samples were taken at –2, 0, 1, 3, 4, 6, 7, 10, and 12 h relative to the first oral dose. All samples were collected into 0.5 mL LoBind Eppendorf tubes and stored at –80 °C until analysis. Bioanalysis and data analysis The MetaQuant microdialysis probe achieved a mean and standard deviation relative in vitro recovery of VES001 of 69.8 ± 3.40%. In contrast, the push-pull microdialysis probe demonstrated a previously established in vitro recovery of PGRN of 26.9 ± 0.86%. The compound and PGRN levels were adjusted based on probe recovery. Compound concentrations of VES001 were quantified using high-performance liquid chromatography (HPLC) coupled with tandem mass spectrometry (MS/MS) detection. For the quantification of rat PGRN, a commercially available ELISA kit was used according to the manufacturer’s guidelines (AdipoGen, AG-45A-0043YEK-KI01). Plasma samples were analyzed in duplicate. Samples that were outside the detection limits were re-analyzed on another plate in a more suitable dilution. In addition, samples for which the %CV of the duplicates was above 30% were re-analyzed on a different plate. Absorbance at 450 nm was measured using a Multiskan FC (Thermo Fisher Scientific, USA). PGRN results were expressed as both the absolute and percentage changes from baseline. Data were processed using Microsoft Excel, and all compound and PGRN data were analyzed and visualized using GraphPad Prism (version 10.1.1). First-in-human phase 1 study: ethical considerations The study adhered to the protocol, ICH GCP, the Declaration of Helsinki, and local regulations (Dutch Medical Research in Human Subjects Act). Ethics approval was obtained prior to initiation, and written informed consent was provided by all participants prior to participating in the study. First-in-human phase 1 study: design This single-center, randomized, double-blind, placebo-controlled first-in-human Phase 1 trial included single ascending dose (SAD) and multiple ascending dose (MAD) cohorts (EU CTR 2023-505847-38-00). The study was conducted at the Centre for Human Drug Research in Leiden, the Netherlands, between 19 Oct 2023 and 26 Jul 2024. The SAD part comprised 6 cohorts of healthy study participants receiving single doses of VES001 (10–990 mg) or placebo (6:2 ratio per cohort; pooled placebo N = 12) (Fig. 3A). Sentinel dosing was used for each cohort. Participants were admitted on Day –1, dosed in a fasted state on Day 1, remained in-house through Day 3, and returned for follow-up on Day 8 (±2 days) (Fig. 3B). Plasma PK and PGRN levels were sampled pre-dose and at designated post-dose time points on Days 1–3. CSF was collected via lumbar puncture pre-dose and at 7 h post-dose in the placebo, 810 mg, and 990 mg VES001 cohorts (Fig. 3C). Cohort 3 completed a second visit after a washout period of at least 4 weeks to assess the effect of food, receiving the same dose in a fed state. A dose escalation review of the fasted visit of cohort 3 and safety review of at least 6 participants of cohort 4 up to at least 48 h was conducted prior to the food effect visit. The MAD study included 3 cohorts dosed with VES001 (180 mg once daily, 180 mg twice daily, or 450 mg twice daily) or placebo for 7 days (8:2 ratio; pooled placebo N = 6) (Fig. 3A). Participants were admitted on Day –1 and remained in-house through 48 h post-final dose (Day 9), with follow-up on Day 15 (±2 days) (Fig. 3B). Plasma PK and PGRN were measured on Days 1–9 and on Day 15. CSF samples were collected on Day –1 (baseline) and Day 7 (Fig. 3C). Participants were screened within 42 days prior to dosing. Dose escalation decisions were based on safety, PK, and plasma PGRN data from previous cohorts, with PGRN evaluation lagging by no more than one cohort. Dose-escalation decisions were based on blinded aggregate data, including safety, PK, and PD summaries at the cohort level that included both active and placebo participants. Individual treatment assignments were not revealed, and no subject-level unblinded PK or PD data were reviewed during interim decision-making. First-in-human phase 1 study: participants Eligible participants were healthy adults aged 18–55 years with a BMI of 18–32 kg/m² and a minimum weight of 50 kg. Effective contraception was required during the study and for 90 days post-treatment. Exclusion criteria included medical conditions or treatments potentially interfering with study assessments, such as neurological disorders, recent malignancy, significant head trauma, severe allergies, abnormal lab results, or infectious diseases (Hepatitis B/C, HIV). Heavy smokers or those unable to abstain were also excluded. The treatment allocation sequence was generated by an unblinded statistician at the study center using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Participants were enrolled by the investigator and assigned to treatment sequentially according to the predefined randomization schedule. The randomization list was kept strictly confidential and was not accessible to investigators, study staff, participants, the sponsor, or blinded study monitors. First-in-human phase 1 study: blinding The study was conducted in a double-blind manner. VES001 and placebo were indistinguishable in appearance and packaging. Investigators, participants, study staff, and the sponsor remained blinded to treatment allocation throughout the study. The randomization list was accessible only to the pharmacist responsible for preparation of study medication and to designated personnel involved in pharmacokinetic bioanalysis and statistical programming for dose-escalation support. These individuals were not involved in participant care, clinical assessments, or endpoint evaluation. Blinded summaries of safety, pharmacokinetic (PK), and pharmacodynamic (PD) data were prepared by unblinded statisticians/programmers in a restricted-access environment and provided to the investigator and sponsor’s medical monitor for dose-escalation decisions. First-in-human phase 1 study: IMP administration VES001 was formulated as a pure solid compound without excipients in gelatin capsules (10, 30, 90, or 180 mg). Placebo capsules were visually identical. All study treatments were prepared at the Leiden University Medical Centre Pharmacy, The Netherlands. Doses were administered orally with 300 mL of still water after a 10-h fast, or, for the food-effect arm, following a high-fat meal. In twice daily regimens, the second dose was given 2 h after fasting. The starting dose (10 mg) was 100-fold below the No Observed Adverse Event Level in dogs (30 mg/kg in dogs, which converts to 1169 mg in a 70 kg human) and 10-fold below the Minimum Anticipated Biological Effect Level in rats (9.29 mg/kg in rats, which converts to 105.4 mg in humans). First-in-human phase 1 study: procedures The primary objective was to assess the safety and tolerability of single and multiple doses of VES001. Safety endpoints included adverse events (AEs), serious unexpected suspected adverse reactions (SUSARs), changes in lab values, ECGs, vital signs, physical/neurological exams, and Columbia-Suicide Severity Rating Scale (C-SSRS) scores. Secondary objectives assessed plasma and CSF PK following single, multiple, and food-effect dosing. PK parameters included AUC0–inf, AUC0–last (SAD/food-effect only), % extrapolated AUC, Cmax, Tmax, Tlag, T½, CL/F, Vz/F, and accumulation ratio (MAD only). CSF endpoints included CSF/plasma ratios in SAD cohorts 5–6 and MAD steady-state cohorts. Exploratory objectives included measuring PGRN concentrations in plasma and CSF and semi-quantitative analysis of VES001 metabolites in plasma and urine. First-in-human phase 1 study: human plasma and CSF PGRN analysis PGRN concentration was quantified in plasma and CSF using a previously developed Meso Scale Discovery (MSD)-based immunoassay38. In brief, blocking buffer (0.5% bovine serum albumin (BSA), in PBS-T (0.05% Tween-20 in 1X phosphate-buffered saline)) was added to a streptavidin-coated MSD plate (MSD, L15SA-1) and incubated overnight at 2–8 °C. The plate was then washed with PBS-T (a step conducted before the addition of each reagent), and biotinylated human PGRN capture antibody added (Biotechne R&D Systems, BAF2420, URL0321101) followed by a 1 h incubation at room temperature (RT). The standard curve (7.813–5000 pg/mL) was prepared using recombinant human PGRN (Biotechne R&D Systems, 2420-PG, MQW1021101) diluted in assay buffer (0.25% BSA in PBS-T). Standards and samples were then added to the plate and incubated for 2 h at RT. The detection antibody (Biotechne R&D Systems, MAB2420, WWO0220081) was added and incubated for 1 h at RT, followed by the addition of the secondary goat anti-mouse antibody with SULFO-TAG (MSD, R32AC-5, W0021235S) and a further 1 h incubation at RT. Finally, the electrochemical signal was generated by the addition of MSD read buffer GOLD (MSD, R92TG-2) and light emission read using a QuickPlex SQ120 plate reader (MSD, 13001801081059). Data was processed with MSD Discovery Workbench® 4.0.12 and SoftMax Pro GxP 7.0.3. Due to variability in baseline PGRN levels, the results were reported as the absolute changes in PGRN from baseline and the percentage change in PGRN from baseline and illustrated using GraphPad Prism (version 10.1.1). Relative Light Unit (RLU) values were reported as whole numbers and analyte concentrations to three decimals, while % Theoretical, % Nominal, % CV, and %RE were reported to one decimal. The acceptance criteria for the individual samples (clinical samples, blank) were as follows: the %CV of duplicates must not exceed 20%, except for the lower limit of quantitation (LLOQ) and the upper limit of quantitation (ULOQ), which must not exceed 25%; %Theoretical must be between 100 ± 20% for each calibration standard, except for the LLOQ and the ULOQ, which must be between 100 ± 25%; and the blank response must be lower than the CS1 response. The sample analysis acceptance criterion was that the CV% of duplicates must not exceed 20% for unknown samples. First-in-human phase 1 study: human plasma and CSF compound analysis VES001 was quantified in human plasma and CSF by an initial protein precipitation followed by LC-MS/MS analysis. In brief, VES001 and an internal standard analog (dexamethasone) were isolated from 20 µL CSF or plasma using protein precipitation. Calibration standards in the range from 1.00 to 5000 ng/mL used for method qualification were freshly prepared in human CSF or plasma on the day of analysis. Chromatographic separation was conducted using an Acquity UPLC column (Waters, HSS T3 1.8 µm 2.1 × 50 mm) with a flow rate of 0.6 mL/min, a sample injection volume of 4 µL, and a 4-min gradient (1.5 min, B 65%; 1.6 min, B 100%; 2.6 min, B 30%). Mobile phases were Buffer A: 0.1% formic acid (FA) in ultra-pure water (UPW) and Buffer B: 0.1% FA 84% acetonitrile (ACN), 0.1% FA. This U-HPLC was coupled to a QTRAP6500+ tandem mass spectrometer equipped with a turbo ion spray probe in positive multiple reaction monitoring mode. QC samples were loaded regularly between runs and data acquisition, peak area integration, and regression were performed using Analyst software (version 1.7.2) from AB Sciex. Concentrations were calculated using 13-point curves with weighted linear regression, according to the following formula: where x = VES001 concentration in ng/mL, y = peak-area ratio, a = intercept, and b = slope. The lower limit of quantitation (LLOQ) for VES001 was 1.00 ng/mL, and the upper limit of quantitation (ULOQ) was 5000 ng/mL. The results were illustrated using GraphPad Prism (version 10.1.1). Each analytical run incorporated duplicate quality control (QC) samples at three concentrations (low, medium, and high), which were prepared freshly in human CSF on the day of sample analysis or in bulk and stored at ≤ –18 °C. The acceptance criteria for the QC samples were as follows: at least four of the six QC samples had to be within ±15.0% of their respective nominal values (%bias), and two of the six QC samples were permitted to be outside ±15.0%, provided they were not at the same concentration. The overall (mean) accuracy of the QC samples for all accepted runs had to be within 85.0–115.0% of the theoretical value. In addition, the overall (mean) coefficient of variation (CV%) of the QC samples for all accepted runs had to be ≤15.0%. First-in-human phase 1 study: statistical analysis All safety and statistical programming were conducted with SAS 9.4 for Windows or newer (SAS Institute, Inc.). PK non-compartmental analysis was conducted with R 3.6.1 for Windows or newer (R Foundation for Statistical Computing/R Development Core Team, Vienna, Austria, 2019). No imputation was performed for missing data. Multiplicity adjustments were not applied. Bioanalysis of PK samples was performed by designated personnel who had restricted access to treatment allocation but were not involved in clinical assessments or dose-escalation decisions. Pharmacodynamic analyses of the progranulin measurements were conducted after completion of clinical data collection and database lock. Investigators and all personnel involved in clinical evaluation and endpoint interpretation remained blinded to treatment allocation throughout the study. Participant disposition and demographic information were summarized for the treated population. Safety analyses were based on all enrolled participants who received at least one dose of study treatment. Mean plasma and CSF concentrations of VES001 were calculated for each time point. The repeatedly measured pharmacodynamic endpoint of PGRN concentration in plasma was summarized by treatment group and time point using descriptive statistics, including sample size (n), mean, geometric mean, standard deviation (s.d.), standard error of the mean (s.e.m.), coefficient of variation (CV), geometric CV%, median, minimum, and maximum values. These data were further illustrated graphically as absolute and % change from baseline over time, with s.d. represented by error bars. Similarly, the single-measure pharmacodynamic endpoint of PGRN concentration in CSF was summarized using the same descriptive statistics and presented graphically via box-and-whisker plots stratified by treatment and time. CSF/plasma ratios were calculated separately for Part A and Part B. In Part A, ratios were obtained by dividing the individual CSF concentration (sampled at approximately 7 h post-dose) by the corresponding plasma concentration collected at 6 h post-dose. In Part B, CSF and plasma samples were collected within a short time interval; therefore, ratios were calculated using the plasma concentration nearest to the CSF sampling time to minimize temporal differences between compartments. Accordingly, plasma samples collected 6 h after Dose 7 (Cohort 1) and 6 h after Dose 13 (Cohorts 2 and 3) were used. For the urine analysis, semiquantitative analysis was performed to assess how many metabolites were present and what their relative abundance was. PD endpoints, including PGRN concentrations in plasma and CSF, were natural log-transformed prior to inferential analyses to stabilize variance and improve approximation to normality. While this does not guarantee that normality assumptions are fully met, the transformed data were considered appropriate for the applied mixed-effects models in this exploratory analysis. After log transformation, repeated PD measurements, such as plasma PGRN concentrations, were analyzed using a mixed-model analysis of covariance (ANCOVA) with fixed effects for treatment, time, and their interaction, adjusting for the (averaged) baseline value as a covariate. The subject was included as a random effect to account for within-subject correlation. Single PD measurements, such as PGRN concentrations in CSF, were analyzed using a mixed-model analysis of variance (ANOVA) with similar fixed and random effects structures. Estimates derived from the models are on the log scale and were back-transformed to the original scale for reporting. These back-transformed estimates represent geometric means or ratios of geometric means, reflecting multiplicative treatment effects. Treatment effects are interpreted as percentage changes relative to baseline. No formal hypothesis testing framework was prespecified, as the study was exploratory in nature. The analyses were intended to characterize pharmacodynamic trends and support dose selection in subsequent patient studies rather than to establish confirmatory statistical significance. Accordingly, no adjustments for multiple comparisons were applied. All reported p values are therefore nominal and should be interpreted from a hypothesis-generating perspective. First-in-human phase 1 study: clinical trial registration The trial was registered prospectively in the European Union Clinical Trials Information System (EU CTR/CTIS) under EU trial number 2023-505847-38-00. The study was conducted at a single site in the Netherlands and was therefore authorized and registered under Regulation (EU) No 536/2014, for which entry in the EU CTR is the mandatory and prospective registration route; CTIS is a World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) primary registry. The protocol and informed consent form were approved by the independent ethics committee on 11 October 2023. The first participant gave informed consent on 19 October 2023, the first participant was dosed on 22 November 2023, and the last participant completed the study on 26 July 2024. All eligibility criteria, endpoints, and analyses reported here were pre-specified in the EU CTR record before the first participant was enrolled. The trial was additionally registered on ClinicalTrials.gov (NCT06226064) so that it would be discoverable to the international frontotemporal dementia research community, to clinicians, and to GRN mutation carriers and their families, for whom ClinicalTrials.gov is the most widely consulted trial resource. This second duplicate registration was submitted on 7 December 2023 and first posted on 26 January 2024, that is, after the first participant had been enrolled and dosed. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting summary Transparent Peer Review file Source data source data Acknowledgements The team would like to thank Manuel Javier Cases-Thomas, who had a critical role in the design, selection, and early characterization of VES001. The preclinical work to characterize VES001 was conducted in collaboration with the following contract research organizations: DomaineX Ltd., Charles River Laboratories, Symeres, Sygnature Discovery, Anapharm, and Ardena. Author contributions M.K., L.K., I.J.S., S.L.F., L.F., G.W., N.S., A.R., O.P., A.N., P.B.L., D.M., S.P., D.D., E.T., J.C., and P.K. contributed to the conception and design of the study. D.D. was responsible for the clinical data collection. M.K., L.K., I.J.S., S.L.F., G.W., N.S., O.P., S.P., D.D., E.T., Y.Y., O.M., and P.K. performed the data analysis. M.K., L.K., I.J.S., S.L.F., G.W., N.S., A.R., O.P., A.N., P.B.L., S.P., D.D., E.T., Y.Y., O.M., and P.K. contributed to the interpretation of the data. M.K., L.K., I.J.S., N.S., D.D., E.T., and P.K. drafted the manuscript. All authors critically reviewed and revised the manuscript, approved the final version for submission, and agree to be accountable for all aspects of the work. Peer review Peer review information Nature Communications thanks Fenghua Hu, Bruno Imbimb and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding PROMEMO is supported by the Danish National Research Foundation (grant no. DNRF133). Data availability The data generated in this study derive from a first-in-human Phase 1 clinical trial (ClinicalTrials.gov NCT06226064; EU CTR 2023-505847-38-00) and from the accompanying preclinical studies. The de-identified participant-level clinical data and the preclinical data underlying every graph presented in this article (Figs. 1A, B, D–F, 4A–G and 5A–F) are made freely available, without restriction, in the Source Data file accompanying this article. No conditions or permissions apply to their access, use, or reuse, and no data-sharing agreement is required. Trial data collected beyond those reported here—the complete study database, case report forms and clinical study report—are not made publicly available, because they contain information that could compromise the privacy and the informed consent of trial participants and because the study sponsor is bound by the trial protocol and by applicable data-protection legislation (the EU General Data Protection Regulation and the Dutch Medical Research Involving Human Subjects Act, WMO). Requests for these data, and for the clinical study protocol, the statistical analysis plan, or the clinical study report, may be directed to the corresponding author, M.K. (mfk@vesperbio.com). Requests are assessed individually by Vesper Bio ApS and answered within 60 days, and where sharing is appropriate, access is granted subject to any ethical and regulatory approvals required. No custom or bespoke computer code was generated in this study. All analyses were performed using commercially or publicly available software as described in the Methods: SAS (version 9.4; SAS Institute), R (version 3.6.1; R Foundation for Statistical Computing) for pharmacokinetic non-compartmental analysis, and GraphPad Prism (version 10.1.1) for data visualization. No restrictions apply. Source data are provided with this paper. Competing interests M.K., A.N., and P.L. are named inventors on the following patent applications, assigned to Vesper Bio ApS (certain earlier filings are recorded under the company name Insusense ApS), which claim small-molecule modulators of sortilin and their therapeutic use and which encompass the chemical series to which VES001, the investigational compound evaluated in this study, belongs: WO 2022/157271, “Pyridine derivatives as modulators of sortilin activity” (EP 4281460 A1; US 2024/0132514 A1); WO 2022/223805, “Modulators of sortilin activity” (EP 4326742 A1; US 2024/0217922 A1); WO 2022/238565, “Modulators of sortilin activity” (EP 4337674 A2; US 2024/0294569 A1); WO 2023/031440, “2-Amino-5,5-dimethylhexanoic acid derivatives as sortilin modulators for use in the treatment of disease of the central nervous system” (EP 4395776 A1; US 2025/0268854 A1); WO 2023/161505, “Amino acid based carbamates and/or ureas for the treatment of sortilin dependent diseases” (EP 4486443 A1; US 2025/0161321 A1); WO 2024/047227, “Sortilin modulators” (EP 4580750 A1); WO 2024/089049, “Modulators of sortilin activity” (EP 4361165 A1); WO 2024/184468, “Modulators of sortilin activity” (EP 4676917 A1); WO 2024/184469, “Modulators of sortilin activity” (EP 4676914 A1); and WO 2024/184471, “Sortilin modulators” (EP 4676915 A1). These filings claim composition of matter and/or medical use of the compound class to which VES001 belongs, and M.K., A.N., and P.L. therefore have a financial interest in the successful clinical development and commercialization of VES001, the subject of this manuscript. L.K., I.J.S., S.L.F., D.N., L.F., S.P., G.M.W., A.R., N.S., P.L., O.P., A.N., and M.K. are employees or consultants of Vesper Bio ApS. M.K. and A.N. hold shares and/or share options in Vesper Bio ApS. A.N. and M.K. are co-founders of Vesper Bio ApS. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Louise Klem, Daniel Dumas. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77638-2. References Associated Data Supplementary Materials Supplementary Information Reporting summary Transparent Peer Review file source data Data Availability Statement The data generated in this study derive from a first-in-human Phase 1 clinical trial (ClinicalTrials.gov NCT06226064; EU CTR 2023-505847-38-00) and from the accompanying preclinical studies. The de-identified participant-level clinical data and the preclinical data underlying every graph presented in this article (Figs. 1A, B, D–F, 4A–G and 5A–F) are made freely available, without restriction, in the Source Data file accompanying this article. No conditions or permissions apply to their access, use, or reuse, and no data-sharing agreement is required. Trial data collected beyond those reported here—the complete study database, case report forms and clinical study report—are not made publicly available, because they contain information that could compromise the privacy and the informed consent of trial participants and because the study sponsor is bound by the trial protocol and by applicable data-protection legislation (the EU General Data Protection Regulation and the Dutch Medical Research Involving Human Subjects Act, WMO). Requests for these data, and for the clinical study protocol, the statistical analysis plan, or the clinical study report, may be directed to the corresponding author, M.K. (mfk@vesperbio.com). Requests are assessed individually by Vesper Bio ApS and answered within 60 days, and where sharing is appropriate, access is granted subject to any ethical and regulatory approvals required. No custom or bespoke computer code was generated in this study. All analyses were performed using commercially or publicly available software as described in the Methods: SAS (version 9.4; SAS Institute), R (version 3.6.1; R Foundation for Statistical Computing) for pharmacokinetic non-compartmental analysis, and GraphPad Prism (version 10.1.1) for data visualization. No restrictions apply. Source data are provided with this paper.