Supramolecular nanostructure mimics GDNF trophic effects in vitro on human dopaminergic neurons

Abstract

Peptide-based supramolecular nanostructures offer a versatile platform with substantial promise for clinical translation in regenerative medicine. These systems allow for the incorporation of biologically active sequences and can be engineered to modulate tissue-specific parameters such as stiffness, diffusivity, and biodegradability. We developed here a bioactive supramolecular nanostructure containing a peptide designed based on glial cell-derived neurotrophic factor. These nanostructures form scaffolds that mimic important trophic effects provided by this growth factor on iPSC-derived human dopaminergic neurons. Our in vitro data show that the nanostructures promote cell viability, confer neuroprotection against 6-hydroxydopamine toxicity, enhance neuronal morphology, facilitate electrophysiological maturation, and induce genes involved in neuronal survival. We also found that the scaffold promoted axonal extension in midbrain human organoids. These findings suggest that the supramolecular system could be useful to improve outcomes in cell-based therapies for Parkinson’s disease, where progressive dopaminergic degeneration is a hallmark.

Introduction

Parkinson´s disease (PD) is recognized as the second most common neurodegenerative disease; the pathological hallmark of this disease is the loss of dopaminergic neurons in the substantia nigra¹. Given the relatively small number of neurons involved and their restricted distribution in the brain, cell replacement therapy has been explored as a therapeutic strategy, particularly for late-stage PD patients^2,3,4. However, the outcomes of clinical trials with transplantation of dopaminergic neurons have been inconsistent^3,5. This discrepancy could be attributed to several factors, particularly the survival of transplanted neurons³. In animal studies, the percentage of dopaminergic neurons that survive transplantation is typically less than 15%^6,7,8,9,10. With the aim of improving the cell viability of grafted neurons, several strategies have been attempted, such as the use of the glial cell line-derived neurotrophic factor (GDNF). In PD animal models the GDNF enhances the survival of endogenous^{11,12,13,14,15,16,17,18,19} or transplanted^20,21 dopaminergic neurons and alleviates motor disabilities. Despite these promising results GDNF administration in humans has yielded mixed results: some patients presented significantly motor improvement, others showed a minor benefit whereas others reported off-state dyskinesias^{3,22,23,24,25}. These reports reflected the high degree of variability in the therapeutic outcomes of using this trophic factor to alleviate symptoms in PD patients²⁶. These challenges underscore the need for innovative approaches to provide GDNF-like support to transplanted dopaminergic neurons.

Supramolecular nanofibers offer highly tunable properties that make them well-suited for in vivo applications. Their chemical structure can be engineered to include peptide sequences that mimic the bioactive regions of biological molecules. Our laboratory has pioneered the development of supramolecular nanostructures based on peptide amphiphiles (PAs) that display bioactive peptide sequences that mimic ECM proteins^27,28,29 as well as growth factors^30,31,32. Typically, these self-assembling molecules developed in the Stupp laboratory contain a peptide terminal domain that is bioactive and internal domains that program molecules to form high aspect ratio and water-soluble nanofibers. These domains are generally sequences that promote β-sheet secondary structure as well as charged amino acids for solubility. The molecular structure also includes aliphatic tails covalently bound to the peptide sequence to drive hydrophobic collapse in aqueous media³³. These one-dimensional nanostructures, which may be cylindrical or ribbon-shaped, mimic fibrils present in natural extracellular matrices (ECMs)³³. In a recent study, our laboratory demonstrated that when a brain-derived neurotrophic factor (BDNF) mimetic peptide was displayed on a peptide amphiphile supramolecular nanostructure, its bioactivity was preserved in mouse cortical neurons³². Additionally, PA systems offer advantages over other alternatives, as their chemical structure can be tailored to mimic the mechanical properties of specific tissues³⁴.

To engineer a supramolecular nanostructure capable of replicating the trophic effects of GDNF on dopaminergic neurons we drew upon prior work by Nielsen et al., who reported the design of GDNF peptide mimetics with the aim of interacting with the neural cell adhesion molecule (NCAM)³⁵. Based on prior studies^36,37,38, Nielsen designed these peptide sequences to cover the finger regions of GDNF, which have been shown to be critical for binding to the GFRα1 co-receptor. Among the sequences described, the sequence ETMYDKILKNLSRSR showed the highest bioactivity by promoting neurite growth in embryonic hippocampal neurons and induced the phosphorylation of the FGFR1 receptor³⁵. Several truncated versions of this sequence were also bioactive. We decided to incorporate the shorter peptide sequence ILKNLSRSR on the structure of PA nanofibers and investigated its bioactivity.

Results

Design and characterization of GDNF mimetic peptide amphiphile nanostructures

We designed a GDNF-mimetic PA by incorporating the sequence ILKNLSRSR at the C terminus of the peptide sequence so that it could be displayed on the surfaces of supramolecular nanostructures upon self-assembly of the monomers. To test the specificity of the bioactive peptide sequence, we also synthesized a PA with a scrambled version of the bioactive sequence (LRNKSRILS) at the C terminus of the PA chemical structure, referred to as the scrambled PA (Scr PA). The PA containing lacking a mimetic peptide sequence consists of a 16-carbon fatty acid tail (palmitic acid), followed by two valine residues, two alanine residues (referred to as the β-sheet domain) and two glutamic acid residues (charged domain) (C₁₆V₂A₂E₂; referred to as E2PA) (Fig. 1a). The bioactive PA nanostructure is shown schematically (Fig. 1b) (see supplementary Figs. 1-14 for PA and peptide characterization). To determine the co-assembly ratio of E2 PA and the GDNF PA that provides an optimal biological effect, we mixed the two PAs at different volume percentages (vol %) to create supramolecular copolymers (Supplementary Table 1). All samples were prepared at 1 weight percent (wt%) (see Methods section for additional details). Representative cryogenic transmission electron micrographs (Cryo-TEM) of 15 vol % GDNF PA, 15 vol % Scr PA and E2 PA (Fig. 1c) revealed the formation of ribbon-shaped supramolecular nanofibers. This was confirmed by atomic force microscopy (AFM) imaging, which shows that E2 PA, 15 vol % Scr PA and 15 vol % GDNF PA have very similar twisted nanoribbons (Supplementary Fig. 14a). We found that the 15 vol % co-assemblies formed longer nanoscale ribbon-like fibers relative to filaments formed by 100% of the Scr and GDNF PAs (compared with Supplementary Fig. 15a). The 100% GDNF PA showed a wide distribution of nanoribbon lengths and 100% Scr PA showed mostly short ribbons that tended to aggregate (Supplementary Fig. 15a), whereas the E2 PA backbone alone formed long nanoribbons (Fig. 1c), which is consistent with previous publications by our group^29,32. In addition, we visualized the formation of nanofibers at different volume percentages (80%, 50%, or 33%) between the GDNF PA and Scr PA with the non-bioactive E2 PA (Supplementary Fig. 15a, Supplementary Table 1). To better understand filament formation at the different volume ratios of Scr PA or GDNF PA with E2 PA, we performed synchrotron experiments including small-angle X-ray scattering (SAXS) and wide-angle X-ray scattering (WAXS). We evaluated the Scr PA and the GDNF PA at 100% (without the diluent E2 PA) and co-assemblies containing the following volume percent ratios with respect to the E2 PA diluent: 75, 50, 25 and 15 of Scr PA or GDNF PA as well as E2 PA alone. The pure PAs and their co-assemblies all formed fibers. The SAXS profiles show differences in the scattering minima and broadening of the features with greater than 50% of the GDNF PA or Scr PA (Fig. 1d). Above this concentration, these materials exhibited aggregation and polydispersity (consistent with TEM imaging)³⁹. We also used wide-angle X-ray scattering (WAXS) to characterize the intermolecular packing within the supramolecular fibers, E2 PA, Scr PA and GDNF PA (Fig. 1e) at the different co-assembly ratios. WAXS revealed a peak consistent with β-sheet hydrogen bonding in 100, 75, 50, 25 and 15 vol % samples. The WAXS diffraction shows reduced crystallinity in the samples with the addition of the GDNF PA or Scr PA. However, the samples do not lose the β-sheet character³⁹. To further characterize the β-sheet structure of 15 vol % co-assembly we performed FT-IR and CD spectroscopies (Supplementary Fig. 14b, c), which corroborated the SAXS and WAXS data (Fig. 1d, e).

Fig. 1: Structural characterization of GDNF mimetic supramolecular nanostructures.

a Chemical structures of E₂ backbone PA, Scr PA and GDNF PA. b Molecular graphics representation of the bioactive supramolecular nanostructure with colors corresponding to the various peptide domains shown in (a) (hydrophobic tail (black), β-sheet domain (red), charged domain (blue) and bioactive domain (green)). This image was generated using Cinema 4 d and ePMVREF. c Cryo-TEM images of E₂ PA, 15 vol% Scr PA and 15 vol % GDNF PA. d Small-angle X-ray scattering patterns of nanostructures containing different co-assembly v/v ratios of Scr PA and GDNF PA in the range 15–100 vol %. e Wide angle X-ray scattering scans shown as intensity versus the scattering vector q (log–log plot) of Scr PA and GDNF PA at 15–100 vol %. f Storage modulus of E₂ PA, 15 vol % Scr PA and 15 vol % GDNF PA. *p < 0.05. One-way ANOVA followed by Tukey´s multiple comparison test. Data are presented as mean ± standard error of the mean. g SEM images of neurons seeded on E₂ PA, 15 vol % Scr PA and 15 vol % GDNF PA coatings. Images shows the interaction of neurons and neurites with the nanofibers.

We next evaluated the mechanical properties of gels formed by 15 vol % Scr PA and 15 vol % GDNF PA co-assemblies and E2 PA, since it has been demonstrated that mechanical properties are critical for the regulation of several biological mechanisms such as synaptic activity⁴⁰ and neuronal growth or branching⁴¹. Rheological analysis indicated that the storage modulus of gels formed by E2 PA was 4.6 ± 0.2 kPa, while co-assemblies with 15 vol % Scr PA and 15 vol % GDNF PA showed storage moduli of 8.7 ± 0.5 kPa and 6.9 ± 1.2 kPa, respectively (Fig. 1f and Supplementary Fig. 16), which are all in the same range as the moduli of human brain tissue⁴².

Next, we used scanning electron microscopy (SEM) to visualize the interaction of neurons with the nanofibers E2, 15 vol % Scr PA and 15 vol % GDNF PA coatings (Fig. 1g) (see Methods section for PA coating procedure). We observed that all samples generated homogeneous coatings (Supplementary Fig. 15b). Low magnification SEM images showed that neurons cultured on the top of E2 PA, 15 vol % Scr PA and 15 vol % GDNF PA formed networks through contacts with other neurons. We observed that neurons seeded on Scr PA and GDNF PA at 15 vol % had a spread morphology compared with cells on E2 PA, which had a rounded morphology (Supplementary Fig. 15c).

Biological response of hiPSC-derived dopaminergic neurons treated with GDNF mimetic PA

To determine the optimal combination of E2 PA and GDNF PA, we coated coverslips³² with different ratios of E2 PA/GDNF PA (see methods section for details) (Supplementary Fig. 17a). We wanted to evaluate first the effect of GDNF PA nanofibers on cell survival, since this growth factor promotes viability of dopaminergic neurons⁴³. The dopaminergic neurons used in this study are derived from human-induced pluripotent cells (hiPSC). The percentage of dopaminergic neurons obtained is approximately 90% (Cellular Dynamics). Therefore, in experiments where we do not show a dopaminergic marker (such as the cell viability assay) we assumed that 10% of the cells are not dopaminergic neurons.

We seeded the cells on PA coatings at the following co-assembly volume percent ratios: 10, 15 and 33 vol % (for additional information on co-assemblies concentration see Supplementary Table 1) of GDNF PA with the E2 backbone PA and assessed viability using Calcein AM and propidium iodide staining at 2 days in vitro (DIV) and 7DIV (Supplementary Fig. 17b, c). These experiments included the following controls: E2 PA coating, E2 PA coating + 100 ng/ml peptide (ILKNLSRSR) (92 nM), E2 PA coating + 3.3 nM peptide, E2 PA coating + 100 ng/ml GDNF protein (3.3 nM) (calculated as homodimer) and E2 PA coating + 10 ng/ml GDNF protein. We decided to use 10 ng/ml and 100 ng/ml of GDNF protein because previous reports indicated that RET, the canonical receptor associated with GDNF in dopaminergic neurons phosphorylates at concentrations in the range of 10 to 100 ng/ml of GDNF protein⁴⁴. We observed that at 2DIV and 7DIV (Supplementary Fig. 17d, e) only the 15 vol % GDNF PA (2DIV 52.57 ± 3.91%, 7DIV 60.47 ± 4.42%) co-assembly reached similar percentages of cell viability to the E2 PA coating with 10 (2DIV 39.27 ± 1.50%, 7DIV 55.34 ± 3.46%) or 100 ng/ml GDNF (2DIV 40.87 ± 1.86%, 7DIV 58.39 ± 3.85%) protein dissolved in the media and to Poly-D-Lysine (PDL) coatings (2DIV 62.02 ± 2.39%, 7DIV 72.65 ± 4.79%), which is the most widely used molecule to promote neuronal adhesion⁴⁵. The other conditions showed significantly lower cell viability. Co-assemblies at 10 (2DIV 29.67 ± 3.59%, 7DIV 44.69 ± 5.02%) and 33 vol % (2DIV 44.40 ± 4.51%, 7DIV 43.37 ± 0.86%); Scr PA at 10 (2DIV 24 ± 2.4%, 7DIV 28.5 ± 2.7%) 15 (2DIV 22.3 ± 3.7%, 7DIV 27.8% ± 2.6) and 33 (2DIV 13.1 ± 3.5%, 7DIV 32.8 ± 2) vol% and E2 PA (2DIV 26.6 ± 2%, 7DIV 34.4% ± 1.9), E2 PA+ Peptide 100 ng/ml (2DIV 30.3 ± 1.6, 7DIV 31 ± 2.1%) and E2 PA+ Peptide 3.3 nM (2DIV 31.8 ± 2.5%, 7DIV 36.2 ± 2.4%). We conclude from this data that 15 vol % GDNF PA co-assembly (0.78 mM of GDNF PA) is able to mimic the trophic effect of GDNF compared to other co-assembly ratios.

In addition to promoting the survival of dopaminergic neurons, GDNF is also known to enhance neurite outgrowth⁴⁶. To determine whether GDNF PA nanostructures can reproduce this effect, we cultured hiPSC-derived dopaminergic neurons on different PA coatings and measured the dopaminergic neurite outgrowth at 2DIV and 7DIV (Supplementary Fig. 18). Tyrosine hydroxylase (TH) is the rate-limiting enzyme for dopamine synthesis and is commonly used as a marker for dopaminergic neurons, and we therefore used an anti-TH antibody to stain the cells. At 2DIV the dopaminergic neurons seeded on PDL substrates (142.14 ± 15.7 µm), 15 vol % GDNF PA (112.81 ± 12.7 µm) and E2 PA + 100 ng/ml (95.6 ± 8.2 µm) exhibited significantly longer TH+ neurites relative to the other conditions (E2 PA 23.36 ± 2 µm, E2 + Peptide 100 ng/ml 26.54 ± 2 µm or 3.3 nM 32.33 ± 2 µm, E2 PA + 10 ng/ml 50.7 ± 4 µm, GDNF PA vol 33% 62.9 ± 15.13 µm and 10% 22.9 ± 1.47 µm, and Scr PA vol 33% 32.2 ± 3 µm 15% 41 ± 5.3 µm and 10% 21.9 ± 2 µm) (Supplementary Fig. 18c). After 7DIV we found that cells on PDL (498.6 ± 34.7 µm) extended significantly longer neurites followed by the 15 vol % GDNF PA co-assembly (382.9 ± 2.8 µm), while the cells exposed to the other conditions displayed shorter TH+ neurites (E2 PA 143.8 ± 10.3 µm, E2 + Peptide 100 ng/ml 156.28 ± 11 µm or 3.3 nM 167.65 ± 11.7 µm, E2 PA + 10 ng/ml 209.4 ± 15.1 µm, GDNF PA vol 33% 322.3 ± 24.3 µm and 10% 207.9 ± 16 µm, and Scr PA vol 33% 109.4 ± 8.2 µm 15% 139.8 ± 16.6 µm and 10% 160.4 ± 11 µm) (Supplementary Fig. 18d).

We hypothesized that at 15 vol % ratio of the two molecules in the supramolecular nanostructures there is a synergistic effect of the bioavailability of the mimetic peptide sequence displayed on the nanostructure surface and the topographical features of the PA nanofibers. This is supported by previous studies which demonstrated that the substrate topography influences neurite growth^47,48. In addition, our group has reported that the topographical distribution of the epitope at the nanoscale influences its bioactivity⁴⁹.

GDNF promotes the morphological differentiation of ventral midbrain dopaminergic neurons with longer neurites and more complex branching of TH+ fiber⁵⁰. To further characterize the branching complexity, we performed a Sholl analysis⁵¹. We used color-scale images of neurons seeded on different substrates to obtain the number of intersections within concentric circles (Supplementary Fig. 18e). We found that the cells seeded on the 15 vol % GDNF PA co-assembly coating had significantly more TH+ dendritic branches within 300 µm of the cell soma relative to the other conditions (Supplementary Fig. 18f). From 300 µm to 365 µm, PDL showed similar values to the 15 vol % GDNF PA co-assembly coating and at 400 µm significant differences were not observed among the groups.

Although previous studies reported that GDNF protein induced dopaminergic neurite branching after 7DIV^50,52, we did not observe this effect in our work. Possibly, the hiPSC-derived dopaminergic neurons take longer to spread their dendrites compared to the rodent primary cultured dopaminergic neurons used in previous studies. This is also consistent with a prior report that demonstrated that dopaminergic neurons derived from human embryonic stem cells (hESC) take three times longer to grow relative to dopaminergic neurons derived from rat embryos⁵³. Earlier work that described the neurite branching effect used dopaminergic neurons derived from rodent embryos, whereas we are using human iPSC-derived dopaminergic neurons. This could explain why we did not observe the GDNF protein branching effect. This data suggests that GDNF PA could accelerate the growth factor branching effects. We speculate that the topography of the nanofibers induced an additive effect besides mimicking the GDNF. This is supported by previous publications that demonstrated the effect of substrate topography on neurite morphology, in particular elongation and branching^54,55. However, we wanted to evaluate the biological effects of GDNF PA without the topography influence of nanofibers. Therefore, we seeded the cells on coverslips coated with poly L-ornithine (PLO) and laminin (Lam) as recommended by the cell manufacturer, and then we added the PA on top (see methods for details) (Fig. 2a). We treated the cells with E2 PA, E2 PA + 100 ng/ml peptide, E2 PA + 10 ng/ml GDNF protein, E2 PA + 100 ng/ml GDNF protein, 15 vol % Scr PA, 15 vol % GDNF PA, and PLO+Lam. We evaluated cell viability and dopaminergic neurite length (Fig. 2b, d), and we found that at 2DIV, the 15 vol % GDNF PA (59 ± 3%) promoted neuron survival similar to PLO+Lam (66 ± 3%) and significantly higher than the other conditions (E2 PA 34 ± 4%, E2 PA + Peptide 45 ± 3%, E2 PA + GDNF 10 ng/ml 40 ± 4% or 100 ng/ml 44 ± 4% and 15 vol% Scr PA 32 ± 4%) (Fig. 2c). In addition, 15 vol % GDNF PA (220 ± 22 µm) induced comparable neurite length to PLO+Lam (216 ± 24 µm) and E2 PA + GDNF 100 ng/ml (242 ± 42 µm); the other conditions showed significantly lower lengths (E2 PA 66 ± 10 µm, E2 + Peptide 95 ± 18 µm and 15 vol % Scr PA 50 ± 4 µm) (Fig. 2e). The consistency of these results with the PA coating experiments supports our conclusion that GDNF PA can improve cell viability and dopaminergic neurite extension, whether it is utilized as a substrate or as nanostructures on the top of the cells and not as an adhesive substrate. It is known that GDNF must bind to a GFRα1 receptor to facilitate the binding to the receptor tyrosine kinase (RET) leading to its dimerization and autophosphorylation⁴⁴. Thus, we carried out Western Blot experiments to evaluate the phosphorylation of the RET receptor using the following treatments: starvation media (Strv), 100 ng/ml GDNF protein (3.3 nM), 100 ng/ml peptide (92 nM), E2 PA, 15 vol % GDNF PA co-assembly and Scr PA (10.73 mM of PA concentration, 3.58 mM final concentration) (Supplementary Table 1 for PA concentrations details). A previous publication showed RET phosphorylation by GDNF after 10 min; therefore, we added the treatments at this time⁴⁴ (Fig. 2f, Supplementary Fig. 25) (see methods for details) (see Supplementary Table 1 for PA concentrations details). The 15 vol % GDNF PA co-assembly induced similar values of phosphorylation of RET and the intracellular ERK1/2 relative to 100 ng/ml of GDNF protein (Fig. 2g–h). This is consistent with a previous report that showed the administration of GDNF in the striatum of mice induces the phosphorylation of the ERK1/2 pathway⁵⁶. After treatment with the bioactive peptide, we did not observe any effect on the activation of the RET receptor. In contrast, the peptide displayed by the supramolecular nanostructures was found to induce RET phosphorylation. This is consistent with previous findings from our laboratory that the presentation of BDNF and FGF2 mimetic peptides on PA supramolecular nanofibers induced the phosphorylation of TrkB and FGFR1, respectively, while treatment with the soluble mimetic peptides alone did not activate the receptors^31,32. A possible explanation could be the optimal adaptive configurations of bioactive peptides displayed by PA nanofibers, which favor the interaction with cell receptors thus enhancing their bioactivity relative to mimetic peptides.

Fig. 2: Effects of GDNF PA on hiPSC-derived dopaminergic neurons.

a Schematic diagram illustrating the addition of PA nanofibers. This image was generated using Adobe Illustrator. b Representative images of neurons treated with PLO+Lam, E₂ PA + 100 ng/ml GDNF protein, E₂ PA + 10 ng/ml GDNF protein, E₂ PA + 100 ng/ml peptide, E₂ PA, 15 vol % GDNF PA, 15 vol % Src PA stained for Calcein AM (green) and propidium iodide (red) at 2DIV. c Percentage of viable neurons at 2DIV. *p < 0.05, ****p < 0. One-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean. d Representative images of neurons immunostained for TH⁺ (green) and Tuj1⁺ (orange). e TH+ neurite length of neurons at 2DIV. ****p < 0.0001. One-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean. (experiments in triplicate, n = 200 cells) (f) Western blot of p-RET, total RET, p-ERK and total ERK 10 min post-treatment in neurons treated with starvation (Strv) media, 100 ng/ml GDNF protein, 100 ng/ml peptide, E₂ PA, 15 vol % GDNF PA and 15 vol % Scr PA. Densitometry analysis of the ratio of p-RET/total RET (g), and pERK1/2/total ERK1/2 (h). **p < 0.01, *** p < 0.001, ****p < 0.0001. One-way ANOVA followed by Tukey post hoc test; data are presented as mean ± standard error of the mean (experiments were done in triplicate). Confocal (i) and 3D-rendered image (j) showing the internalization of labeled 15 vol % GDNF PA-TAMRA (red) into neurons stained with Cholera toxin B (cyan) after 10 min, arrowheads show GDNF PA inside the neuron.

As mentioned above, GDNF interacts with GFRα1 to facilitate the signaling through RET; when this occurs the GDNF-GFRα1 complex is internalized⁵⁷. Consequently, we assessed whether neurons could incorporate the PA nanostructures. We treated 2DIV and 4DIV neurons with a 15 vol % labeled GDNF PA (C16-VVAAEEK (TAMRA) ILKNLSRSR) for 10 min and 1 hour, and we used cholera toxin B to label the cell membrane of neurons. In an effort to enhance the visualization of GDNF PA nanostructures inside the neurons we used the Imaris Software to 3D-render the confocal images. In all the conditions we observed GDNF PA nanostructures inside the neurons (Fig. 2i, j, Supplementary Fig. 19). To monitor the interaction between neurons and GDNF PA in real-time, we utilized confocal live imaging, and we observed the movement of PA nanostructures into the neurons (Supplementary Video 1 and 2). The experiments revealed that GDNF PA nanostructures are internalized by hiPSC-derived dopaminergic neurons within 10 min and 1 hour of exposure. Earlier work conducted in our laboratory demonstrated the cellular uptake of nanofibers lacking any bioactive motif⁵⁸. These findings suggest that the internalization of GDNF PA nanostructures occurs independently of the presence of the GDNF mimetic sequence.

Pharmacological inhibition of RET and NCAM signaling on hiPSC-derived dopaminergic neurons treated with GNDF and GDNF PA

Signaling of the RET receptor by GDNF is known to mediate cell survival, neuroprotection, and neurodegeneration in dopaminergic neurons¹⁹. In addition, it has been shown that GDNF signaling of the NCAM also promotes neurite outgrowth in dopaminergic neurons⁵⁹. To elucidate the connection between the observed cell survival response and dopaminergic neurite extension triggered by GDNF PA via RET or NCAM signaling pathways, we treated the hiPSC-derived dopaminergic neurons with RET and NCAM inhibitors. We administered Vandetanib (V) (1 µM) which has been previously used in a rodent model of PD to inhibit RET⁶⁰. To inhibit NCAM pathway we used PP2 (P) (1 µM) which has been shown to block the downstream signaling protein Fyn⁶¹. Then, we added 100 ng/ml GDNF or 15 vol% GDNF PA (on top of the cells). Although, we found that both inhibitors decreased viability, the cells treated with RET inhibitor showed significantly lower survival with respect to the other conditions (GDNF + P 46.81 ± 2.81%, GDNF + V 29.12 ± 2.21%, GDNF PA + P 36.36 ± 3.46%, GDNF PA + V 24.05 ± 1.99%, GDNF protein 66.47 ± 2.29% and, GDNF PA 57.80 ± 2.96%) (Fig. 3a, b). In agreement with these results, we observed that the RET inhibitor reduced phosphorylated RET and ERK 1/2 (Fig. 3e–g). These data are also consistent with previous studies which demonstrated that the effect of GDNF on cell survival is mediated by the RET receptor^19,62,63. In addition, we found that dopaminergic neurite length was significantly reduced with RET and Fyn inhibitors (GDNF + P 120 ± 9 µm, GDNF + V 64 ± 6 µm, GDNF PA + P 71 ± 9 µm, GDNF PA + V 54 ± 6 µm) compared to GDNF (215 ± 19 µm) and GDNF PA (226 ± 26 µm) (Fig. 2c). Dopaminergic neurite length inhibition by PP2 is in agreement with a previous study which demonstrated the role of NCAM-Fyn signaling in neurite outgrowth in dopaminergic neurons⁵⁹. The PP2 inhibitor was found to significantly reduce the phosphorylation of Fyn protein as expected (Fig. 3e, h). Thus, the neurite length reduction caused by the RET inhibitor could be an indirect effect of poor cell viability. These data indicates that the enhanced cell viability observed upon treatment with the GDNF PA is mainly mediated by the RET pathway, whereas the NCAM pathway plays a more important role in the enhancement of dopaminergic neurite length.

Fig. 3: Effects of RET and NCAM on hiPSC-derived dopaminergic neurons treated with GDNF and GDNF PA.

a Representative images of neurons treated with GDNF protein (100 ng/ml) and 15 vol % GDNF PA, PP2(P) (Fyn inhibitor) (1 µM) + GDNF or GDNF PA, and Vandetanib(V) (RET inhibitor) (1 µM) + GDNF or GDNF PA stained for Calcein AM (green) and propidium iodide (red) at 3DIV. b Percentage of viable neurons treated with previous conditions. ****p < 0.0001. One-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean. c Confocal images of dopaminergic neurons treated with the conditions mentioned previously stained with TH antibody (white). The images below correspond to higher magnification views of the areas indicated by the yellow boxes. d Neurite length of dopaminergic neurons. ****p < 0.0001. One-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean (experiments were done in triplicate, n = 200 cells). e Western blot of p-RET, total RET, p-ERK, total ERK and Fyn 10 min post-treatment in neurons treated with the conditions mentioned in (a). Densitometry analysis of the ratio of p-RET/total RET (f), pERK1/2/total ERK1/2 (g) and pFyn/Fyn (h). *p < 0.05, **p < 0.01, *** p < 0.0001, ****p < 0.0001. One-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean (experiments were done in triplicate).

Transcriptomic profile and electrophysiological maturation of hiPSC-derived dopaminergic neurons treated with GDNF PA

To elucidate the molecular mechanism triggered by GDNF PA nanostructures, we performed RNA sequencing (RNAseq) analysis (see Methods for details) of hiPSC-derived dopaminergic neurons cultured for 4DIV on 15 vol % GDNF PA and 15 vol % Scr PA coatings. Analysis of differentially expressed genes (DEGs) comparing 15 vol % GDNF PA (n = 4 libraries) vs 15 vol % Scr PA (n = 4 libraries) revealed 141 upregulated genes (Fold change cutoff +1, P-value < 0.05), whereas 228 genes were downregulated (fold change cutoff -1and a P-value < 0.05) (Fig. 4a, Supplementary Dataset 1). We analyzed the top 20 most significant up and downregulated transcripts (Supplementary Fig. 20a). The gene SULT1B1, which encodes for one isoform of sulfotransferase enzymes, was among the most highly expressed genes by hiPSC-derived dopaminergic neurons cultured on 15 vol % GDNF PA. These enzymes are implicated in neuronal cell viability⁶⁴ and, interestingly, are downregulated in PD patients⁶⁵. We also found the upregulation of the progesterone receptor gene (PGR); it has been shown that this steroid has a neuroprotective effect in PD models in vitro⁶⁶ and in vivo⁶⁷. The myosin heavy chain4 (MYH4) also known as MYHC IIb was also upregulated, and this gene plays a critical role in the morphological maturation of dendrites⁶⁸ and growth cones⁶⁹ and in synaptic function⁶⁸. We also found the upregulation of TAC3 which encodes the neuropeptide NKB/NK3. This neuropeptide is found in the substantia nigra⁷⁰ and plays an important role in the modulation of dopamine release⁷¹. WNT3A was also significantly upregulated which encodes for a glycoprotein that participates in the development of dopaminergic neurons⁷² and also has a neuroprotective effect against 6-OHDA⁷³. To evaluate the Gene Ontology (GO)⁷⁴ in the synaptic context, we performed an enrichment analysis of the protein-coding DEGs using the gene library SynGO⁷⁵. We found that the postsynaptic specialization GO term was highly enriched (1.47, P-value 0.03375) (Supplementary Table 3) whereas the neurotransmitter uptake (1.76, P-value 0.01703) and reuptake (1.93, P-value 0.01162) GO terms were significantly downregulated (Fig. 4b) (Supplementary Table 3). This is in accordance with the role of GDNF enhancing the establishment of synaptic terminals in dopaminergic neurons⁷⁶. We further analyzed the molecular function GO terms upregulated by GDNF PA nanostructure through the g:Profiler web⁷⁷ (Fig. 4c). Additionally, we compared the upregulated genes by GDNF PA with upregulated genes reported previously by Gantner et al.⁷⁸. In that study, the authors evaluated the gene profile of human pluripotent stem cells-derived ventral midbrain dopamine progenitor grafts into rodent brains exposed to GDNF⁷⁶ (Supplementary Fig. 18c). Curiously, we found the overlapping of genes related to GDNF signaling, such as USP50 and SPDYE2. According to the human gene database GeneCards (https://www.genecards.org/) USP50 is a ubiquitin peptidase. It has been reported that the intracellular domain of RET’s receptor is ubiquitinated after GDNF-RET interaction⁶³. Also, we found that the SPDYE2 gene overlapped; GeneCards database reports that this gene favors protein kinase binding activity. As described previously, the interaction between GDNF and the RET receptor is mediated by kinase activity⁴⁴. We also noted that the IGFN1 gene was overlapped; notably the GeneCards database indicates that this gene is involved in cell adhesion and synapse formation.

Fig. 4: Bulk-RNA transcriptomics and electrophysiological characterization of hiPSC-derived dopaminergic neurons treated with GDNF PA.

a Volcano plot depicting up- and downregulated genes of human dopaminergic neurons cultured on 15 vol % GDNF PA vs. 15 vol % Scr PA for 4DIV (Fold change cutoff ±1, P-value < 0.05). Red dots indicate upregulated genes whereas blue dots indicate downregulated genes. Y-axis denotes −log₁₀ P values while X-axis shows log₂ fold change values. b Bar graph of the most significant up (red) and down (blue) enriched GO terms of genes expressed in human dopaminergic neurons cultured on 15 vol % GDNF PA vs. 15 vol % Scr PA for 4DIV using the SynGO library. c Most significant molecular function GO terms of the upregulated protein-coding genes of human dopaminergic neurons cultured on 15 vol % GDNF PA vs. 15 vol % Scr PA for 4DIV (P-value < 0.05) (n = 4 libraries). Images and representative recordings of cell-attached patched hiPSC-derived dopaminergic neurons seeded on 15 vol % Scr PA (d) and 15 vol % GDNF PA (e) and PDL with GDNF protein (100 ng/ml) (f). Bar graph showing the percentage of cells that revealed spontaneous activity (g). Box plots depicting firing rate (h) and coefficient of variation (CV) (i) of cells seeded on 15 vol % Scr PA and 15 vol % GDNF PA (one-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean) (experiments were done in triplicate, GDNF PA n = 30, Scr PA n = 24 and GDNF protein n = 30 cells).

Next, we investigated if the GDNF PA coating could induce electrophysiological maturation of hiPSC-derived dopaminergic neurons. Previous reports demonstrated that increasing the firing rate in dopaminergic neurons elevates the somatodendritic release of dopamine⁷⁹ and produces an increase in the dendritic calcium signal⁸⁰. This rise in calcium signal changes the propensity of dopamine neurons to generate bursts⁸⁰. It has been described that GDNF also enhances the spontaneous firing rate of midbrain dopaminergic neurons in ex-vivo preparations^81,82,83. To determine whether the GDNF PA can mimic this effect of native GDNF, we studied its impact on the spontaneous firing rate of hiPSC-derived dopaminergic neurons. We cultured dopaminergic neurons on 15 vol% Scr PA, 15 vol% GDNF PA, and PDL with 100 ng/ml GDNF. After 30–35 DIV, we recorded spontaneous activity using cell-attached patch clamp (Fig. 4d–f). We found the 15 vol % GDNF PA co-assembly coating increased the percentage of cells that presented spontaneous firing activity in contrast to the other conditions (15% GDNF PA coating: 67% (n = 24), 15 vol % Scr PA coating: 50% (n = 30) and GDNF protein: 33% (n = 30)) (Fig. 4g).

The neurons seeded on GDNF PA and PDL with GDNF protein presented similar firing rate, in contrast with those cells seeded on the Scr PA (15% vol GDNF PA coating: 6.6 Hz (n = 16), 15% Scr PA: 3.8 Hz, (n = 16), p = 0.009) (Fig. 4h). Regarding the coefficient of variation, we only found a significant difference between GDNF PA with respect to Scr PA (0.4 vs. 0.5, p = 0.02) (Fig. 4i). These results demonstrate that GDNF PA has the same effect of increasing the firing rate of dopaminergic neurons as native GDNF protein, which is supported by previous reports^81,82.

hiPSC-derived dopaminergic neurons in a three-dimensional culture of GDNF PA

To further develop the bioactive scaffold for cell replacement applications, we evaluated the interaction of the nanostructures with the cells in three-dimensional culture. We encapsulated the cells by mixing them with the PA solution followed by the addition of cell media (see methods for details) (Fig. 5a). Using multiphoton microscopy, we observed that the cells were homogeneously distributed within the scaffolds which are 500 µm thick (see Z-depth color scale in Fig. 5b). To compare the 15 vol % Scr PA and the 15 vol % GDNF PA scaffolds, we encapsulated hiPSC-derived dopaminergic neurons and stained the cells with anti-TH and anti-MAP2 antibodies at 15DIV (see Fig. 5c, d for the merged images and Supplementary Fig. 21a–d for the separate color channels). We quantified the number of TH+ cells (dopaminergic neurons) and MAP2+ cells (matured neurons) per mm³. We did not find differences between the conditions for cells stained with the neuronal marker MAP2; however, for TH + /MAP2+ double-stained cells we observed significantly more cells per mm³ in the GDNF PA condition (Fig. 5e, f) demonstrating the bioactive effect of GDNF PA relative to the Scr PA. This suggests that the bioactive epitope displayed by the GDNF PA nanofibers induces the survival of the dopaminergic neuronal population, demonstrating the specificity of the GDNF PA. We also calculated cell volume, including the cell soma and the neurites, indicative of more mature neurons. We found a significantly higher volume of TH + /MAP2+ signal in 15 vol % GDNF PA scaffold relative to 15 vol % Scr PA scaffold (Fig. 5g, h). Additionally, we evaluated the distribution of the cells within the gel using the TH+ fluorescence signal and found that in both conditions the cells were homogeneously distributed (Supplementary Fig. 21e). We used Imaris software to reconstruct in three dimensions the morphology of hiPSC-derived dopaminergic neurons embedded in PA scaffolds using multiphoton microscopy images (see methods for details) (Fig. 5i, j) and measured several morphometric parameters. We evaluated the branching complexity by Sholl analysis, and we found that from 2 µm to 70 µm from the cell soma, the cells encapsulated in the 15 vol % GDNF PA scaffold had significantly more intersections compared to the cells on the 15 vol % Scr PA scaffold (Fig. 5k). We also quantified the TH+ neurite length and found that cells embedded in the 15 vol % GDNF PA scaffold had an average value of 744 ± 43.0 µm compared to 394 ± 36.0 µm for cells within the 15 vol % Scr PA (Fig. 5l). The volume of the TH+ fluorescence signal of cells encapsulated in the 15 vol % GDNF PA scaffold was significantly higher (2239 ± 130.0 µm³) relative to cells encapsulated in 15 vol % Scr PA scaffold (1541 ± 113.0 µm³) (Fig. 5m). In addition, we quantified the number of branching and terminal points per TH+ cell (see the labels in Figs. 5i and 5). The TH+ cells embedded in the 15 vol % GDNF PA scaffold had 33 branching points and 37 terminal points, while the cells in the 15 vol % Scr PA scaffold showed statistically lower values with 16 branching points and 17 terminal points per neuron (Fig. 5n, o). This data suggests that the 15 vol % GDNF PA can form a suitable 3D scaffold for hiPSC-derived dopaminergic neurons that promotes morphological maturation compared to a scaffold that does not present the bioactive signal (Scr PA).

Fig. 5: Dopaminergic neuron morphology is influenced by GDNF PA scaffolds.

a Schematic diagram illustrating the procedure for cell encapsulation. This image was generated using Adobe Illustrator. b Three-dimensional reconstruction of a 500 µm-thick sample using multi-photon microscopy of cells encapsulated in PA scaffolds at 15DIV; the color scale bar indicates the depth at which the cells are located within the PA scaffold. 350 µm thick three-dimensional reconstruction of neurons stained with TH and MAP2 antibodies at 15DIV encapsulated in a 15 vol % Src PA scaffold (c) and a 15 vol % GDNF PA scaffold (d). Quantification of MAP2⁺ (e) and TH⁺/MAP2⁺ (f) neurons within PA scaffolds. Volume of the fluorescence signal of MAP2⁺ neurons (g) and TH⁺/MAP2⁺ neurons (h) within PA scaffolds; *p < 0.05; T-test; data are presented as mean ± standard error of the mean. Digital reconstruction of single TH⁺ neuron encapsulated in 15 vol % Scr PA (i) and 15 vol % GDNF PA scaffolds (j). Sholl analysis of TH⁺ neurons encapsulated in PA scaffolds (k); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; one-way ANOVA followed by Tukey post hoc test; data are presented as mean ± standard error of the mean. Quantification of TH⁺ neurite length (l), volume of TH⁺ fluorescence signal (m), number of neurite branching points (n) and terminal points per TH⁺ neuron (o); **p < 0.01, ****p < 0.0001; T-test; data are presented as mean ± standard error of the mean (experiments were done in triplicate, n = 60 cells).

GDNF PA increases synapse formation

Next, we investigated whether the GDNF PA scaffold influences the synaptic connections of encapsulated hiPSC-derived dopaminergic neurons which reflects functional maturation. We decided to quantify the puncta of the pre-synaptic marker synaptophysin in TH+ neurites of cells encapsulated (three-dimensional cultures) in 15 vol % Scr PA scaffold and 15 vol % GDNF PA scaffold (Fig. 6a, b). We found significantly more synaptophysin puncta in cells embedded in 15 vol % GDNF PA scaffold with an average of 35 ± 4 per 100 µm of TH+ neurite compared to cells embedded in 15 vol % Scr PA scaffold with an average of 21 ± 4 per 100 µm of TH+ neurite (Fig. 6e). The midbrain receives glutamatergic inputs which mediate several functions including reinforcement of learning and motor control⁸⁴. In this regard, previous work demonstrated the presence of glutamatergic synapses in dopaminergic neurons⁸⁴. These synapses can be examined by the presence of the post-synaptic marker PSD95⁸⁴. Thus, we evaluated the presence of PSD95 puncta in neurites of dopaminergic neurons encapsulated in 15 vol % Scr PA and 15 vol % GDNF PA scaffolds (Fig. 6c, d).

Fig. 6: Effect of GDNF PA on the expression of synaptic markers of encapsulated hiPSC-derived dopaminergic neurons.

Representative fluorescence micrographs of TH⁺ (green) and (Syn⁺) (red) neurites in hiPSC-derived dopaminergic neurons encapsulated in 15 vol % Src PA (a) and 15 vol % GDNF PA scaffolds (b) (Syn⁺ puncta indicated by arrow heads). Representative fluorescence images of TH⁺ (green) and (PSD95) (red) neurites in hiPSC-derived dopaminergic neurons encapsulated in 15 vol % Src PA (c) and 15 vol % GDNF PA (d) scaffolds (PSD95⁺ puncta indicated by arrow heads). e Quantification of Syn⁺ puncta in TH⁺ neurites of cells encapsulated in 15 vol % Scr PA scaffolds shown in (a) and in 15 vol % GDNF PA scaffolds shown in (b) (*p < 0.05; T-test) data are presented as mean ± standard error of the mean. f Quantification of PSD95 puncta in TH⁺ neurites of hiPSC-derived dopaminergic neurons encapsulated in 15 vol % Scr PA scaffolds shown in (c) and in 15 vol % GDNF PA scaffolds shown in (d) (****p < 0.0001; T-test; data are presented as mean ± SE). g, h Representative confocal images of dopaminergic (white arrowheads) and non-dopaminergic (yellow arrowheads) neurons embedded in 15 vol % GDNF PA. Number of Syn (i) and PSD95 (j) puncta in TH+ and TH- neurites. (****p < 0.0001; T-test; data are presented as mean ± standard error of the mean) (experiments were done in triplicate, n = 60 cells).

Cells embedded in the 15% vol GDNF PA scaffold showed an average of 10.65 puncta per 100 µm of TH+ neurite which was significantly higher than 5.63 of cells embedded in 15 vol % Scr PA (Fig. 6f). To confirm these findings, we also measured the fluorescence intensity of the synaptic makers and found that the fluorescence of synaptophysin and PSD95 in TH+ neurites of cells embedded in 15 vol % GDNF PA scaffold was statistically higher relative to the fluorescence of these markers in TH+ neurites of cells encapsulated in the 15 vol % Scr PA scaffold (Supplementary Fig. 22).

To determine the specificity of GDNF PA in the enhancement of pre- and post-synaptic vesicles on dopaminergic neurons, we quantified the number of Syn+ puncta on TH+ neurites (35.4 ± 3.8) and TH- neurites (11.7 ± 1.7) (Fig. 6g, i). Additionally, PSD95+ puncta in TH+ neurites (10.7 ± 0.8) with respect to TH- neurites (4.3 ± 0.4) (Fig. 6h, j). This is in agreement with a previous study which demonstrated that GDNF improves the synaptic efficacy in dopaminergic neurons⁷⁶. These results are consistent with those presented in Fig. 5e–h, suggesting that the effects of GDNF PA scaffolds are specific to dopaminergic neurons.

Effect of GDNF PA nanostructure on the axonal extension of human midbrain-like organoids (hMLOs)

Brain organoids recapitulate several features of human brain development and can be used to evaluate potential therapeutic strategies⁸⁵. Previous work demonstrates the role of nanostructures in the maturation of human brain organoids^86,87. Therefore, we decided to investigate whether the GDNF PA scaffold promotes the dopaminergic axonal growth in hMLOs following a protocol reported by Jo et al. ⁸⁵ (see methods for details). The hESC cells used to generate the hMLOs were positive for OCT4, Nanog and SOX2 (data not shown). We observed that 79% of the cells colocalize for TH and the neural precursor marker nestin (Fig. 7a, c), whereas only 12.84% were positive for TH at 45DIV (Fig. 7a and d). This means that at this stage most of the dopaminergic neurons were immature. At 60DIV we found that 39.6% of neurons were TH+ and Tuj1+ (neuronal marker), and 27.2% of TH neurons did not express Tuj1 (Fig. 7b, d). After 80DIV we observed an increase of the dopaminergic population to 34.2% (Fig. 7f). Additionally, the dopaminergic neurons expressed the following maturation markers: 74.8% MAP2, 82.4% GIRK2 (midbrain dopaminergic marker), and 73.8% the dopamine transporter marker DAT (Fig. 7e–g). This data indicates that most of the dopaminergic population reached a mature stage, thus we decided to evaluate the effect of GDNF PA on 80DIV organoids. To maintain the organoids in a three-dimensional environment, we encapsulated them within 15 vol % Scr PA or 15 vol % GDNF PA (Fig. 7h, i and Supplementary Fig. 23a) (see methods for details). The organoids were embedded in the PA scaffold for a week, then fixed and stained with TH, Tuj1 and Hoechst dye to visualize the PA (Fig. 7j). We evaluated the dopaminergic axonal growth from the organoid towards the scaffolds and found that hMLOs embedded in the GDNF PA extended their axons to an average length of 311 µm in contrast to 171 µm in organoids embedded in the Scr PA (Fig. 7k–m). In addition, we counted the number of dopaminergic projections per organoid and found an average of 33.2 projections in those embedded in GDNF PA compared to only 4.4 projections when Scr PA was used for encapsulation, the organoids extended their axons at different planes in a range of 500 µm (Supplementary Fig. 23b). These results are in agreement with those obtained when the GDNF nanostructures were utilized in cultures of hiPSC-derived dopaminergic neurons.

Fig. 7: Effect of GDNF PA on the dopaminergic axonal length of hMLOs.

Cryosection of hMLOs at (a) 45 days stained for Nestin (red) and TH (green) and (b) at 60 days stained for Tuj1 (red) and TH (green). Bottom left: a zoom-in view of the white box. Bottom right: a zoom-in view of the bottom left image. c Percentage of cells showing co-localization of TH + /Hoechst+ and Nestin + /TH+ cells at 45 days, and (d) TH + /Hoechst+ and Nestin + /TH+ at 60 days (data are presented as mean ± standard error of the mean). e Immunostaining of hMLOs at 80 days showing the maturation markers MAP2, GIRK2 and DAT, arrowheads indicate the double-positive cells. Percentage of cells showing co-localization of TH + /Hoechst+ and MAP2 + /TH+ (f), GIRK2 + /TH+ and DAT/TH+ (g) (data are presented as mean ± standard error of the mean). h Schematic diagram illustrating the embedding process of the organoid within the PA scaffold. This image was generated using Adobe Illustrator. i Epifluorescence image to show the top-view of an organoid stained for Tuj1 encapsulated within the PA scaffold. j Organoid encapsulated stained for TH and Tuj1. Representative images of organoids stained for TH embedded in 15 vol % Scr PA (k) and 15 vol % GDNF PA (l). Zoom-in view of the white box shows confocal images and digital reconstruction of dopaminergic axons. m TH+ neurite length and (n) number of TH+ projections of hMLOs embedded within PA scaffolds at 7DIV (****p < 0.0001; T-test; data are presented as mean ± standard error of the mean) (experiments were done in triplicate, n = 9 organoids).

Neuroprotective effect of GDNF PA supramolecular filaments

The neurotoxin 6-hydroxydopamine (6-OHDA) is a hydroxylated analogue of dopamine and is commonly used to generate rodent models of PD due to its ability to selectively damage the dopaminergic neurons⁸⁸. It has been demonstrated that the administration in vitro⁸⁹ and in vivo^11,90 of GDNF has a neuroprotective effect on dopaminergic neurons. Therefore, we investigated whether the GDNF PA has a neuroprotective effect both in two-dimensional and three-dimensional cell cultures. First, we cultured the hiPSC-derived dopaminergic neurons on a PDL substrate with starvation media (see methods for details), PDL substrate with basal media, E2 PA coating with 100 ng/ml GDNF protein, E2 PA coating with 100 ng/ml peptide, E2 PA coating, 15 vol % Scr PA coating and 15 vol % GDNF PA coating. After 15DIV we applied 100 µM of the neurotoxin 6-OHDA in the media (Fig. 8a) for 1 hour. We then carried out immunostaining for TH protein (Fig. 8b–h) and measured the mean intensity fluorescence and the neurite length. We found that the cells seeded on 15 vol % GDNF PA coating reached similar values (79.64 ± 8.92 a.u.) of TH+ mean intensity fluorescence to the cells plated on E2 PA coating with 100 ng/ml GDNF protein (80.56 ± 3.70 a.u.) (Fig. 8i). In addition, the length of dopaminergic neurites of cells seeded on the 15 vol % GDNF PA reached similar values (251.25 ± 12.70 µm) to the cells on the E2 PA coating with 100 ng/ml GDNF protein (219.31 ± 19.11 µm) (Fig. 8j). The other conditions showed significantly lower values for both parameters. Moreover, we measured the expression of Tuj-1 (neuronal marker) and TH (dopaminergic neuronal marker) proteins by Western Blot analysis of two-dimensional cultures at the different conditions (Fig. 8k), and the densitometry showed that there was no statistical difference in the amount of Tuj-1 protein after 6-OHDA treatment (Fig. 8l). In contrast, we observed that the TH protein was higher in E2 PA with 100 ng/ml GDNF protein and 15 vol % GDNF PA coating (Fig. 8m), consistent with the immunostaining results. We also evaluated the neuroprotective effect in three-dimensional cultures by encapsulating the hiPSC-derived dopaminergic neurons in scaffolds of E2 PA, 15 vol % Scr PA or 15 vol % GDNF PA and then adding 100 µM of the neurotoxin 6-OHDA in the media after 15DIV (Fig. 8n–q). To estimate the number of dopaminergic somas and neurites after the neurotoxin treatment in three-dimensional culture, we quantified the volume generated by TH+ fluorescence staining and found that the cells encapsulated in the 15 vol % GDNF PA scaffold showed significantly higher volume (33583.9 ± 2482.0 µm³) relative to E2 PA (16340.9 ± 2429.5 µm³) and 15 vol % Scr PA (24295.6 ± 2789.4 µm³) scaffolds (Fig. 8r). In addition, we did a double immunostaining for TH and cleaved caspase 3 (a marker of apoptosis) and measured their co-localization to identify the percentage of dopaminergic neurons that are undergoing apoptosis. We found that the cells seeded on the 15 vol % GDNF PA coating showed 59 ± 2% colocalization compared to 79 ± 4% localization on E2 PA coating and 68. ±3% on the 15 vol % Scr coating (Supplementary Fig. 24a–c, g). We also evaluated the colocalization of these markers in three-dimensional cultures (Supplementary Fig. 24d–f) and found statistically lower colocalization for the 15 vol % GDNF PA scaffold (51.49 ± 4) relative to E2 PA (85.4 ± 5.00) and 15 vol % Scr PA (71.9 ± 4.60) scaffolds (Supplementary Fig. 24h). From the similar results obtained in two- and three-dimensional cultures, we conclude that 15 vol % GDNF PA coatings are neuroprotective against the 6-OHDA neurotoxin. Although a neuroprotective effect of GDNF-loaded hydrogels against 6-OHDA has been previoiusly reported^20,21,91,92, to our knowledge this is the first biomaterial that provides this effect without any encapsulated growth factor.

Fig. 8: Neuroprotective effect of GDNF PA against the neurotoxin 6-OHDA.

a Schematic diagram illustrating the addition of 6-OHDA to the cells seeded on PA coatings. b–h Representative images of hiPSC-derived dopaminergic neurons seeded on PDL with starvation (Strv) media and with basal media or seeded on E₂ PA + 100 ng/ml GDNF protein, E₂ PA + 100 ng/ml Peptide, E₂ PA, 15 vol % GDNF PA and 15 vol % Scr PA, and then treated with 6-OHDA (100 µM) for 1 h, immunostained for TH. i Mean intensity of TH fluorescence and (j) TH⁺ neurite length of cells seeded on the different conditions (***p < 0.001, ****p < 0.0001; one-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean). k Western blot and densitometry analysis at 15DIV of (l) Tuj1 and (m) TH in neurons treated with 6-OHDA for 1 h exposed to conditions in (b–h) (**p < 0.01, ***p < 0.001, ****p < 0.0001); one-way ANOVA followed by Tukey post hoc test; data are presented as mean ± standard error of the mean. n Schematic diagram illustrating the addition of 6-OHDA to the cells encapsulated within PA scaffolds. Representative 3D fluorescence reconstructions of neurons encapsulated within E₂ PA (o) 15 vol % Scr PA and (p) 15 vol % GDNF PA (q) treated with 6-OHDA (100 µM) for 1 h, stained with TH. r Quantification of TH⁺ fluorescence volume of neurons encapsulated (*p < 0.05, **p < 0.01); one-way ANOVA followed by Tukey´s multiple comparison test; data are presented as mean ± standard error of the mean) (experiments were done in triplicate, n = 60 cells).

Discussion

The goal of this study was the design and synthesis of self-assembling nanostructures capable of improving the viability as well as the morphological and functional maturation of human dopaminergic neurons by mimicking relevant trophic effects of GDNF, without aiming to recreate the complete structure or function of the native protein.

Characterization of the supramolecular bioactive filaments revealed a morphology of nanoscale ribbons, which can form hydrogels with stiffness values similar to the extracellular matrix of the human brain⁴². We determined that the co-assembly at 15 vol % for the GDNF PA led to significantly higher bioactivity relative to the other ratios. We hypothesize that, at this volume percentage ratio, the self-assembling configuration facilitates optimal interaction between the bioactive domain of the nanofibers and the dopaminergic neurons, thereby activating the desired biological signaling pathways. This hypothesis is consistent with previous work from our group, which demonstrated that the bioactivity of nanostructures is influenced by the topographical distribution of the epitopes⁴⁹.

Over the past decade, patient-derived hiPSCs have gained interest as therapeutic tools⁹³, with hiPSC-derived dopaminergic neurons now in clinical trials for PD⁴. To explore the therapeutic potential of our nanoscaffold, we used hiPSC-derived dopaminergic neurons, unlike previous studies that relied on rodent embryos^20,91,92 or hESC-derived cells²¹. Cell viability experiments showed that the GDNF PA significantly improved dopaminergic neuron viability to levels comparable to those achieved by the native protein. In contrast, we observed that the non-bioactive nanofibers (E2 PA) showed reduced viability. This aligns with our previous work on BDNF-mimetic PA³², where E2 PA also showed the lowest survival rates.

Additionally, we performed live imaging to determine if the nanofibers were internalized into neurons, as prior studies have demonstrated GDNF-GFRα internalization into endosomal compartments during signaling⁵⁷. Confocal live imaging and 3D-rendering demonstrated the internalization of GDNF PA nanostructures. Nevertheless, previous publications from our group indicated that nanofibers lacking any bioactive region can also be internalized through endocytosis^58,94. We cannot conclusively determine that the internalization of the nanostructures observed in this study is mediated specifically by the bioactive peptide. We observed that the nanofibers located outside the cell form aggregates of varying sizes, some reaching up to 5 µm in diameter. In contrast, nanofibers internalized by neurons were observed in the form of smaller aggregates. This observation suggests the presence of a size threshold for nanofiber internalization. Additional research is required to elucidate the morphology of the internalized nanofibers, their intracellular half-life, the mechanisms responsible for degrading these structures, and whether their bioactivity is preserved post-internalization.

Nanostructures hold significant promise in nanomedicine, including enhancing dopaminergic neuron viability⁹⁵. However, our previous published work has demonstrated that the combination of nanostructures with a bioactive signal significantly increases their biological effects^{29,30,31,32,96,97}. Here, we focused on developing a bioactive nanofiber that promotes dopaminergic viability and maturation beyond its role as a passive scaffold, and we designed experiments to isolate these bioactive effects. To achieve this, we added the GDNF PA on the top of the cells eliminating the mechanical support provided by nanofibers (PA coating and PA scaffolds). Remarkably, the survival effects and enhancement of morphological maturation observed with the GDNF PA scaffolds (PA coating and PA on the top of the cells) were still observed, indicating that its bioactivity is rooted in its ability to engage in signaling pathways. This suggests that while ECM-mimicking properties are significant, the primary effect results from the bioactive components of the GDNF PA nanofiber.

Our findings indicate that GDNF PA has comparable efficacy to the substrate recommended by the cell provider (PLO) in promoting cell viability and dopaminergic neurite extension. Conventional cell-culture matrices, such as PLO, PDL, or laminin, primarily exert their effects in 2D culture and most do not have the structural capacity to generate 3D ECM mimics, limiting their suitability for translational applications in vivo. In contrast, the GDNF nanostructures developed in this work effectively support both 3D single-cell and 3D midbrain organoid cultures. These findings highlight the potential of GDNF PA nanofibers for in vivo applications, offering an advantage over traditional matrices.

To develop a supramolecular nanostructure that emulates the trophic effects of GDNF on dopaminergic neurons, we leveraged prior work by Nielsen et al., who designed GDNF-mimetic peptides intended to engage the NCAM pathway³⁵. Nielsen et al. reported that these peptide sequences were designed to target the finger domains of GDNF, drawing on prior studies that identified these regions as critical for binding to the GFRα1 co-receptor^36,37,38. Previous reports showed that the GFRα1 co-receptor mediates the interaction between GDNF and NCAM⁹⁸. Nielsen et al concluded that the peptide sequence ETMYDKILKNLSRSR mimics the neurite growth effect of GDNF and this was mediated by FGFR and Fyn which are involved in the GDNF-NCAM signaling pathway³⁵. The effect of this peptide was tested on hippocampal neurons which preferentially utilize NCAM in response to GDNF^36,98. In contrast, dopaminergic neurons exposed to GDNF signal through the RET receptor⁶². Similar to the GDNF-NCAM signaling, the interaction between GDNF and RET is mediated by the co-receptor GFRα1⁶². Here, we used the truncated peptide sequence ILKNLSRSR reported by Nielsen et al. which showed similar bioactivity³⁵. We performed Western Blot experiments to investigate the response of dopaminergic neurons to GDNF PA (C₁₆V₂A₂E₂-ILKNLSRSR). In these experiments we dissolved the GDNF PA into the media rather than seeding the cells on the PA coatings or adding PA on top of the cells. This method was chosen to eliminate any effects that could be simply connected to adhesion of the cells to the substrate. Our data indicates that the GDNF PA indeed phosphorylates RET as well as Fyn. To deeply investigate the signaling of GDNF PA, we employed a pharmacological approach using Vandetanib, a potent tyrosine kinase inhibitor that targets RET, VEGFR2, and VEGFR3 and has been utilized in preclinical models of PD⁶⁰. In parallel experiments, we employed the PP2 inhibitor, which targets Src family kinases, including Fyn⁶¹ a downstream effector in the GDNF-NCAM signaling in dopaminergic neurons⁵⁹. Our findings revealed that RET inhibition primarily impacted the viability of dopaminergic neurons, aligning well with previous studies that established a direct link between GDNF-RET signaling and cell survival^19,60,62. This finding is also consistent with prior research demonstrating that localized ablation of RET leads to a progressive loss of dopaminergic neurons in the substantia nigra⁹⁹. Furthermore, we observed a reduction in neurite length in cells treated with the Fyn inhibitor, consistent with earlier reports demonstrating that PP2 impedes neurite outgrowth⁶¹.

Additional experimental data that supports the signaling through RET include the neuroprotective effects of GDNF PA against the 6-OHDA neurotoxin as prior in vivo research has shown that the neuroprotective effect of GDNF on dopaminergic neurons is mediated through RET¹⁹. Specifically, our data indicates that GDNF could be signaling through RET as well as NCAM pathways. This is plausible because the mimetic peptide used was designed to signal the NCAM pathway through the co-receptor GFRα1⁹⁸, which is also known to participate in the RET⁶² pathway. Additionally, the introduction of a supramolecular nanostructure containing the truncated peptide could improve its interaction with the receptor^32,96,97. This is supported by several publications from our laboratory which demonstrated that the combination of a supramolecular nanostructure with a BDNF mimetic peptide enhances the phosphorylation of TRK receptor compared with the mimetic peptide alone^32,96,97. Two other recent papers demonstrate the same effect in a netrin-1 mimic⁹⁷ as well as a TGFβ-1 mimic⁹⁶. We acknowledge the limitations of the pharmacological approach employed in this study and recognize that further experiments are necessary to fully elucidate the signaling mechanisms of the GDNF PA in dopaminergic neurons. Future studies using genetically modified dopaminergic neurons lacking NCAM, RET, or the GFRα1 co-receptor will be instrumental in clarifying the nanostructure’s mechanism of action.

Moreover, we found that the RNA-seq data robustly correlated with the morphological and functional effects of the GDNF nanostructure on dopaminergic neurons. Notably, SULT1B1, encoding a sulfotransferase that inactivates dopamine⁶⁴, showed the highest upregulation. This enzyme is crucial for cell survival by preventing dopamine-induced oxidative stress¹⁰⁰, suggesting a molecular basis for the improved viability observed with GDNF PA. In addition, the GDNF nanostructure upregulated WNT3a and the progesterone receptor (PGR) gene, both of which protect dopaminergic neurons against toxins⁷³ helping to explain the neuroprotective effect of the GDNF PA. MYH4 (MYHC IIb), significantly upregulated, promotes dendritic spine maturation⁶⁸ supporting our findings on synapse formation in neurons within GDNF PA scaffolds. Overall, RNA data indicate that GDNF PA enhances the viability and accelerates the maturation of human dopaminergic neurons.

Recently, brain organoids have been shown to replicate human brain development while maintaining its genetic background, unlike 2D-differentiated neurons⁸⁵. As an emerging model for testing neurological therapies, we generated 80-day hMLOs from hESC using a published protocol⁸⁵ to assess the bioactivity of our nanostructure. Organoids embedded in the GDNF PA extended longer dopaminergic axons than those in the Scr PA, confirming effects observed in hiPSC-derived dopaminergic neuron cultures.

Cell replacement therapies using human dopaminergic neurons derived from hESC⁶ or hiPSCs⁷ have shown motor improvement in PD models but face challenges, including low cell survival rates (2.6–6%). The GDNF-mimetic nanostructure provides an alternative strategy to provide the trophic effects of GDNF. In addition, the tunability of our supramolecular system allows adjustment of mechanical properties to match those of the target tissue³⁴. This flexibility enables control over stiffness and diffusivity without altering the bioactive sequences, providing a customizable platform for in vivo applications. These features could help overcome the limited diffusion observed previously with native GDNF, which restricts its therapeutic efficacy¹⁰¹. The physical tunability of PAs may also be used to enhance the internalization of the signaling complex even in the situation of high cellular uptake-related stress in PD¹⁰². Properly tuned PA systems in combination with transplanted dopaminergic neurons could be considered as a therapy delivered to injection into the striatum or the substantia nigra.

The aim of this study was to develop supramolecular nanostructures which form a 3D-scaffold that mimics key trophic effects of GDNF. In vitro experiments demonstrated that the GDNF PA scaffold offers some of the critical functionalities provided by GDNF protein. Future experiments beyond the scope of this paper will assess its performance in PD animal models involving cell transplantation.

Methods

Peptide synthesis

Peptide amphiphiles (PA) molecules were synthesized on solid support with Rink amide resin via Fmoc-based chemistry on a CEM Liberty Blue automated microwave peptide synthesizer. The automated coupling reactions were performed using Fmoc-protected amino acid (4 equiv.), DIC (4 equiv.), and Oxyma pure (8 equiv.). Fmoc deprotection was completed using 4-methylpiperidine (20% in DMF). The resin-bound peptides were cleaved using a mixture of TFA (95%), water (2.5%), and TIPS (2.5%). The peptides were precipitated with cold ether followed by HPLC purification using a Waters Prep 150 LC system or a Shimadzu Prominence system in water/acetonitrile gradients containing 0.1% NH₄OH (or 0.1% TFA), both equipped with Phenomenex Gemini C18 columns. Target compounds were identified from fractions using direct injection electrospray ionization mass spectroscopy on an Agilent 6520 Q-TOF system. Evaporation of volatile solvents followed by lyophilization resulted in white powders.

NMR spectroscopy

Proton nuclear magnetic resonance (¹H) spectra were recorded on a Bruker Neo 600 MHz system with a QCI-F cryoprobe (sensitivity: ¹H = 5000, ¹³C = 800). Carbon nuclear magnetic resonance (13 C) spectra were recorded on a Bruker Avance III 500 MHz with a DCH CryoProbe (sensitivity: ¹H = 1800, ¹³C = 1400). NMR spectra were all recorded at 25 °C using DMF-d₇ as solvent with 5% (v/v) D₂O. Chemical shifts are reported in parts per million (ppm) in relation to solvent peak DMF-d₇ (¹H: 8.03, 2.90, 2.75 ppm; ¹³C: 163.1, 34.9, 29.8 ppm).

E2 PA (C16V2A2E2)

NMR spectra in accordance with previously published data¹⁰³. For purity measurements see LCMS chromatogram in Supplementary Fig. 1 (a) with corresponding mass spectrum (b).

GDNF PA (C16V2A2E2ILKNLSRSR)

For purity analysis and structural characterization see Supplementary Figs. 1-7. ¹H-NMR (600 MHz, DMF-d₇): δ = 4.61–4.57 (m, 1H), 4.33–4.28 (m, 2H), 4.26–4.24 (m, 1H), 4.20–4.17 (m, 2H), 4.14–4.10 (m, 1H), 4.08–3.90 (m, 8H), 3.88–3.78 (m, 4H), 3.36–3.29 (m, 4H), 3.09–3.03 (m, 3H), 2.94–2.90 (m, 2H, note: overlap with solvent peak), 2.81–2.68 (m, 3H, note: overlap with solvent peak), 2.54–2.37 (m, 6H), 2.18–1.56 (m, 26H), 1.53 (d, 3H, J = 7.2 Hz, Ala-β), 1.51–1.45 (m, 4H), 1.34–1.22 (m, 24H), 1.10 (d, 3H, J = 6.7 Hz, Val-γ), 1.04–1.02 (m, 6H, 2xVal-γ), 0.95–0.85 (m, 24H, Ile-δ, Ile-γ’, 4xLeu-δ, Pal-ε, Val-γ); ¹³C-NMR (125 MHz, DMF-d₇): δ = 176.7, 176.5, 176.4, 176.1, 175.5, 175.5, 175.5, 175.4, 175.3, 175.2, 175.0, 174.4, 174.2, 174.2, 173.9, 173.0, 172.1, 171.8, 170.5, 157.4 (imine), 157.4 (imine), 64.3–62.5 (several carbons), 61.5, 61.0, 58.6, 57.3–56.6 (several carbons), 54.6 (2 C), 53.3, 53.0–52.5 (several carbons), 41.1, 40.9, 39.6, 39.3 (2 C), 39.19, 36.1–34.9 (several peaks, overlap with solvent peak), 31.9, 30.6–24.9 (multiple carbons, overlap with solvent peak), 24.6, 24.3, 23.1, 23.0, 22.5, 22.4, 21.4, 21.1, 20.8, 20.2, 18.8–15.6 (multiple carbons), 14.7, 13.7, 10.9. (note: several carbon peaks are undetectable due to insufficient signal-to-noise). HRMS (ESI/Q-ToF): calcd for C₈₈H₁₆₀N₂₄O₂₃, 1921.2088; found 1921.2082.

Scr PA (C16V2A2E2LRNKSRILS)

For purity analysis and structural characterization see Supplementary Fig. 1, 3, and 8. ¹H-NMR (600 MHz, DMF-d₇): δ = 4.58–4.61 (m, 1H), 4.31–4.34 (m, 1H), 4.20–4.26 (m, 3H), 4.00–4.14 (m, 9H), 3.91–3.94 (m, 3H), 3.82–3.86 (m, 2H, overlap with HDO-peak), 3.41–3.30 (m, 4H), 3.09–2.99 (m, 3H), 2.94–2.93 (m, 2H, overlap with solvent peak), 2.81–2.77 (m, 3H), 2.48–2.33 (m, 6H), 2.17–1.48 (m, 33H), 1.22–1.34 (m, 24H), 1.09 (d, 3H, J = 6.7 Hz), 1.03 (d, 3H, J = 6.5 Hz), 1.01 (d, 3H, J = 6.8 Hz), 0.84–0.95 (m, 24H). HRMS (ESI/Q-ToF): calcd for C₈₈H₁₆₀N₂₄O₂₃, 1921.2088; found 1921.2102.

Peptide (ILKNLSRSR)

For purity analysis and structural characterization see Supplementary Fig. 1, 2, 9, 10, 11, and 12. ¹H-NMR (600 MHz, DMF-d₇): δ = 4.72–4.68 (m, 1H, Asn-α), 4.52 (dd, 1H, J = 9.6, 5.4 Hz, Leu-α), 4.43–4.34 (m, 6H, 2xArg-α, Leu-α, Lys-α, 2xSer-α), 4.07 (d, 1H, J = 5.7 Hz, Ile-α), 3.89 (dd, 1H, J = 11.3, 6.3 Hz, Ser-β), 3.86 (dd, 1H, J = 11.1, 6.0 Hz, Ser-β), 3.81 (dd, 1H, J = 11.3, 5.1 Hz, Ser-β), 3.76 (dd, 1H, J = 11.1, 5.4 Hz, Ser-β), 3.33–3.25 (m, 4H, 2xArg-δ), 3.03 (t, 2H, J = 7.6 Hz, Lys-ε), 2.84–2.76 (m, 2H, Asn-β), 2.08–2.01 (m, 1H, Ile-β), 2.00–1.92 (m, 2H, Arg-β), 1.84–1.57 (m, 17H, Arg-β, 2xArg-γ, Ile-γ, Leu-β, 2xLeu-γ, Lys-β, Lys-γ, Lys-δ), 1.55–1.42 (m, 2H, Leu-β), 1.36–1.24 (m, 1H, Ile-γ), 1.05 (d, 3H, J = 6.9 Hz, Ile-γ’), 0.92–0.85 (m, 15H, Ile-δ, 4xLeu-δ); ¹³C-NMR (125 MHz, DMF-d7): δ = 174.2 (carbonyl), 173.5(carbonyl), 173.1(carbonyl), 172.6(carbonyl), 172.3(carbonyl), 172.2(carbonyl), 172.2(carbonyl), 171.4(carbonyl), 170.7(carbonyl), 168.4(carbonyl), 157.5 (imine), 157.5 (imine), 62.1 (Ser-β), 61.8 (Ser-β), 57.4 (Ile-α), 56.9 (Lys-α), 56.2 (Leu-α), 53.4, 53.3, 52.8, 52.5, 52.1 (Leu-α), 50.5 (Asn-α), 40.9 (Arg-δ), 40.9 (Arg-δ), 40.7, 40.2, 39.4 (Lys-ε), 36.8 (Ile-β), 36.7 (Asn-β), 31.6, 29.1, 28.6, 26.9, 25.5, 24.6, 24.5, 24.4 (Ile-γ), 23.0 (Leu-δ), 22.9 (Leu-δ), 22.6 (Leu-β), 21.2 (Leu-δ), 21.1 (Leu-δ), 14.5 (Ile-γ’), 10.1 (Ile-δ). HRMS (ESI/Q-ToF): HRMS (ESI/Q-ToF): calcd for C₄₆H₈₈N₁₈O₁₂, 1084.6829; found 1084.6833.

PA purity measurements with LC-MS

LC-MS was performed on an Agilent 1200 system with an Agilent 6520 Q-TOF detector. All gradient methods followed: acetonitrile at 5% for 5 min at 50 µl/min, then 5–95% over 25 min at 50 µl/min followed by 95% for 5 min at 50 µl/min. Ammonium hydroxide (0.1% v/v) or formic acid (0.1%v/v) was added to all basic or acidic solvents respectively. For acidic methods a Phenomenex Jupiter C-12; 100 × 1 mm, 5 µm was used. For basic methods, Phenomenex Gemini C-18, 100 × 1 mm; 5 µm was used.

Accurate mass analysis

High resolution mass spectra (HRMS) were recorded on an Agilent 6210 A LC-TOF mass spectrometer in positive ion mode using electrospray ionization.

PA Preparation

The samples were prepared by dissolving 10 mg of lyophilized peptide powders (E2 PA, GDNF PA and Scr PA) in a 125 mM NaCl and 3 mM KCl solution to a PA concentration of 1 wt%.

The co-assemblies were prepared by mixing the PA solutions in the volume ratios described in Supplementary Table 1, then the solutions were vortexed. All solutions were adjusted to pH of 7.4 by the addition of 1 M NaOH. The sample solutions were further annealed at 80 °C for 30 min in a water bath, then slowly cooled to ambient temperature overnight.

Cryogenic transmission electron microscopy (Cryo-TEM)

Samples for cryo-TEM were dissolved and annealed according to the above-described procedure and held at room temperature until plunge-freezing. 300-mesh copper grids with lacey carbon support (EMS) were glow discharged for 15 seconds using a PELCO easiGLow Glow Discharge Cleaning System (Ted Pella) and used immediately. Samples were diluted 10-fold in milliQ water and vortexed briefly immediately before 7 µL of sample solution was deposited on a copper grid inside a Vitrobot Mark IV (FEI) vitrification robot at room temperature with 95–100% humidity. The grids were blotted to wick away excess sample and plunge frozen into liquid ethane. Samples were transferred into a liquid nitrogen bath for holding until imaging, where they were placed into a Gatan 626 cryo-holder through a cryo-transfer stage. Cryo-TEM was performed using a JEOL 1230 TEM working at 100 kV accelerating voltage. Images were acquired using a Gatan 831 CCD camera.

Small Angle X-Ray Scattering (SAXS), Mid-Angle X-ray Scattering (MAXS), Wide Angle X-Ray Scattering (WAXS)

Experiments were performed at beamline 5-ID-D of the DuPont-Northwestern-Dow Collaborative Access Team (DND-CAT) Synchrotron Research Center at the Advanced Photon Source, Argonne National Laboratory. PA samples were prepared at 2 wt% irradiated for 2 seconds. Data was collected with an X-Ray energy at 17 keV (λ = 0.83 Å). Sample to detector distances were as follows: 201.25 mm for SAXS, 1014.2 mm for MAXS, and 8508.4 mm for WAXS. The scattering intensity was recorded in the interval 0.002390 < q < 4.4578 Å⁻¹. The wave vector q is defined as (4π/λ) sin(θ/2), where θ is the scattering angle. Azimuthal integration (Fit2D) was used to average 2D scattering images to produce 1D profiles of intensity versus q. Background scattering patterns were obtained from samples containing 125 mM NaCl and 3 mM KCl and subtracted from experimental data. All data was analyzed using the Irena software package running on IgorPro software.

Fourier Transformed Infrared (FT-IR) spectroscopy

FT-IR spectra were recorded on a Nexus 870 spectrometer (Thermo Nicolet) FT-IR spectrometer. Samples were prepared in deuterated water (D₂O) and placed between two CaF₂ windows with a spacing of 50 µm. The final spectra result from 32 scans with 1 cm⁻¹ resolution and atmospheric CO₂ and H₂O were background subtracted.

Atomic force microscopy (AFM)

PA solutions were prepared as described in the main text. Sample solutions were diluted 10X in 150 mM NaCl and deposited on freshly cleaved mica surfaces for ~1 min, and the excess solution was rinsed with 150 mM NaCl. The samples were then rinsed with 150 mM NaCl 20 mM CaCl₂ to immobilize the nanostructures on the mica surface, and measurements were performed in the liquid environment. AFM images were captured in PeakForce tapping mode on a Dimension Icon AFM (Bruker) with a silicon nitride cantilever (SNL10-A, Bruker) in a liquid cell. Images were flattened to correct sample tilt before analysis.

Circular Dichroism (CD) spectroscopy

Each PA sample was diluted to concentrations between 0.01–0.04 wt % in either H₂O (no salt samples) or buffer containing 150 mM NaCl and 3 mM KCl (high salt). CD spectra was recorded on a JASCO model J-815 spectropolarimeter using a quartz cell of 0.5 mm optical path length. Continuous scanning mode was used with a scanning speed of 100 nm per minute with the sensitivity set to standard mode. High tension (HT) voltage was recorded for each sample to ensure that the measurement was not saturated. An accumulation of three measurements was used and a buffer sample was background-subtracted to obtain final spectra. The final spectra were normalized to final concentration of each sample using a molar averaged molecular weight.

Scanning electron microscopy (SEM)

All samples were fixed in a 2.5% glutaraldehyde (GTA) 4% paraformaldehyde (PFA) phosphate buffered saline (PBS) solution for at least 20 min. They were dehydrated in a series of ethanol solutions increasing in concentration from 30% to 100% with ten-minute incubations for each incremental increase. Ethanol was then exchanged and removed with critical point drying using a Tousimis Samdri-795. Dehydrated samples were mounted using a carbon glue tape on sample studs and stored under vacuum until the day of imaging. Immediately prior to imaging, samples were coated with 16 nm of osmium (Filgen, OPC-60A) to create a conductive surface. Images were taken using a Hitachi SU8030 instrument using an accelerating voltage of 2–3 kV.

Rheological measurements

An Anton Paar MCR302 Rheometer with a 25 mm cone plate was used for all rheological studies. 150 µl of PA liquid was placed on the sample stage and 30 µl of 150 mM CaCl₂ (final concentration 25 mM) was placed on the plate while in the instrument loading position. The instrument was set to 37 °C to simulate in vitro and in vivo conditions. The plunger was lowered to the measuring position and a humidity collar was added to prevent sample evaporation during the measurement. Each sample was equilibrated for 30 min with a constant angular frequency of 10 rad/s and 0.1% strain. The angular frequency was then decreased from 100 rad/s to 1 rad/s over 21 points and the storage and loss modulus were recorded. The % strain was increased from 0.1 to 100% over 31 points and the storage and loss moduli were recorded.

Cell culture

We used iPSC-derived human midbrain floorplate dopaminergic neurons (iCell® DopaNeurons, Cellular Dynamics). The cells were thawed following manufacturer´s instructions. Briefly, the cells were seeded on coverslips coated with PLO (0.01%) and 3.3 µg/ml laminin solution at 4.0 × 10⁴. We used the media recommended by Cellular Dynamics and followed the instructions provided by the manufacturer. The preparation consists of the addition of two supplements to a basal media (for the starvation condition in WB experiments we used only the basal media without the addition of supplements). The media was changed the next day of seeding, and then we changed the media every three days. We treated the samples with 10 or 100 ng/ml (3.3 nM) of human recombinant GDNF protein (Stemcell Technologies), and with 100 ng/ml or 3.3 nM of the soluble mimetic peptide (ILKNLSRSR). The GDNF or peptide was added every time we changed the media.

PA treatments

PA on the top of cells

We began by removing the cell media and then added 30 µl of PA with a P200 pipette, avoiding any contact between the tip and the cells. The PA was distributed evenly around the coverslip, and we subsequently added fresh cell media slowly.

PA coatings

We used coverslips coated with poly-D-lysine (PDL) 0.1 mg/ml (Sigma-Aldrich). Coverslips were washed once with sterile water and allowed to dry in air inside a laminar flow hood. 15 µl of PA was deposited on a piece of sterile parafilm. The coverslips were placed in contact with the PA for 1 hour. Then, coverslips were placed in a 24-well plate with the side previously in contact with the PA facing upwards. Gelling solution composed of 125 mM NaCl, 3 mM KCl, and 25 mM CaCl₂ was added for 1 min. The gelling solution was removed, and the coverslips were washed with sterile PBS 1X. PBS was withdrawn and cell media was added.

PA scaffolds

iCell® DopaNeurons (Cellular Dynamics) were thawed according to manufacturer instructions to obtain 1.0 × 105 in 10 µl of cell media. Cell suspension was mixed with 20 µl of 1 wt% annealed PA of 15 vol % Scr PA co-assembly or 15 vol % GDNF PA co-assembly by pipetting. This mixed solution (30 µl) was placed in a coverslip and let it rest for 5 min inside the incubator. Then, 500 µl of cell media was added carefully to avoid disrupt the gel.

Cell viability

Cell media with Calcein-AM 1 µM and propidium iodide 2.5 µM (Invitrogen) was added and placed in the incubator for 20 min according to manufacturer’s instructions. The solution was removed, and the samples were washed with PBS 1X. Coverslips were mounted, and the confocal images were acquired immediately. Experiments were performed in triplicate.

PA internalization assay

We prepared the TAMRA-GDNF PA mixing 85 vol % of E2 PA (1 wt%) (without dye) with 15 vol % of C16-VVAAEE-K-TAMRA-ILKNLSRSR (1 wt%) following the procedure described in the PA Preparation section. The cells were seeded on 24 well plates with glass-like polymer bottoms (Cellvis, cat. No. P24-1.5 P) at a density of 4.0 × 10⁴. We followed the cell culture conditions described before. The neurons were incubated with 10 µg/mL cholera toxin B-AlexaFluor647 (Invitrogen, cat. No. C34778) at 37 °C for 1 hour. The neurons were treated with TAMRA-GDNF PA (10.73 mM of PA concentration) for either 10 min or 1 hour at 37 °C prior to being washed and stored in warmed media. Imaging was performed using a Nikon AXR confocal microscope in a live cell chamber at 37 °C under 5% CO₂. For clearer visualization of PA internalization, we utilized Imaris Bitplane software (Oxford Instruments) version 9.2.1. Briefly, we used the surface function, and the threshold was adjusted to match the fluorescence signal of the images obtained with the Nikon AXR confocal. To distinguish PA from neurons, we employed separate surfaces (Imaris function) for each and subsequently superimposed the images. Within Imaris, the neuron surfaces were then adjusted to 70% transparency to better visualize GDNF PA inside the neurons.

Western blot

For receptor activation and downstream signaling experiments 850,000 cells were seeded in 6 well plates precoated with poly-D-lysine (Sigma-Aldrich) until they reached about 70% confluency. All treatments were added for 10 min and then the protein was harvested. To prepare the PA solution, we combined 500 µl of a 10.73 mM PA solution with 500 µl of media collected from the cell culture (preventing receptor activation caused by the introduction of fresh media). To avoid the formation of gel clusters, the mixture was vortexed thoroughly for approximately 10 seconds until no visible gel fragments remained. The resulting solution was then added to the remaining 1 ml of medium in the well, achieving a final concentration of 3.58 mM. We added human recombinant GDNF protein (Stemcell Technologies) to the media to obtain a final concentration of 100 ng/ml (3.3 nM). For neuroprotective experiments with 6-OHDA (Sigma-Aldrich) the cells were seeded on PA coatings except for Starvation (strv) and PDL conditions in which Poly-D-lysine coated wells were used. The neurotoxin 6-OHDA (100 µM) was added for 1 hour and then the protein was extracted. The 25 µg protein was separated by 4–20% SDS-polyacrylamide gel electrophoresis and electrotransferred to a nitrocellulose membrane (Bio-Rad). Membranes were blocked with 5% milk solution (Bio-Rad) for 1 hour and then primary antibodies were added for an overnight incubation at 4 °C. The following primary antibodies were used: rabbit anti-Actin (1:2000, Sigma-Aldrich), rabbit anti-Fyn (1:1000, Cell Signaling), anti-pSrc (1:1000, Cell Signaling), rabbit anti-ERK ½ (1:1000, Cell Signaling), rabbit anti-pERK 1/2 (1:1000, Cell Signaling), rabbit anti-RET (1:1000, Cell Signaling), rabbit anti-pRET (1:500, Cell Signaling), rabbit anti-TH (1:1000, Pel-freez) and mouse anti-Tuj-1 (1:5000, BioLegend), GAPDH (1:2000, Millipore). Subsequently, the membranes were incubated with secondary antibodies (1:1000; ThermoFisher). Densitometry analysis, normalized to actin, GAPDH or total protein (ERK or RET) as a control for protein loading, was accomplished using ImageJ software (National Institutes of Health, USA). The experiments were performed in triplicate.

Pharmacological inhibition of RET and NCAM

We seeded the iPSC-derived human dopaminergic neurons as described in the Cell culture section. For RET receptor inhibition, we incubated the cells (2DIV) for 24 h with Vandetanib (Tocris) (1 µM). In a simultaneous experiment, we incubated the neurons with PP2 (Tocris) (1 µM) for 1 hour to inhibit Fyn (an NCAM downstream signaling protein) (experiments were carried out in, n = 200 cells). Subsequently, we added GDNF protein (100 ng/ml) or 15 vol % GDNF PA (on top of the cells) for two days and did a cell viability assay or fixed the cells for TH immunostaining. In our Western Blot experiments, we used cells at 70% confluency, as previously described. After adding the inhibitors, GDNF protein or GDNF PA was introduced in solution for 10 min, and the protein was then harvested according to the earlier method.

Human embryonic stem cell maintenance and hMLOs generation

This study has been approved by Northwestern University Committee on Human Stem Cell Research. This study did not use data from human participants. Human embryonic stem cell (hESC) line, HUES64 NIH Human Embryonic Stem Cell Registry No. NIHhESC-10-0067 (passage 40)¹⁰⁴ was maintained as previously described¹⁰⁵ in accordance with all relevant ethical regulations. Briefly, the cells were maintained on Matrigel-coated dishes with mTeSR1 or mTeSR plus (Stem Cell Technologies). The cells were dissociated to single cells with 1 mM EDTA or Accumax (Innovative Cell Technologies) for passaging or differentiation, respectively. Matrigel lots have been routinely tested to optimize stem cell maintenance. ROCK inhibitor was used to prevent cell death upon cell dissociation¹⁰⁶. hMLOs were generated by following a previously described protocol⁸⁵. Briefly, we plated 10,000 cells on V-bottomed conical wells (Sumitomo Bakelite) to form the embryonic bodies. We maintained the cells with induction media (DMEM/F12:Neurobasal, 1:1), N2 supplement 1:100 (Invitrogen), B27 without vitamin A1:50 (Invitrogen), 1% Glutamax (Invitrogen), 1% minimum essential media-nonessential amino acid (Invitrogen), and 0.1% β-mercaptoethanol (Invitrogen) with 1 μg/ml heparin (Sigma-Aldrich), 10 μM SB431542 (Stemgent), 200 ng/ml Noggin (Prospec), 0.8 μM CHIR99021 (Reagents Direct), and 10 μM ROCK inhibitor Y27632 (Calbiochem). On day 4, we added 100 ng/ml SHH-C25II (R&D Systems) and 100 ng/ml FGF8 (R&D Systems). After 7DIV we removed the media and added 30 μl of reduced growth factor Matrigel, then we incubated the cells at 37°C for 30 min. We maintained the cells with in Neurobasal medium, N2 supplement 1:100 (Invitrogen), B27 without vitamin A 1:50 (Invitrogen), 1% GlutaMAX (Invitrogen), 1% minimum essential media-nonessential amino acid (Invitrogen), and 0.1% β-mercaptoethanol (Invitrogen) supplemented with 2.5 μg/ml insulin (Sigma-Aldrich), 200 ng/ml laminin (Sigma-Aldrich), 100 ng/ml SHH-C25II (R&D Systems), and 100 ng/ml FGF8 (R&D Systems) for 24 h. The organoids were transferred to low-attachment 6 well-plates (Costar). At this point, we added the differentiation media containing Neurobasal medium, N2 supplement 1:100 (Invitrogen), B27 without vitamin A 1:50 (Invitrogen), 1% GlutaMAX (Invitrogen), 1% minimum essential media-nonessential amino acid (Invitrogen), 0.1% β-mercaptoethanol (Invitrogen), 10 ng/ml BDNF (Peprotech), 10 ng/ml GDNF (Peprotech), 100 μM ascorbic acid (Sigma-Aldrich), and 125 μM db-cAMP (Sigma-Aldrich). Penicillin (100 U/ml) and streptomycin (100 μg/ml) were added. We cultured the organoids in an orbital shaker to promote the nutrient and gas exchange.

Organoid embedding process

At 80DIV, we observed a high population of TH-positive cells; therefore, we used organoids at this stage to evaluate the bioactivity of PAs. We transferred the organoids from the 6 well plates to Lab Tek® 4 well-chambers slide (one organoid per chamber). The organoids were moved carefully to the corner of the chamber with a sterile tip (Supplementary Fig. 13a), then we removed the media and enough PA solution to cover the organoid was added (50 µl) (the PA solution was prepared as previous experiments, 15 vol% of Scr PA and GDNF PA). Immediately, we added gelling solution (25 µl), after 3 min we added media. The media was prepared following the methods described by Jo et al.⁸⁵, we changed the media every three days. After one week we fixed with 4% PFA the samples for two hours. The experiments were performed in triplicate with a total of 9 organoids per condition.

Immunostaining and Imaging

All samples were fixed with PFA 4%, cells seeded on coverslips were fixed for 15 min at RT, cells embedded of PA gels were fixed for one hour at RT, organoids embedded on PA were fixed for 2 h at RT and organoids without PA were fixed overnight at 4 °C. After fixation organoids without PA were transferred in a 30% sucrose solution in PBS overnight at 4 °C, then the organoids were embedded in a O.C.T. solution (Sakura Finetek) and maintain at −80 °C until cryosectioning. The organoids were sliced at a thickness of 10 µm with a Leica CM1520 cryostat. Cells on coverslips and embedded in gels, and slices of organoids were permeabilized with 0.3% Triton 100X and 10% Normal Goat Serum in PBS for 1 hour, organoids embedded in PA were permeabilized for 2 h. Samples were incubated with primary antibodies in PBS with 10% Normal Goat Serum overnight and with secondary antibodies in PBS with 10% Normal Goat Serum for 2 h at RT. After antibodies incubations samples were washed with PBS/Bovine Serum Albumin 0.1% 3X for 5 min for cells on coverslips or 4X for 15 min for cells and organoids embedded in PA gels. The following primary antibodies were used: rabbit anti-TH (Tyrosine Hydroxylase marker 1:1000, Pel-freez), mouse anti-TH (1:1000, RD Systems), mouse anti-MAP2 (marker for mature neurons 1:1000, Thermo Fisher), mouse anti Tuj1 (1:1000, Biolegend), mouse anti-PSD95 (Post-synaptic marker 1:1000, Thermo Fisher), mouse anti-Synaptophysin (Pre-synaptic marker 1:100, Cell Signaling), rabbit anti-Cleaved Caspase-3 (marker for apoptosis 1:500, Cell Signaling), goat anti-Girk2 (midbrain dopaminergic marker 1:200, abcam), mouse anti-nestin (neuronal precursor marker 1:500, Novus Biologicals) mouse anti-Tuj1 (neuronal marker 1:1000, Biolegend), and mouse anti-DAT (dopamine transporter marker 1:200, Atlas antibodies). The following secondary antibodies were used: Alexa 488 or Alexa 568 (1:1000 Thermo Fisher). Stained cells in PA coating were visualized using a Nikon A1C confocal microscope. Cells embedded in PA gels were examined with a Nikon A1R-MP multiphoton microscope (we transferred the coverslips with the PA gels from the 24-well plate to a 35 mm dish for imaging). Organoids images were taken on a Nikon AXR confocal. We used a Zeiss AxioZoom V16 Stereo Microscope to visualize the organoid embedded on the PA. The imaging conditions were identical in all groups in every experiment.

Imaging and morphometric analysis

Confocal images of PA coatings were reconstructed using NIS Elements Advanced Research Microscope Imaging. Images were processed using ImageJ/Fiji (National Institutes of Health) software. Number of dendrites, Sholl analysis and colocalization were analyzed using Fiji. Axonal length was measured using NeuronJ plugin in ImageJ. The experiments were performed in triplicate (in total 200 cells per group). Sholl analysis quantified the number of intersections that the neurites have with concentric circles from the cell soma, for this purpose, we used circles with a 40 µm of distance between them for cell coating experiments and 10 µm for scaffold experiments. Experiments were performed by triplicate, n = 200 cells per condition. 3D images from multiphoton microscopy of PA gels samples were reconstructed and analyzed using Imaris software (version 8.1, Bitplane Scientific software). Volume of fluorescence signal (TH and MAP2) was analyzed using Voxel Counter plugin of ImageJ. For morphology, quantification a minimum of 30 randomly selected neurons were analyzed per condition in at least two independent experiments (in total 60 cells per group). To evaluate the axonal extension from organoids, the confocal images were first converted to 8-bit files then the Tubeness plugin in Fiji software using a 1-sigma value was applied to remove fluorescence noise from the PA. To obtain a 3D digital reconstruction of the axons and measure their lengths, we used the Simple Neurite Tracer plugin. Volume projection images with the depth scale of 3D-cultres and organoids were generated using the volume projection function of NIS Elements software from Nikon.

Figures preparation

Adobe Photoshop CC software version 20.0.7 was used to adjust brightness and contrast and to prepare the figures. We utilized Adobe Illustrator CC version 23.0.1 to create the illustrations presented in Figs. 2a and 5a. The scheme in Fig. 1b was generated using Cinema 4 d and ePMVREF¹⁰⁷.

Electrophysiology

Spontaneous pacemaking activity was recorded on iCell® DopaNeurons (Cellular Dynamics) cultured by 30-35DIV. Cultures were transferred to a recording chamber on a fixed-stage inverted microscope (Diaphot 200; Nikon). Experiments were performed at 32 °C in neurons seeded on Scr PA 15 vol %, GDNF PA 15 vol % and PDL coatings with GDNF (100 ng/ml). The recording chamber was perfused (1–2 ml/min) with HEPES-based solution (in mM): 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES, 10 glucose, 10 sucrose; pH 7.4, osmolarity 300–305 mOsm/L. Patch pipettes (3–5 MW) were filled with internal solution containing the following (in mM): 135 K-MeSO₄, 5 KCl, 5 HEPES, 0.05 EGTA, 10 phosphocreatine-di(tris), 2 ATP-Mg, 0.5 GTP- Na, pH: 7.25–7.30, osmolarity 285–295 Osm/L. Recording patch pipettes were prepared with a horizontal puller (model P-97; Sutter Instruments) using borosilicate glass with filament (outer diameter 1.5 mm, inner diameter 0.86 mm). Somatic cell-attached voltage-clamp recordings were obtained with a Multi-Clamp 700B amplifier (Molecular Devices) interfaced to a Pentium-based PC running pClamp 10.6 (Molecular Devices). The signals were filtered at 1 kHz and digitized at 10 kHz with a Digidata 1440 A (Molecular Devices). Analysis of instantaneous firing frequency was done in Clampfit 10.6 (Molecular Devices). The experiments were performed by triplicate. Dopaminergic neurons were identified based on their characteristic slow, autonomous pacemaking (3.8 Hz)¹⁰⁸.

RNA sequencing libraries

RNA quality was determined with the Agilent Bioanalyzer 2100, accepting RNA integrity numbers (RIN) of > 7 and quantified using Qubit. Directional mRNA libraries were prepared using Illumina TruSeq mRNA Sample Preparation Kits per manufacturer’s instructions. Briefly, polyadenylated mRNAs were captured from total RNA using oligo-dT selection. Next, samples were converted to cDNA by reverse transcription, and each sample was ligated to Illumina sequencing adapters containing unique barcode sequences. Barcoded samples were then amplified by PCR and the resulting cDNA libraries by quantified using qPCR. Finally, equimolar concentrations of each cDNA library were pooled and sequenced on the Illumina HiSeq 4000.

RNA sequencing analysis

The quality of reads, in FASTQ format, was evaluated using FastQC. Reads were trimmed to remove Illumina adapters from the 3’ ends using cutadapt¹⁰⁹. Trimmed reads were aligned to the homo sapiens genome (mm10) using STAR¹¹⁰. Read counts for each gene were calculated using htseq-count¹¹¹ in conjunction with a gene annotation file for mm10 obtained from Ensembl (http://useast.ensembl.org/index.html). Normalization and differential expression were calculated using DESeq2 that employs the Wald test¹¹². The cutoff for determining significantly DEGs was Fold change ±1, P-value < 0.05. Origin2023b software was used to plot the DEGs in the volcano plot.

Gene ontology analysis

To perform the gene ontology, we utilized the protein coding genes with ±1 fold change values (see Supplementary Data 1). We selected the SynGO2020⁷⁵ library to evaluate the DEGs in a synaptic context. We focused on the top ten upregulated and top ten downregulated enriched terms. To identify the most significant molecular function (MF) GO terms of the upregulated protein-coding genes we used the g:Profiler web⁷⁷. We selected the functional profiling option under the following parameters: Organism-Homo sapiens, customed-background list, Significance threshold-g:SCS threshold, User threshold-0.05, Numeric IDs treated as-entrezgene_acc. We used the all the genes shown in the sequencing raw data as a background list (Supplementary Data 1). To compare the upregulated genes by GDNF PA with the upregulated genes reported by Gantner⁷⁸, we requested the information submitted to the GEO database with the number GSE143136 through Pluto-bio software. We used the Venn diagram function from (https://www.biotools.fr) to compare the genes with ≥ 1-fold change values.

Statistical analysis

Statistical analysis was performed using the software GraphPad Prism version 6.01. Data from multiple group experiments were analyzed by One-way ANOVA followed by Tukey post hoc test. Comparison data from two experimental groups were analyzed by T-test. Error bars indicate the standard error of the mean. Differences were considered statistically significant when P < 0.05, and the significance level is shown by *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.