|
Fig. 1. Synapse densities are lower on stiff substrates than on soft substrates.Immunocytochemistry of neurons cultured on a soft (0.1 kPa) and b stiff (10 kPa) substrates, showing (from top to bottom) neurofilament (NF), postsynaptic marker (NRL), vesicular glutamate transporter 2 (Vglut2; excitatory), vesicular GABA transporter (VGAT; inhibitory), and an overlay of all channels. c, d Quantification of synapse densities. Both glutamatergic and GABAergic synapse densities were lower on stiff substrates than on soft ones at DIVs 10 and 14. Boxplots showing detected synapses per mm neurofilament; p-values provided above the plots (two-tailed t-tests). Data from three independent experiments. For DIV 10, 31 fields of view (FOV) were analysed on 0.1 kPa substrates and 37 FOV on 10 kPa substrates. For DIV 14, 35 FOV were analysed on 0.1 kPa substrates and 40 FOV on 10 kPa substrates. Boxplots show the median (central line), the interquartile range (boxes), and whiskers represent 1.5 times the interquartile range. Scale bar: 5 μm.
|
|
Fig. 2. Neuronal activity is lower on stiff substrates than on soft substrates.a Example traces from voltage-clamp recordings showing induced currents in cells on soft (0.1 kPa) and stiff (10 kPa) substrates at DIV 5, 7, 10, and 14. b, c Current densities for voltage-gated sodium (INa) channels were calculated from whole-cell voltage-clamp recordings, as shown in (a). c Peak sodium current densities were significantly lower on stiff substrates, indicating delayed maturation (two-way ANOVA, time points and substrate stiffness as factors). Cell numbers are identical in (b) and (c). Current densities for voltage-gated potassium channels (d, e IK, f, g IKdr) calculated from whole-cell voltage-clamp recordings, as shown in (a). e, g Current densities at +40 mV were similar between the stiffnesses at all time points (two-way ANOVA). Cell numbers are identical in (d–g). h Representative traces of spontaneous action potentials (APs) recorded by whole-cell patch-clamp electrophysiology at DIV 14. i Plot of spontaneous activity. A cell was considered active if it produced at least one action potential during the recording. On soft gels, spontaneous APs were detected from DIV 5 until the end of the culture period, whilst neurons on stiff gels did not show any spontaneous activity until DIV 10. At DIVs 10 and 14, cells showed similar levels of activity regardless of the environmental stiffness, suggesting that the electrical maturation of neurons was delayed on stiffer substrates. j Representative traces of APs evoked in whole-cell current-clamp recordings at DIV 14. k F–I curves showing AP frequencies as a function of the injected current. Neuronal excitability on soft and stiff substrates was evaluated using a two-way ANOVA, with applied current and substrate stiffness as factors. Excitability was significantly higher on soft substrates than on stiff ones between DIV 5–10; this trend reversed at DIV 14. b, d, f, k Line plots represent mean values, error bars the standard error of the mean (SEM). c, e, g Plots represent means, error bars the SEM.
|
|
Fig. 3. Piezo1 delays the electrical maturation of neurons on stiff substrates.a Example traces of voltage-clamp recordings showing induced currents in control and Piezo1 knockdown (P1 KD) cells on soft and stiff substrates at DIV 7. b Potassium current densities on soft and stiff substrates and sodium current densities on soft substrates were similar between control and P1 KD cells. On stiff substrates, control cells showed a trend towards lower peak sodium current densities compared to P1 KD cells, suggesting a Piezo1-dependent delay in maturation (see Fig. 2b), though not statistically significant (ANOVA and Sidak’s multiple comparisons test). c Representative traces of evoked action potentials recorded at DIV 7. d Analysis of spike frequencies as a function of injected current. While control and P1 KD cells displayed similar excitability on soft substrates, control neurons exhibited significantly reduced excitability on stiff substrates compared to P1 KD neurons (two-way ANOVA, factors: applied current and genetic condition), further supporting the idea of a Piezo1-dependent delay in neuronal maturation on stiff substrates. p-values are indicated in the plots. e Image of Rhod-4-loaded neurons, with semi-automatically labelled somata, and soma fluorescence intensities. f Plot of the baseline-corrected intensity of each cell over time. g Filtered signals with automatic peak detection. Colour-coded images showing intensity changes ΔF over 40 s. CTRL cells on h soft substrates showed strong variability, whereas i low fluctuations on stiff substrates indicated little neuronal activity. j Knockdown of Piezo1 rescued the spontaneous activity of neurons on stiff gels. k Active fraction of cells as a function of substrate and time. Cells were considered active if they produced at least one calcium transient (“F” = fields of view, “C” = number of cells). CTRL neurons (red) on soft substrates became active by DIV 5–6 but showed little activity on stiff substrates until DIV 7. Knockdown of Piezo1 (blue and yellow) rescued the activity of neurons on stiff substrates (Kruskal–Wallis test and two-sided Tukey test), confirming that Piezo1 activity delayed the electrical activity of neurons on stiff substrates. Boxplots show the medians (central line) and interquartile ranges (boxes); whiskers represent 1.5 times the interquartile range. Scale bars: 25 μm. See also Supplementary Movies 1–4.
|
|
Fig. 4. Transthyretin (TTR) regulates electrical maturation of neurons downstream of Piezo1.a Volcano plot of RNA-sequencing results of CTRL neurons grown on soft vs. stiff substrates at DIV 7. Foldchange is shown on the x-axis (blue: higher expression on soft, red: higher expression on stiff gels), p-values on the y-axis. Statistical significance was assessed using the DESeq2 two-sided Wald test. The volcano plot displays unadjusted p-values. The most significant change between the conditions was observed in TTR levels. b
TTR RNA counts normalised by sample-specific size factors for CTRL and Piezo1 knockdown neurons on soft and stiff substrates as assessed by RNA sequencing. TTR levels were similarly high in control cells on soft substrates and in Piezo1 KD cells on soft and stiff substrates. However, in control cells cultured on stiff substrates, TTR levels were significantly lower, indicating that knockdown of Piezo1 rescued the effect of a stiff substrate. The pattern of TTR levels in the different groups resembled that of their activity at DIV 7 (low in CTRL neurons cultured on stiff substrates, high in all other conditions; Fig. 3k), suggesting a link between TTR levels and neuronal maturation (ANOVA test followed by a two-sided Tukey post hoc test). c Representative Western blots of CTRL neurons cultured on soft and stiff substrates at DIV 7, stained for TTR and total protein stain (TPS), which was used for normalisation. d Analysis of Western blots of CTRL neurons cultured on soft and stiff substrates at DIV 7. The expression of TTR was significantly higher on soft than on stiff substrates (n = 22 WB bands from 3 biological replicates; one-sided 1-sample t-test). e Proposed pathway linking substrate stiffness-dependent Piezo1 activity to the expression of TTR, synapse formation, and potentially intrinsic excitability. f Plot of the fraction of active cells in TTR KD neurons on soft substrates and CTRL neurons on stiff substrates assessed by calcium imaging. No significant differences were found, supporting the proposed model shown in (e) (n = 44 CTRL cells and 41 TTR KD cells; two-sided Mann–Whitney test). Boxplots show medians (central lines) and interquartile ranges (boxes); whiskers represent 1.5 times the interquartile range.
|
|
Fig. 5. Synapse formation is delayed in stiffer brain tissue in vivo.a Schematic of a Xenopus laevis embryo whose brain was exposed, enabling stiffness measurements by AFM and fluorescence imaging of functional synapses (indicated in pink). b Average stiffness map of Xenopus laevis brains at stages 37–38. c Averaged FM1-43 fluorescence image indicating synapse locations. d Binary maps of average tissue stiffness and synaptic density (see “Methods” section). Stiffness negatively correlated with synaptic density (p = 5 × 10⁻⁶, two-sided Fisher’s exact test). The odds ratio was 0.25, indicating that synapses were four times more likely to form in soft regions than in stiff ones. e, f, j, k Images of exposed Xenopus laevis brains; synapses were stained using FM1-43. e At stage 37–38, a typical staining pattern was found in the area of the supraoptic tract (arrows). f In stiffened brains (cf. Supplementary Fig. 15), this pattern was largely absent. g, h Intensity profiles along the stripes indicated in (e, f). While the untreated brain showed a strong peak, the TG profile was noise-dominated. i The ratio of the maximum over the mean value in each profile was significantly higher in control brains than in stiffened brains (two-sided Mann–Whitney test), indicating more functional synapses in softer brains. j, k After a recovery period of 8 h, strong signals in the area of the supraoptic tract were visible in either condition, suggesting that synapse formation was delayed in the stiffened brains. l, m Line profiles show very similar features between the two conditions at stage 39. n The ratio of the maximum value over the mean value showed no significant difference at stage 39 (two-sided Mann–Whitney test). Boxplots show the median (central line), the interquartile range (box), and whiskers represent 1.5 times the interquartile range. Scale bars: 100 μm.
|
