NEAT1-mediated regulation of proteostasis and mRNA localization impacts autophagy dysregulation in Rett syndrome

Abstract

Rett syndrome (RTT) is a severe neurodevelopmental disorder primarily caused by loss-of-function mutations in the MECP2 gene, resulting in diverse cellular dysfunctions. Here, we investigated the role of the long noncoding RNA (lncRNA) NEAT1 in the context of MeCP2 deficiency using human neural cells and RTT patient samples. Through single-cell RNA sequencing and molecular analyses, we found that NEAT1 is markedly downregulated in MECP2 knockout (KO) cells at various stages of neural differentiation. NEAT1 downregulation correlated with aberrant activation of the mTOR pathway, abnormal protein metabolism, and dysregulated autophagy, contributing to the accumulation of protein aggregates and impaired mitochondrial function. Reactivation of NEAT1 in MECP2-KO cells rescued these phenotypes, indicating its critical role downstream of MECP2. Furthermore, direct RNA-RNA interaction was revealed as the key process for NEAT1 influence on autophagy genes, leading to altered subcellular localization of specific autophagy-related messenger RNAs and impaired biogenesis of autophagic complexes. Importantly, NEAT1 restoration rescued the morphological defects observed in MECP2-KO neurons, highlighting its crucial role in neuronal maturation. Overall, our findings elucidate lncRNA NEAT1 as a key mediator of MeCP2 function, regulating essential pathways involved in protein metabolism, autophagy, and neuronal morphology.

Graphical Abstract

Figure 1.

NEAT1 is highly dysregulated in mutant MeCP2 models, including human neural cells and RTT patient samples. (A) UMAP plot of NPCs [WT and MECP2-KO (clone A33)] across all timepoints, labeled according to sample. (B) Astrocyte signature score across samples. (C) Heatmap of the top differentially expressed genes between WT and MECP2-KO (clone A33) cells, across all time points. (D) (Left) UMAP plot of NPCs (WT and MECP2-KO) across all timepoints, labeled according to MECP2 genotype status; (Right) UMAP plot showing the normalized NEAT1 expression in all NPCs (WT and MECP2-KO). (E) Expression levels of NEAT1 measured by RT-qPCR in WT and MECP2-KO progenitor cells (pg) or after 7 days of free differentiation (7 d). Two different MECP2-KO clones (A33 and 3F) were analyzed. Primers used are depicted in the upper diagram. NEAT1_2 signal corresponds to the long isoform, and NEAT1_1 + 2 signal detects both the short (NEAT1_1) and the long (NEAT1_2) isoforms. Graphs show the mean ± SD of three replicates (****P < .0001, one-way ANOVA). (F) Expression levels of MECP2 and NEAT1 detected by RT-qPCR in WT, MECP2-KO (clone A33) or MECP2-KO cells transfected with MECP2_e1 construct. Graphs show the mean ± SD of three (MECP2 and NEAT1_2 levels) or four (NEAT1_1 + 2) replicates (*P < .05, **P < .01, ****P < .0001, one-way ANOVA). (G) Expression levels of NEAT1 in RTT patient-derived iPS cells (T158M and the isogenic WT control) and in the same cells differentiated into neural progenitor cells (NPCs) for the indicated times. Graphs show the mean ± SD of three (NEAT1_2) or four (NEAT1_1 + 2) replicates (**P < .01, ****P < .0001, ns = not significant, one-way ANOVA). (H) Expression levels of total NEAT1 (both isoforms) in the hippocampus (n = 3), cerebellum (n = 4) and frontal cortex (n = 4) of post-mortem RTT patients or healthy control samples. Graphs show the mean ± SEM of replicates (unpaired t-test, *P < .05, ns = not significant). (I) Expression levels of total NEAT1in the peripheral blood of n = 12 RTT patients. The graph shows the mean ± SEM of replicates (unpaired t-test, *P < .05). See also Supplementary Fig. S1.

Figure 2.

NEAT1 expression is highly variable between cells and its presence correlates with negative regulation of mTORC1. (A) Normalized expression levels for NEAT1 (left) and MECP2 (right) across timepoints in single-cell analysis. (B) (Left) RNA-FISH showing localization of NEAT1 in WT NPCs. Scale bar = 20 μm. (Right) Frequency distribution of NEAT1 RNA-FISH foci per cell. The x-axis represents the number of RNA-FISH foci per cell, and the y-axis indicates the percentage of cells with the corresponding number of foci. Data are represented as the mean ± SD of n = 50 WT NPCs in three independent experiments. (C) Volcano plot showing differentially expressed genes between progenitor cells with high (>60th percentile of normalized NEAT1 expression) and low (<60th percentile) NEAT1 expression. The top differentially expressed genes are highlighted. Dashed lines indicate 0.15 as average log₂ fold change threshold (x axis) and 0.01 as P adjusted value threshold (y axis). For clarity, NEAT1 (log₂FC = 2.4338, P adj. value = 0) and MALAT1 [log₂FC = 0.7312, −log₁₀ (P adj. value = 206.68)] are not included in the plot. (D) (Above) Enriched Gene Ontology terms for the top 20 upregulated genes in cells with high NEAT1 expression as identified by functional clustering (Enrichr). The y axis shows the GO terms and the x axis shows the statistical significance (two-tailed Fisher’s exact test). (Below) Same as above, for the top 20 genes upregulated in cells with low NEAT1 expression. (E) (Left) Volcano plot showing gene expression changes (−10 < log₂FC < 10) in MECP2-R133C point mutant cells relative to WT progenitor cells. Upregulated genes are shown in red, downregulated genes in blue. The dotted vertical line indicates log₂FC = 0, and the dotted horizontal line represents a P-value = .05. (Right) enriched Gene Ontology terms for all differentially expressed genes in R133C cells (P-value < .05) as identified by functional clustering (Enrichr). (F) Enriched Gene Ontology terms for all upregulated genes in R133C cells (FC > 1, P-value < .05) as identified by functional clustering (Enrichr). (G) Expression levels of NEAT1 measured by RT-qPCR in WT and MECP2-R133C progenitor cells (pg) or after 7 days of free differentiation (7 d). NEAT1_2 signal corresponds to the long isoform, and NEAT1_1 + 2 signal detects both the short (NEAT1_1) and the long (NEAT1_2) isoforms. Graphs show the mean ± SD of n = 5 (pg) or n = 3 (7 d) replicates (****P < .0001, one-way ANOVA). See also Supplementary Fig. S2.

Figure 3.

NEAT1-KO neural cells display transcriptomic alterations similar to MECP2-KO cells. (A) NEAT1 depletion levels in MECP2-KO (clones A33 and 3F), NEAT1-KO cells edited by CRISPR/Cas9, and in NEAT1 shRNA-depleted cells were assessed by RT-qPCR. Graphs show the mean ± SD of three replicates (**P < .01, ****P < .0001, one-way ANOVA). Diagram was created in BioRender. Soler (2025) https://BioRender.com/o10b971. (B) (Left) MECP2 mRNA levels measured by RT-qPCR. Graph shows the mean ± SD of three replicates (*P < .05, ***P < .001, ****P < .0001, ns = not significant, one-way ANOVA). (Right) Western blot analysis of MeCP2 in WT or KO progenitor cells. Quantification of band intensity is shown below each lane. (C) (Left) RNA-FISH showing abundance of NEAT1 in WT, MECP2-KO (clones A33 and 3F) or NEAT1-KO NPCs. Scale bar = 10 μm. (Right) Quantification of NEAT1 presence per cell nucleus. Graph represents the mean ± SEM of the number of nuclear NEAT1+ foci in n = 9 cells (*P < .05, **P < .01, ****P < .0001, one-way ANOVA). (D) Heatmaps of differentially expressed transcripts (DETs) (|log₂FC| ≥ 1, q value ≤ 0.05) in NEAT1-KO and MECP2-KO (clone A33) NPCs (top) and 7-day differentiated cells (bottom) relative to WT cells. The number of condition-specific or shared DETs is shown to the left of the heatmaps together with the P-value from the hypergeometric test for the overlap between the conditions. (E) Enriched KEGG pathways for all differentially expressed genes (q value ≤ 0.05) in NEAT1-KO NPCs relative to WT cells, as identified by functional clustering (Enrichr). See also Supplementary Fig. S3.

Figure 4.

NEAT1 downregulation in MECP2-KO cells impacts protein metabolism, vesicle trafficking and mitochondrial function of human neural cells. (A) Western blot analysis of DDIT4 levels in MECP2-KO (clone A33) and NEAT1-KO progenitor cells. Diagram was created in BioRender. Soler (2025) https://BioRender.com/i12q588. (B) Western blot analysis of mTOR, total S6 and phosphorylated-S6 levels in NEAT1-KO, MECP2-KO and MECP2-KO overexpressing NEAT1 progenitor cells. Quantification of band intensity is shown below each lane. (C) (Left) Assessment of protein aggregation in WT or MECP2-KO (clones A33 and 3F) NPCs with ProteoStat^® dye. Graph shows the mean ± SD of three replicates (****P < .0001, one-way ANOVA). (Right) Same assay in NEAT1-KO or NEAT1-shRNA depleted cells. Graph shows the mean ± SD of four replicates (**P < .01, ****P < .0001, one-way ANOVA). (D) Protein aggregation quantified with ProteoStat^® dye in WT, MECP2-KO (clone A33) cells (empty) or MECP2-KO cells with NEAT1 reactivation with three different sgRNAs. The graph shows the mean ± SD of three replicates (****P < .0001, one-way ANOVA). Diagram was created in BioRender. Soler (2025) https://BioRender.com/r99k484. (E) (Left) Representative transmission electron micrographs of WT, MECP2-KO (clones A33 and 3F), NEAT1-KO or MECP2-KO (A33) with NEAT1_2 overexpression or reactivation by means of CRISPRa (sgRNA8) progenitor cells. Mitophagy is indicated by arrows. Scale bars = 1 μm. (Right) Quantification of the number of endosomes in the same conditions. n = 30 micrographs, (**P < .01, ****P < .0001, ns = not significant, one-way ANOVA). (F) Histogram overlay of MitoTracker™ intensity (FITC detection) to compare mitochondria presence in WT, MECP2-KO (clones A33 and 3F) and NEAT1-KO NPCs. (G) Graphs of the mean fluorescence intensity of TMRE (membrane potential), ATP Tracker (ATP production), and MitoSOX (oxidative stress) measured by the PE-A filter, normalized with FITC-A detection of MitoTracker™. Graphs show the mean ± SD of three replicates (ns = not significant, *P < .05, **P < .01, ***P < .001, ****P < .0001, one-way ANOVA). (H) Expression levels of NEAT1, SNCA and MECP2 mRNAs were analyzed by RT-qPCR in NPCs treated with the indicated compounds. Graphs represent the mean ± SD of four independent replicas (ns = not significant, *P < .05, **P < .01, one-way ANOVA). See also Supplementary Figs S4 and S5.

Figure 5.

NEAT1 reactivation in MECP2-KO cells rescues the levels of autophagy-related family of proteins. (A) Top diagram: Autophagic vesicle maturation can be monitored in cells expressing the mCherry-GFP-LC3 reporter, which is sensitive to the increased acidic conditions in autolysosomes relative to autophagosomes. Concomitant expression of GFP- and mCherry-fused LC3 in autophagosomes is visualized as yellow–green, whereas only the expression of mCherry is allowed by acidic lysosomes and thus autolysosomes are visualized as red fluorescence. Bottom: representative images of WT, MECP2-KO (clones A33 and 3F) and NEAT1-KO NPCs transfected with the mCherry-GFP-LC3 reporter after autophagy induction (through mTOR inhibition) by administration of Rapamycin (50 μM) for 16 h. Diagram was created in BioRender. Soler (2025) https://BioRender.com/z98e924. (B) Quantification of the number of autophagosomes and autolysosomes in WT, MECP2-KO (clones A33 and 3F) and NEAT1-KO cells (n ≥ 12 cells per condition, **P < .01, ***P < .001, ns = not significant, one-way ANOVA). (C) Western blot analysis of lysosomal markers LAMP1 (left) or LAMP2 (right) levels in WT or MECP2-KO cells (clone A33) at the progenitor stage or upon free differentiation (times are indicated). (D) Western blot analysis of lysosomal markers LAMP1 and LAMP2 in WT or MECP2-KO (clone A33) NPCs, or upon transfection of MeCP2_e1 isoform. MeCP2 levels are also blotted for reference. (E) (Left) LAMP1 and LAMP2 levels assessed by western blot in WT, NEAT1-KO or MECP2-KO (clone A33) cells. (Right) LAMP1 and LAMP2 levels assessed by western blot in control and NEAT1-shRNA1 depleted cells. (F) (Left) LAMP1 levels assessed by immunofluorescence in WT, MECP2-KO (clones A33 and 3F) or NEAT1-KO NPCs. (Right) Quantification of the LAMP1 immunofluorescence signal, where the graph represents the mean ± SD of n = 6 images (***P < .001, one-way ANOVA) (G) Western blot analysis of autophagy-related proteins in WT or MECP2-KO (clones A33, left, and 3F, right) NPCs transfected with empty vector or MeCP2_e1 isoform. MeCP2 levels are also blotted for reference. (H) RT-qPCR analysis of the mRNA of the same genes shown in (G). Graphs represent the mean ± SD of four independent replicas (*P < .05, **P < .01, ****P < .0001, ns = not significant). (I) Western blot analysis of autophagy-related proteins in WT or NEAT1-KO (left), or in NEAT1-shRNA1 depleted (right) NPCs. (J) Western blot analysis of autophagy-related proteins in WT or MECP2-KO (clones A33, left, and 3F, right) NPCs transfected with empty vector or two different sgRNAs to reactivate NEAT1 expression. (K) RT-qPCR analysis of the mRNA of autophagy-related genes shown in (J). Graphs represent the mean ± SD of three independent replicates (**P < .01, ***P < .001, ****P < .0001, ns = not significant). (L) Western blot analysis of autophagy-related proteins in WT or NEAT1-KO NPCs transfected with empty vector, NEAT1_2 or MALAT1 vectors. Quantification of band intensity is shown below each western blot lane. See also Supplementary Fig. S6.

Figure 6.

NEAT1 directly interacts with mRNAs of autophagy-related genes and promotes their nuclear retention. (A) Diagram summarizing the strategy for the NEAT1 pulldown experiment. Cells were crosslinked and endogenous NEAT1 was retrieved with biotinylated, antisense probes. RNAs bound to NEAT1 were amplified by next-generation sequencing or RT-qPCR analysis. Diagram was created in BioRender. Soler (2025) https://BioRender.com/w62q250. (B) Top 10 enriched functional categories for NEAT1-AS2 probe peak-related genes (threshold q value = 0.05) from RIP-seq analysis. (C) RIP-seq peak signal profiles across ATG16L2 (GRCh38/hg38 chr11:72 814 411–72 829 635). (D) Predicted interaction between ATG16L2 (transcript NM_033388.2) with NEAT1 in the 500 nt region flanking the AS2 probe. The interaction was predicted using IntaRNA (http://rna.informatik.uni-freiburg.de/IntaRNA/) with default parameters, including a minimum seed length of 7 nucleotides. (E) RIP-qPCR analysis of the indicated autophagy-related genes. For each transcript, several amplicons were tested, whose locations are indicated above each graph. The graphs represent the mean ± SD of three independent experiments. The enrichments are normalized by the input material and shown relative to scramble (scr) probe. One-way ANOVA test was used, *P < .05, **P < .01, ****P < .0001, ns = not significant. (F) (Above) RNA-FISH showing localization of NEAT1 and ATG16L2 mRNAs in WT, NEAT1-KO, MECP2-KO (clones A33 and 3F) or MECP2-KO clones overexpressing full-length NEAT1_2 progenitor cells. NESTIN antibody and DAPI staining were used at the same time to visualize the cells. For each cell condition, the white-squared inset is amplified on the right panel. White arrows indicate cytoplasmic signals. Scale bars = 10 μm. (Below) Quantification of ATG16L2 RNA-FISH signal. Total number of foci were counted in each cell and the percentage of cytoplasmic foci per cell is represented. Graphs show the mean ± SEM of n ≥ 18 cells (*P < .05, ***P < .001, ns = not significant, one-way ANOVA). (G) RT-qPCR assessment of the nuclear-cytoplasmic distribution of the indicated autophagy-related and mitochondrial protein-encoding mRNAs. Biochemical fractionation was done in WT, NEAT1-KO and MECP2-KO (clone A33) cells transfected with empty or sgRNAs to reactivate NEAT1. Percentage of the mRNA present in each fraction relative to total mRNA is shown. Graphs represent the mean ± SD of four independent replicas. See also Supplementary Fig. S7.

Figure 7.

NEAT1 reexpression restores the morphological defects of MECP2-KO neurons. (A) Diameter length of neurospheres was measured for WT, NEAT1-KO or MeCP2-KO (clones A33 and 3F) cells (left) and upon transfection of an empty vector or a vector carrying NEAT1_2 (right). Graphs represent the mean ± SEM of at least 12 measurements (****P < .0001, one-way ANOVA). Images correspond to representative spheres of each condition. Scale bar = 0.25 mm. (B) (Left) Control WT neural progenitors, MECP2-KO (clone A33) or NEAT1-KO cells were driven towards glutamatergic differentiation for 7 days, stained with MAP2 and reconstructed in silico from confocal images with the NeuronStudio software. One representative picture for each condition is shown. (Right) Automatic analysis of the cells in (B) allowed total branch points count per each condition (graphs represent mean ± SEM, n = 15 neurons, *P < .05, one-way ANOVA). Scale bar = 10 μm. (C) For the same cells as in (B), the total filopodia count and the abundance of each filopodia subtype was determined with the NeuronStudio software (n = 15 neurons, **P < .01, ****P < .0001, one-way ANOVA). (D) The same quantification in panels (B) and (C) was represented in the histograms as distance from the soma. (E) Morphological analysis as in panels (B) and (C) with WT or MECP2-KO (clone A33) differentiated cells expressing empty vector or sgRNAs to reactivate NEAT1. Graphs represent mean ± SEM, n = 10 neurons (*P < .05, **P < .01, ***P < .001, ****P < .0001, ns = not significant, one-way ANOVA). (F) Morphological analysis as in panel (E) with MECP2-KO (clone A33) differentiated cells expressing NEAT1_2 or an empty vector as control. Graphs represent mean ± SEM, n = 15 neurons (*P < .05, ****P < .0001, two-tailed Mann–Whitney U test). See also Supplementary Fig. S8.

Figure 8.

Schematic summarizing the findings in this work. (Left) In control NPCs, the lncRNA NEAT1 regulates protein metabolism by reducing the activation of S6 protein through regulation of mTORC1 and by directly interacting with autophagy-related transcripts, promoting their biogenesis and translation. NEAT1 responds to stress stimuli and positively influences the autophagic response. (Right) In MECP2-KO cells, the marked downregulation of NEAT1 expression causes hyperactivation of S6 and protein accumulation. At the same time, dysregulation of autophagy-related genes results in the blockage of the autophagic flux, loss of protein homeostasis, and mitochondrial dysfunction, which can be observed at the functional and structural level. Altogether, neurons derived from mutant NEAT1 or MECP2 cells display similar morphological defects, which can be rescued by the reactivation of NEAT1 expression. This suggests a central role for NEAT1 in mediating the metabolic dysregulation typical of RTT. The dashed arrow indicates as yet unclear mechanisms of regulation. Diagram was created in BioRender. Soler (2025) https://BioRender.com/e64d138.