Mitochondria-derived nuclear ATP surge protects against confinement-induced proliferation defects

Abstract

The physical tissue microenvironment regulates cell state and behaviour. How mechanical confinement rewires the subcellular localisation of organelles and affects cellular metabolism is largely unknown. In this study, proteomics analysis revealed that cellular confinement induced a strong enrichment of mitochondrial proteins in the nuclear fraction. Quantitative live cell microscopy confirmed that mechanical cell confinement leads to a rapid re-localisation of mitochondria to the nuclear periphery in vitro, reflecting a physiologically relevant phenomenon in patient-derived tumours. This nucleus-mitochondria proximity is mediated by an endoplasmic reticulum-based net that entraps the mitochondria in an actin-dependent manner. Functionally, the nucleus-mitochondria proximity results in a nuclear ATP surge, which can be regulated by the genetic and pharmacological modulation of mitochondrial ATP production or via alterations of the actin cytoskeleton. The confinement-induced nuclear ATP surge has physiologically significant long-term effects on cell fitness, driven by changes in chromatin state, enhanced DNA damage repair, and cell cycle progression during mechanical cell deformation. Together, our data describe a confinement-induced metabolic adaptation that is required to enable prompt DNA damage repair and cell proliferation under mechanical confinement stress by facilitating chromatin state transitions.

Fig. 1. Proteomic analyses reveal mitochondrial enrichment within the nuclear fraction.

A Schematic of mechanical cell confinement for proteomics analysis, which is based on a 4% agarose pad for confinement through which reagents can be injected. B Schematic of the experimental strategy to apply acute confinement prior to subcellular fractionation. C Enriched and depleted terms when comparing the nuclear fraction of HeLa cells in mechanical confinement versus control conditions. D Comparison of the enriched terms from the nuclear fractions in (C), mapped to organelles based on their known localisation from the hyperLOPIT dataset. Data is obtained from three biological replicates. E Cellular components—GSEA results performed on the nuclear fractions in (C). Light blue background highlights terms related to mitochondria. Concentric circles represent normalised enrichment scores (NES). Box plots represent the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum. P values are FDR adjusted. See also Fig. S1 for the quality control of subcellular fractionation and proteomics results.

Fig. 2. Confinement induces nuclear indentations and an accumulation of mitochondria at the nuclear periphery.

A Cross-section of HeLa cells showing Hoechst (nuclei) and MitoTracker (mitochondria) in suspension and acutely confined cells. B 3D representation of (A). White arrows indicate nuclear indentations where mitochondria localise. C In silico reconstruction of nuclei based on Hoechst in cells from (A, B). Colour bar indicates z-height (µm). D X-Z view of the inner nuclear membrane (LAP2β-GFP) and mitochondria (MitoTracker) of HeLa cells in [top] suspension and [bottom] confinement. White arrows highlight indentations of the nucleus where mitochondria localise. E Schematic of the data analysis to quantify Nucleus-Associated Mitochondria (NAM). Nuclear Regions Of Interest (ROIs) were enlarged by 10%, and NAM formation was quantified as integrated signal intensity within the ROI, averaged across z-slices per cell. F NAM quantification in suspension and confined HeLa cells as outlined in (E). Each point represents a cell, n indicates the number of cells analysed. G Representative images of HeLa cells in suspension and confinement. Solid white line depicts the line profiles used in (H); dashed line indicates the 10% expanded nuclear ROI. H Hoechst and MitoTracker line profiles from the nuclear centre to the 10% expanded ROIs in representative (based on n = 20) suspension and confined HeLa cells as in (G). I Single cells from breast cancer tissue microarrays (TMA) from the tumour body or invasive front were stained for the nucleus (DAPI) and mitochondria (SDHB). White arrows show nuclear indentations and mitochondrial accumulation. Additionally, see Fig. S2N. J NAM quantification (SDHB signal within the nuclear ROI) in NAM-High or NAM-Absent/Low cells in the tumour body or tumour invasive front. Each point represents a cell; n indicates the number of cells analysed. K Penetrance of the NAM phenotype in the tumour body or invasive front. Tumours from 17 individuals were analysed. Barplot represents mean ± SEM. n indicates the number of analysed cells. Statistics in (F, J, K) were performed using the Wilcoxon–Mann–Whitney test. Box plots show the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum. See Fig. S2 for further characterisation and validation in other cell lines.

Fig. 3. Actin, mitochondrial morphology and ER play a role in regulating mitochondrial accumulation at the nuclear periphery.

A NAM levels in HeLa cells treated with Latrunculin A (500 nM), SMIFH2 (100 µM), or Jasplakinolide (500 nM), zero mean normalised with Fig. S3B (see ‘Methods’). Each point represents one cell; n is the number of cells analysed. B Corresponding images of control suspension, control confinement and drug-treated confined cells as in (A). Mitochondria are (MitoTracker) in grey; nuclei (Hoechst) are outlined in cyan. C NAM levels in HeLa cells treated with Nocodazole (10 µM), zero mean normalised. Each point is a cell; n is the number of cells analysed. D Corresponding images of control and Nocodazole-treated HeLa cells under suspension or confinement. Mitochondria are (MitoTracker) in grey; nuclei (Hoechst) are outlined in cyan. E 3D views of confined HeLa cells, untreated or Latrunculin A-treated, stained with Hoechst and MitoTracker. Arrows indicate mitochondria-containing nuclear indentations. In silico reconstructions of F untreated or G Latrunculin A-treated nuclei from (E). H Perinuclear actin (SiR-Actin) quantification in HeLa cells under adhesion, suspension, or confinement, untreated or Latrunculin A-treated. Intensity averaged across z-slices per cell. Suspension and confinement were normalised to respective adhesion. Each point is a cell; n is the number of cells analysed. I Representative images quantified in (H) of HeLa cells stained with SiR-Actin and Hoechst, untreated or Latrunculin A-treated. J NAM levels in wild-type, DRP1 KO, FIS1 KO, or MFN1 KO HeLa cells under suspension or confinement. Each point is one cell; n is the number of cells analysed. K Corresponding images of KO cells as in (J), showing mitochondria (grey) and nuclei (cyan). L Confined HeLa cells co-stained with MitoTracker (see M) and ER marker Turquoise-KDEL (see N) shown as (left) cross-sections and (right) 3D views. See also Fig. S4A for suspension. Single-plane images of the cell in (L) stained with MitoTracker (M) and Turquoise-KDEL (N). O Zoomed 3D (top) and cross-sectional (bottom) views showing mitochondria–ER interaction within a nuclear indentation. Line profiles of ER and mitochondrial signals across P nuclear diameter, or Q 1.5 μm long lines starting from mitochondrial peaks. Generalised additive models are shown with 95% confidence intervals. n is the number of cells analysed. Nuclei were centred, and the measurement distances were standardised. See also Fig. S4 for late endosomes, lysosomes and peroxisomes analyses. Statistics in (A, C, H, J) were performed using the Wilcoxon–Mann–Whitney test. Box plots represent the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum.

Fig. 4. Confinement induces a NAM-fuelled nuclear ATP surge.

A Biological processes—GSEA of proteins enriched in the confined nuclear fraction. Concentric circles represent −log10 (P value). P values are FDR adjusted. B Quantification of nuclear ATP levels in HeLa cells treated with Oligomycin A (1 μM), Pyruvate (3 mM) or combination. Cells were cultured in glucose-containing or glucose-free media as indicated by the hatched boxplot. Each data point represents a cell, and n indicates the total number of cells analysed. C Representative images of nuclear ATP levels in HeLa cells quantified in (B). Nuclear ATP is shown as the YFP/CFP ratio using a pseudo-colour scale to indicate ratio intensity and is applicable throughout this figure for ATP analysis. Brighter colours correspond to higher YFP/CFP ratios. D Nuclear ATP levels in HeLa cells, either untreated or treated with Latrunculin A (500 nM) in suspension and confinement. Each point represents a cell, and n indicates the total number of cells analysed. E Representative images of nuclear ATP in HeLa cells corresponding to the YFP/CFP ratio as quantified in (D). F Nuclear ATP levels in HeLa cells, either untreated or treated with BAM15 (10 μM). Quantification was performed per z-plane and averaged per cell. Each point represents a cell, and n indicates the total number of cells analysed. G Representative images of nuclear ATP in HeLa cells corresponding to the YFP/CFP ratio as quantified in (F). H Nuclear ATP quantification in HeLa wild-type, DRP1 KO, FIS1 KO, and MFN1 KO cells. Each point represents a cell, and n indicates the total number of cells analysed. I Representative images of nuclear ATP in HeLa wild-type or KO cells corresponding to the YFP/CFP ratio as quantified in (H). Statistics in (B, D, F, H) were performed using the Wilcoxon–Mann–Whitney test to compare means. Box plots represent the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum. All nuclear ATP was quantified using a FRET-based nuclear ATP sensor.

Fig. 5. Nuclear ATP regulates chromatin state depending on mechanical confinement.

A Western blot of H3K27me3 (facultative heterochromatin mark) and H3K9me3 (constitutive heterochromatin mark), and total H3 in control or Oligomycin A-treated suspension and acutely confined HeLa cells. α-Tubulin was used as a loading control. Quantification of H3K27me3 and H3K9me3 was normalised to their respective loading controls, and fold change was calculated relative to the Oligomycin A-treated confined condition, for three biological replicates. S indicates suspension cells and C indicates confined cells. Individual points show replicates, and bar charts show mean ± S.E. Quantification of the number of nucleoli per cell (B) and the percentage of centromeres (CENPA) at the nucleolar periphery (C) in control or Oligomycin A-treated, suspension or acutely confined HeLa cells. D Representative images of control or Oligomycin A-treated, suspension or acutely confined HeLa cells expressing a CENPA-GFP reporter, stained with SiR-DNA. E ATAC-seq accessible peaks in control or Oligomycin A-treated HeLa cells. LogFC was calculated for the comparison of acute confinement vs. suspension. Positive LogFC corresponds to differentially more accessible regions in confinement, while negative LogFC corresponds to differentially more accessible regions in suspension. Significant genes are highlighted. P values are FDR adjusted. F Chromosome-wise ATAC-Seq peaks in control or Oligomycin A-treated HeLa cells. The colour scale is based on normalised accessibility per chromosome, which was calculated as the total accessible region (in base pairs) for each chromosome divided by the respective chromosome length. G Over Representation Analysis (ORA) of significantly altered differentially accessible regions (in suspension and confinement) in Oligomycin A-treated HeLa cells. P values are FDR adjusted. H Log2-fold change of promoters of genes detected in the ATAC-Seq dataset corresponding to the gene ontology GO:0031256 leading edge membrane, enriched in (G). Annotations in black highlight genes belonging to either actin cytoskeleton organisation (GO:0030036) or cell cycle (GO:0007049). Statistics in (A) were performed using a Two-way ANOVA followed by the Student’s t-test for specific interactions. Statistics in (B, C) were performed using the Wilcoxon–Mann–Whitney test to compare means. Box plots represent the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum.

Fig. 6. Nuclear ATP during confinement is required for efficient DNA damage response signalling and safeguards S-phase progression after confinement release.

A–D Visualisation and quantification of 53BP1 in U2OS cells expressing a truncated-53BP1-Maroon reporter. Images are maximum intensity projections with nuclear outlines based on transmitted light. Quantification was performed on maximal projections for cells treated with A, B Latrunculin A (500 nM) or C, D Oligomycin A (1 µM) versus controls. Each point represents a cell; n indicates the number of cells analysed. E Protocol to study the effect of acute confinement on cell dynamics and DNA damage repair upon confinement release. U2OS-FUCCI cells were given a 30-min treatment of Oligomycin A, followed by acute confinement for 15 min. Confinement and Oligomycin A treatment were then released, cells allowed to attach under normal growth conditions and single cells were tracked for 36 h. F Quantification of 53BP1 foci in U2OS cells post-confinement release. Data is from 3 biological replicates (number of cells analysed n > 3500). G Relative cell proliferation of unconfined or confined U2OS cells under control or Oligomycin A-treated (1 μM) conditions. Dark lines are best fit lines based on a generalised additive model, and shaded areas are 95% confidence intervals. At timepoint 1, n > 2000 for three biological replicates. H Progression of cells detected in the S-phase at the beginning of tracking, monitored for 36 h after confinement release. Data is shown as the mean of three biological replicates. I Single-cell analysis of FUCCI expression was performed by spatially distributing cells based on standardised signal intensity (see ‘Methods’). S-phase is shown in green. J Single cell representation of cells first detected in the S-phase after confinement release at the first (1st hour) and last (36th hour) time points. K Progression of the S-phase population at 3-h intervals. Statistics in (A, C, F) were performed using the Wilcoxon–Mann–Whitney test to compare means. Statistics in (G, H) were performed using the Wilcoxon test to compare the distribution/ranks of the groups. Box plots represent the lower quartile, median and upper quartile. Whiskers represent the minimum and maximum. See also Fig. S7 for FUCCI, G1-phase and G2 phase.