Nuclear Deformation Causes DNA Damage by Increasing Replication Stress

Abstract

Cancer metastasis, i.e., the spreading of tumor cells from the primary tumor to distant organs, is responsible for the vast majority of cancer deaths. In the process, cancer cells migrate through narrow interstitial spaces substantially smaller in cross-section than the cell. During such confined migration, cancer cells experience extensive nuclear deformation, nuclear envelope rupture, and DNA damage. The molecular mechanisms responsible for the confined migration-induced DNA damage remain incompletely understood. Although in some cell lines, DNA damage is closely associated with nuclear envelope rupture, we show that, in others, mechanical deformation of the nucleus is sufficient to cause DNA damage, even in the absence of nuclear envelope rupture. This deformation-induced DNA damage, unlike nuclear-envelope-rupture-induced DNA damage, occurs primarily in S/G2 phase of the cell cycle and is associated with replication forks. Nuclear deformation, resulting from either confined migration or external cell compression, increases replication stress, possibly by increasing replication fork stalling, providing a molecular mechanism for the deformation-induced DNA damage. Thus, we have uncovered a new mechanism for mechanically induced DNA damage, linking mechanical deformation of the nucleus to DNA replication stress. This mechanically induced DNA damage could not only increase genomic instability in metastasizing cancer cells but could also cause DNA damage in non-migrating cells and tissues that experience mechanical compression during development, thereby contributing to tumorigenesis and DNA damage response activation.

Keywords: DNA damage; cancer; cell compression; confined migration; metastasis; nuclear deformation; nuclear envelope rupture; nuclear mechanobiology; replication stress.

Figure 1:. NE rupture and nuclear deformation lead to DNA damage during confined migration.

(A) Representative image panel showing a HT1080 fibrosarcoma cell co-expressing NLS-GFP and 53BP1-mCherry exhibiting DNA damage following NE rupture during migration through a 1 × 5 μm² constriction in the microfluidic device. Red arrowhead indicates start of NE rupture; the red line indicates the duration of NE rupture; white arrowheads indicate newly occurring 53BP1-mCherry foci. Scale bar: 5 μm (B) Percentage of HT1080 cells with new DNA damage (53BP1-mCherry foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 372 cells) or 15 × 5 μm² control channels (n = 268 cells). **, p < 0.01 based on unpaired t-test with Welch’s correction. (C) Percentage of HT1080 cells in which new DNA damage during migration through ≤ 2 × 5 μm² constrictions was associated with either NE rupture or with nuclear deformation in the absence of NE rupture. n = 372 cells; ***, p < 0.001 based on unpaired t-test with Welch’s correction. (D) Representative image sequence showing a MDA-MB-231 breast cancer cell co-expressing NLS-GFP and 53BP1-mCherry experiencing new DNA damage during migration through a 2 × 5 μm² constriction. White arrowheads indicate newly occurring 53BP1-mCherry foci. Scale bar: 5 μm (E) Percentage of MDA-MB-231 cells with new DNA damage (53BP1-mCherry foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 381 cells) or 15 × 5 μm² control channels (n = 196 cells). *, p < 0.05 based on unpaired t-test with Welch’s correction. (F) Percentage of MDA-MB-231 cells in which new DNA damage during migration through ≤ 2 × 5 μm² constrictions was associated with either NE rupture or with nuclear deformation in the absence of NE rupture. *, p < 0.05 based on unpaired t-test with Welch’s correction. (G) Association of new DNA damage incurred during migration through ≤ 2 × 5 μm² constrictions with either NE rupture (Rupture) or nuclear deformation without NE rupture (Deformation), for a panel of cell lines. The results correspond to the data presented in Figure 1C and 1F, and Figure S1G–I. (H) Representative image sequence of a MDA-MB-231 cell co-expressing NLS-GFP and 53BP1-mCherry incurring DNA damage during migration in a dense (1.7 mg/ml) collagen matrix. White arrowheads indicate newly occurring 53BP1-mCherry foci. Scale bar: 5 μm (I) Percentage of MDA-MB-231 cells with new DNA damage (53BP1-mCherry foci), comparing cells that migrate (n = 48 cells) with those that remain stationary (n = 29 cells) in a collagen matrix (1.7 mg/ml). *, p < 0.0001 based on Fisher’s test. (J) Percentage of MDA-MB-231 cells in which new DNA damage during migration through a collagen matrix was associated with either NE rupture or with nuclear deformation in the absence of NE rupture. n = 33 cells, *, p < 0.05 based on Chi-square test. Data in this figure are presented as mean + S.E.M. See also Videos S1–S3, Figure S1, S2 and S4.

Figure 2:. Nuclear compression is sufficient to cause DNA damage.

(A) Schematic of the custom-built microfluidic compression device with a PDMS piston. The device is connected to a suction source which causes the PDMS piston, with a small circular coverslip attached, to move down onto the cells. Polystyrene beads serve as spacers to ensure a uniform height between the glass coverslip and the glass bottom of the dish. Inset shows cells compressed between the PDMS piston with attached cover slip and the glass surface and the polystyrene beads. (B) Representative image sequence showing the nuclear height of a MDA-MB-231 breast cancer cell expressing H2B-mScarlet, compressed to either 5 μm, 3 μm, or 2 μm height using the compression device. White lines indicate the height of the compressed cell. Scale bar: 5 μm (C) Representative image sequence showing a MDA-MB-231 breast cancer cell expressing 53BP1-mCherry with new DNA damage formation during compression to either 5 μm or 2 μm (bottom) height. White arrowheads indicate newly occurring 53BP1-mCherry foci; black line indicates the duration of compression. Scale bar: 5 μm (D) Percentage of MDA-MB-231 cells with new DNA damage (53BP1-mCherry foci) in unconfined conditions (n = 389 cells) or during compression to 5 μm height (n = 500 cells), 3 μm height (n = 378 cells), or 2 μm height (n = 411 cells). **, p < 0.01; ***, p < 0.001; ****, p < 0.0001, based on ordinary one-way ANOVA with Dunnett’s multiple comparison test. (E) Schematic overview of the AFM-LS system. A micro-mirror is lowered adjacent to a cell of interest and a vertical light sheet propagates out of the objective illuminating a x-z cross-section of the cell. The image plane is raised to intersect the mirror, capturing the virtual image created by the mirror. The AFM cantilever is positioned between the mirror and cell in order to probe the cell from above while imaging the side-view cross section. (F) Representative image sequence showing a MDA-MB-231 breast cancer cell co-expressing NLS-GFP and 53BP1-mCherry experiencing new DNA damage during compression to a height of ~2 μm with an AFM cantilever. Black arrowhead indicates the AFM cantilever; white arrowhead indicates newly occurring 53BP1-mCherry foci. Scale bar: 5 μm (G) Percentage of MDA-MB-231 cells with new DNA damage (53BP1-mCherry foci) during compression by an AFM tip (n = 21 cells) or in uncompressed control conditions (n = 19 cells). **, p < 0.01 based on Fisher’s test. Data in this figure are presented as mean + S.E.M. See also Video S4, Figure S3 and S4.

Figure 3:. Deformation associated DNA damage occurs specifically in S/G2 phase of the cell cycle.

(A) Representative image sequence showing a MDA-MB-231 breast cancer cell expressing FUCCI reporter transitioning from S/G2 to M and to G1 cell cycle phase. Scale bar: 5 μm (B) Representative image sequence of a MDA-MB-231 breast cancer cell co-expressing FUCCI and 53BP1-mCherry experiencing new DNA damage while in S/G2 phase of the cell cycle during migration through a 2 × 5 μm² constriction. White arrowheads indicate newly occurring 53BP1-mCherry foci. Scale bar: 5 μm (C) Representative image sequence of a MDA-MB-231 breast cancer cell co-expressing FUCCI and 53BP1-mCherry experiencing new DNA damage while in G0/G1 phase of the cell cycle during migration through a 2 × 5 μm² constriction. White arrowheads indicate newly formed 53BP1-mCherry foci. Scale bar: 5 μm (D) Percentage of MDA-MB-231 cells with new DNA damage (53BP1-mCherry foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 327 cells) or 15 × 5 μm² control channels (n = 145 cells) as a function of cell cycle phase (G0/G1 or S/G2). **, p < 0.01 based on two-way ANOVA with Tukey’s multiple comparison test. (E) Percentage of HT1080 fibrosarcoma cells with new DNA damage (53BP1-mCherry foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 850 cells) or 15 × 5 μm² control channels (n = 371 cells) as a function of cell cycle phase (G0/G1 or S/G2). *, p < 0.05 based on two-way ANOVA with Tukey’s multiple comparison test. (F) Percentage of MDA-MB-231 cells in G0/G1 or S/G2 phase of the cell cycle in unconfined conditions (n = 3544 cells), or during migration through small (≤ 2 × 5 μm²) constrictions (n = 327 cells) or 15 × 5 μm² control channels (n = 145 cells). Differences were not statistically significant (n.s.) based on two-way ANOVA. (G) Percentage of HT1080 cells in G0/G1 or S/G2 phase of the cell cycle in unconfined conditions (n = 6108 cells) or during migration through small (≤ 2 × 5 μm²) constrictions (n = 850 cells) or 15 × 5 μm² control channels (n = 371 cells). Differences were not statistically significant based on two-way ANOVA. Data in this figure are presented as mean + S.E.M. See also Figure S5.

Figure 4:. Deformation induced DNA damage occurs at replication forks:

(A) Representative image panel of a MDA-MB-231 breast cancer cell expressing 53BP1-mCherry during migration through a 1 × 5 μm² constriction, stained for p-RPA S33 to reveal co-localization between p-RPA S33 foci and 53BP1-mCherry foci. White arrowheads indicate sites of co-localization. Scale bar: 5 μm (B) Representative image panel showing a MDA-MB-231 breast cancer cell co-expressing GFP-PCNA and 53BP1-mCherry experiencing new DNA damage at replication forks during migration through a 2 × 5 μm² constriction. White arrowheads indicate sites of co-localization between newly occurring 53BP1-mCherry foci and GFP-PCNA foci. Scale bar: 5 μm (C) Percentage of MDA-MB-231 cells with co-localization between new DNA damage (53BP1-mCherry foci) and replication forks (GFP-PCNA foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 584 cells) or 15 × 5 μm² control channels (n = 490 cells). **, p < 0.01 based on unpaired t-test with Welch’s correction. (D) Percentage of HT1080 cells with co-localization between new DNA damage (53BP1-mCherry foci) and replication forks (GFP-PCNA foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 986 cells) or 15 × 5 μm² control channels (n = 641 cells). Differences were not statistically significant based on unpaired t-test. (E) Representative image sequence of a MDA-MB-231 cell co-expressing GFP-PCNA and 53BP1-mCherry incurring DNA damage at replication forks during migration in a dense (1.7 mg/ml) collagen matrix. Inset depicts close-up of the region inside the white rectangle to show occurrence of new 53BP1-mCherry foci at replication forks marked by GFP-PCNA foci. White arrowheads indicate site of co-localization between newly occurring 53BP1-mCherry foci and GFP-PCNA foci. Scale bar: 5 μm. Data in this figure are presented as mean + S.E.M. See also Video S5, Video S6 and Figure S6.

Figure 5:. Nuclear deformation leads to increased replication stress.

(A) Representative images of MDA-MB-231 breast cancer cells showing EdU incorporation during either mild (5 μm height, top) or severe (2 μm height, bottom) compression. Scale bar: 20 μm (B) Fluorescence intensity of incorporated EdU per nucleus in MDA-MB-231 cells following compression to either 5 μm height (n = 171 cells) or 2 μm height (n = 184 cells) for 2 hours. *, p < 0.05 based on unpaired t-test with Welch’s correction. (C) Fluorescence intensity of incorporated EdU per nucleus in HT1080 fibrosarcoma cells following compression to either 5 μm height (n = 237 cells) or 2 μm height (n = 195 cells) for 2 hours. Differences were not statistically significant based on unpaired t-test with Welch’s correction. (D) Representative images of a MDA-MB-231 breast cancer cell showing replication forks (p-RPA S33 foci) during compression to either 5 μm or 2 μm (bottom) height. Scale bar: 5 μm (E) Average number of p-RPA S33 foci in MDA-MB-231 cells following compression to 5 μm height (n = 247 cells) or 2 μm height (n = 318 cells) for 2 hours. **, p < 0.01 based on unpaired t-test with Welch’s correction. (F) Average number of p-RPA S33 foci in HT1080 cells following compression to 5 μm height (n = 207 cells) or 2 μm height (n = 159 cells) for 2 hours. Differences were not statistically significant based on unpaired t-test with Welch’s correction. (G) Percentage of MDA-MB-231 and HT1080 cells with increase in replication forks (GFP-PCNA foci) during migration through small (≤ 2 × 5 μm²) constrictions (n = 584 cells for MDA-MB-231; n = 986 cells for HT1080) or 15 × 5 μm² control channels (n = 490 cells for MDA-MB-231; n = 641 cells for HT1080). *, p < 0.05 based on two-way ANOVA with Tukey’s multiple comparison test. Data in this figure are presented as mean + S.E.M.

Figure 6.. Model for DNA damage during confined migration and cell compression.

Nuclei experience severe deformation during cell compression or migration through confined spaces. A subset of cells additionally experience transient NE rupture during these processes. Both nuclear deformation and NE rupture can lead to DNA damage as observed in our experiments, but via separate mechanisms. NE rupture allows uncontrolled exchange of large molecules between the nucleus and cytoplasm. This could lead to a loss of DNA repair factors such as Ku80 or BRCA1 from the nucleus into the cytoplasm, and allow influx of cytoplasmic nucleases such as TREX1 into the nucleus, thereby causing DNA damage. On the other hand, nuclear deformation associated with cell compression or confined migration can alter DNA conformation and make it more difficult to unwind the DNA ahead for replication, leading to an increase in replication stress in the cells. The increased replication stress could be mediated through replication fork stalling or collapse, or aberrant replication stress response by ATR, ultimately resulting in increased DNA damage at replication forks.