Heterochromatin loss as a determinant of progerin‐induced DNA damage in Hutchinson–Gilford Progeria

Abstract Hutchinson–Gilford progeria is a premature aging syndrome caused by a truncated form of lamin A called progerin. Progerin expression results in a variety of cellular defects including heterochromatin loss, DNA damage, impaired proliferation and premature senescence. It remains unclear how these different progerin‐induced phenotypes are temporally and mechanistically linked. To address these questions, we use a doxycycline‐inducible system to restrict progerin expression to different stages of the cell cycle. We find that progerin expression leads to rapid and widespread loss of heterochromatin in G1‐arrested cells, without causing DNA damage. In contrast, progerin triggers DNA damage exclusively during late stages of DNA replication, when heterochromatin is normally replicated, and preferentially in cells that have lost heterochromatin. Importantly, removal of progerin from G1‐arrested cells restores heterochromatin levels and results in no permanent proliferative impediment. Taken together, these results delineate the chain of events that starts with progerin expression and ultimately results in premature senescence. Moreover, they provide a proof of principle that removal of progerin from quiescent cells restores heterochromatin levels and their proliferative capacity to normal levels.

Murray, Warren, Liu, & Shanahan, 2016;Hilton et al., 2017;Wheaton et al., 2017). However, deciphering the causal and temporal links between the different progerin-induced phenotypes remains challenging as the majority of studies have been conducted in patient-derived cells, or cells constitutively expressing progerin, where immediate consequences of progerin expression and secondary effects arising from progerin-induced senescence cannot be distinguished.
We previously reported a doxycycline-inducible system to express physiological levels of progerin in isogenic primary-and TERT-immortalized human dermal fibroblasts (NDF) and found that expression of TERT prevents progerin-induced premature senescence (Chojnowski et al., 2015;Kudlow, Stanfel, Burtner, Johnston, & Kennedy, 2008). However, TERT did not prevent progerin-induced heterochromatin loss and nuclear abnormalities (Chojnowski et al., 2015). This unique system allows us to distinguish what may be a cause or consequence of progerin-induced senescence.
Here, we used this experimental system to temporally restrict progerin expression to particular cell cycle stages and to determine F I G U R E 1 Progerin-dependent loss of H3K9me3 and H3K27me3 heterochromatin marks in G1-arrested confluent NDF. (a,b) Immunofluorescence microscopy of H3K9me3 (a) and H3K27me3 (b) staining in G1-arrested cells in the presence or absence of progerin. V5progerin (V5 antibody) and H3K9me3/H3K27me3 antibodies are indicated, with DAPI overlay. Scale bar: 50 μm. Progerin-expressing cells with extensive loss of H3K9me3 or H3K27me3 are marked by white arrowheads. (c,d) Scatter plot analysis of H3K9me3 (c) or H3K27me3 (d) and progerin levels in confluent or proliferating NDF in the presence (red, orange) or absence (blue, light blue) of progerin. H3K9me3 or H3K27me3 and progerin normalized intensities are plotted on Y and X axis, respectively. A total of ~9 × 10 3 and ~7 × 10 3 nuclei were quantified for H3K9me3 and H3K27me3 analysis, respectively, from 3 independent experiments. (e) Electron microscopy imaging of peripheral heterochromatin in confluent NDF in the presence (lower panel) or absence (upper panel, red arrowheads) of progerin. Nucleus (N) and cytoplasm (C) are indicated, scale bars: 5μm (left panels) and 1μm (right panels). (f) Quantification of perinuclear heterochromatin from TEM images in the absence (−DOX) and presence (+DOX) of progerin (see Figure S1-1a) (*p < .05, n = 11 cells per condition, Mann-Whitney test)  can be reversed upon progerin removal without resulting in any lasting cell proliferation defects.

| Progerin-induced heterochromatin loss is independent of DNA replication and mitosis
We and others previously showed that progerin expression triggers extensive heterochromatin loss, a phenotype observed in both in vitro models and patient cells (Chojnowski et al., 2015;Scaffidi & Misteli, 2005;Shumaker et al., 2006). In addition, we demonstrated that TERT expression prevents progerin-induced senescence, without alleviating heterochromatin loss, suggesting that the heterochromatin loss is not a consequence of cellular senescence (Chojnowski et al., 2015). To further characterize the temporal dynamics of progerin-induced heterochromatin loss and to investigate whether it is contingent upon DNA replication or mitosis, we restricted progerin expression to G1arrested cells and studied heterochromatin and progerin levels by quantitative single-cell immunofluorescence microscopy.

F I G U R E 2
To achieve this, we grew cells to confluence, induced progerin expression and then quantified their heterochromatin levels.
Significantly, the correlation between progerin expression and H3K9me3 and H3K27me3 loss was similar between G1-arrested and control cells allowed to proliferate (Pearson r = −0.44/−0.42 for H3K9me3 and −0.40/−0.56 for H3K27me3 between confluent and proliferating cells, respectively). We then used 3D structured illumination microscopy (3D-SIM) analysis to determine the spatial distribution of H3K27me3 and found that progerin-induced loss of heterochromatin was not due to re-localization of peripheral F I G U R E 3 Preferential accumulation of DNA damage foci in cells with lower levels of H3K9me3 and H3K27me3. (a) Scatter plot analysis of H3K9me3, progerin expression and γH2AX DNA damage foci number per nucleus, in NDF expressing progerin (red) or nonexpressing controls (blue). For each nuclei, H3K9me3 and progerin normalized intensities are plotted on Y and X axis, respectively, while the number of DNA damage foci is represented by colour intensity (light to dark colour: 0, 1, 2, 3 or more DNA damage foci per nucleus). Inset: DNA damage counts of the same data (***p < .001, Student's t test). A total of ~ 1x10 4 cells from 3 independent experiments were analysed. (b) Box plot of data shown in (a), whiskers represent 10-90 percentile (***p < .001, one-way ANOVA with Tukey's post-test). (c) Scatter plot analysis of H3K27me3, progerin expression and γH2AX DNA damage foci number. Inset: DNA damage counts of the same data (***p < .001, Student's t test). A total of ~ 4x10 3 cells from 3 independent experiments were analysed. (d) Box plot of the same data, whiskers represent 10-90 percentile (***p < .001, *p < .05, one-way ANOVA with Tukey's post-test). (e,f) 3D-SIM analysis of DNA damage foci. Illustration of the Eroded Volume Fraction (EVF) index to assess proximity to the nuclear lamina, from closest (0) to furthest (1) (reproduction of Figure S1-1b).
(f,g) Quantification of the average minimum EVF value for 53BP-1 (f) and γH2AX (g) DNA damage foci, ± progerin. A total of 225 53BP-1 and 135 γH2AX DNA damage foci were analysed. (*p < .05, Student's t test) heterochromatin towards the centre of the nucleus ( Figure S1-1b,c). These results demonstrate that progerin expression triggers rapid and extensive loss of peripheral heterochromatin in both quiescent and proliferating cells, without the need for DNA replication or mitosis, suggesting that the mechanism of the heterochromatin loss in proliferative and nonproliferative cells may be similar.

| Progerin-induced DNA damage occurs exclusively in replicating cells
Next, we tested whether the progerin-induced heterochromatin loss and DNA damage could be temporally separated to different stages of the cell cycle. Previous studies showed that progerininduced DNA damage can occur during DNA replication, although significant amounts of DNA damage were also apparent in G1 cells (Wheaton et al., 2017). To precisely determine the timing of progerin-induced DNA damage, we temporally restricted progerin expression to G1-arrested cells using our doxycycline-inducible lentiviral system. We grew NDF harbouring doxycycline-inducible progerin to complete confluence and induced progerin expression by adding doxycycline to the cell culture media, while con- Taken together, these results demonstrate that progerin-induced DNA damage occurs exclusively in replicating cells, despite G1-arrested cells retaining normal DNA damage response capabilities. This is in stark contrast to the heterochromatin loss induced by progerin expression, which also affects G1-arrested cells.

| Progerin-induced DNA damage occurs preferentially in cells with low heterochromatin and specifically during late stages of DNA replication, prior to chromosome condensation
Heterochromatin decompaction renders some cell types more susceptible to DNA damage (Di Micco et al., 2011). To test whether progerin-expressing cells with low levels of heterochromatin were more prone to accumulate DNA damage, we quantified heterochromatin levels (H3K27me3 and H3K9me3), DNA damage foci (by γH2AX) and progerin levels (by v5-tag) in proliferating dermal fibroblasts (NDF) in the presence or absence of progerin ( Figure 3).
We observed an inverse correlation between progerin-induced DNA formation upon treatment with the chemotherapeutic agent doxorubicin. As shown in Figure S4-2a,b, we did not detect any significant delay in the DNA damage response in progerin-expressing cells ( Figure S4-2a,b).
Secondly, we introduced mCherry-BP1-2 (mCherry fused to amino acids 1220-1711 of 53BP-1) ( Together, these results indicate that TERT-preventable progerin-induced DNA damage occurs exclusively during late stages of DNA replication, when heterochromatin is normally replicated, prior to chromosome condensation and mitosis.

| Removal of progerin in G1-arrested cells restores heterochromatin levels and results in no permanent proliferation defect
Our results demonstrate that progerin-induced heterochromatin loss is a dynamic process occurring independently of DNA replication.
However, it remained unclear whether the heterochromatin loss was a permanently acquired phenotype of cells expressing progerin or whether it could be restored upon progerin removal. To address this question, we grew cells to confluence, induced progerin expression for 4 days and subsequently withdrew doxycycline from the cell culture media. After doxycycline removal, we monitored progerin and heterochromatin levels at 3, 6, 9 and 12 days (Figure 5a). Cells expressing progerin throughout the experiment and noninduced cells served as controls. As shown in Figure 5b,d, progerin levels gradually declined upon removal of doxycycline. Importantly, H3K9me3 and H3K27me3 levels were restored concurrently with declining progerin levels (Figure 5c,e). In contrast, cells that were permanently exposed to progerin exhibited low levels of heterochromatin, while uninduced cells showed normal levels. These results demonstrate that (a) heterochromatin levels rapidly increase upon progerin removal and (b) that this process does not require DNA replication or mitosis.
To investigate whether cells transiently exposed to progerin expression exhibit any long-term defects, we expressed progerin in G1-arrested cells for 4 days, upon which doxycycline was removed ( Figure 5a). 8 days after doxycycline removal, we plated cells at low density to allow proliferation and assessed their proliferative capacity. Strikingly, the proliferation rate, DNA damage levels and senescence-associated-β-galactosidase (SA-β-gal) staining of cells transiently exposed to progerin were indistinguishable from cells that were never exposed to progerin (Figure 5f, Figure S5a-d). In contrast, cells constitutively exposed to progerin exhibited impaired proliferation, increased DNA damage and stained positive for SAβ-gal as described previously (Benson, Lee, Aaronson, & a, 2010;Chojnowski et al., 2015;Kudlow et al., 2008). These results demonstrate that removal of progerin from cells prior to DNA replication restores heterochromatin levels without resulting in any lasting proliferative defect. In addition, these results highlight that chromatin compaction upon progerin removal does not require the cell to undergo DNA replication or mitosis.

| D ISCUSS I ON
The loss and disorganization of heterochromatin during aging is now well-documented, from the loss of lamina-associated heterochromatin during normal and premature aging (McCord et al., 2013;Shumaker et al., 2006;Tsurumi & Li, 2012;Whitton et al., 2018;Zhang et al., 2015), to oxidative-stress induced aging (Osanai et al., 2018) and aging of the immune system (Keenan & Allan 2018).
Heterochromatin loss is accompanied by extensive remodelling of the nuclear lamina (Chojnowski, Ong, & Dreesen, 2014;Dreesen et al., 2013), and it has been proposed to be determinant of aging in yeast, flies and worms (Tsurumi & Li, 2012). However, the functional role of the heterochromatin loss in HGPS and how it relates to other progerin-induced defects remains unclear. By temporally restricting the expression of progerin, we demonstrate for the first time that progerin-dependent heterochromatin loss is a rapid and dynamic process, independent of DNA replication or mitosis. Furthermore, we show that this loss can be reversed, as progerin removal restored heterochromatin levels and resulted in no permanent damage or proliferation defect. These results are reminiscent of the reprogramming of HGPS fibroblasts into induced pluripotent stem cells, during which expression of A-type lamins and progerin is silenced and heterochromatin levels are restored (Chen et al., 2017;Zhang et al., 2011).
What is the physiological relevance of the heterochromatin loss in progeric cells? We demonstrate for the first time that progerin-induced DNA damage occurs preferentially in cells with lower levels of heterochromatin. This is in agreement with previous results showing that the relaxation of heterochromatin results in increased DNA damage, a property used by HDAC inhibitors for cancer treatment (Di Micco et al., 2011). In addition, we find that progerin expression per se does not alter the temporal dynamics of the DNA damage response. However, since we analysed the early consequences of F I G U R E 5 Restored heterochromatin levels and cell proliferation upon progerin removal in G1-arrested NDF. (a) NDF were grown to confluence and exposed to progerin expression (+DOX, red circles) for 4 days. Nonexposed cells (−DOX, black circles, blue box) served as a control. After 4 days, progerin expression was either turned off (−DOX; black/red circles, grey box) or left on (+DOX, red circles, red box) and kept confluent for an additional 12 days. Proliferation rates were compared for cells never exposed to progerin (blue), cells in which progerin was removed (grey) and cells constitutively exposed to progerin (red). (b) Western blotting showing doxycycline-dependent progerin expression and removal in confluent primary fibroblasts. Progerin (PG), lamin A (LA) and lamin C (LC), lamin B1 (LB1) and GAPDH are indicated. Progerin intensities quantified by single-cell immunofluorescence microscopy at each day are indicated in panel (d). A total of ~ 7.7 × 10 4 nuclei were quantified for both H3K9me3 and H3K27me3 quantifications, from 2 independent experiments. Whiskers represent 10-90 percentile. (c,e) Box plot representation of H3K9me3 (c) and H3K27me3 (e) levels at the indicated times post doxycycline removal (0, 3, 6, 9, 12 days; grey bars) or in controls never (blue) or constitutively (red) exposed to progerin. A total of ~3.8 × 10 4 and ~3.9 × 10 4 nuclei were quantified for H3K9me3 and H3K27me3, respectively, from 2 independent experiments, ***p < .001, one-way ANOVA with Sidak's post-test. Whiskers represent 10-90 percentile. (d) Box plot representation of progerin normalized intensity per nucleus, in NDF induced to express progerin (grey, red) or noninduced (blue/black). (f) Growth curve of control and primary fibroblasts continuously expressing progerin (red line), not expressing progerin (blue line), or expressing progerin for 4 days while confluent and without induction thereafter (black). Dotted lines indicate SEM (n = 3). Inset: growth rate after 8 days, error bars indicate SEM (*p < .05, one-way ANOVA with Tukey's post-test) Slope (AU/hour).10 -2 * progerin expression, we cannot exclude the possibility that cells with an extended history of progerin exposure may display such impediment (Liu et al., 2005).
Importantly, we demonstrate that heterochromatin loss by itself is not sufficient for progerin to induce DNA damage: progerin-expressing cells also have to undergo DNA replication to accumulate DNA damage, as suggested earlier (Wheaton et al., 2017). Moreover, we provide evidence that the onset of progerin-induced DNA damage occurs exclusively during the late stages of DNA replication, prior to chromosome condensation and mitosis. Interestingly, this corresponds with the replication timing of heterochromatin and telomeres near the periphery, late in S-phase (Arnoult et al., 2010;Rhind & Gilbert, 2013). The timing of the progerin-induced DNA damage is also consistent with the fact that TERT's ability to rescue such DNA damage (Benson et al., 2010;Chojnowski et al., 2015;Kudlow et al., 2008;Li et al., 2019) is only apparent during DNA replication and leaves background DNA damage foci occurring at other stages entirely unaffected. However, the precise mechanisms how TERT, or hTERT mRNA prevent, or repair progerin-induced DNA damage, remains to be determined.
The mechanism by which progerin triggers such rapid and widespread heterochromatin loss remains elusive. Previous studies showed decreased levels of several chromatin-modifying enzymes in progeric cells such as HP1α, Suv39 and EZH2 (McCord et al., 2013;Scaffidi & Misteli, 2005;Shumaker et al., 2006). However, since these experiments were done mostly in HGPS cells, it remains difficult to ascertain whether these changes were direct consequences of progerin expression, preceding senescence or DNA damage-related phenotypes.
Lastly, not all cells with low levels of heterochromatin accumulate progerin-induced DNA damage during DNA replication.
Stochastic processes may account for this variability, but additional factors could also come into play. Hilton and colleagues suggested that progerin or prelamin A sequester the proliferating cell nuclear antigen (PCNA) away from the replication fork, leading to fork collapse and subsequent DNA damage signalling activation (Cobb et al., 2016;Hilton et al., 2017).
Together, we demonstrate that progerin expression in G1arrested cells triggers rapid and extensive heterochromatin loss without causing any DNA damage. Progerin-induced DNA damage accumulates exclusively during late stages of DNA replication, prior to chromosome condensation, and preferentially in cells with low levels of heterochromatin. The necessity for cell proliferation may be more central to the pathophysiology of progeria than expected: studies have shown that multiple cell types undergo proliferation in tissues previously deemed postmitotic. For HGPS patients, whose cardiovascular system is particularly affected, this includes two of the heart's fundamental components: vascular smooth muscle cells and cardiomyocytes (Anversa & Nadal-Ginard, 2002;Bergmann et al., 2009Bergmann et al., , 2015. In addition, the importance of cycling cell types in progeria is underscored by a recent report showing that progerin expression restricted to vascular smooth muscle cells recapitulates many of the HGPS-associated cardiovascular defects (Hamczyk, Campo, & Andrés, 2018). Lastly, we provide evidence that heterochromatin levels are rapidly restored upon progerin removal and that transient exposure to progerin in G1-arrested cells does not result in any lasting proliferative defect. These results define the causal and temporal link between progerin-induced heterochromatin loss, DNA replication and DNA damage. They delineate a chain of events that commences with progerin expression and ultimately results in premature senescence, an essential step towards a better understanding of the pathophysiology of progeria.

| Cell culture & western blotting
NDF culture, progerin expression and Western blotting were conducted as described previously (Chojnowski et al., 2015).

| Culture of cells in silicone inserts
Ibidi cell culture inserts (Ibidi Cat. # 80209) were mounted on 24-mm-diameter coverslips (Nunc Cell Culture, Thermo Fisher). NDF harbouring doxycycline-inducible pTRIPZ-v5-progerin were grown to confluence within culture inserts. NDF were induced ± 1 μg/ml doxycycline for 3 days and culture inserts removed to initiate cell proliferation and migration at the edge. Cells were fixed 72 hr postinsert removal and analysed by immunofluorescence microscopy.

| Immunofluorescence microscopy
Cells were seeded onto chambered cover glass (Nunc) and fixed.
Immunofluorescence staining was conducted as described previously (Chojnowski et al., 2015). Images were acquired using Olympus FV1000 inverted laser scanning confocal microscope, IX-83 inverted wide-field fluorescent microscope (Olympus), Ts2-FL inverted microscope (Nikon) or DeltaVision OMX Blaze microscope (Olympus) (Chojnowski et al., 2015). For time lapse, images were taken every 30min on the Olympus IX-83 fluorescent microscope, for the duration of the experiment. All images were processed and quantified using ImageJ (Schindelin et al., 2012) and CellProfiler (Carpenter et al., 2006). Nuclei were segmented using the Otsu thresholding method, and for each nuclei, average intensities were measured, unless otherwise stated for EVF analysis. EVF analysis was performed using TANGO (Ollion et al., 2013) as an implementation of the EVF described previously (Ballester et al., 2008).

| Transmission electron microscopy
Cells were grown on glass coverslips and processed by flat-embedding for electron microscopy as follows: samples were fixed in 2.5% glutaraldehyde and 4% formaldehyde in 0. were quantified from TEM images using ImageJ, as the proportion of dark pixels along the nuclear lamina. Pixels were considered as dark (electron-dense) when their intensity was below the mean grey value of typical heterochromatin clusters.

| Cell cycle analyses by FACS
Cells were trypsinized at indicated time points, using 0.25% Trypsin/EDTA (GIBCO, Thermo Scientific), and trypsin was neutralized in cell culture media containing 15% FBS. Cells were spun at 300× g for 5min, resuspended in 10% FBS in PBS and fixed by adding ice-cold 100% ethanol, drop-wise, while shaking the tube.
Cells were then kept at −20ºC for at least 4 hr or until they are ready for staining. DNA was stained using 10 µg/ml propidium iodide in PBS with 10 µg/ml RNase A, for 1 hr at room temperature, in the dark. Samples were analysed on the BD TM LSRII (Becton Dickinson) flow cytometer.

| Statistical analysis
Data and statistical analyses were performed using Excel and GraphPad Prism software. Results are shown as mean ± SEM/SD, and box plot whiskers indicate 10-90 percentile, unless otherwise stated.
Data were analysed using one-or two-way ANOVA and Tukey's post hoc test if required, as well as chi-squared, two-tailed Student's t test and Pearson correlation coefficients, as appropriate. p-values below.05 were considered significant. For experiments comprising large sample sizes, significant p-values were reported only when the effect size was at least considered "small" (effect size above 0.2), as described (Chavalarias, Wallach, Li, & Ioannidis, 2016;Sullivan & Feinn, 2012).

ACK N OWLED G M ENTS
We thank Philipp Kaldis, Brian Burke, Graham D. Wright, Chin Yan Lim, Thomas Boudier and Sophie Bellanger and the members of the Dreesen Laboratory (Skin Research Institute Singapore, A*STAR Singapore) for discussions. The authors acknowledge the A*STAR SBIC-Nikon Imaging Centre, Biopolis. This work was supported by the Singapore Biomedical Research Council and the Singapore Agency for Science, Technology and Research (A*STAR).

CO N FLI C T O F I NTE R E S T
Authors declare that there is no conflict of interest.

AUTH O R CO NTR I B UTI O N S
OD designed experiments and wrote the manuscript with AC. AC, PFO, MXRF, DL, LPH and OD performed experiments and analysed the data. CLS reviewed the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.