Recurrent turnover of senescent cells during regeneration of a complex structure

1Cellular senescence was previously identified as a process that permanently halts the proliferation of normal cells in culture following replicative exhaustion. It subsequently became clear that cellular senescence is a stress response that acts both in culture and in vivo to prevent proliferation of cells exposed to oncogenic stress, such as telomere attrition, and various types of DNA damage and oncogene insertions. It therefore acts as an effective anti-tumourigenic mechanism. However, senescent cells can also have detrimental effects on biological processes. Long-term accumulation of senescent cells leads to disruption of tissue structure and function, possibly through the acquisition of a senescence-associated secretory phenotype. This is particularly relevant in the context of ageing, as in most species there is a marked accumulation of senescent cells with time. Indeed, recent studies have uncovered a causative link between cellular senescence and age-related deterioration, underscoring the therapeutic benefits of targeting senescent cells, and establishing cellular senescence as a hallmark of ageing.

2In mammals, the ability to regenerate tissues declines with age. The decline in muscle regeneration with age has recently been linked to the loss of quiescent stem cells through senescence. In contrast, other vertebrates, such as zebrafish and salamanders, are able to accomplish perfect regeneration of a wide range of complex structures throughout their lifespan. It is therefore possible that these organisms exhibit mechanisms to curtail cellular senescence, thus allowing them to undergo indefinite rounds of regeneration. However, the occurrence and regulation of cellular senescence during regeneration have not been addressed in such regeneration-competent species. Here, we analysed the process of cellular senescence in salamanders. Our results demonstrate that cellular senescence is a recurrent process during salamander limb regeneration, and is subject to a highly efficient mechanism of macrophage-dependent elimination.

3To investigate the occurrence of cellular senescence in the salamander system, we first adapted and validated reagents that allow the identification of senescent cells in cell culture and tissue sections. Cellular senescence is a mechanism that irreversibly arrests the proliferation of damaged or dysfunctional cells, and is triggered upon DNA insults such as UV irradiation. We therefore exposed A1 salamander limb cells to UV, while stabilising the level of p53, which is required for the induction of senescence in certain mammalian contexts (Figure 1A). Up to 80% of the irradiated cells entered a state of cell cycle withdrawal within 12 days that, in contrast to quiescence, was accompanied by high levels of Gadd45 expression, permanent γH2AX foci, and high levels of senescence-associated β-galactosidase (SA-βgal), a hallmark of cellular senescence (Figure 1B–F). The timing of SA-βgal induction in salamander cells is similar to that reported in mammalian fibroblasts undergoing stress-induced senescence.

4Senescent cells are characterised by an irreversible cell cycle arrest. While proliferating or quiescent A1 cells re-enter the cell cycle upon serum stimulation (indicated by the incorporation of the nucleotide analogue EdU), SA-βgal positive A1 cells do not, as demonstrated by double SA-βgal/Edu staining (Figure 1G). Furthermore, the cells identified by the SA-βgal staining exhibit other markers of senescence such as sustained production of reactive oxygen species (ROS), and extended mitochondrial and lysosomal networks (Figure 1F,H) relative to non-senescent parallel cultures. In addition, these cells acquire a secretory phenotype comprising a number of cytokines, signalling and matrix remodelling factors (Figure 1I). It is nothworthy that many of these factors, such as the interleukins IL-1a, IL-1b, IL-13, IL-7, IFNγ (Interferon gamma), TNF-α (Tumor Necrosis Factor alpha), M-CSF (Macrophage Colony-Stimulating Factor), the matrix metelloproteases MMP1, 3 and 9, uPA (urokinase-type Plasminogen Activator) and Cathepsin D, have been previously reported to integrate the senescence-associated phenotype (SASP) in various contexts of mammalian senescence. Together, these data suggest that the regulation and properties of the senescent state are comparable in salamanders and mammals.

5In order to determine whether cellular senescence occurs in vivo within normal and regenerating salamander tissues, we adapted the SA-βgal staining to detect senescent cells within salamander tissues and addressed whether senescence occurs during limb regeneration in an adult salamander, the newt (Notophthalmus viridescens). We found that there is a significant induction of cellular senescence during the intermediate stages of regeneration (mid-blastema and palette), but that the number of senescent cells diminishes subsequently (Figure 2A,B). Within the regenerating limb, senescent cells are found at the level of the amputation plane, including cartilage and muscle, and within the blastema, including fibroblast, monocyte-like cells, and epidermal glands (Figure 2—figure supplement 1). To corroborate that SA-βgal+ cells exhibit other hallmarks of cellular senescence in vivo, we performed SA-βgal/EdU stainings in regenerating or normal tissues following systemic EdU injections. We were unable to detect SA-βgal+/EdU+ cells under any condition examined (Figure 2C), suggesting that SA-βgal+ cells are withdrawn from the cell cycle, a characteristic of senescent cells.

6Similar observations were made in another salamander species, the axolotl (Ambystoma mexicanum), both in mature animals as well as at early stages of development (Figure 2D,E). Interestingly, the induction and subsequent disappearance of senescent cells appear to be characteristic of regeneration, as no significant induction of senescence takes place during normal limb development (Figure 2E, Figure 2—figure supplement 2), in contrast to previous findings in amniotic limb development. These observations suggest that cellular senescence is a normal process during salamander limb regeneration, and is subject to dynamic regulation.

7Salamanders have the remarkable ability to undergo multiple rounds of regeneration through their lives. Given the proportion of senescent cells induced during each regeneration cycle (7% of total cells at the mid-blastema stage), it seemed possible that senescent cells could accumulate following multiple rounds. However, we found that the percentage of senescent cells does not increase following sequential amputation–regeneration cycles. The tissues of adult newts that had been subjected to five regeneration cycles over 1.5 years did not show any accumulation of senescent cells compared to intact limbs (Figure 2F). This data suggests that elimination of senescent cells occurs during each round of limb regeneration.

8The proportion of senescent cells within tissues is known to increase as the organism matures in most species studied so far, including mammals. Remarkably, we found that the percentage of senescent cells in heart, spleen and liver does not increase in older salamanders (Figure 2G, Figure 2—figure supplement 3); the latter two organs are known to accumulate senescent cells in mammals as they age. Together, these findings raise the possibility that active mechanisms of senescent cell clearance operate within normal and regenerating salamander tissues.

9To test this possibility, we implanted one thousand senescent or normal cells (labelled with nuclear GFP expressed from an integrated retrovirus) within newt limb tissues and analysed their persistence over time (Figure 3A). After an initial non-specific loss of cells following implantation, the normal cells (A1 GFP+) persist for at least 40 days and contribute to structures within the limb tissues (Figure 3A,B). In contrast, 80% of the implanted senescent cell population is cleared from the limb tissues within 2 weeks post-implantation (Figure 3A,B), consistent with a half-life of 9 days (calculated from the exponential equation corresponding to the senescent dataset, after subtraction of the initial loss observed in normal cells), comparable to the rate of clearance of senescent cells during regeneration (Figure 2). Although this experiment suggests that senescent cells are being cleared from salamander tissues, it cannot exclude the possibility that some cells may reverse their senescence state. To distinguish between these two possibilities, we performed the same experiment using retrovirally labelled nGFP senescent cells (Figure 3B and Figure 3—figure supplement 1). The rate of clearance of nGFP+ senescent cells corresponds to the rate of clearance of unlabelled senescent cells, supporting the clearance hypothesis. In contrast, the implantation of a population of nGFP cells exposed to a UV dose of 1J/m2, 10 times lower than the dose required for senescence induction (Figure 3—figure supplement 1), results in persistence of the implanted cells (Figure 3B). These data suggest that salamander tissues have a significant capacity for senescent cell clearance.

10Interestingly, the implantation of a 1:1 mixture of senescent and normal cells does not result in the persistence of the normal cells, but in the clearance of both populations (Figure 3B, Figure 4A,B), albeit with a delay when compared to the clearance of purely senescent cells. It has been recently shown that senescent cells are able to induce senescence in their neighbours, a phenomenon termed ‘senescence bystander effect’. We found that this property is also present in senescent salamander cells. Cellular senescence can be induced in normal cells by co-culture with senescent cells (Figure 4C), or following incubation with purified media from senescent populations (Figure 4D), supporting the idea that senescence can be transmitted in a paracrine fashion. In addition, implantation of the 1:1 mixture of senescent + normal cells results in the induction of senescence in normal cells (Figure 4E,F), providing an explanation for their eventual clearance. Together, these results suggest that active mechanisms identify and target senescent cells for clearance within salamander tissues.

11A series of elegant experiments have recently described a role for the immune system in the surveillance and clearance of p53 or oncogene-induced senescent hepatocytes in mice. In addition, recent evidence suggests that there is an extensive recruitment of immune cells, and in particular macrophages, to the regenerating limb in salamanders. Hence, we decided to investigate if the macrophage was part of the senescence clearance mechanism. In vivo labelling using TMR-dextran, which is specifically internalised by macrophages (Figure 5A–C), as well as monocyte specific α-naphthyl acetate esterase staining (Figure 5D), reveal that macrophages and senescent cells are found in close proximity within regenerating limbs, suggesting cell-to-cell contact and engulfment in some cases. To determine whether macrophages mediate the clearance of senescent cells, we took advantage of an efficient method for macrophage depletion, based on the intravenous administration of selectively toxic, clodronate salt-filled liposomes (clodrosomes). These are internalised specifically by macrophages via phagocytosis and have been shown to operate effectively in salamanders. We found that treatment with clodrosomes is able to reduce the macrophage population by fivefold, compared with control DiI liposomes (Figure 6A). As with the TMR-dextran, these liposomes are specifically internalised by macrophages, as shown by their incorporation into cells expressing F4/80, a specific macrophage marker (Figure 6B). We therefore used this system to induce systemic macrophage depletion in salamanders prior to the implantation of senescent or normal nGFP+ cells in contralateral regenerating limbs, and followed the fate of the implanted cells with time (Figure 6C). At 20 hr post implantation, we observed a strong accumulation of DiI-labelled macrophages in the vicinity of senescent cells (Figure 6D–F). In contrast, macrophage recruitment to areas of normal cells (control) was not significantly different from their recruitment to other areas of the limb mesenchyme (Figure 6D,F). At 2 weeks post implantation, the normal cells remained within the regenerating limbs (Figure 6G). However, the senescent cells were cleared in animals injected with control DiI liposomes, but remained in animals whose macrophages had been depleted (Figure 6G). This demonstrates that the macrophage is critical for clearance of senescent cells, and constitutes the first evidence for senescence surveillance mechanisms that operate during normal regeneration.

12It is clear that salamanders possess a rapid and efficient mechanism to recognise and clear senescent cells that either arise endogenously, or are introduced from culture. Our study demonstrates that a robust macrophage-dependent surveillance mechanism operates in normal and regenerating tissues of adult salamanders, and this allows them to circumvent the negative effects associated with the long-term accumulation of senescent cells, such as the disruption of tissue structure and function. These surveillance mechanisms are particularly significant for limb regeneration because there is a notable induction of cellular senescence during this process. Consistent with this, recent reports show that systemic macrophage depletion during salamander limb regeneration leads to defects in this process. This has also been observed during zebrafish fin regeneration. Furthermore, cellular senescence has been recently shown to contribute to the decline in muscle regenerative ability upon ageing in mammals. In light of these observations, we propose that the effective immune clearance of senescent cells in salamanders supports their ability to undergo multiple rounds of regeneration through their lifespan (Figure 7).

13The mechanisms underlying the induction of senescent cells and their clearance, for example the involvement of other immune cell types in their surveillance, constitute an important outstanding issue. The analysis of these problems has been limited by lack of reagents, as well as cellular and molecular tools for in vitro modelling of senescent-immune cell interactions as well as gain or loss-of-function approaches to evaluate induction and clearance of endogenous senescent cells. Further progress in this area will enable the exploration of these important matters.

14Nevertheless, it is interesting that we find a consistent induction of cellular senescence during regeneration, suggesting that this could be integral to the mechanism. Senescent cells could play a positive role during regeneration, a proposal that is consistent with recent findings that senescence contributes to tissue remodelling during mouse development and wound healing. Alternatively, senescent cells could contribute to regeneration indirectly, by promoting the efficient recruitment of macrophages, since the latter have been shown to promote tissue remodelling and patterning in various contexts, including regeneration. In either case, the senescence-associated secretory phenotype is likely to play an important role. These are important issues that merit further examination.

15It has recently been suggested that targeting senescent cells could lead to therapeutic strategies for age-related pathologies. Here, we identify an animal with an efficient mechanism for surveillance of senescent cells operating through adulthood. Analysis of this mechanism could lead to the identification of novel therapeutic targets for the amelioration of age-related disorders and extension of healthspan.

16Adult newts (N. viridescens) and axolotls (A. mexicanum) were used in this study, in compliance with the Animals (Scientific Procedures) Act 1986, approved by the United Kingdom Home Office. SA-β-gal activity was determined in cultured cells or tissues fixed with 0.5% glutaraldehyde, using the SA-βgal kit (Cell Signalling) according to the manufacturer's instructions. Immunofluorescence staining, EdU and BrdU labelling of tissues or cultured cells were performed according to standard protocols. Fluorescence and bright-field imaging was performed using a Zeiss Axioscope (Zeiss).

17Procedures for care and manipulation of all animals used in this study were performed in compliance with the Animals (Scientific Procedures) Act 1986, approved by the United Kingdom Home Office.

18Adult newts (N. viridescens) were obtained from Charles Sullivan and Co. (Tennessee, USA) and maintained as described elsewhere. Axolotls (A. mexicanum) were obtained from Neil Hardy Aquatica (Croydon, UK) and maintained in individual aquaria at approximately 18°C. Newts and axolotls were anesthetised in 0.1% tricaine prior to amputation at the mid-humerus level. Animals were allowed to regenerate at 20°C. For repetitive amputation–regeneration cycles, animals were allowed to achieve full limb regeneration prior to re-amputation. Axolotl and newt regeneration stages were defined as previously described.

19A1 cells were previously derived from newt limb mesenchyme. A1 nGFP+ cells were obtained by stable transfection with a pseudotyped retroviral VSV-G vector as described below. Cells were grown on 0.75% gelatin-coated plastic dishes in MEM (Gibco, UK) supplemented with 10% heat-inactivated foetal calf serum (FCS, Gibco), 25% H2O, 2 nM L-Glutamine (Gibco), 10 µg/ml insulin (Sigma, St Louis, MO) and 100U/ml penicillin/streptomycin (Gibco) in a humidified atmosphere of 2.5% CO2 at 25°C. Cell subculture was performed as previously. Human 293GP cells were propagated in DMEM (Gibco) supplemented with 10% heat-inactivated foetal calf serum (FCS, Gibco).

20Pseudotyped retroviral vectors were generated essentially as described. 293GP cells containing the gag and pol MoMLV genes were cotransfected by standard calcium precipitation with the retroviral vector LZRSpBMN-Z—nGFP (encoding an eGFP protein with a nuclear localisation signal between the BglII and BamHI sites) and pCMV-VSV-G (an expression plasmid for Vesicular Stomatitis Virus protein G) in order to generate pseudotyped retrovirus. Medium was collected at 48, 60 and 72 hr post transfection, filtered through a 0.45 mm filter and concentrated 100 fold by ultracentrifugation at 25,000 rpm, 1.5 hr. Stock titers were determined by infecting QT6 cells. A1 cells were grown to 30% confluence and infected with pseudotyped retrovirus stocks (1–2 × 106 PFU) in the presence of 8 µg/ml polybrene (Sigma). After overnight infection, cells were incubated in fresh culture medium with 2 mg/ml G418 until reaching confluence.

21In order to induce cellular senescence in cultures, cells were washed in amphibian PBS (A-PBS, PBS plus 25% dH2O) and exposed to 1, 3 or 10J/m2 of UV irradiation (UV Stratalinker) in 1 ml A-PBS and incubated in growth medium, supplemented with 1 µM (−)nutlin3a (Cayman, Ann Arbor, MI) in a humidified atmosphere of 2.5% CO2 at 25°C. Nutlin3a was added to the growth medium in 1 µM increments every 24 hr for the duration of the experiment. Growth media was changed after 1 week. Cells were then subcultured and senescence induction evaluated using various methods described throughout this manuscript.

22SA-β-gal activity was determined in 0.5% glutaraldehyde fixed cell culture cells using the SA-βgal kit (Cell Signalling, Danvers, MA) according to manufacturer's instructions. For detection of SA-β-gal in animal tissues, salamander embryos or adult tissues were fixed for 30 min or 1 hr respectively in 0.5% glutaraldehyde, washed 3 times in PBS. For whole mount staining, salamander embryos or adult tissues (heart, spleen, liver) were stained O/N using the SA-β-gal kit (Cell Signalling) according to manufacturer's instructions, washed in PBS, fixed in 4%PFA for 4 hr and embedded in Tissue Tek-II. Whole-mount samples were then cryosectioned, washed in PBS and mounted in glycerol for imaging.

23Adult intact or regenerating limbs were cryosectioned in Tissue Tek-II, washed in PBS for 15 min and stained O/N using the SA-β-gal kit (Cell Signalling) according to manufacturer's instructions.

24Culture medium was cleared by centrifugation and concentrated using Ultracel-3K Amicon Ultra centrifugal tubes (Merck Millipore Ltd, Germany). The corresponding cell extracts were prepared by washing cells in PBS and resuspending them in ice-cold 1% NP40, 150 mM NaCl, 50 mM Tris–HCl, pH 7.5, 1 mM NaVO4 and Protease Inhibitor Cocktail (Sigma), incubating for 30 min at 4°C and clearing the debris by centrifugation. Concentrated media and their corresponding cell extracts were analysed using the Proteome Profiler Human Protease and Cytokine Array Kits (R&D systems, Minneapolis, MN) as per manufacturer's instructions. Antibody arrays were developed in an Odyssey scanner (LI-COR, Lincoln, NE), and protein levels quantified using Image Studio software. One Matrix CIM heat maps were built using CIMminer (NCI-NIH, USA).

25For the analysis of DiI liposome incorporation and NAE or NacdE staining of salamander limbs following cell implantation, regenerating or intact limbs were collected by amputation, fixed in 4% ice-cold paraformaldehyde (PFA) for 16–18 hr at 4°C, washed twice in PBS and embedded in Tissue Tek-II. The blocks were serially sectioned longitudinally in a cryostat (Leica, UK) at 12 µm. Sections were collected in Superfrost slides and stored at −30°C until use.

26To determine the percentage of cells in S-phase, proliferating, quiescent or senescent A1 cells were sub-cultured and incubated in growth medium supplemented with 5 µM 5-ethynyl-2′-deoxyuridine (EdU) for 24 hr. Cells were then fixed in glutaraldehyde and SA-β-gal staining performed as described. Next, cells were fixed in 4%PFA for 10 min and stained using Click-iT Edu Alexa Fluor 594 Imaging kit (Life Technologies, UK) according to manufacturer's instructions.

27To detect EdU incorporation in salamander tissues, 10 mM EdU (20 µl per animal) were administered to regenerating newts by i.p. injection. After 48 hr, blastemas were collected, cryosectioned and stained to determine SA-β-gal activity as described. Sections were fixed in 4% PFA for 10 min and EdU incorporation determined using Click-iT Edu Alexa Fluor 594 Imaging kit (Life Technologies).

28Cellular senescence was induced in cultured A1 cells by UV irradiation as described above. Following verification by O/N SA-β-gal staining, senescent cells were trypsinised, spun at 850 rpm for 10 min and the resulting pellet resuspended in A-PBS. Cells were then transferred to a 10 µl Hamilton syringe (Hamilton, Reno, NV) with a 301/2 g, 45° tip needle (Hamilton) attached to a micromanipulator. 1000 cells were injected within the mesenchymal tissues of intact or regenerating (midbud stage) limbs of newts or axolotls. The animals were allowed to recover on a wet tissue for 10 min before transfer to water. Successful implantation was verified by fluorescence stereomicroscopy using a Zeiss Axioscope (Zeiss, Germany).

2917 cm axolotls (3 year old animals at a reproductive stage) were anaesthetised in 0.1% Tricaine for 20 min, prior to i.v. microinjection of 20 µl DiI or clodronate liposome solutions (Encapsula Nano Sciences, Brentwood, TN) as previously described. For long-term depletion experiments, animals were injected every other day 3 times, with a resting period of 3 days before the next injection rounds. The same procedure was applied for macrophage labelling using 20 µl 2MDa tetra-methylrhodamine (TMR) labelled dextran (Invitrogen, UK). All injections were performed 36–48 hr before cell implantation or analysis. Successful compound uptake by macrophages was verified by fluorescence stereomicroscopy using a Zeiss Axioscope (Zeiss).

30Granulocytes -including neutrophils- were detected by specific staining of tissue cryosections with the naphtol AS-D chloroacetate esterase kit (Sigma–Aldrich), while monocytes/macrophages were identified using the α-naphtyl acetate esterase (NAE) kit (Sigma–Aldrich), as previously described18.

31For staining of cultured cells, these were fixed in 2% PFA for 1 min, followed by a 5-min incubation in cold 100% methanol and processed as described elsewhere. Fixed cell samples were incubated overnight with the following primary antibodies: anti-BrdU (Sigma; 1:3000), anti-γH2AX (Upstate, UK; 1:1000), anti-laminB1 (Abcam, UK; 1:500), and anti-F4/80 (ABD Serotec, UK; 1:500).

32In all cases, anti-mouse or anti-rabbit AlexaFluor488 and AlexFluor594 antibodies (Invitrogen; 1:1000) were used for secondary staining. Hoechst 33258 (2 µg/ml) was used for nuclei counterstaining.

33For detection of ROS production, cells were incubated with 10 µM DHR123 (Life Technologies, Invitrogen) for 2 hr at 25°C, fixed and mounted in DAKO fluorescent mounting media.

34Mitochondrial and Lysosomal networks were visualised following incubation of cells with 200 nM MitoTracker red FM and 500 nM LysoTracker red DD99 (Life Technologies, Invitrogen) respectively for 1 hr at 25°C.

35Samples were observed under a Zeiss Axiskop2 microscope and images were acquired with a Hamamatsu Orca camera using Openlab (Improvision) software. Whenever comparative analyses between samples were performed, all images were acquired with identical camera settings and illumination control. Image processing (contrast enhancement) was equally applied to all matched experimental and control samples using Openlab software.

36Cells were labelled for 2 hr by adding 1 µl/ml 5-bromo-2′-deoxyuridine (BrdU) to the growth media. Following the corresponding incubation period, cells were fixed in 4% paraformaldehyde for 1 min at RT followed by 100% methanol for 5 min, and stained for bromodeoxyuridine as previously described.

37RNA was isolated from salamander tissue culture cells using Tri Reagent (Sigma) and random primed cDNA synthesised using Superscript II (Invitrogen). Gene expression was determined by quantitative real time PCR using the following primers: Newt Gadd45-β fwd (AGGGCACAGGAAAGAAGATG); Newt Gadd45-β rev (TCATTGTCGCAGCAGAAGG); Newt L-27 fwd (TACAACCACTTGATGCCA); Newt L-27 rev (CAGTCTTGTATCGTTCCTCA). Rt-PCR was carried out using iQ SYBR Green supermix (Bio-rad, Hercules, CA) on a Chromo 4 instrument running Opticon 3 software (Bio-rad). All reactions were run in triplicate and at least 2 independent RNA preparations were analysed for each sample.

38Newts and axolotls in each sample group were randomly selected. Sample group size (n) is indicated in each figure legend, while all experiments were carried out in at least three biological replicates. Statistical analyses were performed with Prism 4.0 software and unpaired two-tailed t tests were applied unless otherwise stated. Paired two-tailed t tests were carried out to analyse RT-PCR experiments.