Transient Senescence Induces Wound Healing Through SASP Factors: Therapeutic Potentials of Chick Early Amniotic Fluid (ceAF)

family, vascular endothelial growth factor (VEGF), granulocyte macrophage colony stimulating factor (GM-CSF), platelet-derived growth factor (PDGF), connective tissue growth factor (CTGF), interleu-kin (IL) family, and tumor nerosis factor-a family. Currently, patients are treated by three growth factors: PDGF-BB, bFGF,andGM-CSF.OnlyPDGF-BBhassuc-cessfully completed randomized clinical trials in the Unites States. With gene ther-apy now in clinical trial and the discovery of biodegradable polymers, ﬁ brin mesh, and human collagen serving as potential del ivery sys tems other growth factors may soon be available to patients. This review will focus on the speci ﬁ c roles of these growth factors and cytokines during the wound healing procesSenescence


Abstract Background
In ammatory, proliferative and remodelling phases constitute a cutaneous wound healing program. Therapeutic applications/medication are available; however, they commonly comprise forti ed preservatives that might prolong the healing process. Chick early amniotic uids (ceAF) contain native therapeutic factors with balanced chemokines, cytokines and growth-related factors; their origins in principle dictate no existence of harmful agents that would otherwise hamper embryo development.
Instead, they possess a spectrum of molecules driving expeditious mitotic divisions and possibly exerting other functions.

Methods
Employing both in vitro and in vivo models, we examined ceAF's therapeutic potentials in wound healing and found intriguing involvement of transient senescence, known to be intimately intermingled with Senescence Associated Secretory Phenotypes (SASP) that function in addition to or in conjunction with ceAF to facilitate wound healing.

Results
In our in vitro and in vivo cutaneous wound healing models, a low dose of ceAF exhibited the best e cacies; however, higher doses attenuated the wound healing, presumably by inducing p16 expression.

Conclusion
Our studies link an INK4/ARF locus-mediated signalling to cutaneous wound healing, implicate the therapeutic potentials of ceAF exerting functions likely by driving transient senescence and expediting cell proliferation, and conceptualize a homeostatic and/or balanced dosage strategy in medical intervention.

Background
The Graphical Abstract shows the INK4-ARF locus comprising tumour suppressor genes ARF/INK4A(p16)/INK4B that encode proteins with anti-proliferative functions exerted via RB and p53 (1,2). Respectively, RB regulates cell cycle downstream of the signalling with p16 and CDK4/CDK6, and ARF transduces a signal via MDM2, p53 and p21. Stress responses by which proliferative cells lose dividing potentials in an (almost) irreversible fashion are dubbed cellular senescence (2) exerted by diverse intraand extra-cellular stimuli through activating the above pathways in a combinatorial manner (3)(4)(5) Generally, cellular senescence occurs in aging and tumour suppression; disparate biological processes as they seem, there could be an evolutionary logic for such antagonist pleiotropy. Senescent cells can mobilize a cohort of in ammatory cytokines, proteases, chemokines and growth factors as known as senescence-associated secretory phenotypes (SASP) [3], which can be induced upon activation of the INK4-ARF locus. This locus is epigenetically silenced in foetal, embryo and adult stem cells; however, in differentiated cells, it can be reprogramed to become hyper-responsive to mitogenic signals (6). Most senescent cells express p16 (7) that blocks cell cycle by inhibiting CDK4 and CDK6 (8) thus a commonly recognized marker (9). Pathologies aside, senescence also occurs in embryogenesis presumed to play roles in regulating embryonic structures (10,11).
In many species, embryonic amniotic uids are a dynamic milieu that participates in, e.g., cushioning, hydrating and providing immunity to embryos. Recently, we isolated embryonic stem cells in chick early amniotic uids (ceAF) of developing eggs at days 6-8 when embryos expeditiously, if not exponentially, grow, and found that ceAF comprises proliferation-stimulants that at a low-level support robust proliferation of in vitro cultured cells, almost as potently as foetal bovine serum (FBS; results). This characteristic and other biological activities of ceAF beyond conventional growth factors prompted us to establish in vitro and in vivo cutaneous wound healing models to explore its therapeutic potentials and underlying mechanism(s).
Wound healing is an evolutionarily conserved, complex, multicellular process that, in skin, aims at barrier restoration. This process involves the coordinated efforts of several cell types including keratinocytes, fibroblasts, endothelial cells, macrophages, and platelets. The migration, infiltration, proliferation, and differ-entiation of these cells will culminate in an inflammatory response, the formation of new tissue and ultimately wound closure. This complex process is executed and regulated by an equally complex signaling network involving numerous growth factors, cytokines and chemokines. Of particular importance is the epidermal growth factor (EGF) family, transforming growth factor beta (TGF-b) family, fibroblast growth factor (FGF) family, vascular endothelial growth factor (VEGF), granulocyte macrophage colony stimulating factor (GM-CSF), platelet-derived growth factor (PDGF), connective tissue growth factor (CTGF), interleu-kin (IL) family, and tumor nerosis factor-a family. Currently, patients are treated by three growth factors: PDGF-BB, bFGF,andGM-CSF.OnlyPDGF-BBhassuc-cessfully completed randomized clinical trials in the Unites States. With gene ther-apy now in clinical trial and the discovery of biodegradable polymers, fibrin mesh, and human collagen serving as potential del ivery sys tems other growth factors may soon be available to patients. This review will focus on the specific roles of these growth factors and cytokines during the wound healing process Wound healing is an evolutionarily conserved, complex, multicellular process that, in skin, aims at barrier restoration. This process involves the coordinated efforts of several cell types including keratinocytes, fibroblasts, endothelial cells, macrophages, and platelets. The migration, infiltration, proliferation, and differ-entiation of these cells will culminate in an inflammatory response, the formation of new tissue and ultimately wound closure. This complex process is executed and regulated by an equally complex signaling network involving numerous growth factors, cytokines and chemokines. Of particular importance is the epidermal growth factor (EGF) family, transforming growth factor beta (TGF-b) family, fibroblast growth factor (FGF) family, vascular endothelial growth factor (VEGF), granulocyte macrophage colony stimulating factor (GM-CSF), platelet-derived growth factor (PDGF), connective tissue growth factor (CTGF), interleu-kin (IL) family, and tumor nerosis factor-a family. Currently, patients are treated by three growth factors: PDGF-BB, bFGF,andGM-CSF.OnlyPDGF-BBhassuc-cessfully completed randomized clinical trials in the Unites States. With gene ther-apy now in clinical trial and the discovery of biodegradable polymers, fibrin mesh, and human collagen serving as potential del ivery sys tems other growth factors may soon be available to patients. This review will focus on the specific roles of these growth factors and cytokines during the wound healing procesSenescence Major ndings and signi cance We found that senescent broblasts and endothelial cells were enriched at wound areas a few days post skin injury and were more recruited by daubing ceAF to the wound areas, which at a low dose facilitated wound healing. The healing was likely mediated by SASP factors, e.g., PDGF-B, TGF-β and VEGF known to promote myo broblast differentiation and epithelial growth with much-induced sebaceous glands and cellular layers. The ceAF components may function by signalling SASP factors as a marker for transient senescence that in turn promotes faster cell migration and scar-less restoration by, in part, expediting cell proliferation. p16, reported to be able to enhance SASP factors release, was found to attenuate wound healing presumably due to an above-a-threshold expression level. Our studies also provide an inkling that p21, rather than p16, plays roles in orchestrating a cutaneous wound healing program.
Results And Discussion ceAF supports cell growth and promotes wound healing in vitro Given that ceAF is a source of embryonic stem cells, we believed that ceAF comprised biologically balanced secretory growth factors, cytokines and other ligands collectively dubbed proliferationstimulants and, as a prelude to exploring its therapeutic potential, tested its ability to support the growth of cultured HeLa cells using FBS as a control. Starting from ~20% con uence, HeLa cells were cultured in DMEM supplemented with increments of ceAF (0-20%, v/v) or 10% FBS for 48 hours; 5% ceAF optimally, and nearly as potently as 10% FBS, supported cell growth that roughly underwent 2 doubling times ( Figure 1A/1B). Given that cultured HeLa cells exhibited elevated p16 expression in higher doses of ceAF (20% in particular; Figure 1B bottom panel), we proposed that the lower e cacies of higher ceAF doses, as compared to 5% ceAF, for supporting cell growth ( Figure 1B upper panel) were attributable to an antiproliferative activity of p16. This suggests a manifestation of attenuating activities with higher doses of ceAF and emphasizes an importance of balanced (or optimal) dose of ceAF in its application on wound healing.
Prompted by the above, we established an in vitro wound healing model utilizing HaCaT cells, representing an epidermis origin (12); given fastidious nature of ceAF, we employed 5% and 10% doses in our later efforts on the presumption that one or a dose somewhere in between works optimally. Scratches were made on near-con uent monolayers and cells treated with ceAF. Agreeing with cell proliferation assays, 5% ceAF healed the scratches better than 10% ceAF within 24 hours; ectopically expressing p16 hindered the healing in vitro (Figure 1C/1D; also see below).
Encouraged by the above in vitro results, we examined the healing capacities of ceAF on surgically-made cutaneous wounds in a murine model. Mice treated with 5% ceAF healed most expeditiously, manifesting virtually scar-less skin in 10 days (Figure 2A/2B). Dermal application of naturally produced/topically administered exogenous p16 has been a practice in wound healing research, given that this small-size protein can be up-taken by G-protein coupled receptors (13,14). In our hands, p16 when co-applied within a dosage employed by other practices signi cantly attenuated the wound healing process, which served an offset for facilitated wound healing brought about by ceAF particularly at a 5% dose (

p16 impedes wound healing in vitro and in vivo
Under stresses such as those that induce senescence, p16 expression is up-regulated; p16 was reported to be able to drive senescence-associated secretory phenotypes (SASP) that provoke chemokine/cytokine pathways in the in ammatory phase of wound healing preluding the proliferative phase. SASP factors exert functions transiently, but whether this "transient senescence" is p16-dependent remains unresolved.
To examine whether p16 expression in conjunction with ceAF facilitates wound healing, we ectopically expressed Flag-tagged p16 in HaCaT cells supplemented with either 5% or 10% ceAF; p16 unexpectedly and markedly retarded the closure of scratched areas ( Figure 1C/1D) and cellular proliferation (see below). We had serially titrated down p16-expressing plasmid in transfection assays, and were unable to identify a dose in which ectopic p16 expression would additively or synergistically facilitate wound healing by ceAF (data not shown). In the animal models, similar patterns were observed whereby wound healing was attenuated when p16 was co-daubed with ceAF ( Figure 2A/2B). Taken together, these results suggest that, at least in our assay systems, the signalling involving SASP factors and transient senescence might not be strictly dependent on p16. Other pathway(s) such as, notably, a p21-mediated pathway may compensate for and/or play redundant roles in this program (see below). ceAF facilitates wound healing in vivo and manifests increased collagen density and thickened epidermis on healed wounds On day 10, we observed prominent contraction on the ceAF-daubed wounds particularly with a 5% dosage, and signi cant re-epithelialization ( Figure 2A) [signi cance of 5% ceAF as compared to other groups summarized in Table S1(A)]. However, mice treated with ceAF+p16 exhibited wounds with wide epithelial and dermal edges with the gap comprising necrotic brinoid, cell debris, intense in ammation and angiogenesis plus reduced granulation tissues and signi cant lower deposition of collagen as judged by the trichrome staining ( Figure 2B), suggesting a wound healing that was hampered by p16.
Furthermore, histological (H&E) staining showed that, in the 5% ceAF group, wound sections showed a neoformative epidermal layer that was remarkably thicker ( Figure 2C), indicating dense epidermal ridges, sebaceous glands, cellular layers and more formation of primitive hair follicle structures. On the contrary, the ceAF+p16 group wounds exhibited much lower migration of endothelial cells and broblasts as well as signi cantly reduced epidermal thickness of the skin with impaired granulation tissues ( Figure 2C), again indicating a wound healing hampered by p16.
ceAF facilitates an S to G2/M transition and cell migration that is offset by p16 To test whether the healing process facilitated by ceAF and attenuated by p16 expression or administration (Figures 1/2) were related to cell cycle progression, we cultured HaCaT cells with ceAF (5% or 10%) with or without ectopic p16 expression and subjected cells to FACS analyses. Cells in the 5% ceAF group exhibited highest percentile of G2/M cells, but p16 expression on top of that retarded quite some cells in the S-phase that had obvious di culty to enter the G2 phase ( Figure 3A), suggesting that the wound-healing-promoting (by ceAF) and the wound-healing-attenuating (by p16 expression/administration) phenotypes (Figures 1/2) are at least in part due to their functions exerted on cell cycle progression.
Cell migration also plays important roles in wound healing (15)(16)(17) ; we performed trans-well migration assays using HaCaT cells. The 5% ceAF-treated group exhibited a high number of trans-well-passed cells; likewise, and consistent with the cell cycle results ( Figure 3A), expressing p16 on top of ceAF signi cantly attenuated trans-well passaging ( Figure 3B).

ceAF triggers transient senescence
Senescence is generally recognized to be a chronic mechanism associated with, e.g., age-related pathologies; but tissue damages such as skin wound can induce, albeit transiently, senescence dubbed transient senescence [17]. This might be an evolutionarily acquired mechanism; immune surveillance later clears the transiently senescent cells, which manifest certain levels of Senescence Activated βgalactosidase (SA-β-gal) activity that can be detected histochemically, scoring both numbers of positive cells and intensities. Thus, SA-β-gal is a good biomarker for senescent cells in both cell culture and in vivo.
We examined SA-β-gal expression patterns in HaCaT cells incubated in control vs. ceAF-supplemented media with or without a p16 ectopic expression ( Figure 4A). The number of SA-β-gal + cells was highest in the 5% ceAF group ( Figure 4B; upper panel), but more intense SA-β-gal + signals were observed in cells ectopically expressing p16 ( Figure 4B; bottom panel); both criteria were deemed statistically signi cant.
Given that 5% ceAF exerted optimal functions on wound healing and expediting cell division (Figures 1-3), we propose that a proper (homeostatic) SA-β-gal expression level indicates a transient senescence that facilitates wound healing, while a p16-intensi ed SA-β-gal activity may represent senescence at a more permanent scale that is above a threshold and, based on our earlier data, hampers wound healing and blocks cell cycle.

Expression pro ling of factors involved in cutaneous wound healing
In ammation, proliferation and remodelling are three sequential yet overlapped phases of wound healing; upon wounding and thereafter, diverse chemokines, cytokines and growth factors are locally enriched and distantly attracted at wound sites, a signalling cascade triggered by SASP factors transiently released by transient senescent cells. They in turn sequentially or concomitantly orchestrate a healing program. For instance, involving diverse kinds of cells (e.g., epidermal and endothelial cells, broblasts, keratinocytes, neutrophils etc) and acting paracrinally, these factors comprise a host of chemokines that recruit proin ammatory cells including macrophages, which in turn release (additional) cytokines/growth factors to facilitate wound healing (18,19).
We employed wounded/repaired skin tissues from animals (in vivo samples) and HaCaT cells (in vitro samples) treated under speci ed conditions and built an expression pro le of cellular factors as aforeemphasized to be involved in tissue repair. In vivo samples (collected at day 10) were used to examine the expression pattern of genes encoding proteins that dictate the integrity of restored skin boundary, i.e., degree of remodelling. As seen ( Figure 4C, quanti cation; Figure 4D, Heatmap analyses), mRNA expression levels of three tested genes, i.e., that of keratin, keratin-10 and endothelin, were signi cantly up-regulated in the 5% ceAF group (to a lesser degree in the 10% ceAF group); p16 played an offset role (the control group here was the wound daubed with PBS/40% glycerol). In conjunction with the data shown below, we propose that ceAF plays multi-complex roles to facilitate wound healing.
We also established an expression pro le of 12 of ~25 SASP factors reported (20,21) using HaCaT cells incubated with FBS vs. ceAF, with or without ectopic p16 expression. Of the tested SASP factors, the expression of all, save p16, were up-regulated signi cantly in cells incubated in 5% ceAF and to a lesser degree 10% ceAF ( Figure 4C, quanti cations normalized against a house-keeping gene; Figure 4D, Heatmap analyses); matter-of-factly, p16 played offset roles on ceAF-enhanced expression of tested genes, and the ostensible variations of p16 expression per se ( Figure 4C; last panel) were most-likely due to the presence/absence of an ectopic p16 gene.

Physiological interpretation of a ceAF-mediated expression pro le
Structural proteins: For normal skin physiology, the epithelial keratinization is of fundamental importance for counteracting mechanical stress and protecting against pathogenic invasion, and endothelin was suggested to be involved in anabolism of collagen and related proteins (22,23). In addition, endothelin is a mediator for morphogenesis in other systems including bone regeneration and skeleton formation (24). Thus, elevated expression levels of keratin, keratin-10 and endothelin in the cutaneous wound healing model re ect a more completed remodelling phase.
Interleukins, chemokines and growth factors: Interleukins serve key modulators for in ammatory responses including that in skin repair, and is involved in activation, differentiation and proliferation of endothelial cells, broblasts, keratinocytes and leukocytes (25). For instance, IL-6 is released as an early response to tissue injury to induce a signalling of pro-in ammatory cytokines from the resident macrophages and stromal cells (26). In addition, IL-1A is an autocrine regulator for basal keratinocytes proliferation inside the bulge region of hair follicles and an integral agent of the epidermal stem cell population (27); chemokines CCL5/CCL2 can function as attractants for local and distant macrophages to be enriched at wound sites. Furthermore, growth factors such as TGF-b, VEGF, PDGFA/B can regulate direct (cell-cell) and indirect (paracrine) linkage between different cell types and cellular basement and, more relevant to this work, are involved in restoring skin boundary integrity. Therefore, these factors conceivably function at different healing stages covering the in ammatory and proliferative phases and possibly even the remodelling phase. Emphasizing important roles on wound healing, the expression of these factors is quite prominently up-regulated in ceAF-treated cells and offset by p16 ( Figure 4C-D).
PAI-1 and vimentin: Plasminogen activator inhibitor type 1 (PAI-1) is expressed on the surface of keratinocytes, its activation enhances cellular proliferation (18). PAI-1 limits plasmin generation to sustain cell migration and proliferation, a key indicator for re-epithelialization of keratinocytes (28). The intermediate lament protein vimentin plays signi cant role for the "epithelial to mesenchymal transition (EMT)", and studies reveal that vimentin participates in a number of cellular processes including migration and invasion of cells, cell adhesion, cytoskeletal rearrangements, signalling, plasticity and regulation of cell morphology (29,30). These proteins in principle are largely involved in the remodelling phase of a wound healing program.
Cell cycle/senescence regulators: p16 and p21 are well known proteins in regulating cell cycle and senescence (31,32); as far as our assay systems are concerned, p16 is, as opposed to a recognition supported by certain experimentation, not playing a role in establishing the transient senescence that readies wounds to undergo healing. From all criteria, p16 is actually an offsetting factor for therapeutic values of ceAF. Our data ( Figure 4C; second last panel) support a notion that p21 might as well play compensatory role(s) by itself or in conjunction with other factor(s), e.g., a non-redundant function afforded by the third component of the INK4-ARF locus, INK4B.

Further Discussion and Perspectives
Intrinsic to and secreted by developing chick embryos 6-8 days post fertilization, ceAF exhibits powerful wound healing capacities apparent in mammalian cutaneous wound healing models, in which ceAF with no preservatives/forti cation displays potent e cacy both in vitro and in vivo. Naturally, ceAF provides nourishment and protection against environmental stresses including pathogens, and is thought to possess diverse signalling molecules at biologically homeostatic levels to support embryogenesis. This view is especially important given a manifestation of the attenuating activities at high doses ( Figure 1B). Of note, given that the control for establishing the expression pro ling with culture cells (Figure 4C/D) was FBS, ceAF must contain components above and beyond ordinary growth factors and/or other molecules that solely support cell growth and migration as revealed in Figure 1A/B and Figure 3B. Recently, occurrence of senescence and roles of SASP factors during embryogenesis (33) and in limiting brosis upon tissue injury (34,35) were reported. Thus, scar-less tissue repair might take place in embryos, supporting our postulation that ceAF comprises a set of evolutionarily conserved, at least among vertebrates, factors involved in wound healing.
Conceivably, at least in our assay systems and as in many signalling cascades, signals from ceAF, however chick-embryo-derived, obviously can be received and ampli ed by cells in the wound/surrounding areas of a murine model to facilitate the wound healing. From a biochemical, mechanistic and therapeutic perspective, chromatographic isolation of active fractions as well as reconstitution to formulate an e cient cocktail starting from ceAF and, probably much more importantly, efforts to rid of attenuating activities ( Figure 1B) also by biochemical isolation, are a daunting challenge yet obviously a worthy goal.

Conclusion
Our studies link an INK4/ARF locus-mediated signalling to cutaneous wound healing, implicate the therapeutic potentials of ceAF exerting functions likely by driving transient senescence and expediting cell proliferation, and conceptualize a homeostatic and/or balanced dosage strategy in medical intervention.

Materials And Methods
Construction of Flag-1X-p16 and production of Flag-tagged p16 PCR-ampli ed p16 cDNA was cloned into the pXJ40 vector between BamHI/XhoI sites downstream of and in-frame with a Flag-tag epitope to create the Flag-1X-p16 construct; p16 primers are listed in Table  2S(B). HeLa cells were transfected with this construct in a large scale, cell lysates made, Flag-tagged p16 immuno-enriched on M2-agarose beads, eluted with Flag peptide and quanti ed using BSA as a reference.
Chick early amniotic uid (ceAF) preparation Fertilized chick eggs were incubated at 38± 1°C and a 50% humidity, and ceAF was collected from eggs between days 6-8. After centrifugation of the samples at 2500g for 20 min, supernatants were ltered over a 0.22 µm sterilization device (Millipore China) and stored in aliquots at -80 o C after quick-freezing in liquid nitrogen.

Animals
Speci c pathogen-free (SPF) C57BL/6 mice (all 5-week old males weighted at 20-26 g) were purchased from the SIPPR-BK Lab Animals Co. Ltd. Shanghai, China [Certi cate # SCK (hu) 2013-0016]. Grouped randomly (3 mice/cage), they were kept in the Animal Facility, Zhejiang University (Permit #: ZJU20170013) with a 12-hr day/light cycle and access to water and rodent chow. Animal procedures were approved by the institutional animal care use committee.
Step-wisely, procedures were conducted aseptically as: a full-thickness wound (10x10mm) was excised from the dorsum of mice after hair removal; excessive panniculus carnosus layer's contraction was controlled by silicon disks cut/adjusted to wound diameters; the disks were xed by interrupted sutures with no dressing applied. Cocktails, in triplicates, with 0% (control), 5% or 10% ceAF and +/-p16, were made in PBS/40% glycerol and sterilized by ltration. Each surgical wound was evenly/gently daubed with 250 µl of a cocktail; 40% glycerol was used to stabilize proteins in, and enhance the viscosity of, the cocktails for easier topical daubing, which sustains a longer-lasting moist surface after daubing. Flag-tagged p16 or BSA at 1ng/µl was employed when appropriate. Reduction in wound diameters were monitored daily until day 10 when animals were euthanized for histological and other analyses on excised healed/non-healed areas of wounds.

Histological procedures
Excised tissue samples were xed in 10% formalin for minimally 5 days and processed. Brie y, embedded in para n, samples were sectioned in 3 μm thickness, mounted on glass slides and, after depara nizing, stained with haematoxylin-Eosin and Masson's trichrome. The density of in ammatory cells and blood vessels density in dermis were analysed using M-42 system (37). Random elds of the tissue were observed and counted for particular sections with images taken by an optic microscope (Olympus, BX41).

Cell lines and other reagents
HeLa and HaCaT cells, a gift from YU Faxing (Fudan University), were experimentally con rmed free of mycoplasma and assured to have normal/expected morphologies. Cells were cultured in DMEM (Basal cDNA preparation and qPCR cDNA was prepared from isolated total RNA using a cDNA preparation kit. RNase free double distilled water, 4x gDNA wiper mix and 1 µg total RNA were sequentially pipetted into a tube and incubated at 42°C for 2 minutes. Then 5X Hi Script Mix II was added and the samples kept at 50°C for 15 min and 85°C for 5 sec before proceeding to qPCR. Information of PCR primers is in Table 2S(B).

Cell scratch assay
HaCaT cells were cultured in 6-well plates to attain a ~80% con uence, and starved for 24 hours. Cells were grouped in triplicates for transfection: DMEM with no ceAF, i.e., control, or 5%/10% ceAF, and with or without (+/-) the p16 plasmid. In the p16 minus samples was added the vector plasmid. After 24 hours, scratches were drawn with a 10 ul pipette-tip. The readings/cell migration rates at the "wounded" sites were recorded at 0, 24 and 48 hours.

Western blotting
Aspirated of media, cells were rinsed with PBS twice and lysed with HEPES lysis buffer (115 mM NaCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , 2.4 mM K 2 HPO 4 , 20 mM Hepes-KOH, pH7.0, 1% NP40) supplemented with protease and phosphatase inhibitors. Proteins levels were normalized for SDS-PAGE loading, and protein expression further normalized with an internal control (β-actin). Signals were detected by primary antibodies followed by HRP-conjugated secondary antibodies.

SA-β-Gal staining
Solutions were provided by the kit; 48 hours after treatment, cells were rinsed with PBS twice and 1 ml xation solution was added to x the cells for 15 min at room temperature. Cells were rinsed again with PBS and 1 ml β-gal solution added to develop the enzyme activity. The plates were sealed with para n to prevent evaporation and kept in an incubator at 37°C without CO 2 until colour manifestation.
Trans-well migration assay

Statistical analyses
Numerical data were expressed as mean ± SEM and statistical analysis was performed using Prism GraphPad (California). One-way ANOVA was used to reveal signi cance. *p<0.05, **p<0.01 and ***p<0.001 were considered statistically signi cant. All in vitro experiments were independently repeated in triplicates.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. CellDivisionAdditionalSupplementaryFigures.docx