Chronic expression of p16ink4a in the epidermis induces wnt-mediated hyperplasia and promotes tumor initiation

ABSTRACT p16INK4a (CDKN2A) is a central tumor suppressor, which induces cell-cycle arrest and senescence. Cells expressing p16INK4a accumulate in aging tissues and appear in premalignant

lesions, yet their physiologic effects are poorly understood. We found that prolonged expression of transgenic p16INK4a in the mouse epidermis induces hyperplasia and dysplasia, involving

high proliferation rates of keratinocytes not expressing the transgene. Continuous p16INK4a expression increases the number of epidermal papillomas formed after carcinogen treatment.

Wnt-pathway ligands and targets are activated upon prolonged p16INK4a expression, and Wnt inhibition suppresses p16INK4a-induced hyperplasia. Senolytic treatment reduces p16INK4a-expressing

cell numbers, and inhibits Wnt activation and hyperplasia. In human actinic keratosis, a precursor of squamous cell carcinoma, p16INK4a-expressing cells are found adjacent to dividing cells,

consistent with paracrine interaction. These findings reveal that chronic p16INK4a expression is sufficient to induce hyperplasia through Wnt-mediated paracrine stimulation, and suggest

that this tumor suppressor can promote early premalignant epidermal lesion formation. SIMILAR CONTENT BEING VIEWED BY OTHERS ALTERING MYC PHOSPHORYLATION IN THE EPIDERMIS INCREASES THE STEM

CELL POPULATION AND CONTRIBUTES TO THE DEVELOPMENT, PROGRESSION, AND METASTASIS OF SQUAMOUS CELL CARCINOMA Article Open access 07 September 2020 MIR-184 REPRESSES Β-CATENIN AND BEHAVES AS A

SKIN TUMOR SUPPRESSOR Article Open access 26 February 2024 ACTL6A SUPPRESSES P21CIP1 EXPRESSION TO ENHANCE THE EPIDERMAL SQUAMOUS CELL CARCINOMA PHENOTYPE Article 02 July 2020 INTRODUCTION

Activity of tumor-suppressive cellular mechanisms is often associated with cell and organismal aging, yet cancer propensity substantially increases with age1,2,3. p16INK4a, encoded by the

_CDKN2A_ gene (p16 hereafter), represents an important link between cancer, cellular responses to stress, and aging. p16 is a central tumor suppressor, which is among the most commonly

mutated genes in diverse human malignancies4,5. When activated, p16 binds and inhibits CDK4/6-Cyclin D complexes, leading to Rb activation, and thereby induces cell-cycle arrest and

senescence4,6. This pathway represents one of the central mechanisms blocking the proliferation of damaged or oncogene-expressing cells. Whereas p16 is not expressed in most embryonic and

adult cells7, its levels increase in multiple tissues with age8,9,10,11. The specific stimuli underlying age-associated p16 activation have not been directly established. However, a variety

of stresses, including radiation, DNA damaging agents, cigarette smoke, and oncogene activity, were shown to induce p1612,13,14,15. Aged animals lacking p16 show increased replicative and

regenerative capacity in several tissues, indicating that it contributes to the aging-associated decline in these processes1. It was more recently shown that directed genetic elimination of

p16-expressing senescent cells during mouse aging delays the functional deterioration of multiple organs and increases lifespan11. This finding and subsequent studies have highlighted the

negative contribution of senescent cells to age-associated pathologies, and the therapeutic potential for their pharmacologic removal through senolytic drug treatment16,17. Whether senolytic

treatments have potential benefit in cancer therapy is currently largely unknown. The expression of p16 in aging tissues raises the question of whether its activity influences cancer

development. Mice carrying an extra copy of _Cdkn2a_ show increased resistance to cancer, consistent with the known tumor-suppressive role of p1618. In contrast, elimination of

p16-expressing senescent cells reduces cancer mortality rates in mice, suggesting that such cells could contribute to tumor development11. The mechanisms underlying this are not fully known.

It has been suggested that resident senescent cells can promote tumorigenesis during aging by generating inflammation mediated by cytokine secretion, a feature of senescence known as the

senescence-associated secretory phenotype (SASP)3,19. It is, however, unclear whether all cells expressing p16 in vivo achieve a full senescence phenotype, and p16 activity itself appears to

be insufficient to induce the SASP20,21. The functional contributions of p16 to age-associated changes in cancer propensity, therefore, remain poorly characterized. Here we study the

effects of prolonged p16 expression in the epidermis, in order to uncover its effects on tissue structure and cancer development. p16 levels and senescence were reported to increase with age

in the skin dermis and epidermis22,23,24. UV radiation (UVR), the major cause of skin malignancies, activates p1613,25, and p16-expressing cells are detected in premalignant epidermal

lesions such as actinic keratosis26,27,28. The high mutation rates of p16 in cutaneous squamous cell carcinoma and other skin malignancies5,29,30 indicate that it suppresses malignant

progression. However, it is unknown whether the activity of p16 in the normal epidermis and in premalignant lesions influences the development of disease. Furthermore, whether p16-expressing

cells in such early lesions can be targeted by senolytic therapy, and whether this may have therapeutic benefit, has not been tested. Using transgenic mice allowing tissue-specific p16

activation, we demonstrate that the persistent expression of p16 in a subset of cells within the epidermis induces hyperplasia and dysplasia, and promotes tumor formation following

mutagenesis. We show that p16 expression in mice and in cultured keratinocytes leads to Wnt-pathway activation, which contributes to epidermal hyperproliferation, and that senolytic

elimination of p16-expressing cells inhibits hyperplasia. These findings reveal that chronic p16 activity is sufficient to induce premalignant tissue changes through a non-cell-autonomous

mechanism, and uncover a potential tumor-promoting function of this gene during early tumorigenesis. RESULTS EPIDERMAL P16 INDUCTION CAUSES PARTIAL SENESCENCE FEATURES To study the effects

of p16-expressing cells on the adult skin we crossed mice carrying a doxycycline-activated human p16 gene (tet-p16)21 with K5-rtTA mice31, allowing its inducible activation in the basal

epidermis. Transgenic p16 protein was detected in ~40% of basal keratinocytes in the interfollicular epidermis (IFE) after 2 days of doxycycline (dox) treatment at 3 weeks of age (Fig. 

1a–c). Tissues showed reduced phosphorylated Rb levels, consistent with expected Rb activation (Fig. 1d). Staining for Ki67 revealed a dramatic reduction in the number of proliferating IFE

basal cells after 2 days of p16 activation (Fig. 1e, f). However, after 2 weeks of p16 activation, overall proliferation rates returned to those of control animals. Increased numbers of

cells expressing the senescence markers SA-βGal and SenTraGor32 were detected in mice that have undergone 2 weeks of p16 activation (Fig. 1g, h and Supplementary Fig. 1a, b), yet those were

lower than the numbers of p16-expressing cells, suggesting that not all transgene-expressing cells enter a fully senescent state. The average size of keratinocytes was increased in

p16-expressing mice, specifically cells expressing the transgene, resulting in a thickening of the epidermis, with no increase in cell numbers (Fig. 1i, j and Supplementary Fig. 1c–e).

Changes in other known markers of senescence such as p21, Lamin B1, DNA damage, or secreted factors were not detected. Together these findings indicate that p16 expression in epidermal

keratinocytes induces cell-cycle arrest, cell hypertrophy, and some features of senescence. p16 induction led to delayed growth of hair follicles (HF), reduced hair-follicle stem cell (HFSC)

proliferation, and the formation of deformed HFs during the anagen stage initiated at 3 weeks of age (Fig. 1k and Supplementary Fig. 1f, g), reminiscent of the effect of p14ARF

activation33. Upon prolonged p16 expression, mice displayed a decrease in the number of HFs and hair loss (Fig. 1l and Supplementary Fig. 1f, h), and the numbers of CD34+/CD49fhigh HFSCs

were dramatically reduced (Fig. 1m). p16 expression thus blocks the proliferation of HFSCs and leads to their gradual depletion. PROLONGED P16 EXPRESSION INDUCES HYPERPLASIA AND DYSPLASIA To

examine the effects of prolonged p16 expression on the epidermis, we activated the transgene for 6 months. At this timepoint we detected p16 protein in ~20% of IFE cells (Fig. 1c).

Strikingly, p16-expressing mice exhibited a thickened epidermis, in which both cell number and average cell size were increased (Fig. 2a–c). We observed an increased proportion of suprabasal

differentiated cells in p16-expressing skins, while basal layer marker expression (K14, K5) was maintained and expanded in some regions (Fig. 2a, d, e and Supplementary Fig. 2a, b). These

changes contrast with the normal young and aged mouse epidermis, which contains only a thin suprabasal layer (Supplementary Fig. 2c). p16-induced mice also often showed an expansion of the

dermis with dense collagen deposition regions, and a reduction of fat layer width (Supplementary Fig. 2d). Although we did not observe a prominent inflammatory reaction, numbers of epidermal

and dermal T-cells, as well as of macrophages, were increased (Supplementary Fig. 2e, f). Control mice in which GFP was activated instead of p16 for the same period did not show these

changes (Supplementary Fig. 2g). Strikingly, proliferation rates in the epidermis of mice expressing p16 for 6 months were dramatically increased, with 81% of basal keratinocytes expressing

Ki67 relative to 33% in control mice (Fig. 2f, g). Furthermore, some epidermal regions exhibited dysplasia (atypia), exhibiting changes in nuclear morphology and basal layer organization

(Fig. 2f, right panel). Co-staining for p16 and the proliferation marker Mcm7 revealed that 93% of basal IFE keratinocytes in which p16 was not activated (p16–) were proliferating, as

opposed to 26% of basal IFE cells in control mice (Fig. 2h, i). A corresponding increase in BrdU incorporation, marking cells in S-phase, was observed (Fig. 2j). These findings indicate that

prolonged p16 activity in a subset of epidermal cells is sufficient to induce the proliferation of the majority of surrounding keratinocytes. Surprisingly, 30% of p16-expressing (p16+)

cells also stained positive for Mcm7, and 5% stained positive for BrdU, suggesting that some of the transgene-expressing cells possessed the ability to proliferate (Fig. 2i, j). Chronic

epidermal p16 expression thus leads to hyperplasia and regional dysplasia, consistent with paracrine stimulation emanating from the p16+ cells. We tested whether p16-induced hyperplasia

influenced the regenerative capacity of the skin following wounding. Healing was delayed in mice in which p16 was activated for 2 weeks prior to wounding, and a similar delay was observed in

mice in which p16 was activated for 6 months (Supplementary Fig. 3), indicating that the p16-induced hyperplasia does not lead to enhanced regenerative capacity. A SUBSET OF CELLS IN HUMAN

ACTINIC KERATOSIS EXPRESS P16 The appearance of dysplastic regions in p16-expressing mice was reminiscent of early-stage human premalignant epidermal lesions. To establish whether p16+ cells

are detected within such lesions, we stained sections of human actinic keratosis (AK) lesions, the common precursors to cutaneous squamous cell carcinoma (cuSCC)34. Previous studies

reported varying percentages and staining patterns of p16 in these lesions26,27,28. We detected p16 protein in 47 of 72 lesions (65%), collected in three medical centers, with a mean of

11.3% p16+ cells per lesion in those that contained such cells (0.8–41.8%, _n_ = 47) (Fig. 2k, l). p16+ cells were often interspersed among or adjacent to dividing, p16–, cells (Fig. 2m, n

and Supplementary Fig. 4a). An average of 6% of the cells in AK lesions positively stained for SenTraGor, suggesting the presence of senescent cells in them (Supplementary Fig. 4b, c). Rb

protein was also detected in these lesions, excluding Rb deletion or silencing, which would render p16 ineffective (Fig. 2k, right panel). Furthermore, as observed in the p16-induced mice,

some p16+ cells were proliferating, based on Ki67 expression, yet in lower numbers than the surrounding cells (Fig. 2m, n). These findings are consistent with potential crosstalk between

p16+ and p16– subpopulations in human AK lesions, as suggested by the mouse model. P16 PROMOTES FORMATION OF CARCINOGEN-INDUCED PAPILLOMAS The finding that p16 expression was sufficient to

induce epidermal hyperplasia and dysplasia raised the possibility that cells carrying oncogenic mutations would be influenced by a p16-activated environment. We, therefore, tested whether

chronic p16 expression would affect the rates of papilloma formation after carcinogen treatment, using the DMBA/TPA model35. We treated K5-rtTA/tet-p16 mice, as well as sibling control

tet-p16 mice, with a single topical dose of the mutagen DMBA at 3 weeks of age, and 1 week later initiated p16 activation by dox treatment, and simultaneously began twice-weekly topical

treatments with TPA. We tracked papilloma formation for 5 months. Strikingly, p16-expressing mice developed approximately double the number of papillomas than control mice (Fig. 3a, b). This

indicates that p16 generates an environment that supports the ability of keratinocytes bearing an oncogenic mutation to initiate papillomas. Interestingly, the papillomas that developed in

p16-expressing mice were smaller than those in control mice, and contained fewer proliferating cells, which were more restricted to the basal epidermal regions of the lesions (Fig. 3c–g and

Supplementary Fig. 5a). p16+ cells were readily detected within these papillomas (Fig. 3e). SenTraGor staining indicated higher numbers of senescent cells within the p16-induced papillomas,

relative to the adjacent non-transformed epidermis (Supplementary Fig. 5b,c), consistent with previous reports indicating that p16 exerts a stronger pro-senescent effect when activated in

the context of a strong pro-mitotic signal36. Expression profiling of whole dissected papillomas revealed a dramatic elevation in markers of differentiated epidermal layers and reduced

expression of stem cell-associated markers, in addition to reduced expression of cell cycle, signaling, and cytoskeletal genes (Fig. 3h, i and Supplementary Fig. 5d, e). Together these

findings indicate that p16-expressing cells in the non-transformed tissue stimulate neighboring mutated oncogene-expressing cells to form growing premalignant tumors. However, p16 expression

within formed oncogene-expressing lesions restricts proliferation, increasing senescence and differentiation, and thereby reduces tumor growth, consistent with its known tumor-suppressive

roles. P16 EXPRESSION ACTIVATES WNT IN EPIDERMAL KERATINOCYTES We next set out to uncover the mechanisms by which p16-expressing cells stimulate proliferation and induce hyperplasia. We

generated triple-transgenic K5-rtTA/tet-p16/tet-GFP mice, in which p16 and GFP were co-activated, and treated the mice with dox for 6 months (Fig. 4a). Co-expression of GFP did not reduce

the numbers of detected p16+ cells (Supplementary Fig. 6a, b). We then isolated GFP+ keratinocytes from these mice and from K5-rtTA/tet-GFP control mice expressing only GFP (Fig. 4b and

Supplementary Fig. 6c). The transcriptome of GFP+ cells from p16-expressing mice revealed changes in multiple genes relative to GFP+ cells from control mice (624 upregulated and 750 genes

downregulated, Supplementary Data 1), representing pathways known to be involved in senescence, including cell-cycle regulation, cytoskeletal structure, cell adhesion, protein turnover and

metabolism (Fig. 4c). Several cytokine-encoding genes were upregulated, including _Ccl20_, _Cxcl1,_ and _Cxcl9_, as well as _Tgfβ_ ligands, yet these did not constitute a full SASP (Fig. 

4d). Interestingly, we found that the levels of several genes encoding Wnt ligands, as well as known Wnt transcriptional target genes, were elevated, while Fzd receptor-encoding genes showed

reduced mRNA levels (Fig. 4e). _Tcf7_, which encodes the Wnt-target-activating Tcf1 transcription factor (and a Wnt target itself)37,38 was upregulated; conversely, _Tcf7l1_ and _Tcf7l2_,

which respectively encode the Tcf3 and Tcf4 factors and are known to often act as Wnt target repressors37,38, were downregulated, consistent with transcriptional target activation (Fig. 4e).

The proliferation-associated _Ccnd_ genes (encoding Cyclin D proteins) which also respond to Wnt, were among the upregulated genes, as was often previously observed in senescent cells39.

Consistent with Wnt-pathway activation, we detected by immunohistology increased numbers of cells expressing nuclear β-catenin in the epidermis of p16-expressing mice, as well as cells

expressing Tcf1, Cyclin D1 and CD44 (Fig. 4f, g). Isolation and transcriptome analysis of the GFP– cell fraction from p16-expressing mice revealed that, relative to the GFP+ cells from the

same mice, this fraction was enriched for proliferation-associated gene sets (as expected from their hyperproliferative state), and showed reduced expression of senescence-related sets, such

as oxidative stress response, protein translation, and cytoskeletal reorganization (Supplementary Fig. 6c–e). Keratinocyte differentiation genes were somewhat elevated, reflecting the

presence of diverse differentiation states in the expanded epidermis (Supplementary Fig. 6f). We found that, similar to the GFP+ p16-expressing cells, the GFP– fraction showed elevation in

Wnt-associated genes relative to cells from control mice, and that the _Tcf7l1_ gene was further silenced in these cells (Supplementary Fig. 6g). Together these findings indicate that p16

expression leads to activation of Wnt-pathway genes in both the p16-expressing cells and the surrounding IFE cells. The activation of the Wnt pathway by p16 appeared to occur gradually after

prolonged p16 expression, as we did not observe elevation in Wnt-associated genes after 2 days or 2 weeks of p16 induction (Supplementary Fig. 6h). P16-INDUCED WNT SECRETION DRIVES

KERATINOCYTE PROLIFERATION To better dissect p16-induced Wnt activity, we overexpressed p16 in cultured primary mouse keratinocytes by lentiviral transduction. p16 overexpression led to

rapid cell-cycle arrest, acquisition of a senescent morphology, and expression of senescence-associated markers, yet, interestingly, with no evidence of SA-βGal activity (Fig. 5a, b). Within

10 days of p16 expression, the levels of multiple Wnt ligands and targets were upregulated in the p16-expressing cells (Fig. 5c). ELISA measurement indicated that Wnt3a levels in

conditioned medium (CM) of p16-expressing cells was elevated, reaching 3 ng/ml (Fig. 5d). Treatment of the cells with XAV-939, a Tankyrase1/2 inhibitor that causes β-catenin degradation40,

suppressed the upregulation of Wnt target genes, including the genes encoding Cyclin D1 and D2, as well as of Wnt ligands (Fig. 5c). The activator _Tcf7_ gene was upregulated in

p16-expressing cells, whereas the repressors _Tcf7l1_ and _Tcf7l2_ were both downregulated by p16 overexpression, as was observed in the transgenic mice, and XAV-939 treatment reversed this

effect (Fig. 5c). These findings indicate that p16 is sufficient to activate the Wnt pathway in keratinocytes, driving both target and ligand activation in a β-catenin-dependent manner. To

test whether Wnt activation and secretion by p16-expressing keratinocytes can influence naive cells, we treated proliferating primary keratinocytes with conditioned media from p16-expressing

or control (empty vector-expressing) keratinocytes, or, as a positive control, with recombinant Wnt3a. While control-cell CM had no effect on the numbers of treated cells, the CM of

p16-expressing cells led to enhanced proliferation of the treated cells, resulting in a 1.7-fold increase in cell numbers after 4 days of treatment, similar to the effect of Wnt3a (Fig. 5e).

This stimulatory effect was almost completely lost when XAV-939 was administered together with CM or Wnt3a, indicating that proliferation is β-catenin mediated (Fig. 5e). Treated cells

showed upregulation of Wnt target genes and downregulation of _Tcf71,2_, as well as increased levels of genes encoding Wnt ligands (Fig. 5f), mirroring the changes observed in the

p16-expressing mice. These changes were blocked by XAV-939 (Fig. 5f). Together, these findings indicate that Wnt secretion by p16-expressing cells can stimulate the proliferation of naive

keratinocytes, and that Wnt signals act on both the p16-expressing cells themselves and on the non-expressing proliferating cells, activating transcriptional targets as well as Wnt ligands,

in a β-catenin-dependent manner (Supplementary Fig. 7). THE WNT-PATHWAY MEDIATES P16-INDUCED HYPERPLASIA We next tested whether Wnt activation contributes to the hyperplasia observed in

p16-expressing mice. We activated p16 for 6 months, and then treated the mice systemically with XAV-939 for 1 week. XAV-939 treatment led to a significant reduction in epidermal

proliferation rates, as well as a reduction in the expression of Wnt ligand and target genes (Fig. 6a–c). Wnt inhibition was therefore sufficient for a partial reversal of the ongoing

hyperplastic effects of p16 expression. We next tested whether co-activation of the repressive Tcf3 (Tcf7l1) together with p16 would influence Wnt-pathway activation and hyperplasia. We

crossed K5-rtTA/tet-p16 mice with mice carrying an inducible Tcf3 transgene (tet-Tcf3)41 to generate triple-transgenic mice. Co-activation of Tcf3 together with p16 for 6 months resulted in

significantly reduced epidermal hyperplasia, as well as in reduced levels of Wnt ligand and target genes (Fig. 6d–g). Tcf3 expression thus largely blocks the pro-proliferative effects of

p16. Together these results indicate that activation of the Wnt pathway upon chronic p16 expression contributes to the induction of epidermal hyperplasia. SENOLYTIC ELIMINATION OF P16+ CELLS

SUPPRESSES HYPERPLASIA To study whether the continued presence of p16-expressing cells in the epidermis is required for maintenance of hyperplasia, even after prolonged stimulation, we

tested whether these cells can be eliminated by senolytic treatment. The Bcl-2 protein-family inhibitor ABT-737 and its analog ABT-263 (navitoclax) inhibit Bcl-2, Bcl-xl, and Bcl-w, and were

shown to have senolytic activity in various settings16,42,43. We previously showed that ABT-737 eliminates epidermal cells in which the p53 activator, p14ARF, was overexpressed42. However,

it was unclear, in light of the partial induction of senescence by p16, whether ABT-737 would be effective in eliminating p16-expressing cells. Following six months of p16 induction, we

treated mice with ABT-737 six times over nine days and then sacrificed the mice. ABT-737-treated mice showed a significant reduction in the percentage of p16+ cells in the IFE (an

approximate 60% reduction), and a corresponding reduction in basal cell proliferation and epidermal thickness, indicating partial reversal of the hyperplasia (Fig. 7a–d). Furthermore, the

expression of Wnt ligands and targets was significantly reduced in the ABT-737-treated mice (Fig. 7e). p16-expressing cells are thus sensitive to senolytic treatment, and their residence in

the epidermis is necessary for the ongoing maintenance of Wnt stimulation and consequent hyperplasia. DISCUSSION The function of p16 as one of the most important tumor suppressor genes is

well established, and is supported by its recurrent loss-of-function alterations in a wide variety of human malignancies and by numerous studies in which it was shown to block

immortalization and tumorigenesis4. An important additional dimension to p16 function is its activation, often linked to cellular senescence, in a subset of cells within aging tissues.

Recent studies revealed that p16-expressing cells contribute to aging processes and to associated diseases, and have highlighted the question of whether such resident cells influence the

propensity of tissues to develop cancer11,16. Our findings reveal that the continuous expression of p16 itself in the epidermis has pro-proliferative and pro-tumorigenic effects. p16

expression, when activated after carcinogen treatment, promoted the initiation of papillomas, indicating that it generates an environment that supports the ability of cells carrying an

oncogenic mutation to initiate tumors. p16 may thus function in some settings as a double-edged sword, in the short-term blocking proliferation of damaged cells, but, over time, stimulating

the proliferation of neighboring cells and thereby contributing to the formation of premalignant lesions. Our findings indicate that Wnt signaling plays a central role in mediating

p16-induced hyperplasia. Wnt and β-catenin are known to regulate the acquisition of stem cell traits, particularly of bulge hair-follicle stem cells, and also to promote the proliferation of

interfollicular epidermal cells and the formation of epidermal tumors44,45,46,47. p16 overexpression, both in vivo and in cultured primary keratinocytes, led to Wnt ligand secretion and

Wnt-pathway activation, stimulating the hyperproliferation of target cells that do not express p16. Interestingly, the recipient cells not only activated Wnt transcriptional targets, but

also upregulated Wnt ligand expression, thereby generating a positive feedback loop. In both the p16-expressing and the naive cells, Wnt ligand activation was β-Catenin dependent. This mode

of combined autocrine and paracrine stimulation is consistent with previous dissection of Wnt activity in the IFE, in which it was shown that basal cells expressing the Wnt target Axin2 also

secrete Wnt ligands, in a β-Catenin-dependent manner44. Wnt over-activation in the p16-expressing epidermis occurred gradually over a prolonged period. In contrast, in cultured

keratinocytes, Wnt-pathway genes were activated within days of p16 overexpression. The reasons for this difference are unclear, but may be due to more stringent negative feedback loops

active in the living tissue. Wnt activation could be expected to also promote stem cell identity and function in p16-expressing skins. Recent studies have shown that senescence can promote

stem cell-like traits in some settings48,49,50. However, we found that p16 expression blocked the proliferation of hair-follicle stem cells, and led to their gradual depletion from the

tissue, results similar to those observed in a previous study51. Wound healing, a stem cell-dependent process52, was delayed in the p16-expressing mice, even during epidermal hyperplasia. It

is therefore possible that additional actions of p16 counteract the stimulation of stem cell function by Wnt, particularly in the days immediately following p16 activation, in which the

proliferation of HFSCs is arrested and Wnt levels are not yet elevated. p16 expression induced only partial features of senescence. Numbers of SA-βGal+ and SenTraGor+ cells were elevated,

but were lower than those of p16+ cells. We observed an increase in keratinocyte size, a hallmark of senescence, and changes in senescence-associated gene sets, such as genes involved in

cytoskeletal structure and protein translation. Yet various other known markers of senescence53 were unchanged. Several cytokine-encoding genes were upregulated in p16-expressing cells, and

while these could potentially contribute to the observed phenotypes, they did not constitute a full SASP response, which typically involves high levels of multiple cytokines. This is

consistent with previous observations that p16 does not activate the SASP20. While several past studies have linked Wnt ligand activation to senescence, the paracrine actions of Wnt were not

addressed, and in some settings Wnt activity was reported to counteract senescence50,54,55,56. Our work thus reveals a tumor-promoting contribution of p16 that does not necessitate a broad

SASP but acts mainly through Wnt. Accumulation of p16-expressing cells may preferentially occur in epidermal regions exposed to damage. p16 is known to be activated by UVR, the major cause

of human skin malignancies13,25. UV-irradiated mouse skins show a thickening reminiscent of that induced by p16 overexpression57. Expanding previous observations26,27,28, we show that p16 is

indeed often expressed in actinic keratosis lesions, which are among the most common human premalignant growths and are induced by UVR. Furthermore, we show that p16+ cells in these lesions

are in close proximity to p16– dividing cells, in a pattern similar to that observed in our mouse model. This pattern is consistent with paracrine interaction between these cell subsets,

yet such crosstalk between them remains to be directly proven. Mutations in p53 occur very early in skin tumorigenesis, as a direct consequence of UVR34. Interestingly, this may result in

further enhanced compensatory p16 expression and senescence in such lesions9,58, possibly providing a proliferation-promoting environment. Upon progression to cuSCC p16 is often

mutated29,30, suggesting that its counter-proliferative functions outweigh the potential benefit of its activity. Our observation that the expression of p16 restricts proliferation within

growing papillomas suggests that its activity following tumor initiation blocks further growth, as is expected from its tumor-suppressive function (Supplementary Fig. 7). Interestingly, the

major effect of p16 within the papillomas was restriction of proliferation and maintenance of a differentiated state, together with increased senescence. Unlike the untransformed tissue,

p16-expressing papillomas did not show overall increased Wnt activation. The actions of p16 in cells that express an active oncogenic driver may thus be distinct from its effects in normal

cells. Pharmacologic elimination of senescent cells through drugs with senolytic activity has been shown in recent studies to be beneficial in a variety of preclinical disease models, and

some of these drugs are currently being tested in clinical trials16,17,59. Our findings suggest that senolytic treatment may be beneficial for early premalignant epidermal lesions.

Altogether, we reveal in this study a novel dimension of the functions of p16 in skin tumorigenesis, suggesting a non-cell-autonomous contribution to the formation of premalignant lesions,

and a potential avenue for blocking this function through senolytic treatment. METHODS MICE The following transgenic mouse lines were crossed as indicated: tet-p16 (mixed C57Bl6 and 129sv),

K5-rtTA (FVB), tet-GFP (mixed C57Bl6 and 129sv) and tet-Tcf3 (FVB). For transgene induction mice received 2 mg/mL doxycycline (Bio Basic) in their drinking water starting at the age of 3

weeks, and treatment was maintained for the indicated durations. In all experiments, the experimental and control mouse groups were siblings grown and treated together. For short-term

inductions (up to 2 weeks), both males and females were used; for inductions of 6 months and wound-healing assays all presented results are from male mice, whereas for papilloma formation

assays only data of females is shown, due to higher average lesion formation rates. For bromodeoxyuridine (BrdU) labeling mice were injected intraperitoneally with 100 mg/Kg BrdU 2 h prior

to sacrifice. For Wnt-pathway inhibition, mice that received dox for 6 months were injected intraperitoneally with 200 μg XAV-939 (Sigma-Aldrich) in 10% DMSO/90% 0.9% NaCl twice a day for

seven consecutive days. ABT-737 (WuXi AppTec), was prepared at 75 mg/kg in 30% propylene glycol, 5% Tween 80, 3.3% dextrose in water pH 4–5, and mice were injected intraperitoneally 6 times

over 9 days after 6 months of induction. Control mice were injected with vehicle solutions. Upon sacrifice, mice were shaved and their dorsal skins were paraffin-embedded for immunohistology

or frozen in OCT solution for SA-βgal staining. Chemically-induced skin carcinogenesis was conducted as described35. Three-week-old K5-rtTA/tet-p16 and control tet-p16 mice were shaved, and

one day later received a single topical treatment of 50 nmole DMBA (7,12-Dimethylbenz[_a_]anthracene) in 200 μl acetone. After one week dox was added to the drinking water and the mice

began twice-weekly topical treatments with 4 μg TPA (12-O-tetradecanoyl phorbol-13-acetate) in 200 μl acetone. Mice were examined for 23 additional weeks for papilloma detection and

measurement and were sacrificed 10 days after the last TPA treatment. Papilloma numbers were scored weekly, with diameters measured by caliper. Only papillomas with ≥1 mm diameter were

scored. Individual whole papillomas were excised and processed for RNA extraction and profiling, or paraffin embedded. For wound-healing assays dorsal skins were shaved and treated with

depilating cream two days prior to wound formation; mice were anesthetized, their back skin was sterilized with povidone-iodine solution, and 1 cm2 full-thickness excision wounds were

introduced to the upper paravertebral area using sterile scissors. Wounds were measured daily with a caliper. Animals were housed in specific pathogen-free conditions, at 20–24 °C, 30–70%

humidity and a 12:12 h light-dark cycle. The Hebrew University is an AAALAC International accredited institute. All experiments were conducted with approval from the Hebrew University Animal

Care and Use Committee. CULTURED KERATINOCYTES Epidermal keratinocytes were isolated following published procedures60. Dorsal skin was removed from newborn mice, and was incubated with

dispase (Sigma-Aldrich) followed by isolation of the epidermis and treatment with trypsin (Biological Industries). Keratinocytes were plated on fibroblast feeder cells for four passages and

then plated in regular tissue culture dishes. Passage 16 cells were infected with a lentivirus expressing human p16 (pLV-hCDKN2a-GFP-Puro, Vector Builder) or a control empty lentiviral

vector, and selected with puromycin (2 μg/mL). 10 days after infection cells were fixed and stained as indicated, or lysed for RNA extraction. Conditioned medium (CM) was collected from

cells after 3 days of culture and filtered through a 0.45 μm filter. Wnt3a levels in CM were measured using Wnt3a ELISA kit (R&D Systems, DY1324B-05). Naive keratinocytes were treated

with CM in a 1:1 ratio with fresh media, or with 10 ng/ml recombinant Wnt3a (R&D Systems). RNA was collected from treated cells one day after treatment. For proliferation measurements,

104 naive keratinocytes were plated in 24-well plates, treated with CM from p16-expressing or control cells, or with Wnt3a (6 replicates for each), and cell numbers were scored after 4 days.

XAV-939 (Sigma-Aldrich, 5 μM) was added where indicated and control samples were treated with DMSO vehicle. HUMAN ACTINIC KERATOSIS SAMPLES A total of 72 human actinic keratosis lesions,

surgically resected, were collected in three medical centers: 19 patient samples were obtained from the laboratory of Histology-Embryology, Athens Medical School under approval of the

Bio-Ethics Committee of Athens Medical School, in accordance with the Declaration of Helsinki and local laws and regulations, following written consent from patients; 42 lesion samples from

35 patients were obtained from the University of Arizona, USA, approved by the University of Arizona Institutional Review Board protocol, with patient informed consent; 11 patient samples

were obtained from the Hadassah-Hebrew University Medical Center, Jerusalem, Israel under approval of the Hadassah Medical Center Helsinki Committee for usage of fully anonymized archival

specimens, and approved waiver of patient consent. All slides underwent pathologist review to verify lesion type. Patients were aged 38 to 92, with equal representation of males and females.

All available patient samples were included in the analysis, with no selection. IMMUNOHISTOLOGY AND IMMUNOCYTOLOGY Immunohistologic stains were performed on 5 μm paraffin sections of

collected mouse back skins, individual papillomas, or human AKs, according to standard protocols, using Peroxidase Substrate Kit (Vector) or fluorescently-labeled secondary antibodies

(Jackson). The following primary antibodies were used: human p16INK4a (Abcam, ab108349), human p16INK4a (BD Pharmingen, 551153), p-Rb (Santa Cruz, sc-16669-R), Ki67 (Abcam, ab16667), BrdU

(Bio-Rad, MCA2060), Mcm7 (Santa Cruz, sc-9966), E-Cadherin (BD Pharmingen, 610182), K14 (Progen, GP-CK14), K10 (Santa Cruz, sc-23877), K5 (Progen, GP-CK5), Rb (Santa Cruz, sc-102), CyclinD1

(Abcam, ab16663), β-catenin (BD Pharmingen, 610153), Tcf1 (Invitrogen, MA5-14965), Tcf3 (Santa Cruz, sc-8635), CD44 (eBioscience, 14-0441-82), Vimentin (Cell Signaling, 5741), CD3 (Bio-Rad,

MCA1477), and F4/80 (Bio-Rad, MCA497). Masson’s Trichrome stain was conducted using staining kit (Sigma-Aldrich, HT15) according to manufacturer instructions. SenTraGor stain was conducted

as described32. For SA-βGal stains, 12 μm frozen sections of OCT-embedded mouse skins, or cultured keratinocytes, were fixed in 0.5% glutaraldehyde for 15 min and stained using standard

procedures33. For co-stains of cultured keratinocytes for BrdU, p16, and K14, cells were treated for 2 h with 10 μg/mL BrdU and then fixed with cold ethanol, treated with 1 M HCl followed by

0.1 M sodium tetraborate (Na2B4O7) for neutralization, and then stained, as indicated, using PBS + 10% fetal calf serum (FCS) for antibody suspensions. Brightfield images were collected

using an Olympus CX41 and DS-Fi1 camera and processed using NIS-Elements software (Nikon). Fluorescent images were collected using an Olympus FV1000 confocal microscope and processed with

the FV10-ASW software. All mouse skin images shown are of back skin. Percentages of SA-βGal+, SenTraGor+, p16+, Ki67+, Mcm7+, BrdU+, CycD1+, nuclear β-Catenin+, Tcf1+ and epidermal CD3+

cells, out of total IFE or basal IFE cells as indicated, were visually scored from 5 or more microscopic field images per mouse, and 2 or more fields from human AK lesions, with 100–1000

total cells scored per sample. Absolute numbers per microscopic fields of hair follicles, IFE cells, dermal CD3+ and dermal F4/80+ cells were scored from 8 or more fields per mouse (10×

magnification for HFs, 40× for others). Percentage of Ki67+ area in papillomas out of epithelial cell area was measured using NIS-elements software (Nikon). Epidermal thickness was measured

by NIS elements on 36 or more measurements per mouse. Keratinocyte area calculation was performed on skin sections co-stained for E-Cadherin and p16 by measuring cell circumference using

NIS-Elements, with 60 cells or more measured per mouse. FACS ANALYSIS Adult epidermal cells were isolated following published procedures61. Mouse back skins were shaved and excised, and

adipose and muscle tissues were removed. The skin was then incubated with 0.25% trypsin solution overnight at 4 °C, and the epidermis was scraped off, minced and suspended in a trypsin/PBS

with 0.5% BSA solution, followed by filtration through a 70 μm cell strainer. Antibody stains were performed using standard procedures. For hair-follicle stem cell analysis live cells were

stained for Sca-1, CD34, and CD49f, with three mice pooled in each group. Isolation of GFP+ keratinocytes for mRNA-seq profiling was conducted as described above, and live cells were stained

for CD140a, CD45 and CD31 to gate out fibroblasts, immune and endothelial cells, respectively. For scoring of K10+ epidermal cell percentage, adult epidermal cells were isolated from ear

skin samples to obtain improved dissociation. Ear epidermal layers were obtained by digestion with 2 mg/mL Dispase II (Roche Diagnostics) in PBS solution for 40 min at 37 °C, and treated

with 2 mg/mL collagenase type II (Worthington Biochemicals) and 1 mg/mL DNase I (Roche) in PBS with 2% fetal calf serum for 25 min at 37 °C. Samples were then incubated with 0.5 M EDTA for

10 min and filtered through a 70 μm strainer. Cells were treated with fixation and permeabilization solution (BD Pharmingen) for 15 min and stained for K10 and secondary

fluorescent-conjugated antibody. Samples were sorted on a FACS Aria III system using FACSDiva software (BD Biosciences), or were analyzed on a MACSQuant Analyzer (Miltenyi Biotec).

Post-acquisition analysis was conducted with FCS Express (De Novo Software). Antibodies used: CD34 (BD Bioscience, 562608), Sca-1 (Biolegend, 108113), CD49f (eBioscience, 17-0495-80), K10

(Santa Cruz, sc-23877), CD45 (eBioscience, 25-0451-82), CD140a (eBioscience, 17-1401-81), and CD31 (eBioscience, 17-0311-80). RNA EXTRACTION, QRT-PCR, AND EXPRESSION PROFILING Total RNA from

whole skin samples and whole papillomas was extracted using the Tissue RNA purification kit (Norgen Biotek). RNA from isolated epidermis was extracted following epidermal separation using

the RNeasy Minikit (Qiagen). RNA extraction from FACS-sorted cells was done using the Single Cell RNA Purification Kit (Norgen Biotek). RNA was extracted from primary keratinocytes using

Direct-zol kit (Zymo Research, #R2051). cDNA was prepared using iScript cDNA synthesis kit (Bio-Rad), and qRT-PCR was conducted using iTaq Universal SYBR Green Supermix (Bio-Rad) in

triplicates, with expression values normalized to _Hprt_ and _Ppia_ expression levels for mouse and primary keratinocyte samples, respectively. Primer lists are provided in Supplementary

Table 1. All RT-PCR graphs conducted on mouse samples indicate values across individual mice. We conducted two mRNA-seq profilings of isolated keratinocytes. In the first, we analyzed 80,000

GFP+ keratinocytes from K5-rtTA/tet-p16/tet-GFP and control K5-rtTA/tet-GFP mice, after exclusion of CD45+, CD31+ and CD140a+ cells. Cells were pooled from three mice in each replicate,

with four replicates used for each group, obtained in three independent experiments. In the second experiment, we similarly isolated 80,000 GFP+ cells from the same mouse groups (after

CD45+, CD31+ and CD140a+ exclusion), as well as 20,000 GFP− cells from K5-rtTA/tet-p16/tet-GFP mice, pooling 3 mice per each sample. _n_ = 4, 4, 5 for these groups, respectively. mRNA-Seq

transcriptional profiling was conducted using an adapted CEL-Seq2 protocol62: 3′ cDNA was synthesized and barcoded, followed by RNA synthesis, amplification by in vitro transcription, and

library generation for paired-end sequencing. Reads were demultiplexed, quality filtered, and trimmed for adapters and poly-A tail using Cutadapt and aligned with the mouse genome (GRCm38)

using Tophat2. Differentially expressed genes were determined by DESeq2, using a _P_ < 0.05 cutoff. Quality control and principal component analyses were conducted using R: packages

‘RColorBrewer_1.1-2’, ‘pheatmap_1.0.10’, ‘ggplot2_3.1.0’ and ggrepel_0.8.0. Enrichment of gene sets was determined using the hypergeometric test through Metascape63, with adjusted _P_ values

calculated using the Benjamini-Hochberg procedure. Profiling of papillomas was done with the same method, using RNA from whole isolated lesions, 4 control and 3 p16-expressing. Analysis of

gene sets in these comparisons was done for up- or downregulated genes with Log2fold change > 0.6 at _P_ < 0.05. We excluded in this analysis genes not expressed in keratinocytes,

based on our profiling of FACS-sorted populations, to reduce biases introduced by differential inclusion of non-epithelial cell types within the lesions. Gene sets representing epidermal

keratinocyte differentiation states were compiled from Joost et al.64. STATISTICAL ANALYSIS Two-tailed unpaired Student’s _t_ test was used to derive _P_ values in all comparisons stating

usage of _t_ test. Other methods were used as indicated. REPORTING SUMMARY Further information on research design is available in the Nature Research Reporting Summary linked to this

article. DATA AVAILABILITY The mRNA-seq expression profiles were deposited in GEO database, under accession number GSE146979 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi]. All the other

data supporting the findings of this study are available within the article and its supplementary information files and from the corresponding author upon reasonable request. A reporting

summary for this article is available as a Supplementary Information file. REFERENCES * Sharpless, N. E. & DePinho, R. A. How stem cells age and why this makes us grow old. _Nat. Rev.

Mol. Cell Biol._ 8, 703–713 (2007). Article  CAS  PubMed  Google Scholar  * de Magalhaes, J. P. How ageing processes influence cancer. _Nat. Rev. Cancer_ 13, 357–365 (2013). Article  PubMed

  CAS  Google Scholar  * Campisi, J. Aging, cellular senescence, and cancer. _Annu. Rev. Physiol._ 75, 685–705 (2013). Article  CAS  PubMed  Google Scholar  * Kim, W. Y. & Sharpless, N.

E. The regulation of INK4/ARF in cancer and aging. _Cell_ 127, 265–275 (2006). Article  CAS  PubMed  Google Scholar  * Sanchez-Vega, F. et al. Oncogenic signaling pathways in the cancer

genome atlas. _Cell_ 173, 321–337 (2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * He, S. & Sharpless, N. E. Senescence in health and disease. _Cell_ 169, 1000–1011

(2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Zindy, F., Quelle, D. E., Roussel, M. F. & Sherr, C. J. Expression of the p16INK4a tumor suppressor versus other INK4

family members during mouse development and aging. _Oncogene_ 15, 203–211 (1997). Article  CAS  PubMed  Google Scholar  * Krishnamurthy, J. et al. Ink4a/Arf expression is a biomarker of

aging. _J. Clin. Invest._ 114, 1299–1307 (2004). Article  CAS  PubMed  PubMed Central  Google Scholar  * Yamakoshi, K. et al. Real-time in vivo imaging of p16Ink4a reveals cross talk with

p53. _J. Cell Biol._ 186, 393–407 (2009). Article  CAS  PubMed  PubMed Central  Google Scholar  * Burd, C. E. et al. Monitoring tumorigenesis and senescence in vivo with a

p16(INK4a)-luciferase model. _Cell_ 152, 340–351 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Baker, D. J. et al. Naturally occurring p16(Ink4a)-positive cells shorten

healthy lifespan. _Nature_ 530, 184–189 (2016). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic

ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. _Cell_ 88, 593–602 (1997). Article  CAS  PubMed  Google Scholar  * Sorrentino, J. A. et al. p16INK4a

reporter mice reveal age-promoting effects of environmental toxicants. _J. Clin. Invest._ 124, 169–173 (2014). Article  CAS  PubMed  Google Scholar  * Chang, J. et al. Clearance of senescent

cells by ABT263 rejuvenates aged hematopoietic stem cells in mice. _Nat. Med._ 22, 78–83 (2016). Article  ADS  CAS  PubMed  Google Scholar  * Demaria, M. et al. Cellular senescence promotes

adverse effects of chemotherapy and cancer relapse. _Cancer Discov._ 7, 165–176 (2017). Article  CAS  PubMed  Google Scholar  * Childs, B. G. et al. Senescent cells: an emerging target for

diseases of ageing. _Nat. Rev. Drug Discov._ 16, 718–735 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Xu, M. et al. Senolytics improve physical function and increase

lifespan in old age. _Nat. Med._ 24, 1246–1256 (2018). Article  CAS  PubMed  PubMed Central  Google Scholar  * Matheu, A. et al. Increased gene dosage of Ink4a/Arf results in cancer

resistance and normal aging. _Genes Dev._ 18, 2736–2746 (2004). Article  CAS  PubMed  PubMed Central  Google Scholar  * Faget, D. V., Ren, Q. & Stewart, S. A. Unmasking senescence:

context-dependent effects of SASP in cancer. _Nat. Rev. Cancer_ 19, 439–453 (2019). Article  CAS  PubMed  Google Scholar  * Coppe, J. P. et al. Tumor suppressor and aging biomarker

p16(INK4a) induces cellular senescence without the associated inflammatory secretory phenotype. _J. Biol. Chem._ 286, 36396–36403 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar

  * Helman, A. et al. p16(Ink4a)-induced senescence of pancreatic beta cells enhances insulin secretion. _Nat. Med_. 22, 412–420 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar

  * Dimri, G. P. et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. _Proc. Natl Acad. Sci. USA_ 92, 9363–9367 (1995). Article  ADS  CAS  PubMed 

PubMed Central  Google Scholar  * Ressler, S. et al. p16INK4A is a robust in vivo biomarker of cellular aging in human skin. _Aging Cell_ 5, 379–389 (2006). Article  CAS  PubMed  Google

Scholar  * Velarde, M. C., Flynn, J. M., Day, N. U., Melov, S. & Campisi, J. Mitochondrial oxidative stress caused by Sod2 deficiency promotes cellular senescence and aging phenotypes in

the skin. _Aging_ 4, 3–12 (2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Pavey, S., Russell, T. & Gabrielli, B. G2 phase cell cycle arrest in human skin following UV

irradiation. _Oncogene_ 20, 6103–6110 (2001). Article  CAS  PubMed  Google Scholar  * Hodges, A. & Smoller, B. R. Immunohistochemical comparison of p16 expression in actinic keratoses

and squamous cell carcinomas of the skin. _Mod. Pathol._ 15, 1121–1125 (2002). Article  PubMed  Google Scholar  * Salama, M. E. et al. p16INK4a expression in actinic keratosis and Bowen’s

disease. _Br. J. Dermatol_. 149, 1006–1012 (2003). Article  CAS  PubMed  Google Scholar  * Nilsson, K., Svensson, S. & Landberg, G. Retinoblastoma protein function and p16INK4a

expression in actinic keratosis, squamous cell carcinoma in situ and invasive squamous cell carcinoma of the skin and links between p16INK4a expression and infiltrative behavior. _Mod.

Pathol._ 17, 1464–1474 (2004). Article  CAS  PubMed  Google Scholar  * Pickering, C. R. et al. Mutational landscape of aggressive cutaneous squamous cell carcinoma. _Clin. Cancer Res._ 20,

6582–6592 (2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Li, Y. Y. et al. Genomic analysis of metastatic cutaneous squamous cell carcinoma. _Clin. Cancer Res._ 21, 1447–1456

(2015). Article  CAS  PubMed  PubMed Central  Google Scholar  * Diamond, I., Owolabi, T., Marco, M., Lam, C. & Glick, A. Conditional gene expression in the epidermis of transgenic mice

using the tetracycline-regulated transactivators tTA and rTA linked to the keratin 5 promoter. _J. Invest. Dermatol._ 115, 788–794 (2000). Article  CAS  PubMed  Google Scholar  * Evangelou,

K. et al. Robust, universal biomarker assay to detect senescent cells in biological specimens. _Aging Cell_ 16, 192–197 (2017). Article  CAS  PubMed  Google Scholar  * Tokarsky-Amiel, R. et

al. Dynamics of senescent cell formation and retention revealed by p14ARF induction in the epidermis. _Cancer Res._ 73, 2829–2839 (2013). Article  CAS  PubMed  Google Scholar  * Ratushny,

V., Gober, M. D., Hick, R., Ridky, T. W. & Seykora, J. T. From keratinocyte to cancer: the pathogenesis and modeling of cutaneous squamous cell carcinoma. _J. Clin. Invest._ 122, 464–472

(2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Abel, E. L., Angel, J. M., Kiguchi, K. & DiGiovanni, J. Multi-stage chemical carcinogenesis in mouse skin: fundamentals

and applications. _Nat. Protoc._ 4, 1350–1362 (2009). Article  CAS  PubMed  PubMed Central  Google Scholar  * Takahashi, A. et al. Mitogenic signalling and the p16INK4a-Rb pathway cooperate

to enforce irreversible cellular senescence. _Nat. Cell Biol._ 8, 1291–1297 (2006). Article  CAS  PubMed  Google Scholar  * Yi, F. et al. Opposing effects of Tcf3 and Tcf1 control Wnt

stimulation of embryonic stem cell self-renewal. _Nat. Cell Biol._ 13, 762–770 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar  * Lien, W. H. et al. In vivo transcriptional

governance of hair follicle stem cells by canonical Wnt regulators. _Nat. Cell Biol._ 16, 179–190 (2014). Article  CAS  PubMed  PubMed Central  Google Scholar  * Lucibello, F. C., Sewing,

A., Brusselbach, S., Burger, C. & Muller, R. Deregulation of cyclins D1 and E and suppression of cdk2 and cdk4 in senescent human fibroblasts. _J. Cell Sci._ 105(Pt 1), 123–133 (1993).

CAS  PubMed  Google Scholar  * Huang, S. M. et al. Tankyrase inhibition stabilizes axin and antagonizes Wnt signalling. _Nature_ 461, 614–620 (2009). Article  ADS  CAS  PubMed  Google

Scholar  * Nguyen, H., Rendl, M. & Fuchs, E. Tcf3 governs stem cell features and represses cell fate determination in skin. _Cell_ 127, 171–183 (2006). Article  CAS  PubMed  Google

Scholar  * Yosef, R. et al. Directed elimination of senescent cells by inhibition of BCL-W and BCL-XL. _Nat. Commun._ 7, 11190 (2016). Article  ADS  CAS  PubMed  PubMed Central  Google

Scholar  * Ovadya, Y. et al. Impaired immune surveillance accelerates accumulation of senescent cells and aging. _Nat. Commun._ 9, 5435 (2018). Article  ADS  CAS  PubMed  PubMed Central 

Google Scholar  * Lim, X. et al. Interfollicular epidermal stem cells self-renew via autocrine Wnt signaling. _Science_ 342, 1226–1230 (2013). Article  ADS  CAS  PubMed  PubMed Central 

Google Scholar  * Lien, W. H. & Fuchs, E. Wnt some lose some: transcriptional governance of stem cells by Wnt/beta-catenin signaling. _Genes Dev._ 28, 1517–1532 (2014). Article  CAS 

PubMed  PubMed Central  Google Scholar  * Clevers, H., Loh, K. M. & Nusse, R. Stem cell signaling. An integral program for tissue renewal and regeneration: Wnt signaling and stem cell

control. _Science_ 346, 1248012 (2014). Article  PubMed  CAS  Google Scholar  * Sherwood, V. & Leigh, I. M. WNT signaling in cutaneous squamous cell carcinoma: a future treatment

strategy? _J. Invest. Dermatol._ 136, 1760–1767 (2016). Article  CAS  PubMed  Google Scholar  * Mosteiro, L., et al. Tissue damage and senescence provide critical signals for cellular

reprogramming in vivo. _Science_ 354, aaf445 (2016). Article  PubMed  CAS  Google Scholar  * Ritschka, B. et al. The senescence-associated secretory phenotype induces cellular plasticity and

tissue regeneration. _Genes Dev._ 31, 172–183 (2017). Article  CAS  PubMed  PubMed Central  Google Scholar  * Milanovic, M. et al. Senescence-associated reprogramming promotes cancer

stemness. _Nature_ 553, 96–100 (2018). Article  ADS  CAS  PubMed  Google Scholar  * Boquoi, A. et al. Reversible cell cycle inhibition and premature aging features imposed by conditional

expression of p16Ink4a. _Aging Cell_ 14, 139–147 (2015). Article  CAS  PubMed  Google Scholar  * Ito, M. et al. Stem cells in the hair follicle bulge contribute to wound repair but not to

homeostasis of the epidermis. _Nat. Med._ 11, 1351–1354 (2005). Article  CAS  PubMed  Google Scholar  * Gorgoulis, V. et al. Cellular senescence: defining a path forward. _Cell_ 179, 813–827

(2019). Article  CAS  PubMed  Google Scholar  * Ye, X. et al. Downregulation of Wnt signaling is a trigger for formation of facultative heterochromatin and onset of cell senescence in

primary human cells. _Mol. Cell_ 27, 183–196 (2007). Article  CAS  PubMed  PubMed Central  Google Scholar  * Adams, P. D. & Enders, G. H. Wnt-signaling and senescence: A tug of war in

early neoplasia? _Cancer Biol. Ther._ 7, 1706–1711 (2008). Article  CAS  PubMed  Google Scholar  * Basu, D., Reyes-Mugica, M. & Rebbaa, A. Role of the beta catenin destruction complex in

mediating chemotherapy-induced senescence-associated secretory phenotype. _PLoS ONE_ 7, e52188 (2012). Article  ADS  CAS  PubMed  PubMed Central  Google Scholar  * Tavana, O. et al. Absence

of p53-dependent apoptosis leads to UV radiation hypersensitivity, enhanced immunosuppression and cellular senescence. _Cell Cycle_ 9, 3328–3336 (2010). Article  CAS  PubMed  PubMed Central

  Google Scholar  * Leong, W. F., Chau, J. F. & Li, B. p53 deficiency leads to compensatory up-regulation of p16INK4a. _Mol. Cancer Res._ 7, 354–360 (2009). Article  CAS  PubMed  Google

Scholar  * Justice, J. N. et al. Senolytics in idiopathic pulmonary fibrosis: Results from a first-in-human, open-label, pilot study. _EBioMedicine_ 40, 554–563 (2019). Article  PubMed 

PubMed Central  Google Scholar  * Nowak, J. A. & Fuchs, E. Isolation and culture of epithelial stem cells. _Methods Mol. Biol._ 482, 215–232 (2009). Article  CAS  PubMed  PubMed Central

  Google Scholar  * Jensen, K. B., Driskell, R. R. & Watt, F. M. Assaying proliferation and differentiation capacity of stem cells using disaggregated adult mouse epidermis. _Nat.

Protoc._ 5, 898–911 (2010). Article  CAS  PubMed  Google Scholar  * Hashimshony, T., Wagner, F., Sher, N. & Yanai, I. CEL-Seq: single-cell RNA-Seq by multiplexed linear amplification.

_Cell Rep._ 2, 666–673 (2012). Article  CAS  PubMed  Google Scholar  * Zhou, Y. et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. _Nat.

Commun._ 10, 1523 (2019). Article  ADS  PubMed  PubMed Central  CAS  Google Scholar  * Joost, S. et al. Single-cell transcriptomics reveals that differentiation and spatial signatures shape

epidermal and hair follicle heterogeneity. _Cell Syst._ 3, 221–237 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  Download references ACKNOWLEDGEMENTS We thank Safa Hirbawi

for technical assistance, Norma E. Kidess-Bassir for histological preparation, Sharona Elgavish, and Hadar Benyamini for bioinformatics assistance, Dan Lehmann of the Core Research Facility

at Hebrew University for FACS assistance, and Yaron Fuchs for advice. We thank Elaine Fuchs for the tet-Tcf3 mice. This study was supported by grants from the Israel Science Foundation

(1009/13, 2092/18) and from the DKFZ-MOST Fund (CA-161) (I.B.), by a fellowship from the Israel Council for Higher Education (N.A.), and by the Welfare Foundation for Social & Cultural

Sciences (KIKPE) Greece, and NKUA-SARG Greece, grants No 70/3/8916, 70/3/12128 (V.G.G.). AUTHOR INFORMATION AUTHORS AND AFFILIATIONS * Department of Developmental Biology and Cancer

Research, Institute for Medical Research – Israel-Canada, The Hebrew University–Hadassah Medical School, Jerusalem, Israel Narmen Azazmeh, Benjamin Assouline & Ittai Ben-Porath *

Info-CORE, Bioinformatics Unit of the I-CORE Computation Center at the Hebrew University and Hadassah, Jerusalem, Israel Eitan Winter, Shmuel Ruppo & Yuval Nevo * Department of

Pathology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel Alexander Maly, Karen Meir & Eli Pikarsky * Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY,

USA Agnieszka K. Witkiewicz * Department of Cell and Developmental Biology, Sackler Faculty of Medicine Tel Aviv University, Tel Aviv, Israel Jonathan Cohen & Chen Luxenburg * Department

of Histology and Embryology, Medical School, National and Kapodistrian University of Athens, Athens, Greece Sophia V. Rizou & Vassilis G. Gorgoulis * The Lautenberg Center for

Immunology and Cancer Research, Hebrew University of Jerusalem Faculty of Medicine, Jerusalem, Israel Eli Pikarsky * Faculty Institute for Cancer Sciences, Manchester Academic Health

Sciences Centre, University of Manchester, Manchester, UK Vassilis G. Gorgoulis * Biomedical Research Foundation, Academy of Athens, Athens, Greece Vassilis G. Gorgoulis * Center for New

Biotechnologies and Precision Medicine, Medical School, National and Kapodistrian University of Athens, Athens, Greece Vassilis G. Gorgoulis Authors * Narmen Azazmeh View author publications

You can also search for this author inPubMed Google Scholar * Benjamin Assouline View author publications You can also search for this author inPubMed Google Scholar * Eitan Winter View

author publications You can also search for this author inPubMed Google Scholar * Shmuel Ruppo View author publications You can also search for this author inPubMed Google Scholar * Yuval

Nevo View author publications You can also search for this author inPubMed Google Scholar * Alexander Maly View author publications You can also search for this author inPubMed Google

Scholar * Karen Meir View author publications You can also search for this author inPubMed Google Scholar * Agnieszka K. Witkiewicz View author publications You can also search for this

author inPubMed Google Scholar * Jonathan Cohen View author publications You can also search for this author inPubMed Google Scholar * Sophia V. Rizou View author publications You can also

search for this author inPubMed Google Scholar * Eli Pikarsky View author publications You can also search for this author inPubMed Google Scholar * Chen Luxenburg View author publications

You can also search for this author inPubMed Google Scholar * Vassilis G. Gorgoulis View author publications You can also search for this author inPubMed Google Scholar * Ittai Ben-Porath

View author publications You can also search for this author inPubMed Google Scholar CONTRIBUTIONS N.A. and I.B.-P. designed the study and wrote the manuscript, N.A. conducted the

experiments, B.A. assisted in experimentation, E.W., S.R, and Y.N. conducted bioinformatic analysis, A.M., K.M., E.P., A.K.W., and V.G.G. conducted pathological analyses and collected human

lesion samples. S.V.R. conducted senescence stains and quantification. J.C. and C.L. generated and characterized cultured primary keratinocytes. CORRESPONDING AUTHOR Correspondence to Ittai

Ben-Porath. ETHICS DECLARATIONS COMPETING INTERESTS The authors declare no conflict of interest. ADDITIONAL INFORMATION PEER REVIEW INFORMATION _Nature Communications_ thanks Adam Glick,

Eiji Hara and the other, anonymous, reviewer for their contribution to the peer review of this work. PUBLISHER’S NOTE Springer Nature remains neutral with regard to jurisdictional claims in

published maps and institutional affiliations. SUPPLEMENTARY INFORMATION SUPPLEMENTARY INFORMATION DESCRIPTION OF ADDITIONAL SUPPLEMENTARY FILES SUPPLEMENTARY DATA 1 REPORTING SUMMARY RIGHTS

AND PERMISSIONS OPEN ACCESS This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in

any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The

images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not

included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly

from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Azazmeh, N.,

Assouline, B., Winter, E. _et al._ Chronic expression of p16INK4a in the epidermis induces Wnt-mediated hyperplasia and promotes tumor initiation. _Nat Commun_ 11, 2711 (2020).

https://doi.org/10.1038/s41467-020-16475-3 Download citation * Received: 06 November 2018 * Accepted: 06 May 2020 * Published: 01 June 2020 * DOI: https://doi.org/10.1038/s41467-020-16475-3

SHARE THIS ARTICLE Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy to

clipboard Provided by the Springer Nature SharedIt content-sharing initiative