Inactivation of p16 (inhibitor of cyclin-dependent ... - The FASEB Journal

The FASEB Journal • FJ Express Full-length Article

Inactivation of p16INK4a (inhibitor of cyclin-dependent kinase 4A) immortalizes primary human keratinocytes by maintaining cells in the stem cell compartment Riccardo Maurelli,* Giovanna Zambruno,† Liliana Guerra,* Claudia Abbruzzese,* Goberdhan Dimri,‡ Mara Gellini,* Sergio Bondanza,* and Elena Dellambra*,1 * Laboratory of Tissue Engineering and Cutaneous Physiopathology and †Laboratory of Molecular and Cell Biology, I.D.I.–IRCCS, Istituto Dermopatico dell’Immacolata, Rome, Italy; and ‡Division of Cancer Biology, Department of Medicine, Evanston, Illionois, USA Replicative senescence of human keratinocytes is determined by a progressive decline of clonogenic and dividing cells, and its timing is controlled by clonal evolution (i.e., the transition from stem cells to transient amplifying and postmitotic cells). Progressive increase of p16INK4a (inhibitor of cyclin-dependent kinase 4A) expression has been shown to correlate with keratinocyte clonal evolution. Thus, the aim of our study is to understand whether p16INK4a accumulation is a triggering mechanism of epidermal clonal evolution or a secondary event. We show that inactivation of p16INK4a, by an antisense strategy, allows primary human keratinocytes to escape replicative senescence. Specifically, p16INK4a inactivation alone blocks clonal evolution and maintains keratinocytes in the stem cell compartment. Antisense excision is followed by keratinocyte senescence, confirming that persistent p16INK4a inactivation is required for maintenance of clonal evolution block. Immortalization is accompanied by resumption of B-Cell Specific Moloney murine leukemia virus site 1 (Bmi-1) expression and telomerase activity, hallmarks of tissue regenerative capacity. In turn, Bmi-1 expression is necessary to maintain the impairment of clonal evolution induced by p16INK4a inactivation. Finally, p16INK4a down-regulation in transient amplifying keratinocytes does not affect clonal evolution, and cells undergo senescence. Thus, p16INK4a inactivation appears to selectively prevent clonal conversion in cells endowed with a high proliferative potential. These data indicate that p16INK4a regulates keratinocyte clonal evolution and that inactivation of p16INK4a in epidermal stem cells is necessary for maintaining stemness.—Maurelli, R., Zambruno, G., Guerra, L., Abbruzzese, C., Dimri, G., Gellini, M., Bondanza, S., Dellambra, E. Inactivation of p16INK4a (inhibitor of cyclin-dependent kinase 4A) immortalizes primary human keratinocytes by maintaining cells in the stem cell compartment. FASEB J. 20, E742–E756 (2006) ABSTRACT

Key Words: Bmi-1 䡠 telomerase 䡠 keratinocyte clonal evolution 䡠 keratinocyte senescence Replicative senescence, considered the cellular counterpart of in vivo aging, is characterized by the E742

permanent arrest of cell proliferation (1) and is accompanied by gene expression changes that might be regulated by epigenetic mechanisms. Moreover, senescence is considered a tumor suppressor mechanism, and cells need to acquire multiple genetic alterations to overcome it and become immortal (2). Epidermis is a tissue that undergoes continual and rapid self-renewal. Accomplishment of this process relies on the presence of stem and transient amplifying (TA) cells (3, 4). Stem cells have an unlimited capacity for self-renewal and the ability to generate differentiated progeny (5). When stem cells are committed to differentiate, they enter into a transient state of rapid proliferation, giving rise to TA cells, which represent the largest group of dividing cells. On exhaustion of their proliferative potential, the TA cells withdraw from cell cycle and execute the terminal differentiation program (6). Transition from stem cells to TA cells, named clonal evolution, is a continuous unidirectional process that occurs during natural aging, wound healing, and keratinocyte subcultivation (3, 6 – 8). Keratinocyte replicative senescence occurs when all stem cells have completed their clonal evolution and have generated only terminal TA cells, named paraclones (8 –10). The progressive telomere shortening, occurring with each cell division, is the mitotic clock that regulates the onset of replicative senescence in normal somatic cells (2,11,12). There are strong indications that the p16INK4a/Rb pathway is implicated in transducing the telomere shortening signal in human fibroblasts (13– 18). In epidermal tissue, progressive increase of p16INK4a expression has been shown to correlate with keratinocyte clonal evolution (8,19,20). However, it is not clear whether p16INK4a accumulation is a triggering mechanism of epidermal clonal evolution or a secondary event. In fact, p16INK4a can also mediate many stress responses (e.g., DNA damage, culture shock) by a 1 Correspondence: Laboratory of Tissue Engineering and Cutaneous Physiopathology, IDI, Istituto Dermopatico dell’Immacolata, Via dei Castelli Romani, 83/85, Pomezia (Roma), 00040 Italy. E-Mail: [email protected] doi: 10.1096/fj.05-4480fje

0892-6638/06/0020-0742 © FASEB

telomere-independent mechanism called premature senescence or stress or aberrant signaling induced senescence (STASIS; 12). Moreover, it has been reported that replicative senescence in human epidermal keratinocytes is solely dependent on telomere length, suggesting that the p16INK4a/Rb pathway constitutes a stress-induced senescence mechanism consequent to inadequate culture conditions (21, 22). We have previously demonstrated that epidermal keratinocyte clonal evolution can be blocked by antisense-mediated down-regulation of 14 –3-3␴ and that senescence bypass is accompanied by inhibition of p16INK4a and maintenance of telomerase activity. In our model, p16INK4a inactivation appears to precede impairment of clonal conversion and telomerase resumption, suggesting a direct involvement of p16INK4a accumulation in driving clonal evolution (8). It has also been shown that inactivation of p16INK4a/Rb pathway in human keratinocytes, by either human papilloma virus E7 oncoprotein or Cdk4/Cdk4R24C expression, induces bypass of replicative senescence and cells continue to divide only for a finite number of cell doublings, shortening their telomere length. In these conditions, pRb is directly inactivated, but keratinocyte transfection is accompanied by up-regulation of p16INK4a (22–24). Hence, the objective of our study was to directly inactivate p16INK4a in primary human keratinocytes using an antisense strategy to examine its role in epidermal clonal evolution.

MATERIALS AND METHODS Cell culture, colony-forming efficiency, and cell generations 3T3-J2 cells (a gift from Prof. H. Green, Harvard Medical School, Boston, MA), GP⫹E-86, and GP⫹env Am12 packaging cells were grown as described previously (25). Human keratinocytes were obtained from skin biopsies of healthy donors and cultivated on a feeder layer of lethally irradiated 3T3-J2 cells as described previously (26). For serial propagation, cells were passaged at the stage of subconfluence, until they reached senescence (8). For colony-forming efficiency (CFE) assay, cells (100 – 1,000) from each biopsy and from each cell passage were plated onto 3T3 feeder-layers and cultured as above. CFE, and aborted colony values were calculated as described previously (8,9). The number of cell generations was calculated using the following formula: x ⫽ 3.322 log n/no., where n equals the total number of cells obtained at each passage and no. equals the number of clonogenic cells. Clonogenic cells were calculated from CFE data. Clonal analysis was performed as described previously (9, 27). Briefly, single cells were inoculated onto multiwell plates containing a feeder layer of 3T3 cells. After 7 days of cultivation, clones were identified under an inverted microscope; 1/4 of each clone was transferred into a petri dish. The dish was fixed 9 –12 days later and stained with rhodamine B for the classification of clonal type that was determined by the percentage of terminal colonies formed by the progeny of the founding cell. When 0 –5% of colonies were terminal, the clone was scored as holoclone. When ⬎95% of the colonies p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

were terminal, the clone was classified as paraclone. When ⬎5%, but ⬍95%, of the colonies were terminal, the clone was classified as meroclone. Retroviral and adenoviral-mediated gene transfer L(AS-FL)SN was constructed by cloning a 500 kb fragment containing the full-length human p16INK4a cDNA [obtained by polymerase chain reaction (PCR)] in antisense orientation into the EcoRI/BamHI sites of LBSN-LoxP retroviral vector (a gift from G. Ferrari; see ref 28). L(AS-Exo1a)SN was constructed by generating a specific antisense fragment, corresponding to exon1␣ of p16INK4a gene, using PCR with the following primers: 5⬘-ATTCGCCCTCGAGGCATGGAG-3⬘ and 5⬘-CAGAATTCGGATCCGCCTCCG-3⬘. The fragment was cloned into the EcoRI/XhoI sites of of LBSN-LoxP retroviral vector. L(AS-Bmi-1)SN was constructed by cloning a 980 kb fragment containing the full-length human Bmi-1 cDNA in antisense orientation into the EcoRI/BamHI sites of LBSNLoxP retroviral vector. L(p16)SN [encoding the fusion protein p16INK4a-enhanced green fluorescent protein (EGFP)] was constructed by cloning a 500 kb fragment containing the full-length human p16INK4a in sense orientation into the BamHI/BamHI sites of an EGFP expression retroviral vector (LGSN-LoxP vector). The Am12/L(AS-FL)SN, Am12/L(AS-Exo1a)SN, Am12/ L(AS-Bmi-1)SN, and Am12/L(p16)SN producer cell lines were generated by the transinfection protocol, as described previously (25). Producer cell lines showed a viral titer of 0.5–1 ⫻ 106 colony-forming units/ml. Control amphotropic packaging cell lines were generated as above, using the LXSN-LoxP and LGSN-LoxP retroviral vectors. Keratinocyte infections were carried out as described previously (8). Subconfluent cultures were used for further analysis and serial cultivation. For Cre-dependent reversion, antisense p16INK4a-transduced cells after bypass of senescence were plated on coverslips. After 48 h, Ad-Cre element or Ad-EGFP (a gift from C. Gaetano, IDI, Rome, Italy) adenoviruses were introduced into cells at a multiplicity of infection of 1000 plaque-forming units/cell. Epidermal growth factor (EGF) was added to cultures 1 h after infection. Forty-eight hours after infection, cells were used for further analysis and subcultivated. Transient DNA transfection and luciferase reporter assay The reporter vector containing telomerase promoter (pTERT-Luc 800) was a kind gift of R. Dalla-Favera (Columbia University, New York, NY). Primary human keratinocytes were seeded onto 9.6 cm2 wells (60,000 cells/well) and transfected when 80% confluent. Transfection was performed for 6 h in medium containing 10 ml of Lipofectamine 2000 (Invitrogen Life Technologies) and 2.5 mg of total DNA (0.5 mg of the reporter plasmid, 0.5–1.5 mg of p16INK4a-pcDNA3, 0.5 mg of pCMV-lacZ for the control of transfection efficiency, empty-pcDNA3 to a total 2.5 mg). Transfection assays were performed in triplicate. Transfection medium was then removed, and keratinocytes were allowed to recover in fresh medium for 48 h. Cultures were harvested with lysis buffer CCLR (Promega). Luciferase activity was determined and normalized to ␤-galactosidase activity. Organotypic cultures, immunohistochemistry, immunocytochemistry, and Western analysis Dermal equivalents were prepared as described previously (29). Briefly, to obtain a collagen gel, calf skin type I collagen E743

(Symatese Biomateriaux) was mixed with 1 ⫻ 106 young foreskin fibroblasts in Eagle’s minimal essential medium (MEM) supplemented with 10% fetal calf serum. After 3 days, transduced keratinocytes were plated at a cell density of 5 ⫻ 104 cells/ collagen gel. Keratinocyte cultures were performed in submerged conditions for 7 days; then cultures were raised at the air-liquid interface for further 7 days to obtain a fully differentiated epidermis. Hematoxylin and eosin was performed on formalin-fixed, paraffin-embedded specimens. Immunohistochemistry was carried out as described previously (8), using the following antibodies: anti-PCNA (Santa Cruz Biotechnology), anti-LEKTI (a gift from M. D’Alessio, IDI, Rome, Italy), and anti-K10 (Monosan). For immunocytochemistry, keratinocyte colonies were fixed and permeabilized in 3% formaldehyde-0.2% Triton X-100 in PBS, for 10 min. Slides were incubated with antiBmi-1 antibody (Ab; Upstate) for 1 h and 3,3⬘-diaminobenzidine stained after incubation with a peroxidase-conjugated secondary Ab (DAKO). For immunoblots, subconfluent keratinocytes were extracted on ice with lysis radioimmunoprecipitation assay buffer and equal amounts of samples (50 ␮g) were electrophoresed on 12.5% SDS-polyacrylamide gels. Western blot was performed as described (8), using the following antibodies: antip16INK4a (N20), antip53 (DO-1), antip63 (4A4), antiMad1 (C-19), and antic-Myc (N-262) all from Santa Cruz Biotechnology; anti-Cdk4 and anti-Bmi-1 from Upstate; antipRb from BD Pharmingen; antip14ARF from Oncogene; and antip21Waf1 was a kind gift from Kristian Helin (IEO, Milan, Italy).

Northern analysis, in situ hybridization, telomeric repeat amplification protocol assay, and telomere restriction fragment assay For Northern analysis, cellular RNA was extracted with RNeasy Mini Kit (QIAGEN; 10 ␮g of total RNA were sizefractionated through 1% agarose/formaldehyde gels and transferred to nylon membrane (Hybond N⫹, Amersham). Hybridization was performed as described previously (8). For in situ hybridization on keratinocyte colonies, the fixing and the hybridization procedures, with digoxigeninlabeled Neo antisense-riboprobe (1 ng/ml) or a digoxigeninlabeled Bmi-1 specific oligonucleotide probe, were performed as described previously (30). Detection of labeled probes was performed using the DIG RNA labeling kit (Roche). Telomerase activity was detected by telomeric repeat amplification protocol (TRAP) assay using Telo TAGGG Telomerase PCR ELISA and Biotin Luminescent Detection kit (Roche). An internal telomerase assay standard (ITAS) of 150 bp was used (a gift from P. Lacal, IDI, Rome, Italy). Telomere length was determined by telomere restriction fragment (TRF) assay using the North2South Chemiluminescent Hybridization and Detection Kit (Pierce), following the manufacturer’s instruction. After scanning of the signal smears with a GS 670 Imagine densitometer (Bio-Rad), the telomere length was calculated using the following formula: ⌺(optical density (OD)ixLi)/⌺(ODi), where ODi is the densitometer output and Li is the length of the DNA at position i. Quantitative reverse transcriptase-PCR RNA was extracted from cells using the Trizol kit (Invitrogen). Contaminating DNA was eliminated by DNase (Invitrogen) treatments, and RNA was purified using RNeasy kit (QIAGEN). An aliquot (5 ␮g) of total RNA was reverse transcribed by using an oligo(dT) primer. For the thermal E744

Vol. 20

July 2006

cycle reactions, cDNA, was amplified (ABI PRISM 7000; Perkin-Elmer Applied Biosystems) using the SYBR reverse transcriptase (RT)-PCR kit (Applied Biosystems) under the following conditions: 2 min at 50°C and 10 min at 95°C, followed by 45 cycles of 95°C for 15 s and 60°C for 1 min. The controls consisted of amplifications without reverse transcription and reactions without the addition of a cDNA template. The authenticity and sizes of the PCR products were confirmed using a melting curve analysis (using software provided by Perkin-Elmer). mRNA levels were normalized using the GAPDH gene as housekeeping gene. The following primer sets were used: hTERT mRNA, sense (5⬘-GGGAAGCATGCCAAGCTCT-3⬘) and antisense (5⬘-CACGCTCATCTTCCACGTCA-3⬘); and GAPDH mRNA, sense (5⬘-GAAGGTGAAGGTCGGAGTC-3⬘) and antisense (5⬘-GAAGATGGTGATGGGATTTC-3⬘). Primers were chosen using software PrimerExpress2 (Perkin-Elmer Applied Biosystems).

RESULTS p16INK4a inactivation and keratinocyte replicative senescence To investigate the link between p16INK4a accumulation and replicative senescence, we inactivated p16INK4a in primary human keratinocytes. Infections with a defective retrovirus, carrying a full-length human cDNA of p16INK4a in antisense orientation (Fig. 1A), were performed on four different strains (P1, P4, K54, and K79) of primary keratinocytes obtained from healthy donors (Fig. 1B–E). Clonogenic cells were stably transduced with an efficiency of 80 –100%. To investigate the proliferative capacity of antisense p16INK4a-transduced keratinocytes, cells were serially cultivated and the number of cell generations was calculated. As shown in Fig. 1B–E (blue squares), the four strains transduced with an empty vector underwent 128, 188, 140, and 121 cell generations, respectively. Unexpectedly, antisense p16INK4a-transduced keratinocytes bypassed replicative senescence and continued to divide at a rate comparable with that of young keratinocytes (Fig. 1B–E, green circles). All the antisense p16INK4a-transduced cells are still in culture and, to date, have undergone 530, 553, 524, and 499 cell doublings, respectively, and have been serially cultivated for 521, 544, 422, and 504 days, respectively. In addition, these cultures did not show any sign of growth crisis, the so-called “slow growth phase” (31). Northern blot analysis, performed on total RNA extracted from transduced cells at different passages (Fig. 1F, AS-FL lanes), revealed that antisense transcripts were detectable during the entire life span in all the four strains. As expected, Western blot analysis showed that p16INK4a was undetectable in empty vector and antisense p16INK4a-transduced cells at early passages (Fig. 1G, lanes V 28 and AS-FL 27). Thereafter, p16INK4a expression progressively increased during empty vector-transduced keratinocyte subcultivation (Fig. 1G, V lanes). On the contrary, a faint p16INK4a expression was transiently visualized in antisense p16INK4a-transduced

The FASEB Journal

MAURELLI ET AL.

cells (Fig. 1G, lane AS-FL 55). Of note, p16INK4a became again undetectable already before bypass of senescence and p16INK4a inactivation was maintained during subsequent subcultivation (Fig. 1G, AS-FL lanes). Similar results were obtained in all the four strains. The INK4a/ARF locus (see Fig. 1A) encodes, in overlapping reading frames, two proteins, p16INK4a and p14ARF, regulating the pRb and the p53 pathways, respectively (32). To determine whether extension of keratinocyte life span was specifically dependent on p16INK4a inactivation, we repeated the experiments using an antisense retroviral construct that targets exon

1␣ of the INK4a/ARF locus (see Fig. 1A), and thus specific for p16INK4a. Infections with defective retrovirus carrying a 100 bp fragment, corresponding to exon 1␣, in antisense orientation (named Exo1a) were performed on three primary keratinocyte strains (Fig. 1C-E) with a transduction efficiency of 80 –100%. Antisense Exo1a-transduction gave results comparable with those obtained with full-length antisense-p16INK4a in all the transduced strains: 1) Exo1a-transduced cells bypassed replicative senescence (Fig. 1C-E, red circles); 2) antisense transcripts were highly expressed during serial cultivation (Fig. 1F, AS-Exo1a lanes); and 3) p16INK4a levels were undetectable already before bypass of senescence and p16INK4a inactivation was maintained during subcultivation (Fig. 1H, AS-Exo1a lanes). Thus, p16INK4a inactivation appears able to induce replicative senescence bypass in primary human keratinocytes. All the subsequent experiments were carried out with the full-length and the Exo1a antisense p16INK4a retroviral vectors with similar results, and hereafter both the antisense p16INK4a transduced keratinocytes will be generically termed “antisense p16INK4a-transduced cells.” p16INK4a inactivation in TA keratinocytes Western blot results revealed a slight increase of p16INK4a expression in antisense p16INK4a-transduced cells before bypass of replicative senescence (Fig. 1G,

Figure 1. p16INK4a inactivation and keratinocyte replicative senescence. A) Schematic map of the INK4a/ARF locus (top). Open reading frames of p16INK4a (red boxes) and p14ARF (yellow boxes) are indicated. Double lines indicate the position of antisense-p16INK4a full-length (green), directed against INK4a/ARF locus, and antisense Exo1a (red), directed against exon 1␣ of INK4a/ARF locus. Schematic map of LXSN-LoxP provirus (bottom). Green boxes indicate viral LTR, blue box is transgene (antisense p16INK4a full-length or Exo1a), pink box is neomycin phosphotransferase (NeoR) cDNA, and arrowhead-shaped pink box is simian virus 40 early promoter (SV40). 3⬘ LTR contains a LoxP site. Primary P1 (B), P4 (C), K54 (D), and K79 (E) human keratinocytes, transduced with empty LXSN-LoxP vector (blue squares) and with antisense p16INK4a cDNA (green circles), were serially cultivated. Primary P4 (C), K54 (D), and K79 (E) human keratinocytes were also transduced with antisense Exo1a (red circles) and serially cultivated. Number of cell doublings was calculated as described in Materials and Methods. Cumulative number of cell generations per passage was plotted against total time in culture. Arrows indicate keratinocyte infection. F) Antisense transcript expression was assessed by Northern blot using RNA obtained from antisense p16INK4a full-length (AS-FL) and Exo1a (AS-Exo1a) transduced keratinocytes at different passages (indicated by numbers of cell generations) after infection. Empty vector transduced cells (V) were used as negative control. GA3PDH served as a loading control. G, H) p16INK4a expression was assessed by Western blot using cell extracts prepared from empty vector (V-lanes), antisense p16INK4a full-length (AS-lanes) and Exo1a (AS-Exo1a-lanes) transduced cells at different passages (indicated by numbers of cell generations) after infection. GA3PDH served as a loading control. p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

E745

pected, high levels of p16INK4a expression were detected in empty-vector transduced TA cells at the time of transduction (Fig. 2D, lanes V 101). p16INK4a expression resulted down-regulated in antisense p16INK4atransduced TA cells during subcultivation (Fig. 2D, lanes AS 127 and 142). These results suggest that bypass of senescence can be only achieved when p16INK4a is inactivated in human primary keratinocytes endowed with high proliferative potential. Bypass of senescence, clonal evolution, and clonal analysis

Figure 2. Antisense p16INK4a-transduction in TA keratinocytes. Primary P1 (A) and K79 (B) human keratinocytes (light blue squares) were serially cultivated and transduced with empty vector (blue squares), antisense p16INK4a full-length (green circles), or Exo1a (red circles) when approached senescence (arrows). Clonal analysis performed on P1 keratinocyte strain at the moment of infection (93 cd) showed 0% holoclones, 30% meroclones, and 70% paraclones. After transduction, keratinocytes were serially cultivated and number of cell doublings was calculated. C) Antisense transcript expression was assessed by Northern blot using RNA obtained from antisense p16INK4a full-length (AS-FL) transduced TA cells at different passages (indicated by numbers of cell generations) after infection. Empty vector transduced cells (V) were used as negative control. GA3PDH served as a loading control. D) p16INK4a expression was assessed by Western blot using cell extracts prepared from empty vector (V) and antisense p16INK4a (AS-FL) transduced TA cells at different passages (indicated by numbers of cell generations) after infection. GA3PDH served as a loading control.

lane AS-FL 55, and Fig. 1H, lanes AS-Exo1a 75 and 100) that correlates with a very slight decrease of transgene expression in Northern blot analysis (Fig. 1F, lanes AS-FL 55 and AS-Exo1a 75 and 100). This finding could be ascribed to senescence of untransduced or poorly transduced cells but also to the senescence of regularly transduced TA cells. To investigate whether p16INK4a inactivation may contribute to the bypass of senescence in both stem and TA keratinocytes, we transduced keratinocyte cultures approaching senescence, i.e., enriched in TA cells and depleted in stem cells (see Fig. 2). P1 and K79 keratinocytes, which had been shown to undergo 130 and 123 cell generations before senescence (Fig. 2A and B, light blue squares), were transduced at 93 and 95 cell generations, respectively, and then serially cultivated. As shown in Fig. 2, empty vector-transduced P1 and K79-TA cells underwent further 40 and 30 cell generations, respectively (A and B, blue squares). Of note, also antisense p16INK4a-transduced P1 and K79-TA keratinocytes did not bypass replicative senescence, stopping growth after 55 and 41 cell generations, respectively (Fig. 2A and B, green and red circles). Northern blot analysis showed that antisense transcripts were highly expressed during serial cultivation (Fig. 2C). As exE746

Vol. 20

July 2006

To investigate how p16INK4a inactivation affects keratinocyte clonal evolution, we analyzed CFE and clonal evolution of empty vector and antisense p16INK4a-transduced keratinocytes. As shown in Fig. 3A (V); B, blue bars and, empty vector-infected cells showed a progressive decrease of their CFE during serial cultivation and a progressive increase of the percentage of paraclones (Fig. 3C, blue line). In contrast, after an initial decrease of CFE (Fig. 3A, AS-Exo1a 75 cd; B, red bars), which can be due to senescence of both untransduced/poorly transduced cells and regularly transduced TA cells, antisense p16INK4a-transduced keratinocytes showed a progressive increase of CFE. CFE values remained then high during serial cultivation (Fig. 3A, AS-Exo1a 139 – 365 cd; B, red bars), with a percentage of aborted colonies constantly lower than 3% (Fig. 3C, red line). Already before bypass of senescence, antisense

Figure 3. Clonogenic potential of antisense p16INK4a transduced cells. CFE assays were performed at each cell passage of keratinocytes transduced with empty vector (V) or antisense p16INK4a (AS-Exo1a). A) CFE performed at selected passages indicated by numbers of cell doublings. B) CFE values of keratinocytes transduced with antisense p16INK4a (red bars) or with empty vector (blue bars) were expressed as ratio of total number of colonies on number of inoculated cells and were plotted against passages. C) Aborted colony values (% of paraclones) of keratinocytes transduced with antisense p16INK4a (red line) or with empty vector (blue line) were expressed as ratio of aborted colonies on total number of colonies and were plotted against passages.

The FASEB Journal

MAURELLI ET AL.

p16INK4a-transduced keratinocyte cultures formed large and smooth colonies (Fig. 3A, AS-Exo1a 139 cd). Of note, impairment of clonal evolution was observed in antisense p16INK4a-transduced cells several passages before the onset of replicative senescence of empty-vector cells (Fig. 3C, compare blue and red lines). As shown in Fig. 4A, the expression of p63, a marker of high proliferative potential in epithelial cells (27), decreased during empty-vector subcultivation (V lanes). Instead, in antisense p16INK4a-transduced keratinocytes, p63 expression was high in early passages, gradually decreased, then raised again already before bypass of senescence, and remained elevated thereafter (Fig. 4A, AS-Exo1a lanes). This finding suggests a recruitment of cells endowed with high proliferative potential in the p16INK4a-transduced cultures. Colony-forming human epidermal cells are heterogeneous in their growth capacity. Once a clone has been derived from a single cell, its growth potential can be estimated from the resulting colony types. The holoclone is the smallest colony-founding cell, has the highest proliferative capacity, and is considered the surface epithelial stem cell. Holoclones form large colonies with a very smooth and regular perimeter consisting of migrating involucrin negative cells. Meroclones and paraclones are considered young and old TA cells, respectively. In particular, the paraclone is the largest colony-forming keratinocyte and generates aborted colonies containing only terminal cells (3, 9, 10, 25, 27, 33). To assess whether antisense p16INK4a-transduced keratinocytes were maintained in the stem cell compartment, we examined the percentage of stem and TA cells in our transduced cultures. Clonal analysis performed on empty-vector transduced keratinocytes at

Figure 4. Clonal analysis of antisense p16INK4a transduced cells. A) p63 expression was assessed by Western blot using cell extracts prepared from empty vector (V-lanes) and Exo1a (AS-Exo1a-lanes) transduced cells at different passages (indicated by numbers of cell generations) after infection. GA3PDH served as a loading control. B) Clonal analysis, performed on empty vector transduced keratinocytes at early passage (V) and antisense p16INK4a Exo1a transduced cells after bypass of senescence (AS-Exo1a), was expressed as percentage of holoclones (H), meroclones (M), and paraclones (P). p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

early passage (Fig. 4B, V) showed that holoclones represented 10% of total clonogenic cells, meroclones were 80%, and paraclones were 10%. In contrast, epidermal cultures generated from antisense p16INK4atransduced cells, after bypass of senescence, were enriched in holoclones, which represented 50% of total cells. In addition, these cultures were markedly depleted of paraclones (1.25%; Fig. 4B, AS-Exo1a). These findings demonstrate that inactivation of p16INK4a allows primary human keratinocytes to bypass replicative senescence by blocking clonal evolution, and maintains transduced keratinocytes in the stem cell compartment. Reversion of the p16INK4a-dependent phenotype We next investigated whether continued inactivation of p16INK4a is necessary for blocking clonal conversion. A retroviral vector, containing a target site for the Cre recombinase, was used for keratinocyte transduction (see Fig. 1A, LoxP site). On integration of the retrovirus into the genome, the LoxP site is duplicated and the genes carried by the virus are flanked, on both sides, by LoxP sites. Subsequent expression of Cre recombinase causes excision at these LoxP sites, leading to construct removal from the cell genome and to restoration of target gene expression in the cell. After bypass of senescence, antisense p16INK4a-transduced cells were infected with a Cre-expressing adenovirus to ablate antisense expression. Antisense p16INK4a-transduced cells were also infected with an EGFP-expressing adenovirus as a control. To assess deletion of the transgene, we performed an in situ hybridization, using a Neo probe that showed an infection efficiency of 30%. Note that Neo-positive uninfected cells (Fig. 5A, Ad-Cre, red arrows) were small and displayed a regular morphology, while Neonegative infected cells were enlarged and flattened (Fig. 5A, Ad-Cre, black arrows), similar to senescent cells. The transgene deletion was accompanied by re-expression of p16INK4a, indicating that the p16INK4a pathway was not altered in antisense p16INK4a-transduced cells (data not shown). Control EGFP-expressing adenovirus did not affect cell morphology (Fig. 5A, control) and p16INK4a expression (data not shown). The first amplification passage of both AdCre and AdEGFP cells showed a reduction of CFE only in AdCre-infected cultures (Fig. 5B). This finding strongly suggests that in Ad-Cre element treated cultures, Neonegative cells (in which p16INK4a function has been restored) failed to form colonies or generated paraclones, whereas Neo-positive uninfected cells generated normal colonies. Subsequently, AdCre-treated cultures only showed Neo-positive uninfected cells, and their CFE was comparable with those of control culture (data not shown). Thus, persistent inactivation of p16INK4a is required to block clonal conversion. E747

EGF, hydrocortisone, and cholera toxin for optimal growth, and any depletion of these components had a similar effect on both cultures (data not shown). Karyotype performed on antisense p16INK4a-transduced keratinocytes after bypass of senescence revealed that cells had the normal complement of 46 chromosomes (data not shown). It is worth noting, however, that whereas P4-AS-FL, P4-AS-Exo1a, K54-AS-FL, K54AS-Exo1a, and K79-AS-Exo1a cells remained diploid even over 500 cell doublings, K79-AS-FL and P1-AS-FL cells developed ⬵20% of hyperploid metaphases (48 and 47 chromosome number, respectively) at later cell passages (400 and 500 cell doublings, respectively). Control and antisense p16INK4a-transduced keratinocyte tumorigenicity was evaluated by an in vitro assay. None of the cell strains formed colonies after 21 days culture in soft agar, as did the A431 cell line (positive control), indicating that antisense p16INK4a-transduced keratinocyte growth is anchorage dependent (data not shown). Bypass of senescence and differentiative capacity Figure 5. Antisense excision in p16INK4a transduced cells. After bypass of senescence, antisense p16INK4a-transduced cells were infected with a Cre-expressing adenovirus to ablate antisense expression and with an EGFP-expressing adenovirus as a control. A) In situ hybridization was performed using a digoxigenin-labeled Neo antisense-riboprobe on Ad-EGFP (Control) and Ad-Cre element infected keratinocytes, seeded on coverslips. B) CFE assays performed on Ad-EGFP (Control) and AdCre infected keratinocytes.

Growth requirements, karyotype, and tumorigenicity after bypass of senescence To verify whether inactivation of p16INK4a impaired control mechanisms of keratinocyte growth, we performed different assays for keratinocyte growth factor requirements. Specifically, we assessed that both control cells at early passages and antisense p16INK4a-transduced keratinocytes after bypass of senescence needed FCS,

We also evaluated whether p16INK4a inactivation interferes with the coordinated gene expression needed to form a fully differentiated epidermis in skin organotypic cultures. Histological analysis of in vitro reconstituted skin equivalents from empty vector and antisense p16INK4a-transduced keratinocytes showed that antisense p16INK4a-transduced cells were able to stratify into basal and suprabasal layers (Fig. 6, compare A and B). The proliferation marker PCNA was detected in most basal cells and in the first suprabasal cell layers of control organotypic cultures (Fig. 6C). A higher number of suprabasal keratinocytes expressed PCNA in antisense p16INK4a-transduced skin equivalents (Fig. 6D). The differentiation marker cytokeratin 10 (K10) was regularly detected in suprabasal cell layers of the in vitro reconstituted skin equivalents (Fig. 6E–F). However, it appeared more weakly expressed in antisense p16INK4a-transduced keratinocytes (Fig. 6F) than in

Figure 6. Immunohistochemical analysis of organotypic cultures. Organotypic cultures were prepared from empty vector transduced keratinocytes at early passage (A, C, E, G) and antisense p16INK4a Exo1a-transduced cells after bypass of senescence (B, D, F, H). Paraffinembedded sections were analyzed for morphological characteristics (hematoxylin and eosin staining, A, B) and immunostained with antibodies against PCNA (C, D), K10 (E, F), and LEKTI (G, H; magnification, ⫻40).

E748

Vol. 20

July 2006

The FASEB Journal

MAURELLI ET AL.

control cultures (Fig. 6E). In fact, it has been reported that K10 is involved in the control of keratinocyte proliferation (34, 35) and is drastically reduced in skin tumors (36). The serine protease inhibitor LEKTI was normally expressed in keratinocytes of the granular and uppermost spinous layers of organotypic cultures (Fig. 6G and H). However, in antisense p16INK4a-transduced keratinocytes LEKTI expression was extended to several layers of the stratum spinosum (Fig. 6 H), again in keeping with a highly proliferative cell status (37). All these findings indicate that antisense p16INK4atransduced cells maintain their ability to reconstitute a differentiated human epidermis, although signs of hyperproliferative activity are clearly evident. p16INK4a inactivation and pRb pathway It is known that p16INK4a is one of the key components of the pRb pathway. p16INK4a is an inhibitor of Cdk4/ cyclin D complex and maintains pRb in its hypophosphorylated active state, blocking the entry of proliferating cells into S phase (38). We analyzed the effects of p16INK4a inactivation on pRb pathway by Western blot. Cdk4 expression did not significantly vary during empty vector and antisense p16INK4a-transduced serial cultivation (Fig. 7A and B). In empty-vector and antisense p16INK4a-transduced keratinocytes, hypo- and hyperphosphorylated forms of pRb were highly expressed at early passages. pRb expression decreased with serial cultivation of empty-vector cells and the hypophosphorylated form became predominant close to senescence (Fig. 7A and B, V lanes). In contrast, pRb expression slightly decreased in antisense p16INK4a-transduced cells approaching to bypass of senescence. Then, pRb promptly increased and showed a phosphorylation status similar to that of young cells (Fig. 7A and B, AS-FL and AS-Exo1a lanes). Thus, inactivation of p16INK4a inhibits the pRb pathway, maintaining the hyperphosphorylated state of pRb. p16INK4a inactivation and p53 pathway The second product of the INK4a/ARF locus, p14ARF, inhibits MDM2-mediated degradation of p53 (39). To exclude any interference of transgenes on the ARF gene product, we assessed the p14ARF/p53/p21Waf1 expression in antisense p16INK4a-transduced keratinocytes, during subcultivation. p14ARF was barely detectable during empty vectortransduced keratinocyte subcultivation (Fig. 7C and D, V lanes) and in antisense p16INK4a-transduced cells at early passages (Fig. 7C and D, lanes AS-FL 27 and AS-Exo1a 35). Interestingly, p14ARF expression increased in antisense p16INK4a-transduced keratinocytes approaching to bypass of senescence, when p16INK4a was already undetectable (Fig. 7C and D, lanes AS-FL and AS-Exo1a). As shown in Fig. 7, p53 was clearly detected in empty vector-transduced cells at early cell passages (C and D, V 28 and 34), and its expression decreased during serial subcultivation (Fig. 7C and D, V p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

Figure 7. p16INK4a inactivation and pRb and p53 pathways. Cdk4, pRB (A, B), p14ARF, p53, p21Waf1 (C, D) expression was assessed by Western blot using cell extracts prepared from empty vector (V-lanes), antisense p16INK4a full-length (A, C, AS-lanes) and Exo1a (B, D, AS-Exo1a-lanes) transduced cells at different passages (indicated by numbers of cell generations) after infection. HeLa and A431 cells were used as positive control of p14ARF and p53 expression, respectively. GA3PDH served as a loading control. Note hypo- (lower line) and hyperphosphorylated (upper line) forms of pRB.

lanes). In contrast, p53 expression sharply increased in antisense p16INK4a-transduced keratinocytes at bypass of senescence, reaching levels similar or higher than those of the early passage cells (Fig. 7 C and D, AS-FL and AS-Exo1a lanes). The expression of p21Waf1, a p53 transcriptional target, increased in antisense p16INK4atransduced keratinocytes after bypass of senescence. Of note, p21Waf1 levels at the highest passages returned similar to those of early passage cells (Fig. 7C and D, AS and AS-Exo1a lanes). These findings demonstrate that both full-length and Exo1a antisense exclusively inactivate p16INK4a and induce an up-regulation of the p14ARF/p53/p21Waf1 pathway, suggesting that p16INK4a does play the primary role in enforcing replicative senescence in primary human keratinocytes. E749

p16INK4a and telomerase activity Human somatic cells are believed to be immortalized as long as their telomeres are maintained through activation of telomerase or alternative mechanisms (ALT; 40 – 42). As expected, telomerase activity was detected in

empty vector (Fig. 8 A, lane V 34)- and antisense p16INK4a-transduced cells (Fig. 8A, lane AS-Exo1a 35) at early passages and became barely detectable in empty vector transduced cells approaching replicative senescence (Fig. 8A, lane V 174). A slight reduction of telomerase activity was seen in antisense p16INK4a-transduced keratinocytes approaching to bypass of senescence in keeping with CFE analysis and molecular patterns. Then, telomerase activity was resumed (Fig. 8A, AS lanes). Telomere length of empty-vector transduced keratinocytes was 9.3 kb at early passages and 5.7 kb near senescence. After transduction, antisense p16INK4a-transduced keratinocytes showed an average telomere length of 9 kb. Interestingly, before bypass of senescence and in concomitance with CFE increase and paraclone percent decrease (see Fig. 3B, C), we identified two cell subpopulations differing in telomere length: one displaying a decrease of telomere length from 8.9 to 4.3 kb, similar to control cells, and the other one with telomeres that were stabilized at ⬇14 –15.5 kb (Fig. 8B and see Supplemental Fig. 1). Thus, resumption of telomerase activity was accompanied by elongation and stabilization of the telomeres. The shortening of telomere length in a subpopulation of transduced cells is in keeping with a different fate of transduced TA cells as compared with cells endowed with high proliferative potential. It has been reported that p16INK4a regulates telomerase activity by inhibiting the function of Sp1 in human malignant gliomas (43). To evaluate a possible effect of p16INK4a on telomerase activity also in normal human keratinocytes, these cells were cotransfected with a reporter plasmid, containing luciferase driven by telomerase promoter, and a plasmid containing p16INK4a. Luciferase activity was measured 48 h later. Increasing the amount of p16INK4a resulted in a doserelated inhibition of telomerase promoter activity (Fig.

Figure 8. Telomerase activity and Bmi-1 expression. A) TRAP assay was performed on cells transduced with empty vector (V) and antisense p16INK4a-transduced cells (AS-Exo1a lanes) at different passages (indicated by numbers of cell generations) after infection; an ITAS of 150 bp was used. B) TRF assay performed on cells transduced with empty vector (V lanes) and antisense p16INK4a-transduced cells (AS-Exo1a) at different passages (indicated by numbers of cell generations) after infection. C) A reporter plasmid containing telomerase promoter and a plasmid containing p16INK4a were cotransfected, along with a standard amount of cytomeglovirus-lacZ control plasmid. Cells were harvested 48 h after transfection and assayed for luciferase and ␤-galactosidase. Luciferase activity was normalized to corresponding ␤-galactosidase activity. Promoter activities were calculated using empty vector control (column 1) as reference. The ratio promoter: p16INK4a was 1:1 (column 2), 1:2 (column 3), and 1:3 (column 4). Transfection assays were performed in triplicate. D) c-Myc, Mad1, Bmi-1 expression was assessed by Western blot using cell extracts prepared from empty vector (V-lanes) and Exo1a (AS-Exo1a-lanes) transduced cells at different passages (indicated by numbers of cell generations) after infection. HeLa cells were used as positive control of Bmi-1 expression. GA3PDH served as a loading control. E750

Vol. 20

July 2006

The FASEB Journal

MAURELLI ET AL.

8 C). Thus, p16INK4a levels are able to modulate telomerase promoter activity in normal human keratinocytes. To further investigate the link between inactivation of p16INK4a and resumption of telomerase activity in antisense p16INK4a-transduced cells, we assessed the expression of telomerase regulators. c-Myc is widely known to activate the transcription of the telomerase catalytic subunit (hTERT), whereas Mad1, a c-Myc antagonist, represses the same gene (44, 45). No modification of c-Myc expression was detected during empty vector (Fig. 8D, V lanes) and antisense p16INK4a-transduced cell (Fig. 8D, AS-Exo1a lanes) subcultivation. Instead, Mad1 was overexpressed in antisense p16INK4a-transduced cells after bypass of senescence (Fig. 8D, AS-Exo1a lanes). Cell type specific induction of telomerase by Bmi-1, a repressor of the INK4a/ARF locus and a marker of adult stem cells in some tissues, has been reported (46). As shown in Fig. 8D, Bmi-1 was detected in empty vector (lane V 34) and antisense p16INK4a-transduced cells (lane AS 35) at early passages. Interestingly, Bmi-1 expression decreased during senescence of empty vector-transduced keratinocytes (Fig. 8D, V lanes). In contrast, Bmi-1 expression slightly decreased in antisense p16INK4a-transduced keratinocytes approaching bypass of senescence but promptly returned to normal levels thereafter (Fig. 8D, AS-Exo1a lanes). We have shown that antisense p16INK4a transduction of TA keratinocytes does not affect clonal evolution and that transduced TA cells undergo senescence like normal controls. As shown in Fig. 9A, Bmi-1 was detected in empty vector (lane V 101) and antisense p16INK4atransduced TA cells (lane AS 100). Of note, Bmi-1 expression continued to decrease during subcultivation of antisense p16INK4a-transduced TA keratinocytes (Fig. 9A, lanes AS-FL 127 and 142). Telomerase activity was barely detectable in empty vector (Fig. 9B, lane V 101) and antisense p16INK4a (Fig. 9B, lane AS-FL 100) transduced TA cells compared with young keratinocytes (Fig. 9B, lane C 34). Telomerase activity reduction became more evident in antisense p16INK4a-transduced TA keratinocytes during subcultivation (Fig. 9B, lanes AS-FL 127 and 142). Thus, p16INK4a inactivation-induced bypass of senescence is accompanied by resumption of Bmi-1 and by maintenance of telomerase activity, hallmarks of tissue regenerative capacity. On the contrary, down-regulation of p16INK4a in TA cells is unable to resume Bmi-1 and telomerase activity at levels comparable with those of young keratinocytes. These data reinforce the idea that, once p16INK4a is up-regulated during keratinocyte clonal evolution, cells acquire modifications that cannot be reverted by p16INK4a down-regulation. Modulation of Bmi-1 expression and keratinocyte clonal conversion To assess the effects of p16INK4a on Bmi-1 expression, we infected primary keratinocytes with a defective retrovirus carrying a recombinant cDNA encoding the p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

Figure 9. Modulation of Bmi-1 expression and clonal conversion. A) Bmi-1 expression was assessed by Western blot using cell extracts prepared from empty vector (V lane) and antisense p16INK4a (AS-FL-lanes) transduced TA cells at different passages (indicated by numbers of cell generations) after infection. HeLa cells were used as positive control. GA3PDH served as a loading control. B) TRAP assay was performed on untransduced cells (C lane), empty vector (V lane), and antisense p16INK4a (AS-FL-lanes) transduced TA cells at different passages (indicated by numbers of cell generations). C) p16INK4a and Bmi-1 expression was assessed by Western blot using cell extracts prepared from empty vector (V lane) and sense p16INK4a (S lane) transduced cells. GA3PDH served as a loading control. D) TRAP assay was performed on empty vector (V lane) and sense p16INK4a (S lane) transduced cells. E) Bmi-1, p16INK4a, Mad1 expression was assessed by Western blot using cell extracts prepared from antisense p16INK4a -transduced keratinocytes infected with empty vector (ASp16-V lane) and antisense Bmi-1 (ASp16ASBmi-1 lane). HeLa cells were used as positive control of p16INK4a expression. GA3PDH served as a loading control. F) TRAP assay was performed on antisense p16INK4a -transduced keratinocytes infected with empty vector (ASp16-V lane) and antisense Bmi-1 (ASp16-ASBmi-1 lane). In all TRAP assays, an internal telomerase assay standard (ITAS) of 150bp was used.

fusion protein p16INK4a-EGFP. Cells were transduced with an efficiency of 90%. After transduction, keratinocytes immediately underwent senescence. Fluorescence screening, performed on clonal analysis of p16INK4a-E GFP-transduced keratinocytes, showed that EGFP⫹ cells were able to generate only paraclones (data not shown). Western blot and TRAP assay performed on primary keratinocytes transduced with empty vector and p16INK4a-E GFP showed that upregulation of p16INK4a was accompanied by down-regulation of Bmi-1 (Fig. 9C), a decrease of telomerase activity (Fig. 9D) and hTERT expression (see Supplemental Fig. 2, column 1 and 2). Taken together these data demonstrate that modulation of p16INK4a levels affects both keratinocyte clonal evolution and Bmi-1 expression and telomerase activity. To determine the subcellular localization of Bmi-1, we stained normal human keratinocyte colonies using E751

an anti-Bmi-1 Ab. As shown in Supplemental Fig. 3, Bmi-1 showed a prominent nuclear staining at the rim of the colony (A, black arrows), where proliferating keratinocytes are localized, and a cytoplasmic staining in the differentiated cells of the central part of the colony (A, red arrows). Interestingly, in antisense p16INK4a -transduced keratinocytes Bmi-1 expression was mostly nuclear, even in the center of the colony (B, black arrows). We next investigated whether Bmi-1 expression is necessary for maintaining the antisense p16INK4a-mediated impairment of clonal conversion. We infected antisense p16INK4a -transduced keratinocytes, after bypass of senescence, with a defective retrovirus carrying the full-length human cDNA of Bmi-1 in antisense orientation. Cells were transduced with an efficiency of 40%, as evaluated by in situ hybridization (not shown). After transduction, we observed an increased percentage of enlarged and flattened colonies in antisense p16INK4a-antisense Bmi-1-transduced keratinocyte cultures. Immunocytochemistry analysis confirmed that these colonies were not able to express Bmi-1 (Supplemental Fig. 3C). Western blot and TRAP assay, performed on antisense p16INK4atransduced keratinocytes infected with antisense Bmi-1 or an empty control vector, showed that down-regulation of Bmi-1 resulted in a strong increase of Mad1 (Fig. 9E) and a decrease of telomerase activity (Fig. 9F), in the absence of p16INK4a up-regulation (Fig. 9E). Quantitative RT-PCR experiments confirmed that Mad1 up-regulation is accompanied by a decrease of hTERT expression (Supplemental Fig. 2, column 3 and 4). These data indicate that Bmi-1 expression is required for maintaining the block of keratinocyte clonal evolution induced by p16INK4a inactivation.

DISCUSSION p16INK4a inactivation and immortalization The Locus INK4a/ARF encodes two proteins, p16INK4a and p14ARF (p19ARF in the mouse), in overlapping reading frames, which are commonly codeleted in a large variety of human cancers (47). The relative functions of each protein vary in a cell-type- and speciesspecific manner. In murine embryo fibroblasts, for example, loss of p19ARF is sufficient for immortalization, whereas loss of p16INK4a is not. Moreover, in vivo experiments in mouse model systems indicated that p19ARF plays an important protective role in oncogenic transformation and tumorigenicity (48 –50). On the other hand, epidemiological data showing the presence of point mutations and small deletions that specifically affect p16INK4a seem to implicate this protein as the most important tumor suppressor gene product of Locus INK4a/ARF in human (50, 51). Only rare selective inactivation of p14ARF with retention of p16INK4a has been reported in human tumors (38). However, the relevance of the tumor suppressor gene p16INK4a in E752

Vol. 20

July 2006

human keratinocyte replicative senescence is still debated (22, 23). In the present study, we demonstrate that persistent inactivation of p16INK4a in primary human keratinocytes is per se sufficient to: 1) allow bypass of replicative senescence by impairing keratinocyte clonal evolution; 2) maintain primary human keratinocytes in the stem cell compartment; and 3) induce telomere elongation and stabilization. In the adult, telomerase expression is low or absent in most human somatic tissues and is principally restricted to activated lymphocytes, germ cells, and tissue stem cells (2). In the epidermis, telomerase activity is restricted to keratinocytes endowed with high proliferative capacity, and its down-regulation correlates with terminal differentiation. This suggests that telomerase activity plays a role in the lifetime regenerative capacity of normal epidermis (52, 53). Bmi-1, a member of the family of PcG proteins that form repressor complexes, is necessary for the maintenance of adult stem cells in some tissues (46) and is overexpressed in human breast cancer cell lines (54). In addition, a cell type specific induction of telomerase by Bmi-1 has been reported previously (54). Here, we show that p16INK4a inactivation is accompanied by maintenance of Bmi-1 expression and telomerase activity at levels comparable with those of young keratinocytes. Sustained telomerase activity in turn results in telomere elongation and stabilization and consequent immortalization of cultures. On the other hand, overexpression of p16INK4a is associated with irreversible growth arrest, cell flattening, and expression of the senescence-associated marker SA-␤-gal in several cell types (55, 56). We show that also in primary human keratinocytes p16INK4a overexpression determines an abrupt clonal conversion and is associated with Bmi-1 down-regulation and telomerase inactivation. The concomitance of Bmi-1 expression and telomerase activity maintenance after p16INK4a inactivation, on the one hand, and of Bmi-1 down-regulation and telomerase inactivation after p16INK4a overexpression, on the other, indicates that these genes may be regulated by p16INK4a, either directly or as a consequence of the block of clonal evolution. Interestingly, Mad1, a telomerase repressor (45, 57), is overexpressed in antisense p16INK4a-transduced cells after bypass of senescence. Mad1 appears to regulate cell cycle withdrawal and to limit proliferation during in vivo cell differentiation (44, 58). Thus, we speculate that the maintenance of telomerase activity in our cells prompts Mad1 up-regulation in the attempt to repress hTERT transcription and to avoid escape from senescence. Infection of antisense p16INK4a -transduced keratinocyte cultures, with a defective retrovirus carrying an antisense of Bmi-1, shows that inactivation of Bmi-1 in colonies is accompanied by cell enlargement and flattening, features of cell senescence. Importantly, in antisense p16INK4a -antisense Bmi-1 transduced keratinocytes, down-regulation of Bmi-1 is followed by a strong

The FASEB Journal

MAURELLI ET AL.

increase of Mad1 and the subsequent decrease of telomerase expression and activity in the absence of p16INK4a up-regulation. These data demonstrate that Bmi-1 expression is required for maintaining the block of keratinocyte clonal evolution induced by p16INK4a inactivation. It has been reported that hTERT-mediated keratinocyte immortalization is accompanied by p16INK4a expression (21, 22, 30, 59). We hypothesize that, if telomerase is a downstream target of p16INK4a, this upregulation of p16INK4a presumably reflects attempts of feedback compensation. In this case, high levels of p16INK4a would not induce senescence because hTERT is constitutively active. As expected, inactivation of p16INK4a in primary human keratinocytes inhibits the pRb pathway maintaining the hyperphosphorylated state of pRB. This protein is a corepressor of E2F transcription factors, which are required for the regulation of genes essential for DNA replication and cell cycle progression. PCNA, a E2F-target (60), is highly expressed in skin equivalents obtained from antisense p16INK4a-transduced cells, providing evidence of hyperproliferative activity of keratinocytes. Inactivation of pRb pathway is followed by up-regulation of p14ARF expression (61), which prevents p53 degradation (39) in fibroblasts. Here, we show that inactivation of p16INK4a up-regulates p14ARF/p53/ p21Waf1 pathway in primary human keratinocytes. However, up-regulation of the p53 pathway does not appear to interfere with proliferation of antisense p16INK4atransduced keratinocytes. p16INK4a binds to Cdk4 and prevents its association with cyclin D, blocking formation of the active Cdk4/ cyclin D complex that phosphorylates pRb (38). Ectopic expression of Cdk4 or a p16INK4a-insensitive mutant Cdk4R24C in epithelial cells induces bypass of senescence. However, Cdk4 and Cdk4R24C overexpression induces a dramatic up-regulation of p16INK4a and does not lead to keratinocyte immortalization (22, 23). In the same manner, expression of human papilloma virus E7 oncoprotein, which inactivates pRb, can extend the life span of human keratinocytes but results in p16INK4a overexpression and telomere shortening (24). On the basis of our findings, we speculate that in these cells, despite inhibition of pRb pathway, up-regulation of p16INK4a does not allow Bmi-1 expression and telomerase activity resumption and subsequent cell immortalization. p16INK4a expression and clonal evolution Different from p16INK4a, p14ARF is not expressed at detectable levels during keratinocyte replicative senescence and no clear role for p14ARF in the growth arrest of cultured human cells has been demonstrated (19, 23). In addition, it has been recently reported that p14ARF expression is down-regulated during senescence in esophageal epithelial cells (59). These findings suggest that p16INK4a but not p14ARF is instrumental in human keratinocyte senescence. p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

Expression of dominant negative p53 has been demonstrated to extend the life span of human keratinocytes (23, 42). However, p53 pathway is less involved in senescence of human keratinocytes than in fibroblasts and not necessarily inactivated during immortalization of human keratinocytes (20, 24, 30, 59). Here, we confirm that p14ARF remains barely detectable and that p53 decreases during keratinocyte senescence, as described previously (19). Our data also demonstrate that p16INK4a inactivaction impairs keratinocyte clonal evolution in the presence of p14ARF and p53 expression. Altogether, these findings identify p16INK4a as a major regulator of human keratinocyte clonal evolution. We show that Bmi-1 expression and telomerase activity decrease during clonal evolution of normal human keratinocytes. It has been reported that p16INK4a regulates transcription of telomerase activity by inhibiting the function of Sp1 in human malignant gliomas (43). Here, we show that p16INK4a levels affect clonal evolution and modulate telomerase promoter activity also in normal human keratinocytes. The growth of keratinocyte colonies depends on outward migration of the rapidly proliferating cells located in a rim close to the colony perimeter (9). Interestingly, in normal keratinocyte cultures Bmi-1 is mainly expressed in the nuclei of cells located at the periphery of the colony while it is expressed in the cytoplasm of cells that undergo terminal differentiation. Altogether, our data suggest that irreversible growth arrest and terminal differentiation are associated with a decrease of Bmi-1 expression and cytoplasmic distribution. Bmi-1 deletion mutant studies showed that the ability of Bmi-1 to mediate cellular transformation in rodent fibroblasts correlates with its nuclear localization (62). Interestingly, in antisense p16INK4atransduced immortalized keratinocytes, in which we observe a block of clonal evolution, Bmi-1 expression is mostly nuclear. Moreover, inactivation of Bmi-1 in these colonies induces senescence features. These data support the involvement of Bmi-1 in the maintenance of stemness also in human keratinocytes (63). Replicative senescence, a state in which a cell is no longer able to proliferate, is accompanied by gene expression changes, which might be regulated by epigenetic mechanisms. Recently, senescence has been shown to correlate with formation of a distinct chromatin structure called senescence-associated heterochromatic foci (SAHF). SAHF do not contain active transcription sites, and they recruit heterochromatin proteins to the promoters that need to be stably repressed during senescence. In senescent fibroblasts, SAHF formation and silencing of target genes depend on pRb activity (64). In these cells, p16INK4a is crucial for ensuring the irreversibility of the senescence arrest. Once p16INK4a is expressed, unphosphorylated pRb establishes an essentially irreversible repressive chromatin state. Of note, the SV40 large T-antigen, which inactivates pRb by direct binding, stimulates DNA synthesis but not proliferation in p16INK4a-expressing fiE753

broblasts, suggesting that the p16INK4a-related cell proliferation arrest can be independent of continual pRb activity (18). Finally, experiments performed on osteogenic sarcoma cell clones with inducible p16INK4a expression revealed that a sustained period of p16INK4a expression is sufficient to impose a durable block to cell proliferation, which then becomes independent of p16INK4a expression and hypophosphorylation of pRb (55). Our data demonstrate that antisense-p16INK4a transduction allows keratinocytes to bypass replicative senescence only in primary cultures bearing holoclones. On the contrary, antisense-p16INK4a transduction in cultures approaching senescence (only comprising meroclones and paraclones) does not affect clonal evolution, and the onset of senescence is anyway triggered. We hypothesize that irreversible epigenetic changes have already occurred in TA keratinocytes, in which p16INK4a is highly expressed, and that down-regulation of p16INK4a is no more able to revert the proliferative capacities of cells already committed to terminal differentiation. In keeping with our data, down-regulation of p16INK4a in senescent fibroblasts, presenting high p16INK4a expression, does not induce cell proliferation even after concomitant p53 inactivation (18). Thus, inactivation of p16INK4a selectively prevents clonal conversion in keratinocytes endowed with a high proliferative potential. p16INK4a inactivation and self-renewal

We thank Dr. Saveria Pastore for support in quantitative RT-PCR experiments and the Art Department of Istituto Vol. 20

July 2006

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13.

14.

Bypass of replicative senescence and immortalization are key initial steps in cell transformation (65). Tumors may originate from the transformation of normal stem cells. Since self-renewal is the hallmark of stem cells in normal and neoplastic tissues, it seems reasonable that newly arising cancer cells appropriate the machinery for self-renewing cell division expressed in normal stem cells. Indeed, pathways associated with oncogenesis, such as the Notch, Sonic hedgehog, and Wnt signaling pathways, also regulate stem cell self-renewal (66). These findings support the emerging model in which master regulator molecules govern the processes of stem cell self-renewal, normal aging, and tumor development. Specifically to the INK4A/ARF locus, it has been reported that: 1) the locus products are the principal mediators of hematopoietic and neural stem cell longevity in vivo (46); 2) the same proteins are increased in almost all rodent tissues with advancing age (67); and 3) the locus is altered in most human cancers by genetic and epigenetic mechanisms (47). In close agreement with these data, our findings indicate that the tumor suppressor gene p16INK4a regulates keratinocyte clonal evolution and that inactivation of p16INK4a in epidermal stem cells is necessary for maintaining stemness.

E754

Dermopatico dell’Immacolata for the artwork. This work was supported by the Italian Ministry of Health.

15. 16.

17. 18.

19.

20.

21.

22.

23.

Hayflick, L. (1965) The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614 – 636 Sharpless, N. E., and DePinho, R. A. (2004) Telomeres, stem cells, senescence, and cancer. J. Clin. Invest. 113, 160 –168 Watt, F. M. (2001) Stem cell fate and patterning in mammalian epidermis. Curr. Opin. Genet. Dev. 11, 410 – 417 Alonso, L., and Fuchs, E. (2003) Stem cells of the skin epithelium. Proc. Natl. Acad. Sci. U. S. A. 100, 11830 –11835 Lajtha, L. G. (1979) Stem cell concepts. Differentiation 14, 23–34 Barrandon, Y. (1993) The epidermal stem cell: an overview. Dev. Biol. 4, 209 –215 Campisi, J. (1998) The role of cellular senescence in skin aging. J. Invest. Dermatol. Symp. Proc. 3, 1–5 Dellambra, E., Golisano, O., Bondanza, S., Siviero, E., Lacal, P., Molinari, M., D’Atri, S., and De Luca, M. (2000) Down-regulation of 14 –3-3o´ prevents clonal evolution and leads to immortalization of primary human keratinocytes. J. Cell Biol. 149, 1117–1130 Barrandon, Y., and Green, H. (1987) Three clonal types of keratinocyte with different capacites for multiplication. Proc. Natl. Acad. Sci. U.S.A. 84, 2302–2306 Rochat, A., Kobayashi, K., and Barrandon, Y. (1994) Location of stem cells of human hair follicles by clonal analysis. Cell 76, 1063–1073 Sherr, C. J., and DePinho, R. A. (2000) Cellular senescence: mitotic clock or culture shock? Cell 102, 407– 410 Shay, J. W., and Wright, W. E. (2005) Senescence and immortalization: role of telomeres and telomerase. Carcinogenesis 26, 867– 874 Hara, E., Tsurui, H., Shinozaki, A., Nakada, S., and Oda, K. (1991) Cooperative effect of antisense-Rb and antisense-p53 oligomers on the extension of life span in human diploid fibroblasts, TIG-1. Biochem. Biophys. Res. Commun. 179, 528 –534 Shay, J. W., Pereira-Smith, O. M., and Wright, W. E. (1991) A role for both RB and p53 in the regulation of human cellular senescence. Exp. Cell Res. 196, 33–39 Atadja, P., Wong, H., Garkavtsev, I., Veillette, C., and Riabowol, K. (1995) Increased Activity of p53 in Senescing Fibroblasts. Proc. Natl. Acad. Sci. U. S. A. 92, 8348 – 8352 Alcorta, D. A., Xiong, Y., Phelps, D., Hannon, G., Beach, D., and Barrett, J. C. (1996) Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in replicative senescence of normal human fibroblasts. Proc. Natl. Acad. Sci. U. S. A. 93, 13742–13747 Smogorzewska, A., and de Lange, T. (2002) Different telomere damage signaling pathways in human and mouse cells. EMBO J. 21, 4338 – 4348 Beausejour, C. M., Krtolica, A., Galimi, F., Narita, M., Lowe, S. W., Yaswen, P., and Campisi, J. (2003) Reversal of human cellular senescence: roles of the p53 and p16 pathways. EMBO J. 22, 4212– 4222 Munro, J., Stott, F. J., Vousden, K. H., Peters, G., and Parkinson, E. K. (1999) Role of the alternative INK4A proteins in human keratinocyte senescence: Evidence for the specific inactivation of p16INK4A upon immortalization. Cancer Res. 59, 2516 –2521 Dickson, M. A., Hahn, W. C., Ino, Y., Ronfard, V., Wu, J. Y., Weinberg, R. A., Louis, D. N., Li, F. P., and Rheinwald, J. G. (2000) Human keratinocytes that express hTERT and also bypass a p16INK4a-enforced mechanism that limits life span become immortal yet retain normal growth and differentiation characteristics. Mol. Cell. Biol. 20, 1436 –1447 Ramirez, R. D., Morales, C. P., Herbert, B.-S., Rohde, J. M., Passons, C., Shay, J. W., and Wright, W. E. (2001) Putative telomere-independent mechanisms of replicative aging reflect inadequate growth conditions. Genes Dev. 15, 398 – 403 Ramirez, R. D., Herbert, B. S., Vaughan, M. B., Zou, Y., Gandia, K., Morales, C. P., Wright, W. E., and Shay, J. W. (2003) Bypass of telomere-dependent replicative senescence (M1) upon overexpression of Cdk4 in normal human epithelial cells. Oncogene 22, 433– 444 Rheinwald, J. G., Hahn, W. C., Ramsey, M. R., Wu, J. Y., Guo, Z., Tsao, H., De Luca, M., Catricala, C., and O’Toole, K. M. (2002)

The FASEB Journal

MAURELLI ET AL.

24.

25.

26. 27.

28.

29.

30.

31.

32.

33. 34.

35.

36.

37.

38. 39.

40. 41. 42.

A two-stage, p16INK4A- and p53-sependent keratinocyte senescence mechanism that limits replicative potential independent of telomere status. Mol. Cell. Biol. 22, 5157–5172 Kiyono, T., Foster, S. A., Koop, J. I., McDougall, J. K., Galloway, D. A., and Klingelhutz, A. J. (1998) Both Rb/p16INK4a inactivation and telomerase activity are required to immortalize human epithelial cells. Nature 396, 84 – 88 Mathor, M. B., Ferrari, G., Dellambra, E., Cilli, M., Mavilio, F., Cancedda, R., and De Luca, M. (1996) Clonal analysis of stably transduced human epidermal stem cells in culture. Proc. Natl. Acad. Sci. U. S. A. 93, 10371–10376 Rheinwald, J. G., and Green, H. (1975) Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell 6, 331–343 Pellegrini, G., Dellambra, E., Golisano, O., Martinelli, E., Fantozzi, I., Bondanza, S., Ponzin, D., McKeon, F., and De Luca, M. (2001) p63 identifies keratinocyte stem cells. Proc. Natl. Acad. Sci. U. S. A. 98, 3156 –3161 Berghella, L., De Angelis, L., Coletta, M., Berarducci, B., Sonnino, C., Salvatori, G., Anthonissen, C., Cooper, R., ButlerBrowne, G. S., Mouly, V., et al. (1999) Reversible immortalization of human myogenic cells by site-specific excision of a retrovirally transferred oncogene. Hum. Gene Ther. 10, 1607–1617 Bernerd, F., Asselineau, D., Vioux, C., Chevallier-Lagente, O., Bouadjar, B., Sarasin, A., and Magnaldo, T. (2001) Clues to epidermal cancer proneness revealed by reconstruction of DNA repair-deficient xeroderma pigmentosum skin invitro. Proc. Natl. Acad. Sci. U. S. A. 98, 7817–7822 Pellegrini, G., Dellambra, E., Paterna, P., Golisano, O., Traverso, C. E., Rama, P., Lacal, P., and De Luca, M. (2004) Telomerase activity is sufficient to bypass replicative senescence in human limbal and conjunctival but not corneal keratinocytes. Eur. J. Cell Biol. 83, 691–700 Bond, J. A., Haughton, M. F., Rowson, J. M., Smith, P. J., Gire, V., Wynford-Thomas, D., and Wyllie, F. S. (1999) Control of replicative life span in human cells: barriers to clonal expansion intermediate between M1 senescence and M2 crisis. Mol. Cell. Biol. 19, 3103–3114 Quelle, D., Zindy, F., Ashmun, R., and Sherr, C. (1995) Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest. Cell 83, 993–1000 Jones, P., and Watt, F. M. (1993) Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell 73, 713–724 Paramio, J. M., Casanova, M. L., Segrelles, C., Mittnacht, S., Lane, E. B., and Jorcano, J. L. (1999) Modulation of cell proliferation by cytokeratins K10 and K16. Mol. Cell. Biol. 19, 3086 –3094 Santos, M., Paramio, J. M., Bravo, A., Ramirez, A., and Jorcano, J. L. (2002) The expression of keratin K10 in the basal layer of the epidermis inhibits cell proliferation and prevents skin tumorigenesis. J. Biol. Chem. 277, 19122–19130 Roop, D. R., Krieg, T. M., Mehre, l. T., Cheng, C. K., and Yuspa, S. H. (1988) Transcriptional control of high molecular weight keratin gene expression in multistage mouse skin carcinogenesis. Cancer Res. 48, 3245–3252 Bitoun, E., Micheloni, A., Lamant, L., Bonnart, C., TartagliaPolcini, A., Cobbold, C., Al Saati, T., Mariotti, F., MazereeuwHautier, J., Boralevi, F., et al. (2003) LEKTI proteolytic processing in human primary keratinocytes, tissue distribution and defective expression in Netherton syndrome. Hum. Mol. Genet. 12, 2417–2430 Huschtscha, L. I., and Reddel, R. R. (1999) p16INK4a and the control of cellular proliferative life span. Carcinogenesis 20, 921–926 Stott, F., Bates, S., James, M., McConnell, B., Starborg, M., Brookes, S., Palmero, I., Ryan, K., Hara, E., Vousden, K., and Peters, G. (1998) The alternative product from the human CDKN2A locus, p14(ARF), participates in a regulatory feedback loop with p53 and MDM2. EMBO J. 17, 5001–5014 Shay, J. W., and Bacchetti, S. (1997) A survey of telomerase activity in human cancer. Eur. J. Cancer 33, 787–791 Reddel, R. R. (2001) An alternative lifestyle for immortalized oral keratinocytes. J. Clin. Invest. 108, 665– 667 Opitz, O. G., Suliman, Y., Hahn, W. C., Harada, H., Blum, H. E., and Rustgi, A. K. (2001) Cyclin D1 overexpression and p53

p16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

43.

44. 45.

46. 47. 48.

49. 50.

51. 52.

53.

54.

55. 56. 57. 58.

59.

60. 61. 62.

63.

inactivation immortalize primary oral keratinocytes by a telomerase-independent mechanism. J. Clin. Invest. 108, 725–732 Saito, M., Nakagawa, K., Hamada, K., Hirose, S., Harada, H., Kohno, S., Nagato, S., and Ohnishi, T. (2004) Introduction of p16INK4a inhibits telomerase activity through transcriptional suppression of human telomerase reverse transcriptase expression in human gliomas. Int. J. Oncol. 24, 1213–1220 Grandori, C., Cowley, S. M., James, L. P., and Eisenman, R. N. (2000) The Myc/Max/Mad network and the transcriptional control of cell behavior. Ann. Rev. Cell Dev. Biol. 16, 653– 699 Gunes, C., Lichtsteiner, S., Vasserot, A. P., and Englert, C. (2000) Expression of the hTERT Gene Is Regulated at the Level of Transcriptional Initiation and Repressed by Mad1. Cancer Res. 60, 2116 –2121 Park, I.-K., Morrison, S. J., and Clarke, M. F. (2004) Bmi1, stem cells, and senescence regulation. J. Clin. Invest. 113, 175–179 Lowe, S. W., and Sherr, C. J. (2003) Tumor suppression by Ink4a-Arf: progress and puzzles. Curr. Opin. Genet. Dev. 13, 77– 83 Kamijo, T., Zindy, F., Roussel, M. F., Quelle, D. E., Downing, J. R., Ashmun, R. A., Grosveld, G., and Sherr, C. J. (1997) Tumor suppression at the mouse INK4a locus mediated by the alternative reading frame product p19ARF. Cell 91, 649 – 659 Krimpenfort, P., Quon, K. C., Mooi, W. J., Loonstra, A., and Berns, A. (2001) Loss of p16Ink4a confers susceptibility to metastatic melanoma in mice. Nature 413, 83– 86 Sharpless, N. E., Bardeesy, N., Lee, K. H., Carrasco, D., Castrillon, D. H., Aguirre, A. J., Wu, E. A., Horner, J. W., and DePinho, R. A. (2001) Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis. Nature 413, 86 –91 Ruas, M., Peters, G. (1998) The p16INK4a/CDKN2A tumor suppressor and its relatives. Biochim. Biophys. Acta 1378, F115– 177 Harle-Bachor, C., and Boukamp, P. (1996) Telomerase activity in the regenerative basal layer of the epidermis in human skin and in immortal and carcinoma-derived skin keratinocytes. Proc. Natl. Acad. Sci. U. S. A. 93, 6476 – 6481 Bickenbach, J. R., Vormwald-Dogan, V., Bachor, C., Bleuel, K., Schnapp, G., and Boukamp, P. (1998) Telomerase is not an epidermal stem cell marker and is down-regulated by calcium. J. Invest. Dermato. 111, 1045–1052 Dimri, G. P., Martinez, J.-L., Jacobs, J. J. L., Keblusek, P., Itahana, K., van Lohuizen, M., Campisi, J., Wazer, D. E., and Band, V. (2002) The Bmi-1 oncogene induces telomerase activity and immortalizes human mammary epithelial cells. Cancer Res. 62, 4736 – 4745 Dai, C., and Enders, G. (2000) p16 INK4a can initiate an autonomous senescence program. Oncogene 19, 1613–1622 Duan, J., Zhang, Z., and Tong, T. (2001) Senescence delay of human diploid fibroblast induced by anti-sense p16INK4a expression. J. Biol. Chem. 276, 48325– 48331 Oh, S., Song, Y., Yim, J., and Kim, T. (2000) Identification of Mad as a repressor of the human telomerase (hTERT) gene. Oncogene 19, 1485–1490 Foley, K. P., McArthur, G. A., Queva, C., Hurlin, P. J., Soriano, P., and Eisenman, R. N. (1998) Targeted disruption of the MYC antagonist MAD1 inhibits cell cycle exit during granulocyte differentiation. EMBO J. 17, 774 –785 Harada, H., Nakagawa, H., Oyama, K., Takaoka, M., Andl, C. D., Jacobmeier, B., von Werder, A., Enders, G. H., Opitz, O. G., and Rustgi, A. K. (2003) Telomerase induces immortalization of human esophageal keratinocytes without p16INK4a inactivation. Mol. Cancer Res. 1, 729 –738 Bravo, R., Frank, R., Blundell, P., and Macdonald-Bravo, H. (1987) Cyclin/PCNA is the auxiliary protein of DNA polymerase-delta. Nature 326, 515–517 Bates, S., Phillips, A. C., Clark, P. A., Stott, F., Peters, G., Ludwig, R. L., and Vousden, K. H. (1998) p14ARF links the tumour suppressors RB and p53. Nature 395, 124 –125 Cohen, K., Hanna, J., Prescott, J., and Dang, C. (1996) Transformation by the Bmi-1 oncoprotein correlates with its subnuclear localization but not its transcriptional suppression activity. Mol. Cell. Biol. 16, 5527–5535 Claudinot, S., Nicolas, M., Oshima, H., Rochat, A., and Barrandon, Y. (2005) From the cover: Long-term renewal of hair follicles from clonogenic multipotent stem cells. Proc. Natl. Acad. Sci. U. S. A. 102, 14677–14682

E755

64.

Narita, M., Nunez, S., Heard, E., Narita, M., Lin, A. W., Hearn, S. A., Spector, D. L., Hannon, G. J., and Lowe, S. W. (2003) Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113, 703–716 65. Hahn, W. C., Counter, C. M., Lundberg, A. S., Beijersbergen, R. L., Brooks, M. W., and Weinberg, R. A. (1999) Creation of human tumour cells with defined genetic elements. Nature 400, 464 – 468

E756

Vol. 20

July 2006

66.

Reya, T., Morrison, S., Clarke, M., and Weissman, I. (2001) Stem cells, cancer, and cancer stem cells. Nature 414, 105–111 67. Krishnamurthy, J., Torrice, C., Ramsey, M. R., Kovalev, G. I., Al-Regaiey, K., Su, L., and Sharpless, N. E. (2004) Ink4a/ Arf expression is a biomarker of aging. J. Clin. Invest. 114, 1299 –1307

The FASEB Journal

Received for publication June 20, 2005. Accepted for publication March 14, 2006.

MAURELLI ET AL.

The FASEB Journal • FJ Express Summary

Inactivation of p16INK4a (inhibitor of cyclin-dependent kinase 4A) immortalizes primary human keratinocytes by maintaining cells in the stem cell compartment Riccardo Maurelli,* Giovanna Zambruno,† Liliana Guerra,* Claudia Abbruzzese,* Goberdhan Dimri,‡ Mara Gellini,* Sergio Bondanza,* and Elena Dellambra*,1 *Laboratory of Tissue Engineering and Cutaneous Physiopathology and †Laboratory of Molecular and Cell Biology, I.D.I.–IRCCS, Istituto Dermopatico dell’Immacolata, Rome, Italy; and ‡Division of Cancer Biology, Department of Medicine, Evanston, Illionois, USA To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.05-4480fje SPECIFIC AIMS Progressive increase of p16INK4a expression has been shown to correlate with keratinocyte clonal evolution. However, it is not clear whether p16INK4a accumulation is a triggering mechanism of human epidermal replicative senescence or a secondary event. Thus, the aim of our study was to investigate the role of p16INK4a in human keratinocyte clonal evolution.

PRINCIPAL FINDINGS 1. p16INK4a inactivation alone induces replicative senescence bypass in primary human keratinocyte p16INK4a was inactivated in primary human keratinocytes, using a defective retrovirus, carrying the fulllength human cDNA of p16INK4a (named FL), or a 100 bp fragment (named Exo1a), corresponding to exon 1␣ of INK4a/ARF locus, in antisense orientation. Primary keratinocytes transduced with an empty vector underwent senescence, whereas antisense p16INK4atransduced keratinocytes (both FL and Exo1a) bypassed replicative senescence and continued to divide at a rate comparable with young keratinocytes. p16INK4a levels were undetectable already before bypass of senescence and p16INK4a inactivation was maintained during subcultivation. INK4a

2. p16 down-regulation does not induces replicative senescence bypass in TA keratinocytes Keratinocyte cultures approaching senescence, i.e., enriched in transient amplifying (TA) cells and depleted in stem cells, were transduced with the aforementioned retroviral vectors. Antisense p16INK4a-transduced TA keratinocytes did not bypass replicative senescence, although antisense transcripts were highly expressed and p16INK4a expression resulted down-regulated. 1516

3. p16INK4a inactivation impairs keratinocyte clonal evolution Empty vector-transduced cells showed a progressive decrease of their colony-forming efficiency (CFE) during serial cultivation and a progressive increase of the percentage of paraclones. In contrast, antisense p16INK4a-transduced keratinocytes displayed a progressive increase of CFE reaching values ⬎60% (percentage of aborted colonies constantly ⬍3%), which remained then constant during serial cultivation. Of note, this impairment of clonal evolution was observed in antisense p16INK4a-transduced cells several passages before the onset of replicative senescence of empty-vector cells (Fig. 1). Bypass of senescence was accompanied by resumption of p63, a marker of high proliferative potential in epithelial cells. This finding suggests a recruitment of cells endowed with high proliferative potential in antisense p16INK4a-transduced cultures. 4. Inactivation of p16INK4a maintains transduced keratinocytes in the stem cell compartment We examined the percentage of stem and TA cells in our transduced cultures. After bypass of senescence, cultures generated from antisense p16INK4a-transduced cells were enriched in holoclones (considered the surface epithelial stem cells) and markedly depleted of paraclones (terminal TA cells). 5. Persistent inactivation of p16INK4a is required to block clonal conversion After bypass of senescence, antisense p16INK4a-transduced cells were infected with a Cre-expressing adeno1 Correspondence: Laboratory of Tissue Engineering and Cutaneous Physiopathology, IDI, Istituto Dermopatico dell’Immacolata, Via dei Castelli Romani, 83/85, Pomezia (Roma) 00040, Italy. E-Mail: [email protected] doi: 10.1096/fj.05-4480fje

0892-6638/06/0020-1516 © FASEB

companied by elongation and stabilization of the telomeres. Luciferase assays showed that p16INK4a levels are able to modulate telomerase promoter activity in normal human keratinocytes. Cell-type specific induction of telomerase by Bmi-1 has been reported. Interestingly, p16INK4a inactivation-induced bypass of senescence was accompanied by resumption of Bmi-1 expression (Fig. 2). On the contrary, down-regulation of p16INK4a in TA cells was unable to resume Bmi-1 and telomerase activity at levels comparable with young keratinocytes. 10. Bmi-1 expression is required for maintaining the block of keratinocyte clonal evolution induced by p16INK4a inactivation

Figure 1. Clonogenic potential of antisense p16INK4a transduced cells. A) CFE performed at selected cell passages indicated by numbers of cell doublings. B) CFE values of keratinocytes transduced with antisense p16INK4a (red bars) or with empty vector (blue bars) were plotted against cell passages. C) Aborted colony values (% of paraclones) of keratinocytes transduced with antisense p16INK4a (red line) or with empty vector (blue line) were plotted against cell passages.

virus to ablate antisense expression. Ad-Cre infected cells showed senescence features. The transgene deletion was accompanied by re-expression of p16INK4a, and by reduction of CFE. 6. Inactivation of p16INK4a does not impair control mechanisms of keratinocyte growth

Primary keratinocytes were infected with a defective retrovirus carrying a recombinant cDNA encoding the fusion protein p16INK4a-enhanced green fluorescent protein. Keratinocytes immediately underwent senescence and were able to generate only paraclones. Up-regulation of p16INK4a was accompanied by a decrease of telomerase expression and activity, and downregulation of Bmi-1. Inactivation of Bmi-1 in antisense p16INK4a-transduced keratinocyte cultures resulted in increased percentage of Bmi-1 negative enlarged and flattened colonies. Importantly, in antisense p16INK4a-antisense Bmi-1 transduced keratinocytes, down-regulation of Bmi-1 is followed by a strong increase of Mad1, a

Antisense p16INK4a-transduced keratinocytes were shown to 1) need growth factor and FCS; 2) grow only in anchorage-dependent manner; and 3) have a normal complement of 46 chromosomes. 7. Antisense p16INK4a-transduced cells maintain their ability to reconstitute a differentiated human epidermis Histological and immunohistochemical analysis of in vitro reconstituted skin equivalents showed that antisense p16INK4a-transduced cells were able to stratify into basal and suprabasal layers, although signs of hyperproliferative activity were clearly evident. 8. p16INK4a inactivation inhibits pRb and up-regulates p53 pathways Inactivation of p16INK4a inhibited the pRb pathway, maintaining the hyperphosphorylated state of pRb, and induced an up-regulation of the p14ARF/p53/p21Waf1 pathway. 9. p16INK4a inactivation-induced bypass of senescence is accompanied by resumption of Bmi-1 and by maintenance of telomerase activity, hallmarks of tissue regenerative capacity In antisense p16INK4a-transduced keratinocytes, telomerase activity was resumed. This resumption was acp16INK4a AND PRIMARY KERATINOCYTE IMMORTALIZATION

Figure 2. Telomerase activity and Bmi-1 expression. Telomeric repeat amplification protocol (TRAP; A) and TRF (B) assays were performed on cells transduced with empty vector (V lanes) and antisense p16INK4a-transduced cells (AS-Exo1a lanes) at different cell passages. C) Reporter telomerase promoter, p16INK4a, and lacZ control plasmids were cotransfected in primary keratinocytes. Promoter activities were calculated using the empty vector control (column 1) as reference. The ratio promoter:p16INK4a was 1:1 (column 2), 1:2 (column 3), and 1:3 (column 4). D) c-Myc, Mad1, and Bmi-1 expression was assessed by Western blot using cell extracts from empty vector (V-lanes) and Exo1a (AS-Exo1alanes) transduced cells at different cell passages. In A, B, and D, different cell passages were indicated by numbers of cell generations after infection. 1517

Figure 3. Schematic diagram shows that inactivation of p16INK4a in early passage primary human keratinocyte cultures (bearing holoclones) is per se sufficient to induce cellular immortalization by impairing keratinocyte clonal evolution and maintaining cells in stem cell compartment. Inactivation of p16INK4a in late-passage primary human keratinocyte cultures (depleted in holoclones) does not affect clonal evolution, and onset of senescence is anyway triggered. Similar findings are obtained when overexpressing p16INK4a in early passage primary human keratinocytes.

telomerase repressor, and a decrease of telomerase expression and activity in the absence of p16INK4a up-regulation.

CONCLUSIONS AND SIGNIFICANCE In the present study, we demonstrate that direct and persistent inactivation of p16INK4a in primary human keratinocytes is per se sufficient to allow bypass of replicative senescence by impairing keratinocyte clonal evolution and maintaining primary human keratinocytes in the stem cell compartment. p16INK4a inactivation is accompanied by maintenance of Bmi-1 expression and telomerase activity, hallmarks of tissue regenerative capacity. In our transduced cells, sustained telomerase activity in turn results in telomere maintenance and immortalization of cultures. We also show that inactivation of Bmi-1 in antisense p16INK4atransduced keratinocyte colonies is accompanied by senescence features. Importantly, down-regulation of Bmi-1 is followed by a strong increase of Mad1 and the subsequent decrease of telomerase expression and activity in the absence of p16INK4a up-regulation. These data demonstrate that Bmi-1 expression is required for maintaining the block of keratinocyte clonal evolution induced by p16INK4a inactivation. It has been reported that hTERT-mediated keratinocyte immortalization is accompanied by p16INK4a expression. Our data indicate that in those cells upregulation of p16INK4a reflects attempts of feedback compensation but does not induce senescence because

1518

Vol. 20

July 2006

hTERT is constitutively active. We show that, as expected, inactivation of p16INK4a in primary human keratinocytes inhibits the pRb pathway maintaining the hyperphosphorylated state of pRB. It has been reported that inactivation of pRb pathway, by ectopic expression of Cdk4, a p16INK4a-insensitive mutant Cdk4R24C, or human papilloma virus E7 oncoprotein, in epithelial cells can extend the life span but does not lead to keratinocyte immortalization. However, keratinocyte transduction results in p16INK4a overexpression and telomere shortening. On the basis of our findings, we speculate that in those cells, despite inhibition of the pRb pathway, up-regulation of p16INK4a does not allow Bmi-1 expression and telomerase activity resumption and subsequent cell immortalization. Our data also demonstrate that p16INK4a inactivation impairs keratinocyte clonal evolution in the presence of p14ARF and p53 expression, identifying p16INK4a as a major regulator of human keratinocyte clonal evolution. In addition, we show that p16INK4a expression directly affects clonal evolution, modulates telomerase promoter activity, and down-regulates telomerase and Bmi-1 expressions and telomerase activity. We demonstrate that Bmi-1 expression and telomerase activity decrease during clonal evolution of primary human keratinocytes. Immunocytochemistry results suggest that irreversible growth arrest and terminal differentiation are associated with decrease of Bmi-1 expression and cytoplasmic distribution. As reported, the ability of Bmi-1 to mediate cellular transformation in rodent fibroblasts correlates with its nuclear localization. Interestingly, antisense p16INK4a-transduced immortalized keratinocytes, in which we observe a block of clonal evolution, display a predominantly nuclear Bmi-1 expression, whereas inactivation of Bmi-1 in these cells induces senescence features. These data support the involvement of Bmi-1 in the maintenance of stemness also in human keratinocytes. Antisense-p16INK4a transduction in cultures approaching senescence does not affect clonal evolution, and the onset of senescence is anyway triggered. This finding suggests an irreversible role of p16INK4a accumulation in inducing clonal conversion. In keeping with data obtained in senescent fibroblasts, we hypothesize that irreversible epigenetic changes have already occurred in transient amplifying keratinocytes, in which p16INK4a is highly expressed, and that downregulation of p16INK4a is no more able to affect the proliferative capacity of cells already committed to terminal differentiation. Thus, antisense-p16INK4a transduction allows keratinocytes to bypass replicative senescence only in primary cultures bearing holoclones. Altogether, our findings indicate that the tumor suppressor gene p16INK4a regulates keratinocyte clonal evolution and that inactivation of p16INK4a in epidermal stem cells is necessary for maintaining stemness in vitro.

The FASEB Journal

MAURELLI ET AL.