Discussion
Senescence on the cellular level is believed to be a program of irreversible cell cycle arrest, in which the cell stays unresponsive to mitogenic stimulation. However, metabolic activity of the cell is retained, resulting in the appearance of some morphological manifestations that may be defined as a senescent cell phenotype, and include enlargement accompanied by flattening of the cell, enhanced activity of senescence-associated ß-galactosidase, lipofuscin accumulation, as well as increased cytoplasmic granularity and nuclear volume. Some further alterations related to the nucleus and its functions comprise formation of SAHF and, above all, changes in the expression pattern of tumor-suppressor genes, e.g.
p53,
p21
Cip1/Waf1/Sdi1
,
p14
Arf
,
p16
Ink4a
,
pRb[
16‐
19].
However, on the basis of the results obtained in the present study, as well as our previous report [
20] and insight gained from experiments of others, we suggest that there should be a clear distinction between senescence on the cellular level and senescence of the whole population of cells, e.g. cell line models. The first observations regarding these phenomena came from the studies of Hayflick, who presented a restricted proliferative capacity of cellular populations
in vitro[
21]. Later, analogical characteristics were shown to exist in cancer cell lines as a result of chemotherapeutic treatment and other therapeutic modalities [
22‐
24], although these cells were previously regarded to be "immortal" due to unrestrained proliferation and telomerase expression. Moreover, recently the clinical impact of the senescence program, as well as the pros and cons of its induction have been broadly discussed [
25‐
28]. At the same time, the lack of a clear definition and unequivocal markers indicative of this phenomenon, makes it particularly difficult to firmly establish its contribution to the final outcome of therapy. All the more so because some features are misleadingly perceived as crucial determinants of the program. These features could include enhanced senescence-associated ß-galactosidase activity and polyploidy induction.
As regards polyploidy induction, it has been documented that cell population undergoing replicative senescence may feature a defective mitotic checkpoint, which eventually leads to aberrant mitoses [
29]. In view of these findings, it cannot be ruled out that some nuclear alterations accompanying polyploidy reflect a crisis-like event, including some breakage-fusion-bridge cycles that result in increased genomic instability. In our research, a G
2/M population was increasing in parallel with polyploidy induction, which may rather suggest an endoreplication-like response after mitotic slippage. At the same time, the extent of G
2/M accumulating cells as well as SAHF containing cells seemed to be relatively low as compared to those manifesting enhanced SA ß-galactosidase, a fact which may, along with polyploidization, suggest an unstable senescence-like program induced by etoposide. Also results of the TUNEL method imply an accumulation of cells from G
0/G
1 and S phases of the cell cycle into G
2/M and transient arrest followed by some endoreplication/cell death events. Due to the fact that extensive DNA fragmentation accompanied these alterations, chromosomal instability and breakage-fusion-bridge cycles could not be excluded. Similar results were obtained by Sliwinska et al. (2009) after treatment of HCT116 cells with doxorubicin, another topoisomerase II inhibitor. The authors claimed that a substantial fraction of the cells underwent endocycling without entering mitosis, and that some dividing polyploid cells produced aneuploid progeny [
30].
DNA damage signaling has been proved to be coupled to its replication in direct response to topoisomerase inhibitors, including etoposide in the A549 cell population. On the one hand it suggests the induction of DSB following collision of cleavable complexes with replication forks, but some DSB also appeared in non-replicating cells, which may be indicative of other mechanisms of action of this drug, i.e. oxidative stress generation or collision of cleavable complexes with transcription machinery [
31]. On the other hand, it is tempting to consider its influence on mitotic cells and the involvement of the DNA replication process in the long-term response to etoposide. Our results obtained from the TUNEL method confirm the findings of Zhao et al. (2012) regarding the lack of cell-cycle specific effect on DSB induction. However, in our study, G
2/M and polyploid fractions apparently accumulated the highest numbers of damaged cells, which may imply some abnormal/abortive mitoses accompanied by endoreplication-like events. This interpretation is further supported by the fact that in our investigation, regular mitotic figures were only observed exceptionally in etoposide-exposed cells, whereas nuclei presenting atypical morphology were rather abundant. In addition, the TUNEL-positive cells were more abundant after treatment in comparison with the cells presenting biochemical features of cell death in the annexin V/PI assay. There is clear evidence that endo-polyploidization in senescent cell population constitutes a consequence of re-replication of genomically-damaged G
2/M cells [
32]. At the same time, independent karyoplasts/reproductive genome-reduced subcells may form from giant polyploid cells via some kind of reductional division, resulting in increased genetic variation [
33,
34]. What is more, induction of DNA damage in cancer cell lines was proven to be related not only to endopolyploidy and depolyploidization, but also to overexpression of meiosis-specific genes and mobilization of their products to the sites strategic for division [
35‐
37]. More recently it has also been reported that although there are some morphological and biochemical features of senescence as well as growth arrest in the population of A549 cells exposed to topoisomerase II inhibitors [
20,
38], long term effects of the treatment include a re-growth of a fraction of population originating most probably from cancer stem cells [
38]. In this situation, a somewhat more prolonged, but still reversible growth arrest of the population occurred after inhibition of the DNA damage response pathway. Due to the fact that 24 h after exposure to etoposide we still observed a relatively high number of TUNEL-positive cells in comparison with those annexin V/PI-positive, this pathway was most probably active in our experiment as well.
Hence, some alterations in heterochromatin features, like gluey or stick heterochromatin in near-senescent population of cultured cells, seem to be responsible for mitotic failure and chromosomal instability [
39]. Analogically, etoposide may induce impairments in untangling of chromatin, which in turn interferes with the proper formation of chromosomes and the decatenation of chromatids [
5], leading most probably to the same final outcome, i.e. induction of a senescent-like phenotype in the cancer cell population along with increased polyploidy and genetic instability, as well as heterogeneity of the response. What is more, an unstable senescence program induced in cancer cell populations by DNA damage is thought to be related to a phenomenon corresponding to the above-described series of events, and termed neosis [
40,
41]. In fact this process may contribute to chemotherapy resistance, due to the fact that increased instability along with repeated cycles of self-renewal may favour the accumulation of survival-promoting mutations and the expansion of more invasive subclones, especially under continuous contact with a chemotherapeutic drug [
37]. Moreover, polyploidy was suggested to be a transiently appearing response to treatment, followed by differentiation of the genetic material into subnuclei that are either targeted to survival or to autophagic degradation [
42].
As regards the second above-mentioned senescence marker, i.e. senescence-associated ß-galactosidase, the activity of this enzyme reflects changes in lysosomal content and functions, and consequently the altered cellular metabolism of aging cells [
16,
43‐
45], which does not interchangeably mean a cell cycle-arrested cell. This is also supported by our observations of the cell ultrastructure following the treatment, which revealed abundantly present large lysosomes containing numerous inclusions of degradable and non-degradable materials such as lipofuscin, but also some organelles or their components. However, as an increased lysosomal mass may also appear in other processes such as, for example, autophagy, as evidenced by acridine orange staining [
46], an enhanced activity of this enzyme may be indicative of this kind of metabolic shift as well. Therefore, senescence-associated ß-galactosidase activity should not be regarded as the sole marker of the senescence program on the cellular level. Moreover, our results showed an uncoupling between an enhanced activity of this enzyme and cell cycle arrest, as the tendencies of these two parameters seemed not fully consistent. Analogical results were obtained by Dulić et al. (2000), who reported a discrepancy between the senescence-like phenotype and cell cycle block in human diploid fibroblasts transfected with HPV
E6 oncogene that contributed to p53 degradation and in
p21
Cip1/Waf1/Sdi1
nullizygous mouse embryonic fibroblasts [
47]. Thus, being neither a crucial determinant nor an unequivocal reflection of stable cell cycle arrest, an enhanced senescence-associated ß-galactosidase activity rather reflects a metabolic response of the cell to stress. Nevertheless, it has been recently reported that autophagy mediates a transition of mitotic cells to senescence, especially influencing the phenotype, and that autophagic degradation of DNA is an important factor during segregation of subnuclei in polyploid cells [
42,
48]. These suggestions are also consistent with the particularly high activity of this enzyme in polyploid A549 cells resulting from the treatment with etoposide observed in this study. What is more, in our fluorescent microscopic, as well as transmission electron microscopic analyses, we were able to distinguish some diffuse DAPI staining, DNase I staining and some dispersed chromatin-like structures in the cytoplasm of A549 cells after treatment. This supports the hypothesis of DNA degradation. It therefore seems that an increase in the activity of senescence-associated ß-galactosidase is particularly related to phenotypic alterations in the senescent cell population. Morphological changes typical of a senescence-like response were also confirmed by our observations at the ultrastructural level, which was evident from increased heterogeneity of the cytoplasmic compartment, but also progressive alterations of the nuclei. The latter were most probably a phenotypic manifestation of abnormal mitoses/mitotic catastrophe.
Some hallmarks of the senescence phenomenon strictly related to the execution of a cell cycle arrest are believed to be changes in expression of the cell cycle regulatory molecules, including cyclin-dependent kinase inhibitors, cyclin-dependent kinases and cyclins, as well as formation of SAHF [
18,
23,
25,
26,
30]. p21
Cip1/Waf1/Sdi1 was previously reported to be a reliable marker of prematurely induced senescence in many cancer cell lines. Particularly, profound alterations in the expression of the gene encoding this protein and/or in the localization of its product were observed, which may be the major contributor to the G
2/M arrest, but also may activate the G
1 checkpoint [
30,
49]. For example, Sliwinska et al. (2009) found lack of Ki-67 expression along with increased expression of p21
Cip1/Waf1/Sdi1 and cyclin D1 in HCT116 cells after incubation with a low dose of doxorubicin. As regards their fluorescence microscopic observations, the large senescent-like cells manifested evident bright signals for p21
Cip1/Waf1/Sdi1[
30]. To refer more specifically to the A549 population, we would like to mention the work of Shen and Maki (2010). The authors documented there that p21
Cip1/Waf1/Sdi1 induction and its nuclear accumulation were indispensable for maintenance of a senescence-associated G
1 cell cycle arrest of A549 cells with unreduced DNA content (4N) after transient Nutilin-3a treatment, with knock-down of
p21
Cip1/Waf1/Sdi1
or
p53 resulting in endoreduplication [
49]. These findings are particularly interesting in light of our observations, which did not reveal significant changes in p21
Cip1/Waf1/Sdi1, supporting the hypothesis of an unstable senscence-like state induction that was accompanied by endocycling events in our conditions. As reported earlier, insufficient p21
Cip1/Waf1/Sdi1 activity in human fibroblasts with morphological features of a senescence program may lead to uncoupling between phenotypic features of the senescence program and cell cycle arrest [
47], in accordance with our results as well. Further evidences also reinforce this point of view, indicating that in the absence of the proper p21
Cip1/Waf1/Sdi1 function, cells may undergo a transient G2-like arrest, which is followed by replication without regular intervening mitoses [
50].
We think that such a compromised senescence phenotype or senescence-like state may be a consequence of the mutations identified in A549 cells, in particular homozygous deletion of the
Ink4b/Arf/Ink4a locus [
8‐
11]. This, in turn, may render the cells incapable of activation of not only p16
Ink4a, but also p14
Arf, which is an upstream regulator of p53-mediated p21
Cip1/Waf1/Sdi1 induction via HDM2 sequestration. Previous works aimed at elucidation of the functions that are specifically fulfilled by p16
Ink4a and p21
Cip1/Waf1/Sdi1 during senescence execution and suggested their complementation for the most efficient induction and maintenance of the program [
51]. The circumstances vital for the irreversibility of senescence are presently being determined for both the Rb and p53-mediated pathways, in which the above-mentioned proteins constitute essential players. These involve the positive feedback loop of ROS accumulation in case of p21
Cip1/Waf1/Sdi1 and SAHF formation in case of p16
Ink4a -Rb [
52,
53]. Therefore, in order to verify a potential contribution of the p16
Ink4a -Rb arm that may result in the G
0/G
1 arrest, we checked the status of cyclin D1 and SAHF formation in our study.
In agreement with other reports describing accumulation of cyclin D1 in senescent cells and our results with the HL-60 cell line (manuscript in preparation), here we showed some changes in the amount and localization of this protein in a senescence-like state. Increased cyclin D1 level is regarded by some authors as a feature of senescence-like conditions that are accompanied by polyploidization events [
30,
54‐
56]. At the same time, the major role of this protein is to act as an antagonist of Rb-mediated cell cycle arrest and to drive the cell cycle progression through G1 phase [
57]. Therefore, it should not be surprising that an enhanced expression of this protein may be followed by an unfavorable prognosis, especially if it is activated during senescence-like conditions [
30,
56]. Many reports concerning cyclin D1 in senescence, refer in fact to a senescence-like state in the population of heterogeneous by nature cancer cells. The recent report of McGowan et al. (2012) clearly demonstrated the possible latent effects of nuclear accumulation of cyclin D1 in a population of MCF-7 cells presenting typical senescence-like morphology as a consequence of the 14
Arf-p53-p21-Rb pathway activation. These effects were predominantly related to relocalization of Ki-67 into the nucleoli, increased genomic instability and abnormal cell divisions [
56]. Such morphological features and nuclear cyclin D1 accumulation were evident also in our experiment, suggesting, along with low frequency of SAHF, the instability of the induced state. As regards nucleolar sequestration of the cyclin D1 protein observed by us, the phenomenon may have several alternative implications for senescence execution, which presently remain obscure. First of all, sequestration of other cell-cycle regulators contributes to their inactivation [
58]. Bearing in mind the morphology of the cells that accumulated cyclin D1 and only faint, dispersed staining in the cells with SAHF, it may also be hypothesized that sequestration probably serves the special purpose of inactivation of the protein on a cellular route to senescence. However, the phenomenon requires a more in-depth revision, especially in face of the fact that mammalian cyclin D1/Cdk4 complexes have been proven to contribute to endoreduplication, most probably via the stimulation of nucleoli enlargement and ribosomal biogenesis [
59]. In our study the nuclei of polyploid cells showed often intense cyclin D1 staining, sometimes accompanied by intranucleolar aggregations as well.
In this situation, if the cyclin D1 sequestration had been involved in the execution of a cell cycle arrest, we would have been able to observe an increased number of the cells with SAHF manifestation. Derepression of Rb after inactivation of cyclin D1, may have led to supression of the E2F target genes crucial for replication and, finally, to activation of the SAHF pathways essential for irreversibility of the program. In fact, this was not the case in our conditions. Instead, we found the cells with widespread DNA damage and abnormal morphological features indicative of inappropriate DNA replication and aberrant proliferation.
In order to gain a more in-depth insight into the state of A549 cells following exposure to etoposide, we referred also to analyses of some previously described cytoskeletal indicators of senescence, i.e. G-actin and vimentin.
Vimentin is mainly expressed in cells of mesenchymal origin and has been described as a marker of epithelial-to-mesenchymal transition, which is a phenomenon related to increased invasiveness and resistance of some cancer cells, thus manifesting the progression of cancer [
60,
61]. On the other hand, there are studies documenting an abundant vimentin presence, as well as its special role in phenotypic changes that accompany senescence phase in cell populations [
12,
15,
62].
Here we would like to suggest for the first time a special role for vimentin in multinucleated dividing cells resulting from etoposide exposure. Since the protein accumulates especially in the intranuclear spaces of multinucleated cells which are expected to give origin to aneuploid progeny, as well as in invaginations of lobulated nuclei, it is highly probable that it also contributes somehow to the formation, separation, segregation and isolation of nuclei with reduced DNA content, which, according to previous reports [
33,
42], is followed by the budding of individual karyoplasts/paradiploid descendants from the parental nucleus. Darkly stained structures were observed by Walen (2005) in the division furrow during amitotic divisions of polyploid nuclei in a senescent cell population [
33]. Our morphological observations strongly suggest the involvement of vimentin in this process, because the most intense fluorescence signal was also evident in division axis between putative daughter nuclei identified by their aberrant morphology, different sizes and shapes. In these cells, similar to the study of Walen (2005), we noticed the presence of some division-like furrows starting from invaginations that were in close proximity to vimentin-rich areas of the cytoplasm, as well as small distances between sister nuclei, filled with septum-like vimentin structures, which suggests amitotic divisions. Cytoplasmic tails, most probably indicative of the process of karyoplast budding, were also present in both studies, and in our case they were shown to be vimentin-rich. Additionally, in our experiment aggregation of this protein was particularly evident around multiple nuclei of giant cells, often at the expense of its distribution in the perpiheral cytoplasmic areas, which on the whole resembled the process of vimentin retraction into the form of ring-like structures. It has previously been shown that vimentin participates in cytokinesis, being expressed in a cleavage-furrow specific manner. Its mutant form, unresponsive to Aurora B phosphorylation, left daughter cells unseparated [
63]. In light of these findings and our results it seems especially interesting that giant cancer cells appearing after irradiation and expressing catalytically active Aurora B kinase, also retain the capacity to divide [
64]. Some multinucleated cells presented by these authors also had features of abnormal divisions including uneven distribution of chromatin into daughter nuclei, the presence of thin chromatin threads and cytoplasmic tails connecting dividing cells. Besides that, previous studies have indicated that intermediate filaments are crucial for regulation of germ cell nuclear behaviour, participating in meiosis, selection of meiotic products, formation of the gametic pronucleus, its transfer across cell-cell junctions and, finally, in pronuclear fusion during zygote formation in
Tetrahymena[
65]. Intermediate filament-related protein MNS1 is also thought to contribute to the organization of nuclear morphology during meiotic prophase [
66]. In addition, the activity of the isolated head domain of vimentin, released by HIV-1 protease, has been shown to be both necessary and sufficient for alterations of nuclear shape including the formation of invaginations and extensive lobulation, as well as for chromatin reorganization [
67].
In further agreement with our hypothesis, rare cells presenting heterochromatin morphology indicative of SAHF, which may imply more stable cell cycle arrest, presented quite the opposite vimentin organization, i.e. rather regular, faintly-woven, occasionally even dispersed structures. Thus far there have been contradictory indications related to senescence-associated vimentin regulation, as well as its possible function in the program. Firstly, overproduction of vimentin is considered to be indicative of senescence phase in cultured cell population [
12], however, depending on the tissue or cell type, its expression may also be diminished [
68]. Secondly, apart from its anti-oncogenic and suppressive functions during cellular transformation, the possible involvement of this protein in senescence delay and spontaneous transformation has been proposed [
12,
69]. Vimentin is believed to contribute to recombinational repair of nuclear and mitochondrial DNA [
69] and, in line with these observations, this cytoskeletal protein may participate in cytoplasmic anchorage of p53 during senescence [
15], thus making it unavailable for its transcriptional targets and, as a consequence, limiting its nuclear-dependent tumor suppressor functions and downstream pathways of cell death and/or cell cycle arrest. However, as opposed to senescence, which is considered to be a tumor-suppressive mechanism, other reports also suggest an involvement of this protein in immortalization and epithelial-to-mesenchymal transition, both events related to carcinogenesis and tumor progression. However this may only be an apparent contradiction, because a senescent cell population constitutes a source of extended life cells, participating in transformation to immortalized ones [
33]. What is more, since it is engaged in extensive repair processes, this protein has the potential to promote increased plasticity of the genome and invasiveness of cancer cells [
69]. Interestingly, further evidence of vimentin’s contribution to the integrity and repair of the genome is the fact that the destruction of this cytoskeletal component during early stages of apoptosis actually reinforces a pro-apoptotic signaling cascade in the cell, via the product of its proteolysis [
70‐
72], thus making it an attractive candidate among possible „point of no return” sensors. Here we would like to point out the difference in vimentin organization between cells with typical manifestations of SAHF, the multinucleated cells containing abnormally-shaped, irregular nuclei including micronuclei, their putative descendants, as well as the apoptotic cells. Additionally, our cytometric analyses support indications of an increased vimentin content in the senescent cell population as a whole, which may reflect not only a shift in morphology towards an increased cell volume, but also accumulation of vimentin in the cytoplasmic areas crucial for the separation of individual nuclei.
G-actin has previously been described as a marker of senescence, and further studies revealed that especially its nuclear accumulation may be correlated with cell cycle arrest in G
1[
13,
73]. Nuclear G-actin pool was visibly increased not only in replicative, but also in stress-induced premature senescence of human diploid fibroblasts, which was also accompanied by alterations in the expression and distribution of phosphorylated Erk1/2 kinases and small G proteins (belonging to the Rho family). These changes have been interpreted as a reflection of impairments in nuclear export of actin resulting from improper regulation of MEK signaling pathway [
73]. In turn, subsequent observations have shown that the increase in G-actin pool may also be related to the presence of cofilin in the nuclei of senescent G
1-arrested fibroblasts, because cofilin appeared there in its active, dephosphorylated form, capable of F-actin fragmentation [
13]. In support of this, it was also reported that dephosphorylated cofilin-dependent nuclear import of actin occurs in response to stress [
74] and that analogical stress-induced events are regulated independently of apoptosis [
75]. Moreover, there are some indications of possible interrelations between nuclear actin and proliferative capacities of the cell. These suggest that an enforced G-actin export from the nucleus induced not only the reversal of phenotypic and biochemical senescence features, but also stimulated cell cycle progression and replication of previously G
1-arrested cells [
13]. On the other hand, impairments in G-actin export were related to growth arrest and the appearance of senescent morphology [
13,
76]. Furthermore, nuclear G-actin accumulation is considered to be an early and universal marker of senescence in the cell, including senescent cancer cells [
13].
In our study, microscopic quantitative analyses of G-actin content in the nucleus area did not show significant changes after etoposide exposure. However, we believe that this is consistent with previous reports suggesting that dephosphorylated cofilin is a crucial factor for nuclear G-actin accumulation and stable G
1 cell cycle arrest. In fact, we observed an increase only in G
2/M population, in which cofilin levels are supposed to be the lowest. Thus, the lack of nuclear G-actin accumulation in this study further reinforces our assumption that unreduced G
2/M cells were not irreversibly arrested in G
1, but rather underwent endoreplication. However, some cytoplasmic G-actin alterations were evident and we suggest they may be caused by several synergistically acting events, including oxidative stress generation, cell death induction, exocytosis and changes in the cell morphology. Oxidative stress is known to influence the amino acid sequence of actin, via modifications in residues critical for the process of polymerization [
77]. In accordance with this, and with previous studies documenting dose-dependent depolymerization of F-actin under the influence of etoposide [
78], we observed here a more prominent cytoplasmic staining for G-actin with both methods employed. Based only on DNase I labeling, we would not have been able to exclude the possibility that this more intense signal had in fact indicated an enhanced mitochondrial biogenesis and/or autophagic DNA degradation in response to etoposide, as has previously been suggested [
42,
79].
The most intense fluorescence of G-actin was documented in this study for cells presenting apoptotic-like morphology, including nuclear fragmentation and the formation of apoptotic bodies, which was most probably also indicative of F-actin cytoskeleton destruction, but may also imply an involvement of actin in the development of morphological symptoms and/or cell death execution. For example, it has been shown that actin filaments aggregate at sites of apoptotic bodies formation [
80]. Apart from that, early stages of apoptosis are characterized by low expression of actin, subsequently followed by F-actin reorganization and accumulation of G-actin in apoptotic bodies [
81]. In light of these findings, it also seems meaningful that G-actin functions as an inhibitor of DNase I [
82‐
84] until the proteolytic cleavage in advanced stages of apoptosis. Furthermore, inhibition of the proteolytic cleavage of actin results in the suppression of DNA fragmentation in etoposide-treated cells, and mutation of the cleavage-site may actually lead to cellular transformation [
82,
85]. The presence of actin was shown to be crucial for transduction of necrosis-related signaling into mitochondria [
86].
Methods
Cell culture
A549 cells were grown in monolayer in Dulbecco's modified Eagle's medium with Glutamax (DMEM; Gibco) with 50 μg/ml gentamycine (Sigma-Aldrich) and supplemented with fetal bovine serum (FBS; Gibco) at a final concentration of 10%. Cell cultures were maintained at 37°C in a humidified CO2 incubator.
Etoposide (Sintopozid®) was obtained from S.C. SINDAN S.R.L., Bucharest, Romania. The drug was kept in accordance with the supplier's recommendations, and working solutions were prepared in fresh medium prior to use. Cells in the exponential phase of growth, 24 h after seeding, were incubated with increasing concentrations of etoposide (0; 0.75; 1.5; 3 μM) for 72 h. Before further experimental procedures, the cells were cultured for an additional 24 h in a fresh drug-free medium.
Senescence-associated ß-galactosidase assay
Senescence-associated ß-galactosidase activity was determined using the Senescent Cells Staining Kit (Sigma-Aldrich), in accordance with the enclosed instructions. After rinsing with Dulbecco's PBS, the cells were fixed on coverslips with a solution containing 2% formaldehyde/0,2% glutaraldehyde for 6 min at room temperature. After being washed in DPBS, the cells were stained with a freshly prepared solution (pH 6) including X-gal (5-bromo-4-chloro-3-indolyl-ß-galactopyranoside) in the dark at 37°C for 20 h. After that, the coverslips were rinsed again and mounted with Aqua Poly/Mount (Polysciences).
The percentage of cells displaying a visible reaction in the presence of indigogenic substrate was established microscopically (Eclipse E800, Nikon), and was presented as the mean value from five fields for each experimental condition, from at least ten independent repeats.
Cell cycle analysis
Cells growing on 6-well plates, after trypsinization and washing, were suspended in PBS (1–2 × 106 cells/ml). The cell pellet obtained from centrifugation of suspensions (5 min, 300 × g) was resuspended in 1 ml of NSS solution containing 50 μg/ml PI, 0.01% (w/v) sodium citrate (Sigma-Aldrich) and 0.03% (v/v) nonylphenylpolyethylene glycol (Nonidet P40 Substitute; Fluka). Following centrifugation (5 min, 300 × g), the cells were resuspended in 250 μl of NSS and stained at room temperature in the dark for 15 min. Finally, each sample was incubated with 250 μl of RNase A solution (at a concentration of 10 μg/ml in PBS) (Sigma-Aldrich) for 15 min at room temperature in the dark. After the addition of 0.5 ml of PBS, flow cytometric analysis was performed on a FACScan flow cytometer (Becton-Dickinson). Cell cycle distribution was estimated for 20 000 events using CellQuest software (Becton-Dickinson). The whole cell population was divided into fractions with respective DNA contents (G0/G1; S; G2/M; <G1; >G2).
TUNEL assay
A commercially available kit (APO-DIRECT, BD Biosciences Pharmingen) was used to perform the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) method analysis to assess DNA fragmentation rate. Following trypsinization, cells were fixed with 1% (w/v) formaldehyde (Polysciences) in PBS, 15 min, on ice and then washed with PBS. Before staining, the cells were resuspended in 70% (v/v) ice-cold ethanol (POCh) and kept at -20°C for at least 18 h. After fixation and permeabilization, the cells were incubated with staining mixture containing TdT enzyme and substrate (dUTP-FITC) at 37°C. Counterstaining was performed with PI/RNase solution. Finally, the cells were analyzed on FACScan using CellQuest software (Becton-Dickinson). 20 000 events were acquired and non-clumped cells were gated.
Annexin V assay
The extent of apoptotic and necrotic cell death was evaluated based on double staining with annexin V-FITC and PI using the Annexin V-FITC Apoptosis Detection Kit (BD Biosciences Pharmingen). The procedure was performed according to the enclosed protocol. Briefly, after trypsinization and washing of the cells with PBS, they were resuspended in annexin V binding buffer including 2.5% annexin V-FITC (v/v) for a 15-min incubation (dark, RT). After centrifugation (5 min, 300 × g), annexin V binding buffer with 5% addition of PI was incubated with the resulting pellet for 5 min in the above-mentioned conditions. Finally, flow cytometric analysis (FACScan, CellQuest software; Becton-Dickinson) allowed us to determine the percentages of cells presenting biochemical features of early apoptosis (annexin V-positive/PI-negative), late apoptosis (annexin V-positive/PI-positive), necrosis (annexin V-negative/PI-positive), as well as viable cells (annexin V-negative/PI-negative).
Immunofluorescence
Vimentin labeling
Cells grown on coverslips in 12-well plates were immunolabeled according to previously published protocols [
92]. The cells were prefixed (10 min, 37°C) with 0.4 mg/ml bifunctional protein crosslinking reagent DTSP in HBSS [DTSP, 3,3’-dithio-bis(propionic acid
N-hydroxysuccinimide ester); Hank’s balanced salt solution, Sigma-Aldrich] and preextracted in Tsb [0.5% (w/v) Triton X-100 (Serva) in MTSB] containing 0.4 mg/ml DTSP for 10 min at 37°C, were further extracted in Tsb (5 min, 37°C) and fixed with 4% (w/v) paraformaldehyde (Sevra) in MTSB (15 min, 37°C). After incubation with 0.1 M glycine, the cells were blocked with 1% (w/v) bovine serum albumin (both from Sigma-Aldrich) in Tris-buffered saline, pH 7.6 (BSA-TBS), 2 × 5 min. Mouse monoclonal anti-vimentin antibody (clone V9; Sigma-Aldrich) in BSA-TBS (1:200) was used for a 45-min incubation at 37°C in a moist chamber. After being rinsed thoroughly with BSA-TBS (3 × 5 min), the cells were incubated with goat anti-mouse IgG-TRITC (Sigma-Aldrich) (1:85) or goat anti-mouse IgG-BODIPY (Molecular Probes) in BSA-TBS (1:200) in the same conditions. After washing with BSA-TBS and PBS, the coverslips were counterstained with DAPI (diaminophenyl indole; Sigma-Aldrich), washed 3 × 5 min with PBS and, finally, mounted with Aqua Poly/Mount (Polysciences).
G-actin labeling
G-actin fluorescent staining was performed according to two protocols, i.e. using DNase I and vitamin D-binding protein.
In the first case, permeabilization and fixation steps were identical to those described for vimentin staining. After incubation with 0.1 M glycine in DPBS (5 min, RT) and washing with DPBS (2 × 5 min), the cells were incubated with 2 μM phalloidin (Sigma-Aldrich) in DPBS and, subsequently, with DNase I conjugated to Alexa Fluor 488 (Molecular Probes) in DPBS (1:50, 20 min, in the dark). The coverslips were then washed with PBS, stained with DAPI, washed again (3 × 5 min) and mounted on glass slides with Aqua Poly/Mount (Polysciences).
In the second of the above-mentioned methods for indirect staining of non-filamentous actin, previously described procedures were employed, with some slight modifications [
93‐
95]. Coverslips were washed with PBS and fixed with 3.7% (w/v) paraformaldehyde (Serva) in PBS (20 min, RT). Following washing with PBS (3 × 5 min), a 10-min incubation (RT) in PBS with the addition of 1 mg/ml NaBH
4 (Sigma-Aldrich) was performed in order to reduce free aldehyde groups. Then, the cells were rinsed in PBS (3 × 5 min), extracted in 0.1% (w/v) Triton X-100 in PBS, and further incubated with PBS-TB [0.1% (w/v) Triton X-100/0.2% (w/v) BSA in PBS] for 10 min. Incubation steps aiming at G-actin visualization were as follows: 10 μg/ml vitamin D-binding protein (DBP, Calbiochem) in PBS-TB (60 min, RT), goat polyclonal anti-DBP IgG (Santa Cruz Biotechnology) in PBS-TB (1:50, 60 min, RT), Alexa Fluor 488-conjugated donkey anti-goat IgG (Molecular Probes; 1:100, 60 min, 37°C in a moist chamber). All these incubations were split and followed by rinsing in PBS-TB, 3 × 5 min. Finally, the coverslips were washed twice with PBS, counterstained with DAPI, washed with PBS (3 × 5 min) and subsequently mounted with Aqua Poly/Mount (Polysciences).
Cyclin D1 labeling
Cells were briefly washed with PBS, fixed in 4% paraformaldehyde (15 min, RT) and then washed with PBS (3 × 5 min). After that, the cells were incubated in permeabilization solution (0.1% Triton X-100 in PBS, 5 min) and blocked with 1% BSA (30 min). Then, an incubation with mouse monoclonal cyclin D1 antibody (Sigma-Aldrich) (60 min, RT, moist chamber) followed, the cells were washed three times with PBS and incubated with Alexa Fluor 488® goat anti-mouse IgG (Molecular Probes) (45 min, RT). Nuclear staining was performed using DAPI (Sigma-Aldrich). Finally, the cells were washed with PBS and mounted on slides in Aqua Poly/Mount (Polysciences).
p21Waf1/Cip1/Sdi1labeling
The procedure of prefixation, fixation and blocking was like in the above-described protocol for vimentin. Incubation with mouse monoclonal anti-p21 antibody (sc-817; Santa Cruz Biotechnology) in BSA-TBS (1:100) was performed in a moist chamber for 60 min, RT. Then the slides were rinsed with BSA-TBS (3 × 5 min) and incubated with a secondary antibody (Alexa Fluor® 488 goat anti-mouse IgG conjugate from Molecular Probes) in BSA-TBS (1:100) in the same conditions. The final steps were also as described above.
SAHF determination and nuclear morphology evaluation
The estimation of SAHF foci formation was performed by counting the DAPI-stained nuclei with the typically distinguished morphology of clearly separated condensations, enclosed within intact nuclear envelope (microscopic observations). For each sample/concentration, at least 500 cells were taken into account for each repeat of the experiment. Other morphological features were also considered for the quantification of nuclear changes after the treatment: the cells with enlarged nuclei, the cells with apoptotic bodies, the cells with nuclei of a regular size but presenting abnormal morphology and the cells with regular nuclei.
An Eclipse E800 microscope and NIS Elements AR imaging software (both from Nikon) were employed to obtain and analyze documentation.
Flow cytometric analysis of intracellular vimentin
Cells grown on 6-well plates were trypsinized, washed with PBS, centrifuged (5 min, 300 × g) and suspended at a final concentration of 1–2 × 106 cells/ml in 1 ml PBS supplemented with 100 μl of formaldehyde (Polysciences). Samples were incubated for 15 min in the dark and centrifuged (5 min, 300 × g), the resulting pellet was subsequently permeabilized in 2 ml of ice-cold 50% (v/v) methanol (JT Baker). After a 15-min incubation on ice, the cells were washed twice with cold PBS and resuspended in 100 μl of PBS. The cell suspensions were then transferred into flow cytometric tubes containing 20 μl of PE-conjugated mouse monoclonal anti-vimentin IgG1 (clone V9), or normal mouse IgG1-PE as an isotype control (Santa Cruz Biotechnology). After a 30-min incubation on ice, the cells were washed with PBS, centrifuged for 5 min at 500 × g to remove residues of antibody, and resuspended in 200 μl of PBS for flow cytometric analysis on FACScan (Becton-Dickinson). The percentage of vimentin-positive cells, as well as the mean fluorescence intensity of the cell population was established using CellQuest software (Becton-Dickinson).
Quantitative immunofluorescence for nuclear G-actin measurements
In order to prevent contamination by cytoplasmic actin, nuclear G-actin was analyzed after isolation of nuclei. Following trypsinization, washing with PBS and centrifugation (5 min, 300xg), 2 ml of homogenizing buffer [50 mM Tris–HCl, pH 7.5; 0.3 M sucrose; 15 mM CaCl2; 25 mM MgCl2; 10 mM 2-mercaptoethanol (all from Sigma-Aldrich); 0.5% (v/v) Nonidet P40 Substitute] were added to each sample. After homogenization on ice, the suspensions were centrifuged (10 min, 1000 × g) and the supernatant was decanted. The pellet was suspended in 1 ml of homogenizing buffer and gently stratified on top of the buffer for nuclei purification [10 mM Tris–HCl; 0.3 M sucrose; 25 mM MgCl2; 25 mM KCl; 10 mM 2-mercaptoethanol (all from Sigma-Aldrich); 41% (v/v) glycerol (Roth)] in fresh tubes (8 ml). The nuclei obtained by centrifugation (10 min, 1000xg) were washed with PBS, fixed in 3.7% (w/v) paraformaldehyde (Sevra), and finally cytocentrifuged on glass slides. Fluorescent staining was performed with vitamin D-binding protein according to the above-mentioned procedure. Digital images from fluorescence microscopy (Eclipse E800; Nikon) were imported to ImageJ software (National Institutes of Health) and converted to grayscale (8-bit). G-actin fluorescence was colocalized with DAPI and subtraction by the nucleus area was performed. Mean fluorescence intensity was presented as the mean grey value, which is the average grey value for all pixels within the indicated area.
Transmission electron microscopy (TEM)
In order to perform an ultrastructural analysis, cells grown on 6-well plates were trypsinized and fixed with 2.5% (v/v) glutaraldehyde (Merck) in 0.1 M sodium cacodylate (Fluka), pH 7.4 for 30 min at RT. Following washing in cacodylate buffer, the addition of bovine thrombin (Biomed) and fibrinogen (Fluka) resulted in fibrin clot formation, in which cells were entrapped. Then, the samples were postfixed with 2% (w/v) OsO4 (Serva) in 0.1 M cacodylate buffer for 1 h at room temperature, passed through a series of ethanol and acetone solutions, and finally embedded in Epon 812 (Roth). Ultrathin sections were prepared, stained with uranyl acetate/lead citrate (Fluka) and examined with a transmission electron microscope (JEM-100CX; JEOL).
Statistical analysis
Statistical software packages (StatSoft, Tulsa, OK; GraphPad Prism, San Diego, CA) were employed for our analyses. The results obtained were compared using the following tests: the non-parametric Mann–Whitney U test, the Duncan test (for more than 10 repeats, homogeneous variance), as well as the Cochran–Cox test (for non-homogeneous variance). Statistical significance was determined at P<0.05, unless otherwise stated.