Implication of integrin α2β1 in senescence of SK-Mel-147 human melanoma cells

Downregulation of integrin α2β1 enhances the cellular senescence of SK-Mel-147 human melanoma cells

To determine the effects of α2β1 upon the mechanisms of senescence, we assessed the effect of its downregulation on the activity of beta-galactosidase, a typical marker of senescence. Suppression of this integrin was carried out by transducing the cells with a plasmid clone expressing α2-specific shRNA. Western blot and flow cytometry data showed high efficiency of the used clone (Figure 1A, 1B). As shown in Figure 1C, 1D, blocking the expression of α2β1 led to a sharp increase in cellular senescence, as well as a significant decrease in the proliferation of the cells. Inhibition of proliferative activity is a typical trait of cellular senescence.

We have previously shown that knockdown of α2β1 was accompanied by a significant increase in anoikis and in the levels of p53 and p21 proteins in SK-Mel-147 cells [28]. Numerous studies have shown that the oncosuppressor p53 and its downstream target p21 play a key role in the mechanism of cellular senescence and the choice of a cell to respond with either irreversible growth arrest or apoptosis [2]. Decreased proliferation of α2β1-depleted cells (Figure 1C) might be the result of increased activity of p53/p21 that induces both senescence and cell death.

The role of p53/p21 signaling in SK-Mel-147 cell senescence induced by α2β1 integrin downregulation

The p53 tumor suppressor and its downstream effector CDK inhibitor p21 play an essential role in the mechanisms of cell proliferation, apoptosis, and senescence [2, 31]. It has been shown that the initial phase of senescence depends on the high levels of p53 and p21, which then decreases [2, 32] Of importance, the choice of cellular response to a stress stimulus between apoptosis and senescence depends on the level and duration of the stress signal; a weak and transient signal leads to a transient arrest of growth, while senescence requires a stable signal. However, exceeding the threshold level of a stress stimulus leads to cell death [2, 33]. The choice between apoptosis and senescence is also determined by the ratio between p53 and p21; the stimulating effect of p21 on senescence is based on its ability to inhibit apoptosis, while the increased activity of p53 inhibits p21 [2].

Figure 2A shows that α2β1 downregulation was accompanied by a two-fold increase in the expression of p53 and p21 proteins, which is in agreement with our earlier data [28]. Based on this result and the previous citations, it may be inferred that the stress signal induced by downregulating α2β1 in studied cells led to apoptosis and/or senescence. One of the phenotypic features of apoptosis is the accumulation of a subpopulation of cells with low (subG1) DNA content in its total cell population. The data (Figure 2B) indicate that the percentage of subG1 cells in the control SK-Mel-147 population did not exceed 5–7% and did not change in the population with reduced α2β1 expression after 48 hours of growth in culture media. Thus, the increased levels of p53/p21 did not lead to increased apoptotic death of the studied cells.

It could be assumed that the effect of a2β1 downregulation on senescence was not due to increased p53/p21 signaling but instead was initiated by another signaling pathway, such as p16/Rb. To clarify this issue, we investigated the effect of blocking the expression of each of these proteins on the senescence of SK-Mel-147 cells with reduced α2β1 expression. As seen in Figure 2A, in cells lacking α2β1, siRNA-induced silencing of p53 reduced the expression of this protein to the level of control cells and, as evidenced in Figure 2C, suppression of p53 and p21 sharply reduced the ratio of senescent cells in the α2β1-downregulated population. In cells with intact α2β1, the effect of sip53 and sip21 was insignificant most likely because of already low senescence of these cells.

Thus, the deprivation of melanoma cells of one of the β1-family receptors is a stress factor that induces the senescence phenomenon by activating the p53/p21 signaling pathway.

The role of Erk signaling in SK-Mel-147 cell senescence induced by downregulation of α2β1

As noted above, the information on the contribution of MAPK to integrin-dependent signaling mechanisms that control cellular senescence is limited. Thus, it seemed to be of interest to determine the implication of Erk signaling in demonstrated in this study modifications of proliferation and senescence induced by α2β1 suppression. Figures 3A, 3B show that blocking α2β1 significantly reduced the active (phosphorylated) form of Erk in SK-Mel-147 cells. This result may indicate that in the studied cells, Erk signaling is involved in the mechanism of proliferation but not senescence. To verify this suggestion, we investigated the effect of PD98059, a specific inhibitor of MEK/Erk, on proliferation and senescence of the control and α2β1-depleted SK-Mel-147 cells. A 24 h treatment with the Erk inhibitor decreased DNA replication rates of control cells by approximately 30% (Figure 3C), whereas DNA replication of the α2β1-depleted cells was unaffected, most likely because it had already been highly reduced. On the other hand, inhibition of Erk did not affect the proportion of SA-β-Gal positive cells in the populations of control and α2β1-depleted cells (Figure 3D). These results suggest that increase in senescence caused by depletion of α2β1 in melanoma cells is not due to the reduction in Erk activity.

The role of the PI3K/Akt/mTOR pathway in SK-Mel-147 cell senescence induced by downregulation of α2β1

One of the most studied and important signaling mechanisms underlying oncogene-induced senescence is the PI3K/Akt/mTOR pathway [34, 21]. The critical point of this pathway is transmitting the activating signals from Akt through TSC2 to the mTORC1 complex, followed by activation of p53/p21 signaling [34]. This pathway is typical for cell responses to stress challenges, such as DNA damage, oxidative stress, radiation exposure, etc. Under non-stress conditions, Akt performs a well-known (canonical) function aimed at stimulating cell proliferation, tumor progression, overcoming apoptosis, etc., which is realized through the activation of a number of transcription factors, such as NF-kB, cMyc, etc. [35]. A protective function of Akt against cell senescence (particularly in maintaining the integrity of telomeres by the phosphorylation of hTERT and its translocation from the cytoplasm to the nucleus) has previously been described [36, 37]. RASSF1-mediated activation of p21 led to senescence of lung adenocarcinoma cells through the inhibition of Akt [38].

Our recent studies have shown that in SK-Mel-147 cells depleted of integrins α2β1, α3β1, and α5β1, Akt-mediated signals suppressed an in vitro cell invasion and enhanced anoikis (i.e. exhibited non-canonical properties) [28–30]. In this study, we assessed the implication of Akt signaling in senescence of SK-Mel-147 cells caused by the suppression of integrin α2β1.

The data presented in Figure 4A show that blocking α2β1 led to a sharp increase in the phosphorylated (active) form of Akt (pAkt) with no effects on Akt total protein levels and also enhanced the expression of the Akt downstream effector mTOR protein kinase.

We suggested that enhanced activity of these protein kinases is not just a trait accompanying increased senescence but rather can be attributed to their involvement in the mechanisms of senescence. To verify this suggestion, we investigated the effect of inhibitors of the Akt/mTOR pathway on senescence in melanoma cells depleted of α2β1. To block this pathway, the cells were treated with a PI3K inhibitor (LY294002) and an mTORC1 inhibitor (rapamycin). As shown in Figure 4B, 4C, the suppression of PI3K/Akt/mTOR signaling significantly attenuated senescence induced by α2β1 knockdown in SK-Mel-147 cells. Thus, signals transmitted by PI3K/Akt/mTOR play an important role in senescence induced by α2β1 integrin deprivation.

The role of Akt isozymes in SK-Mel-147 cell senescence induced by α2β1 integrin knockdown

In our studies [27, 29], we showed that the non-canonical effect of Akt-induced signals on the invasion and anoikis of SK-Mel-147 cells deficient in α2β1 was due to the activity of the Akt1 isozyme, while other Akt isoforms did not exhibit non-canonical properties. In the present investigation, we attempted to determine the function of Akt isoforms in senescence of these cells.

To this end, we investigated the effect of specific inhibitors of individual Akt isoforms on senescence of control and a2β1-depleted SK-Mel-147 cells. Figure 5 shows that Akt1- and Akt2-specific inhibitors had no significant effect on the senescence of melanoma cells that sustained a high level of α2β1 expression. In cells depleted of α2β1, the Akt1-specific inhibitor reduced the level of the SA-β-Gal-positive population by about 50%, while inhibition of the Akt2 isoform did not affect senescence.

As noted above, the data on the role of integrins in the development of senescence are contradictory, as evidenced by the fact that the same receptor can both stimulate and suppress the development of the senescent phenotype [15–17]. Signaling mechanisms underlying these opposite cellular reactions are poorly studied. Therefore, it is of interest to compare our results on the role of the Akt-mediated pathway in melanoma cell senescence caused by suppressing α2β1 signaling with the observation that the same pathway transmits signals stimulating lung carcinoma cell senescence caused by radiation-induced activation of α6β4 integrin signaling [16]. It is plausible that this difference is due to the strength of a stress signal produced in one case by the suppression of integrin and in the other by the irradiation of cells. This assumption is supported by the authors’ observation that blocking α6β4 switches the path of the cells from senescence to apoptosis. One would suggest that both α6β4 and α2β1 play a role in protecting cells against a worse scenario; in lung carcinoma cells exposed to radiation, α6β4 protects against apoptosis, while in melanoma cells, α2β1 protects against oncogene-induced senescence.

The involvement of individual Akt isoforms in the mechanisms of cellular senescence is also largely unexplored. Of the three Akt isoforms, only Ak1 knockdown led to damage of telomeres in fibrosarcoma cells, which indicates that the function of this kinase is to override senescence [36].

Akt1-null mouse embryonic fibroblasts appeared more sensitive to UV-induced senescence than Akt2-null cells [39]. These findings indicate that, in different cells, the role of Akt1 in senescence complies with its canonical functions, which are to promote cell proliferation and survival, tumor metastasis, protection against apoptosis, etc. However, as shown in this investigation, in melanoma cells lacking α2β1, the Akt1 isoform performs an opposite, non-canonical function. This observation is consistent with our earlier results on the non-canonical activity of Akt1 that were obtained when studying its effect on anoikis and the invasive activity of SK-Mel-147 cells depleted of α2β1, α3β1, and α5β1 integrins [28–30]. We suggest that knockdown of α2β1 is a stress factor that induces the senescence phenotype in SK-Mel-147 cells via the RAS/PI3K/Akt1/mTORC1/p53/p21 pathway, resulting in a decrease in proliferation, invasive activity, and resistance to anoikis of the cells.

A consequence of the non-canonical activity of Akt is that its inhibition did not lead to the death of tumor cells but instead restored their invasive activity [29]. The non-canonical Akt activity raises the question of the risk of using this and other protein kinases as targets in targeted cancer therapies given that anti-tumor therapy that is based on the suppression of protein kinases (including Akt) is currently broadly employed [40]. Further progress in this field will require more advances in studies devoted to the mechanisms that link cell surface receptors (including integrins) and signaling pathways in various types of tumor cells.

Cell culture and chemicals

SK-Mel-147 human melanoma line was obtained from the Memorial Sloan Kettering Cancer Center (USA). Cells were cultured in DMEM medium containing 10% fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin and incubated at 37° C in an atmosphere with 5% CO₂. p53- and p21-specific siRNA were from Santa Cruz Biotech (USA) Polyclonal antibodies against α2-integrin subunit and monoclonal antibodies to the integrin α2β1 were, respectively, from Chemicon and BD PharMingen (USA). Polyclonal antibodies against protein kinases Akt and Erk and their phosphorylated forms (pAkt Ser473 and pErk Thr202/Tyr204), phospho-mTOR (pmTOR Ser2448), and proteins p53, p21 were from Cell Signaling Tech (USA). PI3K/Akt inhibitor, LY294002; Akt1 inhibitor, XXIII; Akt2 inhibitor, XII; ERK inhibitor, PD98059, and mTOR inhibitor, Rapamycin, were from Calbiochem, and other chemicals were from Sigma.

Transduction of cells with shRNA

Bacterial glycerol clones containing lentiviral plasmid vector pLKO.1-puro with shRNA specific for the α2-integrin subunit and pLKO.1-puro lentiviral vector with scramble shRNA were purchased from Sigma. Lentivirus was produced in HEK293T cells by co-transfection with shRNA-containing or scrambled shRNA-containing vectors with packing plasmids as described earlier [41]. Cells were transduced with lentivirus in the presence of polybrene (8 μg/ml) and selected with puromycin (1–2 μg/ml) for 2–4 days.

siRNA transfection

siRNA duplexes, both control and specific to the p-53 and p-21 proteins, as well as siRNA transfection reagent were from Santa Cruz Biotech. (CA, USA). The procedure was performed as described earlier [42]. Briefly, 1–2 x 10⁵ cells in 1 ml of antibiotic-free DMEM + 10% FBS were plated into each well of 6-well culture plates. After reaching ~50% confluence the cells were transfected with 50 nM siRNA (final concentration) for 70 h. Then cells were harvested with trypsin/EDTA, washed and used for subsequent experiments.

Flow cytometry

1x10⁵ cells were fixed with 70% ethanol, rinsed in PBS, resuspended in 1 ml citrate buffer containing 50 μg/ml propidium iodide and 10 μg/ml RNAse A followed by storing on ice for 3 h. Cell surface expression of α2β1 integrin was assessed by treating the cells with anti-α2β1 (BD PharMingen), followed by staining with FITC-conjugated secondary antibody and fixation with 2% formaldehyde. The cells were analyzed using Becton Dickinson flow cytometer.

Senescence assay

Cells were plated in 12-well plates for 24 h, rinsed in PBS, fixed and incubated overnight at 37° C with the staining solution containing the X-Gal substrate (BioVision, USA). The development of blue color was detected visually under the microscope and the cells staining positive for SA-β-Gal were counted as a percentage of the total cell number. The images were captured and processed using Adobe Photoshop CS5.

DNA replication

DNA replication was assayed by incorporation of EdU (5-ethynyl-2′-deoxyuridine) followed by staining of replicating and total DNA, as described in the manufacturer’s protocol (Invitrogen, USA). Briefly, the cells in logarithmic growth phase were seeded on coverslips, incubated at 37° C for 1 h with 10 μM EdU, fixed with 3.7% formaldehyde, permeabilized by Triton X-100 (0.5%) treatment, stained with Alexa Fluor 488 and Hoechst 33342, and viewed in a luminescence microscope at wavelengths (excitation/emission, nm) 495/519 for Alexa Fluor 488 and 350/461 for Hoechst 33342.

SDS-PAGE and western-blotting

The procedures were performed as described in [40]. 30 μg of cell lysate proteins were run on SDA-PAGE and electroblotted onto a PVDF membrane. After reaction with specific primary antibodies, the membrane was probed with HRP-conjugated secondary antibodies, developed in a ECL detection system (Amersham, England) and imaged on ChemiDoc (Bio-Rad; this device pertains to the equipment of “Human Proteome” Core Facility of the Institute of Biomedical Chemistry). Quantification of protein levels from western blot analysis was done using Image Lab software (Bio-Rad).

Statistical analysis

Differences between the groups were assessed using Student’s t-test and considered significant at p < 0.05.