Kathleen
Barrette
a,
Nele
Zutterman
a,
Sofie
Van Kelst
a,
Charlotte
Proby
b and
Marjan
Garmyn
*a
aLaboratory of Dermatology, Catholic University of Leuven, Leuven, Belgium. E-mail: Marjan.Garmyn@med.kuleuven.be
bThe Wellcome Trust Centre for Molecular Medicine, Ninewells Hospital & Medical School, University of Dundee, Dundee, UK
First published on 23rd May 2012
Previous investigations have demonstrated that isogenic cutaneous squamous cell carcinoma cell lines (cSCC), isolated from highly dysplastic skin (PM1), primary invasive SCC (MET1) and its lymph node metastasis (MET4), show an increasing resistance to cisplatin-induced apoptosis in the increasingly malignant MET1 and MET4 cells. To investigate whether cell death sensitivity in progressive stages of skin carcinogenesis is dependent on the kind of stress we examined the sensitivity of PM1, MET1 and MET4 cells to apoptosis in response to a single UVB-dose (mixture of genotoxic and oxidative stress), or to hydrogen peroxide and hypericin photodynamic treatment (both pure oxidative stresses). MET1 cells, followed by the MET4 cells, were more sensitive to UVB, resulting in more cell death and more apoptosis in comparison with the PM1 cells. A similar pattern of sensitivity was observed when we exposed the SCC cells to hydrogen peroxide or hypericin photodynamic treatment, which both generate mainly oxidative stress. The MET1 cells were the most sensitive to all stresses examined. The pattern of cell death sensitivity in a model of progressive cutaneous squamous cell carcinoma is dependent on the kind of stress. While more advanced skin cancer cells like MET1 and MET4 cells lose their sensitivity to the chemotherapeutic agent cisplatin, they remain sensitive to hydrogen peroxide or physical treatments, which induce major oxidative stress. This differential sensitivity could have implications for the treatment of advanced cSCC.
Using a unique model of cell lines representing different tumor progression stages, previous studies in our laboratory have shown that tumor progression parallels increasing resistance to cisplatin, a therapeutic agent that causes DNA damage with inter- and intrastrand DNA crosslinks, used in the clinic for advanced cSCC.9 In this study we investigated the effect of tumor progression on the cell death response to UVB, since apoptosis is an important protection against UVB-induced mutations, involved in the initiation and progression of carcinogenesis. We tested this effect in the same unique model of isogenetic cell lines representing different tumor progression stages, namely dysplastic skin (PM1), a primary tumor (MET1) and its metastasis (MET4), and show that, in contrast to the sensitivity pattern to cisplatin, the primary cancer cells (MET1) cells are more sensitive to UVB than the PM1 cells.10
To investigate if the cell death response at different tumor progression stages is influenced by the type of stress used, we also tested the response of the cells against the pure oxidative stressors hydrogen peroxide and hypericin photodynamic treatment.
For H2O2 treatments, cells were rinsed twice with PBS and respectively incubated for the indicated time period with 1 mM H2O2 (dissolved in the cell culture medium).
For the photodynamic therapy, SCC cells were pre-incubated with hypericin for 24 hours, in subdued light conditions (<1 μW cm−2) and subsequently irradiated (2 J cm−2) in PBS by placing the samples on a plastic diffuser sheet 5 cm above a set of seven L18W30 fluorescent lamps (Osram; maximal emission between 530 and 620 nm, coinciding with the absorption peak of hypericin at 595 nm). At the surface of the diffuser, the uniform fluence rate was 4.5 mW cm−2, as measured with an IL 1400 radiometer (International Light, Newburyport, MA).
24 hours after UVB irradiation (120 mJ cm−2) the cancerous MET1 and MET4 cells show morphologically the most cell death compared to the precancerous PM1 cells (Fig. 1a). UVB decreased cell viability (Fig. 1b) and metabolic activity (Fig. 1c) in all cell lines, with the decrease most pronounced in the primary cancer cells (MET1) and progressively less in, respectively, MET4 and PM1 cells. Since apoptosis is known to be the main cell death modality following UVB stress in keratinocytes, we investigated whether the detected decrease in viability and metabolic activity in the cSCCs after UVB irradiation was due to the induction of apoptosis.6 To this end we measured PARP cleavage and caspase-3 activation, as biochemical markers of apoptosis.
Fig. 1 Cell death response after UVB irradiation differs between progressive/consecutive cSCC cell lines. (a) Morphologic representation of cell death 24 hours after 120 mJ cm−2 UVB. Arrows indicate dead cells. Arrows were drawn proportional to the amount of floating dead cells present. The criteria used to evaluate the morphologic representation of cell death are the amount of floating cells and the amount of cells which display membrane blebbing.23 (b) 24 hours after exposure to a single UVB dose (120 mJ cm−2), cell viability was assessed by trypan blue exclusion assay. A representative graph of at least 3 independent experiments is shown. Columns show mean of an experiment performed in duplicate, bars denote standard deviation; * p < 0.05. Numbers given are percentages of control. (c) PM1, MET1 and MET4 were exposed to a single UVB dose (120 mJ cm−2). Metabolic activity was analyzed by MTT assay and was expressed relatively to the untreated conditions (percent of control). A representative graph of at least 3 independent experiments is shown. Columns show mean of an experiment performed in triplicate, bars denote standard deviation; * p < 0.05. Metabolic activity of PM1, MET1 and MET4 after UVB irradiation was significantly less compared to control (not shown in picture). (d) SCC cells were UVB irradiated (120 mJ cm−2) and cell lysates were collected after 24 hours and were analyzed by immunoblotting with antibodies against PARP, cleaved caspase-3. Percentage of PARP cleavage (pixel total of cleaved product/(pixel total of total PARP + pixel total of cleaved product)) and total pixel of cleaved caspase-3 (pixel total of cleaved caspase 3/pixel total of actin) was calculated using UnScanIt. Actin was used a loading control. A representative blot of at least 3 independent experiments is shown. |
An earlier activation of the apoptotic machinery was detected after 120 mJ cm−2 UVB-irradiation in the MET1 cells (51% PARP cleavage after 6 hours) in comparison with the precancerous PM1 and the MET4 cells (PARP cleavage after 16 hours), indicating that the MET1 cells are the most sensitive to UVB stress, which is in concordance with the data obtained by MTT assay and trypan blue exclusion assay. Complete PARP cleavage was accomplished after 16 hours in the MET1 cells, whereas MET4 cells did not achieve complete PARP cleavage after 24 hours. These data are in agreement with the data obtained on the metabolic activity of the surviving cells and cell death level (MTT and trypan blue assay) suggesting that the main cell death modality is apoptosis.
We can conclude that the primary cancer cells (MET1) and, to a lesser extent, the MET4 cells, are more sensitive to UVB induced cell death than the precancerous PM1 cells, with an earlier activation of the apoptotic machinery. This is in contrast with previous findings, demonstrating that PM1 are most sensitive to CDDP (ESI‡) and that sensitivity to CDDP-induced cell death diminishes as the tumor progresses.9 To investigate whether the difference in cell death response in PM1, MET1 and MET4 is dependent on the stressor we investigated the response of the cells to two additional stressors which cause only oxidative stress and thus indirect DNA damage: H2O2 and hypericin photodynamic therapy.
Fig. 2 Primary SCCs (MET1), followed by metastatic MET4 cells, are more sensitive to H2O2 than precancerous PM1. PM1, MET1 and MET4 cells were treated with 1 mM of hydrogen peroxide (H2O2). (a) Morphologic representation of cell death 24 hours after treatment with H2O2. Arrows indicate dead cells. Arrows were drawn proportional to the amount of floating dead cells present. The criteria used to evaluate the morphologic representation of cell death are the amount of floating cells and the amount of cells which display membrane blebbing.23 (b) Cell viability was assessed by trypan blue exclusion assay after 24 hours of 1 mM H2O2 exposure. A representative graph of at least 3 independent experiments is shown. Columns show mean of an experiment performed in duplicate, bars denote standard deviation; * p < 0.05. Numbers given are percentages of control. (c) Metabolic activity after 24 hours, as measured by MTT assay. A representative graph of at least 3 independent experiments is shown. The metabolic activity was expressed relative to the untreated controls. Columns show mean of an experiment performed in triplicate, bars denote standard deviation; * p < 0.05. Metabolic activity of PM1, MET1 and MET4 cells treated with 1 mM H2O2 was significantly less compared to control (not shown in picture). (d) Lysates were collected at the indicated timepoints and analyzed for PARP and caspase-3 cleavage by Western blot analysis. Actin was used to confirm equal loading of proteins. Percentage of PARP cleavage (pixel total of cleaved product/(pixel total of total PARP + pixel total of cleaved product)) and total pixels of cleaved caspase-3 (pixel total of cleaved caspase 3/pixel total of actin) was calculated using UnScanIt. A representative blot of at least 3 independent experiments is shown. |
Fig. 3 Primary SCCs (MET1), followed by MET4 cells are more sensitive to PDT than precancerous PM1 cells. PM1, MET1 and MET4 cells were exposed to a hypericin-mediated photodynamic therapy by treating the cells 24 hours before illuminating (2 J cm−2) with hypericin (90 nM). (a) Morphologic representation of cell death 24 hours after photodynamic therapy (PDT). Arrows indicate dead cells. Arrows were drawn proportional to the amount of floating dead cells present. The criteria used to evaluate the morphologic representation of cell death are the amount of floating cells and the amount of cells which display membrane blebbing.23 (b) Cell viability was assessed by trypan blue exclusion assay 24 hours after PDT. A representative graph of at least 3 independent experiment is shown. Columns show mean of an experiment performed in duplicate, bars denote standard deviation; * p < 0.05. Numbers given are percentages of control. (c) Metabolic activity was measured after 24 hours by MTT assay. A representative graph of at least 3 experiments is shown. The metabolic activity was expressed relatively to the untreated controls. Columns show mean of an experiment performed in triplicate, bars denote standard deviation; * p < 0.05. Metabolic activity of PM1, MET1 and MET4 after PDT was significantly less compared to control (not shown in picture). (d) Lysates were collected at the indicated time points and analyzed for PARP and caspase-3 cleavage by Western blot analysis. Actin was used to confirm equal loading of proteins. Percentage of PARP cleavage (pixel total of cleaved product/(pixel total of total PARP + pixel total of cleaved product)) and total pixels of cleaved caspase-3 (pixel total of cleaved caspase 3/pixel total of actin) was calculated using UnScanIt. A representative blot of at least 3 independent experiments is shown. |
In this study, we have shown that MET1 cells are the most sensitive to UVB, H2O2 and hypericin PDT induced cell death, followed by the MET4 cells. This is in contrast with the observed sensitivity of the progressive stages of cSCC whereby the dysplastic PM1 cells are the most sensitive and the advanced SCC were most resistant to cisplatin.9
Different factors could be responsible for the loss in cell death sensitivity to cisplatin in cells representative of progressive skin tumorigenesis, including reduced uptake, increased inactivation and increased efflux of cisplatin, which could not be the case with the physical stressor UVB and hydrogen peroxide.13 Cisplatin is a genotoxic agent inducing direct DNA damage by adduct formation. Increased adduct repair and molecular mechanisms, such as activation of the AKT pathway, that inhibit propagation of DNA damage signals that the apoptotic machinery could further be responsible for loss in cell death response.14,15 Of interest in a former study we demonstrated in this model that skin tumor progression parallels enhanced AKT activation and increased resistance to cisplatin-induced apoptosis, while AKT inhibition sensitizes the apoptosis-resistant MET1 and, to a lesser extent, MET4 cells to cisplatin-mediated cell death.9
UVB not only causes direct DNA damage but UVB photons are also absorbed by other chromophores in the cell, resulting in the formation of radicals and reactive oxygen species. We hypothesize that the observed difference in the sensitivity pattern of the cSCC cells detected in cisplatin-treated versus UVB-irradiated cells may be at least partly due to the distinct response of the cells to oxidative versus genotoxic stress, because UVB generates oxidative and genotoxic stress whereas cisplatin predominantly generates genotoxic stress. This hypothesis is supported by our observation that the same pattern of sensitivity (maximal sensitivity in the MET1 cells) is observed when cells are treated with two oxidative stressors, hydrogen peroxide (H2O2) and hypericin photodynamic therapy.
Cells may die by apoptosis or via another form of cell death. After the treatment with H2O2 or hypericin PDT, MET4 cells must have died, at least in part, by another cell death modality since the MET4 cells display excess cell death (morphologically, MTT and TB assay) compared to the observed PARP and caspase-3 cleavage. The type of cell death initiated by UVB irradiation in the MET4 cells is probably apoptosis, since the observed PARP and caspase-3 cleavage concur with the observed cell death (morphologically, MTT and TB-assay). One explanation for the different cell death patterns in MET4 cells in the context of different stressors could be the formation of different forms of ROS (radical versus non-radical) since both oxidative stressors (H2O2 and hypericin PDT) form mostly non-radical ROS, while UVB irradiation leads to the formation of radical and non-radical forms of ROS which are mainly O2 and H2O2.8 The observed alternative cell death could be mediated by a caspase-independent pathway, like necrosis.
Notwithstanding the different forms of ROS formation by UVB irradiation versus H2O2 treatment and hypericin-PDT,16 the sensitivity pattern towards cell death remains the same with the MET1 cells displaying the most cell death, followed by the MET4 cells.
A second explanation for the alternative cell death evident in MET4 cells could be the difference between mainly oxidative stress (H2O2 and hypericin PDT) compared to the UVB stress, which causes oxidative but also genotoxic stress.
Our hypothesis that the observed difference in sensitivity of cSCC cells treated with cisplatin versus UVB is due to the distinct response of the cells to oxidative versus genotoxic stress is further supported by our observation that the same pattern of sensitivity (maximal sensitivity in the MET1 cells) is observed when cells are treated with two oxidative stressors, hydrogen peroxide (H2O2) and hypericin photodynamic therapy, which can however also induce indirect DNA damage. However the underlying mechanisms need to be further investigated.
We could exclude the hypothesis that sensitivity to the different stressors may be a result of differential proliferative rate between the cells since the doubling time is almost the same in het PM1 (34 hours), MET1 (30 hours) and MET4 (29 hours) cells. We also showed that the sensitivity of the PM1, MET1 and MET4 cells is different according to the type of stressor which may indicate that the sensitivity is dependent on the kind of stressor regardless the doubling time of the cells.
Sensitization of cells to ROS-mediated cell death via AKT, as an explanation for the difference in sensitivity pattern between CDDP and UVB, as previously shown in other cells,17 is probably not the mechanism involved in our experimental setting. Although AKT activation is increased in MET1 and MET4,9 we could demonstrate that AKT inhibition sensitizes our cells to UVB-induced cell death (data not shown).
The strength of this study is the use of a unique isogenic tumor progression model in which we can investigate the response of the cells to stress, excluding variations due to interindividual factors. A limitation is the rareness of this model.
Cancer research is focused on identifying new strategies for treating cancer and especially aggressive metastatic cancers. Metastatic cSCCs have a very poor prognosis due to a high recurrence rate after surgery and radiotherapy and resistance to chemotherapy. Transplantation patients who undergo chronic immunosuppression have a major risk to develop these advanced cSCCs with a high cancer burden and a significant risk of death.18 Irrespective of the underlying mechanisms involved, we were able to confirm our hypothesis that the pattern of cell death sensitivity in a model of progressive skin tumor progression is similar in response to agents which cause major oxidative stress (UVB, H2O2 and PDT). This could argue for the use of oxidative stress in chemotherapy, and is further supported by previous findings showing that oxidative stress also exerts antitumorigenic actions as it has been linked to senescence and apoptosis, two major mechanisms that counteract tumor development.19 More advanced cancer cells have a higher metabolic activity rate (Warburg effect) and this could easily translate into higher rates of formation of ROS such as H2O2, features that could be important in the development of new anticancer strategies.20,21 Cancer cells often compensate with overactivation of antioxidant systems, such as superoxide dismutases (SOD), but an overload of oxidative stress could induce more rapid cell death in cancer cells compared to normal cells. Cancer cells with increased oxidative stress are likely to be more vulnerable to damage by further ROS insults induced by exogenous agents compared to normal cells.21 There is experimental evidence that cancer cells are more susceptible to H2O2-induced cell death than normal cells when the concentration of H2O2 becomes toxic.20
However, a low concentration of oxidative stress may have several protumorigenic effects like increasing DNA mutation rate, inducing indirect DNA damage, genome instability and cell proliferation as a low H2O2 concentration.20 Also under persistent intrinsic oxidative stress, many cancer cells become well-adapted to such stress and develop an enhanced endogenous antioxidant capacity, which makes the malignant cells resistant to exogenous stress. Increased ROS stress below a toxic dose correlates in cancer cells with the aggressiveness of tumours and poor patient outcome.21 ROS, such as low levels of H2O2, also appear crucial in regulating several normal cellular functions, whereby regulated changes in ROS concentration has been found critical for functions like respiration, control of enzymatic activities, transcriptional regulation, modulation of signal transduction pathways, cell cycle progression and inflammatory response, resulting in survival of the cancer cells.20
So we believe that the toxicity of ROS is concentration dependent, with low levels of ROS stimulating carcinogenesis, whereas high levels of ROS are cytotoxic to the cancer cell.
Because advanced cancer cells such as the primary (MET1) and metastatic (MET4) SCC cells are more sensitive to oxidative stressors compared to the dysplastic cells, oxidative stress could have therapeutic implications in the treatment of advanced SCC, especially in cells which have lost their sensitivity to genotoxic strategies like cisplatin. Direct administration of H2O2 to cancer patients is not an appropriate therapy. There is increasing evidence that raising the cellular levels of H2O2 using H2O2-generating systems or by inhibition of ROS scavenging may be an efficient way of killing cancer cells. Mangafodipir, which increases H2O2 levels specifically in cancer cells, is undergoing clinical trial in a phase II studies in combination with chemotherapy in liver cancer. Chemotherapy like cisplatin is able to inhibit thioredoxin, which results in disrupted ROS scavenging. So combining cisplatin and thioredoxin inhibitors (such as flavanols) in synergistic combinations may prove to be the most useful approach.22
Footnotes |
† This article is published as part of a themed issue on current topics in photodermatology. |
‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/c2pp25064k |
This journal is © The Royal Society of Chemistry and Owner Societies 2013 |