TL;DR THC is effective and shows promise in causing melanoma cancer cell death, in addition 1:1 ration of THC:CBD seems to be equally effective.
Jane L Armstrong1,2, David S Hill1, Christopher S McKee1, Sonia Hernandez-Tiedra3, Mar Lorente3, Israel Lopez-Valero3,4, Maria Eleni Anagnostou1, Fiyinfoluwa Babatunde1, Marco Corazzari5, Christopher P F Redfern6, Guillermo Velasco3,4,7 and Penny E Lovat1,7
Jane L Armstrong1,2, David S Hill1, Christopher S McKee1, Sonia Hernandez-Tiedra3, Mar Lorente3, Israel Lopez-Valero3,4, Maria Eleni Anagnostou1, Fiyinfoluwa Babatunde1, Marco Corazzari5, Christopher P F Redfern6, Guillermo Velasco3,4,7 and Penny E Lovat1,7
- 1Dermatological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle-upon-Tyne, UK
- 2Faculty of Applied Sciences, University of Sunderland, Sunderland, UK
- 3Department of Biochemistry and Molecular Biology I, School of Biology, Complutense University, Madrid, Spain
- 4Instituto de Investigaciones Sanitarias San Carlos (IdISSC), Madrid, Spain
- 5Department of Biology, University of Rome “Tor Vergata”, Rome, Italy
- 6Northern Institute for Cancer Research, Newcastle University, Newcastle-upon-Tyne, UK
Correspondence:
Penny E. Lovat, Dermatological Sciences, Institute of Cellular
Medicine, The Medical School, Newcastle University, Framlington Place,
Newcastle-upon-Tyne NE2 4HH, UK. E mail: penny.lovat@ncl.ac.uk
7These are joint senior authors.
Received 5 November 2014; Revised 9 January 2015; Accepted 21 January 2015
Accepted article preview online 10 February 2015; Advance online publication 12 March 2015
Accepted article preview online 10 February 2015; Advance online publication 12 March 2015
Top of page
. Novel treatment strategies are therefore urgently required, particularly for patients bearing BRAF/NRAS
wild-type tumors. Targeting autophagy is a means to promote cancer cell
death in chemotherapy-resistant tumors, and the aim of this study was
to test the hypothesis that cannabinoids promote autophagy-dependent
apoptosis in melanoma. Treatment with Δ9-Tetrahydrocannabinol
(THC) resulted in the activation of autophagy, loss of cell viability,
and activation of apoptosis, whereas cotreatment with chloroquine or
knockdown of Atg7, but not Beclin-1 or Ambra1, prevented THC-induced
autophagy and cell death in vitro. Administration of Sativex-like
(a laboratory preparation comprising equal amounts of THC and
cannabidiol (CBD)) to mice bearing BRAF wild-type melanoma xenografts
substantially inhibited melanoma viability, proliferation, and tumor
growth paralleled by an increase in autophagy and apoptosis compared
with standard single-agent temozolomide. Collectively, our findings
suggest that THC activates noncanonical autophagy-mediated apoptosis of
melanoma cells, suggesting that cytotoxic autophagy induction with
Sativex warrants clinical evaluation for metastatic disease.Abstract
Although
the global incidence of cutaneous melanoma is increasing, survival
rates for patients with metastatic disease remain <10 class="mb" span="">%
Abbreviations:
ANOVA,
analysis of variance; BDS, botanical drug substance; CBD, cannabidiol;
ER, endoplasmic reticulum; siRNA, small interfering RNA; THC, Δ9-Tetrahydrocannabinol
Top of page
Introduction
Cutaneous
melanoma incidence continues to increase, and response rates of
patients with metastatic melanoma to current therapy remain poor (Garbe et al., 2011). Identification of driver mutations and development of targeted therapies to BRAF/MEK have revolutionized melanoma therapy, although clinical resistance is inevitable (Chen and Davies, 2014). The emergence of immunotherapies that are able to promote tumor T-cell responses is further changing melanoma management (Wolchok et al., 2010); however, not all patients respond (Prieto et al., 2012).
It is therefore clear that there is no consistently beneficial
treatment for metastatic melanoma and alternative approaches should be
explored.
Autophagy (macroautophagy) is the principal
lysosomal-mediated mechanism for the degradation of damaged or
long-lived organelles and proteins. Under physiological conditions,
autophagy maintains normal turnover of cellular components, as well as
responding to metabolic stress, whereas in pathological settings
autophagy activation mediates defense against extracellular insults and
pathogens (Choi et al., 2013).
The current model for the role of autophagy in cancer is that in the
early stages of tumor development, quality control by autophagy inhibits
tumorigenesis, whereas in advanced cancer autophagy provides energy to
meet the increased demands and a means to resist cell death caused by
cytotoxic therapy (White, 2012).
Preclinical data suggest that lysosomal inhibition can cause tumor
regression, and the lysosome inhibitors chloroquine or
hydroxychloroquine are now being evaluated in clinical trials either
alone or in combination with chemotherapy (Yang et al., 2011). However, recent studies suggest that chloroquine/hydroxychloroquine treatment may promote tumor development (Michaud et al., 2011; Maycotte et al., 2012),
questioning the benefit of autophagy inhibition. An alternative
approach for autophagy modulation is via exacerbation; although
initially this appears counterintuitive to treat advanced cancer, recent
evidence suggests that in particular circumstances a consequence of
autophagy activation is cell death (Ding et al., 2007; Scarlatti et al., 2008; Tomic et al., 2011; Basit et al., 2013).
Therapeutic exploitation of cytotoxic autophagy to drive cancer cell
death is therefore an emerging concept for the development of novel
cancer treatments.
Cannabinoids are a diverse class of compounds derived from Cannabis sativa, with Δ9-tetrahydrocannabinol (THC) the most relevant owing to its high potency and abundance (Pertwee, 2008).
THC exerts its biological effects by mimicking endocannabinoids that
bind to and activate two G protein–coupled cannabinoid receptors: CB1
and CB2 (Howlett et al., 2002). CB1/CB2 receptors are expressed in many cancer cell types (Velasco et al., 2012),
and cannabinoids are currently being investigated as anticancer agents,
including glioblastoma for which THC has shown considerable promise (Velasco et al., 2012).
Preclinical data demonstrate that THC exerts its antitumoral action via
induction of endoplasmic reticulum (ER) stress, upregulation of the
transcriptional coactivator p8 and the pseudo-kinase tribbles homolog 3
(TRIB3), the stimulation of autophagy, and execution of apoptosis (Carracedo et al., 2006; Salazar et al., 2009a, b).
Blockade of autophagy prevents THC-induced apoptosis and cell death,
indicating that autophagy is upstream of apoptosis and demonstrating the
potential of p8/TRIB3-mediated autophagy as a cytotoxic pathway.
Alongside
genetic changes, adaption to ER stress and the aberrant control of
autophagy have emerged as key drivers of malignancy and therapy
resistance in advanced melanoma (Armstrong et al., 2011; Corazzari et al., 2013). The cannabinoid receptors have previously been identified as potential therapeutic targets (Blazquez et al., 2006; Carracedo et al., 2006);
hence, targeting ER stress responses combined with cytotoxic autophagy
using cannabinoids may represent a valuable therapeutic approach for
metastatic melanoma. The aim of this study was to determine whether THC
activates cytotoxic autophagy in melanoma cells in vitro and in vivo.
Our data suggest a noncanonical mechanism of autophagy-mediated
apoptosis, highlighting the potential to harness autophagy for
therapeutic benefit in advanced melanoma.
Top of page
42=3.289, P=0.002) were observed in A375 and CHL-1 cells in response to THC treatment, indicating activation of autophagic flux (Figure 1).
Results
THC activates autophagy and apoptosis in melanoma cells
Characteristic
features of early and late stages of autophagy were assessed using LC3
lipidation (LC3-II) and analysis of autophagic flux using chloroquine
and visualization of tandem mRFP–GFP (monomeric red fluorescent
protein–green fluorescent protein)–tagged LC3 (Kimura et al., 2007),
respectively. LC3-II induction was observed in three human melanoma
cell lines in response to THC, and LC3-II accumulated further in the
presence of chloroquine in both BRAF wild-type (CHL-1) and mutated (A375
and SK-MEL-28) melanoma cells. In addition, increased numbers of
LC3-positive autophagosomes (yellow puncta) and autolysosomes (red
puncta) and total red fluorescence (CHL-1: t43=4.17, P<0 .001="" a375:="" i="">t
Figure 1.
The Δ9-Tetrahydrocannabinol (THC) induces autophagy in melanoma cells. (a–c) A375, SK-MEL-28, and CHL-1 cells were treated with vehicle or THC (4.5 or 5 μM) for 24 hours in the presence or absence of chloroquine (CQ; 10 μM) for the final 2 hours,
and LC3 and β-actin expression was determined by western blotting.
LC3-II expression was quantified and band intensity normalized to
β-actin. Data are expressed as fold change relative to the mean LC3-II/β-actin value for a representative experiment and are shown above the western blot (representative data from n=3 independent experiments). (d, e) mRFP–GFP–LC3 expressing (d) A375 or (e) CHL-1 cells were treated with vehicle or THC (4.5 μM) for 18 hours. Data are representative fluorescent micrographs (bar=20 μm) of three independent experiments. (f)
Total red fluorescence values were generated from ≥20 cells per
treatment condition, from two independent experiments. Pixel intensities
were divided by a factor of 106, and data are shown as mean±SD (**P<0 .01="" and="" class="mb" span="">*
THC
reduced melanoma cell viability in a dose-dependent manner while having
little effect on primary melanocytes at doses up to 6 μM THC (Figure 2a).
Cotreatment of melanoma cells with submaximal concentrations of THC and
the pan-caspase inhibitor ZVAD-fmk significantly increased cell
viability compared with treatment with THC alone (CHL-1: t22=3.962, P=0.0007; A375: t16=3.74, P=0.0018; Figure 2b), suggesting that cell death is caspase dependent. In addition, the cytochrome c–labeled structures present in vehicle-treated cells were substantially reduced in THC-treated cells (Figure 2c), indicating apoptosis activation.
Figure 2.
The Δ9-Tetrahydrocannabinol (THC) induces apoptosis of melanoma cells. (a) Primary melanocytes, A375, and CHL-1 cells were treated with vehicle or THC (3–10 μM) for 24 hours. (b) A375 (i) or CHL-1 (ii) cells were treated with vehicle or THC (5 μM) in the presence or absence of ZVAD (20 μM) for 24 hours.
Cell viability was determined by the
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)
assay. Data generated in triplicate were expressed relative to the mean
of vehicle-treated cells in each experiment, for three independent
experiments, and shown as mean±SD (t-test; **P<0 .01="" and="" class="mb" span="">*THC-induced apoptosis is dependent on autophagy
In glioma, THC activates apoptosis via a mechanism involving TRIB3 and Beclin-1-dependent autophagy (Carracedo et al., 2006; Salazar et al., 2009b).
The small interfering RNA (siRNA)-mediated knockdown of TRIB3
significantly prevented loss of cell viability in response to THC in
A375 cells (one-way analysis of variance (ANOVA); F5, 48=5.053, P=0.001; THC compared with vehicle treatment in siCtrl cells; Tukey’s P<0 .05="" a="" href="http://www.nature.com/jid/journal/v135/n6/suppinfo/jid201545s1.html">Supplementary Figure S1
online), demonstrating that TRIB3 mediates THC-induced cell death.
Furthermore, siRNA-mediated knockdown of Atg7 prevented THC-induced
accumulation of LC3-II in the presence of chloroquine (one-way ANOVA; F11, 24=6.878, P<0 .001="" chloroquine="" class="mb" compared="" span="" with="">+THC-treated siCtrl cells contrast P<0 .05="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig3">Figure 3a and b). THC-induced caspase 3 cleavage was also inhibited by Atg7 knockdown in A375 (Figure 3a), as well as in CHL-1 and SK-MEL-28 cells (CHL-1: one-way ANOVA; F5, 12=6.57, P=0.004; THC compared with vehicle-treated siCtrl cells; Tukey’s P<0 .05="" anova="" f="" one-way="" sk-mel-28:="" sub="">3, 8=4.646, P=0.037; THC compared with vehicle-treated siCtrl cells; Tukey’s P<0 .05="" a="" href="http://www.nature.com/jid/journal/v135/n6/suppinfo/jid201545s1.html">Supplementary Figure S2
online). In addition, THC treatment resulted in a significant loss of
melanoma cell viability only in the absence of chloroquine or Atg7 siRNA
(Figure 3 and Supplementary Figure S2 online; post hoc tests, Games–Howell, or Tukey’s P<0 .01="" a375="" a="" alone="" and="" atg7="" cell="" cells="" chl-1="" downregulation="" effect="" had="" however="" i="" in="" knockdown="" loss="" no="" of="" on="" resulted="" s="" significant="" sk-mel-28="" ukey="" viability="">P<0 .001="" a="" href="http://www.nature.com/jid/journal/v135/n6/suppinfo/jid201545s1.html">Supplementary Figure S2a
online), suggesting that basal autophagy is required to maintain
viability in these cells. Collectively, these data demonstrate that
THC-induced apoptosis of melanoma cells requires TRIB3 and is mediated
by Atg7-dependent autophagy.Figure 3.
The Δ9-Tetrahydrocannabinol (THC)-induced apoptosis requires autophagy. (a–c)
A375 cells were transfected with small interfering RNAs (siRNAs) for
Atg7 (siAtg7) or with a nonsilencing control siRNA (siCtrl) before
treatment with vehicle or THC (4.5, 5 μM) for 24 hours in the presence or absence of chloroquine (CQ; 10 μM) for the (a, b) final 2 hours or for (c) 24 hours. (a, b)
Atg7, LC3, cleaved caspase 3, and β-actin expression were determined by
western blotting. LC3-II expression was quantified and band intensity
normalized to β-actin. Data are expressed as fold change relative to the
mean LC3-II/β-actin value for each experiment, for three separate experiments (mean±SD, n=3). (c)
Cell viability was determined by the
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)
assay. Data generated in triplicate were expressed relative to the mean
of vehicle-treated siCtrl cells in each experiment, for three
independent experiments, and shown as mean±SD (*P<0 .05="" class="mb" span="">*
Beclin-1 promotes autophagy induction and autophagosome formation (Russell et al., 2013); however, Beclin-1-independent autophagy has been reported (Scarlatti et al., 2008; Grishchuk et al., 2011). Unlike in glioma cells, Beclin-1 knockdown did not prevent THC-induced LC3-II accumulation or caspase 3 cleavage (t4=4.494, P=0.011; Supplementary Figure S3a online) in A375 cells (Figure 4a),
and THC treatment resulted in a significant loss of cell viability
under both control and Beclin-1 knockdown conditions (Welch ANOVA; F5, 19.63=94.53, P<0 .001="" cells="" compared="" games="" i="" owell="" shrna="" thc="" treatments="" vehicle-treated="" with="">P
<0 .001="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig4">Figure 4a i) and ii)).
These results suggest that autophagy and subsequent apoptosis occur
independently of Beclin-1. Furthermore, knockdown of the
Beclin-1-interacting protein Ambra1 failed to prevent THC-induced LC3-II
induction and caspase 3 cleavage (t4=6.097, P=0.004; Supplementary Figure S3b online and Figure 4b).
The effect of Ambra1 knockdown alone on A375 viability was variable;
however, THC treatment resulted in a significant loss of cell viability
under both control (t16=9.44, P<0 .001="" ambra1="" and="" conditions="" i="" knockdown="">t16=10.61, P<0 .001="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig4">Figure 4biii).
Collectively, these data suggest that THC activates noncanonical
autophagy-mediated apoptosis of melanoma cells that is dependent on Atg7
but not Beclin-1 or Ambra1.Figure 4.
The Δ9-Tetrahydrocannabinol (THC)-induced autophagy and cell death is not dependent on Beclin-1 or Ambra1. (a)
A375 cells stably expressing short hairpin RNA (shRNA) for Beclin-1
(shBeclin1) or a nonsilencing control shRNA (shCtrl) were treated with
vehicle or THC (4.5 and 5 μM) for 48 hours in the presence or absence of chloroquine (CQ; 10 μM) for the final 2 hours
(i). Beclin-1, LC3, cleaved caspase 3, and β-actin expression were
determined by western blotting (i). Cell viability (ii) was determined
by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
(MTT) assay. Data generated in triplicate were expressed relative to the
mean of vehicle-treated shCtrl cells in each experiment, for three
independent experiments, and shown as mean±SD (***P<0 .001="" b="" cells="" shctrl="" vehicle-treated="" vs.="">bCannabinoid treatment stimulates autophagy and apoptosis and the abrogation of melanoma growth in vivo
A
1:1 mixture of submaximal doses of THC–botanical drug substance (BDS)
and the nonpsychoactive cannabinoid cannabidiol (CBD) –BDS, a laboratory
mimic of the clinical cannabinoid Sativex (Sat-L) (an oromucosal spray
of standardized cannabis extract comprising equal amounts of THC and CBD
(GW Pharmaceuticals)), reduced glioma growth in vivo to the same extent as an identical dose of THC (Torres et al., 2011). Treatment of melanoma cells with THC+CBD resulted in a substantial loss of melanoma cell viability at a concentration of 1 μM THC+1 μM
CBD in CHL-1, A375, and SK-MEL-28 cells compared with equivalent
concentrations of THC, whereas temozolomide had little effect (Figure 5). Temozolomide is an alkylating agent currently indicated as standard single-agent chemotherapy for metastatic melanoma. The in vivo
relevance of these findings was evaluated in the context of BRAF
wild-type melanoma tumors, for which there is a particular demand for
novel therapeutic approach in the absence of targeted therapies in this
tumor group. CHL-1 xenograft tumors were treated for 20 days with
temozolomide, THC, or Sativex-L. Both THC and Sativex-L significantly
inhibited the growth of xenografts (one-way ANOVA; F3, 16=9.347, P=0.001; THC or Sat-L compared with vehicle, Sat-L compared with Temozolomide: P<0 .05="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig6">Figure 6a
).
Tumors removed from animals at the time of killing were processed for
immunohistochemical analysis of proliferative activity (Ki67), apoptosis
(TUNEL), and autophagy (LC3). Ki67 fluorescence differed significantly
between drug treatments (Welch ANOVA; F3, 16.28=61.363, P<0 .001="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig6">Figure 6b)
and was significantly reduced in tumors from animals treated with
temozolomide, THC, or Sativex-L compared with control animals
(Games–Howell; P≤0.001), as well as in tumors from animals
treated with THC or Sativex-L compared with those treated with
temozolomide (Games–Howell; P<0 .001="" also="" anova="" between="" differed="" f="" fluorescence="" one-way="" significantly="" sub="" treatments="" tunel="">3, 32=13.31, P<0 .001="" and="" animals="" control="" from="" higher="" i="" in="" or="" s="" sativex-l="" than="" thc="" treated="" tumors="" ukey="" was="" with="">P≤0.001) and higher in tumors from animals treated with Sativex-L compared with those treated with temozolomide (Tukey’s P<0 .05="" a="" href="http://www.nature.com/jid/journal/v135/n6/full/jid201545a.html#fig6">Figure 6c). Correspondingly, LC3 fluorescence differed between treatments (one-way ANOVA; F3, 32=3.539, P<0 .05="" animals="" compared="" from="" i="" in="" increased="" lc3="" or="" s="" sativex-l="" significantly="" staining="" temozolomide="" treated="" tumors="" ukey="" vehicle="" with="">P≤0.05, Figure 6d).
Staining for Ki67, TUNEL, and LC3 was not significantly different in
tumors from animals treated with THC compared with Sativex-L.
Collectively, these data suggest that THC and Sativex-L are more
effective than temozolomide in terms of apoptosis induction and
antitumor response, further validating the therapeutic relevance of
cannabinoid treatment for melanoma.Figure 5.
Cannabinoids inhibit melanoma cell viability in vitro. (a–c) CHL-1 (a), A375 (b), or SK-MEL-28 (c) cells were treated with temozolomide (Temo; 10–50 μM), Δ9-Tetrahydrocannabinol (THC; 1–5 μM), or THC+CBD (0.5 μM THC+0.5 μM CBD to 2.5 μM THC+2.5 μM CBD) for 48 hours.
Cell viability was determined by the
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)
assay. Data generated in triplicate were expressed relative to the mean
of vehicle-treated cells for each drug treatment in each experiment, for
three independent experiments, and shown as mean±SD for a
representative experiment. CBD, cannabidiol.Full figure and legend (109K)Download Power Point slide (227 KB)
Figure 6.
The Δ9-Tetrahydrocannabinol (THC) and Sativex-like cannabinoids promote autophagy and antitumor responses in melanoma xenografts. Athymic nude mice were injected subcutaneously in the right flank with CHL-1 melanoma cells. When tumors reached a 250 mm3 size, mice were treated daily for 20 days with vehicle, temozolomide (TMZ; 5 mg kg−1; local administration), THC (15 mg kg−1; oral administration), or Sativex-like (Sat-L; 7.5 mg kg−1 THC–BDS+7.5 mg kg−1 CBD–BDS; oral administration). (a)
Tumor volumes were measured daily. Each point is the mean from ≥5
tumors±SD and is expressed relative to the tumor volume on day 1 of
treatment. (b–d) Immunohistochemical analysis of CHL-1 xenograft tumors treated with temozolomide, THC, or Sat-L. (b–d) Micrographs of tumor sections stained for (b) Ki67, (c) TUNEL, or (d)
LC3. In each, green is staining for Ki67, TUNEL, or LC3, and red is the
TO-PRO-3 counterstain. The bar graphs below each set of micrographs
summarize the data analysis from tumor sections (bar=20 μm). Total fluorescence values (total fluorescence (LC3/Ki67/TUNEL)/total
nuclear fluorescence) were generated in triplicate for 3 tumors in each
treatment group. Data are expressed as fold change relative to the mean
value obtained from control animals, from two independent staining
analyses, and shown as mean±SD (**P<0 .01="" and="" class="mb" span="">*
Top of page
Discussion
It
is apparent that autophagy modulation may offer considerable benefit in
cancer treatment; however, potential drawbacks to autophagy inhibition
have recently been identified (Michaud et al., 2011; Maycotte et al., 2012; Takahashi et al., 2012; Rosenfeldt et al., 2013).
Emerging evidence indicates that activation of autophagy can, in some
circumstances, promote cell death; stimulation of cytotoxic autophagy
therefore represents an alternative approach to autophagy modulation (Ding et al., 2007; Scarlatti et al., 2008; Salazar et al., 2009b; Grishchuk et al., 2011; Tomic et al., 2011; Basit et al., 2013).
Here, we show that the cannabinoid THC exerts its antitumor effect on
melanoma cells via activation of noncanonical autophagy and subsequent
apoptosis, suggesting that cannabinoids may be of clinical benefit for
metastatic melanoma.
The molecular mechanisms connecting autophagy to cell death remain poorly understood (Shen and Codogno, 2011); however, reports describing autophagy-dependent apoptosis (Ding et al., 2007; Salazar et al., 2009b; Grishchuk et al., 2011; Tomic et al., 2011) suggest multiple interactions between autophagic and apoptotic machinery. THC exerts its effect via the de novo synthesis of the sphingolipid ceramide, leading to the activation of ER stress, TRIB3-dependent inhibition of Akt/mTORC1 signaling, and autophagy-mediated apoptosis (Carracedo et al., 2006; Salazar et al., 2009b). TRIB3 has been identified as a key switch between cell survival and apoptosis during stress responses (Shimizu et al., 2012),
and the participation of TRIB3 in the melanoma response to THC may
direct cellular fate toward apoptosis in the context of ER
stress–induced autophagy (Salazar et al., 2013).
Autophagy
inhibition using both molecular and pharmacological approaches
prevented THC-induced autophagy and apoptosis of melanoma cells.
However, THC-induced autophagy was not prevented by knockdown of
Beclin-1, suggesting that, in contrast to glioma, noncanonical autophagy
mediates apoptosis in response to THC in melanoma. Beclin-1-independent
autophagy may promote caspase-independent cell death (Scarlatti et al., 2008; Basit et al., 2013) as well as apoptosis (Grishchuk et al., 2011; Tomic et al., 2011),
suggesting that autophagy mechanisms not involving Beclin-1 exist that
interact with cell death machinery. Interestingly, in contrast to
previous studies (Salazar et al., 2009b),
THC-induced autophagy was also independent of Ambra1 in melanoma cells.
Ambra1 is a Beclin-1 interacting protein that promotes autophagy by
stabilizing Beclin-1 complexes (Fimia et al., 2007);
although supporting the concept of Beclin-1-independent autophagy
activation in response to THC, these data highlight the complex
regulation of autophagy that likely occurs in a cell type- and
context-specific manner.
Tumor-selective killing can
be achieved by targeting pathways that are differentially regulated in
cancer cells compared with normal cells. In this respect, we have shown
that normal human melanocytes are resistant to THC at concentrations
that cause cell death in melanoma cells. This is consistent with studies
showing that cancerous cells are more sensitive to THC and other
cannabinoid receptor ligands compared with their nontransformed
counterparts, despite the presence of functional CB receptors (Velasco et al., 2012). Together with previous studies demonstrating an effect of synthetic ligands of cannabinoid receptors in melanoma (Blazquez et al., 2006),
these data support clinical evaluation of cannabinoids in
advanced-stage disease. Furthermore, THC activates autophagy and
apoptosis in both BRAF wild-type and mutated melanoma cell lines,
suggesting that despite autophagy deregulation (Armstrong et al., 2011; Corazzari et al., 2013),
THC is likely effective in melanoma tumors regardless of BRAF mutation
status. Furthermore, our findings show that THC is able to reduce
melanoma cell viability and tumor xenograft growth alone, but when lower
doses of THC are combined with CBD the antitumor effect was enhanced in vitro and was at least equally effective as the higher dose of single-agent THC in vivo. Moreover, CBD induces apoptosis via the production of reactive oxygen species and caspase activation in cancer cells (Massi et al., 2006; Shrivastava et al., 2011), indicating that THC and CBD engage different molecular machineries that cooperate to promote tumor cell death (Shrivastava et al., 2011; Torres et al., 2011).
In
summary, these data highlight the potential for cannabinoid-induced
cytotoxic autophagy as an effective strategy to drive melanoma cell
death, supporting the clinical evaluation of Sativex for the treatment
of metastatic disease.
Top of page
w/w) and CBD–BDS (CBD content 65.4% w/w; THC content 2.5% w/w; other individual plant cannabinoids <1 .7="" class="mb" span="">% w/w) were provided by GW Pharmaceuticals (Cambridge, UK). THC–BDS and CBD–BDS were provided as a resin, dissolved in ethanol (100 mg ml−1), dried, and prepared in DMSO. A 1:1 (w/w) preparation of THC–BDS and CBD–BDS was used to mimic Sativex (Sat-L). For oral administration, THC or Sat-L was solved in 100 μl of sesame oil.Materials and Methods
Cell culture, viability assays, and drug treatment
Melanoma
cell lines CHL-1, A375, and SK-MEL-28 were obtained from the ATCC
(Manassas, VA) in 2006 and cultured as described previously (Armstrong et al., 2007). Cell lines were verified by melan A staining (Flockhart et al., 2009) with B-RAF/NRAS mutational status confirmed using Custom TaqMan SNP genotyping assays (Hiscutt et al., 2010)
(Applera Europe BV, Life Technologies, Paisley, UK) (last tested
February 2014). Before drug treatment, culture medium was changed to 0.5%
fetal bovine serum medium. Temozolomide (OSI Pharmaceuticals, Melville,
NY) and ZVAD-fmk (benzyloxycarbonyl-V-A-D(OMe)-fluoromethylketone)
(Tocris Bioscience, Bristol, UK) were added in DMSO, and chloroquine
(Sigma-Aldrich Ltd, Poole, UK) was added in water. For in vitro
experiments, pure THC (THC Pharm, Frankfurt, Germany) and CBD
(synthesized by Professor Raphael Mechoulam (Hebrew University of
Jerusalem) and kindly provided by Dr Javier Fernandez Ruiz (Complutense
University, Madrid, Spain)) were prepared in DMSO. Control incubations
contained the same amount of DMSO (0.1–0.2% v/v). For treatment with THC+CBD, pure THC and pure CBD were mixed 1:1 (w/w).
Analysis of cell viability was performed using
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT;
Sigma). For in vivo experiments, THC–BDS (THC content 67.6% w/w; CBD content 0.3% w/w; other individual plant cannabinoids <1 .5="" class="mb" span="">%
Western blotting and reverse transcription–PCR analysis
Preparation
of whole-cell lysates and western blotting for LC3B, cleaved caspase 3
(Cell Signaling Technology, Leiden, The Netherlands), Atg7 (Santa Cruz
Biotechnology, Heidelberg, Germany), Beclin-1 (BD Biosciences, Oxford,
UK), and Ambra1 (Novus Biologicals, Cambridge, UK) all diluted 1:4,000,
and β-actin (Sigma) diluted 1:30,000, were performed as described
previously (Armstrong et al., 2007).
Total RNA was isolated from cells using the RNeasy Mini Kit with DNase
digestion (Qiagen, Manchester, UK) according to the manufacturer’s
protocol. Reverse transcription–PCR was performed using the Access
Reverse transcription–PCR system (Promega, Southampton, UK) using
primers for TRIB3 (Carracedo et al., 2006) or β-actin (Armstrong et al., 2005). PCR products were analyzed by electrophoresis on ethidium bromide–stained 2% agarose gels and DNA visualized by exposure to UV light.
siRNA transfections
The
siRNA for Atg7 (HSS116182), TRIB3 (sc-44426, Santa Cruz Biotechnology),
or a nontargeting siRNA (Negative control Low GC Stealth RNAi) (Stealth
RNAi, Life Technologies, Paisley, UK) were transfected into cells in
OPTIMEM containing siRNA (2–4 nM)
using Lipofectamine RNAiMAX (Life Technologies) according to the
manufacturer’s specification for reverse transfection. After 24 hours, the medium was replaced with DMEM containing 0.5% fetal bovine serum and cells treated with drugs as appropriate.
Retroviral or lentiviral infection
Retroviral expression of mRFP–GFP–LC3 (provided by T Yoshimori, Osaka University, Osaka, Japan) (Kabeya et al., 2000; Kimura et al., 2007) was performed as described previously (Armstrong et al., 2011).
Lentiviral expression of shAmbra1 or a nontargeting sequence shCtrl
(MISSION shRNA, Sigma) was performed by cotransfection of 7.5 μg lentivirus vector (pLKO.1-puro) with 2.5 μg
of an expression plasmid for the vesicular stomatitis virus G protein
into 293 cells using the calcium precipitation method. After 48 hours, melanoma cells were incubated with virus-containing supernatant supplemented with polybrene (4 μg ml−1) for 6–8 hours
and selected for puromycin resistance. Lentiviral expression of
shBeclin-1 (V3LHS_349509, Dharmacon, Chalfont St Giles, UK) or a
nontargeting sequence shCtrl (RHS4346, Dharmacon) was performed by
cotransfection of 20 μg lentivirus vector with 5 μg of an expression plasmid for the viral envelope (pMD2.G) and 15 μg packaging plasmid (pCMVdelta8.91) into HEK293T cells using the calcium precipitation method. After 72 hours, melanoma cells were spin transduced (1.5 hours, 1,200 r.p.m., Harrier 15/80, DJB Labcare, Newport Pagnell, UK) with virus-containing supernatant supplemented with polybrene (4 μg ml−1) and selected for puromycin resistance.
Confocal microscopy
Cells were grown on glass coverslips before fixation in 4% paraformaldehyde. For immunolabeling, cells were incubated with 0.2% Triton X-100 before incubation with anti-cytochrome c (BD Biosciences) at room temperature for 1 hour (McGill et al., 2005).
Secondary labeling was performed with Oregon Green 488 conjugated to
anti-rabbit IgG (Life Technologies). Nuclei were counterstained with
TO-PRO-3 iodide (Life Technologies). Cells were imaged under a Leica TCS
SP II laser-scanning confocal microscope with LCS Lite 2.61 software
(Leica Microsystems, Milton Keynes, UK), using a 63 × oil objective.
Xenograft mouse model and immunohistochemical analysis
Athymic nude (nu/nu) 5-week-old male mice (Harlan Iberica Laboratory, Madrid, Spain) were inoculated by subcutaneous injection of 7.5 × 106 CHL-1 cells in 100 μl phosphate-buffered saline containing 0.1% glucose. On establishment of tumors 250 mm3
in volume, mice were randomized into four treatment groups (5–8 mice
per group) and treated by daily administration for 20 days with
temozolomide (5 mg kg−1, local peritumoral injection), THC (15 mg kg−1, oral gavage), or Sativex (7.5 mg kg−1 THC–BDS+7.5 mg kg−1 CBD–BDS, oral gavage). The control group was treated with 100 μl of vehicle (sesame oil). Caliper measurements of tumor length (l) and width (w) were taken each day, and tumor volume was calculated as (4π/3) × (w/2)2 × (l/2).
Mice were humanely killed on the final day of treatment, and tumors
extracted and snap frozen in liquid nitrogen before storage at −80 °C. Frozen sections (6 μm) prepared on (3-Aminopropyl)triethoxysilane (Sigma)-coated glass slides were fixed in 4% paraformaldehyde before staining by TUNEL or with a Ki67 antibody (ab-15580, Abcam, Cambridge, UK) as previously described (Hill et al., 2009). For LC3 immunolabeling, frozen sections were fixed in acetone and incubated with anti-LC3B (ab48394, Abcam) for 1 hour
at room temperature. Secondary labeling was performed with Oregon Green
488 conjugated to anti-rabbit IgG (Life Technologies). Cells were
imaged under a Leica TCS SP II laser-scanning confocal microscope with
LCS Lite 2.61 software (Leica Microsystems), using a 63 × oil objective.
All procedures involving animals were performed according to Spanish
and European regulations and were approved by the ethical committee for
animal experimentation from Complutense University.
Statistical analysis
Images
of LC3, Ki67, and TUNEL staining were analyzed using Velocity (v4.3.1)
(Improvision, Perkin Elmer, Coventry, UK). Total fluorescence was
obtained by multiplying pixel number to the mean pixel intensity, after
appropriate thresholding. TOPRO-3 fluorescence was determined by the
number of pixels with fluorescence above the threshold, which was
proportional to nuclei number. For each tumor (three animals per group,
three randomly selected images each), values were reported as normalized
total fluorescence (total fluorescence (LC3/Ki67/TUNEL)/total
nuclear fluorescence) and expressed as fold change relative to the mean
value obtained from control animals, from two independent staining
analyses. The mRFP fluorescence was analyzed using ImageJ (public domain
licence, http://imagej.nih.gov/ij/); data are total pixel intensities/cell
minus the mean background fluorescence of nuclei. Homogeneity of
variances was checked using Levene’s test, and if variances were equal
(normalized data or after log transformation) data were analyzed by drug
treatment using Student’s t-test or one-way ANOVA with planned contrasts (LC3 expression) or Tukey’s post hoc test for multiple pair-wise comparisons. Where variances were not equal, Welch ANOVA test was used with Games–Howell post hoc tests for pair-wise comparisons (SPSS Statistics 20, SPSS, IBM, Portsmouth, UK).
