mtor inhibition via displacement of phosphatidic acid...

1

mTOR inhibition via displacement of phosphatidic acid 1

induces enhanced cytotoxicity specifically in cancer cells 2

Tra-Ly Nguyen1, Marie-Julie Nokin1,2, Maxim Egorov3,4, Mercedes Tomé5, Clément 3

Bodineau1, Carmelo Di Primo6, Lætitia Minder7, Joanna Wdzieczak-Bakala4, Maria 4

Concepcion Garcia-Alvarez4, Jérôme Bignon4, Odile Thoison4, Bernard Delpech4, Georgiana 5

Surpateanu4, Yves-Michel Frapart8, Fabienne Peyrot8,9, Kahina Abbas8, Silvia Terés1, Serge 6

Evrard10, Abdel-Majid Khatib5, Pierre Soubeyran11, Bogdan I. Iorga4,14, Raúl V. Durán1,13,14, 7

and Pascal Collin4,8,12 8

9

1 Institut Européen de Chimie et Biologie, INSERM U1218, Université de Bordeaux, 2 Rue 10

Robert Escarpit, 33607 Pessac, France 11

2 Metastasis Research Laboratory, GIGA-Cancer, University of Liège (ULiège), Liège, 12

Belgium 13

3 ATLANTHERA, 3 Rue Aronnax, 44821 Saint-Herblain Cedex, France 14

4 Institut de Chimie des Substances Naturelles, CNRS UPR 2301,1 Avenue de la Terrasse, 15

91198 Gif-sur-Yvette, France 16

5 Laboratoire de l’Angiogénèse et du Microenvironnement des Cancers, INSERM U1029, 17

Université de Bordeaux, Allée Geoffroy Saint Hilaire, Bâtiment B2, 33615 Pessac, France 18

6 Université de Bordeaux, Laboratoire ARNA, 146 rue Léo Saignat, Bordeaux, France; 19

INSERM U1212, CNRS UMR 5320, Institut Européen de Chimie et Biologie, CNRS 20

UMS3033/INSERMUS001, 2 Rue Robert Escarpit, 33607 Pessac, France 21

7 Université de Bordeaux, CNRS UMS3033 / INSERM US001, Institut Européen de Chimie 22

et Biologie, 2 Rue Robert Escarpit, 33607 Pessac, France 23

Research. on January 28, 2020. © 2018 American Association for Cancercancerres.aacrjournals.org Downloaded from

Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on July 27, 2018; DOI: 10.1158/0008-5472.CAN-18-0232

http://cancerres.aacrjournals.org/

2

8 Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, CNRS 24

UMR8601, Université Paris Descartes, Sorbonne Paris Cité, 45 Rue des Saints Pères, 25

75006 Paris, France 26

9 Ecole Supérieure du Professorat et de l'Education de l'Académie de Paris, Sorbonne 27

Université, 10 rue Molitor, 75016 Paris, France 28

10 Institut Bergonié, Digestive Tumours Unit, Université de Bordeaux, 229 Cours de 29

l'Argonne, 33076 Bordeaux, France 30

11 Institut Bergonié, INSERM U1218, Université de Bordeaux, 229 Cours de l’Argonne, 31

33076 Bordeaux, France 32

12 Université Paris Diderot, UFR Odontologie, 5 Rue Garancière, 75006, Paris, France 33

13 Leading Author 34

35

Running title: Targeting cancer cells with a new class of mTORC1 inhibitor 36

37

Key words: cancer, cytotoxicity, ICSN3250, mTOR, phosphatidic acid 38

39

Financial support: This work was supported by funds from the following institutions: Centre 40

National de la Recherche Scientifique-CNRS, Institut National de la Santé et de la 41

Recherche Médicale - INSERM, Fondation pour la Recherche Médicale, the Conseil 42

Régional d'Aquitaine, Fondation ARC pour la Recherche sur le Cancer, Ligue Contre le 43

Cancer - Gironde, SIRIC-BRIO, Institut de Chimie des Substances Naturelles -ICSN, Institut 44

Européen de Chimie et Biologie, Université Paris-Descartes, Société d’Accélération de 45

Transfert de Technologie d’Ile de France-SATT IDF-innov, Institut Bergonié, and National 46

Fund for Scientific Research. 47




3

48

14 To whom correspondence should be addressed: 49

Raúl V. Durán. Institut Européen de Chimie et Biologie, 2 Rue Robert Escarpit, 33607 50

Pessac, France. Tlf: +33 (0) 54 000 2259. FAX: +33 (0) 54 000 2215. E-mail: 51

[email protected]. 52

Bogdan I. Iorga. Institut de Chimie des Substances Naturelles, 1 Avenue de la Terrasse, 53

91198 Gif-sur-Yvette, France. E-mail: [email protected] 54

55

Conflict of interests: PC, ME, JWB, BD, JB, and OT are authors of a patent of ICSN3250 56

for the treatment of cancer (patent application WO2014/060366). 57

58

Word count: 5507 59

Total number of Figures: 7 60

Total number of Supplementary Figures: 7 61

Total number of tables: 0 62

Total number of Supplementary Tables: 1 63



mailto:[email protected]



4

Abstract 64

The mammalian target of rapamycin (mTOR) is a central regulator of cell growth and is 65

highly activated in cancer cells to allow rapid tumor growth. The use of mTOR inhibitors as 66

anti-cancer therapy has been approved for some types of tumors, albeit with modest results. 67

We recently reported the synthesis of ICSN3250, a halitulin-analogue with enhanced 68

cytotoxicity. We report here that ICSN3250 is a specific mTOR inhibitor that operates 69

through a mechanism distinct from those described for previous mTOR inhibitors. ICSN3250 70

competed with and displaced phosphatidic acid from the FRB domain in mTOR, thus 71

preventing mTOR activation and leading to cytotoxicity. Docking and molecular dynamics 72

simulations evidenced not only the high conformational plasticity of the FRB domain, but 73

also the specific interactions of both ICSN3250 and phosphatidic acid with the FRB domain 74

in mTOR. Furthermore, ICSN3250 toxicity was shown to act specifically in cancer cells, as 75

non-cancer cells showed up to 100-fold less sensitivity to ICSN3250, in contrast to other 76

mTOR inhibitors which did not show selectivity. Thus, our results define ICSN3250 as a 77

new-class of mTOR inhibitors that specifically targets cancer cells. 78

79

Significance 80

ICSN3250 defines a new-class of mTORC1 inhibitors that displaces phosphatidic acid at the 81

FRB domain of mTOR, inducing cell death specifically in cancer cells but not in non-cancer 82

cells. 83

84




5

Introduction 85

The serine/threonine kinase mTOR (mammalian target of rapamycin) is a master regulator of 86

cell growth, highly conserved among eukaryotes (1,2). mTOR is organised in two structurally 87

and functionally different complexes: the rapamycin-sensitive mTORC1 (mTOR complex 1), 88

and the rapamycin-insensitive mTORC2 (mTOR complex 2) (3–6). mTORC1 is mostly 89

activated by the presence of amino acids, by growth factors, by the bioenergetics status of 90

the cell, and by oxygen availability. In the control of mTORC1 by growth factors, the 91

tuberous sclerosis complex (TSC) and the mTORC1 co-activator Rheb play a crucial role 92

(7,8). One of the mechanisms by which the TSC/Rheb pathway controls mTORC1 involves 93

the production of phosphatidic acid (PA), which binds directly to mTOR at the FRB domain 94

and activates mTORC1 downstream of TSC/Rheb. Indeed, the downregulation of PA 95

production is sufficient to decrease mTORC1 activity (9,10). 96

As a major cell growth regulator, mTORC1 is recurrently upregulated in cancer cells 97

to allow rapid growth of tumors (11). Indeed, the use of rapamycin analogues has been 98

approved as anti-cancer therapy for certain types of cancer. However, the results of these 99

treatment are very modest with respect to patient survival and quality of life (12). Several 100

reasons have been invoked for these modest results in the clinics, including the reactivation 101

of a negative feedback loop downstream of mTORC1 that activates PI3K pathway (13), the 102

absence of mTORC2 inactivation upon rapamycin treatment (5), and recently, the potential 103

features of mTORC1 as a tumor suppressor (14,15). Still, inhibition of mTOR and the design 104

of new compounds that increase cancer cytotoxicity upon mTOR inhibition is an active field 105

of research, with recent reports proposing new-generation mTOR inhibitors that overcome 106

resistance to mTOR inhibition in tumors and effectively induce tumor regression (16). 107

However, to date, most of these mTOR inhibitors tested either showed a limited cytotoxicity 108

towards cancer cells (having mostly a cytostatic effect), or showed an excessive cytotoxicity 109

towards non-cancer cells, thus increasing adverse side effects. 110




6

Recently, we reported the synthesis and cytotoxicity of ICSN3250, an analogue of 111

the cytotoxic marine alkaloid halitulin (see Figure1a for the chemical structure of this 112

compound) (17). Halitulin was firstly reported in 1998 as a bisquinolinylpyrrole isolated from 113

the sponge Haliclona tulearensis, showing cytotoxicity against several tumor cell lines (18). 114

Our previous work concluded with the synthesis of a panel of halitulin analogues through the 115

formation of N-substituted 3,4-diarylpyrroles. Among them, ICSN3250 (also called 116

compound 25) was selected as a very potent derivative, presenting a high cytotoxicity at a 117

nanomolar concentration in a caspase-independent cell death mechanism (17). Our 118

preliminary results indicated an increased autophagy in cancer cells treated with ICSN3250. 119

However, the exact mechanism of action of ICSN3250 underlying its toxicity, and the 120

specificity of this cytotoxicity towards highly proliferative (cancer) cells, were not examined 121

previously. 122

In this report we investigated the molecular mechanism by which ICSN3250 induces 123

toxicity in cancer cells. Starting from a targeted screening analysing several signaling 124

pathways, we identified the mTORC1 pathway as a main target inhibited by ICSN3250 in the 125

nanomolar range. Our results indicated that ICSN3250 inhibited mTORC1 by following an 126

unprecedented mechanism that involved its competition with PA at the FRB domain of 127

mTOR to overcome the TSC negative regulation of mTORC1. This particular mechanism of 128

mTORC1 inhibition conducted to a potent and selective cytotoxicity observed in cancer cells 129

upon ICSN3250 treatment, which was not observed in non-cancer cells. Our results thus 130

defined ICSN3250 as a new-class mTORC1 inhibitor and validated ICSN3250 as a potential 131

anti-cancer drug for future clinical assays. 132




7

Materials and Methods 133

ICSN3250 synthesis 134

ICSN3250 (5,5’(1-(3-(azacyclotridecan-1-yl)propyl)-1H-pyrrole-3,4-diyl)bis(3-nitrobenzene-135

1,2-diol)) was synthesized as described previously (17) and in a published patent application 136

WO2014/060366 (19). Briefly, a new efficient “one-pot” method of unsymmetrically substituted 137

pyrroles synthesis was applied. It includes the condensation of an α-haloketone, first with a 138

primary amine, and then with an aldehyde. Subsequent intramolecular cyclization of this in 139

situ generated β-ketoenamine (enamine onto a ketone) results in formation of pyrrole based 140

ICSN3250 molecule. 141

142

Reagents and antibodies 143

Antibodies against mTOR (#2983, dilution 1:150), S6 (#2217, dilution 1:1000), phospho-S6 144

(Ser235/236) (#4856, dilution 1:1000), S6K (#2708, dilution 1:1000), phospho-S6K(T389) 145

(#9205, dilution 1:1000), 4EBP1 (#9452, dilution 1:1000), phospho-4EBP1(T37/46) (#2855, 146

dilution 1:1000), AKT (#4691, dilution 1:1000), phospho-AKT(Ser473) (#4060, dilution 147

1:1000), phospho-AKT(Thr308) (#13038, dilution 1:1000), AMPKα (#5832, dilution 1:1000), 148

phospho-AMPKα(Thr172) (#2535, dilution 1:1000), p53 (#2524, dilution 1:1000), phospho-149

p53(Ser15) (#9284, dilution 1:1000), p44/42 MAPK (#4695, dilution 1:1000), phospho-p44/42 150

MAPK(Thr202/Tyr204) (#9106, dilution 1:1000), p90RSK (#8408, dilution 1:1000), phosphor-151

p90RSK(Thr359/Ser363) (#9344, dilution 1:1000), phospho-p65(Ser536) (#3033, dilution 152

1:1000), p62 (#5114, dilution 1:1000), LC3 AB (#12741, dilution 1:1000), b-actin (#4967, 153

dilution 1:1000), RAPTOR (#2280, dilution 1:1000), TSC2 (#4308, dilution 1:1000) and Flag 154

(#14793, dilution 1:1000) were obtained from Cell Signaling Technology. Antibodies against 155

p65 (#sc-8008, dilution 1:1000) were obtained from Santa Cruz Biotechnology Inc. Antibody 156

against CD63 (SAB4700215, dilution 1:400) was obtained from Sigma. The secondary 157

antibodies anti-mouse (#7076, dilution 1:1000) and anti-rabbit (#7074, dilution 1:1000) were 158

obtained from Cell Signaling Technology. Phosphatidic acid (PA), Rapamycin (RAP) and 159




8

paraformaldehyde were obtained from Sigma. pcDNA3-FLAG-Rheb plasmid (Addgene 160

#19996) was a gift from Fuyuhiko Tamanoi. 161

162

Cell lines and culture conditions 163

HCT116, U2OS, U87, and K562 cells were obtained from ATCC. GFP-LC3 expressing U2OS 164

cells were kindly provided by Eyal Gottlieb (Cancer research UK, Glasgow, UK). WT and 165

TSC2-/- MEFs were kindly provided by David J. Kwiatkowski (Harvard Medical School, USA). 166

HCT116, U2OS and U87 cells were grown in DMEM high glucose (4.5 g/L) (GIBCO), and 167

K562 cells in RPMI (GIBCO), both supplemented with 10% of fetal bovine serum (Dominique 168

Dutscher), glutamine (2 mM), penicillin (Sigma, 100U/mL) and streptomycin (Sigma, 100 169

mg/mL), at 37° C, 5% CO2 in humidified atmosphere. Human umbilical vein endothelial cells 170

(HUVECs) were obtained from Promocell (Germany) and cultured according to the supplier’s 171

instructions in endothelial cell growth medium 2 containing growth factors and 2% fetal calf 172

serum. Primary normal human dermal fibroblasts (NHDF) derived from adult skin tissue were 173

purchased from Lonza and cultured according to the supplier’s instructions in fibroblast growth 174

medium containing human basic fibroblast growth factor (bFGF), insulin and 2% fetal calf 175

serum. Human follicle dermal papilla cells (HFDPC) isolated from human dermis originating 176

from lateral scalp were purchased fromTebu-Bio (Le Perray en Yvelines, France) and grown 177

in Follicle Dermal Papilla Cells Medium containing 4% FCS, 0.4% bovine pituitary extract, 1 178

ng/mL bFGF and 5 μg/mL of insulin (Tebu-Bio). The cells were maintained at 37°C in a 179

humidified atmosphere containing 5% CO2. Mycoplasma contamination check was carried 180

out using the VenorGeM Kit (Minerva Biolabs GmbH, Germany). When indicated, ICSN3250 181

(dissolved in DMSO before further dilution in assay mixture) was added at the indicated 182

concentration. PA was added to a final concentration of 1, 10 or 100 μM. 183

184

Isolation of patient-derived cancer cells and fibroblasts 185




9

Resected colon tumor tissue was obtained from Bergonié Institute (Bordeaux, France) 186

following written informed consent approved by Bergonié Institute. Patient consent 187

forms were obtained at the time of tissue acquisition. Biopsies were de-identified and 188

processed for cell culture. Briefly, to obtain a single cell suspension, tumor tissue was 189

cut into small fragments and enzymatically/mechanically dissociated with the gentle 190

MACS Octodissociator and the tumor dissociation kit for human (Miltenyi Biotec). 191

Single cells were then plated at 4.5x105 cell/ml in 96-well or 6-well plates previously 192

coated with collagen I at 300ug/ml (Rat tail, Gibco). ICSN 3250 (100nM) was 193

immediately added to the cells. ICSN3250 did not affect the adhesion of primary 194

patient cells in culture (data not shown). Proliferation was measured after 3 days of 195

culture in 96-well plates by cell counting and trypan blue exclusion with at least 3 196

replicates per condition. Photomicrographs of control and treated cells were taken at 197

3 days from 6-well plates (Nikon Eclipse TS100, Archimed software from Microvision 198

Instruments). 199

200

Co-culture experiments 201

GFP-expressing HCT116 or U2OS cancer cells were seeded with GFP-negative HUVEC or 202

NHDF cells in a ratio 1:1 and then treated with ICSN3250 at different concentrations during 203

72h. Cells were then collected and analyzed by flow cytometry. The percentage of GFP-204

positive and negative cells was calculated. 205

206

Plasmids and siRNAs transfections 207

Plasmid and siRNA transfections were carried out using Jetpei and Interferin@ (Polyplus 208

Transfection), respectively, according to the manufacturer’s instructions. Briefly, 70% 209

confluent cells were transfected with 5 μg of plasmid. 24 hours later cells were treated with 210

ICSN3250 for 24 more hours. Cells at 50% of confluence were transfected with siRNA (final 211




10

concentration 10 nM) for 48 h and then treated with ICSN3250 for 24h. All siRNAs were 212

obtained from Dharmacon (on-target plus smartpool siRNA). Sequences of non-targeting 213

control siRNAs (D-001810-02-05) were (1) UGGUUUACAUGUCGACUAA, (2) 214

UGGUUUACAUGUUGUGUGA, (3) UGGUUUACAUGUUUUCUGA, (4) 215

UGGUUUACAUGUUUUCCUA and sequences of TSC2 siRNAs (L-003029-00-0005) were (1) 216

GCAUUAAUCUCUUACCAUA, (2) CGAACGAGGUGGUGUCCUA, (3) 217

GGAAUGUGGCCUCAACAAU, (4) GGAUUACCCUUCCAACGAA. 218

219

Western blot 220

HCT116 cells, U2OS cells, NHDF, HUVEC, TSC2+/+ MEFs, and TSC2-/- MEFs were seeded 221

in 10cm plates. After the treatment, cells were washed with phosphate-buffered saline (PBS 222

1X) and lysed on ice using home-made RIPA buffer (Tris-HCl 50 mM pH 7.5, NaCl 150mM, 223

NP-40 1%, sodium deoxycholate 0.5%, EDTA 2mM, NaF 10mM) supplemented with protease 224

inhibitors (Sigma), phosphatase inhibitors (Sigma) and PMSF 1mM (AppliChem). Protein 225

quantification was performed with BCA assay kit (Thermo Fisher). After electrophoresis, the 226

proteins were transferred to a nitrocellulose membrane (BioRad) with Trans-Blot Turbo 227

Transfer System (Bio-Rad). The membranes were incubated for 30 minutes in PBS 1X with 228

0.01% Tween-20 and 5% bovine serum albumin (BSA). Primary antibodies were incubated 229

overnight at 4° C and secondary antibodies were incubated for 2 hours at room temperature. 230

Finally, membranes were imaged using the Chemi Doc MP Imager (Bio-Rad). 231

232

In vitro kinase assays 233

In vitro kinase assays of mTOR, AKT1, EGFR, PDK1, SRC, PKCα, and PKCε were performed 234

at CEREP Company (France). In vitro kinase assays of PI3Kα, PI3Kβ, PI3Kγ, and PI3Kδwere 235

performed using the PI3 Kinase Activity/Inhibitor ELISA assay from Merck-Millipore (USA). 236

Detailed procedures of these in vitro kinase assays are described in Supplementary Material 237

and Methods. 238

239




11

Immunoprecipitation 240

After treatment, cells were washed twice with cold PBS, then they were lysed with lP lysis 241

buffer (40mM Hepes pH 7.5, 120mM NaCl, 1mM EDTA, 0.3% CHAPS), supplemented with 242

protease inhibitor cocktail and phosphatase inhibitor cocktail (Sigma). Protein extracts were 243

incubated overnight at 4° C with anti-mTOR antibodies and then 4 hours at 4° C with magnetic 244

beads (Pierce Protein A/G Magnetic Beads, Thermo Fisher). Subsequently, beads were 245

washed twice with cold PBS and eluted with Laemmli buffer for Western Blot analysis. 246

247

Cell viability 248

To assess cell viability, 10 000 cells per well were seeded in triplicate in 96-well plates. The 249

number of cells were determined using the TC20 Automated Cell Counter (Bio-Rad) according 250

to the manufacturer’s protocol. Briefly, after the respective treatments cells were detached 251

with trypsin/EDTA and 10 μl of the cells suspension were mixed with 10 μl trypan blue 5% 252

solution (Bio-Rad) and analysed with the cell counter. Alternatively, cell viability was assessed 253

using the CellTiter-Blue Cell Viability Assay (Promega). After the treatment, 20 μl of the 254

reagent was added to each well and the plate was incubated for 1 - 4 hours at 37° C, 5% CO2 255

in humidified atmosphere. The fluorescence was recorded at 560/590nm in a Tristar2 LB942 256

(Berthold) device to determine the cell viability. 257

258

Cell cycle analysis 259

Exponentially growing cancer cells (HCT116 and U2OS) were incubated with ICSN3250 or 260

DMSO for 24 h. Cell-cycle profiles were determined by flow cytometry using propidium iodide 261

on a FC500 flow cytometer (Beckman–Coulter, France). 262

263

Surface Plasmon Resonance (SPR) 264

SPR experiments were performed at 25°C (sample rack at 10 °C) on a Biacore™ T200 265

apparatus (Biacore, GE Healthcare, Uppsala, Sweden) using CM5 sensor chips (Biacore™). 266




12

The surface was first activated with a 7-min pulse of 50 mM NHS / 200 mM EDC (GE 267

Healthcare) aqueous mixture using a flow rate of 5 μL/min. Streptavidin (IBA Lifesciences), 268

full-length (TP320457, Origene) and FRB-containing fragment (10012913, Thermo Fisher) 269

mTOR recombinant proteins were prepared in 10 mM sodium acetate buffer, pH 4.5 for mTOR 270

recombinant proteins, and pH 4.9 for streptavidin, and were injected in flow cells 2, 3 and 4, 271

respectively. 8800 resonance units (RU), 20300 RU and 12300 RU were immobilized, 272

respectively. The surface was then deactivated with a 7-min pulse of 1 M ethanolamine HCl-273

NaOH, pH 8.5 (GE Healthcare) and washed with three 1-min pulses of a mixture of 1 M NaCl 274

/ 50 mM NaOH prepared in mq water. The ICSN3250 compound (10 µM) was prepared in the 275

running buffer (PBS containing 0.05% Tween-20) and was injected for 1-min in duplicate at 276

100 µl/min. The regeneration of the surface was achieved with a 10-sec pulse of SDS 0.05% 277

at 30 µL/min followed by an extra wash with running buffer. One channel left blank was used 278

for referencing of the sensorgrams. The SPR signals were normalized to the moles/mm2 of 279

streptavidin immobilized onto the sensor chip surface (flow cell 2). 280

281

Confocal microscopy 282

Cells were grown on coverslips in 12 wells plates. Subsequently, after the treatments, cells 283

were rinsed with ice-cold PBS and fixed with 4% paraformaldehyde in PBS for 30 minutes at 284

room temperature. After the fixation, cells were permeabilized using PBS with Triton-X 0.05% 285

during 10 minutes, and then blocked with BSA 5% in PBS for 30 minutes. When required, cells 286

were incubated with primary antibody for 1 hour at 37° C. After three washes with PBS, the 287

coverslip was incubated for 1 hour at 37° C with the appropriate secondary antibody (anti-288

rabbit Alexa488, dilution 1:400 or anti-mouse Alexa555, dilution 1:400, obtained from 289

Invitrogen). Finally, coverslips were mounted with Prolong containing DAPI (Invitrogen). 290

Fluorescence was detected using a Leica confocal microscopy. Images were analysed using 291

Image J software. 292

293




13

Molecular modeling 294

Three-dimensional structures of ligands were generated using CORINA version 3.44 295

(http://www.molecular-networks.com). Molecular docking calculations were carried out using 296

GOLD software (20) and GoldScore scoring function, with the protein 2NPU (21) 297

(representative conformer 1) as receptor. The binding site was defined as a sphere with 15 Å 298

radius around a point with coordinates -6.449,6.669,-5.742. In agreement with our previous 299

studies (22–26) showing that an enhanced conformational search is beneficial, especially for 300

large molecules, a search efficiency of 200 % was used to better explore the ligand 301

conformational space. All other parameters were used with the default values. Molecular 302

dynamics simulations were carried out with GROMACS version 4.6.5 (27) using the OPLS-AA 303

(28) force field. Each system was energy-minimized until convergence using a steepest 304

descents algorithm. Molecular dynamics with position restraints for 200 ps was then 305

performed, followed by the production run of 100 ns. During the position restraints and 306

production runs, the Berendsen method (29) was used for pressure and temperature coupling. 307

Electrostatics were calculated with the particle mesh Ewald method (30). The P-LINCS 308

algorithm (31) was used to constrain bond lengths, and a time step of 2 fs was used 309

throughout. Ligand topologies for the OPLS-AA force field were generated using MOL2FF, an 310

in-house developed script, and were deposited into the Ligandbook repository (32) with IDs 311

2929 (https://ligandbook.org/package/2929) and 2930 (https://ligandbook.org/package/2930). 312

DFT calculations were carried out using Gaussian09, version D01 (33). Experimental pKa 313

values were taken from Jencks & Regenstein (1968) (34). All calculations were performed 314

using the High-Performance Computing (HPC) facilities available at the ICSN (Gif-sur-Yvette, 315

France). Images were generated with Pymol, version 1.8.6 (http://pymol.org). 316

317

Statistics 318

The results are expressed as a mean ± SEM of at least three independent experiments. t test 319

comparison was used to evaluate the statistical difference between two groups. One-way 320

ANOVA followed by Bonferroni’s comparison as a post hoc test was used to evaluate the 321



http://www.molecular-networks.com)/

https://ligandbook.org/package/2929)

https://ligandbook.org/package/2930


14

statistical difference between more than two groups. Statistical significance was estimated 322

when p<0.05. 323

324

Data availability 325

The authors declare that all the data supporting the findings of this study are available within 326

the article and its supplementary information files and from the corresponding author upon 327

reasonable request. 328




15

Results 329

ICSN3250 specifically inhibits mTORC1 pathway 330

To better understand the consequences at cell signaling level of ICSN3250 (Figure 1a) in 331

human cells, we treated two human cancer cell lines, the colorectal carcinoma cell line 332

HCT116 and the osteosarcoma cell line U2OS, with increasing concentrations of ICSN3250, 333

and we performed a targeted screening of different signaling pathways. These included -334

AMPK pathway (phospho-Thr172 AMPK), p53 pathway (phospho-Ser15 p53), PI3K pathway 335

(phospho-Thr308 AKT), ERK pathway (phospho-Thr202/Tyr204 p44/42 MAPK; phospho-336

Thr359/Ser363 RSK), NF-κB pathway (phospho-Ser536 p65), mTORC1 pathway (phospho-337

Thr389 S6K), and mTORC2 (phospho-Ser473 AKT). As shown in Figure 1b and 338

Supplementary Figure S1a, the only pathway that was inhibited by ICSN3250 treatment in 339

both cancer cells was mTORC1 pathway. Indeed, some other pathways, such as PI3K, ERK 340

and mTORC2 showed an increase in the phosphorylation of their respective downstream 341

targets. This increase would be in agreement with a specific inhibition of mTORC1 pathway, 342

and the subsequent release of the negative feedback loop that leads to PI3K re-activation 343

(13). To better evaluate the potential feedback role that these activations would have upon 344

ICSN3250 treatment, we investigated the effect of ICSN3250 in both U2OS and HCT116 345

cancer cells through long time-course experiments (24h, 48h and 72h). As shown in 346

Supplementary Figure S1b-c, we observed that ICSN3250 treatment efficiently inhibited 347

mTORC1 activity even at that long term (72h). However, the upregulation of PI3K, ERK and 348

NF-kB pathways observed at short term was robustly attenuated at longer times. These 349

results indeed suggested that the upregulation of these pathways is a transient 350

compensatory response of the cell, probably due to the release of the negative feedback 351

loop downstream of mTORC1, having no major contribution to the long-term effects of 352

ICSN3250 in cancer cells. 353

Dose dependent analysis showed a complete inhibition of mTORC1 (looking at the 354

phosphorylation of 3 well-known targets of mTORC1 pathway: S6K, S6 and 4EBP1) at 355




16

concentrations equal or higher than 50 nM of ICSN3250 (Figure 1c and Supplementary 356

Figure S1d). Time course analysis showed a slow yet efficient inhibition of mTORC1 that 357

reached a maximal inhibition upon 8 - 15 hours of treatment (Figure 1d and Supplementary 358

Figure S1e). This is considerably slower than previously reported mTORC1 inhibitors, such 359

as rapamycin or PP242. Confirming the capacity of ICSN3250 to inhibit mTORC1, we also 360

observed that ICSN3250 treatment induced an increase in autophagy, negatively regulated 361

by mTORC1 (35), as determined by increasing levels of LC3-II, by decreasing levels of the 362

adaptor protein p62, and by the accumulation of GFP-LC3 puncta, all of them standard 363

markers of autophagy (Figure 1e-h and Supplementary Figure S1f-g). Finally, ICSN3250 364

treatment caused cell cycle arrest at G0/G1 phase both in HCT116 cells and U2OS cells, as 365

expected upon mTORC1 inhibition (Figure 1i and Supplementary Figure S1h). 366

To validate ICSN3250 as a direct inhibitor of mTORC1, we investigated if ICSN3250 367

interacted directly with mTOR protein. For this purpose, we performed surface plasmon 368

resonance (SPR) experiments using a recombinant mTOR peptide, including the full length 369

of the protein. As shown in Figure 1j, SPR analysis showed a direct interaction between 370

ICSN3250 and mTOR protein, while no interaction with ICSN3250 was observed when a 371

control protein (streptavidin) was used instead of mTOR, demonstrating the specific 372

interaction between mTOR and ICSN3250. In addition, we repeated the analysis using only 373

the FRB domain of mTOR. Again, SPR analysis showed a specific interaction of ICSN3250 374

with the FRB domain of mTOR (Figure 1j). Altogether, our results indicated that ICSN3250 is 375

an inhibitor of mTORC1 that directly interacts with mTOR at the FRB domain. 376

377

ICSN3250 is not a kinase inhibitor of mTOR 378

Rapamycin and its analogues, as well as dual mTORC1/mTORC2 inhibitors, act as kinase 379

inhibitors of mTOR, with fast time-course kinetics. We analysed if ICSN3250 is a kinase 380

inhibitor of mTOR in vitro. The results shown in Figure 2a indicated that, although ICSN3250 381




17

had a capacity to inhibit the mTOR kinase activity, this effect occurred at concentrations 382

much higher (10 μM) than the observed inhibition of mTORC1 in cells (50 nM), and 105 383

times higher than the IC50 reported for rapamycin (0.1 nM) (36). This result confirmed that 384

ICSN3250 is not a kinase inhibitor of mTOR, suggesting that it operates differently than 385

other mTORC1 inhibitors. Indeed, ICSN3250 did not show any inhibitory capacity neither 386

towards PI3Kα, β, γ, or δ (Supplementary Figure S2a), nor towards other kinases analysed, 387

such as PKCα, PKCε, SRC, AKT1, EGFR and PDK1 (Supplementary Figure S2b). 388

389

ICSN3250 does not prevent lysosomal translocation of mTORC1 390

Next, we investigated if ICSN3250 prevents the translocation of mTORC1 to the surface of 391

the lysosome, a well-known mechanism involved in the activation of mTORC1 by nutritional 392

inputs (37). As shown in Figure 2b, and quantified in Figure 2c, the addition of 100 nM of 393

ICSN3250 (a concentration at which mTORC1 was completely inhibited, see Figure 1b-c) to 394

HCT116 cells did not prevent the co-localization of mTOR with the lysosomal marker CD63, 395

clearly indicating that lysosomal localization of mTORC1 was not impaired by ICSN3250. 396

Similar results were obtained in U2OS cells (Supplementary Figure 2c-d). Furthermore, even 397

when ICSN3250 was not able to prevent the amino acid-induced lysosomal translocation of 398

mTORC1, still ICSN3250 was able to prevent the activation of mTORC1 mediated by amino 399

acid in both cell lines (Figure 2d and Supplementary Figure S2e), again suggesting that the 400

inhibition of mTORC1 occurred once mTORC1 is at the lysosomal surface. To finally discard 401

that lysosomal translocation is involved in the mechanism of action of ICSN3250, we over-402

expressed a de-localized Flag-Rheb, that renders mTORC1 activation outside the lysosome. 403

As expected, Flag-Rheb overexpression induced mTORC1 activation in the absence of 404

amino acids (Supplementary Figure S2f-g). However, de-localized Flag-Rheb did not prevent 405

the inhibitory effect of ICSN3250 towards mTORC1 activity (Figure 2e and Supplementary 406

Figure S2h), finally confirming that ICSN3250 operates after the translocation of mTORC1 to 407

the lysosome. 408




18

409

ICSN3250 does not destabilize mTORC1 410

mTORC1 destabilization has been proposed as a mechanism of mTORC1 inhibition upon 411

certain metabolic stresses (38). Thus, we investigated if ICSN3250 destabilizes mTORC1 as 412

an inhibitory mechanism. For this purpose, we immunoprecipitated mTOR and analysed the 413

presence of the specific mTORC1 component Raptor in the immunoprecipitates. As 414

expected, in the absence of the compound, Raptor was observed upon mTOR 415

immunoprecipitation (Figure 2f). Our results showed that ICSN3250 treatment (at 24h) was 416

not able to prevent the interaction of mTOR with Raptor (Figure 2f). Hence, we concluded 417

that the mechanism of action of ICSN3250 does not affect the integrity of the mTORC1, 418

localized at the lysosome. 419

420

ICSN3250 antagonizes with phosphatidic acid to inhibit mTORC1 421

We next investigated the mechanism that allow mTORC1 activation at the lysosome. These 422

mechanisms are controlled by the Tuberous Sclerosis Protein 1/2 complex (TSC complex), 423

that exerts a negative regulation towards mTORC1 (7). To investigate if TSC complex plays 424

a role in the mechanism of action of ICSN3250, we treated TSC2+/+ MEFs and TSC2-/- MEFs 425

with increasing concentrations of ICSN3250. Similarly to what we observed in cancer cell 426

lines, ICSN3250 induced a complete inhibition of mTORC1 at concentrations higher than 50 427

nM in TSC2+/+ MEFs (Figure 3a). Concomitantly, we observed an activation of autophagy (as 428

determined by increasing LC3II levels), as expected upon mTORC1 inhibition. However, the 429

inactivation of TSC complex in TSC2-/- MEFs induced a complete recovery of mTORC1 430

activity and a lack of autophagy activation even in the presence of ICSN3250 at 100nM 431

(Figure 3a). These results were confirmed by knocking down TSC2 in HCT116 and U2OS 432

cells. As shown in Figure 3b and Supplementary Figure S3a, the efficient silencing of TSC2 433

resulted in the re-activation of mTORC1 in ICSN3250 treated cells. We concluded that the 434




19

regulation of mTORC1 by TSC complex might be involved in the mechanism of action of 435

ICSN3250. 436

The production of phosphatidic acid (PA) by phospholipase D1 (PLD1) has been 437

previously invoked as a mechanism of the regulation of mTORC1 by TSC complex (39), and 438

it is largely known that PA binds to and activates mTORC1 (9). In addition, our previous 439

results indicating that ICSN3250 interacts with the FRB domain of mTOR, where the pocket 440

for PA is located, supported the hypothesis that ICSN3250 could compete with PA in 441

mTORC1 binding, thus displacing PA from its binding site, leading to mTORC1 inhibition 442

downstream of TSC complex. To test this possibility, we first performed a competitive 443

analysis of mTORC1 activation between PA and ICSN3250. For this purpose, we treated 444

HCT116 cells with ICSN3250 (100nM) in the presence of increasing concentrations of PA (0 445

to 100 μM). As we observed previously, ICSN3250 alone induced the inhibition of mTORC1. 446

However, co-incubation of cells with PA induced a dose-dependent mTORC1 reactivation 447

and autophagy inhibition even in the presence of ICSN3250 (Figure 3c-f and Supplementary 448

Figure S3b-c). Conversely, increasing concentrations of ICSN3250 limited the activation of 449

mTORC1 and the inhibition of autophagy induced by PA (Figure 3g-h and Supplementary 450

Figure S3d-e). These results strongly suggest that ICSN3250 antagonizes with PA to inhibit 451

mTORC1. 452

453

ICSN3250 binds to the FRB domain of mTOR and displaces phosphatidic acid 454

To confirm the previous conclusion that ICSN3250 antagonizes with PA, we performed 455

molecular docking calculations to identify the binding modes of ICSN3250 and PA within the 456

FRB domain of mTOR. Three different protonation states of the catechol group in ICSN3250 457

were considered during the docking process (i.e. neutral and deprotonated on either OH 458

group) and the strongest interactions and the best protein-ligand shape complementarity 459

were obtained with the form deprotonated on the OH situated ortho from the NO2 460




20

substituent. We computed the pKa of this OH group using the protocol described by 461

Muckerman et al.(40) (DFT calculations on a simplified analogue of ICSN3250 with implicit 462

solvent and removal of the systematic error) and we found a value of 5.93±0.55, meaning 463

that this group is negatively charged at physiological pH. This is in strong agreement with the 464

docking results, showing interactions between this group and the positively charged side 465

chains of Lys2095 on one side and of Arg2042 on the other (Figure 4a-b). 466

The FRB domain of mTOR (apo form) and the docking complexes with ICSN3250 467

and PA were used for molecular dynamics (MD) simulations (100 ns each), to take into 468

account two factors that were missing in the docking process: protein flexibility and the 469

presence of explicit aqueous solvent. As expected, the apo simulation reached very quickly 470

an equilibrium conformation that is conserved until the end. In contrast, the two complexes 471

evolved slowly towards an equilibrium structure which is attained only after 75-80 ns 472

(Supplementary Figure S4a-c), highlighting the need for relatively long MD simulations in the 473

study of flexible proteins. The representative equilibrium structures from these simulations 474

showed a number of interesting elements. The protein surface is very flexible, changing the 475

shape according to the interaction partner. Consequently, a very good protein-ligand surface 476

complementarity was observed for the two complexes, bringing an important contribution to 477

the ligand affinity, which is complemented by strong ionic interactions between nitrocatechol 478

groups and Lys2095 and Arg2042 in the case of ICSN3250 and between the phosphate 479

group and Arg2109 in the case of PA (Figure 4c-f). The interaction between ICSN3250 and 480

its binding site showed three distinct regions: i) the ionic interaction between the 481

nitrocatechol groups and Lys2095 and Arg2042 that was already mentioned; ii) a π-stacking 482

interaction between the pyrrole ring and Phe2039 and iii) the interaction between the 483

macrocycle and a hydrophobic subpocket composed of residues Trp2101, Tyr2105, 484

Phe2108, Leu2031 and Tyr2104. Ser2035, which was shown to be important for the 485

interaction of mTOR with rapamycin (41), is also part of the binding site (Figure 5a-b). 486

ICSN3250 is relatively flat on the protein surface, whereas PA is deeply buried with its two 487




21

hydrophobic tails that interact with a subpocket containing Trp2101, Tyr2105, Phe2108, 488

Leu2031, Leu2054, Tyr2104, Ser2035, Phe2039, Leu2051, Tyr2038, Val2044 and Met2047. 489

Only the phosphate head is solvent-exposed and interacts with Arg2109 (Figure 5c-d). This 490

orientation is similar to the one previously observed by NMR (21), with the exception of the 491

tail chains that are more deeply buried in our case. 492

Overall, the residues involved in the interaction between mTOR and the two ligands studied 493

in this work clearly show a significant overlapping of the two binding sites. Our results 494

supported that ICSN3250 binds to the FRB domain of mTOR and displaces PA, leading to 495

mTORC1 inhibition. This mechanism defines ICSN3250 as a new-class mTORC1 inhibitor. 496

497

Inhibition of mTORC1 by ICSN3250 is responsible for its cytotoxicity in cancer cells 498

Previously, we reported that ICSN3250 showed an increased cytotoxicity in human cells 499

(17). We have now confirmed that no radical intermediate can be observed by ESR (electron 500

spin resonance) in vitro with ICSN3250 in the presence of superoxide anion, the main ROS 501

in cells (Supplementary Figure S5a). We also observed that ICSN3250 did not react with 502

superoxide (Supplementary Figure S5b), confirming the redox stability of ICSN3250. In 503

addition, as shown in Supplementary Figure S5c-d, ICSN3250 did not induce an increase of 504

ROS levels, neither at the cytosol nor at the mitochondria. In agreement with these results, 505

we also observed that the cytotoxicity of ICSN3250 in HCT116 was not reversed (neither 506

partially reduced) by treating the cells with N-acetylcysteine, a classical antioxidant 507

(Supplementary Figure S5e). These results indicated that the cytotoxicity exerted by 508

ICSN3250 in cancer cells is not mediated by a potential increase in ROS levels. 509

Our results demonstrating that ICSN3250 acts as a new-class mTOR inhibitor led us 510

to investigate if the inhibition of mTORC1 was the primary reason for the cytotoxicity induced 511

by ICSN3250. For this purpose, we investigated if the re-activation of mTORC1 mediated by 512

TSC ablation in TSC2-/- MEFs protected from the cytotoxic effect of ICSN3250. As shown in 513




22

Figure 6a and Supplementary Figure S6a, TSC2-/- MEFs showed an increased protection 514

against cytotoxicity induced by ICSN3250 with respect to TSC2+/+ MEFs (as control, TSC2-/- 515

MEFs did not show any increased viability with respect to TSC2+/+ in the absence of 516

ICSN3250). Similar results were obtained when TSC2 was knocked down in HCT116 or 517

U2OS cells: the efficient silencing of TSC2 was sufficient to restore (at least partially) cell 518

viability in ICSN3250-treated cells (Figure 6b and Supplementary Figure S6b). 519

We previously showed that treatment of HCT116 cells with PA (100 μM) was 520

sufficient to re-activate mTORC1 (see Figure 3b). Now, we confirmed that PA treatment also 521

prevented the cytotoxic effect of ICSN3250 in a dose-dependent manner in HCT116 cells 522

(Figure 6c-e). This result clearly suggested that the inhibition of mTORC1 by ICSN3250 is 523

responsible for its cytotoxicity. Furthermore, the particular mechanism of mTORC1 inhibition 524

induced by ICSN3250, is likely the reason of the increased cytotoxicity showed by this 525

compound with respect to other mTORC1 inhibitors, such as rapamycin. Indeed, while 526

rapamycin induced a stronger inhibition of mTORC1 than ICSN3250 (Figure 6f), it did not 527

cause the cytotoxic effect that we observed upon ICSN3250 treatment (Figure 6g). 528

Compared with a panel of mTOR inhibitors, ICSN3250 was not the most potent mTORC1 529

inhibitor among them as determined by the dephosphorylation of mTORC1-downstream 530

targets (Supplementary Figure S6c-d), but yet it ranked among the most cytotoxic 531

compounds for cancer cells, showing one of the lowest IC50 values (Figure 6h and 532

Supplementary Figure S6e). Hence, we concluded that the qualitative (and not quantitative) 533

differences between the inhibition exerted by ICSN3250 with respect to other mTOR 534

inhibitors are key for the marked cytotoxicity induced by ICSN3250. 535

536

ICSN3250 specifically targets cancer cells both in vitro and ex vivo 537

To validate the potential applicability of ICSN3250 as an anticancer drug, we compared the 538

cytotoxicity of ICSN3250 in a panel of cells including both cancer cells and non-cancer cells. 539




23

As shown in Figure 7a, ICSN3250 showed a cytotoxicity in cancer cells (HCT116, U2OS, 540

U87, and K562) that was 10-100 times more potent than its cytotoxicity in human non-cancer 541

cells (NHDF, HUVEC, and HFDPC; MEFs, as they are highly proliferating, do not really 542

behave as normal cell in culture, and they were sensitive to ICSN3250). Lack of toxicity in 543

non-cancer cells (NHDF and HUVEC) was confirmed in long time-course experiments, at 544

72h of treatment, in clear contrast with cancer cells (HCT116 and U2OS) (Figure 7b-c and 545

Supplementary Figure S7a-b). Importantly, the reduced cytotoxicity of ICSN3250 observed 546

in non-cancer cells correlated with its reduced capacity to inhibit mTORC1 in these cells: the 547

inhibition of mTORC1 in HUVEC and NHDF cells was only reached at concentrations higher 548

than 500 nM (Figure 7d and Supplementary Figure S7c), while full mTORC1 inactivation in 549

U2OS and HCT116 cells was observed at 50 nM (see Figure 1c and Supplementary Figure 550

S1d). 551

To further sustain the notion that ICSN3250 exhibits selectivity for cancer cells over 552

non-cancer cells, we performed co-culture experiments of GFP-labelled cancer cells 553

(HCT116 or U2OS) together with unlabelled non-cancer cells (HUVEC and NHDF) treated 554

with increasing concentrations of ICSN3250 for 72h. Flow cytometry analysis showed a clear 555

decrease in the GFP-positive (cancer cells) population with respect to the GFP-negative 556

(non-cancer cells) population (Figure 7e-g and Supplementary Figure S7d-f). These results 557

corroborated that ICSN3250 exhibits selective cytotoxicity towards cancer cells. Next, in 558

order to validate the capacity of ICSN3250 to target primary cancer cells, we performed ex 559

vivo treatment of cancer cells from a colorectal cancer patient. As shown in Figure 7h-j, we 560

observed a substantial decrease in the viability of these primary cancer cells after 72h of 561

treatment with ICSN3250 ex vivo. Importantly, primary fibroblasts obtained from the same 562

patient failed to show any decrease in viability upon ICSN3250 treatment. These results 563

confirmed the validation and specificity of ICSN3250 to kill primary cancer cells. 564

Finally, compared with other mTOR inhibitors that showed cytotoxicity in cancer cells 565

(such as INK 128, gedatolisib or VS-5584, among others), ICSN3250 was substantially less 566




24

toxic in human primary normal cells (Figure 7k), supporting the concept that ICSN3250 567

presents an action mechanism that makes it particularly interesting to develop anti-cancer 568

strategies. 569




25

Discussion 570

The results shown herein presented ICSN3250 as a new-class of mTORC1 inhibitor that 571

acts through a mechanism that differs from those described by other mTOR inhibitors. 572

ICSN3250 is an analogue of the cytotoxic marine alkaloid halitulin, previously reported to 573

present an increased cytotoxicity (17). However, the mechanism of action underlying this 574

cytotoxicity was not known. Our results showed a specificity of ICSN3250 targeting 575

mTORC1, without inhibiting other signaling pathways, such as AMPK, p53, PI3K, ERK, NF-576

κB, or even mTORC2. Surprisingly, ICSN3250 did not affect the kinase activity of mTOR, 577

neither the stability of mTOR complex. Instead, our results showed that ICSN3250 binds to 578

the FRB domain of mTOR, displacing PA to overcome the TSC negative regulation of 579

mTORC1 as a mechanism for mTORC1 inhibition. Indeed, increasing amounts of 580

exogenously added PA or TSC ablation restored mTORC1 activity. This competition with PA 581

seems to be key for the cytotoxicity of ICSN3250, as exogenously added PA not only 582

restored mTORC1, but also restored cell viability. Of note, ICSN3250 did not show an 583

increased capacity to inhibit mTORC1 with respect to previously reported mTOR inhibitors, 584

but yet it showed a particularly high cytotoxic effect in cancer cells, showing a lower IC50 585

than typical inhibitors such as temsirolimus, accepted by FDA as a treatment against renal 586

cell carcinoma. Importantly, the cytotoxicity of ICSN3250 towards non-cancer cells is 587

substantially lower than the most potent of the other inhibitors of mTOR, placing ICSN3250 588

as a good candidate for future clinical assays. 589

mTOR inhibition has been approved as a cancer therapy for several types of tumor 590

(42). Yet, the efficiency of those treatments is very modest. Rapamycin and analogues 591

showed mostly cytostatic effect, which in the patient results in a mild delay of tumor growth, 592

with little effect (although statistically significant) in patient survival. These modest results 593

have been explained by the re-activation of PI3K pathway as a consequence of the release 594

of negative feedback loop downstream of mTORC1 (13). This is why a new generation of 595

dual mTORC1/mTORC2 inhibitors and PI3K/mTOR inhibitors are being proposed and 596




26

tested. However, these inhibitors still show increased cytotoxicity in non-cancer cells. 597

Besides, the use of monotherapies targeting single signaling pathways to treat cancer is 598

under reconsideration. Due to the intrinsic genetic heterogeneity of tumors and the rapid 599

evolution and adaptation of tumor cells during the progression of the disease (43), 600

developing drug resistance is a recurrent problem during treatment, particularly when 601

monotherapies have been used. Still, the efficacy of ICSN3250 to selectively target tumor 602

cells in vivo remains to be elucidated. 603

As mTORC1 is not the only protein activated by PA, it could be envisioned that other 604

mechanisms or pathways could be involved in ICSN3250-induced cytotoxicity. However, our 605

results showing that mTORC1 re-activation in TSC2-/- cells restored cell viability indicated 606

that mTORC1 inhibition is at the basis of ICSN3250-induced cytotoxicity. The unprecedented 607

mechanism of action of ICSN3250, displacing PA to overcome TSC negative regulation of 608

mTORC1, without affecting mTOR kinase activity, seems to be key to explain the specific 609

cytotoxicity for cancer cells showed by this type of mTORC1 inhibitor. Why this mechanism 610

of action would be more cytotoxic than mTOR kinase inhibition mediated by ATP-competitive 611

inhibitors would require further investigations. As ICSN3250-induced PA displacement from 612

the FRB domain of mTOR would likely occur at the surface of the lysosome (where 613

mTORC1 is located upon activation), it could be hypothesized that this displacement causes 614

a collapse in the lysosomal surface, perturbing lysosomal function and leading to cell death, 615

as proposed for other types of stress (44). Alternatively, the slower inactivation of mTORC1 616

mediated by ICSN3250 as compared with other mTOR inhibitors that we observed could be 617

playing in favour of its cytotoxicity, as our recent results showed that a fast and complete 618

inhibition of mTORC1 upon rapamycin treatment prevents apoptotic cell death during 619

nutritional imbalance (14). 620

Finally, our results make particular emphasis in the control of mTORC1 activity by 621

PA, a regulation that has not received as much attention as the regulation exerted by amino 622

acids or by PI3K signaling. However, our results clearly indicated that interfering with PA 623




27

binding in the FRB domain of mTOR is indeed an effective approach to inhibit mTORC1 624

even in the presence of amino acids and growth factors, underscoring the importance of PA 625

for mTORC1 activity. Besides, as mentioned above, the regulation of mTORC1 by PA 626

seems to be particularly important at the cell physiology level, as the interference with the 627

mTOR-PA interaction resulted in cell death. How exogenously added PA results in mTORC1 628

activation is not clear (9,45,46). In our experiments we needed a substantially higher 629

concentration of PA to compete with ICSN3250, probably reflecting that the concentration of 630

PA at the lysosomal surface do not reach the same concentration than exogenously added 631

PA. 632

In conclusion, ICSN3250 defines a new-class of mTORC1 inhibitors that, due to its 633

particular mechanism of action, induces cell death specifically in tumor cells but not in non-634

cancer cells. Additional researches will determine the applicability of this type of compound 635

for anti-cancer therapy. 636




28

Acknowledgements 637

This work was supported by funds from the following institutions: Centre National de la 638

Recherche Scientifique-CNRS, Institut National de la Santé et de la Recherche Médicale - 639

INSERM, Fondation pour la Recherche Médicale, the Conseil Régional d'Aquitaine, 640

Fondation ARC pour la Recherche sur le Cancer, Ligue Contre le Cancer - Gironde, SIRIC-641

BRIO, Institut de Chimie des Substances Naturelles -ICSN, Institut Européen de Chimie et 642

Biologie, Université Paris-Descartes, Société d’Accélération de Transfert de Technologie 643

d’Ile de France-SATT IDF-innov, Institut Bergonie, and National Fund for Scientific Research 644

of Belgium. MJN is Télévie Post-Doctoral Fellow from the National Fund for Scientific 645

Research (FRNS, Belgium). pcDNA3-FLAG-Rheb plasmid (Addgene #19996) was a gift 646

from Fuyuhiko Tamanoi. GFP-LC3 expressing U2OS cells were kindly provided by Eyal 647

Gottlieb (Cancer research UK, Glasgow, UK). We thank Professor J.Y. Lallemand, director 648

of the ICSN (2000-2009). We extend our thanks to A. Pinault for skillful technical assistance. 649

We would like to remember Dr. C. Marazano (†11/12/2008), who initiated and supervised 650

the initial biomimetic synthesis of ICSN3250, one of the halituline's analogues. We are 651

grateful to the structural biology facility (UMS 3033/ US001) of the Institut Européen de 652

Chimie et Biologie (Pessac, France) for access to the T200 Biacore instrument, which was 653

acquired with the support of the Conseil Régional d’Aquitaine, the GIS-IBiSA and the Cellule 654

Hôtels à Projets of the Centre National de la Recherche Scientifique. We thank Vincent 655

Pitard and the Flow Cytometry Platform (Université de Bordeaux, France) for technical 656

assistance in flow cytometry experiments. 657

658




29

References 659

1. Saxton RA, Sabatini DM. mTOR Signaling in Growth, Metabolism, and Disease. Cell. 660 2017;168:960–76. 661

2. González A, Hall MN. Nutrient sensing and TOR signaling in yeast and mammals. 662 EMBO J. 2017;36:397–408. 663

3. Loewith R, Jacinto E, Wullschleger S, Lorberg A, Crespo JL, Bonenfant D, et al. Two 664 TOR complexes, only one of which is rapamycin sensitive, have distinct roles in cell 665 growth control. Mol Cell. 2002;10:457–68. 666

4. Kim DH, Sarbassov DD, Ali SM, King JE, Latek RR, Erdjument-Bromage H, et al. 667 mTOR interacts with raptor to form a nutrient-sensitive complex that signals to the cell 668 growth machinery. Cell. 2002;110:163–75. 669

5. Sarbassov DD, Ali SM, Kim D-H, Guertin DA, Latek RR, Erdjument-Bromage H, et al. 670 Rictor, a Novel Binding Partner of mTOR, Defines a Rapamycin-Insensitive and 671 Raptor-Independent Pathway that Regulates the Cytoskeleton. Curr Biol. 672 2004;14:1296–302. 673

6. Jacinto E, Loewith R, Schmidt A, Lin S, Rüegg MA, Hall A, et al. Mammalian TOR 674 complex 2 controls the actin cytoskeleton and is rapamycin insensitive. Nat Cell Biol. 675 2004;6:1122–8. 676

7. Inoki K, Li Y, Zhu T, Wu J, Guan KL. TSC2 is phosphorylated and inhibited by Akt and 677 suppresses mTOR signalling. Nat Cell Biol. 2002;4:648–57. 678

8. Durán R V., Hall MN. Regulation of TOR by small GTPases. EMBO Rep. 679 2012;13:121–8. 680

9. Fang Y, Vilella-Bach M, Bachmann R, Flanigan A, Chen J. Phosphatidic Acid-681 Mediated Mitogenic Activation of mTOR Signaling. Science (80- ). 2001;294:1942–5. 682

10. Fang Y, Park IH, Wu AL, Du G, Huang P, Frohman MA, et al. PLD1 Regulates mTOR 683 Signaling and Mediates Cdc42 Activation of S6K1. Curr Biol. 2003;13:2037–44. 684

11. Menon S, Manning BD. Common corruption of the mTOR signaling network in human 685 tumors. Oncogene. 2009;27:S43–51. 686

12. Hudes G, Carducci M, Tomczak P, Dutcher J, Figlin R, Kapoor A, et al. Temsirolimus, 687 interferon alfa, or both for advanced renal -cell carcinoma. N Engl J Med. 688 2007;356:2271–81. 689

13. Um SH, Frigerio F, Watanabe M, Picard F, Joaquin M, Sticker M, et al. Absence of 690 S6K1 protects against age- and diet-induced obesity while enhancing insulin 691 sensitivity. Nature. 2004;431:200–5. 692

14. Villar VH, Nguyen TL, Delcroix V, Terés S, Bouchecareilh M, Salin B, et al. mTORC1 693 inhibition in cancer cells protects from glutaminolysis-mediated apoptosis during 694 nutrient limitation. Nat Commun. 2017;8:1–12. 695

15. Villar VH, Nguyen TL, Terés S, Bodineau C, Durán R V. Escaping mTOR inhibition for 696 cancer therapy: Tumor suppressor functions of mTOR. Mol Cell Oncol. 697 2017;4:e1297284. 698

16. Rodrik-Outmezguine VS, Okaniwa M, Yao Z, Novotny CJ, McWhirter C, Banaji A, et 699 al. Overcoming mTOR resistance mutations with a new-generation mTOR inhibitor. 700 Nature. 2016;534:272–6. 701




30

17. Egorov M, Delpech B, Aubert G, Cresteil T, Garcia-Alvarez MC, Collin P, et al. A 702 concise formation of N-substituted 3,4-diarylpyrroles-synthesis and cytotoxic activity. 703 Org Biomol Chem. 2014;12:1518–24. 704

18. Kashman Y, Koren-Goldshlager G, Gravalos G, Schleyer M. Halitulin, a new cytotoxic 705 alkaloid from the marine sponge Haliclona tulearensis. Tetrahedron Lett. 706 1999;40:997–1000. 707

19. Collin P, Egorov M, Delpech B, Bakala J, Achab M, Bignon J, et al. N-substituted 3,4-708 bis(catechol)pyrrole compounds, and the preparation and use thereof in the treatment 709 of cancer. Patent number: WO2014/060366. France; 2014. 710

20. Verdonk ML, Mortenson PN, Hall RJ, Hartshorn MJ, Murray CW. Protein−Ligand 711 Docking against Non-Native Protein Conformers. J Chem Inf Model. 2008;48:2214–712 25. 713

21. Veverka V, Crabbe T, Bird I, Lennie G, Muskett FW, Taylor RJ, et al. Structural 714 characterization of the interaction of mTOR with phosphatidic acid and a novel class 715 of inhibitor: compelling evidence for a central role of the FRB domain in small 716 molecule-mediated regulation of mTOR. Oncogene. 2008;27:585–95. 717

22. Surpateanu G, Iorga BI. Evaluation of docking performance in a blinded virtual 718 screening of fragment-like trypsin inhibitors. J Comput Aided Mol Des. 2012;26:595–719 601. 720

23. Colas C, Iorga BI. Virtual screening of the SAMPL4 blinded HIV integrase inhibitors 721 dataset. J Comput Aided Mol Des. 2014;28:455–62. 722

24. Martiny VY, Martz F, Selwa E, Iorga BI. Blind Pose Prediction, Scoring, and Affinity 723 Ranking of the CSAR 2014 Dataset. J Chem Inf Model. 2016;56:996–1003. 724

25. Selwa E, Martiny VY, Iorga BI. Molecular docking performance evaluated on the D3R 725 Grand Challenge 2015 drug-like ligand datasets. J Comput Aided Mol Des. 726 2016;30:829–39. 727

26. Selwa E, Elisée E, Zavala A, Iorga BI. Blinded evaluation of farnesoid X receptor 728 (FXR) ligands binding using molecular docking and free energy calculations. J 729 Comput Aided Mol Des. 2017;in press. 730

27. Pronk S, Páll S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, et al. GROMACS 4.5: 731 a high-throughput and highly parallel open source molecular simulation toolkit. 732 Bioinformatics. 2013;29:845–54. 733

28. Jorgensen WL, Schyman P. Treatment of Halogen Bonding in the OPLS-AA Force 734 Field: Application to Potent Anti-HIV Agents. J Chem Theory Comput. 2012;8:3895–735 901. 736

29. Berendsen HJC, Postma JPM, van Gunsteren WF, DiNola A, Haak JR. Molecular 737 dynamics with coupling to an external bath. J Chem Phys. 1984;81:3684–90. 738

30. Essmann U, Perera L, Berkowitz ML, Darden T, Hsing L, Pedersen LG. A smooth 739 particle mesh Ewald method. J Chem Phys. 1995;103:8577–93. 740

31. Hess B. P-LINCS: A Parallel Linear Constraint Solver for Molecular Simulation. J 741 Chem Theory Comput. 2008;4:116–22. 742

32. Domański J, Beckstein O, Iorga BI. Ligandbook: An online repository for small and 743 drug-like molecule force field parameters. Bioinformatics. 2017;33:1747–9. 744

33. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. 745 Gaussian 09, Rev: D01. Pittsburgh, Pa, USA: Gaussian, Inc.,; 2009. 746




31

34. Jencks WP, Regenstein J. Handbook of Biochemistry-Selected Data for Molecular 747 Biology. 1st ed. Sober HA, editor. The Chemical Rubber Co., Cleveland; 1968. 748

35. Villar VH, Merhi F, Djavaheri-Mergny M, Durán R V. Glutaminolysis and autophagy in 749 cancer. Autophagy. 2015;11:1198–208. 750

36. Edwards SR, Wandless TJ. The rapamycin-binding domain of the protein kinase 751 mammalian target of rapamycin is a destabilizing domain. J Biol Chem. 752 2007;282:13395–401. 753

37. Sancak Y, Bar-Peled L, Zoncu R, Markhard AL, Nada S, Sabatini DM. Ragulator-Rag 754 complex targets mTORC1 to the lysosomal surface and is necessary for its activation 755 by amino acids. Cell. 2010;141:290–303. 756

38. Kim SG, Hoffman GR, Poulogiannis G, Buel GR, Jang YJ, Lee KW, et al. Metabolic 757 stress controls mTORC1 lysosomal localization and dimerization by regulating the 758 TTT-RUVBL1/2 complex. Mol Cell. 2013;49:172–85. 759

39. Sun Y, Fang Y, Yoon M-S, Zhang C, Roccio M, Zwartkruis FJ, et al. Phospholipase 760 D1 is an effector of Rheb in the mTOR pathway. Proc Natl Acad Sci U S A. 761 2008;105:8286–91. 762

40. Muckerman JT, Skone JH, Ning M, Wasada-Tsutsui Y. Toward the accurate 763 calculation of pKa values in water and acetonitrile. Biochim Biophys Acta - Bioenerg. 764 2013;1827:882–91. 765

41. Yang H, Rudge DG, Koos JD, Vaidialingam B, Yang HJ, Pavletich NP. mTOR kinase 766 structure, mechanism and regulation by the rapamycin-binding domain. Nature. 767 2013;497:217–23. 768

42. FDA Approval for Everolimus [Internet]. Natl. Cancer Inst. 2013. Available from: 769 https://www.cancer.gov/about-cancer/treatment/drugs/fda-everolimus 770

43. McGranahan N, Swanton C. Clonal Heterogeneity and Tumor Evolution: Past, 771 Present, and the Future. Cell. 2017;168:613–28. 772

44. Aits S, Jaattela M. Lysosomal cell death at a glance. J Cell Sci. 2013;126:1905–12. 773

45. Yoon M-S, Rosenberger CL, Wu C, Truong N, Sweedler J V, Chen J. Rapid mitogenic 774 regulation of the mTORC1 inhibitor, DEPTOR, by phosphatidic acid. Mol Cell. 775 2015;58:549–56. 776

46. Henkels KM, Miller TE, Ganesan R, Wilkins BA, Fite K, Gomez-Cambronero J. A 777 Phosphatidic Acid (PA) conveyor system of continuous intracellular transport from cell 778 membrane to nucleus maintains EGF receptor homeostasis. Oncotarget. 779 2016;7:47002–17. 780

781




32

Figure Legends 782

Figure 1. ICSN3250 specifically inhibited mTORC1 pathway. (a) Chemical structure of 783

ICSN3250. (b) HCT116 cells were treated with the indicated concentration of ICSN3250 784

during 24 h. Cell extracts were analysed by western blot to determine the activation of the 785

indicated pathways. (c - f) mTORC1 and autophagy activation in HCT116 cells treated either 786

with the indicated concentration of ICSN3250 during 24 h (c and e) or with 100 nM of 787

ICSN3250 during the indicated time (d and f). (g-h) Autophagosome formation upon GFP-788

LC3 aggregation in GFP-LC3 expressing U2OS cells treated as indicated for 24 h. The scale 789

bar represents 20 μm. (i) Cell cycle distribution of HCT116 cells treated with the indicated 790

concentration of ICSN3250 during 24 h. (j) SPR analysis of the interaction of ICSN3250 with 791

mTOR (red) FRB domain (blue), and streptavidin as negative control (green). Black arrow: 792

injection of ICSN3250; white arrows: end of the injection. 793

794

Figure 2. ICSN3250 did not act through mechanisms previously described for other 795

mTOR inhibitors. (a) In vitro kinase assay of mTOR in the presence of the indicated 796

concentrations of ICSN3250. (b-c) mTOR localization in HCT116 cells treated with or 797

without 100 nM of ICSN3250 during 24h, as indicated. CD63 was used as a lysosomal 798

marker. (d) mTORC1 activation in HCT116 cells treated with 100 nM of ICSN3250 either in 799

the presence or the absence of amino acids (AA) during 24 h. (e) mTORC1 activation in 800

HCT116 cells transfected with either an empty vector or with a vector expressing Flag-Rheb, 801

and then treated with or without 100 nM of ICSN3250. (f) Immunoprecipitation of mTOR in 802

HCT116 cells treated with or without 100 nM of ICSN3250. 803

804

Figure 3. ICSN3250 antagonized with phosphatidic acid to inhibit mTORC1. (a) 805

mTORC1 and autophagy activation in TSC2+/+ and TSC2-/- MEFs treated with increasing 806

concentrations of ICSN3250 for 24 hours. (b) mTORC1 activation in HCT116 cells 807




33

transfected either with a non-targeting siRNA or with a siRNA against TSC2, and then 808

treated with ICSN3250 for 24h. (c-d) mTORC1 and autophagy activation in HCT116 cells 809

treated with increasing concentrations of PA in the presence of 100 nM ICSN3250. (e-f) 810

Autophagosome formation upon GFP-LC3 aggregation in GFP-LC3 expressing U2OS cells 811

treated with 100 nM of ICSN3250 in the presence or the absence of 100 μM of PA. (g-h) 812

mTORC1 and autophagy activation in HCT116 cells treated with increasing concentrations 813

of ICSN3250 in the presence of 100 μM PA. 814

815

Figure 4. FRB domain of mTOR adopts different conformations in the apo form and in 816

complex with ICSN3250 and PA. (a-b) Representative conformation for FRB domain of 817

mTOR (apo form) extracted from a 100 ns molecular dynamics simulation. (c-f) 818

Representative conformations for complexes between FRB domain of mTOR and ICSN3250 819

(c-d) or PA (e-f) extracted from 100 ns molecular dynamics simulations. The protein and the 820

ligands (ICSN3250 and PA) are shown as surface representations colored in grey, magenta 821

and orange, respectively. 822

823

Figure 5. ICSN3250 and PA have partially overlapping binding sites. (a-d) Residues 824

involved in the interactions with ICSN3250 (a-b) and PA (c-d). The protein is colored in grey 825

and represented in cartoon mode. Protein residues involved in interactions and the ligands 826

ICSN3250 and PA are represented in stick mode and colored in green, magenta and 827

orange, respectively. Ionic interactions and hydrogen bonds are represented as dashed 828

lines. 829

830

Figure 6. Inhibition of mTORC1 by ICSN3250 is responsible for its cytotoxicity in 831

cancer cells. (a) Cell viability of TSC2+/+ and TSC2-/- MEFs treated with or without 832

ICSN3250 100 nM for 72 hours. (b) Cell viability of HCT116 cells transfected either with a 833




34

non-targeting siRNA or with a siRNA against TSC2, and then treated with or without 834

ICSN3250 100 nM for 24 hours. (c) Cell viability of HCT116 cells treated with increasing 835

concentrations of PA in the presence of 100 nM ICSN3250. (d-e) Cell viability (d) and 836

representative microscopy images (e) of HCT116 cells treated with increasing 837

concentrations of ICSN3250 in the presence of 100 μM PA. (f-g) mTORC1 activation (f) and 838

cell viability (g) of HCT116 cells treated either with 100 nM rapamycin or with 100 nM 839

ICSN3250. (h) IC50 values of different mTOR inhibitors in HCT116 cells. 840

841

Figure7. ICSN3250 specifically targets cancer cells both in vitro and ex vivo. (a) Cell 842

viability of both cancer (squares) and non-cancer (circles) cell lines treated with different 843

concentrations of ICSN3250 as indicated. (b-c) Cell proliferation curves of NHDF (b) and 844

HCT116 (c) cells treated with increasing concentrations of ICSN3250. (d) mTORC1 845

activation in NHDF cells treated with increasing concentrations of ICSN3250 for 24h. (e-g) 846

Flow cytometry analysis of GFP-positive HCT116 cells and GFP-negative NHDF or HUVEC 847

cells co-cultured for 72h in the presence of increasing concentrations of ICSN3250. (h-j) 848

Primary cancer cells and primary fibroblast from a colorectal cancer patient were isolated 849

and treated with ICSN3250 100 nM for 72h. Cell viability was determined by microscopy (h) 850

and cell number was quantified (i-j). (k) IC50 values of different mTOR inhibitors in the non-851

cancer cells NHDF. 852




Published OnlineFirst July 27, 2018.Cancer Res Tra-Ly Nguyen, Marie-Julie Nokin, Maxim Egorov, et al. induces enhanced cytotoxicity specifically in cancer cellsmTOR inhibition via displacement of phosphatidic acid

Updated version

10.1158/0008-5472.CAN-18-0232doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2018/07/27/0008-5472.CAN-18-0232.DC1

Access the most recent supplemental material at:

Manuscript

Authorbeen edited. Author manuscripts have been peer reviewed and accepted for publication but have not yet

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/early/2018/07/27/0008-5472.CAN-18-0232To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-18-0232

http://cancerres.aacrjournals.org/content/suppl/2018/07/27/0008-5472.CAN-18-0232.DC1

http://cancerres.aacrjournals.org/cgi/alerts


http://cancerres.aacrjournals.org/content/early/2018/07/27/0008-5472.CAN-18-0232


mtor inhibition via displacement of phosphatidic acid...

Documents