Bendavit_Gabriel_e-thesis

mTOR Transcriptional Regulation by Nrf2 by

Gabriel Bendavit

Principal Investigator

Dr. Gerald Batist

Submitted

April 2015

Department of Experimental Medicine McGill University Montreal, Quebec Canada

A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfillment of the requirements of the degree of

MASTER OF SCIENCE

© Gabriel Bendavit 2015

ABSTRACT___________________________________________

Nuclear Erythroid 2-related factor (Nrf2) is a master transcription factor, and thereby is a

major regulator of cytoprotective responses to oxidative and electrophilic stress. This is

accomplished by recognition and binding to antioxidant response elements (ARE) in the

promoter of target genes, which triggers activation of genes encoding proteins that range

from drug metabolizing enzymes II family to drug efflux pumps. Numerous studies have

shown direct and indirect interactions between Nrf2 and different signaling pathways

including components of the Pi3K/AKT/mTOR signaling pathway.

The potential for a role for Nrf2 in cancer metabolism directed our study towards its

impact on mTOR, the metabolic maestro of this pathway. We observed that modulation

of Nrf2 levels in lung cancer cell lines regulates mTOR protein levels. In order to verify

if this regulation is present at the transcriptional level, we performed both RT-qPCR

analysis and a luciferase assay to functionally analyze the promoter region of this gene

for the presence of functional ARE motifs. We found that transcription of the Mtor

protein was directly modulated by Nrf2 levels in the non small cell lung cancer cell line

A549, as well as in the non-transformed human cell line HEK293. Mutation of the ARE

sequence in the promoter of the mTOR gene, decreased the effect of Nrf2 on an ARE-

luciferase construct’s activity by more than 50%. The physical binding of Nrf2 with the

ARE sequence in mTOR promoter was further confirmed in vitro via DNA pull-down

and EMSA and in vivo via in a ChIP assay. Additional studies show intimate interactions

between other components of the PI3K pathway and Nrf2.

RÉSUMÉ_____________________________________________

Nuclear Erythroid 2-related factor (Nrf2) est un facteur de transcription qui joue un rôle

primordial dans la défense cellulaire contre les stress oxydatif et électrophile. Il régule la

transcription en se fixant sur les éléments de réponse antioxidative (ARE) impliqués dans

la résistance et le métabolisme des médicaments. En outre, plusieurs études montrent des

intercactions directes ou indirectes de Nrf2 avec la voie de signalisation

Pi3K/AKT/mTOR

En se basant sur le rôle de Nrf2 dans le métabolisme du cancer et son interaction avec la

voie de signalisation mTOR, nous avons formulé l'hypothèse selon laquelle Nrf2

régulerait les niveaux de mTOR. Tout D'abord, nous avons observé que la modulation

des niveaux de Nrf2 dans les cellules du cancer du poumon régule mTOR au niveau

protéique. Ensuite, l'utilisation de la PCR quantitative à temps réel et l'essai de

transactivation sur un vecteur rapporteur luciférase contenant le promoteur de mTOR

nous a permis de montrer que Nrf2 régule mTOR au niveau transcriptionnel dans les

cellules HEK293 et A549.

D'autre part, l'introduction des mutations au sein de la séquence de l'ARE du promoteur

de mTOR réduit l'activité luciférase par plus de 50%. Ceci confirme que malgré sa

séquence différente de la séquence consensus, cet ARE est requis pour la liaison et la

régulation de l'expression de mTOR.

l'interaction physique de Nrf2 avec l'ARE du promoteur de mTOR a été confirmé in vitro

par DNA pull down et par retard sur gel (EMSA) et in vivo par immunoprécipitation de la

chromatine. En conclusion, nos résultats suggèrent que le rôle de Nrf2 dans la sensibilité

aux traitements cytotoxiques pourrait découler de sa capacité à réguler l'expression de

mTOR.

TABLE OF CONTENTS_________________________________

Abstract........................................................................................................................... 2

Table of Contents......................................................................................................... 4

1. Introduction......................................................................................................... 7

1.1 Nrf2 and the cap ‘n’ collar (Cnc) family........................................................... 7

1.1.1 Discovering Nrf2………………………................................................... 7

1.1.2 Nrf2 molecular structure ……................................................................... 8

1.2.3 Nrf2 regulation………............................................................................... 9

1.2 Cytoprotective apparatus of cellular detoxification ...................................... 11

1.3 Antioxidant Response Element (ARE)………………………………………12

1.3.1 Discovering the ARE ………………………....…..………....................12

1.4 Nrf2 clinical relevance..………………………………………...................... 13

1.4.1 Nrf2 and carcinogenesis ………………................................................. 14

1.5 Nrf2 cross talk with various pathways involved in cancer………...……….. 15

1.6 The PI3K/Akt/mTOR pathway……………………………………………... 16

1.6.1 Nrf2 interactions with the PI3K pathway….…….……...……..…..….. 18

1.6.2 Clinical relevance of the interaction between Nrf2 and the Pi3K/AKT pathway…………………………………………………………………………. 18

1.7 Nrf2 enhance the PI3K pathway in systems with high metabolic state.……. 19 1.8 mTOR………………………………………………….………………….... 20

1.8.1 mTORC1………………………………………...………..…………... 20

1.8.2 mTORC2……………………………………………..……………….. 21 1.9 Role of Nrf2 on mTOR expression ……..………...…...…………………... 22

2. Hypothesis.......................................................................................................... 22

3. Materials and Methods.................................................................................. 23

3.1 Cell Lines and Tissue Culture/ Transient Transfection…….......................... 23

3.2 Western blot …………………………………………..…..………..………. 24

3.3 Quantitative RT-PCR…………………………………………………………........ 25

3.4 Bioinformatic Analysis………………….…………………………………. 25

3.5 Molecular Cloning and Vector Construction………….……………………. 25

3.6 Nrf2 modulation …………………………………………………..…........... 26

3.7 Luciferase assay constructs……………………………………………...….. 26

3.8 Luciferase Assay…………………………………………….……………… 27

3.9 Electrophoretic Mobility Shift Assay (EMSA)……………………............... 28

3.10 DNA Pull-Down Assay ………….………………………………………... 28

3.11 Chromatin immunoprecipitation ………………………………………..… 29

4. Results................................................................................................................. 30

4.1 Nrf2 modulates mtor expression in A549 cells……..………….….………... 31

4.1.1 mTOR expression when Nrf2 is up-regulated………..……………… 31

4.1.2 mTOR expression when Nrf2 is down-regulated…………………….. 33

4.2 Functional ARE present on mTOR promoter activates its transcription in Nrf2 inducible condition……………………………………………………………… 34

4.3 Nrf2 binds to mTOR promoter region at basal conditions in vitro…………. 36

4.3.1 Nrf2 binding to mTOR promoter region decreases in Nrf2 silencing conditions……………………………………………………………………….. 39

4.3.2 Nrf2 binds to mTOR promoter in vivo at inducible conditions…........ 40

4.4 Expression analyses of the other elements of PI3K pathway due to Nrf2 modulation ……………………………………………………………………... 41

4.4.1 TSC2, S6K and AKT expression when Nrf2 is up-regulated….…… 42

4.4.1.1 TSC2 is a potential indirect Nrf2 transcriptional target at inducible conditions on H460 cells ……………………………………………….. 42

4.4.1.2 At Nrf2 inducible conditions AKT is a possible indirect Nrf2 transcriptional target on H460 cells and posttranslational target on A549 cells……………………………………………………………………... 43

4.4.2 TSC2, S6K and AKT expression when silencing Nrf2…...……..…… 49

4.4.2.1 TSC2, S6K and AKT may be affected post translationally, when Nrf2 is silenced ………………………………………………………… 49

5. Discussion & Conclusions................................................................................. 53

6. Future Directions................................................................................................. 63

7. Acknowledges........................................................................................................ 66

8. References.............................................................................................................. 66

9. Appendix................................................................................................................ 77

1. INTRODUCTION_____________________________________________________

1.1 Nrf2 and the cap ‘n’ collar (Cnc) family

Nrf2 is a basic leucine zipper (bZIP) transcription factor from the cap ‘n’ collar (Cnc)

family. The Cnc domain has 43 conserved amino acids located N-terminal to the DNA

binding domain. Prior to interaction with their target genes, the Cnc family of

transcription factors binds to Maf-recognition elements (MAREs), also known as the

erythroid transcription factor NF-E2 binding sequence(1). Maf (musculo-aponeurotic

fibrosarcoma oncogene) are a family of proteins that lack transcriptional activation

domains. In the nucleus CNC factors function via heterodimerizing with small Maf

proteins, which provide high affinity, sequence-specific DNA-binding activity of the

CNC factors to the MARE element(2).

The Cnc protein family is composed of SKN-1 (Skinhead family member 1) in

Caenorhabditis elegans and Cnc in Drosophila. In vertebrates this family is represented

by, p45 NFE2 subunit(3) and the NFE2-related factors, known as “Nrf” proteins,

Nrf1(NFE2L1/LCRF1/TCF11)(4), Nrf2(NFE2L2) (Itoh et al., 1995)(5), and Nrf3

(NFE2L3)(6). Bach1 and Bach2 (7) are other members of this family, witch however have

no transactivation capacity and instead function as transcriptional repressors. Bach1 is a

truncated isoform of Nrf1, while Bach2 is a caspase-cleaved form of Nrf2. The p45 NFE2

acts during development and is present only in hematopoietic progenitor cells. Besides

their role in early development, the Nrf proteins have a broad and sometimes overlapping

function as stress-activated transcription factors.

1.1.1 Discovering Nrf2

Nrf2 was first isolated and characterized in 1994 by Moi et al,(8) who identified closely

regulated proteins of erythroid-derived 2 (NF-E2). NF-E2, a member of the family of

bZIP transcription factors is a dimeric protein involved in the regulation of the β- globin

gene expression in hematopoietic cells. Nrf2 was named for its ability to bind to the

nuclear factor, NF-E2/ activating protein 1 (AP-1) repeat in the promoter of the β -globin

gene. Tandem Binding of Nrf2 to NF-E2/ AP1 was achieved via expression cloning of

the consensus sequence (5'GCACAGCAATGCTGAGTCATGATGAGTCATGCTG-3')

in K562 erythroid cell line. This repeat sequence is a oligonucleotide containing double-

strand concatemers of the tandem NF-E2/ AP1 repeat of the β- globin locus control

region, DNase I-hypersensitive site 2 (HS2).

1.2.2 Nrf2 molecular structure

The Nrf2 protein, has a molecular weight ranging from 95 to 110 kDa(9), and is

composed of 605 amino acids with 6 functional domains called Neh1-6 (Nrf2-ECH

<chicken Nrf2> homologous domain). The Neh1 holds the CNC homology region and a

basic-leucine zipper domain. It is responsible for heterodimerisation between Nrf2 and

small Maf proteins .The C terminal Neh3 motif is also responsible for Nrf2

transactivation activity (10) The Neh4 and Neh5 are conserved acidic domains that interact

with CBP [CREB cyclic AMP- response element binding protein (CREB) binding

protein], and are responsible for Nrf2 transcription activation strengths(11). Neh6 is a

serine-rich conserved region and serves as a target for a GSK 3 mediated phosphorylation

and consequently proteasomal degradation via ubiquitination(12).

Neh2 is a composite domain that is structurally divisible into two subregions. The

carboxy-terminal of Neh2 (amino acid residues 33–73) is hydrophilic and with no present

functional importance, while the amino-terminal region of Neh2 has 32 amino acids,

which are rich in hydrophobic residues, and shows conservation with Nrf1 and the C.

elegans Skn-1. It is an important functional domain, working as a negative regulator of

Nrf2, proved via domain deletion by Itoh et al(13). They also identified Kelch-like ECH-

associated protein1 (Keap1) responsible for post translational control of Nrf2.

Keap1 is an actin-binding cytoplasmic protein with four main domains, a intervening

region (IVR), double glycine repeat (DGR), C-terminal region (CTR) and broad

complex–tramtrack–bric-a-brac (BTB) domain. The DGR domain, also called Kelch

domain owing to its homology with Drosophila Kelch protein, is important for the

interaction with Nrf2 and for binding to actin. The BTB domain, present in Keap1 C-

terminus, is required for Nrf2 cytoplasmic sequestration and is involved in dimer

formation(14). The IVR domain, which is cysteine-rich protein with 27 cysteine residues,

is important for its reactivity to electrophilic and oxidative stimuli. In the presence of

oxidative stress 10 of these cysteines are activated by positively charged amino acids(15),

which leads to conformational changes in Keap1.

1.2.3 Nrf2 regulation

Keap1 is an important interacting protein of Nrf2 and they form a “hinge and latch”

structure with one another as shown by X-ray crystallography(16). The “hinge” structure is

formed due to a high-affinity interaction of ETGE motif, a stretch of four amino acids

present in the Neh2 domain of Nrf2, with keap1 kelch domain. While the “latch”

structure is generated via low-affinity interaction of DLG motif of nrf2-neh2 domain with

other keap1 monomers(17).

Under basal conditions, the redox–sensitive protein, Keap1 binds Nrf2 to form a

Keap1/Nrf2 complex, and anchors it in the cytoplasm. This cytoplasmic localization was

proved by confocal laser microscopic immunohistochemical analysis, where Keap1 was

shown to be tethered to the actin cytoskeleton(18). As others broad complex–tramtrack–

bric-a-brac (BTB)-containing proteins, Keap1 is an adaptor protein for the Cullin 3

ubiquitin E3 ligase (Cul3) which is a scaffold protein in the E3 ligase complex and forms

a catalytic core complex together with roc1/rbx1/Hrt1.The cognate E2 enzyme is then

recruited by Roc 1. This way, Nrf2 is specifically targeted (Lawah Zellers) for

degradation by the ubiquitin-proteasome pathway by 26 S proteasome(14).

In situations of oxidative stress, Keap1 undergoes conformational changes, which result

in the breakdown of the Nrf2-Keap1 complex. This occurs due to the difference in

affinity of “hinge” and “latch,” interactions, which have a difference of 2 orders of

magnitude, caused by the variance in the number of electrostatic interactions between

each domain and Keap1. This difference in affinity, weakens the interaction of the DLG

motif leading to the Nrf2-Keap1 complex disruption(17,19). This culminates in the release

of Nrf2 and its translocation to the nucleus, where it accumulates and activates the

cytoprotective program. Prior binding to its target genes, Nrf2 forms a heterodimer with

members of the small Maf family. This hetero-dimerization happen in the Nrf2 Neh1

domain. The complex Nrf2/small Maf then binds antioxidant response elements (AREs)

localized in the promoter region of its target genes(20).

Apart from the Keap1 mechanism of post-translational regulation of Nrf2, it is known

that some kinases, such as p38 kinase (21)and PTEN, can inhibit Nrf2. Kensuke Sakamoto

et al(22) showed via chromatin immunoprecipitation of Jurkat human leukemia, baring a

PTEN mutation, that the PI3K inhibitor LY294002 blocks CBP and Nrf2 recruitment to

ARE while it releases Bach1 to ARE. Glycogen synthetase kinase 3 (GSK-3ß) is also a

Kinase that can inhibit Nrf2.(12,23,24)

The serine/threonine GSK-3ß protein regulates glycolytic metabolism and directs the

ubiquitination and proteasomal of a variety of transcription factors(24). GSK-3ß is

involved in metabolic processes such as glycogen metabolism, Wnt signaling and

sensitization to oxidative-stress-mediated apoptosis. GSK-3ß is negatively regulated by

the Ser/Thr kinase Akt(25). AKT phosphorylates GSK-3ß’s Ser-9 in its pseudosubstrate

domain which inactivates GSK-3ß and consequently inhibits apoptosis. In order to

understand the mechanistic connection between the phase II genes’ cyto protection

against oxidative stress and the PI3K survival pathway, Salazar et al (23) focused on

control of nuclear Nrf2 accumulation. They suggested that Nrf2 was negatively regulated

via GSK-3ß phosphorylation in the nucleus post-translation. This study found that Nrf2

contains a consensus sequence for GSK-3ß phosphorylation (S/T)XXX(S/T) which was

confirmed by both immunocytochemistry and subcellular fractionation analyses. In a

following study by Rada P et al(24), it was demonstrated in mouse, that GSK-3ß acts as an

adapter protein for Nrf2 by phosphorylating a group of Ser residues in its Neh6 domain

and consequently targeting it to the SCF/ ß -TrCP SCF protein.

There is thus evidence for interaction between elements of the PI3Kinase pathway and

Nrf2 transcription factor. To date that data demonstrates regulation of Nrf2 by proteins

such as p38 Kinase, PTEN and GSK-3ß

1.2 Cytoprotective apparatus of cellular detoxification

In normal physiological conditions, nuclear factor NRF2 is essential for cell homeostsis

against endogenous and exogenous redox stress. This master cytoprotective transcription

factor is responsible for the activation of phase II detoxifying enzymes, antioxidants,

phase III drug efflux pumps and transporters(26).

The cytoprotective apparatus of cellular detoxification has been stratified into 3

categories phase I, II and III drug metaboling enzymes (DMEs). The phase I and II

enzyme systems are localized in the endoplasmic reticulum (ER) while the phaseIII is

present in the cytoplasmic membrane.(27) Phase I is composed of cytochrome

P450s(CYPs) gene superfamily. These large hydrophobic organic molecules are

responsible for oxidation and reduction by introducing polar functional groups into

nonpolar molecules. This group of enzymes are regulated by, ligand activated, Aryl

hydrocarbon receptor (AHR) transcription factor. DNA sequences called xenobiotic

response elements (XREs) are present in the promoter region of Phase I DMEs and are

essential for the regulation of these classes of enzymes. XREs are the target regions for

AHR binding, which activate transcription, after chaperoning with a nuclear transporter

called ARNT. There are growing evidences that Nrf2 regulates AHR, thus also phase I

DMEs.(27,28)

The phase II DMEs are Nrf2-dependent gene battery that includes enzymes acting on

cellular redox status and cell protection against oxidative damage, cytotoxicity,

mutagenicity and carcinogenicity. Phase II DMEs works synergistically with phase III

DMEs transporters in various metabolic reactions. Together, their functions involves

disposition of xenobiotics, and endogenous substances (26). Some of the phase II DMEs

are glutathione S-transferases (GSTs), sulfotransferases (SULTs) UDP-glucuronosyl

transferases (UGTs) FAD containing flavoprotein NAD(P)H:Quinone

Oxidoreductase(NQO1), Heme oxygenase (HO-1). These are involved in catalyzing

conjugation reactions through covalent linkage of xenobiotics or phase reaction products,

to groups that are more functionally polar (glucuronate, sulfate, amino acids and

glutathione) which occurs via nucleophilic trapping. In this context, GSTs assign

glutathione, a cellular nucleophile, to electrophilic xenobiotics (9). Similar mechanism is

also seen in SULTs and methyltransferases(29,30). The other category of enzymes present

in the DME phase II is represented by UGTs. These conjugate adenosine-containing

cofactors with nucleophilic xenobiotics. Superoxide dismutases, glutathione peroxidase,

and catalase such as the NQO1 function in a similar manner. The detoxification

mechanism of NQO1 involves catalyzing quinone to hydroquinones via two electron

reduction, bypassing the formation of highly reactive semiquinone(30). Phase II DMEs are

also represented by thiol-containing molecules, such as, glutathione and thioredoxin and

HO-1. HO-1 is an essential enzyme in heme catabolism and is responsible for cleaving

heme to form biliverdin, which is ultimately converted to bilirubin. (27)

The third category is composed of membrane efflux transporters such as the multidrug

resistance associated proteins (MRPs 1,2,3 and 4).). The MRPs are adenosine

triphosphate-dependent drug transporters. They are responsible for the excretion of

endogenous substances, such as bilirubin and xenobiotis, together their conjugated

metabolites products from the DME phase II enzymes.

1.3 Antioxidant Response Element (ARE).

The phase II and III DMEs reach their highest level of expression primarily through

activation of a specific enhancer in their respective promoter region. These enhancers are

cis-acting regulatory elements, called antioxidant response element (ARE). Present in

phase II and III enzymes, ARE regulate the expression of genes involved in the cellular

redox status and are present as a single or multiple copies(27).

1.3.1 Discovering the AREs

The ARE pathway was originally observed by Talalay et al(31), when analyzing the

different ways by which some xenobiotics regulate Phase I and Phase II drug-

metabolizing enzymes. This was the first evidence of a Phase II enzyme induction.

Further studies were done(32) in order to identify trans-acting proteins that interact with

these cis-acting regulatory elements. These were classified and characterized by

Rushmore and Pickett(33) after identification of oxidative responsive elements and basal

promoter elements in the rat GST Ya subunit (Gsta2) gene. This novel Cis-acting element

in the 5'-flanking region element, when used in a reporter construct was shown to induce

the activity of the phenolic antioxidant tert-Butylhydroquinone(tBHQ), hence the name

antioxidant response element. The ARE core sequence (cARE), 5′-TGACnnnGC-3′ was

determined via deletion and mutational analysis. Jaiswal el al (34) established the role of

Nrf2 as a transcription factor for genes containing ARE in their promoter region, hence

regulating expression of genes affecting xenobiotic metabolism. The NQO1 induction by

Nrf2 and Nrf1 was shown via supershift assay after transient transfection of these

transcription factors into human hepatoblastoma HepG2 cells. More experiments(35)

involving a broad spectrum of Nrf2 inducers demonstrated the activation of various phase

II DMEs by Nrf2. Sternberg et al 2006 (36) used high-performance liquid chromatography

(HPLC) to show that retinal pigment epithelium (RPE) cells, when treated with zinc,

increased the levels of glutathione synthesis through Nrf2. The cARE motif was further

confirmed as a binding site for Nrf2 via numerous ChIP-seq methodologies followed by

global transcriptional profiling, which demonstrated the variety of Nrf2 proteins

interactions. In recent literature, Biswal et al 2010(37) performed a global Nrf2 ChIP-seq

analysis of mouse embryonic fibroblasts (MEF) with either constitutive nuclear

accumulation (Keap1-/-) or depletion (Nrf2-/-) of Nrf2. Integrating ChIP-Seq and

microarray analyses, they identified 645 basal and 654 inducible direct targets of Nrf2,

with 244 genes overlapping a microarray datasets used to identify Nrf2 direct

transcriptional targets. Also, Chorley et al 2012(38) performed another ChIP-seq analysis

of NRF2-regulated genes utilizing the same cARE motif. Utilizing lymphoid cells with

Nrf2 induced by isothiocyanate, sulforaphane (SFN) they were able to identify 242 high

confidence genomic regions to which Nrf2 binds.

1.4 Nrf2 clinical relevance

There is abundant evidence of Nrf2 involvement in direct protein interactions and

pathway cross talk. This complex regulatory system generated by Nrf2 interactions is

reflected in the clinic by the vast variety of pathologies in which it is involved. In mice

Nrf2 was shown to play a role in carcinogenesis, chronic obstructive pulmonary disease,

obesogenesis, and neurodegeneration(39). Although, Nrf2 knock out was shown to be

nonessential for the normal development in mice(40), the Nrf2/ARE interaction is vital in

humans for normal cell homeostasis promoting cellular antioxidant defenses and

increased capacity to detoxify drugs. Previous studies with Nrf2 _ / _ mouse models(41)

have shown a high sensitivity of mice to chemical and physical insults. As previously

mentioned, these insults have a strong correlation with the incidence of cancer via

oxidative and electrophilic stressors, or drugs that induce the production of free radicals.

It was also shown that Nrf2-deficient mice seemed to be more sensitive to

carcinogenesis,(42,43) and are at an enhanced risk of metastasis(44),(45).Consequently, Nrf2

was considered to work only as tumor suppressor and so the benefits of Nrf2 signaling in

cancer chemoprevention were largely explored (46).

However, this increase in cellular protection, via high Nrf2 levels, leads to unwanted side

effects in some cancer types(47), as constitutive activation or augmented signaling of the

Nrf2 pathway may promote tumorigenesis and be involved in resistance to chemo- and

radiotherapeutic treatments, showing that the transcription factor could have a proto-

oncogenic role(48).

1.4.1 Nrf2 and carcinogenesis

The role of Nrf2 in cancer promotion was first found in an hepatocellular carcinoma

model by Ikeda et al in 2004(49). In this study both levels of Nrf2 and GSTP1, a neoplastic

marker, were elevated. It was also found that Nrf2 was regulating GSTP1 through an

ARE, present in the promoter region of the gene. Additional studies have proven Nrf2

relation to tumorigenesis, chemoresistance, increased cell survival, metastasis, and cell

growth (47)-(49,50)-(51).

While much focus remains on enhancing Nrf2 as a cancer chemoprevention strategy

against genotoxic agents(52), (53) or inflammation(54), participation of Nrf2 in the process of

carcinogenesis is also strongly demonstrated in many papers in the literature. Nrf2

together with its downstream genes, is elevated in many cancers cell lines and human

cancer tissues, resulting in chemoresistance(50) and a poor prognosis in patients (55,56, 59,60)

thus providing the cancer cells an advantage for survival and growth.

One of the principal reasons for the constitutively high levels of active Nrf2 in cancer is

due to loss-of-function mutations in Keap1.(16,55) which causes its inactivation or reduced

expression. This results in increased Nrf2 stability and its translocation to the nucleus and

consequently transcriptional activation of its target genes. Constitutive stabilization of

NRF2, due to Keap1 mutations, was found in various human cancers, with increased Nrf2

activity in lung (~40%), head and neck (~20%), gallbladder (~30%), liver, and breast

cancers(56). There are also some cell lines in which gain-of-function mutations in the Nrf2

gene is observed, (56-58)like in advance Esophageal squamous cancer (ESC) with

occurrence of (18/82, 22%)(50)

In both, in-vivo and in clinical specimens of non-small cell lung cancer (NSCLC)(55),

loss-of-function Keap1 mutations resulted in constitutively high levels of active Nrf2 and

subsequent resistance to chemotherapeutic drugs (taxanes, platinums) and radiotherapy.

Keap1 mutations are reported in up to 60% of papillary lung adenocarcinoma, as well as

in other cancers including ovarian, gall bladder and others(59).

The inverse of the abovementioned is also the case. A low level of Nrf2 within the cancer

cells is responsible for chemo sensitisation. Batist et al 2009(51) found very low Nrf2

levels in breast cancer cell lines and in the majority of a 200-sample tissue microarray,

which is consistent with the high response rates of breast cancer to many cytotoxic

therapies.

1.5. Nrf2 cross talk with various pathways involved in cancer

As mentioned before Nrf2 can block cell damage induced by oxidative and electrophilic

drugs and also reduce their accumulation in the cell via MDR protein. However, Nrf2

chemoresistance can also occur, due to its interaction with other pathways present in

cancer, which are related to metastasis, increase in cell survival and cell growth.

Some Nrf2 target genes, such as HO-1, were shown to be related to cellular metastatic

potential. HO-1 is overexpressed in various solid tumors(60) and is related with

angiogenesis and acceleration of prostate cancer progression(54). The HO-1 protein is also

related with increased cell survival via apoptosis inhibition in chronic myelogenous

leukemia (CML). Nrf2, also, regulates proteins from the Bcl-2 family through

transcriptional control of the antiapoptotic proteins Bcl-2 and Bcl-XL. Additionally, Nrf2

was shown to increase cell survival via inhibition of p53-dependent apoptosis(61). In

response to stress stimuli, the tumor suppressor p53, control the expression of the

cycling-dependent kinase inhibitor p21 via cell cycle G1 arrest(62). Nrf2 is stabilized by

p21 via direct interaction of the DLG and the ETGE Nrf2’s motifs with the KRR motif in

p21, which displaces the Nrf2-Keap1 interaction(63). In a ROS-dependent mechanism,

p53 induces apoptosis via a two-phase Nrf2 response. Under conditions where ROS

levels are low, in a phase called induction, p53 is also low and it enhances the protein

level of Nrf2 transcriptionally via the target gene p21. The other side of this biphasic

regulation is called the repression phase, and it is present when ROS, and consequently

p53 levels, are high. In this phase p53 binds to a sequence near the ARE which repress

Nrf2 transcription by displacing it from the ARE(61,64). P53 was also showed to negatively

regulate TSC2, PTEN, consequently inhibiting the IGF-1-AKT-mTOR axis. (65) This

suggest at least an indirect relationship between Nrf2 biding to its cognate sequence

(ARE) and elements of the PI3K pathway, including mTOR.

1.6 The PI3K/Akt/mTOR pathway

The phosphatidylinositol 3-kinase (PI3K)/Akt/mTOR pathway is important for cell

survival and is involved in metabolism, apoptosis, cell growth, differentiation, calcium

signaling, and insulin signaling(66). In addition to those cited above, a variety of recent

studies suggest that this pathway interacts with Nrf2(67,68). PI3K/AKT pathway has a role

in tumor development and has shown potential in tumor treatment, through the PI3K

pathway inhibitor Wortmannin(66) (69). Multiple molecules that target this pathway are

currently in clinical development.

PI3Ks are part of a lipid Kinase family with main distinctive feature is its capability to

phosphorylate inositol ring 3’-OH group in inositol phospholipids. The mechanism of

action of this signaling pathway starts with PI3K activation. One mode of activation is

through binding of an extracellular growth factor to the RPTK (Receptor Protein

Tyrosine Kinase). Binding of this receptor by growth factors lead to dimerization of

RPTK monomers along with heterologous auto phosphorylation of this receptor

monomers, the IRS-1 (insulin receptor substrate I) then binds to a phosphorylated IGF

receptor. This complex function as binding and activation site for PI3K. Another mode of

activation is via direct binding to a phosphorylated receptor Tyrosine Kinase. This

pathway can also be activated by binding of PI3K to a small membrane bound, active

GTP-Ras(66).

The next step of this pathway involves activation of the second messenger

phosphatidylinositol-3,4,5-trisphosphate (PIP3) and AKT (a serine/treonine kinase

protein also known as protein kinase B). Migration of PI3K to the inner membrane and

binding to PIP2 (Phosphatidylinositol 4,5-bisphosphate) leads to phosphorylation of PIP2

to PIP3 which then activates AKT. The Pi3K pathway is negatively regulated by the

presence of phosphatases capable of dephosphorylating PIP3 back to PIP2. Inhibition of

this pathway can be achieved via chromosome 10 (PTEN) barring a homologue deletion

of phosphatase and tensin. Decrease in PTEN expression indirectly stimulates PI3K

activity and is largely seen in cancer(66).

There are at least four main downstream effects of AKT activation. The first one is the

inhibition of apoptosis via binding with BAX (BCL2-associated X protein) which in-turn

stops BAX from creating holes in the mitochondria inner membrane, responsible for

generating apoptosis by the Caspase cascade. The second effect is the phosphorylation of

Forkhead box O (FoxO) which serves as a substrate for the enzyme ubiquitin ligase,

resulting in its degradation in the proteasome. In the absence of this process FoxO

inhibits cell proliferation. The third effect is the inhibition of Glycogen synthase kinase-

3ß (GSK-3ß). The fourth effect is its role in translation by a multi step protein cascade.

This cascade begins with the activation of Rheb by AKT, which activates the protein

kinase mechanistic target of rapamycin (mTOR; formerly known as mammalian TOR)(70).

Another mechanism of mTOR activation via AKT is by phosphorylation o the mTOR

inhibitor PRAS40 (proline-rich Akt/PKB substrate 40 kDa)(71).

1.6.1 NRF2 interactions with the PI3K pathway

As noted, several studies showed evidence for interactions between the PI3K pathway

and NRF2 using different techniques and models. In the previously mentioned global

mapping of Nrf2 biding sites(37), TSC2 was shown to be a basal target for Nrf2; since the

cells were not “stimulated” in any way with respect to Nrf2 function or nuclear

accumulation, this type of study is mute on Nrf2’s potential role in the transcription of

these proteins in conditions of redox stress.

In an in silico analyses of Nrf2 interactome and regulome, that includes 289 protein–

protein, 7469 TF–DNA and 85 miRNA interactions, shown in a manually curated

network of Nrf2, it was observed that AKT functions as an indirect activator of Nrf2 (67).

Biological evidence of this interaction was also observed in previous studies where

human dopaminergic neuroblastoma SH-SY5Y cells(72) showed PI3K involvement in the

Nrf2 regulation of antioxidative proteins HO-1, Trx, and PrxI, According to the paper,

after treating the cells with hemin, a dose dependent nuclear translocation of Nrf2 was

observed together with PI3K phosphorylation. Also, PI3K inhibitors, wortmannin and

LY294002, lead to inhibition of Nrf2 nuclear translocation. In another study(68), Nrf2 up

regulation via the PI3K and the Extracellular Regulated Kinase (Erk) pathways was

observed after cell treatment with eckol, which is a phlorotannin component of brown

algae such as Ecklonia cava (Laminariaceae), and is known to upregulate ERK and AKT

individually. In this paper it was also shown that treatments with any of the drugs (

U0126, an Erk kinase inhibitor, or LY294002) or short interfering RNAs (Erk1 siRNA,

and Akt siRNA) suppressed Nrf2 activity, which was observed by decrease of HO-1

levels.

1.6.2 Clinical relevance of the interactions between Nrf2 and the Pi3K/AKT

pathway

Interaction between the PI3K/AKT pathway and Nrf2 might well be clinically relevant,

as the pharmacological inhibition of this pathway suppresses the nuclear translocation of

Nrf2 in cancer cells (73,74). This was also shown by Ling Wang et al,(75) who working on

age-related macular degeneration (AMD) caused by accumulated oxidative injury, found

that cultured human retinal pigment epithelium (RPE) cells treated with PI3K inhibitors

were able to decrease Nrf2 levels. Additionally, a study by Papaiahgari et al 2006(76)

showed that PI3K/Akt signaling regulates Nrf2 activation by hyperoxia. Lung injury due

oxygen supplementation (hyperoxia) is currently used in the treatment of pulmonary

diseases such as respiratory distress syndrome (ARDS) and emphysema. PI3K inhibition

blocked hyperoxia-stimulated Akt and ERK1/2 kinase activation, which activate Nrf2

transcriptional activity. Nrf2 regulation by AKT was later shown to occur via inactivation

of GSK-3b(12).

1.7 Nrf2 enhances the PI3K pathway in systems with high metabolic state

There is growing evidence that Nrf2 also enhance the PI3K pathway in systems with a

high metabolic state (74-77). A hyperproliferative phenotype is a fundamental feature of

tumor growth, and this depends on the metabolic reorganization of elements involved in

bioenergetics, macromolecular synthesis, and cell division(77). Besides Nrf2’s role in

cancer cell resistance to cytotoxic agents, it also cross-talks with other pathways

responsible for modulating metabolism and cell growth, including PI3K/AKT/mTOR and

MAP/ERK pathways. In this context, Nrf2 was observed to mediate NSCLC cell

proliferation via activation of the epidermal growth factor receptor EGFR/MEK1-2/ERK

axis. In the NSCLC H292 cell line, which expresses both wild-type EGFR and Keap1,

EGFR ligand was shown to increase Nrf2 levels in a dose-dependent manner via the

MAP/ERK pathway(78). Also, when EGFR is constitutively active, due to gain of function

mutations, Nrf2 is permanently active(78).

Nrf2 was shown to reinforce the metabolic reprogramming triggered by proliferative

signals. Mitsuishi, Y et al(79) has shown that in the presence of active PI3K-Akt signaling,

combined with high Nrf2 levels in the cell, higher than the ones required for the

transcription of antioxidant target genes, Nrf2 redirects glucose and glutamine into

anabolic pathways. Direct Nrf2 transcriptional targets are associated with de novo

nucleotide synthesis via the pentose phosphate pathway (PPP). AKT activation via Nrf2

was observed in another study of liver repair in mice NRF2 KO mice(80) . As expected,

Mitsuishi, Y et al(79) also found AKT to be phosphorylated in a Nrf2 dependent manner,

thus activating the AKT/mTORC1/Sterol Regulatory Element-Binding Proteins (SREBP)

axis. SREBP is a transcription factor known to induce the PPP genes when mTORC1 is

activated (81).

1.8 mTOR

mTor (also known as RAFT1 or FRAP) is a vital cell metabolic regulatory component of

the PI3K pathway, indirectly activated by AKT via Rheb. mTOR plays a central role in

various signaling pathways, is responsible for the intra and extra cellular detection of

nutrients levels, and functions as a metabolic regulator of cellular anabolic and catabolic

processes coupling growth signals to nutrient availability via ribosome biogenesis and

autophagy(82-84)

The mTOR protein has a molecular weight, of 289 kDa and contains 2549 amino acids

with several conserved structural domains. The N terminus possesses 20 tandem

Huntington, EF3, A subunit of PP2A, TOR1 (HEAT) repeats, forming two α helices of

40 amino acids with hydrophobic and hydrophilic residues. These HEAT repeats are

responsible for protein-protein interactions. The kinase domain of mTOR is located in the

C-terminal. The FKBP12-rapamycin-binding (FRB) domain is located upstream of its

catalytic domain and is, responsible for the formation of the rapamycin inhibitory

complex. Near FRB domain a large FRAP, ATM, TRAP (FAT) domain is present. This

FAT domain is essential for mTOR activity because of its interaction with another FAT

domain, present in the end of the C terminal domain, called FATC. The interaction

between those two domains produces a configuration that exposes the catalytic domain.

Between the FATC and the catalytic domain there is a putative negative regulatory

domain (NRD)(82).

1.8.1 mTORC1

mTOR is part of two functionally and structurally distinct complexes, namely,

rapamycin-sensitive mTOR complex 1 (mTORC1) and rapamycin-insensitive mTOR

complex 2 (mTORC2). mTORC1 is related to regulation of translation, autophagy, cell

growth, lipid biosynthesis, mitochondria biogenesis, and ribosome biogenesis. The

downstream effects of mTORC1 are initiated by its interaction with the accessory protein

regulatory-associated protein of mTOR (Raptor). This interaction mediates the

association of this complex to a conserved short sequence called the TOS motif of S6K

and the eukaryotic initiation factor 4E (eIF4E)-binding proteins (4E–BP1 and 2). Once

bound, the raptor–mTORC1 complex phosphorylates S6K, and 4E–BP, which are

markers for mTORC1 activity. S6K is phosphorylated on its Thr389 site, and functions to

enhance the translation of 5′-terminal oligopolypyrimidine (5′-TOP) mRNA’s via

activation of 40S ribosomal subunit. These activated mRNA’s encode anabolic elements

such as, ribosomal proteins, elongation factors and insulin growth factor 2(83,84).

In its non-phosphorylated form 4E-BP binds to eIF4E at the 5 ́-cap of mRNAs, inhibiting

the interaction of eIF4E with eIF-4G protein, consequently arresting initiation of

translation. The 4E-BP/ eIF4E complex is released after 4E-BP phosphorylation by the

raptor–mTOR complex. Therefore enhancing cap-dependent protein translation via eIF4E

activation, resulting in a global boost of cellular protein synthesis and ribosome

biogenesis. Anabolic processes generated by mTORC1 also involve stimulation of

glucose uptake, glycolysis and NADPH production. One of the mechanisms that generate

these effects is the increase in translation of hypoxia-inducible factor 1α (HIF1α),

resulting in higher levels of glucose transporters and glycolytic enzymes(83,84).

1.8.2 mTORC2

The second mTOR complex, mTORC2, interacts with rapamycin-insensitive companion

of mTOR (RICTOR) which is a hydrophobic motif kinase for Akt/PKB activation. Akt is

a vital element of the insulin/PI3K signaling pathway and regulates the influx of nutrients

that activate the raptor–mTOR pathway. The role of mTORC2 in cancer is well

documented (79,80). This complex is hyper activated in cancers via inactivation of the

tumor suppressor PTEN. mTORC2 is known to control cell survival and proliferation by

enhancing the p53-regulator mdm2 and transcription factors from the FOXO family(83,84).

There are a myriad of known mTOR regulators such as growth factors, amino acids,

glucose, energy status, stress (e.g. osmotic stress, DNA damage) and, the tumor

suppressors TSC1 (hamartin) and TSC2 (tuberin). The TSC1/2 complex indirectly

inhibits raptor–mTOR by working as a GTPase-activating protein (GAP) for rheb, a

GTP-binding protein from the ras-family that activates raptor–mTOR by direct biding(69).

mTOR complex 1 activity is also regulated by Rheb via RagD. This member of the

small G-protein family binds directly to the mTOR complex, recruiting it to the

endosomal fraction where mTOR is activated(85).

Using the UCSC genome browser we identified an extended list of ubiquitous

transcription factors acting on mTOR including SP-1, C-MYC and C-FOS. From this list,

the activating factor (ATF-5) was mentioned in the literature. ATF-5 is a member of the

cAMP response element binding (CREB)/ATF subfamily of basic leucine zipper

transcription factors(86). It was shown that the oncoprotein BCR-ABL suppresses

authophagy by up regulating ATF-5 via PI3K/AKT/FOXO4 signaling(87). ATF-5 then

activates mTOR by a direct binding to its promoter, which is in a region between 1560-

2227 bp upstream of the transcription start site, as demonstrated via ChIP assay(87).

Interestingly, a member of the same group of transcription factors, ATF3, is known to

inhibit Nrf2 via direct ATF3-Nrf2 protein-protein interactions(88). Nrf2 belongs to the

same family of transcription factors as ATF and has already been shown to indirectly

interact with mTOR via TSC2 and AKT.

1.9 Role of Nrf2 on mTOR expression

Due to the multi-level interaction of Nrf2 with the PI3K pathway we were interested to

know if Nrf2 could directly act on different components of this pathway. Recent studies

of Nrf2 participation on translation and in cancer anabolism focused our attention to the

metabolic regulator of this pathway, mTOR.

2. HYPOTHESIS________________________________________________________

From literature it is observed that Nrf2 interacts with different components of the PI3K

pathway and regulate specific processes. Recently, Nrf2 has been shown to be involved in

the regulation of metabolic processes in the cell(78) - (80) and hence, we hypothesized that

Nrf2 might also be interacting directly with mTOR, which has not previously been

shown. If demonstrated, this would be one of the possible pathways in which Nrf2

directly regulates the metabolic processes of the cell, positioning it as a link between cell

metabolism and cytoprotection.

To examine this hypothesis, mTOR expression was analyzed using western blot and RT-

PCR in conditions where Nrf2 levels are modulated. Our experiments were focused in

three different cell lines, selected according to the mutations present in them. We used

the non-tranformed Human Embryonic Kidney (HEK293) cells, as well as two human

non-small-cell lung cancer (NSCLC) cell lines. A549 cells have a Kras mutation in

addition to mutations in keap1. Another NSCLC cell line H460, contains a loss of

function mutation on keap1(55) and gain of function mutation on PIK3CA (E545K) and

Kras(89). To further study the Nrf2/mTOR interaction we performed mutation analysis in

dual luciferase assay, as well as DNA pulldown, electroctrophoretic mobility shift assay

(EMSA) and ChIP assay. Additionally, we analyzed the expression of the other elements

of the PI3K pathway (TSC2, S6K and AKT), under Nrf2 silencing and inducing

conditions, via western blot, RT PCR and luciferase assay.

3. MATERIALS & METHODS____________________________________________

3.1 Cell Lines and Tissue Culture/ Transient Transfection

The cell lines A549, HEK293 and H460 (Sigma) were cultured in RPMI (Sigma) media,

supplemented with 10% fetal bovine serum (Sigma), 5% antibiotic/antimycotic (Life

Technologies) and grown in 5% CO2 at 37°C. The cell lines were storage at -80oC in

cryogenic vials containing 106 cell in 1 ml solution of 90% FBS plus 10% DMSO.

Twenty-four hours prior to transfection, 9 X 104 cells were plated in 6 well dish plates

and were transfected when they were approximately 60% confluent. The cells were

incubated with fresh media 1 hour before transfection. The transfections, except for the

ones utilized on ChIP assay, were carried out using Lipofectamine LTX Reagent PLUS™

(Life Technologies) as per manufacturer protocol, utilizing Opti-MEM with a 1:5 ratio of

plasmid to LTX and Plus reagent. The transfection mix was vortexed thoroughly and

incubated for 30 min before addition to cells. Cells were incubated 24 hours before

collection. The internal control used for Luciferase assay, pRL Vector, was co transfected

with the modulatory reagents (pCDNA_Nrf2 or siNrf2 with their respective controls

pCDNA 4.0 or scrambled RNA) and the construct containing the sequence of interest, in

1:1 ratio. 24 hours after transfection the cells were harvested and split for Western blot,

qPCR and luceferace applications.

3.2 Western blot

Protein expression analysis of the cells A549, H460 and HEK293 were performed by

Western blot. Cells were disrupted with lysis buffer (20mM Tris pH 7.5, 420mM NaCl,

2mM MgCl2, 1mM EDTA, 10% glycerol, 0.5% NP-40, 0.5% Triton, 1x P8340 (Sigma),

1mM PMSF, 1mM DTT, 2mM NaF, 10mM BGP) for 30 min on ice followed by a 20

min spin at 13000rpm to pellet debris. The supernatant was then removed and quantified

using the Bradford reagent. The OD595 of each sample was then measured using a

spectrophotometer and compared to a standard curve prepared with bovine serum

albumin. An equal concentration of sample was then separated using standard Sodium

Dodecyl Sulfate-Polyacrilamide Gel Electrophoresis (SDS-PAGE) techniques. 40 µg of

cell protein/lysate per each sample was loaded and run through a 10% SDS-PAGE gel

before transferring electrophoretically at 400mA for 2 hours onto a BioRad

nitrocellulose membrane. For the incubation with antibodies, the membrane was first

blocked with 10% fat-free milk solution in 1x Tris Buffered Saline and 0.1% Tween

(TBS-T) for 1 hour at room temperature and probed overnight at 4oC with the antibodies

listed below at the dilutions provided by the manufactor. The day after, membranes were

washed three times in TBS-T and were then incubated with secondary anti-mouse or anti-

rabbit horseradish-peroxidase for 1hour at room temperature. This was followed by three

additional washes with TBS-T.

The results were documented on x-ray film with ECL detection and autophotography to

capture the differences in protein levels in the cells between samples. The antibodies used

as probes for Western were as follows; Nrf2 (abcam) all the others antibodies, beta-Actin,

TSC2, AKT, S6K and Nqo1 were purchased from Cell signaling.

3.3 Quantitative RT-PCR

Total RNA was isolated from, HEK293, A549 and H460 using EZ-10 DNAaway RNA

Mini-Preps Kit (Bio Basic Canada INC.) according to the manufacturer's protocol.

cDNAs were synthesized from total RNA (1 µg) of each sample using , SuperScript® II

Reverse Transcriptase (Invitrogen™)), diluted 4 times with water. The cDNA was used

as the template for quantitative PCR detection using the GoTaq® qPCR Master Mix

(Promega). The real-time PCR conditions were optimized as 95 °C for 7 min and 40

cycles of 95°C for 10 s, 61°C for 5 s, and 72°C for 20 s followed by melting curve cycle.

The amplification reactions were carried out with the AB Applied Biosystems 7500 Fast

Real-Time PCR System. The primers for amplifying human genes (Nrf2, mTOR,Nqo1,

HMOX1, TSC2, AKT, S6K and Gapdh)appendix(Table 1). The comparative ΔΔCt method was

used for relative quantification of the amount of mRNA in each sample normalized to

GAPDH transcript levels. Fold induction is expressed as the ratio of induction from

treated cells versus untreated. Values represent the mean +- S.E. of three independent

measurements. Statistical analysis (Student’s t test) was performed by comparison of

treated and untreated cells (*, p < 0.05).

3.4 Bioinformatic Analysis

We screened for the presence of the core ARE sequence (TGAxxxxGC) up to 5kb

upstream of the transcription start site of the target genes. This ARE motif analysis was

performed using BlAST, SCOPE and InSilicase algorithms.

3.5 Molecular Cloning and Vector Construction

Primers were designed using the Primer3 software (http://fokker.wi.mit.edu)2,

synthesized by Integrated DNA Technologies, Montreal, QC. PCR was done according to

the Phusion® High-Fidelity DNA Polymerase protocol (Thermo). Sanger DNA

sequencing at the Innovation Centre, located at McGill University, confirmed the

presence of the desired promoters. The restriction enzymes used on molecular cloning

were purchased from Invitrogen™

3.6 Nrf2 modulation

Inducible Nrf2 construct – The inducible construct PC_Nrf2 appendix (figure 1A) containing

1925bp of the Nrf2 coding sequence was obtained by amplifying the coding sequence of

Nrf2 from A549 RNA (cDNA). Restriction sites for BamHI and XbaI were included in

the primers used for Nrf2 amplification, and enabled the insertion of Nrf2 cDNA into the

pCDNA 4.0 plasmid (Life Technologies). The resulting construct, PC_Nrf2, was

sequenced to validate the plasmid identity. Nrf2 induction was generated via transient

transfection of inducible PC_Nrf2 plasmid. pcDNA 4.0 was used as a negative control for

the cells transfected with inducible Nrf2.

siNrf2 – Nrf2 silencing was generated via transient transfection of Small interfering RNA

targeting Nrf2 (siNrf2) NFE2L2HSS181505 (Invitrogen). Scramble RNA (Invitrogen)

was used as a negative control.

3.7 Luciferase assay constructs

pRL – The internal control used was pRL Vector, which is wildtype Renilla luciferase

(Rluc) control reporter vectors that is used for the purpose of normalizing the luciferase

values.

PCR cloning was used to amplify the target regions and clone into PGL3 basic vector. In

short, the constructs were digested with the restriction enzymes Kpn1 and Xho1 with the

exception of Mtor, which was digested by SacI and MluI.

For site directed mutagenesis the TGA portion of the ARE’s analyzed were deleted using

the Quickchange II XL Stie-directed mutagenesis Kit. The primers sequence for Nqo1,

mTOR, TSC2, and S6K mutations are listed at appendix (Table 1).

Molecular Cloning of Nqo1 Promoter – The ARE site at 550bp upstream of start of

transcription is shown to be active on Nqo1(90). This region was cloned on the

PGL3_basic vector used as positive control (Nqo1_PGL3) Appendix (figure 1B). Nqo1_Pgl3

with the deleted TGA sequence (Nqo1_Pgl3 mut) was used as negative control.

Molecular Cloning of mTOR Promoter –The screened mTOR promoter region contained

eight ARE binding sites. I studied the closest ARE site present at 723bp upstream of the

TSS. The promoter region of mTOR, 1231 bp upstream from TSS, was cloned into the

Pgl3 basic vector (mTOR_Pgl3) Appendix (figure 1C) and used in subsequent functional

analyses. For site-directed deletion analyses the mTOR_Pgl3 mut was created.

Molecular Cloning of TSC2 Promoter –The screened TSC2 promoter region contained 6

ARE binding sites. I studied the closest ARE site present at 756bp upstream of the TSS.

The promoter region of TSC2, 1079 bp upstream from TSS, was cloned into the pgl3

basic vector (TSC2 _Pgl3) Appendix (figure 1D) and used in subsequent functional analyses.

For site-directed deletion analyses the TSC2_Pgl3 mut was created.

Molecular Cloning of S6K Promoter –The screened S6K promoter region contained 12

ARE binding sites. I studied the firsts 5 closest ARE sites, present at 255bp, 285bp,

324bp, 432bp and 2543bp upstream of the TSS. The promoter region of S6K, 2660bp

upstream from TSS, was cloned into the pgl3 basic vector (S6K_Pgl3) Appendix (figure 1E) and

used in subsequent functional analyses. For site-directed deletion analyses the two

closest ARE’s to TSS were mutated at the TGA (S6K_Pgl3 mut).

Molecular Cloning of AKT Promoter –The screened AKT promoter region contained 3

ARE binding sites. I studied those Are’s were present at 1191 bp, 1403 bp and 1681 bp

upstream of the TSS. The promoter region of AKT, 2200 bp upstream from TSS, was

cloned into the pgl3 basic vector (AKT _Pgl3) Appendix (figure 1F) and used in subsequent

functional analyses.

3.8 Luciferase Assay

Cells were lysed with Passive Lysis Buffer, and kept at -80ºC overnight. Luciferase

activities were analyzed in 20-µl cell extracts with the dual luciferase assay kit

(Promega).

Firefly and Renilla luciferase activities were then determined in triplicates for each

sample on the EnSpire multimode plate reader (PerKinElmer). The luciferase activities

reported were expressed as a ratio of the pGL3 reporter activity to that of the control

plasmid pRL. -Fold induction (Relative Luciferase activity) is expressed as the ratio of

induction from treated cells(PC_Nrf2 and siNrf2) versus untreated (pcDNA 4.0 and

Scramble RNA) respectively. Values represent the mean +- S.E. of three independent

measurements. Statistical analysis (Student’s t test) was performed by comparing treated

and non-treated cells (*, p < 0.05).

3.9 Electrophoretic Mobility Shift Assay (EMSA)

A549 cells (4 x106), were plated in four 175cm2 flasks with RPMI for 24 hours. The cells,

were transfected with Nrf2 siRNA or scrambled SiRNA and were harvested 24 hours

later. Nuclear extracts of A549 cells were prepared using 1M tris ph 7.5, 100mM Mgcl2,

3M Kcl, 500mM EDTA, 1M sucrose, 100% Glycerol, 1MDTT, 1M orthvanadate, 0.5M

BGlyc-phos, 100mM PMSF and 100x protease cocktail. The annealed primers for Nqo1

wild type, Nqo1 mutant, mTOR wild type, mTOR mutant 1, mTOR mutant 2, and mTOR

mutant 3 composed the probes used for the experiment appendix (table1). The primers were

annealed by heating at 95°C for 10 minutes followed by overnight incubation at 4 °C.

The probes were then labeled with the radioactive isotope g-[32P]ATP at 30°C for 30

minutes following 10 minutes incubation at 65C. For DNA-protein binding reactions, 10

µg of nuclear extract was incubated at room temperature for 30 min with 20 mM HEPES-

KOH (pH 7.9), 60 mM KCL, 1 mM MgCL2, 1 mM EDTA, 1 µg poly(dI-dC)

dithiothreitol, 10% glycerol, 0.2 mM ZnSo4 and 10,000 cpm g-[32P]ATP-labeled probe.

Protein-DNA complexes were resolved through a 4% polyacrylamide gel. The gel was

then dried and subjected to autoradiography with an intensifying screen at -80°C.

3.10 DNA Pull-Down Assay

Tissue culture, transient transfection and the nuclear extraction were performed for both

the DNA pull down as it was for the EMSA assay. This assay was performed via a

modified protocol described by Benoit Grondin et all 2006(91). The biotinylated primers

Nqo1 wild type, NQO1 mutant, Mtor wild type, Mtor mutant appendix (table1) were generated

at IDT (Integrated DNA Technologies). Annealing reaction of the primers was performed

as described for EMSA experiment. For DNA-protein binding reactions, 200 µg of

nuclear protein extraction was incubated over night at – 10C on a shaker with 10 µg of

biotinylated probes on 1 ml of biding/washing buffer (20 mM Tris [pH 8.0], 10%

glycerol, 6.25 mM MgCl2, 5 mM dithiothreitol, 0.1 mM EDTA, 0.01% NP-40) in a final

concentration of 200 mM NaCl. After 1 hour incubation with 50 µl of the magnetic beads

(Dynabeads® MyOne™ Streptavidin C1), immobilized templates were washed three

times with 0.5 ml of binding buffer, dried and resuspended on SDS and loading dye. The

samples were than boiled and resolved on a 10% SDS-PAGE gel for immunoblot

analysis with Nrf2 antibody (abcam).

3.11 Chromatin immunoprecipitation

This experiment was carried with as a modified protocol previously described by Donner

et al 2007(92), 2010(93). Briefly, A549 cells were grown until 80% confluence in 15 cm

plates and were transfected with 15µg of PC_Nrf2 using GenJet Plus transfection reagent

(SignaGen Laboratories). Before harvesting, the cells were cross-linked with 1%

formaldehyde for 10 mins at room temperature on a rocker. The cross-linking reaction

was quenched using 125mM glycine and washed twice with ice-cold phosphate-buffered

saline. The cells were harvested by scraping in RIPA buffer (150mM NaCl;1% v/v

Nonidet P-40;0.5% w/v deoxycholate; 0.1% w/v SDS;50mM Tris pH 8.0;5mM EDTA)

supplemented with protease inhibitor cocktail(Fisher), phosphatase inhibitors and PMSF.

These cells were sonicated on ice with 15 pulses of 15 seconds(20% amplitude) with

30second intervals to obtain an average chromatin length of 500 to 1,000 bps using a

Sonic Dismembrator (Fisher Scientific, Pittsburgh, Pa.) and centrifuged. The supernatant,

containing the chromatin, was collected and quantified alongside BSA standards and

equalized to a final concentration of 1mg/ml. The chromatin (1mg/ml) was pre-cleared

using 25μl of IgG magnetic beads (Dynabeads Invitrogen), previously washed with

RIPA, for 2 hrs at 4°C on a rocker. 10μl of pre-cleared chromatin was reserved as input

sample. The rest was immunoprecipitated with 25μl IgG magnetic beads, blocked with

salmon sperm DNA(0.3mg/ml) and BSA(1mg/ml), and with either anti-Nrf2 antibody

(Santa Cruz, Santa Cruz, Calif.), anti-RNA pol II antibody (Active Motif), or no antibody

overnight at 4°C with rotation. The next day, the beads were washed with RIPA and wash

buffer (100mM TrisHCl pH 8.8;500mM LiCl;1% v/v Nonidet P-40; 1% w/v deoxycholic

acid) and were resuspended in 100μl of 1X TE buffer. To elute the imunocomplexes,

200μl of elution buffer (70mM Tris HCl pH8.0;1mM EDTA;1.5% w/v SDS) was added

and the samples were incubated for 10min at 65°C with occasional vortexing. To reverse

cross-linked chromatin, 200mM NaCl is added to the eluted complexes and input samples

and incubated at 65°C for 6hrs. All the samples were then treated with 20 mg/ml

proteinase K (Fisher) and extracted with phenol-chloroform-isoamyl alcohol (25:24:1).

DNA was precipitated with ethanol and 3M sodium acetate and re-suspended in 100μl of

water. 2μl of purified DNA was used for qPCR appendix (table1).

4. RESULTS__________________________________________

Evidence from the literature shows that Nrf2 interacts with PI3K pathway at different

locations and regulates various functions of the cell(23, (37), (67)-80). The aim of this study

was to determine if Nrf2 transcriptionally controls the expression of the mTOR gene and

to illustrate whether this regulation is through direct or indirect binding of Nrf2 to the

mTOR promoter. To achieve these goals, western blot and qPCR analysis in conditions of

induced and silenced Nrf2 protein levels were performed. This was followed by

luciferase assays to confirm the presence of functionally active AREs in the mTOR

promoter. Lastly, we performed DNA pull down, EMSA and ChIP assays to confirm

direct binding of Nrf2 to elements in the mTOR promoter. The possibility of an Nrf2

impact on the other elements of the PI3K pathway (TSC2, S6K and AKT), was also

analyzed via western blot, qPCR and luciferase assay.

4.1 Nrf2 modulates mTOR expression in A549 cells

4.1.1 mTOR expression when Nrf2 is up-regulated

Expression analysis of mTOR was performed in A549, H460 and HEK293 cell lines.

Induction of Nrf2 was carried out by transiently transfecting Nrf2 cDNA (PC_Nrf2) appendix (figure 1A) for 24h. pcDNA 4.0 was used as a negative control for the cells .

The transiently transfected cell lines (figure 1) have significant increase in Nrf2 mRNA and

protein level, however the basal levels differ amongst the three cell lines. A549 cells have

the lowest basal Nrf2 protein levels such that the effect of transfection was most dramatic

in these cells. In A549 cells, mTOR expression was significantly increased, by

approximately five folds at both transcriptional and protein levels. In HEK 293 cells, an

increase in mTOR transcription was observed while protein levels showed no change. In

H460 cell lines there was 1.6 fold increase in mTOR protein, although thre was no

observable increase in transcriptional activity.

Figure 1. mTOR (Nrf2 inducible) expression analysis- A. mTOR protein levels were not increased in HEK293 cells. B and C. mTOR protein levels were increased five fold in A549 cells and 1.6 folds increased on H460 cells, respectively. D and E. mTOR transcription was increased two folds in HEK 293 cells and four folds in A549 cells. F. No increase in mTOR transcription was observed in H460 cells. G. The relative Luciferase activity of mTOR-WT in HEK293 cells was 20 folds increased and five folds increased in mTOR- mut. H. The A549 cells presented three folds increase of mTOR-WT relative luciferase activity with no change in mTOR-mut. I. The H460 cell lines did not present a significant change of relative Luciferase activity in both mTOR-WT and mTOR-mut. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of induction from treated cells (PC_Nrf2) versus Control (pcDNA 4.0). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05).

Nrf2 Inducible(mTOR)

Western blot

mTOR

β-Actin

mTOR

mTORβ-Actin

1 : 0.96

1 : 1.6

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 1.73

Nqo1 1 : 1.3

Control Pc_Nrf2

1 : 5.3

Nrf2 1 : 1.63

Control Pc_Nrf2

Nqo1 1 : 60

Nrf2 1: 1.5Control Pc_Nrf2

Nqo1 1 : 1.6

qPCR

D)

E)

F)

Control Nrf2 Nqo1 mTOR 0.00.51.01.52.02.5

mR

NA

expr

essi

onle

vels

*

Control Nrf2 Nqo1 mTOR 0.00.51.01.52.0

2345678

mR

NA

expr

essi

onle

vels

Control Nrf2 Nqo1 mTOR 0.00.51.01.52.0

2345678

mR

NA

exp

ress

ion

leve

ls

Luciferase

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.00.51.01.52.0

510152025

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.0

0.5

1.0

1.5

2.02345678

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.0

0.5

1.0

1.5

2.02345678

Rel

ativ

e Lu

cife

rase

Ac

tivity

G)*

*

**

H)*

**

*

*

**

*

**

*

*

I)

4.1.2 mTOR expression when Nrf2 is down-regulated

Figure 2. mTOR (Nrf2 silencing) expression analysis. A, B and C. mTOR protein levels were significantly transiently decreased in the three cell lines. D and E. mTOR transcription was decreased proximately 1.5 folds on HEK293 cells and 2 folds in A549 cells. F. No change was observed on mTOR transcription in H460 cell lines. G, H and I. No change in the luciferase activity was observed for Mtor-WT and mtor mut in all the three cell lines. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of silencing from treated cells (siNrf2) versus Control (Scramble RNA). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05)

Nrf2 Silencing(mTOR)

Western blot

mTOR

β-Actin

mTOR

mTORβ-Actin

1 : 0.03

1 : 0.65

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 0.03

Nqo1 1 : 0.07

Control Si_Nrf2

1 : 0.51

Nrf2 1 : 0.02

Control Si_Nrf2

Nqo1 1 : 0.53

Nrf2 1 : 0.77Control Si_Nrf2

Nqo1 1 : 0.43

qPCR

D)

E)

F)

Luciferase

G)

H)

I)

Control Nrf2 Nqo1 mTOR 0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 mTOR 0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 mTOR 0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

ls

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.0

0.5

1.0

1.5

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.0

0.5

1.0

1.5

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut mTOR - WT mTOR-mut0.0

0.5

1.0

1.5

Rel

ativ

e Lu

cife

rase

Ac

tivity

*

*

*

**

**

*

* *

*

*

In conditions where Nrf2 is silenced (figure 2), all the three cell lines presented a significant

decrease of Nrf2 at both transcriptional and protein levels, with the most significant

effects seen at protein levels in HEK 293 and A549 cells. Silencing Nrf2 transcription

resulted in a two-fold decrease in mTOR, both its transcription and protein levels. In

HEK293 cells, a small decrease in mTOR transcription was observed.

4.2 Functional ARE present on mTOR promoter activates its transcription in Nrf2

inducible conditions.

Luciferase assay was performed in order to verify if the regulation of mTOR gene

expression was due to the presence of a functional ARE binding site in the mTOR

promoter region. Biswal et al(37), performed ChIP-Seq experiment to explore the network

of Nrf2 regulated genes and in this work they used the consensus core ARE sequence

TGANNNNGC. Here the mTOR promoter region was screened for ARE sites that had

the same motif sequence. Biswal et al, also screened 5225 background sequences relative

to the closest gene transcription starting site (TSS) in order to identify ARE sites. They

identified the highest peaks at AREs closest to the genes’ TSS. Similarly, in another Nrf2

ChIP-seq study performed by Chorley BN et al (38), based on 39 currently known

functional human AREs, NRF2-binding sites were found to be cis-acting elements more

commonly located at an average distance of ~1800 bp from the gene TSS. For these

reasons, in this study, from the eight ARE’s found within 5000bp of mTOR promoter

region, the “TGACCAGGC” ARE, located closest to mTOR TSS (723 bp upstream from

TSS), was cloned into an expression vector. The PRL-mTOR vector contained 1231 bp

of the mTOR promoter was then used on Luciferase assay (mTOR WT) appendix (figure1C).

As shown by Biswal et al(37) via alignment of 20 known ARE binding sites and MEME

motif discovery algorithm on their Nrf2 ChIP-Seq dataset, the “TGA” portion of the ARE

is the most recurrent portion of the sequence. For this reason, in this study, site-directed

deletion was performed in the mTOR WT construct where the “TGA” of the ARE biding

site was deleted (mTOR Mut). Both mTOR WT and mTOR Mut constructs were

analyzed by luciferase activity assay at inducible and silencing conditions. Promoter of

the Nqo1 gene, a known target of Nrf2, was used as a positive control for this assay

b

(Nqo1 WT) appendix (figure 1B).

When transfected with the inducible PC_Nrf2 construct Nqo1 was substantially increased

at the protein and transcription level on all the cell lines (figure 1). In Nrf2 inducible

conditions, A549 cell line showed a 60 fold increase in the Nqo1 protein and a three fold

increase in the transcription of Nqo1 gene, compared to basal conditions. Whereas, in

Nrf2 silencing conditions (figure 2), Nqo1 expression was reduced in all the three cell lines.

Both transcription and protein levels of the control were decreased two fold in A549

cells.

The negative control consisted of the same Nqo1 promoter region with a mutated ARE

(Nqo1 Mut). At the basal level (Graph 1), the luciferase assay showed that the negative

control, when compared with Nqo1 WT activity, decreased five fold in A549 cells and

two fold in both of HEK293 and H460 cells. In this same condition, the activity of the

mTOR Mut was two folds lower than the mTOR WT in A549 and HEK293 cells while

no change was recorded on H460 cells.

Analysis of Nrf2 modulation was performed by comparing the fold change of the

luciferase activity of the Pgl3 constructs at basal Nrf2 levels (control) with cells

transfected with the same construct and Pc_Nrf2(figure 1) or Si_Nrf2 (figure 2). Induction or

silencing of Nrf2 was validated with Nqo1 WT activity following Nrf2 up and down

patterns of expression in the three cell lines, with three folds increase and 7 folds

decrease on A549 cells. The negative control was not affected by Nrf2 variations in the

cells. The one exception was HEK293 cells in Nrf2 inducible condition, where there was

a four folds increase. Nevertheless, Nqo1 Mut activity was 6 fold lower than Nqo1 WT

in these conditions in HEK293 cells, so the Nrf2 is playing a regulatory role through its

interaction with ARE. When transfecting the cells with the inducible construct (figure 1) it

was observed that the luciferase activity of the mTOR wild type (mTOR WT) construct

was increase 20 folds in HEK293 and four folds on A549 cells, but there is no change on

H460 cells. mTOR Mut activity remained unchanged during Nrf2 up regulation in A549

but not in HEK293 cells. In silencing conditions (figure 2) no change in activity for the wild

type and mutant mTOR constructs were observed in any of the cell lines. From the cell

lines analyzed, A549 cells presented a clearer correlation between Nrf2 levels and mTOR

expression. For this reason, additional analyses of the Nrf2/mTOR interaction were

performed in this cell line.

Graph1. Nqo1 and mTOR (Nrf2 basal levels) Luciferase activity. A. Nqo1 Mut presented a 2 fold decrease in HEK293 and H460 cells and 4 fold decrease in A549 cells. B. mTOR Mut presented 2 fold decrease in HEK293 and A549 cell and no change on HEK293 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity was represented as the fold change of the ratio from cells transfected with mutant constructs (Nqo1-Mut and mTOR-Mut) versus cells transfected with wild type constructs (Nqo1-WT and mTOR-WT). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of mutant and wild type constructs expression (*, p < 0.05).

4.3 Nrf2 binds to mTOR promoter region at basal conditions in vitro

Nrf2 binding to the mTOR promoter was demonstrated in vitro using DNA pull-down

and EMSA experiments. In the DNA pull down assay the mTOR promoter region was

used as a probe to selectively obtain a protein-DNA complex from an A549 nuclear

extract. The high affinity tag, biotin, was present in both extremities of the probe and the

complex purification was performed with streptavidin magnetic beads. The proteins were

eluted from DNA and detected via western blot (figure 3). Assessment of the biding capacity

of the ARE sequence present in this promoter region was performed via a mTOR probe

with a scrambled ARE site appendix (table 1). Nqo1 promoter region was used as a positive

control, and scrambled ARE site was used as a negative control

Nqo1-W

T

Nqo1-M

ut

Nqo1-W

T

Nqo1-M

ut

Nqo1-W

T

Nqo1-M

ut0.0

0.5

1.0

1.5

2.0HEK A549 H460

**

*

Rel

ativ

e Lu

cife

rase

Ac

tivity

A)

mTOR-WT

mTOR-Mut

mTOR-WT

mTOR-Mut

mTOR-WT

mTOR-Mut

0.0

0.5

1.0

1.5

2.0HEK A549 H460

* *

Rel

ativ

e Lu

cife

rase

Ac

tivity

B)

Figure 3. Western blot from DNA pull-down samples using Nrf2 antibody - Blot analysis of (Input) nuclear extract from A549 cells, (No Probes) negative control comprising of reaction mix alone incubated with magnetic beads and probed samples. The probed samples consisted of (Nqo1 wt) Nqo1 promoter region containing functional ARE which was used as a positive control, scramble ARE from Nqo1 promoter region which was used as a negative control (Nqo1mutant), mTOR promoter region containing ARE (mTOR WT) and scramble ARE from mTOR promoter region which was used as a negative control (mTor mutant). It was observed an 2 folds decrease of Nrf2 protein pulled down with mTOR mutant probe when compered with the amount of protein pulled down with mTOR WT probe, as it was for the controls, Nqo1 mutant and Nqo1 WT.

On western blot analysis, our results suggest that Nrf2 binds to an element(s) in mTOR

promoter region. The fact that the amount of Nrf2 protein pulled down with the WT

mTOR probe was 2 folds higher than the amount pulled down with mutant mTOR probe

and with negative control (no probe), adds our speculation that the ARE is the biding site

of the Nrf2.

EMSA was carried out in order to further verify the Nrf2 biding is at the mTOR’s ARE

located 1231 bp upstream from the TSS (mTOR wild type). For this experiment a

mutation was done by removing the entire ARE sequence TGACCAGGC and adding 5

bp in both 5’ and 3’ prime extremities (mTOR Mut) (figure 4). The mTOR wt, mTOR

mutants as well as the positive(Nqo1 wild type) and the negative control (Nqo1 mutant)

were end labeled with [32P] ATP and incubated with nuclear extract isolated from A549

cells.

It was observed that the predicted Nrf2 site was present in the sample incubated with

mTOR wild type and not in mTOR mutant. It was also observed that an additional biding

was present in the mutated mTOR probe at an adjacent site (figure 5).

Figure 4. mTOR probes used on EMSA assay. mTOR WT sequence containing the “ TGACCAGGC” ARE and mTOR Mut with deleted ARE sequence and 5 bp extension at 5’and 3’ ends. The primers were annealed, with it respective reverse complementary sequence, end labeled with [32P] ATP and used on EMSA experiments.

5’-TTCACCATGTTGACCAGGCTGGTCTCGAC-3’

5’-GGGAATTTCACCATGT********* TGGTCTCGACTCCTC-3’

Figure 5. ARE dependent biding of nuclear components to mTOR-WT- EMSA was performed using labeled promoter fragment of Nqo1-WT (positive control), Nqo1-Mut (negative control), mTOR –WT (mTOR promoter region containg ARE site) and mTOR –Mut (mTOR promoter region containg deleted ARE site plus addiction of 5bp on 5’ and 3’ ends) incubated with nuclear extracts (10 µg per lane ) from A549 cells. Top red arrow indicate shift of predicted Nrf2 biding site and bottom black arrows indicate new and unknown biding appeared on cells incubated with labeled mTOR –Mut probes. Predicted Nrf2 biding site (red arrow) was presented on samples incubated with mTOR-WT and Nqo1-WT and not on samples incubated with Nqo1-Mut and mTOR –Mut

4.3.1 Nrf2 binding to mTOR promoter region decreases in Nrf2 silencing conditions

EMSA assay was also performed in A549 cells in which Nrf2 was silencing (figure 6). After

incubation nuclear extract of the Nrf2 down regulated A549 cell with radioactive labeled

mTOR WT probe a significant decrease in bound protein was observed. Intensity of the

blots present on samples incubated with mTOR WT probes suggests that in basal

conditions the Nrf2-mTOR biding is weak.

Figure 6. Biding of nuclear components to mTOR-WT at Nrf2 silencing conditions. A. EMSA was done on nuclear extract (NE) of transiently transfect A549 cell with SiNrf2 or scrambled RNA (control). SiNrf2 A549cells NE and Scramble A549cells NE were incubated with labeled promoter fragment of Nqo1-WT (positive control), Nqo1-mut (negative control ). The films containing shift of predicted Nrf2 biding site (red arrow) were developed after over nigh or four days gel incubation at -80oC. Once incubated over nigh the SiNrf2 A549cells NE samples that contained Nqo1-WT probes presented decreased blot intensity when compared with Scramble A549 cells NE. After four days incubation, the SiNrf2 A549cells NE samples that contained mTOR-WT probes presented decreased blot

4.3.2 Nrf2 binds to mTOR promoter in vivo at inducible conditions

In order to clarify the in vitro results of the Nrf2/mTOR binding, this interaction was

analyzed in vivo. One of the factors that can influence the assays performed in vitro

assays is the lack of the natural DNA conformational topology on those assays(94). In

order for genomic regulation and recombination to occur, these processes require DNA

bending, twisting, and looping as well as wrapping around histone octamers in order to

occur. Thus, in vitro assays, such as the ones performed in this study, may not give the

precise representation of the actual intracellular processes. Also, in addition to DNA

structure, molecular crowding caused by the presence of particles on the cytoplasmic

microenviroment may influence local and distal interactions(95). Biochemical reactions in

vivo occur at crowding conditions with high concentrations of biomacromolecules. While

the majority of the biochemical reactions in vitro are performed in solutions containing

low concentrations of biomacromolecules.

ChIP assay followed by qPCR amplification enables the capture of protein–DNA

interactions in vivo and is considered a definitive confirmatory method when analyzing

Nrf2 transcriptional targets(96). This assay was used in the past to identify important Nrf2

targets such as antiapoptotic protein Bcl-2, catalytic subunit of glutamylcysteine ligase

(GCLC) and Aldose reductase (AR)(97-99) among others. Nrf2/mTOR biding in vivo

Chromatin ImmunoPrecipitation (ChIP) coupled to detection by quantitative real-time

PCR was performed on A549 cells (Graph 2). The samples were immunoprecipitated with

either anti-Nrf2 antibody, anti-RNA pol II antibody or no antibody. The experiment

compared the fold enrichment, with respect to no antibody control, of crosslinked

protein-DNA complexes in two Nrf2 conditions, basal and inducible. At the basal levels,

the ChIP performed using anti-Nrf2 antibody, showed a 2.5 fold enrichment compered to

no antibody control, of the mTOR promoter, which denotes a weak binding at basal

levels. Whereas in Nrf2 inducible conditions, the enrichment of the same mTOR

promoter was seen to increase to 13 folds. Anti-RNA polII antibody was used as a

positive control antibody to confirm successfulness of the ChIP assay. Nqo1 promoter

region was used as a positive control for the anti-nrf2 antibody, while GAPDH served as

a positive control for anti RNA Pol II antibody and as a negative control for anti Nrf2

antibody.

Graph2. ChIP assay. Crosslinked protein-DNA complexes were immunoprecipitated using either anti-RNA polymerase II antibody (Pol II, positive control), anti-Nrf2 antibody or no antibody in A549 cells transfected with inducible or basal (empty vector) constructs (Nrf2 cDNA containing plasmid). Enrichment was measured as fold increase of antibody vs the no antibody control by q-PCR.

4.4 Expression analyses of the other elements of PI3K pathway due to Nrf2

modulation

Expression of other components of the PI3K pathway elements including TSC2, S6K and

AKT as well as luciferase assay on promoters of these genes, in which ARE core

sequence were identified, were also analyzed in conditions of Nrf2 modulation. The Nrf2

inducible and silencing conditions as well as the control were the same as the ones

ChIP Assay A549 cells

Nrf2-ba

sal

Nrf2-in

ducib

le

Pol II-b

asal

Pol II-i

nduc

ible

0102030405060708090

100110120130140150200400600800

100012001400

MtorNqo1Gapdh

Antibody used for ChIP

Fold

Enr

ichm

ent t

o N

o ab

performed for mTOR. The presence of Nrf2 affected the expression of the targeted

proteins in a very heterogenous fashion across the three cell lines. Also, for some of the

above-mentioned genes, protein expression and transcriptional activity did not followed

the same pattern in all the three cell lines.

Luciferase assay was performed on the promoter regions containing the ARE sites. As

was the case for mTOR, 5000bps upstream from the TSS of each of the respective genes

were screened for the presence of AREs. TSC2 promoter region contained 6 ARE

binding sites. Luciferase assay was performed on the closest ARE present at 756bp

upstream of the TSS (TSC2 WT)(figure1Dappendix). S6K promoter region contained 12

ARE binding sites. The firsts 5 closest AREs, present at 255bp, 285bp, 324bp, 432bp and

2543bp upstream of the TSS were used in this assay (S6K WT) (figure1Eappendix). AKT

promoter region contained 3 ARE’s present at 1191 bp, 1403 bp and 1681 bp upstream of

the TSS witch were cloned and also used for this assay (AKT WT) (figure1Fappendix). For

site-directed deletion analyses the TGA site of the TSC2 ARE was mutated (TSC2 Mut),

and on S6K the two closest ARE’s to TSS were mutated as well ( S6K Mut )(table

1appendix). The activity of the abovementioned AREs showed great variation amongst the

three cell lines and in many cases did not followed the same pattern of the transcription

levels observed via qPCR.

4.4.1 TSC2, S6K and AKT expression when Nrf2 is up-regulated

4.4.1.1 TSC2 is a potential indirect Nrf2 transcriptional target at inducible

conditions on H460 cells

When upregulating Nrf2 (figure 7), TSC2 protein expression was induced only in H460 cells

while transcription was increase in all the three cell lines. The ARE present on TSC2

promoter region (graph 3) showed, in basal conditions, a small decrease in activity for TSC2

mut (A549 and H460). When Nrf2 is induced (figure 7), this ARE driven construct had

increased activity for when the ARE was WT (TSC2 WT) in A549 cells and HEK cells

and also in TSC2 mut in A549 cells. This could indicate that TSC2 is potentially an

indirect Nrf2 transcriptional target of increased Nrf2 as opposed to at basal conditions. In

H460 cells where TSC2 protein levels and transcription were increased. Although TSC2

transcription levels where increased by 10 fold in A549 cells no change was observed at

the protein level, perhaps suggesting a post-translational level of regulation of TSC2 in

these cells.

As observed for TSC2, when Nrf2 is increased (figure 8), S6K transcription is increased in

A549 and H460 cells. However, although, at basal Nrf2 levels (graph 4), luciferase activity

of S6K-mut was decreased 2 fold in the two cell lines, at Nrf2 inducible conditions, both

luciferase activity of S6K-WT and S6K-mut were increased in the A549 cells. This

implies that, while the ARE present on S6K promoter region is important for

transcription at Nrf2 basal levels, it is probable not induced by increased Nrf2 levels.

4.4.1.2 At Nrf2 inducible conditions AKT is a possible indirect Nrf2 transcriptional

target on H460 cells and posttranslational target on A549 cells

In H460 cell lines, at Nrf2 inducible conditions (figure 9), AKT transcription was increased

2 fold and proteins levels by over 5 fold. Since, no change was observed on the AKT

luciferase activity in this cell line, the results suggest that Nrf2 regulates this gene

indirectly, probably at the protein level. The increase of AKT luciferase activity on

HEK293 and A549 cells were also deceptive, since no significant changes were observed

at the transcription and protein levels in HEK293 cells and at the transcription level in

A549 cells. At protein level however, AKT was proximately 2 folds decreased in A549

cells. Hence, high Nrf2 levels affect some post-translational regulation of AKT protein

expression in A549 cells.

Figure 7. TSC2 (Nrf2 inducible) expression analysis. A and B. No significant change on Tsc2 protein levels were observed in HEK293 cells and A549 cells. C. Tsc2 protein levels were 6.78 fold increased in H460 cells. D, E and F. Tsc2 transcription was increased proximately two fold in HEK293 cells, 10 fold on A549 cells and two fold in H460 cells. G. The relative Luciferase activity of TSC2-WT in HEK293 cells was 1.5 fold increased with no change in activity on TSC2- mut. H. The A549 cell lines shown proximately two fold increase of TSC2-WT relative Luciferase activity with 1.5 folds increase on TSC2-mut. I. No significant change was observed in H460 cells for the relative Luciferase activities of TSC2-WT and TSC2-mut. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of induction from treated cells (PC_Nrf2) versus Control (pcDNA 4.0). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05).

Nrf2 Inducible(TSC2)

Western blot

TSC2

β-Actin

TSC2

TSC2β-Actin

1 : 1.16

1 : 6.78

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 1.73

Nqo1 1 : 1.3

Control Pc_Nrf2

1 : 0.91

Nrf2 1 : 1.63

Control Pc_Nrf2

Nqo1 1 : 60

Nrf2 1: 1.5Control Pc_Nrf2

Nqo1 1 : 1.6

qPCR

D)

E)

F)

Luciferase

G)

*

*

*

H)

*

I)

Control Nrf2 Nqo1 TSC20.00.51.01.52.02.53.03.54.0

mR

NA

expr

essi

onle

vels

Control Nrf2 Nqo1 TSC20.00.51.01.52.02.53.03.54.0

69

1215

mR

NA

expr

essi

onle

vels

Control Nrf2 Nqo1 TSC20.00.51.01.52.02.53.03.54.0

69

1215

mR

NA

exp

ress

ion

leve

ls

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.00.51.01.52.02.53.03.54.0

6

9

12

15

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.00.51.01.52.02.53.03.54.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

*

*

**

***

*

**

* *

Graph 3. TSC2 (Nrf2 basal) Luciferase activity. TSC2 Mut presented a small decrease on A549 and H460 cells and no change on HEK cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity was represented as the fold change of the ratio from cells transfected with mutant construct (TSC2-Mut) versus cells transfected with wild type constructs (TSC2-WT). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of mutant and wild type constructs expression (*, p < 0.05).

TSC2-WT

TSC2-Mut

TSC2-WT

TSC2-Mut

TSC2-WT

TSC2-Mut

0.0

0.5

1.0

1.5

2.0HEK A549 H460

* *

Rel

ativ

e Lu

cife

rase

Ac

tivity

Figure 8. S6K (Nrf2 inducible) expression analysis. A, B and C. No significant change was observed on S6K protein levels on the three cell lines D, E and F. S6K transcription did not changed in HEK293 cells and it was 2 fold increased in A549 and H460 cells. G. The relative Luciferase activity of S6K –WT and S6K-mut were 1.5 fold increased in HEK293 cells H.The relative Luciferase activity of S6K –WT and S6K-mut were 2 fold increased in A549 cells I. No change was observed on the relative Luciferase activity of S6K –WT and S6K-mut in H460 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of induction from treated cells (PC_Nrf2) versus Control (pcDNA 4.0). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05).

Nrf2 Inducible(S6K)

Western blot

S6Kβ-Actin

S6K

S6Kβ-Actin

1 : 0.95

1 : 0.85

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 1.73

Nqo1 1 : 1.3

Control Pc_Nrf2

1 : 1.32

Nrf2 1 : 1.63

Control Pc_Nrf2

Nqo1 1 : 60

Nrf2 1: 1.5Control Pc_Nrf2

Nqo1 1 : 1.6

qPCR

D)

E)

F)

Luciferase

G)*

*

*

H)

*

*

I)

Control Nrf2 Nqo1 S6K0.00.51.01.52.02.53.0

mR

NA

expr

essi

onle

vels

Control Nrf2 Nqo1 S6K0.00.51.01.52.02.53.0

369

1215

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 S6K0.00.51.01.52.02.53.0

369

1215

mR

NA

expr

essi

onle

vels

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.00.51.01.52.02.53.0

3

6

9

12

15

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.0

0.5

1.0

1.5

2.0

2.5

3.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.0

0.5

1.0

1.5

2.0

2.5

3.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

*

*

*

*

*

*

*

* *

Graph 4. S6K (Nrf2 basal ) Luciferase activity. S6K Mut presented 2 fold decrease in A549 and H460 cells and no change in HEK293 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity was represented as the fold change of the ratio from cells transfected with mutant construct (S6K-Mut) versus cells transfected with wild type constructs (S6K-WT). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of mutant and wild type constructs expression (*, p < 0.05).

S6K-W

T

S6K-M

ut

S6K-W

T

S6K-M

ut

S6K-W

T

S6K-M

ut0.0

0.5

1.0

1.5

2.0HEK A549 H460

* *R

elat

ive

Luci

fera

se

Activ

ity

Figure 9. AKT (Nrf2 inducible) expression analysis. A, B and C. AKT protein levels were 5.15 fold increased in H460 cells, proximately 2 fold decreased in A549 cells and no significant change was observed in HEK293 cells. D, E and F. No significant change in AKT transcription was observed in HEK293 and A549 cells and it was two folds increased in H460 cell lines. G. Luciferase activity of AKT-WT was 1.5 fold increased in HEK293 cells, proximately 2 fold increased in A549 cells and no change was observed in H460 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of induction from treated cells (PC_Nrf2) versus Control (pcDNA 4.0). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05).

Nrf2 Inducible(AKT)

Western blot

AKTβ-Actin

AKT

AKTβ-Actin

1 : 0.85

1 : 5.15

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 1.73

Nqo1 1 : 1.3

Control Pc_Nrf2

1 : 0.56

Nrf2 1 : 1.63

Control Pc_Nrf2

Nqo1 1 : 60

Nrf2 1: 1.5Control Pc_Nrf2

Nqo1 1 : 1.6

qPCR

D)

E)

F)

Luciferase

G)

*

*

*

H)*

I)

Control Nrf2 Nqo1 AKT0.00.51.01.52.02.53.0

mR

NA

expr

essi

onle

vels

Control Nrf2 Nqo1 AKT0.00.51.01.52.02.53.0

369

1215

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 AKT0.00.51.01.52.02.53.0

369

1215

mR

NA

exp

ress

ion

leve

ls

Control Nqo1 -WT Nqo1-mut AKT-WT0.00.51.01.52.02.53.0

369

1215

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut AKT-WT0.00.51.01.52.02.53.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut AKT-WT0.0

0.5

1.0

1.5

2.0

2.5

3.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

*

*

***

*

*

*

*

4.4.2 TSC2, S6K and AKT expression when silencing Nrf2

4.4.2.1 TSC2, S6K and AKT may be affected post translationally, when Nrf2 is

silenced.

Decreasing Nrf2 has no significant effect on the observed on TSC2 (figure 10) and S6K (figure 11) transcription and luciferase activity. However, Tsc2 protein levels where

decreased 2.64 folds in HEK293 cells and S6K was 5.84 folds increased on A549 cells.

This suggests that at low cellular Nrf2 levels, TSC2 (HEK293 cells) and S6K (A549

cells) protein levels are in some way affected.

When silencing Nrf2 in A549 cells (figure 12), luciferase activity of AKT WT decreased

four folds alongside with two folds decrease in AKT transcription. These findings imply

that AKT could be a direct Nrf2 transcriptional target. However, the small increase in

AKT protein levels suggests that those changes in transcription and luciferase activity

may not be biological relevant. In both H460 and HEK293 cell, the changes in AKT

transcription was also probably misleading since they did not followed the same pattern

observed in the AKT Western blot. However, because AKT protein levels were

proximately two fold decreased in HEK293 cells, we believe that AKT may be a potential

Nrf2 post-translational target.

Figure 10. TSC2 (Nrf2 silencing) expression analysis. A, B and C. TSC2 protein levels were 2.64 fold decreased in HEK293 cells with no significant change observed in A549 and H460 cells. D-I.In all three cell line, no significant change was observed on TSC2 transcription and Luciferase activity of TSC2-WT and TSC2-mut. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of silencing from treated cells (siNrf2) versus Control (Scramble RNA). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05)

Nrf2 Silencing(TSC2)

Western blot

TSC2

β-Actin

TSC2

TSC2β-Actin

1 : 0.34

1 : 0.78

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 0.03

Nqo1 1 : 0.07

Control Si_Nrf2

1 : 1.01

Nrf2 1 : 0.02

Control Si_Nrf2

Nqo1 1 : 0.53

Nrf2 1 : 0.77Control Si_Nrf2

Nqo1 1 : 0.43

qPCR

D)

E)

F)

Luciferase

G)

H)

I)

*

*

Control Nrf2 Nqo1 TSC20.0

0.5

1.0

1.5

2.0

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 TSC20.0

0.5

1.0

1.5

2.0m

RN

A e

xpre

ssio

nle

vels

Control Nrf2 Nqo1 TSC20.0

0.5

1.0

1.5

2.0

mR

NA

exp

ress

ion

leve

ls

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut TSC2-WT TSC2-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

*

*

*

*

*

*

*

Figure 11. S6K (Nrf2 silencing) expression analysis. A, B and C. S6K protein levels were 5.84 fold increased in A549 and no significant change was observed in HEK293 and H460 cells D, E and F. no significant change was detected in S6K transcription on the three cell lines G, H and I. Relative Luciferase activity of both wild type and mutant S6K constructs were 1.7 fold increased in HEK293 and H460 cells, no significant change was observed on A549 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of silencing from treated cells (siNrf2) versus Control (Scramble RNA). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05)

Nrf2 Silencing(S6K)

Western blot

S6K

β-Actin

S6K

S6Kβ-Actin

1 : 1.09

1 : 1.35

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 0.03

Nqo1 1 : 0.07

Control Si_Nrf2

1 : 5.84

Nrf2 1 : 0.02

Control Si_Nrf2

Nqo1 1 : 0.53

Nrf2 1 : 0.77Control Si_Nrf2

Nqo1 1 : 0.43

qPCR

D)

E)

F)

Luciferase

G)

H)

I)

*

*

*

Control Nrf2 Nqo1 S6K0.0

0.5

1.0

1.5

2.0

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 S6K0.0

0.5

1.0

1.5

2.0m

RN

A e

xpre

ssio

nle

vels

Control Nrf2 Nqo1 S6K0.0

0.5

1.0

1.5

2.0

mR

NA

exp

ress

ion

leve

ls

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut S6K-WT S6K-mut0.0

0.5

1.0

1.5

2.0

Rel

ativ

e Lu

cife

rase

Ac

tivity

**

**

**

**

**

Figure 12. AKT (Nrf2 silencing) expression analysis. A, B and C. AKT protein levels were proximately two fold decreased in HEK293 cells and no significant change was observed in A549 and H460 cells. D, E and F. AKT transcription was proximately 1.5 fold increased in HEK293 cells, 2 fold decreased in A549 cells and proximately 2 fold decreased in H460 cells. G, H and I. The Relative Luciferase activity for AKT-WT was four fold decreased on A549 cells and no significant change was observed in HEK293 and H460 cells. The Relative luciferase activity was expressed as a ratio of the pGL3 reporter activity to that of the control plasmid pRL. Relative Luciferase activity and mRNA expression levels were represented as the fold change of the ratio of silencing from treated cells (siNrf2) versus Control (Scramble RNA). Values represent the mean +- S.E. of three independent measurements. Statistical analysis (Student’s t test) was performed by comparison of treated and Control cells (*, p < 0.05)

Nrf2 Silencing(AKT)

Western blot

AKTβ-Actin

AKT

AKTβ-Actin

1 : 0.54

1 : 1.26

β-Actin

HEK293

A549

H460

A)

B)

C)

Nrf2 1 : 0.03

Nqo1 1 : 0.07

Control Si_Nrf2

1 : 1.38

Nrf2 1 : 0.02

Control Si_Nrf2

Nqo1 1 : 0.53

Nrf2 1 : 0.77Control Si_Nrf2

Nqo1 1 : 0.43

qPCR

D)

E)

F)

Luciferase

G)

H)

I)

**

*

Control Nrf2 Nqo1 AKT0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 AKT0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

ls

Control Nrf2 Nqo1 AKT0.0

0.5

1.0

1.5

mR

NA

exp

ress

ion

leve

lsControl Nqo1 -WT Nqo1-mut AKT-WT

0.0

0.5

1.0

1.5

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut AKT-WT0.0

0.5

1.0

1.5

Rel

ativ

e Lu

cife

rase

Ac

tivity

Control Nqo1 -WT Nqo1-mut AKT-WT0.0

0.5

1.0

1.5R

elat

ive

Luci

fera

se

Activ

ity

*

***

**

*

*

**

Nrf2 has been shown to have impact of sensitivity to some cancer cytotoxic drugs, by

virtue of its regulation of a broad cyto-protective gene battery that includes cellular

defense against a variety of chemotoxins reactive oxygen species. A variety of studies

have shown that Nrf2 interacts with a wide variety of cellular proteins in different

pathways, among which is the PI3K pathway. The PI3K pathway is involved in various

mechanisms important for tumor development and growth, such as cell survival,

differentiation, and metabolism. Nrf2 is expressed predominantly in metabolic organs(40)

and there is evidence that Nrf2 enhances the PI3K pathway in systems with a high

metabolic state(79). mTOR is a crucial metabolic regulatory component of the PI3K

pathway, controlling cellular anabolic and catabolic processes coupling growth signals to

nutrient availability, via ribosome biogenesis and autophagy. The biological processes

determined by proteins in this cell-signaling pathway are commonly deregulated in

human cancers. Due to the importance of mTOR in cancer, and the previously described

interactions between Nrf2, PI3K pathway proteins including mTOR, the present studies

aimed to clarify a precise role of Nrf2 in the regulation of the PI3K pathway, including

mTOR.

The most important finding of this study is that modulation of Nrf2 levels regulates the

levels of mTOR at the protein level; further analysis confirmed that Nrf2 regulates the

transcription of mTOR on A549 cells. While the Luciferase assay and DNA pull down

results suggest that this regulation is a direct interaction of Nrf2 with elements in the

mTOR promoter, EMSA and ChIP assay did not allow us to definitely confirm that, as

the binding of Nrf2 to mTOR promoter in all the conditions studied is weak. There is still

much to explore about this interaction since it mechanism of action was not yet

established. An interesting aspect of the Nrf2/mTOR interaction, as well as the Nrf2

interaction with the other elements of the PI3K pathway, is that they may be cell line

specific, as differences were observed between the 3 cell lines used in our study. In A549

cells, mTOR transcription and protein translation correlated with Nrf2 levels. In HEK

cells, while Nrf2 upregulation did not lead to an increase in mTOR pretein levels, despite

the fact that an increase in mTOR transcription was observed. However, Nrf2 silencing

gave a similar expression profile as in A549 cells, which is two-fold decrease of mTOR

transcription along with decrease of its protein translation. In this condition, it was

observed a two-fold decrease of mTOR protein in A549 cells and total silencing of

mTOR in HEK293 cells. In H460 cell lines, which have an activating mutation PI3K and

therefore an activated pathway, as well as higher (than in A549) basal Nrf2 levels, we

observed no change at either mRNA or at Mtor protein levels when Nrf2 expression was

modulated.

These expression and promoter region analysis have shown that A549 cells presented a

more distinct correlation between Nrf2 levels and mTOR expression compared to the

other analyzed cell lines. Therefore, A549 cells were utilized in the subsequent studies of

determining the nature of the regulation of mTOR by Nrf2, using DNA pull down,

EMSA and ChIP assay.

The different results in the three cell lines reflects important heterogeneity amongst them.

HEK293 is a non-transformed Human Embryonic Kidney cell line. The absence of

mutations known to alter Nrf2 expression, can probably explain the low Nrf2

level and the different results when compared with the NSCLC cell lines. In

a study performed by Zhu L et al (100) using BEAS-2B cell line, which is a normal human

bronchial epithelium, Nrf2 up-regulation did not increase mTOR protein levels. On this

basis, it was affirmed that mTOR is not a target of Nrf2 activation. These results,

combined with our observations in HEK cells, suggest that Nrf2 does not cause a direct

increase in mTOR expression in non-cancerous cell lines.

Both A549 and H460 are cancer cell lines derived from NSCLC. The most common

NSCLC histologies include; epidermoid or squamous cell carcinoma, adenocarcinoma

and large cell carcinoma. The cell lines A549 and H460 are derived from

adenocarcinoma and large cell carcinoma respectively. Although diagnosis, staging,

prognosis, and treatment are similar for these different types of NSCLC, distinct set of

mutations, present, even within the same NSCLC histologic group, provide specific

molecular profile for each cell line(101). Both A549 and H460 cell lines have K-RAS and

Keap1 mutations. The K-RAS mutations are present at codon V12 on A549 cells and at

the codon V61 on H460 cells and Keap1 mutations are D236H and G333C in H460 and

A549, respectively (Singh et al 2006). K-RAS is known to generate an oncogene-directed

increased expression of Nrf2(102), while Keap1 when mutated liberates Nrf2 from

proteasomal ubiquitination(7,55). Western blot analysis showed higher Nrf2 protein in the

nucleus of NSCLC cell lines A549 and H460 than in the cytoplasm (55). The combination

of these mutations on K-RAS and Keap1 generates high constitutive levels of active

Nrf2, providing an ideal condition to study this transcription factor. Although, both H460

and A549 cells share similar mutations, they are distinct one from another, at least in part

due to the presence of a PIK3CA gain of function mutation in H460 cells(89). This may

cause a differential regulation of the elements of the PI3K pathway in H460 cells as

suggested by expression analysis assays of mTOR in this cell line as compared to A549.

In a study by ZU-QUAN ZOU(103) these NSCLC cell lines were expose to GDC-0941, a

dual inhibitor of class I PI3K and mTOR. It was found that due activating PIK3CA

mutations, H460 cells were more sensitive to GDC-0941 compared to A549 cells, likely

reflecting cellular “ addiction” to the PI3K pathways, which is driving cell growth. Their

study showed that the molecular profile present in H460 cells generates a distinct

phenotype, when compared to A549 cells, with respect to mTOR regulation. A

constitutively highly expressed PI3K pathway could possibly bypass regulatory

mechanisms of its elements, such as the proposed Nrf2/ mTOR interaction, making this

cell line more Nrf2- independent.

The results found here in A549 cells, in western blot and qPCR studies imply that mTOR

is either a direct or an indirect target of Nrf2. To our knowledge, there is no documented

direct Nrf2 targeting of mTOR transcription. The literature only describes indirect

interactions where the intermediate proteins act post-translationally on mTOR levels.

Shibata T et al (104) described an indirect interaction between Nrf2 mutant and mTOR via

RagD. Utilized gene set enrichment analysis (GSEA) they determined that a mutant Nrf2

induced pathways associated with mTOR signaling; Peng Rapamycin DN, Peng Leucine

DN and Peng Glutamine DN. Additionally they found that mutant Nrf2 indirectly

upregulates the mTOR activator RagD and further that decreased Nrf2 reduced RagD

expression. This could indicate that Nrf2 directly regulates RagD expression, however,

the bioinformatic algorithms we use to identify potential ARE sequences found no such

motifs present in the promoter region of RagD. Thus, the link between Nrf2 and Rag D

remains unknown. Sasaki H et al (56) also found a correlation between mutant Nrf2 and

RagD gene expression in a study that involved 90 cases of surgically-treated lung

squamous cell cancer patients. Among all the patients 14 cases were positive for a Nrf2

gene promoter polymorphism, which generated a gain of function mutation. In this group

of patients RagD expression was 3 times higher, indicating at least a strong correlation

between Nrf2 and RagD.

The somatic mutations present on mutant Nrf2 also occur in its coding region and it

modifies amino acids in the DLG or ETGE motifs, causing abnormal cellular

accumulation of Nrf2, likely because the mutant does not bind efficiently do Keap1, the

chaperone for proteosomal degradation(105). The mutations present in Nrf2 in those

studies may explain why mTOR was not found on GSEA, as structural changes at the

protein level could have affected the detection of mTOR enrichment. Little is known

about the many effects that this mutation cold have on Nrf2 binding and, as it was

observed in our EMSA and ChIP results, the Nrf2/mTOR promoter binding is probably

weak. Perhaps, for this reason, the genome wide profile in the study Shibata T et al (104)

did not observed a direct interaction between Nrf2 and mTOR . Also, as already

described by Shibata et al 2008(105), the Nrf2 present in A549 cells is not mutated. The

Nrf2 mutation described in the literature occurs more frequently in lung cancers that do

not have an EGFR and Kras mutations(105). While, A549 cells express a wild type EGFR

and a mutated Kras gene (106), this NSCLC is an adeneocarcinomic alveolar basal sub-

type, whereas the Nrf2 gene somatic mutation was show to be more prevalent in lung

squamous cell carcinomas(107).

It is possible that an unknown element, which is influenced by Nrf2 levels in the cell is

responsible for the observed mTOR transcription regulation. In order to establish whether

mTOR is a direct Nrf2 transcriptional target, bioinformatics analyses were performed on

this gene promoter region looking for ARE biding sites. The ARE site TGACTCAGC is

believed to be the canonical ARE binding site, although some variations of that canonical

binding site have also shown to be functional. In our study, site directed deletion of the

“TGA” portion of the ARE showed that this ARE is essential for transcription as the

luciferase activity, as the mutated ARE showed a two fold decrease on A549 cells. When

Nrf2 is decreased using SiRNA, no change was observed on mTOR luciferase activity for

this cell line. However, when Nrf2 expression is increased, luciferase activity is increased

three fold in A549 cells transfected with mTOR-WT, whereas there was no change in

cells transfected with mTOR-mut. These data suggest that in A549 cells, while Nrf2 does

not affect basal mTOR expression, when Nrf2 is increased above basal levels it has a

direct effect on mTOR expression.

The nature of the Nrf2/mTOR promoter binding in A549 cell line, however, was unclear

as it contradicted the western blot and qPCR results, which demonstrated that Nrf2 down

regulation results in decreased mTOR protein levels. Various factors could have

contributed to this; one limitation of the luciferase assay is that only a small portion of the

mTOR promoter region was analyzed, excluding further upstream regions of the

promoter that could contain binding sites for interacting proteins and other regulatory

elements that could play a role on mTOR expression. Additionally, not yet identified

Nrf2 targets might bind to the mTOR promoter region and alter its transcription.

Therefore, for more clarification additional analyses on A549 cells .

Through DNA pull-down and EMSA studies, we found that Nrf2 binds to the mTOR

promoter region at the ARE present at 723 bp upstream from TSS under basal Nrf2

conditions. In the DNA pull down assay, the mTOR probe containing a scrambled ARE

site presented two fold decrease of the amount of mTOR protein bound when compared

with the wild type mTOR probe. In EMSA, the addition of the wild type mTOR probe

generated a blot with similar characteristic to the one presented in Nqo1. After removing

the ARE site from the mTOR promoter sequence and adding five bp in the extremities of

the probe, mutant mTOR the predicted Nrf2 binding site disappeared. These results

confirmed that the transcriptional activity observed on A549 cells at Nrf2 basal

conditions in the luciferase assay was truly due to Nrf2 binding. However, the

Nrf2/mTOR binding observed in the EMSA assay suggests that it is a weak binding as it

was only observed after four days of film exposure.

Additionally, It was shown via EMSA that Nrf2 binding to the mTOR promoter region

decreases in Nrf2 silencing conditions. Since no change was observed in the luciferase

assay when Nrf2 was silenced, the decrease in binding observed via EMSA assay, most

likely, does not affect mTOR transcription. Hence, possibly, the decrease in the mTOR

protein and transcriptional levels, observed via immunoblot and qPCR assay at Nrf2

silencing conditions, could be due to an unknown mTOR regulatory element that is

affected by Nrf2 levels in the cell.

An interesting finding, that is outside the scope of this study, is the putative presence of

NF-ΚΒ in mTOR promoter region. In the EMSA assay we observed that protein was

bound to another biding site present on the mTOR mutant probe. This protein or complex

was not present in the wild type mTOR probe, but was seen in the mutant. There is strong

evidence that this protein could be NF-ΚΒ. This transcription factor binds to the

consensus DNA sequence 5’-GGGRNYYYC-‘3 (in which R is a purine, Y is a

pyrimidine and N is any nucleotide) known as the ΚΒ site(108) .The 5’ end of the mutated

probe after the five bp extension is 5’- GGGAATTTC-‘3, which fits the specification of

the ΚΒ biding site. Also, when analyzing the mTOR promoter region using the UCSC

genome browser, NF-ΚΒ is present in the same position of the portion of the promoter

region of this study. The potential for influence of this transcription factor on mTOR

regulation by Nrf2 has never been studied.

Finally, in A549 cells, ChIP analysis confirmed the presence of a weak Nrf2/mTOR

biding under basal Nrf2 conditions along with establishing mTOR as an inducible target

of Nrf2, as mTOR enrichment was seen to increase 13 folds after Nrf2 up regulating in.

The basal cellular level of Nrf2 can be seen as an indication of low cellular stress, due to

redox or chemical toxicity (13,14). The Nrf2 levels are increased in response to any of these

stresses(13,14) Our data thus suggest that cell stressors can result in enhanced mTOR

transcription, which is of interest since mTOR activity can itself generate ROS(82,109).

For the first time it is shown that, on a specific type of NSCLC, Nrf2 directly interacts

with a putative ARE binding site present in the mTOR promoter region. It was

demonstrated that this biding resulted in the induction of mTOR expression at the

transcriptional level. Additionally, it was confirmed via ChIP assay that this binding

occurs at Nrf2 inducible conditions. The next step in order to fully understand this

interaction would be to identify the biological conditions that lead to mTOR

transcriptional activation via Nrf2. There are descriptions in the literature of regulatory

mechanisms between Nrf2 and mTOR(110). Although not yet fully understood, these

interactions were shown to generate a positive loop that regulates autophagy(110). On one

side Nrf2 inactivates mTOR activity through an entirely different, and yet undefined

mechanism that determines mTOR phosphorylation(110) and in the other side this decrease

in mTOR activity indirectly increases Nrf2 expression(111).

Nrf2 is known to participate in autophagy and modulation of hepatic regenerative

response to liver mass loss(112). It is persistently active whenever autophagy is deficient in

the cell, which was shown to be due to Nrf2 up-regulation of Bcl-2(97) and p62(100).

The p62 protein, also called polyubiquitin-binding protein p62/SQSTM1(sequestosome

1), is encoded by the SQSTM1 gene and is responsible for protein aggregation and for

facilitating the passage of those cellular components into autophagosomes for lysosomal

degradation(113) mTOR is also related with autophagy and control of liver architecture.

When activated, this protein decreases lysosomal degradation of intracellular

components. In order to regulate the number and size of maternal hepatocytes, mTOR

needs to be activated in a stage-dependent phosphorylation pattern. When exploring the

role of Nrf2 in the regulation of maternal hepatic adaptation to pregnancy, Yuhong Zou et

al(110) stated that Nrf2 is a negative regulator of mTOR signaling. Their study showed that

the progressive decrease in Nrf2 activity was correlated with a gradual increase in mTOR

phosphorylation in the maternal liver during the second half of pregnancy. Also, Nrf2

deficient mice caused hyperphosphorylation of mTOR in the non-pregnant state. The

mechanism by which Nrf2 inhibit mTOR phosphorylation was not approached in this

study.

The p62 protein along with being transcriptionally up regulated by Nrf2, also increases

Nrf2 stabilization through Keap1 docking, which blocks the Nrf2 binding, thus

generating a positive feedback loop(114). Lerner et al (111) presented mTOR as part of the

p62 interaction with Nrf2 when they showed that reduction in mTOR activity leads to

increased turnover of p62/SQSTM1. Therefore, Nrf2 decreases mTOR activity which in-

turn increases p62/SQSTM1turnover, resulting in increase of Nrf2 activity.

The putative regulatory loop between Nrf2 and mTOR indicates that the two proteins can

interact in a complex biological system and in various ways. These interactions however,

are indirect and occurs at the posttranslational level. In our study we showed that Nrf2

can bind directly to the mTOR promoter region and that it can activate its expression at

the transcriptional level. It is known that depending on the cell type and stress conditions

ROS can serve as a signal for autophagy through various pathways(115). Based on the

literature(77, 116, 117) , we believe that the mTOR transcriptional regulation via Nrf2 is also

dictated by ROS. Both Nrf2 and mTOR activities are altered by ROS levels in the cell.

While regulatory loop of ROS in Nrf2 is well known; an increase in ROS activates Nrf2

which in turn transcriptionally upregulats anti-oxidants genes thereby decreasing ROS.

However, mTOR regulation by ROS is an intricate processes. Chen et al (116) have shown,

in PC12 cells and primary murine neurons, that apoptosis of neural cells via oxidative

stress could be due to the H2O2 inhibition of the mTOR-mediated phosphorylation of

S6K1 and 4E-BP1. This inhibition of mTOR activity occurs indirectly, through the

activation of AMPK and inhibition of AKT and PDK1 phosphorylation. Another

condition where mTOR is inactivated by ROS is via activation of LKB1/AMPK/TSC2 in

Ataxia-telangiectasia mutated (ATM) cells. Mutation of the ATM protein kinase in

ataxia- telangiectasia (AT) is known to generate a high ROS environment in the cell. In

this condition the LKB1 tumor suppressor activate AMPK which activate TSC2 and

therefore inhibiting mTORC1(109) While these are examples of ROS decreasing mTOR

activity, there is much evidence for the opposite effect, and we believe that our findings

regarding Nrf2 here are relevant (118,119).

The role of mTOR as a regulator of metabolism its attributed to its capacity to integrate

signals from the cell environment, such as nutrients and oxygen availability, with protein

translation(77). Arsham et al (118) demonstrated that oxygen is a modulator of the mTOR

pathway signaling with oxygen activating mTOR while hypoxia inhibiting it. Other

studies have shown that O2 byproducts can also activate mTOR, and ROS can even work

as a messenger in nutrient-sensing pathway as it was illustrated by the activation of the

mTOR pathway via Leucine generated ROS(120). In the early stage of incubation with this

amino acid mTOR phosphorylation was independent of the PI3K pathway and at a later

stage mTOR was activated via a ROS mediated IR/IGF-IR phosphorylation which

activate the PI3K/Akt/mTOR and ERK signaling pathways.

The ROS mediated mTOR activation was shown by some groups to be via AKT while

other groups stated that it could also occur as an AKT-independent process. Radisavljevic

et al (121) showed mTOR activation by ROS is a processes that occurs downstream from

the PI3K/Akt signaling pathway, in the study on H2O2 exposed mice type II pneumocytes.

In this study they found that ROS-mediated mitosis was due to H2O2 phosphorylation of

AKT at Ser473, which activated mTOR. Huang C et al (119) demonstrated an AKT-

independent model of ROS mediated mTOR activation in mouse epidermal JB6 Cl41

cells. In their study, the PI3K/TOR /S6K pathway was activated after UV- generated

H2O2 exposure. They found, after combining H2O2 treatment with Rapamycin, that

phosphorylation of p70S6K at Thr389 and Thr421/Ser424 was an AKT independent and

mTOR-dependent processes. The molecular mechanism responsible for mTOR activation

was not revealed(119). The mechanism of AKT-independent of mTOR activation was

shown later by Sanchez Canedo C. et al (122) They showed that PDK1 is required for

PRAS40 phosphorylation in a Leucine mediated activation of the cardiac mTOR/p70S6K

pathway. Our study has shown that mTOR can be activated at the transcriptional level via

Nrf2. In theory, this PI3K independent activation of mTOR might also be mediated by

ROS.

The central role of ROS regulation by both mTOR and Nrf2, their importance in cancer

metabolic reprogramming together with the identification of mTOR as an Nrf2 inducible

transcriptional target, suggest that a ROS mediated feedback loop is possible between

these two proteins. On the one hand, mTORC1 is important for stimulation of

transcription of genes involved in mitochondrial biogenesis, consequently increasing the

levels of ROS in the cell as a byproduct of the respiratory chain(117). On the other hand,

oxidative stress generated by ROS would lead to breakdown of the Nrf2-Keap1 complex

and subsequently transcription of anti oxidants genes and mTOR by Nrf2 in the nucleus.

The ROS inhibition via Nrf2 activation and the mTOR inhibition by ROS could

potentially work as self-regulatory mechanisms of this feedback loop. This novel

interaction between mTOR and Nrf2 could help us better understand the ROS regulatory

system in cancer.

The other elements of the Pi3K pathway were analyzed only in a preliminary fashion,

using western blot, qpcr and luciferase assays, in order to determine if Nrf2 also

interacted with the different elements of the pathway, apart from mTOR. The results

showed varying degrees of regulation of Nrf2 on the other elements of the pathway.

However, further assays need to be performed in order to accurately evaluate the role of

Nrf2 on the regulation of TSC2, AKT and S6K activities and/or expression.

It has been shown previously by Malhotra, D. et al (37) that TSC2 is a possible target of

Nrf2 at basal levels. Our data suggest that in Nrf2-inducible conditions TSC2 is a indirect

Nrf2 transcriptional target. However, when silencing Nrf2, only TSC2 protein levels

where decreased. Therefore, Nrf2 may act on TSC2 expression at the transcriptional level

at inducible conditions and at the protein level when downregulated.

The discordance with Malhotra, D. et al results could be due the cell type specificity of

these effects that we found, since the ChIP-Seq study utilized MEFs cells and our

experiments were done on human cancer cell lines. Furthermore, interacting proteins,

which could be directly or indirectly influenced by Nrf2 levels in the cell, may regulate

TSC2 expression. MAPK is involved in activation of Nrf2 by various pathways and is

also responsible for TSC2 phosphorylation. Perhaps, this indirect interaction observed

between Nrf2 and TSC2 is mediated by MAPK.

To our knowledge, there is no description in the literature showing that Nrf2 directly

interacts with AKT. There are however, examples where Nrf2 enhances, via AKT

interactions, the metabolic reprogramming triggered by proliferative signals. This was

observed in the Nrf2 mediated AKT phosphorylation during PPP(79) and in the redirection

of glucose and glutamine into anabolic pathways in cells with high Nrf2 levels and active

PI3K-AKT signaling. In the present study AKT was shown to be a possible indirect Nrf2

transcriptional target in Nrf2 inducible conditions. This interaction however may be also

cell line dependent as it is the case for mTOR. The higher metabolic state of H460 cells

lines when compared to A549 cell, is primarily due to the PI3K gain of function

mutation, and may explain the presence of a possible interaction between Nrf2 and AKT

in H460 cells. When silencing Nrf2, AKT was two folds decreased at the protein level on

HEK293 cells. In a previous study(80) on ROS mediated insulin/IGF1 resistance, NRF2

KO mice were shown to decrease AKT activity but not at the proteins level. Therefore,

this interaction is yet to be clarified.

Also, to our knowledge, there is no previous mention elsewhere of a direct Nrf2

interaction with S6K. This well-known substrate of mTORC1 is not affected at the

transcription or protein level by Nrf2. Therefore the data obtained in our study are the

result of an unknown interaction. While no significant change was observed on S6K

levels after Nrf2 upregulation, Nrf2 silencing induced a 5.84 folds increased in A549

cells, suggesting that Nrf2 may have a direct or indirect inhibitory effect on S6K at the

protein level. It would be interesting to know in the future how Nrf2 inhibit S6K and if

this inhibition plays a role in the proposed ROS mediated feedback loop between Nrf2

and mTOR.

In order to establish if the Nrf2 transcriptional activation of mTOR is mediated by ROS,

lactate and ROS levels will be measure in a time course experiment together with

proteins activity. This study may utilize AKT inhibitors so as to observe if the PI3K

pathway participates in the proposed feedback loop.

In the future, others transcription factors that may interact with Nrf2 at the mTOR

promoter region will be studied in order to better understand this regulatory system.

It was observed in preliminary ChIP results, but not yet confirmed, (data not shown) that

Nrf2 may interact in the mTOR’s ARE binding site with JundD, a member of the AP-1

family.

Like Nrf2, the AP-1 family of transcription factors also belong to the bZip family and

they are comprised by the Jun (c-Jun, JunB and JunD) and Fos (c-Fos, FosB, Fra-1 and

Fra-2) families. Among the functions from which these transcription factors are involved

are proliferation, differentiation, apoptosis and development(123). The core sequence of

those transcription factors biding site is 5’-TGA(G/C)TCA-3’ and it is called TPA

response element (TRE) (123). Similary, the Nrf2 partner small Maf ( Maf G/F/K) as well

as the other members of the Maf family (c-Maf, MafB, MafA, and Nrl) also bind to the

TRE(124).

The TRE closely resembles the Nrf2 transcription factor binding site ARE. The main

difference between ARE and TRE is the presence of the GC box at the end of the ARE,

which is essential for the ARE-driven gene expression. Jaiswal et al(34) showed that the

human NQO1 ARE contained TRE and TRE like elements and, AP-1 family members,

Jun and Fos, bind to the human Nqo1 ARE. They also demonstrated that overexpression

of combinations of nuclear proteins Jun and c-Fos or Jun and Fra1 downregulated hARE-

mediated gene expression, however, Jun proteins alone did not exert any significant

effect on the hARE. Similar effects of AP-1 proteins were also seen on γ-

Glutamylcysteine synthetase (γ-GCS) by Jaiswal et al. (125) They showed that Nrf2

heterodimerized with c-Jun to significantly upregulate. The binding of Nrf2 and AP-1

proteins to ARE is believed to be competitive. For example, the γ-GCS ARE-mediated

expression is differentially regulated by AP-1(Jun+Fos; negative) and Nrf2(positive

effect) and the dynamic is thought to be determined by the amount of each nuclear

protein present.

It is presumed from preliminary ChIP results that JunD acts as a co-factor for Nrf2 when

binding to mTOR. However, further ChIP analyses are necessary in order to confirm this

interaction. Also, a EMSA super shift experiment will be performed with A549 cells line

in Nrf2 basal, silencing and inducible conditions. In this experiment the mTOR wt and

mTOR mutant probes will be utilized together with JunD. This experiment will further

confirm the presence of Nrf2 on mTOR biding site and also demonstrate the interaction

of this members of the AP1 family with Nrf2 at the mTOR promoter region.

The Nrf2/mTOR interaction explored in this study may also have future application in the

clinic. With various studies focusing on inhibitiors of mTOR activity no studies have

been done on its inhibition at the transcription level. Recent attempts to develop mTOR

inhibitors largely focus on rapamycin analogs. Sadly, clinical trials suggest that

rapamycin is effective in few cancers only (B cell lymphoma, endometrial cancer, and

renal cell carcinoma)(126).This restricted therapeutic effect is due the presence of a feed

back loop, as activation of mTORC1 signaling strongly represses PI3K- AKT signaling

upstream in the PI3K pathway (reviewed in Manning, 2004). This inhibition is done via

mTORC1 dependent S6K activation, which inhibits the insulin receptor substrate 1(IRS-

1) ultimately blocking the PI3K/AKT pathway. This class of drugs not only activates

proliferative effectors, such as AKT, but also incompletely inhibits mTORC1. Therefore,

rapamycin analogs could potentially be responsible for hyperactivation of rapamycin-

resistant mTORC1. Also, mTORC2 works as a PI3K by its direct activation of AKT by

phosphorylation in the hydrophobic Ser473(127). The mTORC2 activity is blocked only at

high toxic levels of Rapamacin(128).

Dual inhibition of the PI3K pathway or other signaling pathways and mTOR could be an

effective strategy (Fan and Weiss, 2006; Wan et al., 2007). Among the drugs that explore

this approach are gefitinib (Iressa, an EGFR inhibitor), imatinib mesylate (Gleevec, a

BCR-ABL inhibitor), tamoxifen (estrogen receptor modulator), cisplatin (DNA damaging

agent), and paclitaxel (microtubule stabilizer) (reviewed in Faivre et al., 2006; Granville

et al., 2006). This line of treatment seems effective in cancer cell lines and tumors with

hyperactivated PI3K pathway. However, presence of K-Ras hyperactivation in Ras-

driven tumorogeneses made necessary the additional blocking of downstream RAS

mediators. This combination of signaling disturbance could lead to toxicity in normal

cells(129).

The abovementioned reflects that the mechanism of action responsible for mTOR

activation is still not fully understood. Possibly, Nrf2 as an mTOR transcriptional

activator could be explored in the future as a new mechanism for the development of a

combined therapy for cancer treatment, with mTOR inhibitors and the blockage of Nrf2

activity.

Lastly, a newly discovered Nrf2 function will be explored alongside with mTOR. Since it

targets mTOR, Nrf2 may also therefore play a role in protein translation (not published

data). In order to better understand the role of Nrf2 on translation and the effects of the

mTOR induced transcription via Nrf2 transactivation, a Polisome fraction assay will be

performed at Nrf2 basal, induction and silencing conditions. Simultaneously, mTOR

inhibitors will also be used so as to evaluate if the translation is due mTOR induction.

7. Acknowledges________________________________________________________

I would like to acknowledge Dr. Batist for giving me the opportunity to work with him in this exhilarating field of research and providing me with his knowledge throughout the project. I am obliged to Dr. Tahar Aboulkassim for his patience and guidance and my lab colleagues Dr. Liu Qiang and Sujay Shah for their assistance. Lastly, I would like to express thanks to the members of Dr. Witcher lab, Dr. Maud Marques, Dr. Khalid Hilmi and Dr. Tiejun Zhao and also the members of Dr. Alaoui-Jamali lab Dr. Sabrina Wurzba and Amine Saad for their support.

Figure 1. Nrf2 inducible construst plus constructs used for Luciferase assay. A. The inducible construct PC_Nrf2 contained 1925bp of Nrf2 coding sequence (green line) cloned on the expression vector pcDNA 4.0. B –F. All the constructs used on Luciferase assay comprised of the promoter region of the gene of interest (green line) cloned on pGL3-basic Luciferase reporter vectors, C terminally to the Luciferace reporter sequence (purple line). B. PGL3-Nqo1 contained 550 bp of Nqo1 promoter region C. PGL3-mTOR contained 1231bp of mTOR promoter region D. PGL3-TSC2 contained 1079 bp of TSC2 promoter region E. PGL3-S6K contained 2660bp of S6K promoter region F. PGL3-AKT contained 2200 bp of AKT promoter region

Table 1. Primers used for qPCR, Luciferse constructs, DNA pulldown, EMSA and ChIP assay

5’-TGAAGGTCGGAGTCAACGGA-3’ 5’-GAGGGATCTCGCTCCTGGAAG-3’5’-GAGAGCCCAGTCTTCATT-3’ 5’-TGCTCAATGTCCTGTTGCAT-3’5’-GCTGGTTGAGCGAGTGTTC-3’ 5’- CTGCCTTCTTACTCCGGAAGG-3’5’-GACTTCGCCCATAAGAGGCA-3’ 5’-TAGCTGTGGAATCTGACGGC-3’5’-GTCTCTTCCGAGATTTTCGACG-3’ 5’-ATCCCCAACATTGTTAAGCGT-3’5’-GGACGCTGGAGAAGTTCAAG-3’ 5’-CGGATTTTTGGTTCAAAGGA-3’5’-TCTATGGCGCTGAGATTGTG-3’ 5’-CTTAATGTGCCCGTCCTTGT-3’

5'-GGTGGGTACCCAAAACCAATTA-'3 5'-GGTG CTCGAG G CCGTTTGAGGTGGACAGCCTA-3'5'-CCTTCCAAATCCGCAGCAGTGACTCAGCA5'-TTCTGCTGAGTCACTGCTGCGGATTTGGAAGG-3'5'-GGTG ACGCGTGTGCCAGGCCCTAGA-3'5'- GTCGAGACCAGCCTGGACATGGTGAAAT5'-GGTGGGTACCGGGAAAAAGGCGCAAGGT5'-CATGCGCCCCGCGTGATGCAAGG-3'5'-GGTG GGTACC GCCAGCAGGGTCCTCTTT GGGCCCGGACCCGCCGC-3' 5'-CCGAACGCTCCAGCCAATGCGCATGCTC-5'-ATATGGTACGCATGCCTGTCCACCGAACG5'-GGTGGGTACCGCCAGCAGGGTCCTCTTTT

5’ -/Biosq/GCAGTCACAGTGACTCAGCAGAA 5’ -/Biosq/TTCTGCTGAGTCACTGTGACTGC- 3’5’ -/Biosq/GCAGTCACAGACTCTCAATAGAAT5’ -/Biosq/ATTCTATTGAGAGTCTGTGACTGC-3’5’- /Biosq/GTCGAGACCAGCCTGGTCAACAT 5’- /Biosq/ATGTTGACCAGGCTGGTCTCGAC-3’5’- Biosq/GTCGAGACCAGCCTGGTAAAT GG 5’ /Biosq/CCATTTACCAGGCTGGTCTCGAC-3’

5'-GCAGTCACAGTGACTCAGCAGAATCTG A5’-CTCAGATTCTGCTGAGTCACTGTGACTGC-3'5’-GCA GTC ACAGACTCTCAATAGAAT-3’ 5’-CTCAGATTCTATTGAGAGTCTGTGACTGC-3’5'-GTCGAGACCAGCCTGGTCAACATGGTGA 5’-TTCACCATGTTGACCAGGCTGGTCTCGAC-3’5’-CAGGAGTCGAGACCAACATGGTGAAATT5’-GGGAATTTCACCATGTTGGTCTCGACTCC-3’

CHIP5'-CGGTGGCTCACGCCCATAATT-'3 5'-CAGCCTCCCGAGTATC-'35-AAGTGTGTTGTATGGGCCCC-'3 5'-GTGGAAGTCGTCCCAAGAGA''3