Download The role of hypoxia in colon cancer cell

The role of hypoxia in colon cancer cell
resistance to cytotoxic antitumour agents and
modulation of Hypoxia-inducible factor-1 as
a strategy to circumvent chemoresistance.
Brecht QUINTENS
Master’s thesis submitted
to obtain the degree of
Master in the Biomedical Sciences
Promoter: Prof. Dr. J. Gettemans
Department of Medical Protein Research,
Ghent University
Co-Promoter: Prof. Dr. E. Monti
Department of Structural and Functional Biology,
University of Insubria
Erasmus Programme
Academic year: 2008-2009
1
The role of hypoxia in colon cancer cell
resistance to cytotoxic antitumour agents and
modulation of Hypoxia-inducible factor-1 as
a strategy to circumvent chemoresistance.
Brecht QUINTENS
Master’s thesis submitted
to obtain the degree of
Master in the Biomedical Sciences
Promoter: Prof. Dr. J. Gettemans
Department of Medical Protein Research,
Ghent University
Co-Promoter: Prof. Dr. E. Monti
Department of Structural and Functional Biology,
University of Insubria
Erasmus Programme
Academic year: 2008-2009
2
3
Preface
This thesis is a part of the Erasmus exchange project among the universities of Insubria and
Ghent. I have performed a 7 month internship at the laboratory of Prof. Dr. E. Monti located
in Busto Arsizio, which is specialized in anti-cancer pharmacology.
First of all I thank my co-promoter Prof. Dr. E. Monti for the excellent guidance throughout
the project as well as Prof. Dr. M. Gariboldi, Dr. R. Ravizza and last but not least Dr. R.
Molteni, each of them possessing great theoretical and practical skills, and, a lot of patience
and devotion to let me acquire the variety of laboratory - and cell techniques. Apparently this
also applies for the other lab-members who even inspired me to learn the Italian language.
I would like to thank Prof. Dr. G. Perletti for his concern as Erasmus responsible, in fact, I
consider him as my ‘Erasmus-dad’; hotel-labo shuttle services, sport material, nothing is too
much for him. His Belgian counterpart, Prof. Dr. J. Gettemans who is Erasmus responsible for
the Biomedical Sciences at UGhent and also my promoter, would I like to thank thoroughly
because of the many arranged essential things, including the traineeship in the first master.
Thanks both of you!
I also want to thank my parents - not for doing my laundry and catering services the last 8
months - , but for giving me this and other opportunities.
It’s not possible to end without mentioning Serge Hoefeijzers, my ‘Erasmus-buddy’. We have
gone through the same things and helped one another where useful, or where compulsory
(laundry for instance) and we strengthened each other as individuals, hence we were able to
finish this academic year.
As a biomedical scientist I am supposed to end with a conclusion:
Personally I experienced whole this Erasmus-programme as enriching, fantastic and intense in
both a cultural and scientific point of view!
Brecht Quintens, Milan, 20th May 2009.
4
Index
Abstract...................................................................................................................................1
Introduction...........................................................................................................................2
1. Colon cancer
1.1. Epidemiology and risk factors
1.2. Genetic basis and pathogenesis....................................................................3
1.3. Therapeutic strategies...................................................................................5
1.3.1. Oxaliplatin
1.3.2. 5-Fluorouracil..............................................................................8
1.3.3. Cetuximab & Bevacizumab.......................................................11
1.3.4. Combination strategies
2. Colon cancer & Hypoxia.......................................................................................12
2.1. Hypoxia
2.2. Hif-1 structure and regulation....................................................................15
2.2.1. Oxygen dependent HIF-1 regulation
2.2.2. Oxygen independent HIF-1 regulation......................................17
2.3. HIF-1 and drug resistance..........................................................................18
3. Modulation of HIF-1.............................................................................................19
3.1. PMX290.....................................................................................................20
3.2. Antisense oligonucleotides
4. Aims........................................................................................................................21
Materials & Methods........................................................................................................23
Results....................................................................................................................................29
1.
2.
3.
4.
Effect of hypoxia on HIF-1α expression
Effect of hypoxia on HIF-1α activity
Effect of hypoxia on cellular response to 5FU and OxPt........................................30
Effect of modulation of HIF-1α on the response of colon
cancer cells to 5FU and OxPt..................................................................................34
5. Effect of EZN2969 treatment on HIF-1α expression, HIF-1
activity and apoptotic response to 5FU in HCT116cells.........................................38
6. Effects of the expression of a degradation-resistant variant
of HIF-1α on the response of HCT116 cells to 5FU...............................................39
Discussion..............................................................................................................................41
1. 5-Fluorouracil
2. Oxaliplatin...............................................................................................................44
3. General Conclusion.................................................................................................46
References.............................................................................................................................47
5
Abstract
Tumour hypoxia represents a major obstacle to the success of chemotherapy. Hypoxiainducible factor 1 (HIF-1) is the pivotal agent of cellular response to low oxygen levels,
which is apparent in colon cancers, and these facts led to the concept that inhibiting HIF-1
activity may sensitise hypoxic colon cancer cells to cytotoxic drugs.
In this study, we investigate the effects of HIF-1 modulation on the response of two human
colon adenocarcinoma cell lines, HCT116 and HT29, which differ in p53 status, to either 5fluorouracil (5FU) or oxaliplatin (OxPt), both currently used in the treatment of colon cancer.
Increasing HIF-1 activity, either by exposing the two cell lines to hypoxia or by forced
expression of a degradation-resistant form of HIF-1α in HCT116 cells, results in poor cell
response to 5FU; conversely, knockdown of HIF-1α by antisense oligonucleotides targeting
the HIF-1α mRNA prevents hypoxia-induced resistance to 5FU. PMX290, a thioredoxin-1
inhibitor, significantly inhibits HIF-1 activity and concomitantly sensitises both HCT116 and
HT29 hypoxic cells to the cytotoxic effect of 5FU and OxPt. Moreover, these results were
confirmed in HCT116 cells grown as three-dimensional spheroids, a model that more closely
reproduces the hypoxic environment of solid tumours. Based on these observations,
downregulation of HIF-1 activity is a potential approach to the circumvention of
chemoresistance in the clinical management of colon cancer.
1
Introduction
1. Colon cancer
1.1.
Epidemiology and risk factors
Colon cancer is the third most common cancer in the world in terms of prevalence and there is
a worldwide mortality/incidence ratio of 52 % [1, 2]. In Flanders (Belgium) the 5-year
survival is 57 %, but this ranges from 65 % in North-America to 30 % in India, indicating a
substantial variation in therapeutic strategies and/or options [3, 4]. So there is a general
relatively good prognosis, nevertheless more than half a million people die each year of
colorectal cancer due to the high incidence. Unlike most locations, this cancer is somewhat
more common in males than in females, with a ratio of 1,2 : 1. In terms of worldwide
incidence, colorectal cancer ranks fourth in frequency in men and third in women [4].
Concerning risk for colon cancer, four major categories of risk factors can be identified.
1) Environmental factors. Studies on migrant populations suggest that colon cancer risk is
determined largely by environmental exposure [5]. There are strong geographical differences
which can be denoted to different environmental exposures, i.e. the higher the living standard
and extent of industrialization, the higher the age-adjusted incidence of colorectal cancer [4].
2) Dietary factors. Diet is definitely the most important exogenous factor identified up to now
in the aetiology of colon cancer. There is a strong positive correlation between risk of colon
cancer and per capita consumption patterns of red meat, animal fat and alcohol, whereas a
negative correlation has been reported for vegetable and fibre intake, as shown -after some
controversy- by the large prospective EPIC study [6, 7].
3) Non-dietary factors. Physical activity and chronic use of NSAID’s has consistently been
associated with a decreased colon cancer risk, whereas tobacco use and a BMI above 30
accounts for an increased risk of colon cancer [6].
4) Genetic factors. Family history definitely plays an important role as a risk factor for
colorectal cancer. Interestingly, the molecular basis for malignant colorectal cancers is
relatively well established, at least in comparison with most other human cancers [5]. This
genetic basis is outlined in the next chapter.
2
1.2.
Genetic basis and pathogenesis
The great majority (> 95 %) of colon cancers is sporadic, but a small tumour subpopulation
arises as a consequence of inherited alleles that create a substantial lifelong risk for this
disease. The colorectal cancers caused by highly penetrating mutations are: Familiar
adenomatous polyposis (FAP) and Hereditary nonpolyposis colorectal cancer (HNPCC) [8].
•
FAP defines a condition where more than 100 polyps can be found within one individual,
or less than 100 polyps in an individual with first degree relative with FAP. The frequency
is 1 % of all colon cancer patients. The causal mutation is located on chromosome 5
(5q21-22), affecting the APC tumour suppressor gene. Of all FAP patients nearly 90 %
possess APC mutations. If the gatekeeper mutation is present there is a penetration rate of
nearly 100 %. More than 95% of these mutations, typically insertions or deletions, will
lead to a truncated protein or nonsense mutations [9, 10]. The APC protein is active in the
Wnt signalling pathway; it carries out the phosphorylation of β-catenin, causing its
subsequent proteosomal degradation. This intracellular protein can interact with the
cellular adhesion molecule E-cadherin, which in turn interacts with the actin cytoskeleton
[11]. APC mutations cause cytoplasmic β-catenin accumulation, leading to increased
DNA binding of TCF transcription factors (T Cell Factors) or Lymphoid Enhancer Factors
(LEF), and modulating the expression of the TCF/LEF responsive genes, including genes
involved in proliferation, differentiation, migration and apoptosis (e.g. c-myc). APC also
plays an important role in cell cycle control by inhibiting the G0/G1 to S phase progression
and by stabilising microtubules, thereby promoting chromosomal stability [12]. Figure 1
illustrates how APC mutations can cause polyp formation.
•
HNPCC (or Lynch syndrome), is associated with germ line mutations in 6 DNA mismatch
repair genes (MMR) [5]. Several criteria allow diagnosing HNPCC. The frequency is at
least 4% of all colon cancer patients. In 70 % of the cases mutations in MLH1 (3p21),
MSH2 (2p21) or MSH6 (2p16) genes, encoding MMR proteins, are present, with a
penetration rate of 80 % [13,14]. MMR genes function as tumour suppressors.
Homozygous mutations cause a marked decrease in DNA repair, which in turn causes
higher mutation rates (mutator phenotype). MMR mutations also cause microsatellite
instability (MSI), resulting in increased genomic instability [15].
3
Fig. 1| β-catenin and the biology of colonic crypts. Obtained from [16].
Colon cancer has provided an useful model for the understanding of the multistep process of
carcinogenesis ‘thanks’ to the existence of causative mutations as described above.
Considering the Knudson two-hit hypothesis, a progression to a malign carcinoma is more
likely to occur in cells already carrying germ line mutations in APC or MMR genes, in
comparison with non-mutated colonic epithelial cells [17]. Furthermore, both these hereditary
syndromes, as well as sporadic colon cancers, undergo a specific stepwise progression,
described by Vogelstein, starting from normal bowel epithelium and ending in a metastatic
carcinoma (Figure 2). Vogelstein et al. examined genetic alterations in colon cancer
specimens at various stages of neoplastic development and found that changes in the 5q
chromosome, in APC and in the KRAS oncogene tend to occur relatively early in the pathway
[18]. Further downstream in the progression to malignancy is the deletion of chromosome 18.
This is frequently deleted in carcinomas and advanced adenomas and is thus named ‘deleted
in colon cancer’ (DCC). Other mutations, including p53, epigenetic changes such as
methylation of CpG islands of the MLH1 promoter and subsequent mismatch repair defects
lead ultimately towards a malignancy. So there is a progressive acquisition of abnormalities
(over various time-frames) of the genome, affecting known proto-oncogenes or tumoursuppressor genes and including epigenetic changes, ultimately leading to a metastatic,
invasive carcinoma [6].
4
Fig. 2| Vogelstein model for the carcinogenesis of colon cancer. Obtained and adapted from [18].
1.3.
Therapeutic strategies
Next to radiation therapy and surgery, (poly-)chemotherapy is commonly used and is outlined
in following chapters.
1.3.1. Oxaliplatin
Introduction
Oxaliplatin (OxPt) is a diaminocyclohexane-containing platinum (DACH-Pt) derivative, that
has been approved by the FDA and is widely used in cancer chemotherapy, usually as part of
combination therapies (Figure 3). The DACH-Pt complex of oxaliplatin can exist as three
isomeric conformations that interact differently with DNA, the trans I (R,R) isomer being the
most effective [19].
Fig. 3| Chemical structure of Oxaliplatin:
1,2-diaminocyclohexaneoxalato platinum.
Obtained and adapted from [20].
This third generation platinum anticancer drug was derived from cisplatin (i.e. first
generation) in a screen aimed at identifying platinum analogs with a broader spectrum of
activity, less prone to encounter resistance and with lower (neuro)toxicity than the parent
compound. Accordingly nowadays oxaliplatin is currently used in the clinical management of
cisplatin-resistant tumours, including colorectal cancers, that are intrinsically resistant to
cisplatin [20, 21].
5
Mechanism of action
OxPt enters the cell by passive diffusion and with the help of high-affinity copper transporters
(CTR1). Hereafter the molecule is converted into an aquated form that is more reactive than
the parent molecule, especially towards DNA, with which it forms different types of adducts
(Figure 4).
Fig. 4| Biotransformation pathway of Oxaliplatin. Obtained from [19].
These adducts arise because the platinum atom of oxaliplatin forms covalent bonds to the N7
positions of purine bases -preferably of nuclear DNA- resulting primarily in 1,2- or 1,3intrastrand crosslinks. In addition, interstrand crosslinks and DNA-protein crosslinks have
also been reported, although with lower frequency. DNA binding can occur by displacement
of the oxalate ligands originally present in the compound [19, 20]. Because of the steric
hindrance of the DACH carrier group, oxaliplatin distorts the DNA duplex, bending it
significantly towards the major groove, which in turn exposes a wide, shallow minor groove
surface to which several classes of proteins can bind and activate several cellular processes
that mediate the cytotoxic effect of this drug [21]. These (interfering) binding proteins are:
• Mismatch repair proteins. The binding of the mismatch repair complex to Pt–DNA
adducts appears to increase the cytotoxicity. This either by activating downstream
signalling pathways that lead to apoptosis or by causing ‘futile cycling’ during
translesion synthesis past Pt–DNA adducts. Although this would seem logic for all PtDNA adducts, these effects appear to be specific for cisplatin but not for OxPt
adducts; MMR deficiency or MMR mutations (hMSH2 and MutS, both components of
the MMR complex) are apparently not a determining factor for the OxPt treatment
outcome [22] .
• Damage recognition proteins such as: i) HMG box proteins (e.g. structure-specific
recognition protein 1 (SRP1) or high-mobility group box protein 1 (HMGB1) as the
most abundant ones), ii) TATA box-binding proteins (TBP) and iii) human upstream
binding factors (UBF). HMGB1 has been linked to several DNA-dependent pathways
6
(i.e. RAG1/2, MAPK and p53 possibly leading to apoptosis via a Bax-dependent
pathway) and modulates the efficiency of nucleotide excision repair [20].
So the DNA damage caused by OxPt modulates several signal transduction pathways (e.g.
AKT, c-ABL, p53, p38MAPK, JNK and ERK pathways) finally determining the cytotoxic
and/or resistance outcome. Next to these pathways, DNA-Pt adducts can also directly block
DNA replication and transcription [20].
DNA is not the only target for OxPt, cellular proteins may also be affected. Knowing that 7585 % of the intracellular OxPt is bound to proteins, it is not surprising that this effect can also
lead to detrimental consequences. Meynard et. al. postulated in 2007 the following
hypothesis. “After entry in the cell, oxaliplatin would especially target the thiol groups (i.e.
cysteine and methionine) of nascent proteins, which would be particularly sensitive to the
reactive oxygen species (ROS) produced by mitochondrial respiration. The resulting protein
oxidation would be at the origin of cell death.” If this hypothesis is correct, both DNA and
protein-mediated damage could lead to apoptosis in response to oxaliplatin treatment [23].
Resistance mechanisms
Some of the resistance mechanisms counteracting OxPt cytotoxicity are drug-specific, while
others more generally affect other drugs and/or act in a cell-type specific fashion (see
Discussion).
•
Translesion (TLS) DNA polymerases. Several translesion DNA polymerases have
been shown to bypass Pt–GG intrastrand adducts. TLS polymerases can bypass the
damage in an error-free or error-prone fashion, the first resulting in a resistance
mechanism, the latter in elevated mutagenesis, possibly leading to apoptosis [22].
• Nucleotide excision repair (NER). NER is the only known mechanism by which bulky
adducts (as those caused by DACH) are removed from DNA in human cells; increased
activity of this repair system is one of the major causes of OxPt resistance [24].
• Nonspecific inactivation of OxPt [20].
• Decreased expression of the copper influx transporter CTR1, and/or, increased
expression of the copper efflux transporter ATP7A mediates OxPt resistance [25].
7
1.3.2. 5-Fluorouracil
Introduction
5-Fluorouracil (5FU) is widely used in the treatment of cancer. Despite the fact that the
mechanisms of action of this drug are clearly understood, resistance remains a significant
limitation to the clinical use of 5FU. This fluoropyrimidine drug has a dual function in
inhibiting the normal cellular metabolism. First, 5FU is an uracil analog in which a fluorine
atom replaces hydrogen at the C-5 position; following activation and phosphorylation to the
corresponding triphosphate nucleotide the drug is misincorporated into RNA end DNA,
disrupting nucleotide-synthesis. Second, 5FU inhibits the nucleotide synthetic enzyme
thymidylate synthase (TS). It might be important to note that in vivo more than 80% of
administered 5FU is normally catabolised primarily in the liver, where dihydropyrimidine
dehydrogenase (DPD) is abundantly expressed (this is obviously not the case in our in vitro
experiments).
Modulation strategies, such as co-treatment with leucovorin and methotrexate, have been
developed to increase the anticancer activity of 5FU. Response rates to 5FU in advanced
colorectal cancer have been dramatically improved by combining the drug with OxPt and
irinotecan (See further) [26].
Mechanism of action
After facilitated transport into the cell, 5FU is converted intracellular into several active
metabolites,
that
are
responsible for the
cytotoxic activity of the compound:
fluorodeoxyuridine monophosphate (FdUMP), fluorodeoxyuridine triphosphate (FdUTP) and
fluorouridine triphosphate (FUTP). DPD, converting 5FU into its inactive metabolite
dihydrofluoruracil (DHFU), is the rate-limiting enzyme in 5FU catabolism. See Figure 5 for
an overview.
8
Fig. 5| 5-Fluorouracil metabolism and chemical structure (Left under). Obtained from [26].
Normally, TS catalyses the conversion of deoxyuridine monophosphate (dUMP) to
deoxythymidine monophosphate (dTMP) with 5,10- methylene tetrahydrofolate (CH2THF) as
the methyl donor. This reaction is the only de novo source of thymidylate. In contrast, when
5FU is present, the active metabolite FdUMP binds to the nucleotide-binding site of TS and
forms a stable ternary complex with TS and CH2THF, thereby blocking the access of dUMP
to the nucleotide-binding site and inhibiting dTMP synthesis. This ultimately results in
deoxynucleotide (dNTP) pool imbalance and increased levels of dexouridine triphosphate
(dUTP), both of which cause DNA damage through a disruptive DNA synthesis and repair
[26]. See figure 6 for an overview.
Fig. 6| Mechanism of thymidylate synthase inhibition by 5FU. Obtained from [26].
9
Resistance mechanisms
Resistance is a major obstacle to the success of 5FU-based therapies. Some resistance
mechanisms specific for 5FU are listed below.
•
dTMP can be salvaged from thymidine through the action of thymidine kinase (TK),
thereby alleviating the effects of TS deficiency; this represents a potential resistance
mechanism [26].
•
Treatment with 5FU has been shown to acutely induce TS expression. This induction
seems to be the result of inhibition of a negative-feedback mechanism in which
ligand-free TS protein binds to, and inhibits the translation of TS mRNA. When stably
bound by FdUMP, TS is no longer able to bind to its mRNA and suppress its own
translation, resulting in increased TS protein expression [26].
•
In vitro studies have shown that DPD overexpression in cancer cell lines confers
resistance to 5FU. Furthermore, high levels of DPD mRNA expression in colon
tumours have been shown to correlate with resistance [27].
•
Overexpression of p53 correlates with resistance to 5FU, although in vitro studies
reported that loss of function of p53 also reduces cellular sensitivity to 5FU [28].
•
Expression levels of mRNA’s encoding the multidrug resistance proteins MDR3/4
were found to significantly correlate with 5FU sensitivity [26].
•
Cells possessing MMR defects (i.e. MLH1-), which is often the case in colon cancer
(particularly in HNPCC), and which causes MSI have been found to be 18-fold more
resistant to 5FU than cells with normally functioning MMR [29]. However, the MSI
phenotype has been associated with excellent survival in patients who receive
adjuvant 5FU based chemotherapy. These contradictory findings can possibly be
explained by intrinsic biological differences between MSI+ (e.g. p53 wt) and MSI(e.g. p53 mutated) tumours [26, 30].
DNA microarray analysis of 5FU-responsive genes playing key roles in resistance will
facilitate the identification of new biomarkers and rational drug combinations. Some
molecular biomarkers that predict tumour sensitivity to 5FU have already been identified,
including mRNA and protein expression levels of TS, indicating that patient selection based
on the molecular profile of their individual tumours might help increase the response rate to
5FU treatment.
10
1.3.3. Cetuximab & Bevacizumab
Because conventional cytotoxic agents, including OxPt and 5FU are administered
systemically, they will affect not only tumour cells, but also normal proliferating cells in the
organism, so that side toxicities are major limiting factors to chemotherapy [31]. An ideal
approach would be to identify pathways that are exclusively altered in tumour cells and to
target them selectively (target-driven therapy) however, this is rarely the case. Although,
context-driven therapies, based on cell intrinsic or extrinsic differences that cause tumour
cells to rely on a specific pathway more than normal cells, represents a way to reduce side
effects. Antibodies have been developed to selectively target a tumour cells based on
quantitative differences in the expression of specific surface antigens, e.g. growth factor
receptors [32]. At present, two monoclonal antibodies have been approved by the FDA for use
in metastatic colon cancer:
•
Cetuximab (Erbitux) is a chimeric monoclonal IgG1 antibody that acts by binding to
the extracellular domain of the Epidermal Growth Factor receptor (EGFR, belonging
to the ErbB family), preventing ligand binding and receptor activation, thereby
blocking the signalling downstream of EGFR and resulting in impaired cell growth
and proliferation. Cetuximab also mediates ADCC (Antibody-Dependent Cellular
Cytotoxicity) [31].
•
Bevacizumab (Avastin) is a humanised monoclonal IgG1 antibody that binds the proangiogenic factor VEGF (Vascular Endothelial Growth Factor) and prevents its
binding to specific tyrosine kinase receptors, inhibiting angiogenesis [33].
1.3.4. Combination strategies
To achieve an optimal clinical outcome several combination strategies with synergistic
cytotoxic effects have been developed, hence 4 regimens can be distinguished these days.
Dependent on the grade of the colon cancer (grade I, IIA, IIB, IIIA, IIIB, IIIC or IV) one
regimen can be chosen and variations in chemotherapy dosing and time schedules are
possible. With the aid of these different chemotherapy regimens an pursuance towards an
optimal and fine-tuned balance between toxicity and resistance is possible [34]. The four most
common regimens are the following, each one apparently causing specific side effects:
11
1. Fluoropyrimidine based:
•
LV5FU2: 5FU and Leucovorin (LV). LV or folinic acid is a reduced
folate that is thought to stabilize fluorouracil’s interaction with
thymidylate synthase. This modulation doubles the response rate with
a statistically significant improvement in disease-free and overall
survival of patients with metastatic colon cancer compared with 5FU
alone [35].
•
There are several hospital-specific (dose) modulations.
2. Oxaliplatin based:
•
FOLFOX: 5FU, LV and oxaliplatin [36].
•
Modified FOLFOX and FLOX: Modulations concerning the dose and
way of administering (bolus/infusion).
3. Irinotecan based:
•
FOLFIRI: 5FU, LV and irinotecan. Irinotecan is a semisynthetic
derivative of the natural alkaloid camptothecin and inhibits
topoisomerase I, an enzyme that catalyzes breakage and rejoining of
DNA strands during DNA replication [37].
•
IFL: Modulations concerning the dose and way of administering [34].
4. Antibody based:
•
Cetuximab + irinotecan and/or 5FU [34].
•
Bevacizumab + irinotecan and/or 5FU [33].
2. Colon cancer & Hypoxia
2.1.
Hypoxia
Hypoxia can be defined as a state of reduced O2 availability or decreased O2 partial pressure
below critical thresholds, thus restricting or even abolishing the function of organs, tissues, or
cells. There is a clear evidence that these hypoxic thresholds can vary widely, although an
upper limit of 35 mmHg can be set [38]. Relative low partial oxygen pressure is a common
feature in many solid tumours, including colon carcinomas. Because of the rapid proliferation
of the tumour mass and the limited diffusion distance (i.e. 100-200 µm; dependent on the
vasculature), cancer cells become hypoxic as they outgrow the standard blood supply [39].
Tumour hypoxia is a powerful driving force for malignant progression and has been identified
12
as an adverse prognostic factor, as clinical and preclinical studies have firmly established that
hypoxia is associated with impaired response to both radiotherapy and chemotherapy [38, 40].
This latter effect is due in part to poor perfusion and restricted drug access to hypoxic areas,
but hypoxia-dependent adaptive changes in gene expression probably play the major role in
reduced drug response [41]. Hypoxic stress, induced by a decrease in O2 partial pressure
below 5% (40 mmHg) activates the transcription factor Hypoxia-inducible factor 1 (HIF-1), a
heterodimer composed of an inducible, oxygen-sensitive α subunit and a constitutively
expressed β subunit, that is considered as the master regulator of the hypoxic world (Figure 7)
[42].
Fig. 7| Degradation and activation of the
HIF-1α transcription factor in normoxia
and hypoxia. Obtained from [42].
Fig. 8| Genes that are transcriptionally
activated by HIF-1. Obtained from [43].
Overall, HIF-1 is responsible for the transcription of more than 100 putative genes (HRE’s;
Hypoxia Responsive Elements) in hypoxic circumstances and this occurs not only in
cancerogenesis, but also during normal development and several pathophysiologic conditions.
[42, 43]. See Figure 8 for an overview.
13
The major HIF-1 dependent pathways are involved in the control of:
•
Cellular proliferation: hypoxia induces the expression of growth factors stimulating
cell proliferation, such as PDGF, TGF-α and IGF2 [44].
•
Metabolism: under hypoxic conditions a switch occurs from aerobic metabolism to
anaerobic glycolysis. HIF-1 regulates the expression of enzymes that are necessary for
the glycolysis and glucose transporters (GLUT1 and GLUT3, mediating glucose
uptake by the cells) [45].
•
Angiogenesis: HIF-1 activates angiogenesis, via enhanced expression of the vascular
growth factors ANG2 (angiopoeitin-2) and VEGF, encoded by one of the best known
HIF-1 target genes [46].
•
Apoptosis: under hypoxic conditions apoptosis is induced following HIF-1-dependent
accumulation of p53, which results in selection of cells carrying mutations in p53
and/or other genes involved in apoptosis control, making tumour cells less prone to
drug-induced cell death. In addition, HIF-1 has been shown to negatively regulate the
expression of pro-apoptotic genes in human colon cancer cells, thereby shifting the
balance even further towards cell survival [47].
•
Immortality: under hypoxic circumstances an increase is observed in telomerase
activity, an enzyme essential to sustain the unlimited proliferative potential of tumour
cells [48].
•
pH regulation: hypoxia-induced activation of anaerobic glycolysis and increased
expression of type IX carbonic anhydrase results in production of lactic acid and
carbon dioxide, respectively, both causing intracellular acidification relative to the
extracellular space. This in turn contributes to tumour invasion by activating a number
of proteases dependent on acidic pH (see further). HIF-1 has been shown to regulate
invasive behaviour in HCC (Human colon carcinoma) cells [49]. So hypoxic cells -in
contrast to normal cells- are able to escape and thus survive from a relatively acidic
microenvironment [50].
•
Drug resistance: HIF-1 has been shown to induce mdr1 gene expression (Multidrug
Resistance gene 1 or Glycoprotein-P) enhancing drug efflux [51].
The activation of one or more of these pathways is a substantial advantage for the fast
growing tumour mass, and indeed, overexpression of the HIF-1α subunit has been
demonstrated in colorectal cancers [52].
14
2.2.
HIF-1 structure and regulation
HIF-1 is the most significant and best studied member of a group of hypoxia inducible factors
(HIFs) and is composed of an oxygen-sensitive HIF-1α subunit and a constitutively expressed
HIF-1β subunit, also known as aryl hydrocarbon nuclear translocator (ARNT). The 2 subunits
have a similar domain structure and they contain a basic helix-loop-helix domain, required for
their dimerization and DNA binding, a Per-ARNT-Sim domain (PAS), that also is important
for dimer formation, and a transactivation domain (TAD). The TAD domain of HIF-1α that
can be subdivided in N-TAD and C-TAD, and has been shown to bind the co-activator
proteins p300/CBP, SRC-1 and TIF2, whereas the TAD of HIF-1β appears to be dispensable
for the activity of the HIF-1 complex [53]. The HIF-1α gene promoter contains recognition
sites for several ubiquitous transcriptional activators, such as Sp-1, AP-1, AP-2 and NF-1,
causing the gene to be constitutively expressed; however, in normal cells under normoxic
conditions HIF-1α is undetectable, due to fast protein degradation. HIF-1α is a 826 amino
acid protein with a molecular weight of 120 kDa [50].
2.2.1. Oxygen dependent HIF-1 regulation
Under normoxic conditions, HIF-1α becomes hydroxylated on two proline residues (402 and
564, located in the so-called oxygen dependent degradation domain, ODDD, overlapped by
N-TAD) by a family of prolyl hydroxylases (PHD1-3) [54]. Due to this hydroxylation, the
von Hippel-Lindau protein (pVHL), a recognition component of an E3 ubiquitin ligase,
recognises the HIF-1α subunit, targeting it for polyubiquitylation and subsequent degradation
of HIF-1α by the 26S proteosomal system. Another so called oxygen sensor is Factor
inhibiting HIF-1 (FIH-1), an oxygen-dependent enzyme that hydroxylates Asn803 within the
C-TAD of HIF-1α, disrupting its interaction with the transcriptional co-activators p300 and
CBP. Thus, the two types of metabolic sensors, PHDs and FIH, by controlling both the
destruction and inactivation of HIF-α subunits, ensure full repression of the HIF pathway in
well-oxygenated cells (Figure 9) [42].
15
Fig. 9| Oxygen sensors contribute to the destruction and inactivation of HIF-1α. Obtained from [50].
These 2 oxygen sensors possess a different oxygen affinity hence a fine tuning of
transcription of HREs is possible; PHD has a much lower affinity for oxygen than FIH, so
that PHD activity is decreased under moderate hypoxic conditions and inactive under
complete hypoxia, whereas severe hypoxia is required to inactivate [50]. This observation,
together with the fact that PHDs cause HIF-1α, thereby silencing both N-TAD and C-TAD
genes, whereas FIH only targets the C-TAD of the HIF-1α protein, can explain the
‘bicephalous’ transcriptional nature of HIF-1α and its ability to differentially regulate two sets
of genes (i.e. the N-TAD and C-TAD genes) (Figure 10).
Fig. 10| Working model of two sets of HIF-1 regulated genes. The further away from blood vessels, the
more hypoxic the cells and the higher the extracellular acidity due to the accumulation of lactate and
CO2. Hypoxia also induces the expression of carbonic anhydrase IX (CA IX), which helps to retain a
relatively neutral intracellular pH, furthermore there is an expression of the proapoptotic protein BNIP-3
under moderately hypoxic conditions, but requires acidosis to promote cell death which occurs in
extreme low pO2 conditions. Obtained from [50].
16
An extra fine-tuning of specific gene activation by HIF may result from isoform specificity,
because three isoforms of HIF-α and several splice variants of each exist and increasing
evidence suggests that specific genes may be activated by one or the other or several
isoforms. Further regulation of HIF-1α is ensured by other post-translational modifications
such as: phosphorylation of HIF-1α, which enhances transcriptional activity, and interaction
with heat shock protein 90 (Hsp90), which regulates its stability [42].
Under hypoxic conditions, the above described degradation and inactivation will not happen
and dimerisation between the HIF-1α and HIF-1β subunit occurs. Together they bind to
hypoxia-response elements (HRE’s) throughout the genome, recruiting transcriptional coactivators and upregulating target gene expression [16].
2.2.2. Oxygen independent HIF-1 regulation
Diverse stimuli including growth factors, cytokines, NO or oncogene activation can activate
HIF-1 under normoxic conditions through the phosphatidylinositol 3-kinase (PI3K) and
mitogen-activated protein kinase (MAPK) pathways in a cell-type-specific manner (whereas
oxygen-dependent HIF-1 regulation occurs in every cell type) (Figure 11) [43]. The most
notable growth factors herein are insulin-like growth factor-2 (IGF2) and transforming growth
factor-α (TGF-α), encoded by HIF-1 target genes itself, thus creating an autocrine-signalling
pathway when binding to their cognate receptors (IGF1R and EGFR, respectively) and
activating signal-transduction pathways that lead to HIF-1α expression and cell
proliferation/survival, both crucial for cancer progression [49]. An improved stabilisation of
HIF-1α in normoxic circumstances can be achieved by activation of oncogenes such as Src or
Ras, and/or inactivation of the tumour supressor genes such as PTEN and VHL. All this
results in an substantial increase in translation of the HIF-1α mRNA through phosphorylation
of the eukaryotic translation initiation factor 4E (eIF-4E), which finally results in HIF-1
protein expression, which is particular sensitive to changes of synthesis-velocity because of
its extremely short half-life in normoxic conditions (< 5 min) [50].
17
Fig. 11| Regulation of O2 independent HIF-1 protein synthesis. Obtained from [43].
2.3.
HIF-1 and drug resistance
Besides the features described above, that concur to tumor progression under hypoxic
conditions, HIF-1 or hypoxia can also contribute to the development of drug resistance in
several ways:
•
Downregulation of the pro-apoptotic proteins Bid and Bax [47].
•
Direct upregulation of Bcl-xL [55].
•
Downregulation of DNA repair proteins [56].
•
HIF-1 promotes the formation of an ‘aggressive’ and abnormal vasculature, which has
an negative influence on drug delivery to the tumour.
•
Due to the acidification, some drugs are retained less efficiently inside the cell and/or
are less cytotoxic, impairing their effectiveness [57].
•
Hypoxia is also associated with resistance to X-Ray therapy, as this therapeutic
modality relies on reactive oxygen species that cannot be efficiently produced under
hypoxic conditions [43].
18
Thus, selection by hypoxia may explain the resistance of many solid tumours not only to
hypoxia-induced-apoptosis, but to radio and chemotherapy as well. Furthermore, through this
selection these cells increase their potential for invasion and metastasis, thereby considerably
worsening patient prognosis [38].
Importantly, HIF-1-dependent tumour cell response can be affected by the p53 tumour
suppressor gene status [58]. After prolonged exposure to hypoxic conditions, p53
downregulates HIF-1α and also the transactivating function of HIF-1α is repressed because of
the competition for the co-activator p300 [59]. This implies that p53 deletion or loss of
function mutations, that are extremely common in tumours derived from epithelial tissues,
including colon carcinomas, will also have an impact on HIF-1α activity, promoting
angiogenesis and other pro-tumour activities [58, 60]. Loss of p53 function may also directly
contribute to drug resistance [61].
From these observations, it can be concluded that induction of HIF-1α and HIF-1 activation
may play a major role in the resistance of hypoxic tumour cells to killing by chemotherapy (or
radiation). Thus, it will be important to determine whether HIF-1α alone or trough crosstalk
with other markers such as p53 can be seen as a valid marker for chemoresistance. Based on
these and other considerations, targeting HIF-1α or HIF pathways may represent an attractive
strategy to potentiate the antitumour effects of conventional cytotoxic agents in colon cancer.
3. Modulation of HIF-1
Based on the multiple roles played in tumour progression and tumour drug response, HIF-1
has become an important therapeutic target. Many compounds already known to act on other
cellular mechanisms or signalling pathways (e.g. including the topoisomerase I inhibitor
topotecan, the natural phytoalexin resveratrol and the guanylyl cyclase activator YC-1) have
been shown to affect HIF-1 and others have been developed or are under development that
target HIF-1, either directly or via modulation of HIF-1α levels [62, 63, 64]. In the present
study, we have used the small molecule inhibitor of thioredoxin PMX290 and the locked
nucleic acid antisense oligonucleotide, described below.
19
3.1.
PMX290
PMX290 is a small molecule inhibitor of the
thioredoxin system (Figure 12). The thioredoxin (Trx)
system includes Trx1 and 2, two low molecular weight
proteins containing thiol (SH) that are oxidized while
providing reducing equivalents to target molecules e.g.
ribonucleotide reductase (which is involved in DNA
Fig. 12| Chemical structure of PMX 290 (=AJM290). Obtained from [66].
synthesis), peroxiredoxin (is a cellular antioxidants) and
various transcription factors. Trx’s are then recycled by
Trx reductase (TrxR) in a NADPH (nicotinamide
adenine dinucleotiede phosphate-oxidase)-dependent reaction. Trx1 is the predominant form
and is localised in the cytosol whereas Trx2 has been identified in mitochondria and executes
functions of the electron-transport chain [65]. It is already apparent that Trx-1 is upregulated
in hypoxic regions of solid tumours, where it is hypothesized to cause HIF-1 activation and to
regulate vascular endothelial growth factor levels and hence angiogenesis. Inhibiting Trx-1
function using PMX290 has been shown to impair HIF-1α CAD transcription activity and
DNA binding; therefore, this compound has been used in this study to verify the role of HIF-1
in drug resistance and to sensitize colon cancer cells to the effects of cytotoxic drugs [66].
3.2.
Antisense oligonucleotides
HIF-1α expression was knocked down by transfection with the
antisense oligonucleotide EZN-2968 directed against HIF-1α
mRNA. The antisense sequence is 5’-TGGcaagcatccTGTa-3’
where
lower
case
letters
represent
“natural”
deoxyribonucleotide residues, whereas capital letters represent
nucleotides featuring a Locked Nucleic Acid (LNA) structure,
with the ribose ring is “locked” by a methylene bridge
connecting the 2’-O atom and the 4’-C atom (Figure 13). LNA
structure is ideally suited for Watson-Crick base pairing and
pairing with a complementary nucleotide strand is more rapid
and the resulting duplex exhibits increased thermal stability as
compared
to
native
oligonucleotide
sequences
[67].
Fig. 13| L.N.A.’s with a clearly
visible CH3 bridge between
2’-O and 4’-C atom. Obtained
from [67].
20
The sequence is also relatively resistant to exo-and endonucleases, due to replacement of
phosphodiester with phosphorothioate internucleotide bonds, and exhibits a high target
specificity. EZN-2968 is complementary to nucleotides 1197 to 1212 in the human HIF-1α
mRNA sequence [68].
4. Aims
I. How does hypoxia affect the cellular response to oxaliplatin and 5FU?
To address this question, HCT116 and HT29 cells were grown as monolayers and exposed to
a range of oxaliplatin and 5FU concentrations under normoxic and hypoxic conditions. At this
stage, the role of HIF-1 in tumour cell response to the drug is not directly addressed, as
hypoxia is known to induce a pleiotropic response that depends for the most part on HIF-1
activation, but that can also involve HIF-1-independent mechanisms. In addition to this
model, HCT116 cells were grown as 3D spheroids. In this case hypoxia will develop
spontaneously once the spheroid size exceeds 100-200 µm in diameter, so the response to
5FU will be compared with the response of normoxic cells grown as monolayers. Drug
sensitivity was assessed by the following methods and results are presented as IC50 values or
apoptotic cell percentages:
1. Cell counting (effect on cell growth and survival)
2. Clonogenic assay (effect on the clonogenic potential)
3. FACS analysis of cells stained with propidium iodide (cytotoxic effect)
II. Is HIF-1 activity directly involved in tumour cell response under hypoxic conditions?
To explain the results obtained in aim I, HIF-1 transcriptional activity was determined by
method of HCT116 transfection with a plasmid containing EGFP cDNA under the control of
an hypoxia-responsive promoter (HRP) and subsequent analysis by FACS. HIF-1
transcriptional activity is expressed as EGFP fluorescence and was compared between
HCT116 spheroids and HCT116 monolayers.
21
III. Is it possible to modify colon cancer cell response to OxPt and 5FU by modulating the
expression of HIF-1α or HIF-1 transcriptional activity, independent of oxygen levels?
This question was addressed by setting up different experimental models whereby HIF-1α
expression or HIF-1 transcriptional activity can be positively or negatively modulated,
irrespective of oxygen conditions. To this aim, we have used HIF-1 modulating agents
(PMX290, oligonucleotides and HIF-1αMUT) and this way we were finally able to assess the
altered effect of 5FU and OxPt in a quantitative and qualitative manner. Results are presented
as apoptotic cell percentages, IC50 values and EGFP fluorescences.
IV. Does p53 status significantly influence the response of colon cancer cells to HIF-1
manipulations?
We have proceeded as described in aims I, II and III, but using the HT29 cell line, expressing
a mutant form of p53. The presence of mutant p53 could lead to a different response to OxPt
or 5FU under normoxic and/or hypoxic conditions as compared to p53wt cells and could
interfere with HIF-1 dependent processes. This issue was investigated by comparing the
effects of HIF-1α manipulation (up- or downregulation) in HT29 cells with those observed in
p53 wt-bearing HCT116 cells. The same methods applied to HCT116 were used.
Thus, the general aim of this project is to assess the effects of HIF-1 modulation
through novel HIF modulating factors, on the response of cultured colon cancer cells
to different cytotoxic agents currently used against colon cancer, namely 5FU and
oxaliplatin. Because of the described crosstalk between HIF-1 and p53, cell lines
characterized by different p53 status were used. HIF modulators might represent an
important, novel approach to circumvent chemoresistance in colon cancer.
The research activities described in this project were performed in the laboratory of anticancer
pharmacology, University of Insubria, Busto Arsizio, Italy, where the main research focus is
currently on the molecular mechanisms of tumour cell response (or the lack thereof) to
combinations of conventional cytotoxic drugs with novel targeted antitumour approaches,
using in vitro and in vivo models.
22
Materials & Methods
Cell lines
We have used different human colon cancer cell lines (Obtained from A.T.C.C., Rockville,
MD);
HT29: Human colon cancer cell line isolated from an adult female Caucasian patient.
These cells express an inactive mutant of the p53 tumour suppressor protein (R237H)
bear a mutation in the APC gene, but do not have defects in the mismatch repair
system [69]. They display an epithelial morphology.
HCT116: Human colon cancer cell line derived from an adult male patient. They bear
a p53 wild type gene. The HCT116 cells have an activating mutation in K-Ras and
bear a defect in the mismatch repair system [69]. They appear as spindle shaped cells
with an epithelial morphology.
HCT116/HRP-EGFP cells were obtained from HCT116 cells by transfection with a
plasmid containing the EGFP (enhanced green fluorescent protein) cDNA under the
control of an artificial hypoxia-responsive
promoter (HRP) consisting of five copies of a
35-bp fragment from the HRE of the human
VEGF gene and a human cytomegalovirus
(CMV) minimal promoter (kindly provided by
Dr. Y. Cao) [70]. Figure 14 shows the
structure of this plasmid. The presence of the
neomycin-resistance gene (Neor) allowed
selection of stably transfected clones after 2
Fig. 14| Plasmid used for transfection of
HCT116 cells to obtain HCT116/HRPEGFP cells.
weeks of growth in media containing 500
µg/ml of the G418 antibiotic; the clone with the highest EGFP induction was used.
Cell cultures
Cells were grown in Dulbecco Modified Eagle’s Medium (DMEM) in the case of HCT116,
whereas for HT29 cells we used McCoy’s Medium, both supplemented with 10 % fetal
bovine serum (EuroClone, Italy), 1 % antibiotic mixture (penicillin/streptomycin), 1 % non
23
essential aminoacids and 1 % L-glutamine, all maintained in 37 °C in a humidified
atmosphere. Cells were incubated in either normoxic (48 h 21 % pO2, 5% pCO2, 74% pN2 ) or
where appropriate in hypoxic conditions (24 h normoxia, subsequently 24 h 1% pO2, 5%
pCO2, 94% pN2). This was achieved by placing the cells in a modular incubation chamber
(Billups-Rothenberg Inc., Del Mar, CA, USA), during the last 24 h.
Three-dimensional spheroids
In contrast with monolayer cultures, multicellular spheroids spontaneously develop hypoxic
areas, thereby reproducing more closely the in vivo situation. To obtain three-dimensional
spheroids, HCT116, HCT116/HRP-EGFP and HT29 cells, grown as monolayers, were
detached by trypsinization and subsequently seeded (5 x 103 cells/well) onto 96-well tissue
culture plates coated with 1,5 % agarose to prevent cell attachment. Complete medium was
used, supplemented with an 1 % extra sodium pyruvate to achieve a better growth of the cells
as spheroids. Cells were incubated at 37 °C in a humidified 5% CO2 atmosphere and grown
for 7 days: at the end of this period of time spheroids are formed, reaching an average
diameter of 600 µm and consisting of approximately 5000 cells (see Figure 18 in the Results
section). Seven-day spheroids were used for flow cytometric and cytotoxicity studies.
Drug treatments
For all the experiments, cells were exposed to chemotherapeutic drugs, either 5FU or OxPt,
with or without HIF-1-modulating agents. PMX290 (formerly AJM290, kindly provided by
Prof. M.F. Stevens, University of Nottingham and Pharminox Ltd., UK) was used at various
concentrations for 48 h. For this HIF-1 modulator stock solutions were prepared in DMSO
(10 mM for PMX290); cellular exposure to DMSO never exceeded 0,025 %.
Transfection with antisense oligonucleotides
Cells were transfected with EZN-2968 (described in the Introduction section); a scrambled
oligonucleotide (EZN-3088) with the sequence 5’-CGTcagtatgcgAATc-3’ was used as
control for non sequence-specific effects. Finally, mock-transfection, using the lipofection
reagent without oligonucleotides, was performed on control cells (indicated as MOCK).
HCT116 and HT29 cells were seeded with a concentration of respectively 40.104 cells/ml and
24
15.104 cells/ml and transfection was performed when a 50 % confluence was obtained (after
72 h) in 100 mm/15 mm petri-dishes. After 2 washes with filtered Opti-Mem (Gibco,
Invitrogen) medium without antibiotics and serum, 6 ml of Opti-Mem, containing 5 µg/ml of
lipofection reagent (Lipofectamin2000, Invitrogen), were added to the dishes. After 15’
incubation at 37°C, an additional 1,5 ml of Opti-MEM containing the oligonucleotides at the
final concentrations of 10 and 100 ng/ml were added. After 4 hours of incubation at 37° C,
cells were washed with Opti-MEM and 10 ml of complete medium was added to terminate
the lipofection reaction. Transfected cells were grown for 24 hours, at the end of which they
were detached and used for the different experiments.
Construction of lentiviral vectors (HIF1α-mut)
Lentiviral particles were generated using a transient expression system, composed of (1) the
pCMV∆R8.74 packaging construct, (2) the pMD2.G envelope expression construct and (3) a
task-specific lentiviral vector: the pWPT-GFP transfer vector, for overexpression of a HIF-1α
degradation-resistant mutant cDNA. The plasmids pWPT/GFP contain a green fluorescent
protein (GFP) cDNA under the transcriptional control of an intronless human elongation
factor 1-α (EF1-α-short) promoter. All constructs were kindly provided by Dr. Didier Trono
(School of Life Sciences, Swiss Institute of Technology, Lausanne, Switzerland). The transfer
vector pWPT/HIF-1αMUT/GFP was generated by cloning a 3100-bp fragment containing the
cDNA of a human mutated form of HIF-1α (kindly provided by Dr. Chris Paraskeva,
University of Bristol, UK) into the pWPT/GFP vector. In HIF-1αMUT, Pro 402 and Pro 564
have been replaced by alanine and glycine, respectively; these modifications prevent oxygendependent prolyl hydroxylations, so that the protein is degradation-resistant under normoxic
conditions [71].
Generation of lentiviral particles and target cell infection.
Lentiviral particles pseudotyped with the VSV envelope glycoprotein were produced by cotransfecting 5.106 293FT cells with 40 µg of total plasmid DNA: the i) pCMV∆R8.74, ii)
pMD2.G and iii) pWPT/HIF-1α MUT vector, with the calcium phosphate precipitation
method, as previously described [72]. Transduction experiments were performed in a medium
containing 4 µg/ml polybrene. Viral titration was performed by flow cytometer-counting
GFP-expressing HCT116 cells 48 h after infection. For in vitro mutant-HIF-1α
25
overexpression experiments, 30 % confluent HCT116 cells were infected for 4 h with 10
MOI lentiviral particles; the particle-containing medium was then replaced with fresh medium
and cells were incubated at 37 ° C for 48 h before use.
Assessment of HIF-1α expression by Western blot analysis.
Western blot analysis was carried out to detect the expression of HIF-1α in whole cell lysates,
following normoxic or hypoxic incubation and/or drug treatment. After harvesting, counting
and centrifugation (1300 rpm, 10’, 4°C), cells were washed with 1 ml of Phosphate Buffer
Saline (PBS). After another centrifugation cells were lysed in a buffer containing NaF 25
mM, EDTA 5 mM, sodium pyrophosphate 25mM in TBS 20 mM, pH 7,4, PMSF 2 mM,
Na3VO4 1 mM, phenylarsine oxide 1 mM, 1% v/v NP-40 and 10 % v/v Protease Inhibitor
Cocktail (Sigma), at the concentration of 100 µl per 107 cells. Subsequently, while holding the
samples in ice, a sonication was performed (2 times 10”, cycle = 1, amplitude % = 100).
Then, a centrifugation 12800 rpm for 20’ was performed and the supernatant was collected.
Protein concentration was determined by the BCA assay (Pierce, Italy) and 100 µg of protein
per sample was loaded onto polyacrylamide gels (8%) and separated under denaturing
conditions. Protein bands were then transferred onto Hybond-P membranes (Amersham
Biosciences, Italy) using Bio-Rad, Trans-Blot SD and Western blot analysis was performed
by standard techniques with mouse monoclonal anti-human HIF-1α antibody (BD,
Biosciences; dilution 1:300). Equal loading of the samples was verified by re-probing the
blots with a mouse monoclonal anti-actin antibody (Santa Cruz Biotechnology Inc.; dilution
1:1000). Protein bands were visualised using a peroxidase-conjugated antimouse secondary
antibody (Sigma-Aldrich; dilution 1:4000) and the Supersignal West Femto Maximum
Sensitivity Substrate (Pierce, Italy).
Effects on cell growth
The antiproliferative effects of 5FU and OxPt were assessed based on cell counts. Cells (i.e.
HCT116, HT29, HCT116 WPT and HCT116/HIF-1α MUT cells) (3,5.103/well) were seeded
onto 24-well plates, allowed to attach and grow for 24 h and subsequently exposed to
different 5FU (1-250 µM) and OxPt (0,5-25 µM) concentrations. After 48 h incubation in the
presence of 5FU under normoxic or hypoxic conditions (for the last 24 h), cells were
26
harvested and counted using a Beckman Coulter Z series cell counter (Beckman Coulter,
Fullerton, CA, USA).
Clonogenic assay
To compare the sensitivity of monolayers and spheroids to 5FU or OxPt, the clonogenic assay
was used. This in vitro assay is based on the ability of a single cell to grow into a colony. It
essentially tests every cell in the population for its ability to undergo “unlimited” division and
determines whether a cell has undergone reproductive death. Following 48 h treatment with
5FU or OxPt, spheroids or monolayer cultures were disaggregated or detached, respectively,
using a Trypsin-EDTA solution, counted and 200 cells/well were plated onto six well plates
and allowed to grow for 8 d. At the end of this period, cell colonies were fixed with 95% v/v
methanol for 5’ at RT, and stained with a solution of methylene blue 0,05 % for 45’. Only
colonies consisting of more than 50 cells were scored and were expressed as fractions of the
number of colonies in control wells (Fu). IC50 values were calculated by the median effect
equation [73].
Flow cytometric analysis: HIF-1 activity
HIF-1α activity was assessed in HCT116/HRP-EGFP cells grown as monolayer cultures
(under normoxic or hypoxic conditions) or as spheroids, with or without PMX290 (0,5 µM for
monolayers; 2,5 µM for spheroids) for 48 h or EZN-2968 (see above); the effects of EZN2968 were compared with those obtained in mock-transfected cells and cells transfected with
the scrambled LNA. At the end of the treatment, cells from monolayers were harvested,
resuspended in PBS and immediately analysed by flow cytometry. Spheroids were
disaggregated as described above, resuspended in PBS and analysed. EGFP fluorescence data
were collected and fluorescence intensity was expressed as mean fluorescence channel
(MFC).
Flow cytometric analysis: apoptosis
Induction of apoptotic cell death under different experimental conditions was also evaluated
by flow cytometry. Monolayer cells were exposed for 48 h to 5FU (10 or 100 µM), or to OxPt
(1 and 10 µM) alone or in combination with PMX290 (0,5 µM, last 48 h), under normoxic or
27
hypoxic conditions. The same pattern for HCT116 cells that were transfected wit EZN-2968,
EZN-3088, or with the vehicle (MOCK), and for HCT116 cells that were transduced with the
HIF-1α MUT or WPT vector. Subsequently cells were
harvested
by trypsinisation
(pooling
adherent
and
detached cells), washed in PBS, centrifugated, and fixed in
70 % v/v ethanol ( -20°C). Spheroids were incubated for
48 h in the presence of 5FU (0,5 - 1mM), with or without
PMX290 (2,5 µM), disaggregated by trypsinisation and
processed as described for monolayers. After a further
Fig15|
Monoparametric
histogram
obtained
by
cytofluorometric analysis of PI
coloured samples to determine
the percentage of apoptotic cells
(i.e. the sub G1 peak).
wash with PBS, DNA was stained with 50 µg/ml
propidium iodide (PI) in PBS in the presence of RNAse A
(30 U/ml) at R.T. for at least 30’. All the samples were
analysed with a FACScan flow cytometer (Becton
Dickinson Mountain View, CA, USA), equipped with a 15
mW, 488 nm and an air-cooled argon ion laser. At least 10.000 events were analysed for each
sample and all data were processed using CellQuest software (Becton Dickinson). Fluorescent
emissions of PI were collected through a 575 nm band-pass filter and acquired in log mode
and the percentage of apoptotic cells in each sample was determined based on the sub-G1
peaks detected in monoparametric histograms (Figure 15).
Statistical analysis
Dose-response curves were analysed by non-linear regression analysis using Calcusyn
(Biosoft, Cambridge, MA), which allowed extrapolating drug IC50 values in parental and
transfected cell lines under different experimental conditions. the IC50 is defined as the drug
concentration inducing a 50% decrease in Fu. Student’s t test was used to evaluate the
difference between 5FU/OxPt IC50 values under normoxic versus hypoxic conditions in
monolayer cultures and in monolayers versus 3D cultures. All the other data were analysed by
ANOVA and Bonferroni’s test for multiple comparisons using Prism 4.03 (GraphPad
Software Inc., San Diego, CA, USA).
28
Results
1. Effect of hypoxia on HIF-1α expression
HC T 116
N
H
HT 29
N
H
Fig. 16| Protein levels of HIF-1α in cell lines
HCT116 and HT29 in normoxic conditions (N)
or hypoxic conditions (H).
Figure 16 shows HIF-1α protein levels in HCT116 and HT29, as assessed by western blot
following 24 h incubation under normoxic or hypoxic conditions. An immunoreactive band at
120 kDa (corresponding to the molecular weight of HIF-1α) is present in both cell lines
maintained under hypoxic conditions, whereas under normoxic conditions the protein is
undetectable, due to its rapid degradation in the presence of oxygen.
2. Effect of hypoxia on HIF-1α activity
Figure 17 shows the results of the flow cytometric analysis performed on HCT116 HRPEGFP cells, in which EGFP expression is regulated by a promoter responsive to HIF-1. In
this cells, the intensity of the
fluorescence emitted by EGFP is directly related to the
transcriptional activity of HIF-1α and increases significantly following 24h incubation under
Cell-number
hypoxic conditions.
Fig. 17| FACS graph representing the
HIF-1α activity in HCT116 HRP-EGFP
cells in normoxia (N) and hypoxia (H). An
increase in HIF-1α activity is clearly
visible
in
hypoxic
circumstances
(i.e. a right shift of the peak).
EGFP-fluorescence
29
3. Effect of hypoxia on cellular response to 5FU and OxPt
3.1.
Determination of IC50 values for colon cancer cells grown as
monolayers
HCT116 and HT29 cells were treated for 48 h with increasing concentrations of 5FU (1-250
µM) or OxPt (0,5-25 µM) under either normoxic or hypoxic conditions (during the last 24h)
and vital cells were counted to obtain IC50 values (Table 1).
Table 1| Effect of 5FU and OxPt (48h) on the growth of HCT116 and HT29 cells as monolayers
under hypoxic or normoxic conditions. Mean IC50 values (Mean ± S.E., n = 3). Grey boxes: p < 0,05
versus normoxic values. Green marked values: p < 0,05 versus HT29 values (in same oxygen
condition and for the same drug).
HCT116
HT29
normoxia
hypoxia
normoxia
hypoxia
5FU
7,15 ± 2,69 µM
13,17 ± 2,57 µM
37,4 ± 3,29 µM
95,91 ± 24,66 µM
OxPt
2,20 ± 0,10 µM
3,07 ± 0,83 µM
0,91 ± 0,07 µM
2,20 ± 0,80 µM
HCT116 cells are less responsive to the cytotoxic effects of 5FU under hypoxic conditions
versus normoxic conditions, as indicated by the nearly twofold value of IC50 whereas no
significant differences can be observed in the response to OxPt. In contrast, hypoxic HT29
cells are significantly less sensitive to both drugs as compared with the results obtained under
normoxic conditions. In addition, the cell lines themselves displayed significant differences in
every case, with the exception of OxPt-treated hypoxic cells.
3.2.
Determination of HIF-1 activity for HCT116 cells grown as
spheroids versus monolayers
Three-dimensional cultures were obtained from HCT116 and HCT116 HRP/EGFP cells.
Following 7-day incubation in non-adherent conditions, spheroids are formed, with diameters
ranging from 500 to 600 µm (Figure 18). These experiments were not performed on HT29
cells, due to the low transfection efficiency obtained in this cell line.
30
Fig. 18| Seven-day spheroids obtained from HCT116 HRP/EGFP cells (Right; visualised by confocal
microscopy, merged fluorescence and clear field images, with visible HIF-1 activity; more obvious in
the centre of the spheroid) and HCT116 cells (Right; visualised by phase contrast microscopy).
Flow cytometric analysis of cells from HCT116 HRP/EGFP spheroids shows that
fluorescence intensity is significantly higher in these cells than in monolayer cells, indicating
higher HIF1- activity (Figure 19).
Fig. 19| HIF-1 transcriptional activity in HCT116 HRP/EGFP monolayers (grey outline) or 3Dspheroids (black peak). There is significantly higher fluorescence intensity in spheroids, indicating
a higher HIF-1 activity.
3.3.
Determination of IC50 values for HCT116 cells grown as spheroids
In agreement with the results obtained in 3.1, 5FU is significantly less potent in inhibiting the
clonogenic potential of cells derived from HCT116 spheroids than these derived from
HCT116 monolayers. Table 2 shows the IC50 values obtained with the clonogenic assay.
31
Table 2| Mean IC50 values (Mean ± S.E., n = 3). The HCT116 cells are treated for 48h with 5FU in
either the context of spheroids or monolayers. Grey box: p < 0,001 versus monolayer values.
HCT116
5FU
3.4.
monolayer
spheroid
12,26 ± 4.76
132.76 ± 41.04
Determination of apoptosis
Figure 20 present histograms of the percentage of apoptotic cells, normalized for the value
obtained in control samples (T/C: drug Treatment/Control) for HCT116 and HT29 cells in
either hypoxic or normoxic conditions. While OxPt induces apoptosis in HCT116 cells in a
concentration-dependent fashion, no significant differences are detectable between normoxic
and hypoxic conditions (Figure 20a), in agreement with the results presented in Table 1.
Fig. 20a| Percentages of apoptotic cells (Treatment/Control) in HCT116 cells treated with Oxaliplatin
(1 and 10 µM) for 48 h in either hypoxic or normoxic conditions.
In contrast, when HCT116 were treated with 5FU, an increase in apoptotic cells was only
observed following exposure to a 100 µM concentration and this was significantly lower
under hypoxic versus normoxic conditions (Figure 20b).
In HT29 cells, OxPt induces a concentration-dependent increase in apoptotic cells under
normoxic conditions, whereas no effect is observed in hypoxic cells (Figure 20c).
In contrast, no effect is observed, in terms of induction of apoptosis, when HT29 cells are
exposed to 5FU, irrespective of oxygen levels (Figure 20d).
32
***
µM
Fig. 20b| Percentages of apoptotic cells (Treatment/Control) in HCT116 cells treated with 5FU (10
and 100 µM) for 48 h under normoxic or hypoxic conditions. Mean ± SE, n ≥ 3; *** p < 0,001.
Fig. 20c| Percentages of apoptotic cells (Treatment/Control) in HT29 cells treated with OxPt (1 and
10 µM) for 48 h in either hypoxic or normoxic conditions. Mean ± S.E., n ≥ 3. *** p < 0,001
versus normoxic value.
µM
Fig. 20d| Percentages of apoptotic cells (Treatment/Control) in HT29 cells treated with 5FU (10
and 100 µM) for 48 h in either hypoxic or normoxic conditions. Mean ± S.E., n ≥ 3.
33
4. Effect of modulation of HIF-1α on the response of colon cancer
cells to 5FU and OxPt
To demonstrate the crucial role of HIF-1 in the decreased responsiveness of colon cancer cells
to chemotherapy under hypoxic conditions, different agents were used to downregulate HIF1α expression and/or HIF-1 transcriptional activity:
1. PMX290, an inhibitor of the thioredoxin system (TRX), that has been shown to
inhibit HIF-1transcriptional activity [66].
2. EZN2968, a LNA antisense oligonucleotide directly targeting the HIF-1α mRNA [68].
In addition, to confirm the ability of HIF to modulate the cellular response to chemotherapy,
HIF-1α was also upregulated, by infecting the cells with a lentiviral vector encoding a
degradation-resistant HIF-1α mutant form.
4.1. Effect of PMX290 treatment on HIF-1 activity and response to 5FU in
HCT116 cells
Figure 21a shows the effect of 48 h exposure to PMX 0,5 µM under hypoxic conditions on the
fluorescence intensity emitted by HCT116/HRP-EGFP cells; the results obtained for these
cells (C) are compared with those obtained for untreated cells under either normoxic (A) or
hypoxic (B) conditions. Whereas the peak of fluorescence of the hypoxic control (B) is
shifted towards right in comparison with the normoxic conditions (A), treatment with PMX
(C) indeed reduces this shift, i.e. HIF-1 transcriptional activity.
A
C
B
Fig 21a| Effect of PMX290 on HIF-1
transcriptional activity in HCT116/HRP-EGFP
cells. Peak A: normoxic untreated control. Peak
B: hypoxic untreated control. Peak C: PMX290
(0.5 µM for 48 h) under hypoxic conditions.
34
Figure 21b shows the percentage of apoptotic cells, obtained by flow cytometric analysis of
HCT116 cells exposed to 5FU and PMX290 under normoxic (Left) and hypoxic (Right)
conditions. The results of these experiments confirm that the cells are more resistant to 5FU
under hypoxic versus normoxic conditions and indicate that treatment with PMX290 alone
increases the percentage of apoptotic cells versus untreated controls, under both normoxia and
hypoxia. When the two agents are used in combination, PMX290 also increases the
percentage of apoptotic cells over the values observed for 5FU alone, and, most importantly,
it significantly restores the ability of 5FU to induce apoptosis in hypoxic cells, almost to the
extent observed in normoxia.
Fig. 21b| Percentages of apoptotic HCT116 cells (Treatment/Control) treated with 5FU 100 µM for 48
h with or without PMX290 (0.5 µM) and incubated in normoxic (Left) or hypoxic conditions (Right).
Mean ± S.E., n ≥ 3; ** p < 0,01 versus no PMX; * p < 0,05 versus no PMX; *** p < 0,001 versus no
PMX.
4.2. Effect of PMX290 treatment on the response of HT29 cells to OxPt and
5FU
The histograms in figure 22 represent the percentage of apoptotic HT29 cells following
treatment with OxPt (1 or 10 µM) for 48 h, with or without PMX290 (0,5 µM), under either
hypoxia or normoxia, in comparison with control (untreated) cells. In contrast to HCT116
cells, HT29 cells do not significantly respond to PMX290 alone, and PMX is unable to
modify the response of these cells to OxPt under normoxic conditions; however, treatment of
hypoxic cells with PMX290 restores the response as achieved in normoxic levels concerning
the 10 µM OxPt treatment.
35
Fig 22| Percentages of apoptotic HT29 cells (Treatment/Control) treated with OxPt 1 or 10 µM for 48
h with or without PMX290 0.5 µM and incubated in normoxia (Left) or in hypoxia (Right). Mean ±
S.E. ≥ 2; *** p < 0,001 versus no PMX.
Similar results are also observed when PMX290 is combined with 5FU (Figure 23).
Fig 23| Percentages of apoptotic HT29 cells (Treatment/Control) following treatment with 5FU 10 or
100 µM for 48 h with or without PMX290 0.5 µM and incubated in normoxia (Left) or in hypoxia
(Right). Mean ± S.E., n ≥ 2; * p < 0,05 versus no PMX.
4.3. Effect of PMX290 treatment on HIF-1 activity and apoptotic response to
5FU in HCT116 cells grown as spheroids
Treatment of HCT116-HRP cells grown as spheroids with PMX290 (2,5 µM for 48 h)
induces a significant reduction in HIF-1 activity. In the histogram (Figure 24), where two cell
36
subpopulations are visible, indicating that a significant fraction of cells has shifted towards
lower fluorescence intensities thus has a significant decrease in HIF-1 activity.
EGFP-FLUORESCENCE
Fig. 24| Flow cytometric analysis of HCT116/HRP-EGFP spheroids without PMX290 treatment
(grey outline) and HCT116/HRP-EGFP spheroids treated with PMX290 (black outline).
Flow cytometric analysis of apoptotic cells from spheroids exposed to 5FU with or without
PMX290 (Figure 25), shows that combined exposure to the quinol (2,5 µM) causes a
significant increase in apoptosis induction by 5FU (0,5 - 1 µM for 48 h) versus exposure to
5FU alone.
no PMX290
PMX290 2.5 µM
% Apoptotic cells
15
*
*
10
5
0
0
0.5
1.0
µM 5FU
Fig. 25| 5FU-induced apoptosis in HCT116/HRP-EGFP spheroid-derived cells. Mean ± S.E.,
n = 3; * p < 0.05 versus cells from spheroids treated with 5FU alone. Empty bars: no PMX290;
filled bars: 2.5 mM PMX290 for 48h.
37
5. Effect of EZN-2968 treatment on HIF-1α expression, HIF-1
activity and apoptotic response to 5FU in HCT116 cells
Figure 26 shows that EZN2968 inhibits HIF-1α expression in HCT116 cells following 24h
incubation under hypoxic conditions. A scrambled (SCR) LNA oligonucleotide, containing
the
same
nucleotides
NORM
SCR
2968
in
random
sequence,
was
used
as
a
negative
control.
HYPO
SCR
2968
Fig. 26| Western blot of HCT116 cell lysates that
were transfected with the A.S.O. EZN2968 10 nM
in normoxic (Left) and hypoxic (Right) conditions.
SCR = scrambled sequence that functions as a
negative control.
Figure 27 shows the results of the flow cytometric analysis of HIF-1α activity in HCT116HRP cells transfected with EZN-2969 or with the scrambled oligonucleotide: under normoxic
conditions no significant difference can be seen between the fluorescent signals emitted by
cells transfected with the two oligonucleotides. However, under hypoxic conditions the
fluorescence detected in the SCR-transfected cells is significantly higher as compared with
FLUORESCENCE
cells transfected with EZN-2968.
Fig. 27| Effect of EZN-2968 (HIF -) on HIF-1 activity in HCT116/HRP-EGFP cells under either
normoxic or hypoxic conditions. Mean ± S.E., n ≥ 2; * p < 0,05 versus normoxia and SCR;
° p < 0,05 versus normoxia.
Figure 28 shows the apoptotic effect of 5FU (10 and 100 µM for 48 h) on HCT116 cells
(Treatment/Control)
transfected
with
EZN-2968
or
with
the
scrambled
control
38
oligonucleotide under normoxic or hypoxic conditions (for the last 24h). Mock-transfected
cells, i.e. HCT116 cells treated only with the transfection reagent without oligonucleotides,
were also used as additional controls (not shown in the figure). The results indicate that both
mock-transfected cells and cells transfected with the scrambled oligonucleotide are less
responsive to 5FU treatment under hypoxic conditions, whereas cells transfected with
EZN2968 exhibit enhanced sensitivity in response to 5FU, as indicated by a significant
increase in the percentage of apoptotic cells as compared to cells transfected with the control
oligonucleotide under the same conditions.
**
Fig. 28| Induction of apoptosis (Treatment/Control) by 5FU (10 or 100 µM for 48 h) in HCT116
cells transfected with either EZN2968 (hif-) or with the scrambled oligonucleotide (scrambled)
and incubated in either hypoxia or normoxia. Mean ± S.E., n ≥ 3; ** p < 0,01 versus normoxia.
6. Effects of the expression of a degradation-resistant variant of
HIF-1α on the response of HCT116 cells to 5FU
To confirm that HIF-1 is indeed responsible for the observed diminished response of HCT116
cells to 5FU under hypoxic conditions, HCT116 cells were infected with lentiviral vectors
encoding a mutant, degradation-resistant variant of HIF-1α. The cell line obtained, called
HCT116/HIF-1α MUT, was compared with HCT116 cells transduced with the control vector
WPT and showed a marked increase in HIF-1α protein levels under normoxic conditions
(Figure 29a).
39
IF
1α
M
UT
H
ve
ct
or
HIF-1α
Fig. 29a|
Effect of HCT 116 cell transduction
with the HIF-1αMUT or WPT (control) lentiviral
vectors under normoxic conditions on HIF-1α protein
levels.
In Table 3 the IC50 values obtained for the two cell lines following treatment with 5FU are
reported, indicating that increased expression of mutant HIF-1α inhibits the cytotoxic activity
of 5FU. HCT116/HIF-1α MUT cells are also refractory to the proapoptotic effect of 5FU, as
indicated by the observation that even relatively high drug concentrations (i.e. 250 – 500 µM)
do not induce a significant increase in apoptotic cells, in contrast with HCT116 WPT cells,
that undergo apoptosis in a concentration-dependent fashion (Figure 29b).
Table 3| IC50 values obtained following 48 h exposure to 5FU under normoxic conditions. Mean ±
S.E., n ≥ 3; p < 0,001.
Normoxia
5FU
HCT116/HIF-1α MUT
HCT116 WPT
658,58 ± 62,2 µM
6,55 ± 1,46 µM
Fig 29b|
Effect of overexpression of a degradation-resistant HIF-1α form on 5FU induced
apoptosis. Mean ± S.E., n ≥ 3; * p < 0.05 versus untreated HCT116/WPT cells; ** p < 0.001 versus
untreated HCT116/WPT cells; ° p < 0.01 versus HCT116/WPT cells at the same 5FU concentration.
40
Discussion
We utilised two different cell lines each with specific characteristics and two
chemotherapeutic drugs, OxPt and 5FU, both used in the first-line treatment of advanced
colorectal cancer. Furthermore HIF-1 modulating agents (potential future drugs) were added
to further clarify the importance of HIF-1 in the obtained results.
5-Fluorouracil
Our data shows that HCT116 cells are more sensitive to 5FU than HT29 cells (for both
oxygen conditions significantly, regarding Table 1), and both become more resistant when
they are incubated under hypoxic conditions.
Why are HT29 cells more resistant to 5FU than HCT116 cells? We know that these
cell lines differ in some important aspects potentially involved in cell response to 5FU. i)
HCT116 cells express a functional p53 protein, whereas HT29 cells carry an inactivating
R237H mutation [69]. Lack of functional p53 protein in HT29 cells may negatively affect
their ability to undergo apoptosis following 5FU treatment, due to reduced expression of proapoptotic proteins (e.g. Bax). This hypothesis is supported by the findings of Boyer et al.,
who demonstrated a five-fold lower IC50 for 5FU in HCT116 p53 wild-type (p53+/+) cells as
compared to an isogenic p53-null (p53-/-) cell line [74]. ii) Peters GJ et al. have observed that
p53, or its lack, can affect the feedback mechanism regulating TS protein synthesis following
5FU treatment, whereby free TS can repress the translation of its own mRNA, whereas
binding to 5FU causes de-repression of translation. The finding that colon carcinoma cells
expressing a mutant (mt) p53 exhibit a higher induction of TS than p53-proficient suggests
that p53 might participate to the downregulation of TS at this level; as TS is the major target
for 5FU activity, this effect might thus contribute to the observed differences in response to
5FU between p53+/+ and p53-/- cells cells [75]. iii) Concerning the number of tandem repeats
of a 28 bp sequence in the enhancer region at the 5’UTR of the TS gene, a well known
polymorphism leading to different levels of TS, both HCT116 and HT29 cells have a
homozygous wildtype genotype (2R/2R) [76]. However, in spite of this common genotypic
feature, Nief et al. have demonstrated that TS expression is higher in HCT116 than in HT29
cells (in a 23,4 : 15,1 ratio), and the ratio between TS activity in the two cell lines is even
higher (39,6 : 2,3 respectively), indicating that other factors are involved in determining this
divergence [77]. iv) In this regard, HCT116 and HT29 cells have been shown to differ for
41
another common polymorphism, consisting in a 6 bp deletion in the 3’UTR of the TS gene,
which has been associated with reduced stability and translation of the TS mRNA. In contrast
with HT29 cells, HCT116 cells exhibit this deletion in homozygous form, and this might
contribute to the higher 5FU sensitivity exhibited by this cell line [77]. v) In addition,
HCT116 cells are known to express undetectable levels of DPD, the major 5FU-inactivating
enzyme [78], and clearly detectable levels of thymidine phosphorylase (TP), one of the
enzymes concurring to anabolic activation of 5FU; these two facts could contribute to the
higher sensitivity to 5FU exhibited by this cell line, at least in comparison with HT29 cells,
which only express low levels of TP activity [71], while exhibiting measurable levels of DPD
expression and activity [79].
All the 5FU-sensitivity modifying properties above described (i-v) do not seem to
have a significant impact on the effect of hypoxia on drug response. The mechanism
underlying decreased drug response under hypoxia is very likely multifactorial, involving
both HIF-1-dependent and -independent effects. In our experimental models, HIF-1 clearly
makes a major contribution to the resistance of hypoxic HCT116 cells to 5FU as shown by
different orders of evidence. First, incubation under hypoxic conditions leads to accumulation
of HIF-1α protein and to an increase in HIF-1 transcriptional activity, as assessed in HCT116
cells stably transfected with a reporter plasmid expressing EGFP under the control of multiple
copies of the HRE derived from the human VEGF gene promoter. Second, HIF-1α
knockdown by transfection of HCT116 cells with the antisense oligonucleotide EZN-2968,
that specifically targets HIF-α mRNA, partially prevents the development of resistance to
5FU under hypoxic conditions. Finally, infection of HCT116 cells with a lentiviral vector
encoding a degradation-resistant form of HIF-1α, lacking the two proline residues that are
crucial for HIF-1α oxygen-dependent hydroxylation and subsequent degradation, leads to
resistance to 5FU under normoxic conditions. In HT29 cells, hypoxia-induced resistance to
the cytotoxic effect of 5FU is also associated with HIF-1α accumulation; however, due to the
low transfection efficiencies obtained with this cell line for both oligonucleotides and
lentiviral vector, it was impossible to assess HIF-1 activity. The observation that treatment
with PMX290 (which inhibits HIF-1 activity, but not HIF-1α accumulation, see below)
partially restores the sensitivity of HT29 cells to 5FU indirectly supports the hypothesis that
increased HIF-1 activity also contributes to 5FU resistance in this cell line.
Why are hypoxic HCT116 and HT29 cells less responsive than normoxic cells to
5FU? Hypoxic cells have the propensity to acquire genetic abnormalities that confer a ‘more
malignant’ phenotype, such as increased expression of the drug exporting proteins MDR1 or
42
BRCP, or proteins that participate in thiol-mediated drug detoxication [51, 80, 81]. However
5FU is not a known substrate for either transporter, and therefore this mechanism is not likely
to play a major role in 5FU resistance. Impairment of mismatch repair activity could be
involved in hypoxia-induced resistance to a number of agents requiring activation of this
DNA repair system to effectively generate a death signal, including 5FU [82, 83]. We know
that HCT116 cells are intrinsically defective in mismatch repair due to a homozygous
mutation in hMLH1, hence they exhibit MSI; however, this is not the case for HT29 cells. In
addition, Meyers et al. have demonstrated that following 5FU treatment there are no
significant differences in apoptosis between MMR-deficient cells (i.e. HCT116) in
comparison with isogenic MMR-proficient cells (i.e. HCT116 3-6) [29]. Thus, as confirmed
by several studies, we cannot consider the status of MMR genes as a major factor determining
the response to 5FU in our cell lines [84]. The most likely hypothesis to explain the reduced
cytotoxicity of 5FU in hypoxic HCT116 and HT29 cells involves critical alterations in the
levels of pro- and antiapoptotic factors. Increasing evidence indicates that HIF-1 activation
modulates the expression of pro- and antiapoptotic genes; for instance, significantly increased
levels of survivin and Bcl-xL, and decreased levels of Bid and Bax were reported [85, 55, 47].
An imbalance between pro-and antiapoptotic signals favouring cell survival could very well
account for a reduced response to a wide range of anticancer agents and could explain the
observed 5FU resistance under hypoxic conditions.
Having established the role of HIF-1 in 5FU resistance, we assessed the ability of
PMX290 to restore the response of our cell lines to 5FU. PMX290 has been shown to
decrease HIF-1α expression and/or HIF-1 activity due to its ability to inhibit Trx-1, a positive
modulator of HIF-1 [66]. Our results indicate that PMX290 decreases HIF-1 activity in
HCT116HRP cells grown both as monolayers under hypoxic conditions and as threedimensional spheroids at normoxic pO2 values. Interestingly, combining PMX290 with 5FU
significantly enhances the antiproliferative effect of 5FU in both cell lines tested (as assessed
by cytotoxicity assays) and significantly increases the percentage of apoptotic cells as
compared with 5FU alone. The results obtained on spheroids confirm the results obtained in
monolayers. Flow cytometric analysis of spheroids obtained from HCT116 HRP/EGFP cells,
indicates that hypoxia develops spontaneously within the cell mass, leading to HIF-1
activation (increased fluorescence is visible at the centre of the spheroid, see Figure 18). The
presence of hypoxic regions causes a significant decrease in the antiproliferative effect of
5FU, as assessed by the clonogenic assays on disaggregated spheroids in comparison with
monolayer cells. However, it should not be neglected that limited diffusion of the drugs into
43
the centre of the spheroid could also contribute to the 10 fold greater IC50 value obtained for
spheroids.
The observation that subtoxic concentrations of PMX290 simultaneously reduce HIF-1
activity and enhance cell response to 5FU indicates that HIF-1 inhibition could be an effective
approach to cell sensitisation in the clinical management of hypoxic tumours.
Oxaliplatin
Significant differences in cell response between HCT116 and HT29 under normoxic
conditions were also observed with OxPt (Table 1); however, in contrast with 5FU, in this
case HT29 cells appear to be more sensitive than HCT116 to the platinum derivative. What
possible mechanism could account for the observed difference?
Reports on the role played by p53 are highly controversial one. Arango et al. found that the
mutational status of the tumour suppressor gene could not predict the apoptotic response to 10
µM OxPt treatment in a panel of 30 different colorectal cancer cell lines, suggesting that
additional factors modulate sensitivity to this agent [86]. The bulk of clinical data regarding
p53 status and sensitivity to platinum compounds has focused on the first generation
compound cisplatin. However, Howels et al. reported that exposure of wild-type p53 HCT116
cells to OxPt results in increased levels of p53, with subsequent upregulation of the p21 gene,
a major target for p53 transcriptional activity [87]. Increased p53 levels can enhance apoptosis
trough different mechanisms, including Bax upregulation, but they also induce cell cycle
arrest, which favors DNA damage repair and could cause failure of OxPt treatment in p53+/+
HCT116 cells, whereas cannot occur in p53 defective HT29 cells. Moreover, Arango et al.
have reported that the concentration of OxPt necessary to cause a 50% growth inhibition after
72 h of exposure was four-fold higher in HCT116 p53-/- cells compared to parental isogenic
HCT116 cells, which seems to rule out a protective antiapoptotic role of p53 in this cell line.
[86]. However, these findings in HCT116 cells cannot be extended to HT29 cells. In fact,
Howells et al. have pointed out that survival signalling in response to OxPt exhibits cell line
specificity; HCT116 cells require p53 for the OxPt induced apoptotic response, whereas
mutated p53 in HT29 cells does not prevent this chemotherapy-induced apoptosis [87].
Another aspect differentiating HCT116 from HT29 cells is MMR status, as HCT116 cells
are deficient in this regard, whereas HT29 cells are not [29]. MMR status is definitely
considered as a major determinant of cell response to cisplatin and carboplatin, but for OxPt
44
this has been shown not to be the case; loss of MMR enzyme activity has been demonstrated
not to be involved in OxPt resistance [88].
A third possible difference is the tetraploïd character of HT29 cells as this might require
more OxPt to reach the same degree of genotoxic damage in comparison with 2n cells such as
HCT116; however, our results seem to point in the opposite direction, so while this feature
can ultimately affect the outcome of OxPt treatment, it does not help explain the observed
difference between HCT116 and HT29 cells.
As already observed for 5FU, HT29 incubated under hypoxic conditions are more
resistant than their normoxic counterparts to the cytotoxic action of OxPt; however, no
significant differences were observed for HCT116 cells, so that we can conclude that hypoxia
only plays a minor role in HCT116 resistance against OxPt.
As already noted in the case of 5FU, the mechanism underlying decreased OxPt response in
hypoxia HT29 cells is very likely multifactorial, involving both HIF-1-dependent and
-independent effects. Unfortunately, direct measures of HIF-1 activity in HT29 cells were not
possible; however, the hypothesis that HIF-1 plays a role in the reduced response to OxPt is
supported by the results following combined treatment with PMX290, as this HIF-1-inhibiting
agent significantly reduces the resistant state observed in hypoxic HT29 cells.
Further investigations will be necessary to identify genes that are predictive of cell
response to OxPt and that may be affected by the HIF-1 transcription factor. At this moment
we can note for example that high expression levels of the nucleotide excision repair gene
ERCC1 predict poor response to OxPt [74]. Another example is the ATP-binding cassette
half-transporter BCRP/ATP-binding cassette G2; Boyer et al. demonstrated overexpression of
this efflux pump in both p53+/+ and p53-/- OxPt resistant cell lines [74]. Cisplatin is not
known to act as a substrate for BCRP, and no clear evidence has been reported that OxPt may
act as such, but given the structural differences, it might be. Interestingly, though, BCRP
expression has been shown to be upregulated under hypoxic conditions through a HIF-1dependent mechanism, so the role of this transporter in hypoxia-induced resistance is
certainly worth investigating [80]. Expression levels of other transport proteins, such as the
copper transporter hCtr1 and the organic cation transporter hOCT1 (these are certainly OxPt
transporters), have been correlated to cytotoxicity of OxPt in HCT116 cells and can also
represent additional factors determining increased OxPt resistance in this cell line in
comparison with HT29 cells [89].
45
In summary, while OxPt resistance is generally attributed to decreased cellular uptake
and/or increased efflux (i.e. intracellular accumulation is detrimental) and to efficient repair of
DNA-Pt adducts, it is not clear how these factors might contribute to hypoxia in OxPt
resistance [89, 90]. It may be important to emphasise that an highly selective hypoxic
environment can facilitate mutational events (such as p53 loss of function) and adaptive
responses (increased resistance) that can decrease susceptibility to both hypoxia and OxPt
induced cell death [20, 91]
General conclusion
To conclude, our data indicate that 1) HIF-1 activation may play an important role in the
response of colorectal cancer cell to 5FU and OxPt, and is in fact a major determinant of
hypoxia-induced drug resistance; 2) agents inhibiting HIF-1 could help improve the
therapeutic index of cytotoxic agents currently used in the clinical management of colorectal
cancer; and 3) the genetic makeup of the individual tumour may affect its response to hypoxia
and its susceptibility to hypoxia-induced resistance, as demonstrated by the different response
observed in HCT116 and HT29 cells, once more underscoring the importance of tumor
genetic profiling in directing the choice of therapeutic options.
46
References
1. Pisani P., Bray F. and Parkin M. (2002). Estimates of the world-wide prevalence of cancer for
25 sites in the adult population. Int. J. Cancer 97: 72-81.
2. Parkin M., Bray F., Ferlay J. and Pisani P. (2001). Estimating the world cancer burden:
globocan 2000. Int. J. Cancer 94: 153-156.
3. Belgian Cancer Registry. http://www.kankerregister.be/
4. Parkin M., Bray F., Ferlay J. and Pisani P. (2005). Global Cancer Statistics, 2002. CA. Cancer.
J. Clin. 55: 74-108.
5. Labianca R., Beretta G., Gatta G., Braud F. and Wils J. (2004). Colon cancer. Crit. Rev.
Oncol. Hematol. 51: 145-170.
6. Johnson I. and Lund E. (2007). Review article: nutrition, obesity and colorectal cancer.
Aliment. Pharamcol. Ther. 26: 161-181.
7. Bingham S. (2006). The fibre-folate debate in colo-rectal cancer. Proc. Nutr. Soc. 65: 19-23.
8. De la Chapelle A. (2004). Genetic predisposition to colorectal cancer. Nat. Rev. Cancer 4:
769-779.
9. Galiatsatos P. and Foulkes W. (2006). Familial adenomatous polyposis. Am. J. Gastroenterol.
101: 385-398.
10. Näthke I. (2004). APC at a glance. J. Cell. Sci. 117: 4873-4875.
11. Darnell J. (2002). Transcription factors as targets for cancer therapy. Nature 2: 740-750.
12. Goss K. and Groden J. (2000). Biology of the adenomatous polyposis coli tumor suppressor. J.
Clin. Oncol. 18: 1967-1979.
13. Jiricny J. and Nystrom-Lahti M. (2000). Mismatch repair defects in cancer. Curr. Opin. Genet.
Dev. 10: 157-161.
14. Fishel R., Lescoe M. and Rao M. (1993). The human mutator gene homolog MSH2 and its
association with hereditary non polyposis colon cancer. Cell 78: 539-542.
15. Vogelstein B. (1996). Lessons from hereditary colorectal cancer. Cell 87: 159-170.
16. Weinberg R. (2007). Tumour suppressor genes. In: The biology of cancer. New York: Garland
Science.
17. Knudson AG. (2001). Two genetic hits (more or less) to cancer. Nat. Rev. Cancer 1: 157-162.
18. Fearon ER. and Vogelstein B. (1990). A genetic model for colorectal tumorigenesis. Cell 61:
759-767.
19. Raymond E., Faivre S., Chaney S., Woynarowski J. and Cvitkovic E. (2002). Cellular and
Molecular Pharmacology of Oxaliplatin. Mol. Cancer Ther. 1: 227-235.
20. Wang D. and Lippard S. (2005). Cellular processing of Platinum anticancer drugs. Nature 4:
307-320.
21. Stordal B., Pavlakis N. and Davey R. (2007). Oxaliplatin for the treatment of cisplatinresistant cancer: A systematic review. Cancer Treat. Rev. 33: 247-357.
22. Chaney S., Campbell S., Bassett E. and Wu Y. (2005). Recognition and processing of
cisplatin- and oxaliplatin- DNA adducts. Crit. Rev. Oncol. Hematol. 53: 3-11.
23. Meynard D., Morvan V., Bonnet J. and Robert J. (2007). Functional analysis of the gene
espression profiles of colorectal cancer cell lines in relation to oxaliplatin and cisplatin
cytotoxicity. Oncol. Rep. 17: 1213-1221.
24. Reardon J., Vaisman A., Chaney S. and Sancar A. (1999). Efficient Nucleotide Excision
Repair of Cisplatin, Oxaliplatin, and Bis-acetoammine-dichloro-cyclohexylamine-platinum
(IV) (JM216) Platinum Intrastrand DNA Diadducts. Cancer Res. 59: 3968-3971.
25. Samimi G., Safaei R., Katano K., Holzer A., Rochdi M. et al. (2004). Increased Expression of
the Copper Efflux Transporter ATP7A Mediates Resistance to Cisplatin, Carboplatin and
Oxaliplatin in Ovarian Cancer Cells. Clin. Cancer Res. 10: 4661-4669.
26. Longley D., Harkin D. and Johnston P. (2003). 5-Fluorouracil: Mechanisms of action and
clinical strategies. Nat. Rev. Cancer 3: 330-338
27. Salonga D., Danenberg K., Johnson M., Metzger R., Groshen S. et al. (2000). Colorectal
tumors responding to 5-fluorouracil have low gene expression levels of dihydropyrimidine
47
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
dehydrogenase, thymidylate synthase, and thymidine phosphorylase. Clin. Cancer Res. 6:
1322–1327.
Ahnen D. Feigl P., Quan G., Fenoglio-Preiser C., Lovata et al. (1998). Ki-ras mutation and
p53 overexpression predict the clinical behavior of colorectal cancer: a Southwest Oncology
Group study. Cancer Res. 58: 1149-1158.
Meyers M. and Wagner M. (2001). Role of the hMLH1 DNA mismatch repair protein in
fluoropyrimidine-mediated cell death and cell cycle responses. Cancer Res. 61: 5193–5201.
Elsaleh H., Powell B., McCaul K., Grieu F., Grant R. et al. (2001). p53 alteration and
microsatellite instability have predictive value for survival benefit from chemotherapy in stage
III colorectal carcinoma. Clin. Cancer Res. 7: 1343-1349.
William G. and Kaelin J. (2005). The concept of syntetic lethality in the context of anticancer
therapy. Nat. Rev. Cancer 5: 689-697.
Schrama D., Reisfeld R. and Becker J. (2006). Antibody targeted drugs as cancer therapeutics.
Nat. Rev. Drug Disc. 5: 147-157.
Iqbal S. and Lenz H. (2004). Integration of novel agents in the treatment of colorectal cancer.
Cancer Chemother. Pharmacol. 54 Suppl 1: S32-39.
Wolpin B., Meyerhardt J., Mamon H. and Mayer R. (2007). Adjuvant Treatment of Colorectal
Cancer. CA Cancer J. Clin. 57: 168-185
Francini G., Petrioli R., Lorenzini L. et al. (1994). Folinic acid and 5-fluorouracil as adjuvant
chemotherapy in colon cancer. Gastroenterology 106: 899-906.
Giacchetti S., Perpoint B., Zidani R., Le Bail N., Faggiuolo R. et al. (2000). Phase III
multicenter randomized trial of oxaliplatin added to chronomodulated fluorouracil-leucovorin
as first-line treatment of metastatic colorectal cancer. J. Clin. Oncol. 18: 136-147.
Iyer L. and Ratain M. (1998). Clinical pharmacology of camptothecins. Cancer Chemother.
Pharmacol. 42 (suppl): S31-S43.
Hockel M. and Vaupel P. (2001). Tumor hypoxia: definitions and current clinical, biologic,
and molecular aspects. J. Natl. Cancer Inst. 93: 266-276.
Brown M. and Giaccia J. (1998). The unique physiology of solid tumours: opportunities (and
problems) for cancer therapy. Cancer res. 58: 1408-1416.
Vaupel P. and Mayer A. (2007). Hypoxia in cancer: significance and impact on clinical
outcome. Cancer Metastasis Rev. 26: 225-239
Trédan O., Galmarini C., Patel K. and Tannock I. (2007). Dug resistance and the solid tumour
microenvironment. J. Natl. Cancer Inst. 99: 144-154.
Brahimi-Horn M., Pouysségur J. (2007). Oxygen, a source of life and stress. FEBS Lett. 581:
3582-3591.
Semenza G. (2003). Targetting HIF-α for cancer therapy. Nat. Rev. Cancer 3: 721-732.
Durai R., Yang W., Gupta S., Seifalian A. and Winslet M. (2005). The role of the insulin-like
growth factor system in colorectal cancer: review of current knowledge. Int. J. Colorectal Dis.
20: 203-220.
Zagórska A. and Dulak J. (2004). HIF-1: the knowns and unknowns of hypoxia sensing. Acta
Biochim. Pol. 51(3): 563-585.
Guba M., Seeliger H., Kleespies A., Jauch K. and Bruns C. (2004). Vascular endothelial
growth factor in colorectal cancer. Int. J. Colorectal Dis. 19: 510-517.
Erler J., Cawthorne J., Williams J., Koritzinsky M., Wauters B. et al. (2004). Hypoxiamediated down-regulation of Bid and Bax in tumours occurs via hypoxia-inducable factor 1dependent and -independent mechanisms and contributes to drug resistance. Mol. Cell. Biol.
24: 2875-2889.
Nishi H., Nakada T., Kyo S., Inoue M., Shay J. et al. (2004). Hypoxia-inducible factor 1
mediates upregulation of telomerase (hTERT). Mol Cell Biol. 24(13): 6076-6083.
Krishnamachary B., Berg-Dixon S., Kelly B., Agani F., Feldser D. et al. (2003). Regulation of
colon carcinoma cell invasion by hypoxia-inducable factor 1. Cancer Res. 63: 1138-1143.
Pouysségur J., Daylan F. and Mazure N. (2006). Hypoxia signalling in cancer and approaches
to enforce tumour regression. Nature 441: 437-443.
48
51. Comerford K., Wallace T., Karhausen J., Lous N., Montalto M. et al. (2002). Hypoxiainducible factor-1-dependent regulation of the multidrug resistance (MDR1) gene. Cancer
Res. 62(12): 3387-3394.
52. Zhong H., De Marzo M., Laughner E., Lim M., Hilton D. et al. (1999). Overexpression of
hypoxia-inducable factor 1 alpha in common human cancers and their metastases. Cancer Res.
59: 5830-5835.
53. Belozerov V. and Van Meir E. (2006). Inhibitors of hypoxia-inducible factor-1 signalling.
Curr. Opin. Investig. Drugs 7(12): 1067-1076.
54. Chan D., Sutphin P., Yen S. and Giaccia A. (2005). Coordinate regulation of the OxygenDependent Degradation Domains of Hypoxia-Inducable Factor 1α. Mol. Cell. Biol. (Vol. 25)
15: 6415-6426.
55. Chen N., Chen X., Huang R., Zeng H., Gong J. et al. (2009). BCL-xL Is a Target Gene
Regulated by Hypoxia-inducible Factor-1α. J. Biol. Chem. (Vol. 284) 15: 10004-10012.
56. Unruh A., Ressel A., Mohamed H., Johsnon R., Nadrowitz R. et al. (2003). The hypoxiainducible factor-1 alpha is a negative factor for tumor therapy. Oncogene 22: 3213-3220.
57. Gerweck L., Vijayappa S. and Kozin S. (2006). Tumor pH controls the in vivo efficacy of
weak acid and base chemotherapeutics. Mol. Cancer Ther. 5: 1275-1279.
58. Ravi R., Mookerjee B., Bhujwalla Z., Sutter C., Artemov D. et al. (2000). Regulation of
tumour angiogenesis by p53-induced degradation of hypoxia-inducable factor 1. Genes Dev.
14: 34-44.
59. Schmid T., Zhou J., Köhl R. and Brüne B. (2004). p300 relieves p53-evoked transcriptional
repression of hypoxia-inducable factor-1 (HIF-1). Biochem. J. 380: 289-295.
60. Bertholon J., Wang Q., Galmarini C. and Puisieux A. (2006). Mutational targets in colorectal
cancer cells with microsatellite instability. Fam. Cancer 5: 29-34.
61. Adamsen B., Kravik K., Clausen O. and De Angelis P. (2007). Apoptosis, cell cycle
progression and gene expression in TP53-depleted HCT 116 colon cancer cells in response to
short-term 5-fluorouracil treatment. Int. J. Oncol. 31 (6): 1491-1500.
62. Brown L., Cowen R., Debray C., Eustace A., Erler J. et al. (2006). Reversing Hypoxic Cell
Chemoresistance in Vitro using Genetic and Small Molecule Approaches Targeting Hypoxia
Inducible Factor-1. Mol. Pharmacol. 69: 411-418.
63. Wu H., Liang X., Fang Y., Qin X., Zhang Y. et al. (2008). Resveratrol inhibits hypoxiainduced metastasis potential enhancement by restricting hypoxia-induced factor-1α expression
in colon carcinoma cells. Biomed. Pharmacother. 62: 613-621.
64. Yeo E., Chun Y., Cho Y., Kim J., Lee J. et al. (2003). YC-1: a potential anticancer drug
targeting hypoxia-inducible factor 1. J. Natl. Cancer Inst. (Vol. 95) 7: 516-525.
65. Mukherjee A. and Martin S. (2008). The thioredoxin system: a key target in tumour and
endothelial cells. Br. J. Radiol. 81: S57-S68.
66. Jones D., Pugh C., Wigfield S., Stevens M. and Harris A. (2006). Novel thioredoxin inhibitors
paradoxically increase hypoxia-inducible factor-alpha expression but decrease functional
transcriptional activity, DNA binding, and degradation. Clin. Cancer Res. 12 (18): 5384-5394.
67. Dwain B. and Corey D. (2000). Locked nucleid acid (LNA) : fine tuning the recognition of
DNA and RNA. Chem. Biol. 8: 1-7.
68. Greenberger L., Horak I., Filpula D., Sapra P. and Westergaard M. (2008). A RNA antagonist
of hypoxia-inducible factor-1α, EZN-2968, inhibits tumor cell growth. Mol. Cancer. Ther. 7
(11): 3598-3608.
69. Comes F., Matrone A., Lastella P., Nico B., Susca F. et al. (2007). A novel cell type-specific
role of p38alpha in the control of autophagy and cell death in colorectal cancer cells. Cell
Death. Differ. 14(4): 693-702.
70. Cao Y., Li C., Moeller B., Yu D, Zhao Y. et al. (2005). Observation of incipient tumor
angiogenesis that is independent of hypoxia and hypoxia inducible factor-1 activation. Cancer
Res. 65: 5498-5505.
71. Mariadason J., Arango D., Shi Q., Wilson A., Corner G. et al. (2003). Gene expression
profiling-based prediction of response of colon carcinoma cells to 5-fluorouracil and
camptothecin. Cancer Res. 63: 8791-8812.
49
72. Osti D., Marras E., Ceriani I., Grassini G., Rubino T. et al. (2006). Comparative analysis of
molecular strategies attenuating positional effects in lentiviral vectors carrying multiple genes.
J. Virol. Methods 136: 93-101.
73. Chou T. and Talalay P. (1984). Quantitative analysis of dose-effect relationships: The
combined effects of multiple drugs or enzyme inhibitors. Adv. Enzyme. Regul. 22: 27-55.
74. Boyer J., McLean E., Aroori S., Wilson P., McCulla A. et al. (2004). Characterization of p53
Wild-Type and Null Isogenic Colorectal Cancer Cell Lines Resistant to 5-Fluorouracil,
Oxaliplatin, and Irinotecan. Clin. Cancer Res. 10: 2158-2167.
75. Peters J., Triest B., Backus H., Kuiper C., van der Wilt C. et al. (2000). Molecular
downstream events and induction of thymidylate synthase in mutant and wild-type p53 colon
cancer cell Lines after treatment with 5-fluorouracil and the thymidylate synthase inhibitor
raltitrexed. Eur. J. Cancer 36: 916-924.
76. Peters J., Backus H., Freemantle S., van Triest B., Codacci-Pisanelli G. et al. (2002).
Induction of thymidylate synthase as a 5-fluorouracil resistance mechanism. Biochim.
Biophys. Acta 1587: 194-205.
77. Nief N., Le Morvan V. and Robert J. (2007). Involvement of gene polymorphisms of
thymidylate synthase in gene expression, protein activity and anticancer drug cytotoxicity
using the NCI-60 panel. Eur. J. Cancer 43: 955-962.
78. Endo M., Miwa M., Eda H., Ura M., Tanimura H. et al. (2003). Augmentation of the
antitumor activity of capecitabine by a tumor selective dihydropyrimidine dehydrogenase
inhibitor, RO0094889. Int. J. Cancer 106: 799-805.
79. Miyazaki K., Shibahara T., Sato D., Uchida K., Suzuki H. et al. (2006). Influence of
chemotherapeutic agents and cytokines on the expression of 5-fluorouracil-associated
enzymes in human colon cancer cell lines. J. Gastroenterol. 41: 40-50.
80. Krishnamurthy P., Ross D., Nakanishi T., Bailey-Dell K., Zhou S. et al. (2004). The stem cell
marker Bcrp/ABCG2 enhances hypoxic cell survival through interactions with heme. J. Biol.
Chem. 279: 24218-24225.
81. Yao K., Gietema J., Shida S., Selvakumaran M., Fonrose X. et al. (2005). In vitro hypoxiaconditioned colon cancer cell lines derived from HCT116 and HT29 exhibit altered apoptosis
susceptibility and a more angiogenic profile in vivo. Br. J. Cancer 93: 1356-1363.
82. Fischer F., Baerenfaller K. and Jiricny J. (2007). 5-Fluorouracil is efficiently removed from
DNA by the base excision and mismatch repair systems. Gastroenterology 133: 1858-1868.
83. Fujita H., Kato J., Hrii J., Harada K., Hiraoka S. et al. (2007). Decreased expression of
hMLH1 correlates with reduced 5-fluorouracil-mediated apoptosis in colon cancer cells.
Oncol. Rep. 18: 1129-1137.
84. Meyers M., Hwang A., Wagner M. and Bootham D. (2004). Role of DNA Mismatch repair in
apoptotic responses to therapeutic agents. Environ. Mol. Mutagen. 44: 249-264.
85. Peng X., Karna P., Cao Z., Jiang B., Zhou M. et al. (2006). Cross-talk between epidermal
growth factor receptor and hypoxia-inducible factor-1α signal pathways increases resistance to
apoptosis by up-regulating survivin gene expression. J. Biol. Chem. 281: 25903-25914.
86. Arango D., Wilson A., Shi Q., Corner G., Aranes M. et al. (2004). Molecular mechanisms of
action and prediction of response to oxaliplatin in colorectal cancer cells. Br. J. Cancer 91:
1931-1946.
87. Howells L., Mitra A. and Manson M. (2007). Comparison of oxaliplatin- and curcuminmediated antiproliferative effects in colorectal cell lines. Int. J. Cancer 121: 175-183.
88. Fink D., Nebel S., Aebi S., Zheng H., Cenni B. (1996). The role of DNA mismatch repair in
Platinum Drug Resistance. Cancer Res. 56: 4881-4886.
89. Kitada N., Takra K., Minegaki T., Itoh C., Tsjujimoto M. et al. (2007). Factors affecting
sensitivity to antitumor platinum derivatives of human colorectal tumor cell lines. Cancer
Chemother. Pharmacol. 62 (4): 577-584.
90. Hector S., Bolanowska-Higdon W., Zdanowicz J., Hitt S. and Pendyala L. (2001). In vitro
studies on the mechanisms of oxaliplatin resistance. Cancer Chemorther. Pharmacol. 48: 398406.
91. Graeber T., Osmanian C., Jacks T., Housmand D., Koch C. et al. (1996). Hypoxia-mediated
selection of cells with diminshed apoptotic potential in solid tumours. Nature 379: 88-91.
50
51