university of zurich - zora.uzh.ch filesuppressor pathway in cancer cells. we showed that genetic...

University of ZurichZurich Open Repository and Archive

Winterthurerstr. 190

CH-8057 Zurich

http://www.zora.uzh.ch

Year: 2008

Strategies for prevention, circumvention, and reversal of drugresistance in tumor cells

Fedier, A

Fedier, A. Strategies for prevention, circumvention, and reversal of drug resistance in tumor cells. 2008, Universityof Zurich, Faculty of Medicine.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:University of Zurich, Faculty of Medicine, 2008.

Fedier, A. Strategies for prevention, circumvention, and reversal of drug resistance in tumor cells. 2008, Universityof Zurich, Faculty of Medicine.Postprint available at:http://www.zora.uzh.ch

Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch

Originally published at:University of Zurich, Faculty of Medicine, 2008.

Strategies for prevention, circumvention, and reversal of drugresistance in tumor cells

Abstract

Both the intrinsic and the acquired anticancer drug resistance presents until today a major drawback inthe management of cancer. The handling with drug resistance is aggravated by the facts that drugresistance is a multifactorial and multifaceted phenomenon. The development of tumor-targetedpharmacological and genetic strategies to overcome intrinsic resistance and to prevent the resistanceacquisition is thus still a pressing need. We put forward the two main objectives: 1) The reversal of thepre-existing drug resistance in p53-deficient cells by either the concomitant abrogation of DNA repairfunctions or by the concomitant blocking of cytoprotective signaling pathways. 2) The identification ofmembers of the novel class of histone deacetylase (HDAC) inhibitors with anticancer activity that have(or do not have) the potential to cause resistance acquisition in tumor cells. Mutations in the p53 geneand dysregulation of p53-dependent cellular processes are often associated with cancer and resistance toradio- and chemotherapy. Many Advances have been taken to reconstitute the (wildtype) p53 tumorsuppressor pathway in cancer cells. We showed that genetic abrogation of the expression of the DNAmismatch repair (MMR) gene Pms2 in cells depleted for the p53 tumor suppressor functionhypersensitized these cells to the cytotoxic effect of some anticancer agents through an increase in therate of apoptosis. Expanding on this, we also showed that genetic depletion of either ATM, BRCA1, orDNA-PK resulted in an increase in the chemosensitivity of p53-deficient cells against a variety ofanticancer agents. These studies indicate that tumor-targeted functional inhibition of either one or theother of these DNA repair genes may reverse resistance in p53-mutated tumors and therefore provide astrategy for increasing chemotherapy efficacy for these tumors. Analogous as well as opposing resultswere reported for loss of the MMR protein MLH1 in p53-deficient cells with Cisplatin. However,neither Radicicol (HSP90 inhibitor) nor LY294002 and LY294005 (Akt inhibitors) were found toreverse Cisplatin resistance specifically in MMR-deficient tumor cells. These Akt inhibitors even furtherdecreased the sensitivity against Cisplatin, suggesting - in addition to its acknowledged cytoprotectivefunction - a cytodestructive function of Akt. On the other hand, we showed the potential of the HDACinhibitor SAHA to induce stable and MDR-independent resistance in HCT116 colon tumor cells. Asthese results differ from those reported with FK-228, we propose a novel mechanism of resistanceacquisition by the SAHA, the nature of which is still obscure. This is under investigation, as is novel andintriguing molecular approach to reverse drug resistance: the liposomal delivery of (exogenous)cytochrome c to tumor cells.

Departement Frauenheilkunde

Universitätsspital Zürich

Klinik für Gynäkologie

(Klinikdirektor: Prof. Dr. med. Daniel Fink)

Strategies for prevention, circumvention, and reversal

of drug resistance in tumor cells

Habilitationsschrift zur Erreichung der

Venia Legendi der Universität Zürich

für das Gebiet

Experimentelle Gynäkologie

Vorgelegt von André Fedier

Herbstsemester 2007

Table of Contents

1. Summary 1

2. Introduction

2.1. Drug resistance as a limiting factor of cancer treatment efficacy 2

2.2. The multiple faces of drug resistance 3

2.3. Resistance due to dysfunction of the p53 tumor suppressor gene 4

2.4. Resistance as a consequence of loss of DNA mismatch repair function 4

2.5. Resistance to histone deacetylase inhibitors 5

2.6. Prevention, circumvention, and reversal of drug resistance 6

2.6.1. Reversal of drug resistance in p53-null cells by genetic deletion of

the Pms2 DNA mismatch repair gene 7

2.6.2. Reversal of drug resistance in p53-null cells by genetic deletion of

DNA double-strand break repair genes 8

2.6.3. Reversal of drug resistance in MMR-deficient cells by pharmacological

blocking cyto-protective signaling pathways

9

2.6.4. Acquired resistance to HDAC-inhibitors and its prevention 10

3. Publication I 12

Increased sensitivity of p53-deficient cells to anticancer agents due to loss of Pms2

Fedier A, Ruefenacht U B, Schwarz V A, Haller U, Fink D

Br. J. Cancer 2002;87:1027-1033

4. Publication II 13

Loss of atm sensitises p53-deficient cells to topoisomerase poisons and antimetabolites

Fedier A, Schlamminger M, Schwarz V A, Haller U, Howell S B, Fink D

Ann. Oncol. 2003;14:938-945

5. Publication III 14

The histone deacetylase inhibitors suberoylanilide hydroxamic (vorinostat) and valproic

acid induce irreversible and MDR1-independent resistance in human colon cancer cells

Fedier A, Dedes K J, Imesch P, von Bueren A O, Fink D

Int. J. Oncol. 2007;31:633-641

6. Conclusion and perspective

6.1. What is the molecular basis of the acquired resistance by HDAC inhibitors? 15

6.2. Cytochrome c as therapeutic tool to overcome drug resistance? 17

7. References 19

Acknowledgements

My deepest gratitude goes to my academic teachers and mentors:

Prof. Dr. Daniel Fink

Prof. Dr. Urs Haller

My gratitude goes also to the members of the laboratory for their great support and assistance:

Among those are Dr. Viola Heinzelmann-Schwarz, Dr. Konstantin J. Dedes, Dr. Patrick Imesch,

and Ursula Schuchter.

This work was supported by the “Swiss National Foundation” (SNF), the “EMDO Stiftung

Zürich”, the “Hartmann-Müller Stiftung”, and the “Swiss Society for Gynecology and

Obstetrics” (SGGG).

Publication I

Increased sensitivity of p53-deficient cells to anticancer agents due to loss of Pms2


Br. J. Cancer 2002;87:1027-1033

Publication II

Loss of atm sensitises p53-deficient cells to topoisomerase poisons and antimetabolites


Ann. Oncol. 2003;14:938-945

Publication III

The histone deacetylase inhibitors suberoylanilide hydroxamic (vorinostat) and

valproic acid induce irreversible and MDR1-independent resistance in human colon

cancer cells


Int. J. Oncol. 2007;31:633-641

1. Summary

Both the intrinsic and the acquired anticancer drug resistance presents until today a major

drawback in the management of cancer. The handling with drug resistance is aggravated by the

facts that drug resistance is a multifactorial and multifaceted phenomenon (1). The development

of tumor-targeted pharmacological and genetic strategies to overcome intrinsic resistance and to

prevent the resistance acquisition is thus still a pressing need. We put forward the two main

objectives: 1) The reversal of the pre-existing drug resistance in p53-deficient cells by either the

concomitant abrogation of DNA repair functions or by the concomitant blocking of

cytoprotective signaling pathways. 2) The identification of members of the novel class of

histone deacetylase (HDAC) inhibitors with anticancer activity that have (or do not have) the

potential to cause resistance acquisition in tumor cells.

Mutations in the p53 gene and dysregulation of p53-dependent cellular processes are often

associated with cancer and resistance to radio- and chemotherapy (2-6). Many Advances have

been taken to reconstitute the (wildtype) p53 tumor suppressor pathway in cancer cells (7). We

showed that genetic abrogation of the expression of the DNA mismatch repair (MMR) gene

Pms2 in cells depleted for the p53 tumor suppressor function hypersensitized these cells to the

cytotoxic effect of some anticancer agents (see PUBLICATION I; 8) through an increase in the

rate of apoptosis. Expanding on this, we also showed that genetic depletion of either ATM (see

PUBLICATION II; 11), BRCA1 (12), or DNA-PK (13) resulted in an increase in the

chemosensitivity of p53-deficient cells against a variety of anticancer agents. These studies

indicate that tumor-targeted functional inhibition of either one or the other of these DNA repair

genes may reverse resistance in p53-mutated tumors and therefore provide a strategy for

increasing chemotherapy efficacy for these tumors. Analogous as well as opposing results were

reported for loss of the MMR protein MLH1 in p53-deficient cells with Cisplatin (9,10).

However, neither Radicicol (HSP90 inhibitor) nor LY294002 and LY294005 (Akt inhibitors)

were found to reverse Cisplatin resistance specifically in MMR-deficient tumor cells (14,15).

These Akt inhibitors even further decreased the sensitivity against Cisplatin, suggesting - in

addition to its acknowledged cytoprotective function - a cytodestructive function of Akt.

On the other hand, we showed the potential of the HDAC inhibitor SAHA (16,17) to

induce stable and MDR-independent resistance in HCT116 colon tumor cells (see

PUBLICATION III; 18). As these results differ from those reported with FK-228 (17,18), we

propose a novel mechanism of resistance acquisition by the SAHA, the nature of which is still

obscure. This is under investigation, as is novel and intriguing molecular approach to reverse

drug resistance: the liposomal delivery of (exogenous) cytochrome c to tumor cells.

1

2. Introduction 2.1. Drug resistance as a limiting factor of cancer treatment efficacy

The management of cancer involves procedures, which include surgery, radiotherapy,

chemotherapy, and endocrine therapy. Despite these in many cases successful therapeutic

options, there are still drawbacks encountered by the oncologists. One of these drawbacks is

drug resistance, which until today is a persistent problem during the treatment of local and

disseminated diseases. Despite the large armentarium that has become available to oncologists

during the past decades, resistance to an anticancer treatment, either present intrinsically (be

inherent in a subpopulation of heterogeneous cancer cells) in tumor cells or acquired during a

treatment (as a cellular response to drug exposure), is a frequently observed and persistent

problem during the cancer treatment and thus presents a major obstacle in the fight against

cancers. It is clear that resistance to chemotherapy limits the effectiveness of anticancer drug

treatment. Resistance to chemotherapy is believed to cause treatment failure in over 90% of

patients with metastatic cancer, and resistant micrometastic tumor cells may also reduce the

effectiveness of chemotherapy in the adjuvant setting. Acquired resistance is a particular

problem, as tumors not only become resistant to the drugs originally used to treat them but may

also become cross-resistant to other drugs with different mechanisms of action.

This is not different with gynecologic cancers. Breast cancer is the most common form of

cancer among women in North America and Europe. It is estimated that each year this disease is

diagnosed in over one million people worldwide and is the cause of more than 400,000 deaths.

Although chemotherapy including endocrine therapy forms part of a successful treatment

regime in many cases, only about 50% of the patients may benefit from this, as a result of

intrinsic or acquired multiple drug resistance or endocrine resistance (21).

Likewise, despite the considerable clinical efficacy of chemotherapy for primary ovarian

cancer as well as among woman with advanced and suboptimal disease following surgery, most

women with advanced stage of ovarian cancer would relapse, including 50% of woman who

have no evidence of disease after primary therapy. Acquired resistance of the recurring tumor

mass to platinum is one important factor for the poor overall prognosis for recurrent ovarian

cancers and helps explaining why ovarian cancer remains one of the most lethal malignancies

(22).

Endometrial cancer, the most common gynecologic malignancy, and cervical cancer, the

third most common type of gynecologic cancers (but much less common in the developed

countries due to the introduction of the Pap smear test) have excellent long-term prognoses due

to early stage diagnosis and efficient surgical and/or radiation treatment. However, decreased

responses to treatment and increased recurrence risk are observed in both malignancies, and the

2

therapeutical options (adjuvant or neoadjuvant chemotherapy, surgical resection) remain limited

(23). It is believed that chemotherapy-resistance is an important factor; in particular in HPV-

positive cervical cancers, which constitute more than 95% of cervical carcinomas, seem to have

an inherent, HPV-associated resistance to chemotherapy-induced apoptosis (24).

2.2. The multiple faces of drug resistance

Drug resistance is a multifaceted phenomenon and can occur at many levels. Extensive review

papers and books on the issue of drug resistance in cancers are available (1). Three main classes

of resistance may be defined. Physiological resistance describes a pathophysiologic

phenomenon characterized by the interrelationship between dysregulated angiogenesis,

abnormal oxygenation, hypoxia and acidosis, and unstable perfusion. These factors help

maintaining a tumor microenvironment that promotes treatment resistance and increases

propensity for invasion and metastasis. Biological resistance is characterized by the presence of

a subpopulation of non-curable tumor stem cells which drives cancer development and by the

presence of a tumor microenvironment that promotes metastasis and therefore significantly

affects the success of therapy. Parameters defining such a tumor microenvironment include the

interactions between cell-cell interface, cell-extracellular matrix interface, and between cells

and cytokines/soluble factors. Biochemical resistance is another well studied phenomenon.

Principal mechanisms include altered membrane transport involving the P-glycoprotein product

of the multidrug resistance (MDR) gene as well as other associated proteins (e.g. decreased drug

uptake, increased drug efflux); altered drug targets (e.g. mutated drug targets such as

topoisomerase II and tubulins); decreased drug activation, increased drug degradation, drug

inactivation, or increased subcellular redistribution and sequestration (e.g. altered expression of

drug-metabolizing enzymes, conjugation with increased glutathione and metallothioneins);

enhanced DNA repair or DNA damage tolerance (e.g. loss of DNA mismatch repair MMR);

failure to apoptose as a result of mutated cell cycle or cell suicide proteins (e.g. mutations in the

p53 tumor suppressor gene); and covalent epigenetic modifications (DNA methylation, histone

(de-)acetylation) leading to transcriptional silencing of genes relevant to drug responses.

In particular, the drug resistance arisen (i) from the functional loss of the tumor suppressor

gene p53, (ii) from mutations in the MMR genes, and (iii) the acquired resistance to histone

deacetylase inhibitors have been in the focus in our laboratory.

3

2.3. Resistance due to dysfunction of the p53 tumor suppressor gene

The p53 tumor suppressor gene encodes a transcription factor that regulates the expression of

targets genes which induce cell cycle arrest, apoptosis, senescence, DNA repair or alter

metabolism in response to a variety of stresses, including DNA damage, overexpressed

oncogenes and various metabolic limitations (4,25-28). Inactivation of tumor suppressor genes

can lead to a lack of proper control, especially under stress, leading to clonal outgrowth and

tumor progression that, in turn, usually involves blockage of normally regulated cell cycle

control and apoptosis mediated by tumor suppressor genes. Two well established consequences

of loss of the appropriate function of p53, either by mutations or by transcriptional down-

regulation, are tumorigenesis and progression and the frequently observed reduced

responsiveness (i.e. resistance) to genotoxic agents.

On the one hand, p53 is the prototype tumor suppressor gene in human cancer due to its

pro-apoptotic and antiproliferative function in response to oncogenic stress (4,25,28,29).

Depending on the severity of damage to the genome, p53 can activate genetic programs that halt

cell proliferation transiently (G1 and G2 cell cycle arrest) or permanently (senescence), or

eliminate the cell altogether (apoptosis). The p53 pathway is inactivated in the majority of

human malignancies (2-4). Increased levels of its negative regulator MDM2 downregulate p53

function in many of the rest (30), and therefore the p53 pathway is most likely disrupted also in

a large fraction of wild-type p53-carrying tumors. Li-Fraumeni syndrome is an inherited cancer

predisposition caused by p53 mutations, giving rise to early onset of cancers (31). p53 also

plays a role in inflammatory responses and in non-neoplastic diseases (32).

On the other hand, the p53-status of the tumor is an important determinant of the

sensitivity of tumors to genotoxic agents. p53 mutations or inactivation of its function often

associate with chemoresistance in tumors, e.g. against antracyclines, antimetabolites, and

cisplatin (5,6,33). In the last decade much effort has been taken to reconstitute the p53 tumor

suppressor pathway in cancer cells (7). Despite some promising advances, the list of p53-based

therapeutic strategies under preclinical development is growing and novel p53-based therapeutic

strategies may also be combined with conventional cancer therapy.

2.4. Resistance as a consequence of loss of DNA mismatch repair function

The DNA mismatch repair (MMR) system functions in the elimination of biosynthetic errors

that arise during DNA replication, in DNA damage surveillance, and in the prevention of

recombination between non-identical sequences. It therefore substantially contributes to the

4

maintenance of genome integrity (34-37). It is thus not surprising that loss of MMR function

may lead to cancer (38-40). Inherited defects due to germline mutations in the MMR genes

underlie the hereditary non-polyposis colon cancer (HNPCC) syndrome in humans, and

epigenetic silencing of the hMLH1 gene accounts for sporadic cancers, including those of the

endometrium and ovaries. Even more important, another hallmark of MMR is its capacity to

elicit DNA damage-induced cell death (41). Although this might seem to make MMR a useful

target for anticancer agents, it has become clear that tumor cells with defective MMR can

display reduced sensitivity to DNA damaging agents. This was in fact the case with some types

of DNA damaging agents (42). Previous studies, including those by our own laboratory (18,43-

47) have identified a variety of commonly used anticancer agents against which MMR-deficient

tumor cells are resistant, including the alkylating agents Temozolomide and Busulfan, the

platinum compounds Cisplatin, Carboplatin, and Lipoplatin (liposomal Cisplatin; Regulon Inc.,

Mountain View, CA), the topoisomerases I poisons Camptothecin and Topotecan, the

topoisomerase II poisons Doxorubicin, Epirubicin, Mitoxantrone, and Etoposide, and the

antimetabolites 5-Fluorouracil and 6-Thioguanine. In addition and not less important, anticancer

agents and a light-based therapy were also identified against which these cells retain sensitivity.

These include the taxanes Docetaxel and Paclitaxel, the platinum-derivative Oxaliplatin and its

liposomal formulation Lipoxal (Regulon Inc., Mountain View, CA), the DNA minor groove

binder Brostallicin, the histone deacetylase (HDAC) inhibitors SAHA, Trichostatin A, and

Valproate, and the photodynamic therapy. Chemoresistance due to loss of MMR would predict

reduced therapeutic efficacy in the treatment of these tumors, and this is even exacerbated by

the potential of some agents to cause de novo generation of MMR-resistant variants (48).

Although still on debate, the potential clinical relevance of this usually moderate degree (2 to 4-

fold) of resistance has been demonstrated in clinical studies (49-51).

2.5. Resistance to histone deacetylase inhibitors

Histone deacetylase (HDAC) inhibitors are a relatively novel and promising class of potent

anticancer agents, as they have strong anticancer properties and show promising results with

respect to efficacy and tolerability (52-55). By affecting gene expression, HDAC inhibitors,

either by themselves or in combination with other anticancer agents, induce cell cycle arrest and

apoptosis specifically in tumor cells (56-58). They also counteract invasion, vascularization, and

angiogenesis (59).

Vorinostat (SAHA, Zolinza) is one of the lead compounds and is the first of the new

HDAC-inhibitors to be approved by Food and Drug Administration for the clinical use in cancer

5

patients, namely the treatment of cutaneous T-cell lymphoma (CTCL) (16,60-62). Vorinostat is

currently tested in more than 30 clinical trials. At least 14 different HDAC-inhibitors (e.g.

LBH589, FK-228, MS-275, Valproic acid) are in clinical trials as monotherapy or in

combination in patients with hematologic and solid tumors.

However, resistance to HDAC-inhibitors has also been observed in preclinical studies

(17,63,64) and even in clinical trials (61). The basis of resistance is not well understood, but

overexpression of anti-apoptotic Bcl-2 (65), of thioredoxin (66), of the ROS-reducing

peroxiredoxin (67), and of MDR may play a role (68). Moreover, the HDAC-inhibitor FK-228

(depsipeptide) induced reversible expression of MDR in tumor cells, rendering them transiently

resistant to FK-228 (19,20). This suggests the potential of HDAC inhibitors to induce

resistance. Whereas the mode of action of HDAC inhibitors has been intensively studied (58),

the mechanisms of resistance is much less understood.

2.6. Prevention, circumvention, and reversal of drug resistance

The development of pharmacological and genetic strategies to overcome intrinsic resistance and

the prevention of resistance acquisition during chemotherapy are a pressing need for researchers

and oncologists. However, this is often compromised by the fact that drug resistance is a

multifactorial and multifaceted phenomenon that is characterized by the physiology and the

microenvironment of the tumor, by the efficacy of drug delivery to the tumor, by the drug

activation/inactivation and the availability of the drug target in the tumor, and perhaps also by

the different clinicopathological characteristics of different subgroups of patients.

Drug resistance may not be restricted to a certain (chemical) class of anticancer agents but

could be observed for almost any molecule with anticancer activity. With the development of

novel and efficient anticancer agents, researches and oncologists should consider the possibility

that a novel and promising anticancer agent could potentially turn out to be associated with

reduced responsiveness against some but not other tumor types and/or with a certain

subpopulation of cancer patients, and hence to be an inducer of a stable drug resistance

phenotype. Great efforts are being taken to identify cancer-relevant molecular targets that can be

targeted by potent molecules with a high antitumor activity. In addition to the anticancer agents

already widely and successfully used for cancer treatment, a variety of novel classes of potential

targets and anticancer agents arose in the last decade.

Approaches to handle with drug resistance in tumor cells mainly include the following

strategies: i) the use of combination drug therapy using different classes of drugs with

minimally overlapping toxicities to allow maximal dosages and with narrowest cycle intervals;

6

(ii) the identification of anticancer agents to which tumor cells are intrinsically resistant and

therefore may not be used for therapies; iii) the identification of anticancer agents that retain

sensitivity also in drug-resistant cells (i.e. no cross-resistance) and therefore may be suitable for

a therapy; iv) the application of molecular (genetic and pharmacologic) strategies that reverse

the drug-resistance phenotype either by further increasing cellular stress or by downregulating

cytoprotective functions; v) the identification of anticancer agents that have the potential to

cause acquisition of resistance in initially sensitive tumor cells (and of those which don’t) and

therefore may (or may not) compromise a therapy.

With our interest in establishing novel and clinically applicable strategies to cope with

drug resistance in tumor cells, we became increasingly interested in both genetic and

pharmacological strategies that reverse (pre-existing) drug resistance on the one hand and in the

identification of novel and potential anticancer agents that result – or that do not result - in the

development (acquisition) of a drug resistance phenotype in tumor cells on the other. The

former is of increasing interest, because to date no clinically suitable compounds has been

identified to which preferentially MMR-deficient (but not MMR-proficient) tumor cells are

hypersensitive. Likewise, despite the promising achievements in restoring wildtype p53 function

in (resistant) tumors, there is still potential to be fully tapped in this regard.

With regard to the reversal of the (pre-existing) drug resistance phenotype arisen either by

loss of p53 function or by loss of MMR function, we put forward the two following strategies:

(i) the reversal of p53 defect-associated resistance through the genetic abrogation of MMR

function on the one hand (2.6.1) and of DNA double-strand break repair on the other (2.6.2); (ii)

the reversal of MMR defect-associated resistance through the pharmacological blocking of

cytoprotective signaling pathways (2.6.3).

With regard to the prevention of resistance development, the strategy has been put

forward that aimed at identifying members of the relatively novel class of potential anticancer

agents (the HDAC inhibitors) that either could or could not lead to development of a resistance

phenotype in tumor cells (2.6.4).

2.6.1. Reversal of drug resistance in p53-null cells by genetic deletion of the Pms2 DNA

mismatch repair gene

It is generally acknowledged that cells with defective p53 function or defective MMR exhibit

reduced responsiveness to some genotoxic agents. Thus, one might think that both defects

present in the cell at the same time would result in an even higher magnitude of resistance. Or

one might think that the absence of proper repair of mismatched DNA introduced by genotoxic

7

agents (due to loss of MMR) combined with the failure to properly induce cell cycle checkpoint

and apoptosis activation (due to loss of p53 function) in the presence of DNA damage would

cause the cell to become highly sensitive to the DNA damaging agent, i.e. would lead to

reversal of drug resistance in p53-deficient cells. This latter idea was pursued. The strategy was

to achieve hypersensitivity and drug resistance reversal in (chemoresistant) p53-deficient cells

through the additional (genetic) deletion of the function of the MMR gene Pms2.

Indeed, we were able to show (see PUBLICATION I, 8) hat loss of the MMR protein

Pms2 in p53-deficient cells results in hypersensitivity to the anticancer agents Cisplatin,

Doxorubicin, Etoposide, Docetaxel and Gemcitabine. Moreover, this hypersensitizing effect

resulting from Pms2 loss correlated with increased apoptosis but not with alterations in cell

cycle checkpoint control. These results indicated that p53-deficient (resistant) tumor cells can be

re-sensitized to anticancer agents by the concurrent loss of one specific type of DNA repair

function, i.e. MMR-function. This suggests that tumor cells which are defective in both

functions at the same time are more chemosensitive than the cells defective in p53 alone. It

therefore seems that Pms2 is a positive modulator of cell survival in p53-deficient cells (as

opposed to its cell death-inducing property in p53-wildtype cells), although the molecular basis

of this hypersensitizing effect is not fully understood. Analogous results were reported for loss

of the MMR protein MLH1 in p53-deficient cells with Cisplatin (9). However, opposing data

were also reported in a recent study (10): cells deficient in MLH1 and p53 function were even

more resistant than cells deficient in either one or the other function, and this correlated with

enhanced mutagenic translesion synthesis and increased homologous recombination.

2.6.2. Reversal of drug resistance in p53-null cells by genetic deletion of DNA double-

strand break repair genes

The aforementioned results with the Pms2/p53 doubly-deficient cells prompted us to

incorporate the strategy of “double-impact” into a more general concept. Accordingly, we

proposed that p53-deficient (chemoresistant) tumor cells could be re-sensitized to DNA

damaging agents through the additional loss of their ability to properly repair anticancer agents-

introduced DNA damage. The idea behind this was that cells accumulate high levels of

irreparable DNA damage (due to reduced repair) and at the same time fail to properly respond

(i.e. halting cell cycle progression and allowing for repair) to the presence of this damaged DNA

due to the lack of p53 function, finally leading to extensive and rapid cell death.

One of the target genes was ATM (Ataxia Telangiectasia Mutated). The ATM gene is

mutated in patients with ataxia telangiectasia, an autosomal recessive genetic disorder

characterized by progressive cerebellar ataxia, oculomotor apraxia, immunodeficiency,

8

radiohypersensitivity, and increased risk of malignancies including breast cancer (69). ATM

also play a central role in the repair of DNA double-strand break in that it is recruited to sites of

DNA damage (and thus acts as DNA damage sensor) where it initiates a signaling cascade

through phosphorylation of multiple DNA damage response and cell cycle proteins (70).

In line with our proposal, we were able to show that ATM/p53 doubly-deficient cells are

hypersensitive to the topoisomerase I poisons Camptothecin and Topotecan, to the

topoisomerase II poisons Doxorubicin, Epirubicin, and Etoposide, and to the antimetabolites 5-

Fluorouracil and Gemcitabine, as compared to ATM-expressing p53-deficient counterparts (see

PUBLICATION II; 11). This hypersensitizing effect correlated with increased necrosis rather

than with increased apoptosis and additional cell cycle alterations. We thus conclude that the

concomitant loss of the ATM function increases the sensitivity of p53-deficient cells to these

anticancer agents, and that tumor-targeted functional inhibition of ATM may provide a valuable

strategy for increasing the efficacy of anticancer agents in the treatment of p53-mutant and

chemoresistant cancers. No hypersensitization was observed with Cisplatin, Carboplatin,

Oxaliplatin, Docetaxel, and Paclitaxel.

We speculated that this hypersensitizing effect may not be restricted to the loss of the

ATM function, and in fact we found that this strategy applied also to the loss of BRCA1 (12) or

DNA-PK (13) in p53-deficient cells. The BRCA1 (Breast Cancer Type 1 Susceptibility) protein

is important in maintaining genomic stability by promoting efficient and precise repair of

double-strand breaks and by mediating signal transduction to a variety of downstream events

(cell cycle checkpoint control and apoptosis induction) upon DNA damage thought the

interaction with many other repair and repair-associated proteins (71). Carriers of heterozygous

mutations in BRCA1 are strongly predisposed to breast and ovarian cancers (72,73). DNA-PK

(DNA-Dependent Protein Kinase) belongs to the PI3-K related kinase family, is involved in the

non-homologous end joining (NHEJ) process that corresponds to the major activity responsible

for cell survival after ionizing radiation or chemotherapeutic treatment producing DNA double

strand breaks, and is believed to be a molecular sensor for DNA damage that enhances the

signal via phosphorylation of many downstream targets (74). Despite the function of DNA-PK

in the protection from genomic instability, its role in cancer prevention is still unclear (75).

2.6.3. Reversal of drug resistance in MMR-deficient cells by pharmacological blocking

cyto-protective signaling pathways

The other strategy was to achieve re-sensitization of MMR-deficient, Cisplatin-resistant colon

tumor cells through the concomitant downregulation of cytoprotective signaling pathways by

means of target-specific small molecule inhibitors against key factors of these pathways. It was

9

anticipated that this would overrule the drug-resistance phenotype. Two targets have been

focused on: the heat shock protein HSP90 and the Akt protein kinase.

HSP90 is a molecular chaperone required for the stress-survival response, protein

refolding, and the conformational maturation of a variety of signaling proteins. An association

of HSP90 overexpression and cancer has been proposed, and HSP90 inhibitors have antitumor

activity when given alone and in combination with cytotoxics (76,77). We were able to show

that the inhibition of the HSP90-function by Radicicol re-sensitizes (Cisplatin-resistant) MMR-

deficient cells to Cisplatin; but this re-sensitizing effect was also observed to even a greater

extent for the corresponding MMR-proficient cells (14). We therefore conclude that the HSP90

inhibitor Radicicol sensitizes cells to Cisplatin and that the presence of MLH1 even aggravates

this sensitizing, suggesting that HSP90-inhibition does indeed increase the cytotoxic effect of

Cisplatin but that it is not a suitable way to render in particular Cisplatin-resistant cells

hypersensitive to Cisplatin.

The Akt protein kinase transduces signals from growth factors and oncogenes to

downstream targets that control crucial elements in tumor development. The Akt pathway is one

of the most frequently hyperactivated signaling pathways in human cancers, and therefore

down-regulation of this pathway may be a viable strategy to prevent cancer (78,79). Rather

unexpectedly, we found that inhibition of the proliferative Akt-signaling pathway by the novel

Akt-specific inhibitor LY294005 decreased (rather than increased) Cisplatin and Lipoplatin

sensitivity in both MMR-deficient and MMR-deficient cells and that this effect occurred to a

greater extent and hence preferentially with MMR-deficient cells (15). Lipoplatin is a novel

liposomal formulation of Cisplatin to which MMR-deficient cells are also resistant (46). These

results appear not consistent with our concept of drug resistance reversal. Instead, a perhaps

novel function of the protein kinase Akt was uncovered (at least with platinum compounds)

which seems to increase rather than decrease cell survival. The molecular basis is still unclear

and is subject to further investigation in our laboratory.

2.6.4. Acquired resistance to HDAC-inhibitors and its prevention

With regard to the apparently only limited availability of clinically applicable strategies to

circumvent or even reverse the drug resistance phenotype seen in tumor cells, the prevention of

drug resistance development by anticancer agents is of increasing interest. In other words: the

generation of drug-resistant variants during the treatment of initially drug-sensitive tumor cells

with anticancer agents should be prevented (acquisition of resistance). This means that each

novel and potential anticancer agent should be subject to testing whether it has drug resistance-

causing potential. In this respect, we showed that the novel HDAC-inhibitor SAHA can induce

10

moderate and MDR-independent resistance in tumor cells (see PUBLICATION III; 18),

suggesting that SAHA is a potential inducer of drug resistance. In addition, these SAHA-

resistant cells were also cross-resistant to the HDAC-inhibitors TSA and VPA, but retained

sensitivity to conventional anticancer agents such as Docetaxel, Cisplatin, and Temozolomide.

These results expand on recent findings by two other groups that the HDAC-inhibitor FK-228

also induces resistance in tumor cells but in a reversible and MDR-dependent manner (19,20).

However, the molecular basis of this putatively novel type of acquired resistance is still unclear

and is thus subject to current investigation in our laboratory. Although not conclusive at this

point in terms of clinical relevance, this resistance-inducing potential of (at least some) HDAC-

inhibitors and the associated cross-resistance may be kept in mind prior the use in for cancer

treatment. At least, the data suggest that HDAC inhibitor-resistant tumors may be treated with

conventional anticancer agents. Moreover, the identification of HDAC-inhibitors (as well of

other anticancer agents) that do not cause development (acquisition) of resistance in tumor cells

is an important issue in this context.

11

3. Publication I Increased sensitivity of p53-deficient cells to anticancer agents due to loss of Pms2


Br. J. Cancer 2002;87:1027-1033 (Impact factor = 4.459), (ref 8)

12

Increased sensitivity of p53-deficient cells to anticancer agentsdue to loss of Pms2

A Fedier1, UB Ruefenacht1, VA Schwarz1, U Haller1 and D Fink*,1

1Department of Obstetrics and Gynaecology, Division of Gynaecology, University Hospital of Zurich, CH-8091, Switzerland

A large fraction of human tumours carries mutations in the p53 gene. p53 plays a central role in controlling cell cyclecheckpoint regulation, DNA repair, transcription, and apoptosis upon genotoxic stress. Lack of p53 function impairs thesecellular processes, and this may be the basis of resistance to chemotherapeutic regimens. By virtue of the involvement ofDNA mismatch repair in modulating cytotoxic pathways in response to DNA damaging agents, we investigated the effects ofloss of Pms2 on the sensitivity to a panel of widely used anticancer agents in E1A/Ha-Ras-transformed p53-null mousefibroblasts either proficient or deficient in Pms2. We report that lack of the Pms2 gene is associated with an increasedsensitivity, ranging from 2 – 6-fold, to some types of anticancer agents including the topoisomerase II poisons doxorubicin,etoposide and mitoxantrone, the platinum compounds cisplatin and oxaliplatin, the taxanes docetaxel and paclitaxel, and theantimetabolite gemcitabine. In contrast, no change in sensitivity was found after treatment with 5-fluorouracil. Cell cycleanalysis revealed that both, Pms2-deficient and -proficient cells, retain the ability to arrest at the G2/M upon cisplatintreatment. The data indicate that the concomitant loss of Pms2 function chemosensitises p53-deficient cells to some types ofanticancer agents, that Pms2 positively modulates cell survival by mechanisms independent of p53, and that increasedcytotoxicity is paralleled by increased apoptosis. Tumour-targeted functional inhibition of Pms2 may be a valuable strategy forincreasing the efficacy of anticancer agents in the treatment of p53-mutant cancers.British Journal of Cancer (2002) 87, 1027 – 1033. doi:10.1038/sj.bjc.6600599 www.bjcancer.comª 2002 Cancer Research UK

Keywords: drug sensitivity; p53; DNA mismatch repair; PMS2

The tumour suppressor gene p53 plays a central role in controllingcell cycle checkpoint regulation, DNA repair, transcription, andapoptosis upon genotoxic stress. p53 is one of the most frequentlymutated genes in human cancers and plays a critical role in theregulation of cell cycle and apoptosis (Levine, 1997). A numberof studies have suggested that loss of p53 function may be a majorreason underlying failure to respond to chemotherapy (Ferreira etal, 1999). Evidence for this notion is emerging from studies ofestablished human cancer cell lines and from knockout micemodels.

DNA mismatch repair (MMR) proteins play an important rolein the maintenance of genomic stability since it corrects replicativemismatches that escape DNA polymerase proofreading. Loss ofMMR results in genomic instability characterised by small insertionand deletion mutations in repetitive sequences throughout thegenome. As well as being involved in carcinogenesis, loss of theMMR activity is of concern with respect to the use of chemother-apeutic agents to treat established tumours. Loss of MMR has beenreported to cause resistance to some types of anticancer agentsincluding cisplatin (Fink et al, 1996) and the topoisomerase IIpoisons (Fedier et al, 2001). Interestingly, loss of MLH1 functionhas recently been reported to result in an increased sensitivity tocisplatin in p53-mutated human colorectal adenocarcinoma cells(Vikhanskaya et al, 1999; Lin et al, 2001). However, the cell linesused in these previous studies were not truly isogenic, since in

order to restore MMR activity, a whole copy of chromosome 3carrying a wild-type copy of MLH1 was inserted into theHCT116 cells. Thus, it is conceivable that the differences in cispla-tin sensitivity observed in these cell lines could have been due toone of the many genes on the inserted chromosome other thanthe wild-type copy of the missing MMR gene. So far no informa-tion is available on the effect of loss of PMS2, another MMR gene,to chemotherapeutic agents in p53-deficient cells.

Using transformed primary fibroblasts established from E1A/Ha-Ras-transfected knockout mice, we report here that p53-deficientcells are sensitised to some anticancer agents by the additional lossof the Pms2 gene, indicating that Pms2 positively regulates cellsurvival by a p53-independent mechanism. Tumour-targeted func-tional inhibition of Pms2 may thus be an adjunct to anticanceragents in the treatment of p53-mutant cancers.

MATERIALS AND METHODS

Cell lines

The Pms27/7/p537/7 and Pms2+/+/p537/7 cell lines, establishedfrom E1A/Ha-Ras-transformed knockout mice primary fibroblasts,were generously provided by Dr PM Glazer (Zeng et al, 2000).These primary fibroblasts have been produced by breeding miceheterozygous for Pms2 to either wild-type or p53-null mice, andthen by intercrossing progeny animals to generate mice eitherwild-type or null for Pms2 in a p53-nullizygous background. Thecells were maintained in DMEM medium supplemented with2 mM L-glutamine (Life Technologies, Basel, Switzerland), 10%

Exp

erim

enta

lT

her

apeu

tics

Received 7 May 2002; revised 15 August 2002; accepted 23 August 2002*Correspondence: D Fink; E-mail: [email protected]

British Journal of Cancer (2002) 87, 1027 – 1033

ª 2002 Cancer Research UK All rights reserved 0007 – 0920/02 $25.00

www.bjcancer.com

heat inactivated foetal calf serum (Oxoid, Basel, Switzerland) andpenicillin/streptomycin (100 U ml71/100 mg ml71, Life Technolo-gies) at 378C in a humidified atmosphere containing 5% carbondioxide. The Pms27/7/p537/7 cells as well as the Pms2+/+/p537/7

cells form defined individual colonies when seeded sparsely onstandard culture dishes. All the cell lines tested negative forcontamination with Mycoplasma spp.

Reagents

The following drugs were generous gifts: docetaxel (Aventis,Zurich, Switzerland), cisplatin, etoposide and paclitaxel (Bristol-Myers Squipp, Baar, Switzerland), mitoxantrone (Lederle, Zug,Switzerland), doxorubicin (Pharmacia & Upjohn, Dubendorf, Swit-zerland), 5-fluorouracil (Roche, Reinach, Switzerland), gemcitabine(Eli Lilly, Vernier, Switzerland), and oxaliplatin (Sanofi-Synthelabo,Meyrin, Switzerland).

Antiproliferative and cytotoxicity assays

Cisplatin, oxaliplatin, doxorubicin, mitoxantrone and 5-fluoroura-cil were dissolved immediately before use in 0.9% NaCl solution.Etoposide and paclitaxel were prepared in DMSO, whereas doce-taxel was dissolved in methanol. The final concentration ofDMSO or methanol in the cultures was 50.1% at all drug concen-trations and in controls. Previous experiments (data not shown)have shown that neither 0.1% DMSO nor 0.1% methanol affectsthe viability or growth of these cell lines.

The antiproliferative effect in response to drug treatment wasdetermined by the colorimetric MTT-assay (Mosmann, 1983).Briefly, cells growing in the log phase were harvested by brief tryp-sinisation and washed once with medium containing 10% foetalcalf serum. Using 96 well plates, 1000 cells were plated 24 h priorto incubation with or without the drug for 72 h at 378C in a humi-dified atmosphere containing 5% carbon dioxide. A volume of20 ml MTT in PBS to a final concentration of 0.5 mg ml71 wasadded, followed by incubation at 378C for 4 h, aspiration of themedium, and addition of 200 ml DMSO. Optical density wasmeasured by the Emax microplate reader E9336 (Molecular Devices,Clearwater, MN, USA) at 540 nm setting the value of the cell linesin medium to 1.0 (control) and the value of the no cells blank tozero. Differences in drug sensitivity of the respective cell lines weredetermined from at least four independent experiments withcontinuous drug exposure and are reported as the concentrationrequired to suppress proliferation by 50% (IC50).

In addition, one member of each drug class was tested for cyto-toxic effects by means of colony forming assay and trypan blueexclusion assay. Clonogenic survival in response to drug treatmentwas performed by plating 1000 cells in 60 mm cell culture dishes.After 24 h, the drug was added, followed by incubation for 6 daysat 378C in a humidified atmosphere containing 5% carbon dioxide.

Exp

erimen

talT

herap

eutics

Cisplatin Oxaliplatin

100

10

1

Cel

l gro

wth

(%

)

0 1 2 3 4 5 6µM µM

100

10

1

Cel

l gro

wth

(%

)

0 2.5 5 7.5 10

Doxorubicin Etoposide

100

10

1

Cel

l gro

wth

(%

)

100

10

1

Cel

l gro

wth

(%

)

0 20 40 60 80 0 100 200 300

100

10

1

Cel

l gro

wth

(%

)

0 15 30 45 60

100

10

1

Cel

l gro

wth

(%

)

0 1 2 3 4 5µM

nM

nM

nM

Docetaxel 5-Fluorouracil

Pms2+/+/p53–/–

Pms2–/–/p53–/–

Figure 1 Antiproliferative effect to a continuous exposure to cisplatin,oxaliplatin, doxorubicin, etoposide, docetaxel, and 5-fluorouracil forPms2+/+/p537/7 and Pms27/7/p537/7 cells as determined by theMTT-assay. Each point represents the mean+s.d. of at least four indepen-dent experiments.

Cisplatin Doxorubicin

µM

100

10

1

Per

cent

sur

viva

l

100

10

1

Per

cent

sur

viva

l

100

10

1

Per

cent

sur

viva

l

100

10

1

Per

cent

sur

viva

l

µM

Docetaxel 5-Fluorouracil

0 0.2 0.4 0.6 0.8 1 0 5 10 15 20nM

nM

0 5 10 15 0 0.2 0.4 0.6 0.8 1

Pms2+/+/p53–/–

Pms2–/–/p53–/–

Figure 2 Clonogenic survival curves in response to a continuous expo-sure to cisplatin, doxorubicin, docetaxel, and 5-fluorouracil for the Pms2+/

+/p537/7 and the Pms27/7/p537/7 cell lines. Each point represents themean+s.d. of at least four independent experiments.

Chemosensitivity in p53-deficient cells

A Fedier et al

1028

British Journal of Cancer (2002) 87(9), 1027 – 1033 ª 2002 Cancer Research UK

Cells were fixed with 25% acetic acid in ethanol and stained withGiemsa. Colonies of at least 50 cells were scored visually. Eachexperiment was performed a minimum of four times using tripli-cate cultures for each drug concentration. The logarithm of relativecolony formation was plotted against the concentration of thedrug. The IC50 was estimated by linear interpolation of the loga-rithmic transformed relative plating efficiencies. For trypan blueexclusion assay, cells were grown to 60% confluence and incubatedwith or without the drug for 24 h, 48 h or 72 h. At the time pointsindicated, floating and adherent cells were collected. Cells werethen incubated with trypan blue solution at 0.1% final concentra-tion for 1 min, and the number of trypan blue-positive and-negative cells was determined using a haematocytometer.

TUNEL apoptosis assay

Cells were grown to 70% confluence in 60 mm dishes in triplicatecultures and then treated with 5 mM cisplatin. Adherent and float-ing cells were collected at the time points indicated and washed inPBS. Cells were then fixed by dropwise addition of ice-cold 70%ethanol. Samples were stored at 48C until further use. Sampleswere washed twice with PBS and resuspended in the TUNELreaction mixture and incubated at 378C for 2 h, according to themanufacturer’s protocol (Roche Molecular Biochemicals, Basel,Switzerland). Cells were analysed by flow cytometry (EPICS ELITE,Beckmann-Coulter, Hialeah, FL, USA).

Cell cycle analysis

Cells were grown to 70% confluence in 60 mm dishes in triplicatecultures and then incubated with or without cisplatin (1.0 mM).This drug concentration induces about a 70% growth inhibitionof Pms2-deficient cells in the MTT-assay and about a 95% reduc-tion of clonogenic survival. Cells were then harvested by brieftrypsinisation and washed once in ice-cold PBS. The pellet wasresuspended in 200 ml PBS and cells were fixed by dropwise addi-tion of 4 ml ice-cold 70% ethanol and then stored at 48C until use.After removing ethanol by centrifugation at 30006g and washingtwice in PBS, cells were stained in 1 ml of propidium iodide stain-ing solution (50 mg ml71 propidium iodide and 100 U ml71

RNAse A in PBS) by incubation at room temperature for 60 minin the dark and then washed once in PBS containing 0.2% BSA.Samples were analysed for their DNA content by flow cytometry(EPICS ELITE, Beckmann-Coulter, Hialeah, FL, USA), and thepercentage of cells in each phase was determined using the Multi-Cycle for Windows Software (Phoenix Flow Systems, San Diego,CA, USA).

Statistical analysis

Mean+s.d. values were calculated for all data sets. The two-sidedpaired t test was used to compare the effects of loss of Pms2 func-tion on drug sensitivity. P50.05 was considered to be a statisticallysignificant difference.

RESULTS

Growth inhibition of Pms2+/+/p537/7 and Pms27/7/p537/7 cells after treatment with anticancer agents

We addressed the question as to whether the concomitant loss ofPms2 function in p53-deficient cells is associated with changes insensitivity to a panel of clinically relevant anticancer agents. Thepresence or absence of p53 and Pms2 in the respective cell lineswas verified by immunoblot analysis.

The antiproliferative effect of the platinum compounds cisplatinand oxaliplatin, the topoisomerase II poisons doxorubicin, mitox-

antrone and etoposide, the taxanes docetaxel and paclitaxel, andthe antimetabolites 5-fluorouracil and gemcitabine was tested incell lines proficient or deficient in Pms2 in a setting of p53 nulli-zygosity. Figure 1 shows the antiproliferative effect to a continuousexposure to cisplatin, oxaliplatin, doxorubicin, etoposide, docetaxeland 5-fluorouracil for Pms2+/+/p537/7 and Pms27/7/p537/7 cellsas determined by the MTT-assay. The Pms2-deficient Pms27/7/p537/7 cells were 6.4-fold more sensitive to cisplatin(0.7+0.3 mM vs 4.7+0.6 mM, P=0.0002), 2.5-fold more sensitiveto oxaliplatin (1.4+0.2 mM vs 3.5+1.1 mM, P=0.03), 5.3-fold moresensitive to doxorubicin (13.9+2.2 nM vs 73.4+18.8 nM,P=0.002), 3.0-fold more sensitive to mitoxantrone (2.7+0.8 nM

vs 8.1+1.8 nM, P=0.005), 6.1-fold more sensitive to etoposide(17+3 nM vs 105+30 nM, P=0.01), 4.8-fold more sensitive todocetaxel (13+2 nM vs 64+13 nM, P=0.0005), 4.3-fold moresensitive to paclitaxel (15+4 nM vs 66+14 nM, P=0.004), and

Exp

erim

enta

lT

her

apeu

tics

Cisplatin

90

80

70

60

50

40

30

20

10

0

% C

ell k

ill

0 24 48 72

Time (h)

Doxorubicin

90

80

70

60

50

40

30

20

10

0

% C

ell k

ill

0 24 48 72

Time (h)

5-Fluorouracil

90

80

70

60

50

40

30

20

10

0

% C

ell k

ill

0 24 48 72

Time (h)

Figure 3 Sensitivity to cell kill of Pms27/7/p537/7 (open field)and Pms2+/+/p537/7 (closed field) cells in response to cisplatin (3 mM),doxorubicin (50 nM) or 5-fluorouracil (5 mM) as a function of time deter-mined by trypan blue exclusion. Each point represents the mean+s.d. ofat least four independent experiments.


A Fedier et al

1029

ª 2002 Cancer Research UK British Journal of Cancer (2002) 87(9), 1027 – 1033

Exp

erimen

talT

herap

eutics

396

330

264

196

132

66

00 32 64 96 128 160 192 224 256

Cel

l num

ber

DNA content

286

238

190

143

95

47

0

400

333

266

200

133

66

0

399

333

266

199

133

66

0

191

159

127

95

63

31

0

209

174

139

104

69

34

0

176

147

117

88

58

29

0

342

285

228

171

114

57

0

Cel

l num

ber

Cel

l num

ber

Cel

l num

ber

Cel

l num

ber

Cel

l num

ber

Cel

l num

ber

Cel

l num

ber

0 32 64 96 128 160 192 224 256

DNA content

0 32 64 96 128 160 192 224 256

DNA content

0 32 64 96 128 160 192 224 256

DNA content

0 32 64 96 128 160 192 224 256

DNA content

0 32 64 96 128 160 192 224 256

DNA content

2 34 66 98 130 162 194 226 258

DNA content

0 32 64 96 128 160 192 224 256

DNA content

0 h

24 h

48 h

72 h

2N 4N

Pms2–/–/p53–/–

2N 4N

Pms2+/+/p53–/–

Figure 4 Representative cell cycle phase distribution profile of the DNA content for Pms2+/+/p537/7 cells (left panel) and Pms27/7/p537/7 cells (rightpanel) as a function of time (0 h, 24 h, 48 h, 72 h) in response to treatment with 1.0 mM cisplatin. 2N represent cells accumulated in G1, 4N represent cellsaccumulated in G2/M.


A Fedier et al

1030


5.5-fold more sensitive to gemcitabine (39+21 nM vs 213+65 nM,P=0.001) as compared to the Pms2+/+/p537/7 cells. In contrast, nodifference in sensitivity was found for 5-fluorouracil (1.5+0.2 mM

vs 1.7+0.6 mM, P=0.43). Thus, our data indicate that the concomi-tant loss of Pms2 in p53-deficient cells results in a hypersensitivityto platinum compounds, topoisomerase II poisons, taxanes, andgemcitabine.

Clonogenic cell survival and trypan blue exclusion assays

Colorimetry-based short-term cytotoxicity assays such as theMTT-assay may underestimate overall cell killing owing to thenotion that cells may not die immediately after treatment butmay instead remain in a transient state of arrest for several daysprior to dying (Brown and Wouters, 1999). Therefore, the clono-genic assay was performed for one member of each drug class inaddition to the MTT-assay. Figure 2 shows the survival curves forthe Pms2-proficient Pms2+/+/p537/7 cell line as well as therepair-deficient Pms27/7/p537/7 cell line as a function of drugconcentration. In comparison to the Pms2+/+/p537/7 cell linethe Pms2-deficient Pms27/7/p537/7 cell line was 1.6-fold moresensitive to cisplatin (0.68+0.10 mM vs 0.41+0.07 mM, P=0.002),1.7-fold more sensitive to doxorubicin (12.1+1.4 nM vs7.3+0.4 nM, P=0.004) and 1.8-fold more sensitive to docetaxel(7.5+1.0 nM vs 4.0+0.2 nM, P=0.005). In contrast, no differencein sensitivity was observed in response to 5-fluorouracil(0.57+0.06 mM vs 0.63+0.07 mM, P=0.25). Thus, the clonogenicsurvival data support the results observed in the MTT-assay,indicating that loss of Pms2 function in p53-deficient cellsdecreases clonogenic survival to these agents.

The trypan blue exclusion assay demonstrates that Pms2-defi-cient cells have a substantially higher percentage of trypan blue-stained cells than Pms2-proficient cells in response to equi-antipro-liferative doses (i.e. doses that inhibited proliferation by 90% inPms2-deficient cells) of doxorubicin (50 nM) or cisplatin (3 mM)at each time point after treatment (Figure 3). This effect, however,is not observed in response to 5-fluorouracil (5 mM). These dataindicate that Pms2 deficiency reduces the threshold for cell killto doxorubicin and to cisplatin, but not to 5-fluorouracil. Thus,the higher sensitivity to cell kill of Pms2-deficient cells to doxoru-bicin or cisplatin confirms the increased antiproliferative effect andthe decreased clonogenic survival of these cells to these agents.

Pms2-deficiency and cisplatin-induced apoptosis

Using cisplatin as a representative of compounds that displayhypersensitivity in clonogenic survival as well as in proliferationinhibition, the TUNEL apoptosis assay was performed to determinewhether the increased sensitivity to cisplatin due to loss of Pms2function was accompanied by increased apoptosis. Treatment with5 mM cisplatin, which inhibited growth of Pms2-deficient cells bymore than 95%, produced more apoptotic Pms27/7/p537/7 cellsthan Pms2+/+/p537/7cells. The respective values were 37+5% vs5+1% at 30 h after cisplatin treatment and 43+4% vs 20+14%at 60 h after treatment. These results indicate that the hypersensi-tivity to the cytotoxic effect of cisplatin in Pms2-deficient cells isassociated with increased apoptosis.

Cell cycle analysis of Pms2+/+/p537/7 andPms27/7/p537/7 cells after treatment with cisplatin

The question was addressed as to whether the increased sensitivityto, for instance, cisplatin in p53-deficient cells is accompanied byalterations in triggering cell cycle checkpoint activation. Cisplatinhas been shown to be a potent inducer of MMR-dependentG2/M arrest (Brown et al, 1997). Figure 4 shows that 1 mM

cisplatin, which inhibited growth of Pms2-deficient cells by 70%

in the MTT-assay, resulted in response to cisplatin in a sustainedaccumulation at the G2/M transition of Pms2-proficient and-deficient cells. The data demonstrate that both Pms2-proficientand -deficient cells retain the ability to arrest at the G2/Mupon cisplatin treatment. Thus, the observed hypersensitivity inPms2-deficient cells seems not to be accompanied by alterationsin the characteristic changes in cell cycle distribution profileinduced by drug treatment.

DISCUSSION

The present study indicates that p53-null mouse fibroblasts arerendered hypersensitive to cell killing in response to platinumcompounds, topoisomerase II poisons, taxanes and the antimetabo-lite gemcitabine by the concomitant loss of Pms2 function, andthat both Pms2-deficient and -proficient cells retain the ability toarrest at G2/M upon cisplatin treatment. This observation is impor-tant for several reasons: First, it identifies PMS2 as another putativeprotective mediator of cell survival in p53-deficient cells, and itthus expands on the previous finding reporting increased sensitivityto cisplatin by the concomitant loss of MLH1 function (Vikhans-kaya et al, 1999; Lin et al, 2001). Second, it identifies, inaddition to the platinum compounds, two other clinically impor-tant classes of anticancer agents that result in hypersensitivity inp53-deficient cells by the additional loss of MMR. Third, it furthersupports the notion that MMR proteins such as PMS2 or MLH1(Vikhanskaya et al, 1999; Lin et al, 2001) modulate protectiveDNA damage response pathways independently of p53 function.This finding supports the concept of a novel role of MMR, in addi-tion to mediating drug-induced cytotoxicity (Fink et al, 1998;Fedier et al, 2001), as a positive modulator of cell survival afterDNA damage. A precedent for this hypothesis has already beenfound as it has been reported that MLH1-deficiency is associatedwith hypersensitivity to mitomycin C in p53-proficient cells(Fiumicino et al, 2000). Fourth, the observed hypersensitivity ofPms2-deficient cells seems to be independent of the cell cycleand seems to be due to increased apoptosis. The observed p53-independence of cytotoxicity goes along with the previous observa-tion that MMR can trigger apoptosis in a p53-independentpathway (Zeng et al, 2000).

Mutations that disable p53 are frequently found in humancancers (Hollstein et al, 1991), often in association with tumourprogression or high grade malignancy (Carder et al, 1993). There-fore, much effort has gone into determining the effects of p53inactivation on the response of cancer cells to therapeutic agents.The results vary with the cell line under study and the experimentalset-up, and thus both decreased and increased sensitivities havebeen observed in different model systems (Blandino et al, 1999;Bunz et al, 1999). Recently, it has been reported that loss ofMLH1 is associated with a hypersensitivity to cisplatin in p53-defi-cient cells in the HCT116 model (Vikhanskaya et al, 1999; Lin et al,2001). However, the repair-deficient member of these cell lines wasnot truly isogenic since in order to restore MMR activity a wholecopy of chromosome 3 carrying a wild-type copy of MLH1 wasinserted into the HCT116 cells (Koi et al, 1994). The present studyuses isogenic primary fibroblasts established from E1A/Ha-Ras-transfected knockout mice. These cells are genetically well definedand therefore constitute a more meaningful test system thancancer-derived cell lines, not only because they are truly isogenicbut also because the cancer-derived cells are likely to contain anumber of other accumulated mutations and abnormalities thatcould influence response to anticancer agents. It is unlikely thatthe hypersensitivity seen in the Pms27/7/p537/7 cells was dueto the experimental set-up rather than being caused by the lackof Pms2 function, because the increased sensitivity was alsoobserved, though to a lesser extent, in the clonogenic survival assayand in the trypan blue exclusion assay. This observation indicates

Exp

erim

enta

lT

her

apeu

tics


A Fedier et al

1031


that the increased sensitivity to the platinum compounds, thetopoisomerase II poisons and the taxanes arises through enhancedcell killing rather than only through decreased proliferation ability.The increased cell killing by cisplatin in cells lacking Pms2 is paral-leled by increased apoptosis.

The mechanism by which the MMR proteins may influencedamage response is not yet fully understood. One hypothesisproposes that the MMR proteins recognise base damage andinitiate a cycle of futile repair, leading to gaps and breaks thatmay ultimately signal apoptosis (Vaisman et al, 1998). It is alsopossible that the recognition of damage by the MMR proteinsdirectly initiate a signal transduction pathway (Fink et al,1998). Direct evidence supporting the latter hypothesis includesa requirement for MLH1 function in cisplatin induction of c-abl kinase activity (Nehme et al, 1997) and of p73 accumulation(Gong et al, 1999).

An increased cytotoxic effect due to loss of Pms2 was observedin p53-deficient cells to a panel of drugs that have differentmodes of action and introducing different types of damage. Themolecular basis for the putative protective role of MMR in thep53-independent response to platinum compounds, topoisomeraseII poisons and to taxanes is not yet clear. Loss of either Pms2 orMLH1 (Vikhanskaya et al, 1999; Lin et al, 2001) results in p53-deficient cells in hypersensitivity to cisplatin, whereas loss ofMLH1 or PMS2 results in p53-proficent cells in resistance tocisplatin (Aebi et al, 1996; Fink et al, 1997). Likewise, cell killby topoisomerase II poisons, but not by taxanes, has beenreported to be affected by loss of MMR (Fedier et al, 2001; Linet al, 2001), indicating that MMR-dependent damage responseis, at least partially, modulated by p53. Indeed, cross talks betweenMMR- and p53-dependent pathways have been reported (Duckettet al, 1999; Wu et al, 1999). Cells defective in p53 function retainthe ability to mediate apoptosis by a p73-dependent pathway(Gong et al, 1999), which may be modulated by certain MMRproteins. It is possible that PMS2 and MLH1 protect cells fromexcessive cell death by counteracting p73-mediated apoptosis ina MMR-dependent manner upon DNA damage introduced bycisplatin or topoisomerase II poisons. In a p53-independent path-way PMS2 may, as in the case of cisplatin, decrease adducttolerance and damage accumulation on this basis. Oxaliplatinand cisplatin produce different types of DNA damage and thushave been shown to display a differential response in MMR-defi-cient cells proficient for p53 (Fink et al, 1996). Our data thussuggest a role for PMS2 in a DNA damage response pathway inaddition to that in MMR. However, the possibility can not beexcluded that p53 modulates oxaliplatin damage differently fromthat of cisplatin damage. Likewise, PMS2 may modulate the effi-cacy of the repair machinery to process stalled replication forksarisen by blocked DNA-topoisomerase intermediates and gapsintroduced by the chain-terminating antimetabolite gemcitabine.Additional loss of PMS2 thus abolishes the damage removal activ-ity, giving rise to excessive DNA damage and cytotoxicity. The

finding that MMR-deficiency does not alter 5-fluorouracil cyto-toxicity in a p53-deficient setting may be of clinical interest forthe treatment of colorectal cancer, despite the previously reportedin vitro association of MMR with resistance to 5-fluorouracil inp53-proficient cells (Carethers et al, 1999).

p53 either mediates growth arrest, both in G1 or G2 phases ofthe cell cycle, or directs cells to apoptosis. These two cellular deci-sions are distinctive end points of p53 induction, depending on thecellular context and the type of DNA damage. The present studydemonstrates that both Pms2-proficient and -deficient cells retainthe ability to arrest at the G2/M upon cisplatin treatment and thatloss of Pms2 is not accompanied by substantial alterations in thecharacteristic changes in cell cycle distribution. A similar resulthas also been reported in p53-deficient cells after loss of Msh2(Strathdee et al, 2001) or MLH1 (Lin et al, 2001). Therefore, itseems unlikely that MMR is a major trigger for a G2/M arrest incells that lack functional p53.

The feature that p53 and MMR proteins modulate cellularresponses upon DNA damage and the availability of geneticallyengineered human and murine cells present a potential means todevelop strategies to circumvent the reduced responsiveness byre-sensitising or by hypersensitising p53-mutant tumours to thera-peutic regimens. Candidate approaches include the restoration ofthe wild-type p53 function, the activation of cytotoxic pathwaysthat operate independently of p53, or the concurrent disruptionof genes implicated in DNA damage response pathways (Mcgilland Fisher, 1999).

Since both, MMR and p53 status affect the mechanism of cyto-toxicity, the genotype-based predictions may require that the MMRstatus as well as the p53 status of the tumour are taken intoaccount. In summary, the present data show that p53-deficient cellsare sensitised to the platinum compounds, the topoisomerase IIpoisons and the taxanes by the concurrent loss of Pms2 function.Although PMS2 mutations or mutations in both PMS2 and p53 arenot frequently found in human cancers and thus may be of minorclinical importance, our study may nevertheless contribute tofostering the concept that tumour-targeted functional inhibitionof PMS2 may be an adjunct to chemotherapy in the treatmentof tumours unresponsive to therapeutic regimens due to mutationsin the p53 gene.

ACKNOWLEDGEMENTS

The authors are grateful to Eva Niederer (Institute for BiomedicalEngineering, ETH/University of Zurich, Switzerland) for assistancein cell cycle analysis and Dr PM Glazer (Yale University School ofMedicine, New Haven, CT, USA) for kindly providing the celllines. This work has been supported by a grant from the SwissNational Science Foundation (No. 31-52531.97) and the HermannKlaus Foundation Zurich.

REFERENCES

Aebi S, Kurdi-Haidar B, Gordon R, Cenni B, Zheng H, Fink D, Christen RD,Boland CR, Koi M, Fishel R, Howell SB (1996) Loss of DNA mismatchrepair in acquired resistance to cisplatin. Cancer Res 56: 3087 – 3090

Blandino G, Levine AJ, Oren M (1999) Mutant p53 gain of function: differ-ential effects of different p53 mutants on resistance of cultured cells tochemotherapy. Oncogene 18: 477 – 485

Brown R, Hirst GL, Gallagher WM, McIlwrath AJ, Margison GP, van der ZeeAG, Anthoney DA (1997) hMLH1 expression and cellular responses ofovarian tumour cells to treatment with cytotoxic anticancer agents. Onco-gene 15: 45 – 52

Brown MJ, Wouters BG (1999) Apoptosis, p53, and tumor cell sensitivity toanticancer agents. Cancer Res 59: 1391 – 1399

Bunz F, Hwang PM, Torrance C, Waldman T, Zhang Y, Dillehay L, WilliamsJ, Lengauer C, Kinzler KW, Vogelstein B (1999) Disruption of p53 inhuman cancer cells alters the responses to chemotherapeutic agents. J ClinInvest 104: 263 – 269

Carder P, Wyllie AH, Purdie CA, Morris RG, White S, Piris J, Bird CC (1993)Stabilised p53 facilitates aneuploid clonal divergence in colorectal cancer.Oncogene 8: 1397 – 1401

Carethers JM, Chauhan DP, Fink D, Nebel S, Bresalier RS, Howell SB, BolandCR (1999) Mismatch repair proficiency and in vitro response to 5-Fluor-ouracil. Gastroenterology 117: 123 – 131

Exp

erimen

talT

herap

eutics


A Fedier et al

1032


Duckett DR, Bronstein SM, Taya Y, Modrich P (1999) hMutSa- andhMutLa-dependent phosphorylation of p53 in response to DNA methyla-tor damage. Proc Natl Acad Sci USA 96: 12384 – 12388

Fedier A, Schwarz VA, Walt H, Delli Carpini R, Haller U, Fink D (2001)Resistance to topoisomerase poisons due to loss of DNA mismatch repair.Int J Cancer 93: 571 – 576

Ferreira CG, Tolis C, Giaccone G (1999) p53 and chemosensitivity. AnnOncol 10: 1011 – 1021

Fink D, Nebel S, Aebi S, Zheng H, Cenni B, Nehme A, Christen R, Howell SB(1996) The role of DNA mismatch repair in platinum drug resistance.Cancer Res 56: 4881 – 4886

Fink D, Nebel S, Aebi S, Nehme A, Howell SB (1997) Loss of DNA mismatchrepair due to knockout of MSH2 or PMS2 results in resistance to cisplatinand carboplatin. Int J Oncol 11: 539 – 542

Fink D, Aebi S, Howell SB (1998) The role of DNA mismatch repair in drugresistance. Clin Cancer Res 4: 1 – 6

Fiumicino S, Martinelli S, Colussi C, Aquilina G, Leonetti C, Crescenzi M,Bignami M (2000) Sensitivity to DNA cross-linking chemotherapeuticagents in mismatch repair-defective cells in vitro and in xenografts. Int JCancer 85: 590 – 596

Gong JG, Costanzo A, Yang HO, Melino G , Kaelin Jr WG, Levrero M, WangJY (1999) The tyrosine kinase c-Abl regulates p73 in apoptotic response tocisplatin-induced DNA damage. Nature 399: 806 – 809

Hollstein M, Sidransky B, Vogelstein B, Harris CC (1991) p53 mutations inhuman cancers. Science 253: 49 – 53

Koi M, Umar A, Chauhan DP, Cherian SP, Carethers JM, Kunkel TA, BolandCR (1994) Human chromosome 3 corrects mismatch repair deficiency andmicrosatellite instability and reduces N-methyl-N’-nitro-N-nitrosoguani-dine tolerance in colon tumor cells with homozygous hMLH1 mutation.Cancer Res 54: 4308 – 4312

Levine AJ (1997) p53, the cellular gatekeeper for growth and division. Cell 88:323 – 331

Lin X, Ramamurthi K, Mishima M, Kondo A, Christen RD, Howell SB (2001)p53 modulates the effect of loss of DNA mismatch repair on the sensitivityof human colon cancer cells to the cytotoxic and mutagenic effects ofcisplatin. Cancer Res 61: 1508 – 1516

McGill G, Fisher DE (1999) p53 and cancer therapy: a double-edged sword. JClin Invest 104: 223 – 225

Mosmann T (1983) Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J Immunol Methods 65:55 – 63

Nehme A, Baskaran R, Aebi S, Fink D, Nebel S, Cenni B, Wang JYJ, HowellSB, Christen RD (1997) Differential induction of c-Jun NH2-terminalkinase and c-Abl kinase in DNA mismatch repair-proficient and -deficientcells exposed to cisplatin. Cancer Res 57: 3253 – 3257

Strathdee G, Samson OJ, Sim A, Clarke AR, Brown R (2001) A role formismatch repair in control of DNA ploidy following DNA damage. Onco-gene 20: 1923 – 1927

Vaisman A, Varchenko M, Umar A, Kunkel TA, Risinger JI, Barrett JC,Hamilton TC, Chaney SG (1998) The role of hMLH1, hMSH3, andhMSH6 defects in cisplatin and oxaliplatin resistance: correlation withreplicative bypass of platinum-DNA adducts. Cancer Res 58: 3579 – 3585

Vikhanskaya F, Golella G, Valenti M, Parodi S, d’Incalci M, Broggini M(1999) Cooperation between p53 and hMLH1 in a human colocarcinomacell line in response to DNA damage. Clin Cancer Res 5: 937 – 941

Wu JX, Gu LY, Wang HX, Geactinov NE, Li GM (1999) Mismatch repairprocessing of carcinogen-DNA adducts triggers apoptosis. Mol Cell Biol19: 8292 – 8301

Zeng M, Narayanan L, Xu XS, Prolla TA, Liskay MR, Glazer PM (2000)Ionizing radiation-induced apoptosis via separate pms2- and p53-depen-dent pathways. Cancer Res 60: 4889 – 4893

Exp

erim

enta

lT

her

apeu

tics


A Fedier et al

1033


4. Publication II Loss of atm sensitises p53-deficient cells to topoisomerase poisons and antimetabolites


Ann. Oncol. 2003;14:938-945 (Impact factor = 5.179) (ref 11)

13

Annals of Oncology 14: 938–945, 2003

Original article DOI: 10.1093/annonc/mdg240

© 2003 European Society for Medical Oncology

Loss of atm sensitises p53-deficient cells to topoisomerase poisons and antimetabolitesA. Fedier1, M. Schlamminger1, V. A. Schwarz1, U. Haller1, S. B. Howell2 & D. Fink1*

*Correspondence to: Dr D. Fink, Department of Obstetrics and Gynaecology, University of Zurich, CH-8091 Zurich, Switzerland. Tel: +41-1-255-5327; Fax: +41-1-255-4553; E-mail: [email protected]

1Department of Obstetrics and Gynaecology, Division of Gynaecology, University Hospital of Zurich, Switzerland; 2Department of Medicine and the Cancer Centre, University of California at San Diego, La Jolla, CA, USA

Received 10 January 2003; revised 7 February 2003; accepted 17 February 2003

Background: Ataxia-telangiectasia is a pleiotropic autosomal recessive disorder caused by mutations in the

ATM gene. In addition to a profound cancer predisposition, another hallmark of ataxia-telangiectasia is radio-

sensitivity. Recently, p53-null mouse fibroblasts have been reported to be radiosensitised by the concurrent loss

of ATM.

Materials and methods: We compared the sensitivity of atm+/+/p53–/– and atm–/–/p53–/– mouse embryonic

fibroblasts to different classes of chemotherapeutic agents using the MTT assay, Trypan Blue exclusion and

fluorescence-activated cell sorting for cell cycle and apoptosis analyses.

Results: Loss of ATM function in p53-deficient cells resulted in a 2- to 4-fold increase in sensitivity to the

topoisomerase I poisons camptothecin and topotecan, to the topoisomerase II poisons doxorubicin, epirubicin

and etoposide, and to the antimetabolites 5-fluorouracil and gemcitabine, but not to the platinum compounds

cisplatin, carboplatin and oxaliplatin, the taxanes docetaxel and paclitaxel, or to busulfan. Loss of ATM

function did not result in increased apoptosis, but resulted in increased Trypan Blue staining in response to

epirubicin, suggesting that processes other than apoptosis may mediate cytotoxicity. ATM deficiency did not

alter the extent of G1/S or G2/M cell cycle phase accumulation produced by epirubicin, suggesting that enhanced

sensitivity was not due to failure of checkpoint activation.

Conclusions: We provide further evidence that ATM is involved in regulating cellular defences against some

cytotoxic agents in the absence of p53. Tumour-targeted functional inhibition of ATM may be a valuable

strategy for increasing the efficacy of anticancer agents in the treatment of p53-mutant cancers.

Key words: antimetabolites, ATM, cancer, drug sensitivity, p53, topoisomerase poisons

Introduction

The p53 tumour suppressor gene is the most frequently mutatedgene in human cancers [1]. p53 is involved in regulating cell cyclecheckpoints, apoptosis and DNA repair in response to DNAdamage, and thus plays a pivotal role in maintaining the integrityof the genome [2]. Lack of functional p53 is associated withabrogation of cell cycle checkpoint control and of apoptosisinduction. This contributes to resistance to radiotherapy and tocertain chemotherapeutic agents in some tissues [3]. The failure ofp53-mutant tumours to respond to therapeutic regimens is ofincreasing concern. Strategies that cause cells to override theDNA damage checkpoints, to block protective pathways or totrigger apoptosis via p53-independent pathways are predicted tosensitise cells to killing by genotoxic agents.

ATM is mutated in patients with ataxia-telangiectasia, a pleio-tropic autosomal recessive disorder characterised by progressiveneurodegeneration, premature senescence, immunodeficiency,predisposition to cancer, chromosomal instability and hyper-

sensitivity to γ-irradiation [4, 5]. ATM is a PI-3-like protein kinaseoperating upstream of p53 [6] and it has been shown to bind tofree DNA ends produced by DNA double-strand breaks that occurduring normal replication and recombination, or in response toexogenous genotoxic stress [7]. It is known to interact with avariety of other targets in addition to p53, including c-Abl,BRCAl, CHK2, MDM2 and DNA-PK [8, 9]. ATM plays a keyrole in sensing DNA damage and in propagating signals thatmodulate protective cellular responses to genotoxic agents. Itserves a surveillance function that helps maintain genomic integ-rity by promoting cell cycle arrest and damage repair, and possiblyby recruiting repair proteins to the site of damage to preventdouble-strand break repair from entering an error-prone pathway[10]. Ataxia-telangiectasia cells display defective p53 induction,abrogation of G1/S and G2/M cell cycle checkpoints, and hyper-sensitivity to γ-irradiation [5], suggesting that ATM and p53interact in a common pathway in response to this type of DNAdamage. However, some p53-null mouse tissues have been shownto be rendered radiosensitive by the concurrent loss of the ATMgene [11], suggesting the existence of an ATM effector pathwaythat is activated in response to ionising radiation but which doesnot depend on p53.

939

Because a very large fraction of human cancers is functionallyp53 deficient, and hence may be resistant to chemotherapeuticagents, a means of sensitising p53-deficient tumours assumesgreat importance. Therefore, we were interested in determiningwhether p53-deficient cells could be sensitised to a panel of clinic-ally important chemotherapeutic agents by the additional lossof ATM function. We report here that loss of ATM function inp53-null mouse embryonic fibroblasts results in hypersensitivityto a panel of chemotherapeutic agents, indicating that ATM playsa role in protective responses to DNA damage independently ofp53. The increased drug sensitivity of ATM-deficient cells wasnot accompanied by increased apoptosis or by further alteration ofcell cycle checkpoint activation. Our data support the concept thattumour-targeted functional inhibition of ATM may be a valuableapproach to improving the effectiveness of chemotherapeuticagents in the treatment of p53-mutant cancers.

Materials and methodsCell lines and culture conditions

The atm+/+/p53–/– and atm–/–/p53–/– mouse embryonic fibroblasts were gener-ously provided by Dr P. Leder. Mice were generated by crossing p53–/– micein a FVB background to atm–/– mice [12]. The cells were maintained inDulbecco’s modified Eagle’s medium supplemented with 4 mM L-glutamine(Life Technologies, Basel, Switzerland), 15% heat-inactivated fetal calf serum(Oxoid, Basel, Switzerland) and penicillin/streptomycin (100 U/ml, 100 µg/ml;Life Technologies) at 37°C in a humidified atmosphere containing 5% carbondioxide. The human breast adenocarcinoma cell line MCF-7 was obtainedfrom the American Type Culture Collection (ATCC HTB 22; Manassas, VA,USA) and maintained in Opti-MEM (Life Technologies supplemented with10% heat-inactivated fetal calf serum. All cell lines tested negative for con-tamination with Mycoplasma spp.

Drugs

Camptothecin and busulfan were purchased from Sigma (Buchs, Switzerland).The following drugs were generous gifts: epirubicin and doxorubicin (Pharmacia& Upjohn, Dubendorf, Switzerland), topotecan (SmithKline Beecham, Thoris-haus, Switzerland), docetaxel (Aventis, Zurich, Switzerland), oxaliplatin (Sanofi-Synthelabo, Meyrin, Switzerland), etoposide, paclitaxel, cisplatin and carbo-platin (Bristol-Myers Squibb, Baar, Switzerland), 5-fluorouracil (Roche, Reinach,Switzerland) and gemcitabine (Eli Lilly, Vernier, Switzerland).

Immunoblot analysis

To provoke p53 induction, exponentially growing cells were exposed to25 nM epirubicin for 24 h and then collected as described. Cells were washedin cold phosphate-buffered saline (PBS) and lysed on ice in 150 mM NaClcontaining 5 mM EDTA, 1% Triton X-100, 10 mM Tris–HCl (pH 7.4), 5 mMdithiothreitol (DTT), 100 µg/ml phenylmethylsulfonyl fluoride and 1 µg/mlaprotinin, followed by centrifugation at 14 000 g for 20 min at 4°C. The proteinamount was determined using the Bio-Rad protein assay dye (Bio-Rad,Glattbrugg, Switzerland). After centrifugation, 150 µg (for ATM) or 50 µg(for p53) of protein was boiled in an equal volume of 100 mM Tris–HCl(pH 6.8) containing 20% glycerol, 200 mM DTT, 4% sodium dodecylsulphate(SDS) and 0.2% bromophenol blue. The proteins were separated using SDS–PAGE on a 5% gel for ATM and on a 10% gel for p53 analysis, followed byblotting onto a polyvinylidene difluoride membrane (Amersham PharmaciaBiotech, Little Chalfont, UK). ATM protein was detected using a mousemonoclonal antibody directed against amino acids 2577–3056 (ATM-2CI;GeneTex, Inc., San Antonio, TX, USA), whereas p53 protein was detected

using the mouse monoclonal antibody pAb 240 (Santa Cruz BiotechnologyInc., Basel, Switzerland). After washing the blots, horseradish peroxidase-conjugated antimouse antibody (BD Biosciences, Allschwil, Switzerland) wasadded and the complexes were visualised by enhanced chemiluminescence(Amersham Pharmacia Biotech).

Growth inhibition assay

Epirubicin, doxorubicin, topotecan, cisplatin, carboplatin, oxaliplatin,5-fluorouracil and gemcitabine were diluted in 0.9% NaCl immediatelybefore use. Stock solutions of etoposide, camptothecin, paclitaxel and busul-fan were prepared in dimethyl sulphoxide (DMSO), whereas docetaxel wasdissolved in methanol. The final concentration of DMSO or methanol inthe cultures was <0.1% at all drug concentrations and in controls. Previousexperiments (data not shown) have established that neither 0.1% DMSOnor 0.1% methanol affects the viability or growth of these cell lines. Growthinhibition in response to drug treatment was determined by the MTT assay[13]. Cells growing in the log phase were harvested by brief trypsinisation andwashed once with medium containing 15% fetal calf serum. MTT assays wereperformed by seeding 500 (atm–/–/p53–/–) or 1000 (atm+/+/p53–/–) cells into96-well plates 24 h before incubation without or with the drug for 5 days. Avolume of 20 µl MTT in PBS was added to a final concentration of 0.5 mg/ml,followed by incubation at 37°C for 4 h, aspiration of the medium and additionof DMSO 200 µl. Optical density was measured by the Emax microplatereader E9336 (Molecular Devices, Clearwater, MN, USA) at 540 nm, settingthe value of the cell lines in medium to 1.0 (control) and the value of the nocells blank to zero. Differences in drug sensitivity of the respective cell lineswere determined from at least four independent experiments and are reportedas the concentration required to suppress growth by 50% (IC50).

TUNEL apoptosis assay

Cells were grown to 70% confluence in 60 mm dishes in triplicate cultures andthen treated with 25 nM epirubicin. Adherent and floating cells were collectedat the time points indicated and washed in PBS. Cells were then fixed in 4%paraformaldehyde on ice for 30 min, followed by dropwise addition of ice-cold 70% ethanol. Samples were stored at 4°C until further use. Samples werewashed twice with PBS and resuspended in the TUNEL reaction mixture andincubated at 37°C for 2 h, according to the manufacturer’s protocol (RocheMolecular Biochemicals, Basel, Switzerland). Cells were analysed by flowcytometry (EPICS ELITE; Beckmann-Coulter, Hialeah, FL, USA).

Trypan Blue exclusion assay

Cell viability after drug treatment was determined by means of the TrypanBlue exclusion assay. Cells were grown to 60% confluence and incubatedwithout (controls) or with 25 nM epirubicin for 24, 48 or 72 h. At the indicatedtime points, floating and adherent cells were collected. Cells were thenincubated with Trypan Blue solution (0.1% final concentration) for 1 min, andthe number of Trypan Blue-positive and -negative cells was determined usinga haematocytometer.

Cell cycle analysis

Cells were grown to 50% confluence in 60 mm dishes. Cells were then incu-bated with or without (controls) epirubicin 10 nM. This dose inhibited growthof ATM-deficient cells by at least 70% in the MTT assay. Before collection bytrypsinisation at the times indicated, cells were incubated with BrdU 10 µM(Serva, Heidelberg, Germany) at 37°C for 8 h. Samples were washed in ice-cold PBS, resuspended in PBS 100 µl, fixed by dropwise addition of ice-cold70% ethanol and stored at 4°C until use. Then, ethanol was removed bycentrifugation (3000 g) and cells were resuspended and incubated in 2 N HCl/0.5% Triton X-100 for 30 min at room temperature, followed by centrifugation(3000 g) and resuspension in 0.1 M Na2B4O7 to neutralise the acid. After

940

collection of the pellet by centrifugation, cells were resuspended and incu-bated in PBS 500 µl /0.5% Tween-20/1% bovine serum albumin (BSA) andfurther incubated with FITC-conjugated anti-BrdU antibody (5 µg/ml; BDBiosciences) at room temperature for 30 min. Nuclei were then collected bycentrifugation (3000 g) and incubated in 1 ml of propidium iodide stainingsolution (50 µg/ml propidium iodide and 100 U/ml RNase A in PBS) at roomtemperature for 1 h, followed by washing once in PBS containing 0.5% BSA.All light-sensitive steps were carried out in twilight. Samples were analysedfor BrdU incorporation and DNA content by flow cytometry (EPICS ELITE;Beckmann-Coulter), and the percentage of cells in each phase was determinedusing the MultiCycle for Windows Software (Phoenix Flow Systems, SanDiego, CA, USA).

Statistical analysis

Mean ± SD are indicated for all data sets. The two-sided paired t-test wasperformed to compare the effects of loss of ATM function on drug sensitivity.P <0.05 was considered to be statistically significant.

Results

Loss of ATM and increased sensitivity to topoisomerase I and II poisons and to antimetabolites

Loss of the ATM gene has previously been reported to renderp53-null mouse tissues more sensitive to γ-irradiation [11]. Usinga pair of mouse knockout cell lines, we determined the effect ofloss of ATM function in a p53-null background on sensitivity toa panel of anticancer agents. atm+/+/p53–/– cells are replete withrespect to ATM but lack expression of p53, whereas atm–/–/p53–/– cells express neither ATM nor p53 protein (Figure 1).

Figure 2 shows the extent of growth inhibition for both cell linesas a function of drug concentration in response to continuousexposure to the topoisomerase II poisons epirubicin and etopo-side, to the topoisomerase I poisons camptothecin and topotecan,and to the antimetabolites 5-fluorouracil and gemcitabine. TheIC50 values for all the drugs tested are summarised in Table 1.The atm–/–/p53–/– cell line was 3.2-fold (P = 0.0004, n = 5) moresensitive to epirubicin and 3.1-fold (P = 0.003, n = 5) more sensi-tive to etoposide than the atm+/+/p53–/– cell line. Likewise, a3-fold (P = 0.0001, n = 5) greater sensitivity was also found fordoxorubicin. In addition, atm–/–/p53–/– cells displayed a 3.6-foldincreased sensitivity to camptothecin (P = 0.0001, n = 4) and a2.7-fold increased sensitivity to topotecan (P = 0.01, n = 4) ascompared with the ATM-proficient cells. They were also 2-foldhypersensitive to 5-fluorouracil (P = 0.0001, n = 5) and 2.1-foldhypersensitive to gemcitabine (P = 0.003, n = 5) as comparedwith atm+/+/p53–/– cells. Thus, the additional loss of the ATM genein a p53-deficient genetic background resulted in a 2- to 4-foldincreased sensitivity to the topoisomerase I and II poisons, and tosome types of antimetabolites.

Loss of ATM and unaltered sensitivity to platinum compounds, taxanes or the alkylating agent busulfan

Figure 3 shows the extent of growth inhibition for atm+/+/p53–/–

and atm–/–/p53–/– cells in response to a continuous exposure tothe platinum compounds cisplatin, carboplatin and oxaliplatin, tothe alkylating agent busulfan, and to the microtubule poisons

docetaxel and paclitaxel. The IC50 values for these drugs arereported in Table 1. There was no difference in sensitivitybetween the atm+/+/p53–/– and atm–/–/p53–/– cells with respect totreatment with cisplatin (P = 0.38, n = 5), carboplatin (P = 0.36,n = 5), oxaliplatin (P = 0.15, n = 5), docetaxel (P = 0.81, n = 4),paclitaxel (P = 0.51, n = 4) or busulfan (P = 0.14, n = 5). Thus, lossof the ATM function in p53-null cells does not alter the sensitivityto platinum compounds, to taxanes, or to the alkylating agentbusulfan.

ATM deficiency and increased apoptosis

The TUNEL apoptosis assay was performed to determine whetherthe increased sensitivity to epirubicin due to loss of ATM functionwas accompanied by increased apoptosis. Treatment with 25 nMepirubicin, which inhibited growth of ATM-deficient cells bymore than 85%, produced 3- to 4-fold more apoptotic atm+/+/p53–/– cells than atm–/–/p53–/–cells (Figure 4). The respectivevalues were 9 ± 6% versus 2 ± 1% at 34 h after epirubicin treat-ment, and 13 ± 3% versus 4 ± 1% at 57 h after treatment. Theseresults indicate that the hypersensitivity to the cytotoxic effect ofepirubicin in ATM-deficient cells is not associated with increasedapoptosis.

Loss of ATM and overall cell viability in response to the topoisomerase II poison epirubicin

The Trypan Blue exclusion assay was performed to clarify theapparently discrepant results observed with epirubicin in thedifferent assay systems. The data indicate that loss of ATM func-tion was associated with a 2- to 3-fold increase in the number ofTrypan Blue-positive cells following exposure to 25 nM epiru-bicin (Figure 5). The changes in the proportion of Trypan Blueexcluding cells with the time after drug treatment were alsomonitored. The decrease in viability was faster in the ATM-deficient cells than in the ATM-proficient cells (55 ± 8% versus20 ± 8% at 48 h, 71 ± 11% versus 42 ± 12% at 72 h). This parallelsthe data obtained in the MTT assay. Thus, both the growth rateinhibition assay and the Trypan Blue assay indicate that loss of

Figure 1. Immunoblot analysis of the expression of ATM and/or p53 protein in atm+/+/p53–/– and atm–/–/p53–/– cells. The human breast cancer cell line MCF-7 was used as a positive control. atm+/+/p53–/– cells lack the expression of p53, whereas atm–/–/p53–/– cells express neither ATM nor p53.

941

Figure 2. Sensitivity to a continuous exposure to epirubicin, etoposide, camptothecin, topotecan, 5-fluorouracil and gemcitabine in atm+/+/p53–/– (filled squares) and atm–/–/p53–/– (open squares) cells as determined by the MTT assay. Each point represents the mean of at least four independent experiments. Bars = SD.

942

ATM function reduces the threshold for cell kill by a process thatdoes not depend on apoptosis.

Loss of ATM and cell cycle phase distribution after treatment with the topoisomerase II poison epirubicin

To determine whether the differential sensitivity of ATM-proficient and -deficient cells to epirubicin was due to alterationin the ability to trigger G1/S phase arrest, the fraction of cells inS phase was quantified as a function of drug exposure. Table 2shows that 10 nM epirubicin reduced the fraction of ATM-proficient and -deficient cells in S phase after 24 h to about thesame extent.

An alternative hypothesis is that the differential sensitivity toepirubicin is due to altered ability to activate the G2/M checkpoint.Cells were exposed to 10 nM epirubicin, a concentration thatinhibited the growth of ATM-deficient cells in the MTT assay byat least 70%, and the fraction of cells in G2/M was determined atvarious time points. The data presented in Table 3 demonstratethat epirubicin produced a transient 1.3-fold accumulation ofATM-deficient cells in G2/M at 24 h, largely at the expense ofcells in G1. Thus, ATM-deficient p53-null cells seemed to retain,at least partially, the ability to arrest in G2/M in response toepirubicin. A comparable analysis of the response of ATM-proficient cells to epirubicin could not be interpreted due to therapid formation of hyperploid cells.

Taken together, the cell cycle data suggest that ATM does notplay a central role in the ability of epirubicin to trigger either theGl/S or G2/M checkpoint, and that the differential sensitivity ofATM-proficient and -deficient p53-null cells cannot be accountedfor by loss of either of these cell cycle control elements.

Discussion

We report here that p53-null cells are sensitised by the loss of theATM gene to topoisomerase I and II poisons and several types ofantimetabolites, but not to the major platinum compounds, themajor taxanes and one alkylating agent. The data permit severalconclusions to be drawn. First, they extend the previousobservation that p53-deficient mouse tissues are rendered moresensitive to γ-irradiation by the concurrent loss of ATM [11] byidentifying classes of anticancer agents to which these cellsare also hypersensitive. Secondly, they support the existence ofan ATM-dependent, p53-independent DNA damage-inducedeffector pathway and the role of ATM as a sensor for specifictypes of DNA damage [5, 14]. Thirdly, the effector pathwayactivated by ATM in the absence of p53 must involve some as yetunravelled cellular defence mechanism(s), because the increasedchemosensitivity of ATM-deficient cells was not accompanied byincreased apoptosis or by altered activation of either the G1/S orG2/M checkpoints.

ATM may function as a sensor of DNA double-strand breaksarising from γ-irradiation [11] and, as reported here, from sometypes of anticancer agents including topoisomerase poisons orantimetabolites. However, it does not sense lesions produced bymicrotubule poisons or platinum compounds. This is perhaps notsurprising for docetaxel and paclitaxel, since they are not knownto interfere with DNA. Perhaps, as for UV radiation-introducedDNA damage [14], ATR, another member of the PIK family[9, 15], is responsible for sensing damage produced by platinumdrugs. ATR and ATM are partially redundant, and they act inparallel but overlapping DNA damage signalling pathways, butrespond primarily to different kinds of lesions [7, 15].

ATM regulates cell cycle checkpoint activation upon DNAdamage causing cells to arrest in G1/S and/or G2/M, ensuring thatcells delay entry into mitosis and putatively permitting time fordamage repair prior to the onset of mitosis [16]. Although stressmay cause overriding of the G1/S and the G2/M arrests in p53-mutant cells [2], damage-mediated activation of ATM may stopcells at the G2/M checkpoint in these cells [16]. However, loss ofATM does not substantially affect either G1/S or G2/M checkpointactivation in p53-deficient cells after γ-irradiation [17]. Our datashow that the increased sensitivity of these cells to epirubicincannot be explained by the lack of ATM-mediated cell cyclecheckpoint control. Indeed, ATM, unlike ATR, does not seem tomodulate the length and magnitude of the G2 arrest induction[18, 19].

The primary mechanism by which ATM exerts its protectiveeffect may be through modulating damage repair and decreasingthe threshold for cell kill [10]. ATM immunoprecipitates withDNA damage repair proteins including MLH1, MSH2, MSH6and BRCA1, and may facilitate recruitment of repair proteinsto the site of the lesion [20]. When such damage remains unre-paired as the cell enters mitosis, the increased rate of chromosomebreaks and aberrant chromosome segregation results in increasedcell killing [11]. Our results indicate that the increased chemo-sensitivity of the ATM-deficient cells may be a result of processesother than apoptosis.

Table 1. IC50 for ATM-proficient or -deficient p53-null cells determined by MTT assaya

aData are expressed as the mean ± SD of at least four independent experiments.bS represents the sensitisation factor.cP values for statistical significance were determined by the two-sided paired t-test.

Drug atm+/+/p53–/– atm–/–/p53–/– Sb Pc

Epirubicin (nM) 12.0 ± 2.0 3.7 ± 0.6 3.2 0.0004

Doxorubicin (nM) 34.4 ± 5.3 11.6 ± 4.8 3.0 0.0001

Etoposide (nM) 16.2 ± 5.3 5.2 ± 1.5 3.1 0.003

Camptothecin (nM) 26.4 ± 3.2 7.4 ± 1.1 3.6 0.0001

Topotecan (nM) 91.8 ± 24.6 34.6 ± 8.8 2.7 0.01

5-Fluorouracil (µM) 3.1 ± 0.4 1.6 ± 0.2 2.0 0.0001

Gemcitabine (nM) 19.1 ± 5.1 8.9 ± 1.6 2.1 0.003

Cisplatin (µM) 0.37 ± 0.14 0.44 ± 0.14 0.9 0.38

Carboplatin (µM) 5.4 ± 1.5 6.1 ± 1.3 0.9 0.36

Oxaliplatin (µM) 2.9 ± 0.3 2.4 ± 0.6 1.2 0.15

Busulfan (mM) 0.20 ± 0.07 0.14 ± 0.03 1.4 0.14

Docetaxel (nM) 5.4 ± 1.5 5.0 ± 2.4 1.1 0.81

Paclitaxel (nM) 35.6 ± 5.5 33.0 ± 6.3 1.1 0.51

943

Figure 3. Growth inhibition in response to a continuous exposure to cisplatin, carboplatin, oxaliplatin, busulfan, docetaxel and paclitaxel in atm+/+/p53–/– (filled squares) and atm–/–/p53–/– (open squares) cell lines as determined by the MTT assay. Each point represents the mean of at least four independent experiments. Bars = SD.

944

This study supports the concept that tumour-targeted functionalinhibition of ATM increases the efficacy of some anticanceragents in the treatment of p53-deficient cancers, which comprise alarge proportion of human tumours [1].

Acknowledgements

The authors would like to express their appreciation to Dr PhilipLeder (Harvard Medical School, Boston, MA, USA) for provid-ing the cell lines, to Dr Martin Pruschy (Department of Radiology,University of Zurich, Switzerland) for assistance in immunoblotanalysis and to Eva Niederer (Institute for Biomedical Engineer-ing, ETH/University of Zurich, Switzerland) for assistance in cellcycle analysis. This work was supported by the Swiss NationalScience Foundation (grant no. 31-52531.97), the Hermann Klaus-Foundation, and grant CA78648 from the National Institutes ofHealth.

References

1. Greenblatt M, Harris C. Mutations in the p53 tumor suppressor gene:

clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;

54: 4855–4878.

2. Levine AJ. p53, the cellular gatekeeper for growth and division. Cell

1997; 88: 323–331.

3. Bunz F, Hwang PM, Torrance C et al. Disruption of p53 in human cancer

cells alters the responses to therapeutic agents. J Clin Invest 1999; 104:

263–269.

4. Savitsky K, Bar-Shira A, Gilad S et al. A single ataxia-telangiectasia gene

with a product similar to PI-3 kinase. Science 1995; 268: 1749–1753.

5. Lavin MF, Shiloh Y. The genetic defect in ataxia-telangiectasia. Annu

Rev Immunol 1997; 15: 177–202.

6. Kastan MB, Zhan Q, El-Deiry WS et al. A mammalian cell cycle check-

point pathway utilizing p53 and GADD45 is defective in ataxia-

telangiectasia. Cell 1992; 71: 587–597.

7. Smith GC, Cary RB, Lakin ND et al. Purification and DNA binding

properties of the ataxia-telangiectasia gene product ATM. Proc Natl Acad

Sci USA 1999; 96: 11134–11139.

8. Rotman G, Shiloh Y. ATM: from gene to function. Hum Mol Genet 1998;

7: 1555–1563.

9. Khanna KK, Jackson SP. DNA double-strand breaks: signaling, repair

and the cancer connection. Nature Genet 2001; 27: 247–254.

10. Morrison C, Sonoda E, Takao N et al. The controlling role of ATM in

homologous recombinational repair of DNA damage. EMBO J 2000; 19:

463–471.

Figure 4. Effect of loss of ATM on the frequency of apoptotic cells in response to treatment with 25 nM epirubicin as a function of time. Open bars represent atm–/–/p53–/– cells, whereas shaded bars represent atm+/+/p53–/– cells. The data are the mean of two independent experiments. Bars = SD.

Figure 5. Effect of ATM deficiency on the frequency of Trypan Blue including atm–/–/p53–/– (open bars) and atm+/+/p53–/– (shaded bars) cells after treatment with 25 nM epirubicin. The data are the mean of two independent experiments. Bars = SD.

Table 2. Percentage of ATM-proficient or -deficient p53-null cells in S phase in response to treatment with epirubicin determined by BrdU incorporationa

aValues are the mean ± SD of two independent data sets.

Time (h) Epirubicin (10 nM)

atm+/+/p53–/– atm–/–/p53–/–

0 29 ± 1 15 ± 3

24 7 ± 5 3 ± 1

48 0 ± 0 0 ± 0

72 0 ± 0 1 ± 1

Table 3. Percentage of ATM-deficient p53-null cells accumulated in the different cell cycle phases in response to treatment with epirubicina

aValues are the mean ± SD of two independent data sets.

atm–/–/p53–/– Epirubicin (10 nM)

G1 G2/M

Control 46 ± 10 37 ± 10

24 h 35 ± 5 47 ± 8

48 h 42 ± 6 31 ± 4

72 h 40 ± 4 28 ± 4

945

11. Westphal CH, Hoyes KP, Canman CE et al. Loss of atm radiosensitizes

multiple p53 null tissues. Cancer Res 1998; 58: 5637–5639.

12. Westphal CH, Rowan S, Schmaltz C et al. Atm and p53 cooperate in

apoptosis and suppression of tumorigenesis, but not in resistance to acute

radiation toxicity. Nature Genet 1997; 16: 397–401.

13. Mosmann T. Rapid colorimetric assay for cellular growth and survival:

application to proliferation and cytotoxicity assays. J Immunol Methods

1985; 65: 55–63.

14. Xu Y, Baltimore D. Dual roles of ATM in the cellular response to

radiation and in cell growth control. Genes Dev 1996; 10: 2401–2410.

15. Gatei M, Zhou BB, Hobson K et al. Ataxia telangiectasia mutated

(ATM) kinase and ATM and Rad3 related kinase mediate phosphoryla-

tion of Brca1 at distinct and overlapping sites: in vivo assessment using

phospho-specific antibodies. J Biol Chem 2001; 276: 17276–17280.

16. Falck J, Mailand N, Syljuasen RG et al. The ATM-Chk2-Cdc25A check-point pathway guards against radioresistant DNA synthesis. Genes Dev2001; 15: 1067–1077.

17. Westphal CH, Schmaltz C, Rowan S et al. Genetic interactions betweenATM and p53 influence cellular proliferation and irradiation-induced cellcycle checkpoints. Cancer Res 1997; 57: 1664–1667.

18. Pincheira J, Bravo M, Navarrete MH et al. Ataxia-telangiectasia: G(2)checkpoint and chromosomal damage in proliferating lymphocytes.Mutagenesis 2001; 16: 419–422.

19. Cliby WA, Lewis KA, Lilly KK, Kaufmann SH. S phase and G2 arrestsinduced by topoisomerase I poisons are dependent on ATR kinase func-tion. J Biol Chem 2002; 277: 1599–1606.

20. Wang Y, Cortez D, Yazdi P et al. BASC, a supercomplex of BRCA1-associated proteins involved in the recognition and repair of aberrantDNA structures. Genes Dev 2000; 4: 927–939.

5. Publication III The histone deacetylase inhibitors suberoylanilide hydroxamic (vorinostat) and valproic

acid induce irreversible and MDR1-independent resistance in human colon cancer cells


Int. J. Oncol. 2007;31:633-641 (Impact factor = 2.556) (ref 18)

14

Abstract. Histone deacetylase (HDAC) inhibitors such assuberoylanilide hydroxamic acid (SAHA, Vorinostat), valproicacid (VPA), and FK228 are members of a relatively novelclass of small molecular weight chemicals that have high anti-neoplastic activity. They cause growth inhibition and apoptosisspecifically in tumor cells, and they act also as chemo- andradio-sensitizers. In the present study, the potential of SAHAand VPA to induce resistance was studied. To that aim HDACinhibitor-resistant sublines were generated by stepwiseexposure of colon tumor cells to increasing concentrationsof these compounds. Clonogenic data demonstrated that theSAHA- and VPA-induced sublines were 2-fold resistant tothese compounds. This resistance was non-reversible, as itwas maintained even when the sublines were cultured in theabsence of SAHA or VPA. The SAHA- and VPA-inducedresistant sublines were also stably cross-resistant to VPAand SAHA, respectively, but retained sensitivity against non-HDAC inhibitor-type anticancer agents. The SAHA-inducedresistance correlated with loss of the G2/M checkpoint but itwas not accompanied by reduced induction of the endogenouscell cycle inhibitors p21 and p27. Furthermore, SAHA-inducedresistance was not due to reduced apoptosis, and it was neitherdependent on MDR expression nor was it due to increasedexpression of HDAC1 and HDAC3. Taken together, thesedata demonstrate the potential of SAHA and VPA to induceresistance. This resistance was not dependent on MDRexpression, did not involve MMR, and seemed to underlie amechanism that differs from that underlying the previouslyobserved FK228-induced resistance. The finding that SAHAand VPA induce only modest resistance despite continuous

treatment and that the resistance is MDR-independent suggestsa preference for these two drugs over FK228 for use incombination treatment with classic anticancer agents.

Introduction

Enzymes modifying the activity of histones, such as histoneacetyltransferases (HATs) and histone deacetylases (HDACs),are crucial to proliferation, apoptosis, development, angio-genesis, and carcinogenesis. The balance between theseactivities regulates the expression of genes controlling theseprocesses, mainly by regulating the accessibility of DNA-interacting proteins for the DNA. HDAC-mediated silencing oftumor suppressor genes plays a role in cancer pathophysiology.HDACs are subdivided in four classes: class I (HDAC1, -2, -3,-8), class II (HDAC4, -5, -6, -7, -9, -10), class III (alsoreferred to as sirtuins: SIRT1 through SIRT7), and class IV(HDAC11) (1-3).

Inhibitors of HDACs counteract the removal of acetyl-groups from histones and render the DNA available for DNA-interacting proteins. HDAC inhibitors have strong anticancerproperties and many of them have moved forward into clinicaltrials, and Vorinostat (suberoylanilide hydroxamic, SAHA)has even been granted market approval for the indication ofcutaneous T-cell lymphoma (4-7). HDAC inhibitors inducecell cycle checkpoint activation and apoptosis specifically intumor cells (8-12). They also radio- and chemo-sensitizetumor cells (13,14). HDAC inhibitors have also been linkedto some characteristics of DNA repair (15,16), suggestingthat HDAC inhibitors can induce abnormal DNA structuresthat may be recognized by DNA repair proteins (17).

Recent studies have shown that the powerful HDACinhibitor FK228 (depsipeptide) can induce reversible FK228resistance in tumor cells by reversible induction of MDR1(18,19), a multidrug resistance transporter that functions byextrusion of cytotoxic drugs from the cell and by mediatingsequestration of the drugs into intracellular compartments,both leading to a reduction in effective intracellular drugconcentrations (20). Expanding on this issue, we investigatedwhether the HDAC inhibitors SAHA and valproic acid (VPA),which are structurally unrelated to FK228, are potentialinducers of HDAC inhibitor resistance in tumor cells andwhether HDAC inhibitor-induced resistance is associated

INTERNATIONAL JOURNAL OF ONCOLOGY 31: 633-641, 2007 633

The histone deacetylase inhibitors suberoylanilide hydroxamic(Vorinostat) and valproic acid induce irreversible and

MDR1-independent resistance in human colon cancer cells

ANDRE FEDIER1, KONSTANTIN J. DEDES1, PATRICK IMESCH1, ANDRE O. VON BUEREN2 and DANIEL FINK1

1Department of Gynecology, University Hospital Zurich; 2Division of Oncology, University Children's Hospital Zurich, Zurich, Switzerland

Received February 7, 2007; Accepted March 27, 2007

_________________________________________

Correspondence to: Dr André Fedier, Department of Gynecology,University Hospital of Zurich, Frauenklinikstrasse 10, CH-8091Zurich, SwitzerlandE-mail: [email protected]

Key words: cell cycle, apoptosis, drug resistance, histone deacteylaseinhibitors, multidrug resistance, p21 and p27, suberoylanilidehydroxamic acid

633-641 24/7/07 11:02 Page 633

with cross-resistance to non-HDAC inhibitor-type anticanceragents, inducible MDR1 expression, reduced expression ofthe HDAC-responsive gene p21, altered expression and/oracetylation status of HDACs, and impaired cell cycle check-point and apoptosis activation. To this aim and to determinea possible involvement of DNA mismatch repair (MMR),respective sublines of a pair of human adenocarcinoma celllines either expressing the MMR protein MLH1 (MLH1-proficient HCT116ch3) or lacking MLH1 expression (MLH1-deficient HCT116ch2) were generated by stepwise exposureof these cell lines to increasing concentrations of HDACinhibitors.

Our results identify SAHA and VPA as potential inducersof a non-reversible and MDR1-independent HDAC inhibitorresistance phenotype. This resistance seems to be different tothat observed with FK228 and not dependent on the MMR-status of the tumor cells.

Materials and methods

Drugs and chemicals. Suberoylanilide hydroxamic acid(SAHA; Alexis Biochemicals, Lausen, Switzerland) andvalproic acid (VPA; Sigma, Buchs, Switzerland) werepurchased, as were cisplatin, docetaxel, and 6-thioguanine(Sigma). Temozolomide was a generous gift (Schering-Plough,Kenilworth, NJ). Stock solutions were prepared in DMSO(SAHA, temozolomide), in ethanol (docetaxel), or in H2O(cisplatin, 6-thioguanine, VPA). All stock solutions were storedat -20˚C.

Cell culture and generation of HDAC inhibitor-resistantsublines. A pair of an MLH1-deficient human colorectal adeno-carcinoma cell line (designated HCT116ch2, complementedwith chromosome 2) and its MLH1-proficient counterpart(designated HCT116ch3, complemented with chromosome 3),which were derived from the MLH1-deficient parental humancolorectal adenocarcinoma cell line HCT116 (American TypeCulture Collection; ATCC CCL 247), were used. The charac-teristics of the cell lines (e.g. chromosome complementation)and the culturing conditions have been described previously(21-23). A HeLa cell line (provided by Dr G. Marra, Institute ofMolecular Cancer Research, University of Zurich, Switzerland)was also used. When seeded sparsely on tissue culture plates,all the cell lines and sublines formed well-defined individualcolonies.

Similar to the method described previously (18), the HDACinhibitor-resistant cell sublines, hereafter designated asHCT116ch3/SAHA, HCT116ch2/SAHA, HCT116ch3/VPA,or HCT116ch2/VPA, respectively, were generated by step-wise exposures of the MLH1-proficient HCT116ch3 cell lineand the MLH1-deficient HCT116ch2 cell line to increasingconcentrations of either SAHA or VPA, starting with 1 μMfor SAHA or 2.5 mM for VPA. A similar protocol was usedto generate the HeLa/SAHA subline. The principle of selectionwas the clonal growth in the presence of increasing concen-trations of the HDAC inhibitor, based on the idea that cellsare altered by chronic HDAC inhibitor exposure in that theyacquire new features in an irreversible fashion. Basically,cells (100,000) were plated in cell culture flasks and treatedwith SAHA or VPA 24 h after plating. After another 48 h,

the HDAC inhibitor-containing medium was exchanged forinhibitor-free medium, followed by incubation for another 6days to allow recovery of the surviving cells. These were thenharvested by trypsinization, transferred into new flasks, andexpanded to confluence. One fraction was stored at -80˚C(for protein analysis), the other (100,000 cells) was re-seededin culture flasks and was subjected to treatment with SAHA orVPA 24 h later, to medium exchange, recovery and harvestingas described. This protocol was repeated (10 times for SAHA,8 timed for VPA), and for each cycle, the concentrations ofSAHA or VPA were incremented, resulting in a 6-fold totalincrement for SAHA (6 μM) and 60-fold for VPA (150 μM).The growth rates of the cell lines and the sublines werecalculated from the doubling times from one passage to thesubsequent and averaged for a period of two months. MLH1gene expression in the cell lines and sublines was determinedby immunoblotting.

Drug sensitivity assay. In a typical clonogenic assay, 400cells in medium were plated onto 60-mm cell culture dishes,followed by drug addition after 24 h. For a 24-h treatmentthe medium was replaced by drug-free medium 24 h afterdrug addition; for a continuous treatment (8 days) the drug-containing medium was maintained. Cells were incubatedfor 7 days to allow colony formation, fixed with 25% aceticacid in ethanol and stained with Giemsa. Colonies of at least50 cells were scored. Each experiment was performed atleast three times using triplicate cultures. The relative colonyformation (% clonogenic survival) was plotted against thedrug concentrations, and the IC50 concentrations were calc-ulated by linear extrapolation.

Apoptosis and cell cycle analyses by flow cytometry. Analysesof apoptosis (TUNEL DNA fragmentation) and cell cyclewere performed by flow cytometry on a FACSCalibur flowcytometer (BD Biosciences; Allschwil, Switzerland) withCELLQuest software (BD Biosciences). Data analyses forapoptosis and cell cycle distribution were performed on linearPI histograms using the mathematical software ModFit LT 2.0(Verity Software House; Topsham, ME, USA). For samplepreparation, cells were grown to 70% confluence in 60-mmdishes and treated with the concentration of SAHA whichreduced clonogenic survival by at least 95%. At the timesindicated, adherent and floating cells were collected andprepared for apoptosis and cell cycle analysis as describedpreviously (23).

Immunoblot analysis. After cells had grown to 70% confluencein 60-mm dishes, they were treated with SAHA or TSA andcollected 3, 7, or 14 h later, washed in PBS, and prepared forimmunoblot analysis performed following standard protocolsfor PAGE gel electrophoresis. Protein (20 μg) was separatedusing 10% SDS-PAGE, followed by blotting onto a poly-vinylidene difluoride membrane (Amersham Biosciences,Otelfingen, Switzerland). Expression of MLH1 protein wasdetected by the mouse antibody (550838; BD BiosciencesPharmingen, Basel, Switzerland) and the anti-mouse secondaryhorseradish peroxidase-conjugated antibody (M15345;Transduction Laboratories, Lexington, KY). p21 and p27proteins were detected by the rabbit antibody (05-345; Upstate,

FEDIER et al: SAHA-INDUCED RESISTANCE IN HUMAN COLON TUMOR CELLS634

633-641 24/7/07 11:02 Page 634

Lucerna Chem AG, Lucerne, Switzerland) and the rabbitantibody (2552; Cell Signaling, BioConcept, Allschwil,Switzerland), respectively. p53 and MDR1 were detectedby the respective rabbit antibodies (sc-6243, sc-13131;Santa Cruz Biotechnology Inc., LabForce AG, Nunningen,Switzerland). HDAC1, -3, -4, -5, -6, and -7 were detected byrabbit antibodies (Kit 9928; Cell Signaling), acetyl-histoneH3 (Lys9) and acetyl-histone H4 (Lys8) by the respectiverabbit antibodies (6971, 2594; Cell Signaling). The anti-rabbitsecondary horseradish peroxidase-conjugated antibody (7074;Cell Signaling) was used. As a sample loading control, themouse antibody against ß-actin (A5441; Sigma) was used. Allthe complexes were visualized by enhanced chemiluminescence

(Amersham Biosciences). Quantitative analysis of thecomplexes (intensity on the autoradiograph) was performedby densitometry (normalized against ß-actin) using the ScionImage 4.01 Win software (Scion Corporation, Frederick, MD).

Statistical analysis. The mean ± SD values were calculatedfor all data sets. A p value <0.05 was considered statisticallysignificant (paired, two-tailed Student's t test).

Results

Resistance induction by SAHA and VPA. Clonogenic datarevealed that the MLH1-deficient cell line (HCT116ch2) wasas sensitive as the MLH1-proficient cell line (HCT116ch3) toa continuous (8 days) treatment with SAHA or VPA (Fig. 1Aand B). For the generation of the SAHA- and VPA-inducedsublines, the respective IC90 concentrations were determined,being 1 μM for SAHA and 2.5 μM for VPA. Expression ofMLH1 protein was present in both sublines derived from theHCT116ch3 cell line and was absent in those derived fromthe HCT116ch2 cell line (Fig. 1C). The growth rates of theHDAC inhibitor-induced sublines were similar to those ofthe non-induced cell lines (Table I).

It was determined whether these HDAC inhibitor-inducedsublines are on the one hand resistant to the agent by whichthey were induced and on the other cross-resistant to VPA, acarboyxylate HDAC inhibitor which is structurally differentfrom SAHA, a member of the hydroxamic acids class ofHDAC inhibitors. The results (Fig. 2; Table II) showed thatboth the MLH1-proficient and the MLH1-deficient SAHA-induced sublines were 2-fold (p<0.01) resistant to SAHA(Fig. 2A) and also 2-fold (p<0.01) cross-resistant to VPA(Fig. 2B). Likewise, VPA treatment also induced a 2-fold(p<0.01) resistance to VPA (Fig. 2C) as well as a 2-fold(p<0.01) cross-resistance to SAHA (Fig. 2D), irrespective ofthe MLH1 status of the cells.

To consider a possible contribution of the complementedchromosomes, resistance induction by SAHA was alsoinvestigated for the parental HCT116 cell line. The resultsshowed that 8 cycles of SAHA treatment also induced a 2-fold(p<0.01) resistance with this cell line. The respective IC50


Figure 1. Clonogenic survival of MLH1-proficient HCT116ch3 (ƒ) andMLH1-deficient HCT116ch2 (∫) human colorectal adenocarcinoma cellsafter an 8-day treatment with the HDAC inhibitors SAHA (A) and VPA (B).Each point is the mean ± SD of three independent experiments performedin triplicate cultures. (C) The presence or absence of MLH1 expression inthe HCT116ch3 (ch3) or HCT116ch2 (ch2) cells, respectively, and in thecorresponding HDAC inhibitor-induced sublines (ch3/SAHA, ch3/VPA,ch2/SAHA, ch2/VPA). ß-actin is the sample loading control.

Table I. Comparison of the cell doubling time.–––––––––––––––––––––––––––––––––––––––––––––––––Cell lines Doubling time (h)a

–––––––––––––––––––––––––––––––––––––––––––––––––HCT116+ch3 21.4±1.1HCT116+ch3/SAHA 23.4±0.5HCT116+ch3/VPA 23.7±0.5

HCT116+ch2 21.6±0.9HCT116+ch2/SAHA 22.4±1.3HCT116+ch2/VPA 23.8±1.1–––––––––––––––––––––––––––––––––––––––––––––––––aThe doubling times of the cell lines and the respective sublineswere calculated from one passage to the subsequent and averagedfor a period of two months. The values represent the mean ± SD of5 independent data sets.–––––––––––––––––––––––––––––––––––––––––––––––––

633-641 24/7/07 11:02 Page 635

values were 0.6±0.1 μM (HCT116) vs. 1.3±0.3 μM (HCT116/SAHA). Notably, SAHA (80-fold increment, 7 cycles) failedto induce resistance in HeLa cells. The respective IC50 valueswere similar: 1.1±0.1 μM (HeLa) vs. 1.3±0.1 μM (HeLa/SAHA).

It is noteworthy that all the HDAC inhibitor-induced sub-lines maintained resistance for at least 6 months (>30 passages,maximum period of time tested) even when cultured in HDACinhibitor-free medium. In addition, the SAHA-resistant sublineswere not cross-resistant to non-HDAC inhibitor-type anti-


Figure 2. Clonogenic survival after an 8-day treatment with HDAC inhibitors for the MLH1-proficient HCT116ch3 (ƒ) and MLH1-deficient HCT116ch2 (∫)cell lines (straight lines) and the respective SAHA-induced (A, B) or VPA-induced (C, D) sublines (dashed lines). (A, C) Resistance to the HDAC inhibitorsthemselves; (B, D) cross-resistance of SAHA-induced sublines to VPA. Shown is also the cross-resistance of SAHA-induced sublines to the non-HDACinhibitor-type antitumor agents cisplatin (E), docetaxel (F), 6-thioguanine (G), and temozolomide (H). Each point is the mean ± SD of 4 independentexperiments performed in triplicate cultures.

633-641 24/7/07 11:03 Page 636

cancer agents such as cisplatin, docetaxel, 6-thioguanine, andtemozolomide, regardless of the MLH1 status of the cells(Fig. 2E-H).

SAHA-induced sublines and G2/M checkpoint activation. It wasdetermined whether loss of cell cycle checkpoint activationaccounted for the observed SAHA-induced resistance. Flow


Figure 3. G2/M cell cycle checkpoint induction in the SAHA-induced sublines and the corresponding non-induced cell lines in response to a 24-h exposure to 5 μMSAHA. (A) Representative cell cycle phase distribution profiles of the DNA content obtained by flow cytometry for the MLH1-proficient (HCTch3) and theMLH1-deficient (HCTch2) HCT116 cell lines and the respective sublines induced by SAHA (HCTch3/SAHA, HCTch2/SAHA). X-axis (channels): position ofcells accumulated in G1, S, or G2/M. Y-axis; number of events per channel. (B) Quantitative presentation of primary flow cytometry data captured 24 h post-treatment of the cells with 5 μM SAHA. The changes in the proportion of cells accumulated at the G2/M (black) transition, in the S phase (white), and at theG1/S transition (grey) are presented. Each point is the mean ± SD of 3 independent experiments. *p<0.05.

Table II. IC50 values for the MLH1-proficient HCT116ch3 cell line and the MLH1-deficient HCT116ch2 cell line and therespective sublines derived from stepwise exposure to SAHA.–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––Inhibitor HCT116ch3 HCT116ch3/SAHA Fold changea HCT116ch2 HCT116ch2/SAHA Fold changea

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––SAHA (μM) 1.7±0.1 3.9±0.2 2.3b 1.6±0.1 3.2±0.6 2.0b

VPA (mM) 2.1±0.3 4.7±1.3 2.2b 2.1±0.2 4.2±0.5 2.0b

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––HCT116ch3 HCT116ch3/VPA Fold changea HCT116ch2 HCT116ch2/VPA Fold changea

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––VPA (mM) 1.4±0.2 2.6±0.3 1.9b 1.4±0.2 2.6±0.3 1.9b

SAHA (μM) 0.7±0.1 1.3±0.3 1.9b 0.7±0.1 1.4±0.3 2.0b

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––aFold change is referred to as resistance or cross-resistance and is defined as the ratio of the IC50 values for the sublines and those for the

non-induced cell lines. bp<0.01.–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

633-641 24/7/07 11:03 Page 637

cytometry data analysis (Fig. 3) demonstrated that 5 μM SAHA(reduced clonogenic survival by >95%) produced an arrestat the G2/M transition of the cell cycle in both the MLH1-proficient HCT116ch3 and the MLH1-deficient HCT116ch2cell lines. This, however, was not observed for the respectiveSAHA-induced sublines, HCT116ch3/SAHA and HCT116ch2/SAHA. This demonstrated that the ability to induce the G2/Mcheckpoint was substantially (2-fold, p<0.05) reduced in theSAHA-induced sublines as compared to the non-induced celllines, indicating that loss of this checkpoint contributes to theSAHA-induced resistance observed for these sublines.

SAHA-induced resistance and regulation of p21 or p27. Itwas determined whether the SAHA-induced resistance wasaccompanied by reduced induction of p21 and/or p27, twoendogenous cell cycle inhibitors induced by HDAC inhibitors(24-26). Immunoblot data showed that SAHA strongly inducedp21 and p27 expression in all samples, i.e. in both sublines(MLH1-proficient and MLH1-deficient) and in the respectivesublines (Fig. 4A), demonstrating that SAHA-inducedresistance and loss of the G2/M checkpoint were not due toloss of p21 and p27 induction in the resistant sublines.

SAHA-induced resistance and apoptosis. DNA fragmentationdata (Fig. 4B) showed that susceptibility to SAHA-imposedapoptosis was nearly the same for the SAHA-induced (resistant)sublines and for the non-induced (sensitive) cell lines, regardlessof the presence or absence of MLH1 expression in the cells,demonstrating that SAHA-induced resistance did not correlatewith loss of apoptosis.

SAHA-induced resistance and MDR1 expression. Therecently reported resistance to FK228 has been attributed tothe inducible and reversible expression of MDR1 (18,19). Incontrast, our observations (Fig. 5) showed that MDR1 wasnot expressed in any of the cell lines or the respective SAHA-induced sublines (untreated or treated with 15 μM SAHA for14 h), demonstrating that SAHA-induced resistance was notassociated with induction of MDR1. Expression of MDR1was also not detected in the VPA-induced sublines.

SAHA-induced resistance and HDAC1 and HDAC3 expression.HDAC inhibitors cause hyperacetylation of histones and it wasthus determined whether the SAHA-induced, SAHA-resistantsublines displayed hypoacetylation of the histones H2B, H3,and H4 (relative to the non-induced cell lines). The immunoblotresults are shown (Fig. 6A). SAHA produced acetylation ofH3 and H4 in both non-induced cell lines (i.e. MLH1-proficientand MLH1-deficient) as well as in the respective SAHA-induced sublines, demonstrating that SAHA exerts its effectin a target-specific manner. For the MLH1-proficient setting,SAHA produced a higher acetylation of H3 and H4 in theSAHA-induced, resistant subline (HCT116ch3/SAHA) thanin the non-induced cell line (HCT116ch3), indicating thatSAHA-induced resistance in MLH1-expressing cells was notaccompanied by hypoacetylation. This was different for theMLH1-deficient setting. First, the level of acetylated H4 inthe SAHA-induced subline (HCT116ch2/SAHA) was similarto that in the non-induced cell line (HCT116ch2). Second,the level of acetylated H3 in the SAHA-induced sublinewas lower than in the respective non-induced cell line. Thisdemonstrated that SAHA-induced resistance was accompaniedby hypoacetylation of only H3 and only in MLH1-deficientcells. In addition, detection of acetylated histone H2B waspoor.

It was reasoned that SAHA-induced resistance could bedue to increased levels of HDACs present in the respectivesublines. This was not the case, as the SAHA-induced sublinesexpressed levels of HDAC1, -3, -6, and -7 that were similarto those expressed in the respective non-induced cell lines(Fig. 6B). Expression of HDAC4 and HDAC5 was not


Figure 4. (A) Induction of p21 and p27 by SAHA in untreated controlsamples (no SAHA) and in samples captured 7 or 14 h after treatment with 5or 15 μM SAHA. Actin is the sample loading control. Representative of twoindependent experiments. (B) Induction of apoptosis (DNA fragmentation)presented as the percentage of TUNEL-positive (apoptotic) cells as a functionof treatment with 15 μM SAHA (captured 24 h post treatment). Mean ± SDof 3 independent data sets.

Figure 5. Expression of MDR1 as a function of a 14-h treatment with 15 μMSAHA in the non-induced cell lines and the respective SAHA- or VPA-induced sublines. A control lysate (sc-2284; Santa Cruz Biotechnology Inc.)was used as a positive control for MDR1 (Co).

633-641 24/7/07 11:03 Page 638

detected. In addition, SAHA produced increases in the levelsof HDAC1 and HDAC3 in all the cell lines and the respectivesublines relative to untreated samples (Fig. 6C) and to acomparable extent (Fig. 6D). This demonstrated that resistanceto SAHA was not accompanied by elevated expression ofHDAC1 and HDAC3 in these sublines.

Discussion

We observed the following: i) that HDAC inhibitors SAHAand VPA induced HDAC inhibitor (cross-) resistance whichwas not associated with cross-resistance to some none-HDACinhibitor-type anticancer agents; ii) that this type of HDACinhibitor-induced resistance was stable, MDR1-independent,and not associated with elevated expression of HDAC1 andHDAC3; iii) that the SAHA-induced resistance correlatedwith defective activation of the G2/M checkpoint but not withloss of p21 and p27 induction and apoptosis; and iv) thatMLH1 was irrelevant for the cytotoxic effect of SAHA andVPA. We may conclude that a novel mechanism of induceddrug resistance, which is different to that observed withFK228, underlies the herein described resistance induced bythese HDAC inhibitors.

The antineoplastic activity of HDAC inhibitors is anunquestionable property of these compounds. However, ourobservations with SAHA and VPA, together with thosereported for FK228 (18,19), may shed some light on another

less well-described aspect of HDAC inhibitors, namely theirpotential to induce resistance in tumor cells. In general terms,resistance induction may mean that existing cells are alteredby HDAC inhibitor exposure in that they acquire new featuresin an irreversible fashion that renders them resistant (clonalgrowth in the presence of an HDAC inhibitor) or it mightmean selection of those cells that are a priori more resistantto cell killing by HDAC inhibitors. The former principle wasthe basis of our experimental design, although a possiblecontribution of the latter cannot be ruled out. Despite someshared findings, our observations differ from those reported forFK228 in several ways, indicating that different mechanismsmay underlie these resistance phenomena.

First, the SAHA- and VPA-induced resistance phenotypeis likely to be stable, as it was maintained for at least 6 months(maximum period of observation) even in the absence of thedrugs in the culture medium, and unlikely to be due to differ-ential growth rates of the HDAC inhibitor-induced sublines.In contrast, FK228-induced resistance required the presenceof FK228 and correlated with slower growth rates (19).

Second, our results demonstrate that the SAHA- and VPA-induced resistance is not dependent on the MDR1 transporter.This is consistent with the findings that SAHA is not substratefor MDR1 (9,18) and also explains why we did not observecross-resistance to ‘classic’ anticancer agents such as cisplatinand docetaxel which are substrates for MDR1 (20). This isopposed to the FK228-induced resistance that correlated with


Figure 6. (A) Acetylation of histones H2B, H3, and H4. (B) Expression of HDACs in the non-induced cell lines and the SAHA-induced sublines. (C)Expression of HDACs as a function of treatment with 15 μM SAHA in the non-induced cell lines and the respective SAHA-induced sublines captured 14 hpost treatment. (D) Quantitative presentation of the changes in HDAC1 and HDAC3 expression (relative to untreated controls and normalized against ß-actinloading control).

633-641 24/7/07 11:03 Page 639

reversible MDR1 induction and cross-resistance to doxorubicinand etoposide but not to SAHA and TSA (18,19). We alsofound cross-resistance among members of structurally differentclasses of HDAC inhibitors (e.g. the hydroxamic acid-likeSAHA versus the carboxylate-like VPA). It therefore seemsthat mechanisms of resistance induction other than thosemediated by MDR1 are involved, and this may also explainthe rather low magnitude of resistance (2-fold) for SAHAand VPA, as compared to the high degree of resistance (up to1,700-fold) reported for FK228 (18). The finding that MDR1is not involved in SAHA- and VPA-induced resistance suggeststhat, despite their potential to induce modest resistance, theseHDAC inhibitors can be used in combination with otherantitumoral agents that are substrates for MDR1. In addition,it remains to be seen whether this low-level resistance issufficient to accumulate a stably resistant subpopulation oftumor cells and therefore to impair tumor treatment in in vivomodels, although this has been shown for the 2-fold cisplatinresistance in MMR-deficient cells (27).

Candidate mechanisms of resistance may include differ-ential induction of cell cycle checkpoints and apoptosis,overexpression of HDACs, loss of HDAC expression due tomutations, or even ‘off-target’ (i.e. HDAC-independent)effects. Our particular interest was directed towards SAHA,as it is considered one of the most potent and promisingHDAC inhibitors and is the first of its class to be grantedmarket approval. Our observations indeed demonstrated thatSAHA-induced resistance was accompanied by loss of theG2/M checkpoint, meaning that this defective checkpoint mayprovide these cells with a growth advantage over the SAHA-sensitive cells. The involvement of HDAC inhibitors in G2/Mcheckpoint control has been shown previously (26,28-30).Other studies have shown that HDAC inhibitors regulate theG1/S checkpoint, and this is via induction of p21, an HDACinhibitor-inducible endogenous cell cycle inhibitor (24,31).Our observations indicate that neither the loss of the G1/Scheckpoint nor the failure to induce p21 accounted for SAHA-induced resistance. In addition, loss of the G2/M checkpointwas not accompanied by reduced induction of p27, a proteinalso involved in regulation of HDAC inhibitor-mediated check-point control (26,32) or by altered p53 expression. Furthermore,SAHA-imposed apoptosis (8) was observed in the non-induced(sensitive) cell lines as well as in the SAHA-induced (resistant)sublines, indicating the SAHA resistance in these cells did notcorrelate with loss of apoptosis.

It was reasoned that resistance to SAHA was due to anincreased expression of its specific targets. Consistent with aprevious study reporting HDAC inhibitor-imposed increasesin HDAC levels (33), SAHA-imposed increases in HDAC1and HDAC3 expression were found, but these increases werenot more pronounced in the resistant sublines, suggesting thatresistance may not be caused by increased expression ofHDACs. Likewise, the acetylation status of histones H3 andH4 was not generally lower in the resistant sublines, meaningthat SAHA specifically affected its targets and that resistancemay not be explained by the failure of SAHA to inhibit HDACactivity.

It has been suggested that HDAC inhibitors may induceabnormal DNA structures that can be recognized by DNArepair proteins (16,17). Recently, HDAC inhibitor resistance

associated with loss of HDAC2 protein due to a truncatingmutation in the HDAC2 gene was reported in MLH1-deficient,microsatellite-instable tumor cells (34). Therefore, we haveaddressed the issue of whether the lack of MLH1 expressionreduces the cytotoxic effect of HDAC inhibitors, and ourresults showed that this was not the case. We also questionedwhether the absence of MLH1 is a determinant for resistanceinduction. Our observations showed that this was not the casewith SAHA and VPA. In addition, the presence of the extrachromosomes within the cells used in this study is unlikely tobe critical for resistance induction by these HDAC inhibitors,as we observed SAHA-induced resistance also for the parentalHCT116 cell line.

Although inconclusive, our data suggest a novel mechanismof HDAC inhibitor-induced resistance that is not due to MDR1expression, elevated expression of HDAC1 and HDAC3,and reduced apoptosis. It is unclear whether SAHA-inducedresistance is a cell line- or tissue-specific phenomenon (itwas not seen with HeLa cells), and therefore it may thus notbe generalizable. In addition, the molecular bases of theunderlying mechanisms of this possibly novel type of HDACinhibitor induced resistance are still unknown and moredetailed studies are required to identify them. These includemicroarray studies, the use of a larger set of cell lines, andthe consideration of other cellular drug detoxification systems.The possibility should also be considered that this low-foldresistance may be ascribed to subtle, chronic treatment-induceddifferences in gene expression profiles that may be hard toidentify experimentally rather than being clearly ascribed to afew single factors.

Acknowledgements

This study was supported by the EMDO Foundation Zurich.

References

1. Cress WD and Seto E: Histone deacetylases, transcriptionalcontrol, and cancer. J Cell Physiol 184: 1-16, 2000.

2. Marks P, Rifkind RA, Richon VM, Breslow R, Miller T andKelly WK: Histone deacetylases and cancer: causes and therapies.Nat Rev Cancer 1: 194-202, 2001.

3. de Ruijter AJ, van Gennip AH, Caron HN, Kemp S andvan Kuilenburg AB: Histone deacetylases (HDACs): characteriz-ation of the classical HDAC family. Biochem J 370: 737-749,2003.

4. Rosato RR and Grant S: Histone deacetylase inhibitors inclinical development. Expert Opin Investig Drugs 13: 21-38,2004.

5. Dokmanovic M and Marks PA: Prospects: histone deacetylaseinhibitors. J Cell Biochem 96: 293-304, 2005.

6. Minucci S and Pelicci PG: Histone deacetylase inhibitors andthe promise of epigenetic (and more) treatments for cancer. NatRev Cancer 6: 38-51, 2006.

7. Bolden JE, Peart MJ and Johnstone RW: Anticancer activities ofhistone deacetylase inhibitors. Nat Rev Drug Discov 5: 769-784,2006.

8. Ruefli AA, Ausserlechner MJ, Bernhard D, et al: The histonedeacetylase inhibitor and chemotherapeutic agent suberoyl-anilide hydroxamic acid (SAHA) induces a cell-death pathwaycharacterized by cleavage of Bid and production of reactiveoxygen species. Proc Natl Acad Sci USA 98: 10833-10838,2001.

9. Ruefli AA, Bernhard D, Tainton KM, Kofler R, Smyth MJ andJohnstone RW: Suberoylanilide hydroxamic acid (SAHA)overcomes multidrug resistance and induces cell death in P-glycoprotein-expressing cells. Int J Cancer 99: 292-298,2003.


633-641 24/7/07 11:03 Page 640

10. Marks PA, Richon VM, Miller T and Kelly WK: Histonedeacetylase inhibitors. Adv Cancer Res 91: 137-168, 2004.

11. Villar-Garea A and Esteller M: Histone deacetylase inhibitors:understanding a new wave of anticancer agents. Int J Cancer112: 171-178, 2004.

12. Catalano MG, Fortunati N, Pugliese M, et al: Valproic acidinduces apoptosis and cell cycle arrest in poorly differentiatedthyroid cancer cells. J Clin Endocrinol Metab 90: 1383-1389,2005.

13. Kim MS, Blake M, Baek JH, Kohlhagen G, Pommier Y andCarrier F: Inhibition of histone deacetylase increases cytotoxicityto anticancer drugs targeting DNA. Cancer Res 63: 7291-7300,2003.

14. Camphausen K, Burgan W, Cerra M, et al: Enhanced radiation-induced cell killing and prolongation of gammaH2AX fociexpression by the histone deacetylase inhibitor MS-275. CancerRes 64: 316-321, 2004.

15. Krajewski WA: Effect of in vivo histone hyperacetylation on thestate of chromatin fibers. J Biomol Struct Dyn 16: 1097-1106,1999.

16. Ju R and Muller MT: Histone deacetylase inhibitors activatep21(WAF1) expression via ATM. Cancer Res 63: 2891-2897,2003.

17. D'Amours D and Jackson SP: The Mre11 complex: at thecrossroads of DNA repair and checkpoint signalling. Nat RevMol Cell Biol 3: 317-327, 2002.

18. Xiao JJ, Huang Y, Dai Z, et al: Chemoresistance to depsipeptideFK228 [(E)-(1S,4S,10S,21R)-7-[(Z)-ethylidene]-4,21-diisopropyl-2-oxa-12,13-dithia-5,8,20,23-tetraazabicyclo[8,7,6]-tricos-16-ene-3,6,9,22-pentanone] is mediated by reversibleMDR1 induction in human cancer cell lines. J Pharmacol ExpTher 314: 467-475, 2005.

19. Yamada H, Arakawa Y, Saito S, Agawa M, Kano Y andHoriguchi-Yamada J: Depsipeptide-resistant KU812 cells showreversible P-glycoprotein expression, hyper-acetylated histones,and modulated gene expression profile. Leuk Res 30: 723-734,2006.

20. Gottesman MM and Pastan I: Biochemistry of multidrug resistancemediated by the multidrug transporter. Annu Rev Biochem 62:385-427, 1993.

21. Koi M, Umar A, Chauhan DP, et al: Human chromosome 3corrects mismatch repair deficiency and microsatellite instabilityand reduces N-methyl-N'-nitro-N-nitrosoguanidine tolerance incolon tumor cells with homozygous hMLH1 mutation. CancerRes 54: 4308-4312, 1994.

22. Boyer JC, Umar A, Risinger JI, et al: Microsatellite instability,mismatch repair deficiency, and genetic defects in human cancercell lines. Cancer Res 55: 6063-6070, 1995.

23. Fedier A, Poyet C, Perucchini D, Boulikas T and Fink D: MLH1-deficient tumor cells are resistant to lipoplatin, but retain sensitivityto lipoxal. Anticancer Drugs 17: 315-323, 2006.

24. Komatsu N, Kawamata N, Takeuchi S, et al: SAHA, an HDACinhibitor, has profound anti-growth activity against non-smallcell lung cancer cells. Oncol Rep 15: 187-191, 2006.

25. Li H and Wu X: Histone deacetylase inhibitor, Trichostatin A,activates p21WAF1/CIP1 expression through downregulation ofc-myc and release of the repression of c-myc from the promoterin human cervical cancer cells. Biochem Biophys Res Commun324: 860-867, 2004.

26. Heider U, Kaiser M, Sterz J, et al: Histone deacetylase inhibitorsreduce VEGF production and induce growth suppression andapoptosis in human mantle cell lymphoma. Eur J Haematol 76:42-50, 2006.

27. Fink D, Zheng H, Nebel S, et al: In vitro and in vivo resistanceto cisplatin in cells that have lost DNA mismatch repair. CancerRes 57: 1841-1845, 1997.

28. Qiu L, Burgess A, Fairlie DP, Leonard H, Parsons PG andGabrielli BG: Histone deacetylase inhibitors trigger a G2checkpoint in normal cells that is defective in tumor cells. MolBiol Cell 11: 2069-2083, 2000.

29. Mitsiades N, Mitsiades CS, Richardson PG, et al: Molecularsequelae of histone deacetylase inhibition in human malignant Bcells. Blood 101: 4055-4062, 2003.

30. Bali P, Pranpat M, Swaby R, et al: Activity of suberoylanilidehydroxamic acid against human breast cancer cells withamplification of her-2. Clin Cancer Res 11: 6382-6389, 2005.

31. Richon VM, Sandhoff TW, Rifkind RA and Marks PA: Histonedeacetylase inhibitor selectively induces p21WAF1 expressionand gene-associated histone acetylation. Proc Natl Acad SciUSA 97: 10014-10019, 2000.

32. Sakajiri S, Kumagai T, Kawamata N, Saitoh T, Said JW andKoeffler HP: Histone deacetylase inhibitors profoundly decreaseproliferation of human lymphoid cancer cell lines. Exp Hematol33: 53-61, 2005.

33. Bradbury CA, Khanim FL, Hayden R, et al: Histone deacetylasesin acute myeloid leukaemia show a distinctive pattern ofexpression that changes selectively in response to deacetylaseinhibitors. Leukemia 19: 1751-1759, 2005.

34. Ropero S, Fraga MF, Ballestar E, et al: A truncating mutationof HDAC2 in human cancers confers resistance to histonedeacetylase inhibition. Nat Genet 38: 566-569, 2006.


633-641 24/7/07 11:03 Page 641

6. Conclusion and perspective If drug resistance could be overcome, the impact on survival would with no doubt be highly

significant. A large variety of different strategies aimed at overcoming drug resistance

associated with tumor cells have been developed in the passed years. These include

pharmacologic, genetic, or viral restoration of their ability to properly respond to anticancer

therapies. Some strategies have even advanced into clinical applications. The advances in DNA

microarray and proteomic technology, the researchers’ ability to characterize the signaling

pathways involved in regulating tumor cell response to chemotherapy more completely than

ever before, and the ongoing development of new targeted therapies will open up new

opportunities to combat drug resistance. The ability to predict response to chemotherapy and to

modulate this response with targeted therapies will permit selection of the best treatment for

individual patients.

The prevention of drug resistance acquisition, the overcoming of intrinsic drug resistance,

and the understanding of the underlying mechanisms of drug resistance (development) are of

continuing interest in our research laboratory. Two of the key questions are:

6.1. What is the molecular basis of the acquired resistance by HDAC inhibitors?

We and others have shown that HDAC inhibitors can cause drug resistance development in

tumor cells. However, the mechanisms underlying the above described and putatively novel

type of acquired resistance by the HDAC inhibitor SAHA is still not understood at this point.

Ongoing studies in our laboratory are thus aimed at uncovering the molecular basis of this type

of acquired resistance. The data accumulated so far indicate that SAHA-induced resistance in

HCT116 colorectal adenocarcinoma cells correlates with histone hypoacetylation. This, together

with the observation that multidrug-resistance transporters are not affected (see

PUBLICATION III; 18), suggests that SAHA may not do its job, i.e. may not inhibit HDAC

activity, and thus does not result in acetylation of the histones in the SAHA-resistant cells. One

of the questions is what is the reason why SAHA does not produce accumulation of acetylated

histones in the resistant cells. Several possibilities are plausible and are summarized in a model

chart (Figure 1). The failure of histone acetylation (hypoacetylation) may arise (i) from reduced

uptake of SAHA, (ii) from increased sequestration/detoxification of SAHA, (iii) from reduced

transport of SAHA into the nucleus (its compartment of action), or (iv) from decreased activity

of histone acetyltransferases. In addition, resistance could also arise from the dysregulation of

SAHA-mediated signaling pathways that favor survival and counteract cell death processes.

15

Figure 1: Simplified model showing the various histone and non-histone targets that are affected by histone deacetylase inhibitors (blue). SAHA enters cells either via passive diffusion or via active inward transport. Inside the cells SAHA binds to its substrates, i.e. HDACs, and blocks the catalytic site of the HDACs, resulting in accumulation of acetylated histones, in the decondensening of the chromatin (allowing for gene transcription), and in the accumulation of acetylated proteins (histone level). SAHA is proposed to also modulate mediators of signal transduction pathways (“off-target” level). The net effect of all this networking is the promotion of cellular responses that lead to cell death and growth inhibition and the abrogation of those that are survival-prone and promote growth. Also shown are the possible targets and mediators of (acquired) resistance to SAHA (red). Resistance to SAHA associated with histone hypoacetylation may be due 1) to reduced influx or 2) to increased efflux of SAHA (both processes reduce the intracellular concentration of SAHA); 3) to reduced transport of SAHA from the cytoplasm into the nucleus; 4) to the loss acetyltransferases HATs (prevents gene transcription); or 5) to mutations in HDAC genes (may prevent binding of SAHA). 6) It is also possible that resistance is due to dysregulation of signaling pathways responsive to HDAC inhibitors (shifts balance from cytodestructive towards cytoprotective properties).

16

It is also not clear why the expression of cell cycle regulation-associated proteins is

apparently similar in SAHA-resistant cells and in SAHA-sensitive cells, suggesting that the loss

of the G2/M checkpoint in the SAHA-resistant cells is not reflected by changes in the

expression pattern of these proteins. Pilot data also indicate that acquired resistance to TSA

seems to depend on the absence of MLH1 protein, suggesting that MLH1 protects via yet

unknown mechanisms from resistance acquisition by TSA. Another open issue is whether

SAHA-mediated resistance acquisition is a tumor cell line- and/or a tumor type-specific

phenomenon.

These issues are currently under investigation, which also includes microarray technology

in order to identify genes or groups of genes that are differentially regulated and expressed in

the sensitive and in the resistant cells. However, we cannot exclude at this point whether a

multifactorial alteration of different cell regulating pathways that probably arises due to the

epigenetic targeting of the HDAC inhibitors rather than one well understood single pathway

may underlie SAHA resistance.

6.2. Cytochrome c as therapeutic tool to overcome drug resistance?

An intriguing approach pursued in our laboratory is aimed at delivering cytochrome c (from

outside by a suitable carrier) directly into the cytoplasm where it associates with the

apoptosome and eventually induces apoptosis. Cytochrome c is a protein with a highly

conserved primary structure across different species localized in the mitochondrial

intermembrane space. Normally it transfers electrons between complexes III and IV as part of

the respiratory chain. Cytochrome c is also a critical protein in the intrinsic apoptotic pathway,

inducing apoptosis when it is accumulated in the cytosol in response to pro-apoptotic stimuli.

Cytosolic cytochrome c, together with APAF-1 and procpase-9, forms the apoptosome, which

then activates the effector caspases-3 and -7 (80,81). Earlier studies have shown that direct

microinjection of cytochrome c into the cytoplasm of a cell or delivery by electroporation

activates apoptosis (82,83). As these applications are hardly suitable for treatments in patients,

we use cytochrome c encapsulated in liposomes. This bio-compatible cytochrome c (drug)

delivery would also have an additional benefit: the reduction of the severe adverse side effects

often seen with chemotherapies using "classic" apoptosis-inducing agents. Preliminary data

from our laboratory are promising, as they indicate that cytochrome c-encapsulated liposome

but not liposomes without encapsulated cytochrome c induce apoptosis in HeLa cervical tumor

cells.

17

This approach may also be useful for drug resistance reversal (Figure 2). We propose that

cytochrome c delivered this way would induce apoptosis even in tumor cells resistant to

apoptosis-inducing anticancer agents. This is because cytochrome c, coming into action

downstream of apoptosis-inducing signaling cascades and independently of additional pro-

apoptotic stimuli, would overcome the lack of these apoptosis-signaling processes. Such

processes are frequently absent in drug-resistant tumor cells, for instance due to MDR

overexpression or due to the defective apoptosis initiation machinery.

Figure 2: Tentative chart of restoration of drug sensitivity by delivering exogenous cytochrome c (cyto c) into the cytoplasm by a liposomal carrier. (Right insert) Tumor cell that has become drug-resistant due to overexpression of multidrug resistance transporters (e.g. MDR), drug-inactivating proteins (e.g glutathione/glutathione-transferase), or anti-apoptotic factors (e.g. Bcl-2), DNA damage tolerance (e.g. MMR), or dysregulated stress-responsive pathways (e.g. p53) and therefore displays insufficient (endogenous) apoptosis signaling. Consequently, cyto c is not released from the mitochondria, the apoptosome is not formed, and caspase-mediated apoptosis is not initiated (A). (Left) Liposome-encapsulated exogenous cyto c delivered to the tumor cell by endocytosis (1), forming an endosome-encapsulated/cyto c-liposome “hybrid” (2). Cyto c is released from the endosome/liposome into the cytoplasm (3), thereby bypassing the absence of mitochondrial cyto c release due to lack of or insufficient (endogenous) apoptosis signaling. This exogenously delivered, cytoplasmic cyto c is incorporated together with APAF-1 and pro-caspase-9 into the apoptosome (4). Subsequent activation of pro-caspase-9 cleaves (activates) the pro-caspases-3 and/or -7 which, in turn, cleave PARP-1 and other substrates, finally leading to the destruction of the tumor cell by apoptosis (5).

18

7. References

1. Teicher BA. Cancer drug resistance (Book). Humana Press, Totowa, NJ, 2006.

2. Oren M. p53: the ultimate tumor suppressor gene? FASEB J. 1992;6:3169-76.

3. Vousden KH, Lu X. Live or let die: the cell's response to p53. Nat Rev Cancer. 2002;2:594-

604.

4. Harris SL, Levine AJ. The p53 pathway: positive and negative feedback loops. Oncogene.

2005;24:2899-908.

5. Geisler S, Lonning PE, Aas T, Johnsen H, Fluge O, Haugen DF, Lillehaug JR, Akslen LA,

Borresen-Dale AL. Influence of TP53 gene alterations and c-erbB-2 expression on the

response to treatment with doxorubicin in locally advanced breast cancer. Cancer Res.

2001;61:2505-12.

6. Bunz F, Hwang PM, Torrance C, Waldman T, Zhang Y, Dillehay L, Williams J, Lengauer

C, Kinzler KW, Vogelstein B. Disruption of p53 in human cancer cells alters the responses

to therapeutic agents. J Clin Invest. 1999;104:263-9.

7. Wiman KG. Strategies for therapeutic targeting of the p53 pathway in cancer. Cell Death

Differ. 2006;13:921-6.

8. Fedier A, Ruefenacht UB, Schwarz VA, Haller U, Fink D. Increased sensitivity of p53-

deficient cells to anticancer agents due to loss of Pms2. Br J Cancer. 2002;87:1027-33.

9. Vikhanskaya F, Colella G, Valenti M, Parodi S, D'Incalci M, Broggini M. Cooperation

between p53 and hMLH1 in a human colocarcinoma cell line in response to DNA damage.

Clin Cancer Res. 1999;5:937-41.

10. Lin X, Howell SB. DNA mismatch repair and p53 function are major determinants of the

rate of development of cisplatin resistance. Mol Cancer Ther. 2006;5:1239-47.

11. Fedier A, Schlamminger M, Schwarz VA, Haller U, Howell SB, Fink D. Loss of atm

sensitises p53-deficient cells to topoisomerase poisons and antimetabolites. Ann Oncol.

2003;14:938-45.

12. Fedier A, Steiner RA, Schwarz VA, Lenherr L, Haller U, Fink D. The effect of loss of

Brca1 on the sensitivity to anticancer agents in p53-deficient cells. Int J Oncol.

2003;22:1169-73.

19

13. Fedier A, Moawad A, Haller U, Fink D. p53-deficient cells display increased sensitivity to

anthracyclines after loss of the catalytic subunit of the DNA-dependent protein kinase. Int J

Oncol. 2003;23:1431-7.

14. Fedier A, Stuedli A, Fink D. Presence of MLH1 protein aggravates the potential of the

HSP90 inhibitor radicicol to sensitize tumor cells to cisplatin. Int J Oncol. 2005;27:1697-

705.

15. Fedier A, Erdmann R, Boulikas T, Fink D. Potential of the Akt inhibitor LY294005 to

antagonize the efficacy of Cisplatin against HCT116 tumor cells in a DNA mismatch

repair-dependent manner. Int J Oncol. 2006;29:1303-10.

16. Marks PA. Discovery and development of SAHA as an anticancer agent. Oncogene.

2007;26:1351-6.

17. Xu W, Ngo L, Perez G, Dokmanovic M, Marks PA. Intrinsic apoptotic and thioredoxin

pathways in human prostate cancer cell response to histone deacetylase inhibitor. Proc Natl

Acad Sci U S A. 2006;103:15540-5.

18. Fedier A, Dedes KJ, Imesch P, Von Bueren AO, Fink D. The histone deacetylase inhibitors

suberoylanilide hydroxamic (Vorinostat) and valproic acid induce irreversible and MDR1-

independent resistance in human colon cancer cells. Int J Oncol. 2007;31:633-41.

19. Xiao JJ, Huang Y, Dai Z, Sadee W, Chen J, Liu S, Marcucci G, Byrd J, Covey JM, Wright

J, Grever M, Chan KK. Chemoresistance to depsipeptide FK228 [(E)-(1S,4S,10S,21R)-7-

[(Z)-ethylidene]-4,21-diisopropyl-2-oxa-12,13-dithia-5,8,20,23-tetraazabicyclo[8,7,6]-

tricos-16-ene-3,6,9,22-pentanone] is mediated by reversible MDR1 induction in human

cancer cell lines. J Pharmacol Exp Ther 2005;314:467-75.

20. Yamada H, Arakawa Y, Saito S, Agawa M, Kano Y, Horiguchi-Yamada J. Depsipeptide-

resistant KU812 cells show reversible P-glycoprotein expression, hyper-acetylated histones,

and modulated gene expression profile. Leuk Res 2006;30:723-34.

21. O'Driscoll L, Clynes M. Biomarkers and multiple drug resistance in breast cancer. Curr

Cancer Drug Targets. 2006;6:365-84.

22. Kurzeder C, Sauer G, Deissler H. Molecular targets of ovarian carcinomas with acquired

resistance to platinum/taxane chemotherapy. Curr Cancer Drug Targets. 2006;6:207-27.

23. Awtrey CS, Cadungog MG, Leitao MM, Alektiar KM, Aghajanian C, Hummer AJ, Barakat

RR, Chi DS. Surgical resection of recurrent endometrial carcinoma. Gynecol Oncol. 2006

Sep;102:480-8.

20

24. Padilla LA, Leung BS, Carson LF. Evidence of an association between human

papillomavirus and impaired chemotherapy-induced apoptosis in cervical cancer cells.

Gynecol Oncol. 2002;85:59-66.

25. Vousden KH. p53: death star. Cell. 2000;103:691-4.

26. Harms K, Nozell S, Chen X. The common and distinct target genes of the p53 family

transcription factors. Cell Mol Life Sci. 2004;61:822-42.

27. Mashima T, Tsuruo T. Defects of the apoptotic pathway as therapeutic target against

cancer. Drug Resist Updat. 2005;8:339-43.

28. Green DR, Chipuk JE. p53 and metabolism: Inside the TIGAR. Cell. 2006;126:30-2.

29. Dimri GP. What has senescence got to do with cancer? Cancer Cell. 2005;7:505-12.

30. Toledo F, Wahl GM. Regulating the p53 pathway: in vitro hypotheses, in vivo veritas. Nat

Rev Cancer. 2006;6:909-23.

31. Li YJ, Sanson M, Hoang-Xuan K, Delattre JY, Poisson M, Thomas G, Hamelin R.

Incidence of germ-line p53 mutations in patients with gliomas. Int J Cancer. 1995;64:383-7.

32. Royds JA, Iacopetta B. p53 and disease: when the guardian angel fails. Cell Death Differ.

2006;13:1017-26.

33. Aas T, Borresen AL, Geisler S, Smith-Sorensen B, Johnsen H, Varhaug JE, Akslen LA,

Lonning PE. Specific P53 mutations are associated with de novo resistance to doxorubicin

in breast cancer patients. Nat Med. 1996;2:811-4.

34. Modrich P, Lahue R. Mismatch repair in replication fidelity, genetic recombination, and

cancer biology. Annu Rev Biochem. 1996;65:101-133.

35. Jiricny J. Replication errors: cha(lle)nging the genome. EMBO J. 1998;17:6427-36.

36. Jiricny J. The multifaceted mismatch-repair system. Nat Rev Mol Cell Biol. 2006;7:335-46.

37. Kolodner RD, Marsischky GT. Eukaryotic DNA mismatch repair. Curr Opin Genet Dev.

1999;9:89-96.

38. Fishel R, Kolodner RD. Identification of mismatch repair genes and their role in the

development of cancer. Curr Opin Genet Dev. 1995;5:382-95.

39. Jiricny J, Nyström-Lahti M. Mismatch repair defects in cancer. Curr Opin Genet Dev.

2000;10:157-61.

21

40. Peltomaki P. Role of DNA mismatch repair defects in the pathogenesis of human cancer. J

Clin Oncol. 2003;21:1174-9.

41. Bellacosa A. Functional interactions and signalling properties of mammalian DNA

mismatch repair proteins. Cell Death Diff. 2001;8:1076-92.

42. Fedier A, Fink D. Mutations in DNA mismatch repair genes: Implications for DNA damage

signaling and drug sensitivity. Int J Oncol. 2004;24:1039-47.

43. Fink D, Aebi S, Howell SB. The role of DNA mismatch repair in drug resistance. Clin

Cancer Res. 1998;4:1-6.

44. Fedier A, Schwarz VA, Walt H, Delli Carpini R, Haller U, Fink D. Resistance to

topoisomerase poisons due to loss of DNA mismatch repair. Int J Cancer. 2001;93:571-6.

45. Fedier A, Fowst C, Tursi J, Geroni C, Haller U, Marchini S, Fink D: Brostallicin (PNU-

166196) - a new DNA minor groove binder that retains sensitivity in DNA mismatch

repair-deficient tumour cells. Br J Cancer. 2003;89:1559-65.

46. Fedier A, Poyet C, Perucchini D, Boulikas T, Fink D. MLH1-deficient tumor cells are

resistant to lipoplatin, but retain sensitivity to lipoxal. Anticancer Drugs. 2006;17:315-23.

47. Schwarz VA, Hornung R, Fedier A, Fehr MK, Walt H, Haller U, Fink D. Photodynamic

therapy of DNA mismatch repair-deficient and -proficient tumour cells. Br J Cancer.

2002;86:1130-5.

48. Friedman HS, Johnson SP, Dong Q, Schold SC, Rasheed BK, Bigner SH, Ali-Osman F,

Dolan E, Colvin OM, Houghton P, Germain G, Drummond JT, Keir S, Marcelli S, Bigner

DD and Modrich P: Methylator resistance mediated by mismatch repair deficiency in a

glioblastoma multiforme xenograft. Cancer Res. 1997;57:2933-6.

49. Samimi G, Fink D, Varki NS, Husain A, Hoskins WJ, Alberts DS, Howell SB. Analysis of

MLH1 and MSH2 expression in ovarian cancer before and after platinum drug-based

chemotherapy. Clin Cancer Res. 2000;6:1415-21.

50. Mackay HJ, Cameron D, Rahilly M, Mackean MJ, Paul J, Kaye SB, Brown R. Reduced

MLH1 expression in breast tumors after primary chemotherapy predicts disease-free

survival. J Clin Oncol. 2000;18:87-93.

51. Drake AC, Campbell H, Porteous ME, Dunlop MG. The contribution of DNA mismatch

repair gene defects to the burden of gynecological cancer. Int J Gynecol Cancer.

2003;13:262-77.

22

52. Dokmanovic M, Marks PA. Prospects: histone deacetylase inhibitors. J Cell Biochem.

2005;96:293-304.

53. Liu T, Kuljaca S, Tee A, Marshall GM. Histone deacetylase inhibitors: multifunctional

anticancer agents. Cancer Treat Rev. 2006;32:157-65.

54. Bolden JE, Peart MJ, Johnstone RW. Anticancer activities of histone deacetylase inhibitors.

Nat Rev Drug Discov. 2006;5:769-84.

55. Glaser KB. HDAC inhibitors: Clinical update and mechanism-based potential. Biochem

Pharmacol. 2007;74:659-71.

56. Villar-Garea A, Esteller M. Histone deacetylase inhibitors: understanding a new wave of

anticancer agents. Int J Cancer 2004;112:171-8.

57. Minucci S, Pelicci PG. Histone deacetylase inhibitors and the promise of epigenetic (and

more) treatments for cancer. Nat Rev Cancer. 2006;6:38-51.

58. Xu WS, Parmigiani RB, Marks PA. Histone deacetylase inhibitors: molecular mechanisms

of action. Oncogene. 2007;26:5541-52.

59. Michaelis M, Michaelis UR, Fleming I, Suhan T, Cinatl J, Blaheta RA, Hoffmann K,

Kotchetkov R, Busse R, Nau H, Cinatl J Jr. Valproic acid inhibits angiogenesis in vitro and

in vivo. Mol Pharmacol. 2004;65:520-7.

60. Duvic M, Vu J. Vorinostat: a new oral histone deacetylase inhibitor approved for cutaneous

T-cell lymphoma. Expert Opin Investig Drugs. 2007;16:1111-20.

61. Duvic M, Talpur R, Ni X, Zhang C, Hazarika P, Kelly C, Chiao JH, Reilly JF, Ricker JL,

Richon VM, Frankel SR. Phase 2 trial of oral vorinostat (suberoylanilide hydroxamic acid,

SAHA) for refractory cutaneous T-cell lymphoma (CTCL). Blood. 2007;109:31-9.

62. Garber K. HDAC inhibitors overcome first hurdle. Nat Biotechnol. 2007;25:17-9.

63. Butler LM, Agus DB, Scher HI, Higgins B, Rose A, Cordon-Cardo C, Thaler HT, Rifkind

RA, Marks PA, Richon VM. Suberoylanilide hydroxamic acid, an inhibitor of histone

deacetylase, suppresses the growth of prostate cancer cells in vitro and in vivo. Cancer Res.

2000 ;60:5165-70.

64. Richon VM, Sandhoff TW, Rifkind RA, Marks PA. Histone deacetylase inhibitor

selectively induces p21WAF1 expression and gene-associated histone acetylation. Proc

Natl Acad Sci U S A. 2000;97:10014-9.

23

65. Mitsiades N, Mitsiades CS, Richardson PG, McMullan C, Poulaki V, Fanourakis G,

Schlossman R, Chauhan D, Munshi NC, Hideshima T, Richon VM, Marks PA, Anderson

KC. Molecular sequelae of histone deacetylase inhibition in human malignant B cells.

Blood. 2003;101:4055-62.

66. Ungerstedt JS, Sowa Y, Xu WS, Shao Y, Dokmanovic M, Perez G, Ngo L, Holmgren A,

Jiang X, Marks PA. Role of thioredoxin in the response of normal and transformed cells to

histone deacetylase inhibitors. Proc Natl Acad Sci U S A. 2005;102:673-8.

67. Rosato RR, Grant S. Histone deacetylase inhibitors: insights into mechanisms of lethality.

Expert Opin Ther Targets. 2005;9:809-24.

68. Glaser KB. Defining the role of gene regulation in resistance to HDAC inhibitors--

mechanisms beyond P-glycoprotein. Leuk Res. 2006 ;30:651-2.

69. Ahmed M, Rahman N. ATM and breast cancer susceptibility. Oncogene. 2006;25:5906-11.

70. Shiloh Y. The ATM-mediated DNA-damage response: taking shape. Trends Biochem Sci.

2006;31:402-10.

71. Gudmundsdottir K, Ashworth A. The roles of BRCA1 and BRCA2 and associated proteins

in the maintenance of genomic stability. Oncogene. 2006;25:5864-74.

72. Turner N, Tutt A, Ashworth A. Targeting the DNA repair defect of BRCA tumours. Curr

Opin Pharmacol. 2005;5:388-93.

73. Levy-Lahad E, Friedman E. Cancer risks among BRCA1 and BRCA2 mutation carriers. Br

J Cancer. 2007;96:11-5.

74. Collis SJ, DeWeese TL, Jeggo PA, Parker AR. The life and death of DNA-PK. Oncogene.

2005;24:949-61.

75. Vilenchik MM, Knudson AG. Endogenous DNA double-strand breaks: production, fidelity

of repair, and induction of cancer. Proc Natl Acad Sci U S A. 2003;100:12871-6.

76. Solit DB, Rosen N. Hsp90: a novel target for cancer therapy. Curr Top Med Chem.

2006;6:1205-14.

77. Zhao R, Houry WA. Molecular interaction network of the Hsp90 chaperone system. Adv

Exp Med Biol. 2007;594:27-36.

78. Vivanco I, Sawyers CL. The phosphatidylinositol 3-Kinase AKT pathway in human cancer.

Nat Rev Cancer. 2002;2:489-501.

24

79. Crowell JA, Steele VE, Fay JR. Targeting the AKT protein kinase for cancer

chemoprevention. Mol Cancer Ther. 2007;6:2139-48.

80. Garrido C, Galluzzi L, Brunet M, Puig PE, Didelot C, Kroemer G. Mechanisms of

cytochrome c release from mitochondria. Cell Death Differ. 2006;13:1423-33.

81. Riedl SJ, Salvesen GS. The apoptosome: signalling platform of cell death. Nat Rev Mol

Cell Biol. 2007;8:405-13.

82. Li F, Srinivasan A, Wang Y, Armstrong RC, Tomaselli KJ, Fritz LC. Cell-specific

induction of apoptosis by microinjection of cytochrome c. Bcl-xL has activity independent

of cytochrome c release. J Biol Chem. 1997;272:30299-305.

83. Gabriel B, Sureau F, Casselyn M, Teissie J, Petit PX. Retroactive pathway involving

mitochondria in electroloaded cytochrome c-induced apoptosis. Protective properties of

Bcl-2 and Bcl-XL. Exp Cell Res. 2003;289:195-210.

25

university of zurich - zora.uzh.ch filesuppressor pathway in cancer cells. we showed that genetic...

Documents