Báo cáo khoa học: Reduced FAS transcription in clones of U937 cells that have acquired resistance to Fas-induced apoptosis ppt

Clones with such reduced Fas expression also displayed partial cross-resistance to tumour necrosis factor-a stimulation, but the mRNA expression of tumour necro-sis factor receptors was

Fas (CD95⁄ Apo-1) is a cell surface receptor that is

important for the mediation of cell death, and is one

of eight different death receptors that have been

char-acterized to date [1] The role of the Fas ligand (FasL)

and receptor interaction has been emphasized in the

function of cytotoxic T lymphocytes and in the control

of immune cell homeostasis [2,3] Oligomerization of

Fas via binding of its cognate ligand (FasL) induces a signalling cascade that culminates in the controlled degradation of cellular components [4] The apical caspases-8⁄ 9, together with the downstream effector caspase-3, have been documented to be crucial players

in the mediation of death receptor-induced apop-tosis [5]

Correspondence

J Blomberg, Department of Molecular

Biology, Umea˚ University, S-90187 Umea˚,

Sweden

Fax: +46 90 771420

Tel: +46 90 7852535

E-mail: jeanette.blomberg@molbiol.umu.se

(Received 16 July 2008, revised 6

November 2008, accepted 12 November

2008)

doi:10.1111/j.1742-4658.2008.06790.x

molecular changes that occur during selection for insensitivity to Fas-induced apoptosis, a resistant variant of the U937 cell line was established Individual resistant clones were isolated and characterized The most frequently observed defect in the resistant cells was reduced Fas expression, which correlated with decreased FAS transcription Clones with such reduced Fas expression also displayed partial cross-resistance to tumour necrosis factor-a stimulation, but the mRNA expression of tumour necro-sis factor receptors was not decreased Reintroduction of Fas conferred susceptibility to Fas but not to tumour necrosis factor-a stimulation, sug-gesting that several alterations could be present in the clones The reduced Fas expression could not be explained by mutations in the FAS coding sequence or promoter region, or by silencing through methylations Protein kinase B and extracellular signal-regulated kinase, components of signalling pathways downstream of Ras, were shown to be activated in some of the resistant clones, but none of the three RAS genes was mutated, and experi-ments using chemical inhibitors could not establish that the activation of these proteins was the cause of Fas resistance as described in other systems Taken together, the data illustrate that Fas resistance can be caused by reduced Fas expression, which is a result of an unidentified mode of regulation

Abbreviations

AKT, protein kinase B; CpG, cytosine-phosphate-guanine; ERK, extracellular signal-regulated kinase; FasL, Fas ligand; FLIP, Flice-like inhibitory protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MAP-kinase, activated protein kinase; MEK, mitogen-activated ERK-activating kinase; P.U cells, parental U937 cells; PARP, poly (ADP-ribose) polymerase; PI3-kinase, phosphoinositide 3-kinase; qRT-PCR, quantitative RT-PCR; SOCS-1, suppressor of cytokine signalling 1; TNFR, tumour necrosis factor receptor; TNF-a, tumour necrosis factor-a; TRAIL, TNF-related apoptosis-inducing ligand.

Expression of Fas and susceptibility to FasL

activa-tion are common traits in most tissues [6] In contrast,

tumour cells frequently display impaired death

recep-tor functions [7] Nonfunctional Fas signalling has

been implicated in the resistance to apoptosis induced

by chemical stimuli Moreover, a lack of Fas function

enables tumour cells to evade surveillance by the

immune system and facilitates metastatic progression

[8,9] Several different mechanisms that contribute to

impaired Fas signalling have been described in a range

of tumours, such as suppressed expression of Fas at

both the mRNA and protein levels [10,11] Epigenetic

silencing of the FAS promoter has recently been shown

to be regulated by oncogenic Ras [12,13] Furthermore,

mutations or deletions in FAS have been found to

cause an autoimmune lymphoproliferative syndrome

[14] The Fas signalling pathway can be modified by

an array of proteins, e.g Flice-like inhibitory proteins

(FLIPs), Bcl-2 family members, Fas-associated protein

tyrosine phosphatase 1 and inhibitors of apoptosis

proteins [15]

To establish which molecular alterations occur

dur-ing the acquisition of Fas resistance, we established

Fas-resistant U937 cells by prolonged growth in

pro-gressively increasing concentrations of

apoptosis-induc-ing Fas antibody We have demonstrated previously

that the resistant phenotype is associated with multiple

molecular changes, such as reduced Fas expression,

increased cFLIP expression and altered activities of

both protein tyrosine kinases and protein tyrosine

phosphatases Moreover, selection for Fas resistance

results in coselection for resistance to other death

receptor ligands [16] In this study, individual clones

derived from the Fas-resistant population were used to

dissect the resistance mechanisms in the heterogeneous

resistant population described earlier The results

showed that a decrease in Fas expression was the most

prominent reason for resistance, and that the

reintro-duction of Fas expression abolished Fas resistance

Notably, the reduced amount of Fas in the resistant

clones was a result of impaired FAS transcription

Results

Downregulation of Fas as a means to develop

Fas resistance

Previously, we have developed Fas-resistant U937 cells

by prolonged growth in progressively increasing

con-centrations of stimulating Fas antibody [16] To

inves-tigate the molecular basis that underlies the resistance

to Fas-induced apoptosis, the resistant cells were

seeded out as single cells in microtitre wells and 39

different clones were established These clones showed various degrees of Fas sensitivity, ranging from 50% sensitivity to complete resistance compared with paren-tal U937 (P.U.) cells (Fig 1A) (Apoptosis was detected as illustrated in Fig S1.) When the surface expression of Fas was determined by flow cytometry in the 39 resistant clones, it was evident that 82% of the clones contained a reduced amount of Fas (Fig 1B)

As no correlation between Fas expression and Fas sen-sitivity was detected in this experiment (R = 0.1446), four clones with low Fas expression (F7, F23, F30 and F35) and two with high Fas expression (F1 and F33) within the more resistant area (< 10% apoptosis) were randomly selected All six clones were resistant for up

to 48 h of Fas stimulation, which demonstrated that a persistent resistance had been obtained (data not shown) Four of the six clones contained reduced expression of Fas protein when assayed both with flow cytometry and immunoblot (Fig 1C,D) The decrease

in Fas expression at the cell surface was approximately 60% When activation of the Fas signalling cascade was investigated with immunoblot on cleaved caspase-8 and its downstream caspase target, poly(ADP-ribose) polymerase (PARP), no cleavage was detected in any

of the resistant clones compared with parental cells (Fig 1E) In summary, the data illustrate that four of the six resistant clones (F7, F23, F30 and F35) exhibit reduced Fas expression, whereas clones F1 and F33 have acquired alternative abnormalities that abolish caspase-8 activation

A decreased expression of Fas could be the result of alterations at many different levels, including transcrip-tion, mRNA half-life, translation and protein turnover

As an initial attempt to elucidate the mechanism behind Fas resistance, conventional RT-PCR and quantitative RT-PCR (qRT-PCR) were performed to explore whether FAS transcription was reduced in the resistant clones Three different primer pair sets (Fig 2A) were used: one control set designed against the mRNA encoding the ubiquitous enzyme gly-ceraldehyde-3-phosphate dehydrogenase (GAPDH), another set encompassing exon–exon boundaries of the Fas mRNA, with which steady-state mRNA levels could be measured, and a third set encompassing the exon–intron boundary of the Fas pre-mRNA, with which newly synthesized transcripts could be mea-sured As shown in Fig 2B, clones F7, F23, F30 and F35 exhibited a reduced steady-state level of Fas tran-script compared with clone F1, clone F33 and P.U cells The reduction in clones F7, F23, F30 and F35 measured 40% or less by qRT-PCR (Fig 2C) Inter-estingly, the same four clones also showed an 80% decrease in the amount of Fas pre-mRNA (Fig 2D,E),

suggesting that the reduced amount of Fas protein

was a result of events regulating FAS transcription

qPCR on genomic FAS did not reveal any differences

between parental and resistant cells (data not shown)

Thus, the decrease in FAS transcription did not

depend on a reduced amount of FAS gene copies in

the resistant clones Mutations in both the promoter

and coding sequences of FAS have been identified in a

vast range of tumours (reviewed in [8]) However, no

acquired mutations in the Fas cDNA or FAS

promoter, spanning the region from)1781 to )22 bp,

were detected in our resistant clones when compared with P.U cells (data not shown)

Fas-resistant clones with reduced Fas expression are partially cross-resistant to tumour necrosis factor-a (TNF-a) independent of restored Fas transcription

Cross-resistance to the activation of other members of the death receptor family was investigated in the clones with reduced Fas expression Rather surprisingly, the

Fig 1 Reduced Fas expression in resistant clones Thirty-nine individual clones were isolated from the Fas-resistant variant of U937 described earlier (A) Fas sensitivity was monitored by propidium iodide staining and flow cytometry after 15 h of stimulation with

20 ngÆmL)1a-Fas in P.U cells and resistant clones (numbered) Surface expression of Fas was investigated by flow cytometry in 39 Fas-resistant clones (B) and six Fas-resistant clones that were randomly selected for further studies (C) Fas was stained with a nonapoptosis-induc-ing antibody, as described in Experimental procedures, and the secondary antibody alone served as a negative control (D) Immunoblottnonapoptosis-induc-ing was performed to determine the total amount of Fas protein in the six Fas-resistant clones (E) Cells were treated with or without

15 ngÆmL)1stimulating Fas antibody for 15 h before the processing of caspase-8 and PARP was investigated by immunoblot Actin was used as a loading control in all immunoblot experiments.

sensitivity to TNF-a was reduced in all four clones

when compared with P.U cells (Fig 3A, P < 0.02)

In contrast, susceptibility to TNF-related

apoptosis-inducing ligand (TRAIL) was similar to that of P.U

cells in all clones, except clone 23, which showed a

70% decrease in the apoptotic response (P = 0.03)

(Fig 3B) This cross-resistance prompted us to

investi-gate whether a general defect in apoptosis had been acquired We therefore treated U937 cells and the Fas-resistant clones with the anticancer drug etoposide, which is a known activator of caspases and inducer of apoptosis, independent of Fas expression [17,18] As the sensitivity to etoposide was similar in all clones and P.U cells (Fig 3C), it was concluded that there was no general apoptosis defect in the resistant clones

In addition, because only one clone displayed cross-resistance to TRAIL, we hypothesized that the decreased response to TRAIL represented a second independent event, whereas the reduced sensitivity to TNF-a was associated with acquired Fas resistance U937 cells have been reported to express both tumour necrosis factor receptor 1 (TNFR1) and TNFR2 [19] To investigate whether the reduced sensi-tivity to TNF-a in the Fas-resistant clones could be explained by a concomitant decreased expression of TNFRs, qRT-PCRs were performed with isoform-specific primers No drastic decrease in either TNFR1

or TNFR2 mRNAs was detected in the Fas-resistant clones compared with P.U cells which correlated with the reduced sensitivity to TNF-a stimulation (Fig 3D,E) Thus, the partial cross-resistance to TNF-a-induced apoptosis is not caused by a general suppression of death receptor transcription

The significance of the reduced Fas expression in the resistant clones was studied by the reintroduction of Fas by transfection into two of the clones The success

of Fas expression was confirmed by immunoblot anal-ysis (Fig 4A) of the total lysate and flow cytometry of Fas expressed on the cell surface (Fig 4B) Stimulation

of Fas confirmed that the re-established Fas expression mediated susceptibility to Fas-induced apoptosis in the resistant cells (Fig 4C) The restored apoptotic response to Fas stimulation also correlated with a normal activation of caspase-8 (Fig 4D) Despite the fact that the Fas-resistant clones were cross-resistant

to TNF-a, the reintroduction of Fas did not mediate renewed sensitivity to TNF-a This suggests that either the FAS gene and components of TNF signalling are downstream of a common regulator mutated in these clones, or these two resistance mechanisms have arisen independently of each other (Fig 4E)

Potential regulators of Fas expression

It has been shown that oncogenic factors, such as Ras and p53, are important regulators of Fas expression Ras inhibition of Fas expression is associated with hypermethylation of cytosine-phosphate-guanine (CpG)-rich regions in the FAS promoter [12,13], and demethylation of CpG sites in the first intron of FAS

Fig 2 Decreased amount of Fas pre-mRNA in the resistant clones.

(A) Schematic presentation of the location of the primers used for

RT-PCR and qRT-PCR studies GAPDH was used as a control and

the primers were designed to allow the determination of

contami-nating genomic DNA RT-PCR (B) and qRT-PCR (C) were performed

with primer pair I to analyse Fas mRNA expression in the different

clones Primer pair II was used for the study of Fas pre-mRNA

expression with RT-PCR (D) and qRT-PCR (E) RT-PCR products

were visualized with ethidium bromide staining after separation on

a 1% agarose gel.

has been illustrated to enhance p53-induced Fas

expression [11] As the Fas-resistant cells contained a

reduced amount of Fas pre-mRNA (Fig 2D,E), a

potential repression of FAS expression through

epige-netic silencing was investigated The FAS gene

con-tains a 650 bp CpG island in the 5¢-flanking region of

the transcriptional start site (as illustrated in Fig 5A)

However, none of the unmethylated CpGs (black

boxes) in the CpG island were altered in any of the

resistant clones when assayed with genomic sequencing

of bisulfite-modified DNA (Fig 5B) In addition,

methylation-specific PCR of a CpG region in the first

intron of FAS revealed no methylations in any of the

clones (Fig 5C) Methylation-specific PCR of

suppres-sor of cytokine signalling 1 (SOCS-1) was used as a

positive control, as it has been reported to be

methy-lated in U937 (Fig 5C) [20] Treatment with a

deme-thylating agent, 5-aza-2¢-deoxycytidine, did not restore

either Fas surface expression or Fas sensitivity (data

not shown) Furthermore, no activating mutations of

codons 12, 13 and 61 in the genes encoding H-,

K- and N-Ras [21] could be detected in any of the

clones (data not shown), which would have provided

an explanation for the altered FAS transcription In

summary, this clearly shows that epigenetic silencing through CpG methylations does not account for the resistance to Fas-induced apoptosis in our system Survival signalling pathways downstream of Ras, such as the phosphoinositide 3-kinase (PI3-kinase) and the mitogen-activated protein kinase (MAP-kinase) sig-nalling cascades, have been suggested to regulate death receptor-induced apoptosis [22,23] Before we analysed the RAS genes for mutations, we performed immuno-blot analyses and used specific inhibitors to determine whether the PI3-kinase and MAP-kinase signalling cas-cades contributed to Fas resistance The Fas-resistant clones with decreased Fas expression (F7, F23, F30 and F35) contained elevated levels of phosphorylated protein kinase B (pAKT), whereas resistant clones with normal Fas expression (F1 and F33) contained less pAKT as well as total AKT protein (Fig 6A) Quanti-fication showed that the increase in pAKT was slight,

as it reached statistical significance only for clones F7 and F23 (Fig S2) Inhibition of the PI3-kinase path-way with wortmannin for 1.5 h completely abolished AKT phosphorylation (Fig 6B) However, wortman-nin pretreatment for 1 h before Fas stimulation did not restore the sensitivity to Fas stimulation in the

Fig 3 Partial cross-resistance to TRAIL and

TNF-a, but not etoposide, treatment in the

resistant clones Cells were stimulated with

2 ngÆmL)1TNF-a (A) and 2 ngÆmL)1TRAIL

(B) for 20 h (C) Cells were treated with

dif-ferent concentrations of etoposide for 16 h.

Apoptosis was assayed with propidium

iodide staining and flow cytometry in all

experiments Normal expression of TNFR1

and TNFR2 was detected in the resistant

clones when mRNA expression of TNFR1

(D) and TNFR2 (E) was measured with

qRT-PCR.

Fig 5 Methylation of the FAS promoter does not account for the reduced expression in the resistant clones (A) Schematic illustration of CpGs in the 650 bp 5¢-flanking region of the FAS promoter Black boxes are individual CpGs and the numbers in the white boxes represent the number of nucleotides between each site (B) Schematic illustration of the sequenced CpGs in the FAS promoter in parental and resis-tant clones, where the boxes represent unmethylated CpG sites (black), methylated CpG sites (grey, none present) and other nucleotides (white) (C) Primers that recognized either unmethylated or methylated sequences in the first intron of FAS were used to determine the methylation status with methyl-specific PCR Methylated primers for suppressor of cytokine signalling 1 (SOCS-1) were used as a positive control The PCR products were visualized with ethidium bromide staining after separation on a 2% agarose gel.

Fig 4 Reintroduction of Fas abolishes the insensitivity to Fas, but not TNF-a-induced apoptosis The introduced Fas expression was con-firmed by immunoblot (A), and the increased surface Fas expression was investigated by flow cytometry of surface-expressed Fas (B) (C) Vector control and Fas-expressing cells were stimulated with 15 ngÆmL)1of anti-Fas for 15 h, and apoptosis was assayed with propidium iodide staining and flow cytometry (D) Restored processing of caspase-8 on Fas activation was studied by immunoblot (E) TNF-a sensitivity was monitored in vector control and Fas-expressing cells by stimulation with 2 ngÆmL)1of TNF-a for 15 h, followed by propidium iodide staining and flow cytometry.

resistant clones and did not reduce the apoptotic

response in parental cells (Fig 6C) The same results

were obtained with another PI3-kinase inhibitor,

LY294002 (data not shown) These experiments

demonstrate that the altered level of pAKT is not

responsible for the Fas resistance

By performing immunoblots, we identified resistant

clones containing increased levels of phosphorylated

extracellular signal-regulated kinase 1 (pERK1),

possi-bly as a consequence of increased expression of the

ERK1 protein (Fig 6D) As these clones were not

amongst those with lower Fas expression, we ruled out

ERK1 as a mediator of FAS silencing ERK2

exhib-ited variable expression between experiments, and

there was no correlation between the presence of

ele-vated phosphorylation and increased Fas resistance In

addition, the upstream ERK kinases,

mitogen-acti-vated ERK-activating kinase 1 (MEK1) and MEK2,

were not excessively phosphorylated in any of the

clones, suggesting that whatever causes the irregular

hyperphosphorylation of ERK1⁄ 2 represents a

nonca-nonical pathway (Fig 6E) (For the quantification of

total and phosphorylated levels of ERK1⁄ 2 and

MEK1⁄ 2, see Fig S2.) To lend further support to this

notion, the inhibition of MEK1⁄ 2 for 1.5 h with

PD98059 resulted in increased ERK1⁄ 2

phosphoryla-tion in both Fas-resistant clones and P.U cells

(Fig 6F), which is opposite to what is expected with

this inhibitor The PD98059-induced phosphorylation

of ERK1⁄ 2 was stable for at least 16 h after

adminis-tration (Fig S3A) These obscure results were not caused by dysfunction of PD98059, as it blocked the fetal bovine serum-induced activation of ERK1⁄ 2 in starved HeLa cells and P.U cells (Fig S3B) Neverthe-less, PD98059 did not confer Fas resistance to P.U cells (Fig 6G), indicating that increased phosphoryla-tion of ERK1⁄ 2 is not sufficient to mediate the resis-tance to Fas-induced apoptosis In conclusion, this suggests that the canonical MAP-kinase signalling cascade does not have a direct regulatory role in the resistant cells

Discussion

Tumour development is a multistep process that evolves with time It is driven by a progressive increase in the acquisition of mutations and genetic aberrations The acquisition of genetic lesions that abrogate sensitivity to cell death signals is an important part of tumour devel-opment and progression, as it is needed to support increased proliferation [24] In addition, resistance to cell death enables tumour cells to avoid elimination triggered by both cytotoxic immune cells and thera-peutic agents [25] Through an increased understanding

of the mechanisms that mediate resistance to apoptosis, important improvements in therapeutic interventions can be made In this article, we have described a model system in which acquired Fas resistance is shown to be dependent on more than one mechanism; reduced Fas receptor expression was studied in detail

Fig 6 Altered activities of the PI3-kinase and MEK ⁄ ERK signalling pathways in resistant clones do not contribute to the reduced susceptibil-ity to Fas (A) Lysates were subjected to immunoblot of phosphorylated and total levels of AKT (B) Reduced phosphorylation of AKT with

1 l M wortmannin after 1.5 h of treatment was confirmed by immunoblot (C) Cells were pretreated with or without 1 l M wortmannin for

1 h before 10 ngÆmL)1of anti-Fas was added for 15 h Apoptosis was detected by propidium iodide staining and flow cytometry Phosphory-lated and total levels of ERK1 ⁄ 2 (D) and MEK1 ⁄ 2 (E) were investigated by immunoblot (F) Cells were treated with or without 50 l M

PD98059 for 1.5 h before the phosphorylated levels of ERK1 and ERK2 were investigated by immunoblot (G) Parental cells were pretreated with or without 50 l M PD98059 for 1 h before 10 ngÆmL)1of a-Fas was added Cells were harvested after 15 h of Fas stimulation, and apoptosis was determined by propidium iodide staining and flow cytometry.

The reduced expression of death receptors has been

detected in a considerable number of different

tumours, but the mechanism contributing to this

phe-notype is still not well defined and needs to be

deter-mined [8] In this article, we have shown that a

decrease in Fas expression is the major phenotype in

U937 cells selected for resistance to Fas, making these

cells a tractable model for the identification of the

pathways involved in the regulation of Fas expression

and resistance Interestingly, many of the clones

exhi-bit cross-resistance to other death receptor signals

Indeed, communication between different death

recep-tors has been postulated as they share certain

intracel-lular signalling molecules [26,27] In addition, the

presence of TNFRs is believed to be important for

appropriate susceptibility to Fas stimulation in mouse

T cells and macrophages [28–30] However, a

depen-dence of Fas expression on TNFR signalling has not

been reported, which would have provided an

explana-tion for our data if it had not been for the results

showing that the re-expression of Fas did not reverse

TNF-a resistance The partial cross-resistance to TNF-a

is puzzling and highlights the complexity of cell death

signalling

Death receptor activation has been illustrated to

mediate nonapoptotic signalling, which is most

promi-nent in TNF signalling [31] The balance between

pro-and anti-apoptotic signalling is tightly regulated by

important transcription factors, such as nuclear

factor-jB and c-Jun, which have also been demonstrated to

regulate FAS transcription [32–34] The interplay and

regulation of different transcription factors are highly

complex, and the cellular context is critical for

deter-mining the contribution of different factors for FAS

transcription [35] Preliminary data have illustrated

that there is reduced protein expression of c-Jun in our

resistant clones with low Fas expression when

com-pared with P.U cells (data not shown) The precise

function of c-Jun, however, is still controversial, as it

has different effects depending on the type of cell and

expression levels [36] Thus, the binding pattern of

several different transcription factors to the FAS

promoter is an interesting topic for further study

Apoptosis is only one of several cell death pathways,

e.g necrosis and autophagy [37] Reports by others

have shown that, if the apoptosis signalling pathway

via Fas is defective, an alternative route leading to

necrosis can be activated [38,39] Fas has also been

shown recently to stimulate autophagy when activated

by autoantibodies in neuroblastoma [40] However, the

importance and interplay between the different cell

death pathways are not fully understood, and the

involvement of death receptor expression has not been

investigated thoroughly Thus, the cooperation of death receptors in cell death is an interesting topic to explore further

Studies have implied that, even if tumour cells con-tain multiple genetic and epigenetic alterations, they can be completely dependent on only one or a few important gene alterations for survival and prolifera-tion [41] It has been shown that oncogenic factors exist which deregulate Fas expression [12,13,42–44] However, the sequencing of RAS did not reveal any alterations in our resistant cells (data not shown), and

it is unlikely that dysfunction of the tumour suppressor p53 can account for the resistant phenotype in our sys-tem, as U937 cells are TP53-null cells (data not shown [45]) Somatic mutation of FAS itself was first detected

in lymphoid tumours, and it has been reported to be present in considerable proportions of non-Hodgkin’s lymphoma [8] Sequencing of the coding region and the promoter, containing the core, enhancer and silen-cer region, of FAS did not reveal any mutations in the resistant cells Because tumour progression is tightly coupled to genetic changes, complete genomic sequenc-ing, comparative genomic hybridization or single nucleotide polymorphism analysis of resistant clones would be the most straightforward way to elucidate the genetic aberrations that occur during the acquisi-tion of Fas resistance The genomic sequencing of tumours has been shown to be a powerful method of identifying reoccurring alterations in the complex heterogenicity of different tumours [46]

Epigenetic changes, through promoter methylations

or other global chromatin modifications, have recently received increasing attention in cancer research, as it has become evident that alterations in methylation status are one of several important tumour character-istics [47] We could not detect methylation of any CpG sites in a 650-bp 5¢-flanking region of the FAS promoter by genomic sequencing of bisulfite-modified DNA This region has been studied by others, because it is recognized as a CpG island that contains many CpG sites [11] In addition, regions upstream

of the FAS promoter and in intron 1 have also been shown to be methylated [11,12] However, as the inhibition of DNA methylation did not result in an elevated level of Fas expressed on the surface, and did not restore sensitivity to Fas stimulation in resis-tant cells (data not shown), we conclude that epige-netic changes through methylation cannot account for the reduced expression of Fas

Survival pathways, such as the PI3-kinase and MAP-kinase pathways, regulate a multitude of cellular responses and have a major impact on cell viability Importantly, signalling molecules in these pathways

in ERK1⁄ 2 phosphorylation by PD98059 suggests that

this alternative pathway is negatively regulated by

MEK, but nonessential, as its activation is insufficient

to confer Fas resistance to P.U cells Future studies

need to be performed to identify this potentially

inter-esting mode of activation by the PD98059 ‘inhibitor’,

as studies like this may provide additional knowledge

on how the MEK⁄ ERK pathway is regulated by

scaf-fold proteins, kinases and phosphatases Taken

together, the observed alterations in the AKT- and

MAP-kinase pathways illustrate that the selection

pres-sure imposed on the Fas-resistant cells affects them in

a profound way, requiring several important signalling

cascades to be activated in order to support the

devel-opment of resistance This is an example of one of the

major challenges in the elucidation of tumour

develop-ment, namely to discriminate between secondary

sup-portive alterations and important tumour-maintaining

aberrations [48]

In other studies on acquired resistance to

Fas-induced apoptosis, it has been illustrated that the

resistant phenotype is associated with a loss of Fas

function through mutations in FAS or a dysfunctional

activation of the sphingomyelin–ceramide pathway

[49–51] In this article, we have reported that resistance

to Fas-induced apoptosis is caused by a decrease in

FAS transcription by an incompletely understood

mechanism The clones described here may therefore

be utilized in unbiased screens for components

involved in acquired Fas resistance, which may

gener-ate new targets for anticancer treatments

Experimental procedures

Cell culture and reagents

The human monocytic cell line U937 was obtained from the

American Type Culture Collection, Manassas, VA, USA

Cells were cultured in RPMI1640 medium supplemented with

from Sigma PD98059 was obtained from Calbiochem (EMD Biosciences Inc., Darmstadt, Germany)

Generation of FasL-resistant cells and clones

Fas-resistant U937 cells were established as described previ-ously [16] In order to isolate individual resistant clones from the resistant cell population, 0.1 and 0.3 cells per well were seeded out into microtitre wells and 39 single colonies were expanded Four clones with low Fas expression and two with high Fas expression within the more resistant area (< 10% apoptosis) were randomly selected for further studies

Apoptosis detection with flow cytometry

Cells were stimulated with different concentrations of Fas, TNF-a, TRAIL and etoposide, and analysed at the time points indicated in the figures For inhibition studies, cells were pretreated with either 1 lm wortmannin or 50 lm PD98059 for 60 min before stimulation with 15 ngÆmL)1 Fas antibody for 15 h Cell pellets were dissolved in propi-dium iodide solution (0.1% v⁄ v Nonidet P-40, 20 m Tris

pH 7.5, 100 mm NaCl, 50 lgÆmL)1 propidium iodide and

20 lgÆmL)1 RNAse) and incubated for 30 min at 4C before analysis was performed with a FACSCalibur flow cytometer equipped with cell quest pro software (BD Biosciences, San Jose, CA, USA) At least 1· 104

cells were acquired per sample and the sub-G1 population was scored as apoptotic cells Live and apoptotic cells were identified on the FL-2- and fetal bovine serum–height plots, and particles smaller than approximately 1000-fold relative

to live cells were excluded Doublets were discriminated for

in the FL-2 area versus FL-2 width plots (see Fig S1)

Fas staining for flow cytometry

Cells (1· 106) were labelled with 0.4 lg of anti-Fas DX2 for 60 min, followed by Alexa Fluor 488 mouse anti-body for 60 min Fas expression was quantified with a

FACSCalibur flow cytometer The secondary antibody

alone was used as negative control

Immunoblot analysis

For studies on protein expression of Fas, caspase-8 and

PARP, cells were lysed in SDS lysis buffer (2% w⁄ v SDS,

100 mm Tris pH 6.8) supplemented with CompleteMini

Pro-tease Inhibitor Cocktail (Roche Diagnostics, Mannheim,

Germany) For immunoblotting of the total and

phosphory-lated levels of AKT, MEK1⁄ 2 and ERK1 ⁄ 2, cells were lysed

in RIPA lysis buffer [50 mm Tris⁄ HCl (pH 7.4), 150 mm

NaCl, 0.1% SDS, 1% NP-40 and 0.5% sodium

deoxy-cholate] containing CompleteMini Protease Inhibitor

Cock-tail, 2 mm phenylmethylsulfonyl fluoride, 1 mm sodium

orthovanadate, 10 lgÆmL)1 p-nitrophenyl phosphate, 5 mm

b-glycerophosphate and 50 lgÆmL)1 Glycine max(soybean)

inhibitor These lysates were kept on ice for 15 min and

cleared by centrifugation at 20 000 g for 15 min SDS-PAGE

was performed as described previously [16] In brief, 30 lg

of protein was separated by SDS-PAGE and transferred to

nitrocellulose membranes with a semidry blot (Bio-Rad

Lab-oratories, Richmond, CA, USA) Primary antibodies were

visualized with horseradish peroxidase-conjugated secondary

antibodies and enhanced chemiluminescent substrates

(Pierce Biotechnology, Rockford IL, USA) For the

quanti-fication of arbitrary units, a fluor-s multi imager and

quantity onesoftware were used (Bio-Rad Laboratories)

RT-PCR and qRT-PCR

Total RNA was isolated with a Nucleospin RNA II Kit

(Macherey-Nagel, Du¨ren, Germany) First-strand cDNA

was synthesized from 1 lg of total RNA with Superscript II

RT (Invitrogen Life Technologies, Paisley, Renfrewshire,

UK) The following primers were used: Fas mRNA,

5¢-AGATCTAACTTGGGGTGGCT-3¢ and 5¢-ATTTATT

GCCACTGTTTCAGGAT-3¢; Fas pre-mRNA, 5¢-GGACC

CAGAATACCAAGTG-3¢ and 5¢-GTCAGTGTTACTTC

TNFR2, 5¢-AAACTCAAGCCTGCACTC-3¢ and 5¢-GGA

described previously [16]

For RT-PCR amplification, 25 ng of template was

amp-lified with HotMaster Taq polymerase (Eppendorf,

Hamburg, Germany) under the following conditions: 95C

for 15 min; 94C for 45 s; 56 C for 30 s; 72 C for 45 s

The PCRs were run in a MastercyclerNN (Eppendorf)

qRT-PCRs were performed on 25 ng of template using a

QuantiMix EASY SYG KIT (Biotools⁄ Techtum Lab AB,

Umea˚, Sweden); the conditions applied were 95C for

3 min, 95C for 10 s and 60 C for 45 s Forty cycles were

performed qRT-PCRs were run in an iCycler thermal

cycler and the data were analysed using icycler iq

software (Bio-Rad Laboratories) The expression of each target transcript was normalized to GAPDH Duplicates were made on three independent RNA isolations and the data represent the means ± SD

Cell transfection

The expression construct was prepared in the pCEP4 expression vector (Invitrogen Life Technologies) by insert-ing Fas cDNA (IMAGE: 5202648) cDNA was extracted from the pCMV-SPORT6 vector and cloned into pCEP4 at the XhoI and KpnI restriction sites Cells (2· 107) were electroporated with 20 lg of plasmid at 240 V and 950 lF using a Bio-Rad Gene Pulser Stable transfectants were selected with 300 lgÆmL)1 hygromycin B (Roche Diagnos-tics) for 3–4 weeks and subsequently analysed Diagrams represent the triplicates of two transfections, and similar data were obtained in more than three experiments

Sequencing of bisulfite-modified genomic DNA and methylation-specific PCR

Genomic DNA was purified with a DNeasy Kit (Qiagen Nordic, Solna, Sweden) DNA (1.5 lg) was denatured in 0.3 m NaOH at 42C for 30 min, and subsequently treated with 3.3 m sodium bisulfite plus 0.5 mm hydroquinone for

15 h at 55C DNA was purified in spin columns (Pro-mega, Madison, WI, USA), denatured with 0.3 m NaOH and neutralized with 3 m ammonium acetate Finally, DNA was precipitated with ethanol, washed and reconstituted in

TE buffer (10 mm Tris⁄ HCl, pH 7.6; 1 mm EDTA) For sequencing, PCR amplification from)575 to +8 of the Fas promoter was carried out with the primers described else-where [11] HotStart Taq polymerase (Qiagen Nordic) was used and the conditions applied were as follows: 95C for

15 min; 94C for 45 s; 56 C for 1 min; 72 C for 2 min; for 35 cycles The PCRs were run in a MastercyclerNN (Eppendorf) The PCR products were cloned into the pCR 4-TOPO vector with the TOPO TA Cloning Kit (Invitrogen Life Technologies) before sequencing

Methylation-specific PCR of CpGs in the first intron of Fas was performed on bisulfite-treated DNA, as described elsewhere [11]

Statistical analysis

All data represent three independent experiments if not stated otherwise For statistical analyses, Student’s t-test was applied and P < 0.05 was considered to be statistically significant

Acknowledgements

We would like to thank Professor Staffan Bohm, Umea˚ University, for provision of the cDNA of FAS