 |
INTRODUCTION |
Characteristic chromosomal translocations are commonly present in
specific types of mesenchymal tumors and frequently involve genes
encoding transcription factors (1-3). The clonal nature of these
chromosomal abnormalities, their association with specific cell types,
and their presence in somatic cells but not germ cells suggests an
important role in the development of neoplasia. Transcription factors
are typically composed of multiple domains with different functions
including DNA binding, dimerization, and transcriptional activation.
The combination of specific domains from unrelated transcription
factors may result in the generation of chimeric proteins with activity
distinct from either of its components (3).
The chromosomal translocation t(12;22)(q13;q12) associated with Clear
Cell sarcoma (CCS)1 was first
described by Bridge et al. (3, 4) and has since been shown
to be a characteristic finding of diagnostic value. The balanced
translocation gives rise to a fusion protein in which the N-terminal
325 amino acids of the Ewing's sarcoma protein (EWS) replace the
N-terminal 65 amino acids of the activating transcription factor 1 (ATF1) (3, 4). The reciprocal ATF1/EWS translocation does
not produce a fusion product in transformed cells due to the presence
of an in-frame stop codon immediately C-terminal to the ATF1 sequence.
The biologic role of EWS is not well understood. EWS functions as an
RNA-binding protein and may play a role in RNA processing (5).
Bertolotti et al. (6) suggested EWS may also function as a
transcription factor due to a high degree of homology with the
TBP-associated factor hTAFII68. EWS was shown in studies by
Pan et al. (7) to possess multiple determinants that
cooperate synergistically to activate transcription, but by itself EWS
was not capable of binding to DNA. Protein-protein interactions may
occur between transcription factors, independent of DNA binding, and
are important in transcriptional activation. EWS also participates in
fusion proteins with other DNA-binding proteins, most notably FLI (8),
a member of the ETS family, and EWS/FLI is overexpressed in Ewing's sarcoma.
ATF1 is a member of the CREB/ATF subfamily of bZIP transcription
factors that also includes CREB and CREM (9-11). These inducible transcription factors regulate transcription through binding as homodimers or heterodimers to cyclic AMP response elements (CRE) following activation of certain pathways such as protein kinase A. ATF1
is a weaker transactivator in vitro than CREB (12, 13). EWS/ATF1 is predicted to bind to CREs via the bZIP domain provided by
the C-terminal region of ATF1, but it does not retain cAMP-inducible activation due to partial deletion of the kinase-inducible domain located in the N-terminal 65 amino acids of ATF1(14).
Brown et al. (15) have shown in a heterologous cell type
that EWS/ATF1 is a strong constitutive activator of some CRE-containing promoters and a repressor of others. A plausible mechanism for transformation in Clear Cell sarcoma involves the deregulated activation of CRE-containing promoters by the fusion protein. Other
chimeric proteins, including the PAX/FKHR chimeric protein found in
rhabdomyosarcoma, are capable of transforming cells in culture, and
EWS/ATF1 may function in a similar manner to initiate tumor cell
proliferation (16). The development of cancer is believed to be a
multistep process, and downstream events may occur that render the
tumor independent of the initiating event (14). It is not known whether
EWS/ATF1 or other chimeric proteins resulting from translocations are
essential for maintenance of cell proliferation.
In studies aimed at addressing the cellular roles of ATF1, a single
chain Fv fragment (scFv) was derived from an anti-ATF1 monoclonal
antibody (mAb). The mAb, designated mAb41.4, inhibited ATF1 binding and
transcriptional activation from CRE-containing promoters in
vitro, and its epitope was mapped to the first 15 residues
N-terminal to the DNA binding domain (13, 17). The single chain
variable fragment (scFv) derived from mAb41.4 (designated scFv4)
retained in vitro activity, but unlike the mAb, it could be
expressed intracellularly to target ATF1 activity (18). The availability of the anti-ATF1 scFv4 provided a unique means to explore
the importance of DNA binding by EWS/ATF1 and evaluate its role in the
neoplastic process. In this study, we demonstrate that the C-terminal
region of EWS/ATF1 retains the mAb4 epitope and that this epitope is
accessible for binding by scFv4. In electrophoretic mobility shift
assays (EMSA), the presence of mAb4 was capable of inhibiting CRE
binding by EWS/ATF1. Intracellular expression of scFv4 reduced the
activational effect of EWS/ATF1 on CRE-containing reporters when
transfected into HeLa cells expressing EWS/ATF1 but had no effect on
cell viability. Additional studies were performed in 293T cells
transfected with EWS/ATF1 and in a cell line termed SU-CCS-1, derived
from a human CCS tumor. Levels of EWS/ATF1 in both of these cell lines
were lower than the level of EWS/ATF1 in primary tumor tissue as
determined by Western blot. Following transfection into the
EWS/ATF1-expressing SU-CCS-1 cell line, scFv4 reduced activity of a
CRE-containing reporter, and this inhibitory effect could be reversed
by overexpression of ATF1. These studies suggested that the role of
EWS/ATF1 in maintenance of the malignant phenotype could be
investigated directly if scFv4 intracellular expression could be
achieved in a high percentage of SU-CCS-1 cells. When introduced via
retroviral vector into SU-CCS-1 cells, intracellular expression of
scFv4 reduced viability to 10% compared with controls. Evidence of
tumor cell death occurring through an apoptotic mechanism was obtained
by TUNEL and flow cytometry.
| |
MATERIALS AND METHODS |
Preparation of mAbs and scFv--
Preparations of mAb4 were
affinity purified on a protein G column and quantitated by absorbance
at 280 nm (assuming 1 mg IgG/ml of IgG = 1.4 A280) with the Bradford assay (Bio-Rad). Soluble scFv4 was produced and quantitated as described previously (38). Escherichia coli HB21 cells, incubated until reaching an
A600 of 0.6, were induced with
isopropyl-D-thiogalactopyranoside and incubated an
additional 4 h at 25 °C. The periplasm was extracted in a high
salt lysate buffer, clarified, and dialyzed. scFv4 quantitation was
performed through slot blotting of the periplasmic extract and a
peptide standard. The slot blots were stained with an
anti-c-myc tag Ab (murine 9E10 hybridoma, ATCC) and an
alkaline phosphatase (AP)-conjugated anti-mouse IgG heavy and light (H
& L) chain Ab (Jackson ImmunoResearch Laboratories, West Grove, PA). A
standard curve (1-100 ng) using c-Myc peptide 1 (Oncogene Research
Products, Cambridge, MA) was generated, and the signal of scFv wells
was visually compared for determination of approximate concentration and digitally scanned for densitometric analysis. Following
normalization for mass (mass of c-Myc peptide = mass of scFv/8),
the average periplasmic concentration of scFv was observed to be 5 ng/ml.
Reverse Transcriptase-Polymerase Chain Reaction and Isolation of
EWS/ATF1 cDNA--
Total RNA mini-preps were prepared from 100-mm
dishes of SU-CCS-1 cells using Qiagen RNeasy and QuiaShredder columns
(Qiagen, Valencia, CA). The protocol for the isolation of total RNA
from animal cells was followed as supplied by the manufacturer. 50 ng
of total RNA was reverse-primed with an oligonucleotide poly(dT) primer
and extended with SuperscriptTM reverse transcriptase (Life
Technologies, Inc.) according to established protocols. The
EWS/ATF1 (15) fusion was amplified from the product of the
cDNA synthesis by polymerase chain reaction with appropriately
designed primers based on the GenBankTM ATF1 and
EWS sequences. A polymerase chain reaction product of approximately 1600 base pairs was obtained and ligated into the T/A
cloning vector (Invitrogen, Carlsbad, CA) for screening and sequencing.
Multiple colonies were screened using mini-prep spin columns (Qiagen),
and those containing the properly sized insert were submitted for
automated sequencing.
DNA Constructs--
For intracellular expression assays, the
cDNA of EWS/ATF1 was cloned into pCMV4(18). The
EcoRI-HindIII fragment from T/A-EWS/ATF1 was
inserted into the BglII-HindIII sites of pCMV4 to
generate the vector referred to as pEWS/ATF1 and used to generate
protein in 293T cells. The vectors pATF1 and pFv4 are as described
previously (18). The EWS/ATF1 cDNA was inserted into the
EcoRI site of pET29(b) (Novagen, Madison, WI) which had the
NcoI-EcoRV fragment removed. This construct,
pET-EWS/ATF1, was screened for orientation and used for the in
vitro generation of recombinant protein in E. coli BL21.
Preparation of Recombinant Proteins--
Recombinant EWS/ATF1
was generated by in vitro transcription-translation using
the TnT® T7 Quick Coupled Transcription/Translation System (Promega,
Madison, WI) as per the instructions of the manufacturer. Both
35S-labeled and unlabeled recombinant proteins were
generated for use as markers in Western blot and EMSA. Recombinant
EWS/ATF1 and ATF1 were also generated through
isopropyl-D-thiogalactopyranoside induction of ATF1
cDNA and EWS/ATF1 cDNA containing pET vectors in E. coli BL21 (19). ATF1-expressing bacteria were boiled for 20 min as
described previously (19). EWS/ATF1 was isolated as the insoluble
protein fraction of induced bacteria according to established protocols
(20). Additionally, EWS/ATF1 was generated in 293T cells following
transfection with pEWS/ATF1 and isolation of the nuclear extract using
established protocols.
Electrophoretic Mobility Shift Assays--
EMSA were performed
as described previously (13, 17). Incubations were conducted at
30 °C after determining that EWS/ATF1 forms more intense complexes
with the CRE at this temperature. 32P-Labeled
oligonucleotide containing the consensus CRE 5'-AGA GAT TGC
CTG ACG TCA GAG AGC TAG-3' was
incubated with 50 ng of full-length recombinant ATF-1 from E. coli BL21 or EWS/ATF1 from 293T cells. The binding reactions were
done in the presence or absence of mAb4, mAb5, EWS-N, and species and
isotype-matched controls. Following electrophoresis, the bound and
unbound fractions of labeled oligonucleotide were quantitated by
autoradiography for 12 h using a PhosphorImager (Molecular
Dynamics). The PhosphorImager data were exported as TIFF files and used
to prepare Fig. 1, B and C.
Immunoblot Assays--
Protein extractions from HFF and SU-CCS-1
cell lines were made using triple detergent saline (TDS) lysis buffer
(1.0% Triton X-100, 0.5% deoxycholate, and 0.1% lauryl sulfate
(SDS)). Protein extraction efficiencies were determined by examining
the relative amount of EWS/ATF1 and/or ATF1 in the insoluble cell
membrane fraction as compared with the TDS-soluble fraction. The
insoluble fraction remaining from the original TDS extraction was
re-solubilized in 1% SDS, and DNA was sheared by sonication. The
samples were boiled for 10 min and analyzed by SDS-polyacrylamide gel
electrophoresis. Immunoblots (Western) were performed as described
previously (13). Protein extraction from a Clear Cell sarcoma tumor was
performed by mechanical homogenization in the presence of TDS lysis
buffer. Protein concentrations were determined for each extract using the Bradford Assay Kit (Bio-Rad). Immunoprecipitation was performed using mAb1 and mAb5 concurrently and 20 µl of protein A-Sepharose (6 µg/µl) incubated with 150 ng of cellular or tumor extract for 150 min at 4 °C. Efficiency of immunoprecipitation was determined by
comparison of pre- and post-immunoprecipitation and supernatant fractions by SDS-polyacrylamide gel electrophoresis and transfer to
nitrocellulose. Membranes were incubated with either 1 µg/ml mAb5
followed by an alkaline phosphatase (AP)-conjugated goat anti-mouse
heavy and light (H & L) chain secondary antibody (Jackson ImmunoResearch) or EWS-N (Santa Cruz Biotechnology) followed by an
AP-conjugated mouse anti-goat antibody (Santa Cruz Biotechnology). The
stained Western blots were digitally scanned using a UMAX Astra 610s
scanner to generate transfer image file format (TIFF) images that were
imported into Canvas version 5.0.3 and used to prepare Fig.
1D. In vitro 35S-labeled EWS/ATF1
analyzed by autoradiography migrated identically to the presumed
EWS/ATF1 band generated by Western blot, thus confirming the identity
of the EWS/ATF1 band. Analysis of band intensity was performed on the
stained blots using a densitometer (Molecular Dynamics).
Transient Cotransfections and Luciferase/
-Galactosidase
Assays--
Transient cotransfections of HeLa cells were performed
according to established protocols using calcium phosphate
precipitation (18). The transfections were performed in duplicate 35-mm
wells containing 5 µg of the CMV-luciferase reporter construct and an Rous sarcoma virus--galactosidase construct (2 µg) to control for
variations in transfection efficiency. Cotransfections included increasing amounts of the EWS/ATF1 vector at 0, 5, 10, and 20 µg and
the presence of plasmids pFv4 and pATF1. Additionally, a molar
equivalent of parent vector (without cDNA insert) was used to
maintain an equal number of promoter units in each transfection. The
cells were harvested at 48 h post-transfection, and the reporters were assayed. Transient cotransfections of SU-CCS-1 cells were performed using a similar approach of increasing amounts of pFv4. To
facilitate efficient transfection of SU-CCS-1 cells, liposome-mediated transfection was used with the LipofectAMINE PLUS system (Life Technologies, Inc.), and cells were harvested at 72 h. Measurement of reporter activity of firefly luciferase was determined as described relative to an internal -galactosidase standard. Following
transfection, cell extracts were prepared by freeze-thaw lysis in a
potassium phosphate buffer. ATP and luciferin were added, and light
emission was measured with a Luminoskan RS (Lab Systems/Denley,
Franklin, MA) microplate luminometer. -Galactosidase expression was
quantitated through the addition of
o-nitrophenyl--D-galactopyranoside, and the
absorbance at 405 nm was measured on an enzyme-linked immunosorbent assay plate reader. The luciferase value of each well was normalized to
the internal -galactosidase reporter. Results of three to five
experiments were then averaged to generate the data depicted in Figs. 2
and 3.
Production of Retrovirus and Infection of Cells--
The
retroviral vectors were produced by inserting the
EcoRI-HindIII fragment of pFv4 that contains the
cDNA of scFv4 into the SR
-PN retrovirus at the corresponding
sites. SR-PN was kindly provided by Dr. T. Smithgall and is the
SRMStkneo vector described by the laboratory of Dr. O. Witte
(21-23) with the pCMV5 polylinker inserted into the HindIII
site. To infect SU-CCS-1 cells, the SR-Fv DNA construct was
cotransfected into 293T cells with the amphotrophic packaging vector
psi(
) ampho. 10 µg of each was transfected using the LipofectAMINE
system. The cellular supernatant was collected every 12 h between
24 and 72 h post-infection and pooled. The retroviral titer was
determined by colony-forming assay in 3Y1 cells grown in minimum
Eagle's medium containing 5% bovine calf serum and 800 mM
G418 (Geneticin). Typical yields of retrovirus were 104
cfu/ml. Infection of cells was performed using 3 ml of retroviral stock/well in a 6-well plate in the presence of 4 mg/ml hexadimethrine bromide (Polybrene). Plates were spun at 1250 × g in a
refrigerated centrifuge at 18 °C.
Cell Viability Determinations, Trypan Blue Exclusion, and MTS
Assays--
The viability of SU-CCS-1 cells infected by SR-Fv4 or
control SR-PN was determined by trypan blue stain exclusion. Cells were harvested from 35-mm dishes with a rubber policeman, suspended in
minimum Eagle's medium, and transferred to centrifuge tubes. The cells
were washed in PBS and resuspended in 1 ml of PBS. An equal amount of
cell suspension was added to 2× trypan blue stain, and the cells were
counted in a hemocytometer. Grids were counted to quantitate blue cells
and white cells until a minimum of 400 was obtained. In order to avoid
the mechanical harvesting which could interfere with viability
measurements, an MTS assay was performed using the CellTiter 96 Aqueous
non-radioactive proliferation assay (Promega) which is a colorimetric
method for determining the number of viable cells in proliferation
assays. The assay is composed of the tetrazolium compound 3-(4,
5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) and the electron coupling reagent phenazine methosulfate. MTS is
bioreduced by cells into a formazan that is soluble in tissue culture
medium. The absorbance of the formazan at 490 nm can be measured
directly from 96-well assay plates without additional processing. The
conversion of MTS into the aqueous soluble formazan is accomplished by
dehydrogenase enzymes found in metabolically active cells. Therefore,
the quantity of formazan is directly proportional to the number of
living cells in culture. For this assay, SU-CCS-1 cells were plated at
2 × 104 cells/well and infected with SR-Fv4 or
controls (0.2 ml/well). Cell infections were conducted over 7 days to
generate a time course of viability. On day 7, 96-well plates were
incubated with the MTS assay reagents, and the absorbance was measured.
The results of 3-6 experiments were normalized and plotted as percent
viable cells versus time. The same MTS procedure was used to
study the effect of SR-Fv4 and control treatments on HeLa cell
viability over a 4-day time course.
Apoptosis Measurements, Flow Cytometry, and TUNEL
Staining--
50 µl of the washed cell suspensions described above
were plated on glass slides, air-dried, and fixed in 50% acetone, 50% methanol. The remaining cell suspension was pelleted and fixed in 70%
ethanol. The ethanol-fixed cells were prepared for DNA content analysis
and apoptosis measurement by flow cytometry by washing in PBS and
staining with propydium iodide (Telford reagent) overnight (24).
Measurements were made using a Becton Dickinson FACStarPLUS
flow cytometer, and the data set was analyzed using ModFit DNA modeling
software (Versity Software, Topsham, ME). The slides of fixed cells
were stained for apoptosis by in situ labeling of DNA breaks
using terminal deoxynucleotidyltransferase (TdT)-mediated dUTP-biotin
nick end labeling (TUNEL) (25). TdT was used to incorporate
biotinylated deoxyuridine at sites of DNA breaks, and the signal was
amplified by avidin-peroxidase and photographed under light microscopy.
| |
RESULTS |
Anti-ATF1 mAbs Recognize the EWS/ATF1 Fusion Oncoprotein and
Inhibit Its Binding to a CRE--
EWS/ATF1 incorporates the C-terminal
region of ATF1 containing the epitopes of the two anti-ATF1 mAbs used
in these studies (Fig. 1A).
Although both mAb4 and mAb5 recognize epitopes adjacent to the DNA
binding domain of ATF1, mAb4 interferes with DNA binding by ATF1 in
EMSA, and mAb5 super-shifts ATF1 without disrupting its DNA binding
activity (18). The contribution of EWS to the overall conformation of
the chimeric protein is unknown, and it was important to determine
whether the addition of the EWS domain would block the epitope of mAb4
and mAb5. EWS/ATF1 and ATF1 were used in gel shift assays with
radiolabeled CRE DNA to evaluate the ability of mAb4 and mAb5 to bind
EWS/ATF and determine the effect of mAb4 and mAb5 on complex formation.
EWS/ATF1 binding to CRE DNA has been previously demonstrated (14, 15,
26); however, it is not known whether CRE sequences are the primary target in cells or whether other related DNA sequences are capable of
being bound (13, 17). For these studies, a consensus CRE (TGACGTCA) as
occurs in the somatostatin promoter was utilized. EWS/ATF1 was
expressed in 293T cells rather than bacteria to control for possible
effects of post-translational modification (13, 17). The presence of
mAb4 inhibited EWS/ATF1 complex formation (lane 5, Fig.
1C), whereas mAb5 supershifted the EWS/ATF1 complex (lane 7). This effect on complex formation was similar to
that of mAb4 on ATF1·CRE complexes (compare lanes 2 and
5, Fig. 1B) and the supershift of ATF1·CRE
complexes by mAb5 (lane 7). The EWS-N Ab (Santa Cruz
Biotechnology), which recognizes the N-terminal region of EWS, was used
to verify the identity of the EWS/ATF1 complex and this antibody was
capable of producing a partial supershift (lane 6, Fig.
1C). As expected, EWS-N had no effect on ATF1·CRE complexes (lane 6, Fig. 1B). Specificity of
EWS/ATF1 for the CRE was demonstrated with the addition of 100-fold
excess of unlabeled AP1 and CRE competitors. Competition with unlabeled
CRE resulted in a loss of ATF1 complexes, whereas competition using AP1
did not diminish the intensity of the complex (lanes 3 and
4, Fig. 1, B and C). AP1 is useful as
a control for specificity since it differs from a consensus CRE by only
one G-C base pair at its center. Isotype-matched control Abs had no
effect on complex formation (data not shown). These studies indicated
that although the EWS domain is considerably larger than the deleted
amino portion of ATF1, it did not interfere with binding of specific
epitopes by either mAb4 or mAb5.
View larger version (45K):
[in this window]
[in a new window]
|
Fig. 1.
A, schematic representation of ATF1
(above), EWS (middle), and EWS/ATF1
(below). Structural and functional domains are shown
(NTR, N-terminal region; , -helical region; ,
-sheet region, PKC and PKA, protein kinase C
and A phosphorylation sites; NTD, N-terminal domain;
RGG, RGG RNA binding box; basic, basic DNA
binding region; and zip, leucine zipper dimerization
domain). The epitopes of antibodies are identified, and the breakpoint
fusing the EWS N-terminal region to the ATF1 C-terminal region is
indicated. B, EMSA comparing effects of mAb4, mAb5, and
EWS-N on complex formation with recombinant ATF1 (rATF1).
Shifted complexes are indicated by arrows. Each reaction
mixture contained 50 ng of rATF1, 2 µg of antibody, 4.0% glycerol,
and 0.1% gelatin. Additional controls included isotype-matched mAbs
(data not shown). C, EMSA comparing effects of mAb4, mAb5,
and EWS-N on 293T-expressed EWS/ATF1. Shifted complexes are indicated
by arrowheads. Each reaction mixture contained 5 µg of
293T-EWS/ATF1, 2 µg of antibody, 4.0% glycerol, and 0.1% gelatin
and was incubated at 30 °C. Additional controls included isotype
matched mAbs (data not shown). D, immunoblot comparing
intracellular levels of ATF1 and EWS/ATF1. Protein extracts were
prepared from HFF cells, pEWS/ATF1 transfected 293T cells, SU-CCS-1
cells, and a primary human CCS tumor immunoprecipitated with mAb1 and
mAb5. ATF1 (35 kDa) and EWS/ATF1 (80 kDa) were detected with the
anti-ATF1 antibody (mAb5) and an AP-conjugated secondary antibody, and
the protein bands were analyzed by densitometry. Recombinant ATF1 and
293T-EWS/ATF1 were used as markers. HFF cell extracts served as a
non-transformed control cell line.
|
|
The EWS/ATF1 Fusion Protein Is Expressed at Greater Levels Than
Endogenous ATF1--
The EWS/ATF1 fusion protein is hypothesized to be
the primary genetic event leading to CCS; however, the level of
EWS/ATF1 expression in primary tumor tissue has not been demonstrated
previously. Extracts from SU-CCS-1 cells, a primary CCS tumor, and a
primary human fibroblast cell termed HHF were immunoprecipitated and
analyzed by Western blotting (Fig. 1D). Efficiencies of
protein extraction and immunoprecipitation were both shown to be
greater than 95% (data not shown). HHF cells were utilized to
represent non-transformed control cells of mesenchymal origin.
Recombinant ATF1 expressed in E. coli BL21 and EWS/ATF1
expressed in 293T cells were used as markers for the proteins of
interest. The EWS-N Ab (Santa Cruz Biotechnology) which recognizes the
N-terminal region of EWS was again used to confirm identity of the
presumed EWS/ATF1 band. Due to the t(12;22) translocation, only one
normal ATF1 allele remains in SU-CCS-1 cells and the CCS tumor (4).
However, levels of ATF1 were similar to those of nontransformed HFF
fibroblasts with two alleles. The EWS/ATF1 band was considerably darker
in comparison with the endogenous ATF1 band in the SU-CCS-1 cell line
and the CCS tumor (lane 3 and 4, Fig.
1D). Densitometric analysis indicated that EWS/ATF1 levels
were 3.0-fold greater than those of ATF1 in the SU-CCS-1 cell extract
and 10.6-fold greater than ATF1 in the CCS tumor extract. As expected,
EWS/ATF1 was not present in the control HHF cell extract.
scFv4 Inhibits EWS/ATF1 Activation of a CRE Reporter in HeLa
Cells--
In our past studies, inhibition of specific complex
formation in vitro by mAb4 was predictive of decreased
reporter expression in transfected cells (18). Since EWS/ATF1 binding
to a CRE was inhibited in vitro by mAb4, a similar effect on
transactivation was expected in cells following transfection of scFv4.
HeLa cells were chosen for their relatively higher level of ATF1
versus CREB expression and their well documented history of
CRE reporter activation (18). Transient cotransfection assays of HeLa
cells were performed using a CRE luciferase reporter and constructs
expressing scFv4 (pFv4) and EWS/ATF1 (pEWS/ATF1). The reporter
construct incorporated the strong CMV immediate early gene promoter
which contains five CRE sequences. To normalize results for variation
in transfection efficiency between experiments, an internal Rous
sarcoma virus--galactosidase control was included in the
transfection system.
The number of promoter elements present in each transfection was held
constant by the addition of equimolar amounts of parental vectors.
Transfection of 5 µg of pEWS/ATF1 per 106 HeLa cells
produced a 3.3-fold increase in CRE-luciferase expression and use of 10 µg pEWS/ATF1 per 106 cells produced a 6.5-fold increase
(Fig. 2). Cotransfection of pFv4 (10 µg
per 106 cells) into this system reduced the observed
6.5-fold increase in reporter expression to less than 3-fold, thus
suggesting that scFv4 was capable of inhibiting CRE activation by
EWS/ATF1 in HeLa cells. The levels of CRE reporter expression in
response to EWS/ATF1 were similar to those previously described (14, 15, 26). Expression of EWS/ATF following transfection was confirmed
using immunofluorescently labeled antibodies (data not shown).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 2.
Effect of scFv4 expression on CRE and
EWS/ATF1-regulated transcription in HeLa cells. Transient
co-transfection experiments in HeLa cells were performed using 5-10
µg of pEWS/ATF1 and a CRE-luciferase (CRE-luc) reporter.
Relative light units (RLU) are depicted on the y
axis with results of reporter alone set to 1.0 for comparison.
Error bars representing the average of at least three
experiments are indicated. A constant number of CRE promoter elements
was maintained in all experiments. Transfection efficiency was
controlled by co-transfection of a -galactosidase reporter.
|
|
scFv4 Inhibits the Activity of Endogenous EWS/ATF1 in SU-CCS-1
Cells--
The SU-CCS-1 cell line was derived from a CCS tumor that
expresses endogenous EWS/ATF1 (27) and optimal transfection conditions were unknown. Therefore, a green fluorescent protein (GFP)-expressing construct was used to determine the optimal transfection method and
time course to be used. A higher level of transfection efficiency was
achieved using the liposome-mediated system than with calcium phosphate. Expression of CRE-luciferase reporter measured over a
24-96-h time course demonstrated the peak level occurred at 72 h.
Therefore, to evaluate the effect of scFv4 on endogenous EWS/ATF1
activity, transient transfections of SU-CCS-1 cells were performed
using the liposome-mediated method, and luciferase activity was
measured at 72 h with CRE-luciferase reporter and increasing amounts (2.5-10 µg per 106 cells) of pFv4. Luciferase
reporter activity decreased proportionately as increasing amounts of
pFv4 were transfected into the SU-CCS-1 cells. Activity was reduced by
80% when 10 µg of pFv4 per 106 cells was used (Fig.
3). Previously, we have observed that 10 µg of pFv4 per 106 cells decreased reporter activity by
only 20% in the non-EWS/ATF1-expressing HeLa cell line (18).
Therefore, the significantly greater decrease in reporter activity in
SU-CCS-1 cells was likely to be due to the inhibition of the strong
EWS/ATF1 activator by scFv4 and not inhibition of endogenous ATF1
activity. However, since the decrease in CRE reporter activity was
reversed by overexpression of ATF1, either possibility remained. 1 µg
of pATF1 cotransfected with 2.5 µg of pFv4 per 106
SU-CCS-1 cells restored luciferase expression to near base-line levels
(Fig. 3), indicating that ATF1 competed for scFv4 binding and allowed
free EWS/ATF1 or endogenous factors to activate the CRE reporter. In
HeLa cells, the small effect on reporter activity may be due to the
presence of other strong activating proteins that regulate expression
as well as regulatory elements other than CRE. Although in
vitro assays may not accurately reflect all aspects important to
transcriptional regulation, the level of inhibition by scFv4 was
predictive of results when cell viability was determined.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 3.
Effect of scFv4 on endogenous
EWS/ATF1-regulated CRE-reporter activity in SU-CCS-1 cells.
Transient co-transfection experiments in SU-CCS-1 cells were performed
using increasing amounts (0-10 µg) of pFv4 and a CRE-luciferase
(CRE-luc) reporter. Results from the addition of 10 µg of
pFv4 are shown in lane 5. Relative light units
(RLU) are depicted on the y axis with results of
reporter alone set to 1.0 for comparison. Error bars
representing the average of at least three experiments are indicated. A
constant number of CRE promoter elements was maintained in all
experiments. Transfection efficiency was controlled by co-transfection
of a -galactosidase reporter.
|
|
Long Term Expression of scFv4 in SU-CCS-1 Cells Leads to Loss of
Cell viability and Death--
Our next goal was to deliver the scFv4
to a majority of SU-CCS-1 cells to determine whether the inhibition of
EWS/ATF1 activity would affect cell viability. Earlier work with the
GFP constructs demonstrated that less than 10% of the SU-CCS-1 cells
were transfected by the liposome-mediated system. The ability of a
Moloney sarcoma retrovirus system (SRMStkneo) to transduce the
SU-CCS-1 cells was examined (21-23). An SR retrovirus capable of
expressing GFP demonstrated a transduction efficiency of 80% or
greater (data not shown). Therefore, to attain widespread delivery of
scFv4 to the SU-CCS-1 cells, the SR retroviral system was utilized and modified to express scFv4. The cDNA of scFv4 was placed into the SR-PN construct and used to produce infectious amphotropic retrovirus. SU-CCS-1 cells were transduced with 104 cfu of
either SR retrovirus expressing scFv4 (SR-Fv4), the parental
SR-PN with no insert, or a mock media preparation that simulated the
infection conditions (control). The SU-CCS-1 cells were visually
inspected daily following treatment. Control cells showed no decrease
in density, grew to confluence, and showed no reduction in viability
(Fig. 4A). Cells exposed to
SR-Fv4 demonstrated membrane blebbing and cell nuclear condensation
beginning at day 3, and these changes subsequently became apparent
throughout the population. By day 5, the cytotoxic effects reached
maximum, and cell density began to decrease substantially. At day 7, viable cells were sparse and examination under × 100 magnification showed considerable cellular debris (Fig. 4C).
The experiments were repeated on three occasions with similar results.
Conversely, less than 1% of cells exposed to SR-PN demonstrated
focal cytotoxic effects apparent at day 5, characterized by a reduction
in cell size and focal membrane blebbing. However, the remaining cells
continued to grow and proliferate to day 10 with no progressive loss in viability (Fig. 4B).
View larger version (101K):
[in this window]
[in a new window]
|
Fig. 4.
Photomicrographs of SU-CCS-1 cells 10 days
after exposure to mock (A),
SR-PN (B), and
SR-Fv4 retroviral preparations
(C).
|
|
To correlate the physical appearance of cells with an objective
measurement, the percentage of viable cells was determined by two
different methods as follows: trypan blue dye exclusion and the MTS
assay (CellTiter AQueousTM, Promega). The cells transduced
with SR-Fv4 showed a pronounced decrease in viability as measured by
trypan blue dye exclusion, beginning at day 2 which became prominent by
day 5 with only one-third of the cells remaining viable. Corresponding
to our visual observations, only 10% of the SU-CCS-1 cells remained
viable as determined by dye exclusion at day 10 (Fig.
5A). Control SR-PN infected
cells and mock-transduced cells had similar percentages of viable cells throughout the course of study. Since the levels of viability in the
control cells was 60% rather than the expected 90-100%, we
investigated the effect of cell-harvesting procedures on overall viability when measured by the dye exclusion method. The impact of
harvesting cells by scraping was examined by comparison of results with
the MTS assay that requires minimal cell manipulation. The viability of
SR-Fv4-transduced cells declined to 60% on day 3 and continued to
decrease to 30% at day 7 as determined by MTS assay. In comparison,
both the mock-transduced cells and those transduced by SR-PN
demonstrated similar results with the percentage of viable cells
starting at 100% and decreasing to 60% at day 7 (Fig. 5B).
Since the results by both trypan blue exclusion and MTS assay were
similar and corresponded to the visual appearance, we concluded that
the expression of scFv4 had a significant effect on SU-CCS-1 cell
viability, and based on the morphologic appearance, we postulated that
cell death may be occurring through a process of apoptosis.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 5.
Determination of cell viability.
A, trypan blue dye exclusion. Mock-, SR-PN-, and
SR-Fv4-treated cells were collected over a 10-day period, processed,
and stained for viability. Cells were counted in a hemocytometer until
reaching a minimum of 400. Data points representing results
from at least three experiments were plotted as % viability on the
y axis. Error bars are indicated. B,
MTS assay for SU-CCS-1 cell viability. Mock-, SR-PN-, and
SR-Fv4-treated SU-CCS-1 cells were collected over a 7-day period,
assayed by the MTS method, and cell survival was calculated. Data
points representing results from at least three experiments were
plotted as % cells surviving on the y axis. Error bars are
indicated. C, MTS assay for HeLa cell viability. Mock-,
SR-PN-, and SR-Fv4-treated HeLa cells were collected over a 4-day
period, assayed by the MTS method, and cell survival was calculated.
Data points representing results from at least three experiments were
plotted as % cells surviving on the y axis. Error
bars are indicated.
|
|
The Intracellular Expression of scFv4 Induces Apoptosis in SU-CCS-1
Cells--
The process of SU-CCS-1 cell death could occur through
either necrosis or apoptosis or a combination of both mechanisms (28, 29). The visual observations described above suggested that apoptosis
was occurring in the SR-Fv4-infected cells. In order to confirm
these observations, aliquots of SU-CCS-1 cells from the same time
course as the viability study were first submitted for DNA content
analysis by flow cytometry (Fig.
6A). Differences between
controls and SR-Fv4-infected cells were apparent at day 3 and
continued to increase throughout the remainder of the 10-day time
course. Transfection by SR-Fv4 resulted in 25% apoptosis at days
5-7 which increased to 33% on day 10. At similar time points of days
5 and 10, 15% (p < 0.05) and 18% (p < 0.00005) of the mock-transduced cells were apoptotic, respectively,
and 10% (p < 0.005) and 22% (p < 0.0005) of the SR-PN-transduced cells were apoptotic, respectively
(Fig. 6). Although values for the measurements of apoptosis induced by
SR-Fv4 made by flow cytometry are significantly different, the
processes of harvesting, centrifugation, washing, and staining could
contribute to cell damage and death. Therefore, to minimize the effect
of processing on apoptosis, cells were also fixed to slides and
analyzed by TUNEL (25). SR-Fv4, SR-PN, and control cells were
analyzed at days 1, 3, 5, and 10. A progressive increase in both the
number and intensity of TUNEL-positive SU-CCS-1 cells following
transduction by SR-Fv4 was apparent beginning at day 3 and became
extensive between day 5 and day 10 (Fig. 6B). At day 10, 30% of cells were TUNEL-positive. No intensely dark-staining nuclei
were observed in the control preparations at day 1.
View larger version (61K):
[in this window]
[in a new window]
|
Fig. 6.
Measurement of apoptosis in SU-CCS-1
cells. A, flow cytometric analysis of scFv4-treated
cells. SU-CCS-1 cells treated with either mock, SR-PN, or SR-Fv4
were collected over a 10-day span, stained with Telford reagent, and
analyzed by flow cytometry. B, TUNEL staining for apoptotic
cells. SU-CCS-1 cells treated with either mock-, SR-PN-, or
SR-Fv4 were collected over a 10-day span and stained for apoptosis
by in situ labeling of DNA breaks using terminal
deoxynucleotidyltransferase (TdT)-mediated dUTP-biotin nick end
labeling (TUNEL). Black-staining nuclei are indicative of
apoptosis.
|
|
Expression of scFv4 in HeLa Cells Does Not Alter Cell
Viability--
Since the intracellular expression of scFv4 could
potentially induce cell death due to cross-reactivity with ATF1 or
CREB, retroviral transduction experiments were performed in HeLa cells in which ATF1 and CREB are readily detectable. HeLa cells were transduced with 104 cfu of SR-Fv4, SR-PN, or a mock
media (control) preparation and assayed by the MTS method. Although
transient effects were again seen at day 1, no significant differences
in cell viability were observed between the scFv4 and control treated
cells (Fig. 5C). The absence of any reduction in the
percentage of cells remaining viable indicated that scFv4 is not toxic
to HeLa cells and support the conclusion that apoptosis in SU-CCS-1
cells was due to specific targeting of the EWS/ATF1 fusion protein
rather than the inhibition of other transcription factors.
| |
DISCUSSION |
Malignant transformation is believed to be a multistep process,
and chromosomal translocations that generate chimeric proteins such as
EWS/ATF1 may initiate a cascade of events leading to cancer (3, 29,
34). The exquisite specificity of antibodies for defined targets
presents numerous opportunities for disrupting protein-protein or
protein-DNA interactions, particularly when the targeted structures are
complex and not amenable to blockade by small molecules. Recently,
scFvs have been used to achieve phenotypic knockout of cell surface or
cytoplasmic target proteins involved in neoplasia such as Ki-Ras,
ErbB2, epidermal growth factor receptor, and the interleukin-2 receptor
(20, 31-33). We considered whether a similar approach could be used to
disrupt activity of a nuclear protein and demonstrate its role in the neoplastic process. In SU-CCS-1 cells, interference with the activity of EWS/ATF1 could theoretically eliminate the initiating process leading to neoplasia and yet have no effect on tumor growth since other
pathways may become dominant following transformation. Interference with DNA binding and transcriptional activity by the ATF1-inhibitory scFv demonstrated EWS/ATF1 is important for maintenance of tumor cell
viability in addition to its previously proposed role in initiating the
neoplastic process (38). Although DNA binding was blocked, the EWS/ATF1
protein remained available for interactions with other proteins of the
transcriptional apparatus (16).
Two plausible explanations for the effect of scFv4 on EWS/ATF1 have
been considered as follows: either scFv4 inhibits DNA binding of
EWS/ATF1 in the nucleus of cells or scFv4 binds to EWS/ATF1 in the
cytoplasm or nucleus leading to premature degradation or
immunodepletion. We believe the former mechanism is most likely as the
inhibitory effect has been studied extensively in vitro, and
both mAb4 and scFv4 are capable of inhibiting complex formation whether
added before or after ATF1 is complexed with CRE DNA (13, 18). As such,
mAb4 could inhibit EWS/ATF1 binding through either steric or allosteric
mechanisms. The predicted interactions between CRE DNA and ATF1 are
based on the structural studies of GCN4 bound to CRE DNA and AP-1 DNA
(35, 36). A conformational change in a critical domain of EWS/ATF1 may
occur following binding by scFv4, or presence of the antibody may
destabilize the important amino acid side chain interactions with the
phosphate-DNA backbone. When EWS/ATF1 is not bound to DNA, the antibody
may prevent binding of transcription factor to DNA by occupying a
region adjacent to the DNA binding domain. Although the binding
kinetics of EWS/ATF1 are not known, scFv4 has been shown to disrupt
ATF1·DNA complexes, and the presence of scFv4 in the region between
the -helices may also prevent rebinding of the factor to DNA. If
immunodepletion is the mechanism, then the inhibitory effect of scFv4
on EWS/ATF1 may be due to the removal of transcription factor from the
cellular pool by altering its intracellular processing or nuclear
transport. Additional studies are in the process of examining the
complex issue of intracellular processing of scFv4 and/or EWS/ATF1 at the single cell level.
Fujimura et al. (26) have proposed that EWS is a negative
regulator of ATF1 binding activity based on relatively lower intensity of recombinant protein complexes in gel shift assays and results from
deletion mutant experiments. We also noted a significant difference in
the relative binding affinity of recombinant EWS/ATF1 to the CRE as
compared with recombinant ATF1 when measured by band intensity on EMSA.
However, the intensity of EWS/ATF1·CRE complexes using cellular
extracts from either 293T or SU-CCS-1 cells was roughly equivalent to
that seen with recombinant ATF1 (data not shown). Therefore
post-translational modification of EWS/ATF1 may be important for
regulating binding activity as has been shown for EWS/FLI (37). In
direct comparison with ATF1, EWS/ATF1 greatly increases gene expression
when measured by reporter assay (26, 38). The increased activity of
EWS/ATF1 is thought to result from either the loss of regulatory
elements by truncation of ATF1 or the contribution of the potent EWS
transactivation domain (14). A quantitative comparison of EWS/ATF1 to
other intracellular proteins in human tumors has not been previously demonstrated. Since the chimeric protein is not produced in the absence
of the translocation between chromosomes 12 and 22, expression levels
must be compared with other endogenous protein. As determined by
cytogenetic analysis, a single allele of the wild type EWS and ATF1 genes remains intact in SU-CCS-1 cells. Our Western
blot experiments indicate that EWS/ATF1 is present in considerable excess to the endogenous levels of ATF1 in the SU-CCS-1 cell line and a
CCS tumor (Fig. 1B). Densitometric analysis indicated that EWS/ATF1 is expressed at a 3.0-fold greater level than ATF1 in the
SU-CCS-1 cell line and a 10.6-fold greater level in a CCS tumor. As
originally suggested for Ewing's sarcoma, the EWS/ATF1 fusion protein
may achieve transformation through both overexpression and strong
transcriptional activation capability (39). Similar explanations have
been proposed for alveolar rhabdomyosarcoma associated with
translocations of the Pax3 and FKHR protein genes (40).
EWS/FLI, EWS/ATF1, and other chimeric proteins resulting from specific
translocations in leukemias, lymphomas, and sarcomas can be considered
true tumor-specific proteins, and the joining region, in particular,
could serve as a unique epitope for derivation of antibodies. However,
molecular modeling of the EWS/ATF1 chimeric protein suggested that the
fusion junction was not an exposed surface and unlikely to be available
for binding by antibody. As shown here with mAb5, binding of
transcription factors by antibody does not necessarily result in loss
of function in vitro. We previously determined that
intracellular expression of scFv4 reduced activity of the
CRE-containing proliferating cell nuclear antigen promoter by
approximately 60%, but no loss of cell viability was seen when compared with controls (18). Although the proposed mechanism of scFv4
action is through prevention of CRE binding and transcriptional activation by EWS/ATF1, cross-reactivity with ATF1 or even CREB may
occur. These transcription factors have been shown to be intimately involved with various proliferative and differentiation processes in
many cell types, but their role in the SU-CCS-1 cell line is unknown.
As demonstrated by Bar-Eli and associates (41, 42), inhibition of CREB
by a dominant negative mutant leads to loss of viability in melanoma
cells. Since a similar phenomenon could be occurring in the SU-CCS-1
cells, the HeLa cell transfections were important in verifying that
scFv4 expression was not cytotoxic in cells without EWS/ATF1 (Fig.
5C). No loss in viability was observed in transfected HeLa
cells, and we concluded that scFv4 induced cell death in SU-CCS-1 cells
by disruption of EWS/ATF1 activity and not through inhibition of
endogenous ATF1 activity.
The process of cell death in SU-CCS-1 cells exposed to scFv4 appears to
have occurred through an apoptotic mechanism (25). However, cell death
involves multiple pathways, and ultrastructural studies may be helpful
in determining whether evidence of necrosis is present (29). The
finding that 30% of cells exposed to SR-Fv4 were apoptotic as
compared with controls (p < 0.005) is comparable to
results observed by others (43, 44) in studies of apoptosis.
Disruption of key molecular processes responsible for neoplastic
transformation and reversal of malignant phenotypes are important goals
in developing new cancer therapeutics (45). The targeted disruption of
EWS/ATF1 activity via the ATF1 epitope of scFv4 reduced SU-CCS-1 cell
viability but had little effect on HeLa cells not expressing the
oncogenic fusion protein. By demonstrating activity in this tumor cell
type, we demonstrate the importance of chimeric proteins with
transcriptional activity in maintenance of tumor cell viability. These
findings may have broad applications to leukemias, lymphomas, and other
sarcomas with characteristic chromosomal translocations involving
transcription factors such as the EWS/FLI-1 in Ewing's sarcoma and
Pax3/FKHR in alveolar rhabdomyosarcoma. However, as demonstrated here,
the level of the oncogenic EWS/ATF1 protein is higher in primary tumors
than in established cell lines, and in vivo studies are
needed to determine the therapeutic potential for disruption of
fusion protein transcriptional activity by antibodies.
,
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed: Dept. of Pathology
and Microbiology, 986495 Nebraska Medical Center, Omaha, NE 68198. Tel.: 402-559-4116; Fax: 402-559-4077; E-mail:
shinric@mail.unmc.edu.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
