Bcr-Abl induction Hsp70 - The Journal of Biological Chemistry

48 Pages·2004·1.77 MB·English

JBC Papers in Press. Published on May 21, 2004 as Manuscript M401851200
BCR-ABL transactivation of Hsp70

Genomic mechanisms of p210BCR-ABL signaling: induction of Heat
shock protein 70 through the GATA response element confers
resistance to paclitaxel-induced apoptosis

Sutapa Ray1, Ying Lu1, Scott H. Kaufmann2, W. Clay Gustafson3 , Judith E. Karp4, Istvan D
Boldogh5 , Alan P. Fields6 and Allan R. Brasier1,3. ow
n
lo
a
d
FGraolmve sthtoen 1,D TeXpa 7r7tm55e5n-ts1 0o6f 0I;n t 2eDrniavli sMioend oicfi nOen, cUonloivgeyr Rsietys eoafr cThe,x Masa Myoe Cdilcinali cB, rRaoncchhe (sUteTr,M MBN), ed from
h
55905; 3Sealy Center for Cancer Cell Biology, UTMB, Galveston, TX, 77555-1048; 4Sidney ttp
Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD 21231; the ://w
w
5Department of Microbiology and Immunology, UTMB, Galveston, TX, 77555-1019; and the w
.jb
6Mayo Clinic Comprehensive Cancer Center, Jacksonville, FLA 32224. c.o
rg
b/
y
Running title: BCR-ABL transactivation of Hsp70 gu
e
s
t o
n
A
p
ril 13
, 2
0
Keywords: Chronic myelogenous leukemia/BCR-ABL/ heat shock protein/ GATA/apoptosis 19

3To whom requests for reprints should be addressed, at:
Division of Endocrinology
MRB 8.138
The University of Texas Medical Branch
301 University Blvd.
Galveston, TX 77555-1060
Ph: [phone]
FAX: [phone]
Email: [email protected]

1
Copyright 2004 by The American Society for Biochemistry and Molecular Biology, Inc.
BCR-ABL transactivation of Hsp70
SUMMARY

Chronic myelogenous leukemia (CML) results from a t(9,22) translocation, producing the
p210BCR-ABL oncoprotein, a tyrosine kinase that causes transformation and chemotherapy
resistance. To further understand mechanisms mediating chemotherapy resistance, we identified
556 differentially regulated genes in HL-60 cells stably expressing p210BCR-ABL vs those
expressing an empty vector using cDNA macro- and oligonucleotide microarrays. These BCR-
ABL regulated gene products play diverse roles in cellular function including apoptosis, cell
cycle regulation, intracellular signaling, transcription, and cellular adhesion. In particular, we
identified upregulation of the inducible form of Heat shock protein 70 ( Hsp70), and further
explored the mechanism for its upregulation. In HL-60/BCR-ABL and K562 cells (expressing
D
o
p210BCR-ABL), abundant cytoplasmic Hsp70 expression was detected by immunoblot analysis. wn
lo
a
d
Moreover, cells isolated from bone marrow aspirates of patients in different stages of CML ed
fro
(chronic, aggressive and blast crisis) express Hsp70. Expression of p210BCR-ABL in BCR-ABL m
h
ttp
negative cells induced transcription of the proximal Hsp70 promoter. Mutational analysis ://w
w
mapped the major p210BCR-ABL responsive element to a high affinity 5’(A/T)GATA(A/G)-3’ w.jb
c
.o
“GATA” response element (GATA-RE) that binds GATA-1 in CML cells. The GATA-RE was rg
b/
y
sufficient to confer p210BCR-ABL and p185BCR-ABL mediated trans-activation to an inert promoter. gu
e
s
t o
siRNA mediated “knockdown” of Hsp70 expression in K562 cells induced marked sensitivity to n
A
p
paclitaxel-induced apoptosis. Together these findings indicate that BCR-ABL confers ril 1
3
, 2
chemotherapeutic resistance through intracellular signaling to the GATA-RE element found in 0
1
9
the promoter region of the anti-apoptotic Hsp70 protein. We suggest that downregulation of the
GATA-Hsp70 pathway may be useful in the treatment of chemotherapy-resistant CML.
2
BCR-ABL transactivation of Hsp70
INTRODUCTION
CML, a diagnosis given for 15 % of adult leukemias, represents a clonal expansion of
granulocytic progenitor cells (1). CML is caused by a reciprocal t(9;22) translocation of c-abl
from chromosome 9 to the breakpoint cluster region (bcr) gene on chromosome 22 (1;2). In 95
% of CML cases, translocation involves the major bcr, leading to generation of the 210 kDa
BCR-ABL fusion oncoprotein (p210BCR-ABL ). A similar BCR-ABL translocation is also be
found in 20-30 % of cases of adult acute lymphoblastic leukemia (ALL) and 5 % of childhood
ALL (3;4). In both cases, the serine-threonine kinase domain from BCR partially replaces the
inhibitory NH -terminal SH domain of ABL, producing a fusion protein with constitutive ABL
2 3
tyrosine kinase activity (3;5;6).
Dysregulated BCR-ABL signaling induces pleiotropic phenotypic changes in
D
o
w
granulocytic cells, including resistance to chemotherapy-induced apoptosis, disruption of cell n
lo
a
d
cycle checkpoints, induction of growth factor-independent proliferation (7), cellular ed
fro
transformation in vitro (3), and production of a CML-like leukemia in vivo (8;9). Moreover, m
h
ttp
BCR-ABL signaling is required to actively maintain cellular viability. For example, ://w
w
w
downregulation of BCR-ABL expression (10;11) or inhibition of its ABL tyrosine kinase activity .jb
c
.o
using the 2-phenlyaminopyridmidine inhibitor, imatinib (STI571) induces apoptosis in vitro and rg
b/
y
remissions during the chronic phase of CML in the clinic (10;12). A body of work has shown gu
e
s
that BCR-ABL regulates diverse signaling pathways including p21ras, PI3 kinase, protein kinase t on
A
p
C [PKC, (13)], Jak-STAT (3;14), and NF-κB (15;16). The relationships of these signaling ril 1
3
, 2
pathways to specific cellular responses are still being elucidated. Perhaps best understood is the 01
9
p21ras /Raf pathway in mediating cellular transformation by BCR-ABL. Inhibition of p21ras,
either by administration of antisense oligonucleotides or microinjection of blocking antibodies,
prevents expression of the c-myc proto-oncogene, thereby blocking cellular transformation
[reviewed in (17)].
A hallmark of BCR-ABL-expressing cells is the profound resistance to chemotherapeutic
agent-induced apoptosis (13;16;18;19). Chemotherapy-induced cell death involves proteolytic
cleavage of cysteine proteases, called caspases, whose activation can be triggered by cytochrome
c release from mitochondria (20). In the cytosol, cytochrome c binds apoptotic protease
activation factor (Apaf) -1, inducing its oligomerization to form the “apoptosome”(21), an
enzymatically competent Apaf-1/procaspase-9 complex that cleaves procaspase –3, thereby
committing the cell to autolysis.
3
BCR-ABL transactivation of Hsp70
BCR-ABL causes chemotherapy resistance, at least in part, by interfering with the
caspase activation pathway at multiple steps. Previous studies have shown that Ph+ cells
isolated either during CML blast crisis or after ectopic BCR-ABL expression release diminished
amounts of cytochrome c into the cytosol after exposure to high dose Ara-C or etoposide (22;23).
At the biochemical level, BCR-ABL transformed cells express increased amounts of the outer
mitochondrial membrane protein Bcl-x which prevents Bax insertion into the outer
L
mitochrondial membrane, reducing chemotherapy-induced cytochrome c release (23). In
addition, we have recently shown that BCR-ABL signaling through the PKCι−NF-κB pathway is
also required for resistance to paclitaxel (Taxol)-induced apoptosis (16). NF-κB is an inducible
transcription factor that activates expression of several inhibitors of apoptosis (IAP) proteins,
polypeptides that bind and inactivate caspases as well as inducing ubiquitin-mediated
D
o
w
degradation of the RHG proteins [Reaper, HID, and Grim, Ref (24)]. Together these findings n
lo
a
d
suggest that BCR-ABL may induce chemotherapy resistance by controlling expression of ed
fro
proteins that antagonize caspase activation at multiple points. m
h
ttp
Several studies have begun to identify BCR-ABL activated genomic programs, focusing ://w
w
w
on BCR-ABL activated genes important in cell cycle regulation or molecular signatures (25-27). .jb
c
.o
However, whether BCR-ABL activates other genetic targets controlling anti-apoptosis has not rg
b/
y
been completely resolved. Here we apply discovery-based tools to investigate the effects of gu
e
s
t o
BCR-ABL on gene expression. Differential expression of a variety of signaling kinases, n
A
p
transcription factors, cell surface receptors and adhesion molecules was observed in HL-60 cells ril 1
3
stably expressing p210BCR-ABL vs empty vector controls. Further investigation focused on the , 20
1
9
BCR-ABL induced upregulation of Heat shock protein (Hsp)- 70 kDa isoform, an anti-apoptotic
protein known to bind and inhibit Apaf-1, apoptosis inducing factor (AIF) and inhibit the
apoptosis signal-inducing kinase [ASK1, Ref(28-30)]. After Hsp70 upregulation was validated
in K562 cells and bone marrow aspirates from CML patients, subsequent experiments mapped
sequences responsible for BCR-ABL-mediated transactivation to a high affinity GATA response
element (GATA-RE) located in the proximal Hsp70 promoter between –82 to –58 nt. Finally,
short interfering RNA (siRNA)-mediated downregulation of Hsp70 expression sensitized K562
cells to paclitaxel-induced apoptosis. Together these findings indicate that p210BCR-ABL
transactivates the GATA-RE in the Hsp70 promoter, leading to upregulation of the anti-apoptotic
Hsp70 gene.
4
BCR-ABL transactivation of Hsp70

EXPERIMENTAL PROCEDURES

Cell culture and treatment- Human K562 erythroleukemia (16), HepG2 hepatocellular
carcinoma (31) and HL-60 acute myeloid leukemia cells stably transfected with BCR-ABL (HL-
60 /BCR-ABL ) or empty vector ( HL-60/Neo) were cultured as described (23). Where
indicated, paclitaxel (Sigma Aldrich) was resuspended in DMSO and added to a final
concentration of 1 µM.
In some experiments, cells were isolated from leukemic bone marrow aspirates. Bone
marrow aspirates were obtained after informed consent was obtained under an institutional
review board-approved protocol at the time of diagnosis or treatment on the Adult Leukemia D
o
w
Service at the Johns Hopkins Hospital. Fractions of mononuclear cells or granulocytes were nlo
a
d
e
isolated on double Ficoll-Hypaque gradients (32), washed with RPMI 1640 medium containing d fro
m
10 mM HEPES (pH 7.4), lysed in 6 M guanidine hydrochloride under reducing conditions, h
ttp
reacted with iodoacetamide, dialyzed into 0.1% (w/v) SDS, and lyophilized as described (33). ://w
w
w
Membrane based cDNA Macroarrays- Total RNA was extracted from control (HL- .jb
c
.o
60/Neo) or p210BCR-ABL –expressing (HL-60 /BCR-ABL) cells by acid guanidium-phenol brg/
y
g
u
extraction (TRI -reagent, Sigma, St. Louis, MO) and treated with Dnase. Five micrograms were e
s
t o
n
reverse transcribed in the presence of 35 µCi of [α33P]dATP and cDNA purified by column A
p
chromatography (Chroma spin-200, Clontech). Atlas 1.2 Arrays (Clontech) were hybridized ril 13
, 2
0
with 106 cpm/ml of probe at 68 °C and washed as recommended by the manufacturer. 19
Membranes were exposed to a PhosphorImager cassette and relative changes in hybridization
intensity determined by AtlasImage 1.01 software (Clontech, Palo Alto, CA ). Comparisons of
mRNA populations between control and p210BCR-ABL –expressing cells were performed with two
different sets of Atlas Array membrane lots in two independent experiments.
For each gene, local background was subtracted from the total hybridization intensity and
the average signal intensity determined for duplicate spots. To compare differences in gene
expression between arrays, background subtracted average intensity was normalized that of
housekeeping genes. Only those genes which showed an average 3.5-fold upregulation or
downregulation across duplicate membranes were further considered.
High density oligonucleotide arrays- Four independent RNA samples were prepared
5
BCR-ABL transactivation of Hsp70
from the HL-60/Neo and the HL-60/BCR-ABL transfected cells for hybridization to the Hu95A
GeneChip (Affymetrix, Santa Clara, CA). First-strand cDNA synthesis was performed using
total RNA (10-25 µg), a T7 – (dT) oligomer and SuperScript II reverse transcriptase (Life
24
Technologies). Second strand synthesis, target RNA labeling and hybridization were as
previously described (34). Gene Chip arrays were scanned using a Gene Array Scanner (Hewlett
Packard) and analyzed using the Gene Chip Analysis Suite 4 software (Affymetrix Inc). The
average difference statistic was retrieved for quantification of mRNA abundance in those
samples in which the absolute call indicated that the gene was present.
Oligonucleotide Microarray Data analysis- Reproducibility of the four independent
microarrays was determined by calculating the correlation coefficient for the log-transformed
average difference values for the probe sets in each array (34). For each pairwise comparison,
D
o
w
the mean correlation coefficient was 0.945 ± 0.024 (n = 6) in the HL-60/Neo data sets, and 0.977 n
lo
a
d
± 0.010 ( n = 6 ) for the HL-60/BCR-ABL data sets, indicating that the measurements were ed
fro
m
highly reproducible ( Table I ). For comparison of the fluorescence intensity (Average h
ttp
Difference) values among multiple experiments, the Average Difference values for each ://w
w
w
“experimental” GeneChip were scaled to that of the “base” GeneChip (34). Genes differentially .jb
c
.o
expressed were identified by one way ANOVA with replicates comparing the Average brg/
y
g
Difference values of a probe sets in HL-60/Neo vs HL-60/BCR-ABL. Genes (probe sets) with a u
e
s
t o
p-value [Pr(F)] <0.0001 were selected for further analysis. Agglomerative hierarchical n A
p
clustering using the Unweighted Pair-Group Method with Arithmetic mean [UPGMA, Ref (34)] ril 1
3
, 2
was performed on the indicated genes (Spotfire Array Explorer, v. 7, Spotfire Inc., Cambridge 01
9
MA). Data are graphically presented as heat maps in which fluorescence intensity is represented
by a color gradient.
Northern blot analysis- RNA was isolated from K562, HL-60 Neo and HL-60 BCR-
ABL cells using RNAqueous (Ambion). 20µg of total RNA was fractionated by electrophoresis
on a denaturing formaldehyde 1% agarose gel (13), transferred to BrightStar membrane
(Ambion) and immobilized by UV crosslinking. Full length, linearized Hsp70 cDNA, GAPDH
control probe (Ambion), and 18S rRNA control probes were radiolabeled with 32P using the X
kit (Amersham), hybridized and washed (13).
Western blot analysis- Western blots were performed as described previously (13).
Briefly, cells were counted and equal numbers were harvested and lysed directly into Laemmli
sample buffer. Equal volumes of sample were then fractionated by 5-20% SDS-PAGE gradient
6
BCR-ABL transactivation of Hsp70
gel (Invitrogen), transferred to nitrocellulose membranes and the blocked with 5% nonfat dry
milk in PBS/ 1% Tween 20. Blots were then probed with the indicated primary antibody;
immune complexes were detected by enhanced chemiluminescence (ECL, Amersham ) or, where
indicated, near–infrared fluorescence (Odyssey Imaging System, LiCor BioSciences).
Plasmids- The -259/ +37 Hsp70 /LUC promoter-driven luciferase reporter plasmid was
constructed by PCR of HL60 genomic DNA using the upstream primer – 259 5’-
ACGGATCCCACCGCCA CTCCCCCTTC-3’, and the Hsp70 downstream primer 5’-
AAAAAGCTTGTGGACT GTCGCAGCAGCTC-3’ (Hind III site underlined). The restricted
gel purified PCR product was ligated into pOLUC reporter digested with the same endonucleases
(35). A series of 5’ deletions were constructed by PCR using Hsp 70 (-259/ +37) /LUC as
template with the upstream primers -200, 5’-GTGGATCCCAG AAGACTCTG-3 ’; -163, 5’-
D
o
GCGGATCCCTGG CCTCTGATT-3’; -123, 5’-GGGGATCCAC GGGAGGCGAAA-3’; -82, w
n
lo
a
5’-CCTGGATCCCTCA TCGAGCTC-3’; -58, 5’-GATTGGATCCGA AGGGAAAAGG-3’ and de
d
fro
the Hsp70 downstream primer. Each 5’ deletion was digested with Bam H1 (site is underlined) m
h
ttp
and Hind III, gel purified and ligated into pOLUC. The site directed mutation of the GATA ://w
w
binding element was constructed by PCR SOEing (36) using the sense primer 5’- w
.jb
c
.o
GAGCTCGGTCTCAGG CTCAGGATA-3’ and the antisense primer 5’- TCTGAGCCTGAGA rg
b/
y
CCGAG-3’ (mutations underlined) with the Hsp70 downstream primer. g
u
e
s
The multimeric GATA (WT) and GATA (Mut)-driven reporter was constructed by t on
A
p
annealing the sense and antisense oligonucleotides (tabulated below). ril 1
3
GATA (WT): GATCGAGCTCGGTGATTGGCTCAGAA , 20
1
9
CTCGAGCCACTAACCGAGTCTTCTAG
GATA (Mut): GATCGAGCTCGGTCTCAGGCTCAGAA
CTCGAGCCAGAGTCCGAGTCTTCTAG
Duplex oligonucleotides were then phosphorylated, ligated with T4 DNA ligase, 3 copies ligated
into Bam HI linearized –59/+22 rAGT/LUC, driven by the inert angiotensinogen TATA box
(35). pCMV-FLAG BCR-ABL expression vectors were constructed by ligating the BCR-ABL
coding sequences into pCMV-Tag plasmid (37). All plasmids were purified by ion exchange
chromatography (Qiagen) and cloned inserts sequenced to confirm authenticity.
Cell transfection and reporter assays- Transient transfections in K562 cells were carried
out using DMRIE-C reagent (Gibco, BRL). For reporter plasmids, 4 to 10 µg of plasmid DNA
was resuspended into 0.5 ml OPTI-MEM® I reduced serum medium (Gibco, BRL), mixed with
7
BCR-ABL transactivation of Hsp70
an equal volume of medium containing 10 µl DMRIE-C Reagent and incubated at room
temperature (RT) for 15 min to allow lipid-DNA complex formation. Logarithmically growing
K562 cells (2 x 106 cells) were centrifuged, resuspended in 0.5 ml reduced serum medium and
added to the lipid-DNA complex and incubated at 37 oC in a CO incubator for 5 h. After
2
transfection, growth medium (containing 10% FBS) was added and the cells were incubated
overnight. For reporter assays, cells were harvested at the indicated time points and washed two
times with cold PBS. Cytoplasmic lysates were prepared and independently measured for
luciferase and β-galactosidase activity (Promega, Madison, WI) as described previously (35).
Luciferase reporter activity was normalized to the internal control of β-galactosidase activity to
control for differences in transfection efficiency.
Electrophoretic Mobility Shift Assay (EMSA)- Nuclear proteins were purified over a
D
o
w
sucrose cushion (31) normalized for protein concentration by the Coomassie G250 assay (Bio- n
lo
a
d
Rad, Hercules, CA). The GATA (WT) monomeric duplex was radiolabeled with Klenow ed
fro
m
polymerase and purified by gel filtration chromatography. DNA-binding reactions were carried h
ttp
out in a mixture of 20 µg of nuclear proteins, 12 mM HEPES (pH 7.9), 40 mM KCl, 120 mM ://w
w
w
NaCl, 0.2 mM EDTA, 0.2 mM EGTA, 0.4 mM dithiothreitol, 0.5 mM PMSF, 8% glycerol, 2 µg .jb
c
.o
of poly-(dA-dT), and 20,000 cpm of α-32P-labeled double-stranded GATA (WT) probe in a total brg/
y
g
u
volume of 20 µl. The reaction mixture was incubated on ice for 15 min and then fractionated by es
t o
n
6% nondenaturing polyacrylamide gel electrophoresis (PAGE) containing 2% glycerol. Gels A
p
were dried and subjected to autoradiography using Kodak X-AR film at -70°C. Competition was ril 13
, 2
0
1
performed by the addition of a 100-fold molar excess of nonradioactive double-stranded 9
oligonucleotide competitor at the time of addition of the radioactive probe. For supershift, anti-
GATA-1 antibody was added to the gel shift reaction and incubated on ice for 1 h prior to
fractionation by nondenaturing PAGE.
Apoptosis- Cells were stained with Annexin V-phycoerythrin (PE) using an Annexin V-
PE Apoptosis Detection Kit I (BD Pharmingen). Briefly, 5x105 washed cells were collected by
centrifugation and resuspended in Annexin V-binding buffer. Cell suspensions were then
incubated with Annexin V-Phycoerythrin (PE) at 1 µg/ml and 7 aminoactinomycin D (7AAD) at
a final concentration of 1 µg/ml for 25 min at room temperature in the dark. The percentage of
apoptotic cells was determined by flow cytometry (Becton-Dickinson FACScan) analysis (38).
8
BCR-ABL transactivation of Hsp70
siRNA mediated Hsp-70 “knockdown”- Various concentrations from 50-200 nM Hsp
70 or LaminA/C siRNA ( custom SMART pool, HSPA1L-NM_005346, and siRNA CONTROl,
Dharmacon Research Inc., Lafayatte, CO, USA ) were substituted for the reporter plasmid and
transfected into logarithmically growing K562 cells using DIMRIE-C as described above. After
5 h, growth medium was added and cells returned to culture in the absence or presence of
paclitaxel (1 µM final concentration) for the times indicated.

RESULTS
Identification of genes downstream of the BCR-ABL pathway- In this study, we
initially determined gene expression profiles in HL-60 cells stably expressing p210BCR-ABL
because these cells show enhanced Bcl-x expression and decreased cytochrome c release in
L D
o
w
response to chemotherapeutic agent administration (23). To confirm stable p210BCR-ABL nlo
a
d
e
expression, Western immunoblot analysis was perfomed using an anti-c-Abl antibody. HL- d fro
m
60/BCR-ABL cells express amounts of p210BCR-ABL comparable to Ph+ K562 cells, whereas the h
ttp
HL-60/Neo controls have no detectable BCR-ABL antigen (Fig. 1). To identify genomic targets ://w
w
w
of p210BCR-ABL we systematically examined differences in mRNA expression by cDNA .jb
c
.o
rg
macroarrays and high density oligonucleotide microarrays. Using membrane-based cDNA b/
y
g
macroarrays containing 1,176 sequenced human gene probes, we identified 25 genes whose ue
s
t o
n
expression was increased by 3.5-fold (or greater), and 34 genes whose expression was reduced A
p
by 3.5-fold (or greater) in duplicate pairwise experiments (Table II). The major biochemical ril 1
3
, 2
0
functions of the differentially expressed genes were apoptosis (Bcl-x , Hsp70), cellular signaling 1
L 9
[cell surface receptors (ApoE, PAF) and intracellular signaling kinases (PKCβ-1, PLCγ, RAP1
GAP)], DNA synthesis/repair, transcription (Id), cell adhesion/and immune recognition, and
extracellular matrix turnover/regulation. Importantly, we noted that c-Abl expression
(representing signal from ectopic p210BCR-ABL expression ) was ~ 7.99 ± 2.11-fold upregulated-,
and that Bcl-x was 4-fold upregulated in the BCR-ABL transfected relative to control cells,
L
confirming the findings of others (23). Together these indicated a valid data set was generated
by the macroarray. The 23 -fold upregulation of inducible heat shock protein (Hsp)-70 kDa
isoform in the HL-60/BCR-ABL cells (Table II, Fig. 2) was particularly noteworthy because Hsp
70 inhibits apoptosis by interfering with oligomerization of Apaf-1 (39;40) and function of
apoptosis inducing factor [AIF, Ref(28)].
9
BCR-ABL transactivation of Hsp70
Differences in gene expression were further explored using high density oligonucleotide
microarrays representing 12,626 sequenced human gene probes. In this experiment, 4 replicates
(each representing an independent isolation and hybridization) of HL-60/Neo and 4 for HL-
60/BCR-ABL were performed, and genes differentially expressed identified by one way
ANOVA. This analysis identified 487 unique genes whose expression was changed by BCR-
ABL at a p value [Pr(F)] of <0.001 (293 were upregulated) and 36 at p<0.000001 (22 were
upregulated). These latter genes were classified by primary biochemical function (Table III).
The most highly regulated group included the pim-1 oncogene, signaling kinases (a guanine
nucleotide exchange factor and Ras homolog), transcription factors (SRY and Kruppel-like
factors), a family of cell surface antigens originally identified in melanoma cells ( the GAGE
antigens 5-7 ), and alpha catenin. For visualization of the pattern of gene expression,
D
o
hierarchical clustering and heat map analysis was performed. This analysis identified 22 genes w
n
lo
a
whose expression was undetectable in the HL-60/Neo cells yet strongly upregulated by p210BCR- de
d
ABL expression as well as 16 genes whose expression was inhibited by p210BCR-ABL expression from
h
ttp
(Fig. 3). A high concordance was found in expression patterns for genes represented in both the ://w
w
spotted cDNA macroarray and the high density oligonucleotide arrays. In particular, Hsp-70 w
.jb
c
.o
expression was called “Absent” in the HL-60/Neo data set, but was strongly detected in all four rg
b/
y
of the HL-60/BCR-ABL microarrays (26,159 ± 6,178 scaled units, Pr(F) < 0.000031), supporting g
u
e
s
the macroarray finding that Hsp 70 expression is upregulated by p210BCR-ABL . t on
A
p
Validation of differential expression of Hsp70- To validate the microarray studies, ril 1
3
, 2
Northern blot analysis was performed using radiolabeled cDNA to human Hsp70. A strong 0
1
9
induction of the 2.4 kb Hsp70 mRNA was seen in the HL-60/BCR-ABL and K562 cells when
compared to HL-60/Neo cells (Fig. 4A). Immunoblot analysis confirmed that the increased
Hsp70 mRNA resulted in increased Hsp70 protein. As seen in Fig. 4B, strong cytoplasmic
staining of Hsp 70 was seen in the HL-60/BCR-ABL and K562 cells, but was undetectable at
this exposure in HL-60/Neo cells. Detection of cytosolic Hsp70 is particularly relevant because
Hsp70 complexes with Apaf-1 in this compartment to inhibit cytochrome c- induced apoptosome
formation.
Hsp70 expression in clinical CML samples- To determine whether Hsp70 expression
also occurs in clinical CML, enriched fractions of granulocytic and mononuclear cells purified
from bone marrow aspirates of patients at various stages of the disease were assayed for Hsp70
expression by Western immunoblots. As seen in Fig. 5, Hsp70 staining was detectable in
10