| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Laboratory of Molecular Oncology, MSZ, Universität Würzburg, Germany;
Klinische Molekularbiologie und Tumorgenetik, GSF, Munich, Germany;
IPF, Hannover, Germany; and
§ M. D. Anderson Cancer Center, Houston, Texas 77030, USA
1Correspondence: Laboratory of Molecular Oncology, MSZ, Versbacher Str. 5, D-97078 Würzburg, Germany. E-mail stephan.feller{at}mail.uni-wuerzburg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: CRKL • adapter protein • Bcr-Abl • MAP kinase (MAPK)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
). They are found in over 90% of
human chronic myeloid leukemias. Previous work has shown that the
tyrosine kinase activity of Bcr-Abl is greatly elevated compared to
c-Abl and that this elevation directly correlates with the
transformation potency of Bcr-Abl (5)
. Bcr-Abl expression
in cells affects a plethora of signal transduction pathways, including
the Ras-MAP kinase (MAPK) cascade, JNK/SAPK family kinases, and PI3
kinase (1
2
3
4
, 6)
. Despite this wealth of data, it has been
difficult to distinguish primary signaling events set off by Bcr-Abl
from secondary and tertiary effects and to determine essential
signaling cascades that are truly specific for Bcr-Abl. Hints for
possible candidates of Bcr-Abl signaling initiators have come from the
identification of Bcr-Abl substrates like CRKL, Paxillin, and c-Cbl as
well as proteins that can bind to Bcr-Abl, including again CRKL and
c-Cbl, but also adapter proteins like Grb2, SHC, and others (2
, 4
, 6)
.
CRKL was discovered as a gene near one of the chromosomal breakpoints
evident in Bcr-Abl-expressing cells and as a prominent
tyrosine-phosphorylated protein in human chronic myeloid leukemia (CML)
cells (2
, 6)
. From its sequence, it was obvious that CRKL
is an adapter protein and shares high homology with the previously
discovered Crk adapter proteins. Besides a great sequence similarity of
the SH2 and SH3 domains, a homologous region exists in c-Crk-II and
CRKL around a tyrosine between the SH3 domains (6)
. This
tyrosine is phosphorylated by Bcr-Abl and is assumed to play a role in
the regulation of protein conformation in c-Crk-II and CRKL.
Of the two SH3 domains found in c-Crk-II and CRKL, only the first
ones [SH3(1)] have been documented to function as mediators of highly
selective protein–protein interactions. The SH3(1) domains of c-Crk-II
and CRKL bind with high affinity to certain proline-rich sequences that
conform to the consensus P-x-x-P-x-K (6
, 7)
. A
high-resolution crystal structure revealed that the lysine following in
position + 2 of the SH3 typical P-x-x-P motif is crucial for high
binding affinity and selectivity (8)
. Mutational analyses
of naturally occurring binding motifs indicated that residues outside
the consensus motif also influence the binding affinity of these
sequences (7
, 9)
. In general, SH3 binding affinities of
short synthetic peptides corresponding to naturally occurring sequences
are in the micromolar range. In a previous study we generated chimeric
peptides that bind with nanomolar affinities to the SH3(1) domains of
Crk and CRKL (9)
. This study also showed that the
high-affinity Crk/CRKLSH3(1) binding peptides (HACBPs) have a unique
selectivity for Crk/CRKL when compared to thousands of cellular
proteins from lysates of metabolically labeled K562 (CML blast) cells.
This remarkable degree of HACBP binding selectivity is an important
prerequisite for reliable results when these peptides are used for
studies with living cells or for future in vivo studies with
animals.
In the current study we attempted to determine whether the SH3(1)-mediated binding of CRKL to Bcr-Abl is important for the function of this leukemic oncogene. Thus, shuttle tags that allow a rapid, receptor-independent uptake of peptides into cells were attached to the HACBPs. Furthermore, affinities of the HACBPs were again improved. The newly developed HACBPs are capable of disrupting Bcr-Abl–CRKL complexes in CML cells, resulting in a decreased MAP kinase activity and a strong reduction of proliferation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Fluorescence spectrometry
Measurement of binding affinities based on the interaction of
the peptides with aromatic residues (predominantly tryptophan) in the
SH3 domains was done as described (9)
on a LS50B
spectrometer (Perkin Elmer) with a water-cooled cuvette chamber.
We were unable to obtain usable results with the Trp-containing
peptides due to the high background fluorescence of the peptides. These
peptides were therefore characterized by other assays (details below).
The disulfide bond-containing peptides (see Table 1B
) could be measured only prior to coupling to the
Trp-containing Antp sequence.
|
Cell culture and protein complex formation inhibition
Human Bcr-Abl(p210)-positive CML cells K562 (ATCC, Rockville,
Md.) and BV173 and human AML cells KG-1a (gifts of S. Frühauf,
University Heidelberg) were cultured in RPMI1640 with 5% fetal bovine
serum (FBS). B15 cells (Bcr-Abl-positive human acute lymphoblastic
leukemia; ref 10
) were supplemented with 15% FBS. Daudi
and Raji (Burkitt’s lymphoma), HL60 (promyelocytic leukemia), and
Ba/F3 (mouse pre-B) cells were cultured with 10% FBS. Ba/F3 cells were
also supplemented with interleukin 3-containing conditioned WEHI-3
medium. For serum deprivation, K562 cells were washed three times with
serum-free medium and then cultured for 3 days. Cell lysis was
previously described (9)
. PC12 pheochromocytoma cells were
cultured in DMEM with 5% FBS and 3% horse serum and antibiotics.
In vitro inhibition of CRKL complexes by Antennapedia- and
integrin peptides (nomenclature in Table 1B
) in Fig. 1A
was carried out with total cell lysates. Twenty micrograms
of GST-tagged HACBP fusion peptide (9)
was preincubated
for 2 h at 4°C in 500 µl IP buffer containing 2% ovalbumin
and protease inhibitors (9)
, Escherichia coli
protein extract (1 mg; ref 9
), and glutathione-Sepharose
beads. K562 protein (1 mg) and the different peptides (30 µM) were
then added and incubated overnight. Precipitates were washed three
times with RIPA buffer (9)
. After sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and blotting,
CRKL bound to GST-HACBP fusion peptide was detected with anti-CRKL
(Santa Cruz, sc-319, Santa Cruz, Calif.) and 25 µCi
35S-protein A (Amersham Pharmacia, Little
Chalfont, U.K.).
|
In vitro inhibition of Grb2–SoS complex formation was done
similarly: K562 protein (200 µg) and the different peptides (10 µM)
were mixed; 25 µg of GST-Grb2 (9)
immobilized on
glutathione-Sepharose in the presence of 500 µg E. coli
protein extract was then added and the samples were incubated
overnight. Precipitates were washed three times with RIPA buffer
(9)
, separated by SDS-PAGE, blotted, and probed with
SoS-antiserum (Upstate Biotechnology Inc., #06–246, Lake Placid,
N.Y.).
For inhibition assays with living cells in Fig. 3
, the K562 cells were
washed and resuspended in medium supplemented with DNase I (20 µg/ml;
Sigma, D4513, St. Louis, Mo.). Cells were seeded into 24-well plates at
a density of 1 x 105 per 0.2 ml and
incubated for 30 min at 37°C. Thereafter, the peptide solution that
had been mixed with 0.2 ml of medium was slowly added to the cells. On
the next day cells were again treated with DNase I and peptide.
Evaluation on day 3 was done by counting the viable cells after
staining with trypan blue dye. Activity of MAP kinase was evaluated
with activation-specific anti-phosphoMAPK (New England Biolabs, #9101,
Beverly, Mass.) following the manufacturer’s instructions.
|
Immunofluorescence
PC12 cells were seeded onto glass coverslips. On the next day
they were serum-starved for 16 h and treated with DNase and 100
µM of the indicated peptides, as described above, for 5 h. Cells
were then rapidly washed twice with cold phosphate-buffered saline
(PBS) containing 2 mM Na2EDTA and high salt (500
mM NaCl) and fixed with an ice-cold ethanol-acetic acid solution
(90/10) for 5 min. After washing with PBS, fixed cells were blocked for
30 min with PBS supplemented with 10% FBS. Cells were then stained
with DTAF-streptavidin (Jackson ImmunoResearch, #016–010-084, West
Grove, Pa.; 1:200 diluted in block buffer) and analyzed by fluorescence
microscopy. Cells without peptide addition were photographed for the
same length of time as peptide-treated cells to detect nonspecific
background fluorescence of DTAF-streptavidin.
Peptide stability assays
Evaluation of Bio-Antp-HACBP half-life in K562 cells was done by
peptide addition as described above. At indicated time points, cells
were washed with cold PBS and then lysed with RIPA buffer. Biotinylated
peptide was precipitated with streptavidin-agarose. Precipitates were
washed three times with RIPA and subjected to SDS-PAGE. 10 µl of
culture supernatant was also analyzed. After blotting onto PVDF,
peptides were immobilized with 4% glutaraldehyde in PBS for 20 min and
detected with horseradish-labeled streptavidin (AmershamPharmacia,
RPN1231).
Isolation and peptide treatment of patient CML blast cells
Peripheral blood from CML patients with active disease was
diluted with 1 volume of PBS; 10 ml of this diluted blood was layered
onto 5 ml of Histopaque-1077 (Sigma, #1077–1). A low-density gradient
was obtained by centrifugation in a swing-out rotor (400 g,
20 min, 20°C). The low-density cell fraction containing the blasts
was aspirated with a Pasteur pipette. Blasts were then washed twice
with PBS (2000 g, 10 min, 4°C) and cultured in RPMI1640
with 10% FBS. Cell-penetrating peptides (20 µM) were added as
described above for K562 cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Furthermore, complexes of Bcr-Abl and CRKL are detectable by
coimmunoprecipitation (reviewed in refs 1
, 2
, 4
, 6
).
Molecular details of the Bcr-Abl–CRKL complex formation were not known
previously. Besides a CRKLSH3(1)-dependent interaction, binding of the
phosphorylated CRKLTyr207 to the SH2 domain of Abl and other binding
modes seemed possible.
|
To test whether complexes between Bcr-Abl and CRKL depend on the SH3(1)
domain of CRKL and whether they are functionally important in CML
cells, cell-penetrating peptides to selectively disrupt this
interaction were generated. Peptides used in the experiments are
summarized in Table 1B
. The starting point for the experiments was an
HACBP generated in a previous study by chimerization of naturally
occurring CRKLSH3(1) binding sequences in C3G and SoS1
(9)
. This peptide (designated P2; see Table 1B
) displays a
unique binding selectivity for the SH3(1) domains of Crk and CRKL.
To use these peptides in living cells, it is necessary to attach
an epitope that allows them to enter cells efficiently. Several reports
of peptide sequences that perform a receptor-independent cell entry
have been published in recent years. The best-characterized ‘shuttle
tag’ is the third 
-helix of the Drosophila transcription
factor Antennapedia (reviewed in ref 11
), but sequences
from the human ß3-integrin chain
(12)
, the HIV Tat protein (13)
, a synthetic
amphipathic helix (14)
, and several other sequences have
been reported.
To successfully disrupt the constitutive Bcr-Abl–CRKL complexes in CML blast cells, several requirements must be met. The shuttle must be sufficiently stable and cannot be toxic when cells are incubated for extended periods. It should also not interfere with the binding of the ‘cargo’ peptides to the target. Moreover, the shuttle should transport its cargo in significant amounts to the cytoplasm where the Bcr-Abl protein is located.
Three shuttle sequences were tested in our study. Peptides containing
the synthetic shuttle sequence described by Oehlke et al. (P3 and P4)
were toxic for K562 CML blast cells when the incubation was carried out
for 12 h or more (data not shown); therefore, they were not used
further. The ß3-integrin shuttle
sequence-containing HACBP peptide (P5) showed a significantly reduced
binding to the CRKLSH3(1) domain (compare affinities of P2 and P5; also
see competition assay in Fig. 1A
) and therefore was not used
further.
Finally, peptides coupled to a mutated Antennapedia shuttle
(abbreviated Antp) were tested. In this sequence a glutamine essential
for DNA binding is mutated to a proline, strongly reducing the DNA
binding and preventing an accumulation in the nucleus (Pro50 mutant;
10). As shown in Fig. 1A
, Antp-HACBP (P8) displayed no
reduced CRKL binding when compared to the HACBP alone (P2) in an
in vitro competition assay.
Accumulation of Antp (P7) and Antp-HACBP (P8) in the cytoplasm was
confirmed for several cell lines by immunofluorescence using
biotinylated peptides and detection with streptavidin-DTAF. In
Fig. 2
, representative results are shown for the adherent cell line PC12,
where cytoplasm and nucleus are easily recognizable as separate
regions, different from most hematopoietic cells. Furthermore, CRKL
protein could be precipitated with a biotinylated form of the peptide
P8 from cell lysates (not shown).
|
Point mutant peptides were made to distinguish nonspecific toxicity
from specific effects resulting from the HACBP binding to CRKL. Mutants
were initially analyzed in vitro by fluorescence
measurements and competition assays (Fig. 1A
). While
mutation of a crucial proline or lysine residue results in a marked
drop in affinity for CRKL (7
8
9)
, all single point mutant
peptides tested retained a significant residual binding activity (data
not shown).
Thus, for the longest HACBP with the highest affinity (P14;
Kd=35 nM), a 4-site mutant control peptide (P15) was
synthesized. This mutant peptide was void of any detectable residual
binding (Table 1B)
. The affinity measurements of peptides toward the
SH3 domains by tryptophan fluorescence had to be carried out in the
absence of the Antp shuttle peptide since this contains tryptophans,
resulting in a very high background fluorescence of the peptide.
Determination of peptide stabilities in cultured cells
To analyze the degradation of biotinylated Antp-HACPB (P8) in
living cells, a single 10 µM peptide dose was added to K562 cells for
up to 48 h (Fig. 1B
). This continuous Antp-HACBP
molecule (without a disulfide bond between shuttle and cargo) is
expected to shuttle persistently back and forth between culture medium
and cellular interior until it is bound or degraded. At indicated
times, aliquots of medium and cells were frozen in SDS-PAGE sample
buffer. Samples were analyzed for remaining peptides and fragments by
SDS-PAGE and detection with streptavidin-peroxidase/ECL. Within the
lysates, peptide remains detectable until 36 h, with a half-life
of ~12 h. Significant degradation of peptide in the K562 culture
medium is visible after 6 h (fragments indicated by filled
arrowheads). These fragments were not detected in the lysate,
indicating that they do not enter cells efficiently. Serum protease
activity clearly accounted for a prominent portion of the extracellular
degradation, according to our experiments with cell-free medium
containing different concentrations of fetal bovine serum (unpublished
data). Based on these results, a protocol with a once daily application
of new peptide was chosen for the subsequent studies with K562 cells.
Analysis of peptide effects on cell proliferation
From several previous studies, it is known that Bcr-Abl and CRKL
exist in a constitutive complex. The blocker peptides thus are only
expected to work on complexes that spontaneously disassemble and on
complexes that are newly formed due to protein turnover and cell
proliferation. Taking this into account, K562 cells were maintained in
the presence of Antp-HACBP (P8), Antp alone (P7), or mutant control
peptide (P11) for 48 h, allowing several rounds of cell division.
Without HACBP addition, the total number of cells increased ~ninefold
under the conditions used (from 1x105 to
9x105 cells/well). Cell proliferation was
monitored by counting viable cells. Antp shuttle alone and mutant
peptide showed little inhibition of cell growth up to a concentration
of 200 µM. By contrast, Antp-HACBP (P8) induced a highly reproducible
partial growth inhibition (Fig. 3A
). We did not see signs of an increased apoptosis based on
cell morphology (membrane blebbing, cell fragmentation; data not
shown).
Since Antp-HACBP is degraded within the K562 cells after ~48 h, the block of cellular proliferation should be reversible if there is not some sort of ‘poisoning’ leading to a permanent growth arrest. This recovery of the proliferation was indeed observed on prolonged cultivation without new peptide.
To further exclude nonspecific effects of the peptides on cell growth,
the Bcr-Abl negative cell lines Ba/F3, HL60, KG-1a, Daudi and Raji
cells were tested. No significant difference in proliferation was seen
with Antp-HACBP (P8) or the control peptide (P11) (Fig. 3B
).
At this stage, we were not satisfied with the concentration of SH3 blocker peptide required for a partial inhibitory effect on proliferation. To make the peptides more potent, two strategies were tested. First, peptides with a disulfide bond instead of a peptide bond between shuttle and cargo were synthesized. The disulfide bond is expected to open up under the reducing conditions inside of cells, allowing the shuttle peptide to exit the cell but trapping the cargo peptide within the cell. Therefore, an intracellular accumulation of peptide beyond the extracellular concentration is expected. Second, HACBP was stepwise elongated to increase the potential binding surface with the SH3 domain.
As shown in Fig. 3A
, introduction of a disulfide bond
dramatically increased the potency of the Antp-HACBP (compare peptides
P8 and P12), arguing for a release of the cargo peptide in the cells
and thus the desired intracellular accumulation. Furthermore, peptide
affinities in vitro (see Table 1B
) and activity in cultured
cells (Fig. 3A
, P12 to P14) were further increased by
stepwise elongation of the flanking regions surrounding the core motif
contained in P2. The flanking sequences were derived from the regions
next to the ‘CB1’ motif in C3G (15)
. The rationale
behind this choice was that sequences from a naturally occurring Crk
and CRKL binding protein should be less likely to cause nonspecific
toxicity in cells and should not negatively interfere with the core
sequence. Elongation of the SH3 binding peptide from 11 amino acids to
28 amino acids resulted in a fourfold higher in vitro
affinity as determined by tryptophan fluorescence measurement (Table 1B
; compare peptides P2 and P13, P14) and a significant increase in
antiproliferative potency (Fig. 3A
). This suggests to us
that affinities may be improved even further by enlargement of the
peptide binding surface for the SH3 domain and continuing optimization
of the peptide sequence.
The disulfide bond-containing 4-site mutant control peptide (P15)
did not display significant nonspecific toxicity. Similar inhibition
results for P14 and P15 to those obtained with K562 cells were also
seen with the Bcr-Abl(p185)-positive cell line B15 (cells described in
ref 10
) and BV173 cells (results not shown).
Molecular events triggered by HACBPs
If the HACBPs function in cells as thought, a significant decrease
of Bcr-Abl–CRKL complexes should result. This was demonstrated by
incubating K562 cells with the different peptides, followed by the
detection of Bcr-Abl–CRKL complexes with coimmunoprecipitation and
Western blot (Fig. 4A
).
|
Control peptide (P15) had no effect on CRKL–Bcr–Abl complexes. A strong reduction of these complexes was seen with Antp-SS-HACBP (P14). Disruption of Bcr-Abl–CRKL complexes was also seen with the continuous Antp-HACBP (P8) at 100 µM (not shown).
Several lines of evidence document that the biological activity of HACBPs does not result from a lack of binding selectivity of HACBPs and therefore a simple ‘spillover’ effect onto the mitogenic cascade Grb2-SoS-Ras-Raf-MEK-MAPK.
The fluorescence measurements summarized in Table 1B
show that HACBPs
cannot strongly bind to the Grb2SH3(N) in vitro. To confirm
this with full-length proteins, SoS was precipitated from K562 extracts
using immobilized GST-Grb2 in the presence or absence of HACBP (P14) or
4-point mutant control peptide (P15). As shown in Fig. 4B
,
complex formation between SoS and Grb2 is not abolished through 10 µM
of Antp-SS-HACBP or control peptide, but completely prevented by a
mutant peptide with the crucial lysine changed to arginine (Antp-SS-R
mut), which switches the binding specificity from Crk/CRKLSH3(1) to the
Grb2SH3(N) domain (for details, see refs 7
8
9
). In
addition, coimmunoprecipitation experiments with peptide-treated K562
cells clearly documented that Antp-SS-HACBP does not affect Grb2–SoS
complexes, which are efficiently disrupted by the Grb2SH3(N)-specific
peptide Antp-SS-Rmut (not shown).
To gain insight into the molecular signals resulting from disruption of
Bcr-Abl–CRKL complexes, peptide effects on MAP kinases were
investigated. Using activation-specific phosphoMAPK antibodies, it was
observed that Antp-HACBP (P11) but not the control peptide (P8) led to
a strong reduction of phosphoMAPKs in K562 cells (Fig. 4C
).
It was also observed that a complete withdrawal of serum from the K562
culture medium for 3 days barely affected cell proliferation.
PhosphoMAPK was not significantly reduced in these totally
serum-deprived cells (Fig. 4D
). Therefore, we speculate that
Bcr-Abl allows K562 cells to proliferate well with little exogenous
growth factors and that adapter proteins like CRKL are used to keep
MAPKs in their active forms.
HACBP inhibits proliferation of primary CML blast cells from
patients
The results obtained with the human Bcr-Abl-positive CML cell
lines raised the question how these peptides affect the growth of CML
blast cells from patients. CML is characterized by a considerable
genetic heterogeneity during blast crisis that is mirrored by the
lineage diversity of circulating blast cells. About one-third of the
patients have blasts of lymphoid type whereas the rest have different
types of myeloid cells (4)
. Blast cells from patients were
isolated from peripheral white blood cells by density gradient
centrifugation and incubated with 20 µM HACBP (P14) or control
peptide (P15) ex vivo. The results are summarized in
Table 2
. Of 16 patients tested, cells from 11 patients showed a significant
inhibition of proliferation. All of the patients received or had
previously received one or more drugs and often represented a
relatively late stage of the disease in which additional mutations are
likely to exist in a significant number of cases. We conclude that
HACBP-type peptides and structurally related molecules may be a novel
option to interfere with the proliferation of CML blast cells.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Besides standard CML
therapeutics like interferon, which significantly prolong the overall
survival of some patients, several substances are currently in clinical
trials, including a modified form of interferon (peggylated IFN) and a
Bcr-Abl kinase inhibitor compound (CGP 57148B; recently renamed to STI
571), which were also used to treat some of the patients relevant to
our study. While such novel therapies are expected to lead to an
improved survival for some patients, there is clearly plenty of room
for novel therapeutic strategies.
CRKL was shown to be a major substrate for tyrosine phosphorylation by
Bcr-Abl in CML cells and to also form stable complexes with the Bcr-Abl
oncoprotein (reviewed in refs 2
, 6
). The functional
significance of CRKL in CML was not obvious, however. The data
presented here support the idea that CRKL is an important player in
aberrant signaling events triggered by Bcr-Abl in CML blast cells.
Previous studies have investigated the role of specific binding sites
for the CRKLSH3(1) domain on Bcr-Abl by creating Bcr-Abl deletion
mutants that abolish at least some of the proline-rich, direct binding
sites on Bcr-Abl (17
and references therein). As shown in
Table 1A
, four potential CRKLSH3(1) binding sites exist in Bcr-Abl, not
all of which have been studied yet (17)
. Furthermore, the
formation of indirect (ternary) complexes between CRKL and Bcr-Abl that
depend on the CRKLSH3(1) domain is certainly a possibility
(6)
. Direct and indirect complexes of CRKL and Bcr-Abl
that depend on the CRKLSH3(1) domain should be similarly affected by
the HACBPs used in this study whereas Bcr-Abl mutant proteins with
deleted P-x-x-P-x-K motifs will only abolish direct interactions.
In principle, ternary complexes could also occur via the CRKL SH2
domain since the hyperactive Bcr-Abl tyrosine kinase phosphorylates
many proteins in CML cells, including several large multisite docking
proteins (paxillin, c-Cbl, p130Cas-family proteins, etc. (reviewed in
ref 6
), that contain several CRKL SH2 domain binding sites
when massively tyrosine phosphorylated.
Since the CRKL–Bcr-Abl complexes are strongly decreased by the CRKLSH3(1) blocker peptides, our data argue against a prominent role of the CRKL SH2 domain in mediating CRKL–Bcr–Abl complexes.
The effects seen in cultured cell lines and freshly isolated patient
CML blasts are just a first, if promising, step in testing the idea of
a protein interaction blocker-based CML therapy. An obvious next step
is to use animal models for CML—for example, NOD/SCID mice with
xenografted human CML blast cells (18
19
20
21
22
and references
therein) or recently established mouse models created by infection of
mouse bone marrow with retroviruses or other approaches
(23
24
25
and references there)—for further tests. One
critical aspect that needs to be addressed in this context is the
peptide stability in vivo. From various experiments (not
shown here), we know that serum proteases are a major factor in the
degradation of Antp-coupled HACBPs in cultured cells and that the
Antennapedia portion of the disulfide-bonded Antp-HACBPs is strongly
affected. Fortunately, the analysis of proteolytic cleavage sites by
mass spectrometry is well established. Hence, it should be possible to
find out which residues are critical. The Antennapedia shuttle peptide
can be mutated in many positions without losing its biological activity
(11)
. This may result in novel shuttle sequences with
better stability. Other shuttle sequences can also be tested. To our
surprise, in initial experiments, an all D-amino acid Antp sequence
that is known to be functional (11)
was quite rapidly
degraded in serum (unpublished data).
From comparing inhibitor concentrations required to down-regulate other
signal transduction pathways in cultured cells vs. animal models, we
expect that the biological activity of the HACBPs still needs to be
improved 20- to 50-fold in order to accomplish major effects in animal
models with peptide concentrations below 100 µM. Besides an improved
stability, an increased CRKLSH3(1) binding affinity under retention of
the great selectivity is therefore highly desirable. Recent in
vitro studies suggest that major increases in binding affinity can
be accomplished by using non-natural derivatives of amino acids
(26)
or consolidated ligands that will interact with more
than one Src homology domain (27)
.
Besides a detailed analysis of downstream targets of the Bcr-Abl–CRKL complex, a major future goal is thus the generation of peptides and structural analogs with picomolar affinities and greatly improved stability.
What are the advantages of selecting CRKL rather than other
signaling proteins—for example, Grb2—as a protein interaction blocker
target for CML? Grb2 is obviously a prominent contributor to Ras
activation that subsequently leads to MAP kinase (Erk) activation and,
finally, to cell proliferation. As a component of the central mitogenic
signaling cascade, it is expected to inhibit the proliferation of
nearly all cells. Experiments performed by us (unpublished data) and
others (28
, 29)
indeed show that Grb2 blocker peptides
inhibit the proliferation of many cell types. It is therefore at least
questionable whether Grb2-blocking molecules will be very useful as
clinical therapeutics, unless they are used in combination with other
more selective drugs. By contrast, Crk and CRKL have not yet been shown
to play a crucial role in normal cell proliferation. The highly
specific Crk/CRKL SH3(1) domain blocker peptides may therefore
preferentially affect tumor cells where CRKL has acquired a
proliferation-regulating function.
An interesting strategy to improve the effectiveness of drugs and to
reduce undesired side effects by using short targeting peptides was
recently reported by the Ruoslahti group (30
31
32
33
34)
. These
peptides are selected by phage display in animals to find sequences
that preferentially bind in a specific organ or to a specific cell
type. By coupling such a sequence to a chemotherapeutic compound, its
biological activity can be significantly increased (34)
.
Such a strategy could also be useful in order to enhance the binding of
HACBPs to blast cells and to reduce the amount of peptide lost in
erythrocytes, endothelial cells, etc.
Beyond Bcr-Abl-induced signaling, other potential therapeutic targets
for Crk/CRKLSH3(1) blockers are currently surfacing. c-Crk binds to a
mutated form of the Ret receptor tyrosine kinase involved in certain
cancers (multiple endocrine neoplasia type 2B; 35
). Work
from our laboratory indicates that bombesin, an autocrine growth factor
produced by small cell lung cancer cells, also depends on Crk/CRKL for
signal transmission in certain cell types (unpublished results).
Additional therapeutic targets for CRKLSH3(1) specific inhibitors may
surface as our knowledge of the molecular signaling cascades involving
Crk family adapters increases.
|
ACKNOWLEDGMENTS |
|---|
Received for publication October 13, 1999.
Revision received January 3, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. I. Lopez, S. A. Belmonte, G. A. De Blas, and L. S. Mayorga Membrane-permeant Rab3A triggers acrosomal exocytosis in living human sperm FASEB J, December 1, 2007; 21(14): 4121 - 4130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Sitko, C. I. Guevara, and N. A. Cacalano Tyrosine-phosphorylated SOCS3 Interacts with the Nck and Crk-L Adapter Proteins and Regulates Nck Activation J. Biol. Chem., September 3, 2004; 279(36): 37662 - 37669. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. V. Melo, T. P. Hughes, and J. F. Apperley Chronic Myeloid Leukemia Hematology, January 1, 2003; 2003(1): 132 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nishihara, M. Maeda, A. Oda, M. Tsuda, H. Sawa, K. Nagashima, and S. Tanaka DOCK2 associates with CrkL and regulates Rac1 in human leukemia cell lines Blood, December 1, 2002; 100(12): 3968 - 3974. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Hemmeryckx, A. van Wijk, A. Reichert, V. Kaartinen, Ron de Jong, P. K. Pattengale, I. Gonzalez-Gomez, J. Groffen, and N. Heisterkamp Crkl Enhances Leukemogenesis in BCR/ABL P190 Transgenic Mice Cancer Res., February 1, 2001; 61(4): 1398 - 1405. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
