| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2
* Cardiovascular Research Group, Department of Research, and
Medical University Outpatient Division, University Hospital, 4031 Basel, Switzerland; and
§ Nervous System Therapeutic Area, Novartis, 4002 Basel, Switzerland
2Correspondence: Head Laboratory of Vascular Biology and Hypertension Clinic, Department of Research and Medical Outpatient Division, University Hospital, Petersgraben 3, CH-4031 Basel, Switzerland. E-mail: ebattegay{at}uhbs.ch
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: receptor for activated PKCß • signaling • angiogenic process • endothelium • cancer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5)
. To identify genes that are
specifically up- or down-regulated in angiogenesis and thus may play a
role in this process, we applied mRNA fingerprinting (differential
display) to in vitro and in vivo models of
angiogenesis and the corresponding quiescent endothelial cells and
tissues. The model of angiogenesis in vitro consists of
specific strains of cloned bovine aortic endothelial cells that
spontaneously form capillary-like cords without addition of exogenous
matrix substances (4
, 6
7
8
9
10)
. In contrast, other strains
of cloned bovine aortic endothelial cells form monolayers only
(4
, 6
7
8
9
10)
. These different strains thus represent
different phenotypes of endothelium with a global shift in expressed
genes. We and others have previously used this in vitro
model to demonstrate that cord-forming endothelial cells show distinct
responses to platelet-derived growth factor BB (PDGF-BB) (8
, 11)
, Transforming growth factor ß (10)
, and
express specific angiogenesis-related patterns of extracellular
matrices (4
, 7
, 12)
. To ensure that in vitro
findings correlate with angiogenesis in vivo
(13)
, we simultaneously applied mRNA fingerprinting to
bovine corpora lutea classified to the early, mid, and late stages of
the ovarian cycle. These stages represent active and regressive
angiogenesis and served as an in vivo model of angiogenesis
(5
, 14
15
16)
.
Protein kinase C ß (PKCß) plays an important role in angiogenesis
(17
, 18)
. Here we report the isolation of the bovine
receptor for PKCß, i.e., the receptor for activated C kinase 1
(RACK1), which we found to be up-regulated in active angiogenesis.
Bovine RACK1 cDNA was completely cloned and its nucleotide sequence
determined. Expression studies revealed that RACK1 was distinctly
up-regulated in angiogenesis in vitro and in
vivo.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 6
7
8
9
10
, 12)
.
Fresh bovine ovaries with corpora lutea were obtained at the local
slaughterhouse. Pregnancy was excluded by macroscopic examination of
uterine horns. Corpora lutea were dissected from the cortical tissue
and classified by macroscopic appearance to three estrous cycle stages:
early (days 2–4 of development), mid (days 5–17 of secretion), and
late (days 18–21 of regression) (14
, 15)
.
Mouse ovaries were staged by analysis of vaginal smears
(19)
and bought from BRL (Füllinsdorf, Switzerland).
Human non-small cell lung, colon, and breast carcinomas were obtained from patients at the University Hospital, Basel, Switzerland, in accordance with regulations of the hospital ethics committee. All tissues were frozen in liquid nitrogen immediately after dissection and stored at -80°C.
Cell culture
Clones of BAEC were grown on uncoated plastic dishes (TPP,
Trasadingen, Switzerland) in Dulbecco’s modified Eagle’s medium
(Seromed, Berlin, Germany) supplemented with 10% (v/v) fetal bovine
serum (Seromed), antibiotics-antimycotics (GibcoBRL, Merelbeke,
Belgium), nonessential amino acids (Seromed), and 1 mM sodium pyruvate
(Seromed) in a 37°C humidified atmosphere of 5%
CO2. At 90% confluence, cells were trypsinized
(Seromed) and split at a ratio of 1:3 until they were seeded on 15 cm
dishes. Cells of the monolayer-forming BAEC clone were grown until
contact inhibition and confluence as described before (7)
.
Cells of cord-forming BAEC clones were grown 2 wk beyond confluence
until a dense network of cord-like structures formed on top of the
monolayers as described before (7)
.
mRNA fingerprinting and differential display
Differential display of messenger RNA by means of polymerase
chain reaction (PCR) was carried out as described (20)
with (dT)12CG as reverse anchored primer. Total
RNA of tissues was extracted as described (21)
. RNA
quality was checked on a denaturing formaldehyde-agarose gel (Bio-Rad,
Glattbrugg, Switzerland). First strand cDNA synthesis was carried out
using the SuperScript Preamplification System (GibcoBRL), 5 µg total
RNA, RNAsin (Promega, Madison, Wis.), 200 µM dNTPs, and 3.3 µmol
(dT)12CG (Intron Laboratory for Molecular
Biology, Kaltbrunn, Switzerland). Generation and radioactive labeling
of differently sized cDNA 3'-ends was facilitated by PCR and use of 1
µl first strand cDNA reaction, 2.5 µCi
[
-32P]dCTP (Amersham Pharmacia Biotech,
Dübendorf, Switzerland), 2 µM dNTPs (Boehringer, Mannheim,
Germany), 1.5 µM (dT)12CG, and 1.5 µM of 24
different decanucleotide forward primers (Intron) selected from a list
of random sequences and tested for absence of self-complementarity of
more than two nucleotides (22)
. Amplification conditions
were 94°C for 5 min, 40 cycles of 94°C for 1 min, 40°C for 1 min,
and 72°C for 1 min, with the elongation time at the last cycle
extended to 8 min. Two microliters PCR reaction per lane was
fractionated on a denaturing 6% polyacrylamide gel
(acrylamide-bisacrylamide ready-mix, GibcoBRL) containing 50% (w/v)
urea (Sigma, Buchs, Switzerland) and 1 x TBE. A gel was run at
1600 V for 1.5 h or 3 h and exposed to a MR-1 BioMax film
(Kodak, Rochester, N.Y.) overnight.
Full-length cDNA cloning
Gel pieces corresponding to bands of interest were excised and
cast into an agarose gel (Bio-Rad). cDNAs were transferred to activated
DEAE cellulose membranes (Schleicher & Schuell, Feldbach, Switzerland)
and eluted. Reamplification was carried out by PCR using 10 µl cDNA
dialysates, 20 µM dNTPs, and the identical primers and amplification
conditions of the differential PCR. Amplification products were
separated on a 2% agarose gel (Bio-Rad), and DNA fragments of expected
size were purified using the QIAquick Gel Extraction Kit (Qiagen,
Hilden, Germany). Purified cDNA 3'-ends were ligated into the pGEM-T
Vector (Promega) and transformed into Escherichia coli
strain JM109 (Promega). Bacteria were selected on LB agar plates
containing ampicillin (Sigma), IPTG (Promega), and X-Gal (Promega).
White colonies were amplified for plasmid preparations using the
QIAprep Spin Plasmid Kit (Qiagen). Double-strand sequencing was
performed using the Sequenase DNA Sequencing Kit (Amersham) and
[-35S]dATP (Amersham). Reaction products
were separated on a denaturing 6% polyacrylamide gel
(acrylamide-bisacrylamide ready-mix, GibcoBRL) containing 50% (w/v)
urea (Sigma) and 0.5 x TBE. A gel was run at 70 W for 2 h or
3 h and exposed to a MR-1 BioMax film (Kodak) overnight. Sequences
were determined and matched against the DDBJ/EMBL/GenBank Nucleotide
Sequence Database with the homology search program fastA.
The RACK1 cDNA 3'-clone was selected for full-length cloning using the
5' RACE System for Rapid Amplification of cDNA Ends (GibcoBRL)
(23)
, 5 µg total RNA of cord-forming BAEC, and a RACK1
specific reverse primer (Microsynth, Balgach, Switzerland). The 5' RACE
reaction product was ligated into the pGEM-T Vector, sequenced, and
identified as RACK1 cDNA 5'-clone. Clones were combined and both
full-length cDNA strands were sequenced (Microsynth).
Northern blot analysis
A RACK1 specific DIG-labeled antisense riboprobe was generated
by in vitro transcription and use of 1 µg linearized cDNA
3'-clone, the RiboMax RNA Production System (Promega), 7.5 mM rNTPs,
and 750 µM DIG-UTP (Boehringer). The probe was purified by gel
filtration on a CentriSpin column (Princeton Separations, Adelphia,
N.J.). Preparation of magnetic oligo (dT) particles has been described
(24)
. Poly (A)+ RNAs from tissues
were purified using magnetic oligo (dT) particles as described
(25)
. Two micrograms of photometrically determined poly
(A)+ RNAs was glyoxalated in glyoxal solution
(Fluka, Buchs, Switzerland) previously deionized with Resin AG
501-X8(D) (Bio-Rad), fractionated on a 1.2% agarose gel containing 10
mM sodium phosphate, transferred to an Electran nylon membrane (BDH,
Poole, England), and cross-linked. Blots were hybridized with 200 ng
RACK1 antisense riboprobe per ml DIG-EasyHyb hybridization solution
(Boehringer) at 68°C overnight. 1.1 kb RACK1 transcripts were
detected by chemoluminescence using DIG Wash & Block Buffer Set
(Boehringer), alkaline phosphatase-conjugated anti-DIG
Fab fragments (Boehringer), and CDP-Star
(Promega). Blots were exposed to a MR-1 BioMax film (Kodak) for 10 min.
In situ hybridization
Immediately after resection, tissue samples were transferred
into Tissue-Tek Cryomolds (Miles, Elkhart, Ind.), embedded with
Tissue-Tek O.C.T. Compound (Sakura Finetek, Torrance, Calif.), and
floated on isopentane cooled by solid CO2. Twelve
micrometer slices were cut on a cryostat and transferred to silanized
slides. Cells of BAEC clones were grown on culture slides (Nalge Nunc,
Naperville, Ill.) under the conditions described (Cell culture
section). Tissue sections and BAEC cultures were fixed in 4%
paraformaldehyde (Merck, Dietikon, Switzerland) and acetylated in
0.25% acetic anhydride (Fluka). For hybridization, the RACK1 antisense
riboprobe was used (Northern blot analysis section) and a RACK1 sense
riboprobe as negative control. Aliquots of dilution series of sense and
antisense riboprobes were spotted on a nylon membrane, cross-linked,
and detected as described (see Northern blot Analysis) to serve for
normalization of different anti-DIG Fab fragment
binding affinities. Each slide was hybridized with normalized amounts
of sense or antisense riboprobes in presence of 50% [w/v] formamide
(Fluka) at 72°C overnight. Slides were processed using DIG Wash &
Block Buffer Set (Boehringer), alkaline phosphatase-conjugated anti-DIG
Fab fragments (Boehringer), and a substrate
buffer containing NBT (Boehringer), BCIP (Sigma), and levamisole
(Sigma). Development was carried out in the dark until a color reaction
could be specifically observed on slides hybridized with the antisense
riboprobe. Fluorescence counter staining of human endothelial cells was
performed with FITC-conjugated lectin from Ulex europaeus
(Sigma) (26)
and of nuclei with Hoechst 33342 stain
(Sigma). Slides were mounted with Aquamount Mountant (BDH) and
photographed (Ektachrome 400x film, Kodak).
Quantitative PCR
Poly (A)+ or total RNAs from tissues were
purified as described (25
, 21)
. First strand cDNA
synthesis was carried out using the SuperScript Preamplification System
(GibcoBRL), either 275 ng poly (A)+ RNA or 1–5
µg total RNA, 500 µM dNTPs, and 25 ng/µl
oligo(dT)12–18 (GibcoBRL). Real time
quantitative PCR was performed on an Abi Prism 7700 machine and
software (Perkin Elmer Biosystems, Foster City, Calif.). For each
species, an optimal RACK1 amplicon, spanning the TaqMan probe (5'-FAM
and 3'-TAMRA modified, Perkin Elmer, and Eurogentec, Seraing, Belgium)
and two flanking primers (Microsynth), was identified by matching the
criteria of the Primer Express software (Perkin Elmer). In 25 µl
reactions, 12.5 µl master mix (Perkin Elmer), 8 pmol forward primer,
8 pmol reverse primer, and 16 pmol TaqMan probe were used. 40 cycle
PCRs were performed in optical 96-well plates (Perkin Elmer) under
standard conditions with threshold cycles Ct set
automatically. For absolute quantifications, a dilution series of
101 to 106 RACK1 plasmid
copies served as standard curve.
Immunoblot analysis
Cells were lysed in 1x SDS buffer (0.28 M Tris-Cl, pH 6.8, 45%
glycerin, 0.1 M SDS) supplemented with a protease inhibitor mixture
(Roche Molecular Biochemicals, Rotkreuz, Switzerland). Protein load was
equalized by quantifying protein with the BCA Protein Assay Reagent
(Pierce, Rockford, Ill.). Equal amounts of protein per lane were
subjected to SDS-gel electrophoresis and electroblotting. Protein load
was visualized by Amido Black staining (Sigma, Buchs, Switzerland). For
detection, monoclonal IgM against RACK1 (Transduction Laboratories, San
Diego, Calif.) and HRPO-conjugated IgG (Transduction Laboratories) were
used to visualize RACK1 protein on BIOMAX films (Kodak) by a
chemiluminescence reaction.
Statistical analysis
Differences between groups were analyzed with Mann-Whitney rank
sum test or Kruskal-Wallis one way ANOVA on ranks. Results were
expressed as box plots and considered statistically significant at
*P < 0.05 and **P < 0.01.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
to bovine in
vitro and in vivo models of angiogenesis. Gene
expression was analyzed by differential display in a confluent,
contact-inhibited monolayer-forming clone of BAEC and in a cord-forming
BAEC clone at subconfluent, confluent, and cord-forming stages.
Expression of a 357 bp cDNA, later identified as RACK1 (see below), was
observed to increase with progressive stages of endothelial cord
formation in cord-forming BAEC, whereas only limited expression was
detected in the monolayer-forming BAEC (Fig. 1a
). In parallel, differential gene expression was displayed
in bovine corpora lutea of the early, mid, and late stages of the
ovarian cycle that are associated with progressive, active, and
regressive stages of angiogenesis in vivo. A moderate and
strong expression of the same cDNA was present in corpora lutea of the
early and mid stages, respectively, but no expression was found in the
late-stage corpora lutea (Fig. 1a
). A close look at the
autoradiographs revealed three bands that seem to be regulated in
endothelial cells (Fig. 1a
, left panel, second to fourth
lane). Restriction analysis of six clones and sequencing of one clone
of the three bands all proved to consist of RACK1 cDNA (see below).
Thus, the three bands might be due to hybridizing of the 10 mer
downstream primer with mismatches.
|
Northern blot analysis of the above-mentioned cells or tissues
performed with a probe derived from the reamplified 357 bp cDNA
confirmed differential expression of the cDNA. Angiogenically active
tissues and cells proved to express substantially more of a 1.1
kb-sized mRNA than inactive cells or tissues both in vitro
and in vivo (Fig. 1b
).
To ascertain up-regulation of RACK1 during angiogenesis in
vitro, immunoblot analysis of RACK1 was performed in sequential
stages of endothelial cord development (Fig. 1c
). Similar to
RACK1 mRNA expression (Fig. 1a
, left panel), RACK1 protein
expression was highest in confluent and cord-forming stages of
cord-forming BAEC as compared to subconfluent cord-forming BAEC (Fig. 1c
).
These experiments suggested a potential role for this cDNA in angiogenesis. We therefore proceeded to establish the full-length cDNA sequence and to identify this cDNA.
Bovine RACK1 full-length cDNA cloning
To identify the differentially expressed cDNA, fragments were
extracted from the gel, reamplified, cloned, and sequenced. The clone
proved to consist of homogenous DNA and had a length of 357 bp.
Homology search identified the clone to be highly homologous to the
3'-end of RACK1 cDNA, and to include its stop codon and polyadenylation
signal. This clone served as a template for full-length cDNA cloning
using 5' RACE, as outlined in Materials and Methods. The derived
full-length cDNA sequence from bovine tissue was determined on a length
of 1091 nucleotides (Fig. 2
).
|
The bovine RACK1 cDNA has an open-reading frame of 951 nucleotides,
coding for 317 amino acids. The coding region starts with the ATG start
codon at nucleotide position 99 and ends with the TAA stop codon at
nucleotide position 1050. The sequence of the isolated full-length cDNA
clone was found to be 99% homologous from nucleotides 4 to 1091 to rat
RACK1 cDNA (accession no. U03390) (27)
, which is shown for
comparison (Fig. 2)
. There are eight mismatches within the 5'- and
3'-noncoding regions with a common length of 137 nucleotides, and three
mismatches within the coding region with a length of 951 nucleotides,
resulting in a 19-fold higher mismatch incidence in the noncoding
regions. From these data, we concluded that we have isolated bovine
RACK1 cDNA.
The bovine RACK1 cDNA shows also a high nucleotide sequence
identity of 95% to mouse (accession no. D29802), 89% to human
(accession no. M24194), and 82% to chicken (accession no. M24193)
RACK1 cDNAs. The corresponding amino acid sequence identity is 100% to
RACK1 of rat (accession no. A36986), mouse (accession no. I49700),
human (accession no. P25388), and chicken (accession no. P25388) (not
shown). These high degrees of homology imply an extremely high degree
of conservation, which suggests that RACK1 may be essential for
cellular functions (27)
, possibly also in angiogenesis and
endothelial proliferation.
RACK1 is up-regulated in subconfluent EC and angiogenesis in
vitro
To investigate expression of RACK1 during endothelial
proliferation, one of the most prominent aspects of angiogenesis
(1
, 2)
, and angiogenesis in vitro, we assessed
RACK1 mRNA expression by in situ hybridization and
quantitative PCR. For this purpose, we again compared monolayer-forming
and cord-forming BAEC clones during subconfluence and contact
inhibition. RACK1 was ubiquitously expressed in cord-forming BAEC
(Fig. 3c
, subconfluent, and Fig. 3d
, confluent,
continuous growth; arrows, brown staining). Also, strong RACK1 mRNA
expression was detected in subconfluent cultures of monolayer-forming
BAEC (Fig. 3a
, b
, arrows, brown staining).
|
Quantitative PCR was performed to better assess the changes occurring
in contact-inhibited vs. proliferating monolayer-forming BAEC and in
progressive stages of cord formation in cord-forming BAEC (Fig. 4
). During confluence, i.e., growth arrest, there was very little RACK1
expression in monolayer-forming BAEC. A 34-fold induction of RACK1 was
observed in subconfluent, proliferating BAEC compared to
growth-arrested monolayer-forming BAEC (Fig. 4)
.
|
In cord-forming BAEC, RACK1 expression increased dramatically from
subconfluence to stages where cords began forming (77,000-fold in
confluent in comparison to subconfluent cord-forming endothelial cells)
or formed (41,000,000-fold in cord-forming in comparison to
subconfluent cord-forming endothelial cells) (Fig. 4)
. This latter
increase of RACK1 expression during endothelial cord formation vastly
exceeds the changes seen in rates of cell proliferation and does not
necessarily reflect the rate of DNA synthesis (8)
. Indeed,
DNA synthesis in response to fetal calf serum, the conditions under
which the cell cultures were harvested in this study, decreases
continuously as the cells that differentiate into cords
(8)
. In contrast, the response of cells localized in cords
to PDGF-BB increases as the cords form (8)
. The massive
increase in RACK1 expression during endothelial cord formation, also in
comparison to proliferating monolayer-forming EC, therefore indicated
an innate association of RACK1 expression with the changing phenotype
of endothelial cells. However, increased expression was not specific
for the cord-forming phenotype of endothelial cells because increased
RACK1 expression could also be seen in subconfluent vs. confluent
monolayer-forming endothelial cells (Fig. 4)
. The difference of RACK1
expression seen between subconfluent monolayer- vs. cord-forming clones
in quantitative PCR (Fig. 4)
was not detectable by in situ
hybridization (Fig. 3a
vs. 3c
).
Taken together, the data suggest that RACK1 can be associated with endothelial cord formation, i.e., angiogenesis in vitro.
RACK1 is expressed in cancer vessels and human carcinomas in
vivo
Next we asked whether RACK1 might also be expressed in
angiogenesis in vivo, e.g., in cancer. To assess RACK1 mRNA
in different cell populations of human cancers in vivo, we
performed RACK1 in situ hybridizations on human non-small
cell lung and colon carcinomas. Endothelia (arrows) of a human
non-small cell lung carcinoma (Fig. 5a
, c
) and human normal colon tissue (Fig. 5e
) were
specifically stained by fluorochrome-labeled Ulex europaeus
agglutinin (yellow staining). RACK1 in situ hybridizations
of adjacent slices were performed. The specific in situ
signal is brownish representing weak expression and bluish reflecting a
strong expression. Expression of RACK1 mRNA in the tumor endothelium of
large (Fig. 5d
; arrow, blue staining) and small vessels
(Fig. 5b
; arrow, blue staining) was detected. In contrast to
large tumor vessels, expression of RACK1 mRNA seemed to be patchy in
small vessels of non-small cell lung carcinomas, with some endothelial
cells not staining at all (data not shown). In comparison to the
endothelium of a normal colon mucosa vessel (Fig. 5f
; arrow,
brown staining), the RACK1 signal was stronger in the large vessel of a
non-small cell lung carcinoma (Fig. 5d
; arrow, blue
staining). The difference in the strength of signals between control
and tumor vessels was not obvious in small tumor vessels. A visible
difference in the strength of the in situ signal was only
present in the large vessels of the specific tumor samples investigated
(Fig. 5d
) vs. the normal control vessel (Fig. 5f
).
|
To better assess differences in RACK1 mRNA levels in human carcinomas
in vivo, quantitative PCR was performed in human non-small
cell lung, colon, and breast carcinomas in comparison to the patients’
normal tissues. Compared with the patients’ normal lung and colon
tissues, RACK1 mRNA levels were found to be up-regulated 4.4-fold in
human non-small cell lung carcinoma (Fig. 6a
) and 18.3-fold in human colon carcinoma (Fig. 6b
). RACK1 mRNA levels tended to be elevated in human breast
carcinoma as compared to levels in the patients’ normal breast tissues
(Fig. 6c
), but to a lesser extent and without attaining
statistical significance. From these observations we concluded that
RACK1 is more strongly expressed in some cancers, including their
vasculature, as compared with control tissues.
|
RACK1 is also associated with nonendothelial cells in
angiogenically active tissues
We next asked whether RACK1 mRNA expression was exclusively
localized to vessels in angiogenically active tissues in
vivo. RACK1 in situ hybridizations were performed on
slices of staged mouse ovaries representing the proestrus, estrus, and
diestrus stages of the ovarian cycle. In this model, the estrus stage
is known to be associated with high endothelial and nonendothelial
rates of proliferation as compared to other stages of the ovarian
cycle.
Strongest RACK1 mRNA expression was observed in follicles and
follicular cells of the estrus stage (Fig. 5g
; arrow, dark
blue staining) as compared to other stages of the ovarian cycle (data
not shown). These observations are consistent with our findings in
bovine corpora lutea, and confirm that RACK1 expression occurs in
active phases of the ovarian cycle that are known to be associated with
angiogenesis in vivo. However, these experiments also show
that RACK1 expression is associated with proliferation of
nonendothelial cells such as follicular cells.
Similarly, substantial RACK1 mRNA expression was detected in epithelial
cells of a human colon carcinoma (Fig. 5h
; arrow, brown
staining) and in proliferating epithelial cells of normal colon mucosa
(data not shown).
These data suggest that RACK1 expression is not only associated with active angiogenesis, but also with the activity of processes associated with angiogenesis.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. We
have performed mRNA fingerprinting to search for molecules up- or
down-regulated in angiogenesis and found a massive up-regulation of the
receptor for activated C kinase 1 (RACK1) in active angiogenesis
in vivo and in vitro. However, expression of
RACK1 was not particular to angiogenesis. Thus, we found RACK1 to be
expressed in mouse follicles and human carcinoma cells in
vivo, in normal human colon epithelial cells, and in endothelial
cells not engaged in angiogenesis, albeit at lower levels.
RACK1 is a pivotal molecule directly associated with the signal
transduction pathway of protein kinase C (PKC). Specifically, PKCß,
the ligand of RACK1, is known to play an important role in angiogenesis
(17
, 18)
. Because PKCß requires RACK1 for signaling, our
findings suggest the possibility of a specific and effective control
for PKCß activities at the level of RACK1 expression, i.e., a control
of angiogenesis associated endothelial processes at the level of the
receptor for PKCß, RACK1.
Our data clearly demonstrate that RACK1 is up-regulated in different angiogenically active tissues in vitro and in vivo. In the monolayer-forming clone of bovine aortic endothelial cells (BAEC), a low RACK1 expression is associated with contact inhibition. In cord-forming BAEC, RACK1 expression increases dramatically during sequential stages of cord-formation, i.e., during angiogenesis in vitro. To corroborate these findings in tissues in vivo, we investigated RACK1 expression in tumor angiogenesis in vivo. RACK1 was up-regulated during tumor angiogenesis. Overall, levels of RACK1 expression were much higher in human non-small cell lung and colon carcinomas than in the corresponding normal tissues. Expression of RACK1 was not limited to the angiogenically active endothelium. It was also observed in colon cancer cells, normal colon mucosa, and ovarian follicles. Normal vessels and subconfluent monolayer-forming endothelial cells also expressed RACK1 mRNA, although to lesser levels than large tumor vessels or advanced stages of cord-forming endothelial cells.
Taken together, the up-regulation of RACK1 in angiogenesis and in tissues associated with angiogenesis suggests a potential role for RACK1 in angiogenesis and processes associated with angiogenesis such as cancer. In view of the current knowledge of PKCß (see below), these results implicate that RACK1 may be important for the maintenance of tumor angiogenesis and tumor growth.
To date, the PKC family consists of 12 different isozymes
(28)
. RACKs bind to activated PKC isozymes via a
nonsubstrate binding site in a saturable and specific manner
(29)
. Each PKC isozyme is hypothesized to bind to its
specific RACK upon activation. Thus, RACKs are activation-induced
translocation targets for PKC isozymes and are located at various
intracellular sites including plasma membranes, intra- and perinuclear
structures, and cytoskeletal elements (29)
. Hence, a given
RACK would selectively direct its specific activated PKC isozyme to a
subcellular compartment. RACKs may thus contribute to substrate
specificity and selective function of its PKC isozyme
(30)
. RACK1 has been cloned as the first protein
displaying the characteristics of RACKs and is a homologue of the
ß-subunit of heterotrimeric G-proteins (27)
. The
structure of RACK1 is composed of seven internal repeats forming a
circular, propeller-like structure with seven blades (31
, 32)
. It belongs to the WD-repeat adaptor protein family known to
regulate diverse cellular functions by simultaneously interacting with
different classes of proteins (33)
. RACK1 is likely to be
essential for cellular functions because its amino acid sequence is
100% identical in human, rat, chicken (27)
, mouse
(34)
, and (as we have shown here) cow.
RACK1 has been found to colocalize with activated PKCßI and PKCßII,
suggesting specific binding of RACK1 to PKCß (35)
, and
was recently characterized as a shuttling protein that moves PKCßII
from one intracellular site to another (36)
. PKC is a key
enzyme involved in signal transduction for diverse cellular functions
including growth, differentiation, and gene expression
(37)
, and has also been shown to play an important role in
angiogenesis. The specific sites where activated PKCß resides either
before or after being shuttled may define a specific substrate to be
phosporylated, and thus a specific cellular feature to be induced.
Hypothetically, increased shuttling of PKCßII by up-regulation of
RACK1 may result in increased phosphorylation of a near-by substrate
that signals proliferation and angiogenesis and/or in decreased
phosphorylation of a substrate that signals growth arrest or apoptosis.
Thus, in addition to regulation of PKCß expression and activation,
alterations in availability of RACK1 might be another important level
at which PKCß signaling is controlled in angiogenesis and
carcinogenesis.
PKC-activating phorbol esters induce angiogenesis in vitro
and in vivo (38
, 39)
, and angiogenesis induced
by cytokines or ethanol is PKC dependent (40
, 41)
. Signal
transduction of VEGF, an important angiogenic molecule, has recently
been shown to be PKC dependent in endothelial cells (42)
.
Specifically, the PKCßI and PKCßII isozymes, derived from
alternative splicing of the same mRNA (43)
, have been
shown to be involved in angiogenesis. PKCßI is one of two predominant
PKC isozymes known to be present in BAEC; overexpression of PKCßI in
these cells promotes proliferation (44)
, a process crucial
for angiogenesis. Activated PKCßII translocates to different
subcellular compartments of human microvascular endothelial cells on
either phorbol ester vs. collagen treatment (17)
. This
difference in localization of PKCßII after stimulation may be of
functional relevance, i.e., explain why collagen treatment induces
endothelial tube formation, whereas phorbol ester treatment induces a
conversion of endothelial shape (17)
. Obviously, both
processes are necessary for angiogenesis (17
, 1
, 2)
. Also,
VEGF-induced tumor angiogenesis and tumor growth in vivo
have recently been shown to be PKCß dependent (18)
;
inhibition of PKCß significantly suppressed VEGF-induced
neovacularizations in a mouse model of hepatocellular carcinoma cells
(18)
. PKCß inhibition reduced the size and number of
VEGF-induced tumors both at initial stages of tumor development and
after the tumors had attained a large size (18)
. Taken
together, these findings suggest that PKCßI and PKCßII, the
specific ligands for RACK1, are crucial for angiogenesis.
Our findings support the importance of PKCßI and PKCßII in angiogenesis and carcinogenesis, and suggest that their signaling may also be regulated at the level of their specific receptor RACK1. We speculate that the availability of RACK1 may be rate-limiting for the activities of PKCß whereby the amount of activated PKCßII shuttled from one intracellular site to another may depend on the amount of RACK1 available. This speculation assumes that substantial expression of RACK1 is required for one or many RACK1-mediated processes to occur and that lower levels of RACK1 expression, which we observed in angiogenically inactive tissues or endothelial cells, are insufficient to mediate specific effects. It therefore remains to be clarified whether the differences in expression we observed translate into functional differences.
In conclusion, RACK1 was up-regulated during angiogenesis in vitro and in vivo, and was also expressed in tumor angiogenesis. RACK1 expression was higher in human non-small cell lung and colon carcinomas than in the corresponding normal tissues. RACK1 was also expressed in non-angiogenically active endothelia, although at lower levels, and nonendothelial components of angiogenically active or inactive tissues, i.e., ovarian follicles and colonic mucosa. In conjunction with data of other groups on PKCß, these results suggest that RACK1 might contribute to angiogenesis and tumor growth. Future studies will address the mechanism of RACK1 gene regulation and the functional role of RACK1 in angiogenesis.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication December 13, 1999.
Revision received May 26, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Kraus, D. Gioeli, T. Vomastek, V. Gordon, and M. J. Weber Receptor for Activated C Kinase 1 (RACK1) and Src Regulate the Tyrosine Phosphorylation and Function of the Androgen Receptor. Cancer Res., November 15, 2006; 66(22): 11047 - 11054. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Kiely, D. O'Gorman, K. Luong, D. Ron, and R. O'Connor Insulin-Like Growth Factor I Controls a Mutually Exclusive Association of RACK1 with Protein Phosphatase 2A and {beta}1 Integrin To Promote Cell Migration. Mol. Cell. Biol., June 1, 2006; 26(11): 4041 - 4051. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Zhang, C. S. Zong, U. Hermanto, P. Lopez-Bergami, Z. Ronai, and L.-H. Wang RACK1 Recruits STAT3 Specifically to Insulin and Insulin-Like Growth Factor 1 Receptors for Activation, Which Is Important for Regulating Anchorage-Independent Growth Mol. Cell. Biol., January 15, 2006; 26(2): 413 - 424. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mourtada-Maarabouni, L. Kirkham, F. Farzaneh, and G. T. Williams Functional expression cloning reveals a central role for the receptor for activated protein kinase C 1 (RACK1) in T cell apoptosis J. Leukoc. Biol., August 1, 2005; 78(2): 503 - 514. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Kiely, M. Leahy, D. O'Gorman, and R. O'Connor RACK1-mediated Integration of Adhesion and Insulin-like Growth Factor I (IGF-I) Signaling and Cell Migration Are Defective in Cells Expressing an IGF-I Receptor Mutated at Tyrosines 1250 and 1251 J. Biol. Chem., March 4, 2005; 280(9): 7624 - 7633. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. C. Nery, D. O. Passos, V. S. Garcia, and J. Kobarg Ki-1/57 Interacts with RACK1 and Is a Substrate for the Phosphorylation by Phorbol 12-Myristate 13-Acetate-activated Protein Kinase C J. Biol. Chem., March 19, 2004; 279(12): 11444 - 11455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Shor, J. Calaycay, J. Rushbrook, and M. McLeod Cpc2/RACK1 Is a Ribosome-associated Protein That Promotes Efficient Translation in Schizosaccharomyces pombe J. Biol. Chem., December 5, 2003; 278(49): 49119 - 49128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D.-S. Choi, H. Young, T. McMahon, D. Wang, and R. O. Messing The Mouse RACK1 Gene Is Regulated by Nuclear Factor-{kappa}B and Contributes to Cell Survival Mol. Pharmacol., December 1, 2003; 64(6): 1541 - 1548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Cox, D. Bennin, A. T. Doan, T. O'Toole, and A. Huttenlocher RACK1 Regulates Integrin-mediated Adhesion, Protrusion, and Chemotactic Cell Migration via Its Src-binding Site Mol. Biol. Cell, February 1, 2003; 14(2): 658 - 669. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. McCahill, J. Warwicker, G. B. Bolger, M. D. Houslay, and S. J. Yarwood The RACK1 Scaffold Protein: A Dynamic Cog in Cell Response Mechanisms Mol. Pharmacol., December 1, 2002; 62(6): 1261 - 1273. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Wallqvist, A. A. Rabow, R. H. Shoemaker, E. A. Sausville, and D. G. Covell Establishing Connections between Microarray Expression Data and Chemotherapeutic Cancer Pharmacology Mol. Cancer Ther., March 1, 2002; 1(5): 311 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Saito, G Fujii, Y Sato, M Gotoh, M Sakamoto, G Toda, and S Hirohashi Detection of genes expressed in primary colon cancers by in situ hybridisation: overexpression of RACK 1 Mol. Pathol., February 1, 2002; 55(1): 34 - 39. [Abstract] [Full Text] [PDF]
|
