| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Chair of General Pathology and Immunology, and
* Unit of Histology and Cytology, Department of Biomedical Sciences and Biotechnology, School of Medicine, University of Brescia, 25123 Brescia, Italy;
Istituto di Ricerche Farmacologiche ‘Mario Negri’, 24125, Bergamo, Italy;
International Center for Genetic Engineering and Biotechnology, 34012 Trieste, Italy; and
§ Istituto Nazionale per la Ricerca sul Cancro, 16132 Genova, Italy
1Correspondence: Department of Biomedical Sciences and Biotechnology, Via Valsabbina 19, 25123 Brescia, Italy. E-mail: presta{at}med.unibs.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: AIDS • extracellular matrix • angiogenesis • oncogenesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. The HIV-1 Tat protein is an 86–101 amino acid
polypeptide, depending on the virus strain, which is encoded by two
exons and translated from multiply spliced 2 kb mRNAs (3)
.
Tat is released from HIV-1-infected cells (4)
.
Extracellular Tat has the ability to enter the cell and nucleus in an
active form where it stimulates the transcriptional activity of HIV-LTR
(5)
and trans-activates endogenous genes (ref
6
and references therein). Moreover, extracellular Tat
acts as a pleiotropic molecule inducing several biological effects on
target uninfected cells. Tat appears to be responsible, at least in
part, for the observed increased incidence of tumors in AIDS patients
(7
, 8)
. Indeed, Tat-transgenic mice develop
tumors with different histotypes (9
10
11)
. Also, Tat
protein trans-activates cellular proto-oncogenes and the
genome of oncogenic viruses (12
13
14)
, induces cell
immortalization (15)
, prevents apoptosis in different cell
types (16
, 17)
, and induces their proliferation
(17
18
19)
. Finally, Tat stimulates angiogenesis
(19)
that is required for tumor growth and metastasis. Tat
has been implicated in the development of Kaposi’s sarcoma (KS)
(20
, 21)
, a hypervascularized lesion frequently observed
in male AIDS patients (22)
.
Tat interacts with at least three classes of cell surface receptors
present on different target cells: cell adhesion receptors of the
integrin family (18)
, the vascular endothelial growth
factor receptors Flt-1 and Flk-1/KDR (19
, 23)
, and the
chemokine receptors CCR2 and CCR3 (24)
. Interaction
of extracellular Tat with these receptors may activate various
intracellular signaling pathways responsible for the biological
responses elicited by this protein in target cells
(24
25
26
27
28)
. Moreover, Tat is a heparin binding growth factor
(29
30
31)
, and heparan sulfate (HS) proteoglycans (HSPGs)
mediate the interaction of Tat with the cell surface (32)
and its accumulation in the extracellular matrix (ECM)
(33)
. HSPGs are also required for Tat internalization and
consequent HIV-LTR trans-activation (29
30
31
; M.
Tyagi et al., unpublished observations). Conversely, free heparin
inhibits the uptake of extracellular Tat and its HIV-LTR
trans-activating activity (29
30
31)
, modulates
the angiogenic activity of Tat in vitro and in
vivo (32)
, and affects cell surface interaction,
mitogenic activity, and protease-inducing activity exerted by Tat
protein in murine adenocarcinoma T53 cells (6
, 30
, 31)
.
Finally, other ECM components, including fibronectin, fibrinogen, and
vitronectin, modulate the biological activity of Tat in different cell
types (34
, 35)
.
Platelet thrombospondin-1 (TSP) is a heparin binding trimeric protein
of 440 kDa secreted by several cell types including endothelial and
inflammatory cells (36
37
38
39)
. In vivo TSP is
found as a free molecule or in a cell-associated, HSPG-bound form
(40
41
42)
. Multiple and sometimes opposite functions have
been attributed to TSP. Indeed, it both promotes and inhibits cell
adhesion, motility, and proliferation of different cell types
(43
44
45
46)
. Also, TSP production appears to be inversely
correlated to the angiogenic and tumorigenic potential of neoplastic
cells (47
, 48)
and TSP inhibits tumorigenesis and
angiogenesis in different in vitro and in vivo
models (48
49
50
51
52)
. The pleiotropic effects of TSP are
probably the consequence of its modular structure, which enables TSP to
bind to specific cell receptors and to growth factors and enzymes, such
as basic fibroblast growth factor (FGF2) (53)
, hepatocyte
growth factor/scatter factor (HGF) (54)
, transforming
growth factor-ß (TGF-ß) (55)
, platelet derived growth
factor (PDGF) (56)
, and several proteinases
(57)
.
Recent observations have shown that TSP also inhibits the mitogenic and
chemotactic activity exerted by Tat protein on endothelial cells in
culture and modulates its angiogenic activity in vivo
(58)
. In the present paper, we investigated the capacity
of TSP to bind directly to HIV-Tat protein and to modulate its
bioavailability to target cells. The results demonstrate that TSP binds
with high affinity to Tat. This interaction sequesters Tat protein in
the extracellular environment and prevents its biological activity in
different in vitro and in vivo experimental
systems. Thus, TSP appears to be an important regulator of the
biological functions of extracellular Tat protein.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The purity was higher than
90% as assessed by sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE), with TGF-ß contamination being lower than
0,02%. The rabbit polyclonal anti-TSP antibody has been described
(53)
. Briefly, it recognizes TSP (0,034–1 pmol) but not
fibronectin or type IV collagen at 1:1000 dilution in enzyme-linked
immunoassay (ELISA). The neutralizing anti-Tat antiserum was provided
by A. Corallini (University of Ferrara, Italy). Recombinant HIV-1 Tat
has been expressed in Escherichia coli as
glutathione-S-transferase (GST) fusion protein. The corresponding
plasmid construct derives from pGST-Tat 2E, originally obtained by
cloning the coding region of both exons of HIV-1HXB2 Tat in the
commercial vector pGEX2T (60)
. This construct codes for
the 86 amino acid Tat protein. Recombinant GST-Tat was also fused at
its carboxyl terminus to the green fluorescent protein (GFP) as already
described (31)
. Purification of recombinant GST, GST-Tat,
and GST-Tat-GFP proteins was performed by glutathione-agarose affinity
chromatography (60
, 61)
. The purity (>95%) and integrity
of the proteins were routinely checked by SDS-PAGE and silver staining.
Bacterial lipopolysaccharide contamination was 
1% and heating
completely destroyed the biological activity of GST-Tat preparations in
all the assays. Previous studies have shown that the GST and GFP
moieties do not interfere with the LTR trans-activating
activity of Tat and its capacity to bind heparin (29
, 31
,
and data not shown).
Recombinant GST-Tat was labeled with [125I] (17
Ci/mg, New England Nuclear, Boston, Mass.) using Iodogen (Pierce
Chemical Co., Rockford, Ill.) to a specific radioactivity of 400
cpm/fmol as described (30)
.
Gel permeation chromatography
TSP (2.3 pmol), GST-Tat (23 pmol), or a mixture of 16 pmol of
TSP plus 23 pmol of GST-Tat, each in a final volume of 100 µl in 10
mM phosphate buffer, pH 7.3, containing 0.15 M NaCl (PBS), was applied
on a gel filtration fast protein liquid chromatography Superose-12
column (Pharmacia, Uppsala, Sweden). The column was then eluted with
the same buffer at a flow rate of 0.7 ml/min and 0.7 ml fractions were
collected; 350 µl of each fraction was spotted onto a nitrocellulose
membrane and subjected to dot blot analysis using a 1:400 dilution of
anti-TSP antiserum or a 1:100 dilution of anti-Tat antiserum.
Immunocomplexes were visualized by chemiluminescence using the ECL
Western blotting kit (Amersham International, Little Chalfont, U.K.)
and quantified by computerized image analysis. Control experiments
demonstrated that anti-TSP and anti-Tat antibodies do not cross-react
(data not shown).
GST-Tat affinity chromatography
TSP (0.6 pmol) in a final volume of 200 µl of PBS was applied
onto columns (5x10 mm) of GST or GST-Tat immobilized to
glutathione-agarose beads. After extensive washing with PBS, columns
were eluted at a flow rate of 0.1 ml/min with a linear gradient of NaCl
in PBS. Fractions (0.3 ml) were collected, spotted onto a
nitrocellulose membrane, and probed by dot blot analysis with anti-TSP
antibody (1:400). Immunocomplexes were visualized and quantified as
described above.
125I-GST-Tat binding to TSP- or heparin-coated plastic
TSP was adsorbed to 96-wells polystyrene nontissue culture
microplates as described (53)
. Briefly, wells were
incubated for 2 h at 37°C with 40 µl aliquots of TSP (100 nM)
in TBS (25 mM Tris-HCl pH 7.4 plus 130 mM NaCl). Then wells were washed
and overcoated with 1% bovine serum albumin (BSA) in TBS. Under these
experimental conditions, up to 7.5% of TSP originally added to the
well remain immobilized to the plastic (53)
. In parallel
experiments, nontissue culture microplates were incubated overnight at
4°C with 50 µl aliquots of heparin (30 µM in TBS). After three
washes with TBS, wells were incubated for another 30 min at room
temperature with 1% BSA in TBS. Experiments performed by using
3H-heparin as a tracer revealed that up to 37
pmol of heparin are immobilized to the plastic under these experimental
conditions. For binding experiments, 100 µl aliquots of
125I-GST-Tat (0.54 nM in TBS) were incubated for
2 h at 4°C into TSP- or heparin-coated wells in the absence or
presence of different competitors. At the end of incubation, wells were
washed three times with TBS; Tat-associated radioactivity was
solubilized by incubating the wells for 30 min at 50°C with 2% SDS
in H2O, collected, and measured in a liquid
scintillation counter. The amount of 125I-GST-Tat
bound to BSA-coated wells was used as nonspecific binding and was
subtracted from all the values.
To determine the Kd of the interaction of
125I-GST-Tat with TSP, 100 µl aliquots of TBS
containing different concentrations of
125I-GST-Tat were added to TSP-coated wells. Then
samples were processed exactly as described above. Binding data were
analyzed by the Scatchard plot procedure (62)
.
Cell cultures and TSP cDNA transfection
HL3T1 cells are derived from HeLa cells and contain integrated
copies of pL3CAT, a plasmid in which the chloramphenicol
acetyltransferase (CAT) bacterial gene is driven by HIV-1 LTR
(63)
. The T53 cell line was established from
adenocarcinoma of skin adnexa of Tat-transgenic mice. They
produce and secrete biologically active Tat, which in turn stimulates
their proliferation (16
, 64)
. T53 TAT-AS clone 10 cells
were obtained by transfection of T53 cells with a plasmid containing an
antisense HIV-1 Tat cDNA. This clone is characterized by a 70–80%
decrease of the production of Tat protein when compared to parental T53
cells (6)
. T53 Tat-less cells were obtained by
subcloning parental T53 cells by the limiting dilution method. Dot blot
analysis of the conditioned medium of T53 Tat-less cells
performed with anti-Tat antibodies reveals that this clone does not
produce detectable amounts of extracellular Tat when compared to
parental T53 cells (31)
. However, T53 Tat-less
cells retain their capacity to respond to extracellular Tat in terms of
cell proliferation (see Results).
Parental T53 cells were transfected with the expression vector
pCMVNeoTHBS-1 obtained by inserting the 3567 bp human TSP cDNA into
pCMVBamNeo under the control of the CMV promoter (a kind gift from
P. S. Steeg, NCI, NIH, Bethesda, Md.) (48)
. To obtain
stable transfectants, T53 cells were plated at 2.5 x
103 cells/100 mm plates and transfected with a
calcium phosphate precipitate containing 20 µg of plasmid DNA and 20
µg of salmon sperm DNA. After 72 h, 500 µg/ml of G418 sulfate
antibiotic (Sigma. St. Louis, Mo.) was added to cell cultures.
G418-resistant clones were isolated and tested for Tat and TSP protein
expression by dot blot analysis or ELISA assay of the cell
supernatants, respectively.
All cell types were grown and maintained in Dulbecco’s modified Eagle medium (DMEM) (Life Technologies, Inc., Grand Island, N.Y.) with 10% fetal calf serum (FCS; Life Technologies).
LTR/CAT trans-activation assay
LTR/CAT trans-activation assay was performed as
described (29)
. Briefly, HL3T1 cells were incubated for
24 h in DMEM containing 10% FCS and 100 µM chloroquine in the
absence or presence of GST-Tat (5.4 nM) and of increasing
concentrations of TSP. Then conditioned medium was removed and cell
cultures were incubated for another 24 h in DMEM containing 10%
FCS. At the end of incubation, the amount of CAT protein present in the
cell extracts was determined by ELISA using the CAT ELISA kit
(Boehringer, Mannheim, Germany) according to manufacturer’s
instructions.
Internalization of 125I-GST-Tat and GST-Tat-GFP in
HL3T1 cells
The evaluation of 125I-GST-Tat or of
GST-Tat-GFP cell internalization was performed as already described
(30
, 31)
. Briefly, for 125I-GST-Tat,
cells were incubated in the absence or presence of TSP for 16 h at
37°C in medium containing 0.15% gelatin and 20 mM HEPES, pH 7.5, and
a mixture of 125I-GST-Tat (0.54 nM) and unlabeled
GST-Tat (5.4 nM) used as a carrier (30)
. At the end of
incubation, cells were washed and lysed. Radioactivity was then
measured in the cell extract. Under these experimental conditions, up
to 90% of the radioactivity remains associated to the cells after a
wash with 2.0 M NaCl in 20 mM sodium acetate pH 4.0, thus demonstrating
the intracellular localization of cell-associated
125I-GST-Tat (31)
. Nonspecific
radioactivity was evaluated by incubating the cells at 4°C with
125I-GST-Tat (0.54 nM) and a 200-fold molar
excess of unlabeled GST-Tat, and subtracted from each experimental
point.
For GST-Tat-GFP, HL3T1 cells adherent to glass coverslips were incubated for 6 h at 37°C in DMEM containing 10% FCS and GST-Tat-GFP (5.4 nM) in the absence or presence of different competitors. At the end of incubation, cells were washed twice with 2.0 M NaCl in PBS to remove cell surface-bound fluorescence and fixed. Observations were carried out under a Nikon photomicroscope equipped for epifluorescence and GST-Tat-GFP internalization was quantified by computerized image analysis.
Image analysis
The Image Pro-Plus analysis system (Media Cybernetics, Silver
Spring, Md.) was used to quantify the amount of GST-Tat-GFP
internalized in HL3T1 cells. Three to six fields were chosen randomly
for each experimental condition and input via a TV camera (Sensicam,
CCD imaging, Kelheim, Germany) mounted on the microscope, digitalized
on a high resolution monitor, and stored within the Pro-Plus analysis
system’s memory. The amount of GST-Tat-GFP internalized by HL3T1 cells
was quantified by counting the intracellular fluorescent granules
corresponding to cell lysosomes loaded with the GFP fusion protein.
T53 cell proliferation assays
Parental T53 cells, T53 TAT-AS clone 10 cells, and T53
Tat-less cells were seeded in 96-well dishes at 10,000
cells/cm2 in DMEM containing 10% FCS. After
24 h, cell cultures were washed twice with DMEM and incubated for
24 h in fresh medium containing 10% FCS in the absence or
presence of GST-Tat (2.7 nM) and/or different competitors. At the end
of incubation, cells were trypsinized and counted in a Burker chamber.
Chick embryo chorioallantoic membrane (CAM) assay
Fertilized chicken eggs were incubated under constant
humidity at 37°C. On the third day of incubation, a window was opened
in the eggshell after removal of 2–3 ml of albumen to detach the shell
from the developing CAM. The window was sealed with a glass and the
eggs were returned to the incubator. On day 8, 10
mm3 sterilized gelatin sponges (Gelfoam; Upjohn
Co., Kalamazoo, Mich.) were implanted on the top of the CAMs under
sterile conditions and adsorbed with 10 µl of 10% FCS in DMEM alone
(negative control) or containing parental T53 cells, TSP-transfected
cells, or FGF2-overexpressing mouse aortic endothelial (MAE) pZipFGF2
cells, used here as a positive control (65)
. All cell
types were tested at 6 x 104 or 2 x
104 cells/sponge (5–6 embryos per group). CAMs
were examined daily until day 12 and photographed in ovo
under a Zeiss SR stereomicroscope. The angiogenic response was scored
by two investigators without knowledge of the samples tested and graded
on an arbitrary scale of 0–4+, with 0 representing no angiogenic
response and 4+ representing the strongest activity.
Subcutaneous (s.c.) tumor growth assay
Female NCr-nu/nu mice were obtained from the National
Cancer Institute-Frederick Cancer Research and Development Center
(Frederick, Md.) and used when 6–8 wk old. Parental and
TSP-transfected T53 cells were harvested by a brief exposure to 0.25%
trypsin/0.02% EDTA, washed twice, and resuspended in DMEM. Mice were
given a s. c. injection of 1 x 106 cells
resuspended in 0.2 ml DMEM in the dorsal scapular region (5 animals per
group). The tumor mass was measured twice a week with calipers and
tumor weight (in grams) was estimated by the formula: (length x
width2)/2. At death, lesions were removed and
divided in two parts: one part was paraffin embedded and tissue
sections were stained with hematoxylin and eosin; the other one was
processed for immunohistochemical analysis as described below.
Immunohistochemistry and assessment of microvessel density
Tissues were embedded in OCT compound, frozen, and 5 µm
sections were obtained by a cryostat microtome. For immunolocalization
of TSP, sections were rinsed in PBS and incubated for 20 min with 0.3%
H2O2 in absolute methanol
to block endogenous peroxidase and for a further 20 min with 0.2%
Triton X-100 in PBS. Then a 30 min preincubation with diluted normal
serum was followed by incubation at 4°C with undiluted monoclonal
anti-TSP antibody (58)
in a humidified chamber. Sections
were then exposed to biotinylated secondary antibody (Vector
Laboratories. Burlingame, Calif.) and to avidin-biotin-peroxidase
complex (DAKO ABComplex HRP) for 30 min. Peroxidase color reaction was
developed with 3-amino-9-ethyl-carbazole (Sigma) and the sections were
lightly counterstained with Mayer’s hematoxylin.
To detect blood vessels, sections were processed exactly as described
above except for the use of a rat anti-murine CD31 antibody
(66)
replacing the monoclonal anti-TSP antibody. To
evaluate microvessel density, sections from each tumor were examined at
low-power magnification to identify the areas with the highest density
of CD31-positive vessels in the parenchymal and stromal compartments.
For both compartments, the most vascularized area was selected and
microvessels in a x400 field were counted. Microvessels were defined
as any immunoreactive endothelial cell(s) that were separate from
adjacent microvessels. Vessel lumens were not necessary for a structure
to be defined as a microvessel. No significant differences in
microvessel counts were observed between paired sections of individual
tumors.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
54
55
56)
. Since HIV-1 Tat protein acts like an angiogenic
factor (19
, 20
, 32)
and TSP inhibits angiogenesis induced
by Tat (58)
, we assessed the capacity of TSP to bind Tat
protein. TSP or GST-Tat was first independently subjected to size
exclusion chromatography onto a Superose 12 column. Under these
experimental conditions, TSP elutes with the void volume of the column
(8 ml) whereas GST-Tat elutes with a retention volume of 14 ml
(Fig. 1A
). In contrast, when GST-Tat is incubated with TSP before
chromatography and then loaded onto the column, it elutes as a
TSP/GST-Tat complex in the void volume (Fig. 1A
). Under the
same experimental conditions, ferritin does not affect the
chromatographic behavior of GST-Tat and TSP does not alter the
chromatographic behavior of cytochrome c, thus supporting
the specificity of TSP/Tat interaction (data not shown).
|
When comparing the elution profiles of GST-Tat samples preincubated with increasing amounts of TSP, we observed the complete disappearance of the peak of free GST-Tat at a TSP:GST-Tat molar ratio equal approximately to 1:3, suggesting that each monomer of the trimeric TSP protein binds one molecule of GST-Tat.
To confirm the capacity of TSP to interact with Tat protein, GST-Tat
was immobilized onto agarose-glutathione beads to generate a GST-Tat
affinity column, which was then assessed for the capacity to bind free
TSP. As shown in Fig. 1B
, TSP binds with high affinity to
immobilized GST-Tat, from which it is eluted at 0.5 M NaCl. The
specificity of the interaction was demonstrated by the incapacity of
TSP to bind to immobilized GST protein devoid of the Tat moiety (Fig. 1B
).
TSP exists in both a free and an immobilized form
(39
40
41
42)
. To assess whether immobilized TSP retains its
Tat binding capacity, TSP was adsorbed to plastic and evaluated for its
capacity to bind to 125I-GST-Tat. As shown in
Fig. 2A
, 125I-GST-Tat interacts with
TSP-coated plastic in a dose-dependent manner, maximal binding obtained
when wells were coated with TSP at the concentration of 50–100 nM. The
binding of 125I-GST-Tat to immobilized TSP is
specific since it was inhibited by a molar excess of unlabeled GST-Tat
and by anti-Tat antibodies (Fig. 2A
). On this basis, to
determine the relative affinity of TSP/Tat interaction, increasing
concentrations of 125I-GST-Tat were incubated
onto TSP-coated wells. At the end of incubation, the amount of
125I-GST-Tat bound to TSP was measured and the
binding data were analyzed by the Scatchard plot procedure
(62)
. As shown in Fig. 2B
, the binding of
125I-GST-Tat to immobilized TSP is dose dependent
and saturable. Scatchard regression of the binding data (Fig. 2B
, inset) reveals a single component binding with a
Kd equal to 25 nM.
|
Effect of heparin on TSP/Tat interaction
Both Tat protein (29
30
31
32)
and TSP
(39
40
41
42)
are heparin binding molecules. It was thus
interesting to assess whether heparin affects Tat/TSP interaction.
Heparin was added to TSP-coated wells together with or 1 h after
125I-GST-Tat. As shown in Fig. 3A
, free heparin prevents the binding of
125I-GST-Tat to TSP-coated plastic, being
ineffective when added 1 h after the radiolabeled molecule.
|
In a second set of experiments, heparin was immobilized to plastic and
evaluated for its ability to bind 125I-GST-Tat in
the presence of TSP in solution. Preliminary experiments in the absence
of TSP had shown that 125I-GST-Tat binds to
immobilized heparin but not to chondroitin-4-sulfate (Ch-4-S) (data not
shown). Also, the interaction is inhibited by 2 µM unlabeled heparin,
a concentration able to compete for the binding of
3H-heparin to immobilized GST-Tat (29
, 30)
. Again, free Ch-4-S was ineffective (data not shown). As
shown in Fig. 3B
, TSP in solution competes for the binding
of 125I-GST-Tat to immobilized heparin when added
together with the trans-activating factor, but is unable to
dissociate the already formed Tat/heparin complex.
TSP inhibits HIV-1 LTR trans-activation induced by
extracellular Tat in HL3T1 cells
Extracellular Tat protein can enter the cell and reach the nucleus
where it stimulates the transcriptional activity of HIV-LTR
(5)
. On this basis, we investigated the possibility that
TSP interaction may affect the LTR trans-activating activity
of extracellular Tat. To this purpose, HL3T1 cells containing the
CAT gene under the control of HIV-1 LTR were treated with
GST-Tat in the presence of increasing concentrations of TSP. At the end
of treatment, the amount of CAT protein produced by the cells,
proportional to the LTR trans-activating activity exerted by
GST-Tat, was measured by a CAT ELISA assay. As shown in Fig. 4A
, TSP inhibits the trans-activating activity
exerted by GST-Tat in a dose-dependent manner, with an
ID50 equal to 10 nM. TSP inhibited with similar
potency the trans-activating activity of GST-Tat when the 86
amino acid Tat moiety was replaced by the 101 amino acid Tat form, a
widely distributed form of Tat produced by most primary HIV-1 strains
(67)
(Fig. 4A
). TGF-ß, potentially present in
TSP preparations as a contaminant (55)
, does not affect
the LTR trans-activating activity of GST-Tat in HL3T1 cells
when tested at doses of up to 10 ng/ml (data not shown).
|
To investigate the mechanism(s) of inhibition of the LTR
trans-activating activity of Tat, TSP was administered to
HL3T1 cells together with GST-Tat or at different times before/after
GST-Tat administration. As shown in Fig. 4B
, TSP inhibits
the LTR trans-activating activity of Tat when added up to
3 h before the administration of GST-Tat and maintained throughout
the whole experimental period or when added within the first 15 min
after the beginning of GST-Tat treatment. In contrast, the addition of
TSP to the cell culture medium 30 min after the beginning of GST-Tat
treatment was ineffective, indicating that TSP affects an early event
in the interaction of Tat with HL3T1 cells. Note that different Tat
antagonists, such as the polyanionic heparin and suramin, retain the
capacity to inhibit the trans-activating activity of
extracellular Tat when added up to 3 h after the administration of
the trans-activating factor (30)
.
We then evaluated the effect of TSP on cell internalization of Tat
protein, a phenomenon that depends on the interaction of the
trans-activating factor with cell surface HSPGs (refs
29
30
31
; M. Tyagi et al., unpublished observations). HL3T1
cells were treated with 125I-GST-Tat in the
presence of increasing concentrations of TSP. At the end of treatment,
free 125I-GST-Tat was removed, cells were lysed,
and the amount of intracellular radioactivity was measured. In parallel
experiments, HL3T1 cells were treated with the fluorescent chimera
GST-Tat-GFP in the presence of increasing concentrations of TSP. At the
end of treatment, free and cell surface-associated GST-Tat-GFP was
removed, cells were fixed, and the amount of intracellular fluorescence
was quantified by computerized image analysis (31)
. As
shown in Fig. 5A
, TSP inhibits cell internalization of
125I-GST-Tat and GST-Tat-GFP in a dose-dependent
manner with an ID50 equal to 10–30 nM.
|
TSP loses its capacity to inhibit the trans-activating
activity of Tat when added to HL3T1 cells 30 min after the
trans-activating factor (see Fig. 4B
). We
evaluated the effect of a delayed TSP administration on cell
internalization of GST-Tat-GFP. As shown in Fig. 5B, C
, TSP
inhibits the internalization of GST-Tat-GFP only when added together
with the fluorescent protein, being ineffective when added 1 h
after GST-Tat-GFP administration. Under the same experimental
conditions, heparin inhibits the internalization of GST-Tat-GFP both
when added together with the fluorescent protein or 1 h after its
administration (Fig. 5B
). These data indicate the ability of
TSP to prevent but not disrupt the binding of Tat with cell surface
HSPGs, which is required for cell internalization and LTR
trans-activation (refs 29
30
31
; M. Tyagi et al.,
unpublished observations). This agrees with the in vitro
observations showing that TSP inhibits the interaction of Tat with
substrate-immobilized heparin but does not dissociate the protein from
the glycosaminoglycan once the interaction has occurred (see Fig. 3B
).
TSP inhibits the mitogenic activity of Tat in murine adenocarcinoma
T53 cells
To further characterize the Tat-antagonist activity of TSP,
we evaluated its capacity to modulate the mitogenic activity exerted by
extracellular Tat on T53 Tat-less cells. These cells were
obtained in our laboratory (31)
by subcloning the
Tat-overexpressing T53 cell line originally established from
adenocarcinoma of skin adnexa of Tat-transgenic mice
(16
, 64)
. T53 Tat-less cells do not produce
significant amounts of extracellular Tat when compared to parental T53
cells, but still retain their capacity to proliferate when exposed to
GST-Tat (Fig. 6A
). T53 Tat-less cells were treated with GST-Tat
in the absence or presence of increasing concentrations of TSP. At the
end of treatment, cells were trypsinized and counted. As shown in Fig. 6A
, TSP inhibits the mitogenic activity of GST-Tat in a
dose-dependent manner (ID50 equal to 3 nM)
without affecting the basal proliferation induced in T53
Tat-less cells by 10% FCS. To investigate the time
dependency of the inhibitory activity of TSP, the molecule was added to
T53 Tat-less cells together with GST-Tat or 1 h after
its administration. As shown in Fig. 6B
, TSP exerts its
inhibitory effect only when added to the cells together with the
trans-activating factor. Under the same experimental
conditions, heparin inhibits the mitogenic activity of GST-Tat
independently of the time of administration.
|
In agreement with these observations, 100 nM TSP inhibits the binding of GST-Tat-GFP to T53 Tat-less cells (data not shown). As observed for Tat/HL3T1 cell interaction (see above), TSP exerts its inhibitory activity when added together with but not 1 h after GST-Tat-GFP administration. Under the same conditions, heparin inhibits the interaction of GST-Tat-GFP with T53 Tat-less cells independently of the time of administration (data not shown).
Our data indicate that exogenous recombinant GST-Tat exerts a mitogenic
effect on T53 Tat-less cells that is inhibited by free TSP.
Parental T53 cells are characterized instead by a high rate of basal
proliferation due to the autocrine stimulation exerted by the
endogenously synthesized Tat protein (16
, 64)
. This
autocrine loop of stimulation can be interrupted by neutralizing
anti-Tat antibodies or by Tat antagonists, including heparin and
suramin (30
, 31)
. We investigated whether TSP can
interfere with the autocrine loop of stimulation exerted by endogenous
Tat in T53 cells. Parental T53 cells were incubated for 24 h in
the absence or presence of increasing concentrations of TSP and then
counted. As shown in Fig. 7A
, TSP inhibits T53 cell proliferation in a dose-dependent
manner, with an ID50 equal to 10 nM. The
inhibitory effect appears to be specific, since a different ECM
component such as fibronectin does not affect significantly cell
proliferation when tested at doses up to 100 nM. Also, TSP does not
affect the basal proliferation of a clone of T53 cells (T53 TAT-AS
clone 10) obtained by transfection with an antisense Tat cDNA (Fig. 7B
). These cells are characterized by a very limited
production of Tat protein and a reduced, Tat-independent rate of cell
proliferation when compared to parental T53 cells. Taken together, the
data indicate that TSP inhibits T53 cell proliferation by acting on
endogenously synthesized Tat protein released by the cells in the
extracellular compartment.
|
TSP overexpression inhibits cell proliferation in T53 cells
T53 cells do not synthesize significant amounts of TSP (see
below). To assess the effect of endogenous TSP on Tat-dependent cell
proliferation, parental T53 cells were stably transfected with an
expression vector harboring the full-length human TSP cDNA.
Transfectants were first screened for Tat protein production and
secretion by dot blot analysis of their conditioned media using
specific anti-Tat antibodies. Only clones characterized by a production
of Tat protein comparable to that of parental T53 cells were then
subjected to a second screening for TSP expression and secretion by
ELISA. Two clones, named T53-TSP-7 and T53-TSP-10 cells, were then
selected and characterized further.
In detail, parental T53 cells and the two TSP-transfected clones secrete comparable amounts of Tat protein (0.6–0.7 ng/ml/106 cells in 24 h). Also, T53-TSP-7 cells and T53-TSP-10 cells secrete in the medium 14 ng/ml and 65 ng/ml of TSP per 106 cells in 24 h, respectively, whereas parental T53 cells do not release detectable amounts of TSP.
The two TSP-expressing clones were then compared to parental T53 cells
for their rate of growth in vitro under basal culture
conditions (Fig. 8A
). Both T53-TSP-7 and T53-TSP-10 cells are characterized by
a rate of growth significantly lower than that of parental T53 cells,
indicating that endogenously synthesized TSP has the capacity to
inhibit the autocrine loop of stimulation exerted by Tat in T53 cells,
as observed for exogenously added TSP.
|
TSP overexpression inhibits the angiogenic and tumorigenic capacity
of T53 cells
T53 cells induce in vivo neovascularization that is
inhibited by anti-Tat antibodies (32)
. Moreover, the
angiogenic activity of recombinant Tat can be inhibited by exogenously
added TSP (58)
. Also, T53 cells induce tumorigenesis when
injected in nude and syngeneic mice with several T53-derived clones
showing a tumorigenic capacity comparable to that of parental cells
(16)
. Since TSP has been demonstrated to inhibit
angiogenesis and tumorigenesis in different in vivo models
(48
49
50
51
52)
, we compared the angiogenic and tumorigenic
potential of parental T53 cells to that of T53-TSP-10 cells that
express the highest levels of TSP.
In a first series of experiments, gelatin sponges adsorbed with
parental T53 cells and T53-TSP-10 cells were implanted on the top of
8-day-old chick embryo CAMs. Sponges adsorbed with vehicle alone or
with FGF2-overexpressing pZipFGF2 MAE cells (65)
were used
as negative and positive controls, respectively. CAMs were examined
daily until day 12 when the angiogenic response was scored (Fig. 9
). T53 cells exert an angiogenic response similar to that exerted by
FGF2-overexpressing cells when tested at 2 x
104 or 6 x 104
cells/sponge. In contrast, neovascularization induced by T53-TSP-10
cells is strongly reduced and comparable to that observed in
vehicle-treated CAMs (Fig. 9)
.
|
To evaluate the effect of TSP overexpression on the tumorigenic
activity of T53 cells, parental T53 cells and T53-TSP-10 cells were
injected s.c. in nude mice. As shown in Fig. 8B
, tumors
appeared rapidly (median 5 days, range 5–9 days) in mice injected with
parental T53 cells and reached an average weight of 1.0 g at 5 wk.
In contrast, injection of the same number of T53-TSP-10 cells resulted
in the later appearance of slow-growing tumors (median 26 days, range
19–33 days) that reached the weight of 
1.0 g at 7 wk.
Histological analysis performed on 5-wk-old lesions revealed that
parental T53 tumors are characterized by tightly packed, poorly
differentiated spindle-shaped cells with a limited stromal component
and occasional small necrotic foci (Fig. 10a
, c
). In contrast, T53-TSP-10 lesions show alveolar and
tubular structures formed by more differentiated adenocarcinoma cells,
with abundant stromal component and extensive areas of necrosis (Fig. 10b
, d
). Immunohistochemical analysis showed a
strong TSP immunoreactivity in T53-TSP-10 lesions, mainly associated
with the stromal component (Fig. 10f
). A very faint TSP
immunoreactivity was instead detected in tumors originated by parental
T53 cells (Fig. 10e
).
|
CD 31 immunostaining followed by microvessel counting demonstrated that the density of blood vessels in the stromal component of the tumors was reduced in T53-TSP-10 lesions when compared to parental T53 lesions (23±1 vs. 31±2 blood vessels/field, respectively; P<0.001, Student’s t test), with no major differences in vascularity of the neoplastic parenchyma (57±1 and 65±2 blood vessels/field for T53 and T53-TSP-10 tumors, respectively).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
0.5 M NaCl) and by anti-Tat antibody or
heparin. Instead, no binding was observed to GST devoid of the Tat
moiety, thus confirming the specificity of the interaction of TSP with
the trans-activating protein.
Previous observations had shown the ability of TSP to bind various
growth factors and proteinases, including FGF2 (53)
, HGF
(54)
, TGF-ß (55)
and PDGF
(56)
, with an affinity similar to that measured for Tat
interaction. Note that most of the TSP ligands share with Tat the
capacity to bind heparin/HS (54
, 68
69)
and that heparin
prevents their binding to TSP (53
, 54)
. Our finding that 1
mol of TSP binds 3 mol of Tat suggests that each TSP monomer binds one
Tat molecule. Preliminary experiments have shown that FGF2 competes for
the binding of GST-Tat to TSP, raising the possibility that similar
molecular determinants are responsible for TSP interaction. Also,
experimental evidences suggest that the TSP binding domain of Tat is
located within the COOH terminus of the protein (M. Rusnati,
unpublished observations), a region of the protein that contains the
integrin binding domain of Tat (18)
but is distinct from
the basic domain responsible for KDR receptor (19)
and
heparin/HSPG (30
, 33)
interactions. Additional experiments
are required to identify unambiguously the molecular bases of TSP/Tat
interaction.
TSP inhibits cell internalization and HIV-LTR
trans-activating activity of GST-Tat in HL3T1 cells. Also,
TSP inhibits cell interaction and mitogenic activity of GST-Tat in T53
Tat-less cells. In all the assays, TSP exerts its antagonist
activity with an ID50 close to the
Kd of TSP/Tat interaction (25 nM). Cell surface
binding of Tat occurs mainly through HSPGs (32
, and M.
Tyagi et al., unpublished observations) and here we have shown that TSP
competes for the binding of GST-Tat to immobilized heparin. Taken
together, our data strongly support the hypothesis that TSP exerts its
inhibitory effect by binding and sequestering the
trans-activating factor in the extracellular environment,
thus hampering its capacity to interact with the cell surface and to
exert its biological activity. We have recently demonstrated that TSP
inhibits Tat-induced proliferation, chemotaxis, migration of
endothelial cells in culture, and angiogenesis in vivo
(58)
. The mitogenic activity exerted by Tat in endothelial
cells is inhibited by TSP with an ID50 equal to
28 nM, in keeping with the proposed extracellular sequestration of Tat
protein by TSP. In contrast, Tat-induced chemotaxis, migration, and
angiogenesis are inhibited at much lower concentrations of TSP
(0.02–0.2 nM), raising the possibility of alternative mechanisms of
action. For instance, both Tat and TSP bind
vß3 integrin
(18
, 36)
, suggesting that TSP may also interfere with
Tat/vß3 interaction.
Moreover, the ability of TSP to bind several proteases
(57)
may affect the capacity of Tat to modulate
plasmin-dependent proteolysis (6)
required during the
neovascularization process.
TSP competes for the binding of GST-Tat to immobilized heparin when
added together with the trans-activating factor, but is
unable to dissociate the already formed Tat/heparin complex.
Accordingly, TSP completely loses its antagonist ability when added to
cell cultures 30 min after Tat administration. This capacity of TSP to
prevent but not disrupt the interaction of Tat with heparin and cell
surface HSPGs is of interest. Recently, we have demonstrated that Tat
establishes a cooperative interaction with heparin in which the binding
of the first Tat molecule facilitates the interaction of more Tat
molecules to adjacent Tat binding regions of the glycosaminoglycan
chain. This causes the formation of highly packed multimeric
Tat/heparin complexes with a 100-fold increase of the affinity of the
interaction (31)
. Thus, it is possible to hypothesize that
TSP can prevent the formation of these complexes by binding to Tat and
hampering its cooperative interaction with heparin/HS. However, TSP
becomes ineffective once the multimeric high-affinity
Tat/glycosaminoglycan complexes are formed. Accordingly, free heparin
can prevent but not disrupt the interaction of immobilized TSP with
Tat.
Although poorly present in plasma, TSP is found in serum after platelet
activation and bound to ECM and cell surface HSPGs of many human
tissues (36
37
38
39
40
41)
. Both free and immobilized TSP bind Tat.
Heparin can destroy this interaction and the interaction of Tat with
the cell surface (30
31
32)
. Finally, TSP prevents but does
not disrupt the binding of Tat to heparin and cell surface. Thus, a
scenario can be envisaged in which extracellular Tat may bind to free
and cell-associated TSP and/or HSPGs. This will affect the
bioavailability and biological activity of extracellular Tat
differently depending on their relative concentrations and timing of
interaction. It is worth noting that the release of HSPGs from the cell
surface is strictly regulated by the plasmin system (68)
,
which in turn can be modulated by TSP (46
, 57)
and Tat
(6)
. Thus, a complex and strictly regulated interplay may
occur among Tat and the free or immobilized forms of TSP and HSPGs.
This kind of spatial control has been already demonstrated for
heparin/FGF2 interactions (68)
and TSP/plasminogen
interactions (57)
.
T53 cells produce and secrete biologically active Tat protein, which in
turn stimulates their proliferation in vitro in an autocrine
manner and is responsible for their angiogenic and tumorigenic capacity
in vivo (64)
. T53 cells do not express
detectable amounts of TSP. Here we have shown that their stable
transfection with an expression vector harboring the full-length human
TSP cDNA results in the production of TSP-expressing cells
characterized by a reduced angiogenic potential in the CAM assay. This
is in keeping with previous observations showing the capacity of TSP to
exert an anti-angiogenic activity in different in vitro and
in vivo models by interacting with angiogenic growth factors
(53
, 54)
and endothelial cell receptors, including CD36
(70)
, and to inhibit angiogenesis induced by recombinant
Tat protein in the murine Matrigel plug assay (58)
.
We have also observed that TSP transfectants showed a reduced
tumorigenic capacity when injected s.c. in nude mice. Xenografts were
characterized by a longer latency and a reduced rate of growth when
compared to lesions originated by parental T53 cells. This appears to
be the consequence of the capacity of endogenous TSP to inhibit the
autocrine loop of stimulation exerted by Tat protein in T53 cells.
Indeed, TSP transfectants showed a reduced rate of growth in
vitro similar to that observed in T53 Tat-less cells,
in antisense Tat cDNA-transfected T53 TAT-AS clone 10 cells, and in
parental T53 cells treated with exogenous TSP or with specific anti-Tat
antibodies (64)
. However, we cannot rule out the
possibility that the anti-tumorigenic effect of TSP may also occur
through mechanism(s) other than a direct, specific effect on
extracellular Tat activity. Indeed, TSP has been demonstrated to reduce
the tumorigenic potential of a wide spectrum of transformed cells
mainly by inhibiting the neovascularization required for solid tumor
growth (48
, 49
, 71)
. Accordingly, TSP inhibits
angiogenesis induced by Tat protein (58)
as well as by a
variety of angiogenic growth factors (54)
. Thus, the
moderate reduction of the microvascular density observed in the stroma
of the xenografts originated by TSP transfectants may contribute to the
reduced rate of growth of these tumors. Immunohistochemical analysis
has shown that TSP accumulates in the stroma of the lesions induced by
TSP transfectants. This may explain the restriction of the
anti-angiogenic effect of TSP to the stromal component of these tumors.
A stromal association of TSP has also been observed in
TSP-overexpressing human squamous cell carcinoma xenografts
(71)
, in keeping with the ability of TSP to interact with
ECM components (40
41
42)
. Finally, T53-TSP-10 tumors showed
a more differentiated morphology when compared to T53 lesions, in
keeping with the ability of TSP-overexpression to restore a normal
phenotype in transformed endothelial cells (72)
.
In conclusion, our data demonstrate that TSP interacts in
vitro with extracellular Tat protein, exerting a potent antagonist
activity on various biological activities of the
trans-activating factor. Experimental evidences suggest that
this interaction may be of importance also in vivo:
1) TSP inhibits the angiogenic and tumorigenic activity of
Tat in different in vivo models (ref 56
and
present study); 2) the Kd of TSP/Tat
interaction and the potency of the Tat-antagonist activity of TSP
measured in vitro are consistent with the concentration of
TSP found in human serum [from 1 nM, as measured in normal serum
(73)
, to 100 nM after platelet activation and in
various pathological conditions (G. Taraboletti, unpublished
observations)], thus indicating that the Tat-antagonist activity of
TSP may be biologically relevant during AIDS where abnormal platelet
activation occurs (74)
; 3) TSP inhibits with
similar potency the 86 amino acid form of Tat and its 101 amino acid
variant, with the latter representing the form most widely distributed
among the different virus strains in vivo (67)
.
Tat has been implicated in various AIDS-associated pathologies,
including KS. Recent observations have demonstrated that KS lesions do
not express TSP, leading to the hypothesis that TSP down-regulation
might be permissive for the development of KS in AIDS patients
(58
, 75)
. Also, Tat protein has been proposed as a target
for AIDS therapy (76
, 77)
, leading to the development of
anti-Tat vaccines (78)
and synthetic Tat antagonists
(30
, 31
, 79
, 80)
. Besides its Tat antagonist activity, TSP
has been shown to block HIV infection by binding to gp120
(81
82)
. Even though additional studies are required to
demonstrate unambiguously the capacity of TSP to interact with Tat
in vivo and the relevance of this interaction for the
biological activity of extracellular Tat in AIDS, our study suggests
the potential use of TSP-derived peptides and/or peptido-mimetics as
‘multi-target’ compounds able to affect different aspects of HIV
infection and AIDS-related disorders.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 8, 1999.
Revision received March 23, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), 23 pmol of GST-Tat alone (
), or a mixture of 16 pmol of
TPS plus 23 pmol of GST-Tat (•) was chromatographed on a Superose-12
column. Eluted fractions were analyzed for TSP and GST-Tat content by
dot blot analysis using anti-TSP (
). At the end of incubation, radioactivity
associated to immobilized TSP was extracted and measured. Each point is
the mean of 2–3 determinations in duplicate. SE never
exceeded 15% of the mean value. The arrow indicates the amount of
nonspecific binding evaluated by incubating 125I-GST-Tat
onto BSA-coated wells. B) Increasing concentrations of
125I-GST-Tat were incubated for 2 h at 4°C onto
wells coated with TSP at the concentration of 100 nM. At the end of
incubation, the amount of radioactivity associated to the immobilized
TSP was measured. Radioactivity bound to BSA-coated wells was
subtracted from all values. The experiment shown is representative of 2
experiments that gave similar results. Inset: Scatchard plot regression
of the 125I-GST-Tat binding data to TSP.
) of GST-Tat treatment (5.4 nM). At the
end of GST-Tat treatment, cell extracts were assayed for the levels of
CAT antigen by ELISA and data were expressed as percent of the
trans-activating activity measured in control cultures
treated with GST-Tat alone. Each point is the mean of 2–3
determinations in duplicate. SE never exceeded 12% of the
mean value.
