(Received for publication, October 29, 1996, and in revised form, May 21, 1997)
From the Department of Biological and Technological Research, Although several functions have been suggested
for chromogranin A, a glycoprotein secreted by many neuroendocrine
cells, the physiological role of this protein and of its proteolytic
fragments has not been established. We have found that mixtures of
chromogranin A fragments can inhibit fibroblast adhesion. The
anti-adhesive activity was converted into pro-adhesive activity by
limited trypsin treatment. Pro-adhesive effects were observed also with
recombinant N-terminal fragments corresponding to residues 1-78 and
1-115 and with a synthetic peptide encompassing the residues 7-57.
These fragments induced adhesion and spreading of fibroblasts on plates coated with collagen I or IV, laminin, fetal calf serum (FCS) but not
on bovine serum albumin. The long incubation time required for adhesion
assays (4 h) and the FCS requirements for optimal adhesion suggest that
the adhesive activity is likely indirect and requires other proteins
present in the FCS or made by the cells.
These findings suggest that chromogranin A and its fragments could play
a role in the regulation of cell adhesion. Since chromogranin A is
concentrated and stored within granules and rapidly released by
neuroendocrine cells and neurons after an appropriate stimulus, this
protein could be important for the local control of cell adhesion by
stimulated cells.
Chromogranin A (CgA)1
was originally discovered as the major soluble protein of adrenal
medullary chromaffin granules. This protein was found later to be a
member of a family of regulated secretory proteins present in the
electron-dense granules of many other neuroendocrine tissues and of the
nervous system (1, 2).
The physicochemical properties and tissue distribution of CgA have been
extensively characterized (2-5). For instance, CgA is co-stored with
various hormones in the secretory vesicles of cells of the
gastrointestinal tract (4), the adeno- and neurohypophysis (5), the
parathyroid (6), the endocrine pancreas (7), the thyroid C-cells (8),
the immune system (9), and the atrial myocardium (10). In addition, it
is a component of dense core synaptic granules in many areas of the
central nervous system (11, 12). Human CgA is a hydrophilic protein of
439 amino acids, characterized by low isoelectric point and by several
post-translational modifications, including glycosylation, sulfation,
and phosphorylation (3, 13, 14). Within the secretory vesicles, CgA may
form dimers and tetramers as a function of pH and Ca2+
concentration (15-17). A large body of evidence suggests that tissue-specific patterns of proteolytic processing generate different fragments of CgA (18-21). Moreover, multiple forms having different hydrodynamic sizes of 600, 100, and 55 kDa have been detected in the
serum of cancer patients (22).
As far as the physiological role is concerned, CgA has been suggested
to be involved in hormone packaging within secretory granules and in
modulating the secretory granule functions by binding Ca2+
and ATP. Moreover, it has been proposed that CgA represents a precursor
of biologically active peptides with endocrine, paracrine, and
autocrine functions (23). For instance, CgA residues 248-293 were
found to be homologous to pancreastatin, a pancreatic peptide that
inhibits insulin secretion (24), whereas chromostatin, a peptide
corresponding to residues 124-143, inhibits secretion of
catecholamines from chromaffin cells (25). Retrogradely perfused and
nerve-stimulated bovine adrenal medullae release the CgA
fragments corresponding to amino acids 1-76 and 1-113. These
fragments have been named vasostatin I and II (VS-1 and VS-2),
respectively, because of their inhibitory effects on vascular tension
(26-28).
Despite the above proposed functions, the physiological roles of this
protein remain largely undefined. In this work, we show that CgA
fragments from natural and recombinant sources can modulate the
adhesion of fibroblasts on various substrates and that a pro-adhesive domain is present in the N-terminal portion of the molecule.
96-Well polyvinyl chloride
microtiter plates (Falcon Micro Test III flexible assay plates) were
obtained from Becton Dickinson and Co. (Oxnard, CA). Bovine serum
albumin (BSA, fraction V), polyoxyethylene sorbitan monolaurate (Tween
20), paraformaldehyde, goat anti-mouse IgG horseradish peroxidase
conjugate, goat anti-rabbit IgG-peroxidase conjugate, normal goat
serum, o-phenylenediamine dihydrochloride and
streptavidin-peroxidase, laminin and collagen type IV (isolated from
Englebreth-Holm-Swarm mouse sarcoma) and poly-L-lysine were
from Sigma. Crystal violet was from Fluka-Chemica (Milan, Italy). Rat
tail type I collagen was from ICN Biomedicals, Inc. (Costa Mesa, CA).
D-Biotinyl-6-aminocaproic acid
N-hydroxysuccinimide ester was from Societá Prodotti
Antibiotici S.p.A (Milan, Italy). Enhanced chemiluminescence (ECLTM)
Western blotting kit was from Amersham Italia SRL (Milan, Italy). Milk
"Humana 3" was from Humana Italia S.p.A (Milan, Italy).
Nitrocellulose membranes were from Schleicher & Schuell (D-3354,
Dassel, Germany). Mouse monoclonal antibodies (mAb) A11 and B4E11
(anti-CgA) were described previously (29, 30).
SK-N-BE and CHP-134 neuroblastoma cells (obtained from Dr. G. Della
Valle, University of Pavia, Italy) were cultured in RPMI, 20% fetal
calf serum (FCS), 2 mM glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, 0.25 µg/ml amphotericin B, at 37 °C, 5% CO2; PC-12 (obtained from American Tissue Culture
Collection, ATCC CRL1721), LB6 and NIH-3T3 mouse fibroblasts (obtained
from Prof. F. Blasi, San Raffaele Scientific Institute), and human foreskin fibroblasts HFSF-132 (obtained from Dr. J. Bizik, San Raffaele
Scientific Institute) were cultured in Dulbecco's modified Eagle's
medium (DMEM) supplemented with 2 mM glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, 0.25 µg/ml
amphotericin B, 1 mM sodium pyruvate (DMEM, 1 ×) and 10%
FCS, at 37 °C, 5% CO2.
Natural human CgA was purified
from pheochromocytoma tissues, as follows. The tumor was frozen in
liquid nitrogen immediately after surgical excision, lyophilized, and
homogenized in distilled water. The homogenate was boiled for 6 min and
centrifuged at 120,000 × g for 30 min. The supernatant
containing CgA, which is heat stable (1), was kept at Various recombinant fragments of human CgA (rCgA) were
prepared including the following: (a) rCgA-(7-439),
(b) NH2-Ser-Thr-Ala-rCgA-(1-78) (VS-1);
(c) NH2-Ser-Thr-Ala-rCgA-(1-115) (VS-2);
(d) a variant of the rCgA-(7-439) in which residues 46-48
(RGD) were replaced with RGE (RGE-rCgA-(7-439)) (numbering according
to Konecki et al. (13)). Cloning, purification, and
characterization of all these recombinant polypeptides are described
elsewhere.2 Since all
products contain the residues 68-70 recognized by mAb B4E11 (30) and
were able to bind to B4E11-agarose, all products were purified from
recombinant Escherichia coli strains using this
immunoadsorbent, essentially as described for natural CgA.
Various 23-25-mer
peptides spanning most of the CgA sequence were simultaneously
synthesized on pre-loaded NovaSyn TGA resin (Novabiochem) using an
apparatus for manual multiple peptide synthesis and a solid phase
method based on Fmoc chemistry (31). Tyrosine and glycine spacers were
added at the N or the C termini to those peptides not containing
chromophoric amino acids, to allow spectrophotometric detection. Each
peptide was purified by reverse-phase HPLC and lyophilized.
The peptide 7-57 was synthesized by the solid phase Fmoc method using
an Applied Biosystems model 433A peptide synthesizer. After peptide
assembly the side chain protected peptidyl resin was de-blocked as
described (32) and purified to apparent homogeneity by reverse-phase
chromatography. The peptide, containing two cysteines, was oxidatively
folded by overnight treatment with 5-fold excess of oxidized
glutathione (33) and purified by ion exchange followed by reverse-phase
chromatography.
Electrospray ionization mass spectrometry
of VS-2 was carried out by PRIMM s.r.l. Laboratories (Milan,
Italy).
Polyvinylchloride
microtiter plates were coated with B4E11 (10 µg/ml in PBS, 50 µl/well, overnight at 4 °C). All subsequent steps were carried out
at room temperature. After washing three times with PBS, the plates
were blocked with 3% BSA in PBS (200 µl/well, 2 h at room
temperature) and washed with PBS again. CgA standard or sample
solutions, diluted 1:2 in PBS containing 0.5% BSA, 0.05% (v/v) Tween
20, and 2.5% normal goat serum ("assay buffer") were added (50 µl/well) and incubated for 1.5 h at 37 °C. The plates were
washed eight times by emptying and filling with PBS containing 0.05%
(v/v) Tween 20 and incubated with rabbit anti-rCgA-(7-439) antiserum,
1:1000 in assay buffer (50 µl/well, 1.5 h at room temperature).
The plates were washed again as above and further incubated with goat
anti-rabbit IgG-peroxidase conjugate (1:3000 in assay buffer, 50 µl/well, 1 h at room temperature). After the final wash, the
plates were incubated with 0.4 mg/ml o-phenylenediamine
solution in 0.05 M phosphate-citrate buffer, pH 5.0, containing 3.5 mM hydrogen peroxide (100 µl/well, 30 min). The reaction was stopped by adding 10% (v/v) sulfuric acid (100 µl/well), and the absorbance of each well was read at 492 nm. Each
assay was calibrated with eight rCgA-(7-439) solutions at various
concentrations.
Sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE) was carried out in a Phast System
apparatus (Pharmacia) using ready made polyacrylamide gels (Phast Gels
12.5% or 20%, Pharmacia). Samples were 2-fold diluted with 20 mM Tris-HCl, pH 8.0, containing 2 mM EDTA, 5%
(w/v) SDS, 10% (v/v) HSFS-132 cells were detached by
treatment with a prewarmed trypsin/EDTA solution (0.2 µg/ml trypsin,
0.2 mg/ml EDTA) for 5 min at 37 °C, and mixed with DMEM 1 × supplemented with 10% FCS. After three washings with 0.9% (w/v)
sodium chloride to remove residual FCS, by centrifugation at 200 × g for 10 min, the cells were resuspended at 3 × 105 cell/ml in DMEM 1 ×, containing 1% FCS.
Ninety-six-well polystyrene cell culture plates (Nunc) were coated with
50 µl/well of CgA fraction solutions in PBS (90 min at 37 °C).
After coating, 100-µl aliquots of a 3% (w/v) BSA solution in DMEM
was added to each well and left to incubate for 30 min at 37 °C. The
plates were then washed two times by emptying and filling with 0.9%
(w/v) sodium chloride and filled with 50 µl/well cell suspension.
After 4 h incubation in a 5% CO2 incubator at
37 °C, 200-µl aliquots of DMEM were added to each well and
aspirated using a glass pipette connected to a vacuum pump. The cells
were washed again with DMEM and fixed by adding 100 µl/well of 3%
(w/v) paraformaldehyde (5 min at room temperature). The solution was
then removed, and the fixed cells were stained by adding 50 µl/well
of 0.5% (w/v) crystal violet for 10 min.
This assay was performed using solid
phases coated with various solutions containing collagen type I (10 µg/ml) or type IV (10 µg/ml), or laminin (10 µg/ml), or 1% FCS,
or poly-L-lysine (50 µg/ml) in PBS (90 min at 37 °C).
After coating, the plates were washed and blocked with BSA as described
for assay 1. Cells were resuspended at 3 × 105
cell/ml in DMEM containing 4 mM glutamine, 200 units/ml
penicillin, 200 µg/ml streptomycin, 0.5 µg/ml amphotericin B, 2 mM sodium pyruvate solution (DMEM, 2 ×) and 2% (w/v) BSA.
Twenty-five µl of cell suspension was then added to each well and
mixed with 25 µl of CgA fractions at various concentrations in 0.9%
(w/v) sodium chloride. After 4 h incubation, adherent cells were
washed, fixed, and stained as described for assay 1.
Human CgA was isolated from the heat stable fraction
(HSF) of homogenized pheochromocytomas by immunoaffinity
chromatography, using a Sepharose-bound monoclonal antibody directed
against the CgA-(81-90) epitope (mAb A11) (30). Two sequential
fractions containing immunoreactive CgA material were recovered by
eluting the column with a pH 2.0 buffer (fraction I) and then with PBS (fraction A). Reducing SDS-PAGE and Western blotting analyses were then
performed using an antibody that recognizes an epitope located within
residues 68-70 of CgA (mAb B4E11) (30). These studies showed that both
fractions consist of heterogeneous mixtures of CgA fragments (Fig.
1, panels A and
B).
The effect of fraction I and fraction A on the adhesive properties of
HFSF-132 human foreskin fibroblasts was examined using these fractions
adsorbed onto a solid phase (adhesion assay 1). In the presence of 1%
FCS in the culture medium, the pattern of cell adhesion to fraction I-
or fraction A-coated plates was opposite, being enhanced by fraction A
and inhibited by fraction I (Fig. 2,
left panels). Interestingly, microscopic inspection revealed that in fraction A-coated wells both cell adhesion and spreading were
increased, whereas virtually no cells were found in fraction I-coated
wells (Fig. 2, right panels). In the absence of FCS, HFSF-132 fibroblasts were still able to adhere to fraction A which were
totally unable to adhere to wells coated with BSA or with fraction I. Similar results were observed also with mouse NIH-3T3 fibroblasts,
indicating that these effects are not species-specific (not shown).
The inhibitory effect of fraction I was further tested using a
different type of assay, in the absence of FCS. Plates were coated with
nonsaturating amounts of various adhesive proteins, such as type I or
type IV collagen or fibronectin, further incubated with various amounts
of fraction I, and then blocked with a large excess of BSA. As shown in
Fig. 3, 10 µg/ml fraction I was
sufficient to inhibit NIH-3T3 fibroblast adhesion to these
substrates.
Size-exclusion HPLC of fraction I and fraction A, followed by detection
of CgA antigen and adhesion-promoting activity in the chromatographic
fractions, indicated that the hydrodynamic properties of the
pro-adhesive and anti-adhesive materials were different, being eluted
at Mr 50-150 × 103 and
Mr 300-700 × 103,
respectively (Fig. 4).
To characterize the structural correlates of the adhesive and
anti-adhesive effects, CgA was then produced by recombinant DNA
technology. The cDNA coding for residues 7-439 of human CgA was
cloned and expressed in E. coli. The product was isolated from the heat stable fraction of E. coli extract by
immunoaffinity chromatography on mAb B4E11-agarose, followed by
reverse-phase HPLC (Fig. 5). Again, two
fractions were obtained, named rCgA-fraction A and rCgA-fraction I,
characterized by adhesive and anti-adhesive effects. SDS-PAGE and
Western blot analysis of the products (Fig. 1, panels A and
B) revealed various immunoreactive bands of 70, 60, 45, 32, and 30 kDa and several other <20-kDa bands in both products. The
N-terminal sequence of the 70- and 60-kDa bands was identical to that
of residues 7-20 of CgA (not shown). This pattern suggested that
extensive proteolysis occurred also in rCgA, probably starting from the
C terminus, and that both the isolated fractions are highly
heterogeneous. Since rCgA-fraction A was lacking the 70-kDa band and
was apparently more degraded than rCgA-fraction I, we assume that
proteolysis was a critical event in the generation of the
pro-adhesive activity.
To verify this hypothesis, we carried out various digestions of
rCgA-fraction I with trypsin-agarose, followed by removal of the
immobilized enzyme and analysis of the adhesion activity. Limited
trypsin treatment (5-10 min) was accompanied by a change from
anti-adhesive to pro-adhesive effects, whereas extensive treatment
(1-2 h) completely abolished both activities (Fig.
6). Various bands corresponding to
<20-kDa products were observed by SDS-PAGE and Western blotting after
limited digestion (not shown). These results strengthen the assumption
that increased CgA fragmentation is associated with a change from
anti-adhesive to pro-adhesive activity.
In conclusion, both natural and recombinant CgA behave like precursors
that generate polypeptides endowed with opposite adhesive activities.
To assess the potential role of integrins in
rCgA-fraction A-mediated cell adhesion, we performed competition
experiments using a GRGDS-peptide, commonly used to compete the
interaction with integrin and RGD-containing substrates, and an
anti-
These results indicate that the rCgA-fraction A-mediated cell adhesion
can be inhibited by affecting the Three
N-terminal fragments were prepared by recombinant DNA technology and
peptide synthesis. In particular, we produced (a) the
recombinant NH2-Ser-Thr-Ala-rCgA-(1-78) fragment (VS-1);
(b) the recombinant
NH2-Ser-Thr-Ala-rCgA-(1-115) fragment (VS-2); and
(c) the synthetic CgA-(7-57) peptide.
The homogeneity and authenticity of recombinant products were assessed
by SDS-PAGE, Western blotting, and in the case of VS-2, also by mass
spectrometry.
Reducing SDS-PAGE of the purified VS-1 and VS-2 showed bands of about
15 and 10 kDa (Fig. 1, panels C and D). Moreover,
electrospray ionization mass spectrometry of VS-2 indicated a molecular
mass of 13247.38 ± 0.54 daltons, which differs of only 0.33 daltons from the calculated mass. These results are consistent with the expected sequences of VS-1 and VS-2.
The effects of VS-1 and VS-2 and the synthetic CgA-(7-57) peptide on
the adhesive properties of various cell lines were assessed by two
adhesion assays in which these products were either adsorbed onto a
solid phase (assay 1) or in the liquid phase (assay 2). The cell lines
tested included both CgA-producing and non-producing cell lines. The
results, shown in Fig. 8, can be
summarized as follows. (a) VS-1 and VS-2 adsorbed onto the
solid phase have little or no effect on the adhesion of human HFSF-132
in the absence of FCS in the medium (panel A), while these
polypeptides can significantly increase adhesion and spreading of these
cells in the presence of 1% FCS (panel B). (b)
All soluble polypeptides, including CgA-(7-57), at concentrations
greater than 3 µM, increase human and mouse fibroblast
adhesion to solid phases coated with 1% FCS or with poly-L-lysine, even in the absence of FCS in the liquid
phase (panels C and D, and E and
F). (c) None of the polypeptides are effective on
the adhesion of human CgA-producing SK-N-BE and CHP-134 neuroblastoma
or rat PC12 cells (panels G-N). (d) similar
amounts of control proteins such as BSA, E. coli protein
extract, and invertase had no effects. No effects were observed also
with other control proteins spanning a wide range of isoelectric points
such as the mouse IgG1 B4E11, transferrin, and avidin (not shown). Of
note, FCS was apparently necessary for VS-1 and VS-2 adhesive activity
in assay 1 (Fig. 8, panels A and B) while it was
not necessary for that of fraction A (Fig. 7). Possibly, VS-1 and VS-2
peptides stick to the plastic less efficiently than CgA or these
peptides need a FCS component to work.
The adhesive effects of N-terminal CgA polypeptides were not restricted
to FCS or poly-L-lysine-coated plates. As shown in Fig.
9, the same effects were observed also on
the adhesion of human fibroblasts to plates coated with types I and IV
collagens or laminin but not with BSA.
To assess the capability of soluble N-terminal CgA fragments to
interact with cell membrane components, we compared the adhesion of
human fibroblasts after preincubation with or without VS-2 or
CgA-(7-57) (30 min) and after washing out unbound ligands. As shown in
Fig. 10, even in this case both CgA
fragments were able to promote adhesion to various solid phase
extracellular matrix (ECM) proteins, suggesting that CgA fragments can
interact with cell surface molecules and trigger pro-adhesive
effects.
Various 18-22-mer peptides spanning the entire CgA
sequence were produced by chemical synthesis. These peptides included
CgA 1-20, 25-46, 37-57, 47-68, 68-91, 91-113, 107-130, 130-153,
163-187, 187-210, 222-244, 231-255, 254-275, 275-297, 297-319,
315-337, 331-352, 353-375, 374-396, 395-417, and 416-439. None of
these peptides, employed at 50 µM concentration,
exhibited significant effects on the adhesion of LB6 and HFSF-132
cells. The fact that CgA-(7-57) was sufficient to exert pro-adhesive
effects whereas no effects were observed with the 1-20, 25-46,
37-57, and 47-68 peptides suggests that conformational constraints in
the N-terminal domain of CgA are necessary for activity.
In neuroendocrine cells and in neurons at least two secretory
pathways are active: the regulated pathway, where the secreted products
are concentrated and stored in secretory granules and released in
response to external stimulation, and the constitutive pathway, where
proteins are continuously transported to the cell surface without prior
concentration or storage (34).
Chromogranins/secretogranins are a family of regulated secretory
proteins present in the electron-dense granules of a variety of
endocrine and neuroendocrine cells. These proteins are released to the
extracellular environment together with the co-resident hormones and
reach the bloodstream via the capillaries or the lymph vessels (2).
Although several functions have been proposed for chromogranins, the
intracellular and extracellular functions of these proteins and of
their proteolytic fragments are still unclear.
In this work we have shown that different mixtures of natural and
recombinant chromogranin A fragments (CgA-fraction A and -fraction I)
when adsorbed onto solid phases, can exert different effects on
fibroblast adhesion: pro-adhesive (fraction A) or anti-adhesive (fraction I). These activities are related to antigenic forms with
different hydrodynamic properties, since fraction I was eluted earlier
then fraction A in gel filtration experiments. Moreover, limited
proteolysis of the anti-adhesive fraction is associated with the
appearance of a pro-adhesive effect. We also found that recombinant
N-terminal fragments of CgA corresponding to VS-1 and VS-2 can exert
pro-adhesive effects either when bound to a solid phase or when added
in solution in various fibroblast adhesion assays. These results
suggest that CgA and/or its N-terminal fragments may be involved in the
regulation of cell adhesion.
Analysis of the primary structure of CgA has revealed in the N-terminal
domain the presence of integrin binding motifs and sequence
similarities with other adhesive molecules (Fig.
11). First, an RGD sequence, often
present in ECM proteins involved in adhesive processes, e.g.
fibronectin, vitronectin, and collagen (35), is present at residues
43-45. Second, a KGD sequence, also present in the integrin binding
domain of barbourin, a snake venom disintegrin (36), is present at
residues 9-11. Third, the sequence of the disulfide-loop located
between the KGD and RGD regions (residues 19-37) is characterized by
32% identity and 64% similarity with that of a portion of the
fibronectin type III-9 domain of tenascin, a tumor- and
development-associated ECM protein (37, 38). These sequence
similarities with other molecules involved in controlling adhesion make
it attractive to speculate that the KGD-disulfide loop-RGD region
(residues 8-45) plays a role in the adhesive effects observed with
N-terminal fragments. Interestingly, we found that a synthetic peptide
encompassing this region (peptide 7-57) is sufficient for mediating
fibroblast pro-adhesive effects. The importance of this region in CgA
is also suggested by the high degree of conservation (>82%) among various species (Fig. 11). However, while the KGD sequence is highly conserved within human, mouse, rat, porcine, bovine, and ostrich CgA
(3, 39), the RGD sequence is replaced with QGD in the mouse and rat
(40, 41), arguing against its functional importance. Accordingly, we
found that an RGE-rCgA-(7-439) mutant was still able to induce
anti-adhesive and pro-adhesive effects, after trypsin treatment, just
as the wild type RGD-rCgA-(7-439) (not shown).
Another protein of the granin family, i.e. secretogranin
I/chromogranin B (CgB), has been shown to be a heparin-binding adhesive protein (42). It has been proposed that CgB becomes associated with the
ECM after secretion from endocrine cells or neurons and can thus
mediate local cell-substrate adhesion. However, the adhesive properties
of CgA and CgB are likely to rely on different mechanisms. Indeed,
putative heparin binding sequences are not included in CgA. Although in
a previous study an RGD-peptide was unable to block 3T3 cell adhesion
to CgB (42), here we demonstrate that a similar peptide as well as an
anti- CgA and CgB are often co-stored in the secretory granules of
neuroendocrine cells. Considering that CgA N-terminal fragments were
found to induce adhesion and spreading of fibroblasts on various ECM
proteins, it is possible that CgA fragments cooperatively regulate not
only cell adhesion to ECM proteins but also the adhesive properties of
ECM-bound CgB. Although we were unable to identify the structural
correlates of the anti-adhesive activity of fraction I, the finding
that it can be converted into pro-adhesive material by trypsin
treatment suggests that proteolytic processing of CgA could be critical
for its adhesive functions. Tissue-specific patterns of proteolytic
processing and different fragments thereof have been observed by many
investigators (18-21). Thus, the modulation of both the intra-granular
and/or the extracellular proteases involved in these processes could
represent a mechanism by which secretory cells control the adhesion of
themselves and/or the adjacent cells.
The adhesive effects of CgA were observed with human and mouse
fibroblasts, i.e. with cells that do not secrete CgA. When we added exogenous CgA fragments to SK-N-BE or CHP-134 neuroblastoma cells no pro-adhesive effects were observed. Several explanation are
possible. For instance, it is possible that these cells do not express
the proper receptors for CgA-mediated adhesion. Moreover, considering
that these tumor cells continuously secrete CgA in the culture medium,
as we monitored by enzyme-linked immunosorbent assay, it is possible
that secreted CgA has saturated the receptor/target molecules and
produced a maximum adhesive effect that cannot be further increased by
an extra addition of CgA fragments. Although no clear conclusions can
be drawn on these points, the lack of effects on these cell lines
strengthens the concept that the positive effects observed with
fibroblasts are dependent on specific mechanisms.
What is the physiological relevance of the modulation of fibroblast
adhesion by CgA? In our assay systems, the adhesive and anti-adhesive
effects on fibroblasts occurred in the 1-10 µM range, depending on the assay type. It is well known that CgA can reach a very
high concentration in secretory granules, approaching the millimolar
range (44), and that intact as well as proteolytic fragments can be
released together with the co-stored hormones. For instance, N-terminal
peptides are generated naturally in the adrenal medulla (45), the
parathyroid glands (46), and the endocrine pancreas (47). It is also
known that, after secretion, CgA can reach the bloodstream and that
various forms of the CgA antigen circulate at nanomolar levels (2, 22).
Micromolar concentration of CgA and its fragments are, therefore,
likely reached in the extracellular environment, at least in close
proximity to the secreting cells, and therefore may affect adhesion of
bystander cells. Fibroblasts and fibroblast-like cells are present
within endocrine tissues and the peripheral nervous system and are
known to play a role in the organization of the extracellular matrix, which, in turn, is important not only for the physiological function of
the neuroendocrine cells but also for the development and the tissue
architecture. Since factors that change the adhesion of fibroblasts
markedly change their physiology, the adhesive activity of CgA may be
important for the regulation of neuroendocrine tissue development and
remodeling. One attractive hypothesis is that modulation of
fibroblast-like cell adhesion by CgA plays a role during embryogenesis,
e.g. during axonal path finding. Another possibility is that
the co-release of adhesion-effective molecules together with hormones
represents a mechanism for the regulation of diffusion of the latter
molecules from the site of production to their targets and, thus, for
the regulation of hormone action, at least for large polypeptide
hormones. The CgA-adhesive activity may also have implications in tumor
pathology. Since CgA is abnormally produced by many endocrine and
neuroendocrine tumor tissues (23), it is possible that the adhesive
effects of CgA may play a role in the progression and metastatization
of neuroendocrine tumors.
In conclusion, our findings have revealed that in addition to ECM and
cell surface membrane proteins, which are released by constitutive
secretion, also proteins stored in the secretory vesicles of the
regulated pathway, such as CgA, can play major roles in cell adhesion.
Since CgA is present in high amounts within secretory granules and is
released in bulk after an appropriate stimulus, in contrast to
constitutive secretory proteins that are continuously secreted without
concentration, its effects might be more discretely delimited.
We thank Barbara Colombo and Angelina Sacchi
for excellent technical assistance and Drs. Ivan De Curtis and J. Meldolesi for helpful discussions.
Department of Biomedical Sciences and Human Oncology,
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
Materials and Cell Lines
20 °C until
use. CgA was then purified from the pheochromocytoma heat stable
fraction (HSF) by immunoaffinity chromatography on mAb A11-Sepharose as
follows: 3 mg of mAb A11 was coupled to 0.5 g of activated
CH-Sepharose (Pharmacia Biotech Inc.), using 0.1 M sodium
carbonate, pH 8.0, as coupling buffer (overnight at 4 °C), and 0.1 M Tris-HCl, pH 8.0, as blocking agent (2 h at room
temperature). After column washings (three times with 0.5 M
sodium chloride, 0.1 M Tris-HCl buffer, pH 8.0, and with
0.5 M sodium chloride, 0.1 M sodium acetate buffer, pH 4.0) pheochromocytoma HSF, 1 mg in 1 ml of sodium chloride, sodium phosphate buffer, pH 7.3 (PBS), was loaded onto the mAb A11-agarose column and washed with PBS until the absorbance of the
effluent reached the base line. The column was then eluted with 3 ml of
0.5 M sodium chloride, 0.2 M glycine, pH 2.0 (fraction I), and with 10 ml of PBS (fraction A). The protein content
in the starting material and in the purified fractions was quantified using the "Bio-Rad Protein Assay" kit. After pH neutralization, the
CgA fractions were mixed with 2 volumes of acetone, incubated overnight
at
20 °C, and centrifuged 45 min at 2500 × g. The
pellet was dried using a SpeedVac concentrator, resuspended in water, extensively dialyzed against distilled water (three changes), and kept
at
20 °C.
-mercaptoethanol, 0.02% (w/v) bromphenol blue
and boiled for 5 min prior to electrophoresis. Western blot analysis
was carried out essentially as follows: proteins, after SDS-PAGE, were
electrophoretically transferred to nitrocellulose membranes using 20 mA
for 25 min and 25 mM Tris, 150 mM glycine, 20%
(v/v) methanol, pH 8.9, as transfer buffer. The nitrocellulose
membranes were rinsed twice with PBS and were incubated overnight at
4 °C in PBS containing 1% (w/v) BSA and 3% (w/v) milk. The
membranes were then incubated for 2 h with anti-CgA or anti-CgB
mAbs (2 µg/ml) in PBS containing 1% (w/v) BSA, 3% (w/v) milk, 1%
(v/v) normal goat serum ("binding buffer"). After washing with PBS
containing 0.02% (v/v) Tween 20, the membranes were further incubated
for 2 h with goat anti-mouse IgG horseradish peroxidase conjugate
(1:1000) in binding buffer. After the final wash, the visualization
reaction was carried out with "ECLTM-Western blotting" kit
(Amersham Corp.), based on chemiluminescence of a luminol substrate.
Natural and Recombinant CgA Affect Human Fibroblast
Adhesion
Fig. 1.
SDS-PAGE (panels A and
C) and Western blot analyses (panels B and
D) of natural and recombinant CgA, VS-1, and VS-2 under reducing (+ME) and nonreducing (
ME)
conditions. CgA fraction A and fraction I (lanes a and
b); rCgA-(7-439) fraction A and fraction I (lanes
c and d); VS-2 (lanes e); VS-1 (lanes
f). SDS-PAGE was carried out in a Phast System apparatus
(Pharmacia) using ready-made PhastGels 12.5% (panels A and
B) and PhastGels 20% (panels C and
B).
[View Larger Version of this Image (50K GIF file)]
Fig. 2.
Adhesion of HFSF-132 fibroblasts to solid
phases coated with BSA (panel A), fraction A (panel
B), fraction I (panel C) (left and
right panels), or pheochromocytoma HSF (tiss.
extr.) or FCS (left panels). The assay was
carried out according to the assay 1 protocol (see "Experimental
Procedures"). Bar, 30 µm.
[View Larger Version of this Image (44K GIF file)]
Fig. 3.
Adhesion of mouse NIH-3T3 fibroblasts to
solid phases coated with non-saturating amounts of collagen I (0.3 µg/ml), collagen IV (3 µg/ml), or fibronectin (0.2 µg/ml) and
overcoated with CgA fraction I or with BSA. Microtiter wells were
coated with solutions of each adhesive protein in PBS (50 µl,
1.5 h at 37 °C) and further incubated with CgA fraction A or
BSA at the concentration indicated in the abscissa (50 µl, 1.5 h, 37 °C). The plates were further blocked with 3% BSA in DMEM (200 µl, 0.5 h, 37 °C). NIH-3T3 cells (40,000/well, in DMEM 1 × without FCS) were then added and left to adhere for 3 h at
37 °C, 5% CO2. The plates were then washed and stained
as described for assay 1 (see "Experimental Procedures").
[View Larger Version of this Image (15K GIF file)]
Fig. 4.
Size-exclusion HPLC of fraction A
(upper panel) and fraction I (lower
panel). Size-exclusion HPLC was carried out at room
temperature using a Bio-Sil 250 Guard column joined to a Bio-Sil
SEC-250 column (Bio-Rad) as follows. The column was equilibrated and
eluted with PBS (flow rate 0.6 ml/min). Fractions (0.3 ml) were
collected and stored at 20 °C until analysis. The column was
calibrated using thyroglobulin (670 kDa), IgG (158 kDa), bovine serum
albumin (66 kDa), ovalbumin (44 kDa), myoglobin (17 kDa), and
cyanocobalamin (1.3 kDa) as molecular markers. Fractions were analyzed
by CgA-enzyme-linked immunosorbent assay and adhesion assay 1.
[View Larger Version of this Image (24K GIF file)]
Fig. 5.
Reverse-phase HPLC of rCgA-(7-439).
rCgA-(7-439), purified by immunoaffinity chromatography on mAb
B4E11-agarose, was loaded onto a reverse-phase Pro-RPC10/15 column
(Pharmacia) and eluted as follows: buffer A, 0.1% trifluoroacetic acid
in water, buffer B, 0.1% trifluoroacetic in acetonitrile; 0% B for 20 min; linear gradient 0-70% B for 20 min; flow rate 0.3 ml/h.
Fractions corresponding to A and I bars were
collected and termed rCgA-fraction A and rCgA-fraction I,
respectively.
[View Larger Version of this Image (12K GIF file)]
Fig. 6.
Adhesion of HSFS-132 cells to solid phases
coated with rCgA-fraction I pretreated for various times with
trypsin-agarose. The horizontal line indicates the
adhesion level to BSA-coated solid phase. Trypsin-agarose was prepared
by coupling 1 mg of trypsin to 0.3 g of activated CH-Sepharose
(Pharmacia), using 0.1 M sodium carbonate, pH 8.0, as
coupling buffer (overnight at 4 °C), and 0.1 M Tris-HCl
as blocking agent (2 h at room temperature). After washing (three times
with 0.5 M sodium chloride, 0.1 M Tris-HCl buffer, pH 8.0, and with 0.5 M sodium chloride, 0.1 M sodium acetate buffer, pH 4.0), 0.1 ml of gel was mixed
with 0.1 ml of rCgA-fraction I (0.73 mg/ml) and incubated at room
temperature. The supernatant was withdrawn at various times and tested
by adhesion assay 1.
[View Larger Version of this Image (20K GIF file)]
1-integrin Antiserum Inhibits rCgA-Fraction A-mediated Cell
Adhesion
1 integrin antiserum. As shown in Fig.
7, the adhesion of HFSF-132 fibroblasts to rCgA-fraction A-coated plates, was efficiently competed by GRGDS and
by the anti-
1 antiserum and poorly or not at all by control GRGESP
and BSA. Interestingly, cell adhesion to fraction A was virtually
abolished by fraction I suggesting that these CgA mixtures
contain materials that compete with each other in cell
adhesion.
Fig. 7.
Adhesive effects of rCgA-fraction A in the
presence of soluble competitors. The assay was carried out by
coating plates with rCgA-fraction A (50 µg/ml in PBS, 90 min at
37 °C) and, after washing three times with 0.15 M sodium
chloride, by plating HFSF-132 cells in DMEM 1 × (without FCS) in
the presence of various competitors, including GRGDS peptide, GRGESP
peptide, rCgA-fraction I; BSA (all at 500 µg/ml) and goat anti-1
integrin subunit antiserum (1/100). Cells were incubated for 4 h
at 37 °C (5% CO2) and stained as described for assay
1.
[View Larger Version of this Image (16K GIF file)]
1-integrin function, hinting at a
specific adhesion mechanism. However, we are unable from these data to
speculate whether
1-integrin interact directly or indirectly with
CgA fragments. Considering the long incubation time required for this
adhesion assay (4 h), it is possible that the adhesive activity is
indirect and requires other proteins made by the cells.
Fig. 8.
Effects of CgA N-terminal fragments on
adhesion of human and mouse fibroblasts and other CgA-producing cells,
as measured by assay 1 and assay 2. Assay 1 (panels A
and B), HFSF-132 human fibroblasts in DMEM 1 × containing 1% FCS (panel B) or in DMEM 1 × without
FCS (panel A) were plated in wells coated with various
amounts of CgA N-terminal fragment or with control proteins and
analyzed as described under "Experimental Procedures." Assay 2 (panels C-N), cell suspensions in DMEM 1 × containing
various amounts of soluble proteins (without FCS) were plated for
4 h onto solid phases coated with poly-L-lysine
(panels C, E, G, I, and M) or FCS (panels
D, F, H, L, and N) (see "Experimental Procedures"). Cells tested included human HFSF-132 fibroblasts (panels C
and D), mouse LB6 fibroblasts (panels E and
F), human SK-N-BE neuroblastoma cells (panels G and
H), human CHP-134 neuroblastoma cells (panels I-L), and rat PC-12 pheochromocytoma cells (panels M
and N).
[View Larger Version of this Image (32K GIF file)]
Fig. 9.
Pro-adhesive effects of CgA fragments on
human HSFS-132 fibroblasts as measured by assay 2 (see "Experimental
Procedures") using various substrates. Solid phase proteins are
indicated within the frame of each panel, and the proteins added in the liquid phase are indicated on the abscissa. Protein concentrations in
the liquid phase were as follows: 70 µg/ml VS-2 or VS-1, 200 µg/ml
CgA-(7-57), 500 µg/ml BSA. Panels A-L show the cell
adhesion to the various substrates in the presence of BSA (left
panels) or CgA-(7-57) peptide in the liquid phase, as observed by
microscopy inspection. Solid phases were coated with type I collagen
(A and B), type IV collagen (C and
D), laminin (E and F), FCS
(G and H), and poly-L-lysine
(I-L). Bar, 10 µm.
[View Larger Version of this Image (88K GIF file)]
Fig. 10.
Adhesion of human HSFS-132 fibroblasts to
various solid phase proteins after preincubation with CgA-(7-57),
VS-2, BSA. Cells were preincubated with 200 µg/ml CgA-(7-57),
100 µg/ml VS-2, 200 µg/ml BSA for 30 min at 37 °C, washed once
with 0.15 M sodium chloride, and plated for 4 h at
37 °C in DMEM 1 × containing 1% (w/v) BSA in microtiter
plates precoated with types I and IV collagen, laminin, and FCS as
described in the assay 2 protocol.
[View Larger Version of this Image (31K GIF file)]
Fig. 11.
Comparison of amino acid sequences of human
(h), bovine (b), porcine (p), mouse
(m), rat (r), and ostrich (o)
CgA-(7-57) (3, 13, 39-41, 48).
[View Larger Version of this Image (12K GIF file)]
1 integrin antiserum block HFSF-132 adhesion to recombinant
fraction A. These results suggest that CgA may modulate adhesion
mechanisms by regulating either the binding capacity or the affinity of
integrins for their ligands, possibly via inside-out or outside-in
signals (43). Several hypotheses can be made on how this phenomenon
occurs: for instance, CgA could interact with other adhesive molecules
of the extracellular matrix and thus modulate their interaction with
integrins. Alternatively, CgA could interact with membrane receptors
that trigger regulatory signals that, in turn, affect either
cytoskeletal organization or the interaction between integrins and ECM
proteins. Further work is necessary to verify these hypotheses. The
long incubation time for adhesion assays (4 h) and the FCS requirements
suggest that the adhesive activity is more likely indirect and requires other proteins present in the FCS or made by the cells.
*
This work was supported by a grant from the Associazione
Italiana per la Ricerca sul Cancro (AIRC).The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
Recipient of a fellowship from ICGEB-UNIDO. Present address: CGEB,
Centre for Genetic Engineering and Biotechnology, Havana, Cuba.
**
To whom correspondence should be addressed: DIBIT, San Raffaele
Scientific Institute, via Olgettina 58, 20132 Milan, Italy. Tel.: 39 2 26434802; Fax: 39 2 26434786; E-mail: cortia{at}dibit.hsr.it.
1
The abbreviations used are: CgA, chromogranin A;
CgB, chromogranin B; mAb, monoclonal antibody; HSF, heat stable
fraction; ECM, extracellular matrix; VS-1, vasostatin-1, VS-2,
vasostatin-2; PBS, 0.15 M sodium chloride, 0.05 M sodium phosphate buffer, pH 7.3; BSA, bovine serum
albumin; HPLC, high performance liquid chromatography; FCS, fetal
calf serum; DMEM, Dulbecco's modified Eagle's medium; PAGE,
polyacrylamide gel electrophoresis; Fmoc, N-(9-fluorenyl)methoxycarbonyl.
2
A. Corti, L. Perez-Sanchez, A. Gasparri, F. Curnis, R. Longhi, A. Brandazza, A. G. Siccardi, and A. Sidoli,
manuscript in preparation.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.