GSK3 and PKB/Akt are associated with integrin-mediated regulation of PTHrP, IL-6
and IL-8 expression in FG pancreatic cancer cells
John J. Grzesiak1, Kathy C. Smith2, Douglas W. Burton2, Leonard J. Deftos2and Michael Bouvet1*
1Department of Surgery, University of California, San Diego and Veterans Affairs San Diego Healthcare System,
San Diego, CA, USA
2Department of Medicine (Endocrinology), University of California, San Diego and Veterans Affairs San Diego Healthcare System,
San Diego, CA, USA
We have demonstrated recently that PTHrP is upregulated in
pancreatic adenocarcinoma and that the ECM exerts regulatory
control, at least in part, over PTHrP expression. In our present
study, we examined the potential signaling interactions between
these 2 pathways. Our results demonstrate that, under serum-free
conditions, adhesion of FG pancreatic adenocarcinoma cells on Fn
is mediated by the ?5?1integrin, whereas adhesion to Type I
collagen is mediated by the ?2?1integrin. ?5?1integrin-mediated
adhesion to Fn results in a phenotype that includes a reduction in
cell proliferation, increased E-cadherin localization in cell–cell
contacts, increased ?-catenin localization throughout the cell, in-
hibition of haptokinetic cell migration, and increased expression of
PTHrP, IL-6 and IL-8 relative to ?2?1integrin-mediated adhesion
on Type I collagen. A phosphoprotein immunoblotting screen of
FG pancreatic cancer cells grown on either Fn or Type I collagen
indicates that GSK3 and PKB/Akt are differentially phosphory-
lated on these 2 substrates. These results implicate GSK3 and
PKB/Akt in the integrin-mediated regulation of PTHrP, IL-6 and
IL-8 in pancreatic cancer.
© 2004 Wiley-Liss, Inc.
Key words: ?-catenin; Lef/Tcf; fibronectin; Type I collagen; ECM
Pancreatic adenocarcinoma is a devastating disease character-
ized by the progressive accumulation of genetic mutations includ-
ing those of K-ras, CDKN2A, p53, BRCA2 and SMAD4/DPC4.1
Mutations of the SMAD4 gene in particular, result in the consti-
tutive activation of TGF? signaling and are thought to be respon-
sible for the extensive proliferation of stromal fibroblasts and
deposition of ECM (desmoplasia) that are hallmarks of this dis-
ease.1The ECM, interacting with cells through cell adhesion
receptors, in particular the integrins, has been shown to be a
critical regulator of important cell processes including angiogen-
esis, mitogenesis, migration and differentiation.2,3We have dem-
onstrated recently that specific ECM proteins, known to be aber-
rantly expressed in pancreatic adenocarcinoma, affect the
expression of the PTHrP axis.4
PTHrP is an onco-fetal protein expressed in normal tissues as
well as many malignancies, including pancreatic, breast, colon,
gastric, melanoma, lung and prostate cancers.5–11PTHrP seems to
play a role in cell growth, proliferation and angiogenesis,12–15
functioning primarily, but not exclusively, through a member of
the 7 membrane spanning and G protein-coupled cell surface
receptor family, the PTHrPR.16,17
Fn is an ECM protein found normally with a relatively even
distribution in the basement membrane of pancreatic ductal
cells.18–24In pancreatic cancer, Fn expression becomes down-
regulated in the basement membrane and sparsely expressed
throughout the ECM.19Our previous results demonstrated that
growth of FG pancreatic adenocarcinoma cells on Fn resulted in
a relative decrease in proliferation and increased expression of
PTHrP and the PTHrPR compared to Type I collagen after 96 hr
in serum-free culture.4Type I collagen, which is normally
expressed in the interstitium, is profoundly upregulated in pan-
creatic cancer.18–21,23,24The differential effects mediated by
these 2 ECM substrates were also characterized by morpholog-
ical changes. Whereas growth of FG cells on Fn substrates was
accompanied by “strong” cell–cell contacts, growth on Type I
collagen resulted in “loose” cell–cell contacts.4
We extend our initial observations by examining the potential
molecular mechanisms responsible for the regulation of PTHrP
expression on these 2 important ECM components in pancreatic
cancer. Our results indicate that GSK3 and PKB/Akt intracellular
signaling is associated with the integrin-mediated regulation of
PTHrP, IL-6 and IL-8 in FG pancreatic cancer cells.
Material and methods
FG cells are a fast-growing (FG), metastatic variant of the
pancreatic adenocarcinoma cell line, COLO-357. FG cells were
used for these studies because their integrin profile is known,25,26
they express relatively high levels of PTHrP (unpublished obser-
vations) and they were used in our previous studies.4Cells were
cultured in DMEM supplemented with 10% FBS in a humidified
atmosphere containing 5% CO2at 37°C.
Cell culture assays
FG cells were serum-starved 24 hr before assay in DMEM
supplemented with 1 mg/ml BSA (Sigma, St. Louis, MO). Six-
well culture plates (Becton Dickinson, Franklin Lakes, NJ), not
treated for tissue culture, were coated with either Fn (1–48 ?g/ml)
or Type I collagen (3–25 ?g/ml) in PBS for 24 hr at RT (Chemi-
con, Temecula, CA). These ECM coating concentrations have
been shown previously to promote adhesion or migration of FG
cells.4,25,26ECM-coated wells were then washed twice with PBS
and blocked with 1 mg/ml BSA in PBS for 1 hr at 37°C. Cells were
trypsinized, treated with 1 mg/ml soybean trypsin inhibitor
(Sigma), followed by 3 PBS washes. Cells were resuspended in
serum-free medium and seeded at 2.5 ? 105/ml, 2 ml/well. Cul-
tures were then incubated and harvested at the indicated time
Inhibition of cell attachment assays
Inhibition of attachment assays were carried out as described
previously.27Briefly, 96-well microtiter plates were previously
Abbreviations: AES, amino terminal enhancer of split; Bit1, Bcl-2
inhibitor of transcription; BSA, bovine serum albumin; BRCA2, breast
cancer 2; CDKN2A, cyclin-dependent kinase inhibitor 2A; CDK1, cyclin-
dependent kinase 1; DMEM, Dulbecco’s modified Eagles’ medium; ECM,
extracellular matrix; ELISA, enzyme-linked immunosorbent assay; FBS,
fetal bovine serum; Fn, fibronectin; GSK3, glycogen synthase kinase 3; IL,
interleukin; ILK, integrin-linked kinase; Lef/Tcf, lymphoid enhancer fac-
tor/T-cell factor; mAb, monoclonal antibody; PKB?, protein kinase B?;
PTHrP, parathyroid hormone-related protein; PTHrPR, parathyroid hor-
mone-related protein receptor; Rb, retinoblastoma; RT, room temperature;
SMAD4/DPC4, mother’s against decapentaplegic, deleted in pancreatic
carcinoma, locus 4; TGF?1, transforming growth factor ?1.
Grant sponsor: Department of Veterans Affairs; Grant sponsor: National
Institute of Health; Grant number: DK60588, AR47347.
*Correspondence to: Department of Surgery (112-E), University of
California, Veterans Affairs San Diego Healthcare System, 3350 La Jolla
Village Drive, San Diego, CA 92161. Fax: ?858-552-4352.
Received 18 June 2004; Accepted after revision 15 September 2004
Published online 17 December 2004 in Wiley InterScience (www.
Int. J. Cancer: 114, 522–530 (2005)
© 2004 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
coated with Type I collagen (10 ?g/ml) or Fn (25 ?g/ml) and
unbound sites blocked with 1 mg/ml BSA as described above. FG
cells (5 ? 104) for Type I collagen or 105FG cells for Fn were
added to each well in serum-free DMEM supplemented with 1
mg/ml BSA. Purified monoclonal antibodies were added at a final
concentration of 25 ?g/ml in the serum-free medium described
above. After 1 hr incubation at 37°C, media were removed. At-
tached cells were fixed with 4% paraformaldehyde in PBS and
stained with 0.5% toluidine blue in 3.7% formaldehyde. Stained
cells were subsequently solubilized with 20% SDS and the absor-
bance read at 595 nm.
Immunofluorescence studies were conducted as described pre-
viously.28Briefly, 13 mm glass coverslips were coated with Type
I collagen (10 ?g/ml) or Fn (25 ?g/ml) overnight at RT, and
subsequently blocked with 1 mg/ml BSA in PBS for 1 hr at 37°C.
Serum-starved FG cells were plated on coated coverslips in serum-
free DMEM supplemented with 1 mg/ml BSA at a density of 2.5 ?
105cells/well for 24 hr at 37°C. The cells were then fixed with 4%
paraformaldehyde in PBS for 5 min at RT and permeabilized with
0.1% Triton X-100 in PBS for 10 min at RT. After incubation with
1% normal goat serum/1% BSA in PBS for 10 min at RT, the
coverslips were incubated with anti-E-cadherin or anti-?-catenin
mAbs (Chemicon) at 10 ?g/ml in 1% normal goat serum/1% BSA
in PBS for 30 min at RT. After rinsing with PBS, coverslips were
incubated with fluorescein-isothiocyanate-conjugated secondary
antibody (goat anti-mouse IgG at 1:100 in 1% normal goat se-
rum/1% BSA in PBS) (Jackson ImmunoResearch Laboratories,
Inc., West Grove, PA) for 10 min at RT. Coverslips were rinsed,
mounted and fluorescence microscopy was carried out with an
Olympus BX-60 microscope equipped with a Spot Digital Imaging
package (Diagnostic Instruments, Inc., Burlingame, CA).
Migration assays were conducted using the modified Boyden
chamber as described previously.27Briefly, the chamber consists
of 2 compartments separated by a filter and migration was mea-
sured by counting the number of cells crossing the membrane
through pores of defined size. Lower chambers were filled with
serum-free DMEM (Invitrogen, Carlsbad, CA) supplemented with
1 mg/ml BSA pore polycarbonate membrane filters (Neuro Probe,
Inc., Gaithersburg, MD), that were coated with either Type I
collagen (10–25 ?g/ml) or Fn (3–25 ?g/ml), were then placed on
top of the lower chambers, and the upper chambers were secured
in place. Upper chambers were filled with 5 ? 104FG cells that
were serum-starved 24 hr before assay, in the same media de-
scribed above. Lower chamber final volumes were 27 ?l and the
upper chambers were 50 ?l. The entire apparatus was then incu-
bated for 17–22 hr at 37°C. After the incubation period, the filters
were fixed in methanol and stained with 0.5% toluidine blue in
3.7% formaldehyde. Excess stain was washed away with water, the
attached cells on the upper side of the filters were mechanically
removed using wet, cotton-tipped applicators. The migratory cells
on the underside of the filters were quantitated by counting 4 high-
powered fields (100? magnification) per well using an inverted light
microscope (Olympus BH 2). Photomicrographs were imaged using a
Nikon Eclipse TE 300 inverted light microscope (Nikon Inc.,
Melville, NY) equipped with a Spot Digital Imaging package.
At the indicated time points, media were harvested and cell
extracts were prepared by sonication in lysis buffer containing 0.25
M Tris, pH 7.4, 0.25% Nonidet P-40 and 2 mM EDTA. The
insoluble fractions were pelleted by centrifugation at 16,000g for
15 min and the soluble fractions were transferred to a fresh tube.
Total cell protein was then measured using a modified Bradford
protein assay with BSA as the standard (Bio-Rad Laboratories,
Inc., Hercules, CA). PTHrP was measured with an in-house radio-
immunoassay, using modifications described previously.29Human
PTHrP 1-34 peptide (Bachem, King of Prussia, PA) was used as
the assay standard. Rabbit antisera raised against the peptide was
used in a non-equilibrium immunoassay format. PTHrP 1-86
(Bachem) was used to prepare tracer by chloramine-T-radioiodi-
nation. Lack of cross-reactivity in the assay for at least a 100-fold
excess of peptide was demonstrated for non-corresponding PTHrP
peptides, calcitonin, calcitonin gene-related peptide and human
and rat atrial natriuretic peptide and bone natriuretic peptide. All
samples were assayed in multiple dilutions that paralleled the
corresponding PTHrP standard. The intra- and interassay varia-
tions were 7 and 12%, respectively.30
To assess IL-6 and IL-8 levels, media from the cell culture
experiments described above were assayed using specific sand-
wich ELISA according to manufacturer’s recommendations as
described previously (BioSource International, Camarillo, CA).31
Briefly, 96-well microtiter plates (Immulon 4 HB) were coated
with anti-human IL-6 or IL-8 mouse monoclonal antibodies (0.6
?g/ml) and incubated overnight at 4°C. The captured plates were
blocked for 2 hr and washed 4? with PBS containing 0.1% Tween
20. Serial dilutions of recombinant IL-6 and IL-8 and test samples
were added to the wells followed immediately by the addition of
50 ?l/well of the biotinylated anti-human IL-6 or IL-8 detection
monoclonal antibody (0.4 ?g/ml). After subsequent incubation
with streptavidin ?-galactosidase conjugate (Calbiochem), the flu-
orescent substrate for ?-galactosidase, methyl umbelliferone ga-
lactopyranoside (MUG), was added. The fluorescence signal gen-
erated by the final product, methyl umbelliferone (4 MU), was
detected by reading microtiter plates with filters for 355 nm
FIGURE 1 – FG pancreatic cancer cell adhesion
on Fn and Type I collagen is mediated by the
?5?1and ?2?1integrins, respectively. Inhibition
of FG cell attachment was conducted on Fn (25
?g/ml) or Type I collagen (10 ?g/ml) coated
welts with purified monoclonal antibodies di-
rected against the indicated specific integrin sub-
units at a final concentration of 25 ?g/ml (see
Material and Methods). Black bars ? Type I
collagen; grey bars ? Fn. Data are expressed as
the percent maximum for each substrate and rep-
resent the mean ? SEM of 3 experiments carried
out in triplicate. *p ? 0.05.
ECM REGULATION OF PTHrP, IL-6 AND IL-8 IN FG CELLS
excitation and 460 nm emission wavelengths. The minimal detec-
tion limit for both assays was 3 pg/ml. The cytokine concentrations
were normalized to cellular protein.
Phosphoprotein screening studies
Cell culture assays on Type I collagen (25 ?g/ml) or Fn (3
?g/ml) were conducted using 6-well culture dishes as described
above. At 1 hr and 24 hr time points, cells were harvested by
addition of 0.5 ml lysis buffer (PBS, 2 mM EGTA, 5 mM EDTA,
30 mM sodium fluoride, 40 mM ?-glycerophosphate, pH 7.2, 10
mM sodium orthovanadate, 0.5% Nonidet P-40) per plate, scraping
and collecting cells, transferring to the next well and repeating
until the entire plate was harvested. Lysates were sonicated for 2 ?
15 sec and frozen at ?20°C. Lysates were subsequently thawed,
spun for 30 min at 16,000g and the supernatants transferred to
fresh tubes. After protein content was determined using the Bio-
Rad protein assay according to manufacturer’s instructions, sam-
ples were combined with 4? Laemmli sample buffer under reduc-
ing conditions and adjusted to a final concentration of 1 mg/ml,
followed by boiling for 5 min. Samples were analyzed in a KPSS
1.3 phospho-site screen of 31 known signaling proteins (Kinexus,
Vancouver, Canada). The immunoblotting analyses involved prob-
FIGURE 2 – E-cadherin and ?-catenin immuno-
localization are differentially regulated by the
ECM in FG pancreatic cancer cells. FG cells
were cultured for 24 hr at 37°C on either Type I
collagen (10 ?g/ml) or Fn (25 ?g/ml) coated
glass coverslips, and immunofluorescence studies
for E-cadherin and ?-catenin were conducted
(see Material and Methods). Representative pho-
tomicrographs taken under the exact same con-
ditions for each substrate are shown. Broad ar-
rows indicate ?-catenin immunolocalization in
the cytoplasm on Fn and narrow arrows indicate
differences in nuclear localization of ?-catenin
on the 2 substrates. Secondary antibody only
controls are also shown in the lower panels. This
experiment was conducted twice yielding similar
results. Bar ? 50 ?m.
FIGURE 3 – FG pancreatic cancer cell hap-
tokinetic migration is promoted on Type I
collagen and inhibited on Fn. Eight ?m pore
polycarbonate filters coated with either 25
?g/ml Type I collagen or 3 ?g/ml Fn. These
ECM components were used as substrates to
measure haptokinetic FG cell migration (see
Material and Methods). Photomicrographs are
shown for each substrate. Bar ? 50 ?m. Data
were quantified by counting 4 high-power
fields per well (100? magnification). Data are
expressed as % Max ? SEM, with 100% ?
16.5 ? 2.2 cells/high-power field. Results
from 3 experiments carried out at least in
triplicate are shown. *p ? 0.05.
GRZESIAK ET AL.
ing with mixes of in-house validated primary antibodies from
commercial sources and the application of each mix into a separate
lane of a 20-lane multiblotter as described previously.32Detailed
protocols for the Kinetworks™ analyses can be found at the
Kinexus Bioinformatics website (www.kinexus.ca). Changes in
the intensity of phosphorylation signals between the 2 ECM sub-
strata at 1 hr and 24 hr after plating of FG cells were compared to
T ? 0 control FG cultures prepared after 24 hr culture in serum-
free medium and just before plating on ECM substrates. Quantity
One software from Bio-Rad was used to quantify the data.
Changes ?30% of control were considered significant.
Statistical significance (p ? 0.05) was determined using two-
tailed Student’s t-tests.
FG pancreatic cancer cell adhesion on Fn and Type I collagen
substrates is mediated by the ?5?1and ?2?1integrins,
To determine the specificity of integrin involvement and thus,
the initiation of the signaling cascade upon adhesion of FG pan-
creatic adenocarcinoma cells to Fn and Type I collagen, we con-
ducted inhibition of adhesion assays using function-blocking
monoclonal antibodies directed against specific integrin subunits.
Figure 1 demonstrates that antibodies directed against the ?2and
?1integrin subunits inhibited FG cell attachment on Type I col-
lagen by 69% and 87%, respectively. The antibody directed
against the ?5integrin subunit and a control antibody directed
against the ?v?5integrin had essentially no effect on FG cell
adhesion on Type I collagen. Figure 1 also demonstrates that FG
cell adhesion on Fn was inhibited by anti-?5and anti-?1antibodies
by 90% and 83%, respectively, whereas anti-?2and anti-?v?5
antibodies had no statistically significant effect compared to con-
trol. It has been demonstrated previously that the anti-?5 mono-
clonal (clone P1D6) blocks arg-gly-asp (RGD)-dependent binding
to fibronectin.33These data demonstrate that the signaling cascade
initiated by FG pancreatic cell adhesion on Fn is mediated pre-
dominantly by the RGD-dependent ?5?1integrin, and adhesion on
Type I collagen is mediated predominantly by the ?2?1integrin.
E-cadherin immunolocalization is increased in FG pancreatic
cell–cell contacts on Fn compared to Type I collagen
Previous observations by our laboratory demonstrated that FG
cell adhesion on Type I collagen resulted in “loose” cell–cell
contacts whereas Fn adhesion resulted in “strong” cell–cell con-
tacts,4similar to previous observations demonstrating that Type I
FIGURE 4 – PTHrP, IL-8 and IL-6 are increased on Fn and decreased on Type I collagen in FG pancreatic adenocarcinoma cells. Cell lysates
were prepared and conditioned media from FG cells grown on Fn or Type I collagen in triplicate wells were harvested after 96 hr (see Material
and Methods). Secreted (a) and intracellular (b) PTHrP levels were measured utilizing an in-house radioimmunoassay developed using antibodies
to the amino-terminus (amino acids 1-34) of PTHrP. Conditioned media were also tested for IL-6 (c) and IL-8 (d) secretion using in-house ELISA
(see Material and Methods). Black bars ? Type I collagen, grey bars ? Fn. Results represent the mean ? SEM of triplicate wells from 5
experiments. *p ? 0.05.
ECM REGULATION OF PTHrP, IL-6 AND IL-8 IN FG CELLS
collagen promoted the epithelial-mesenchymal transition in pan-
creatic cancer cells.28,34–36These results suggested that E-cad-
herin-mediated cell–cell adhesion was differentially regulated by
these 2 ECM substrates. To test this hypothesis, immunofluores-
cence studies were conducted to evaluate the localization of E-
cadherin and its cytoplasmic anchor, ?-catenin, in FG cells after 24
hr in culture on Fn vs. Type I collagen. Figure 2 demonstrates that
Fn adhesion resulted in increased E-cadherin localization to cell–
cell contacts relative to Type I collagen, where E-cadherin ap-
peared diffusely expressed throughout the cells. These results are
in agreement with previous studies demonstrating the immunolo-
calization of E-cadherin in epithelium cultured on Type I collagen
or fibronectin using this monoclonal antibody.28,37,38Additionally,
?-catenin expression was apparent throughout the cell on Fn;
including localization in the cytosol and nucleus. Conversely, on
Type I collagen, ?-catenin localized only sporadically to the cell
surface with reduced expression in the nucleus. These data dem-
onstrate an integrin-dependent differential localization of cell–cell
adhesion architecture in FG pancreatic cancer cells.
Haptokinetic FG pancreatic cell migration is promoted on Type
I collagen and inhibited on Fn
We hypothesized that the relative differences in E-cadherin-medi-
ated cell–cell adhesion observed when FG cells were cultured on Fn
vs. Type I collagen might also affect their integrin-mediated migra-
tion. In the modified Boyden chamber migration assay, Figure 3
demonstrates clearly that FG cells exhibit haptokinetic migration on
Type I collagen but not on Fn. Even when Fn coating concentrations
were raised to 25 ?g/ml, no haptokinetic FG cell migration was
observed (data not shown). These data collectively demonstrate that
integrin-dependent adhesion of FG pancreatic cells on Fn vs. Type I
collagen results in very different phenotypes; Fn adhesion yields a
slower-growing, non-migratory phenotype, whereas Type I collagen
adhesion yields a faster-growing, migratory phenotype.
PTHrP, IL-8 and IL-6 expression in FG pancreatic cancer cells
are upregulated on Fn compared to Type I collagen
We have demonstrated previously that Type I collagen pro-
moted increased FG cell proliferation and decreased PTHrP ex-
FIGURE 5 – Kinetworks™ KPSS 1.3 phosphoprotein analysis of the phosphorylation states of known signaling proteins from FG pancreatic
cancer cells 1 hr and 24 hr after culture on Fn vs. Type I collagen. (a) Immunoblot analysis of the phosphorylation state of known
phosphoproteins 1 hr and 24 hr after plating FG pancreatic cancer cells on Fn vs. Type I collagen. Cell lysates were prepared and analyzed (see
Material and Methods). Cell lysates were electrophoresed into multiple lanes and immunoblotted with different panels of pre-validated
antibodies. Numbered bands shown in (a) correspond to those shown in table form in (b). (b) Percent increase or decrease in the phosphorylation
of the indicated phosphoproteins on Fn vs. Type I collagen at 1 hr and 24 hr compared to T ? 0 control lysates prepared just before plating are
shown. Grey shading indicates differences in the phosphorylation states of known phosphoproteins in response to integrin-mediated adhesion
on Type I collagen or Fn. For example, gray shading of the boxes for GSK3?(S9) show that at 1 hr, phosphorylation is decreased by 40% on
Fn, but increased by 61% on Type I collagen compared to T ? 0 control cultures. This indicates a 101% difference in phosphorylation between
the 2 substrates. The remainder of the gray-shaded boxes should be evaluated in a similar manner.
GRZESIAK ET AL.
pression relative to Fn, where proliferation was decreased and
PTHrP expression was increased after 96 hr in culture. Figure 4a
extends our initial observations by demonstrating a nearly 2-fold
and statistically significant (p ? 0.05) difference in PTHrP secre-
tion after 96 hr on the 2 substrates. Growth of FG cells on Fn also
resulted in elevated and statistically significant differences (p ?
0.05) in intracellular PTHrP expression relative to Type I collagen
(Fig. 4b). To determine whether the increase in PTHrP expression
observed on Fn was dependent on the concentration of the sub-
strate, we coated wells with Fn concentrations ranging from 0–48
FIGURE 5 – CONTINUED.
ECM REGULATION OF PTHrP, IL-6 AND IL-8 IN FG CELLS
?g/ml and found that, although PTHrP expression was promoted
on all Fn concentrations tested, maximal PTHrP expression oc-
curred between 3–6 ?g/ml (data not shown).
It has been demonstrated previously that increased IL-6 produc-
tion in pancreatic cancer cell lines increases liver metastasis for-
mation in nude mice and that both IL-6 and IL-8 are increased in
response to endotoxin or TNF?.39–41We were interested to know
whether the ECM might also play a regulatory role in IL-6 and
IL-8 expression in FG pancreatic cancer cells. Using the FG cell
conditioned media from the studies discussed above, IL-6 (Fig. 4c)
and IL-8 (Fig. 4d) were also upregulated on Fn compared to Type
I collagen, with significant differences noted for both cytokines.
These results collectively demonstrate that ?2?1integrin-mediated
attachment and growth of FG pancreatic cancer cells on Type I
collagen substrates results in decreased expression of PTHrP, IL-6
and IL-8, compared to ?5?1integrin-mediated attachment and
growth on Fn, where PTHrP, IL-6 and IL-8 are increased.
GSK3 and PKB/Akt are differentially phosphorylated in FG
pancreatic cancer cells cultured on Fn vs. Type I collagen
To begin to understand how integrin-mediated signaling regu-
lates de novo PTHrP, IL-6 and IL-8 expression, we conducted
phosphoprotein expression profiling studies of FG pancreatic can-
cer cells 1 hr and 24 hr after plating on Fn vs. Type I collagen.
Figure 5 shows the results of that screen. Notable differences in the
phosphorylation of GSK3 between the 2 substrates at both time
points are apparent. Specifically, FG cell adhesion on Type I
collagen resulted in increased phosphorylation of GSK3?(S9) by
61% at 1 hr and 297% at 24 hr compared to T ? 0 control lysates
prepared at the time of initial plating. Fn adhesion, by contrast,
resulted in decreased phosphorylation of GSK3?(S9) by 40% at 1
hr and 49% at 24 hr compared to T ? 0 control. The phosphory-
lation of GSK3?(S21) was decreased by 38% after 1 hr and
increased by 22% after 24 hr of adhesion on Fn, compared to T ?
0 control, whereas FG adhesion to Type I collagen resulted in
essentially no change after 1 hr and an increase of 91% after 24 hr
compared to control. Although essentially no differences were
noted on either substrate after 1 hr compared to T ? 0 control, the
phosphorylation of GSK3?(Y216) was increased by 169% on Fn
after 24 hr compared to a 53% increase on Type I collagen at the
same time point. Similar results were noted for the phosphoryla-
tion state of GSK3?(Y279), with essentially no differences noted
on either substrate at 1 hr compared to T ? 0 control, whereas a
235% increase was noted on Fn and a 110% increase was observed
on Type I collagen compared to control after 24 hr.
Differences in the phosphorylation state of PKB?(S473) were
also observed on the 2 ECM substrates. Fn adhesion elicited a 68%
and 22% decrease after 1 hr and 24 hr, respectively, whereas Type
I collagen adhesion resulted in a 22% decrease at 1 hr and a 42%
increase in the phosphorylation of PKB?(S473) after 24 hr com-
pared to control. CDK1(Y15), Src(Y418), Rb(S780 and S807/
811), adducin ?(S724), and adducin ?(S662) were also increased
on Type I collagen after 24 hr compared to Fn.
Some similarities were also observed between the 2 substrates at
both time points. Notably, STAT3, MEK1/2, CREB, ERK1 and
ERK2 were all heavily dephosphorylated, whereas p70 S6K was
significantly phosphorylated on both substrates at both the 1 hr and
24 hr time points.
In the present study, we extend our understanding of the rela-
tionship between the ECM and PTHrP expression in pancreatic
cancer. ?5?1integrin-mediated signaling in FG pancreatic adeno-
carcinoma cells upon adhesion to Fn resulted in a phenotype that
included a reduction in cell proliferation, increased E-cadherin-
mediated cell–cell adhesion, increased ?-catenin expression
throughout the cell, inhibition of haptokinetic cell migration, and
increased expression of PTHrP, IL-6 and IL-8 compared to the
signaling initiated by ?2?1integrin-mediated adhesion to Type I
The desmoplastic reaction associated with pancreatic adenocar-
cinoma, including changes in Type I collagen and fibronectin
expression, has been appreciated for some time now.18–24Recent
analyses using microarray and serial analysis of gene expression
(SAGE) technology, confirm those initial findings and point out
the strong expression of many other extracellular matrix genes in
pancreatic cancer as well.42,43There is growing evidence that the
desmoplastic response associated with pancreatic cancer is repre-
sentative of dysregulated normal injury repair processes that in-
clude TGF?1-stimulated expression of collagen and fibronectin.44
In fact, it has been shown recently that TGF?1 increases the
desmoplastic response in pancreatic cancer cells, including up
regulated fibronectin and Type I collagen expression.45It is clear
that these 2 ECM components play crucial roles in pancreatic
cancer, and it will be important to determine their contributions to
the growth, progression and metastases of pancreatic cancer.
An intriguing aspect of our signaling studies was the differential
phosphorylation of GSK3, in particular, on these 2 ECM compo-
nents. It has been shown previously that ?2?1integrin-mediated
adhesion results in the proteolytic breakdown of E-cadherin-me-
diated cell–cell adhesion and an increase in migration on Type I
collagen in HaCaT epithelial cells as well as pancreatic cancer
cells, suggesting that there exists a competitive interplay between
?2?1integrin-mediated adhesion on Type I collagen and E-cad-
herin-mediated cell–cell adhesion.28,35,38Besides its function in
the stabilization of E-cadherin-mediated cell–cell adherens junc-
tions on the cytoplasmic side of the plasma membrane,37?-catenin
has also been shown to play a critical role in intracellular signal-
ing,46where GSK3 is a key component in the ?-catenin signaling
Phosphorylation of ?-catenin by GSK3 leads to its dissociation
from the adherens complex and transfer of the protein to the cytosol,
where it exists in a soluble, monomeric state.46Cytosolic ?-catenin
may be subsequently degraded or translocate to the nucleus, where it
binds with a member of the Tcf/Lef family of transcription factors to
form a complex that activates target genes by binding to their pro-
moter regions.46,48The E-cadherin, Fn and IL-8 genes all contain
consensus Lef/Tcf binding domains in their promoter regions, and
have been shown previously to be direct targets of ?-catenin/Lef/Tcf
transcriptional activator complexes.49–51
The consensus Lef/Tcf DNA binding sequence is (A/T)(A/
T)CAAAGG.50–52Interestingly, the PTHrP gene also contains a
potential Lef/Tcf DNA binding sequence between exons 1A and
1B.52In support of the involvement of Lef/Tcf and ?-catenin in
PTHrP expression, PTHrP signaling has been shown to regulate
the epidermal and mesenchymal expression of Lef-1 and ?-catenin
during embryonic breast development.53This indicates that these
changes in cell fate involve an interaction between the PTHrP and
?-catenin signaling pathways.
In light of our results, it seems that Fn could be promoting a
slower-growing, non-migratory phenotype, at least in part, by
activation of Lef/Tcf target genes that seem to be upregulated in
FG pancreatic cancer cells grown on Fn. Type I collagen, in
contrast, may promote a faster-growing, migratory phenotype, at
least in part, by downregulation of cell–cell adhesion architecture,
leading to the downregulation of Lef/Tcf target genes. It is also
noteworthy that the PTHrPR has been localized to the basolateral
membrane in rat enterocytes; an observation that also correlates
with the downregulation of PTHrP observed when cell–cell con-
tacts are down regulated on Type I collagen substrates.54
Changes in the expression and phosphorylation state of GSK3,
PKB? and CDK1 also suggest a role for integrin-linked kinase
(ILK) in the integrin-mediated regulation of PTHrP and cytokine
expression in pancreatic cancer, as ILK is a crucial upstream
regulator of PKB?, GSK3 and cyclin D1 expression.55In vivo
studies have shown that overexpression or constitutive activation
GRZESIAK ET AL.
of ILK results in oncogenic transformation and progression to an
invasive and metastatic phenotype via this pathway.55Increased
cell migration and phosphorylation of PKB?, GSK3 and CDK1
suggest that binding of the Type I collagen-binding integrin, ?2?1,
could promote an invasive/metastatic phenotype via the ILK sig-
naling cascade in pancreatic cancer.
Differences in the activities of other phosphoproteins were also
found in our studies (Fig. 5). Rb is a marker for proliferation
whose phosphorylation is increased on both substrates, but is
markedly higher on Type I collagen. Phosphorylation of Rb results
in the progression of cell cycle from G1-S phase.56These results
provide independent confirmation of our previous results showing
that FG cells proliferate better on Type I collagen than Fn.4
Adducin is a membrane skeletal protein localized at cell–cell
contacts in epithelial cells that functions to bind calmodulin and in
the assembly of the spectrin-actin network.57The relative increase
in the phosphorylation of both adducin isoforms on Type I colla-
gen and the resultant loss of activity are consistent with a role for
?2?1in the Type I collagen-dependent downregulation of cell–cell
adhesion architecture observed in our studies.
In the present study, we extend our understanding and ap-
preciation of the importance of desmoplasia in pancreatic can-
cer. During desmoplasia, Type I collagen is upregulated and the
basement membrane becomes discontinuous and irregular; this
exposes the primary tumor mass to Type I collagen in the
underlying interstitium.18,19Our results indicate that Type I
collagen can, in the absence of any other exogenous factors,
downregulate the cell–cell adhesion architecture and stimulate
proliferation and migration. These observations indicate that
?2?1integrin-mediated adhesion to Type I collagen could be
playing a critical role in the growth, progression and metastasis
of pancreatic cancer. In contrast, ?5?1integrin-mediated adhe-
sion to Fn results in upregulated cell–cell adhesion architecture,
reduced proliferation, inhibition of cell migration and poten-
tially, the upregulation of Lef/Tcf target genes. These genes
may participate in the survival of pancreatic cancer cells that
interact with Fn at reduced expression levels in disrupted and
irregular basement membranes.18,19In support of this postulate,
a recent study indicates that Fn participates in cell survival, at
least in part, by counter-acting the apoptotic effect of Bit1 and
AES.58Interestingly, AES has been shown to interact with
Tcf-4, resulting in repression of pituitary growth.59Fn may also
be important to the survival of distant metastases, as lymph
nodes are a primary target for pancreatic adenocarcinoma and
express high levels of a multimerized, insoluble Fn that seems
to enhance adhesiveness and reduce cell migration.60,61
In conclusion, integrin-mediated adhesion to the ECM can
elicit profound phenotypic changes in pancreatic cancer cells,
resulting in a relatively slower growing, non-migratory pheno-
type (Fn), or increased proliferation and migration (Type I
collagen). Such phenotypic changes seem to be achieved by
integrin-mediated regulation of cell–cell adhesion architecture
and downstream effector molecules, including GSK3 and PKB/
Akt, that are involved in intracellular signaling to Lef/Tcf target
genes, including IL-8, Fn, E-cadherin and, potentially, PTHrP.
Future studies will be aimed at confirmation of these results
utilizing biochemical and functional assays and by determining
the applicability of our observations with other pancreatic can-
cer cell lines and in vivo.
We thank Drs. E. Ruoslahti and K. Vuori for helpful discussions
and critical readings of the manuscript, and D. Cornelius and J.
Pache for assistance in preparation of the figures. Our study was
supported by VA Merit grants from the Department of Veterans
Affairs (M.B. and L.J.D.) and National Institute of Health grants
DK60588 and AR47347 (L.J.D.).
1. Bardeesy N, DePinho RA. Pancreatic cancer biology and genetics.
Nat Rev Cancer 2002;2:897–909.
Juliano RL. Signal transduction by cell adhesion receptors and the
cytoskeleton: functions of integrins, cadherins, selectins, and immu-
noglobulin-superfamily members. Annu Rev Pharmacol Toxicol
Ruoslahti E. RGD and other recognition sequences for integrins.
Annu Rev Cell Dev Biol 1996;12:697–715.
Grzesiak JJ, Clopton P, Chalberg C, Smith K, Burton DW, Silletti S,
Moossa AR, Deftos LJ, Bouvet M. The extracellular matrix differen-
tially regulates the expression of PTHrP and the PTH/PTHrP receptor
in FG pancreatic cancer cells. Pancreas 2004;29:85–92.
Abdeen O, Pandol SJ, Burton DW, Deftos LJ. Parathyroid hormone-
related protein expression in human gastric adenocarcinomas not
associated with hypercalcemia. Am J Gastroenterol 1995;90:1864–7.
Carron JA, Fraser WD, Gallagher JA. PTHrP and the PTH/PTHrP
receptor are co-expressed in human breast and colon tumours. Br J
Yeung SC, Eton O, Burton DW, Deftos LJ, Vassilopoulou-Sellin R,
Gagel RF. Hypercalcemia due to parathyroid hormone-related protein
secretion by melanoma. Horm Res 1998;49:288–91.
Deftos LJ. Granin-A, parathyroid hormone-related protein, and calci-
tonin gene products in neuroendocrine prostate cancer. Prostate Suppl
Bouvet M, Nardin SR, Burton DW, Behling C, Carethers JM, Moossa
AR, Deftos LJ. Human pancreatic adenocarcinomas express para-
thyroid hormone-related protein. J Clin Endocrinol Metab 2001;86:
10. Bouvet M, Nardin SR, Burton DW, Lee NC, Yang M, Wang X,
Baranov E, Behling C, Moossa AR, Hoffman RM, Deftos LJ. Para-
thyroid hormone-related protein as a novel tumor marker in pancreatic
adenocarcinoma. Pancreas 2002;24:284–290.
11. Hastings RH, Burton DW, Quintana RA, Biederman E, Gujral A,
Deftos LJ. Parathyroid hormone-related protein regulates the growth
of orthotopic human lung tumors in athymic mice. Cancer 2001;92:
12. Henderson JE, Amizuka N, Warshawsky H, Biasotto D, Lanske BM,
Goltzman D, Karaplis AC. Nucleolar localization of parathyroid hor-
mone-related peptide enhances survival of chondrocytes under con-
ditions that promote apoptotic cell death. Mol Cell Biol 1995;15:
13. Insogna KL, Stewart AF, Morris CA, Hough LM, Milstone LM,
Centrella M. Native and a synthetic analogue of the malignancy-
associated parathyroid hormone-like protein have in vitro transform-
ing growth factor-like properties. J Clin Invest 1989;83:1057–60.
14. Massfelder T, Dann P, Wu TL, Vasavada R, Helwig JJ, Stewart AF.
Opposing mitogenic and anti-mitogenic actions of parathyroid hor-
mone-related protein in vascular smooth muscle cells: a critical role
for nuclear targeting. Proc Natl Acad Sci USA 1997;94:13630–5.
15. Bakre MM, Zhu Y, Yin H, Burton DW, Terkeltaub R, Deftos LJ,
Varner JA. Parathyroid hormone-related peptide is a naturally occur-
ring, protein kinase A-dependent angiogenesis inhibitor. Nat Med
16. Kronenberg HM, Lanske B, Kovacs CS, Chung UI, Lee K, Segre GV,
Schipani E, Juppner H. Functional analysis of the PTH/PTHrP net-
work of ligands and receptors. Recent Prog Horm Res 1998;53:283–
17. Segre GV. Principles of bone biology. In: Bilezikian JP, Raisy LG,
Rodan FA, eds. New York: Academic Press, 1996. 389–406.
18. Lohr M, Trautmann B, Gottler M, Peters S, Zauner I, Maillet B,
Kloppel G. Human ductal adenocarcinomas of the pancreas express
extracellular matrix proteins. Br J Cancer 1994;69:144–51.
19. Linder S, Castanos-Velez E, von Rosen A, Biberfeld P. Immunohis-
tochemical expression of extracellular matrix proteins and adhesion
molecules in pancreatic carcinoma. Hepatogastroenterology 2001;48:
20. Mollenhauer J, Roether I, Kern HF. Distribution of extracellular
matrix proteins in pancreatic ductal adenocarcinoma and its influence
on tumor cell proliferation in vitro. Pancreas 1987;2:14–24.
21. Kuehn R, Lelkes PI, Bloechle C, Niendorf A, Izbicki JR. Angiogen-
esis, angiogenic growth factors, and cell adhesion molecules are
upregulated in chronic pancreatic diseases: angiogenesis in chronic
pancreatitis and in pancreatic cancer. Pancreas 1999;18:96–103.
22. Shimoyama S, Gansauge F, Gansauge S, Oohara T, Beger HG.
Altered expression of extracellular matrix molecules and their recep-
tors in chronic pancreatitis and pancreatic adenocarcinoma in com-
parison with normal pancreas. Int J Pancreatol 1995;18:227–34.
23. Tani T, Lumme A, Linnala A, Kivilaakso E, Kiviluoto T, Burgeson
ECM REGULATION OF PTHrP, IL-6 AND IL-8 IN FG CELLS
RE, Kangas L, Leivo I, Virtanen I. Pancreatic carcinomas deposit
laminin-5, preferably adhere to laminin-5, and migrate on the newly
deposited basement membrane. Am J Pathol 1997;151:1289–302.
24. Tempia-Caliera AA, Horvath LZ, Zimmermann A, Tihanyi TT, Korc
M, Friess H, Buchler MW. Adhesion molecules in human pancreatic
cancer. J Surg Oncol 2002;79:93–100.
25. Klemke RL, Yebra M, Bayna EM, Cheresh DA. Receptor tyrosine
kinase signaling required for integrin alpha v beta 5-directed cell
motility but not adhesion on vitronectin. J Cell Biol 1994;127:859–
26. Leavesley DI, Ferguson GD, Wayner EA, Cheresh DA. Requirement
of the integrin beta 3 subunit for carcinoma cell spreading or migra-
tion on vitronectin and fibrinogen. J Cell Biol 1992;117:1101–7.
27. Grzesiak JJ, Davis GE, Kirchhofer D, Pierschbacher MD. Regulation
of alpha 2 beta 1-mediated fibroblast migration on Type I collagen by
shifts in the concentrations of extracellular Mg2? and Ca2?. J Cell
28. Grzesiak JJ, Pierschbacher MD. Changes in the concentrations of
extracellular Mg?? and Ca?? down-regulate E-cadherin and up-
regulate alpha 2 beta 1 integrin function, activating keratinocyte
migration on Type I collagen. J Invest Dermatol 1995;104:768–74.
29. Deftos LJ, Burton DW, Brandt DW. Parathyroid hormone-like protein
is a secretory product of atrial myocytes. J Clin Invest 1993;92:727–
30. Deftos LJ, Gazdar AF, Ikeda K, Broadus AE. The parathyroid hor-
mone-related protein associated with malignancy is secreted by neu-
roendocrine tumors. Mol Endocrinol 1989;3:503–8.
31. Gujral A, Burton DW, Terkeltaub R, Deftos LJ. Parathyroid hormone-
related protein induces interleukin 8 production by prostate cancer
cells via a novel intracrine mechanism not mediated by its classical
nuclear localization sequence. Cancer Res 2001;61:2282–8.
32. Zhang H, Shi X, Zhang QJ, Hampong M, Paddon H, Wahyuningsih
D, Pelech S. Nocodazole-induced p53-dependent c-Jun N-terminal
kinase activation reduces apoptosis in human colon carcinoma
HCT116 cells. J Biol Chem 2002;277:43648–58.
33. Wayner EA, Carter WG. Identification of multiple cell adhesion
receptors for collagen and fibronectin in human fibrosarcoma cells
possessing unique alpha and common beta subunits. J Cell Biol
34. Grisanti S, Guidry C. Transdifferentiation of retinal pigment epithelial
cells from epithelial to mesenchymal phenotype. Invest Ophthalmol
Vis Sci 1995;36:391–405.
35. Hodivala KJ, Watt FM. Evidence that cadherins play a role in the
downregulation of integrin expression that occurs during keratinocyte
terminal differentiation. J Cell Biol 1994;124:589–600.
36. Yi JY, Hur KC, Lee E, Jin YJ, Arteaga CL, Son YS. TGFbeta1-
mediated epithelial to mesenchymal transition is accompanied by
invasion in the SiHa cell line. Eur J Cell Biol 2002;81:457–68.
37. Takeichi M. Cadherin cell adhesion receptors as a morphogenetic
regulator. Science 1991;251:1451–5.
38. Menke A, Philippi C, Vogelmann R, Seidel B, Lutz MP, Adler G,
Wedlich D. Down-regulation of E-cadherin gene expression by col-
lagen Type I and Type III in pancreatic cancer cell lines. Cancer Res
39. Blanchard JA 2nd, Barve S, Joshi-Barve S, Talwalker R, Gates LK Jr.
Cytokine production by CAPAN-1 and CAPAN-2 cell lines. Dig Dis
40. Friess H, Guo XZ, Nan BC, Kleeff O, Buchler MW. Growth factors
and cytokines in pancreatic carcinogenesis. Ann NY Acad Sci 1999;
41. Saito K, Ishikura H, Kishimoto T, Kawarada Y, Yano T, Takahashi T,
Kato H, Yoshiki T. Interleukin-6 produced by pancreatic carcinoma
cells enhances humoral immune responses against tumor cells: a
possible event in tumor regression. Int J Cancer 1998;75:284–9.
42. Iacobuzio-Donahue CA, Maitra A, Olsen M, Lowe AW, van Heek
NT, Rosty C, Walter K, Sato N, Parker A, Ashfaq R, Jaffee E, Ryu B,
et al. Exploration of global gene expression patterns in pancreatic
adenocarcinoma using cDNA microarrays. Am J Pathol 2003;162:
43. Iacobuzio-Donahue CA, Ashfaq R, Maitra A, Adsay NV, Shen-Ong
GL, Berg K, Hollingsworth MA, Cameron JL, Yeo CJ, Kern SE,
Goggins M, Hruban RH. Highly expressed genes in pancreatic ductal
adenocarcinomas: a comprehensive characterization and comparison
of the transcription profiles obtained from three major technologies.
Cancer Res 2003;63:8614–22.
44. Menke A, Adler G. TGFbeta-induced fibrogenesis of the pancreas. Int
J Gastrointest Cancer 2002;31:41–6.
45. Lohr M, Schmidt C, Ringel J, Kluth M, Muller P, Nizze H, Jesnowski
R. Transforming growth factor-beta1 induces desmoplasia in an ex-
perimental model of human pancreatic carcinoma. Cancer Res 2001;
46. Wong NA, Pignatelli M. Beta-catenin—a linchpin in colorectal car-
cinogenesis? Am J Pathol 2002;160:389–401.
47. Doble BW, Woodgett JR. GSK-3: tricks of the trade for a multi-
tasking kinase. J Cell Sci 2003;116:1175–86.
48. Brantjes H, Barker N, van Es J, Clevers H. TCF: Lady Justice casting
the final verdict on the outcome of Wnt signaling. Biol Chem 2002;
49. Gradl D, Kuhl M, Wedlich D. The Wnt/Wg signal transducer beta-
catenin controls fibronectin expression. Mol Cell Biol 1999;19:5576–87.
50. Levy L, Neuveut C, Renard CA, Charneau P, Branchereau S, Gauthier
F, Van Nhieu JT, Cherqui D, Petit-Bertron AF, Mathieu D, Buendia
MA. Transcriptional activation of interleukin-8 by beta-catenin-Tcf4.
J Biol Chem 2002;277:42386–93.
51. Huber O, Korn R, McLaughlin J, Ohsugi M, Herrmann BG, Kemler
R. Nuclear localization of beta-catenin by interaction with transcrip-
tion factor LEF-1. Mech Dev 1996;59:3–10.
52. Mangin M, Ikeda K, Dreyer BE, Broadus AE. Identification of an
up-stream promoter of the human parathyroid hormone-related pep-
tide gene. Mol Endocrinol 1990;4:851–8.
53. Foley J, Dann P, Hong J, Cosgrove J, Dreyer B, Rimm D, Dunbar M,
Philbrick W, Wysolmerski J. Parathyroid hormone-related protein
maintains mammary epithelial fate and triggers nipple skin differen-
tiation during embryonic breast development. Development 2001;
54. Gentili C, Morelli S, de Boland AR. Characterization of PTH/PTHrP
receptor in rat duodenum: effects of ageing. J Cell Biochem 2003;88:
55. Troussard AA, Mawji NM, Ong C, Mui A, St.-Arnaud R, Dedhar S.
Conditional knock-out of integrin-linked kinase demonstrates an es-
sential role in protein kinase B/Akt activation. J Biol Chem 2003;278:
56. Cowgill SM, Muscarella P. The genetics of pancreatic cancer. Am J
57. Kimura K, Fukata Y, Matsuoka Y, Bennett V, Matsuura Y, Okawa K,
Iwamatsu A, Kaibuchi K. Regulation of the association of adducin
with actin filaments by Rho-associated kinase (Rho-kinase) and my-
osin phosphatase. J Biol Chem 1998;273:5542–8.
58. Jan Y, Matter M, Pai JT, Chen YL, Pilch J, Komatsu M, Ong E,
Fukuda M, Ruoslahti E. A mitochondrial protein, Bit1, mediates
apoptosis regulated by integrins and Groucho/TLE corepressors. Cell
59. Brinkmeier ML, Potok MA, Cha KB, Gridley T, Stifani S,
Meeldijk J, Clevers H, Camper SA. TCF and Groucho-related
genes influence pituitary growth and development. Mol Endocrinol
60. Pasqualini R, Bourdoulous S, Koivunen E, Woods VL Jr, Ruoslahti E.
A polymeric form of fibronectin has antimetastatic effects against
multiple tumor types. Nat Med 1996;2:1197–203.
61. Castanos-Velez E, Biberfeld P, Patarroyo M. Extracellular matrix
proteins and integrin receptors in reactive and non-reactive lymph
nodes. Immunology 1995;86:270–8.
GRZESIAK ET AL.