Vol. 81, No. 10
Journal of Clinical Endocrinologyand Metabolism
Copyright0 1996by The EndocrineSociety
Gastrin-Releasing
Present in Human
Membranes*
Prmted
Peptide-Like
Immunoreactivity
Maternal
and Fetal Placental
Q. XIAO, X. HAN, J. R. G. CHALLIS,
K. AKAGI, AND T. J. MCDONALD
D. J. HILL,
E. R. SPINDEL,
in U.S.A.
Is
C. J. PRASAD,
Departments
of Medicine (T.J.M., D.J.H., Q.X.), Pharmacology
and Toxicology (T.J.M.), the Robarts
Research Institute
(T.J.M., Q.X.), Lawson Research Institute (D.J.H.), the University of Western
Ontario, London, N6A 5A5 Canada; the Division of Neuroscience,
Oregon Regional Primate Center
(E.R.S., C.J.P.), Beaverton, Oregon 97006; Department
of Obstetrics and Gynecology, Tohoku
University, School of Medicine (K.A.), Sendai, Japan; Department
of Physiology,
University of Toronto
(J.R.G.C., X.H.), Toronto, M5S lA8 Canada
ABSTRACT
Extracts
of human
term amnion,
placenta,
and chorion/decidual
tissue (n = 5) contained
gastrin-releasing
peptide-like
immunoreactivity (GRPLI)
in amounts
of 4.7 2 2.9 (pmol/g wet wt; mean f SEM),
3.6 2 1.1 and 2.9 + 1.5, respectively.
Using C-terminally
directed
antisera
and gel filtration
chromatography
and reverse-phase
highperformance
liquid
chromatography
(HPLC),
each tissue contained
molecular
forms consistent
with the presence of GRP,.,,
and GRP,,.,,
but also contained
larger
amounts
of two GRPLI
peaks, which apparently
are novel GRP-like
peptides.
In contrast,
tissue extracts
of
human
fetal lung contained
only GRP,-a,,
GRP,,.,,,
and GRP,,-a,.
Using RT-PCR
and specific GRP primers
and probes, messenger
RNA
I
N PREVIOUS STUDIES (1, 2, 3), we have reported the
unexpected presence of gastrin-releasing peptide
(GRP) in ovine maternal and fetal circulations and in fluids
associated with ovine pregnancy. We also reported that
ovine pregnant endometrial epithelium and glands and
myometrial cells contained GRP-like immunoreactivity
(GRPLI) and large amounts of messenger RNA (mRNA)
that encoded for a typical GRP-like peptide (2). These
results were, surprising, as in the adult mammal. GRP is
a neuropeptide that does not appear in significant
amounts in the circulation under physiological conditions
(4). Also surprising was the finding that the GRPLI present
in the ovine plasmas and fluids of pregnancy and in pregnant uterine tissue was of apparently larger molecular size
(1, 3) than the mature neuropeptide GRPi.,,, the largest
molecular form known to be bioactive in mammalian tissues (5).
GRP and the structurally related amphibian peptide,
bombesin, have potent trophic effects on certain maternal
and fetal mammalian tissues; for example, mouse and human fetal lung (6), ovine fetal chondrocytes (7), and human
Received
March 20, 1996.
1996.
Address
correspondence
Donald,
Room 5L2, University
don, Canada, N6A SA.5.
* The present study was
Medical Research Council to
Revised
May
and requests
Hospital,
17, 1996. Accepted
May
22,
for reprints
to: Dr. T. J. Mc339 Windermere
Road, Lon-
supported
by grants from the Canadian
Dr. T. J. McDonald
and Dr. J. R. G. Challis.
3766
encoding
for GRP was readily
demonstrable
from g-weeks
gestation
throughout
pregnancy
to term in full-thickness
membranes,
placental
villi, and decidua.
Positive
immunohistochemical
staining
for GRP
occurred
in extravillous
trophoblasts
in decidual
septa and fetal membranes,
cytotrophoblasts,
syncytiotrophoblast,
and certain
stromal
cells in placental
villi and amniotic
epithelium.
GRPLI
and GRP
messenger
RNA were present
from the earliest
dates examined
(6-9
weeks) throughout
pregnancy
to term. Given the proven
trophic
nature of GRP and related
peptides,
these peptides
may play important
roles in maternal,
placental,
and fetal development
during
human
pregnancy.
(J Clin Endocrinol
Metab
81: 3766-3773, 1996)
endometrium (8). Given these potent effects of GRP on ruminant and nonruminant fetal and maternal tissuesand the
finding that the GRP-like entity apparently appears specifically during ovine pregnancy (2), we hypothesized that GRP
may appear in similar fluids and tissues during primate
pregnancy. This study reports that human placenta and fetal
membranes contain: 1) mRNA encoding for GRP; 2) GRPLI
molecular forms consistent with the presenceof GRP,-,, and
GRP,,.,,; and 3) GRPLI forms that are apparently novel GRPlike entities.
Tissue and fluid
Material
collection
and Methods
At parturition
of normal-term
human
pregnancies,
freshly
obtained
placentae
and membranes
were separated
into chorionidecidua, amnion,
and placenta,
and aliquots
of the tissues were frozen
in liquid
nitrogen
within
5 min of delivery.
The tissues were stored
at -70 C until extraction
for peptide
or RNA analysis.
Portions
of the
tissues were fixed in 100 mL PBS containing
4% paraformaldehyde0.2% glutaraldehyde
for immunohistochemistry.
First-trimester
abortion archival
tissue for immunohistochemistry
was obtained
from Dr.
C. Goodyer
(Montreal
Children’s
Hospital,
Montreal,
Canada).
Blood
from umbilical
veins and arteries
was sampled
simultaneously
by
venipuncture
in 89 patients
at uncomplicated
vaginal
delivery.
The
blood samples
were subjected
to immediate
centrifugation
at 4 C and
the plasma frozen at -70 C until RIA for GRP. All blood samples were
collected
with informed
consent
under
protocols
approved
by the
Human
Ethics Committees
at Tohoku
University,
School of Medicine,
Japan.
GRP IN HUMAN
Tissue extraction
and RIA procedures
Tissues were thawed,
boiled immediately
to inactivate
proteolytic
enzymes
(95 C, 10 min) using 10 vol boiling water per unit wet wt tissue.
The extraction
solutions
were cooled; trifluoracetic
acid (TFA) and acetonitrile
were added to make the solutions
2% and 10% (vol to vol)
respectively.
The tissues in the extraction
solutions
were then homogenized using a hand-held
(OMNI
1000) homogenizer
(Omni
International, Waterbury,
CT) and the extraction
mixture
stirred for 1 hr at 4 C.
The extracted
tissues were removed
by centrifugation
and the extraction
fluids lyophilized.
RIA for GRP was performed
using two antisera, designated
LR-148
and LR-16,
which were raised in rabbits
against
synthetic
porcine
(Bachem,
Torrance,
California);
the characterization
of both
GRP,.,,
antisera has been described
in detail (1, 2, 9). Briefly,
for both antisera,
the assay buffer
was 0.06 mol/L
disodium
phosphate/monosodium
phosphate,
pH 8.0, containing
0.05% (wtlvol)
sodium azide, 0.01 mol/L
EDTA, and 0.3% (wt/vol)
BSA (Sigma, St. Louis, Missouri).
The total
volume
of incubation
was 1 mL. All samples were analyzed
at multiple
dilutions
added to the incubation
mixture
in a volume
of 0.1 mL. The
final dilutions
of the antisera used in RIA were 1:80,000 for the LR-148
and 1:60,000 for the LR-16 antiserum.
Radiolabeled
ligand was prepared
as described
previously
(9) with the exception
that the radiolabeled
ligand was purified
using reverse-phase
HPLC. Under
these RIA conditions,
the LR-148
and LR-16 react equipotently
with GRP,,,
and
GRP,,.,,.
The LR-16 and LR-148 have, respectively,
the following
crossreactivities
with other bombesin
peptide
family members:
with neuromedin B, less than 0.01% and less than 0.3%; with [Se?, Arg’“, Phei3]bombesin
(SAP-bombesin
(lo)), less than 0.1% and 1.9%; and with Phe’abombesin,
12.9% and 7% (10). A 10% drop of the bound/free
ratio from
initial binding
conditions
(no addition
of unlabeled
ligand)
occurred
with the addition
of 1.25 (LR-148)
and 0.625 (LR-16)
fmol GRP,.,,
or
bombesin
per assay tube, and the dose-inhibition
50% level occurred
with 10 (LR-148)
and 7.5 (LR-16) fmol GRP,-,,
or bombesin
added per
assay tube. Bound from free radiolabeled
peptide was separated
using
a charcoal-Dextran
technique
as described
(9). All tissue extract samples
were assayed using the LR-148,
and chromatographic
samples
were
assayed with both the LR-16 and LR-148 antisera. Identical
chromatographic profiles were seen with both antisera.
Chromatography
Gel filtration
chromatography
was performed
using Sephadex
G-50
superfine
(Pharmacia,
Montreal,
P.Q.) columns
(1 X 120 cm), equilibrated and eluted with a 1-mol/L
acetic acid solution
containing
0.15
mol/L
sodium chloride.
The void volumes
(V,) and salt-peak
elution
volumes
(V,) were defined by determining
the elution volumes
of Dextran Blue (Pharmacia,
Montreal,
P.Q.) and sodium
I’*” (Amersham,
Toronto, Ontario).
The elution positions
of human GRP,.,,
and GRP,,-,,
(Bachem, Torrance,
California)
were determined
by chromatography
of
the synthetic
peptides
that were applied
to the columns
in the acetic
acid/sodium
chloride
eluting
solution.
Aliquots
of tissue extract samples (2-4 different
samples of each of the three tissue types) containing
sufficient
GRPLI were loaded onto the Sephadex
G-50 columns
in 1 mL
quantities.
One-milliliter
fractions
were collected
and lyophilized.
For
RIA, each fraction was dissolved
in 1 mL GRP assay buffer and appropriate aliquots
were assayed for GRP.
Before performing
HPLC
procedures,
portions
of the lyophilized
tissue extracts
were dissolved
in 0.5 mol/L
acetic acid and partially
purified
by loading
onto prewashed
Sep-Pak
C-18 cartridges
(Waters
Associates,
Toronto,
Ontario),
as described
(1). The loaded cartridges
were washed
with 0.1% (vol/vol)
TFA (Pierce Chemical
Company,
Rockford,
Illinois)
in water and the peptides
subsequently
eluted with
an acetonitrile
(Fisher Chemical Company,
Toronto, 0ntario):TFA:water
mixture
(80:0.1:19.9;~01).
The peptides
were recovered
from eluates by
lyophilization.
Reverse-phase
HPLC of partially
purified
(Sep-Pak
C-18 cartridge)
chorion/
decidua,
amnion,
and placental
tissue extracts (2-4 different
samples of each of the three tissue types) was performed
using a Beckman System Gold HI’LC
and a Phenomenex
C-18 (3.9 X 300 mm, 10
microns)
reverse-phase
column.
The solvent
systems employed
were
0.12% (vol/vol)
TFA in water (A) and 0.1% (vol/vol)
TFA in acetonitrile
(8); the solvent
flow rate was 1 mL per min. The partially
purified
3767
PREGNANCY
extracts were injected onto the reverse-phase
column and eluted with
sequential
linear gradients
consisting
of: 1) 18-32%
solvent B over 60
min; 2) 32-65% solvent B over 30 min; 3) 65100%
solvent B over 10 min;
and finally,
4) a lo-min
isocratic
elution
period
at 100% solvent
B.
One-minute
fractions
were collected
into tubes containing
0.1 mg BSA
and the solvents
removed
by lyophilization.
Fractions
were reconstituted with GRP assay buffer and appropriate
aliquots taken for RIA. The
columns
were calibrated
with GRP,,.,,,
human
GRP,.,,,
and human
GRP ,4.27 and the elution peaks detected by monitoring
absorbance
at 215
nm. Before experimental
procedures,
the columns
and apparatus
were
thoroughly
washed, and multiple
blank runs were performed
before and
after calibration
procedures
until the blank runs did not contain detectable amounts
of GRPLI.
Because a common
technical
artifact occurs by oxidation
of certain
amino acid residues
in peptides
(ie. methionine),
we established
the
elution
positions
of oxidized
forms of GRP,.,,
and GRP,,.,,
in our
reverse-phase
HPLC systems.
Twenty-microgram
aliquots
of GRP,->,
and GRP,,-,,
were dissolved
in 0.05 mol/L
acetic acid and oxidized
by
incubation
with hydrogen
peroxide.
Peptides were oxidized
for 15,45,
and 60 min using a 0.05% (vol/vol)
hydrogen
peroxide
solution
in water
and under harsher
conditions
with 0.5% (vol/vol)
hydrogen
peroxide
for 15,45, or 60 min. At the end of the incubation,
the reaction mixtures
were diluted 5-fold with water and the peptides
recovered
by lyophilization.
The retention
times of the products
of oxidation
together
with
the nonoxidized
molecules
were determined
using the reverse-phase
HPLC conditions,
as outlined
above, by measuring
absorbancy
at 215
nm. Under
these conditions,
the nonoxidized
forms of GRP,,.,,
and
GRP,.,,
have retention
times (n = 6) of 19.5 2 0.09 min (mean ? SEM)
and 45.1 -t 0.07 min. A very small peak of oxidized
GRP,,.,,
appeared
at 10.8 min, using 0.05% peroxide
for 60 min; this became much larger
using 0.5% peroxide
for 60 min. Very small peaks of oxidized
GRP,-,,
occurred
at 38.4 and 34.9 min, using 0.05% peroxide
for 15 min. Approximately
equal-size
peaks of nonoxidized
GRP,.,,
and oxidized
GRP i-z7 molecules
occurred
at 45.1, 38.4, 34.9, and 27.2 min, using 0.5%
peroxide
for 45 min.
Immunohistochemistry
Samples of chorion/decidua,
amnion,
and placenta,
fixed in a 4%
paraformaldehyde
-0.2% glutaraldehyde
solution,
were embedded
in
paraffin
and sectioned
at a thickness
of 5 pm (11). Immunohistochemistry, using the antiserum
designated
LR-148 at a final dilution
of 1:2,000,
was performed
by the avidin-biotin
peroxidase
method
(Vectastain,
Vector Laboratories,
Burlingame,
California),
as previously
described
(11). The specificity
of the immunohistochemical
staining was tested by
preabsorption
of antibodies
with GRP,.,,
and bombesin
(incubated
with
100 pg/mL
peptide
overnight
at 4 C). Further
controls
included
the
primary
antibody
being replaced
with nonimmune
rabbit serum.
RNA extraction
and RT-PCR
Total RNA was prepared
by homogenization
of tissues in guanidine
thiocyanate,
followed
by centrifugation
through
CsCl (12). Ten pg total
RNA was reverse transcribed
with 25 pmol oligo(dT’s),
200 U of M-MLV
RT (BRL), 5X buffer (250 mmol / L Tris-HCL,
pH 8.3; 375 mmol / L KCL,
15 mmol/L
MgC12,50
mmol/L
DTT, 2.5 mmol/L
dNTP’s) in20 FL total
vol at 37 C for 1 h. Ten PL of RT was used in a lOO-PL PCR reaction, using
100 pmol of the 5’ primer
and 3’ primers
selected to amplify
mRNA
sequences
corresponding
to the human GRP gene entire exon three, as
previously
described
by Li et al (13). The 5’ primer had the sequence(5’
> 3’) GTAGACTCTCTGCTCCAG;
the 3’ primer
had the sequence (5’
> 3’) TCGTCGTAGAAGACCAAA.
PCR conditions
were 1 cycle of 92
C X 2 min, 50 C X 2 min, 72 C X 5 min for second strand synthesis,
followed
by 35 cycles of 92 C X 1 min, 55 C X 1 min, 72 C X 2 min, using
2.5 U of Tuq polymerase
(Promega).
Twenty
PL of the reaction mixture
was separated
on a 1% agarose gel, blotted and hybridized
to an internal
3ZP-end-labeled
oligonucleotide
probe
(sequence
= (5’ > 3’)
TCGTAGTCAAGATGCCT).
Data analysis
GRPLI
concentrations
mean values of GRPLI
in tissue extracts
were calculated
from the
obtained
by assaying each sample in multiple
3768
XL40
dilutions. All values are presented as mean + SEM, pmol/g wet wt. A
paired t test was used to test for a statistically significant difference
between umbilical artery and vein plasma GFS’LI concentrations. A p
value of less than 0.05 was considered significant.
ET AL.
JCE & M . 1996
Vol81
. No 10
1 2 3 4
5 6
7 8 9 10 11 12 1.3 14 15
Results
RIA and RT-PCR
GRPLI was found in each of the human chorion/ decidua,
amnion, and placental tissue extracts. The highest amounts
were found in the amnion extracts, 4.7 + 2.9 (pmol/g tissue
wet wt; mean t SEM; n = 5). The placental extracts contained
3.6 ~fr1.1 (n = 5) and the chorion/decidua extracts 2.9 t 1.5
(n = 5). GRPLI was present in simultaneously sampled umbilical artery and vein blood obtained at uncomplicated vaginal delivery (Fig. 1). There was a significantly higher (P <
0.05)level of GRPLI in the umbilical artery plasmas.RT-PCR
performed on RNA extracted from the three tissue sources
provided readily detectable mRNA encoding for GRP (Fig.
2) in the decidua, fetal membranes, and placental villi at the
earliest time sampled (6 weeks), at midgestation (17 weeks),
and at term.
Gel filtration
chromatography
and HPLC
Figure 3 (A, B, and C) portrays the gel filtration patterns
of the GRPLI present in the chorionldecidua, amnion, and
placental tissue extracts. GRPLI appeared at the elution positions of GRP1-2,and GRP1seZ7;
in a minority of samples, no
peak of immunoreactivity was seen at the GRP,,, elution
position. The GRPLI peak, which occurred close to the
GRP,,,, elution position, was not symmetrical but somewhat broadened, suggesting heterogeneity of immunoreactive entities present (Fig. 3).
Figure 4 (A, B, and C) portrays the reverse-phase HPLC
elution patterns of the GRPLI present in extracts of the three
tissues.A peak of immunoreactivity occurred at the retention
time (45 min) of human GRP,,, and a smaller GRPLI peak
at the elution position of GRP,s-,, (19.5 min). In all samples,
a predominant GRPLI peak occurred at a retention time (29
min) that did not correspond to the retention time of human
GRP14-2,(Fig. 4D) or that of any oxidized product of human
GRP,.,, (see Material and Methods). Further, a major peak of
80
1
20
0
1. Portrays GRPLI concentrations present in simultaneously
sampled umbilical vein and arterial blood obtained during cesarean
sections in 89 patients. The umbilical artery blood samples contained
significantly (P < 0.05) greater elevations of GRPLI than the umbilical vein samples.
FIG.
FIG. 2. Represents agarose gel chromatography
of the RT-PCR reaction mixture. The lane numbering represents: 1, Human fetal lung;
2, term decidua; 3, full thickness term membrane; 4,17-week decidua;
5, 17-week villi; 6, l7-week full-thickness membrane; 7, 16-week
decidua; 8, Is-week villi; 9, IS-week full-thickness membrane; 10,
g-week decidua; 11, 6-week villi; 12, la-week decidua; 13, la-week
villi; 14, RT control; and 15, PCR control.
immunoreactivity occurred at retention time 17 min, consistently earlier than the retention time of GRI’,,,, but later than
that of any oxidation product of GRP18-27.Rechromatography of the GRPLI HPLC peaks at retention times 29 min (Fig.
3D) and 17 min (data not shown) on gel filtration columns
demonstrated that both peaks eluted near the GRP,,.,, molecular size marker. Reverse-phase HPLC of human fetal
lung extracts (Fig. 4D) demonstrated a small GRPLI peak at
the retention time of human GRP,,, and larger GRPLI peaks
at the retention times of GRP,,-,, and human GRP,,.*,; no
GRPLI was seenat retention times 17 or 29 min. In all cases,
the reverse phase HPLC and gel filtration profiles were precisely the samewhether measured by the LR-16 or the LR-148
antibody.
Immunohistochemistry
The immunohistochemical locations of GRPLI in intrauterine tissue from human pregnancy are portrayed in Fig.
5. In early gestation (9.5 weeks), intense cytoplasmic staining
for GRPLI occurred in both cytotrophoblast cells and in the
syncytiotrophoblast of chorionic villi with little staining occurring in the fetal stroma (Fig. 5A). Preabsorption of the
antiserum with GRP,.,, abolished the staining (Fig. 5B). Also
early in pregnancy (6-9.5 weeks), cells in the decidual septa
with the characteristic features of extravillous trophoblasts
and epithelial cells in the maternal glands and amniotic epithelium also stained positively for GRPLI (data not shown).
At term, there was intense positive cytoplasmic staining for
GRPLI in extravillous trophoblasts of decidual septa, which
was greatly reduced but not completely abolished by preabsorption of the antiserum with GR!? (Fig. 5C and D). In
term chorionic villi, the syncytiotrophoblast stained positively for GRPLI but so did a number of different cell types
in the fetal stroma (Fig. 5C); positive staining was abolished
or greatly reduced by preabsorption of antiserum with GRP
(Fig. 5D). In term human fetal membranes,intensely positive
cytoplasmic staining for GRPLI was seenin the extravillous
trophoblasts, which in certain caseshad formed binucleate
cells (Fig. 5E). The term fetal membranes also contained
positive cytoplasmic staining for GRPLI in amniotic epithelial cells, but this was somewhat more variable than the
GRP IN HUMAN
2
2
PREGNANCY
3769
600
NaI ’ 25
2
8
400
200
0
0
10
20
30
40
50
ELUATE
60
70
VOLUME
80
90
100
0
110
(ML)
ELUATE
VOLUME
(ML)
ELUATE
VOLUME
(ML)
GRP, -2,
800
5.
GRP,e-2,
F
a
600
0
4
0
IO
20
30
40
ELUATE
50
60
VOLUME
70
80
90
100
110
(ML)
FIG. 3. Portrays
the gel filtration
patterns
reverse-phase
HPLC GRPLI
peak at retention
with a 1 mol/L acetic acid/O.15
mol/L sodium
and GRP1s-a7 are portrayed
in each panel.
of GRPLI
in tissue
extracts
time 29 min. Chromatography
chloride
solution.
The elution
staining seen in the extravillous trophoblasts. Positive staining for GRPLI was present in fibroblasts of the amniotic
stroma and in certain maternal decidual cells (Fig. 5E). As in
other areas, preabsorption of the antisera greatly reduced or
abolished the immunoreactivity
(Fig. 5F). Tissues obtained at
22 weeks gestation produced results similar to those seen in
early pregnancy and at term (data not shown).
Discussion
Because GRP was first characterized as a neuropeptide
present in the mammalian enteric nervous system, the earliest studies of GRP’s actions emphasized its role as a neurotransmitter/modulator (4). Consistent with GRP being a
neuropeptide, there had been no previous demonstration of
significant levels of GRP in the adult mammalian circulation
under physiological circumstances (4). Our recent demonstration of GRP’s presence in maternal and fetal circulations
during ovine pregnancy (1, 2) was unexpected, and the
present study demonstrates that GRP and novel GRP-like
peptides also are present during human pregnancy.
of: A), chorion/decidua;
B), amnion;
C), placental
villi; and D),
was performed
on Sephadex
G-50 columns
(1 X 120 cm), eluted
positions
of Dextran
Blue (v,), Na1125 (VJ, and synthetic
GRP,.,,
Considerable amounts of GRPLI are present in the chorion/decidua, amnion, and placenta of normal term human
pregnancy. Messenger RNA encoding for human GRP is
demonstrable in the placenta, fetal membranes, and decidual
tissues as early as 6 weeks of gestation and remains present
throughout pregnancy. Gel filtration of the term tissue extracts demonstrates GRPLI peaks eluting at the position of
GRP,_,, and a broad peak eluting at the position of GRP1smZ7.
Reverse-phaseHPLC confirmed the presenceof GRP,,, and
of smaller amounts of GRP,,-,, but the largest GRPLI peaks
occurred at retention times different from known molecular
forms of GRP. Rechromatography of these two apparently
novel GRPLI peaks on gel filtration revealed that they were
of similar molecular size to that of GRP,,-,,. Conceivably, the
earliest eluting GRPLI peak (retention time 17 min) may be
an altered form of the apparently novel GRPLI eluting at 29
min. Hence, although authentic human GRP,.,, and GRP,,,,
are present in tissue extracts of term human maternal and
fetal membranes, the predominant GRPLI peaks present are,
at this time, unidentified.
XL40
3770
2000
ETAL.
JCE & M . 1996
Volt31 . No 10
A)
1
1500:
z
7
GRP, a-2,
human
i
GRP,-,,
4
/..’
1000:
.:’
:
:’
&
500:
.,,
0
0
.. ..
IO
,.....
20
,,......
/.........
30
40
RETENTION
50
60
70
80
TIME (MINUTES)
90
100
1
GRP, a--27
human
0
10
20
human
GRP, --27
GRP,-,,
30
40
50
60
70
80
RETENTION
TIME (MINUTE)
90
100
118
0
10
20
80
30
40
50
60
70
RETENTION
TIME (MINUTE)
90
,r)o
,,o”
FIG. 4. Represents reverse-phase HPLC profiles of GRPLI contained in tissue extracts of: A), chorion/decidua; B), amnion; C), placental villi;
and D), human fetal lung. The elution positions of human GRP,-a,, human GRP,,-a,, and GRP,,,, are indicated by arrows. The bars indicate
the concentrations of GRPLI (PM) and the dotted lines represent the gradient elution conditions. See text for details.
Possibleexplanations of the identities of these potentially
novel GRP-like peptide
include the following. First, both
may prove to be technical artifacts altering the structure of
GRP18-27.However, neither of these unidentified GRPLI
forms appears in extracts of human fetal lung (Fig. 4D) or in
any of the extracted ovine maternal or fetal tissues(3), despite
being subjected to identical procedures. Somewhat less
likely, but possible,these peptides could represent molecular
forms generated by different, but asyet unrecognized, posttranslational processing mechanisms. Second, this peptide
may be a unique peptide having structural similarity to
known forms of bombesin. GRP is not the mammalian analogue of the amphibian skin peptide bombesin (14). Both
bombesin and GRP are present in frogs (14,15); the frog GRP
prohormone has significantly greater homology to the human GRP prohormone than to the bombesin prohormone
(14,16), and distinct bombesin- and GRP-preferring receptor
subtypes are present in frog (17). These findings have led to
the suggestion that GRP and bombesin may have separated
in evolution before the vertebrate radiation (16). Recently, a
unique bombesin receptor subtype has been characterized in
frog brain and designated the BB-4 receptor (18). The BB-4
receptor’s preferred ligand is Phe13-bombesin,a recently
characterized peptide present in frog brain (10). It is possible
that one of the human peptides identified in these studies
may be a bombesin-like peptide, such as Pher3-bombesin.
The recently characterized SAP-bombesin (10) and a neuromedin B-like peptide are not possibilities becausethese entities are not recognized by the GRP antisera used in this
study. Third, based on the recognition of a phyllolitorinpreferring receptor in gastrointestinal and lung neoplasms
(19), and on phyllolitorin’s distinct pharmacological activity
profile (20), there has been speculation as to the existence of
a mammalian analogue of phyllolitorin. It is possiblethat the
novel GRP-like peptides identified in this study may be a
mammalian form of phyllolitorin. Fourth, and lessprobable,
these GRP-like peptides may be unique peptides defining a
fourth subgroup of the bombesin family. We currently are
engaged in the isolation and characterization of these novel
GRP-like peptides to address the above possibilities.
Gorbulev et al. (21) have identified and characterized a
unique bombesin receptor subtype present in the pregnant,
but not the nonpregnant, guinea pig uterus. They noted that
this unique receptor was GRP- rather than NMB-preferring,
GRP IN HUMAN
PREGNANCY
3771
CS
int
FIG. 5. Immunohistochemical
fetal membranes
at term
phoblast;
C, cytotrophoblasts;
staining
for GRPLI
in A, human
placental
villi
(39 weeks).
Panels B, D and F show corresponding
int, intermediate
(extravillous)
cytotrophoblasts;
but that the binding affinity of the new receptor for GRP was
considerably lower than expected. The authors suggested
that the natural ligand for this receptor might be a previously
uncharacterized
peptide member of the bombesin family.
Subsequently, the same receptor was characterized in other
mammalian tissues (22) but was also demonstrated
to be
present in the gravid uterus (23). It is conceivable that the
novel peptides described in this study may be the natural
at 9.5 weeks;
preabsorption
a, amnion;
C, human
controls.
c, chorion;
placenta
at term (39 weeks); E, human
IVS, intervillous
space; S, syncytiotrod, decidua.
The bar represents
60 Frn.
ligands for the newly described BRS3 (also designated BB-3)
receptor subtype.
Messenger RNA encoding for GRP was found to be
present in both maternal and fetal placental tissue in early
pregnancy (6 weeks), in midgestation
(17 weeks), and at
term. Similarly, GRPLI was seen to be present, by immunohistochemistry,
as early as 9.5 weeks and remained present
throughout pregnancy. In early pregnancy, GRPLI is present
JCE &
XIAO ET AL,.
in both cytotrophoblasts
and the syncytiotrophoblast
of placental villi. With increasing age of gestation, decreased
amounts of GRPLI staining seemed to be present in the
cytotrophoblast
layer, consistent with the decrease in prominence of this layer at term. The specific subset of cells in the
decidual septa that stain positively for GRPLI resemble those
that also stain positively with cytokeratin and, hence, have
the characteristics of extravillous trophoblasts. These GRPLIpositive extravillous trophoblasts
are present throughout
pregnancy both in the decidual septa adjacent to placental
villi and in the chorion/decidua
of the fetal membranes.
Cytoplasmic staining for GRPLI in the amniotic epithelium
also was present throughout pregnancy. Later in pregnancy,
positive staining occurred in cells of the fetal stroma. Preabsorption of the GRP antisera with GRP greatly reduced or
abolished positive staining for GRPLI in all tissues. Of interest, the staining of the extravillous trophoblasts
in the
chorion and decidual septa did not disappear completely in
any specimen examined. This result is consistent with the
presence of a novel GRP-like peptide that has structural
homology with, but is not identical to, GRP.
The functions of GRP and the novel GRP-like peptides
present in human pregnancy are unknown. There are a number of possibilities. GRP is mitogenic to human endometrium
(8), and it is possible that GRP, and probably the novel
GRP-like peptides, are trophic to the placenta, maternal tissues and possibly the fetus. Cytotrophoblast
cells and the
syncytiotrophoblast
produce a number of paracrine and hormonal agents, classical growth factors, and prostaglandins,
all of which are thought to be of import for the development
and maintenance of normal pregnancy and in mechanisms
governing parturition (11, 24, 25). GRP administration
into
adult mammals results in the release into the circulation of
a number of neuropeptides and hormones (4). Conceivably,
GRP and GRP-like peptides present in cytotrophoblasts
and
syncytiotrophoblast
act to regulate the synthesis and release
of bioactive agents. Similarly, because extravillous trophoblast cells of the chorion and decidual stroma produce a
number of bioactive factors (11, 24, 25), GRP also may act as
a regulatory agent in these cells. Certain GRPLI positive cells
present in the fetal stroma of placental villi resemble macrophages. There have been a number of reports of specific
GRP/bombesin
receptors being present on mammalian immune competent cells on which GRP/ bombesin has potent
effects (for example, see references 26 and 27). GRP may act
during pregnancy to modulate the function of such cells.
We were not able to obtain maternal tissue containing
myometrium
and, hence, cannot comment as to whether
GRP and GRP-like peptides are present in myometrium during human pregnancy, as is the case in pregnant sheep (1,2,
3). Of interest, GRPLI is present in term cord blood and the
observed gradient of GRPLI concentration suggests a transfer of GRPLI from fetus to placenta or mother; the significance of this gradient is not known at this time. Further, it is
not known whether GRP is present in the fetal circulation at
other times during gestation. In this study, we documented
that human fetal lung contains GRPLI consistent with the
known forms: GRP,.,,, GRP,,.,, and GRP,,,,.
That is, in
contrast to human maternal and placental tissue, human fetal
lung samples contained only peptides consistent with the
Vol81
M . 1996
. No 10
recognized forms of GRP; none of the novel GRP-like forms
were seen. This situation is similar to that seen in the ovine
fetus VS. ovine maternal tissues (3). The significance of this
finding is at present unknown but is probably of functional
importance and requires further investigation.
In summary, the present study has demonstrated
that
tissues associated with human pregnancy contain both GRP
and potentially novel GRP-like peptides, which appear early
in pregnancy and remain present until term. Given the
proven trophic nature of GRP and the close structural similarity of the potentially novel GRP-like peptides to GRP,
there is a strong probability that GRP and these novel peptides play important roles in maternal, placental, and fetal
development and maintenance during human pregnancy.
Acknowledgments
The provision
of archival
pathology
tissue by Dr. C. Goodyer,
the
technical expertise of Ms. Christine
Moogk,
and the secretarial
assistance
of Ms. G. Kellett are gratefully
acknowledged
by the authors.
References
1. Fraser
M, Carter
AM,
Challis
JRG,
McDonald
peptide immunoreactivity
is present
maternal circulations. Endocrinology.
2. Fraser
Gastrin
nology.
M, McDonald
TJ, Spindel
releasing peptide is produced
135:2440-2445.
TJ. 1992
in ovine amniotic
131:2033-2035.
ER, Fahy
M, Hill
in the pregnant
Gastrin releasing
fluid and fetal and
D, Challis
JRG.
1994
ovine uterus. Endocri-
Q, Fraser
M, Wlodek
ME, et al. 1996 Gastrin-releasing
peptide-like
immunoreactive
entities in ovine pregnancy: locations and molecular forms.
Peptides. 17489-495.
McDonald
TJ. 1988 The gastrin releasing polypeptide.
In: Mutt V, ed. Advances in metabolic disorders, gastrointestinal
hormones, vol 11. San Diego:
Academic Press; 199-250.
McDonald,
TJ. 1991 Gastroenteropancreatic
regulatory peptide structures: an
overview. In: Daniel EE, ed. Neuropeptide
function in the gastrointestinal tract.
Boca Raton: CRC Press; 19-86.
Sunday ME, Hua J, Dai HB, Nusrat A, Torday JS. 1990 Bombesin increases
fetal lung growth and maturation in utero and in organ culture. Am J Respir
Cell Mol Biol. 3:199-205.
Hill DJ, McDonald
TJ. 1992 Mitogenic action of gastrin-releasing
polypeptide
on isolated epiphyseal growth plate chondrocytes
from the ovine fetus. Endocrinology.
130:2811-2819.
Endo T, Fukue H, Kanaya
M, et al. 1991 Bombesin and bradykinin
increase
inositol phosphates and cytosolic free Ca2+ and stimulate DNA synthesis in
human endometrial
stromal cells. J Endocrinol. 131:313-318.
3. Xiao
4.
5.
6.
7.
8.
9. McDonald
10.
11.
12.
13.
14.
15.
16.
17.
TJ, Christofi
FL, Brooks
BD, Bamett
W, Cook
MA.
1988 Char-
acterization of content and chromatographic
forms of neuropeptides
in purified nerve varicosities prepared from guinea pig myenteric plexus. Regul
Pept. 21:69-83.
Nagalla
SR, Barry BJ, Falick
AM, et al. 1996 There are three distinct forms of
bombesin: identification
of [Leu13] bombesin, [Phez3] bombesin and [Se?,
Arg”,
I’he13] bombesin,
in the frog Bombina
orientalis.
J Biol Chem.
271:7731-7737.
Sangha RK, Walton
JC, Ensor CM, Tai H-H, Challis
JRG. 1994 Immunohistochemical localization, messenger ribonucleic acid abundance, and activity of
15-hydroxy-prostaglandin
dehydrogenase
in placenta and fetal membranes
during term and preterm labor. J Clin Endocrinol Metab. 78:982-989.
Chirgwin
JM, Przybyla
AE, MacDonald
RJ, Rutter
WJ. 1979 Isolation of
biologically
active ribonucleic
acid from sources enriched in ribonuclease.
Biochemistry.
18:5294-5299.
Li K, Nagalla
SR, Spindel
ER. 1994 A rhesus monkey model to characterize
the role of gastrin-releasing
peptide (GRP) in lung development.
Evidence for
stimulation
of airway growth. J Clin Invest. 94:1605-1615.
Nagalla
SR, Gibson
SW, Tang D, Reeve Jr JR, Spindel
ER. 1992 Gastrinreleasing peptide (GRP) is not mammalian
bombesin. J Biol Chem.
267:6916-6922.
Conlon JM, D’Harte F, Vaudry
H. 1991 Primary structures of the bombesinlike neuropeptides
in frog brain show that bombesin is not the amphibian
gastrin-releasing
peptide. Biochem Biophys Res Commun. 178:526-530.
Spindel
ER, Giladi
E, Segerson TP, Nagalla
S. 1993 Bombesin-like
peptides:
of ligands and receptors. Recent Prog Harm Res. 48:365-391.
Nagalla
SR, Barry BJ, Creswick
K, Spindel
ER. 1994 Cloning of cDNAs
encoding a bombesin-preferring
receptor distinct from the GRP receptor from
the frog Bombina orientalis.
Dig Dis Sci. 39:123 (Abstract 1762).
GRP IN HUMAN
18. Nagalla
SR, Barry BJ, Creswick
KC, Eden P, Taylor
JT, Spindel
ER. 1995
Cloning
of a receptor
for amphibian
[Phel3]
bomb&n
distinct
from the receptor for gastrin-releasing
peptide:
identification
of a fourth
bombesin
receptor subtype
(884). Proc Nat1 Acad Sci USA. 92:6205-6209.
19. Cardona
C, Reeve JG, Bleehen
NM. 1991 Evidence
for the existence
of separate
receptors
for gastrin
releasing
peptide,
neuromedin
B and phyllolitorin
in
human tumour
cell lines. Br J Cancer.
62(Suppl
13):14 (Abstract).
20. Negri L, Improta
G, Broccardo
M, Melchioni
P. 1988 Phyllolitorins:
a new
family of bombesin-like
peptides.
Ann NY Acad Sci. 547~415-428.
21. Gorbulev
V, Akhundova
A, Biichner
H, Fahrenholz
F. 1992 Molecular
cloning
of a new bombesin
receptor
subtype
expressed
in uterus
during
pregnancy.
Eur J Biochem.
208:405-410.
22. Fathi Z, Cojay
MH, Shapira
H, et al. 1993 BRS-3: a novel bombesin
receptor
subtype
selectively
expressed
in testis and lung carcinoma
cells. J Biol Chem.
268:5979-5984.
23. Battey JF, Benya RV, Kroog GS, Sainz E, Weber C, Jensen RT. 1994 Molecular
PREGNANCY
24.
25.
26.
27.
3773
genetic analysis
of mammalian
bombesin
receptors.
Dig Dis Sci. 39:120 (Abstract 1762).
Strauss
III JF, GBfvels
M, King BF. 1995 Placental
hormones.
In: Degroot
LJ,
ed. Endocrinology.
Philadelphia:
W. B. Saunders
Company;
2171-2206.
Irving
JA, Lysiak
JJ, Graham
CH, Heam S, Han VKM,
Lala PK. 1995 Characteristics
of trophoblast
cells migrating
from first trimester
chorionic
villus
explants
and propagated
in culture.
Placenta.
16:413-433.
Van To1 EAF, Kraemer
CVE, Verspaget
HW, Masclee
AAM, Lamers
CBHW.
1991 Intravenous
administration
of bombesin
in man stimulates
natural
killer
cell activity
against
tumour
cells. Neuropeptides.
18:15-21.
De la Fuente M, Del Rio M, Hemanz
A. 1993 Stimulation
of natural
killer and
antibody-dependent
cellular
cytotoxicity
activities
in mouse leukocytes
by
bombesin,
gastrin-releasing
peptide
and neuromedin
C: involvement
of cyclic
AMP, inositol
1,4,5+isphosphate
and protein
kinase C. J Neuroimmunol.
48:143-150.