Uteroglobin Binding Proteins: Regulation of
Cellular Motility and Invasion in Normal
and Cancer Cells
GOPAL C. KUNDU,a,b ZHONGJIAN ZHANG,b GIUDITTA MANTILE-SELVAGGI,b,c
ASIM MANDAL,b CHIUN-JYE YUAN,b,d AND ANIL B. MUKHERJEEb
bSection on Developmental Genetics, Heritable Disorders Branch, National Institute of
Child Health and Human Development, National Institutes of Health,
Bethesda, Maryland 20892-1830, USA
ABSTRACT: Uteroglobin (UG) is a multifunctional, secreted protein with anti-
inflammatory and antichemotactic properties. While its anti-inflammatory
effects, in part, stem from the inhibition of soluble phospholipase A2 (sPLA2)
activity, the mechanism(s) of its antichemotactic effects is not clearly under-
stood. Although specific binding of UG on microsomal and plasma membranes
has been reported recently, how this binding affects cellular function is not
clear. Here, we report that recombinant human UG (hUG) binds to both nor-
mal and cancer cells with high affinity (20–35 nM, respectively) and specificity.
Affinity cross-linking studies revealed that 125I-hUG binds to the NIH 3T3 cell
surface with two proteins of apparent molecular masses of 190 and 49 kDa,
respectively. UG affinity chromatography yielded similar results. While both
the 190- and 49-kDa proteins were expressed in the heart, liver, and spleen, the
lung and trachea expressed only the 190-kDa protein. Some cancer cells (e.g.,
mastocytoma, sarcoma, and lymphoma) expressed both the 190- and 49-kDa
proteins. Further, using functional assays, we found that UG dramatically sup-
pressed the motility and extracellular matrix invasion of both NIH 3T3 and
some cancer cells. In order to further characterize the anti-ECM-invasive
properties of UG, we induced expression of hUG into cancer cell lines derived
from organs that, under physiological circumstances, secrete UG at a high
level. Interestingly, it has been reported that a high percentage of the adeno-
carcinomas arising from the same organs fail to express UG. Our results on
induced hUG expression in these cells show that inhibition of motility and
ECM invasion requires the expression of both UG and its binding proteins.
Taken together, our data define receptor-mediated functions of UG in which
this protein regulates vital cellular functions by both autocrine and paracrine
aPresent address for correspondence: National Center for Cell Science, NCCS Complex, Pune
411007, India. Fax: 91-20-5672259.
cPresent address: Hackensack University Medical Center, 1 Pavilion East, Room 1928,
30 Prospect Avenue, Hackensack, NJ 07601.
dPresent address: Department of Biological Sciences and Technology, National Chiao Tung
University, 70 Po-Ai Street, 300 Hsinchu, Taiwan.
235KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
Cellular migration and extracellular matrix invasion are critical for embryonic
development, wound healing, inflammation, and cancer cell metastasis. However,
the exact molecular mechanisms that regulate these processes are not well under-
stood. Blastokinin1 or uteroglobin (UG)2 is a steroid-dependent, multifunctional,
secreted protein with anti-inflammatory and antichemotactic activities (reviewed in
reference 3). Depending on the status of expression of this protein in different tissues
or body fluids, UG has been identified by a number of names, including Clara cell
10-kDa (CC10) protein,4,5 progesterone-binding protein,6 urine protein-1,7–10
retinol-binding protein,11 and polychlorinated biphenyl–binding protein.12
Structurally, UG is a covalently linked, homodimeric protein in which the 70
amino acid subunits are connected to each other by two disulfide bonds in an anti-
parallel orientation. Each UG monomer forms four α-helices (α-helix 1–4) and a β-
turn between α-helices 2 and 3, but there is no β-structure. Human UG (hUG)13,14
is the counterpart of rabbit UG.15,16 Both rabbit and human UGs are potent inhibi-
tors of low-molecular-weight group I and group II extracellular phospholipase A2
(E.C.188.8.131.52).17–19 Several years ago, Robinson et al. described an active transport
of UG into the blastocoele of preimplanted rabbit embryos by a carrier-mediated
process.20 Moreover, Nieto and coworkers have shown the specific binding of UG in
microsomal and plasma membranes.21
In this paper, we summarize our findings on the characterization of the hUG-
binding proteins demonstrating novel effects of recombinant, homodimeric hUG in
suppressing the ability of NIH 3T3 cells to invade the extracellular matrix (ECM).
This effect of hUG appears to be mediated via its high-affinity binding protein
(putative receptor). This binding protein has a molecular mass of 190 kDa and is ex-
pressed on several cell types including the NIH 3T3 cells. Since UG exists as a
homodimer with two interchain disulfide bonds, we sought to determine whether the
reduced, monomeric form of UG, like its dimer, shows the same biological effect as
well as interacts with the same 190-kDa protein. Interestingly, we discovered that
monomeric UG not only interacts with the 190-kDa protein, but also binds with the
49-kDa protein in several normal (NIH 3T3) and cancer (mastocytoma, sarcoma, and
lymphoma) cells with high affinity and specificity. Moreover, the reduced hUG, like
dimeric hUG, also suppresses the ECM invasion in those normal and cancer cells
that express the UG-binding proteins. The anti-ECM-invasive properties of hUG
were further confirmed by inducing the expression of hUG into several cancer cells
by stable transfection using hUG cDNA. These cell lines were derived from the
adenocarcinomas of the lung, prostate, breast, and uterus, which all (under physio-
logical conditions) secrete UG at a high level, but fail to do so when malignant trans-
formation occurs. The data showed that only one of the cell lines was derived from
an adenocarcinoma of the uterus on which the anti-invasive activity of UG could be
demonstrated. It turned out that this was the only cell line, in which hUG expression
was induced, that also expressed the UG-binding proteins. These results clearly
showed that the anti-invasive effect of hUG is mediated via its putative receptor.
Data on tissue-specific expression indicated that both the 190- and 49-kDa pro-
teins are detectable in the heart, liver, and spleen, whereas only the 190-kDa protein
is present in the lung and trachea. Neither the 190- nor 49-kDa protein was present
236ANNALS NEW YORK ACADEMY OF SCIENCES
in the aorta. Pretreatment of the NIH 3T3 cells with IL-6 and LPS followed by 125I-
UG-binding showed a significant increase in expression of both the 190- and 49-kDa
binding proteins. Purification of these proteins by UG affinity chromatography from
bovine lungs and analysis by SDS-PAGE followed by silver staining identified two
protein bands with molecular masses of 180 and 40 kDa, respectively. Taken togeth-
er, these data demonstrate that UG plays critical roles in regulating the cellular
motility and ECM invasiveness by interacting with these binding proteins.
MATERIALS AND METHODS
NIH 3T3, mouse mastocytoma, sarcoma, lymphoma, fibrosarcoma, and human
adenocarcinoma cell lines HEC-1A (uterus), HTB-174 (lung), HTB-81 (prostate),
and HTB-30 (breast) were obtained from the American Type Culture Collection
(Rockville, MD). All cell culture–grade reagents were purchased from GIBCO-
BRL. These cell lines were grown according to standard procedures. Recombinant
hUG was expressed and purified as described.15 BioCoat Matrigel invasion cham-
bers were obtained from Collaborative Research. Disuccinimidyl suberate (DSS)
was from Pierce.
In Vitro Invasion Assay
The in vitro invasion assay was performed as described earlier.22 Briefly, semi-
confluent cells were harvested with trypsin and EDTA, centrifuged, and washed with
PBS containing 0.1% BSA. The cells were resuspended in either DMEM or Mc-
Coy’s medium containing 0.1% BSA and seeded in the upper compartment of the
prehydrated, Matrigel-coated invasion chamber. The cells were treated in the
absence or presence of nonreduced or reduced hUG and then incubated at 37°C for
24–36 h in a humidified incubator. The lower compartment was filled with fibroblast
conditioned medium, which acted as a chemoattractant. The noninvaded cells and
the Matrigel were scraped and removed from the upper surface of the filter. The in-
vaded cells were fixed with methanol, stained with Giemsa, and washed with PBS.
The invaded cells were counted under an inverted microscope and the percentage of
invaded cells was calculated.
The radioiodination of hUG and the radioreceptor assay were performed as
described earlier.23,24 Briefly, the recombinant hUG was radioiodinated by using
[125I]-NaI (2 mCi; carrier-free) and IODO-BEADS. The 125I-UG was purified by G-
25 column chromatography and the specific activity of the purified, carrier-free 125I-
hUG was about 20–25 µCi/µg. The cells were grown to confluence in 12-well plates
and washed with PBS. The cells were incubated with varying concentrations of non-
reduced or reduced 125I-hUG in the absence or presence of excess unlabeled hUG in
1 mL Hank’s balanced salt solution (HBSS), pH 7.6, containing 0.1% bovine serum
albumin (BSA) at 4°C for 2 h. In some experiments, the cells were incubated with
nonreduced or reduced 125I-hUG in the same HBSS, pH 7.6, in the absence or pres-
237 KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
ence of increasing concentrations of unlabeled hUG under the same conditions as
described above. The reaction was stopped by rapid removal of unbound 125I-hUG.
The cells were washed with PBS, pH 7.4, solubilized in 1 N NaOH followed by
addition of 1 N HCl and the radioactivity was counted by a gamma-counter. The
monomeric form of hUG was prepared by treating the dimeric hUG with 10 mM
dithiothreitol at 37°C for 15 min. The specific binding was calculated by subtracting
the nonspecific binding from the total binding, and the data were analyzed by Scat-
chard plot using the LIGAND computer program.25
Affinity Cross-linking Experiments
The cells were grown in 6-well plates and, when semiconfluent, they were incu-
bated with nonreduced or reduced 125I-hUG in the absence or presence of excess un-
labeled hUG in 2 mL HBSS, pH 7.6, containing 0.1% BSA at 4°C for 2 h. The cells
were washed with PBS, pH 7.4, and incubated further with 0.2 mM DSS in 2.0 mL
HBSS, pH 7.6, for 20 min. The cells were collected by centrifugation (10,000g for
15 min) and lysed in 40 µL of lysis buffer (1% Triton X-100 containing 1 mM PMSF,
20 µg/mL leupeptin, and 2 mM EDTA). The supernatants were resuspended in sam-
ple buffer containing 5% β-mercaptoethanol, boiled, and resolved by electrophoresis
on 4–20% gradient sodium dodecyl sulfate (SDS)–polyacrylamide gel (Bio-Rad).
The gels were stained, dried, and autoradiographed using Kodak X-Omat AR X-ray
The membrane preparations from bovine spleen, trachea, heart, aorta, lung, and
liver were prepared as described before.23 Briefly, the membrane samples containing
equal amounts of total proteins were incubated with reduced 125I-hUG in HBSS,
pH 7.6, containing 0.1% BSA in the absence or presence of unlabeled hUG for bind-
ing as described before. The samples were cross-linked with DSS, solubilized,
electrophoresed, and autoradiographed. To check the regulation of hUG-receptor ex-
pression, the NIH 3T3 cells were treated with various cytokines and other agents.
The treated cells were incubated with 125I-hUG for binding, cross-linked with DSS,
electrophoresed, and autoradiographed.
UG Receptor Purification
The hUG receptor was purified from bovine spleen by hUG affinity chromatog-
raphy. The tissue was homogenized in 10 mM NaHCO3 buffer, pH 8.0, and the
homogenate was centrifuged at 600g for 10 min at 4°C. The supernatant was centri-
fuged at 24,000g for 60 min and the pellets were collected and solubilized in solu-
bilization buffer [50 mM Tris-HCl buffer, pH 7.4, containing 1% Triton X-100,
0.4 mM phenylmethylsulfonyl fluoride (PMSF), 10 mg/mL leupeptin, and 2 mM
EDTA] by stirring at 4°C for 6 h. The supernatant was collected by centrifugation
and loaded to the CNBr-activated Sepharose 4B–coupled hUG affinity column. The
hUG receptor protein was eluted by using 0.1 M glycine-HCl buffer, pH 3.0, con-
taining 0.1% Triton X-100, 0.4 mM PMSF, 10 mg/mL leupeptin, and 2 mM EDTA.
All fractions were neutralized with 2 M Tris-HCl, pH 8.0. The fraction containing
238ANNALS NEW YORK ACADEMY OF SCIENCES
receptor protein was analyzed by 125I-hUG-binding and affinity cross-linking assays
and the purity was checked by SDS-PAGE followed by silver staining.
Construction of pRC/RSV-hUG and Transfection
The full-length hUG cDNA, cloned in pGEM 4Z,15 was excised by EcoRI diges-
tion and subcloned into the TA vector. The orientation of the hUG cDNA construct
in the TA vector was checked by DNA sequencing. The fragment was excised from
the TA vector by using HindIII and XbaI and ligated further into the pRC/RSV
expression vector (Invitrogen) to make the pRC/RSV-hUG construct. The cell lines
derived from the adenocarcinomas of human uterus (HEC-1A), lung (HTB-174),
prostate (HTB-81), and breast (HTB-30) were cultured as described before.26 Both
the pRC/RSV-hUG and pRC/RSV (mock) were individually transfected to each of
the tumor cell lines by electroporation. G418 (400 µg/mL) was added to the cells
after 24 h of the transfection. The stably transfected clones were isolated, expanded,
and used for further experiments.
Detection of hUG mRNA by RT-PCR
In order to detect the level of hUG mRNA in different adenocarcinoma cells, total
RNAs were isolated from the cells by using the RNAzol method according to the
manufacturer’s instructions (Tel Test, TX). Reverse transcription (RT) using equal
amounts of total RNA and cDNA amplification were performed using DNA Thermal
Cycler 480 (Perkin Elmer, Norwalk, CT). The sequences of hUG-specific primers
were as follows—hUG-L: 5′-ATG AAA CTC GCT GTC ACC C-3′; hUG-R: 5′-TAC
ACA GTG AGC TTT GGG C-3′. The sequences of GAPDH-specific primers were
as follows—GAPDH-L: 5′-CCA TGG AGA AGG CTG GGG-3′; GAPDH-R: 5′-
CAA AGT TGT CAT GGA TGA CC-3′. The Gene Amp Thermostable rTth Reverse
Transcriptase PCR kit (Perkin Elmer) was used for the reaction according to the
manufacturer’s instructions. The PCR products were subjected to electrophoresis on
agarose gel, transferred to nylon membranes, and UV cross-linked. The membranes
were hybridized using hUG- and GAPDH-specific probes and autoradiographed.
The sequences of the probes were as follows—hUG-P: 5′-TGA AGA AGC TGG
TGG ACA CC-3′; GAPDH-P: 5′-TCC TGC ACC ACC AAC TGC TT-3′.
Detection of hUG by Immunoprecipitation and Western Blot Analysis
The confluent adenocarcinoma cells were washed with PBS, lysed in lysis buffer
(50 mM Tris-HCl, pH 7.5, containing 150 mM NaCl, 1% Nonidet P-40, 15 µg/mL
leupeptin, and 0.5 mM PMSF), and immunoprecipitated by using a kit according to
the manufacturer’s instructions (Boehringer Mannheim). Briefly, the cell lysates
were centrifuged and the supernatants were incubated with rabbit hUG antibody for
1 h and incubated further with protein A-agarose at 4°C overnight. Bound complexes
were collected by centrifugation, washed, and eluted by boiling in SDS-sample buff-
er. The samples were resolved on SDS-PAGE and transferred to the nitrocellulose
membranes. In case of Western blot analysis, the membranes containing hUG pro-
tein were blocked, washed, and incubated with goat hUG antibody at room temper-
ature for 1 h. The membranes were washed, incubated further with rabbit anti-goat
horseradish peroxidase (HRP)–conjugated IgG, and detected by enhanced chemi-
239KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
luminescence (ECL) according to the manufacturer’s instructions (Amersham
Soft Agar Assay
The anchorage-independent growth of cells on soft agar was performed as
described.26 Briefly, wild-type, pRC/RSV-hUG-, and pRC/RSV (mock)–transfected
adenocarcinoma cells were trypsinized. Usually, transfected cells with 3–4 passages
were used for this assay. These single cell suspensions were stained with neutral red
to check the cell viability. About 10,000 viable cells were suspended in 2.5 mL of
McCoy’s 5A medium containing 10% FBS and 0.3% Noble agar (Difco) and plated
on 60-mm petri dishes. These cells were then layered over the agar-coated petri dish-
es. The medium used to prepare the top agar contained G418 (200 µg/mL) and was
added for both the pRC/RSV-hUG- and pRC/RSV (mock)–transfected cells. These
plates were incubated at 37°C for 3 weeks in an atmosphere of 5% CO2 and 95% air.
The colonies were stained with neutral red, counted, and photomicrographed.
RESULTS AND DISCUSSION
We have previously reported that recombinant, dimeric hUG suppresses ECM
invasion of NIH 3T3 and human trophoblast cells by interacting with a specific
hUG-binding protein (putative receptor).22 Because hUG exists as a homodimer in
nature with two interchain disulfide bonds, we sought to determine whether the re-
duced, monomeric form of hUG also suppresses the ECM invasion in normal and
cancer cells. Accordingly, ECM invasion assays were performed by treating each of
the cell lines (NIH 3T3, mastocytoma, sarcoma, lymphoma, and fibrosarcoma cells)
with the monomeric form of hUG. The data showed that the monomeric hUG (1 µM)
suppresses the ECM invasion in NIH 3T3 (82%), mouse mastocytoma (77%), sarco-
ma (79%), and lymphoma (75%), but there was no such effect when fibrosarcoma
cells were used (FIG. 1a). Myoglobin, a nonspecific protein, was used as a negative
control (FIG. 1a). In order to ascertain the possible roles of hUG in reversing the
transformed phenotype of hUG-transfected adenocarcinoma cells, we tested them
for ECM invasion and anchorage-independent growth on soft agar assays. We select-
ed human cancer cell lines derived from the adenocarcinomas of the uterus, lung,
prostate, and breast. All of these adenocarcinoma cell lines did not express hUG,
although the organs in which these cancers arose constitutively expressed high levels
of hUG. To determine the effects of endogenous hUG production in these cells, we
induced its expression by stably transfecting the cells with an hUG cDNA construct,
pRC/RSV-hUG. The results of ECM invasion using pRC/RSV-hUG-transfected
HEC-1A cells showed a significant suppression of ECM invasion (88%) compared
with nontransfected HEC-1A cells (FIG. 1b). The wild-type HEC-1A cells when
treated with recombinant hUG yielded 79% suppression of ECM invasion (FIG. 1b).
There were no differences in the monolayer growth properties among the hUG-
transfected, mock-transfected, and nontransfected HEC-1A cells. Neither the wild-
type nor the hUG-transfected adenocarcinoma-derived prostate cell lines (HTB-81)
showed any suppression of ECM invasion (data not shown). We have also examined
the levels of hUG mRNA and protein in hUG-transfected HEC-1A and HTB-81 cells
240 ANNALS NEW YORK ACADEMY OF SCIENCES
as compared to nontransfected cells. Both RT-PCR and immunoprecipitation
followed by Western blot analysis indicated that nontransfected (wild-type) cells
neither express the hUG mRNA (FIG. 2a, upper left panel) nor the hUG protein
(FIG. 2b), whereas the hUG-transfected HEC-1A and HTB-81 cells express both the
hUG mRNA (FIG. 2a, upper middle and right panels) and the hUG protein (FIG. 2b),
respectively. The expression of GAPDH, a housekeeping gene, is shown in the lower
panel (FIG. 2a). The soft agar assay data demonstrated that the anchorage-
independent growth of the pRC/RSV-hUG-transfected clones (HEC-1A) on soft
agar was drastically reduced in size compared with the nontransfected or mock-
transfected clones, although the numbers of cells seeded in each soft agar plate were
equivalent (FIGS. 3a–3c). Both the nontransfected (FIG. 3a) and mock-transfected
(FIG. 3b) cells grew a mixture of large and small colonies, whereas the cells trans-
fected with pRC/RSV-hUG (FIG. 3c) showed very small and uniform-type colonies
on soft agar plates after 4 weeks of incubation. The pRC/RSV-hUG-transfected
colonies (HEC-1A) were transferred to the monolayer culture and these cells grew
equally well as compared to nontransfected or mock-transfected cells, which indi-
cates that the growth inhibition on soft agar was not because of nonviability of the
cells, but because of the presence of hUG. Neither the nontransfected nor the hUG-
transfected prostate adenocarcinoma cells (HTB-81) showed any effect on colony
formation on soft agar plates (data not shown).
To delineate the mechanism(s) of suppression of cellular invasiveness, we first
determined whether these normal and cancer cells express any functional receptor
(binding protein) for hUG. Since hUG exists as a homodimer and earlier data have
FIGURE 1. (a) Effect of recombinant hUG on extracellular matrix (ECM) invasion by
NIH 3T3, mouse mastocytoma, sarcoma, lymphoma, and fibrosarcoma cells. The hUG-
induced suppressions of ECM invasion were significant when NIH 3T3, mouse mastocyto-
ma, sarcoma, and lymphoma cells were used, but there was no effect on fibrosarcoma cells.
Myoglobin was used as a control. Data represent the average of three experiments. (b) Effect
of hUG on ECM invasion by nontransfected and hUG-transfected HEC-1A cells. Note that
there was significant suppression of ECM invasion when using either wild-type HEC-1A
cells treated with 1 µM recombinant hUG or hUG-transfected HEC-1A cells. Data represent
the average of three experiments. Reprinted with permission from reference 26.
241KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
shown that radiolabeled dimeric hUG interacts with the surface of NIH 3T3 and
trophoblast cells with a molecular mass of 190 kDa, we sought to determine whether
monomeric hUG also interacts with the same 190-kDa protein in these normal and
cancer cells. Thus, the reduced form of 125I-hUG was incubated with NIH 3T3 cells
and Scatchard analysis of the binding data using this 125I-hUG as a ligand showed a
single class of specific binding in these cells with a dissociation constant (Kd) of
20 nM (FIG. 4a). The nonreduced and reduced forms of radiolabeled hUG were
individually incubated with several cancer cells, such as mastocytoma, sarcoma,
lymphoma, and fibrosarcoma, and the Kd values were in the range of 30–35 nM
(nonreduced) and 20–25 nM (reduced), respectively (data not shown), except that no
such binding was observed in the case of fibrosarcoma cells. In order to delineate the
molecular size of the hUG-binding protein, we performed affinity cross-linking
experiments by incubating the reduced form of 125I-hUG with NIH 3T3 and other
cancer cells (mastocytoma, sarcoma, lymphoma, and fibrosarcoma) in the absence
or presence of excess unlabeled hUG and then cross-linked with DSS. A new 49-kDa
radioactive protein band, in addition to the previously identified 190-kDa band, was
detected when NIH 3T3,22 mastocytoma, sarcoma, and lymphoma cells were used
(FIG. 4b; lanes 2, 4, 6, and 8). However, both the 49- and 190-kDa bands were virtu-
ally undetectable when 1 µM hUG was added to the cells prior to 125I-hUG-binding
and affinity cross-linking (FIG. 4b; lanes 3, 5, 7, and 9). As expected, no protein
bands were detected in the absence of DSS (FIG. 4b; lane 1). The 190-kDa protein
band was also detected when the nonreduced form of 125I-hUG was used as a ligand
and incubated with each of the mastocytoma, sarcoma, and lymphoma cells (data not
FIGURE 2. (a) Reverse transcription of total RNA isolated from wild-type and pRC/
RSV-hUG-transfected HEC-1A (uterus) and HTB-81 (prostate) cells. PCR products were
resolved on agarose gel, transferred to nylon membrane, and hybridized with hUG-specific
oligonucleotide probe (hUG-P). GAPDH was used as a positive control in order to check the
integrity and quality of the RNA. Left panels: wild-type HEC-1A and HTB-81 cells. Right
panels: pRC/RSV-hUG-transfected HEC-1A and HTB-81 cells. Lanes 1 and 2 represent two
independently derived pRC/RSV-hUG-transfected clones, each of HEC-1A (left) and HTB-
81 (right), respectively. (b) Detection of hUG in nontransfected and hUG-transfected HEC-
1A and HTB-81 cells by immunoprecipitation followed by Western blot analysis: hUG (d),
hUG dimer; UG (m), hUG monomer. Reprinted with permission from reference 26.
242 ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 3. Anchorage-independent growth on soft agar: (a) nontransfected (wild-type), (b) pRC/RSV (mock)–transfected, and (c) pRC/RSV-
hUG-transfected HEC-1A cells. Note that there was striking suppression of anchorage-independent growth of pRC/RSV-hUG-transfected HEC-1A
cells. Magnification: ×20. Reprinted with permission from reference 26. [Figure reduced to 75%.]
243KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
shown). However, the 49- and 190-kDa protein bands were not observed when either
the nonreduced or reduced form of 125I-hUG was incubated with fibrosarcoma cells
(data not shown). To investigate the mechanism(s) of hUG-mediated suppression of
ECM invasion and anchorage-independent growth in nontransfected and pRC/RSV-
hUG-transfected HEC-1A cells, we first detected the expression of hUG-binding
protein (receptor) in both cells by receptor binding and affinity cross-linking exper-
iments. The results of the binding data using reduced 125I-hUG as a ligand showed
specific, high-affinity binding with a Kd value of 25 nM in nontransfected HEC-1A
cells (FIG. 4c). Affinity cross-linking data indicated the presence of 190- and 49-kDa
bands in nontransfected HEC-1A cells using reduced, radiolabeled hUG as a ligand
and these bands were specifically displaced when excess unlabeled hUG was added
FIGURE 4. (a) Scatchard plot of specific binding of 125I-hUG (reduced) on NIH 3T3
cells. The data are from three experiments and each point represents the mean of triplicate
determinations. (b) Affinity cross-linking of hUG receptor(s) on NIH 3T3 (lanes 1–3), mas-
tocytoma (lanes 4 & 5), sarcoma (lanes 6 & 7), and lymphoma (lanes 8 & 9) cells. Lane 1,
−DSS; lane 2, +DSS; lane 3, +unlabeled hUG, +DSS; lane 4, +DSS; lane 5, +unlabeled hUG,
+DSS; lane 6, +DSS; lane 7, +unlabeled hUG, +DSS; lane 8, +DSS; lane 9, +unlabeled hUG,
+DSS. Note that both the 190- and 49-kDa protein bands are present and those bands are
specifically displaced in the presence of excess unlabeled hUG. (c) Scatchard analysis of
specific binding of reduced 125I-hUG in nontransfected HEC-1A cells. The data are from
three experiments and each point represents the mean of triplicate determinations. Affinity
cross-linking of 125I-hUG in (d) HEC-1A and (e) HTB-81 cells. Lane 1, −DSS; lane 2,
+DSS; lane 3, +unlabeled hUG, +DSS. Note the absence of hUG-binding protein(s) in HTB-
81 cells. Reprinted with permission from reference 26.
244ANNALS NEW YORK ACADEMY OF SCIENCES
prior to the binding assays (FIG. 4d). Similar results were obtained when pRC/RSV-
hUG-transfected HEC-1A cells were used for these studies (data not shown). In con-
trast, both the binding and affinity cross-linking data demonstrated that there was no
such detectable specific, cross-linked protein band in nontransfected (FIG. 4e) and
transfected (data not shown) prostate adenocarcinoma cells (HTB-81). These data
led us to conclude that induced hUG expression or pretreatment of the adenocarci-
noma cells with purified hUG inhibits ECM invasion of those cells that also express
the hUG-binding protein(s). These results raise the strong possibility that hUG reg-
ulates the cellular invasiveness of uterine adenocarcinoma cells by both autocrine
and paracrine pathways.
To determine the tissue-specific expression of hUG-binding proteins, the re-
duced, radiolabeled hUG was incubated with several bovine tissues (heart, spleen,
trachea, aorta, lung, and liver) for binding and then cross-linked with DSS. The data
indicated that the heart, spleen, and liver expressed both the 190- and 49-kDa bind-
ing proteins, whereas only the 190-kDa protein was present in the trachea and lung
(FIG. 5a). Neither the 190- nor 49-kDa protein was present in the aorta (FIG. 5a).
These results showed that the 190- and 49-kDa proteins are expressed in a tissue-
specific manner. Since hUG is an anti-inflammatory protein, we sought to determine
whether hUG-binding protein(s) is up- or downregulated by different inflammatory
cytokines and other agents such as lipopolysaccharide (LPS). The NIH 3T3 cells
were pretreated with cytokines and LPS alone and in combination followed by bind-
ing and affinity cross-linking with 125I-hUG. The data showed that there was a sub-
stantial increase in the expressions of both the 190- and 49-kDa protein bands when
the cells were pretreated with interleukin-6 (IL-6) and LPS as compared to nontreat-
FIGURE 5. (a) Tissue-specific expression of hUG-binding protein(s) using different
bovine tissues by affinity cross-linking technique. Note that both the 190- and 49-kDa bind-
ing proteins are detectable in bovine heart, spleen, and liver. The trachea and lung express
only the 190-kDa protein, while none of these proteins are present in the aorta. (b) Effect of
different cytokines and other agents on the expression of hUG-binding protein(s) by NIH
3T3 cells. The intensities of both the 190- and 49-kDa protein bands were considerably
higher when the cells were treated with LPS and IL-6 compared to the control. However,
there was no increase in intensities of these protein bands when other agents were used. Re-
printed with permission from reference 23.
245 KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
ed cells (FIG. 5b). We found that both IL-6 and LPS mediated an apparent increase
in the expression of hUG receptor(s) by these agents. These results suggest that the
homeostatic mechanism to control the inflammatory response may involve not only
hUG production, but also the dynamics of the interaction of hUG with its binding
proteins that is regulated by proinflammatory agents.
In an attempt to purify the hUG-binding protein(s), the bovine spleen tissues were
homogenized and the solubilized extract of the homogenate was passed through the
FIGURE 6. Purification of the hUG-binding protein(s) by hUG affinity chromatogra-
phy. The affinity-purified samples were resolved by SDS-PAGE under denaturing and reduc-
ing conditions and then stained with silver staining. Note that two protein bands with
apparent molecular masses of 180 and 40 kDa were visualized. In addition, a faint band with
a molecular mass of 32 kDa was also seen. However, the 32-kDa band did not interact with
the radiolabeled hUG, indicating that this may be either a degradation product of upper
bands that lacks the hUG-binding epitope or an artifact. Reprinted with permission from
246ANNALS NEW YORK ACADEMY OF SCIENCES
Sepharose 4B–linked hUG affinity column. The bound protein was eluted by lower-
ing the pH of the elution buffer. The affinity-purified fraction was examined by bind-
ing and affinity cross-linking assays (data not shown). The purity was checked by
SDS-PAGE followed by silver staining. Two major protein bands with apparent
molecular masses of 180 and 40 kDa were detected (FIG. 6). It is not clear if these
two protein bands are the subunits of the hUG-binding protein or if the lower band
is the degradation product of the upper one with the putative receptor–binding
epitope remaining the same. Molecular cloning and further characterization will an-
swer which of the two alternatives is correct. In addition, a faint protein band with
an apparent molecular mass of 30 kDa was also noticed that did not interact with
radiolabeled hUG. Data suggested that this 30-kDa band might be a degradation
product of a high-molecular-weight hUG-binding protein band that does not recog-
nize the receptor-binding epitope or it might be an artifact.
Earlier, UG was thought to be a steroid-inducible secretory protein.1,2 However,
now it is reported that a nonsteroid hormone like prolactin enhances the progesterone-
induced expression of the hUG gene.27,28 Both progesterone and prolactin have been
suggested to possess anti-inflammatory/immunomodulatory effects and it is possible
that hUG may act as the effector molecule for both of these hormones for exerting
these cellular effects. Although the biochemistry, molecular biology, and structural
biology of hUG have been studied extensively, the physiological functions of this
protein remained unsolved until recently. To understand the physiological function
of hUG, we performed targeted disruption of the hUG gene and generated hUG-
deficient mice. These hUG-deficient mice were found to develop multimeric
fibronectin-deposited renal glomerular disease, distal tubular hyperplasia, and renal
parenchymal fibrosis.29 Recent studies also demonstrated that both UG knockout as
well as UG antisense transgenic mice have developed immunoglobulin A (IgA)
nephropathy.30 The molecular mechanisms by which UG prevents abnormal fibro-
nectin (Fn) deposition in the glomeruli at least in part is due to the formation of Fn-
UG heteromers that compete with Fn homomerization, required for abnormal depo-
sition of Fn.31–33 As UG/hUG is an anti-inflammatory protein,34 the development of
glomerulonephritis, an inflammatory disease,35 and renal parenchymal fibrosis, a se-
quela of the inflammatory process, is understandable. However, the mechanism(s) of
distal tubular hyperplasia is not yet clear. In summary, we have demonstrated for the
first time the receptor-mediated anti-ECM-invasive function of hUG in specific nor-
mal and cancer cells in an autocrine and paracrine manner. In addition, hUG also
showed the inhibition of anchorage-independent growth of uterine-derived adeno-
carcinoma cells on soft agar and this is a characteristic property of most of the cancer
cells. Therefore, it is likely that UG/hUG is a novel cytokine3 that functions as a
tumor suppressor via its interaction with the cell surface binding proteins.
We thank I. Owens, J. Chou, J. D. Butler, and S. W. Levin for critical review of
the manuscript and helpful suggestions, and Mr. Tiawari and S. Philip for editorial
assistance. We also thank Rick Dreyfuss and Shauna Everett of Medical Arts and
Photography, NIH, for their expert photomicrographic assistance.
247KUNDU et al.: UTEROGLOBIN BINDING PROTEINS
1. KRISHNAN, R.S. & J.C. DANIEL, JR. 1967. “Blastokinin”: inducer and regulator of
blastocyst development in the rabbit uterus. Science 158: 490–492.
2. BEIER, H.M. 1968. Uteroglobin: a hormone-sensitive endometrial protein involved in
blastocyst development. Biochim. Biophys. Acta 160: 289–291.
3. MUKHERJEE, A.B., G.C. KUNDU, G. MANTILE-SELVAGGI et al. 1999. Uteroglobin: a
novel cytokine? Cell. Mol. Life Sci. 55: 771–787.
4. SINGH, G., J. SINGH, S.L. KATYAL et al. 1988. Identification, cellular localization, iso-
lation, and characterization of human Clara cell–specific 10 kD protein. J. Histo-
chem. Cytochem. 36: 73–80.
5. SINGH, G. & S.L. KATYAL. 1997. Clara cells and Clara cell 10 kD protein (CC10). Am.
J. Respir. Cell Mol. Biol. 17: 141–143.
6. BEATO, M. 1976. Binding of steroids to uteroglobin. J. Steroid Biochem. 7: 327–334.
7. JACKSON, P.J., R. TURNER, J.N. KEEN et al. 1988. Purification and partial amino acid
sequence of human urine protein 1: evidence for homology to rabbit uteroglobin. J.
Chromatogr. 452: 359–367.
8. BERNARD, A.M., R. LAUWERYS, A. NOEL et al. 1989. Urine protein 1: a sex-dependent
marker of tubular or glomerular dysfunction. Clin. Chem. 35: 2141–2142.
9. BERNARD, A., R. LAUWERYS, A. NOEL et al. 1991. Determination by latex immuno-
assay of protein 1 in normal and pathological urine. Clin. Chim. Acta 201: 231–245.
10. OKUTANI, R., Y. ITOH, H. HIRATA et al. 1992. Simple and high-yield purification of
urine protein 1 using immunoaffinity chromatography: evidence for the identity of
urine protein 1 and human Clara cell 10-kilodalton protein. J. Chromatogr. 577: 25–35.
11. LOPEZ DE HARO, M.S., M. PEREZ MARTINEZ, C. GARCIA & A. NIETO. 1994. Binding of
retinoids to uteroglobin. FEBS Lett. 349: 249–251.
12. GILLNER, M., J. LUND, C. CAMBILLAU et al. 1988. The binding of methylsulfonyl-
polychloro-biphenyls to uteroglobin. J. Steroid Biochem. 31: 27–33.
13. DHANIREDDY, R., T. KIKUKAWA & A.B. MUKHERJEE. 1988. Detection of a rabbit utero-
globin-like protein in human neonatal tracheobronchial washings. Biochem. Bio-
phys. Res. Commun. 152: 1447–1454.
14. KIKUKAWA, T., B.D. COWAN, R.I. TEJADA & A.B. MUKHERJEE. 1988. Partial character-
ization of a uteroglobin-like protein in the human uterus and its temporal relationship
to prostaglandin levels in this organ. J. Clin. Endocrinol. Metab. 67: 315–321.
15. MANTILE, G., L. MIELE, E. CORDELLA-MIELE et al. 1993. Human Clara cell 10-kDa
protein is the counterpart of rabbit uteroglobin. J. Biol. Chem. 268: 20343–20351.
16. PERI, A., E. CORDELLA-MIELE, L. MIELE & A.B. MUKHERJEE. 1993. Tissue-specific
expression of the gene coding for human Clara cell 10-kD protein, a phospholipase
A2–inhibitory protein. J. Clin. Invest. 92: 2099–2109.
17. LEVIN, S.W., J.D. BUTLER, U.K. SCHUMACHER et al. Uteroglobin inhibits phospholipase
A2 activity. Life Sci. 38: 1813–1819.
18. MIELE, L., E. CORDELLA-MIELE, A. FACCHIANO & A.B. MUKHERJEE. 1988. Novel anti-
inflammatory peptides from the region of highest similarity between uteroglobin and
lipocortin I. Nature (London) 335: 726–730.
19. MIELE, L., E. CORDELLA-MIELE & A.B. MUKHERJEE. 1990. High level bacterial expres-
sion of uteroglobin, a dimeric eukaryotic protein with interchain disulfide bridges in
its natural quaternary structure. J. Biol. Chem. 265: 6427–6437.
20. ROBINSON, D.H., K.L. KIRK & D.J. BENOS. 1989. Macromolecular transport in rabbit
blastocyst: evidence for a specific uteroglobin transport system. Mol. Cell. Endo-
crinol. 63: 227–234.
21. DIAZ GONZALEZ, K. & A. NIETO. 1995. Binding of uteroglobin to microsomes and plas-
matic membranes. FEBS Lett. 361: 255–258.
22. KUNDU, G.C., G. MANTILE, L. MIELE et al. 1996. Recombinant uteroglobin suppresses
cellular invasiveness via novel class of high-affinity cell surface binding site. Proc.
Natl. Acad. Sci. U.S.A. 93: 2915–2919.
23. KUNDU, G.C., A.K. MANDAL, Z. ZHANG et al. 1998. Uteroglobin (UG) suppresses
extracellular matrix–invasion by normal and cancer cells that express the high-
affinity UG-binding proteins. J. Biol. Chem. 273: 22819–22824.
248ANNALS NEW YORK ACADEMY OF SCIENCES Download full-text
24. KUNDU, G.C. & A.B. MUKHERJEE. 1997. Evidence that porcine pancreatic phospho-
lipase A2 via its high-affinity receptor can stimulate extracellular matrix invasion by
normal and cancer cells. J. Biol. Chem. 272: 2346–2353.
25. MUNSON, P.J. & D. RODBARD. 1980. Ligand: a versatile computerized approach for
characterization of ligand binding systems. Anal. Biochem. 107: 220–239.
26. ZHANG, Z., G.C. KUNDU, D. PANDA et al. 1999. Loss of transformed phenotype in can-
cer cells by overexpression of the uteroglobin gene. Proc. Natl. Acad. Sci. U.S.A. 96:
27. CHILTON, B.S., S.K. MANI & D.W. BULLOCK. 1988. Servomechanism of prolactin and
progesterone in regulating uterine gene expression. Mol. Endocrinol. 2: 1169–1175.
28. RANDALL, G.W., J.C. DANIEL, JR. & B.S. CHILTON. 1991. Prolactin enhances utero-
globin gene expression by uteri of immature rabbits. J. Reprod. Fertil. 91: 249–257.
29. ZHANG, Z., G.C. KUNDU, C-J. YUAN et al. 1997. Severe fibronectin-deposit renal glo-
merular disease in mice lacking uteroglobin. Science 276: 1408–1412.
30. ZHENG, F., G.C. KUNDU, Z. ZHANG et al. 1999. Essential role of uteroglobin in prevent-
ing IgA-nephropathy. Nat. Med. 5: 1018–1025.
31. RUOSLAHTI, E. 1988. Fibronectin and its receptors. Annu. Rev. Biochem. 57: 375–413.
32. ZHANG, Q. & D.F. MOSHER. 1996. Cross-linking of the NH2-terminal region of
fibronectin to molecules of large apparent molecular mass. J. Biol. Chem. 271:
33. HYNES, R.O. 1990. Fibronectins. Springer-Verlag. New York/Berlin.
34. MUKHERJEE, A.B., G.C. KUNDU, A.K. MANDAL et al. 1998. Uteroglobin (UG): its
physiological role in normal glomerular function uncovered by targeted disruption of
the UG gene in mice. Am. J. Kidney Dis. 32: 1106–1120.
35. BORDER, W.A. & E. RUOSLAHTI. 1992. Transforming growth factor-beta in disease: the
dark side of tissue repair. J. Clin. Invest. 90: 1–7.