| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Departments of Biochemistry and Molecular Biology, Saint Louis University School of Medicine, St. Louis, Missouri, USA;
Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan; and
Department of Biochemistry, Taipei Medical University, Taipei, Taiwan
1 Correspondence: Department of Biochemistry and Molecular Biology, Saint Louis University School of Medicine, 1402 South Grand Blvd., St. Louis, MO 63104, USA. E-mail: huangss{at}slu.edu or huangjs{at}slu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2M receptor as shown by MALDI-TOF analysis of tryptic peptides of TßR-V purified from bovine liver. In addition, 125I-IGFBP-3 affinity-labeled TßR-V in Mv1Lu cells is immunoprecipitated by antibodies to LRP-1 and TßR-V. RAP, an LRP-1 antagonist, inhibits binding of 125I-TGF-ß1 and 125I-IGFBP-3 to TßR-V and diminishes IGFBP-3-induced growth inhibition in Mv1Lu cells. Absent or low levels of LRP-1, as with TßR-V, have been linked to the malignant phenotype of carcinoma cells. Mutagenized Mv1Lu cells selected for reduced expression of LRP-1 have an attenuated growth inhibitory response to TGF-ß1 and IGFBP-3. LRP-1-deficient mouse embryonic fibroblasts lack a growth inhibitory response to TGF-ß1 and IGFBP-3. On the other hand, stable transfection of H1299 human lung carcinoma cells with LRP-1 cDNA restores the growth inhibitory response. These results suggest that the LRP-1/TßR-V/IGFBP-3 receptor is required for the growth inhibitory response to IGFBP-3 and TGF-ß1.—Huang, S. S., Ling, T.-Y., Tseng, W.-F., Huang, Y.-H., Tang, F.-M., Leal, S. M., Huang, J. S. Cellular growth inhibition by IGFBP-3 and TGF-ß1 requires LRP-1.
Key Words: type V TGF-ß receptor • IGFBP-3 receptor • growth inhibition • epithelial cells • carcinogenesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
70% sequence identity and exhibit distinct functions in vivo (1
, 2)
. All three TGF-ß isoforms are physiologically important. Null mutations in a gene encoding any one of the TGF-ß isoforms cannot be corrected by other family members (3)
. TGF-ß isoforms regulate multiple biological processes, including proliferation, extracellular matrix synthesis, angiogenesis, immune response, apoptosis, and differentiation (3)
. They have been implicated in the pathogenesis of tissue fibrosis, autoimmune diseases, cancer, and other disorders (3)
.
The various biological activities of TGF-ß isoforms (collectively referred to as TGF-ß) are mediated by specific cell surface receptors in responsive cells. Multiple cell surface receptors of various sizes have been identified in cultured cells and tissues by cross-linking of 125I-labeled TGF-ß (125I-TGF-ß) to these molecules in the presence of bifunctional cross-linking reagents. These include type I (TßR-I, M.W. 53,000), type II (TßR-II, M.W. 70,000), type III (TßR-III, M.W. 280,000–370,000), type IV (TßR-IV, M.W. 60,000), type V (TßR-V, M.W. 400,000), and type VI (TßR-VI, M.W. 180,000) receptors as well as several membrane-associated binding proteins (M.W. 38,000–190,000) (4
, 5)
. TßR-I and TßR-II are Ser/Thr-specific protein kinases and are believed to be primarily responsible for TGF-ß-induced cellular responses (6
, 7)
. TßR-III is a proteoglycan-containing membrane glycoprotein that presents the ligand to other TGF-ß receptor types and has recently been reported to regulate signaling mediated by the TßR-I/TßR-II heterocomplex (8
9
10)
. The identity of TßR-IV has not been confirmed by independent studies (11
, 12)
. TßR-V coexpresses with TßR-I, TßR-II, and TßR-III in most cell types (13)
and serves as the insulin-like growth factor binding protein-3 (IGFBP-3) receptor mediating IGF-independent (TGF-ß antagonist sensitive) growth inhibition upon IGFBP-3 stimulation (14
15
16)
. The TßR-VI and other membrane-associated TGF-ß binding proteins are expressed only in specific cell types (4
, 5)
.
One prominent activity of TGF-ß is transcriptional activation of genes coding for extracellular matrix proteins and their regulatory proteins (e.g., collagen, fibronectin, and plasminogen activator inhibitor-1). Another activity is cellular growth regulation; it inhibits the growth of most cell types including epithelial cells, endothelial cells, embryonic fibroblasts, and hematopoietic cells and stimulates growth of certain mesenchymal cells (e.g., fibroblasts) and some other specific cell types. These two activities are uncoupled in some cell types under certain experimental conditions (14
, 17
, 18)
. The segregation of the activities cannot be easily interpreted with a simple model of TßR-I/TßR-II complex formation followed by Smad2/Smad3/Smad4 signaling (6
, 7)
. Increasing evidence indicates that other signaling cascades in addition to the TßR-I/TßR-II signaling cascade are involved in the growth inhibitory response to TGF-ß (17
18
19
20
21
22)
. The TßR-V is expressed in most cell types used to investigate the TGF-ß-induced growth regulation and signaling via the TßR-I/TßR-II heterocomplex (6
, 7
, 23)
. It would logically be involved in the growth inhibitory response to TGF-ß. The finding that the TßR-V is identical to the IGFBP-3 receptor (14
15
16)
, which mediates IGF-independent growth inhibition induced by IGFBP-3, highlights the potential importance of TßR-V in TGF-ß-induced growth regulation. A pivotal role of TßR-V in this important activity is also supported by the observation that cells expressing little or no TßR-V do not exhibit the growth inhibitory response to TGF-ß1 and IGFBP-3 (14
, 24)
. To elucidate the role of TßR-V in TGF-ß- and IGFBP-3-induced growth suppression, we studied the structure and function of TßR-V purified and expressed in cultured cells. Unexpectedly, these studies demonstrated that the TßR-V/IGFBP-3 receptor is identical to the low density lipoprotein receptor-related protein-1 (LRP-1/2M receptor) (25)
, providing evidence for a new and previously unreported function of LRP-1. They also showed that stable transfection of human lung carcinoma cells with LRP-1 cDNA confers sensitivity to growth inhibition by either TGF-ß1 or IGFBP-3.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 14)
. Anti-TßR-V serum and anti-human LRP-1 light chain serum (carboxyl-terminal 15-residue peptide) were raised in rabbits according to published procedures (14
, 26)
. Anti-human LRP-1 IgG and anti-human LRP-1 serum and human receptor-associated protein (RAP) were provided by Drs. Guejun Bu, Joachim Herz, and Dudley Strickland. GST-RAP (a fusion protein of glutathione S-transferase and RAP) was expressed in Escherichia coli using pGEX-KG-RAP (6.4 kb) plasmid and purified according to the procedure of Herz et al. (27)
. pGEX-KG-RAP, pcDNA 3.1(–)neo, and pcDNA 3.1(–)neoLRP-1 plasmids were provided by Dr. Joachim Herz. Wheat germ lectin-Sepharose 4B was prepared as described (28)
. Protein A-Sepharose was obtained from Pharmacia LKB Biotech (Piscataway, NJ, USA). ß125(41-65), a specific TGF-ß peptide antagonist, was prepared as described previously (29)
. Disuccinimidyl subserrate (DSS) was obtained from Pierce (Rockford, IL, USA). Human TGF-ß1 was purchased from Austral Biologicals (Santa Clara, CA, USA) and R&D Systems (Minneapolis, MN, USA). Human IGFBP-3 (expressed in E. coli, M.W. 35,000) was obtained from Upstate (Charlottesville, VA, USA). Mink lung epithelial cells (Mv1Lu cells), mouse embryonic fibroblasts (MEF), homozygous LRP-1-deficient mouse embryonic fibroblasts (PEA-13 cells), human colorectal carcinoma cells (HCT116 and RII37 cells), human lung carcinoma cells (H1299 cells), human hepatocarcinoma cells (HepG2 and H3B cells), and human osteosarcoma cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS). PEA-13 cells were derived by Pseudomonas exotoxin A (PEA) selection of mouse embryonic fibroblasts that were heterozygous for the LRP-1 knockout mutation. Selection pressure by PEA-13 resulted in gene conversion of the functional LRP-1 allele, thus generating a homozygous LRP-1-deficient cell line (30)
. MEF and PEA-13 cells were obtained from American Type Culture Collection (Rockville, MD, USA). Human bone marrow-derived mesenchymal stem cells (HBMC) were kindly provided by Dr. Su-Li Cheng.
Purification of TßR-V from bovine liver plasma membranes
The TßR-V was purified by DEAE-cellulose column chromatography after Triton X-100 extraction of bovine liver plasma membranes and wheat germ lectin-Sepharose 4B affinity column chromatography as described (28)
. 125I-TGF-ß1 affinity labeling was used to locate TßR-V in the chromatographic fractions. The TßR-V was clearly identified as a 400 kDa Coomassie blue-stained protein band on SDS-PAGE in the N-acetylglucosamine eluents of wheat germ lectin-Sepharose 4B affinity column chromatography and in the NaCl eluents of DEAE-cellulose column chromatography.
MALDI-TOF analysis
The TßR-V purified from the DEAE-cellulose chromatography or wheat germ lectin-Sepharose 4B (28)
was subjected to 5% SDS-PAGE under reducing conditions, stained with Coomassie blue and digested with trypsin. MALDI-TOF analysis of the tryptic digests was carried out at Applied Biosystems (Foster City, CA, USA) and the Biotechnology Resource Laboratory, HHMI Biopolymer Laboratory/M. Keck Foundation, Yale Cancer Center Mass Spectrometry Resource (New Haven, CT, USA). The results provided by the two institutions appeared to be the same.
Western blot analysis
Equal amounts of protein from each cell type were subjected to 5% or 7.5% SDS-PAGE under nonreducing conditions (for using antisera or antibodies to LRP-1 heavy chain) or reducing conditions, followed by electrophoretic trans-blotting onto nitrocellulose membranes. The antigens on the nitrocellulose membranes were reacted with antisera or antibodies to LRP-1 heavy chain and light chain, followed by incubation with the second antibody-conjugated with horse radish peroxidase, and visualized using the ECL system (Santa Cruz).
Northern blot analysis
RNA analysis of plasminogen activator inhibitor-1 (PAI-1), glyceraldehyde-3-phosphate dehydrogenase (G3PDH), and LRP-1 was carried out as described previously (29)
. The relative levels were estimated based on the ratio of PAI-1 mRNA and G3PDH mRNA levels or of LRP-1 mRNA and rRNA levels. The relative intensities of the mRNAs on the autoradiograms were quantitated by a PhosphorImager.
[Methyl-3H] thymidine incorporation assay
Cells were plated on 24-well clustered dishes (0.5–1x105/well) and incubated with various concentrations of TGF-ß1 or IGFBP-3. After incubation at 37°C for 16 h, the cells were pulse labeled with 1 µCi of [methyl-3H]thymidine at 37°C for 4 h (14)
. The cells were then washed twice with 1 mL of 10% trichloroacetic acid and once with 0.5 mL of ethanol:ether (2:1, v/v) and dissolved in 0.2 N NaOH for scintillation counting. To examine the effect of GST-RAP on IGFBP-3-induced inhibition of DNA synthesis, cells were incubated with various concentrations of IGFBP-3 and GST-RAP (100 µg/mL). During incubation, GST-RAP (100 µg/mL) or the solvent vehicle was added to the medium hourly for 8 h. The assays were performed in quadruplicate.
Mutagenesis and Pseudomonas exotoxin selection of Mv1Lu cells
The mutagenesis and Pseudomonas exotoxin selection of Mv1Lu cells were performed according to Fitzgerald et al. (31)
. Briefly, Mv1Lu cells were grown in DMEM containing 10% FCS and treated with 5 mM ethyl methanesulfonate in the DMEM medium. After 21 h, the cells were split at a ratio of 1:50 in 10 cm Petri dishes and grown for 4 days. Cells were then treated with Pseudomonas toxin (100 ng/mL) for 1 wk. The clones were selected and grouped into two classes: one expressed very low levels of LRP-1 (a representative clone was PEA-C11) and the other expressed LRP-1 levels comparable to those of parent cells (a representative clone was PEA-B1 cells). These were found to have alterations (e.g., accelerated ligand degradation after LRP-1-mediated ligand binding and internalization) in post-LRP-1 events.
Stable transfection of H1299 cells with LRP-1 cDNA
Cells were plated at a cell density of 7 x 105/10 cm plate. Twelve hours later, the cells were transfected with pcDNA3.1(–)neoLRP-1, pcDNA3.1(–)neo vector using the calcium phosphate method. Briefly, 20 µg of pcDNA 3.1(–)neoLRP-1 or of pcDNA 3.1(–)neo vector was mixed with 417.5 µL H2O. CaCl2 (2 M in H2O, 62.5 µL) was slowly added to the DNA solution. This CaCl2 and DNA solution was then slowly added to 0.5 mL of 2x HEPES buffer (50 mM HEPES, pH 7.05, 280 mM NaCl, 10 mM KCl, 1.5 mM Na2HPO4, and 1.2 mM glucose). After 15–30 min at room temperature, the solution was added to the medium of each 10 cm Petri dish. After 12 h at 37°C, the cells were washed with phosphate-buffered saline and incubated with fresh medium. Twenty-four hours later, the cells were split at a ratio of 1:10 and plated on 6-well clustered plates containing 2 mL medium. After incubation at 37°C for 24 h, the cells were selected with 800 µg/mL of G418. After 14 days, cells expressing LRP-1 and vector only were selected and named H1299/LRP-1 and H1299/vector cells, respectively. The expression of the transfected LRP-1 cDNA was determined by Western blot analysis.
125I-Labeling of cell surface LRP-1
Cell surface LRP-1 was labeled with 125I using the lactoperoxidase method as described (26)
. The cell surface 125I-labeled LRP-1 in the cell lysates was immunoprecipitated with antisera to LRP-1 heavy chain or light chain (2–5 µg) and analyzed by 5 or 7.5% SDS-PAGE under reducing conditions. Both the LRP-1 heavy chain or intact LRP-1 and LRP-1 light chain were labeled with 125I.
125I-TGF-ß1- and 125I-IGFBP-3 affinity labeling
Purified TßR-V and cells were affinity labeled with 125I-TGF-ß1 or 125I-IGFBP-3 according to the published procedures (14
, 15
, 24)
. The 125I-IGFBP-3 affinity-labeled TßR-V was immunoprecipitated with anti-LRP-1 serum, anti-LRP-1 IgG, anti-TßR-V serum, or nonimmune serum (2–10 µg) as described (14
, 15
, 24)
.
Cell growth
Cells were plated on 24-well clustered dishes at a density of 1–2 x 104 cells/well in DMEM containing 1% FCS. The cell number was counted every day or after a 3 or 4 day incubation using a hematocytometer. The assays were performed in quadruplicate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, but we recently found that the kinase activity is not intrinsic to the TßR-V but is due to a casein kinase II-like kinase associated with it. To determine its structure, TßR-V was first purified from bovine liver plasma membranes as described (28)
. The purified TßR-V migrated as a single band on 5% SDS-PAGE. The Coomassie blue-stained band was excised and subjected to MALDI-TOF analysis after trypsin digestion. As shown in Table 1
, MALDI-TOF analysis revealed that the molecular masses of 23 tryptic peptides of bovine TßR-V were identical to those of the corresponding tryptic peptides of human LRP-1 (accession number NP 002323) originally cloned by Herz et al. (25)
. This suggests that bovine TßR-V is the homologue of human LRP-1.
|
Finding that bovine TßR-V and human LRP-1 are homologous prompted us to examine the effects of Ca2+ and LRP-1 ligands [activated 2macroglobulin (2M*), lactoferrin] (33
, 34)
and an antagonist (receptor-associated protein, RAP) (27)
on 125I-IGFBP-3 and 125I-TGF-ß1 binding to TßR-V. TßR-V was purified from bovine liver plasma membranes and ligand affinity labeled (28)
. Both the 125I-IGFBP-3-TßR-V complex and the 125I-TGF-ß1-TßR-V complex were cross-linked by the bifunctional reagent DSS after the binding of 125I-IGFBP-3 or 125I-TGF-ß1 to purified TßR-V had been carried out in the presence and absence of Ca2+ and of RAP. Ca2+ is known to be required for the ligand binding activity of LRP-1. RAP is an LRP-1 antagonist that blocks binding of all known ligands to LRP-1 (27
, 35
36
37)
. As shown in Fig. 1
A, binding of 125I-IGFBP-3 to purified TßR-V required the presence of Ca2+ whereas Ca2+ was not required for (but did enhance) binding of 125I-TGF-ß1 to TßR-V (lanes 4 vs. 2 and lanes 3 vs. 1, respectively). The binding of 125I-IGFBP-3 and 125I-TGF-ß1 to purified TßR-V was blocked by a TGF-ß1 peptide antagonist, ß125(41-65) (29)
(lanes 5 and 6). As shown in Fig. 1B
, 125I-IGFBP-3 bound to TßR-V in a Ca2+ concentration-dependent manner. Ca2+ also enhanced 125I-TGF-ß1 binding to TßR-V in a concentration-dependent manner (Fig. 1C
). Maximal binding of 125I-IGFBP-3 and 125I-TGF-ß1 occurred at 1–6 mM concentrations of Ca2+. The two bands of the 125I-IGFBP-3-TßR-V complex on the SDS-polyacrylamide gel represent 125I-IGFBP-3 dimer and monomer complexes (15)
(Fig. 1B
, lane 10). The LRP-1 antagonist RAP appeared to inhibit 125I-IGFBP-3 and 125I-TGF-ß1 binding to TßR-V (Fig. 1D
, lanes 11–13 and lanes 3–5, respectively). Lactoferrin did not have a significant effect on either 125I-IGFBP-3 or 125I-TGF-ß1 binding to TßR-V (Fig. 1D
, lanes 15 and 7, respectively). 2M* did not inhibit 125I-IGFBP-3 binding to TßR-V but blocked binding of 125I-TGF-ß1 to TßR-V (Fig. 1D
, lanes 14 and 6, respectively). This inhibition was due to the fact that 2M* itself forms complexes with 125I-TGF-ß1 (38)
. These results indicate that, like LRP-1, TßR-V requires the presence of Ca2+ for optimal ligand binding and that this ligand binding is sensitive to RAP inhibition.
|
RAP inhibits binding of IGFBP-3 and TGF-ß1 to cell surface receptors and blocks IGFBP-3-induced growth inhibition
Mv1Lu cells are a well-established model system for investigating TGF-ß activities and receptor functions. We therefore studied the effect of RAP on 125I-IGFBP-3 and 125I-TGF-ß1 binding to TßR-V in Mv1Lu cells using affinity labeling (binding and cross-linking). As shown in Fig. 2
, RAP inhibited 125I-IGFBP-3 binding to TßR-V in a concentration-dependent manner with an IC50 of 5 µg/mL (Fig. 2A
), whereas RAP weakly inhibited 125I-TGF-ß1 binding to TßR-V and TßR-III in Mv1Lu cells (Fig. 2B
). Since RAP strongly inhibited 125I-IGFBP-3 binding to TßR-V in Mv1Lu cells, we anticipated that RAP would block IGFBP-3-induced growth suppression in these cells. We therefore examined the effect of repeated doses of GST-RAP (a fusion protein of glutathione S-transferase and RAP) on DNA synthesis of Mv1Lu cells. We found that a single dose (100 µg/mL) of GST-RAP was unable to block DNA synthesis of Mv1Lu cells during an 18 h incubation. This was consistent with a report that RAP was no longer effective in blocking 2M* association and degradation in cells after a >1 h incubation time, presumably due to efficient cellular binding and degradation of RAP under culture conditions (39)
. For this reason, Mv1Lu cells were incubated with various concentrations of IGFBP-3 and GST-RAP (100 µg/mL) or the solvent vehicle, each (GST-RAP or the solvent vehicle) added to the culture medium hourly for 8 h. After further incubation for 10 h, DNA synthesis of the cells was determined. As shown in Fig. 2C
, repeated doses of GST-RAP effectively blocked growth suppression induced by 0.5 µg/mL IGFBP-3 in these epithelial cells. RAP was also used for the same experiment and yielded similar results (data not shown). Since RAP is a well-known LRP-1 antagonist, these results support the notion that TßR-V is functionally identical to LRP-1.
|
The 125I-IGFBP-3 affinity-labeled TßR-V is immunoprecipitated by antibodies to LRP-1
In a previous paper we showed that 125I-IGFBP-3 affinity-labeled TßR-V could be immunoprecipitated by antiserum to TßR-V (14
, 15)
. If TßR-V is identical to LRP-1, 125I-IGFBP-3 affinity-labeled TßR-V should be immunoprecipitated by antisera to either LRP-1 or TßR-V. To test this, TßR-V in Mv1Lu cells was 125I-IGFBP-3 affinity labeled, then the 125I-IGFBP-3 affinity-labeled TßR-V was immunoprecipitated by antibodies to LRP-1 or TßR-V. As shown in Fig. 3
,TßR-V was affinity labeled with 125I-IGFBP-3 in the presence of DSS (prior to immunoprecipitation) (lane 1). The 125I-IGFBP-3 affinity labeling of TßR-V was blocked in the presence of 100-fold excess of unlabeled IGFBP-3 (lane 2). The 125I-IGFBP-3 affinity-labeled TßR-V was immunoprecipitated by anti-LRP-1 IgG, anti-LRP-1 serum, and anti-TßR-V serum (lanes 3, 6, and 4, respectively) but not by nonimmune serum (lane 5). Anti-LRP-1 IgG and anti-LRP-1 sera were provided by Drs. Guojan Bu, Joachim Herz and Dudley Strickland. With the results described above, these results suggest that TßR-V is identical with LRP-1.
|
Cells lacking or expressing low levels of TßR-V also express no to low levels of LRP-1
TßR-V coexpresses with TßR-I, TßR-II, and TßR-III in all normal cell types examined (13
14
15
, 24)
. Many carcinoma cells (e.g., HCT116, H1299, HepG2, MCF-7, and H3B cells) do not express detectable TßR-V or express very low levels of TßR-V (13
, 14)
. If TßR-V is LRP-1, one should see correspondingly undetectable or very low levels of expression of LRP-1 in these carcinoma cells. To test this, we performed Western blot analysis using antiserum to the LRP-1 light chain. The light chain of LRP-1 contains the transmembrane domain, is stable and is therefore appropriate to use as an indicator for the measurement of LRP-1 expression. The recovery of the LRP-1 heavy chain varies depending on experimental conditions because it noncovalently associates with the LRP-1 light chain (25)
. As shown in Fig. 4
A, carcinoma cells (HepG2, H1299, and H3B cells) and osteosarcoma cells expressed low levels of LRP-1 (lanes 3–6) whereas normal or nontransformed cells such as mouse embryonic fibroblasts (MEF), Mv1Lu cells, and Mv1Lu mutant cells (PEA-B1 cells, which have acceleration of degradation of internalized LRP-1 ligands) exhibited high levels of LRP-1 (lanes 1, 2, and 7). On the other hand, human colorectal carcinoma cells (HCT116 and RII37 cells), originally identified as TßR-II-deficient cells but later also found to be deficient in TßR-V (14
, 18)
, did not express LRP-1 as demonstrated by Northern blot analysis (Fig. 4Ba
, lanes 2 and 3). PEA-13 cells (a homozygous LRP-1-deficient cell line) also showed no expression of TßR-V as determined by 125I-IGFBP-3 affinity labeling (Fig. 4C
, lane 2). These results support the notion that LRP-1 is identical to TßR-V.
|
Reduced expression of LRP-1 attenuates the growth inhibitory response to IGFBP-3 and TGF-ß1 in Mv1Lu cells
Mv1Lu cells are a standard model system to investigate TGF-ß activity and functions of TGF-ß receptors, including TßR-V. To define the role of LRP-1 in IGFBP-3-induced growth inhibition, we created Mv1Lu cell mutants using ethyl methane sulfonate mutagenesis (31)
. We then selected those expressing low levels of LRP-1 by Pseudomonas exotoxin selection using published procedures (31)
. Pseudomonas exotoxin selection yields mutant cells with reduced expression of LRP-1. A representative clone was PEA-C11 cells. Western blot analysis (Fig. 5
A) revealed that the PEA-C11 cells produced <5% (based on the sensitivity of this Western blot) of the amount of LRP-1 produced by the parent cells (lane 4 vs. lane 3) and that the control cells (for comparison) mouse embryonic fibroblasts (MEF cells) and LRP-1-deficient mouse embryonic fibroblasts (PEA-13 cells) expressed high levels of and no LRP-1, respectively (lanes 1 and 2). At the steady state, 90–95% of LRP-1 is localized intracellularly (36)
. To evaluate the cellular distribution of LRP-1, we examined the cell surface expression of LRP-1 in these mutant cells by 125I-cell surface labeling followed by immunoprecipitation. As shown in Fig. 5B
, PEA-C11 cells expressed the LRP-1 light chain (which contains the transmembrane domain of LRP-1) at levels comparable to that in Mv1Lu cells (Fig. 5Ba
, lane 1 vs. lane 3). However, the amount of the heavy chain of LRP-1 that noncovalently associates with the transmembrane light chain in PEA-C11 cells was greatly reduced compared with that found in Mv1Lu cells (Fig. 5Bb
, lane 2 vs. lane 4). This suggests that PEA-C11 cells have less functional LRP-1 than the parental cells. Approximately 10% of the heavy chain remained associated with the light chain of cell surface LRP-1 in these mutant cells under these experimental conditions as determined by 125I-cell surface labeling and immunoprecipitation.
|
The cell surface expression of TßR-V/LRP-1 in Mv1Lu and PEA-C11 cells was also examined by cell surface 125I-TGF-ß1 affinity labeling. As shown in Fig. 5C
, the amount of 125I-TGF-ß1 affinity-labeled TßR-V in PEA-C11 cells was less than that found in Mv1Lu cells (Fig. 5Cc
, lanes 6 vs. 7). PEA-C11 cells contained 15% as much TßR-V as the parent Mv1Lu cells. It is of interest to note that concomitant attenuation of TßR-III expression was also observed in these mutant cells (Fig. 5Cd
vs. Fig. 5Cb
). These results suggest that the PEA-C11 cells possess 15% as much cell surface TßR-V/LRP-1 as the parent cells.
We then examined the growth inhibitory response to TGF-ß1 and IGFBP-3 in Mv1Lu and PEA-C11 cells. PEA-C11 cells showed a diminished response to TGF-ß1- and IGFBP-3-induced growth inhibition as determined by measuring [methyl-3H] thymidine incorporation into cellular DNA (Fig. 6
A) and by counting cell number (Fig. 6B
). At 0.5 µg/mL, IGFBP-3 inhibited DNA synthesis and cell growth in PEA-C11 cells by 10% and 45%, respectively, compared with 30% and 70% in Mv1Lu cells (Fig. 6Aa
and Fig. 6Ba
). TGF-ß1 (20 pM) blocked DNA synthesis and cell growth in PEA-C11 cells by 70% as compared with 90–100% inhibition in Mv1Lu cells (Fig. 6Ab
and Fig. 6Bb
). By contrast, PEA-C11 cells exhibited a level of TGF-ß1-induced transcriptional activation of plasminogen activator inhibitor-1 (PAI-1) comparable to that observed in Mv1Lu cells (Fig. 6C
). PEA-C11 cells also exhibited a growth response to basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) equal to that observed in Mv1Lu cells (data not shown). These results indicate that decreased expression of LRP-1 induced by mutagenesis of Mv1Lu cells leads to attenuation of their growth inhibitory response (to IGFBP-3 and TGF-ß1) without significant effect on the TGF-ß1-induced transcriptional activation of PAI-1 or growth regulation by other growth factors such as bFGF and EGF.
|
LRP-1-deficient mouse embryonic fibroblasts lack the growth inhibitory response to TGF-ß1
All of the Mv1Lu cells mutant cells we obtained exhibited attenuated expression of LRP-1. We attempted to generate mutant epithelial cells lacking expression of LRP-1 but were unable to produce them. Such cells lacking LRP-1 would provide an ideal epithelial cell system for defining the role of LRP-1/TßR-V in IGFBP-3- and TGF-ß-induced growth suppression. We then determined the effects of IGFBP-3 and TGF-ß1 on DNA synthesis and cell growth in MEF cells and homozygous LRP-1-deficient mouse embryonic fibroblasts (PEA-13 cells) (30)
. As shown in Fig. 7
, TGF-ß1 (1.25 to 20 pM) inhibited DNA synthesis (Fig. 7A
) and cell growth (Fig. 7B
) in MEF cells but not in PEA-13 cells. TGF-ß1 (20 pM) exhibited 30–40% inhibition of DNA synthesis and cell growth in MEF cells. In contrast, TGF-ß1 (20 pM) stimulated DNA synthesis and cell growth by up to 20% in PEA-13 cells. IGFBP-3 did not appear to have a significant effect on DNA synthesis and cell growth in MEF cells or PEA-13 cells at concentrations up to 1 µg/mL (data not shown). On the other hand, TGF-ß1 was capable of transcriptional activation of PAI-1 equally well in both MEF and PEA-13 cells (Fig. 7C
). These results support the notion that LRP-1/TßR-V is important for TGF-ß1-induced growth suppression. They also suggest that in PEA-13 cells, the absence of LRP-1/TßR-V does not affect TGF-ß1-induced PAI-1 expression, which is known to be primarily mediated by the TßR-I/TßR-II complex signaling in the cell systems studied so far. Like MEF cells, PEA-13 cells express TßR-I, TßR-II, and TßR-III (data not shown).
|
LRP-1 expression by stable transfection with LRP-1 cDNA restores the growth inhibitory response to IGFBP-3 and TGF-ß1 in a human lung carcinoma cell line
To prove that growth suppression is mediated by LRP-1/TßR-V, we attempted to restore the growth inhibitory response to TGF-ß1 in PEA-13 cells. However, we were unable to generate the stable clones of PEA-13 cells expressing LRP-1 at levels comparable to those in MEF cells after transfection with LRP-1 cDNA. For this reason, we turned to H1299 cells (human lung carcinoma cells) for rescue experiments. H1299 cells were chosen for two reasons: 1) H1299 cells are derived from lung epithelial cells that in general express 
30% as much LRP-1 as fibroblasts, (e.g., MEF cells). H1299 cells express very low levels of endogenous LRP-1. 2) IGFBP-3 and TGF-ß1 do not inhibit DNA synthesis or cell growth in H1299 cells. H1299 cells were stably transfected with LRP-1 cDNA or vector only, cloned under G418 selection, and named H1299/LRP-1 and H1299/vector cells, respectively. The expression of LRP-1 was determined by Western blot analysis using antisera to LRP-1 light chain (Fig. 8
Aa) and heavy chain (Fig. 8Ab
). The effects of IGFBP-3 and TGF-ß1 on DNA synthesis and cell growth in these cells were then examined. As shown in Fig. 8B
, C both IGFBP-3 and TGF-ß1 inhibited DNA synthesis (Fig. 8B
) and cell growth (Fig. 8C
) of H1299/LRP-1 cells but are mitogens or growth stimulators for H1299/vector cells. TGF-ß1 induced transcriptional activation of PAI-1 in H1299/LRP-1 cells and H1299/vector cells (data not shown). These results indicate that stable transfection of H1299 cells with LRP-1 cDNA can restore the growth inhibitory response to IGFBP-3 and TGF-ß1 without significantly altering TGF-ß1-induced transcriptional activation of PAI-1 in these cells.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
34
35
36
37)
. The embryonic lethality of the LRP-1 null mutation strongly suggests that LRP-1 may have important biological functions other than its involvement in catabolism of ligand molecules (30
, 36
, 37)
. The LRP-1 ligand 2M* has been shown to induce signaling in several cell types (36
, 37)
. The finding that TßR-V is identical to LRP-1 presented here has disclosed a novel growth regulatory function of LRP-1 that may be important, even necessary, during embryonic development. It also raises the question of how LRP-1, a well-known endocytotic receptor, mediates IGFBP-3-induced growth inhibition and why it is required for TGF-ß1-induced growth inhibition. LRP-1 binds many structurally unrelated ligand molecules. A few have been reported to regulate cell growth, but the mechanisms underlying the regulation remain unknown (36
, 37)
. It is possible that IGFBP-3 and TGF-ß1 bind to TßR-V/LRP-1 at specific sites that are distinct from those for binding other ligands and cause LRP-1 activation (specific conformational change), resulting in cellular signaling that leads to growth inhibition. This is supported by the observation that the LRP-1 ligands lactoferrin and 2M* did not inhibit binding of 125I-IGFBP-3 and 125I-TGF-ß1 to the purified TßR-V. It is also possible that IGFBP-3 and TGF-ß1, both dimeric proteins (noncovalently and covalently bound, respectively), may be capable of activating LRP-1 by inducing dimerization or oligomerization.
Unlike the case in wild-type cells (Mv1Lu cells), high concentrations of TGF-ß1 (50–100 pM) inhibit DNA synthesis only weakly in Mv1Lu mutants DR26 and R1B cells, which lack functional TßR-II and TßR-I, respectively (14
, 24
, 40)
. In fact, TGF-ß1 at <10 pM is ineffective in inhibiting growth of these mutant cells. This suggests that TßR-I and TßR-II are obligatory for the growth inhibitory response to TGF-ß1, especially at low concentrations. The moderate effect of high concentrations of TGF-ß1 on DNA synthesis in DR26 and R1B cells is presumably mediated by TßR-V/LRP-1, known to be present in these cells (24)
. In contrast to TGF-ß1, IGFBP-3 is a potent growth inhibitor in DR26 and R1B cells (14)
. It inhibits DNA synthesis in DR26 cells more strongly than in wild-type Mv1Lu cells. The ability of IGFBP-3 to induce growth inhibition in DR26 and R1B cells suggests that the TßR-V/LRP-1/IGFBP-3 receptor can mediate growth inhibition in the absence of TßR-I or TßR-II.
Here we demonstrate that the Mv1Lu mutant cells (PEA-C11 cells) express only 10–15% as much cell surface LRP-1 as Mv1Lu cells and exhibit an attenuated growth inhibitory response to IGFBP-3 and TGF-ß1. This is consistent with the notion that LRP-1/TßR-V mediates the IGFBP-3- and TGF-ß1-induced growth inhibitory response in responsive cells. The requirement of LRP-1/TßR-V for IGFBP-3- and TGF-ß1-induced growth inhibition is further evidenced by the observation that homozygous LRP-deficient mouse embryonic fibroblasts (PEA-13 cells) fail to respond to growth inhibition induced by TGF-ß1 as wild-type MEF cells do. Furthermore, stable transfection with LRP-1 cDNA of H1299 cells, a human lung carcinoma cell line that expresses very low levels of LRP-1 and is insensitive to IGFBP-3 and TGF-ß1 growth inhibition, restores the sensitivity to both IGFBP-3 and TGF-ß1.
The molecular basis for the requirement of LRP-1 for growth inhibition induced by TGF-ß1 is unknown. We suspected that in addition to its potential signaling functions (41
, 42)
, the endocytic function of LRP-1 might be indirectly involved in signaling mediated by the TßR-I/TßR-II heterocomplex. It has recently been reported that the endosomal localization of the TßR-I/TßR-II heterocomplex-mediated signaling complex assembly is important for signaling that leads to cellular responses (43)
. TßR-V has been shown to physically associate with TßR-I (24)
. It might influence endocytosis of the TßR-I/TßR-II heterocomplex after stimulation by low concentrations of TGF-ß1. However, we did not find significant differences in the endocytosis and degradation rates of cell surface receptor-bound 125I-TGF-ß1 and TGF-ß1-stimulated phosphorylation of Smad2/3 between MEF and PEA-13 cells or between H1299/LRP-1 and H1299/vector cells (unpublished results). Nevertheless, the similarity in TGF-ß1-stimulated transcriptional activation of PAI-1 between PEA-13 and MEF cells and between H1299/LRP-1 and H1299/vector cells suggests that TßR-I/TßR-II-mediated signaling (which leads to expression of PAI-1 and other genes) is still functional in all these cells.
The TßR-I/TßR-II heterocomplex-mediated signaling generally believed to be primarily responsible for TGF-ß-induced cellular responses has been studied extensively (6
, 7
, 23)
. After ligand binding, TßR-II and TßR-I form heterocomplexes resulting in activation of the cytoplasmic kinase activity of TßR-I in the heterocomplex. The activated TßR-I then phosphorylates and activates Smad2 and Smad3. The activated Smad2/Smad3 forms oligomers with Smad4 and these translocate to the nucleus to regulate expression of target genes. The expression of the target genes directs the cellular responses to TGF-ß stimulation. The growth inhibitory response to TGF-ß1 has been studied in a variety of in vitro cultured cell systems. It is generally thought that the TGF-ß-activated Smad proteins target the promoters of the c-myc gene and cyclin-dependent kinases and repress its transcription in cooperation with nuclear corepressors. The various Smad protein and transcriptional coactivator complexes are also thought to activate the transcription of three major cell cycle inhibitors, the cyclin-dependent kinase inhibitors (6
, 7
, 23
, 44)
. These inhibit cyclin-dependent kinase activities associated with the G1 to S phase progression, prevent phosphorylation of RB by cyclin-dependent kinases, and arrest cells in G1. Currently, the exact molecular bases of Smad protein corepressor and coactivator complex formation are not well understood. The signaling involving LRP-1/TßR-V may function in concert with TßR-I/TßR-II-mediated signaling (Smad2/Smad3/Smad4) and possibly others (19
20
21)
, resulting in transcriptional repression of cell cycle progression-related genes and activation of cell cycle arrest-related genes and eventual growth inhibition.
TGF-ß is the most potent known growth inhibitor for epithelial cells. Loss of the growth inhibitory response to TGF-ß is believed to contribute to malignancy of many human carcinoma cells and other cancer cell types (45
, 46)
. Lack of TßR-I or TßR-II can explain in part why these carcinoma cells do not exhibit the growth inhibitory response to TGF-ß. However, stable transfection by TßR-I or TßR-II cDNA of some of these carcinoma cells failed to restore the growth inhibitory response (14
, 18)
, suggesting that other alterations, including concomitant loss or attenuation of expression of other receptor types (e.g., TßR-V) and postreceptor signaling defects, might have occurred in these carcinoma cells. We hypothesize that loss or diminished expression of TßR-V/LRP-1 is an important step in carcinogenesis. This hypothesis is supported by several observations. 1) Cancer cells have greatly decreased or undetectable expression of LRP-1 in comparison with their normal counterparts (47
48
49
50)
. 2) Mv1Lu mutants R1B and DR26 cells, which express TßR-V and lack functional TßR-I and TßR-II, respectively, respond to IGFBP-3-induced growth inhibition and exhibit normal cell properties (14
, 24
, 40)
. 3) No normal cells have been found to lack LRP-1/TßR-V as determined by 125I-TGF-ß1 (or 125I-IGFBP-3) affinity labeling and Western blot analysis (13
, 14
, 24)
. 4) Many carcinoma cells lack LRP-1/TßR-V or express low levels of LRP-1/TßR-V but express TßR-I and TßR-II (13
, 14
, 49)
, and 5) Carcinoma cells (e.g., H1299 cells), which express low levels of LRP-1/TßR-V, exhibit mitogenic or growth stimulatory response to TGF-ß1 and IGFBP-3 and increased malignancy.
Thus, accumulating evidence indicates that LRP-1/TßR-V expression inversely correlates with malignancy and invasiveness of carcinoma cells and other cancer cell types, supporting the importance of LRP-1/TßR-V in the tumor biology of carcinoma cells and possibly other cancer cells (13
, 47
48
49
50)
. On the other hand, LRP-1 overexpression has been found in glioma and other cancer cells (51)
. These mesenchymal cell-derived cancer cells and their normal counterparts generally have a growth stimulatory response to IGFBP-3 and TGF-ß. The association of increased expression of LRP-1 with malignancy of these cancer cells is consistent with the notion that LRP-1 can play a stimulatory or inhibitory role in determining the malignant behavior of different cancer cells (46)
. Together with angiogenesis factors (FGF-3 and VEGF), LRP-1 and IGFBP-3 have recently been identified as a group of hypoxia-induced genes of tumor cells (52)
. The autocrine cell growth suppression mediated by LRP-1 and IGFBP-3 and the angiogenesis stimulated by FGF-3 and VEGF may enable tumor cells to survive under hypoxic conditions. Investigations of the complex mechanisms by which LRP-1/TßR-V regulates cell growth promise to increase our understanding of tumor biology of carcinoma cells and possibly other cancer cell types.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 22, 2003. Accepted for publication June 26, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
