TFPI-2 (tissue factor pathway inhibitor-2) has recently been recognized as a new tumour suppressor gene. Low expression of this protein in several types of cancers allows for enhanced tumour growth, invasion and metastasis. To investigate the molecular mechanism responsible for the tumour-suppressor effects of TFPI-2, we performed yeast two-hybrid analysis and identified PSAP (prosaposin) as a TFPI-2-interacting partner. This interaction was confirmed by co-immunoprecipitation and immunofluorescence. The region of TFPI-2 that interacts with PSAP is located in the KD2 (Kunitz-type domain 2). Further study showed that PSAP does not affect the function of TFPI-2 as a serine proteinase inhibitor, but that TFPI-2 could inhibit the invasion-promoting effects of PSAP in human HT1080 fibrosarcoma cells. The results of the present study revealed that TFPI-2 interacts with PSAP, which may play an important role in the physiology and pathology of diseases such as cancer.
TFPI (tissue factor pathway inhibitor)-2 is a 32-kDa Kunitz-type serine proteinase inhibitor consisting of three Kunitz-type proteinase inhibitor domains homologous with TFPI. TFPI-2 is secreted into the ECM (extracellular matrix) by a wide variety of human cells, including endothelial cells , monocytes , fibroblasts , epithelial cells , smooth muscle cells , syncytiotrophoblast cells [6,7], and several human tumour cells [8–11]. It is now commonly recognized that TFPI-2 can protect the matrix from degradation, thereby counteracting tumour invasion and metastasis [8,12–15]. Our previous studies found that TFPI-2 is associated with fibrosarcoma, breast cancer, ALI (acute lung injury) and pre-eclampsia [16–19]. The down-regulation of TFPI-2 has been found to be related to numerous pathophysiological processes, such as inflammation, angiogenesis, atherosclerosis and tumour growth/metastasis [20–23]. TFPI-2 also strongly inhibits trypsin, plasmin, chymotrypsin, cathepsin G, plasma kallikrein, factor VIIa–tissue factor complex, and the trypsin- and plasmin-mediated activation of the proMMPs (matrix metalloproteinases) proMMP-1 and proMMP-3, and MMP-2 and MMP-9 [15,24–28].
In 2006, Torres-Collado et al.  identified TFPI-2 as a binding partner and a novel substrate of ADAMTS1 (a disintegrin and metalloproteinase with thrombospondin motifs 1). Cleavage that removed the protease-sensitive C-terminal region in TFPI-2 could alter its binding properties . Although some studies have addressed the structural basis for molecular recognition and protein interaction with other proteins by the KD (Kunitz-type domain) 1 of TFPI-2, the interaction of proteins with KD2, KD3 or the C-terminus of TFPI-2 has not been reported to date.
In the present study, we performed a yeast two-hybrid screen to identify proteins that interact with TFPI-2 using human fetal brain cDNA and identified a series of potential partners of TFPI-2, including PSAP (prosaposin). Finally, we confirmed that KD2 of TFPI-2 is responsible for the binding to PSAP, and the interaction between these two proteins has been validated by GST (glutathione transferase) pull-down, co-immunoprecipitation and co-localization, and this interaction inhibits the invasion-promoting effects of PSAP in human HT1080 fibrosarcoma cells. We also demonstrated that PSAP does not affect the function of TFPI-2 as a serine proteinase inhibitor.
The HEK (human embryonic kidney)-293T, COS-7, HeLa, HT1080, HT1080-TFPI-2, HUVEC (human umbilical vein endothelial cell) and HASMC (human aortic smooth muscle cell) cell lines were cultured in DMEM (Dulbecco's modified Eagle's medium; Invitrogen) supplemented with 10% heat-inactivated FBS (fetal bovine serum; Bioind) in a humidified atmosphere containing 5% CO2 at 37°C and were passaged every 2–4 days by trypsinization. Saccharomyces cerevisiae strains AH109 (MATa trp1-901 leu2-3, 112 ura3-52 his3-200 gal4Δgal80 LYS2::GAL1UAS-GAL1TATA-HIS3 GAL2UAS-GAL2TATA-ADE2 URA3:: MEL1UAS-MEL1TATA-lacZ) and Y187 (MATa ura3-52 his3-200 ade2-101 trp1-901 leu2-3,112 gal4Δmet_gal80ΔURA3::GAL1UAS-GAL1TATA-lacZ) were obtained from BD Biosciences Clontech. Medium was prepared as described in the BD Biosciences Clontech Yeast Protocols Handbook. Matrigel™ was purchased from BD Biosciences.
pGBKT7 (TRP1, 2 μm) Gal4p DNA-binding domain vector and pGADT7 (LEU2, 2 μm) Gal4p activation domain vector were obtained from Clontech. pGBKT7-TFPI-2 yeast two-hybrid bait plasmids were prepared by standard molecular biology procedures using the following cDNAs: TFPI-2 (amino acids 23–235), TFPI-2ΔKD1 (amino acids 92–235 of TFPI-2), TFPI-2/KD1 (amino acids 23–91 of TFPI-2), TFPI-2/KD3/C (amino acids 154–235 of TFPI-2), TFPI-2ΔKD2 (amino acids 23–91 and 154–235 of TFPI-2), TFPI-2/KD2 (amino acids 92–153 of TFPI-2) and TFPI-2/C (amino acids 209–235 of TFPI-2). Yeast two-hybrid bait plasmids were generated by introducing the corresponding cDNAs into pGBKT7. The C-terminal coding sequence of human PSAP was subcloned into pGADT7. The human PSAP open reading frame was obtained by PCR amplification using HEK-293T cDNA and was subcloned into the pcDNA3.1(+) (Invitrogen) vector to direct the expression of PSAP tagged with the Myc epitope and the His epitope at its C-terminus. TFPI-2 was subcloned into pGEX-4t-2 (Amersham Biosciences) for expression of a GST–TFPI-2 fusion protein in Escherichia coli and was subcloned into pcDNA3.0(−) to direct the expression of TFPI-2 tagged with the HA (haemagglutinin) epitope at its C-terminus. All constructs were verified by sequencing.
Yeast two-hybrid assay
The MATCHMAKER GAL4 two-hybrid system 3 (Clontech) was used according to the manufacturer's instructions. The bait plasmid, pGBKT7-TFPI-2, was stably expressed in yeast strain AH109 and did not have any self-transcriptional activity. The prey plasmid, human fetal brain MATCHMAKER cDNA library/pGADT-7 (Clontech), was transformed into yeast strain Y187. The corresponding AH109 yeast strains (MATα) were mated with the Y187 yeast strain (MATa). Positive clones were selected on medium lacking adenine, histidine, leucine and tryptophan to identify diploids containing a pGADT7-cDNA library plasmid and the pGBKT7-TFPI-2 plasmid and expressing the ADE2 and HIS3 two-hybrid reporter genes. Positive clones were further screened for expression of an additional reporter gene, MEL1-encoding α-galactosidase, using X-α-Gal (5-bromo-4-chloroindol-3-yl α-D-galactopyranoside; Clontech). Double-positive clones were isolated and characterized by DNA sequencing.
GST-fusion protein expression and pull-down assays
GST–TFPI-2 fusion proteins were expressed in E. coli BL21 cells by incubation in LB (Luria–Bertani) medium supplemented with 0.1 mM IPTG (isopropyl β-D-thiogalactopyranoside) and ampicillin (100 μg/ml) for 12 h at 20°C. The cells were lysed by sonication on ice in lysis buffer [1×PBS (pH 7.4), 1 mM PMSF and protease inhibitor mixture], followed by the addition of Triton X-100 to a final concentration of 1% and subsequent incubation for 30 min on ice. The lysate was partially purified by centrifugation at 15000 g for 20 min at 4°C. The resulting supernatant was used as a GST–TFPI-2 probe in a pull-down assay. GST alone was also produced to use as a negative control in the assay. The GST–TFPI-2 probe was incubated with glutathione–Sepharose beads (Amersham Biosciences) for 2.5 h at 4°C in lysis buffer. GST–TFPI-2 glutathione—Sepharose beads were washed three times with lysis buffer to remove unbound GST–TFPI-2 and then aliquoted for incubation with cell lysates. For GST pull-down assays, HEK-293T cells were transiently transfected with pcDNA3.1/PSAP/Myc using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. At 48 h later, the cells were washed three times with ice-cold PBS and resuspended in 500 μl of lysis buffer [25 mM Tris/HCl (pH 7.4), 150 mM NaCl, 5 mM MgCl2, 1% Igepal, 1 mM dithiothreitol, 5% glycerol and protease inhibitors]. Lysates were briefly vortex-mixed, incubated on ice for 30 min, and then centrifuged at 16000 g for 20 min at 4°C. The resulting supernatants were incubated with the beads coupled to the GST–TFPI-2 probe at 4°C for 2 h with rotation. After incubation, the beads were washed three times with lysis buffer, resuspended in 2×Laemmli buffer containing 5% 2-mercaptoethanol, and boiled for 10 min. Samples were analysed by SDS/PAGE and immunoblotted using an anti-Myc mouse monoclonal antibody (Invitrogen) to detect Myc-tagged PSAP proteins.
Cell lysates or conditioned media were preabsorbed to Protein G–agarose (Roche) for 1 h at 4°C on a rotation wheel. The material was immunoprecipitated with 2 μg of anti-HA antibody (Clontech), anti-TFPI-2 antibody (made in the laboratory), anti-PSAP antibody (Abcam) or normal rabbit IgG (Santa Cruz Biotechnology) and 30 μl of Protein G–agarose (Roche) for 6 h or overnight at 4°C and washed five times with ice-cold lysis buffer; final pellets were resuspended in 2×SDS loading buffer and analysed by SDS/PAGE. The antibodies used were monoclonal anti-Myc, polyclonal anti-HA, polyclonal anti-PSAP and polyclonal anti-TFPI-2.
Confocal microscopy studies
For the immunofluorescence studies, COS-7 and HeLa cells were seeded on to 13-mm circular glass coverslips and transfected 16 h later with an HA-tagged TFPI-2 expression construct and a Myc-tagged PSAP expression construct using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. At 36 h later, the cells were washed with PBS, warmed to 37°C, and fixed with 4% paraformaldehyde for 20 min at room temperature (20–25°C). Cells were washed three times with PBS and permeabilized with 0.2% Triton X-100 in PBS for 5 min. Non-specific staining was reduced by blocking with 5% goat serum overnight at 4°C. Cells were incubated with an anti-HA antibody (Clontech) or an anti-Myc antibody (Invitrogen) in PBS containing 5% goat serum overnight at 4°C, followed by incubation with Alexa Fluor® 488 anti-mouse secondary antibody or Alexa Fluor® 594 anti-rabbit secondary antibody (Invitrogen) at a dilution of 1:1000 for 1 h at room temperature. The coverslips were then washed several times with PBS and mounted with Moviol (Sigma–Aldrich). The same protocol was used to process untransfected HUVECs and HASMCs using polyclonal anti-PSAP (Abcam) and monoclonal anti-TFPI-2 (R&D Systems) as the primary antibodies, and anti-rabbit IgG-FITC and anti-mouse IgG-R (Santa Cruz Bitechnology) as the secondary antibodies.
Expression and purification of recombinant human PSAP in HEK-293T cells
For the purification of PSAP, 20 (10-cm diameter) culture dishes of subconfluent HEK-293T cells were transfected with a His-tagged PSAP expression construct using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. At 5 h later, the cells were washed with DMEM and incubated with DMEM without serum for 60 h. Then, samples of serum-free conditioned medium were collected, and insoluble material was removed by centrifugation at 3000 g for 15 min. The supernatant was concentrated by centrifugal ultrafiltration (Millipore), and the buffer of the concentrate was changed by adding a 10-fold volume of buffer A [50 mM NaH2PO4, 300 mM NaCl and 10 mM imidazole (pH 8)]. After reconcentration by centrifugal ultrafiltration (Millipore), the samples were applied to a nickel spin column [Ni-NTA (Ni2+-nitrilotriacetate); Qiagen] that had been pre-equilibrated with buffer A. The column was washed with buffer B (buffer A, containing 20 mM imidazole) until the A280 returned to the baseline value. PSAP–His was then eluted with 5 ml of buffer C (buffer A, containing 250 mM imidazole) and dialysed at 4°C against PBS. The size and purity of purified PSAP–His were assessed by SDS/PAGE and Western blotting with a monoclonal anti-Myc antibody.
siRNA (small interfering RNA) knockdown
For down-regulation of PSAP in HT1080 and HT1080-TFPI-2 cells, subconfluent cells in the culture dish were transfected with PSAP siRNA or control siRNA using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. At 5 h later, the transfection medium was removed and cells were incubated with complete medium for 48 h before performing functional assays or harvesting cell lysate for protein expression analyses. The expression of PSAP was assessed by semi-quantitative RT (reverse transcription)–PCR and Western blotting. The fragments of siRNA were synthesized by Integrated RNA Technologies (GenePharma), and the sequences were as follows: PSAP, 5′-GCUGGUGGGUUAUUUGGAUTT-3′ (sense) and 5′-AUCCAAAUAACCCACCAGCTT-3′ (antisense); control, 5′-UUCUCCGAACGUGUCACGUTT-3′ (sense) and 5′-ACGUGACACGUUCGGAGAATT-3′ (sense).
Cell invasion and migration assays
For transwell migration assays, (2.5–5)×104 HT1080 or HT1080-TFPI-2 cells were plated in the top chamber with a non-coated membrane (24-well insert; pore size, 8 μm; Corning Life Science). For invasion assays, 1.25×105 cells were plated in the top chamber with a Matrigel™-coated membrane (24-well insert; pore size, 8 μm; Corning Life Science). In both assays, cells were plated in serum-free medium without or with rhPSAP (recombinant human PSAP) at concentrations of 7.5, 15, 30 or 60 nM, and medium supplemented with serum was used as a chemoattractant in the lower chamber. The cells were incubated for 24 h, and cells that did not migrate or invade through the pores were removed by a cotton swab. Cells on the lower surface of the membrane were stained with Crystal Violet staining solution and counted. HT1080 or HT1080-TFPI-2 cells transfected with control siRNA or PSAP siRNA were also used for transwell migration and invasion assays as described above.
Trypsin inhibition assay
Trypsin inhibition assays were performed as described previously . In brief, 50 nM trypsin was incubated with or without various concentrations of rhTFPI-2 (recombinant human TFPI-2; American Diagnostica) and rhPSAP for 15 min at room temperature. S-2251 (American Diagnostica) was then added. The reaction was carried out at 37°C in 15 mM Tris/HCl (pH 7.4) containing 100 mM NaCl, 5 mM CaCl2 and 0.01% Tween 80 in a total volume of 200 μl. The trypsin activity was measured as the change in absorbance at 405 nm, and the inhibition constant (Ki) was calculated according to the method of Bieth et al. [29a].
MMP-2 and MMP-9 enzymatic activities were assayed by gelatin zymography . In brief, samples of serum-free conditioned medium with 5× loading buffer [0.25 M Tris/HCl (pH 6.8), 50% glycerol, 5% SDS and Bromophenol Blue) were run on 10% acrylamide gels impregnated with 0.1% gelatin (Sigma) at 120 V for 2.5 h. After incubation in wash buffer [2.5% Triton X-100, 50 mM Tris/HCl and 5 mM CaCl2 (pH 7.6)] for 1 h at room temperature with agitation and two brief washes in collagenase buffer [50 mM Tris/HCl, 200 mM NaCl, 10 mM CaCl2 and 1 μM ZnCl2 (pH 7.6)], the gels were incubated for 16 h in collagenase buffer at 37°C. Gelatinolytic activity was detected by a single-step stain/destain method using 0.02% Coomassie Blue in 1:3:6 acetic acid/methanol/water. No bands other than MMP-2 and MMP-9 were identified on gelatin zymography.
Enzymatic activity assay
The activities of MMP-2 and MMP-9 in serum-free conditioned medium were assayed using a Cell Active MMP-2 Fluorescent Assay kit and Cell Active MMP-9 Fluorescent Assay kit (Genmed Scientifics) respectively, according to the manufacturer's instructions. The relative fluorescence units were determined with an excitation wavelength of 330 nm and an emission wavelength of 400 nm. The consistency of the fluorescent polypeptide segments was calculated on the basis of the relative fluorescence units; MMP-2 and MMP-9 activities were expressed as nmol/mg per min.
RNA extraction, cDNA synthesis, semi-quantitative RT–PCR, real-time quantitative PCR
RNA from HT1080 and HT1080-TFPI-2 cells was isolated using TRIzol® reagent (Invitrogen). For cDNA synthesis, the template (5 μg of RNA per sample) was reverse-transcribed using SuperScript II RNase H− reverse transcriptase and oligo(dT)25 as a primer (Invitrogen). Semi-quantitative RT–PCR was performed in a total volume of 20 μl, including 1 μl of cDNA, 0.2 μM dNTPs, 0.4 μM primers and 0.4 μM Taq DNA polymerase (Promega). Primers were synthesized by Integrated DNA Technologies (Invitrogen), and the sequences were as follows: PSAP (ID 5660), 5′-ATGTACGCCCTCTTCCTCCT-3′ (sense) and 5′-CTAGTTCCACACATGGCGT-3′ (antisense); TFPI-2 (ID 7980), 5′-ATGGACCCCGCTCGCCCCCTG-3′ (sense) and 5′-TTAAAATTGCTTCTTCCGAAT-3′ (antisense); and GAPDH (glyceraldehyde-3-phosphate dehydrogenase; (ID 2597), 5′-ACCACAGTCCATGCCATCAC-3′ (sense) and 5′-TCACCACCCTGTTGCTGTA-3′ (antisense). After cDNA synthesis, PCR was completed under the following conditions: a denaturation cycle at 94°C for 5 min, 94°C for 30 s, annealing at 56°C for 30 s and elongation at 72°C for 100 s, and a final extension at 72°C for 10 min. The sizes of the amplified cDNA fragments and the number of PCR cycles were as follows: 1575 bp/26 cycles for PSAP, 708 bp/26 cycles for TFPI-2 and 450 bp/22 cycles for GAPDH. Real-time PCR amplification was carried out using an ABI Prism 7300 sequence detector (Applied Biosystems) and SYBR Green reagent. The specific primers were as follows: MMP-2 (ID 4313), 5′-AGAAGGATGGCAAGTACGGCTTCT-3′ (sense) and 5′-AGTGGTGCAGCTGTCATAGGATGT-3′ (antisense); MMP-9 (ID 4318), 5′-ATTTCTGCCAGGACCGCTTCTACT-3′ (sense) and 5′-TGTCATAGGTCACGTAGCCCACTT-3′ (antisense); and GAPDH, 5′-CATGTTCGTCATGGGTGTGAACCA-3′ (sense) and 5′-AGTGATGGCATGGACTGTGGTCAT-3′ (antisense). All amplification reactions were carried out under the following conditions: an initial stage of 95°C for 30 s, then a two-step program of 95°C for 5 s and 60°C for 31 s over 40 cycles; experiments were performed in triplicate. The relative target mRNA levels were analysed with ABI Prism 7300 software and normalized to that of the internal control GAPDH.
PSAP is a novel TFPI-2-interaction protein isolated in a yeast two-hybrid screen
To identify binding partners of TFPI-2, we performed a yeast two-hybrid screen with a human fetal brain cDNA library using the full-length human TFPI-2 protein without the signal peptide as bait. One of the clones isolated from the screen contained the C-terminal coding sequence of PSAP (GenBank® accession number NM_002778). The C-terminal cDNA of PSAP was isolated from the clone, and the two-hybrid interaction with TFPI-2 was confirmed by co-transforming the yeast reporter strain AH109 and assaying for growth on SD/−Ade/−His/−Leu/−Trp medium containing X-α-Gal (Figure 1A). To map the region of TFPI-2 responsible for PSAP binding, four deletion mutants of TFPI-2 were generated as follows: TFPI-2ΔKD1 (amino acids 92–235 of TFPI-2), TFPI-2/KD3/C (amino acids 154–235 of TFPI-2), TFPI-2ΔKD2 (amino acids 23–91 and 154–235 of TFPI-2) and TFPI-2/C (amino acids 209–235 of TFPI-2) (Figure 1B). We found that only TFPI-2ΔKD1 was able to interact with the C-terminus of PSAP in the yeast two-hybrid assay (Figure 1B). Several papers have reported that TFPI-2/KD1 appears to contain all of the structural elements necessary for serine proteinase inhibition . Two truncated forms of TFPI-2 comprising amino acids 23–91 (TFPI-2/KD1) and amino acids 92–153 (TFPI-2/KD2) were generated to determine the domain responsible for the PSAP interaction. As shown in Figure 1(B), only TFPI-2/KD2 interacted with the C-terminus of PSAP in the yeast two-hybrid assay, suggesting that the KD2 of TFPI-2 is critical for the interaction with PSAP.
Identification of PSAP as a novel TFPI-2-interacting protein
GST pull-down and co-immunoprecipitation assays confirmed the interaction between TFPI-2 and PSAP
First, we preformed a pull-down assay using a GST-fusion protein of TFPI-2 (GST–TFPI-2) to confirm the specific interaction of TFPI-2 with PSAP in vitro. GST–TFPI-2 was immobilized on glutathione–Sepharose beads and used as a probe to precipitate Myc-tagged human PSAP (Myc–PSAP) transiently expressed in HEK-293T cells. As shown in Figure 1(C), Myc–PSAP was specifically precipitated by GST–TFPI-2, but not by the GST protein alone. Secondly, we performed co-immunoprecipitation assays to confirm that TFPI-2 can interact with PSAP in vivo under normal biological conditions. We transiently co-expressed HA-tagged TFPI-2 and Myc-tagged PSAP in HEK-293T cells and immunoprecipitated the proteins using an anti-HA antibody. We also immunoprecipitated the proteins using an anti-TFPI-2 or anti-PASP antibody in HUVECs. Rabbit IgG (Santa Cruz Biotechnology) was used as a normal control. As shown in Figure 1(D), Myc–PSAP was detected in the samples immunoprecipitated by the anti-HA antibody, but not in those immunoprecipitated by rabbit IgG; and in Figure 1(E), PSAP or TFPI-2 was detected in the samples immunoprecipitated with anti-TFPI-2 or anti-PSAP antiobody, but not in those immunoprecipitated by rabbit IgG. Taken together, the results of these experiments confirm the specific interaction of TFPI-2 with PSAP in vitro and in vivo under normal biological conditions.
Co-localization of TFPI-2 and PSAP in cells
Given that TFPI-2 can interact with PSAP in vitro and in the cellular context, we next determined whether these two proteins co-localize in cells by confocal immunofluorescence microscopy. HeLa and COS-7 cells were co-transfected with HA–TFPI-2 and Myc–PSAP. At 36 h post-transfection, the cells were fixed, permeabilized and immunostained with primary antibody (monoclonal anti-Myc and polyclonal anti-HA antibodies) and secondary antibody (Alexa Fluor® 488-conjugated anti-mouse secondary antibody and Alexa Fluor® 594-conjugated anti-rabbit secondary antibody) to determine the subcellular localization of the TFPI-2 and PSAP proteins. As shown in Figures 2(A), 2(B), 2(E) and 2(F), both TFPI-2 and PSAP were localized in the cytoplasm, forming a ring-like pattern. The merged images (Figures 2D and 2H) revealed the clear co-localization of the two proteins in HeLa and COS-7 cells. The endogenous co-localization of TFPI-2 and PSAP was also evident in HUVECs and HASMCs (Figures 2L and 2P).
Co-localization of TFPI-2 and PSAP in cells
TFPI-2 inactivates the PSAP-induced enhancement of fibrosarcoma cell invasion and migration
A previous paper has reported that the restoration of TFPI-2 expression in a highly aggressive fibrosarcoma cell line, HT1080, dramatically reduced its growth and metastasis in athymic mice , and therefore we examined the possible roles of TFPI-2 and PSAP in stably transfected HT1080 cells expressing human TFPI-2. To evaluate the biological activities of PSAP, we generated a mammalian expression vector to express the biologically active full-length rhPSAP (Figure 3C). As shown in Figures 3(A) and 3(B), treatment of HT1080 cells with rhPSAP at 60 nM significantly stimulated their invasion and migration abilities. Also, PSAP in HT1080 and HT1080-TFPI-2 cells was down-regulated by siRNA knockdown (Figure 3D). Probably, the invasion and migration abilities of HT1080 cells transfected with PSAP siRNA were significantly weakened (Figures 3A and 3B). However, rhPSAP or PSAP siRNA did not induce dramatic changes in the invasion and migration abilities of HT1080-TFPI-2 cells, which stably expressed human TFPI-2 (Figures 3A and 3B). Quantification of PSAP mRNA and protein by RT–PCR and Western blotting respectively indicated that PSAP expression at the mRNA or protein level was not different between HT1080 cells and HT1080-TFPI-2 cells (Figure 3E). TFPI-2 markedly inhibited trypsin activity in a concentration-dependent manner; however, rhPSAP had no influence on the activity of trypsin, it did not affect the ability of TFPI-2 to inhibit the activity of trypsin (Figure 4). Overall, these results show that PSAP can stimulate HT1080 cell invasion and metastasis, and that these effects can be inhibited by TFPI-2.
TFPI-2 inactivates the PSAP-induced enhancement of HT1080 cell invasion and migration
rhPSAP did not affect the ability of TFPI-2 to inhibit the activity of trypsin
TFPI-2 blocks the PSAP-induced enhancement of MMP-2 activity in HT1080 human fibrosarcoma cells
MMP-2 and MMP-9 are often involved in the invasion and migration of tumour cells. The effects of PSAP on the activities of both MMPs were evaluated using gelatin zymography and fluorigenic substrate. When HT1080 cells were incubated with rhPSAP for 3 days, rhPSAP markedly increased MMP-2 activity in a concentration-dependent manner (Figures 5A and 5B), but did not show any effect on MMP-9 activity (Figure 5A and 5B). However, neither MMP-2 nor MMP-9 activity was changed by rhPSAP when HT1080-TFPI-2 cells were incubated with rhPSAP for 3 days (Figures 5A and 5B). At the same time, rhPSAP had no influence on the mRNA (Figure 5C) or protein level of MMP-2 or MMP-9 (Figure 5D). These findings suggest that PSAP can increase MMP-2 activity in HT1080 cells, and that PSAP cannot influence MMP-2 activity in the presence of TFPI-2.
Effects of PSAP on the activities of MMPs
In the present study, we demonstrated a novel interaction between TFPI-2 and PSAP, and determined that the binding sites are the KD2 of TFPI-2 and the C-terminus of PSAP. The outcome of this interaction is the inhibition of PSAP; PSAP does not affect the function of TFPI-2.
It is known that the first KD of TFPI-2 appears to contain all of the structural elements necessary for serine proteinase inhibition . KD2 of TFPI-2 largely mediates the interaction of TFPI-2 with gC1qR, a ubiquitously expressed cellular protein involved in modulating complement, coagulation and the kinin cascade . In the present study, we showed that KD2 of TFPI-2 contributes to the interaction with the C-terminus of PSAP. Using co-immunoprecipitation binding assays, we revealed that TFPI-2 interacted with PSAP in transfected HEK-293T cells and non-transfected HUVECs overexpressing TFPI-2 and PSAP. The same extracellular interaction was also observed in the medium of transfected HEK-293T cells. This intracellular association was also confirmed by co-localization assays, which showed that TFPI-2 and PSAP were positioned in close proximity in mammalian cells, specifically co-localizing in the cytoplasm, where they formed a ring-like pattern. The endogenous association was also verified by co-localization assays, which demonstrated that TFPI-2 was located in the nucleolus and cytoplasm, whereas PSAP was located in the cytoplasm; these proteins co-localized in the cytoplasm. These results support the hypothesis that TFPI-2 interacts specifically with PSAP in vitro and in vivo under normal biological conditions, potentially allowing for functional modulation of these two proteins.
PSAP is a 53-kDa protein that is modified post-translationally to give a 65-kDa form and is further glycosylated to yield a 70-kDa secretory product . As a multifunctional protein, the precursor of four lysosomal activator proteins known as saposins in the lysosome, PSAP plays an important role in the hydrolysis of sphingolipids. Mutations in PSAP can cause lysosomal storage disease and sphingolipidoses, such as Gaucher's disease, metachromatic leukodystrophy and Krabbe disease [33–35]. It has also been demonstrated that PSAP is overexpressed in conditioned media of ER (oestrogen receptor)-positive MCF-7 and ER-negative MDA-MB-231 breast cancer cell lines, as well as in a human SV40-transformed breast epithelial cell line, HBL100 . Previous studies have shown that the close functional association between PSAP and ProCathD (pro-cathepsin D) may eliminate tissue barriers by facilitating proteolytic degradation of basement membrane glycoproteins [36,37]. To evaluate the biological activities of PSAP in HT1080 cells, we treated HT1080 cells with rhPSAP or down-regulated PSAP by siRNA knockdown. Our results are consistent with a recent report by Hu et al.  demonstrating that PSAP down-regulation decreases metastatic PCa (prostate cancer) cell adhesion, migration and invasion, whereas another recent report by Kang et al.  demonstrated that PSAP inhibits tumour metastasis via paracrine and endocrine stimulation of stromal p53 and Tsp-1 . With respect to PCa, PSAP is involved in PCa invasion and migration by a co-ordinated regulation of ceramide levels, CathD and β1A-integrin expression, and attenuation of the ‘inside-out’ integrin signalling pathway . The results of the present study revealed that rhPSAP promoted HT1080 cell migration and invasion. The invasion and migration abilities of HT1080 cells transfected with PSAP siRNA were significantly weakened. Furthermore, we confirmed that rhPSAP markedly increased MMP-2 activity in a concentration-dependent manner. The findings of the present study suggest that PSAP can faciliate the metastatic process via increasing MMP-2 activity. Therefore the roles of PSAP in invasive and metastatic progression of malignancies might be quite different and require additional detailed investigations. However, in HT1080-TFPI-2 cells stably expressing human TFPI-2, rhPSAP or PSAP siRNA did not induce dramatic changes in cell invasion and migration. In addition, it was found that rhPSAP did not affect the ability of TFPI-2 to inhibit the activity of trypsin, and the PSAP expression level did not change in the absence or presence of TFPI-2.
Restoration of TFPI-2 can decrease the invasion and metastatic abilities of the malignant meningioma cell line IOMM-Lee, the human glioblastoma cell line SNB19 and the fibrosarcoma cell line HT1080 [16,40,41]. In athymic mice, HT1080 cells expressing wild-type TFPI-2 produced considerably smaller subcutaneous tumours and exhibited a lower metastatic rate . The precise mechanism whereby genetically engineered expression of functional TFPI-2 by TFPI-2-null cells reduces tumour size and its aggressive phenotype in vivo is unclear. Several studies indicate that TFPI-2 is a potent inhibitor of plasmin, which activates MMPs, which are involved in ECM degradation, thus regulating tumour cell invasion and migration [26,27,42–44]. However, plasmin associated with the ECM or the membranes of cultured cells is resistant to inhibition by known physiologically relevant proteinase inhibitors [45–47], and it has been suggested that metastatic tumour cells generate ‘unregulated’ plasmin activity that potentiates metastatic behaviour [48,49]. In the context of our present findings, it is tempting to speculate that, through the interaction with PSAP, TFPI-2 inactivates PSAP-induced enhancement of HT1080 cell invasion and migration, thereby contributing to tumour inhibition. The results of the present study may provide the basis for better understanding of the role of TFPI-2 and PSAP in cancer progression.
In conclusion, we have shown in the present study that PASP is a novel TFPI-2-interacting protein and that the binding sites are the KD2 of TFPI-2 and the C-terminus of PSAP. On the basis of the results of the present study, we hypothesize that TFPI-2 acts as an inhibitor of PSAP, leading to the inhibited abilities of invasion, migration and even metastasis of malignant tumour cells.
Dulbecco's modified Eagle's medium
human aortic smooth muscle cell
human embryonic kidney
human umbilical vein endothelial cell
recombinant human PSAP
small interfering RNA
tissue factor pathway inhibitor
Duan Ma and Zuohua Mao designed the experiment, analysed the data, and contributed to writing the paper. Chundi Xu and Fenge Deng carried out most of the experiments and analyses, and helped in paper preparation and scientific discussions. Jing Zhang, Huijun Wang, Jiping Wang, Jingui Mu and Shanshan Deng performed experiments and contributed data analyses. All authors have read and approved the final paper.
We thank the laboratory technician Yuliang He for help with confocal microscopy.
This work was supported by the National Key Basic Research Program of China [973 Program, grant number 2009CB941704] and the National Natural Sciences Foundation of China [grant number 30670687].
These authors contributed equally to the present study.