Tumor dormancy induced by downregulation of urokinase receptor in human carcinoma involves integrin and MAPK signaling

Abstract

Mechanisms that regulate the transition of metastases from clinically undetectable and dormant to progressively growing are the least understood aspects of cancer biology. Here, we show that a large ( approximately 70%) reduction in the urokinase plasminogen activator receptor (uPAR) level in human carcinoma HEp3 cells, while not affecting their in vitro growth, induced a protracted state of tumor dormancy in vivo, with G(0)/G(1) arrest. We have now identified the mechanism responsible for the induction of dormancy. We found that uPA/uPAR proteins were physically associated with alpha5beta1, and that in cells with low uPAR the frequency of this association was significantly reduced, leading to a reduced avidity of alpha5beta1 and a lower adhesion of cells to the fibronectin (FN). Adhesion to FN resulted in a robust and persistent ERK1/2 activation and serum-independent growth stimulation of only uPAR-rich cells. Compared with uPAR-rich tumorigenic cells, the basal level of active extracellular regulated kinase (ERK) was four to sixfold reduced in uPAR-poor dormant cells and its stimulation by single chain uPA (scuPA) was weak and showed slow kinetics. The high basal level of active ERK in uPAR-rich cells could be strongly and rapidly stimulated by scuPA. Disruption of uPAR-alpha5beta1 complexes in uPAR-rich cells with antibodies or a peptide that disrupts uPAR-beta1 interactions, reduced the FN-dependent ERK1/2 activation. These results indicate that dormancy of low uPAR cells may be the consequence of insufficient uPA/uPAR/alpha5beta1 complexes, which cannot induce ERK1/2 activity above a threshold needed to sustain tumor growth in vivo. In support of this conclusion we found that treatment of uPAR-rich cells, which maintain high ERK activity in vivo, with reagents interfering with the uPAR/beta1 signal to ERK activation, mimic the in vivo dormancy induced by downregulation of uPAR.

Figure 1

In vivo growth and cell cycle analysis of uPAR-rich (tumorigenic) and uPAR-deficient (dormant) cell lines. (A) Western blot of uPAR (R2 antibody, top) and ERK (loading control) (MK12 antibody, bottom) protein in lysates (50 μg protein each) of T-HEp3 (lane 1), D-HEp3 (lane 2), LK25 (lane 3), AS24 (lane 4), LK5 (lane 5), AS33 (lane 6), and AS48 (lane 7) cells. Lanes 8–10, purified soluble uPAR (suPAR) at 12, 60, and 120 ng, respectively (experiment was repeated three times). (B) Growth of HEp3 cells on CAMs. T-HEp3 (empty squares) or D-HEp3 cells (filled squares) were inoculated on CAMs and at the indicated times live tumor cells were counted (see Materials and Methods). Mean and SD (n = 4) of the number of cell divisions shown (experiment done twice). (C and D) Cell cycle analysis. T-HEp3 (C) and D-HEp3 (D) were inoculated on CAMs at 1–2 × 10⁶/CAM, and at the indicated times single tumor cell suspensions were prepared and processed for FACS^® analysis based on DNA content (see Materials and Methods). The percentage of cells in each stage of the cell cycle is indicated: G₀/G₁ (empty bars), S phase (filled bars) and G₂/M (striped bars). Each result represents the mean and SEM for at least three CAMs. Similar results were obtained in three additional experiments. *P < 0.005, ^# P < 0.015, as determined by Kruskal-Wallis statistics. A comparison of G₀/G₁ and S phases of T-HEp3 and D-HEp3 cells after 3 d of growth on CAMs showed statistically significant differences, P = 0.000 and P = 0.001, respectively.

Figure 1

In vivo growth and cell cycle analysis of uPAR-rich (tumorigenic) and uPAR-deficient (dormant) cell lines. (A) Western blot of uPAR (R2 antibody, top) and ERK (loading control) (MK12 antibody, bottom) protein in lysates (50 μg protein each) of T-HEp3 (lane 1), D-HEp3 (lane 2), LK25 (lane 3), AS24 (lane 4), LK5 (lane 5), AS33 (lane 6), and AS48 (lane 7) cells. Lanes 8–10, purified soluble uPAR (suPAR) at 12, 60, and 120 ng, respectively (experiment was repeated three times). (B) Growth of HEp3 cells on CAMs. T-HEp3 (empty squares) or D-HEp3 cells (filled squares) were inoculated on CAMs and at the indicated times live tumor cells were counted (see Materials and Methods). Mean and SD (n = 4) of the number of cell divisions shown (experiment done twice). (C and D) Cell cycle analysis. T-HEp3 (C) and D-HEp3 (D) were inoculated on CAMs at 1–2 × 10⁶/CAM, and at the indicated times single tumor cell suspensions were prepared and processed for FACS^® analysis based on DNA content (see Materials and Methods). The percentage of cells in each stage of the cell cycle is indicated: G₀/G₁ (empty bars), S phase (filled bars) and G₂/M (striped bars). Each result represents the mean and SEM for at least three CAMs. Similar results were obtained in three additional experiments. *P < 0.005, ^# P < 0.015, as determined by Kruskal-Wallis statistics. A comparison of G₀/G₁ and S phases of T-HEp3 and D-HEp3 cells after 3 d of growth on CAMs showed statistically significant differences, P = 0.000 and P = 0.001, respectively.

Figure 2

Phosphorylation of ERK1/2 in uPAR-rich and uPAR-deficient cells and its role in cell growth in culture. (A). Activation of ERK. Confluent cultures were serum-starved for 24 h, incubated with fresh serum-free medium for 5 min, scraped, lysed, and analyzed by Western blotting (50 μg protein/lane) and tested for phospho-ERK (top), and after stripping of the membrane, for ERK1/2 (bottom) (experiment repeated six times). (B) Inhibition of growth by MEK-1 inhibitor. LK25 (empty squares) or AS24 (filled squares) were seeded in 24-well plates (1.5 × 10⁵ cell/well), grown in 10% FBS overnight, washed, and incubated with daily medium changes for 72 h, with or without the indicated doses of MEK-1 inhibitor-PD98059. The cells were detached and counted in a Coulter counter. Shown are mean and SD of triplicate determinations. Inset: LK25 and AS24 cell monolayers, grown as in B, treated for 1 h without (control) or with 10 μM of PD98059. Phospho-ERK levels shown: lanes 1 and 3, controls for LK25 and AS24 cells, respectively; lanes 2 and 4, PD98059 treatment of LK25 and AS24 cells, respectively (both experiments were repeated three times).

Figure 2

Phosphorylation of ERK1/2 in uPAR-rich and uPAR-deficient cells and its role in cell growth in culture. (A). Activation of ERK. Confluent cultures were serum-starved for 24 h, incubated with fresh serum-free medium for 5 min, scraped, lysed, and analyzed by Western blotting (50 μg protein/lane) and tested for phospho-ERK (top), and after stripping of the membrane, for ERK1/2 (bottom) (experiment repeated six times). (B) Inhibition of growth by MEK-1 inhibitor. LK25 (empty squares) or AS24 (filled squares) were seeded in 24-well plates (1.5 × 10⁵ cell/well), grown in 10% FBS overnight, washed, and incubated with daily medium changes for 72 h, with or without the indicated doses of MEK-1 inhibitor-PD98059. The cells were detached and counted in a Coulter counter. Shown are mean and SD of triplicate determinations. Inset: LK25 and AS24 cell monolayers, grown as in B, treated for 1 h without (control) or with 10 μM of PD98059. Phospho-ERK levels shown: lanes 1 and 3, controls for LK25 and AS24 cells, respectively; lanes 2 and 4, PD98059 treatment of LK25 and AS24 cells, respectively (both experiments were repeated three times).

Figure 3

Analysis of uPA/uPAR–dependent signaling to ERK. (A) Effect of scuPA on phospho-ERK. Subconfluent cultures of LK25 or AS24 in 6-well plates were serum-starved for 24 h, acid-stripped (to remove uPAR-bound uPA), and incubated for 5 min with the indicated doses of scuPA and 200 KIU/ml of aprotinin (aprotinin was also used in B and C). The levels of phospho-ERK1/2 and ERK (lower panels) were determined by Western blot (see Materials and Methods); the exposure time in upper left panel is 1 s and in the upper right panel is 10 s (the experiment was repeated three times); the graphs show the Phospho-ERK/ERK ratio (R) obtained by densitometric scanning. (B) Effect of MEK-1 inhibitor on scuPA-induced ERK activation. LK25 or AS24 serum-starved cells were pretreated for 15 min with 10 μM of MEK-1 inhibitor PD98059, or medium alone. The monolayers were acid-stripped, incubated with 10 nM scuPA or medium alone, for 5 min, with or without 10 μM PD98059, lysed, and analyzed for phospho-ERK (upper panels) or ERK levels (lower panels) (experiment was repeated twice). (C) Time dependence of ERK activation by scuPA. LK25 or AS24 cell monolayers were acid-stripped and incubated for the indicated times with 10 nM scuPA. Phospho-ERK (upper panels) and ERK (lower panels) levels were determined (experiment was repeated twice).

Figure 4

Adhesion to FN and analysis of integrin expression and activation in tumorigenic and dormant cells. (A, left) Adhesion to FN. Cells (2.5 × 10⁴ per well of 96-well plates, four wells per experimental point) were inoculated into wells coated with FN, incubated at 37°C for 30 min, fixed, and stained with crystal violet (see Materials and Methods), the dye was extracted, and the absorbance was measured at 570 nm. The inter-well absorbance differences were <15%. (A, right) Adhesion to FN of different tumorigenic (T-HEp3 and LK25) or dormant (AS24, AS33, and D-HEp3) cell lines (10⁴ cells/well) into FN-coated (10 μg/ml) 96-well plates. Adhesion was measured after 20 min at 37°C as in A (left). (B) Surface expression of α5 and β1 integrins. Expression was measured using FACS^® analysis as described in Materials and Methods. The antibodies used: isotype-matched mouse IgG-control for α5; anti-α5 (P1D6); anti-β1 (AIIB2). Cells treated with rat IgG similar to mouse IgG (not shown). (C). Stimulation and inhibition of cell adhesion to FN. Cells were incubated either with β1 or α5β1 blocking antibodies (AIIB2 or BIIG2, respectively) or with stimulating (TS2/16) anti-β1 integrin antibody, or without antibodies, and inoculated (10⁴ cells/well) into FN-coated (10 μg/ml) 96-well plates. Adhesion was measured after 20 min at 37°C as in A. The effect of 1.5 mM MnCl₂ was tested without preincubation. Adhesion in medium without antibodies or MnCl₂ was considered 100% (experiment was repeated three times).

Figure 5

Modulation by FN of ERK activation and cell growth in tumorigenic and dormant cells. (A) Activation of ERK by FN. 6-well plates were coated with PL (4 μg/ml) or with 0.4 and 4 μg/ml FN. HEp3 cells (left) or D-HEp3 cells (right) (1 × 10⁶ cells/well) were plated, allowed to attach for 20 min, lysed, and analyzed for phospho-ERK (upper panels) and ERK (lower panels) levels (experiment was repeated three times). (B) Time-course of FN-induced ERK activation. Cells plated into wells precoated with 4 μg/ml of PL or FN were incubated for 10, 20, 40, or 90 min, lysed, and examined for phospho-ERK and ERK levels (experiment was repeated three times). LK25 and AS24 cells plated on CLI and examined after 10 and 20 min showed no difference in ERK activation. The ratio of phosphoERK to ERK determined by densitometry was 0.34 and 0.43 for LK25, 10 and 20 min, and 0.37 and 0.39 for AS24 at 10 and 20 min, respectively. (C) Analysis of phospho-JNK. Cells (10⁶) were plated in 6-well plates as follows: LK25 on PL (lane 1), on FN (lane 2), AS24 on PL (lane 3), on FN (lane 4), and analyzed after 20 min of incubation for phospho-JNK by Western blotting (the intensities of the bands remained constant for up to 90 min, not shown; experiment was repeated twice). (D and E) Effect of FN on cell growth. LK25 (D) or AS24 (E), serum-starved cells were plated at 0.8 × 10⁵ cells/well on 24-well plates coated with 0.04 (diamonds), 0.4 (circles), and 4 (triangles) μg/ml FN and at the indicated times, detached, and counted in a Coulter counter. Data shown as mean and SD of triplicate determinations (experiments were done three times). * P < 0.01 for all concentrations of FN versus uncoated (squares) surface, as determined by ANOVA test.

Figure 5

Modulation by FN of ERK activation and cell growth in tumorigenic and dormant cells. (A) Activation of ERK by FN. 6-well plates were coated with PL (4 μg/ml) or with 0.4 and 4 μg/ml FN. HEp3 cells (left) or D-HEp3 cells (right) (1 × 10⁶ cells/well) were plated, allowed to attach for 20 min, lysed, and analyzed for phospho-ERK (upper panels) and ERK (lower panels) levels (experiment was repeated three times). (B) Time-course of FN-induced ERK activation. Cells plated into wells precoated with 4 μg/ml of PL or FN were incubated for 10, 20, 40, or 90 min, lysed, and examined for phospho-ERK and ERK levels (experiment was repeated three times). LK25 and AS24 cells plated on CLI and examined after 10 and 20 min showed no difference in ERK activation. The ratio of phosphoERK to ERK determined by densitometry was 0.34 and 0.43 for LK25, 10 and 20 min, and 0.37 and 0.39 for AS24 at 10 and 20 min, respectively. (C) Analysis of phospho-JNK. Cells (10⁶) were plated in 6-well plates as follows: LK25 on PL (lane 1), on FN (lane 2), AS24 on PL (lane 3), on FN (lane 4), and analyzed after 20 min of incubation for phospho-JNK by Western blotting (the intensities of the bands remained constant for up to 90 min, not shown; experiment was repeated twice). (D and E) Effect of FN on cell growth. LK25 (D) or AS24 (E), serum-starved cells were plated at 0.8 × 10⁵ cells/well on 24-well plates coated with 0.04 (diamonds), 0.4 (circles), and 4 (triangles) μg/ml FN and at the indicated times, detached, and counted in a Coulter counter. Data shown as mean and SD of triplicate determinations (experiments were done three times). * P < 0.01 for all concentrations of FN versus uncoated (squares) surface, as determined by ANOVA test.

Figure 6

Physical and functional association between uPAR and α5β1 integrin. (A) Coimmunoprecipitation of β1-integrin and uPAR. T-HEp3, D-HEp3, or AS24 cells were lysed in Triton X-100 buffer (see Materials and Methods), centrifuged, and the pellets were extracted with RIPA buffer. Cell proteins (400 μg) were mixed with anti–β1 antibody (TS2/16) or G beads (G-control) or isotype-matched mouse IgG (IgG-control) and the resulting IPs were analyzed by Western blotting with anti–β1 integrin antibodies (upper panel) or anti–uPAR antibodies (lower panel) (experiment was repeated twice). (B, left) Cells were surface-biotinylated and the cell extracts were subjected to IP with anti-α5β1 (BIIG2), anti-uPAR (R2), anti-α₅ (P1D6), or with isotype-matched IgG antibodies and analyzed by streptavidin-HRP binding to biotinylated proteins after SDS-PAGE and transfer to PVDF membranes. The arrows on the left indicate the positions of α5, β1, and uPAR. (B, right) Surface-biotinylated proteins were immunoprecipitated with anti–β1 antibodies (TS2/16) and the biotinylated coimmunoprecipitating proteins were detected as described above. The arrows indicate the position for α integrin, β1 integrin, and uPAR. (C) Western blot for β1 integrin from whole cell lysates from the two tumorigenic (T-HEp3 and LK25) or the two dormant (D-HEp3 and AS24) cell lines. The arrows indicate the position of β1 and pre-β1 integrin.

Figure 7

Effect of β1-integrin–uPAR complex disruption and suPAR treatment on ERK activation. (A) Effect of soluble uPAR on ERK activation. LK25 (top panels) or AS24 (bottom panels) monolayers were serum-starved for 24 h and incubated with suPAR for 5 min in the presence of 200 KIU/ml of aprotinin. The cells were lysed and the proteins were analyzed for phospho-ERK and ERK levels (experiment was repeated three times). (B) Effect of anti–uPAR antibodies on ERK activation. T-HEp3 cells were incubated in suspension with medium alone (lanes 1 and 2), 7 μg/ml of isotype-matched IgG (lane 3), or with 7 μg/ml of anti–uPAR antibody (R2, lane 4) at 37°C for 35 min and inoculated into plates coated with 4 μg/ml PL (lane 1) or FN (lanes 2–4), allowed to adhere for 10 min, lysed, and analyzed for phospho-ERK (upper panel) and ERK (lower panel) levels. The numbers on top of each lane show the ratio of phospho-ERK/ERK determined by densitometry. (Experiment repeated twice). (C) Effect of peptide 25 (P25) on ERK phosphorylation. Subconfluent monolayers of T-HEp3 cells were treated with peptide 25 or its scrambled version (SP25) for 10 min, the cells were lysed, and phospho-ERK and ERK levels were determined. (D) Effect of anti–β1 antibody and peptide 25 on adhesion-activated ERK. Suspensions of T-HEp3 cells were incubated with or without anti–β1 antibodies (AIIB2) or isotype-matched IgG for 30 min, the cells plated on plastic plates coated with 4 μg/ml of PL or a combination of PL + 4 μg/ml FN and after 15 min tested for phospho-ERK and ERK levels. In addition, T-HEp3 cells were plated on FN in the presence or absence of 5 μM peptide 25 and tested as above.

Figure 8

Analysis of ERK activation and the role of β1-integrin/uPAR signaling in vivo. (A) HA-ERK expression and phosphorylation in culture. LK25 (lanes 1, 2, and 4) or AS24 (lane 3) cells were either transfected (lanes 2, 3, and 4) or mock-transfected (lane 1) with 10 μg of expression vector coding for hemagglutinin-tagged ERK2 (HA-ERK2). After 48 h in culture, cell lysates were subjected to IP with an anti–HA antibody (lanes 1–3) or isotype-matched IgG (lane 4). IPs were analyzed by Western blotting using either anti–HA antibodies (lower panel) or anti phospho-ERK antibodies (upper panel) (experiment was repeated three times). (B) HA-ERK expression and phosphorylation in vivo. T-HEp3 (lanes 1, 2, and 6), LK25 (lane 3), D-HEp3 (lane 4), or AS24 (lane 5) cells were transfected as in A but, after 24 h in culture, the cells were detached and inoculated on the CAMs (2.5 × 10⁶ cells/CAM, see Materials and Methods). After 24 h of growth in vivo, the CAMs containing tumor cells were analyzed by IP using anti–HA antibodies (lanes 1–5) or isotype-matched IgG (lane 6). HA-ERK2 was determined by Western blotting with anti–HA (lower panel) and its level of phosphorylation with anti–phospo-ERK antibody (upper panel). The numbers below lanes 2 and 3 in A and 2, 3, 4, and 5 in B indicate the ratio of phospho-HA-ERK2/HA-ERK2 determined by densitometry. (C) The effect of anti-β1 blocking antibody or anti-uPAR (domain 3) antibody on T-HEp3 proliferation in vivo. T-HEp3 cells were incubated for 40 min with medium alone (C), 15 μg/ml of isotype-matched IgG (IgG), 15 μg/ml of anti–CD55/DAF mAb (CD55), 7 μg/ml of monoclonal anti-uPAR (R2) antibody, 10 μg/ml of anti-β1 blocking mAb (AIIB2) or R2 + AIIB2 (same concentrations, R2/AIIB2). Cells (10⁶ HEp3 cells/CAM) were inoculated on CAMs of 9-d-old chick embryos and, after 3 d, the number of cells per tumor was determined (see Materials and Methods). (R2, AIIB2, and antibody combination treatment was repeated twice.)

Figure 8

Analysis of ERK activation and the role of β1-integrin/uPAR signaling in vivo. (A) HA-ERK expression and phosphorylation in culture. LK25 (lanes 1, 2, and 4) or AS24 (lane 3) cells were either transfected (lanes 2, 3, and 4) or mock-transfected (lane 1) with 10 μg of expression vector coding for hemagglutinin-tagged ERK2 (HA-ERK2). After 48 h in culture, cell lysates were subjected to IP with an anti–HA antibody (lanes 1–3) or isotype-matched IgG (lane 4). IPs were analyzed by Western blotting using either anti–HA antibodies (lower panel) or anti phospho-ERK antibodies (upper panel) (experiment was repeated three times). (B) HA-ERK expression and phosphorylation in vivo. T-HEp3 (lanes 1, 2, and 6), LK25 (lane 3), D-HEp3 (lane 4), or AS24 (lane 5) cells were transfected as in A but, after 24 h in culture, the cells were detached and inoculated on the CAMs (2.5 × 10⁶ cells/CAM, see Materials and Methods). After 24 h of growth in vivo, the CAMs containing tumor cells were analyzed by IP using anti–HA antibodies (lanes 1–5) or isotype-matched IgG (lane 6). HA-ERK2 was determined by Western blotting with anti–HA (lower panel) and its level of phosphorylation with anti–phospo-ERK antibody (upper panel). The numbers below lanes 2 and 3 in A and 2, 3, 4, and 5 in B indicate the ratio of phospho-HA-ERK2/HA-ERK2 determined by densitometry. (C) The effect of anti-β1 blocking antibody or anti-uPAR (domain 3) antibody on T-HEp3 proliferation in vivo. T-HEp3 cells were incubated for 40 min with medium alone (C), 15 μg/ml of isotype-matched IgG (IgG), 15 μg/ml of anti–CD55/DAF mAb (CD55), 7 μg/ml of monoclonal anti-uPAR (R2) antibody, 10 μg/ml of anti-β1 blocking mAb (AIIB2) or R2 + AIIB2 (same concentrations, R2/AIIB2). Cells (10⁶ HEp3 cells/CAM) were inoculated on CAMs of 9-d-old chick embryos and, after 3 d, the number of cells per tumor was determined (see Materials and Methods). (R2, AIIB2, and antibody combination treatment was repeated twice.)