Proliferation and invasion: plasticity in tumor cells

Abstract

Invasive and proliferative phenotypes are fundamental components of malignant disease, yet basic questions persist about whether tumor cells can express both phenotypes simultaneously and, if so, what are their properties. Suitable in vitro models that allow characterization of cells that are purely invasive are limited because proliferation is required for cell maintenance. Here, we describe glioblastoma cells that are highly invasive in response to hepatocyte growth factor/scatter factor (HGF/SF). From this cell population, we selected subclones that were highly proliferative or displayed both invasive and proliferative phenotypes. The biological activities of invasion, migration, urokinase-type plasminogen activation, and branching morphogenesis exclusively partitioned with the highly invasive cells, whereas the highly proliferative subcloned cells uniquely displayed anchorage independent growth in soft agar and were highly tumorigenic as xenografts in immune-compromised mice. In response to HGF/SF, the highly invasive cells signal through the MAPK pathway, whereas the selection of the highly proliferative cells coselected for signaling through Myc. Moreover, in subcloned cells displaying both invasive and proliferative phenotypes, both signaling pathways are activated by HGF/SF. These results show how the mitogen-activated protein kinase and Myc pathways can cooperate to confer both invasive and proliferative phenotypes on tumor cells and provide a system for studying how transitions between invasion and proliferation can contribute to malignant progression.

Fig. 1.

Comparison of DB-P and selected subclones. (A) Thymidine incorporation by DB-P and subclones. Cells were serum-starved for 48 h and then left untreated (SS) or supplemented with 100 ng/ml HGF/SF (HGF) and assayed as described in Materials and Methods. (B) Western blotting showing Met protein level and activation of downstream signaling. (C) Invasion assay in Matrigel chambers. Cells (2 × 10⁴ per well) were seeded into Matrigel inserts. Control medium (SS) or medium containing 100 ng/ml HGF/SF (HGF) was added into the lower chamber. Cells that migrated through Matrigel and attached to the under surface of the filter were counted. The mean values from triplicate experiments (plus 1 SD) are presented. (D) HGF/SF-mediated uPA activation in DB-P and DB-A2 cells. The mean values (plus 1 SD) from three experiments are displayed.

Fig. 2.

Decreased branching morphogenesis in highly proliferative subclones. (A) Cells growing in 3D Matrigel were treated with 100 ng/ml HGF/SF (HGF) or without HGF/SF (C) for 3 d. The number of branching cells per 200 cells was counted, and the mean (plus 1 SD) from three separate counts is presented. (B) Branching morphogenesis after cell growing in 3D Matrigel for 10 d was observed under a phase-contrast microscope.

Fig. 3.

Anchorage-independent cell growth and in vivo tumorigenic assay. (A) Colony formation assay. DB-P or DB-A2 cells (2.5 × 10⁴ cells per well) were grown in soft agar with 100 ng/ml HGF/SF (HGF) or without HGF/SF (C) for 16 d. (Left) Representative areas from triplicate experiments are presented. Visible colonies (>0.5 mm in diameter) were counted after crystal violet staining. (Right) The mean value (plus 1 SD) of three independent experiments is shown. (B) Enhanced tumorigenic activity of DB-A2. DB-P or DB-A2 cells (3 × 10⁶ cells per mouse) were s.c. injected into 6-week-old athymic female (nu/nu) mice. Tumor size was estimated by caliper measurements after 8 d and then at 3-d intervals. The mean volumes (plus 1 SD) of tumors from nine mice for each group are presented as a function of time after injection.

Fig. 4.

Different signaling pathways are induced by HGF/SF in DB-P and proliferative subclones. (A) HGF/SF-induced phosphorylation of Met receptor in DB-P and DB-A2 cells. Total Met immunoprecipitates recovered with polyclonal antibody C-28 were analyzed by immunoblotting with antibodies directed against Met (25H2) or phosphotyrosine (4G10). (B) The activation of signal pathways in DB-P, DB-A2, and DB-A6 cells. (C) Myc expression is enhanced during DB-P tumorigenesis. DB-P or DB-A2 cells before (C) or after (T) growing in nude mice were analyzed by immunoblotting to estimate levels of total cellular Met, Myc, and β-actin. DB-P tumors (T20, T28, and T29) were collected after 49 d, and DB-A2 tumors (T5, T8, and T12) were collected after 31 d of growth.

Fig. 5.

Consequences of Myc overexpression in DB-P cells. (A) Ectopic expression of Myc mRNA in transfected DB-P clones. The expression of ectopic myc (cmv-myc) and endogenous myc (Endo-myc) was examined by RT-PCR. (B) Expression of Myc protein in transfected clones assayed by immunoblotting. (C) Thymidine incorporation assay showing increased DNA synthesis in Myc-expressing clones. Cells cultured with (FBS) or without serum (SS) were left untreated (SS and FBS) or treated with HGF/SF (SS-HGF and FBS-HGF) for 12 h. (D) In vitro invasion assay of DB-Pc and myc-expressing clones. (E) Enhanced colony formation activity on soft agar in Myc-expressing clones. Cells (1 × 10⁴ cells per well) were grown in soft agar and feed with DMEM supplemented with serum alone (C) or with HGF/SF (HGF) for 16 d. Visible colonies >0.5 mm in diameter were counted after crystal violet staining. The mean value (plus 1 SD) of three independent experiments is presented.

Fig. 6.

Characterization of highly invasive revertants from DB-A2 cells. (A) Branching morphogenesis assay. Cells were treated with 100 ng/ml HGF/SF (HGF) or without treatment (C) for 3 d. The number of branching cells per 200 cells was counted under a microscope, and the mean (plus 1 SD) from three separate counting is presented. (B) Branching morphogenesis of A2-BH7 cells after cell growing for 7 d in 3D Matrigel was observed under a phase-contrast microscope. (C) In vitro invasion assay. Control medium (SS) or medium containing 100 ng/ml HGF/SF (HGF) was added into the lower chamber. The mean number of invading cells from triplicate experiments (plus 1 SD) is presented. (D) HGF/SF-mediated uPA activation in DB-A2 revertant. The mean values (plus 1 SD) from three experiments are displayed. (E) Thymidine incorporation assay. Cells were serum-starved and untreated (SS) or supplemented with 100 ng/ml HGF/SF (HGF) for 12 h. (F) The activation of signal pathways in DB-P, DB-A2, and A2-BH7 cells. Total cell lysates were resolved by electrophoresis and analyzed by immunoblotting.