p38gamma mitogen-activated protein kinase integrates signaling crosstalk between Ras and estrogen receptor to increase breast cancer invasion

Abstract

Ras is believed to stimulate invasion and growth by different effector pathways, and yet, the existence of such effectors under physiologic conditions has not been shown. Estrogen receptor (ER), on the other hand, is both anti-invasive and proliferative in human breast cancer, with mechanisms for these paradoxical actions remaining largely unknown. Our previous work showed an essential role of p38gamma mitogen-activated protein kinase in Ras transformation in rat intestinal epithelial cells, and here, we show that p38gamma integrates invasive antagonism between Ras and ER to increase human breast cancer invasion without affecting their proliferative activity. Ras positively regulates p38gamma expression, and p38gamma in turn mediates Ras nonmitogenic signaling to increase invasion. Expression of the Ras/p38gamma axis, however, is trans-suppressed by ER that inhibits invasion and stimulates growth also by distinct mechanisms. Analysis of ER and its cytoplasmic localized mutant reveals that ER additionally binds to p38gamma protein, leading to its specific down-regulation in the nuclear compartment. A p38gamma-antagonistic activity of ER was further shown in a panel of breast cancer cell lines and was shown independent of estrogens by both ER depletion and ER expression. These results revealed that both Ras and ER use distinct pathways to regulate breast cancer growth and invasion, and that p38gamma specifically integrates their antagonistic activity to stimulate cell invasion. Selective targeting of p38gamma-dependent invasion pathways may be a novel strategy to control breast cancer progression.

Figure 1

Ras positively regulates p38γ protein expression independent of ERK phosphorylation, whereas both Ras and p38γ protein expression is trans-suppressed by ER in human breast cancer cells. A, Ras inhibition suppresses p38γ expression without affecting ERK/p38 phosphorylation. ER⁻ 231 cells were infected with control adenovirus (Ad-β-gal) or virus expressing dominant-negative Ras (Ad-N17) for 4 hours and incubated for an additional 24 hours before analyzed for protein expression by Western blot. B, Ras activation induces p38γ expression without stimulating ERK phosphorylation. Cells were infected with Ad-β-gal or virus expressing oncogenic H-Ras (Ad-L61) as in (A) and examined for protein expression. C and D, ER inhibits Ras/p38γ protein expression dependent of its transcription activity. Cells were cultured with and without Tet for the indicated time to induce wild-type (C) and the mutant ER expression (D), and their effects on Ras/p38γ protein expression were examined by Western. Representative from three separate experiments.

Figure 2

Endogenous Ras activity is required for both proliferation and invasion, whereas p38γ only transmits nonmitogenic Ras signaling to stimulate invasion. A, adenovirus-mediated N17 expression and small interfering RNA–induced p38γ depletion. Cells were infected with either adenovirus (Ad-β-gal or Ad-N17) or retrovirus (pSR-Lucif, control retrovirus; pSR-sip38γ, small interfering RNA retrovirus) and incubated for 24 hours before Western analysis. B, both Ras inhibition and p38γ depletion inhibit cell invasion. Cells were plated for invasion assay after infection, and invaded cells were counted and normalized to respective controls. Columns, mean of 15 fields (P < 0.01 for N17 and sip38γ versus their respective controls); bars, SD. Similar results were obtained from additional two experiments. C, Endogenous Ras but not p38γ is required for cell proliferation. Cells were infected as above, and cell proliferation was estimated by thymidine incorporation. Columns, mean of three experiments (P < 0.05 only for Ad-N17 versus Ad-Vect); bars, SD. D, N17 and p38γ overexpression. ER⁻ 231 cells were infected either with adenovirus (Vect, N17) and/or pLHCX retrovirus (Vect, p38γ) and analyzed for protein expression 24 hours later. E, p38γ overexpression rescues N17-mediated invasion inhibition. Cells were coinfected with Ad-N17 with and without pLHCX-p38γ and assessed for cell invasion as described above. Columns, mean of 13 fields (P < 0.01 for p38γ versus Vect, N17 versus Vect, and for N17 versus p38γ + N17); bars, SD. Similar results are obtained from one additional experiment. F, high levels of p38γ protein expression do not overcome N17-induced growth inhibition. Columns, mean of three separate experiments (P < 0.05 for N17 versus Vect, but P > 0.05 for p38γ versus Vect and N17 + p38γ versus N17); bars, SD.

Figure 3

ER inhibits invasion and stimulates growth by distinct mechanisms. A, ER protein expression in Tet-on cells. Tet-on 231 cells were incubated with and without Tet for 24 hours and examined for ER and ER/T311A protein expression. B, Thr³¹¹ is required for ER transcription activity. Cells were transfected with ERE-Luc, incubated with and without Tet for 24 hours, and assessed for luciferase activity. Columns, mean of five independent experiments; bars, SD. C, ER requires its transcription activity to inhibit invasion. Cells were cultured ± Tet for 22 hours in the invasion chamber for the assay (see Materials and Methods). Columns, mean from 16 fields in one experiment; bars, SD. Similar results were obtained in two additional experiments. The absorption at A _{600 nm} (OD600 ) is mean of three separate experiments. D, ER does not require its transcription activity to stimulate DNA synthesis. Cells were cultured for 24 hours with or without Tet and cell growth was estimated by thymidine incorporation.

Figure 4

ER antagonizes nuclear p38γ activity through direct binding. A, ER is nuclear while ER/T311A is predominantly cytoplasmic. Cells were incubated for 24 hours (±Tet) and double stained for ER (detected with Cy3-labeled second antibody) and for p38γ expression (detected with FITC-labeled second antibody) with 4′,6-diamidino-2-phenylindole (DAPI) staining as a control for nuclear signals. B, ER disrupts nuclear p38γ protein through Thr³¹¹-dependent binding. Cells were prepared for nuclear and cytoplasmic fractions, which were analyzed for protein expression (top). p38γ binding activity of ER was assessed by immunoprecipitation with a p38γ antibody followed by Western analyses (middle). For pulldown assays (bottom), the purified GST, GST-ER, and GST-ER/T311A proteins were incubated with equal amounts of lysates prepared from 293T cells transfected with flag-p38γ and flag-p38γ/AGF, and the precipitates were analyzed for p38γ-bound ER protein. C, overexpressed p38γ is cytoplasmic in ER⁺ cells. ER⁺ 231 cells (Tet-on cells incubated with Tet for 24 hours) were infected with Ad-Vect (Vect) or Ad-p38γ for 24 hours and stained with anti-flag antibody for transfected flag-p38γ. D, higher levels of p38γ proteins increase invasion (right) but not DNA synthesis (left) in ER⁺ breast cancer cells. Tet-on 231 cells were plated for invasion and thymidine incorporation after infection. Right, columns, mean from 18 fields of one representative experiment; bars, SD. Similar results were obtained from one additional experiment. Left, columns, mean of four experiments; bars, SD.

Figure 5

ER negatively regulates p38γ expression in a panel of human breast cancer cell lines. A, ER⁺ phenotype correlates with lower levels of p38γ RNA expression. Total RNA was prepared, separated, and transferred to a nitrocellulose membrane that was hybridized with a specific radioactive-labeled probe for p38γ expression. B, ER depletion increases p38γ expression. ER⁺ cells were infected with pSR-Lucif (Lucif) or pSR-siER (siER) for 24 hours and examined for p38γ RNA (left) or protein (right) expression. C, ER negatively regulates p38γ expression by ligand-independent mechanisms. Cells in steroid-depleted medium were incubated for 24 hours (±10 nmol/L E2) after the pSR infection (left) or in the presence or absence of Tet (right) and analyzed by Western analysis for protein expression.

Figure 6

An experimental model illustrates that ER and Ras regulate breast cancer progression through distinct invasive and proliferative pathways where p38γ only integrates their antagonistic activity to increase invasion. ER requires transcription activity to suppress Ras/p38γ expression and to bind p38γ protein, whereas it increases DNA synthesis by ERE-independent mechanisms. Ras, on the other hand, increases invasion and DNA synthesis through p38γ-dependent and p38γ-independent pathways in ER⁻ cells. Although loss of ER expression may reduce cell growth, resultant Ras activation will provide mitogenic signaling, leading to increased DNA synthesis [and possible cell proliferation (?)]. Induced p38γ following ER inactivation and/or Ras activation will specifically mediate Ras nonmitogenic signaling to stimulate invasion. Through these two mechanisms, Ras/p38γ activation will lead to an increased breast cancer progression. This model suggests a feasibility to individually target invasive and proliferative signal transduction pathways in human breast cancers.