Interaction of oestrogen receptor with the regulatory subunit of phosphatidylinositol-3-OH kinase

Abstract

Oestrogen produces diverse biological effects through binding to the oestrogen receptor (ER). The ER is a steroid hormone nuclear receptor, which, when bound to oestrogen, modulates the transcriptional activity of target genes. Controversy exists, however, concerning whether ER has a role outside the nucleus, particularly in mediating the cardiovascular protective effects of oestrogen. Here we show that the ER isoform, ER alpha, binds in a ligand-dependent manner to the p85alpha regulatory subunit of phosphatidylinositol-3-OH kinase (PI(3)K). Stimulation with oestrogen increases ER alpha-associated PI(3)K activity, leading to the activation of protein kinase B/Akt and endothelial nitric oxide synthase (eNOS). Recruitment and activation of PI(3)K by ligand-bound ER alpha are independent of gene transcription, do not involve phosphotyrosine adapter molecules or src-homology domains of p85alpha, and extend to other steroid hormone receptors. Mice treated with oestrogen show increased eNOS activity and decreased vascular leukocyte accumulation after ischaemia and reperfusion injury. This vascular protective effect of oestrogen was abolished in the presence of PI(3)K or eNOS inhibitors. Our findings define a physiologically important non-nuclear oestrogen-signalling pathway involving the direct interaction of ER alpha with PI(3)K.

Figure 1

Activation of eNOS by oestrogen is mediated by PI(3)K. a, b, Concentration-dependent (a) and time-dependent (b) effects of E₂ and wortmannin (WM, 30 nM) on eNOS activity (fold induction versus baseline) in human vascular endothelial cells. Asterisk indicates P < 0.05 compared with unstimulated or E₂ stimulation. c, E₂-stimulated NOS activity in murine p85α^+/+ and p85α^-/- fibroblasts (FB) transfected (Tx) with vector (pcDNA3), eNOS, ERα, p85α or dominant-negative p85α (Δp85α) cDNAs. Asterisk indicates P < 0.05 compared with transfection with eNOS cDNA alone; two askerisks indicate P < 0.05 compared with transfection with ERα and eNOS cDNAs.

Figure 2

Oestrogen stimulates ERα-associated PI(3)K activity. a, Effect of vehicle (ethanol 0.01% v/v), E₂ (10 nM) or insulin (100 nM) on endogenous PtdIns-3,4,5-P₃ levels. Asterisk indicates P < 0.05 compared with vehicle. b, Time-dependent effect of E₂ on ERα, p85α and PI(3)K activity (PIP₃) in ERα immunoprecipitate (IP). c, Effect of ICI (10 μM) or WM on E₂ or 17α-oestradiol (αE₂)-stimulated ERα-associated PI(3)K activity. Cells were pre-treated with ICI or WM for 30 min. d, Effect of tamoxifen (TM, 1 μM), PD 98059 (PD, 5 μM) and actinomycin D (ACT, 5 μM) on ERα-associated PI(3)K activity. Inhibitors were added 2 h before E₂ stimulation. e, Effect of E₂ or insulin (Ins) on p-Tyr- and IRS-1-associated PI(3)K activity. f, Effect of E₂, progesterone (Prog, 10 nM), testosterone (Test, 10 nM), thyroid hormone (Thyr, 10 nM), dexamethasone (Dex, 1 μM), WY14643 (WY, 100 μM) and 15-deoxy-Δ,-prostaglandin J₂ (PGJ₂, 100 μM) on PI(3)K activity in the corresponding steroid hormone nuclear receptor immunoprecipitates.

Figure 3

Ligand-dependent interaction of ERα with p85α. a, b, Effect of E₂ on ERα—p85α co-immunoprecipitation in non-transfected human endothelial cells (a) and murine p85α^-/- fibroblasts (b) transfected (Tx) with ERα and p85α, alone or in combination. c, d, Affinity purification using agarose-conjugated GST or GST—p85α (c), or GST—p85α amino-terminal SH2 domain (NSH2, amino acids 321–470), carboxy-terminal SH2 domain (CSH2, 576–724), or SH3 domain (NSH3, 1–80) fusion protein and human recombinant (hr) ERα (d). e, E₂- or insulin (Ins)-stimulated Akt kinase activity. Asterisk indicates P < 0.05 compared with no stimulation. f, Effect of E₂ on eNOS activity (fold induction over baseline) in endothelial cells transfected with adenovirus containing no Akt (vector), constitutively active (myr), or a dominant-negative (dn) Akt. Asterisk indicates P < 0.05 compared with vector alone.

Figure 4

PI(3)K and NO mediate the vascular protective effects of oestrogen. Cumulative histograms of leukocyte rolling velocities before (-) and after (+) ischaemia and reperfusion (I/R) are shown. a, b, Effect of superfused WM (100 nM) or L-nitroarginine methylester (L-NAME, 0.1 mM) on leukocyte rolling velocity (a) and leukocyte adhesion and eNOS activity in the murine cremaster muscle (b). Data are expressed as fold increase over baseline before I/R in the same paired venules. Asterisk indicates P < 0.001 compared with untreated after I/R (None). c, Representative video images showing the same venules before (-) and after (+) I/R with the indicated treatments. Scale bar, 40 μm.