Distinct roles of Akt1 and Akt2 in regulating cell migration and epithelial-mesenchymal transition

Abstract

The Akt family of kinases are activated by growth factors and regulate pleiotropic cellular activities. In this study, we provide evidence for isoform-specific positive and negative roles for Akt1 and -2 in regulating growth factor-stimulated phenotypes in breast epithelial cells. Insulin-like growth factor-I receptor (IGF-IR) hyperstimulation induced hyperproliferation and antiapoptotic activities that were reversed by Akt2 down-regulation. In contrast, Akt1 down-regulation in IGF-IR-stimulated cells promoted dramatic neomorphic effects characteristic of an epithelial-mesenchymal transition (EMT) and enhanced cell migration induced by IGF-I or EGF stimulation. The phenotypic effects of Akt1 down-regulation were accompanied by enhanced extracellular signal-related kinase (ERK) activation, which contributed to the induction of migration and EMT. Interestingly, down-regulation of Akt2 suppressed the EMT-like morphological conversion induced by Akt1 down-regulation in IGF-IR-overexpressing cells and inhibited migration in EGF-stimulated cells. These results highlight the distinct functions of Akt isoforms in regulating growth factor-stimulated EMT and cell migration, as well as the importance of Akt1 in cross-regulating the ERK signaling pathway.

Figure 1.

IGF-IR hyperstimulation induces morphological changes in mammary epithelial cells. (A) MCF-10A cells overexpressing human IGF-IR or vector control (Neo) were grown in monolayer cultures containing EGF ± IGF (100 ng/ml). Lysates were immunoblotted with antibody against phosphorylated IGF-IR or total receptor. (B) Phase-contrast images of IGF-IR or Neo cells grown in 3D Matrigel cultures for 16 d in media containing 2% horse serum and EGF ± IGF-I (100 ng/ml). B (C) IGF-IR, Neo, or MCF-10A cells overexpressing a variant IGF-IR (Y950F) were grown in 3D Matrigel cultures for 7 or 16 d in 2% horse serum, EGF, and 100 ng/ml IGF-I. (D) Lysates of IGF-IR, Neo, or Y950F cells grown in monolayer cultures in 2% horse serum, EGF, and 100 ng/ml IGF-I were immunoblotted with antibody against phosphorylated Akt or phosphorylated, activated ERK. Bars, 50 μM.

Figure 2.

Isoform-specific down-regulation of Akt results in distinct phenotypes. (A) IGF-IR cells were superinfected with shRNA vectors targeting Akt1 and/or -2. Empty vectors (V) were used as controls. Immunoblotting with antibody against actin was used to confirm equal loading. Cells in monolayer cultures were lysed and immunoblotted with Akt isoform-specific antibodies. (B) IGF-IR cells superinfected with empty vector, Akt1 shRNA, Akt2 shRNA, or both were grown in triplicate monolayer cultures for the indicated days in media containing EGF and IGF-I (100 ng/ml), with initial seeding of 2.5 × 10⁴ cells. Cells were trypsinized, stained with Trypan blue, and counted. A representative experiment of three independent experiments is shown. Error bars represent means ± SD. (C) IGF-IR cells in which either Akt1 and/or -2 were down-regulated were cultured in monolayer (top) or 3D Matrigel/collagen (50:50) cultures for 8 d (middle and bottom). Monolayer cultures were grown in the presence of EGF and IGF-I (100 ng/ml). 3D cultures were grown in the presence of 5 ng/ml EGF and 100 ng/ml IGF-I. Bars: (top and middle) 50 μM; (bottom) 100 μM.

Figure 3.

Akt2-specific down-regulation inhibits IGF-IR–induced changes in 3D cultures. IGF-IR cells expressing Akt2 shRNA vector were cultured in 3D Matrigel cultures for 8 or 16 d in the presence of EGF and IGF-I (100 ng/ml). Structures were stained with antibodies to cleaved caspase-3 or Ki-67 (green), α6 integrin (red), and TOPRO (blue). Equatorial confocal images are shown. Bars, 50 μM.

Figure 4.

Down-regulation of Akt1 in IGF-IR cells induces migration and EMT. (A, top) The motility of IGF-IR cells with isoform-specific down-regulation of Akt was assessed using transwell assays. Cells were starved overnight in 2% serum in the absence of EGF and IGF. Migration was assessed after 16–20 h in the presence or absence of 100 ng/ml IGF-I. The histogram displays the mean number of migrated cells obtained by counting 10 separate fields in three independent experiments. (bottom) IGF-IR cells with down-regulation of both Akt1 and -2 were starved and migration in the presence or absence of IGF-I was assessed. Shown is a representative experiment displaying the mean number of migrated cells in five separate fields. Error bars represent means ± SD. (B) IGF-IR cells infected with empty vector control (V) or shRNA vectors targeting Akt1 or -2 were grown in 2% horse serum and 100 ng/ml IGF-I and lysed in NP-40 (for E- and N-cadherin) or RIPA buffer (Vimentin). Lysates were immunoblotted with the indicated antibodies.

Figure 5.

ERK activation is enhanced by Akt1 down-regulation. (A) Neo or IGF-IR cells with isoform-specific down-regulation of Akt were grown in monolayer cultures, starved, and stimulated with 2% horse serum ± 100 ng/ml IGF-I in the absence of EGF. Lysates were immunoblotted with the indicated antibodies. (B) IGF-IR cells overexpressing empty vector control, Akt1 shRNA, or both Akt1 and -2 shRNA vectors were starved and stimulated with 100 ng/ml IGF-I in the absence of EGF. Lysates were immunoblotted as indicated.

Figure 6.

Expression of constitutively active MEK in IGF-IR cells results in invasive structures and repression of E-cadherin. (A) IGF-IR or Neo cells overexpressing vector control or constitutively active MEK2 (MEKDD) were grown in monolayer cultures containing EGF and IGF-I (100 ng/ml). Bar, 50 μM. (B) Transwell migration assay and western analyses of IGF-IR or Neo cells overexpressing MEKDD were performed. Cells were starved in the absence of EGF for 24 h. Migration was assessed after 16–20 h in media containing 2% horse serum and 100 ng/ml IGF-I. Lysates were immunoblotted as indicated. Error bars represent means ± SD. (C) IGF-IR or Neo cells overexpressing MEKDD were grown in 3D Matrigel/collagen (50:50) cultures for 12 d. All cells were maintained in 2% horse serum, EGF, and 100 ng/ml IGF-I. Phase-contrast images are shown. Bar, 50 μM. (D) IGF-IR or Neo cells overexpressing vector control (V) or MEKDD were grown in monolayer cultures with 2% horse serum and 100 ng/ml IGF-I. Cells were lysed in NP-40 lysis buffer and lysates were immunoblotted with the indicated antibodies.

Figure 7.

Inhibition of ERK signaling inhibits migration of Akt1 down-regulated cells. (A) Migration of IGF-IR cells expressing Akt1 shRNA or empty vector (V) control treated with DMSO, 2–10 μM UO126, or 10 μM SB 202190 was assessed using transwell migration assay. Cells were starved in the absence of EGF. Migration was assessed after 16–20 h in media containing 2% horse serum, 100 ng/ml IGF-I, and DMSO or inhibitor. The histogram displays the mean percentage migration relative to Akt1 down-regulated IGF-IR cells treated with DMSO within the same experiment. The mean values were derived from three independent experiments. Error bars represent means ± SD. (B) Akt1 down-regulated IGF-IR cells were grown in 2% serum, 100 ng/ml IGF-I, and DMSO, 2–10 μM UO126, or SB 202190 for 72 h, with the inhibitor replaced after 48 h. Lysates were immunoblotted as indicated. (C) IGF-IR cells expressing Akt1 or empty vector shRNA were cultured in 2% serum, IGF-I, and DMSO or 2–10 μM UO126 for 72 h, lysed in NP-40 lysis buffer, and immunoblotted with antibodies against E-cadherin. (D) IGF-IR cells expressing Akt1 or empty vector shRNA and cultured in 2% serum, IGF-I, and DMSO or 2–10 μM UO126 for 72 h were lysed in RIPA lysis buffer and immunoblotted with the indicated antibodies.

Figure 8.

Down-regulation of Akt1 enhances EGF-stimulated migration and ERK activation. (A) MCF-10A cells overexpressing empty vector or Akt1 or -2 shRNA vectors were generated. Isoform-specific down-regulation was confirmed by Western analysis and migration in response to EGF stimulation was assessed. Before plating in transwell assays, cells were starved overnight in the absence of EGF. Cells were stimulated with EGF at the indicated concentrations and migration was assessed. The histogram displays a representative experiment with mean values obtained by counting 10 independent fields. Error bars represent means ± SD. (B) MCF-10A cells overexpressing empty vector control or Akt1 or -2 shRNA vectors were starved in the absence of EGF. Cells were stimulated with the indicated concentrations of EGF, lysed, and immunoblotted with the indicated antibodies.

Figure 9.

Overexpression of Akt1 suppresses EGF-stimulated migration and ERK activation. (A) MCF-10A cells overexpressing HA-tagged wild-type Akt1 or -2 were generated. (top) Levels of overexpression were confirmed using antibodies against the HA tag or a panAkt antibody that recognizes all isoforms. (bottom) Levels of activated, phosphorylated Akt and phosphorylated GSK3β were assessed after EGF stimulation. (B) Migration of MCF-10A cells overexpressing vector control or wild-type HA-tagged Akt1 or -2 was assessed by transwell assay. Cells were starved in the absence of EGF for 24 h. Migration was assessed after 16–20 h in media with 2% horse serum ± 5 ng/ml EGF. A representative experiment is shown with mean values obtained by counting 10 separate fields. Error bars represent means ± SD. (C) MCF-10A cells overexpressing vector control or wild-type Akt1 or -2 were starved for 24 h and stimulated with media containing 2% horse serum ± 5 ng/ml EGF. Cells were lysed and immunoblotted with the indicated antibodies.