The hVps34-SGK3 pathway alleviates sustained PI3K/Akt inhibition by stimulating mTORC1 and tumour growth

Abstract

We explore mechanisms that enable cancer cells to tolerate PI3K or Akt inhibitors. Prolonged treatment of breast cancer cells with PI3K or Akt inhibitors leads to increased expression and activation of a kinase termed SGK3 that is related to Akt. Under these conditions, SGK3 is controlled by hVps34 that generates PtdIns(3)P, which binds to the PX domain of SGK3 promoting phosphorylation and activation by its upstream PDK1 activator. Furthermore, under conditions of prolonged PI3K/Akt pathway inhibition, SGK3 substitutes for Akt by phosphorylating TSC2 to activate mTORC1. We characterise 14h, a compound that inhibits both SGK3 activity and activation in vivo, and show that a combination of Akt and SGK inhibitors induced marked regression of BT-474 breast cancer cell-derived tumours in a xenograft model. Finally, we present the kinome-wide analysis of mRNA expression dynamics induced by PI3K/Akt inhibition. Our findings highlight the importance of the hVps34-SGK3 pathway and suggest it represents a mechanism to counteract inhibition of PI3K/Akt signalling. The data support the potential of targeting both Akt and SGK as a cancer therapeutic.

Keywords: NanoString; PI3K and NDRG1; SGK3; mTORC1; mTORC2; protein kinase inhibitors; signal transduction inhibitors.

Figure 1. Prolonged treatment with Akt inhibitors leads to upregulation of SGK3

1. A, B

   ZR‐75‐1 and CAMA‐1 cell lines were treated with 1 μM MK‐2206 or 1 μM AZD5363 for the indicated time periods. Cell lysates were subjected to (A) immunoblot analysis with the indicated antibodies or (B) mRNA isolation followed by cDNA preparation. Real‐time PCR was performed on cDNA samples using specific primers against SGK1, SGK2 and SGK3 isoforms. Relative mRNA levels were calculated using 2^(−ΔΔ) C _t method using DMSO‐treated samples as a calibrator. Results are presented as relative mRNA level means ± SD for triplicates.

2. C

   ZR‐75‐1 and CAMA‐1 cells were treated with 1 μM MK‐2206 or 1 μM AZD5363 for one hour (1 h) or 10 days. After 7‐day treatment, SGK3 was knocked down by using three different shRNA probes, named SGK3 A, B and C and compared to a control shRNA probe, named sh scramble. Cells were maintained in the presence or absence of the indicated inhibitor during this period. At day 10, cells were lysed and subjected to immunoblot analysis with the indicated antibodies.

Source data are available online for this figure.

Figure 2. Prolonged treatment with Class I PI3K inhibitors leads to upregulation of SGK3

1. The indicated cell lines were treated with either 1 μM MK‐2206, 1 μM AZD5363, 1 μM GDC0941 or 1 μM BKM120 for the indicated times. Cell lysates were subjected to immunoblot analysis with the indicated antibodies.

2. The indicated cells were treated as in (A) and SGK3 was immunoprecipitated from the lysates using an anti‐SGK3 antibody. The immunoprecipitates (IP) were subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction (upper panel) followed by immunoblot analysis with the indicated antibodies (lower panel). Kinase reactions are presented as means ± SD for triplicate reaction.

Source data are available online for this figure.

Figure EV1. Prolonged treatment with Class I PI3K and Akt inhibitors leads to upregulation of SGK3 mRNA

ZR‐75‐1 (left panel), CAMA‐1 (middle panel) and T47D (right panel) cells were treated for 5 days with either 1 μM MK‐2206, 1 μM AZD5363, 1 μM GDC0941 or 1 μM BKM120. mRNA isolation was followed by cDNA preparation. Real‐time PCR was performed on cDNA samples using specific primers against SGK3 isoform. Relative mRNA levels were calculated using 2^(−ΔΔ) Ct method using DMSO‐treated samples as calibrator. Results are presented as relative mRNA level means ± SD for triplicates.

Figure 3. SGK3 activity induced by inhibition of PI3K/Akt is regulated by hVps34 and mTORC2

1. ZR‐75‐1 cells were treated with 1 μM MK‐2206 for 5 days and then, 1 h prior to cell lysis, cells were further treated with increasing doses of VPS34‐IN1. Cell lysates were subjected to immunoblot analysis with the indicated antibodies.

2. ZR‐75‐1 cells cultured in serum in the absence of Akt inhibitor were treated for 1 h with the indicated concentrations of VPS34‐IN1. The cell lysates were analysed by immunoblot using the indicated antibodies.

3. ZR‐75‐1 cells were treated for 5 days with either 1 μM MK‐2206, 1 μM AZD5363, 1 μM GDC0941 or 1 μM BKM120. One hour prior to lysis, the cells were incubated in the presence or absence of 1 μM VPS34‐IN1. SGK3 was immunoprecipitated from lysates and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction (upper panel). Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) and lysates were analysed by immunoblot with the indicated antibodies. One‐hour (1‐h) treatment with the PDK1 inhibitor GSK2334470 (Najafov et al, 2011) (1 μM) was used as a control for SGK3 inhibition.

4. ZR‐75‐1 cells were stably transfected with either a control shRNA vector (scrambled) or a shRNA vector that targets Rictor expression (shRictor). The cells were grown in the presence or absence of 1 μM MK‐2206 or 1 μM GDC0941 for 5 days. SGK3 was immunoprecipitated from the lysates and subjected to in vitro kinase assay as in (C). Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) and lysates (lower panel) were also subjected to immunoblot analysis with the indicated antibodies.

5. ZR‐75‐1 cells were cultured in the absence or presence of 1 μM MK‐2206 for 5 days. Cells were then treated in the absence or presence of 0.1 μM AZD8055 or 0.1 μM rapamycin for 1 h. SGK3 was immunoprecipitated and subjected to in vitro kinase assay as in (C). Kinase reactions are presented as means ± SD for triplicate reaction. The immunoprecipitates (IP) and lysates were analysed with the indicated antibodies.

Figure EV2. Further evidence that SGK3 activity induced by inhibition of PI3K/Akt is regulated by hVps34

1. ZR‐75‐1 cells were treated with 1 μM MK‐2206 for 5 days, and then, 1 h prior to cell lysis, cells were further treated with increasing doses of SAR405.

2. ZR‐75‐1 cells cultured in serum in the absence of Akt inhibitor were treated for 1 h with the indicated concentrations of SAR405. The cell lysates were analysed by immunoblot using the indicated antibodies.

3. ZR‐75‐1 cells were treated for 5 days with 1 μM MK‐2206 or 1 μM GDC0941. One hour prior to lysis, the cells were incubated in the presence of absence of 0.3 μM SAR405. SGK3 was immunoprecipitated from lysates and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction (upper panel). Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) and lysates were also analysed by immunoblot with the indicated antibodies.

Figure 4. PtdIns(3)P binding to PX domain of SGK3 promotes phosphorylation and activation by PDK1

1. A, B

   SGK3 [S486E]‐GST and SGK3 [R90A S486]‐GST were purified from HEK293 cells transiently overexpressing these enzymes. One hour prior to lysis, cells were treated with 5 μM VPS34‐IN1 and 5 μM GSK2334470 to ensure that the SGK3 was in its inactive dephosphorylated form. SGK3[S486E]‐GST (A) or the non‐PtdIns(3)P‐binding mutant SGK3[R90A, S486E]‐GST (B) was incubated with lipid vesicles comprising phosphatidylcholine (PC) and phosphatidylserine (PS) containing the indicted concentrations of PtdIns or PtdIns(3)P in the presence or absence of added recombinant PDK1 (50 ng) and kinase reactions were initiated by addition of MgATP. After 30 min at 30°C, PDK1 was inhibited by addition of the 1 μM GSK2334470 PDK1 inhibitor and the reaction mixture was supplemented with 2 μg GST‐NDRG1 SGK3 substrate. After another 30 min at 30°C, the reaction was terminated by addition of SDS sample buffer. The reaction mixtures were subjected to immunoblot analysis with the indicated antibodies.

Figure 5. 14h selectively suppresses both the activity and activation of SGK3 by PDK1 and mTORC2

1. Chemical structure of the Sanofi‐14h SGK inhibitor.

2. IC₅₀ values of Sanofi‐14h SGK inhibitor on the indicated recombinant kinases.

3. Protein kinase profiling undertaken against the Dundee panel of 140 protein kinases in the presence of 1 μM Sanofi‐14h at the International Centre for Protein Kinase Profiling. The result for each kinase is presented as a mean kinase activity of the reaction taken in triplicate relative to a control reaction where the inhibitors were omitted. Abbreviations and assay conditions are described at http://www.kinase-screen.mrc.ac.uk.

4. ZR‐75‐1 cells were treated for 1 h with the indicated concentrations of 14h. The cell lysates were analysed by immunoblot analysis using the indicated antibodies.

5. ZR‐75‐1 cells were treated for 1 h with the indicated concentrations of 14h. SGK3 was immunoprecipitated from cell lysates and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction (upper panel). Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) were also analysed by immunoblot with the indicated antibodies.

6. The effect of the indicated concentration of 14h on the ability of SGK3[S486E]‐GST to be activated by PDK1 in the presence of PtdIns(3)P was assessed as described in Fig 4.

Figure 6. SGK3 counteracts inhibition of the PI3K/Akt pathway by phosphorylating TSC2 and stimulating S6K1

ZR‐75‐1 cells were treated with 1 μM MK‐2206 for the indicated times. SGK3 (upper panel) and S6K1 (middle panel) were immunoprecipitated and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction. Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) were also analysed by immunoblot with the indicated antibodies. The cell lysates were also analysed by immunoblot using the indicated antibodies (lower panel).

Figure 7. SGK3 counteracts inhibition of the PI3K/Akt pathway by phosphorylating TSC2 and stimulating mTORC1

ZR‐75‐1 cells were treated for 1 h or 5 days with 1 μM MK‐2206, 1 μM GDC0941 or 3 μM 14h inhibitors, alone or in combination, as indicated.

1. The cell lysates were analysed by immunoblot using the indicated antibodies.

2. SGK3 (upper panel), Akt1 (middle panel) and S6K1 (lower panel) were immunoprecipitated from the same cell lysates and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide for kinases in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min, 30°C reaction. Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) were also analysed by immunoblot with the indicated antibodies.

3. SGK3 was knocked down in ZR‐75‐1 cells by using shRNA probe B and compared to a control shRNA probe, named sh scramble. After infection, the cells were kept for 2 days in puromycin selection media and then seeded for the experiment. The cells were treated with 1 μM MK‐2206 for 1 h or 5 days. The cell lysates were subjected to immunoblot analysis with the indicated antibodies.

Figure EV3. Further evaluation of Class I PI3K/Akt and hVps34 inhibition on mTORC1 activity and phosphorylation of Akt substrates

1. ZR‐75‐1 cells were treated for 1 h with 1 μM MK‐2206, 1 μM GDC0941 or 3 μM 14h inhibitors, alone or in combination, as indicated. The cell lysates were analysed by immunoblot using the indicated antibodies.

2. ZR‐75‐1 cells were treated for 1 h or 5 days with 1 μM MK‐2206, 1 μM GDC0941 or 1 μM VPS34‐IN1 inhibitors, alone or in combination, as indicated. The cell lysates were analysed by immunoblot using the indicated antibodies.

3. ZR‐75‐1 cells were treated for 1 h or 5 days with 1 μM MK‐2206, 1 μM GDC0941 or 3 μM 14h inhibitors, alone or in combination, as indicated. The cell lysates were analysed by immunoblot using the antibody against phosphorylated Akt phosphorylation consensus motif (RxRxxpS/pT). Red asterisk indicates potential dual Akt and SGK3 substrates whose phosphorylation is induced following 5 day MK‐2206 or GDC0941 treatment and suppressed by a subsequent 1‐h treatment with 14h. Blue circles indicates potential Akt selective substrates whose phosphorylation is inhibited by MK‐2206 or GDC0941 treatment and but do not become re‐phosphorylated following 5‐day incubation with these inhibitors under conditions where SGK3 is upregulated.

Figure 8. Dual treatment with Akt and SGK inhibitors reduces tumour growth in BT‐474 xenograft model

1. BT‐474 cells were treated for the indicated times with 0.3 μM MK‐2206. The cell lysates were analysed by immunoblot using the indicated antibodies.

2. BT‐474c cells were treated with inhibitors as indicated either as monotherapy or in combination and cell confluency measured on the Incucyte ZOOM every 4 h for up to 4 weeks.

3. BT‐474 cells were injected subcutaneously into athymic Foxn1 ^nu nude mice. Mice were treated with either vehicle (8 mice) or MK‐2206 (100 mg/kg) (10 mice) or 14h (25 mg/kg) (6 mice) or both, MK‐2206 and 14h (10 mice) for 24 days. The tumour volume was measured twice a week. Tumour growth is represented as the fold change mean ± SEM.

4. All mice were weighed at the end of the treatments. Results are presented as mean ± SD.

5. Plasma concentrations of MK‐2206 and 14h were analysed in samples obtained 2–3 h after the administration of the last dose on the 24^th day of treatment. Results are presented as a mean ± SD from three different samples.

6. Tumours were harvested at the end of the experiment, 4 h after the last dosage and subjected to immunohistochemical analysis using cleaved caspase‐3 antibody (clCasp3). Apoptotic cells were counted in 25 fields per condition (left panel) and quantified as clCasp3‐positive cells/field (right panel). The results are presented as mean ± SD number of cleaved caspase‐3 positive cells. Representative images are shown. Scale bar, 100 μm.

7. Tumours were harvested at the end of the experiment, 4 h after the last dosage and subjected to immunohistochemical analysis with the indicated antibodies. Representative images are shown. Scale bar, 100 μm.

8. Tumours were obtained the same way as in (G) and subjected to immunoblot analysis with the indicated antibodies. Six different tumours were analysed from each treatment group and each line represents one tumour sample.

Figure EV4. Akt inhibition induces dose‐dependent inhibition of cell growth in BT‐474c cells in vitro and stimulates SGK3 to activate mTORC1

1. A, B

   BT‐474c cells were treated with DMSO, 3 μM 14h and the indicated doses of MK‐2206 (A) or AZD5363 (B) inhibitors. Cell confluency was measured on the Incucyte ZOOM every 4 h for up to 6 days.

2. C

   BT‐474c cells were treated for 1 h or 5 days with 1 μM MK‐2206 or 3 μM 14h, as indicated. SGK3 (upper panel) and S6K1 (middle panel) were immunoprecipitated and subjected to in vitro kinase assay by measuring phosphorylation of the Crosstide substrate peptide in the presence of 0.1 mM [γ‐³²P]ATP in a 30 min 30°C reaction. Kinase reactions are presented as means ± SD for triplicate reaction. Immunoprecipitates (IP) were also analysed by immunoblot with the indicated antibodies. The cell lysates were subjected to immunoblot analysis using the indicated antibodies (bottom panel).

Source data are available online for this figure.

Figure 9. Prolonged exposure to Akt or Class I PI3K inhibitors induces transcriptional changes in human kinome

1. A–C

   ZR‐75‐1 cells were treated for 5 days with 1 μM MK‐2206 (A), 1 μM AZD5363 (B), 1 μM GDC0941 (C) or DMSO. Human mRNAs were hybridised to NanoString human kinome and control code sets, then subjected to quantification using NanoString software. Results are presented as mRNA change of each kinase relative to mRNA isolated from control sample treated with DMSO. To permit data compaction, and simple kinome‐wide comparisons, the fold changes are log2 transformed. The kinase mRNAs exhibiting prominent changes are annotated, with SGK3 highlighted in red. Similar profiles were obtained in second independent experiment.