Detection of PI3K Inhibition in Human Neuroblastoma Using Multiplex Luminex Bead Immunoassay

NB Cell Lines and Culture

An established panel of nine NB cell lines was used. Human NB cell lines SK-N-SH, SH-SY5Y, SK-N-BE(2), SK-N-AS, SK-N-FI, SK-N-MC, and IMR-32 were purchased from the American Type Culture Collection (ATCC; Manassas, VA). Cell lines NB5 ¹² and NB1691 ¹³ were established at St. Jude Children’s Research Hospital. All cell lines were maintained as monolayer culture at 37 °C in a humidified atmosphere with 5% CO₂ according to ATCC recommendations. NB5 and NB1691 were cultured and maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and nonessential amino acids.

Chemotherapeutic Agents

OSI-906 (OSI; insulin-like growth factor–1 receptor inhibitor) and NVP-BEZ235 (BEZ; PI3K/mTOR inhibitor) were purchased from ChemieTek (Indianapolis, IN), NVP-BKM120 (BKM; PI3K inhibitor) was obtained from Novartis (Basel, Switzerland), and everolimus (mTOR inhibitor) was purchased from Selleck (Houston, TX).

Cell Treatments

All compounds were dissolved in DMSO to obtain stock solutions of 20 mM, which were diluted appropriately at the time of use in culture medium. The final concentration of DMSO did not exceed 0.5% and did not alter cell viability.

Cell Viability Dose-Response Curves

Exponentially growing cells were plated in 96-well microtiter plates at 2.5 × 10³ cells/well. The day after plating, the medium was replaced with medium containing increasing concentrations of specific PI3K compounds along with the appropriate controls (DMSO only at 0.5% final concentration as a negative control and staurosporine at 10 µM as a positive control). Cells were incubated in the presence of drugs for 72 h followed by addition of an equal volume CellTiter-Glo reagent (Promega, Madison, WI) according to the manufacturer’s protocol. Plates were rocked on a rotator for 2 min, and total luminescence was measured 10 min later on a BioTek Synergy 2 plate reader (BioTek, Winooski, VT). Readings were converted to percentage inhibition relative to the DMSO-only control, dose-response (DR) curves were constructed, and IC₅₀ values were calculated using a four-parameter logistic model curve fit built into the GraphPad Prism software package. Each point on the curves represents the mean value of readings from eight wells.

Assessment of PI3K Pathway by WB and Luminex Bead Assays

Whole-cell extracts of untreated and treated NB cell lines (SK-N-AS, SK-N-FI, IMR-32, and NB5) were prepared by lysis on ice using the following buffer: 10 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) sodium dodecyl sulfate, and 10% (v/v) glycerol. Immediately before use, one protease inhibitor cocktail tablet (Roche, Indianapolis, IN) and one PhosSTOP tablet (Roche) were added per 10 mL of buffer, and phenylmethanesulfonyl fluoride was added to 1 mM final concentration. Total protein (40 µg) was electrophoresed at 160 V for 65 min (PowerPac 200; Bio-Rad, Hercules, CA) on 10% sodium dodecyl sulfate–polyacrylamide gels followed by transfer to nitrocellulose membranes (Li-Cor Biosciences, Lincoln, NE) in a Mini Trans-Blot cell (Bio-Rad) at 80 mA for 60 min. Nonspecific binding was blocked with Tris-buffered saline containing 5% (w/v) dry milk for 1 h at room temperature. Blots were then incubated overnight at 4 °C with primary antibodies (Cell Signaling Technology, Danvers, MA) specific for AKT, glycogen synthase kinase-3β (GSK-3β), p70S6K, and PRAS40 (both total protein and phosphorylation site–specific forms). All primary antibodies were used at conditions recommended by the manufacturer. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal loading control. The membranes were washed and incubated with secondary antibodies coupled to IRDye 680LT or IRDye 800 (Li-Cor Biosciences). Immunoreactive proteins were visualized and signal intensity was quantified using Odyssey Infrared Imaging System (Li-Cor Biosciences).

Luminex xMAP assays were performed on a Bio-Plex 200 instrument (Bio-Rad). Analyte-specific bead kits and required reagents were obtained from Invitrogen (Carlsbad, CA) or EMD Millipore (Darmstadt, Germany), and assays were performed as recommended by the manufacturer. Single-analyte kits (AKT[pS473], PRAS40[pT246], p70S6K[pT389], and GSK-3β[pS9]; total protein or phosphorylation site–specific forms) were used. All kits used were obtained from Invitrogen with the exception of p70S6K[pT389], which was obtained from EMD Millipore. In preliminary tests, we determined that 10 µg of protein was the optimum input amount per assay to be able to detect low amounts of the protein of interest and stay below signal saturation. The data output gives the mean fluorescence intensity (mFI) as a measure of protein abundance. We calculated the ratio of phosphoprotein mFI to total protein mFI as a measure of PI3K pathway activity.

To demonstrate the equivalency of WB and Luminex xMAP data, a Pearson correlation calculation was performed with the aid of a function built into the GraphPad Prism software package.

Baseline Level of PI3K Pathway Activation in NB Cell Lines

Extracts from nine NB cell lines were tested for the presence and activation of key components of the PI3K pathway with WB. Figure 1A shows the results of probing with antibodies specific for the survival kinase AKT, ¹⁴ the fully activated form of AKT phosphorylated at serine 473 (pS473-AKT), and GAPDH, which was used as an internal normalization standard. Although all cell lines contained significant levels of total AKT protein, the level of AKT activation as measured by phosphorylation at serine 473 varied significantly among the NB cell lines. We quantified the signal intensity for total AKT and pS473-AKT and adjusted that intensity to the strength of the GAPDH signal. The adjusted signals for total protein were then normalized to the highest signal and plotted to get the relative total AKT protein level for all nine cell lines ( Fig. 1B ). The adjusted signals for the phosphorylated protein were then divided by the corresponding total protein signal for each cell line to obtain a measure of phosphorylation. The ratios were also normalized to the highest ratio and plotted to obtain the relative level of AKT phosphorylation for all cell lines ( Fig. 1C ). The same experiments and analyses were also completed for the mTORC1 regulator and AKT target PRAS40 ¹⁵ and the regulatory kinases GSK-3β, ¹⁶ also an AKT target, and for p70S6K, ¹⁷ a target of mTORC1 and indicator of mTORC1 activity. ¹⁸

OPEN IN VIEWER

Figure 1.

AKT pathway activation in neuroblastoma (NB) cell lines. (A) Western blot (WB) analysis of AKT activation in nine NB cell lines. Blot was probed with antibodies specific for total AKT protein, pS473 AKT, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Signals were quantitated and normalized to the GAPDH signal. The obtained values are plotted in panels B and C. (B) Relative levels of four key components of the PI3K/AKT/mTOR kinase (PI3K) cascade. For AKT, WB was probed with antibodies directed against total PRAS40, GSK-3β, and p70S6K. After quantitation and normalization to the GAPDH signal, the relative intensities for each protein in all cell lines were plotted with the highest intensity set to 1.0. (C) Relative phosphorylation levels of four key components of the PI3K cascade. The intensity of the signal for the phosphorylated form was divided by the intensity of the total protein signal for each of the four proteins shown in B. The ratios were plotted with the highest ratio set to 1.0.

As shown in Figure 1B , the levels of total AKT and p70S6K protein varied up to approximately twofold among the cell lines. In contrast, GSK-3β levels vary 8-fold, and those of PRAS40 vary more than 10-fold. A distinct theme emerged when examining the relative phosphorylation levels ( Fig. 1C ). Phosphorylation of p70S6K and GSK-3β showed relatively minor variations (up to twofold), whereas the phosphorylation levels of PRAS 40 varied about eightfold. The most dramatic phosphorylation fluctuation was exhibited by AKT, with a robust 50-fold difference between IMR32 and SK-N-FI.

DRs of NB Cell Lines to Inhibitors Targeting the PI3K Pathway

To identify NB cell lines that are either sensitive or resistant to PI3K pathway inhibitors, DR curves for single agents (OSI, BEZ, BKM, and everolimus) were established using the panel of nine NB cell lines. The IC₅₀ values obtained from these experiments are listed in Table 1 . From this test cohort, we selected four representative cell lines that biochemically reflected the high and low end of the spectrum of both DR to the drugs used and activation levels of key PI3K components. The DR curves for the four cell lines selected are shown in Supplementary Figure 1.

OPEN IN VIEWER

Table 1.

Cytotoxicity of PI3K pathway inhibitors to NB cell lines.

Cell Line	Drug	OSI-906	BEZ-235	BKM-120	Everolimus
IMR-32		7.8 ± 5.5 µM	47 ± 17 nM	0.43 ± 0.14 µM	>10 µM
SK-N-SH		>10 µM	5.0 ± 3.3 nM	1.8 ± 1.3 µM	>10 µM
SH-SY5Y		0.19 ± 0.11 µM	0.25 ± 0.22 µM	0.18 ± 0.33 µM	>10 µM
SK-N-FI		1.4 ± 0.9 µM	25 ± 7.1 nM	0.43 ± 0.13 µM	>10 µM
SK-N-AS		2.9 ± 2.1 µM	16 ± 6.5 nM	0.31 ± 0.10 µM	>10 µM
SK-N-MC		>10 µM	40 ± 33 nM	0.94 ± 0.26 µM	>10 µM
NB5		0.71 ± 0.60 µM	27 ± 16 nM	0.26 ± 0.11 µM	>10 µM
SK-N-BE(2)		>10 µM	0.32 ± 0.15 µM	1.4 ± 0.2 µM	>10 µM
NB1691		2.2 ± 1.4 µM	4.0 ± 2.3 nM	0.96 ± 0.43 µM	>10 µM

The cytotoxicity of PI3K pathway inhibitors to neuroblastoma (NB) cells was determined using CellTiter-Glo assays. Cells were incubated in the presence of compounds for 72 h before adding the CellTiter-Glo reagent. Dose-response curves were constructed and fitted to a four-parameter logistic model in GraphPad Prism (curves shown in Suppl. Fig. 1), and the resulting IC₅₀ values were tabulated. Results were calculated from three independent replicates.

All NB cell lines were particularly sensitive to the dual PI3K/mTOR inhibitor BEZ, with IC₅₀ values ranging from 4 nM (1691) to 260 nM (SK-N-BE(2); Table 1 ). They were less sensitive to the broad-spectrum PI3K inhibitor BKM, with IC₅₀ values ranging from 0.18 µM (SH-SY5Y) to 1.8 µM (SK-N-SH). NB5, SK-N-AS, and SK-N-FI were sensitive to the IGF-1R inhibitor OSI, whereas IMR-32 was relatively insensitive to it ( Table 1 ). All NB cell lines were insensitive to everolimus, judging solely by the IC₅₀ values ( Table 1 ). However, examination of the DR curves showed that all NB cell lines were sensitive to everolimus over a wide range of concentrations. All lines exhibit a reduction in growth to 60% to 70% control, with the drug becoming cytotoxic only at concentrations greater than 10 µM (Suppl. Fig. 1).

Changes in Phosphorylation Levels of PI3K Pathway Effectors in NB Cells upon Drug Application

The alteration of phosphorylation levels of key PI3K components was next evaluated in response to treatment with the PI3K/mTOR dual inhibitor BEZ. We treated NB5 cells with increasing drug doses from 0.01 µM to 10 µM for 4 h and analyzed the changes in the levels of phosphorylation of AKT, PRAS40, p70S6K, and GSK-3β by WB. Results are shown in Supplementary Figure 2 (left panel). We quantified these signals, calculated a ratio of the intensity of the phosphorylated protein signal to the total protein signal, normalized each ratio to the DMSO-only control ratio, and plotted these values in Figure 2A (top left panel).

OPEN IN VIEWER

Figure 2.

Dose response of NB5 to PI3K pathway inhibitors. (A) NB5 cells were treated with 0, 0.01, 0.1, 1, and 10 µM BEZ, BKM, and everolimus. Cells were harvested after 4 h of exposure to the drugs, lysed, and analyzed by Western blot (WB) for AKT, PRAS40, p70S6K, and GSK-3β phosphorylation levels (for WB blot images, see Suppl. Fig. 1). Signal intensities (mean fluorescence intensity [mFI]) were quantitated and used to calculate the ratio of phosphorylated mFI to total mFI as a measure of phosphorylation level. These values were then normalized to no-drug control and plotted. (B) Same as in A, except quantitation was done by xMAP assay. (C) Correlation plot of WB versus Luminex xMAP data.

Whereas the signals for the level of the total protein in cell extracts remained unchanged, the phosphorylation level decreased with increasing drug concentrations for all four proteins. AKT phosphorylation at serine 473 decreased most dramatically and was nearly undetectable at higher concentrations. A decreased phosphorylation pattern was to be expected, because BEZ inhibits mTOR kinase and serine 473 is phosphorylated by the mTOR kinase complex mTORC2. ¹⁸ Similarly, a decline in p70S6K phosphorylation, which is a target of the mTOR kinase complex mTORC-1, ¹⁸ was also observed when cells were treated with BEZ.

Both PRAS40 and GSK-3β are targets of regulation by AKT, ¹⁴ and, as expected, reduction of both phosphoproteins was observed upon treatment with increasing BEZ concentrations. Phosphorylation of Ser473 is required for full AKT activation; therefore, a lack phosphorylation results in less active AKT and a subsequent decline in phosphoproteins that are direct AKT substrates. PRAS40 phosphorylation was more affected than GSK-3β phosphorylation, probably because, unlike PRAS40, GSK-3β can also be targeted by other cellular kinases.

Given that WB procedures are laborious and time-consuming and that our experimental outline required the processing and testing of a large number of samples, we validated the use of the Luminex xMAP assay using a BioPlex-200 Suspension array system from Bio-Rad with the same samples used in Figure 2A . The Luminex xMAP assay is a microplate-based multiplex immunoassay system that can simultaneously quantify multiple analytes in a single sample, combining aspects of ELISA, immunoprecipitation, and flow cytometry. The assay uses spectrally coded beads coupled with antibodies for capture of specific target proteins. Fluorescent tags conjugated to detection antibodies for the same targets are added in a second step. The fluorescence intensity readout measures the abundance of those targets. We measured the abundance of total and phosphorylated proteins with Luminex xMAP assays and processed the results in the same manner as described for Figure 2B above. Those results are shown in Figure 2B (top panel). Next, NB5 cells were treated with the PI3K inhibitor BKM ( Fig. 2A , B , middle panels) and the mTORC1 inhibitor everolimus ( Fig. 2 , bottom panels). As with BEZ, we observed a decrease in phosphorylation levels with increasing concentrations of BKM. Higher concentrations of BKM were required to produce similar levels of inhibition, correlating well with the higher IC₅₀ value for BKM. As expected, treatment of the cells with the mTORC1 inhibitor everolimus inhibited only p70S6K phosphorylation, leaving AKT, PRAS40, and GSK-3β phosphorylation unchanged ( Fig. 2A , B , bottom panels).

We then tabulated the values for the phosphorylation levels obtained by WB versus those obtained with xMAP and performed a correlation analysis. The resulting plot is shown in Figure 2C . A Pearson correlation calculation gave us a correlation coefficient of 0.9371 with a two-tailed p value of <0.0001. Overall, the WB data correlate very well with the xMAP data. Although both methods give a decrease in both p70S6K and PRAS40 phosphorylation levels, the apparent decrease seen by WB is higher than that observed by xMAP assays. This may reflect a much higher signal-to-noise ratio for xMAP assays than for WB. The WB signals for the phosphorylated form of p70S6K and PRAS40 were very low and barely above background, whereas the signals obtained by xMAP were much stronger, which may very well explain the apparent discrepancy in the magnitude of the decrease in signal. Based on these observations, Luminex xMAP multiplex assays were used for the following DR studies.

DR Studies Using PI3K Cascade Inhibitors in NB Cell Lines

The NB cell lines NB5, SK-N-FI, SK-N-AS, and IMR32 were exposed to increasing doses of single agents (BEZ, BKM, OSI, and everolimus) at 0.01, 0.1, 1, and 10 µM concentrations, and the levels of AKT, PRAS40, p70S6K, and GSK-3β, both phosphorylated and total protein, were measured by xMAP 2, 4, 8, and 24 h after drug treatment. We calculated the ratio of the signal intensity of phosphorylated protein to total protein and normalized all values to the 2 h, 0.01 µM value to obtain a measure of change in the level of protein phosphorylation in response to an increase in drug concentration and time of exposure. The results for the treatment of NB5 with all four compounds are shown in Figure 3 .

OPEN IN VIEWER

Figure 3.

Pharmacodynamics of NB cell line NB5 exposed to PI3K pathway inhibitors. NB5 cells were treated with increasing doses of NVP-BEZ235, NVP-BKM120, OSI-906, and everolimus at 0.01, 0.1, 1, and 10 µM final concentrations. After 2, 4, 8, and 24 h of exposure to the inhibitors, cells were harvested and cell extracts prepared. The abundance of total and phosphorylated protein for AKT (blue bars), PRAS40 (red bars), p70S6K (green bars), and GSK-3β (purple bars) in those extracts was measured by xMAP assay, and ratios of phosphorylated protein to total protein were calculated and normalized to the solvent-only control for each analyte at each time point.

The level of pS473-AKT ( Fig. 3 , blue bars) was affected the most by increasing doses of BEZ. At a 0.1 µM concentration of the inhibitor, a 90% loss of phosphorylation was observed and showed clear dose dependence. This effect was immediate and did not change much from 2 to 24 h. AKT serine 473 was phosphorylated by mTORC2, ¹⁸ and the decrease in phosphorylation was likely a reflection of the inhibition of the mTOR kinase by BEZ. Inhibition of the second mTOR kinase containing complex mTORC1 by BEZ was apparent due to loss of phosphorylation at threonine 389 of p70S6K, a direct mTORC1 target ( Fig. 3 , green bars). The level of phospho-PRAS40 ( Fig. 3 , red bars) also decreased with increasing drug concentration. PRAS40 is a direct substrate of AKT, ¹⁵ and therefore a decrease in cellular AKT phosphorylation (and its kinase activity) should result in the observed decrease of PRAS40 phosphorylation as expected. GSK-3β is also an AKT substrate and, as expected, we observed a decrease in its phosphorylation at the 2 and 4 h time points ( Fig. 3 , purple bars). However, at later time points, the GSK-3β phosphorylation levels rebounded. Because GSK-3β serine-9 can also be phosphorylated by cellular kinases other than AKT (e.g., p90RSK), the rebound in GSK-3β phosphorylation levels may reflect this activity.

A similar pattern emerged from treatment of NB5 with the PI3K inhibitor BKM ( Fig. 3 , second panel from top). Higher concentrations of BKM were needed to achieve a similar decrease in phosphorylation of all four proteins tested, not surprising considering the higher IC₅₀ value of this compound. With the exception of AKT, phosphorylation levels recovered slightly at the 24 h time point. Unlike what was observed for treatment with BEZ, GSK-3β phosphorylation levels did not recover as much at the 8 and 24 h time points when cells were treated with BKM, suggesting other pathways targeting serine-9 of GSK-3β are also inhibited by BKM.

Likewise, it would be expected that the IGF-1R inhibitor OSI would inhibit the PI3K pathway as well ( Fig. 3 , third panel from top). However, only AKT phosphorylation at Ser473 was strongly suppressed initially (by more than 80%). Although both p70S6K and GSK-3β were reduced by as much as 60%, PRAS40 phosphorylation was practically not affected at the measurement time points. Although reduced initially at 2 h, the abundance of GSK-3β phosphoprotein recovered at 4 h and increased at 8 and 24 h, perhaps due to secondary activation of a compensating pathway.

The mTORC1 inhibitor everolimus strongly reduced p70S6K phosphorylation, as would be expected, at all concentrations ( Fig. 3 , bottom panel). This effect was immediate and persisted at all time points. Aside from diminutive fluctuations, little effect on the phosphorylation levels of PRAS40 and GSK-3β was seen. AKT phosphorylation was not affected initially but decreased in a dose-dependent manner at 8 h, and the nadir was at 24 h.

Next, we treated the NB cell line SK-N-FI with the four PI3K antagonists. The results are shown in Supplementary Figure 2. SK-N-FI and NB5 share the common biochemical signature characterized by elevated levels of AKT phosphoprotein (Suppl. Fig. 2). Overall, the results obtained for SK-N-FI reflect those for NB5. BEZ treatment led to a rapid dose-dependent reduction of AKT and p70S6K phosphorylation by 80% and 90%, respectively. PRAS40 and GSK-3β phosphorylation levels decreased as well and, as was seen for NB5, the GSK-3β phosphorylation level recovered after 24 h. During this time interval, minimal AKT phosphorylation was observed. BKM, OSI, and everolimus treatment resulted in a duplicative manner of dose- and time-dependent phosphorylation levels in SK-N-FI and NB5. A notable exception was the level of GSK-3β phosphorylation, which did not fluctuate in SK-N-FI as much as in NB5 when the cells were treated with OSI.

A distinctively different and unpredictable phosphoprotein pattern emerged when SK-N-AS and IMR32 were subjected to anti-PI3K treatment. Both cell lines showed attenuated basal levels of activated AKT expression, and although the phospho-AKT signal was well above background in SK-N-AS cells, it was barely detectable by WB ( Fig. 1 ) or xMAP in IMR32 cells. For these reasons, phospho-AKT levels in IMR32 are not displayed in Supplementary Figure 4. Only p70S6K phosphorylation levels responded substantially to BEZ, BKM, and everolimus treatment of SK-N-AS (Suppl. Fig. 3). In contrast, AKT or GSK-3β phosphorylation levels did not respond to PI3K antagonists (BEZ, BKM, or OSI). A dose-dependent decrease of PRAS40 phosphorylation was observed only for treatment with BEZ at all time points and with BKM or OSI at 24 h. Phosphorylation levels in IMR32 cells were also largely nonresponsive to PI3K inhibitors (Suppl. Fig. 4). Everolimus suppressed p70S6K phosphorylation by nearly 60% and surprisingly had the same effect on GSK-3β phosphorylation. Given there is no known target or pathway to explain the latter result, more investigation into the mechanistic basis for this observation is needed.

Additive Effects Using Combinations of PI3K Pathway Inhibitors

The additive effects of dual combinations of the four drugs on NB5 cell growth were evaluated. Based on our IC₅₀ results shown in Supplementary Figure 1, we chose a concentration for each drug above its IC₅₀ value to treat NB5 cells with one drug alone or in combinations of two. Cell growth was measured 72 h after drug application by CellTiter-Glo assay. The results were expressed as a percentage of growth compared with a no-drug control and are shown in Table 2 . Treating NB5 cells with 300 nM BKM resulted in a 6.4% decrease in growth, and treatment with 300 nM OSI resulted in a 20.8% decrease in growth. Assuming Bliss independence ¹⁹ (i.e., the effect of OSI is independent of the presence of BKM), we would predict a decrease in growth by 1–0.936*0.792, or 25.9%, when combining the two drugs at the same concentrations. Yet we observed a decrease of 46.1%, nearly double what was predicted. We also observed similarly significant decreases in growth when combining BEZ with BKM, OSI, or everolimus. Only minor differences were observed when combining everolimus with BKM or OSI.

OPEN IN VIEWER

Table 2.

Synergistic effects of PI3K/mTOR pathway inhibitors.

% Control
Drug	300 nM BKM	300 nM OSI	10 nM BEZ	100 nM Everolimus
None	93.6 ± 5.9	79.2 ± 5.1	63.6 ± 3.8	38.8 ± 1.6
300 nM BKM		53.9 ± 6.0 (74.1)	29.7 ± 1.4 (59.5)	33.0 ± 1.3 (36.3)
300 nM OSI			26.4 ± 1.1 (50.4)	26.9 ± 1.5 (30.7)
10 nM BEZ				15.8 ± 0.52 (24.7)

PI3K pathway inhibitors were tested alone or in dual combinations at the concentrations shown in the table. Cells were incubated in the presence of these inhibitors for 72 h, followed by addition of CellTiter-Glo and measuring luminescence. Results were calculated from three independent replicates and tabulated as percentage of control grown with no inhibitor added. For all combinations, we calculated the values expected assuming Bliss independence. ¹⁹ These are shown in parentheses next to the value actually measured for that particular combination.

Comparative Phosphoproteomic Profiles of NB5 Cells Treated with PI3K Antagonist Combinations

NB5 cells were treated with double and triple combinations of all four drugs at the same concentrations used in Table 2 for 4 and 24 h. Quantitative information on NB5 PI3K phosphoprotein profiles in response to combinatorial PI3K blockade is shown in Figure 4 . At 4 h, AKT phosphorylation was the most sensitive to treatment with three of the four PI3K inhibitors, with no response to everolimus. This is in agreement with the results shown in Figure 3 . Further blockage of AKT phosphorylation was achieved by combining anti-PI3K drugs with the exception of everolimus. The most effective combination at reducing AKT phosphorylation was BKM and OSI (>95%). This confirmed the results we obtained measuring the cytotoxicity of these combinations ( Table 2 ). Phosphorylation of PRAS40 largely followed the pattern for AKT and decreased by up to 80%. GSK-3β phosphorylation was reduced by only up to 50%, even when using inhibitor combinations, suggesting other PI3K-independent pathways are involved in its regulation. After 24 h, AKT phosphorylation remained suppressed; however, the phosphorylation levels of the other three analytes had rebounded or even increased slightly. This was certainly the case for PRAS40 and to a lesser extent for p70S6K, whose phosphorylation levels increased twofold over the levels at the 4 h time point.

OPEN IN VIEWER

Figure 4.

Comparative phosphoproteomic profiles of NB5 cells treated with PI3K antagonist combinations. NB5 cells were incubated with NVP-BEZ235, NVP-BKM120, OSI-906, and everolimus alone or in double and triple combinations at concentrations below the IC₅₀ value for each drug. The following concentrations were used: 10 nM NVP-BEZ235 (Z), 300 nM NVP-BKM120 (K), 300 nM OSI-906 (O), and 10 nM everolimus (E). Extracts were prepared 4 or 24 h after compound addition to the growth medium. The abundance of total and phosphorylated protein for AKT (blue bars), PRAS40 (red bars), p70S6K (green bars), and GSK-3β (purple bars) in each sample was measured by xMAP assay, and ratios of phosphorylated protein to total protein were calculated and normalized to the solvent-only control for each analyte at both time points.