The EBC-1 cell line was purchased from JCRB and cultured in DMEM supplemented with 10% fetal bovine serum (BioWest, Nuaillé, France), 20 mM HEPES pH 7.3, 100 IU/ml penicillin and 100 $\mu $g/ml streptomycin. All culture media were purchased from Lonza (Breda, Netherlands) except when otherwise stated. Cells were maintained in 25-75 cm$^{2}$ flasks (Greiner Bio-One GmbH, Frickenhausen, Germany) and incubated at 37$^{\circ}$C in a humidified atmosphere of 5% CO$_{2}$.

Crizotinib was purchased from SelleckChem (Houston, USA), dissolved in DMSO, aliquoted out and stored at $-$80$^{\circ}$C until further use. Immediately before use, crizotinib was diluted in PBS.

Cell growth inhibition was determined using the sulforhodamine B assay (SRB) [16]. Briefly, cells were seeded in 96-well flat bottom plates (Greiner Bio-One GmbH, Frickenhausen, Germany) and were allowed to attach overnight. Next day, cells were treated with a range of crizotinib concentrations (0-5 $\mu $M). A dilution of 0.1% DMSO served as control. After 72 h, cells were fixed with 50% trichloroacetic acid, incubated at 4$^{\circ}$C for one hour and washed 5 times with water. Cells were stained with a 0.4% SRB solution in 1% acetic acid and incubated for 15 min at room temperature. To remove unbound stain, the plates were washed 5 times with 1% acetic acid and were allowed to dry. Cellular SRB stain was dissolved in 10 mM Tris-base and the optical density was determined at 540 nm with the HTX Synergy microplate reader (BioTek). The IC$_{50}$ concentration of crizotinib was determined with Graphpad Prism (v5).

Parental EBC-1 cells were maintained in a 25 cm$^{2}$ flasks and first treated with a crizotinib concentration representing the IC$_{50}$ value. The growth medium was refreshed once a week until cells started to grow at the parental growth rate. Drug concentrations were increased when cells continued to proliferate. When cells were growing normally at a concentration of 1 $\mu $M of crizotinib, drug treatment was terminated and cells were grown at least four weeks in crizotinib-free medium prior to SRB assays. Next, the crizotinib IC$_{50}$ values were determined while comparing the parental EBC-1 cell line with its EBC-CR daughter subline.

Analysis of tyrosine kinase inhibition was performed as previously described [17]. Cells were seeded in 25 cm$^{2}$ flasks and treated for 24 h with their respective IC$_{50}$ concentrations of crizotinib (25 nm for the EBC-1 and 7 $\mu $M for the EBC-CR) or 0.1% DMSO as a vehicle control. Lysis buffer was made by adding phosphatase and protease inhibitors (Thermo Scientific, Rockford, IL) to M-PER buffer (mammalian protein extraction reagent, pH 7.6) (Thermo Scientific, Rockford, IL). Cells were lysed on ice and harvested using a cell scraper. Lysates were spun at 10,000 $\times$ g for 15 min at 4${^\circ}$C, supernatant was collected and either immediately used or stored at $-$80$^{\circ}$C until analysis. 5 $\mu $g aliquot of total protein lysate was incubated with reaction buffer 1$\times$ABL buffer (Westburg), 100 $\mu $M ATP (Sigma-Aldrich), and fluorescein-labeled antibody PY20 (Exalpha, Maynard, MA). The arrays were blocked with 2% BSA before loading of the sample and next incubated at 30$^{\circ}$C for 60 cycles on a PamStation 12 (Pamgene International, `s Hertogenbosch, The Netherlands). A 12-bit CCD camera was used for fluorescence imaging, monitoring the fluorescence in real-time. Spot intensities and time-resolved curves were fit to calculate the amount of phosphorylation by Bionavigator software version 6.1 (PamGene International).

The intracellular concentration of crizotinib was determined as previously described [18]. In brief, cells were seeded and treated for 24 h with 5 $\mu $M crizotinib. Bafilomycin A1 was used to disrupt the acidic lysosomal pH and thus prevent lysosomal sequestration of crizotinib. A final concentration of 50 nm bafilomycin A1 was added to the cells 30 min prior to crizotinib treatment. Next, cells were sedimented by centrifugation and washed 3 times with PBS. Pellets were snap-frozen and stored at $-$80$^{\circ}$C until further analysis. After addition of 160 $\mu $l of cold phosphate buffer, cell pellets were homogenized and protein concentration was determined with the Bio-rad Bradford protein assay according to manufacturer's instructions. A sample of 100 $\mu $l was added to 400 $\mu $l acetonitrile, incubated for 20 min on ice, vortexed and centrifuged (21,000 $\times$ g, 10 min at 4$^{\circ}$C). Next, 100 $\mu $l were used for LC injection. Statistical analysis was performed with graphpad prism v5 using one-way ANOVA with Tukey post-hoc testing.

Cells were plated in a chambered cover glass (#1.5, Nunc Lab-Tek, Thermo Scientific, Rockford, IL) and treated for 24 h with 5 $\mu $M crizotinib or 0.1% DMSO as the vehicle control. The next day, the growth medium was replaced by an IMDM medium devoid of a pH indicator. In the first step, sunitinib (5 $\mu $M, LC laboratories, Woburn, MA) and Lysotracker red (0.5 $\mu $M, Thermo Scientific) were added to the wells. The cells were incubated in the dark for 30 min at 37$^{\circ}$C. In the second step, pHRodo green was added according to manufacturer's instructions, final concentration 1$\times$ (Life Technologies) and cells were further incubated in the dark at 37$^{\circ}$C for 30 min. A single drug stock solution was made for treatment of all wells. Cells were then washed 3 times with PBS and imaged using a Leica TCS SP8 STED 3$\times$ microscope. A 3D representation of the cells was obtained by imaging the cells on several planes, leading to a z-stack. Z-stacks were imaged at bright field, 445 nm (laser power 2) for sunitinib, 488 nm (laser power 0.7) for pHRodo green and 561 nm (laser power 1.5) for Lysotracker Red.

Images were analyzed with FIJI-software[19,20,21]. Z-stacks per sample were imported with Bio-Formats. For each z-plane the following analysis was performed on 10 representative cells per sample. In the first step, lysosomes were identified using the Lysotracker Red channel (Fig.1A). The image of the red channel was copied and made binary. With the `analyze particle' tool of FIJI, the lysosomes were selected using the automatic threshold `triangle'. The regions of the lysosomes were saved. Next, the image of the pHRodo green channel was copied and the regions of the 10 cells were selected (Fig.1B). Firstly, the regions of the lysosomes were overlaid (Fig.1C) with this image and the intensity of these regions was erased (Fig.1D). Then, the total intensity of the remaining regions of the cells was analyzed and saved as "intensity of cytoplasm". Secondly, the regions of the lysosomes were selected and the intensity of the outside region was erased (Fig.1E). The intensity of the lysosomes was saved as "intensity of lysosomes". This analysis was performed for all planes of the z-stack. We summed up these intensities and corrected for the area of the cells or lysosomes respectively, and the amount of the z-planes.

Fig. 1.

FIJI analysis: partial images of the red and green channels are shown. Yellow markings represent regions of interest (ROIs). (A) Lysotracker Red with markings for 6 selected cells for analysis; (B) pHRodo green with 6 selected cells; (C) pHRodo Green with 6 selected cells, containing markings for lysosomes; (D) pHRodo Green with 6 selected cells with deleted intensities for lysosomes; (E) pHRodo Green with 6 selected cells representing only the lysosomes.

Selection of EBC-1 NSCLC cells in gradually increasing concentrations of crizotinib led to the establishment of the EBC-CR daughter cell line (Fig.2). This drug resistant subline was cultured for at least 4 weeks without drugs prior to drug sensitivity assessment. The sensitivity of both the parental and drug-resistant tumor cell lines to crizotinib was determined using the SRB-assay (Fig. 2). The crizotinib IC$_{50}$ of the parental cell line was 28.5 $ \pm $ 1.1 nM, whereas the IC$_{50}$ value of the EBC-CR subline markedly increased to 7.4 $ \pm $ 0.3 $\mu $M, thus representing a 260-fold crizotinib resistance.

Fig. 2.

Growth inhibition by crizotinib: cells were treated with a crizotinib concentration range of 0-10 $\mu $M for 72 h. Values represent mean $\pm $ SEM of triplicates of 3 separate experiments. Horizontal dashed line represents a growth inhibition of 50%. EBC-CR: crizotinib resistant line; par: parental EBC-1 cell line.

To determine the effect of crizotinib on downstream signaling, the kinase activity was determined in both cell lines after exposure of intact cells to crizotinib (Fig. 3) using an equally cytotoxic concentration of crizotinib. Basal tyrosine kinase activity of the EBC-CR versus parental EBC-1 cells showed an overall higher kinase activity as compared to the basal activity (black) in the parental cell line, which was set at 100% for each kinase. This increased activity was approximately 2-fold higher for the following kinases: B3AT (Band-3 cotransporter, an anion transporter), c-Cbl (Casitas B-lineage lymphoma), CRK, DYR1A (Dual specific tyrosine phosphorylation-regulation kinase), Fes, LCK and p85A. In the parental cell line its kinase activity was not inhibited in response to crizotinib treatment, whereas in the EBC-CR resistant cell line, this kinase activity dropped below the background level [22]. C-Cbl is the E3-ligase that is responsible for marking several tyrosine kinase receptors, including cMET and EGFR, for breakdown [23]. This kinase was inhibited after crizotinib treatment in both the parental and resistant cell lines. CRK [24], Fes [25] and LCK [26] are downstream signaling proteins that were all inhibited after crizotinib treatment in both cell lines. The PI3K subunit p85A was also inhibited by crizotinib treatment. After crizotinib treatment, a 2-fold increase in EphA1 (Ephrin receptor type A) activity was observed in EBC-CR cells. Activation of this receptor leads to increased adhesion to the extracellular matrix [27]. cMET activity was inhibited in both cell lines after crizotinib treatment. This inhibition was stronger in the EBC-CR as compared to the parental cell line.

Fig. 3.

Tyrosine kinase activity. Cells were treated for 24 h with their respective IC$_{50}$ values of crizotinib (25 nM for EBC-parental (orange) and 7 $\mu $M for the EBC-CR (blue)) or 0.1% DMSO as control. Cells were lysed and tyrosine kinase activity was determined with the pamgene kinase array. (A) Kinase activity of EBC-CR vs parental is represented in this graph. (B) Inhibition of kinase activity after crizotinib treatment compared to the respective basal activity of both cell lines.

In the parental cell line, 24 h treatment with 5 $\mu $M crizotinib resulted in a mean intracellular crizotinib concentration of 52.5 pmol/$\mu $g protein (Fig. 4). Whereas, in the resistant cells, this intracellular concentration dropped by 55% with an intracellular level of 23.5 pmol/$\mu $g protein ($p <$ 0.001). The addition of the V-ATPase inhibitor bafilomycin A1 (which alkalinizes the native acidic pH of lysosomes) prior to crizotinib treatment led to a $\sim $40% decrease in the intracellular concentrations of crizotinib in both the parental cells and their drug resistant subline with values of 34.1 and 14.7 pmol/$\mu $g protein, respectively.

Fig. 4.

Intracellular crizotinib concentrations: cells were treated for 24 h with 5 $\mu $M crizotinib. Pellets were snap-frozen and analysis was performed by liquid chromatography coupled to a tandem mass-spectrometer. EBC-par: parental EBC-1 cell line; EBC-CR: crizotinib resistant EBC-1 daughter cell line. Baf: 50 nM Bafilomycin, added 30 min before crizotinib treatment. *: $p <$ 0.05; **: $p <$ 0.01; ***: $p <$ 0.001.

The pH sensitive pHRodo Green probe was used to assess the intracellular pH in the different cellular compartments. Bright staining correlates with acidic pH whereas alkalinization results in markedly decreased fluorescence. Lysotracker Red specifically stains the lysosomes and sunitinib was used to visualize lysosomal sequestration of weak-base small molecule protein tyrosine kinase inhibitors. The first panel of Fig. 5 shows representative images from the different channels and overlays of the parental EBC-1 cell line treated with DMSO. The images of Lysotracker red (Fig.5B) and sunitinib (Fig.5C) are shown to overlap in the first panel of Fig. 5G. The bright field image also shows that these lysosomes are located in the peri-nuclear zone (Fig.5G). An overlay of Lysotracker red and pHRodo Green co-localizes at the lysosomes in the brightest spots of pHRodo Green (Fig. 5D, 5H). The second panel shows the parental EBC-1 cell line after treatment with 5 $\mu $M crizotinib. The number of fluorescent events in the lysotracker red and sunitinib channel were increased in comparison with the control. The EBC-CR crizotinib resistant cell line (third and fourth panels) showed the same effect after crizotinib addition. Quantification of the average fluorescence intensity of 10 cells throughout the stacks is shown in Fig. 6. The intensity of pHRodo Green was quantified in the cytoplasm and in the lysosomes in both cell lines. The parental EBC-1 cell line showed a basal intensity of 1637 RFU (relative fluorescence units) in the cytoplasm and 3862 RFU in the lysosomes. After treatment with crizotinib, the intensity of pHRodo Green doubled in the cytoplasm and in the lysosomes. pHRodo Green fluorescence intensity in the cytoplasm and in the lysosomes is similar to the values in the parental cell line after crizotinib treatment. Treatment with crizotinib did not increase the fluorescence intensity in the cytoplasm but did significantly decrease the fluorescence intensity of the lysosomes. In summary, crizotinib increased the fluorescence intensity in both the cytoplasm and lysosomes in parental cells, whereas in the resistant cell subline the fluorescence intensity of the lysosomes was significantly decreased. This points towards a general acidification in the parental cell line, whereas in the resistant subline the lysosomes became more alkaline when pulse treated with crizotinib.

Fig. 5.

Intracellular effect of crizotinib: cells were seeded in a chambered cover glass and treated with drugs for 24 h. Cells were stained with sunitinib and Lysotracker red for 1h and with pHRodo Green for 30 min. Cells were imaged in depth by dividing them in different layers (z-stack). Z-stacks were imaged using a Leica TCS SP8 STED 3$\times$ microscope. (A) pHRodo Green; (B) Lysotracker red; (C) Sunitinib; (D) pHRodo Green $+$ lysotracker red overlay; (E) Bright field; (F) Bright field $+$ lysotracker red overlay; (G) Bright field, lysotracker red and sunitinib overlay; (H) Bright field, lysotracker red and pHRodo Green overlay. Representative images of one z-plane.

Fig. 6.

Subcellular pHRodo Green fluorescence intensity: cells were treated with crizotinib for 24 h and stained with pHRodo Green, lysotracker red and sunitinib as described in Fig. 5. Analysis with FIJI was performed, selecting the lysosomes based on lysotracker red staining and determining the intensity of pHRodo Green in the cytoplasm and the lysosomes respectively for all z-planes of the stack, while taking into account the cell area and number of stacks for each sample. EBC par: parental cell line; EBC CR: crizotinib resistant daughter cell line; ctrl: DMSO treated cells; criz: treated with 5 $\mu $M crizotinib. *: $p<$ 0.05; **: $p<$ 0.01; ***: $p<$ 0.001.