Transcription factor EB regulates phosphatidylinositol-3-phosphate levels that control lysosome positioning in the bladder cancer model

Cell culture and treatments

Bladder cancer cells lines RT4, MGHU3, RT112, KU19-19, JMSU1, T24 and TCCSup were grown in RPMI-1640 medium (Life Technologies, Carlsbad, CA, USA), supplemented with 10% Fetal Bovine Serum (FBS; Eurobio, Courtaboeuf, France). Normal human urothelium (NHU) cells were from Jennifer Southgate (University of York, UK). NHU were grown in KSFMC medium65. For experiments with inhibitors, as per the experiment either the day after cell seeding or after transfection respective drugs were added for incubation time of 24 h or as indicated and cells were incubated, at 37 °C. The concentration of inhibitors used were as follows: rapamycin (10 μM), wortmannin (1 μM, 2 h), ML-SI1 (20 μM, 3 h), BAPTA AM (10 µM, 3 h) and cycloheximide (20 µg/mL). For starvation experiments, the day after cell seeding, the medium was removed and cells were washed once with EBSS (Earle’s Balanced Salt Solution) and incubated in EBSS for 4 or 24 h, as per the experiment, before lysate preparation or cell fixation with 4% PFA.

Cell transfection

For RNA interference studies, 200,000 cells were transfected in 12 well plate with 25 pmol siRNA (siTFEB: ON-TARGETplus Human TFEB, L-009798-00-0005, DharmaconTM) using Lipofectamine RNAiMAX Transfection Reagent (1:200; Life Technologies). Cells were incubated 72 h at 37 °C prior to further manipulation or drug treatment. Efficiency of gene silencing was verified by western blot of cell lysate after three days of transfection.

For plasmid transfection, 200,000 cells were transfected in a 12 well plate. Transfection was performed using Lipofectamine LTX with Plus reagent (Invitrogen) using 1 μg of plasmid. pEGFP-N1-TFEB plasmid was a gift from Shawn Ferguson (Addgene plasmid # 38119; n2t.net/addgene:38119; RRID:Addgene_38119n36) or EGFP-2X FYVE plasmid (kind gift from B. Payrastre, Toulouse). 48 h post transfection, cells were trypsinized and transferred to sterilized coverslips (12 mm) in 1 mL medium in 12 well plate. Cells were fixed with 4%PFA 72 h after transfection and used for immunofluorescence and imaging.

Micro-array analysis

Micro array data were analyzed with R (3.5.2). The annotation was performed using affy package (1.58.0) with a custom CDF (Chip Description File) from brain array (huex10st, genome version 23). Normalization was done with RMA algorithm using affy library66 and batch effect corrected with ComBat67. The PCA was computed from these normalized and corrected data.

qRT-PCR

Total RNA was isolated from JMSU1 cells using Trizol reagent (Life technologies). cDNA was prepared using one microgram total RNA with high capacity cDNA reverse transcriptase kit and the manufacturer’s protocol (Life technologies). Quantitative PCR was performed using KAPA SYBR FAST (Sigma-Aldrich) and the following primers for the PIK3C3 gene: forward, 5ʹ-GACAGGCCGATGATGAGGATTT-3ʹ and reverse 5ʹ-AAGGGGGCTGGTTATAATTTGGG-3ʹ. To normalize the expression beta actin gene was used as normalizing control with the following primers: forward 5ʹ-CATGTACGTTGCTATCCAGGC-3ʹ and reverse 5ʹ-CTCCTTAATGTCACGCACGAT-3ʹ. Samples were run in triplicate and gene of interest expression was normalized to human beta actin.

Micropatterned coverslips preparation and cell seeding

Micropattern production was performed using photo-lithography methods16,17. Briefly, coverslips were coated with Poly-L-Lysine(20)-grafted[3.5]-Polyethyleneglycol(2) (PLL-g-PEG) from SuSoS (Dübendorf, Switzerland) at a final concentration of 0.1 mg/mL in 10 mM HEPES (pH 7,3) solution. Coverslips were exposed to deep UV during 5 min using a photomask containing arrays of crossbows (37 μm diameter, 7 μm thick). Prior to cell seeding, the patterned surface was incubated for 1 h with a mixture of 50 μg/mL fibronectin (Sigma-Aldrich, St. Louis, MO, USA), 5 μg/mL concanavalin A (Sigma-Aldrich, St. Louis, MO, USA) and 1 μg/mL fibrinogen–Cy5 (Invitrogen). Cells were seeded on micropatterns in RPMI medium supplemented with 20 mM HEPES (Life Technologies) for 4 h prior the experiment.

Invasion assay

Cells were trypsinized and 104 cells/ml were re-suspended in RPMI medium containing 10% FBS and 1% Penicillin-Streptomycin (Life Technologies). Then 100 µl of cell suspension was plated in 48-well plates coated with 1% agarose (Life Technologies) and incubated for 3 days. In each well, a spheroid was formed from 103 cells. Next, the spheroids were plated on Lab-Tek chambers (Sigma), in a mixture of collagen I from rat tail (Corning) at a final concentration of 2 mg/ml, PBS, sodium hydroxide (NaOH) and serum-free medium. The spheroids were monitored for 5 consecutive days by using an inverted Leica microscope (Wetzlar, Alemanha) equipped with camera device using 4x objective.

Immunofluorescence

Cells were fixed with 4% formaldehyde for 15 min at room temperature, washed three times with PBS and permeabilized in PBS containing 0.2% BSA and 0.05% saponin. Cells were then incubated with the primary antibodies in PBS containing 0.2% BSA and 0.05% saponin (anti-Lamp1/CD107a, 555798, BP PharmingenTM, 1:1000; anti-mTOR, 7C10, #2983, Cell Signaling Technology, 1:1000; anti-EEA1, 610456, BD Biosciences, 1:500; protrudin/ ZFYVE27, 12680-1-AP, Proteintech, 1:500) and Alexa Fluor 488, Alexa Fluor 647 or Cy3-coupled secondary antibodies (Jackson ImmunoResearch, 1:400 in PBS containing 0.2% BSA and 0.05% saponin) for 1 h. Actin was visualized by FluoProbes 547H (557/572 nm) coupled Phalloïdin (Interchim) and nuclei with 0.2 μg/mL 4’,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Coverslips were mounted in Mowiol (Sigma-Aldrich).

Western blot

In total, 250,000 cells were seeded in a 12 well plate one day prior to the experiment. Drug treatments or knock-down experiments were performed as mentioned before. Equal volumes of lysate from each cell line was loaded on a 10% or 12% polyacrylamide gel, resolved by SDS-PAGE and electrotransferred to nitrocellulose membranes. Membranes were incubated with primary antibodies at 4 °C overnight: Phospho P-70 (Thr389)-S6K (CST: 9205 S, 1:1000 in 5% BSA in TBST), P-70 S6K (CST: 9202 S, 1:1000 in 5% milk in TBST), Phospho (Ser65)-4EBP1 (CST: 9451, 1:1000 in 5% BSA in TBST), 4EBP1(CST: 9452, 1:1000 in 5% milk in TBST), GAPDH (Sigma: G9545, 1:10,000 in 5% milk in TBST), EEA1(610456, BD Biosciences, 1:500 in 5% milk in TBST), protrudin (ZFYVE27, Proteintech 12680-1-AP), TFEB (CST: 4240, 1:1000 in 5% milk in TBST), Lamin B1 (Abcam: ab16048, 1:500 in 5% milk in TBST) and species specific HRP secondary antibodies (1:10,000) for 1 hour at room temperature, following ECL western blotting substrate.

Image acquisition

Images for immunolabelled cells on micropatterns were acquired with an inverted wide field Deltavision Core Microscope (Applied Precision) equipped with highly sensitive cooled interlined charge-coupled device (CCD) camera (CoolSnap Hq2, Photometrics). Z-dimension series were acquired every 0.5 µm.

Images for non-pattered immuolabelled cells were acquired with a spinning disk confocal microscope (Inverted Eclipse Ti-E (Nikon) + spinning disk CSU-X1 (Yokogawa) integrated with Metamorph software by Gataca Systems). Cells were imaged as Ζ-stacks with 0.2 µm distance and 12 µm total height.

Image processing and analysis

For cells on micropatterns, several tens of single cell images were aligned using the coordinates of the micropattern17,68 determined via an ImageJ-based macro (Bethesda, MD, USA). To extract the 3D spatial coordinates of lysosomes, images were segmented with the multidimensional image analysis (MIA) interface on MetaMorph (Molecular Devices, Sunnyvale, CA, USA) based on wavelet decomposition. The coordinates of the segmented structures were processed for density estimation17 programmed in the ks library in R. For visualizing kernel density estimates, probability contours were visualized using the extension libraries mvtnorm, rgl, and miscd. The lysosome volume was estimated using the sum of all pixels in all 2D planes of one segmented lysosome (Z-planes were acquired every 0.5 µm). This value has been obtained from the MIA segmentation interface on MetaMorph (Molecular Devices, Sunnyvale, CA, USA) used in batch on all cells and represents arbitrary units (a.u.) of lysosome volume measurement.

Levels of lysosome dispersion in non patterned MGHU3, RT112, KU19-19 and JMSU1 cells were measured using statistical inertia (=averaged squared distance to the center of mass). To control for variations in cell size differences, normalization to cell size has been applied. Lysosome coordinates have been divided by the coordinates of the center of the mass (setting the center mass at x = 1, y = 1). This quantifies the dispersing of the lysosome structures independently of homogeneous dilations due to cell size. To test statistical significance, a Kruskal-Wallis test with Dunn post-hoc test with Sidak correction for multiple comparisons has been applied.

Image analysis for the Figs. (2B, 4H, J, 5B, D, and S4B, G, S5B, E, G, K, N) was done using CellProfiler (version: 3.1.9) on one Ζ-plane of the images. The pipelines for different analysis were prepared as follows:

To detect the total and membrane bound intensities of protein of interest (labeled as total integrated intensity or spots/total, respectively, in the figures) or intensities of co-localized proteins the pipeline was created as follows:

Step 1: Module ‘EnhanceorSuppressFeatures’ was applied to channels where the objects needs to be segmented, either to obtain their intensities or objects for the intensities of co-localized proteins, to get sharp and defined objects which makes segmentation easier (for eg. On channels with LAMP1 or EEA1 or GFP-FYVE).

Step 2: Nucleus was identified in the DAPI channel using the ‘IndentifyPrimaryObject’ module

Step 3: Module ‘IndentifyPrimaryObject’ was used again on the images obtained from Step 1 to segments objects whose measurements are required (such LAMP1, EEA1, EGFP-FYVE)

Step 4: Cells were segmented using the ‘IdentifySecondaryObject’ module with nucleus as the ‘primary object’ (identified in step 2) and using phalloidin or another cytoplasmic protein channel to recognize the cell boundaries.

Step 5: Module ‘RelateObjects’ was used to relate the objects obtained in step 3 to each cell obtained in Step 4. Output of this channel was saved as another object which gives the objects of protein of interest per cell.

Step 6: Objects from step 3 were masked on the channel whose co-location or membrane bound fraction had to be calculated using the ‘MaskImage’ module (for e.g.: to calculate EGFP-FYVE on lysosomes in Fig. 5B, Lysosomes were segmented in step 3 and the output objects were masked on EGFP-FYVE channel or to calculate membrane bound EGFP-FYVE, segmentation of EGFP-FYVE objects from step 3 was masked on EGFP-FYVE channel). Output of this step was saved as a new image in the pipeline.

Step 7: ‘MeasureObjectIntensity’ module was used to obtain total ‘per cell intensity’ and ‘intensity on spots’ of protein of interest. Intensities were picked from images from step 6 and raw images of channel of interest using cells from step 4 as the objects.

Step 8: Cell size was obtained using the module ‘MeasureObjectSizeandShape’ on the cells segments in Step 4 as the objects.

Step 9: Finally, all the measurements were exported to the excel sheet using the module ‘ExporttoSpreadsheet’.

Step 10: The final values were exported to a csv file named ‘cell’. This file had the values of cell size (in pixels), total intensity of protein of interest per cell, intensity of protein of interest on spots and intensity of co-localized protein on the object of interest (e.g.: GFP-FYVE on lysosomes).

Integrated intensities were used for the analysis and to plot the graphs. For measurements of EEA1 or EGFP-FYVE intensities on endosomes (Fig. 5D; S5D–G; S5J–N) the ‘intensity on spots’ was normalized to the total fluorescence (‘per cell intensity’) of the cell analyzed.

Cell fragmentation

1.5 million cells were seeded in a 10 cm dish 2-3 days before the experiment and allowed to grow. After 3 days, cells were washed once in ice cold PBS, pH 7.4 and scrapped using a plastic scrapper in 1.5 mL ice cold PBS. Cells were centrifuged for 10 sec in a tabletop pop-spin centrifuge and the supernatant were discarded. The pellet was resuspended in 1.2 mL of ice cold 0.1% NP40 (Sigma Aldrich) in PBS. 400 µL of mix was removed as the whole cell lysate, 100 µL of 5X Laemmli buffer was added, and the sample was boiled for 1 min. The rest of the sample was centrifuged as above. 400 µL of supernatant was collected as the cytosolic fraction, 100 µL of 5X Laemmli buffer was added, and the sample was boiled for 1 min. The rest of the supernatant was discarded. The pellet containing the nuclear fraction was resuspended in 400 µL of 1X Laemmli buffer, and the sample was boiled for 1 min. The protocol was adapted from Suzuki et al., 201069.

GSEA analysis

GSEA has been made using fgsea() function from fgsea package. High-grade condition agglomerates 2 TCCSup, 2 T24, 1 JMSU1 and 1 KU1919 transcriptomic data, and low-grade agglomerates 1 RT4, 3 MGHU3 and 2 RT112 transcriptomic data.

Statistics and reproducibility

The statistical analysis of endolysosome volume, number and normalized NND was performed with R (3.6.0). For NND analysis, the centroids distance between structures was calculated from a constant number of lysosomes that was randomly sampled from each cell. Therefore, variation in NNDs cannot be imputed to variation in the number of lysosomes but to bona-fide variation of their spatial organization. The statistical analysis was a Kruskal-Wallis test with Dunn test for multiple comparisons correction.

For all experiments, a large number of cells were monitored from 3 to 6 independent experiments. Two-sided Mann-Whitney U test were performed for 2 conditions comparisons. For multiple comparisons, a Kruskal-Wallis has been used with Dunn’s test for multiple comparisons. Additionally, to compare the global distribution of cell population, χ² tests were performed (R function “chi-square()”) and Benjamini-Hochberg multiple comparison correction has been applied. For the statistical analysis on the data from CellProfiler, GraphPad Prism 9.3.1 was used. Mann-Whitney U test or paired t test was applied for the two conditions comparison (indicated in respective figures) or Kruskall-Wallis test with Dunn test for multiple comparison.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Read more here: Source link