Shear Stress Increases V–H${}^{+}$-ATPase and Acidic Vesicle Number Density, and p-mTORC2 Activation in Prostate Cancer Cells

Abstract

Introduction

Cells in the tumor microenvironment experience mechanical stresses, such as compression generated by uncontrolled cell growth within a tissue, increased substrate stiffness due to tumor cell extracellular matrix (ECM) remodeling, and leaky angiogenic vessels which involve low fluid shear stress. With our hypothesis that shear stress increases V–H${}^{+}$-ATPase number density in prostate cancer cells via activation of the mTORC1 and mTORC2 pathways, we demonstrated and quantified such a mechanism in prostate cancer cells.

Methods

Moderately metastatic DU145 and highly metastatic PC3 prostate cancer cells were subjected to 0.05 dynes ${\text{cm}}^{-2}$ wall shear stress for 24 h, followed by immunocytochemistry and fluorescence measurements of $\beta$1 integrin, endosome, lysosome, V–H${}^{+}$-ATPase proton pump, mTORC1, and p-mTORC2 antibodies. Post shear stress migration assays, and the effects of vacuolar proton pump inhibitor Bafilomycin A1 (60 nM, 24 h) as well as shear stress on the ICC fluorescence intensity of the proteins of interest were conducted with DU145 cells.

Results

Low fluid shear stress increases the fluorescence intensity of $\beta$1 integrin, endosome, lysosome, V–H${}^{+}$-ATPase, mTORC1, and p-mTORC2 antibodies in PC3 and DU145 cells, and also increased cell migration. However, Bafilomycin A1 decreased fluorescence intensity of all of these proteins in DU145 cells exposed to shear stress, revealing that V–H${}^{+}$-ATPase controls the expression of these proteins.

Conclusions

Prostate cancer cell mechanotransduction increases endosomes, lysosomes, and proton pumps—where increases have been associated with enhanced cancer aggressiveness. We also show that the prostate cancer cell’s response to force promotes the cancer drivers mTORC1 and mTORC2.

Introduction

Cells in the tumor microenvironment are subject to several mechanical stresses, including compressive stresses generated by uncontrolled cell growth in a confined space, increased tissue stiffness due to tumor cell ECM remodeling, and increased interstitial fluid pressure due to the angiogenic growth of new leaky blood vessels. Increased fluid pressure triggers growth of lymphangiogenic vessels for drainage, subjecting tumor cells to shear stresses31 on the order of 0.01–0.1 dynes ${\text{cm}}^{-2}$, two to three orders of magnitude smaller than blood vessel wall shear stresses.18,31 Some studies have shown that low shear stress renders prostate and ovarian cancer cells more invasive.18,28 Here we show how fluid shear stress leads to biochemical changes in prostate cancer cells. This can help develop novel treatment strategies.

Endosomes and lysosomes, inherent transport and degradation vehicles in epithelial cells, recycle and dispose of cellular waste consisting of macromolecules, proteins, and organelles, and digest extracellular materials delivered to cells by endocytosis, phagocytosis, and autophagy.4 Cancer cells use endosomes and lysosomes, and other membrane bound vesicles, to sustain cell growth and enable migration via integrin trafficking.5,6 Lysosomes are overexpressed near the plasma-membrane by 2–3 times in cancer cells in comparison with normal cells.5 Vesicles are transported to and from the plasma membrane within the cell by active transporters such as microtubule motor kinesins and myosin,33 as well as by random dispersion.12 Several types of myosin are overexpressed in breast1,38 and prostate cancer,24 among others, and kinesins are overexpressed in many cancers.27 This increases the number of lysosome plasma membrane fusion events which can deliver V–H${}^{+}$-ATPase proton pumps to the plasma membrane.4,20

V–H${}^{+}$-ATPases acidify endosomes and lysosomes, from pH $\approx$ 6 in early endosomes to pH $\approx$ 4.5–5 in mature lysosomes.20 When located at the plasma membrane, as typically found in metastatic cancer cells and specialized acid-secreting cells, proton pumps acidify the extracellular environment and activate proteases that degrade extracellular matrices (ECM)—facilitating metastasis.3,8,21,22 V–H${}^{+}$-ATPase controls mTORC1 (mechanistic target of rapamycin complex 1) in the presence of amino acids, by stimulating Rag (recombination activating gene) to recruit mTORC1 to the lysosome surface, thence activated by Rheb (Ras homolog enriched in brain). The mTORC1 pathway regulates cell growth, proliferation, and metabolism,3,8,22 and is frequently overexpressed in cancer.10 At the cell plasma membrane, V–H${}^{+}$-ATPase activates Rac1 in breast and pancreatic cancer cells. Rac1 is a Rho-GTPase enzyme that regulates the actin cytoskeleton and cell morphology,36 and Rac1 can coordinate mTORC1 and mTORC2 (mechanistic target of rapamycin complex 2) in osteoclasts.7 mTORC2 is an upstream regulator of mTORC1 via the PI3K–Akt pathway, glucose and lipid metabolism, and cytoskeletal changes.17 Despite its prevalence,10 very little is known about how activation of mTORC2 occurs in cancer. We suggest that V–H${}^{+}$-ATPase activates both mTORC2 and mTORC1 in cancer via the known lysosomal membrane mechanism for mTORC1 described above, and via V–H${}^{+}$-ATPase’s interaction with Rac1.

We hypothesize that lysosome transport to the plasma membrane increases in response to shear stress transmitted by integrins to cytoskeletal actins,9,13 which remodel in response to the applied force.13 Actin remodeling can increase microtubule motor activity,37 transporting more acidic vesicles to fuse with the plasma membrane. Hence, as explained above, shear stress can increase V–H${}^{+}$-ATPase at the plasma membrane, and activate mTORC1 and mTORC211—all consistent with a more metastatic phenotype. Using indirect immunocytochemistry (ICC) we show that a low shear stress—0.05 dynes ${\text{cm}}^{-2}$—increases: (i) endosome and lysosome, (ii) V–H${}^{+}$-ATPase, (iii) $\beta$1 integrin, and (iv) mTORC1 and mTORC2 antibodies’ fluorescence in moderately metastatic DU145 and highly metastatic PC3 prostate cancer cells. These features indicate increased metastatic potential. We also show that shear stress increases cell migration speed, where more metastatic cells are expected to migrate faster.30 These new findings promise better understanding of a cancer-enhancing mechanotransduction process in the tumor environment, and aid in identifying treatment strategies from early to late stage cancer progression.

Materials and Methods

Cell Culture

DU145 and PC3 prostate cancer cells were kindly provided by Dr. Stephanie Filleur (Texas Tech University Health Sciences Center, Lubbock, TX), and were cultured at 37${}^{\circ }\mathrm{C}$ and 5% CO${}_2$. The cells were cultured in RPMI 1640 medium with L-glutamine (Gibco 10-040-CV, Thermo Fisher Scientific, Massachusetts, USA), supplemented with 10% fetal bovine serum (FBS) (Gibco 10437-028) and 1% penicillin-streptomycin (Gibco, 15140-122). For every passage and for experiments, cells were harvested by incubating with trypsin (Gibco by Life Technologies, 25302-062—5 mL per flask) at 37${}^{\circ }\mathrm{C}$ for 3 min. Trypsin was inhibited by adding 500 $\mu$L FBS. Cells were centrifuged for 5 min at 1000 revolutions per minute (Eppendorf Centrifuge 5804R, Hauppauge, New York, USA), the supernatant was removed, cells were suspended in 2–4 mL of fresh medium, and counted with trypan blue (Corning, 25900CI) using an automated cell counter (TC20, Bio-Rad, Berkeley, CA, USA).

Cell Shear Stress Experiments

Tygon tubings with a 500 $\mu$m ± 5 $\mu$m inner diameter, 40 cm long, were incubated with collagen (Corning, 354249) according to the manufacturer’s instructions to promote cell adhesion to the tubing walls. Variations in the tubing diameter were quantified by carefully cutting the tubing perpendicular to its length, photographing the tubing on a Nikon Eclipse microscope (Eclipse Ti, Nikon) at × 10 magnification, and performing circular fits to the inner diameter. After incubation tubings were flushed with phosphate buffered saline (PBS), cells were pipetted into the tubings, and the tubings were submerged in culture media. When confluent, shear stress was applied with cell culture media in 60 mL syringes (BD, 309628) using Chemyx Fusion syringe pumps at a flow rate Q of 3.54 mL/h for 24 h. Wall shear stress $\tau$ was calculated as

$$\begin{array}{c}\hfill \tau =4\eta Q{\pi }^{-1}{r}^{-3}\end{array}$$

where $\eta =$ 1 mPas is the media viscosity and r is the inner radius of the tubing. Control cells were cultured in tubing immersed in media. Cells were trypsinized by applying 200 $\mu$L of trypsin to the tubing with a pipette, incubating for 3 min as described above, aspirating the trypsinized cells and placing in a sterile microcentrifige tube, and inhibited with 20 $\mu$L of FBS. Cells were pipetted onto sterile cover glasses for immunocytochemistry, or 96 well plates for migration assays, and cultured with media for 24 h of culture prior to further experiments.

Many more cells were needed for western blots; therefore, cells were cultured on 1.56 mm inner diameter and 50 cm long tygon tubings (Saint-Gobain, ADF00003) treated with collagen as described above. When confluent, shear stress was applied with cell culture media in 140 mL syringes (Covidien Monoject, 140cc) at a flow rate of 6.71 mL/h for 24 h. Cells were trypsinized by applying 1 mL of trypsin to the tubing with a pipette, incubating for 3 min as described above, aspirating the trypsinized cells and placing in a sterile microcentrifige tube, and inhibited with 100 $\mu$L of FBS. The samples were then handled as described below in Cell fractionation and western blots.

Cell Migration Assay

Cells exposed to shear stress and control cells seeded on 96 well plates were treated with NucBlue (R37605, Thermo Fisher Scientific) fluorescent nuclear stain according to the manufacturer’s protocol, and imaged with a Nikon Eclipse microscope (Eclipse Ti, Nikon) at × 30 magnification using a DAPI filter set. Images were recorded with a Zyla 4.2 sCMOS (scientific Complementary-metal-oxide-semiconductor) camera (Andor) every 5 min for 24 h. Cells could not be tracked beyond 12 h due to photobleaching. The experiments were performed at 37${}^{\circ }\mathrm{C}$ with 5% CO${}_2$. Individual isolated cells were tracked by cropping the cells from a 12-h image stack in ImageJ, applying a threshold, and using ImageJ’s automated particle identification package. Centroids were measured and used to generate a trajectory for each cell. Cells that came close to another cell, or divided, were dropped from analysis at the time of the occurrence. Cell trajectories and displacements were plotted and calculated with Matlab. Inter-frame cell displacements were used to calculate migration speeds.

Bafilomycin A1 Treatment

Moderately metastatic DU145 prostate cancer cells were cultured on collagen coated 500 $\mu$m diameter tubes until confluent. Bafilomycin A1 (LC Laboratories, B-1080) treated cells were exposed to 60 nM for 24 h delivered by syringe pumps as described above. Control (CTRL) cells were exposed to shear stress with media containing the equivalent concentration of DMSO as Bafilomycin treated cells for 24 h—85 $\mu$M.

Immunocytochemistry

Cells were plated on coverslips until sub-confluent, fixed with 4% paraformaldehyde, quenched with Phosphate Buffer Saline (PBS) and 50 mM NH${}_4$Cl (Sigma-Aldrich, 213330) for 10 min, permeabilized with 0.1% PBS-Triton for 10 min, and blocked with 1% PBS-BSA (Sigma-Aldrich, A2153) for 15 min. Samples were incubated with primary antibodies against V–H${}^{+}$-ATPase subunit c (Developmental Studies Hybridoma Bank or DSHB, ATP6V0c, 1:10 dilution), EEA1 (DSHB, CRP-EEA1-1F8, 1:50), LAMP2 (DSHB, H4B4, 1:50), RAPTOR (Santa Cruz Biotechnology or SCB, sc-81537, 1:50), and p-mTORC2 phosphorylated Ser 2448 primary antibody (SCB, sc-293133, 1:50) for 1 h at room temperature. The coverslips were then washed 7 times and incubated with secondary antibodies (anti-mouse R37117, or anti-rat A-11007, ThermoFisher Scientific) conjugated with Alexa Fluor 594, and Alexa Fluor phalloidin 488 (dilution 1:100 in PBS-BSA, Molecular probes by Life Technologies, A12379) that specifically stains F-actin, for 1 h. After staining the coverslips were washed 7 times with PBS and mounted on glass slides using Prolong Gold antifade reagent with DAPI nuclear stain (Molecular probes by Life Technologies, P36935). Images were obtained at × 90 magnification using a Nikon Eclipse microscope. The corrected cell fluorescence density, taking into account background fluorescence, was calculated as: $\text{CCFD}=\text{CID}-\text{MBI}\times \text{Area}$, where CCFD is the corrected cell fluorescence density, CID is the integrated cell fluorescence density (the product of average cell grayscale and cell area), MBI is the average background intensity (averaged over 8 regions near the cell), and Area is the cell area.2CCFD was calculated from raw fluorescence images. Images were analyzed using ImageJ.

Cell Fractionation and Western Blots

PC3 cells were trypsinized (see above), washed twice with cold PBS and resuspended in homogenizer buffer—1.3 M sucrose (Sigma-Aldrich S7903i), 1 mM MgCl2 (Sigma-Aldrich M4880), 10 mM potassium phosphate buffer pH 6.8 (Sigma-Aldrich P5379) supplemented with a protease and phosphatase inhibitor cocktail (Thermo-Fisher Scientific 78444) following manufacturer recommendation. Cells were homogenized in a teflon/glass dounce homogenizer 20 times. All samples were stored at $-80{}^{\circ }\mathrm{C}$. Prior to the Western blots, protein concentration was determined with a spectrophotometer/fluorometer (DeNovix DS-11 FX+).

Proteins—30–50 $\mu$g diluted 1:1 in laemmli sample buffer (Bio-Rad 161-0737) containing 2-mercaptoethanol (Bio-Rad161-0710)—were separated by SDS-PAGE in 4-20%gradient acrylamide gels (Bio-Rad 456-8093), and transferred to nitro-cellulose membranes (Bio-Rad 170-4158) using the Trans-Blot system (Bio-Rad). The membranes were incubated for 1 h in a 2% BSA, Phosphate Buffer Saline (PBS), Tween solution on a plate shaker (blocked) or a 2% BSA, Tris Buffer Saline (TBS), Tween solution for p-mTORC2. The blocking solution was made from 2 g of BSA (Sigma Aldrich), dissolved in 100 mL of PBS or TBS. At the time of the experiment Tween was added to PBS or TBS (Bio-Rad, 170-6531, 1:1000 dilution). Membranes were washed 3 times in the appropriate saline-Tween solution for 5 min. Then membranes were incubated with primary antibodies against p-mTORC2 (SCB, sc-293133, 1:250 dilution), Raptor (Santa Cruz Biotechnology or SCB, sc-81537, 1:250), V–H${}^{+}$-ATPase subunit c (Developmental Studies Hybridoma Bank or DSHB, ATP6V0c, 1:250 dilution), and $\beta$-actin (was used as a loading control, Sigma-Aldrich A3854, 1:25,000 dilution) for 2 h at 37${}^{\circ }\mathrm{C}$. Finally, the membranes were incubated with horseradish peroxidase-conjugated anti-mouse IgG (Jackson ImmunoResearch 115035068) secondary antibody. The dilution used was 1:2000. The blots were revealed using ECL solution (Bio-Rad 170-5060) and were visualized using the Image Analyzer system (Bio-Rad ChemiDoc MP). Expression levels were quantified using ImageJ software and normalized by $\beta$-actin loading control levels. Statistical significance was determined using Matlab’s two-sample Kolmogorov–Smirnov test.

Results

Prostate Cancer Cell Speed/Aggressiveness is Enhanced by Low Shear Stress

Recall that moderately metastatic DU145 and highly metastatic PC3 prostate cancer cells were cultured on collagen coated 500 $\mu$m diameter tubes (40 cm long) until confluent and subjected to 0.05 dynes ${\text{cm}}^{-2}$ of wall shear stress—on their exposed cell surfaces when the cells are attached to the sidewall—for 24 h using syringe pumps. The shear stress was calculated from the set flow rate through the tube (see Materials and Methods). Control (CTRL) cells were cultured in identical tubes for the same amount of time. After 24 h cells were trypsinized and cultured in 96 well plates for 24 h, treated with NucBlue fluorescent DNA (deoxyribonucleic acid) stain to enable automated cell tracking, and imaged every 5 min for 24 h with a fluorescent microscope at 37${}^{\circ }\mathrm{C}$ and 5% CO${}_2$. Individual nuclei of isolated cells were tracked for 12 h (this duration due to photobleaching of the nuclear stain). Representative images for DU145 cells are shown in Fig. 1a, with the DU145 trajectories shown in Fig. 1b (CTRL) and Fig. 1c (SS), and PC3 trajectories given in Fig. 1d (CTRL) and Fig. 1e (SS). The average DU145 SS cell migration speed increases by 143% of the CTRL cells, and the average PC3 SS migration speed increases by 11% of the CTRL cells (Fig. 1f); therefore, low wall shear stress significantly increases prostate cancer cell aggressiveness, confirming a prior study.18

Figure 1.

Low-level shear stress increases prostate cancer cell migration. (a) Left panel: representative images of control (CTRL) and cells subjected to 0.05 dynes/${\text{cm}}^2$ of wall shear stress for 24 h (SS), imaged at 0 and 12 h. Cells are stained with NucBlue. Grayscale values in the range of [8, 51] are linearly rescaled to [0, 255]. Right panel: raw images from the right panel were inverted and grayscale values in the range of [183, 246] are rescaled to [0, 255]. Scale bars are 20 $\mu$m. (b) Trajectories of six representative DU145 CTRL cells over 12 h. The origin is the starting position of each cell. (c) Trajectories of six representative DU145 SS cells over 12 h. Note that distances travelled are much larger in (c). (d) Trajectories of six representative PC3 CTRL cells over 12 h. (e) Trajectories of six representative PC3 SS cells over 12 h. (f) Average ± SE CTRL and SS cell migration speeds from 21 DU145 CTRL, 22 DU145 SS, 32 PC3 CTRL, and 34 PC3 SS cells, obtained at 5 min intervals, from 4 independent experiments yielding 2963 velocity samples for DU145 CTRL, 3075 for DU145 SS, 1896 for PC3 CTRL, and 1633 for PC3 SS. ***$p\ll 1\times {10}^{-6}$, **$p\ll 1\times {10}^{-3}$, the statistical comparison used the two-sample Kolmogorov–Smirnov test.

Low Wall Shear Stress Increases the Fluorescence of Endosome, Lysosome, $\beta$1 Integrin, F-Actin, V–H${}^{+}$-ATPase, and mTOR Markers

DU145 and PC3 prostate cancer cells were exposed to 0.05 dynes ${\text{cm}}^{-2}$ of wall shear stress for 24 h (SS), and control (CTRL) cells were cultured under static conditions as described above. Indirect immunocytochemistry (ICC) was performed using primary and secondary (fluorescent) antibodies, and cell fluorescence density was measured. This is calculated as cell area$\times$(mean cell greyscale − mean background), where the background is averaged over 8 locations; see Materials and Methods. Using phalloidin, a specific stain for F-actin, we find that polymerized F-actin (filamentous actin) fluorescence density significantly increases in SS cells compared to CTRL (by $\sim$ 4.7 times for DU145 and by $\sim$ 2.1 times for PC3; Figs. 2a and 2c). We also find that $\beta$1 integrin fluorescence density increases with shear stress in comparison with control, suggesting increased expression (by $\sim$ 2.7 times for DU145 and $\sim$ 1.4 times for PC3; Figs. 2b and 2c). This is consistent with the widely-held view that integrins and actin in cells respond to mechanical forces.9,13 Indirect ICC also shows that the fluorescence density of endosome and lysosome markers, EEA1 and LAMP2, increase fluorescence density with shear stress (EEA1 by $\sim$ 2.5 times for DU145 and by $\sim$ 1.8 times for PC3, LAMP2 by $\sim$ 1.8 times for DU145 and by $\sim$ 1.9 times for PC3; Fig. 3). V–H${}^{+}$-ATPase also increases (by $\sim$ 3 times for DU145 and by $\sim$ 2.2 times for PC3). Raptor, that binds with mTOR protein to form mTORC1, increases (by $\sim$ 2.9 times for DU145 and $\sim$ 1.2 times for PC3), as also p-mTORC2 increases (by $\sim$ 1.7 times for DU145 and $\sim$ 1.4 times for PC3; Fig. 4). These suggest that V-type proton pumps activate both mTOR complexes, and that proton pumps, acidic vesicles, and mTOR are activated by mechanotransduction.

Figure 2.

Low-level shear stress increases F-actin stain and $\beta$1 integrin antibody fluorescence in moderately metastatic DU145 and highly metastatic PC3 prostate cancer cells. (a) to (b): upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are 20 $\mu$m. (a) Control cells (CTRL), and cells subjected to 0.05 dynes ${\text{cm}}^{-2}$ of wall shear stress for 24 h (SS), stained with phalloidin. Inverted image grayscale values in the range of [100, 255] were linearly rescaled to [0 255]. (b) $\beta$1 integrin antibody fluorescence for CTRL and SS cells. Grayscale values in the range of [75, 255] in inverted images were linearly rescaled to [0 255]. (c) Average cell fluorescence density ± SE. F-actin cases: DU145 CTRL n = 383, DU145 SS n = 216, PC3 CTRL n = 242, PC3 SS n = 122. $\beta$1 integrin cases: DU145 CTRL n = 340, DU145 SS n = 229, PC3 CTRL n = 283, PC3 SS n = 149. ***$p\ll 1\times {10}^{-6}$ for CTRL-SS comparisons for F-actin and $\beta$1 integrin. Statistical comparisons used the two-sample Kolmogorov–Smirnov test. Data was averaged from three independent experiments.

Figure 3.

Shear stress increases endosome and lysosome antibody fluorescence in DU145 and PC3 cells. (a) to (b): Upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are $20\mu \text{m}$. (a) Early endosome EEA1 antibody fluorescence for control cells (CTRL), and cells subjected to 0.05 dynes ${\text{cm}}^{-2}$ of wall shear stress for 24 h (SS). Grayscale values in the range of [110, 255] in inverted images were rescaled to [0 255]. (b) Lysosome LAMP2 antibody fluorescence for CTRL, and SS cells. Inverted image grayscale values in the range of [110, 255] were linearly rescaled to [0 255]. (c) Average cell fluorescence density ± SE. EEA1 cases: DU145 CTRL n = 364, DU145 SS n = 251, PC3 CTRL n = 220, PC3 SS n = 190. LAMP2 cases: DU145 CTRL n = 381, DU145 SS n = 163, PC3 CTRL n = 162, PC3 SS n = 168. ***$p\ll 1\times {10}^{-6}$, statistical comparisons used the two-sample Kolmogorov–Smirnov test. Data was obtained from three independent experiments.

Figure 4.

Shear stress increases vacuolar proton pump, Raptor, and p-mTORC2 fluorescence in DU145 and PC3 cells. (a) to (c): upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are 20 $\mu$m. (a) V–H${}^{+}$-ATPase (V-ATPase) antibody fluorescence for control (CTRL), and shear stress exposed cells (SS). Grayscale values in the range of [110, 255] in inverted images were rescaled to [0 255]. (b) Raptor antibody fluorescence for CTRL, and SS cells. Inverted image grayscale values in the range of [120, 255] were linearly rescaled to [0 255]. (c) p-MTORC2 antibody fluorescence for CTRL and SS cells. Grayscale values in the range of [150, 255] were linearly rescaled to [0 255] for inverted images. (d) Average cell fluorescence density ± SE. V–H${}^{+}$-ATPase cases: DU145 CTRL n = 250, DU145 SS n = 231, PC3 CTRL n = 155, PC3 SS n = 128. Raptor cases: DU145 CTRL n = 295, DU145 SS n = 163, PC3 CTRL n = 253, PC3 SS n = 168. p-mTORC2 cases: DU145 CTRL n = 423, DU145 SS n = 253, PC3 CTRL n = 203, PC3 SS n = 132. *p = 0.01, **$p=0.0046$, ***$p\ll 1\times {10}^{-6}$, statistical comparisons used the two-sample Kolmogorov–Smirnov test. Data was obtained from three independent experiments.

To confirm these observed trends, western blots were performed for key proteins V-ATPase, Raptor, and p-MTORC2 in highly metastatic PC3 cells, as shown in Fig. 5. PC3 cells were selected since they grow to confluence more quickly on tubing, due to their highly metastatic nature and significantly more cells are needed for blots in comparison with ICC. While quantified levels of all of these proteins were found to increase with shear stress—$\sim$ 87% for V–H${}^{+}$-ATPase, $\sim$ 155% for Raptor, and $\sim$ 242% for p-MTORC2 (Fig. 5d)—only p-MTORC2 expression levels increased significantly. This is likely due to fewer samples of the blots (n = 3) in comparison with ICC results ($n>100$).

Figure 5.

Western blots of V-ATPase, mTORC1 component Raptor, and p-mTORC2 from PC3 cells exposed to 0.05 dynes ${\text{cm}}^{-2}$ of wall shear stress for 24 h (SS) and control (CTRL) cells. (a) to (c): left panel, raw images of western blots. Right panel, contrast enhanced (CE) images. (a) Representative blots for V-ATPase’s c-subunit and $\beta$-actin loading control. In CE images, grayscale values from [42, 163] were rescaled to [0, 255]. (b) Representative blots for Raptor and $\beta$-actin loading control. Grayscale values in CE images were rescaled from [73, 241] to [0, 255]. (c) Representative blots for p-mTORC2 and $\beta$-actin loading control. Grayscale values in CE images were mapped from [39, 193] to [0, 255]. (d) Blot quantifications of band densities normalized by loading control. n = 3 from three independent experiments, *p = 0.033, two sample Kolmogorov–Smirnov test.

Proton Pump Inhibitor Bafilomycin A1 Decreases Fluorescence of Endosome, Lysosome, $\beta$1 integrin, F-actin, V–H${}^{+}$-ATPase, and mTOR Markers

Moderately metastatic DU145 cells were cultured on tygon tubes (500 $\mu$m diameter) and treated with 60 nM Bafilomycin A1 (BAF) under 0.05 dynes ${\text{cm}}^{-2}$ shear stress for 24 h, or media containing the equivalent volume of dimethyl sulfoxide (DMSO) as the BAF treated cells (CTRL) under 0.05 dynes ${\text{cm}}^{-2}$ shear stress for 24 h. Bafilomycin A1 (BAF), a known proton pump inhibitor, binds to subunits a and c of the proton pump’s V${}_0$ domain inhibiting proton transport, in addition to reducing expression of three V–H${}^{+}$-ATPase genes.23 DU145 cells were selected for proton pump inhibitor treatment since they showed the strongest increase in V–H${}^{+}$-ATPase fluorescence under shear stress.

Figures 6a and 6c show that cells exposed to both BAF and SS significantly decrease F-actin stain phalloidin fluorescence in comparison with CTRL (with $\sim$ 15% of CTRL value). F-actin fluorescence density can decrease with BAF treatment due to decreasing V–H${}^{+}$-ATPase’s effect of polymerizing and stabilizing F-actin, as observed in tobacco hornworm midgut cells.34 Decreased F-actin in cells due to proton pump inhibition may also result in tumor cell rupture and death due to higher levels of shear stress,35 reducing metastasis.

Figure 6.

V–H${}^{+}$-ATPase inhibition decreases F-actin and $\beta$1 integrin fluorescence in shear stressed DU145 cells. Cells were exposed to 60 nM Bafilomycin A1 (BAF) dissolved in DMSO, or the equivalent volume of DMSO as BAF cells (CTRL) under 0.05 dynes ${\text{cm}}^{-2}$ shear stress for 24 h. (a) to (b): Upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are 20 $\mu$m. (a) Phalloidin (a specific stain for F-actin) fluorescence for CTRL and BAF treated cells. Inverted image grayscale values in the range of [60, 255] were linearly rescaled to [0 255]. (b) $\beta$1 integrin antibody fluorescence for CTRL and BAF treated cells. Inverted image grayscale values in the range of [60, 255] were also linearly rescaled to [0 255]. (c) Average F-actin and $\beta$1 integrin antibody cell fluorescence densities ± SE for CTRL and BAF treated cells. F-actin cases: CTRL n = 153, BAF n = 138; $\beta$1 integrin cases: CTRL n = 103, BAF n = 102. ***$p\ll 1\times {10}^{-6}$, two-sample Kolmogorov–Smirnov test. Data was obtained from 4 independent experiments.

The observed decrease in $\beta$1 integrin fluorescence ($\sim$ 37% of CTRL, Figs. 6b and 6c) in comparison with control suggests that the proton pump inhibitor reduces $\beta$1 integrin transport to the plasma membrane by endosomes and lysosomes in response to shear stress.6 Supporting this, Fig. 7 shows reduced fluorescence of EEA1 ($\sim$ 15% of CTRL) and LAMP2 ($\sim$ 28% of CTRL) in BAF treated DU145 cells compared with SS controls, indicating fewer endosomes and lysosomes are present in cells with BAF treatment.

Figure 7.

V–H${}^{+}$-ATPase inhibition decreases endosome and lysosome fluorescence in shear stressed DU145 cells. Cells were exposed to 60 nM Bafilomycin A1 (BAF) dissolved in DMSO, or the equivalent volume of DMSO as BAF cells (CTRL) under 0.05 dynes ${\text{cm}}^{-2}$ for 24 h. (a) to (b): upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are $20\mu \text{m}$. (a) Endosome antibody EEA1 fluorescence for CTRL and BAF treated cells. Grayscale values in the range of [130, 240] were linearly rescaled to [0 255] in inverted images. (b) Lysosome antibody LAMP2 fluorescence for CTRL and BAF treated cells. Inverted image grayscale values in the range of [130, 245] were linearly rescaled to [0 255]. (c) Average EEA1 marker cell fluorescence density ± SE for CTRL (n = 140) and BAF (n = 124) treated cells, and average LAMP2 marker fluorescence density ± SE for CTRL (n = 109) and BAF (n = 157) treated cells. ***$p\ll 1\times {10}^{-6}$, two-sample Kolmogorov–Smirnov test. Data was obtained from 4 independent experiments.

There was also a significant decrease in V–H${}^{+}$-ATPase fluorescence in cells treated with BAF compared with CTRL ($\sim$ 14% of CTRL, Figs. 8a and 8c), as expected. The significant reduction in both Raptor ($\sim$ 28% of CTRL) and p-mTORC2 ($\sim$ 72% of CTRL) marker fluorescence in BAF treated cells compared with CTRL cells (Figs. 8b and 8c) supports our hypothesis that proton pumps activate both mTOR complexes in response to mechanotransduction, thereby making cancer more aggressive under low shear stress.

Figure 8.

V–H${}^{+}$-ATPase inhibition decreases mTORC1 component Raptor, and p-mTORC2 fluorescence in shear stressed DU145 cells. Cells were exposed to 60 nM Bafilomycin A1 (BAF) dissolved in DMSO, or the equivalent volume of DMSO as BAF cells (CTRL) under 0.05 dynes ${\text{cm}}^{-2}$ for 24 h. (a) to (c): Upper panel, raw fluorescence images. Lower panel, contrast enhanced inverted images. Scale bars are $20\mu \text{m}$. (a) V–H${}^{+}$-ATPase (V-ATPase), (b) Raptor, and (c) p-mTORC2 antibody fluorescence for CTRL and BAF treated cells. Inverted images were contrast enhanced with the following linear rescalings: (a) [130, 240] to [0, 255], (b) [100, 245] to [0, 255], and (c) [130, 250] to [0. 255]. (d) Average V–H${}^{+}$-ATPase, Raptor, and p-mTORC2 cell fluorescence densities ± SE for CTRL and BAF treated cells. V–H${}^{+}$-ATPase: CTRL n = 166, BAF n = 140; Raptor: CTRL n = 93, BAF n = 132; p-mTORC2: CTRL n = 116, BAF n = 186. **$p=3.5\times {10}^{-4}$, ***$p\ll 1\times {10}^{-6}$, two-sample Kolmogorov–Smirnov test. Data was obtained from 4 independent experiments.

To confirm these trends of decreased protein levels with Bafilomycin A1 (BAF) treatment, as revealed by ICC, highly metastatic PC3 cells were cultured on tygon tubes and treated with 60 nM BAF under 0.05 dynes ${\text{cm}}^{-2}$ shear stress for 24 h. CTRL cells were cultured with media containing the equivalent volume of dimethyl sulfoxide (DMSO) as the BAF treated cells under 0.05 dynes ${\text{cm}}^{-2}$ shear stress for 24 h. PC3 cells were used for western blots for key proteins V–H${}^{+}$-ATPase (V-ATPase), Raptor, and p-mTORC2, shown in Fig. 9, to enable comparisons with Fig. 6 and due to their faster growth to confluence on the tubes. We observe that shear-stressed PC3 cells treated with BAF replicate the ICC trends for DU145 cells—V-ATPase BAF has 87% of CTRL level, Raptor BAF has 80% of CTRL level, and p-mTORC2 BAF has 79% of CTRL level—see Fig. 9d. Of these decreases, only the p-mTORC2 finding is statistically significant (with p < 0.05).

Figure 9.

Western blots of V-ATPase, mTORC1 component Raptor, and p-mTORC2 from PC3 cells treated with V-ATPase inhibitor BAF and shear stress. Cells were exposed to 60 nM Bafilomycin A1 (BAF) dissolved in DMSO, or the equivalent volume of DMSO as BAF cells (CTRL) under 0.05 dynes ${\text{cm}}^{-2}$ for 24 h. (a) to (c): Left panel, raw images of western blots. Right panel, contrast enhanced images. (a) Representative blots for V-ATPase’s c-subunit and $\beta$-actin loading control. In contrast enhanced (CE) images, grayscale values from [30, 205] were rescaled to [0, 255]. (b) Representative blots for Raptor and $\beta$-actin loading control. Grayscale values in CE images were rescaled from [73, 231] to [0, 255]. (c) Representative blots for p-mTORC2 and $\beta$-actin loading control. Grayscale values in CE images were mapped from [36, 238] to [0, 255]. (d) Blot quantifications of band densities normalized by loading control. n = 3 from three independent experiments, *p = 0.033, two sample Kolmogorov–Smirnov test.

Discussion

This study was designed to demonstrate that proton pumps and mTOR complexes 1 and 2, factors that increase cancer aggressiveness, directly affect prostate cancer cell mechanotransduction. To this end, we have applied low levels of shear stress to moderately and highly metastatic prostate cancer cells cultured on the inner sidewalls of narrow tubing—a low throughput technique with spatially homogeneous and accurately calculated shear stress using a physiologically relevant geometry similar to lymphatic vessels. A low throughput technique necessitates demonstrating cell changes using indirect immunocytochemistry techniques on individual cells to estimate changes in levels of protein expressions. Western blots were also used, in upscaled experiments, to quantify key protein relative expressions—V–H${}^{+}$-ATPase, Raptor (a component of mTORC1), and p-mTORC2. Such blots confirmed the trends of increases due to shear stress, though with less statistical significance due to smaller sample sizes.

We show that, as expected, low shear stress (0.05 dynes ${\text{cm}}^{-2}$ for 24 h) significantly increases peripheral F-actin polymerization relative to non-shear stressed control prostate cancer cells (Fig. 2a), since actin responds to physical forces applied to cells.9,13 Indeed, fluid shear stress has been reported to increase cortical F-actin in laryngeal squamous carcinoma and mesenchymal stem cells16,19; however, a prior prostate cancer cell study reported that peripheral F-actin did not increase in highly metastatic PC3 prostate cancer cells exposed to 0.05 dynes ${\text{cm}}^{-2}$ of shear stress for 3 h.18 Our report may be the first to demonstrate that prostate cancer cells exposed to shear stress increase cortical F-actin polymerization. This can be due to possibly increased V–H${}^{+}$-ATPase at the plasma membrane, delivered by increased acidic vesicle transport in response to shear stress (Fig. 10), since V–H${}^{+}$-ATPase promotes F-actin polymerization.34 Increased F-actin polymerization can also be due to increased mTORC2 activation, as this pathway is attributed to cytoskeletal remodeling.17

Figure 10.

Schematic of the hypothesized signaling pathway whereby V–H${}^{+}$-ATPase (V-ATPase), mTORC1, and mTORC2 increase due to shear stress. (a) Shear stress activates integrin, which causes a rise in F-actin polymerization, stimulating increased transport of acidic vesicles to the plasma membrane by molecular motors. Such vesicles can transport integrins and V-ATPase pumps to the plasma membrane upon fusion; mTORC1 is additionally activated on lysosome membranes. mTORC2 is likely to be activated by Rac1, which is stimulated by V-ATPase. (b) Inhibiting V-ATPase with BAF can decrease F-actin polymerization and stability. Inhibition also disrupts acidic vesicles, whose transport to the plasma membrane is reduced due to decreased polymerized actin, resulting in fewer V-ATPase pumps and integrins at the plasma membrane, and reduced mTORC1 and mTORC2 activity.

We have also shown that shear stressed prostate cancer cells increase $\beta$1 integrin at the cell periphery (Fig. 2b). While this effect has not been previously reported for prostate cancer cells, to our knowledge, similar observations have been made in vascular endothelial cells and human umbilical vein endothelial cells exposed to shear stress.14,15 This increase can occur in response to shear stress due to increased lysosome transport, as we hypothesize for V–H${}^{+}$-ATPase, as lysosomes can traffic integrin to the plasma membrane6,26 (Fig. 10). Our results also indicate that endosome and lysosome markers increase in prostate cancer cells exposed to shear stress relative to control (Fig. 3)—perhaps due to increased production and in response to shear stress, and also to the more metastatic phenotype.5 While it has been demonstrated that proximal tubule cells increase endocytosis in response to fluid shear stress,25 this is the first report to our knowledge of endosome and lysosome changes in prostate cancer cells in response to shear stress.

Our finding that shear stress significantly increases V–H${}^{+}$-ATPase marker fluorescence in moderately metastatic DU145 and highly metastatic PC3 cells (Fig. 4a), and non-significantly increases V–H${}^{+}$-ATPase expression levels in blots for highly metastatic PC3 cells (Fig. 5a), indicates that proton pumps may be mechanosensors. Our report is the first to demonstrate this pump’s involvement in prostate cancer mechanotransduction, and the increased vacuolar proton pumps may raise cancer aggressiveness via translocation to the plasma membrane, acidifying the tumor microenvironment and activating metalloproteases to degrade ECM.3,8,21,22 We also show, indirectly using ICC, that mTORC1 component Raptor increases significantly in DU145 and PC3 prostate cancer cells due to shear stress (Fig. 4b), and increases non-significantly in western blots for PC3 cells (Fig. 5b). This was expected, as V–H${}^{+}$-ATPase activates mTORC1 on lysosome membranes.8,39 We have not shown that the proton pumps in cells exposed to shear stress are functional and active; however, prior work39 also did not demonstrate that the vacuolar pumps were functional and actively acidifying endosomes and lysosomes to activate mTORC1 on lysosome membranes in human embryonic kidney cells. Our planned future studies should establish that shear stress exposed cancer cells translocate V–H${}^{+}$-ATPase to the plasma membrane by measuring intracellular and extracellular pH.

Our results show that shear stress also significantly increases p-mTORC2 marker expression in prostate cancer cells (Figs. 4c and 6c). Interestingly, the ICC images show that the majority of the antibody fluorescence is in the nucleus (Fig. 4c). This may indicate a role for mTORC2 in transcription in prostate cancer cells,29 that is increased by mechanotransduction. Though this complex is responsible for cytoskeletal remodeling,17 this is the first evidence of shear stress-induced increases and activations of mTORC2 and V–H${}^{+}$-ATPase in prostate cancer cells. Prior studies18,32 suggest that YAP (yes-associated protein) pathway activation may be responsible for this increased aggressiveness, due to mTORC1 and mTORC2 regulation. However, our specific V–H${}^{+}$-ATPase inhibitor experiments show that it is not only YAP, but also proton pumps that can regulate mTORC1 and mTORC2. Possible connections between the YAP pathway and proton pumps warrant further study. Also, our results suggest that proton pump inhibition may effectively treat cancer.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.