Caveolin‐1 promotes mitochondrial health and limits mitochondrial ROS through ROCK/AMPK regulation of basal mitophagic flux

Caveolin‐1 (CAV1), the main structural component of caveolae, is phosphorylated at tyrosine‐14 (pCAV1), regulates signal transduction, mechanotransduction, and mitochondrial function, and plays contrasting roles in cancer progression. We report that CRISPR/Cas9 knockout (KO) of CAV1 increases mitochondrial oxidative phosphorylation, increases mitochondrial potential, and reduces ROS in MDA‐MB‐231 triple‐negative breast cancer cells. Supporting a role for pCAV1, these effects are reversed upon expression of CAV1 phosphomimetic CAV1 Y14D but not non‐phosphorylatable CAV1 Y14F. pCAV1 is a known effector of Rho‐associated kinase (ROCK) signaling and ROCK1/2 signaling mediates CAV1 promotion of increased mitochondrial potential and decreased ROS production in MDA‐MB‐231 cells. CAV1/ROCK control of mitochondrial potential and ROS is caveolae‐independent as similar results were observed in PC3 prostate cancer cells lacking caveolae. Increased mitochondrial health and reduced ROS in CAV1 KO MDA‐MB‐231 cells were reversed by knockdown of the autophagy protein ATG5, mitophagy regulator PINK1 or the mitochondrial fission protein Drp1 and therefore due to mitophagy. Use of the mitoKeima mitophagy probe confirmed that CAV1 signaling through ROCK inhibited basal mitophagic flux. Activation of AMPK, a major mitochondrial homeostasis protein inhibited by ROCK, is inhibited by CAV1‐ROCK signaling and mediates the increased mitochondrial potential, decreased ROS, and decreased basal mitophagy flux observed in wild‐type MDA‐MB‐231 cells. CAV1 regulation of mitochondrial health and ROS in cancer cells therefore occurs via ROCK‐dependent inhibition of AMPK. This study therefore links pCAV1 signaling activity at the plasma membrane with its regulation of mitochondrial activity and cancer cell metabolism through control of mitophagy.


| INTRODUCTION
Cancer cell development requires adaptive metabolic reprogramming to maintain cell survival and proliferation.Mitochondrial dynamics, biogenesis, and recycling mediate energy levels in the cell and their homeostasis is a central part of metabolic function in cancer cells. 1 Damaged or dysfunctional mitochondria can lead to excess ROS production, which has long been studied in cancer as a cause of DNA damage and genetic instability. 2Eukaryotic cells have diverse mechanisms in which they regulate mitochondrial health and function, one of which is the selective degradation of damaged mitochondria, a process known as mitophagy. 3Defining molecular mechanisms that cancer cells use to manage mitochondrial health and ROS production through mitophagy represents a valuable approach to target select populations of cancer cells and reverse the cancer cell metabolic phenotype.
][6][7][8][9][10] However, CAV1 has also been ascribed tumor suppressor functions, and in some tumor types, CAV1 expression is associated with improved prognosis. 11The varied impact of CAV1 expression on cancer prognosis highlights the diverse roles this protein plays in cancer progression such that targeting this protein for cancer therapy is particularly challenging.
Canonical pro-cancer roles of CAV1 are related to Srcdependent phosphorylation of CAV1 on tyrosine 14 (Y14) promoting focal adhesion (FA) turnover and tension, and promote tumor cell migration and metastasis. 12][15][16][17][18] Furthermore, Y14 phosphorylated CAV1 (pCAV1) has been shown to promote invasion and extravasation of MDA-MB-435 and B16F10 melanoma cells. 19,20These mechanosensitive signaling functions of pCAV1 at focal adhesions are caveolae independent as they occur in PC3 prostate cancer cells that lack the caveolae adaptor proteins CAVIN1 and, therefore, caveolae. 4,17][23][24] Similarly, CAV1 shows diverse roles in regulation of cellular metabolism and specifically mitochondrial function.For example, CAV1 knockout mice present altered mitochondrial function in adipocytes and CAV1 knockout MEFs present proteasomal degradation of mitochondrial complexes I, III, IV, and IV. 25,26However, in metastatic cancer cells, CAV1 was found to directly inhibit mitochondrial complex IV function. 27In A549 lung cancer cells, CAV1 promotes Parkin-mediated mitophagy 28 while in lung epithelial Beas-2B cells CAV1 was shown to directly associate with the ATG5/ATG12 complex preventing autophagosome formation and inhibiting autophagy. 29onsistently, autophagy was shown to be upregulated in adipocytes of CAV1-deficient mice. 30Interestingly, recent studies showed that CAV1 promotes mitochondrial dysfunction in a pCAV1-dependent manner in metastatic cells and that CAV1 regulates mitophagy by mediating mitochondrial fission-fusion dynamics. 27,31However, the relationship between mechanosensitive CAV1 signaling and CAV1 metabolic regulation remains unclear.
ROCK is associated with diverse metabolic functions depending on tissue type, and several studies have focused on ROCK as a therapeutic target for metabolic diseases such as diabetes and obesity.Recently, ROCK was shown to be a potential target for Parkinson's disease therapeutics, as ROCK inhibition promoted mitophagy in neurons. 32Other metabolic disorders such as hypercholesterolemia, glucose intolerance, and obesity in mice fed high-fat diets were also shown to be ROCK dependent.Interestingly, these phenotypes were lost upon inhibition of ROCK, which increased oxygen consumption and inhibited by CAV1-ROCK signaling and mediates the increased mitochondrial potential, decreased ROS, and decreased basal mitophagy flux observed in wildtype MDA-MB-231 cells.CAV1 regulation of mitochondrial health and ROS in cancer cells therefore occurs via ROCK-dependent inhibition of AMPK.This study therefore links pCAV1 signaling activity at the plasma membrane with its regulation of mitochondrial activity and cancer cell metabolism through control of mitophagy.

K E Y W O R D S
AMPK, caveolin-1, mitochondria, mitophagy, reactive oxygen species, rho kinase AMPK activation. 33AMPK is activated in response to low energy levels and cellular stress and has been shown to regulate mitochondrial number and quality through biogenesis and mitophagy. 346][37] Therefore, we hypothesized that CAV1 may be implicated in the regulation of mitochondrial health and ROS production of cancer cells through ROCK/AMPK signaling.Here, using CRISPR/Cas9 knockout of CAV1 in metastatic MDA-MB-231 breast cancer and PC3 prostate cancer cells, we show that pCAV1 signaling via ROCK mediates mitochondrial health and ROS production by inhibiting AMPK activation and thereby limiting mitophagy.
The MDA-MB-231 clonal cell lines and CRISPR/Cas9 CAV1 knockout MDA-MB-231 cell line used in this study were as described. 38To generate a CAV1 KO PC3 cell line, parental PC3 cells (ATCC) were first cloned and a wild-type clone chosen based on morphological similarly to parental PC3 cells and CAV1-dependent migration. 38CRISPR/CAS CAV1 KO was performed using sgRNA designed based on previously published gRNA1 sequence targeting the initiation codon of CAV1 in exon 1. 38 Briefly, RNP (Ribo-Nucleo Protein) complex was prepared with Cav1 initiation codon targeting custom sgRNA dissolved in 1xIDTE pH 7 buffer, (Cat#11-01-02-02, IDT), Alt-R® S.p.Cas9 Nuclease V3 (Cat#216350294, IDT), Cas9plus (Cat#CMAX00001, Invitrogen), and OptiMEM (Cat#31985070, Invitrogen) following the protocol by IDT and 40K PC3 cells were transiently reverse transfected using Lipofectamin CRISPRmax (cat#CMAX00001 Invitrogen) for 48 h.Sets without or with sgRNA were washed with PBS and genomic DNA was extracted using Lucigen QuickExtract DNAextraction solution (Cat# LGN-QE0905T, Lucigen) to perform Genomic Cleavage Detection assay (A24372, Invitrogen, USA) to check the cleavage efficiency.Based on optimized cleavage efficiency, reverse transfection was repeated and post 48 h incubation, cells were trypsinized and singly plated in 96 well plates by limiting dilution and pates were incubated till the colonies were formed.
Replicating colonies, screenings and INDEL detections were done as described previously. 38o generate the stable CAV1-WT-Myc rescue cell lines (rescue #19, #20, #23) in CAV1 KO cells, approximately 700 000 CAV1 KO MDA-MB-231 cells at passage 1 were plated in a 10 cm dish and incubated at 37°C in a CO 2 incubator.After 24 h, cells were transiently transfected with CAV1-WT-Myc mammalian expression plasmid using Effectene (Qiagen) transfection reagent and following the manufacturer's protocol.Transfected cells were incubated for 72 h at 37°C.Cells were washed with warmed phosphate-buffered saline (PBS) and subjected to geneticin selection (500 μg/mL) in fresh media for a minimum of 2 weeks.During this selection period, cells were washed with warmed PBS and media supplemented with geneticin was changed every 48 h.This was repeated until the resistant cell population emerged forming colonies.At least 20 colonies were hand-picked under the microscope in a biosafety cabinet and plated in 12-well plates.These colonies were replicated, a set was frozen in freezing media for storage in liquid nitrogen for future use, and other cells were harvested for western blotting as described below.
Transfection of plasmids, siRNA, double siRNA, or combinations of siRNA and plasmids was performed using Lipofectamine 2000 (Invitrogen, USA) with OptiMEM (ThermoFisher, USA) transfection media, incubated for 5 h.The cells were then washed and incubated for 48 h with fresh media.
Where indicated, cells were treated with 20 μM Y-27632 or an equivalent volume of DMSO as a control for 1 h prior to imaging.Furthermore, where indicated, cells were treated with 1X CCCP or an equivalent volume of DMSO as a control for 24 h prior to imaging.

| Seahorse XF96 stress test assay
The oxygen consumption rate of cells was measured using the Seahorse XF96 mitochondrial stress test (Seahorse Bioscience, North Billerica, MA). 25 000 cells/well were plated in growth media for 18 h prior to testing.OCR was measured in Seahorse XF base medium containing minimal DMEM, 10 mM glucose, 4 mM L-glutamine, and 2 mM pyruvate, as per the test protocol.The cells and sensor cartridge were washed and incubated in assay medium and calibrant solution for 1 h at 37°C, 0% CO 2 for equilibration.Injector ports were loaded to final concentrations of 2 μM oligomycin, 3 μM FCCP, and 1 μM rotenone/antimycin, respectively.The Pierce Micro BCA Protein Assay Kit (ThermoFisher) was used to measure total protein, which was then used to normalize OCR. 3 replicate experiments were completed with 7-8 wells per cell line per experiment.

| MV633 and MitoSOX red live cell
imaging and analysis 30 000 cells were seeded into μ-Slide glass bottom ibidi chambers (#80827) and the indicated transfections and/ or treatments were performed.MV633 and MitoSOX Red labeling were done by incubating the cells for 30 min at working concentrations of 50 nM and 2.5 μM, respectively, prior to imaging, as previously described. 39Cells were then gently washed thrice with Live Cell Imaging Solution (A14291DJ) and incubated with this solution while imaging.
Confocal images taken of live cells were taken with the 100×/1.40Oil HC PL APO CS2 objective of a Leica TCS SP8 confocal microscope (Wetzlar, Germany) in an environmental chamber system set to 37°C.Image acquisition was done in a temperature-controlled system set to 37°C.Confocal images of MV633 and MitoSOX were analyzed using ImageJ analysis software.
The maximum brightness of the images was increased to visualize and draw a region of interest (ROI) around the cell outline.The brightness of the image was then adjusted so that the background had an integrated density (IntDen) of 0. Mean gray value of segmented mitochondria was measured for each ROI.

| Mito-Keima imaging and analysis
20 000 cells per well were seeded in Ibidi chambers for 24 h prior to transfection with the mito-Keima plasmid.Transfection and treatment protocol was as described above.Cells were washed 3x and then incubated with Live Cell Imaging Solution and imaged on a Leica TCS SP8 confocal microscope equipped with a 100×/1.40Oil HC PL APO CS2 objective (Leica, Wetzlar, Germany), white light laser and HyD detectors (Leica, Wetzlar, Germany), as previously described. 39At physiological pH of mitochondria (pH 8.0), mito-Keima was detected by excitation at 470 nm and in an acidic environment (pH 4.5), it was detected by excitation at 561 nm.Quantification of detected mitolysosomes was done using with the mito-QC Counter macro installed in FIJI analysis software. 40The following settings for ratio analysis were determined to report on mitolysosome expression most accurately across all data sets: For 231 cells, radius for smoothing = 8, ratio threshold = 1, red channel threshold = 2 and for PC3 cells, radius for smoothing = 6, ratio threshold = 11, red channel threshold = 2.In some images, the software detected mitolysosomes where they were not present.For consistency, images with the largest and smallest number of mitolysosomes from each group were removed from the analysis.

| Western blots
Confluent cells were harvested in PBS using a cell scraper and spun at 5000 rpm for 5 min in a microcentrifuge.The resulting cell pellet was lysed using M2 (20 mM Tris HCl (pH 7.6); 0.5% NP-40/IGAPAL; 250 mM NaCl; 3 mM EGTA (pH 8.0); and 3 mM EDTA (pH 8.0) with added protease and phosphatase cocktail inhibitor tablets (Roche) and incubated for 40 min at 4°C.Samples blotted for pAMPK were lysed with M2 lysis buffer as described above and supplemented with activated 0.2 M Na3VO4 and 0.5 M NaF to ensure no phosphatase activity.Cells were then centrifuged for 15 min at 13 200 RPM and 4°C to separate protein from cell debris.The protein concentration was determined using a Bradford Assay and equal amounts of protein were loaded onto the sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) gel and allowed to run at 135 V and 2.00A for approximately 1 h.The proteins were transferred to a nitrocellulose membrane with the semi-dry BioRad transfer system.The membrane was blocked with 5% milk product in PBS for 1 h and then incubated with appropriate concentrations of the respective primary antibodies overnight at 4°C.The following day the membranes were washed and incubated with either rabbit or mouse horse radish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature.The membrane was washed three times with PBS-T for 10 min each.Finally, enhanced chemiluminescence (ECL) was used to visualize the proteins.

| Statistical analysis
The statistical analysis of data presented within this thesis was performed using GraphPad Prism 8.0 software.Student's t-test or one-way analysis of variance with Tukey's multiple comparisons test was used.p < .05 is considered significant.

| CAV1 knockout increases mitochondrial activity in MDA-MB-231 cells
The metastatic, triple-negative breast cancer cell line MDA-MB-231 expresses high levels of CAV1 and pCAV1. 4D volumetric analysis of CRISPR/Cas9 CAV1 knockout MDA-MB-231 clones 38 labeled for the mitochondrial protein TOM20 shows a significant increase in mitochondrial volume (Figure 1A).To explore the role of CAV1 in regulating mitochondrial metabolic activity, we measured oxygen consumption rates (OCR) using the Seahorse XF96 stress test in CRISPR/Cas9 CAV1 knockout MDA-MB-231 clones.Basal, ATP-linked, maximum respiration, and reserve capacity all showed a significant increase in OCR in CAV1 KO cells (Figure 1B).
Consistently, labeling of live cells with the fluorescent, potential dependent dye, MitoView 633 (MV633) showed that CAV 1 KO significantly increased the intensity of MV633 relative to WT cells, indicating higher mitochondrial potential in the KO cells (Figure 2A,B).Rescue of CAV1 KO cells by stable expression of CAV1 (rescue clones #19, #20, and #23) (Figure S1A) reduced mitochondrial potential compared to KO cells (Figure 2A,B).Live cell labeling with the mitochondria-targeted fluorogenic ROS reporter, MitoSOX Red, showed that CAV1 KO significantly decreased MitoSOX fluorescence compared to WT cells and the three CAV1 rescue clones (Figure 2A,C).Overall, CAV1 expression reduces oxidative respiration and mitochondrial potential of MDA-MB-231 breast cancer cells increasing mitochondrial ROS production.

| Caveolae-independent CAV1-ROCK signaling decreases mitochondrial potential and increases ROS
To determine if pCAV1 signaling impacts mitochondrial health, CAV1 WT, phosphomimetic Y14D, or non-phosphorylatable Y14F red fluorescent protein (RFP)-tagged mutants were exogenously expressed in MDA-MB-231 CAV1 KO cells and the cells incubated with MV633.CAV1 mutant-expressing cells were identified by their RFP fluorescence and cells chosen for analysis displayed comparable levels of RFP expression.Mitochondria from cells expressing CAV1 WT or Y14D exhibit a noticeable decrease in fluorescence intensity compared to their non-transfected CAV1 KO neighbors (Figure 3A).Expression of the non-phosphorylatable Y14F mutant did not reduce MV633 labeling intensity that remained significantly higher than WT cells (Figure 3A).These data suggest that reduced mitochondrial potential exhibited by CAV1 expressing cells is pCAV1 dependent.
To test whether pCAV1 regulates mitochondrial potential through downstream ROCK signaling, CAV1 WT, KO, and CAV1 rescue #23 MDA-MB-231 cells were treated with the ROCK1/2 inhibitor Y-27632 for 1 h.Using MV633, Y-27632 treatment increased mitochondrial potential of WT and CAV1 rescue #23 but not CAV1 KO MDA-MB-231 cells (Figure 3B).Consistently, knockdown of ROCK1 or ROCK2 in CAV1 WT MDA-MB-231 cells significantly increased MV633 mitochondrial potential and reduced mitochondrial ROS production while CAV1 KO cells were not significantly affected (Figure 3C).Furthermore, siROCK1 and siROCK2 reduced mitochondrial ROS production in CAV1 WT MDA-MB-231 cells while CAV1 KO cells were not significantly affected (Figure 3C).To determine whether pCAV1 reduction of mitochondrial potential was mediated through ROCK activity, we rescued MDA-MB-231 CAV1 KO cells by transfection with CAV1 WT, Y14D, or Y14F and also transfected the cells with siRNA against ROCK1 or ROCK2 (Figure 3D).Knockdown of either ROCK1 or ROCK2 prevented reduction of mitochondrial potential in CAV1 KO cells by expression of either CAV1 WT or Y14D.
PC3 cells have high CAV1 expression but lack Cavin1, the adaptor protein that interacts with CAV1 and is required for formation of caveolae. 41Therefore, CAV1 exists in the membrane in non-caveolar scaffold domains.CAV1 was knocked out in PC3 cells using CRISPR/Cas9 (Figure S2A).As observed for MDA-MB-231 cells, incubation of CAV1 WT and KO PC3 cells with MV633 and MitoSOX showed that CAV1 KO PC3 cells exhibited increased mitochondrial potential and reduced ROS compared to WT PC3 cells (Figure 4A).Expression of cavin-1 in PC3 WT cells reversed the phenotype, reducing ROS expression (Figure S2B,C).Treatment of PC3 CAV1 WT and KO cells with Y-27632 increased mitochondrial potential in both cell lines to the same level (Figure 4B).As for MDA-MB-231 cells, mitochondria from PC3 CAV1 KO cells transiently expressing CAV1 WT or phosphomimetic CAV1 Y14D, but not non-phosphorylatable CAV1 Y14F, showed significantly decreased MV633 intensity compared to the non-transfected CAV1 KO cells (Figure 4C).Overall, these results suggest that pCAV1 activation of ROCK decreases mitochondrial potential and increases ROS production independently of caveolae via CAV1 scaffolds. 42

| CAV1 activation of ROCK limits mitophagic flux
To assess whether mitophagy played a role in the enhanced mitochondrial health of CAV1 KO cells, we knocked down the autophagy protein ATG5 and mitophagy sensor PINK1 in CAV1 WT and KO MDA-MB-231 cells (Figure S3).Both ATG5 and PINK1 knockdown decreased MV633 mitochondrial potential and increased MitoSOX ROS levels of CAV1 KO MDA-MB-231 cells to the levels of WT MDA-MB-231 cell (Figure 5A).Mitochondrial fission is critical for mitochondrial health and can participate in mitophagy by isolating damaged mitochondria. 43Knockdown of the mitochondrial fission inducer, Drp1, reduced mitochondrial potential and increased mitochondrial ROS selectively in CAV1 KO MDA-MB-231 cells (Figure 5B).These results show that the enhanced basal mitophagy and mitochondrial fusion in CAV1 KO MDA-MB-231 cells are critical to the improved mitochondrial health of these cells, as reflected in the increased mitochondrial potential and reduced ROS upon knockout of CAV1.
To better define CAV1 regulation of basal mitophagy, we next employed the mito-Keima pH-sensitive fluorescent mitophagy probe.At the neutral pH of mitochondria, mito-Keima is excited by shorter wavelengths.When mitochondria are engulfed within the acidic lysosome, the probe is excited by longer wavelengths and therefore allows for the visualization of mitolysosome formation. 44ive cell images of CAV1 WT, KO, and rescue #23 MDA-MB-231 cells transiently transfected with the mito-Keima probe were analyzed using an ImageJ plugin 40 to detect mitochondria-associated lysosomes (Figure 6A).The number and area of mitolysosomes were significantly higher in the CAV1 KO MDA-MB-231 cells compared to WT or rescue #23.Treating the cells with Y-27632 significantly increased mitolysosome number and area in CAV1-expressing cells to the level of CAV1 KO cells (Figure 6A).Cells were treated with the mitochondrial depolarizing agent, carbonyl cyanide m-chlorophenylhydrazone (CCCP) as a positive control to induce mitophagy.Application of the mito-Keima probe to PC3 cells showed that CAV1 KO PC3 cells exhibited more mitolysosomes and larger areas than WT cells (Figure 6B), consistent with CAV1 regulation of mitophagy in these cells.Treatment of CAV1 expressing PC3 cells with Y-27632 increased mitolysosome number to levels of CAV1 KO PC3 cells (Figure 6B).To further validate that CAV1 activation of ROCK limits mitophagy, CAV1 WT and KO MDA-MB-231 cells were co-transfected with siROCK1 or siROCK2 and mito-Keima (Figure 7).As observed for Y-27632 treatment, siRNA knockdown of ROCK1 or ROCK2 in CAV1-expressing cells increased mitolysosome number and area to the levels of CAV1 KO cells (Figure 7).Overall, these data suggest that CAV1 inhibits mitophagy via downstream ROCK signaling.

| ROCK-dependent activation of AMPK in CAV1 KO cells increases mitophagy and improves mitochondrial health
AMPK is a crucial regulator of mitochondrial homeostasis, impacting mitochondrial biogenesis, dynamics, and mitophagy, whose activation is inhibited by ROCK. 33,45revious studies have investigated the relationship between CAV1 and AMPK and multiple studies report that absence of CAV1 results in AMPK activation. 35,36,46Western blot analysis showed that CAV1 KO MDA-MB-231 cells have increased levels of active AMPK (pAMPK) (Figure 8A).To test whether CAV1 KO cells present increased mitophagy due to increased active AMPK levels, we knocked down AMPK with siRNA in CAV1 WT and KO MDA-MB-231 cells and incubated the cells with MV633 and MitoSOX.Knockdown of AMPK did not impact mitochondrial potential or ROS production in MDA-MB-231 cells but selectively reduced MV633 intensity and increased MitoSOX intensity of CAV1 KO MDA-MB-231 cells to the levels of WT cells (Figure 8B).Use of the mito-Keima probe showed that knockdown of AMPK in CAV1 KO cells reduced basal mitophagic flux in these cells (Figure 8C).These results suggest that CAV1 KO cells utilize active AMPK as a regulator of mitochondrial health.
To determine whether ROCK signaling downstream of CAV1 regulated AMPK activation, we assessed pAMPK levels in MDA-MB-231 WT and CAV1 KO cells transfected with siRNA to ROCK1 or ROCK2 (Figure 9A).Knockdown of ROCK1 or ROCK2 increased pAMPK levels selectively in MDA-MB-231 WT cells demonstrating that in these breast cancer cells CAV1/ROCK signaling limits pAMPK activation.To test whether AMPK activation downstream of CAV/ROCK signaling regulated mitochondrial health and mitochondrial ROS, we assessed the impact of double knockdown experiments for AMPK and either ROCK1 or ROCK2 on MV633 and MitoSox labeling.As observed in Figure 9B, AMPK knockdown alone impacted neither mitochondrial potential nor mitochondrial ROS; however, double knockdown of AMPK prevented the increased mitochondrial potential and reduced ROS due to ROCK1 or ROCK2 knockdown (Figure 9B).pCAV1 activation of ROCK in CAV1 scaffolds of cancer cells therefore inhibits AMPK activation, thereby reducing mitophagy resulting in reduced mitochondrial health and increased ROS production in cancer cells (Figure 10).

| DISCUSSION
Tyrosine phosphorylated pCAV1 has well-described signaling and mechanosensory function that activate Rho/ ROCK signaling, focal adhesion turnover and tension, and promote cancer cell invasion and metastasis. 12CAV1 has been shown to regulate both glycolysis and mitochondrial respiration impacting cancer cell metabolism 26,[47][48][49][50] as well as mitochondria fusion-fission and ER-mitochondria contact sites. 51More recently, pCAV1 was shown to inhibit mitochondrial function and promote ROS production in cancer cells via the mitophagy regulators Parkin/PINK and PTP1B dephosphatase known to target pCAV1. 27,31ere, we show that pCAV1 inhibition of mitochondrial activity and promotion of ROS is a direct consequence of pCAV1 inhibition of basal mitophagic flux and occurs downstream of ROCK inhibition of the autophagy regulator AMPK.This study therefore directly links pCAV1 signaling activity at the plasma membrane with its regulation of mitochondrial activity and cancer cell metabolism through control of mitophagy.
The term mitophagy was first used in 2005 and is defined as the selective lysosomal-associated degradation of damaged mitochondria. 52Mitophagy is categorized as either ubiquitin-mediated or receptor-mediated. 34][55] Recently, ubiquitin-mediated basal mitophagy, via the Gp78 ubiquitin ligase, was shown to promote basal mitophagy, increasing mitochondrial health and reducing mitochondrial ROS. 39The PINK1 and Parkin-ubiquitin pathway is one of the most well-defined ubiquitin-mediated mechanisms of mitophagy and has been linked to ROS production in cancer cells and more specifically to CAV1 regulation of mitochondrial health. 56Use of the fluorescent mitoKeima mitophagy probe 44 allowed us to show that CAV1 knockout in two metastatic cancer cell lines increased mitophagic flux and concomitantly mitochondria potential, reducing ROS.This effect was phenocopied by siRNA knockdown of the autophagy protein ATG5, that forms a complex with ATG12 and is essential in autophagosome formation during, 57 of PINK1, that accumulates in the membrane of damaged mitochondrial and recruits Parkin, 58 and of DRP1, a mitochondrial fission protein critical for some forms of ROS-induced mitophagy. 43These studies therefore define a role for basal mitophagic flux in the reduced mitochondrial health of CAV1 expressing cancer cells.CAV1 inhibition of mitophagic flux in cancer cells therefore reduces mitochondrial health and represents a critical determinant of CAV1 regulation of cancer cell metabolism and promotion of ROS production.
We show here that CAV1 inhibition of mitophagic flux in cancer cells is mediated by ROCK activation and inhibition of AMPK signaling.CAV1 knockdown in cancer cells induces Parkin-mediated mitophagy that mediated by pCAV1 and paralleled by ROCK inhibition in lung cancer cells. 27,28As previously reported for pCAV1 regulation of ROCK1-dependent motility and of focal adhesion tension, 4,17 pCAV1 inhibition of mitophagy occurs in PC3 prostate cancer cells lacking CAVIN1 and lacking caveolae.This argues that the functional CAV1 domain mediating downstream ROCK inhibition of mitophagy is a scaffold domain. 59,60The ability of cavin-1 to reduce ROS expression in PC3 cells, that lack cavin-1 and caveolae, argues that scaffolds mediate pCAV1 signaling to ROCK and AMPK, as previously observed for pCAV1 focal adhesion signaling in PC3 cells. 23][23] Collectively, these data argue that ROCK is a critical signaling hub integrating pCAV1 scaffold regulation of cancer cell invasion and metabolism (Figure 10).
ROCK inhibition induces both autophagy and mitophagy. 32,61,62ROCK inhibition has been shown to induce mitophagy by promoting recruitment of HK2, a positive Parkin regulator, to mitochondria. 32Interestingly, the ROCK dependence of metabolic disorders such as hypercholesterolemia, glucose intolerance, and obesity in mice fed high-fat diets is associated with ROCK inactivation of AMPK signaling. 33AMPK is a cellular energy biosensor that recognizes shifts in the AMP:ATP ratio.When ATP is depleted either by impaired ATP synthesis or elevated ATP consumption, AMPK stimulates mitochondrial proliferation via transcriptional regulation of nuclear genes. 34,63Mitochondrial stress can activate AMPK, which in turn activates autophagy-initiating kinase ULK1/2, promoting mitophagy and loss of either AMPK or ULK1 results in aberrant mitophagy. 64,6566,67 The demonstration here that AMPK is required for the induction of mitophagy and mitochondrial potential in CAV1 KO MDA-MB-231 cells argues that ROCK inhibition of AMPK signaling plays a critical downstream signaling role in CAV1 inhibition of mitophagy.pCAV1 signaling via ROCK therefore contributes to cancer progression via both focal adhesion signaling and cancer cell migration/invasion 4,15-17 as well as through inhibition of AMPK and basal mitophagy, resulting in accumulation of damaged mitochondria, reduced mitochondrial respiration, and increased ROS production.

F I G U R E 1
CAV1 KO increases mitochondrial volume and OCR in MDA-MB-231 cells.(A) Wild-type (WT) and CAV1 knockout (CAV1 KO) were fixed and labeled for mitochondrial TOM20 and 3D confocal image stacks were acquired.Representative maximum projections of the image stacks and mitochondrial volume quantified.(Mean ± SEM; n = 3; >40 cells per condition; **p < .01;Scale Bar: 10 μm).(B) The Seahorse XF96 mitochondrial stress test was used to measure oxygen consumption rates (OCR).Oligomycin (oligo) inhibits ATP synthase/complex V and hyperpolarizes mitochondria.Carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) disrupts the proton gradient and allows for the measurement of maximum OCR.Antimycin/rotenone (A/Rot) inhibit complex III/I, respectively, and inhibit all mitochondrial respiration.OCR was measured for MDA-MB-231 WT and CAV1 KO cell lines before and after addition of these mitochondrial inhibitors.Values displayed are mean ± SD.Calculations for basal OCR, ATP-linked OCR, maximum OCR, and reserve capacity are as shown.(Mean ± SEM; 7-8 wells per cell line per treatment; n = 3 **p < .01;****p < .0001).

F I G U R E 4
Cav1 reduces mitochondrial potential and increases ROS in a caveolae-independent manner.(A) PC3 WT and CAV1 KO cells were incubated with 50 nM MV633 (top) or 2.5 μM MitoSOX (bottom) for 30 min at 37°C and imaged at 37°C by confocal microscopy.Quantified mean gray value of mitochondria dyed with MV633 or MitoSOX (left and right, respectively) of PC3 WT and CAV1 KO cells are shown in the bar graphs.(B) Quantified mean gray value of MV633 and MitoSOX (left and right, respectively) of DMSO and Y-27632 treated PC3 WT and CAV1 KO cells.(C) CAV1 KO PC3 cells were transfected with RFP-tagged CAV1 WT, CAV1 Y14D, or CAV1 Y14F and incubated with MV633.Transfected cells are outlined by the white dotted line with the MV633 signal (top) and RFP signal (bottom) shown.Quantified mean gray value of MV633 and CAV1-RFP of PC3 WT, CAV1 KO, and CAV1 mutant-expressing cells is shown in the bar graphs.(Mean ± SEM; 10-15 cells per replicate; n = 3, replicates performed independently; **p < .01;***p < .001;****p < .0001)Scale bar: 10 μm.