p66ShcA functions as a contextual promoter of breast cancer metastasis
The p66ShcA redox protein is the longest isoform of the Shc1 gene and is variably expressed in breast cancers. In response to a variety of stress stimuli, p66ShcA becomes phosphorylated on serine 36, which allows it to translocate from the cytoplasm to the mitochondria where it stimulates the formation of reactive oxygen species (ROS). Conflicting studies suggest both pro- and anti-tumorigenic functions for p66ShcA, which prompted us to examine the contribution of tumor cell-intrinsic functions of p66ShcA during breast cancer metastasis.
We tested whether p66ShcA impacts the lung-metastatic ability of breast cancer cells. Breast cancer cells characteristic of the ErbB2+/luminal (NIC) or basal (4T1) subtypes were engineered to overexpress p66ShcA. In addition, lung-metastatic 4T1 variants (4T1-537) were engineered to lack endogenous p66ShcA via Crispr/Cas9 genomic editing. p66ShcA null cells were then reconstituted with wild-type p66ShcA or a mutant (S36A) that cannot translocate to the mitochondria, thereby lacking the ability to stimulate mitochondrial-dependent ROS production. These cells were tested for their ability to form spontaneous metastases from the primary site or seed and colonize the lung in experimental (tail vein) metastasis assays. These cells were further characterized with respect to their migration rates, focal adhesion dynamics, and resistance to anoikis in vitro. Finally, their ability to survive in circulation and seed the lungs of mice was assessed in vivo.
We show that p66ShcA increases the lung-metastatic potential of breast cancer cells by augmenting their ability to navigate each stage of the metastatic cascade. A non-phosphorylatable p66ShcA-S36A mutant, which cannot translocate to the mitochondria, still potentiated breast cancer cell migration, lung colonization, and growth of secondary lung metastases. However, breast cancer cell survival in the circulation uniquely required an intact p66ShcA S36 phosphorylation site.
This study provides the first evidence that both mitochondrial and non-mitochondrial p66ShcA pools collaborate in breast cancer cells to promote their maximal metastatic fitness.
KeywordsBreast cancer Lung metastasis p66ShcA Reactive oxygen species
The ShcA gene encodes three isoforms (p46, p52, and p66), which together integrate mitogenic and oxidative stress responses to dynamically regulate cell fate decisions (as reviewed in [1, 2, 3, 4]). p46/p52ShcA are encoded from a single transcript and arise through alternate translational start sites . In contrast, p66ShcA is more variably expressed and encoded by its own promoter . ShcA isoforms exert diverse biological functions. Whereas p46/p52ShcA transduce mitogenic signals [4, 5], p66ShcA induces oxidative stress by facilitating mitochondrial-dependent reactive oxygen species (ROS) production .
ShcA isoforms share an amino-terminal phospho-tyrosine-binding (PTB) domain, a carboxy-terminal Src-homology 2 (SH2) domain, and a central collagen-homology 1 (CH1 domain) harboring three tyrosine phosphorylation sites . However, p66ShcA uniquely possesses a CH2 domain at its amino terminus, containing a serine residue (S36) that is essential for its biological function as a redox protein. Phosphorylation of S36 by stress kinases permits binding of the Pin1 prolyl isomerase, facilitating p66ShcA mitochondrial translocation [8, 9]. In the mitochondria, p66ShcA stimulates ROS production by binding to cytochrome c and facilitating the transfer of electrons from cytochrome c to molecular oxygen .
The role of p66ShcA in cancer development is complex and context dependent. Both mitochondrial and non-mitochondrial p66ShcA pools influence cancer progression, and the variability in how p66ShcA influences cancer cells is consistent with the fact that ROS functions as a double-edged sword in cancer [11, 12]. In lung cancer, increased p66ShcA levels are associated with improved patient outcome . Aggressive lung cancers upregulate Aiolos, a lymphocyte-lineage restricted transcription factor that epigenetically silences p66ShcA . In addition, p66ShcA reduced the metastatic potential of lung cancers in mouse models . The tumor-suppressive properties of p66ShcA in lung cancer are associated with several mechanisms. For example, p66ShcA restrains Ras signaling in lung cancer cells by reducing activation of Grb2/SOS signaling complexes [6, 14]. In addition, p66ShcA suppresses an epithelial-to-mesenchymal transition (EMT) in lung cancer cells  and increases anoikis [16, 17].
Paradoxically, p66ShcA largely confers pro-tumorigenic properties in breast, ovarian, and prostate cancers. p66ShcA is overexpressed in each of these cancers compared to benign tissue [18, 19, 20]. In breast cancer, independent studies provide opposing data regarding the relationship between p66ShcA levels and patient outcome. In one study, breast tumors with elevated p66ShcA levels combined with reduced tyrosine phosphorylation of the p46/52 ShcA isoforms were associated with good outcome . However, an independent study showed that p66ShcA is overexpressed in breast cancer cell lines and primary tumors with increasing metastatic properties .
Multiple mechanisms may explain the increased tumorigenic potential associated with p66ShcA in these cancers. For example, p66ShcA overexpression increases the proliferative rate of ovarian and prostate cancers [20, 22]. Moreover, p66ShcA increases the migratory properties of prostate and breast cancer cells [1, 23, 24] by its recruitment to focal adhesion complexes, thereby regulating Rac1-mediated actin remodeling [16, 25]. Furthermore, p66ShcA activates the Arf6 monomeric G protein in breast cancer cells to potentiate Ras signaling . We recently demonstrated that p66ShcA induces an EMT in breast cancer cells . Finally, a unique role for p66ShcA in hypoxia survival and the acquisition of stem-like features has been described in breast cancer cells .
We established a fundamental role for the p46/52ShcA isoforms in breast cancer progression to metastatic disease [28, 29]; however, the precise role of p66ShcA in breast cancer metastasis remains poorly defined. Indeed, p66ShcA plays a complex role in transducing biomechanical signals that promote anchorage-dependent proliferation of cancer cells while paradoxically increasing anoikis following matrix detachment . In this study, we explore the biological significance of p66ShcA during breast cancer metastasis.
Mouse tumor grafts and resections for spontaneous metastasis
All animal studies were approved by the Animal Resources Centre at McGill University and complied with guidelines set by the Canadian Council of Animal Care. Orthotopic injections into the mammary fat pads were performed in BALB/c mice for 4T1-derived cell populations and FVB mice for NIC cell lines. Breast cancer cells were injected at a 1:1 mixture with Matrigel. Tumor size was determined by caliper measurements of two dimensions and volume calculated as follows: 4/3π × width × length2 (smaller measurement is always width). Tumors were resected once they reached 500 mm3.
The lung-metastatic 4T1 derivatives (537 cell population) were injected into the tail veins of BALB/c mice. Mice were examined daily for signs of respiratory distress, and after the indicated time periods, cohorts were euthanized, and lungs subjected to histological and IHC analysis. For lung colonization assays, the indicated cell populations were stained in vitro with 1 μM of Cell Tracker Red CMPTX dye (Cat. #: C34552, Invitrogen) for 45 min in serum-free media. The cells were then washed twice in PBS prior to tail vein injection. The lungs were removed at 1 h or 24 h post-injection, and the left lobe from each animal was whole-mounted for imaging. Images were captured on the Zeiss Axiozoom.V16 microscope at × 50 magnification using 5 fields of view to cover the lung area. The auto-fluorescence in the green channel was used to obtain a topography image of the lung area. Cells were quantified using ImageJ by applying an Otsu threshold and scoring the number of signals with a square pixel size larger than 0.001.
Cells were seeded onto μ-slide 8 well coverslips (Cat. #: 80821, ibidi), at a density of 1500 cells/cm2, approximately 16–24 h prior to imaging. Coverslips were coated with human plasma fibronectin diluted to a concentration of 5 μg/cm2 with 1× PBS for 1 h at 37 °C. Images were acquired in phase contrast every 10 min for 24 h on a Zeiss AxioObserver fully automated inverted microscope equipped with a Plan-Neofluar 10x/0.3NA Ph1 objective, Axiocam 506 camera (Carl Zeiss, Jena, Germany), and Chamlide TC-L-Z003 top-stage incubator system (Live Cell Instrument, Seoul, South Korea). Cells were then manually tracked using MetaXpress analysis software (v. 22.214.171.124; Molecular Devices, Sunnyvale, CA). X, Y position data for each cell track was exported to MATLAB (v. 8.6.0, rel 2015b; The MathWorks, Natick, MA) for further analysis. Rose plots depicting cell movement were created by superimposing the starting position of each cell track on the origin (0, 0). The average speed of each cell was calculated by determining the mean distance traveled between each time point over the imaging interval (10 min). A histogram of cell speed was created in Prism 7 (v. 7.0a; GraphPad Software, La Jolla, CA) by binning data into 5 μm/h intervals ranging from 10 to 70 μm/h. The data was subsequently smoothed using a spline curve and plotted against relative frequency.
Imaging cellular adhesions
Cells were seeded onto 35-mm cover-glass bottom cell culture dishes (Cat. #: FD35-100, World Precision Instruments) and transfected with 1 μg of pmCherry paxillin (Cat. #: 50526, Addgene). Media were changed 18–24 h after transfection, and cells were allowed to recover for 24 h. Images were acquired every 20 s for 20 min on a Total Internal Reflection Fluorescence (TIRF)-Spinning Disk Spectral Diskovery System (Spectral Applied Research, Richmond Hill, ON) based on a Leica DMI 6000 microscope stand (Quorum Technologies, Puslinch, ON) equipped with a Leica Plan-Apochromat 63x/1.47NA oil DIC objective, ImagEM X2 EM-CCD camera (Hamamatsu Photonics K.K., Hamamatsu City, Japan), and Chamlide CU-501 top-stage incubator system (Live Cell Instrument, Seoul, South Korea). Each cell was illuminated with a 561-nm diode laser set to 22.7% (or ~ 74 μW power). An ET600/50m emission filter (Chroma, Bellows Falls, VT) was used to capture mCherry fluorescence. The camera exposure time was set to 500 ms with an EM gain of 255 and read speed of 22 MHz. A total internal reflection fluorescence (TIRF) prism was used to limit fluorescence excitation to a depth of 80 nm.
Measuring adhesion dynamics
Images collected with the TIRF microscope were processed in Imaris (v. 8.3.1; Bitplane, Zurich, Switzerland) using the Surfaces function. Briefly, the leading edge of each cell was manually selected using the region of interest tool. Surface detail was smoothed and set to 0.300 μm with a local background subtraction of 0.300 μm. Adhesions were then masked by adjusting the threshold and splitting touching objects with a seed point diameter of 0.700 μm. Finally, adhesions were tracked over time using an autoregressive algorithm with a max distance of 2 μm and gap size of 3 time points. Surfaces smaller than 10 voxels and tracks shorter than 2 min (or 6 time points) were removed with filters.
Mean intensity data for each adhesion was then exported to MATLAB for further analysis. In the first part of the algorithm, a spline curve was fitted to each intensity trace to identify segments of assembly and disassembly. The difference in intensity between each time point was calculated and changes greater than 20% were considered significant. A string of 5 or more upward points was interpreted as assembly, while 5 or more downward points were interpreted as disassembly. In the second part of the algorithm, a log-linear fitting method was used to determine the rate for each event. Fits with an R2 value greater than 0.7 were significant. Assembly and disassembly rates from three independent experiments were combined to determine the mean for each condition.
Characterizing adhesion size and aspect ratio
A representative image of each analyzed cell for adhesion dynamics was chosen to be re-processed in Imaris. Surface detail was smoothed and set to 0.200 μm with a local background subtraction of 0.200 μm. All adhesions within a cell were masked by adjusting the threshold and splitting touching objects with a seed point diameter of 0.700 μm. Surfaces smaller than 10 voxels were removed with filters. The semi-minor and semi-major axis lengths for each adhesion were then compared in MATLAB to determine aspect ratio. The data shown represents the mean ± SEM for all cells analyzed in three independent experiments. Similarly, area data for each adhesion was exported from Imaris into Prism 7 to determine mean ± SEM for all cells analyzed in three independent experiments. A histogram of adhesion size was generated by binning data into 1-μm intervals ranging from 0 to 8.5 μm. A line plot for p66-CR (VC) was subsequently added to emphasize the change in distribution. For further information regarding experimental procedures, please see the supplemental materials and methods (Additional file 13).
p66ShcA is not sufficient to increase the lung-metastatic potential of breast cancer cells
To assess the impact of p66ShcA on lung metastasis, a separate experiment was performed in which primary tumors were resected at a tumor volume of 500 mm3 and lung-metastatic burden quantified 14 days (4T1) or 21 days (NIC) post-resection. We observed no significant differences in the lung-metastatic burden between mice bearing VC- and p66ShcA-expressing mammary tumors in either the NIC or 4T1 models (Fig. 1e, Additional file 1: Figure S1E). These findings indicate that p66ShcA overexpression is not sufficient to increase breast cancer lung metastasis.
p66ShcA is required for breast cancer lung metastasis
These findings do not preclude the possibility that p66ShcA may be required for dissemination of breast cancer cells that have already acquired increased lung-metastatic potential. We examined endogenous p66ShcA expression levels in cell populations that were in vivo selected to aggressively metastasize to the lung, liver, or bone [32, 33, 34], three common metastatic sites for human breast cancer. Breast cancer cell lines established following in vivo passage through the mammary fat pad served as negative controls. Endogenous p66ShcA levels were increased in lung- and liver-metastatic breast cancer populations relative to explants that were in vivo passaged through the bone or mammary fat pad (Additional file 2: Figure S2A). Moreover, 33% (12/36) of single cell clones derived from parental 4T1 cells exhibit elevated endogenous p66ShcA levels, suggesting that these cell populations pre-existed in the tumor and were specifically enriched in breast cancer cells that metastasized to the lung or liver (Additional file 2: Figure S2B). This is consistent with our interrogation of publically available datasets showing that metastases isolated from breast cancer patients displayed significantly higher p66ShcA mRNA levels compared to primary breast tumors (Additional file 2: Figure S2C). Therefore, we explored whether p66ShcA was necessary for the increased formation of breast cancer lung metastases by stably deleting p66ShcA from a lung-metastatic 4T1 variant (4T1-537) using CRISPR/Cas9 approaches. Individual p66ShcA null clones were pooled to generate a polyclonal cell population (Additional file 2: Figure S2D, E).
Non-mitochondrial p66ShcA pools potentiate the later stages of the metastatic cascade
To better define the steps of the metastatic cascade that are influenced by p66ShcA, breast cancer cells were injected directly into bloodstream via the lateral tail vein. An initial experiment comparing parental 537 cells with p66-CR (VC) cells employed a single endpoint (21 days) at which time mice were analyzed for lung-metastatic burden. Loss of p66ShcA [p66-CR (VC)] negatively impacted the metastatic ability of breast tumors, as revealed by a 10-fold reduction in the lung-metastatic burden relative to parental cells (Fig. 3d). Given the aggressive nature of 4T1-537 cells, we wished to distinguish between effects due to inefficient lung colonization versus impaired outgrowth to form macroscopic metastases. To test this, we performed a second experimental metastasis assay whereby mice injected with parental 4T1-537 cells were sacrificed early (21 days) while mice injected with p66-CR (VC), p66-CR (WT), or p66-CR (S36A) cells were followed for longer prior to sacrifice (26 days). Again, we observed a reduced metastatic burden (~ 2.2-fold) in mice injected with p66-CR (VC) cells compared to 4T1-537 controls (Fig. 3e). Surprisingly, both p66CR (WT)- and p66-CR (S36A)-expressing cells rescued the metastatic ability of p66ShcA-null cells (Fig. 3e). Combined, these data suggest that p66ShcA is essential for the early (prior to intravasation) and late (following intravasation) stages in the metastatic cascade. Our data further suggests that mitochondrial-localized p66ShcA is required either before, during, or immediately following intravasation, whereas non-mitochondrial p66ShcA pools potentiate extravasation, colonization, and/or secondary growth of breast cancer lung metastases.
p66ShcA supports efficient breast cancer lung metastasis independently of an EMT
We examined whether the ability of p66ShcA to induce an EMT was associated with an increased lung-metastatic potential. Consistent with our published studies , p66ShcA increases Vimentin expression in ErbB2+ luminal breast tumors (NIC) whereas E-Cadherin levels are largely unaffected compared to control tumors (Additional file 7: Figure S7). Despite this fact, we observed no differences in the lung-metastatic potential of VC- or p66ShcA-expressing NIC tumors (Additional file 1: Figure S1E). Moreover, whereas p66ShcA-S36A expressing tumors were significantly debilitated in their lung-metastatic potential, they showed a profound increase in Vimentin levels and a corresponding reduction in E-Cadherin expression (Additional file 7: Figure S7). These data suggest that although non-mitochondrial p66ShcA pools increase the mesenchymal properties of luminal breast cancers, the ability of p66ShcA to induce an EMT is not enough to promote breast cancer lung metastasis. We also examined E-Cadherin and Vimentin levels in 4T1-537 parental, p66-CR (VC), p66-CR (WT), and p66-CR (S36A) mammary tumors, which require p66ShcA for spontaneous breast cancer lung metastasis. We do not observe appreciable differences in Vimentin or E-Cadherin levels in any of these tumors (Additional file 5: Figure S5A, D, E). Thus, p66ShcA supports breast cancer lung metastasis in this TNBC model, even in the absence of an EMT.
Non-mitochondrial p66ShcA pools enhance breast cancer cell migration by promoting adhesion dynamics
p66ShcA is dispensable for the formation of functional invadopodia and ECM degradation
Finally, we assessed whether p66ShcA altered the proteolytic property of breast cancer cells by plating them onto fluorescently-labeled gelatin and measuring the degree of gelatin degradation. Surprisingly, we found that p66-CR (VC) cells showed the greatest increase in gelatin degradation (Additional file 8: Figure S8A), even though they displayed the lowest metastatic potential (Fig. 3b, c). However, both p66-CR (WT)- and p66-CR (S36A)-expressing cells displayed a similar degree of gelatin degradation relative to parental controls (Additional file 8: Figure S8A). Examination of the degraded surface area in these studies revealed two distinct degradation patterns: (1) “punctate” areas of gelatin degradation suggestive of individual invadopodia and (2) larger “patches” of gelatin degradation that may arise due to sustained invadopodia formation in cells that are less migratory. Indeed, most of the degradation patterns observed with p66-CR (VC) cells (80%) resembled these larger degradation patterns (Additional file 8: Figure S8B), which is consistent with their reduced migratory speed (Fig. 4), larger focal adhesion structures (Fig. 5c, d), and impaired rates of adhesion assembly and disassembly (Fig. 5e). In contrast, 4T1-537 (Par)-, p66-CR (WT)-, and p66-CR (S36A)-expressing cells form a higher frequency of more “punctate” degradation patterns (Additional file 8: Figure S8B), consistent with their increased migratory properties and adhesion dynamics (Figs. 4 and 5). Finally, we show that the degree of gelatin degradation was comparable across all groups when specifically analyzing areas with punctate degradation patterns, resembling invadopodia (Additional file 8: Figure S8C). These data suggest that neither p66ShcA loss [p66-CR (VC)], nor expression of p66ShcAWT or p66ShcAS36A alleles, appreciably altered the degradative properties of breast cancer cells. However, p66-CR (VC) cells degraded a significantly larger surface area of gelatin, compared to p66ShcAWT and p66ShcAS36A-expressors, when specifically examining larger patches of degradation (Additional file 8: Figure S8D). These data argue the increased degradative behavior of p66ShcA-null cells may be an indirect effect of their reduced cell migration and increased adhesion to the extracellular matrix, resulting in the formation of larger and more stable degradative structures.
Distinct p66ShcA pools support breast cancer cell survival in circulation and subsequent lung colonization
Following intravasation into the bloodstream, breast tumor cells must successfully extravasate and seed the secondary organ to support their metastatic dissemination. To test this, we enumerated the ability of fluorescently labeled cancer cells to colonize the lung immediately following tail vein injection. The lungs of mice injected with parental, p66-CR (VC), p66-CR (WT), and p66-CR (S36A) breast cancer cells were examined by whole-mount fluorescent microscopy either 1 h or 24 h following tail vein injection. At 1 h post-injection, the number of cancer cells in the lungs of mice from each cohort was similar, although there were slightly fewer cells in the lungs of mice bearing p66ShcA (S36A)-expressing cells (Fig. 6c). By comparing the ratio of fluorescently labeled cells remaining in the lung 24 h post-injection, relative to the 1-h time point, we calculated the ability of breast cancer cells to colonize and survive in the lung immediately following extravasation. Loss of p66ShcA significantly reduced the proportion of surviving cells relative to p66ShcA-proficient controls (Fig. 6d). Moreover, both wild-type p66ShcA and p66ShcA (S36A) alleles rescued lung colonization in p66ShcA-null cells (Fig. 6d). These data suggest that non-mitochondrial p66ShcA pools promote efficient breast cancer cell lung colonization.
Non-mitochondrial p66ShcA pools increase the growth potential of breast cancer lung metastases
Given that p66ShcA is a redox protein and the lung microenvironment has a high oxidative potential, we also assessed the degree of oxidative damage along with induction of an apoptotic response. We show that wild-type p66ShcA was dispensable for controlling breast cancer cell survival, both within the lung metastases (Fig. 7b) and primary tumors (Additional file 6: Figure S6). Similarly, loss of p66ShcA did not change the degree of lipid peroxidation (4HNE) in either the primary tumor (Additional file 9: Figure S9) or lung metastases (Fig. 7c). Re-expression of the non-phosphorylatable form of p66ShcA (S36A) modestly reduced the degree of lipid peroxidation in breast cancer lung metastases but had no appreciable impact on their apoptotic rate (Fig. 7b, c).
Increased AKT/mTOR and Src family kinase signaling is associated with the pro-metastatic properties of non-mitochondrial p66ShcA pools
Overall, we demonstrate an essential role for p66ShcA in increasing the lung-metastatic potential of aggressive TNBC cells. Our data using a p66ShcA mutant that cannot translocate to the mitochondria reveal that this adaptor protein potentiates several steps of the metastatic cascade independently of its ability to induce mitochondrial ROS formation. This includes increased cell migration, lung colonization, and growth of secondary metastases. However, our data suggests a unique role for mitochondrial p66ShcA in facilitating the intravasation into and/or survival of circulating tumor cells within the bloodstream. In order to elucidate potential mechanisms that contribute to p66ShcA-induced lung metastasis, we assessed the status of several signaling pathways that have previously been implicated in this process. In order to model stress responses that breast cancer cells are likely to encounter in vivo, we cultured the 4T1-537 cells that differ in their p66ShcA status (p66ShcA-null, p66ShcA-WT, or p66ShcA-S36A) under conditions that alter nutrient availability (10% vs 0% FBS), anchorage dependency (adherent versus suspension), and oxidative stress (phenformin, arsenic).
We first focused on signaling pathways controlling energy balance, given the central requirement for increased bioenergetics in promoting breast cancer lung metastasis [37, 38]. Indeed, p66ShcA has been shown to support catabolic metabolism by favoring a shift towards oxidative phosphorylation . We therefore examined whether modulating p66ShcA function altered AMPK activation, a ser/thr kinase that is a central regulator of metabolic reprogramming towards catabolic reactions . Interestingly, we observed increased AMPK phosphorylation, specifically in p66ShcA-S36A-expressing cells, under nutrient replete conditions (Additional file 10: Figure S10A). However, in response to nutrient deprivation, cell detachment, or oxidative stress conditions, AMPK is activated in a p66ShcA-independent manner within all cell lines (Additional file 10: Figure S10A). Moreover, we do not observe appreciable pAMPK levels in any of the mammary tumors or lung metastases that we analyzed (Additional file 10: Figure S10B). These data suggest that although non-mitochondrial p66ShcA pools may regulate AMPK to control cellular bioenergetics under normal growth conditions, deregulation of this pathway is not likely to contribute to the increased ability of p66ShcA to promote breast cancer lung metastasis.
We next examined whether Src family kinase (SFK) signaling was altered in the various breast cancer cell lines that differ in their lung-metastatic potential, given their multi-faceted role in controlling several stages of the metastatic cascade . We show that p66ShcA-S36A-expressing cells specifically exhibit increased tyrosine phosphorylation of SFKs in vitro (Additional file 11: Figure S11A) and in mammary tumors in vivo (Additional file 11: Figure S11B). However, SFK activation is further potentiated in response to nutrient deprivation or oxidative stress, but in a p66ShcA-independent manner. Moreover, SFK Y416-phosphorylation is severely reduced in all cell lines examined following matrix detachment (Additional file 11: Figure S11A). Finally, we also do not observe significant differences in pSFK levels in any of the lung metastases irrespective of p66ShcA status (Additional file 11: Figure S11B). Taken together, these observations suggest that the elevated SFK activation, increased cell migration and focal adhesion dynamics mediated by non-mitochondrial p66ShcA pools (Figs. 4 and 5) may contribute to the ability of p66ShcA to potentiate the early stages of the metastatic cascade (Fig. 3).
Breast cancers are lethal primarily due to metastatic progression. We now show that the p66ShcA adaptor protein, which is variably expressed in breast cancers, enhances the metastatic potential of breast cancers. Whereas p66ShcA overexpression is not sufficient to increase breast cancer lung metastasis in two independent models, we show that breast cancer cells with increased lung-metastatic potential upregulate endogenous p66ShcA levels and that p66ShcA is necessary for these cells to retain their lung-metastatic fitness. These data support the idea that p66ShcA is one of many pro-metastatic genes that may be coordinately upregulated and collaboratively increase the lung-metastatic potential of breast cancers. Identification of such coordinately regulated genes is beyond the scope of this study but warrant further investigation.
Even though p66ShcA induces an EMT in NIC tumors, which are models for luminal breast cancer , our data suggest that a p66ShcA-induced EMT is not sufficient to further promote lung metastasis of NIC tumors. Moreover, we do not observe differences in expression of epithelial or mesenchymal markers in 4T1 tumors, which is a model for triple-negative breast cancer. Despite this fact, p66ShcA is required to support the lung-metastatic potential of 4T1 tumors. These data are consistent with independent studies demonstrating that an EMT may be dispensable for the ability of pancreatic and breast cancer cells to metastasize to the lung [52, 53]. Nevertheless, these observations do not rule out the possibility that breast cancer cells undergoing an EMT can be pro-metastatic. For example, following an EMT, cancer cells can act in paracrine to increase the metastatic properties of neighboring cells . Our data suggests that a causal relationship between EMT and breast cancer lung metastasis is likely context dependent.
To define the mechanisms by which p66ShcA enhances metastasis, we therefore decided to examine the role of p66ShcA in breast cancer cell motility, partly guided by previous research showing that it can localize to focal adhesions and mediate signaling [14, 16, 17]. Our results demonstrate that p66ShcA loss severely reduces breast cancer motility in vitro due to the formation of larger and more stable cellular adhesions. It is well documented that preventing FAK-mediated adhesion turnover results in larger adhesions with reduced cell speed [55, 56]. Recent research also demonstrates that newly forming adhesions have smaller areas and more circular shapes . These data suggest that the reduced speed of p66ShcA-deficient breast cancer cells is a consequence of mature adhesion formation, which impairs cell migration. Indeed, the assembly and disassembly rates of adhesions in p66ShcA-deficient breast cancer are significantly slower than cells expressing wild-type p66ShcA. Furthermore, p66ShcA (S36A)-expressing cells migrate even faster than their wild-type counterparts, which is consistent with increased Src activation in these cells. Non-mitochondrial p66ShcA pools therefore function to increase cell motility by enhancing adhesion dynamics, potentially through the ability of Src family kinases to modulate actin dynamics. This conclusion is supported by evidence that p66ShcA mediates FAK signaling through RhoA [16, 17].
We also found that non-mitochondrial p66ShcA improved the lung colonization ability of breast cancer cells and subsequently increased their proliferative potential to form overt metastases. One possibility is that p66ShcA may also be mediating integrin signaling at the metastatic site, as it has been found that p66ShcA can facilitate proliferative signaling in adherent cells through RhoA and the YAP/TAZ transcription factors . p66ShcA may also regulate signaling through other interactions. For example, p66ShcA has been reported to activate Rac1 , which has important roles in promoting breast cancer cell motility, invasion, and survival .
Not all pro-metastatic functions of p66ShcA can be attributed to the non-mitochondrial pool of p66ShcA. Tumors expressing a p66ShcA mutant that cannot translocate to the mitochondria  revealed a significantly reduced lung-metastatic burden concomitant with a decreased number of circulating tumor cells and an impaired ability to survive following ECM detachment. The involvement of p66ShcA in anoikis has been well documented for lung cancers, in which p66ShcA is often repressed or silenced [13, 14]. Previous studies have shown that upon recruitment to cellular adhesions, p66ShcA/FAK complexes promote RhoA activation by recruiting Rho-specific GEFs, including GEF-H1 and p115-RhoGEF . In turn, increased RhoA activity initiates anoikis following matrix detachment . Whereas serine 36 phosphorylation of p66ShcA is essential for mitochondrial ROS production, opening of the permeability transition pore and apoptosis [7, 10], paradoxically, p66ShcA serine 36 phosphorylation is protective against anoikis [14, 16]. Indeed, a phospho-mimetic p66ShcA (S36E) mutant was previously shown to promote anoikis resistance in lung cancer cells, thereby promoting their ability to colonize and grow in the lung parenchyma . This is consistent with our observations, which reveal that loss of p66ShcA serine 36 phosphorylation in metastatic breast cancer cells abrogates their ability to survive in circulation and sensitizes them to anoikis. While serine 36 phosphorylation of p66ShcA is likely to be protective of anoikis upon matrix detachment, the precise mechanisms controlling this phenotype remain poorly defined and warrant further investigation.
The dichotomous roles of p66ShcA indicate that this protein has a range of functions. While ShcA acts as a metastasis suppressor in lung cancer, it is a metastasis promoter in breast cancer. The impact of such functions on cancer cell phenotypes is likely to result from a combination of internal and external factors that cooperatively influence which role p66ShcA plays in cancer progression. Future research to better understand how distinct interacting molecules govern the function of p66ShcA, in a context-dependent manner, may lead to the identification of more suitable prognostic indicators or novel therapeutic targets for breast cancer metastases.
We identify the p66ShcA redox protein as novel mediator of breast cancer lung metastasis. We show that p66ShcA is not sufficient to improve the metastatic capabilities of models representing triple-negative or luminal breast cancer. However, breast cancer cells that have already acquired increased lung- and liver-metastatic potential upregulated endogenous p66ShcA levels and require p66ShcA to promote efficient lung metastasis. These data suggest that p66ShcA acts collectively with a group of pro-metastatic genes, representing cooperative adaptations selected for in breast cancer cells with an increased propensity to metastasize. Moreover, p66ShcA re-expression increased breast cancer metastasis both in spontaneous and experimental metastasis assays, suggesting that p66ShcA acts upon multiple stages of the metastatic cascade. Interestingly, however, tumors expressing a p66ShcA mutant that cannot localize to the mitochondria had reduced spontaneous metastatic burden but retained the ability to seed and colonize the lung following intravasation.
We thank The George and Olga Minarik Research Pathology Facility (Lady Davis Institute) and the McGill Histology Core Facility (Goodman Cancer Research Centre) for preparation of tissue for histological analysis. Imaging experiments from the author’s laboratories were performed at the McGill Life Sciences Complex Advanced BioImaging Facility (ABIF).
JH, KL, AK, JS, JH, EV, MA, VS, RA, RL, ST, MS, ML, ECC, and MD performed the experiments. CMB and IRW provided expertise in method development. PS and JUS conceived the experiments and wrote the manuscript. All authors edited the manuscript. All authors approved the final manuscript.
J.H., A.K., R.A., and R. L. acknowledge support in the form of an FRQS doctoral studentship. JRH was funded by a CIHR doctoral award. PMS is a McGill University William Dawson Scholar, and JUS is a Senior Research Scholar of the FRQS. This research was supported by an i2I grant to PMS and CMB from the Canadian Cancer Society (Grant #: 705838) and a project grant to JUS from the Canadian Institutes of Health Research (Grant #: MOP-133670).
Ethics approval and consent to participate
The Lady Davis Institute Animal Care Committee approved all the animal studies. No additional ethical approvals or consents were required.
Consent for publication
We confirm that this is original work that is not being considered for publication elsewhere and consent to its publication in Breast Cancer Research.
The authors declare that they have no competing interests.
- 6.Migliaccio E, Mele S, Salcini AE, Pelicci G, Lai KM, Superti-Furga G, Pawson T, Di Fiore PP, Lanfrancone L, Pelicci PG. Opposite effects of the p52shc/p46shc and p66shc splicing isoforms on the EGF receptor-MAP kinase-fos signalling pathway. EMBO J. 1997;16(4):706–16.PubMedPubMedCentralCrossRefGoogle Scholar
- 9.Pinton P, Rimessi A, Marchi S, Orsini F, Migliaccio E, Giorgio M, Contursi C, Minucci S, Mantovani F, Wieckowski MR, et al. Protein kinase C beta and prolyl isomerase 1 regulate mitochondrial effects of the life-span determinant p66Shc. Science. 2007;315(5812):659–63.PubMedCrossRefPubMedCentralGoogle Scholar
- 10.Giorgio M, Migliaccio E, Orsini F, Paolucci D, Moroni M, Contursi C, Pelliccia G, Luzi L, Minucci S, Marcaccio M, et al. Electron transfer between cytochrome c and p66Shc generates reactive oxygen species that trigger mitochondrial apoptosis. Cell. 2005;122(2):221–33.PubMedCrossRefPubMedCentralGoogle Scholar
- 17.Wu RF, Liao C, Fu G, Hayenga HN, Yang K, Ma Z, Liu Z, Terada LS. p66Shc couples mechanical signals to RhoA through FAK-dependent recruitment of p115-RhoGEF and GEF-H1. Mol Cell Biol. 2016;36(22):2824–37.Google Scholar
- 27.Sansone P, Storci G, Giovannini C, Pandolfi S, Pianetti S, Taffurelli M, Santini D, Ceccarelli C, Chieco P, Bonafe M. p66Shc/Notch-3 interplay controls self-renewal and hypoxia survival in human stem/progenitor cells of the mammary gland expanded in vitro as mammospheres. Stem Cells. 2007;25(3):807–15.PubMedCrossRefPubMedCentralGoogle Scholar
- 28.Northey JJ, Dong Z, Ngan E, Kaplan A, Hardy WR, Pawson T, Siegel PM. Distinct phosphotyrosine-dependent functions of the ShcA adaptor protein are required for transforming growth factor beta (TGFbeta)-induced breast cancer cell migration, invasion, and metastasis. J Biol Chem. 2013;288(7):5210–22.PubMedCrossRefPubMedCentralGoogle Scholar
- 34.Tabaries S, Dong Z, Annis MG, Omeroglu A, Pepin F, Ouellet V, Russo C, Hassanain M, Metrakos P, Diaz Z, et al. Claudin-2 is selectively enriched in and promotes the formation of breast cancer liver metastases through engagement of integrin complexes. Oncogene. 2011;30(11):1318–28.PubMedCrossRefGoogle Scholar
- 35.Orsini F, Migliaccio E, Moroni M, Contursi C, Raker VA, Piccini D, Martin-Padura I, Pelliccia G, Trinei M, Bono M, et al. The life span determinant p66Shc localizes to mitochondria where it associates with mitochondrial heat shock protein 70 and regulates trans-membrane potential. J Biol Chem. 2004;279(24):25689–95.PubMedCrossRefPubMedCentralGoogle Scholar
- 37.Andrzejewski S, Klimcakova E, Johnson RM, Tabaries S, Annis MG, McGuirk S, Northey JJ, Chenard V, Sriram U, Papadopoli DJ, et al. PGC-1alpha promotes breast cancer metastasis and confers bioenergetic flexibility against metabolic drugs. Cell Metab. 2017;26(5):778–87 e775.PubMedCrossRefPubMedCentralGoogle Scholar
- 47.Finetti F, Pellegrini M, Ulivieri C, Savino MT, Paccagnini E, Ginanneschi C, Lanfrancone L, Pelicci PG, Baldari CT. The proapoptotic and antimitogenic protein p66SHC acts as a negative regulator of lymphocyte activation and autoimmunity. Blood. 2008;111(10):5017–27.PubMedCrossRefPubMedCentralGoogle Scholar
- 50.Le Gal K, Ibrahim MX, Wiel C, Sayin VI, Akula MK, Karlsson C, Dalin MG, Akyurek LM, Lindahl P, Nilsson J, et al. Antioxidants can increase melanoma metastasis in mice. Sci Transl Med. 2015;7(308):308re308.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.