The Cytoplasmic Region of SARAF Reduces Triple-Negative Breast Cancer Metastasis through the Regulation of Store-Operated Calcium Entry

Triple-negative breast cancer has a poor prognosis and is non-responsive to first-line therapies; hence, new therapeutic strategies are needed. Enhanced store-operated Ca2+ entry (SOCE) has been widely described as a contributing factor to tumorigenic behavior in several tumor types, particularly in breast cancer cells. SOCE-associated regulatory factor (SARAF) acts as an inhibitor of the SOCE response and, therefore, can be a potential antitumor factor. Herein, we generated a C-terminal SARAF fragment to evaluate the effect of overexpression of this peptide on the malignancy of triple-negative breast cancer cell lines. Using both in vitro and in vivo approaches, we showed that overexpression of the C-terminal SARAF fragment reduced proliferation, cell migration, and the invasion of murine and human breast cancer cells by decreasing the SOCE response. Our data suggest that regulating the activity of the SOCE response via SARAF activity might constitute the basis for further alternative therapeutic strategies for triple-negative breast cancer.


Introduction
Breast cancer is the most common type of cancer and the leading cause of death in women worldwide [1,2]. It is characterized as a heterogeneous disease at the molecular level, and both prognosis and treatment depend on a subclassification determined by the histological and molecular features of the tumor [3]. The triple-negative breast cancer (TNBC) subtype represents approximately 15% of the total breast cancers and is characterized by an aggressive subtype that has a 40% recurrence rate owing to metastatic spread [4]. Since TNBC tumors lack estrogen and progesterone receptors, and HER2 expression, they are not susceptible to first-line therapies, such as hormone-based therapy. Hence, there is an urgent need to develop new therapeutic strategies to overcome the poor prognosis of this cancer subtype.
Ca 2+ signaling is deregulated in several types of cancers [5,6]. In different types of cancers including breast cancer, changes in the expression and activity of entities, such as ion channels, transporters, and their regulators [7,8] have been proposed as tumorigenic events that lead to malignancy. Altered Ca 2+ homeostasis increases cell motility and invasion, proliferation, apoptosis resistance, and changes in gene transcription (as reviewed by Monteith et al. (2017) [6]). The store-operated Ca 2+ entry (SOCE) response is a ubiquitous and essential cellular homeostatic mechanism that regulates Ca 2+ levels in both excitable and non-excitable cells. This response is mainly mediated by Orai1, a highly selective Ca 2+ channel located in the plasma membrane (PM). Its activity relies on its interaction with STIM1, a type I transmembrane protein located mainly in the endoplasmic reticulum (ER). In the absence of Ca 2+ in the ER lumen, STIM1 leads a clustering process that allows the formation of the Orai1-STIM1 complex at the ER-PM junctions, activating Orai1 channels. The SOCE response is relevant in different types of cancer such as melanoma [9][10][11], glioblastoma [12], clear cell renal carcinoma [13], hepatocellular carcinoma [14], prostate cancer [15], cervical cancer [16] and breast cancer [17][18][19][20]. These types of cancer exhibit an augmented SOCE response and overexpression of STIM1 and/or Orai1. The inhibition of SOCE decreases the proliferation and metastasis of cancer cells [21]. Similarly, increased Orai1-STIM1 activity enhances the formation of podosomes and invadopodia, thus contributing to the invasive phenotype of breast cancer cells [22,23]. These reports suggest a robust and novel role of Orai1-STIM1 activity, characterized by an increased expression of Orai1-STIM1 which contributes to cancer progression and metastasis. Thus, understanding the molecular mechanisms that regulate Orai1-STIM1 activity could help in the development of new therapeutic strategies to treat TNBC, particularly by taking advantage of the upregulated SOCE response.
As such, the SARAF-STIM1 interaction constitutes the main inhibitory molecular checkpoint for SOCE regulation. It is proposed that SARAF plays a crucial role in the development and progression of breast cancer, although no significant differences have been observed in SARAF protein levels between pre-neoplastic and neoplastic breast cancer cell lines [33]. The silencing of the SARAF protein improves SOCE component expression in pre-neoplastic and TNBC cells. At the same time, in the luminal subtypes, the overexpression of SARAF decreases the expression of SOCE components. Furthermore, SARAF silencing reduced proliferation and migration in MCF7 and MDA-MB-231 cells, whereas MCF10A cells were unaffected, suggesting its cell-type-dependent differential role [33]. Therefore, the identification of novel modulators of the SARAF-STIM1 interaction and their impact on Orai1-STIM1 function is an interesting biomedical research topic. Currently, the lack of drug specificity remains a challenge, despite the attempts to find selective pharmacological modulators of Orai1-STIM1 in different models [34]. It is known that the cytoplasmic C-terminal region of SARAF protein is sufficient to exert its activity as negative regulator of the SOCE response [30]. Here, we determined the effect of the cytoplasmic C-terminal region of SARAF on the SOCE response in breast cancer and in the critical cellular events leading to the malignant phenotype, such as tumor cell proliferation, motility, and metastasis [35]. Our data suggest that regulating the activity of the SOCE response via SARAF activity might constitute the basis for further alternative therapeutic strategies for triple-negative breast cancer.

SARAF Expression Is Reduced in TNBC
To evaluate whether canonical SOCE components including the Orai1, STIM1, and SARAF proteins, are correlated with breast cancer metastatic potential, we consulted The Cancer Genome Atlas Program (TCGA) NIH database using the GEPIA platform [36]. Results indicate that no statistically significant changes in Orai1, STIM1, and SARAF expression were observed in tumorous and non-tumorous tissues ( Figure 1A-C). SARAF expression is diminished in the triple-negative subtype compared to the Luminal A subtype of tumors ( Figure 1C). The Luminal A subtype manifests less metastatic potential and thus presents a better prognosis than TNBC. In addition, we evaluated the endogenous protein levels of the SOCE components in different breast cancer cell lines ( Figure 1D,E). Relative quantification indicated that the expression of the components depended on the aggressiveness of the cell line ( Figure 1E). As the MCF10A line is a non-neoplastic cell line, we observed a decrease in STIM1 expression in MCF7 and MDA-MB-231 cells. No differences were observed in Orai1 expression in MDA-MB231 cells. However, we observed a diminished expression of the Orai1 protein in T47D and MCF7. Notably, SARAF expression diminished in all malignant cell lines. Therefore, since triple-negative tumors have high metastatic potential compared to other tumors, decreased SARAF expression can be correlated to a higher migration and invasive tumor phenotype. It also suggests that the overexpression of SARAF protein could be a potential strategy to counteract the highly invasive and proliferative behavior of TNBC cell lines.

SARAF Expression Is Reduced in TNBC
To evaluate whether canonical SOCE components including the Orai1, STIM1, and SARAF proteins, are correlated with breast cancer metastatic potential, we consulted The Cancer Genome Atlas Program (TCGA) NIH database using the GEPIA platform [36]. Results indicate that no statistically significant changes in Orai1, STIM1, and SARAF expression were observed in tumorous and non-tumorous tissues ( Figure 1A-C). SARAF expression is diminished in the triple-negative subtype compared to the Luminal A subtype of tumors ( Figure 1C). The Luminal A subtype manifests less metastatic potential and thus presents a better prognosis than TNBC. In addition, we evaluated the endogenous protein levels of the SOCE components in different breast cancer cell lines ( Figure 1D,E). Relative quantification indicated that the expression of the components depended on the aggressiveness of the cell line ( Figure 1E). As the MCF10A line is a non-neoplastic cell line, we observed a decrease in STIM1 expression in MCF7 and MDA-MB-231 cells. No differences were observed in Orai1 expression in MDA-MB231 cells. However, we observed a diminished expression of the Orai1 protein in T47D and MCF7. Notably, SARAF expression diminished in all malignant cell lines. Therefore, since triple-negative tumors have high metastatic potential compared to other tumors, decreased SARAF expression can be correlated to a higher migration and invasive tumor phenotype. It also suggests that the overexpression of SARAF protein could be a potential strategy to counteract the highly invasive and proliferative behavior of TNBC cell lines.

The C-Terminal SARAF Fragments Reduce Store-Operated Calcium Entry
It has previously been reported that the C-terminal region of SARAF is sufficient to decrease the SOCE response [30]. As a decrease in SOCE response affects malignancy of breast cancer cells, we generated a C-terminal fragment of SARAF fused to mScarlet fluorescent protein (mScarlet-C-SARAF) ( Figure 2A) and evaluated the effect of its overexpression on the SOCE response. Our initial approach was to determine the SOCE response via Ca 2+ imaging, using Fura-2-AM in HEK293 cells. We observed that the overexpression of mScarlet-C-SARAF decreased the SOCE response without affecting thapsigargin-induced Ca 2+ release ( Figure 2B-D). Moreover, we found that mScarlet-C-SARAF interacts with STIM1 ( Figure 2E). Once the effect of this construct was validated in HEK293 cells, we performed Ca 2+ imaging by overexpressing mScarlet or mScarlet-C-SARAF in the human TNBC cell line MDA-MB-231. We observed that the overexpression of mScarlet-C-SARAF resulted in similar results to those obtained in HEK293 cells ( Figure 2F-I). We also evaluated the effect of the overexpression of the mScarlet-C-SARAF fragment on the formation of the STIM1-Orai1 complex in MDA-MB-231 cells. For this purpose, we performed a bimolecular fluorescence complementation (BiFC) approach [37,38], using STIM1 and Orai1 constructs, which are fused to either non-fluorescent N or C terminal fragments of the Venus protein [39] ( Figure 2J,K). We observed that mScarlet-C-SARAF overexpression decreased puncta formation upon thapsigargin stimulation, indicating that the interaction between STIM1 and Orai1 is reduced. This also correlates with a diminished nuclear translocation of NFAT, indicating that a diminished SOCE response elicited by mScarlet-C-SARAF expression affects the signaling associated with this response ( Figure 2L,M). These results together indicate that mScarlet-C-SARAF overexpression reduces the SOCE response.

The C-Terminal SARAF Fragment Reduces Malignancy Features in Human TNBC
The SOCE response regulates proliferation, apoptosis resistance, migration, and invasion in different types of cancer [21]. We evaluated the effect of C-SARAF overexpression in MDA-MB-231 cells using MTT cell proliferation assays. The results indicate that the overexpression of the mScarlet-C-SARAF fragment had no effect on cell proliferation compared to the control between 16 and 48 h ( Figure 3A). Scratch assay results indicate ( Figure 3B) a decrease in cell migration when overexpressing the mScarlet-C-SARAF fragment compared to the control condition (mScarlet overexpression) ( Figure 3B,C). We also performed Transwell ® invasion assays in MDA-MB-231 cells to determine if the overexpression of the mScarlet-C-SARAF fragment was sufficient to reduce cell invasion. We observed that mScarlet-C-SARAF-overexpression reduced cell invasion, confirming that SOCE inhibition by the C-terminal fragment of SARAF could reduce the metastatic potential of TNBC cells ( Figure 3D,E). The above results suggest that the mScarlet-C-SARAF fragment regulates the SOCE response, reduces Ca 2+ entry, and modulates the migratory and invasive potential of TNBC cell lines, but not their proliferation.
performed Transwell ® invasion assays in MDA-MB-231 cells to determine if the overexpression of the mScarlet-C-SARAF fragment was sufficient to reduce cell invasion. We observed that mScarlet-C-SARAF-overexpression reduced cell invasion, confirming that SOCE inhibition by the C-terminal fragment of SARAF could reduce the metastatic potential of TNBC cells ( Figure 3D,E). The above results suggest that the mScarlet-C-SARAF fragment regulates the SOCE response, reduces Ca 2+ entry, and modulates the migratory and invasive potential of TNBC cell lines, but not their proliferation.

The C-Terminal SARAF Fragment Affects Dynamic Focal Adhesion and Cell Retraction
Since mScarlet-C-SARAF reduces the migratory and invasive potential of the human TNBC cell line, MDA-MB-231, we evaluated if the overexpression of cytoplasmic C-terminal SARAF could affect cellular processes, such as focal adhesion (FA) dynamics and cytoskeleton rearrangement, both involved in cell migration and invasion. To explore these processes and to determine the adhesion capacity of cells overexpressing C-SARAF, we performed cell spreading assays. We observed that mScarlet-C-SARAF reduced the area of MDA-MB-231 cells after 45 min of seeding ( Figure 3F-H).
FA dynamics is a cellular process involving coordination between FA turnover and actin cytoskeleton rearrangement, and it is the key for cell migration [40]. To determine if C-SARAF affects focal adhesion turnover, we co-transfected MDA-MB-231 cells with EGFP-paxillin and mScarlet or mScarlet-C-SARAF and analyzed the FA dynamics using TIRF microscopy ( Figure 4A). The results suggest that the mScarlet-C-SARAF fragment increased the assembly rate ( Figure 4B) without affecting the disassembly rate ( Figure 4B-E). In addition, we observed a decrease in cell retraction ( Figure 4E,F), indicating that the overexpression of C-SARAF impacts focal adhesion dynamics and cell motility. A decrease in cell movement is likely due to a decrease in cell retraction upon serum stimulation. In summary, mScarlet-C-SARAF overexpression increased the assembly rate of FA and reduced cell retraction.
TIRF microscopy ( Figure 4A). The results suggest that the mScarlet-C-SARAF fragment increased the assembly rate ( Figure 4B) without affecting the disassembly rate ( Figure 4B-E). In addition, we observed a decrease in cell retraction ( Figure 4F-H), indicating that the overexpression of C-SARAF impacts focal adhesion dynamics and cell motility. A decrease in cell movement is likely due to a decrease in cell retraction upon serum stimulation. In summary, mScarlet-C-SARAF overexpression increased the assembly rate of FA and reduced cell retraction.  To evaluate if the overexpression of mScarlet-C-SARAF in the human TNBC cell line MDA-MB-231 has the same effects in a murine model, we performed proliferation, migration, and invasion assays in the triple-negative murine cell line, 4T1. The results suggest that overexpression of mScarlet-C-SARAF resembled the effects of C-SARAF overexpression in MDA-MB-231 cells (Figure 3), showing a reduced SOCE response, diminished translocation of NFAT from the cytoplasm to the nucleus, cell migration, and invasion in the 4T1 murine cell line ( Figure 5A-H). To assess the effect of mScarlet-C-SARAF overexpression in a pathophysiological context, we used a murine model of tumor growth and metastasis. We used an in vivo syngeneic murine model of tumor growth. 4T1 cells (BALB/c strain) stably transfected with mScarlet or mScarlet-C-SARAF were injected into the fourth inguinal mammary gland to generate the primary tumor ( Figure 5N). No significant difference was observed in primary tumor size between both conditions ( Figure 5N,O). However, mice injected with cells expressing the mScarlet-C-SARAF fragment had a lower percentage of liver vessels affected by infiltrating tumor cells compared to the control ( Figure 5P,Q). metastasis. We used an in vivo syngeneic murine model of tumor growth. 4T1 ce (BALB/c strain) stably transfected with mScarlet or mScarlet-C-SARAF were injected in the fourth inguinal mammary gland to generate the primary tumor ( Figure 5N). No si nificant difference was observed in primary tumor size between both conditions ( Figu  5N,O). However, mice injected with cells expressing the mScarlet-C-SARAF fragment ha a lower percentage of liver vessels affected by infiltrating tumor cells compared to th control ( Figure 5P,Q).

Discussion
Several studies have shown that in breast cancer, the regulation of Ca 2+ signaling is crucial for tumorigenesis and cellular processes, such as cell growth, proliferation, migration, metastasis, and resistance to apoptosis (reviewed in [5,41]). In Ca 2+ signaling, remodeling of the composition and activity of the protein complexes involved is a common phenomenon. Proteins involved in the SOCE response, such as Orai [20,42], TRPC1 [43,44], TRPC6 [45,46] and STIM [20,[47][48][49] act as modulators of the Ca 2+ homeostatic response. In fact, evidence indicates that the increased SOCE response is key for tumorigenic processes, such as proliferation, metastasis, tumor formation, and the avoidance of antitumor immunity [50]. Nevertheless, although only few regulators of the SOCE response are involved in this process, the development of compounds targeting SOCE has been studied as a potential strategy for cancer therapy, including breast cancer [51]. As these compounds are specific and exhibit side effects, the study of the interactors or regulators of the SOCE response has become an interesting pharmacological target. Herein, we propose the role of the C-terminal region of SARAF as a potential negative regulator of the SOCE response in TNBC.
When SARAF was first identified, Palty et al. (2012) reported that overexpression of the C-terminal region of SARAF reduced the SOCE response efficiently [30] and in our research, it reduced Ca 2+ entry during Fura-2-AM recordings in HEK293 (Figure 2A-D), MDA-MB-231 ( Figure 2F-I) and 4T1 cells ( Figure 5A-C). Furthermore, the mScarlet-C-SARAF fragment induced a reduction in the SOCE response and led to a reduced STIM1-Orai1 complex formation, as evidenced by the BiFC assay in a human TNBC cell line. These data indicate that a reduction in the SOCE response in breast cancer cells could be related to the inhibition of STIM1-Orai1 complex assembly, as reported previously [30].
Different reports have shown the relevance of the SOCE response and its effect on cell migration and the invasion of neoplastic cells [20,52,53]. A reduction in the expression of STIM1 and Orai1 diminishes cell migration and invasion, and decreases metastasis in a murine model [12,20]. Our results suggest that the mScarlet-C-SARAF fragment reduced the SOCE response and resulted in lower migration and the invasion of MDA-MB-231 and 4T1 cells without affecting cell proliferation. This effect can be explained by an alteration in the regulation of focal adhesion formation.
Currently, the molecular mechanisms involved in the SOCE-dependent regulation of focal adhesions need further investigation. D'Souza et al. (2020) showed that the Orai1-STIM1 complex clusters near the FAs of MDA-MB-231 cells and Ca 2+ influx via SOCE leads to activation of the GTPase ARF5 via the Ca 2+ -activated GEF IQSec1, which is essential for promoting focal adhesion disassembly [54]. In an osteosarcoma model, Lin et al. (2021) showed that constitutively active STIM1 increased Ca 2+ influx and the turnover of FA proteins, which impeded cell migration [55]. In contrast, dominant-negative STIM1 decreased FA turnover and promoted cell migration. Another protein has been reported to regulate the conformational change in STIM1, promoting its activation in TNBC. Lee et al. (2022) determined that PGRMC1, a protein that is expressed in TNBC, promotes the SOCE response, in addition to regulating the turnover of focal adhesions and the formation of actomyosin [56]. These data suggest that cancer cells require sufficient Ca 2+ to control the assembly and disassembly of focal adhesions. Our data show an increase in the rate of assembly of focal adhesions and a decrease in cell retraction in the mScarlet-C-SARAF overexpressed condition in MDA-MB-231 cells, consistent therefore with a lower migration potential. Thus, the mScarlet-C-SARAF fragments could interact directly with STIM1, regulating its conformational change, which may result in reduced cell migration and the invasion of cancer cells.
During cell invasion, STIM-Orai1 is essential for orchestrating the activation of metalloproteinases [11,57] and the formation of podosomes [23] that allow the degradation and migration of neoplastic cells through the extracellular matrix. Sun et al. (2014) described that oscillations of Orai1 activity have an impact on invadosome formation and the activation of the metalloproteinase MT1-MPP [11]. In addition, Rizaner et al. (2016) reported the high metastatic potential of MDA-MB-231 cells due to spontaneous Ca 2+ oscillations, whereas such oscillations were not observed in MCF7 cells, which had a reduced metastatic potential [58]. It has been reported that in MDA-MB-231 cells, the expression of Orai1 predominates over Orai2, and this is associated with the activation of NFATc1 with low concentrations of agonists [59]. Taken together, these data indicate that mScarlet-C-SARAF mainly regulates Orai1, which could influence the activation of MPP and the formation of podosomes, implying diminished cell invasion.
Chakraborty et al. (2016) described the use of phemindole, a synthetic dietary indole that exhibits antitumor properties and is proposed as a potential compound for pharmacological use in chemotherapy. Phemindole treatment reduces STIM1 expression by reducing the interaction between STIM1 and Orai1, resulting in decreased cell migration in vitro through the regulation of FAK; a reduction in tumor growth in vivo was also observed [60].
In summary, our data suggest that mScarlet-C-SARAF fragments are a potential alternative to treat breast cancer subtypes such as TNBC due to their inhibitory activity upon the SOCE response and the subsequent reduction in cell migration and invasion. Thus, methods to selectively deliver mScarlet-C-SARAF fragments to tumors are necessary to prevent cancer cell dissemination.
Stable lines of 4T1 cells that overexpress the fluorescent protein mScarlet and the fused proteins mScarlet-C-SARAF were generated using the antibiotic G418 (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, Cat#11811-031) at a concentration of 2 mg/mL. These cells were subsequently sorted at the REDECA facility (Faculty of Medicine, Universidad de Chile) using a FACSAria™ III Cell Sorter by selecting cells that exhibited red fluorescence for use in the in vivo metastasis experiments.

Plasmids, Antibodies and Reagents
The plasmids used are summarized in Table 1. The mScarlet-C-SARAF plasmid was generated from SARAF-myc-DDK (Origene #RC201864) as a template for the amplification of the C-SARAF sequence (amino acids 195 to 339, according to the Uniprot database). Primers with restriction sites 5 -GAG CTC AAG ACG GGC AGT ATT CTC CTC CAC-3 and 5 -CTG CAG TTA TCG TCT CCT GGT AC-3 were designed and the C-terminal SARAF sequence was incorporated into Sacl and Pstl sites of the pmScarlet-H_C1 plasmid (Addgene #85043).
The antibodies used for immunoblot are listed in Table 2.

Ca 2+ Imaging
Cells were loaded with 5 µM Fura-2-AM (Thermo Fisher Scientific, Waltham, MA, USA, Cat#F1201) for 30 min at room temperature. Then, cells were washed once and recorded using Ringer's modified medium at pH 7.4 containing 140 mM NaCl, 2.5 mM KCl, 10 mM Glucose, 2 mM CaCl 2 , 1 mM MgCl 2 and HEPES 10 mM. ER Ca 2+ depletion was induced with 2 µM thapsigargin (TG) (Merck, Darmstadt, Germany, Cat # 67526-95-8) in free-Ca 2+ Ringer's modified medium containing 1 mM EGTA in the absence of CaCl 2 . Samples were excited using an halogen light source with alternated 340 nm and 380 nm excitation filters (Chroma). Fluorescence emissions at 510 nm were captured using a Chameleon Camera (Chameleon CM3-U3-31S4, FLIR, Richmond, BC, Canada), and an Eclipse Ti2-U inverted microscope (Nikon, Tokio, Japan). The SOCE protocol in 4T1 and MDA-MB-231 consisted of 3 min of basal recording with Ringer's modified with 2 mM CaCl 2 , 7 min of reticular Ca 2+ depletion using Ca 2+ -free Ringer's with 2 µM TG to finally replace the milieu with modified Ringer's 2 mM CaCl 2 solution, and recording for 5 min. Fluorescence intensity was quantified using the Image J2/FIJI v.2.9.0 software (National Institute of Mental Health, Bethesda, MD, USA) and data were normalized using the initial ratio of fluorescence 340 nm/380 nm.

Immunoprecipitation
Immunopurification assays were performed as previously described [61]. Briefly, HEK293 transfected with mScarlet and mScarlet-C-SARAF were solubilized in lysis buffer containing 1% v/v Triton X-100, 150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl (pH 8.0), 1 mM NaVO 4 , 5 mM NaF, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF) and protease inhibitor cocktail for 15 min at 4 • C, followed by centrifugation at 11,000× g for 10 min at 4 • C. The supernatants were incubated with anti-RFP antibodies ( Table 2) for 3 h at 4 • C on rotation, followed by the addition of protein G sepharose beads (Cytiva, Marlborough, MA, USA, Cat #17061802) for 1 h at 4 • C on rotation. The beads were washed six times in lysis buffer and immunopurified complexes were eluted by boiling in Reducing Sample Buffer for 5 min, followed by size fractionation on SDS-PAGE and analysis by immunoblot.

Focal Adhesion Dynamics
Focal adhesion (FAs) turnover was studied via live-cell imaging recordings using total internal reflection fluorescence microscopy (TIRFM) [64,65]. This approach was performed in MDA-MB-231 cells transfected with EGFP-Paxillin and mScarlet or mScarlet-C-SARAF. The cells were transfected in suspension and immediately seeded in fibronectin-coated coverslips (5 µg/mL). The cells were depleted from serum 48 h post-transfection for at least 3 h. Imaging was performed at room temperature. FA turnover dynamics were stimulated by the addition of 10% v/v FBS in Ringer's solution modified with 2 mM CaCl 2 . Images were acquired every 1 min for 30 frames using a TIRFM 60× objective (N.A. = 1.45) in an Olympus IX71 microscope equipped with an FLIR camera (Backfly S, BFS-U3-51S5M, FLIR, Richmond, BC, Canada). Images were processed using Image J2/FIJI v.2.9.0 software. FA assembly and disassembly were visualized and quantified as the appearance or loss of fluorescence in a region of interest, according to Goetz et al. (2008) [66]. The assembly and disassembly rates were calculated by plotting the values of ln I/I 0 versus time as described [65].

Transwell Chamber Invasion Assays
MDA-MB-231 and 4T1 cells (5 × 10 4 cells) transfected with mScarlet or mScarlet-C-SARAF were plated into 8 µm pore Transwell ® chambers (Sigma-Aldrich, San Louis, MO, USA, Cat#CLS3422). For the invasion assay, Transwell ® chambers were preincubated with Matrigel (Sigma-Aldrich, San Louis, MO, USA, Cat#E1270) overnight at 4 • C and then they were placed for 30 min at 37 • C before seeding the cells. Cell invasion was induced by adding 10% v/v FBS in the lower chamber for 16 h at 37 • C. The non-invasive cells were removed, and the invading cells were fixed and stained with 0.2% w/v crystal violet dissolved in 10% v/v methanol. The invading cells were counted and expressed as percentage of control.

Scratch Assay
MDA-MB-231 and 4T1 cells transfected with mScarlet or mScarlet-C-SARAF were seeded on coverslips pre-coated with 5 µg/mL of Human plasma Fibronectin (Sigma-Aldrich, San Louis, MO, USA, Cat#FC010), forming a monolayer. After 24 h, cells were depleted of serum for 4 h and then scratches were made with a sterile micropipette tip (p200). Detached cells were removed by washing three times with DPBS, and then the medium was replaced with DMEM 2% FBS and three images were obtained from each wound (t 0 ) with an inverted microscope. After 16 h, cells were fixed for 15 min at 4 • C in fixative solution [4% w/v formaldehyde (freshly prepared from paraformaldehyde, Sigma-Aldrich, San Louis, MO, USA, Cat#158127), 4% w/v sucrose (Sigma-Aldrich, San Louis, MO, USA, Cat#S0389) in Dulbecco's phosphate-buffered saline (DPBS), pH 7.4], and three images were obtained from each wound (t 16h ). The relative sizes of the wounds at t 0 and t 16h were obtained using a wound-healing size plugin of the Image J2/FIJI v.2.9.0 software and results were expressed as percentage of closing ([(Area t 0h − Area t 16h )/ Area t 0h ] × 100).

In Vivo Metastasis Model
Stable cell lines of 4T1 that overexpress mScarlet or mScarlet-C-SARAF were injected to generate two animal models associated with the animal protocol 20382-MED-UCH, which was approved by the Bioethics Committee of the Faculty of Medicine, Universidad de Chile. Female Balb/c mice were orthotopically injected into the fatty tissue of the fourth mammary gland (inguinal area) with stable cell line 4T1. Cells were resuspended in Dulbecco's modified phosphate-buffered saline (DPBS) with 0.04% trypan blue (2 × 10 5 cells, 50 µL) [67]. The general condition of the animals was evaluated daily according to the modified monitoring protocols of Morton and Griffiths (Veterinary Record, 116: 431- 36,1985). The mice were maintained for 21 days or when the score was equal to or greater than 15. Tumor growth was assessed postmortem. Once the tumor was extracted, three diameters of the ellipsoid were measured using a caliper and the following formula was used to calculate the volume [68]: Additionally, organs were removed, fixed, and processed for histological analysis and the evaluation of metastatic sites.

Histology
The main organs that were affected during cancer metastasis, either by indirect effect of primary tumor growth, were removed, washed with DPBS and fixed in fixative solution (4% w/v formaldehyde-freshly prepared from paraformaldehyde-4% w/v sucrose in DPBS, pH 7.4) for 24 h at 4 • C. Subsequently, the tissue was processed with increasing concentrations of alcohol, cleared with xylol and embedded in paraffin. Serial sections with a thickness of 3 µm were obtained and adhered to slides for histological hematoxylin and eosin and Masson's trichrome staining. Images were acquired with a Leica DFC425C camera installed on a Leica DM2500 CCD microscope with 10× and 40× objectives (Leica, Wetzlar, Germany).