The communication between tumor cells and their neighboring stromal cells is essential to sustain tumor growth and promote invasion and metastasis (Becker et al., 2016; Follain et al., 2020). Notably, this communication allows tumors to indoctrinate their microenvironment and switch the phenotypes of various cell types, such as endothelial cells, fibroblasts, or immune cells to the benefit of tumor growth, invasion, immune escape and metastasis. Such communication occurs with organs distant of the primary tumors and favors the formation of pre-metastatic niches where the modified microenvironment can help settling metastatic tumor cells (Peinado et al., 2017). Seeding of this favorable metastatic environment can be mediated by soluble molecules (Kaplan et al., 2005; Wang et al., 2017) or by extracellular vesicles (EVs) secreted by tumor cells (Costa-Silva et al., 2015; Hoshino et al., 2015; Jung et al., 2009; Peinado et al., 2012). EVs are lipid bilayered vesicles of nanometric diameters containing a complex mixture of RNA and protein cargoes, including a repertoire of surface receptors (Mathieu et al., 2019). They can be directly secreted from the plasma membrane and called microvesicles or originate from an endosomal compartment, the multi-vesicular body (MVB), and then called exosomes (van Niel et al., 2018). The levels of circulating tumor EVs tend to correlate with tumor progression (Baran et al., 2010; Galindo-Hernandez et al., 2013; Logozzi et al., 2009). Accordingly, inhibition of key components of the EV secretion machinery often correlates with decreased metastasis (Hyenne et al., 2017). For instance, Rab27a, which directs exosome secretion by controlling the docking of MVBs to the plasma membrane (Ostrowski et al., 2010), promotes breast and melanoma tumor growth and metastasis in mice (Bobrie et al., 2012; Peinado et al., 2012) and predicts poor survival in human pancreatic cancer (Wang et al., 2015). In addition to the levels of secreted tumor EVs, their content, and in particular their set of surface adhesion proteins equally orchestrates metastasis formation. For instance, the presence of tetraspanins CD151 and Tspan8 on the surface of pancreatic adenocarcinoma EVs favors metastasis in rats by enhancing their adhesive capacities and controlling their biodistribution (Yue et al., 2015). Moreover, integrin receptors exposed by tumor EVs dictate their organotropism and thereby tune/control the seeding of a premetastatic niche in specific and distant organ (Hoshino et al., 2015). Therefore, accumulating evidence show that both the levels and the content of secreted tumor EVs are instrumental in promoting metastasis.

However, the molecular mechanisms coordinating these processes remain elusive. In particular, how the machinery governing EV secretion can impact the pro-metastatic properties of tumor EVs deserves in-depth characterization. To address this issue, we focused on the members of the Ral family, RalA and RalB (collectively referred to as RalA/B), acting downstream of RAS and promoting metastasis of different tumor types in both mice and human (Gentry et al., 2014; Yan and Theodorescu, 2018). We recently found that these versatile proteins are evolutionarily conserved regulators of exosome secretion (Hyenne et al., 2015). We originally observed that, in the nematode C. elegans, the Ral GTPase ortholog RAL-1 controls exosome secretion by acting on the biogenesis of MVBs. Importantly, we further showed that RalA/B modulate the levels of secreted EVs in models that are relevant to human breast cancer (Hyenne et al., 2015) suggesting that these GTPases could influence disease progression through EVs release. Here, we exploited 4T1 cells, an aggressive mammary tumor model that mimics human triple-negative breast cancer (Kaur et al., 2012) to further decipher how RalA/B tune EV secretion mechanisms and thereby control metastatic progression of the disease.

In this study, we first provide a detailed dissection of the impact of the Ral GTPases on EV secretion levels and unravel the mechanisms by which they control the homeostasis of MVBs. We have discovered that RalA/B directly acts through the phospholipase D1 (PLD1), which, as we show, also promotes EVs secretion, to favor the maturation of MVBs. We further demonstrate that RalA and RalB promote lung metastasis without affecting the invasive potential of breast carcinoma. Importantly, RalA/B are crucial for the organ targeting of tumor EVs, and, as a consequence, for the seeding of pre-metastatic niches. Finally, we identify the adhesion protein CD146/MCAM as a key EV cargo controlled by RalA and RalB and demonstrate that it conveys, in part, the pro-metastatic function to EVs by controlling the lung tropism of breast cancer EVs.

The therapeutic limitations of breast cancer metastasis warrant a deeper understanding of its molecular machinery. Our findings highlight the exosome-mediated priming of metastatic niches by Ral GTPases as a critical requisite for lung metastasis during breast cancer progression. We show that RalA and RalB promote the secretion of exosomes by maintaining a high number of multi-vesicular bodies, likely through the PLD1-PA axis. Furthermore, we demonstrate that RalGTPases favor the secretion of CD146-rich exosomes, which accumulate in metastatic organs, notably in lungs, where they establish premetastatic niches (Figure 5i). Finally, we show that high levels of RalA and RalB correlated with poor prognosis suggesting a unified mechanism for human breast cancer metastasis.

This work, together with our previous study of RAL-1 in C. elegans (Hyenne et al., 2015), establishes Ral GTPases as major evolutionarily conserved mediators of exosome secretion. Our experiments suggest that RalA/B contribute to exosome secretion in several tumor cell lines, of different origins, implying that they might function pleiotropically over various cancers. Our results suggest that RalA/B and their effector PLD1 affect the levels of secreted exosomes by tuning the levels of cytoplasmic MVBs. While Ral GTPases, partially localized at the plasma membrane, could also affect microvesicle secretion, our data indicate that they function in exosome biogenesis upstream of PLD1. Similarly, a direct correlation between MVB density and levels of secreted EVs was recently suggested by studies showing that chemical or electric stimulation of MVB biogenesis results in increased EV secretion (Kanemoto et al., 2016; Yang et al., 2020). The formation of MVBs results from dramatic biochemical transformations of endosomes involving multiple protein and lipid switches (Huotari and Helenius, 2011; Scott et al., 2014). Understanding the steps at which RalA/B and PLD affect this endosome maturation program is critical and remains to be fully deciphered. Our results from mice and C. elegans suggest that biogenesis of ILVs, which is a key step in MVB maturation and the initial phase of the exosome secretion pathway, could as well be controlled by RalA/B. Our work further identifies PLD as an effector acting downstream of Ral to control exosome secretion. Whether other Ral effectors contribute to EV secretion remains to be addressed. Interestingly, while PLD2 was found to impact exosome secretion by governing ILV biogenesis in a different breast carcinoma cell line (Ghossoub et al., 2014), our data rather suggest that PLD1 controls exosome biogenesis in 4T1 cells. Indeed, PLD1 localizes on MVBs and its inhibition, but not the inhibition of PLD2, decreases MVB density. Nevertheless, it should be noted that we measured EV secretion levels and MVB density based on PLD inhibition at previously-published high concentrations of the inhibitors (compared to their respective IC50) and that off-target effect can not be ruled out. By contrast, PLD2 is essentially localized at the plasma membrane of 4T1 cells and its inhibition reduces EV secretion suggesting that PLD2 could rather promote microvesicle secretion in 4T1 cells. Therefore, we speculate that RalA/B-PLD1 control ILV biogenesis in 4T1 cells, possibly through the regulation of PA levels. Alternatively, they could impact the homeostasis of a subclass of MVBs, for instance by controlling their stability or their degradation.

Priming of metastatic niches by (soluble or) EV-mediated factors takes central stages in cancer progression (Gao et al., 2019; Peinado et al., 2017) and identification of molecular machineries that underlie this condition could point to new therapeutic or diagnostic targets. Our study demonstrates that Ral GTPases enhance the formation of lung metastasis in mouse models, by promoting the secretion of exosomes within primary tumors, while RalA/B expression levels correlates with metastasis in human breast cancer. While the pro-tumoral activity of Ral GTPases was so far mostly attributed to their capacity to promote anchorage-independent cell growth (for RalA) or cell invasion (for RalB) (Yan and Theodorescu, 2018), we now show that Ral GTPases have additional non-cell autonomous functions, and that these functions are important contributors to metastasis. Indeed, in 4T1 cells, depletion of either RalA or RalB alters the levels, content and functionality of secreted EVs, without decreasing cell migration or proliferation. Depending on the cell type or the biological process, RalA and RalB can display redundant, synergistic or even antagonist activities (Gentry et al., 2014). Since RalA and RalB mostly share similar phenotypes regarding EV secretion, content and function, they likely function in the same pathway. Interestingly, both Ral proteins appear to be essential for exosome secretion, revealing that their functions are not fully redundant. Therefore, both GTPases are required for the generation of a specific subpopulation of EVs with enhanced pro-metastatic properties and further work is needed to fully unravel the downstream molecular pathways. With this work, RalA and RalB add to the list of proteins known to control exosome secretion and to affect tumor progression, such as Rab27a (Bobrie et al., 2012; Kren et al., 2020; Peinado et al., 2012), Alix (Monypenny et al., 2018), syntenin (Das et al., 2019), and components of the ESCRT machinery (Mattissek and Teis, 2014). These studies demonstrate that the number of EVs secreted by a primary tumor is an essential element determining the efficiency of metastasis. However, it is important to keep in mind that all these proteins regulating EV trafficking, including RalA/B, contribute to tumor progression through both exosome dependent and exosome independent functions. Altogether, despite pointing to additional functions of RAL GTPases, our study is the first to identify new molecular machinery from its function in EV biogenesis up to its pro-metastatic function in breast cancer lung metastasis.

Priming of metastatic niches by EVs has, so far, mostly been attributed to increased levels of pro-metastatic EVs with pro-metastatic functions (Becker et al., 2016; Bobrie et al., 2012; Peinado et al., 2012). In addition to controlling the levels of secreted EVs, we show that RalA/B affect their function by enhancing their capacity to induce endothelial permeability in vitro and pre-metastatic niches in vivo. These two observations could be linked, as RalA/B-dependent EVs could promote endothelial permeability locally in the primary tumor or at distance in lungs, thereby favoring both tumor intravasation and extravasation. Content analysis revealed that RalA/B control the identity and levels of RNAs and proteins present in secreted EVs. Interestingly, Ras, which is known to activate RalA/B (Gentry et al., 2014), also controls the protein and RNA cargo of tumor EVs (Cha et al., 2015; Demory Beckler et al., 2013; McKenzie et al., 2016), although its effect on the levels of secreted EVs is unclear (Demory Beckler et al., 2013; McKenzie et al., 2016). As McKenzie and collaborators identified a MEK-ERK-Ago2 pathway downstream of Ras (McKenzie et al., 2016), it would be interesting to determine how this pathway connects with the Ral-PLD-PA axis described in our study. Among the few proteins significantly enriched in RalA/B-dependent EVs, we identified CD146, a molecule known to modulate cell-cell adhesion (Wang and Yan, 2013). We showed, using functional inhibition, that CD146 present on pro-metastatic EVs controls their lung targeting efficiency thereby impacting their biodistribution and niche-promoting function. Accordingly, we and others show that high expression of CD146 correlates with poor prognosis in human breast carcinoma (Garcia et al., 2007; Zeng et al., 2012). CD146 functions as an adhesion molecule involved in homophilic and heterophilic interactions (Wang and Yan, 2013), promoting for instance monocyte transmigration (Bardin et al., 2009). CD146 can perform trans-homophilic interactions via its immunoglobulin-like extracellular domain (Taira et al., 1994; Taira et al., 2005). It also binds to extracellular matrix proteins or other transmembrane proteins, such as VEGFR2 (Wang and Yan, 2013). Therefore, it is tempting to speculate that CD146 affects the biodistribution and organ targeting efficiency of circulating tumor EVs by mediating their interaction with specific ligands present on the luminal side of endothelial cells of metastatic organs. Other adhesion molecules, such as integrins and tetraspanins were shown to affect the biodistribution of tumor EVs and ultimately the formation of metastasis (Hoshino et al., 2015; Yue et al., 2015). Therefore, it is likely that the combination of these receptors at the surface of tumor EVs, combined with the differential expression of their ligands on endothelial cells throughout the organism will dictate their homing. More work will be needed to characterize this organ specific EV zip code and to identify relevant endothelial ligands for circulating EVs and develop inhibitory strategies to impair their arrest and uptake at metastatic sites. In addition, the presence of other cell types in the circulation, such as patrolling monocytes, which take up large amounts of circulating EVs, could also contribute to the accumulation of tumor EVs in specific organs (Hyenne et al., 2019; Plebanek et al., 2017). Finally, other factors, such as the vascular architecture and hemodynamic patterns could be involved (Follain et al., 2020; Hyenne et al., 2019) and the interplay between these mechanical cues and the surface repertoire of metastatic EVs should be a fertile ground for future research. Precisely dissecting the mechanisms by which tumor EVs reach specific organs would allow to understand the priming of premetastatic niches.

Overall, our study identifies RalA/B GTPases as a novel molecular machinery that regulates the formation and shedding of pro-metastatic EVs. We also discovered CD146 as an EV cargo whose targeting could inspire new therapeutic strategies to impact the progression of metastatic breast cancer.

Sequencing and mass spectrometry data have been deposited in the EV-Track database.

1. 1. The Cancer Genome Atlas Network

   (2012) TCGA breast invasive carcinoma cohort (1097 patients)

   Comprehensive molecular portraits of human breast tumours.

   https://doi.org/10.1038/nature11412

1. 1. Anders S
   2. Pyl PT
   3. Huber W

   (2015) HTSeq--a Python framework to work with high-throughput sequencing data

   Bioinformatics 31:166–169.

   https://doi.org/10.1093/bioinformatics/btu638

   • PubMed
   • Google Scholar
2. 1. Bankhead P
   2. Loughrey MB
   3. Fernández JA
   4. Dombrowski Y
   5. McArt DG
   6. Dunne PD
   7. McQuaid S
   8. Gray RT
   9. Murray LJ
   10. Coleman HG
   11. James JA
   12. Salto-Tellez M
   13. Hamilton PW

   (2017) QuPath: open source software for digital pathology image analysis

   Scientific Reports 7:5.

   https://doi.org/10.1038/s41598-017-17204-5

   • PubMed
   • Google Scholar
3. 1. Baran J
   2. Baj-Krzyworzeka M
   3. Weglarczyk K
   4. Szatanek R
   5. Zembela M
   6. Barbasz J
   7. Czupryna A
   8. Szczepanik A
   9. Zembala M

   (2010) Circulating tumour-derived microvesicles in plasma of gastric Cancer patients. Cancer immunology

   Immunotherapy 59:841–850.

   https://doi.org/10.1007/s00262-009-0808-2

   • Google Scholar
4. 1. Bardin N
   2. Blot-Chabaud M
   3. Despoix N
   4. Kebir A
   5. Harhouri K
   6. Arsanto JP
   7. Espinosa L
   8. Perrin P
   9. Robert S
   10. Vely F
   11. Sabatier F
   12. Le Bivic A
   13. Kaplanski G
   14. Sampol J
   15. Dignat-George F

   (2009) CD146 and its soluble form regulate monocyte transendothelial migration

   Arteriosclerosis, Thrombosis, and Vascular Biology 29:746–753.

   https://doi.org/10.1161/ATVBAHA.108.183251

   • PubMed
   • Google Scholar
5. 1. Becker A
   2. Thakur BK
   3. Weiss JM
   4. Kim HS
   5. Peinado H
   6. Lyden D

   (2016) Extracellular vesicles in Cancer: cell-to-cell mediators of metastasis

   Cancer Cell 30:836–848.

   https://doi.org/10.1016/j.ccell.2016.10.009

   • PubMed
   • Google Scholar
6. 1. Bobrie A
   2. Krumeich S
   3. Reyal F
   4. Recchi C
   5. Moita LF
   6. Seabra MC
   7. Ostrowski M
   8. Théry C

   (2012) Rab27a supports exosome-dependent and -independent mechanisms that modify the tumor microenvironment and can promote tumor progression

   Cancer Research 72:4920–4930.

   https://doi.org/10.1158/0008-5472.CAN-12-0925

   • PubMed
   • Google Scholar
7. 1. Bruntz RC
   2. Lindsley CW
   3. Brown HA

   (2014) Phospholipase D signaling pathways and phosphatidic acid as therapeutic targets in Cancer

   Pharmacological Reviews 66:1033–1079.

   https://doi.org/10.1124/pr.114.009217

   • PubMed
   • Google Scholar
8. 1. Cha DJ
   2. Franklin JL
   3. Dou Y
   4. Liu Q
   5. Higginbotham JN
   6. Demory Beckler M
   7. Weaver AM
   8. Vickers K
   9. Prasad N
   10. Levy S
   11. Zhang B
   12. Coffey RJ
   13. Patton JG

   (2015) KRAS-dependent sorting of miRNA to exosomes

   eLife 4:e07197.

   https://doi.org/10.7554/eLife.07197

   • PubMed
   • Google Scholar
9. 1. Collot M
   2. Ashokkumar P
   3. Anton H
   4. Boutant E
   5. Faklaris O
   6. Galli T
   7. Mély Y
   8. Danglot L
   9. Klymchenko AS

   (2018) MemBright: a family of fluorescent membrane probes for advanced cellular imaging and neuroscience

   Cell Chemical Biology 26:9.

   https://doi.org/10.1016/j.chembiol.2019.01.009

   • Google Scholar
10. 1. Corrotte M
    2. Chasserot-Golaz S
    3. Huang P
    4. Du G
    5. Ktistakis NT
    6. Frohman MA
    7. Vitale N
    8. Bader MF
    9. Grant NJ

    (2006) Dynamics and function of phospholipase D and phosphatidic acid during phagocytosis

    Traffic 7:365–377.

    https://doi.org/10.1111/j.1600-0854.2006.00389.x

    • PubMed
    • Google Scholar
11. 1. Costa-Silva B
    2. Aiello NM
    3. Ocean AJ
    4. Singh S
    5. Zhang H
    6. Thakur BK
    7. Becker A
    8. Hoshino A
    9. Mark MT
    10. Molina H
    11. Xiang J
    12. Zhang T
    13. Theilen TM
    14. García-Santos G
    15. Williams C
    16. Ararso Y
    17. Huang Y
    18. Rodrigues G
    19. Shen TL
    20. Labori KJ
    21. Lothe IM
    22. Kure EH
    23. Hernandez J
    24. Doussot A
    25. Ebbesen SH
    26. Grandgenett PM
    27. Hollingsworth MA
    28. Jain M
    29. Mallya K
    30. Batra SK
    31. Jarnagin WR
    32. Schwartz RE
    33. Matei I
    34. Peinado H
    35. Stanger BZ
    36. Bromberg J
    37. Lyden D

    (2015) Pancreatic Cancer exosomes initiate pre-metastatic niche formation in the liver

    Nature Cell Biology 17:816–826.

    https://doi.org/10.1038/ncb3169

    • PubMed
    • Google Scholar
12. 1. Cox J
    2. Hein MY
    3. Luber CA
    4. Paron I
    5. Nagaraj N
    6. Mann M

    (2014) Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ

    Molecular & Cellular Proteomics 13:2513–2526.

    https://doi.org/10.1074/mcp.M113.031591

    • PubMed
    • Google Scholar
13. Book

    1. Das SK
    2. Sarkar D
    3. Emdad L
    4. Fisher PB

    (2019)

    MDA-9/Syntenin: An Emerging Global Molecular Target Regulating Cancer Invasion and Metastasis

    Elsevier Inc.

    • Google Scholar
14. 1. Demory Beckler M
    2. Higginbotham JN
    3. Franklin JL
    4. Ham AJ
    5. Halvey PJ
    6. Imasuen IE
    7. Whitwell C
    8. Li M
    9. Liebler DC
    10. Coffey RJ

    (2013) Proteomic analysis of exosomes from mutant KRAS Colon cancer cells identifies intercellular transfer of mutant KRAS

    Molecular & Cellular Proteomics 12:343–355.

    https://doi.org/10.1074/mcp.M112.022806

    • PubMed
    • Google Scholar
15. 1. Deutsch EW
    2. Bandeira N
    3. Sharma V
    4. Perez-Riverol Y
    5. Carver JJ
    6. Kundu DJ
    7. García-Seisdedos D
    8. Jarnuczak AF
    9. Hewapathirana S
    10. Pullman BS
    11. Wertz J
    12. Sun Z
    13. Kawano S
    14. Okuda S
    15. Watanabe Y
    16. Hermjakob H
    17. MacLean B
    18. MacCoss MJ
    19. Zhu Y
    20. Ishihama Y
    21. Vizcaíno JA

    (2020) The ProteomeXchange consortium in 2020: enabling “big data” approaches in proteomics

    Nucleic Acids Research 48:gkz984.

    https://doi.org/10.1093/nar/gkz984

    • Google Scholar
16. 1. Follain G
    2. Herrmann D
    3. Harlepp S
    4. Hyenne V
    5. Osmani N
    6. Warren SC
    7. Timpson P
    8. Goetz JG

    (2020) Fluids and their mechanics in tumour transit: shaping metastasis

    Nature Reviews Cancer 20:107–124.

    https://doi.org/10.1038/s41568-019-0221-x

    • PubMed
    • Google Scholar
17. 1. Galindo-Hernandez O
    2. Villegas-Comonfort S
    3. Candanedo F
    4. González-Vázquez MC
    5. Chavez-Ocaña S
    6. Jimenez-Villanueva X
    7. Sierra-Martinez M
    8. Salazar EP

    (2013) Elevated concentration of microvesicles isolated from peripheral blood in breast Cancer patients

    Archives of Medical Research 44:208–214.

    https://doi.org/10.1016/j.arcmed.2013.03.002

    • PubMed
    • Google Scholar
18. 1. Gao Y
    2. Bado I
    3. Wang H
    4. Zhang W
    5. Rosen JM
    6. Zhang XH

    (2019) Metastasis organotropism: redefining the congenial soil

    Developmental Cell 49:375–391.

    https://doi.org/10.1016/j.devcel.2019.04.012

    • PubMed
    • Google Scholar
19. 1. Garcia S
    2. Dalès JP
    3. Charafe-Jauffret E
    4. Carpentier-Meunier S
    5. Andrac-Meyer L
    6. Jacquemier J
    7. Andonian C
    8. Lavaut MN
    9. Allasia C
    10. Bonnier P
    11. Charpin C

    (2007) Poor prognosis in breast carcinomas correlates with increased expression of targetable CD146 and c-Met and with proteomic basal-like phenotype

    Human Pathology 38:830–841.

    https://doi.org/10.1016/j.humpath.2006.11.015

    • PubMed
    • Google Scholar
20. 1. Gentry LR
    2. Martin TD
    3. Reiner DJ
    4. Der CJ

    (2014) Ral small GTPase signaling and oncogenesis: more than just 15minutes of fame

    Biochimica Et Biophysica Acta (BBA) - Molecular Cell Research 1843:2976–2988.

    https://doi.org/10.1016/j.bbamcr.2014.09.004

    • Google Scholar
21. 1. Ghossoub R
    2. Lembo F
    3. Rubio A
    4. Gaillard CB
    5. Bouchet J
    6. Vitale N
    7. Slavík J
    8. Machala M
    9. Zimmermann P

    (2014) Syntenin-ALIX exosome biogenesis and budding into multivesicular bodies are controlled by ARF6 and PLD2

    Nature Communications 5:3477.

    https://doi.org/10.1038/ncomms4477

    • PubMed
    • Google Scholar
22. 1. Heymann D
    2. Kerdraon O
    3. Verriele V
    4. Verhille E
    5. Veron V
    6. Vitre M
    7. Delmas F
    8. Henry C
    9. Gouy Y
    10. Amiand M
    11. Bard J-M

    (2020) Centre de ressources Biologiques-Tumorothèque: bioresources and associated clinical data dedicated to translational research in oncology at the institut de cancérologie de l’Ouest, France

    Open Journal of Bioresources 7:62.

    https://doi.org/10.5334/ojb.62

    • Google Scholar
23. 1. Hoshino A
    2. Costa-Silva B
    3. Shen TL
    4. Rodrigues G
    5. Hashimoto A
    6. Tesic Mark M
    7. Molina H
    8. Kohsaka S
    9. Di Giannatale A
    10. Ceder S
    11. Singh S
    12. Williams C
    13. Soplop N
    14. Uryu K
    15. Pharmer L
    16. King T
    17. Bojmar L
    18. Davies AE
    19. Ararso Y
    20. Zhang T
    21. Zhang H
    22. Hernandez J
    23. Weiss JM
    24. Dumont-Cole VD
    25. Kramer K
    26. Wexler LH
    27. Narendran A
    28. Schwartz GK
    29. Healey JH
    30. Sandstrom P
    31. Labori KJ
    32. Kure EH
    33. Grandgenett PM
    34. Hollingsworth MA
    35. de Sousa M
    36. Kaur S
    37. Jain M
    38. Mallya K
    39. Batra SK
    40. Jarnagin WR
    41. Brady MS
    42. Fodstad O
    43. Muller V
    44. Pantel K
    45. Minn AJ
    46. Bissell MJ
    47. Garcia BA
    48. Kang Y
    49. Rajasekhar VK
    50. Ghajar CM
    51. Matei I
    52. Peinado H
    53. Bromberg J
    54. Lyden D

    (2015) Tumour exosome integrins determine organotropic metastasis

    Nature 527:329–335.

    https://doi.org/10.1038/nature15756

    • PubMed
    • Google Scholar
24. 1. Huotari J
    2. Helenius A

    (2011) Endosome maturation

    The EMBO Journal 30:3481–3500.

    https://doi.org/10.1038/emboj.2011.286

    • PubMed
    • Google Scholar
25. 1. Hyenne V
    2. Apaydin A
    3. Rodriguez D
    4. Spiegelhalter C
    5. Hoff-Yoessle S
    6. Diem M
    7. Tak S
    8. Lefebvre O
    9. Schwab Y
    10. Goetz JG
    11. Labouesse M

    (2015) RAL-1 controls multivesicular body biogenesis and exosome secretion

    Journal of Cell Biology 211:27–37.

    https://doi.org/10.1083/jcb.201504136

    • Google Scholar
26. 1. Hyenne V
    2. Lefebvre O
    3. Goetz JG

    (2017) Going live with tumor exosomes and microvesicles

    Cell Adhesion & Migration 11:173–186.

    https://doi.org/10.1080/19336918.2016.1276694

    • PubMed
    • Google Scholar
27. 1. Hyenne V
    2. Ghoroghi S
    3. Collot M
    4. Bons J
    5. Follain G
    6. Harlepp S
    7. Mary B
    8. Bauer J
    9. Mercier L
    10. Busnelli I
    11. Lefebvre O
    12. Fekonja N
    13. Garcia-Leon MJ
    14. Machado P
    15. Delalande F
    16. López AA
    17. Silva SG
    18. Verweij FJ
    19. van Niel G
    20. Djouad F
    21. Peinado H
    22. Carapito C
    23. Klymchenko AS
    24. Goetz JG

    (2019) Studying the fate of tumor extracellular vesicles at high spatiotemporal resolution using the zebrafish embryo

    Developmental Cell 48:554–572.

    https://doi.org/10.1016/j.devcel.2019.01.014

    • PubMed
    • Google Scholar
28. 1. Jiang H
    2. Luo JQ
    3. Urano T
    4. Frankel P
    5. Lu Z
    6. Foster DA
    7. Feig LA

    (1995) Involvement of ral GTPase in v-Src-induced phospholipase D activation

    Nature 378:409–412.

    https://doi.org/10.1038/378409a0

    • PubMed
    • Google Scholar
29. 1. Jung T
    2. Castellana D
    3. Klingbeil P
    4. Cuesta Hernández I
    5. Vitacolonna M
    6. Orlicky DJ
    7. Roffler SR
    8. Brodt P
    9. Zöller M

    (2009) CD44v6 dependence of premetastatic niche preparation by exosomes

    Neoplasia 11:1093–1105.

    https://doi.org/10.1593/neo.09822

    • PubMed
    • Google Scholar
30. 1. Kanemoto S
    2. Nitani R
    3. Murakami T
    4. Kaneko M
    5. Asada R
    6. Matsuhisa K
    7. Saito A
    8. Imaizumi K

    (2016) Multivesicular body formation enhancement and exosome release during endoplasmic reticulum stress

    Biochemical and Biophysical Research Communications 480:166–172.

    https://doi.org/10.1016/j.bbrc.2016.10.019

    • PubMed
    • Google Scholar
31. 1. Kaplan RN
    2. Riba RD
    3. Zacharoulis S
    4. Bramley AH
    5. Vincent L
    6. Costa C
    7. MacDonald DD
    8. Jin DK
    9. Shido K
    10. Kerns SA
    11. Zhu Z
    12. Hicklin D
    13. Wu Y
    14. Port JL
    15. Altorki N
    16. Port ER
    17. Ruggero D
    18. Shmelkov SV
    19. Jensen KK
    20. Rafii S
    21. Lyden D

    (2005) VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche

    Nature 438:820–827.

    https://doi.org/10.1038/nature04186

    • PubMed
    • Google Scholar
32. 1. Kaur P
    2. Nagaraja GM
    3. Zheng H
    4. Gizachew D
    5. Galukande M
    6. Krishnan S
    7. Asea A

    (2012) A mouse model for triple-negative breast Cancer tumor-initiating cells (TNBC-TICs) exhibits similar aggressive phenotype to the human disease

    BMC Cancer 12:120.

    https://doi.org/10.1186/1471-2407-12-120

    • Google Scholar
33. 1. Kebir A
    2. Harhouri K
    3. Guillet B
    4. Liu JW
    5. Foucault-Bertaud A
    6. Lamy E
    7. Kaspi E
    8. Elganfoud N
    9. Vely F
    10. Sabatier F
    11. Sampol J
    12. Pisano P
    13. Kruithof EK
    14. Bardin N
    15. Dignat-George F
    16. Blot-Chabaud M

    (2010) CD146 short isoform increases the proangiogenic potential of endothelial progenitor cells in vitro and in vivo

    Circulation Research 107:66–75.

    https://doi.org/10.1161/CIRCRESAHA.109.213827

    • PubMed
    • Google Scholar
34. 1. Kim D
    2. Langmead B
    3. Salzberg SL

    (2015) HISAT: a fast spliced aligner with low memory requirements

    Nature Methods 12:357–360.

    https://doi.org/10.1038/nmeth.3317

    • PubMed
    • Google Scholar
35. 1. Klumperman J
    2. Raposo G

    (2014) The complex ultrastructure of the endolysosomal system

    Cold Spring Harbor Perspectives in Biology 6:a016857.

    https://doi.org/10.1101/cshperspect.a016857

    • PubMed
    • Google Scholar
36. 1. Kren N
    2. Michaud D
    3. Bagchi S
    4. Greene K
    5. Pylayeva-Gupta Y

    (2020) Rab27a plays a dual role in metastatic propensity of pancreatic Cancer

    Scientific Reports 10:e64248-1.

    https://doi.org/10.1038/s41598-020-64248-1

    • Google Scholar
37. 1. Lewis JA
    2. Scott SA
    3. Lavieri R
    4. Buck JR
    5. Selvy PE
    6. Stoops SL
    7. Armstrong MD
    8. Brown HA
    9. Lindsley CW

    (2009) Design and synthesis of isoform-selective phospholipase D (PLD) inhibitors. part I: impact of alternative halogenated privileged structures for PLD1 specificity

    Bioorganic & Medicinal Chemistry Letters 19:1916–1920.

    https://doi.org/10.1016/j.bmcl.2009.02.057

    • PubMed
    • Google Scholar
38. 1. Logozzi M
    2. De Milito A
    3. Lugini L
    4. Borghi M
    5. Calabrò L
    6. Spada M
    7. Perdicchio M
    8. Marino ML
    9. Federici C
    10. Iessi E
    11. Brambilla D
    12. Venturi G
    13. Lozupone F
    14. Santinami M
    15. Huber V
    16. Maio M
    17. Rivoltini L
    18. Fais S

    (2009) High levels of exosomes expressing CD63 and caveolin-1 in plasma of melanoma patients

    PLOS ONE 4:e5219.

    https://doi.org/10.1371/journal.pone.0005219

    • PubMed
    • Google Scholar
39. 1. Lou Y
    2. Preobrazhenska O
    3. auf dem Keller U
    4. Sutcliffe M
    5. Barclay L
    6. McDonald PC
    7. Roskelley C
    8. Overall CM
    9. Dedhar S

    (2008) Epithelial-mesenchymal transition (EMT) is not sufficient for spontaneous murine breast Cancer metastasis

    Developmental Dynamics 237:2755–2768.

    https://doi.org/10.1002/dvdy.21658

    • PubMed
    • Google Scholar
40. 1. Love MI
    2. Huber W
    3. Anders S

    (2014) Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

    Genome Biology 15:550.

    https://doi.org/10.1186/s13059-014-0550-8

    • PubMed
    • Google Scholar
41. 1. Luo JQ
    2. Liu X
    3. Frankel P
    4. Rotunda T
    5. Ramos M
    6. Flom J
    7. Jiang H
    8. Feig LA
    9. Morris AJ
    10. Kahn RA
    11. Foster DA

    (1998) Functional association between arf and RalA in active phospholipase D complex

    PNAS 95:3632–3637.

    https://doi.org/10.1073/pnas.95.7.3632

    • PubMed
    • Google Scholar
42. 1. Mathieu M
    2. Martin-Jaular L
    3. Lavieu G
    4. Théry C

    (2019) Specificities of secretion and uptake of exosomes and other extracellular vesicles for cell-to-cell communication

    Nature Cell Biology 21:9–17.

    https://doi.org/10.1038/s41556-018-0250-9

    • PubMed
    • Google Scholar
43. 1. Mattissek C
    2. Teis D

    (2014) The role of the endosomal sorting complexes required for transport (ESCRT) in tumorigenesis

    Molecular Membrane Biology 31:111–119.

    https://doi.org/10.3109/09687688.2014.894210

    • PubMed
    • Google Scholar
44. 1. McKenzie AJ
    2. Hoshino D
    3. Hong NH
    4. Cha DJ
    5. Franklin JL
    6. Coffey RJ
    7. Patton JG
    8. Weaver AM

    (2016) KRAS-MEK Signaling Controls Ago2 Sorting into Exosomes

    Cell Reports 15:978–987.

    https://doi.org/10.1016/j.celrep.2016.03.085

    • Google Scholar
45. 1. Monypenny J
    2. Milewicz H
    3. Flores-Borja F
    4. Weitsman G
    5. Cheung A
    6. Chowdhury R
    7. Burgoyne T
    8. Arulappu A
    9. Lawler K
    10. Barber PR
    11. Vicencio JM
    12. Keppler M
    13. Wulaningsih W
    14. Davidson SM
    15. Fraternali F
    16. Woodman N
    17. Turmaine M
    18. Gillett C
    19. Franz D
    20. Quezada SA
    21. Futter CE
    22. Von Kriegsheim A
    23. Kolch W
    24. Vojnovic B
    25. Carlton JG
    26. Ng T

    (2018) ALIX regulates Tumor-Mediated immunosuppression by controlling EGFR activity and PD-L1 presentation

    Cell Reports 24:630–641.

    https://doi.org/10.1016/j.celrep.2018.06.066

    • PubMed
    • Google Scholar
46. 1. Ombrato L
    2. Nolan E
    3. Kurelac I
    4. Mavousian A
    5. Bridgeman VL
    6. Heinze I
    7. Chakravarty P
    8. Horswell S
    9. Gonzalez-Gualda E
    10. Matacchione G
    11. Weston A
    12. Kirkpatrick J
    13. Husain E
    14. Speirs V
    15. Collinson L
    16. Ori A
    17. Lee JH
    18. Malanchi I

    (2019) Metastatic-niche labelling reveals parenchymal cells with stem features

    Nature 572:603–608.

    https://doi.org/10.1038/s41586-019-1487-6

    • PubMed
    • Google Scholar
47. 1. Ostrowski M
    2. Carmo NB
    3. Krumeich S
    4. Fanget I
    5. Raposo G
    6. Savina A
    7. Moita CF
    8. Schauer K
    9. Hume AN
    10. Freitas RP
    11. Goud B
    12. Benaroch P
    13. Hacohen N
    14. Fukuda M
    15. Desnos C
    16. Seabra MC
    17. Darchen F
    18. Amigorena S
    19. Moita LF
    20. Thery C

    (2010) Rab27a and Rab27b control different steps of the exosome secretion pathway

    Nature Cell Biology 12:19–30.

    https://doi.org/10.1038/ncb2000

    • PubMed
    • Google Scholar
48. 1. Oxford G
    2. Owens CR
    3. Titus BJ
    4. Foreman TL
    5. Herlevsen MC
    6. Smith SC
    7. Theodorescu D

    (2005) RalA and RalB: antagonistic relatives in Cancer cell migration

    Cancer Research 65:7111–7120.

    https://doi.org/10.1158/0008-5472.CAN-04-1957

    • PubMed
    • Google Scholar
49. 1. Peinado H
    2. Alečković M
    3. Lavotshkin S
    4. Matei I
    5. Costa-Silva B
    6. Moreno-Bueno G
    7. Hergueta-Redondo M
    8. Williams C
    9. García-Santos G
    10. Ghajar C
    11. Nitadori-Hoshino A
    12. Hoffman C
    13. Badal K
    14. Garcia BA
    15. Callahan MK
    16. Yuan J
    17. Martins VR
    18. Skog J
    19. Kaplan RN
    20. Brady MS
    21. Wolchok JD
    22. Chapman PB
    23. Kang Y
    24. Bromberg J
    25. Lyden D

    (2012) Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET

    Nature Medicine 18:883–891.

    https://doi.org/10.1038/nm.2753

    • PubMed
    • Google Scholar
50. 1. Peinado H
    2. Zhang H
    3. Matei IR
    4. Costa-Silva B
    5. Hoshino A
    6. Rodrigues G
    7. Psaila B
    8. Kaplan RN
    9. Bromberg JF
    10. Kang Y
    11. Bissell MJ
    12. Cox TR
    13. Giaccia AJ
    14. Erler JT
    15. Hiratsuka S
    16. Ghajar CM
    17. Lyden D

    (2017) Pre-metastatic niches: organ-specific homes for metastases

    Nature Reviews Cancer 17:302–317.

    https://doi.org/10.1038/nrc.2017.6

    • PubMed
    • Google Scholar
51. 1. Plebanek MP
    2. Angeloni NL
    3. Vinokour E
    4. Li J
    5. Henkin A
    6. Martinez-Marin D
    7. Filleur S
    8. Bhowmick R
    9. Henkin J
    10. Miller SD
    11. Ifergan I
    12. Lee Y
    13. Osman I
    14. Thaxton CS
    15. Volpert OV

    (2017) Pre-metastatic Cancer exosomes induce immune surveillance by patrolling monocytes at the metastatic niche

    Nature Communications 8:e01433-3.

    https://doi.org/10.1038/s41467-017-01433-3

    • Google Scholar
52. 1. Scott SA
    2. Selvy PE
    3. Buck JR
    4. Cho HP
    5. Criswell TL
    6. Thomas AL
    7. Armstrong MD
    8. Arteaga CL
    9. Lindsley CW
    10. Brown HA

    (2009) Design of isoform-selective phospholipase D inhibitors that modulate cancer cell invasiveness

    Nature Chemical Biology 5:108–117.

    https://doi.org/10.1038/nchembio.140

    • Google Scholar
53. 1. Scott CC
    2. Vacca F
    3. Gruenberg J

    (2014) Endosome maturation, transport and functions

    Seminars in Cell & Developmental Biology 31:2–10.

    https://doi.org/10.1016/j.semcdb.2014.03.034

    • Google Scholar
54. 1. Szklarczyk D
    2. Gable AL
    3. Lyon D
    4. Junge A
    5. Wyder S
    6. Huerta-Cepas J
    7. Simonovic M
    8. Doncheva NT
    9. Morris JH
    10. Bork P
    11. Jensen LJ
    12. Mering CV

    (2019) STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets

    Nucleic Acids Research 47:D607–D613.

    https://doi.org/10.1093/nar/gky1131

    • PubMed
    • Google Scholar
55. 1. Taira E
    2. Takaha N
    3. Taniura H
    4. Kim CH
    5. Miki N

    (1994) Molecular cloning and functional expression of Gicerin, a novel cell adhesion molecule that binds to neurite outgrowth factor

    Neuron 12:861–872.

    https://doi.org/10.1016/0896-6273(94)90338-7

    • PubMed
    • Google Scholar
56. 1. Taira E
    2. Kohama K
    3. Tsukamoto Y
    4. Okumura S
    5. Miki N

    (2005) Gicerin/CD146 is involved in neurite extension of NGF-treated PC12 cells

    Journal of Cellular Physiology 204:632–637.

    https://doi.org/10.1002/jcp.20365

    • PubMed
    • Google Scholar
57. 1. Timpson P
    2. Mcghee EJ
    3. Erami Z
    4. Nobis M
    5. Quinn JA
    6. Edward M
    7. Anderson KI

    (2011) Organotypic collagen I assay: a malleable platform to assess cell behaviour in a 3-Dimensional context

    Journal of Visualized Experiments 13:e3089.

    https://doi.org/10.3791/3089

    • Google Scholar
58. 1. Tominaga N
    2. Kosaka N
    3. Ono M
    4. Katsuda T
    5. Yoshioka Y
    6. Tamura K
    7. Lötvall J
    8. Nakagama H
    9. Ochiya T

    (2015) Brain metastatic Cancer cells release microRNA-181c-containing extracellular vesicles capable of destructing blood-brain barrier

    Nature Communications 6:6716.

    https://doi.org/10.1038/ncomms7716

    • PubMed
    • Google Scholar
59. 1. Treps L
    2. Edmond S
    3. Harford-Wright E
    4. Galan-Moya EM
    5. Schmitt A
    6. Azzi S
    7. Citerne A
    8. Bidère N
    9. Ricard D
    10. Gavard J

    (2016) Extracellular vesicle-transported Semaphorin3A promotes vascular permeability in glioblastoma

    Oncogene 35:2615–2623.

    https://doi.org/10.1038/onc.2015.317

    • PubMed
    • Google Scholar
60. 1. van Niel G
    2. D'Angelo G
    3. Raposo G

    (2018) Shedding light on the cell biology of extracellular vesicles

    Nature Reviews Molecular Cell Biology 19:213–228.

    https://doi.org/10.1038/nrm.2017.125

    • PubMed
    • Google Scholar
61. 1. Vennin C
    2. Chin VT
    3. Warren SC
    4. Lucas MC
    5. Herrmann D
    6. Magenau A
    7. Melenec P
    8. Walters SN
    9. Del Monte-Nieto G
    10. Conway JR
    11. Nobis M
    12. Allam AH
    13. McCloy RA
    14. Currey N
    15. Pinese M
    16. Boulghourjian A
    17. Zaratzian A
    18. Adam AA
    19. Heu C
    20. Nagrial AM
    21. Chou A
    22. Steinmann A
    23. Drury A
    24. Froio D
    25. Giry-Laterriere M
    26. Harris NL
    27. Phan T
    28. Jain R
    29. Weninger W
    30. McGhee EJ
    31. Whan R
    32. Johns AL
    33. Samra JS
    34. Chantrill L
    35. Gill AJ
    36. Kohonen-Corish M
    37. Harvey RP
    38. Biankin AV
    39. Evans TR
    40. Anderson KI
    41. Grey ST
    42. Ormandy CJ
    43. Gallego-Ortega D
    44. Wang Y
    45. Samuel MS
    46. Sansom OJ
    47. Burgess A
    48. Cox TR
    49. Morton JP
    50. Pajic M
    51. Timpson P
    52. Australian Pancreatic Cancer Genome Initiative (APGI)

    (2017) Transient tissue priming via ROCK inhibition uncouples pancreatic Cancer progression, sensitivity to chemotherapy, and metastasis

    Science Translational Medicine 9:eaai8504.

    https://doi.org/10.1126/scitranslmed.aai8504

    • PubMed
    • Google Scholar
62. 1. Vitale N
    2. Mawet J
    3. Camonis J
    4. Regazzi R
    5. Bader M-F
    6. Chasserot-Golaz S

    (2005) The small GTPase RalA controls exocytosis of large dense core secretory granules by interacting with ARF6-dependent phospholipase D1

    Journal of Biological Chemistry 280:29921–29928.

    https://doi.org/10.1074/jbc.M413748200

    • Google Scholar
63. 1. Wang Q
    2. Ni Q
    3. Wang X
    4. Zhu H
    5. Wang Z
    6. Huang J

    (2015) High expression of RAB27A and TP53 in pancreatic Cancer predicts poor survival

    Medical Oncology 32:372.

    https://doi.org/10.1007/s12032-014-0372-2

    • PubMed
    • Google Scholar
64. 1. Wang D
    2. Sun H
    3. Wei J
    4. Cen B
    5. DuBois RN

    (2017) CXCL1 Is Critical for Premetastatic Niche Formation and Metastasis in Colorectal Cancer

    Cancer Research 77:3655–3665.

    https://doi.org/10.1158/0008-5472.CAN-16-3199

    • Google Scholar
65. 1. Wang Z
    2. Yan X

    (2013) CD146, a multi-functional molecule beyond adhesion

    Cancer Letters 330:150–162.

    https://doi.org/10.1016/j.canlet.2012.11.049

    • Google Scholar
66. 1. Wieczorek S
    2. Combes F
    3. Lazar C
    4. Giai Gianetto Q
    5. Gatto L
    6. Dorffer A
    7. Hesse A-M
    8. Couté Y
    9. Ferro M
    10. Bruley C
    11. Burger T

    (2017) DAPAR & ProStaR: software to perform statistical analyses in quantitative discovery proteomics

    Bioinformatics 33:135–136.

    https://doi.org/10.1093/bioinformatics/btw580

    • Google Scholar
67. 1. Yan C
    2. Liu D
    3. Li L
    4. Wempe MF
    5. Guin S
    6. Khanna M
    7. Meier J
    8. Hoffman B
    9. Owens C
    10. Wysoczynski CL
    11. Nitz MD
    12. Knabe WE
    13. Ahmed M
    14. Brautigan DL
    15. Paschal BM
    16. Schwartz MA
    17. Jones DN
    18. Ross D
    19. Meroueh SO
    20. Theodorescu D

    (2014) Discovery and characterization of small molecules that target the GTPase ral

    Nature 515:443–447.

    https://doi.org/10.1038/nature13713

    • PubMed
    • Google Scholar
68. 1. Yan C
    2. Theodorescu D

    (2018) RAL GTPases: biology and potential as therapeutic targets in Cancer

    Pharmacological Reviews 70:1–11.

    https://doi.org/10.1124/pr.117.014415

    • PubMed
    • Google Scholar
69. 1. Yang Z
    2. Shi J
    3. Xie J
    4. Wang Y
    5. Sun J
    6. Liu T
    7. Zhao Y
    8. Zhao X
    9. Wang X
    10. Ma Y
    11. Malkoc V
    12. Chiang C
    13. Deng W
    14. Chen Y
    15. Fu Y
    16. Kwak KJ
    17. Fan Y
    18. Kang C
    19. Yin C
    20. Rhee J
    21. Bertani P
    22. Otero J
    23. Lu W
    24. Yun K
    25. Lee AS
    26. Jiang W
    27. Teng L
    28. Kim BYS
    29. Lee LJ

    (2020) Large-scale generation of functional mRNA-encapsulating exosomes via cellular nanoporation

    Nature Biomedical Engineering 4:69–83.

    https://doi.org/10.1038/s41551-019-0485-1

    • Google Scholar
70. 1. Yue S
    2. Mu W
    3. Erb U
    4. Zöller M

    (2015) The tetraspanins CD151 and Tspan8 are essential exosome components for the crosstalk between Cancer initiating cells and their surrounding

    Oncotarget 6:2366–2384.

    https://doi.org/10.18632/oncotarget.2958

    • PubMed
    • Google Scholar
71. 1. Zago G
    2. Veith I
    3. Singh MK
    4. Fuhrmann L
    5. De Beco S
    6. Remorino A
    7. Takaoka S
    8. Palmeri M
    9. Berger F
    10. Brandon N
    11. El Marjou A
    12. Vincent-Salomon A
    13. Camonis J
    14. Coppey M
    15. Parrini MC

    (2018) RalB directly triggers invasion downstream ras by mobilizing the wave complex

    eLife 7:e40474.

    https://doi.org/10.7554/eLife.40474

    • PubMed
    • Google Scholar
72. 1. Zeng GF
    2. Cai SX
    3. Wu GJ

    (2011) Up-regulation of METCAM/MUC18 promotes motility, invasion, and tumorigenesis of human breast Cancer cells

    BMC Cancer 11:113.

    https://doi.org/10.1186/1471-2407-11-113

    • PubMed
    • Google Scholar
73. 1. Zeng Q
    2. Li W
    3. Lu D
    4. Wu Z
    5. Duan H
    6. Luo Y
    7. Feng J
    8. Yang D
    9. Fu L
    10. Yan X

    (2012) CD146, an epithelial-mesenchymal transition inducer, is associated with triple-negative breast Cancer

    PNAS 109:1127–1332.

    https://doi.org/10.1073/pnas.1111053108

    • PubMed
    • Google Scholar
74. 1. Zhou W
    2. Fong MY
    3. Min Y
    4. Somlo G
    5. Liu L
    6. Palomares MR
    7. Yu Y
    8. Chow A
    9. O'Connor ST
    10. Chin AR
    11. Yen Y
    12. Wang Y
    13. Marcusson EG
    14. Chu P
    15. Wu J
    16. Wu X
    17. Li AX
    18. Li Z
    19. Gao H
    20. Ren X
    21. Boldin MP
    22. Lin PC
    23. Wang SE

    (2014) Cancer-secreted miR-105 destroys vascular endothelial barriers to promote metastasis

    Cancer Cell 25:501–515.

    https://doi.org/10.1016/j.ccr.2014.03.007

    • PubMed
    • Google Scholar

Acceptance summary:

Overall, we find that this a very complete and thorough study of the role of 2 small GTPases, RalA and RalB, in extracellular vesicle (EV) release and breast carcinoma progression. We believe the main advance of this manuscript is to describe a signaling role for the GTPases RalA and RalB in regulating phospholipase D (PLD) and multivesicular body function to regulate exosome biogenesis and organ-targeting of exosomes, with specific relevance to lung cancer metastasis. Notably, the work has clinical relevance, given the finding that RalA, RalB and CD146 levels are indicators poor prognosis for breast cancer patients. Overall these are interesting observations that should impact multiple fields of study.

Decision letter after peer review:

Thank you for submitting your article "Ral GTPases promote metastasis by controlling biogenesis and organ colonization of exosomes" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Suzanne Pfeffer as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Aaron Hobbs (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

As the editors have judged that your manuscript is of interest, but as described below that additional experiments are required before it is published, we would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). First, because many researchers have temporarily lost access to the labs, we will give authors as much time as they need to submit revised manuscripts. We are also offering, if you choose, to post the manuscript to bioRxiv (if it is not already there) along with this decision letter and a formal designation that the manuscript is "in revision at eLife". Please let us know if you would like to pursue this option. (If your work is more suitable for medRxiv, you will need to post the preprint yourself, as the mechanisms for us to do so are still in development.)

Summary:

The role of extracellular vesicles (EVs) such as exosomes as factors potentiating metastasis by solid tumors has attracted considerable recent interest. The article by Ghoroghi et al. is a very complete and thorough study of the role of 2 small GTPases, RalA and RalB, in extracellular vesicle (EV) release and breast carcinoma progression. The main advance of this manuscript is to describe a signaling role for the GTPases RalA and RalB in regulating phospholipase D (PLD) and multivesicular body function to regulate exosome biogenesis. They also show that RalA and RalB regulate expression of MCAM/CD146 on EVs, and that reduced levels of CD146 on EVs affects efficiency of lung cancer metastasis. Finally, they show RalA, RalB and CD146 levels are indicators poor prognosis for breast cancer patients. Overall these are interesting observations with clinical relevance. The study is extremely carefully performed, in general, with appropriate controls and conclusions. There are a limited number of weaknesses that need to be addressed.

Essential revisions:

1) Figure 3C shows that RALA and RALB have functions where they do not always act in series with each other. They may for exosome excretion, but not necessarily for proliferation. For Figure 3—figure supplement 1C, this experiment needs to be longer than 72 hours for a proliferation assay to be truly convincing. While interesting that proliferation results differ between mice and growth on plastic, a 10 day growth experiment would lead to greater conviction that this is a sustained difference. Figure 3D better agrees with Figure 3—figure supplement 1C, where loss of RALB gives greater proliferation over control, albeit not as great as RALA loss. Was Figure 3D performed at 12 days? The variable time points make it difficult to assess the role of RALA/B on growth.

2) Figure 4 carefully describes why RALA/B deficient EVs fail to prime regions for metastasis. As the EVs are not directed to those locations, the contents fail to increase permeability of the regions. However, statistically significant does not mean biologically meaningful; specifically, looking at Figure 4F and 4G, knockdown of RALA/B had little effect on EV internalization. One question that is not answered is whether RALA/B are acting in series or parallel. Concurrent depletion of both RALA and RALB in the types of experiment performed in Figure 4 would help answer this question. If depletion of both isoforms is greater than either alone, then it implies they act in parallel.

3) How are RALA/B activated in the context of 4T1 cells? The use of the RAL inhibitor in Figure 1B is most effective in cell lines without RAS mutations. Panc1 and MD231 cells have KRAS mutations, yet RALGEF inhibitors are least effective. At the concentrations used, all RALGEF activity should be inhibited. Is RAL expression levels similar between cell lines?

4) The logic for targeting PLD1 seems rushed. There are several RAL effectors, and while the data on PLD1 are compelling, it is unclear why only this effector was selected. Were other effectors tested? The Hyenne paper, which is cited throughout, provided stronger logic for connecting these pathways. The statement in the Discussion, "Our work further identifies PLD as the most likely effector acting downstream of RAL to control exosome section" is incorrect as stated. This work identified PLD1 as A downstream effector important for exosome secretion. To state that it is the most likely effector is an overstatement as no other RAL effectors were tested.

5) The use of either PLD inhibitor at 10 μm will have significant inhibition of both PLD1 and PLD2 regardless which inhibitor is being tested, as well as additional off-target effects. Given both inhibitors have nanomolar IC50 values for their specific PLD isoform and low μM IC50 values for the related isoform, these experiments are inconclusive as presented. This section of the text should be reworked to account for the lack of isoform selectivity at the concentrations of inhibitor used.

a) siRNA or shRNA could be used to validate the PLD1/2 inhibitor data.

6) In all cases when using inhibitors, there are no western blots or additional validation showing target inhibition in any cell line. In this case, the use of such high concentrations of inhibitor likely resulted in inhibition of the target proteins. However, such high concentrations often result in off-target effects. Determining the lowest efficacious dose to inhibit target function should have been performed and used throughout the experiments.

7) Would loss of MCAM expression, via direct knockdown, result in EVs that are unable to permeabilize HUVEC cells, as in Figure 4A/B? Alternatively, does loss of MCAM expression (or blocking via anti-CD146) result in decreased metastasis, similar to RALA/B knockdown? While Figure 5G shows a decrease in EV localization to lungs on treatment with Anti-CD146, the rate of metastatic lesion formation was not assessed when CD146 was blocked.

8) In a previous study (Hyenne et al., 2015) the authors already demonstrated a role for RalA and RalB in controlling MVB and EV secretion in the 4T1 breast cancer model, in a process evolutionarily conserved through nematodes. They also wrote a follow-up review article describing considerable evidence in the literature for RalA controling PLD and ARF6 function to control EV secretion. This diminishes the novelty and interest of the first part of the study, which could be reduced. The more novel and thought-provoking parts of the study are in the definition of the relationship between the RAL proteins and CD146, and the identification of unique properties of the exosomes produced in cells with manipulated RAL proteins (for instance, in regard to influencing exosome permeability). These could be better emphasized.

9) Almost the entire mechanistic model is based on the work of a single breast cancer cell line, 4T1, which has been described by some as triple negative. This is a significant weakness, as there may be features of that line that make it uncharacteristic of breast cancer in general. At least some of the key conclusions should be functionally confirmed in an additional cell model.

10) In addition, breast cancer cells fall into multiple different subtypes, which have different metastatic propensities and gene expression patterns. Is the functional relationship between the expression of RALA/B and CD146 in exosomes observed in just triple negative cells, or in other subtypes?

11) In Figure 5H, K-M analysis of TCGA shows a weak relationship between MCAM expression and survival. Were the tumors analyzed segregated by tumor subtype? Were data corrected for tumor stage? This is important, as the MCAM staining pattern may reflect a propensity for MCAM expression in tumors at a late stage, or subtypes with a poorer prognosis.

12) The authors provide observations suggesting that RalA and B control primarily the EV secretion pathway involving Multivesicular bodies, hence leading to exosome secretion. This is mainly demonstrated by observation of a decrease in MVBs upon knock-down of RalA/B, demonstrated by thorough electron microscopy analyses. This is correlative, rather than truly demonstrative, but the best one can do so far. In most experiments, the authors use EVs isolated by a relatively crude method, ultracentrifugation, that co-isolates non-specific components, and they do not analyse larger EVs that can be recovered at lower centrifugation speed, thus an effect of RalA/B on these non-exosomal EVs cannot be excluded. The EVs are only characterized by their number (NTA counting), which is not very precise, and not consistent with guidelines of ISEV (MISEV 2018, J Extracell Vesicles 2018, 7: 1535750).

Maybe the authors can argue that they did perform more complete analyses of their 4T1 EVs in a previous article? Did they use the EV-TRACK website to verify their experimental EV isolation and characterization set up? Of note, the authors also perform quantitative proteomic and RNomic analyses, which gives a better characterization of EVs. The protein composition and its change upon RalA/B KD could have been used to try to confirm (or not) the MVB origin of the EVs controlled by RalA or RalB, but it is not crucial for the message.

Alternatively, to demonstrate that the CD146-bearing EVs that carry the prometastatic function are bona fide exosomes, the authors could have shown its localization in the cells, upon or not RalA/B depletion, and show if a drastic change in ratio of localization in MVBs vs the PM occurs, but this would be an additional study, not necessary for the current paper.

13) The authors insist that RalA/B control exosome secretion, and discuss the bases and limitations of their demonstration properly. The summarizing schemes (Figure 2F and Figure 5I) of their model show the release of RalA/B-depent pro-metastatic exosomes, and RalA/B-independent exosomes which are not pro-metastatic. However, EVs that are released in the absence of RalA/B could instead be formed at the plasma membrane, and correspond to ectosomes. Nothing in this study demonstrates the origin of RalA/B-independent EVs, thus PM-derived EVs should be represented in the scheme.

Essential revisions:

1) Figure 3C shows that RALA and RALB have functions where they do not always act in series with each other. They may for exosome excretion, but not necessarily for proliferation. For Figure 3—figure supplement 1C, this experiment needs to be longer than 72 hours for a proliferation assay to be truly convincing. While interesting that proliferation results differ between mice and growth on plastic, a 10 day growth experiment would lead to greater conviction that this is a sustained difference. Figure 3D better agrees with Figure 3—figure supplement 1C, where loss of RALB gives greater proliferation over control, albeit not as great as RALA loss. Was Figure 3D performed at 12 days? The variable time points make it difficult to assess the role of RALA/B on growth.

As the reviewer pointed out, our initial manuscript reports that in vivo tumor growth (Figure 3D; at 41 days) differs form in vitro proliferation assays (Figure 3—figure supplement 1C; over 72h). In details, while loss of RalA consistently shows a trend to increased growth in both conditions, loss of RalB shows a slight increase in cell proliferation in vitro but a reduced tumor growth in vivo. To better understand this discrepancy, we carefully analyzed the cell cycle of shRal cells by flow cytometry. We observed a significant decrease in the proportion of cells in the sub-G1 phase in shRalA cells compared to control, potentially explaining the overall increased proliferation observed in this condition. However, the proportion of cells in each phase of the cycle was similar in shRalB and shCtl cells. Overall, this analysis further illustrates the importance of in vivo experiments, and the difficulty to reconcile in vitro with in vivo situations (where microenvironment and non-cell intrinsic parameters are likely to significantly alter these behaviors and impact tumor growth). We have now included the cell cycle analysis in Figure 3—figure supplement 1D and mentioned it in the manuscript.

2) Figure 4 carefully describes why RALA/B deficient EVs fail to prime regions for metastasis. As the EVs are not directed to those locations, the contents fail to increase permeability of the regions. However, statistically significant does not mean biologically meaningful; specifically, looking at Figure 4F and 4G, knockdown of RALA/B had little effect on EV internalization. One question that is not answered is whether RALA/B are acting in series or parallel. Concurrent depletion of both RALA and RALB in the types of experiment performed in Figure 4 would help answer this question. If depletion of both isoforms is greater than either alone, then it implies they act in parallel.

We would like to split the reviewer’s comment in 2 parts:

1) how do we explain the relatively modest defects in shRal EVs internalization in Figure 4F/G (in vitro-in zebrafish) compared to the (more important) effect on metastasis priming (Figure 4D) ?

2) Do RalA and RalB act in series or in parallel?

Regarding the (i) first question, we would like to insist on two aspects that are likely to be driven by the inherent differences of the experimental approaches used: first, the shRal EVs targeting defects are stronger in mice lungs (Figure 4D/E) than in vitro (Figure 4F) or in zebrafish (ZF, Figure 4E). This read-out (targeting of EVS to the lung parenchyma) is not exactly comparable to the assays used in vitro and in ZF (internalization in an endothelium). Nevertheless, the experimental approach used in such homologous model is very close to the one used for the metastatic niche priming model. Interestingly, in both cases, shRal EVs show comparably high deficiencies (60-70% decrease) which demonstrates that EV targeting/internalization in mice is directly linked to its effect on metastatic niche priming. In contrast, while assessing internalization of EVs in HUVECs or ZF endothelium offers a level of resolution that can’t be reached in mouse models, comparing these different assays of EV internalization also suggests that the mechanisms involved in the internalization of circulating EVs in mice are more complex that the one taking place in heterologous systems. We hypothesize that HUVECs or ZF take up EVs in a less specific manner and are therefore less sensitive to targeting of specific adhesion receptors. Second, Ral depletion can affect adhesive-independent properties of breast tumor EVs able to influence premetastatic niche priming. Indeed, several proteins and multiple RNAs show alteration of their expression levels in shRal EVs (Figure 5A-D). Some of these cargoes could contribute to the formation of a premetastatic niche without affecting the adhesion properties of the EVs.

The (ii) second question regarding the potential redundancy of RalA and RalB has puzzled Ral experts for more than two decades. Depending on the context and biological effect, Ral GTPases can have redundant, parallel or antagonist functions. To address this question and determine whether RalA and RalB act in series or in parallel in EV secretion and function, we first generated 4T1 cells with a double knock-down (KD) for both RalA and RalB . We then determined how their co-depletion affects EV secretion. Co-depletion of both RalA and RalB further decreases EV secretion (when compared to single RalA or RalB depletion). To better understand how they could function, we decided to analyze the phenotypes of MVBs in double KD cells by electron microscopy (based on high pressure freezing). Here again, we observed that co-depleting Ral GTPases further decreases the density of MVBs in the cytoplasm when compared to single KD. Therefore, co-depletion of Ral GTPases worsens individual phenotypes, which implies that RalA and RalB act partially in parallel (or independent) ways. However, the decrease observed does not correspond to the sum of their individual depletion, suggesting that they still act on a similar exosomal secretion pathway.

Understanding precisely how RalA and RalB cooperate in exosome secretion and function would imply a much deeper epistatic analysis. While we have now initiated such analysis, which will certainly require a significant amount of work, we hope the reviewers understand that such analysis goes beyond this study. While we are thankful to the reviewers for raising and confirming this important question, we would like to ask for their understanding as we favor a more thorough analysis and thus exclude these incomplete data from the current study. At this stage, and without an extensive study, it is yet impossible to claim any definitive conclusion, which our future work will allow to solve confidently.

3) How are RALA/B activated in the context of 4T1 cells? The use of the RAL inhibitor in Figure 1B is most effective in cell lines without RAS mutations. Panc1 and MD231 cells have KRAS mutations, yet RALGEF inhibitors are least effective. At the concentrations used, all RALGEF activity should be inhibited. Is RAL expression levels similar between cell lines?

Regarding the first question, we currently ignore how RalA/B are activated in 4T1 cells. However, we speculate that they could be activated by the RAS pathway, which is active in 4T1 cells and required for 4T1 cells aggressivity. Indeed, 4T1 cells express KRAS and HRAS (Xi et al., 2017; Hu et al., Nuc. Acid Res. 2019) and the RAS pathway is active in 4T1 cells, as assessed by BRAF and erk phosphorylation status (Phan et al., 2013). Importantly, both 4T1 migration abilities and tumor progression depend on KRAS expression levels (Xi et al., 2017; Booth et al., 2017).

To address the reviewer’s last question, we measured RAL expression levels in several cell lines by western blot (Author response image 1). Interestingly, the 4T1 cell lines thoroughly used in our study displays the highest Ral expression levels. In addition, as pointed out by the reviewer, a higher expression of RalA could explain that the effect of Ral inhibitors on EV secretion is higher in 4T1 cells than in Panc1 or MDAMD231 cells.

Author response image 1

Download asset Open asset

4) The logic for targeting PLD1 seems rushed. There are several RAL effectors, and while the data on PLD1 are compelling, it is unclear why only this effector was selected. Were other effectors tested? The Hyenne paper, which is cited throughout, provided stronger logic for connecting these pathways. The statement in the Discussion, "Our work further identifies PLD as the most likely effector acting downstream of RAL to control exosome section" is incorrect as stated. This work identified PLD1 as A downstream effector important for exosome secretion. To state that it is the most likely effector is an overstatement as no other RAL effectors were tested.

When dissecting the mechanistic effects of RalA/B on exosome secretion, we have indeed used a strategy that is based on published evidence for targeting PLDs as Ral effectors, in the context of exosome secretion. Thus, while this reviewer is partly correct since we have only tested PLDs and not other Ral effectors, such decision had been carefully thought based on arguments that are described in the manuscript and summarized here.

Among all known Ral effectors, PLDs had been uniquely connected to exosome secretion, through the work of Pascal Zimmermann’s group (Ghossoub et al., 2014), and were therefore the most obvious candidates to be targeted. Interestingly, recent independent studies from Borel and colleagues confirmed that PLDs affect EV secretion levels in a Phosphatidic Acid dependent manner in prostate cancer cells (Borel et al., BBA 2020). While other Ral effectors (such as RalBP1, Filamin or exocyst) have, to our knowledge, never been shown to alter EV secretion, they are indeed potential additional candidates whose targeting would require additional work that would go beyond this study.

Taking this into account, we have now modified the way we introduce PLD1 and toned down our conclusions.

5) The use of either PLD inhibitor at 10 μm will have significant inhibition of both PLD1 and PLD2 regardless which inhibitor is being tested, as well as additional off-target effects. Given both inhibitors have nanomolar IC50 values for their specific PLD isoform and low μM IC50 values for the related isoform, these experiments are inconclusive as presented. This section of the text should be reworked to account for the lack of isoform selectivity at the concentrations of inhibitor used.

a) siRNA or shRNA could be used to validate the PLD1/2 inhibitor data.

We have now added a sentence in the Discussion to emphasize the potential lack of isoform selectivity. Nevertheless, we would like to emphasize that the concentration that have been used had been published before (Figure 6 of Ghossoub et al., 2014) when demonstrating that PLD2 is specifically involved in ARF6 exosome biogenesis.

Furthermore, as suggested by the reviewer, we undertook experiments with siRNA against PLD1/2. Unfortunately, these experiments failed as we did not manage to specifically detect PLD isoforms by western blotting using commercially available antibodies (data not shown).

Finally, to further demonstrate that we can indeed observe defects related to each PLD isoform, we subjected cells to decreasing concentrations of PLD inhibitors (in order to reach nanomolar IC50 values) and observed, as expected, a dose-dependent effect of the inhibition on EV secretion levels (Author response image 2, see below, #6). These results show that EV secretion levels can be tuned with nanomolar concentrations of PLD inhibitors. Thus, although both PLDs are involved in controlling EV secretion levels (validated here), only PLD1 localization to endosome is perturbed in absence of RalA/B and thus responsible of Ral dependent EV secretion. Nevertheless, in the Discussion, we acknowledge the possibility of off target effects at high inhibitor concentrations.

Author response image 2

Download asset Open asset

6) In all cases when using inhibitors, there are no western blots or additional validation showing target inhibition in any cell line. In this case, the use of such high concentrations of inhibitor likely resulted in inhibition of the target proteins. However, such high concentrations often result in off-target effects. Determining the lowest efficacious dose to inhibit target function should have been performed and used throughout the experiments.

While we appreciate the fact that high concentrations might lead to off-target effects, we would like to emphasize again that these concentrations (for both Ral and PLD inhibition) have been used and published in the past (Yan et al., 2014; Ghossoub et al., 2014) and demonstrated their efficacy and selectivity. As suggested by the reviewer, we now used different concentrations of PLD inhibitors (10nM, 100nM, 1µM and 10µM) and analyzed their effect on EV secretion by NTA. As shown on Author response image 2, there is a clear dose dependent effect, using both inhibitors. These results show that EV secretion levels can be tuned with nanomolar concentrations of PLD inhibitors. Thus, although both PLDs are involved in controlling EV secretion levels (validated here), only PLD1 localization to endosome is perturbed in absence of RalA/B and thus responsible of Ral-dependent EV secretion.

7) Would loss of MCAM expression, via direct knockdown, result in EVs that are unable to permeabilize HUVEC cells, as in Figure 4A/B? Alternatively, does loss of MCAM expression (or blocking via anti-CD146) result in decreased metastasis, similar to RALA/B knockdown? While Figure 5G shows a decrease in EV localization to lungs on treatment with Anti-CD146, the rate of metastatic lesion formation was not assessed when CD146 was blocked.

We thank the reviewer for raising this point and agree that such aspect is determinant to validate our model. In order to test whether blocking CD146 would tune tumor metastasis, we set out a challenging experimental approach allowing to explore the role of EV-bound CD146 in metastatic niches priming. To do so, we set up a classical “priming” experiment where 4T1 EVs are blocked with the anti-CD146 antibody. As shown in Figure 5H and Figure 4C, pre-injection of 4T1 EVs increases lung metastasis of 4T1 cells compared to PBS treatment. However, when pre-injecting a similar amount of 4T1 EVs treated with the blocking anti-CD146 antibody, we observed a striking reduction in lung metatstatic burden, nearing PBS-treated levels (14 mice per group in two independent experiments). This experiment solidly demonstrates that CD146, exposed at the surface of EVs and whose enrichment is tuned by RalA/B, is a key molecular determinant of EV-mediated metastasis. Together with our previous results, it strongly suggests that CD146 molecules present at the surface of tumor EVs favor lung colonization by EVs, likely by favoring binding to the endothelium, thereby allowing pre-metastatic niche formation. As this result provides a new and solid mechanistic insight into the molecular machineries tuned by RalA/B levels and thus consolidates and improves our earlier observations, we decided to add it in Figure 5I of our manuscript and discuss it in the main text.

8) In a previous study (Hyenne et al., 2015) the authors already demonstrated a role for RalA and RalB in controlling MVB and EV secretion in the 4T1 breast cancer model, in a process evolutionarily conserved through nematodes. They also wrote a follow-up review article describing considerable evidence in the literature for RalA controling PLD and ARF6 function to control EV secretion. This diminishes the novelty and interest of the first part of the study, which could be reduced. The more novel and thought-provoking parts of the study are in the definition of the relationship between the RAL proteins and CD146, and the identification of unique properties of the exosomes produced in cells with manipulated RAL proteins (for instance, in regard to influencing exosome permeability). These could be better emphasized.

We can’t agree more with this reviewer when he claims that the most novel aspect of our study is the relationship between Ral and CD146 levels and role on EVs. In turn, and with the newly added mechanistic data (Figure 5H), we’ve further put the emphasis on this novel aspect of our work.

Moreover, we wish to maintain the second additional exciting part of this work, which is the careful molecular and cell biological dissection of Ral function in exosome secretion. Our previous description of PLD involvement downstream of Ral (Hyenne at al, 15 + 16) was mostly speculative while, we now provide, within this manuscript, experimental evidences of the molecular links between Ral and PLD1. In the light of the new experiments, we have now modified the Abstract accordingly and have incorporated our latest results on CD146 in priming metastasis as Figure 5I (See Comment #7).

9) Almost the entire mechanistic model is based on the work of a single breast cancer cell line, 4T1, which has been described by some as triple negative. This is a significant weakness, as there may be features of that line that make it uncharacteristic of breast cancer in general. At least some of the key conclusions should be functionally confirmed in an additional cell model.

With all due respect, we wish to point out that we also had observed that Ral inhibition leads to a significant decrease in EV secretion levels in pancreatic, melanoma and breast cancer cells (Figure 1B). Thus, together with our previous work performed in C. elegans (Hyenne et al., 2015), this suggests that the role of Ral in EV secretion can be extended is likely to be universal or can be at least extended to other models and is conserved in evolution. In addition, additional evidence had been obtained when correlating Ral levels with metastasis in our cohorts (Figure 3A, B) or via access to TCGA public databases (that contain various breast cancer subtypes, see #10), which further demonstrates that high Ral levels (or CD146) correlate with metastasis and poor survival. Thus, altogether, this demonstrates that our discovery can be extended to other breast cancer subtypes. While we appreciate that our in vivo mouse approaches are restricted to the TNBC 4T1 model, such experiments are heavy and challenging and should be seen as an advantage, rather than a pitfall of our work. Further in vivo experiments performed by us (or others) will provide additional insights on the universality of such pathway, but we humbly feel that this beyond the scope of this study. Finally, we now provide additional demonstration that Ral inhibition is effective in other model, including three other breast cancer cell lines (see Point 10; Figure 1—figure supplement 1C).

10) In addition, breast cancer cells fall into multiple different subtypes, which have different metastatic propensities and gene expression patterns. Is the functional relationship between the expression of RALA/B and CD146 in exosomes observed in just triple negative cells, or in other subtypes?

To address this question, we analyzed the role of Ral GTPases in EV secretion by three other breast cancer cell lines with distinct characteristics: MCF7 (human, estrogen receptor⁺, progesterone receptor⁺, HER2^-), SKBR3 (human, HER2⁺) and D2A1 (murine, estrogen receptor⁺, HER2⁺). In all cases, we found that Ral inhibition with RBC8 drug induces a significant decrease in EV secretion (Figure 1—figure supplement 1C). This further demonstrates that all breast cancer types release Ral dependent EVs. This new data is now added as Figure 1—figure supplement 1C and the associated description added in the manuscript accordingly.

11) In Figure 5H, K-M analysis of TCGA shows a weak relationship between MCAM expression and survival. Were the tumors analyzed segregated by tumor subtype? Were data corrected for tumor stage? This is important, as the MCAM staining pattern may reflect a propensity for MCAM expression in tumors at a late stage, or subtypes with a poorer prognosis.

We followed the reviewer recommendation and re-analyzed TCGA data, segregating the patients by tumor subtypes and by tumor stages (Author response image 3). This analysis indeed revealed differences between tumor subtypes. High CD146 expression tends to correlate with decreased survival probability in three subtypes (basal, Her2+ and Luminal A) and with increased survival in one subtype (Luminal B). However, statistical significance was only reached for the luminal A subtype. Importantly, high CD146 expression correlates with a decreased survival probability in stage III breast cancer, while stages I and II are inconclusive, due to low patients number.

We performed similar analysis for RalA and RalB and found that elevated expression of both GTPases show a tendency to correlate with decreased survival in all breast cancer subtypes and stages (with the notable exceptions of basal subtype where high RalB correlates with enhanced survival and Luminal where no clear correlation is observed for RalB). In conclusion, at this stage, while Ral levels clearly correlate with poor survival, such correlation for MCAM/CD146 will require additional work.

Author response image 3

Download asset Open asset

12) The authors provide observations suggesting that RalA and B control primarily the EV secretion pathway involving Multivesicular bodies, hence leading to exosome secretion. This is mainly demonstrated by observation of a decrease in MVBs upon knock-down of RalA/B, demonstrated by thorough electron microscopy analyses. This is correlative, rather than truly demonstrative, but the best one can do so far. In most experiments, the authors use EVs isolated by a relatively crude method, ultracentrifugation, that co-isolates non-specific components, and they do not analyse larger EVs that can be recovered at lower centrifugation speed, thus an effect of RalA/B on these non-exosomal EVs cannot be excluded. The EVs are only characterized by their number (NTA counting), which is not very precise, and not consistent with guidelines of ISEV (MISEV 2018, J Extracell Vesicles 2018, 7: 1535750).

Maybe the authors can argue that they did perform more complete analyses of their 4T1 EVs in a previous article? Did they use the EV-TRACK website to verify their experimental EV isolation and characterization set up? Of note, the authors also perform quantitative proteomic and RNomic analyses, which gives a better characterization of EVs. The protein composition and its change upon RalA/B KD could have been used to try to confirm (or not) the MVB origin of the EVs controlled by RalA or RalB, but it is not crucial for the message.

Alternatively, to demonstrate that the CD146-bearing EVs that carry the prometastatic function are bona fide exosomes, the authors could have shown its localization in the cells, upon or not RalA/B depletion, and show if a drastic change in ratio of localization in MVBs vs the PM occurs, but this would be an additional study, not necessary for the current paper.

As pointed by this reviewer, we indeed partially characterized 4T1 EVs in a previous study (Hyenne et al., 2015). More precisely, we characterized and quantified 4T1 EVs (secreted by control, shRalA and shRalB cells) by electron microscopy and western blots using different markers (TSG101, CD63, Hsc70 and Alix). In the present study, we used a complementary quantification method, Nanoparticle Tracking Analysis, to assess EV size and concentrations, and characterized the content of these EVs by mass spectrometry, RNA sequencing and lipidomics (measuring PC and PA levels). Therefore, we humbly disagree with the reviewer and rather think that we deeply characterized 4T1 EVs secreted by control and Ral depleted cells. Our data were indeed submitted to EV-Track during the revision process. Regarding the MVB origin of EVs secreted in a Ral-dependent manner, we agree with the reviewer that our observations are mostly correlative. We cannot fully exclude that some of these EVs could originate from the plasma membrane and we now acknowledge this in the manuscript.

However, the mass spectrometry analysis revealed a large proportion of endosomal proteins, and this aspect is now further emphasized in the Results section. Following the reviewer’s recommendations, we now carefully assessed the localization of CD146 in 4T1 cells by immunofluorescence. We found that CD146 is present both at the plasma membrane and in CD63+ MVBs/late endosomes. Since we had shown that both RalA and RalB colocalize with CD63 (Figure 1C), this suggests that CD146 present on EVs could originate from MVBs, where its loading would depend on Ral GTPases. However, as pointed by the reviewer, describing such trafficking in details would require an independent study. We have added CD146 endosomal localization in 4T1 cells in Figure 5—figure supplement 2C.

13) The authors insist that RalA/B control exosome secretion, and discuss the bases and limitations of their demonstration properly. The summarizing schemes (Figure 2F and Figure 5I) of their model show the release of RalA/B-depent pro-metastatic exosomes, and RalA/B-independent exosomes which are not pro-metastatic. However, EVs that are released in the absence of RalA/B could instead be formed at the plasma membrane, and correspond to ectosomes. Nothing in this study demonstrates the origin of RalA/B-independent EVs, thus PM-derived EVs should be represented in the scheme.

We agree with the reviewer and modified our scheme in Figure 2F accordingly. However, for the sake of clarity of our final take-home message, we would appreciate if the reviewers agree that we focus on exosome secretion (Ral dependent and Ral independent) in the final model from Figure 5J, and thus exclude Ral-independent microvesicles from this graph (as these have not been tested in our study).

Author details

1. Shima Ghoroghi

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-5715-4559

2. Benjamin Mary

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0815-842X

3. Annabel Larnicol

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

4. Nandini Asokan

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0505-6487

5. Annick Klein

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

6. Naël Osmani

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1195-753X

7. Ignacio Busnelli

   1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
   2. Université de Strasbourg, Strasbourg, France
   3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

8. François Delalande

   Laboratoire de Spectrométrie de Masse BioOrganique (LSMBO), IPHC UMR 7178, CNRS, Université de Strasbourg, Strasbourg, France

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3590-9874

9. Nicodème Paul

   1. Université de Strasbourg, Strasbourg, France
   2. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France
   3. INSERM UMR_S1109, Genomax, Strasbourg, France

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4680-3012

10. Sébastien Halary

    CNRS, UMR 7245 MCAM, Muséum National d’Histoire Naturelle de Paris, Paris, France

    Contribution

    Data curation, Formal analysis

    Competing interests

    No competing interests declared

11. Frédéric Gros

    1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-6252-4323

12. Laetitia Fouillen

    Université de Bordeaux, CNRS, Laboratoire de Biogenèse Membranaire, UMR 5200, Villenave d'Ornon, France

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-1204-9296

13. Anne-Marie Haeberle

    Centre National de la Recherche Scientifique, Université de Strasbourg, Institut des Neurosciences Cellulaires et Intégratives, Strasbourg, France

    Contribution

    Investigation

    Competing interests

    No competing interests declared

14. Cathy Royer

    Plateforme Imagerie In Vitro, CNRS UPS 3156, Strasbourg, France

    Contribution

    Investigation

    Competing interests

    No competing interests declared

15. Coralie Spiegelhalter

    IGBMC Imaging Center CNRS (UMR7104)/ INSERM (U1258)/ Université de Strasbourg, Illkirch, France

    Contribution

    Investigation

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-6372-1647

16. Gwennan André-Grégoire

    1. Team SOAP, CRCINA, INSERM, CNRS, Université de Nantes, Université d’Angers, Nantes, France
    2. Integrated Center for Oncology, ICO, St-Herblain, France

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

17. Vincent Mittelheisser

    1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France
    4. Nanotranslational laboratory, Institut de Cancérologie Strasbourg Europe, Strasbourg, France

    Contribution

    Investigation

    Competing interests

    No competing interests declared

18. Alexandre Detappe

    1. Nanotranslational laboratory, Institut de Cancérologie Strasbourg Europe, Strasbourg, France
    2. Équipe de synthèse pour l’analyse (SynPA), Institut Pluridisciplinaire Hubert Curien (IPHC), UMR7178, CNRS/Université de Strasbourg, Strasbourg, France

    Contribution

    Supervision

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-9364-1621

19. Kendelle Murphy

    1. Vincent's Clinical School, Faculty of Medicine, University of New South Wales, Sydney, Australia
    2. The Kinghorn Cancer Centre, Garvan Institute of Medical Research, Sydney, Australia

    Contribution

    Formal analysis, Investigation

    Competing interests

    No competing interests declared

20. Paul Timpson

    1. Vincent's Clinical School, Faculty of Medicine, University of New South Wales, Sydney, Australia
    2. The Kinghorn Cancer Centre, Garvan Institute of Medical Research, Sydney, Australia

    Contribution

    Supervision

    Competing interests

    No competing interests declared

21. Raphaël Carapito

    1. Université de Strasbourg, Strasbourg, France
    2. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France
    3. INSERM UMR_S1109, Genomax, Strasbourg, France

    Contribution

    Formal analysis, Supervision

    Competing interests

    No competing interests declared

22. Marcel Blot-Chabaud

    C2VN, INSERM 1263, Inrae 1260, Aix-Marseille Université, Marseille, France

    Contribution

    Resources

    Competing interests

    No competing interests declared

23. Julie Gavard

    1. Team SOAP, CRCINA, INSERM, CNRS, Université de Nantes, Université d’Angers, Nantes, France
    2. Integrated Center for Oncology, ICO, St-Herblain, France

    Contribution

    Resources, Formal analysis, Supervision

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-7985-9007

24. Christine Carapito

    Laboratoire de Spectrométrie de Masse BioOrganique (LSMBO), IPHC UMR 7178, CNRS, Université de Strasbourg, Strasbourg, France

    Contribution

    Formal analysis, Supervision

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-0079-319X

25. Nicolas Vitale

    Centre National de la Recherche Scientifique, Université de Strasbourg, Institut des Neurosciences Cellulaires et Intégratives, Strasbourg, France

    Contribution

    Formal analysis, Supervision

    Competing interests

    No competing interests declared

26. Olivier Lefebvre

    1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

    Contribution

    Data curation, Formal analysis, Supervision, Investigation

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-9130-9174

27. Jacky G Goetz

    1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France

    Contribution

    Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Writing - original draft, Project administration, Writing - review and editing

    Contributed equally with

    Vincent Hyenne

    For correspondence

    jacky.goetz@inserm.fr

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-2842-8116

28. Vincent Hyenne

    1. INSERM UMR_S1109, Tumor Biomechanics, Strasbourg, France
    2. Université de Strasbourg, Strasbourg, France
    3. Fédération de Médecine Translationnelle de Strasbourg (FMTS), Strasbourg, France
    4. CNRS SNC5055, Strasbourg, France

    Contribution

    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing - original draft, Writing - review and editing

    Contributed equally with

    Jacky G Goetz

    For correspondence

    hyenne@unistra.fr

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-1254-2814

• 3,836

  Page views

• 615

  Downloads

• 38

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.