Metastasis is defined as the spread of cancer cells from the primary site of formation to distant tissues and/or organs. Such spatiotemporal cellular dissemination that occurs through the blood and lymphatic system is considered to be responsible for the vast majority of cancer deaths worldwide [1]. A deeper understanding of metastasis might identify new mechanisms that could be intercepted, leading to new anti-cancer therapeutic strategies. The classic hypothesis of the metastatic process is based on Stephen Paget’s “Seed and Soil” model, according to which the distribution of metastases is not casual, but organ-specific: the “soil”, the proper tissue or organ environment, allows the growth of the “seed”, i.e., certain tumor cells with metastatic potential, owing to the interaction between the cancer cells and the host organ [2]. It was shown that the mechanism by which metastatic cancer cells implant and grow at distant sites involves cross-communication between them and resident stromal/immune cells, via direct physical cell–cell contact or soluble factors, secreted locally or carried systemically by nanosized extracellular membrane vesicles (abbreviated here as EVs) ^{[3][4][5][6][7][8][9][10][11][12]}[3,4,5,6,7,8,9,10,11,12].

The role attributed to EVs as mediators of intercellular communication is now emerging in the contexts of embryogenesis, tissue regeneration and the immune system, as well as in various diseases, notably in cancer, where EVs can favor the formation of metastases. Different types of EVs have been described, the most common being those released after the fusion of multi-vesicular bodies with the plasma membrane, as exosomes, with a size below 120 nm, or directly released from the plasma membrane of donor cells, as ectosomes/microvesicles (100–1000 nm) ^{[9][13][14][15]}[9,18,19,20]. In addition, larger EVs such as apoptotic bodies (1–5 μm), released upon cell fragmentation during apoptotic cell death, and large oncosomes (1–10 μm), released from non-apoptotic membrane blebs of migrating cancer cells harbouring an amoeboid phenotype, have also been described ^{[16][17][18][19][20][21]}[21,22,23,24,25,26]. Another type of large EVs are migrasomes (0.5–3 μm), released upon the degradation of cell’s retraction fibers left behind by migrating cells or during membrane cell retraction ^[22][23][27,28]. The latter may transport chemokines and cytokines and thus play some role in the dissemination of cancer cells and/or interfere with the immune system ^[24][25][29,30].

Once released into a given bodily fluid, EVs can encounter other cells and transfer their contents to them. In this process, the membrane constituents of EVs and those associated with the plasma membrane of recipient cells could determine the selective targeting to a particular recipient cell type and the mechanism of uptake or internalization (reviewed in Refs. ^[15][26][20,37]). The uptake of EVs, notably those derived from cancer cells, can play a major role in the metastatic process. First, through hemo-lymphatic circulation, EVs could reach distant sites, where they can stimulate non-neoplastic stromal/immune cells to support tumor growth, thus creating the appropriate “soil” of the pre-metastatic niche (PMN) ^{[4][5][27][28]}[4,5,38,39]. Thus, EVs could induce neo-angiogenesis ^[29][30][31][40,41,42] and immunosuppression ^[32][33][34][43,44,45]. 

The impact of EVs on a given physiological process is modulated by their content. In cancer, especially in metastasis, certain proteins associated with the surface of EVs contribute to cell invasion. As examples, the highly glycosylated form of the extracellular matrix metalloproteinase inducer (EMMPRIN), present at high levels in EVs from metastatic breast cancer patients, contributes to tumor invasion in the surrounding tissue ^[35][48]. Similarly, the tetraspanin protein CD9, also highly expressed on EVs, is thought to be associated with cancer cell invasiveness by promoting EV internalization, which stimulates cellular transformation (reviewed in Ref. ^[36][49]), while some integrin proteins expressed on the surface of EVs allow their selective uptake by organ-specific cells, leading to PMN preparation [5]. In addition to protein content, various types of RNAs were reported in EVs. Among them were messenger RNA (mRNA), microRNA (miRNA) and long non-coding RNA (lncRNA). The latter may induce permanent changes in chromatin structure and the regulation of gene expression ^[37][38][50,51], acting as inducers of pro-metastatic transformation ^{[38][39][40][41][42]}[51,52,53,54,55]. Cancer cell-derived apoptotic bodies and large oncosomes may also carry large fragments of DNA ^[16][17][43][21,22,56]. Even smaller EVs such as exosomes have been shown to contain large fragments (up to 10 kb) of double-strand DNA, and carry mutations of parental cancer cells ^[44][45][46][57,58,59]. They can also carry mutated DNA fragments, possibly harboring cancer driver mutations ^[44][47][57,60]. EVs isolated from the serum of tumor-bearing mice were found to contain the DNA, which reflects the genetic status of tumor-donor cells, including the amplification of the oncogene c-Myc as well as retrotransposon transcripts ^[48][61]. 

The universality of the “Seed and Soil” model has been challenged over the years ^{[49][50][51][52]}[64,65,66,67]. For instance, James Ewing proposed that the patterns of tumor metastases could be attributed to the anatomy of vascular and lymphatic drainage from the primary tumor ^[53][68]. Accordingly, Ewing’s hypothesis maintains that tumor cells follow the circulatory route, draining from the primary tumor, and stop non-specifically in the first organ encountered. Although there is evidence to support this theory, it cannot be opposed to the “Seed and Soil” model; the two are not mutually exclusive and may depend on the tissue origin and cancer types. Note that the Ewing hypothesis can account for the distribution of circulating EVs, which do not have active mobility, as described for migrating cells (see below), and are driven by circulatory flows. Decades later, alternative theories of metastasis formation were elaborated, such as metastasis by cellular fusion, genometastasis and, more recently, the horizontal transfer of malignant traits (HTMT) ^[54][55][56][69,70,71]. 

2. Model of Primary Tumor Cell Migration and Growth: The “Seed and Soil” Model

At least five major sequential steps can be identified in the “Seed and Soil” model, stemming from the original Paget’s hypothesis, which allows the cancer cells to metastasize to an anatomic site where the local microenvironment is favorable, just like a seed will only grow if it lands on fertile soil, owing to a concurrent action of circulating cancer cells and microenvironmental cues at the secondary sites [2]. Organotropism supports the idea that the colonization of circulating cancer cells depends on local conditions at remote sites, rather than solely on passive diffusion or random distribution ^[57][58][46,79]. Various chemokine- and growth-factor-mediated mechanisms have been described to explain tumor organotropism. A classic example is the CXCR4–CXCL12 chemotaxis axis ^[59][81]. CXCR4, a chemokine receptor expressed by most cancer types, including cancers of epithelial, mesenchymal and hematopoietic origin ^[60][82], has a critical role in cell migration and metastasis to organs that secrete its ligand, CXCL12, also known as stromal cell-derived factor-1 (SDF-1) ^[61][83]. CXCR4 overexpression is associated with poor prognosis in many types of cancer ^[62][84].  The PMN is also induced by signals released by primary tumors into the general circulation that promote molecular and cellular changes in the microenvironment, enabling circulating cancer cells to seed and give rise to metastatic lesions. In a landmark publication ^[63][88], Kaplan and colleagues demonstrated that a specific sub-population of bone marrow-derived hematopoietic progenitor cells expressing vascular endothelial growth factor receptor 1 (VEGFR1) and VLA-4 (i.e., integrin α4β1) moved to tumor-specific PMN, where they formed cellular clusters before the arrival of tumor cells; in this case, Lewis lung carcinoma cells.  Vascular alterations are then followed by changes in resident cells and the recruitment of monocytes and metastasis-associated macrophages that, in a positive feedback loop, promote the further extravasation and survival of metastasizing cancer cells ^[64][65][93,94]. In addition to these phenomena, extracellular matrix (ECM) proteins, such as tenascin C and periostin, can be produced by tumor cells themselves or by tumor-associated cells and play a pleiotropic role in metastasis progression by promoting invasive cell behavior, cancer migration and growth at the metastatic sites and neo-angiogenesis and cancer cell viability under stress ^[66][67][95,96]. These molecular and cellular processes seem to confirm that cancer-induced factors are important cues in determining the metastatic process.

3. Role of EV-Mediated Intercellular Communication in the “Seed and Soil” Model

EVs released by cancer cells are one of the unexpected factors that could determine and influence the cellular transformation, vascular permeability and the establishment of PMNs, thus preparing the “soil” in the target organs for metastasis. As mentioned above, and regardless of their subcellular origins, EVs and their cargoes play various roles in intercellular communication under physiological conditions, as well as in the pathogenesis of many diseases [10]. Interestingly, the total number of EVs in blood plasma increases in cancer, indirectly highlighting their potential contribution to cancer progression and metastasis ^{[68][69][70][71]}[99,100,101,102]. In metastases, integrin molecules associated with EVs can determine the organotropism ^[5][72][5,108]. The interaction of EVs with cell-surface-associated ECM in a given tissue/organ may allow their specific retention and uptake by resident cells at the predicted metastatic destination, activating intracellular pathways ^[5][73][5,109]. For examples, EVs harboring integrins α6β1 and α6β4 were associated with lung metastasis, while those with integrin αvβ5 were associated with liver metastasis due to preferential fusion with different types of resident cells at their predicted destination, lung fibroblasts and liver Kupffer cells, respectively [5]. Interestingly, the targeting of the specific integrins resulted in decreased EV uptake at the level of the organs site of metastasis [5]. The integrins (e.g., αvβ6) were also shown to be transferred from prostate cancer cells to recipient cells via exosomes and remained active in the host cells, where the αvβ6-mediated signaling pathway could modify the tumor microenvironment ^{[74][75][76][77]}[110,111,112,113] Hence, through EVs, cancer cells communicate between themselves and with cancer-associated fibroblasts (CAFs), or with other stromal cells, such as mesenchymal stromal cells (MSCs), vascular endothelial cells and surrounding immune cells, to promote their own growth and spreading ^[78][79][80][97,115,116]. In the context of an intercellular communication mechanism between cancer cells, Schillaci and colleagues showed that metastatic SW620 colon carcinoma cells were able to transfer, via EVs, an aggressive amoeboid phenotype to isogenic non-metastatic SW480 cells and to induce endothelial permeability, while non-metastatic cell-derived EVs were unable to induce such a transformation ^[81][117]. The mechanism underlying the observed cellular transformation has not yet been fully pinpointed ^[82][118], but certain differentially represented proteins in EVs derived from metastatic cancer cells have been suggested as potential candidates for the stimulation of pro-metastatic features. Cancer cells can also prime non-neoplastic cells via their EVs to facilitate cancer growth and metastasis through various cargoes. Evidence that EVs can mediate the transfer of cancer traits to normal MSCs derived from the same tissue/organ was provided by an in vitro study on colorectal cancer ^[83][129]. EVs isolated from cultured primary or metastatic cancer cell lines were able to induce morphological and functional changes in colonic MSCs. These changes included the formation of atypical microvilli and pseudopods, and an increase in EV release and augmented proliferation, migration and invasion rate. This evidence confirmed a clear involvement of EVs in phenotypic transformation at a local level, paving the way for the possibility of the induction of phenotypical transformation at a distance.  Collectively, these selected datasets highlight the contribution of cancer EVs in the modulation of various pathways associated with the metastatic process and suggest that they may play a central role in the transmission of messages to metastatic sites to promote the “homing” of circulating cancer cells. Thus, EVs have added another level of complexity to the Seed and Soil model, which may in some ways address some of the limitations of this classical hypothesis based solely on cancer cell migration and the appropriate metastatic niche.

4. Limitations and Challenges of the Seed and Soil Model

As mentioned in the Introduction, the universal validity of the original Seed and Soil model has, nonetheless, been questioned over the years due to certain shortcomings or a lack of experimental data supporting the entire underlying process of metastases ^{[49][50][51][52]}[64,65,66,67]. These shortcomings include (i) the low number of tumor cells that can actually circulate in the blood, e.g., less than 0.1% cells remain viable and <0.01% surviving circulating cells can produce metastasis ^[84][85][137,138]; (ii) the lack of definitive evidence that a single cancer cell is capable of sequentially performing all steps of the metastatic process, i.e., separation from the primary tumor, intravasation, survival in the circulation, extravasation and successful colonization ^[86][139]; (iii) the contrast between the concept of tumor dormancy, i.e., a prolonged latent state of asymptomatic micro-metastatic disease prior to overt metastasis formation ^[87][140], and uncontrolled proliferation ^[88][141], considered one of the main hallmarks of cancer; (iv) the poor correlation between bone marrow micro-metastases and their clinical manifestation ^[89][142]; and (v) differential gene expression profiles between primary cancer and metastatic cells ^[90][143]. Furthermore, the passage of cancer cells from the circulation to the metastatic site involves the ability to overcome physical constraints determined, for instance, by the presence of the tight capillaries’ junctions. This capability entails the reacquisition of new features such as reduced actin cytoskeleton anisotropy, cell stiffness and focal adhesion density. Moreover, once in the bloodstream, circulating tumor cells must acquire the ability to survive a variety of stresses, such as avoiding anoikis (a form of apoptosis due to the loss of integrin-dependent anchorage to the ECM), evading the immune system, and overcoming hemodynamic shear forces ^[91][144]. The acquisition of these abilities, according to the conventional model, might be determined by the expression of epithelial-to-mesenchymal transition (EMT) transcription factors such as Snail family transcriptional repressor (Snail), Twist, Zinc finger E-box binding homeobox 1 (Zeb1) and different miRNAs, as well as epigenetic and post-translational regulators. Upon reaching the metastatic site, the same cell would need to undergo a reversal of state, i.e., mesenchymal-to-epithelial transition (MET), to be able to home in the metastatic niche. The strongest element supporting the classical model of metastasis is the immunohistochemical similarity between primary cancer cells and the metastatic deposits. However, microarray analyses have shown that the gene expression patterns of primary breast tumors differ from those of their respective lymph node metastases, and that there are genes that are characteristically different in metastatic cells compared with their counterpart in primary tumors ^{[90][92][93][94]}[143,153,154,155].

5. Metastasis by Cellular Fusion

Fusion between cells is involved in many physiological processes, including fertilization and myogenesis, among others ^[95][157]. Such cellular processes could also occur in cancer, and explain their progression and metastases. As proposed by the physician Otto Aichel more than a century ago, metastasis could result from the fusion of cancer cells with healthy cells (e.g., leukocytes), resulting in hybrid cells that retain the properties of both parent cells ^[96][158] (reviewed in Refs. ^[97][98][99][159,160,161]). One of the first supports for the fusion came from a set of experiments by Barski and Cornefert, who mixed two distinct tumor cell lines and co-injected them into host mice, producing hybrid clones that, upon extraction and reinjection, gave rise to tumors in secondary recipient animals ^[100][162]. More recently, a group reported that aggressive breast cancer cells MDA-MB-231 (hereafter MDA) and MA-11 spontaneously fused with MSCs, while less aggressive MCF-7 or benign mammary epithelial cells did not ^[101][102][122,165]. Hybrids showed predominantly mesenchymal morphological characteristics. An analysis of single-nucleotide polymorphisms revealed genetic contributions from both parental partners to hybrid tumors and metastasis. Both MDA and MA-11 hybrids were tumorigenic in immunodeficient mice, and some MDA hybrids had an increased metastatic capacity compared to parental MDA ^[102][165]. Both in culture and as xenografts, hybrids underwent DNA ploidy reduction and reversal to a breast carcinoma-like morphology, while maintaining a mixed breast cancer–mesenchymal expression profile. The fusogenic activity of CD9 was shown to be essential for the fusion of MDA or MA-11 cells ^[101][122].

6. The Genometastatic Hypothesis

In 1905, Ehrlich and Apolant reported the sudden appearance of sarcomatous tissue within a mouse mammary tumor with the replacement of carcinoma tissue with pure spindle-cell tumors ^[103][166]. These observations were followed several decades later by the finding mentioned above, among others, where the subcutaneous implantation of human ovarian carcinoma cells grafted into nude mice resulted in two cancer populations, a human carcinoma and a murine sarcoma, suggesting the possibility of the transfer of factors from the human cancer cells to murine mesenchymal cells that would induce malignant transformation ^[104][163]. Following the reports that the horizontal transfer of genes in simpler organisms such as bacteria and fungi induced antibiotic resistance and adaptation to new environments ^[105][106][170,171], Holmgren and colleagues raised the question of whether DNA could be transferred from one somatic cell to another via the phagocytosis of apoptotic bodies ^[107][172]. They showed that the cultivation of apoptotic bodies derived from Epstein–Barr Virus (EBV)-carrying cell lines with either fibroblasts, monocytes or endothelial cells resulted in the uptake of DNA and the expression of EBV-specific markers in the recipient cells. The expression of the EBV markers, Epstein–Barr nuclear antigen 1 as well as EBV-encoded small RNAs, could be detected up to five weeks after the beginning of co-cultivation experiments.

7. Horizontal Transfer of Malignant Traits Model

7.1. HTMT and Preneoplastic Lesions

The potential of EVs circulating in the blood of patients in precancerous stages to promote malignant transformation has been suggested experimentally. For instance, EVs isolated from the sera of patients with dysplastic lesions or carcinomas in situ, which by definition are pre-cancerous lesions made of cells that have not yet acquired the ability to metastasize, caused malignant transformation in immortalized human HEK-293 embryonic kidney cells, which have an insertion of approximately 4.5 kb adenoviral DNA in chromosome 19 ^[17][22]. Upon transplant into NOD-SCID mice, transformed HEK-293 gave rise to tumors whose histology was compatible with poorly differentiated carcinomas with a high mitotic index ^[108][181], suggesting that cancer EVs circulate in the blood before the full neoplastic transformation is complete and cell invasion has occurred. Furthermore, such data reinforce the concept brought forward by the genometastatic and HTMT models that the metastatic process might be partially or completely independent of cell migration ^[92][153]. The discovery that, according to the HTMT model, EVs circulating in patients with neoplastic disorders, ranging from dysplastic to metastatic lesions, have a transforming ability on oncosuppressor-KO cells (see below), is the basis of a cancer screening test that can distinguish healthy patients from patients with cancer or at risk of developing cancer ^{[108][109][110]}[181,182,183]. This ability displayed by oncosuppressor-mutated cells to be transformed by cancer EVs has been incorporated in a unique biological platform for cancer screening called MATERD (Metastatic And Transforming Elements Released Discovery platform) ^{[108][109][110]}[181,182,183]. Among the peculiarities of this system is its intrinsic ability to detect neoplastic disorders even in patients who had only dysplastic lesions or cancers in situ, and its ability to detect cancer factors even years after the resection of the primary tumor to predict metastatic recurrence. In this regard, MATERD sensitivity is higher than the latest liquid biopsy tests based on the detection of circulating cancer cells, circulating DNA and circulating miRNA ^{[111][112][113]}[184,185,186]. The clinical results obtained with the MATERD test confirm that (i) cancer factors circulate in the blood prior to the complete cancerous transformation of the cells, undermining the concept that cell circulation is paramount to metastasis formation in distant organs ^{[108][109][110]}[181,182,183], and (ii) cancer factors are still found circulating in the serum after primary cancer has been removed and for several years afterwards, therefore strengthening the concept that these factors do not require the presence of cancer cells to induce metastatic disease ^{[108][109][110]}[181,182,183].

7.2. Oncosuppressor Genes As Gatekeepers

From a histopathological viewpoint, cells must acquire a series of mutations and transition through the phases of metaplasia, anaplasia and dysplasia before becoming fully cancerous ^[114][180]. Therefore, HTMT would occur at the metastatic site where local cells harbor mutations that predispose them to cancerous phenotypes by impairing DNA repair and cell cycle control and/or by regulating the uptake of oncogenic factors from the extracellular milieu ^[115][179]. A key distinction between the Seed and Soil model and the genometastatic hypothesis/HTMT is that in the revisited concepts of the Seed and Soil model, cells in the PMN would be expected to respond to signals from the tumor cells by altering the nutrients, immune response, and ECM in a way that facilitates the arrival and subsequent establishment of circulating tumor cells ^[116][187]. For example, when immortalized cell lines such as human HEK293 kidney embryonic cells, PNT-2 prostate cells ^[117][188], which exhibit alterations in cell cycle control, MCF10 cells carrying a mutated oncosuppressor gene, like phosphate and tensin homolog (PTEN), or the human fibroblast cell line carrying a CRISPR-Cas9-based deletion of the tumor suppressor breast cancer gene 1 (BRCA1) are exposed to blood serum of cancer patients or EVs derived therefrom, they transform into malignant cells which, upon subcutaneous injection into immunosuppressed mice, give rise to cancer masses ^{[108][109][110][114][118]}[178,180,181,182,183].

7.3. The Malignant Metastatic Phenotype Is Dependent on the Cell Type That Uptakes the Cancer EV-Derived Oncogenic Message

Another important distinction between the Seed and Soil and the HTMT models is the role of the PMN in the development of overt metastases. When BRCA1-KO fibroblasts (see above) were exposed to different types of cancer sera, regardless of their origin (e.g., breast cancer, lung cancer or lymphoma), they always acquired a malignant epithelial phenotype compatible with a lower gastrointestinal tract differentiation (CK7⁻/CK20⁺/CDX2⁺), whereas PTEN-deficient MCF10A cells exposed to the same patients’ sera consistently acquired a malignant phenotype suggestive of upper gastrointestinal tract differentiation (CK7⁺/CK20⁻/CDX2⁻) ^[110][183]. According to the HTMT model, the occurrence of two types of primary cancer in the same patient (30% of cancer cases) would be secondary to the uptake of the same oncogenic message by two different cell types and its subsequent expression at different points in time rather than being caused by chronic exposure to the same risk factors ^{[110][119][120][121]}[183,190,191,192].

7.4. EV-Mediated Interactions of Cancer Cells with the Immune System

Both the Seed and Soil and the HTMT models require that the immune system fails to recognize developing metastatic tumors. According to the HTMT hypothesis, the transfer of HLA proteins is particularly important in the context of immune tolerance; in fact, whole-genome sequencing of DNA isolated from BRCA-1 KO fibroblasts (see above), before and after their transformation in colon cancer cells by colon cancer EVs, showed that gene variants codifying for the extra-membrane portion of HLA proteins were transferred through EVs ^[92][115][153,179].

7.5. Cancer Dormancy According to the HTMT Model

Cancer cell dormancy postulates that, although uncontrolled proliferation is a distinct feature of cancer cells, they can stop proliferating and hibernate for long periods until reactivated by unknown signals ^{[87][122][123]}[140,194,195]. The HTMT model, involving an interplay of cancer EVs, immune cells, and other supportive cells in determining the course of the neoplastic disease, is consistent with the concept of cancer cell dormancy. In fact, it theorizes that cancer EVs released by the primary carcinoma travel through the lymphatics to the regional nodes, where they interact with lymphoid cells ^[124][196]. The interplay between EVs and lymphoid cells determines either the destruction of the cancer genetic material contained in EVs or immune tolerance to these oncofactors. If tolerance ensues, EVs can freely travel through the bloodstream, undetected by the immune system, and reach a PMN. There, the cancer genes integrate in the genome of the cell and might be either expressed, determining malignant transformation of the cell, or remain silent and become activated later in time, determining the phenomenon of latency.

7.6. Molecular Mechanisms Involved in the HTMT Model

The nature of the molecular mechanisms responsible for the HTMT process has not yet been clarified. The whole genomic sequencing and transcriptome analysis of cancer EVs and the recipient BRCA1-KO fibroblasts prior to and after exposure to cancer EVs confirmed that the active transfer of nucleic acids notably mutated cancer genes and their active transcription ^[115][125][179,205]. Xeno-transplants of colon cancer EV-transformed BRCA1-KO fibroblasts displayed the epithelial colorectal cancer phenotype, indicating MET. This in vivo evidence correlates at a molecular level with the differential expression of genes involved in the MET process, as well as with cell growth and cell death at both transcript and protein expression levels ^[115][179]. Thus, a decrease in the mesenchymal markers Snail1, Snail2, Zeb1, Zeb2, N-cadherin (CDH2) and fibronectin was observed, while the expression of the epithelial marker CDH1 was increased.

8. Unanswered Questions Raised by the HTMT Model

Although the laboratory evidence gathered so far about the HTMT model has been based on cancer-derived small EVs, whether exposure to large oncosomes or cancer-derived apoptotic bodies could obtain the same effect needs to be investigated. Additionally, while it has been described that the same cancer EVs induce different cancer phenotypes depending on the type of target cell (epithelial vs. mesenchymal), whether the type of oncosuppressor mutation might also have a role in the expression of the phenotype needs to be ascertained. Most importantly, the EV cargo and its molecular mechanisms that trigger the malignant transformation need to be defined. So far, the experiments performed to explain the HTMT dynamics have confirmed the presence of cancer-derived DNA, mRNA and miRNA in the transformed cells. Which EV-contained genes and/or epigenetic mechanisms cause the malignant transformation needs to be established. Little is known about the true composition of the malignant cargo that, once it has been taken up into the cells, is able to perform such dramatic changes. It cannot be ruled out that distinct EVs and their cargoes act synergistically when they are taken up. The evidence that circulating free DNA alone can induce undifferentiated malignant transformation triggers questions regarding the role that other genetic moieties and entities, transferred through the EVs, might have in the process ^[117][188]. The malignant transformation of a healthy cell might be secondary to a process similar to the induction of pluripotency in a skin cell induced by the simple epigenetic activation of only four genes ^[126][206].

9. A Novel Intracellular Pathway Accounting for the Nuclear Transport of EV Cargo

In the modified Seed and Soil and HTMT models, the fate of EVs after internalization into recipient cells should have a major impact on their functional role, but unfortunately little is known about this, including about their subcellular distribution in the cytoplasmic compartment and how their cargoes reach a specific molecular target. After endocytosis, the most common cellular entry route for EVs, internalized EVs traffic through endosomal compartments, where the fusion of late endosomes with lysosomes causes the degradation of EV-derived components (see Ref. ^[15][20] and references therein). In this recognized and accepted pathway, the EV cargo would not elicit a cellular response. Alternatively, the acidic environment of late endosomes may promote EV membrane fusion with the limiting endosomal membrane, allowing soluble EV cargoes to reach the cytoplasm, and perhaps their targets ^[127][207]. Such an exciting scenario remains difficult to conceive, unless there is a massive uptake of EVs, as EV cargoes contain limited amounts of biomaterials. In search of a model that could address this limitation, scholars recently described a novel nuclear pathway that delivers EVs to the nucleoplasmic reticulum (NR), composed of invaginations of the inner (type I) or both inner and outer nuclear membrane (type II) ^[128][208]. This novel pathway consists of a fraction of Rab7⁺ late endosomes carrying endocytosed EVs that translocate by a microtubule-dependent mechanism into the NR, specifically to type II nuclear envelope invaginations (NEIs) ^[129][130][72,73]. Through interactions of the outer nuclear membrane-associated vesicle-associated membrane protein-associated protein A (VAP-A), cytoplasmic oxysterol-binding protein (OSBP)-related protein-3 (ORP3) and small GTPase Rab7, a tripartite protein complex (named VOR, an acronym for the three proteins involved) allowed the docking of late endosomes onto the outer nuclear membrane and their translocation into NEIs. After the fusion of EVs and the endosomal membrane, the discharge of EV cargoes such as soluble proteins and nucleic acids into the cytoplasmic core of NEIs could allow their concentration therein and/or their translocation into the nucleoplasm through the local nuclear pores ^[129][130][72,73]. Thus, the NEI may play a dual role in the transfer of EV information, either by increasing the probability that EV cargoes (proteins and mRNAs) will meet or bind their host targets as newly synthesized mRNAs exported out of the nucleus, or by facilitating the nuclear import of EV cargoes. The entry of EV cargoes (proteins and nucleic acids) into the nucleus has been observed in numerous studies, by ourselves and others ^{[129][130][131][132]}[72,73,209,210].

10. VOR Complex Inhibition As a Means to Intercept the EV-Mediated Nuclear Transfer of Oncogenic Factors

The disruption of EV-mediated bidirectional communication between cancer cells themselves at the primary tumor site and/or non-neoplastic stromal/immune cells at the metastatic site may be a novel approach to inhibit cellular transformation and PMN development. There are currently no anti-cancer drugs known to target communication between cancer-cell-derived EVs and the nuclear compartment of host cells, in particular the nuclear transfer of EV cargoes. A group hypothesized that compounds that interfere with protein interactions of the VOR complex would prevent such nuclear transfer of EV cargoes, and thereby inhibit the downstream pathways leading to the cellular transformation ^[133][212]. In search for such inhibitors, it came across FDA-approved antifungal drug itraconazole (ICZ) that had previously displayed an inhibitory activity on enterovirus and hepatitis C virus replication through binding to OSBP and ORP4, members of the same family to which ORP3 belongs ^[134][213].

11. Conclusions

The different roles attributed to EVs in many pathological conditions, including cancer, are revolutionizing our understanding of the metastatic process, among other things. Their function in the development of the PMN and recent data on the EV-induced malignant transformation of oncosuppressor-KO cells deserve further investigation. Targeting the uptake of EVs and the intracellular pathways they use upon internalization could lead to new therapies that interfere with metastasis. Interestingly, the word “metastasis”, originally from Greek, appeared in the English language in the late 16th century as a rhetorical term, implying ‘rapid transition from one point to another’, while in a medical context, metastasis simply means the transference of the seat of disease ^[135][136][214,215]. Therefore, its original meaning may still encompass all possible mechanistic explanations for the “spreading” of the cancer disease from the original site to distant organs.