Lung cancer is one of the most common cancers and is still the leading cause of cancer-related deaths. Indeed, as reported by GLOBOCAN, lung cancer is the second most prevalent neoplasm and the main cause of cancer mortality, with 2.2 million newly diagnosed cases and 1.8 million deaths in 2020. In most countries, the 5-year survival rate of lung cancer patients is only 10% to 20%: stratifying by disease stage, only 60% of stage IB, 50% of stage II, and 40% of stage IIIA patients survive more than 5 years (1). The lack of symptoms in the initial stages makes early diagnosis difficult, often delaying lung cancer detection at an advanced or widely metastatic stage (2).

Several efforts have been made to improve the therapeutic management of advanced lung cancer including specific target therapies against driver mutations and, more recently, the introduction of immunotherapies alone or in combination with standard chemotherapy. Immune-checkpoint inhibitors (ICI) have demonstrated efficacy in the treatment of advanced lung cancer even though a sizable proportion of patients, unfortunately, do not respond to treatment or have a durable response. Despite treatment improvement, the 5-year survival rate for advanced lung cancer remains very low overall (6%) (3).

The survival rate of lung cancer patients is strongly affected by the presence of advanced or metastatic disease which often occurs quickly or is already present at diagnosis. It has been estimated that almost 70% of the total lung cancer-related deaths are a consequence of metastatic spread, thus representing an extremely relevant clinical and social issue (1). Moreover, despite the common belief that metastatic dissemination might be related to late-stage disease, cancer cell dissemination could be an early event in cancer progression (4, 5).

One of the earliest events of the metastatic process is the escape of cancer cells from the primary tumor, often linked to the acquisition of mesenchymal and invasive properties through the induction of the epithelial-to-mesenchymal transition process (EMT), allowing cancer cells to intravasate into the lymphatic or blood circulatory systems. Then, cancer cells traveling in the circulation, alone or within a cluster of cells, can reach distant sites where they can extravasate and grow as metastatic lesions. All these processes require a complex network of communications among different cell types and tissues. In the attempt to fully elucidate the metastatic process and the crosstalk among different cellular players, experimental studies have often focused on cancer dissemination to the lungs since it represents one of the most frequent sites of metastasis from many tumor types, including carcinomas of the breast, colon, kidney, melanoma, and lung cancer itself (6). In fact, it must be emphasized that primary lung cancer also frequently metastasizes to the lung: indeed, lung cancer is the tumor that contributes most to the incidence rate of synchronous lung metastases (41%), followed by colon and rectum (10%), kidney (7%), pancreas (7%) and breast (6%) (7).

The preferential sites for lung cancer-derived metastasis at diagnosis are lung (49.0-59.2%), brain (29.9-41.9%), bone (28.5-38.8%), liver (13.2-26.3%), and adrenal gland (10.1-24.1%) (8). In detail, among the most frequent form of lung cancer, non-small cell lung cancer (NSCLC), different histological subtypes can preferentially metastasize to specific organs: squamous cell lung carcinoma (SCLC) often spreads to the liver whereas adenocarcinoma most frequently spreads to brain. Moreover, specific genetic alterations could influence the metastatic pattern as in the presence of EGFR mutations which have been linked to brain metastasis in NSCLC (9). These differences in metastatic tropism for lung cancer subtypes may also reflect different interactions between tumor cells and specific microenvironments (6).

This clinical evidence underlines the need to further elucidate the mechanisms behind lung cancer tropism to certain organs. Additionally, information gathered on the pulmonary ‘soil’ as permissive or restrictive for the growth of cancer cells of different origins may also provide important insights into understanding lung cancer metastatization to the lung. Therefore, besides discussing evidence focused on lung cancer metastasis, we will also review studies on the main mechanisms regarding lung metastasis originating from different primary tumors.

The first hypothesis suggesting that metastatic dissemination was not a “matter of chance”, was postulated in 1889 by Stephen Paget who noted in 735 cases of advanced breast cancer that not all organs were equally apt to receive “particles” from the primary tumor and develop metastasis (10).

Starting from this first evidence, Ewing and colleagues tried to explain the link between metastatic dissemination and the mechanical dynamics of hematogenous flow (11). Then, other studies have been conducted in the field concluding that certain organs were generally more susceptible to metastases than others (12). Finally, the now well-known “seed and soil” theory was put forward: tumor cells (“the seeds”) require an appropriate local microenvironment (“the soil”) to effectively grow as metastatic lesions.

Several efforts were then made to address the many open questions about how the primary tumor interacts with distant organs during the metastatic process (13). In this regard, a seminal work by Kaplan et al. in 2005 introduced the concept of “pre-metastatic niche” (PMN). This work demonstrated for the first time that the early recruitment of bone marrow-derived cells (BMDC) to the lungs caused microenvironmental changes and the creation of a pro-metastatic environment before the homing of circulating tumor cells (CTC) (14). These findings support the concept that primary tumors can shape the microenvironment of distant organs before tumor cell colonization. Although similar to Paget’s and Ewing’s theories, the concept of the PMN is more complex since it postulates the potential of some factors derived from primary tumors to precondition specific organs, making them suitable sites for metastases. Indeed, it’s now clear that primary tumors actively prime stromal cells at distant organs to generate a supportive environment for the recruitment, implantation, survival, and outgrowth of tumor cells (14–17).

In a general view, the establishment of the PMN is a multifaceted and multi-step process characterized by the requirement of fine-tuned communication between different sites and microenvironment players (11, 15). As recently reviewed by Peinado et al. (11), the establishment of PMN occurs through sequential steps: first, the blood vessels of the future site of metastasis lose their integrity in a phenomenon called “vascular leakage”, causing increased permeability and allowing the entrance of cells and macromolecules, which usually cannot otherwise penetrate the endothelial barrier. Next, other local stromal cells undergo a wide range of alterations; among them, fibroblasts play a central role through the deposition and remodeling of the extracellular matrix (ECM) or the secretion of molecules for immune cell recruitment and/or modulation (18, 19). The ensuing vascular and stroma rearrangement induces the recruitment of non-resident cells, mostly BMDC, that in combination with other deregulated immune cells generate a supportive environment for disseminated tumor cells and for subsequent metastasis outgrowth (11, 15, 20).

Several studies have described molecular and cellular effectors of this “systemic effect” connecting communication among different tissues (the site of the primary tumor, the bone marrow, and the site of metastasis) and proving the central role of tumor-derived extracellular vesicles (EVs) in PMN establishment (11). Since the role of other soluble factors, such as cytokines and growth factors, in the formation of the PMN was already well described (21), in this review we will focus on EVs, with a particular interest in the metastatic lung cancer setting.

Originally described simply as “waste operators”, EVs have recently gained attention for their role as key players in cell-cell communication (22). EVs are a large and heterogeneous group of cell-derived membranous vesicles secreted by almost all cell types (23, 24). Their family comprises several vesicles different in features (mainly in their size) and biogenesis. Despite the controversial and still evolving classification, EVs can be broadly subdivided into small EVs (sEV, also known as exosomes) derived from multivesicular bodies of late endosomes (~50–150 nm in diameter), and microvesicles (MVs or ectosomes) which originate through extracellular membrane budding (from ~100 nm up to 1 µm in diameter) (22, 25).

The bioactive cargo of EV comprises mostly transmembrane proteins, lipids, and nucleic acids as DNA and RNA (mRNAs, long non-coding RNAs, and miRNAs) (26–28). Although it is still unclear how proteins and nucleic acids are actively sorted or packaged into EVs, it is well established that the cargo of EVs is systematically integrated into the EVs through a strictly controlled process. In this regard, it has been proposed the involvement of the proteins of the Endosomal Sorting Complex Required for Transport (ESCRT) both in EV biogenesis and in particular during multivesicular bodies (MVBs) and intra-luminal vesicles (ILVs) formation and release (29–31). Indeed, the ESCRT-associated proteins Alix and TSG101 were found inside EVs, corroborating the role of ESCR complexes during cargo packaging. Even though the entire process is not yet fully understood, there is consensus about the requirement for some post-transcriptional modifications such as ubiquitination, SUMOylation, ISGylation, Phosphorylation, and Glycosylation (reviewed in detail by Sushma Anand and colleagues (32)) during the selective incorporation of cargo components. Indeed, the ESCRT complexes were found to be actively involved in the incorporation of ubiquitinated proteins inside EVs (33, 34) and the aforementioned Alix appears to play an active role in this process, sorting both proteins and miRNAs inside EVs (35, 36).

The content of EVs is cell-specific and reflects the physiological status of the cell of origin (37). For this reason, the presence of specific markers on the surface of EVs can be used to identify the cell type responsible for their release. For example, the tumor antigen 5T4 was detected in EVs from prostate cancer cell cultures and patients’ urinary-EV, but not in urinary EVs from healthy donors highlighting their specific tumor origin (38). In this scenario multiplex bead-based platforms allow simultaneous detection of several markers on EVs surface linked to the cell of origin. For instance, it has been demonstrated that NK cells-derived EVs are characterized by CD2, CD8, and CD56 markers whereas platelet-derived EVs lack CD2 and CD8 but are enriched in platelet markers such as CD41b, CD42a, and CD61 (39).

EVs have been successfully purified from many body fluids such as blood, urine, pleural effusions, ascites, and bronchoalveolar fluid (40), so they represent an interesting and non-invasive source of biomarkers for specific disease detection

The EV’s cargo is protected from enzymatic degradation by the phospholipidic bilayer allowing delivery of their content to recipient cells without any alteration (41). Once released, EVs can modify the physiological state of near or distant recipient cells by adhesion, fusion, and transfer of cargo components. The process of EV uptake by recipient cells is not yet fully understood but it occurs mainly in two possible ways: internalization by endo- and/or phagocytosis or by direct fusion with the membrane of the target cell. In both cases, the EV content is released into the cytosol of recipient cells. As a potential third mechanism of bioactive action, EVs can activate downstream signaling via receptor-ligand interaction on target cells (42). Several EV components (tetraspanins, integrins, or adhesion molecules) have been described as mediators of the binding between EVs and recipient cells (43). Indeed, the presence of specific integrins on the surface of cancer cell-derived EVs has been highlighted as responsible for the EV-cell interactions in different organs also during the PMN formation (28). Moreover, a specific pattern of tetraspanins on EVs has been shown to drive EV uptake by selected target cells, as indicated for example by the requirement of Tspan8 for the interaction between EVs and endothelial cells (44).

Following EVs uptake, their biological active cargo is released intracellularly to regulate in an autocrine and/or paracrine fashion multiple cellular processes including cell proliferation, survival, and potentially even transformation (43). Indeed, due to their ability to ‘deliver’ messages between different cells, EVs have been linked to all stages of cancer development, progression, and particularly to metastasis formation (45). Tumor-derived EVs indeed regulate different features of malignant cells such as proliferation (46), EMT (47), abnormal apoptosis (48), and metastatic spread (49).

EVs are involved in the acquisition of mesenchymal and more aggressive phenotypes by lung cancer cells ( Table 1 , Figure 1 ). In this context a prominent role is played by cells endowed with increased tumor-forming potential: the existence of a subset of cancer stem-like cells (CSC) able to sustain primary tumors and initiate distant metastasis has been extensively demonstrated in both hematological and solid tumors, including lung cancer in which CSC have been identified by the expression of CD133 marker by different groups (84, 85). Expression of CSC markers in primary tumors has been shown to correlate with worst prognosis and metastasis occurrence. Recently, some evidence also highlights the importance of CSC-derived EVs in mediating aggressive phenotype of tumors or in corrupting surrounding and distant stroma cells to promote tumor progression (86). Indeed, in different tumor types, it has been demonstrated that CSC- EV cargo can induce non-tumor cells to gain stem-like features through the induction of EMT which in turn promotes chemotherapy resistance and metastasis formation (83, 87, 88). Different bioactive molecules carried in CSC-derived EVs have been shown to mediate such effects, among which stemness-related proteins or activators of stem-related signaling (89, 90) and RNA molecules (91).

In lung cancer, it has been demonstrated that CSC-derived exosomes can enhance the invasive and pro-metastatic properties of lung cancer cells by transferring miR-210-3p that binds and down-modulates fibroblast growth factor receptor-like 1 (FGFRL1), inducing a mesenchymal phenotype (74).

Interactions among different subsets of cancer cells can also be mediated by EVs: metastatic cell-derived EVs can stimulate the acquisition of aggressive behavior in recipient cells. Indeed, exosomes isolated from metastatic small-cell lung cancer cells increased cancer cell proliferation and migration compared to non-metastatic cells through the release of EV-derived TGF-β and IL-10 (75). Moreover, exosomes released by metastatic lung cancers can induce EMT in normal epithelial cells by transferring vimentin (47). miRNA-499a-5p, highly enriched in EVs from metastatic lung cancer cells, modulates the mTOR pathway in recipient lung cancer cells with concomitant induction of proliferation, EMT, and migration (76). The presence of miR-1260b in lung cancer-derived EVs promotes the invasive capacity of lung tumor cells through the modulation of the sFRP1 and SMAD4 pathway (77). Additionally, Tetraspanin 8 on the surface of EVs stimulates the invasiveness of both human and mouse lung cancer cells in in vitro studies (78). Interestingly, EVs can also be able to promote proliferation and migration of lung cancer cells by transferring HGF and concomitant activation of c-Met (79). Tumoral-derived exosomes detected in the pleura exudates carry enzymes for leukotriene (LT) biosynthesis that can act on lung cancer cells promoting their migration and proliferation (92).

Exosomes from cancer cells can be exploited as a communication tool with local and distant normal cells to generate a micro-environment suitable for their proliferation/invasion. For instance, lung cancer cell-derived exosomes were able to induce a pro-inflammatory phenotype in mesenchymal stem cells that in turn support tumor growth (93).

Similarly, exosomes from immune/stroma cells can in turn contribute to the induction of aggressive properties in recipient tumor cells. Microvesicles released by activated platelets in patients with lung cancer are also involved in metastatic outgrowth. Indeed, these vesicles are enriched in miR-223 compared to healthy donors and this miRNA promotes cell invasion through the inhibition of erythrocyte membrane protein band 4.1-like 3 (EPB41L3) (80). Another study revealed that microvesicles from tumor-educated platelets enhance the migration of lung cancer cell lines and increase the release of angiogenic factors such as VEGF, MMP-9, and IL-8. The authors demonstrated also that the presence of integrin CD41 inside EVs can enhance the metastatic potential of murine lung cancer cells (81). In addition, activated CD8 T cells release exosomes expressing FasL that can modulate MMP-9 expression in cancer cells and consequently lung metastasis (82).

Primary tumors constantly shed millions of cells per day into the bloodstream, an event that also occurs at the early stages of the disease (94). Circulating tumor cells (CTC) represent the seed of metastasis and their targeting in the circulation, or once landed at the PMN, represents a challenging but crucial requirement to counteract metastasis formation. Despite the very low number of CTCs that can be detected in the blood of patients at specific time points (1 CTC:10⁷ leukocytes), these cells are characterized by extreme heterogeneity, detected at genomic, transcriptomic, and functional levels which greatly limits their in-depth study and characterization (95).

Considering the great number of CTCs released by primary tumors during development and expansion, the efficiency of metastatic seeding is extremely rare, with an estimated success rate of 0,001% (96). This observation subtends that only very few CTCs can survive in the circulation, reach PMN and initiate metastasis. Rising evidence is emerging highlighting that the metastatic potential of CTCs is largely confined to the cell subset endowed with stemness and mesenchymal traits that could succeed to carry out all the steps of the metastatic cascade (97–99). Recent technological advances in single-cell sequencing strategies have allowed a better understanding of CTC heterogeneity in different tumor types and confirmed, in preclinical models and patient samples, the activation of stemness programs in CTCs, strongly supporting the existence of a rare subpopulation of stem-like cancer cells guiding tumor spread and metastasis initiation (100–102).

CTCs can enter the circulation as single cells or, occasionally, as clusters of tumor cells alone or associated with immune cells or stroma cells, such as neutrophils or cancer-associated fibroblasts (103). Clusters of tumor cells, held together by cell-cell junctions (104) or cytoskeletal proteins (105) can be shed from primary tumors; alternatively, clusters can be formed in the circulation by single CTC aggregation through homophilic interactions (106). Several studies have demonstrated in different tumor types that clusters and in particular heterotypic clusters, comprising CTCs and immune/stroma cells, possess the highest metastatic ability (104) and their detection in patients’ blood predicts worst outcome (104, 107–109).

Neutrophils are the most common and ‘dangerous’ travel companions of CTCs. It has been demonstrated that neutrophils can escort circulating tumor cells and support their proliferation by enabling cell cycle progression: as a result, once landed at the PMN, active CTCs have a higher chance to initiate metastasis (109). Besides neutrophils, other immune cells and stroma cells can travel in circulation with CTC. For instance, platelets bound to CTCs can have multiple protective effects, by preventing immune cells’ recognition and attack of CTCs and by providing factors able to support CTCs’ survival in unfavorable conditions or conferring malignant traits, such as acquisition of CSC phenotype, invasiveness, and drug resistance (110). Macrophages that bind to CTCs play a crucial role in tumor cell intra and extravasation steps, supporting CTC seeding and survival at distant sites (111). Finally, Duda et al. demonstrated that also cancer-associated fibroblasts can travel into heterotypic CTC cluster that spontaneously spread from the primary tumor and this interaction greatly help CTCs to establish distant metastasis (112).

Once landed at the PMN, CTC can interact with (pre)activated endothelial cells and immune cells that can facilitate their entrapment and survival. In particular, it has been demonstrated that neutrophil extracellular traps (NETs) can capture CTC through a β1-integrin-mediated mechanism and this interaction can sustain CTC potential in metastasis development (113).

Even though no direct evidence has been reported regarding the role of EVs in dictating the metastatic behavior of CTCs, it is conceivable that EVs released by both immune/stroma cells and cancer cells can mediate the process leading to CTCs escape from the primary tumor, survival in circulation and seeding at distant sites. Indeed, as discussed in the previous paragraphs, EVs released from stroma and immune cells may participate in the process of cancer cell intravasation and formation of the metastatic clusters as well as could mediate the activation of stemness/survival/proliferative pathways in CTCs. Finally, CTCs attracted to EVs-primed PMN can find the most favorable soil to initiate metastasis.

After the arrival at metastatic sites, disseminating tumor cells could be eliminated by immune cells or enter a state of dormancy by exiting proliferative cell cycle (54). Remodeling of lung stromal cells is a key step to generate a fully competent metastatic niche before or also after the arrival of CTCs ( Figure 2 ). The recruitment of immune cells with an immunosuppressive and pro-metastatic phenotype appears, however, crucial for the development of metastasis. In a murine model of lung carcinoma, tumor cells activated through TLR7 are able to recruit myeloid-derived suppressor cells (MDSCs) via the release of cytokines such as CCL2 and GM-CSF (114). The expansion of the MDSC pool with an immunosuppressive phenotype in the lungs leads in turn to metastatic progression.

Neutrophils emerge as one of the most important immune cells that could control the metastatic process in lung. The recruitment of neutrophils by lung epithelial cells activated by tumor-derived exosomal RNAs-TLR3 binding was recently shown to be an important step in PMN formation by supporting the growth of lung cancer cells (59). Interestingly, it has been also demonstrated that neutrophils reactivate dormant cancer cells by switching their polarization status. Indeed, stress-activated PMN-MDSCs release S100A8/S100A9 which can support the reactivation of dormant tumor cells (115).

In some cases, dormant cells can activate bone-marrow-derived endothelial progenitor cells by promoting an angiogenic switch that allows angiogenesis-mediated progression of micro-metastasis to overt metastasis (116). Since the formation of the metastatic niche is a process that involves several cell types, further studies are needed to elucidate the cross-talk between EVs, stroma, and immune cells at distant organs. Little is known about the role of EVs derived from other cell types at secondary organs in modulating the environment for metastatic outgrowth. It could be also possible that EVs control tumor cell dormancy although this process is not yet fully understood. An interesting work showed that the presence of miR-210 and miR-193 inside EVs released by hypoxic bone marrow-derived cells (BMDC) can increase lung cancer cells invasion and EMT through the activation of STAT3 pathway (117). Recruited BMDC contribute to the formation of liver metastasis from lung cancer by releasing EVs enriched in miR-92a (118). In the brain, EVs shed by endothelial cells exert a protective effect on the survival of SCLC cells through the up-regulation of S100A16 (119).

Even though it appears clear that EVs derived from cells at the metastatic niche could regulate the fate of disseminating cells, it is hard to establish the exact origin of these EVs. Moreover, the majority of published studies were focused on the ability of EVs to induce an aggressive phenotype in lung cancer cells while very little is known regarding EVs-induced mechanisms that regulate the transition from DTC to full metastasis.

Metastatic disease remains the first cause of lung cancer-related deaths. The complexity of the metastatic cascade steps and the intricate connections and cross-talk between cancer cells and the microenvironment constitute challenges for a deeper understating of advanced disease and the possibility to prevent/counteract it. It’s becoming clear that the generation of specialized microenvironments at distant sites, named the pre-metastatic niche (PMN), guides the propensity of primary cancer cells to colonize specific distant organs and dictates their success to initiate metastasis.

Primary tumor-secreted factors are able to prime stromal and immune cells at distant sites thus creating a permissive niche for circulating tumor cells landing and metastatic colonization. Among secreted factors, extracellular vesicles have been recently recognized as central mediators for PMN formation. EVs cargo is composed of a plethora of bioactive molecules that impact all different steps of the metastatic cascade. EVs released by primary tumors can induce vascular leakiness at distant sites, activation of fibroblasts, and recruitment of BMDC, events that all together concur in PMN formation. On the other hand, at primary tumor sites EVs released by both cancer and stromal/immune cells can impact on the acquisition of invasiveness and stem-like cancer phenotypes and can favor tumor cells’ intravasation and dissemination.

Circulating tumor cells can be directed by stimuli within the primary tumor to reach distant organs, where they can interact with EVs-primed stromal/immune cells which favor their extravasation and govern cancer cells’ dormancy/awakening, dictating metastatic success.

Here we have reviewed available evidence highlighting the involvement of EVs in all the different steps of metastasis formation in the context of lung cancer. Although the precise role of EVs has yet to be elucidated for a comprehensive view of the process, it is clear that they contribute to the successful development of PMNs and metastasis formation. Therefore, a deeper knowledge of EVs’ origin and cargo can shed new light on our understanding of metastatic disease and may allow the identification of novel targets to prevent lung cancer progression and metastasis formation.

Writing—original draft preparation: FP, LR, OF, and GB. Writing—review and editing: FP, LR, OF, and GB. Funding acquisition: OF and GB. All authors have read and agreed to the published version of the manuscript. All authors contributed to the article and approved the submitted version.