The Immunomodulatory Role of Hypoxic Tumor-Derived Extracellular Vesicles

Simple Summary Hypoxia, a characteristic of many cancer types, can suppress the antitumor effector functions of the adaptive and innate immune system. Tumor-cell-derived extracellular vesicles, which function as a mechanism of communication between tumor cells and immune cells, are also affected by hypoxia, and may drive immunosuppression. The aim of this review is to summarize the current knowledge on hypoxic cancer-cell-derived extracellular vesicles in immunosuppression, and to provide an overview of enriched factors (i.e., miRNA and proteins) in hypoxic tumor-derived EVs and their role in immunomodulation. This complete overview may indicate relevant directions for future research into the role of hypoxia in immunosuppression during cancer. Abstract Tumor-associated immune cells frequently display tumor-supportive phenotypes. These phenotypes, induced by the tumor microenvironment (TME), are described for both the adaptive and the innate arms of the immune system. Furthermore, they occur at all stages of immune cell development, up to effector function. One major factor that contributes to the immunosuppressive nature of the TME is hypoxia. In addition to directly inhibiting immune cell function, hypoxia affects intercellular crosstalk between tumor cells and immune cells. Extracellular vesicles (EVs) play an important role in this intercellular crosstalk, and changes in both the number and content of hypoxic cancer-cell-derived EVs are linked to the transfer of hypoxia tolerance. Here, we review the current knowledge about the role of these hypoxic cancer-cell-derived EVs in immunosuppression. In addition, we provide an overview of hypoxia-induced factors (i.e., miRNA and proteins) in tumor-derived EVs, and their role in immunomodulation.


Introduction
Cancer is a devastating disease caused by the uncontrolled proliferation of cells, and is the second leading cause of death globally. According to the World Health Organization, approximately 1 in 6 deaths are due to cancer, with an estimated 10 million deaths worldwide in 2020 [1]. Furthermore, the incidence and mortality rates are expected to steadily increase over the coming years. During tumor development, cancer cells acquire 14 biological properties, known as the hallmarks of cancer which, among other traits, allow them to proliferate uncontrollably, spread to other organs, prevent cell death, adapt to the extreme changes in the TME, and evade the immune system [2].
In addition to providing protection against infections, the immune system is involved in the clearance of debris, initiation of tissue repair, and preventing tumor growth through Figure 1. Suggested mechanisms for hypoxic tumor-derived EV-mediated immunosuppression to support tumor growth: Tumor hypoxia induces a more aggressive and therapy-resistant phenotype in cancer cells. This is associated with the release of immunosuppressive EVs, among other effects. These vesicles are taken up by a variety of immune cells, and prevent their differentiation into functional effector cells. In addition, these EVs induce the formation of regulatory immune cells such as Tregs and MDSCs, which further dampen the efficiency of the antitumor immune response. Furthermore, these EVs induce the formation of tumor-supportive M2 macrophages. Image created with BioRender.com, (accessed on 7 June 2022).

Neutrophils
Neutrophils are the most abundant population of white blood cells in the human circulatory system, reflecting their role as first responders during infection or tissue injury. Elevated neutrophil counts are observed in the peripheral blood and tumors of patients with advanced cancers. In these patients, both tumor-supportive and tumor-suppressive neutrophil subsets have been described (Table 1) [25,26]. The direction of the tumor-associated neutrophil function is dependent on the types of soluble mediators produced by cancer cells. For example, in a transforming growth factor β (TGF-β)-rich environment, neutrophils typically display a tumor-promoting phenotype, whereas they acquire antitumor properties in the presence of interferon β (IFN-β) [25]. To date, no direct effect of hypoxic EVs on neutrophil function has yet been described. Nevertheless, hypoxic cancerderived EVs have been shown to be enriched in factors known to play a role in neutrophil chemotaxis, including CXCL8 (IL-8) and MIR-451 ( Figure 2A). CXCL8 is a pro-inflammatory cytokine that is enriched in EVs of hypoxic glioma cells [27]. It mediates the recruitment of leukocytes-including neutrophils and myeloid-derived suppressor cells (MDSCs)-to tumors through binding the receptors CXCR1 and/or CXCR2 [28]. Moreover, through activation of MAPK signaling, CXCL8 signaling inhibits neutrophil apoptosis and enhances proliferation [29]. An opposing role has been observed for MIR-451, which inhibits the p38 MAPK signaling pathway, suppressing neutrophil chemotaxis [30]. Given the known diversity in pro-or antitumor neutrophil subpopulations and the varying effects of the different hypoxic-EV-associated factors, further research is warranted to clarify the effects of hypoxic tumor-derived EVs on the influx of neutrophils with either antitumor or tumor-suppressive properties. Hypoxia induces the secretion of immunosuppressive factors in EVs: Under hypoxia, cancer cells increase the secretion of various proteins and microRNAs via EVs. By themselves, these factors have been described to suppress immune reactions via inhibition of immune cells' differentiation, proliferation, and effector function, and by inducing the differentiation of immunosuppressive cells. Secretion of these factors via EVs from hypoxic cancer cells might provide novel mechanisms by which these cells can manipulate their surroundings into providing a tumor-promoting environment. Image created with BioRender.com, (accessed on 7 June 2022).

Neutrophils
Neutrophils are the most abundant population of white blood cells in the human circulatory system, reflecting their role as first responders during infection or tissue injury. Elevated neutrophil counts are observed in the peripheral blood and tumors of patients with advanced cancers. In these patients, both tumor-supportive and tumorsuppressive neutrophil subsets have been described (Table 1) [25,26]. The direction of the tumor-associated neutrophil function is dependent on the types of soluble mediators produced by cancer cells. For example, in a transforming growth factor β (TGF-β)-rich environment, neutrophils typically display a tumor-promoting phenotype, whereas they acquire antitumor properties in the presence of interferon β (IFN-β) [25]. To date, no direct effect of hypoxic EVs on neutrophil function has yet been described. Nevertheless, hypoxic cancer-derived EVs have been shown to be enriched in factors known to play a role in neutrophil chemotaxis, including CXCL8 (IL-8) and MIR-451 ( Figure 2A). CXCL8 is a pro-inflammatory cytokine that is enriched in EVs of hypoxic glioma cells [27]. It mediates the recruitment of leukocytes-including neutrophils and myeloid-derived suppressor cells (MDSCs)-to tumors through binding the receptors CXCR1 and/or CXCR2 [28]. Moreover, through activation of MAPK signaling, CXCL8 signaling inhibits neutrophil apoptosis and enhances proliferation [29]. An opposing role has been observed for MIR-451, which inhibits the p38 MAPK signaling pathway, suppressing neutrophil chemotaxis [30]. Given the known diversity in pro-or antitumor neutrophil subpopulations and the varying effects of the different hypoxic-EV-associated factors, further research is warranted to clarify the effects of hypoxic tumor-derived EVs on the influx of neutrophils with either antitumor or tumor-suppressive properties.

Macrophages
Macrophages, as phagocytosing cells of the innate immune system, play an important role in maintaining tissue homeostasis. A variety of macrophage phenotypes have been identified, each with different effector functions [31]. Importantly, tumor-associated macrophages (TAMs) are functionally distinct from the normal and tissue-resident populations [32]. A commonly used paradigm categorizes macrophages into pro-inflammatory (i.e., antitumor) and an anti-inflammatory (i.e., tumor-promoting) phenotypes, also known as the 'M1' and 'M2' subtypes, respectively [33,34]. These macrophage subtypes are characterized by different metabolic profiles, reflecting their functions [35,36]. M1 activated macrophages perform aerobic glycolysis, with reduced activity of their respiratory chain which, in turn, allows the production of reactive oxygen species (ROS) to support their pro-inflammatory function [35,36]. Conversely, M2 macrophages rely mostly on oxidative metabolism for their energy production [35,36]. Forcing macrophage metabolism towards either glycolysis or oxidative metabolism also skews their phenotype towards M1 or M2, respectively [35]. TAMs share many phenotypic and functional characteristics with the M2 subtype, as they facilitate tumor cells' proliferation, invasion, intravasation, and metastatic dissemination. Moreover, TAM-derived factors have proangiogenic and immunosuppressive effects [37]. In vitro stimulation can enrich cells with these phenotypes on either side of the spectrum. However, in vivo macrophage phenotypes are more dynamic and complex [38,39]. Accumulating evidence suggests that tumor-derived EVs mediate M2 polarization of macrophages, and thereby contribute to cancer progression [40,41].
Other factors that can alter macrophage behavior include MIR-21, MIR-23, MIR-92A, MIR-127, MIR-135, MIR-210, MIR-494, MIR-1246, and carbonic anhydrase IX (CAIX). These factors are enriched in hypoxic EVs (Table 2), and their effects may therefore be mediated through EV secretion, although this has not been reported directly ( Figure 2B). CAIX is an enzyme that plays an important role in the acidification of the extracellular environment [49]. This extracellular acidosis is known to reduce the function of various immune cells, including T cells, natural killer (NK) cells, and dendritic cells [50]. In macrophages, this acidification causes metabolic reprograming, pushing them towards the oxidation of fatty acids [51]. This, in turn, skews them towards the tumor-promoting M2 phenotype [51]. MIR-21, MIR-23, and MIR-494 affect macrophage behavior by downregulating the expression of PTEN [52][53][54][55]. The subsequent suppression of NF-κB and AKT activity pushes the macrophages towards an M2 phenotype, with reduced production of tumor necrosis factor alpha (TNF-α) and IL-1β, and increased production of IL-10, CD206, and PD-L1 [52][53][54][55]. In addition, these macrophages induce T-cell apoptosis [54]. Several other factors enriched in EVs derived from hypoxic cancer cells have the potential to alter the secretome of macrophages to support a pro-tumor phenotype. Macrophage-derived pro-inflammatory factors, such as TNF-α and ROS, may be suppressed by EV-derived MIR-135 [56]. Tumor-EV-derived MIR-92A has also been shown to alter macrophages to increase their production of IL-6 which, in turn, stimulates cancer cells' proliferation, migration and invasion [57]. For other factors, the direction of stimulation is less clear, and may be dependent on the context or additional signaling. For example, MIR-127 can target the expression of CD64 and TRAF1 to support an anti-inflammatory phenotype in macrophages with decreased production of pro-inflammatory TNF-α, IL-1β, and IL-6 [58,59]. On the other hand, MIR-127 has also been associated with M1 macrophage polarization via targeting of BCL-6 and subsequent regulation of DUSP1 expression and JNK activity [60]. Further research is thus required to clarify its role in macrophage polarization.
In addition to altering macrophage function as described above, induction of macrophage cell death or prevention of full macrophage differentiation are alternative mechanisms that can suppress macrophage-mediated antitumor reactions. Under hypoxia, cancer cells secrete increased levels of MIR-210 and MIR-1246 via their EVs [44,61]. MIR-210 targets DECR1 and promotes mitochondrial dysfunction and necroptosis in macrophages [62]. MIR-1246 targets the expression of caveolin-1, which is required for the differentiation of monocytes into macrophages [63,64]. Downregulation of caveolin-1 in macrophages increases their VEGF-A/VEGFR1 signaling activity and downstream expression of matrix metalloproteinase 9 (MMP9) and colony-stimulating factor 1 (CSF1), thereby facilitating angiogenesis and metastasis formation [65].
In addition to miRNAs, hypoxic cancer-cell-derived EVs are enriched in pyruvate kinase M2 (PKM2). PKM2 is a key enzyme in the glycolytic pathway whose functionality changes depending on its polymerization [66]. As a tetramer, it has a high affinity for its substrate (phosphoenolpyruvate), and stimulates glycolysis [66]. However, as a dimer, it has lower enzymatic activity, but is able to translocate to the nucleus, where it acts as a protein kinase of several transcription factors [66]. It has been described that EV-associated PKM2 can determine macrophage cell fates. THP1 monocytes cultured with normoxic EVs containing PKM2 display increased glycolysis through regulation of hypoxia-inducible factor 1α (HIF-1α), which induces the differentiation of monocytes to macrophages [67]. Moreover, macrophage differentiation is further controlled by the translocation of PKM2 to the nuclei of monocytes, where it phosphorylates STAT3 [67]. Phosphorylated STAT3 is required for the production of the differentiation-associated transcription factors MAFB, C-MAF, and EGR-1. In these macrophages, M1 markers are downregulated and M2 markers are upregulated [67]. Similarly, PKM2 was shown to inhibit LPS-induced pro-inflammatory M1 macrophage polarization, while promoting M2 macrophage traits through the inhibition of IL-1β production and boosting the production of IL-10 [68]. The increased levels of PKM2 found in the EVs of hypoxic cancer cells might further enhance the abovementioned effects, and thereby induce a stronger immunosuppressive phenotype in recipient monocytes compared to normoxic EVs. In conclusion, hypoxic cancer-cell-derived EVs are enriched in numerous factors able to affect macrophage differentiation, polarization, and migration and, as such, contribute to cancer progression ( Figure 2B).

Myeloid-Derived Suppressor Cells
MDSCs are a heterogeneous group of immature myeloid immune cells that can increase in number during pathological conditions such as chronic inflammation and cancer [69]. They have attracted great scientific interest due to their ability to suppress Tcell-mediated tumor clearance and foster tumor progression through the production of inducible nitric oxide synthase (iNOS) and arginase, among others [70]. MDSCs have high plasticity, and can further differentiate into dendritic cells, tumor-associated macrophages, and granulocytes within tumor environments [71]. In addition to their T-cell-suppressive function, MDSCs have a pro-angiogenic secretory profile, and stimulate cancer-supporting inflammation [72]. It has been shown that cancer-derived EVs can stimulate the development of MDSCs as well as their tumor-supportive effector functions [73,74]. Interestingly, vesicles derived from hypoxic glioma cells have a greater MDSC-inducing capacity compared to normoxic EVs, associated with increased production of immunosuppressive factors such as TGF-β, IL-10, nitric oxide (NO), ROS, and arginase [73,74]. Consequently, these MDSCs inhibit CD8 + T-cell proliferation more efficiently compared to MDSCs stimulated with normoxic control EVs (Table 3) [73,74]. These immunosuppressive effects of hypoxic glioma EVs have been attributed to MIR-10a and MIR-21, which reduce the expression of RAR-related orphan receptor alpha (Rora) and Pten, respectively [73]. The subsequent translocation of NF-κB into the nucleus and increased AKT activation induce MDSC activation [73]. In addition, increased levels of MIR-29a and MIR-92a within the hypoxic glioma-derived EVs decrease the expression of the high-mobility group box transcription factor 1 (Hbp1) and protein kinase CAMP-dependent type I regulatory subunit alpha (Prkar1α) genes in MDSCs, respectively [74]. Hbp1 inhibits cell-cycle progression in the G1 phase, while Prkar1α is part of the PKA/p-STAT3 pathway, where it regulates the production of ROS, TGF-β, and IL-10 [74]. Targeting these genes through miRNAs results in increased proliferation, and increases the production of immunosuppressive factors by MDSCs [74]. In conclusion, tumor-derived EVs-especially hypoxic populations-play important roles in the development and function of pathological MDSCs, which protect tumor cells from antitumor immune reactions ( Figure 1B).
In addition, various other factors known to modulate MDSC function are increased in EVs derived from hypoxic cancer cells, including MIR-210, MIR-494, C-C motif chemokine ligand (CCL)2, and insulin-like growth-factor-binding protein 3 (IGFBP-3) ( Table 3) [75][76][77]. MIR-210 enhances the immunosuppressive effects of MDSCs via increases in Arg1 expression, ARG activity, and NO production [75]. Consequently, overexpression of MIR-210 in MDSCs results in increased tumor growth due to increased T-cell suppression [75]. MIR-494 reduces PTEN expression in MDSCs, subsequently activating AKT, NF-κB, and mammalian target of rapamycin (mTOR) which, in turn, potentiate their immunosuppressive function, thereby supporting tumor growth [76]. Murine models have shown that CCL2 stimulates the production of ROS and NO by MDSCs, which subsequently suppress T-cell proliferation [78]. In humans, stimulation of CCL2 production in monocytes results in enhanced differentiation towards MDSCs [79]. IGFBP-3, enriched in hypoxic cancer-cell-derived EVs, is a factor known to induce a subpopulation of CD38 high MDSC cells with an even more immature phenotype compared to MDSCs lacking CD38. Subsequently, these CD38 high MDSCs have a greater capacity to inhibit activated T cells [80]. Transfer of these factors from hypoxic cancer cells via EVs may induce the development of MDSCs with a great T-cell-suppressive capacity, thereby supporting tumor growth ( Figure 2C). Yes: a direct EV-mediated effect on the immune cell has been identified. No: no direct EV-mediated effects have yet been described, but the miRNA or protein has known immunomodulatory functions.

Dendritic Cells
Conventional antigen-presenting dendritic cells (cDCs) are a crucial bridge between the innate and adaptive immune responses. DCs originate from hematopoietic stem cells. These progenitor cells differentiate into immature DCs, which infiltrate peripheral tissues to sample antigens [81]. Upon encountering a foreign antigen, immature DCs transition into mature DCs [82]. Three major subtypes of DCs have been described: The cDC1s are able to cross-present antigens to CD8 + T cells, resulting in the activation of the cytotoxic T cells and, subsequently, the activation of Th1 CD4 + T cells [83,84]. The cDC2s are preferentially involved in the activation of CD4 + T cells [85]. Therefore, successful DC maturation upon encountering an antigen is important, since DC-mediated priming of T cells is required for unleashing antitumor T-cell responses. The plasmacytoid DCs (pDCs) are the third and rarest subset of DCs; pDCs differ from the other cDC subsets in terms of development and antigen-presenting function. However, the production of high levels of type I IFNs by pDCs, which is observed in response to viral infections, may also play a role in antitumor immunity [86,87]. The currently available studies mainly demonstrate the effects of hypoxia on the cDC subsets.
Tumor-derived EVs exert immunosuppressive effects by reducing the ability of DCs to orchestrate the adaptive antitumor immune response (Table 4). Breast-cancer-derived EVs block the differentiation of myeloid precursor cells into DCs, and this effect is at least partially mediated via the induction of IL-6 in recipient myeloid precursor cells/monocytes [88]. Moreover, melanoma-derived EVs suppress DCs' maturation in vitro, as indicated by impaired expression of CD83 and CD86, as well as decreased expression of chemokines inducing Th1 polarization. Although various factors known to affect DC differentiation have been identified in these EVs, the exact cargo responsible for the observed effects remains to be elucidated [89]. Furthermore, the contribution of hypoxia to this EV-mediated suppression of DC differentiation is currently unknown. Reduces production of the pro-inflammatory cytokines IL-6, IL-12, and TNF-α. Expression of miR-301 in DC represses the release of IFN-γ from DC-primed CD8 + and CD4 + responder cells. No
No 1 Yes or no refers to the current knowledge regarding the respective hypoxic EV-enriched proteins or miRNAs. Yes: a direct EV-mediated effect on the immune cell has been identified. No: no direct EV-mediated effects have yet been described, but the miRNA or protein has known immunomodulatory functions.
Hypoxic cancer-cell-derived EVs are enriched in MIR-301 and MIR-451. Although no direct evidence for their role in EV-mediated immunosuppression exists, both miRNAs have been shown to reduce the production of pro-inflammatory cytokines such as IL-6, IL-12, CCL3/MIP1a, CCL5/RANTES, and TNF-α in dendritic cells, which could result in less activation of Th1 adaptive antitumor immune responses [90,91]. Furthermore, expression of MIR-301 in DCs represses the release of IFN-γ from DC-primed CD8 + and CD4 + responder cells [90]. In addition, increased levels of CCL2 in hypoxic EVs could affect the differentiation of monocytes into DCs [92]. Supplementation of a standard DC differentiation cocktail (GM-CSF and IL-4) with CCL2 resulted in lower production of IL-12 which, in turn, hampered effective IFN-γ-mediated T-cell cytotoxicity [92]. Collectively, these data show how hypoxic cancer-derived EVs, via transfer of MIR-301, MIR-451, and CCL2, may be able to interfere with dendritic cells' functioning, resulting in reduced antitumor immune reactions ( Figure 2D).

NK Cells
NK cells are part of the innate immune system, and are among the first responders in the inflammatory cascade. They have the capacity to recognize and kill infected, neoplastic, and/or malignantly transformed cells. NK cells recognize a potential target cell through interaction between their membrane-associated receptors and the activating or inhibitory ligands expressed by the target cell [93]. If the signaling through activating NK cell receptors is higher compared to the signaling through inhibitory receptor-ligand interactions, NK cells' effector functions are triggered. These include NK cell degranulation and release of their toxic granules containing perforin and granzymes, killing of target cells via death receptors, and the secretion of IFN-γ. NK cells are controlled by a very broad panel of inhibitory and activating receptors, although the most important group of inhibitory ligands are the major histocompatibility complex class I (MHC-1) molecules, which are expressed in virtually every healthy cell [93].
Cancer cells have been shown to evade NK-mediated killing. In several cancer models, EVs have been shown to regulate NK cells' antitumor effector functions and, depending on the exact makeup of the vesicles, this results in either stimulation or an inhibition of NK cells' antitumor functions, as reviewed in [94]. Moreover, tumor hypoxia has been identified as one of the important factors contributing to tumor cells' resistance to NK cell killing [95]-presumably because hypoxia enhances tumor cell autophagy, which in breast cancer cells has been shown to reduce NK cell killing via the breakdown of NK-secreted granzymes in the tumor cells [96]. Although the impact of (hypoxic) EVs on NK cells is only very poorly understood, one study showed that hypoxic lung-cancer-cell-derived EVs may reduce NK-cell-mediated cytotoxicity via transfer of the immunosuppressive cytokine TGF-β, which decreases IFN-γ expression and secretion, and decreases surface expression of the activating receptor NKG2D. Additionally, these EVs transfer MIR-23A, which directly targets the expression of degranulation-associated CD107a, resulting in a lower proportion of degranulating NK cells when encountering a target cell (Table 5) [97] (Figure 1C). Table 5. EV-enriched factors and their effects on NK cells.

Factor
Effect on NK Cells Proven to Be EV-Mediated? 1

TGF-β [97]
Inhibits NK cell function by decreasing surface expression of the activating receptor NKG2D and decreasing IFN-γ production. Yes

MIR-23a [97]
Decreases the percentage of degranulating NK cells. Yes 1 Yes or no refers to the current knowledge regarding the respective hypoxic EV-enriched proteins or miRNAs. Yes: a direct EV-mediated effect on the immune cell has been identified. No: no direct EV-mediated effects have yet been described, but the miRNA or protein has known immunomodulatory functions.

T Cells
T cells are part of the adaptive arm of the immune response, and express either the CD4 or CD8 glycoprotein, which roughly subdivides them into helper T cells and cytotoxic T cells. Upon activation, these T cells undergo clonal expansion to generate a large pool of antigen-specific lymphocytes, which then further differentiate into specialized effector T cells [98]. Naïve CD4 + T-helper cells can further differentiate into subsets, which are distinguishable based on their surface marker expression and cytokine profile. The principal subsets of CD4 + T cells that play key roles in cancer development are T-helper (Th)1, Th2, Th17, and regulatory T (Treg) cells [99]. These CD4 + T-cell subsets can play opposing roles in cancer progression. Th1 cells secrete pro-inflammatory type 1 cytokines (including IFN-γ, TNF-α, CCL2, and CCL3), which support antitumor immune reactions and CD8 + T-cell activation. In contrast, Th2 cells secrete type 2 cytokines (such as IL-4, -5, and -13), which promote type II immune responses (e.g., M2 macrophage polarization), tissue remodeling, and tumor growth. The role of Th17 in tumor progression remains controversial, with reports of both tumor-promoting angiogenic effects and antitumor immune reactions [99,100]. Lastly, Treg cells suppress immune responses by producing cytokines (e.g., IL-10, TGF-β) that inhibit the activation of lymphocytes, dendritic cells, and macrophages [98]. Furthermore, they may render antigen-presenting cells unable to provide the co-stimulatory signals needed for T-cell activation, thus suppressing antitumor immune reactions [98]. Cancer cells modulate these different T-cell subsets in various ways to support tumor growth, including via the release of EVs (Table 6) [101,102]. EVs derived from nasopharyngeal carcinoma cells have been shown to suppress the proliferation of CD4 + and CD8 + T-cell subsets and inhibit the differentiation towards Th1 and Th17 subtypes [103]. In addition, they increase the percentage of FOXP3-positive Treg cells, skewing the TME towards tumor-supportive immunosuppression [103]. These effects can be attributed to vesicular MIR-24-3p, and are more pronounced in hypoxic conditions [103] ( Figure 1D).

Hypoxic EVs May Limit Differentiation and Regulate Polarization toward Tumor-Supportive Subtypes via Transfer of Factors with Known Immunomodulatory Roles
Hypoxic EVs are enriched in a number of factors for which immunomodulatory roles have been described independent of EV-mediated communication. Factors that can alter T-cell differentiation and maturation include MIR-125, MIR-210, and thrombospondin-1 (TSP-1) [61,73,104]. MIR-125 suppresses CD4 + T-cell differentiation and maintains a naïve T-cell state by decreasing the expression of Ifn-γ, TNF-α, IL-2Rβ, IL-10Rα, PR domain zinc finger protein 1 (BLIMP-1), Stat3, and Il-13 [104][105][106]. Furthermore, MIR-125-mediated suppression of STAT3, IL-13, and IFN-γ stabilizes Treg lineage commitment, which contributes to the formation of an immunosuppressive TME [105]. Similarly, TSP-1 induces the differentiation of immunosuppressive Treg cells [107]. HIF-1α plays an important role in the polarization of T cells and regulates the balance between Th17 and Treg polarization [108,109]. MIR-210, upregulated in the EVs from hypoxic cancer cells, reduces the expression of HIF-1α, which subsequently suppresses Th17 differentiation and reduces inflammation [110]. Conversely, PKM2 promotes Th17 differentiation through STAT3 activation [111]. However, as mentioned above, the role of Th17 cells in cancer immunity is still paradoxical, and further research is necessary to understand their contribution to pro-versus antitumor immune reactions. In summary, hypoxic EVs can contribute to the immunosuppressive TME by interfering with the development of functional T cells, or by skewing their polarization into tumor-supportive subtypes.

Hypoxic EVs May Decrease the Abundance of T Cells (Infiltration and Proliferation) in the Tumor and Limit Antitumor Immunity via Transfer of Factors with Known Immunomodulatory Roles
Reducing T-cell-mediated cytotoxic efficacy yields efficient suppression of antitumor immune reactions. This reduced cytotoxic efficacy could be induced through hypoxic-EVmediated communication between tumor cells and target cells. As such, MIR-23, TGF-β, and CAIX are enriched in hypoxic EVs, and these factors can limit cytotoxic T-cell effector function [97,[112][113][114][115][116]. MIR-23 blunts the expression of multiple key cytotoxic T-lymphocyte (CTL) effector molecules, including IFN-γ, thereby suppressing T-cell-mediated cytotoxicity and, subsequently, accelerating tumor progression and increasing the tumor burden [117]. These effects are mediated by suppression of BLIMP-1-a key regulator in the activation of cytotoxic T-cell immune reactions [117][118][119]. TGF-β increases the expression of MIR-23 in CTLs, thereby further contributing to immunosuppression [97,114,117]. CAIX, as described above, plays an important role in extracellular acidification [49]. In T cells, this acidification inhibits glycolysis which, in turn, induces anergy, with impaired proliferation and reduced cytolytic activity and cytokine release [120][121][122]. In addition, extracellular acidification has also been postulated to suppress T-cell differentiation through the suppression of mTORC1 [50]. mTORC1 plays an important role in the differentiation of effector T cells, and its inhibition skews their differentiation towards immunosuppressive Treg cells [123,124]. Similarly, high CAIX expression in tumors is correlated with high infiltration of FOXP3 + Treg cells [125].
In addition to a reduction in T-cell-mediated killing efficacy, suppression of T-cell proliferation also severely reduces the efficacy of the immune response. MMP9 and Let-7a are enriched in hypoxic EVs, and have the capability to inhibit T-cell proliferation [42,126]. MMP9 mediates shedding of the IL-2 receptor-α from the surface of T cells, thereby preventing CD4 + CD8 + T-cell proliferation [127]. Let-7a reduces both the infiltration of T cells into the tumor and CD3 + T-cell proliferation [128,129]. Moreover, Let-7a reduces the secretion of IFN-γ by T cells via decreased STAT3 expression [129]. This indicates that, in addition to the crude number of cells present, EV-associated factor Let-7a inhibits immune cells' effector function, presenting another approach via which EV-associated factors can modulate tumor immunology.
Alternatively, various factors enriched in EVs of hypoxic cancer cells have been shown to limit the infiltration of immune cells into the tumor, including IGFBP-3, TSP-1, and a disintegrin and metalloprotease with thrombospondin motif 1 (ADAMTS1) [27,113]. In mice with mammary tumors deficient in IGFBP-3, the intratumoral gene expression levels of Ifn-γ, Cd8, and Tnf-α were elevated, indicating a higher infiltration of CD8+ cytotoxic T cells into the tumor [130]. In triple-negative breast cancer, TSP-1 expression was observed to be inversely correlated with the infiltration of CD8+ lymphocytes. Moreover, this was recapitulated by TSP-1 knockdown in the 4T1 metastatic mouse model [131]. Transfer of IGFBP-3 via (hypoxic) cancer-derived EVs might thus suppress immune infiltration into the tumor and, as such, stimulate tumor growth. Another hypoxia-enriched factor in EVs that affects both the infiltration and effector function of T cells is ADAMTS1-an extracellular protease for which both anti-and pro-tumorigenic functions have been described [132]. Knockout of ADAMTS1 in mammary and melanoma cancer models resulted in reduced primary and secondary tumor burden, which correlated with increased infiltration of cytotoxic lymphocytes and increased expression of genes involved in antitumor immune reactions [133,134]. However, this response may be dependent on the expression and abundance of ADAMTS1 substrates such as the matrix proteoglycan versican (VCAN) in the TME. Intact VCAN triggers the secretion of cytotoxic T-cell-inhibitory cytokines by antigen-presenting cells, whereas its proteolytic product versikine triggers the secretion of cytotoxic T-cell-activating cytokines. However, the balance between ADAMTS1 s proand anti-inflammatory effects is currently not fully understood, leaving the exact immune response in these settings unknown [135]. In conclusion, hypoxic cancer-cell-derived EVs may play a role in the evasion of antitumor immune reactions by reducing T-cell functionality and by limiting the proliferation and infiltration of cytotoxic T cells, while stimulating the recruitment of immunosuppressive regulatory T cells ( Figure 2F). Table 6. EV-enriched factors and their effects on T cells.
Increases the FOXP3 + Treg cell population.
Hinders T-cell differentiation via extracellular acidification. No

NKT Cells
NKT cells are a developmentally and functionally distinct lineage of T cells that recognize lipid antigens, such as glycolipids and glycerol lipids, presented by CD1dan MHC class Ib molecule. Different NKT cell subtypes with opposing functions have been described. Type I NKT cells have effective antitumor effector functions. On the other hand, type II NKT cells are regulators of immunosuppression [136]. Hypoxic EVassociated factors may regulate the skewing of NKT-cell subsets (Table 7). While there are no studies directly investigating the effects of hypoxic EVs on NKT cells, it is known that hypoxic cancer cells increase their EV-mediated secretion of MIR-92A [126]. In another study, MIR-92A has been shown to induce the expression of IL-6 and IL-10 in NKT cells, reducing their direct antitumor effects and leading to NKT-mediated suppression of CD8 + T-cell proliferation [137]. As such, increased loading of MIR-92A into EVs under hypoxic conditions may promote tumor growth via the induction of immunosuppressive cells ( Figure 2F). Table 7. EV-enriched factors and their effects on NKT cells.

Factor
Effect on NKT Cells Proven to Be EV-Mediated? 1 MIR-92a [137] Induces immunosuppressive NKT cells with reduced antitumor effects via increased IL-6 and IL-10 expression.
No 1 Yes or no refers to the current knowledge regarding the respective hypoxic EV-enriched proteins or miRNAs. Yes: a direct EV-mediated effect on the immune cell has been identified. No: no direct EV-mediated effects have yet been described, but the miRNA or protein has known immunomodulatory functions.

B Cells
B cells play a key role in immunological memory of infectious diseases as long-lived and antibody-producing plasma cells or memory B cells that are capable of responding to reinfections. High abundances of tumor-infiltrating B cells in solid cancers are commonly found [138]. However, their exact role in the TME is not yet clear, as both tumor-growthpromoting and antitumor immune reactions have been described [138]. Several factors known to suppress B-cell functioning are enriched in the EVs of hypoxic cancer cells, including MIR-210 and MIR-125 (Table 8) [61,73]. MIR-210 acts as a negative regulator of B-cell-mediated immune responses, and fine-tunes the balance between pathogen clearance and autoimmunity [139]. Its overexpression results in impaired B-cell proliferation and antibody production [139]. In addition, MIR-125 prevents B cells' maturation and release from the bone marrow [73,140]. As such, delivery of MIR-210 and MIR-125 to B cells via EVs may suppress B-cell-mediated immune reactions ( Figure 2G). However, as the role of B cells in tumor immunity is yet to be fully elucidated, further research is required [138]. Table 8. EV-enriched factors and their effects on B cells.

Factor
Effect on B Cells Proven to Be EV-Mediated? 1 MIR-210 [139] Impairs B-cell proliferation.

Reduces antibody production. No
MIR-125 [73,140] Prevents B cells' maturation and release from the bone marrow.

Immune Stimulation by Hypoxia-Upregulated Factors in EVs
Various factors enriched in hypoxic cancer-cell-derived EVs also stimulate the immune system, and could therefore induce antitumor immune reactions (Table 9). MIR-181 enhances the development of NK cells by downregulating the expression of nemo-like kinase (NLK)-a negative regulator of the Notch signaling pathway [141,142]. Activation of the Notch signaling pathway stimulates both early and late phases of NK cells' development [143][144][145]. In addition, MIR-155 targets the expression of SH2 domain-containing inositol 5 -phosphatase 1 (SHIP1) in NK cells [146]. SHIP1 is a negative regulator of pro-inflammatory IFN-γ production, and its downregulation by MIR-155 could therefore enhance antitumor immune responses [146]. Furthermore, MIR-155 is required to elicit an effective antitumor immune response in macrophages, as it targets the expression of IL-13Rα1. This reduces the capacity to respond to M2-inducing signals, and pushes the macrophages towards the M1 phenotype via targeting of interferon regulatory factor 4 (Irf4) [147][148][149][150][151][152]. In CD8 + T cells, MIR-155 plays an important role in antitumor immunity, as it regulates the expression of several critical activation and effector genes, including Ifn-γ, Granzyme B, Perforin, Klrg, Cd62l, Ctla4, and Pdcd1 [153]. Furthermore, MIR-155 reduces Socs1 expression, which modulates CD8 + T cells' responsiveness to IL-2, IL-7, and IL-15 stimulation [154]. Taken together, these observations show the dual nature of hypoxiaupregulated factors in EVs in both immunosuppression and immune stimulation. These contradictory roles could be explained by the existence of different EV subpopulations, which contain different protein and miRNA cargoes [155,156]. Consequently, these different subpopulations might be selectively taken up by specific recipient cells; as such, one batch of EVs can elicit different effects, depending on the receptor cells. The complexity and heterogeneity of EVs warrants further investigation into the different subpopulations of EVs and their pathological roles, as a more thorough understanding of their biology will pave the way towards new therapeutic strategies and biomarker development. Table 9. Immune stimulation by hypoxia-upregulated factors in EVs.

Factor
Immune-Stimulating Effect

MIR-181
Enhances the development of NK cells.
Reduces the capacity of macrophages to respond to M2-inducing signals.

Importance of EV Isolation Methodology and Experimental Setup
Aiming to increase the reproducibility and reliability of published EV results, the International Society for Extracellular Vesicles has established the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines [16]. These guidelines provide recommendations on EV isolation, characterization, and functional studies, and emphasize the importance of reporting requirements specific to the EV field [16]. Despite these efforts, many studies still base their conclusions on experimental procedures that may not be optimal for their purpose. One of the main problems is the use of EV isolation techniques that disregard EVs' purity and morphology/functionality. For example, ultracentrifugationone of the most widely used methods to isolate EVs-is known to suffer from various drawbacks, such as the incomplete sedimentation of EVs, co-isolation of non-EV protein/RNA impurities and aggregates, and possible damage to EVs [157][158][159][160][161]. The same holds true for precipitation-based isolation strategies [157]. The impacts of the different isolation methodologies on the obtained results have been extensively described previously [157][158][159][160][161][162][163][164]. In addition, the field of hypoxia research also suffers from diversity in experimental setup, with different oxygen concentrations and chemical models such as CoCl 2 being applied to simulate tumor hypoxia. As results obtained using different EV isolation methods and experimental setups may introduce variability, we summarize the details of the discussed studies in Tables 10 and 11.   200× g 30 min), 300 kDa MWCO centrifugation at 4000 g, centrifugation (12,000× g 30 min), flotation on sucrose cushion (5.5% sucrose, 100,000× g 15 h).

Conclusions
In conclusion, tumors use EVs to communicate with immune cells. Under the influence of hypoxia, the composition of these EVs is altered. In these EVs, increased levels of immunomodulatory proteins are frequently observed, resulting in immunosuppression and tumor persistence. Moreover, evidence reveals that modulation by miRNAs plays an important role in regulating the immune compartment of the tumor, where the delivery of modulators by hypoxic EVs mostly results in immunosuppression. In contrast, some EV-enriched factors may support antitumor immune responses. However, other regulatory mechanisms may be at play that prevent efficient uptake by the specific recipient immune cells, or the effects induced by these factors may not be sufficient to prevent tumor outgrowth.