To address some of these challenges in terms of cost-effective production and lack of ethical issues, plants might offer an alternative source of exosomes and EVs. As a result, a growing number of studies are looking into their potential health benefits. For example, the effect of plant extracellular vesicles loaded with curcumin are currently being tested in clinical trials (NCT01294072) to evaluate their impact on surgery of newly diagnosed colon cancer patients (https://clinicaltrials.gov/ct2/show/NCT01294072).

Unfortunately, plant unconventional protein secretion pathways have attracted only late interests and our current knowledge of plant UPS and EVs is growing but still limited (Ding et al., 2014a; Robinson et al., 2016; Hansen and Nielsen 2017; Cui et al., 2019). The presence of leaderless proteins in apoplastic extracellular vesicles has confirmed that these EVs represent genuine plant UPS pathways involved in cell wall remodelling and resistance to infection (Delaunois et al., 2013, 2014). Investigations around these vesicular mechanisms have uncovered the existence of at least three pathways that result in the release of extracellular vesicles in plants: exocyst-positive organelle mediated secretion (EXPO), vesicle budding from the PM (including microvesicles), and multivesicular body (MVB)-PM fusion (Wang et al., 2010; Regente et al., 2012; Cui et al., 2019). A growing number of studies report the beneficial effect of crude and pure extracts of plant EVs on human health (Akuma et al., 2019; Alfieri et al., 2021; Urzì et al., 2021). To evaluate their potential as drug delivery tools, the current state of the field in terms of plant EVs classification, purification, and biomedical applications is presented below.

The processes described to isolate plant EVs depend on the nature of the plant material. Apoplastic fluids are usually extracted from leaves, while blending/juicing is performed on fruits or roots. EVs can also be isolated from liquid plant exudates (Araya et al., 2015). Although the purities of different EV fractions will vary, they have all been found to have therapeutic potential in biomedicine.

The apoplast is the space outside the plasma membrane of plant cells where material can freely move (Sattelmacher, 2001). Although it is unknown how EVs cross the cell wall, their presence in the apoplast has been confirmed (Regente et al., 2012; Rutter et al., 2017; He et al., 2021). To recover these vesicles, a standard technique based on vacuum-infiltration and ultracentrifugation is performed (O’Leary et al., 2014). Applying sequential rounds of negative and atmospheric pressure onto leaves forces a buffer into the apoplastic space that can be recovered after centrifugation of the leaf. This method ensures that plant cells remain mostly undamaged and results in a relatively pure fraction containing EVs but depleted of intracellular components. It has been mostly used to purify EVs from leaf material (Arabidopsis thaliana, Nicotiana benthamiana) or seeds (sunflower) (Regente et al., 2009; Rutter and Innes, 2017; Zhang et al., 2020). Additional purification steps will allow further isolation of different types of EVs as described above (Regente et al., 2009; Rutter and Innes, 2017; He et al., 2021). Recently, a comparative analysis of two major methods for isolating EVs from apoplastic wash fluids has provided a guide into the selection of the right method adapted to the type of downstream applications desired (Huang et al., 2021).

Enriched EV fractions have been obtained through blending plant matter such as ginger roots, herbs, wheat, and dandelion (Mu et al., 2014; Xiao et al., 2018; Chen et al., 2019). Juicing of citrus fruits, pears, grapefruit, watermelons, and coconut water has also been used to prepare EV extracts (Liang et al., 2015; Raimondo et al., 2015; Xiao et al., 2018; Zhao et al., 2018). However, unless they are subjected to further purification steps, these methods often result in a mix of EVs and intracellular content (vesicles, organelles, membranes), meaning they are not solely products of UPS (Pinedo et al., 2021). They are, therefore, referred to as Plant-derived nanovesicles (PDNVs) or Exosome-like nanovesicles (ELNs) rather than EVs which refer to the purer fractions. There is increasing evidence that these PDNVs have significant biological effects on human cells and have brought new hope into novel forms of natural drug delivery systems (Di Gioia et al., 2020; Alfieri et al., 2021; Urzì et al., 2021).

Plant exudates are substances excreted from plants that include liquids flowing through and out of plants. This includes sap, gum, resins or root exudates. They have been used for many years in traditional medicine. Exudates contain many bioactive compounds, amongst them peptides, with beneficial effects on human health such as reduction of oedema and inflammation (Licá et al., 2018). Plant EVs derived from exudates are a relatively new research topic. EVs isolated from the sap of two plants (namely Dendropanax morbifera, and Pinus densiflora) have shown cytotoxic and anti-metastatic effects on human tumour cells (Kim et al., 2020a; Kim et al., 2020b). Furthermore, EVs from a hydroponic solution containing tomato (Solanum lycopersicum L.) root exudates were shown to inhibit the spore germination of three fungal phytopathogens (Fusarium oxysporum, Botrytis cinerea and Alternaria alternata) suggesting an antifungal activity in plants (De Palma et al., 2020). Whether this activity can be applied to mammalian fungal pathogens has not been tested. More research is needed to understand if exudates EVs could hold promising therapeutic applications.

Plants have been known for centuries to be beneficial for human health. Yet the identification of extracellular vesicles and their molecular content shed a new light on our understanding of cross-kingdom interaction and transfer of bioactive molecules.

In the past decade, numerous reports have described the beneficial effects of plant PDNVs/EVs in mammalian health. While PDNV proteomes from various plant origins have been characterised and some common proteins frequently identified in these vesicles, the variety of PDNVs and the lack of specific protein markers limits their classification which may prove problematic for large scale good manufacturing practices (GMPs). Nevertheless, PDNVs contain a range of bioactive molecules such as proteins, lipids, or metabolites with therapeutic effects summarised in (Woith et al., 2019; Di Gioia et al., 2020; Kocak et al., 2020; Alfieri et al., 2021; Urzì et al., 2021). Amongst the most studied plant PDNVs are those originating from ginger. These EV-containing PDNV isolates have many natural therapeutic potentials and can induce physiological changes in mammals. They were shown to influence the human gut microbiota (Teng et al., 2018), inhibit inflammasome activation (Chen, Zhou and Yu, 2019), and found to have a positive effect on inflammatory bowel disease and colitis-associated cancer (Zhang et al., 2016). They have also been shown to be taken up by, and inhibit the pathogenicity of, the periodontitis-causing Porphyromonas gingivalis (Sundaram et al., 2019). In parallel, wheat derived nanovesicles have been shown to aid in vitro wound healing by promoting proliferation and migration of dermal fibroblasts, endothelial, and epithelial cells (Şahin et al., 2018). Nanovesicles derived from various fruits and vegetables were also shown to inhibit cancer cell growth (Kameli et al., 2021). Despite their numerous health benefits, it is unclear however, if this positive impact is attributable to the combined action of various bioactive components in the crude fraction or to particular compounds that may be isolated from purer EVs preparations.

Research on EVs (obtained from the apoplast of plants) as potential drug delivery systems is far more restricted than those on PDNVs. So far, to our knowledge, only one study has shown that purified apoplastic small EVs (sEVs) are efficiently taken up by human ovarian cancer cells OVAR5 (Liu et al., 2020). This paper compared the uptake of apoplastic sEVs (purified from the apoplast of Arabidopsis leaves) and nanovesicles (obtained from disrupted leaf material). OVAR5 cells were found to be significantly more susceptible to apoplastic sEV uptake than leaf nanovesicle uptake, based on elevated numbers of fluorescent cells. These results suggest that pure EV samples have the same, if not greater, drug delivery potentials than PDNV isolates have, and that EVs may be the contributing factor to PDNV success. Unfortunately, to our knowledge, this is the only study that uses purified apoplastic EVs in human cells and more data is required to conclude. In addition, an assessment of immunogenicity and toxicity should be undertaken to validate pure plant EVs as a drug delivery system.

Based on the successes of PDNVs, efficient uptake of sEVs, and the potential of engineering exosomes in plants, biopharming is an attractive solution to produce cheap pharmaceuticals with a rapid turnover. Biopharming, or plant molecular farming, refers to the use of genetic tools to produce a wide range of pharmaceuticals. Plants have already been used to produce antibodies and vaccines for humans, animals, and aquaculture (Shoji et al., 2012; Takeyama et al., 2015; Yao et al., 2015; Lefebvre and Lécuyer, 2017; Zahara et al., 2017; Su et al., 2021). Recently, plants have been explored as a rapid alternative biofactory for the production of COVID vaccines through the expression of Virus-like particles exposing an immunogenic part of the Spike S protein (Dhama et al., 2020; Maharjan and Choe, 2021). Regarding clinical trials, intravenous administration of β-glucocerebrosidase protein expressed in carrots has been approved as being safe and efficient and successfully used for 2 decades (Shaaltiel et al., 2015). The advantages of using plants as Biofactories include their ability to produce functional proteins in large amounts, and at lower costs (Shaaltiel et al., 2007). One additional advantage is the possibility of relatively simple engineering associated with plants, potentially allowing in vivo packaging of exogenous cargo into EVs, ready for extraction. More data on the mechanisms of loading into plant EVs is still required, but with this possibility in mind, and given that delivery of therapeutic molecules by mammalian EVs has already been demonstrated by several studies (Alvarez-Erviti et al., 2011; Batrakova and Kim, 2016; Elsharkasy et al., 2020), biopharming plants to isolate therapeutic pure EVs is a very exciting avenue that needs to be explored.

If plant EVs are to be potential drug delivery systems, their administration and bioavailability must be considered. The first strong evidence of cross kingdom effects was provided when isolated PDNVs were fed to mice and found to reach intestinal macrophages. The vesicle uptake in these cells increased the expression of interleukins and alleviated colitis symptoms (Ju et al., 2013; Mu et al., 2014). This study has demonstrated that PDNVs are able to resist gastric and intestinal digestion, suggesting oral administration methods of plant nanoparticles are suitable for targeting these organs. In order to reach other organs, alternative administration methods have been investigated. In particular, intravenous injection is normally considered to have the advantage of avoiding the first-pass effect of hepatic metabolism, producing the highest bioavailability. When intravenous administration of edible tea flower nanoparticles was compared to oral administration, no difference was noted in terms of body weight and main pro-inflammatory cytokines levels. However, a sharp increase of complement C3 concentrations was detected, suggesting a slight immune reaction induced by these nanoparticles when they are administered intravenously (IV) (Chen et al., 2022). Other studies have suggested that IV administration of ginger derived exosome-like nanovesicles (GDELN) did not promote an immune reaction, though only body weight was examined (Li et al., 2018). The slight immune reaction induced by repetitive intravenous injection of EVs appears non-specific to plant EVs since a mild immune response has also been reported for human EVs (Saleh et al., 2019). The authors found that EVs purified from different sources could induce different responses. Therefore, this could also be the case for plant EVs, and more information needs to be collected before a conclusion could be drawn on intravenous injections of plant EVs. In parallel, one study has reported that intranasal administration of engineered grapefruit-derived nanovectors (GNVs) could slow down tumour brain progression in mice (Zhuang et al., 2016). This brings hope for the use of plant EVs as therapeutic tools in neurodegenerative diseases. It is noticeable that EV biodistribution changes with the administration method. While intravenous injection of mammalian and plant EVs results in the wide uptake by various organs (including spleen, liver, kidney, lung, heart, and brain) (Lai et al., 2014; Garaeva et al., 2021), the gut is more specifically targeted in oral administration of edible EVs (Ju et al., 2013; Mu et al., 2014; Zhang et al., 2016; Deng et al., 2017; Teng et al., 2018). In addition, plant EVs have been shown to penetrate a human skin model, which encourages their consideration for skin care treatments (Lee et al., 2020). Altogether, the data accumulated suggests that specific administration methods would have to be developed depending on the pathology targeted and that plant EVs present a lot of potential in therapeutic processes.