**4. Recommendation—EVs as drug carriers**

In providing a recommendation to engineer EVs as DDSs, EV engineering methods to overcome specific barriers can be deployed. Natural evasion of immune cells is a highly favourable quality and should mark all engineered EVs regardless of the barrier(s) they target. CD47, CD24, CD31 and PD-L1 are prominent surface proteins that achieve this quality and should be incorporated into engineered EVs in high amounts if not originally present [129–134]. Fusing EVs with liposomes to create hybrid DDSs can also increase their drug loading capacity without risking cargo aggregation [198, 199].

The choice of the source of EVs depends on its availability and the ability of its EVs to overcome respective biological barriers associated with the disease. Milk and plant-derived EVs, which are highly available in nature and able to overcome gastrointestinal barriers [76, 77, 85–105], can be engineered for drug delivery *via* the oral route to treat IBD. The patient's own EVs might also be used as a form of personalised medicine. EVs from the patient may be chosen based on whether their cell of origin matches the target cell for better selectivity, but there are exceptions. For instance, immune and intestinal cell-derived EVs can be internalized by placental cells [103, 128]. Breast milk might also be a possible EV drug carrier source to treat

both gastrointestinal and neurological conditions, as milk-derived EVs can cross gastrointestinal barriers [76, 85–90] and the BBB [87, 96] respectively. Human Type O RBC EVs loaded with antisense oligonucleotides were also found to target human leukaemia and breast cancer cells *in vitro* and *in vivo*. This is advantageous as RBCs are widely accessible from blood banks and lack DNA, which ensures that no oncogenic material is transferred from EVs to target cells [200]. This offers diverse compatible EV sources to choose from for a single ailment, enhancing the flexibility of the engineering process.

A summary of the recommendation is shown in **Figure 4**.

#### **Figure 4.**

*Recommendation on how to engineer extracellular vesicles (EVs) to overcome various biological barriers for drug delivery applications. (1) Gastrointestinal barriers can be overcome by deploying milk and plant-derived EVs to withstand the harsh acidic and enzymatic conditions, while also ensuring that EVs possess surface glycoproteins to enable trans-endocytosis across the small intestinal barrier. (2) The placental barrier can be overcome directly via the use of placental and THP-1 monocyte EVs, or indirectly via engineering watermelon EVs to target intestinal epithelial cells (IECs) which can communicate with the placenta. (3) Immunological barriers can be overcome by having a high proportion of CD47, CD24, CD31 and PD-L1 to produce the "don't eat me" signal, and probably a low level of phosphatidylserine (PS) on the surface of EVs to minimize the chances of being engulfed by macrophages. (4) Neurological barriers can be overcome minimally by incorporating high amounts of CD46 and CD63 into EVs as the quantity of these tetraspanins are positively correlated with the ability of EVs to cross the blood-brain barrier (BBB). (5) The lymphatic barrier can be overcome by EVs naturally as their lipidic nature enables them to cross the reticular network of the blood-lymph barrier (BLyB). Adding* α*-D-mannose and PEGylating EVs may also enhance their passage across the barrier. (6) The pulmonary barrier may be targeted by EVs derived from bronchoalveolar lavage fluid (BALF), which possess MHC classes I and II, CD54, CD63 and CD86. (7) The renal barrier can be overcome by EVs naturally, probably due to their small size and fluid lipid membranes which might allow them to squeeze through the tiny pores of the glomerular filtration barrier. (8) Cellular barriers can be overcome by EVs naturally via retrograde trafficking, endocytic recycling, direct fusion with the plasma membrane or other mechanisms. Engineering EVs with a size of <100 nm might help to reduce the chances of EVs being internalised via macropinocytosis which tends to lead to lysosomal degradation more than other mechanisms accessible to smaller EVs. Incorporating syncytin-1 and -2 into EVs might also enable them to fuse with the plasma membrane directly, allowing them to evade the endosomal and lysosomal membranes completely (created with BioRender.com).*
