Milk exosomes in nutrition and drug delivery

Abstract

Exosomes are natural nanoparticles that originate in the endocytic system. Exosomes play an important role in cell-to-cell communication by transferring RNAs, lipids, and proteins from donor cells to recipient cells or by binding to receptors on the recipient cell surface. The concentration of exosomes and the diversity of cargos are high in milk. Exosomes and their cargos resist degradation in the gastrointestinal tract and during processing of milk in dairy plants. They are absorbed and accumulate in tissues following oral administrations, cross the blood-brain barrier, and dietary depletion and supplementation elicit phenotypes. These features have sparked the interest of the nutrition and pharmacology communities for exploring milk exosomes as novel bioactive food compounds and for delivering drugs to diseased tissues. This review discusses the current knowledgebase, uncertainties, and controversies in these lines of scholarly endeavor and health research.

INTRODUCTION

Milk exosomes (MEs) have attracted considerable attention in both the nutrition of infants and adults and their usage in the delivery of drugs to diseased tissues. The interest was sparked by reports that MEs and their cargos resist degradation by harsh conditions (low pH, RNases) in the gastrointestinal tract and during the processing of milk in dairy plants, are absorbed and accumulate in various tissues following oral administrations, cross the blood-brain barrier, and dietary depletion elicits phenotypes of depletion (1–8). This review discusses the current knowledgebase in these lines of scholarly endeavor and is based on a survey of papers listed in PubMed and research grants supported by the National Institutes of Health (NIH RePORTER, https://reporter.nih.gov/) as well as the senior author’s leadership role and knowledge of publications in the field of MEs.

OVERVIEW: CELL-TO-CELL COMMUNICATION BY EXOSOMES

Exosomes constitute one of the four classes of extracellular vesicles (EVs) identified in mammals to date, including apoptotic bodies, microvesicles, exosomes, and exomeres (Table 1). The four classes have in common that the EVs are released from donor cells into the extracellular space through which they travel to adjacent or distant recipient cells. In 2014, The International Society for Extracellular Vesicles released strict guidelines regarding the authentication of exosomes and the society suggested that authors use “EV” rather than “exosome” in publications unless the endosomal origin of EVs is documented (12). Subsequently, ISEV has loosened the stringency of its recommendations regarding terminology and suggested that authors clearly describe the protocols used for EV isolation and authentication, so that readers may form their own opinion whether the preparations constitute a heterogenous mixture of EVs, exosome-rich EVs, or exosomes (13). In this review, we use exosomes when discussing milk because most studies followed ISEV guidelines when authenticating vesicle preparations. Readers are encouraged to use their own judgment as to the identity and purity of preparations studied. Milk is a challenging matrix to work with because small casein and fat globules, in addition to small microvesicles may be difficult to separate and distinguish from exosomes (14).

Table 1.

Extracellular vesicles in mammals

Extracellular Vesicle	Origin	Size	References
Apoptotic body	Dying cells	1,000–5,000 nm	(9)
Microvesicle	Blebbing off plasma membrane	100–1,000 nm	(10)
Exosome	Endocytic system	∼100 nm	(10)
Exomere	Unknown	∼35 nm	(11)

Exosomes may be separated from other EVs and exosome-sized complexes by using a variety of technologies that take advantage of differential sizes, densities, and surface markers such as differential ultracentrifugation, iodixanol gradient centrifugation, size exclusion chromatography, tangential flow filtration and affinity chromatography. The Extracellular RNA Communication program, funded by the National Institutes of Health Common Fund, has endorsed select purification protocols and made them publicly available in form of a decision-making tree (15). Recommendations for authenticating exosome preparations include analysis by immunoblotting and transmission electron microscopy (12, 13). Immunoblots should include positive markers such as the tetraspanins CD9, CD63, and CD81 and luminal proteins such as Alix and Tsg101, and the absence of markers for contamination with microvesicles and lipoproteins and, in case of milk, casein and fat globules. Exosome isolation and authentication protocols may be deposited in the public domain through EV-TRACK to enhance transparency and reproducibility (16). A report that spin columns used for RNA isolation are contaminated with bacterial short RNA could not be reproduced in independent studies and was attributed to a possible laboratory-specific contamination (17–19).

This review focuses on exosomes because of their unique features in cell-to-cell communication. Exosome biogenesis is achieved through two pathways (Fig. 1A): the endosomal complex required for transport (ESCRT) and a ceramide-dependent pathway (21–24). Biogenesis of exosomes initiates through the inward budding of vesicles (endocytosis) at the plasma membrane, which leads to the formation of early endosomes. Biogenesis continues with the invagination of the late endosomal limiting membrane to form multivesicular bodies and the sorting of cargos into these bodies (10). In donor cells, exosomes are loaded with regulatory cargos such as various biotypes of RNA, lipids, and proteins by using regulated, yet incompletely understood, processes that lead to preferred sorting of distinct cargos into exosomes (25–27). Sorting motifs (nucleotide sequences) have been reported for microRNAs (25, 28, 29). Exosomes may be degraded in lysosomes or secreted into the extracellular space upon fusion of multivesicular bodies with the plasma membrane. The homing features that determine the localization of exosomes to their recipient cells are poorly understood. Once exosomes reach their destination, they may regulate the metabolism in recipient cells through various, not mutually exclusive, lines of communication including binding to receptors on outer surface of recipient cells and triggering of signaling cascades, fusion with the recipient cell membrane and release of cargo into the cell interior, and cargo delivery by endocytosis of exosomes (Fig. 1B) (10, 30). Evidence is compelling that exosomes transfer mRNA and microRNA from donor cells to recipient cells (31). The usage of exosomes in cell-to-cell communication is not restricted to the animal kingdom but extends to plants and bacteria (32, 33). Prokaryotic EVs are beyond the scope of this review and readers are referred to a recent review regarding EVs in both Gram-positive and Gram-negative bacteria and an original research paper demonstrating that mammalian hosts absorb vesicles secreted by gut bacteria (34, 35). Plant EVs also have biological activities in animals (36).

Figure 1.

Cell-to-cell communication by exosomes. A: biogenesis in donor cells. Note that the orientation of the plasma membrane (outer and inner leaflet) and membrane proteins remains unchanged in biogenesis. B: signaling in recipient cells. Endocytosis may occur by clathrin-dependent endocytosis, phagocytosis, lipid raft-mediated endocytosis, micropinocytosis, or caveolin-mediated endocytosis (20). ESCRT, endosomal sorting complex required for transport; MVB, multivesicular body (late endosome).

EXOSOME AND CARGO CONTENT IN MILK

Exosomes

Exosomes are present in virtually all body fluids including milk in humans and animals (37). The exosome content differs greatly between species. For example, human and bovine milk contain ∼2.2 × 10¹¹ and 1.4 × 10¹⁴ exosomes/mL, respectively (18, 38, 39). It has been proposed that MEs originate in the mammary gland, but the evidence in support of this conclusion is circumstantial. Three papers arrived at this conclusion based on the observation that the majority of cells in human milk were epithelial cells, microRNA profiles were similar in lactating mammary glands and bovine milk, and MAC-T cells secreted exosomes that have a size, morphology, and microRNA content similar to bovine MEs (40–42). MAC-T cells were derived from primary bovine mammary glands in lactating Holstein cows and immortalized using simian virus 40 T antigen (43).

mRNA

Between 3,600 and 11,000 mRNAs have been identified in bovine MEs (1, 44, 45). The difference among these studies is due to experimental conditions including the use of raw milk versus processed milk and the use of microarrays versus RNA-sequencing analysis. The presence of mRNA in exosomes and its translation into protein is controversial, particularly as it relates to mRNA in MEs (45–47). It would be important to know whether natural MEs (as opposed to genetically or chemically altered MEs) harbor truncated or full-length mRNAs that are translated into peptides or proteins, respectively, in recipient cells which might contribute to the development of food allergies or immune tolerance (48–50).

Noncoding RNA

MEs contain a diverse pool of noncoding RNAs (ncRNAs). Across multiple studies, more than 200 and 1,500 microRNAs have been identified in human and bovine MEs, respectively (1, 18, 44, 51–53). The content of microRNAs has also been assessed in MEs from the giant panda and pigs (54, 55). The microRNA content of MEs is driven by a small number of transcripts in both humans and cows. Ten microRNAs accounted for ∼70% of the total microRNA content in human MEs, consistent with microRNA sorting events during ME biogenesis (18). Likewise, 5 to 10 microRNAs accounted for ∼50%–70% of the total microRNA content in bovine MEs (41, 53, 56). The identity of the 10 most abundant microRNAs is different between human and bovine MEs. Little is known about factors that affect the levels of microRNA cargos in exosomes in biological fluids including milk (Table 2). A report on differential expression of microRNAs in distinct populations (Utah residents vs. Yoruba people in Nigeria) might be of limited relevance for milk, because the data were collected in lymphoblastoid cells (64).

Table 2.

Factors that affect the microRNA content in exosomes isolated from biological fluids and study limitations

Factor	Observation	Limitation	References
Bacterial infection	Staphylococcus aureus infection alters microRNA profiles in MEs	Cow as model	(44)
Lactational stage	Immune-related microRNAs are higher in exosomes from colostrum than mature milk	Pig as model	(57)
Processing	Loss of microRNAs during homogenization, microwaving, and high pressure	MicroRNAs as sole marker	(2, 58)
BMI and lactational stage	MicroRNA levels depend on maternal weight and lactational stage	Whole milk, not MEs, was analyzed	(59)
Diet	MicroRNA levels depend on dietary habits	Plasma as model	(60, 61)
Diet	MicroRNA levels depend on feed	Cow as model	(62)
Maturity and diet	Milk maturity and maternal diet impact microRNA levels	Milk, not MEs, was analyzed; milk was frozen	(63)

BMI, body mass index; EV, extracellular vesicle; ME, milk exosome.

To date, most studies of ncRNA in MEs have focused on microRNAs and disregarded other ncRNAs. Presumably, this bias was motivated by reports that microRNAs regulate more than 60% of human genes, loss of microRNA maturation in Dicer knockout mice is embryonic lethal, and the first papers demonstrating the bioavailability of food RNAs (rice and milk) focused on microRNAs (4, 36, 65, 66). The investigator bias in favor of microRNAs warrants a more compelling rationale because all ncRNA have important biological functions (Table 3) and both human and bovine MEs contain ncRNA biotypes other than microRNAs in appreciable amounts (44, 77). In addition to microRNAs, the following biotypes of ncRNA have been identified in bovine MEs in that order of abundance: tRNAs, rRNAs, snRNAs, snoRNAs, repetitive sequences, and nonannotated sequences (44). Fifty-five lncRNAs have been identified in human MEs by using a PCR screen of 87 lncRNAs, i.e., a targeted protocol that might missed transcripts (77). We have detected more than 90,000 unique sequencing reads that mapped to human ncRNAs in human MEs, including all the biotypes in Table 3 (T. Chen, J. Cui, and J. Zempleni, unpublished observation). MicroRNAs accounted for only 1.5% of the sequencing reads, whereas circRNAs accounted for close to 94%, highlighting the importance of conducting a comprehensive RNA-sequencing analysis of ncRNAs in human MEs. These observations shed a new light on the stoichiometric analysis of the microRNA content of exosomes from plasma, seminal fluid, dendritic cells, mast cells, and ovarian cancer cells, which suggested that a typical exosome contains 0.008 molecules of a given microRNA (78). The authors concluded that exosomes “do not carry biologically significant numbers of microRNAs and are unlikely to be functional as vehicles of microRNA-based communication.” CircRNAs are abundant in MEs and might contribute toward gene regulation by microRNAs (75), especially when considering the enormous amount of MEs consumed by a breast-fed infant daily (see Dietary Depletion and Supplementation).

Table 3.

Biological functions of noncoding RNAs

Biotype	Main Biological Function	References
microRNA	Gene silencing	(67)
piRNA	Silencing of retrotransposons	(68)
tRNA	Translation of Mrna	(69)
snoRNA	Guide chemical modifications of rRNAs, tRNAs, and snRNAs	(70–72)
snRNA	Processing of pre-mRNA	(73,74)
circRNA	microRNA sponges	(75)
lncRNA	Miscellaneous, including modulation of chromatin function	(76)

circRNA, circular RNA; lncRNA, long noncoding RNA; piRNA, PIWI-interacting RNA; snoRNA, small nucleolar RNA; snRNA, small nuclear RNA; tRNA, transfer RNA.

DNA

Exosomes that originated in cancer cells may also contain double-stranded DNA (79). It has been proposed that exosomal DNA may be used as biomarker in cancer detection. Healthy cells may use exosomes to secrete harmful cytoplasmic DNA fragments, thereby maintaining cellular homeostasis (80).

Proteins and Lipids

Up to 2,000 proteins have been identified in bovine MEs by using mass spectrometry (53, 81). The detection of milk fat globule proteins in one of the reports suggests that ME preparations might have been contaminated with fat globules leading to an artificially large number of proteins (81). This concern does not apply to a study of the proteome in human EVs which authenticated milk EVs by immunoblot analysis of five marker proteins and comparison of proteomics data to previously published studies. The authors identified close to 2,000 proteins and chose the term EV, not ME, for their preparations (82). Approximately 2,400 lipid signatures and close to 400 lipids were identified in bovine and human MEs, respectively (53, 83). Readers are referred to ExoCarta, which is a searchable database of exosome cargos including proteins and lipids (84). The most recent update to ExoCarta appears to date back to 2016. The hosts of the database encourage members of the EV community to submit updates.

Effects of Industrial Processing

Ultraheat treatment (≥135°C for ≥1 s) decreased the content of exosomes (denoted EVs in that study) in bovine milk to below detectable quantities, whereas pasteurization (72–78°C for ≥1 s) had no effect on exosome count but altered exosome appearance, protein signatures, and RNA content (85). Another study used select microRNAs as proxy for ME and took bovine raw milk through fat adjustment, homogenization, pasteurization, cold storage (4°C), and heating in the microwave oven; the study also assessed the microRNA concentration in dairy products such as yoghurt and cheese (2). The content of microRNAs in milk decreased during pasteurization and heating in the microwave oven. The concentrations of microRNAs in yoghurt and cheese were lower than in milk; encapsulation of microRNAs in MEs was not assessed.

MILK EXOSOMES IN HUMAN NUTRITION

Intestinal Transport, Bioavailability, and Distribution

MEs have attracted considerable attention in the nutrition of both infants and adults. As aforementioned, a significant fraction of MEs and their cargos resists degradation in the gastrointestinal tract and the processing of milk in dairy plants (1–3). The uptake of bovine MEs by intestinal cell lines follows first-order kinetics and has been modeled by using a lipophilic exosome marker, FM 4–64, applying the Michaelis-Menten equation(5). The Michaelis constant was 55.5 ± 48.6 µg exosomal protein/200 µL media in human colon carcinoma Caco-2 cells, and 152 ± 39.5 µg exosomal protein/200 µL media in rat small intestinal IEC-6 cells; maximal rates were 0.08 ± 0.06 ng exosomal protein × 81,750 cells⁻¹ × h⁻¹ in Caco-2 cells, and 0.14 ± 0.01 ng exosomal protein × 36,375 cells⁻¹ × h⁻¹ in IEC-6 cells. Inhibitors of endocytosis and vesicle trafficking, and carbohydrate competitors caused a significant decrease in the uptake of bovine MEs by Caco-2 cells, consistent with uptake by glycan-dependent endocytosis. MicroRNAs, delivered through bovine MEs at the apical side, were transported across Caco-2 cell monolayers and accumulated in the extracellular space at the basal side in a dual chamber system; reverse transport was negligible. Human umbilical vein endothelial cells also endocytosed bovine MEs, consistent with the possibility of ME transport from blood vessels to tissues (6).

Studies in humans, pigs, and mice provided experimental evidence that MEs and their cargos are bioavailable. When adults consumed cow’s milk in nutritionally relevant amounts in randomized dose-response time course studies, a dose-dependent increase in plasma levels of microRNAs was evident; plasma levels peaked 3–4 h after milk consumption and returned to baseline levels $\sim$9 h after milk consumption (4). MEs deliver cargos that are biologically active because the addition of bovine MEs to cultures of human embryonic kidney HEK-293 cells decreased the expression of microRNA reporter genes (4). It remains to be demonstrated whether microRNAs or other cargos or MEs per se repressed the reporter genes.

Studies in transgenic pigs and mice provided experimental evidence in support of the observations made in humans. When wild-type piglets were nursed by sows secreting MEs containing a fluorescent protein, ZsGreen1, full-length ZsGreen1 accumulated in the cerebellum and, to a lesser extent, in the spleen (7). The same study used Exosome and Cargo Tracking (ECT) mice, which synthesize exosomes, including MEs, that express enhanced green fluorescent protein (EGFP) fused with the exosome marker CD63 (in the absence of cre recombinase) or near-infrared protein fused with CD63 (in the presence of cre). When wild-type pups were nursed by ECT dams for 17 days, EGFP-labeled MEs accumulated primarily in the brain, intestinal mucosa, liver, and spleen in the suckling pups (7, 8). Z-stack confocal microscopy and serial two-photon tomography were used to demonstrate that murine MEs left the vasculature and accumulated in many brain regions including the cortex, cerebellum, and hippocampus in suckling mice (8). One study estimated the apparent bioavailability of bovine MEs to be ∼25%, but that was a crude estimate, based on using one microRNA as a marker and relying on the hepatic accumulation of that microRNA at one time point after oral and intravenous administration in mice (7). Independent laboratories have confirmed that MEs and their RNA cargos are bioavailable (4, 5, 7, 86–98).

The bioavailability of milk microRNAs has been contested by Laubier et al., Auerbach et al. and Title et al., and Kang et al. (99–102). Despite the limitations of these studies, they contribute to the knowledge base on MEs and their RNA cargo. Laubier et al. (99) fostered wild-type pups to transgenic mice that overexpressed miR-30b and failed to see a substantial increase in tissue levels of miR-30b in pups. The failure to observe an increase in miR-30b in pup tissues was probably because the miR-30b in overexpression dams was not encapsulated in MEs, evidenced by the decrease in miR-30b levels in the stomach. This observation is consistent with reports that encapsulation in MEs protects microRNAs against degradation by low pH and RNases in the gastrointestinal tract (1, 3). Title et al. (100) detected only trace amounts of miR-375 in the plasma of miR-375 knockout mouse pups fostered to wild-type dams. Subsequent studies suggest that miR-375 in milk is subject to first pass elimination in intestinal mucosa and liver and therefore its concentrations in circulation and peripheral tissues are low (7, 103). The paper by Title et al. is consistent with a report that plasma levels of some, but not all microRNAs, increase after milk consumption in humans (19) Auerbach et al. (101) reported a failure to detect bovine miR-29b and miR-200c in human plasma following a milk meal. The authors acknowledged that the dry ice used for shipping had sublimated by the time the samples arrived at the study site. Subsequent studies suggest that the integrity of the samples used in that study was compromised and the RNA was degraded (93). The paper by Auerbach et al. may serve as an important reminder that great care needs to be taken when collecting and storing biological samples for subsequent RNA analysis. Kang et al. (102) conducted a meta-analysis of RNA-sequencing data and concluded that the abundance of dietary microRNAs in body fluids is very low and possibly due to assay artifacts. The paper allowed for a greater number of mismatches when mapping sequencing reads to human microRNAs compared with food microRNAs, thereby creating a bias toward calling plasma microRNAs human. The successful use of mRNA vaccines to combat the COVID-19 pandemic suggests that RNAs, encapsulated in lipid nanoparticles, are bioactive (104). That said there are differences in RNA delivery by MEs and synthetic nanoparticles such as route of administration, vehicle composition, and dose of RNA. The controversies surrounding the first paper on cross-kingdom transfer of microRNAs in rice are of limited interest in the context of this review on MEs and their cargos and readers are referred to a previous publication (105).

Dietary Depletion and Supplementation

Rodent diets, defined by their content of MEs and RNA cargos, have been developed and used to determine whether endogenous microRNA synthesis can compensate for dietary depletion and assess phenotypes of depletion in mice. Briefly, (milk) exosome and RNA-depleted (ERD) and (milk) exosome and RNA-sufficient (ERS) diets are prepared by substituting ultrasonicated and nonultrasonicated bovine milk, respectively, for casein in the AIN-93G formulation before pelleting (53). The diets provide a nutritionally relevant amount of milk and differ only by their content of milk exosomes and their RNA cargos. The content of bioavailable MEs and microRNAs is ∼85% and 99% lower, respectively, in the ERD diet compared with the ERS diet. The content of RNA other than microRNAs was not assessed. The diets mimic the situation encountered in formula-fed and breast-fed infants; formulas contain only trace amounts of MEs and microRNAs, whereas human milk contains ∼1,011/mL MEs and more than 200 microRNAs (18). When ERD feeding was initiated in mice at weaning and continued for ∼4 wk, microRNA levels in plasma, liver, skeletal muscle, intestinal mucosa, and placenta decreased by up to 60% compared with controls fed the ERS diet, i.e., endogenous synthesis was insufficient to compensate for dietary depletion (4, 106–109). Consumption of the ERD diet elicited phenotypes of depletion (Table 4). It will be important to establish cause-and-effect and generate mechanistic insights by fostering wild-type pups to dams in which ME or RNA biogenesis is impaired in future studies. Likewise, it seems worthwhile to assess whether MEs and their cargos have a direct effect on gene expression and metabolism in humans and animals or whether effects are indirect through ME-dependent changes in the host microbiome (110).

Table 4.

Phenotypes observed in juvenile mice fed an (milk) exosome- and RNA-depleted (ERD) diet compared with controls fed an (milk) exosome- and RNA-sufficient (ERS) diet

Phenotype of ERD feeding	References
Aberrant metabolism of purines1	(106)
Moderate decrease in muscle grip strength2	(107)
Increased severity of symptoms in a mouse model of inflammatory bowel disease	(108)
Small litter size and decreased survival to weaning	(109)
Decreased spatial learning and memory and increased severity of kainic acid-induced seizures	(8)
Changes in bacterial communities in the cecum	(110)
Changes in mRNA expression	(8, 106–109)

¹The study confirmed the same phenotypes in formula-fed infants compared with breast-fed infants.

²Similar observations were made in rats fed the ERD diet (111).

Information is scarce regarding effects of ME supplementation. ME supplementation had positive effects such as increased villus height and crypt depth in the murine intestinal mucosa, decreased severity of symptoms in murine models of necrotizing enterocolitis, and improved bone health in a murine model of osteoporosis (88, 96, 112). When soy formulas were supplemented with bovine MEs, the absorption of microRNA cargos was negligible in adult subjects, probably because of ME agglutination by soy lectins (113). Theoretically, microRNA cargos in MEs might have adverse health effects such as aberrant DNA methylation events but the reports are limited to theoretical considerations without providing experimental evidence (114, 115).

Even though MEs are small and can hold only a limited amount of cargos, the mass transferred from mother to infant is substantial, e.g., breast-fed infants consume ∼176 trillion MEs in 800 mL milk/day (18, 116). This drain on maternal resources during lactation may warrant investigation just as much as the potential harm in formula-fed infants who miss out on this transfer.

Conclusions – Human Nutrition

Taken together, MEs and their RNA cargos meet the definition of bioactive food compounds by the National Cancer Institute which is “A type of chemical found in small amounts in plants and certain foods […]. Bioactive compounds have actions in the body that may promote good health” (117). It remains to be determined if this holds true for ncRNAs other than microRNAs, and whether MEs and their ncRNA cargos are conditionally essential nutrients that “can usually be synthesized in adequate amounts endogenously, but may require exogenous supplementation during some circumstances” (118).

MILK EXOSOMES IN DRUG DELIVERY

Features Conducive to Successful Application

Lipid nanoparticles and cell culture-derived exosomes are promising tools for delivering therapeutics, including the delivery of small interfering RNA to sites of disease such as tumors (119, 120). Nanotechnology offers a variety of advantages compared with the conventional method of delivering free drugs. For example, nanoparticles may facilitate the transfer of therapeutics across the blood-brain barrier and protect unstable RNA cargos against degradation (121, 122). MEs offer the same and several additional advantages compared with lipid nanoparticles and culture-derived exosomes, including stability, protection of cargo, oral bioavailability, scalability, low immunotoxicity, and transfer across the blood-brain barrier (1–4, 7, 8, 19, 89). Low immunogenic potential is an important consideration when pursuing federal approval of a drug (123). The production of BMEs is scalable. On average, a cow produced 10,800 kg of milk annually in the US in 2020, and bovine milk contains 10¹²–10¹⁴ exosomes per milliliter (1, 53, 124). Some studies of exosome bioavailability and distribution need to be interpreted with caution because the exosomes were labeled with lipophilic dyes which are known to transfer from the exosome membrane to other lipophilic structures (89, 125–127). Alternative labeling technologies are available such as loading exosomes with fluorophore-labeled RNA or tracer RNA and expression of fluorescent fusion proteins (7, 42, 128, 129). The pharmaceutical industry has recognized potential of using MEs for delivering RNA therapeutics to brain tumors and committed $1 billion to licensing the technology (130).

Previous Uses of MEs in Drug Delivery

The feasibility of delivering drugs by MEs has been demonstrated. For example, when bovine MEs were loaded with paclitaxel and administered orally to a nude mouse model of human lung cancer, tumor growth decreased by 60% compared with controls (90). Studies with MEs, loaded with locked antisense nucleotides and delivered orally to mice had no significant effect on target gene knockdown in mice, probably because the accumulation of oligonucleotides in target tissues was low (131).

MEs have in common with lipid nanoparticles and culture-derived exosomes a few limitations. For example, the particles are rapidly cleared by tissue macrophages, thereby decreasing their half-life and the half-life of their therapeutic cargo (7, 125, 126). The uptake of MEs by murine bone marrow-derived macrophages is facilitated by Class A scavenger receptor-1/2 and perhaps other transporters (132). The ME uptake by bone marrow-derived macrophages was not saturated at the highest ME concentrations tested. Approaches to develop smart MEs, e.g., MEs that evade elimination by macrophages, could include expressing a don’t eat me signal like CD47 on the ME surface (133). The advent of MAC-T cells affords an opportunity to explore genetic editing of MEs in cell cultures (42). That would be for research purposes because MAC-T cells do not secrete exosomes in amounts sufficient for treating patients. Cell culture-derived exosomes have been successfully used to deliver proteins like Cre recombinase to the brain in mice but the feasibility of protein delivery by MEs remains to be tested (134).

Conclusions – Drug Delivery

Taken together, MEs afford a vehicle for delivering therapeutic agents to diseased tissues but the full potential of the technology has yet to be realized. The optimization of cargo loading protocols and macrophage evasion technologies might make important next steps. There are additional hurdles that need to be overcome before MEs may be used in clinical applications, e.g., loading the desired cargo in proper amounts and achieving desirable pharmacokinetic properties.

Conclusions

MEs make for a promising novel line of investigation in nutrition and pharmaceutical sciences. The groundwork has been laid through the development of ME labeling technologies, transgenic animals and feeding protocols, identification of depletion and supplementation phenotypes, and the development of cell lines amenable to genetic editing. Major unknows include the content and biological activity of ncRNA cargos other than microRNAs, factors that alter ME cargos and technologies that improve the loading of MEs with therapeutic cargos and their homing to target sites.

GRANTS

This work was supported by the National Institutes of Health Grants P20GM104320 and OD028749; the National Institute of Food and Agriculture Grants 2016-67001-25301, 2020-67017-30834, and 2022-67021-36407; the U.S. Department of Agriculture Grant Hatch and W-4002; and the SynGAP Research Fund (to J. Zempleni).

DISCLOSURES

J. Zempleni serves as consultant for PureTech Health, Inc. in Boston, MA, and declares no conflict of interest. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

J.Z. prepared figures; J.Z. drafted manuscript; A.N., S.W., H.W., A.K., and J.Z. edited and revised manuscript; A.N., S.W., H.W., A.K., and J.Z. approved final version of manuscript. All authors have read and approved the final version of the manuscript.