Extracellular vesicles in dry eye disease and Sjögren’s syndrome: A systematic review on their diagnostic and therapeutic role

Abstract

Extracellular vesicles (EVs), defined as membrane-bound vesicles released from all cells, are being explored for their diagnostic and therapeutic role in dry eye disease (DED). We systematically shortlisted 32 articles on the role of EVs in diagnosing and treating DED. We cover the progress in the last 2 decades on the classification and isolation of EVs and their role in DED. The diagnostic predictability of exosomes was evaluated in Sjögren syndrome (SS) patients’ tears, plasma, and saliva, where upregulation of inflammatory proteins was reported uniformly across studies. Also, we evaluate the therapeutic effects of MSC-derived EVs in in vitro and in vivo studies of SS and DED mouse models. A significant response occurs at a functional level with improved tear production and saliva flow rate and at a cellular level with reduced lymphocyte infiltration, improved corneal structural integrity, decreased epithelial cell apoptosis, and dampening of the inflammatory cytokine response. The proposed mechanisms of EV action include PD-L1, PRDM, NLRP-3, and Nf-kb pathways, and an increase in M2 macrophage phenotype. Current use of exosomes in DED is limited due to their cumbersome isolation techniqus. Further research on human subjects is needed, in addition to optimizing exosome isolation and delivery methods.

1. Introduction

Exosomes are nanosized, bilayered extracellular vesicles (EVs) released from cells. They contain biological material like proteins, nucleic acids, and lipids from their parent cell. Exosomes assist in intercellular communication and transmit macromolecules between the cells.³⁴ They are commonly found in various body fluids, including serum, urine, cerebrospinal fluid, saliva, and bronchiolar lavage fluid.^29,56 Recently, exosomes’ potential therapeutic role was explored in various diseases, including autoimmune disease, malignancy, glaucoma, and dry eye disease (DED)^1,15 Exosomes in the eye have been found in aqueous humor, tears, and vitreous humor. Tear proteins have been studied for more than a decade for their diagnostic role, especially in picking up early DED cases. As DED has many subtypes and the severity of symptomatology varies despite similar signs, many research groups have focused on tear proteomics as an approach for differentiating the subtypes and as reliable markers for severity grouping. Tears are an easily accessible ocular fluid; hence, an attempt to isolate and characterize them has been explored in recent years.

Exosome-derived proteins and micro RNA (miRNA) are being explored for their potential role as novel diagnostic biomarkers of Sjögren syndrome (SS), a systemic autoimmune condition with lymphocytic proliferation of exocrine glands, including salivary and lacrimal glands. SS has an unclear etiology, with a combination of genetic, immune, and environmental factors SS predominantly affects middle-aged women, with a prevalence between 0.5 % and 5 %.⁶² Immune-mediated damage of the exocrine glands results in glandular atrophy and reduced tear volume. The clinical features and course can be highly varied, but involves dry surfaces of mucous membranes, including mouth, eyes, nose, throat, and vagina. DED is a chronic progressive disease involving an unstable tear film and disturbance in the homeostatic balance of the ocular surface microenvironment. The pathogenesis of DED is multifactorial, but may involve inflammation and immune response at the level of the Meibomian glands, lacrimal gland, and conjunctival goblet cells that produce the different components of the tear film.²⁴ There may also be neurosensory changes from trauma, surgery or systemic diseases such as diabetes that interfere with normal reflex tear production. DED presents with ocular surface symptoms, including dryness, irritation, stinging, burning, epiphora, pain, foreign body sensation, red eyes, and blurry vision. It is estimated that 10 % of patients with DED have underlying SS, but two-thirds remain undiagnosed. ⁵⁰

The clinical manifestations of SS-related dry eye and non-SS dry eye are similar and difficult to differentiate, and many patients with dry eye do not have a work-up for SS. There are limitations to the current ACR/EULAR 2016 diagnostic criteria⁷ for SS as it has low importance given to ocular signs and remains cumbersome, with criteria requiring tests such as minor salivary gland (MSG) biopsy, ocular surface staining, Schirmer test, saliva flow rate, and serum autoantibodies. The saliva flow rate, Schirmer’s test, and ocular surface staining are nonspecific to SS; hence, many cases of dry eye secondary to SS remain undiagnosed. Also, SS has a highly variable clinical course, and current diagnostic tools fail to identify patients likely to have a more severe clinical course.⁵ This raises the pertinent clinical need for a more efficient, practical, non-invasive, and specific diagnostic test for SS with prognostic value. Exosomes may also be involved in the pathogenesis of SS. Salivary gland epithelial cells are found to release exosomes containing Ro/SSA and La/SSB autoantigens, which may then be presented to the immune system to stimulate autoimmunity.¹⁴ Hence, isolating exosomes from tears and exploring their role in diagnosing or treating DED seems promising. This systematic review summarizes the current state of exosome research in DED, including exosome isolation and identification and their diagnostic and therapeutic role in DED in animals and humans.

1.1. Exosome − an extracellular vesicle

1.1.1. Definition and types

Extracellular vesicles (EVs) are small-sized, lipid-bound vesicles released by the cells into the extracellular environment that was initially, thought of as cellular waste, wherein the cells selectively eliminate proteins, lipids, and RNA. In recent years, EVs have gained importance as they have been found to facilitate cell-to-cell communication. The EVs are categorized into 3 subtypes: microvesicles, exosomes, and apoptotic bodies, based on the biogenesis, size, content, and function.^55,60 Microvesicles measure 0.1−1.0 μm in size and are formed by the outward budding of the plasma membrane. Apoptotic bodies are the largest EVs, measuring 500nm−2μm in size, and are released by cells undergoing apoptosis. On the other hand, exosomes are small-sized extracellular vesicles (EVs) with 40−150nm diameter. They are formed by the plasma membrane’s inward budding, leading to early endosomes forming. They then become multivesicular bodies (MVBs) containing exosomes, released into the intraluminal space upon fusion with the plasma membrane.^12,20,21

1.1.2. Isolation of exosomes

Current isolation techniques include differential ultracentrifugation, density gradient separation, ultrafiltration, size exclusion chromatography, precipitation, immunocapture, and microfluidics.⁴³ Among all the techniques mentioned above, differential ultracentrifugation is considered the gold standard (Tables 1 and 2).¹⁸ The following methods are widely employed, each offering unique advantages and challenges.

Table 1. Summary of different Exosome Isolation Techniques.

Technique	Time (min)	Purity	Yield	Advantages	Limitations
Ultracentrifugation (UC)	140−600	Moderate	High	Established, reliable for research	Time-consuming, large sample volume, risk of damage to EVs
Size-Based Isolation (SEC)	15−130	High	Moderate	High purity, preserves exosome integrity	Expensive, time-consuming
Immunoaffinity Capture	240	Very High	Low	Excellent for tissue-specific isolation	Expensive, low throughput
Polyethylene-based Precipitation (PEG)	30−120	Low−Moderate	High	Simple, scalable, compatible with commercial kits	Contamination risks, lower purity
Microfluidic Devices	30−120	High	High	Portable, rapid, minimal sample requirement	Technical expertise needed, limited scalability and expensive.

Table 2. Details of Extracellular vesicle isolation and characterization techniques for their diagnostic role.

Study Name	Sample	Source of EVs	EV isolation technique	EV characterisation techniques used	EV concentration	EV size (mean, in nanometers)	Markers for EV validation
Ferrant J et al.1	SS and healthy controls	Plasma	Size-exclusion chromatography	NTA, flow cytometry	5×1010 (approx from graph)	110−150	NA
Aqrawi et al.2	SS and healthy controls	Tears and saliva	Size-exclusion chromatography	NTA, flow cytometry	Tears SS 1.09×109 ± 1.06×108 Control, 1.54×109 ± 3.08×108 Saliva SS 5.46×1010 ± 1.43×1010Controls 2.41×109 ± 3.98×109	Tears SS 171 ± 6.9Control, 163 ± 9.6Saliva SS 189 ± 4.1Controls 189 ± 4.4	CD9
Aqrawi et al.3	SS, non-SS and healthy controls	Tears and saliva	Size-exclusion chromatography	NTA, immunoaffinity capture, flow cytometry	Tears (SS, 5.0×108, Non-SS, 1.2×109Controls, 7.1×108) Saliva (SS, 1.9×1010± 0.7×109, Non-SS, 1.0×1010± 1.6×109 Controls, 7.9×109± 1.6×109	Tears (SS, 255 ± 40,Non-SS, 204 ± 8Controls, 215 ± 9) Saliva (SS, 233 ± 17,Non-SS, 231 ± 13Controls, 264 ± 6	CD9, CD−44
Kakan SS et al.4	NOD mice model	Plasma	Differential ultracentrifugation for discovery experiments and with SEC	RNA isolation, TEM, Western blotting, NTA	N/A	Median 136 nm (male NOD serum with UC), 131 nm (BALB/c mouse serum with UC).SEC-median diameter of 122 nm in both strains.	TSG101
Li F et al.5	SS and non-SS	MSGs and plasma	Total Exosome Isolation Kit (Invitrogen)	Not done	NA	NA	NA
Yamashiro K et al.6	SS and healthy controls	Mouth rinse	Ultracentrifugation	NTA, western blot	NA	145	CD9, CD63
Cross T et al.8	SS and healthy controls	Saliva	Size-exclusion chromatography and modified Qiagen exoRNeasy EV isolation	NTA, TEM, Immunoaffinity capture and Western blot	2.0 x109 particles/mL	154	CD9, CD63, Hsp70
Pucker et al.9	DED and non-DED	Tears	Polyethylene glycol (PEG) 8000	TEM, ELISA	550,000 in DED 600,000 in non DED	NA	CD63
Cross T et al.10	DED	Tears	Modified Qiagen exoRNeasy column approach	NTA, TEM, Flow cytometry, western blot	~108 particles/mL	100−350	Tetraspanin CD9 (flow cytometry), CD9, CD−63, Hsp70

SS − Sjögren’s Syndrome, DED − Dry Eye Disease, NTA − Nanoparticle Tracking Analysis, NOD − Non-Obese Diabetic, TEM − Transmission Electron Microscopy, SEC − Size Exclusion Chromatography, UC - Ultracentrifugation, MSG − Minor Salivary Gland

1.1.3. Ultracentrifugation (UC)

Ultracentrifugation (UC) is the traditional method for exosome isolation firstly reported by Raposo and coworkers.⁴⁰ It separates extracellular vesicles based on size and density through high-speed spins varying from 70,000 to 200,000 ×g. While it effectively isolates small extracellular vesicles (EVs), UC is time-intensive, requires large volumes, and risks mechanical damage to exosomes.⁶¹ These limitations sought for advancements like the addition of density gradients (e.g., sucrose) to enhance purity, but that increased the complexity and costs. Paolini and coworkers³⁷ isolated exosomes from patients’ serum, including methods incorporating density-gradient centrifugation with iodixanol and sucrose. The exosomes yield was of significantly higher purity compared to differential centrifugation followed by ultracentrifugation or single-step precipitation kits.

1.1.4. Size-based isolation

These techniques include ultrafiltration (UF), sequential filtration, and size-exclusion chromatography (SEC).¹⁰ They isolate exosomes using membranes or columns with specific pore sizes (30−200 nm). UF is cost-effective, but risks mechanical damage and clogging.⁶¹ SEC offers high purity while preserving exosome integrity, though it is slower and more expensive. In biological samples, size-based isolation techniques encounter significant challenges with smaller volumes, as limited sample quantities often lead to pore clogging, thereby reducing the yield of exosomes.

1.1.5. Immunoaffinity capture

Immunoaffinity capture isolates exosomes by targeting unique surface proteins such as CD81, CD63, and CD9. These methods often use biotinylated antibodies to bind these proteins, and avidin-biotin interactions to isolate the exosomes, enabling high specificity. Several commercial kits facilitate this method, including the EpCAM isolation reagent (Thermo Fisher Scientific), CD63 isolation reagent (Thermo Fisher Scientific), and CD81/CD63 isolation kit (Miltenyi Biotec).⁵⁷ Immuno-magnetic bead-based techniques, such as EXOBead, have demonstrated higher exosome purity compared to size-exclusion methods, as shown by Benecke and coworkers.⁴ However, their efficiency can be limited by factors like elution buffer effects on exosome integrity and the availability of binding molecules. Yoshida and co-workers⁵⁸ developed the TIM4-affinity isolation technique, which targets phosphatidylserine (PS) on exosome membranes, using calcium-dependent TIM4-PS binding and enables the elution of intact exosomes with ethylenediaminetetraacetic acid (EDTA), improving purity while overcoming the drawbacks of traditional immunoaffinity approaches.

1.1.6. Precipitation techniques

Exosome precipitation is a simple and efficient method for isolating small extracellular vesicles (sEVs) using polymers like polyethylene glycol (PEG) or natural alternatives such as chitosan. PEG is widely used owing to its low cost, while natural polymers are more biocompatible and suitable for clinical applications. These methods are simple and scalable but often yield lower purity due to contamination risks. Commercial kits like ExoQuick from System Biosciences, the Total Exosome Isolation Kit from Thermo Fisher Scientific, and ExtraPEG simplify the process for clinical and research use.^28,41

1.1.7. Microfluidic devices

Lab-on-chip systems use physical, electrical, or immunoaffinity properties to isolate exosomes. These devices are compact, fast, and require minimal sample volume, achieving high yields and specificity. Microfluidic devices, such as the ExoChip and ion exchange membranes²⁶ offer solutions for isolating exosomes with high yield and minimal damage; however, they require technical expertise and may have lower throughput.

1.1.8. Contents of the exosomes and mechanism of action

Exosomes are known to contain membrane, cytosolic and nuclear proteins, extracellular matrix proteins, metabolites, and nucleic acids, namely mRNA, noncoding RNA, and DNA, as shown in Figure 1.^38,48 They are more stable, have greater safety, higher permeability, and less tumorigenicity compared to mesenchymal stromal cells. Due to these properties, they can pass through biological barriers, including the blood-brain barrier, making them ideal drug carriers.¹³ Additionally, the exosome’s miRNA differentiates hematopoietic stem cells into the myeloid and lymphoid lineage and in the pathogenesis of various autoimmune diseases, including SS.³⁵ The function of these exosomes varies mainly depending on their cell source. It has been proposed that exosomes can be designed to carry therapeutic molecules such as proteins, nucleic acids, or low molecular weight drugs to act on target cells,⁵³ allowing for targeted cell therapy.

Fig. 1. Schematic representation of exosome formation and its contents.

2. Methods

A review using an online database search (PubMed) was conducted for the studies on exosomes and dry eyes using the keywords: “Exosomes” OR “Extracellular vesicles” AND “Dry Eye Disease” OR “Lacrimal Gland” OR “Salivary Gland” OR “Dry Eye” OR “Sjögren’s syndrome.” No restrictions were used during the search; all years were included. Articles other than English were included if an English translation was available online. Reference lists from retrieved articles and existing reviews were manually searched for additional studies.

2.1. Screening of studies

Two authors (P.C & A.F) independently reviewed the list of identified articles to assess eligibility for inclusion. Any disagreements were resolved by consensus with the senior author (S.S). The following inclusion criteria were used for this review:

• 1)Published as an original article.
• 2)Articles describing exosomes and their content in a diagnostic or therapeutic role applied to dry eye disease and/or Sjögren syndrome.
• 3)Both animal and human studies concerning exosomes
• 4)Both in vivo and in vitro studies

All the relevant articles were screened at the following stages: title, abstract, and full text. Non-peer-reviewed articles were excluded. Articles that explored the role of exosomes in the pathogenesis of DED or SS were excluded, as were all articles on exosome use outside of DED and SS.

2.2. Data extraction & analysis

After screening, the data were extracted by 2 authors (PC and AF) together. The data extraction from the articles included the author, study title, type of study, experimental target, exosome source, isolation, characterization techniques, concentration, size and exosomal markers, outcome measures, and the main findings. In addition, we have included the exosome delivery method for therapeutic studies. All these variables were entered into an MS Excel spreadsheet (Version 16.54), and similar articles were compiled in the review.

3. Results

3.1. Data screening and identification of articles

Search using the keywords yielded 184 articles, of which 127 were excluded by title and abstract screening based on the inclusion criteria (Figure 2). The remaining 57 articles were screened for full text based on the inclusion criteria for the original research articles, which included 32 articles. The other 25 articles were review articles and were excluded. The exosomes derived from various sources have been studied in DED animal models and DED patients as biomarkers for early diagnosis or their therapeutic effects (Tables 3 to 5). We have divided the findings into exosome isolation, diagnostic, and therapeutic use.

Fig. 2. PRISMA flow chart showing the screening of articles for inclusion in the review.

Table 3. Articles on the outcomes of exosome treatment on the tear film of DED animal models.

Study	Experimental target	Source of exosomes	Exosome volume	Exosome size	Exosome delivery method	Exosome content/ markers	Outcomes measured	Results
Yu et al. 2020	Mice subject to 5 days of desiccating stress exposure	Human adipose tissue MSCs	NA	100 nm	Topically 4 times to murine cornea, after being labelled with PKH67	TSG101, CD63, ALIX	Fluorescein staining, tear production and PAS staining	CFS decreased to 5.2 (+/- 0.84) from 12.0 (+/−2.44). Tear production improved from 2.40 + /− 0.49, to 6.10 + /−0.75. increased PAS-stained goblet cells
Wang et al. 2022	BAC model of Male C57BL/6 mice (sham, PBS, positive control, 12.5, 25 and 50 mg/mL mADSC- Exos)	Mouse adipose derived mesenchymal stem cell exosomes (mADSC-Exos)	12.5, 25 and 50 mg/mL	40-600 nm, average 134 nm	5 μL topically 3 times a day for 7 days	CD9, CD63, CD81	Corneal fluoresceine staining, tear secretion, tear BUT TUNEL, Flow cytometry, qRT-PCR, Western blotting for corneal cytokines	Increased tear volume (2.2 mL), increased TBUT (2.3 s) in the 50 mg/mL exos group, reduced corneal cell apoptosis, decreased levels of IL−1β, IL−6, IL−1α, caspase−1, IL-18, IFN-γ and TNF-α, increased levels of IL−10, and downregulation of NLRP3 inflammasome
Li et al. 2022	Adult female white rabbits with autoimmune dacryoadenitis-PBS group (n = 9) and exosome group (n = 9).	Human umbilical cord MSCs	NA	50−150 nm	Subconjunctival injection of 30μg on days 1,3,5,7 and 9 after adoptive transfer of lymphocytes	CD81, CD9, CD63, TSG101	TBUT, corneal fluoresceine staining, tear volume, histopathological assessment of lacrimal gland and conjunctiva was performed at 8 weeks	Improved CFS (9 vs 12), tear secretion (6 vs 4 mm), TBUT (7 vs 4 s); reduced lymphocytes in the lacrimal glands; reduced M1 macrophage markers (NOS2, IRF5, TNF-a, IL−1b and IL−6) and increased M2 macrophage markers (Arg1, CD206, KLF4, IL-10, TGF-b).
Guo et al. 2022	Female C57BL/6 mice, 3 groups with 4 mice per group; control, PBS and exosome groups	Human umbilical cord MSCs	1μg/1μl	80−180 nm	1μl given topically QID for 11 days	CD9, CD63 and CD81 (Calnexin was negatively expressed)	Tear volume, corneal fluorescein staining	Improved tear production compared to PBS solution (2.7 vs 1.8 p = 0.0476). Reduced CFS; Reduced mRNA expression of TNF-a, IL1b and IL−6 in the MSC-EV treated group.
Tian et al. 2023	C57BL mice, BAC model, the left eye served as control. 3 treatment groups 10 μL of PBS, MSC-Exo or MSC-Exo-Ce	Murine bone marrow, Ce (NO3)3-6H2O was incorporated into the MSC-Exos to form The MSCExo-Ce	20−40μ/mL	100−150 nm	Topical, 10μL of PBS, MSC-Exo, or MSCExo-Ce twice daily	CD 48, CD 63, TSG101	Tear production, fluorescein staining, ELISA of tear samples, histological assessment of cornea with H&E staining, in vivo ROS scavenging assay, distribution of MSC-Exo and MSC-Exo-Ce in vivo.	Reduced CFS from 14 in to 1 (MSCExo-Ce), and 6 (MSC-Exo group). Improved epithelial integrity, central corneal thickness in the MSC-Exo-Ce group compared to MSC-Exo or PBS group.
Wang et al. 2023	Female C57BL/6 mice aged 8−10 weeks subject to a desiccating environment and scopolamine administration. 8 groups (4 eyes per group); FML, PBS, control and 5 groups given varying concentrations of MSC-Exo of	Human umbilical cord MSCs	0.5, 1.0, 2.0, 3.0 and 5.0 mg/mL	Average 107.5 nm	Topically 5pL given 4 times/day for 21 days	CD9, CD63, CD81, Alix and calreticulin was negative	Tear secretion, corneal fluorescein staining, cytokine profiles, TUNEL assay	CFS decreased in all concentrations of exos (p < 0.05); improved tear volume in all groups; 3.0 mg/mL group had the lowest corneal fluorescein score (2.0 vs 9.3 p < 0.0001) and highest tear volume (3.5 vs 1.1 p < 0.001); decreased inflammatory cytokines in tears of exosomes group.
Ma et al. 2023	Divided into 5 groups of 6 mice each; control group with no BAC, saline group, mExo group, AA group, mExo@AA group. All groups treated with BAC, with 5 μL of 0.2% BAC twice/day for 7 days	Mouse mesenchymal stem cell derived exosomes with and without ascorbic acid coupling	1 mg/mL	100−150 nm	Topically twice/day for 7 days	ALIX, beta-actin, TSG 101 positive, negative for GM130	Corneal fluorescein staining quantitative measurement, fluorescence DHE probe for ROS, CD206 fluorescence quantitative measurement, ELISA measurement of IL−6 and IL−1β in the tears, histological study of the cornea	Less CFS with exo, saline and mExo@AA showed the least ROS with the DHE probe, Tears in the mExo@AA group had the lowest IL−6 and IL-1β, higher CD206, reduced corneal damage on histology. Tear production in mExo was 3.5 mm, mEXo-AA 5.5 mm, saline group 2 mm, control group 6.5 mm.

MSC − Mesenchymal Stromal Cell, PAS - Periodic acid-Schiff, CFS- Corneal Fluorescein Score, BAC − Benzalkonium Chloride, PBS − Phosphate Buffer Solution, TBUT − Tear Break Up Time, TUNEL - terminal deoxynucleotidyl transferase dUTP nick end labeling, qRT-PCR − Qualitative Reverse Transcription Polymerase Chain Reaction, ELISA - Enzyme-Linked Immunosorbent Assay, ROS- Reactive Oxygen Species, FML - Fluorometholone, AA- Ascorbic Acid, DHE − Dihydroethidium

Table 5. Articles on the outcomes of exosome treatment on the salivary glands of Sjögren’s syndrome animal models.

Study	Experimental target	Source of exosomes	Exosome volume	Exosome size	Exosome delivery method	Exosome content/ markers	Outcomes measured	Results
Xing et al. 2022	NOD female mice, positive control group with intragastric infusion of Hydroxychloroquine, negative control infused with PBS, treatment group had either LGMSCs or LGMSC-Exos	Human labial gland mesenchymal stem cells (LGMSCs)	50μg/mL in 200μL PBS of LGMSC-Exos or LGMSCs (1×10^6 in 200μL PBS)	Median 121 nm, range 80−497 nm	Intravenous on alternate days for 14 days	CD9, CD63, Psg101	Flow cytometry on splenic lymphocytes for CD19, CD27 and CD138. RNA isolation and qRT-PCR. Saliva flow rates, lymphocyte infiltration in submandibular glands.	Improved saliva flow rate & reduced lymphocytes in Exos group, similar saliva flow rates and lymphocyte infiltration between the LGMSC-Exos and LGMSC groups; reduced PRDMI (a gene regulating B cell differentiation) expression in PBMCs.
Kim et al. 2021	Mice model of primary SS, with female NOD mice over 4 months old	Human induced pluripotent stem cells from MSCs, exosomes were collected from early passage P5 and late passage P15 iMSCs	1.5×10^10 particles in 100μL PBS	50−300 nm	IV into tail vein twice a week for 2 weeks	CD9, CD63, CD81	Submandibular gland and serum collected for lymphocyte infiltration	P5 iMSC EVs led to significantly decreased lymphocyte infiltration but P15 iMSC EVs had minimal difference. Only P5 EVs decreased serum anti-Lo and anti-Ro52, decreased mRNA levels of Th1 and Th17 cells and increased regulatory cytokines IL−10 and TGF-b1.
Li et al. 2021	Female NOD and BALB/c mice. Positive control group treated with HCQ intragastrically and BMMSC injection into the tail vein, negative control group (PBS injection), treatment groups LGMSCs or LGMSC-Exos injection.	Human Labial gland mesenchymal stem cell exosomes (LGMSC-Exos)			Tail vein injection-LGMSC-Exos 50μ/mouse, or LGMSC 10^6/mouse	CD 9, CD 63, CD 81	Saliva flow rate, histological assessment of salivary gland, serum cytokines	LGMSC-Exos treated mice showed a significantly increased salivary flow rate by 1.4g/10min compared to control mice with PBS. Salivary gland lymphocyte infiltration was significantly reduced in the LGMSC and LGMSC-Exos groups compared to negative control.
Rui et al. 2021	Female C57BL/6 mice immunised with 200 μg of SG proteins via SC injection in the neck on day 0 and 7 and then booster injection on day 14, to stimulate ESS.	Olfactory ectodermal mesenchymal stem cells (C57BL/6 mice). Bone marrow mesenchymal stem cells (BMMSCs) from the tibia and femur of these mice.	60 μg/mL	50−150 nm	Two IV injection of 100 μg of OE-MSC-Exos or BM-MSC-Exos on days 18 and 25 after the first immunisation	CD9, CD63, IL−6, TLR4	Quality real time-PCR, saliva flow rates, histological analysis of submandibular glands for lymphocytic foci	Improvement in the saliva flow rates following OE-MSC-Exos or BM-MSC-Exos Rx by 120microL/15min. Reduced serum autoantibodies against SSA, M3R and SG antigens in the OE-MSC-Exos group compared to BM-MSC-Exos and control. Reduced histological scores of SG destruction in OE-MSC-Exos compared to BM-MSC-Exos.
Rui et al. 2022	Female C57BL/6 mice immunised with 200 μg of SG proteins via SC injection in the neck on day 0 and 7 and then booster injection on day 14, to stimulate ESS.	Olfactory ectodermal mesenchymal stem cells, from the nasal cavity of C57BL/6 mice	Not mentioned	Not tested	Intravenous exosomes injection-100 μg/mouse at 18- and 25-days post first immunisation	PD-L1	Saliva flow rate, salivary gland histology, Serum antibodies.	OE-MSC-Exos treated mice had smaller spleens, cervical LNs; significantly restored saliva flow by 80microL/15min with OE-MSC-Exos treatment at 35 days; reduced serum anti-SG, anti-M3R and anti-SSA
Chu et al. 2023	14-week-old NOD mice	Human exfoliated deciduous teeth		137 + / −6.2nm	50μg in 25μL of PBS exosomes injected in submandibular gland	CD9, CD63 and HSP70	Stimulated saliva flow, lymphocyte infiltration in SMGs, TUNEL staining for apoptotic cells, capase− 1 activity, ROS levels	7 weeks after exosome treatment, the saliva flow rate was increased from 38 in age matched NOD and PBS groups to 55μL/min/100g Reduced lymphocyte foci in treated SMG, reduced apoptotic cells, cleaved caspase−3 and decreased levels of ROS
Du et al. 2023	7- and 14-week-old NOD mice and age matched BALB/c mice	Human exfoliated deciduous teeth		126.5 + /−5.7	50μ g in 25μ L of PBS exosomes injected in submandibular gland or infused through intraoral duct orifice(OD) of SMG at 14 weeks age	CD9,CD63, CD81, HSP70	Stimulated saliva flow rate, lymphocyte infiltration of SMG with focus score and ratio index	Saliva flow rate improved from 22 to 55μL/min/100g in the NOD mice at 21 weeks compared to Exosomes group at 21 weeks. Focus score and ratio score of lymphocyte infiltration were both decreased in the exosomes group compared to age matched untreated and PBS treated groups.
Hu et al. 2023	NOD 2 female mice, 6 positive control with hydroxychloroquine gastric infusion, 6 negative control with PBS gastric infusion, 6 treatment group	Human dental pulp stem cell	5, 20 and 80μg/mL	30−150 nm	Weekly intravenous injections for 10 weeks	CD63, ALIX, TSG101	Saliva flow rate, immunohistochemistry and immunofluorescence, ELISA of venous blood for anti-SSA/Ro, anti-SSB/La	Increased saliva flow rate by 0.2g/10min, reduced lymphocyte foci in salivary glands, increased AQP5 and GPER expression; HCQ and DPSC-Exos downregulated anti-SSA/Ro and anti-SSB/La serum levels
Ogata et al. 2023	Female NOD mice, treatment group given exosomes, control group given PBS	iPSC with different graded concentrations of secreted HGF and TGF-b1	300μ/mL	200 nm	30 μg delivered once IV through the tail vein	CD9, CD81	Saliva flow rates, serum anti-SSA antibody, salivary gland inflammatory cell infiltration, mRNA of cytokines produced in the salivary glands	iPS-EVs (higher HGF and TGF-b1) improved saliva flow rates (17.5μl/15min), PBS group (10.5μl/15min) and the iPS-EVs (lower HGF and TGF-b1) 13μl/15min; decreased salivary gland inflammatory cell infiltration;
Zhao et al. 2023	4-month-old female NOD.B10.H2 mice, model of primary SS	Human induced pluripotent stem cells from MSCs, early passage PD15 and late passage PD45	1.5×10^10 particles in 100μL PBS		IV infusion of EVs dyed with DiR (near infrared fluorescent dye)		Mice imaged in vivo 1, 3 and 24hrs after injection, major organs visualised for presence of EVs	decreased serum anti-SSA, IL−6 and TNF-a; increased anti-inflammatory cytokines IL−10 and TGF-b1 DiR was strongest in the liver and spleen, but not present in cervical lymph nodes or SMG collected at 24hrs after injection. miR− 125b inhibitors incubated ageing iMSCs reduced leucocyte infiltrate in SMG, not with ageing EVs (with miR-125b acitivty).
Zhou et al. 2023	ESS mice model made by immunising C57BL/6 mice with SG proteins on day 0, 7 and a booster on day 14	Myeloid derived suppressor cells from the spleen of tumour bearing C57BL/6 mice		50−150 nm	100 μg/mouse delivered IV on days 18 and 25 after first immunisation	CD63, CD19, TSG101, but not calnexin	Saliva flow rates, serum autoantibodies against SG antigen, ANA, anti-M3R, size of the spleen, cervical lymph nodes and salivary glands	Improved saliva flow rate by 75microL/15min with exosomes compared to ESS alone. Reduced serum level of autoantibodies against SG antigen, ANA and anti-M3R antibodies; smaller sized spleen, cervical lymph nodes and salivary glands in the exosome treated mice

NOD-Non-Obese Diabetoc, PBS − Phosphate Buffer Solution, qRT-PCR- Qualitative Reverse Transcription Polymerase Chain Reaction, PBMC- Peripheral Blood Mononuclear Cell, HCQ- Hydroxychloroquine, MSC − Mesenchymal Stem Cell, BMMSC − Bone Marrow MSC, SG − Salivary Gland, SC − Subcutaneous, LN − lymph node, SMG − Submandibular Gland, TUNEL - terminal deoxynucleotidyl transferase dUTP nick end labeling, ROS − Reactive Oxygen Species, ELISA - Enzyme-Linked Immunosorbent Assay, AQP5 − Aquaporin 5, GPER − G-protein coupled estrogen receptor, iPSC − induced Pluripotent Stem Cells, HGF- Hepatocyte Growth Factor, SS- Salivary Gland, ESS − Experimental Sjögren’s Syndrome, ANA − Anti-nuclear Antibody

3.2. Diagnostic use of exosomes

3.2.1. Plasma, tears, and saliva EVs protein profile

The diagnostic predictability of exosomes has been evaluated in SS patients’ tears, plasma, and saliva. Differences are noted in SS patients’ plasma, tear, and saliva EVs versus healthy controls. The differences between the three body fluids vary, but pro-inflammatory molecules were identified in both tears and saliva. Ferrant and coworkers¹⁶ reported significant differences in the protein profile of the plasma EVs of SS (n = 12) and healthy subjects (n = 8). The neutrophil-derived CD15 but not T cell-derived CD3, B cell-derived CD19, nor myeloid-derived CD18 and CD68 EVs were increased in SS patients. The EV size and concentration did not differ between the two groups. SS is driven by T and B cells, but surprisingly, the exosomes did not differ in their respective EVs. Among protein markers, RAB10, CD36, olfactomedin-4 (OLFM4), and integrin alpha M (ITGAM)/CD11 were expressed more in SS than in controls. Other than plasma, tear EVs also demonstrated differences between SS and healthy subjects. Aqrawi and coworkers³ studied the tear and saliva exosomes from SS patients (n = 27) and healthy controls (n = 32), and found that the proteins responsible for innate immunity, cell repair, and wound repair, such as neutrophil gelatinase-associated lipocalin (LCN2), granulins (GRN), calmodulin (CALM), epididymal secretory protein 1 (NPC2), and calmodulin-like protein 5 (CALML5) were upregulated in the saliva of SS when compared to healthy controls. Among tear EVs, the tear proteins such as DNA (apurinic or apyrimidinic site) lyase (APEX1), thioredoxin-dependent peroxidase reductase (PRDX3), copine (CPNE1), aconitate hydratase (ACO2), and LIM domain only protein 7 (LMO7), responsible for TNF-α signaling and B cell survival were upregulated in SS tears compared to healthy controls. Another study² on the salivary gland tissue, tear fluid, and saliva of non-SS (n = 15), SS (n = 10), and healthy controls (n = 10) revealed upregulation of Ficolin-1 (FCN1), CD44 and ANXA4 and LIM domain only protein 7 (LMO7), E3 ubiquitin-protein ligase HUWE1 (HUWE1) and Tumour protein D52 (TPD52) in both saliva and tears of SS compared to healthy controls. The identified proteins are responsible for proinflammatory pathways, ubiquitination, B cell differentiation, innate MHC class I cellular regulation, and T cell activation. In non-SS subjects, none of the EV proteins (identified in SS) showed expression either from tears or saliva, but mild inflammation in the salivary gland tissue on histopathology compared to SS patients. The sample size in these studies is small, and these results must be tested in patients with varying disease severity and correlated with other standard diagnostic methods.

3.2.2. Plasma, tears, and saliva EVs miRNA profile

Micro RNA profile has also been explored in plasma, tears, and saliva. Different miRNAs are expressed in the tears and saliva of DED patients. RNA profile of exosomes shows different micro-RNA (miRNA) and circular RNAs (circRNAs). One study has looked at circular RNAs (circRNAs) besides miRNAs. Li et al.³¹ compared the non-invasive cir-RNA profile from plasma EVs and MSG in primary SS (n = 4) and non-SS subjects (n = 4). They found 10 upregulated cir-RNAs in MSG, Circ-IQGAP2, and circ-ZC3H6 were upregulated in both plasma and MSG EVs.

Further, these two cir-RNAs correlated with clinical features, serum IgG level, and MSG focus scores and can be explored as a diagnostic marker for the SS diagnosis. Yamashiro and coworkers⁵⁴ first studied the EV miRNA in the mouth rinse samples of SS patients (n = 24) and compared them with healthy subjects (n = 24). As saliva secretion is less or nil in these patients, mouth rinse samples were used, which are noninvasive and easy to obtain in larger quantities. Interestingly, EVs were identified in these samples. RNA profile revealed that of 12 miR-NAs, 4 were overexpressed: let-7b-5p, miR-1290, miR-34a-5p, and miR-3648. ROC analysis for the diagnostic ability of the biomarker revealed that a combination of miR-1290 and let-7b-5p produced higher diagnostic ability with an AUC of 0.85. Recently, Cross and coworkers⁸ studied the RNA profile of the salivary EVs in SS patients (n = 11) to determine if it can serve as biomarkers for the disease. The transcripts of mRNA such as AC087392.1 and MTRNR2L2 (seysnoy gene), miRNA such as MIR4472−2 and MIR3135A, tRNA such as tRF tRNA-Gly-GCC-5 were significantly higher in the SS compared to controls. Pucker and coworkers³⁹ studied whether the inflammatory RNAs are present in the tear film of dry eye disease subjects (n = 5) defined based on the presence of signs and symptoms compared to non-DED subjects (n = 5). Tears were collected using eye wash with PBS, and EVs were isolated and characterized. It was found that, among the inflammatory RNAs identified, 9 of them were found to be upregulated, which includes miR-127−5p, miR-1273h-3p, miR-1288−5p, miR-130b-5p, miR-139−3p, miR-1910−5p, miR-203b-5p, miR-22−5p, and miR-4632−3p. Further validation of these results is needed as the sample size was small.

Similarly, Cross and coworkers⁹ studied the tear EV RNA profile from dry eye patients with unstable (n = 5) and stable tear films (n = 5). This study collected tears using Schirmer strips, and the EVs were isolated and characterized. Interestingly, the mRNA sodium channel modifier 1 (SCNM1) and the immature miRNA-130b were expressed 3.8 times lower and 1.5 times higher, respectively, in the unstable tear film group as compared to the stable tear film group, indicating that the EV RNAs are secreted variably within the dry eyes too. It will be interesting to compare the EV yield using ocular surface wash versus the Schirmer strip-based isolation method.

3.2.3. Animal studies on diagnostic use

There is little evidence available from DED animal model studies. Only one study by Kakan and coworkers^{25; 42; 51} studied the serum exosomes of 5 male NOD mice (a model for early-intermediate SS) and found that 5 miRNA (miR-127−3p, miR-329−5p, miR-409−3p, miR-410−3p, and miR-541−5p) were upregulated in NOD mice when compared to BALB/c mice. These miRNAs might be involved in the disease pathogenesis and may be used as diagnostic biomarkers of SS; however, their expression within tears and salivary glands still needs to be determined.

3.3. Therapeutic role of MSC-derived exosomes on DED

Current treatments for SS include nonsteroidal anti-inflammatory drugs, corticosteroids, and potent immunosuppressants such as metho-trexate, mycophenolate, and biological agents.²³ Recent developments in stem cell therapy have brought to light the potential benefits of stem cells in dampening the immune response and reversing the damage done in DED by immunomodulation, tissue repair, and anti-inflammatory properties. Mesenchymal stem cells (MSCs) have properties of immune regulation, tissue regeneration, and suppressing inflammation. Using MSCs directly in DED comes with the risks of cell rejection, minor vessel damage, and reduced permeability through the blood-retinal barrier. Exosomes produced by MSCs will have the same functions and can avoid these risks. MSC-derived exosomes have been experimentally tested in animals (in vivo studies) and in vitro human studies for treating DED. We identified 18 studies that tested exosomes for treating non-SS DED (6) and SS (12). The source of exosomes included murine bone marrow MSCs,⁴⁷ murine adipose tissue MSCs,⁴⁹ murine olfactory ecto-MSCs,^44,45 murine myeloid-derived suppressor cells,⁶⁴ human umbilical cord MSCs,^19,30,50 human adipose tissue MSCs,⁵⁹ human labial gland MSCs,^{32, 52} human dental pulp MSCs,²² human exfoliated deciduous teeth MSCs^{6, 11} and human induced pluripotent stem cells.^27,36,63 The exosome de- livery methods were topical in 6 studies and via subconjunctival injection in 1 study. In SS animal models, there were nine studies on intravenous use and two studies on submandibular gland injection.^6,11 The in vitro studies co-incubated exosomes with salivary gland epithelial cells,²² human corneal epithelial cells,^19,33,47,59 human peripheral blood mononuclear cells,^32,52 murine CD4+T cells,⁴⁴ CD19+B cells,⁶⁴ murine activated splenocytes,^27,64 human macrophages,³⁰ human labial gland cells,⁶ and human submandibular gland tissue.¹¹ Tables 3–5 summarize the outcomes of exosome treatment in different in-vitro and in-vivo studies on animals and humans.

3.3.1. Exosomes in non-SS DED animal models

Six studies were conducted on the non-SS DED model. Non-SS mouse models include the BAC model (1 week of 0.2 % BAC BID), desiccating model, or scopolamine model.^{2,19,23,25,36,63} Exosome delivery was topical 2−4 times/day in all 6 studies for a duration ranging from 5 to 21 days. Co-incubation of human corneal epithelial cells with exosomes varied from 4 to 24 h in vitro. Individual study details are summarized in Table 3.

3.3.1.1. Changes in tear film parameters

The outcomes measured across the studies included tear production, corneal fluorescein score, histological assessment, corneal integrity, and conjunctival goblet cell concentration. Additionally, the in vitro studies evaluated the apoptotic cells and their markers, cytokine profiles, reactive oxygen species activity, macrophage immunophenotype, dendritic cells, and Th17 cells. Tear volume and corneal fluorescein staining improved in all six studies compared to phosphate buffer solution (PBS) or saline treatment. A mean absolute increase of 2.4 mm was noted in Schirmer strips (Table 6). A direct comparison in tear volume production between studies to determine the best exosome type is unreliable. There are too many variables between studies, including duration and intensity of dry eye disease stimulation, concentration and duration of exosome treatment, and different sample sizes. Overall, it can be concluded that exosomes improved tear production in DED mice models. The mean absolute improvement across the studies in CFS with exosomes compared to PBS or saline was 6.5 (Tables 1 and 2). Of note is that the coupling of exosomes with ascorbic acid and cerium oxide⁴⁷ led to an even further reduction in CFS compared to exosomes alone. This is due to the increased reactive oxygen species scavenging ability.

Table 6. Changes in tear production in different animal models.

Study	DED model type and treatment route	Exosomes group	PBS/saline group	Absolute difference between exosome and PBS group	Healthy control	Absolute difference between exosome and healthy control
Tian et al., 2023*	C57BL/6 mice BAC model, topical	MSC-Exo 3.5 mm, MSC- Exo-Ce 4.5 mm	1.5 mm	2-3 mm	5.2 mm	−1.7 to −0.7 mm
Ma et al., 2023 *	C57BL/6 mice BAC model, topical	MSC-Exo 3.5 mm, MSC- Exo-AA 5.5 mm	2 mm	1.5-3.5mm	6.5 mm	−1 to −3 mm
Wang et al., 2022*	C57BL/6 mice BAC model, topical	5 mm	2.5 mm	2.5 mm	6 mm	−1mm
Guo et al., 2022	C57BL/6 mice BAC model, topical	2.663 mm	1.850 mm	0.813 mm	4.375 mm	−1.712mm
Wang et al, 2023 *	C57BL/6 mice BAC model, topical	4 mm	2 mm	2 mm	1 mm	3 mm
Yu et al., 2020	C57BL/6 mice BAC model, topical	6.10 mm	2.40 mm	3.7 mm	8.36 mm	−2.26

^*Values obtained from the graph as no absolute values given in the article

3.3.1.2. Histological changes

The histological assessment also demonstrated improved corneal and conjunctival epithelial integrity in exosome-treated groups. Tian and coworkers⁴⁷ and Ma and coworkers³³ demonstrated improved corneal epithelial integrity through fluorescein staining and increased central corneal thickness and improved stromal architecture seen on histological assessment with exosomes derived from murine MSCs in vivo with the BAC model of C57BL mice compared to PBS. In vitro, Tian and coworkers⁴⁷ demonstrated better reepithelialization of human corneal epithelial cells with a smaller percentage of wound surface area remaining with murine bone marrow-derived MSC-Exo-Ce, compared to MSC-Exo alone or PBS (26.63 % vs. 34.83 % vs. 77.53 %). PAS-stained conjunctival goblet cell density improved with exosome treatment.^19,50,59 The TUNEL assay detected apoptotic cells,^47,49,59 and found that exosome treatment inhibited corneal cell apoptosis. Wang and coworkers⁴⁹ demonstrated a statistically significant reduction in corneal epithelial cell apoptosis (p < 0.01) in the 50 mg/mL exosomes group compared to the model group, and Yu and coworkers⁵⁹ similarly demonstrated reduced TUNEL positive cells under hyperosmotic stress with the addition of exosomes (p < 0.05).

3.3.1.3. Molecular changes

Proinflammatory cytokines were reduced in vitro and in vivo upon exosome treatment in all 6 studies. The corneal epithelium responds to ocular surface damage, such as from tear film hyperosmolarity, and this stimulates the innate immune system activation with the release of NLRP3 inflammasome. Wang and coworkers⁴⁹ and Yu and coworkers⁵⁹ demonstrated exosome-mediated inhibition of NLRP3 inflammasome activation. One of the downstream effects of NLRP3 is the release of pro-inflammatory cytokine IL-1β. Topical exosome treatment reduced IL-1β levels compared to PBS or saline groups across five studies.^{19,33,47,49,59} Other pro-inflammatory cytokines shown to be suppressed by exosomes include IL-6, caspase-1, IL-18, IFN-γ, and TNF-α.⁵⁰ Guo et al.¹⁹ demonstrated that topical administration of human umbilical cord MSC-derived exosomes decreased Th17 helper T cells in the lymph nodes of mice and decreased dendritic cells in the cornea compared to PBS groups. Th17 cells are known to be involved as a bridge between innate and adaptive immunity, and dendritic cells function as antigen-presenting cells for B cell activation. Ma and coworkers³³ cultured murine MSC exosomes with murine peritoneal macrophages and found an increase in macrophages’ M2 (anti-inflammatory) phenotype. Immunofluorescence of the corneal epithelial cells by Wang and coworkers⁵⁰ found reduced corneal CD4 T cells after topical appli- cation of human umbilical cord MSC-derived exosomes compared to the dry eye model in C57BL mice. NF-κB is a known signaling pathway stimulated by desiccating stress in the cornea. Wang et al.⁵⁰ established that NF-κB was downregulated by several abundant exosomal miRNAs, including miR-125 b, let-7b, and miR-6873. Together, these studies affirm exosomes’ significant role in directly dampening the immune response in DED.

3.3.2. Exosomes in SS animal models

We identified 12 studies exploring the exosome effect in SS, including animal and human in vitro models (Tables 4, 5). SS was replicated in animals using female NOD mice in 8 studies,^{6,11,22,27,32,36,52,63} autoimmune sialadenitis female C57BL/6 mice in 3 studies,^43,45,63 and an autoimmune dacryoadenitis rabbit model in 1 study.³⁰ Exosomes were delivered intravenously with varying frequency and concentration in 9 studies, injected into the submandibular gland, and given as a subconjunctival injection in 1 study each.³⁰ No study compared the differences between different routes of administration. As SS is a multisystem disease, the results discussed are tears and salivary flow rate changes.

Table 4. Summary of in-vitro studies assessing the exosome effect on human corneal cell line, salivary glands and mononuclear cells.

Study	Experimental target	Source of exosomes	Exosome volume	Exosome size	Exosome delivery method	Exosome content/ markers	Outcomes measured	Results
Yu et al. 2020	Human corneal epithelial cells exposed to hyperosmolar medium for 24 h	Human adipose tissue MSCs	1μg/μL	100 nm	Co-incubated with HCECs for 4h	TSG101, CD63, ALIX	TUNEL assay, cell viability assay, NLRP3, ASC, capase− 1 and IL− 1b levels	TUNEL positive cells increased in hyperosmotic medium (17.42 +/- 4.20% vs 1.55 +/−0.55 %), that decreased after treatment with exosomes, with TUNEL positive cells reducing to 4.88 + /−2.49 %. Increased mRNA expression of NLRP3, ASC, capase−1 and IL−1b post hyperosmotic stress reduced by exosome treatment
Rui et al.2021	Mouse CD4 + T cells	Olfactory ectodermal mesenchymal stem cells (OE-MSC), Bone marrow mesenchymal stem cells (BM-MSCs) derived from the tibia and femur		50−150 nm	OE-MSC-Exos or BM-MSC-Exos were co-incubated with MDSCs for 2 days, and then co-cultured with mouse CD4 + T cells.	CD9, CD63	CFSE fluorescence to evaluate CD4 + T cells by flow cytometry, western blot analysis	OE-MSC-Exos promoted expansion of MDSCs much more than BM-MSC-Exos. OE-MSC-Exos led to enhanced suppressive effect of MDSCs on CD4 + T cell proliferation.
Kim et al. 2021	TLR−4 stimulated splenocytes from adult male BALB/c mice	Human induced pluripotent stem cells from MSCs, exosomes from P5 and P15 iMSCs	1.5×10^10 particles in 100 μL PBS	50−300 nm	Co-incubation	CD9, CD63, CD81	Cytokines secreted by activated T lymphocytes, mRNA expression of markers for IL−17, IL−6, IL-1, TGF-b1, IFN-gamma. Protein profiling of P5 and P15 iMSC EVs	P5 iMSC EVs were more effective than P15 iMSC EVs in suppressing the secretion of pro-inflammatory Th1 and Th17 cytokines, IFN-gamma, IL− 6, IL−17 and increasing TGF-b1.
Li et al. 2022	Differentiated human macrophages	Human umbilical cord MSCs	na	50−150 nm	Differentiated macrophages were co-incubated with exosomes 5 μg/mL stained with PKH26 for 48hrs.	CD81, CD9, CD63, TSG101	mRNA and protein markers	mRNA levels of M2 markers, protein markers of M2 macrophages and expression of Treg markers were all increased in the exosomes group compared to control. Exosome educated macrophages when cultured with CD4 + T cells led to significant repression of the CD4 + T cells.
Xing et al. 2022	Human peripheral blood mononuclear cells (PBMCs)	Human labial gland mesenchymal stem cells (LGMSCs)	30μg/mL	Median 121 nm, range 80−497 nm	Co-culture of PBMCs and LGMSCs at a ratio of 10:1, in a subset an exosome inhibitor GW4869 was also added.	CD9, CD63, Psg101	Flow cytometry for B cell markers CD19, CD20, IgD, CD38, CD27, CD24. RNA isolation and qRT-PCR	PBMCs with LGMSC-Exos led to reduced CD19, CD20, CD24, CD38 plasma cells, this effect was less pronounced with the exosome inhibitor GW4869
Guo et al. 2022	Human corneal epithelial cells cultured in hyperosmotic culture medium to simulate DED	Human umbilical cord MSCs	na	80−180 nm	Co-culture of MSC-EVs with HCECs	CD9, CD63 and CD81 (Calnexin was negatively expressed)	Flow cytometry to detect Th17 cells, RT-PCR to detect mRNA expression of inflammatory cytokines, immunofluorescence to detect dendritic cells	MSC-EVs were protective to HCECs under hyperosmotic stress by suppressing inflammatory cytokines TNF-alpha, IL−6, IL−1b. MSC-EVs at concentrations ranging between 10 and 50μg/mL were protective for HCECs
Tian et al. 2023	Human corneal epithelial cells	Murine bone marrow, Ce (NO3)3·6H2O was incorporated into the MSC-Exos to form MSCExo-Ce	20−40μ/mL	100−150 nm	Co-incubation	CD 48, CD 63, TSG101	ROS scavenging activity assay e, uptake of MSC-Exo on HCECs via Cy5 staining, in vitro scratch wound assay, mitochondrial ROS scavenging, cytotoxicity study of MSC-Exo-Ce	MSC-Exo-Ce demonstrated superior ROS scavenging ability compared to MSC-Exos. HCEC uptake of exosomes was not impaired by Ce coupling. HCEC re-epithelialisation was significantly improved with MSC-Exo-Ce, with a smaller wound surface area (27.63) compared to MSC-Exo (34.83) or PBS (77.53).
Li et al. 2021	Peripheral blood mononuclear cells from SS patients	Human Labial gland mesenchymal stem cell exosomes (LGMSC-Exos)	30μ/mL		LGMSCs and LGMSC-Exos co-cultured with PBMSCs for 72hrs (1:10)	CD 9, CD 63, CD 81	Percentage of Th17 and Treg, cytokines in the supernatant	LGMSC and LGMSC-Exos significantly suppressed the cytokines IL−17A, IFN-γ, IL−6, and TNF-α and increased TGF-β and IL-10.
Chu et al. 2023	Damaged labial gland tissue from SS patients	Human exfoliated deciduous teeth	na	137 + / −6.2nm	Co-cultured for 24 hrs	CD9, CD63 and HSP70	Cell death markers ACSL4, KEAP1, RIPK3	Cell death markers expression downregulated in SS labial glands. Exosomal miRNA associated with cell death pathways were identified to be KRAS, HRAS, MEK1/2 and the downstream signalling molecule p-ERK1/2
Du et al. 2023	Human submandibular gland tissue	Human exfoliated deciduous teeth	200μ/mL	126.5 + / −5.7	Co-cultured for 24hrs	CD9, CD63, CD81, HSP70	Exosome uptake into the glandular cells	PKH− 26-stained exosomes were found in the cytoplasm of SMGC6 cells, acinar cells and duct cells.
Hu et al. 2023	IFN-γ treated salivary gland epithelial cells	Human dental pulp stem cell (DPSC)	5, 20 and 80μg/mL	30−150 nm	Different concentrations of DPSC-Exos were added to SGECs for 24, 48 and 72 h	CD63, ALIX, TSG101	AQP5, GPER expression	Increased proliferation, AQP5 and GPER expression in the IFN-γ treated SGECs with 20 and 80μg/mL DPSC-Exos.
Zhou et al. 2023	CD19 + B cells were cultured under Germinal Centre B polarising conditions	Myeloid derived suppressor cells from the spleen of tumour bearing C57BL/6 mice		50−150 nm	Co-cultured for 96 h with CD19 + B cells	CD63, CD19, TSG101, but not calnexin	Differentiation of B cells	Exosome treatment suppressed differentiation of B cells, decreased mRNA and protein levels of Bcl−6. miR− 10A− 5p is an exosome miRNA thought to be involved in regulating Bcl−6 expression
Zhao et al. 2023	Isolated splenocytes from 4-month-old female NOD.B10. H2 mice, model of primary SS	Human induced pluripotent stem cells from MSCs, early passage PD15 and late passage PD45	3×10’9 particles/mL		Early and late passage EVs were labelled with PKH26 and co-incubated with splenocytes for 3 hrs		Flow cytometry analysis for PKH26, macrophage, T cell and B cell markers	Macrophages were the major population > 70 % uptaking iEVs both early and late passage, Young EVs (early passage PD15) promoted M2 polarisation of splenic macrophages that was maintained 2 weeks after the last IV treatment in PD15 (young) but not with PD45 (aged) EVs
Ma et al. 2023	Human corneal epithelial cells monolayer with scratch assay. Incubated with AA, mExo, mExo@AA and saline for 12 h	Mouse mesenchymal stem cell derived exosomes with and without ascorbic acid coupling	0.01−1.0 mg/mL	100−150 nm	Unknown	ALIX, beta-actin, TSG 101 positive, negative for GM130	Re-epithelialisation of the monolayer of HCECs, ROS scavenging ability, macrophage immunophenotype	mExo@AA accelerated wound closure compared to saline, mExos and AA alone, increased ROS scavenging ability. mExo@AA showed the greatest increase in M2 phenotype (anti-inflammatory) of macrophages; mExo and mExo@AA reduced pro-inflammatory cytokines IL−6 and IL-1β, with mExo@AA and mExo having similar effects.
Zhou et al. 2023	CD19 + B cells were cultured under Germinal Centre B polarising conditions	Myeloid derived suppressor cells from the spleen of tumour bearing C57BL/6 mice		50−150 nm	Co-cultured for 96 h with CD19 + B cells	CD63, CD19, TSG101, but not calnexin	Differentiation of B cells	Exosome treatment suppressed differentiation of B cells, decreased mRNA and protein levels of Bcl−6. miR− 10A− 5p is an exosome miRNA thought to be involved in regulating Bcl−6 expression

MSC-Mesenchymal Stem Cell, PBS − Phosphate Buffer Solution, SS − Sjögren’s Syndrome, DED-Dry Eye Disease, HCEC − Human Corneal Epithelial Cell, TUNEL - terminal deoxynucleotidyl transferase dUTP nick end labeling, ASC - Apoptosis-associated speck-like protein containing a caspase recruitment domain, CFSE - Car-boxyfluorescein succinimidyl ester, MDSC − Myeloid Derived Stem Cells, qRT-PCR - Qualitative Reverse Transcription Polymerase Chain Reaction, ROS − Reactive Oxygen Species, SGEC − Salivary Gland Epithelial Cell, AQP5 − Aquaporin 5, GPER − G-protein coupled estrogen receptor, NOD − Non-Obsese Diabetic, AA − Ascorbic Acid

3.3.2.1. Changes in tear film parameters

The outcomes measured included saliva flow rate, lymphocyte infiltration in submandibular glands, SS antibodies, AQP5/GPER expression, exosomal miRNA profiling, lymphocyte profiling with flow cytometry, macrophage phenotype, TUNEL assay for apoptotic cells, and cytokine profiling. The stimulated saliva flow rate was measured by injecting anesthetized mice with pilocarpine and then collecting saliva into a glass capillary kept under the tongue for 10 or 15 min. A comparison could be made between studies using the same units of measurement, producing a mean absolute increase in saliva flow rate by 1.1g/10min^22,32,52 mean absolute increase of 70.5μL/15min, ^36,44,45,64 and a mean absolute increase by 25μL/min/100g^6,11 following exosome treatment compared to PBS control.

3.3.2.2. Histological changes

Ten studies, in varying ways, measured lymphocyte infiltration. Hu and coworkers²² found reduced lymphocyte foci in the salivary glands of NOD/ltj mice after intravenous delivery of human dental pulp MSC-derived exosomes but did not measure this quantitatively. Salivary gland lymphocyte infiltration area in NOD mice was measured as a ratio to normal glandular tissue area and found to decrease by 0.013³² and 0.018⁵² in intravenous exosome-treated groups compared to PBS. The effect on lymphocyte foci was variable depending on the culture age. Kim and coworkers.²⁷ found that early passage exosomes from human induced pluripotent stem cells produced a significant decrease in the percentage area of lymphocyte infiltration in salivary glands of NOD.B10.H2^b mice from 4 % in the PBS group to 1 % in the early passage group; this effect was almost ameliorated with late passage exosomes having no significant difference compared to PBS. Ogata and coworkers³⁶ tested exosomes from induced pluripotent stem cells, modified to produce less HGF (hepatocyte growth factor) and TGF-b1, which are involved in tissue regeneration and immunosuppression. The focus score (number of inflammatory cells per 10 mm² in salivary glands of NOD mice decreased from 8 in the PBS group to 4.8 in the less HGF and TGF-b1 producing exosomes and decreased further to 2.1 in the group of exosomes modified to produce more HGF and TGF-b1. It indicates that TGF-b1 and HGF may be important in reducing immune system activation in SS mice. The remaining 5 studies^{6,11,30,44,63} used a common method of assessing lymphocyte infiltration in the salivary glands of NOD mice or lacrimal gland of rabbits³⁰ with a focus score measured as the number of foci with inflammatory infiltrates of at least 50 cells per 4 mm,² the mean focus score in PBS group was 2.21, and the mean absolute decrease between PBS and exosome groups was 1.5. Zhou and coworkers⁶⁴ demonstrated that the size of the spleen, cervical lymph nodes, and salivary glands in the ESS mice model was significantly smaller in the intravenous exosome-treated mice than in the control. These are organs containing lymphatic tissue with germinal centers, and it is known that the germinal center B cell population correlates positively with the severity of experimental SS.⁶⁴ Intravenous use of exosomes in SS mice models reduced the serum antibody levels of anti-SSA/Ro.^22,36,45

3.3.2.3. Molecular changes

Aquaporin 5 (AQP5) is a transmembrane channel expressed in salivary and lacrimal gland epithelial cells that promotes saliva secretion; the inflammatory process in SS is known to downregulate AQP5 to decrease saliva secretion directly. GPER (G protein-coupled estrogen receptor) is a gene that stimulates cAMP and causes increased intracellular Ca concentration; this, in turn, leads to AQP5 transmembrane channels being transported to the cell membrane to be utilized in saliva secretion. Hu and coworkers²² demonstrated that human dental pulp stem cell-derived exosomes increased AQP5 and GPER expression in vivo in NOD/ltj mice and in vitro in IFN-γ treated salivary gland epithelial cells. The direct molecular pathway that allows exosomes to target AQP5 and GPER function remains unknown. Du and coworkers¹¹ further demonstrated that SS patients had decreased AQP5 expression compared to controls. Still, the in vitro study of human SMG tissue from SS patients could not show the impact of co-culturing exosomes for 24 h on AQP5 levels. Interestingly, Du and coworkers¹¹ did show improved glandular function in NOD mice with submandibular injection of human exfoliated deciduous teeth-derived exosomes through increased levels of ZO-1 (zonula occudins-1), which is involved in paracellular permeability.

3.3.3. Proposed mechanisms of action

3.3.3.1. Micro-RNAs

Exosomes are known to contain miRNA, which is non-coding regulatory RNA that is involved in post-transcriptional modulation of gene expression. A commonly studied mi-RNA is miR-125b in SS models. Xing et al.⁵² identified PRDM1, a gene promoting B cell differentiation. PRDM1 has increased expression in peripheral blood mononuclear cells of NOD mice, which decreased following intravenous exosome treatment. They further found through dual luciferase reporter assays that miR-125b (an exosomal miRNA) targets PRDM1 by reducing its expression. miR-125b was selectively over-expressed in one group of exosomes and selectively knocked down in another group of exosomes; they discovered that over-expression of miR-125b led to a reduction in mRNA levels of PRDM1 and the proportion of plasma cells was also significantly reduced, knockdown of miR-125b had the opposite effects. Zhao and coworkers⁶³ also studied miR-125b, but found its effects on the immune system contradict those described by 2 other papers^52,63 that studied the use of exosomes from induced pluripotent stem cell-derived MSCs. They directly compared exosomes from the early passage (young) iMSCs to exosomes from the late passage (aging) iMSCs. They demonstrated that young iEVs but not intravenous aging iEVs promoted M2 (anti-inflammatory) macrophage phenotype in NOD mice’s spleen and inhibited Th17 cell differentiation. They went further to find that miR-125b was more abundant in aging iEVs, and suppression of miR-125b in aging iEVs enabled the aging iEVs to suppress Th17 cells and promote M2 macrophages, as did the young iEVs. Therefore, Xing and coworkers⁵² found that miR-125b expression has immunosuppressive effects in reducing plasma cells. On the contrary, Zhao and coworkers⁶³ found miR-125b inhibition to produce immunosuppressive effects with fewer Th17 cells and more M2 macrophages.

Li and coworkers³⁰ found that miR-100−5p was abundant in the exosomes isolated from human umbilical cord MSCs. Overexpression of miR-100−5p led to an increased polarisation of the macrophages from the M1 (pro-inflammatory) to M2 (anti-inflammatory) phenotype; this was confirmed both in vivo in the rabbit lacrimal glands and in vitro with induced macrophages, and also by knockdown of miR-100−5p producing the opposite effect of increased M1 macrophages. Furthermore, Kim and coworkers²⁷ explored early and late passage exosomes from human induced pluripotent stem cells and discovered that miR-21−5p and miR-125b-5p varied in concentration between these two exosome lines. Silencing miR-21−5p led to decreased inhibitory effect of exosomes on IL-6, IL-17, and IFN-gamma levels in serum and SMG of NOD mice; in contrast, silencing miR-125b-5p led to enhanced immu-nosuppressive effects. These effects were confirmed by Kim and co-workers²⁷ with in vitro studies in TLR-4 stimulated spleenocytes from adult male BALB/c mice. Zhou and coworkers⁶⁴ explored the effect on Bcl-6 expression (involved in B cell differentiation) from myeloid-derived suppressor cells exosomes when co-incubated with activated CD19 + B cells in vitro. They concluded that inhibition of miR-10A-5p leads to Bcl-6 upregulation and increased B cell differentiation. These studies have described some of the underlying molecular pathways exosomes employ in immunomodulation. They have also highlighted their complexity and how the same miRNA can have opposing effects in different situations.

3.3.3.2. PD-L1 mediated T cell inhibition

Another molecular pathway exosomes employ is PD-L1, a programmed cell death ligand that inhibits the activation of T follicular helper cells (Tfh). Rui and coworkers⁴⁵ found a reduction in Tfh and germinal center B cell concentrations in the spleen and lymph nodes of C57BL/6 mice after intravenous treatment with olfactory ectodermal MSC-derived exosomes. Selective knockdown of PD-L1 in exosomes led to amelioration of the immunosuppressive effects of exosomes, with the spleen and lymph nodes remaining enlarged, as well as salivary flow rate and serum antibodies remaining comparable to SS mice without exosome treatment.

3.3.3.3. Macrophages

In the lacrimal glands of rabbits treated with human dental pulp MSC-derived exosomes,³⁰ the gene expression of M1 macrophage markers (NOS2, IRF5, TNF-a, IL-1β and IL-6) was significantly reduced, and the gene expression of M2 macrophage markers (Arg1, CD206, KLF4, IL-10, TGF- β) was significantly increased. Exo-somes co-cultured with differentiated macrophages in vitro converted them into an anti-inflammatory M2 phenotype, suppressing CD4 + T cells and promoting Treg cells. Regulatory T cells have an immuno-suppressive function, and TGF-β1 and HGF are known to induce Treg differentiation. Li and coworkers³⁰ found the proportion of Treg cells and their markers, including IL-10 and TGF-β, to increase in the lacrimal gland with subconjunctival exosome treatment compared to PBS. Mirroring this, Li and coworkers³² found Treg cells and their markers TGF-β and IL-10 to be increased in the spleens of NOD mice given intravenous exosomes.

Similarly, Ogata and coworkers³⁶ found an increase in IL-10 and TGF-β1 and increased Foxp3, a Treg marker, in the saliva of NOD mice treated with intravenous exosomes. Th17 cells are known to be a bridge between innate and adaptive immunity, and Th1 cells also play an essential role in autoimmunity. Li et al.³² demonstrated decreased Th17 cells and their markers IL-6 and IL-17 in the spleens of mice with exosome treatment. This was further corroborated by Rui and coworkers⁴⁴ by demonstrating decreased Th17 and Th1 cells and their markers IL-17 and IFN-γ in the serum of C57BL/6 mice following intravenous exosome treatment. Kim and coworkers²⁷ concluded that intravenous early passage exosomes were more effective than late passage exosomes in suppressing the pro-inflammatory cytokines of Th1 and Th17 cells, IFN-γ, IL-6, and IL-17 in NOD mice. Ogata and coworkers³⁶ further showed that exosome treatment led to suppression of IL-17a, and exosomes modified to produce more TGF-β1 and HGF had enhanced suppression of IL-17a.

3.3.3.4. Apoptosis inhibition

In the later stages of inflammation involved in SS, there is apoptosis of epithelial glandular tissue. Chu and coworkers⁶ found exosomes cocultured with labial gland tissue from SS patients decreased apoptotic cells and their markers (ACSL4, KEAP1, and RIPK3) compared to baseline tissue levels. These same human-exfoliated deciduous teeth-derived exosomes did not affect apoptosis in labial gland tissue from healthy patients. The target genes for exosomal miRNA involved in apoptosis were identified to be KRAS, HRAS, MEK1/2, and the downstream signaling molecule p-ERK1/2. Submandibular gland injection of exosomes in NOD mice led to decreased expression of these target genes and the target molecule p-ERK1/2, showing various pathways in which exosomes decrease cell apoptosis in SS.

4. Future research implications and roadblocks

4.1. Existing limitations

There are limitations in using exosomes as a therapy, including the need for a standard protocol for isolation, purity during extraction, and the differential yield of exosomes from the samples. There is currently no standardized protocol for the isolation of exosomes, although they are derived from similar cell lineages such as MSCs or induced pluripotent stem cells, and the process typically involves a series of differential centrifugation, ultracentrifugation, and storage at low temperatures. There is a need to create standard manufacturing kits and protocols for achieving a consistent exosome concentration and size that is sustainable and cost-effective. The challenge in creating standardized exosome populations lies in their innate heterogeneity based on their parent cell population, species of origin, and complex interactions with the cellular environment, which constantly changes in response to autoimmunity, malignancy, and inflammation. For example, adipose-derived MSC exosomes will have genetic and molecular characteristics different from bone marrow-derived MSC exosomes. Each population of exosomes may be better suited to certain therapeutic functions. The heterogeneity is unfathomable, and without a complete understanding of the role and function of all exosomal molecular pathways, the risk of detrimental off-target effects is significant. The limited knowledge of exosomal miRNA and their unique targets in the immune system is only the tip of the iceberg; much more extensive research and whole genome sequencing are needed to gain a deeper understanding of these molecular pathways. This may explain why the same miRNA (such as miR-125b) may have variable and opposing effects on the immune system in different circumstances.

4.2. Personalized exosome therapy

EVs/Exosomes can be used as carriers for targeted RNA or protein or drug delivery and as a target that can be inhibited with different drug molecules.⁴⁶ Exosomes have a hydrophilic core that can be used for loading water-soluble drugs. With a deeper understanding of miRNA-based molecular pathways, exosomes can be tailor-made with the required molecular targets upregulated or downregulated to produce the maximal immunosuppressive outcomes in humans. Based on the underlying disease process, the exosome delivery method can be best tailored, for example, topical or subconjunctival delivery of exosomes for DED and injection into the lacrimal and/or salivary glands for SS. By directly targeting certain molecular pathways and combining this with tailored delivery methods best suited for the disease process, the required exosome concentration will also be less, allowing for more efficacy and reproducibility.

4.3. Regulatory concerns

Exosomes are cell-free therapy. Hence, their categorization under stem cell regulatory guidelines is unclear in some countries. In the USA and Europe, EVs are considered as drugs or biological medicinal products, requiring approval from the governing bodies.¹⁷ In countries such as Japan, EVs are not covered under the Act on the Safety of Regenerative Medicine. International Society for Extracellular Vehicles has issued guidelines for safety issues associated with unproven EV therapies marketed as antiaging or for hair growth, as reports of sepsis were reported.¹⁷ There is a pressing need to develop regulations worldwide so that EVs as a therapy is controlled and any adverse reactions should be reported to authorities. There may also be challenges with corporate involvement and financial interest, which could lead to patenting the use of certain miRNAs or molecular pathways in exosomes.

5. Conclusion & future research implications

The studies on exosome use in DED demonstrate a significant response measured at a functional level with improved tear production and saliva flow rate, at a cellular level with reduced lymphocyte infiltration, enhanced structural integrity, and decreased epithelial cell apoptosis as well as at a molecular level with dampening of the inflammatory cytokine response. These changes were consistent across the studies despite the high variability in the source of exosomes, the concentration, and the duration of exosome use; however, different studies have shown different effects of the same miRNAs and used mainly BAC models for their effect demonstration. Hence, research is needed to delineate the underlying molecular pathways, including the genes and molecules targeted by exosomes, to confirm their positive effects. The real-world translation of exosome therapy in DED is not imminent because of the barriers of cumbersome exosome extraction processes, heterogenous miRNAs detected in patients, unclear mechanisms of action, and limited human studies.

Table 7. Changes in corneal fluorescein score following exosomes treatment.

Study	DED model type and treatment route	DED treated with Exosomes	DED given PBS/saline	Absolute difference between PBS group and exosomes	Healthy control	Absolute difference between exosome and healthy control
Tian et al. 2023 *	C57BL/6 mice BAC model, topical	MSC-Exo; 6MSC-Exo- Ce; 2	12.5	10.5−6.5	0.5	1.5−5.5
Ma et al.2023 *	C57BL/6 mice BAC model, topical	mExo; 7.5mExo@AA; 4	15	7.5−11	1	3−6.5
Wang et al. 2022*	C57BL/6 mice BAC model, topical	1.8	3.5	1.7	0.5	1.3
Guo et al. 2022 *	C57BL/6 mice BAC model, topical	3	5	2	0.5	2.5
Wang et al.2023 *	C57BL/6 mice BAC model, topical	2 in 3.0 mg/mL Exo group	8	6	N/A
Yu et al.2020	C57BL/6 mice BAC model, topical	5.2	12.0	6.8	1.2	4

^*Values obtained from the graph as no absolute values given in the article

DED- Dry Eye Disease, PBS − Phosphate Buffer Solution, BAC − Benzalkonium Chloride, MSC − Mesenchymal Stem Cell

Table 8. Summary of implicated miRNAs and their targeted pathways from Sjögren’s syndrome and dry eye disease studies.

Study	Disease	Exosomal miRNA	Targeted Pathway
Xing et al. 2022	SS	miR−125b	Suppression of PRDM1 regulated B cell differentiation
Zhao et al. 2023	SS	miR−125b	M2-M1 macrophages imbalance, thereby affecting Th17 cells
Kim et al. 2021	SS	miR−125b	M2-M1 macrophages imbalance, Th1 and Th17 cytokines
Wang et al. 2023	DED	miR−125b	Downregulation of NF-κB
Kim et al. 2021	SS	miR−21−5p	Inhibits secretion of IL−6, IFN-γ and IL−17
Wang et al. 2023	DED	miR−6873	Downregulation of NF-κB
Li et al. 2022	SS	miR−100−5p	Increased polarisation of macrophages from M1 to M2 phenotype
Zhou et al. 2023	SS	miR−10A−5p	Inhibits Bcl−6 and B cell differentiation