Orally Administered Exosomes Alleviate Mouse Contact Dermatitis through Delivering miRNA-150 to Antigen-Primed Macrophages Targeted by Exosome-Surface Antibody Light Chains

We previously discovered suppressor T cell-derived, antigen (Ag)-specific exosomes inhibiting mouse hapten-induced contact sensitivity effector T cells by targeting antigen-presenting cells (APCs). These suppressive exosomes acted Ag-specifically due to a coating of antibody free light chains (FLC) from Ag-activated B1a cells. Current studies aimed at determining if similar immune tolerance could be induced in cutaneous delayed-type hypersensitivity (DTH) to the protein Ag (ovalbumin, OVA). Intravenous administration of a high dose of OVA-coupled, syngeneic erythrocytes induced CD3+CD8+ suppressor T cells producing suppressive, miRNA-150-carrying exosomes, also coated with B1a cell-derived, OVA-specific FLC. Simultaneously, OVA-immunized B1a cells produced exosome subpopulation, originally coated with Ag-specific FLC, that could be rendered suppressive by in vitro association with miRNA-150. Importantly, miRNA-150-carrying exosomes from both suppressor T cells and B1a cells efficiently induced prolonged DTH suppression after single systemic administration into actively immunized mice, with the strongest effect observed after oral administration. Current studies also showed that OVA-specific FLC on suppressive exosomes bind OVA peptides, suggesting that exosome-coating FLC target APCs by binding to Ag-major histocompatibility complexes. This renders APCs able to inhibit DTH effector T cells. Thus, our studies described a novel immune tolerance mechanism mediated by FLC-coated, Ag-specific, miRNA-150-carrying exosomes that are particularly effective after oral administration.


38
Recent studies have shown that immune regulation in vivo is more diverse than previously 39 appreciated. A unique antigen (Ag)-specificity of T cell immunosuppression was described 40 previously in contact sensitivity (CS) induced by epicutaneous immunization of mice with reactive 41 hapten Ags [1][2][3][4][5]. This form of Ag-specific T cell tolerance is systemically generated by intravenous elsewhere. In addition, when OVA-specific Ts cell-derived exosomes were separated on OVA-linked 145 affinity column, they also strongly suppressed adoptively transferred DTH response (Figure 2A).

146
In summary, all of these findings concerning phenotypic and descriptive properties of the OVA 147 Ts exosomes, fit with the recent most thorough description of true classical exosome characteristics 148 [13]. We also have linked this phenotype to the suppressive function of EVs by testing of their 149 biological activity after separation on OVA Ag and anti-CD9 affinity columns ( Figures 1F and 2A).   Ts cell-derived exosomes that had bound to OVA Ag-linked sepharose columns were found to 186 suppress DTH effector cells (Figure 2A), suggesting that APCs may be targeted by these exosomes.

187
Accordingly, previous studies on suppression of hapten-induced CS showed that Ag-presenting Figure 3C).

212
In kinetic testing, we found that the OVA-specific, Ts cell-derived exosomes strongly suppressed 213 subsequent daily continuing DTH responses at 48, 72, 96 and 120 hours after IP administration of a 214 single physiological dose ( Figure 3D). This experiment showed the clinically applicable in vivo 215 suppressive activity of the OVA-specific Ts cell-derived exosomes. 216 2.6. In vivo testing of Ag-specificity of the Ts cell-derived exosomes 217 Similar Ts cell-derived exosomes, from a comparable control DTH system induced by the 218 antigenically non-cross reactive protein keyhole limpet hemocyanin (KLH) showed that the OVA 219 exosome suppression was Ag-specific. In this case, KLH-specific exosomes harvested from lymphoid 220 cells of mice identically tolerized to KLH, and similarly IP injected into actively immunized mice just 221 after the peak of the active 24 hour OVA-induced DTH response, were totally inactive ( Figure 3E). In    Outstandingly, the most efficient suppression was observed after PO administration of the 245 single dose of the inhibitory exosomes ( Figure 3F, red line). Finally and interestingly, when treatment 246 with the suppressive exosomes was by the ID route in the torso, this also caused suppression of the 247 DTH responses ( Figure 3F, dashed black line). Note that this was via the same route used for Ag 248 immunization, but at different skin sites that also were distant from DTH elicited in the ears.   Ts cell exosomes that were also dependent on their ability to deliver miRNA-150 to targeted cells.

281
In the CS system, the suppressive exosome from hapten Ag tolerized mice, additionally were 282 activated to be able to associate with a given miRNA in vitro. Accordingly, we tested these non-        We hypothesized that exosomes need two properties to be suppressive: the Ag specificity due 315 to surface FLC and delivery of miRNA-150. Accordingly, we tested the ability of non-suppressive,

316
Ag-specific B1a cell-derived exosomes from optimally immunized mice, that were non-suppressive, 317 likely because they contained no miRNA-150, to be rendered suppressive by similar association with 318 miRNA-150 (Protocol, Figure 5A). For this experiment, B1a cell-derived exosomes were harvested 319 from culture supernatants derived from lymph node and spleen cells of mice simply immunized ID 320 with OVA. They were harvested at only 2 days after ID immunization when B1a cells produce Ag-321 specific IgM antibodies and Ab FLC [9,11]. The B1a cell exosomes in this supernatant were 322 hypothesized to likely be OVA Ag-specific due to membrane anti-OVA BCR, and/or a surface coating 323 with anti-OVA Ab FLC, as shown in Figure 5B [8]. We determined that these immune B1a cell-derived

329
As expected, these exosomes derived from optimally immunized, but non-tolerized donors were 330 non-suppressive, likely because they did not naturally contain endogenously generated miRNA-150 331 because they were not from tolerized mice ( Figure 5D, Group C vs. B). In contrast, when these B1a 332 cell-derived, Ag-specific exosomes were in vitro associated with miRNA-150, they became 333 suppressive ( Figure 5D, Group D vs. C).

334
In summary, these data of the OVA protein DTH system confirmed that Ag-specific, suppressive

359
Overall, these studies confirmed the ability of miRNA-150 to associate with OVA Ag-specific 360 immune B1a cell-derived exosomes, as we found previously with these type of exosomes also from 361 early immune, but hapten-Ag-specific B1a cells [10]. The limiting dose here was much greater than 362 in the hapten system, and biotinylating the 5' end of miRNA-150 prevented association or function 363 in targeted cells.

403
where each was rendered into its Ab heavy and Ab FLC. We compared the ability of these constituent 404 chains to whole IgG in dose response binding to native OVA Ag, adhering to wells of enzyme-linked 405 immunosorbent assay (ELISA) plates and serially diluted from 2.5 to 0.078 µg per ml ( Figure 6A).

406
Over the dilution range employed, binding to native OVA adsorbed to plastic ELISA wells, 407 occurred with the whole IgGs as well as the separate Ab heavy and light chains. The anti-OVA mAb 408 were numbered #1, 2, 3 and 4, and a rank order of binding was determined. Binding of the isolated 409 heavy and light chains to native OVA Ag was in the same rank order for a given anti-OVA whole 410 monoclonal IgG, and differed in strength of OVA-binding comparing the four mAb ( Figure 6A).

411
Binding was best with whole IgG and next the heavy chains and least the FLC ( Figure 6A).

412
Another related ELISA experiment showed that the binding to native OVA by Ab FLC could be

417
In another ELISA assay we have compared the four mAb IgG-derived FLC for their binding to 418 native OVA and OVA tryptic peptides. FLC #1, 2 and 4 bound with a similar strength to both OVA 419 preparations, while binding of FLC #3 to OVA tryptic peptides was weaker than to native OVA 420 ( Figure 6C). Furthermore, all assayed FLC were shown to bind to OVA 323-339 antigenic determinant 421 but, however, with a very low efficacy ( Figure 6D).

422
As mention above, antigen-activated B1a cells provide FLC for coating of Ts cell-derived 423 exosomes ( Figure 1C). Thus, we speculated that these B1a cell-derived FLC can be found in serum 424 collected from mice immunized with OVA 2 days earlier. Prior studies in the CS system and in a 425 model of early resistance to pneumococcal pneumonia indicated that such sera contain polyclonal anti-hapten IgM Ab, and also Ab FLC with diverse Ag-specificities due to germ line V-region mutated 427 DNA sequences in a subpopulation of these ordinarily germ line expressing early B1a cells [14][15][16].

428
Thus, to evaluate this assumption, latex beads coated with native OVA antigen were incubated with 429 OVA immune sera, washed, stained against mouse kappa light chains, and subjected to flow 430 cytometric analysis, that revealed the presence of Ab FLC, which bound to OVA-coated beads, in 431 tested sera ( Figure 6E). As expected, the amount of detected anti-OVA FLC was much greater in sera 432 of OVA-immunized mice, when compared to naive mouse sera ( Figure 6E).  ]. Thus, we tested the relative ability of the four 455 mAb-derived FLC to restore suppressive action of exosomes, when used to coat exosomes derived 456 by Ts cells from OVA Ag-tolerized JH neg/neg mice. We found that the JH neg/neg mouse Ts cell exosomes 457 coated with FLC # 2, 3 and 4 mediated significant suppression of adoptively transferred DTH effector 458 cells ( Figure 7B). Simultaneously, we performed an ELISA-based assay to compare the binding of 459 JH neg/neg mouse Ts cell-derived exosomes coated with four mAb-derived FLC to native OVA and OVA 460 tryptic peptides (Protocol, Figure 7C, left), which was found similar in the case of FLC #1, 2 and 4 461 when tested as freely dispersed in assay buffer ( Figure 6C). This assay showed the significantly 462 greater amount of JH neg/neg mouse Ts cell exosomes that bound to OVA tryptic peptides in comparison 463 with binding to native OVA, when these exosomes were coated with FLC #1, 2 and 4, but not with

471
The yielded column eluates containing FLC that had strongly bound to OVA tryptic peptides, were 472 then used to coat JH neg/neg mouse Ts cell-derived exosomes that have been incubated with DTH 473 effector cells prior to their adoptive transfer. We observed that coating of JH neg/neg mouse Ts cell 474 exosomes with OVA tryptic peptide-binding FLC #2 and 3, and OVA-immune serum FLC, led to 475 significant suppression of DTH response ( Figure 7D). This confirmed that the ability of FLC to bind 476 the Ag indeed determines its suppressive activity when coating the Ts cell exosomes.

477
It has been previously shown that coating of exosomes from Ts cells generated by Ag induced

584
In the current study, OVA Ag-induced tolerance was also found to be mediated by CD3 + CD8 + 585 cells producing Ag-specific, suppressive exosomes, similarly coated with Ab FLC and particularly 586 transferring miRNA-150 cargo. Further, the OVA-specific suppressive exosomes similarly targeted 587 the APC companions of the finally suppressed OVA-specific DTH effector T cells ( Figure 2B).

588
Other EV-mediated mechanisms of immune suppression have already been described and 589 claimed to have Ag-specific activity. However, this has often been tested only in one-way Ag-activity

596
In great contrast, our prior studies on CS suppression definitively established the strict hapten 597 Ag-specificity of exosomes by dual reciprocal testing [4,8,10]. In the current study, we have also 598 reciprocally tested the Ag-specificity of exosome-mediated tolerance to a pair of unrelated protein 599 Ag; i.e. OVA and KLH. Accordingly, OVA-specific suppressive exosomes strongly inhibited effector 600 DTH responses to OVA protein Ag in actively immunized mice. Reciprocally, KLH-specific 601 suppressive exosomes failed to suppress OVA-induced DTH (Figures 3D and 3E). This proved that 602 protein-Ag-specific suppressive exosomes acted in a strictly Ag-specific manner.

616
we employed exosomes derived from B1a cells generated at only two days after ID immunization alone with OVA [11,30]. Although these B1a cell-derived exosomes were OVA Ag-specific due to the 618 surface expression of OVA Ab FLC and/or anti-OVA BCR, they were not suppressive, since they 619 lacked a miRNA-150 cargo, because the donors were not Ag-tolerized. Thus, of crucial significance, 620 as in CS suppression [10], we found that in vitro association of these B1a cell-derived, OVA Ag-621 specific exosomes with miRNA-150 alone rendered them suppressive of the OVA-specific effector T 622 cells via the APC. Further, these miRNA-150-associated, anti-OVA B1a cell-derived exosomes were 623 strongly suppressive over subsequent days, when injected systemically at the 24 hour peak of the 624 OVA-specific DTH response ( Figure 5F), and this was comparable to suppression mediated by the 625 OVA-specific Ts cell-derived exosomes (Figures 3D and 3F) [10]. In summary, we hold that achieving

716
Acquisition of activation characteristics is true for exosomes of mice undergoing strong and 717 prolonged exposure to high doses of Ag in tolerogenesis, or even mere two day ID immunization with proteins without an adjuvant, and to exosomes from various Ag high dose-tolerized Ab-719 deficient or miRNA KO mice as well.

720
The association of the mAb FLC was determined by ability to bind the surface of "activated", 721 but not normal, exosomes. This was verified by cytometric visualization of FLC on the exosome 722 surface (Figures 1C and 5B) [4,8]. Such surface and other activations seemed to account for acquisition 723 of both, the ability to bind specific Ag via surface Ab FLC, that enabled mediation of Ag-specific 724 suppressive function, and also to associate and then transfer selected miRNA for alteration of 725 intracellular functional effects.

726
This Ag-specificity was demonstrated by OVA Ag-affinity column separation of the total 727 suppressive exosome population that yielded two subpopulations (Figure 2A) [4]. These were 728 specifically Ag-binding, Ab-coated, therefore considered "activated", classical small exosomes that 729 had the suppressive functional activity vs. Ag-non-binding and functionally inactive exosomes 730 ( Figure 2A). This suggested that, in this case, when employing Ag-affinity column separation, only 731 around 10% of the total exosomes from immunized or tolerized animals is actually "activated" and

749
Further, study of cellular uptake of exosomes showed that interactions with targeted cells can be 750 modified by changing the lipid composition of their membranes [41,42].

751
Concerning membrane lipids and Ab FLC binding, our ideas also came from studies showing 752 that binding of immunoglobulin FLC to human mononuclear cell subsets depends on surface lipids 753 [43][44][45]. These ideas were sustained by our preliminary unpublished studies examining lipid binding

763
Studies are underway to similarly examine miRNA interactions with activated vs. normal 764 exosomes. The amounts of miRNA involved in association with "activated" OVA suppressive 765 exosomes are likely minute, as we showed before in CS [10], and here in DTH ( Figure 5E). It was 766 hoped that biotinylation of miRNA-150 would be helpful for such analysis, since it could aid 767 detection of the binding of miRNA to the exosome surface or associated within the "activated" 768 exosomes by pull down techniques [46]. Unfortunately, this biotinylation caused inactivation of

771
Lipid changes of "activated" exosome membranes may account for their other exceptional 772 properties, like our unique finding that oral systemic suppressive therapy with Ag-specific exosomes 773 can affect a strong local immune reaction at a specific cutaneous anatomical site. This may relate to 774 the physiologically transferred functions of numerous exosomes particularly contained in mothers' 775 milk and delivered to neonates [48][49][50]. The special property of these exosomes is resistance to harsh 776 conditions in the neonatal stomach, like very low pH plus digestive enzymes [51,52], that also could 777 be properties of "activated" exosomes in the immunological systems we described [53]. RNA cargo 778 of milk exosomes may transfer epigenetic information to neonates after intestinal absorption [54,55], 779 perhaps via intestinal epithelial cell endocytosis [56,57], benefitting the neonatal intestine [58,59], and 780 allowing postulated subsequent systemic transfer [60], that may influence neonatal immunity 781 [48,49,[61][62][63][64], bone formation [65,66], and development of the nervous and endocrine systems in the

786
Of further great relevance, there are numerous claims that among miRNAs taken orally in foods, 787 in some cases these mixtures contain exosomes that also can pass stomach digestion for uptake 788 systemically to affect host processes [68][69][70]. Again, there is controversy with some compelling data 789 against this concept [67,[71][72][73][74]. Note, however, that these are complex issues with many technical

826
To investigate this further, we made a tryptic digested peptide mixture from native OVA and 827 tested its ability to block suppression mediated by these exosomes, using different protocols. In the

925
In some instances prior to culture, mouse lymph node and spleen cells from Ag-tolerized donors, 926 were depleted of either CD3 + or CD8 + cells by incubation with, respectively, anti-CD3 or anti-CD8 927 monoclonal IgG antibodies and rabbit complement for 60 minutes in 37 o C water-bath. Afterwards, 928 dead cells were removed by discontinuous gradient centrifugation on Ficoll (1.077g/mL, GE 929 Healthcare).

930
In some experiments, actively tolerized mice were ear challenged on day 11, and subsequent ear 931 swelling responses were measured as described below. In some instances, tolerized and actively

953
The average ear swelling was expressed as the delta ± standard error (SE), after subtraction of the 954 negative control value. The two-tailed Student t test or one-way Analysis of Variance (ANOVA) with 955 post hoc RIR Tukey test were used to assess the significance of differences between groups, with p 956 values of less than 0.05 taken as the minimum level of statistical significance.

957
Exosome counts were estimated by Nanoparticle Tracking Analysis (Nanosite). The dose used 958 for treatment of actively immunized mice was considered physiological, since this is the average 959 number of exosomes per ml of blood in normal unmanipulated mice.

971
OVA DTH-effector cells were obtained from mixed spleen and lymph node cells of mice actively 972 sensitized by ID multiple injections of a total of 100 µg OVA in plain 0.9% NaCl (as described above) and harvested at day 4 after immunization [4,11]. OVA Ag-specific Ts-cell-derived, suppressive 974 exosomes were used at the same dose of 1×10 10 EVs for in vitro pulsing of a mixture of DTH-effector 975 T cells and APC, at 7×10 7 total lymphoid cells to be transferred per eventual recipient. Then, the 976 exosome pulsed DTH effector cells were adoptively transferred into naive recipients, in which 24 977 hours later DTH ear swelling responses were elicited by ID injection of 10 µl of OVA 0.9% NaCl 978 solution (0.5 mg/ml, thus 5 µg per ear). Ears were measured for thickness with an engineer's 979 micrometer (Mitutoyo, Japan), at 24 hours after challenge. Results were statistically analyzed as

991
In other experiments, exosomes from Ag-tolerized miRNA-150 neg/neg mice were treated in a 37°C

1024
The resulting pellet containing enriched Ag-specific B1a cell-derived exosomes was  (S1A) Protocol for ID immunization and eliciting ear swelling responses after inducing OVA Ag DTH. CBA male mice were immunized ID twice with OVA in saline injected into four different sites of the abdomen in 50µL each, for a total dose of 100µg OVA. Four days later they were skin challenged ID in the ears with 5µg OVA and the subsequent ear swelling was measured with an engineer's micrometer at 2 and 24 hours, and, where indicated, daily up to 120 hours after challenge, by a well experienced reader, unaware of their experimental status. The data were recorded by an assistant who randomly selected mice for measurements and administered ether inhalation anesthesia for the ear readings. All readings were done in duplicate for each of both ears to obtain the single mean ear swelling determination for that mouse, at that single time point. Mice were kept fed and watered in a bacteriology hood in the laboratory for the duration of the experiment to protect them from infection and from the biological changes that might happen during transit to and from the animal quarters far away.