Lactobacillus johnsonii was grown in de Man, Rogosa, Sharpe (MRS) media as previously described (8). NVs and membrane isolation were performed from L. johnsonii cultures grown in exosome-free MRS media (15). Isolated NVs were immediately quantified using Nanosight NS300 (Malvern instruments Ltd, Malvern, UK) and aliquots were stored at -80°C until use.

The pancreatic cell line βlox5 was obtained from Dr. Fred Levine (Sanford Children’s Health Research Center, La Jolla, CA) (18). Caco-2, Jurkat, and THP-1 cell lines were obtained from ATCC (Gaithersburg, MD, USA). βlox5 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) medium supplemented with 10% heat inactivated FBS (Sigma-Aldrich, Saint Louis, MO), 1% penicillin and streptomycin solution containing 10,000 units of penicillin and 10 mg of streptomycin/ml (Sigma-Aldrich, Saint Louis, MO), 0.2% of BSA, HEPES (15 mM), pH=7, and 5 ml of MEM non-essential amino acids (Sigma-Aldrich, Saint Louis, MO). Caco-2 cells were cultured in Eagle’s minimum essential medium (EMEM) supplemented with 15% heat inactivated FBS and 2% of penicillin and streptomycin solution. Cells were cultured in RPMI 1640 medium supplemented with 10,000 units of penicillin, 10 mg of streptomycin, and 10% or 20% heat inactivated FBS (for Jurkat or THP-1 cells, respectively). Where indicated, cytokines were added at TNFα (1,000 U/mL), IFNγ (750 U/mL) and IL-1β (75 U/mL). For the THP-1/βlox5 co-cultures, THP-1 cells were seeded in the inserts at 5 x 10⁵ cells/well and activated by adding phorbol 12-myristate-13-acetate (PMA) at 100 nM, for 48 h, at 37°C. Media was then removed, and inserts were transferred to a six well plate containing 6 x 10⁵ βlox5 cells. Vehicle control (10 µL of media) or NVs at 10¹⁰ particles/mL were added to the top chamber containing the THP-1 cells at 37°C. After 5 h, cells from both chambers were harvested and used for qRT-PCR and Western blot analysis.

Monocyte THP-1 cells were co-cultured with βlox5 cells using 4 µm pore size cell culture inserts (Thermo Scientific, Waltham, MA). THP-1 cells were seeded in the inserts at 5 x 10⁵ cells/well and activated by adding phorbol 12-myristate-13-acetate (PMA) at 100 nM, for 48 h, at 37°C in 5% CO₂ incubator. Media was then removed and inserts were transferred to a 6 well plate containing 6 x 10⁵ βlox5 cells. Vehicle control (10 µl of complete media) or NV at 10¹⁰ particles/ml were added to the top chamber containing the THP-1 cells at 37°C in 5% CO₂ incubator. After 5 h, cells from both chambers were harvested and individually used for qRT-PCR and Western blot analysis.

RNA was isolated from cell lines and human pancreatic islets via Trizol. Briefly, 300 µl of Trizol (Zymo Research, Irvine, CA) was added to each well. After homogenizing, the Trizol reagent was transferred to an Eppendorf tube containing 300 ul of 95% ethanol and thoroughly mixed. The mixture was transferred to a spin column (Zymo Research, Irvine, CA) and RNA isolation was performed following the manufacturer’s protocol. DNA was removed by treatment with DNase (QIAGEN, Germantown, MD) according the manufacturer’s protocol. RNA quality was monitored on 1% agarose gels, and RNA quantification was performed using Thermo Scientific Nanodrop One Microvolume UV-vis spectrophotometer (Thermo Fisher Scientific, Grand Island, NY). qRT-PCR was performed as described (19). Primer sequences used to determine relative transcript abundance are listed in Table S1 .

After extraction of total RNA, library construction and sequencing were performed by Novogene, (Novogene Co., Davis, CA, USA). 1 µg of RNA from each sample was used for the library preparation. Sequencing libraries were generated using NEBNext^® Ultra TM RNA Library Prep Kit for Illumina^® (NEB, USA) following manufacturer’s recommendations. mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5X). First strand cDNA was synthesized using random hexamer primers and M-MuLV Reverse Transcriptase (RNase H-). Second strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase H. In order to select cDNA fragments of preferentially 150~200 bp in length, the library fragments were purified with AMPure XP system (Beckman Coulter, Beverly, USA). Then 3 µl USER Enzyme (NEB, USA) was used with size-selected, adaptor ligated cDNA at 37°C, for 15 min, followed by 5 min at 95°C before PCR. PCR was performed with Phusion High-Fidelity DNA polymerase (NEB, USA), Universal PCR primers and Index (X) Primer. PCR products were purified (AMPure XP system) and library quality was assessed on the Agilent Bioanalyzer 2100 system. The clustering of index-coded samples was performed on a cBot Cluster Generation System using PE Cluster Kit cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina platform and paired-end reads were generated. Raw data (raw reads) of FASTQ format were firstly processed through fastp (20) to generate clean data (clean reads). Paired-end clean reads were aligned to the reference genome using the Spliced Transcripts Alignment to a Reference (STAR) software. Novogene Co., Ltd Quantification FeatureCounts (21) was used to count the read numbers mapped for each gene, and the RPKM of each gene was calculated based on the length of the gene and read counts mapped to this gene. Differential expression analysis between two conditions/groups was performed using DESeq2 R package (17). The resulting p values were adjusted using the Benjamini-Hochberg method for controlling the False Discovery Rate (FDR). Genes with an adjusted p value < 0.05 found by DESeq2 were assigned as differentially expressed. GO term enrichment analysis was used to analyze the biological significance of the differently expressed genes (DEGs) and graphs were generated by R package “GOplot” (22). One-way ANOVA was used to identify genes with significant differences between all groups. Tukey honestly significant difference (HSD) test was used to determine the significance between two groups.

Approximately 1x10⁶ βlox5 cells were incubated with NVs at a concentration of 10⁸ or 10¹⁰ particles (1:100 or 1:10000 cells to nanovesicles, respectively), and cytokines [TNFα (1 U/ml), IFNγ (1 U/ml), and IL-1β (0.056 U/ml)], followed by incubation at 37°C, under 5% CO_2, for 6 h. Then, cells attached and in the supernatant were combined by centrifugation at 700 x g for 5 min and washed once with 2 ml of cold βlox5 complete media. Cells were labeled according to the specifications of the BD Annexin V FITC Assay using the BD FACSVerse™ System Kit. Cells were counted using the Accuri C6 flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). The analysis was performed using FCS Express Version 7 software (DeNovo software, Pasadena, CA, USA), and the gating strategy is shown in Supplementary Figure 3 .

Human pancreatic islets were purchased from ProdoLabs (Aliso Viejo, CA). Islets were cultured in CMRL 1066 medium, supplemented with 10% heat inactivated FBS (Sigma-Aldrich, Saint Louis, MO), 5,000 units of penicillin and 5 mg of streptomycin (Sigma-Aldrich, Saint Louis, MO) at 37°C, in 5% CO2 incubator. After a transport recovery period of 48 h, islets were incubated in starvation media containing a 50:50 mixture of DMEM (no glucose) and Ham’s media, 7.1 mM sodium bicarbonate, 750 µM calcium chloride dehydrated, 2.5% BSA, 1% penicillin/streptomycin solution and 3 mM D-glucose overnight at 37°C, under 5% CO₂. After 12 h, the starvation media was replaced by KREBB’s media containing either 0.5, 3, 10, or 15 mM D-glucose, and incubated at 37°C, under 5% CO₂. Samples from the supernatant were collected after 1, 2, and 3 h of incubation, and insulin concentration was determined by ELISA (Mercodia, Winston Salem, NC). The insulin concentration obtained was normalized to the DNA concentration of the islets. DNA was extracted using DNeasy blood and tissue kit (QIAGEN, Germantown, MD) according to the manufacturer’s protocol. Insulin secretion was expressed as mU Insulin/L/ng of DNA. For qRT-PCR analysis, 250 islets equivalent (IEQ) were cultured with 6 x 10⁹ NVs for 5 h in CMRL 1066 culturing medium, supplemented with 10% heat inactivated FBS, 1% penicillin and streptomycin solution (Sigma-Aldrich, Saint Louis, MO) at 37°C in 5% CO2 incubator. RNA was extracted as described above. Each experiment was conducted in triplicates with islets purified from two non-diabetic donors. Demographics are provided in Table S2 . Details of batch glucose-stimulated insulin secretion (GSIS) are provided in the Supplementary Materials section. Each experiment was conducted in triplicate with islets purified from donors without a diagnosis of diabetes.

Western blot assays were used to test the presence of STAT3, P-S727-STAT3 and P-Y705-STAT3. β actin quantification was used as the loading control. All antibodies were purchased from Abcam (Cambridge, MA, USA). Total proteins were extracted from βlox5 cells using Radio Immunoprecipitation Assay Buffer (RIPA) containing 150 mM NaCl, 50 mM Tris (pH 8), 1% Triton X-100, and 0.1% sodium dodecyl sulfate (SDS), with Halt™ protease inhibitor cocktail (Thermo Fisher, Waltham, MA, USA). The cell homogenates were centrifuged at 12,000 g for 10 min, at 4°C and the protein concentration was measured using Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, USA). Fifteen micrograms of proteins were separated using 12.5% (v/v) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio Rad, Hercules, CA, USA) using a semi dry transfer system. The membranes were blocked with 5% non-fat milk in 0.1% Tween 20, in saline Tris buffer pH=7.4 (TBS-T) for 1 h at RT. PVDF membranes were incubated with anti-STAT3 and anti-phospho S727 STAT3 at 1/1,000 dilution factor, anti-phospho Y705 STAT3 at 1/500 dilution factor, and anti-actin at 1/10,000 primary antibodies (Abcam, Cambridge, MA, USA) over night at 4°C. Membranes were washed with tris buffered saline containing 1% tween 20 (TBS-T), three times, for 5 minutes at RT after blocking and incubation with primary antibodies. Subsequently, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody at 1/10,000 for 1 h at RT and washed three times with TBS-T. The enhanced chemoluminescence method (ProSignal^® ECL Reagent; Genesee Scientific, San Diego, CA, USA) was used for visualization via the automatic imager FluorChem R (Proteinsimple, San Jose, CA, USA). The relative intensity of the bands visualized in the membranes were quantified with ImageJ software (Free Java software provided by the National Institutes of Health, Bethesda, Maryland, USA). In all cases, β-actin band intensity was used to normalize the bands of each sample. All experiments were done in biological triplicate.

To visualize changes in the cellular localization of AHR, 3x10⁵ βlox5 cells were cultured on coverslips previously coated with poly D-lysine (Sigma-Aldrich, Saint Louis, MO) for 3 h. AHR translocation assays were initiated by addition of NVs at 10¹⁰ particles/ml or 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) at 100 nM for 0 and 45 minutes, at 37°C, under 5% CO₂. Following incubation, the cells were washed three times with ice-cold PBS (pH 7.4), fixed with 4% paraformaldehyde for 10 min, and washed three times with ice cold PBS. Cells were permeabilized with PBS solution containing 0.1% Triton X-100 for 15 min, at 25°C and rinsed three times with PBS, 5 minutes each wash. Cells were incubated with rabbit anti-AHR antibody tagged with R-phycoerythrin (PE)-texas red (Abcam, Cambridge, MA) for 1h at 25°C and subsequently washed three times with PBS. Fluorescence was acquired using a confocal fluorescence microscope (Carl Zeiss, Berlin, Germany) and 3D reconstructions generated by acquiring Z-stacks of 18 images.

Supernatants from cell/islet cultures were used to determine protein levels of IL10, IL1β, TNFα (BD biosciences, Franklin Lakes, NJ, USA) and insulin (Mercodia, Winston Salem, NC, USA) according to the manufacturer’s protocols.

R studio (RStudio Team, Boston, MA) as well as GraphPad Prism 5.01 software (GraphPad Software, La Jolla, CA, United States) were used for data analysis and visualization. Statistical tests were performed using one-way analysis of variance (ANOVA) to evaluate the effects of treatments on cell lines or human islets. Significance of model terms and treatment comparisons were considered significant at a level of α=0.05, and degrees of freedom were adjusted using the Kenward–Rogers correction. Finally, post hoc multiple comparisons were performed with least significant differences.

The RNA sequencing data generated in this study are deposited in the NCBI Sequence Read Archive (SRA) under BioProject PRJNA804404.

Before exhibiting its typical clinical symptoms, type 1 diabetes follows a slow-progressing autoimmune process, comprised of a cascade of immunological steps, culminating with apoptosis of beta cells. It has been shown that secretion of IL-1β, TNFα and IFNγ by local antigen presenting cells (APCs) and infiltrating T cells, are responsible for induction of transcription factors and apoptosis-related proteins, such as NF-κB, caspase-3, p38 and c-Jun N-terminal kinase (JNK) (23). To investigate the effect of L. johnsonii on inflammation-induced apoptosis, human βlox5 cells were incubated in presence or absence of NVs at 10⁸ or 10¹⁰ particles/ml (NV8 or NV10, respectively) for 2 h, at 37°C. Addition of 10⁸ or 10¹⁰ NV/mL did impact cell viability compared to the vehicle control (4.3%, 3.5%, and 4.5%, respectively). The addition of cytokines significantly (p<0.05) increased apoptosis of βlox5 cells. Addition of NV with cytokines prevented apoptosis ( Figures 1A, B ). Similarly, a reduction in pre-apoptotic cells was observed (p=0.053). The reduction in apoptotic cells with NV10 also resulted in an increase in live cells, when compared to the control group ( Figure 1C ). In contrast, the addition of NV8 did not affect the number of apoptotic, pre-apoptotic or live cells, suggesting the ability for NVs to protects against apoptosis in the pancreatic beta cell line is dose dependent ( Figures 1A–C ).

In human islets, apoptosis may be triggered by intrinsic and extrinsic pathways, as well as through perforin/granzyme pathway (24). To evaluate the effect of NVs on apoptosis pathways, expression of the pro-apoptotic genes BAX, BIM, FADD and the anti-apoptotic genes BCL-2 and BCL-XL were quantified in βlox5 cells following the ame treatment protocol as immediately above. Expression of these genes was not significantly modified by the addition of NVs or the pool of cytokines (data not shown).

To identify the mechanism by which NVs reduce apoptosis in βlox5 cells, we performed RNAseq analysis of βlox5 cells treated with NVs from L. johnsonii at 10⁸ and 10¹⁰ particles/ml (NV8 and NV10, respectively), in presence or absence of the pool of cytokines. We have recently reported that different proteins and lipids are enriched in NVs when compared to the membranes (MEMB) isolated from L. johnsonii (15). To compare the differential host response to NVs or MEMB components, βlox5 cells were challenged with 1 or 5 µg/mL MEMB (MEMB1 and MEMB5, respectively) in the presence or absence of the cytokine combination. MEMB1 and MEMB5 are equivalent to NV concentrations NV8 and NV10, respectively.

The gene expression profile was mostly driven by the addition of cytokines ( Figure 1D ). All differentially expressed genes were classified into functional categories using gene ontology enrichment analyses (GO terms) ( Supplementary Figures 1A–D ).

In absence of an inflammatory environment, the addition of NV8, as well as MEMB1, did not result in significant changes in gene expression. Treatments with NV10 or MEMB5 did not repress the expression of any gene. Interestingly, 21 genes were up regulated by addition of NV10, while MEMB5 up regulated the expression of two genes ( Table 1 ). The expression of Cytochrome P450 Family 1 Subfamily B Member 1 (CYP1B1) was the only significantly upregulated transcript that overlapped between the N10 and MRMB5. The GO categories enriched by the addition of NV10 were defense response to virus and cellular response to type 1 interferons ( Table 1 ; Supplementary Figure 1 ).

Addition of the pool of cytokines induced a strong shift in the expression profile of βlox5 cells. It was found that 1110 genes were significantly upregulated while 419 were downregulated. As expected, GO pathway analysis indicated that most of the upregulated genes are classified in categories related to the regulation of cytokine production and inflammatory responses ( Supplementary Figure 1B ; Supplementary Table 3 ). Under these inflammatory conditions, co-treatment with NV10 (CYT+NV10) was found to upregulate the expression of an additional 119 genes and downregulate the expression of 20 genes. Interestingly, many of the upregulated genes belong to the initial response to viral infection categories, similar to the results observed in absence of the pool of cytokines ( Figure 1A ; Table 1 ). In contrast, the addition of the cytokine pool to MEM5 treated cells resulted in the enrichment of different GO term categories, where upregulated genes belonged to categories related to regulation of vascular endothelial growth, response to toxic substances, and regulation of hormone levels, while the downregulated genes belong to the I band, supramolecular complex/polymer and fiber categories ( Supplementary Figure 1 ). The differential effect observed on βlox5 cells following treatment with different cellular components of L. johnsonii, indicate that NVs are able to generate a specialized host response when compared to cell membranes from the same bacterium.

In absence of inflammatory conditions, the addition of NV10 significantly induced the 2’,5’-oligoadenylate synthetase (OAS) and the AHR pathways ( Table 1 ). The OAS pathway is an IFN-stimulated antiviral response activated by viral or bacterial RNA (25). In humans, the OAS family is composed of three enzymatically active enzymes, OAS1, OAS2 and OAS3, all of which were significantly induced in presence of NV10, but not by NV8, MEMB1, or MEMB5. As expected, all three OAS genes were significantly induced upon addition of the pool of cytokines (where INFγ is a component). Interestingly, the expression of those genes was further induced by the addition of NV10 from L. johnsonii, with the exception of OAS2 ( Table 1 ). Similar results were observed on the enzymatically inactive OASL, which is induced by RIG-1 and MAVS oligomerization, and is crucial in the stabilization of the OAS complex ( Table 1 ). Taken together, these findings suggest that nucleic acids present in NVs are sensed by the βlox5 cells. Other genes involved in the sensing and response to nucleic acids, such as IFI44L, MX1, MX2 and DDX60, followed a similar pattern of stimulation in presence of NV10 ( Table 1 ).

Genes in the AHR pathway were among the most upregulated genes induced by NV10. AHR is a ubiquitous transcription factor that is activated in response to environmental and metabolic cues. In the inactive form, AHR is located on the cytoplasm, associated in a multiprotein complex that includes the chaperone heat-shock protein 90 (HSP90), the co-chaperone p23, and the hepatitis B virus X-associated protein (XAP2). Upon binding to a ligand, conformational changes in AHR result in its translocation into the nucleus, where it changes interaction partners by binding to the AHR nuclear translocator (ARNT). The AHR/ARNT complex binds to specific sequences in the promoter thereby controlling the expression of several genes (26). Transcription of cytochrome P450 (CYP1) family members, including subfamily A member 1 (CYP1A1) and subfamily B member 1 (CYP1B1), is regulated by AHR activation. Our data shows that 10¹⁰ NV particles/ml from L. johnsonii significantly increased the expression of CYP1A1 and CYP1B1. The addition of MEM5 significantly increased the expression of CYP1B1 and AHRR (p<0.05), while the changes induced in CYP1A1 were more variable and did not reach statistical significance. Interestingly, addition of the pool of inflammatory cytokines failed to induce activation of the AHR pathway on βlox-5 cells, where higher expression levels were observed only in the presence of NV10 or MEMB5, highlighting the specificity of the results obtained with the bacterial components ( Figures 1E–H ).

Once AHR is active, it is negatively regulated by binding of AHRR or TIPARP (TCDD-induced poly-ADP-ribose isomerase). It was observed that addition of NV10 significantly increased the expression of both repressors in presence or absence of the pool of cytokines ( Figures 1G, H ).

Activation of the canonical AHR pathway observed in presence of NVs was evaluated following the nuclear translocation of AHR, via immunofluorescence. As a positive control, cells were also incubated with 100 nM of TCDD, a well-known AHR inducer (27). To confirm the localization of AHR inside the nucleus, we analyzed the colocalization of both stains, Texas red (staining the AHR) and DAPI (staining the nucleus). It was observed that βlox5 cells incubated with NV10 showed an increase in the AHR nuclear translocation after 45 minutes of incubation ( Figures 2A–F ). The AHR nuclear translocation observed with NV10 treatment was comparable to the translocation observed when cells were incubated with 100 nM of TCDD ( Figures 2G–L ). For both treatments, a significant increase in the colocalization levels was observed in βlox5 cells after 45 min of incubation ( Figures 2M–P ). The colocalization coefficient increased from 0.08 to 0.18 when cells were incubated with TCDD, whereas the colocalization coefficient increased from 0.02 to 0.11 when cells were incubated with NVs ( Figures 2M–P ).

While we observed effects from NVs in different pathways, we could not evaluate their activities on insulin secretion in βlox5 cells due to their inability to produce insulin. Expression of the gene encoding insulin (INS), and the key regulator of the transcription of insulin, the pancreatic and duodenal homeobox 1 (PDX1, also known as insulin promoter factor 1), are not expressed in βlox5 cells (28). To investigate the effect of NVs from L. johnsonii on insulin secretion by pancreatic beta cells, we evaluated the levels of insulin secreted by primary pancreatic islets isolated from human donors. Under starvation conditions (0.5 mM of D-glucose), there was no significant difference between the insulin levels from islets incubated with NVs when compared to the vehicle control group ( Figures 3A, B ). In presence of 3 mM D-glucose, insulin levels increased over time in both treatment groups, with more insulin being released when cells were incubated with NVs ( Figures 3A, B ). The highest concentration of insulin was observed when islets were cultured under stimulating glucose concentrations (10 or 15 mM), after 2 or 3 h in presence of NVs. These results indicate that NVs from L. johnsonii may be able to either increase glucose uptake in human pancreatic islets or stimulate insulin secretion over time.

The expression of genes involved in glucose uptake, as well as regulation of insulin synthesis and secretion, were analyzed on pancreatic islets. Islets treated with NV10 showed an 8-fold increase in the expression of glucose transporter Solute Carrier Family 2, Member 6 (SLC2A6), also known as glucose transporter 6 (GLUT6) ( Figure 3C ), and a 2-fold increase in the Sterol Regulatory Element Binding Transcription Factor 1 (SREBF1) ( Figure 3D ), while expression of SLC2A1, PDX1, and INS were not affected after 5 h of incubation (data not shown). Glucagon-like peptide 1 receptor (GLP1R) mRNA levels were elevated in islets incubated with NVs (p=0.06) ( Figure 3E ). Interestingly, the expression of protein kinase catalytic subunit α (PRKACA), which is part of the cAMP insulin signaling pathway, was significantly increased in islets treated with NVs ( Figure 3F ). To investigate the role of NVs in insulin secretion, via Ca²⁺ modulation, we evaluated the mRNA levels of calcium/calmodulin dependent protein kinase II gamma (CAMK2G), where NVs were found to induce a slight increase in CAMK2G expression, however, these values did not reach statistical significance ( Figure 3G ). Altogether, these results indicate that under high glucose stimulation, NVs can induce glucose uptake by the pancreatic islets, thereby increasing insulin secretion.

Previous studies have suggested glucose-dependent insulin secretion may be directly linked to reactive oxygen species (ROS) and oxidative stress in pancreatic beta cells (29). Beta cells are inherently susceptible to oxidative stress due to high endogenous production of ROS combined with low expression levels of antioxidative enzymes (30). Metastasis Associated 1 Family Member 2 (MTA2) and Stanniocalcin-2 (STC2) play a central role in protection against oxidative stress. Upon AHR activation, MTA2 is recruited to the STC2 promoter, resulting in histone acetylation and activation of STC2 transcription, where STC2 subsequently attenuates reticulum oxidative stress and stress-induced apoptosis (31). Therefore, we evaluated the effects of NVs from L. johnsonii on the expression levels of genes involved in antioxidant responses. In agreement with these reports, it was found that MTA2 and STC2 mRNA levels were significantly induced (p <0.05) following treatment with NV10 ( Figures 3H, I ). To evaluate the role of NVs on oxidative stress markers, the mRNA levels of SOD1 was also determined, where NVs were found to significantly repress the expression of SOD1 ( Figure 3J ). Taken together, these findings suggest that NVs from L. johnsonii may prevent intra-islet oxidative stress.

To evaluate whether activation of AHR by NVs is tissue specific, we tested their effects on different cell lines. THP-1 macrophages treated with NV10 showed a significant increase in the mRNA levels of CYP1A1, while the expression of CYP1B1 and AHRR was not affected by the addition of NVs ( Supplementary Figures 2A–C ). Similarly, in Caco-2 cells, the expression of CYP1A1 was significantly induced, while the expression of CYP1B1 and AHRR was not detected after 5 hours of incubation ( Supplementary Figure 2D ). In contrast, Jurkat T cells did not show significant changes in the expression of CYP1A1 and CYP1B1 after 5h of incubation with NVs ( Supplementary Figures 2E, F ). In summary, the results obtained suggest that NVs differentially affect host cell types either through specific cell receptors that facilitate entry to the host cell or through regulating expression levels of target host proteins.

It has been reported that AHR activation in murine macrophages results in the polarization towards the tolerogenic M2b phenotype, through the production of IL10 (32). To evaluate the impact of NV addition on THP-1 polarization, THP-1 cells were incubated in presence or absence of NV10 for 3 and 8 h, and subsequently used to determine the mRNA and protein levels of markers of M1 or M2 phenotypes (TNFα, IL-1β, and IL-10 as well as the mRNA levels of TLR2, TLR3, TLR4, TLR7, and TLR9). We observed that cells incubated with NV10 significantly increased the mRNA and protein levels of TNFα, IL-1β and IL10 over time ( Figures 4A–F ). These results are indicative of the polarization of macrophages towards the M2b tolerogenic phenotype. In agreement with these observations, the transcription levels of TLR2 and TLR7 were significantly increased after 3 h of NV treatment while the mRNA levels of TLR4 were not affected after 8 h ( Figures 4G–I ). These data suggest that NVs from L. johnsonii can induce a shift in the phenotype of macrophages towards an anti-inflammatory profile.

The mRNA expression of IL10 is controlled by multiple transcription factors including STAT3 and AHR (32). To determine the involvement of the STAT3 pathway in NV-induced IL10 expression, the activation of STAT3 was quantified by of S727 and Y705 phosphorylation (pS-STAT3 and pY-STAT3, respectively) as well as total STAT3 (T-STAT3) in THP-1 cells exposed to NV10. It was found that addition of NV10 resulted in a significant increase in the T-STAT3/actin ratio as well as pY-STAT3/T-STAT3, while pS-STAT3/T-STAT3 was not affected ( Figures 5A–D ). These results indicate that the addition of NVs promote the transcriptional activity of STAT3 as Y705 phosphorylation is required for the role of STAT3 as a transcription factor and binding to specific promoters in the nucleus (33). It was also observed that the increased pY-STAT3 positively correlated with a significant increase in the mRNA expression of AHR ( Figures 5E, F ). These results indicate that the induced expression of the CYP genes is consequence of an indirect effect of NVs on the AHR pathway, and that STAT3 is the primary target of NVs in macrophages. However, the observed increase in expression of IL10 may be explained by the combined function of STAT3 as well as AHR, since binding sites for both transcription factors have been reported to activate the IL10 promoter (32).

Macrophages have an important role in the pancreatic tissue, sensing danger associated molecular patterns (DAMPs) released by injured tissue, resulting in the activation of an inflammatory response. Alternatively, they can act as mediators of anti-inflammatory responses through the secretion of IL10 (34). Therefore, we determined the effects of the cytokines secreted by NV-treated THP-1 cells on βlox5 cells. Inserts containing 6x10⁵ PMA-activated THP-1 cells were transferred to 6-well plates containing 6x10⁵ βlox5 cells/well. Stimulation was performed by addition of NVs at 10¹⁰ particles/mL to the top chamber of the inserts (empty inserts or PMA-activated THP-1 cells), and the expression of AHR activation markers was quantified after 5 h. It was found that treatment of THP-1 cells with NVs did not significantly affect the expression of CYP1A1, CYP1B1 and AHRR in βlox5 ( Figures 6A–C ). In contrast, the expression of IL10 mRNA was significantly increased in βlox5 cells that were co-cultured with THP-1 ( Figure 6D ). These results are significant since our transcriptome results showed that the IL10 gene is not expressed in βlox5 cells in absence of THP-1 cells.

These results suggest that different mechanisms are involved in NV signaling depending on the cell lines tested and/or cells in their environment within a tissue. We hypothesize that THP-1-secreted IL10 acts as the signaling molecule that stimulates the STAT3 pathway in βlox5 cells resulting in further IL10 expression. First, the levels of T-STAT3 as well as S-STAT3 and Y-STAT3 were evaluated in βlox5 cells after 5 h of co-culture with THP-1, in presence or absence of NV10 ( Figure 6E ) as described above. The stimulation of βlox5 cells with NV10 in absence of THP-1 cells did not significantly affect the levels of T-, S- or Y-STAT3 ( Figures 6F–H ). Interestingly, the co-culture of THP-1 with βlox5 cells resulted in a significant increase in Y-STAT3 phosphorylation, while further stimulation was not observed upon addition of NVs to THP-1. The levels of S-STAT3 were not affected under any conditions tested.

To evaluate whether IL10 secretion by THP1 is involved in STAT3 activation in βlox5 cells, IL10 at 100 ng/ml was added to an empty insert, and anti-IL-10 at 1µg/ml was added to a co-culture containing NV-stimulated THP-1 and βlox5 cells, where the levels of Y-STAT3 were significantly increased in βlox5 cells in presence of IL10. In contrast, levels of Y-STAT3 were significantly decreased following neutralization of IL10 with monoclonal antibodies ( Figures 6E–H ).