EXPERIMENTAL MODEL AND SUBJECT DETAILS
Ethical Statement
Animal handling, maintenance and experimentation were conducted in accordance with the guidelines of the Helsinki Declaration and with the regulations stated by the European Legislation for the Protection of Animals Used for Scientific Purposes. The animal procedures were approved by the Committee on the Ethics of Research in Animal Experimentation of the UdL (Protocol #: CEEA 02-04/16).
Housing and Husbandry
Mice were bred and maintained in the rodent animal house of the UdL. Being kept under specific pathogen-free (SPF) conditions, animals were provided autoclaved food (Envigo, Cat#2018S) and water ad libitum and their light-dark cycle was controlled in a 12:12h format. Sacrificing by isoflurane inhalation anesthesia and/or cervical dislocation was performed when mice developed diabetes or at the end of the studies.
Mouse Models
NOD mouse strain was originally purchased as NOD/ShiLtJ from The Jackson Laboratory (Bar Harbor, ME, Cat#JAX:001976). NOD.RAG-2-/- mouse strain was obtained from Dr. P. Santamaria (University of Calgary, Alberta, Canada). The 116C-NOD mouse model,transgenic for a β-cell-autoreactive B lymphocyte, was generated on the NOD strain genetic background in our laboratory. As well, the 116C-NOD.RAG-2-/- mouse model, which harbors monoclonal transgenic β-cell-autoreactive B cells, was generated on the NOD.RAG-2-/- strain genetic background by our team.11 C57BL/6J mice were purchased from The Jackson Laboratory (Charles River, Europe, Cat#JAX:000664).
Experimental design of animal groups
NOD and 116C-NOD mice were obtained from 116C-NOD breeder pairs (116C-NODxNOD). Weaned females were distributed in cages depending on two experimental conditions: isolation and cohousing between non-transgenic and transgenic siblings. Mice were grouped in cages of isolation which only contained NOD or 116C-NOD females, and in cages of cohousing with a balanced number of NOD and 116C-NOD littermates.
A cage change experiment was undertaken to confirm the results of the isolation and cohousing conditions. NOD females were housed in cages previously occupied by NOD or 116C-NOD female donor counterparts. The cages containing the donors’ fecal pellets were renewed thrice weekly.
In parallel, NOD.RAG-2-/- and 116C-NOD.RAG-2-/- were obtained from 116C-NOD.RAG-2-/- breeder pairs (116C-NOD.RAG-2-/-xNOD.RAG-2-/-). In this case, weaned females were grouped in cages that only contained NOD.RAG-2-/- or 116C-NOD.RAG-2-/-.
C57BL/6J female mice were used as non-T1D-prone controls.
Specific groups of mice were subjected to different in vivo and in vitro assays and were maintained until the appropriate age for those (see the sections below).
Diabetes incidence assessment
Diabetes in mice from NOD and 116C-NOD (isolation, cohousing, and cage change groups), NOD.RAG-2-/- and 116C-NOD.RAG-2-/- strains was followed up weekly and for ten months by measuring glycosuria with Medi-Test Glucose urine test strips (Macherey-Nagel, Cat#93001). Animals were considered diabetic after two consecutive positive readings of values ≥ 50 mg/dL and by monitoring glycemia levels with Accu-Chek Performa Glucose blood test strips (Roche, Cat#06454011) after obtaining values ≥ 250 mg/dL.
Insulitis scoring
Pancreas from isolated and cohoused female NOD and 116C-NOD mice, aged six and 12 weeks, were embedded in Tissue Freezing Medium (Electron Microscopy Sciences, Cat#72592-C) and snap frozen in a ≤ -75ºC cooling bath of dry ice and isopentane (Sigma-Aldrich, Cat#M32631). Pancreatic cryosections of 8 µm were rapidly collected on standard glass slides and immediately fixed in cold 95% ethanol for 10 minutes, dried for another 10 minutes, and frozen at -80ºC until hematoxylin and eosin (H/E) staining. H/E staining consisted of the following steps: thawing in air at 4ºC for 5 minutes and in PBS at 4ºC for another 5 minutes, staining for 5 minutes in filtered hematoxylin solution (5.3 g/L 1-hydrate Gurr (VWR Chemicals, Cat#340374T), 70.24 g/L Al2(SO4)3 (VWR Chemicals, Cat#100103M), 300 mL/L glycerol and 0.5 g/L sodium iodate (Honeywell, Cat# 71702)), washing in tap water for 5 minutes, differentiation in 0.5-1% HCl 70% ethanol solution and 0.001% ammonia solution, washing in tap water for another 5 minutes, staining in eosin solution (10 g/L eosin Y Gurr (VWR Chemicals, Cat#341972Q) and 80% ethanol) for 5 minutes, brief washing in tap water, dehydration in 80%, 85%, 90%, 95%, and 100% ethanol solutions (1 minute in each), 3 minutes in absolute xylene, and finally mounting in DPX. Insulitis degree was determined by means of a blind analysis of 15-30 islets/mouse based on the following scoring criteria: score 0 (no cell infiltration in the islet); score 1 (peri-insulitis); score 2 (mononuclear cell infiltration in <25% of the islet), score 3 (mononuclear cell infiltration in 25-75% of the islet); and score 4 (mononuclear cell infiltration in >75% of the islet). The mean insulitis score of each pancreas was calculated via the equation: insulitis score=[(0xA)+(1xB)+(2xC)+(3xD)+(4xE)]/TNI]; where A, B, C, and D are the number of islets belonging to score 0, 1, 2, 3, and 4, respectively, and TNI is the total number of islets.
Lymphocyte stimulation
Under sterile conditions, spleens harvested from 12-week-old isolated and cohoused female NOD mice (four animals per group) were mechanically disrupted with glass slide frosted ends in HBSS (Ddbiolab, Cat#X0509-500) containing 1% heat-inactivated fetal bovine serum or hiFBS (Gibco, Cat#10270106) and converted into single-cell suspensions by passing splenocytes through 40 µm nylon filters. T and B lymphocytes were then separately purified via negative selection using isolation kits specific for each population: Mouse Pan T Cell Isolation Kit II (Miltenyi Biotec, Cat#130-095-130) and Mouse B Cell Isolation Kit (Miltenyi Biotec, Cat#130-090-862); as well as the AutoMACS Pro Separator magnetic cell sorter (Miltenyi Biotec, Cat#130-092-545), following manufacturer’s instructions. Yield and purity of T and B cells were assessed by staining CD3 and CD19 cell surface markers with the monoclonal antibodies FITC anti-CD3 (BD Pharmingen, Cat#561798) at 2 µg/mL and violetFluor 450 anti-CD19 (Tonbo Biosciences, Cat#75-0193-U100) at 0.8 µg/mL in PBS with 1% hiFBS at 4ºC for 20 minutes, and by using the flow cytometer FACSCanto II (BD Biosciences, Cat#640806). Lymphocyte purity was only accepted when values were greater than 90%.
Purified T and B lymphocytes were cultured in vitro with different stimuli at 37ºC and 5% CO2 in complete culture media (CCM) consisting of RPMI 1640 (Biowest, Cat#L0501-500) supplemented with 10% hiFBS, 2 mM L-glutamine (Corning, Cat#25-005-CI), 1 mM sodium pyruvate (Gibco, Cat#11360-070), 50 µM 2β-mercaptoethanol (Sigma-Aldrich, Cat#M6250-100ML), 100 U/mL benzylpenicillin sodium (Normon, Cat#602896.4), and 100 µg/mL streptomycin sulfate (Normon, Cat#624569.9).
T lymphocyte stimulation involved a three-day incubation of the purified T cells plated at 3x105 cells/well and cultured either: (1) alone; (2) with soluble anti-CD3 (sαCD3); (3) with well-coated anti-CD3 or fixed anti-CD3 (FαCD3); or (4) with sαCD3 in the presence of purified B cells, also plated at 3x105 cells/well (co-culture at a 1:1 ratio). Under the fourth condition, all possible combinations between T and B cells from isolated and cohoused NOD mice were made. When soluble, purified anti-CD3 monoclonal antibody (BD Pharmingen, Cat#553057) was added at 5 µg/mL. In the coated-plate condition, wells were incubated with 40 µL of 10 ug/mL anti-CD3 diluted in TBS buffer (pH=9.4) at 37ºC for 2 hours, left at 4ºC for ≤ 24h and washed thrice with PBS before use. Important note: The two types of anti-CD3 stimulation: sαCD3 and FαCD3 exert a different effect on T cells. SαCD3 induces a low activation of T cells when these have no other stimuli in the culture. When sαCD3 is present in the co-culture of T and B cells, it acts as a linker between both lymphocytes, allowing their interaction. FαCD3 exerts a potent activation of T cells since the well-bound antibody molecules can cross-link T cell receptors (TCR).
Cell culture was performed in Nunclon Delta round-bottom 96-well plates (Nunc, Cat#163320) for all conditions, except for the coated-plate one, where Immulon 4 HBX flat-bottom 96-well plates (Nunc, Cat#047612) were used. Cells and stimuli were added to a final CCM volume of 200 µl/well.
In vitro B lymphocyte stimulation was carried out by plating the purified B cells at 3x105 cells/well for two days in Nunclon Delta round-bottom 96-well plates and adding either: (1) no stimuli (ns), (2) lipopolysaccharide (LPS) at 10 µg/mL (Sigma-Aldrich, Cat#L3012-5MG), (3) purified anti-B cell receptor or IgM (αBCR) monoclonal antibody (Jackson Immunoresearch, Cat#715-006-020) at 5 µg/mL, or (4) purified anti-CD40 (αCD40) monoclonal antibody (BD Pharmingen, Cat#553787) at 10 µg/mL plus IL-4 (R&D Systems, Cat#404-ML-010/CF) at1 ng/mL, to a final CCM volume of 200 µl/well.
Stimulated lymphocytes were analyzed in terms of cytokine profiling and transcription factor analysis, as described in the following sections. Each study was performed from different plate wells as their fluorescent staining were not compatible.
Cytokine profiling
After lymphocyte purification and stimulation, culture supernatants were collected from both T and B cells assay plates and stored at -20ºC until use (for ≤ 1 week). The supernatants were screened for lymphocyte-derived secretion of IFN-ɣ, TNF-α, IL-17A, IL-6, IL-10, and IL-4 by means of the Cytometric Bead Array (CBA) Mouse Th1/Th2/Th17 Cytokine Kit (BD Pharmingen, Cat#560485), following manufacturer’s instructions and acquiring samples in the flow cytometer FACSCanto II. Cytokine concentrations were calculated using the FCAP Array Software v3.0 (BD Biosciences).
Transcription factor analysis
The major transcription factors T-bet, GATA3, RORɣT, and FOXP3 were analyzed in T cell subpopulations after in vitro stimulation. CD4+ and CD8+ T cells were stained with the monoclonal antibodies PerCP anti-CD4 (BD Pharmingen, Cat#553052) at 0.8 µg/mL and PE anti-CD8 (BD Pharmingen, Cat#553033) at 0.8 µg/mL, in PBS with 1% hiFBS at 4ºC for 20 minutes. Intracellular staining of the transcription factors was achieved using the FOXP3/Transcription Factor Staining Buffer Set (eBioscience, Cat#00-5523-00), as well as the monoclonal antibodies PE-Cy7 anti-T-bet (eBioscience, Cat#25-5825-82) at 2 µg/mL, Alexa Fluor 488 anti-GATA3 (eBioscience, Cat#53-9966-42) at 0.5 µg/mL, APC anti-RORɣT (eBioscience, Cat#17-6988-82) at 2 µg/mL, and eFluor 450 anti-FOXP3 (eBioscience, Cat#48-5773-82) at 2 µg/mL, following the instructions of the intracellular staining buffer set manufacturer. Samples were acquired with the flow cytometer FACS Canto II and the corresponding FMO (fluorescence minus one) staining was used as a control. Flow cytometry data were analyzed with FlowJo 10.0.7 (BD Biosciences).
Lymphocyte transfer
NOD.RAG-2-/- females at six weeks of age were intravenously injected (by retro-orbital sinus) with purified T cells or B cells separately, or total splenocytes from six-week-old NOD donors (or NOD.Rag2-/- donors as controls). The animals reached the end of the study at 12 weeks of age.
16S rRNA gene amplification and sequencing
Fecal samples from NOD mice (isolated and cohoused), 116C-NOD, NOD.RAG-2-/-, 116C-NOD.RAG-2-/- and C57BL/6J mice were freshly collected from the same animals at six, 12, and 20 weeks of age as part of a longitudinal study. In the case of isolated NOD, cohoused NOD, and 116C-NOD mice, fecal samples at six, 12, and 20 weeks of age were collected from non-diabetic animals, both from mice that become diabetic at a later time (future diabetic) and mice that remained resistant to the disease until the end of the study at 40 weeks (future resistant). Fecal samples were also isolated at any age of diabetic onset from isolated NOD, cohoused NOD, and 116C-NOD mice. Moreover, fecal samples of recipient NOD.RAG-2-/- females transferred with purified T cells, B cells, total NOD spleen, and total NOD.Rag2-/-spleen were also collected.
Microbiome analysis was performed as previously reported in Lleal et al.22 After collection, stool samples were stored at -80ºC until DNA extraction. Genomic DNA was extracted following the recommendations of the International Human Microbiome Standards (IHMS; http://www.microbiome-standards.org).35 Briefly, a frozen aliquot (100 mg) of each sample was suspended in 250 mL of guanidine thiocyanate, 40 mL of 10% N-lauryl sarcosine, and 500 mL of 5% N-lauryl sarcosine. Mechanical disruption of the microbial cells with beads was applied, and nucleic acids were recovered from clear lysates by alcohol precipitation. The V4 hyper-variable region of the 16S rRNA gene was amplified by PCR for each sample using the following primers: forward (V4F_515_19: 5’-GTGCCAGCAMGCCGCGGTAA-3’) and reverse (V4R_806_20: 5’-GGACTACCAGGGTATCTAAT-3’) primers as previously described.36 The amplicons were sequenced with the Illumina MiSeq system at the Autonomous University of Barcelona.
Sequence data analysis
The QIIME 2.0 bioinformatics pipeline was used to process the sequences. Briefly, sequences were demultiplexed, denoised, and dereplicated into amplicon sequence variants (ASVs) using the dada2 tool. Each sequence read was trimmed to a length of 298 bp. A total of 19,826,052 high-quality sequences of the 16S rRNA gene were generated from 442 fecal samples, with a mean of 45,161 sequences per sample. Three samples with very few reads were removed from further analysis. A feature table was generated for all samples with a minimum of 2,710 sequences per sample. The feature table was then used to perform taxonomic classification and to measure alpha- and beta-diversity. Taxonomy was assigned to each ASV using the 16S Greengenes database (gg_13_8_99 release), which contains 202,421 bacterial and archaeal sequences.
Quantification and statistical analysis
Statistical parameters including the value of “n” (where n represents the number of mice per group), the expression of data with central values and dispersion measures (mean±SD (standard deviation) or mean±SEM (standard error of mean)), the statistical methods applied to data and the degree of statistical significance are described in the corresponding figure legends.
GraphPad Prism 9.0.0 software was used to analyze survival curves of diabetes incidence with the Log-rank (Mantel-Cox) test (where the two-tailed and one-tailed t-test were applied) and their hazard ratio with the Mantel-Haenszel test; as well as to compare groups of insulitis scores, cytokine production, transcription factor subpopulations, and the Chao1 and Shannon indexes with the Mann-Whitney test (where the two-tailed t-test was applied). The statistical significance degree was considered as follows: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, **** p ≤ 0.0001.
UniFrac distances were analyzed through the PERMANOVA test. False discovery rate (FDR) corrected two-tailed p values were taken into account to consider significant results and the statistical significance degree was expressed as follows: #q ≤ 0.1, *q ≤ 0.05, **q ≤ 0.01, ***q ≤ 0.001, **** q ≤ 0.0001. To determine the association between microbiome data and biological variables, we used linear mixed models as implemented in the MaAsLin2 (Microbiome Multivariable Association with Linear Models) package.37 MaAsLin2 was set up with the following parameters: normalization = “TSS”, transform = “LOG”, correction = “BH”, analysis_method = “LM”, max_significance = 0.25 (default significance threshold), min_abundance = 0.0001, min_prevalence = 0.1. Normalized taxa were modeled with a fixed effect of treatment group and random effects of timepoint and mouse ID. Results with a false-discovery rate (FDR) lower than 0.10 were considered significant.