Nutrient-derived signals regulate eosinophil adaptation to the small intestine

Significance In addition to protecting the host from pathogenic microbes, immune cells perform noncanonical physiological functions which require tissue-specific adaptation. Eosinophils are immune cells that are particularly abundant in the nutrient-rich small intestine (SI), but whether nutrients guide their adaptation to this tissue is largely unexplored. In this study, we found that eosinophils gradually migrate along the crypt–villus axis and develop into a transcriptionally distinct villus-resident subpopulation. We determined that this adaptation depends on retinoic acid, a metabolite of vitamin A. Furthermore, we unexpectedly found that high levels of dietary amino acids limit the accumulation of villus-resident eosinophils by accelerating eosinophil turnover. This study highlights the important role of dietary nutrients in shaping the immune cell populations of the SI.


Flow cytometry
Single cell suspensions were first stained with Zombie Yellow viability dye (Biolegend) according to the manufacturer's instructions.Next, cells were incubated with anti-CD16/32 (Fc block) and fluorochrome-conjugated antibodies directed at cell surface antigens for 20 minutes at 4° C (see Table S2 for full list of antibodies).BrdU staining was performed with a FITC BrdU Flow Kit (BD Biosciences) according to the manufacturer's instructions.Staining for transcription factors was performed using the Foxp3 / Transcription Factor Staining Buffer set (Ebioscience). 123count eBeads (Invitrogen) counting beads were added to the cell suspensions to facilitate absolute cell number quantification.Flow cytometric analysis was performed on a BD Biosciences LSR II or FACSymphony cell analyzer with FACSDiva software.Downstream data processing and analysis was done in FlowJo (FlowJo, LLC).

Cell sorting
Eosinophils were sorted from single cell suspensions of proximal small intestine lamina propria.Cells were stained with Zombie Yellow and surface makers as in flow cytometry.Following staining, DAPI was added (0.05 ug/ml) so that cells that died after the initial staining could be excluded.Sorting was performed on a BD Aria II using a 100 µM nozzle and eosinophil subsets were sorted into 100% fetal bovine serum at 4°C.Each subset sample contained 225,000-415,000 sorted eosinophils from 2 pooled mice.After sorting, RNA was isolated immediately.

Isolation of immune cells from other tissues
For some organs, tissue eosinophil frequency (shown as "percent of non-circulating CD45+ cells") was determined by exclusion of circulating eosinophils.Briefly, mice were injected intravenously with 3 µg anti-CD45 antibody conjugated to Alexa Fluor 700 approximately 3 minutes prior to sacrifice to label blood-exposed cells, and these cells were identified by flow cytometry and excluded from the subsequent analysis.
Adipose tissue: Perigonadal adipose tissue was isolated and minced with scissors, then incubated in Ham's F-10 media containing 2 mM L-glutamine, 15 mg/mL bovine serum albumin (BSA), and 1 mg/mL Type 1 Collagenase (Worthington) for 30 minutes at 37°C, shaking.The resulting cell suspension was then briefly vortexed, passed through a 70 µm strainer, and washed twice with Ham's F-10 media prior to staining for flow cytometry.
Blood: Blood was collected by retro-orbital puncture into EDTA-coated tubes and lysed with Ammonium-Chloride-Potassium (ACK) buffer (6 minutes at room temperature), then washed with PBS prior to staining for flow cytometry.
Liver: After intravenous labeling with anti-CD45 (see above), the whole liver was isolated and minced with scissors, then incubated in Hank's Balanced Salt Solution (HBSS) containing 0.2 mg/mL BSA, 0.01 mg/mL DNase I, and 1 mg/mL Type IV Collagenase (Worthington) for 30 minutes at 37°C, shaking.The resulting cell suspension was then briefly vortexed, passed through a 70 µm strainer, and centrifuged.The pellet was then resuspended in 33% Percoll (diluted with PBS) and centrifuged for 20 minutes at 800 g, room temperature, without brake.The resulting pellet was then lysed with ACK buffer (6 minutes at room temperature), then washed with PBS twice before staining for flow cytometry.
Mesenteric lymph nodes: The entire chain of mesenteric lymph nodes was dissected from surrounding tissue and manually disrupted through a 70 µm strainer to create a single cell suspension.
Peritoneal cavity: To isolate peritoneal cells, the peritoneal cavity was exposed and injected with 4.5 mL of RPMI media containing 3% FBS using a 30 gauge needle.The peritoneal cavity was vigorously agitated for one minute to dislodge cells, then the fluid (containing peritoneal cells) was collected.
Stomach and colon: Single cell suspensions of stomach and colon lamina propria were isolated in the same manner as the small intestine.Spleen: Spleens were manually disrupted through a 70 µm strainer, incubated with ACK lysis buffer for 5 minutes at room temperature, resuspended in fresh media, and filtered again as needed prior to downstream applications.
Thymus: Thymus was dissected, gently blotted briefly on a Kimwipe to remove excess blood or fat if present, shredded with tweezers to create a single cell suspension, and filtered through a 70 µm strainer prior to staining for flow cytometry.

Histology, immunofluorescence/immunohistochemistry, and TUNEL
For most procedures, freshly isolated small intestine was fixed in 10% formalin overnight at room temperature.The tissue was then embedded in paraffin, sectioned at 5 μm thickness, deparaffinized with xylene and rehydrated through graded concentrations of ethanol in water.Sections were then stained with hematoxylin and eosin (H&E), or processed for immunofluorescence/immunohistochemistry or TUNEL.For MBP staining, antigen retrieval was performed with Digest-All™ 3 pepsin for 30 minutes at 37° C, followed by blocking (2.5% normal goat serum and 2.5% normal donkey serum) for 1 hour at room temperature, and then incubation with rat anti-MBP antibody (clone MT2-14.7.3, 1 μg/mL, purchased from the laboratory of Elizabeth Jacobsen, Mayo Clinic) overnight at 4° C.This was followed by incubation with secondary goat anti-rat IgG conjugated to Alexa Fluor 568 (or similar) for 30 minutes at room temperature.Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) was done with In Situ Cell Death Detection Kit (Roche) per the manufacturer's instructions.For pretreatment of the tissue sections, the permeabilization solution was used.DAPI staining was performed at 300 nM for 3 minutes.Samples were mounted with Fluoromount G (Southern Biotech).
BrdU immunohistochemistry to quantify epithelial migration was performed by Yale Pathology Tissue Services.Small intestine sections were deparaffinized and rehydrated as described above, then denatured by incubation with 1N HCl at 37° C for 30 minutes.Epitope retrieval was then performed by incubation with trypsin.Next, the sections were incubated with H 2O2 to quench endogenous peroxidase activity.Then the sections were incubated with primary anti-BrdU antibody (Sigma #B2531), followed by incubation with secondary 'MACH 2' antibody reagent (Biocare Medical) conjugated to horseradish peroxidase.The sections were developed by applying 3,3′-Diaminobenzidine (DAB), followed by counterstaining with hematoxylin, dehydration, and mounting with resinous mounting media.
For CD22 and BrdU immunofluorescence staining in eosinophils, freshly isolated proximal small intestine was immediately frozen in optimal cutting temperature (OCT) compound (Sakura Finetek).The tissue was sectioned at 10 μm thickness, air dried for 30 minutes, and fixed with 10% formalin for 20 minutes.For CD22 staining, the sections were then blocked and permeabilized with Tris-buffered saline (TBS) containing 0.025% Triton X-100, 1% bovine serum albumin, 2.5% normal donkey serum, and 0.3 M glycine for 1 hour at room temperature.This was followed by incubation with goat anti-CD22 (R&D Systems, #AF2296, 0.67 μg/mL) and rat anti-MBP (0.5 μg/mL) antibodies in the same buffer as above (without glycine) overnight at 4° C. Secondary antibody incubation, DAPI staining, and sample mounting was performed as described earlier.For BrdU staining, fixed sections were incubated with 2N HCl for 30 minutes at 37° C, neutralized by two washes in 0.1 M borate buffer (pH 8.5), followed by permeabilization with 0.2% Triton X-100 in PBS for 30 minutes at room temperature.The samples were then blocked with 2.5% normal donkey serum and 2.5% normal goat serum for 1 hour at room temperature.This was followed by incubation with anti-BrdU antibody conjugated to APC (BD Pharmingen #51-23619, 1:200) and rat anti-MBP antibody (1 μg/mL) for 2 hours at room temperature.Secondary antibody incubation, DAPI staining, and sample mounting was performed as described earlier.
Imaging of CD22 staining (in combination with MBP and DAPI) was performed with a Leica Stellaris 8 confocal microscope, using 40X/1.3or 63X/1.4oil objectives.Imaging of other immunofluorescence staining was performed with a Leica DMI6000 B widefield microscope, using 10X, 20X, or 40X air objectives.LAS X software (Leica) was used for image acquisition and export in tiff format.ImageJ/Fiji software (NIH) was used for image analysis.
Eosinophils were identified based on the presence of cytoplasmic MBP staining.Eosinophils were quantified per 'region', defined as an area of the tissue section that corresponds to 1 mm of longitudinal intestinal length.Apoptosis was indicated by overlap between TUNEL and DAPI staining.BrdU+ eosinophils were identified first by finding overlap of MBP and BrdU signal (at 20X on the widefield microscope).BrdU incorporation was then visually verified by detecting overlap between BrdU staining and nuclear DAPI staining in the eosinophil.Only eosinophils with BrdU signal above a certain threshold (equivalent to signal observed on day 3 after BrdU labeling) were counted in the analysis in order to exclude eosinophils marked by residual BrdU remaining in progenitors after termination of the BrdU pulse.Eosinophil position along the crypt-villus axis was determined by measuring its relative distance from the villus base (position '0') and villus tip (position '100').Eosinophils located below the villus base (in the vicinity of the crypts) were identified as crypt eosinophils, while those located above were identified as villus eosinophils.BrdU incorporation analysis was performed in a blinded manner.
Imaging of H&E staining and BrdU immunohistochemistry for quantification of epithelial migration was done with an Olympus BX40 microscope, using 10X/0.25 or 20X/0.40air objectives.Images were exported in JPG format.ImageJ software (NIH) was used to quantify villus length, area, and epithelial migration.Villus length was measured as the distance from the villus base to the villus tip.Villus cross-sectional area was measured between the villus tip and the base.Epithelial migration distance was measured as the distance from the villus base to the cluster of BrdU+ epithelial cells located closest to the villus tip.At least 30 villi were measured in each sample to obtain a mean value.All image analysis and quantification were done in a blinded manner.

RNA isolation and quantitative RT-PCR (qPCR)
Epithelial cell pellets or whole tissue pieces were homogenized in RNA-Bee or RNA STAT-60 reagents (both AMSBIO).Cell pellets were vortexed and tissue pieces were placed in Omni bead tubes and homogenized using a Bead Ruptor Elite bead mill homogenizer (Omni Intl).RNA was isolated with a Direct-Zol RNA MiniPrep Plus kit (Zymo Research) according to the manufacturer's instructions.RNA concentration was measured with a NanoDrop Eight Spectrophotometer.cDNA synthesis was performed with 1 μg of isolated RNA, 0.5 μg of oligo(dT)20 primer, 1 mM dNTPs, 10 mM DTT, and 50 units of RT SMART MMLV Reverse Transcriptase in First-Strand Buffer (Takara Bio).qPCR was performed with PerfeCTa SYBR Green (Quanta Bio) using the BioRad CFX384 platform.qPCR primer sequences used in this study are listed in Table S3.
The threshold value (CT) was used to calculate the relative abundance of mRNA.The mRNA abundance in each sample was expressed relative to the abundance of Rpl13 (the reference gene).
To isolate RNA from sorted eosinophils, cells were lysed in 0.5 ml of Tri Reagent (Sigma-Aldrich), incubated at room temperature for 5 minutes, followed by addition of 100 µl of chloroform.The mixture was vigorously shaken by hand, incubated at room temperature for 8 minutes, then centrifuged at 12,000 g for 15 minutes at 4° C. The aqueous phase was isolated and mixed with an equal volume of ethanol.This mixture was added to a Zymo-Spin™ IC Column and further RNA isolation was conducted with the Direct-zol RNA Microprep Kit (Zymo Research) according to the manufacturer's instructions.RNA amount and quality was measured with the High Sensitivity RNA ScreenTape assay (Agilent) according to the manufacturer's instructions.The RNA integrity number (RIN) of the samples used for sequencing ranged from 6.3 to 7.6.(B-C) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in germfree and SPF mice after 15 days of feeding with control ("C") or high protein ("HP") diet, representative of 3 experiments with 4-6 mice/group.(D-E) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice treated with a cocktail of broad spectrum antibiotics for two weeks prior to and during feeding with the indicated diets (15 days), representative of 3 experiments with 4 mice per group.(F-G) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice fed a diet containing the AHR ligand I3C or a control diet (AIN76) for 15 days, data pooled from 2 experiments with 4-5 mice/group.(H-I) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice treated (daily) with either BMS493 or vehicle control during the last 12 of 15 days of feeding with the indicated diets, data pooled from 2 experiments with 2-3 mice/group.

Figure S2 .
Figure S2.Eosinophil-deficient ΔdblGata mice exhibit decreased villus area and reductions of other immune cell populations in small intestine lamina propria.(A-D) Representative microscopy images of proximal SI sections from WT and ΔdblGata mice, stained for BrdU (brown) and hematoxylin (light blue), after 48 hours of BrdU labeling (A; scale bar =, 100 μm).Quantification of villus cross-sectional area (B), villus length (C), and epithelial migration based on BrdU labeling (D) correspond to the same experiment.Representative of 2 experiments with 3-5 mice per group.(E) Expression of Maf and c-Maf-dependent genes, relative to the reference gene Rpl13, in proximal SI epithelial fraction of WT or ΔdblGata mice, pooled from two experiments with 4-5 mice per group.

Figure
Figure S3.IL-33 signaling is not required for accumulation of CD22+ villus-resident eosinophils, and expression of RARs in SI eosinophil subsets.(A-B) Eosinophil frequency (A, % of total live CD45+ cells) and percent of eosinophils that were CD22+ (B, by flow cytometry) in proximal SI lamina propria (LP) of WT and Il1rl1 -/- (St2 -/-) mice.Representative of 3 experiments with 3-6 mice per group.(C) Expression level (as normalized counts obtained from DESeq2) of Rara, Rarb, and Rarg in sorted α4β7+, α4β7-CD22-, and CD22+ eosinophil subsets.Data in A-B are presented as mean ± SEM and were analyzed by Student's t-test.ns = not significant (p > 0.05).

Figure S4 .
Figure S4.Food restriction does not reduce SI eosinophils, and the effect of high protein diet is limited to eosinophils in SI, stomach, and thymus.

Figure S5 .
Figure S5.The effect of high protein diet on SI eosinophils is likely to be indirect and independent of several known regulators of SI eosinophil adaptation.(A) SI lamina propria (LP) cells were cultured ex vivo with the indicated concentrations of amino acids (1X is equivalent to the standard concentration of amino acids in MEM media).Graph shows the viability of eosinophils compared to the average of all other immune cells after 48 hours of culture, representative of 2 experiments.(B-C) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in germfree and SPF mice after 15 days of feeding with control ("C") or high protein ("HP") diet, representative of 3 experiments with 4-6 mice/group.(D-E) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice treated with a cocktail of broad spectrum antibiotics for two weeks prior to and during feeding with the indicated diets (15 days), representative of 3 experiments with 4 mice per group.(F-G) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice fed a diet containing the AHR ligand I3C or a control diet (AIN76) for 15 days, data pooled from 2 experiments with 4-5 mice/group.(H-I) SI LP eosinophil frequencies and percent of eosinophils that were CD22+ in mice treated (daily) with either BMS493 or vehicle control during the last 12 of 15 days of feeding with the indicated diets, data pooled from 2 experiments with 2-3 mice/group.