Mice

Six- to eight-week-old C57BL/6 J mice were purchased from Koatech (Gyeonggi-do, South Korea). B6.129S7-Rag1tm1Mom/J (Rag1KO) mice, B6(C)-Il5tm1.1(icre)Lky/J (IL-5 tdTomato reporter [Red5] and IL-5 deficient) mice, and C.129S1(B6)-Gata1tm6Sho/J (dblGATA1) mice, which lack eosinophils, were purchased from Jackson Laboratories (Bar Harbor, ME, USA). C57BL/6NTac. Cg-Rag2tm1Fwa Il2rgtm1Wjl (Rag2–/– IL2rg–/– double-KO [DKO]) mice were purchased from Taconic (Germantown, NY, USA). All animals were housed in the SPF animal care facility at Seoul National University Hospital (SNUH). All murine studies were performed at the Research Institute of SNUH, which is accredited by AAALAC International. The murine experiments were also approved by the IACUC of SNUH Biomedical Research Institute (IACUC, # 35-2019-0135).

Animal experiments

Recombinant mouse IL-33 (rmIL-33; BioLegend, 580508) was used to induce EoE in the mice. A 250 ng dose of rmIL-33 per mouse was instilled intranasally into lightly anesthetized mice. To mimic acute and chronic EoE, the mice were administered IL-33 on days 0, 1, and 2 and days 0, 7, and 14, respectively, and were sacrificed on days 7 and 28 after the initial injection. Control IgG (R&D Systems) or recombinant mouse Areg (rmAreg; R&D systems, 989-AR-100/CF) was administered intraperitoneally (2 µg per mouse) to wild-type C57BL/6 J mice on days 0, 1, and 2, and the mice were sacrificed on day 7. Goat IgG control (R&D Systems, AB-108-C), anti-Areg (R&D Systems, AF989) intraperitoneal injection or the EGFR tyrosine kinase inhibitor erlotinib hydrochloride (MCE, HY12008) via oral gavage (10 mg/kg) were used on days 2, 4, and 6. The rats were sacrificed on day 7.

Mouse histology, immunohistochemistry, and immunofluorescence image analysis

All H&E-stained images and IHC images were captured via an Olympus IX53 microscope (Center Valley, PA, USA). To analyze esophageal histology, 3 randomly selected regions of each mouse esophagus were captured at 200X magnification and quantified via ImageJ software (NIH, MD, USA). Esophageal epithelium thickness was measured by drawing 3 straight lines from the bottom to the top of the esophageal epithelium and obtaining the average length. The basal cell layer thickness was measured by drawing 3 straight lines from the basal membrane to the end of unstratified basal cells in the basal cell layer. Lamina-propria thickness was measured by drawing 3 straight lines from the top to the bottom of the lamina propria.

To investigate eosinophil infiltration, we used immunohistochemistry image analysis. The antibodies and IHC staining kits used for staining were as follows: Rb-anti-M MBP1 (ab187523) and TripleStain IHC Kit: M&R&Rt on rodent tissue (DAB, HRP/Green & AP/Red) (ab183297).

To investigate basal-cell hyperplasia and IL-5+ ILC2 accumulation in the esophageal mucosa of mice, we used immunofluorescence image analysis. The antibodies used for staining were as follows: Rb-anti-M/H p-EGFR (Y1068) (1:200; Abcam, ab40815), Rat-anti-M/H Ki67 (1:200; Invitrogen, 14-5698-82), and Rat-anti-M/H CD3 (1:200; Abcam, ab11089). Primary antibody-labeled sections were incubated with the secondary antibodies Dk-anti-Rb AF594 (1:500; Invitrogen, A-21207) or Dk-anti-Rat AF488 (1:500; Invitrogen, A-21208). One drop of ProLong diamond antifade mountant with DAPI (Invitrogen, P36961) was added, images were acquired via Nikon confocal A1 microscopy (Nikon, Tokyo, Japan), and the analysis was performed via the Nikon NIS-Elements Viewer (Nikon, Tokyo, Japan).

Immune cell preparations

The esophagus was incubated in RPMI 1640 with 2 mM EDTA (Sigma, E7889-100 ml) at 37 °C for 20 min and then minced. The minced tissue was then digested with collagenase type 4 (Worthington, LS004189) and DNase1 (Sigma, DN25-1G) in RPMI 1640 at 37 °C for 1 h. The single-cell suspension was filtered through a 40 µm cell strainer, and red blood cells were removed by treating the cells with RBC lysis buffer (BioLegend, 420301).

For single-cell isolation, mechanically minced lung was digested with collagenase type 4 and DNase 1 in RPMI 1640 at 37 °C for 1.5 h. The resulting single-cell suspension was filtered through a 40 µm cell strainer, and red blood cells were removed as described above.

Flow cytometry analysis

Single cells were suspended in cold PBS, stained with the Zombie Aqua Fixable Viability Kit (BioLegend, 423102) and treated with a mouse CD16/32 FC blocking antibody (BioLegend, 156604). To assess cytokine production, single cells were restimulated with 100 ng/ml PMA (Sigma, P8139-1MG), 1 µg/ml ionomycin (Sigma, I0634-1MG), and 0.7 µl/ml Golgi stop (BD biosciences, 554715) for 3 h. Single cells were then fixed and permeabilized via a fixation/permeabilization solution kit (BD biosciences, 554715) and a Foxp3 transcription factor staining buffer set (Invitrogen, 00-5523-00). The absolute numbers of cultured ILC2s were calculated with Precision Count Beads™ (BioLegend, 424902) according to the manufacturer’s instructions. All flow cytometry experiments were performed via a BD LSR FortessaTM X-20 (BD, NJ, USA) and analyzed via FlowJo (V10) software (BD, NJ, USA). The mouse antibodies used for flow cytometry included anti-CD45 (30-F11), anti-CD19 (1D3), anti-CD49b (DX5), anti-F4/80 (BM8), anti-FcεRIα (MAR-1), anti-CD127 (A7R34), anti-CD90.2 (30-H12), anti-CD25 (PC61), anti-IFN-γ (XMG1.2), anti-IL-17A (TC11-18H10.1), anti-IL-5 (TRFK5), anti-ICOS (C398.4 A), anti-KLRG1 (2F1), anti-SCA1 (D7), PerCP-Cy5.5-streptavidin (405214), anti-CCR3 (J073E5), anti-CCR9 (9B1), anti-CCR7 (4B12), anti-CCR4 (2G12), anti-CCR6 (29-2L17), anti-SiglecF (S17007L), anti-Gata3 (L50-823), anti-Tbet (4B10), and anti-Rorγt (B2D). Additional antibody information can be found in supplemental Table 2.

Culture of the CP-A and HET-1A human esophageal epithelial cell lines

The hTERT-immortalized human Barrett’s esophagus epithelial cell lines CP-A (ATCC, KR-42421) and HET-1A (ATCC, CRL-2692) were employed, and subculture and subculture and handling procedures were followed with the ATCC cell line culture manufacturer in the BEGM BulletKit (Lonza, CC-3170) (BEBM: Lonza, CC-3171 and BEGM™ Bronchial Epithelial Cell Growth Medium SingleQuots™ Supplements and Growth Factors: Lonza, CC-4175). The CP-A cell line was established from an area of nondysplastic metaplasia, and HET-1A was isolated from the normal esophagus. To quantify cell proliferation, CP-A and HET-1A cell lines were treated with recombinant mouse EGF (rmEGF; 10 ng/ml; Corning, CB-40001), rmAreg (100 ng/ml), and erlotinib (200 ng/ml) in BEBM (Lonza, CC-3171) for 72 h. The numbers of CP-A and HET-1A cells were subsequently quantified.

Western blot

CP-A and HET-1A were starved overnight in fresh BEBM and then treated with 100 ng/ml Areg (R&D systems, 989-AR-100/CF) or 10 ng/ml EGF (R&D systems, 2028-EG-200) in BEBM for 5, 10, 15, 60, and 180 min. To verify that Areg induces the EGFR signaling pathway, starved CP-A and HET-1A cells were challenged with 100 ng/ml rmAreg for 10 min with or without 200 ng/ml erlotinib hydrochloride (MCE, HY12008).

The CP-A and HET-1A cells were lysed and prepared for western blot analysis. The primary antibodies used for immunoblotting were as follows: rabbit anti-H/M beta-actin (1:2000, Thermo Fisher Scientific, BS-0061R), Rb anti-H/M p-EGFR (Y1068) (1:1000, Cell Signaling Technology, 2234S), Rb anti-H/M EGFR (1:1000, Cell Signaling Technology, 4267S), Rb anti-H/M p-AKT (S473) (1:1000, Cell Signaling Technology, 4060S), Rb anti-H/M AKT (1:1000, Cell Signaling Technology, 9272S), Rb anti-H/M p-ERK1/2 (1:1000, Cell Signaling Technology, 9101S), and Rb anti-H/M ERK1/2 (1:1000, Cell Signaling Technology, 9102S). Anti-Rb HRP was used as the secondary antibody (1:2000, Thermo Fisher Scientific, 7074P2). All densitometric analyses were performed via ImageJ (USA, NIH).

Quantitative real-time PCR

Mouse whole esophageal tissues were homogenized with TRIzol (Invitrogen, USA) Reagent. cDNA was synthesized via a SeniFAST cDNA synthesis kit (Bioline, UK). qPCR was performed with a SeniFAST SYBR Lo-ROX kit (Bioline, UK) or Probe Lo-ROX kit (Bioline, UK) with a CFX96 Real-Time PCR Detection System (Bio-Rad, California, USA) according to the manufacturer’s instructions. All primers were obtained from Integrated DNA Technologies (IdT; Iowa, USA). The relative expression of the target genes was normalized to that of Gapdh via the 2-ΔΔCT method.

Isolation of mouse primary esophagus basal cells and 3D organoid formation

Esophageal epithelial cells were successfully isolated from C57BL/6 J mice via EpCAM+ microbeads following enzymatic dissociation with TrypLE Express. Single-cell suspensions displayed high viability and purity, as confirmed by flow cytometry. The seeded cells (2 × 10³ cells per well) embedded in Matrigel formed well-structured organoids by day 4 of culture in YEE-MEOM medium. The organoids were spherical in shape, with diameters exceeding 20 µm. E-MEOM facilitated efficient organoid growth, and the Matrigel droplets remained intact throughout the culture period.

ILC2 coculture with esophageal organoids

Lung-derived ILC2s were cocultured with esophageal organoids in YE-MEOM supplemented with IL-2, IL-7, and IL-33, along with either anti-Areg (100 ng/mL) or erlotinib (10 µM). The coculture conditions supported the continued expansion and viability of both ILC2s and organoids over the 6-day period. Organoid size increased notably in the presence of ILC2s, with significant differences observed between the anti-Areg and erlotinib treatment groups. On day 6, organoid diameters were measured, and those cocultured with ILC2s in anti-Areg media presented an average diameter increase of 25% compared with that of the control conditions (p < 0.05). Conversely, the organoids treated with erlotinib were smaller than those in both the control and anti-Areg groups were, suggesting that EGFR inhibition adversely affected organoid growth in the ILC2 coculture system. Organoid size quantification was performed by measuring the diameter under an Olympus IX53 microscope.

Studies on human samples

Tissue microarrays (TMAs) were obtained from human esophageal tissues from two cohorts of patients. The first cohort included normal tissues from healthy controls (n = 8), esophageal mucosa tissues from patients with GERD (n = 10), and esophageal mucosa tissues from patients with EoE (n = 22) at Jeju National University Hospital and Hanyang University Hospital, South Korea (Supplementary Table 1). This study was approved by the Jeju University Hospital Institutional Review Board (IRB number 2020-11-003). All patients consented in writing to the use of their samples for scientific research.

The primary antibodies used for immunofluorescence image analysis were as follows: Rb-anti-H amphiregulin (1:100, Abcam, ab234750), Rat-anti-M/H CD3e (1:200, Abcam, ab11089) and PerCP-Cy5.5 M/H KLRG1 (1:200, Biolegend, 138418). The primary antibody-labeled sections were then incubated with the secondary antibodies DK-anti-Rat AF488 (1:500; Invitrogen, A-21208) and Dk-anti-Rb AF594 (1:500; Invitrogen, A-21207). Nuclei were stained with Hoechst (1:2000, Invitrogen, H3570) and mounted with VECTASHIELD® Antifade Mounting Medium (Vector Lab, H-1000-10). Images were acquired via Nikon confocal A1 microscopy.

To investigate p-EGFR expression in human esophageal samples, we used immunohistochemistry image analysis. The antibodies and IHC staining kits used for staining were as follows: Rb-anti-p-EGFR (Cell Signaling Technology, 2234S) and TripleStain IHC Kit: M&R&Rt on rodent tissue (DAB, HRP/Green & AP/Red) (ab183297). All the IHC images were captured via an Olympus IX53 microscope (Center Valley, PA, USA).

Analysis of public bulk mRNA-seq data and single-cell RNA-seq data

For validation of our EoE mouse model, we selected a public IL-33-overexpressing (IL33 OE) mouse esophagus dataset, which was obtained from Masuda et al. (GSE238122) and generated via the Illumina HiSeq platform. The differentially expressed genes (DEGs) with an adjusted q < 0.05 were considered statistically significant, and a multiple variable plot was generated via log2 FC, adjusted q-scores and log2 RNA expression values to visualize the DEGs.

To determine the clinical importance of ILC2-derived AREG in human EoE, we utilized public single-cell RNA-seq data from human EoE patients provided by Jiarui Ding et al. (Single Cell Portal SCP1242) [12]. Processed gene count matrices and embeddings were examined to explore AREG expression and its associated pathways in the context of EoE.

Statistical analysis

All the statistical analyses were performed via GraphPad Prism 9.5.1 software (Graph Pad, La Jolla, CA, USA). For comparisons of statistical significance, the Mann‒Whitney U test or one-way ANOVA with post hoc Tukey’s multiple comparisons test was used. For comparisons involving more than two datasets, two-way ANOVA was used. The data are presented as the mean ± standard error of the mean (SEM), and p values less than 0.05 were considered statistically significant. The numbers of samples per group and replicate experiments are indicated in the figure legends.