Group 2 innate lymphoid cells are elevated and activated in chronic rhinosinusitis with nasal polyps

Patients and tissue collection

Patients with CRS were recruited from the Otolaryngology clinic and the Northwestern Sinus Center of Northwestern Medicine. All patients met the criteria for CRS as defined by the European Position Paper on Rhinosinusitis and Nasal Polyps 2012 1. Patients with an established immunodeficiency, pregnancy, coagulation disorder or diagnosis of Churg-Strauss syndrome or cystic fibrosis were excluded from the study. During endoscopic sinus surgery, ethmoid tissue was collected from CRS without NPs (CRSsNP) and CRSwNP, and NP tissue was collected in patients diagnosed with CRSwNP. Disease-free ethmoid sinus tissues from normal control patients without a history of CRS were obtained during transnasal endoscopy skull base procedures. Tonsil tissue was obtained when patients underwent tonsillectomy for chronic tonsillitis, recurrent acute tonsillitis, or obstructive sleep apnea as collected by faculty of the Otolaryngology Department of Northwestern Medicine. Patients were skin-tested for pollens, dust mites, pets, molds, and cockroach using Hollister-Stier (Spokane, WA) extracts. Several patients were taking a variety of medications, including corticosteroids. Details of patients’ characteristics are included in Table 1. All subjects signed informed consent and the study was approved by the Institutional Review Board of Northwestern University Feinberg School of Medicine (IRB Project Number: STU00200416). Human buffy coats (American Red Cross, St. Paul, MN) and human peripheral blood leuko paks (STEMCELL Technologies, Vancouver, BC, Canada) were obtained from healthy subjects for isolation of peripheral blood mononuclear cells (PBMC).

Cell isolation and flow cytometric analysis

Cells were isolated from tonsil and sinus tissue using a previously described method 12, 13. We obtained 759 ± 296 million live cells (tonsil, n = 21), 0.35 ± 0.21 million live cells (control, n = 5), 1.21 ± 0.69 million live cells (CRSsNP, n = 13), 3.96 ± 1.68 million live cells (CRSwNP ethmoid, n = 9) and 42.4 ± 10.4 million live cells (NP, n = 25) from 1847 ± 360 mg, 107 ± 48 mg, 108 ± 19 mg, 206 ± 64 mg, and 1630 ± 391 mg tissue, respectively.

Cells were first treated with Aqua LIVE/DEAD fixable dead cell staining reagent (Invitrogen, Carlsbad, CA, USA) as a live/dead discriminator. Cells were then incubated with an Fc Block reagent (Miltenyi Biotec, San Diego, CA, USA) for 10 min at 4°C in the dark. All antibodies were obtained from BioLegend (San Diego, CA, USA), unless otherwise stated. The following antibodies and dilutions were used to stain the surface of the cells: FITC anti-human Lineage Cocktail (CD3, CD14, CD16, CD19, CD20, CD56, 1:20), 2 μg/ml FITC anti-FcϵRIa (AER-37), 4 μg/ml FITC anti-CD11c (Bu15), 0.25 μg/ml Alexa Fluor 700 anti-CD45 (HI30), 5 μg/ml Alexa Fluor 647 anti-CRTH2 (Bim16), 2.25 μg/ml PE/Cy7 anti-CD127 (A019D5), 5 μg/ml APC/Cy7 anti-CD161 (HP-3G10), PE/CF594 anti-CD117 (yb5.b8, 1:20, BD Biosciences, San Jose, CA, USA), 2.5 μg/ml PE anti-NKp44 (P44-8), 5 μg/ml Brilliant Violet 421 anti-KLRG1 (2F1/KLRG1), 2 μg/ml PerCP/Cy5.5 anti-ICOS (C398.4A), 2.5 μg/ml PerCP/Cy5.5 anti-CD4 (RPA-T4), and 1.25 μg/ml PerCP/Cy5.5 anti-CD5 (L17F12). Cells were stained for 30 min at 4°C in the dark, and washed with MACS buffer (Miltenyi Biotech). Cells were fixed with a BD Cytofix/Cytoperm Kit, resuspended in MACS buffer and stored at 4°C in the dark before analysis on an LSRII (BD Biosciences). For the detection of intracellular CD127, fixed cells were incubated with 0.6 μg/ml PE anti-CD127 (ebioRDR5, eBioscience) in Perm/wash buffer (BD Biosciences) for 30 min at 4°C in the dark. A total of 4,983,000 ± 538,915 events (PBMC), 3,981,000 ± 322,800 events (tonsil), 146,800 ± 61,400 events (control), 151,900 ± 41,600 events (CRSsNP), 499,700 ± 199,100 events (CRSwNP ethmoid) or 1,418,000 ± 243,200 events (CRSwNP) were collected. A minimum of 3,000,000 (PBMC), 1,000,000 (tonsil), 19,000 (control), 15,000 (CRSsNP), 57,000 (CRSwNP ethmoid), or 30,000 (NP) events was collected for each sample. The frequency of CD45⁺ cells within the live cell population was 99.8 ± 0.05% (PBMC), 99.9 ± 0.02% (tonsil), 77.3 ± 6.81% (control), 85.2 ± 2.12% (CRSsNP), 91.8 ± 2.21% (CRSwNP ethmoid) and 95.2 ± 0.65% (NP) and it was significantly higher in CRSwNP (p < 0.05) and in NP (p < 0.001) than in CRSsNP. The frequency of granulocytes (SSC^high, eosinophils plus neutrophils) in CD45⁺ cells was 0.488 ± 0.154% (PBMC), 0.454 ± 0.073% (tonsil), 23.9 ± 6.60% (control), 12.2 ± 2.97% (CRSsNP), 37.1 ± 5.59% (CRSwNP), and 59.0 ± 2.35% (NP). All analysis and compensation were performed with FlowJo software, version 10.1 (TreeStar, Ashland, OR), and each experiment contained the proper single-stained control beads (BD Biosciences and eBioscience) and fluorescence minus one (FMO) negative controls.

ILC2 sorting and real-time RT-PCR

Details can be found in the Materials and methods section in the Supporting information.

Cell culture

PBMC were suspended in RPMI 1640 medium (Invitrogen) supplemented with 10% FBS (Invitrogen), 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen), and were cultured in the presence or absence of 10 ng/ml TSLP (R&D systems, Minneapolis, MN) or 10 ng/ml IL-7 (R&D systems) for two days. The expression of CD127 on ILC2 was detected by a CytoFLEX (Beckman Coulter, Indianapolis, IN). Data analysis was performed with FlowJo software.

Sorted ILC2 (10,000 cells/ml) were suspended in RPMI 1640 medium supplemented with 25 IU/ml IL-2 (Prometheus, San Diego, CA), 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin and were cultured in the presence or absence of 10 ng/ml IL-33 (BioLegend) for 4 days. The concentrations of IL-5 and IL-13 in cell-free supernatants were measured using a MILLIPLEX MAP Human Cytokine/Chemokine Panel from EMD Millipore (Billerica, MA). The minimal detection limits for IL-5 and IL-13 are 3.2 pg/ml.

Statistics

All data were reported as the median (25–75% interquartiles) or as the mean ± SEM. Differences between groups were analyzed using the 1-way ANOVA Kruskal–Wallis Dunn's multiple comparison test, the Mann-Whitney test, or the Paired t-test. A p-value of less than 0.05 was considered significant.

ILC2 are elevated in CRSwNP

To characterize ILC subsets, we collected ethmoid tissue from CRSsNP, NPs from CRSwNP, tonsil tissue from patients undergoing tonsillectomy and PBMC from healthy subjects, and determined the presence and frequency of ILC in the CD45⁺ cell population by flow cytometry based on a gating strategy developed by Hazenberg and Spits 4 with some modifications. We first selected the singlets and the Aqua⁻ live cell population and then selected the CD45⁺ population, excluded granulocytes (side scatter (SSC) high), and selected the Lin (CD3, CD11c, CD14, CD16, CD19, CD20, CD56, FcϵR1a) negative population (Fig. 1). In order to determine a cut off line for the Lin⁻ population, we used anti-CD4 or anti-CD5 instead of anti-ICOS and anti-KLRG1 in a second panel (not shown). We then identified ILC subsets using the following markers; ILC1 (CD127⁺, CRTH2⁻, CD161⁺, CD117⁻, NKp44⁻), ILC2 (CD127⁺, CRTH2⁺, CD161⁺), NCR⁻ ILC3 (CD127⁺, CRTH2⁻, CD161⁺, CD117⁺, NKp44⁻) and NCR⁺ ILC3 (CD127⁺, CRTH2⁻, CD161⁺, CD117⁺, NKp44⁺) (Fig. 1). We found that all ILC subsets were present in PBMC, tonsil and CRS but the frequency was tissue dependent (Fig. 1). ILC2 were the major type in PBMC and NP while NCR⁺ ILC3 were dominant in tonsil (Fig. 1). In contrast, the frequency of ILC subsets was similar, except for NCR⁺ ILC3, in CRSsNP (Fig. 1).

We next compared the frequency of ILC between groups. We found that the frequency of ILC1 was significantly elevated in CRSsNP compared to PBMC, tonsil, and NP (Fig. 2A). The frequency of NCR⁻ ILC3 was also significantly higher in CRSsNP than in PBMC and tonsil (Fig. 2A). In contrast, NCR⁺ ILC3 were significantly elevated in tonsils (Fig. 2A). We then focused on ILC2 which are predominant in NP. We found that the frequency of ILC2 was significantly elevated in NP compared to PBMC, tonsil, and CRSsNP (Fig. 2A).

Next we compared the frequency of ILC subsets between NP tissue and ethmoid sinus tissues from control, CRSsNP and CRSwNP in a small cohort. We found that there were no differences in the frequency of ILC1 or ILC3 between groups (Fig. 2B). In contrast, the frequency of ILC2 was elevated in NP compared to ethmoid tissues from control, CRSsNP or CRSwNP (Fig. 2B). Since inflammatory cells were more accumulated in CRSwNP, we also normalized the data by weight of the tissue. We found that ILC2 numbers were also significantly elevated in NP (Fig. 2C). Although CRSwNP ethmoid tissues showed 23-fold higher levels of ILC2 compared to control ethmoid tissue, this difference did not reach significance (Fig. 2C). This was probably due to the sample size of the current study and it will require further study with a larger cohort to test the hypothesis that ILC2 are also elevated in non polyp areas of CRSwNP. Taken together, these results confirmed published studies 5-9 that ILC2 are highly accumulated in CRSwNP.

To test whether accumulation of ILC2 in CRSwNP patients can be found systemically, we also isolated PBMC from CRSwNP patients and compared them to control subjects. We found that there was no significant difference in the frequency of ILC subsets when PBMC from healthy subjects and patients with CRSwNP were compared (Fig. 2D). This suggests that the frequency of ILC subsets in CRS is regulated locally. However, a recent study showed that peripheral blood ILC2 were increased during pollen season in patients with allergic rhinitis 14. The determination of seasonal changes of blood ILC2 may require in future a study with a larger cohort. Since 52% of CRSwNP patients took oral and/or nasal steroid, 68% of CRSwNP were comorbid with asthma and 24% have aspirin exacerbated respiratory disease (AERD), we examined whether this affected the frequency of ILC2 in CRSwNP. However, we found no significant difference in ILC2 levels based on the status of steroid treatment, status of asthmatics or presence of AERD (Fig. S1).

Differential expression of CD127 and ICOS on ILC2 in CRSwNP

Since CRSwNP is characterized by type 2 inflammation and ILC2 can produce large amounts of type 2 cytokines, we hypothesized that ILC2 are not only elevated but also activated in CRSwNP. Recently, Salimi et al. found the up-regulation of killer cell lectin like receptor G1 (KLRG1) on ILC2 in atopic dermatitis, and Maazi et al. reported that inducible T-cell co-stimulator (ICOS) on ILC2 controlled type 2 inflammation 15, 16. Therefore, we first examined the expression of KLRG1, ICOS and ILC2 markers in ILC2 from PBMC, tonsil, and CRS. When we compared NP ILC2 to PBMC, tonsil and CRSsNP ILC2, we found that the levels of side scatter (SSC) and ICOS were increased, CD127 (except in comparison with CRSsNP) was down-regulated but CRTH2 did not show a difference (Fig. 3). In addition, KLRG1 was weakly up-regulated in NP ILC2 compared to tonsil and CRSsNP ILC2 (Fig. 3B). In contrast, there was no difference in the level of marker expression on ILC2 between PBMC, tonsils and CRSsNP (Fig. 3). This result suggests that ILC2 in CRSwNP might be subject to several micro environmental factors within the inflamed area potentially changing expression levels of cell surface molecules.

TSLP and IL-33 down-regulate cell surface expression of CD127 on ILC2

CD127 (also known as IL-7Ra) is the α subunit of the receptor for both IL-7 and thymic stromal lymphopoietin (TSLP) and the TSLP receptor complex is expressed on ILC2 4. We recently found that TSLP is upregulated in CRSwNP 17. We investigated the potential role of TSLP and IL-7 on the cell surface expression of CD127 using blood ILC2. We found that TSLP and IL-7 were able to down-regulate cell surface expression of CD127 in blood ILC2 (Fig. 4A and B). Interestingly, they also increased the level of SSC on blood ILC2 (Fig. 4A and B). In addition to TSLP, IL-33 has been reported present in CRSwNP tissue 6. IL-33 is well known to act as a potent stimulus on ILC2 inducing the production of type 2 cytokines including IL-5 and IL-13 accompanied by morphological changes 5, 18. We next investigated the effect of IL-33 on the levels of CD127 and SSC. We found that IL-33 down-regulated CD127 and enhanced levels of SSC in blood ILC2 (Fig. 4C). These results indicate that the levels of CD127 and SSC are changed in activated ILC2, and ILC2 in NPs may be activated in part due to the presence of TSLP and IL-33 in NPs.

ILC2 are activated in CRSwNP

To further investigate whether ILC2 are truly activated in NP in vivo, we isolated ILC2 from the blood of healthy subjects and from NP tissue of CRSwNP patients and compared the production of type 2 cytokines. Interestingly, NP ILC2 but not blood ILC2 spontaneously released IL-5 and IL-13 without additional stimulation, although a common ILC2 activator, IL-33, was able to induce production of type 2 cytokines from both blood and NP ILC2 (Fig. 5A). This result indicates that ILC2 are already activated in CRSwNP and indeed produce type 2 cytokines in vivo.

We next investigated whether the accumulation of ILC2 affects type 2 inflammation in NPs. We isolated RNA from the same NP tissues and examined the levels of type 2 cytokines. We found that concentration of ILC2 correlated with tissue IL-5 and IL-13 in NPs (Fig. 5B). We also investigated whether the observed changes of SSC, CD127, and ICOS in ILC2 affect type 2 inflammation in NPs. We found that levels of SSC and CD127 but not ICOS, CRTH2 nor KLRG1 on ILC2 correlated with tissue IL-5 and IL-13 in CRSwNP (Fig. 5C and S2). We also found that levels of SSC and CD127 on sorted ILC2 were significantly correlated with spontaneous production of IL-5 and IL-13 (Fig. S3), although this data was from only a small cohort. These results indicate that ILC2 are activated in NPs and level of CD127 on ILC2 may be used as a marker for activation status of ILC2 in CRSwNP, although the use of SSC as a phenotypical cut off for the ILC2 population might be difficult.