DOI: 10.1002/JLB.2MA0820-320RR

TISSUE/SYSTEM: BLOOD

Cytokine trafficking of IL-9 and IL-13 through TfnRc+ vesicles in activated human eosinophils

Sarah Almas1 Nawell Fayad1 Ojas Srivastava1 Mujtaba Siddique1 Nicolas Touret2 Paige Lacy1

1Alberta Respiratory Centre, Department of Medicine, University of Alberta, Edmonton, Abstract Alberta, Canada Eosinophils are granulocytes that are elevated in lung mucosa in approximately half of patients 2Department of Biochemistry, University of with allergic asthma. These highly granulated cells can synthesize and secrete many cytokines, Alberta, Edmonton, Alberta, Canada including IL-9 and IL-13. We hypothesized that IL-9 and IL-13 are found as preformed mediators Correspondence in crystalloid granules and secreted using distinct trafficking pathways. Human eosinophils were Paige Lacy, Alberta Respiratory Centre, Depart- purified from peripheral venous blood, adhered to coverslips, and stimulated with platelet acti- ment of Medicine, University of Alberta, 559 HMRC, Edmonton, AB, T6G 2S2, Canada. vating factor (PAF). Cells were immunolabeled with antibodies to IL-9 or IL-13 and colocalized E-mail: [email protected] with markers for secretory organelles, using CD63 for crystalloid granules and transferrin recep- tor (TfnRc) for vesicles. Fixed cells were imaged using super-resolution microscopy and quantified by colocalization using Pearson’s correlation coefficient. IL-9 immunofluorescence increased in a time-dependent manner to PAF, whereas colocalization of IL-9 and CD63 significantly increased from 0.52 to 0.67 after 5 min PAF. Colocalization of IL-9 with TfnRc significantly increased at 60 min of stimulation with PAF (0.54 at 0 min to 0.60 at 60 min). IL-13 showed lower colocaliza- tion with CD63 (0.55) than TfnRc (0.63) in unstimulated cells. Upon PAF stimulation, IL-13 inten- sity transiently decreased at 5 and 60 min, whereas colocalization of IL-13 with CD63 decreased throughout stimulation to 0.43. While colocalization of IL-13 with TfnRc transiently increased to 0.66 at 5 min PAF, it returned to near baseline levels (0.64) after 15 min PAF. Our results suggest that IL-9 and IL-13 are stored in crystalloid granules as well as endosomal structures, and that IL-9 is primarily trafficked to the cell surface via TfnRc+ endosome-like vesicles.

KEYWORDS allergy/asthma, cytokines, degranulation, secretory granules, super-resolution microscopy, vesicles

1 INTRODUCTION highly active lipid mediator that acts on PAF receptor and other mech- anisms to induce eosinophil activation.6 Eosinophils are highly granulated white blood cells and are prominent Eosinophils synthesize and store several Th2 cytokines, includ- in asthma.1 These granulocytes are rarely observed in healthy airways, ing IL-4, IL-5, IL-9, and IL-13, many of which exhibit granular dis- but in approximately 50% of asthmatics, they become substantially tribution in immunostained cells.5 Both IL-9 and IL-13 skew the elevated in blood and tissue.2,3 Increased eosinophilic inflammation immune response to a Th2 phenotype in eosinophilic inflammatory of the airways broadly correlates with increased asthma severity and diseases.7–10 Although the importance of these Th2 cytokines in immu- is often accompanied by degranulation.4 When activated, eosinophils nity and asthma is well established, the specific intracellular traf- rapidly release immunomodulatory cytokines.5 A well-known stimu- ficking mechanisms of these cytokines in eosinophils have not been lant of eosinophil degranulation is platelet activating factor (PAF), a characterized. Cytokine secretion pathways in innate immune cells are dependent on soluble NSF attachment receptors (SNAREs) that are Abbreviations: PAF, Platelet activating factor; SNARE, Soluble NSF attachment protein involved in the docking and fusion of vesicles to the cell membrane for receptor; TfnRc, ; VAMP, Vesicle-associated .

Received: 31 October 2019 Revised: 17 August 2020 Accepted: 21 August 2020 JLeukocBiol.2020;1–10. www.jleukbio.org ○c 2020 Society for Leukocyte Biology 1 2 ALMAS ET AL. cargo release.11–14 A SNARE complex comprising vesicle-associated Franklin Lakes, NJ, USA). Differential counts of whole blood were per- membrane protein (VAMP)-7 on crystalloid granules and SNAP23 and formed using Kimura stain and C-Chip hemocytometers (Bulldog Bio- syntaxin-4 at the plasma membrane has been proposed to mediate NanoEnTek, Portsmouth, NH, USA). the release of cargo in cytoplasmic eosinophil compartments.15–17 Our recent study demonstrated that eosinophils from an eoCRE trans- 2.3 Human blood eosinophil isolation and genic mouse model with eosinophil-specific VAMP-7 deficiency preparation for immunolabeling showed decreased degranulation in correlation with diminished airway hyperresponsiveness in allergen-sensitized and challenged Eosinophils were purified from whole blood using MACSxpress mice.18 Eosinophils from eoCRE/V7 mice also exhibited selectively human eosinophil isolation kits (Miltenyi Biotec, Bergisch Gladbach, decreased secretion of IL-9 but not other cytokines, suggesting that Germany).23 Isolations resulted in ≥95% purity and ≥99% viability, as VAMP-7-mediated exocytosis contributes to IL-9 release.18 Immuno- measured with Kimura stain and Trypan Blue stain, respectively. Puri- labeling of IL-9 in eosinophils exhibits a granular pattern of staining,19 fied eosinophils were then washed once with serum-free RPMI 1640 and we have previously reported that IL-13 is stored in crystalloid (Gibco Life Technologies, Waltham, MA, USA), centrifuged, and resus- granules in human eosinophils.20 However, the membrane transport pended in this media. Borosilicate glass coverslips (#1, Thermo Fisher pathways responsible for IL-9 and IL-13 trafficking and release have Scientific, Waltham, MA, USA) were cleaned with 95% ethanol and not been determined. In macrophages, recycling endosomes have been placed in 6-well plates (Thermo Fisher Scientific). Per well, 1 × 106 shown to be important traffickers of newly synthesized cytokines from eosinophils/ml were added. Cells were incubated at 37◦C for 30 min the trans-Golgi network.11,12 Recycling endosomes, bearing membrane to allow for adherence to coverslips. To stimulate cells, 5 µMPAF bound markers such as Arf6, Rab11, VAMP-3, and transferrin receptor (C18; Sigma-Aldrich, St. Louis, MO, USA) was added for 0, 5, 15, (TfnRc), canonically recycle membrane and receptors such TfnRc and 60 min. Next, coverslips were washed thrice with PBS (HyCline to the cell surface.11,12 These recycling endosomes have also been Labs, GE Healthcare Life Sciences, Chicago, IL, USA). Eosinophils implicated in cytokine trafficking in NK and microglial cells,21,22 but were then fixed with 4% PFA for 30 min, washed, and stored in PBS no reports have yet suggested a role for them in eosinophil cytokine at 4◦C. trafficking. Our study investigated intracellular storage sites of cytokines IL-9 2.4 Immunofluorescence staining and IL-13 in eosinophils and pathways involved in their release. We hypothesized that the Th2 cytokines, IL-9 and IL-13, are found as Cells were permeabilized and blocked with 0.1% saponin (Sigma- preformed cytokines stored in eosinophil crystalloid granules, and Aldrich), 2% BSA, and 10% heat-inactivated human serum in PBS for other intracellular organelles, such as small endosome-like vesicles, 30 min. Eosinophils were immunolabeled with Alexa Fluor 647 mouse and are secreted using distinct trafficking pathways. We propose to 𝛼-human IL-9 (BD Biosciences) or Alexa Fluor 647 mouse 𝛼-human + examine the presence of TfnRc vesicles in eosinophils, which have IL-13 (R&D Systems, Minneapolis, MN, USA) and Alexa Fluor 488 not been explored in these cells before. In this study, we used human rabbit 𝛼-human TfnRc (Bioss, Woburn, MA, USA) or Alexa Fluor 488 eosinophils isolated from venous blood of allergic and/or asthmatic anti-human CD63 (Invitrogen, Waltham, MA, USA)/FITC anti-human voluntary donors. Immunofluorescence analysis was coupled with CD63 (Bio-Rad, Hercules, CA, USA). Isotype controls were included super-resolution microscopy to determine IL-9 and IL-13 trafficking for all antibodies (Alexa Fluor 647-conjugated mouse IgG1𝜅 [BD Bio- following stimulation with PAF. sciences], rabbit IgG Alexa Fluor 488-conjugated [Bioss], mouse IgG1

Alexa Fluor 488-conjugated [R&D Systems], and mouse IgG1 FITC- conjugated [Bio-Rad], respectively) (Supplementary Fig. S1). Rho- 2 MATERIALS AND METHODS damine phalloidin (Thermo Fisher Scientific) was used to label F-actin and DAPI (Thermo Fisher Scientific) for nuclear staining. After label- ing, coverslips were washed and mounted to slides using ProLong Glass 2.1 Donor recruitment Antifade Mountant (Thermo Fisher Scientific). Eosinophil immunoflu- Volunteers with self-reported asthma and/or allergies not taking any orescence analysis was conducted in cells obtained from at least three form of medication for at least 2 wk prior to sampling were recruited. donor samples. All participants were 18 yr or older and provided informed consent prior to blood collection. Donors with over 3% circulating eosinophils 2.5 Image collection and analysis were selected for the study. This study was approved by the University of Alberta human ethics review board. Cells were imaged using a DeltaVision OMX super-resolution micro- scope. In all experiments, images were collected from isotype and test antibodies to correct for endogenous fluorescence and nonspecific 2.2 Human whole blood collection binding by antibodies. Colocalization was quantified through Pearson’s A total of 30 ml of human blood was collected from each donor in EDTA correlation coefficient using Volocity image analysis software (Version (K2) vacutainer blood collection tubes (Becton Dickinson Biosciences, 6.5.1, Puslinch, Ontario, Canada). We found Pearson’s correlation ALMAS ET AL. 3 coefficient to be best suited to determine colocalization for our and Supplementary Figs. S5, S6).27 The stretched morphology of PAF- experiments.24,25 An automated process used to automatically define stimulated eosinophils was accompanied by the appearance of numer- the region of interest and subsequent analysis was employed, as ous filopodia that were often polarized at one end of the cell, particu- manual selection of cells yielded no overall statistical difference from larly after 60 min of stimulation. Interestingly, we observed increased the automated process (Supplementary Figs. S2 and S3). At least 50 IL-9 immunofluorescence intensity at each time point of PAF stim- cells from each time point were analyzed. Object-based colocalization ulation (P < 0.05, Fig. 1B), which may be reflective of de novo syn- of granules and vesicles was carried out using Imaris software (Version thesis. Although variability between different donors occurred, an 8.1, Zurich, Switzerland) (Supplementary Fig. S4). overall increasing trend was consistently observed across five donors (Fig. 5A). With 5, 15, and 60 min of PAF stimulation, colocalization of IL-9 with 2.6 Statistical analysis CD63 transiently increased at 5 min of stimulation before returning to All analysis was performed using GraphPad Prism software (Version 8, baseline values (P < 0.05, Fig. 1C), and this trend was reflected in data La Jolla, CA, USA). Statistical comparisons between time points were from multiple donors (Fig. 5B). To explore this from a different angle, performed with a Kruskal-Wallis analysis followed by Dunn’s multiple we used object-based colocalization method to determine proportions comparisons test. Data from all experimental replicates were analyzed of IL-9 granule- or vesicle-associated cytokine pools (Fig. 6). Interest- with a two-way ANOVA followed by Dunnett’s multiple comparisons ingly, object-based colocalization showed that greater proportions of test. A minimum of three donors were analyzed for all statistical tests. CD63+ granules containing IL-9 were observed only at 5 min of PAF Significance was set at P < 0.05. stimulation (Fig. 6C). In contrast, IL-9 increasingly colocalized with TfnRc over 60 min (P < 0.01, Figs. 2B and 5C). Similarly, greater proportions of TfnRc+ vesicles containing IL-9 were observed at 15 and 60 min of PAF stimu- 3 RESULTS AND DISCUSSION lation (Fig. 6E).

3.1 IL-9 and IL-13 colocalize with eosinophil CD63+ + crystalloid granules and TfnRc vesicles 3.3 IL-13 colocalization with CD63+ crystalloid

To determine the intracellular distribution of IL-9 and IL-13 in granules decreases throughout PAF stimulation eosinophils, adherent cells were fixed and immunolabeled for super- In marked contrast to IL-9, IL-13 immunofluorescence intensity con- resolution microscopy imaging. Both IL-9 and IL-13 exhibited punctate sistently showed a modest decrease after 60 min of PAF stimula- staining throughout the cytoplasm in unstimulated cells, suggesting tion, although in one donor it declined at 5 min as well (P < 0.01, granular storage of preformed cytokines (Figs. 1–4). We assessed Figs. 3B and 5D). Overall, our findings suggest that IL-13 was not the colocalization of these Th2 cytokines with membrane-bound newly synthesized during PAF stimulation. IL-13 colocalization with markers of intracellular organelles, CD63 and TfnRc. CD63 labeling CD63 decreased over the entire PAF stimulation time course (Figs. 3C represents crystalloid granules, whereas TfnRc staining includes and 5E). This decrease in IL-13 colocalization with CD63 suggests endosomal populations (recycling, early, and late), referred to here that IL-13 may be shuttled from its storage site in crystalloid gran- + as TfnRc vesicles. CD63 staining appeared doughnut shaped, ules for mobilization to the cell surface as the cell is activated. Con- indicative of membrane localization of this in crystal- versely, IL-13 colocalization with TfnRc+ vesicles transiently increased 18,26 loid granules (Figs. 1, 3), as previously reported. The staining after 5 min PAF stimulation and then returned to baseline levels by for TfnRc followed a dense, granular pattern throughout the cell 15 min (Figs. 4B and 5F). Although IL-13 colocalized with TfnRc+ cytoplasm with fewer ring-shaped structures (Figs. 2, 4). Pearson’s vesicles, there was no overall change in colocalization during stim- correlation coefficient was used to evaluate colocalization of IL-9 ulation, suggesting that TfnRc+ vesicles are not the main traffick- + + and IL-13 with CD63 crystalloid granules and TfnRc vesicles upon ing compartment for this cytokine. As IL-13 did not follow the same PAF activation of eosinophils. Our findings confirm earlier obser- pattern of colocalization compared to IL-9, our findings suggest an vations that IL-9 and IL-13 are stored as preformed cytokines in independent pathway of IL-13 trafficking to the cell periphery for 19,20 eosinophils. release. This is the first report of TfnRc+ vesicles in eosinophils, suggesting 3.2 IL-9 immunofluorescence and colocalization that eosinophils, like many other cells in the body, utilize membrane recycling pathways for trafficking of receptors and granule contents. with TfnRc+ vesicles increase following PAF Moreover, TfnRc+ vesicles colocalized with IL-9 and IL-13, suggesting stimulation a role for these vesicles in cytokine trafficking in eosinophils. We found Following stimulation by PAF (5 µM), we observed morphologic increased colocalization of IL-9 with TfnRc following stimulation with changes from compact, rounded eosinophils to stretched and elon- PAF, indicating elevated IL-9 content in these vesicles following stim- gated shapes by 15 and 60 min, as previously reported (Figs. 1–4, ulation and that TfnRc+ vesicles play a key role in IL-9 trafficking for 4 ALMAS ET AL.

FIGURE 1 IL-9 intensity and colocalization with CD63+ crystalloid granules changes dynamically in platelet activating factor (PAF)- stimulated eosinophils.(A) Human eosinophils were adhered to glass coverslips and stimulated with PAF for 0, 5, 15, and 60 min. Cells were fixed, permeabilized, and labeled with anti-IL-9 (red) and anti-CD63 (green) as a marker for crystalloid granules. Actin was stained with rhodamine-phalloidin (gray) and nuclei with DAPI (blue). Yellow regions indicate colocalization between intracellular cytokine and crystalloid gran- ules. High magnification insets of the merged images are shown at right. Scale bar represents 5 µm. Using Volocity software, immunofluorescence intensity values for intracellular cytokine levels (B) and Pearson’s correlation coefficients for IL-9 and CD63 colocalization (C) were measured. A minimum of 50 eosinophils were analyzed in each plot. Data are shown from a representative donor. Means are indicated by red lines. Error bars indicate SD.*P < 0.05 and **P < 0.01 eventual release from eosinophils. Our findings on IL-9 and IL-13 colo- A limitation of our study is that there is an inherent bias in our analy- calization with the endosomal marker TfnRc and crystalloid granule sis, as only cells that adhered to coverslips were evaluated. This subset marker CD63 provide insights into mechanisms for trafficking of these of adherent eosinophils may display different cytokine levels and traf- cytokines in eosinophils. ficking compared to the population in suspension. However, it was not ALMAS ET AL. 5

FIGURE2 Increased IL-9 colocalization with transferrin receptor (TfnRc)+ vesicles in platelet activating factor (PAF)-stimulated eosinophils. Human eosinophils were stimulated and immunolabeled as described in Figure 1 with anti-IL-9 (red) and anti-TfnRc (green) as a marker of vesicles. Scale bar represents 5 µm. Data are shown from a representative donor. Means are indicated by red lines. Error bars indicate SD.**P < 0.01

feasible to employ a different technique to analyze the entire sample of To further elucidate the trafficking pathways of IL-9 and IL-13, cells, such as flow cytometry, as flow cytometry does not provide suf- we attempted to stain for additional markers of recycling endosomes ficient intracellular resolution required for colocalization of intracel- (VAMP-3 and Rab11) and crystalloid granules (Rab27a and VAMP- lular crystalloid granules and vesicles. Additionally, only atopic donors 7), as well as other intracellular compartments, including the Golgi were tested in this study as healthy individuals do not provide suffi- apparatus (giantin), endoplasmic reticulum (calreticulin), and SNARE cient numbers of eosinophils for isotype and test antibodies at all time (SNAP-23 and syntaxin-4). However, these markers did not points to conduct our experiments. generate a signal greater than their respective isotype control when 6 ALMAS ET AL.

FIGURE 3 IL-13 intensity and colocalization with CD63+ crystalloid granules changes in platelet activating factor (PAF)-stimulated eosinophils. Human eosinophils were stimulated and labeled as described in Figure 1 with anti-IL-13 (red) substituting for anti-IL-9 and anti-CD63 (green) as a marker for crystalloid granules. Scale bar represents 5 µm. Data are shown from a representative donor. Means are indicated by red lines. Error bars indicate SD.*P < 0.05 and **P < 0.01 labeled using our immunostaining protocol (data not shown). These detection of supernatant-derived cytokines by ELISA. Previous ELISA challenges in immunostaining were likely due to high levels of endoge- measurements of mouse eosinophil IL-9 and IL-13 used cells from nous eosinophilic autofluorescence and nonspecific antibody labeling. IL-5 transgenic mice with 15-fold more cells per sample and showed Cytokine secretion could not be detected at levels above the higher levels of these cytokines, whereas assays on human eosinophils limit of quantification (>30 pg/ml) for ELISAs when we examined detected negligible quantities (<10 pg/ml) of released IL-9 and supernatants from 1 × 106 eosinophils stimulated by 5 µMPAF IL-13.6,18 or positive control calcium ionophore (data not shown). This was Future studies will benefit from testing additional cytokines and presumably due to limited numbers of eosinophils, which prevented chemokines such as CCL5/RANTES, a key proinflammatory chemokine ALMAS ET AL. 7

FIGURE 4 Colocalization of IL-13 with transferrin receptor (TfnRc)+ vesicles in platelet activating factor (PAF)-stimulated eosinophils. Human eosinophils were stimulated and labeled as described in Figure 1 with anti-IL-13 (red) and anti-TfnRc (green). Scale bar represents 5 µm. Data are shown from a representative donor. Means are indicated by red lines. Error bars indicate SD.**P < 0.01

detected in eosinophils.28 Understanding the mechanisms for IL-9 and AUTHORSHIP IL-13 release will contribute to our understanding of cytokine secre- tion from these cells. Coordinated efforts may link the mechanistic S.A. designed, performed, and analyzed experiments, and wrote the understanding of cytokine release with clinical practice in treating original draft of the manuscript; N.F., O.S., and M.S. performed and patients with eosinophilic asthma. Our current findings combined with analyzed experiments and reviewed and revised the manuscript; N.T. future studies will provide insight into potential targets for therapeutic and S.A. designed the image analysis protocols; and P.L. supervised the intervention. study, designed experiments, and edited the manuscript. 8 ALMAS ET AL.

A B IL-9 and CD63 colocalization C IL-9 and TfnRc colocalization

✱✱ ✱✱

✱✱ ✱✱ ✱✱ 5000 ✱✱ 0.8 0.8

4000 ✱✱ D1 0.6 0.6 D2 3000 D3 0.4 0.4 2000 D4

D5 0.2 0.2 1000

0 0.0 0.0 Pearson's correlation coefficient Pearson's correlation coefficient

IL-9 immunofluorescence intensity 0 5 15 60 051560 051560

D E IL-13 and CD63 colocalization F IL-13 and TfnRc colocalization ✱✱

✱✱

✱✱ ✱✱ 5000 ✱✱ 0.8 0.8

4000 0.6 0.6 3000 0.4 0.4 2000

0.2 0.2 1000

0 0.0 0.0 Pearson's correlation coefficient 0 5 15 60 Pearson's correlation coefficient IL-13 immunofluorescence intensity 051560 051560 PAF stimulation (min)

FIGURE 5 Intensity and colocalization of IL-9 and IL-13 with CD63+ granules and transferrin receptor (TfnRc)+ vesicles demonstrates a similar pattern of release from all donors. Immunofluorescence (A and D) and Pearson’s correlation coefficients (B, C, E,andF) were measured on Volocity software from grouped donor data. Each plot had a minimum of three donors, from which 50 eosinophils were analyzed and averaged for each platelet activating factor (PAF) stimulation time points. Donor numbers represent replicates for each data set and are not assigned to an individual. Mean values of 250 measurements from each donor at each time point are shown by a data point. Error bars indicate SEM.*P < 0.05 and **P < 0.01

ACKNOWLEDGMENTS DISCLOSURES

The authors acknowledge all blood donors, Dr. Vivek Gandhi for his The authors declare no conflicts of interest. support in optimizing techniques, and Eduardo Reyes-Serratos for his assistance with experimental protocols. The authors also thank ORCID Greg Plummer and Dr. Stephen Ogg (Cell Imaging Centre, University of Alberta) for their expertise with super-resolution microscopy, Paige Lacy https://orcid.org/0000-0001-8885-6011 image analysis, and image presentation. They also acknowledge

Dr. Xuejun Sun at the Cell Imaging Facility (Cross Cancer Institute, REFERENCES University of Alberta) for immense help in the development and use of the Imaris protocol. Experiments were performed at the 1. Hogan SP, Rosenberg HF, Moqbel R, et al. Eosinophils: biological prop- erties and role in health and disease. Clin Exp Allergy. 2008;38:709-750. University of Alberta Faculty of Medicine & Dentistry Cell Imag- 2. Wenzel SE. Asthma phenotypes: the evolution from clinical to molecu- ing Centre, which receives financial support from the Faculty of lar approaches. Nat Med. 2012;18:716-725. Medicine & Dentistry, the Department of Medical Microbiology and 3. Wenzel SE, Schwartz LB, Langmack EL, et al. Evidence that severe Immunology, and Canada Foundation for Innovation (CFI) awards asthma can be divided pathologically into two inflammatory subtypes with distinct physiologic and clinical characteristics. Am J Respir Crit to contributing investigators. S.A. was supported by the Natural Care Med. 1999;160:1001-1008. Sciences and Engineering Research Council of Canada (NSERC) 4. Bousquet J, Chanez P, Lacoste JY, et al. Eosinophilic inflammation in Undergraduate Summer Research Awards (USRAs), AllerGen Summer asthma. NEnglJMed. 1990;323:1033-1039. Studentships, Alberta Innovates Summer Research Studentship, and 5. Davoine F, Lacy P. Eosinophil cytokines, chemokines, and growth fac- CSACI Summer Studentship. O.S. was supported by NSERC USRA. tors: emerging roles in immunity. Front Immunol. 2014;5:570. 6. Dyer KD, Percopo CM, Xie Z, et al. Mouse and human eosinophils N.F. was supported by NSERC and the Lung Association of Alberta & degranulate in response to platelet-activating factor (PAF) and NWT. This project was supported by P.L.’s NSERC Discovery Grant LysoPAF via a PAF-receptor-independent mechanism: evidence for a (2015-2021). novel receptor. J Immunol. 2010;184:6327-6334. ALMAS ET AL. 9

A

B C

D E

PAF stimulation (min)

FIGURE 6 Greater proportion of intracellular IL-9 in transferrin receptor (TfnRc)+ endosomes compared to CD63+ crystalloid granules. Intracellular IL-9, CD63+ granules, and TfnRc+ endosomes were defined and quantified on a single cell basis as described in Supplementary Figure S4. IL-9-defined spots co-located within CD63+ granules or TfnRc+ endosomes were then quantified on a per cell basis. A number of IL- 9-defined spots (A), percentage of total IL-9 within CD63+ granules (B) or TfnRc+ endosomes (D), and percentages of CD63+ granules or TfnRc+ endosomes containing IL-9 (C, E) were measured per cell. At each time point, 50 eosinophils were analyzed from an individual donor. Means are represented as red lines and error bars indicate SD. Kruskal-Wallis test followed by Dunn’s multiple comparisons test was performed. *P < 0.05 and **P < 0.01 10 ALMAS ET AL.

7. Zhou Y, McLane M, Levitt RC. Th2 cytokines and asthma—interleukin- 21. Reefman E, Kay JG, Wood SM, et al. Cytokine secretion is distinct from 9 as a therapeutic target for asthma. Respir Res. 2001;2:1-5. secretion of cytotoxic granules in NK cells. J Immunol. 2010;184:4852- 8. Sehra S, Yao W, Nguyen ET, et al. TH9 cells are required for tissue mast 4862. cell accumulation during allergic inflammation. J Allergy Clin Immunol. 22. Hulse RE, Swenson WG, Kunkler PE, et al. Monomeric IgG is neuro- 2015;136:433-440.e1. protective via enhancing microglial recycling endocytosis and TNF-𝛼. 9. Schmid-Grendelmeier P, Altznauer F, Fischer B, et al. Eosinophils J Neurosci. 2008;28:12199-12211. express functional IL-13 in eosinophilic inflammatory diseases. J 23. Son K, Mukherjee M, McIntyre BAS, et al. Improved recovery of func- Immunol. 2002;169:1021-1027. tionally active eosinophils and neutrophils using novel immunomag- 10. Spencer LA, Szela CT, Perez SAC, et al. Human eosinophils constitu- netic technology. J Immunol Methods. 2017;449:44-55. tively express multiple Th1, Th2, and immunoregulatory cytokines that 24. Dunn KW, Kamocka MM, McDonald JH. A practical guide to evalu- are secreted rapidly and differentially. J Leukoc Biol. 2009;85:117-123. ating colocalization in biological microscopy. AmJPhysiol,CellPhysiol. 11. Lacy P, Stow JL. Cytokine release from innate immune cells: associa- 2011;300:C723-C742. tion with diverse membrane trafficking pathways. Blood. 2011;118:9- 25. Adler J, Parmryd I. Quantifying colocalization by correlation: the Pear- 18. son correlation coefficient is superior to the Mander’s overlap coeffi- 12. Stanley AC, Lacy P. Pathways for cytokine secretion. Physiology cient. Cytometry A. 2010;77:733-742. (Bethesda). 2010;25:218-229. 26. Kim JD, Willetts L, Ochkur S, et al. An essential role for Rab27a GTPase 13. Stenmark H. Rab GTPases as coordinators of vesicle traffic. Nat Rev in eosinophil exocytosis. J Leukoc Biol. 2013;94:1265-1274. Mol Cell Biol. 2009;10:513-525. 27. Lacy P, Willetts L, Kim JD, et al. Agonist activation of f-actin-mediated 14. Jahn R, Scheller RH. SNAREs—engines for membrane fusion. Nat Rev eosinophil shape change and mediator release is dependent on Rac2. Mol Cell Biol. 2006;7:631-643. Int Arch Allergy Immunol. 2011;156:137-147. 15. Logan MR, Lacy P, Bablitz B, et al. Expression of eosinophil target 28. Lacy P, Mahmudi-Azer S, Bablitz B, et al. Rapid mobilization of intra- SNAREs as potential cognate receptors for vesicle-associated mem- cellularly stored RANTES in response to interferon-gamma in human brane protein-2 in exocytosis. J Allergy Clin Immunol. 2002;109:299- eosinophils. Blood. 1999;94:23-32. 306. 16. Logan MR, Odemuyiwa SO, Moqbel R. Understanding exocytosis in immune and inflammatory cells: the molecular basis of mediator secre- SUPPORTING INFORMATION tion. J Allergy Clin Immunol. 2003;111:923-932. quiz 933. 17. Logan MR, Lacy P, Odemuyiwa SO, et al. A critical role for vesicle- Additional information may be found online in the Supporting Informa- associated membrane protein-7 in exocytosis from human eosinophils and neutrophils. Allergy. 2006;61:777-784. tion section at the end of the article. 18. Willetts L, Felix LC, Jacobsen EA, et al. Vesicle-associated membrane protein 7-mediated eosinophil degranulation promotes allergic airway How to cite this article: Almas S, Fayad N, Srivastava O, Sid- inflammation in mice. Commun Biol. 2018;1:83. 19. Shimbara A, Christodoulopoulos P, Soussi-Gounni A, et al. IL-9 and its dique M, Touret N, Lacy P. Cytokine trafficking of IL-9 and IL- receptor in allergic and nonallergic lung disease: increased expression 13 through TfnRc+ vesicles in activated human eosinophils. J in asthma. J Allergy Clin Immunol. 2000;105:108-115. Leukoc Biol. 2020;1–10. https://doi.org/10.1002/JLB.2MA0820 20. Woerly G, Lacy P, Younes AB, et al. Human eosinophils express and -320RR release IL-13 following CD28-dependent activation. J Leukoc Biol. 2002;72:769-779.