Eosinophils are involved in several inflammatory processes including allergic inflammation. It has been shown that eosinophil functions may be regulated by activating or inhibitory receptors. Hypoxia is a feature of inflamed tissues and has recently been shown to regulate eosinophil viability and pro-angiogenic potential. In this study, the effect of hypoxia and GM-CSF on the inhibitory receptor CD300a in human peripheral blood eosinophils was investigated.
CD300a expression on eosinophils was analyzed by flow cytometry and evaluated by immuno-fluorescence; mRNA levels were evaluated by RT-PCR.
An increase in the expression of CD300a was observed in hypoxic eosinophils compared to the normoxic ones. GM-CSF strongly induced CD300a increase also after 3 h in culture. In addition, hypoxia augmented mRNA levels of CD300a. Inhibition of hypoxia-inducible factor (HIF)-1 abolished the hypoxia-/GM-CSF-induced CD300a increase.
CD300a expression is up-regulated by hypoxia, and GM-CSF where HIF-1 might play an important role. These results are important for the understanding of eosinophils behavior in inflamed tissue and suggest a new effect on their function in allergic inflammation. Taken together our data point out CD300a as a novel target for the treatment of allergy.
Eosinophil infiltration characterizes a variety of allergic inflamed tissues where their reactivity is orchestrated by an interplay between stimulatory and inhibitory signals of various types. We previously found that eosinophil functions are regulated by receptors known to be expressed mainly by T cells and NK cells such as the inhibitory receptors CD300a and Siglec7 from the IgG-like superfamily and stimulatory receptors such as CD48 and 2B4 belonging to the CD2 family [1, 2].
In allergic inflamed tissue, high levels of granulocyte–macro-phage colony-stimulating factor (GM-CSF) are present that recruit, sustain, and activate the eosinophils. Hypoxia, due to an inadequate blood supply and high consumption of oxygen by the infiltrated cells, is also often evident. Inflammatory cells, such as the eosinophils, must be able to maintain viability and function under hypoxic conditions. Hypoxia causes an alteration in gene expression due to the up-regulation of a number of transcription factors, among them the hypoxia-inducible factor (HIF)-1 that functions as a master regulator of oxygen homeostasis. Recently we demonstrated that eosinophils respond to hypoxia by HIF-1 up-regulation and by increasing their viability and pro-angiogenic potential (3). Indeed, eosinophil prolonged survival and activation are ‘hallmarks’ of allergic diseases. In the present study, the effect of hypoxia on CD300a expression was investigated on human peripheral blood eosinophils cultured in normoxic or hypoxic conditions with or without GM-CSF. The inhibitory receptor CD300a, containing four immunoreceptor tyrosine-based inhibitory motifs (ITIM) (4, 5), is of particular interest to us as a possible immunopharmacological tool to down-regulate allergic responses. In fact, CD300a was demonstrated to effectively down-regulate eosinophil functions in vitro and to inhibit inflammation in a model of eosinophilic allergic murine asthma (6).
Material and methods
Eosinophils were purified as previously described  from the peripheral blood of untreated, mildly atopic volunteers (blood eosinophil levels 5–10%), who were asymptomatic and therefore not taking any drug for their condition, and on the day of blood donating, any other drug. Written informed consent was obtained according to the guidelines of the Hadassah-Hebrew University Human Experimentation Helsinki Committee. Eosinophils collected at a purity of >98% (Kimura staining), with a viability of >98% (trypan blue staining) were re-suspended (5 × 106 or 106 cells/ml) in culture medium consisting of RPMI 1640 supplemented with l-glutamine (300 mg/l), 10% heat-inactivated FCS, and penicillin–streptomycin solution (100 u/ml) (Biological Industries, Beit Haemek, Israel).
Eosinophil culture in hypoxic or normoxic condition
Eosinophils were cultured in 96-well u-shaped plates (Nunc, Rochester, NY, USA) in 200 μl of medium alone or supplemented with 20 ng/ml of GM-CSF, IL-3, or IL-5 (PeproTech, Rocky Hill, NJ, USA) according to the experiment. For experiments on HIF-1-binding inhibition, cells were cultured at the indicated times with or without 10 nM echinomycin (Calbiochem, Damstadt, Germany). For experiments with the different inhibitors, cells were cultured at the indicated times with or without 10 μM PD98059 (Calbiochem), brefeldinA or cycloheximide (Sigma Aldrich, Rehovot, Israel). For experiments in hypoxia, plates were placed in a closed humidified chamber at 37°C, with a continuous flow of gas mixture of 95% N2 and 5% CO2. The oxygen percentage in the medium was monitored with a dissolved oxygen meter before each time point, checked and found to be ≤3% (MettlerToledo, Schwerzenbach, Switzerland). At this time point, eosinophils were added to the wells. All the assays were performed on eosinophils cultured for a further 1 h, to permit full re-equilibration of the oxygen in the chamber (<3%). For experiments in normoxia, plates were placed at 37°C in a humidified incubator with 5% CO2 and air.
Flow cytometry analysis
For flow cytometry analysis, eosinophils (106 cells/ml) were incubated for 10 min on ice in Hanks balanced salt solution supplemented with 5% goat serum and 0.1% bovine serum albumin (Biological Industries, Kibbutz Beit Haemek, Israel). Eosinophils were then incubated with Abs recognizing CD300a (mAbs from clone #12) kindly provided by Prof. O. Mandelboim (Department of Immunology, The Hebrew University of Jerusalem, Jerusalem, Israel) , CD48 (e-bioscience, San Diego, CA, USA), or IgG1 isotype control all at a final concentration of 5 μg/ml followed by 40-min incubation on ice with goat anti-mouse FITC (1 : 500; ICN Pharmaceuticals, Costa Mesa, CA, USA).
After staining, the cells were analyzed on a Becton Dickinson FACSCalibur System (Becton Dickinson, San Jose, CA, USA). For each staining, at least 10 000 events were collected, and data analysis was performed using CellQuest software (BD, Mansfield, MA, USA).
Total RNA (15 μg) from fresh eosinophils was extracted using TriReagent (Sigma), and reverse transcription was performed using SuperScript First-Strand kit (Invitrogen, Carlsbad, CA, USA). The cDNA was amplified using primers for human VEGF (sense: 5′-CGC CGA CCA AGG AAA ACT C-3′, antisense: 5′-GGG TAC TGG ATG TCA GGT CTG C-3′, 569 bp), CD300a (sense: 5′-AGCATCCAGGAGGAAACTGA-3′, antisense: 5′- GAGGCCTCTGAGCAGCTATC-3′), or CD48 (sense: 5′-GGCTCTGGAATTGCTACT-3′, antisense: 5′-CTGTTCTGGAGCTCCTTTG-3′). Beta-actin was used as control. Thirty-five cycle-PCR products were visualized by Gel-red (Biotium, Hayward, CA, USA)-stained gel electrophoresis. Quantitative analysis was performed calculating the OD of the band relatively to the OD of the corresponding actin band.
For immunofluorescence (IF), 105 eosinophils per sample were blocked in 5% goat serum in Hanks balanced salt solution supplemented with 0.1% bovine serum albumin for 10 minutes on ice. Cells were then incubated with anti-CD300a antibody or IgG1 isotype control (5 μg/ml) for 1 h on ice and, after washing, with goat anti-mouse Alexa fluor 647 (1:500; Molecular Probes, Eugene, OR, USA) for 40 min on ice. Stained cells were fixed with 4% PFA and transferred to a glass slide precoated with poly-l-Lysine (0.01%; Sigma). Analysis was performed with a FluoView FV1000 confocal microscope (Olympus) at 10× and 60× magnifications.
Promoter region was identified using UCSC genome browser as 1000 nt upstream to the transcription start site (TSS). Promoter analysis was performed using Genomatix software (Munich/Germany; http://www.genomatix.de/), and a total of 169 promoters binding sites were found. The HIF-1 promoter was located at position-927 from the TSS.
The control and experimental groups were compared by paired t-test for the evaluation of significance (*P < 0.05; **P < 0.05). The data are expressed as mean ± standard error of measurement (SEM) of at least three independent experiments performed in triplicate. The KyPlot™ analysis tool-pack was used to perform the statistical analysis.
Results and discussion
We first evaluated the influence of the survival cytokines interleukin (IL)-3, interleukin-5, and GM-CSF on the receptor expression. Flow cytometry analysis shows that after 18 h, GM-CSF and IL-3 increased CD300a cell surface expression, while IL-5 had no effect (Fig. 1A). Short-term exposure (3 and 6 h) of eosinophils to GM-CSF also significantly increased CD300a expression (Fig. 1B). In an attempt to better understand this phenomenon, the protein synthesis inhibitor cycloheximide was used. Cycloheximide had no effect on the short-term regulation of CD300a (Fig. 1C). Also, Brefeldin A, a membrane transport inhibitor, did not inhibit membrane CD300a up-regulation. ERK1/2 phosphorylation is a feature of eosinophil activation, and it was found to be induced by GM-CSF. The specific ERK1/2 inhibitor PD98059 partially inhibited the enhancing effect of GM-CSF on CD300a at 18 h, while no effect was observed on the short-term increase of the receptor (3 h; Fig. 1C). It can be concluded from this evidence taken together that in eosinophils while the overnight increase of CD300a would imply de novo synthesis probably through ERK1/2 phosphorylation, the short-term increase would rather involve a rapid mechanism of receptor transport to the membrane. In previous works, it has also been shown that GM-CSF leads to the rapid up-regulation of CD300a expression on neutrophils membrane in an event which does not involve new protein synthesis . The notion that an intracellular pool of CD300 might be maintained for fast translocation suggests the importance of this receptor for the prompt modulation of eosinophil function. However, the mechanism has yet to be elucidated.
We then evaluated the effect of hypoxia in the presence or absence of GM-CSF on CD300a expression. Hypoxia increased CD300a expression after 18 h culture (MFI: normoxia 14.08 ± 2.94; hypoxia 17.98 ± 2.34; P = 0.045; n = 5). The addition of GM-CSF caused a further significant up-regulation in CD300a (normoxia P = 0.0027; hypoxia P = 0.0005; Fig. 1D, E). Interestingly, the expression of other eosinophil receptors, such as Siglec-7, CD48, and 2B4 was not significantly influenced by either GM-CSF or hypoxia (not shown).
To further study the effect of hypoxia, RT-PCR for CD300a genes was performed on mRNA from eosinophils cultured in normoxia or hypoxia with or without GM-CSF. Hypoxia increased mRNA levels of CD300a (Fig. 2A band a) and VEGF (positive control) but not CD48 (used as negative control). Surprisingly, in the CD300a product, an additional PCR product, shorter than the expected one, appeared under hypoxic conditions (Fig. 2A band b). Notably, GM-CSF had no effect on the two CD300a products. The two PCR products were sequenced and according to a BLAST analysis both corresponded to CD300a. The shorter one resulted as a spliced transcript missing the exon-containing sequences of two of the four CD300a ITIMs [1, 5] (Fig. 2B).
Hypoxia-inducible factor-1, a heterodimer of HIF-1α and HIF-1β, binds to the hypoxia responsive element (HRE) present in the promoter regions of hypoxia responsive genes, such as VEGF. Taking into consideration the responsiveness of the CD300a gene to hypoxia, HREs presence in the promoter was investigated using a genomatix promoter analysis on the first 1000 bases upstream of the transcription initiation site of the CD300a gene (Fig. 2C). Having identified different HREs, the involvement of HIF-1 in CD300a up-regulation under hypoxia/GM-CSF was investigated by testing the effect of the HIF-1 inhibitor echinomycin. Echinomycin partially inhibited CD300a increase (Fig. 2D). Particularly this inhibition was observed for the GM-CSF-driven increase in normoxia suggesting the involvement of HIF-1 in the GM-CSF-mediated regulation of CD300a. Notably, the short-term increase of CD300a was not affected by echinomycin (not shown). Importantly, in the presence of echinomycin, the shorter CD300a PCR product was significantly decreased (Fig. 2E).
We have previously shown a cooperation between GM-CSF and hypoxia in the up-regulation of HIF-1 . In light of our novel results, we suggest a role for the interplay between hypoxia and GM-CSF in the up-regulation of CD300a.
Notably, other studies have shown that hypoxia induced up-regulation of CD300a mRNA level in synovial fibroblasts and monocytes. In these studies, protein expression level was not characterized [8, 9]. Here, we provide the first evidences for CD300a regulation in hypoxic eosinophils at both the mRNA and protein level. Whether different eosinophil receptors can undergo a regulation similar to the one described for CD300a, and whether CD300a regulation takes place in a similar fashion on eosinophils in vivo, has yet to be further investigated.
We wish to thank Tgst Levi and Rivka Zach-Magid for the technical support, Dr. Tamar Kahan for the bioinformatics consultations and Dr. Saar Mizrahi for many helpful discussions.
A.H.N.B.E. designed the research, performed all experiments, analyzed data, and wrote the manuscript; L.K. designed the primers and critically revised the manuscript; M.B.Z. contributed to the conception and the interpretation of the data and critically revised the manuscript; and F.L.S. designed the research, received the grant supports, provided overall supervision, analyzed the data and edited the manuscript.
Conflicts of interest
The authors have no financial relationships to disclose.
This work was funded by Aimwell Charitable Trust (UK).