Eotaxin (CCL11) and Eotaxin-2 (CCL24) Induce Recruitment of Eosinophils, Basophils, Neutrophils, and Macrophages As Well As Features of Early- and Late-Phase Allergic Reactions Following Cutaneous Injection in Human Atopic and Nonatopic Volunteers¹

There is a well-known association between the clinical features of atopic asthma and rhinitis and the numbers of eosinophils at the site of allergic inflammation (1). Eosinophil accumulation is partially dependent on the local generation of soluble chemoattractant chemical signals, prominent among which are a family of chemotactic cytokines: chemokines. Chemokines are 8- to 10-kDa proteins that exhibit similar tertiary structures consisting of three β-pleated sheets and a C-terminal α helix. Chemokines are subdivided into four subfamilies based on the arrangement of four conserved cysteine residues within the protein (2). In the CC subfamily the first two cysteines are adjacent. Many members of the CC family have been described, which are variously active on T lymphocytes, eosinophils, basophils, and monocytes and play important roles in the regulation of many inflammatory diseases.

Guinea pig allergic airways inflammation is associated with the selective recruitment of eosinophils, and using HPLC and microsequencing, the activity responsible for this recruitment was identified as a novel chemokine, named eotaxin (3). Eotaxin causes the rapid accumulation of eosinophils in guinea pig lung (4) and has been identified as an important eosinophil chemoattractant in both mouse and man (5, 6). Eotaxin was initially found to be selective for eosinophils (7), but has now been shown to be chemotactic for basophils (8), Th2 lymphocytes (9), and tryptase-chymase mast cells (10). Eotaxin has important physiological actions beyond chemotaxis, including the induction of eosinophil degranulation (11) and IgE-independent degranulation of basophils (12).

A functional homologue of eotaxin has been described and named eotaxin-2 (13). This chemokine acts on the same receptor as eotaxin and shows the same cellular selectivity, although it is only 39% homologous at the protein level (14). A standardized chemokine nomenclature has recently been introduced, with eotaxin being termed CCL11, and eotaxin-2 being called CCL24. Together these proteins show a potency and selectivity for eosinophils greater than that of any known chemoattractant (15). The exact interplay of these two chemokines in vivo has yet to be elucidated. An early indication of differing roles comes from work examining the allergen-induced cutaneous late phase reaction (16), suggesting that eotaxin has a role in the early recruitment of eosinophils, while eotaxin-2 appears to be involved in eosinophil infiltration at a later time point.

An eotaxin receptor designated CCR3 has been identified and shown to be present in high numbers on human eosinophils (6, 17). Until recently it was believed that both eotaxin and eotaxin-2 only signal via CCR3. However, eotaxin has also been reported to be both an agonist (18) and an antagonist for CCR2 and an agonist for CCR5 (19).

Verification of the biological activity of chemokines in man is difficult, and to date only the CC chemokines RANTES and macrophage inflammatory protein-1α (MIP-1α)³ have been injected into human volunteers. Following RANTES injection a dose- and time-dependent recruitment of eosinophils, CD45RO⁺ cells, and CD3⁺ cells was observed (20). Injection of MIP-1α induced significant infiltration of monocytes, lymphocytes, eosinophils, and neutrophils (21). Our present study addresses the biological activities of eotaxin and eotaxin-2 in vivo in man by injecting these chemokines into the skin of atopic and nonatopic volunteers.

This study was approved by the ethics committee of the Royal Brompton and Harefield National Health Service trust, and all volunteers gave informed consent. Nine atopic and nine nonatopic volunteers were recruited, aged 18–47 yr. Atopy was defined as the presence of a history of seasonal and/or perennial allergic rhinitis and/or asthma with a positive skin prick test to an aeroallergen in the presence of a positive histamine and negative vehicle control. The absence of atopy in the normal volunteers was described as a negative history plus a negative skin prick testing. Any medication that could interfere with the early or late phase response was discontinued for an appropriate length of time before study entry.

Eotaxin for injection was prepared by Leukosite (Cambridge, MA), using a baculovirus expression system. The chemically synthesized eotaxin-2 for injection was a gift from Dr I. Clark-Lewis (Vancouver, Canada). Both chemokines were shown to be pure proteins by HPLC, and endotoxin testing demonstrated <0.01 Eu LPS/10 μg of chemokine. The chemokines were aliquoted under sterile conditions in the pharmacy of the Royal Brompton Hospital and stored at −80°C prior to use. All chemokine injections were made up under aseptic conditions to a final volume of 50 μL with sterile isotonic saline and used on the same day. Lyophilized allergen extracts of the relevant allergens were obtained from ALK (Aquagen; Horsholm, Denmark) and one batch was used throughout.

The first three atopic and nonatopic subjects participated in a dose-ranging study to determine the doses of eotaxin and eotaxin-2 that produced a significant eosinophil infiltrate without adverse effects. Initially eotaxin was injected into two atopic and two nonatopic volunteers. The following were injected intradermally into the forearm extensor surface: 0.01, 0.1, and 1 μg eotaxin and an equal volume of diluent (isotonic saline). The magnitude of any cutaneous reaction was measured after 15, 30, and 60 min and 3 and 6 h. After 6 h all injection sites were biopsied with a 4-mm punch biopsy under 1% plain lignocaine anesthesia. We then repeated the dose-ranging study in one atopic and one nonatopic volunteer, increasing the maximum dose to 10 μg. Following this, we repeated the dose-ranging study with eotaxin-2 in one atopic and one nonatopic volunteer who had previously been injected with eotaxin.

We then performed a time-course study with six atopic and six nonatopic volunteers using the 10-μg dose of eotaxin and eotaxin-2. All volunteers were injected with diluent at one site, with 10 μg eotaxin at three sites, and with 10 μg eotaxin-2 at three sites. The atopic volunteers were also injected with 30 biological units of a relevant allergen at one site. All injections were placed in the extensor aspect of the forearm. The magnitude and the nature of the cutaneous response at the injection sites were measured after 15, 30, and 60 min and 3, 6, and 24 h. Skin biopsies were taken from the eotaxin and eotaxin-2 injection sites after 1, 6, and 24 h. The diluent and allergen injection sites were biopsied after 24 h.

All skin biopsy specimens were divided with a scalpel into two equal parts. Half the biopsy was immediately mounted in OCT (optimum cutter temperature) medium (Bayer, Basingstoke, U.K.) on a cork tile before being snap-frozen in isopentane (BDH, Leicester, U.K.) cooled to nearly freezing with liquid nitrogen. The other half was fixed in 4% paraformaldehyde (BDH) before mounting in OCT. Sections (8 μm) of the biopsies were captured onto 0.1% poly-l-lysine-coated slides (BDH) before staining.

Eosinophil counts were performed on Congo Red-stained sections as described previously (22, 23). To further characterize the inflammatory cell infiltrate, immunohistochemistry was performed on the sections using a modification of the alkaline phosphatase-anti-alkaline phosphatase technique as described previously (16). Briefly, mouse mAbs against CD3, CD4, CD8, CD68, mast cell tryptase, and neutrophil elastase (all from DAKO, Ely, U.K.) and BB1 (gift from Dr. A. F. Walls, Southampton, U.K.) were used as primary Abs under appropriate conditions. Following incubation with a rabbit anti-mouse secondary Ab (DAKO), alkaline phosphatase-mouse anti-alkaline phosphatase (DAKO) was added before incubation with Fast Red solution (Sigma-Aldrich, Gillingham, U.K.). Appropriate isotype controls were included in the staining runs.

The number of cells per square millimeter was determined by counting whole sections in a blinded fashion at ×40 magnification with an eyepiece graticule using an Olympus BH-2 microscope (New Hyde Park, NY).

To examine neutrophil surface CCR3 expression, whole blood was obtained by venipuncture, placed into EDTA-containing Vacutainers (BD Biosciences, Mountain View, CA), and processed immediately. One hundred microliters of whole blood was incubated with saturating concentrations of CD16 FITC (BD Biosciences) and CCR3 PE (R&D Systems, Abingdon, U.K.) or a PE-labeled isotype (R&D Systems). RBC were removed with FACSLyse (BD Biosciences) before fixing in 0.5% formaldehyde (BDH) in isoton (Beckman Coulter, High Wycombe, U.K.). Samples were analyzed on a BD FACScan using CellQuest software (BD Biosciences). CCR3 surface staining on neutrophils was measured by first identifying granulocytes by their physical characteristics before analyzing the CD16-positive neutrophil population. Color compensation was performed before each analysis.

To examine surface and intracellular CCR3 expression, neutrophils were purified from 50 ml peripheral blood collected into a heparinized syringe. After dextran (Life Technologies, Paisley, U.K.) sedimentation the granulocyte pellet was collected following centrifugation across a density gradient (Histopaque 1.077; Sigma-Aldrich). Following RBC lysis by osmotic shock the granulocytes were stained as described above for cell surface CCR3. The level of intracellular CCR3 was measured using saponin (Sigma-Aldrich) to allow intracellular staining. The samples were run on a BD Biosciences FACScan and were analyzed using CellQuest software as described above.

The GAFS assay was performed as described previously (24). Briefly, 10-μl aliquots of agonist, IL-8 (CXCL8), or eotaxin (both from PeproTech, Totam Biologicals, Northampton, U.K.) or buffer were placed in each 1.2-ml polypropylene cluster tube. Blood was collected from four volunteers, and 90 μl was immediately added to each aliquot of agonist or buffer in the cluster tubes. The samples were mixed gently and incubated at 37°C for 4 min in a shaking water bath. The tubes were then placed on ice, and 250 μl ice-cold fixative solution (1× Cellfix; BD Biosciences; diluted 1/4 in PBS) was added to each tube for 1 min. The samples were then pipetted into cold NH₄Cl and left on ice for 20 min for RBC lysis. Samples were analyzed according to the GAFS protocol using a BD FACSCalibur and CellQuest software.

The effects of eotaxin and IL-8 on CD11b expression on neutrophils were determined as described previously (21). Briefly, 4 ml blood was collected from four volunteers. Blood was prewarmed at 37°C for 5 min before adding increasing concentrations of eotaxin or IL-8 (0.01–1000 nM). After 15 min the tubes were placed into an ice bath, and 1 ml cold PBS containing 2% FCS and 0.2% sodium azide was added. After washing, the cells were stained with saturating concentrations of CD16 FITC and CD11b PE (R&D Systems) or isotype PE (R&D Systems). The level of CD11b expression was then measured on the CD16-positive granulocyte population as described above.

Statistical analysis was performed using the Wilcoxon signed rank test for intragroup comparisons and the Mann-Whitney test for between-group comparisons. A p < 0.05 was considered significant. Data are expressed as the mean ± SEM.

Intracutaneous injection of either eotaxin or eotaxin-2 produced a wheal and flare reaction that peaked at 15 min. The reaction was demonstrable in a dose-dependent fashion at higher concentrations in the dose-ranging study (mean diameters of 7.00 and 17.00 mm following injection of 1 μg eotaxin and eotaxin-2, respectively, and of 17.75 and 25.35 mm following injection of the 10-μg doses). In the time-course study (six atopic and six nonatopic volunteers) there were no differences in the size or intensity of the wheal and flare response in atopic and nonatopics. In the majority of volunteers the cutaneous reaction resolved within 1 h (Fig. 1). However, one nonatopic volunteer developed a cutaneous late phase reaction (mean diameter, 11.5 mm) 6 h after the injection of eotaxin, and five (three atopic and two nonatopic) developed a cutaneous late phase reaction following the injection of eotaxin-2 (mean diameter, 18.0 mm; Fig. 1). The one subject who developed a late cutaneous reaction following the injection of eotaxin also developed a reaction following the injection of eotaxin-2. Thus, a total of five of 12 subjects developed late cutaneous reactions following injection of the chemokines. However, these late phase reactions were smaller than those observed with 30 biological units of allergen (mean diameter, 51.4 mm). No other adverse reactions were observed with eotaxin and eotaxin-2 over the dose range studied (i.e., 0.01–10 μg).

Intracutaneous injection of eotaxin and eotaxin-2 induced a dose-dependent eosinophil infiltrate from 0.01–10.00 μg (Fig. 2). The 10-μg dose was used for the subsequent time-course study in which both chemokines were injected at three sites, and skin biopsies were taken after 1, 6, and 24 h. Eotaxin produced a significant eosinophil infiltrate in atopic and nonatopic subjects within 1 h of injection compared with diluent (Fig. 3). This increased at the 6 and 24 h points, at which time the tissue eosinophilia was of a similar magnitude to that observed following allergen injection in the atopics. Eotaxin-2, at the same concentration, produced a significant, but proportionally smaller, eosinophil infiltrate at all three time points (Fig. 3). Using Congo Red staining there was microscopic evidence of eosinophil degranulation 6 h after the injection of eotaxin, but not with eotaxin-2 (Fig. 4).

There was no difference in the speed of eosinophil infiltration between atopics and nonatopics following the injection of both eotaxin and eotaxin-2 (Fig. 3). As there were no differences for any of the other inflammatory cell types that accumulated, all subsequent data are presented as a combination of both atopic and nonatopic volunteers.

Both eotaxin and eotaxin-2 induced a small, but significant, infiltrate of BB1⁺ basophils at all three time points (Fig. 5). The basophil numbers were smaller than those observed following allergen injection in the atopic subjects.

Injection of both eotaxin and eotaxin-2 also induced a significant elastase-positive neutrophil infiltrate at all three time points, comparable to that observed in the atopic individuals following the injection of allergen (Fig. 5). Neutrophil recruitment occurred in a dose-dependent fashion (Fig. 2).

Intradermal injection of both chemokines induced a significant CD68⁺ macrophage infiltration at the 24 h point, and the magnitude of the infiltrate was comparable to that seen 24 h after injection of allergen in the atopic volunteers (Fig. 5).

There was no significant increase in the number of CD3⁺ (Fig. 5), CD4⁺, or CD8⁺ cells (results not shown) following the injection of eotaxin or eotaxin-2.

There was a decrease in tryptase-positive mast cell numbers following chemokine injection; this reached significance 24 h after the injection of eotaxin-2 (Fig. 6).

Neutrophils are generally thought to be CCR3 negative, yet were recruited to skin sites injected with eotaxin or eotaxin-2, but not with diluent. To determine whether neutrophils expressed surface CCR3 that was down-regulated during purification, we examined neutrophils within whole blood immediately after venipuncture. We were able to consistently demonstrate low level cell surface expression of CCR3 using flow cytometry (Fig. 7 A) with a mean standard mean fluorescence of 2.05 ± 0.47 (n = 6) in both atopic and nonatopic individuals.

Flow cytometric analysis of neutrophils following purification demonstrated no cell surface expression of CCR3. However, following saponization there was a strong intracellular signal (standard mean fluorescence, 58.66; Fig. 7 B).

After demonstrating the presence of surface CCR3 on neutrophils, we attempted to demonstrate functionality in whole blood assays. Analysis of freshly collected peripheral blood samples following incubation with increasing doses of eotaxin failed to show any neutrophil shape change in the GAFS assay. In contrast, IL-8 was effective at inducing neutrophil shape change at physiological concentrations (Fig. 8 A).

We also attempted to demonstrate functionality by examining whether incubation of whole blood with increasing concentrations of eotaxin was able to up-regulate the expression of CD11b. There was no evidence of an increase in CD11b on neutrophils after incubation of whole blood with up to 1000 nM eotaxin. However, incubation with IL-8 consistently induced CD11b up-regulation on neutrophils (Fig. 8 B).

We have demonstrated for the first time that both eotaxin and eotaxin-2 induce a significant infiltration of eosinophils and basophils in human volunteers following intradermal injection. In contrast to the findings following intradermal injection of RANTES (20), there were no differences in the rate of inflammatory cell infiltration between atopic and nonatopic individuals following chemokine injection. This lack of differential response may be due to the increased chemotactic potency of eotaxin and eotaxin-2 compared with that of RANTES (15). An alternative explanation is that, as there was no significant eosinophil accumulation until 24 h after the injection of RANTES in normal volunteers, this infiltration was secondary to an indirect signal, unlike the early infiltration seen following the injection of eotaxin and eotaxin-2. Previous investigators reported that eotaxin was capable of inducing eosinophil degranulation and eosinophil-derived neurotoxin release (11). At the 10-μg dose, intradermal injection of eotaxin produced histological evidence of eosinophil degranulation with Congo Red staining, in keeping with the previous in vitro findings.

The magnitude of the basophil infiltration was smaller than that observed with allergen (in atopic subjects). This can be explained by the fact that allergen injection induces the release and production not only of eotaxin, but also of many other CC chemokines believed to be important in basophil recruitment, such as monocyte chemoattractant protein-4, MIP-1α, and RANTES (25), therefore producing a stronger signal than the injection of eotaxin alone.

The production of an acute wheal and flare and the progression to a late phase reaction following injection were unexpected and were not seen with the injection of RANTES (20) or MIP-1α (21). Tryptase-chymase mast cells are known to express CCR3 (10), and basophils can be induced to degranulate by eotaxin via IgE-independent mechanisms (12). It is therefore reasonable to speculate that eotaxin and eotaxin-2 induce mast cell degranulation via CCR3. Evidence to support this comes from our finding that mast cell numbers decrease up to the 24-h point following chemokine injection. The generation of the cutaneous late phase reaction following chemokine injection is likely to be a sequelae of mast cell degranulation. It is possible that basophils play a role in its development, as eotaxin can induce basophil degranulation (12). However, there was no correlation between basophil recruitment at 6 h and the magnitude of the late phase reaction (results not shown).

The recent report that eotaxin is a ligand for CCR5 (19) may explain the macrophage accumulation that occurred at 24 h. However, infiltrating leukocytes may often be sources of pro-inflammatory chemokines, and it is more likely that the delayed recruitment of macrophages into the tissue reflects secondary processes dependent upon the induction of other chemokines rather than being dependent directly upon the action of eotaxin or eotaxin-2.

The early and significant neutrophil infiltration was unexpected. Given the recent discovery of CCR1 on neutrophils (21), we investigated whether neutrophils in whole blood expressed functional CCR3. Bonnechi et al. (26) have previously demonstrated CCR3 mRNA expression in neutrophils following incubation with IFN-γ, although others have claimed that this was due to eosinophil contamination (27). We were able to demonstrate low level surface expression of CCR3 on neutrophils, which was lost upon purification. A similar phenomenon was seen with CCR1 (21), and here the authors attributed it to chemokine release by platelets activated during the purification process.

Although we did demonstrate CCR3 expression on neutrophils, we found no evidence that it was functional. The GAFS assay is a sensitive marker for response to chemokines (15), and we were unable to demonstrate any neutrophil shape change in whole blood following incubation with physiological concentrations of eotaxin. In contrast, incubation with IL-8 induced significant neutrophil shape change. It is impossible to accurately predict the local chemokine concentration following intradermal injection. We therefore performed the GAFS assay at presumed physiological concentrations of eotaxin; it is possible that higher concentrations of eotaxin were generated in the skin following chemokine injection.

Lee et al. (21) demonstrated functionality of CCR1 on neutrophils by incubating whole blood with MIP-1α and demonstrating CD11b up-regulation on neutrophils. Following a similar protocol we saw no evidence of CD11b up-regulation on neutrophils when incubating whole blood with eotaxin. Incubation with IL-8 consistently induced up-regulation of CD11b as expected.

The rate of neutrophil accumulation following eotaxin and eotaxin-2 injection suggests that it may be the result of a direct chemotactic signal. Our inability to demonstrate any functional CCR3 response suggests that neutrophil accumulation may occur secondary to eotaxin-induced mast cell degranulation with release of neutrophil-specific chemokines such as IL-8. However, there was no correlation between neutrophil accumulation after 1 h and the magnitude of the early cutaneous reaction following the injection of either eotaxin or eotaxin-2 (results not shown). The importance of mast cells in neutrophil accumulation is supported by the findings that human mast cell tryptase induces infiltration of neutrophils in guinea pig skin (28) and also stimulates the release of an IL-8-dependent neutrophil chemotactic activity from HUVEC (29). Work on a mouse model of bullous pemphigoid supports the hypothesis, as mast cells play an essential role in neutrophil recruitment during subepidermal blister formation (30).

We were unable to demonstrate any increase in T lymphocyte numbers following the injection of eotaxin and eotaxin-2. However, because only 1% of peripheral blood lymphocytes express CCR3 (9), it is possible that the changes were too subtle to be identified by immunohistochemistry.

In conclusion, we have demonstrated in vivo chemotactic activity of eotaxin and eotaxin-2 for both eosinophils and basophils in man. We have also demonstrated that eotaxin and eotaxin-2 may play significant roles in perpetuating allergic inflammation by inducing mast cell degranulation and subsequent development of the late cutaneous reaction.