Chronic intestinal nematode infection induces Stat6-independent interleukin-5 production and causes eosinophilic inflammatory responses in mice

Abstract

The role of Stat6 (signal transducers and activators of transcription) in the recruitment and activation of eosinophils has been studied in detail in asthma and other allergic diseases. In this study, we demonstrated that eosinophil responses occur in a Stat6-independent manner in mice infected with the intestinal nematode, Nippostrongylus brasiliensis. Stat6-deficient (Stat6^–/–) mice cannot expel N. brasiliensis and establish chronic infections. Prominent blood and intestinal eosinophilia were induced after day 14 postinfection (p.i.) and maintained at this level in Stat6^–/– mice, whereas in wild-type mice eosinophil responses reached a peak on day 10 p.i. and declined thereafter. The introduction of a secondary infection of N. brasiliensis into wild-type mice induced rapid and exaggerated eosinophilia, whereas secondary infection in Stat6^–/– mice resulted in almost the same eosinophil responses as those of the primary infection, suggesting a lack of memory responses. Blood eosinophilia was also induced in Stat6^–/– mice implanted with N. brasiliensis in the small intestine, suggesting that intestinal exposure to parasitic antigen is sufficient to induce eosinophil responses. Furthermore, this prominent eosinophil response of Stat6^–/– mice after day 14 was closely associated with an increase of interleukin (IL)-5 production in serum and intestine. Neither IL-4 nor eotaxin were significantly induced in Stat6^–/– mice after infection with N. brasiliensis. We also found that mRNA for IL-5, GATA-3 and eosinophil peroxidase (EPO) are induced in the intestine of Stat6^–/– mice on day 14 p.i. Taken together, these results provide evidence for Stat6-independent IL-5 production and subsequent eosinophil responses after chronic infection with N. brasiliensis.

Introduction

Nippostrongylus brasiliensis, an intestinal nematode, induces strong T helper type 2 (Th2) responses, such as peripheral and tissue eosinophilia, mastocytosis and production of a high level of immunoglobulin E (IgE).¹ The primary cytokines driving the production of IgE and eosinophilia are interleukin (IL)-4 and IL-5, respectively. IL-4 and IL-5 are produced mainly by Th2 cells that develop via the IL-4 receptor-α (IL-4Rα)/Stat6 pathway. The role of Stat6 in the development of Th2 cells has been elegantly shown using N. brasiliensis-infected Stat6^–/– mice.² It was demonstrated that in Stat6^–/– mice infected with N. brasiliensis, Th2 cytokines (including IL-4 and IL-5) from splenic T cells of infected mice were reduced on day 8 postinfection (p.i.), implying a defective development of Th2 cells in these mice.²N. brasiliensis are not expelled in these Stat6^–/– mice, where they establish a chronic infection.^3,4

Eosinophilia in peripheral blood and tissue, observed during gastrointestinal infection with nematodes, is also a characteristic feature of airway-allergic diseases.⁵ It has been shown that Stat6 is required to develop airway eosinophilia and airway hyper-responsiveness in mice.^6,7 Stat6^–/– mice also failed to induce eosinophilic inflammation in an intestinal allergy model.⁸ Thus, Stat6 has been considered to play a vital role in the eosinophilic response, in addition to Th2 development. However, compelling evidence for the role of Stat6 in the eosinophil response has been recently reported using a chronic airway allergic model induced by fungal infection⁹ and an ovalbumin-inhalation challenge model of chronic asthma.¹⁰ Furthermore, it has been shown that nu/nu mice infected with Toxocara canis induce eosinophilia as a result of IL-5 production from CD4^– CD8^– T cells.^11,12 The role of Stat6 in eosinophilic inflammation and IL-5 production caused by intestinal nematode infection has not been fully delineated.

We herein examined the kinetics of eosinophil responses after N. brasiliensis infection in wild-type and Stat6^–/– mice, the latter as a model for chronic intestinal nematode infection. We found that chronic intestinal helminth infection induces eosinophil responses in tissue and peripheral blood in a Stat6-independent manner. The kinetics of eosinophil responses were dramatically different between wild-type and Stat6^–/– mice, including the responses to secondary infection. We also found that these eosinophil responses are induced by Stat6-independently produced IL-5 in infected tissue and intestine, and we highlighted the possibility that innate eosinophils produce IL-5 via the autocrine pathway in Stat6^–/– mice after infection with N. brasiliensis. The significance of the current findings, Stat6-independent IL-5 production and subsequent eosinophil responses, will be discussed in relation to the clinical situation.

Materials and methods

Mice

C57BL/6 mice were purchased from Charles River Japan, Inc. (Yokohama, Japan). Breeding pairs of Stat6^–/– mice on a C57BL/6 background were kindly provided by Professor S. Akira (Osaka University) and used as described previously.¹³ All mice were male and used at 8–10 weeks of age at the beginning of the experiments in this study. Four to five mice per group were used for each experiment. The experiments were performed conforming to the code of the Committee of Animal Experiment in Miyazaki Medical College.

Parasite infection

N. brasiliensis larvae were prepared as described previously.¹³ Mice were infected subcutaneously with 500 third-stage larvae (L3). Before secondary infection, mice were treated with 1 mg/kg body weight of pyrantel pamoate (Pfizer, New York, NY) for 2 days to purge the primary infection of intestinal adult worms. Worm implantation (250 worms/mouse) into small intestine was performed using worms obtained from rat small intestine on day 3 p.i., as described previously.¹³ Faecal egg output was regarded to represent the degree of N. brasiliensis infection.

Eosinophil count in the peripheral blood

Peripheral blood was collected from the tail vein. To count the eosinophils, the blood sample was stained in Hinkelmann's solution (0·5% Eosin Y/0·5% phenol/0·5% formalin in H₂O), or in 5% acetic acid/H₂O to count leucocytes.

Histology

The small intestine was removed and cut 7 cm from the pylorus. Tissues were fixed with 4% paraformaldehyde (Nacarai Tesque, Kyoto, Japan)/in phosphate-buffered saline (PBS) (pH 7·4) and embedded into paraffin wax. For eosinophil staining, sections (4 µm) were stained with Congo-red and counterstained with methyl-green solution.¹⁴ For detection of mucosal mast cells (MMC), sections (4 µm) from Carnoy's fixed tissue were stained with alcian blue (pH 0·3) and safranin O (pH 0·1), as previously described.¹⁵ The number of mast cells was counted in 50 villus crypt units (VCU) and expressed as mast cell numbers per 10 VCU.

Immunohistochemistry for IL-5

For IL-5 immunohistochemistry, sections (4 µm) were blocked with 0·3% H₂O₂ for 30 min at room temperature. Non-specific immunoglobulin interactions were further blocked with 3% rabbit serum for 30 min. Sections were then incubated overnight with anti-IL-5 immunoglobulin (H-85; Santa Cruz Biotechnology, Santa Cruz, CA) at 4°. Biotin-conjugated anti-rabbit immunoglobulin G (IgG) (Dako, Carpintera, CA) was used at a 1 : 250 dilution as a secondary antibody (for 1 hr at room temperature). After washing with PBS, streptavidin-conjugated horseradish peroxidase (Dako) was added, followed by reaction with 3′,3′-diaminobenzidine (Metal-enhanced DAB substrate kit; Pierce Biotechnology, Rockford, IL). The sections were counterstained with haematoxylin solution.

Quantification of cytokines and mouse mast cell protease-1 (mMCP-1) by enzyme-linked immunosorbent assay (ELISA)

Organ extracts for cytokine assays were prepared as described previously.¹⁶ Briefly, tissues were homogenized with a Teflon homogenizer in 1% CHAPS (Calbiochem, San Diego, CA)/RPMI-1640, then centrifuged at 600 g for 15 min at 4°. Aliquots of supernatant were stored at −70° until assayed for cytokines. IL-5 levels in sera and tissue extracts were determined by microplate ELISA, as follows. First, monoclonal anti-mouse IL-5 immunoglobulin (TRFK4; Endogen, Woburn, MA) was coated onto a 96-well plate (Nunc, Rochester, NY). After blocking with PBS containing 1% bovine serum albumin (BSA), samples or recombinant mouse IL-5 (Sigma, St. Louis, MO), as a standard, was applied to each well. Following the addition of biotin-conjugated monoclonal anti-mouse IL-5 immunoglobulin (TRFK5; Endogen), streptavidin-peroxidase was added. Then, 3,3′,5,5′-tetramethylbenzidine (TMB; Dako) was reacted as the substrate. The absorbance was determined at 450 nm by using an EL800 (Bio-Tek, Winoski, VT). IL-4 and eotaxin levels were determined by ELISA (BD Pharmingen, San Diego, CA), according to the manufacturer's protocols. mMCP-1 concentrations in serum were assayed using an mMCP-1 ELISA (Moredun, Midlothian, UK), according to the manufacturer's protocol.

Semiquantitative reverse transcription–polymerase chain reaction (RT–PCR)

The small intestine was removed and a 1-cm length excised at a point 7 cm distal from the pylorus after removal of Peyer's patches. Tissues were homogenized in 1 ml of TRIzol^® (Invitrogen, Carlsbad, CA) on ice. RNA was extracted with chloroform and precipitated using isopropanol. The purity and quantity of RNA were assessed by measuring the UV absorbance at 260 and 280 nm. At each time-point, three mice were killed and tissue samples extracted from the small intestine for RNA purification. Equal amounts of RNA obtained from three individual mice were mixed and used for RT–PCR. RT–PCR reactions were carried out with the iCycler 170–8720JA (Bio-Rad, Tokyo, Japan). RNA (1 µg) was used for cDNA synthesis by using the SuperScriptII^®, RNase H Reverse TranscriptaseII kit (Invitrogen Corp., Carlsbad, CA), according to the manufacturer's protocol. The cDNA was used as a template for amplification in the PCR.

The primers used for PCR analysis are shown in Table 1. The primers of MBP-1 and eosinophil peroxidase (EPO) were as referred to in a previous report.¹⁷ Cycling was performed according to the following profile: 30 seconds at 95°, 60 seconds at the temperatures shown in Table 1, and 60 seconds at 72°. The optimal number of cycles was determined so that the amplified cDNA was below the plateau level. The amplified products were separated on 1·8% agarose gels containing ethidium bromide and analysed by FUJIX DF-20 (FujiFilm, Tokyo, Japan). The amplified products were compared with CD45, a marker of leucocytes.

Table 1.

Primer sequences for semiquantitative reverse transcription–polymerase chain reaction (RT–PCR)

	Sense or antisense	Sequence	Amplified band (bp)	Annealing temp. (°C)	No. of cycles
IL-5	Sense	ACGCAGGAGGATCACATACC	200	64	33
	Antisense	GGCTCTCATTCACACTGCAA
GATA-3	Sense	GGAGGACTTCCCCAAGAGCA	169	67	34
	Antisense	CATGCTGGAAGGGTGGTGAG
EPO	Sense	ACTGTTTCCTGCTAGAGCTTTTGC	295	64.5	36
	Antisense	AGAGTGCTGCTGTTCCTTCAGG
MBP-1	Sense	TCTACTTCTGGCTCTTCTAGTCGGG	629	65	35
	Antisense	GACACAGTGAGATAGACGCCAGTG
CD45	Sense	AAATGGAGATGCAGGGTCCA	452	62	37
	Antisense	CAATGGGACCACTGAAGAAG

Statistical analysis

Statistically significant differences were identified using the Student's t-test.

Results

Unique kinetics of eosinophils in Stat6^–/– mice after primary and secondary infection with N. brasilensis

We first examined the kinetics of faecal egg output and the level of blood eosinophils after subcutaneous infection with N. brasiliensis (500 L3) in wild-type and Stat6^–/– mice. Egg output was observed continuously in Stat6^–/– mice until the end of the experiment (day 21 p.i.), while wild-type mice had ceased the egg output by day 12 p.i. (Fig. 1a). No worms were detected in the small intestine of wild-type mice at the cessation of egg output (data not shown). Thus, Stat6^–/– mice failed to expel N. brasiliensis, as shown previously.³ The percentage of blood eosinophils among leucocytes in wild-type mice was transiently increased on day 10 p.i., and had decreased by day 18 p.i. (Fig. 1b). In Stat6^–/– mice, blood eosinophils showed an increase at day 14 p.i., reached a peak of ≈ 25% on day 18 p.i., and the high eosinophil level persisted thereafter. These results suggest that blood eosinophilia with unique kinetics can be induced in N. brasiliensis-infected Stat6^–/– mice.

Figure 1.

Kinetics of the nematode burden and eosinophilia of mice after primary and secondary infection with Nippostrongylus brasiliensis. The kinetics are shown of egg output in faeces (a), or blood eosinophils (b), of wild-type mice or Stat6-deficient (Stat6^–/–) mice subcutaneously infected with 500 third-stage larvae (L3) of N. brasiliensis. The kinetics are shown of egg output in faeces(c), or eosinophil levels of peripheral blood(d), of wild-type mice or Stat6^–/– mice after a secondary infection with N. brasiliensis, following treatment with 1 mg/kg body weight of pyrantel pamoate for 2 days.(e)Kinetics of blood eosinophils when worm implantation (250 worms/animal) into small intestine was performed using worms obtained from rat small intestine on day 3 p.i. Results are representative of data from three experiments (n = 5).

As previously reported, the induction of Th2 cells which produce IL-4, IL-5 and IL-13 are disturbed in Stat6^–/– mice after infection with N. brasiliensis or Heligomosoides polygylus.^2,3,18 Therefore, we next examined the eosinophil response to secondary infection with N. brasiliensis. When primary infection was artificially terminated by treatment with pyrantel pamoate, blood eosinophil levels were found to return to normal levels 7 days after the treatment. Mice were then reinoculated with the same dose of larvae as given in the primary infection. In wild-type mice, egg production was not detected (Fig. 1c), and the blood eosinophil level rose more quickly than in the primary infection, reaching a level twice as high as that seen in primary infection (Fig. 1d). This augmented eosinophil response, of the timing and magnitude observed in wild-type mice after secondary infection, can be interpreted as being the result of secondary memory responses mediated by presensitized N. brasiliensis-specific Th2 cells that produce IL-4/IL-5 vigorously. In sharp contrast, both the kinetics of egg output and blood eosinophil levels were similar to those of primary infection in Stat6^–/– mice, owing to a lack of memory responses in these mice.

When N. brasiliensis larvae are infected via the subcutaneous route, they migrate to the lungs, up the trachea, down the oesophagus and finally to the small intestine.¹⁹ Thus, infected worms can stimulate host systemic immune responses, in addition to those of the gut mucosal immune system. To examine whether antigen exposure in the small intestine is sufficient to induce blood eosinophilia, we surgically implanted worms into the small intestine of naive mice. Implanted worms are usually expelled from recipient mice at around day 10 after implantation.¹³ The elevation of peripheral eosinophils was low after the implantation of worms in wild-type mice. On the other hand, in Stat6^–/– mice, the eosinophil level was significantly elevated, reaching a plateau on day 14 after worm implantation (Fig. 1e). Stat6^–/– mice were unable to expel the implanted worms until the end of this experiment (data not shown). These results suggest that Stat6-independent eosinophil responses, seen in Stat6^–/– mice, can be induced by exposure to local antigens in the intestine and lack a memory response after secondary infection.

Blood eosinophilia does not always correlate with eosinophil infiltrations in tissues. Therefore, we next counted eosinophils in the small intestine where adult worms reside. Eosinophils were seen in the lamina propria and submucosa in Stat6^–/– mice, as well as in wild-type mice (Fig. 2a). As shown in Fig. 2(b), the number of eosinophils was significantly higher in wild-type mice on day 7 p.i., and had returned to baseline levels by day 14 p.i. In contrast, eosinophils in the intestines of Stat6^–/– mice showed a significant increase by day 14 p.i. and remained at a high level thereafter, confirming again that Stat6-dependent and -independent eosinophil accumulations in intestine have different kinetics, as revealed in blood eosinophil kinetics. Collectively, these results clearly demonstrated that blood and intestinal eosinophilic responses could be induced by chronic intestinal nematode infection in a Stat6-independent manner. The different kinetics of eosinophil responses, and the different responses to secondary infections between wild-type and Stat6^–/– mice, raise the possibility that unique cells, other than Th2 cells, may play a role in eosinophil responses in Stat6^–/– mice.

Figure 2.

Accumulation of eosinophils in sites infected with Nippostrongylus brasiliensis. Small intestine histology showing eosinophil infiltration (a, arrowhead) in the lamina propria in Stat6^–/– mice on day 14 after primary infection with N. brasiliensis (magnification × 400). Sections were stained with Congo red and counterstained with methyl-green solution. (b) Kinetics of eosinophils in the small intestine of wild-type mice and Stat6^–/– mice after infection with N. brasiliensis (n = 5).

Stat6 independent induction of IL-5 in intestine after N. brasiliensis infection

To gain insight into the possible mechanism of Stat6-independent eosinophil responses seen in chronic N. brasiliensis infection in Stat6^–/– mice, we first measured serum IL-5 levels, together with IL-4 and eotaxin, after infection with N. brasiliensis. IL-5 is considered to be critical for eosinophil responses in peripheral blood and tissue, because IL-5 neutralization inhibits N. brasiliensis-induced eosinophilia,²⁰ and IL-5-deficient mice failed to induce eosinophil responses in an airways allergic model and helminth infection in mice.^21,22 As shown in Fig. 3(a), the serum IL-5 level of wild-type mice peaked on day 7 p.i., while that of Stat6^–/– mice had increased significantly by day 14 p.i. As eosinophil levels increase not only in peripheral blood but also at infected sites, we then measured IL-5 levels in the intestine by whole-tissue ELISA. As shown in Fig. 3(d), the protein levels of intestinal IL-5 increased to a maximum by day 14 in Stat6^–/– mice, whereas intestinal IL-5 peaked on day 7 in wild-type mice after N. brasiliensis infection. We also measured IL-4 and eotaxin levels by ELISA. The levels of IL-4, one of the key cytokines for the induction of Th2 cells, were undetectable in the serum of wild-type or Stat6^–/– mice, at least by commercially available ELISA (Fig. 3b). We were unable to detect a significant increase of IL-4 levels in the intestines of either wild-type or Stat6^–/– mice (Fig. 3e). The eotaxin levels in serum and intestine did not change much during the course of infection, either in wild-type or in Stat6^–/– mice (Fig. 3c, 3f). These results suggest that IL-5 produced locally by infected intestine may play a role in inducing eosinophilic accumulation in the intestine of Stat6^–/– mice after infection with N. brasiliensis.

Figure 3.

Kinetics of interleukin (IL)-4/IL-5/eotaxin levels in serum and intestine during the course of primary infection with Nippostrongylus brasiliensis. Kinetics of serum IL-5(a), eotaxin(b)and IL-4(c)of wild-type mice or Stat6^–/– mice during the course of subcutaneous infection with 500 third-stage larvae (L3) of N. brasiliensis. Kinetics of intestinal IL-5(d), eotaxin(e), and IL-4(f)were also shown. Data are representative of two separate experiments and error bars represent the standard error of the mean (SEM) of data from n – 3 – 5 mice.

To confirm that intestine is the site of IL-5 production in Stat6^–/– mice after infection with N. brasiliensis, we next determined IL-5 levels in various tissue extracts together with IL-4 and eotaxin. As shown in Fig. 4, the baseline levels of IL-5 in lungs, small intestine, spleen and mesenteric lymph nodes (mLN) were almost the same in wild-type and Stat6^–/– mice. In wild-type mice, the level of IL-5 in each tissue was increased at day 7 p.i., most significantly in the lungs, intestine and mLN. In contrast, in Stat6^–/– mice, the IL-5 levels showed a marked increase only in small intestine on day 14 p.i., and the IL-5 levels in lungs or mLN changed little, if any, during the course of infection. The kinetics of IL-5 production of wild-type and Stat6^–/– mice may well explain the different eosinophil kinetics: a peak at around day 10 in wild-type mice and a peak after day 14 in Stat6^–/– mice.

Figure 4.

Kinetics of interleukin (IL)-5 (a and b), IL-4 (c and d) and eotaxin (e and f) in various organs/tissues of wild-type and Stat6^–/– mice after infection with Nippostrongylus brasiliensis. Data are representative of three independent experiments and error bars represent the standard error of the mean (SEM) of data from n = 3 – 5 mice. mLN, mesenteric lymph node.

We also measured tissue eotaxin and IL-4 levels by ELISA. As shown in Fig. 4, the level of IL-4 was significantly increased in the mLN and lungs of wild-type mice on day 7 p.i. In Stat6^–/– mice, the levels of IL-4 did not show significant elevation in lungs, spleen or intestine, but did in mLN. The profile of IL-4 production was correlated with IL-5 production in wild-type mice, while in the Stat6^–/– mice there was a clear dissociation between the profile of IL-4 and IL-5 production, suggesting that there may be a unique population which produce IL-5 independently of IL-4 or Stat6 signalling pathways. The kinetics of eotaxin in various organs in wild-type mice also showed a peak on day 7 p.i., and the increase of eotaxin was prominent in the lung but less prominent in mLN. In Stat6^–/– mice, the eotaxin level did not change much during the course of infection in any of the organs examined, or in the mLN. Taken together, the kinetics of eosinophil responses were strongly correlated with intestinal IL-5 levels, but not with eotaxin or IL-4, in Stat6^–/– mice. Stat6-independent IL-5 production in the small intestine may cause the high level of blood and intestinal eosinophilia seen in Stat6^–/– mice after N. brasiliensis infection.

Detection of IL-5, GATA-3 and EPO mRNA expression in the small intestine by semiquantitative RT–PCR

To further confirm that IL-5 is produced in the small intestine where worms reside, we determined IL-5 mRNA expression in the small intestine of mice infected with N. brasiliensis. As shown in Fig. 5, the intestinal IL-5 mRNA level of wild-type mice peaked on day 7 p.i., while that of Stat6-deficient mice peaked on day 14 p.i. This finding confirms that Stat6-independent production of IL-5 can occur during chronic infection with N. brasiliensis in mice and that this IL-5 production from infected intestine may be responsible for the peripheral and intestinal eosinophil responses.

Figure 5.

Semiquantitative reverse transcription–polymerase chain reaction (RT–PCR) of interleukin-5 (IL-5), GATA-3, eosinophil peroxidase (EPO) and major basic protein-1 (MBP-1) of intestine from wild-type and Stat6^–/– mice on days 0, 7 and 14 after infection with Nippostrongylus brasiliensis. At each time-point, three mice were killed and RNA was extracted from the small intestine. Equal amounts of RNA from three individual mice were mixed and analysed by RT–PCR for the expression of IL-5, GATA-3, EPO, MBP-1 and CD45 mRNA. RT–PCR for each sample was conducted using specific primers, as shown in Table 1. Data are representative of two independent experiments.

GATA-3 is a transcription factor essential for Th2 polarization and known to activate the IL-5 promoter.²³ Importantly, GATA-3 is not only expressed in Th2 cells, but also in lung eosinophils.^24,25 We found that the expression of GATA-3 mRNA was greatly increased in the intestine of Stat6^–/– mice on day 14 after infection with N. brasiliensis.

EPO has been used successfully as a marker for the degree of activation of eosinophils. In order to determine whether intestinal eosinophils, seen in Stat6^–/– mice on day 14 after infection with N. brasiliensis, are actually activated, we analysed the mRNA level of EPO and major basic protein-1 (MBP-1) in the small intestine after infection. We found that the EPO mRNA level showed a dramatic increase on day 14 of infection in Stat6^–/– mice after infection with N. brasiliensis, whereas in wild-type mice the mRNA level of MBP-1 peaked on day 7. Increased EPO mRNA in the intestine of Stat6^–/– mice after N. brasiliensis infection indicated that eosinophils induced in intestine by Stat6-independent mechanisms are activated. Dissociation of EPO and MBP-1 mRNA increases in Stat6^–/– may reflect the different cytokine requirements for the induction of EPO and MBP-1. It has been suggested that IL-5 is essential for the induction of EPO, and IL-4 for MBP-1.²⁶ In Stat6^–/– mice, only IL-5 (and not IL-4) is produced, and these unique cytokine environments in the intestine may cause the dissociation of EPO and MBP-1 in Stat6^–/– mice. Taking into account these findings of IL-5, GATA-3 and EPO mRNA expression in the intestine of Stat6^–/– mice, it can be speculated that intestinal eosinophils observed in Stat6^–/– mice on day 14 after infection with N. brasiliensis may be activated by parasitic stimulation and are producing IL-5 via the autocrine pathway, which has been reported previously by several investigators.^27–29

Expression of IL-5 in small intestine

To directly identify the IL-5-producing cells in the intestine, seen in Stat6^–/– mice, we performed immunohistochemistry for IL-5 using serial sections of intestines from Stat6^–/– mice on day 14 after N. brasiliensis infection. As shown in Fig. 6, we detected IL-5-positive cells in the lamina propria of Stat6^–/– mice, and some of the IL-5-positive cells were considered to be eosinophils by Congo-red/methyl-green staining of serial sections.

Figure 6.

Detection, by immunohistochemistry, of interleukin-5 (IL-5)-positive cells and eosinophils, in the small intestine of Stat6^–/– mice, on day 14 after infection with Nippostrongylus brasiliensis. Serial sections were stained with Congo red/methyl green to detect eosinophils. Arrows indicate the double-positive cells.

Number of intestinal MMC and serum mMCP-1 concentration

To exclude the possibility of IL-5 production by MMC, we analysed the numbers and activation status of MMC by measuring serum mMCP-1. As shown in Table 2, the increase in MMC and mMCP-1 levels of Stat6^–/– mice were less than those of wild-type mice. Mast cells of Stat6^–/– mice on day 14 after N. brasiliensis infection were mainly seen in the intraepitherial area (data not shown). On the other hand, IL-5-positive cells were found mainly in the lamina propria (Fig. 6). Taken together, these results suggest that mucosal mast cells were unlikely to be a cellular source for Stat6-independent IL-5.

Table 2.

Intestinal mast cell numbers and mouse mast cell protease-1 (mMCP-1) levels in the serum of wild-type and Stat6^–/– mice infected with Nippostrongylus brasiliensis

	No./10VCU ± SEM		mMCP-1 ± SEM (ng/ml)
Day p.i.	Wild type	Stat6–/–	Wild type	Stat6–/–
0	1 ± 0.5	2 ± 0.8	17 ± 1.4	53.6 ± 18.7
14	103 ± 32.4	22 ± 8.1	702.8 ± 84.0	183.1 ± 50.0

p.i., postinfection; SEM, standard error of the mean; VCU, villus crypt units.

Discussion

One of the notable findings in this study was the different kinetics of eosinophil responses and IL-5 production between Stat6-dependent and -independent pathways. After infection with N. brasiliensis, wild-type mice showed eosinophilic accumulation that peaked at around day 10 and decreased thereafter, whereas in Stat6^–/– mice the magnitude of the eosinophil response kept increasing and reached a plateau after day 14. We found that serum IL-5 levels correlated well with the kinetics of eosinophils, both in wild-type and in Stat6^–/– mice. Detailed analysis of tissue cytokines showed that both IL-5 and IL-4 production are greatly increased in mLNs on day 7 in wild-type mice (Fig. 4a,4c), indicating Th2 cells that have developed via Stat6-dependent pathway after N. brasiliensis infection. In sharp contrast to this, after N. brasiliensis infection, the production of IL-5 in Stat6^–/– mice did not correlate with the production of IL-4 (Fig. 4b, 4d) and only IL-5 production was evident in the intestine (Fig. 4d), indicating that in Stat6^–/– mice, IL-5 seemed to be produced by some cells localized in the intestine, other than typical Th2 cells seen in wild-type mice. The findings that secondary infection with N. brasiliensis in these Stat6^–/– mice did not induce any augmenting effects on eosinophil responses also strongly suggest that Stat6-independent IL-5 production is not from typical Th2 cells.

Th2 development has been considered to be essentially an IL-4Rα/Stat6-dependent process, but recent analysis of Bcl-6/Stat6 double-knockout mice suggests that Th2 development could be partly Stat6 independent.³⁰ Ouyang and colleagues established that GATA-3 is part of a master switch for Th2 commitment in T cells.³¹ This response to GATA-3 is independent of Stat6. Therefore, interpretation of IL-5 and GATA3 mRNA expression in the intestine of Stat6^–/– mice should be undertaken with caution and it remains possible that the IL-5 seen in Stat6^–/– mice is derived from Th2 cells that developed in a Stat6-independent manner. However, as previously reported by Takeda et al., we could not detect any significant level of IL-5 or IL-4 production from in vitro-stimulated splenic T cells obtained from Stat6^–/– mice on days 14 or 21 after infection with N. brasiliensis (data not shown). Combined with the lack of secondary memory responses in Stat6^–/– mice, these results suggest that IL-5 is unlikely to be produced by Stat6 independently developed Th2 cells in this study. In addition to Th2 cells, mast cells are well known as a cellular source of IL-5. Intestinal mast cells have been reported to require IgE receptor cross-linking to produce IL-5.^32,33 However there were no IgE responses after infection with N. brasiliensis in these Stat6^–/– mice.³ Thus, it is unlikely that mast cells are producing IL-5 after N. brasiliensis infection in Stat6^–/– mice. Taken together, our results strongly suggest that unique cells in intestine, other than mast cells or Th2 cells, may play a role in Stat6-independent IL-5 production induced after chronic intestinal nematode infections.

Recently, evidence has accumulated that eosinophils can release a significant amount of IL-5 which may contribute to local eosinophil recruitment and activation.^27–29 Furthermore, it has been demonstrated that not only Th2 cells, but also eosinophils, express GATA-3 upon activation.^24,25 Therefore, we believe that the increase of GATA-3 mRNA expression in intestine of Stat6^–/– mice after chronic infection with N. brasiliensis may be derived from eosinophils, not from Th2 cells. It may be speculated that intestinal eosinophils are activated by local stimulation of the parasitic burden and start to express GATA-3 mRNA. As supporting evidence for this, we found that EPO mRNA increased in intestine obtained from Stat6^–/– mice on day 14 after N. brasiliensis infection (Fig. 5), indicating that these eosinophils in intestine on day 14 are activated and can potentially express GATA-3. Furthermore, we demonstrated, by immunohistochemistry, that some intestinal eosinophils of Stat6^–/– mice on day 14 p.i. were actually positive for IL-5. Taken together, our data highlight the possibility of the autocrine pathway of eosinophil differentiation and activation in Stat6^–/– mice after chronic infection with N. brasiliensis.

As previously reported, the IL-4Rα/Stat6 pathway is required for the expulsion of the nematode, N. brasiliensis.³ Infected Stat6^–/– mice revealed a sustained level of intestinal worms and suffered from chronic nematode infections, and we found that chronic nematode infections can induce IL-5 production in a Stat6-independent manner and cause eosinophil inflammation. This occurred when N. brasiliensis was implanted directly into the intestine of Stat6^–/– mice (Fig. 1e), indicating that sustained parasitic exposure in intestine is the cause of Stat6-independent IL-5 production. This Stat6-independent pathway for IL-5 production and subsequent eosinophil responses has important clinical implications because similar mechanisms may play a role in human chronic allergic diseases. Blease et al. has recently reported that Stat6-independent eosinophil responses can occur in a chronic airways hypersensitivity model using chronic fungal exposure.⁹ Another study also supports this Stat6-independent airway hyper-reactivity in an inhalant challenge model of chronic asthma.³⁴ Taken together, it can be stated with confidence that Stat6-independent IL-5 production exists, which is caused by chronic antigenic exposure to the mucosal immune system, and that it may play a role in the pathogenesis of chronic allergic/inflammatory responses in the lungs and intestine. Future elucidation of cell types responsible for this Stat6-independent production of IL-5, and its induction mechanism, may provide a new direction for the improved understanding and treatment of chronic allergic/inflammatory diseases.

Abbreviations

ELISA

enzyme-linked immunosorbent assay

EPO

eosinophil peroxidase

IgE

immunoglobulin E

IL

interleukin

IL-4Rα

IL-4 receptor-α

L3

third-stage larvae

MBP-1

major basic protein-1

mLN

mesenteric lymph nodes

MMC

mucosal mast cell

mMCP-1

mouse mast cell protease-1

N. brasiliensis

Nippostrongylus brasiliensis

p.i.

postinfection

PBS

phosphate-buffered saline

RT–PCR

reverse transcription–polymerase chain reaction

Stat6

signal transducers and activators of transcription family

Th2

T helper 2

VCU

villus crypt units