IL-5 drives eosinophils from bone marrow to blood and tissues in a guinea-pig model of visceral larva migrans syndrome

This study was undertaken to evaluate the role of IL-5 in eosinophil migration and in the maintenance of eosinophilia in a guinea-pig model of visceral larva migrans syndrome. The results show that the infection of animals with Toxocara canis induced an early increase in serum IL-5 levels that might be essential for eosinophil differentiation and proliferation and for the development of eosinophilia. When infected guinea-pigs were treated with mAb anti-IL-5 (TRFK-5) given at the same time or 1 or 3 days after infection, there was a high percentage of reduction of eosinophil counts 18 days after infection. However, when the mAb was administered during the peak of eosinophilia, there was high inhibition in blood, no inhibition in bronchoalveolar lavage fluid (BALF) or peritoneum and an increase in eosinophil numbers in bone marrow. Thus, a basic level of IL-5 may be essential to drive eosinophils from bone marrow to blood and tissues, and for the maintenance of eosinophilia in infected animals. We may also conclude that when eosinophils have already migrated to the lungs, TRFK-5 has no power to inhibit eosinophilia, which is also under control of local lung cells producing IL-5. In this way, only one later TRFK-5 treatment may not be sufficient to modify the lung parenchyma microenvironment, since T. canis antigens had already stimulated some cell populations to produce IL-5.


Introduction
Eosinophilia has been associated with parasitic diseases, particularly when the parasites invade the tissues or injure the mucosal surfaces. Toxocara canis is an intestinal parasite of dogs, and is the most common aetiologic agent of visceral larva migrans syndrome (VLMS). In humans, VLMS results from the ingestion of embryonated eggs of T. canis, that eclode in the small intesfine. The infective larvae invade the mucosa, move into the liver via the portal circulation, and from there to the lungs. 2 Beaver et al., 3 who were the first to describe this syndrome, noted the intense eosinophilia which reaches more than 90% of total leucocyte counts. However, there are few studies regarding the mechanisms involved in the blood and tissue eosinophilia obseeeed in VLMS.
Several investigators have suggested a direct correlation between eosinophilia and interleukin-5 (IL-5) in human helminth infections 4'5 and in 67 experimental animal models. Inhibition of eosinophilia has been demonstrated by anti-IL-5 24 Mediators of Inflammation Vol 5 1996 treatment in mice infected with Nipostrongylus brasiliensis, 8 $chistosoma mansoni, Toxocara canis 7 and Heligmosomoidespolygyrus. 1   also been shown to support the terminal differentiation, proliferation of eosinophil precursors 11'12 and eosinophil activation. 13 Although IL-5 does not demonstrate eosinophil chemotactic activity in vivo 14 there is some evidence suggesting that this cytokine may modulate a selective eosinophil accumulation at the site of inflammation. Moreover, Sehmi et a/.15 reported that IL-5 has a selective priming effect on eosinophil migratory response to nonselective chemoattractant mediators in vitro. Also, Moser et a/. 16 have demonstrated that in order to acquire the ability to transmigrate, eosinophils must be primed with IL-5, IL-3 and GM-CSF. Thus, the involvement of IL-5 in eosinophilia is not fully understood.
In the present study we have used a guinea-pig model of VLMS to investigate the involvement of IL-5 in eosinophil migration and in the maintenance of eosinophilia in blood, bone marrow, lung and peritoneal cavity. carried out using diluting fluid in a Neubauer of peroxidase-labelled streptavidin (1/1000, Kirchamber. Differential countings were obtained kegaard & Perry Laboratories Inc., Maryland, using Rosenfeld-stained cytocentrifuge prepara-USA) were added to each well. Following incubations, tion for l h at 37C and further washing, the enzyme was developed using the TMB substrate Bronchoalveolar lavage fluid. The guinea-pigs peroxidase for 5 min. The reaction was stopped were killed by an overdose of sodium pento-by adding 501.tl of 2.0 N HCl, and the optical barbitone and 5ml of phosphate-buffered saline densities were read at 490nm using an auto-(PBS) containing 0.5% sodium citrate (PBS/SC), mated plate reader. The sensitivity of the assay at room temperature, were instilled through a was 0.15 ng/ml and the upper limit 100ng/ml. polyethylene cannula introduced into the trachea. The cells present in the bronchoalveolar lavage Monoclonal antibodies: The rat monoclonal antifluid (BALF) were recovered immediately. The body TRFK-5 was a generous gift from Dr P. procedure was repeated once. The leucocyte Minoprio, Institut Pasteur, Paris. The neutralizing counts in the BAI_ were determined as descriantibody was purified by precipitation with bed above, ammonium sulfate (45%) from ascites prepared in CD1 nude mice (Charles River, St Aubin les Peritoneal cells: The cells from the peritoneal Elbeuf, France) inoculated 1 week before the cavities were harvested by injection of 10ml of injection of hybridoma cells, with I ml of pris-PBS/SC into the peritoneum. Only 5-8 ml of the tane (Sigma). After precipitation and dialysis of fluid was withdrawn for cell counts, as described the ascite fluid overnight against PBS, the dialyabove, sate was further purified on a Protein G1 column (HiTrapTM, Pharmacia Upsala, Sweden).
Bone marrow cells: Bone marrow cells were collected by flushing the contents of the guinea-pig Eosinophil and cytokine depletion: Guinea-pigs femur with 10 ml of PBS/SC. Total cell numbers were injected i.p. with TRFK-5 or with the irrelewere determined as above. In the differential cell vant antibody (rat IgG against total anti-human counts the cell populations were divided into IgG) once, 2mg/animal, at the time of infection mature neutrophils, mature eosinophils and or at different intervals (1, 3, 12 or 17 days) others (mainly precursors and mononuclear thereafter. The animals in this group were sacricells), riced 18 days after infection.
Histopathological studies: Tissues were removed Recovery of larvae from liver: One lobule of from guinea-pigs at various times post-infection each liver was used to determine the larval and immediately fixed in 10% formalin. Tissues counts from infected guinea-pigs. Larval recovery was evaluated as described by Kayes and Oaks, 18 with minor modifications. Briefly, the tissue was chopped and digested with pepsin-HC1 (pH 1.  . Asterisks indicate a significant difference between infected and noninfected animals (n 5 6). *p < 0.05 and **p < 0.01. IL-5 drives eosinophils in guinea-pig sively until day 24, 12.44 4-2.72 x 105) (Fig. 1D).
The percentage of eosinophils in some animals reached 55% at the peak of infection. No increase in the number of mononuclear cells was seen in any compartment analysed.
Larval counts: The percentage of inoculated T. canis larvae recovered by peptic digestion of the liver of experimental animals 4 h and 1, 2, 3, 4, 9, 12 and 18 days after inoculation of 500 eggs per animal is shown in Fig. 2. Most of the larvae were recovered 2 to 4 days after infection and 10% recovery was also observed on day 18  drastically reduced, even when determined 18 days after infection (Table 1). No inhibition of eosinophil counts was observed when the animals were inoculated with the irrelevant antibody at the time of infection (Table 1). Fig. 4 shows the comparative results of eosinophilia obtained when the antibody was given 3 days or 17 days after egg inoculation. The antibody given at 3 days after infection induced a high percentage of inhibition in eosinophil counts in all the compartments analysed 18 days after infection (Fig. 4A). However, when TRFK-5 was administered to the infected animals on day 17 post-infection (thus 1 day before sacrifice), a significant inhibition in number and percentage of eosinophils was observed only in the blood (p=0.030) (Fig. 4B). A small non-significant decrease was seen in BALF (p 0.790) and peritoneum (p= 0.222). Moreover, the number of mature eosinophils in bone marrow increased by 140% (p 0.038).
As demonstrated in Fig. 4B, the behaviour of eosinotShilia in BALF was completely different from that observed in blood. Thus, to better understand the eosinophilia in the lungs of infected animals, we monitored eosinophil numbers in BALF after administration of TRFK-5 at the same time, or 1, 3, 12 or 17 days after infection. The animals were sacrificed 18 days after infection. In another group, TRFK-5 was administered 18 days post-infection and the animals were sacrificed 6 days later. When the mAb was administered at the same time or 1 or 3 days postinfection there was a significant inhibition in the number of eosinophils (Fig. 5). These data show L. H. Faccioli et al. that the inhibition of the first peak of IL-5 which appeared at 1 to 3 days after infection as shown in Fig. 3, is also very important for the establishment of eosinophilia in the lungs. However, when the mAb was administered 12, 17 or 18 days after infection there was no significant inhibition in the numbers of eosinophils in BALF heart; data not shown) and muscle, as reported by other investigators, 8 were infiltrated. The factors responsible for in vivo eosinophil accumulation at inflammatory sites have been poorly defined, although T lymphocytes and mast cells appear to be involved in eosinophilia. 9'2 IL-5, a T cell-derived factor that regulates B cell functions, is an eosinophil differentiation factor 11 as well as a stimulating and survival-prolonging factor specific for eosinophils in vitro. 2 Also, several investigators have demonstrated that sTstemic eosinophilia in mice infected with parasites is mediated by IL-5 produced in response to the infection. 2'22 In the present study, the i.p. administration of the TRFK-5 antibody markedly inhibited the widespread eosinophilia observed the intestinal wall had migrated into the liver within 72h after inoculation as demonstrated nophilia persists in lungs, probably by the secrehere and elsewhere. 2 It is apparently during this tion of IL-5 from cells localized in the lung interval that the worm provides the signals to microenvironment, cytokine-producing cells, which in turn trigger increased serum levels of specific cytokine as Histopathological analysis: The treatment of T. demonstrated here for IL-5, 24 to 72 h after infeccanis-infected animals with irrelevant antibody tion. The signals may be provided directly by the showed a widespread eosinophilic infiltration as invading parasite or by cells in response to the in untreated animals (Fig. 6A,B). However, the parasite. The cytokine pattern that develops at treatment of animals with TRFK-5 at the same this early stage, probably induced by a T-cell time of infection, or 1 day or 3 days later ted to independent pathway, may also influence the a complete inhibition of eosinophil infiltration in pattern of T cell differentiation into a Th2 type, the lung parenchyma (Fig. 6C). By contrast, the which may be responsible for the second peak mononuclear cell infiltration in the lungs was of IL-5 observed in our experimental model not modified. When the infected guinea-pigs ( Fig. 3), although a second cycle of larval invareceived TRFK-5 1 day before sacrifice (or 17 sion (Fig. 2) with a rapid peak of IL-5 liberation days post-infection), eosinophil infiltration in the cannot be ruled out. lung parenchyma was also inhibited (Fig. 6D) Thus, our results suggest that the eosinophilia but not to the same extent as observed in the against helminth larvae may be initiated by the group receiving TRFK-5 given at the time of release of IL-5 when the parasites migrate from infection or 3 days later. Thus, the histological the intestine to the liver by stimulation of specific determination of eosinophil infiltration in these cell populations. Then, an early release of IL-5 lungs corroborates a reduction but not a sizequickly induces eosinophil recruitment, probably able inhibition of eosinophil numbers as first from the stored mature eosinophil pool observed in the BALF of the same infected from vascular endothelium or by the mobilizaanimals, tion of eosinophils from extravascular sites to the blood. This fact could explain why we found Discussion increased eosinophils first in blood and later in other compartments. The early IL-5 release may The results of the present study show that in also serve as a signal for eosinophil differentiaour experimental model widespread eosinophilia tion and maturation in bone marrow. The time follows the infection of guinea-pigs with second inteeeal observed between the first peak of IL-5 stage eggs from T. canis, as also noted in release and the increase of eosinophils in blood humans and in other experimental animals. 7'7 T. coincides with that reported to be necessary for canis is a potent stimulus for systemic eosino-eosinophil differentiation and maturation in 12 philia, since blood, BALF, peritoneum and all vitro. Increased eosinophil production and libtissues examined (kidney, eyes, spleen, thymus, eration into blood and other tissues occurs thereafter. Thus, early and later IL-5 release provides a necessary level of this cytokine, which is involved in the maintenance of eosinophilia. We may assume that the inhibition of the first peak of IL-5 release by TRFK-5 does not permit the subsequent T cell stimulation and differentiation. This may explain the long-lasting effect of TRFK-5 treatment observed here and also reported by others. 8 In agreement with our results, there is an important observation of Svetic et al. 24 showing that a specific and highly reproducible IL-5 gene expression pattern is detectable in Peyer's patches by 6 to 12h after Heligmosomoides polygyrus infection. The early increase in IL-5 gene expression after infection was probably T cell-independent, inasmuch as it was obseeeed in Peyer's patches of congenitally athymic mice and of conventional mice treated with anti-CD4 30 Mediators of Inflammation Vol 5 1996 and anti-CD8 mAb. Moreover, Kusama eta/. 25 have observed two peaks of eosinophilia in normal and athymic mice, and suggested that IL-5 observed in the first peak was produced by cells other than CD4 T cells, since anti-CD4 and anti-CD3 mAb reduced only the second peak of eosinophilia in normal mice and slightly reduced the first peak of eosinophilia in both normal and nu/nu mice. The local lung cells producing IL-5 may also help us to explain the reason why 12, 17 or 18 days post-infection TRFK-5 treatment only partially inhibits, or does not inhibit eosinophil infiltration into the lungs, as demonstrated in Figs 5 and 6. We may suggest that when eosinophils have already migrated to the lungs, TRFK-5 has no power to inhibit eosinophilia, which is also under control of local lung cells producing IL-5. In this way, only one later TRFK-5 treatment may not be sufficient to modify the lung parenchyma microenvironment, since T. canis antigens have already stimulated some cell populations to produce IL-5, as demonstrated by Kusama et aL25These results suggest that eosinophilia in lungs is under the control of different factors when compared to that observed in blood and the peritoneal cavity. One of the most important results obtained here was the inhibition of circulating eosinophil numbers by the different mAb treatments, even when the antibody was given at the peak of blood eosinophilia, which was accompanied by an increase of mature eosinophils in bone marrow. This suggests that IL-5, apart from being required for the terminal differentiation of eosinophils in bone marrow, 26 is also likely to drive eosinophils from the bone marrow to the blood and then to the tissues, probably by upregulating VLA-4 expression in eosinophils.

Moser et aL have demonstrated that in order to
acquire the ability to transmigrate, eosinophils must be primed with cytokines such as IL-5, IL-3 or GM-CSF for expression of adhesion molecules such as VI-4. Recently, Pretolani et al. 27 have indeed shown that an anti-VLA-4 antibody suppresses eosinophil recruitment to lung in the guinea-pig and, as a consequence, inhibits the accompanying bronchopulmonary hyperresponsiveness.