Amelioration of Influenza-Induced Pathology in Mice by Coinfection with Trichinella spiralis

Illness due to respiratory virus infection is often inducedby excessive infiltration of cells into pulmonary tissues, leadingto airway occlusion. We show here that infection with Trichinellaspiralis results in lower levels of tumor necrosis factor inbronchoalveolar lavage fluid and inhibits cellular recruitmentinto the airways of mice coinfected with influenza A virus.Infiltration of neutrophils and CD4⁺ and CD8⁺ lymphocytes wasreduced, resulting in animals gaining weight more rapidly followingthe initial phase of infection. Influenza resulted in a generalizedincrease in vascular permeability in pulmonary tissues, andthis was suppressed by parasite infection, although the effectswere restricted to the early phase of trichinosis. Moreover,the number of cells producing interleukin-10 (IL-10), and thelocal levels of this cytokine, were reduced, suggesting thatamelioration of pulmonary pathology by parasite infection occursindependently of IL-10 production.

INTRODUCTION

Top

Introduction

Coinfection with different pathogens is a common occurrencewhich can alter the progression of disease (33, 36, 38). Theserelationships are complex, as the immune response can regulatesurvival and proliferation of the pathogen and be responsiblefor tissue damage. The effect of helminths on concurrent infectionsis of particular interest, given their prevalence in human populations,their ability to drive strong type 2 responses, and a wide rangeof immunomodulatory effects (27). As type 1 and type 2 responsesare generally antagonistic, it has been suggested that helminthparasites may result in increased progression of disease whenassociated with bacteria and viruses, for which type 1 responsesare generally regarded as beneficial in the control of infection(7, 22). There has been considerable debate on the impact ofhelminth infection on management of other diseases and whetherthis would be improved by anthelminthic treatment of the targetpopulations (10, 11, 29).

Studies in animal models have in some cases correlated helminthcoinfection with a shift in lymphocyte polarization, resulting,for example, in defective induction of type 1 cytokines andcytotoxic T-cell activity and a delay in viral clearance (1).However, the impact of one infectious agent upon another interms of resultant immune responses, clearance of pathogens,and development of pathology varies greatly dependent upon theorganisms involved, the timing of infection, the anatomicaldistribution of the pathogens, and their mode of replicationwithin the host (13, 16, 26, 32). Indeed, recent studies haveshown that treatment of Ugandan individuals with an antischistosomaldrug resulted in transient increases in human immunodeficiencyvirus type 1 loads and sustained decreases in CD4⁺ T-cell counts(8).

We investigated the effects of Trichinella spiralis on influenzaA infection. The rationale is, first, that these organisms colonizedistinct anatomical sites, allowing an analysis of systemicrather than local bystander effects. T. spiralis organisms initiallyinvade intestinal epithelia, later disseminating to skeletalmuscle, whereas influenza virus is localized to the respiratorytract. Approximately 2 million children are estimated to dieof acute respiratory infection per annum, 70% of these in Africaand Southeast Asia, where helminth infections are also highlyprevalent (40). Influenza occurs in seasonal epidemics whichaffect 5 to 15% of the world's population, and there is considerableconcern that the recent explosion of the H5N1 avian influenzastrain in Asia may result in a global pandemic in the humanpopulation (39).

The enteric phase of T. spiralis infection generally inducesa type 2-biased response, particularly in fast-responder (resistant)strains of mice, which becomes mixed during the systemic phase(18, 23). In contrast, replication of influenza virus in epithelialcells lining the airways is initially controlled by innate immunemechanisms, prior to specific adaptive responses which ultimatelyresult in subsequent protection against reinfection. A priorset of experiments indicated that coinfection of mice with Nematospiroidesdubius (Heligomosomoides polygyrus) and influenza A resultedin lower antihemagglutinin titers, reduced lung consolidation,and an indication of lower viral titers than in mice infectedwith influenza alone, although the mechanisms underlying theseeffects were not investigated (9). We show here that infectionwith T. spiralis inhibits cellular recruitment into the airwaysof mice infected with influenza, resulting in a more rapid gainof weight during the adaptive phase of the antiviral immuneresponse. Interestingly, this effect was observed only duringthe early phase of parasite infection, not during long-termresidence in skeletal muscle, and was not accompanied by localexpansion of interleukin-10 (IL-10)-producing cells.

MATERIALS AND METHODS

Top

Materials and Methods

Infection procedures. Male NIH mice (6 to 8 weeks old) were infected by oral gavagewith 500 infective larvae. At different times after parasiteinfection, or independently, mice were inoculated intranasallywith 50 hemagglutinin (HA) units of influenza A virus X31 (HAtiter, 4,096) diluted in phosphate-buffered saline. Animalswere monitored and their weights determined on a daily basis.

Influenza-specific plaque assay. The plaque assay was performed as previously described (19).Cells were removed from the upper left lobe of the mouse lung,and viral plaques were determined on Madin-Darby canine kidney(MDCK) cells using anti-influenza nucleocapsid A antibody (cloneAA5H; Serotec).

Recovery of cells, flow cytometry, and histochemical staining. Single-cell suspensions from mediastinal lymph nodes (MdLN),lungs, and bronchoalveolar lavage (BAL) fluid were preparedas previously described (21) and then stained with allophycocyanin-conjugatedrat anti-mouse CD8a, 1 µg ml^–1 (clone 53-6.7; BDBiosciences); phycoerythrin (PE)-Cy5-conjugated rat anti-mouseCD4, 1 µg ml^–1 (clone RM4-5; Caltag Laboratories);rat anti-mouse CD49b/pan NK cell marker, 2.5 µg ml^–1(clone DX5; BD Biosciences); PE-conjugated rat anti-mouse IL-10,2 µg ml^–1 (clone JES5-2A5; Caltag); or PE-conjugatedrat anti-mouse gamma interferon (IFN- ${gamma}$ ), 2 µg ml^–1(clone XMG1.1; Caltag). Suspensions were then analyzed on aFACSCalibur flow cytometer using CELL-Quest software.

Cell suspensions from BAL fluid were enumerated, centrifugedonto glass slides, dried, fixed, and stained in hematoxylinand eosin (VWR). At least 400 cells per slide were counted todetermine the relative percentages of lymphocytes, macrophages,eosinophils, and neutrophils. For histochemistry, lungs wereremoved from mice at different time points postinfection andfixed in 10% normal buffered formalin; 4-µm paraffin-embeddedsections were cut and stained with hematoxylin and eosin.

Measurement of albumin in bronchoalveolar lavage fluid. Samples and murine-albumin standards were mixed with an equalvolume of the fluorescent probe 8-anilino-1-napthalenesulphonicacid (Sigma) at 0.2 mM, 0.25 M sucrose, and 0.05 M Tris-HCl(pH 7.4) at room temperature. Fluorescence was measured at 380nm excitation and 460 nm emission wavelengths, and albumin concentrationswere determined from a standard curve (37).

Cytokine assay. Enzyme-linked immunosorbent assay plates were coated overnightat 4°C with antimouse IL-10 (BD Biosciences) at 4 µgml^–1 or antimouse IFN- ${gamma}$ (R&D Systems) at 6 µgml^–1. Plates were washed and blocked, 30 µl of BALfluid was added, and plates were incubated at room temperaturefor 2 h. Samples were washed prior to addition of either biotinylatedantimouse IL-10 at 1 µg ml^–1 (BD Biosciences) orbiotinylated antimouse IFN- ${gamma}$ at 400 ng ml^–1 and developedwith avidin-horseradish peroxidase conjugate (BD Biosciences)-3,3',5,5'-tetramethylbenzidine(TMB) substrate and then read at 450 nm.

Apoptosis and in vivo cellular proliferation. Cells from MdLN were stained with PE-conjugated anti-annexin-V(BD Biosciences), used neat, and either PE-Cy5-conjugated ratanti-mouse CD4 (clone RM4-5; Caltag) or peridinin-chlorophyl-proteincomplex-conjugated rat anti-mouse CD8a (clone 53-6.7; BD Biosciences),both used at 1:200. Cells were analyzed by flow cytometry afteraddition of 1 µM To-pro-3 (Molecular Probes) to the finalsuspension. Cellular proliferation in vivo was determined aspreviously described (19). Mice were given 0.8 mg ml^–1bromodeoxyuridine (BrdU) in their drinking water from the initialday of influenza A infection until sacrifice; single-cell suspensionswere prepared from MdLN, stained with fluorescein isothiocyanate(FITC)-conjugated anti-BrdU (BD Biosciences), and analyzed byflow cytometry.

Statistical analysis. Groups of five mice were used, and experiments were repeatedat least twice. Data are shown as means ± 1 standarddeviation (SD). Data were analyzed using two-way analysis ofvariance and Bonferroni's posttest. P values of $≤$ 0.05 were consideredto be statistically significant (*, P $≤$ 0.05; **, P $≤$ 0.01; ***,P $≤$ 0.001).

RESULTS

Top

Results

Coinfection of mice during the enteric phase of trichinosis results in reduced lung pathology and accelerated recovery of weight. In order to assess the effects of T. spiralis at different stagesof development on the course of influenza infection, male NIHmice were inoculated with virus during the peak of the entericphase coincident with release of newborn larvae, i.e., day 7,and also during the systemic phase when parasites were encystedin skeletal muscle, i.e., day 60 postinfection (p.i.). A doseof 50 HA units administered intranasally to control mice resultedin animals dropping to approximately 80% of their initial weightat day 4, followed by progressive recovery (Fig. 1A). Infectiousviral particles were recovered from the lungs at days 2 and4 p.i., but viral clearance was complete by day 7 (Fig. 1B).In mice coinfected with T. spiralis, there was no alterationin weight loss in the first 4 days, and these animals also lost20% of their body mass over this time period (Fig. 1A). Afterday 4, however, the coinfected mice gained weight more rapidlyand were significantly heavier than controls from day 7 onward.Nevertheless, the number of infectious viral particles in thelungs was not significantly different on days 2 and 4, and therate of viral clearance was unaffected. In contrast, coinfectionwith influenza virus during the later phase of trichinosis (day60) did not affect the dynamics of weight loss and recovery,and viral replication and clearance were similarly unaffected(Fig. 1C and D). These observations proved to be highly reproducible.

View larger version (13K):
[in this window]
[in a new window]
FIG. 1. Coinfection with enteric-stage T. spiralis infection enhances recovery from influenza infection but has no effect on viral clearance. Mice were infected orally with T. spiralis and at two later distinct time points inoculated intranasally with 50 HA units of influenza A virus (X31). Weight loss was monitored daily and expressed as a percentage of the weight prior to influenza infection (influenza administered at day 7 [A] and at day 60 [C] of parasite infection). Viral load per lung was assessed by plaque assay (influenza administered at day 7 [B] and at day 60 [D] of parasite infection). Points in panels B and D represent viral titers from individual mice. In all panels, data for mice infected with influenza virus alone are shown as filled circles and data for coinfected animals shown as open circles; data represent mean values ± 1 SD. Significance values are shown relative to mice infected solely with influenza. **, P $≤$ 0.01; ***, P $≤$ 0.001.

The concentration of serum albumin in BAL fluid reflects vascularintegrity in pulmonary tissues and permeability of the epitheliallining of the airways and was thus measured as an indicatorof lung pathology. In control mice singly infected with influenzavirus, the albumin concentration in BAL fluid remained relativelylow in the first 4 days but was then sharply elevated to over2 mg ml^–1 at days 7 and 10 p.i. By contrast, albumin inBAL fluid of mice coinfected during early trichinosis remainedunder 1 mg ml^–1, which was significantly lower than thecontrol group at both of these time points (Fig. 2A). In micecoinfected with influenza virus during the later phase of trichinosis,the concentration of albumin in BAL fluid was slightly reducedat days 7 and 10 p.i. but was not significantly different fromthat of controls (Fig. 2B).

View larger version (13K):
[in this window]
[in a new window]
FIG. 2. Coinfection results in lower albumin concentrations in BAL fluid. Mice were infected orally with T. spiralis and 7 days (A) or 60 days (B) later inoculated intranasally with influenza A virus. The albumin concentration in lavage fluid was determined at different time points after viral infection. Filled bars indicate data for mice infected with influenza alone and open bars represent data for coinfected animals. Mean values ± 1 SD are shown, and significant differences relative to mice infected solely with influenza are highlighted. ***, P $≤$ 0.001.

Altered characteristics of cellular infiltration. Mice with early-phase T. spiralis infection did not displaygreatly altered kinetics of cellular infiltration into the pulmonaryenvironment following inoculation with influenza virus but hadreduced cell numbers in the BAL fluid at day 2 p.i. and in thelungs at day 7 p.i. (Fig. 3). As with other parameters, therewas no significant alteration in cellular infiltration wheninfluenza virus was administered at day 60 of parasite infection,and this again was highly reproducible (data not shown). Wetherefore focused our attention on phenotypic alterations observedduring early-phase trichinosis. Early influx of neutrophilsinto BAL fluid is a characteristic response during influenza,and this was recapitulated in these experiments, peaking atday 2 but reduced by over 50% as a result of parasite coinfection.Levels of lymphocytes and macrophages were also elevated atlater time points, and although the numbers of lymphocytes werelower in coinfected mice, there was no statistically significantdifference, whereas macrophage numbers were unaffected. Eosinophilnumbers were very low but were slightly enhanced by parasiteinfection at day 10 (Fig. 4).

View larger version (15K):
[in this window]
[in a new window]
FIG. 3. Cellular infiltration into the lungs is suppressed by T. spiralis. Mice were inoculated with influenza virus 7 days following infection with T. spiralis. Total cells recovered from the BAL fluid (A) and lungs (B) were enumerated over the course of viral infection. Values represent means ± 1 SD, and significant differences to mice infected with influenza virus alone are shown. Filled circles, influenza infection alone; open circles, coinfected animals. *, P $≤$ 0.05; ***, P $≤$ 0.001.

View larger version (19K):
[in this window]
[in a new window]
FIG. 4. Parasite infection inhibits the early influx of neutrophils. Cells recovered from BAL fluid at different times after influenza infection were stained with hematoxylin and eosin following cytospin preparation. Neutrophils, eosinophils, lymphocytes, and macrophages were enumerated. Data are presented from mice infected solely with influenza virus and from those infected 7 days earlier with T. spiralis. The values shown are means ± 1 SD, and significance values for coinfected mice (open circles) are expressed in comparison to mice infected with influenza virus alone (filled circles). ***, P $≤$ 0.001.

Coinfection did result in reduced numbers of lymphocytes inthe lungs at day 7. This was not restricted to a single cellsubset, as numbers of CD4⁺, CD8⁺, and DX5⁺ (NK) cells were allsignificantly lower than controls at this time (Fig. 5). Interestingly,the kinetics of NK cell influx was altered, peaking at day 4,and it was notable that there was already an elevated residentpopulation in parasite-infected mice prior to inoculation withinfluenza virus. Expression of CD45RB progressively decreasedon CD4⁺ lymphocytes during influenza infection as would be expectedduring activation, but the kinetics of expression of CD4⁺ CD45RB^hidid not differ between groups. In a similar manner, the proportionof CD8⁺ CD45RB^hi cells was decreased at day 10 after influenzainfection in both singly infected and coinfected animals (datanot shown).

View larger version (21K):
[in this window]
[in a new window]
FIG. 5. Reduction and altered kinetics of lymphocyte infiltration. Mice were infected with T. spiralis and with influenza A virus 7 days later, or influenza A virus alone, as described. Single-cell suspensions from the lungs were analyzed by flow cytometry, and total numbers of lymphocytes and CD4⁺, CD8⁺, and DX5⁺ cells were determined. Numbers are expressed as means ± 1 SD, and significant differences between singly infected (filled circles) and coinfected (open circles) mice are shown. **, P $≤$ 0.01; ***, P $≤$ 0.001.

Histological examination confirmed that pulmonary inflammationwas ameliorated by coinfection with T. spiralis. Mice whichhad been infected with parasites for 1 week did not show anynoticeable alterations (Fig. 6A and B), and the same appliedfor animals infected for 2 days with influenza virus (Fig. 6C and D).By day 7 after influenza infection, however, there wasan intense infiltration of leukocytes surrounding the airways,and a cellular exudate was evident in the lumen of the bronchioles(Fig. 6E). This was primarily mononuclear at this stage, withsome polymorphonuclear cells. Examination at higher magnificationrevealed that the alveolar spaces frequently contained erythrocytes(Fig. 6G). In contrast, animals coinfected with T. spiralisshowed a greatly reduced cellular infiltration in pulmonarytissues (Fig. 6F), with both the bronchial and alveolar spacesremaining largely clear (Fig. 6H).

View larger version (126K):
[in this window]
[in a new window]
FIG. 6. Amelioration of cellular infiltration and airway occlusion. Lungs were fixed, sectioned, and stained with hematoxylin and eosin. (A, C, E, and G) Tissue from mice infected with influenza virus alone is shown; (B, D, F, and H) tissue from mice preinfected with T. spiralis is shown. Representative sections from each group are shown at day 0 (A and B), day 2 (C and D), and day 7 (E and F) after influenza infection at 40x magnification. Sections of lungs at day 7 p.i. are also shown at 100x magnification (G and H). Bronchioles (b), alveoli (a), and blood vessels (bl) are shown. Infiltrating cells (i) are indicated, as are erythrocytes (e) in the alveolar spaces.

We had previously observed that infection of NIH mice with T.spiralis resulted in high levels of IL-10 produced by splenocytesand mesenteric lymph nodes at day 7 p.i. (15), and thus we examinedthe phenotype of T cells recruited into the lungs during coinfectionby intracellular cytokine staining. There was no alterationin the relative proportion of IL-10- or IFN- ${gamma}$ -positive lymphocytes,however. The numbers of CD4⁺ cells secreting these cytokinesand CD8⁺ cells secreting IL-10 were reduced (Fig. 7). The concentrationof IL-10 in BAL fluid was also significantly lower at this timepoint (Fig. 8). Early production of tumor necrosis factor alpha(TNF- ${alpha}$ ) was inhibited (Fig. 8), which may be significant as thishas been shown to be a contributory factor in illness causedby influenza (20).

View larger version (24K):
[in this window]
[in a new window]
FIG. 7. Infiltration of both IFN- ${gamma}$ - and IL-10-secreting lymphocytes is suppressed by parasite infection. Samples were analyzed by flow cytometry and the numbers of IFN- ${gamma}$ - and IL-10-positive lymphocytes in the lung determined by intracellular cytokine staining. Numbers are expressed as means ± 1 SD, and significant differences between singly infected (filled circles) and coinfected (open circles) mice are shown. ***, P $≤$ 0.001.

View larger version (13K):
[in this window]
[in a new window]
FIG. 8. Prior infection with T. spiralis suppresses IL-10 and TNF- ${alpha}$ in BAL fluid. Concentrations of cytokines were determined by enzyme-linked immunosorbent assay. The values shown represent the means ± 1 SD, and significant differences between singly infected (filled bars) and coinfected (open bars) mice are shown. ***, P $≤$ 0.001.

Cellular turnover during coinfection. It was clear that the rapid recovery of coinfected mice wasassociated with lower numbers of a broad spectrum of cells infiltratingpulmonary tissues, but the mechanisms underlying this phenomenonwere unclear. We therefore examined the rate of apoptosis andproliferation in cells isolated from the MdLN, the primary siteof lymphocyte expansion in influenza infection, as alterationsin these parameters had previously been observed to contributeto decreased immunopathology in murine influenza (19). The proportionof apoptotic CD4⁺ cells was increased in influenza virus-infectedmice at 7 and 10 days after inoculation relative to uninfectedmice, but it was not significantly different in those animalscoinfected with T. spiralis (Fig. 9A). The proportion of apoptoticCD8⁺ lymphocytes in naive mice was extremely low and was notsignificantly affected in singly infected or coinfected mice(Fig. 9B). The possibility that decreased cell numbers in coinfectedanimals might be due to lower levels of lymphocyte proliferationwas investigated by administering mice with BrdU in their drinkingwater. The total numbers of proliferating cells in MdLN wereelevated in both singly infected and coinfected groups at day4 and day 7, but no significant difference between these groupswas evident (Fig. 9C).

View larger version (10K):
[in this window]
[in a new window]
FIG. 9. Reduced numbers of T lymphocytes in the lungs are not accounted for by enhanced apoptosis or suppression of proliferation. Mice were infected orally with T. spiralis and 7 days later with influenza A virus as described. Single-cell suspensions from the MdLN were analyzed by flow cytometry and apoptotic cells defined as annexin⁺ and To-pro-3^–. The proportions of apoptotic CD4⁺ (A) and CD8⁺ (B) lymphocytes were determined. Cellular proliferation was assessed by incorporation of BrdU and flow cytometry (C). No significant differences were observed between singly infected (filled bars) or coinfected (open bars) animals in any case.

DISCUSSION

Top

Discussion

This study demonstrates that coinfection of mice with influenzavirus during the early phase of trichinosis results in a reducedinflammatory infiltrate in the lungs, decreased concentrationsof cytokines in the BAL fluid, and faster recovery of weight,but it does not impede upon the mechanisms of viral clearance.The latter point indicates that generalized pulmonary inflammationdoes not contribute to viral clearance per se but representsan immunological background within which antiviral immunityoperates or that the inflammatory response is magnified severalfoldover what is necessary and sufficient for effective clearance(19). With respect to parasite infection, the timing was observedto be crucial for amelioration of pathology, as there was noeffect when parasites were encysted in skeletal muscle.

An integral component of the adaptive immune response to viralinfection is the generation of specific cytotoxic T cells. Itmight be predicted that helminth infection would impair theirdevelopment via inhibition of IL-2 and IFN- ${gamma}$ production, andexperimental evidence suggests that this can occur in givensituations (1). In the current experiments, viral clearancewas unimpaired even though the total numbers of NK⁺ and CD8⁺lymphocytes in the lungs were decreased. NK cells have beenshown to play a role in immunity to influenza (4). Multiplemechanisms contribute to viral clearance, however, and eventhe total absence of IFN- ${gamma}$ or IFN- ${alpha}$ /ß is not sufficientto prevent eradication of influenza virus in mice (17, 31),indicative of considerable redundancy in the response (19).

For the large part, parasite infection did not induce alterationsin the pulmonary cell infiltrate prior to inoculation with influenzavirus, although the numbers of DX5⁺ cells were elevated. Thisis consistent with previous reports of enhanced NK cell cytolyticactivity in the lungs during the period in which newborn larvaeare migrating from the intestine to skeletal muscle (2, 30).Although this site is not actively occupied by the parasite,Korten et al. (24) observed that numbers of DX5⁺/CD3^–NK cells and DX5⁺/CD3⁺ T cells were expanded systemically duringLitomosoides sigmodontis infection, and thus it is possiblethat the systemic release of parasite products and/or localcytokine and chemokine activity resulting from migration ofnewborn larvae through the capillary bed lining the lungs invokesrecruitment and activation of this population.

The suppression of pulmonary pathology and cellular infiltrationobserved here could be due to a variety of factors, such asa state of immune exhaustion as a result of ongoing intestinalinflammation, although if this were a contributing factor, thenthe same effect might have been expected during the systemicphase of T. spiralis infection. Alternatively, lower numbersof cells could have been accounted for by increased apoptosisor reduced levels of cellular proliferation, via interferencewith lymphocyte activation in the regional lymph nodes. OX40-immunoglobulinfusion proteins block the interaction of OX40 (CD134) on activatedT cells with its ligand on antigen-presenting cells. Directedblockade of this costimulatory pathway via administration ofOX40-immunoglobulin prevented the loss of weight which accompaniesmurine influenza without impairing viral clearance and was associatedwith reduced proliferation and enhanced apoptosis of T cellsin the lungs and MdLN (19). The similarities with the currentstudy prompted us to examine these two parameters, and althoughthe proportion of apoptotic CD4⁺ cells increased following inoculationwith influenza virus, this was not perturbed by concomitantparasite infection.

IL-10 is a potent anti-inflammatory cytokine which operateson many levels, including the inhibition of lymphocyte activationand macrophage function. Given that IL-10 is a dominant featureof helminth infection and has been shown to prevent necrosisand inflammation during T. spiralis infection (3, 5), this seemeda likely candidate for amelioration of the lung pathology seenhere. Perhaps somewhat surprisingly, reduced numbers of IL-10-secretinglymphocytes were observed in the lungs and lower amounts ofcytokine were present in the BAL fluid of coinfected mice. Thelower amounts of cytokine could simply be due to reduced numbersof local T cells, although we do not know to what extent otherIL-10-producing cells contribute to the situation. Parasiteinfection did not downregulate the capacity of cells to respondto influenza, as incorporation of BrdU indicated elevated numbersof proliferating cells in the MdLN, with no evident differencein coinfected mice. Other studies have shown that helminth infectioncan result in suppression of inflammation in an IL-10-independentmanner. Thus, H. polygyrus has been documented to inhibit establishedcolitis in IL-10-deficient mice. Resolution of inflammationcould be adoptively transferred by T cells from mesenteric lymphnodes, and enhanced expression of Foxp3 in these cells impliesthat they may display regulatory activity (12). IL-10-deficientmice injected intravenously with T. spiralis newborn larvaeshow a marked increase in the numbers of inflammatory cellsaround parasites in skeletal muscle. This peaked at around 20days but had resolved by 55 days p.i., indicative of an IL-10-independentmechanism for resolution of inflammation during late-stage infection(3).

Production of proinflammatory cytokines contributes to the symptomsobserved during acute influenza, which can be ameliorated bythe use of blocking antibodies against IFN- ${alpha}$ /ß, IL-1 ${alpha}$ ,and TNF- ${alpha}$ (20, 25). In the current study, TNF- ${alpha}$ in BAL fluid peakedearly at day 2 p.i., which is consistent with both murine andhuman studies (14, 28), and it was significantly suppressedin coinfected mice. Depletion of TNF- ${alpha}$ during infection withrespiratory syncytial virus or influenza A virus abolished illnessand weight loss, resulted in decreased inflammatory infiltrationin the lungs, and reduced the expression of IFN- ${gamma}$ on CD4⁺ cells,but it did not alter viral clearance (20). It is therefore possiblethat the decreased levels of TNF- ${alpha}$ observed in this study resultin decreased expression of adhesion molecules on the vascularendothelium and/or reduced chemokine and chemokine receptorexpression, resulting in a net suppression of inflammation.Alleviation of the general increase in vascular permeabilityin pulmonary tissues was also evidenced by suppression of serumalbumin in BAL fluid and the lack of erythrocytes in the alveolarspaces of coinfected mice.

Trichinella pseudospiralis has been observed to suppress inflammationin coinfection studies and also results in reduced cellularrecruitment around subcutaneously implanted material (35). Earlydownregulation of both the latter response and delayed-typehypersensitivity to trinitrochlorobenzene was followed laterduring infection by normalization of these parameters relativeto uninfected mice, leading to the conclusion that the immunomodulatoryeffects were associated with the presence of migratory larvaein the host (34). A similar conclusion was reached in a studyin which inflammatory demyelination in a rat model for multiplesclerosis was suppressed by T. pseudospiralis (6). The courseof influenza infection in our experiments is coincident withthe production and dissemination of newborn parasite larvae,and thus, the immunomodulatory effects observed here may haveevolved to minimize tissue damage during parasite migrationand to protect the relatively vulnerable larval stages whenthey are most exposed to immune attack. This is conceptuallysimilar to the immunomodulation seen during the late phase ofinfection with most other helminth parasites, which, due todifferences in parasite biology, also coincides with dispersalof eggs and larval stages (27).

ACKNOWLEDGMENTS

This work was supported by the Wellcome Trust (no. 059532).R.C.F. is the recipient of a Prize Studentship (no. 064610).

We thank Lorraine Lawrence for histological processing and staining.

FOOTNOTES

* Corresponding author. Mailing address: Division of Cell and Molecular Biology, Biochemistry Building, Imperial College London, London SW7 2AZ, United Kingdom. Phone: (44) 20 7594 5214. Fax: (44) 20 7594 5207. E-mail: m.selkirk{at}imperial.ac.uk.

Editor: W. A. Petri, Jr.

REFERENCES

Top