INTRODUCTION
Host protective immunity against gastrointestinal (GI) nematodes, such as the common murine mouse model Heligmosomoides bakeri (previously named Heligmosomoides polygyrus; see Cable et al. Reference Cable, Harris, Lewis and Behnke2006) is mediated by the Th2 cytokines, IL-4 and IL-13, as well as IL-3, 5, 6, 9, 10 (Urban et al. Reference Urban, Noben-Trauth, Donaldson, Madden, Morris, Collins and Finkelman1998; Gause et al. Reference Gause, Urban and Stadecker2003) that induce Th2 effectors including serum IgE, eosinophils and mucosal mast cells (Finkelman et al. Reference Finkelman, Shea-Donohue, Goldhill, Sullivan, Morris, Madden, Gause and Urban1997). Recent studies have also demonstrated that the timing of peak fluid secretion, ion flux and mucus production correspond with loss of worms from the intestine (Shea-Donohue et al. Reference Shea-Donohue, Sullivan, Finkelman, Madden, Morris, Goldhill, Pineiro-Carrero and Urban2001), that IL-13 regulated expression of a Goblet cell derived factor RELM-β disrupts the chemosensory apparatus of GI–dwelling nematodes (Artis et al. Reference Artis, Wang, Keilbaugh, He, Brenes, Swain, Knight, Donaldson, Lazar, Miller, Schad, Scott and Wu2004) and that IL-13 induced intestinal smooth muscle contractility (Vallance et al. Reference Vallance, Blennerhassett and Collins1997; Au Yeung et al. Reference Au Yeung, Smith, Zhao, Madden, Elfrey, Sullivan, Levander, Urban and Shea-Donohue2005) and accelerated intestinal epithelial cell turnover (Cliffe et al. Reference Cliffe, Humphreys, Lane, Potten, Booth and Grencis2005) may also play an important role in nematode expulsion. Thus, evidence is growing to suggest that Th2-dependent inflammatory responses and altered gut physiology also contribute to worm expulsion.
It is also known that protein deficiency impairs the Th2 responses (Ing et al. Reference Ing, Su, Scott and Koski2000). We have previously reported reduced IL-4 along with several Th2 effectors including eosinophils, mast cells, and serum IgE, but elevation of the Th1 cytokine, IFN-γ, in protein-deficient mice infected with H. bakeri (Ing et al. Reference Ing, Su, Scott and Koski2000). At that time, we had presumed that the prolonged survival of the parasite in these mice was due to impaired Th2 immunological responses. However, given the recent work indicating that inflammatory and/or pathophysiological effectors secondary to the reduced Th2 response may also contribute to worm expulsion (Shea-Donohue et al. Reference Shea-Donohue, Sullivan, Finkelman, Madden, Morris, Goldhill, Pineiro-Carrero and Urban2001; Artis et al. Reference Artis, Wang, Keilbaugh, He, Brenes, Swain, Knight, Donaldson, Lazar, Miller, Schad, Scott and Wu2004; Au Yeung et al. Reference Au Yeung, Smith, Zhao, Madden, Elfrey, Sullivan, Levander, Urban and Shea-Donohue2005; Cliffe et al. Reference Cliffe, Humphreys, Lane, Potten, Booth and Grencis2005), the possibility that the impairment of these secondary responses might contribute to prolonged parasite survival in protein-deficient mice must be considered. Moreover, because pathophysiological responses to infection are normally stronger at higher infection doses, we hypothesized that worm expulsion should occur more rapidly when immunocompetent protein-sufficient mice are infected with higher doses of GI nematodes. Thus the objectives of the current study were to determine (1) if H. bakeri were expelled more rapidly at higher infection doses, (2) if protein deficiency altered the dose-dependent expulsion, and (3) if delayed expulsion of H. bakeri in protein-deficient mice was linked with dysregulated inflammatory responses and/or impaired gut pathophysiological responses.
MATERIALS AND METHODS
Mice and diets
Female, 25-day-old BALB/c mice purchased from Charles River Breeding Laboratories (Quebec, Canada) were housed individually in Nalgene cages (Fisher Scientific, Montreal, Canada) with stainless steel covers in a temperature controlled mouse room (22–25°C) with a 10:14 h light:dark cycle.
Egg-white based semi-purified powdered diets were formulated (Tu et al. Reference Tu, Koski, Wykes and Scott2007) such that energy from protein made up 24% of total calories of the protein-sufficient (PS) diet, providing a diet sufficient for maintenance of normal physiology, metabolism and growth of mice (NRC, 1995). In contrast, a 3% protein diet was used as the protein-deficient (PD) diet; previous experiments in our laboratory have shown that H. bakeri-infected mice fed this diet have reduced body weight gain, increased parasite burdens (Tu et al. Reference Tu, Koski, Wykes and Scott2007) and suppressed Th2 immune responses (Ing et al. Reference Ing, Su, Scott and Koski2000). Both diets were identical for all nutrients except for protein and carbohydrate, the latter of which was adjusted to ensure that both diets were isoenergetic. Diet composition was based on National Research Council mouse requirements (NRC, 1995) to ensure that a 30% reduction in food intake due to infection or protein deficiency would not generate other nutrient deficiencies. Food and tap water were provided ad libitum. All procedures were approved by the McGill Animal Care Committee according to the Canadian Council on Animal Care (1993).
Experimental design and protocol
Using a 2×3 factorial design, a total of 120 mice were randomized into 2 diet groups (PS and PD) at the beginning of the experiment. They were further subdivided into 3 groups at the time of challenge infection (challenged with either 100 L3 or 200 L3, or a sham challenge with no larvae). Ten mice per subgroup were killed at both day 7 and day 14 post-challenge infection (pci).
Initially, all mice were fed commercial Mouse Chow 5015 (Agribands Canada, Ontario, Canada) for 3 days and then acclimatized for 4 days to the PS diet provided in Mouse Powder Feeders (Lab Products, Quebec, Canada) designed to reduce spillage. Following acclimatization to the powder feeders, mice were fed either the PD or PS diet for 1 week and then each mouse was given 100 L3 by oral gavage. L3 were obtained by faecal culture of stock parasites maintained in outbred CD-1 mice (Charles River) and counted in 10 sham doses for accuracy before infecting the mice. At days 9 and 14 post-primary infection (ppi), each mouse was treated with pyrantel pamoate (172 mg/kg bwt, Pfizer Canada Inc., Montreal, Canada) to kill the adult worms. One week later, mice were re-infected with 0, 100 or 200 L3. This drug-abbreviated immunizing protocol normally induces a strong protective immune response (Behnke and Robinson, Reference Behnke and Robinson1985) leading to parasite expulsion within 18 to 21 days after challenge infection (Ing et al. Reference Ing, Su, Scott and Koski2000). Just prior to necropsy on days 7 and 14 pci, mice were anaesthetized using Ketamine/Xylazine (50/5 mg/kg body weight), then exsanguinated by cardiac puncture; sera were frozen at −20°C until later analysis for cytokines and chemokines. At both times, 7 mice per subgroup were anaesthetized 1 h before they were killed and 0·1 ml of 2% Evan's Blue in saline was injected into the tail vein for later measurement of fluid leakage into the duodenum (Ohishi and Odagiri, Reference Ohishi and Odagiri1984). Intestinal tissues from the 3 remaining mice per subgroup were used for histopathology and measurement of gut myeloperoxidase (MPO) activity. Body weight was recorded weekly and food intake was measured every 4 days.
Parasites and fluid leakage into the intestine
At necropsy, the intestine of mice injected with Evan's Blue was removed and cut into 8 equal length sections. The first 3 cm of duodenum was rinsed with 1 ml of saline, any adult worms were removed, and then the flushed fluid was frozen at −20° for later detection of Evan's Blue (Ohishi and Odagiri, Reference Ohishi and Odagiri1984), an index of intestinal mucosal leakage. Briefly, thawed extravasation fluid was made up to 3 ml by adding 0·5% sodium dodecyl sulphate and 10% Na2CO3 to dissolve the mucus. The solution was then transferred to a 1 cm cuvette and the absorbance was measured using a Beckman spectrophotometer (Bio-Tek Instruments, Inc. Winooski, Vermont) at 620 nm. One extravasation unit was defined as 0·001 absorbance units. Each sample was measured in duplicate.
Fourth-stage larvae (L4) embedded in the intestinal mucosa and adult worms were counted and their spatial distribution along the intestine was recorded. At day 14 pci, worm viability was recorded based on response of non-moving worms to the touch of a needle. The median location of L4 and adult worms, the percentage recovery of L4 and adults, and percentage viability were expressed as mean±s.e.m.
Gut histopathology and MPO activity
Histopathology was assessed in the proximal and distal duodenum, defined by Solaiman et al. (Reference Solaiman, Jonah, Miyazaki, Ho and Bhattacharyya2001) as 0–1·0 and 3·5–4·5 cm from the pyloric sphincter, respectively. In each region, two 0·5 cm segments were removed from 3 mice per subgroup immediately after the mouse was killed, and fixed in either 4% paraformaldehyde or Carnoy's fixative. The fixed tissues were dehydrated and embedded in paraffin following standard histological procedures. Sections (5 μm) of paraformaldehyde-fixed tissues were stained with haematoxylin and eosin; eosinophils, Goblet cells, and Paneth cells were counted and villus height and crypt depth were recorded. Toluidine blue staining was done on sections from Carnoy-fixed tissues and mucosal mast cells were counted. Each indicator was expressed as the mean±s.e.m. per 10 villus crypt units (vcu).
In addition, two 2 cm segments (1·5 and 4·5 cm along the intestine from the pyloric sphinter) were immediately removed, blotted dry and weighed. Gut MPO activity was detected using an MPO assay kit (Cytostore, Calgary, Alberta, Canada, Cat. ASA-001) as a biochemical marker of intestinal inflammation (Bradley et al. Reference Bradley, Priebat, Christensen and Rothstein1982). Briefly, the segment was put in hexadecyltrimethyl ammonium bromide buffer in a ratio of 50 mg of tissue per 0·5 ml, homogenized and then centrifuged. Then 20 μl of supernatant was added to each of 2 wells of the ELISA plate, followed by 200 μl of development reagent (4·175 mg o-dianisidine dihydrochloride in 25 ml of potassium phosphate buffer and 12·5 μl of H2O2). The absorbance at 450 nm in each well was measured twice using a Beckman spectrophotometer, once immediately after adding the developing reagent and 1 min later in order to record the change of absorbance. One unit of MPO activity was defined as the quantity catalysing the decomposition of 1 μmol of H2O2 to water per min at 25°C (Bradley et al. Reference Bradley, Priebat, Christensen and Rothstein1982) and the activity was expressed in units per mg of tissue.
Serum cytokines and chemokines
We selected 2 cytokines to measure quantitatively by ELISA: monocyte chemotactic protein-5 (MCP-5) as an inflammatory cytokine recently linked with nematode infections (Sarafi et al. Reference Sarafi, Garcia-Zepeda, MacLean, Charo and Luste1997) and leptin as a chemokine tied both to food intake (Flier, Reference Flier1997) and Th1 responses (Lord et al. Reference Lord, Matarese, Howard, Baker, Bloom and Lechler1998). The mouse MCP-5 immunoassay kit (R&D Systems, Inc. MN, USA, Cat. MCC120) was used according to the manufacturer's instructions. Briefly, 50 μl of standard, control or sample was added to anti-mouse MCP-5 pre-coated ELISA plates that were then incubated for 2 h at room temperature. After 5 washes, wells were incubated for 2 h with 100 μl of mouse MCP-5 conjugate, followed by a 30-min incubation with the substrate 3,3′,5,5′-tetramethylbenzidine (TMB). The reaction was stopped with diluted HCl. The plates were read with an EL 309 Microplate Autoreader (Bio-Tek Instruments, Inc. Winooski, Vermont) at both 450 nm and 540 nm. The reported OD value was obtained by subtracting readings at 540 nm from the readings at 450 nm. Each sample was measured in duplicate.
Serum leptin was measured using a mouse leptin ELISA kit (Linco Research, Missouri, USA, Cat. EZML-82K) following the manufacturer's instructions. Briefly, 10 μl of samples were added to the microtitre plate and then 50 μl of pre-titred anti-rodent leptin serum was added to each well. After incubation for 2 h, followed by 3 washes, 100 μl of pre-titred biotinylated anti-mouse leptin antibody were added to each well and the plate was then incubated for 1 h. Then 100 μl of pre-titred streptavidin-horseradish peroxidase conjugate were added and the plate was incubated for 30 min. Substrate TMB was then added, the plate was incubated for 30 min and the reaction was stopped by addition of 0·3 m HCl. The plates were read with an EL 309 Microplate autoreader at 450 nm and 590 nm. The OD value was calculated by subtracting the readings at 590 nm from the readings at 450 nm. Each sample was measured in duplicate.
In addition, we screened for broader effects of protein deficiency and challenge dose on a wide range of cytokines and chemokines in nematode-infected mice on day 14 pci. An equal amount of serum from each mouse within each subgroup was pooled and diluted 1:10 in blocking buffer and tested in duplicate. The pooled sera were assayed, according to the manufacturer's instructions, using a RayBioTM Mouse Cytokine Antibody Array II (RayBiotech Inc., Norcross, Ga, USA). The intensity of each spot was calculated by subtracting the average intensity of the background around each spot from the intensity of the spot itself as read by densitometry (Versa Doc Imaging system, Bio-Rad, USA). The intensity of the positive control was assigned a value of 100, and the intensity of the lightest spot was assigned a value of 0. These were used as reference values to determine the relative intensity for each cytokine, then duplicate values were averaged. If the relative intensity of the sample from mice challenged with 100 larvae was at least 50% lower than the sample from unchallenged mice, we considered the cytokine to be down-regulated whereas if the relative intensity was at least double that of unchallenged mice, we considered it to be up-regulated (Klein et al. Reference Klein, Paul, Angele, Popp, Pfister and Koedel2006). Similarly, we compared responses between mice challenged with 200 versus 100 larvae, and between PS and PD mice at each dose of larval challenge.
Statistics
Data on the number of worms, percentage viable adults, percentage recovery of worms and spatial distribution of parasites were analysed using 2×2 ANOVA with main effects of diet (PD vs PS) and larval dose (100 vs 200) whereas all other data were analysed using 2×3 ANOVA with main effects of diet (PD vs PS) and larval dose (0, 100 and 200), with Scheffe's post-hoc test. When there was no significant diet∗dose interaction, the data were pooled according to main effects, and re-analysed. Intestinal mucosal leakage was analysed using linear regression with log worm numbers as the independent variable. All analyses were done with SAS 9·01 and the analyses were considered significant at P<0·05.
RESULTS
Initial body weights were similar, but PD mice gained less weight than PS mice, despite higher cumulative food intake (Table 1). Cumulative food intake and weight gain were unaffected by dose of challenge infection (data not shown).
Table 1. Comparison of nutritional indicators and parasite numbers between protein-deficient (PD) and protein-sufficient (PS) mice at days 7 and 14 days post-challenge infection (pci)
(Values are mean±s.e.m. Means with different superscripts differ significantly between PD and PS mice (P<0·05).)
1 Values pooled across the 3 challenge infection doses.
Parasite outcomes
Worm establishment and/or survival over the first week of challenge infection was higher in PD mice challenged with 200 L3 compared with those challenged with 100 L3 (Table 1). Also, as expected based on our previous work (Ing et al. Reference Ing, Su, Scott and Koski2000), a higher percentage of total worms and viable worms were recovered on day 14 pci in PD mice challenged with both 100 and 200 L3, compared with PS mice (Fig. 1B).
Fig. 1. The influence of dose of infection and diet on Heligmosomoides bakeri location in the intestine and survival on day 7 and day 14 post challenge infection (pci). (A) Median location of worms along the intestinal tract (main effect of diet: F1,24=21·35, P<0·0001; main effect of dose: F1,24=8·65, P=0·0071; interaction of diet∗dose: F1, 24=8·65, P=0·0071); (B) Left axis – percentage viable worms (main effect of diet: F1,24=4·63, P=0·04); right axis – percentage recovery of worms (main effect of diet: F1,24=5·05, P=0·03). Mice were fed either a protein-sufficient (PS, ■) or protein-deficient (PD, ▲) diet and given a challenge infection with either 100 (cross-hatched bars) or 200 (solid bars) larvae. Different letters indicate significant differences within a time-point (P<0·05).
Our hypothesis that worm expulsion would be more rapid in PS mice challenged with 200 compared with 100 L3 was in part supported by the more posteriorad location of worms on day 14 pci in PS mice challenged with 200 L3 (Fig. 1A), but not by the percentage viable adults nor percentage recovery (Fig. 1B). In contrast, we found no evidence for dose-dependent worm expulsion in PD mice. Regardless of challenge dose, parasites at day 14 pci were found in the anterior small intestine (Fig. 1A) and the percentage viable adults or percentage recovery (Fig. 1B) were similar between PD mice challenged with 100 and 200 L3.
Gut histopathology
Compared with unchallenged mice, challenge with 100 or 200 L3 induced elevation in numbers of mucosal mast cells (Fig. 2A), eosinophils (Fig. 2C), Goblet cells (Fig. 2E) and Paneth cells (Fig. 2G) in the proximal duodenum, at both day 7 and day 14 pci. Neither diet (PD vs PS) nor challenge dose (100 vs 200 L3) influenced cell populations in this region. A broadly similar pattern was observed in the distal region of the duodenum (Fig. 2B, D, F, H), although eosinophilia was not detected in challenged PD mice on day 7 pci (Fig. 2D), significant elevation of Goblet cell numbers only occurred in mice challenged with 200 L3 (Fig. 2F), and Paneth cell numbers were unaffected by challenge infection on both day 7 and day 14 pci (Fig. 2H). In the distal duodenum, mucosal mast cells (Fig. 2B) and eosinophils (Fig. 2D) were less abundant in PD compared with PS mice, but only on day 7 pci.
Fig. 2. Histopathology of the proximal (A, C, E, G) and distal (B, D, F, H) duodenum induced by protein deficiency and Heligmosomoides bakeri challenge infection on days 7 and 14 post-challenge infection (pci). (A, B) Mucosal mast cells (MMC); (C, D) eosinophils; (E, F) Goblet cells; (G, H) Paneth cells. Cell populations are reported as mean numbers per 10 villus crypt units. PS: protein sufficient; PD: protein deficient; challenge dose of 0 L3 (open bars), 100 L3 (cross-hatched bars) or 200 L3 (solid bars). Different uppercase letters represent significant differences among challenge doses, based on post-hoc analyses following a significant main effect of dose. A significant main effect of diet is represented by a∗. Different lowercase letters represent significant differences among dose and diet groups within a time-point, based on Scheffe's post-hoc comparison, if there was a significant interaction. (A) Day 7 pci – main effect of dose: F2,12=34·88, P<0·0001; interaction of diet∗dose: F2,12=4·42, P=0·0364; day 14 pci – main effect of dose: F2,12=8·74, P=0·0045; (B) day 7 pci – main effect of diet: F1,12=5·30, P=0·04; main effect of dose: F2,12=8·92, P=0·0042; day 14 pci – main effect of dose: F2,12=9·18, P=0·0038; (C) day 7 pci – main effect of dose: F2,12=13·82, P=0·0008; day 14 pci – main effect of dose: F2,12=12·46, P=0·0012; (D) day 7 pci – main effect of dose: F2,12=46·26, P=0·0001; day14 pci – main effect of dose: F2,12=10·25, P=0·0025; (E) day 7 pci – main effect of dose: F2,12=7·79, P=0·0068; main effect of diet; F1,12=49·06, P<0·0001; interaction diet∗dose: F2,12=12·42, P=0·0012; day14 pci – main effect of dose: F2,12=9·14, P=0·0039; (F) day 7 pci – main effect of dose: F2,12=7·13, P=0·0091; day 14 pci – main effect of dose: F2,12=8·30, P=0·0055; (G) day 7 pci – main effect of dose: F2,12=4·19, P=0·0417; day 14 pci – main effect of dose: F2,12=9·37, P=0·0035.
Although gut MPO was higher in challenged mice than unchallenged mice, indicating higher intestinal inflammation in both regions of the duodenum (Fig. 3A, B), a graded response of MPO to challenge dose was detected only in the distal duodenum on day 7 pci (Fig. 3B). In addition, PD reduced gut MPO in the distal duodenum at both time-points (Fig. 3B).
Fig. 3. Gut myeloperoxidase (MPO) activity (A, B) and villus height (C, D) of the proximal (A, C) and distal (B, D) duodenum induced by protein deficiency and Heligmosomoides bakeri challenge infection on days 7 and 14 post challenge infection (pci). PS: protein sufficient; PD: protein deficient; challenge dose of 0 L3 (open bars), 100 L3 (cross-hatched bars) and 200 L3 (solid bars). Different letters represent significant differences among challenge doses, based on post-hoc analyses following a significant main effect of dose. A significant main effect of diet is represented by a *. (A) Day 7 pci – main effect of dose: F2,12=3·99, P=0·047; day 14 pci – main effect of dose: F2,12=7·58, P=0·0074; (B) day 7 pci – main effect of diet: F1,12=5·61, P=0·0355; main effect of dose: F2,12=22·90, P<0·0001; day 14 pci – main effect of diet: F1,12=5·65, P=0·0349; main effect of dose: F2,12=23·42, P<0·0001; (C) day 7 pci – main effect of dose: F2,12=11·07, P=0·0019; day 14 pci – main effect of diet: F1,12=5·39, P=0·0387; main effect of dose: F2,12=12·95, P=0·001.
We had hypothesized that protein deficiency would reduce villus height (Deo et al. Reference Deo, Sood and Ramalingaswami1965; Dudley, Reference Dudley, Wykes, Dudley, Fiorotto, Burrin, Rosenberger, Jahoor and Reeds1997). However, villi in the proximal duodenum were longer (not shorter) in PD mice than in PS mice on day 14 pci (Fig. 3C). Interestingly, challenge infection induced a similar degree of villus atrophy on both day 7 and day 14 pci in the proximal (Fig. 3C) duodenum, regardless of challenge dose (100 and 200). However, further along the intestine, there was no evidence of villus atrophy in response to either infection or protein deficiency (Fig. 3D). Crypt depth was not affected by either infection or diet in either region of the duodenum (data not shown). Together, these data demonstrate that only gut MPO was more severe in mice infected with 200 compared with 100 L3. In addition, the presence of a challenge infection reversed the detrimental effect of protein deficiency on villus height in the proximal duodenum.
Analysis of extravasation units of Evan's Blue in duodenal washes on days 7 and 14 pci revealed no main effect of protein deficiency or of dose of larval challenge (data not shown). However, linear regression showed that fluid leakage on day 14 pci increased with increasing worm burdens in PS mice (Fig. 4), a result consistent with the more posteriorad distribution of worms in the intestines of PS mice infected with 200 compared with 100 L3. In contrast, fluid leakage decreased with increasing worm burdens in PD mice (Fig. 4).
Fig. 4. Relationship between gut permeability as indicated by extravasation units of Evan's Blue in the duodenum and number of Heligmosomoides bakeri in protein-sufficient (□) and protein-deficient (▲) mice using data pooled across infection doses of 100 and 200 on day 14 post-challenge infection. Linear regression equation for PS mice: Y=0·4249X+0·5107, R2=0·2947, P=0·04. Linear regression equation for PD mice: Y=−0·9664X+2·517, R2=0·492, P=0·0075. The slopes differed significantly (P=0·0025). One outlier in PD mice (filled spot) was excluded from the analyses.
Cytokines
Quantitative comparison of serum leptin revealed an overall diet effect with higher concentrations in PD mice than in PS mice, but no effect of dose of larval challenge (Fig. 5A). On the other hand, serum MCP-5 was significantly affected by both dose and diet; MCP-5 was higher in challenged mice than unchallenged mice at both time-points in both diet groups (Fig. 5B). On day 14 pci, MCP-5 was further elevated in mice infected with higher numbers of larvae, and levels were higher in PS than in PD mice.
Fig. 5. Serum concentrations of leptin (A) and monocyte chemoattractant protein 5 (MCP-5) (B) in protein-deficient (PD) and protein-sufficient (PS) mice on days 7 and 14 post-challenge infection (pci) with Heligmosomoides bakeri. Challenge dose of 0 L3 (open bars), 100 L3 (cross-hatched bars), and 200 L3 (solid bars). For leptin (A), a significant main effect of diet is represented by a * on day 7 (main effect of diet: F1,54=4·75, P=0·03) and day 14 pci (main effect of diet: F1,54=5·57, P=0·02). For MCP-5 (B), different lowercase letters represent significant differences (P<0·05) among challenge doses within diet groups, based on post-hoc analyses following a significant main effect of dose (F2,54=21·61, P<0·0001). The significant main effect of diet is represented by a ∗ (F1,54=4·25, P=0·04).
We used a multiple cytokine array to screen pooled serum samples in both experiments. Only 1 cytokine was differentially expressed between unchallenged PD and PS mice; leptin was higher in PD mice than in PS mice (data not shown). In addition, only 3 cytokines were differentially expressed between challenged PD and PS mice challenged with 100 L3; GCSF was higher in PD mice but values for IL-5 and MCP-5 were reduced in PD mice (data not shown).
In contrast to the restricted effect of protein deficiency on just a few cytokines in unchallenged mice, challenge infection of PD mice modified a wide range of cytokines (Table 2). Comparison between unchallenged PD mice and PD mice challenged with 100 L3 showed that infection up-regulated a very broad range of cytokines, including leptin, GCSF, GM-CSF, eotaxin, IL-9, IL-12P40P70, IL-12P70, IL-13, IL-17, IFN-γ, KC, MIP-2 and MIP-3β, and down-regulated only 2 chemokines (MCP-1 and VEGF). The effects of challenge infection were more limited in PS mice, where expression of only IL-5, CTACK, MCP-1, TPO and VEGF were down-regulated. Also, IL-4 and IL-13 levels were not detectably elevated in challenged PS mice compared with unchallenged mice, perhaps because of persistence of these cytokines from the primary infection in the unchallenged mice.
Table 2. Effect of challenge dose on cytokine expression in protein-sufficient (PS) and protein-deficient mice (PD) challenged with Heligmosomoides bakeri
(Cytokines were measured using a RayBioTM Mouse Cytokine Antibody Array, on day 14 post-challenge infection. ↑: increase of at least double in first of the two comparison groups; ↓: decrease of at least 50% in first of the two comparison groups.)
1 IL: interleukin; CTACK: cutaneous T cell-attracting chemokine; MCP-1: monocyte chemotactic protein 1; TPO: thrombopoietin; VEGF: vascular endothelial growth factor; TIMP-1: tissue inhibitor of metalloproteinases-1; GCSF: granulocyte-colony stimulating factor; GM-CSF: granulocyte-macrophage colony stimulating factor; INF: interferon; KC: keratinocyte-derived chemokine; MIP-2: macrophage inflammatory protein 2; MIP-3β: macrophage inflammatory protein 3β.
Interestingly in both diet groups, higher dose challenge (comparison between doses of 100 and 200 L3) did not further up-regulate expression of most cytokines (Table 2). Rather, with the exception of VEGF in PS mice, those cytokines responding to dose were down-regulated at the higher infection dose (IL-6 in both PS and PD mice and TIMP-1, GCSF, GM-CSF and IL-17 in PD mice).
DISCUSSION
The immune expulsion of GI nematodes is usually associated with a dominant Th2 immune response (Gause et al. Reference Gause, Urban and Stadecker2003). Recently, however, there has been growing interest in other mechanisms that may be involved in expulsion including intestinal inflammation (Lawrence et al. Reference Lawrence, Kennedy, Garside, Kennedy and Harnett2001), mastocytosis (McDermott et al. Reference McDermott, Bartram, Knight, Miller, Garrod and Grencis2003), increased epithelial cell turnover (Cliffe et al. Reference Cliffe, Humphreys, Lane, Potten, Booth and Grencis2005), hyperplasia of Goblet cells (Artis et al. Reference Artis, Wang, Keilbaugh, He, Brenes, Swain, Knight, Donaldson, Lazar, Miller, Schad, Scott and Wu2004), increased mucous production (Khan and Collins, Reference Khan and Collins2004), and elevated gut permeability, ion flux and luminal fluid secretion (Shea-Donohue et al. Reference Shea-Donohue, Sullivan, Finkelman, Madden, Morris, Goldhill, Pineiro-Carrero and Urban2001; Madden et al. Reference Madden, Whitman, Sullivan, Gause, Urban, Katona, Finkelman and Shea-Donohue2002). In H. bakeri challenge infection, expulsion of worms is normally completed within 18 to 21 days pci (Ing et al. Reference Ing, Su, Scott and Koski2000), and does not begin before the fourth-stage larvae have moved from the serosal musculature into the intestinal lumen. This migration to the intestine is delayed compared with a primary infection (Behnke and Parish, Reference Behnke and Parish1979) but is usually completed by day 12 pci (Urban et al. Reference Urban, Maliszewski, Madden, Katona and Finkelman1995; Gause et al. Reference Gause, Urban and Stadecker2003). We had hypothesized that more rapid expulsion would occur in PS mice challenged with 200 larvae than with 100 larvae, but that a dose-dependent effect would not be evident in PD mice given previous reports that protein deficiency induces immunosuppression in mice and delays expulsion (Ing et al. Reference Ing, Su, Scott and Koski2000). In addition, we wished to relate the active expulsion to potential inflammatory and histopathological changes in the mouse. Therefore, we examined responses on day 7 pci, when the parasites would still be embedded in the serosal musculature, and day 14 pci when active expulsion of H. bakeri has been reported (Urban et al. Reference Urban, Maliszewski, Madden, Katona and Finkelman1995; Ing et al. Reference Ing, Su, Scott and Koski2000) in Balb/c mice.
Although our data on percentage worm recovery on day 14 pci did not support the hypothesis of dose-dependent expulsion in PS mice, H. bakeri were located more distally along the intestine in the mice challenged with 200 compared with 100 larvae, and were thus displaced from the preferred location in the proximal duodenum (Bansemir and Sukhdeo, Reference Bansemir and Sukhdeo1996). This may indicate that the more distally located adults in PS200 mice were in the process of being expelled, and this is consistent with the reduced viability of adult worms in PS mice compared with PD mice. Nevertheless, we cannot discount the possibility that at least some of these worms might have remained in the mice for several more days. In PD mice, our data are entirely consistent with the hypothesis that dose-dependent expulsion would not be evident in PD mice.
Our infected mice fed the protein-deficient diet had elevated food intake, and lower body weights. PD mice are known to increase their total food intake as a physiological response to low dietary protein (Swick and Gribskov, Reference Swick and Gribskov1983; Colombo et al. Reference Colombo, Cervantes, Kokorovic, Pfister and Perritaz1992; Deschepper and de Groote, Reference Deschepper and De Groote1995; Du et al. Reference Du, Higginbotham and White2000) or to increase their food intake relative to body weight (Rothwell and Stock, Reference Rothwell and Stock1987; Zhao et al. Reference Zhao, Jorgensen, Gabert and Eggum1996) in order to meet their needs for protein, nitrogen and amino acids (Webster, Reference Webster1993). However, despite increased food intake, low protein intake is unlikely to be sufficient to maintain lean body mass; instead the relative propotion of body fat increases (Meyer, Reference Meyer1958; Swick and Gribskov, Reference Swick and Gribskov1983; Du et al. Reference Du, Higginbotham and White2000). Although we did not measure body fat directly, we measured leptin, a chemokine that is predominantly released by adipocytes and is typically used as an index of body fat (Flier, Reference Flier1997; Singhal et al. Reference Singhal, Farooqi, O'Rahilly, Cole, Fewtrell and Lucas2002). Leptin was elevated in our PD mice, an observation consistent with deposition of body fat driven by increased intake of a PD diet observed in uninfected rats (Du et al. Reference Du, Higginbotham and White2000).
One of the more surprising results from this study was the up-regulation in infected PD mice of a wide range of cytokines, predominantly pro-inflammatory cytokines and chemokines associated with a Th1 response that is known to promote H. bakeri survival (Ing et al. Reference Ing, Su, Scott and Koski2000; Gause et al. Reference Gause, Urban and Stadecker2003). Researchers have reported a variety of adaptations to a protein-deficient diet associated with conserving protein, including reduced whole body protein synthesis and breakdown (Golden et al. Reference Golden, Waterlow and Picou1977; Wykes et al. Reference Wykes, Fiorotto, Burrin, Del Rosario, Frazer, Pond and Jaboor1996), urea secretion and increased urea recycling (Badaloo et al. Reference Badaloo, Boyne, Reid, Persaud, Forrester, Millward and Jackson1999). The concurrent influence of infection has been reported to increase demand for protein, thus accelerating protein turnover and inducing a negative protein balance (Berclaz et al. Reference Berclaz, Benedek, Jequier and Schutz1996). Neither the conservation of protein nor the increased protein demand applies to all proteins, however (Golden, Reference Golden1982; Jahoor et al. Reference Jahoor, Bhattiprolu, Del Rosario, Burrin, Wykes and Frazer1996), and we found several examples where infection elevated protein levels in PD mice but not in PS mice. Elevated IL-12 p70 promotes a Th1 immune response (Trinchieri, Reference Trinchieri1998), elevated MIP-2 mediates mast cell-dependent inflammation (Biedermann et al. Reference Biedermann, Kneilling, Mailhammer, Maier, Sander, Kollias, Kunkel, Hultner and Rocken2000), and elevated IL-17 induces production of GM-CSF (Cai et al. Reference Cai, Gommoll, Justice, Narula and Fine1998) which has been postulated as a potential mediator of worm expulsion (Shea-Donohue et al. Reference Shea-Donohue, Sullivan, Finkelman, Madden, Morris, Goldhill, Pineiro-Carrero and Urban2001). Others, such as IL-9 (Khan et al. Reference Khan, Richard, Akiho, Blennerhasset, Humphreys, Grencis, Van Snick and Collins2003;Leech and Grencis, Reference Leech and Grencis2006) and eotaxin (Mir et al. Reference Mir, Benahmed, Igual, Borras, O'Connor, Moreno and Rulld2006), are associated with generalized intestinal inflammation. However, we are most intrigued by the possible interrelationships between protein deficiency, leptin, IFN- γ, MPO and MCP-5.
Leptin is now considered as a pro-inflammatory Th1 cytokine as it upregulates IFN-γ production along with a variety of other Th1 cytokines (Faggioni et al. Reference Faggioni, Feingold and Grunfeld2001; Busso et al. Reference Busso, So, Chobaz-Peclat, Morard, Martinez-Soria, Talabot-Ayer and Gabay2002; Fenton et al. Reference Fenton, Hursting, Perkins and Hord2007). Leptin is believed to promote T cell activity at least in vitro, through leptin receptor signalling on naïve T cells (Lord et al. Reference Lord, Matarese, Howard, Baker, Bloom and Lechler1998). Moreover a shift to a predominantly Th1 response is observed when leptin is added to T cells from leptin deficient mice (Lord et al. Reference Lord, Matarese, Howard, Baker, Bloom and Lechler1998). Although we have shown that H. bakeri normally induces a strong Th2 response with low levels of IFN-γ in well-nourished hosts, this cytokine profile is disturbed in PD mice, such that PD H. bakeri-infected mice have elevated IFN-γ as reported by Ing et al. (Reference Ing, Su, Scott and Koski2000) and seen in our multiple cytokine arrays. Thus, based on the enhanced, generalized Th1 response seen in cytokine profiles of challenged PD mice in the present study, we suggest that elevated leptin resulting from increased consumption of a PD diet may promote a Th1 environment with elevated levels of IFN-γ that in turn suppress the Th2 responses necessary for parasite expulsion, leading to prolonged parasite survival in the PD host.
From our data, it is also possible to suggest that the leptin-induced IFN-γ rich environment interacts with epithelial cell renewal in the villus to promote villus growth and parasite survival. Villus height is a critical factor for this nematode, as the parasites preferentially move to the region of the intestine with the longest villi, they maintain their position in the intestine by twisting around the villi, and the villi also provide a food resource for the parasite (Bansemir and Sukhdeo, Reference Bansemir and Sukhdeo1996). Moreover, recent data have shown that intestinal epithelial cell turnover mediates nematode worm expulsion and is up-regulated by IL-13 and down-regulated by an IFN-γ induced chemokine (Cliffe et al. Reference Cliffe, Humphreys, Lane, Potten, Booth and Grencis2005). In our mice, villi in the proximal small intestine were longer in challenged PD mice than in challenged PS mice, and parasite survival was better in mice with minimal villus atrophy in the duodenum, namely in PD mice. It has been reported that villus growth is induced by IFN-γ which promotes epithelial cell proliferation (Dignassn, Reference Dignassn2001). We have previously shown that elevated IFN-γ is associated with prolonged nematode survival in PD mice (Ing et al. Reference Ing, Su, Scott and Koski2000), and propose that the up-regulation of IFN-γ in infected PD mice may facilitate epithelial cell proliferation, slow the rate of epithelial cell turnover and thus minimize the villus atrophy normally induced by protein deficiency in the absence of infection (Deo et al. Reference Deo, Sood and Ramalingaswami1965; Dudley, Reference Dudley, Wykes, Dudley, Fiorotto, Burrin, Rosenberger, Jahoor and Reeds1997). Thus the prolonged parasite survival in PD mice may be, in part, a result of the relatively normal villus length in these mice. In contrast, the Th2 environment of the PS infected mouse provides a microenvironment rich in IL-13 (Zhao et al. Reference Zhao, McDermott, Urban, Gause, Madden, Yeung, Morris, Finkelman and Shea-Donohue2003) which induces villus atrophy (Zou et al. Reference Zou, Coulomb, Venet, Foussat, Berrebi, Beyer, Crevon, Minty, Couedel-Courteille, Vivier, Capron, Galanaud and Emilie1998) that impairs the ability of the worms to maintain their position and to feed.
A further link with leptin is seen in our data on gut MPO, which was lower in PD mice and also showed a dose-dependent response to larval challenge on day 7 pci in the distal small intestine. Interestingly, this response was evident during the phase when larvae are embedded in the gut tissues, but not when they had emerged into the gut lumen. Gut MPO is commonly observed in intestinal inflammation in Trichinella spiralis (Khan et al. Reference Khan, Blennerhasset, Varghese, Chowdhury, Omsted, Deng and Collins2002; Venkova and Greenwood-van Meerveld, Reference Venkova and Greenwood-van Meerveld2006), a nematode that lives within the tissues of the intestine, and both MPO (Gay et al. Reference Gay, Fioramonti, Garcia-Villar and Bueno2000) and MCP-5 (Sarafi et al. Reference Sarafi, Garcia-Zepeda, MacLean, Charo and Luste1997) are elevated during infections with the lumen-dwelling phase of Nippostrongylus brasiliensis. Cakir et al. (Reference Cakir, Bozkurt, Ercan and Yegen2004) reported that exogenous leptin induced less gut damage in a model of intestinal colitis and also reversed the elevated MPO levels. This is consistent with the possibility that elevated leptin in PD mice established a Th1 environment where inflammation was reduced and indicators of intestinal inflammation, such as gut MPO, were also reduced.
Of all the indicators measured in this study, MCP-5 was the most consistent with worm expulsion data, as it responded to both diet and challenge dose on day 14 pci, and to challenge infection on day 7 pci. MCP-5 is expressed in the lung during N. brasilienses infection (Sarafi et al. Reference Sarafi, Garcia-Zepeda, MacLean, Charo and Luste1997), it is up-regulated in the small intestine by IL-4 (Ruth et al. Reference Ruth, Warmington, Shang, Lincoln, Evanoff, Kunkel and Chensue2000) and IL-13 (Cho et al. Reference Cho, Kang, Homer, Kang, Zhang, Lee, Elias and Lee2006), and it is involved in inflammation through induction of monocyte migration and activation of macrophages (Sarafi et al. Reference Sarafi, Garcia-Zepeda, MacLean, Charo and Luste1997). The role of MCP-5 in H. bakeri infection has not previously been investigated, but we suggest it may be critical through its effect on alternatively activated macrophages. It is believed that delayed larval development and stunting of parasites (Cypess et al. Reference Cypess, Lucia, Dunsford and Enriquez1988) involves alternatively activated macrophages that accumulate around the larvae (Morimoto et al. Reference Morimoto, Morimoto, Whitmire, Xiao, Anthony, Mirakami, Star, Urban and Gause2004), as in vivo depletion of these macrophages delayed worm expulsion (Anthony et al. Reference Anthony, Urban, Alem, Hamed, Rozo, Boucher, Van Rooijen and Gause2006). If MCP-5 is important in recruitment of these cells to the site of the larvae, then the reduced serum levels of MCP-5 in PD mice on day 7 pci and in PD mice infected with lower doses of H. bakeri on day 14 pci may provide a partial explanation for prolonged parasite survival in these mice. In contrast, in PS immunocompetent mice, the elevated IL-4 and IL-13 induced by H. bakeri (Zhao et al. Reference Zhao, McDermott, Urban, Gause, Madden, Yeung, Morris, Finkelman and Shea-Donohue2003) would promote MCP-5 expression leading to infiltration of alternatively activated macrophages that would damage the developing parasites. To our knowledge, our study is the first to examine MCP-5 during H. bakeri infection and also the first to report that protein deficiency suppresses serum levels of MCP-5. This would need to be confirmed by studies that associate tissue levels of MCP-5 to recruitment of alternatively activated macrophages around the larval parasite.
It has been assumed that intestinal inflammation results from infiltration of mucosal mast cells, eosinophils, Goblet cells, and Paneth cells into the villi of the infected host (Fakae et al. Reference Fakae, Harrison and Sewell2000; Shea-Donohue et al. Reference Shea-Donohue, Sullivan, Finkelman, Madden, Morris, Goldhill, Pineiro-Carrero and Urban2001). The elevated d14 pci cellular infiltration observed in our infected PS mice, compared with unchallenged PS mice, and the reduced eosinophil and mucosal mast cell populations in the distal small intestine of PD mice are consistent with this hypothesis. However, none of these cell populations differed between PS mice challenged with 100 and 200 larvae, and thus it is unlikely that these cells were directly responsible for the apparently more rapid expulsion in PS mice challenged with 200 L3.
In summary, we have previously assumed that impaired expulsion in PD mice was due to suppression of Th2 immune effectors, and that the elevation of IFN-γ was a secondary consequence (Ing et al. Reference Ing, Su, Scott and Koski2000). It is difficult to know, however, whether protein deficiency down-regulates Th2 responses with the down-stream consequence of up-regulated Th1 responses, or whether protein deficiency up-regulates Th1 responses leading, in turn, to down-regulated Th2 responses. Our current data are consistent with the latter. We have shown that infected PD mice increase their food intake which leads to elevated leptin concentrations, as reported by others (Du et al. Reference Du, Higginbotham and White2000) and observed in this study. Leptin upregulates IFN-γ (Faggioni et al. Reference Faggioni, Feingold and Grunfeld2001) and thus establishes a pro-inflammatory Th1 environment in the PD mice. Our multiple cytokine arrays and our previous work (Ing et al. Reference Ing, Su, Scott and Koski2000) both demonstrate elevated IFN- γ in PD infected mice. The reduced serum concentration of MCP-5 observed in this study may reduce infiltration of macrophages around the larval parasites, as shown in other systems (Sarafi et al. Reference Sarafi, Garcia-Zepeda, MacLean, Charo and Luste1997), thus permitting normal larval development. Additionally, the elevated levels of IFN-γ in the PD mice could account for the reduced villus atrophy in the duodenum of the PD mice, which would in turn provide a more permissive luminal environment for the adult worms. Thus, larvae would be able to develop normally in PD mice, adult worms would be able to remain twisted around the villi, intestinal inflammation would be reduced, and worm survival would be prolonged. We suggest that the roles of leptin as an initiator and MCP-5 as an effector of prolonged parasite survival in PD mice be explored further.
We thank Dr Peter Lee for guidance with the multiple cytokine arrays. This study was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grants and Fonds Québécois de la recherche sur la nature et les technologies (FQRNT). Research at the Institute of Parasitology was supported by a Regroupement stratégique from FQRNT.
