BALB/c and MIF-deficient mice on the BALB/c background (18) were bred in-house and housed in individually-ventilated cages (IVCs) according to UK Home Office guidelines. Mice on the C57BL/6 background expressing Cre recombinase under the LysM promoter (19), and carrying flanking loxP (flox) sites either side of the Stat3 locus, were bred as described previously (20).

Infections employed 200 L3 larvae of H. polygyrus in 200 μl water by oral gavage. Parasite lifecycles and collection of excretory-secretory (HES) products were conducted as previously described in CBAxC57BL/6 F1 mice (21). Granuloma and adult worm counts were conducted after small intestines were removed and sliced longitudinally. Three to four fecal pellets were weighed and dissolved in 2 ml dH₂0; 2 ml of saturated salt solution (400 g NaCl in 1 L dH₂0) was then added and eggs enumerated using a McMaster egg counting chamber. Egg counts are represented as eggs/g fecal material. For secondary infections, mice were cleared of worms after infection with H. polygyrus with pyrantel embonate Strongid-P paste (Elanco Animal Health), given in 2 doses of 2.5 mg dissolved in 200 μl dH₂0 given on days 28 and 29 by oral gavage. After 2 weeks, mice were re-infected with 200 L3 larvae.

One mg of MIF inhibitor, 4-IPP (Tocris Bioscience #3249) (22) dissolved in DMSO, or DMSO alone, was administered intraperitoneally (i.p.) in 50 μl every other day, during H. polygyrus infection [adapted from (17)]. Fifty ng of recombinant MIF (R&D) in 50 μl PBS, or PBS alone, was administered i.p. every other day, during H. polygyrus infection. rIL-33 (R&D) was administered intranasally (200 ng in 50 μl PBS) to sedated mice on days 0, 1 and 2, and lung tissue taken at day 3 for analysis. Alternaria alternata antigen (Greer) was administered intranasally (10 μg in 50 μl PBS) to sedated mice. BALF was harvested 1 h later (adapted from (23). For vaccination, mice were immunized with 5 μg of HES i.p. in alum adjuvant, and boosted on days 28 and 35 with 1 μg in alum before challenge with H. polygyrus at day 42 (24).

Mesenteric lymph node (MLN) cell suspensions were prepared directly by passage through 70 μm nylon filters (BD) and placed in RPMI1640 (Gibco) containing 10% FCS, 1% PenStrep (Gibco) and 1% L-glutamine (Gibco) (complete RPMI). Cells were restimulated for 72 h at 37°C with either media alone or HES at a final concentration of 1 μg/ml with 1 × 10⁶ cells, in triplicate. Peritoneal exudate cells were collected by washing the peritoneal cavity with 2 × 5 ml RPMI1640 using a 23 gauge needle. Red blood cells were removed by adding 3 ml red blood cell (RBC) lysis buffer (Sigma) for 4 min, and washing with complete RPMI. Peritoneal lavage used for ELISA analysis consisted of the supernatant from the first 5 ml wash following centrifugation to pellet cells. Bronchoalveloar lavage was collected by washing the lungs with 1 ml ice-cold PBS. Lung tissue was digested in HBSS (Gibco) supplemented with 4 U/ml Liberase TL (Roche) and 160 U/ml DNAse 1 (Sigma). Tissue was incubated at 37°C for 25 min, passed through 70 μm nylon filters (BD) and RBC-lysed before cells were used for flow cytometric analysis.

Cells were stained in 96-well round-bottomed plates. Prior to antibody staining, cells were washed in PBS and stained with LIVE/DEAD Fixable Blue (Invitrogen) at a 1/1000 dilution in 100 μl PBS for 20 min at 4°C. Then, Fc receptors were blocked in 50 μl of FACS buffer containing 100 μg/ml of naïve rat IgG (Sigma) for 20 min at 4°C. Samples were then surface stained for 20 min in 20 μl of FACS buffer containing a combination of the antibodies detailed below. Lineage markers for ILC2 negative gating: CD3 (Biolegend 17A2), CD4 (Biolegend RM4-5), CD8α (Biolegend 53-6.7), CD19 (Biolegend 6D5) CD49b (eBioscience DX5), Gr1 (Biolegend RB6-8C5), CD11c (Biolegend N418); F4/80 (Biolegend BM8), CD11b (Biolegend M1/70), SiglecF (BD E50-2440); ICOS staining for ILC2 used BioLegend C398.4A. To measure intracellular IL-5, cells were first stimulated for 4 h at 37°C in the presence of PMA (50 ng/ml), Ionomycin (1 μg/ml), and Brefeldin A (10 μg/ml) (all from Sigma). For pSTAT6 staining, clone pY641 conjugated to AF647 (BD Cat No 558242) was used. Following surface staining, cells were permeabilised for 30 min at 4°C in Cytofix/Cytoperm solution (BD), and then washed twice in 200 μl of Perm/Wash (BD). ILC2s were stained for intracellular cytokine expression in Perm/Wash (BD) using anti-IL-5 (eBioscience TRFK5). For Foxp3 (eBioscience, FJK-16s), Arginase-1 (R&D Systems IC5868P), RELM-α [R&D Systems 226033, labeled with AF647 (Invitrogen)] and Chil3 [R&D biotinylated goat anti-mouse combined with Streptavidin PeCy7 (Biolegend)], samples were stained for surface markers after which cells were permeabilised for 12 h at 4°C in Fix/Perm solution (eBioscience Foxp3 staining set), and then washed twice in 200 μl of Perm/Wash (eBioscience Foxp3 staining set). After staining, cells were washed twice in 200 μl of FACS buffer before acquisition on the LSR II or Canto flow cytometers (BD Bioscience) and subsequently analyzed using FlowJo (Tree Star).

Cytokine levels were detected in culture supernatants and BALF by ELISA using monoclonal capture and biotinylated detection antibody pairs as follows, used at concentrations optimized previously: IL-4 [11B11 + BVD6-24G2 (BD Pharmingen)]; IL-13 [eBio13A + eBio1316H (eBioscience)]; IL-33 (R&D Duoset). p-nitrophenyl phosphate (pNPP, 1 mg/ml, Sigma) was used as a substrate. OD was measured at 405 nm on a Precision microplate reader (Molecular Devices) and data analyzed using Softmax Pro software.

Serum antibodies to HES were measured by ELISA as previously described (7). Briefly, plates were coated with 1 μg/ml HES in carbonate buffer, blocked with 10% BSA in carbonate buffer, and incubated with serial dilutions of sera. Antibody binding was detected using HRP-conjugated goat anti-mouse IgA or IgG1 (Southern Biotech 1070–50 and 1040–50) and ABTS Peroxidase Substrate (KPL), and read at 405 nm.

Approximately 1 cm small intestine was homogenized in 500 μl 1x lysis buffer (Cell Signaling Technology Inc) plus 5 μl phenylmethanesulfonyl fluoride solution (PMSF) (Sigma) using a TissueLyser (Qiagen). Samples were centrifuged at 12,000 rpm for 10 min to remove debris and supernatants added to ELISAs, at a 1:10 dilution, to measure RELM-α (Peprotech) and Chil3 (R&D). Levels were normalized to total protein content measured using a Bradford assay. The same ELISA sets were used to analyse peritoneal lavage levels of RELM-α and Chil3.

Transverse sections were made from 2 cm of paraffin-embedded small intestine, at a thickness of 4 μm using a cryostat. For MIF staining, sections were deparaffinized by immersing slides in Histoclear (Brunel Microscopes Ltd) for 5 min, and then hydrated through 100, 95, and 70% ethanol successively. Antigen retrieval was undertaken with citrate buffer (20 mM citric acid + 0.05% Tween 20 at pH6) warmed to 95°C for 20 min. Sections were blocked in 1x PBS with 1% BSA, 2% normal rabbit serum, 0.1% Triton X-100 and 0.05% Tween 20 for 30 min at room temperature and then incubated with rabbit α-MIF (Invitrogen) at 1:2000 dilution in block buffer, and left overnight at 4°C. Slides were immersed in 3% H₂O₂ for 10 min at room temperature, and washed in PBS. Goat α-rabbit conjugated to biotin (Vector Laboratories) at 5 μg/ml in PBS was added for 1 h at room temperature, in the dark. Following 2 washes in PBS, several drops of ABC Vectastain (Vector Laboratories) were added and slides left for 30 min at room temperature, in the dark. Slides were washed twice in PBS and DAB peroxidase solution (Vector Laboratories) was added for 5 min (until a brown stain had developed). With water washes in between, the following were added successively to counterstain the sections: Harris hemotoxylin solution (Sigma), acid alcohol (75% ethanol + 1% HCl) and Scott's Tap Water Substitute (ddH2O + 42 mM NaHCO3 and 167 mM MgSO4). Slides were dehydrated through 75, 95, and 100% ethanol and then Histoclear added for 5 min. Coverslips were added with DPX mountant (Sigma) and slides were left to dry overnight, in the dark. Pictures were taken using a Leica DFC290 compound microscope and Leica Application Suite software.

For fluorescent staining of macrophages, proximal small intestinal tissue was harvested and longitudinally opened. Any food matter was then gently removed by scraping and the tissue rolled onto a toothpick. Tissue was immediately immersed in OCT compound (Tissue-Tek), frozen on dry ice and stored at −80°C. Then, 13 μm thick sections were cut using a cryotome (Thermo Fisher), attached to positively charged microscope slides (VWR), dried for 15 min at room temperature and then stored at −80°C. For the staining procedure, tissue sections were thawed at room temperature and dried for 15 min under airflow, followed by fixation with 4% paraformaldehyde in PBS for 15 min at room temperature. The sections were then rinsed twice in PBS, permeabilised in PBS/0.1% saponin (Sigma-Aldrich) and consecutively blocked using a solution of 10% donkey serum (Abcam) and 0.3 M glycine (Fisher Scientific) in PBS for 60 min at room temperature. After 2 washes in PBS/0.1% saponin, primary staining was performed overnight at 4°C in PBS/1%BSA/0.1% saponin containing an antibody cocktail of: rat anti-mouse CD68-FITC (FA-11, Biolegend used at 5 ug/mL), rat anti-mouse EpCam-PE (G8.8, Biolegend, 2.5 ug/mL) and sheep anti-human/mouse arginase 1 (R&D, 5 ug/mL). Stained tissue sections were then washed 3 times in PBS/0.1% saponin and secondary antibody staining was performed with donkey anti-sheep AF647 (Abcam, 1/500) for 40 min at room temperature. This was followed by 3 washes in PBS/0.1% saponin and 2 PBS washes. DAPI containing Vectashield mounting media (Vector) and coverslips were applied prior to imaging using an EVOS FL Auto 2 fluorescence microscope (Invitrogen). Mean fluorescence intensity (MFI) of Arginase-1 staining in granulomas was analyzed using the image analysis software Fiji (SciJava).

To isolate mRNA from MLN and duodenal tissue, samples were first immersed in 1 ml of Trizol (Invitrogen) and disrupted using a TissueLyser (Qiagen) for 2 min at 25 Hz and then stored at −80°C until mRNA isolation was performed with the Qiagen mRNA easy kit (Qiagen) according to manufacturer's instructions. For duodenal analysis, ~0.5 cm of the uppermost part of the duodenum was sampled. Briefly, tissue was first disrupted using a TissueLyser (Qiagen), then 200 μl chloroform was added and samples were centrifuged at 12,000 g for 15 min at 4°C. The upper aqueous layer was recovered and added to 500 μl of isopropanol, mixed, and stood at room temperature for 10 min. The sample was then centrifuged again at 12,000 g for 10 min at 4°C. Pelleted RNA was washed once in 70% ethanol, and allowed to air dry before being dissolved in 50 μl of DEPC-treated water; 15 μl RNA was treated with DNAse (DNAFree kit, Ambion), concentrations were determined using a Nanodrop 1000 (Thermo Scientific), and samples reverse-transcribed using 1–2 μg of RNA with M-MLV reverse transcriptase (Promega). A PCR block (Peltier Thermal Cycler, MJ Research) was used for the transcription reaction at 37°C for 60 min. Gene transcript levels were measured by real-time PCR on a Roche Lightcycler 480 II, in 10 μl total volume made up of 4 μl cDNA, 5 μl SYBR Green (Roche), 0.3 μl of each primer (10 μM), and 0.4 μl DEPC treated water (Ambion) using standard conditions for 60 cycles. Target gene expression levels were normalized against the housekeeping gene GAPDH.

Primer sequences were as follows:

Two independent array experiments were conducted; in one, wild type BALB/c or MIF-deficient mice were infected with H. polygyrus for 5 days with tissues from the duodenum and MLN being collected and stored in RNA Later (Ambion) prior to processing. Duodenal tissues from uninfected BALB/c and MIF-deficient mice were also taken and stored (4 mice per group for each condition). MLNs from uninfected mice were too small to include in this experiment. In the second experiment, duodenal tissues were taken from naive mice (day 0) as well as days 3 and 7 of H. polygyrus infection for both BALB/c and MIF-deficient mice (4 mice per group for each condition).

Total RNA from tissues was extracted by firstly placing tissue in RLT buffer (Invitrogen) and then homogenized using a Tissue Lyser II (Qiagen) set for 2 min at 25 Hz. RNA isolation was performed with the RNeasy mini kit (Qiagen) according to manufacturer's instructions. RNA amplification and biotinylation prior to array hybridization was performed using the Illumina TotalPrep RNA Amplification kit (Ambion) according to manufacturer's instructions. All samples were checked for RNA quality prior to hybridization by Agilent 2100 Bioanalyzer (Agilent).

Data were generated at the Wellcome Trust Clinical Research Facility (WTCRF) located at Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU. A total of 48 Illumina MouseWG6_V2_0_R3_11278593_A arrays were QC analyzed using the arrayQualityMetrics Bioconductor package to identify sub-standard and/or outlier arrays. Three arrays were identified as outliers and were removed from subsequent analyses.

All statistical analyses were performed using Prism (Graphpad Software Inc.). Error bars on graphs display mean and standard error the mean (SEM). Student's t-test was used to compare groups. n.s., not significant, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. Results are combined from several similar experiments unless otherwise stated in the figure legend.

A widely used model system for helminth infection is that of H. polygyrus in which parasitic larvae invade the intestinal tract and mature to luminal-dwelling adult worms releasing eggs into the environment (25, 26). BALB/c mice are initially susceptible to infection but are able to gradually reduce their worm load over several weeks through a macrophage-dependent mechanism (7), and are almost fully clear of adult worms (Figure 1A) and fecal eggs (Figure 1B) by day 28 post-infection. We tested MIF-deficient BALB/c mice and found that they were unable to reduce adult worm burdens or egg output following a primary H. polygyrus infection.

We then tested resistance of MIF^−/− mice to parasite infection in two models of acquired immunity. In the first, immunity to infection can be accelerated by a prior episode of abbreviated infection, terminated by anthelmintic therapy (27); in this case, alternatively-activated (M2) macrophages have been shown to be essential for protection (28). We found that MIF-deficient mice are unable to expel adult worms, which are mostly cleared by day 21 in the wild-type mice (Figure 1C). Secondly, we used a vaccine model in which sterilizing immunity is elicited by immunization with H. polygyrus excretory-secretory (ES) products in alum adjuvant (24). In this setting, BALB/c mice show complete protection but MIF-deficient animals fail to clear the parasites (Figure 1D).

We also reproduced the phenotype using a pharmacological inhibitor of MIF, 4-iodo-6-phenylpyrimidine (4-IPP), which acts as a “suicide substrate” by covalently binding the N-terminal proline required for catalytic activity (17, 22). Mice receiving this inhibitor showed significantly greater susceptibility than vehicle-treated mice to H. polygyrus, in terms of both adult worm burdens and egg output (Figures 1E,F).

Although immunity generally correlates with the formation of intestinal granulomas (29, 30), we found that MIF^−/− mice developed normal numbers of granulomas despite being completely susceptible to infection (Figure 1G).

Immunity to H. polygyrus following either drug-mediated clearance, or HES vaccination, has been shown to be antibody-dependent, in particular requiring IgG1 (24, 31). We therefore compared serological responses to infection in wild-type and MIF^−/− mice but found no difference in serum IgG1 titer (Figure 2A). In addition, both genotypes responded with equally high anti-HES antibody titers following vaccination (Figure 2B), although only the MIF^−/− animals failed to expel parasites. These findings implicated a deficiency in the cellular arm of the response in the absence of MIF.

We then compared parasite-specific T cell responses, as immunity to H. polygyrus is strongly Th2 dependent (32), by challenging small intestine draining mesenteric lymph node (MLN) cells from H. polygyrus-infected BALB/c and MIF^−/− mice with HES antigens. We found comparably robust IL-4 and IL-13 responses in both strains (Figures 2C,D); no induction of antigen-specific IFNγ responses above background was seen in either strain (data not shown), indicating the susceptibility of the MIF-deficient mouse cannot be explained by a switch to the Th1 mode of immunity.

Regulatory T cells (Treg) expressing the Foxp3⁺ transcription factor are known to expand during H. polygyrus infection (33, 34) and render mice susceptible (35, 36). MLN cell populations were analyzed by flow cytometry at 14 and 28 days post-infection, and similar increases in Foxp3⁺ Treg numbers were seen in both wild-type and MIF^−/− mice infected with H. polygyrus (Figure 2E). Increases in Treg frequency (as percentage of total CD4⁺ cells) and Foxp3 intensity were also similar between the two strains (data not shown), indicating that increased Treg activity is not contributing to greater susceptibility in the gene-targeted mice.

We then analyzed innate immune cell responses in BALB/c and MIF^−/− mice at day 7 post-H. polygyrus infection. In the wild-type animals, infection provokes a sharp increase in cell numbers within the MLNs (Figure 3A), which is diminished in the MIF^−/− mice. Infection also results in activation of ILC2s to express IL-5 which is almost totally ablated in the MIF-deficient animals (Figure 3B), within an overall reduction of ILCS in the MLN (Figure 3C). Likewise, cellular expansion in the peritoneal cavity provoked by infection is profoundly reduced in MIF-deficient mice (Figure 3D), as is eosinophilia (Figure 3E). The loss of eosinophils in the absence of MIF has previously been observed in both helminth infection and airway asthma models (13, 37, 38). We also tested the MIF-dependence of eosinophilia by administering the 4-IPP MIF inhibitor at the time of H. polygyrus infection. Wild-type mice receiving this inhibitor showed significantly reduced peritoneal eosinophilia compared to animals treated with the DMSO vehicle alone (Figure 3F).

To ascertain whether ILC2 differentiation was intrinsically compromised in MIF-deficient mice, we first tested the effects of exogenous IL-33 injection on the activation of ILCs in the lung; IL-33 drove equivalent IL-5⁺ ILC2 responses irrespective of MIF genotype (Figure 3G). We then tested the response of mice to airway challenge with Alternaria allergen, a potent stimulator of the ILC2 population through provoking rapid release of IL-33 from the airway epithelium (39). The introduction of exogenous Alternaria antigen elicited equivalent levels of IL-33 into the bronchoalveolar lavage after 1 h in BALB/c and MIF-deficient animals (Figure 3H), arguing that both the release of ILC2-stimulating alarmins and the development of ILC2 responses to these cytokines are intact in the MIF-deficient setting.

The diminished eosinophil responses in MIF-deficient mice cannot readily account for their increased susceptibility, as eosinophil-deficient ΔdblGATA mice retain their ability to expel H. polygyrus following vaccination (24). In addition, the poor ILC2 response observed did not translate into any shortfall in the Th2 response that develops to parasite antigens (Figures 2C,D), although it is possible that abated ILC2 production of IL-5 explains the deficient eosinophil responses in MIF^−/− mice.

We next analyzed myeloid subpopulations, which play critical roles in mediating immunity in many helminth settings (7, 28, 40). To establish whether the absence of MIF resulted in significant differences within the myeloid compartment, we compared the phenotype of CD11b⁺ F4/80⁺ macrophages in MIF-sufficient and -deficient mice. We found that, following H. polygyrus infection, few viable lamina propria cells could be recovered from either BALB/c or MIF-deficient mice and hence populations were assayed from the peritoneal cavity, in which there is extensive expansion and alternative activation of macrophages during the first week of infection (41). Notably, the increase in macrophage numbers was muted in the peritoneal cavity of MIF-deficient mice (27% above naïve levels) compared to wild-type animals (90% increased) (Figure 4A).

Because MIF has previously been shown to promote the alternative activation of macrophages, alongside IL-4Rα-binding cytokines (16), we measured expression of key alternatively-activated macrophage (AAM)-associated gene products Arginase-1, Chil3 (Ym1) and RELMα by a combination of flow cytometry of peritoneal cell populations, and ELISA for soluble proteins in peritoneal lavage fluids. By each of these measures MIF^−/− mice showed significant impairment of alternative activation. Thus, the proportions of peritoneal macrophages staining for Arginase-1 (Figures 4B,C) and RELMα (Figures 4D,E) were significantly reduced in MIF-deficient animals, as were levels of detectable Chil3 and RELMα protein in the lavage following H. polygyrus infection (see below). We also compared the level of IL-4-stimulated STAT6 phosphorylation in bone marrow-derived macrophages from MIF-sufficient and deficient mice, and found no significant differences (Figure 4F), indicating that the requirement for MIF acts downstream, or independently of, IL-4 receptor ligation.

We next examined the in vivo effects of pharmacological MIF inhibition on the expression of AAM markers; administration of 4-IPP significantly reduced the number of CD11b⁺ Arginase-1⁺ (Figure 4G) and Chil3⁺ (Figure 4H) peritoneal macrophages after H. polygyrus infection, as well as the levels of both Chil3 and RELMα protein in the peritoneal lavage fluid of BALB/c mice (Figure 4I).

To test whether MIF is directly responsible for the alternative activation of macrophages, we evaluated the effects of administering recombinant MIF into the peritoneal cavity of MIF-deficient mice. Such treatment restored the proportions of Chil3⁺ AAM in this site after H. polygyrus infection to levels comparable with wild-type mice (Figure 5A), but did not rescue the significant deficit in ILC2 cells in the same location (Figure 5B). Exogenous MIF was able to partially restore protein levels of Chil3 and RELM-α in the peritoneal lavage fluid of MIF^−/− mice (Figures 5C,D), although remaining significantly below those of the wild-type mice, and no eosinophilia was elicited (data not shown). Furthermore, these products were also upregulated in small intestinal tissues of H. polygyrus-infected MIF-deficient mice (Figures 5E,F). As intraperitoneal delivery of MIF did not restore resistance to the parasite infection (data not shown), it is likely that localized production and release within the intestinal tract may be required for effective recruitment and activation of tissue macrophages at the site of infection.

While peritoneal macrophages may mirror the phenotype of the intestinal population, it is important to also study those cells closely associated with larval parasites in the submucosa of the small intestine, where H. polygyrus is found for the first 8 days of infection. We used immunofluorescence imaging to characterize the patterns of macrophage activation and accumulation around larval parasites, and their expression of Arginase-1 which is known to be required for immunity to this helminth (28). Surprisingly, local macrophage infiltration and overall Arginase-1 expression did not significantly differ in infected MIF^−/− mice (Figure 6A), and although the intensity of Arginase-1 staining in gene-deficient tissues was marginally weaker at day 4 of infection (Figure 6B), by day 6 it was as ubiquitous as in the wild-type controls (Figure 6C). In both examples, Arginase-1 is disseminated throughout the granuloma, indicating that it is either or both expressed by cells other than macrophages, and/or released extensively into the extracellular milieu from those cells which express it. Immunohistochemical staining was also used to identify widespread expression of MIF in intestinal tissues in vivo; in particular, MIF was intensely expressed within the granulomas centered around immobile larvae (Figure 6D), at the foci of the local immune response to intestinal helminth infection.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.