FALC B2 cells differentiate into plasma cells in infected mice

To further investigate the role of FALCs in B-cell activation, we performed whole-mount staining and confocal analysis of mediastinal FALCs from Day 11 Ls-infected mice. Staining of proliferative cells with the nuclear marker Ki67 revealed intense proliferation of IgM⁺ and B220⁺ B cells in FALCs of the infected mice (Fig. 2a confocal pictures and bar chart quantification of Ki67⁺ pixels within FALCs). IgM staining revealed that a large proportion of B cells show accumulation of intra-cellular IgM, characteristic of IgM-secreting plasma cells (Fig. 2a). This was further confirmed by the induction of the plasma cell marker CD138 in FALCs of the infected mice (Fig. 2b).

Open in a separate window

Figure 2

FALC B2 cells differentiate into plasma cells in parasite infected mice.

(a,b) Representative immunofluorescence staining of mediastinal FALCs of naive and day 11 Ls-infected C57BL/6 mice showing Ki67 (red), B220 (blue) and IgM (green in a) or CD138 (green in b) and quantification of the percentage area of Ki67 staining (a). Scale bar, 50μm. (c–e) Flow-cytometric analysis of digested pericardial FALC cells and PLEC from naive and day 11 Ls-infected C57BL/6 mice showing Ki67^high B2 cells (c, upper panel); and SSC^highIgD⁻ plasma cells (c, lower panel). Quantification of % of Ki67^high and Ki67^int (d) and plasma cells (e, left graph) in B2 cells and plasma cell number (e, right graph) in pericardial FALCs, PLEC and lymph node (LN) are shown. (f) Flow-cytometric analysis of digested pericardial FALC cells from naive and day 11 Ls-infected C57BL/6 mice showing Ki67^highGL-7⁺ germinal centre (GC) B cells and quantification of germinal centre B cells as a % of the total B2-cell population. Data in a–f are combined from two independent experiments, symbols in the graph in a represent individual clusters, data in d–f represent individual mice, n=12 per time point. ns not significant, **P<0.01, ****P<0.0001 (normally distributed data analysed by one-way ANOVA with Sidak multiple comparisons post test, non-normally distributed data analysed using the Kruksal–Wallis test with Dunn's multiple comparisons post test). Error bars represent mean with s.e.m. in all graphs.

As B2 cells produced the vast majority of Ls-specific IgM in FALCs, we focused our flow-cytometric analysis on the emergence of plasmablasts in the B2-cell subset from pericardial FALCs, PLEC and mediastinal LNs at day 11 post Ls infection. B2 plasmablasts were defined by the high expression of Ki67, accompanied by loss of membrane IgD (Fig. 2c, upper panel). Because CD138 is cleaved during collagenase digestion of FALCs, we identified plasma cells as SSC^highIgD⁻ (Fig. 2c, lower panel). FALCs were the only compartment where active proliferation and B2-plasma cell differentiation occured (Fig. 2c–e). B2 cells of the PLEC and mediastinal LNs in contrast, were not within active cell cycle (Fig. 2c,d). The large plasma cell population (30% of B2-cell subset) found in pericardial FALCs was not present in the draining LNs or the PLEC (Fig. 2c,e). Furthermore, we found a small population of GL7⁺Ki67⁺ germinal centre-like B2 cells within the pericardium at day 11 post infection (Fig. 2f). Taken together, these results provide evidence that pleural FALCs enable the formation of plasma cells-secreting antigen-specific IgM at the site of infection, outwith a classical secondary lymphoid organ.

Parasite-specific IgM within pleural FALCs is dependent upon IL-33R

IL-33, a cytokine central to the induction of type 2 responses has previously been shown to activate B1 cells to proliferate and secrete IgM in vitro and in vivo following intra-peritoneal injection of recombinant IL-33 (ref. ²⁰). However, no direct in vivo link between type 2 immunity and IL-33-dependent activation of innate B-cell responses has been demonstrated so far. We thus assessed the role of the IL-33R in the activation of FALC B cells during Ls infection, which induces a type 2 immune response. Whole-mount immunofluorescence staining showed that the induction of B-cell proliferation in mediastinal FALCs of BALB/c mice was impaired in IL-33R-deficient BALB/c mice (Il1rl1^−/−) at day 11 post infection (Fig. 3a). Immunofluorescence staining indicated that mediastinal FALCs of naive Il1rl1^−/− mice are smaller than their BALB/c counterparts, flow cytometric analysis of digested pericardial FALCs showed a trend for fewer cells in the mediastinal adipose but this did not reach significance (Fig. 3b). Next, we showed that in Il1rl1^−/− mice, none of the pericardial FALC B-cell subsets increase in number following Ls infection, unlike their wild type BALB/c counterparts (Fig. 3c). This suggested that IL-33R signalling is specifically needed for the accumulation of B cells in FALCs. Even though B2 cells failed to accumulate in pericardial FALCs of Il1rl1^−/− mice, the differentiation of B2 cells into SSC^highIgD⁻plasma cells was not completly abrogated (Fig. 3d,e), suggesting that IL-33R is not, or only partially, involved in plasma cell differentiation. However, overall the total number of plasma cells found in FALCs was severely reduced (Fig. 3e). ELISA analysis confirmed the defect in pleural FALC B-cell activation with impaired IgM accumulation in the pleural lavage of Il1rl1^−/− mice (Fig. 3f). Taken together, these results revealed that IL-33R signalling is essential for pleural B-cell antigen-specific IgM secretion during infection with the filarial nematode Ls.

Open in a separate window

Figure 3

Parasite-specific IgM produced in FALCs is dependent upon IL-33R.

(a) Representative immunofluorescence staining of mediastinal FALCs from naive and day 11 Ls-infected BALB/c and Il1rl1^−/− mice showing IgM (Green) Ki67 (Red) and B220 (blue). Scale bar, 50μm. (b,c) Analysis of digested pericardial FALCs from naive and day 11 Ls-infected BALB/c and Il1rl1^−/− mice showing total cell number and the number of B1a, B1b and B2 cells (gated as shown in Supplementary Fig. 1). (d,e) Flow cytometric analysis of digested pericardial FALCs from naive and day 11 Ls-infected BALB/c and Il1rl1^−/− mice showing frequency of SSC^highIgD⁻ B2 plasma cells (d) and quantification of frequency and number of SSC^highIgD⁻ B2 plasma cells (e). (f) ELISA of total and Ls antigen-specific IgM in the pleural lavage of naive and day 11 Ls-infected BALB/c and Il1rl1^−/− mice. Data are combined from two independent experiments; symbols represent individual mice, n≥8 per time point. ns not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (normally distributed data analysed by one-way ANOVA with Sidak multiple comparisons post test, non-normally distributed data analysed using the Kruksal–Wallis test with Dunn's multiple comparisons post test). Error bars represent mean with s.e.m. in all graphs.

IL-33 is produced by FALC stromal cells

Since B-cell activation in FALCs was so highly dependent on IL-33, we investigated whether pleural FALCs and milky spots could be a physiological source of IL-33. We compared the quantity of IL-33 released spontaneously by naive lung tissue (a known source of IL-33), the mediastinum, the omentum (peritoneal cavity adipose tissue containing milky spots) and the gonadal adipose tissue (GAT) (adipose tissue with few FALCs)¹³, using a 1h in vitro culture assay (Fig. 4a). Per gram of tissue, the mediastinum released 17 times more IL-33 than the GAT, and the omentum 7 times more IL-33 than the GAT. This suggested that the lymphoid clusters associated with adipose tissue of the visceral cavities are themselves a source of IL-33. The levels of IL-33 released by the mediastinum were comparable to the levels released by the lung. Since IL-33 expression has been previously reported in lymph node fibroblastic reticular cells²⁷, we performed whole-mount immunofluorescence staining and confocal analysis of mediastinal FALCs. The staining revealed that Gp38⁺ stromal cells from mediastinal FALCs express high levels of nuclear IL-33 (Fig. 4b). Notably, we could also detect low levels of extra-nuclear IL-33 that colocalized with membranous Gp38 suggesting a basal release of IL-33 by FALC stromal cells. The same pattern of IL-33 expression was found in stromal cells of pericardial FALCs and omental milky spots (data not shown). We next addressed whether FALC stromal cells contained IL-33 during Ls infection (Fig. 4c). Flow cytometric analysis of digested pericardial FALCs confirmed that >95% of CD45⁻GP38⁺ stromal cells from C57BL/6 mice expressed intracellular IL-33 (Fig. 4d) compared with only ~2% of CD45⁺ cells, when gated based on a secondary antibody only control (Fig. 4d). At day 11 following Ls infection, no difference in the levels of IL-33 within stromal or haematopoietic cells was detected by flow cytometry (Fig. 4d), ELISA analysis of spontaneous IL-33 release during 1h in vitro culture of the mediastinum (Fig. 4e) or whole-mount immunoflouresence staining as compared with naive controls (Fig. 4f).

Open in a separate window

Figure 4

IL-33 is produced by FALC stromal cells.

(a) Pieces of lung, mediastinum, omentum or gonadal adipose tissue (GAT) were cultured for 1h at 37°C and spontaneous release of IL-33 in the supernatant was measured by ELISA. Data are representative of four independent experiments, symbols represent individual mice, n=5. (b) Representative whole-mount immunofluorescence staining of mediastinal FALCs from BALB/c mice showing IL-33 (red), 4,6-diamidino-2-phenylindole (DAPI; blue) and Gp38 (green). Higher magnification shows expression of IL-33 in nucleus (n) and cytoplasmic/membranous area (*). Scale bar, 50μm. (c,d) Flow cytometric analysis of digested pericardial FALCs from naive and day 11 Ls-infected C57BL/6 mice showing intracellular IL-33 expression by CD45⁻Gp38⁺ stromal and CD45⁺ haematopoietic cells (c), and % of stromal cells and haematopoietic cells expressing IL-33 (d). (e) Naive and day 11 Ls-infected C57BL/6 mediastinum were cultured for 1h at 37°C and spontaneous release of IL-33 in the supernatant was measured by ELISA. (f) Representative whole-mount immunofluorescence staining of mediastinal FALCs from naive and day 11 Ls-infected C57BL/6 mice showing IL-33 (red), DAPI (blue) and Gp38 (green) and quantification of the % area of IL-33 staining. Scale bar, 50μm. Data in (c–f) are pooled from two experiments, symbols in d,e represent individual mice, n=8 or 10 per group, symbols in f represent individual clusters. ns not significant, ***P<0.001, ****P<0.0001 (normally distributed data analysed by one-way ANOVA with Sidak multiple comparisons post test, non-normally distributed data analysed using the Kruksal–Wallis test with Dunn's multiple comparisons post test). Error bars represent mean with s.e.m. in all graphs.

IL-33 controls pleural IgM secretion upon airway inflammation

To address whether our findings with Ls were more broadly relevant, we investigated the activation of pleural FALCs in a distinct model in which involvement of the pleural cavity has never previously been addressed. Inhalation of inflammatory agents results in increased expression and secretion of IL-33 by lung epithelial cells²⁸ and intra-nasal (i.n.) instillation of the fungal allergen Alt is known to increase pulmonary IL-33 expression^28,29,30,31. Surprisingly, i.n. delivery of Alt induced a noticeable release of IL-33 in the pleural lavage at 48h (Fig. 5a) revealing continuity of the inflammatory IL-33 signal between the lungs and the pleural space.

Open in a separate window

Figure 5

FALC activation during acute airway inflammation is IL-33 dependent.

(a) ELISA of IL-33 in the pleural lavage (pl Lav) of naive and Alt-treated C57BL/6 mice. Data are combined from two independent experiments, symbols represent individual mice, n=6 or 9 per group. (b) Representative whole-mount immunofluorescence staining of mediastinal FALCs from naive and Alt-treated BALB/c and Il1rl1^−/− mice showing IgM (green), Ki67 (red) and B220 (blue) and quantification of the percentage area of Ki67 staining within individual clusters, scale bar, 50μm. (c,d) Flow cytometric analysis showing proliferating Ki67^high and Ki67^int (c, left panel) and quantification of % Ki67^high (c, right panel) and number (d) of B1a (first row), B1b (second row) and B2 (third row) cells from digested pericardial FALC cells and PLEC from naive and Alt-treated BALB/c and Il1rl1^−/− mice. (e) ELISA of total IgM in the pleural lavage (pl Lav), peritoneal lavage (p Lav) and serum of naive and Alt-treated BALB/c and Il1rl1^−/− mice. Data in c–e are combined from two independent experiments, symbols represent individual mice, n=8 or 10 per time point. (f) ELISA of total IgM in the supernatant of overnight culture of PLEC and digested FALC cells from naive and Alt-treated BALB/c mice. Data are combined from three independent experiments, symbols represent individual mice, n= 9 or 14 per time point. ns not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (normally distributed data analysed by one-way ANOVA with Sidak multiple comparisons post test, non-normally distributed data analysed using the Kruksal–Wallis test with Dunn's multiple comparisons post test). Error bars represent mean with s.e.m. in all graphs.

We thus assessed whether the pleural IL-33 release induced by Alt instillation was associated with mediastinal B-cell activation. BALB/c and IL-33R-deficient (Il1rl1^−/−) mice were subjected to i.n. delivery of Alt, and B-cell activation in FALCs was assessed 48h after instillation. Whole-mount immunofluorescence microscopy showed a marked induction of B-cell proliferation in mediastinal FALCs of BALB/c mice, which was impaired in IL-33R-deficient mice (Fig. 5b). We next quantified the precise effect of Alt instillation on B-cell proliferation in collagenase digested pericardial FALC cells and PLEC by flow cytometry using the nuclear marker Ki67. B cells in pericardial FALCs but not the PLEC showed high expression levels of Ki67 characteristic of cells actively in cell cycle 48h after Alt instillation (Fig. 5c). Proliferation was more markedly increased in B1a and B1b cells (30% of Ki67^high B cells) than in B2 cells (3% of Ki67^high B2 cells) (Fig. 5c). Absence of Ki67^high B cells in PLEC demonstrated the importance of FALC to sustain pleural B-cell proliferation. Lack of IL-33R led to complete impairment in the induction of B1b and B2-cell proliferation and a fourfold reduction in the level of proliferation of FALC B1a cells in Il1rl1^−/− mice (Fig. 5c). The marked proliferation of B1a and B1b-cell subsets in FALCs following Alt delivery in BALB/c mice was correlated with a significant increase in their number in pericardial FALCs but not in the PLEC (Fig. 5d). Absence of IL-33R abrogated the accumulation of B1a and B1b cells in FALCs after Alt instillation in Il1rl1^−/− mice (Fig. 5d).

We then assessed the level of total IgM secreted within the pleural lavage, with the levels found within peritoneal lavage, a non-related site, and the serum. Alt instillation induced a significant increase in total IgM within the pleural lavage, but not in the peritoneal lavage or serum (Fig. 5e), demonstrating that the response induced is highly localized. Analysis of Il1rl1^−/− mice showed that induction of IgM secretion upon Alt instillation in the pleural space is dependent on IL-33R. To confirm that FALC B cells were the source of the IgM produced in response to Alt, we placed the same number of wild-type pleural and mediastinal FALC elicited cells in culture overnight, and determined the production of total IgM within the supernatant by ELISA. In all, 200,000 FALC cells from Alt exposed animals secreted micrograms of IgM overnight (Fig. 5f). In contrast, PLEC were unable to secrete IgM (Fig. 5f) confirming in another setting that serous B cells do not secrete antibodies³². Taken together, these results show that FALCs of the pleural cavity respond to lung perturbation by supporting localized B1-cell proliferation and IgM production within the pleural cavity. Furthermore, IL-33R has a central role in the pleural B-cell response to acute lung inflammation controlling both FALC B1-cell proliferation and IgM secretion.

FALC B cells require IL-5 for activation

Peritoneal B cells were previously shown to express low levels of IL-33R and directly respond to IL-33 in vitro²⁰, suggesting that during acute lung inflammation the IL-33 released would act directly on pleural B cells. To test this, we labelled total PLEC from BALB/c and Il1rl1^−/− donor mice with carboxyfluorescein succinimidyl ester (CFSE) and CellTrace Violet (CTV) respectively. Labelled PLEC were co-injected into the pleural space of a recipient BALB/c animal, Alt instilled 18h later and the CFSE (wild type) and CTV (Il1rl1^−/−) labelled B-cell populations compared after 48h (Fig. 6a). The injected PLEC were composed on average of 60% B cells of which 60% were B1 cells and 40% B2 cells. After transfer, between 2 and 4% of the PLEC were of donor origin. We could detect both CFSE and CTV-positive CD45⁺CD19⁺CD11b⁺ B1 and CD45⁺CD19⁺CD11b⁻ B2 cells within the PLEC and in the FALCs, indicating that IL-33R signalling was not necessary for the recruitment of pleural B cells into FALCs (Fig. 6b and not shown). Following Alt induction, the transferred B1 cells located within pericardial FALCs showed comparable dilution of CFSE and CTV, and the same increase in Ki67 expression (Fig. 6b), whereas the transferred cells found in the PLEC had not divided (not shown). This indicated that FALC resident Il1rl1^−/− B1 cells had no impairment in the induction of proliferation and that intrinsic IL-33R signalling is not necessary for B1-cell proliferation upon Alt instillation.

Open in a separate window

Figure 6

FALC B1-cell activation requires IL-5 but does not require eosinophils.

(a) Schematic of the i.pl. transfer experiment of CTV-labelled Il1rl1^−/− and CFSE-labelled BALB/c control PLEC into BALB/c recipient mice before Alt treatment. (b) Flow cytometric analysis of digested pericardial and mediastinal FALCs showing gating of CFSE and CTV-labelled B1 cells (gated as CD45⁺CD19⁺CD11b⁺). CFSE and CTV intensities (upper histograms) and Ki67 expression (lower histogram) are shown for CFSE (left) and for CTV-labelled B1 cells (right). Quantification of %Ki67 expression within CFSE⁺ BALB/c and CTV⁺ Illrl1^−/− B1-cell populations. Data combined from two independent experiments, symbols represent individual recipient mice, n=4. (c–h) Schematic of αIL-5 i.pl. blocking experiment with mice that received: intra-nasal (i.n.) and intra-pleural PBS, i.n. Alt and intra-pleural IgG control, i.n. Alt and intra-pleural anti-IL-5 blocking IgG (c). Flow cytometric analysis of digested pericardial FALCs, %Ki67^high B1a and B1b cells in pericardial FALCs (d). Representative whole-mount immunofluorescence staining of mediastinal FALCs showing IgM (green), Ki67 (red) and CD45 (blue) (e). Scale bar, 50μm. Quantification of % area of Ki67 staining (f). ELISA of total IgM secreted during 4h of in vitro culture of whole mediastina (g). Quantification of eosinophil numbers within PLEC and digested pericardium (h). Data in d–h combined from 3 (d–g) and 2 (h) independent experiments, symbols represent individual mice, n= 13 or 15 (d–g) and n=5 or 10 (h) per group. (i–k) BALB/c and ΔdblGATA mice received i.n. Alt for 48h. Flow cytometric analysis of digested pericardial FALCs showing %Ki67^high B1a and B1b cells (i). Quantification of the % area of Ki67 staining as assessed by whole-mount immunofluorescence staining of mediastinal FALCs (j). ELISA of total IgM within the pleural lavage (k). Data in i–k combined from two independent experiments, symbols represent individual mice, n=8 or 9 per group, symbols in j represent individual clusters. ns not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (normally distributed data analysed by one-way ANOVA with Sidak multiple comparisons post test, non-normally distributed data analysed using the Kruksal–Wallis test with Dunn's multiple comparisons post test). Error bars represent mean with s.e.m. in all graphs.

It has been proposed that IL-33's action on B1 cells is largely mediated by IL-5 (ref. ²⁰). We thus tested whether the action of IL-33 on FALC B1 cells was dependent on IL-5 in our model. At the time of Alt instillation, mice were injected intra-pleurally (i.pl.) with blocking anti-IL-5 antibody (Fig. 6c). Flow-cytometric analysis of FALC pericardial cells 48h later showed that blocking of IL-5 completely abrogated B1a and B1b-cell proliferation compared with mice injected with control antibody (Fig. 6d). Confocal analysis confirmed that the level of proliferation induced by Alt in mediastinal FALCs was markedly reduced following i.pl. injection of blocking anti-IL-5 antibody (Fig. 6e,f). Blocking of IL-5 also led to impaired IgM secretion by B cells within the mediastinum (Fig. 6g). Taken together, these results indicate that IL-5 is critical for the activation of FALC B1 cells during Alt induced lung inflammation. As eosinophils are the other main target of IL-5^33,34, we wanted to assess the contribution of eosinophils to the induction of B-cell proliferation and IgM secretion. First, we analysed the impact of the delivery of anti-IL-5 antibody in the pleural cavity. As expected, we found a reduction in the number of eosinophils within the PLEC and a trend for reduced eosinophilia within the pericardial FALCs. However, this did not reach significance (Fig. 6h). To completely rule out that the effect we were seeing on B cells was dependent on eosinophils, we performed Alt experiments in ΔdblGATA mice that lack eosinophils. At 48h following delivery of Alt, pericardial FALC B1a and B1b cells of ΔdblGATA mice were proliferating significantly more than their BALB/c counterparts (Fig. 6i), there was enhanced proliferation within the mediastinum as assessed by immunofluorescence staining (Fig. 6j) and there was no defect in the secretion of IgM within the pleural lavage (Fig. 6k). These data indicated that the induction of B-cell proliferation and IgM secretion was independent of eosinophils. However, since both B cells and eosinophils are dependent on IL-5, they may be in competition for its access. In the absence of eosinophils, B cells would have more IL-5 available, providing an explanation for the enhanced B-cell proliferation we found in ΔdblGATA mice.

Cell isolation and culture

Peritoneal exudate cell and PLEC were isolated by flushing the peritoneal and pleural cavities with RPMI 1640 (Sigma), respectively. Pericardium and mediastinum were enzymatically digested with 1mgml⁻¹ Collagenase D (Roche) for 35min at 37°C in RPMI 1640 containing 1% fetal bovine serum (Sigma). PLEC, lymph node cells or FALC cells, were cultured overnight and supernatant used for ELISA. Equivalent weights of lung, whole mediastinum, whole omentum and GAT were cultured for 1h before determination of IL-33 in supernatant by ELISA. For analysis of intracellular IL-5 and IL-33, pericardium and PLEC cells were used directly ex vivo or rested overnight in RPMI 1640 (Sigma) containing 10% fetal bovine serum (Sigma), 50Uml⁻¹ Penicillin (Sigma) 50μgml⁻¹ Streptomycin (Sigma), 2mM L-glutamine (Sigma), cells were then stimulated for 4h at 37°C with 50ngml⁻¹ phorbol-myristate acetate (Sigma) and 1μgml⁻¹ Ionomycin (Sigma) including 1 × Brefeldin A (eBioscience) for the final 3h of incubation.

Flow cytometry

Cells were surface stained, fixed and permeabilized as below. Cells were stained with LIVE/DEAD (Invitrogen), blocked with mouse serum and FcR-Block (clone 2.4G2, Biolegend), stained for cell surface markers (See Table 1 for list of antibodies used), and then fixed and permeabilized for intra-cellular staining with 1 × fixation-permeabilization buffer (eBioscience). Ki67 was detected using antibody clone B56 from BD Biosciences or antibody clone REA183 from Miltenyi Biotech, both of which allow the detection of Ki67^hi cells. In Fig. 7: Lineage= CD19/TCRβ/Ly6G/Ly6C/F4/80/CD11c/FcR1α/NK1.1/Ter119/CD5/CD11b/CD49b. Samples were acquired using a BD Fortessa and analysed with FlowJo software (Tree Star). Fluorescence activated cell sorting was performed using a BD FACSAria.

Table 1

List of antibodies used in the flow cytometry and immunofluorescence stainings

Reagent	Clone	Conjugate	Source	Dilution
Armenian hamster anti-CD11c	N418	PE	Biolegend	1 in 200
		APC-Cy7		1 in 200
		BV605		1 in 100
Armenian hamster anti-FcRIα	Mar_1	PE-Cy7	Biolegend	1 in 200
Armenian hamster anti-TCRb	H57–597	Pacific Blue	eBiosciences	1 in 200
Donkey anti-Goat	Polyclonal	AlexaFluor488	Invitrogen	1 in 300*
Donkey anti-mouse IgM	Polyclonal	Rhodamine Red	Jackson Laboratories	1 in 500*
Goat anti-IL33	Polyclonal	Unconjugated	R&D Systems	1 in 25*
Golden Syrian hamster anti-Podoplanin (Gp38)	eBio1.8.1	Biotin	eBiosciences	1 in 100*
Mouse anti-CD3	eBio500A2	PE	eBiosciences	1 in 200
		APC-Cy7		1 in 200
Mouse anti-Ki67	B56	FITC	BD Pharmingen	1 in 100
	REA183	FITC	Miltenyi	1 in 100
	REA183	PE		1 in 100
	SolA15	APC (IF)	eBiosciences	1 in 100*
Mouse anti-NK1.1	PK136	PEVio770	Miltenyi	1 in 200
Rat anti-I-A/I-E	M5/114.15.2	AF700	Biolegend	1 in 800
		AF780	eBiosciences	1 in 800
Rat anti-B220	RA3-6B2	eFluor450	eBiosciences	1 in 200*
Rat anti-CD11b	M1/70	eFluor450	eBiosciences	1 in 200*
		PerCP-Cy5.5		1 in 1,200
		APC		1 in 800
		BV711	Biolegend	1 in 1,000
		PE Dazle −594		1 in 1,000
Rat anti-CD138	281-2	Biotin	Biolegend	1 in 200*
Rat anti-CD19	eBio1D3	PE-Cy7	eBiosciences	1 in 200
	6D5	PE	Biolegend	1 in 200
	6D5	BV421		1 in 200
Rat anti-CD4	GK1.5	eFluor660	eBiosciences	1 in 200
	RM4–5	BV650	Biolegend	1 in 200
Rat anti-CD45.2	104	FITC	eBiosciences	1 in 100*
		BV650	Biolegend	1 in 100
		PerCP		1 in 100
Rat anti-CD49b	DX5	APC-Cy7	Biolegend	1 in 200
Rat anti-CD5	53-7.3	APC-Cy7	eBiosciences	1 in 100
		PE-Cy7		1 in 200
Rat anti-CD90.2	30-H12	Pacific Blue	Biolegend	1 in 100
		PerCP	Miltenyi	1 in 50
Rat anti-F4/80	BM8	PE	eBiosciences	1 in 200
		PE-Cy7		1 in 200
Rat anti-IgD	11–26c	eFluor450	eBiosciences	1 in 200
Rat anti-IL-5	TRFK5	PE	Biolegend	1 in 100
Rat anti-Ly-6C	HK1.4	BV570	Biolegend	1 in 50
		PE-Cy7		1 in 200
Rat anti-Ly6G	1A8	BV421	Biolegend	1 in 200
		PE	BD Pharmingen	1 in 200
Rat anti-Siglec-F	E50–2440	BV421	BD Pharmingen	1 in 200
Rat anti-ST2	DJ8	Biotin	Mdbiosciences	1 in 100
	RMST2-2	APC	eBiosciences	1 in 100
Rat anti-Ter119	TER-119	PE-Cy7	eBiosciences	1 in 200
Streptavidin	AlexaFluor555	Invitrogen	1 in 400*
	PerCP	Biolegend	1 in 200
	APC		1 in 200
	BV711		1 in 200

Open in a separate window

APC, allophycocyanin; FITC, fluorescein isothiocyanate; IF, immunofluorescence; PE, phycoerythrin.

^*Antibodies also used in immunofluorescence.

Intra-pleural injections

Total BALBc or Il1rl1^−/− PLEC were labelled with 5μgml⁻¹ CFSE (Invitrogen) or 2μM CellTrace Violet (Molecular probes), respectively, before injection of 200,000 labelled cells at a 1:1 ratio into the pleural cavity in 100μl PBS. Fifty μg of Alt was delivered i.n. 18h later. To block IL-5 within the pleural cavity, 30μg of either purified functional grade anti-human/mouse IL-5 (eBioscience, clone TRFK5), 30μg of Rat IgG1 (eBioscience, clone eBRG1) in 100μl PBS (Sigma), or PBS (Sigma) alone was injected i.pl., immediately following injection 50μl PBS or 50μg of Alternaria extract (Greer) in 50μl PBS (Sigma) was delivered i.n.

Antibody and cytokine ELISAs

IgM antibodies were detected using goat anti-mouse IgG/A/M (10100) from AbD Serotec diluted at 1 in 1000, followed by secondary goat anti-mouse IgM horseradish peroxidase (102005) from Southern Biotech diluted at 1 in 2000. Purified mouse IgM isotype control (553472) from BD Pharmingen was used as standard. For antigen-specific antibody detection, ELISA plates (Nunc) were coated with Ls Ag or Alternaria (Greer) at a concentration of 5μgml⁻¹ in carbonate buffer. IL-33 was detected using the IL-33 duoset ELISA (DY3626) from R&D systems following the manufacturers instructions. Plates were developed using two parts TMB reagent (KPL & Biolegend).

Whole-mount immunofluorescence staining and confocal images

Mediastinal adipose tissues were fixed in 10% neutral buffered formalin (NBF) for 1h on ice and permeabilized in PBS 1% Triton for 30min on ice prior staining with primary antibodies for 2h at room temperature in PBS 0.5% BSA 0.5% Triton. After wash in PBS, tissues were stained with secondary antibodies for one hour at room temperature in PBS 0.5% BSA 0.5% Triton. Antibodies used are listed in Table 1. Confocal images were acquired using a Leica SP5 laser scanning confocal microscope. Image analysis was performed using ImageJ.

Statistical analysis

Power calculations showed that for our most commonly measured parameters (lymphoid cluster number and cell number) six mice per group provide sufficient power (90%) to detect at least a twofold difference between the groups, which we regard as an acceptable cutoff for identifying important biological effects. No randomization and no blinding was used for the animal experiments. Whenever possible, the investigator was partially blinded for assessing the outcome (cluster counts). All data were analysed using Prism 6 (Graphpad Prism, La Jolla, CA, USA).

We would like to thank E. Mohr, J. Caamano and members of the IIIR for their advice on the work, A. Fulton for maintaining the parasite life cycle, M. Waterfall for flow cytometry assistance, A. McKenzie and C. Lloyd for kind provision of mouse lines and the University of Edinburgh animal house staff and veterinarians for their husbandry. This work was supported by MRC UK grants to J.E.A. (MR/K01207X/1) and C.B. (MR/M011542/1).