Enhanced protection from fibrosis and inflammation in the combined absence of IL-13 and IFN-γ

Abstract

Persistent or dysregulated IL-13 responses are key drivers of fibrosis in multiple organ systems, and this identifies this cytokine as an important therapeutic target. Nevertheless, the mechanisms by which IL-13 blockade leads to the amelioration of fibrosis remain unclear. Because IFN-γ exhibits potent anti-fibrotic activity, and IL-4Rα signalling antagonizes IFN-γ effector function, compensatory increases in IFN-γ activity following IL-13/IL-4Rα blockade might contribute to the reduction in fibrosis. To investigate the role of IFN-γ, we developed novel IL-13^−/−/IFN-γ^−/− double cytokine deficient mice and examined disease progression in models of type 2 driven fibrosis. As predicted, we showed that fibrosis in the lung and liver are both highly dependent on IL-13. We also observed increased IFN-γ production and inflammatory activity in the tissues of IL-13-deficient mice. Surprisingly however, an even greater reduction in fibrosis was observed in IL-13/IFN-γ double deficient mice, most notably in the livers of mice chronically infected with Schistosoma mansoni. The increased protection was associated with marked decreases in Tgfb1, Mmp12, and Timp1 mRNA expression in the tissues, reduced inflammation, and decreased expression of important pro-inflammatory mediators such as TNF-α. Experiments conducted with neutralizing monoclonal antibodies to IL-13 and IFN-γ validated the findings with the genetically deficient mice. Together, these studies demonstrate that the reduction in fibrosis observed when IL-13 signalling is suppressed is not dependent on increased IFN-γ activity. Instead, by reducing compensatory increases in type-1-associated inflammation, therapeutic strategies that block IFN-γ and IL-13 activity simultaneously can confer greater protection from progressive fibrosis than IL-13 blockade alone.

Introduction

Type-2 cytokine responses, defined by the cytokines interleukin-4 (IL-4), IL-5, IL-9, and IL-13, and type-1 responses characterized by IL-12 and IFN-γ[1], play distinct roles in immunity, inflammation, and fibrosis and both responses undergo significant cross-regulation or suppression by the opposing pathway [2,3]. In addition to suppressing type-1 driven inflammatory responses, type-2 cytokine responses are intimately involved in the activation of allergic inflammation, anti-helminth immunity, and tissue repair [4,5]. In particular, the type-2 cytokine IL-13, produced by a variety of cell types including CD4⁺ Th2 cells, type 2 innate lymphoid cells, eosinophils, mast cells, basophils, and NK T cells, has emerged as a key cytokine in numerous type 2 driven diseases [1,6,7].

In addition to its well-defined roles in allergic disease and immunity to gastrointestinal parasites [8,9], chronic IL-13 responses are also directly involved in the development of pathological fibrosis [10,11]. Indeed, dysregulated IL-13 responses have been shown to contribute to the development of fibrosis in multiple organ systems [12,13]. Consequently, several clinical trials are currently underway to investigate the therapeutic potential of IL-13 blockade in idiopathic pulmonary fibrosis and moderate-to-severe asthma [14,15]. Nevertheless, the mechanisms by which IL-13 blockade leads to the amelioration of fibrosis and tissue remodelling remains unclear. Some studies have suggested that IL-13 directly activates the pro-fibrotic functions of macrophages and myofibroblasts [16–18]. Alternatively, because IFN-γ exhibits potent anti-fibrotic activity [19], and type 2 cytokines antagonize IFN-γ effector function [2], compensatory increases in IFN-γ expression or activity following IL-13 blockade might also contribute to the reduction in fibrosis.

To investigate the potential protective role of IFN-γ in the development of IL-13-dependent fibrosis, we developed novel IL-13^−/−/IFN-γ^−/− double cytokine deficient mice and examined disease progression in two pulmonary and hepatic models of type 2 driven fibrosis. In one set of studies, C57BL/6 wild type, IFN-γ^−/−, IL-13^−/−, and IL-13^−/−/IFN-γ^−/− mice were examined in a well-established model of type 2 driven granulomatous inflammation in which live Schistosoma mansoni eggs are injected intravenously into i.p. egg-sensitized mice [20]. In this model, granuloma formation and fibrosis peak in the lung on day 7, with IL-4 and IL-13 playing collaborative roles in granulomatous inflammation and IL-13 playing the dominant role in the development of fibrosis [20,21]. In a second series of experiments, mice were infected with 35 infectious S. mansoni cercariae and type-2 cytokine dependent granuloma formation and fibrosis were evaluated in the liver at acute (wk 8), early chronic (wk 12), and late chronic (wk 20) time points following infection [16].

As predicted, we showed that fibrosis in the lung and liver are both highly dependent on IL-13 [22]. We also observed increased IFN-γ production and activity in the tissues of IL-13-deficient mice, suggesting that the protective anti-fibrotic effects of IL-13 deficiency might indeed be in part due to the increased IFN-γ response. Surprisingly however, we observed an even greater reduction in fibrosis in the IL-13/IFN-γ double deficient mice, particularly in the livers of chronically infected mice. The increased protection observed in IL-13^−/−/IFN-γ^−/− mice was associated with marked decreases in Tgfb1, Mmp12, and Timp1 mRNA in the tissues, reduced inflammation, and suppression of other pro-inflammatory mediators like TNF-α. Moreover, studies conducted with neutralizing mAbs to IFN-γ and IL-13 confirmed these findings. Together, these studies demonstrate that the protective anti-fibrotic activity observed in IL-13 deficient mice is completely independent from increased IFN-γ activity. Instead, our findings suggest that by reducing compensatory increases in type-1-associated inflammation, simultaneous reductions in IL-13 and IFN-γ signalling might confer greater protection from pathological fibrosis than IL-13 blockade alone.

Methods

Mice and Parasites

C57BL/6 mice were obtained from Taconic Farms (Germantown, NY). IL-13KO mice were backcrossed into B6 background for 13 generations and intercrossed with IFN-γ KO mice to generate IL-13/IFN-γ double KO mice. All mice were housed under specific pathogen-free conditions at the National Institutes of Health in an American Association for the Accreditation of Laboratory Animal Care approved facility. All experiments were performed on approved animal study protocol LPD 16E. S. mansoni eggs were extracted from the livers of infected mice (Biomedical Research Institute, Rockville, MD). Mice were infected percutaneously via the tail with 25–35 cercariae of a Puerto-Rican Strain of S. mansoni (NMRI) that were obtained from infected Biomphalria glabrata snails (Biomedical Research Institute). All animals underwent perfusion at the time of sacrifice so that worm and tissue egg burdens could be determined.

Infection with S. mansoni

Mice were infected by exposing their tails to water containing 35 cercariae of S. mansoni of Puerto Rican strain. 9, 12 or 20 wk later, infected mice were euthanized with high dose pentobarbital and their portal circulation perfused to enumerate worm numbers. In some studies, infected C5BL/6 mice (Taconic) were treated twice weekly with anti-IL-13 (200 microgram/dose, gift from Centocor), anti-IFN-γ (400 microgram/dose) (clone XMG1.2, BioXCell) or with both antibodies. Additional control mice were treated with isotype matched control antibodies at the same dose. All antibodies were administered via the intraperitoneal route in 0.5 ml saline.

Pulmonary Granuloma model

For the induction of secondary lung granulomas, mice were sensitized intraperitoneally (i.p.) with 5,000 S. mansoni eggs, and then challenged 14 days later with 5,000 live S. mansoni eggs i.v. Lungs were harvested 7 d post challenge to quantify immune and histopathological parameters.

Histopathology and Fibrosis

The sizes of pulmonary and hepatic granulomas were determined on histological sections stained with Wright’s Giemsa stain (Histopath of America, Clinton, MD). Around 30 granulomas per mouse were included in all analyses. The percentages of eosinophils, mast cells and other types of cells were evaluated in the same sections. The number of schistosome eggs in the liver and the collagen content of the liver, as measured by a hydroxyproline assay, were determined.

RNA isolation and purification and reverse transcription (RT)-qPCR

Total RNA was extracted from lung and liver tissue samples placed individually in 1 mL TRIzol reagent (Invitrogen, Carlsbad, CA). The sample was homogenized using a tissue polytron (Omni International Inc.) and total RNA was extracted according to the manufacturer’s recommendations and further purified using RNeasy Mini Kit from Qiagen (Qiagen Sciences, Germantown, MD). Individual sample RNA (1 µg) was reverse transcribed using Superscript II (Invitrogen, Carlsbad, CA) and a mixture of oligo (dT) and random primers. Real- time polymerase chain reaction amplification was performed on an ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA). Relative abundance of mRNA for several genes was determined using SYBR Green PCR Master Mix (Applied Biosystems) and by the comparative threshold cycle method as described by Applied Biosystems for the ABI Prism 7700/7900 sequence detection systems. In this method, mRNA levels for each sample were normalized to hypoxanthine guanine phosphoribosyl transferase (Hgprt) mRNA levels and then expressed as a relative increase or decrease (‘fold change’) compared with levels in naive controls. Primers were designed using Primer Express software (Applied Biosystems).

Ex vivo intracellular cytokine staining

WBCs were purified from lungs of infected mice by separation on a Percoll gradient. The separated WBCs were cultured in the presence of PMA, ionomycin and BFA for 3 hours. Cells were washed surface stained with anti-CD4, fixed with 2% paraformaldehyde, permeabilized with 0.1% saponin and stained for intracellular IFN-γ, IL-4 and IL-13. The stained cells were acquired with a FACS Calibur (BD Immunocytometry, CA) and data analyzed with Flowjo (Treestar Inc., OR).

Statistics

Data between multiple groups were compared by one-way analysis of variance with Bonferroni post hoc testing, and where applicable, data between two groups were compared by a two-tailed Student’s t-test. P-values less than 0.05 were considered statistically significant.

Results

IFN-γ activity increases in the lungs of IL-13^−/− mice

To determine whether the type-1 cytokine IFN-γ exhibits regulatory activity in models of IL-13 driven fibrosis, C57BL/6 wild type (WT), IL-13^−/−, and IFN-γ^−/− mice were i.p. sensitized with 5,000 live Schistosoma mansoni eggs and then challenged intravenously on day 14 with a second batch of live eggs to induce IL-4 and IL-13-dependent granuloma formation in the lungs [23]. As expected, WT mice undergoing secondary granulomatous inflammation displayed marked increases IL-4 and IL-13 in leukocytes isolated from the lung on day 7, with the majority of the cytokine restricted to CD4⁺ T cells (Fig 1A). Although S. mansoni granuloma formation is often described as type-2 cytokine polarized inflammatory response [24], we also observed significant IFN-γ expression in both CD4⁺ and CD4⁻ cells on day 7, confirming that the lung granulomatous response is more accurately characterized by a mixed type 1/type 2 cytokine response. Egg exposed IFN-γ^−/− mice displayed modest increases in CD4⁺ T cell derived IL-13 and IL-4 compared with WT mice, confirming an important cross-regulatory role for type-1 IFN-γ responses in the modulation of type 2 immunity. Consistent with these observations, egg-challenged IL-13^−/− mice displayed a similar but even more marked increase in IFN-γ in CD4⁺ T cells (Fig. 1A, middle panels and Fig. 1B). Additionally, the heightened IFN-γ activity, likely due to the increase in IFN-γ secreting CD4⁺ T lymphocytes, was associated with increased production of the type-1 driven chemokines Cxcl10 and Cxcl9 (Fig. 1B). As expected, no difference in Ifng expression was observed between naïve groups (not shown). Together, these data demonstrate that IL-13 and IFN-γ responses are simultaneously induced but reciprocally regulated in the lung following exposure to S. mansoni eggs.

Figure 1. Increased IFN-γ activity in the absence of IL-13.

(A) Flow cytometric analysis of ex vivo intracellular cytokine staining of lung leukocytes isolated 7 days post intravenous S. mansoni egg challenge into primed mice. Gate frequencies indicate percentage of cytokine positive live lymphocytes. (B) RT-qPCR analysis of lung tissue to assess downstream IFN-γ activity by assessing to chemokines that regulate type 1 inflammation. Y-axis represents fold change over naïve WT mice. Data are representative of two similar experiments with an n of 3–5 mice per group per iteration. ** p < 0.01; *** p < 0.005

IFN-γ is not required for the reduction in pulmonary fibrosis in IL-13^−/− mice

IFN-γ has been shown to exhibit anti-fibrotic activity by negatively regulating TGF-beta induced extracellular matrix (ECM) production by myofibroblasts [25]. Nevertheless, whether type 1 immunity or IFN-γ more specifically plays a critical anti-fibrotic role in models of IL-13-driven fibrosis was unclear. To investigate the potential regulatory role of IFN-γ, we bred IL-13^−/− mice with IFN-γ^−/− mice, and compared granuloma development and fibrosis in the double knockout (DKO) mice with WT, IL-13-, and IFN-γ-deficient animals. Although IFN-γ^−/− mice displayed no significant change in granuloma formation or fibrosis compared with WT mice, granuloma size was reduced approximately 50% (Fig. 2A, B) and fibrosis was reduced more than 65% in the IL-13^−/− mice (Fig. 2C, D). Surprisingly however, the DKO mice developed significantly larger granulomas than IL-13^−/− mice, while fibrosis was unchanged or even slightly decreased in the DKO mice when compared with IL-13^−/− mice (Fig. 2C, D). Here again, the IFN-γ-inducible chemokines Cxcl9 (MIG) and Cxcl10 (IP-10) were significantly increased in the lungs of IL-13^−/− mice relative to WT mice but both returned to baseline levels in DKO mice (Fig. 2E). In contrast, while there was relatively little change in Tgfb1 expression, there was a modest decrease in Tgfb1 in IL-13^−/− mice that was restored when IFN-γ was simultaneously knocked out (Fig. 2E). Nevertheless, this pattern did not correlate with the degree of fibrosis (Fig. 2C). Consequently, these data confirm that IL-13, but not TGF-β1, functions as the dominant pro-fibrotic mediator in this model [26]. The findings also unequivocally demonstrate that the reduction in pulmonary fibrosis observed in IL-13^−/− mice is not dependent on increased IFN-γ effector function.

Figure 2. IFN-γ regulates granuloma size but not pulmonary fibrosis in IL-13^−/− mice.

Secondary lung granulomas were induced in mice that were previously primed with S. mansoni eggs. On day 7, lungs were harvested and processed for histology. Granuloma size (A) was measured from Giemsa stained sections (B) and lung fibrosis was scored (C) from picrosirius red stained sections (D) by a trained pathologist unaware of the experimental groups. (E) RNA extracted from a lung lobe used for RT-qPCR analysis of transcripts for Cxcl9, Cxcl10 and Tgfb1. Data are representative of two similar experiments with an n of 3–7 mice per group per iteration. * p < 0.05

Liver fibrosis is maximally reduced in the combined absence of IL-13 and IFN-γ

The previous studies examined the roles of IL-13 and IFN-γ in an acute 21-day sensitize and challenge model of pulmonary fibrosis. Next, to determine whether IFN-γ and IL-13 play similar roles in another organ system and in a more chronic model of fibrosis, we infected WT, IL-13^−/−, IFN-γ^−/−, and DKO mice with the helminth parasite S. mansoni and granulomatous inflammation and fibrosis were evaluated in the liver at acute (8 wk) and chronic (12 wk) time points post-infection. In contrast to the acute lung granuloma model in which IL-13^−/− mice displayed an approximate 50% reduction in granuloma size on day 7, granulomatous inflammation increased slightly but significantly at wk 8 post-infection in the livers of IL-13^−/− mice (Fig. 3A), suggesting a pro-inflammatory role for type 1 immunity in the liver, which was confirmed in the DKO mice. Nevertheless, despite displaying a modest increase in type-1-driven inflammation, fibrosis was markedly decreased in the IL-13^−/− mice (Fig. 3B). In contrast to IL-13^−/− mice, IFN-γ^−/− mice displayed no difference in fibrosis on wk 8 when compared with WT mice. However, their granulomas were slightly but consistently smaller, again supporting a pro-inflammatory role for IFN-γ in the liver. Strikingly however, granuloma size and fibrosis were both maximally and significantly reduced in the DKO mice on wk 8 (Fig. 3A, B). A similar attenuation of inflammation and fibrosis was also observed in the DKO mice at the 12 wk chronic time point (Fig. 3 C–E), although fibrosis was even more significantly reduced in the DKO mice relative to IL-13^−/− mice at this time point (Fig. 3D). It is important to note that worm burdens and tissue egg burdens were similar in all groups at all time points (Table 1), so the observed pathological changes were not attributed to changes in infection intensity or granuloma number.

Figure 3. Liver fibrosis and granuloma formation are diminished in the combined absence of IL-13 and IFN-γ.

Livers from mice infected with S. mansoni for 8 (A, B) or 12 (C–E) weeks were processed for histology and granuloma size (A) and (C, D) was enumerated from Giemsa-stained sections by a pathologist unaware of the treatment group. Collagen content of the liver was assessed by quantifying the hydroxyproline levels (B, E) and expressed after normalization to liver egg burden. Data are representative of two similar experiments with an n of 7–13 mice per group per iteration. ** p < 0.01; *** p < 0.001; **** p < 0.0001

Table 1.

Groups of WT and the respective KO mice (n = 9–13) were infected with 35 S. mansoni cercariae and sacrificed at 8 and 12 wk post infection. Worm numbers and liver egg burdens enumerated as described in methods, and presented as mean ± SD. The two time points represent independent experimental infections and are representative of 2 similar iterations.

		WT	IL-13−/−	IFN-γ−/−	13/γ DKO
8 Wk	WP	9 ± 3.2	12.6 ± 4.9	8.6 ± 2.7	10.2 ± 3.9
	Eggs (×103)	39.2 ± 20	38.9 ± 12	31.9 ± 12	41.2 ± 15
12 Wk	WP	4.5 ± 1.9	4.2 ± 2	4.6 ± 1.7	4.6 ± 1.9
	Eggs (×103)	40.7 ± 18	29 ± 9.2	32 ± 13	33.9 ± 17

Mediators of inflammation and fibrosis are maximally reduced in DKO mice

To understand why fibrosis was maximally ablated in the combined absence of IL-13 and IFN-γ, several genes that have been previously linked with important inflammatory and pro-fibrotic pathways were evaluated in the liver by quantitative PCR analysis [27–31]. Interestingly, although fibrosis was significantly decreased in the livers of IL-13^−/− mice, Tgfb1 expression was maximally elevated in this group, thus supporting previous studies that concluded TGF-β1 plays little to no role in the pathogenesis of schistosomiasis in WT mice [26] (Fig. 4). We theorize the increased TGF-β1 activity in IL-13^−/− mice might be an attempt to control the increased type-1 associated inflammation in these mice, as TGF-β1 is well known to regulate type 1 immunity [32]. Consistent with this hypothesis, Tgfb1 expression returned to baseline levels in the DKO mice (Fig. 4). The combined decrease in type-1-associated inflammation, Il13 expression, and TGF-β1 activity likely explains why fibrosis was maximally reduced in this group. Several other pro-inflammatory and pro-fibrotic mediators were also significantly reduced in the livers of the DKO mice, including Tnf, Chil3, Timp1, Mmp9, and Mmp12, further supporting this conclusion (Fig. 4).

Figure 4. Differential expression of granuloma associated cytokine and matrix modulators in liver.

Liver biopsies of 100–150 mg each obtained from mice infected with S. mansoni for 12 wks were used to extract RNA and RT-qPCR was performed to assess the relative abundance of transcripts for various granulomatous inflammation associated genes. Y-axis represents fold change relative to naïve WT liver samples. Data are representative of two similar experiments with an n of 5–7 mice per group per iteration. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001

Sustained amelioration of inflammation and fibrosis by dual IL-13/IFN-γ deficiency

To determine whether other pro-fibrotic mechanisms might compensate for the absence of IL-13 and IFN-γ, WT, IL-13^−/−, IFN-γ^−/−, and DKO mice were chronically infected with S. mansoni and granuloma size and fibrosis were assessed at a late 20 wk chronic time point. As seen at earlier time points, IFN-γ^−/− mice developed fibrosis at comparable levels with WT mice, while IL-13^−/− mice showed a significant reduction in fibrosis (Fig. 5A). Nevertheless, the DKO mice again displayed the most significant reduction in fibrosis. The granulomas in DKO mice were also significantly smaller than the other three groups (Fig. 5B). However, the cellular composition of the granulomas was mostly indistinguishable between the groups, although IL-13^−/− mice displayed a modest increase in eosinophils that returned to WT levels in the DKO mice (Fig. 5D). The increase in granuloma associated parenchymal necrosis observed in IL-13^−/− mice was also ameliorated in DKO mice (Fig. 5C), suggesting a critical role for IFN-γ. Finally, liver transaminase levels were also maximally reduced in DKO mice (Fig. 5E and 5F), again supporting the conclusion that DKO mice are better protected from egg-induced pathology than IL-13^−/− mice. Finally, further evidence that IFN-γ plays an important regulatory role was obtained by monitoring expression IL-13Rα2, which is induced by IL-13/Stat6 mediated signalling and suppressed by IFN-γ [33]. As expected, serum levels of total (Fig. 5G) and IL-13-complexed (Fig. 5H) Il13ra2 were both upregulated in infected WT mice and substantially reduced in IL-13^−/− mice. Strikingly, however serum levels of Il13ra2 were significantly elevated in IFN-γ^−/− mice, confirming an active regulatory role for IFN-γ even in the late chronic stages of the disease. Together, these data illustrate the important collaborative roles played by IL-13 and IFN-γ in the initiation and maintenance of S. mansoni egg induced pathology.

Figure 5. Diminution of liver fibrosis and granulomatous inflammation persists chronically in the absence of IL-13 and IFNγ.

Collagen content of liver samples from mice infected for 20 wk with S. mansoni assessed by hydroxyproline assay (A) and granuloma size (B), parenchymal necrosis (C) and eosinophil prevalence (D) were determined from Giemsa-stained histology sections from each mouse, scored by a pathologist unaware of the treatment group. Serum transaminases (alanine (E) and aspartate (F)) to assess liver function were measured from the same cohort at 12 wk post infection. Total serum Il13ra2 levels (G) and endogenously IL-13 bound Il13ra2 levels (H) were measured by a modified sandwich ELISA.

Dual blockade of IL-13 and IFN-γ provides superior protection from fibrosis

Monoclonal antibodies targeting IL-13 are currently being tested in numerous clinical trials [41–43]. As our studies with IL-13^−/− and DKO mice provided evidence that blocking IL-13 in combination with IFN-γ might provide superior protection from both inflammation and fibrosis, in a final series of experiments we investigated whether neutralizing mAbs to IL-13 and IFN-γ could duplicate the results obtained with the transgenic mice. In these experiments, WT C57BL/6 mice were infected with 35 S. mansoni cercariae and 5 mice per group were treated twice weekly with anti-IL-13, anti-IFN-γ, or both monoclonal antibodies between week 7 and 11 following infection and inflammation and hepatic fibrosis were evaluated on day 80.

Similar to mice with genetic deletions, anti-IFN-γ treated mice developed fibrosis at comparable levels as mice treated with isotype control antibodies, while anti-IL-13 treated mice showed a significant reduction in fibrosis (Fig. 6A, B). Mice treated with both antibodies displayed the most consistent and marked reduction in fibrosis (Fig. 6A, B). Despite the significant changes in fibrosis, anti-IL-13 treated mice developed granulomatous inflammation comparable to isotype controls (Fig. 6C, D) concurrent with a significant increase in Tnf production by CD4⁺ T-cells (Fig. 6E). In contrast, the granulomatous inflammatory response was significantly reduced in anti-IFN-γ treated mice and reduced even further in mice treated with both neutralizing mAbs (Fig. 6C, D). The marked inflammation in WT and anti-IL-13 treated mice was associated with greater parenchymal damage as assessed by serum transaminase levels (Fig. 6F, G) and measures of granuloma-associated necrosis (Fig. 6H, I). In contrast, neutralizing IFN-γ a□□□□ resulted in significantly reduced serum transaminase levels (Fig 6 F, G) and hepatocellular necrosis (Fig. 6H, I), with dual blockade significantly reducing all measures of parenchymal damage as well as fibrosis.

Figure 6. Diminution of liver fibrosis and granulomatous inflammation persists chronically in the absence of IL-13 and IFNγ.

Mice were infected percutaneously with 35 S. mansoni cercariae and treated with anti-IL-13, anti-IFN-γ, both neutralizing antibodies, or with isotype matched control antibodies starting on day 48 post-infection. Antibodies were administered twice weekly for a total of 4.6 weeks. Collagen content of liver samples from mice infected for 80 days with S. mansoni assessed by picrosirius red staining (A), hydroxyproline assay (B) and, Wright’s Giemsa-stained histology sections (C) from each mouse were used to determine granuloma volumes (D) by a pathologist unaware of the treatment group. Tnf⁺ CD4 T-cells numbers (E) were determined by flow cytometry. Serum transaminases (alanine (F) and aspartate (G)) to assess liver function were measured from the same cohort at 12 wk post infection. Representative examples of liver necrosis are shown in Giemsa-stained slides (H) and quantified as shown on a scale of 0–4 (I). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Discussion

Although the hypothesis tested in this study was straightforward, the findings generated were surprising and unexpected. We originally theorized that the reduction in fibrosis observed in the absence of IL-13 might at least be partially due to increases in the expression of the anti-fibrotic cytokine IFN-γ. Indeed, numerous reports have demonstrated that type 1 (IFN-γ) and type-2 (IL-4/IL-13) responses are strongly cross-regulatory [34,35]. Studies have also supported the concept that fibroblast activation, proliferation, and collagen deposition are reciprocally regulated by type 1 and type 2 cytokine responses [36–39], with IFN-γ suppressing myofibroblast function and IL-4 and IL-13 augmenting their activity [16,40]. Nevertheless, whether increases in IFN-γ activity play a protective or pathogenic role when IL-13 activity is suppressed was not previously investigated in vivo. Although we detected marked increases in IFN-γ and type 1-associated inflammation in IL-13^−/− and anti-IL-13 mAb treated mice, the elevated IFN-γ response was not required for the reduction in pulmonary and hepatic fibrosis induced by schistosome eggs. In fact, we observed a more significant amelioration of both fibrosis and inflammation when IL-13 and IFN-γ were simultaneously ablated.

This hypothesis was important to investigate because therapeutic antibodies targeting IL-13 are currently in multiple phase II/III clinical trials for moderate to severe asthma and idiopathic pulmonary fibrosis [41–43]. Our findings together with a recent study of allergic airway disease suggest that IL-13 blockade alone, while highly efficacious for the treatment of type 2 cytokine driven asthma and fibrosis, may lead to compensatory increases in type 1 associated inflammation that could complicate treatment over time [3]. Early studies investigating the function of IL-13 identified the cytokine as a potent inhibitor of inflammatory cytokine production in human peripheral blood monocytes and mouse macrophages [44] [45]. Consistent with these observations, we detected marked increases in Tnf and inflammatory chemokine expression in the lungs and livers of IL-13-deficient mice, which were markedly decreased when IL-13 and IFN-γ were simultaneously ablated. The inflammatory response in the livers of the DKO mice, assessed by granuloma size, was also reduced more than 50% relative to IL-13-deficient mice at both the acute and chronic time points, which likely contributed to the reduction in fibrosis. Tissue necrosis and liver transaminase levels were also significantly increased in IL-13^−/− mice but decreased when IFN-γ was simultaneously eliminated, further illustrating the emergent pathogenic role of IFN-γ when IL-13 activity is blocked. Eliminating the increased IFN-γ response also led to marked reductions in Tgfb1, Tnf, Timp1, Mmp9, and Mmp12 expression, which likely contributed to the marked decrease in collagen deposition in the DKO mice, as all of these mediators have been directly linked with the development of fibrosis [11,28,46,47]. Similar results were obtained when infected wild type mice were treated with neutralizing mAbs to IL-13 and IFN-γ. Given these findings, it will be important to determine whether patients receiving neutralizing antibodies to IL-13 develop similar compensatory increases in type-1 associated inflammation. If so, therapeutic efficacy might be improved with a bi-specific, bi-functional, or combination therapy that reduces IL-13 and IFN-γ activity simultaneously.

The role of type-1 and type-2 responses in the pathogenesis of schistosomiasis in mice and humans has been debated for years, with some studies suggesting that dysregulated type-1/IFN-γ responses are critical to the development of hepatosplenic disease and other studies supporting the concept that type-2 immunity is essential to the progression of hepatic fibrosis and the life threatening complications associated with advanced cirrhosis [48–53]. Although early studies in mice suggested that once egg laying commences, granuloma formation is characterized by a strongly polarized type 2 response [54], our findings support studies that have also identified important regulatory roles for IFN-γ. Perhaps the best evidence supporting this conclusion was the striking increase in serum Il13ra2 levels observed in IFN-γ^−/− mice at 20 weeks post-infection, as IL13Rα2 is induced by IL-13/Stat6-mediated signalling and inhibited by IFN-γ [33]. These findings confirmed that IL-13 and IFN-γ exhibit potent cross-regulatory activity at all stages of schistosomiasis, emphasizing the importance of targeting both mediators simultaneously so that type 1-associated inflammation, which increases when IL-13 is targeted individually, and IL-13-driven fibrosis are both maximally ameliorated.

Therapeutic drugs for fibrotic diseases are still in the early stages of development. Although IL-13 has emerged as a promising therapeutic target for several diseases that are characterized by persistent or dysregulated type 2 cytokine responses [1,55,56], our studies suggest that in some cases rebound inflammation driven by the type 1 cytokine IFN-γ may be an unwanted complication resulting from IL-13 blockade. In fact, it is tempting to speculate that rebound inflammation may be common side effect associated with anti-fibrotic therapy as wound healing and fibrosis are tightly linked with the suppression of inflammation [5,10]. Indeed, the pro-fibrotic cytokine TGF-β also serves as a potent anti-inflammatory mediator [32,57], with deficiencies in Tgfb1 or TGF-β1 signalling linked with the development of numerous autoimmune and inflammatory diseases [58,59]. Consequently, therapeutics targeting core pro-fibrotic pathways will need to be carefully evaluated for this potential undesired side effect. Our findings suggest that highly effective anti-fibrotic therapies may need to be combined with therapeutics that mitigate this rebound inflammation as this will likely lead to a more substantial and sustained reduction in fibrosis.