Effect of TLR ligands in nitrofen-induced CDH

Initially, we tested several TLR2 ligands in a rat model of CDH induced by nitrofen. Nitrofen was administered to pregnant rats at embryonic stage E9.5 to induce CDH and 3 days later (E12.5), in order to allow organ alterations to appear, several TLR ligands were injected. On day 21st of gestation, the rats were sacrificed and the embryos were examined for the presence or absence of diaphragmatic hernia (CDH + or CDH-, respectively). Pregnant rats exposed to the herbicide nitrofen had between 60–85% of fetuses with CDH (73% on average in Fig. 1A) in line with many previous studies [35]. In this model, the severity is high and hernia is prominent. Thus, we found that all herniated fetuses have left involvement with liver in the thorax and 75% both left and right side in agreement to what it has been described [36]. Treatment of the pregnant mothers with a single dose of TLR2/1 or TLR2/6 ligands, 3 days after nitrofen administration reduced on average CDH + fetuses from 73% in saline only treated animals to 46% and 52% respectively (Fig. 1A). This represents a relative decrease (or improvement) in CDH incidence caused by nitrofen of 37% and 29% respectively. More importantly, atypical LPS which are a new class of dual TLR2-TLR4 ligands from soil bacteria (rhizobiales) [29] were much better at reducing this CDH incidence. Thus, the LPS from Rhizobium rhizogenes K-84 (from now on referred as CS1) reduced CDH incidence to only 16% whereas LPS from Ochrobactrum intermedium (referred as CS2) resulted in a 21% incidence. This represents a relative decrease of 79% and 72% in hernia incidence respect to the nitrofen with saline treatment (Fig. 1A). Treatment with LPS from E.coli was toxic to the fetuses, and no fetuses were recovered.

Fig. 1

Effect of TLR ligands in diaphragmatic hernia induced by nitrofen. A Pregnant Wistar rat females were administered nitrofen (n = 6), or olive oil as placebo (control) (n = 2), at E9.5 and 3 days later injected intraperitoneally with 100 μg/Kg of the indicated immunomodulators or only saline (None or Untreated). Results are represented as percentages of all fetuses pregnant rats with hernia (CDH +) at E21. B and C Pregnant Wistar rat females were administered nitrofen at E9.5 and 5 h, 1 day or 3 days later injected intraperitoneally with 100 μg/Kg of CS1 or saline. B Diaphragmatic hernia percentage. Results are represented as percentages of all fetuses with hernia (CDH +) at E21. C The weight of the fetuses recorded at E21. **p > 0.001, *p > 0.01 respect to untreated nitrofen group (control). In this case, all fetuses from 3 rats from each group, independent or not of the presence of hernia, were analyzed

Full severe hernia may take some time to appear after nitrofen. Thus, adding the compounds at different times aftrer nitrofen, may distinguish “preventive” from “curative” effects of compounds. We treated with CS1, 5 h, 1 or 3 days after nitrofen exposure at day E9.5. Interestingly, the efficiency in reducing the percentage of herniated CDH + fetuses was higher when the treatment with CS1 was applied 1 day and especially 3 days after nitrofen administration, reaching then the maximum healing of the hernia (Fig. 1B). Average embryo weight showed a slight increase following treatment with CS1 compared to saline-treated rats, (Fig. 1C). Although, both CS1 and CS2 compounds behaved very similarly in all biological assays, only those for the best one (CS1) are shown below.

The histological analysis of fetuses from pregnant mothers treated with CS1 were analyzed at E21 (Fig. 2). In healthy fetuses, a complete diaphragm is clearly observed with a normal morphology and homogeneous muscle fibers separating the abdomen from the thorax. As expected, fetuses from mothers that were administered nitrofen presented with an invasion of the thorax by the liver, due to partial or total absence of the diaphragm, combined with amorphous morphology. Importantly, treatment with CS1 immunomodulator prevents damage to the diaphragm. Thus, at E21, most of fetuses of nitrofen administered mothers and treated with the immunomodulator, presented a complete diaphragm although with a more disorganized and hypertrophied morphology (Fig. 2 D,E,F).

Fig. 2

Effect of CS1 on the diaphragms of nitrofen-treated fetuses. Pregnant Wistar rat females were administered nitrofen, or olive oil (control), at E9.5 and 3 days later injected intraperitoneally with 100 µg/Kg of CS1 immunomodulator or only saline. Histological analyses of the fetuses were performed at E21. The representative images show fetuses from the healthy (Control), nitrofen with hernia (Nitro CDH +) and CS1-treated nitrofen without hernia (Nitro + CS1 CDH-) groups. (A, C, E) photographic composition of a cross-section of the fetuses of the different groups. A complete diaphragm is clearly observed in the control group while in the fetuses treated with nitrofen the liver (Li) has invaded the thorax due to the absence of the diaphragm. In the nitrofen fetus treated with the CS1 a complete diaphragm is also observed. (B, D, F) magnification of the diaphragm (*). The control fetus diaphragm (B) shows a normal morphology and homogeneous muscle fibres separating the abdomen from the thorax (A, arrow). In contrast, the incomplete diaphragm of the (Nitro CDH +) group (D) presents an amorphous morphology. Although there was no hernia in Nitro + CS1 CDH- group disorganised muscle fibre structure was also observed (F)

Alteration and retarded development of the lung is a characteristic of CDH [1, 3]. Histological analysis of the embryonic lungs showed clear differences among the groups treated with CS1 at E18 or E21 (Additional file 1: Fig. S1). Those lungs from nitrofen CDH + fetuses were notably smaller than the rest including the CS1-treated CDH + fetuses (with herniation), and even smaller than CS1-treated CDH- fetuses (without herniation). Furthermore, the lungs of CDH + nitrofen fetuses showed altered morphology with less defined lobes. Importantly, CS1-treated nitrofen fetuses exhibited lung morphology resembling that of normal mature lungs in 21-day-old fetuses with an increased number of respiratory bronchioles, alveolar ducts and alveoli (Additional file 1: Fig. S1), while saline treated CDH + fetuses displayed a more immature lung appearance, with underdeveloped and less clearly defined alveolar structures. The differences in alveolar ducts, alveoli, and respiratory bronchioles were evident as well. The untreated nitrofen group showed reduced airspace, especially in CDH + fetuses, when compared to the treatment group and healthy controls. In contrast, the treatment group demonstrated better lung development in terms of size and an increased number of alveolar ducts especially in CDH- fetuses.

Quantification of the alveolar volume indicates that it was reduced around 40% in CDH + nitrofen fetuses, whereas was similar to control placebo olive oil control rats in nitrofen CS1-treated CDH- fetuses (Fig. 3A). Morphometric analyses were conducted to assess pulmonary vascular hypertrophy. Quantification of radial alveolar count demonstrated a substantial reduction in animals CDH + after nitrofen treatment that was greatly substantially recovered in nitrofen CS1-treated CDH (Fig. 3B). In contrast, alveolar wall thickness increased after nitrofen and again was reversed in CS1-treated CDH- animals (Fig. 3C).

Fig. 3

Effect of TLR ligand in lung development. Pregnant Wistar rat females were treated with nitrofen at E9.5 and 3 days later injected intraperitoneally with 100 µg/Kg CS1. A Percentage of empty alveolar to cellular space in lungs. This was analyzed from 4 different regions of 4 different fetuses in each group. B Radial alveolar counts were performed analyzing several images of 4 fetuses from each group as escribed in methods. C alveolar thickness was measured as the mean septal wall thickness of terminal alveoli per field of the field measured. D Wall thickness/vessel radius ratio. Analysis of E18 lung vessels from the different groups. 3 independent calibrations for different lung sections of 3 different fetuses were staining with van Giemson. E Representative van Giemson staining of the indicated groups:. **p > 0.001 and *p > 0.01, respect to control healthy group; ☨☨p > 0.001 and ☨p > 0.01 respect to untreated CDH + nitrofen group

Healthy controls exhibited a highly stained media layer of the aorta wall with Van Giemson staining. In contrast, the nitrofen-CDH + showed reduced staining of elastic fibers, indicating areas of tissue injury (Fig. 3D-E). However, CS1 immunomodulatory treatment reversed this effect, supporting the beneficial impact of CS1 treatment in mitigating the negative effects of pulmonary hypoplasia and associated pulmonary vascular changes in CDH.

Effect of TLR ligands in WT1 conditional KO (G2-GATA4Cre;Wt1fl/fl) mice

Next, we decided to confirm the effects of CS1 treatment, using a genetic model, the G2-GATA4Cre;Wt1fl/fl mice, with conditional ablation of the Wilms’ tumor suppressor gene (Wt1) in lateral plate cells expressing GATA4 under the control of the G2 enhancer [10]. The use of this WT1 conditional knockout overcomes the early embryonic death caused by systemic deficiency of WT1 and it constitutes a valuable animal model of CDH. In this model, WT1 is involved in the generation of the mesenchyme of the ST/PHMP/PPFs continuum through epithelial-mesenchymal transition. Defect in the PHMP in G2-Gata4Cre;Wt1fl/fl mutant embryos and a strong reduction of the ST mesenchyme could be observed as early as E10.5 G2-Gata4Cre;Wt1fl/fl mutant embryos as previously described [10]. Thus, we treated G2-GATA4Cre;Wt1fl/fl pregnant mice mothers with CS1 or PBS intraperitoneally twice at E9.5 and E10.5 about 2 days later than the WT1 deletion in the embryos was genetically induced. Embryos were analyzed at E15.5 (Fig. 4 and Additional file 1: Fig. S2). In absence of treatment about 70% of the embryos developed typical Bochdalek-type CDH, with diaphragmatic defect, liver invasion of the left pleural cavity and hypoplasia of the left lung. Very remarkably, CS1 treatment rescued the CDH phenotype in the G2-GATA4Cre;Wt1fl/fl model. Of the 11 mutant fetuses analyzed, only 1 had diaphragmatic hernia (only 9% incidence) (Fig. 4). In all but one CS1-treated mutants the pleural cavities were completely closed in contrast to the untreated mutant mice (Additional file 1: Fig. S2). Very importantly, the sinus venosus of all CS1-treated embryos was abnormal indicating that WT1 was in fact deleted in all CS1-treated animals (Additional file 1: Fig. S2). This is an internal control to confirm the deletion of WT1 and indicates that hernia was healed by CS1 and that the observed decrease on CDH was not due to a deficient deletion of the WT1 gene in those animals treated with CS1. Although the diaphragm in most cases does not appear normal (as in immunomodulator-treated nitrofen CDH + rats), being more irregular, with less organized musculature and hypertrophied, it was completely closed (Additional file 1: Fig. S3). It is also important to mention that control embryos treated with CS1 were normal (Fig. 4 and Additional file 1: Fig. S3). Treatment with CS1 was neither toxic to the normal pregnant mothers.

Fig. 4

Effect of CS1 in G2-GATA4Cre;Wt1fl/fl mice. Pregnant mice mothers were treated with 100 µg/Kg of CS1 or saline intraperitoneally, twice at E9.5 and E10.5. Embryos were analysed at E15.5. Images representatives from WT1 mutant embryo of mothers treated with PBS or CS1. Embryos from control not mutant mothers are also shown

Immune cell infiltration in damaged organs in CDH

First, we measured immune cell infiltration through the expression markers of macrophages (CD68 +), B (CD20 +) and T (CD3 +) lymphocytes and neutrophils (p67 +) with immunohistochemistry techniques (Fig. 5). In the CS1 treated groups, both CDH- and CDH + fetuses were analyzed separately.

Fig. 5

Immune infiltration in lung and diaphragm induced by nitrofen and CS1 treatments. Pregnant Wistar rat females (3 per group) were treated with nitrofen or olive oil (control) at E9.5 and 3 days later injected intraperitoneally or with 100 µg/Kg CS1 or saline. A–D Indicated cell infiltration in the diaphragm, left and right lung was determined by immunohistochemistry at E21. In the CS1 treated groups, both CDH-negative and CDH-positive fetuses were analysed separately. E CD206 expression was determined by immunofluorescence at E18. **p > 0.001 and *p > 0.01, respect to control normal group; ☨☨p > 0.001 and ☨p > 0.01 respect to untreated CDH + nitrofen group

In rats with CDH induced by nitrofen, a significant and high increase in CD3 T cell infiltration was observed in both the left and right lungs as well as in the diaphragm of the embryos at E21. However, in fetuses from pregnant rats treated with CS1, this infiltration was notably reduced, even in the few cases where the diaphragm did not close completely. This reduction was particularly remarkable in the “cured” CDH- fetuses (Fig. 5). No relevant effect was observed in B cell infiltration (Fig. 5). Examining the percentage of p67 + cells, a marker for neutrophils, in lung tissue, we found a lower number of p67 + cells in both the control and CS1-treated CDH- groups (Fig. 5). This indicates a higher neutrophil infiltration in CDH + pups. No p67 + infiltrating cells were detected in the diaphragms.

We observed a significant increase of macrophage infiltration (CD68 + cells in all CDH + groups compared to the control or CDH- groups. Interestingly, although nitrofen induced an infiltration of CD68 + macrophages in left lung and diaphragm, CS1 significantly increased this infiltration, especially in the CDH- fetuses (Fig. 5). To corroborate this, we specifically studied the repairing macrophage population by using the M2 macrophage specific marker CD206 (Fig. 5E). A great infiltration by M2 macrophages was observed in all nitrofen-treated animals. However, in CS1-treated CDH- animals the infiltration of this M2 population to the diaphragm was 2–threefold higher than in nitrofen CDH + animals. This indicates that there is an active recruitment of these M2 cells to facilitate the repair of the damaged tissue in CS1-treated CDH- animals. Infiltration in the lungs and diaphragm in the CS1-treated but still with hernia CDH + , animals is modified with the same tendency than Nitro + CS1 CDH- animals, although in almost all cases (except for CD206 macrophages and CD3 T cells) did not reach statistical significance.

Gene expression regulation by CS1 in CDH

To gain a better understanding of the effects of CS1 in tissues, we analyzed the gene expression of immune, and developmental or RA pathway genes, known to be related to CDH [12, 13], in the lungs, diaphragm, and spleen. Figure 6 provides a summary of two independent experiments, each including all fetuses from pregnant rats treated with nitrofen at E9.5, independent of the presence or absence of hernia. The results were normalized for comparison with the untreated control group using real time QC-PCR. Fetuses from control animals treated with CS1 at E12.5 showed no significant changes in gene expression at E21 compared to fetuses from untreated control rats. Nitrofen administration alone induced few changes in gene expression in the analyzed organs at E21. The most notable ones were the induction of Arg1 in the spleen and diaphragm, Ccl2 in the diaphragm, and Wt1 and Stra6 in the lungs. Interestingly, treatment with CS1 of nitrofen animals resulted in a significant increase in Arg1, Aldh1a2, Rbp1 and Rarb in the diaphragm. The spleens from fetuses of CS1-treated rats also showed increases in Arg1, Aldh1a2, Rara, Ccl2, Pparg, and Slit3. In the lungs, however, only Arg1 showed a significant and high induction following CS1 treatment. Additionally, there was a tendency for increased expression in retinoic pathway genes Aldh1a2, Rara and Rarb, while the inductions of Stra6 and Wt1 caused by nitrofen were reduced by CS1. These results indicate a significant increase in Arg1 and the RA pathway due to CS1 treatment in the affected organs and spleen.

Fig. 6

Effect of TLR ligand CS1 on gene expression in vivo. Pregnant Wistar rat females were treated with nitrofen or olive oil (placebo) at E9.5 and 3 days later injected intraperitoneally with 100 µg/Kg CS1 or saline. RT-PCR of mRNAs of the indicated proteins in the Diaphragm, Spleen or Lung organs was carried out at E21. Results are the mean of 2 independent experiments with 5 fetuses per group each. Results are shown normalized to controls of untreated rats. *p > 0.01, respect to untreated CDH + nitrofen group

CS1 induce retinoic pathway genes in macrophages

To further understand the effect of the treatments in the immune activation, we investigated the ability of CS1 to activate macrophages in vitro (Fig. 7). CS1 induced RA genes Aldh1a2 and Rbp1, Arg1, a marker of M2 macrophages and also weakly induced Slit3 which has recently been found to be also expressed in M2 mouse peritoneal macrophages [37]. In addition to those genes, CS1 also induced Il12. Interestingly, by using macrophages from TLR2 or TLR4 deficient macrophages, we found that those effects were dependent on both TLR2 and TLR4 receptors. These in vitro effects align with the observed infiltration of macrophages and gene expression profile primarily observed in the diaphragm, but also in spleen and lungs, of the CS1-treated fetuses shown above.

Fig. 7

Effect of TLR ligands on gene expression by macrophages. Peritoneal macrophages from WT, TLR2 or TLR4 deficient mice were stimulated with CS1 at the indicated doses and 14 h later the gene expression Aldh1a2, Slit3, Rbp1, Arg1 and Il12 was analysed by RT-PCR. (*) p > 0.001