Prmt2 Regulates the Lipopolysaccharide-Induced Responses in Lungs and Macrophages

The LPS after intranasal instillation induces a complex response in the lungs that is tightly regulated (1). The LPS binds adaptors and activates the TLR4, leading to the activation of different signaling pathways that will elicit the lung and the inflammatory responses. Part of the control is due to specific adaptor proteins that contribute to the integrated response. In the classical TLR4/MyD88 pathway, NF-κB is normally sequestered in the cytoplasm by an IκB complex (2–6), and the phosphorylation of IκB by inducible kinases, such as IκB kinase 1 (IκK1) and IκK2, leads to subsequent ubiquitination and degradation of these proteins. The released NF-κB then translocates to the nucleus, where it stimulates the transcription of genes involved in immune and inflammatory responses (7). NF-κB DNA binding and NF-κB–dependent transcription are attenuated by newly synthesized IκB-α in the nucleus, which associates with NF-κB/RelA complexes. As IκB-α accumulates in the nucleus, there is a progressive reduction of both NF-κB DNA binding and NF-κB–dependent transcription (8), presumably by export of NF-κB/IκB-α complexes from the nucleus (9–11). An additional pathway, independent of MyD88/MyD88-adaptor like and NF-κB, is regulated by TRAM and moderately by Toll/IL-1R domain containing adaptor inducing IFN-β (TRIF), leading to the late activation of NF-κB (12), and activates type I IFN synthesis (13).

LPS-induced inflammation is the consequence of the activation of the reticuloendothelial system leading to the synthesis of various mediators that have a local but also a general action. The principal proinflammatory cytokines are the IL-1, the IL-6, and the TNF-α. Expressions of IL-6 and TNF-α are directly regulated by NF-κB (14, 15), and these molecules are responsible for the amplification of the inflammatory response and part of the systemic responses (16).

In the course of our study to isolate mouse dosage-sensitive genes homologous to human chromosome 21 (Hsa21) genes, we developed a new mouse model of monosomy for the Prmt2-Col6a1 (noted Ms1Yah) genetic interval found on the mouse chromosome 10 (17). The Ms1Yah model is deleted for 14 genes corresponding to the telomeric end of Hsa21, which was shown to display several copy number variants, in particular some that affect the Prmt2, S100B, and the genes. The only phenotype of the Ms1Yah mice is an impaired airway response and an increased inflammatory response after LPS stimulation. Further investigations showed macrophages (MΦ) as an important cellular compartment for the increased production of proinflammatory cytokines. Several genes of the Pmrt2-Col6a1 region were found to be dosage sensitive in lungs and MΦ, and are probably involved in phenotypes observed in Ms1Yah mice (17).

To further explain the increased production of IL-6 and TNF-α in Ms1Yah mice in response to LPS, we focused our attention on the Prmt2, which modifies arginine residues during posttranslational modification of proteins (18). PRMT2 is known to act as a negative regulator of the NF-κB pathway in a dose-dependent manner. In fibroblasts, Prmt2 exerts its effect by causing nuclear accumulation of IκB-α, which concomitantly decreases nuclear NF-κB DNA binding (19). Thus, we hypothesized that the decrease in the copy number of this gene is at the origin of the increase of the inflammatory response observed in the Ms1Yah model. Indeed, the expression of proinflammatory cytokines IL-6 and TNF-α is NF-κB dependent and increases in Ms1Yah mice after stimulation with LPS. In this study, we demonstrated the involvement of Prmt2 in the inflammatory response and determined how Prmt2 heterozygotes could recapitulate the defects observed in Ms1Yah mice. The conclusion of this study is that Prmt2 is a dosage-sensitive gene contributing to the control of the LPS-induced inflammatory response by regulating nuclear level of NF-κB. More surprisingly, we found an inverse pulmonary function in Prmt2 heterozygous mice compared with that of Ms1Yah mice, suggesting that Prmt2 should interact with other dosage-sensitive genes, located in the Col6a1-Prmt2 region to recapitulate the complete Ms1Yah phenotypes.

The Ms1Yah monosomic mice, Del(10Col6a1-Prmt2)Yah, were generated by chromosomal engineering and correspond to a deletion of 14 genes located in the Col6a1-Prmt2 interval. Ms1Yah mice were genotyped by Southern blot analysis as described previously (1). The Prmt2^tm1Enbl allele carries a G119X mutation and a neocassette replacing exons 4 and 5 and part of exon 6 (18, 20). The presence of the mutation is identified by PCR in standard conditions with three primers: Prmt2<Nab>primer A: 5′-CTGAGGTATTACCAGCAGACA-3′; Prmt2<Nab>primer B: 5′-CTCTCTGATGCAGGTCTAC-3′; Prmt2<Nab>primer C: 5′-CCGGTGGATGTGGAATGTGT-3′. The primers A and B identified the wild-type (wt) allele that corresponds to a 190-bp fragment, whereas the primers B and C enable the amplification of the mutant allele with a 280-bp amplicon. All the lines used in the following experiments were maintained on C57BL/6J (B6) with >10 generations of advanced backcross level. Y.H., as the principal investigator in this study, was granted the accreditation 67-369 to perform the reported experiments.

Animals (n = 10) were treated by intranasal instillation with either isotonic saline solution or 10 μg LPS (Escherichia coli, serotype O55B5, 10 mg; Sigma-Aldrich, St. Louis, MO) in deep anesthesia. Airway response was investigated over a period of 6 h after treatment using whole-body plethysmography and measurements of the parameter of enhanced pause (Penh), a dimensionless parameter that accounts for the respiratory profile, taken in consideration the period of expiration and the variations of pressure measured in a closed chamber during the respiratory cycle (EMKA Technologies, Paris, France). PenH can be conceptualized as the phase shift of the thoracic flow and the nasal flow curves. Increased phase shift correlates with increased respiratory system resistance. PenH is calculated by the formula PenH = (Te/RT − 1) × PEF/PIF, where Te is expiratory time, RT is relaxation time, PEF is peak expiratory flow, and PIF is peak inspiratory flow (21). Data are analyzed using Datanalyst software (EMKA Technologies) and expressed as mean ± SEM. Twenty-four hours after stimulation, mice were euthanized and bronchoalveolar lavages fluids (BALFs) were analyzed for cell composition and cytokine quantification as described previously (22, 23). An aliquot was stained with trypan blue solution and analyzed to determine cellular content. After centrifugation on microscopic slides, air-dried preparations were fixed and stained with Diff-Quick (Merz & Dade) using a May–Grünwald–Giemsa coloration. Two hundred cells were counted twice for the determination of the differential counts of each cell type in the BALF. Part of the lung was stored at −80°C for the myeloperoxidase (MPO) assay.

IL-6 and TNF-α concentration were estimated by ELISA test, in standard condition with the protocol of the supplier (R&D Systems). The sera were diluted at one fourth and the supernatant of MΦ at half in PBS/BSA 1%. The 96-well plate was read by the reader EL 800 (BioTek Instruments). For the MPO measurement, the right-heart ventricle was perfused with saline to flush the vascular content, and lungs were frozen at −80°C until use. Lung was homogenized by polytron and centrifuged, and the supernatant was discarded. The pellets were resuspended in 1 ml PBS containing 0.5% hexadecyltrimethyl ammonium bromide and 5 mM EDTA. After centrifugation, 50 μl supernatants were placed in test tubes with 200 μl PBS-hexadecyltrimethyl ammonium bromide EDTA, 2 ml HBSS, 100 μl o-dianisidine dihydrochloride (1.25 mg/ml), and 100 μl H2O2 0.05%. After 15 min of incubation at 37°C in an agitator, the reaction was stopped with 100 μl NaN3 1%. The MPO activity was determined as absorbance at 460 nm against medium.

Total RNA was extracted after LPS stimulation from the MΦ and from the lung of mice using the RNAeasyR mini-kit (Qiagen). The concentration of RNA was measured by Nanodrop, and the quality was evaluated by the Agilent Bioanalyzer 2000 (Agilent). We kept all the samples that have an RNA integrity number ≥6 and with a ratio [28S/18S] ≥1.5 (24). The synthesis of cDNA was done with the Absolute 2-step QRT-PCR SYBR Green (Abgen). To determine the relative expression of IL-6 and TNF-α, we used primers pairs from Qiagen. Each primer pair was tested successfully with the efficiency ranging from 90 to 110%. The quantitative PCR was performed with 15 ng cDNA and 200 nM of each primer in a 15 μl final reaction in a Stratagene Mx4000 with a standard amplification procedure. In parallel, similar experiments were carried out for eight housekeeping genes, Actb, β2m, Gapdh, Pgk1, Rpl13a, Tbp, Tubb4, and 18SRNA, used for normalization through the Genorm procedure (25). All the tested samples were performed in triplicate, and the results were reported as the mean ± SEM.

Ten mice of each genotypes (Ms1Yah, Prmt2^+/−, Prmt2^−/−, and B6) were euthanized, and bone marrow cells were isolated from femurs and differentiated into MΦ (17). For the experiment, cells were plated in 96-well plates at 10⁵ cells/well. Cells from five mice of each genotype were stimulated with NaCl or LPS (serotype O111B4 at 100 ng/ml; Sigma-Aldrich) and IFN-γ. Supernatants were harvested after 24 h and stored at −20°C for cytokine quantification.

After differentiation, MΦ were transferred on permanox four-well slides, at 10⁵ cells/well. Cells from two mice of each genotype were stimulated with NaCl or LPS (serotype O111B4 at 30 ng/ml; Sigma-Aldrich) during 4 h. The cells were washed twice with PBS for 5 min and fixed in paraformaldehyde 4% overnight. Cells were further washed on PBS twice and twice in PBS plus glycine 0.1 M before 15 min of permeabilization in Triton 0.5%. After two washes in PBS with BSA 1%, cells were incubated with NF-κB p65 Ab (SC-109, dilution 1/50 in PBS plus BSA 1%; Santa Cruz) for 90 min. Cells were washed three times in PBS plus BSA 1% and incubated with the second Ab (Alexa 488, 711-486-156 dilution 1/100 in PBS, BSA 1%; Jackson Immunoresearch) for 1 h. Cells were washed three times in PBS, BSA 1% and mounted with Mowiol mounting medium containing DAPI (diluted 1/100). To determine the effect of Prmt2 on the nuclear accumulation of NF-κB, the presence of the transcription factor was evaluated in wt and Prmt2^−/− cells. Confocal microscopy was performed using a Leica confocal microscope. Quantification of nuclear NF-κB in individual cells from several fields was done as follows. The nucleus was identified by DAPI staining, and the outline of cell nuclei was drawn in a field with ImageJ. NF-κB pixel intensity from the nucleus of each individual cell in the field was measured as pixel intensity per square micrometer using ImageJ. For each condition, the data from 10 fields were compiled (with at least 4–6 cells/field).

Mice of each genotype (Ms1Yah, Prmt2^+/−, Prmt2^−/−, and control) received an i.p. injection of 100 μg LPS (serotype O55B5; n = 10) or saline control solution (n = 6) to evaluate the systemic inflammatory response in vivo (17). After 90 min, mice were euthanized, the blood was collected through the femoral vein and centrifuged at 2000 rpm for 15 min, and serum was collected and stored at −20°C for cytokines assays.

Statistical analysis was performed using either the parametric Fisher Student t test when applicable or the nonparametric Wilcoxon–Mann–Whitney U test via the Statgraphics software (Centurion XV, Sigma plus, Levallois Perret). Values are presented as mean ± SEM, and the significant threshold was p < 0.05 or otherwise indicated.

Using whole-body plethysmography and the PenH that reflects the respiratory pattern, we previously demonstrated that Ms1Yah mice carrying a single copy of Prmt2 have a reduced respiratory response after LPS instillation (17). To investigate the precise role of Prmt2 in this phenotype, we used the loss-of-function allele of Prmt2 engineered previously (18), and we compared control (B6), heterozygote (Prmt2^+/−), and homozygote (Prmt2^−/−) indiiduals instilled with LPS (10 μg) or saline control solution, and the experiment was carried out twice independently. As done previously in the study of the Ms1Yah model (17), the airway response was monitored by whole-body plethysmography. The instillation of saline solution did not affect the respiratory function as assessed by Penh in wt and mutant mice (Fig. 1), whereas an intranasal administration of LPS induced a strong respiratory response characterized by an increase of the Penh values within 90–120 min, which lasted 3–4 h and decreased slowly, in the control mice. Contrary to Ms1Yah mice, the Prmt2^+/− mice developed a delayed airway hyperresponse (AHR) compared with the control animal (from 80 to 145 min) with a higher level of the response and no real restoration of the initial PenH level even after 6 h (Fig. 1A). This delayed response is comparable with the AHR observed in Prmt2^−/− (Fig. 1B). The respiratory response induced by LPS in the Prmt2^+/− mice showed that the absence of only one copy of Prmt2 is sufficient to modify the pulmonary function, and that the LPS-induced response of the Ms1Yah mouse airways is controlled by one or more additional gene(s) because these mice do not present an AHR. Moreover, we demonstrated in this study a dual role of Prmt2 with a delayed response during the early phase and an increased amplitude and length of duration of the AHR.

We previously determined that Ms1Yah mice instilled by LPS have an enhanced recruitment of inflammatory cells and production of TNF-α and IL-6 both in the lungs and in the bronchoalveolar space in comparison with wt mice (17). Thus, we wanted to explore the involvement of Prmt2 in the local inflammatory response. Twenty-four hours after the instillation of LPS or saline control solution, wt, Prmt2^+/−, and Prmt2^−/− mice were euthanized, and the inflammation induced by LPS was followed by neutrophil recruitment, MPO activity, and concentration of TNF-α and IL-6 in BALFs. The mice instilled with the saline control solution did not present any changes in the type of inflammatory cell recruitment. Similar to Ms1Yah mice (17), the Prmt2^+/− and Prmt2^−/− mice showed an increase in the number of neutrophils when compared with control mice (Fig. 2A). Accordingly, the MPO activity, an enzyme specifically expressed in neutrophil cells, is increased in both mutant mice versus control (Fig. 2B). The data are in agreement with the increase in the number of neutrophils.

To go further in the investigation of the inflammatory response in the lung of Prmt2^+/− and Prmt2^−/− mice, we measured the concentration of TNF-α and IL-6 in the BALFs 24 h after inflammatory induction. LPS significantly increased the concentration of the two proinflammatory cytokines in Prmt2^+/− and Prmt2^−/− (Fig. 2C). Consequently, loss of function of one copy of Prmt2 is sufficient to deeply affect the respiratory and inflammatory response toward LPS, highlighting a potential dual role for Prmt2. Prmt2 might certainly contribute to the early LPS-induced responses but also mostly contributes later to inhibit the TLR4/NF-κB pathway.

To better understand the role of Prmt2 in the NF-κB pathway, we decided to study the impact of the partial and complete deletion of Prmt2 on the expression of TNF-α and IL-6.

First, the expression of Prmt2 was evaluated by quantitative real-time PCR in total RNA isolated from lungs of B6, Ms1Yah, Prmt2^+/−, and Prmt2^−/− mice 90 min after LPS or NaCl administration. B6 mice stimulated with LPS were taken as the reference condition, and data were normalized using the Genorm method (25). Prmt2 is expressed with a basal level in control mice and is further downregulated after LPS stimulation, as already shown (17). In absence of LPS instillation, the expression of Prmt2 is reduced approximately by a factor 0.5 in Ms1Yah, strongly reduced in Prmt2^+/−, and close to zero in Prmt2^−/− (Fig. 3A). Prmt2 appeared to be a dosage-sensitive gene, and the mutant Prmt2^−/−, used in the study, completely abolished its expression. Interestingly, the expression of Prmt2 is strongly decreased in Ms1Yah, Prmt2^+/−, and Prmt2^−/− individuals after LPS instillation. We conclude that the LPS pathway acts directly on Prmt2 transcription. Such an LPS-induced Prmt2 gene downregulation is in agreement with the hypothesis that Prmt2 acts as an inhibitor of the NF-κB pathway in the lung.

To check the impact of Prmt2 on the NF-κB pathway, we studied the pattern of expression of TNF-α and IL-6, which are direct targets of this pathway, in the lungs from B6, Prmt2^+/−, and Prmt2^−/− mice. The saline control solution did not affect the expression of TNF-α or IL-6. The LPS induced an increase in the expression of IL-6 and TNF-α in the lungs of Prmt2^+/− and Prmt2^−/− mice in a manner that is inversely correlated to the number of Prmt2 copies present in the mice (Fig. 3B). Altogether, these results support that Prmt2 acts on the expression of IL-6 and TNF-α in a dose-dependent manner, certainly by interfering with NF-κB activity.

TNF-α and IL-6 are primarily produced from the activated MΦ in response to LPS. Therefore, we next asked whether the loss-of-function of Prmt2 in MΦ could influence directly the secretion of proinflammatory cytokines, such as TNF-α and IL-6, after LPS treatment. MΦ from Prmt2^+/−, Prmt2^−/−, and control mice were compared for their response toward LPS by measuring the production of TNF-α and IL-6 in vitro 24 h after stimulation by LPS or saline buffer (NaCl).

We found that LPS-stimulated MΦ derived from Prmt2^−/− mice produced significantly higher amounts of TNF-α and IL-6 than those derived from control animals (Fig. 4). The level of TNF-α and IL-6 secreted in the culture medium of LPS-stimulated Prmt2^−/− MΦ was equivalent to that found in the medium of Ms1Yah-derived MΦ. Surprisingly, the amount of both cytokines in Prmt2^+/− MΦ was equivalent to that in B6-treated MΦ, suggesting that in Ms1Yah, additional genes should contribute to the Ms1Yah phenotypes.

We looked further at the expression of proinflammatory cytokines in MΦ 24 h after the LPS stimulation. As expected, control treatment (NaCl) did not induce expression of IL-6 and TNF-α genes in MΦ (Fig. 4). On the contrary, LPS induced a significant increase in the expression of the IL-6 and TNF-α genes in the Ms1Yah, Prmt2^+/−, and Prmt2^−/− mice, which is coherent with the increase in the concentration of the cytokines in the supernatant of the MΦ. Nevertheless, clear differences are noted. Indeed, the expression of TNF-α is not affected when one copy of Prmt2 is functional in heterozygotes, whereas similar level of expression is achieved when there is a complete loss of function of Prmt2. On the contrary, the maximum level of expression of IL-6 in LPS-induced MΦ is observed when only one copy of Prmt2 is inactivated. The MΦ derived from Ms1Yah mice showed a stronger increase in the expression of IL6, similar to that observed in Prmt2 complete loss of function, confirming the hypothesis that another gene from the region deleted in Ms1Yah controls TNF-α expression, but not IL-6 (Fig. 4).

We found that the absence of Prmt2 induces an increase in expression and production of TNF-α and IL-6. According to Ganesh et al. (20), we hypothesize that the loss of Prmt2 will lead to an increase of NF-κB accumulated into the nucleus of Prmt2^−/− cells, leading to a stronger and long-lasting NF-κB–dependent transcription. To test this hypothesis, we stimulated MΦ from wt and Prmt2^−/− mice with saline control solution or LPS for 4 h. Immunofluorescence was performed on the cells, revealing an accumulation of NF-κB in nucleus of MΦ from both wt and Prmt2^−/− mice (Fig. 5A). To determine the effect of Prmt2 on the nuclear accumulation of NF-κB, the presence of the transcription factor was evaluated in wt and Prmt2^−/− cells by fluorescence intensity (measured as pixel intensity per square micrometer). LPS induced nuclear accumulation of NF-κB in wt and Prmt2^−/− MΦ, but the quantity of fluorescent signal for the NF-κB transcription factor was more important in Prmt2^−/− cells (Fig. 5B).We thus confirmed that Prmt2 inhibits nuclear export of NF-κB in MΦ as shown previously in fibroblasts (20). The increase of NF-κB in the nucleus of Prmt2^−/− cells enlightens the increase of cytokine expression and production in Prmt2^−/− lungs and MΦ compared with wt mice.

Systemic injection of LPS induced production of proinflammatory cytokines such as TNF-α and IL-6 as observed in the sera of Ms1Yah mice (17). We supposed that the inhibitory role of Prmt2 in the NF-κB signaling could explain this phenotype. To test whether systemic LPS could lead to the same enhanced inflammatory response in Prmt2^+/− or in Prmt2^−/− mice as in Ms1Yah mice, we administered an i.p. injection of 100 μg LPS or saline control solution in 10 mice of each genotype plus B6 mice as control. The proinflammatory response was assayed by measuring the concentration of cytokines and chemokines in the serum 90 min after LPS stimulation. In particular, we selected the TNF-α and the IL-6, which are secreted during the inflammatory response mediated through the innate immune system by the TLR4 and the MyD88/NF-κB signaling pathway (22, 26). The Prmt2^−/− mice had the same increase in the TNF-α concentration as Ms1Yah mice, whereas sera of Prmt2^+/− mice presented the same concentration of TNF-α as B6 treated mice (Fig. 6). The concentration of IL-6 is increased in Ms1Yah mice and in both Prmt2^+/− and Prmt2^−/− compared with control mice (Fig. 6). Haploinsufficiency of Prmt2 is sufficient to induce a change in the production of IL-6, whereas it has no impact on the release level of serum TNF-α, suggesting different levels of control mediated by Prmt2 for the regulation of the LPS response in vivo in different compartments.

We report in this article a new role for the protein Prmt2 in the regulation of pulmonary inflammatory and airway distress syndrome induced by LPS, but also in MΦ and after systemic stimulation.

Aerogenic exposure to LPS induces pulmonary inflammation in mice, characterized by recruitment and activation of MΦ and neutrophils in the airways, local TNF-α production, and direct AHR (27). LPS first makes a complex with the LPS-binding protein and the cell membrane CD14 coreceptor to activate TLR4 (28). Then cell activation can occur via two pathways, both leading to NF-κB activation (29–31). The first one is MyD88 dependent and need both the MyD88 and TIRAP adaptors. MyD88 and TIRAP-deficient mice are both resistant to systemic LPS-induced shock (21, 32). Moreover, LPS-induced bronchoconstriction, neutrophil recruitment, and TNF-α production in the airways are abrogated in both lungs and MΦ of MyD88-and TIRAP-deficient mice (21, 33–35). However, LPS stimulation still activates NF-κB in MyD88- and TIRAP-deficient cells, but with a delayed kinetics compared with wt cells (32), and preserve their ability to induce IFN-inducible genes (36, 37). The MyD88-independent pathway involves the TRIF and TRAM. TRIF- and TRAM-deficient MΦ were impaired in the LPS-induced inflammatory cytokine production but have normal NF-κB activation, suggesting that cooperation between both the MyD88-dependent and -independent pathways is required for the TLR4-mediated inflammatory cytokine production (38, 39). Whereas bronchoconstriction is abrogated in MyD88-deficient mice, TRIF-deficient mice have normal bronchoconstriction in response to LPS exposure. This observation indicates that bronchoconstriction induced by instillation of LPS is strictly dependent on the MyD88-dependent pathway (22, 40). Ms1Yah mice showed the same absence of bronchoconstriction as the MyD88- and TIRAP-deficient mice (17). This observation led us to think that at least one gene of the Prmt2-Col6a1 region is responsible of the bronchoconstriction associated with the MyD88-dependent pathway. As opposed to mutants of the MyD88-dependent signaling, Prmt2^+/− and Prmt2^−/− mice showed a delayed but enhanced AHR maintained in time compared with wt after LPS exposure. Prmt2 behaves as an activator in the early phase of AHR, mainly as a consequence of its own downregulation by LPS, and as an inhibitor during the regulation of the response amplitude and length of the AHR. Recently, the expression of the Prmt1-6 was studied in lungs and spleen from a rat model of asthma (41). Prmt1, Prmt2, Prmt3, and Prmt5 were overexpressed in lungs of all animals, whereas Prmt4 was decreased. In the spleen, only Prmt2 and Prmt5 were overexpressed. This study suggests that PRMTs play an important role in the posttranslational modification process of asthma-related genes that is reinforced by our results demonstrating the involvement of Prmt2 in the AHR induced by LPS. Prmt2 is a new actor of the LPS/TLR4/MyD88-dependent induced bronchoconstriction. Nevertheless, at least another gene of the Prmt2-Col6a1 region is involved in the reduced AHR observed in Ms1Yah mice. Finding such a gene is of interest because it should also have the capacity of altering the increased AHR observed in Prmt2-deficient mice found in this report.

Prmt2 has been shown to inhibit NF-κB–dependent transcription by causing a nuclear accumulation of IκB-α that prevents DNA binding in a dose-dependent manner (20). Even if the exact mechanism is unclear, Prmt2 interacts with IκB-α by its ankyrin domain, which also mediates the interaction with NF-κB. The Prmt2/IκB-α complex still binds NF-κB molecules and prevents them from binding DNA (20). All these observations led us to hypothesize that Prmt2 loss of function can modify the inflammatory signaling mediated by NF-κB activation after LPS treatment. The analysis of the lung inflammation of the Prmt2^+/− and Prmt2^−/− mice after LPS stimulation showed an increase in the number of neutrophils, the MPO activity, and the concentrations of both IL-6 and TNF-α. We then studied the expression of Prmt2, TNF-α, and IL-6 in lungs. In absence of LPS, the basal expression level of Prmt2 is more important in control lung and MΦ. Somehow, in absence of stimulus, the NF-κB factor is sequestered in the cytoplasm and the NF-κB–dependent transcription is “off”. Tam et al. (11) proposed a new model in which the maintenance of the p65/NF-κB subunit in the cytoplasm of unstimulated cells requires a continuous flow of nuclear protein. Accordingly, the NF-κB/IκB-α complexes are formed in the nucleus and then exported to the cytoplasm via the CRM1 chaperone protein. This leads to the accumulation of NF-κB/IκB-α in the cytoplasm, which cannot go back to the nucleus without degradation of IκB-α, which masks the nuclear localization signal of NF-κB. The expression level of Prmt2 in unstimulated conditions could facilitate the formation of NF-κB/IκB-α dimers. Indeed, Prmt2 promotes the nuclear accumulation of IκB-α (20). After LPS stimulation, expression of Prmt2 decreases, allowing the decrease of nuclear IκB-α, and thus the binding of NF-κB on DNA, leading to the expression of the target genes like IL-6 and TNF-α. This decrease is more important in Prmt2 mutants, almost to a null level, and can explain the increased expression level of TNF-α and IL-6 in lungs and MΦ. In parallel, we demonstrated that Prmt2-deficient MΦ showed an increased nuclear accumulation of NF-κB after LPS treatment. Overall, our results stressed the role of Prmt2 in controlling the NF-κB–dependent inflammatory response and particularly the expression and secretion of both IL-6 and TNF-α. Moreover, in lungs, there is a good correlation between the number of copies of Prmt2 and expression of the two cytokines. Taken together, all these observations confirm a role of Prmt2 as an inhibitor in the control of the NF-κB–dependent inflammatory response in lungs, induced by the activation of the TLR4 signaling.

The systemic response observed is different in Prmt2^+/− as expected. In fact, the loss of one copy of Prmt2 is sufficient to modify the systemic concentration of IL-6, but not TNF-α. This observation reinforces the involvement of Prmt2 in the inflammation signaling pathway and suggests the involvement of another gene or a different mechanism to regulate TNF-α production. After LPS stimulation, the activated MΦ is one of the major sources of TNF-α production. In MΦ, the loss of the two copies of Prmt2 is necessary to find the same increase in TNF-α and IL-6 expression and production as in Ms1Yah mice. In fact, expression of the cytokines seems to be regulated by two different mechanisms. One is dependent on Prmt2; indeed, in absence of the two copies of Prmt2, we observed at the same time an increase of the expression and the concentration of both cytokines, showing the same phenotype as the monosomic mice. In Prmt2^+/− MΦ, TNF-α expression is not altered compared with WT, and no change in TNF-α concentration is observed in the supernatant, whereas expression of Il-6 is upregulated in MΦ from Prmt2^+/− mice, suggesting a different regulation between IL-6 and TNF-α expressions. Activation of TLR4 by LPS induces the expression of IL-10 by MΦ to regulate excessive production of inflammatory cytokines (42). IL-10 has been shown to inhibit TNF-α at early (30 min) and late (120 min) time points, and IL-6 at 120 min (43), which is consistent with the identification of these genes as primary and secondary response genes (44). Because of their different kinetic, IL-10 might act at different stages of their induction (42) and could explain the differences observed between TNF-α and IL-6 production by MΦ and during the systemic response. Although the expression of IL-6 is increased in Prmt2^+/− MΦ, the concentration of the IL-6 protein is lower than in WT MΦ, suggesting a posttranscriptional regulation.

Prmt2 clearly interferes with different pathways. Indeed, it interacts with the retinoblastoma gene product, which is an important regulator of the transcription factor E2F (18, 45). Prmt2 can also directly regulate transcription factor activity like the estrogen receptor-α (45), Stat3 (46) and the androgen receptor (47). Moreover, Prmt2 directly modifies the structure of the chromatin, mediating an asymmetric methylation of histone H3(R8) after recruitment by β-catenin on its targeted genes, establishing a poised chromatin architecture necessary and sufficient to regulate expression of key organizer genes (48, 49). Thus, we could hypothesize an another mode of action of Prmt2, through the modification of the chromatin structure, altering the long-term consequence of NF-κB–dependent transcription as suggested, but in opposition to the effect of the H3 lysine 4 methyltransferase, Seth 7/9, on NF-κB and inflammatory gene expression (50).

LPS induces asymmetric dimethylarginine (ADMA) formation in lung and MΦ (51–54). Increased ADMA level leads to both decreased NO production and an increased O₂⁻ formation by uncoupling NO synthase (55, 56). Finally, high level of ADMA induced cell damage, AHR, and protein dysfunction. The type I Prmts (1–4, 6) are directly involved in ADMA formation and localize to bronchial epithelial cells, endothelial cells, and smooth muscle cells. Prmt2 is a type I methyltransferase that participates in ADMA formation (51). Consequently, the loss of one or both copies of Prmt2 should lead to a decrease in ADMA level, and thus an increase in NO level. Accordingly, the concentration of NO was increased in MΦ from Ms1Yah after LPS stimulation (17). We can reasonably suppose that the loss of functional copies of Prmt2 leads to a decrease in ADMA concentration in our Prmt2^+/− and Prmt2^−/− models versus control mice. This induces an increase in NO with a reduced or no ADMA response, which can explain the enhanced AHR observed in Prmt2^+/− and Prmt2^−/− mice (56).

In this report, we demonstrated that Prmt2 plays a direct role in inflammation induced by LPS certainly through the control of the NF-κB pathway. Its sensitivity to gene dosage depends on the compartments but is clearly observed in lungs and in MΦ. The staining of NF-κB in wt and Prmt2^−/− MΦ (Fig. 5) clearly showed that Prmt2 allows a nuclear accumulation of NF-κB, supporting the NF-κB–dependent transcription. Taking into account our data and those previously described (11, 20) led us to propose the following model for Prmt2 controlling NF-κB signaling. In unstimulated cells (Fig. 7), there is a constitutive production of IκB-α waiting for new synthesis of the NF-κB subunit p65. Prmt2 controls the nuclear accumulation of IκB-α. IκB-α/p65 complexes are formed and are actively exported to the nucleus through the CRM1 chaperon protein. After LPS treatment, Prmt2 is downregulated and the phosphorylated IκB-α is degraded, which allows the translocation of NF-κB to the nucleus, where it binds the promoters of its target genes. This leads to the inflammatory response. IκB-α is newly synthesized and goes to the nucleus to enter in competition with DNA to bind NF-κB. In presence of LPS, we demonstrated that the expression of Prmt2 is decreased, but the produced quantity is sufficient to participate in the regulation of the NF-κB transcription, allowing a return to physiological conditions. We showed that in absence of Prmt2, the inflammatory response is more important, supporting the fact that Prmt2 facilitates the formation of the IκB-α/NF-κB complex. In lungs, Prmt2 exerts its effects in a dose-dependent manner. Prmt2 is also involved in the systemic and MΦ responses, but there are other mechanisms such as posttranscriptional and/or translational modifications.

In conclusion, our results clearly demonstrate that Prmt2 is involved in the regulation of the inflammatory response induced by LPS, certainly via its inhibitory role of the NF-κB–dependent transcription. Given the importance of inflammation in a great number of lung diseases, Prmt2 appeared as a new potential therapeutic target to modulate the inflammatory response. Furthermore, it reinforces the role of the histone methyltransferase in contributing to the regulation of NF-κB–dependent transcription (50).

We thank members of the research group, the Institut de Génétique Biologie Moléculaire et Cellulaire laboratory, the Institut Clinique de la Souris, and the AnEUploidy consortium for helpful comments (http://www.aneuploidy.org). We are grateful to the animal caretakers of the Centre National de la Recherche Scientifique UPS44 Transgénèse et Archivage d'Animaux Modèles unit and the Institut Clinique de la Souris. We also thank Pamela Gasse and Hughes Jacobs for useful discussion.

The authors have no financial conflicts of interest.