Deficient arginase II expression without alteration in arginase I expression attenuated experimental autoimmune encephalomyelitis in mice

Summary

In the past there have been a multitude of studies that ardently support the role of arginase II (Arg II) in vascular and endothelial disorders; however, the regulation and function of Arg II in autoimmune diseases has thus far remained unclear. Here we report that a global Arg II null mutation in mice suppressed experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis. During EAE, both Arg I and Arg II were induced in spinal cords, but only Arg II was induced in spleens and splenic dendritic cells (DCs). DC activation by lipopolysaccharide (LPS), CD40L or TLR8 agonist significantly enhanced Arg II expression without affecting Arg I expression. Conversely, DC differentiating cytokines [IL‐4 and granulocyte macrophage‐colony‐stimulating factor (GM‐CSF)] yielded opposite effects. In addition, Arg I and Arg II were regulated differentially during Th1 and Th17 cell polarization. Arg II deficiency in mice delayed EAE onset, ameliorated clinical symptoms and reduced myelin loss, accompanied by a remarkable reduction in the EAE‐induced spinal cord expression of Th17 cell markers (IL‐17 and ROR γt). The abundance of Th17 cells and IL‐23⁺ cells in relevant draining lymph nodes was significantly reduced in Arg II knockout mice. In activated DCs, Arg II deficiency significantly suppressed the expression of Th17‐differentiating cytokines IL‐23 and IL‐6. Interestingly, Arg II deficiency did not lead to any compensatory increase in Arg I expression in vivo and in vitro. In conclusion, Arg II was identified as a factor promoting EAE likely via an Arg I‐independent mechanism. Arg II may promote EAE by enhancing DC production of Th17‐differentiating cytokines. Specific inhibition of Arg II could be a potential therapy for multiple sclerosis.

Introduction

Vertebrates are known to express two isoforms of arginases, arginase I (Arg I) and Arg II, both of which catalyse the same reaction that converts l‐arginine into urea and ornithine. However, they are encoded by distinct genes, and differ in cellular expression, regulation, subcellular localization and possibly physiological function.1, 2, 3 Arg I is cytoplasmic and highly expressed in the liver. Arg II, on the other hand, is found in mitochondria and expressed in virtually all mitochondria‐containing extrahepatic cells, with the highest level of expression being in the kidneys.1 The main classic function of arginases is to eliminate ammonia through the urea cycle in the liver. Additionally, arginases also exert other cellular functions via downstream products of the initial reaction, such as polyamines and proline.

Besides their classic functions described above, rapidly emerging evidence suggests arginases as key players in the mammalian immune system.2, 4, 5 In this regard, arginase has been described as an old enzyme with new tricks.4 The role of arginases in the pathogenesis of various animal disease models has been investigated by using both genetic and inhibitor approaches, reviewed in.2, 4, 5, 6 Arg I global knockout (KO) mice die within 10–12 days of age from hyperammonaemia.7 In humans, patients with Arg I deficiency are seemingly normal at birth but gradually develop the symptoms of brain damage, manifested by spasticity, loss of cognitive function and growth retardation.8 In contrast to Arg I KO mice, homozygous Arg II KO mice are apparently indistinguishable from wild‐type (WT) mice, except for an elevated plasma arginine level.9 This dramatically different outcome of Arg I and Arg II global null mutation is a clear indication that the functions of Arg I and Arg II may not entirely overlap, and may potentially be dependent on their tissue and cellular distribution. Compared with Arg I, studies on the function of Arg II are lagging but are rapidly emerging.10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 These studies have primarily been focused on the role of Arg II in the pathogenesis of diseases that are related to endothelial dysfunction.12, 16, 17, 19, 20 There is a general consensus that Arg II acts to impair endothelial function via reducing free radical production in endothelial cells, and promotes the polarization of macrophages towards the proinflammatory M1 macrophage phenotype.

The currently available information on the regulation and function of Arg II in autoimmune diseases is limited. To our knowledge, only one study has thus far been published on the functional role of arginases in autoimmune diseases using the mouse experimental autoimmune encephalomyelitis (EAE) model.21 In this study, mice treated with a non‐selective arginase inhibitor ABH [2(S)‐amino‐6‐boronohexanoic acid] showed a significant delay in EAE onset and ameliorated disease symptoms thereafter.21 Although this study clearly demonstrates a destructive role of Arg I and Arg II in central nervous system (CNS) autoimmunity, the respective contributions of Arg I and Arg II in EAE remain unknown because the non‐selective inhibitor abrogates the activity of both isozymes. In this regard, knowledge on the specific role of these two arginases in EAE is particularly important, as there is convincing evidence that in macrophages Arg I and Arg II are regulated differentially and function in the opposite manner.^{2,10,13,14,22–25}

Because arginase inhibitors developed thus far inhibit both Arg I and Arg II, it is important to use mouse models with a targeted knockout of Arg I and Arg II to elucidate if the two arginases play a similar or distinct role in EAE pathogenesis. Because Arg I KO mice die within 10–12 days after birth,7 it makes it impossible to use this model to evaluate the role of Arg I in EAE. We chose to focus on studying the role of Arg II in EAE because Arg II KO mice are phenotypically normal9 and commercially available from the Jackson Laboratory.

The results from this study demonstrate that Arg I and Arg II are differentially regulated both in vivo and in vitro in response to inflammatory stimuli. Importantly, a global Arg II deficiency in mice significantly delayed EAE onset, ameliorated disease symptoms and improved CNS histopathology. To our knowledge, we are the first to report on the function of a specific arginase isoform in any autoimmune diseases.

Materials and methods

Mice and EAE induction

Arg2^−/−(Arg2 ^tm1Weo /J) mice in C57BL/6J genetic background and control mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Animals were housed in pathogen‐free conditions. To prevent mortality, paralysed mice bearing severe EAE were given easy access to food and water by providing soft food at the bottom of the cages. No animal death occurred under our EAE induction protocol. All experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee at the J.L Pettis Memorial VA Medical Center.

Experimental autoimmune encephalomyelitis was induced by immunization with myelin oligodendrocyte glycoprotein peptide (MOG_35–55) in 9‐week‐old female mice per instructions provided by Hooke Laboratories (Lawrence, MA). Unless indicated otherwise, the mice were subcutaneously injected in two locations on their backside with a total of 0·2 ml of the antigen emulsion (0·1 ml/location), which contains 200 μg MOG_35–55 and 0·5 mg Mycobacterium tuberculosis H37Ra emulsified in complete Freund's adjuvant. The mice were then injected twice intraperitoneally with 200 ng of pertussis toxin (provided in the kits) in 0·1 ml water at 4 hr and at 48 hr post‐antigen emulsion injection.

EAE clinical symptom evaluation and motor function test

The mice were monitored for clinical symptoms of disease, and the disease severity was then scored on a numerical scale from 0 to 4 according to the following criteria: 0, no disease; 1, weak tail or wobbly walk; 2, hind limb paresis; 3, hind limb paralysis; 4, hind and forelimb paralysis.26 Motor function was determined by the hanging wire grip test.27 The mice were briefly placed on top of a wire cage lid that was shaken gently three times, causing them to grip the wire. The wire cage lid was then inverted at a height of 20 cm above the cage floor to prevent the animal from easily climbing down. Latency to fall was recorded three times.

Histological analysis

Spinal cords were pushed out from the spine using a 10‐ml syringe filled with phosphate‐buffered saline (PBS). The proximal spinal cords (5 mm from the hindbrain) were cut, fixed with 10% formalin overnight, and embedded in paraffin. Eight‐micrometre‐thick sections were stained with H&E (Sigma‐Aldrich, St Louis, MO) to evaluate spinal cord lesions. The lesion area displaying aggregated nuclei and the total spinal cord section area were measured using the Image‐Pro Plus image processing software. The relative lesion area was then calculated by (total lesion area/total spinal cord section area) × 100.

Demyelination in the lesions was evaluated by Luxol fast blue staining and myelin basic protein (MBP) immunostaining, respectively. The tissue sections were deparaffinized with xylene, and cleared in 100% alcohol followed by hydration in 95% alcohol. Sections were stained in 0·1% Luxol fast blue (AC212170250, Acros Organics, Fisher Scientific) at 60° for 2 hr, followed by rinses with 95% ethanol and distilled water to remove the excess blue stain. The sections were differentiated by using 0·05% lithium carbonate (193360100, Acros Organics, Fisher Scientific); differentiation was continued in 70% alcohol followed by the rinse and wash in distilled water. Sections were then counterstained with haematoxylin (MHS32 Sigma Aldrich) and eosin (E4009, Sigma Aldrich).

For MBP immunostaining, spinal cord sections were baked at 60° for 1 hr, followed by deparaffinization in xylene, 100%, 90% and 70% ethanol, and then rehydration in water. The sections were treated with BLOXALL endogenous peroxidase and alkaline phosphatase blocking solution (Vector Lab, Cat# SP‐6000, Vector Laboratories, Inc., Burlingame, CA) followed by washing with PBS. Sections were treated with 1% bovine serum albumin (BSA) to prevent non‐specific staining. The sections were incubated with mouse anti‐MBP primary antibodies (1 : 200; Monoclonal Mouse IgG1 Clone # 932908; MAB42282‐SP, R&D systems) overnight at 4°. After staining with primary antibody, the sections were washed three times with PBS and incubated in biotin‐conjugated anti‐mouse IgG (Vector Labs, Cat# MB‐9100, Vector Laboratories, Inc., Burlingame, CA) for 30 min at room temperature, followed by the wash in PBS. The sections were incubated in streptavidin‐horseradish peroxidase (HRP; Vector Lab, Cat# SA‐5004, Vector Laboratories, Inc., Burlingame, CA), 1 : 250 dilutions in 0·1% BSA in the dark. After 30 min, sections were washed three times in PBS (10 min/wash), incubated in DAB solution (DAB Peroxidase Substrate Kit, SK‐4100 Vector Labs) until staining was evident microscopically. The tissue sections were counterstained with haematoxylin and mounted with clear hard mounting solution CYTOSEAL 60 (ThermoFisher Scientific, Waltham, MA). Staining with DAB results in the deposition of a brown insoluble precipitate, and was visualized using the light microscope. Immunostaining negative controls were prepared without the primary antibody.

Dendritic cell (DC) isolation, purification and treatment

The femur and tibia were cleaned to remove muscle tissues, fibres and ligaments, and then crushed to release bone marrow. Collected cells were then spun down and red blood cells lysed. The remaining marrow cells were then cultured in IMDM (Gibco^™/ThermoFisher Scientific) with 10% fetal bovine serum, 0·1% 2‐mercaptoethanol, 50 ng/ml mouse granulocyte macrophage‐colony‐stimulating factor (GM‐CSF) and 40 ng/ml mouse IL‐4. Forty‐eight hours later, non‐adherent cells were removed and adherent cells were cultured for an additional 5 days. Immature CD11c⁺ DCs in detached cells were purified with CD11c MACS microbeads (Miltenyi Biotech). Purified DCs were seeded in 12‐well plates (5 × 10⁵ cells/well) and incubated with various factors indicated in the figure legends. In ex vivo studies, splenic DCs were purified with CD11c MACS beads from single‐cell suspensions prepared from the spleens of control or EAE mice.

T‐cell isolation and polarization assay

Isolation and polarization of the naive T‐cells were performed as previously described.27 Naïve T‐cells were purified from splenic single‐cell suspension using naïve T‐cell isolation kits per the manufacturer's instruction (Miltenyi Biotec, Auburn, CA). Polarization of Th1 and Th17 cells was induced using a commercial kit as per the manufacturer's instruction (Miltenyi Biotech, Th1 polarization kit: Cat# 130‐107‐758; Th17 polarization kit: 130‐107‐758). Cells were harvested 4 days later for analysis of arginase and T‐cell marker gene expression by real‐time polymerase chain reaction (PCR) or flow cytometry.

FACS analysis of polarized T‐cells and lymphocytes

Cervical lymph nodes (LNs), inguinal LNs and brachial LNs were isolated and pooled to obtain an adequate number of lymphocytes for FACS analysis. These LNs were used in particular due to the fact that they are either CNS‐draining LNs or LNs draining the MOG‐immunization sites on the back. Pooled LNs were gently pressed through cell strainers to produce single‐cell suspensions. After lysing red blood cells, the mononucleated cells were cultured in the presence of cell stimulation cocktail (Cat#00‐4970‐03, eBioscience/ThermoFisher Scientific), and protein transporter inhibitor (Cat# 00‐4980‐03, eBioscience/ThermoFisher Scientific) was added 30 min later. After 4 hr of incubation, cells were stained with viable dye and then antibodies that recognize various cell surface markers (eBioscience, San Diego, CA) indicated in figure legends. For intracellular cytokine staining, cells were fixed and permeabilized in Perm‐Fix solution (eBioscience) for 1 hr at 4°, washed twice in Perm‐Wash buffer (eBioscience) and stained with the appropriate monoclonal antibody for 2 hr. Isotype primary conjugated antibodies served as a negative control. All analyses were performed following gating of viable cells.

Analysis of tissue Arg I and Arg II by ELISA

Spleens collected from control or EAE mice were weighed and homogenized in eight volumes (V/W) of native RIPA buffer. Additional non‐ionic detergent (0·2% Chaps) was added to the RIPA buffer to ensure complete release of Arg II from mitochondria. Protease inhibitor cocktail was added to RIPA buffer to prevent proteolytic degradation of intracellular arginases. After centrifugation at 12 000 g for 15 min, the supernatant was collected and protein concentrations were measured using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific). Concentrations of Arg I and Arg II in the supernatants (pg/μg cellular protein) were measured using ELISA kits in accordance to the manufacturer's instruction (LifeSpan Biosciences, Seattle, WA). According to the manufacturer, the Arg I and Arg II ELISA do not have cross‐reactivity between Arg I and Arg II. The specificity of the Arg II ELISA was further verified by showing that the absorbance measured in Arg II KO spleen homogenates was similar to that measured in the sample buffer alone.

Real time (RT)‐quantitative (q)PCR analysis

RNA extraction, cDNA synthesis, and RT‐qPCR (RT‐qPCR) were completed as described previously.28 Specific primers for each gene‐of ‐interest are given in Table 1. RT‐qPCR analyses were carried out using Sybr Green RT PCR master mix (ThermoFisher Scientific) in accordance with the manufacturer's instructions. Relative expression was normalized to the ubiquitously expressed GAPDH and calculated using the ΔΔCT method.

Table 1.

Sequences of primes used in this study

	Forward	Reverse
GAPDH	TGGCAAAGTGGAGATTGTTGCC	AAGATGGTGATGGGCTTCCCG
CD11c	CTGGATAGCCTTTCTTCTGCTG	GCACACTGTGTCCGAACTCA
CD86	TTGTGTGTGTTCTGGAAACGGAG	AACTTAGAGGCTGTGTTGCTGGG
IFNγ	GATGCATTCATGAGTATTGCCAAGT	GTGGACCACTCGGATGAGCTC
IL‐1β	TTGACGGACCCCAAAAGAT	GAAGCTGGATGCTCTCATCAG
IL‐4	GGTCTCAACCCCCAGCTAGT	GCCGATGATCTCTCTCAAGTGAT
IL‐5	TGCCTGGAGCAGCTGGAT	GTGGCTGGCTCTCATTCACA
IL‐6	AGGATACCACTCCCAACAGAC	CAAGTGCATCATCGTTGTTCA
IL‐12	AAATGAAGCTCTGCATCCTGC	TCACCCTGTTGATGGTCACG
IL‐17a	CTCCAGAAGGCCCTCAGACTAC	GGGTCTTCATTGCGGTGG
IL‐23	TGGCATCGAGAAACTGTGAGA	TCAGTTCGTATTGGTAGTCCTGTTA
MHC Class II	GAGGCTCAACTTGTCCCAAAAC	GCAGCCGTGAACTTGTTGAAC
Arg I	TGGCTTGCGAGACGTAGAC	GCTCAGGTGAATCGGCCTTTT
Arg II	ACCAGGAACTGGCTGAAGTG	TGAGCATCAACCCAGATGAC
TNFα	GAACTCCAGGCGGTGCCTAT	TCGGCTGGCACCACTAGTTG
RORyt	ACCTCCACTGCCAGCTGTGTGCTGTC	TCATTTCTGCACTTCTGCATGTAGACTGTCCC
IL‐10	TATGCTGCCTGCTCTTACTG	CTCCACTGCCTTGCTCTTAT
CD11b	GGATCATAGGCGCCCACTT	TCCTTACCCCCACTCAGAGACT
F4/80	CTCTGTGGTCCCACCTTCAT	GATGGCCAAGGATCTGAAAA

Statistical analysis

Results are expressed as mean ± SEM, and statistically analysed by Student's t‐test or one‐way or two‐way anova followed by the Tukey post hoc test. A value of P < 0·05 was considered statistically significant.

Results

Regulation of Arg I and Arg II differs in CNS versus peripheral immune tissue during EAE

In our efforts to identify novel genes involved in the pathogenesis of EAE, we performed genome‐wide microarray analysis of gene expression in the spinal cord of EAE mice versus healthy control mice. We found that both Arg I and Arg II were significantly induced by EAE, although the magnitude of induction of Arg II was lower compared with that of Arg I (data not shown). This result was confirmed by RT‐PCR analysis of Arg I and Arg II mRNA levels in the spinal cord (Fig. 1a). We also analysed the expression levels of the two isoforms in several other tissues. In muscles, kidneys and intestines, mRNA levels of either isoforms were similar between control and EAE mice (data not shown). To our surprise, we found that Arg II but not Arg I was significantly upregulated in the peripheral immune organ spleen (Fig. 1b), despite the similarity in the inductions of pro‐inflammatory cytokines (i.e. IL‐1β, IL‐6, IL‐17 and IFNγ) in both spinal cord and spleen (data not shown). Consistent with this result, the concentration of Arg I, but not Arg II, in spleen extracts measured by ELISA was significantly increased in the EAE mice compared with control mice (Fig. 1c). The concentration of Arg II in Arg II KO mouse spleens was undetectable, confirming the specificity of the Arg II ELISA (data not shown).

Figure 1.

Expression of both Arg I and Arg II was significantly induced in spinal cords, whereas only Arg II expression was elevated in spleen or splenic dendritic cells (DCs) during experimental autoimmune encephalomyelitis (EAE). In (a) and (b), EAE in 9‐week‐old female C57BL/6J mice was induced by immunization with myelin oligodendrocyte glycoprotein peptide (MOG) _35–55 using an EAE induction kits (Materials and methods). On day 14 post‐MOG immunization, mice were killed, and spinal cords and spleen were collected for total RNA isolation. The mRNA level of the Arg I and Arg II in spinal cords (a) and spleens (b) was determined by quantitative polymerase chain reacyion (qPCR) and adjusted for GAPDH expression. In (c) and (d), spleens were collected from control or EAE mice in an independent experiment (day 13 post‐EAE induction). One‐third of each spleen was used to extract cellular proteins for quantitation of Arg I and Arg II by ELISA (c). The remaining two‐thirds of each spleen was used to purify CD11c⁺ for analysis of Arg I and Arg II mRNA levels (d). Mean ± SEM (n = 4–5). ***P < 0·001 (EAE versus healthy control).

It has been reported that activation of macrophages by lipopolysaccharide (LPS; a functional component of inactivated bacterial included to activate DCs in EAE) induced expression of Arg II but not Arg I.10 Because DCs and macrophages are derived from the same precursors and both cell types possess antigen‐presenting function, we determined if DCs in the spleens contributed to the differential regulation of Arg I and Arg II in this tissue. As shown in Fig. 1d, expression of Arg II but not Arg I in purified CD11c⁺ DCs from spleens was remarkably induced by EAE.

Expressions of Arg I and Arg II are differentially regulated in DCs and T‐cells

Because our ex vivo studies clearly demonstrated that expression of Arg I and Arg II in splenic DCs were differentially regulated during EAE, we further examined if the expression of the two isoforms responded differently upon treatment of DCs with various immunoregulators. CD11c MACS bead‐purified bone‐marrow‐cell‐derived DCs received one of the following treatments: PBS (cntr), LPS, CD40L, TLR8 agonist R848, IL‐4 or GM‐CSF for 36 hr at indicated concentrations (Fig. 2). Treatment with DC‐activating agents, i.e. LPS, CD40L or R848, significantly increased the expression of Arg II as compared with the controls. However, the expression of Arg I was not significantly affected by these factors (Fig. 2a,b). Conversely, treatment with IL‐4 or GM‐CSF, which induced DC differentiation from myeloid cells,29 significantly increased the expression of Arg I, but had an opposite effect on the expression of Arg II (Fig. 2a,b). These results demonstrate a differential regulation of Arg I and Arg II during DC differentiation and activation.

Figure 2.

Arg I and Arg II expressions were differentially regulated during dendritic cell (DC) differentiation and activation. In (a) and (b), DCs were generated from C57BL/6J mouse bone marrow cells in the presence of IL‐4 and granulocyte macrophage‐colony‐stimulating factor (GM‐CSF) 27 (Materials and methods). On day 5, DCs were purified with CD11c‐MACS beads, seeded in 12‐well plates and treated with vehicle (cntr), lipopolysaccharide (LPS; 0·5 μg/ml), CD40L (1 μg/ml), TLR8 agonist R848 (1 μg/ml), IL‐4 (20 ng/ml) or GM‐CSF (50 ng/ml) for 36 hr. The mRNA level of Arg I (a) and Arg II (b) was determined by quantitative polymerase chain reaction (qPCR) and adjusted for GAPDH expression. In (c) and (d), purified DCs were treated with either vehicle (cntr) or LPS (0·5 μg/ml) or IL‐4 (20 ng/ml) for 36 hr. Data represent the means ± SEM (n = 4). *P < 0·05, **P < 0·01, ***P < 0·001: versus cntr.

Next, we determined if the differential regulation of Arg I and Arg II in DCs was associated with a specific regulation of cytokine expression. To address this question, LPS was selected from the list of the tested DC activators, because it is a functional component of the MOG emulsion, to activate DCs in this EAE model. IL‐4 was also selected because it represents the DC‐differentiating cytokine that showed the strongest induction of Arg I expression in DCs (Fig. 2a). Among a panel of DC cytokines analysed, LPS treatment dramatically increased the expression of the key Th17 cell differentiating cytokines IL‐23 and IL‐6 (Fig. 2c,d), whereas IL‐4 treatment had no effect. Also, treatment with IL‐4 but not LPS moderately increased the Th2 cytokine IL‐4 expression in DCs (Fig. 2c). In contrast, treatment of either IL‐4 or LPS had no significant effect on the expression of the Th1‐differentiating cytokine IL‐12p35 (Fig. 2c).

Lastly, we determined if the expression of Arg I and Arg II differed during T‐cell polarization in the absence of DCs. As expected, naïve T‐cells cultured under Th17 and Th1 conditions showed a dramatic increase in IL‐17 and IFN‐γ expression, respectively (Fig. S1a). Under the Th1 polarization condition, Arg I expression was significantly decreased, whereas Arg II expression remained unchanged compared with that measured in Th0 cells (Fig. S1b). Interestingly, under the Th17 polarization condition, Arg I expression was reduced but Arg II was significantly upregulated. Collectively, our data demonstrate convincingly that the expressions of Arg I and Arg II are differentially regulated during activation and differentiation of both DCs and T‐cells, and that Arg II upregulation appears to be predominantly associated with Th17 differentiation.

Evidence that Arg II plays an imperative role in promoting CNS autoimmune inflammation in EAE

The differential regulation of Arg II versus Arg I by factors that modulate activation or differentiation of DCs and T‐cells suggests that Arg II could play an important role in the pathogenesis of EAE. To test this hypothesis, we performed a series of in vivo experiments to determine if Arg II deficiency in mice could modulate neuroinflammation in the CNS during EAE.

Both WT mice and Arg II KO mice were subjected to EAE induction using an EAE induction kit per the manufacturer's instructions. Out data showed that Arg II KO mice exhibited a significant delay in the onset of disease compared with WT mice (Fig. 3a). Also, disease severity, as indicated by clinical score and motor function test (Fig. 3a, insert), was significantly ameliorated in Arg II KO mice. To evaluate the long‐term impact of Arg II deficiency on EAE, an independent experiment was conducted and clinical score was recorded for a period of 4 months following EAE induction (Fig. 3b). In this experiment, the doses of MOG and mycobacterium tuberculosis for the EAE induction were reduced by 50% to prevent mortality without the necessity to provide soft food after the peak disease phase. At a reduced dosage, no clinical symptoms were observed in both groups at day 22 post‐EAE induction. Although EAE onset was overlooked due to missing observations between day 22 and day 35, anova analysis of the data indicated a highly significant difference in average clinical score between the two groups (P < 0·001; Fig. 3b). For most of the mid‐to‐late time points, the clinical score of the Arg II KO mice was significantly lower than that of the WT mice (Fig. 3b). Towards the end of the experiment, i.e. 4 months post‐EAE induction, the majority of the WT mice were still paralysed with at least one hind limb. However, none of the Arg II KO mice showed even the mildest clinical symptom criteria in this model (tail down).

Figure 3.

Arg II deficiency in mice delayed experimental autoimmune encephalomyelitis (EAE) onset and promoted disease recovery. (a) (Short‐term experiment) Arg II knockout (KO) mice and wild‐type (WT) mice (n = 5) were subjected to EAE induction using a full dosage of antigen emulsion and pertussis toxin per manufacturer's instruction. The data show an average clinical score within the observation period. P < 0·001, WT versus KO (anova test). Insert: motor function was determined by wire hang test as described27 at day 19 when the clinical score peaked in WT mice. ***P < 0·001, WT versus KO group (mean ± SEM). (b) (Long‐term experiment) WT and Arg II KO mice were subjected to EAE induction using 50% of antigen emulsion and pertussis toxin suggested by the manufacturer. *P < 0·05, **P < 0·01, WT versus KO.

To evaluate histopathology, an additional independent short‐term experiment was conducted. At day 18 post‐EAE induction when clinical score in WT mice reached the maximum, spinal cords were collected and examined by H&E staining. As expected, no lesion was observed in the white matter of the spinal cords from WT and Arg II KO mice that were not subjected to EAE induction (Fig. 4a,b). In the spinal cords of WT mice bearing EAE, severe damage in the white matter was obvious, characterized as clusters of lesions with aggregated nuclei of inflammatory cells (Fig. 4c, circled area). Although such damage was also present in the spinal cord of Arg II KO mice bearing EAE, its extent was largely attenuated (Fig. 4d). The severity of the tissue damages, quantitated as the percentage of lesion areas in the total cross‐spinal cord section area, was significantly reduced from 17·4% in the WT EAE mice to 3·5% in the Arg II KO mice (Fig. 4e). To confirm the difference in the white matter damage between the two groups, the degree of myelin loss in the lesions was evaluated by both MBP immunostaining (Fig. 5a) and Luxol Fast Blue staining (the dye binds specifically to the myelin sheath; Fig. 5b). Both MBP and LFB stains revealed that the loss of myelin in white matter lesions was reduced in the EAE‐bearing Arg II KO mice compared with EAE‐bearing WT mice.

Figure 4.

Arg II deficiency in mice reduced lesion areas in the white matter of spinal cords. Wild‐type (WT; six mice) and Arg II knockout (KO; six mice) were subjected to experimental autoimmune encephalomyelitis (EAE) induction as described in Materials and methods. Non‐immunized WT (n = 3) and Arg II KO (n = 3) mice were also included as controls. (a–d) Sections of spinal cords collected at day 18 post‐myelin oligodendrocyte glycoprotein peptide (MOG) immunization were subjected to H&E staining. Lesion areas displaying aggregated nuclei of inflammatory cells were circled. (e) The percentages of lesion areas in the analysed total spinal cord section were determined as described in Materials and methods. Data represent mean ± SEM (n = 6), *P < 0·05: KO EAE versus WT EAE group.

Figure 5.

Arg II deficiency in mice reduced the myelin loss in the white matter. Spinal cord sections of wild‐type (WT) and Arg II knockout (KO) mice with or without experimental autoimmune encephalomyelitis (EAE) were prepared from the experiment described in Fig. 4 and subjected to myelin basic protein (MBP) immunostaining (a) and Luxol Fast Blue (LFB) staining (b). An area of typical myelin loss was indicated by arrows. Representative images of three mice per group are shown.

Arginase II deficiency remarkably reduced the EAE‐induced increase in Th17 cell marker gene expression in the spinal cords, and decreased the abundance of CD3⁺CD4⁺IL‐17⁺ Th17 cells and IL‐23⁺ cells in the draining LNs

To begin to understand the cellular mechanism by which Arg II deficiency attenuates EAE, we analysed the expression of the marker genes for various types of immune cells in the spinal cords. At sub‐peak disease stage (day 13 post‐EAE induction), all the five WT mice immunized with MOG developed EAE with an average score of 2·43. In contrast, only one of the five KO mice developed mild EAE (tail‐down, score of 1). Because expression of all the genes analysed did not differ significantly between non‐immunized WT and non‐immunized Arg II KO mice, data are presented as the fold increase of the mRNA levels measured in non‐immunized WT mice. Among various T‐cell cytokines (Fig. 6a), the expression of the Th17 cytokine IL‐17 and Th1 cytokine IFN‐Ƴ in the spinal cord was greatly induced by EAE. The EAE induction of Th1 cytokine IFN‐Ƴ in the spinal cords was not suppressed in Arg II KO mice. In contrast, greater than 90% of the EAE‐induced Th17 cytokine IL‐17 expression measured in the WT EAE mice was blocked in the Arg II KO mice. Consistent with these data, Arg II deficiency essentially blocked the EAE‐induced expression of RORγt, which is a transcription factor indispensable for Th17 cell differentiation and is also a Th17 cell marker.30 It is important to note that Arg II suppression of IL‐17 expression in the spinal cord was sustained during the peak and recovery phase of EAE (Fig. S2) without significantly affecting the expression of other T‐cell cytokines (data not shown). Besides T‐cell makers, we also analysed the spinal cord expression of markers for other immune cells (Fig. 6b). At the sub‐peak disease stage, the EAE‐induced increase in the expression of monocytes markers (CD11b), macrophages (F4/80), DC markers CD86 and MHC II were not significantly (P > 0·05) altered in Arg II KO mice.

Figure 6.

Arg II deficiency significantly and specifically reduced the experimental autoimmune encephalomyelitis (EAE)‐induced increase in spinal cord expression of the Th17 cell markers IL‐17 and ROR‐Ƴt. Wild‐type (WT) mice or Arg II knockout (KO) mice (n = 5) were subjected to myelin oligodendrocyte glycoprotein peptide (MOG) immunization at a full MOG dosage recommended by the manufacturer. Non‐immunized WT and Arg II KO mice (n = 4) were included as controls. At sub‐peak disease stage (day 13 post‐EAE induction), spinal cords were isolated for analysis of gene expression by quantitative polymerase chain reaction (qPCR). Because expression of all the genes analysed did not differ significantly between non‐immunized WT and Arg II KO control mice, data are expressed as the fold increase of the mRNA levels measured in non‐immunized WT mice. (a) mRNA level of cytokines or marker genes for various T‐cell subtypes. (b) mRNA levels of marker genes for monocytes, macrophages and dendritic cells (DCs). Mean ± SEM, *P < 0·05, **P < 0·01, ***P < 0·001, WT EAE versus KO EAE.

Thus far, our data on the analysis of spinal cord gene expression indicate that Th17 cells are in fact the major immune cells that are greatly suppressed in Arg II KO mice during EAE. It is known that the development of pathogenic T‐cells occurs mainly in the LNs where activated immunogen‐specific DCs are recruited to the T‐cell area of the draining LNs to activate and guide T‐cell differentiation.31, 32 Therefore, we analysed by FACS if the abundance of Th17 cells and the IL‐23⁺ cells that promote Th17 cell differentiation was reduced in the relevant draining LNs. Consistent with reduced expression of Th17 markers (IL‐17 and RORγt) in the spinal cords, the abundance of CD3⁺CD4⁺IL‐17⁺ Th17 cells in the LNs was significantly reduced in Arg II KO mice compared with WT mice (Fig. 7a,b). Importantly, the reduced Th17 cell abundance is accompanied with reduced abundance of the cells expressing each subunit of IL‐23 (p19 and p40; Fig. 7c,d).

Figure 7.

Arg II deficiency significantly reduced the abundance of Th17 cells and IL‐23⁺ cells in the draining lymph nodes (LNs). Wild‐type (WT) and Arg II knockout (KO) mice (n = 5) were subjected to myelin oligodendrocyte glycoprotein peptide (MOG) immunization (same experiment as described in Fig. 6). Mononucleated cells isolated from draining LNs of WT experimental autoimmune encephalomyelitis (EAE) mice and Arg II KO EAE mice were subjected to various cell surface marker or intracellular cytokine staining as indicated. (a). Representative Th17 FACS histograms were shown here. (b) Statistical analysis of the gated viable and CD3⁺/CD4⁺ IL‐17⁺ lymphocytes (three WT EAE mice and three Arg II KO mice). Mean ± SEM, *P < 0·05 (WT EAE versus Arg II KO EAE). (c and d) Lymphocytes that were pooled from two WT EAE mice or two Arg II KO mice were subjected to FACS analysis of IL‐23p19 (c) or IL‐23/IL‐12p40 (d).

Arg II deficiency impaired the expression of the key Th17‐differentiating cytokines in DCs

Data on the analysis of Th17 markers in both spinal cord and LNs indicate that Arg II deficiency compromised Th17 development. To begin to determine the cause of this effect, we first determined if Arg II deficiency in T‐cells compromised the commitment of naïve T‐cells to Th17 cells. To address this question, naïve T‐cells isolated from WT and Arg II KO mice were treated with or without cytokines that induce Th17 differentiation. Our data showed that the expression level of Th17 cytokine IL‐17 under the Th17 polarization condition was not affected by Arg II deficiency (Fig. S3a). Consistent with these data, FACS analysis revealed that the percentages of CD3⁺ CD4⁺ IL‐17⁺ Th17 cells were similar in WT and Arg II KO naïve T‐cells cultured under Th17 polarization condition (Fig. S3b). In addition, polarization of Th1 cells (CD3⁺ CD4⁺ IFN‐Ƴ⁺), as analysed by FACS, was also not compromised by Arg II deficiency (Fig. S3c).

We then determined if Arg II deficiency may have suppressed Th17 cell development via alteration of the Th17‐differentiating cytokine expression in activated DCs. Because the expression of all the genes analysed did not differ in non‐activated WT versus non‐activated Arg II KO DCs, data are expressed as the fold increase of the mRNA levels measured in non‐activated WT DCs. As shown in Fig. 8a, the LPS‐induced expression of DC activation marker CD86 was significantly reduced in Arg II KO DCs compared with WT DCs. The LPS‐induced expression of IL‐23 and IL‐6, both of which are required for Th17 cell differentiation, was significantly suppressed in Arg II KO DCs (Fig. 8b,c). The mRNA levels of other T‐cell‐differentiating cytokines analysed were similar between activated WT and Arg II KO DCs (Fig. 8d–f). These data suggest that while Arg II deficiency suppressed general DC activation, it specifically inhibited the expression of the key Th17‐differentiaing cytokines.

Figure 8.

Arg II deficiency inhibited lipopolysaccharide (LPS) activation of dendritic cells (DCs) and specifically suppressed LPS‐induced DC expression of the Th17‐differetiating cytokines. DCs derived from wild‐type (WT) and Arg II knockout (KO) mice were treated with or without 0·5 μg/ml LPS for 24 hr. mRNA levels of the indicated genes (a–f) were determined by quantitative polymerase chain reaction (qPCR) and adjusted for GAPDH expression. Because the mRNA levels of all analysed genes did not differ significantly between untreated WT DCs and untreated Arg II KO DCs, data were presented as fold induction of the WT untreated DCs. Mean ± SEM (n = 4). *P < 0·05, **P < 0·01, LPS‐treated WT DCs versus LPS‐treated Arg II KO DCs.

Arg II deficiency did not cause a compensatory increase in Arg I expression both in vivo and in vitro

Our data on the differential regulation of Arg I and Arg II suggest that the benefit of Arg II deficiency on EAE is unlikely to be associated with Arg I. To obtain further evidence to support this notion, we determined if Arg II deficiency resulted in a compensatory in Arg I expression both in vivo and in vitro. We found that the mRNA levels of Arg I in both CNS (Fig. S4a) and immune periphery spleens (Fig. S4b) were similar in Arg II KO mice and WT mice bearing EAE. To further evaluate the effect of Arg II deficiency on Arg I expression, we generated DCs from bone marrow of WT and Arg II KO mice. Regardless of the DC activation, there was no significant difference in the mRNA levels of Arg I in DCs derived from WT versus Arg II KO mice (Fig. S4c).

Discussion

Since the late 1980s, the immune regulatory role of arginases has been extensively studied.1, 2, 3, 4, 5, 6 However, most information on the expression and function of arginases in the immune system concerns Arg I. Although the importance of Arg II has been appreciated in recent years, a void in the research regarding the regulation and function of Arg II in autoimmune diseases exists. Our findings demonstrate that Arg I and Arg II expression was regulated in an opposite manner in DCs and T‐cells. Importantly, we found that Arg II deficiency alone with the Arg I gene intact in mice effectively suppressed EAE. Our findings demonstrate for the first time that Arg II acts to promote CNS inflammation and demyelination, possibly via increasing DC‐mediated Th17 development in the peripheral immune organs.

Interestingly, our in vivo and in vitro experiments demonstrate a distinct pattern of the expression of Arg I versus Arg II in response to various immunoregulators. While both Arg I and Arg II were upregulated in the spinal cord (Fig. 1a), only Arg II was significantly increased in the spleen, a peripheral immune tissue (Fig. 1b,c). This tissue‐specific expression pattern may reflect the difference in the immune cell profile between CNS tissue and peripheral immune organs at the disease stage analysed during EAE. In this regard, it has been shown that both Arg I‐expressing M2 macrophages and Arg II‐expressing M1 macrophages are recruited to the CNS.33, 34 The increased presence of these two forms of macrophages could account for the increased Arg I and Arg II mRNA levels in the CNS. Besides macrophages, other infiltrating inflammatory cells may also contribute to the increased expression of arginases in the CNS. For example, we have observed that Th17 cells, which are the primary pathogenic T‐cells recruited to CNS during EAE, expressed a much higher level of Arg II once they differentiated from naïve T‐cells (Fig. S1). The upregulation of Arg II but not Arg I in the spleen could be explained by the distinct response of the two genes in splenic DCs upon MOG immunization. There are two lines of evidence to support this notion. First, our ex vivo experiments clearly demonstrated a significant increase in expression of Arg II but not Arg I in the splenic DCs derived from the EAE mice compared with that of control mice (Fig. 1d). Second, our in vitro data demonstrated that factors (LPS, CD40L and TLR8 agonist) that activate DCs significantly increased Arg II but not Arg I expression (Fig. 2a,b).

Our finding that Arg II, but not Arg I, is specifically upregulated during DC activation and Th17 cell differentiation points to a potentially more significant role of Arg II than Arg I in these immune cells. Although Arg II deficiency did not affect tissue or cellular expression of any genes analysed in our studies under non‐challenged conditions, we have obtained several lines of compelling evidence that Arg II is critically involved in the pathogenesis of EAE. Firstly, compared with WT mice, Arg II KO mice showed a delay in the onset of EAE (Fig. 3a). Second, the overall disease severity was significantly attenuated in Arg II KO mice, as indicated by the reduction in clinical score (Fig. 3a,b), percentage of lesion area (Fig. 4c–e) and MBP loss in the white matter of the spinal cord (Fig. 5a,b). Third, our long‐term experimental data showed that Arg II deficiency not only delayed the onset of EAE but also appeared to facilitate disease recovery (Fig. 3b).

Our data indicate that the beneficial effect of Arg II deficiency on EAE seems to be independent of Arg I. It has previously been shown that the expression of Arg II is moderately (twofold increase) elevated in the kidney of Arg I KO mice compared with control mice.7 In our studies, no compensatory increase in the expression of Arg I was observed in either the spleens or spinal cords of EAE‐bearing Arg II KO mice compared with the EAE‐bearing WT mice (Fig. S4a,b). Consistent with these in vivo data, Arg II deficiency had no significant effect on the expression of Arg I in activated DCs (Fig. S4c).

Although the cellular mechanism by which Arg II deficiency protects against EAE needs to be clarified further, the data we obtained thus far suggest that Arg II deficiency may have attenuated the infiltration and/or development of Th17 cells, the most crucial pathogenic T‐cell subtype leading to EAE.35 Analysis of the abundance of the mRNAs of various immune cell markers in the CNS tissue indicated that the Th17 cells are the primary immune cells suppressed by Arg II deficiency during EAE (Fig. 6a). The mild reduction in expression of markers for other immune cells such as macrophages could be secondary to suppression of Th17 cells that are known to recruit these cells to the CNS. The reduced Th17 cell abundance in the CNS tissue could be due to impairment of the ability of Arg II‐deficient Th17 cells to migrate to the CNS and/or impairment of Th17 development in the peripheral immune tissues. While the former possibility has not yet been examined, the latter is likely to be the case. The development of pathogenic T‐cells upon antigen immunization mainly occurs in the draining LNs where activated and immunogenic DCs interact with naïve T‐cells.31, 32 In this regard, surgical excision of the CNS‐draining LNs ameliorated EAE,36 while vaccination via injecting MOG‐microparticles into CNS‐draining LNs induced immunotolerance and prevented EAE.37 Our FACS analysis of lymphocytes isolated from the draining LNs relevant to the EAE model revealed a significant reduction in the abundance of the IL‐17‐producing T‐cells (Fig. 7a,b). Such reduction was accompanied with a significant reduction in the number of cells that express the key Th17‐differentiating cytokine IL‐23 (Fig. 7c,d). These findings suggest that development of Th17 cells was impaired in LNs as result of the diminished number of DCs expressing the key Th17‐differentiating cytokine IL‐23. This premise was further supported by our evaluation of the effect of Arg II deficiency on Th17 polarization and DC function as discussed below.

The impaired Th17 cell development in the LNs does not seem to be caused by a defect in T‐cells per se in the Arg II KO mice. Under optimized Th17 polarization conditions that provide excessive amounts of exogenous IL‐23 and other assisting Th17‐differentiating cytokines, Arg II‐deficient naïve T‐cells and WT naïve T‐cells exhibited similar capability to differentiate into Th17 cells (Fig. S3). Thus, the signal transduction involved in Th17 differentiation remains normal in Arg II‐deficient T‐cells. Physiologically, the differentiation of Th17 cells requires several DC cytokines, which include TGF‐β, IL‐6, IL‐23, IL‐22, IL‐1, IL‐21 and IL‐9. These cytokines regulate Th17 differentiation at different stages via different mechanisms.38 Initial stimulation with a combination of IL‐6 and IL‐1β induces the commitment of naïve CD4⁺ precursors toward Th17 cells. In the late stage, IL‐23 promotes terminal differentiation and expansion of committed Th17 cells.38 Among LPS‐induced Th17‐differentiating cytokines, IL‐6 and IL‐23 were significantly suppressed by Arg II deficiency (Fig. 8a,b). These data, combined with reduced IL‐23⁺ cells observed in the LNs (Fig. 7c,d), strongly suggest that Arg II deficiency attenuates EAE via compromising the ability of DCs to induce Th17 cell differentiation.

To summarize, our study has identified Arg II as a new factor that contributes to CNS autoimmunity and inflammation. The beneficial effect of Arg II deficiency on EAE is likely to be caused by the impairment of the Th17 development as a result of suppression of the key Th17‐differentiating cytokine production in DCs. Future studies are needed to explore if Arg II attenuates EAE also through other mechanisms. For example, Arg II deficiency is known to be beneficial in maintaining normal vascular‐endothelial function via increasing nitric oxide production and decreasing reactive oxygen species production.16, 39 The damage in the blood–brain barrier (BBB) facilitates transmigration of the inflammatory cells from the periphery to the CNS.40 Thus, it is important to examine if permeability of the BBB to immune cells is reduced in Arg II KO mice. Moreover, the effect of Arg II deficiency on CNS residential cells and their interaction with infiltrating immune cells cannot be overlooked.

Disclosures

The authors declare no competing conflict of interest.

Supporting information

Figure S1. Arg I and Arg II expression was differentially regulated during Th1 and Th17 cell polarization.Figure S2. Arg II deficiency decreased mRNA level of the Th17 cytokine IL‐17 in the spinal cord during the disease peak and recovery stages.Figure S3. Arg II deficiency did not inhibit polarization of Th17 cells and/or Th1 cells in the presence of exogenous T‐cell‐differentiating cytokines.Figure S4. Arg II deficiency did not lead to compensatory increase in Arg I expression in vivo and in vitro.