Endothelial nitric oxide synthase inhibits the development of autoimmune-mediated vasculitis in mice




Many different genes or mediators have been implicated in promoting the development of vasculitis, although little is known regarding the mechanisms that normally act to suppress lesion formation. Endothelial nitric oxide synthase (eNOS) has been shown to inhibit vascular inflammation in many different model systems, but its roles in the pathogenesis of vasculitis have not been elucidated. This study was undertaken to determine the functions of eNOS in the initiation and progression of vasculitic lesion formation.


MRL/MpJ-Faslpr mice lacking the gene for eNOS (Nos3−/−) were generated and comprehensively evaluated and compared to controls with regard to the development of autoimmune disease, including vasculitic lesion formation and glomerulonephritis.


Nos3−/− MRL/MpJ-Faslpr mice exhibited accelerated onset and increased incidence of renal vasculitis compared to Nos3+/+ controls. In contrast, no significant differences in severity of glomerulonephritis were observed between groups. Vasculitis was also observed in other organs of eNOS-deficient mice, including in the lungs of several of these animals. Ultrastructural analyses of renal lesions revealed the presence of electron-dense deposits in affected arteries, and IgG, IgA, and C3 deposition was observed in some vessels in the kidneys of Nos3−/− mice. In addition, Nos3−/− MRL/MpJ-Faslp mice showed increased levels of circulating IgG–IgA immune complexes at 20 weeks of age, compared to Nos3+/+ MRL/MpJ-Faslpr and Nos3−/− C57BL/6 mice.


These findings strongly indicate that eNOS serves as a negative regulator of vasculitis in MRL/MpJ-Faslpr mice and further suggest that NO produced by this enzyme may be critical for inhibiting lesion formation and vascular damage in human vasculitic diseases.

“Vasculitis” is a general term used to describe a heterogeneous group of disorders characterized by inflammatory processes leading to destruction of blood vessels (1). It can result in vessel necrosis, occlusion, and subsequently, tissue ischemia. Vasculitis is the primary pathologic manifestation of several different diseases, such as granulomatosis with polyangiitis (Wegener's), giant cell arteritis, and polyarteritis nodosa, and can also be observed in patients with systemic lupus erythematosus (SLE) and other connective tissue diseases (2–5).

It has been proposed that during the initiation of vasculitis, stimuli such as infectious agents, anti–endothelial cell antibodies (AECAs), immune complexes, complement proteins, cytokines, and other factors activate endothelial cells, leading to leukocyte adhesion and infiltration of the vessel wall (5–7). Endothelial and smooth muscle cell damage may then occur through a variety of mechanisms, including neutrophil release of granular contents and reactive oxygen species or T cell– and macrophage-mediated immune mechanisms (8, 9). Priming of neutrophils is also thought to be an important event in the development of vasculitis in some disorders (10). Antineutrophil cytoplasmic antibodies (ANCAs) and other inflammatory mediators may partially activate neutrophils, which can result in increased interactions of these leukocytes with endothelial cells, promoting their respiratory burst and ultimately leading to endothelial damage (5, 8, 10).

Nitric oxide (NO) is an important regulator of various physiologic and inflammatory responses and has been previously implicated in the development of vasculitis (11–13). NO is produced during the conversion of L-arginine to L-citrulline by different isoforms known as NO synthases (NOS) (14, 15). Endothelial nitric oxide synthase (eNOS) (Nos3) is a constitutively active enzyme that is expressed in endothelial cells and has important roles in regulating vasodilatation, inhibiting smooth muscle proliferation and platelet aggregation, modulating leukocyte–endothelium adhesion events, and controlling other key vascular functions (15, 16). Neuronal nitric oxide synthase (nNOS) is the predominant source of NO in neurons and functions in neurotransmission events, but is additionally expressed in muscle and blood vessels (14, 17). Finally, inducible nitric oxide synthase (iNOS) is expressed in many different cells, including macrophages, hepatocytes, and endothelial cells (14, 18). Expression and activity of iNOS are significantly up-regulated in response to inflammatory stimuli (19), and NO produced from this isoform is critical for host defense and other cellular processes (20–22). Reports of studies of iNOS in vasculitis models suggest that this enzyme significantly contributes to vessel damage (11, 23); however, the role of eNOS or nNOS in relevant in vivo models of vasculitis has yet to be elucidated.

To examine the possible involvement of eNOS in the context of vasculitis, we assessed Nos3-deficient MRL/MpJ-Faslpr mice (24) for vasculitic lesions in the kidneys and other organs. We found that Nos3−/− mice developed severe renal vasculitis and had significantly increased lesion severity scores compared to nonmutant controls. Loss of eNOS expression also led to accelerated onset of vasculitis, with renal lesions observed as early as 11 weeks of age. In addition, vasculitis was observed in other organs of Nos3−/− MRL/MpJ-Faslpr mice, including increased lesion formation in the lung. Finally, onset of glomerulonephritis was earlier in Nos3−/− MRL/MpJ-Faslpr mice, but these animals did not exhibit a significant increase in the overall severity of glomerular disease at later time points. Thus, our findings suggest that eNOS plays an important role in regulating the development of vasculitis, acting to prevent or restrict the onset and progression of vascular inflammation and damage.



MRL/MpJ-Faslpr and Nos3-deficient C57BL/6 mice were obtained from The Jackson Laboratory (25). MRL/MpJ-Faslpr mice deficient in eNOS expression were generated by backcrossing the Nos3 mutation for 8 generations onto the MRL/MpJ-Faslpr strain, and homozygotes were then generated by intercrossing. Mice were genotyped for the Nos3 mutation by polymerase chain reaction, and in some cases, homozygosity was confirmed by Western blot analysis of liver tissue. Nos3+/+ (MRL/MpJ-Faslpr)N8 littermates or inbred MRL/MpJ-Faslpr mice were used as controls. Approximately equal numbers of male and female animals were used for all studies. Animal care and experimental manipulations were conducted according to the Guide for the Care and Use of Laboratory Animals and with approval of the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham.

Histologic analysis and measurement of serum creatinine.

Kidneys were collected and fixed in buffered 10% formalin, processed for paraffin sectioning, sectioned at 5 μm, and stained with hematoxylin and eosin. Duplicate sections were stained with periodic acid–Schiff (PAS) and hematoxylin. Vasculitis and glomerulonephritis were scored by a pathologist who was blinded with regard to the animal's age and genotype. Vasculitis was assessed by examining the entire area of each section for vascular lesions and recording the type of vessel (arterioles, muscular arteries, elastic arteries, venules, and veins). For each affected vessel, adventitial mononuclear cells, neutrophils, necrosis, and fibrosis; medial mononuclear cells, neutrophils, and necrosis; and intimal mononuclear cells, neutrophils, necrosis, and proliferation were each graded 0, 1, 2, or 3, for absent, mild, moderate, or severe, respectively. Vasculitis scores for individual vessel lesions were calculated from these values according to the formula adventitial mononuclear cells + adventitial neutrophils + adventitial necrosis + adventitial fibrosis + (2 × medial mononuclear cells) + (2 × medial neutrophils) + (4 × medial necrosis) + (2 × intimal mononuclear cells) + (2 × intimal neutrophils) + (4 × intimal necrosis) + (2 × intimal proliferation). Total and average vasculitis scores for each mouse were calculated as the sum of individual vessel lesion scores and the mean score for affected vessels, respectively.

Glomerulonephritis was evaluated as previously described (26). Specific changes assessed included glomerular cellularity, necrosis, crescent and synechia formation, neutrophil accumulation, capillary basement membrane thickening and reduplication, mesangial sclerosis, capsular and periglomerular fibrosis, tubular changes, interstitial inflammatory cell accumulation, and interstitial fibrosis. Each was scored 0, 1, 2, or 3 for normal, mild, moderate, or severe, respectively. At least 6, and up to 15, glomeruli and adjacent tubules and interstitium were evaluated from both hematoxylin and eosin–stained and PAS-hematoxylin–stained sections from each mouse, and equal numbers of glomeruli from superficial, middle, and deep cortex were examined. Only glomeruli sectioned through the approximate center of the tuft and including the base of the tuft were assessed. Glomerulonephritis scores for each mouse were calculated as the mean of the summed component lesion scores for each glomerulus, with scores for necrosis and crescent formation each weighted by a factor of 2. Lymph nodes, spleen, stomach, intestines, liver, heart, and lungs were also collected from 20-week-old Nos3−/− mice (n = 10), stained with hematoxylin and eosin, and examined for vasculitis.

For ultrastructural analyses, kidney tissue was fixed in 2% glutaraldehyde, embedded in plastic, and examined by transmission electron microscopy. Toluidine blue–fuschin–stained semithin sections were examined by light microscopy prior to performing transmission electron microscopy to ensure that the tissue being examined was representative of renal parenchyma. Renal function was assessed by measuring serum creatinine, by liquid chromatography tandem mass spectrometry.

Immunohistochemistry analysis, determination of serum immunoglobulin levels, and autoantibody analysis.

For detection of T cells, B cells, and macrophages, kidney sections were first deparaffinized in xylene and rehydrated in phosphate buffered saline (PBS), and then microwave antigen retrieval was performed. The sections were then blocked with 10% serum, stained with rat anti-mouse CD3, B220, or Mac-2 antibodies (Cedarlane), washed, incubated with biotin-conjugated anti-rat IgG, and processed with avidin–biotin–peroxidase (Vector). Sections were then developed with diaminobenzidine (Vector) and counterstained with hematoxylin. For analysis of IgG deposition, kidneys were embedded in OCT compound and snap-frozen in liquid nitrogen. For IgA and C3 antibody staining, the kidneys were fixed in 4% paraformaldehyde in PBS, washed in PBS, placed in a 30% sucrose–PBS solution overnight, and then embedded in OCT compound. Five-micrometer sections were fixed in ice-cold acetone (IgG) or 4% paraformaldehyde (IgA and C3), washed in PBS, treated with 0.5% Triton X-100, washed in PBS, blocked with 3% bovine serum albumin–2% donkey serum, and then incubated with either Cy5-conjugated goat anti-mouse IgG (Jackson ImmunoResearch), fluorescein isothiocyanate (FITC)–conjugated goat anti-mouse IgA (SouthernBiotech), or FITC-conjugated goat anti-mouse C3 (MP Biomedicals) overnight at 4°C. Antibody-stained slides were then washed several times with PBS, and glass coverslips were mounted to the slides with Vectashield mounting medium (Vector). IgG anti-DNA autoantibody titers and total serum IgM, IgA, IgG, IgG1, IgG2a, IgG2b, and IgG3 levels were determined as previously described (26).

Analysis of glomerular C3, IgG, and IgA deposition.

Glomerular deposition was analyzed by histomorphometry using Image Pro Plus, version 6.2 (Media Cybernetics). Color images of at least 6 glomeruli were made at 20× objective magnification. The background of each image was thresholded to black (RGB values of 0, 0, and 0), glomeruli were selected as regions of interest, and areas of staining and density sums of areas of staining were recorded. The data were exported to Excel, and average areas and densities calculated for each mouse.

Determination of IgG–IgA immune complexes in serum samples. Sera were fractionated using a calibrated Superose 6 column. The levels of IgG–IgA immune complexes in each fraction were determined by capture enzyme-linked immunosorbent assay. Briefly, plates were coated with 2 μg/ml goat anti-mouse IgG, and the captured IgG–IgA complexes were then detected with horseradish peroxidase–conjugated goat anti-mouse IgA. The peroxidase chromogenic substrate o-phenylenediamine–H2O2 (Sigma-Aldrich) was then added. The color reaction was stopped with 1M sulfuric acid, and absorbance at 490 nm was measured using a microplate reader.

Statistical analysis.

Vasculitis and glomerulonephritis scores were analyzed by general linear model analysis of variance supplemented with Tukey's test for mean comparisons, using Statistix 9.0. P values less than or equal to 0.05 were considered significant.


Generation of eNOS-deficient MRL/MpJ-Faslpr mice.

MRL/MpJ-Faslpr mice spontaneously develop immune-complex–based glomerulonephritis, dermatitis, and vasculitis in many different organs (24). To investigate whether eNOS mediates vasculitis and related disease processes, we backcrossed a Nos3-null mutation onto the MRL/MpJ-Faslpr strain background for 8 generations (25). Heterozygotes were then intercrossed to generate Nos3−/− mice. Control and eNOS-deficient N8 mice displayed the expected clinical features characteristic of the inbred MRL/MpJ-Faslpr strain, including the progressive development of dermatitis and lymphadenopathy. We also analyzed serum immunoglobulin and autoantibody levels in Nos3−/− and Nos3+/+ MRL/MpJ-Faslpr mice. No significant differences in levels of total serum IgG, IgG1, IgG2a, IgG2b, IgG3, IgM, or IgA, or in anti–double-stranded and anti–single-stranded DNA antibody levels, were observed at age 20 weeks (data not shown).

Histologic features of renal vasculitis and glomerulonephritis.

Previous studies of Nos3-deficient mice in other renal inflammation models, including anti–glomerular basement membrane glomerulonephritis, diabetic nephropathy, and remnant kidney systems, have shown that eNOS deficiency leads to increased severity of disease, suggesting a regulatory role of this enzyme (27–29). However, with the exception of the anti–glomerular basement membrane glomerulonephritis model, different pathogenic mechanisms are thought to mediate kidney injury, and the development of vasculitis has not been reported in any of these model systems.

To determine whether loss of eNOS expression might also lead to similar alterations in the onset or severity of renal inflammation in the MRL/MpJ-Faslpr model, we performed a comprehensive analysis of vasculitis and glomerulonephritis in Nos3−/− and Nos3+/+ mice. Mice were killed at age 11–13 weeks (12-week age group), 16 weeks, 19–22 weeks (20-week age group), or 28 weeks or when moribund (28-week age group). These groups were selected because we expected, based on our previous analyses of this model, to see no vasculitis or glomerulonephritis in mice <16 weeks old, minimal disease in 16-week-old mice, mild-to-moderate disease in 20-week-old mice, and more advanced disease in 28-week-old and older mice (26, 30). Interestingly, we found that loss of eNOS expression resulted in a significant increase in severity of renal vasculitis, but not glomerulonephritis, in mice in the 20- and 28-week age groups (Figure 1). The increase in total vasculitis score was due to increased numbers of affected vessels in Nos3−/− mice, as demonstrated by the finding that total vasculitis scores closely paralleled the numbers of affected vessels.

Figure 1.

Renal vasculitis and glomerulonephritis scores in Nos3+/+ (wild-type [WT]) and Nos3−/− (endothelial nitric oxide synthase [eNOS]–deficient) MRL/MpJ-Faslpr mice. Total renal vasculitis scores, numbers of affected renal vessels, and glomerulonephritis scores were determined in kidneys from mice killed at different ages. The numbers of Nos3+/+ and Nos3−/− mice, respectively, evaluated at each time point were as follows: at 12 weeks, n = 7 and n = 7; at 16 weeks, n = 5 and n = 5; at 20 weeks, n = 22 and n = 24; and at 28 weeks (or moribund), n = 6 and n = 11. Values are the mean ± SD. ∗ = P ≤ 0.05 versus the 12- and 16-week time points.

To determine whether partial loss of eNOS expression could also alter renal disease in this model, we evaluated vasculitis and glomerulonephritis in kidneys of Nos3+/ MRL/MpJ-Faslpr mice at 19–22 weeks of age. We found that Nos3+/ mice had an intermediate vasculitis phenotype, although differences between the mean score in heterozygotes and those in Nos3−/− and Nos3+/+ mice were not statistically significant (data not shown). In addition, there were no significant differences in the glomerulonephritis scores among the different genotypes (data not shown). Finally, vasculitis was not observed in the kidneys of Nos3−/− C57BL/6 mice, even at 40 weeks of age (results not shown), suggesting that eNOS deficiency alone is not sufficient for the development of lesions.

Renal vasculitis involved mainly medium-sized arteries, with no significant involvement of small blood vessels (Figure 2). Vasculitis in both eNOS-deficient and control MRL/MpJ-Faslpr mice was multifocal and affected segmental, arcuate, and interlobular muscular arteries. In Nos3+/+ mice, lesions were typical of those in the polyarteritis nodosa–like vasculitis previously described in MRL/MpJ-Faslpr mice (31), with intense adventitial accumulation of lymphocytes, macrophages, and neutrophils, medial necrosis and neutrophil accumulation, and intimal inflammatory cell accumulation (Figures 2A and B). Occasionally these changes were accompanied by slight-to-mild intimal proliferation. In contrast, in Nos3−/− mice, intimal proliferation and inflammation were strikingly prominent in many lesions (Figures 2C and D), a feature present in few lesions in Nos3+/+ mice even at 28 weeks of age. Immunohistochemical staining of selected sections showed that cellular accumulations contained CD3+, B220+, and Mac-2+ leukocytes (Figures 3A–C). Although the differences between mean scores were not statistically significant, it was noteworthy that vasculitis and glomerulonephritis were present in Nos3−/− mice in the 12-week age group whereas such lesions were absent in Nos3+/+ mice of that age (Figure 1), suggesting that eNOS deficiency additionally leads to earlier onset of renal disease.

Figure 2.

Renal arteritis in Nos3+/+ and Nos3−/− MRL/MpJ-Faslpr mice. A and B, Typical granulomatous adventitial inflammation and necrotizing medial and intimal inflammation in samples from representative Nos3+/+ mice. C and D, Similar adventitial and medial inflammation in samples from representative Nos3−/− mice, but with extensive intimal proliferation and inflammation. Hematoxylin and eosin stained; original magnification × 20.

Figure 3.

Immunohistochemical staining of renal artery lesions in a representative 20-week-old Nos3−/− MRL/MpJ-Faslpr mouse. Vasculitic lesions were stained with anti-CD3 (A), anti-B220 (B), and anti–Mac-2 (C) antibodies. Original magnification × 20.

The appearance of glomerular lesions was as would be expected in this SLE model and did not differ in character between Nos3+/+ and Nos3−/− MRL/MpJ-Faslpr mice (results not shown). Mild glomerulonephritis was characterized primarily by mild-to-moderate mesangial thickening (sclerosis) and proliferation, with slight-to-mild neutrophil accumulation. Features of more severe lesions in older mice included moderate-to-severe mesangial thickening and proliferation, moderate-to-severe neutrophil accumulation, and, in the most severe instances, segmental tuft necrosis, epithelial crescent formation, synechia (adhesion of the tuft to Bowman's capsule), pericapsular inflammatory cell accumulation, and tubular atrophy. Finally, no significant differences in serum creatinine levels were observed between mutant and control mice at 16 or 20 weeks of age (data not shown), even though Nos3−/− mice exhibited both earlier onset and increased severity of vasculitis.

Ultrastructural characterization.

Electron microscopic examination of the larger muscular arteries in the kidney revealed marked contrasts between Nos3−/− and Nos3+/+ MRL/MpJ-Faslpr mice (Figure 4). Muscular arteries from Nos3+/+ mice had no significant ultrastructural abnormalities. The endothelium, elastic layer, muscular layer, and adventitia exhibited typical features of a normal muscular artery, without intimal hyperplasia, disruption of the elastic membrane, muscular hypertrophy, adventitial inflammation, or electron-dense deposits (Figure 4A). In contrast, the muscular arteries from Nos3−/− mice showed intimal thickening with an intact elastic membrane (Figure 4B). There was disruption of the muscular layer, and readily identified large electron-dense deposits were seen in the area of muscular layer disruption and adventitial tissue. In addition, smaller electron-dense deposits were visualized in the subendothelium immediately above the elastic layer. Occasional lymphocytes were present in the adventitia and disrupted muscular layer.

Figure 4.

Ultrastructural analyses of renal vasculitic lesions and glomerulonephritis. A, Electron micrograph of a normal muscular artery from a representative 20-week-old Nos3+/+ MRL/MpJ-Faslpr mouse, with no ultrastructural abnormalities in the artery structure or adventitial tissue. B, Electron micrograph of a muscular artery from a representative age-matched Nos3−/− MRL/MpJ-Faslpr mouse, exhibiting electron-dense deposits beneath the endothelial layer adjacent to the elastic layer (a) and within the adventitial tissue (b). In addition, there is disruption of the muscular layer (c). C, Electron micrograph of mesangium from a representative 20-week-old Nos3+/+ MRL/MpJ-Faslpr mouse, with expanded matrix and readily identified intermediate-to-large electron-dense deposits (asterisk). D, Electron micrograph of mesangium from a representative age-matched Nos3−/− MRL/MpJ-Faslpr mouse, with reduced, fine, small electron-dense deposits (asterisk). Original magnification × 5,000.

Consistent with these findings, a small number of the affected vessels in the kidneys of Nos3−/− mice showed deposition of both IgG and IgA (Figures 5A and B), and C3 staining was also evident in these lesions (Figure 5C). However, such deposition was sparse and difficult to find, both in these mice and, especially, in Nos3+/+ MRL/MpJ-Faslpr mice, due to the low frequency of vasculitis in the latter. Based on these observations, we initiated studies of circulating IgG–IgA immune complexes to determine whether eNOS deficiency led to any alterations in their levels or other properties. We found that levels of IgG–IgA immune complexes were elevated in Nos3−/− MRL/MpJ- Faslpr mice compared to Nos3+/+ MRL/MpJ-Faslpr controls and Nos3−/− C57BL/6 mice at 20 weeks of age (Figure 5D). These complexes were large, with a molecular mass of >670 kd, in MRL/MpJ-Faslpr mice of both genotypes. In contrast, in eNOS-deficient C57BL/6 mice these complexes occurred in low levels and had a smaller molecular mass. These findings suggest that circulating immune complexes may play a role in promoting vasculitis in the kidneys of eNOS-deficient MRL/MpJ-Faslpr mice, although additional mechanisms for initiating lesion formation must also exist.

Figure 5.

A–C, Antibody staining of kidney sections, showing arterial wall IgG and IgA deposition. Serial sections from a representative 20-week-old Nos3−/− MRL/MpJ-Faslpr mouse are shown. Sections were stained with Cy5-conjugated goat anti-mouse IgG (A), fluorescein isothiocyanate (FITC)–conjugated goat anti-mouse IgA (B), or FITC-conjugated goat anti-mouse C3 (C). Original magnification × 40. D, Levels of circulating IgG–IgA immune complexes. Serum IgG–IgA immune complex levels were elevated in Nos3−/− MRL/MpJ-Faslpr mice (shaded circles; n = 5) compared to Nos3+/+ MRL/MpJ-Faslpr mice (open circles; n = 6) and Nos3−/− C57BL/6 mice (solid circles; n = 3). Values are the mean ± SD. OD = optical density.

Ultrastructural analysis of glomeruli from 20-week-old Nos3+/+ MRL/MpJ-Faslpr mice showed a moderate increase in mesangial cells and matrix, without segmental lesions or crescents, and there were readily identifiable intermediate-to-large electron-dense deposits within the mesangium and paramesangium, as well as large deposits in the glomerular basement membranes (Figure 4C). The deposits in the basement membrane were subepithelial and often extended into the intramembranous areas and tended to be distinct and not confluent. The basement membrane was markedly thickened in the areas with membranous deposits, and the foot processes were attenuated. The visceral epithelial cells were prominent, but lacked cytoplasmic vacuolization. There were no tubular reticular inclusions, and tubules and peritubular capillaries lacked electron-dense deposits. Finally, there were occasional aggregates of lymphocytes and plasma cells within the interstitium.

In contrast, kidneys from age-matched Nos3−/− MRL/MpJ-Faslpr mice showed a mild increase in mesangial matrix, and only focal areas with minimally increased mesangial cells (Figure 4D). The mesangium contained fine, small electron-dense deposits with mildly increased mesangial matrix. Fine, small paramesangial deposits were also observed. The basement membrane was of appropriate thickness. The foot processes were well maintained. There were rare small, subendothelial deposits adjacent to the basement membrane. The visceral epithelial cells were not prominent and lacked vacuolization, and no tubular reticular inclusions were observed. Similar to findings in Nos3+/+ mice, the tubules and peritubular vessels lacked deposits, and occasional aggregates of lymphocytes and plasma cells were seen.

Glomerular C3, IgG, and IgA deposition.

Glomerulonephritis in MRL/MpJ-Faslpr mice has been previously shown to be promoted in part due to immune complex deposition and complement fixation, which activates the inflammatory response and leads to glomerular injury (24, 32). Our ultrastructural analyses suggested that there were some differences between eNOS-deficient and control mice in the extent of electron-dense deposits in the mesangium, paramesangium, and glomerular basement membrane. Therefore, we performed antibody staining and morphometric analysis of kidney sections from 20-week-old Nos3−/− and Nos3+/+ MRL/MpJ-Faslpr mice, to compare the pattern and intensity of staining for glomerular C3, IgG, and IgA. We found that staining patterns varied considerably among glomeruli within and between individual mice of both genotypes, with no statistically significant differences in staining areas or densities as determined by histomorphometry (results not shown).

Vasculitis in other tissues.

In a preliminary assessment of vasculitis in other viscera in 20-week-old Nos3−/− MRL/MpJ-Faslpr mice, we found necrotizing inflammation of small-to-medium arteries in the spleen, lymph nodes, and digestive and reproductive tracts, as we and others have previously observed in MRL/MpJ-Faslpr mice (33–35). However, findings in the lung were strikingly different. Seven of the 10 mice had multifocal pulmonary arteritis with endarteritis similar to that in renal vessels (Figure 6). In our experience, true vasculitis, evidenced by degeneration, necrosis, and/or inflammatory cell accumulation within the vessel wall, is rare in pulmonary vessels of MRL/MpJ-Faslpr mice. These findings, along with those in the kidney, suggest that loss of eNOS expression can significantly promote the development of vasculitis in multiple tissues.

Figure 6.

Histopathologic findings in a lung section from a representative 20-week-old Nos3−/− MRL/MpJ-Faslpr mouse. Pulmonary endarteritis with intimal proliferation and accumulation of lymphocytes and neutrophils is observed. Hematoxylin and eosin stained; original magnification × 20.


Mechanistic studies of the immune and inflammatory events leading to the development of vasculitis in different disorders have been hampered by several factors, including the heterogeneity of the diseases themselves, poor understanding of the genetic and environmental factors involved, and a shortage of appropriate animal models (36). Thus, for many different vasculitic diseases the key genes or molecules that promote or inhibit lesion formation remain largely unidentified, which has further hindered the generation of new agents that may be used for treatment. Previous investigations of eNOS have shown that this enzyme is an important source of NO in vessels for regulating blood pressure, smooth muscle proliferation, inflammatory responses, angiogenesis, and other vascular functions (37, 38). In addition, alterations in eNOS expression or activity have been correlated with the pathogenesis of diseases such as atherosclerosis, diabetes, and asthma (16, 37, 39). However, studies specifically analyzing the role of this enzyme in the initiation and progression of vasculitis have not been reported.

Herein we have shown, using the MRL/MpJ-Faslpr mouse model of SLE and vasculitis, that loss of eNOS expression accelerates the onset of disease and increases the number and distribution of affected vessels in the kidney. Renal vasculitic lesions were detected as early as age 11 weeks in Nos3−/− MRL/MpJ-Faslpr mice and primarily affected medium-to-large arteries. Although not the main focus of the current studies, vasculitis was additionally observed in several other organs of the Nos3−/− MRL/MpJ-Faslpr mice, including increased expression in the lung compared to Nos3+/+ controls. Nos3+/ mice presented with an intermediate renal phenotype compared to homozygous mutant and control mice, suggesting that the overall level of eNOS expression is a critical determinant in regulating lesion formation. Spontaneous development of vasculitis has not been demonstrated in Nos3-deficient animals in previous studies, which have largely utilized mice on non–autoimmune strain backgrounds (25, 40). In accordance with those investigations, we did not observe lesions in the kidneys of eNOS-deficient C57BL/6J mice up to 40 weeks of age. Thus, these findings suggest that decreased eNOS expression does not by itself lead to vasculitis, but requires additional genetic factors to promote lesion formation.

Several mediators are thought to be critical in promoting endothelial injury leading to vasculitis in MRL/MpJ-Faslpr mice, including immune complex deposition, ANCAs, and AECAs (41–43). However, studies to determine the individual roles of these factors in promoting lesions in MRL/MpJ-Faslpr mice are lacking, and other inflammatory stimuli may also contribute. In the present study we observed increased levels of IgG–IgA immune complexes in the circulation and identified both IgG and IgA immune deposits in some affected blood vessels in the kidneys. Furthermore, we found evidence of complement activation, based on C3 staining in these same areas. These results suggest that eNOS deficiency may render blood vessels more vulnerable to vasculitic lesion formation at sites of immune complex deposition. In addition, they indicate that reduced eNOS expression may alter the formation or clearance of circulating immune complexes. As this was an unexpected observation, understanding of the eNOS-dependent mechanisms that restrict vasculitis development requires further study. Thus, the Nos3-deficient MRL/MpJ-Faslpr mouse model presents a unique opportunity for detailed investigation of mechanisms of endothelial cell functions and vasculitis and for identifying possible novel functions of eNOS in controlling autoimmunity.

Decreased Nos3 expression may also predispose specific areas of the vasculature to inflammation and damage in response to binding of endothelial cell antigens by AECAs, or indirectly as a consequence of ANCA activation of neutrophils. Previous studies of eNOS suggest that this enzyme modulates leukocyte–endothelial cell interactions by down-regulating endothelial adhesion molecule and chemokine expression through inhibition of NF-κB (44–46). Thus, the increased incidence and severity of vasculitic lesions in Nos3−/− MRL/MpJ-Faslpr mice may result from increased leukocyte activation and/or infiltration of vessels at sites of immune complex deposition, or in areas where the endothelium has undergone stimulation by AECAs or other stimuli. An additional contributing factor may involve the loss of eNOS-mediated suppression of intimal proliferation, which might lead to augmented luminal occlusion and reduced blood flow following initial lesion formation (47, 48). However, further studies are necessary to specifically identify the mechanisms by which eNOS limits or suppresses vasculitis in the MRL/MpJ-Faslpr mouse model, as well as to determine whether this enzyme plays similar roles in the pathogenesis of human vasculitides.

Earlier studies have linked NO to the progression of inflammatory disease in MRL/MpJ-Faslpr mice, although those investigations focused primarily on the role of iNOS or used nonspecific NO synthase inhibitors. It has been shown that MRL/MpJ-Faslpr mice overexpress iNOS and overproduce NO in an age-dependent manner that parallels expression of autoimmune disease (11). In addition, blocking of NO production by oral administration of the inhibitor NG-monomethyl-L-arginine reduced the severity of arthritis, glomerulonephritis, and vasculitis in this model (11, 49, 50). Interestingly, Nos2−/− MRL/MpJ-Faslpr mice exhibited significant decreases in both the incidence and the severity of renal vasculitis compared to wild-type mice (23), suggesting that iNOS may contribute to lesion formation through NO production in infiltrating leukocytes. In contrast, iNOS deficiency did not inhibit the development of glomerulonephritis in this model (23). In our studies, we observed that glomerulonephritis developed earlier in Nos3−/− MRL/MpJ-Faslpr mice than in controls; however, at later time points, eNOS-deficient mice did not show any significant increases in the severity of glomerular disease, although some minor ultrastructural differences were noted.

In summary, we have described a novel and inhibitory role of eNOS in the context of vascular lesion formation and severity in a well-characterized model of SLE and vasculitis. These findings may facilitate studies to gain further insights into the pathogenesis of vascular inflammation and to test novel targets for the treatment of vasculitis.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Bullard had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Schoeb, Jarmi, Hicks, Novak, Agarwal, Bullard.

Acquisition of data. Schoeb, Jarmi, Hicks, Henke, Zarjou, Suzuki, Kramer, Novak, Bullard.

Analysis and interpretation of data. Schoeb, Jarmi, Hicks, Henke, Zarjou, Suzuki, Kramer, Novak, Agarwal, Bullard.


We would like to acknowledge the University of Alabama at Birmingham–University of California, San Diego O'Brien Core Center and the University of Alabama at Birmingham Mucosal Human Immunodeficiency Virus and Immunobiology Center and Animal Resources Program Comparative Pathology Laboratory for their technical assistance and support of these studies.