Systemic lupus erythematosus (SLE) is associated with an increased risk of development of cardiovascular events (1, 2), and it has been hypothesized that accelerated development of atherosclerotic lesions would explain the cardiovascular vulnerability of patients with this disease (3). Our quantitative histopathologic postmortem analysis of the arteries of the SLE patient described herein, who had fatal occlusive large vessel disease, represents an approach to defining histomorphologic criteria for distinguishing arteriopathy with predominant arterial inflammation, i.e., arteritis, from conventional atherosclerosis. Accurate typing of large vessel disease may assist in determining the best treatment in individual patients with SLE.
A woman with a history of systemic lupus erythematosus presented with extensive bilateral strokes due to acute inflammatory, occlusive large vessel disease affecting several aortic branches including the carotid, subclavian, renal, and iliac arteries. We quantitatively characterized the arterial inflammation in this patient and compared it with the inflammatory infiltrates from 22 patients with conventional atherosclerosis. Profound histomorphologic differences from conventional atherosclerosis (predominance of CD8-positive lymphocytes, relative absence of macrophages, no ectopic neovascularization, no signs of plaque hemorrhage, concentric instead of eccentrical stenosis) suggest that this patient's accelerated arteriopathy was precipitated by pathogenic events other than conventional atherosclerosis.
Clinical and laboratory findings.
The patient, a 53-year-old woman, was admitted to our hospital in November 2004, after she was found at home in distress, with marked orthopnea and rapid respiratory failure requiring mechanical ventilation. She had a history of SLE, first diagnosed in 1993, when she developed a typical facial butterfly rash and arthralgias. She was treated with steroids and resorcinol for a few months. In 1997 the patient developed polyserositis, generalized polyarthritis, and acute lupus nephritis with renal failure. Treatment with methylprednisolone and cyclophosphamide was initiated. The treatment was well tolerated, and after clinical remission was achieved, steroids were tapered and eventually stopped in 2001. The patient was a heavy smoker (90 pack-years) and had given birth to 3 children.
At the initial clinical examination, the patient was unconscious and afebrile (36.0°C), and her blood pressure was 80/50 mm Hg. During the next few hours, her blood pressure recovered, and loud vascular bruits were heard on both carotid arteries and the left supraclavicular, upper abdominal, and periumbilical areas (Figure 1A). A significant blood pressure difference was noted between the left arm (110/55 mm Hg) and the right arm (230/120 mm Hg). Chest radiography revealed pulmonary edema and bilateral pleural effusions. Echocardiography showed left ventricular hypertrophy and reduced ejection fraction (30%), but no pericarditis. Hyporegenerative Coombs-negative normochromic, normocytic anemia and leukocytosis with 35% band forms but no toxic signs were present.
Laboratory studies showed an elevated erythrocyte sedimentation rate (56 mm/hour), a mildly elevated C-reactive protein level (23 mg/liter), and normal levels of electrolytes and liver and muscle enzymes. Despite normal creatinine values, renal function may have been slightly impaired, given the patient's low body mass index (16.5 kg/m2) and putatively poor muscle mass. The protein and albumin levels were profoundly reduced (45 gm/liter and 18 gm/liter, respectively), suggesting a catabolic state. Urinary sediment showed leukocytes and erythrocytes but no glomerular forms. Moreover, round epithelia, suggestive of acute tubular injury, as well as hyaline and granular casts were found. A 24-hour urine collection of 3.8 liters revealed proteinuria (2.8 gm). Serologic analyses revealed mildly increased antinuclear antibody and anti–double-stranded DNA titers, while antineutrophil cytoplasmic antibodies, anti–β2-glycoprotein, and antiphospholipid antibodies were negative. Levels of C3 and C4 complement components were slightly diminished. There were no signs of bacterial or viral infection. Results of a serologic test for syphilis were negative, and examination of cerebrospinal fluid showed a clear, colorless liquid without cells, and normal glucose, protein, and lactate concentrations.
A bilateral Babinski's sign and hyperactive deep tendon reflexes were noted, and computed tomography (CT) of the head (Figure 1B) disclosed extensive bilateral frontoparietal ischemic areas. CT angiography demonstrated concentrically stenotic arteries at the sites of vascular bruits (Figure 1A); bilateral carotid arteries (Figures 1C and D) and the left subclavian, both renal, and the iliac arteries at the bifurcation were affected with short, concentric stenoses. With the exception of the abdominal aorta, no arterial calcifications were detected. Duplex sonography of the carotid arteries confirmed the significant bilateral concentric and hypodense wall thickening suggestive of vasculitis (Figure 1C).
Treatment with high-dose methylprednisolone (1 gm/day) was initiated. The patient's neurologic condition deteriorated, and she died on the fourteenth day of hospitalization, in persistent coma.
Histopathologic examination of the arteries.
An autopsy was performed 20 hours after the patient's death. The aorta and its branches were examined macroscopically for overt stenoses, aneurysms, or occlusions. Segments (0.5 cm long) of the left carotid artery, the occluded left subclavian artery, and the left renal and left iliac arteries were removed, formalin fixed, and paraffin embedded, and 3-μm–thick sections were cut. For pathologic examination of the tissue, either hematoxylin and eosin or elastic–van Gieson staining was performed. Immunohistochemical staining was performed as previously described (4) to detect the following antigens: CD68 (M-0876; 1:200) as a macrophage marker, CD45 (M-0701; 1:100) as a lymphocyte marker, CD3 (A-0452; 1:100) as a T lymphocyte marker, CD8 (M-7103; 1:100) as a marker of the cytotoxic T lymphocyte subset, CD20 (M-0755; 1:200) as a marker of B lymphocytes, and von Willebrand factor antibody (M-0616, 1:10) as a marker of vascular endothelial cells (all from Dako, Glostrup, Denmark).
Quantitative analysis of arterial inflammation.
The tissue sections were analyzed microscopically, and calibrated digital images covering the intima, media, and adventitia of the artery were acquired (Diskus, Hilgers, Germany). For each digital image, the area of the intima, the media, and the adventitia was measured (ImageJ; NIH, Bethesda, MD). The positively stained cells or vascular profiles per area were counted. The inflammatory infiltrate and arterial microvascular network in the patient were compared with those found in conventional atherosclerosis, using previously described arterial tissue microarrays (4). Data on 22 patients who had had cardiovascular events were obtained for comparison (Table 1).
|Conventional atherosclerosis (n = 22)*||SLE-associated arteriopathy (n = 1)†|
|Common iliac artery|
|Common carotid artery|
|Cytotoxic T cells (CD8)|
|Common iliac artery|
|Common carotid artery|
At autopsy, the stenotic lesions of the carotid, subclavian, renal, and iliac arteries that had been identified by CT angiography were confirmed (Figure 1A). Furthermore, a small aneurysm of the left iliac artery was detected. Histopathologic examination of the affected arterial sectors did not reveal the classic hallmark lesions of vasculitis or of atherosclerosis. There was no dense infiltrate of granulocytes or arterial wall destruction as typically seen in polyarteritis nodosa, Kawasaki disease, or lupus vasculitis. No giant cells or granuloma was detected (typically seen in polymyalgia rheumatica, giant cell arteritis, Takayasu arteritis, or Wegener's granulomatosis). The American Heart Association (AHA) plaque types (5) ranged from I to V, but the most advanced AHA plaque types were not found at the sites of stenosis. Even the completely occluded left subclavian artery (Figure 1E) showed no typical signs of an advanced atherosclerotic plaque such as calcification or hemorrhage (5, 6), but rather exhibited concentric lumenal narrowing due to intimal hyperplasia and inflammation.
Quantitative histopathologic examination revealed extensive inflammation consisting mostly of mononuclear cells. In a previous report, we described the macrophage content of the intima and the arterial microvascular network in patients with symptomatic atherosclerosis (4). We now extend our analysis to other leukocyte subsets and to all 3 arterial wall layers.
The left carotid, left renal, and left iliac arteries of the patient were examined, and the microvascular network, macrophages, and CD8 T cells were compared with those observed in conventional atherosclerosis (Table 1). A remarkably high number of CD8-positive lymphocytes was found at all 3 arterial sites (Figure 2B). In conventional atherosclerosis, the media layer generally is free of leukocytes (Table 1). In contrast to these findings, macrophage counts in the present patient were unusually low compared with those in patients with conventional atherosclerosis (Table 1 and Figure 2C). Furthermore, and also in contrast to conventional atherosclerosis, neither ectopic neovascularization (4) nor plaque hemorrhage (6) was observed at the sites of stenosis. Since immune complexes play a pathogenic role in SLE, we examined the patient's arterial wall for the presence of B lymphocytes. B cells were virtually absent from the intima and media, but there were focally dense accumulations of B cells in the adventitia. Only a few patients with conventional arteriosclerosis exhibited similar infiltrates; the majority had no B lymphocytes infiltrating the arterial wall.
The patient described herein presented with an unusual course of occlusive large vessel disease: she developed bilateral, extensive cerebral strokes leading to deep coma as a consequence of arterial stenoses affecting the carotid, subclavian, renal, and iliac arteries. Left ventricular forward and backward failure precipitated this fatal course and presumably was caused by uncontrolled renovascular hypertension. Large vessel occlusion due to vasculitis has been described as an uncommon complication of SLE (7). However, recent duplex sonographic studies suggest that subclinical arteriosclerosis is a common event in patients with SLE (3, 8). If large vessel occlusion due to inflammation evolves from subclinical arteritis, accelerated arteriosclerosis observed in SLE may represent (and follow the pattern of) arteritis rather than conventional atherosclerosis.
The arteriopathy in this 53-year-old woman differed from conventional atherosclerosis in several ways. First, the coronary circulation, a vascular bed that usually has a high plaque burden (9), was not affected by the disease. Second, T lymphocytes, particularly CD8-positive cells, and not macrophages were the dominant cell subset found in the arterial wall involved in the disease. Third, ectopic microvascularization, which is typically seen in advanced atherosclerotic lesions (4), was not detectable in this patient, even in the arterial segments affected by extensive concentric intimal thickening. Finally, intimal hyperplasia, rather than plaque rupture (5, 10), was involved in the lumenal occlusion of the affected arteries. Our findings are shared at least partially with findings in transplant-associated arteriopathy: endothelialitis, the subendothelial accumulation of activated CD8-positive T lymphocytes observed in small arteries of transplanted organs, is the hallmark histologic lesion of acute vascular rejection (11), and chronic vascular rejection, which is characterized by intimal hyperplasia, is thought to be a T cell–driven process (12). Interferon-γ, a cytokine that is produced by antigen-specific, activated T cells, is capable of inducing intimal hyperplasia (13) and may also have been the main cytokine driving arterial obliteration in this patient with arteritis. The fact that professional antigen-presenting cells, such as macrophages, were virtually absent from the lesions suggests that residual vessel wall cells, such as endothelial or smooth muscle cells, could be the immediate targets of infiltrating lymphocytes. The peripheral release of self proteins in the course of vascular injury mediated by cytotoxic T cells may favor the generation of autoantibodies typically seen in SLE (14).
In conventional atherosclerosis, it is thought that ectopic neovascularization facilitates the recruitment of leukocytes into the intima (15) and supports lesional growth due to improved nutritional supply (4, 16). In the patient described herein, ectopic neovascularization of the media and intima was not present; despite this, however, lymphocyte counts at all 3 arterial sites exceeded those found in atherosclerosis. In a blood vessel explant model of arteritis, we have observed a similar degree of CD8 T cell infiltration when the tissue is treated with interferon-γ and tumor necrosis factor (Biedermann BC, et al: unpublished observations), suggesting that proinflammatory cytokines are sufficient, and ectopic neovascularization not necessary, to induce the type of arteritic inflammation found in this patient.
It should be emphasized that the data presented in this report were collected from a single patient with SLE. Although similar findings were consistently observed at all arterial sites analyzed, we cannot confirm that they are representative of the majority of patients with SLE and accelerated atherosclerosis. Therefore, these criteria—predominance of CD8-positive lymphocytes, relative absence of macrophages, lack of ectopic neovascularization, absence of signs of plaque hemorrhage, and concentric rather than eccentric stenosis—need to be validated in prospective studies of more patients with SLE. Despite its limitations, this quantitative analysis comparing the inflammatory infiltration in an arteriopathy associated with SLE with conventional atherosclerosis may have diagnostic and therapeutic implications for the management of SLE. Noninvasive methods to detect intramural changes in the course of inflammatory arteriopathy should be evaluated in controlled clinical trials, to assess their capacity to identify SLE patients at risk of developing this fatal complication. It has been reported that immunosuppressive treatment limited the progression of carotid arteriopathy in patients with SLE (3). The predominance of T cells in the arterial inflammatory infiltration observed in this patient and the histopathologic similarity to transplant-associated arteriopathy suggest that immunosuppressive treatment specifically targeting T lymphocytes or approaches aimed at interfering with peripheral T cell recruitment might be particularly effective in the prevention and treatment of this specific complication of SLE (17).
We thank Reto Krapf for helpful discussions and critical comments on the manuscript.