Clinicopathologic correlations of renal microthrombosis and inflammatory markers in proliferative lupus nephritis

From a large cohort of SLE patients followed up at two University Hospitals between 1986 and 2010, we selected those with available kidney biopsy specimens to perform histologic and immunohistochemical studies. All patients fulfilled at least four of the American College of Rheumatology criteria for the diagnosis of SLE [18]. The study received approval by the Ethics Committee of Hospital 12 de Octubre. This study was performed on previously collected and filed pathology materials obtained after informed patient consent. Kidney biopsies had been performed for diagnostic purposes in patients with proteinuria higher than 500 mg/day, abnormal urinary sediment, or elevated serum creatinine level. We considered that clinicopathologic correlations could be independently analyzed in different episodes, and therefore, biopsies from different episodes of the same patient were included when available. Kidney histology was evaluated and classified according to the ISN/RPS classification [19] by a pathologist who was blinded to patients' clinical and IHC data. Hematoxylin-eosin and phosphotungstic acid-hematoxylin (PAS) stained sections were evaluated. The whole series included 86 patients. In total, 101 histologic samples were analyzed. Among them, we selected 65 (64%) samples with proliferative nephritis (ISN/RPS types II, III, and IV) from 58 patients for further study. Their demographic, clinical, laboratory, and pathology characteristics before and at renal biopsy are detailed in Table 1.

APS was diagnosed according to the Sapporo criteria [20]. Definite APS was considered if at least one of the following clinical and laboratory criteria was satisfied. Clinical criteria included the presence of either vascular thrombosis or pregnancy morbidity. Laboratory criteria included the presence of aPL on two or more occasions at least 6 weeks apart, as demonstrated by one or more of the following: IgG and/or IgM anticardiolipin (aCL) antibodies in moderate or high titer and lupus anticoagulant (LA) activity. LA was detected after updated guidelines of the Subcommittee for the Standardization of Lupus Anticoagulant of the International Society of Thrombosis and Haemostasis [21], whereas aCLs were detected with enzyme-linked immunosorbent assay (ELISA) by following the manufacturer's instructions (Quanta Lite ACA IgG III and IgM III; Inova Diagnostics Inc, San Diego, CA, USA).

The following demographic data were recorded: sex, age, age at SLE diagnosis, age at renal disease, and time from kidney biopsy to the end of follow-up. Clinical and laboratory SLE data, including the disease-activity index SLEDAI [22], cardiovascular risk factors (high blood pressure (HBP), diabetes, hypercholesterolemia and/or hypertriglyceridemia, smoking, and hormonal contraception), APS criteria, and previous therapy were recorded.

Renal manifestations at biopsy and during the follow-up, specific therapy, response and time to response, and relapse and time to relapse were analyzed. Renal manifestations included HBP, renal failure defined as serum creatinine ≥1.5 mg/dl and/or creatinine clearance ≤65 ml/min, 24-hour proteinuria, hematuria, leukocyturia, and cellular casts. Complete clinical response to therapy was defined by 24 hours of proteinuria lower than 500 mg, normal urine sediment, and normal renal function (serum creatinine, < 1.5 mg/dl, and creatinine clearance, > 75 ml/min). Partial response to treatment was defined by improvement in all these parameters not reaching complete-response criteria. The presence of aPL (LA, IgG, or IgM aCL) and APS criteria before and after kidney disease was also recorded.

Immunohistochemical studies were performed on paraffin-embedded kidney sections. Immunostaining with either monoclonal mouse anti-human CD68 (CD68, clone KP1; DakoCytomation Denmark A/S, Glostrup, Denmark), monoclonal mouse anti-human gpIIIa (CD61, sz21; Immunotech, Marseille, France), or polyclonal rabbit anti-human C4d (C4dpAb; Biomedica Medizinprodukte GmbH & Co KG, Vienna, Austria) were performed. Renal sections from patients with primary thrombotic microangiopathy (TMA) were used as positive controls for CD61. Renal sections from a normal kidney were used as negative controls.

Sections were deparaffinized and pretreated with 10 μg/ml proteinase K (P6556; Sigma-Aldrich Quimica SA, Madrid, Spain) in the case of CD61 detection, or boiled in 1 mM EDTA or citrate buffer 10 mM, pH 6, for CD68 and C4d detection, respectively. The slides were incubated overnight at 4°C with primary antibodies. Immunohistochemical staining was performed with a peroxidase avidin-biotin complex technique (SK-4100; Vector Laboratories Inc., Burlingame, CA, USA). Diaminobenzidine was used as chromogen substrate, and sections were counterstained with hematoxylin.

Glomerular and extraglomerular CD68 and C4d staining was quantified by using the public domain image-processing program, Image J software [23]. In brief, the percentage of stained area in 10 glomeruli and 10 extraglomerular areas randomly selected from each sample was calculated. Two blinded and independent observers analyzed all slides.

The associations between categoric variables were tested by using the χ² or Fisher Exact test, where appropriate. The odds ratios (ORs) with the corresponding 95% confidence intervals (95% CIs) were calculated. For continuous variables, the comparisons were carried out by using the t test for two independent samples. A Spearman rank correlation was used to detect correlations among different study parameters. P values less than 0.05 were considered significant. The analysis was performed by using advanced SPSS software version 15.

In healthy, histologically normal kidney sections used as controls for IHC, CD61-positive platelet aggregates were not detected (Figure 1a). CD61-positive microthrombi were detected in 52% of the proliferative LN tissues (79% in glomerular and 62% in extraglomerular vessels) (Figure 1d, g, j).

CD68-positive macrophages were identified in the mesangial areas of both normal (Figure 1b) and lupus glomeruli (Figure 1e), as well as in extraglomerular areas (Figure 1h). The density of CD68-positive macrophages in both glomerular and extraglomerular areas was greatly increased in lupus patients. The mean stained area was 2.37% ± 3.02% and 0.72% ± 0.98% in glomerular and extraglomerular regions of lupus sections, respectively, compared with 0.0004% ± 0.0005% and 0.0003% ± 0.0007% in normal kidney.

C4d deposition was undetectable in normal kidney sections (Figure 1c), whereas it was detected in most SLE biopsies, in either glomerular (78%) or extraglomerular (86%) regions (Figure 1f, i). The mean C4d-stained glomerular and extraglomerular areas were 2.76% ± 3.73% and 1.04% ± 1.32%, respectively. The main extraglomerular pattern was diffuse linear C4d deposition along peritubular capillaries (98%), although some focal arteriolar C4d staining was also found in interstitial vessels (19%) (Figure 1l).

Patients with glomerular CD61-positive microthrombi had significantly higher glomerular (P = 0.03) macrophage infiltration. C4d extraglomerular deposition positively correlated with extraglomerular macrophage infiltration (r = 0.4; P = 0.004), but mean C4d deposition was similar in patients with or without CD61-positive microthrombi.

Both glomerular and extraglomerular macrophagic infiltrations positively correlated with SLE manifestations. Glomerular macrophagic infiltration correlated with SLEDAI (r = 0.5; P < 0.001) and was significantly higher in patients with positive anti-dsDNA antibodies (P < 0.001) or low complement levels (P < 0.001) at biopsy. Regarding renal manifestations, glomerular macrophagic infiltration positively correlated with 24-hour proteinuria (r = 0.4; P = 0.002) and was associated with the presence of microhematuria (P < 0.001), and cellular casts (P = 0.001). Extraglomerular macrophage infiltration correlated with SLEDAI (r = 0.3; P = 0.01) but was not associated with the presence of anti-dsDNA antibodies or low complement levels. Regarding renal manifestations, extraglomerular macrophage infiltration correlated with 24-hour proteinuria (r = 0.4; P = 0.001) but not with other analytic parameters. No correlation was found between glomerular or extraglomerular macrophage infiltration and renal function defined by creatinine clearance at biopsy.

Extraglomerular C4d deposition was significantly higher in patients with positive anti-dsDNA antibodies (P = 0.01) and correlated with SLEDAI (r = 0.4; P = 0.003) and 24-hour proteinuria (r = 0.3; P = 0.03). No differences in extraglomerular C4d deposition were found when we considered other clinical or laboratory parameters of renal function and SLE activity. Glomerular C4d deposition did not show any specific trend stratified by all clinical and laboratory variables. Clinical and laboratory correlations of platelet aggregates, macrophage infiltration, and C4d deposition are detailed in Table 2.

As initial therapy for LN, all patients received high doses of steroids (≥1 mg/kg/day) as well as one of the following immunosuppressive drugs: cyclophosphamide (79%), mycophenolate mofetil (17.7%), or azathioprine (1.6%). In addition, 54.2% received aspirin, and 5.1%, oral anticoagulants. Most patients (77%) had been treated with steroids before LN development for other manifestations at variable doses (Table 1). Twenty patients were treated with low-dose aspirin, and two patients were taking oral anticoagulation before LN development. In most patients treated with aspirin, indication was positive for LA and/or aCL antibodies (n = 12). Some patients received aspirin for HBP. The two patients receiving oral anticoagulation had previous venous or arterial thrombosis. Different groups of patients stratified by aspirin or steroid dose did not show any trend toward a different expression of LN or platelet aggregates incidence, or differences in the long-term outcomes. Fifty (83.3%) patients responded to treatment, and the mean time to partial or complete response was of 15.2 months (range, 1 to 101 months). To analyze the potential long-term effect of immunohistochemical findings on clinical and renal outcomes, we selected only patients homogeneously treated with intravenous cyclophosphamide according to the NIH protocol after excluding repeated episodes of LN [24]. In brief, induction therapy consisted of six monthly boluses of intravenous cyclophosphamide (0.5 to 1 g per square meter, further adjusted by leukocyte count) and glucocorticoids. After 6 months, patients were treated with intravenous cyclophosphamide boluses every 3 months and for at least 12 months after remission or stability. Maintenance therapy was variable and, according to clinical practice, included azathioprine, mycophenolate mofetil, or antimalarials with glucocorticoids. Taking into account only those patients (n = 43) who had received the same induction treatment, we did not found significant correlations between the presence of CD61⁺ microthrombi, macrophagic infiltration, or C4d deposition, and clinical outcomes including percentage of response to therapy, mean time to response, percentage of relapse, and long-term renal function.