Clinicopathologic Associations of Anti-endothelial Cell Antibodies in Immunoglobulin A Nephropathy and Lupus Nephritis

Clinicopathologic Associations of Anti-endothelial Cell Antibodies in Immunoglobulin A Nephropathy and Lupus Nephritis Min-Xia Wang, MB, MD, Rowan G. Walker, MD, FRACP, MBBS, and Priscilla Kincaid-Smith, MBBCh, FRACP, MD,FRCP, FRCPA • Sera from 45 patients with lupus nephritis (LN), 63 patients with immunoglobulin A nephropathy (lgA N), and 71 glomerulonephritic controls (including 44 mesangial proliferative glomerulonephritis cases, 14 membranous glomerulonephritis cases, and 13 focal segmental glomerular sclerosis cases), and from 33 normal control subjects were tested by a cellular enzyme-linked immunoabsorbent assay for their anti-endothelial cell antibody (AECA) activity. Compared with normal controls, AECAs of the IgG subtype (AECA-lgG) were detected in LN (P < 0.001) and AECAs of the IgA subtype (AECA-lgA) were detected in both IgA Nand LN (P = 0.018 and P < 0.001, respectively). Binding activity of AECA to endothelial cells was inhibited by endothelial cell lysate and fibroblast lysate but not by lymphocyte lysate, double stranded-DNA, or bovine serum albumin. Anti-endothelial cell antibody-positive sera also reacted with fibroblasts. In IgA N, associations were found between the presence of AECA and younger age (P = 0.036), proportion of crescents greater than 10% (P = 0.016), fibrin crescents (P = 0.016), and focal and segmental necrotizing lesions (P = 0.047). In LN, inverse associations were found between the presence of AECA and the duration of disease (P = 0.021), elevated serum creatinine levels (P = 0.020), decreased creatinine clearance (P = 0.043), and frequency of chronic renal failure (P = 0.036). Positive associations were observed between the presence of AECA and active lupus (P = 0.017), anti-nuclear antibodies (P = 0.015), and anti-DNA antibodies (P = 0.041). Our results suggest that AECA may be linked with the pathogenesis of LN and IgA N. © 1993 by the National Kidney Foundation, Inc. INDEX WORDS: Anti-endothelial cell antibodies; immunoglobulin A nephropathy; lupus nephritis.

I

N 1971, Lindquist and Osterland first reported that anti-endothelial cell antibodies (AECAs) were present in several inflammatory diseases. I Shortly thereafter, Tan and Pearson 2 adapted an immunofluorescent technique using a mouse kidney substrate and were able to show the presence of AECAs in a variety of rheumatic diseases. Since then, AECAs have been demonstrated in several disease situations, such as HLA-matched graft rejection, 3 thrombotic thrombocytopenic purpura,4 systemic sclerosis,5 hemolytic uremic syndrome,6 pre-eclampsia/ autoimmune hypoparathyroidism,s and recently in Wegener's granulomatosis and microscopic polyarteritis. 9 Although the pathogenic role of these antibodies remains unclear, it is evident that AECAs From the Department of Nephrology and the University of Melbourne Department of Medicine, Royal Melbourne Hospital, Parkville, Australia. Received July 9, 1992; accepted in revised form April 8, 1993. Dr Wang is the recipient of an 1SN International Fellowship Training Award (International Society of Nephrology). Her work was also supported in part by the Renal Research Institute and a grant from the Victor Hurley Medical Research Fund. Address reprint requests to Rowan G. Walker, MD, FRACP, MBBS, Department ofNephrology, Royal Melbourne Hospital, Parkville, Australia 3050. © 1993 by the National Kidney Foundation, Inc. 0272-6386/93/2203-0003$3.00/0 378

have been associated particularly with autoimmune disorders, especially immune vasculitis. Many renal diseases, especially glomerulonephritis, are considered to be autoimmune disorders, and it is conceivable that the endothelial cell (EC) in the human glomerulus is actively involved in the immune processes leading to glomerular injury. 10 In 1988, Yap et al first reported the detection of AECAs in patients with primary glomerulonephritis: IgA nephropathy (IgA N). II To further determine the presence or absence of AECAs and to evaluate the clinical significance of AECAs in different types of glomerulonephritis, we have used a cellular enzyme-linked immunosorbent assay (ELISA) method for the detection of AECAs. The association between the presence of AECAs and various clinical, laboratory, and renal biopsy features has been evaluated. MATERIALS AND METHODS

Patients Sera from 45 patients with biopsy-proven lupus nephritis (LN) (six men and 39 women with a mean age of39.2 ± 15.3 years), 63 patients with 19A N (45 men and 18 women with a mean age of 39.5 ± 14.7 years), and 71 glomerulonephritic controls, including 44 mesangial proliferative glomerulonephritis (lgA-negative) (MPGN) cases (19 men and 25 women with a mean age of 38.1 ± 11.8 years), 14 membranous glomerulonephritis (MGN) cases (nine men and five women with a mean age of 45.7 ± 16.8 years), and 13 focal and segmental

American Journal of Kidney Diseases, Vol 22, No 3 (September), 1993: pp 378-386

AECA IN IgA NEPHROPATHY AND lUPUS NEPHRITIS

379

Table 1. Associations Between the Presence of Anti-endothelial Cell Antibodies and Clinical and laboratory Parameters LN (n

Patient

Age <40 yr Sex (M/F) Duration of disease < 10 yr MAP> 110 mm Hg Urine red blood cell count >8,OOO/ml > 105 ml Proteinuria > 0.15 g/d >1.0 g/d Serum creatinine> 1.2 mg/dl Creatinine clearance < 1.5 ml/s ANA positive Anti-DNA-positive ACA positive lupus anticoagulant Serum complement C3 < 0.083 gIL C4 < 0.15 g/l CRF Disease activity'

AECA· Positive

= 45)

IgA N (n

P

AECA· Negative

= 63) P

Value

AECA· Positive

AECA· Negative

Value

20/30 (67%) 3/27 21/30 (70%) 6/29 (21%)

7/15 (47%) 3/12 5/15 (33%) 6/15(40%)

NS NS 0.021 NS

14/19 (74%) 14/5 18/19 (95%) 1/18 (6%)

20/44 (45%) 31/13 36/44 (82%) 2/41 (17%)

0.036 NS NS NS

14/25 (56%) 4/25 (16%)

8/15 (53%) 2/15 (13%)

NS NS

16/19 (84%) 10/19 (53%)

35/41 (85%) 19/41 (46%)

NS NS

16/27 (59%) 6/27 (22%) 4/30 (13%) 13/27 (48%) 19/28 (68%) 14/27 (52%) 4/12 (33%) 8/29 (28%)

10/13 (77%) 5/13 (38%) 7/15 (47%) 12/25 (80%) 3/12 (25%) 2/12 (17%) 1/6 (16%) 3/13(23%)

NS NS 0.02 0.043 0.015 0.043 NS NS

16/19 (84%) 6/19 (32%) 9/19 (47%) 12/19 (62%) 1/8 (13%) 1/7 (14%)

34/41 (81%) 9/41 (21%) 22/43 (51%) 23/27 (62%) 0/17 0/6

NS NS NS NS NS NS

6/19 (32%) 9/19 (47%) 1/30 (3%) 15/30 (50%)

0/10 4/10 (40%) 4/15 (27%) 2/15 (13%)

NS NS 0.036 0.017

1/19 (5%) 6/13 (46%)

6/44 (14%) 2/23 (9%)

NS 0.017

Abbreviations: NS, not significant; MAP, mean arterial pressure; ANA, anti-nuclear antibodies; Anti-DNA, anti-doublestranded DNA antibodies; ACA, anti-cardiolipin antibodies; CRF, chronic renal failure. CRF, chronic renal failure; persistantly elevated serum creatinine > 2.3 mg/dl. • Disease activity: active IN was defined by recent clinical, renal , and/or extrarenal manifestations of lupus, together with serologic and/or histologic evidence of activity (Correia et al 30); active IgA N was defined by the histologic features of crescents and focal necrotizing lesions (Droz3 ').

hyalinosis and sclerosis (focal glomerular sclerosis) (FSGS) cases (six men and seven women with a mean age of 38.2 ± 13.7 years) were collected, aliquoted, and stored at ±70°C until assayed.

Normal Control Subjects Sera were also collected from 33 healthy adult volunteers (17 men and 16 women with a mean age of 33.8 ± 7.9 years). These volunteers served as normal control subjects (NCs).

Cellular Enzyme-Linked Immunosorbent Assay for the Determination ofAnti-endothelial Cell Antibodies Endothelial cells were obtained from human umbilical cord veins by an established method and cultured under standard conditionsY Endothelial cells between the second to third passages were detached by exposure to trypsin-EDTA (Gibco, Grand Island, NY) and seeded onto gelatin-coated, 96-well, flat-bottomed tissue culture plates (Flow Laboratories, Mclean, VA) at 2 X 104 cells per well in EC culture medium. ' 2 Cells reached a confluent monolayer in approximately 24 hours and were fixed to the culture plates by exposure to a

hair drier at room temperature for 20 minutes. Cells fixed in this way could be stored at 4 °C for at least 2 weeks without affecting the AECA-binding activity. The detection of AECA in serum samples was carried out using a modification of the cellular ELISA method described by Rosenbaum et al.13 In brief, the test sera were diluted I: 300 for the IgG-specific assay and I :50 for the IgA assay, and were incubated for 2 hours at 37°C in triplicate wells containing cultured ECs. The plates were washed four times with T-PBS (0.05% Tween in phosphate-buffered saline), then incubated at 37°C for 90 minutes with 100 ilL/well of alkaline phosphate-conjugated, affinity-purified goat anti· human IgG (I :3,000) or IgA (I :2,000) (Tago, Burlingame, CAl for the IgG and [gA assays, respectively. After an additional five washes, 100 ILL of substrate was added to each well and the optical density (OD) read at 405 nm (reference, 610 nm) on a microELISA auto-reader (Dynatech, Chantilly, VA). In accordance with methods described by most investigators, the results were expressed as the percentage binding of a test sample against a standard laboratory positive controI. 9 •13-'7 A positive control serum was obtained from an LN patient for the AECA-IgG subclass assay and from an IgA N patient for the AECA-lgA subclass assay. Samples were recorded as being

WANG, WALKER, AND KINCAID-SMITH

380

Table 2. Associations Between the Presence of Anti-endothelial Cell Antibodies and Histologic Features LN (n AECAPositive

Patient Characteristics

Pattern MP DP Crescents> 10% Fibrin crescents FSN lesions GS> 10% IFfTA Arterial and/or arteriolar abnormality Immunofluorescence

~

18)

IgA N (n

P

AECANegative

6/12 (50%) 5/12 (42%) 3/12 (25%) 2/12 (17%) 3/12 (25%) 12/12 (100%)

AECAPositive

Value

1/6 (17%) 2/6 (33%) 1/6 (17%)

NS NS NS

1/6 (17%) 4/6 (67%) 6/6 (100%)

NS NS NS

~

36)

P

AECANegative

Value

5/13 (38%)

14/22 (61%)

NS

5/13 6/13 4/13 5/13 9/13

1/23 (4%) 2/23 (9%) 1/23 (4%) 13/23 (57%) 20/23 (87%)

0.016 0.016 0.047 NS NS

(38%) (46%) (31%) (38%) (69%)

NS 12/12 (100%) 10/13 (77%) 18/23 (78%) 6/6 (100%) The incidence of positive staining of IgG, A, M, C1q, and C3 on glomeruli was not significantly different between AECA-positive and -negative patients.

NS

Abbreviations: MP, mesangial proliferation glomerulonephritis; DP, diffuse proliferative glomerulonephritis; NS, not significant; FSN, focal and segmental necrotizing; GS, glomerular sclerosis; IF/TA, interstitial fibrosis/tubular atrophy.

positive if the percentage binding (as a proportion of the positive control) was greater than the mean ± 2 SO of the NC population.

Inhibition Studies Endothelial cell lysate, human fibroblast (HES; derived from fetal skin; Commonwealth Serum Laboratories, Victoria, Australia) lysate, lymphocyte (LC; from healthy donors) lysate, double-stranded DNA from calf thymus (Sigma, St Louis, MO), and bovine serum albumin (BSA) were used to inhibit the binding of AECA in patients' sera to ECs.

The EC, HES, and LC Iysates were prepared by solubilizing the cells with Triton X-100 solution. IS In brief, EC and HES were harvested by exposure of the cells to 0.2% EDT A in PBS for 3 minutes at 37°C. The LCs were isolated by standard methods. Cells were lysed by cooled lysis buffer (0.5% Triton X-100, 1 mmol/L EDTA, and I mmol/L phenyl-methyl-sulphonyl fluoride in PBS, pH 7.4). The protein concentration of cell lysate was measured by Micro BCA Protein Assay Reagent (Pierce, Rockford, IL). Endothelial cell lysate, HES lysate, LC lysate, DNA, and BSA were diluted in T-PBS to make different concentrations of the various inhibitors. Sera positive for both AECA-IgG and double-stranded DNA, diluted 1:200, and sera positive for AECA-IgA, diluted

p
110

,-----------,

100

100

••

90

90

BO

BQ

.t.•

70

g> -0 0::::

IlJ ~

p-O.018

110

•

60

:.• ".:.

50 40

70 C> 0::::

-0 0::::

IlJ

'.

*-

•

30

.-•hh..

NC

LN

n-45

,--1

50

"

.,

40

•••

30

10

•••• n-33

p
20

20 10

60

IgAN n-63

MPGN n-44

MGN n-14

FSHS n-13

Fig 1. Distribution of AECA-lgG binding activity (%binding) in patient groups (see text) and NCs. The broken line indicates the upper limit of the normal range.

NC

n-33

LN

n-45

IgAN n-63

MPGN n-44

MGN n-14

FSHS n-13

Fig 2. Distribution of AECA-lgA binding activity (%binding) in patient groups and NCs. The broken line indicates the upper limit of the normal range.

381

AECA IN IgA NEPHROPATHY AND LUPUS NEPHRITIS 1:50, were incubated with differing concentrations of the inhibitors for 2 hours at 4°C. After incubation, the mixtures were assayed for AECA-IgG or AECA-IgA binding activity by the same ELISA method as described above. The degree of inhibition was determined as the percentage reduction in binding after incubation with the respective inhibitors compared with incubation with T-PBS alone.

Enzyme-Linked Immunosorbent Assay on Fibroblasts Human fibroblasts were cultured in RPMI 1640 with 10% fetal calf serum and fed twice a week. The method for HES coating and fixation and ELISA procedures on HES was the same as that for ECs. Binding activities of sera on HES cells were expressed as the 00 reading at 405 nm (reference, 610 nm).

Serum Immunoglobulin G and Immunoglobulin A Concentrations Serum IgG and IgA levels were measured by T-Metric IgG and IgA kits (Silenus Laboratories, Victoria, Australia) with modification.

Clinicopathologic Associations The presence of AECA in patients with LN and IgA N were correlated with a wide range of clinical, laboratory, and renal histologic features (Tables I and 2). The only data used were those obtained within 2 weeks of the time the serum samples were collected for the AECA test.

Statistics Comparisons between study groups were carried out using the Wilcoxon rank sum test. Correlation analyses (serum IgG and IgA) were calculated using Spearman's test. Fisher's exact test was used to analyze the associations between the presence or absence of AECA and the clinical, laboratory, and histologic parameters.

RESULTS

Frequ'"ency ofAnti-endothelial Cell Antibodies Compared with NCs, AECA-IgG (Fig 1) were detected in LN (P < 0.001) and AECA-IgA (Fig 2) were detected in both LN (P < 0.001) and IgA N (P = 0.018). In LN patients, 62% (28 of 45) had an elevated AECA-IgG binding activity compared with only 6% (two of33) ofNCs. Of the glomerulonephritic controls, an elevated AECA-IgG binding activity was found in 6% (fourof63) of the IgA N patients, in 14% (six of 44) of the MPGN patients, in 7% (one of 14) of the MGN patients, and in 23% (three of 13) of the FSGS patients; the levels were not significantly different from those found in the NCs (Fig 1).

An elevated AECA-IgA binding activity was detected in 29% (13 of 45) of the LN patients and in 29% (18 of63) of the IgA N patients, compared with only 3% (one of33) of the NCs. For AECAIgA an elevated binding activity was present in only a small proportion of glomerulonephritic controls: in 11% (five of 44) of the MPGN patients, in 7% (one of 14) of the MGN patients, and in 15% (two of 13) of the FSGS patients. These levels were not significantly different from the NCs (Fig 2). In 11 LN patients, three IgA N patients, three MPGN patients, and two FSGS patients, both AECA-IgG and AECA-IgA were detected.

Specificity ofAnti-endothelial Cell Antibodies Repeated serial dilutions of multiple positive samples showed linear assay results when expressed as percentage binding (data not shown). Inhibition assay. Figure 3 shows the degree of inhibition of the different inhibitors used on AECA-IgG binding activity in the sera from five LN patients who were positive for both AECAIgG and anti-DNA antibodies. Of the inhibitors, both EC lysate and HES lysate demonstrated a higher percentage of inhibition of AECA-IgG binding compared with LC lysate, doublestranded DNA, and BSA. The EC and HES lysates also showed a higher percentage of inhibition of AECA-IgA binding compared with BSA in a serum positive for AECA-IgA (Fig 4). Enzyme-linked immunosorbent assay on fibroblasts. Higher IgG binding activity on HES was detected in 16 AECA-IgG-positive patients compared with 15 NC patients and 16 AECAIgG-negative patients (Fig 5). Higher IgA binding activity on HES also was detected in 14 AECAIgA-positive patients compared with 14 NCs and 14 AECA-IgA-negative patients (Fig 6). Serum immunoglobulin G and immunoglobulin A levels. No correlations were found between serum IgG concentrations and AECA-IgG levels or serum IgA concentrations and AECAIgA levels (data not shown).

Clinicopathologic Associations Tables 1 and 2 show the clinicopathological associations of AECA in LN and IgA N. In IgA N, significant associations were found between the presence of AECA and younger age, crescents greater than 10%, fibrin crescents, and focal and

382

A

WANG, WALKER, AND KINCAID-SMITH

B

70

c

o

60

60

50

50

40

c

40

o

......

."!::: .0

.0

-§

70

30

..s:::: 30 c

20

20

10

10

o

o .01

.1

10

.01

o

70

c::;,

1

10

HES lysate (mg/ml)

EC lysate (mg/ml)

c

.1

70

60

60

50

50

40

c

o

o

......

40

.0

.0

-§

:E 30 c

30

20

20

10

10

o

o .01

.1

1

LC lysate (mg/ml)

10

.01

.1

1

10

dsDNA (ug/ml)

Fig 3. Inhibition curves derived from five LN sera positive for AECA-lgG. Values are percentage inhibition (%inhibition) ± SEM. Inhibitors: A, EC lysate; B, HES lysate; C, LC lysate; D, double-stranded DNA (dsDNA); and E, BSA (see text).

AECA IN IgA NEPHROPATHY AND LUPUS NEPHRITIS

segmental necrotizing lesions (active IgA N). In LN, inverse associations were found between the presence of AECA and duration of disease, incidence of impaired renal function, and frequency of chronic renal failure, and positive associations were found between the presence of AECA and active lupus (Table 1), as well as antinuclear and anti-double-stranded DNA antibodies. DISCUSSION

In the present study, AECA-IgG were mainly found in patients with LN and AECA-IgA were mainly found in both LN and IgA N patients. We have previously reported in preliminary form the presence of AECA-IgA in IgA N patients and the partial characterization of some of the antigens on ECs recognized by AECA-IgA.19 However, we cannot exclude the possibility that some reactivity might be due to other factors, such as IgA rheumatoid factor interactions with AECA-IgG after these have bound to the plasticattached ECs.

E

70

60

50

40 c. 0

.....

P- 30 .s= e eft.

20

10

0 .01

.1

BSA (mg/ml) Fig 3 (cont'd).

10

383

It is widely accepted that AECAs are common in the sera from patients with systemic lupus erythematosus (SLE) and other connective tissue diseases. The frequency of detection of AECAIgG in SLE patients with nephritis in the present study is in agreement with the findings of most other studies. Shingu and Hurd20 detected AECAIgG in 50% of SLE sera from by an immunofluorescence technique, but failed to find AECAs of either IgA or IgM subclass. We have demonstrated that AECA-IgA were present in SLE sera and, although these findings differ from the observations of Shingu and Hurd 20 and Hashemi et al,21 it is possible that the differences are explained by the variations in techniques used to detect AECA-IgA. Apart from connective tissue diseases, AECAs also have been detected in both experimental and human nephritis. In 1987, Matsuo et al22 induced a model of experimental glomerulonephritis by injecting AECA directed against angiotensinconverting enzyme into rabbits. Granular deposits of these antibodies appeared on the glomerular endothelium and were associated with an increase in vascular permeability leading to "planting" of the heterologous IgG on the subepithelial part of the glomerular capillary and the pathologic features resembled those seen in the classical Heymann glomerulonephritis. In Matsuda's experiments,23 guinea pigs immunized with EC membrane products exhibited mild proteinuria. Histologically, glomeruli showed mesangial proliferation and mild mononuclear cell infiltration, and immunofluorescence staining showed finely granular, peripheral capillary loop and mesangial staining for IgG and C3. Serologically, titers of AECAs and circulating immune complexes were significantly elevated. The results suggested that antibody-mediated immune mechanisms might be important in this immunization with EC plasma membrane products model of glomerular tissue injury. In primary glomerulonephritis in humans, Yap et al ll described AECA-IgA in up to 32% ofIgA N patients; Frampton et aIlS reported similar findings. We have detected AECA-IgA in 29% of IgA patients, compared with 3% of NCs and 7% to 15% of other types of primary glomerulonephritis, findings essentially in agreement with those of Yap et alII and Frampton et alY

WANG, WALKER, AND KINCAID-SMITH

384 2.0

---EC

70

--0-

HES

--<>- BSA

0-

50 c

~

-=

1.6

•

~ 1.4

~ ." 1.2

40

<: G>

:;;

:.c

p
1.8

60

o

30

t+

1.0

as .5:! 0.8

11'

Q.

20

0

0.6 0.4

10

0.2

0 .01

.1

0.0

10

Inhlbl10r concen1ration (mu/ml)

Fig 4. Percentage inhibition of AECA-lgA binding in a single patient positive for AECA-lgA. Inhibitors: EC lysate, HES lysate, and eSA (see text).

It has been shown that resting human ECs express class I HLA antigens. 24 Anti-endothelial cell antibody binding activity therefore could be explained by antibodies directed to HLA class I determinants. II However, the results from the inhibition studies, which showed that LC lysate from several healthy donors did not significantly inhibit AECA-binding activities, make this possibility very unlikely. Furthermore, to rule out the possibility that the AECA activity we detected might represent antibodies to other antigens, such as anti-DNA and anti-ribonuclear protein antibodies (commonly detected in SLE), we chose fibroblasts and LCs as cell controls and BSA as a protein control in the inhibition studies. The results indicated that DNA did not inhibit the positive binding of test sera on ECs, although the same sera were also positive for anti-DNA antibodies. Similar observations have been reported by Rosenbaum et al. 13 As the ECs, fibroblasts, and LCs were lysed by the same method and the protein concentrations in the cell lysates were kept constant across the various inhibition assays, the inhibition produced by the EC lysate was most unlikely to be due to the nuclear components or ribonuclear proteins contained in the cell lysate; otherwise, the LC lysate would have produced an inhibition curve similar to that produced by EC lysate.

• •••

T*•

•• •

•

-rt!.:

~

NC n-15

AECA ·ve n-16

:t· •

AECA +ve n-16

Fig 5. Immunoglobulin G binding (expressed as 00 reading) on HES in NCs and in AECA-lgG-negative (AECA -vel and AECA-lgG-positive (AECA +ve) sera. The solid lines indicate the median values.

The results obtained in the inhibition studies suggested that AECA activity can be removed by EC lysate, therefore confirming the specificity of the AECAs. However, the fibroblast (HES) lysate also significantly inhibited AECA activity and, although the percentage inhibition achieved using 2.0 1.8 p
1.6

,

1.4

00

1.2

!:' 1.0 .;;; <: G>

0

a;

0.8

.!:! 0.6 Q. 0

0.4 0.2 0.0

•• * it. • NC n-14

•

•••*

•• • • ••••

•

•• •

AECA ·ve n-14

AECA +ve n-14

Fig 6. Immunoglobulin A binding (expressed as 00 reading) on HES in NCs and in AECA-lgA-negative (AECA -vel and AECA-lgA-positive (AECA +ve) sera. The solid lines indicate the median values.

AECA IN IgA NEPHROPATHY AND LUPUS NEPHRITIS

385

fibroblast lysate was less than that for EC lysate, the difference did not reach statistical significance. Furthermore, when AECA-positive sera were tested for fibroblast-binding activities, they showed higher binding activities on fibroblasts than were demonstrated in AECA-negative sera. These results lend support to the hypothesis, proposed by Rosenbaum et al 13 and Heurkens et al, 16 that AECAs represent a group of antibodies that react with surface antigens on both ECs and fibroblasts. Therefore, these cells may share certain antigens. The clinical significance of AECAs and their associations with other autoantibodies reported by some investigators are controversia1. 9 ,16,25-29 We have demonstrated significant associations between the presence of AECAs and the pathologic activity of IgA N (crescents> 10%, fibrin crescents, and focal and segmental necrotizing lesions), which suggests that AECAs may playa role (at least in part) in the pathogenesis ofIgA N. However, this study was cross-sectioned in design, and more powerful information on clinicopathologic correlations might be obtained from serial observations of AECA levels in appropriate patients. In LN, although we demonstrated an association between AECAs and anti-DNA antibodies, our inhibition studies carried out in five LN sera have not supported the hypothesis that there was cross-reactivity between AECAs and anti-DNA antibodies. Perhaps what has been suggested by these findings and the findings of others is that AECAs consist of a heterogeneous population of antibodies, in which the subpopulation cross-reacting with other antibodies varies strongly between different patients. Other investigators have tested only limited numbers of patients28 ,29 and perhaps individual patient differences account for apparently contradictory results between different studies. We also have found a significant association between the presence of AECAs and LN disease activity. By contrast, Zhou et al26 have shown no association between AECAs and disease activity. This discrepancy may again represent a slightly different patient population. Although the 21 patients studied by Zhou et al 26 had SLE, unlike the patients in the current study, they did not have renal involvement. Thus, we might speculate that the presence of AECAs, like anti-DNA,

may reflect disease activity ofLN and that crossreactivity with ANA and anti-DNA is possible. In conclusion, this study has demonstrated the presence of AECAs in LN and IgA N patients. Endothelial cells may share some common antigens with fibroblasts. Anti-endothelial cell antibodies, like anti-DNA, may reflect disease activity of LN and there may be cross-reactivity with anti-nuclear and anti-DNA antibodies. The AECAs tended to be present in the earlier, active stages of disease and may presumably playa role in the pathogenesis ofIgA Nand LN.

REFERENCES 1. Lindquist RJ, Osterland CK: Human antibodies to vascular endothelium. Clin Exp ImmunoI9:753-760, 1971 2. Tan EM, Pearson CM: Rheumatic disease sera reactive with capillaries in the mouse kidney. Arthritis Rheum 15:2328,1972 3. Cerilli J, Holliday JE, Fesperman DP, Folger MA: Antivascular endothelial cell antibody-Its role in transplantation. Surgery 81:132-138, 1977 4. Burns ER, Zucker-Franklin D: Pathologic effects of plasma from patients with thrombotic thrombocytopenic purpura on platelets and cultured vascular endothelial cells. Blood 60:1030-1037, 1982 5. Penning CA, Cunningham J, French MAH, Harrison G, Rowell NR, Hughes P: Antibody-dependent cytotoxicity of human vascular endothelium in systemic sclerosis. Clin Exp ImmunoI57:548-556, 1984 6. Leung DYM, Moake JL, Havens PL, Kim M, Prober JS: Lytic antiendothelial cell antibodies in haemolytic uraemic syndrome. Lancet 2: 183-186, 1988 7. Rappaport VJ, Hirata G, Yap HK, Jordan SC: Antivascular endothelial cell antibodies in severe preeclampsia. Am J Obstet GynecoI162:138-146, 1990 8. Fattorossi A, Aurbach GD, Sakaguchi K, Cama A, Marx SJ, Streeten EA, Fitzpatrick LA, Brand ML: Antiendothelial cell antibodies: Detection and characterization in sera from patients with hypoparathyroidism. Proc Nat! Acad Sci USA 85:4015-4019, 1988 9. Frampton G, Jayne DRW, Perry GJ, Lockwood CM, Cameron JS: Auto-antibodies to endothelial cells and neutrophil cytoplasmic antigens in systemic vasculitis. Clin Exp ImmunoI82:227-232, 1990 10. Salant DJ: Immunopathogenesis of crescendi glomerulonephritis and lung purpura. Kidney Int 32:408-425, 1987 II. Yap HK, Sakai RS, Bahn L, Rappaport V, Woo KT, Ananthuraman V, Lim CH, Chiang GSC, Jordan SC: Antivascular endothelial cell antibodies in patients with IgA nephropathy: Frequency and clinical significance. Clin Immunol ImmunopathoI49:450-453, 1988 12. Vismara A, Meroni PL, Tincani A, Harris EN, Barcellini W, Brucato A, Khamashta M, Hughes GRV, Zanussi C, Balestriere G: Relationship between anti-cardiolipin and anti-endothelial cell antibodies in systemic lupus erythematosus. Clin Exp Immunol 74:247-253, 1988 13. Rosenbaum J, Pottinger BE, Woo P, Black CM, Loizou S, Byron MA, Pearson D: Measurement and characterisation

386 of circulating anti-endothelial cell IgG in connective tissue diseases. Clin Exp Immunol 72:450-456, 1988 14. Van Der Zee JM, Siegert CEH, De Vreede TA, Daha MR, Breedveld FC: Characterization of anti-endothelial cell antibodies in systemic lupus erythematosus (SLE). Clin Exp Immunol 84:238-244, 1991 15. Frampton G, Walker RG, Perry GJ, Harada T, Cameron JS: IgA affinity to ssDNA or endothelial cells and its deposition in glomerular capillary walls in IgA nephropathy. Nephrol Dial Transplant 5:841-846, 1990 16. Heurkens AHM, Hiemstra PS, Lafeber GJM, Daha MK, Breedveld FC: Antiendothelial antibodies in patients with rheumatoid arthritis complicated by vasculitis. Clin Exp Immunol 78:7-12, 1989 17. Savage COS, Pottinger B, Gaskin G, Lockwood CM, Pusey CD, Pearson JD: Vascular damage in Wegener's granulomatosis and microscopic polyarteritis: Presence of antiendothelial cell antibodies and their relation to anti-neutrophil cytoplasm antibodies. Clin Exp Immunol 85:14-19, 1991 18. Johnstone A, Thorpe R: Isolation fractionation of lymphocytes. Solubilisation of cells and membranes, in Immunochemistry in Practice (ed 2). London, UK, Blackwell, 1987, p 105 19. Wang M-X, Walker RG, Kincaid-Smith P: Endothelial cell antigens recognized by 19A auto-antibodies in patients with IgA nephropathy: Partial characterisation. Nephrol Dial Transplant 7:805-810, 1992 20. Shingu M, Hurd ER: Sera from patients with systemic lupus erythematosus react to human endothelial cellS'. J Rheumatol 8:581-586, 1981 21. Hashemi S, Smith CD, Izaquirre CA: Antiendothelial cell antibodies: Detection and characterization using a cellular enzyme linked immunoabsorbent assay. J Lab Clin Med 109: 435-440, 1987

WANG, WALKER, AND KINCAID-SMITH

22. Matsuo S, Fukatsu A, Taub ML, Caldwell PRB, Brentjens JR, Andres G: Glomerulonephritis induced in the rabbit by anti-endothelial antibodies. J Clin Invest 79: 17981811,1987 23. Matsuda M: Experimental glomerular tissue injury induced by immunization with cultured endothelial cell plasma membrane. Acta Pathol Jpn 38:823-839, 1988 24. Moraes JR, Stastny P: A new antigen system expressed in human endothelial cells. J Clin Invest 60:449-454, 1977 25. Harris EN, Gharavi AE, Tincani A, Chan JKH, Englert H, Mantelli P, Allegro F, Ballestrieri G, Hughes GRV: Affinity purified anti-cardiolipin and anti-DNA antibodies. J Clin Lab ImmunoI17:155-162,1985 26. Zhou SF, Wang XM, Zhong HP, Li A, Wang MX, Wang SL, Zhang L, Du XH: Pathogenically role of anti-vascular endothelial cell antibodies in systemic lupus erythematosus. Chinese Med J 102:505-509, 1989 27. Homeland M, Rekvig OP, Jorgensen L: Cultured human endothelial cells display an antigen that is recognized by certain human antichromatic auto-antibodies. Clin Exp Immunol 54:373-377, 1983 28. Hasselaar P, Derksen RHWM, Blokzijl L, de Groot PG: Cross-reactivity of antibodies directed against cardiolipin, DNA, endothelial cells and blood platelets. Thromb Haemost 63:169-173,1990 29. Faaber P, Rijke TPM, van de Putte LBA, Capel PJA, Berden JHM: Cross-reactivity of human and murine antiDNA antibodies with heparan sulfate. J Clin Invest 77:18241830, 1986 30. Correia P, Cameron JS, Ogg CS, Williams DG, Bewick M, Hicks A: End-stage renal failure in systemic lupus erythematosus with nephritis. Clin Nephrol 22:293-302, 1984 31. Droz D: IgA nephropathy: Clinicopathologic correlations, in Clarkson AR (ed): 19A Nephropathy. Boston, MA, Martinus Nijhoff, 1987, p 98