Effects of Anti-Endothelial Cell Antibodies in Leprosy and Malaria

Abstract

As a result of damaging endothelial cells (ECs), Mycobacterium leprae triggers the production of antibodies (Abs). These anti-EC Abs (AECAs) can be divided into two types. The first type nonspecifically reacts with components of the cytosol (CY) and can be detected by enzyme-linked immunosorbent assay (ELISA). The second specifically reacts with the EC membrane (MB) and requires fluorescence-activated cell sorter (FACS) analysis to be detected. The presence of both types of AECAs was determined in 68 leprosy patients. The ELISA was positive for 35 of them but also for 30 of 34 malaria patients and 17 of 50 healthy African controls. However, whereas FACS analysis showed MB reactivity in only three malaria patients and four controls, this reactivity was found in 27 leprosy patients, more of those having the lepromatous than the tuberculoid form. Specificity for MB, which we failed to absorb by incubation with CY lysates, predominated over that for CY in leprosy, unlike malaria, where the EC reactivity was restricted to the CY. Western blot analysis and two-dimensional electrophoresis revealed that calreticulin, vimentin, tubulin, and heat shock protein 70 were targeted by AECAs from leprosy patients, but other proteins remained unidentified. These auto-Abs, but not those from malaria patients, did activate ECs, as indicated by the E-selectin and intercellular adhesion molecule 1 upregulation, and/or induced them into apoptosis, as documented by four different methods. Our findings suggest that, in some but not all leprosy patients, AECAs may play a role in pathogenesis.

Leprosy is caused by intracellular infection with Mycobacterium leprae. This chronic disease encompasses various clinical presentations (28) ranging between tuberculoid leprosy (TL) and lepromatous leprosy (LL). Between these polar forms, immunologically overlapping cases are classified as borderline tuberculoid (BT), midborderline (BB), and borderline lepromatous (BL).

TL is characterized by a vigorous T-helper (Th) type 1 protective response reducing bacteria to paucity in tissues, whereas LL mounts such an ineffectual cell-mediated immune defense that an abundance of bacilli remains detectable within the lesions. This Th type 2 form of the disease initiates, however, a prominent antibody (Ab) response to the M. leprae antigens (Ags), along with hyperglobulinemia, immune complexes, and a flurry of auto-Abs. They include rheumatoid factor (24) and antinuclear (22), antiphospholipid (2), antineutrophil cytoplasmic (19), and antimitochondrial (10) auto-Abs. Since none of them has unequivocally been proven to generate autoimmune complications in leprosy, it has been tempting to incriminate polyclonal activation of B lymphocytes, rather than specific Ag stimulation, in their appearance.

Although colonization of endothelial cells (ECs), most notably those lining epineurial and perineurial blood vessels, by M. leprae has long been recognized (9), the integration of this process into a model of the mechanisms by which ECs contribute to the development of the disease is new (31). Given that these bacteria reside and multiply inside ECs, immune reactivity to these cells, which has never been previously appreciated in leprosy, warrants being set apart from other auto-Abs found in this disease. Furthermore, due to vascular injury, target Ags for anti-EC Abs (AECAs) may indeed be engendered anew, and cryptic Ags may be exposed and then expressed or released (15), thereby becoming immunogenic.

Insights into the production and clinical relevance of AECAs are only beginning. The diversity of conditions associated with them (38) is so extensive that AECAs represent an extremely heterogeneous family of auto-Abs (21). Thus, their presence does not even imply a causal relationship with any condition. Indeed, the production of AECAs may follow, rather than precede, EC damage, and attempts to demonstrate their pathogenicity have had mixed results (21, 38, 39). A recent experimental model of systemic vasculitis has, based on auto-Ab idiotype, provided compelling evidence suggesting that some AECAs are pathogenic (5). Current efforts have focused on the EC activation of type II, which would be elicited by this or another group of AECAs (4). Evidence for such an activation includes upregulation of adhesion molecules, e.g., E-selectin and intercellular adhesion molecule 1 (ICAM-1). In this respect, it is important that the level of circulating ICAM-1 is elevated in leprosy patients (29). In addition, recent studies have shown that some AECAs are capable of inducing apoptosis in ECs (3).

All in all, the above-cited observations support the contention that AECAs may be influential in the pathophysiology of leprosy, depending on their specificity. To be pathogenic, AECAs should bind to structures expressed on the membrane (MB) of ECs, rather than penetrate through the MB, and encounter candidate Ags in the cytosol (CY). This is the case in non-organ-specific autoimmune diseases, particularly systemic lupus erythematosus (SLE), where, in addition to Ag-driven AECAs, auto-Abs may be generated by various components of the CY that are present in all the cells. There are no reasons for ECs to be specific for their CY. Hence, the production of this fraction of the auto-Ab population may be due simply to polyclonal B-cell activation, as stated previously, not only in SLE but, importantly, also in malaria (1). Malaria patients were actually selected as disease controls in this study, due to the high prevalence of auto-Abs (6) in this infectious disease.

This is the standpoint from which we attempted to determine whether one form of leprosy was associated with pathogenic AECAs, as opposed to malaria. We present evidence herein that AECAs in leprosy, where anti-MB AECAs predominate, are distinct from those in malaria, where all of them recognize CY determinants, and constitute purely and simply a marker of the disease. In contrast, anti-MB AECAs might play a role in the pathogenesis of some, but not all, cases of leprosy in view of their activating effect on the cells and the possibility of inducing them into apoptosis.

MATERIALS AND METHODS

Patients and controls.

Sixty-eight consecutively registered leprosy patients (37 males and 31 females, aged 30.0 ± 15.3 years) were drawn from the Institut de Léprologie Appliquée in Dakar, Senegal. The diagnostic criteria for this disease were standard clinical signs. Biopsy samples were taken from active lesions to confirm leprosy and determine the bacterial index. Thirty-four patients each belonged to the multibacillary and paucibacillary groups. At the time of inclusion in our study, 56 of them were classified into 18 LL, 8 BL, 6 BB, 18 BT, and 6 TL patients, but the remaining 12 cases could not be classified. These patients participated in a therapeutic trial of a monthly dose of 600 mg of rifampin, 400 mg of ofloxacin, and 100 mg of minocycline (18). Sera had, however, been collected before the initiation of their antimycobacterial treatment and stored at −20°C until use. None of the patients showed clinical symptoms of autoimmune diseases. Sera with anti-human immunodeficiency virus Abs were excluded from the study, and in this series, there were only one serum sample with Abs to hepatitis C virus and eight with Abs to HBs Ag (25).

Thirty-four patients (19 males and 15 females, aged 30.5 ± 15.4 years) with chronic malaria, as defined by the presence of Plasmodium falciparum on a thick blood film, were also enrolled in the study as disease controls. Their sera served to distinguish AECAs specific for leprosy from those associated with other infectious diseases. Fifty African blood donors (31 males and 19 females, aged 31.2 ± 18.7 years), with no known exposure to leprosy, were selected as healthy controls to assign AECAs to leprosy, rather than merely ascribe them to chronic infections endemic in the area. Leprosy patient contacts were, therefore, systematically excluded. Serum samples from one AECA-positive patient fulfilling four American College of Rheumatology criteria for SLE (34) were used as reference patterns in the two-dimensional (2D) electrophoresis of Ags targeted by AECAs from leprosy and malaria. Informed consent was received from study patients and disease controls for the therapeutic trial and blood withdrawal for this particular study.

To set the threshold for positive AECAs in the cell enzyme-linked immunosorbent assay (ELISA), we selected 81 European AECA-negative sera among those forwarded to our laboratory to be routinely tested in this ELISA for the presence of AECAs. None of them appeared to be positive for an AECA-associated disease (27). Additional sera from 13 volunteers among workers in the laboratory and from clinical staff served as healthy controls to set the cutoff level in the fluorescence-activated cell sorter (FACS) analysis.

Cell culture.

Cells from the EA.hy926 cell line (kindly donated by Cora-Jean S. Edgell, University of North Carolina, Chapel Hill) were employed as substrate cells in the ELISA for the detection of AECAs (7). This cell line was originally derived by fusing human umbilical vein ECs (HUVECs) with human lung epithelioma A549/8 cells. The EA.hy926 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS), 2 μM glutamine, 100 μM hypoxanthine, 0.4 μM aminopterin, and 16 μM thymidine (Sigma Chemical Co., St. Louis, Mo.). The A549/8 cells (gift of Cora-Jean S. Edgell) were cultured in the presence of 0.1 μM 6-thioguanine in DMEM supplemented with 10% FCS. The cells were harvested using 0.25% trypsin-EDTA, transferred to 96-well microtiter plates (Nunc, Roskilde, Denmark) precoated with 1% fibronectin, and cultured to confluence. They were monitored by phase-contrast microscopy to ensure confluency throughout the procedure.

Human ECs were dissociated from umbilical veins by 0.1% collagenase (Sigma) digestion, as previously described (13), identified as such through the expression of von Willebrand factor, and grown to confluence in RPMI 1640 (Gibco, Paisley, Scotland) containing 10% FCS, 2 mM glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin/ml. HUVECs were then passaged twice, harvested, and suspended at a concentration of 10⁴ cells in 100 μl of DMEM supplemented with 10% FCS and 100 IU of polymyxin B/ml.

The human epithelial line 2 (HEp-2) cells (American Type Culture Collection, Manassas, Va.) were cultured in DMEM supplemented with 2 ml l-glutamine, 100 U of penicillin/ml, 100 μg of streptomycin/ml, and 10% FCS. These HEp-2 cells were those used routinely for the detection of antinuclear Abs (ANAs) in our laboratory.

AECA tests.

The cell ELISA for AECAs has been described previously (3, 4, 26, 27). Briefly, EA.hy926 cells were fixed with 1% glutaraldehyde for 10 min at 4°C. Nonspecific binding was blocked with 200 μl of RPMI 1640 supplemented with 3% bovine serum albumin (BSA) by incubation for 2 h at room temperature (RT). After three washes with phosphate-buffered saline (PBS) supplemented with 3% BSA (PBS-BSA), 100 μl of the coded sera was added in triplicate to the coated wells. Samples were diluted 1:100 in PBS supplemented with 10% FCS to minimize artifactual recognition by heterophile Abs to BSA nonspecifically retained by ECs (26). After a 2-h incubation at 37°C and another three washes with PBS-BSA, horseradish peroxidase (HRP)-conjugated F(ab′)₂ anti-human immunoglobulin G (IgG; Jackson ImmunoResearch Laboratories, West Grove, Pa.) was added and incubated for 1 h at 37°C. Standard curves were constructed with positive internal reference sera.

In view of the known limitation of the ELISA for AECAs (39), all the sera were also evaluated by FACS analysis, with 2 × 10⁵ unfixed HUVECs as substrate cells and fluorescein isothiocyanate (FITC)-conjugated rabbit F(ab′)₂ anti-human IgG (Dakopatts, Glostrup, Denmark) as the revealing agent. After passing through a nylon mesh filter, stained ECs were enumerated in an ELITE flow cytometer (Coulter, Hialeah, Fla.). Appropriate settings of forward and side scatter gates were used to examine 10,000 cells per experiment. The FACS settings selected were controlled by using unstained cells and the FITC-conjugated developing reagent alone.

In both of the AECA assays, the results were expressed as AECA binding indices (BIs), which were calculated from 100 × (S − A)/(B − A), where S is the optical density (OD) of the test serum and A and B are those of a negative and a positive control, respectively (30). Samples were recorded as positive if the BI was greater than the mean of 81 and 13 healthy European sera for the ELISA and the FACS analysis, respectively, plus 3 standard deviations (SDs). This definition of positivity had been suggested by the participants in the AECA workshop (39), given that the distributions of the data were roughly unimodal (26).

Detection of auto-Abs to endothelial CY.

To distinguish the specific binding to the MB from nonspecific Abs directed to endothelial CY structures, ECs and A549/8 cells were permeabilized by adding 0.5% saponin (Sigma) to their culture medium. Eighteen sera from leprosy patients and 24 sera from malaria patients were evaluated in the FACS assay before and after permeabilization of the cells. AECAs detected before permeabilization are seemingly reactive with MB motifs, while those detected after this treatment combine some of these particular AECAs with others directed towards CY components.

Sera were referred to as CY⁺ when positive in the ELISA and MB⁺ when positive in the FACS assay. They could thus be categorized on the basis of positivity for these two assays into CY⁺ MB⁺, CY⁻ MB⁺, and CY⁺ MB⁻ sera.

Specificity of AECAs for CY components, either exclusive (CY⁺ MB⁻ sera) or in addition to their specificity for MB structures (CY⁺ MB⁺ sera), was also determined by indirect immunofluorescence (IIF) with HEp-2 cells plated on a glass slide as the substrate. The slides were examined by confocal microscopy (Leica Microsystems, Wetzlar, Germany).

Two CY⁺ MB⁺ and two CY⁻ MB⁺ sera from leprosy patients and two CY⁺ MB⁻ sera from malaria patients were selected at random for inhibition experiments. Once diluted down to 50% of their maximal activity in the ELISA, they were distributed into aliquots, triplicates of which were incubated with 0 to 5 mg of EA.hy926 CY preparation/ml for 18 h at 4°C. After centrifugation, supernatants were tested again in the ELISA, and percentages of inhibition were calculated according to the formula 100 × (OD of the test serum − OD of the inhibited serum)/OD of the test serum.

CY and MB extracts were concomitantly subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gradient electrophoresis and transferred onto polyvinylidene difluoride (PVDF) MBs (Amersham Pharmacia Biotech, Orsay, France), with a semidry blotting chamber (Bio-Rad Laboratories, Hercules, Calif.) and according to the protocol that we have described previously (27). They were then probed with the same six sera as in the ELISA inhibition experiments, before and after they had been incubated with 1 mg of CY extract/ml, as described above.

Western blotting.

CY and MB from EA.hy926 cells were dissociated according to the method described by Pasquali et al. (23). The cells were removed from culture flasks by using a cell scraper and resuspended in a homogenization buffer (1 M sucrose, 100 mM Tris-HCl, 100 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 100 mM KCl, 50 mM MgCl₂, 1 μM leupeptin, 1 μM pepstatin, and 1 μM aprotinin). They were then subjected to two freezing-thawing cycles in liquid nitrogen and sonicated three times for 15 s each time. Nuclei were removed by centrifugation at 1,000 × g for 10 min.

Supernatants were ultracentrifuged at 100,000 × g for 30 min, and CY fractions were collected and stored at −70°C until use. The MB-enriched extract-containing pellets were suspended in solubilization buffer containing 7 M urea, 2 M thiourea, 100 mM dithiothreitol (DTT), and 4% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate. They were left rocking overnight at 4°C. Solubilized MB proteins were recovered by centrifugation at 11,000 × g, and their concentration was determined by the Micro BCA protein assay kit (Pierce, Rockford, Ill.).

The lysates were subjected to SDS-5 to 15% polyacrylamide gel electrophoresis and transferred onto PVDF MBs. These were blocked with 2% nonfat milk in 0.05% Tween 20 by incubation for 2 h and Western blotted overnight with sera diluted 1:50 in PBS containing 1% nonfat milk in 0.05% Tween 20. Biotinylated goat F(ab′)₂ anti-human IgG (Beckman Coulter, Villepinte, France) was then added onto the MBs, with a second layer of HRP-conjugated streptavidin (Dakopatts) developed using the diaminobenzidine detection system (Sigma).

The pattern of reactivity of each serum was determined with a Bio-Rad model G-700 imaging densitometer and the Molecular Analyst software. The molecular weight of the migrated proteins was evaluated, and the ODs of those recognized by AECAs were recorded.

2D immunoblotting.

For isoelectric focusing, the protein samples were loaded onto nonlinear pH 3 to 10 immobilized pH gradient strips (Pharmacia, Uppsala, Sweden) and focused for 30,000 V · h at RT. The strips were incubated for 15 min at RT with 50 ml of a primary equilibration solution. This contained 0.375 M Tris-HCl (pH 8.8), 6 M urea, 2% SDS, 35% glycerol, and 70 mM DTT. They were then incubated for 30 min at RT with a second equilibration solution, which was identical to the first, except that 70 mM DTT was replaced with 135 mM iodoacetamide (Sigma). In-house 5 to 15% acrylamide-bisacrylamide gradient gels were used to resolve the proteins, based on their molecular weight. The strips were loaded onto the top of the slab gel, electrophoresed for 4 h at 30 mA/gel, and transferred to PVDF MBs. At the end of their electrophoresis run, these were probed with AECA-positive sera at a dilution of 1:50, and the blots were probed with biotinylated anti-human IgG Abs (Jackson ImmunoResearch), with a second layer of HRP-labeled streptavidin (Amersham Pharmacia Biotech). Auto-Ag spots were identified in reference to the EA.hy926 2D gel map from the Swiss-Prot database (http://www.harefield.nthames.nhs.uk/nhli/protein/eahy).

Other auto-Ab tests.

ANAs were determined by a standard IIF method with ethanol-fixed HEp-2 cells as the substrate. The IgG-specific ELISA for antiphospholipid Abs (aPLs), with the use of plates coated with 50 μg of cardiolipin (Sigma)/ml, has been thoroughly described elsewhere (2).

Induction and assessment of activation.

Sera from eight leprosy patients, three malaria patients, and five AECA-negative European controls were passed through a protein G-Sepharose column (Pharmacia), and IgG fractions were eluted with HCl-glycine, pH 2.8. The resulting IgG concentration was measured by an in-house sandwich ELISA, and concentrations were made the same in all preparations.

HUVECs were distributed into four triplicates of wells of a 96-well microtiter plate for AECA IgG-induced E-selectin, non-AECA IgG (as a control for the AECA induction of E-selectin), AECA IgG-induced ICAM-1, and non-AECA IgG (as a control for the AECA induction of ICAM-1), respectively. AECA and non-AECA IgG were adjusted to a concentration of 320 μg/ml in the culture medium of the four triplicates. One hundred microliters of the preparations was poured into each well to reach a final concentration of 160 μg of IgG/ml. On each plate, three separate wells containing 10 IU of interleukin-1β (IL-1β; Genzyme, Cambridge, Mass.) in 100 μl of medium served as internal controls for activation. After a 6-h incubation at 37°C, the cells were harvested using 0.25% trypsin-EDTA. Those from the three wells of each triplicate were pooled and washed with PBS-BSA.

We ended up with two pairs of cell pools: the first pool of each pair contained AECA IgG and the second contained non-AECA IgG. One pair was incubated with saturating concentrations of FITC-conjugated anti-ICAM-1 monoclonal Ab (MAb), and the other was incubated with unconjugated anti-E-selectin MAb (both from Beckman Coulter), followed by FITC-conjugated goat F(ab′)₂ anti-mouse IgG (Tago, Burlingame, Calif.). After two final rinses with PBS-BSA, all cell preparations were analyzed by FACS assay. An irrelevant FITC-conjugated mouse IgG1 MAb was used as a negative control (Beckman Coulter). ICAM-1 and E-selectin Ag densities were measured by assessing the mean fluorescence intensity (MFI) of cells analyzed in each test. The results were expressed as an MFI variation which was equal to 100 × (MFI of AECA IgG-treated cells − MFI of non-AECA IgG-treated cells)/non-AECA IgG-treated cells.

Identification of apoptotic cells.

Time course and dose-effect curve experiments were carried out to determine the optimal conditions for AECA-induced apoptosis, as reported previously (3). Following a 24-h incubation of ECs with 320 μg of AECA IgG or non-AECA IgG/ml, apoptosis was documented by four methods. We first examined the morphology of apoptotic cells. Triplicates of EC suspensions were cytocentrifuged for 1 min at 300 × g onto microscope slides and stained in a May-Grünwald-Giemsa solution (Merck, Darmstadt, Germany). Percentages of cells exhibiting a morphology typical of apoptosis were evaluated, and the cutoff level was set at 14% of the cells (mean + 3 SDs of five AECA-negative European sera).

As a second method, the phosphatidylserine (PS) translocation to the outer face of the MB was established through the binding of annexin V coupled with FITC (Beckman Coulter). Propidium iodide (PI) diluted to 10 μg/ml enabled us to exclude dead cells. Percentages of annexin V-positive cells were thus calculated within the PI-negative population of cells, and the cutoff level was set at 22% (mean + 3 SDs of five AECA-negative European sera).

Enumeration of hypoploid cells was the third method employed. The cell suspension was washed in citrate buffer (0.1 M sodium citrate and 0.1% Triton X-100) and incubated in 250 μl of citrate buffer containing 10 μg of PI/ml overnight at 4°C in the dark. Reduction in PI staining intensity, compared with that of control cells determined by FACS analysis, was taken as a measure of hypoploidy. Again, the threshold was set at the mean ± 3 SDs of five AECA-negative European sera, i.e., 13%.

The fourth method used was DNA fragmentation analysis. ECs were washed three times in Hanks' balanced salt solution (Eurobio, Les Ulis, France) and once in lysis buffer (250 mM sucrose, 50 mM Tris [pH 7.5], 25 mM KCl, and 5 mM MgCl₂). Cell pellets were resuspended and incubated for 8 min on ice in a solution of 0.25% Triton X-100 (Sigma) added to 500 μl of lysis buffer. After centrifugation for 5 min at 500 × g and 4°C, nuclei were resuspended in 500 ml of lysis buffer supplemented with 25 μl of 0.5 M EDTA, 70 μl of 10% SDS, and 0.2 mg of proteinase K (all from Sigma) and incubated for 3 h at 37°C. The DNA was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), followed by two extractions with chloroform (vol/vol). After 5 min of washing and centrifugation at 400 × g and 4°C, a 1:10 volume of sodium acetate, followed by 2 volumes of absolute ethanol, was added. DNA was kept for 12 h at 4°C, centrifuged for 10 min at 450 × g, dissolved in 10 mM Tris (pH 7.5)-1 mM EDTA (pH 8) for 12 h after evaporation of ethanol, loaded into wells of a 1.5% agarose gel, electrophoresed at 75 mV using 100 mM Tris-100 mM boric acid-0.2 mM EDTA as running buffer, and visualized by ethidium bromide staining.

Statistical analysis.

All the results quoted below are means ± SDs. Comparisons were made using the chi-square test, with Yates' correction when required, and the Mann-Whitney U test for unpaired data.

RESULTS

Anti-EC activity.

The mean BIs of AECAs in the ELISA were 29.3% ± 41.3% for the leprosy patients, 32.7% ± 26.4% for the control African cohorts (nonsignificant [NS] compared with the leprosy patients), and 50.5% ± 27.6% for the malaria patients (P < 0.01, compared with the leprosy patients and the control African cohort). Based on the cutoff value of 20% (mean value for the European controls + 3 SDs), positive AECAs were found (Table 1) in 35 leprosy patients (51.5%), 17 control African cohorts (34.0%; NS compared with the leprosy patients), and 30 malaria patients (88.2%; P < 10⁻⁴ compared with the leprosy patients and with the control African cohort).

TABLE 1.

Detection of AECAs by ELISA and by FACS analysis in leprosy patients, compared with malaria patients and healthy African controls

Sample group	No. (%) positive
	ELISA	FACS
Leprosy patients (n = 68)	35 (51.5)a	28 (41.2)b
Malaria patients, (n = 34)	30 (88.2)	3 (8.8)
Healthy African controls (n = 50)	17 (34.0)	4 (8.0)

^aP < 10⁻³ compared with malaria patients and NS compared with healthy controls.

^bP < 10⁻⁴ compared with malaria patients and with healthy controls.

Because ECs are permeabilized by fixation with glutaraldehyde in the ELISA, auto-Abs to non-EC-specific CY components may also be detected in this assay. The sera were therefore evaluated by FACS analysis (36). By this assay (Table 1), reactivity was found in 28 leprosy patients (41.2%), four African controls (8%; P < 10⁴ compared with leprosy patients), and three malaria patients (8.8%; P < 10⁴ compared with leprosy patients, and NS compared with the control African cohort). In leprosy patients, more of the group with the LL and BL forms than of that with the TL and BT forms scored positive (15 of 26, compared with 6 of 24; P < 0.02). The results of the former group were comparable to those obtained in the patients with the BB form (one of six) and in the nonclassifiable patients (6 of 12). Interestingly, ANAs were associated with AECAs in leprosy (11 of 28 AECA⁺ sera, compared with 6 of 40 AECA⁻ sera; P < 0.03), while aPLs were only occasionally detected (two sera in the AECA⁺ group and none in the AECA⁻ group). Table 2 summarizes the results of ELISA and FACS analysis according to each subgroup of leprosy patients.

TABLE 2.

Detection of AECAs using an ELISA and by FACS analysis in 68 patients according to each subgroup of leprosy

Subgroup of patients	No. (%) positive
	ELISA	FACS
LL (n = 18)	13 (72.2)	10 (55.6)
BL (n = 8)	4 (50.0)	5 (62.5)
BB (n = 6)	2 (33.3)	1 (16.7)
BT (n = 18)	5 (27.8)	4 (22.2)
TL (n = 6)	2 (33.3)	2 (33.3)
Unclassified (n = 12)	9 (75.0)	6 (50.0)
Total (n = 68)	35 (51.5)	28 (41.2)

Overall, the truly specific auto-Abs for MB of ECs were preferentially associated with the multibacillary group and the lepromatous form of leprosy. In contrast, auto-Abs exclusively specific for the CY characterized malaria.

Analysis of AECAs.

Sera found positive in the ELISA were referred to as CY⁺, and those found positive in the FACS assay were referred to as MB⁺. Thus, there appear to be CY⁺ MB⁺, CY⁺ MB⁻, and CY⁻ MB⁺ auto-Ab profiles in the two groups of patients, as well as the control African cohort (Table 3). However, reactivity in both assays was detected in 27 leprosy sera (39.7%), compared with as few as three malaria sera (8.8%) and four healthy control sera (8.0%). The low percentage of CY⁺ MB⁻ patterns in the lepromatous sera relative to that in malarian and healthy sera (10.3% versus 79.4%, P < 10⁻⁹, and versus 20.0%, P < 0.03, respectively) suggests that the majority of leprosy patients display a first group of AECAs which bind to the MB and a second group which bind to the CY, whereas AECAs from the malaria patients recognize CY Ags almost exclusively.

TABLE 3.

Relationship between detection of AECAs measured by ELISA and by FACS analysis in the sera of leprosy patients, malaria patients, and healthy African controls

ELISA/FACS result	No. of individuals
	Leprosy	Malaria	Controls
+/+	27	3	4
+/−	7	27	13
−/+	1	0	0
−/−	33	4	33
Total	68	34	50

This interpretation was substantiated by our finding that FACS analysis of 24 of the 27 sera from the malaria patients that were positive in the ELISA (CY⁺) but negative in the FACS assay (MB⁻) turned positive in the latter assay, once ECs had been permeabilized with saponin (top row in Fig. 1A), compared with four of the seven sera from the malaria patients and 12 of the 13 African controls. Furthermore, all 24 malaria sera, 4 leprosy sera, and 12 control sera scored positive when tested with permeabilized A549/6 cells in the FACS assay, as well as with permeabilized HEp-2 cells in the IIF test (top right panel in Fig. 1B). Clearly, the target Ags of these pseudo-AECAs are not specific for ECs. With regard to the leprosy patients, two of the seven sera positive in the ELISA (CY⁺) but negative in the FACS assay (MB⁻) became positive in the FACS assay (middle row in Fig. 1A and middle right panel in Fig. 1B), while 3 of the 10 sera positive in the ELISA (CY⁺) and the FACS assay (MB⁺) were confirmed to contain two fractions in the population of auto-Abs. One fraction included auto-Abs which bound to the MB, and another fraction included auto-Abs which bound to the CY (bottom row in Fig. 1A and bottom right panel in Fig. 1B).

FIG. 1.

Detection of AECAs in three representative sera by ELISA and FACS analysis. (A) FACS analysis was employed to detect AECAs with FITC-conjugated rabbit F(ab′)₂ anti-human IgG as the revealing agent. HUVECs were used live or permeabilized with saponin as a first substrate (left), and A549/6 epithelial cells were used as an alternative substrate (right). From top to bottom are sera from one malaria patient positive by ELISA (which detects anti-CY Abs) and negative by FACS analysis (which detects anti-MB Abs), one leprosy patient negative by ELISA and positive by FACS analysis, and another leprosy patient positive by both ELISA and FACS analysis, respectively. Dotted lines indicate the control fluorescence obtained with FITC-conjugated rabbit F(ab′)₂ anti-human IgG only. (B) HEp-2 cells served as the substrate, the same three sera as in panel A were evaluated, and their Ab binding was revealed by IIF.

To confirm the predominant specificity of AECAs for MB in leprosy patients, as opposed to their exclusive specificity for CY in malaria, two CY⁺ MB⁺ and two CY⁻ MB⁺ sera from leprosy patients and two CY⁺ MB⁻ sera from malaria patients were incubated with increasing amounts of CY extracts before being retested for AECAs in the ELISA (Fig. 2A). Most of the reactivity of the sera from malaria was absorbed in a dose-dependent manner, whereas again that of leprosy sera showed two fractions of auto-Abs. This was evident from the observation that the reactivity of CY⁺ MB⁺ lepromatous sera was partially reduced, with the inhibition reaching a plateau despite the increased concentration of the CY extract. In contrast, CY⁻ MB⁺ lepromatous sera were unaffected by incubation with the CY extracts. These same six sera were also analyzed by Western blotting (Fig. 2B), before and after incubation with the CY extracts. The 122-, 80-, 65-, 39-, 26-, and 22-kDa bands recognized by Abs in the CY⁺ MB⁻ sera from malaria patients were absorbed, and their binding to 50- and 47-kDa bands was weakened. With regard to the two CY⁺ MB⁺ sera from leprosy patients, the 65-kDa band disappeared, while reactivity with the 91-kDa band was reduced. In contrast, the Ag recognition profile of two CY⁻ MB⁺ sera was unaffected by incubation with CY extracts.

FIG. 2.

Six AECA-containing sera were absorbed by incubation with increasing amounts of CY extract: two CY⁺ MB⁺ and two CY⁻ MB⁺ sera from leprosy patients and two CY⁺ MB⁻ sera from malaria patients (see Fig. 1 legend for definitions). (A) AECAs were seen by ELISA. (B) AECAs from the same sera were analyzed by Western blotting before and after incubation with 1 mg of CY extract/ml. Numbers to the right of each subpanel are molecular masses in kilodaltons.

Identification of some target auto-Ags for AECAs.

To distinguish between target Ags for AECAs in leprosy and malaria, the MBs were electrophoresed and blotted with eight CY⁺ MB⁺ sera from leprosy patients and eight CY⁺ MB⁻ sera from malaria patients. As can be inferred from their densitometric profiles (one lepromatous patient profile, one malarian patient profile, and the healthy control profiles are presented at the top of Fig. 3A), AECAs from the leprosy patients did not exhibit the same pattern of reactivity as did those from the malaria patients. AECAs from the former patients recognized the 180-, 100- to 105-, 95-, and 67-kDa proteins, while they bound to a 51-kDa protein band in the sera from the latter patients. A 70-kDa protein also acted as an auto-Ag for AECAs from the leprosy patients and, to a much lower degree, for those from the malaria patients. Shown at the bottom of Fig. 3A is the comparable Ag recognition obtained with the eight sera from leprosy patients.

FIG. 3.

Analysis of target Ags of AECAs. (A) EC MB extracts were electrophoresed and blotted with sera from eight leprosy patients, eight malaria patients, and two healthy European controls. A representative example is shown for each at the top, and the eight patterns from leprosy patients are superimposed at the bottom. (B) 2D electrophoresis of one serum sample each from a healthy European control and a patient with SLE as negative and positive controls, respectively. One representative serum from eight leprosy patients and one representative serum from eight malaria patients are shown.

After 2D electrophoresis, MB proteins of EA.hy926 cells were transferred onto PVDF MBs and probed. MB Ag specificities of AECAs contained in CY⁺ MB⁺ sera from eight leprosy patients contrasted with those identified by sera from eight malaria patients. In the SLE serum, there was a combination of auto-Ags resembling those of leprosy and auto-Ags resembling those of malaria. In leprosy, AECAs (Fig. 3B), calreticulin (pI 4.3; 67 kDa), vimentin (pI 4.7; 51 kDa), tubulin (pI 4.9; 55 kDa), and heat shock protein 70 (HSP-70) (pI 5.2; 70 kDa) could be discerned, among other auto-Ags. There were also a number of unidentified spots with sera from leprosy patients that were seen neither with the sera from malaria nor with the serum from SLE.

Effects of AECAs.

IgG AECAs from three of the five LL and one of the three TL sera tested activated the ECs. They reproducibly enhanced the expression of E-selectin (Table 4) from 0.3 (serum 5) to 63.7% (serum 7), as well as that of ICAM-1 from 2.9 (serum 5) to 36.3% (serum 4), compared with 70.7% for E-selectin and 93.5% for ICAM-1 following incubation of the cells with IL-1β. In contrast, IgG from the remaining four sera from leprosy patients and from three randomly chosen AECA-positive sera from malaria patients failed to influence the adhesion molecule expression.

TABLE 4.

Modulation of E-selectin and ICAM-1 following incubation of ECs with AECAs^a

Patient no.	% of MFI variation		Induction of activation
	E-selectin	ICAM-1
1	−18.0	−9.9	No
2	−5.6	−10.5	No
3	+2.6	−9.9	No
4	+20.8	+36.3	Yes
5	+0.3	+2.9	No
6	+8.9	+8.2	No
7	+63.7	+29.3	Yes
8	−10.1	+28.4	Yes
IL-1β	+70.7	+93.5	Yes

^aConfluent ECs were incubated with AECA IgG at the final concentration of 160 μg/ml. E-selectin and ICAM-1 expression levels were measured by flow cytometry 6 h later. The variation of MFI was calculated from 100 × (AECA IgG-treated EC MFI − control IgG-treated EC MFI/control IgG-treated EC MFI). Activation with IL-1β was used as a positive control.

In addition, IgG AECAs from four of the four LL and none of the four TL sera tested accelerated apoptosis of ECs as suggested by cell morphology (Fig. 4A), whereas those from sera of malaria patients exerted a negligible effect. Only one of the four activating sera generated apoptosis. This was confirmed by the translocation of PS and its subsequent binding to annexin V in 32.3 to 44.0% of ECs, following incubation of the cells with AECAs from leprosy patients (Fig. 4B). In addition, 15.2 to 21.1% of ECs became hypoploid (Fig. 4C), and DNA exhibited fragmentation (Fig. 4D).

FIG. 4.

AECAs from leprosy patients, but not from malaria patients, trigger apoptosis of ECs. Two representative examples are shown in each assay, one from leprosy and the other from malaria. Apoptosis was determined as follows: morphology of the cells (apoptotic cells are shown with arrows) (A); binding of annexin V to PI-negative HUVECs (B); hypoploidy, as determined by PI staining of permeabilized cells (C); and agarose gel electrophoresis showing DNA degradation (D). MW, molecular weights in thousands.

DISCUSSION

The present study set out to compare AECAs from leprosy patients to those from malaria patients. Because of the controversy over their pathogenicity, it was also aimed at demonstrating that such auto-Abs could play an active role in either of these diseases. Although the baseline production of Abs (12) and auto-Abs (22) has been reported to be relatively high in our healthy African controls (as would be expected for an area where parasitic infection is endemic), significant EC reactivity was, nevertheless, found across the whole spectrum of leprosy. By use of the ELISA, AECAs were identified in 52% of the leprosy patients, compared with 88% of the malaria patients. However, as many as 41% of the leprosy patients reacted with ECs in the FACS analysis, in contrast to as few as 9% of the malaria patients.

One trivial explanation for this discrepancy could relate to technical differences between the two AECA assays. Owing to their fixation with glutaraldehyde, ECs are inevitably permeabilized in the ELISA, whereas these substrate cells are tested live in the FACS assays. Given their detectability in the latter assay, some auto-Abs from leprosy patients may be assumed to bind to the MB. As such, these are “genuine AECAs.” Conversely, those that tend to recognize CY components in malaria patients behave as “pseudo-AECAs.” This interpretation is aided by the fact that permeabilization of ECs rendered most of the CY⁺ MB⁻ sera from malaria patients FACS assay positive but eliminated the reactivity of most of the sera from leprosy patients. To provide further evidence that AECAs are not of the same specificity in leprosy and malaria patients, we incubated sera with CY lysates to show that EC reactivity was absorbed in a dose-dependent manner in malaria, but not in leprosy. Interestingly, AECAs from malaria patients, but again not those from leprosy patients, were retained by similar CY structures in permeabilized epithelial cells.

These findings suggest that genuine AECAs are specific for ECs in leprosy patients, unlike malaria patients, where pseudo-AECAs are directed to ubiquitously expressed CY determinants. AECAs should thus represent one of many serological hallmarks of polyclonal B-cell activation in malaria (1). The failure of LL patients to localize their infection and association of this form with a type 2 cytokine profile (37), which is typical of a dominant humoral response, are consistent with their greater frequency of AECAs compared with TL patients. The same holds true for ANAs (22; this study), aPLs (2), and antineutrophil (19) and antimitochondrial (10) Abs. Yet, AECAs from leprosy, even in its LL form, cannot be equated with the possible concomitant polyclonal B-lymphocyte activation.

Mycobacteria have been shown to be detained and ingested by ECs in animal models and cell cultures, although a high multiplicity of infection is necessary for complete saturation (32). Because, once triggered, AECAs keep being generated, they may in turn serve a role as EC-injuring effectors. As a result, neighboring noninfected ECs could be impacted by shed AECAs. Morphological evidence of activation has been previously described in infected ECs, particularly those associated with peripheral nerves (32). Promotion of AECAs by M. leprae-induced alterations of ECs and the capability of some of them to perpetuate the vicious circle by activating the cells raise the possibility that such EC damage facilitates adhesion of M. leprae-infected macrophages. Thus, spreading of the disease would be encouraged. On the basis of our observations, the possibility of activation being a prerequisite for the advent of apoptosis remains unlikely, given that there was only one EC-activating serum in the four that generated apoptosis. However, the possibility still exists that EC apoptosis is due to Ab-dependent cell-mediated cytotoxicity, as recently described (33). In this respect, it is predictable that ECs enter several apoptotic programs, implying that distinct categories of MB-associated AECAs do exist. Similar data have been reported in cytomegalovirus (17) and dengue virus (16) infections.

Subsequently, AECAs act as opsonins by facilitating removal of M. leprae-injured cells. Indeed, following its translocation to the outer leaflet of their plasma MB, PS becomes one of the most efficient markers for phagocytosis (8). This phenomenon sets in motion events that prevent inflammation. It follows that bacterial debris is released into macrophages, which is another mechanism for AECAs to promote dissemination of the disease.

Inasmuch as ECs from various tissues differ with respect to their surface phenotype, a major concern regarding the specificity of AECAs remains. By analogy with published 2D EC maps, that of the EC MB proteins has uncovered a few candidate auto-Ags, such as calreticulin, vimentin, tubulin, and HSP-70. Their residence within the CY does not cast doubt on the specificity of this determination for MB Ags. These proteins are indeed kept adherent to the MB throughout its preparation.

In addition, CY, MB, and MB-bound CY Ags may resemble mycobacterial epitopes. For instance, human HSP-70, which is a chaperone molecule, reproduces the C-terminal half of the M. leprae HSP-70 (14). Accordingly, it may initiate cross-reactive auto-Abs, either singly or through interaction with any chaperoned auto-Ag. The resulting complexes are released from dying cells, so that they express cryptic epitopes (15) and become immunogenic. Vimentin and tubulin have also been endowed with sufficient immunogenicity to initiate autoimmune processes after Ag spreading (35). This hypothesis corresponds with the presence of auto-Abs to tubulin in 94% of leprosy patients (11) and that of anticalreticulin auto-Abs (20) in SLE and various parasitic infections, including leishmaniasis and onchocerciasis. In contrast, little or no increase in these auto-Ab levels was seen in sera from malaria patients. To complement these studies, MB immunogenic motifs are currently being dissected, using mass spectrometry.

Taken together, the present observations suggest divergent behavior of AECAs in leprosy and malaria. They may play a role in the pathogenesis of some cases of leprosy, particularly in its LL form. They refine our understanding of the mechanisms operating in this disease. Their determination could, hopefully, help predict which leprosy patients will experience an aggressive outcome and allow workers to devise unique therapeutic strategies.