Sulphatide-binding properties are shared by serum amyloid P component and a polyreactive germ-line IgM autoantibody, the TH3 idiotype

Abstract

Serum amyloid P component (SAP) concentration was elevated in sera from leprosy patients, significantly so above endemic controls in lepromatous cases. In the sera of lepromatous leprosy (LL) patients who experienced an erythema nodosum leprosum (ENL) episode the SAP fell at the onset of ENL and remained low throughout, in two of three cases. Changes in SAP concentration parallel anti-sulphatide IgM concentrations. TH3, a monoclonal IgM germ-line antibody derived from a LL patient, and SAP share similar binding patterns. In this study we demonstrate binding to heparin and sulphatide. Moreover, SAP inhibited the binding of TH3 to sulphatide, as well as anti-sulphatide IgM found in a range of sera, and anti-sulphatide IgG in the only sera sample in which it was found. The observation that anti-TH3 idiotype monoclonal and polyclonal anti-SAP antibodies both inhibited the binding of TH3 and IgM in sera (but not IgG) to sulphatide without binding to sulphatide themselves further demonstrated similar binding specificities. The observations of similarity in binding reinforce ideas that SAP may function as a primitive opsonin, but the clear ability to inhibit binding of autoantibodies suggests that SAP may play a role in ameliorating tissue and particularly nerve damage in leprosy patients.

INTRODUCTION

Serum amyloid P component (SAP) is a serum glycoprotein found in all forms of amyloidosis. Although it has structural similarities to plant lectins [1] and binds to a variety of carbohydrate structures, its function is not understood. SAP and IgM autoantibodies with the TH3 or related idiotype, as defined by a human MoAb, bind to similar ligands. For example, they both bind to single-stranded and double-stranded DNA [2–4]. Autoantibodies have been demonstrated to bind human cytoskeletal glycoproteins or mitochondria [3] and capsular glycolipids from Mycobacterium leprae [4]. SAP binds to many ligands, in a calcium-dependent manner, including glycolipids from Klebsiella rhinoscleromatis and Streptococcus pyogenes [5]. Thus, it has been suggested that the presence of heparan sulphate in glomerular basement membrane may be responsible for the deposition of SAP and autoantibody complexes at this site [6,7], since both bind to this sulphated carbohydrate. Also SAP bound to sulphatide (cerebroside-3-sulphate) amongst a range of sulphated and phosphorylated carbohydrate ligands [8]. Subsequently, we showed binding of the MoAb TH3, as well as IgM antibodies in leprosy sera, to solid-phase sulphatide [9]. Since binding of IgM antibodies to heparin had not been investigated, we determined whether binding to this ligand, too, was shared with SAP.

The acute-phase reactants C-reactive protein (CRP) and fibronectin [10], amongst others, are elevated in sera from patients with leprosy. Like SAP, these reactants bind a wide range of ligands, so we investigated some interactions between the two pentraxins, SAP and CRP. Further, anti-sulphatide IgM (but rarely IgG) is elevated in leprosy, in relation to bacterial load [9]. SAP is not considered an acute-phase reactant. However, given that SAP has some functional similarities to both anti-sulphatide IgM and some acute-phase reactants, we investigated the SAP content of some leprosy sera. Since antibodies to sulphatide are associated with a number of autoimmune diseases [11,12], including neuropathies [13,14], we wondered if interactions of SAP and TH3 idiotype with sulphatide could play a part in the neural pathogenesis in leprosy.

Thus in the present study our major interest was to investigate whether SAP could interfere with the binding of TH3 and related antibody to sulphatide. We also investigated, vice versa, whether TH3 interfered with the binding of SAP to sulphatide. We were also interested to determine whether antibodies raised to SAP could react with antibodies with the TH3 idiotype, indicating possible functional similarities and even similar conformational and charge patterns in the binding site of SAP and TH3.

MATERIALS AND METHODS

Production of SAP and CRP

SAP was isolated from human serum by affinity chromatography on phosphorylethanolamine Sepharose 4B using elution with 10 mm EDTA. SAP was further purified using anion exchange chromatography (DEAE-cellulose) followed by zinc chelate chromatography [8]. CRP was isolated on the basis of its affinity for phosphorylcholine-Sepharose. Further purification was done using anion exchange followed by gel filtration [8]. Proteins were at least 98% pure as assessed by PAGE.

Production of antibodies

TH3 and PR4 idiotype MoAbs

These MoAbs have similar binding properties, in particular to a range of charged ligands [3]. They were prepared as described previously [3] by fusing monocytes from patients with lepromatous leprosy (LL) with the lymphoblastoid human cell line GM 4672.

Anti-idiotype TH3

Anti-idiotype was produced by collecting polyclonal rabbit serum and performing affinity chromatography with a final step of absorption to a TH3 affinity column followed by elution at pH 2.5 to yield TH3 idiotype-specific antibody. The method was described in full previously [4].

Antibodies against SAP and CRP

Antibodies were raised in goats using purified protein initially in Freund's incomplete adjuvant (FIA) and subsequently in saline. The antibody was affinity-purified by immunoaffinity chromatography on the respective protein coupled to cyanogen bromide activated-Sepharose 4B (Pharmacia) according to the manufacturer's instructions.

Monoclonal antibodies against SAP

MoAbs were prepared by immunization of 6–8-week-old BALB/c mice with native SAP (5 μg subcutaneously and 25 μg intraperitoneally) in FIA. At 14 days mice were reimmunized by the same procedure. At day 21 antibody titres were assayed by ELISA. Each mouse was immunized intravenously with 30 μg SAP in physiological saline. SP2/0-Ag 14 myeloma cells were fused to mouse spleen cells on day 3 post-i.v. immunization (day 24) by the method of Galfre & Milstein [15] with modifications by Lane [16]. Cells positive for antibody to SAP by ELISA were cloned twice at one cell per well. These cells were expanded and subsequently grown in Serum-Free Medium HL-1.

Antibodies against immunoglobulins

These were purchased from Sigma (Poole, UK). Alkaline phosphatase-linked goat anti-human IgM (μ-chain), alkaline phosphatase-linked goat anti-human IgG (γ-chain) and alkaline phosphatase-linked goat anti-mouse IgG were used in ELISA.

Screening ELISA for monoclonals raised against SAP

SAP (1 μg/ml) in 0.13 m borate pH 9.6 was coated onto Immulon 1 plates (Dynatech, Chantilly, VA) and incubated overnight at 4°C. Wells were blocked with 10 mm Tris–HCl pH 8.0 containing 0.15 m NaCl (TBS) containing 1% bovine serum albumin (BSA) for 2 h at 37°C. Plates were washed with TBS containing 0.05% Tween-20 (TBS–T) and culture supernatant (50 μl) was added and plates incubated for 1 h at 37°C. The plates were washed three times with PBS–T and probed with 60 μl/well goat anti-mouse IgG alkaline phosphatase.

Human sera used in this study

All sera used for experiments involving sulphatide-binding antibodies in this study were from patients with LL from Papua New Guinea. Sera coded R9 (a male, 37 years, amyloid) and R139 (a male, 40 years, not amyloid) were from patients suffering from erythema nodosum leprosum (ENL) reactions; serum coded AL108 (a female, 25 years, not amyloid) was from a patient not in reaction. In all leprosy sera used, anti-sulphatide IgM was elevated to between three and seven times control level. Anti-sulphatide IgG was only elevated in R9 and was rare, as described previously [9].

For SAP determinations, sera coded AL108 (above), AL109 (a female, 25 years, not amyloid) and AL184 (a male, 31 years, amyloid) were taken from patients before and during ENL reactional episodes. Additionally, further sera from LL, borderline tuberculoid (BT) and tuberculoid (TT) patients were used for determination of mean SAP values; these patients had a range of 21–35 years and male/female amyloid/non-amyloid were represented in each group. Possible renal dysfunction was indicated by the detection, using Labstix, of proteinuria, in three of the six LL patients and three of the four TT patients from whom serum was assayed for SAP in this study.

All sera were stored at −40°C. Dilutions, in 5% BSA–PBS [9], were used fresh or after freezing or rethawing no more than once.

ELISA for sulphatide-binding antibodies

Immediately before use, Immulon 1 plates were washed three times with PBS (from tablets; Sigma) with 0.1 mm CaCl₂ (PBS–Ca). Plates were coated with, per well: sulphatide (1 μg), phosphatidylcholine (2.5 ng), cholesterol (2.5 ng) in 50 μl methanol. All lipids were from Sigma, the sulphatide was from bovine brain. Control wells included phosphatidylcholine and cholesterol alone. The methanol was allowed to dry in air at room temperature. Subsequently plates were blocked for 1 h at 37°C with 5% BSA in PBS–Ca. After blocking, without further washing, serum samples or MoAbs (from stock solutions at 500 μg TH3 or PR4 per ml) diluted at least 100-fold in 5% BSA in PBS–Ca were added to wells. Edge wells were not used. After 1 h at 37°C, the plates were washed three times with 5% BSA in PBS–Ca and twice with 5% BSA in PBS, then alkaline phosphatase-linked goat anti-human IgM (μ-chain, Sigma; diluted 1:1500 with 5% BSA in PBS) or alkaline phosphatase-linked goat anti-human IgG (γ-chain, Sigma; diluted 1:1000 with 5% BSA in PBS) was added to all wells. After a further hour, plates were washed twice with 5% BSA in PBS, twice with PBS, and twice with carbonate/bicarbonate buffer pH 9.6. Then, plates were developed with paranitrophenol phosphate (Sigma 104) in 0.1 m carbonate/bicarbonate buffer pH 9.6 containing 2 mm MgCl₂ and absorbance at 405 nm was read in the wells using an ELISA plate reader (Dynatech MR5000).

ELISA with mixtures of anti-sulphatide and either SAP, anti-SAP, anti-CRP or anti-TH3

TH3 was used at 1 μg/ml, PR4 and sera were diluted to give a similar rate of colour development to 1 μg TH3/ml (± 20%) for anti-sulphatide IgM, and 55–80% the rate of colour development of 1 μg TH3/ml for anti-sulphatide IgG. For all these sulphatide-binding antibodies, these concentrations lay at points where colour development was proportional to log concentration. Actual concentrations of antibodies used are shown in Results. Antibody was mixed with either SAP, anti-SAP, anti-CRP or anti-TH3 or buffer control and incubated at 25°C for 1 h. Then both mixtures and controls were added to wells in the same blocked ELISA plate, after which plates were developed as described in the section above on the simple ELISA method. All agents were added to reactions as solutions in 5% BSA–PBS.

Preincubation with solid-phase sulphatide of either SAP, anti-SAP, anti-CRP or anti-TH3 followed by ELISA of anti-sulphatide

Either SAP, anti-SAP, anti-CRP or anti-TH3, as a solution in 5% BSA in PBS–Ca, were added to sulphatide-coated wells in blocked plates and incubated at 37°C for 1 h. Wells were then washed three times in 5% BSA in PBS–Ca. Then anti-sulphatide, at the same concentrations as used in mixtures (see above section), was added to wells, after which plates were developed as described above.

ELISA: determination of SAP in serum

SAP was detected using a sandwich ELISA using antibodies described in this study. Immulon 2 plates (Dynatech) were coated with a MoAb to human SAP (G2) at 0.1 μg/ml for 16 h and blocked with PBS including 0.1% Tween-20 (PBS–T) containing 1% BSA for 1 h at room temperature. The sample (diluted 1:4000 in TBS–T) and standards (100–0.5 ng/ml) were added to the wells and incubated for 1 h. Plates were washed and affinity-purified goat polyclonal anti-SAP conjugated to alkaline phosphatase (made using the single-step glutaraldehyde method) added and incubated for 1 h at room temperature before substrate addition as described above. Standards were purified human SAP. Normal human sera was shown to have concentrations within the normal range (see Table 1; actual range 21–78 μg/ml; coefficient of variation 4.1%). Further dilutions of up to 10-fold gave similar concentrations for the SAP concentration. Nearly all (> 95%) SAP reactivity was removed with PE-Sepharose.

Table 1.

Serum amyloid P (SAP) content in leprosy and control sera

ELISA: binding of antibody to heparin-coated wells

Immulon 1 plates were coated with poly-l-lysine (1 mg/ml in PBS pH 7.4) for 90 min at 37°C. Heparin or heparan sulphate (100 μl from 2 μg/ml solutions) was added to the wells for 2 h at 37°C and then the plates were blocked with PBS containing BSA 1% and 1 mm CaCl₂ (1 h, 25°C). Then antibody was added to wells and incubated for 2 h followed by detection as described above. For antibody binding, background optical density (OD) values were no more than 0.08 for the experiments that are reported.

Determination of binding of SAP to sulphatide

Wells of Immulon 1 plates were coated with sulphatide as described for antibody binding. The plates were blocked with PBS containing BSA 1% and 0.5 mm CaCl₂ (1 h, 25°C), incubated with SAP and TH3 in combination, including controls with either no SAP or no TH3 (1 h, 25°C). The SAP bound was quantified as in ELISA for SAP determination described above.

Analysis of amino acid sequence homology

No significant sequence homology was observed between SAP and the sequences of members of the TH3-related immunoglobulin family using analysis with FastA and bestfit at the Clinical Research Centre human genome mapping project facilities.

Statistical analysis

Percent inhibition was calculated as follows:

where ‘inhibitor’ is either SAP, anti-SAP, anti-CRP or anti-TH3.

Student's t-test was used for tests of significance and t-values were used in calculating 95% confidence limits. Regression analysis was done by least mean squares, through the Fig P package.

RESULTS

SAP content of leprosy sera

SAP was significantly elevated in sera from patients with LL, 65% over the level in endemic controls (Table 1). LL patients had an apparently lower SAP concentration during ENL (Table 1), though this was not statistically significant. Although hard to anticipate, sera taken before, at the onset, and during ENL were available from three patients. In these patients' sera (Table 2) the fall in SAP from the pre-ENL value was significant in the two patients whose serum anti-sulphatide IgM was shown to fall 2.5–5-fold in titre as described in a previous paper [9], at the onset of ENL. There was no relationship between possible renal dysfunction, which was indicated by proteinuria (see the section ‘Human sera used in this study’) and SAP content, but such a relationship has been established elsewhere [17] and it might become evident on leprosy patients if studied systematically, since renal function is known to be impaired in patients with LL [18], and one of the main line leprosy drugs, rifampin, is nephrotoxic in some patients [19].

Table 2.

Serum amyloid P (SAP) content in sera from individual leprosy (LL) patients before and during erythema nodosum leprosum (ENL)

Effect of SAP on sulphatide-binding antibodies

The criteria for selecting sera for binding inhibition and competition studies were elevated anti-sulphatide IgM giving the most reproducible OD values; clear documentation as to whether they were taken either during ENL episodes or definitely not from reactional episodes (see Materials and Methods); sufficient quantity to do all the experiments. R9 was the only serum available with a strong anti-sulphatide IgG response, though this binding activity could just be detected in R139, too.

SAP inhibited binding of antibody to sulphatide in a dose-dependent fashion (Figs 1,2 and 3). Each well in the ELISA had been coated with 1 μg sulphatide, as it was optimal in studies of inhibition of anti-sulphatide by SAP: it gave the highest percentage inhibition values for anti-sulphatide IgM. For anti-sulphatide IgG, a double reciprocal plot showed SAP to inhibit binding of antibody to sulphatide competitively, though SAP gave only slightly higher percentage inhibition values at 1 μg sulphatide than 0.1 or 0.3 μg sulphatide per well.

Fig. 1.

Inhibition of TH3 binding to sulphatide by serum amyloid P (SAP). The monoclonal antibody, TH3, was incubated at 1 μg/ml together with varying concentrations of SAP (•) or C-reactive protein (CRP; ○). Error bars show s.e.m.; n = 4–6 determinations for each value in the figure.

Fig. 2.

Inhibition of serum IgM binding to sulphatide by serum amyloid P (SAP). Serum R9 (▪, □) was used at 1:2000, serum R139 (▴, ▵) at 1:1000. Diluted sera and either SAP (▪, ▴) or C-reactive protein (CRP; □, ▵) were incubated together. Error bars show s.e.m.; n = 4–6 determinations for each value in the figure.

Fig. 3.

Inhibition of serum IgG binding to sulphatide by serum amyloid P (SAP). Serum R9 (▪, □) was used at 1:1000, serum R139 (▴, ▵) at 1:100. Diluted sera and either SAP (▪, ▴) or C-reactive protein (CRP; □, ▵) were incubated together. Error bars show s.e.m.; n = 4–6 determinations for each value in the figure.

Inhibition of antibody binding to sulphatide by SAP was shown with a polyspecific IgM MoAb, TH3, which bound to sulphatide (Fig. 1). Dose-dependent inhibition of sulphatide binding in sera was also shown, for both anti-sulphatide IgM (Fig. 2) and IgG (Fig. 3). When SAP was tested at just two concentrations against four other sources of anti-sulphatide IgM, it gave about 50% inhibition of binding of a further MoAb, PR4, and antibody in serum AL108, at 30 μg SAP/ml (Table 3) and antibody in two other sera (results not shown) at 3 μg SAP/ml. SAP did not inhibit binding of human IgM (Sigma) to alkaline phosphatase-linked goat anti-human IgM in a typical assay for IgM (see [9]), showing that the effect of SAP was not a general effect on IgM antibody–antigen binding).

Table 3.

Mechanism of inhibition by serum amyloid P (SAP) of autoantibody binding to sulphatide

In contrast, inhibition of SAP binding to solid-phase sulphatide by anti-sulphatide IgM could not be shown. When SAP at 0.01, 0.1 and 1 μg/ml was mixed with up to 5 μg/ml TH3 or PR4 (IgM MoAbs with sulphatide binding as one of their polyreactive properties), no effect on the amount of SAP bound to sulphatide was observed.

In order to show if the mechanism of inhibiting binding of IgM antibodies to sulphatide by SAP was competing by binding to sulphatide, SAP alone was added to sulphatide-coated wells followed by washing and subsequent addition of antibody. The SAP that remained bound after washing inhibited antibody binding. These experiments were performed with concentrations of SAP which gave 50% inhibition of antibody binding (Table 3) when SAP and antibody were mixed together. These results show that SAP must bind to sulphatide and thus compete with antibody for sulphatide. However, for two anti-sulphatide IgM tested (in AL108; TH3), less inhibition was obtained in experiments when SAP was preincubated with solid-phase sulphatide followed by washing excess SAP, suggesting that there may be some interaction between SAP and antibody. Alternatively, some bound SAP may be lost during plate washing and this is sufficient to lessen its inhibitory effect in the cases of TH3 and AL108. At 100 μg/ml, which was approximately the highest concentration observed in serum, SAP inhibited TH3 binding to sulphatide by 65 ± 1% when mixed with TH3 and by 45 ± 5% when preincubated with solid-phase sulphatide followed by washing. CRP—used as a control—did not inhibit the binding of antibody to sulphatide except weakly at the very high concentration of 100 μg/ml, when it inhibited only anti-sulphatide in sera to the same extent as 1 μg SAP/ml (Fig. 2). CRP had no effect on TH3 (Fig. 1) or on any anti-sulphatide when preincubated with solid-phase sulphatide followed by washing (results not shown).

Requirement of Ca²⁺ for effect of SAP

When Ca²⁺ was omitted from buffers and EDTA was added to chelate divalent cations, antibody still bound to sulphatide but there was no statistically significant effect of SAP (Table 4). Ca²⁺ was used at subphysiological concentrations to avoid precipitation; in the single experiment when it was added at 2 mm in ELISA similar colour development appeared to that in the wells including 0.1 mm Ca²⁺ (results not shown).

Table 4.

Calcium dependency of inhibition by serum amyloid P (SAP) of antibody binding to sulphatide

Binding site similarities between anti-sulphatide IgM and SAP

In order to search for more evidence of similarity of binding site the ability of antibodies to SAP to bind anti-sulphatide was investigated. Anti-SAP at 100 μg/ml inhibited anti-sulphatide IgM whenever it was tested against the material shown in Fig. 4, as well as antibody in the sera AL108 and R139 by 30–34%. However, anti-SAP did not inhibit binding to sulphatide by anti-sulphatide IgG when tested using the sera R9, R139 or AL108. Anti-SAP and anti-sulphatide must be mixed together to show inhibition by anti-SAP, since preincubation of anti-SAP with sulphatide-coated plates had no effect on antibody binding. This suggests that it acts by interacting with sulphatide-binding antibody. Typically for an antibody reaction, Ca²⁺ was not required for the effect of anti-SAP reported above, as when Ca²⁺ was omitted from buffers and EDTA included, % inhibition values for anti-sulphatide IgM TH3 and in R9, R139 and AL108 were not significantly different from those shown above and in Fig. 4. Anti-CRP, used as a control ‘irrelevant’ antibody, did not inhibit sulphatide binding. Anti-idiotype TH3 acted only against sulphatide binding by the MoAb TH3, but not in R9. However, anti-TH3 does inhibit anti-sulphatide IgM in sera in the usual circumstance where it is related to an increase in the expression of the TH3 idiotype [9].

Fig. 4.

Inhibition of IgM binding to sulphatide by anti-serum amyloid P (SAP) and anti-TH3. The MoAb TH3 (•, ○, ▾) was incubated at 1 μg/ml; serum R9 (▪, □, ♦) was used at 1:1000. Diluted sera or TH3 were incubated together with either anti-SAP (○, □), anti-TH3 (•, ▪) or anti-C-reactive protein (CRP; ▾, ♦). Error bars show s.e.m.; n = 4 determinations for each value in the figure.

The above results suggest the possibility of similar (sulphatide-) binding sites on the IgM antibody and on SAP. However, the sequence of SAP and antibodies of the TH3 and PR4 family which have been sequenced do not show any sequence similar to that shown by the group of sulphatide-binding proteins that have the sequence CSVTCGXXXRXR typified by properdin, thrombospondin and others [20]. Another protein sequence Trp-Ser-X-Trp has also been suggested to be involved in sulphated glycoconjugate binding [21], and this sequence is not found in SAP either.

Binding of TH3 to heparin

Since we were interested in binding similarities of SAP and TH3-like antibodies to sulphatide, we tested the binding of TH3 to heparin. TH3 at 1 μg/ml consistently gave a positive, though low, OD reading (0.095 ± 0.024; four determinations). Wells on the same plate, coated with 1 μg sulphatide and incubated at the same time with 1 μg TH3/ml gave an OD value of 1.45. Attempts to increase OD values by increasing the concentration of TH3 failed because background OD values, from uncoated wells, became very high and variable. However, heparin-binding IgM was more readily detected in serum R9, which also had high anti-sulphatide: binding to heparin could be detected down to a 1:2000 dilution (OD = 0.25 ± 0.01, six determinations). Although it had anti-sulphatide IgG, no heparin-binding IgG was detected in R9. Heparin inhibited neither binding of TH3 (0.3 and 1 μg/ml) to sulphatide nor anti-sulphatide IgM in serum R9 (1:1000 and 1:2000). In these experiments, heparin at 0.3–48 μg/ml was mixed with antibody. As previously determined [8], SAP bound to heparin and heparan sulphate.

DISCUSSION

SAP not only binds to sulphatide but also inhibits the binding of antibodies to sulphatide at concentrations of SAP found in sera. SAP inhibits principally by binding to sulphatide, so it might have a role in inhibition of autoantibody binding to sulphatide, which is a normal constituent of human cells and predominantly occurs on nerve cells. In this respect it is worth noting that CRP has been reported to be able to reduce the concentration of IgG autoantibodies to histones, DNA and DNP in a mouse model of systemic lupus erythematosus (SLE) [22]. CRP has previously been suggested to have sequence similarity to immunoglobulins in the CH2 region, and similarities in CRP binding properties to that shown by phosphorylcholine binding autoantibodies have been well documented. In the same way SAP may have the same binding characteristics as another set of autoantibodies of the family that includes TH3 and PR4.

It has been observed previously that purified SAP has a tendency to autoaggregate in the presence of calcium [23] and that ligands for SAP such as heparin can prevent this aggregation [24]. However, we used concentrations of calcium and carrier BSA at concentrations designed to minimize this risk. We were unable to observe any alteration in molecular weight by gel filtration of radiolabelled SAP under these conditions in the course of other studies. It is also clear that sulphatide binding also occurs in whole undiluted serum [25]. In these experiments [25], sulphatide was adsorbed to octyl-Sepharose and serum passed over the column. SAP and albumin were the major proteins bound. The amount of SAP bound was greatly in excess of what bound to octyl-Sepharose alone. This was also the case when sulphatide was absorbed to plastic. In addition, SAP was bound when whole serum from mice or hamsters was used (mouse SAP does not autoaggregate). Finally, since SAP binds to terminal sulphated galactose in a variety of different carbohydrate structures it is unlikely that binding to plastic alters its structure.

Anti-TH3 idiotype appears to be an anti-SAP binding site antibody, since both antibodies inhibit the binding of autoantibodies to sulphatide, without binding to sulphatide itself. The affinity of SAP binding to sulphatide may be higher than TH3 and related sulphatide-binding antibodies, since excess TH3-type antibody does not displace SAP when SAP is only present as bound to sulphatide. Thus, TH3 did not inhibit the binding of SAP to sulphatide.

Apart from binding sulphatide, we found that TH3 also shares with SAP [26] the property of binding heparin, another sulphated sugar. The binding of TH3-like antibodies to heparin appeared to be very weak, however, and heparin did not inhibit anti-sulphatide IgM, so that it would be possible to detect the sulphatide-binding activity in plasma.

In some diseases such as chronic infections there is a persistence of ‘starter’ autoreactive IgM antibodies accounting for the high levels of autoantibodies associated with infectious diseases: this is illustrated here in leprosy. SAP and CRP share several features with those IgM molecules encoded in the germ-line and produced initially in response to infection. These ‘starter’ antibodies are polyreactive, often autoreactive antibodies which evolve by the process of somatic mutation to higher affinity specific IgG antibodies. Since the pentraxins CRP and SAP phylogenetically predate the existence of immunoglobulins, they represent primitive recognition structures. They also share with immunoglobulins the ability to activate complement via the classical pathway and may act as opsonins for bacteria, parasites and immune complexes. SAP has also been shown to bind to DNA, chromatin and histones and may help their clearance and prevent nuclear antigen-stimulated autoimmune responses. Their persistence through evolution provides evidence for their importance, being available or produced rapidly following microbial challenge and acting as an early acute-phase defence system before specific antibodies are produced. Alternatively, their production during the acute phase may reflect the need to remove nuclear antigens before they cause inappropriate responses in the immune system and as we have shown here may inhibit binding of these antibodies to ligand. It is possible that this may prevent further stimulation of these antibodies in either quantity or affinity. It is not clear if this property has any influence on the pathogenesis of leprosy and if SAP could reduce nerve damage associated with anti-leprosy responses.

It remains to demonstrate the biological or clinical significance of the interplay of polyreactive antibodies and SAP. Although an association has been shown between the presence of autoantibodies to sulphatide and neuropathies, it has not been possible to prove a causal relationship with regard to pathology. Other potential significance of the interaction may relate to the observation that sulphatide has been shown to be a ligand on cells recognized by several pathogens, including Helicobacter pylori and certain herpes viruses, and that sulphatide can affect macrophage activation and interact with selectins. We have shown here that concentrations of SAP and autoantibody with similar ligand specificities parallel each other under certain conditions during clinical leprosy, and indeed SAP and autoantibody compete for sulphatide. We underline in this study the importance of non-protein antigens, particularly containing sulphated sugars, as ligands. The demonstration of interactions of sulphatide, SAP and autoantibodies to sulphatide in leprosy should allow the significance of the interactions to be followed, particularly during the acute, autoimmune reactional episodes in patients [27], in animal models of neural damage in experimental leprosy [28], and possibly in renal dysfunction in lepromatous leprosy [18].