Scami. J. Immunol. 35, 735-744. 1992

Western Blot Analysis of Human IgG Reactive with the CoUagenous Portion of Clq: Evidence of Distinct Binding Specificities u. MARTENSSON, A. G. SJOHOLM. G. STURFELT*, L. TRUEDSSON & A -B. L A U R E L L Department of Medical Microbiology. Lund University, and *Dcpartmcnl of Rheumatology, University Hospital. Lund. Sweden

Martenssoii U. Sjoholm AG, Slurfell G. Truedsson L. Laurel! A-B. Western Blot Analysis of Human IgG Reaetive wilh ihe CoUagenous Portion of Clq: Evidence of Distinct Binding Specificities. Scand J Immunol iy92;35'735 44 An enzyme-linked immunosorbent assay (ELISA) with purifiedcollagenous Clq fragments in the solid phase was used for detection of CI q-specific immunoglobulins in the sera of twelve palients with systemic lupus erythcmatosus (SLE) or the SLH-like disease hypocomplemenlcmic urticarial vaseulitis syndrome (HUVS). By elinical criteria, four patieiils had SLE. and three HUVS. Five patients had overlap syndromes. All patienls demonstrated high concentrations of Clq-specilic IgG and markedly low concentrations of circulating Clq. Detection of Clq-specilic IgG m SLE sera was facilitated by employment of saturating eoncentrations of collagenous Clq fragments in the solid-phase ELISA. When added to SLE serum, immune complex-fixed C1 q inhibited binding of IgG to the Clq fragments, whereas addition of Clq alone had limited inhibitory effecis. Under similar conditions, using approximately equimolar amounts of Clq relative to solid-phase Clq fragments, no ELISA inhibition was obtained after addition of Clq or immune complex-fixed Clq to a HUVS serum. Even in large excess, purified Clq did not inhibit binding of HUVS-IgG to solid-phase Clq fragments. Thus, possible interactions between HlJVS-lgG and nalive Clq are probably of low affinity. By Western blot analysis. IgG reactive with the B and C chains of Clq was found in Ihe eight patients with evidence of HUVS. five of whom also showed IgG binding to C C'and A' B'dimers of collagenous Clq fragments. Sera from SLE patients were negative by Western blot analysis. It seems likely that Clq-specific IgG in SLE primarily recognizes assembled Clq molecules or collagenous Clq fragments expressing conformational epitopes of bound Clq. Interestingly, patienls with evidence of HUVS fairly consistently had zymogen (ClrCIs): complexes in their serum, while patients with SLE showed high concentrations of complexes containing CI inhibitor. Clr and Cls. Different binding specificites of Clq-reactive IgG could be of importance wilh regard to pathogenctic mechanisms in SLE and HUVS. There was no correlation between findings of CI q-specific IgG and a variety of autoantibodies associated with SLE and SLE-likc disease. Anders G. Sjoholm. Deparimcnl of Medical Microhii)hgy, Solvegatan 23, S-223 62 Lund. Sweden

A.s first reported by Agnello et at. [1]. the sera of some patients with systemic lupus crythcmatosus (SLE) and other hypocomplemcntcniic states are capable of precipitating purified Clq in agarose gel. The C'lq-precipitating activity was ascribed to IgG complexes of high molecular weight (19S) and to low molecular weight (7S) components. Low molecular weight Clq precipitins have particularly been found in patients with very low Clq concentrations and a rare SLE-like syndrome characterized by fairly distinctive clinical features

Clinical Immiiiiolof^v Section.

including urticarial vaseulitis, obstructive lung disease and recurrent uvcitis [2]. The disease is usually referred to as the hypocomplementemic urticarial vaseulitis syndrome (HUVS). even though other designations are sometimes used [3]. Low molecular weight Clq precipitins have been shown to be polyclonal IgG [4, 5]. Recent reports of F(ab');-mediated binding to Clq clearly suggest Clq autoanlibodies to be responsible for the reactivity [6-8]. Analysis of sera from patients with SLE have provided evidence of antibodies 735

736

U. Martensson et al.

recognizing the collagenous portion of CI q [6. 7], In Ihc present study, we employed an cnzymclinked itnmtinosorbent assay (ELISA) for dcteclion of immunoglobulins reactive with the collagenous portion ofClq.Toensure that reactivities did not involve specific or non-specific immunoglobitlin binding to the globular part of Clq, collagenous Clq fragments [9\ were used in the assay. On the basis of high values in the ELISA, the sera of twelve patients with SLE or HUVS were selected for further investigation of binding specificities. Western blot analysis was performed with Clq and collagenous Clq fragments together with ELISA inhibition experiments using Clq and immune complex-bound Clq. The sera were also investigated with regard to some of the autoantibodies associated with SLE and SLElike conditions.

MATERIALS A N D METHODS Patients. The patient sera were originally referred to the laboratory for diagttostic coniplemeti! analysis. Samples were slored itt aliquolstit —80 C, In principle, one serum sample from each patient was used in the investigation. Patient I was mainly treated in the Departtnent of Nephrology, and patients X and XI in the Department of Rheumatology, University Hospital, Lund, Basic information concerning patients III, IV and V was kindly provided by Professor R, Hiillgren, Academic Hospital, Uppsala, Patient VIM was treated in the Departments of Opthalmology and Rheumatology, and patieni XII in the Department of Internal Medicine, Sahlgren's Hospital, Gothenburg, Patient II was treated in the Department of Internal Medicine, Regional Hospital, Orcbro. and patient VII in the Hospital of Ostersund, Patient VI was treated in the Department of Dermatology, Central Hospital, Kristianstad, and patient IX in the Department of Internal Medicine, Central Hospital, Viixjo. For each patieni, records were reviewed in collaboration with the patient's physician. Revised ACR criteria for the dtagnosts of SLE were used for classification [10], Clinical data are summarized in Table I, C(«»/j/c/jif/j//iffy/f//i,v, The concentrations orCtq.C4, C2. C 3 a n d C l inhibitor (CI INH) in serum or EDTA plasma were delermmed by clcctroimmunoassay [II], Values were given in mg/1 assuming thai pooled reference serum or plasma contained Clq at 70 mg/1. C4 at 600mg.l,C2at26mg/l,C3at 13(H)mg/l.and CI INHat ISO mg.l [12], CI subcomponent complexes were studied by crossed immunoelectrophoresis with semiquantitativcestitiialion of(Clr C"ls); in ihe/^i regton and CI INH-conlaining complexes comprising CI INH CIr C l s a n d C I INH-Clr CIs CI INH in the a^ region [13, Purified Clq and collagenous Clq fragments. Highly purified preparations of human Clq and the collagenous Clq fragment obtained by pepsin digestion of Clq

were prepared according to published methods [9, 15], Concentrations of purified Clq diluted in C!q-depleted sertim [16] were quantified with electroimniunoassay [11]. Concentrations of collagenous Clq fragments were delermined by measurement of ahsorbanee at 275 n m ( E ' ' / l em = 2,l)[17]. Solid-phase ELISA for measurement of immunoglohulin binding to collagenous Cli/ friigmenis. Most experiments were carried out with flat-bottomed microlitre plates (Multibind itnmunoassay plates, high binding capacity, Greiner Gmbh, Krcmsmunster, Austria). The plales were coated wilh 0,05 ml of collagenous Clq fragments per well al a concentration of 0.01 gperl (5()0 ng) in pliosphate-buffered saline (PBS), With the reported binding capacity of the plates (6t)l} ng per cm-), 400 ng of the protein or less was assumed to be bound per well at saturation. Other llat-bottomed mierotitre plates (Nunc-Immunoplates, Maxisorp, A/S Nunc, Roskilde, Denmark) and variation of the coating doses were tried in some experiments, Wilh the latter plates, coating was performed with 0.1-ml volumes assuming saturation at about 500 ng of Ihe protein per well. The coating step was consistently allowed to proceed for about 18 h at 4 C, The weiis were then washed three times with PBS eontaining 0,05"/i. Tween 20 (PBS-Tween), Patient sera (0,05 ml) at a dilution of 1/200 in PBS-Tween were applied in duplicate to the coated wells and to an uncoated control well. The plates were incubated at room temperature (20 C) Tor 2 h. After washing three times with PBS-Twecn, 0,1 ml of alkaline phosphataseconjugated rahbit F(ab'): anti-IgG (Ec-fragmenl specific, Cappel, Organon Teknika, Durham, NC, USA) was added to each well. Conjugates with alkaline phosphatase (Type Vll-S, Sigma Chemical Co,, St Louis, MO, USA) were prepared as described by Vollcr et al. [IS]. Alkaline phosphatase-conjugatcd rabbit F-''(ab')2 antiIgM and goat F(ab'): anti-IgA (Cappel) were used in some experiments. Peroxidase-cotijugated rabbit IgG to kappa or lambda light chains (Dakopalts A/S, Glostrup. Denmark) were also used. Binding was allowed to proceed for 18 h at 4 C, The plates were washed three times with PBS-Tween hefore addition of appropriate enzyme substrates. Absorhance was measured in a Multiskan Plus photometer (Labsystems Ltd, Helsinki, Finland). A reference curve was obtained from mean values after repeated titration ofa patient serum said to contain Clq-reaetive IgG at 1000 arbitrary units (AU) per litre, A positive control yielding an absorbance of about 1,0 was included in each plate. After adjusting ahsorbanee values according to the positive control and correction for non-specific binding in each sample, values were read from the reference curve and given in All per 1. The limit of detection was 16 AU per 1. ELISA inhihilion experitnents. Inhibition of IgG binding to solid-phase CI q fragments was studied using preformed immune complexes, purified Clq, and purified Clq bound to immune complexes. Immune complexes were prepared at equivalence from Ovalhumin (OA) (Grade V. A 5503. Sigma Chemical Company) and rabbit anti-OA [19], Atler incubation of OA with anti-OA for 1 h at 37 C and IS h at 4 C the precipitate was washed three times in Tris 25 mmol/1. NaCI 150 mmol/1, pH 7,4, The protein concentration of Ihe complexes was 18,5 g/i as determined hy colorimetric assay (Bio-Rad Protein Assay, Bio-Rad Laboratories,

Autoantibodies Richmond, CA. USA) after solubilizalion of the cottiplexesiti NaOH at lOOinmoM. Purified Clq at 70 mg/l was fixed lo inimiiiie complexes at I g.ldiiring40min at 4 CinTris20mmo!/l.NaCI I20mmol;i. wilhCa-'' at 5 mmol/l. pH 7.(1 No Clq remained unbound a.s assessed by ujimunochemical measurement of Clq [II] in the supernatant after centrifugation of the mixltire (Bctknuin mierofugc E. 10,000 rpm. 15 s). In experiments such as those shown in Fig. 4. 0.1-ml volumes of sera from patients II. III. Vlll. IX. X and XI diluted 1/50 1/200 in Ca-*-containing Tris bulTer were pretreated with equal volumes of Clq (70 mg/l), immune complexes, or Ciq bound to immune complexes. After 2(1 h at 4 C the mixtures were centrifuged at 10.000 rpm for 15 s. Aliquots of the supernatants were diluted 1/2 in PBS-Twecn yielding fmal scrum dilutions ranging between l.'2(K) and i/1600. and then 0,05-ml volumes were analysed with ELISA for IgG binding to solidphase Clq fragments (coating dose 500 ng). With the procedure used, the IgG in each 0.05-ml volume finally analysed was estimated lo have been exposed to 875 ng of added C1 q. The molecular weight of collagenous C1 q fr;tgn"tents is considered lo be about 4()".li of that of Clq [17. 20). This implies that the molar eoni.entration of Clq added was probably similar to the molar concentration of solid-phase Clq fragments in the assay. In other experiments, purified Clq in up to 100 X molar excess relative to solid-phase Clq fragments was added to the serum of a HUVS (patient VIII). Pretreatment with Clq was carried out either for 20 h at 4 C as described above, or for I h at 37 C. Western hint analysis. Polyacrylamide gel electrophoresis (PAGE) of Clq and collagenous Clq fragments was performed in \1"A, gels (Acrylamide/Bis. 19:1. BioRad) according to the procedure of Laemmli [21] using an LKB Midget Electrophoresis Unit (LKBProdukter. Bromma. Sweden). The gel dimension was 0,75 X S5 X 70 mm. Clq was reduced in the presenee of 2-mercaptoethanol (2-ME 5'!',i vol/vol, 100 C. 5 min or 37 C. 20 min) before PAGE analysis. A totul of 0.75/Jg lif the protein was applied per 5 mm of the gel. PAGKof collagenous Clq fragments was perfonned under nonreducing conditions with application of 1.1 /ig of the protein. Transfer of electrophoresed proteins [22] to nitrocellulose membranes (0.45 micron. Bio-Rad Laboratories) wus performed with semi-dry electroblotk-r equipment (AN