Oct 19, 1993 - K. Whaley. Expression of the components and regulatory proteins of the classical pathway of complement in normal and diseased synovium.
Rheumatol Int (1994) 14:13-19
9 Springer-Verlag 1994
P. Gulati 9 D. Guc 9 C. Lemercier 9 D. Lappin K. Whaley
Expression of the components and regulatory proteins of the classical pathway of complement in normal and diseased synovium
Received: 19 October 1993 / Accepted: 23 February 1994
Abstract We studied the synthesis of the classical pathway complement components in synovial membrane. Ribonucleic acid was extracted from the synovial membranes of patients with rheumatoid arthritis (RA) or osteoarthritis (OA), as well as from normal synovial membrane. Northern blot and dot blot analysis showed that the mRNAs for all classical pathway complement components (ClqA chain, ClqB chain, CIqC chain, Clr, C1s, C4 and C2) and the fluid-phase regulatory components (Cl-inhibitor, C4-bp and factor I) were present in all three types of synovial membrane. Thus, all the components of the classical pathway were expressed in normal and diseased synovium. In an attempt to determine which components were synthesised by each cell type, monocytes (mononucIear phagocytes), human umbilical vein endothelial cells (HUVEC), synovial membrane fibroblasts (from normal, OA and RA synovial membrane) and peripheral blood lymphocytes were cultured in vitro and secretion rates of individual components were measured and total cellular RNA was analysed by Northern blotting. Monocytes secreted Clq, Clr, Cls, C4, C2, Cl-inhibitor and C4-bp but not factor I. Fibroblasts secreted Clr, Cls, C2, C3, Cl-inhibitor and factor I but not Clq, C4 or C4-bp. HUVEC secreted Cls, C2, Cl-inhibitor and factor I but not Clq, Clr, C4 or C4-bp. Lymphocytes did not secrete any of these components. In three instances mRNA was detected in the absence of secreted protein: mRNAs for the ClqA and CIqC chains were detected in HUVEC, whereas the mRNA for the ClqB chain was not, and C4 mRNA was detected in both fibroblasts and HUVEC. In the former the hybridisation signal was strong, whereas in the latter it was weak. These data indicated that, at least in fibroblasts, there may be a block in C4 translation. The results of these studies sug-
P. Gulati 9C. Lemercier 9D. Lappin 9K. Whaley ([]) Department of Immunology, Leicester Royal Infirmary, Leicester, LE1 5WW, UK D. Guc Beatson Institute for Cancer Research, Glasgow, UK
gested that all of the classical pathway components are synthesised in normal, RA and OA synovial membrane, and this may be explained at least in part by synthesis in mononuclear phagocytes, endothelial ceils and fibroblasts. They also showed that there are important cellspecific differences in the expression of the genes for the proteins of the classical complement pathway that require further investigation. Key words Classical pathway complement components.
Fluid-phase regulatory components 9 Synovial membrane 9 Cell type
Introduction The complement system comprises a group of proteins that promote the inflammatory response and destroy micro-organisms. The classical pathway comprises C1, C4 and C2. The C1 macromolecular complex (C1 q:Clr2 : C1 s2) is the product of five genes, three encode the A, B and C chains of Clq [1], while Clr and Cls are each encoded by a single gene [2]. C1 is activated on binding antigen-antibody immune complexes containing IgM or IgG antibody. Activated C1 then activates C4 and C2, by limited proteolysis. Activated C4 (C4b) binds covalently to suitable target surfaces. C2, the rate-limiting component of the classical pathway, binds to C4 prior to activation by C1 to form C4b2a, the classical pathway C3 convertase, which cleaves C3, the bulk protein of the system, into C3a (anaphylatoxin) and C3b. C3b binds covalently to acceptor groups and that which binds to the C4b component of C4b2a converts its specificity to that of a C5 convertase, C4b2a3b [reviewed in 3]. Fluid-phase regulatory proteins that modulate activation by the classical pathway include Cl-inhibitor (Cl-inh), which prevents spontaneous activation of C 1 and inhibits activated C1, and C4-binding protein (C4-bp), which acts as a c0factor for factor I[I] in the degradation of C4b [3].
14
During classical pathway activation a number of important pro-inflammatory products are generated. These include the anaphylatoxins C4a, C3a and C5a, which are cleaved from the N-termini of C4, C3 and C5, respectively. They activate different cell types including neutrophils, macrophages and mast cells, and C5a is a powerful chemotaxin [reviewed in 4]. C3b and iC3b ligate the complement receptors CR1 and CR3, respectively, and ligation of CR3 results in phagocytosis [reviewed in 5]. At sublytic concentrations the C5b-9 membrane attack complex (MAC) activates a variety of different cell types including neutrophils, macrophages and synoviocytes, resulting in secretion of cytokines and arachadonic acid and reactive oxygen metabolites [reviewed in 6]. Rheumatoid arthritis (RA) is a chronic disease affecting primarily synovial joints and is characterised by chronic inflammation and joint destruction [7]. Antigenantibody complexes are present within the synovial fluid and joint tissues in association with intense classical pathway activation [8], with the generation of anaphylatoxins [9] and the MAC [10]. Thus, intra-articular complement activation could contribute significantly to the inflammatory process in RA. Although the liver is the major site of synthesis for most circulating complement components, extra-hepatic synthesis of C3 and factor B occurs in most cells, while other complement components are also synthesised by mononuclear phagocytes, fibroblasts, endothelial cells, epithelial cells and adipocytes cultured in vitro [11]. Whether these cells synthesise complement components in vivo is unknown. Local synthesis of complement components may be important in host defence in the tissues and may contribute to the inflammatory process. Although exudation is characteristic of acute inflammation, the synovitis of RA is a chronic inflammatory process in which cellular proliferation/recruitment is the dominant feature. Evidence that chronic inflammation is responsible for joint damage is provided by (a) the clinical observation that joint destruction often proceeds in the absence of effusion and (b) the histological observation that at the leading edge of the pannus and at sites of cartilage erosion, fibroblasts, macrophages and newly formed non-patent capillary binds (comprising endothelial cells) are present [12]. As all three of these cell types are capable of synthesising complement components, it is expected that in RA local synthesis occurs in the synovium. This could contribute significantly to the amount of complement within the joint tissues and play a role in the mediation of joint damage. In this context it is important to note that a study of C3 metabolism in humans has shown that up to half of the C3 present in the inflamed joints of patients with RA is synthesised locally [13]. In the present study, we determined the expression of the classical pathway components in normal and inflamed synovial membrane and determined which were expressed by mononuclear phagocytes, lymphocytes, fibroblasts and endothelial cells, all of which are present at sites of chronic inflammation.
Materials and methods Reagents The following reagents were purchased from the sources shown. Linbro multiwell tissue culture dishes, trypsin-EDTA solution in Puck's saline (trypsin EDTA), Linbro 75 cm2 tissue culture flasks, Dulbecco's modified Eagle's (DMEM) and RPMI 1640 were purchased from Flow Laboratories, Rickmansworth, Herefordshire, United Kingdom. Fetal calf serum (FCS), antibiotic, Hank's balanced salt solution (HBSS) and Nunclon tissue culture flasks 25 cm 2, 75 cm 2 and 175 cm 2 were purchased from GIBCO BRL, Paisley, Renfrewshire, United Kingdom. Hybond-N membranes were purchased from Amersham International, Buck, United Kingdom. Random primed DNA labelling kit, DNase I (RNase free) and RNase inhibitor were purchased from Boehringer Mannheim, Mannheim, Germany. Diethylpyrocarbonate, gelatin (2% (w/v) solution) and endothelial cell growth supplement were purchased from Sigma Chemical Co, Poole, Dorset, United Kingdom. RNAzol was purchased from Biogenesis, Bournemouth, Dorset, United Kingdom. Human AB serum (ABS) was supplied by the Scottish Blood Transfusion Service (Law Hospital, Carluke, United Kingdom). FCS and ABS were heat-inactivated (56~ for 211) prior to use.
Preparation of cDNA probes Plasmids containing the following cDNAs were used: ClqA, ClqB, ClqC and C4-bp (pCIlA, pClqB, pClqC and PB-8, respectively; K. B. M. Reid, MRC Immunochemistry Unit, Department of Biochemistry, University of Oxford, Oxford, United Kingdom) [1, 14], CIr and Cls (pHClr5, pHCls22; M, Tosi, Institut Pasteur, Paris, France) [15, 16], C4 and C2 (Alu7, pC201; R. D. Campbell, MRC Immunochemistry Unit, Department of Biochemistry, University of Oxford) [17, i8], Cl-inh (pC1; P. Carter, Department of Bitchemistry, University of Aberdeen, Scotland) [19] and factor I (psp64; R. B. Sire, MRC Immunochemistry Unit, Oxford) [20]. With the exception of C4-bp, the cDNA inserts were excised from their plasmid vectors with the appropriate restriction enzymes. The inserts were isolated by electrophoresis in low-melting point agarose gels and purified by phenol-chloroform extraction and ethanol precipitation [21]. The C4-bp cDNA plasmid insert was amplified by using the polymerase chain reaction (PCR) using pAT t 53 primers (5' > CTC ATG TTT GAC AGC TTA TC < 3' and 5' > CAC GAT GCG TCC GGC GTA GA < 3') at a final concentration of 0.5 gM with 20 ng plasmid DNA. The PCR conditions were 1 min at 96 ~ (denaturation), 2 rain at 60 ~ (annealing) and 2min at 72~ (extension). Radiolabelled (~32p dCTP) cDNA probes were prepared by random priming [22].
Tissues and cells Synovial tissue Specimens of synovium were collected at the time of surgery from the knee joints of three patients with RA and three with osteoarthritis (OA) and from three individuals undergoing menisectomy who did not have OA or chronic inflammatory joint disease. Immediately after collection the tissue specimens were snap-frozen in liquid nitrogen and stored at -70 ~ until required. Finely ground frozen tissue fragments were homogenised in RNAzol (2 ml/100 mg tissue) before adding 1/10 volume of chloroform and mixing vigorously for 15 s and standing on ice for 5 rain. The suspension was centrifuged (120009 for 15 min at room temperature) and the aqueous phase was transferred to a fresh tube and an equal volume of ice-cold isopropanol was added. After standing on ice for 15 min
15 the tube was centrifuged (12000 g for 15 min at room temperature), the supernatant was removed and the pellet washed twice with ice-cold 75% ethanol before being dissolved in water (100 gl). The concentration of RNA preparations was determined spectrophotometrically (OD26o). RNA samples that were contaminated with genomic DNA were incubated (5 min at 37 ~ with RNase inhibitor (10 units/gg RNA) prior to incubation (10 rain at 25~ with RNase-free DNase 1 (5 units/100 gl). The sample was then re-extracted with RNAzol and chloroform and precipitated with isopropanol as described above. RNA extracted from synovial membrane was analysed in formaldehyde-denaturing agarose gels and the relative abundances of mRNA species determined by dot-blot analysis using 5 gg RNA per dot [23]. Hybridisation reactions, washing and autoradiography were performed as described for Northern blots (see below). Blots were standardised for RNA loading as described previously [24].
Syno vial fibroblasts Specimens of synovium were collected at the time of surgery and transferred into DMEM. Isolated adherent synovial cells were prepared according to the method described by Dayer et al. [25]. Briefly, tissues were washed in Dulbecco's calcium and magnesiumfree PBS, cut into small 1- to 2-mm3 fragments and washed a further three times in PBS. The tissue was digested for 4 h in HBSS containing collagenese (1 mg/ml) and DNase 1 (100 gg/ml). The incubation mixture was centrifuged (400 0 for 10 rain at room temperature), the supernatant discarded and the pellet resuspended in PBS containing trypsin (500 gg/ml) and EDTA (200 gg/ml) before being incubated at 37 ~ for 1 h. Tissue clumps were removed by filtering through sterile gauze. The filtrate containing the individual cells was centrifuged, and the cells were washed thrice in DMEM before being transferred to a 25-cm2 tissue culture flask and cultured in DMEM containing 10% FCS at 37~ in a humidified 5% CO 2 air atmosphere. When confluent, the monolayer was detached by trypsinisation and then cultured in a 75-cm2 flask until confluent. At this stage the cells were detached, divided into three equal aliquots, each of which was cultured in a 75-cm2 flask. Experiments were performed when these cultures had become confluent. Synovial fibroblasts stained positively for vimentin when stained using the immuno-alkaline phosphatase technique.
After 3 days cells were washed five times and the medium changed to RPMI 1640 containing 20% FCS, and the cells were incubated under the same conditions for 24 h before any experiments were performed.
Lymphocytes The non-adherent cells from the mononuclear leucocyte suspensions used for the preparation of monocyte monolayers were washed in RPMI, resuspended to 2.5 x 106 cells/ml in RPMI containing 10% FCS and incubated at 37 ~ in a humidified 5% CO:/ air atmosphere. Total cellular RNA and Northern blotting Total cellular RNA was prepared from cells that had been cultured for 7 days using RNAzol, and Northern blotting was performed as described previously [23]. Blots were hybridised (incubation at 42 ~ overnight) to the 32p-labelled cDNAs (106 cpm/ml hybridisation fluid), washed to high stringency [0.1 x SSC containing 0.1% (w/v) SDS] at 65 ~ and subjected to autoradiography. Autoradiographs were scanned using a Joyce-Loebl chromoscan-3 (Joyce-Loebl, Gateshead, Tyne and Wear, United Kingdom). An arbitrary value of 1.00 was assigned to the control level of expression, Northern blots were stripped and reprobed with a synthetic oligonucleotide probe for 28S rRNA as described for dot-blot analysis (see above).
Measurement of proteins in culture fluids A set of each type of cell culture was incubated for 7 days. On days 1, 3, 5 and 7, the entire culture supematant was replaced and the medium stored at -70 ~ until used. On day 7 the cells were washed and adherent cells detached by trypsinisation. An aliquot was used for determining cell number and the remainder was used for RNA extraction (see above). The concentration of C1 q, C lr, C1 s, C4, C2, Cl-inh and C4-bp were determined by ELISA and, as the levels of complement components increased linearly with time, secretion rates of the proteins were calculated and expressed in molecules per cell per minute [23].
Endothelial cells
Statistics
Primary cultures of human umbilical vein endothelial cells (HUVEC) were prepared from freshly collected umbilical cords. Collagenase 0.1% (w/v) was used for treatment of the umbilical vein instead of 0.2% as originally described by Jaffe et al. [26]. Cells were grown initially at 37 ~ in a humidified 5% CO2/air atmosphere in RPMI 1640 containing 15% FCS and 10% ABS, 30 ng/ml endothelial cell growth supplement, 10 units/ml heparin, 2 m M t,-glutamine, 2.5 ~tg/ml fungizone and 100 g/ml penicillin-streptomycin. Cells were identified by their characteristic morphology and by the expression of factor VIII antigen. They were grown to confluence in 25-, 75- and 175-cm2 flasks coated with 1% (w/v) gelatin. The cells were subcultured by treating the monolayer with trypsin-EDTA, washing the freed cells and reseeding. The cells were used for experimental work during the fourth passage. ABS was omitted from the culture medium from the third passage onwards.
Differences between the relative mean values of the abundances of each mRNA in the different types of synovial membrane and the secretion rates of each complement component in the different cell types were analysed by Student's t-test.
Monocytes Human monocyte monolayers were prepared from the buffy coats of blood donations from normal individuals in 24-well Linbro tissue culture plates [23] and the cells were cultured in RPMI 1640 containing 10% ABS at 37 ~ in a humidified 5% CO2/air atmosphere.
Results Synovial tissue m R N A s D o t - b l o t s o f R N A f r o m R A , O A a n d n o r m a l synovial tissue gave positive h y b r i d i s a t i o n signals for all the m R N A s studied [ C l q A , C l q B , C l q C , C l r , C l s , C4, C2 (Fig. 1 a), C l - i n h , C 4 - b p a n d factor I (Fig. 1 b)]. The relative a b u n d a n c e s of the m R N A s for C l q A , C l q B , C l q C a n d C2 were similar in the three types o f tissue (Table 1). However, the a b u n d a n c e s o f C l r (t = 2.91, P < 0.05), C l s (t = 4.39, P < 0.01) a n d C4 (t = 3.26, P < 0.05) were increased in R A synovial m e m b r a n e c o m p a r e d with n o r m a l tissues. The fluid-phase c o n t r o l p r o t e i n s C l - i n h , C4-
16 Table 1 Relative abundance of mRNA for complement components in synovial tissue (OA osteoarthritis, RA rheumatoid arthritis)
Fig. l a, b Dot blot analysis of RNA from normal (N), osteoarthritis (OA) and rheumatoid arthritis (RA) synovial membrane. The dots (5 pg RNA) were probe a for mRNAs for the ClqA
(qA), ClqB (qB) and ClqC (qC) chains and Clr (lr), Cls (Is), C4 (4) and C2 (2) and b for mRNAs for Cl-inhibitor (inh) and C4-binding protein (bp). For each species of mRNA, single strips of Hybond N membrane carrying single dots of RNA from normal, osteoarthritis and rheumatoid arthritic synovial membrane were hybridised with the appropriate cDNA probe. Thus, intensities of the hybridisation signal for each species of mRNA could be compared directly
ClqA CIqB ClqC Clr Cls C4 C2 Cl-inh b C4-bp b Factor I b
Normal
OA
RA
1.00_+0.09 a 1.00-+-0.48 1.00+0.64 1.00_+0.35 1.00_+0.18 1.00• 1.00_+0.35 1.00 1.00 1.00
0.58_+0.38 0.63___0.08 0.84___0.12 1.62_+0.43 1.27_+0.13 1.20_+0.15 1.06-+0.43 1.80 1.10 0.70
1.25_+0.51 1.65_+0.15 0.46+0.21 3.78_+1.62" 1.68_+0.20"* 1.78_+0.06" 1.43_+0.50 4.40 3.00 4.30
* Difference between this value and normal was significant P < 0.05 ** Difference between this value and normal was significant P < 0.01 Results represent the mean _+ SEM of three experiments b Due to shortage of RNA only one value was obtained for the relative abundance of Cl-inh, C4-bp and factor I mRNAs a
bp and factor I appeared to be increased in R A synovium compared with OA and normal synovium (Table 1). Due to a shortage of R N A , only one sample in each group could be analysed for the presence of m R N A for each of the regulatory proteins, so these results must be interpreted with caution.
Cell m R N A s Single species o f m R N A were observed for C I q A (1.5 kb), C l q B (1.6 kb), C l q C (1.6 kb), C l r (2.4 kb), C l s (2.4kb), C2 (2.9kb; Fig. 2a), Cl-inh (2.1 kb), C4-bp (2.5 kb) and factor I (2.4 kb; Fig. 2b). Two bands were observed for C4 (5.4 kb, 3.0 kb) on Northern blots performed on R N A isolated from the cells.
Fibroblasts
Fig. 2a, b Northern blot analysis of RNA from monocytes, syn.ovial membrane fibroblasts (from normal synovial membrane) and umbilical vein endothelial cells (HUVEC). The blots were probed a for mRNAs for ClqA (qA), ClqB (qB) and ClqC (qC) chains and Clr (lr), Cls (Is), C4 (4) and C2 (2) and b for the mRNAs for Cl-inhibitor (inh), C4-binding protein (bp) and factor I. Although RNA preparations from fibroblasts from RA and OA synovial membrane were also analysed, they are not shown as they gave identical results to those of RNA from normal synovial membrane. The mRNAs encoding the ClqA and ClqC chains appear to be partially degraded. The hybridisation signal for factor I mRNA in HUVEC was weak and does not appear on the photograph. The sizes of the mRNAs were as follows: ClqA, (1.5 kb), ClqB (1.6 kB), ClqC (1.6kb), Clr (2.4 kb), Cls (2.4 kb), C4 (5.4 kb, 3.0 kb), C2 (2.9 kb), Cl-inhibitor (2.1 kb), C4-bp (2.5 kb) and factor I (2.4 kb)
Messenger R N A s ( m R N A s ) encoding Clr, Cls, C4, C2 (Fig. 2a), Cl-inh and factor I (Fig. 2b) were detected by Northern blots of R N A prepared from synovial fibroblasts isolated from all three types o f patient, m R N A s for C l q A , C l q B and C l q C (Fig. 2a) and C4-bp (Fig. 2b) were not detected. There did not appear to be any significant differences in the abundances of any of the m R N A s in fibroblasts from the three different sources.
Endothelial cells m R N A s encoding C I q A , C l q C , Cls, C4, C2 (Fig. 2a), Cl-inh and factor I (Fig. 2b) were detected on the Northern blots prepared from R N A isolation from three separate H U V E C cultures. The hybridisation signal for factor I m R N A was weak and did not p h o t o g r a p h well. This
17 C2) and the~ fluid-phase regulatory proteins (Cl-inh, C4-bp and factor I) by intact synovial membrane and by four types of cells (mononuclear phagocytes, fibroblasts, endothelial cells and lymphocytes) that are present in normal tissues and at sites of chronic inflammation. Our Component Cells data showed that mRNAs for each of these proteins were expressed in normal, OA and RA synovial membrane. In Monocyte Fibroblast" HUVEC a separate study, we have shown that C3, factor I, CR1, C1q 160,+27 b ND ND DAF and MCP mRNAs are also present [27]. Thus, the Clr 27-+ 9 218,+15 ND mRNAs for all the classical pathway components and the Cls 127-t-13 140-t-ll 189+30 secreted and cell-bound classical pathway regulatory C4 180+ 6 ND ND proteins are expressed within normal and inflamed synC2 42+_ 2 45_+ 5 40-+3 ovium. The biological importance of such synthesis has Cl-inh 161 • 30 418 • 80 20,+ 6 C4-bp 23,+ 3 ND ND yet to be determined. Although our data suggested that Factor I ND 39,+ 120 40 • 12 the abundances of some mRNAs were increased in RA synovium (Table 1) they must be interpreted cautiously Results for fibroblasts from normal synovialmembrane. The results for fibroblasts from RA and OA synovialmembraneswerenot as the tissue samples were small, and may not necessarily reflect events occurring in the entire synovial membrane. significantlydifferentfrom these However, as synthesis of complement components is b Data represents the mean and SEM of three determinations known to be modulated by cytokine action [11], and as cytokines are produced locally within joints [reviewed in may be due to instability of factor I mRNA in HUVEC. 28], it is probable that rates of synthesis of complement The mRNAs encoding ClqB chain, Clr (Fig. 2a) and components in chronically inflamed synovium differ from those in the normal tissue. However, the one report C4-bp (Fig. 2 b) were not detected. of an in vivo metabolic study that showed that half of the C3 present in the synovial fluid of a RA patient was synthesised locally [13] did not contain data on normal Monocytes individuals or OA patients, and we were unable to meamRNAs encoding ClqA, ClqB, ClqC, Clr, Cls, C4, C2 sure confidently secretion rates of the individual proteins (Fig. 2a), Cl-inh and C4-bp (Fig. 2b) were expressed by in synovial membrane fragments as it was impossible to suppress consistently the release of proteins from synmonocytes (Fig. 2). ovial tissue fragments with cycloheximide (2.5 ~tg/ml; data not shown). Thus, in our experiments the majority Secretion rates of complement components of the proteins accumulating in the culture supernatants could have resulted from contamination with plasma The secretion rates of Clq, Clr, Cls, C4, C2, Cl-inh, proteins. In a previous study [29] we were able to show C4-bp and factor I were linear throughout a 7-day period that cycloheximide suppressed synthesis of Cl-inh, C4, of culture. Fibroblasts did not secrete Clq, C4 or C4-bp, C2, C3 and B. We were unable to account for this disendothelial cells did not secrete Clq, Clr, C4 or C4-bp, crepancy, but minor differences in sample preparation but monocytes secreted all the components studied (C1 q, and washing were probably responsible. Further studies Clr, Cls, C4, C2, Cl-inh, C4-bp and factor I; Table 2). are required to determine whether the levels of expression Fibroblasts secreted higher levels of Clr than monocytes of complement components are different in normal, RA (t = 10.92, P < 0.001). The secretion rates of C2 and Cls and OA synovium. In addition, we did not have data on were similar in each cell type. Fibroblasts had a higher the disease activity or medication in these patients. The synthesis rate of Cl-inh than monocytes ( t = 3.01, effects of these on complement gene expression also merP < 0.05), while the rate in monocytes was higher than its investigation with larger numbers of patients. that in HUVEC (t = 4.61, P < 0.01). The molar ratio of There are a variety of cells in normal synovial memClr : Cls was approximately 1 : 1 in fibroblasts and 1 : 5 branes including type A (macrophage-like) and B (fiin monocytes (Table 2). The synthesis rates of comple- broblast-like) synoviocytes, fibroblasts, macrophages, ment components were similar in all three types of fibrob- endothelial cells and adipocytes [7]. In OA, hyperplasia of last. Lymphocytes did not secrete any of the complement synoviocytes occurs and may be associated with the accuproteins investigated in this study. mulation of small lymphoid aggregates [30]. In RA, hyperplasia of the synovial lining cells is associated with marked chronic inflammation that is characterised by Discussion infiltration with macrophages, lymphocytes and fibroblasts and the formation of new blood vessels [12]. Thus, We used a combination of immunochemical and molecu- in synovium from normal individuals and patients with lar biological techniques to study the synthesis of the OA or RA, there are several cell types that are capable of classical pathway components (Clq, Clr, Cls, C4 and synthesising complement components. Previous studies
Table 2 Secretion rates of complement components (molecules/ minute/cell)in differentcell types. Results of statistical analysisare given in the text (ND not detected, HUVEC human umbilicalvein endothelial cells)
18 have shown that monocytes express Clq, Clr, Cls, C4, C2, Cl-inh and C4-bp, fibroblasts express Clr, Cls, C2 and Cl-inh and endothelial cells express C3 and Cl-inh [reviewed in 11]. Endothelial cells failed to synthesise Clq, probably because the CIqB chain m R N A was not expressed, whereas the failure of fibroblasts to synthesise this protein was associated with lack of expression of all three C l q mRNAs. These data suggested that there may be cell-specific differences in the expression of the three C l q genes, and that those for the C l q A and C chains must be able to be expressed independently of that for the C l q B chain. As the gene encoding the C l q B chain is located 3' to those encoding the A and C chains [1], independent regulation of the B chain gene expression is a possibility. Although in these experiments we were unable to detect C l r secretion by HUVEC, we have since been able to detect low levels of C l r in the supernatants o f cultures that have been unchanged for 72 h. Although endothelial cells and fibroblasts both expressed C4 m R N A , C4 protein was not detected in the culture supernatants. Fibroblast R N A gave a strong C4 hybridisation signal with C4 cDNA, whereas the signal with H U V E C R N A was weak. As it has been shown previously that synovial membrane fibroblasts do not incorporate 35S-methionine into immunoprecipitable C4 [31], the failure to express C4 must be due to a translational block. The failure to detect C4 protein in H U V E C culture supernatants could be simply a problem of sensitivity of the ELISA procedure (lower limit of sensitivity 0.5 ng/ml). Although factor I secretion values were similar in H U V E C and fibroblasts, the hybridisation signal obtained with factor I c D N A was very weak in H U V E C , whereas it was very strong in endothelial cells. This difference in abundance could be due to transcriptional or post-transcriptional effects and the possibility o f cellspecific differences in factor I m R N A translational efficiency. Synovial fibroblasts, like skin fibroblasts and HepG2 cells (unpublished observations), secreted approximately equimolar amounts of C1 r and C I s, whereas monocytes secreted about five-fold more Cls and C l r and endothelial cells appeared to synthesise only Cls. These observations suggested that although the genes encoding C l r and Cls are reduplications, and are located within a 50 kb stretch of D N A on chromosome 12p13 [2], they must be capable of being regulated independently. Their tail-totail orientation supports this notion. In summary, the data presented in this paper showed that all the components and the regulatory proteins of the classical pathway are synthesised within normal, OA and RA synovial membrane, with mononuclear phagocytes, endothelial cells and fibroblasts being the cells most probably involved, although confirmation of this conclusion must await the results of combined in situ hybridisation and immunohistochemical studies on intact synovial membrane [32]. This is particularly important with respect to complement synthesis by endothelial cells as H U V E C may not be representative of vascular endotheli-
um in general. As the range of components synthesised by each cell type and the level of expression of each component are not always the same in each cell type, cell-specific mechanisms that regulate complement gene expression must exist. These mechanisms and those involved in the regulation of complement gene expression by cytokines synthesised within the joint require investigation. Acknowledgements This work was supported by a grant from the Arthritis and Rheumatism Council for Research and the Scottish Home and Health Department. C. Lemercier was in receipt of an EEC bursary.
References 1. Sellar GC, Blake DJ, Reid KBM (1991) Characterisation and organisation of the genes encoding the A-, B- and C-chains of human complement subcomponent Clq. The complete derived amino acid sequence of human Clq. Biochem J 274:481-490 2. Nguyen VC, Tosi M, Gross MS, Cohen-Haguenauer O, JejouFoubert, Cole Tarid ME Meo T, Frezal J (1988) Assignment of the complement serine protease genes Clr and Cls to chromosome 12 region 12p13. Hum Genet 4:363-368 3. Weiler JM (1993) Introduction to complement. In: Whaley K, Loos M, Weiler JM (eds) Complement in health and disease, 2nd edn. Kluwer, London New York, pp 1-37 4. Kohl J, Bitter-Suermann D (1993) Anaphylatoxins. In: Whaley K, Loos M, Weiler J (eds) Complement in health and disease, 2nd edn. Kluwer, London New York, pp 299-324 5. Law SKA (1993) Complement receptors. In: Horton MA (ed) Blood cell biochemistry, 5, macrophages and related cells. Plenum Press, New York, pp 223-259 6. Morgan BP (1993) Cellular responses to the membrane attack complex. In: Whaley K, Loos M, Weiler J (eds) Complement in health and disease, 2nd edn. Kluwer, London New York, pp 325-351 7. Sledge CB (1989) Biology of the joint. In: Kelley WN, Harris ED, Ruddy S, Sledge CB (eds) Textbook of rheumatology. Saunders, Philadelphia London Toronto Montreal Sydney Tokyo, pp 1-21 8. Ruddy S, Austen KF (1973) Activation of the complement system in rheumatoid synovium. Fed Proc 32:134-137 9. Moxley G, Ruddy S (1985) Elevated C3 anaphylatoxin levels in synovial fluids from patients with rheumatoid arthritis. Arthritis Rheum 28:1089-1095 10. Morgan BP, Daniels RH, Williams BD (1988) Measurement of terminal complement complexes in rheumatoid arthritis. Clin Exp Immunol 73:473-478 11. Colton HR, Strunk RC (1993) Extrahepatic synthesis of complement. In: Whaley K, Loos M, Welter JM (eds) Complement in health and disease, 2rid edn. Kluwer, London New York, pp 127-158 12. Gardner DL (1992) Rheumatoid arthritis, cell and tissue pathology, in: Gardner DL (ed) Pathological basis of connective tissue disease. Edward Arnold, London, pp 444-526 13. Ruddy S, Colten HR (1974) Biosynthesis of complement proteins by synovial tissue. N Engl J Med 290:1284-1288 14. Chung LP, Bentley DR, Reid KBM (1985) Molecular cloning and characterisation of the cDNA coding for C4b-binding protein a regulatory protein of the classical pathway of the human complement system. Biochem J 230:133-141 15. Journet A, Tosi M (1986) Cloning and sequencing of full length cDNA encoding the precursor of human complement component Clr. Biochem J 240:783-787
19 16. Tosi M, Duponchel C, Meo T, Julier C (1987) Complete cDNA sequence of human complement Cls and close physical linkage of the homologous genes Cls and Cir. Biochemistry 26: 85108524 17. Carroll MC, Parker RR (1983) Cloning of a human complement component C4 gene. Proc Natl Acad Sci USA 80: 264269 18. Bentley DR, Porter RR (1984) Isolation of eDNA clones for human complement component C2. Proc Natl Acad Sci USA 81:1212-1215 19. Carter PE, Dunbar B, Fothergill JE (1988) Genomic and eDNA cloning of the human C1 inhibitor intron-exon junctions and comparison with other serpins. Eur J Biochem 173:163169 20. Catteral CF, Lyons A, Sinn RB, Day AJ, Harris TJ (1987) Characterisation of primary amino acid sequence of human complement control protein factor I from an analysis of eDNA clones. Biochem J 242:849-856 21. Sambrook J, Fritsch EF, Manniatis T. Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 22. Feinberg AP, Vogelstein A (1984) A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 137:266-267 23. Lappin DF, Birnie GD, Whaley K (1990) Modulation by interferons of the expression of monocyte complement genes. Biochem J 268:387-392 24. Barber V, Dautry F (1989) Northern blot normalisation with a 28S rRNA oligonucleotoide probe. Nucleic Acids Res 17:7115
25. Dayer JM, Krane SM, Russell RG (1976) Production of collagenase and prostaglandins by isolated adherent rheumatoid synovial cells. Proc Natl Acad Sci USA 73:845-849 26. Jaffe EA, Nachmann RL, Becker CB, Minick CR (1973) Culture of human endothelial cells derived from umbilical vein: identification by morphologic and immunologic criteria. J Clin Invest 52:2743-2756 27. Guc D, Gulati P, Lemercier C, Lappin D, Birnie GD, Whaley K (1994) Expression of the components and regulatory components of the alternative complement pathway and the membrane attack complex in normal and diseased synovium. Rheumatol Int 13:139-146 28. Arend WP, Dyer JM (1990) Cytokines and cytoline inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum 33: 301-315 29. Moffat GM, Lappin D, Birnie G, Whaley K (1989) Complement biosynthesis by synovial tissue. Clin Exp Immuno178: 5460 30. Athanosou NA, Quinn J, Aeryet A, Pudle B, Woods CG, McGee JO'D (1988) The immunohistochemistry of synovial lining cells in normal and inflamed synovium. J Pathol 155: 133-142 31. Katz Y, Strunk RC (1988) Synovial fibroblast like cells synthesise seven proteins of the complement system. Arthritis Rheum 31:1365-1370 32. Firestein GS, Pane MM, Littman BH (1991) Gene expression (collagenase, tissue inhibitor of metalloproteinase, complement and HLA-DR) in rheumatoid arthritis and osteoarthritis synovium. Arthritis Rheum 34:1094 1105
