Research Paper Mediators of Inflammation, 7, 347–353 (1998)

We investigated the serum concentrations of interleukin-6 (IL-6) and two IL-6 family of cytokines (leukaemia inhibitory factor (LIF) and ciliary neurotrophic factor (CNTF) as well as IL-6 soluble receptor (sIL-6R) using an enzyme-linked immunosorbent assay (ELISA) in 66 patients with rheumatoid arthritis (RA) and 24 healthy controls. We examined a possible association between the serum levels of these peptides and RA activity according to the Mallya and Mace scoring system and Ritchie's index. We also evaluated the correlation between the serum levels of IL-6, LIF, CNTF and sIL-6R and duration of the disease and calculated sIL-6R/IL-6 ratio in RA patients and in the control group. IL-6 and sIL-6R were detectable in all 66 patients with RA and 24 normal individuals. LIF was also found in the serum of all patients with RA and in 16 (66.7%) normal individuals. In contrast CNTF was measurable only in 15 (22.7%) patients with RA and 24 (33.3%) normal individuals. The highest IL-6 and sIL-6R levels were found in the patients with Stages 3 and 4 of RA activity and the lowest in the control group. In contrast there were no statistically significant differences between the LIF and CNTF levels in RA patients and normal individuals. We found positive correlation between IL-6 and sIL-6R concentrations and Ritchie's index and a lack of such correlation with LIF and CNTF. IL-6 serum level correlated positively with the disease duration, but sIL-6R, LIF and CNTF did not. Serum sIL-6R/IL-6 ratio was significantly lower in RA patients than in healthy controls. In conclusion, an increase in the serum levels of IL-6 and sIL-6R, but not LIF and CNTF concentrations, may be useful markers for RA activity.


Introduction
Rhe umatoid arthritis (RA) is a syste mic autoimmune disease localized pre fere ntially in the synovial joints, re sulting in joint de struction and pe rmane nt disability. 1 ,2 The re is grow ing evide nce suggesting that some cytokines, particularly proinflammatory cytokines such as tumour ne crosis fac tor a (TNF-a ), inte rleukin-1 (IL-1), inte rferon g (IFN-g ) and inte rleukin-6 (IL-6), play an important role in the pathogene sis of this disease. 3 ,4 The se inflammatory c ytokine s are pre se nt in the rhe umatoid synovial membrane and participate in c ell proliferation as w e ll as in the synthesis of prostaglandins, metallop rote inases and other cytokine s. 5 -7 IL-6 is a pleiotrop ic, immunomodulatory cytokine produce d by a variety of ce ll types, including fibroblasts, endothe lial c ells, monoc ytes and both be nign and malignant lymp hocyte s of B and T ce ll origin. 8 It is a multifunc tional cytokine, w hich plays a ke y role in the differe ntiation and grow th of hae matopoie tic ce lls, B-ce lls, T-c ells, keratino c ytes, ne uronal cells, oste oclasts and e ndothelial cells. 9 More ove r, IL-6 modulates the transcription of se veral liver-specific gene s during ac ute inflammatory state s. Pe rtinent to inflammation is the ability of IL-6 to induc e ac ute phase protein synthesis in hep atocytes. 1 0 IL-6 is a membe r of a family of cytokines w hich also include s le ukaemia inhibitory factor (LIF), oncostatin M (OSM), c iliary ne urotrophic fac tor (CNTF), and inte rleukin-11 (IL-11). 1 1,1 2 These c ytokine s are a group of evolutionary re late d prote ins characte rize d by a common te rtiary framew ork, w ith a distinc tive, four helix bundle topology. 1 3,1 4 IL-6 typ e cytokines induc es grow th or differe ntion via a re c eptor syste m that involves a spec ific re c eptor and the use of a share d signalling subunit, gp 130. 15 Ce lls is olated from the synovium of RA patients ex pre sse d mRNA for LIF, OSM and IL-11 at higher levels than did synovial ce lls from oste oarthritis patients, and spontaneously re le as ed gre ater quantities of this protein in culture. 1 6 Elevated le vels of LIF, OSM and IL-11 w ere also found in the synovial fluids of RA patie nts. 16 -18 How ever OSM, LIF and IL-11 have be e n re ported to be undete c table in periphe ral blood of these patients. 1 6 To our know ledge the le ve l of CNTF in the serum and synovial fluid of RA patients has be e n not inve stigate d, so far. How e ver, w e have found a dete ctable level of this cytokine in the serum of the majority of syste mic lupus e rythe matosus (SLE) patie nts. 1 9 The ce llular IL-6 re ce ptor complex c onsists of tw o differe nt prote ins, an 80-kDa ligand binding glyc oprote in (IL-6R) and gp 130 involve d in ce llular signal transduction. 20 These tw o subunits of the IL-6R complex are prote olytic ally cleave d and re le as ed from the cell as soluble re c eptor prote ins. 21 Soluble forms of the IL-6R (sIL-6R) and gp 130 are found in diffe re nt body fluids in patie nts w ith various inflammatory diseases, including SLE and RA. 22 ,23 In the pre sent study w e me as ure d the serum conc entration of IL-6, LIF, CNTF and sIL-6R in patients w ith RA using ELISA as say. We c orrelated the serum levels of these prote ins w ith dise ase activity ac cording to the Mallya and Mac e sc oring syste m and Ritchie 's index as w ell as w ith disease duration and type of tre atment. We also e valuate d the corre lation be tw e en the serum le vels of these thre e c ytokine s and IL-6R as w ell as IL-6 w ith LIF and CNTF.

Patients
The study involve d 66 patients (57 females and nine males). The ir mean age w as 51.1 ye ars (range 20 -78 ye ars). The y fulfille d the re vised crite ria of the American Rheumatism Association (ARA), crite ria for diagnosis of RA. 2 4,2 5 The me an duration of the ir dise as e w as 8.7 years (range from 1 month to 30 years). Forty-six of them w ere found positive for the rheumatoid factor (RF) and 20 w ere ne gative. Most of the RA patients w ere re c eiving non-ste roidal anti-inflamatory drugs (NSAID). In addition, 26 of the patients w ere treate d w ith pre dnisone , e ight w e re taking D-pe ncillamine, nine -gold salts, e ight methotrex ate and e ight sulfasalazine. On the day of blood sampling, e ach patient's history w as re corde d and a physic al ex amination w as performed. Clinic al variable s included dise as e duration, applie d tre atment and a joint c ount for pain/te nderness. Ritchie 's articular index (RI), 2 6 duration of the morning stiffness and vis ual analogue pain score (VAS) w ere re c orded by the same observer (A.G.).
Dise ase ac tivity w as score d during visits to the outpatient clinic ac cording to the me thod de scribe d by Mallya and Mac e. 2 7 Our group of patients included 16 patients in Stage 1 or 2, 34 in Stage 3 and 16 in Stage 4 of the dise as e ac tivity, ac cording to this classific ation. The clinical and laboratory feature s of RA patie nts are pre se nte d in Table 1.
The control group c onsiste d of 24 healthy individuals, 19 w omen and 5 me n, aged from 24 to 68 years (mean 51.2 years). Each unde rw e nt a thorough physical ex amination pe rforme d by one of the authors (A.G.).

Laboratory tests
On the day of blood sampling for c ytokines the follow ing laboratory parame te rs w ere analysed: complete blood cell count (CBC), We ste rgre n erythrocyte sedimentation rate (ESR), urinalysis blood ure a nitrogen and c re atinine leve ls, fibrinog e n le ve l, liver function te sts (GOT, GPT, bilirubin ), C-re active prote in (CRP), serum iron and rheumatoid fac tor (RF). Che st radiographs, abdomen ultrasonography and EKG w ere also performed.

Serum sampling and cytokine determination
Venous blood samples for IL-6, IL-6 re lated cytokines (LIF and CNTF) and sIL-6R w e re collecte d at the time of clinic al asse ssme nt into p yroge n free tubes, allow e d to clot at -4°C for 1 h and c entrifuged at 20003 g for 10 min. The obtained serum w as divide d into aliqots and stored at -25°C until assayed for IL-6, LIF, CNTF and sIL-6R. The se ra w ere randomly code d and te sting w as c arried out w ithout know le dge of the clinical status of the subje ct or of re late d laboratory data. The c ytokine serum c onc entration w as assaye d by sp ecific, c ommercially available , enzyme linke d (ELISA) assay kits (Quantikin e, R&D Syste ms Inc, USA) in ac cordance w ith the manufac ture r's instructions and analysed w ith an ELISA re ader at 492 mm.
The p roc edure w as desc ribe d in details e lsew here . 1 9,28 In brie f, the monoclonal antib odie s, spe cific for the c ytokines w ere place d onto the mic rotitre plates provide d w ith the kits. Standard and the samples w ere pipetted into the w ells and any pre sent cytokines w ere bound by the immobilizing antib ody. After rinsing e nzyme -linked p olyclonal antib odie s, spe cific for cytokines, w e re added to the w ells to sandw ich the cytokines immobilize d during the first incubation. After the nex t rinsing the substrate solution w as adde d to the w ells and the colour developed proportionally to the amount of cytokines bound in the first ste p. In e ach assay the appropriate re c ombinant human cytokine w as use d to ge ne rate the standard c urve. The c oncentration of cytokines and sIL-6R in the samples w ere dete rmine d by interpolation from the standard curve . Serum sIL-6R conce ntration measurement w as diluted 40 times and its le vel w as me as ure d be tw ee n 7.8 and 500 pg/ ml. Se nsitivity of the assay for IL-6 w as 0.7 pg/ml, for LIF 2.0 pg/ml and for CNTF 8.0 pg/ml.

Statistical analysis
The mean value s w ere compare d in Kruskal Wallis and Mann-Whitne y te sts . Diffe re nc es in paramete rs be tw e en the groups w ere evaluate d w ith Student's t-te st. Statistical analysis for the frequenc y of de te c table cytokines w as performed using chi-squared te st. The linear c orre lations be tw e en serum interle ukin levels compared w ith e ach othe r or w ith Ritchie 's index w ere e valuated using the Sp earman rank-sum correlation coefficie nt and linear re gre ssion calc ulate d w ith the le as t-square s me thod. Results are pre se nte d w ith R 2 coefficie nts. Comparison and correlation w ere cons idere d signific ant w hen P < 0.05. . How e ver, LIF w as more freque ntly dete ctable in RA patie nts than in the control group (P < 0.03). We found a positive correlation be tw ee n IL-6 as w ell as sIL-6R c onc entrations and Ritchie 's index (R 2 = 0.1404, P < 0.002 and R 2 = 0.0952, P < 0.02, re spe ctively) but no such correlation be tw een LIF and CNTF se rum le ve ls w ith Ritchie 's index (R 2 = 0.0035, P > 0.05 and R 2 = 0.002, P > 0.05 re spec tively) (Fig. 1). We also analyse d the re lationship be tw ee n se rum conc entrations of IL-6, sIL-6R, LIF and CNTF w ith the duration of the disease (Fig. 2). We obse rved a positive correlation be tw e en the se paramete rs only in the case of IL6 (R 2 = 0.1401, P < 0.002), but no such correlation in the c ase of sIL-6R, LIF and CNTF (R 2 = 0.0420, P > 0.05, R 2 = 0.0001, P > 0.05 and R 2 = 0.0020, P > 0.05 re sp ective ly).

Results
The re lationship be tw e en the serum le vels of sIL-6R w ith e valuated c ytokines, as w ell as be tw ee n particular c ytokine s alone w as also analysed (Fig. 3). We observe d no c orre lation be tw ee n any of the compared paramete rs. The c alc ulation of the ratio of sIL-6R to IL-6 in patients w ith RA ac cording to disease activity is pre se nted in Table 3.
We found a significantly low e r ratio of sIL-6R to IL-6 in RA patie nts compare d w ith normal pe rsons (1406.3 and 8806.0 re spec tively). The re w as also significant differe nc e be tw e en the sIL-6R to IL-6 ratio in Stages 1 and 2 and Stage 4 RA ac tivity according to Mallya and Mace (P < 0.05).
The influe nce of the tre atment schedule on the se rum conce ntrations of IL-6, sIL-6R, LIF and CNTF has be e n also analyse d (data not pre sented). How e ve r w e found no statistically significant differe nce s be tw e en the levels of dete cte d cytokines or sIL-6R in the patients treated w ith NSAID only, pre dnisone , me thotrex ate , D-pe nicillamine, gold salts or sulfasalazine.

Discussion
The role of cytokines in the pathogene sis of RA and their signific anc e in c linic al monitoring of the disease advance ment has be e n attac ting much attention since a few years. 3 -6 In this disease the articular synovial me mbrane is infiltrate d w ith inflammatory cells diffusing into the synovial fluid. The inflammatory proce ss spre ads from the synovial membrane to the cartilage and bone tissue c ausing the ir damage . 3 Cytokine s p lay a crucial role in sustaining an inflammatory proce ss w ithin synovial membranese. So far, most atte ntion has be en paid to the role of inflammatory cytokines in RA pathogene sis, espe cially TNF-a , IL-1, IFN-g and IL-6. 3,5 ,2 2 The ac tivity of othe r cytokines c onstituting IL-6 group, like LIF, OSM and IL-11 in RA w as a subject of single re ports and the role of CNTF in this disease has not be en investigated so far. 16 -18 ,29 Unlike the soluble TNF re ce ptors the role of sIL-6R in RA and the prognostic meaning of this re c eptor are also poorly investigate d. 3

0,31
In our studies w e as sessed se rum conce ntration of IL-6, LIF, CNTF and sIL-6R in 66 patients w ith RA in differe nt stages of dise as e advanc eme nt and duration as w e ll as in 24 healthy controls. IL-6, sIL-6R w ere dete cte d in the serum of all patie nts w ith RA and all he althy pers ons. De te c table le vels of LIF w ere found in all RA patients but only in 16 (66.7%) he althy c ontrols, and CNTF in 15 (22.7%) patients w ith RA and e ight (33.3%) individuals from the control group.
The c once ntration of IL-6 in blood serum of RA patients w as six -fold higher than in healthy pe rsons and corre late d w ith the dise ase activity and its duration. The se re sults are c onsiste nt w ith the observations made by Madhok e t a l. 32 and van Lee uw e n e t a l. 33 Moreove r, high c oncentration of IL-6 in the synovial fluid of RA patients and the corre lation be tw e en the c onc entration of this cytokine and the inte nsity of bone le sions seen in radiograms, as w ell as its signific ant role in the joint destruction, w ere demonstrate d. 3 0,3 4 Also clinic al improve ment and the decre as e of C-re active prote in in the serum of the Mediators of Inflammation · Vol 7 · 1998 patients treated w ith antibodies against CD4+ lymphocytes, and then antibodies ne utralizing IL-6, may prove the significanc e of IL-6 in the pathoge ne sis of RA. 35 The re sults of our studies may indicate that tw o othe r cytokines re late d to IL-6 i.e . LIF and CNTF, unlike IL-6 are of smaller pathogene tic signific anc e and of no use in de te rmining RA activity. Their conc entration in the se rum of patients did not correlate w ith the de gree of RA activity and w as similar to their se rum conc entration in healthy individuals. Our re sults diffe r from those re porte d by Okamoto e t a l. 1 6 w ho show ed that the isolated from the synovium of RA patients cell produce gre ater quantities of LIF than patie nts w ith oste oarthritis, ne ve rthe less, the y w ere unable to dete ct this cytokine in blood serum of patients w ith RA. These differe nc es may ste m from differe nt sens itivity of ELISA te sts applied in both studie s. In the studies of Okamoto e t a l. the sensitivity re ached 15 pg/ml, w hile in our te sts the sens itivity w as much higher and amounte d to 2 pg/ml. High conce ntration of LIF in the synovial fluid of patie nts w ith RA w as also observe d by Waring e t a l. 1 7 It should be emphasize d that w e dete cte d LIF pre viously in the se rum of some SLE patients and the conc entration of this cytokine corre late d w ith the disease activity w hen using the same ELISA method. 19 Ele vated le ve ls of LIF w ere also observe d in the se rum of patie nts w ith non-Hodgkin's lymphoma, Hodgkin's dise ase , chronic lymphoc ytic leukae mia as w ell as in tissue shock. 3 6,37 Although it should be emphasize d that in those studies the ELISA method w ith p olyclonal antirabbit antib odie s w as employe d, w hich is probably more sens itive than the c ommercially available te sts used in our studie s i.e. Quantikin e, R&D Syste ms Inc. w ith monoclonal antib odies. 3 8 To our know le dge CNTF has so far not be en asse sse d in the serum of RA patients. In our studies this cytokine w as dete ctable only in 15 out of 66 (22.7%) patients w ith RA and in e ight out of 24 (33.3%) of he althy individuals. This c ytokine w as pre viously as sessed by us in SLE patie nts, and w as pre sent in 52 out of 64 of the ex amined patients. 19 In SLE w e also de monstrated a c orrelation be tw ee n the conc entration of CNTF and the activity of the disease. It may indicate that CNTF has a greater significance in the pathogene sis of SLE than in RA.
The dete rmination of the role of IL-6 group cytokines in the pathogene sis of RA re quires furthe r investigation. How ever, the already available data indicate that the role of LIF, CNTF, OSM and IL-11 in the initializing and sustaining inflammation is le ss important than the role of IL-6 alone . Gabay e t a l. 3 9 show ed that LIF and IL-11 stimulate human hepatocyte s to produc e the acute phas e protein much poore r than IL-6, although in the ex pe riments on mic e it w as found that CNTF stimulate s production of the hepatic ac ute phase proteins to the same ex te nt as IL-6. 40 It is not clear, how e ve r, w he ther this cytokine has a similar biological activity in ex perimental animals and humans. The re sults of our re se arch are cons is te nt w ith the obse rvation s of Okamoto e t a l. 1 6 and indicate that despite the func tional similarity of IL-6 group c ytokine s, IL-6 is in this group the main me diator of inflammation in RA.
Soluble IL-6 re c eptor (sIL-6R), unlike other soluble cytokine re cep tors, has the unique property of ac ting agonistically w ith its ligand and e nhanc es the stimulation of this c ytokine on gene rating acute phase prote ins (APP) by human hepatocytes, 4 1 as w ell as prolife ration of mye loma ce lls 4 2 and synovial fibroblasts. 43 In our studies w e found p ositive c orre lation be tw e en the se rum c once ntration of sIL-6R in patients w ith RA and Ritchie 's index . We also demonstrate d significantly low er values of sIL-6R/IL-6 ratio in RA p atie nts than in he althy subje cts and low er value s of this index in the most active form of the disease (Stage 4 of RA activity according to Mallya and Mac e) than in less ac tive forms (Stages 1 and 2). We did not re veal how e ve r, c orrelation be tw ee n the c oncentrations of IL-6 and sIL-6R in the serum of RA patients. These observations indicate a c omplex and not ve ry clear re lation be tw e en these tw o prote ins in this disease.
Studie s of Kotake e t a l. show ed highe r c onc entrations of IL-6 and sIL-6R in the synovial fluid in RA patients than in p atie nts w ith oste oarthritis and correlation of these pe ptides w ith the degre e of joint destruc tion. 30 Though contradictory re sults w ere obtain ed in patie nts w ith juvenile rheumatoid arthritis, w he re ne gative corre lation be tw een the conc entration of sIL-6R and IL-6 in serum w as obse rved, and the c oncentration of sIL-6R in those patie nts w as low er than in he althy persons . 44 How e ver it should be emphasize d that the conc entration of sIL-6R in patients w ith SLE, AIDS and multiple mye loma w as significantly higher than in healthy pe rsons. 42 ,4 5 Neverthele ss, like in the c ase of our RA patie nts, no correlation be tw ee n the serum conc entration of sIL-6R and IL-6 w as found in multiple mye loma. 46 These data indic ate that differe nt age nts influenc ing the production of both prote ins, may be pre sent in differe nt dise as es.
In conclusion w e c an state that the serum conc entration of IL-6 and sIL-6R in patients w ith RA is higher than in healthy pe rsons, and these p eptides may se rve as markers of this dise as e activity. The c oncentration of othe r c ytokine s of IL-6 group (LIF and CNTF) in se rum of RA patients and he althy individuals ne ither differs significantly nor correlates w ith the ac tivity of the disease.