ARTHRITIS & RHEUMATISMVol. 43, No. 9, September 2000, pp 1951–1959© 2000, American College of Rheumatology

SYNOVIAL MACROPHAGE DEPLETIONWITH CLODRONATE-CONTAINING LIPOSOMES

IN RHEUMATOID ARTHRITIS

PILAR BARRERA, ARJEN BLOM, PETER L. E. M. VAN LENT, LOUIS VAN BLOOIS, JOS H. BEIJNEN,NICO VAN ROOIJEN, MAARTEN C. DE WAAL MALEFIJT, LEO B. A. VAN DE PUTTE,

GERT STORM, and WIM B. VAN DEN BERG

Objective. To assess whether intraarticular (IA)administration of clodronate liposomes results in localmacrophage depletion in patients with rheumatoid ar-thritis (RA). Primary goals were to address both theimmunohistologic and potential toxic effects of thisapproach. Moreover, the correlation between immuno-histologic findings and clinical assessments of diseaseactivity and cartilage damage were assessed.

Methods. An open study was conducted in consec-utive RA patients who were scheduled for knee jointreplacement in our department. Synovial biopsy tissuewas obtained from the knee joint at 2 weeks before andat the time of surgery. This protocol was controlled forsafety and immunohistologic concordance in 6 patients.One week before surgery, 10 patients received a singleIA dose of clodronate liposomes. Staining of synovialtissue for cell markers (CD68, CD14, CD3, CD38) andadhesion molecules (vascular cell adhesion molecule 1[VCAM-1], intercellular adhesion molecule 1 [ICAM-1]) was assessed by 2 blinded observers. Local andsystemic parameters of disease activity were measuredbefore each intervention. Cartilage damage was scoredusing standard radiologic techniques at baseline andduring surgery.

Results. A single IA dose of clodronate liposomessignificantly reduced the number of CD68-positive cells(P 5 0.005) and the expression of ICAM-1 and VCAM-1in the synovial lining (P 5 0.013 and P 5 0.039,respectively). The intervention did not affect fibroblast-like synoviocytes, T cells, or plasma cells. No immuno-histologic changes were observed in the control group.The procedure was well tolerated. The levels of ICAM-1and VCAM-1 in the sublining layers correlated with theextent of macroscopic synovitis (P < 0.0005 and P <0.005, respectively). The expression of ICAM-1 andCD14 in the sublining correlated with the levels ofC-reactive protein (P < 0.0005 and P < 0.01, respec-tively). Cartilage destruction was correlated only withthe expression of CD68 in the sublining (P 5 0.02).

Conclusion. A single IA administration ofclodronate liposomes leads to macrophage depletionand decreased expression of adhesion molecules in thesynovial lining in patients with longstanding RA. Theprocedure is well tolerated, and its therapeutic potentialis currently under investigation. The expression of ad-hesion molecules in the sublining layers reflects ongoinginflammation.

Rheumatoid arthritis (RA) is a common auto-immune disorder that is characterized by chronic arthri-tis leading to irreversible joint destruction. The patho-genesis of RA is not fully understood, but currentmodels are based on the recognition of peptides boundto HLA class II molecules on macrophages and otherantigen-presenting cells by T cells. Rheumatoid synovialtissue is characterized by accumulation of inflammatoryand mesenchymal cells (1), hyperplasia and activation ofsynovial fibroblasts in the lining (2), and neovasculariza-tion (3) and pannus formation resulting in cartilagedestruction. Unfortunately, specific approaches aimed at

Supported by the Dutch League Against Rheumatism (grant97.1.101).

Pilar Barrera, MD, PhD, Arjen Blom, MSc, Peter L. E. M.van Lent, PhD, Maarten C. de Waal Malefijt, MD, PhD, Leo B. A. vande Putte, MD, PhD, Wim B. van den Berg, PhD: University HospitalNijmegen, Nijmegen, The Netherlands; Louis van Bloois, MSc, GertStorm, PhD: Utrecht University, Utrecht, The Netherlands; Jos H.Beijnen, PhD: Slotervaart Hospital, Amsterdam, The Netherlands;Nico van Rooijen, PhD: Free University, Amsterdam, The Nether-lands.

Address reprint requests to Pilar Barrera, MD, PhD, Depart-ment of Rheumatology, University Hospital Nijmegen, PO Box 9101,6500 HB Nijmegen, The Netherlands.

Submitted for publication February 24, 2000; accepted inrevised form May 18, 2000.

1951

preventing cell influx and neovascularization or normal-izing fibroblast differentiation are not yet available.Moreover, trials targeting T cells in RA have yieldeddisappointing results (4,5).

Evidence accumulated in recent years stronglysuggests that the presence of synovial macrophages isessential for the initiation and maintenance of arthritis(6,7) and that the number of synovial macrophages atthe target organ may predict joint destruction (8,9). Thishypothesis is supported by the clinical improvementobserved after specific neutralization of macrophage-derived cytokines such as tumor necrosis factor (TNF)and interleukin-1 (IL-1) in patients with RA. It shouldbe kept in mind that in addition to proinflammatorycytokines, activated monocyte/macrophages produce alarge array of mediators, including metalloproteinases,reactive oxygen radicals, and arachidonic acid deriva-tives (7). Counteracting the monocyte/macrophage sys-tem itself could therefore be more effective than block-ing single macrophage-derived mediators.

Because of their phagocytic capacity, cells fromthe monocyte/macrophage lineage can be targeted withthe use of conventional liposomes. These liposomeshave been used to deliver antimicrobials, cytostaticagents, immunomodulators, antigens, and genes directlyinto macrophages (10). Such macrophage-directed drug-delivery systems can also be used to yield a selectivedepletion of these cells in several organs. This can beachieved using liposomes encapsulating clodronate (di-chloromethylene bisphosphonate [Cl2MDP]) (11). Clo-dronate is a bisphosphonate that has been used in thetreatment of skeletal diseases because of its capacity toprevent bone resorption (12). This drug also has antiin-flammatory effects in experimental arthritis (13) andreduces the production of several proinflammatory me-diators in vitro (14–16). Clodronate is highly watersoluble and poorly lipid soluble. The antiinflammatoryeffects and intracellular concentrations of this drug areenhanced by encapsulating the drug in liposomes (16–18).

Clodronate induces apoptosis in monocyte/macro-phages and osteoclasts both in vitro and in vivo (19–23).This is due to inhibition of the binding of transcriptionfactors such as nuclear factor kB to DNA (15,20), andprobably due to cytoplasmic events requiring proteinkinase C (22) and iron (24). Unlike other amino-containing bisphosphonates, clodronate neither inter-feres with the mevalonate metabolism nor activatescaspase 3–like proteases (12,19).

A large number of experimental animal studieswith clodronate liposomes have shown that macrophage

depletion can be achieved in the joints and several otherorgans, depending on the administration route (25). Wehave extensively demonstrated that selective depletionof synovial lining macrophages can be achieved afterintraarticular (IA) administration of clodronate lipo-somes in experimental arthritis. Before disease onset,this intervention prevents arthritis (26–29), and it alsoprevents arthritic flares (30). Furthermore, it amelio-rates ongoing chronic arthritis (30–32) and preventscartilage destruction (29).

Based on our preclinical data with the use ofclodronate liposomes in experimental arthritis models,we performed an open study in patients with RA. Theprimary end points of this study were, first, to assesswhether a single IA administration of clodronate lipo-somes results in local macrophage depletion in patientswith advanced RA, and second, to determine the poten-tial systemic and local toxic effects of this approach inhumans. Additional goals were to study the relationshipbetween immunohistologic parameters and clinical mea-surements of disease activity and cartilage damage.

PATIENTS AND METHODS

Patients. Eligibility criteria for the study were thatpatients had to be diagnosed as having RA according to theAmerican College of Rheumatology (formerly, the AmericanRheumatism Association) criteria (33), older than 18 years ofage, and scheduled for a knee joint prosthesis at our center.Concomitant therapy with antirheumatic drugs and/or low-dose oral steroids (,10 mg/day) was kept at a constant dosage4 weeks prior to and during the study. In this period, no IAtherapy other than the study medication was allowed. Patientswith a history of bleeding or infectious disorders or bisphos-phonate allergy were excluded from the study. Other exclusioncriteria were pregnancy/breast-feeding, severe medical condi-tions involving abnormalities in the liver or kidney or inhematologic parameters, and history of alcohol or drug abuse.

Study protocol. Consenting RA patients awaiting aselective knee arthroplasty were enrolled in the study. Thebiopsy protocol encompassed the use of a Parker-Pearsonneedle to obtain percutaneous synovial biopsy samples (.20)from the suprapatellar pouch at 2 weeks before surgery (day 0)and open biopsy samples obtained at the same site duringsurgery (day 14). After initial testing of the safety of thisprotocol, patients received a single IA administration of clo-dronate liposomes 1 week before surgery (day 7). This timeschedule was chosen because, in experimental arthritis, mac-rophage depletion is maximal 1 week after liposomalclodronate administration (28,32). The clodronate dose usedin this study (mean 6 SD 160 6 35 mg) was also extrapolatedfrom effective doses used in animal studies, and amounted to;50% of the daily intravenous dosages used for malignanthypercalcemia (34).

Knee joint examination was performed by a singleobserver (PB) on days 0, 7, and 14. Joint tenderness (0 5 none,

1952 BARRERA ET AL

1 5 response on questioning, 2 5 spontaneous response, 3 5withdrawal) and swelling (0 5 none, 1 5 thickening withoutloss of bony contours, bulge sign, 2 5 loss of bony contours,palpable but not tightly distending effusions, 3 5 bulgingsynovial proliferation, tightly distending effusions) were as-sessed. Total macroscopic scores in the knee joint werecalculated by adding up the tenderness and swelling scores.Clinical assessments of disease activity (Disease Activity Score[35], Ritchie Articular Index [36], swollen joint score, visualanalog scales [VAS] for pain and general health), and functionalassessment scores (Health Assessment Questionnaire [37])were recorded at baseline and on days 7 and 14. At the sametime points, hemoglobin concentrations, platelet counts, theerythrocyte sedimentation rate, and C-reactive protein (CRP)levels were measured. Adverse events were recorded daily, andserum calcium concentrations were measured at baseline andon days 1 and 7.

At surgery, a single orthopedic surgeon (MCW) scoredmacroscopic synovitis and cartilage damage on a VAS scale.Radiologic damage in the knee, hands, and feet was measuredwith the Larsen (38) and modified Sharp (39) scores, respec-tively.

The study was approved by the Ethics Committee ofthe University Hospital Nijmegen.

Immunohistochemical analysis of synovial biopsy tis-sue. Biopsy samples were immediately fixed in 10% formalinand embedded in paraffin. Serial 7-mm microtome sectionswere mounted on superfrost slides and stained with hematoxy-lin and eosin. For immunohistochemistry, tissue was embed-ded in Tissue Tek OCT (Miles, Elkhart, IN) and snap frozen inliquid nitrogen. Cryostat sections were mounted on superfrostslides and fixed in acetone. Immunohistochemical staining wasperformed as previously described (40–42). For each marker,all sections were stained in the same run to minimize interassayvariations.

Serial sections were stained with the following mono-clonal antibodies: anti-CD68 (for macrophages, EBM 11;Dako, Glostrup, Denmark), anti-CD14 (for monocytes, TUK4;Dako), anti-CD3 (for T cells, UCHT1; Dako), anti-CD38 (forplasma cells, HIT2; PharMingen, San Diego, CA), anti–inter-cellular adhesion molecule 1 (anti–ICAM-1) (anti-CD54,PME12; kind gift of R. de Waal, Department of Pathology,University Hospital Nijmegen), and anti–vascular cell adhe-sion molecule 1 (anti–VCAM-1) (anti-CD106, 1G11B1;Monosan, Uden, The Netherlands). This primary step wasfollowed by incubation with normal horse serum and withbiotinylated horse anti-murine IgG. Slides were stained withavidin–peroxidase (Elite kit; Vector, Burlingame, CA), devel-oped with diaminobenzidine, and counterstained with hema-toxylin for 3 minutes. Controls consisted of 1) irrelevantprimary isotype-specific IgG1 and IgG2a antibodies (Dako),and 2) omission of the secondary antibodies.

All areas of each section were randomly analyzed by 2blinded observers (AB and PLEML), who used semiquantita-tive 5-point scales (from 0 to 4) to score the immunohistologicstaining results. Differences of 1 point in the readings weretaken as the average, and differences exceeding 1 point wereresolved by mutual agreement.

Preparation of clodronate-containing liposomes.Unilamellar liposomes encapsulating clodronate (Cl2MDP,Bonefos; Schering, Weesp, the Netherlands) were elaborated

according to the “film method.” Briefly, 1,1-egg-phosphatidyl-choline (Lipoid, Ludwigshafen, Germany) and 98 mg choles-terol (Sigma, St. Louis, MO) were dissolved in 15 ml ethanol at50°C in a sterile round-bottomed flask. Vacuum rotary evap-oration was then applied at 50°C to obtain a thin lipid film onthe flask wall. Clodronate (10 ml; concentration 60 mg/ml) wasadded to the lipid film, which was then dispersed by hand-shaking and vortexing. The liposomes were sized by extrusionthrough sterile polycarbonate filters (0.2, 0.1, and 0.05 mm;Poretics, Livermore, CA) using a sterile Liposofast-100 ex-truder (Avestin, Ottawa, Ontario, Canada). The encapsulationefficiency was 10–14%.

Removal of unencapsulated clodronate was achievedby gel filtration over sterile PD-10 columns (Pharmacia Bio-tech, Uppsala, Sweden) with phosphate buffered saline aseluent. Eluted fractions were pooled, filtered (through 0.2 mmfilters, Minisart; Sartorius, Gottingen, Germany), and stored insterile vials at 4°C until administration. The average particlesize and polydispersity index, as determined by dynamic lightscattering measurements (Malvern 4700, Malvern, UK), wererequired to be 120–160 nm and ,0.3 nm, respectively. Theconcentration of encapsulated clodronate was determinedspectrophotometrically at a wavelength of 240 nm afterphenol–chloroform extraction and binding to Cu (requirement4–10 mg/ml). Pyrogenicity testing was performed using theLimulus amebocyte cell lysate assay (BioWhittaker, Walkers-ville, MD), and sterility for aerobic and anaerobic pathogenswas controlled using standard techniques.

Statistical analysis. Statistical analysis was performedusing the SAS statistical package (PC version 6.04; SASInstitute, Cary, NC). Data were expressed as the mean 6 SDor as the median (range) as appropriate. Within-group com-parisons were analyzed by Student’s paired t-test or Wilcoxon’ssigned rank test. Baseline comparisons between groups wereperformed using the t-test, Mann-Whitney U test, or chi-square test as appropriate. Correlations were expressed usingthe Spearman’s rank correlation coefficient.

RESULTS

Clinical results. From April 1997 to April 1999,16 consecutive patients were enrolled in the study. Thesafety of the biopsy protocol and the consistency ofimmunohistologic findings in percutaneous and openbiopsy samples was initially tested in a control group(n 5 6). Subsequently, 10 patients received a single IAdose of clodronate liposomes 7 days before surgery. Asshown in Table 1, the demographic and clinical charac-teristics were similar in the control and treated patients.As expected from the inclusion criteria, most patientshad a long disease duration, had taken several antirheu-matic drugs in the past, and had severe knee osteoar-thritis (median Larsen score 3).

The biopsy procedures were well tolerated anddid not interfere with elective surgery in any of thepatients. Some patients had a mild hematoma at thebiopsy site which waned within 7 days after percutaneous

SYNOVIAL MACROPHAGE DEPLETION IN RA 1953

biopsy. The biopsy procedures and clodronate liposomeadministration were considered to be painless. On ques-tioning, no patient presented any objections to under-going repeated biopsies or clodronate liposome admin-istration if needed.

No clinically relevant complications were ob-served during or after the study. Pain and swelling of thebiopsied joint and the clinical assessments of diseaseactivity, function, and subjective well-being did not show

significant changes after clodronate liposome adminis-tration. The same was true for the levels of acute-phasereactants and hemoglobin concentrations (Table 2). Theserum concentrations of calcium were not affected bythis therapy (median differences from baseline were 0and 0.01 mmoles/liter at days 1 and 7, respectively).

Immunohistologic findings. As shown in Table 3,baseline immunohistologic scores in the treated andcontrol groups were similar. Baseline CD68 staining in

Table 1. Patient characteristics at baseline*

Control(n 5 6)

Treated(n 5 10)

Female, % 50 80Age, years 66 (47–73) 65 (25–78)Rheumatoid factor positive, % 83 60Shared epitope positive, % 66 70Disease duration, years 11 (4–32) 13 (1–22)Previous DMARDs, number 2 (0–6) 3 (0–7)Disease Activity Score 3.47 (2.07–4.87) 2.79 (1.57–4.45)Ritchie Articular Index 9.5 (5–19) 5.5 (1–13)Swollen joint score 7 (1–15) 2.5 (0–16)Global pain assessment, VAS 51 (19–75) 47 (0–83)Global health assessment, VAS 42 (15–76) 47 (2–80)HAQ score 0.44 (0.30–2.26) 1.1 (0.52–2.41)Sharp damage score 52 (7–105) 100 (0–279)Knee joint pain 1 swelling score, 0–6 scale 3.5 (1–5) 2.5 (1–4)Larsen damage score for knee joint, 0–5 scale 3 (3–4) 3 (1–5)Macroscopic synovitis, VAS 61 (49–100) 41 (0–93)Macroscopic cartilage damage, VAS 68 (50–100) 74 (54–99)ESR, mm/hour 22 (5–64) 20 (6–51)C-reactive protein, mg/liter 11 (0–33) 8 (0–107)Hemoglobin, mmoles/liter 8 (6.4–9.2) 8 (6.3–8.6)Platelets, 3109/liter 246 (200–309) 269 (151–363)

* Except where indicated otherwise, values are the median (range). DMARDs 5 disease-modifyingantirheumatic drugs; VAS 5 visual analog scale, HAQ 5 Health Assessment Questionnaire; ESR 5erythrocyte sedimentation rate.

Table 2. Changes in clinical and laboratory variables on day 7 (liposome administration) and day 14 (surgery)*

Control (n 5 6) Treated (n 5 10)

At day 7 At day 14 At day 7 At day 14

Knee joint pain 1 swelling score, 0–6 scale 0 (22, 2) 0.5 (0, 2) 0 (22, 2) 0 (22, 3)Disease Activity Score ND 0.32 (20.65, 0.57) 20.14 (21.33, 0.52) 20.17 (21.3, 1.45)Ritchie Articular Index ND 1.5 (25, 10) 20.7 (25, 3) 0 (25, 7)Swollen joint score ND 1.5 (22, 11) 20.5 (25, 5) 20.5 (25, 10)Global pain assessment, VAS ND 5.5 (23, 64) 1 (251, 32) 1 (270, 15)Global health assessment, VAS ND 18 (259, 56) 21 (252, 10) 28.5 (272, 31)HAQ score ND 0.09 (20.43, 0.66) 0.03 (20.52, 0.22) 0 (20.42, 0.39)ESR, mm/hour 2 (24, 5) 4 (224, 13) 21 (210, 10) 23.5 (217, 13)C-reactive protein, mg/liter 26 (235, 3) 214.5 (242, 9) 21 (272, 35) 0 (220, 65)Hemoglobin, mmoles/liter 20.1 (20.3, 0.7) 0.6 (20.4, 1.4) 0.1 (20.4, 0.6) 0.2 (20.2, 1.5)Platelets, 3109/liter 23 (221, 6) 223 (298, 18) 218 (217, 8) 232 (2109, 30)

* Values are the median (range) change from baseline in percutaneous biopsy samples. ND 5 not done (see Table 1 for other definitions).

1954 BARRERA ET AL

the lining was, by chance, slightly higher in the treatedgroup, but this did not reach statistical significance.

Within-group comparison showed that immuno-histologic scores in the untreated control patients werehighly concordant. In this group, scores for cellularmarkers and adhesion molecules in percutaneous biopsysamples were similar to those in samples obtained atsurgery. In contrast, in the treated group, a significantdecrease in CD68-positive cells in the lining layer wasobserved after clodronate liposome administration (me-dian score decreased from 4 to 0.45 in the treated group[P 5 0.005] and remained unchanged in the controlgroup). The staining for CD14 in the lining was muchless intense than that of CD68 and was not scoredseparately. The scores for CD68- and CD14-positivecells in the sublining were correlated (r 5 0.54; P 50.002), and the staining for CD68 in the sublining layerswas slightly decreased in treated patients (median scoredecreased from 2.65 to 1.65; P 5 0.04).

Clodronate liposome administration did not alterthe scores for T lymphocytes or plasma cells (CD3- andCD38-positive cells) in synovial tissue (Table 3). Thehistologic appearance of the lining after clodronateliposome administration showed a predominance offibroblast-shaped synoviocytes and sporadic CD68-positive cells (Figure 1).

Administration of clodronate liposomes also re-sulted in significant reduction in the expression ofICAM-1 (P 5 0.013 ) and VCAM-1 (P 5 0.039) in thelining layer (Table 3 and Figure 2). The reduction inexpression of these adhesion molecules in the lining washighly correlated (r 5 0.9, P 5 0.0005) and also relatedto the changes in the lining CD68 expression (ICAM-1and VCAM-1 r 5 0.71 and r 5 0.65, respectively, P 50.03 and P 5 0.06, respectively).

Relationship between immunohistologic mark-ers, disease activity parameters, and cartilage damage.The relationship between immunohistologic markersand local/systemic parameters of disease activity andradiologic/macroscopic cartilage damage was also exam-ined. The expression of ICAM-1 and VCAM-1 in thesublining layers was strongly correlated with synovitis asscored at surgery (r 5 0.83 and r 5 0.75, respectively,P , 0.0005 and P , 0.005, respectively) and, to a lesserextent, with scores for macroscopic swelling (r 5 0.56and r 5 0.46, respectively, P 5 0.001 and P 5 0.01,respectively). The staining for CD14 and ICAM-1 in thesublining correlated with CRP measurements (r 5 0.6and r 5 0.5, respectively, P , 0.0005 and P , 0.01,respectively).

Concerning the degree of joint destruction, wefound a positive correlation between staining for CD68in the sublining layers and cartilage damage as scored atsurgery (r 5 0.59, P 5 0.02); however, no other corre-

Figure 1. Detail of the lining layer before and after clodronate-liposome administration in the same patient Note the sparsemacrophage-like synoviocytes and the normal features of thefibroblast-like synoviocytes after therapy. (Hematoxylin and eosinstained; original magnification 3 250.)

Table 3. Immunohistologic features in percutaneous and surgical biopsy specimens*

Control (n 5 6) Treated (n 5 10)

Percutaneousday 0

Surgeryday 14 P

Percutaneousday 0

Surgeryday 14 P

CD68 lining 2 (1, 2) 2 (1, 2) NS 4 (2.5, 4) 0.45 (0.2, 1.1) 0.005CD68 sublining 2 (2, 2) 2 (2, 3) NS 2.65 (2, 4) 1.65 (1, 2) 0.04CD14 sublining 1.5 (0, 2) 1.5 (1, 3) NS 1.3 (0.8, 2.5) 1 (0.5, 1.3) NSCD3 1 (1, 1) 1.5 (1, 3) NS 1.3 (0.8, 1.8) 0.95 (0.6, 1.7) NSCD38 2 (1.5, 2) 2 (1, 3) NS 0.85 (0.5, 1) 1.05 (0.4, 1.5) NSICAM-1 lining 2.5 (2, 4) 2 (1, 4) NS 2.8 (1.5, 3.5) 0.5 (0.2, 1) 0.013ICAM-1 sublining 2.5 (1, 3) 2 (2, 4) NS 1.8 (1.5, 2) 1.5 (1.1, 2) 0.068VCAM-1 lining 2 (0, 4) 3.5 (1, 4) 0.16 1.5 (1, 3) 0.7 (0.4, 1) 0.039VCAM-1 sublining 1 (0.5, 2) 1.5 (1, 2) 0.04 1.3 (0.5, 2) 0.7 (0.3, 1.1) NS

* Values are the median (range) score (scale 0–4) for immunohistologic staining. NS 5 not significant; ICAM-1 5 intercellular adhesion molecule1; VCAM-1 5 vascular cell adhesion molecule 1.

SYNOVIAL MACROPHAGE DEPLETION IN RA 1955

lations between histologic markers and macroscopicscores (surgery) or radiologic damage (Larsen andSharp scores) were observed.

DISCUSSION

This is, to our knowledge, the first study oforgan-targeted macrophage removal in humans. Ourresults show that a single IA administration of clo-dronate liposomes results in depletion of synovial liningmacrophages in patients with longstanding RA. Theobserved effect was selective for macrophage-like type Asynoviocytes. These phagocytic cells are predominant inthe synovial lining, derived from bone marrow, andexpress CD68, a differentiation marker associated withthe presence of lysosomes (7). Earlier studies revealedthat type A synoviocytes are an important source ofchemokines and selective elimination of these cells hasprophylactic and therapeutic effects in experimentalarthritis (26–32). Lining fibroblast-like type B synovio-cytes, T lymphocytes, and plasma cells in synovium werenot affected by the treatment.

As in previous studies, the synovial expression ofCD14 was low in the lining, but abundant and highlycorrelated with CD68 staining in the sublining layers (8).Sublining monocyte/macrophages staining for CD14/CD68 are probably newly recruited from the circulationand are in the process of maturation and migration tothe lining (43). Administration of clodronate liposomesyielded a slight decrease in CD68 expression in thesublining layers. This effect might be due to an increasedturnover to replenish the pool of CD68-positive cells inthe lining and, less likely, to a direct effect of theseliposomes in deeper layers.

Our findings in RA are corroborated by studies inexperimental arthritis. The preferential effect of clo-dronate liposomes on lining macrophages is explainedby the high phagocytic capacity of the macrophages. Thecolloidal size of these liposomes probably prevents deeppenetration in synovial tissue. Such a combination of cellselectivity and limited penetration makes these lipo-somes an ideal system for macrophage depletion injoints and other cavitated organs.

Clodronate liposome administration also yieldeda significant reduction in the expression of ICAM-1 andVCAM-1 in the lining. These adhesion molecules arewidely expressed in RA synovium and localize to liningand sublining macrophages, fibroblast-like synoviocytes,and endothelial cells (44,45). Binding of ICAM-1 andVCAM-1 to their respective ligands, lymphocytefunction–associated antigen 1 and very late activationantigen 4 (45,46), has been implicated in the interactionof fibroblast-like synoviocytes with lining macrophages(47), activated T cells (48), and matrix components (2).

The down-regulation of ICAM-1 and VCAM-1

Figure 2. Rheumatoid synovial tissue from the same patient beforeand after clodronate liposome administration. Immunohistochemicalanalyses involved hematoxylin and eosin staining alone (A) andstaining for CD68 (macrophages) (B), CD14 (monocytes) (C), inter-cellular adhesion molecule 1 (ICAM-1) (D), and vascular cell adhesionmolecule 1 (VCAM-1) (E). Note the marked decrease in CD68-positive cells and ICAM-1 and VCAM-1 expression in the lining.(Original magnification 3 100.)

1956 BARRERA ET AL

after clodronate liposome therapy correlated with thedecrease in CD68 expression in the lining and maytherefore reflect a drop in macrophage-like synoviocytesafter treatment with clodronate liposomes. Further-more, depletion of lining macrophages may reduce thelocal production of proinflammatory cytokines such asIL-1, TNF, and chemotactic factors (26), which arepotent inducers of ICAM-1 and VCAM-1 in a wide arrayof cells (45). Our results suggest that, either directly orindirectly, macrophage-like synoviocytes are responsiblefor most of the expression of ICAM-1 and VCAM-1 inthe lining layers. In patients with RA, decreased synovialexpression of adhesion molecules has been reportedafter administration of systemic therapies, includingmethotrexate (49), pulse steroids (50), leflunomide (51),anti-CD4 monoclonal antibodies (52), and TNFa block-ing strategies (53). Such effects are therefore not drug-specific and probably reflect the amelioration of synovi-tis. This is supported by the positive correlation observedin the present study between ICAM-1 and VCAM-1expression in the sublining and the extent of macro-scopic synovitis. Of note, these correlations were muchstronger than those observed between synovitis andmacrophage markers in this and other studies (8,41).

Prospective studies in early RA have previouslyshown that macrophage counts in synovial tissue arepredictive of radiologic damage (8,9). Cartilage damageis the result of inflammation in the foregoing years anddoes not necessarily have to correlate with synovialcharacteristics at the time of surgery. In our cross-sectional comparison, only the expression of CD68 in thesublining layers showed a moderate correlation with thecartilage destruction observed at surgery. Nevertheless,even in chronic phases of the disease, macrophages mayplay a role in ongoing cartilage damage.

This is the first feasibility and proof of principlestudy with clodronate liposomes in humans. Our resultsshow that clodronate liposome administration results inselective depletion of lining macrophages and decreasedexpression of adhesion molecules in the synovial lininglayer. The time schedule and the dose of liposomalclodronate used in the present study were extrapolatedfrom experimental arthritis models (28,32), and, forobvious reasons, neither the duration of the immunohis-tologic effects nor the optimal dose of liposomal clo-dronate were investigated.

In experimental arthritis, the depletion of liningmacrophages after clodronate liposome administrationis reversible, and complete repopulation of the macro-phages takes 2–4 weeks (28,32). Despite this temporaryeffect, a single administration of these liposomes is

effective in prophylactic settings and has therapeuticeffects in chronic ongoing arthritis (6). Perhaps moreimportantly, the depletion of the lining macrophagesdoes decrease cartilage destruction in chronic experi-mental arthritis (29). Finally, this form of therapy seemsto increase the efficiency of gene transfer in vivo (54).Therefore, this approach could not only provide a proofof principle on the role of macrophages, but also be oftherapeutic interest in patients with arthritis and in theoptimization of future therapies. Both the immunohis-tologic findings of this study and the lack of toxicity ofclodronate liposomes encourage further study of thisapproach. Future investigations will need to address theduration of the effects at the synovial level, the optimaldosages of liposomal clodronate, the therapeutic poten-tial of this approach in arthritis, and the feasibility ofrepeated administration, if needed, of clodronate lipo-somes.

ACKNOWLEDGMENTS

We are indebted to Esther van Kesteren-Hendrikx forthe clodronate measurements, and to Renee van Soesbergen, arheumatologist in the Slotervaart Hospital in Amsterdam, forher collaboration in the clinical study.

REFERENCES

1. Zvaifler NJ, Marinova L, Burger JA, Mutafchieva, Evans C,Adams G, et al. Mesenchymal stem cells in the blood [abstract].Arthritis Rheum 1999;42 Suppl 9:S155.

2. Firestein GS. Invasive fibroblast-like synoviocytes in rheumatoidarthritis: passive responders or transformed aggressors? ArthritisRheum 1996;39:1781–90.

3. Paleolog EM, Fava RA. Angiogenesis in rheumatoid arthritis:implications for future therapeutic strategies. Springer SeminImmunopathol 1998;20:73–94.

4. Van den Berg WB. Role of T cells in arthritis: lessons from animalmodels. In: Miossec P, van den Berg WB, Firestein GS, editors. Tcells in arthritis. Basel: Birkhauser Verlag AG; 1998. p. 75–92.

5. Epstein WV. Expectation bias in rheumatoid arthritis clinicaltrials: the anti-CD4 monoclonal antibody experience. ArthritisRheum 1996;39:1773–80.

6. Van Lent PL, Holthuysen AE, van Rooijen N, van de Loo FA, vande Putte LB, van den Berg WB. Phagocytic synovial lining cellsregulate acute and chronic joint inflammation after antigenicexacerbation of smouldering experimental murine arthritis.J Rheumatol 1998;25:1135–45.

7. Burmester GR, Stuhlmuller B, Keyszer G, Kinne RW. Mononu-clear phagocytes and rheumatoid synovitis: mastermind or work-horse in arthritis? Arthritis Rheum 1997;40:5–18.

8. Mulherin D, FitzGerald O, Bresnihan B. Synovial tissue macro-phage populations and articular damage in rheumatoid arthritis.Arthritis Rheum 1996;39:115–24.

9. Yanni G, Whelan A, Feighery C, Bresnihan B. Synovial tissuemacrophages and joint erosion in rheumatoid arthritis. AnnRheum Dis 1994;53:39–44.

10. Storm G, Crommelin DJA. Liposomes: quo vadis? Pharm Sci TechToday 1998;1:19–31.

SYNOVIAL MACROPHAGE DEPLETION IN RA 1957

11. Bakker J, Sanders A, van Rooijen N. Effects of liposome-encapsulated drugs on macrophages: comparative activity of thediamidine 49,6-diamidino-2-phenylindole and the phenanthri-dinium salts ethidium bromide and propidium iodide. BiochimBiophys Acta 1998;1373:93–100.

12. Rogers MJ, Frith JC, Luckman SP, Coxon FP, Benford HL,Monkkonen J, et al. Molecular mechanisms of action of bisphos-phonates. Bone 1999;24:73S–9S.

13. Osterman T, Kippo K, Lauren L, Pasanen I, Hannuniemi R,Sellman R. A comparison of clodronate and indomethacin in thetreatment of adjuvant arthritis. Inflamm Res 1997;46:79–85.

14. Giuliani N, Pedrazzoni M, Passeri G, Girasole G. Bisphosphonatesinhibit IL-6 production by human osteoblast-like cells. ScandJ Rheumatol 1998;27:38–41.

15. Makkonen N, Salminen A, Rogers MJ, Frith JC, Urtti A,Azhayeva E, et al. Contrasting effects of alendronate andclodronate on RAW 264 macrophages: the role of a bisphospho-nate metabolite. Eur J Pharm Sci 1999;8:109–18.

16. Makkonen N, Hirvonen MR, Teravainen T, Savolainen K,Monkkonen J. Different effects of three bisphosphonates on nitricoxide production by RAW 264 macrophage-like cells in vitro.J Pharmacol Exp Ther 1996;277:1097–102.

17. Monkkonen J, Pennanen N, Lapinjoki S, Urtti A. Clodronate(dichloromethylene bisphosphonate) inhibits LPS-stimulated IL-6and TNF production by RAW 264 cells. Life Sci 1994;54:PL229–34.

18. Pennanen N, Lapinjoki S, Urtti A, Monkkonen J. Effect ofliposomal and free bisphosphonates on the IL-1 beta, IL-6 andTNF alpha secretion from RAW 264 cells in vitro. Pharm Res1995;12:916–22.

19. Benford HL, Frith JC, Auriola S, Monkkonen J, Rogers MJ.Farnesol and geranylgeraniol prevent activation of caspases byaminobisphosphonates: biochemical evidence for two distinctpharmacological classes of bisphosphonate drugs. Mol Pharmacol1999;56:131–40.

20. Frith JC, Monkkonen J, Blackburn GM, Russell RG, Rogers MJ.Clodronate and liposome-encapsulated clodronate are metabo-lized to a toxic ATP analog, adenosine 59-(beta, gamma-dichloromethylene) triphosphate, by mammalian cells in vitro.J Bone Miner Res 1997;12:1358–67.

21. Van Rooijen N, Sanders A, van den Berg TK. Apoptosis ofmacrophages induced by liposome-mediated intracellular deliveryof clodronate and propamidine. J Immunol Methods 1996;193:93–9.

22. Selander KS, Monkkonen J, Karhukorpi EK, Harkonen P, Han-nuniemi R, Vaananen HK. Characteristics of clodronate-inducedapoptosis in osteoclasts and macrophages. Mol Pharmacol 1996;50:1127–38.

23. Schmidt Weber CB, Rittig M, Buchner E, Hauser I, Schmidt I,Palombo Kinne E, et al. Apoptotic cell death in activated mono-cytes following incorporation of clodronate-liposomes. J LeukocBiol 1996;60:230–44.

24. Monkkonen J, Heath TD. The effects of liposome-encapsulatedand free clodronate on the growth of macrophage-like cells invitro: the role of calcium and iron. Calcif Tissue Int 1993;53:139–46.

25. Van Rooijen N, Sanders A. Liposome mediated depletion ofmacrophages: mechanism of action, preparation of liposomes andapplications. J Immunol Methods 1994;174:83–93.

26. Van Lent PLEM, Holthuysen AEM, van den Bersselaar LAM, vanRooijen N, Joosten LAB, van de Loo FAJ, et al. Phagocytic liningcells determine local expression of inflammation in type IIcollagen-induced arthritis. Arthritis Rheum 1996;39:1545–55.

27. Van Lent PL, van den Hoek AE, van den Bersselaar LA,Spanjaards MF, van Rooijen N, Dijkstra CD, et al. In vivo role ofphagocytic synovial lining cells in onset of experimental arthritis.Am J Pathol 1993;143:1226–37.

28. Van Lent PL, van den Bersselaar L, van den Hoek AE, van deEnde M, Dijkstra CD, van Rooijen N, et al. Reversible depletionof synovial lining cells after intra-articular treatment withliposome-encapsulated dichloromethylene diphosphonate. Rheu-matol Int 1993;13:21–30.

29. Van Lent PL, Holthuysen AE, van Rooijen N, van de Putte LB,van den Berg WB. Local removal of phagocytic synovial lining cellsby clodronate-liposomes decreases cartilage destruction duringcollagen type II arthritis. Ann Rheum Dis 1998;57:408–13.

30. Van Lent PL, Holthuysen AE, van Rooijen N, van de Loo FA, vande Putte LB, van den Berg WB. Phagocytic synovial lining cellsregulate acute and chronic joint inflammation after antigenicexacerbation of smouldering experimental murine arthritis.J Rheumatol 1998;25:1135–45.

31. Van Lent PL, Holthuysen AE, van den Bersselaar L, van RooijenN, van de Putte LB, van den Berg WB. Role of macrophage-likesynovial lining cells in localization and expression of experimentalarthritis. Scand J Rheumatol Suppl 1995;101:83–9.

32. Van Lent PL, van den Bersselaar LA, Holthuysen AE, vanRooijen N, van de Putte LB, van den Berg WB. Phagocyticsynovial lining cells in experimentally induced chronic arthritis:down-regulation of synovitis by Cl2MDP-liposomes. RheumatolInt 1994;13:221–8.

33. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,Cooper NS, et al. The American Rheumatism Association 1987revised criteria for the classification of rheumatoid arthritis.Arthritis Rheum 1988;31:315–24.

34. List A. Malignant hypercalcemia: the choice of therapy. ArchIntern Med 1991;151:471–6.

35. Van der Heijde DM, van ’t Hof MA, van Riel PL, Theunisse LA,Lubberts EW, van Leeuwen MA, et al. Judging disease activity inclinical practice in rheumatoid arthritis: first step in the develop-ment of a disease activity score. Ann Rheum Dis 1990;49:916–20.

36. Ritchie DM, Boyle JA, McInnes JM, Jasani MK, Dalakos TG,Grieveson P, et al. Clinical studies with an articular index for theassessment of joint tenderness in patients with rheumatoid arthri-tis. QJM 1968;37:393–406.

37. Fries JF, Spitz PW, Kraines RG, Holman HR. Measurement ofpatient outcome in arthritis. Arthritis Rheum 1980;23:137–45.

38. Larsen A, Dale K, Eek M. Radiographic evaluation of rheumatoidarthritis and related conditions by standard reference films. ActaRadiol 1977;18:481–91.

39. Van der Heijde DM, van Riel PL, van Leeuwen MA, van ’t HofMA, van Rijswijk MH, van de Putte LB. Prognostic factors forradiographic damage and physical disability in early rheumatoidarthritis: a prospective follow-up study of 147 patients. Br JRheumatol 1992;31:519–25.

40. Barrera P, Joosten LA, den Broeder AA, van de Putte LB, van denBerg WB. Effect of a fully human anti-TNFa monoclonal antibodyon the local and systemic expression of TNFa and IL-1b [abstract].Ann Rheum Dis. In press.

41. Tak PP, Smeets TJM, Daha MR, Kluin PM, Meijers KAE, BrandR, et al. Analysis of the synovial cell infiltrate in early rheumatoidsynovial tissue in relation to local disease activity. Arthritis Rheum1997;40:217–25.

42. Youssef PP, Smeets TJ, Bresnihan B, Cunnane G, Fitzgerald O,Breedveld F, et al. Microscopic measurement of cellular infiltra-tion in the rheumatoid arthritis synovial membrane: a comparisonof semiquantitative and quantitative analysis. Br J Rheumatol1998;37:1003–7.

43. Edwards JC. The synovium. In: Klippel JH, Dieppe PA, editors.Rheumatology. London: Mosby; 1998. p. 5.6.1.

44. Hale LP, Martin ME, McCollum DE, Nunley JA, Springer TA,Singer KH, et al. Immunohistologic analysis of the distribution ofcell adhesion molecules within the inflammatory synovial micro-environment. Arthritis Rheum 1989;32:22–30.

1958 BARRERA ET AL

45. McMurray RW. Adhesion molecules in autoimmune disease.Semin Arthritis Rheum 1996;25:215–33.

46. Szekanecz Z, Haines GK, Lin TR, Harlow LA, Goerdt S, RayanG, et al. Differential distribution of intercellular adhesion mole-cules (ICAM-1, ICAM-2, and ICAM-3) and the MS-1 antigen innormal and diseased human synovia: their possible pathogeneticand clinical significance in rheumatoid arthritis. Arthritis Rheum1994;37:221–31.

47. Ishikawa H, Hirata S, Andoh Y, Kubo H, Nakagawa N, Nishiba-yashi Y, et al. An immunohistochemical and immunoelectronmicroscopic study of adhesion molecules in synovial pannus for-mation in rheumatoid arthritis. Rheumatol Int 1996;16:53–60.

48. Van Dinther Janssen AC, Kraal G, van Soesbergen RM, ScheperRJ, Meijer CJ. Immunohistological and functional analysis ofadhesion molecule expression in the rheumatoid synovial lininglayer: implications for synovial lining cell destruction. J Rheumatol1994;21:1998–2004.

49. Dolhain RJ, Tak PP, Dijkmans BA, De Kuiper P, Breedveld FC,Miltenburg AM. Methotrexate reduces inflammatory cell num-bers, expression of monokines and of adhesion molecules insynovial tissue of patients with rheumatoid arthritis. Br J Rheu-matol 1998;37:502–8.

50. Youssef PP, Triantafillou S, Parker A, Coleman M, Roberts-Thomson PJ, Ahern MJ, et al. Effects of pulse methylprednisoloneon cell adhesion molecules in the synovial membrane in rheuma-toid arthritis: reduced E-selectin and intercellular adhesion mole-cule 1 expression. Arthritis Rheum 1996;39:1970–9.

51. Kraan MC, Reece RJ, Barg EC, Breedveld FC, Emery P, FarnellJ, et al. Expression of ICAM-1 and MMP-1 in rheumatoid synovialtissue after treatment with leflunomide or methotrexate [abstract].Arthritis Rheum 1999;42 Suppl 9:S162.

52. Tak PP, van der Lubbe PA, Cauli A, Daha MR, Smeets TJM,Kluin PM, et al. Reduction of synovial inflammation after anti-CD4 monoclonal antibody treatment in early rheumatoid arthritis.Arthritis Rheum 1995;38:1457–65.

53. Tak PP, Taylor PC, Breedveld FC, Smeets TJM, Daha MR, KluinPM, et al. Decrease in cellularity and expression of adhesionmolecules by anti–tumor necrosis factor a monoclonal antibodytreatment in patients with rheumatoid arthritis. Arthritis Rheum1996;39:1077–81.

54. Kuzmin AI, Finegold MJ, Eisensmith RC. Macrophage depletionincreases the safety, efficacy and persistence of adenovirus-mediated gene transfer in vivo. Gene Ther 1997;4:309–16.

SYNOVIAL MACROPHAGE DEPLETION IN RA 1959