Enhanced Production of Monocyte
Chemoattractant Protein-1 in Rheumatoid Arthritis
Alisa E. Koch, * " Steven L. Kunkel,11' Lisa A. Harlow,"t Bruce Johnson, **Holly L. Evanoff,11' G. Kenneth Haines,11t
Marie D. Burdick,*l Richard M. Pope,**" and Robert M. Strieter*l
Departments of*Medicine and 11 Pathology, tNorthwestern University Medical School, and §Veteran's Administration Lakeside Medical
Center, Chicago, Illinois 60611; and'The University ofMichigan Medical Center, Ann Arbor, Michigan 48109
Cells withinthe synovial tissuemay recruitmononuclearphago-
cytes into the synovial fluid and tissues ofarthritic patients. We
investigated the production of the chemotactic cytokine mono-
cyte chemoattractant protein-i (MCP-1) using sera, synovial
fluid, synovial tissue, as well as macrophages and fibroblasts
isolated from synovial tissues from 80 arthritic patients. MCP-
1 levels were significantly higher (P < 0.05) in synovial fluid
from RA patients (mean 25.5±8.1 ng/ml ISEI) compared to
synovial fluid from osteoarthritis (OA) patients (0.92±0.08),
or from patients with other arthritides (2.9±1.5). MCP-1 lev-
els in RA sera (8.44±2.33) were significantly greater than
MCP-1 innormal sera(0.16±0.06).Thequantities ofRA syno-
vial fluid IL-8, which is chemotactic for neutrophils and lym-
phocytes, and MCP-1 were strongly positively correlated (P
< 0.05). To examine the cellular source of MCP-1, RA syno-
vial tissue macrophages and fibroblasts were isolated. Synovial
tissue fibroblasts did not express MCP-1 mRNA, but could be
induced to produce MCP-1 by stimulation with either IL-1#,
tumor necrosis factor-alpha (TNF-a), or LPS. In contrast, un-
like normal peripheral blood monocytes or alveolar macro-
phages, RA synovial tissue macrophages constitutively ex-
pressed MCP-1 mRNA and antigen. Immunohistochemical
analysis of synovial tissue showed that a significantly greater
percentage ofRA macrophages (50±8%) as compared to either
OA macrophages (5±2) or normal macrophages (1±0.3)
reacted with anti-MCP-1 antibodies. In addition, the synovial
lining layer reacted with MCP-1 in both RA and OA synovial
tissues. In contrast, only a minority of synovial fibroblasts
(18±8%) from RA synovium were positive for immunolocaliza-
tion ofMCP-1. These results suggest that synovial production
of MCP-1 may play an important role in the recruitment of
mononuclear phagocytes during inflammation associated with
RA and that synovial tissue macrophages are the dominant
source of this cytokine. (J. Clin. Invest. 1992. 90:772-779.)
Key words: mononuclear phagocytes * chemotaxis * cytokine-
Synovial tissue macrophages are important in mediating RA
joint destruction, mainly due to their ability to process antigen
Addresscorrespondence toAlisa E. Koch, M. D., Northwestern Univer-
sity Medical School, Department ofMedicine, 303 East Chicago Ave-
nue, Ward Building 3-315, Chicago, IL 6061 1.
Receivedforpublication 17 January 1992.
The Journal ofClinical Investigation, Inc.
Volume 90, September 1992, 772-779
and release a variety ofcytokines, including IL-1, -6, -8, trans-
forming growth factor-beta, tumor necrosis factor-a (TNF-
a),' and the colony-stimulating factors, macrophage colony-
stimulating factor-I andgranulocyte-macrophage colony-stim-
ulating factor (GM-CSF) (1-15). In addition, macrophages
mediate the fibroproliferative phase ofRA by producing angio-
genic activity (16, 17). Furthermore, synovial macrophages
may influence synovial fibroblasts to liberate the cytokines IL-
6, -8, and GM-CSF, via TNF-a and IL-1, thus creating an
operative cytokine network in the joint.
The mechanism by which synovial monocytes are recruited
into synovial tissues and fluids has not been fully elucidated.
Synovial tissue blood vessel endothelial cells express adhesion
molecules like vascular cell adhesion molecule-I, which can
mediate the binding ofmonocytes to blood vessels ( 18, 19). In
addition, it is likely that chemoattractants released by cells in
the synovial tissue and fluid recruit mononuclear phagocytes
into the joint.
Recently a chemotactic cytokine, termed monocyte che-
moattractant protein-I (MCP- 1) has been identified (20-23).
This chemotaxin is expressed by a variety ofcell types, includ-
ing leukocytes, smooth muscle cells, endothelial cells, fibro-
blasts, epithelial cells, and tumor cell lines (20, 24-32). More-
over, this cytokine appears to have selective chemotactic activ-
ity for mononuclear phagocytes. The role of this cytokine in
the inflamed RA joint has, as yet, not been defined.
In this study, we demonstrated significantly greater MCP- 1
levels from synovial fluids of patients with RA as compared
with osteoarthritis (OA) or other inflammatory and nonin-
flammatory arthritides. Patients with RA had increased levels
of MCP-1 in their serum as compared to normal volunteers.
Furthermore, levels ofMCP- 1 and another chemotaxin, IL-8,
were strongly and positively correlated from synovial fluids
and sera ofRA patients. Isolated RA synovial tissue fibroblasts
expressed MCP-1 mRNA and protein in response to either
LPS, IL-1fI, or TNF-a stimulation. Isolated RA tissue macro-
phages constitutively expressed both MCP- 1 mRNA and anti-
genic MCP- 1. Finally, we identified MCP-1 by immunohisto-
logy in a significantly greater percentage ofRA than OA syno-
vial tissue macrophages.
Reagent preparation. Human recombinant IL- I,
U/ng was a gift from The Upjohn Co. (Kalamazoo, MI). Human
recombinant TNF-a with a sp act of 22 U/ng was a gift from Genen-
tech (San Francisco, CA). Human recombinant MCP- i and IL-8 were
with a sp act of 30
1. Abbreviations used in thispaper:GM-CSF, granulocyte macrophage
colony-stimulating factor; MCP-1, monocyte chemoattractant pro-
tein- 1; OA, osteoarthritis; TNF-a, tumor necrosis factor-a.
Koch et al.
purchased from Peprotech, Inc. (Rocky Hill, NJ). Lipopolysaccharide
(Escherichia coli 0111; B4) was obtained from Sigma Chemical Co.
(St. Louis, MO). Polyclonal antihuman IL-8 and MCP-1 were pro-
duced by immunization of rabbits with recombinant IL-8 or MCP-l
with CFA. MAb Leu-M5 (anti-CD 1 Ic, p150,95, CR4 receptor) detects
monocytes and macrophages (Becton Dickinson & Co., Mountain
View, CA) as does mAb HAM56 (Enzo Biochem Inc., New York,
NY). MAb FVIII detects Factor VIII-related antigen present on endo-
thelial cells (Dakopatts, Carpinteria, CA).
Patient population. Synovial fluids were isolated from patients with
either RA, OA, or other arthritides during therapeutic arthrocentesis.
Serum specimens were obtained from some ofthe RA patients as well
as from healthy volunteers. Synovial tissue was obtained from patients
undergoing total joint replacements who met the American College of
Rheumatology criteria for RA orOA (32, 33). Normal synovial tissues
were obtained from fresh autopsies. Synovial tissues from these pa-
tients were snap frozen in OCT (Miles Laboratories Inc., Elkhart, IN).
Alternatively, fresh RA tissues were used for isolation ofmacrophages
or fibroblasts (see below). All specimens were obtained with Institu-
tional Review Board approval. Patient demographic information in-
cluded synovial fluid leukocyte and differential counts.
Isolation ofhuman RA synovial tissuefibroblastsandmacrophages
and preparation of conditioned media. Fresh synovial tissues were
minced and digested in a solution ofdispase, collagenase, and DNase,
as previously described (16-18, 34). Synovial fibroblast cells were cul-
tured in RPMI plus 10% FCS containing 1 mM glutamine, 25 mM
Hepes,100 U/ml penicillin, l00 ng/ml streptomycin (Gibco Laborato-
ries, Grand Island, NY), and 1% nonessential amino acids (complete
media) in 75-mm tissue culture flasks (Costar Corp., Cambridge,
MA). Upon reaching confluence, the cells were passaged by brieftryp-
sinization as previously described (16). The cellswere used at passage 4
or older, atwhich time theywereahomogenous population offibroblas-
tic cells. The cells were plated at a concentration of8.8 X I04 cells/well
in 24-well plates (Costar Corp.) in
concentrations of IL-1/, TNF-a, or LPS in RPMI were added, and
conditioned media harvested.
For isolation of synovial tissue macrophages, the tissues were
minced and digested as described above. The resultant single-cell sus-
pensions were fractionated into density-defined subpopulations by iso-
pyknic centrifugation through continuous preformed Percoll gradients
(Pharmacia Inc., Piscataway, NJ). Macrophages were enriched by ad-
herence to fibronectin-coated collagen gels and selective trypsinization
(incubation with trypsin:EDTA for 5-10 min) (16, 17, 34, 35). Macro-
phages were harvested from the collagen gels by treatment with clostri-
dial collagenase and found to be 2 90% pure, as assessed byFcreceptor-
mediated phagocytosis ofIgG opsonized sheep red blood cells, esterase
staining, and staining with commercial antimacrophage mAbs (16,
17). For preparation ofconditioned medium, freshly isolated macro-
phages were incubated in DME + gentamicin at 1 x 106 cells/ml, and
supernatants collected after 24 h.
MCP-I ELISA. Antigenic MCP- I was measured using a modifica-
tion of a double ligand method as previously described (36). In brief,
96-well plates (Nunc, Kamstrup, Denmark) were coated with 50 i1/
well rabbit anti-MCP- 1 (3.2 Mig/ml in 0.6 M NaCl, 0.26 M H3BO4, and
0.08 N NaOH, pH 9.6) for 16 h at 4°C and then washed in PBS, pH
7.5, 0.05% Tween-20 (wash buffer). Nonspecific binding sites were
blocked with 2% BSA in PBS (200/l),and the plates incubated for 90
min at 37°C. Plates were rinsed (3x) with wash buffer and diluted
(neat, 1:5, and 1:10) test sample (50 Id) in duplicate was added, fol-
lowed by incubation for 1 h at 37°C. Plates were washed (4x) and 50
,Ml/well biotinylated rabbit anti-MCP-1 (6 vg/ml in PBS, pH 7.5,
0.05% Tween-20, 2% FCS) added for 45 min at 37°C. Plates were
washed (4x), streptavidin-peroxidase conjugate (100 ,ug/ml) (Dako
patts) added, and the plates were incubated for 30 min at 37°C. The
plates were washed (3x ) and 100I
orthophenylenediamine dichloride) (Dako patts) added. The plates
were incubated at 25°C for 6 min, and the reaction terminated with 50
dl/wellof 3 MH2SO4solution in wash bufferplus2% FCS. Plates were
I ml serum-free RPMI. Various
l chromogen substrate (0.67 mg/ml
read at 490 nm in an ELISA reader. Standards were 1/2 log dilutions of
recombinant MCP-1 from 1,000 ng/ml to 1 pg/ml (50 Ml/well). The
ELISA consistently detected MCP-1 concentrations > 0.05 ng/ml.
IL-8 ELISA. Antigenic IL-8 was measured using a modification of
a double ligand method as previously described using polyclonal anti-
IL-8 (1). The assay was standardized using human recombinant IL-8.
Northern blot analysis. Total cellular RNA was obtained from 2.5
x 106 macrophages or confluent fibroblasts in 100-mm tissue culture
dishes using a modification ofChirgwin et al. and Jonas et al. (37-39).
Briefly, cells were scraped into a solution containing 25 mM Tris, pH
8.0, 4.2 M guanidine isothiocyanate, 0.5% Sarkosyl, and 0.1 M 2-mer-
captoethanol. After homogenization, the suspension was added to a
solution containing an equal volume of 100 mM Tris, pH 8.0, 10 mM
EDTA, and 1% SDS. The mixture was then extracted with chloroform-
phenol and chloroform-isoamyl alcohol. TheRNA was alcohol precipi-
tated and the pellet dissolved in diethylpyrocarbonate-treated H20.
Total RNA was separated by Northern analysis using formaldehyde,
1% agarose gels, transblotted onto nitrocellulose, baked, prehybridized,
and hybridized with a 32P-5' end-labeled oligonucleotide probe. A 30-
mer oligonucleotide probe was synthesized using the published cDNA
sequence for human-derived MCP- 1 (20). The probe was complemen-
tary to nucleotides 256-285 and had the sequence 5'-TTG-GGT-TTG-
CTT-GTC-CAG-GTG-GTC-CAT-GGA-3'. Blotswerewashedand au-
toradiographs were quantitated using laser densitometry (Ultrascan
XS; LXB Instruments, Inc., Houston, TX). Equivalent amounts of
total RNA/gel were assessed by monitoring 28s and 18s rRNA.
Bioassay for chemotactic activity for monocytes. Chemotaxis of
monocytes was performed as previously described (24). Normal hu-
man mononuclear cells were obtained from peripheral blood by Ficoll-
Hypaque density gradient centrifugation. Monocytes were suspended
in HBSS with calcium and magnesium (Gibco Laboratories) at 3 x 106
cells/ml with > 95% viability by trypan blue exclusion. In brief, 150Al
of synovial fluid or synovial tissue macrophage conditioned medium
which was diluted 1:1 with HBSS, 10-8 M FMLP (Sigma Chemical
Co.), or HBSS alone were placed in duplicate bottom wells of a blind-
well chemotaxis chamber. A 5-,um pore size polyvinylpyrrolidone-free
polycarbonate filter (Nuclepore Corp., Pleasanton, CA) was placed in
the assembly and 250 ul ofmonocyte suspension placed in each ofthe
top wells. Chemotaxis chambers were incubated at 37°C in humidified
95% air/5% CO2 for 2 h. The filters were removed, fixed in absolute
methanol, and stained with 2% toluidine blue (Sigma Chemical Co.).
Monocytes that had migrated through to the bottom ofthe filter were
counted in 10 high power fields (x 1,000).
Immunoperoxidase staining. 4-tm sections of frozen tissues were
cut, and immunoperoxidase stained using an avidin-biotin technique
(Vector Laboratories, Burlingame, CA) (34, 40-44). Slides, air dried
for 2-16 h, were fixed in cold acetone for 20 min. Endogenous peroxi-
dase activity was quenched by incubating the slides for 30 min in 0.3%
hydrogen peroxide in methanol. All subsequent incubations were per-
formed for 15 min at 37°C in a moist chamber. The tissue sections were
pretreated with 50 M1 diluted normal horse serum ( 135 Ml horse serum
in 10 ml 1% PBS-BSA), incubated with either rabbit anti-human
MCP-1, preimmune rabbit serum, mAb Leu-M5, mAb HAM56, or
control mAb and washed (2x). The slides were incubated with a 1:400
dilution of anti-mouse biotinylated antibody in PBS-BSA, washed
(2x) with PBS, incubated with avidin/biotinylated horseradish perox-
idase complex, and washed with PBS (2x). Slides were then stained
with diaminobenzidine tetrahydrochloride substrate for 5 min at room
temperature, rinsed in tap water for 2min, counterstained with Harris'
hematoxylin, and dipped in saturated lithium carbonate solution for
bluing. Serial tissue sections were examined to determine the percent-
age ofeach cell type expressing antigenic MCP- I ( 18, 34, 43).
Cytospin preparations of isolated RA synovial tissue macrophages
were made using 105 cells per slide in a cytospin (Shandon I; Shandon
Inc., Swickley, PA). Slides were stained using immunohistochemistry
as described above.
Statistical analysis. Statiscal analysis was performed using analy-
sis of variance(45, 46). P values < 0.05 were considered significant.
Monocyte Chemoattractant Protein-] in Arthritis
Figure 1. Antigenic MCP-1 levels in synovial fluid from various
arthritides as well as RA and normal serum. Results represent the
mean±SE. Six replicate determinations per patient were performed.
N represents the number of patients studied. *; synovial fluid; m,
MCP-J is elevated in both RA synovialfluid and sera. In initial
experiments, MCP-1 was measured by ELISA using samples
obtained from 27 patients (Fig. 1). The patients with RA had
the greatest levels of MCP-1 (mean 25.5±8.1 ng/ml). OA pa-
tients as compared with RA patients had significantly less
MCP-l in their synovial fluids (0.92±0.08 ng/ml, P < 0.05).
Patients with other forms of inflammatory and noninflamma-
chronic lymphocytic leukemia, and mixed connective tissue
disease had a mean of2.9±1.5 ng/ml ofsynovial fluid MCP-1,
which was also significantly less than patients with RA (P
To determine whether peripheral blood from patients with
RA also contained significant quantities of MCP-1 compared
to normal volunteers, serum MCP- 1 levels were determined by
ELISA using 19 RA samples (Fig. 1). Serum levels of RA
MCP-1 ranged from <0.08 ng/ml to 179.2 ng/ml (mean
22.4±8.9). In contrast normal sera obtained from 10 volun-
arthritis including polymyositis,
RA Serum MCP- I (ng/ml)
Figure 2. Positive correlation between serum antigenic MCP-1 and
Figure 3. Time-dependent generation ofsynovial fibroblast MCP-1 by
LPS, and IL-1IB stimulation. RA fibroblasts (8.8 X 104 cells/well)
were cultured in serum-free RPMI for various time periods. Results
represent the mean±SE from two patient samples. Six replicates of
each sample were assayed. -* -, LPS (1 ,Ig/ml);-v -, IL-1d
(20 ng/ml);- A -, nontreated.
teers contained significantly less MCP-1 (< 0.05 to 0.72 ng/
ml) (mean 0.16±0.06, P < 0.05). The high levels of MCP-1
present in the RA sera were not attributable to rheumatoid
factor in these samples, since there was no correlation between
rheumatoid factor positivity (using rabbit IgG) and MCP-1
Positive correlation between synovial fluid and serum
MCP-I andIL-8 levels. Since cells ofthe synovial milieu might
concomitantly recruit both monocytes and other inflamma-
tory leukocytes, we wished to determine whether those patients
who produced large quantities of MCP-1 also produced large
quantities of IL-8. ELISA assays were performed on the sera
and synovial fluids of arthritic patients to determine the quan-
tities of IL-8 produced. Comparing RA synovial fluid MCP-1
and IL-8 levels, a significant positive correlation (r = 0.57, P
Figure 4. (A) Inducible
RA fibroblast MCP- I
gene expression. Repre-
sentative Northern blot
of RA synovial fibro-
blasts which were cul-
tured for 8 h in serum-
free media with or with-
IL-1I3 (20 ng/ml), or
TNF-a (20 ng/ml).
MCP- I message was ab-
sent without stimula-
tion. Ctrl, control un-
treated fibroblasts. (B)
Laser densitometry of
MCP- I mRNA. (C)
18S and 28S rRNA to
loading ofRNA in A.
Koch et al.
Figure 5. (A) A representative Northern
blot showing MCP-1 gene expression
from freshly isolated RA synovial tissue
macrophages from two patients. (B) 18S
and 28S rRNA demonstrating equivalent
loading of total RNA in A. (C) Immuno-
peroxidase stained cytocentrifuge prepa-
ration of freshly isolated RA synovial
macrophages showing MCP- 1 antigen ex-
pression (arrows) (X8 16).
< 0.05) was found. Similarly, synovial fluid from other arthri-
tides contained MCP- 1 and IL-8 levels which were also
strongly positively correlated (r = 0.96, P < 0.05). RA serum
MCP- 1 and IL-8 levels were correlated, implying a parallel re-
lationship between the amounts of chemotactic cytokines re-
leased specific for differing leukocyte populations in this dis-
order (Fig. 2, r = 0.67, P < 0.05).
RAsjynovial/ibroblast production ofMCP-J
whether cellular constituents ofthe synovial tissue were able to
produce MCP-1 in vitro, isolated cell populations from RA
synovial tissues were examined. RA fibroblast cells appeared to
be a homogenous cellular population of cells with elongated
processes. No rounded cells were present. Moreover, RA fibro-
blasts were esterase negative, and nonreactive with the antima-
crophage mAbs Leu-M5 and HAM 56, excluding the presence
ofcontaminating macrophages. Cytocentrifuge preparations of
fibroblasts did not react immunohistochemically with mAb
FVIII, excluding the presence of contaminating endothelial
cells. LPS induced fibroblast MCP- 1 release at concentrations
of 100 ng/ml and above. In contrast, as little as 0.02 ng/ml
IL-1: or TNF-a induced fibroblast MCP-1 release. Maximal
stimulatory concentrations of IL-1f and TNF-a were 0.2 ng/
ml and 20 ng/ml, respectively. Time-dependent generation of
MCP- 1 is shown in Fig. 3. LPS-induced fibroblast MCP- 1 in-
creased steadily over 24 h as did IL- 1-/ induced MCP- 1. While
LPS induced 1.62±0.8 ng/ml (SE) MCP-1, IL-l13 induced
5.7±0.78 ng/ml MCP-1. Nonstimulated fibroblasts released
0.29±0.78 ng/ml MCP-1. Fibroblast MCP-1 gene expression
Monocyte Chemoattractant Protein-I in Arthritis
@. $i, .;
Table I. Chemotaxis ofMonocytes in Response to RA Synovial
Fluids and RA Synovial Tissue Macrophage Conditioned
Medium Compared to MCP-1 Protein Levels Determined by
FMLP (10-8 M)
RA synovial fluids and synovial tissue macrophage-conditioned me-
dia were assayed for their ability to induce chemotaxis of monocytes.
The results represent analysis of 10 high power fields per sample.
Negative control migration in response to HBSS = mean of five
cells/high power field. MCP- 1 was assayed by ELISA (see Methods).
was absent in nonstimulated RA fibroblasts, but readily induc-
ible upon LPS (1 l
g/ml), IL-tlo(20 ng/ml), or TNF-a (20
ng/ml) treatment of cells (Fig. 4).
RA synovial fluid and macrophage generated monocyte
chemotactic activity. We then determined whether RA synovial
fluids and conditioned medium from synovial tissue macro-
phages produced biologically active chemotactic activity for
monocytes. Chemotactic activity for normal human peripheral
blood monocytes was detected in the RA synovial fluids exam-
ined (Table I). Conditioned medium from nonstimulated mac-
rophages obtained from six patients also contained chemotac-
tic activity for monocytes. Measurable antigenic MCP- 1
ranged from < 0.05 to 7.4 ng/ml in these samples. The
amounts of chemotactic activity for monocytes and antigenic
MCP-1 did not correlate, implying the presence of additional
chemotactic factors for monocytes in the synovial fluids and in
the RA synovial tissue macrophage supernatants.
Synovial tissue macrophage production ofMCP-1. To de-
termine if RA synovial tissue macrophages constitutively ex-
pressed MCP- 1, MCP-1 gene expression was determined using
isolated RA synovial tissue macrophages (Fig. 5). In contrast
to RA fibroblasts, RA macrophages did not require exogenous
stimulation to express MCP-1 mRNA. To confirm macro-
phage constitutive MCP-1 production, macrophages were
freshly isolated from the synovial tissues ofRA patients, cyto-
spun, and stained for immunolocalization of MCP- 1. Fig. 5 c
shows RA synovial tissue macrophage MCP-1 antigen expres-
Immunohistochemical localization ofMCP-J to RA syno-
vial tissue macrophages. To immunolocalize the cells responsi-
ble for MCP-1 production in vivo, we examined synovial tis-
Figure 6. Immunohistochemical demonstration of MCP- 1 expression
in synovial tissues. Mphage, macrophage; smooth m, vascular smooth
muscle; fib, fibroblast; lymph, lymphocyte; and endo, endothelial
cell. _, RA (n= 13);
smooth m lymph
, OA (n= 11);
, NL (n=3).
sues by immunohistochemistry. Synovial tissues from 27 sub-
jects were examined for antigenic MCP- 1. The results of
immunolocalization are shown in Figs. 6 and 7. In general, the
RA synovial tissues had higher inflammatory scores (mean
2.4±0.2) as compared to either the OA tissues (mean 1.5±0.2)
or the normal tissues (mean 1.0) ( 18, 44). The specificity of
anti-MCP- 1 reactivity with tissues was confirmed by the use of
nonimmune serum as well as preadsorbtion with recombinant
human MCP- 1. The MCP-1 positive cells were located in the
macrophage-rich synovial lining layer, in both RA and OA
synovial tissues. MCP-1 immunolocalized to a mean of76±6%
ofthe RA lining cells (Fig. 7 A) and 60±10% ofthe OA lining
cells (Fig. 7 B), but only to 1±0.3% ofthe lining cells found in
normal synovial tissues (Fig. 7 D) (P < 0.05 for either RA or
OA compared to normal). RA synovial tissue macrophages
located in the subsynovial areas also expressed MCP- 1, with
50±8% of macrophages being MCP-1 positive (Fig. 7 C). In
contrast, only 5±2% ofthe OA Leu-M5 positive macrophages
were MCP-1positive (P < 0.05) (Fig. 7). Similarly, RA macro-
phage antigenic MCP-1 expression was significantly greater
than the expression found in normal synovial tissues (P
As compared to the reactivity ofthe majority ofRA macro-
phages with anti-MCP- 1, MCP-1 expression was found in a
minority of fibroblasts from both RA (mean of 18±8%) and
OA (mean of 9±3%). In addition, a small percentage (mean
9.4±6.3%) of synovial tissue blood vessel smooth muscle cells
were MCP-1 positive.
MCP-1is a 76-amino acid basic protein with selective chemo-
tactic activity for mononuclear phagocytes (23, 47). MCP-1
exists in two forms: MCP-1a, with a mol wt of 13,000, and
MCP-13, with a mol wt of 15,000 (20, 47). These two forms of
MCP-1are functionally identical, differing only by the extent
of posttranslational modification. MCP-1 belongs to a su-
pergene family that includes LD78, ACT-2, RANTES, and
The role of MCP-1 in the inflamed RA joint may be the
recruitment ofmononuclear phagocytes. There are a variety of
chemotactic factors which may attract mononuclear phago-
cytes in the RA joint. These include thrombin, C5a, platelet
Koch et al.
i' * )
Monocyte Chemoattractant Protein-i in Arthritis
activating factor, leukotriene B4, lymphocyte-derived chemo-
tactic factor, and the cytokines transforming growth factor-
beta, GM-CSF, and platelet-derived growth factor (PDGF)
(49, 50). In this study, we demonstrated that the cytokine
MCP- 1 may also serve as a potent stimulus for recruitment of
mononuclear phagocytes in RA.
We examined synovial fluids from patients with various
arthritides forMCP- 1. MCP- 1 levels were significantly elevated
in synovial fluids from patients with RA as compared toOA or
other arthritides. Moreover, while the serum levels ofMCP- 1
were 0.16±0.06 ng/ml from normal subjects, MCP-1 RA
serum levels were significantly increased (22.4±8.9 ng/ml, P
The process of synovial inflammation likely results from
the influx ofa series ofinflammatory cells that release a cascade
of inflammatory mediators. We reasoned that synovial fluids
and sera containing chemotactic factors for mononuclear
phagocytes such as MCP- 1 might also contain chemotactic fac-
tors for neutrophils and lymphocytes. Indeed, we found a
strong positive correlation between MCP-1 and IL-8 in both
synovial fluid and serum of patients with RA. These results
suggested thatthe same factorswhich might upregulate produc-
tion of MCP-1 also influence the production of IL-8 (1).
Hence, we examined whether RA fibroblasts produced MCP- 1
in response to the same agonists which upregulated production
of synovial fibroblast IL-8. As described by DeMarco et al.,
who showed that RA synovial fibroblasts expressed MCP-1
mRNA in response to IL- 1 and TNF-a, we demonstrated that
RA fibroblasts respond to these signals as well as to LPS and
produce both MCP- 1 mRNA and protein (5 1 ). While RA fi-
broblasts spontaneously produced either nondetectable or low
levels of MCP- I, TNF-a, or IL-13f, and to a lesser extent LPS
induced increased levels offibroblast MCP- 1. These same stim-
uli serve as inducing agents for RA fibroblast IL-8 production.
It is likely that in the inflamed RA synovium, macrophages
produce IL- 1 and TNF-a, both ofwhich act on synovial fibro-
blasts to stimulate release of IL-8 and MCP-1, which in turn
selectively recruit neutrophils, lymphocytes, and mononuclear
phagocytes into the joint.
While MCP- 1 may be produced by a variety ofcells such as
blood mononuclear cells, endothelial cells, smooth muscle
cells, fibroblasts, epithelial cells, and tumor cell lines, we have
identified the synovial tissue macrophage as a major cell pro-
ducing MCP- 1 within the RAjoint. Macrophages isolated from
RA synovial tissues produced MCP- 1 mRNA, antigenic MCP-
1, and biologically active chemotactic activity for peripheral
blood monocytes. In addition, we immunolocalized MCP- 1 to
isolated RA synovial tissue macrophages. RA macrophages
constitutively produce MCP- 1 in greater numbers as compared
to either OA or normal synovial tissue macrophages. Interest-
ingly, normal human peripheral blood monocytes or alveolar
macrophages do not express MCP-1 mRNA either constitu-
tively or when stimulated with LPS (52). Perhaps, in the in-
flamed synovial milieu thesynovial macrophage assumes aphe-
notype of chronic activation with subsequent inflammatory
cytokine release. Indeed, MCP- 1 has been shown to "activate"
monocytes, causing cytostatic activity against tumor cell lines,
stimulation of the leukocyte respiratory burst, and lysosomal
enzyme release (23, 47). It may be thatMCP- 1 itselfmay func-
tion to activate newly recruited mononuclear phagocytes,
hence perpetuating the inflammatory response in the RA syno-
In summary, synovial fluids from RA patients containsig-
nificantly more MCP-1 than fluids from OA patients or pa-
tients with other arthritides. RA serum MCP- 1 levels were sig-
nificantly elevated compared to normal serum MCP levels.
There was a strong positive correlation between RA synovial
fluid or serum levels ofMCP-1 and IL-8. Immunohistological
examination revealed that a significantly greater proportion of
RA than OA or normal macrophages expressed antigenic
MCP- 1. RA synovial fibroblasts were induced to produce
MCP-1 mRNA and protein by IL-1,B, TNF-ca, or to a lesser
degree LPS. RA macrophages constitutively produced MCP- 1
mRNA and protein. These resultsmay helpelucidate the mech-
anism by which mononuclear phagocytes enter both the syno-
vium and synovial fluid in RA.
We thank our colleagues, Dr. S. D. Stulberg, C. Schwartz, J. Galante,
N. Rana, B. Briggs, and J. Lessard for supplying the synovial tissues.
We would also like to thank Drs. Herbert Rubinstein and Frank
Schmid forhelpfuldiscussions. Wealsowishtoacknowledgetheinvalu-
able help ofDr. James Sinacore in performing statistical analysis ofthe
data. We thank Ms. Corinne Washington for expert secretarial assis-
This work was supported by National Institutes of Health grants
AR30692 (A. E. Koch and R. M. Pope), AR41492 (A. E. Koch), HL-
02401 (R. M. Strieter), HL-31693 (S. L. Kunkel), an Arthritis Foun-
dation grant (R. M. Pope), an Arthritis Foundation fellowship (A. E.
Koch), Veteran's Administration Merit Reviews (A. E. Koch and R.
M. Pope), and an American Lung Association research grant. R. M.
Strieter is an RJR-Nabisco Research Scholar.
1. Koch, A. E., S. L. Kunkel, J. C. Burrows, H. L. Evanoff, G. K. Haines,
R. M. Pope, and R. M. Strieter. 1991. Synovial tissue macrophage as a source of
the chemotactic cytokine IL-8. J. Immunol. 147:2187-2195.
2. Wood, D. D., E. J. Ihrie, and D. Hamerman. 1985. Release ofinterleukin- I
from human synovial tissue in vitro. Arthritis Rheum. 28:853-862.
3. Miyasaka, N., K. Sato, M. Goto, M. Sasano, M. Natsuyama, K. Inoue, and
K. Nishioka. 1988. Augmented interleukin-l production and HLA-DR expres-
sion in the synovium of rheumatoid arthritis patients. Possible involvement in
joint destruction. Arthritis Rheum. 31:480-486.
4. Wood, D. D., E. J. Ihrie, C. A. Dinarello, and P. L. Cohen. 1983. Isolation
of an interleukin-l-like factor from human joint effusions. Arthritis Rheum.
5. Yocum, D. E., L. Esperanza, S. Dubry, J. B. Benjamin, R. Volz, and P.
Scuderi. 1989. Characteristics oftumor necrosis factor production in rheumatoid
arthritis. Cell. Immunol. 122:131-145.
6. Husby, G., and R. C. Williams, Jr. 1988. Synovial localization of tumor
necrosis factor in patients with rheumatoid arthritis. J. Autoimmun. 1:363-371.
7. Lipsky, P. E., L. S. Davis, J. J. Cush, and N. Oppenheimer-Marks. 1989.
The role ofcytokines in the pathogenesis ofrheumatoid arthritis. SpringerSemin.
8. Arend, W. P., and J. M. Dayer. 1990. Cytokines and cytokine inhibitors or
antagonists in rheumatoid arthritis. Arthritis Rheum. 33:305-315.
9. Dayer, J.-M., B. deRochemonteix, B. Burrus, S. Demczuk, and C. A. Din-
arello. 1986. Human recombinant interleukin 1 stimulates collagenase and pros-
taglandin E2 production by human synovial cells. J. Clin. Invest. 77:645-648.
10. Dayer, J.-M., B. Beutler, and A. Cerami. 1985. Cachectin/tumor necrosis
factor exerts endocrine, paracrine, and autocrine control of inflammatory re-
sponses. J. Exp. Med. 162:2163-2168.
11. Firestein, G. S., J. M. Alvaro-Gracia, and R. Maki. 1990. Quantitative
analysis of cytokine gene expression in rheumatoid arthritis.
12. Firestein, G. S. 1991. The immunopathogenesis ofrheumatoid arthritis.
Curr. Opin. Rheumatol. 3:398-406.
13. Firestein, G. S., and N. J. Zvaifler. 1990. How important are T cells in
chronic rheumatoid synovitis? Arthritis Rheum. 33:768-773.
14. Alvaro-Gracia, J. M., N. J. Zvaifler, C. B. Brown, K. Kaushansky, and
G. 5. Firestein. 1991. Analysis of the synovial cells involved in granulocyte-
macrophage colony stimulating factor production and gene expression in rheu-
Koch et al.
matoid arthritis and its regulation by IL-I and tumor necrosis factor-a. J. Im- Download full-text
15. Alvaro-Gracia, J. M., N. J. Zvaifler, and G. S. Firestein. 1989. Cytokines
in chronic inflammatory arthritis. J. Exp. Med. 170:865-875.
16. Koch, A. E., P. J. Polverini, and S. J. Leibovich. 1986. Stimulation of
neovascularization by human rheumatoid synovial tissue macrophages. Arthritis
17. Koch, A. E., S. J. Leibovich, and P. J. Polverini. 1988. Functional heteroge-
neity of human rheumatoid synovial tissue macrophages. J. Rheum. 15:1058-
18. Koch, A. E., J. C. Burrows, G. K. Haines, T. M. Carlos, J. Harlan, and S. J.
Leibovich. 1991. Immunolocalizationofleukocyteandendothelialadhesion mol-
ecules in human rheumatoid and osteoarthritic synovial tissues. Lab. Invest.
19. Carlos, T., and J. M. Harlan. 1990. Membrane proteins involved phago-
cyte adherence to endothelium. Immunol. Rev. 114:1-28.
20. Yoshimura, T., E. A. Robinson, S. Tanaka, E. Appella, and E. J. Leonard.
1989. Purification and amino acid analysis oftwo human monocyte chemotrac-
tants produced by phytohemaglutinin-stimulated human blood mononuclear
leukocytes. J. Immunol. 142:1956-1962.
21. Furutani, Y., H. N. Nomura, M. Notake, Y. Oyamada, T. Fukui, M.
Yamada, C. G. Larsen, J. J. Oppenheim, and K. Matsushima. 1989. Cloning and
sequencing of the cDNA for human monocyte chemoattractant and activating
factor (MCAF). Biochem. Biophys. Res. Commun. 159:249-255.
22. Yoshimura, T., E. A. Robinson, S. Tanaka, E. Appella, and E. J. Leonard.
1989. Purification and amino acid analysis oftwo human glioma-derived cyto-
kines. J. Exp. Med. 169:1449-1459.
23. Matsushima, K., C. G. Larsen, G. C. DuBois, and J. J. Openheim. 1989.
Purification and characterization ofa novel monocyte chemotactic and activat-
ing factor produced by a human myelomonocyte cell line. J. Exp. Med.
24. Streiter, R. M., R. Wiggins, S. H. Phan, B. L. Wharram, H. J. Showell,
D. G. Remick, S. W. Chensue, and S. L. Kunkel. 1989. Monocyte chemotactic
protein gene expression by cytokine-treated human fibroblasts and endothelial
cells. Biochem. Biophys. Res. Commun. 162:694-700.
25. Sica, A., J. M. Wang, F. Colotta, E. Dejana, A. Mantovani, J. J. Oppen-
heim, C. G. Larsen, C. 0. Zachariae, and K. Matushima. 1990. Monocyte che-
motactic and activating factor gene expression induced in endothelial cells by
IL-1 and tumor necrosis factor. J. Immunol. 144:3034-3038.
26. Larsen, G. C., C. 0. Zachariae, J. J. Oppenheim, and K. Matsushima.
1989. Production of monocyte chemotactic and activating factor (MCAF) by
human dermal fibroblasts in response to interleukin- 1 or tumor necrosis factor.
Biochem. Biophys. Res. Commun. 160:1403-1408.
27. Yoshimura, T., and E. J. Leonard. 1990. Secretion by human fibroblasts
of monocyte chemoattractant protein-I, the product of gene JE. J. Immunol.
28. Graves, D. T., Y. L. Jiang, M. J. Williamson, and A. J. Valente. 1989.
Identification of monocyte chemotactic activity produced by malignant cells.
Science (Wash. DC). 245:1490-1493.
29. Elner, S. G., R. M. Streiter, V. M. Elner, B. J. Rollins, M. A. Del Monte,
and S. L. Kunkel. 1991. Monocyte chemotactic protein gene expression by cyto-
kine treated human retinal pigment epithelial cells. Lab. Invest. 64:819-825.
30. Rollins, B. J., P. Stier, T. Ernst, and G. G. Wong. 1989. The human
homolog of the JE gene encodes a monocyte secretory protein. Mol. Cell. Biol.
31. Rollins, B. J., T. Yoshimura, E. J. Leonard, and J. S. Pober. 1990. Cyto-
kine-activated human endothelial cells synthesize and secrete a monocyte che-
moattractant, MCP-1 /JE. Am. J. Pathol. 136:1229-1233.
32. Arnett, F. C., S. M. Edworthy, D. A. Bloch, D. J. McShane, J. F. Fries,
N. S. Cooper, L. A. Healey, S. R. Kaplan, M. H. Liang, H. S. Luthra, et al. 1988.
The American Rheumatism Association 1987 revised criteria for the classifica-
tion ofrheumatoid arthritis. Arthritis Rheum. 31:315-324.
33. Altman, R., E. Asch, D. Bloch, G. Bole, D. Borenstein, K. Brandt, W.
Christy, T. D. Cooke, R. Greenwald, and M. Hochberg. 1986. Development of
criteria for the classification and reporting of osteoarthritis: classification of
osteoarthritis ofthe knee. Arthritis Rheum. 29:1039-1049.
34. Koch, A. E., J. C. Burrows, A. Skoutelis, R. Marder, P. Domer, and S. J.
Leibovich. 1991. Monoclonal antibodies detecting monocyte/macrophage acti-
vation and differentiation antigens and identifying functionally distinct subpopu-
lations of human rheumatoid synovial tissue macrophages. Am. J. Pathol.
35. Koch, A. E., P. J. Polverini, and S. J. Leibovich. 1986. Induction of
neovascularization by activated human monocytes. J. Leukocyte Biol. 39:233-
36. Evanoff, H. L., M. D. Burdick, S. A. Moore, S. L. Kunkel, and R. M.
Streiter. 1992. A sensitive ELISA forthe detection ofhuman monocyte chemoat-
tractant protein-I (MCP-I). Immunol. Invest. 21:39-45.
37. Standiford, T. J., S. L. Kunkel, M. A. Basha, S. W. Chensue, J. P. Lynch
III, G. P. Toews, J. Westwick, and R. M. Strieter. 1990. Interleukin-8 gene expres-
sion by a pulmonary epithelial cell line: a model for cytokine networks in the
lung. J. Clin. Invest. 86:1945-1953.
38. Chirgwin, J. M., A. E. Przybyca, R. J. MacDonald, andW. J. Rutter. 1979.
Isolation ofbiologically active ribonucleic acid from sources enriched in ribonu-
clease. Biochemistry. 18:5294-5299.
39. Jonas, E., T. D. Sargent, and I. B. Davis. 1985. Epidermal keratin gene
expressed in embryos of xenopus laevis. Proc. Natl. Acad. Sci. USA. 82:5413-
40. Koch, A. E., G. K. Haines, R. Rizzo, J. A. Radosevich, R. M. Pope, P. G.
Robinson, and W. H. Pearce. 1990. Human abdominal aortic aneurysms: im-
munophenotypic analysis suggesting an immune-mediated response. Am. J.
41. Robinson, P. G., E. W. Eggena, A. E. Koch, and J. A. Radosevich. 1989. A
method for immunohistochemical staining with an iron counterstain. Lab. Med.
42. Hsu, S. M., L. Raine, and H. Fanger. 1981. Use ofavidin-biotin-peroxi-
dase complex (ABC) in immunoperoxidase techniques: a comparison between
ABC with unlabeled antibody (PAP) procedures. J. Histochem. Cytochem.
43. Koch, A. E., J. A. Radosevich, P. G. Robinson, and R. M. Pope. 1990.
Distribution ofCD45RA+ and CD45RO+ T lymphocyte subsets in rheumatoid
arthritis synovial tissue. J. Clin. Immunol. 10:192-199.
44. Koch, A. E., J. C. Burrows, R. M. Marder, P. Domer, and S. J. Leibovich.
1990. Reactivity ofhuman tissues with monoclonal antibodies to myeloid activa-
tion and differentiation antigens. Pathobiology. 58:241-248.
45. Winer, B. J. 1971. Statistical principles in experimental design. McGraw-
Hill Inc., New York.
46. Carmer, S. G., and M. R. Swanson. 1973. An evaluation often pairwise
multiple comparison procedures by Monte Carlo methods. J. Am. Stat. Assoc.
47. Matsushima, K., and J. J. Oppenheim. 1989. Interleukin-8 and MCAF:
novel inflammatory cytokines inducible by IL-1 and TNF. Cytokines. 1:2-13.
48. Schall, T. J. 1991. Biology of the RANTES/SIS cytokine family. Cyto-
49. Verghese, M. W., and R. Snyderman. 1989. Chemotaxis and chemotactic
factors. In Human Monocytes. M. Zembala and G. L. Asherton, editors. Aca-
demic Press Inc., San Diego, CA. 167-175.
50. Harris, E. D. 1988. Pathogenesis ofrheumatoid arthritis: a disorder asso-
ciated with dysfunctional immunoregulation. In Inflammation: Basic Principles
and Clinical Correlates. J. I. Gallin, I. M. Goldstein, and R. Snyderman, editors.
Raven Press, Ltd., New York. 751-773.
51. DeMarco, D., S. L. Kunkel, R. M. Streiter, M. Basha, and R. B. Zurier.
1991. Interleukin-1 induced gene expression ofneutrophil activating protein (in-
terleukin-8) and monocyte chemotactic peptide in human synovial cells. Bio-
chem. Biophys. Res. Commun. 174:411-416.
52. Streiter, R. M., S. W. Chensue, M. A. Basha, T. J. Standiford, J. P. Lynch
III, M. Baggiolini, and S. L. Kunkel. 1990. Human alveolar macrophage gene
expression of interleukin-8 by tumor necrosis factor-a, lipopolysaccharide, and
interleukin-lfI. Am. J. Respir. Cell Mol. Biol. 2:321-326.
Monocyte Chemoattractant Protein-i in Arthritis