Osalle borrelioosiin sairastuneista, erityisesti geneettisesti alttiille (HLA-DR4) henkilöille kehittyy krooninen niveltulehdus joka saattaa aiheuttaa esim. rustojen ja luiden eroosiota.
Live Borrelia burgdorferi Preferentially Activate Interleukin-1f. Gene Expression and Protein Synthesis over the lnterleukin-1 Receptor Antagonist
Laurie C. Miller, Sana Isa, Edouard Vannier, Kostis Georgilis, Allen C. Steere, and Charles A. Dinarello
Departments ofPediatrics and Medicine, New England Medical Center and
Tufts University School ofMedicine, Boston, Massachusetts 02111
Lyme arthritis is one of the few forms of chronic arthritis in which the cause is known with certainty. Because cytokines are thought to contribute to the pathogenesis of chronic arthritis, we investigated the effect of the Lyme disease spirochete, Borrelia burgdorferi, on the gene expression and synthesis of IL-1,8 and the IL-1 receptor antagonist (IL-lra) in human pe-ripheral blood mononuclear cells. Live B. burgdorferi induced fivefold more IL-1l@ than IL-la and sevenfold more IL-1l8 than IL-lra; LPS or sonicated B. burgdorferi induced similar amounts of all three cytokines. This preferential induction of IL-1,B was most dramatic in response to a low passage, virulent preparation of B. burgdorferi vs. three high passage avirulent strains. No difference in induction of IL-lra was seen between these strains. The marked induction of IL-1iB was partially di-minished by heat-treatment and abrogated by sonication; IL-lra was not affected. This suggested that a membrane compo-nent(s) accounted for the preferential induction ofIL-1i6. How-ever, recombinant outer surface protein ,B induced little IL-1i6.
By 4 h after stimulation, B. burgdorferi induced sixfold more
IL-1,8 protein than LPS. In contrast to LPS-induced IL-l/i mRNA which reached maximal accumulation after 3 h, B. burgdorferi-induced IL-1A6 mRNA showed biphasic elevations at 3 and 18 h. B. burgdorferi-induced IL-lra mRNA peaked at 12 h, whereas LPS-induced IL-lra mRNA peaked at 9 h. IL-1,8 synthesis increased in response to increasing numbers ofspiro-chetes, whereas IL-lra synthesis did not. The preferential in-duction by B. burgdorferi ofIL-1i8 over IL-lra is an example of excess agonist over antagonist synthesis induced by a microbial pathogen, and may contribute to the destructive lesion ofLyme arthritis. (J. Clin. Invest. 1992. 90:906-912.) Key words: an-tagonist * cytokine * lipopolysaccharide * Lyme arthritis
Lyme disease is a multisystem disease caused by infection with the tick-borne spirochete Borrelia burgdorferi ( 1). Arthritis is a prominent feature ofthe disorder. Early in the illness, the spir-ochete probably spreads hematogenously to joints and may cause vague, migratory joint pain. After several months, many
Address reprint requests to Dr. Miller, Box 67, New England Medical
Center, 750 Washington Street, Boston, MA 021 1 1.
Receivedfor publication 3 October 1991 and in revisedform 21 February 1992.
patients begin to have brief attacks of arthritis in large joints (2). During the second and third years ofillness, a small subset ofpatients, primarily those with HLA-DR4 or HLA-DR2 (3, 4), develop chronic arthritis which may lead to erosion ofcarti-lage and bone (5). The synovial lesion in these patients is simi-lar to that in rheumatoid arthritis. The development ofchronic arthritis coincides with the appearance of a humoral immune response to two prominent outer surface proteins (Osp)' ofB. burgdorferi, Osp A and B, (6).
Cytokines, in particular interleukin-1 (IL-1) and tumor ne-crosis factor-a (TNF), are thought to contribute to the patho-genesis of the synovial lesion in rheumatoid arthritis (7-9). The potent biological effects ofthese cytokines are tightly regu-lated at many levels, including gene transcription, translation, protein processing and secretion, as well as naturally occurring inhibitors such as the IL-1 receptor antagonist (IL-1 ra) (10, 11) which occupies IL-1 receptors without inducing signal transduction.
Because Lyme arthritis is one of the few forms of chronic inflammatory arthritis in which the cause is known with cer-tainty (1), this illness presents a unique opportunity to study the direct effects ofthe pathogen on cytokine gene expression. In contrast to IgG, granulocyte/macrophage colony-stimulat-ing factor, and experimental endotoxemia which preferentially induce IL-1ra ( 12-15), we now demonstrate that live B. burg-dorferi preferentially induce synthesis of the agonist IL-1,8 compared to the antagonist IL-lra. This preferential induction was most marked in response to a low-passage, virulent prepa-ration of B. burgdorferi. Furthermore, the selective induction of IL- 1: appears to reside in a membrane component of B. burgdorferi whose activity is partially diminished by heat treat-ment but abrogated by sonication.
Sample collection. Whole heparinized blood was obtained from 23 healthy volunteers, age 22-40 yr, who had not ingested nonsteroidal anti-inflammatory drugs ( 16) within the previous 2 wk, and had no history or clinical symptoms suggestive ofLyme disease. The investiga-tion was approved by the Human Investigation Review Board ofNew England Medical Center. Informed consent was obtained from each donor.
All reagents and glassware were sterile and endotoxin free. Culture media and sterile water used to prepare Ficoll-Hypaque were passed through a hollow-fiber polysulfone capillary ultrafilter to remove mi-crobial products and endotoxin (17). RPMI 1640 was supplemented only with L-glutamine (10 mM) (Gibco Laboratories, Grand Island, NY) and 10 mM Hepes (M.A. Biologicals, Walkersville, MD). No antibiotics were included in any media used. In preliminary experi-
1. Abbreviations used in this paper: HP and LP, high and low passage (strains); Osp, outer surface protein; ra, receptor antagonist; TNF, tu-mor necrosis factor-a.
906 Miller et al.
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 90, September 1992, 906-912
ments, the effects on cytokine induction ofheat-inactivated (56°C for 1 h) 10% FCS (Hyclone Laboratories, Inc., Logan, UT) and 1% human
AB serum were compared. Human serum alone induced IL-Ira, whereas no induction ofeither IL-I or IL-Ira was detected in the pres-ence of 10% FCS (data not shown). Therefore, for these experiments,
FCS was used as a culture supplement.
B. burgdorferi. B. burgdorferi were grown in Barbour-Stoenner-
Kelly medium at 32°C in a humidified atmosphere containing 5%
CO2. Three different strains were used: 297, isolated from the cerebro-spinal fluid of a patient with Lyme disease; N40, an Ixodes dammini tick midgut isolate; and G39/40, initially isolated from an L dammini tick, but passaged for many years in the laboratory. Both high-passage (HP) and low-passage (LP) preparations of strain 297 were used; only the LP 297 strain retained infectivity in mouse inoculation studies as previously reported (18). Organisms were used at log phase growth. The spirochetes were pelleted by centrifugation at 10,000 g for 20 min at room temperature, then washed three times in RPMI and resus-pended at the desired concentration. In some experiments, spirochetes killed by sonication or heat (57°C for 2 h) were used. After pelleting by centrifugation at 10,000 g for 20 min at 4°C and four washes in cold 0.01 M PBS/5 mM magnesium chloride (pH 7.3), organisms were sonicated on ice by eight 15-s pulses, setting 6, ofa cell sonicator (Bran-son Sonic Power Co., Danbury, CT). The supernatant was clarified by
centrifugation at 10,000 gfor 30 min at4°C; protein content was deter-mined by optical density at 280 nm in a spectrophotometer (Gilford Instrument Laboratories, Inc., Oberlin, OH). Aliquots were stored at
-70°C. In other experiments, recombinant Osp B protein (a gift of
Drs. John Leong and Robert Kalish, New England Medical Center, Boston, MA) was used to stimulate cytokine production from PBMC. To investigate whether a soluble product of B. burgdorferi could con-tribute to the marked induction of IL-1I, the spirochetes were exten-sively washed, then cultured (25 x 106) in antibiotic-free RPMI for 24 h at 37°C or 32°C. After ultracentrifugation, the spirochete-free super-natant was filtered twice through 0.22-jim filters. The absence ofspiro-chetes or spirochetal fragments was verified by darkfield microscopy. The spirochete-conditioned media was then incubated with PBMC for 24 h and resulting cytokine production measured by radioimmunoas-says (RIAs).
PBMC cultures. Heparinized whole blood was fractionated by den-sity gradient centrifugation using Ficoll (Sigma Chemical Co., St. Louis, MO)-Hypaque (Winthrop Laboratories, New York, NY). Mononuclear cells were cultured (2.5 x 106 cells/ml) in 1.0 ml of complete RPMI with 10% heat-inactivated endotoxin-free FCS in 2.5-ml polypropylene tubes (Falcon Plastics, Oxnard, CA) with B. burg-dorferi or LPS (Escherichia coli 055:B5, Sigma Chemical Co.), or me-dium alone. Cultures were incubated for 24 h in a humidified, 5% CO2 atmosphere at 37°C. In experiments measuring total cytokine produc-
tion, cells were lysed by three freeze/thaw cycles (19). After centrifuga-tion at 400 gfor 15 min, the supernatants were removed, and the pellets containing cellular and spirochetal debris were discarded. In experi-ments measuring secreted vs. cell-associated cytokines, cell-free culture supernatant (containing secreted cytokines) was removed and replaced by an equal volume offresh media. The cell-containing tubes were then subjected to three freeze/thaw cycles to release cell-associated cyto-kines, centrifuged at 400 g for 15 min, and harvested as above.
Cvtokine RIAs. Specific RIAs for IL-1IB, TNF, IL-la, IL-6, and
IL-Ira were used (14, 19-22). lodination of IL-1,B and IL-Ira used the
Bolton-Hunter method; all other cytokines were iodinated by the chloramine-T method. Sensitivities for each of the RIAs was< 80 pg/ ml except for the IL-lra RIA which had a sensitivity of 300 pg/ml.
RNA isolation and Northern analysis. After incubation with B. buirgdorferi (five spirochetes per PBMC), LPS (10 ng/ml), or control media, total cellular RNA was extracted by lysis with 4 M guanidine isothiocyanate, followed by ultracentrifugation on a 5.7 M cesium chlo-ride cushion. Total RNA (20,g) was subjected to electrophoresis in
6.6% formaldehyde (Sigma Chemical Co.) and 1.2% agarose (Interna-tional Biotechnologies Inc., New Haven, CT), and then transferred to nylon membranes (Hybond-N, Amersham Corp., Arlington Heights,
IL) by capillary blotting. For quantitation ofmRNA levels, serial dilu-tions of RNA (2.50, 1.25, and 0.62 jug) were directly applied to nylon membranes using a filtration manifold apparatus (Schleicher & Schuell, Inc., Keene, NH). The membranes were exposed to short-wave UV light for 5 min to fix the RNA to the nylon matrix, and treated at 42°C for 2 h with prehybridization solution containing 10 mg/ml salmon sperm DNA. Membranes were then treated overnight with prehybridization solution containing 10 mg/ml ofsalmon sperm DNA and 32P-labeled nucleic acid probe. The probes used were a
1,075-bp fragment of human IL-I/3 precursor cDNA subcloned into pGEM2, 800-bp fragment of human IL-Ira precursor cDNA sub-cloned into pUC8, and the full length (2,000-bp) chicken ,B-actin cDNA subcloned in pGEM3. The DNA was labeled using [32P]dCTP (3,000 Ci/mmol, New England Nuclear, Boston, MA) and a random primed DNA labeling kit (Boehringer Mannheim, Mannheim, FRG).
After incubation, membranes were washed in 0.1% SDS, 1 X SSC at
42°C. Washed membranes were exposed overnight to Kodak KAR5
X-ray film (Eastman Kodak Co., Rochester, NY) at -70°C with an intensifying screen.
Statistical analysis. Total cytokine levels were expressed as mean±SEM of the indicated number ofdonors. Differences were ana-lyzed for significance using Student's t test for paired samples or analy-sis of variance using the computer program StatView (BrainPower, Inc., Calabasas, CA) on a Macintosh SE computer.
Cytokine synthesis induced by B. burgdorferi. Live B. burgdor-feri (297 LP), sonicated B. burgdorferi, and LPS were com-pared for their ability to induce cytokine synthesis from PBMC (Fig. 1). Live B. burgdorferi induced 65±11 ng/ml of IL-1I. Large amounts of TNF were also induced (82±18 ng/ml), whereas production of IL-ia and IL-6 were stimulated to a lesser extent (12±2 and 2±1 ng/ml, respectively). Live B. burgdorferi induced significantly more IL-113 (P < 0.01) and TNF (P <0.001 ) than sonicated B. burgdorferi or LPS. Synthe-sis of IL-1 a, IL-6, and IL-1 ra induced by all three stimuli was similar. Unexpectedly, live B. burgdorferi stimulated fivefold more IL-113 than IL-1 a (P < 0.0001 ) and sevenfold more IL-1I3
3- M TNF
'r t L
0- E3 ~~~~~IL-1ira
live sonicated LPS
Figure 1. Induction of IL-1I3, IL-la, TNF, IL-6, and IL-lra by live B. burgdorferi (LP 297), sonicated B. burgdorferi, or LPS. PBMC from 23 donors were cultured for 24 h, and the resulting supernatants tested in cytokine-specific RIAs. Live B. burgdorferi induced signifi-cantly more IL-1IB (*P < 0.01) and TNF (**P < 0.001) than the amounts induced by sonicated B. burgdorferi or LPS.
Borrelia burgdorferi Preferentially Induce IL-1j3 over IL-I Receptor Antagonist 907
U IL-1 p
150- [l IL-lra
297 LP 297 HP N40 G39/40
Figure 2. Induction of IL-1I13 and IL-Ira by four preparations of B. burgdorferi, strains 297 (LP and HP), N40, and G39/40. Data are mean±SEM of six donors. Significantly more IL-1I3 was induced by the virulent strain 297 LP compared to the other three strains (** vs. *, P < 0.001I); no strain differences in IL-Ira induction were seen.
than IL-Ira (P < 0.001I). In contrast, the ratios of IL- 113 to IL- 1 a and IL-113l to IL-lIra induced by sonicated B. burgdorferi or LPS were close to 1.
Strain differences and Borrelial factors. Because of the markedly high levels of IL- 113 induced by LP 297 live B. burg-dorfieri, the cytokine-inducing capacity ofother strains was also tested. LP 297 live B. burgdorferi induced 148±16 ng/ml IL-I11, significantly more than the avirulent strains HP 297,
N40, or G39/40 (P < 0.001I) (Fig. 2). However, no difference was observed in the amount ofIL- 1ra induced by these strains.
To elucidate further the microbial factor(s) responsible for the preferential induction ofIL- 113l, four preparations oflive vs. heat-treated B. burgdorferi were compared. Heating signifi-cantly decreased induction of IL-1I3 by all four B. burgdorferi preparations (P < 0.05) (Fig. 3, left), whereas no differences were seen after heating in the induction of IL- Ira (Fig. 3, right).
The possibility that live B. burgdorferi, particularly LP 297, produced a soluble factor which could stimulate IL-113l produc-tion was investigated. Because BSK medium (which contains neopeptone, yeastolate, tryptone, gelatin, bovine serum albu-min, and rabbit serum) is itself a potent inducer of cytokines (20-30 ng/ml ofIL-1I3 or TNF), we prepared spirochete-con-ditioned RPMI to test as a stimulant for cytokine induction.
svv * ~~~~IL-1, O v
80 - O heated
.0 40* ; i
Spirochete-conditioned RPMI was incubated with PBMC from four donors. Levels of IL-113, IL-Ira, and TNF by this conditioned media were below limits of detection by RIAs (data not shown).
Because immunoaffinity purified B. burgdorferi lipopro-teins (including a mixture ofOsp A and B) induce TNF (23), the contribution ofOsp B, a major outer surface protein ofB. burgdorferi, to the induction ofIL-I,B and IL-Ira was then ex-amined (Fig. 4). At the highest concentrations tested, Osp B induced 1.0±0.2 ng/ml of IL-1,B, whereas at lower concentra-tions, Osp B induced <0.10 ng/ml IL-1,B. In contrast, the highest concentration ofOsp B tested induced 4.1± 1.1 ng/ml IL-Ira, whereaslowerconcentrationsinduced 1-2 ng/ml. Poly-myxin B had no effect on the response to this stimulus. Thus, Osp B made little contribution to the marked induction of
IL-11, but could contribute to some extent to induction of
Secretion ofIL-1a, IL-1I1, and IL-i ra. The proportions of cell-associated vs. secreted IL-I a, IL-113, and IL-Ira induced by five spirochetes per PBMC (live LP 297) were compared to those induced by LPS (10 ng/ml) in PBMC of 18 donors. After stimulation with B. burgdorferi, 34±2% of total IL-la and
78±1% of total IL-11 were secreted. After LPS stimulation,
21±4% (P = 0.01) of IL-la and 63±3% of total IL-1I8 (P
< 0.001) were secreted. No differences were seen in the amounts or proportions ofsecreted IL-Ira induced by B. burg-dorferi or LPS (72±3% and 76±5%).
Dose-dependent effect ofB. burgdorferi on cytokineproduc-tion. Increasing the number of spirochetes (live LP 297) per PBMC (from 0.08 to 60 organisms per PBMC) resulted in a dose response for IL- 11 synthesis (Fig. 5, top). In contrast, increasing the number of spirochetes from 0.08 to 0.74 per PBMC resulted in a shallow dose-response curve for IL-Ira, which reached a plateau thereafter despite an increase to 60 spirochetes per PBMC. No dose response was seen for IL-Ia. The ratio of IL-13 to IL-Ira increased to 14:1 (inset). When LPS (0.001-1000 ng/ml) was used as a stimulus, production ofIL-1I3 and IL-Ira increased to maximal amounts in response to 10 ng/ml LPS, then decreased at higherconcentrations (Fig. 5, bottom). The ratio of IL-1,8 to IL-Ira increased to a maxi-mum at a concentration of 1 ng/ml LPS, then remained at a plateau (inset). IL-I a production increased linearly in re-sponse to LPS concentrations from 0.001 to 0.10 ng/ml, then
297LP 297 HP N40 G39/40 297 LP 297 HP N40 G39/40
Figure 3. Induction ofIL-1I3 (left) and IL-Ira (right) by paired samples offour preparations oflive or heated B. burgdorferi in PBMC from five donors. For all preparations of B. burgdorferi, heating significantly decreased the synthesis of IL-1B (all pairs, P < 0.05), but had no effect on induction ofIL-Ira.
908 Miller et al.
500 ng/ml 50 ng/ml 5 ng/ml 500 pg/ml
Figure 4. Induction of IL-1:3 and IL-lra by recombinant Osp B pro-tein. Log dilutions ofrecombinant Osp B were incubated with PBMC of four donors (mean±SEM).
remained unchanged despite increasing the concentration of
LPS to 1,000 ng/ml.
Kinetics ofIL-IA, IL-la, and IL-Ira synthesis. No differ-ences were seen in the time of production of IL-la or IL-Ira induced by LP 297 B. burgdorferi or LPS (Fig. 6). In contrast, by 4 h, B. burgdorferi induced 24±5 ng/ml of IL- fl, whereas
LPS induced 2±3 ng/ml (P < 0.03). Compared to LPS, more striking differences in B. burgdorferi-induced IL-l1B protein synthesis were seen at 12 h (95±18 vs. 23±13 ng/ml, P < 0.01) and 24 h (109±6 vs. 19±8 ng/ml, P < 0.006). By 4 h, B. burgdorferi induced the synthesis of more than fivefold more
IL-iI3 than IL-Ira, whereas no difference was seen in LPS-in-duced synthesis of IL- 3l and IL-Ira. For both stimulants, ra-tios of IL- to IL- ra remained unchanged at 12 and 24 h.
Accumulation ofmRNAfor IL-JI andIL-Ira. The kinetics ofsteady-state IL- 1l3 and IL- ra mRNA induced by LP 297 B. burgdorferi or LPS were compared by Northern and dot blot analysis. Levels ofmRNA for IL- (l: and IL- ra were nondetec-table in unstimulated cells. The time course for IL- 13 mRNA levels were similar during the first 4 h after B. burgdorferi or
LPS stimulation; IL- 1:3 mRNA accumulation reached a peak at 4 h (Fig. 7). Protein production showed eightfold more IL- 13 stimulated by B. burgdorferi compared to LPS at 4 h ( 17 vs. 2.1 ng/ml). In a longer time course study, IL-1(3 mRNA reached a peak at 3 h and declined until 12 h after stimulation by B. burgdorferi or LPS (Fig. 8 A). However, IL-1B mRNA levels again increased 18 h after exposure to B. burgdorferi, and remained elevated at 24 h. In contrast, LPS-induced IL-1,B mRNA continued to decline until 24 h. B. burgdorferi-in-duced IL- 1: protein continued to rise at 24 h; LPS-induced IL-1O protein reached near-peak levels by 6 h.
.01 .1 1 10 100 spirochetes/PBMC
.001 .01 .1 1
Figure 5. Dose-dependent effect of B. burgdor-feri on the synthesis ofIL-lIB. (Top) Synthesis
of IL- I,, IL- a, and IL- I ra in response to in-
IL-la creasing numbers of live spirochetes per
PBMC. (Bottom) Synthesis of IL- Id, IL- a, and IL- I ra in response to increasing concen-trations of LPS. Data are mean±SEM of six
10 100 1000 donors. Insets show the ratios of IL-1,B to IL-
Ira over these different concentrations of stimuli.
Borrelia burgdorferi Preferentially Induce IL-1f3 over IL-I Receptor Antagonist 909
I BILI3I I I I
0 4 8 12 16 20 24 C
0- _ I I I I I I 4 8 12 16 20 24
_ai - tQ 4:1
.. bgd_- f-_i
Figure 7. Early kinetics of IL- 1,B protein and mRNA induced by LPS and B. burgdorferi. (A) The synthesis ofIL-1O protein at 30 min and
1, 2, and 4 h. (B) The corresponding IL-lI3 mRNA. (C) The same blot probed for (-actin.
Figure 6. Kinetics of production of IL-la, IL-113, and IL-Ira in re-sponse to live B. burgdorferi or LPS. Data (mean±SEM) for three donors are shown. No differences were seen in the kinetics of pro-duction of IL-lca or IL-Ira, whereas significantly more IL-I was in-duced by live B. burgdorferi at 4, 12, and 24 h than LPS (*P < 0.03, **P < 0.01, and ***P < 0.006).
Similar to IL-1( mRNA, the accumulation of IL-1ra mRNA induced by B. burgdorferi or LPS were comparable until 4 h (data not shown). Thereafter, LPS-induced IL-1ra
mRNA continued to increase until reaching peak levels at 9 h, then decreasing gradually until 24 h (Fig. 8 B). In contrast, B. burgdorferi-induced IL- 1ra mRNA peaked at 12 h, and then gradually declined. No difference in IL- 1 ra protein induced by the two stimuli was seen over 24 h.
By using specific RIAs, we have shown a remarkable preferen-tial induction ofIL- 1l: and TNF by B. burgdorferi. Unlike LPS, live B. burgdorferi preferentially induced IL- over the IL-I ra, as well as IL-Ia. This selectivity was most dramatic when the virulent strain LP297 was used. Habicht et al. (24) reported that B. burgdorferi stimulate marked IL-1 biologic activity from murine macrophages, P388Dl cells, and human PBMC. However, they used nonspecific fibroblast and T cell prolifera-tion assays which do not distinguish between IL- 1(3 and IL-I a, and may respond to other cytokines such as IL-2, IL-4, IL-6, or TNF, alone or in combination (25-27). Furthermore, it is known that the bioassays for IL- 1 are affected by cytokine an-tagonists such as IL- 1 ra and soluble cytokine receptors (28-30).
The present studies show that the component(s) ofB. burg-dorferi associated with the preferential induction of IL- 1l: is heat resistant but eliminated during the sonication procedure. We were unable to demonstrate that spirochete-conditioned
910 Miller et al.
RNA ".11 t",
0 .0 -.
h 3 6 9 12 18 24 3 6 9 12 18
LPS B. burgdorferi
* * **: * *#90ikk t * *
_ .+. ;t *@,0f.9..
_r- 10 -
h 3 6 9 12 18 24 3 6 9 12 18 24
LPS B. burgdorferi
eAiL*&* *** * # 2.50
* ., 1.25
X * -S@- - *------2.50
Figure 8. Late kinetics of IL-1O protein and mRNA induced by LPS and B. burgdorferi. (A) The synthesis ofIL-1O protein at 3, 6, 9, 12, 18, and 24 h, with the corresponding levels of IL- 1O mRNA shown by dilutional analysis below. (B) IL-Ira protein and mRNA induced by LPS and B. burgdorferi. (C) The same blot probed for /-actin.
medium contained soluble borrelial product(s) which induced IL-1:. Therefore, we conclude that the component(s) of B. burgdorferi responsible for IL-1/3 induction resides in the cell wall, and is not readily secreted, but rather is associated with the structural integrity ofthe wall. This suggests a requirement for intact cells. Although one could speculate that such a puta-tive spirochetal wall structure satisfies a simple "particulate size" requirement for monocyte stimulation, the data on strain and passage differences suggest another mechanism. The com-ponent(s) which preferentially trigger IL-1/3 gene expression and protein synthesis may be a complex ofcomponents which varies among strains and/or is lost with repeated in vitro pas-sage ofB. burgdorferi. Extensive passage of strain 297 renders the spirochete noninfectious ( 18) and as we show, its ability to induce high levels of IL-1/3. Preferential induction of IL-i over IL- Ira may relate to infectivity and production ofdisease in vivo.
The onset of chronic arthritis in Lyme disease coincides with the appearance of a humoral immune response to Osp A and Osp B (6). Furthermore, Osp B may be lost during serial laboratory passage ofB. burgdorferi; in some strains, this corre-
sponds to loss of infectivity in mice (31, 32). We found that recombinant Osp B induced little IL-1o.
The relative synthesis of the related but distinct genes IL-
13, IL-1 a, and IL-1 ra is stimulus dependent ( 12, 33, 34). Solu-ble stimuli such as LPS, phytohemagglutinin, and toxic shock syndrome toxin-1 induce nearly equal amounts of IL-l1a, IL-1/, and TNF (5-15 ng/ml) with somewhat more IL-la pro-duced by most donors (35, 36). Particulate stimuli such as Staphylococcus epidermidis induce threefold more IL-1O than
IL-1a. However, there appears to be more IL-1ra than IL-1: synthesized by PBMC using various stimuli (37). Soluble IgG or granulocyte/macrophage colony-stimulating factor induce IL-lra but not IL-la or IL-l1/ ( 12-14). More IL-lra than IL-I is found in children with systemicjuvenile rheumatoid arthritis (38). In experimental human endotoxemia, at least 100-fold more IL-Ira is found in the circulation than IL-1/3( 15). Thus, live B. burgdorferi, ofany stimuli studied to date, appears to be unique in its marked preferential activation of IL-1/. In our dose-response experiments, increasing amounts of LPS in-duced parallel increases in IL-1/ and IL-1 ra. In contrast, B. burgdorferi induced larger amounts of the agonist IL- I than the antagonist IL-1 ra. By 4 h after stimulation with B. burgdor-feri, IL-1/ protein exceeded IL-1 ra protein by greater than five-fold. Some reports suggest that a 10-50 molar excess ofIL-Ira is needed to inhibit 50% of IL-I binding to T cells (39). Changes in the relative proportions of cytokines and their an-tagonists in vivo at different sites in the body or different times after infection clearly has important biological implications.
The kinetics of IL-1/ mRNA production induced by B. burgdorferi resemble the pattern observed with LPS (40) within the first 12 h after stimulation. Nonetheless, increased protein production is seen as early as 4 h after stimulation, suggesting that B. burgdorferi enhances translational efficiency ofIL- /3 rather than affecting transcription. The second peak of IL-1/ mRNA induced by B. burgdorferi may reflect failure of B. burgdorferi to induce a repressor which down-regulates or destabilizes this mRNA, as described after LPS or PMA stimu-lation (41 ).
We believe our findings are relevant to the pathogenesis of Lyme disease. Early in the illness, low-grade fever, malaise, and marked fatigue are common symptoms, which may be me-diated by IL-I and other cytokines (42). However, the specific immune response to the organism seems to be suppressed early in the illness, and patients commonly experience only vague joint pain despite the presumed early spread of B. burgdorferi to synovial tissue. After several months, as the specific cellular and humoral immune responses expand to multiple spiroche-tal polypeptides, patients have briefattacks ofarthritis in large
Alterations in the balance of cytokines and their antagonists are likely involved in the sudden turning on and offofthe inflammatory response in these patients, but the mechanisms controlling such alterations are not yet known. At the time of maximal expansion of the immune response, which usually occurs during the second or third years ofillness, a genetically susceptible subset of patients, particularly those with HLA-DR4, may develop chronic arthritis (4). As in rheumatoid arthritis, erosion of cartilage and bone with elevated levels of collagenase and PGE2 may occur (5). The preferential induc-tion by B. blurgdorferi ofthe agonist IL-1/ , over the antagonist IL-1 ra, may contribute to the destructive lesion of Lyme ar-thritis.
Borrelia burgdorferi Preferentially Induce IL-ifi over IL-1 Receptor Antagonist 911
The contributions ofB. Reinhardt, J. Mitchell, Dr. J. G. Schaller, and
Dr. M. S. Klempner are gratefully acknowledged. This work was sup-ported in part by the Charles A. Hood Foundation and Massachusetts Arthritis Foundation (Dr. Miller), and National Institutes of Health grants AR-20358 andAR-40576 (Dr. Steere) and AI-15614 (Dr. Dina-rello).
1. Steere, A. C., R. L. Grodzicki, A. N. Komblatt, J. E. Craft, A. G. Barbour, W. Burgdorfer, G. P. Schmid, E. Johnson, and S. E. Malawista. 1983. The spiro-chetal etiology ofLyme disease. N. Engl. J. Med. 308:733-40.
2. Steere, A. C., R. T. Schoen, and E. Taylor. 1987. The clinical evolution of Lyme arthritis. Ann. Intern. Med. 107:725-3 1.
3. Steere, A., A. Gibosky, M. Patarroyo, R. Winchester, J. Hardin, and S. Malawista. 1979. Chronic Lyme arthritis: clinical and immunogenetic differen-tiation from rheumatoid arthritis. Ann. Intern. Med. 90:896-901.
4. Steere, A., E. Dwyer, and R. Winchester. 1990. Association of chronic Lyme arthritis with HLA-DR4 and HLA-DR2 alleles. N. Engi. J. Med. 323:219-223.
5. Steere, A., C. Brinckerhoff, D. Miller, H. Drinker, E. Harris, and S. Mala-wista. 1980. Elevated levels ofcollagenase and prostaglandin E2 from synovium associated with erosion of cartilage and bone in a patient with chronic Lyme arthritis. Arthritis Rheum. 23:591-9.
6. Kalish, R., J. Leong, and A. Steere. 1991. Delay in the immune response to outer-surface proteins (OSP) A and B ofB. burgdorferi: correlation with arthritis and treatment failure in susceptible patients with Lyme disease. ArthritisRheum.
7. Arend, W., and J.-M. Dayer. 1990. Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum. 33:305-315.
8. Miller, L. C., and C. A. Dinarello. 1987. Biologic activities ofinterleukin-1 relevant to rheumatic diseases. Pathol. Immunopathol. Res. 6:22-36.
9. Eastgate, J. A., J. A. Symons, N. C. Wood, F. M. Grinlinton, F. S. diGio-vine, and G. W. Duff. 1988. Correlation of plasma interleukin 1 levels with disease activity in rheumatoid arthritis. Lancet. ii:706-9.
10. Arend, W., H. Welgus, R. Thompson, and S. Eisenberg. 1990. Biological properties ofrecombinant human monocyte-derived interleukin 1 receptorantag-onist. J. Clin. Invest. 85:1694-1697.
11. Eisenberg, S., R. Evans, W. Arend, E. Verderber, M. Brewer, C. Hannum, and R. Thompson. 1990. Primary structure and functional expression from com-plementary DNA ofa human interleukin-I receptor antagonist. Nature (Lond.).
12. Arend, W., F. Joslin, R. Thompson, and C. Hannum. 1989. IL-1 inhibitor from human monocytes: production and characterization ofbiological activities. J. Immunol. 143:1851-1854.
13. Arend, W., M. Smith, R. Janson, and F. Joslin. 1991. IL-I receptorantago-nist and IL- I# production in human monocytes are regulated differently. J. Im-munol. 147:1530-1536.
14. Poutsiaka, D., B. Clark, E. Vannier, and C. Dinarello. 1991. Production in interleukin- 1ra and interleukin- I, by peripheral blood mononuclear cells is dif-ferentially regulated. Blood. 78:1275-81.
15. Granowitz, E., A. Santos, D. Poutsiaka, J. Cannon, D. Wilmore, S. Wolff, and C. Dinarello. 1991. Production of interleukin- 1 receptor antagonist during experimental endotoxaemia. Lancet. ii:1423-1424.
16. Endres, S., R. Ghorbani, V. Kelley, K. Georgilis, G. Lonnemann, J. Van derMeer, J. Cannon, T. Rogers, M. Klempner, P. Weber, et al. 1989. The effect of dietary supplementation with n-3 polyunsaturated fatty acids on the synthesis of interleukin- 1 and tumor necrosis factor by mononuclear cells. N. Engl. J. Med.
17. Schindler, R., and C. Dinarello. 1989. A method for removing interleu-kin- 1 and tumor necrosis factor-inducing substances from bacterial cultures by ultrafiltration with polysulfone. J. Immunol. Methods. 116:159-165.
18. Georgilis, K., A. Steere, and M. Klempner. 1991. Infectivity ofBorrelia burgdorferi correlates with resistance to elimination by phagocytic cells. J. Infect. Dis. 163:150-155.
19. Endres, S., R. Ghorbani, G. Lonnemann, J. van der Meer, and C. Dina-rello. 1988. Measurement ofimmunoreactive interleukin-lI from human mono-nuclear cells: optimization ofrecovery, intrasubject consistency, and comparison with interleukin-la and tumor necrosis factor. Clin. Immunol. Immunopathol.
20. Lonnemann, G., S. Endres, J. Van der Meer, J. Cannon, K. Koch, and C. Dinarello. 1989. Differences in the synthesis and kineticsofreleaseofinterleukin-la, interleukin-l,B and tumor necrosis factor from human mononuclear cells. Eur. J. Immunol. 19:1531-1536.
21. van der Meer, J., S. Endres, G. Lonnemann, J. Cannon, T. Ikejima, S. Okusawa, J. Gelfand, and C. Dinarello. 1988. Concentrationsofimmunoreactive human tumor necrosis factor alpha produced by human mononuclear cells in vitro. J. Leukocyte Biol. 43:216-223.
22. Schindler, R., J. Mancilla, S. Endres, R. Ghorbani, S. Clark, and C. Dina-rello. 1990. Correlations and interactions in the production of IL-6, IL-1 and TNF in human blood mononuclear cells. Blood. 75:40-46.
23. Radolf, J., M. Norgard, M. Brandt, R. Isaacs, P. Thompson, and B. Beutler. 1991. Lipoproteins of Borrelia burgdorferi and Treponema pallidum activate cachectin/tumor necrosis factor synthesis. J. Immunol. 147:1968-1974.
24. Habicht, G. S., G. Beck, and J. L. Benach. 1988. The role ofinterleukin-I in the pathogenesis ofLyme disease. Ann. N. Y. Acad. Sci. 539:80-6.
25. Mandrup-Poulsen, T., K. Bendtzen, C. Dinarello, and J. Nerup. 1987. Potentiation ofIL-1 mediated cell killing by TNF: human tumor necrosis factor potentiates human interleukin-l mediated rat pancreatic , cell cytotoxicity. J. Immunol. 139:4077-4082.
26. Movat, H., C. Burrowes, M. Cybulsky, and C. Dinarello. 1987. Acute inflammation and a Shwartzman-like reaction induced by interleukin-1 and tu-mor necrosis factor: synergistic action ofthe cytokines in theinduction ofinflam-mation and microvascular injury. Am. J. Pathol. 129:463-467.
27. Shalaby, M., A. Waage, L. Aarden, and L. Espervik. 1989. Endotoxin, TNF, and IL-1 induce IL-6 production in vivo. Clin. Immunol. Immunopathol.
28. Nophar, Y., 0. Kemper, C. Brakebusch, H. Engelmann, R. Zwang, D. Aderka, H. Holtmann, and D. Wallach. 1990. Soluble forms oftumor necrosis factor receptors: the cDNA for the Type I TNF-R, cloned using amino acid sequence data ofits soluble form, encodesboth the cell surface and a soluble form ofthe receptor. EMBO (Eur. Mol. Biol. Organ.) J. 9:3269-78.
29. Novick, D., H. Engelmann, M. Revel, 0. Leitner, and M. Rubinstein. 1991. Monoclonal antibodies to the human IL-6 receptor: affinity purification, ELISA, and inhibition ofligand binding. J. Chromatogr. 10:137-146.
30. Engelmann, H., D. Novick, and D. Wallach. 1990. Two TNF-binding proteins purified from human urine: evidence forimmunological cross-reactivity with cell surface TNF receptors. J. Biol. Chem. 265:1531-1536.
31. Schwan, T., and W. Burgdorfer. 1987. Antigenic changesofBorrelia burg-dorferi as a result ofin vitro cultivation. J. Infect. Dis. 156:852-56.
32. Schwan, T., W. Burgdorferi, and C. Garon. 1988. Changes in infectivity and plasmid profile of the Lyme disease spirochete, Borrelia burgdorferi, as a result ofin vitro cultivation. Infect. Immun. 56:1831-1836.
33. Dinarello, C. 1991. Interleukin-1 and interleukin-1 antagonism. Blood. 77: 1627-1652.
34. Arend, W. 1991. Interleukin-1 receptor antagonist. J. Clin. Invest. 88:1445-1451.
35. Endres, S., J. Cannon, R. Ghorbani, T. Dempsey, S. Sisson, G. Lonne-mann, J. Van der Meer, S. Wolff, and C. Dinarello. 1989. In vitro production of IL-lI#, IL-la, TNF, and IL-2 in healthy subjects: distribution, effect ofcyclooxy-genase inhibition and evidence ofindependent gene regulation. Eur. J. Immunol.
36. Ikejima, T., S. Okusawa, J. van der Meer, and C. Dinarello. 1988. Induc-tion by Toxic-shock syndrome toxin-I ofa circulating tumor necrosis factor-like substance in rabbits and ofimmunoreactive tumor necrosis factor and interleu-kin-l from human mononuclear cells. J. Infect. Dis. 158:1017-1025.
37. Dinarello, C., and R. Thompson. 1991. Blocking IL-1: interleukin 1 re-ceptor antagonist in vivo and in vitro. Immunology Today. 12:404-409.
38. Prieur, A.-M., M.-T. Kaufmann, C. Griscelli, and J.-M. Dayer. 1987. Specific interleukin-l inhibitor in serum and urine of children with systemic juvenile chronic arthritis. Lancet. ii:1240-1242.
39. Granowitz, E., B. Clark, J. Mancilla, and C. Dinarello. 1991. Interleukin-l receptorantagonist competitively inhibitsthe bindingofinterleukin- 1 to the type II interleukin-l receptor. J. Biol. Chem. 266:14147-14150.
40. Fenton, M., B. Clark, K. Coflins, A. Webb, A. Rich, and P. Auron. 1987.
Transcriptional regulation ofthe human prointerleukin-l betagene. J. Immunol.
41. Fenton, M., M. Vermeulen, B. Clark, A. Webb, and P. Auron. 1988. Human prointerleukin-1 gene expression in monocytic cells is regulated by two distinct pathways. J. Immunol. 140:2267-2273.
42. Smith, J., W. Urba, R. Steis, J. Janik, B. Fenton, W. Sharfman, K. Conlon, M. Sznol, S. Creekmore, N. Wells, et al. 1990. Interleukin-1 alpha: results of a phase I toxicity and immunomodulatory trial. Am. Soc. Clin. Oncol. 9:717.
43. Steere, A. 1989. Lyme disease. N. Engl. J. Med. 321:586-596.