• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of clinexpimmunolLink to Publisher's site
Clin Exp Immunol. Jul 1999; 117(1): 63–69.
PMCID: PMC1905466

Protective role of NK1.1+ cells in experimental Staphylococcus aureus arthritis


In a model of Staphylococcus aureus-induced septic arthritis in C57Bl/6 mice we investigated the role of natural killer (NK) cells in the development of disease. Depletion of NK1.1+ cells was achieved by repeated injections of the PK136 antibody, whereas control mice received an irrelevant monoclonal antibody, O1C5.B2. Both groups of mice then received injections intravenously with 2 × 107 live S. aureus LS-1 secreting toxic shock syndrome toxin-1 (TSST-1). The mice were evaluated for 16 days with regard to weight, mortality and arthritis. Nine days after bacterial injection, 9/19 mice depleted of NK cells had developed arthritis compared with 1/17 in the control group (P = 0.01). The experiment was repeated twice with the same outcome. NK cell-depleted and control mice displayed the same degree of histopathological signs of arthritis at day 16. Depletion of NK cells did not affect uptake of bacteria by phagocytic cells in vitro, or bacterial clearance in vivo. In NK cell-depleted mice there was a tendency to increased levels of antibodies to TSST-1, whereas total immunoglobulin levels were similar to those in controls. NK cell depletion of non-infected mice did not affect the magnitude of inflammatory response during the T cell-dependent cutaneous DTH reaction to oxazolone, or during granulocyte-mediated inflammation. However, specific antibody responses to oxazolone were greatly increased in depleted animals. In conclusion, our study demonstrates that NK cells protect against arthritis during S. aureus infection. This outcome does not seem to be due to an influence on bacterial clearance, but could be due to an interaction with the host anti-inflammatory mechanisms.

Keywords: NK1.1+ cells, Staphylococcus aureus, septic arthritis


Natural killer (NK) cells are known to respond vigorously in some experimental models of inflammatory and infectious conditions, determining a favourable outcome in Leishmania infection [1] but a fatal one in the Shwartzman reaction, a model of lipopolysaccharide (LPS)-induced septic shock [2]. In both cases, the contribution of NK cells to the disease process is dependent on the swift production of interferon-gamma (IFN-γ), before antigen-specific responses have come into effect [1, 3]. In vitro studies have shown that upon stimulation of human lymphocytes with Staphylococcus aureus antigens, primarily NK cells become activated and produce IFN-γ [4, 5].

An animal model for S. aureus-induced septic arthritis through i.v. injection of live bacteria has recently been established [6, 7]. The role of lymphocytes in disease development has been extensively studied. Abdelnour et al. have shown that T lymphocytes play a major pathogenic role in the disease [8, 9]. Also, B lymphocytes contribute to arthritis, as shown in studies demonstrating that B cell-deficient Xid mice display a less severe arthritis [10]. In experimental S. aureus arthritis, IFN-γ is known to influence disease outcome. Thus, administration of IFN-γ is harmful regarding the onset as well as the progression of arthritis, whereas treatment with anti-IFN-γ MoAbs ameliorates the disease [11].

In the present study, we wished to investigate the role of NK1.1+ cells in the development of S. aureus arthritis. NK cells have several potential ways to influence disease outcome in addition to IFN-γ production, as they have the capacity to present superantigens [12] and to influence B cell reactivity, either by promoting antigen-specific responses [13] or by abrogating B cell-mediated disease manifestations [14].



C57Bl/6 mice were bought from Bomhåltgård (Ry, Denmark) and maintained in the animal facility at the Departments of Rheumatology and Clinical Immunology, University of Göteborg. Mice were housed 5–10 in each cage under standard conditions of temperature and light and fed laboratory chow and water ad libitum.

Depletion of NK cells

MoAbs were produced by the mouse IgG2a hybridoma PK136 reactive with the NK1.1 antigen and used to deplete NK cells. Efficacy of depletion was checked with FACS and an NK cell in vitro assay as described below. MoAbs from the IgG1 hybridoma O1C5.B2 recognizing a herpes simplex virus antigen were used as control antibodies. NK cell depletion started 3 days prior to induction of septic arthritis by i.p. injection of 100 μg of either MoAb, and continued by bi-weekly i.p. injections of 200 μg of respective MoAbs after bacterial inoculation.

In vitro NK cell activity assay

One hundred micrograms of PK136 or control O1C5.B2 antibody were administered intraperitoneally to C57Bl/6 mice and after 24 h an assay for in vitro cytotoxic activity of spleen cells was performed as previously described [15]. Briefly, a suspension containing 107/ml spleen cells was serially diluted. One hundred microlitres of each dilution were set in triplicates on a 96-well round-bottomed dish to give effector:target ratios of 200-100-50-25:1. Target 51Cr-labelled YAC-1 mouse lymphoma cells were suspended to 5 × 104/ml and 0.1 ml was added to each well. After incubation for 4 h at 37°C, supernatants containing released 51Cr were collected and counted in a Packard Cobra gamma counter. Specific lysis was calculated by the formula: specific lysis = (experimental value − spontaneous value)/(maximal value − spontaneous value) × 100%, where spontaneous release was derived from wells without effector cells and maximal release from wells where detergent (SDS) was added.

Flow cytometry

In order to assess the efficacy of NK cell depletion, flow cytometry and MoAb stainings were applied. C57Bl/6 mice were administered one i.p. injection of 200 μg NK cell-depleting (n = 3) or control antibody (n = 4). Spleen cells were obtained after 24 h, washed, counted and 1 × 106 cells were suspended in 75 μl PBS–bovine serum albumin (BSA) and incubated at 4°C for 45 min with PE-labelled anti-NK1.1 (Pharmingen, San Diego, CA) at a dilution of 1:20, together with either FITC–anti-NK/5E6 (PharMingen) (the 5E6 epitope is expressed on C57Bl NK cells and a subset of T cells) or FITC–anti-NK/2B4 (PharMingen) (expressed on a subset of C57Bl NK cells) at 1:125. After washing three times cells were suspended in 250 μl PBS–BSA and counted in a FACstar (Becton Dickinson, San Jose, CA). The frequency of NK1.1+ T cells was analysed by two-colour staining with combinations of antibodies to NK1.1, CD4, CD8 and CD3.

Bacterial strain and culture

Staphylococcus aureus strain LS-1 was originally isolated from a swollen joint of a spontaneously arthritic NZB/W mouse [6]. This bacterial strain is coagulase- and catalase-positive and produces large amounts of toxic shock syndrome toxin-1 (TSST-1). Bacteria were cultured on blood agar for 24 h, then reincubated on blood agar for another 24 h. Bacteria were kept frozen at −20°C in PBS containing 5% BSA and 10% dimethylsulfoxide (C2H6OS) until use. Before each experiment, the bacterial solution was thawed, washed in PBS twice, and diluted in PBS to achieve the desired concentration of bacteria.

Administration of bacteria

On day 0, 0.2 ml of an S. aureus suspension containing 1 × 108 colony-forming units (CFU)/ml in physiological saline was injected in a lateral tail vein of each mouse. Viable counts in the leftover solution were determined to ascertain the number of bacteria injected.

Clinical evaluation of arthritis

Mice were followed individually to obtain a record of weight, mortality and arthritis. Limbs were inspected on days 0, 1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 14, 15 and 16 after bacterial inoculation by two independent blinded observers. Arthritis was defined as visible erythema and/or swelling of at least one joint. To evaluate the intensity of arthritis, a clinical scoring (arthritic index) was used as previously described [16]. Briefly, macroscopic inspection of mice yielded a score of 0–3 points for each limb (1 point = mild swelling and/or erythema; 2 points = moderate swelling and erythema; 3 points = marked swelling and erythema). In exceptional cases a mouse could be scored as arthritic with erythema in only one joint, but in all cases it displayed significant swelling of that joint the following day or two. The arthritic index was constructed by dividing the total score by the number of animals used in each experiment group.

Bacteriological examination

Samples for bacteriological analysis of joints were obtained using charcoaled sticks, after dissection of left talocrural and left radiocarpal joints, and transferred to 5% horse blood agar. The culture was considered positive if > 10 S. aureus colonies were present after 48 h of incubation at 37°C. The left kidney was aseptically removed, homogenized at 4°C, diluted in nutrient culture medium and inoculated on horse blood agar in serial dilutions to estimate the bacterial growth. All bacterial isolates were tested for catalase and coagulase activity.

Histopathological examination

At sacrifice, histological examination of joints was performed after routine fixation, decalcification and paraffin embedding. Tissue sections stained with haematoxylin and eosin (H–E) were prepared from upper extremities (elbow, wrist, carpal joints, toes) and lower extremities (knee, ankle, tarsal joints, and toes). The identities of the joint sections were coded, and sections were then examined with regard to synovial hypertrophy, defined as synovial membrane thickness of more than two cell layers [17], pannus formation (synovial tissue overlaying joint cartilage) and cartilage and subchondral bone destruction. Further, from each mouse the right kidney was taken for H–E staining. Infiltrates of mononuclear cells on kidney sections were evaluated on a semiquantitative basis as previously described [18]. Briefly, infiltrates were categorized as either focally distributed of small size (size smaller than one glomerulus), medium (equal to one to five glomeruli) or heavy (larger than five glomeruli), diffusely distributed, perivascular or glomerular. For each category, the extent of infiltration was judged as mild, moderate or severe.

Immunofluorescent staining of joints and spleens

In a repeat experiment, the lower extremity of control mice was taken at sacrifice on day 10 and immediately frozen. Sections were then stained with PE-labelled anti-NK1.1 antibody (PharMingen) and examined in a fluorescence microscope. Similarly, spleens obtained on day 16 from NK cell-depleted mice and controls were frozen, then stained with PE-labelled anti-NK1.1 and FITC-labelled anti-5E6 (PharMingen).

IL-6 assay

The cell line B9, which is dependent on IL-6 for growth, has been previously described [19]. B9 cells were harvested from tissue culture flasks, seeded into microtitre plates (Nunc, Roskilde, Denmark) at the concentration of 5000 cells per well, and cultured in Iscove's medium supplemented with 5 × 10−5 m 2-mercaptoethanol (2-ME), 5% fetal calf serum (FCS; Seralab, Sussex, UK), penicillin 100 U/ml, and streptomycin 100 μg/ml, and serum samples were added. 3H-thymidine was added after 68 h of culturing, and the cells were harvested 4 h later. The samples were tested in two-fold dilutions and compared with a recombinant mouse IL-6 standard (Genzyme, Cambridge, MA). B9 cells were previously shown not to react with several recombinant cytokines, including IL-1α, IL-1β, IL-2, IL-3, IL-5, granulocyte-macrophage colony-stimulating factor (GM-CSF), tumour necrosis factor-alpha (TNF-α), and IFN-γ.

IFN-γ assay

Serum IFN-γ was assayed using an ELISA system. Polystyrene flat-bottomed plates were coated overnight at 4°C with 2 μg/ml of purified anti-mouse IFN-γ (PharMingen). The coated plates were washed with PBS–0.05% Tween 20, and saturated with 1% BSA for 1 h at 37°C. After washing, the plates were incubated for 2 h at 37°C with different dilutions of samples. Biotinylated anti-mouse IFN-γ (PharMingen) (2 μg/ml) was added and incubated overnight at 4°C. The plates were washed again and incubated with extravidin alkaline phosphatase (Sigma, St Louis, MO). The enzyme substrate was then added and optical density (OD) was determined in a Titertek Multiscan photometer (Flow Labs, McLean, VA). The concentration of IFN-γ was calculated using a standard curve based on known quantities of recombinant mouse IFN-γ (Genzyme).


Serum levels of total IgG1, IgG3 and IgM were measured by the radial immunodiffusion technique as previously described [7, 20]. Antisera and immunoglobulin standards specific for IgG1, IgG3, and IgM were purchased from Sigma.

Anti-TSST-1 antibodies

Serum levels of IgG and IgM antibodies to TSST-1 were estimated by an ELISA using 0.5 μg/ml of highly purified TSST-1 (Toxin Technology, Sarasota, FL) as a solid-phase coating as previously described [7].

In vitro stimulation of spleen cells for IFN-γ production

Mouse spleen mononuclear cells (MNC; 1 × 106/ml) were cultured in Iscove's complete medium (10% FCS, 2-ME 5 × 10−5 m, 2 mml-glutamine, and 50 μg/ml of gentamycin) and stimulated with 10 μg/ml of highly purified TSST-1 (Toxin Technology), or 107 cells/ml of formalin-killed S. aureus strain LS-1 or LPS 10 μg/ml (Sigma). The cultures were maintained on 24-well plates (Nunc) at 37°C in 5% CO2 and 95% humidity. The supernatants were collected after 72 h for detection of IFN-γ with the ELISA method described above.

Analysis of the phagocytic activity of leucocytes

Freshly obtained heparinized whole blood was vortexed and aliquoted on the bottom of a 5-ml tube. Pre-cooled FITC-conjugated bacteria were added (1 × 109/ml) and incubated for 20 min at 37°C. Ice-cold quenching solution was then added to remove cell surface-bound FITC. The erythrocytes were lysed and leucocyte membranes solubilized to permit detection of intracellular FITC-bacterial deposits (Pharma, Heidelberg, Germany). Measurements were performed with FACScan (Becton Dickinson).

Olive oil inflammation

In order to assess the capacity for granulocyte-mediated, but T lymphocyte-independent inflammation in animals depleted of NK cells as indicated above, the olive oil inflammation assay was used [21]. In ether anaesthesia, right hind footpad thickness was registered by an Oditest spring caliper (Kröplin, Schlüchtern, Germany) and 30 μl of olive oil (Apoteksbolaget, Göteborg, Sweden) was injected intradermally in the footpad. After 24 h, footpad swelling was measured.

DTH to oxazolone

To assess T cell-dependent inflammation, mice depleted of NK cells were sensitized by epicutaneous administration of 100 μl oxazolone 30 mg/ml in ethanol:acetone (3:1) solution. After 7 days the right ear was challenged with topical application of 20 μl oxazolone 10 mg/ml in olive oil [22]. Ear thickness was measured prior to and 24 h after challenge with an Oditest spring caliper.

Assay for antibodies to oxazolone

An ELISA method was applied to measure serum antibody activity against oxazolone as previously described [22]. Oxazolone coupled to dog serum albumin (Sigma) 10 μg/well was used to coat wells, followed by addition of serially diluted sera. To quantify antibodies bound to oxazolone, wells were first incubated with heavy chain-specific goat anti-mouse IgG or IgM antibodies (Cappel Labs, Cochraneville, PA), then with horseradish peroxidase (HRP)-conjugated rabbit anti-goat antibodies (Dakopatts, Glostrup, Denmark), after which an enzyme substrate was added. Absorbance levels were recorded in a Titertek scan at 405 nm. OD data at serum dilutions 1:10 000 and 1:1000 were chosen for IgG and IgM, respectively.

Statistical analysis

The differences between parametric and non-parametric values in all treatment groups were tested for significance by use of the two-tailed Student's t-test and the Wilcoxon signed rank test, respectively. Differences between groups regarding the occurrence of arthritis were analysed by the χ2 test. Results are presented as means ± s.d. P ≤ 0.05 was considered statistically significant.


Depletion of NK cells

Spleen cells from PK136- and O1C5.B2-injected C57Bl/6 mice were recovered and the efficacy of the NK cell depletion procedure was analysed.

Flow cytometry analysis revealed that in mice injected with a single dose of PK136 MoAb, the population of NK1.1+ spleen cells was reduced from 5.5% to 0.6%. Similarly, NK/5E6+ cells were reduced from 5.7% to 2.2% and NK/2B4+ cells from 4.9% to 3.4%.

The functional status of NK cells was tested in an in vitro cytotoxicity assay. At an effector:target ratio of 200:1 specific lysis was 3.3% in the PK136 group but 21% in controls, and at a 50:1 ratio the figures were 0.8% and 11%, respectively (P < 0.0001 at both effector:target ratios). Standard error was always < 4%.

NK1.1+ T cells

By flow cytometry, spleen cells from naive C57Bl/6 mice were analysed for co-expression of the NK1.1 epitope and T cell markers. The frequency of CD4+NK1.1+ splenocytes was ≤ 0.5%; CD8+NK1.1+ were ≤0.5% and CD3+NK1.1+ cells were ≤1%.

Effect of NK cell depletion on the clinical course of arthritis

C57Bl/6 mice were continuously depleted of NK cells by bi-weekly injections with anti-NK1.1 antibodies. Three days after start of injections mice were inoculated intravenously with 2 × 107 CFU of S. aureus LS-1. In one pilot study followed by an experiment including 10 mice per group, NK cell-depleted mice developed more frequent and severe arthritis. The same outcome was seen in a third study including 19 mice depleted of NK cells and 17 controls receiving an irrelevant herpes antibody. As shown in Fig. 1, weight loss during the first week of infection was slightly more pronounced in the NK cell-depleted mice compared with controls. One week after bacterial inoculation, one mouse died in each group. At sacrifice, spleen weight was 179 ± 48 mg in the depleted group compared with 201 ± 88 mg in controls, equal to 0.93% and 1.03% of body weight (NS). The appearance of arthritis was similar during the first 3 days, but at day 6 post-inoculation the number of arthritic mice was significantly higher in the NK cell-depleted group (Fig. 2a). The difference in prevalence of arthritis was most pronounced on day 9, when one out of 17 control mice had arthritis compared with 9/19 in the depletion group (P = 0.01). When the number of arthritic joints and their severity were used to construct an arthritic index of all mice in each group, the difference between the groups became statistically significant throughout the study (Fig. 2b). On day 9, the mean arthritic index of the nine arthritic mice in the depletion group was 1.7, while the single control mice had an index of 1.

Fig. 1
Female 10-week-old C57Bl/6 mice were administered MoAb PK136 (NK cell-depleted, [filled square]) or MoAb O1C5.B2 (controls, ○), and inoculated intravenously with 2 × 107 colony-forming units (CFU)/mouse of Staphylococcus aureus LS-1 on day 0. ...
Fig. 2
The prevalence of arthritis (a) and severity of arthritis (b) in female 10-week-old C57Bl/6 inoculated intravenously with 2 × 107 colony-forming units (CFU)/mouse of Staphylococcus aureus LS-1 are recorded during a period of 16 days. Before and ...

Recovery of staphylococci in joints and kidneys

In the NK cell-depleted group growth of S. aureus was found in the joints of three mice and in the kidneys of six out of 19 mice. In the controls, the corresponding figures were none for joints and seven for kidneys out of 17 mice.

Histopathological findings

No differences were observed as to histopathological signs of arthritis when joints obtained at sacrifice (day 16) were examined. Thus, in both groups most joints had thickening of synovial layer, whereas half of the examined joints showed pannus and erosions of bone and cartilage. In kidney sections, NK cell-depleted mice had a tendency to heavier mononuclear cell infiltrates of all categories than the controls, though the difference was not statistically significant (Table 1).

Table 1
Histopathological findings in kidney sections from female C57Bl/6 mice obtained at sacrifice 16 days after inoculation with Staphylococcus aureus LS-1

Immunofluorescence of joints and spleens

In the joints of control mice, only very few cells could be seen which stained positively with antibody to the NK1.1 antigen. In the spleens of control mice killed on day 16, NK1.1+ cells and 5E6+ cells could readily be seen, comprising approx. 10% and 5%, respectively, of splenocytes. In the NK cell-depleted group, only a few scattered positively stained cells were seen.

Serological findings

Serum levels of immunoglobulins 16 days post-inoculation were found to be similar in both treatment groups (data not shown). However, with regard to TSST-1-specific responses, there was a tendency for NK cell-depleted mice to display higher levels compared with controls of both IgG, 35.6 ± 30 versus 21.9 ± 14 (ELISA units), respectively, and IgM antibodies 40.7 ± 26 versus 28.7 ± 22. IL-6 levels were similar in the NK cell-depleted group compared with the controls (data not shown). Serum IFN-γ was detectable in < 10% of mice.

In vitro IFN-γ production by splenocytes

The spleen cells from non-infected C57Bl/6 mice NK cell-depleted by a single injection with MoAb PK136 were stimulated with formalin-killed S. aureus LS-1, TSST-1 and LPS for evaluation of IFN-γ production and compared with control mice. IFN-γ production was reduced in NK cell-depleted mice in response to LPS, whereas exposure to whole bacteria and TSST-1 resulted in a similarly high IFN-γ production in both study groups (Table 2).

Table 2
IFN-γ production (units ± s.d.) from spleen cells of 10-week-old non-infected female C57Bl/6 mice after in vitro stimulation with formalin-killed Staphylococcus aureus, toxic shock syndrome toxin-1 (TSST-1) and lipopolysaccharide (LPS), ...

In vitro phagocytosis

After a single injection of NK cell-depleting MoAb or control MoAB to naive C57Bl/6 mice, no differences in the capacity of peripheral blood phagocytic cells to phagocytose bacteria could be detected (data not shown).

Olive oil inflammation

Non-infected 6-week-old male C57Bl/6 mice were administered PK136 or OIC5.B2 intraperitoneally twice a week. Twenty-four hours after the first injection of MoAb, mice were challenged with olive oil intradermally in the right hind footpad. Both groups displayed a similar degree of inflammatory response, as measured by increase of footpad thickness (data not shown).

OXA-DTH, OXA antibodies and immunoglobulin levels

The DTH response to oxazolone was not affected by NK cell depletion (data not shown). In contrast, the OD levels of specific antibodies to oxazolone of both IgG and IgM isotypes were significantly increased in the NK cell-depleted group compared with controls (IgG: 0.545 ± 0.153 versus 0.067 ± 0.085, P < 0.001; IgM: 0.685 ± 0.109 versus 0.223 ± 0.060, P < 0.001). Total immunoglobulin levels were not affected by the NK cell depletion procedure.


In the present study of haematogenously induced S. aureus infection, we show that depletion of NK1.1+ cells increased the frequency and severity of arthritis. However, at sacrifice on day 16 after bacterial inoculation, both groups displayed equally extensive proliferative and erosive changes in joints at histopathological examination. Weight loss and spleen weight of mice at sacrifice did not differ substantially between the groups, which indicates that systemic disease was similarly severe in both groups. NK cells have previously been found in the synovium and in synovial fluid of patients with autoimmune inflammatory joint diseases [23, 24] as well as in osteoarthritis [25], though their role in pathogenesis is not established. In experimental Lyme disease in C57Bl/6 mice, NK cell depletion did not affect development of arthritis [26]. Thus, we are the first to ascribe a role to NK cells in the development of arthritis. This occurred without the presence of NK cells in the joints by the end of the experiment.

At sacrifice 16 days after inoculation with S. aureus we did not find a significantly increased presence of bacteria systemically or in joints in NK cell-depleted animals. A previous study has indicated that human NK cells can phagocytose S. aureus bacteria and display bactericidal activity in vitro [27]. However, in our study NK cell depletion of non-infected animals did not impede phagocytosis of bacteria by peripheral blood mononuclear cells (PBMC) when assayed in vitro. Thus, our present study did not give any clear evidence that NK cells contribute to the control of the non-erosive inflammatory component of the arthritis manifestation through substantial effect on bacterial load. However, our data do not totally exclude that NK cells could reduce the early spread of bacteria resulting in suppression of arthritis.

Hence, to investigate what mechanism governed the differential outcome of arthritis, we made in vitro studies as well as serological analyses of blood samples obtained at sacrifice to examine the effects of NK cell depletion on immune function.

In vitro, stimulation of spleen cells from non-infected mice for 3 days with TSST-1 and formalin-killed bacteria yielded similar levels of IFN-γ in supernatants from NK cell-depleted animals as from controls. This latter finding is in contrast to the findings in human PBMC of Yoshihara [5], where S. aureus stimulation in vitro induced IFN-γ production exclusively from cells of NK phenotype. Similarly to the findings of Heremans [2], IFN-γ production in response to LPS was completely abolished by anti-NK1.1 antibody. Unfortunately, with our ELISA method we could detect IFN-γ only in a minority of serum samples. Therefore we cannot in the present study delineate the in vivo importance of NK cell IFN-γ production for S. aureus-mediated septic arthritis.

Earlier studies have demonstrated the pivotal role of CD4+ T cells in the progression of septic arthritis [8, 9]. In the light of a recent report demonstrating that NK cells can play an important role in down-regulating Th1-mediated experimental colitis by controlling the responses of effector T cells to gut bacteria [28], we analysed the effect of in vivo NK cell depletion on T cell-dependent inflammation. The results showed no changes in cutaneous DTH reactivity. In contrast, depletion of NK cells significantly increased the antigen-specific but not the polyclonal antibody responses. These findings do not totally rule out an interaction between NK cells and T cells in septic arthritis, but indicate a possible interaction between NK cells and subsets of B lymphocytes. In mice inoculated with S. aureus we found similarly elevated levels of serum IL-6 and of serum IgG1, IgG3 and IgM levels in both NK cell-depleted mice and controls. However, there was a tendency to increased serum IgG and IgM antibody levels specific for TSST-1 in NK cell-depleted animals. In agreement with a suppressive effect of NK cells on antigen-specific antibody production, we found in another in vivo experiment that antibody titres to oxazolone were significantly higher in non-infected NK cell-depleted C57Bl/6 mice. In previous experiments investigating the importance of B cell function in S. aureus-induced septic arthritis, Zhao et al. [10] have shown that expression of arthritis was less severe in Xid mice, which display disturbances of B cell functions [29], than in healthy controls. As these mice differed from the controls not only by showing lower levels of antibodies to S. aureus-specific antigens, but also in the total levels of immunoglobulin subclasses and as well as in spleen cytokine mRNA expression, it is not possible to point out a single B cell-dependent arthritis accelerating factor present both in the Xid and in our NK cell depletion model.

In conclusion, we show that NK cell depletion promotes the development of the non-erosive arthritic component of systemic infection with S. aureus LS-1. This occurs without evidence of significant late phase decrease of bacterial growth in animals. We thus believe that the arthritogenic effect of NK cell depletion mainly occurs through loss of NK cell-mediated interaction with the host inflammatory response rather than through loss of antibacterial activity.


We thank Mrs Lena Svensson, Mrs Ing-Marie Nilsson and Mrs Margareta Verdrengh for their expert technical assistance. This study was supported by grants from the Göteborg Medical Society, the Swedish Society of Medicine, the Swedish Association against Rheumatism, the Nanna Svartz' foundation, the King Gustav V's 80 years foundation, the Börje Dahlin foundation, the Medical Faculty of the University of Göteborg (LUA) and the Swedish Medical Research Council.


1. Scharton TM, Scott P. Natural killer cells are a source of interferon gamma that drives differentiation of CD4+ T cell subsets and induces early resistance to Leishmania major in mice. J Exp Med. 1993;178:567–77. [PMC free article] [PubMed]
2. Heremans H, Dillen C, van Damme J, et al. Essential role for natural killer cells in the lethal lipopolysaccharide-induced Shwartzman-like reaction in mice. Eur J Immunol. 1994;24:1155–60. [PubMed]
3. Ozmen L, Pericin M, Hakimi J, et al. Interleukin 12, interferon gamma, and tumor necrosis factor alpha are the key cytokines of the generalized Shwartzman reaction. J Exp Med. 1994;180:907–15. [PMC free article] [PubMed]
4. D'Orazio JA, Burke GW, Stein SJ. Staphylococcal enterotoxin B activates purified NK cells to secrete IFN-gamma but requires T lymphocytes to augment NK cytotoxicity. J Immunol. 1995;154:1014–23. [PubMed]
5. Yoshihara R, Shiozawa S, Fujita T, et al. Gamma interferon is produced by human natural killer cells but not T cells during Staphylococcus aureus stimulation. Infect Immun. 1993;61:3117–22. [PMC free article] [PubMed]
6. Bremell T, Lange S, Svensson L, et al. Outbreak of spontaneous staphylococcal arthritis and osteitis in mice. Arthritis Rheum. 1990;33:1739–44. [PubMed]
7. Bremell T, Abdelnour A, Tarkowski A. Histopathological and serological progression of experimental Staphylococcus aureus arthritis. Infect Immun. 1992;60:2976–85. [PMC free article] [PubMed]
8. Abdelnour A, Bremell T, Holmdahl R, et al. Clonal expansion of T lymphocytes causes arthritis and mortality in mice infected with toxic shock syndrome toxin-1 producing staphylococci. Eur J Immunol. 1994;24:1161–6. [PubMed]
9. Abdelnour A, Bremell T, Holmdahl R, et al. Role of T lymphocytes in experimental Staphylococcus aureus arthritis. Scand J Immunol. 1994;39:403–8. [PubMed]
10. Zhao Y, Abdelnour A, Holmdahl R, et al. Mice with the Xid B cell defect are less susceptible to Staphylococcus aureus-induced arthritis. J Immunol. 1995;155:2067–76. [PubMed]
11. Zhao YX, Nilsson IM, Tarkowski A. The dual role of IFNγ in experimental Staphylococcus aureus septicemia versus arthritis. Immunol. 1998;93:80–85. [PMC free article] [PubMed]
12. D'Orazio JA, Stein SJ. Human natural killer (NK) cells present staphylococcal enterotoxin B (SEB) to T lymphocytes. Clin Exp Immunol. 1996;104:366–73. [PMC free article] [PubMed]
13. Wilder JA, Koh CY, Yuan D. The role of NK cells during in vivo antigen-specific antibody responses. J Immunol. 1996;156:146–52. [PubMed]
14. Takeda K, Dennert G. The development of autoimmunity in C57Bl/6 lpr mice correlates with disappearance of natural killer type 1-positive cells: evidence for their suppressive action on bone marrow stem cell proliferation, B cell immunoglobulin secretion, and autoimmune symptoms. J Exp Med. 1993;177:155–64. [PMC free article] [PubMed]
15. Wigzell H, Ramstedt U. Natural killer cells. In: Weir DM, editor. Handbook of experimental immunology: III Cellular immunology. Vol. 60. Oxford: Blackwell Scientific Publications; 1986. pp. 1–11.
16. Abdelnour A, Arvidson S, Bremell T, et al. The accessory gene regulator (agr) controls Staphylococcus aureus virulence in a murine arthritis model. Infect Immun. 1993;61:3879–85. [PMC free article] [PubMed]
17. Goldenberg DL, Cohen AS. Synovial membrane histopathology in the different diagnosis of rheumatoid arthritis, gout, pseudogout, systemic lupus erythematosus, infectious arthritis and degenerative joint disease. Med. 1978;57:239–52. [PubMed]
18. Verba V, Tarkowski A. Participation of Vβ4+-, Vβ7+-, and Vβ11+-T lymphocytes in hematogenously acquired Staphylococcus aureus nephritis. Scand J Immunol. 1996;44:261–6. [PubMed]
19. Aarden LA, Groot ERD, Shaap OL, et al. Production of hybridoma growth factor by human monocytes. Eur J Immunol. 1987;17:1411–6. [PubMed]
20. Mancini G, Carbonara AO, Heremans JF. Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochem. 1965;2:235–54. [PubMed]
21. Josefsson E, Tarkowski A, Carlsten H. Anti-inflammatory properties of estrogen. Cell Immunol. 1992;142:67–78. [PubMed]
22. Carlsten H, Holmdahl R, Tarkowski A, et al. Oestradiol suppression of delayed-type hypersensitivity in autoimmune (NZB/NZW) F1 mice is a trait inherited from the healthy NZW parental strain. Immunol. 1989;67:205–9. [PMC free article] [PubMed]
23. Armstrong RD, Panayi GS. Natural killer cell activity in inflammatory joint disease. Clin Rheumatol. 1983;2:243–9. [PubMed]
24. McQueen FM, Skinner MA, Krissansen GW, et al. Natural killer cell function and expression of beta 7 integrin in psoriatic arthritis. J Rheumatol. 1994;21:2266–73. [PubMed]
25. Kummer JA, Tak PP, Brinkman BM, et al. Expression of granzymes A and B in synovial tissue from patients with rheumatoid arthritis and osteoarthritis. Clin Immunol Immunopathol. 1994;73:88–95. [PubMed]
26. Barthold SW, de Souza M. Exacerbation of Lyme arthritis in beige mice. J Infectious Dis. 1995;172:778–84. [PubMed]
27. Abo T, Sugawara S, Amenomori A, et al. Selective phagocytosis of gram-positive bacteria and interleukin 1-like factor production by a subpopulation of large granular lymphocytes. J Immunol. 1986;136:3189–97. [PubMed]
28. Fort MM, Leach MW, Rennick DM. A role for NK cells as regulators of CD4+ T cells in a transfer model of colitis. J Immunol. 1998;161:3256–61. [PubMed]
29. Kaplan RB, Quintans J. Phosphorylcholine-specific helper T cells in mice with an X-linked defect of antibody production to the same hapten. J Exp Med. 1979;149:267–72. [PMC free article] [PubMed]

Articles from Clinical and Experimental Immunology are provided here courtesy of British Society for Immunology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...