Published online before print February 24, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Rheumatology and Inflammation Research, University of Göteborg, Sweden; and
Department of Clinical Chemistry, Sahlgrenska University Hospital, Göteborg, Sweden
1Correspondence: Department of Rheumatology and Inflammation Research, University of Göteborg, Guldhedsgatan 10A, 41346 Göteborg, Sweden. E-mail: vincent.collins{at}rheuma.gu.se
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B stimulation resulted in the production of tumor necrosis factor 
, a potent, arthritogenic factor. Finally, extracellular mtDNA was detected in the synovial fluids of rheumatoid arthritis patients but not of control subjects. We conclude that endogenous mtDNA displays inflammatogenic properties as a result of its content of unmethylated CpG motifs and oxidatively damaged adducts.
Key Words: inflammation • rheumatoid arthritis • monocytes/macrophages
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
As mtDNA is a target of free radical damage as a result of its proximity to the electron transfer chain reactions and as free radical-mediated mtDNA damage is repaired with low efficiency, we investigated the contribution of oxidatively damaged DNA to arthritogenicity. Indeed, an oligonucleotide that lacked CpG motifs but that contained a single 8-hydroxy-2'-deoxyguanosine (8-oxodG) residue was inflammatogenic in vivo. In contrast, an oligodeoxynucleotide (ODN) with exactly the same sequence, except that it lacked the oxidized residue, was totally inert in vivo.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
DNA injections in mice
mtDNA, nDNA, and ODN samples (20 µl vol containing 5 µg DNA or 10 nmol ODN) were injected i.a. in the knees of female, 6- to 8-week-old mice [BALB/c mice from ALAB, Stockholm, Sweden; CB17 and severe combined immunodeficiency (SCID) mice from M&B, Bomholtvej, Denmark; NMRI mice from B&K, Universal AB, Sollentuna, Sweden]. All animals were housed in the animal facility of the Department of Rheumatology and Inflammation Research, University of Göteborg (Sweden), under standard conditions. The i.a.-injected mice were killed after 3 or 14 days, and the joints were removed for histopathology or immunohistochemistry.
In vivo depletion of immune cells
Female NMRI mice were depleted of peripheral blood monocytes by treatment with etoposide (Bristol Myers Squibb AB, Bromma, Sweden), which was injected subcutaneously (s.c.) at a dose of 12.5 mg/kg daily, starting 2 days before DNA injection, and continued for the course of the experiment (3 days). This treatment selectively depletes monocytes, as previously demonstrated [9
]. The control mice received s.c. injections of PBS. Female BALB/c mice (6–8 weeks old) were depleted of granulocytes by intraperitoneal (i.p.) pretreatment with 1 mg of the monoclonal antibody (mAb) RB6-8C5 [10
] 2 h before the i.a. injections of mtDNA. Control mice were pretreated with an immunoglobulin G (IgG) rat anti-ovalbumin (anti-OVA) mAb.
Histopathology and immunohistochemistry of mouse joints
The histopathological and immunohistochemical examinations of mouse joints were performed as described previously [4
]. The severity of arthritis in knee-joint sections was evaluated by a blinded observer and assessed on a scale of 0–3, where 0 = no signs of inflammation, 1 = mild inflammation characterized by hyperplasia of the synovial lining layer, and 2–3 = increasing levels of inflammation characterized by influx of inflammatory cells into the synovial tissue. For reference purposes, the joint shown in Fig. 1a
rates a score of 2, and that seen in Figure 1b
is scored as 0.
 View larger version (102K): [in a new window] |
Figure 1. Histopathological and immunohistochemical analysis of MDIA. (a) Histopathology of a knee joint of an NMRI mouse 3 days after i.a. injection of 5 µg murine mtDNA originating from muscle tissue. Infiltration of mononuclear cells is apparent. (b) Histopathology of a knee joint of an NMRI mouse 3 days after i.a. injection of 5 µg nDNA. Joint tissues appear normal. (c) Immunohistochemistry of a knee joint that was treated as above (a), showing synovial expansion of cells expressing Mac1 (brown). JC, Joint cavity; V, blood vessel; ST, synovial tissue; arrows, inflammatory cells in the synovium.
|
BB was blocked in female NMRI mice by i.p. administration of 900 µg antisense ODN, 5'-GAAACAGATCGTCCATGGT-3', or the mismatched control ODN, 5'-GAAACAGATCGTCTATGGT-3', 2 days before i.a. injection of mtDNA. These phosphorothioate-modified ODNs were synthesized by CyberGene AB (Huddinge, Sweden). No arthritis was observed in histopathological sections of the joints of mice injected i.p. with the antisense or the mismatched ODN alone at a dose of 900 µg per mouse.
Proliferation and tumor necrosis factor (TNF-) measurements
Naïve, murine spleen cells from four individual NMRI mice were incubated at a concentration of 2 x 106 cells/ml in Iscove’s medium containing 10% fetal calf serum. The cells were cultured at 37°C in 5% CO2 in the presence of 10 µg/ml mtDNA, 10 µg/ml nDNA, or medium alone for 24 h (TNF- assay) or for 68 h (proliferation assay). The 24-h culture supernatants were assayed for TNF- production by enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN). To study proliferative responses, the 68-h cultures were pulsed with 3H-thymidine for 4 h, the cells were harvested onto filters, and 3H-thymidine incorporation was measured in a ß-counter.
Analysis of human synovial fluids
PCR analysis for the presence of free (extracellular) mtDNA was performed on SF samples aspirated from 54 patients presenting with RA at Sahlgrenska University Hospital (Göteborg, Sweden). The control SF samples (n=17) were collected at autopsy from control subjects (none of whom showed any signs of RA or osteoarthritis). The samples were centrifuged at 500 g for 10 min and then at 2000 g for 10 min, and the supernatants were boiled for 10 min. Following centrifugation in a microfuge (11,000 g for 15 min at 4°C), the supernatants were analyzed by PCR using the human mtDNA-specific ODN primers, 5'-GCTCTCCATGCATTTGGTAT-3' (forward primer) and 5'-TTGTATTGATGAGATTAGTA-3' (reverse primer). The reaction mixtures contained 4 µM each forward and reverse primers in a total volume of 25 µl, which consisted of 2.5 µl-extracted SF in 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 4.0 mM MgCl2, 0.2 mM deoxy-unspecified nucleoside 5'-triphosphates (deoxy-adenosine 5'-triphosphate, -cytidine 5'-triphosphate, -guanosine 5'-triphosphate, and -thymidine 5'-triphosphate), and 1.25 U Taq DNA polymerase (AmpliTaq Gold, Perkin-Elmer, Branchburg, NJ). All SF samples were subjected to 45 cycles of PCR amplification, each consisting of a 12-min hot start at 94°C, 30 s denaturation at 94°C, 30 s annealing at 60°C, and extension at 72°C for 1 min. Samples that showed the 453-bp PCR product on silver-stained sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels were considered to be positive for the presence of mtDNA.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (16K): [in a new window] |
Figure 2. Effects of i.a. injection of whole mtDNA and mtDNA fragments on arthritis development in mice. Incidence of arthritis in knee joints of NMRI mice injected i.a. with PBS (n=10), 5 µg mouse liver nDNA (n=13), 5 µg mouse muscle mtDNA (n=43), or 5 µg each of the PCR-amplified fragments of the human mitochondrial genome: PCR-I (n=5) or PCR-II (n=5). Mice were killed 3 days after DNA injection, and knee joints were assessed histopathologically for severity of synovitis. Statistically significant differences (***, P<0.001) between nDNA and mtDNA and PCR-I and PCR-II were determined by Fisher’s exact test.
|
B is important in CpG-mediated immunostimulation [13
]. We used an antisense ODN to block NF-B activity [14
] and found that it reduced MDIA frequency by 40% and the severity of arthritis by 59% (Fig. 3
) compared with animals treated with a mismatched ODN. One of the main products triggered by stimulation of NF-B is the proinflammatory and proarthritogenic cytokine TNF- [15
]. When mtDNA was added to splenocyte cultures, which were prepared from four individual mice, TNF- production increased by an average of 88% over the level induced by nDNA treatment (P=0.03; Fig. 4
). Conversely, there were no significant differences between the TNF- levels in the supernatants of cells that were exposed to nDNA or medium alone. Furthermore, treatment with mtDNA induced a one- to 2.5-fold increase in splenocyte proliferation compared with treatment with nDNA (P=0.003). |
View this table: [in a new window] |
Table 1. The Role of Immunocompetent Cells in mtDNA-Induced Arthritis
|
 View larger version (14K): [in a new window] |
Figure 3. Antisense therapy directed against NF-B affects the severity of MDIA. NMRI mice were injected i.p. with NF-B antisense or mismatched ODN (n=12 per group) 2 days before i.a. injection in the knee joint with mouse muscle mtDNA (5 µg). Mice were killed 3 days after DNA injection, and knee joints were examined histopathologically for scoring of arthritis severity. Statistically significant differences (*, P<0.05) in arthritic index were determined by the unpaired Student’s t-test.
|
 View larger version (9K): [in a new window] |
Figure 4. Induction of TNF- production in mouse splenocyte cultures by the addition of mtDNA. The TNF- levels (pg/ml) were measured by ELISA in the 24-h culture supernatants of murine splenocytes from four individual mice, where the cells were untreated ( ) or stimulated with 10 µg/ml mtDNA (•) or 10 µg/ml nDNA ( ).
|
i.a. injection of an ODN containing an oxidized nucleotide
To study the impact of oxidative damage on arthritogenic potential, we compared the inflammatogenicity of two synthetic 20-mer ODN, one with a central deoxyguanosine (GpC–ODN) and the other with an 8-oxodG residue substituted for this deoxyguanosine (oxoGpC–ODN). When injected i.a., the GpC–ODN, as expected, did not trigger inflammation in mouse joints, whereas the oxoGpC–ODN provoked arthritis in 25/27 cases (Fig. 5
). Notably, in a small number of cases, there was severe joint inflammation characterized by pannus formation and/or bone destruction. Therefore, the incorporation of a single oxidatively damaged nucleotide had a dramatic effect on the ability to induce inflammation. As mtDNA usually contains oxidized nucleotides, the 8-oxodG residues in released mtDNA fragments or those excreted during damage repair probably contribute to this type of inflammation.
 View larger version (13K): [in a new window] |
Figure 5. Effects of i.a. injection in mice of ODN that contain or lack a single 8-oxodG residue. NMRI mice were injected i.a. with 10 nmol nonarthritogenic GpC–ODN (n=10) or oxoGpC–ODN (n=27). The mice were killed 3 days later, the knee joints were examined histopathologically, and the incidence of arthritis was determined. The differences in incidence of arthritis between the two groups were statistically significant (***, P<0.001), as assessed using Fisher’s exact test.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B, which regulates the expression of several proinflammatory cytokines, including TNF-, ameliorated MDIA. Macrophages respond to immunostimulatory signals by producing proinflammatory cytokines. Indeed, TNF- production was increased in splenocyte cultures by the addition of mtDNA but not by the addition of nDNA. Therefore, it seems likely that mtDNA is taken up by macrophages that are resident in the joints, which then transduce signals, at least in part, mediated by NF-B, as shown in the present report, thereby inducing the production of proinflammatory mediators. The reduction in terms of frequency and severity of MDIA in granulocyte-depleted mice was less pronounced than the effect of depleting monocytes, which indicates that neutrophils play a less-prominent role in this inflammation. B and T cells do not appear to participate in this condition, based on the finding that MDIA could be induced in SCID and CB17 mice with the same frequency and severity. Furthermore, histochemical staining of sections of immunocompetent NMRI mouse joints that were injected with mtDNA revealed a total absence of CD4- and CD8-expressing T cells.
That mtDNA induces arthritis in the murine model raises several intriguing questions regarding the involvement of self-DNA in inflammation. It can be envisaged that pathologic states resulting in tissue necrosis would lead to the release of nuclear and mtDNA from cells. We found that nDNA extracted from mouse liver cells was in contrast to mtDNA, nonarthritogenic. It is important to identify the components of mtDNA that contribute to the proinflammatory effect. Bacterial DNA and synthetic ODNs, which contain unmethylated CpG motifs, display arthritogenic properties [4
, 5
]. The clinical characteristics of the arthritis induced by mtDNA were very similar to those seen with bacterial DNA [4
], and the mean severity index of arthritis induced in NMRI mice was slightly higher for a phosphodiester-backbone, CpG-containing ODN (1.1±0.6; n=8) than for mtDNA (0.69±0.16; n=8) (data not shown). CpG dinucleotides are under-represented in nDNA and mtDNA compared with bacterial genomes [16
, 17
]. However, it should be remembered that mtDNA is present in multiple copies (100–1000 per cell), and most significantly, the CpG motifs in mtDNA are nonmethylated, a property shared with the bacterial genome. Although both of the PCR-amplified mtDNA fragments induced MDIA following i.a. injection in mice, it remains to be elucidated whether nonmethylated CpG motifs in the mitochondrial genome are recognized by the innate immune system, leading to the type of immune activation seen with bacterial DNA [1
2
3
4
, 18
]. With respect to immune cell recognition, the splenocytes of MyD88 knockout mice did not produce TNF- or show any cell proliferation in response to mtDNA or a CpG-containing ODN, whereas the splenocytes of the congeneic wild-type mice showed TNF- induction and proliferative responses to these DNA species (data not shown). Therefore, MyD88 appears to play a crucial role in transducing the mtDNA-induced signal.
Previously, it was noted that mtDNA preparations contained intact, circular, genomic DNA and short fragments of nDNA and mtDNA, the latter sequences being enriched for adducts, such as 8-oxodG [19 ]. Our purified mtDNA preparations consistently showed high molecular weight circular mtDNA and low molecular weight DNA fragments in agarose gels (data not shown). We compared the arthritogenicity of two ODN, whose only difference was that a single 8-oxodG replaced one deoxyguanosine in the GpC–ODN. Whereas GpC–ODN was noninflammatogenic, the oxoGpC–ODN provoked arthritis. We conclude that oxidatively damaged bases in DNA are major contributors to arthritis development, independent of CpG motifs. Since the presence of 8-oxodG in DNA strands, e.g., as a result of free radical damage, is detrimental to cells, these adducts are normally excised by repair mechanisms and excreted. When oxidative damage to DNA is excessive, repair is not possible, and apoptotic mechanisms are induced. Although nDNA and mtDNA undergo oxidative damage [20 ], it is intriguing to note that mtDNA incurs far higher free radical damage than nDNA [21 ] as a result of its proximity to the respiratory chain, lack of histones, and less-efficient DNA repair systems, and mitochondria are instrumental in signaling cellular apoptosis in response to overwhelming oxidative stress [22 , 23 ]. It is noteworthy that conditions of high oxidative stress have been associated with the rheumatoid joint [24 , 25 ].
We considered whether endogenous DNA could promote destructive processes in the host by the initiation of autoinflammation. Recently, it has been postulated that development of autoimmunity might occur in response to the release of autoantigens during cell death [26 ]. Endogenous mtDNA released from cells undergoing necrosis could activate macrophages to produce proinflammatory cytokines and thereby participate in the arthritic process. Our survey of SF samples from patients suggests that mtDNA is present in the majority of RA joints and is absent from the joints of nonarthritis subjects. Although it may be premature to suggest that the mtDNA detected in these joints is participating in the inflammation, there are strong indications that endogenous, cellular components contribute to inflammatory diseases. Further work is required to adjudge a definitive role for mtDNA in human arthritis.
To date, much effort has been expended in identifying bacterial DNA in inflamed human joints [27 , 28 ], and our recent report shows conclusively that bacterial DNA can induce arthritis in mice [4 ]. In this study, we have clearly shown that mtDNA is immunostimulatory and also capable of inducing arthritis in mice. It seems likely that the presence of cell-free mtDNA exacerbates inflammation by stimulating the production of proinflammatory cytokines, thereby creating a vicious circle of inflammation and cell destruction and accelerating the release of additional endogenous, proinflammatory DNA. In the particular case of arthritis, therapies, such as nuclease or antisense treatments, might be envisaged that minimize or counteract the impact of extracellular mtDNA in joints.
Received July 16, 2003; revised January 11, 2004; accepted January 13, 2004.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-mediated shock Eur. J. Immunol. 27,1671-1679[Medline]B abrogates established experimental colitis in mice Nat. Med. 2,998-1004[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  L. Busconi, C. M. Lau, A. S. Tabor, M. B. Uccellini, Z. Ruhe, S. Akira, G. A. Viglianti, I. R. Rifkin, and A. Marshak-Rothstein DNA and RNA autoantigens as autoadjuvants Innate Immunity, December 1, 2006; 12(6): 379 - 384. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
