The majority of discoveries made during this project have been published in well-known scientific journals (
https://insight.jci.org/articles/view/125191#SEC2(s’ouvre dans une nouvelle fenêtre) and
https://journals.sagepub.com/doi/full/10.1177/0022034519898144(s’ouvre dans une nouvelle fenêtre)) and the main findings are summarised below:
I) Using a gold-standard mouse model for inflammatory arthritis, I found that infection with P. gingivalis, but not with the benign bacterium Bacteroides thetaiotaomicron, caused a significant worsening of joint inflammation. Moreover, P. gingivalis caused a breakdown of the normal gut-barrier function, resulting in increased invasion of gut bacteria into gut tissue and from there several other organs. Importantly, I observed that this was only the case in arthritic animals, not if non-arthritic mice were infected with the bacterium. This suggested that arthritis may change the gut, thus making it vulnerable to the harmful actions of P. gingivalis. Indeed, when I next compared the function of the gut-barrier between healthy and arthritic mice, in the absence of any P. gingivalis, I found that during arthritis, amounts of pro-resolving lipid mediators in the gut and particularly the mediator RvD5n-3 DPA were significantly lower. In addition, several components that make up the gut-barrier were deregulated, including the production of the protective mucus layer and immune cells that normally guard the barrier against invading bacteria.
II) Due to changes in the way patients are referred to our clinic, sufficient stool samples from RA patients could not be collected quickly enough for me to conduct the planned experiments during this fellowship. In order to investigate the relevance of P. gingivalis in humans, however, I used a fraction of the blood called serum, from RA patients and healthy human volunteers to investigate whether specific modifications that occur on some of the proteins from which P. gingivalis is made but not at all in other known bacteria, may trigger an antibody immune response that could then turn the immune system to wrongfully attack the patients’ tissues. The latter is known to happen in early RA. Although this is a popular hypothesis, there are conflicting opinions and reports as to its accuracy. My findings from this project indicate that these modifications on P. gingivalis structures are not likely to act as such triggers of autoimmune antibody responses in early RA.
III) As described above, I have discovered that during arthritis, amounts of the lipid mediator RvD5n-3 DPA are significantly decreased in the gut. This is associated with a weakening of the gut-barrier, making it vulnerable to P. gingivalis, which in turn leads to worsening of joint inflammation. I next investigated whether giving back this lipid mediator would have protective effects. Indeed, treatment of arthritic mice that have been infected with P. gingivalis with RvD5n-3 DPA reversed the dysregulation of the different components of the gut-barrier and preserved its function, and prevented the worsening of joint disease despite the infection with P. gingivalis (see attached images).
Experimental work with P. copri is ongoing, thus findings at this stage are confidential, but has already enabled me to secure a 3-year lectureship position and is expected to result in a high-profile publication. Collection of stool samples from RA patients continues and will allow me to carry out the planned work, in the near future, as part of my next grant/fellowship.
In summary, this project has discovered that inflammatory arthritis affects not only the joints but also, by causing a decrease in intestinal amounts of pro-resolving lipid mediator RvD5n-3 DPA, leads to a weakening of the gut-barrier. This makes the barrier vulnerable to the pathogenic actions of P. gingivalis, which in turn causes a further loss of gut-barrier function, invasion of gut-bacteria into the tissues and significant worsening of joint disease. Administration of RvD5n-3 DPA protected the barrier and prevented the worsening of inflammatory arthritis by the bacterium.
Figure Legends for attached images:
Figure A: Image of mouse intestines. In arthritic mice without P. gingivalis (left image), bacteria (green) did not invade into mucus layer (black, marked by dotted lines) or host tissues (blue). If arthritic mice were infected with P. gingivalis (middle image), bacteria invaded into host tissues (arrows). This was prevented by treatment with RvD5n-3 DPA (right image).
Figure B: Joint disease severity over time. P. gingivalis alone (red) significantly increased severity. Treatment with RvD5n-3 DPA (blue) prevents this increase in severity.
