Article Text

Download PDFPDF

Mr Outside and Mr Inside: classic and alternative views on the pathogenesis of rheumatoid arthritis
  1. Georg Schett1,
  2. Gary S Firestein2
  1. 1University of Erlangen, Erlangen, Germany
  2. 2Division of Rheumatology, Allergy and Immunology, University of California, San Diego, La Jolla, California, USA
  1. Correspondence to Professor Georg Schett, University of Erlangen, Erlangen D-91054, Germany; georg.schett{at}


Different strategies can be applied to overcome an obstacle: one may either overcome it from the ‘outside’ or break it from the ‘inside’. Similar strategies have been successfully used by two American football players in the 1940s, known as ‘Mr Outside’ and ‘Mr Inside’. In this article, the authors propose similar concepts for initiating joint inflammation. Arthritis may start ‘outside’ as a primary inflammation of the synovial membrane (synovitis), which later spreads over to adjacent structures resulting in penetration into the bone marrow (outside-in concept). Alternatively, arthritis may start ‘inside’, in the bone marrow space, and later encroaches upon the synovium (inside-out concept). The authors discuss these two fundamentally different viewpoints on the pathogenesis of arthritis, each one of which bears attractive explanations for the mystery of the pathogenesis of arthritis.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

American football is something of a cultural phenomenon in the USA and is looked upon with mild amusement by the rest of the world. While contemplating possible alternative explanations for the initiation of rheumatoid arthritis (RA), the authors were struck that a couple of famous American football players provided us with some unexpected clues. In the 1940s, two US college stars, Felix ‘Doc’ Blanchard and Glenn Davis led the army to two national championships. They were national heroes, receiving many honours and were even pictured on the cover of Time magazine. Their success was partly due to their uniquely complementary talents: Blanchard carried the ball straight towards the middle of the defensive line (‘Mr Inside’) while Davis ran around the defence (‘Mr Outside’). How does this relate to inflammatory arthritis? As a parable for American football and these famous athletes, the authors suggest that similar inside-out and outside-in patterns of disease might apply to RA.

Conventional wisdom dictates that rheumatoid synovitis results from the siren song of chemoattractants that produce a gradient to tempt blood inflammatory cell migration into the synovium. Due to the irresistible call of the synovium, they pause in their travels to roll, firmly adhere and then transmigrate into the tissue. In addition to cell ingress, hyperplasia of resident synovial fibroblast-like synoviocytes (FLS) occurs as a result of a combination of proliferation and insufficient apoptosis. FLS contribute to the increasing mass of cells and to the complex tissue pathology that results in damage to cartilage, ligaments, tendons, bone and other support structures. Severely altered synovial biology ultimately leads to structural derangement, dysregulated mechanics and disability.

Whereas the final common pathway in diseases such as RA seems unquestionable, the same is not true for its origin. Is inflammatory arthritis really a primary disease of the synovial membrane, or does it begin elsewhere and then expand to its final destination? For instance, the juxta-articular bone marrow and subchondral bone represents an alternative location for the immune abnormalities that eventually coalesce as synovial inflammation. At first glance this possibility might appear counterintuitive because synovitis is considered synonymous with arthritis. Nevertheless, accumulating evidence suggests that osteitis in the bone marrow can occur early and perhaps even antedate synovial membrane changes.1,,3 The two hypotheses, that is, that arthritis begins outside in the synovial lining (‘outside-in’) or inside in the bone marrow (‘inside-out’), are not mutually exclusive and could even synergise in the development of full-blown disease (figure 1). This novel view of arthritis will be considered below, and could have implications for how the authors define early disease and how they might intervene before irreversible changes occur.

Figure 1

Diagram illustrating the outside-in (arthritis begins outside the joint) and inside-out (arthritis begins in the synovial lining itself) hypotheses.

The outside-in hypothesis

The traditional ‘outside-in’ hypothesis, as exemplified by the speedy Glenn Davis (‘Mr Outside’) who followed the path of least resistance, posits that arthritis originates in the synovial membrane. An altered synovial membrane would thus be the primary pathogenic event, followed by the recruitment of immune cells. Synovial membranes along diarthrodial joints are delicate structures composed of one or two cell layers that are loosely associated and lack a well-defined basement membrane. The initial steps that ‘prepare’ the synovium for the onslaught of inflammatory cells, cell accumulation due to unbalanced proliferation and death, and a synovial immune response are not fully defined. At least two scenarios seem plausible. Each one is a combination of predetermined influences (eg, genetic) that interact with stochastic events to create a synovial environment that is ripe for establishing a widespread synovial inflammatory process.

  1. (1) A triad of environmental factors (eg, repeated activation of innate immunity in the joint), genetic factors (eg, the shared epitope or the myriad of recently described genes) and nascent breakdown of tolerance (eg, anti- citrullinated peptide antibodies) directly leads to changes of the synovial membrane and paves the way for the initiation of arthritis. This concept is attractive and is supported by the observation of an altered cellular composition of the synovial membrane before clinically overt joint inflammation (prearthritis).4 Also, elevated acute phase reactants and even serum cytokines might be reflectors of this early (inflammatory) change of the synovial membrane.5

  2. (2) Rather than a purely ‘joint-centric’ process, initiation might be systemic and related to activities in the central lymphoid tissues or elsewhere. A second tissue-specific ‘hit’ is needed to focus disease to the synovium as the final pathway. The strength of this concept is that it is consistent with the observation that well-defined triggers, such as autoantibodies, are present long before clinically apparent disease emerges.6 Furthermore, smoking is clearly an extra-articular stimulus that can generate citrullinated peptides in tissues far from the joint.7 Individually and even in combination, none of these are sufficient to explain the actual timing of disease onset, making a second hit plausible. One major problem of this concept, however, is that no tissue-specific factor has yet been defined.

As the outside-in hypothesis is based on a ‘synovitis first’ concept, the inflamed synovial membrane would initiate and promote invasion of inflammatory tissue into cartilage, bone and finally into the bone marrow space. This concept is supported by numerous studies, especially performed in RA, suggesting that the inflamed synovial membrane has a highly invasive potential. Matrix enzyme production by FLS, which is enhanced by mononuclear cells, can thus degrade joint cartilage and other non-calcified tissues.8 9 Synovitis also promotes osteoclast differentiation, which allows the resorption of mineralised cartilage and cortical bone necessary to fracture the tiny barrier between the synovial and the marrow compartment.10 This process allows a direct communication of the synovium with the bone marrow. Imaging studies using radiographs, computed tomography and magnetic resonance imaging (MRI) scans as well as histopathological analysis revealed clear signs of cortical breaks, and histopathological studies in particular show direct communication between synovial tissue and the bone marrow space.11 12

The putative role of the synovium in the outside-in hypothesis is supported by the distribution of cortical breaks in RA. They are not randomly found along the articular surfaces, but have a clear predilection for regions close to the insertion sites of the synovial membrane, which backs the outside-in hypothesis.13 Close contact between synovium and bone thus seem to be critical. Also, synovitis of the tendon sheaths might significantly contribute to the localisation of bone destruction. Inflamed tendon sheaths passing by joints can potentially transduce synovitis from tendon to synovium and primarily affect bone near the sheaths.

The inside-out hypothesis

Synovitis is the most intensively studied feature of arthritis, which is mainly because of the relative accessibility of tissue and effusions. Our knowledge on synovitis makes the outside-in hypothesis attractive, although many features are based on assumptions rather than data. We might consider entirely different pathogenic mechanisms, such as the possibility that disease starts in the bone marrow and then migrates to the lining of the joint. We term this hypothesis the inside-out hypothesis, which, like ‘Doc’ Blanchard (‘Mr Inside’), begins deep in the interior of the bone and must migrate through seemingly impregnable obstacles.

Involvement of the bone marrow in RA was recognised over a century ago. Baker14 described bone cysts in the mid 19th century and suggested that they protect joints. Much later, raised intra-articular pressure during synovitis was implicated in causing bone cysts, thus indirectly supporting an outside-in hypothesis.15 Bone cysts were considered pressure-regulated escape mechanisms for the synovium into the marrow space, thereby mediating a protective effect to the joint.16 Whether these cysts actually form before they communicate with the joint space is still a matter of speculation. The introduction of modern imaging techniques has suggested an alternative explanation: MRI studies in the earliest stages of disease reveal bone marrow alterations before a ‘pressure phenomenon’ can occur or before any obvious communication with the synovium is detectable.17

Signal changes in the bone marrow depicted by MRI are based on the enrichment of water and reduction of fat. These ‘inside’ lesions are termed osteitis and they indeed fuel the hypothesis that ‘arthritis’ might even start in the bone marrow, eliciting destruction of the inner cortical bone surface ultimately leading to communication with the synovium. Mesenchymal elements may migrate from the bone marrow through cortical bone pores prior to synovitis and provide the stimulus for the initiation of subsequent joint inflammation. Bone channels communicating between the synovium and the bone marrow might thus in fact allow the transduction of inflammation from the outside to the inside and visa versa. In support of this concept, widening of subchondral bone channels has been observed in the course of experimental arthritis.18

Recently, the main corpus delicti for the inside-out hypothesis, MRI-based osteitis, has been characterised in more detail by histopathological analysis. At sites of positive signals, bone marrow fat is replaced by inflammatory tissue dominated by lymphocytes.11 12 19 Moreover, several studies have shown that these lesions are associated with structural damage of joints and even predict bone erosion in RA patients.1 2 Cytological changes in the bone marrow of RA patients were, in fact, reported over 50 years ago by Duthie et al,20 summarising 18 studies performed between 1938 and 1955. Most of them revealed bone marrow plasmocytosis. In fact, plasmocytosis was more frequent in early disease and was associated with hyperglobulinaemia. Later, mast cell infiltration was described in the bone marrow.21 Moreover, several investigations have described an expansion of mononuclear cells manifesting as nodular lymphoid hyperplasia.11 22 Modern immunohistochemical techniques have shown a preponderance of mature B cells and activated T cells, forming aggregates, suggesting immune activation in the bone marrow of RA patients.23 Some data indeed support a functional change of the bone marrow in RA patients. (1) Bone marrow mesenchymal cells from RA patients support B-cell growth and the survival of T cells.24 25 (2) CD14 monocytes from the bone marrow of RA patients selectively induce the production of IgM rheumatoid factor in vitro.26 (3) Bone marrow CD34 cells have a high potential to form endothelial cells, which may contribute to synovial vascularisation.27 (4) Proinflammatory cytokines such as tumour necrosis factor, IL-6 and IL-8 are locally expressed in the bone marrow of RA patients.28 On the other hand, the role of bone marrow in RA might not be entirely negative, as signs of endosteal bone formation have been described adjacent to such lesions. In particular, B-cell aggregates can support bone formation and thus indirectly limit structural damage.29


While we admit that the football analogy is somewhat contrived, Mr Inside and Mr Outside provide a convenient foil for thinking about RA. In principle, two separate pathophysiological concepts of arthritis parallel the running styles of our heroes. One is that arthritis is a primary synovial disease, which invades neighbouring tissue and finally also bone marrow. The other is that of a primary bone marrow disease, which later migrates to the synovial membrane. Evidence in support of both hypotheses can be found when studying arthritis, and at the present time neither of the hypotheses can be confirmed. Different forms of arthritis may preferentially use either the outside-in or the inside-out mechanism. It could well be that RA is indeed a true synovial disease and reaches bone marrow early by using pre-existing bone channels or small bone erosions. Other arthritides, however, may preferentially follow the inside-out concept. Therefore, inflammation in anterior spondylitis dominates the bone marrow and also affects sites with no adjacent synovium, such as the intervertebral spaces. The latter observation would mandate an inside-out process as the only plausible explanation for lesions such as anterior spondylitis. Osteitis, rather than synovitis, might in fact preferentially stimulate bony spur formation rather than erosions. For instance, bone marrow inflammation itself could be a major anabolic signal, which is transduced by osteocytes through cortical bone.30

The term ‘arthritis’ may in fact reflect fundamentally different conditions: diseases that emerge from primary synovitis and follow the outside-in concept and those starting with primary osteitis following the inside-out concept. These concepts are not exclusive and can be used by the same disease. However, preference for one of these two mechanisms may better explain some of the clinical differences of the various forms of arthritis.



  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.