An aggrecan fragment drives osteoarthritis pain through Toll-like receptor 2.

Pain in surgical joint instability models of osteoarthritis (OA) is challenging to measure. Such models have modest inflammation and painful behavior can be subtle. Previous studies have defined three distinct patterns of painful behavior after joint destabilization in mice. Distal mechanical allodynia (measured by von Frey filaments) occurs immediately post-operatively and persists beyond 16 weeks (Miller et al., 2012); joint hyperalgesia (assessed using the Pressure Application Measurement tool) occurs immediately in the post-operative period and resolves between 12 and 16 weeks (Miller et al., 2017); and spontaneous painful behaviour (measured by incapacitance testing) occurs briefly in the post-operative period (for around one week), and then develops from 12 weeks onwards (Inglis et al., 2008).

In this very elegant study by Miller et al., compelling evidence is presented that early pain sensitization in the destabilized joint is due to generation of a specific 32-amino-acid aggrecan fragment (32-mer) peptide derived by proteolysis of aggrecan, the principle proteoglycan in cartilage. This peptide is generated by the action of two different proteases and induces pain responses through Toll-like receptor 2 (TLR2). TLR2 null mice or mice that are unable to generate the peptide (by a knock-in point mutation that changes one of the cleavage sites on aggrecan) are protected from osteoarthritis-induced pain even though they still get cartilage breakdown and other hallmarks of disease.

The group shows that the 32mer stimulates dorsal root ganglion (DRG) neurons through TLR2 to induce the chemokine CCL2, a known pain sensitizer. They speculate that this mediates the transient regulation of CCL2 that is detected in DRGs eight weeks following surgery according to their previous study (Miller et al., 2012).

A number of interesting questions arise from this work. Firstly, why does joint hyperalgesia resolve by 16 weeks? Is this because there is no longer generation of the 32mer due to the accrued loss of cartilage (substrate) over this time? Or does the joint somehow become refractory to its stimulation? As 12 weeks is when spontaneous painful behavior is measured in the model by incapacitance testing (Inglis et al., 2008), this raises the possibility that the sensitization they describe is necessary for the second phase of pain. As the latter is mediated by nerve growth factor (NGF), a validated target in human and murine osteoarthritis pain (Lane et al., 2010; McNamee et al., 2010), it would be interesting to test whether TLR2 blockade early in disease prevents the NGF-dependent late phase.

The concept that cartilage damage gives rise to pain in osteoarthritis is not new (Driscoll et al., 2016) but it nonetheless challenges historical paradigms. Whilst other tissues of the joint such as the synovium and bone may also contribute to generation of pain in patients with complex osteoarthritis pathology, this paper highlights the value of careful molecular dissection of pain pathways in non-inflammatory models of disease.