Wilkinson et al. identified the neutrophil elastase (NE) as a novel and potent activator of MMP-13. The authors could show that in vitro incubation of human cartilage samples with NE leads to an MMP-dependent release of collagens into the culture medium. Glycosaminoglycans (GAGs) are also released upon this treatment, however, this could not be inhibtied by the broad MMP inhibitor GM6001. NE alone was able to strongly activate recombinant MMP-13 in vitro, while activation of MMP-1 was only modest and MMP-8 was not at all activated. The cleavage site of NE to activate MMP-13 was identified by N-terminal sequencing. The authors could detect both NE and MMP-13 in synovial fluid of OA patients further stengthening the potential role of NE in MMP-13 activation during OA progression. Interestingly, NE and MMP-13 actvity seems to correlate with synovitis. Using a set of elegant experiments the authors come up with a nice model of how NE activates MMP-13 and itself is regulated by AAT. A regulatory loop that could well be relevant in OA.

This manuscript contains a lot of interesting data that will definitely be interesting for a broader readership. The authors used a large variety of different methods and very nicely combined in vitro and in vivo data. The results presented are of high quality and the experiments are straight forward. The paper is well written, very easy to follow and the conclusions drawn are fully supported by the data.
I only have a couple of minor comments that the authors might want to address to further improve the quality of their manuscript:

In the introduction, the authors mention that clinically the success of MMP inhibition is limited due to low selectivity. However, in some of the in vitro experiments performed in the present study (e.g. Figure 1) the use of highly selective MMP-13 inhibitors (for example one that is freely availble from Boehringer Ingelheim) would have been interesting to be able to assign the collagenolytic and GAGolytic activity to MMP-13.

In Figure 1, collagen and GAG release was measured under five different conditions. Is there any reason why MMP activity was not determined for NE alone?

In the methods section on page7 line 13, the authors wrote that 20nM NE was used for in vitro experiments. However, in the treatment scheme in Figure 1 200 nM NE are given. The authors should clarify if this is only a typo or, if true, why a tenfold higher concentration was used.

I like the kinetic of activation shown in Figure 3. I assume that also here the used NE concentration was 20 nM? Could it be interesting to show a concentration-dependent activation also on a gel? In a patient, the concentrations might be lower but, in particular in chronic diseases, act much longer. Could the concentrations from the ELISA be given in molar concentrations?

It is nice that NE could also be detected in synovial fluid (and I can fully understand that synovial fluid form healthy individuals was not available. However, it would be interesting to know if NE levels under more inflammatory conditions like rheumatoid arthritis are similar or higher. Further, is there a correlation between NE and MMP-13 levels? Do the patients with high(er) NE levels always show high(er) MMP-13 levels? Does the ELISA used detect total or active MMP-13?

In the discussion, I learned that elastase might cleave some collagens directly (though most likely not collagen II in cartilage). Is there anything known about cleavage of the minor cartilage collagens IX and XI (or purified collagen II)?

The authors addressed all points appropriately.