The research on PGC-1 and inflammation has now rapidly expanded on several experimental levels.
- On cultured muscle cells (c2c12): It has been clearly demonstrated (Mormoneo et al PLoS One 2012) that muscle cells adenovirally infected with PGC-1 express higher levels of proinflammatory markers.
STRONG PART: The researchers were very smart and used equal multiplicity of infection (MOI) of GFP and PGC-1. The effects they see are therefore real and not due to different infection degrees of the different viruses.
WEAK PART: The overexpression by adenoviruses (AD) is too big. One AD strain expresses only GFP, the other GFP and PGC-1. The second strain thus even has to preduce the double amount of proteins, which is associated with enormous cellular stress.
c2c12 are immortalized cells. It has to be answered to which extent the cancer phenotype, which is probably further favored by overexpression of PGC-1 (hypermetabolic state) is responsible for the increase in inflammatory markers.

- On isolated myoblasts: This is maybe the best and strongest experiment. It has been clearly demonstrated (Olesen et al Medicine & Science in Sports & Exercise 2010) that myoblasts from PGC-1 transgenic animals express higher levels of proinflammatory markers, especially of TNFalpha
STRONG PART: the overexpression of PGC-1 is physiological and not too high. The artefacts discussed above can be excluded

- In mice directly: this article by Olesen at al finishes the story about PGC-1 and inflammation by delivering the last piece of evidence from mice. Moreover, as mentioned by others, further studies deliver the explanations for the pro-inflammatory role of PGC-1. As mentioned in the comment of Dr. Hasanova, Romanino, K et al PNAS 2012 showed in the same rodent model of muscle specific overexpression of PGC-1 increases in skeletal muscle Akt protein levels. Akt, also known as Protein kinase B (PKB) is known to phosphorylate p65 (shown by two independent groups: Viatour, P et al Trends Biochem Sci 2005 and Chen, LF et al Nat Rev Mol Cell Biol 2004). Therefore, there is increased phosphorylation of p65 in muscle-specific transgenic mice. In addition, Summermatter, S et al AJP 2012 demonstrated in the same model that calcineurin activity is increased. Calcineurin activates p65 and this has directly been shown in muscle cells (shown by two independent groups: Harris, CD et al Cellular and Molecular Life Science 2005 and Alzuherri, H and Cheng KC Cell Signal 2003).
STRONG PART: The story is very complete and convincing. It was important that LPS was used to induce inflammation rather than TNF directly. LPS is a inflammatory stimulus only, while TNFalpha is a inflammatory stumulus, but also leads to very severe atrophy and muscle degradation. That would completely prevent interpretation of the results, because a catabolic model would be studied.
It is very important, too, that LPS was injected intraperitoneally and not directly into the muscle. Injection into muscle would have resulted in attraction of macrophages. Such recruitment could differ between control and transgenic animals, because of the high vascularization in transgenic animals. The higher surface of blood vessels allows faster infiltration, but also faster removal of macrophages.
WEAK PART: In general the question hast to be asked, if muscle is important for inflammation. Compared to adipose tissue, the contribution in cytokine release by muscle is neglectible.

All the articles together are in perfect line and strongly support the pro-inflammatory role of PGC-1