LETTER TO THE EDITOR Does CNTF mediate the effect of intraocular inflammation on optic nerve regeneration?

Sir, we wish to comment on the paper ‘Astrocyte-derived CNTF switches mature retinal ganglion cells (RGCs) to a regenerative state following inflammatory stimulation’ (Muller et al., 2007). RGCs, the projection neurons of the eye, are normally unable to regenerate their axons if the optic nerve is injured. However, they can be stimulated to do so by inducing an inflammatory reaction in the eye (Leon et al., 2000; Yin et al., 2003). The evidence presented later argues against CNTF playing a central role in this transformation. The authors do not demonstrate that CNTF can directly stimulate RGCs to regenerate axons. They show that coculturing retinal explants with lens or zymosan increases CNTF expression in retinal glia, but they do not show that this CNTF is secreted nor whether this stimulation, or even exogenous CNTF, is sufficient to induce axon outgrowth in these cultures. Two studies have shown that it cannot (Cohen et al., 1994; Cen et al., 2007). In dissociated cultures, CNTF stimulates axon outgrowth in immature RGCs provided cAMP levels are elevated (Jo et al., 1999), but it has only a weak effect on mature RGCs (which is also cAMP-dependent) (Yin et al., 2003, 2006). A new paper in Brain shows that CNTF by itself has barely any effect on RGC5 cells (Lingor et al., 2008). The effects exerted by CNTF in vivo appear to be indirect. Very high concentrations of CNTF (41000 ED50) can stimulate RGCs to regenerate injured axons through a peripheral nerve graft, and, if expressed via viral infections, into the optic nerve (Cui et al., 1999; Cui and Harvey, 2000; Park et al., 2004; Leaver et al., 2006). The present study shows that 2 injections of CNTF (41000 ED50) stimulate axon regeneration in the optic nerve; Lingor et al. (2008) report a much weaker effect. Importantly, however, CNTF is a chemoattractant for blood-borne macrophages, and systemic depletion of macrophages eliminates most of the axon-promoting effects of CNTF in vivo (Cen et al., 2007). Studies using inhibitors also fail to support the contention that CNTF is the primary link between intravitreal inflammation and axon regeneration. In earlier studies, antibodies against CNTF or its receptor failed to diminish axon regeneration following lens injury (Leon et al., 2000; Lorber et al., 2002). In the present study (Fig. 5H), injecting an anti-CNTF antibody after lens injury reduced neurite outgrowth by only 30% when retinas were explanted into culture several days later; an inhibitor of the Jak-STAT signalling pathway reduced growth by 40%. Thus, although CNTF and/or Jak-STAT signalling may play an ancillary role (e.g. in cell survival), they do not appear to be essential for outgrowth per se. Significantly, the authors do not report whether these reagents affected regeneration in vivo. Müller et al. argue that oncomodulin, another protein that has been proposed to mediate the effects of intravitreal inflammation, does not play a significant role in this regard. In the presence of elevated cAMP, oncomodulin enhances the ability of RGCs to regenerate axons in culture and in vivo (Yin et al., 2006). Müller et al. cite data from their lab that oncomodulin is not elevated in the eye after lens injury (Hauk et al., 2007), but a prior study showed that lens injury strongly upregulates oncomodulin mRNA and protein levels (Yin et al., 2006), and these results have been confirmed using quantitative real-time PCR and an affinitypurified antibody (Y. Yin, Q. Cui and L. Benowitz, in preparation). Müller et al. report that injecting a polyclonal anti-oncomodulin antibody into the eye did not diminish the ability of RGCs to extend axons in culture after being exposed to lens injury in vivo. The antibody that they used can eliminate the axon-promoting activity of Brain (2008) Page E1 of E2