Kv1 Channel Localization and Function in Demyelinated and Remyelinating Axons

Kv1 channel localization after demyelination and during remyelination was described using the lysolecithin model of peripheral demyelination (Rasband et al., 1998). In these experiments, lysolecithin was injected directly into the sciatic nerve, resulting in activation of macrophages, and the disruption and eventual phagocytosis of myelin. In this model, complete, focal demyelination occurs at the site of injection within about 1 week. After demyelination, Schwann cells proliferate and are able to remyelinate the injured region. However, the structure of remyelinated axons is different than before the drug injection: there are fewer layers of myelin, and the internodal length is decreased to about one-fourth the length found in uninjured animals. As a result, new nodes of Ranvier form in regions that were formerly internodal with low densities of ion channels. Finally, it is relatively easy to isolate the sciatic nerve and record action potentials in the demyelinated/remyelinating nerve. Thus, this model provides an excellent system in which to examine the localization and relative contribution of ion channels to nervous system dysfunction after disease or injury. This model has also been used to study the mechanisms of Nav channel localization after demyelination and during remyelination (Dugandzija-Novakovic et al., 1995). An indepth discussion of Na+ channel clustering during remyelination is given by Shrager et al. (see Chapter 8).

One week after injection of lysolecithin myelin is cleared from the injected site, and broad regions of denuded axons are apparent. Double-immunostaining with antibodies against Kv1.1 and Nav channels reveals former nodal sites, as these can be easily distinguished based on the presence of a focal Nav channel cluster (Fig. 3A,B, arrowheads). In con-strast, Kv1 channels were not retained at juxtaparanodes, but instead were very labile in the axolemma and were able to diffuse into formerly paranodal and internodal zones. A quantitative analysis of this dispersion showed that 6 days after injection, 60% of nodal regions had some detectable Kv1 channel immunoreactivity. Just 1 day later, however,

Kv1.1 NaCh

Kv1.1 NaCh

Kv1.1 MAG

Figure 3 Kv1 channel localization in demyelinated and remyelinating rat sciatic nerve fibers. Demyelinated and remyelinating nerve fibers doublelabeled for Kv1.1 (A, C, E, G, I, K) and Nav channels (B, D, F) or the myelin associated glycoprotein (MAG; H, J, L). 6 days postinjection (A, B), 9 days postinjection (G, H), 18 days postinjection (C, D, I, J), and 24 days postinjection (E, F, K, L). (M, N) Immunoelectron microscopy showing Kv1.1 in nodal (M) and paranodal (N) regions at 19 and 27 days postinjection, respectively. Scalebars: A-H = 25 |m; I-L =12 |m. (Figure modified from Rasband et al. 1998, copyright 1998 by the Society for Neuroscience.)

Figure 3 Kv1 channel localization in demyelinated and remyelinating rat sciatic nerve fibers. Demyelinated and remyelinating nerve fibers doublelabeled for Kv1.1 (A, C, E, G, I, K) and Nav channels (B, D, F) or the myelin associated glycoprotein (MAG; H, J, L). 6 days postinjection (A, B), 9 days postinjection (G, H), 18 days postinjection (C, D, I, J), and 24 days postinjection (E, F, K, L). (M, N) Immunoelectron microscopy showing Kv1.1 in nodal (M) and paranodal (N) regions at 19 and 27 days postinjection, respectively. Scalebars: A-H = 25 |m; I-L =12 |m. (Figure modified from Rasband et al. 1998, copyright 1998 by the Society for Neuroscience.)

immunoreactivity was reduced to only 20% of sites. These results suggest that either Kv1 channels are rapidly internalized on demyelination, or that they are able to freely diffuse in the axolemma in the absence of an overlying myelin sheath. Since Caspr2 and TAG-1 deficient mice have dramatically reduced amounts of juxtaparanodal channels, but no decrease in the total numbers of channels (Poliak et al., 2003; Traka et al., 2003), it is more likely that in the absence of myelin, the juxtaparanodal K+ channel complex is unable to interact with glial TAG-1 and remain restricted to the jux-taparanode.

At the start of the second week after injection (7 to 9 days after injection) Schwann cells associate with axons, extend processes, and begin to remyelinate damaged regions. Although Nav channels can be detected at the edges of these extending Schwann cell processes (Dugandzija-Novakovic et al., 1995), Kv1 channels are undetectable and do not accumulate at these sites. For example, Figs. 3G and 3H show a remyelinating nerve fiber double-labeled for Kv1.1 (Fig. 3G) and the myelin-associated glycoprotein (MAG, a marker of early myelination; Fig. 3H). This figure shows that newly forming nodes of Ranvier (arrowhead) are devoid of any Kv1 channel immunoreactivity.

Although initially new nodes of Ranvier lack Kv1 channels, they rapidly acquire a high density as remyelination progresses. These Kv1 channels co-localize with nodal Nav channels (Figs. 3C, D, arrowheads) and are restricted to the gap between adjacent MAG-labeled, myelinating Schwann cells (Figs. 3I, J, arrowheads). By the end of the third week after injection, as many as 60% of nodal sites have Kv1 channel immunoreactivity. This result is surprising and significant, as it suggests that although remyelination may occur with appropriate Nav channel clustering, the aberrant localization of Kv1 channels may position them in a region where they can block normal conduction. Further, this result suggests that in contrast to Nav channels clustered at new nodes of Ranvier (which derive from a preexisting pool of axolemmal channels [Tzoumaka et al., 1995]), the Kv1 channels are manufactured de novo and inserted at nascent nodes of Ranvier. The nodal localization of these channels has been confirmed by immunoelectron microscopy: Fig. 3M shows a new node of Ranvier with a dense, dark precipitate at the node (arrows).

At the start of the fourth week after lysolecithin injection, the formerly nodal Kv1 channels begin to redistribute into paranodal and finally their normal, juxtaparanodal zones. Figures 3E and F show the initial stage where these Kv1 channels begin to split and become restricted beneath the myelin sheath. Interestingly, the split occurs precisely at the site where Nav channels are clustered at the highest density (Figs. 3E, F, arrowhead). Double-immunostaining with anti-MAG shows that the point where Kv1 channels begin to be removed from the node corresponds to the edges of MAG immunoreactivity (compare Figs. 3K and L, arrowhead), suggesting that myelin sheath may actively participate in determining the localization of Kv1 channels. Immunoelectron microscopy of remyelinating sciatic nerve fibers shows the redistribution of Kv1 channels into paranodal and juxtaparanodal domains (Fig. 3N, arrows).

The results described here indicate that demyelination and the lack of glial binding partners, possibly TAG-1, results in the redistribution of Kv1 channel proteins throughout the axolemma due to the latter's intrinsic mobility within the membrane. Remyelination also initiates the clustering of Nav channels (Dugandzija-Novakovic et al., 1995) at the edges of elongating processes. As new nodes form and are defined, newly synthesized Kv1 channels are inserted at nascent nodes by mechanisms that remain unknown. At this stage of remyelination, Kv1 channels are transient in their localization to nodes of Ranvier; they soon become sequestered beneath the myelin sheath in paranodal and juxtaparanodal domains. Of importance, if remyelination is delayed by addition of the mitotic inhibitor mito-mycin-C, Schwann cells do not proliferate, axons remain unmyelinated, and channels fail to cluster. Although there is an apparent increase in the amount of Kv1 channels in these demyelinated axons at 3 to 4 weeks after lysolecithin injection, the channels remain diffusely distributed throughout the axolemma. Under these conditions, remyelination eventually occurs and Kv1 channels eventually become localized first at nodal, then juxtaparanodal domains (Rasband et al., 1998).

Significantly, the localization of Kv1 channels during remyelination closely parallels the degree of pathophysiology and sensitivity to K+ channel blockers such as 4-AP. Thus, before demyelination and when Kv1 channels are sequestered beneath the myelin sheath in juxtaparanodal

Figure 4 The sensitivity of remyelinating nerve fibers to 4-aminopyridine (4-AP) correlates closely with the fraction of nodes that have Kv1 channels. Kv1 channels are transiently detected at nodes of Ranvier with a peak at about 18 days postinjection. Boxed insets show, for the times after lysolecithin injection indicated, compound action potentials before and after exposure to 1 mM 4-AP. The boxed inset on the left shows compound action potentials from noninjected, control sciatic nerves before and after exposure to 1 mM 4-AP. (Figure modified from Rasband et al., 1998, copyright 1998 by the Society for Neuroscience.)

Figure 4 The sensitivity of remyelinating nerve fibers to 4-aminopyridine (4-AP) correlates closely with the fraction of nodes that have Kv1 channels. Kv1 channels are transiently detected at nodes of Ranvier with a peak at about 18 days postinjection. Boxed insets show, for the times after lysolecithin injection indicated, compound action potentials before and after exposure to 1 mM 4-AP. The boxed inset on the left shows compound action potentials from noninjected, control sciatic nerves before and after exposure to 1 mM 4-AP. (Figure modified from Rasband et al., 1998, copyright 1998 by the Society for Neuroscience.)

domains, peripheral myelinated nerve fibers are insensitive to 4-AP (Fig. 4). However, 2 weeks after demyelination these fibers become sensitive to 4-AP. The most dramatic effects of 4-AP occur when the majority of new nodes have Kv1 channels. Addition of 4-AP at these times significantly increased both the amplitude and duration of the action potential. It is important to note that at this stage of remyeli-nation axons are well-wrapped, and there are significant densities of nodal Nav channels, suggesting that the pathophysiology is a consequence of the misplaced Kv1 channels rather than too little myelin or too few nodal Nav channels. As Kv1 channels become sequestered beneath the myelin sheath, the sensitivity to 4-AP decreased (Fig. 4A). Taken together, these results suggest that subsequent to demyelina-tion and during remyelination, axolemmal Kv1 channels are exposed, aberrantly localized to nodes of Ranvier, and block conduction by their activity at these sites.

0 0

Post a comment