[page 79↓]

4.  Discussion

Potassium channels are important components of the signal transduction machinery of the nervous system and virtually all other cells of the mammalian body. They are involved in a wide variety of functions such as setting and stabilizing the resting potential of most cell types or regulating the depolarization time course in pacemaker cells. Other parameters like action potential duration, firing frequencies, and interspike intervals are also determined by the activity of potassium channels1. A functional channel is formed by the homomeric or heteromeric aggregation of four subunit proteins, usually from members of one channel family. The basic structural organisation of such a subunit consists of a short amino acid sequence, which forms a “loop” into the membrane and is flanked by two transmembrane domains (see Fig. 1B). The loop takes part in forming the channel pore. This basic structural unit may be supplemented by four additional membrane spanning domains, or two units may be combined to a single protein, thus forming three structural classes of potassium channels2. With respect to sequence homologies, each of these classes consists of at least six families, again each represented by one to nine members, each possibly represented by more than five splice variants. Summing up to about 90 individual genes with even more functional gene products from different splice variants and the formation of heterooligomers, there is an impressive amount of potassium channels, large enough to cover the complex range of channel functions in living cells.

This complexity explains the fact that a variety of quite distinct diseases can be traced back to ion channel defects. Thus, Andersen’s syndrome (AS), a rare inherited disorder characterized by periodic paralysis, cardiac arrhythmias, and dysmorphic features, is due to mutations of the Kir2.1 potassium channel3. Furthermore, an increasing number of neurological diseases such as hereditary deafness syndromes, episodic ataxia type-1, acquired neuromyotonia, and others are the results of compromised potassium channel functions4. On the other hand, the large molecular diversity of the channels should allow the development of highly specific drugs, which may selectively target cell types, groups, or systems throughout the body, correcting or at least improving disturbed functions. The [page 80↓]great potential of this approach is documented by the effective treatment of diabetes or cardiac arrhythmias with drugs, which are primarily directed to ATP-sensitive or HERG-type potassium channels, respectively. The impact of this strategy for improving neuropsychiatric pharmacotherapy is under consideration5. Systematic development of further neuroactive drugs, however, will depend on information about the involvement of molecularly defined potassium channels in identified neuronal circuits. In this respect it is important to understand the regional, cellular, and subcellular localizations of identified potassium channel subunits in clinically important brain nuclei such as the basal ganglia.

In the present study we generated highly specific, affinity-purified rabbit antibodies to study the detailed distribution of Kir2 inwardly rectifying potassium channel subunits in the striatum. With respect to a potential involvement of Kir2 channels in motor output of the basal ganglia, we focused on the striatal key structures for movement regulation. In particular, we asked whether any of the Kir2 channels are selectively distributed in the patch or matrix compartment of the striatum or preferentially localized at cholinergic interneurons.

4.1. 

4.1.1. Kir2 channel proteins are differentially expressed throughout the rat brain

The monospecific antibodies allowed the characterization of Kir2 protein localization in the rat central nervous system. The widespread presence of all four Kir2 channel subunits in the rat brain may indicate their important role in central signal processing and neural transmission. Our data agree with earlier in situ hybridization (ISH) experiments (see chapter 3.3.) and electrophysiological studies, confirming the basic distribution patterns of Kir2.1, Kir2.2 and Kir2.3 subunits. However, ISH experiments only indicate which neurons can synthesize these subunits. Our immunocytochemical experiments extend these data, presenting information on the subcellular localization in the different types of neurons. Exemplary in the basal ganglia, all Kir2 subunits are localized in the somata and dendrites of striatal neurons. Data obtained so far suggest that Kir2.2 is the only member of this channel family, which is additionally sorted to the axonal plasma membrane.


[page 81↓]

Regional and cellular localizations of the Kir2.4 subunit, as presented here, contrast an earlier in situ hybridization study by Töpert et al. 1998, where Kir2.4 mRNA was reported to be primarily expressed in neurons of cranial nerve motor nuclei. The present investigation reveals a widespread Kir2.4 immunoreactivity among neurons in many brain areas (see chapter 3.3.). In all areas the signal was blocked by preabsorption of antibodies with its cognate recombinant protein, ruling out background labeling. Another difference is the Kir2.4 localization in the rat striatum, where no Kir2.4 transcripts were detected by in situ hybridization6. We found a signal in most, if not all striatal neurons. With the exception of the cholinergic interneurons, however, Kir2.4 immunoreactivity is less intense as compared to the other members of the Kir2 subfamily.

The apparent contradiction may be, however, simply due to aspects of sensitivity. We also find very strong labeling of brain stem motor neurons (Fig. 17-VII,J), where Kir2.4 mRNA was reported to be primarily expressed, but we detect weaker expression in most neurons throughout the brain. Our data, therefore, support the prominent localization of Kir2.4 protein in motor nuclei. Interestingly enough, motor neurons are cholinergic as are the giant interneurons in the striatum. This raises the possibility that cholinergic neurons in general may preferentially use the Kir2.4 subunit to control their resting potentials. Further investigation of other cholinergic cell groups in the vertebrate CNS should provide more decisive information.

The sensitivity aspect is supported by experiments with the rat retina. Kir2.4 transcripts7 and proteins8 were clearly detected in the retina, while Töpert et al. did not find any Kir2.4 mRNA expression in this organ. These data strengthen our hypothesis of a more widespread Kir2.4 localization as supported by Töpert’s study.

The detailed channel protein distribution of all members of the Kir2 subfamily in the rat brain is summarized in table 4. Excellent agreement of the Kir2.1-Kir2.3 subunit (for Kir2.4 see above mentioned comments) with former in situ hybridization experiments9 was found in most brain regions, for example olfactory bulb, olfactory tubercle, neocortex, hippocampus, caudate putamen, nucleus accumbens, lateral olfactory tract nucleus, hypothalamus, habenula and superior colliculus (Fig. 17-I – 17-IX). But in some areas [page 82↓]there are differing distribution patterns of mRNA and channel protein expression. The most obvious difference was observed in the hindbrain, where the Kir2.1 and Kir2.3 subunits were reported to be virtually absent10. In contrast to their report and in strong agreement with a previous mRNA study on the Kir2.3 subunit11, we found the distribution of the Kir2.3 channel subunit extended to cerebellar Purkinje cells and to neurons of the deep cerebellar nuclei (Fig. 17-VIII,E/I). The Kir2.1 subunit was also detected in the cerebellum (molecular layer and deep nuclei), which is supported by a former localization study12. Nevertheless, the cerebellar expression of the Kir2.2 protein with very high signals in the granule cell layer is confirmed in our investigation.

Another difference was found in the thalamus (Fig. 17-V). Although the Kir2.2, Kir2.3 and Kir2.4 channel distribution is similar to the previously described mRNA localization pattern, the Kir2.1 protein was found at significant levels in various thalamic nuclei, thereby contrasting former investigations13. Finally, we observed expression of all four members of the Kir2 subfamily in the substantia nigra, where only the Kir2.2 subunit was reported to be expressed14. In fact, the Kir2.2 channel protein was present in markedly elevated levels, but also the Kir2.3 subunit displayed specific signals mainly in the pars reticulata.

4.1.2. The Kir2.3 subunit is preferentially expressed in striatal matrix neurons

All members of the Kir2 subfamily are differentially distributed throughout the rat striatum, thus confirming the fundamental localization pattern as reported from previous in situ hybridization experiments (see table 5). The broad distribution of all Kir2 channels suggests important roles for these channels in regulating neuronal excitation and transmission within the basal ganglia circuitry.

The present study demonstrates an inhomogeneous distribution of potassium channels in the striatum. Thus, one member of the inwardly rectifying Kir2 family, namely the Kir2.3 channel protein in the rat striatum, displays a heterogeneous pattern that can neither be [page 83↓]explained by varying cell densities nor by the numerous fiber bundles. Instead, the density of the inwardly rectifying potassium channels nicely correlates with the striatal compartments. The Kir2.3 channel tends to avoid the patch compartment and is predominantly expressed in the striatal matrix.

An inhomogeneous distribution of Kir2.3 channels in the striatum was inferred from Surmeier’s group15 using single-cell RT-PCR and whole-cell patch clamp techniques. They found a preferential expression of Kir2.1, Kir2.2 and Kir2.3 channels in either SubP-positive or Enk-positive medium spiny neurons. The Kir2.3 channel was expressed in all SubP-positive neurons but only in a minority of Enk-positive cells. Thus, the discovered distribution pattern refers to the distinction between direct and indirect pathway of striatal signal transduction (see chapter 1.2.3.). Extending these data, our results illuminate a difference between the striatal patch and matrix compartment.

Fig. 26: Ratios of patch neurons to matrix neurons across embryonic ages. Ratios above the solid horizontal line indicate a predominance of patch neuron birthdates and ratios below the line indicate a predominance of matrix neuron birthdates. Consequently, at the time of beginning Kir2.3 mRNA expression most of the neurons end up in the matrix compartment.

The preferential localization of the Kir2.3 subunit in the striatal matrix compartment is in agreement with developmental studies. Thus, most of early born neurons (becoming postmitotic between E12-E16) end up in the patch, while the majority of later born neurons [page 84↓](E18-postnatal day P2) mature in the matrix compartment16. Correspondingly, onset of Kir2.3 channel expression begins late, and Kir2.3 mRNA is absent from the brain until E2117. Apparently, at the time when striatal neurons start expressing Kir2.3 channel proteins, these neurons are still in development and there is an extraordinarily high probability that they will end up in the matrix compartment (Fig. 26). Of course, not only developing neurons express Kir2.3 channels. But Kir’s evident influence in development of trophic interaction and synaptogenesis18 suggests an electrophysiological involvement as early as necessary.

Inwardly rectifying Kir2 channels dominate the electrophysiological behavior in the down-state of medium spiny neurons19 (see chapter 1.2.1) by keeping the membrane potential close to the K+ equilibrium potential. Based upon our studies, in matrix neurons this is mainly done by the Kir2.3 channel. Thus, the amount and activity of Kir2.3 channels determine the membrane conductance in matrix medium spiny cells. It is well known that the transition to the up-state and following spike generation depends on temporally coherent excitatory synaptic input. If the cells express a high number of Kir2.3 potassium channels, massive convergent input is necessary to generate the up-state. Consequently, it may be expected that pharmacologically blocked Kir2.3 channels fail to stabilize the neuronal down-state.

The matrix-related basal ganglia loops are thought to transmit effectory signals, i.e. the motor output. In contrast, patch-related circuitry mainly accounts for the regulation of the motor effects. Application of a selective Kir2.3 channel blocking drug could facilitate spiking within the matrix compartment by excitatory input while the patch might be virtually unaffected due to their absent Kir2.3 expression. Therefore, blocking of Kir2.3 channels may specifically improve the motor outcome (for example bradykinesia) in Parkinson’s disease.


[page 85↓]

4.1.3.  Kir2.4 immunoreactivity in the striatum is most prominently displayed by the giant cholinergic interneurons

Although anti-Kir2.4 antibodies react with many neurons in the striatum, they most intensely stain the giant interneurons. The cholinergic nature of these neurons was confirmed by double labeling experiments at the light and electron microscopic level. Cholinergic interneurons also display the Kir2.2 subunit, but only at a level comparable to that in the spiny projection neurons. Our results are supported by the Kir2.2 mRNA distribution as reported recently20.

Bearing in mind the pivotal role of cholinergic neurons for the motor outcome of neuronal computations in basal ganglia, Kir2.4 subunit containing potassium channels, due to their restricted localization, may be especially important. They will take part not only in controlling the activity of the interneurons, but may be further involved in regulating the fine tuning between the two basal ganglia pathways. Thus, cholinergic interneurons are known21 to activate the spiny projection neurons of the indirect pathway via M1 receptors (see introduction). M1-based stimulation of these neurons should be additionally facilitated by the concomitant destabilization of the resting potential due to an inhibition of Kir2.3 channel activity inherent to M1 receptor activation22.

4.1.4. Can Kir channel subunits be targets for novel therapeutic strategies?

As mentioned above, a variety of distinct diseases can be traced back to ion channel defects. This is also confirmed in the special case of inwardly rectifying potassium channels. Mutations in genes encoding Kir channels have been shown to cause hereditary diseases. Salt-wasting and metabolic acidosis in Bartter’s syndrome23 often is due to the malfunction of ROMK1 (Kir1.1) channels in the kidney. Degeneration of cerebellar granule cells with resulting ataxia is due to Kir3.2 missense mutations in weaver mice24. However, only little progress has been made in supporting recent therapeutic strategies by targeting Kir channel subunits. This may be primarily due to the fact that our knowledge on the detailed [page 86↓]biological functions of Kir channels in the nervous system is still limited. Some natural drugs affecting Kir channel activity like the Kir1.1/Kir3.1-specific channel inhibitor tertiapin25 and the spermine-like Kir blocker Philanthotoxin26 are available. These are hopeful beginnings on the way to specifically treating Kir channel-based diseases.

Inwardly rectifying potassium channels can be attractive candidates for novel therapeutic regimes only if the channels are selectively expressed in those nervous structures that are important for any neuropsychiatric disorder or for side effects during established treatment. In that case, activation or inhibition of Kir channels may influence the clinical situation of affected patients. The present study reports the Kir2 channel localization to defined structures of the basal ganglia circuits, which are basically involved in the regulation of movements. Patch and matrix striatal compartments as well as the cholinergic interneurons represent these key structures and display a specific distribution pattern of Kir2 channel subunits.

Pharmacological manipulations of potassium channels will extend available interactions with transmitter systems. Thus, selective activation of Kir2.4-subunit containing channels in the striatum will stabilize the resting potential of giant cholinergic interneurons. Consequently, action potential firing and cholinergic transmission will be reduced in a manner comparable with the application of an anticholinergic drug. Because of the influence restricted to the interneuron, however, diverse side effects of anticholinergic therapy may be absent or reduced. It could not be expected, of course, that simply targeting Kir2.4 channels in the striatum by channel openers or blockers would affect only motor behavior without side effects on other neurons. However, it is conceivable to use such drugs at subthreshold doses. Thus, on their own, such treatments may remain without obvious pharmacological effects. Combination with again subthreshold doses of a second drug, however, which affect striatal ChAT neurons via a distinct mechanism, for instance other channels27, should cause an effect above threshold.

Pharmacological intervention with neurotransmission within the matrix and patch compartments of the striatum seems to be more complex with respect to the predominant Kir2.3 expression in the matrix. Increasing membrane excitability by inhibiting Kir channels may partly compensate the problems of dopamine-deficiency such as akinesia, tremor, or rigor in parkinsonian states.

Matrix inputs in most areas of the striatum are predominantly related to sensorimotor processing, whereas patch input has a strong relation to the limbic system. Keeping in [page 87↓]mind these functional compartmental differences, a Kir2.3-mediated activation or inhibition of matrix neurons may predominantly influence motor circuits within the basal ganglia, leaving limbic circuits mainly unaffected. Thus, targeting Kir2.3 channels may participate in an improved neuroleptic therapy. Aiming at motor and not at limbic circuits may disconnect the antipsychotic effects of neuroleptics from their severe motor side effects, such as early and late dyskinesias as well as parkinsonoid syndromes.

Additional examination of limbic circuits will be required to provide further information. The nucleus accumbens seems to connect motor outcome with the limbic system (“limbic/motor interface”) and plays a major role in mediating motivation and reward28. It contains two subdivisions, shell and core, which are cytoarchitectonically and pharmacologically distinct and differ also in their afferent and efferent connections29. The present investigation revealed the expression of all four Kir2 subfamily members in the nucleus accumbens (Fig. 17-IV). Interestingly, the Kir2.3 subunit displayed a patchy distribution in the core part resembling the striatal expression pattern and supporting the strong histochemical relation between core and striatum30. Moreover, particularly elevated signals of the Kir2.3 channel protein in the shell part of the nucleus accumbens suggest profound involvement in the neuronal excitability and regulation at this site of limbic/motor interaction. Therefore, a more detailed analysis of ventromedial striatal areas and the nucleus accumbens is in progress in our laboratory.


Footnotes and Endnotes

1 Hille, B. 1995

2 for more information see http://nt-salkoff.wustl.edu

3 Plaster, N.M. 2001, Felix, R. 2000

4 Benatar, M. 2000

5 Allen, J.W. and Etcheberrigaray, R. 1998

6 Töpert, C. 1998

7 Hughes, B.A. 2000

8 Skatchkov, S.N. 2001

9 Morishige, K.-I. 1993; Horio, Y. 1996; Karschin, C. 1996; Bredt, D.S. 1995; Falk, T. 1995.

10 Karschin, C. 1999; Isomoto, S. 1997

11 Falk, T. 1995

12 Liao 1996; Miyashita and Kubo 1997; Signorini 1997

13 Horio, Y. 1996; Isomoto, S. 1997.

14 Karschin 1996

15 Mermelstein, P.G. 1998

16 van Vulpen, E.H.S. 1998; van der Kooy, D. 1987

17 Karschin, C. 1997

18 Karschin, C. 1999

19 Mermelstein, P.G. 1998, Nicola, S.M. 2000

20 Karschin, C. 1996

21 Di Chiara, G. 1994; Calabresi, P. 2000

22 Chuang, H. 1997

23 Scheinman, S.J. 1999

24 Patil, N. 1995

25 Jin, W. and Lu, Z. 1998

26 Lee, J.K. 1999

27 Thomzig, A. 2003

28 Haber, S.N. 2000; White, N. 1998

29 Voorn, P. 1996; Haber, S.N. 2000; Groenewegen, H.J. 1996; Joel, D. 2000; Jongen-Rêlo, A.L. 1994

30 Groenewegen, H.J. 1996; Haber, S.N. 2000



© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
15.06.2004