healthmedicinet

In the present study we report that GIRK1 is expressed in a majority of DRG neurons
of different sizes and thus associated with different modalities, whereas GIRK2 is
limited to a small subpopulation of nociceptors. Both are down-regulated after peripheral
nerve injury. These results provide histochemical support for involvement of GIRKs
in pain processing at the spinal level as reported in earlier studies 29], 30], 38], and complement a large number of studies showing a similar regulation, and presumable
function, of voltage gated K
⁺
channels, which also are down-regulated by peripheral nerve injury 31], 33], 54]–62].

GIRK1 and -2 mRNAs in DRGs

In the current study, we focused on the expression of GIRK1 and -2 in DRGs and spinal
cord under native and neuropathic pain conditions. Our qPCR and ISH results show that
GIRK1 and -2 mRNAs are expressed in rat DRGs, which is in agreement with a previous
study on rat based solely on RT-PCR 39]. The GIRK1 and -2 mRNA signals were robust (Ct values around 24 and 28, respectively,
data not shown). Also with ISH we could detect transcripts for GIRK1 and -2, both
of which undergo alternative splicing 3], 4].

GIRK1 and -2 proteins in DRGs

So far, little is known about the neurochemical phenotype of GIRK1 and -2-IR neurons
in rat DRGs and spinal cord. Here we observed a patchy, cytoplasmic and perinuclear
GIRK1-IR staining around the nucleus, suggestive of endoplasmic reticulum (ER) localization,
while a distinct association with the neuronal cell membrane was rarely observed.
This fits with the notion that GIRK1 lacks an ER export signal, and must associate
with another GIRK subunit for expression in the plasma membrane 8], 63]. Here GIRK2 would be suitable, which was extensively expressed in both cytoplasm
and on membranes, however in a fairly small population. Almost 73% of the GIRK2
⁺
and ~50% of the GIRK1
⁺
neurons were IB4
⁺
, indicating that nociceptors might express GIRK1/2 heterotetrameric channels. In
contrast, around 27% of the GIRK1
⁺
, but no GIRK2
⁺
, neurons expressed CGRP; thus it is unlikely that heterotetramers between these two
GIRKs exist in peptidergic nociceptors 40].

The large myelinated neurons, identified by NF200, give rise to A-fibers, including
A? fibres, and mostly subserve mechansensory functions 64]. In this cell population, GIRK1-LI was expressed in ~39% of the NPs throughout the
cytoplasm. GIRK2-LI was found in around one-third of these neurons, and the immunoreactivity
was always associated with the cell membrane. Thus, in these neurons functional GIRK1/2
heterotetramers may exist.

Further phenotyping showed that different small GIRK1
⁺
neuron subpopulations express all four CaBPs studied, the largest one expressing PV
(~30%). GIRK2
⁺
neurons only expressed CB. The functional implications of these results remain to
be resolved.

GIRK1 and -2 interaction with inhibitory GPCRs

G protein-induced opening of GIRK channels typically results in a several-fold increase
in potassium conductance at the resting membrane potential. GIRK opening is one of
the fastest metabotropic effector mechanisms, owing to its short signaling pathway,
with the G protein as the single link between the receptor and the channel. As such,
it well suited for rapid responses to neuromodulators 1].

Double immunofluorescence staining of DRG neurons showed that GIRK1 co-exists with
three neuropeptide receptors linked to inhibitory neurotransmission and neuropathic
pain: the Y1, SST1 and SST2A receptors 65]. All these receptors showed a clear membrane association, and in some cases an overlap
with GIRK1 staining was seen, possibly indicating functional interaction. GIRK2 was
only associated with SST1
⁺
neurons.

GIRK channels are downstream effectors of SST4 for its analgesic effects in rat DRGs
16], 66], and somatostatin may induce slow inhibitory, postsynaptic currents (IPSCs) through
activation of GIRK channels 12]. However, it has been reported that the SST1 receptor, in contrast to SST2-5, does
not open GIRK channels but might instead decrease their currents 67], 68]. Therefore, somatostatin may play different roles in different neuronal populations
in DRGs.

Furthermore, NPY may via Y1 receptors, and GIRK channels cause postsynaptic hyperpolarization
in tonic firing neurons in substantia gelatinosa 69]. Also in other regions of the central nervous system NPY acts via GIRK channels,
e.g., inhibitory effects of NPY in lateral amygdala have been shown to be mediated
through Y1 receptor and GIRK1, 2, and 3 70].

Also for galanin actions GIRK channels are important, since the anticonvulsant effect
of a GalR1 agonist was abolished by the GIRK channel inhibitor tertiapin Q 71]. In addition, galanin has been associated with GIRKs at the spinal level 72]. Unfortunately there is no reliable antiserum recognizing any of the three galanin
receptors, GalR1-3, but both GalR1 and GalR3 mediated inhibition via opening of potassium
channels 17], 73]. ISH has shown that 40% of the small and medium-sized NPs in rat DRGs express GalR1
mRNA (and ~75% GalR2) 74].

GIRKs have mainly been associated with postsynaptic receptors (see below), but DRG
neuron cell bodies are not considered to be ‘innervated’ by nerve endings and thus
do not seem to represent a postsynaptic structure. However, it was previously shown
that the neuropeptide substance P can be released from DRG neuron cell soma 75], and thus influences adjacent ‘postsynaptic’ DRG cell bodies in a paracrine way,
or acts on the releasing cell itself (‘autocrine’ signaling). We have proposed that
also NPY and galanin, especially after nerve injury-induced up-regulation, in a similar
fashion can be released from DRG cell soma 76], and thus be involved in chemically mediated cross-excitation in DRGs as proposed
by Amir and Devor 77]. Here, GIRKs expressed at the DRG neuron cell membrane may be involved in NPY Y1
and GalR1 signaling.

GIRK1 and -2 in spinal cord

In lamina II of mouse dorsal spinal horn, GIRK1- and -2-LI have previously been detected
in excitatory interneurons (“almost exclusively in postsynaptic membranes”) by immunoelectron
microscopy 28], 37], 38]. However, expression of these subunits has not been explored in the rat spinal cord.
Here, we demonstrate presence of a dense plexus of GIRK1
⁺
and -2
⁺
neuronal processes and cell bodies in, mainly, lamina II in the rat dorsal horn, as
well as in nerve terminals in the skin. The decrease in staining in the dorsal horn
after axotomy and rhizotomy, suggests that GIRK
⁺
processes not only originate from local neurons but also represent primary afferents,
i.e. GIRK1 and -2 have a presynaptic localization. This is also supported by the co-localization
of GIRK1 or -2 with VGLUT1 in a few boutons in the dorsal horn, since VGLUT1 in this
region is only presented in, even if few, primary afferents 78]. Further support comes from the fact that the GIRKs accumulate around a ligation
of the sciatic nerve, i.e. the channels are transported in axons.

A presynaptic localization has been shown in some studies 79], 80]. Ladera et al. reported, using immunogold electron microscopy, presence of GABA
_B
receptors and GIRK2 and -3 in presynaptic boutons in cerebral cortex, and a GABA
_B
-mediated reduction in glutamate release that could be reversed by the GIRK channels
blocker tertiapin-Q 80]. A similar situation exists in the cerebellum, here involving parallel fibers and,
again, GABA
_B
receptors 11]. Taken together, GIRKs may play a role in presynaptic control of excitatory signaling
also in the dorsal horn (and skin) and via such mechanisms influence pain signaling.

GIRK1-IR cell bodies were also found in other spinal layers, some multipolar neurons
in lamina V and white matter and fusiform neurons in the deep layers. Thus, GIRK1
⁺
neurons exhibit diverse morphological properties in deep dorsal horn layers. They
may also represent projection neurons. In addition, an extensive expression of GIRK1-LI
in the ventral horn was seen from lamina VI to X, some of which were co-localized
with CGRP and thus represent motor neurons. Similarly, GIRK2-IR cells bodies were
found in different layers of the spinal dorsal horn, but none in the ventral horn.

Inhibitory interneurons constitute 30–40% of neurons in rat lamina I–III, many of
which serve an important anti-nociceptive function, and virtually all are GABAergic,
some using glycine as co-transmitter 81]–83]. Around half of the inhibitory interneurons in lamina I–II possess SST2A 52], all of which are GABA
⁺
in lamina I–II of rat 84]. Here, we did not find any SST2A
⁺
interneurons in rat lamina I–II that were GIRK1
⁺
, and only a few were GIRK2
⁺
. Thus, rat GIRK interneurons may in general, as in mouse, be excitatory. Also, GIRK
channels may be of limited importance in pain signaling through somatostatin receptors
at the spinal level, although GIRK3 and -4 should be considered. In agreement, GIRK
channels influence signaling from myelinated low-threshold mechanical nociceptive
afferents as well 85].

Axonal transport of GIRK1 and -2

GIRK1- and -2-LIs were observed in control rat sciatic nerve at apparently comparable
levels. This is somewhat surprising, since there are almost 10 times as many GIRK1
⁺
as GIRK2
⁺
NPs in the DRGs. Sciatic nerve ligation caused an accumulation of both GIRKs on the
proximal and distal side of the lesion, indicating both antero- and retrograde axonal
transport of the channels. We also found GIRK1
⁺
and GIRK2
⁺
fibers in the dermis layer of hind paw. Twenty-one days after dorsal rhizotomy, there
was a reduction of both GIRK1- and -2-IR processes in the spinal dorsal horn, providing
further support also for central centrifugal transport from the DRG cell bodies.

Regulation of GIRK1 and -2 by peripheral nerve injury

Persistent hyperexcitability of peripheral nociceptors caused by changes in receptor
and ion channel activity is an important mechanism in chronic pain conditions 33], 86]–88]. In the peripheral sensory system, ion channels like Na
⁺
and Ca
²⁺
channels, contribute to excitation of sensory neurons under chronic pain conditions,
either by up-regulation or enhancement of activity 89], 90]. K
⁺
channels set the resting membrane potential in neurons, and thus control the excitability
of sensory neurons under physiological conditions 1]. Here, we found that nerve injury caused a significant down-regulation of the K
⁺
channels GIRK1 and -2 in DRG neurons, both at the mRNA and protein levels, as well
as in the dorsal horn, 14 days after peripheral nerve injury.

These results are in line with previous studies showing that other types of K
⁺
channels also are down-regulated after various types of peripheral nerve injury 31], 33]. An immunohistochemical analysis of rats subjected to the Chung model of neuropathic
pain has demonstrated expression of various Kv1 family subunits in distinct DRG neuron
populations, including nociceptors, as well as a distinct reduction in channel protein
levels 55]. Moreover, axotomy decreased mRNA levels for certain Kv1 subunits with up to 80%,
in parallel decreasing K
⁺
currents 54]. These findings have in general been interpreted to show that a reduction of voltage
gated K
⁺
channel activity after trauma causes nerve hyperexcitability, and thus may contribute
to peripheral and perhaps also central mechanisms underlying neuropathic pain 33]. The present and other studies suggest that a similar scenario then can be advanced
for GIRK1 and -2. The conductance of inwardly rectifying potassium channels is greatest
at membrane potentials close to the resting voltage 1]. Hence, GIRK channels are thought to be a key regulator of excitability, increasing
the stimulation current needed for eliciting an action potential. In contrast, other
K
⁺
channels which open at more depolarized voltages, have less impact on the firing threshold
91].

It has been suggested that many of the changes in protein expression in DRGs after
nerve injury serve to counteract pain 47], for example, the up-regulation of the inhibitory transmitters galanin and NPY, and
the down-regulation of excitatory neuropeptides CGRP and substance P result in attenuated
pain signaling in the dorsal horn. Thus, this apparently contrasts the consequences
of the decreased expression of various K
⁺
channels which enhance pain sensation. However, it may be speculated that a reduced
number of presynaptic GIRK1 channels may allow facilitated release of up-regulated
pain-inhibiting molecules like galanin from primary afferents in the dorsal horn and
thus contribute to pain relief. In summary our data show a window of opportunities
to restore or enhance inhibitory signaling by locally produced analgesic peptides
by restoring or enhancing downstream GIRK mediated suppression of excitability.