46 A
the compensation of the static symptoms, the biochemical selective CaMKII inhibitor was not employed. The present
basis of this resting activity recovery, and the other experiment attempted to further investigate the role of
neuronal changes responsible for compensation, are poorly protein kinases in static compensation by comparing the
understood see Refs. [6,38] for a review. effects of PKC and CaMKII inhibitors, delivered into the
21
Protein kinase C PKC and Ca calmodulin-depen-
ipsi-VNC at the time of the UVD in guinea pig. dent kinase II CaMKII are major serine threonine pro-
tein kinases found in particularly high concentrations in brain [42]. There is good evidence that these kinases play
2. Methods
important roles in several types of neuronal plasticity e.g., [3,13]; see Ref. [40] for a recent review. A number of
2.1. Animals studies have shown that vestibular compensation in frog
and guinea pig is accompanied by changes in the phos- Fifty-four adult male and female pigmented guinea pigs
phorylation of specific brain proteins in vitro [10,22– Cavia porcellus; 290–725 g were used in these experi-
24,35]. In guinea pig, the phosphorylation of at least one ments, twenty four each for the PKC inhibitor and CaMKII
probable PKC substrate in the bilateral medial vestibular inhibitor experiments in which cannulae were placed
nucleus prepositus hypoglossi, increased following com- stereotaxically in the right VNC see Table 1. In addition,
pensation of the static symptoms compared to sham-oper- data for the two doses of the PKC inhibitor were obtained
21
ated controls [35]. This contrasts with a lack of Ca from six animals n53 for each dose in which the cannula
calmodulin-stimulated protein phosphorylation changes in was placed in the IVth ventricle as an additional control.
the same tissue. There is also some evidence that the in Although the animals used in these experiments varied in
vitro phosphorylation changes that occur in frog whole weight, they were all adult guinea pigs and in our previous
brains, removed at different stages of compensation, are studies we have not observed differences in vestibular
mediated by an endogenous PKC and CaMKII [21,22]. compensation for adult animals of different weight [39].
These results suggest that the phosphorylation state of specific endogenous proteins in areas of the brain involved
2.2. Cannula implantation in processing vestibular information, is changed during
vestibular compensation in vivo. Several potential mecha- All surgical procedures were approved by the University
nisms could account for these changes, including de novo of Otago Committee on Ethics in the Care and Use of
synthesis of substrate proteins, and or a change in kinase Laboratory Animals. The cannula implantation procedure
activity. Direct evidence that the initial stages of static has been described in detail elsewhere [34]. Briefly,
compensation in guinea pig are not associated with a animals were anaesthetised with fentazin 0.4 ml kg, i.m.;
change in de novo protein synthesis was recently reported fentazin contains 0.4 mg ml fentanyl citrate, 58.3 mg ml
by Ris et al. [31], who found that systemic injections of a xylazine HCl and 3.2 mg ml azaperone; Parnell Lab-
protein synthesis inhibitor cycloheximide had no effect oratories, Auckland and a 21-gauge metal cannula was
upon the early stages of resting activity recovery in the lowered through the cerebellar vermis into the right VNC
ipsi-VNC of the guinea pig. However, the possibility that later stages of compensation might involve de novo protein
Table 1
synthesis cannot be excluded [4]. The hypothesis that
Experimental conditions
mammalian vestibular compensation might involve a
Inhibitor Concentration
Total dose N
change in PKC activity is further supported by recent
mM nmol
immunohistochemical experiments, which have shown that
PKC Inhibitors
early stages of compensation in rat are accompanied by
Bis I 5
0.02 4
changes in the expression of vestibulo-cerebellum PKC-
50 0.2
4 200
0.8 2
alpha, -gamma and or -delta [2,12] or VNC PKC-delta
Bis V 5
0.02 4
[28]. Taken together, these data suggest that vestibular
50 0.2
4
compensation of the static symptoms is correlated with
200 0.8
2
changes in PKC CaMK pathways in the VNC and or
Vehicle –
– 4
cerebellum; however, it is not clear whether these changes
5 DMSO in mACSF CaMKII Inhibitors
are causally involved in the compensation process or
myr-AIP 10
0.04 4
merely coincident with it.
100 0.4
4
Despite the correlative evidence implicating protein
500 2.0
3
phosphorylation in the VNC in the mechanisms of vestibu-
AIP cell impermeable 10
0.04 3
lar compensation [23,35], to date there has been only one
100 0.4
4 500
2.0 3
study in rat [1] which has examined the effects of kinase
Vehicle –
3
inhibitors on the compensation process, and in this case
mACSF
drugs were delivered into the IIIrd ventricle only and a
A .J. Sansom et al. Brain Research 882 2000 45 –54
47
complex stereotaxic co-ordinates: 12.5 mm caudal; 1.5 autocamtide-2 related inhibitory peptide AIP; N-Lys-Lys-
mm lateral; 10.8 mm ventral to the skull surface and Ala-Leu-Arg-Arg-Gln-Glu-Ala-Val-Asp-Ala-Leu-OH; MW
secured to the skull with dental cement. Animals n56 1498, were purchased from Biomol Research Laboratories
with IVth ventricle placements were used as additional USA. These peptides are analogs of the synthetic CaM-
controls; in these animals the cannula had penetrated the KII substrate, autocamtide-2. myr-AIP is reputedly cell-
cerebellum and there was clear evidence of dye infusion permeable due to the addition of a myristoyl group a
see below in that region, suggesting that the drug may saturated fatty acid on the N-terminus [16,19,20]. Since
have also diffused to some extent into the cerebellum. AIP lacks this myristol group it was used as a cell-
Some animals displayed transient SN and mild postural impermeable control. The concentrations of AIP used in
asymmetries during the first few hours after cannula the present experiments were based upon those used in an
implantation. As a necessary part of the cannula implanta- in vitro phosphorylation assay of post-synaptic density
tion procedure, some part of the cerebellum and VNC must proteins [15], since to the best of our knowledge, there are
be lesioned in order to position the cannula in the VNC. In no published reports of its use in vivo. Each peptide was
our experience, transient vestibular symptoms indicate reconstituted in sterile filtered millipure water to a stock
only a minor lesion and our previous studies indicate that solution of 1000 mM, and stored as frozen aliquots at
such minor damage, while confirming that the cannula is in 2808C for a maximum of 1 month to ensure optimal
the right place, does not significantly affect vestibular stability. On the day of each experiment, frozen aliquots
function thereafter [5,34]. However, animals which failed were thawed at room temperature and diluted to working
to compensate for these symptoms in less than 24 h were concentrations with sterile mACSF pH 7.0. Vehicle
excluded from the experiment. controls for both the Bis 5 DMSO in mACSF and AIP
mACSF groups were run in the same way as the 2.3. Drugs
experimental groups see Table 1. The PKC inhibitor 3-[1-3-dimethylaminopropyl-1H-
2.4. Drug delivery indol-3 -yl]-4-1H-indol-3-yl-1H-pyrrole-2,5-dione, HCl
bisindolylmaleimide I, HCII GF 109203X, HCl and the Guinea pigs were randomly assigned to one of four
less selective
analogue, 2,3-bis1H-indol-3-yl-N-
experimental groups see Table 1. Two 25-ml Hamilton methylmaleimide bisindolylmaleimide V, were purchased
syringes were attached to a 1m length of polyethylene from LC Laboratories USA. Bisindolylmaleimide I HCl
tubing and placed into a syringe pump Harvard Instru- Bis I is selective for PKC in the nM range in vitro
ments, MA, USA. Distal to the syringes, each infusion IC 531 nM [41] but will inhibit other protein kinases at
line was attached to an infusion needle constructed of
50
much higher concentrations; for example, the IC ’s of Bis 27-gauge stainless steel tubing. Drug or vehicle solutions
50
I for phosphorylase kinase, PKA, the endothelial growth were backfilled into the tubing prior to the infusion needle
factor receptor and the insulin receptor are: 0.7 mM, 2 mM, being placed inside the guide cannula in the animal. The
.100 mM, .100 mM, respectively [26,41]. In contrast, the syringe pump delivered drug solutions at a constant rate of
affinity of bisindolylmaleimide V Bis V for PKC is 2 ml h for a total of 2 h, beginning 1 h pre-, and finishing
greatly reduced IC .100 mM [8] and was used as a 1 h post-UVD. The total infusion volume was verified after
50
negative control for possible non-specific effects of Bis I. each infusion. The group mean infusion volumes ranged
The doses of Bis I were chosen based upon data from from 4.21 to 4.34 ml. Each animal was given a single dose
whole cell preparations in vitro [40] see Ref. [16] for a of the sedative muscle relaxant, xylazine 12 mg kg, i.m.,
review. Bis I and V were dissolved in 100 dimethylsul- 10 min prior to the start of the cannula infusion. Approxi-
phoxide DMSO to stock solutions of 1 mM and stored at mately 30 min after the start of the infusion, animals were
2208C. On the day of each experiment, frozen stock anesthetised for UVD surgery. For PKC inhibitor experi-
solutions were diluted to working concentrations with ments, all drug and vehicle solutions were administered
sterile filtered
modified artificial
cerebrospinal fluid
blind; solutions were coded by a third party and codes mACSF [34] and pH adjusted to approximately 7.0 if
were not revealed to the experimenter until after the required with 0.1 M HCl. The concentration of DMSO in
experiments were completed. Drug and control conditions working solutions was 5, except for 200 mM solutions in
were run together on the same day where possible. which case the DMSO concentration was raised to 10
CaMKII inhibitor experiments were not run blind because due to solubility problems. For this reason, the 200 mM
AIP peptides have a limited shelf-life in solution approxi- groups were not compared with other dose groups in
mately 1 month at 2808C, therefore control and ex- statistical analysis.
perimental conditions could not be run in parallel. The pseudosubstrate CaMKII inhibitors, myristoylated
Since parts of the VNC border on the IVth ventricle, we autocamtide-2 related inhibitory peptide myr-AIP; N-Myr-
took the following precautions to minimise drug spread Lys-Lys-Ala-Leu-Arg-Arg-Gln-Glu-Ala-Val-Asp-Ala-Leu-
from the injection site into the IVth ventricle, surrounding OH; MW 1708 and the non-myristoylated analogue,
brainstem nuclei and cerebellum. First, the slow rate of
48 A
drug infusion 2 ml h means that drug concentrations at In addition, no statistical comparisons were made between
sites distant from the target nuclei will be low compared to Bis I and V at the 200 mM dose, because of the small
the injection site [36,44]. The infusion rate used in this sample sizes involved n52 per group.
study is comparable to that of some osmotic minipumps. Second, the diameter of the injection needle and inner
2.8. Histology cannula were identical in order to minimise drug backflow
up the sides of the injection needle. Cannula tip positions were verified for each animal in
these experiments. At the conclusion of the experiment, 2.5. Unilateral vestibular deafferentation UVD
animals were given an overdose of pentobarbital followed by a 0.5–1.0 ml cannula injection of dye Alcian blue
One week following cannula surgery, all animals were 8GX, 5 w v to mark the infusion site. Animals were
given a right global surgical labyrinthectomy UVD, as then transcardially perfused with 0.9 saline followed by
described in detail elsewhere [34]. Briefly, guinea pigs a phosphate-buffered formaldehyde 10 solution. The
were anesthetised and prepared for surgery in an identical brainstem and cerebellum were removed and stored in a
manner to that described for the cannula surgery above. A formaldehyde solution see above until sectioned. The
horse-shoe-shaped incision was made over the right ear fixed tissue was mounted on a microtome stage, frozen
and the temporal bone exposed following blunt dissection with CO and 80–90 mm coronal sections cut serially with
2
of the temporalis muscle and its subsequent removal with a Leitz microtome. The coronal sections were mounted on
surgical scissors. The bony labyrinth was exposed with a gelatin-coated slides and air dried for 24 h before staining
high speed drill, the horizontal and anterior semicircular with thionin blue and coverslipping. At least one slide
canal ampuliae and otoliths were subsequently destroyed from each animal was photographed Kodak Ektachrome
under microscopic control using a fine burr, and their 64T film.
contents aspirated. Our previous studies have shown that this UVD procedure results in complete destruction of the
vestibular receptors [5,11,23,33–35,39].
3. Results
