Biphasic requirement for geranylgeraniol in hippocampal long-term potentiation

Abstract

Mice deficient in cholesterol 24-hydroxylase exhibit reduced rates of cholesterol synthesis and other non-sterol isoprenoids that arise from the mevalonate pathway. These metabolic abnormalities, in turn, impair learning in the whole animal and hippocampal long-term potentiation (LTP) in vitro. Here, we report pharmacogenetic experiments in hippocampal slices from wild-type and mutant mice that characterize the dependence of LTP on the non-sterol isoprenoid, geranylgeraniol. Addition of geranylgeraniol to slices from 24-hydroxylase knockout mice restores LTP to wild-type levels; however, farnesol, a chemically related compound, does not substitute for geranylgeraniol nor does another animal model of impaired LTP (apolipoprotein E deficiency) respond to this isoprenoid. The requirement for geranylgeraniol is independent of acute protein isoprenylation as judged in experiments employing cell-permeable inhibitors of protein farnesyl transferase and geranylgeranyl transferase enzymes and in mutant mice hypomorphic for geranylgeranyltransferase II. Time course studies show that geranylgeraniol acts within 5 min and at 2 different times during the establishment of LTP: just before electrical stimulation and approximately 15 min thereafter. Localized delivery of geranylgeraniol to the dendritic trees of CA1 hippocampal neurons via the recording electrode is sufficient to restore LTP in slices from 24-hydroxylase knockout mice. We conclude that geranylgeraniol acts specifically and quickly to affect LTP in the Schaffer collaterals of the hippocampus.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Schematic of cholesterol synthesis and turnover in the brain showing selected pathway intermediates and inhibitors used in this study. Statins inhibit HMG CoA reductase early in the biosynthetic pathway, whereas FTI-277, GGTI-2133, and 3-PEHPC block the addition of isoprenyl groups to protein targets. The 24-hydroxylase turns over cholesterol by converting it to 24(S)-hydroxycholesterol, a membrane-permeable end-product. Synthesis of new cholesterol to replace that turned over ensures a constant flow of intermediates through the pathway, one of which, geranylgeranyl diphosphate or a derivative of this polyisoprenoid, is required for LTP. Apolipoprotein E (ApoE) facilitates the movement of cholesterol between different cell types, either as a component of lipoprotein particles or in the form of ApoE-cholesterol complexes.

Fig. 2.

Protein isoprenylation inhibitors do not reduce LTP in hippocampal slices from gunmetal mice that are hypomorphic for geranylgeranyltransferase II. (A) LTP in slices from gunmetal mice (●, n = 20) was indistinguishable from that in wild-type slices (○, n = 19). (B) Treatment of hippocampal slices from gunmetal mice with ACSF alone (●, n = 20) or ACSF containing 10 μM GGTI-2133 (○, n = 19), an inhibitor of geranylgeranyltransferase I (GGTI), did not reduce LTP. Inhibitor administration started at time 0 of the experiment. (C) Treatment of hippocampal slices from gunmetal mice with ACSF alone (●, n = 20) or ACSF containing 1.5 mM 3-PEHPC (○, n = 18), an inhibitor of geranylgeranyltransferase II, did not reduce LTP. (D) Treatment of hippocampal slices from gunmetal mice with an inhibitor of HMG CoA reductase (statin, 12.5 μM) reduced LTP (○, n = 25); this inhibition was prevented by inclusion of 0.2 mM geranylgeraniol in the ACSF (●, n = 21).

Fig. 3.

Mice heterozygous for an induced mutation in the 24-hydroxylase gene show an intermediate phenotype in LTP and spatial learning assays. (A) LTP was induced by theta-burst stimulation in hippocampal slices from mice of the indicated 24-hydroxylase genotypes. Slices from animals homozygous for the mutation (−/−, ●, n = 14) showed reduced LTP compared to those from wild-type animals (+/+, ○, n = 18), while slices from heterozygotes (+/−, shaded circles, n = 14) showed an intermediate phenotype in which LTP was induced normally at tetanization but began to decline at approximately 40 min after the start of the experiment and by approximately 60 min reached the level of potentiation observed in homozygous slices. (B) Inclusion of 0.2 mM geranylgeraniol (GG) in the ACSF throughout the experiment restored LTP in heterozygous slices (○, n = 12). (C) Spatial learning in mice of the indicated 24-hydroxylase genotypes was assessed in Morris water maze tests. Mean time to find a submerged platform (latency) is shown as a function of trial day. Data were collected in four different experiments involving 40 +/+, 20 +/−, and 30 −/− male mice. Results from the wild-type and homozygous mutant animals were reported previously (9), and are reproduced here for comparison purposes. (D) Probe trial responses. The submerged platform was removed from the tank on the indicated days of the Morris water maze experiments and the number of times animals of different 24-hydroxylase genotypes swam across the former location of the platform was recorded. Differences between wild-type and heterozygous mice were statistically significant (P <0.05) on each of the four probe days.

Fig. 4.

Brief treatment with geranylgeraniol restores LTP in hippocampal slices from 24-hydroxylase heterozygous mice. (A) Slices from heterozygous mice (+/−) were treated with ACSF (●, n = 14) or ACSF supplemented with geranylgeraniol (GG) (○, n = 9) for the 5 min interval (time = 35–40 min) marked by the open rectangle. The transient presence of geranylgeraniol prevented the decline in potentiation observed between 40 and 50 min in heterozygous slices treated with ACSF alone. (B) Geranylgeraniol treatment delivered between time = 50–55 min did not maintain LTP in heterozygous slices (●, n = 11). Data from an experiment with wild-type slices (+/+, n = 15) are shown for comparison purposes. (C) Hippocampal slices from wild-type mice (+/+) were treated with ACSF alone (○, n = 12) or ACSF supplemented with an HMG CoA reductase inhibitor (statin, ●, n = 18) for the 15 min interval (time = 30–45 min) marked by the closed rectangle. The presence of the drug during this interval caused a decline in potentiation. (D) Slices (●, n = 11) from wild-type mice were treated with statin before, during, and after high-frequency stimulation (black rectangle, time = 17–22 min). The presence of the inhibitor during this period prevented the initiation of LTP.

Fig. 5.

Brief treatment with geranylgeraniol restores LTP in hippocampal slices from 24-hydroxylase homozygous mice. (A) Slices from homozygous mice (−/−) were treated for 5 min with geranylgeraniol (GG) during tetanization (●, n = 13, closed rectangle, time = 17–22 min). Potentiation was induced transiently in slices by this treatment but decayed over the course of the experiment versus that measured in wild-type slices (○, n = 8). The pattern observed resembled that measured in slices from untreated heterozygous mice (e.g., Fig. 4A). (B) Treatment of hippocampal slices (○, n = 10) from knockout mice with geranylgeraniol during stimulation and at time = 35–40 min produces a wild-type level of LTP. Little potentiation was observed in untreated slices (●, n = 18). (C) Geranylgeraniol treatment at time = 35–40 min does not affect potentiation in knockout slices (○, n = 11) versus ACSF-alone treated controls (●, n = 17).

Fig. 6.

Presence of geranylgeraniol, statin, or prenyltransferase inhibitors in recording electrode restores, inhibits, or has no effect on LTP, respectively. (A) LTP was measured in hippocampal slices from 24-hydroxylase knockout mice (−/−) using a recording electrode containing ACSF (●, n = 22) or ACSF supplemented with 0.2 mM geranylgeraniol (GG) (○, n = 10). LTP in the knockout slices was similar to that measured in wild-type slices. (B) LTP was measured in slices from wild-type mice (+/+) using a recording electrode containing ACSF (○, n = 14) or ACSF supplemented with 12.5 μM statin (●, n = 11). Delivery of the inhibitor via the recording electrode decreased LTP to levels observed in untreated knockout slices. (C) LTP was measured in wild-type slices using a recording electrode containing ACSF (○, n = 12), ACSF supplemented with 10 μM FTI- 277 (shaded circles, n = 11), or ACSF supplemented with 10 μM GGTI-2133 (●, n = 12). Neither prenyltransferase inhibitor affected potentiation.