The Kinase Function of MSK1 Regulates BDNF Signaling to CREB and Basal Synaptic Transmission, But Is Not Required for Hippocampal Long-Term Potentiation or Spatial Memory

MSK1 KD mice display a deficit in basal synaptic transmission which is not due to increased GABAergic inhibition. Basal synaptic transmission was measured in the CA1 region of hippocampal slices from wild-type (black squares) or MSK1 KD (white circles) mice. A, Plot of fibre volley amplitude versus fEPSP slope demonstrates that MSK1 KD mice show reduced synaptic transmission at higher stimulus strengths. Inset is representative fEPSPs at 10 to 300 µA stimulus strengths for both genotypes (n = 27 pathways from 18 wild-type mice and 17 pathways from 12 MSK1 KD mice). This deficit is also observed when fEPSP slope is plotted against stimulus strength (B; ***p = 0.0001; genotype x stimulus strength interaction). In contrast, a plot of fibre volley amplitude against stimulus strength (C) showed no significant difference between the genotypes. D, The deficit in fEPSP was maintained when the experiments were repeated in the presence of 50 µM of the GABA_A receptor antagonist picrotoxin (n = 5 pathways from four wild-type mice and seven pathways from five MSK1 KD mice; **p = 0. 001; genotype x stimulus strength interaction). Error bars represent SEM.

Electrophysiological characterization of MSK1 KD mice shows no deficit in paired-pulse facilitation, mGluR-LTD, or LTP. A, Paired-pulse facilitation in area CA1 was measured in hippocampal slices prepared from adult wild-type (black squares; 21 pathways from 18 animals) or MSK1 KD (white circles, 22 pathways from 19 animals) mice. Representative families of paired fEPSPs taken over the 50 to 250 ms interstimulus interval range are shown in the upper panel, and quantification of the paired-pulse ratios in the lower panel. Data are expressed as mean and error bars (where visible) represent the SEM. B, mGluR-LTD was induced in area CA1 of hippocampal slices from wild-type (black squares) and MSK1 KD (white circles) mice. The GI mGluR agonist DHPG (100 µM) was applied from time 0-10 min (as indicated by the black bar on the graph) in the presence of both picrotoxin (50 µM) and the NMDA receptor glycine site antagonist L689,560 (5 µM). Inset are representative fEPSPs taken 10 min before and 50 min after the end of DHPG application in wild-type and MSK1 KD slices (n = 15 and 10 slices from 11 wild-type and 8 MSK1 KD mice, respectively). Data are expressed as mean, and error bars represent the SEM. C, D, Tetanus- (C) and theta-burst stimulation-induced LTP (D) were measured in the area CA1 of hippocampal slices from adult mice. Inset are representative fEPSPs taken 10 min before and 120 min after induction of LTP (at time zero) in the stimulated (left) and control (right) pathways for each genotype. Data are expressed as mean, and error bars represent the SEM. For tetanus-induced LTP, data are taken from single slices from nine wild-type and eight MSK1 KD mice, while for theta-burst-induced LTP, individual slices from seven wild-type and seven MSK1 KD mice were used. E, TBS-LTP is sensitive to the transcription inhibitor Act-D. Graphs show the effects of DMSO vehicle (0.08%) applied throughout the experiment (black squares and dark grey bar; n = 5; three wild-type and two MSK1 KD mice); 40 µM Act-D applied for the duration of the experiment (dark grey squares and dark grey bar; n = 5; four wild-type and one MSK1 KD mice); and 25 µM Act-D (0.05% DMSO) applied for 15 min before and after TBS (light grey squares and light grey bar; n = 9; three wild-type mice and six slices from four MSK1 KD mice). Both concentrations of Act-D affect LTP similarly, but the lower concentration and shorter duration of application of 25 µM has less of an effect on the control pathway. Data are expressed as mean, and error bars represent the SEM. F, The inhibitory effects of Act-D on late TBS-LTP are independent of genotype. Data are replotted from E but with respect to genotype instead of concentration: black squares, seven slices from seven wild-type mice; open circles, seven slices from five MSK1 KD mice. Data are expressed as mean, and error bars represent the SEM.

MSK1 KD mice display no deficit in the watermaze test for spatial reference memory. A, The ability to locate a visible platform in a 2-m pool was assessed by a 3-d visual cue task as described in Daumas et al. (2008). Both wild-type (WT) and MSK1 KD mice improved their latency to the platform with time at an equivalent rate. B, Spatial reference memory was assessed using a standard 5-d protocol (four spaced trials/day, 10-cm hidden platform, probe test after 24 h). No learning deficit was observed during training. C, The retention “probe” test indicated good memory in both groups, revealed as greater time spent searching in the training quadrant. Broken line, chance level = 25 %.

MSK1 KD mice display no deficit in the watermaze test for serial spatial learning. A, A serial spatial learning task was used to assess memory flexibility as described in Daumas et al. (2008). In this task, the platform location changes between up to eight separate locations and between days, with each location trained until the animals reached a fixed criterion of performance (latency < 20 s over three trials). The measure of learning now is not latency, but the number of trials (#) required to reach criterion. B, Both groups showed similar performance. After reaching criterion for each of five serial locations, memory was assessed at 10 min (C) and 24 h (D). The proportion of time spent in the relevant platform zone (extended to 20 cm in diameter) relative to the other seven possible platform locations was analysed (broken line, chance level = 12.5 %).