Input-specific regulation of hippocampal circuit maturation by non-muscle myosin IIB

Abstract

Myh9 and Myh10, which encode two major isoforms of non-muscle myosin II expressed in the brain, have emerged as risk factors for developmental brain disorders. Myosin II motors regulate neuronal cytoskeletal dynamics leading to optimization of synaptic plasticity and memory formation. However, the role of these motor complexes in brain development remains poorly understood. Here, we disrupted the in vivo expression of Myh9 and/or Myh10 in developing hippocampal neurons to determine how these motors contribute to circuit maturation in this brain area important for cognition. We found that Myh10 ablation in early postnatal, but not mature, CA1 pyramidal neurons reduced excitatory synaptic function in the Schaffer collateral pathway, whereas more distal inputs to CA1 neurons were relatively unaffected. Myh10 ablation in young neurons also selectively impaired the elongation of oblique dendrites that receive Schaffer collateral inputs, whereas the structure of distal dendrites was normal. We observed normal spine density and spontaneous excitatory currents in these neurons, indicating that Myh10 KO impaired proximal pathway synaptic maturation through disruptions to dendritic development rather than post-synaptic strength or spine morphogenesis. To address possible redundancy and/or compensation by other Myosin II motors expressed in neurons, we performed similar experiments in Myh9 null neurons. In contrast to findings in Myh10 mutants, evoked synaptic function in young Myh9 KO hippocampal neurons was normal. Data obtained from double Myh9/Myh10 KO neurons largely resembled the MyH10 KO synaptic phenotype. These data indicate that Myosin IIB is a key molecular factor that guides input-specific circuit maturation in the developing hippocampus. Non-muscle myosin II is an actin binding protein with three isoforms in the brain (IIA, IIB and IIC) encoded by the myh9, myh10, and myh14 genes in mice, respectively. We have studied the structure and the function of hippocampal CA1 neurons missing NMIIB and/or NMIIA proteins at different times during development. We have discovered that NMIIB is the major isoform regulating Schaffer collateral inputs, and that this regulation is restricted to early postnatal development.

Keywords: hippocampus; myh10; myh9; myosin; neurodevelopment; spines.

Figure 1. NM IIB ablation disrupts baseline excitatory synaptic function in developing, but not in adult CA1

(A) Experimental design of the study. (B) Representative western blots from hippocampi of PND8 and PND14 Myh10^fl/fl animals injected with 1μl AAV-Cre at PND1 show reduction in relative Myh10 expression compared to control animals ( PND8 t(8)=3.85, p=0.005; PND14 t(7)=6.32, p=0.0004). Representative DIC and fluorescent images of (C) PND1 and (G) PND42 injections show td tomato expression in the hippocampus which reports Cre expression. (D) Representative field recording traces and (E) I/O curves show reduction in synaptic transmission in PND1-injected Myh10^fl/fl hippocampus compared to wild-type (t(36)=2.29, p=0.028; n=22 wt, n=16 ko slices from n=5 wt, n=4 ko mice) . (F) Normal paired pulse ratio (interpulse interval 100 ms) in PND1 injected animals (t(28)=0.71, p=0.48, n=16 wt, n=14 ko slices from n=5 wt, n=4 ko mice). (H) Representative field recording traces and (I) I/O curves show normal synaptic transmission in PND42-injected Myh10^fl/fl hippocampus compared to wild-type (t(21)=0.41, p=0.69, n=11 wt, n=12 ko slices from n=3 wt, n=3 ko mice). (J) Normal paired pulse ratio (interpulse interval 100 ms) in PND42-injected animals (t(21)=0.50, p=0.62, n=11 wt, n=12 ko slices from n=3 wt, n=3 ko mice ). Traces reflect responses at 30% of maximal stimulation intensity. Error bars represent SEM.

Figure 2. NM IIB ablation cell-autonomously disrupts the pathway-specific development of synaptic excitation in CA1 neurons

(A) The illustration shows the design of these experiments with electrode replacement either in proximal (C, I) or in distal (F) stimulation locations in separate sets of experiments. (B) Representative DIC and fluorescent images of Myh10^fl/fl hippocampus show sparse Cre expression. Representative AMPA (inward, recorded at −70 mV) and NMDA currents (outward, recorded at +40 mV) from (C, I) Myh10^fl/fl and (F) Myh10+/+ animals injected with Cre. Arrows point to where AMPA and NMDA measurements were made. Recordings from neighboring Cre negative and Cre positive neurons (Myh10 KO) show (D) reduced AMPA (paired t(20)=2.33, p=0.03, n=21 pairs from n=9 mice ) and (E) reduced NMDA currents (paired t(14)=2.16, p=0.048, n=15 pairs from n=7 mice) at proximal stimulation locations, but (G) normal AMPA (paired t(17)=0.99, p=0.33, n=18 pairs from n=3 mice) and (H) increased NMDA currents (paired t(15)=2.57, p=0.021, n=16 pairs from n=3 mice) at distal stimulation locations. (I-K) Control experiments at proximal locations show that Cre positive neurons (wt/td) have (G) normal AMPA (paired t(15)=0.50, p=0.63, n=16 pairs from n=5 mice) and (H) normal NMDA currents (paired t(11)=0.11, p=0.92, n=12 pairs from n=5 mice) from Myh10^+/+ animals. Dashed black line is identity line. Filled circle is the sample mean. Error bars represent SEM.

Figure 3. NM IIB ablation selectively disrupts the development of proximal dendritic branches in CA1 neurons

(A) Photomicrograph multiphoton image of a portion of hippocampal CA1 showing labeled pyramidal neurons. (B-C) Representative 3D pyramidal hippocampal cell reconstruction using Neurolucida software. Neuronal complexity was measured applying the Sholl ring analysis method; (D) Bar graphs showing surface extension of traced apical arbors in both wt (14 neurons, n= 7 slices, n=5 mice) and ko (12 neurons, n= 6 slices, n=4 mice); p < 0.05. (E-F) Graphs exhibiting arbor complexity of pyramidal neurons in both wt and ko, measured as a number of intersections [RMANOVA, F1,12 = 2.4170; p < 0.01]) or cumulative length [RMANOVA, F(1,12) = 2.0649; p < 0.05] in relation to the distance from soma (insets depicting the total number of intersections and the total length, respectively); A LSD post-hoc test was applied where appropriate; # p<0.05 wt vs. ko comparisons. Error bars represent SEM.

Figure 4. NM IIB ablation has no effect on spine density or postsynaptic strength

(A-C) Example traces (A), cumulative probability distribution of amplitude (B) and frequency (C) of spontaneous EPSC recordings indicate slightly reduced frequency of sEPSCs in the absence of NMIIB. (K-S test on amplitude Z=0.95, p=0.33; Bar graph inset for amplitude, Student's t-test, t(33)=0.22, p=0.83; K-S test on frequency Z=1.58, p=0.013; n=18 cre−, n=17 cre+ from n=6 mice; Bar graph inset for frequency, Student's t-test t(33)=0.90, p=0.38) (D-F) Example traces (D), cumulative probability distribution of amplitude (E) and frequency (F) of miniature EPSC recordings indicate normal mEPSCs in the absence of NMIIB (K-S test on amplitude Z=0.940, p=0.34; Bar graph inset for amplitude, Student's t-test, t(26)=0.1, p=0.92 ; K-S test on frequency Z=0.664, p=0.77; n15 cre−, n=13 cre+ from n=7 mice; Bar graph inset for frequency. Student's t-test, t(26)=0.1, p=0.93). (G, K) Representative examples of proximal (G) and distal (K) dendrites from wt and ko CA1 neurons. (H, L) Bar graphs show normal spine density in proximal (H) and distal (L) dendrites of Myh10 ko CA1 neurons (Proximal branches, wt = 9 cells, 4 mice, ko = 10 cells, 4 mice. [Student t-Test, t(17)=0.1397 p=0.6087]; Distal branches wt = 10 cells, 2 mice, ko = 11 cells, 3 mice [Student t-Test, t(19)=0.1909, p=0.8506]) (I,M) Cumulative frequency curves show normal spine width in proximal (I) and in distal (M) dendrites of Myh10 ko neurons. (Proximal branches, wt = 1512 spines, 4 mice; ko = 1512 spines, 4 mice. [K-S test, p > 0.05]; Bar graph inset for spine width, Student t-test [t(6)=0.370, p=0.7237]; Distal branches wt = 2507 spines, 2 mice; ko = 2507 spines, 3 mice [K-S test, p < 0.005]; Bar graph inset for spine width, Student t-test [t(3)=0.389, p=0.7225]). (J, N) Cumulative frequency curves show normal spine length in proximal (J) and increased spine length in distal (N) dendrites of Myh10 ko neurons. (Proximal branches, wt = 1512 spines, 4 mice; ko = 1512 spines, 4 mice. [K-S test, p < 0.001]; Bar graphs inset for spine length, Student t-test [t(6)=1.469, p=0.192]) (Distal branches wt = 2507 spines, 2 mice; ko = 2507 spines, 3 mice [K-S test, p < 0.001]; Bar graph inset for spine length, Student t-test [t(3)=3.681, p=0.034]). Error bars represent SEM.

Figure 5. NM IIA ablation has no impact on excitatory synaptic function in CA1

(A-D) Representative western blots show no reduction in Myh9 levels in hippocampus (A) (t(6)=0.80, p=0.45)(A), but robust reduction in liver (B) (t(6)=8.95, p=0.0001). Glial enriched cultures show higher levels of Myh9 (C) (t(4)=6.42, p=0.003). Driving Cre expression in hippocampus using CMV promoter results in Myh9 knockdown (D) (t(4)=4.21, =0.014). (E) Representative AMPA (inward, recorded at −70 mV) and NMDA currents (outward, recorded at +40 mV) from Myh9^fl/fl animals injected with Cre. Recordings from neighboring Cre negative and Cre positive neurons (Myh9 KO) show (F) normal AMPA (paired t(30)=0.68, p=0.50, n=31 pairs from n=8 mice) and (G) normal NMDA currents (paired t(30)=0.70,p=0.49, n=31 pairs from n=8 mice) at proximal stimulation locations. Dashed black line is identity line. Filled circle is the sample mean. (H-J) Example traces (H), cumulative probability distribution of amplitude (I) and frequency (J) of spontaneous EPSC recordings indicate normal synaptic function in the absence of NMIIA (K-S test on amplitude Z=0.97,p=0.30; Bar graph inset for amplitude, Student's t-test t(23)=1.42, p=0.17; K-S test on frequency Z=1.34,p=0.054; Bar graph inset for frequency, Student's t-test t(23)=0.47, p=0.64; n=11 cre−, n=14 cre+ from n=4 mice). (K-M) Example traces (K), cumulative probability distribution of amplitude (L) and frequency (M) of miniature EPSC recordings (+TTX) indicate normal synaptic function in the absence of NMIIA (K-S test on amplitude Z=1.28,p=0.08; Bar graph inset for frequency, Student's t-test t(30)=0.1, p=0.94; K-S test on frequency Z=0.60, p=0.87; Bar graph inset for frequency, Student's t-test t(30)=0.1, p=0.95; n=15 cre−, n=17 cre+ from n=5 mice). Error bars represent SEM.

Figure 6. Ablation of both NM IIB and NM IIA in developing CA1 neurons results in a phenotype similar to that of NMIIB null neurons

(A) Representative EPSCs from proximal and distal stimulation locations in Myh10^fl/fl; Myh9^fl/fl animals injected with Cre. Recordings from neighboring Cre negative and Cre positive neurons (Myh10/Myh9 double KO) show (B) reduced proximal (paired t(13)=2.49,p=0.026) and (C) normal distal EPSCs (paired t(13)=0.49, p=0.62) (n=14 pairs from n=6 mice). Dashed black line is identity line. Filled circle is the sample mean. (D-F) Example traces (D), cumulative probability distribution of amplitude (E) and frequency (F) of spontaneous EPSC recordings indicate reduced synaptic function in the absence of NMIIA and NMIIB (K-S test on amplitude Z=2.41, p<0.001; Bar graph inset for amplitude, Student's t-test t(20)=1.61, p=0.12; K-S test on frequency Z=3.23, p<0.001; Bar graph inset for frequency, Student's t-test t(20)=1.51, p=0.14; n=11 cre−, n=11 cre+ from n=5 mice). (G-I) Example traces (G), cumulative probability distribution of amplitude (H) and frequency (I) of miniature EPSC recordings (+TTX) indicate reduced quantal content in the absence of NMIIA and NMIIB(K-S test on amplitude Z=1.57, p=0.014; ; Bar graph inset for amplitude, Student's t-test t(24)=1.30, p=0.21, K-S test on frequency Z=0.78, p=0.58; Bar graph inset for frequency, Student's t-test t(24)=0.79, p=0.44; n=13 cre−, n=13 cre+ from n=7 mice). (J-L) Current injection spike frequency response curves indicate normal spiking responses in Myh10 ko neurons (J), slightly non-significant depolarized responses in Myh9 neurons (K) and altered slope of spiking responses in Myh10/Myh9 double ko neurons (L)(Myh10 RM ANOVA genotype F(1,37)=0,p=0.98, genotype*current F(6,222)=0.83,p=0.55, n=21 cre−, n=18 cre+ from n=7 mice; Myh9 RM ANOVA genotype F(1,25)=1.63,p=0.21, genotype*current F(6,150)=0.24,p=0.96 n=12 cre−, n=15 cre+ from n=4 mice, Myh10/Myh9 double ko RMANOVA genotype F(1,28)=0.24, p=0.63 genotype*current F(6,168)=3.67, p=0.002, n=15 cre−, n=15 cre+ from n=7 mice. Error bars represent SEM.