Validation of AAV expression and confirmation of calcium dependent Ng-CaM binding.
Our injury, AAV administration, behavioral and molecular timeline was optimized to evaluate the effect of increased Ng on hippocampal post-synaptic molecular signaling and learning and memory behaviors. Sprague-Dawley rats received a moderate to severe Sham or CCI injury and 30 minutes later, control-AAV or Ng-AAV was administered over 20 minutes starting at 30 min post injury. 4 weeks later, ipsilateral hippocampal tissue was collected for synaptosomal isolation (Fig. 1a).
The AAV9 virus was generated from a plasmid with a CaMKII promoter and a T2A multicistronic vector to co-express Ng and GFP (Ng-AAV) (Fig. 1b). A control AAV9 virus was also generated with an identical promoter to only express GFP (Control-AAV). Four weeks after injury and injection, immunofluorescence shows expression of Ng and GFP in the ipsilateral hippocampus (Fig. 1c). Although the histological evaluation was not powered for a statistical assessment, it provided confirmation of injection coordinates and a preliminary readout of AAV expression in preparation for subsequent immunblot and Golgi histological experiments.
Pooled hippocampal lysate samples were incubated with CaM Sepharose beads with and without calcium to determine whether both exogenous (viral) and endogenous Ng binds to CaM in whole cell and synaptosomal lysates (Fig. 1d). Input samples show two distinct bands representing exogenous Ng at 17 kD and endogenous Ng at the predicted molecular weight of 15 kD. This separation in the exogenous form is due to the T2A linker peptide which upon cleavage, remains bound to the upstream protein.38 Our findings demonstrate that in the absence of calcium, both exogenous and endogenous Ng are bound to CaM, showing that the added T2A peptide does not interfere with key binding capabilities. Additionally, Ng is absent in the calcium enriched fraction when CaM is bound to CaMKII. This confirms proper protein interactions of exogenous Ng in our AAV treated animals in both Sham and CCI conditions.
Variable expression of Ng AAV and Control AAV after CCI in whole cell hippocampal lysates.
Immunoblot of whole cell lysates were conducted to quantify hippocampal protein expression of GFP and Ng 4 weeks after Sham or CCI injury and AAV injections (Fig. 2a). All Sham animals demonstrated viral expression for both Ng and Control AAV. Whole cell expression of GFP was not observed in 5/16 animals in the CCI-Control AAV group and 5/15 in the CCI-Ng AAV group. These animals had GFP optical densities of 0% ± 2% once background was removed and OD were normalized to Sham-Control AAV group, which defined the cut-off point. CCI-Ng AAV samples showing no GFP expression, also exhibited no expression of exogenous Ng. These animals are referred to as “non-expressors” and were separated from subsequent molecular statistical analysis (Figs. 2, 3 and 4).
One-way ANOVA shows a significant group effect in the whole protein expression of GFP (F3, 44 = 10.25, p < 0.0001, Fig. 2b). Tukey’s post-hoc analysis showed no significant differences between the two Sham AAV groups or between the two CCI AAV groups (p > 0.05). Compared to the Sham-Ng AAV and the Sham-Control AAV group, both the CCI-Control AAV and CCI-Ng AAV groups had significantly lower GFP expression (CCI-Con: p < 0.0001 and p < 0.0001, respectively; CCI-Ng: p = 0.0002 and p = 0.0001, respectively).
One-way ANOVA shows a significant group effect in the whole protein expression of total Ng, which encompasses both exogenous and endogenous forms (F3, 44 = 16.74, p < 0.0001, Fig. 2c). Tukey’s post-hoc analysis showed the Sham-Ng AAV group has significantly higher total Ng expression than Sham-Control AAV and CCI-Control AAV groups (p < 0.0001 and p < 0.0001, respectively). The CCI-Ng AAV also showed significantly higher total whole cell Ng expression compared to the Sham-Control AAV and Sham-CCI AAV groups (p < 0.0111 and p < 0.0147, respectively). There are no significant differences in total Ng expression between Sham-Ng AAV and CCI-Ng AAV (p > 0.05) and no significant differences between Sham-Control AAV and CCI-Control AAV groups (p > 0.05). One-way ANOVA showed no significant group effect in whole cell expression of endogenous Ng (F3, 44 = 2.571, p = 0.0662, Fig. 2d).
Ng AAV increases exogenous synaptic Ng in Sham and CCI.
Immunoblot of synaptosomes were conducted in animals identified as expressors to quantify hippocampal synaptic protein expression of total and endogenous Ng 4 weeks after Sham or CCI injury along with Ng or Control AAV injection (Fig. 3a). One-way ANOVA shows a significant group effect in the synaptic expression of total Ng (F3, 44 = 18.03, p < 0.0001, Fig. 3b). Tukey’s post-hoc analysis showed Sham-Ng AAV group had significantly higher total synaptic Ng expression than Sham-Control AAV and CCI-Control AAV and CCI-Ng AAV groups (p < 0.0001, p < 0.0001, and p = 0.0075, respectively). There were no significant differences in total Ng expression between the Sham-Control AAV group and the CCI-Control AAV or CCI-Ng AAV group (p > 0.05). The CCI-Ng AAV group has significantly higher total Ng expression than the CCI-Control AAV group (p = 0.0314).
One-way ANOVA shows a significant group effect in the synaptic expression of endogenous Ng (F3, 44 = 3.322, p = 0.0280, Fig. 3c). There were no significant differences between expression of Sham-Ng AAV to all three groups (p > 0.05). There were no significant differences between Sham-Control AAV and CCI-Control AAV groups. There was a significant decrease in endogenous Ng expression in the CCI-Ng AAV group compared to the Sham-Control AAV (p = 0.0222). There were no differences in endogenous Ng expression between CCI-Control and Ng AAV groups (p > 0.05).
Increasing Ng by AAV does not change P-Ng/Ng ratio in Sham or CCI injured animals.
Immunoblots of synaptosomes were conducted to quantify hippocampal synaptic protein expression of total and endogenous P-Ng and P-Ng/Ng ratio 4 weeks after Sham or CCI injury along with Ng or Control AAV injection (Fig. 3a). One-way ANOVA shows a significant group effect in the expression of total P-Ng (F3, 44 = 14.86, p < 0.0001, Fig. 3d). Tukey’s post-hoc analysis showed Sham-Ng AAV group had significantly higher total P-Ng expression than Sham-Control AAV and CCI-Control AAV groups (p < 0.0001, p < 0.0001, respectively). The CCI-Ng AAV group had significantly higher total P-Ng expression than the Sham and CCI-Control AAV groups (p = 0.0447 and p = 0.0159, respectively). There were no significant differences in total P-Ng expression between the Sham-Ng AAV group and the CCI-Ng AAV group, although there was a trend (p = 0.0701). There were no significant differences in total P-Ng expression between the Sham-Control AAV group and the CCI-Control AAV group (p > 0.05). One-way ANOVA showed no significant group effect in synaptic expression of endogenous P-Ng (F3, 44 = 1.835, p = 0.1548, Fig. 3e). One-way ANOVA showed no significant group effect in synaptic expression of total P-Ng/Ng ratio (F3, 44 = 1.211, p = 0.3170, Fig. 3F) and no significant group effect in synaptic expression of endogenous P-Ng/Ng ratio (F3, 44 = 0.5019, p = 0.6829, Fig. 3g).
Increasing synaptic Ng reduces αCaMKII expression without altering P-CaMKII/αCaMKII ratio.
Immunoblot of synaptosomes were conducted to quantify hippocampal synaptic protein expression of P-CaMKII (Thr286), αCaMKII and P-CaMKII/αCaMKII ratio 4 weeks after Sham or CCI injury along with Ng or Control AAV injection (Fig. 4a). One-way ANOVA showed no significant group effect in synaptic expression of P-CaMKII (F3, 44 = 1.110, p = 0.3554, Fig. 4b). One-way ANOVA showed a significant group effect in the expression of α-CaMKII (F3, 44 = 3.148, p < 0.0343, Fig. 4c). Tukey’s post-hoc analysis showed Sham-Ng AAV group had significantly lower α-CaMKII expression than the Sham-Control AAV group (p < 0.0404). All other comparisons were not significant. One-way ANOVA showed no significant group effect in synaptic expression of P-CaMKII/αCaMKII ratio (F3, 44 = 0.6633, p = 0.5790, Fig. 4d).
Ng-AAV increases mushroom spine density in Shams and increases filopodia density in CCI.
Golgi-stained neurons were reconstructed into 3-dimensional computer models using Neurolucida 360 software to determine the effect of Ng AAV and CCI injury on apical dendritic spine density and morphology in the CA1 of the hippocampus 4 weeks-post injury and AAV administration (Fig. 5a-l). One-way ANOVA showed no significant group effect in total spine density (F3, 217 = 1.349, p = 0.2596, Fig. 5m). One-way ANOVA showed no significant group effect in thin and stubby spine density (F3, 217 = 1.234, p = 0.2984, Fig. 5n, F3, 217 = 0.4828, p = 0.6946, Fig. 5o, respectively).
One-way ANOVA showed a significant main group effect for mushroom spine density (F3, 217 = 3.639, p = 0.0136, Fig. 5p). Tukey’s post-hoc test showed the Sham-Ng AAV group had a significantly higher mushroom spine density than the CCI-Control AAV and CCI-Ng AAV group (p = 0.0289 and p = 0.0293). There was also a trending increase in the Sham-Ng AAV group compared to Sham-Control AAV group (p = 0.0818). There were no significant differences between the Sham-Control AAV group and either CCI group and no significant differences between CCI-Control and CCI-Ng AAV groups.
One-way ANOVA showed a significant main group effect for filopodia spine density (F3, 217 = 2.2954, p = 0.0335, Fig. 5q). Tukey’s post-hoc test showed the CCI-Ng AAV had a significantly higher filopodia spine density than the Sham-Ng AAV group (p = 0.0194). No other statistical comparisons were significant.
Increasing Ng expression by AAV does not change motor behaviors after CCI.
Beam-balance and beam-walking tasks assessed if Ng or Control AAV injection alters sensorimotor and vestibulomotor function after Sham or CCI injury. These supplement the classical measures of injury severity, such as the righting reflex, given the length of time animals remain under isoflurane anesthesia after Sham or CCI injury and confirm that that injury severity was consistent between the CCI-injured groups. On the beam-balance task, repeated-measures ANOVA showed a main group effect for latencies over 5 days (F5, 95 = 11.429, p < 0.001, Fig. 6a). Tukey’s post-hoc test shows significantly decreased latency on the balance beam in CCI-Control and CCI-Ng AAV groups compared to the Sham-Control and Sham-Ng AAV groups (CCI-Con: p = 0.035 and p = 0.028, respectively; CCI-Ng: p < 0.001 and p < 0.001 respectively).There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
Similarly, on the beam-walking task, repeated-measures ANOVA showed a main group effect for latencies over 5 days (F5, 95 = 62.747, p < 0.001, Fig. 6b). Tukey’s post-hoc test shows significantly increased latency to traverse the beam in CCI-Control and CCI-Ng AAV groups compared to the Sham-Control and Sham-Ng AAV groups (CCI-Con: p < 0.001 and p < 0.001, respectively; CCI-Ng: p < 0.001 and p < 0.001, respectively). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
Increasing hippocampal Ng expression by AAV does not change learning and memory behaviors after CCI.
Performance in the MWM task for spatial learning and memory was evaluated days 21 to 27 after Sham or CCI injury along with Ng or Control AAV injection. Submerged latency measured the length of time an animal takes to reach a hidden platform. Repeated-measures ANOVA showed a significant group main effect (F5, 95 = 20.935, p < 0.001, Fig. 7a). Tukey’s post-hoc test shows significantly increased latency to reach submerged platform in CCI-Control and CCI-Ng AAV groups compared to the Sham-Control and Sham-Ng AAV groups (CCI-Con: p < 0.001 and p < 0.001, respectively; CCI-Ng: p < 0.001 and p < 0.001, respectively). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
The hidden platform was removed for the probe trial on day 26 post-injury and AAV injection. One-way ANOVA showed a significant main effect for the measure of percent time (normalized to total 60 second trial) spent in the target quadrant (F3, 95 = 1.881, p = 0.0081, Fig. 7b). Tukey’s post-hoc test showed the CCI-Ng AAV group spent significantly less % time in the target quadrant compared to the Sham-Control AAV group (P = 0.0340) indicating an impairment in spatial memory retention. There was no significant difference in % time between the CCI-Control AAV group and the Sham-Control AAV group, although there was a modest trend (p = 0.0915). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
An additional method to measure probe trial performance is to measure the latency an animal takes to first reach the designated “platform zone.” One-way ANOVA showed a significant main effect (F3, 95 = 6.772, p = 0.0003, Fig. 7c). Tukey’s post-hoc test showed the CCI-Control and CCI-Ng AAV groups had a significantly longer latency to reach the platform zone compared to the Sham-Control AAV group (p = 0.0007 and p = 0.0171, respectively). The CCI-Control group was also significantly higher than the Sham-Ng AAV group (p = 0.0151). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
As shown in Fig. 2, approximately 40% of CCI animals did not express the AAV and were classified as non-expressors. Therefore, MWM performance was analyzed in expressors to determine if there is an association between behavior and hippocampal synaptic total Ng expression. The results of this sub-set were similar to the results of the entire group analysis.
For the submerged trial, repeated-measures ANOVA showed a significant group main effect (F3, 44 = 20.935, p < 0.001, Fig. 7d). Tukey’s post-hoc test shows significantly increased latency to reach submerged platform in CCI-Control and CCI-Ng AAV groups compared to the Sham-Control and Sham-Ng AAV groups (p < 0.001, p < 0.001, respectively). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.
In the probe trial, one-way ANOVA of the percent time (normalized to total 60 second trial) spent in the target quadrant (Fig. 7e) showed no significant main effect. For the latency to first reach the designated “platform zone,” one-way ANOVA showed a significant main effect (F3, 44 = 4.212, p = 0.0105, Fig. 7f). Tukey’s post-hoc test showed the CCI-Ng AAV group had a significantly longer latency to reach the platform zone compared to the Sham-Control AAV group (p = 0.0112). There was a trending difference between the CCI-Control AAV group and the Sham-Control AAV group (p = 0.0524). There were no significant differences between Sham-Control and Sham-Ng AAV groups and no significant differences between CCI-Control and CCI-Ng AAV groups.