C57BI/6J mice were trained in the CFC paradigm. The fear acquisition group and respective control animals (fear acquisition, n = 18; CTRL-no shock, n = 8) were sacrificed 2 to 4 h after fear conditioning, in the window of fear memory consolidation. The fear memory group and respective controls (fear memory, n = 21; CTRL-no shock, n = 10) performed a fear memory retrieval test 24 h after fear training, to assess the retention of fear memory. Here, animals were sacrificed 2 to 4 h after fear memory retrieval, in the window of reconsolidation. Sacrifice of control and test animals was randomized to avoid bias in sample collection. Our goal was to study alterations in NT-3/TrkC and its intracellular signalling during the formation of a contextual fear memory. To that end, we first quantified the levels of TrkC expression and activation (as measured by phosphorylation) in the brain fear circuit, comprising the amygdala, PFC and hippocampus, of fear conditioned mice during the periods of fear memory consolidation (fear acquisition group) and reconsolidation (fear memory group).

Consolidation of a Fear Memory Correlates with Decreased TrkC Activation in the Amygdala and PFC

For the fear acquisition group (Fig. 1a), repeated measures two-way ANOVA performed on the freezing levels revealed a statistically significant effect of US presentations (US1-US5) (F (2.974, 71.38) = 8.36, p < 0.0001), treatment (CTRL-no shock vs CFC, F (1, 24) = 14.32, p = 0.0009) and US x treatment interaction (F (5, 120) = 8.591, p < 0.0001). Post hoc comparisons revealed that freezing levels increased with successive shock presentations (CFC: US1 vs US5, t (17) = 5.387, p = 0.0007) and that the percentage of time freezing in the fear acquisition animals was increased as compared to CTRL-no shock animals (US5: CTRL-no shock vs CFC, t (17.43) = 6.049, p < 0.0001), confirming that a fear response was successfully induced in the animals in the experimental condition.

Fig. 1

Expression and activation of TrkC in the fear circuit during contextual fear memory consolidation. (a) Fear acquisition mice (n = 18) underwent contextual fear conditioning, while CTRL-no shock mice (n = 8) did not receive any foot-shocks. The percentage of time spent freezing was assessed during the initial 2 min before US presentation (basal) and in the 15 s after each shock (US1-US5). (b, f, j) Representative western blot images of pTrkC and TrkC performed in (b) hippocampus, (f) amygdala and (j) prefrontal cortex total protein extracts from fear acquisition (n = 18) and CTRL-no shock (n = 8) mice sacrificed during fear consolidation. Quantification of (c, g, k) pTrkC/full-length TrkC ratio, (d, h, l) levels of phosphorylated TrkC and (e, i, m) total full-length TrkC levels. β-actin was used as a loading control. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. CTRL, control; pTrkC, phosphorylated TrkC; US, unconditioned stimulus

During fear memory consolidation, no significant differences were found in the expression and phosphorylation levels of TrkC between fear acquisition and CTRL-no shock mice in the hippocampus (pTrkC, t (22) = 0.6664, p = 0.5121; full-length TrkC, t (22) = 0.8627, p = 0.3976; pTrkC/full-length TrkC ratio, U = 59, p = 0.8362; Fig. 1b-e). However, in the amygdala (Fig. 1f-i) we observed a significant decrease in the relative phosphorylation of TrkC in fear acquisition animals when compared to CTRL-no shock (pTrkC/full-length TrkC ratio, t (23) = 2.583, p = 0.0166) and a trend for decrease in the total levels of phosphorylated TrkC (t (23) = 1.945, p = 0.0641), while the levels of full-length TrkC showed no alterations (U = 46, p = 0.3261). Likewise, in the PFC (Fig. 1j-m) contextual fear conditioning is associated with a decrease in the pTrkC/full-length TrkC ratio (t (7.513) = 2.531, p = 0.037), accompanied by a significant increase in the levels of full-length TrkC (t (23) = 2.184, p = 0.0394), but no differences in total phosphorylated TrkC (t (23) = 1.850, p = 0.0772).

Reconsolidation of a Fear Memory Correlates with Decreased TrkC Activation in the Hippocampus

In the fear memory group, during the fear acquisition phase (Fig. 2a), we observed a significant effect of US presentations (F (5, 145) = 8.919, p < 0.0001), treatment (F (1, 29) = 6.166, p = 0.0191) and US x treatment interaction (F (5, 145) = 7.810, p < 0.0001) in the percentage of time spent freezing. Again, successive US presentations increased the time spent freezing in the fear conditioned animals (US1 vs US5, t (145) = 8.684, p < 0.0001) and the percentage of time freezing in the fear memory animals was increased as compared to CTRL-no shock animals (US5: CTRL-no shock vs CFC, t (174) = 5.318, p < 0.0001). Twenty-four hours later, animals were placed back into the training chamber to assess fear memory retrieval. Here, conditioned mice displayed significantly higher freezing levels than CTRL-no shock mice (Fig. 2b, t (20.33) = 7.869, p < 0.0001), demonstrating proper retention and recall of fear memory.

Fig. 2

Expression and activation of TrkC in the fear circuit during contextual fear memory reconsolidation. (a) Fear memory mice (n = 21) underwent contextual fear conditioning, while CTRL-no shock mice (n = 10) did not receive any foot-shocks. The percentage of time spent freezing was assessed during the initial 2 min before US presentation (basal) and in the 15 s after each shock (US1-US5). (b) Percentage of time spent freezing was measured in conditioned and control mice during a 2-min fear retrieval session 24 h after CFC. (c, g, k) Representative western blot images of pTrkC and TrkC performed in (c) hippocampus, (g) amygdala and (k) prefrontal cortex total protein extracts from fear memory (n = 21) and CTRL-no shock (n = 10) mice sacrificed during fear reconsolidation. Quantification of (d, h, l) pTrkC/full-length TrkC ratio, (e, i, m) levels of phosphorylated TrkC and (f, j, n) total full-length TrkC levels. β-actin was used as a loading control. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. Panel c), non-contiguous lanes from the same membrane. CFC, contextual fear conditioning; CTRL, control; pTrkC, phosphorylated TrkC; US, unconditioned stimulus

During fear memory reconsolidation, we observed a decrease in relative TrkC phosphorylation levels in the hippocampus of conditioned animals (t (30) = 3.018, p = 0.0051) and a trend for decrease in total levels of phosphorylated TrkC (t (29) = 1.780, p = 0.0856), with no differences observed in the levels of full-length TrkC (t (28.74) = 1.319, p = 0.1974; Fig. 2c-f). In the amygdala (Fig. 2g-j), overall no differences were detected in the expression or activation levels of TrkC (pTrkC/full-length TrkC ratio, t (28.84) = 1.992, p = 0.0559; full-length TrkC, t (29) = 0.1586, p = 0.8751; pTrkC, U = 89, p = 0.5189). In the PFC (Fig. 2k-n), the total levels of phosphorylated TrkC were significantly decreased in conditioned animals (U = 33, p = 0.0042), with no differences found in the relative phosphorylation (U = 88, p = 0.7899) or in the levels of full-length TrkC (t (28) = 1.43, p = 0.1639).

Overall, our data points to a downregulation of TrkC signalling in key brain regions of the fear network during the formation of a fear memory.

Fear Memory-Related Decrease in TrkC Phosphorylation in the Brain Fear Circuit is not Associated with a Reduction in NT-3 Levels or Increase in Truncated TrkC Isoform

Next, we used ELISA to measure the levels of NT-3, which binds TrkC with high affinity, in lysates from brain regions highly implicated in the processing of fear and where the relative activation of TrkC was found to be altered at specific timepoints, i.e. amygdala during the fear consolidation phase (Fig. 1f, g) and hippocampus during fear reconsolidation (Fig. 2c, d). Statistical analysis did not reveal differences in the NT-3 levels between fear acquisition and CTRL-no shock animals in the amygdala (U = 60, p = 0.5308; Fig. 3a) or between fear memory and CTRL-no shock animals in the hippocampus (t (29) = 0.014, p = 0.9889; Fig. 3f). Although we did not observe differences in total NT-3 levels, both during consolidation and reconsolidation of fear, we cannot exclude changes in the activity-dependent secretion of NT-3 that are more likely to account for synaptic plasticity effects.

Fig. 3

Expression levels of different modulators of TrkC activation. (a, f) Quantification by ELISA of NT-3 levels (pg/mL) in protein extracts from (a) the amygdala of fear acquisition animals sacrificed during fear consolidation and (f) the hippocampus of fear memory animals sacrificed during fear reconsolidation, and respective controls. (b, c, g, h) Quantification of (b, g) full-length TrkC/truncated TrkC ratio and (c, h) truncated TrkC levels in total protein extracts from (b, c) the amygdala of fear acquisition animals sacrificed during fear consolidation and (g, h) the hippocampus of fear memory animals sacrificed during fear reconsolidation. Representative western blot images showing full-length TrkC and truncated TrkC under CTRL-no shock and fear acquisition/fear memory conditions are shown in Figs. 1f and 2c. (d, i) Representative images of PTP1B western blot performed in total protein extracts from (d) the amygdala of fear acquisition animals sacrificed during fear consolidation and (i) the hippocampus of fear memory animals sacrificed during fear reconsolidation. (e, j) Quantification of expression levels of PTP1B. β-actin was used as a loading control in western blots. CTRL, control; ELISA, enzyme-linked immunosorbent assay; NT-3, neurotrophin 3; WB, western blot

The truncated isoform of TrkC, which cannot be phosphorylated, can act as a dominant negative by sequestering NT-3 and preventing the activation of full-length TrkC receptors [44, 45]. To investigate the possibility that the overall decrease in TrkC phosphorylation, observed in the brain regions of the fear circuit of fear conditioned animals, is accompanied by an increase of the dominant negative truncated TrkC isoform, we measured its expression levels. During fear consolidation, no differences were detected between fear acquisition and CTRL-no shock animals in the ratio of full-length TrkC/truncated TrkC (U = 60, p = 0.8826) or in the expression levels of truncated TrkC (t (23) = 1.059, p = 0.3007) in the amygdala (Figs. 1f and 3b, c). Likewise, no differences were found between groups in the hippocampus (full-length TrkC/truncated TrkC, t (23) = 0.0354, p = 0.9721; truncated TrkC, t (23) = 0.2729, p = 0.7874; Fig. 1c; supplementary Fig. 1a, b) or in the PFC (full-length TrkC/truncated TrkC, U = 46, p = 0.1597; truncated TrkC, t (24) = 1.349, p = 0.19; Fig. 1j; supplementary Fig. 1c, d).

During fear reconsolidation, also no differences were observed in the total or relative expression levels of truncated TrkC between fear memory and CTRL-no shock animals in the hippocampus (full-length TrkC/truncated TrkC, t (30) = 0.5184, p = 0.6080; truncated TrkC, t (28.88) = 1.319, p = 0.1974; Figs. 2c and 3g, h), amygdala (full-length TrkC/truncated TrkC, t (29) = 0.6821, p = 0.5006; truncated TrkC, t (29) = 1.270, p = 0.2143; Fig. 2g; supplementary Fig. 1e, f) or PFC (full-length TrkC/truncated TrkC, U = 87, p = 0.7558; truncated TrkC, t (27.21) = 0.8677, p = 0.3932; Fig. 2k; supplementary Fig. 1 g, h).

Overall, there is no evidence for an association between the observed decrease in TrkC phosphorylation levels, during the consolidation and reconsolidation of a contextual fear memory, and alterations in the expression of TrkC truncated isoform.

No Alterations in the Expression Levels of the Trk-Targeting Phosphatase PTP1B

PTP1B is a phosphatase that targets, among others, Trk receptors, dephosphorylating them [46]. To investigate the possibility that the observed decrease in TrkC activation during the (re)consolidation of a contextual fear memory is associated with an increased expression of the TrkC-targeting PTP1B phosphatase, after fear acquisition or fear memory retrieval, we measured its expression levels, by western blot, in brain regions and timepoints where a decrease in TrkC activation was observed. We did not observe any differences between conditioned and CTRL-no shock animals in the levels of PTP1B in the amygdala during fear consolidation (U = 47, p = 0.4551; Fig. 3d, e) or in the hippocampus during fear reconsolidation (t (22) = 0.5175, p = 0.61; Fig. 3i, j). These results provide no evidence for an association between the observed decrease in TrkC phosphorylation levels, during the consolidation and reconsolidation of a contextual fear memory, and alterations in the expression of PTP1B.

Downregulation of Hippocampal NT-3/TrkC-ERK Pathway During Reconsolidation of Contextual Fear Memory

Given the results described above showing alterations in TrkC activation associated with the formation of a contextual fear memory, we aimed at investigating possible alterations in intracellular signalling pathways activated by the NT-3/TrkC system. To this end, we measured the expression and phosphorylation levels of the endpoint molecules Erk, Akt and PLC-γ in brain regions of the fear network during fear consolidation and fear reconsolidation timepoints.

During the consolidation phase, in the amygdala no differences were found between fear acquisition and CTRL-no shock animals in the expression and phosphorylation of Erk (Fig. 4a, d; pErk-1/Erk-1, t (23) = 0.4727, p = 0.6409; total Erk-1, U = 55, p = 0.4747; pERK-1, t (23) = 0.7439, p = 0.4645; pErk-2/Erk-2, U = 40, p = 0.1104; total Erk-2, t (23) = 0.4921, p = 0.6273; pErk-2, U = 60, p = 0.834), Akt (Fig. 4b, d; pAkt/Akt, t (23) = 0.1627, p = 0.8722; total Akt, t (22) = 0.2265, p = 0.8229; pAkt, U = 65, p = 0.8867), or in the expression of total PLC-γ (Fig. 4c, d; t (23) = 1.5, p = 0.1473).

Fig. 4

Expression levels of endpoint molecules of TrkC-recruited signalling pathways in the brain fear network during contextual fear consolidation. Representative western blot images of (a, e, i) pErk-1/2 and Erk-1/2, (b, f, j) pAkt and Akt, and (c, g, k) PLC-γ performed in the (a-c) amygdala, (e–g) prefrontal cortex and (i-k) hippocampus total protein extracts from fear acquisition (n = 18) and CTRL-no shock (n = 8) mice, sacrificed during fear consolidation. (d, h, l) Quantification of the ratio of phosphorylated/total protein levels, levels of phosphorylated proteins and total protein levels in (d) amygdala, (h) prefrontal cortex and (l) hippocampus. β-actin was used as a loading control. *p ≤ 0.05. CTRL, control

In the PFC, there were no differences between fear acquisition and CTRL-no shock animals in the expression and phosphorylation of Erk-1 (Fig. 4e, h; pErk-1/Erk-1, t (23) = 0.9776, p = 0.3385; total Erk-1, U = 57, p = 0.4285; pErk-1, t (24) = 0.1666, p = 0.8691). We did observe an increase in the pErk-2/Erk-2 ratio in fear acquisition animals as compared to CTRL-no shock (Fig. 4e, h; U = 24, p = 0.0131), though we did not detect differences between groups in the levels of total Erk-2 or phosphorylated Erk-2 (Fig. 4e, h; total Erk-2, t (24) = 0.0991, p = 0.9219; pErk-2, t (22) = 1.739, p = 0.096). No differences between groups were observed in the expression and phosphorylation of Akt (Fig. 4f, h; pAkt/Akt, t (21) = 0.2019, p = 0.842; total Akt, U = 41, p = 0.3411; pAkt, t (21) = 1.18, p = 0.2512) or in the expression of total PLC- γ (Fig. 4g, h; t (22.94) = 0.8122, p = 0.425).

In the hippocampus, we did not observe any differences between fear acquisition and CTRL-no shock experimental groups in the expression and phosphorylation of Erk-1/2 (Fig. 4i, l; pErk-1/Erk-1, U = 55, p = 0.367; total Erk-1, U = 55, p = 0.367; pErk-1, t (24) = 1.416, p = 0.1696; pErk-2/Erk-2, U = 42, p = 0.1021; total Erk-2, t (24) = 0.0333, p = 0.9737; pErk-2, U = 52, p = 0.2852), Akt (Fig. 4j, l; pAkt/Akt, t (22) = 0.1373, p = 0.8921; total Akt, t (23) = 0.1541, p = 0.8789; pAkt, t (22) = 0.9498, p = 0.3525) or in the expression of total PLC- γ (Fig. 4k, l; t (20.96) = 0.9924, p = 0.3323).

In sum, during the consolidation phase we observed an increase in the pErk2/Erk2 ratio in the PFC of fear acquisition animals as compared to CTRL-no shock animals, where TrkC changes were not detected, indicating that two pathways are most probably unrelated in these conditions.

During the reconsolidation window, in the amygdala we did not detect differences between fear memory animals and CTRL-no shock in any of the molecules studied. In detail, there were no differences between groups in the expression and phosphorylation levels of Erk-1/2 (Fig. 5a, d; pErk-1/Erk-1, t (24) = 0.5526, p = 0.5856; total Erk-1, t (24) = 0.984, p = 0.3349; pErk-1, t (24) = 1.649, p = 0.1121; pErk-2/Erk-2, t (24) = 0.2231, p = 0.8254; total Erk-2, t (24) = 0.1558, p = 0.8775; pErk-2, t (24) = 0.0318, p = 0.9749), Akt (Fig. 5b, d; pAkt/Akt, t (24) = 1.129, p = 0.27; total Akt, U = 77, p = 0.8971; pAkt, t (24) = 1.081, p = 0.2904) or in the expression of total PLC- γ (Fig. 5c, d; t (24) = 0.8548, p = 0.4011).

Fig. 5

Expression levels of endpoint molecules of TrkC-recruited signalling pathways in the brain fear network during contextual fear reconsolidation. Representative western blot images of (a, e, i) pErk-1/2 and Erk-1/2, (b, f, j) pAkt and Akt and (c, g, k) PLC-γ performed in the (a-c) amygdala, (e–g) prefrontal cortex and (i-k) hippocampus total protein extracts from fear memory (n = 16–21) and respective CTRL-no shock (n = 10) mice, sacrificed during fear reconsolidation. (d, h, l) Quantification of the ratio of phosphorylated/total protein levels, levels of phosphorylated proteins and total protein levels in the (d) amygdala, (h) prefrontal cortex and (l) hippocampus. β-actin was used as a loading control. *p ≤ 0.05, **p ≤ 0.01. CTRL, control; panel i), non-contiguous lanes from the same membrane

In the PFC, no differences were found in the expression and phosphorylation of Erk-1/2 between fear memory and CTRL-no shock groups (Fig. 5e, h; pErk-1/Erk-1, t (21) = 0.0792, p = 0.9376; total Erk-1, U = 48, p = 0.3686; pErk-1, t (21) = 1.145, p = 0.2652; pErk-2/Erk-2, t (21) = 0.1265, p = 0.9005; total Erk-2, t (19) = 0.816, p = 0.4246; pErk-2, t (19) = 1.002, p = 0.3292). We observed a decrease in total Akt expression in the fear memory group (Fig. 5f, h; total Akt, t (20) = 2.311, p = 0.0316), without differences in the phosphorylation of Akt (Fig. 5f, h; pAkt/Akt, t (19) = 1.085, p = 0.2916; pAkt, t (19) = 0.0506, p = 0.9601). We did not observe differences in the expression levels of total PLC- γ (Fig. 5g, h; t (19) = 0.273, p = 0.7878).

In the hippocampus, we observed a significant decrease in the expression levels of total Erk-1 (Fig. 5i, l; t (28) = 3.368, p = 0.0022) and in the levels of phosphorylated Erk-1 (t (28) = 3.545, p = 0.0014) in fear memory animals, although no differences were found in the pErk-1/Erk-1 ratio (t (28) = 0.1893, p = 0.8512). In addition, we observed a decrease in the pErk-2/Erk-2 ratio in fear memory animals (t (28) = 2.062, p = 0.0486), without differences observed in the expression levels of total Erk-2 (t (28) = 0.7121, p = 0.4823) or levels of phosphorylated Erk-2 (t (28) = 1.424, p = 0.1655) between the two conditions. Furthermore, no differences were found in the expression or phosphorylation levels of Akt (Fig. 5j, l; pAkt/Akt, t (29) = 1.569, p = 0.1274; total Akt, U = 102, p = 0.9173; pAkt, t (29) = 1.16, p = 0.2555) or in the expression levels of total PLC-γ (Fig. 5k, l; U = 85, p = 0.4164).

In sum, we observed an overall decrease in the activation of Erk-1/2 in the hippocampus during fear reconsolidation concurrent with a decrease in the activation of TrkC. Also, the observed reduction in the expression of Akt in the PFC should not be related to TrkC signalling as no changes in TrkC activation were detected in the PFC and this pathway is not expected to affect Akt expression levels.