**2. Evidence that TrkB-Shc alternative transcripts are selectively increased in the hippocampus during severe, late stage AD**

In Wong et al. (2012) [38], we measured changes in TrkB alternate transcript levels in control and AD postmortem human brain tissue derived from the hippocampus, temporal cortex, oc‐ cipital cortex, and cerebellum (Braak stages V and VI) [38]. By quantitative real-time PCR, us‐ ing primers specific for each TrkB alternative transcript, we found significant increases in TrkB-Shc mRNA expression in the hippocampus but not in any other brain region (Figure 3).

Considering that brain homogenates contain a mixed population of cells, we determined whether the changes found in TrkB transcript expression using the hippocampal tissue ho‐ mogenates also occur in neurons exposed to an amyloidogenic environment. Here, changes in TrkB transcripts were assessed by incubating differentiated SHSY5Y cells (a human neu‐ roblastoma cell-line which express TrkB) with different species of amyloid beta 1-42 (Aβ42) peptides at various stages of aggregation. Oligomers and fibrils were prepared as described in Ryan et al. [39] and characterized by western blotting and atomic force microscopy imag‐ ing [38]. A significant increase in TrkB-Shc mRNA levels was found when cells were incu‐ bated with preparations of Aβ<sup>42</sup> containing fibrils compared to controls (Figure 4). The small magnitude of change was expected as the Aβ<sup>42</sup> fibril preparation contained mixed Aβ<sup>42</sup> spe‐ cies and the absolute amount of fibrils would be low (fibrils were absent in the monomer and oligomer Aβ<sup>42</sup> preparations). Further, in comparison to the Aβ42 monomer and oligomer preparations, the Aβ42 fibril preparations would be most representative of all Aβ42 species present in the AD hippocampus as this preparation comprises a mix of all three species [38]. These results were consistent with findings of increased TrkB-Shc mRNA levels in the AD hippocampus (Figures 3 and 4).

**Figure 2. TrkB receptor dimer combinations.** All dimer combinations of TrkB receptors can bind to BDNF. However, only a homodimer of TrkB-TK+ can initiate second messenger signaling. This figure utilizes modified ProteinLounge

**2. Evidence that TrkB-Shc alternative transcripts are selectively increased**

In Wong et al. (2012) [38], we measured changes in TrkB alternate transcript levels in control and AD postmortem human brain tissue derived from the hippocampus, temporal cortex, oc‐ cipital cortex, and cerebellum (Braak stages V and VI) [38]. By quantitative real-time PCR, us‐ ing primers specific for each TrkB alternative transcript, we found significant increases in TrkB-Shc mRNA expression in the hippocampus but not in any other brain region (Figure 3). Considering that brain homogenates contain a mixed population of cells, we determined whether the changes found in TrkB transcript expression using the hippocampal tissue ho‐ mogenates also occur in neurons exposed to an amyloidogenic environment. Here, changes in TrkB transcripts were assessed by incubating differentiated SHSY5Y cells (a human neu‐ roblastoma cell-line which express TrkB) with different species of amyloid beta 1-42 (Aβ42) peptides at various stages of aggregation. Oligomers and fibrils were prepared as described in Ryan et al. [39] and characterized by western blotting and atomic force microscopy imag‐ ing [38]. A significant increase in TrkB-Shc mRNA levels was found when cells were incu‐ bated with preparations of Aβ<sup>42</sup> containing fibrils compared to controls (Figure 4). The small magnitude of change was expected as the Aβ<sup>42</sup> fibril preparation contained mixed Aβ<sup>42</sup> spe‐

graphics created using the Pathway Builder Tool (www.proteinlounge.com).

184 Trends in Cell Signaling Pathways in Neuronal Fate Decision

**in the hippocampus during severe, late stage AD**

**Figure 3. TrkB alternative transcript expression in various control and AD brain regions.** Expression of TrkB alter‐ native transcripts in the (A) hippocampus, (B) temporal cortex, (C) occipital cortex, and (D) cerebellum of control (CON) (white) and Alzheimer's disease (AD) (black) postmortem human brain tissue were measured by qPCR. \*\* P=0.004; # P=0.07. Figure is from Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expres‐ sion with implications for Alzheimer's disease. Neuroscience 2012;210 363-374.

In regards to the brain regions examined, AD brain pathology from most severe to least af‐ fected is: hippocampus>temporal cortex>occipital cortex>cerebellum. The selective increase in TrkB-Shc transcripts observed only in the hippocampus suggested that the elevated TrkB-Shc transcript levels were occurring in brain regions that are most severely affected in the diseased state and that the observed increase may likely be influenced by the neuronal cell population present. This was supported by our *in vitro* AD cell culture model showing that TrkB-Shc mRNA levels in the differentiated SHSY5Y neuronal cells can be increased by ex‐ posure to preparations of Aβ42 containing fibrils. Aβ42 fibril species are increased in the ad‐ vanced stages of AD [40]. While it is widely accepted that soluble oligomers are the more neurotoxic of the Aβ42 species, evidence also suggest that the neurotoxicity of Aβ42 requires its aggregation in the fibrillar form, particularly in the form of protofibrils [41, 42]. In our Aβ42 preparations, we detected various sizes of fibrils in the Aβ<sup>42</sup> fibril preparations, includ‐

duced but little change was observed for phosphorylated TrkB:total TrkB-TK+ and phos‐

Neurotrophin Signaling and Alzheimer's Disease Neurodegeneration − Focus on BDNF/TrkB Signaling

http://dx.doi.org/10.5772/53364

187

**Figure 5.** Effect of TrkB-Shc overexpression on BDNF-stimulated TrkB-TK+ second messenger signaling in differentiat‐ ed SHSY5Y neuronal cells. Differentiated SHSY5Y cells were transfected with either empty vector (EV; blue) or myctagged TrkB-Shc (pink) for 24 h and treated with increasing concentrations of BDNF for 15 min and harvested. Proteins were separated by Western blotting and immunoprobed. (A) Representative Western blot images. (B) Bands were quantitated by densitometry and presented as protein expression ratios of phosphorylated protein to non-phos‐ phorylated protein (pTrkB-TK+:TrkB-TK+; pAKT:AKT; pERK1:ERK1; pERK2:ERK2) -/+ SEM. \* P=0.02 for pTrkB:TrkB-TK+ at 25 ng, P=0.03 for pERK1:ERK1 at 5 ng BDNF and P=0.049 for pERK1:ERK1 at 15 ng BDNF. Figure is from Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheim‐

In a parallel study [43], we determined how TrkB-Shc exerts its dominant negative effect on BDNF-stimulated TrkB-TK+ signaling. Using a non-neuronal cell-line, Chinese Hamster Ovary K1 (CHOK1) cells (which do not express endogenous TrkB receptors), we transiently overexpressed TrkB-Shc protein and examined its effect on TrkB-TK+ protein stability. We used cycloheximide to block protein synthesis as this would prevent newly synthesized pro‐ tein from replacing protein that has been degraded. From this, we found that TrkB-Shc pro‐ tein levels were rapidly decreased when cells were exposed to exogenous BDNF. Moreover, in co-expression experiments where TrkB-Shc and TrkB-TK+ were co-expressed, cyclohexi‐ mide treatment revealed increased protein degradation of phosphorylated TrkB-TK+ pro‐ tein, a process that is accelerated by BDNF exposure (Figure 6A and B) [43]. Interestingly, while the reduction of phosphorylated TrkB-TK+ protein was more pronounced in the pres‐ ence of TrkB-Shc following BDNF exposure, the stability of TrkB-Shc protein itself was in‐

Our recent findings have important implications in regards to the role of TrkB-Shc and its impact on BDNF-mediated TrkB-TK+ signaling in neurons in AD. MEK signaling through

phorylated AKT:total AKT ratios (Figure 5).

er's disease. Neuroscience 2012;210 363-374.

creased (Figure 6C).

**4. Discussion**

**Figure 4.** Effect of different structural forms of Aβ<sup>42</sup> on TrkB alternative transcript expression in differentiated SHSY5Y neuronal cells. SHSY5Y cells were differentiated for 9 d and incubated in the absence (white bars) or presence of 1 μM Aβ42 monomers (AβM) (gray bars), oligomers (AβO) (black bars), and fibrils (AβF) (hatched bars) for 6 h and harvested. Expression of TrkB alternative transcripts were then measured by qPCR. Data are expressed as mean + SEM relative to the control (white bars) condition that was set to 1. \* P=0.03. Figure is from Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheimer's disease. Neuroscience 2012;210 363-374.

ing protofibrils which can be ~100 nm in length. If we consider that Aβ<sup>42</sup> fibrils are predomi‐ nant in the advanced stages of AD [40], that the hippocampus is the most affected brain region, and that the CA1 subregion (the hippocampal region assessed) is the most severely impacted subregion of the hippocampus in AD, altogether, our findings suggest that the se‐ lective increase in the TrkB-Shc alternative splice transcript in the AD hippocampus may be specific to severe, late stage pathology.

### **3. TrkB-Shc inhibits TrkB-TK+ function and BDNF/TrkB-TK+ signaling**

The elevated levels of TrkB-Shc in AD suggested that it may have a functional impact on cellular TrkB-TK+ signaling *in vivo*. Previous studies have demonstrated that TrkB-Shc can function as a dominant negative receptor by inhibiting TrkB-TK+ phosphorylation [28]. In Wong et al. (2012) [38], we confirmed this finding and extended it by showing that TrkB-Shc can decrease downstream second messenger signaling linked to TrkB-TK+, supporting a dominant negative function of TrkB-Shc on BDNF/TrkB-TK+ signaling. In particular, we found that when TrkB-Shc was overexpressed in SHSY5Y cells (which express endogenous TrkB-TK+), BDNF/TrkB-TK+ stimulated MEK pathway signaling was selectively attenuated. The ratio of phosphorylated ERK1/2:total ERK1/2 (a measure of ERK1/2 activity) was re‐ duced but little change was observed for phosphorylated TrkB:total TrkB-TK+ and phos‐ phorylated AKT:total AKT ratios (Figure 5).

**Figure 5.** Effect of TrkB-Shc overexpression on BDNF-stimulated TrkB-TK+ second messenger signaling in differentiat‐ ed SHSY5Y neuronal cells. Differentiated SHSY5Y cells were transfected with either empty vector (EV; blue) or myctagged TrkB-Shc (pink) for 24 h and treated with increasing concentrations of BDNF for 15 min and harvested. Proteins were separated by Western blotting and immunoprobed. (A) Representative Western blot images. (B) Bands were quantitated by densitometry and presented as protein expression ratios of phosphorylated protein to non-phos‐ phorylated protein (pTrkB-TK+:TrkB-TK+; pAKT:AKT; pERK1:ERK1; pERK2:ERK2) -/+ SEM. \* P=0.02 for pTrkB:TrkB-TK+ at 25 ng, P=0.03 for pERK1:ERK1 at 5 ng BDNF and P=0.049 for pERK1:ERK1 at 15 ng BDNF. Figure is from Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheim‐ er's disease. Neuroscience 2012;210 363-374.

In a parallel study [43], we determined how TrkB-Shc exerts its dominant negative effect on BDNF-stimulated TrkB-TK+ signaling. Using a non-neuronal cell-line, Chinese Hamster Ovary K1 (CHOK1) cells (which do not express endogenous TrkB receptors), we transiently overexpressed TrkB-Shc protein and examined its effect on TrkB-TK+ protein stability. We used cycloheximide to block protein synthesis as this would prevent newly synthesized pro‐ tein from replacing protein that has been degraded. From this, we found that TrkB-Shc pro‐ tein levels were rapidly decreased when cells were exposed to exogenous BDNF. Moreover, in co-expression experiments where TrkB-Shc and TrkB-TK+ were co-expressed, cyclohexi‐ mide treatment revealed increased protein degradation of phosphorylated TrkB-TK+ pro‐ tein, a process that is accelerated by BDNF exposure (Figure 6A and B) [43]. Interestingly, while the reduction of phosphorylated TrkB-TK+ protein was more pronounced in the pres‐ ence of TrkB-Shc following BDNF exposure, the stability of TrkB-Shc protein itself was in‐ creased (Figure 6C).

#### **4. Discussion**

ing protofibrils which can be ~100 nm in length. If we consider that Aβ<sup>42</sup> fibrils are predomi‐ nant in the advanced stages of AD [40], that the hippocampus is the most affected brain region, and that the CA1 subregion (the hippocampal region assessed) is the most severely impacted subregion of the hippocampus in AD, altogether, our findings suggest that the se‐ lective increase in the TrkB-Shc alternative splice transcript in the AD hippocampus may be

**Figure 4.** Effect of different structural forms of Aβ<sup>42</sup> on TrkB alternative transcript expression in differentiated SHSY5Y neuronal cells. SHSY5Y cells were differentiated for 9 d and incubated in the absence (white bars) or presence of 1 μM Aβ42 monomers (AβM) (gray bars), oligomers (AβO) (black bars), and fibrils (AβF) (hatched bars) for 6 h and harvested. Expression of TrkB alternative transcripts were then measured by qPCR. Data are expressed as mean + SEM relative to the control (white bars) condition that was set to 1. \* P=0.03. Figure is from Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheimer's disease. Neuroscience

**3. TrkB-Shc inhibits TrkB-TK+ function and BDNF/TrkB-TK+ signaling**

The elevated levels of TrkB-Shc in AD suggested that it may have a functional impact on cellular TrkB-TK+ signaling *in vivo*. Previous studies have demonstrated that TrkB-Shc can function as a dominant negative receptor by inhibiting TrkB-TK+ phosphorylation [28]. In Wong et al. (2012) [38], we confirmed this finding and extended it by showing that TrkB-Shc can decrease downstream second messenger signaling linked to TrkB-TK+, supporting a dominant negative function of TrkB-Shc on BDNF/TrkB-TK+ signaling. In particular, we found that when TrkB-Shc was overexpressed in SHSY5Y cells (which express endogenous TrkB-TK+), BDNF/TrkB-TK+ stimulated MEK pathway signaling was selectively attenuated. The ratio of phosphorylated ERK1/2:total ERK1/2 (a measure of ERK1/2 activity) was re‐

specific to severe, late stage pathology.

186 Trends in Cell Signaling Pathways in Neuronal Fate Decision

2012;210 363-374.

Our recent findings have important implications in regards to the role of TrkB-Shc and its impact on BDNF-mediated TrkB-TK+ signaling in neurons in AD. MEK signaling through

**Figure 6. Effect of TrkB-Shc on the stability of phosphorylated TrkB-TK+.** CHOK1 cells co-transfected with either empty vector + TrkB-TK+ or TrkB-TK+ + TrkB-Shc for 24 h were treated with 15 ng BDNF for 15 min and then incubated with cycloheximide for 3 h before harvest. Proteins were separated by Western blotting and immunoprobed. (A and C) Representative Western blot images. (A) and (C) are from the same blot and have the same exposure. (B) Bands in (A) were quantitated by densitometry and presented as protein expression normalized to β-actin + SEM. \*\*P = 0.007. Figure is from Wong, J and Garner, B. Biochemical and Biophysical Research Communications 2012; 420 331–335.

**Figure 7. Summary diagram.** Upper panel: In the normal, non-diseased state, binding of BDNF to TrkB leads to autophosphorylation of TrkB-TK+ homodimers (but not other TrkB dimer combinations). This leads to activation of down‐ stream second messenger signaling pathways including PLCγ (phospholipase C-gamma), MEK (mitogen-activated protein kinase kinase), and PI3K (phosphatidyl inositol 3 kinase), which are critical for neuronal viability and function. TrkB-TK+ homodimers can also be auto-phosphorylated in the absence of BDNF, although activation of downstream signaling pathways are less intense. BDNF-stimulation of TrkB receptors also leads to their degradation. The arrow thickness indicates the magnitude of effect. The magnitude of decrease in protein levels are greatest for the following receptor combinations when stimulated with exogenous BDNF: TrkB-TK+/TrkB-Shc>TrkB-TK+/TrkB-TK+>TrkB-Shc/ TrkB-Shc. Lower panel: In AD, there is Aβ plaque accumulation. Neurons are exposed to mixed species of Aβ, including Aβ<sup>42</sup> fibrils and oligomers at various stages of aggregation. BDNF protein expression has been reported to be reduced in AD (indicated by red arrow) and TrkB-Shc levels have been reported by us to be increased in AD (hippocampus, CA1) (indicated by green arrow) and in response to Aβ42 fibril exposure. The increase in TrkB-Shc is predicted to in‐ crease heterodimer combinations of TrkB-TK+/TrkB-Shc and homodimer combinations of TrkB-Shc, which would lead to an overall reduction in downstream signaling. Moreover, the increase in TrkB-Shc dimer combinations also increas‐ es TrkB receptor degradation. Thus, the combination of reduced BDNF expression and increased TrkB-Shc expression in the AD hippocampus would likely result in an overall decrease in BDNF/TrkB-TK+ signaling. Figure utilizes modified ProteinLounge graphics created using the Pathway Builder Tool (www.proteinlounge.com). Insert panel: Atomic force microscopy image of the Aβ42 fibril preparation used in Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheimer's disease. Neuroscience 2012;210 363-374. This

Neurotrophin Signaling and Alzheimer's Disease Neurodegeneration − Focus on BDNF/TrkB Signaling

http://dx.doi.org/10.5772/53364

189

contains mixed species of Aβ42 monomers, oligomers, and fibrils of various sizes. Scale represents 1 μm.

ERK1/2 phosphorylation is increased in vulnerable neurons in AD and is implicated in the abnormal phosphorylation of tau and neurofilament proteins [44]. Our finding of a selective attenuation in MEK signaling activity in neurons with elevated levels of cellular TrkB-Shc implicates that elevated levels of this neuron-specific truncated TrkB receptor in AD may oc‐ cur as a response to the disease. However, our finding that cells may increase TrkB-Shc pro‐ tein levels in response to BDNF stimulation to regulate TrkB-TK+ activity by increasing degradation of activated receptor complexes also has ramifications for BDNF/TrkB-TK+ sig‐ naling in AD. In the non-diseased or control state, this process is akin to feedback regulatory loops observed in metabolic pathways (Figure 7). While BDNF/TrkB-TK+ signaling is impor‐ tant in multiple aspects related to neuronal viability and differentiation, overactivation or inappropriate temporal and spatial activation of BDNF/TrkB-TK+ signaling during brain de‐ velopment or "leakage" of BDNF to adjacent neurons or brain regions can negatively impact brain function. However, in AD, elevated levels of TrkB-Shc in association with reduced BDNF protein levels (which is well documented) and no change in TrkB-TK+ expression may also result in an overall increase in the degradation of phosphorylated TrkB-TK+ recep‐ tors, and thus, reduce overall BDNF/TrkB-TK+ activity in neurons in AD.

**Figure 7. Summary diagram.** Upper panel: In the normal, non-diseased state, binding of BDNF to TrkB leads to autophosphorylation of TrkB-TK+ homodimers (but not other TrkB dimer combinations). This leads to activation of down‐ stream second messenger signaling pathways including PLCγ (phospholipase C-gamma), MEK (mitogen-activated protein kinase kinase), and PI3K (phosphatidyl inositol 3 kinase), which are critical for neuronal viability and function. TrkB-TK+ homodimers can also be auto-phosphorylated in the absence of BDNF, although activation of downstream signaling pathways are less intense. BDNF-stimulation of TrkB receptors also leads to their degradation. The arrow thickness indicates the magnitude of effect. The magnitude of decrease in protein levels are greatest for the following receptor combinations when stimulated with exogenous BDNF: TrkB-TK+/TrkB-Shc>TrkB-TK+/TrkB-TK+>TrkB-Shc/ TrkB-Shc. Lower panel: In AD, there is Aβ plaque accumulation. Neurons are exposed to mixed species of Aβ, including Aβ<sup>42</sup> fibrils and oligomers at various stages of aggregation. BDNF protein expression has been reported to be reduced in AD (indicated by red arrow) and TrkB-Shc levels have been reported by us to be increased in AD (hippocampus, CA1) (indicated by green arrow) and in response to Aβ42 fibril exposure. The increase in TrkB-Shc is predicted to in‐ crease heterodimer combinations of TrkB-TK+/TrkB-Shc and homodimer combinations of TrkB-Shc, which would lead to an overall reduction in downstream signaling. Moreover, the increase in TrkB-Shc dimer combinations also increas‐ es TrkB receptor degradation. Thus, the combination of reduced BDNF expression and increased TrkB-Shc expression in the AD hippocampus would likely result in an overall decrease in BDNF/TrkB-TK+ signaling. Figure utilizes modified ProteinLounge graphics created using the Pathway Builder Tool (www.proteinlounge.com). Insert panel: Atomic force microscopy image of the Aβ42 fibril preparation used in Wong J, et al. Amyloid beta selectively modulates neuronal TrkB alternative transcript expression with implications for Alzheimer's disease. Neuroscience 2012;210 363-374. This contains mixed species of Aβ42 monomers, oligomers, and fibrils of various sizes. Scale represents 1 μm.

ERK1/2 phosphorylation is increased in vulnerable neurons in AD and is implicated in the abnormal phosphorylation of tau and neurofilament proteins [44]. Our finding of a selective attenuation in MEK signaling activity in neurons with elevated levels of cellular TrkB-Shc implicates that elevated levels of this neuron-specific truncated TrkB receptor in AD may oc‐ cur as a response to the disease. However, our finding that cells may increase TrkB-Shc pro‐ tein levels in response to BDNF stimulation to regulate TrkB-TK+ activity by increasing degradation of activated receptor complexes also has ramifications for BDNF/TrkB-TK+ sig‐ naling in AD. In the non-diseased or control state, this process is akin to feedback regulatory loops observed in metabolic pathways (Figure 7). While BDNF/TrkB-TK+ signaling is impor‐ tant in multiple aspects related to neuronal viability and differentiation, overactivation or inappropriate temporal and spatial activation of BDNF/TrkB-TK+ signaling during brain de‐ velopment or "leakage" of BDNF to adjacent neurons or brain regions can negatively impact brain function. However, in AD, elevated levels of TrkB-Shc in association with reduced BDNF protein levels (which is well documented) and no change in TrkB-TK+ expression may also result in an overall increase in the degradation of phosphorylated TrkB-TK+ recep‐

**Figure 6. Effect of TrkB-Shc on the stability of phosphorylated TrkB-TK+.** CHOK1 cells co-transfected with either empty vector + TrkB-TK+ or TrkB-TK+ + TrkB-Shc for 24 h were treated with 15 ng BDNF for 15 min and then incubated with cycloheximide for 3 h before harvest. Proteins were separated by Western blotting and immunoprobed. (A and C) Representative Western blot images. (A) and (C) are from the same blot and have the same exposure. (B) Bands in (A) were quantitated by densitometry and presented as protein expression normalized to β-actin + SEM. \*\*P = 0.007. Figure is from Wong, J and Garner, B. Biochemical and Biophysical Research Communications 2012; 420 331–335.

188 Trends in Cell Signaling Pathways in Neuronal Fate Decision

tors, and thus, reduce overall BDNF/TrkB-TK+ activity in neurons in AD.
