Background and Significance

Due to the dramatic rise in life expectancy during the 20^th century, Alzheimer's disease (AD) has emerged as the most prevalent form of dementia in humans.

Apolipoprotein E and Alzheimer’s disease:

From a molecular perspective, AD is more than one disorder, and several genetic associations of AD have been identified (52,69,75). Among these genetic associations, apolipoprotein E (ApoE) genotype has recently been reported as a major genetic risk factor for the development of AD (69). Specifically, homozygous possession of the є4 allele of ApoE has been found to be consistently associated with increased risk of late-onset AD, although the mechanisms underlying this association are not yet understood (11,19,20,42,44,46,81).

The ApoE gene is localized to chromosome 19q at the region previously found to show genetic linkage to Alzheimer's disease in late-onset families (81). The genetic analysis also indicates that the e4 allele of ApoE is over represented in subjects with AD compared with the general population. Possession of the e4 allele is also found to decrease the age of onset of AD (11,70,72,81). Rubinsztein and Easton (70) have estimated that 60% of late-onset and 92% of early-onset AD may be attributable to ApoEe4. On the other hand, there is some indication that possession of ApoEe2 may have a protective effect and may delay the onset of dementia (11,12,73,90).

Involvement of ApoE in the pathogenesis of late-onset AD was also suggested by immunohistochemical localization of ApoE to senile plaques and neurofibrillary tangles (60). The ApoE polymorphism has also been found to affect the response to head trauma as well as cerebral hemorrhage (50).

What may be the mechanism underlying the association of ApoEe4 with AD?

Several hypotheses have been proposed to account for the isoform-specific association of ApoE with AD (for review, see 48A). Since the ApoE e4 allele is also associated with cardiovascular pathology, it is possible that it contributes to neurodegenerative changes through vascular mechanisms. Several isoform-specific properties of ApoEe4 allele on binding to Aβ proteins, binding to [tau] protein, cholinergic deficits, neuronal morphology, neuronal degeneration, activation of microglia etc. have been reported. The amount of deposited amyloid-β peptide, Aβ, and its vascular load in the brains of patients with AD has been found to increase in proportion to the number of copies of ε4 alleles (48A).

ApoEe4 versus ApoEe3 or ApoEe2 is less efficient in neuronal repair as well as in the prevention of microtubular polymerization. ApoEe4 is also suggested to contribute to neuronal degeneration due to its less effective lipid transportation, which can cause disturbances in brain lipoprotein metabolism (15). ApoEe4 may also affect the blood-brain barrier transport of proteins like Aβ (48A).

Does thrombin play a role in the pathogenesis of Alzheimer’s disease?

Accumulation of higher immunoreactivity of thrombin in neuritic plaques in brains of patients with Alzheimer’s disease (1,2) and the demonstration of thrombin-induced neurotoxicity (39,67) in vitro suggests that thrombin may play a critical role in the pathogenesis of AD. Moreover, levels of the major brain thrombin inhibitor, protease-nexin-1, were reported to be lower in postmortem AD brains, (91) in particular around blood vessels (89), implicating increased thrombin activity in AD etiology.

Thrombin, in addition to playing a central role in the formation of blood clots, is a serine protease with diverse bioregulatory activity (Reviewed in 86). Thrombin has been demonstrated to induce neuronal cell death and to inhibit neurite retraction and branching of neuronal cells in in vitro studies (26,37,39,82,97). Thrombin treatment has also been shown to elevate [Ca²⁺] levels in vitro (78), which can lead to destabilization of the neuronal cytoarchitecture, leading to cell damage and eventual death (77,95). In our preliminary study, i.c.v. administration of thrombin induced significant deficits in working and reference memory in rats, which suggests that thrombin impairs cognitive function.

Interaction between thrombin and ApoE: implications in neurodegeneration

Head trauma, in which nerve cells are exposed to abnormally high levels of thrombin, is associated with a 1.8-fold risk for developing AD (59). Traumatic brain injury may also decrease the age of AD onset (61). Boxers, most of whom are subjected to repeated head trauma, have been shown to develop cognitive disorders like dementia pugilistica, characterized pathologically by diffuse Aβ deposits and neurofibrillary tangles [85]. These observations suggest that traumatic brain injury is a risk factor for AD.

Studies have also shown that both history of head trauma (with abnormally high levels of thrombin) and inheritance of apolipoprotein E e4 allele are associated with increased risk for development of AD (48,56,73,96). Also, AD risk is significantly greater in Ee4 carriers who have experienced head trauma than in those without a history of head trauma (48). Ee4 is also associated with neurological deficits in boxers (19) and is a negative risk factor for recovery from head trauma (22,83). Older football players possessing ApoEe4 have been reported to show lower cognitive performance (48). In spite of several studies showing the association between the risk factors such as ApoE and traumatic brain injury, and neurological deficits, the cellular and behavioral consequences of the interaction between traumatic brain injury and ApoEe4 are not yet known.

There is some in vitro evidence, which shows that thrombin, one of the vascular factors released in traumatic brain injury, cleaves ApoE into neurotoxic peptides (54). Truncated ApoE is found to increase intracellular calcium levels, which may mediate ApoE neurotoxicity (84). Interestingly, the 22-kDA ApoEe-4 derived fragment is found to be more toxic than the fragments of ApoEe2 or ApoEe3 (55). In spite of this in vitro evidence, there is no information about the effects of the interaction between thrombin and different ApoE isoforms on cognitive function.

Therefore, to have a better understanding of the interaction between thrombin and ApoEe4, we will determine whether possession of Apoe4 allele enhances thrombin’s neurotoxic effects (Aim 1). Our hypothesis-1 is that ApoEe4 dose-dependently enhances thrombin’s degenerative effects on working and reference memory. On the contrary, ApoEe2 appears to confer some protection to neurons from injury and delay the onset of dementia (11,12,73,90). Therefore, we will also examine whether ApoE e2 dose-dependently protects neurons from thrombin’s degenerative effects tested using a task involving reference and working memory.

What is the mechanism underlying neuronal injury in Alzheimer's disease?

There is a growing body of evidence suggesting that oxidative injury may be a major mechanism underlying neuronal damage and death in the pathogenesis of AD (3,6,41,94). The high lipid content and unusually high concentration of polyunsaturated fatty acids in the brain makes it particularly susceptible to oxidative injury. The high lipid content in the brain is suggested to promote the formation of additional reactive oxygen species and to result in protein and DNA oxidative damage causing neuronal injury.

Increasing evidence also indicates that most of the risk factors for the pathogenesis of AD such as Down's syndrome (68,88), vascular disease (79), head injury (61) are also associated with increased free radical formation (8,66,71,76). Today, although it is generally accepted that AD is a heterogeneous and multifactorial disease with multiple causes and pathogenetic mechanisms, oxidative stress is as yet considered to be a common mechanism underlying the pathogenesis of this disease (3,6,41,94).

Does thrombin-induced neurotoxicity involve oxidative injury? What is the role of apolipoprotein E in thrombin-induced oxidative injury?

We observed significant number of apoptotic cells in brain sections of thrombin-treated rats (Fig.8; Preliminary study). This indicates that thrombin-induced neuronal injury may involve oxidative mechanism. In Aim 2, we will determine the effects of ApoEe2 and ApoEe4 alleles on thrombin-induced oxidative injury. This will suggest whether effects of ApoEe4 isoform on thrombin-induced neurotoxicity are mediated by an increase in oxidative injury.

Is thrombin-induced neurotoxicity due to its proteolytic action?

Thrombin's neurotoxic effects are suggested to be mediated by proteolytic activation of thrombin receptor (86). In addition, Marques and coworkers (54,55) have proposed that ApoE, escaped and accumulated in the extracellular space, may be susceptible to in situ proteolysis by thrombin (0.1-1.0 micromole) like proteases, resulting in the production of the neurotoxic peptides (45,54). These neurotoxic peptides may play a direct role in AD pathology (Crutcher 13). Furthermore, full-length ApoEe4 has been shown to exhibit greater neurotoxicity than ApoEe3, an effect that is associated with production of truncated ApoE (54). Truncated Ee4, specifically Ee4-derived 22kDa fragment is reported to be significantly more toxic than the E2-derived fragment (54,55).

Recent studies have shown that a 10kDa C-terminal fragment of ApoE is complexed to Aβ in neuritic plaques [reviewed in (21)]. Studies have also shown that ApoE peptides can modulate amyloid formation in vitro (21). Moreover, thrombin cleavage of ApoE has been found to generate a similar C-terminal fragment that can form amyloid-like fibrils (21). Protease inhibitors have been found to reduce both the toxicity of full-length ApoEe4 and the fragmentation of ApoEe4, which suggests that proteolysis of ApoE may mediate its toxic effects (84). Though the mechanism by which ApoE-related peptides elicit toxic effects is unknown, there is consensus regarding the proteolysis of ApoE playing a role in generating neurotoxicity.

In Aim 3, we will determine whether administration of protease inhibitors suppresses degenerative effects of thrombin or combined effects of thrombin and ApoE on reference and working memory. This will suggest a probable mechanism underlying the interaction between thrombin and ApoE and will also suggest a therapeutic strategy for the prevention of head-trauma-associated neurodegeneration.

This proposal will lead to an improved understanding of the cellular and biochemical mechanisms leading to neural injury after head trauma and may facilitate the development of improved pharmacotherapy for head trauma-associated neurodegenerative disorders.

Significance: The better understanding of interrelationships of multiple substrates of dementia has major implications for neuroprotective and disease slowing therapies. The understanding of the interaction between thrombin and ApoEe4 will lead to prevention or a better management of neurodegenerative disorders associated with traumatic brain injury.

PRELIMINARY STUDIES

Our laboratory has extensively studied the involvement of cerebral vasculature in the pathogenesis of AD (9, 17,27-36,58,62). Recent data from our laboratory have demonstrated that both brain microvessels isolated from AD patients and biochemically disturbed rat brain endothelial cells in culture, secret factors that cause lethal injury specifically to neurons (33). Our laboratory also demonstrated that thrombin (one of the vascular factors released by inflamed endothelial cells) induces neuronal cell death in vitro (67). Thrombin has been demonstrated to inhibit neurite growth and branching of neuronal cells in “in vitro” studies (37-39). We also found that intracerebroventricular (i.c.v.) administration of thrombin causes cognitive deficits (57).

Experiment 1. Intracerebroventricular administration of thrombin causes cognitive deficits in rats. An eight-arm radial maze was used to investigate thrombin's short-term effects on working and reference memory in young male rats. Subjects were initially trained to retrieve bait from either odd or even arms without revisiting an arm.

Treatment: Group 1 (n=6) received the vehicle (0.35% BSA in saline) for 28-days delivered by subcutaneously implanted osmotic pump (flow rate = 0.25ml/hr). Group 2 (n=6) received thrombin (25 nM) for 28 days. Group 3 (n=6) received thrombin (100 nM) for 28 days.

Testing: Rats were tested on the 5^th, 10^th, 15^th, 20^th, 25^th and 30^th days following the beginning of thrombin treatment, for the retention of working and reference memory in the radial arm maze. Working memory was assessed by rat’s abilities to enter each baited arm only once. Reference memory was assessed by rat’s abilities to enter baited arms and avoid arms with no baits.

Results:

Body Weight: Body weight differences at the end of the treatment between vehicle-treated (370.35 ± 20.39gm), thrombin (25 nM) (372.25 ± 20.48gm) and (100nM) (374.83±20.35gm) –treated rats were not statistically significant.

FIG. 1A. Radial maze performance of subjects during training. Mean errors/four sessions (±SE). Rats did not receive any treatment in this phase

Behavioral Testing:

*

**

**

*

*

Pretreatment training: Performance of all the three groups during the training is shown in Fig.1A and 2A. The differences between groups along the twenty training sessions were not statistically significant for both kinds of memory (reference memory, p>0.05, working memory: p>0.05).

Tests during thrombin treatment:

FIG. 1A. Mean errors/four sessions (±SE)

Reference Memory: Mean reference memory errors scores of 6 test sessions of vehicle-, thrombin (25nM)-, thrombin (100nM)-treated rats are shown in Fig. 1B. The comparison between thrombin (100nM)-treated and vehicle-treated groups showed a significant group X session interaction. Therefore, multiple comparisons were done to evaluate the group effect. Significant differences between the performance of different

Fig.1B. Deficits in Reference memory in rats receiving thrombin treatment. *P<0.05, **p<0.01

groups were found for sessions on the 5^th (p<0.01), 20^th(p<0.0.05), 25^th(p<0.01) and 30^th (p<0.05) day. This suggests that thrombin (100nM) treatment affects the performance of subjects. The comparison between the groups treated with vehicle and lower concentration of thrombin (25nM) showed no significant group X session interaction, showing that thrombin (25nM) treatment has no effect on reference memory of subjects.

FIG. 2A. Working Memory performance of subjects during training. Mean errors/four sessions (±SE). Rats did not receive any treatment in this phase.

**

*

**

**

*

*

*

Working memory: There was a significant decrease in the number of errors over training sessions in all the groups. The working memory performance during training is shown in Fig. 2A, where mean number of errors/block of 4 sessions/group is plotted.

The comparison between the vehicle- and thrombin (100nM)-treated groups showed a significant group X session interaction (Fig.2B). Therefore, each session was separately examined. Significant differences between thrombin-treated group and vehicle-

treated group were found in sessions on the 5^th (p<0.05), 10^th (p<0.05) and 20^th day (p<0.05) of thrombin treatment.

Fig.2B. Deficits in Working memory in rats receiving thrombin treatment. *p<0.05, **p<0.01

	**

Fig.3: Thrombin (25, 100 nM) treatment increases mean latency. *p<0.05

Mean latency for completion of the task:

Fig.3 shows that even lower concentration of thrombin (25nM) significantly increases the average time taken to complete the task. Response latency was significantly increased following thrombin (25 and 100nM) treatment (p<0.05).

Fig.4: Thrombin treatment impairs the cognitive performance biphasically. **p<0.05

Thrombin treatment causes biphasic changes in the performance. Over the six recall tests conducted in 30 days (thrombin treatment duration-28 days), there were differential effects of thrombin treatment. The best performance of rats, which received thrombin (25 or 100 nM) treatment, was on the day 15. Rats, who received thrombin (100nM), made significantly more number of errors on day 5 and day 25.

We do not have any explanation of these biphasic effects of thrombin. It appears that the effects seen on day 5 are due to direct thrombin-induced neurotoxicity. However, we do see adaptation of the rats to this toxicity. The effects seen on day 20 and day 25 could be due to a secondary mechanism initiated by thrombin. Rats show a similar biphasic curve in the total time taken to complete the task. Interestingly, the mean latency for completion of the task is higher on the day 30, when rats are no longer receiving thrombin treatment. Currently, we are looking at the mechanism underlying thrombin’s biphasic effects on cognitive function.

Fig.5: Thrombin treatment during the training phase increases Time/Error ratio. (p<0.05)

FIG. 2A. Working Memory performance of subjects during training. Mean errors/four sessions (±SE). Rats did not receive any treatment in this phase.

Thrombin treatment during the training Phase: Thrombin (100nM) treatment during the training phase significantly increased the average number of errors per session and the time/error ratio (Fig.5), which suggests that thrombin also affects processes involved in learning.

Histology: After completion of the experiment, each animal was transcardially perfused with a phosphate-buffered saline solution (pH 7.4), followed by 10% formalin under anesthesia. Brains were post-fixed overnight at 4°C and were placed in 20% and 30 %sucrose in 0.1 mol/L PBS, pH 7.4, for 24 hours each. The brains were subsequently frozen in isopentane cooled to -30°C and cut into 20µm segments with a cryostat microtome. The 20-mm sections of the brain were stained with hematoxylin/eosin and silver stain to determine gross changes in the morphology and injection locations. The morphology of the coronary sections revealed that thrombin treatment causes significant enlargement of ventricles.

Brain section of thrombin-treated and vehicle-treated rats

Fig. 7. Brain section of section of vehicle-treated rat.

Fig.6. Brain section of thrombin-treated rat

TUNEL ANALYSIS

Brain sections from thrombin (100 nM) treated rats showed significant number of apoptotic cells (Fig.8). There were negligible numbers of TUNEL positive cells in the brain sections at the same coordinates of vehicle-treated rats (Fig.9).

TUNEL-positive cells in thrombin-treated rats

Fig.8. TUNEL positive cells in the brain section of thrombin-treated rat.

Fig.9: No TUNEL-positive cells in the brain section of vehicle-treated rat.

Conclusions:

1. Thrombin treatment causes significant deficits in the reference and working memory.

2. Thrombin treatment causes significant enlargement of ventricles.

3. Thrombin-induced neuronal injury may involve oxidative mechanism.