Research Report 2001-2003:

1. Top-Down perturbations using animal models of stress

An effective strategy to investigate factors accounting for the striking differences between hippocampal and amygdaloid function comes from behavioral studies in rodents in which repeated stress facilitates aversive learning but impairs spatial learning. While chronic stress produces hippocampal degeneration and impairs hippocampal-dependent learning, the basolateral amygdala (BLA) has been shown to be essential for stress-induced facilitation of aversive learning. In other words, the inherent differences between the outputs of these two structures can be amplified further by using stress as a “top-down” behavioral perturbation.

a. Effects of chronic stress on structural plasticity in hippocampal and amygdaloid neurons
Ajai Vyas, Rupshi Mitra, Savita Bernal, Abhishek Mukhopadhyay, Aditi Sarkar and Garga Chatterjee

We studied the effects of two different models of chronic stress on hippocampal and amygdaloid neuronal morphology in rats (Vyas and Mitra et al., 2002). In agreement with previous reports, chronic immobilization stress (CIS) induced dendritic atrophy in CA3 pyramidal neurons of the hippocampus. In striking contrast, pyramidal and stellate neurons in the basolateral amygdala (BLA) exhibited enhanced dendritic arborization or hypertrophy in response to the same CIS (Figure 1). Interestingly, while hippocampal atrophy was reversible, BLA hypertrophy persisted well beyond the duration of the CIS protocol (Figure 2b). Chronic unpredictable stress (CUS), however, had little effect on CA3 pyramidal neurons and induced atrophy only in BLA bipolar neurons. We have extended these studies and shown (Vyas et al., 2003) that similar patterns of structural plasticity are elicited in the bed nucleus of the stria terminalis (BNST), but not the central nucleus (CeA). These results indicate that the same chronic stress can cause contrasting patterns of dendritic remodeling in neurons of the amygdala and hippocampus. Our findings raise the possibility that certain forms of chronic stress, by affecting specific neuronal elements in the amygdala, may lead to behavioral manifestations of enhanced emotionality. There is also increasing evidence for contrasting effects of stress on both anxiety and learning in males versus females. Hence, we are extending our studies to examine the effects of gender difference and estrogen on structural plasticity in the hippocampus and amygdala.

Collaborators: Bruce McEwen and Gwen Wood, Rockefeller University, U.S.A.

b. Effects of chronic stress on behavioral models of anxiety and depression
Ajai Vyas, Anup G., Aparna Suvrathan and Adarsh S. Reddy
What may be the behavioral consequences of stress-induced morphological changes in the amygdala? We have initiated studies to explore the potential relationship between these neuronal changes and affective behavior in chronic stress-treated rats by measuring anxietylike behavior in CIS and CUS animals (Vyas et al., 2003). In these studies, CIS, but not CUS, caused enhanced anxiety (Figure 2a). These data suggest a correlation between stressinduced enhancement of anxiety and consolidation of aversive learning on the one hand, and stress-induced hypertrophy in the basolateral amygdala at the other. Moreover, the CISinduced increase in anxiety, like BLA hypertrophy, persists even after a 21-day period of stress-free rehabilitation (CIS+REHAB, Figure 2b). Taken together these findings provide a putative cellular model in the amygdala for affective disorders triggered by chronic stress. This model, along with its links to depression, is now being explored further using pharmacological and genetic manipulations.

Collaborator: Hikal Ltd., India.

Figure 2. a. CIS, but not CUS, caused greater anxiety compared to control animals. This was manifested as a significant reduction in open-arm exploration in the elevated plus-maze (**, p<0.01, t-test). b. CISinduced BLA hypertrophy and enhanced-anxiety persist after 21 days of rehabilitation (CIS+REHAB).

c. Effects of chronic stress on hippocampal synaptic plasticity
Rupshi Mitra, Shantanu Jadav and Anupratap Tomar

The most widely studied synaptic mechanisms underlying neuronal information storage are long-term synaptic potentiation (LTP) and depression (LTD), which are robust forms of persistent activity-dependent modifications of synaptic transmission. Our electrophysiological studies in hippocampal slices indicate that a 10-day CIS protocol does not affect LTP or its reversal in area CA3 of the hippocampus (Figure 3a). However, when CIS is continued for 21 days, it specifically prevents LTD of a previously potentiated input (Figure 3b). Interestingly, both CIS and CUS enhance paired-pulse facilitation (PPF), a measure of presynaptic release of neurotransmitters. In view of the auto-associative network architecture of the CA3 area, the observed changes in basal synaptic transmission may be indicative of the first hints of recurrent hyperexcitability and eventual excitotoxicity leading to neuronal atrophy and death of CA3 pyramidal neurons. These possibilities are now being explored with further electrophysiological studies as well as molecular biological studies on neurodegeneration and cell death. Finally, consistent with the effects on basal synaptic transmission, we have also observed complex and contrasting patterns of modulation in spine-density following chronic stress (Figure 1).

Figure 3. a. Theta burst stimuli (TBS) elicited LTP of field EPSPs in hippocampal commissural/ associational-CA3 synapses. The potentiated response undergoes depotentiation or reversal of LTP via LTD following low-frequency stimuli (LFS: 900 pulses at 1 Hz). CIS for 10 days did not impair LTP or its reversal. b. CIS, continued for 21 days, prevented LTD of the previously potentiated input (Post-TBS baseline).

d. Computational analysis of the effects of stress on hippocampal excitability
Rishikesh Narayanan and Anusha Narayan

In light of the findings described above, we are performing computational simulations using NEURON (Hines and Carnevale, 1997) to analyze the effects of stress-induced dendritic remodeling on hippocampal excitability at both single-cell and network levels. We developed a statistical algorithm that systematically prunes dendritic arbors of three-dimensional CA3 pyramidal cell reconstructions (Duke-Southampton Archive) to replicate our experimental data on stress-induced atrophy. Next, we imposed passive and active membrane properties on these reconstructions. Our simulations indicate that dendritic atrophy leads to: (i) enhanced cell excitability as manifested by increased firing frequencies (Figure 4a); (ii) transition from a bursting (Figure 4c) to regular-spiking pattern (Figure 4b); and (iii) differential increase in back-propagating action potential amplitudes along basal and apical dendrites. We also studied the effects of these single-neuron changes on emergent properties at the population level. These network-level simulations suggest that seizure susceptibility of the CA3 region increases with stress.

2. Bottom-Up genetic perturbations using transgenic and knockout mice

a. Effects of chronic stress on transgenic mice with over-expression of brainderived neurotrophic factor (BDNF)
Sumantra Chattarji, B. S. Shankaranarayana Rao and Deepti Nair
Investigations into the molecular mechanisms underlying stress-induced hippocampal pathology have provided evidence for a possible role of neurotrophins in stress and depression. For example, certain paradigms of chronic stress in rats result in decreased levels of brain-derived neurotrophic factor (BDNF) mRNA, an effect which is blocked by antidepressants. Certain antidepressants also prevent stress induced atrophy. Together these findings have contributed to the “neurotrophic hypothesis” of depression which states that a deficiency in neurotrophic support may contribute to hippocampal pathology during depression, and that reversal of this deficiency by antidepressant treatments may contribute to the resolution of depressive symptoms. However, a direct link between BDNF, stress induced hippocampal atrophy and depression has not been made. Thus, we have examined if an increase in BDNF activity will counteract the effects of stress with respect to CA3 dendritic atrophy. For this purpose, we used a transgenic mouse, generated by Prof. S. Tonegawa and colleagues, with forebrain specific overexpression of BDNF. Our data show that CIS failed to induce atrophy in apical dendrites of CA3 neurons in BDNF-Tg mice (Figure 5). BDNF over-expression did not affect CIS-induced increase in stress hormones or body weight loss. Moreover, as predicted by the neurotrophic hypothesis, the BDNF-Tg mice exhibited a decrease in depression-like behaviour in the forced swim test.

Collaborators: Susumu Tonegawa and Arvind Govindarajan, Massachusetts Institute of Technology, U.S.A.

b. Cortical spine plasticity and memory consolidation in p21-activated kinase (PAK) transgenic mice
Sumantra Chattarji, B.S. Shankaranarayana Rao and Shantanu Jadav

Memory formation and consolidation are thought to involve changes in the number and structure of synaptic connections. Ever since Cajal’s discovery of dendritic spines, the search for a structural basis of long-term memory has focused on plasticity of spines. Actin dynamics comprises a key determinant of spine plasticity and is regulated by p21-activated kinase (PAK), a major downstream effector of the p21 small GTPases Rac/Cdc42. In neurons, Prof. Tonegawa and colleagues observed that active PAK is associated with the postsynaptic density of spines. To further investigate the role of structural plasticity of spines in memory, they generated transgenic mice in which endogenous PAK activity is specifically impaired in the postnatal forebrain. In a collaborative project, we observed that cortical neurons of these mice have decreased spine density, increased mean spine size, and a higher percentage of shorter perforated spines. These structural alterations correlated with enhanced cortical LTP and impaired consolidation of long-term memory. These findings reveal a critical role for PAK in spine morphogenesis underlying memory consolidation in cortical networks.

Collaborators: Susumu Tonegawa and Mansuo Hayashi, Massachusetts Institute of Technology, U.S.A.