When she began this work five to six years ago as a postdoc in Sheng’s lab at MIT, there was no solid evidence for a link between caspases and synaptic plasticity, said Li, who has moved to the National Institute of Mental Health in Bethesda, Maryland. And while caspases had long been implicated in Alzheimer disease and other neurodegenerative disorders, their main role, presumably, was mediating cell death. Yet researchers had detected caspase activation at early stages of AD (Gamblin et al., 2003; Rissman et al., 2004), even at synapses (Yang et al., 1998; Mattson et al., 1998; Mattson et al., 1998), in the absence of apoptosis. These observations led Sheng, who is now at Genentech, South San Francisco, and colleagues to wonder whether caspases may in fact “have nothing to do with cell death in neurons,” Li told ARF. “That’s where I started.”

Several research groups have since published non-apoptotic roles for caspases in the nervous system—in dendrite pruning of Drosophila sensory neurons (Kuo et al., 2006; Williams et al., 2006) and in guidance of growth cones (Campbell and Holt, 2003). Now, Sheng, Li, co-first author Jihoon Jo, University of Bristol, UK, and colleagues provide evidence that caspases are critical for regulating synaptic strength during long-term depression (LTD), a temporary decline in a neuron’s responsiveness to neurotransmitter signal, which is important for development and learning.

The researchers did a series of electrophysiology experiments to examine how caspase activation affects neuronal function. In one analysis, they prepared hippocampal slices from three- to four-week-old rats, treated them with various caspase inhibitors, and measured excitatory post-synaptic potentials in the CA1 region. These studies showed that two inhibitors—DEVD-FMK (selective against caspases 3, 7, and 8) and LEHD-FMK (caspase-9-prefering with weaker activity against caspases 6, 8, and 2)—blocked LTD induced by low-frequency stimulation, but had no effect on long-term potentiation (LTP) or basal synaptic strength. Inhibitors selective for caspases 1 and 8 did not block LTD, suggesting that caspases 3/7 and 9 were likely responsible for the synaptic plasticity effects.

The findings with peptide inhibitors found support in a second set of experiments using molecular approaches to block caspases with greater specificity. Here, the researchers overexpressed several caspase-inhibiting and anti-apoptotic proteins in CA1 neurons of hippocampal slice cultures made from six- to eight-day-old rats. In whole-cell patch-clamp recordings, they found decreased LTD in cells transfected with proteins that block caspase-3/7 (XIAP-Bir3) or caspase-3 and -9 (Bcl-xL). These proteins had no effect on LTP, though, indicating their overexpression did not harm synaptic function and plasticity in a non-specific manner.

The researchers confirmed the pharmacological and overexpression findings by measuring LTD in CA1 neurons of hippocampal slices from two- to four-week-old caspase-3 knockout mice, which reach adulthood with minimal brain pathology (Leonard et al., 2002). Low-frequency stimulation triggered LTD in wild-type but not caspase-3-deficient cells. Both had normal LTP induction and basal synaptic transmission, again suggesting the caspase-mediated effect on neuronal physiology is quite specific.

In line with its impact on LTD, blockage of caspase activity prevented AMPA receptor endocytosis, the molecular phenomenon that makes neurons less activatable. Sheng and colleagues demonstrated this in immunohistochemical experiments on hippocampal neurons either treated with caspase inhibitor peptides or transfected with caspase-inhibiting factors, and in caspase-3-deficient cells. Furthermore, the researchers found that the same NMDA treatment that induces LTD or AMPA receptor internalization in the neurons could activate caspase-3 and caspase-9, its upstream activator in the mitochondrial apoptotic pathway. The activation was modest, transient, not associated with eventual cell death, and able to be blocked with an NMDA receptor antagonist or calcium chelator. Moreover, the scientists showed that the pro-survival kinase Akt3, a substrate of caspase-3 during apoptosis, is also key for caspase-3-mediated LTD. All told, the data suggest that “the upstream signals activating the mitochondrial caspases in cell death and in LTD are very similar. They depend on calcium and on protein phosphatases,” Li said. “Yet the cellular consequences of activating caspases in these contexts are very different. We are curious as to why you have such different outcomes from activation of the same molecule.”

Steve Barger of the University of Arkansas for Medical Sciences, Little Rock, was similarly intrigued. “Here’s something that seems, on the face of it, to be degradative and harmful, and yet it’s an important component of physiology,” he said of the newly proposed function of caspases in LTD. One could imagine caspases playing a role in removing spines or whole branches of dendrites, which may be important for certain kinds of learning and memory but “wouldn’t lead to death of the entire cell,” Barger said.

Before this presumed killing of spines, it is conceivable that activated caspases have subtler physiological effects that might eventually contribute to neurodegenerative disease. Though the current study does not directly address this, Li raised the possibility that in AD, caspases may be active in early, pre-symptomatic stages, but “instead of leading to cell death, they may be involved in decreasing synaptic transmission.” If so, early intervention with caspase inhibitors “may help slow the deterioration of neuronal function,” she suggested. (Rohn, 2010 describes potential AD therapeutic approaches targeting caspases.)

A recent report suggests that soluble Aβ facilitates LTD (Li et al., 2009 and ARF related news story), consistent with the authors’ speculation that caspase-3-mediated LTD may contribute to synapse loss in AD. And in multiphoton imaging studies, caspase activity was dissociated from cell death and showed up prior to tangle formation in the brains of pathogenic tau-overexpressing transgenic mice, supporting the idea of caspase activation as an early, non-apoptotic event in AD (de Calignon et al., 2010 and ARF related news story).

Elliott Mufson, Rush University Medical Center, Chicago, speculated that, given all the molecular dysfunction in AD (e.g., hyperphosphorylation of tau, overexpression of neurotrophins), there could be a “shift in the equilibrium of how caspases are held in check as you get older, that might even get more out of whack when you develop AD.”

Andrea LeBlanc thinks the real test for whether caspase-3 is indispensable for LTD would be to check behavior and cognition in caspase-3 knockout mice. However, these animals have vision and hearing problems (Takahashi et al., 2001) that prevent their use in behavioral studies, Li said.—Esther Landhuis.

Reference:
Li Z, Jo J, Jia JM, Lo SC, Whitcomb DJ, Jiao S, Cho K, Sheng M. Caspase-3 Activation via Mitochondria is Required for Long-Term Depression and AMPA Receptor Internalization. Cell. 2010 May 28;141:859-871. Abstract