26 June 2010. Scientists know presenilins act as key proteins in the pathogenesis of Alzheimer disease (AD), given the role of mutant presenilin-1 (PS1) and PS2 in familial AD (FAD), but what precisely do presenilins do? A growing body of evidence agrees that FAD-PS mutants are connected to dysregulated calcium signaling in the brain. But there is sharp disagreement about the mechanism, with multiple competing hypotheses under investigation by different groups. In a paper in the June 23 Journal of Neuroscience, researchers led by Ilya Bezprozvanny of the University of Texas Southwestern Medical Center, Dallas, provide more evidence in support of one of these hypotheses, namely, that one of the normal functions of presenilins is to act as a calcium leak channel, slowly releasing calcium stores from the endoplasmic reticulum (ER). When this function is disrupted, excessive calcium stores build up in the ER, leading to aberrant calcium signaling that may contribute to AD pathology.
The current study builds on previous work by the same group, which used in vitro lipid bilayers and mouse fibroblast cells to show that presenilins act as leak channels that allow calcium to escape from the ER (see ARF related news story on Tu et al., 2006). The calcium leak channels balance the calcium influx activity of the smooth endoplasmic reticulum calcium ATPase (SERCA) pumps, maintaining homeostasis in the cell.
To extend this finding, first authors Hua Zhang and Suya Sun showed they could get the same results using hippocampal neurons, cells that are directly relevant to AD. They used two transgenic mouse strains: one was a PS1/PS2 double-knockout, the other, an FAD PS mutant knock-in. Calcium release from the ER ryanodine receptors was enhanced in hippocampal cells from both strains, agreeing with findings from the lab of Grace Stutzmann of Rosalind Franklin University, Chicago, Illinois (see ARF related news story on Chakroborty et al., 2009), and suggesting a swollen ER calcium pool. The authors then showed they could restore normal calcium signaling by adding wild-type presenilins or mutant presenilins with intact leak function, but not with mutant presenilins lacking leak function. The findings indicate that normal PS is needed to maintain ER calcium homeostasis.
The enhanced calcium release through the ryanodine receptors implied that the ER calcium pool in mutant neurons was larger than in wild-type mice. The authors tested this by using the ion carrier ionomycin to dump all the calcium from the ER, allowing them to directly measure the size of the calcium pool. Consistent with their hypothesis, they found that the ER pool was almost threefold larger in mutant hippocampal cells compared to wild-type. The authors also directly measured calcium leakage by comparing the size of the ER calcium pool after incubating the neurons in calcium-free media to draw out ER calcium. In the mutant neurons, the pool size decreased by only 7 percent, compared to 65 percent in wild-type neurons, indicating that not much calcium leaked out in neurons that lacked presenilins or expressed the FAD mutant form. All of these data pointed to presenilins acting as calcium leak channels in the ER.
In addition, the authors further investigated the role of ryanodine receptors. In agreement with previous work in the field (Chan et al., 2000), they found that ryanodine receptors were upregulated in neurons lacking functional presenilins. They hypothesized that this compensates for the loss of presenilin leak channels. To test this hypothesis, they blocked ryanodine receptors with the inhibitor dantrolene, or used RNAi to knock down the receptor expression, and found that the size of the ER calcium pool in mutant hippocampal neurons increased another threefold. Finally, the authors fed dantrolene for several months to APP/PS1 transgenic mice, which resulted in more severe neuropathology, including higher amyloid load, more synaptic loss, and more neuronal atrophy.
“This is the most interesting and unexpected finding in the paper,” Bezprozvanny said, “that we are able to make this AD mouse model much worse by blocking the ryanodine receptor.” Bezprozvanny said his group is now investigating the mechanisms behind the increase in amyloid load in the dantrolene-fed AD mice. “It indicates that changing the calcium levels in the ER has a big effect on amyloid.” This suggests that the connection between calcium levels and amyloid load is bidirectional, Bezprozvanny said. “You can drive amyloid pathology by calcium, or you can drive calcium pathology by amyloid. People usually think of these two pathways as parallel, but actually they are very closely linked.”
These observations on mutant presenilins may relate to late-onset AD by pointing research in the direction of pathways that also affect sporadic AD, Bezprozvanny said. Microarray and proteomic studies have shown there are massive changes in the expression level of calcium signaling proteins in patients with sporadic AD. “Any drug that will help reduce ER calcium levels in aging neurons may potentially have therapeutic value for AD,” Bezprozvanny said. He suggested that ryanodine receptors might be a useful therapeutic target as well, since they appear to help compensate for calcium dysregulation by helping unload excess calcium from the ER.
Reaction to these new findings is mixed. Stutzmann called the work solid and internally consistent, and said the data “further validate and extend Bezprozvanny’s previous findings, which were met with a certain amount of skepticism.”
Kevin Foskett of the University of Pennsylvania, Philadelphia, however, expressed doubt that presenilins act as calcium leak channels. Work in his lab instead shows that presenilins act by modulating calcium release through the IP3 receptor (see ARF related news story on Cheung et al., 2008; see also Cheung et al., 2010). Bezprozvanny said that their experiments were designed to bypass IP3R activation, so that IP3 receptor activity is not likely to contribute to the observed effects.
The current paper is unlikely to lay the controversy to rest. Stutzmann pointed out the technical difficulty of conclusively proving that presenilins act as leak channels, due to the limitations of electrophysiology technology. Present technology has a low signal-to-noise ratio when looking at such small channels, Stutzmann said, making it difficult to be certain of what you’re seeing. Stutzmann believes that several ER channels may contribute to calcium signaling, and said “a more comprehensive look at all of the ER calcium channels should divvy up relative contributions to pathology.”
Some researchers also questioned the interpretation of the in vivo dantrolene experiments, and recommended following up on these results with rigorous biochemistry and investigation into the mechanisms involved. This would include measurement of soluble Aβ and APP levels in the brain, for example. Bezprozvanny is planning to do those exact experiments.
Despite the controversy, the new findings add to the evidence that calcium signaling plays a key role in the pathogenesis of AD. “I think it represents a very significant paradigm shift away from [a focus on] the late-stage plaques, tangles, and histopathology, to early signaling mechanisms that can hopefully be normalized,” Stutzmann said. Most current therapies “try to clear plaques or restore cognitive function, and usually all they’re capable of doing is just delaying the disease.” By contrast, addressing calcium signaling might help prevent neuropathology, Stutzmann speculated, because it could be addressing one of the fundamental mechanisms behind AD, not just a late-stage symptom.—Madolyn Bowman Rogers.
Zhang H, Suya S, Herreman A, De Strooper B, Bezprozvanny I. Role of Presenilins in Neuronal Calcium Homeostasis. J Neurosci. 2010 June 23;30(25):8566-8580. Abstract