The epidemiologic link between elevated blood homocysteine (HC) and Alzheimer disease (see ARF related news story) has spurred the search for the mechanisms by which this ubiquitous amino acid might damage neurons. HC has been directly linked to AD by experiments showing it sensitizes neurons to the toxic effects of soluble amyloidβ (Aβ) peptides (see, e.g., Ho et al., 2001; Kruman et al., 2002; and Alzforum live discussion). Now, to add insult to injury, new results suggest that a metabolite of HC, homocysteic acid (HA), can act via γ-secretase to raise levels of intracellular Aβ peptides.
In a paper published May 16 in the Journal of Neuroscience Research online, Tohru Hasegawa in Saga, Japan, and collaborators show that high concentrations of HA increase intracellular Aβ1-42, but not Aβ1-40, in cultured cortical neurons and in APP-expressing CHO cells. The researchers also report that a γ-secretase inhibitor can protect neurons from the toxic effects of HA, suggesting a role for Aβ production in the neurotoxicity of HA. The physiologic significance of the finding, however, remains to be seen, as the increases in Aβ are elicited at concentrations that exceed the levels of HA the researchers could measure in plasma or CSF.
For the study, Hasegawa treated cultured cortical neurons with homocysteine and homocysteic acid, and measured Aβ1-42 and A β1-40 in the culture medium or in cell lysates. In cells treated with 1 microM or more HA, they detected an increase in Aβ42, but not Aβ40, associated with cells. In CHO cells expressing human APP (Swedish mutation), intracellular Aβ1-42 was elevated after exposure to 10 microM HA. Cortical neuron cultures treated with 1 microM HA lost about half the neurons within 48 hours, but the cells could be completely protected by pretreatment with the γ-secretase inhibitor LY-411,575, suggesting a role for Aβ production in neuronal toxicity of HA.
When the researchers measured the concentration of HA in normal individuals and AD patients, they found both groups had CSF levels in the 100 nM range. Plasma levels were undetectable. They did see the expected difference in both plasma and CSF levels of HC, where AD patients had CSF levels of HC around 700 nM, nearly twice as high as unaffected individuals.
Despite the widely different concentrations of HA present in vivo and in vitro, the results are intriguing. HA is produced from HC in the brain, and can promote calcium influx via NMDA receptors, a signaling pathway recently shown to increase Aβ production (Pierrot, 2004). Increased oxidative stress could also increase Aβ production. Compared to the cortical cultures studied here, HA has been shown to be more potently toxic for hippocampal neurons (Lockhart, 2000). If the results hold up, they may suggest a “potential therapeutic benefit of agents that modify the production and neurotoxic actions of HA and homocysteine,” the authors conclude.—Pat McCaffrey