But that may soon change. For the first time, researchers have been able to measure the production and clearance of Aβ peptides in the central nervous system of human volunteers. In a clinical research tour de force, Randall Bateman, David Holtzman, and colleagues at Washington University School of Medicine in St. Louis, Missouri, used stable carbon isotope labeling and mass spectrometry analysis of Aβ from the CSF of normal human volunteers to quantify protein turnover. Surprisingly, they found that Aβ is made and degraded very rapidly in healthy volunteers, with the second-fastest in vivo turnover rate recorded for any protein so far.

The new method will be useful for studying perturbations in Aβ metabolism in AD, and for rapidly assessing the effect of therapies aimed at blocking production or increasing clearance of Aβ. Also, the same approach can be used to measure the turnover of any CNS protein that ends up in the cerebrospinal fluid, a boon for studying the metabolism of specific proteins—tau, prions, α-synuclein—that cause or contribute to other neurodegenerative diseases.

Holtzman and colleagues previously showed that steady-state levels of Aβ in the CSF predict how much Aβ is in the brain (see ARF related news story). CSF Aβ is nearly all derived from neuronal production, and the peptide enters the CSF on its way to the bloodstream to be degraded in peripheral sites. But measuring steady-state levels does not give much information about Aβ dynamics, and what the researchers really wanted to know was how fast Aβ was being made and cleared in the CNS, and how risk factors for AD might affect those parameters in people.

To get those answers, the researchers developed a method for stable 13C isotope labeling coupled with mass spec analysis. After intravenously infusing volunteers with 13C-labeled leucine, they collected CSF by lumbar puncture. Aβ was immunoprecipitated with a pan-peptide specific antibody, and tryptic fragments analyzed by LC-MS (liquid chromatography followed by mass spectrometry). By this procedure, the researchers could separately quantify unlabeled Aβ and labeled Aβ, which stood out by the mass shift in the tryptic fragment Aβ17-28. The labeled peptide was six units larger than the unlabeled—one mass unit for each of the six heavy carbons in the peptide’s lone leucine residue.

By optimizing dosage of 13C leucine and sampling times, the researchers were able to quantify both the production and clearance of Aβ. Taking CSF samples every hour, they measured the ratio of labeled to unlabeled Aβ. No label was detected for the first 4 hours, after which the fraction of heavy peptides increased between hours five and 13, then remained steady from 13 to 24. From 24 to 36 hours, the fraction of labeled protein decreased. Separately tracking the build-up and decay of the labeled protein allowed the investigators to calculate fractional rates of protein synthesis and clearance.

In six healthy subjects, Aβ dynamics were consistent. The average fractional synthesis rate was 7.6 percent, meaning that that much of the total CNS Aβ pool was produced each hour. Similarly, the fractional clearance rate was 8.3 percent—the production and clearance rates were not significantly different. This rapid turnover means that it would only take 6 to 7 hours to replace half the β amyloid found in a person’s CNS, according to a press release from the investigators. The fast turnover is surprising, given the slow accumulation of Aβ that occurs over years. The rate is comparable, though a bit faster, to the turnover of Aβ measured in mouse brain by in vivo microdialysis (see ARF related news story).

The technique, while safe, is intensive, requiring 36 hours of CSF collection and 2 days in the clinic. Of 10 people enrolled in the initial study, two stopped because of post-puncture headache. And the length of the procedure may present practical and ethical issues for studies of demented patients. On the positive side, when Holtzman presented the results in a seminar in May at the Merck Research Labs in Boston, he estimated that researchers would have to analyze just five people to determine accurately if a drug treatment changed Aβ synthesis or degradation. Holtzman also mentioned a technique in the works to double-label subjects with an early pulse of leucine, and a later pulse of another amino acid, like 13C phenylalanine. This would allow measurement of rates of synthesis at two different times in one subject, an ideal setup for testing, for example, γ-secretase inhibitors.

Work is continuing with the new method, testing candidate and current therapies for effects on Aβ turnover. Also, the researchers want to determine if changes in Aβ production or clearance can be detected in people who have already accumulated amyloid as opposed to those who have not, and in normal people with different ApoE phenotypes, Holtzman said in his presentation. They are currently studying a small group of people with PS mutations, hoping to get a look specifically at Aβ42 dynamics. They are looking at other metabolites of APP, as well. For example, soluble APP α fragment enters CSF in much higher rates than Aβ, and may provide an additional readout to follow amyloid production. With more results soon to follow, Bateman and Holtzman’s technique opens a welcome new window on the comings and goings of Aβ in the brain.—Pat McCaffrey.

Reference:
Bateman RJ, Munsell LY, Morris JC, Swarm R, Yarasheski KE, Holtzman DM. Human amyloid-beta synthesis and clearance rates as measured in cerebrospinal fluid in vivo. Nat Med. 2006 Jun 25; [Epub ahead of print] Abstract