In a poster presentation, Yinxing Liu, from the laboratory of Lou Hersh at the University of Kentucky, Lexington, reported that boosting neprilysin in the circulation of transgenic mice can reduce Aβ deposits in the brain. Liu used a retroviral approach to inject a neprilysin-expressing construct into the hind leg of nine-month-old 3xTG mice (Oddo et al., 2003). The secreted form of mouse neprilysin reached levels of 400-1,000 nmol product/min/ml as measured in an assay using an artificial substrate (activity in normal plasma is ~0.2 nmol/min/ml). After three months of this, the plasma levels of Aβ dropped from about 7 pM to 4.5 pM. Interestingly, brain Aβ also fell. This might be explained by the peripheral sink hypothesis, which suggests that lowering Aβ in the blood eventually pulls it out of the brain as well. This has been seen with other therapeutic approaches (see ARF related news story), including vaccines (see ARF related news story). Liu reported that in addition to Aβ deposits being cut in half, soluble Aβ in the brain was down by almost a third compared to untreated transgenic controls, but he reported no changes in behavior.

Researchers at Eliezer Masliah’s lab at the University of California, San Diego, also described a viral-neprilysin approach. In a slide talk, Brian Spencer reported a lentiviral construct that targets neprilysin for passage across the blood-brain barrier (BBB). The construct fuses a secreted form of neprilysin with the low-density lipoprotein receptor-binding domain of apolipoprotein B. In theory, this domain should help ferry the neprilysin in and out of cells and into the brain.

Spencer reported that even though it is conjugated to the ApoB domain, the secreted neprilysin degrades Aβ in vitro. In vivo, three months after a single intraperitoneal injection of the vector into transgenic mice expressing APPSwe under the Thy 1 promoter, the fusion protein was present in the brain and brain neprilysin activity went up. The increase was accompanied by a reduction in Aβ deposits and soluble Aβ monomers. APP and oligomeric Aβ levels were not changed. Spencer said he was not sure why oligomers were unaffected, but that other neprilysin studies had reported the same thing. “There may not be a direct line from the monomeric Aβ to oligomeric Aβ to plaques,” Spencer told ARF via e-mail. Rather, the plaques might be one outcome of accumulation of monomeric Aβ and free oligomers may be another outcome. “Further studies need to be performed to determine the exact relationship neprilysin plays in the accumulation of other Aβ species,” he said.

Spencer had no behavioral data to show whether this approach improves learning and memory in the transgenic animals, but he did show that the construct was found predominantly near neurons and glia in the dentate gyrus of the hippocampus.

“One of the problems with current methods of delivering neprilysin is that they are all intracranial, which would not be feasible for humans,” Spencer said. Dave Morgan of the University of South Florida, Tampa, agrees. “At best, using intracranial injection, we can get neprilysin into about one third of the mouse brain,” Morgan commented. “The human brain is 1,000 times the volume of the mouse brain, which means we’d have to turn it into a pin cushion if we wanted to use intracranial delivery,” he said. Instead, Morgan’s lab has come up with a novel way of sneaking neprilysin into the brain—by expressing it in monocytes. If practical in humans, this approach would be used beyond AD. “Any CNS disorder that has a significant macrophage activation component could be theoretically amenable to treatment using this technique,” Morgan said.

In her poster presentation, Lori Lebson from Morgan’s lab demonstrated how monocytes expressing a secreted form of neprilysin prevent buildup of Aβ plaques in transgenic mice. Lebson isolated GFP-expressing macrophages from a mouse line and transfected them with a plasmid that expresses a secreted form of neprilysin (the membrane-binding domain is replaced with a secretory signal). First Lebson injected these neprilysin-secreting monocytes directly into the cortex and hippocampus of 15-month-old double transgenic mice (APP/PS1) to determine if they would do any good. She found a drop in soluble Aβ and in Congo red-positive deposits after one week, whereas a control experiment using an inactive neprilysin construct did not.

Next Lebson tested if monocytes injected into the blood could enter the brain and degrade Aβ. Several labs have shown that circulating immune cells can cross the BBB in AD mouse models (see ARF related news story). Morgan noted that it has been theorized that monocyte therapy could work, though no one has been able to prove it. Lebson injected five million monocytes twice weekly into nine-month-old transgenic mice via a microvascular port attached into the jugular vein; this gives a more consistent injection pattern than trying to inject into mouse tail veins, for example, and allows for repeated injections over weeks. After two months, the researchers found that the monocytes completely reduced the buildup of new Aβ plaques but that it had no effect on plaques that were already in the brain. “That’s a very important point,” said Morgan. “Very few people are measuring Aβ load at initiation, but without that data point we would be saying we reduced amyloid load by half.” Over the two months, plaque load doubled in untreated mice.

Intriguingly, while the researchers found that monocytes entered the brain of transgenic mice and congregated in the vicinity of plaques, they found that absolutely no monocytes found their way into the brain of control animals after injection. This indicates that there has to be some signal from the brain, or perhaps damage to the BBB, which facilitates the entry of monocytes into the brain of transgenic models.

Morgan said the advantage of this method is that the therapy can be directly targeted to the site where it is needed. In addition, it could be used to deliver other genes or treat other diseases. “We are not so interested in neprilysin as in showing that the monocyte therapy itself can work,” he said. In the end, neprilysin may not be the best therapeutic approach because as a protease it is fairly “promiscuous,” he said, and so may have untoward effects.

Morgan suggested that the way forward using this methodology is first to show that it would work in an acute setting, where a patient’s own blood cells are transfected and re-introduced. Because monocytes are short-lived, this approach would be relatively safe. In the mouse circulation, for example, the GFP monocytes were undetectable within 90 minutes and lasted only about a week in the brain. If that holds true in humans as well, then treatment could be withdrawn easily. For a long-term therapy, Morgan predicted that a viral approach would have to be used to transfect patients’ own stem cells to keep an Aβ-targeting monocyte population going over the long run. Monocytes are known to change phenotype when they enter the brain. Therefore, the transfected, therapeutic gene should be driven by those promoters that get activated when the cells make this phenotypic switch, suggested Morgan.—Tom Fagan.