This means that the fate of cholinergic neurons may depend on a delicate balance between NGF, proNGF, and their respective receptors, tyrosine receptor kinase A (TrkA) and the neurotrophin receptor p75NTR. In fact, Margaret Fahnestock at McMaster University, Hamilton, Canada, recently showed that there is very little NGF in the human brain, whereas proNGF is easily detectable and gradually increases as patients progress from no cognitive impairment (NCI) to mild cognitive impairment (MCI) and on to mild and severe AD (see Peng et al., 2004). In addition, many groups, including Elliott Mufson’s at Rush University Medical Center, Chicago, have demonstrated that levels of tyrosine receptor kinase A (TrkA), the “pro survival” receptor for NGF, are reduced in both the nucleus basalis and cortex of people with mild to severe AD (see Hock et al., 1998, Counts et al., 2004, and Mufson et al., 1997), and that levels of TrkA correlate quite well with cognitive function.

These findings suggest that the loss of the pro-survival TrkA receptor, coupled with the increase in the pro-apoptotic proNGF could create an environment that contributes to the disappearance of cholinergic neurons with AD progression.

In Sorrento, Mufson gave an update on his ongoing analysis of samples taken through the Religious Order Study on aging and Alzheimer disease. This study conducts yearly cognitive and neurological examinations of hundreds of catholic priests, nuns, and brothers at over 40 religious communities throughout the U.S. who have all agreed to donate postmortem tissue. It has already provided invaluable information on cognitive decline (see, for example, ARF related news story on the link between AD and Parkinson disease), or more recently Bennett et al., 2005).

In his latest examination of postmortem samples, Mufson focused on the deterioration of basal forebrain cholinergic neurons and their cortical projections, and the relationship between that deterioration and NGF/proNGF signaling. Last October, Mufson and colleagues reported that in contrast to TrkA, cortical levels of p75NTR, which activates a neuronal cell death pathway, were unchanged in people with mild cognitive impairment (MCI) or mild to severe AD (see Counts et al., 2004). In Sorrento, he reported that in the cholinergic basal forebrain (CBF), neurons staining for both p75NTR and TrkA are reduced in both MCI and mild AD compared to NCI, whereas numbers of neurons expressing choline acetyltransferase (ChAT) are similar among the three groups. This suggests that dysfunctional neurotrophin signaling may precede the loss of cholinergic transmission in the CBF, said Mufson.

Corroborating this protein and immunocytochemistry data were the results from expression analyses on custom microarrays of single cholinergic neurons aspirated from tissue samples. This work is a collaboration with Steven Ginsberg at the Nathan Kline Institute in New York. Mufson reported that there are no significant differences between gene expression of p75NTR in MCI, early AD, or NCI, but levels of TrkA transcripts did turn out to be down in samples from MCI patients. Expression of ChAT was no different among the three groups, again pointing to neurotrophin dysfunction as a forerunner of cholinergic transmission problems. The loss of p75NTR in the basal forebrain neurons, despite the apparent normal expression of the protein, suggests that perhaps degradation of the receptor is accelerated in this particular region of the brain. In addition, the fact that p75NTR levels are maintained in cortical projection sites despite the phenotypic reduction in the expression of p75NTR in neurons of the basal forebrain indicates that other cells may be compensating by increasing expression of p75NTR in the cortex, suggested Mufson. All told, the data indicate that there may be a progressive loss of the “good” NGF signaling while the “bad” proNGF signals are maintained as patients progress.

Further support for the “too-little-NGF, too-much-proNGF” hypothesis comes from Antonio Cattaneo at Lay Line Genomics, Rome, and the International School for Advanced Studies, Trieste, Italy. Cattaneo has developed a line of transgenic mice, the AD11 mice, which are deficient in basal forebrain NGF because they express a high-affinity anti-NGF antibody in the central nervous system (see ARF related news story). These mice exhibit AD-like symptoms including long-term potentiation and behavioral deficits, and cortical degeneration. In addition, they have elevated levels of Aβ1-40 and Aβ1-42, plaque pathology, and tau pathology.

How much of this pathology might go back to changes in the cell survival versus cell death pathways activated by mature versus proNGF? Cattaneo reported that in the AD11 mice, the anti-NGF antibody reacts poorly with proNGF in the first place, and when it does bind it dissociates faster than it does from the mature trophin. Perhaps the antibody is neutralizing the mature form of NGF while leaving the pro-apoptotic immature form free to act through the p75NTR receptor and its recently identified binding partner sortilin? (See Nykjaer et al., 2004 and ARF related Sorrento news on the relationship of two other sorting proteins, SorLA and sorting nexin 30 to Aβ precursor protein processing).

To test this idea, Cattaneo bred AD11 mice with p75NTR knockout animals. Because these animals are missing the extracellular domain of p75NTR, signaling through the receptor is lost while signaling through the TrkA receptor should still work. These AD12 crosses have a less severe phenotype than their AD11 parents, Cattaneo revealed. Though cholinergic neurons begin to die in AD12 mice at two months—as they do in AD11 mice—by six months the cholinergic deficit has been completely rescued in AD12 animals. Amyloid deposits are also much reduced: At six months, amyloid deposition is down 20-fold compared to AD11 mice and at 15 months, AD12 animals have almost no amyloid deposition.

The tau pattern is not so simple, however. At two months, AD11 mice show little hyperphosphorylated tau (P-tau), but AD12 mice contain a large number of P-tau neurons in both the cortex and the hippocampus, reported Cattaneo. He has not yet worked out how to explain this. Even so, he suggested that the two forms of NGF can have profound consequences for the well-being of neurons depending on how their processing is modulated, and how their availability and signaling changes.

Margaret Fahnestock also addressed the role of mature versus immature trophic factors, in the progression of AD, but focused on brain-derived growth factor (BDNF). Researchers have known for a decade that BDNF mRNA levels are down in the hippocampus (see Murray et al., 1994). That study, using just a handful of postmortem samples from AD patients, has stood up well as numerous others have confirmed and extended it, including a report from Fahnestock’s group showing that levels of proBDNF are also reduced, by up to 40 percent, in the parietal cortex of AD patients (see Michalski and Fahnestock, 2003).

But are all BDNF transcripts to be tarred with the same brush? As Fahnestock pointed out, humans produce at least seven different BDNF transcripts and different promoters regulate their expression. This raises the possibility that there may be differential expression of the BDNF family members in response to the challenges posed in the AD brain.

To address this, Fahnestock used reverse-transcriptase PCR to measure levels of specific transcripts in human brain tissue. She reported that transcripts 1, 2, and 3 are all downregulated in the AD brain compared to controls. In keeping with this, levels of proBDNF were reduced in samples from the Religious Order Study by about a third in both MCI and AD patients, compared to NCI.

This downregulation might be caused by none other than Aβ. Fahnestock reported that when she treated SY5Y neuroblastoma cells with Aβ1-42, it downregulated expression of the BDNF 3 transcript and led to a reduction in the total amount of BDNF in the cells. In contrast to NGF, it appears that people with MCI and early AD lose both the mature and immature forms of BDNF, and that Aβ may play at least a partial role in reducing transcription of the trophin. This suggests yet another avenue whereby the amyloid peptide might compromise neuronal function.—Tom Fagan.