One of the shortcomings of most mouse models of Alzheimer disease is that they fail to recapitulate the rampant neurodegeneration that is seen in mid- to late stages of the disease in humans. This has led to the idea that in mice at least, Aβ alone is not sufficiently toxic to cause neuron loss—additional factors, such as tau pathology or even beyond tau, may be a co-requisite. In this week’s Journal of Neuroscience, researchers led by Michael Lee at Johns Hopkins University, Baltimore, Maryland, offer a slightly different view. They report that in PS/APP double transgenic mice, there is significant neuron loss—of brainstem monoaminergic neurons, that is, which project into the forebrain from the locus ceruleus (LC). Because LC neurons that do not project to the cortex are unaffected, one explanation is that pathology in the cortex—be it Aβ deposition, inflammation, or some other unknown entity unleashed in this model—leads to degeneration of the brainstem monoaminergic neurons that project into the cortex. The findings suggest that at least some mouse neurons are susceptible to robust and progressive neurodegeneration in the absence of tau pathology.

Loss of monoaminergic neurons—serotonergic neurons in the raphe nuclei and noradrenergic neurons in the LC—has been well documented in AD patients and may even occur early in the disease (see Grudzien et al., 2007). Some LC noradrenergic neurons are also lost in both humans (see Marien et al., 2004) and mice (see Leslie et al., 1985; Sturrock et al., 1985) during normal aging, but whether these neurons are affected in AD mouse models has been more controversial. One study suggests minimal effect on LC neurons in single transgenic (V717F APP) mice (see German et al., 2005), but more recent studies suggest significant loss of tyrosine hydroxylase (TH), a marker of noradrenergic neurons, in the LC of APP/PS1 double transgenic mice (see O’Neil et al., 2007) and even in single transgenic Tg2576 mice by as early as eight months (see Guerin et al., 2007). In showing a temporal loss of both TH and LC cell bodies, this new paper by Lee and colleagues seems to confirm that LC neurons are vulnerable in an AD-like setting.

Lead author Ying Liu and colleagues first examined four- to 18-month-old APP/PS1 transgenic mice for tyrosine hydroxylase. They found that in the cortex and hippocampus there was progressive loss of TH-positive afferent axons. Four-month-old TG mice appeared normal, but in 12-month-old animals there was significant TH loss in the motor and barrel cortices, and in the CA1 and dentate gyri of the hippocampus. The amygdala, which is spared Aβ deposition until much later in these animals, had normal TH levels even in 18-month-old mice. In single APP or single PS1 transgenic animals, the researchers found no loss of TH, even in 18-month-old mice.

To test if frank neuronal loss accompanied this TH loss, the authors used stereomicroscopy to examine neurons in the LC and the dorsal raphe nuclei. In 12-month-old animals the numbers of neurons were similar to those in wild-type animals, but by18 months there was a 50 percent reduction—only in the LC. The results support the idea of a progressive neurodegenerative process that starts in the NA axons that innervate the cortex and hippocampus, and eventually leads to loss of NA neurons in the LC. “The results of the present study demonstrate that the APPswe/PS1ΔE9 mouse model of AD recapitulates the progressive degeneration of MAergic neurons occurring in AD,” write the authors.

“This is an elegant study that clearly shows damage to noradrenergic afferents in the cortex, which is consistent with previous work,” said Doug Feinstein, University of Illinois Chicago, in an interview with ARF. Feinstein was not involved in this work but has studied noradrenergic loss in AD. One thing he questioned is whether the timing accurately mimics what is seen in AD patients, which can have noradrenergic loss very early in the disease (see Grudizen et al., 2007). “One important question is when does loss of, or damage to, these neurons happen?” asked Feinstein. He said that since NA neuron loss has been documented in very mild AD, it suggests that what goes on in these older double transgenic mice, which already have rampant Aβ deposits, could be different. “That, or there could be more subtle damage occurring earlier that they didn’t see,” he suggested.

Michael Heneka, University of Bonn, Germany, agrees. “It’s a bit surprising that the LC degeneration appears so late,” he told ARF (and see also comment below). This raises the possibility that the direct cause of the damage might not be Aβ, but some secondary mechanisms, such as damage of synapses and axons through inflammatory mediators or excitotoxic stimuli, he suggested.

Both Feinstein and Heneka agreed that how and why noradrenergic axons are damaged requires further study. Why do these neurons in particular degenerate, when other cortical and hippocampal neurons are relatively preserved in these transgenic mice despite the abundance of Aβ? Liu and colleagues ruled out the direct involvement of Aβ in the LC by immunostaining for both Aβ deposits (4G8 antibody) and soluble Aβ (Aβ42-specific antibody). Likewise, they found no evidence of phosphorylated tau (AT8 and PHF1 immunoreactivity) within NA cell bodies, suggesting that toxic tau also has no or limited role in noradrenergic loss.

Feinstein said that the LC neurons are notoriously sensitive. Studies have shown that upon injection of the neurotoxin DSP4, the LC NA neurons are damaged or die but other NA neurons are spared. “So there may be selective vulnerability,” he said. The authors agree. “The lack of Ach neuron loss in Tg mice may reflect species difference in cellular vulnerability,” they write. They also propose that axon length may be a factor. “Alternatively, longer cortical afferents on MAergic neurons, compared with the Ach neurons, may increase vulnerability of MAergic neurons to defects in retrograde support,” they write. Feinstein expressed a similar sentiment. He suggested that loss of trophic support due to damage to glia and neurons in the cortex and hippocampus might play a role. “NA afferents can pick up neurotrophic factors, which, if not being produced, could cause damage in other brain regions,” he said.

One other aspect of AD attributable that may be linked to noradrenergic loss is heightened anxiety. Liu and colleagues found that loss of noradrenergic LC neurons precedes increased anxiety in these mice, as judged in an open field test, which lends some support to that hypothesis.—Tom Fagan


  1. This is an important and well-conducted study that may shed some light on the mechanisms that underlie the degeneration of the locus ceruleus (LC) in the course of Alzheimer disease (AD).

    LC degeneration is a very early and substantial feature of AD, as evidenced by studies from Grudzien et al., 2007 and many others. Importantly LC degeneration is greater than degeneration of the nucleus basalis Meynert in AD (Zarow et al., 2003). Before the loss of LC neurons, however, one could imagine a period of neuronal dysfunction and thus the starting point, its cause, nor the dynamics of LC degeneration in humans are known. In the present analysis, Liu and colleagues describe the degeneration of axons within LC projection areas at 12 months of age causing a significant 30-40 percent loss of norepinephrine (NE) levels in LC projection areas such as the cortex and hippocampus at 18 months of age. It is important to note that the LC itself is intact at 12 months and shows a 50 percent loss at 18 months of age. Given the massive deposition of β amyloid at this age in the APPswe/PS11d9 transgenic mouse model as well as the toxicity that has been attributed to the β amyloid oligomers and fibrils, it seems rather surprising that axonal and neuronal LC degeneration appears so late. In fact, this may point to the possibility that β amyloid itself is not the direct cause of the observed phenomenon but rather secondary mechanisms such as damage of synapses and axons through inflammatory mediators or excitotoxic stimuli. Furthermore, the tyrosine hydroxylase (TH) staining for LC axons and neurons could have been supported by additional immunostainings, e.g., against dopamine-β hydroxylase, in order to substantiate that induced neuronal loss, but not downregulation of TH itself is being observed.

    Since LC degeneration in AD is observed as early as in clinical pre-AD stages, e.g., in mild cognitive impairment (Grudzien et al., 2005), but is present only in late stages of the APPswe/PS11d9 transgenic mouse model, the authors’ conclusion, that this model recapitulates the progressive degeneration of monoaminergic nuclei in AD, seems a bit far-fetched and hypothetical. Of note, early LC degeneration is accompanied by neurofibrillary changes that have not been found in the present set of experiments. The study, however, nicely describes the fact that it may well be possible that LC neurons die by a retrograde mechanism initiated in its projection area. Given the neuroprotective and anti-inflammatory properties of NE, one can easily imagine how axonal loss, decreased NE, β amyloid deposition and neuroinflammation establish a vicious cycle that contributes to the progression of pathology and the clinical picture of AD.


    . Locus coeruleus neurofibrillary degeneration in aging, mild cognitive impairment and early Alzheimer's disease. Neurobiol Aging. 2007 Mar;28(3):327-35. PubMed.

    . Neuronal loss is greater in the locus coeruleus than nucleus basalis and substantia nigra in Alzheimer and Parkinson diseases. Arch Neurol. 2003 Mar;60(3):337-41. PubMed.

  2. I really appreciate the positive comments regarding our present work on degeneration of monoaminergic neurons in APP/PS1 Tg model. I agree that defining the mechanism(s) by which these neurons degenerate will be important. However, it is also highly important that we finally have a robust system to study this aspect of AD in vivo. I also want to expand on a few of the comments in the overview.
    First, while the "overview" focuses on LC and NA neurons, I want to stress that we observe degeneration of multiple monoaminergic systems. In particular, degeneration of 5-HT system in both Tg model and in AD cases is very significant since this system can modulate memory, emotion, and BDNF signaling. It is possible that the degeneration of 5-HT system, as with the degeneration of LC neurons, contributes to the feed-forward nature of AD pathology.
    Second, given the 40-50 percent loss of fibers and significant neuronal atrophy at 12 months of age, it is likely that neurodegeneration starts well prior to 12 months of age (considered rather late by Michael Heneka). Consistent with this view, the loss of monoaminergic fibers is clearly apparent in the eight-month-old Tg mice, indicating that neurodegeneration starts much earlier than 12 months of age.

  3. It was gratifying to read the paper by Liu et al. [1], at a time when our hypothesis on the possible role of brainstem neurons in Alzheimer disease (AD) pathobiology was posted on the SWAN Alzheimer Knowledge Base (hosted on the Alzheimer Research Forum website). [Editor's note: The SWAN database is under development.]

    In their paper, Liu et al. nicely describe the progressive neurodegeneration of neurons in the locus coeruleus (LC), in a mouse model of AD, and conclude that the LC neurons, which project into the brain regions affected by AD pathology, die by retrograde mechanisms. Abnormalities of LC neurons have been described in other mouse models of AD at early ages, several months before amyloid-β (Aβ) deposits are detected in the regions normally affected by plaque deposition (see [2], for example). As Dr. Heneka pointed out in his comment on the Liu et al. paper, degeneration of LC neurons is detected very early in the human disease too. Surprisingly, this significant degeneration occurs in the absence of Aβ pathology in the subcortical brain regions. Based on observations that degeneration appears to affect first the distal projections of the LC neurons, it was assumed that in AD, degeneration of LC neurons proceeds retrogradely, from the distal parts of the projections towards the cell bodies. The cause for this retrogradely advancing degeneration may not be the cortical Aβ deposits, since in most cases the deposits are largely absent at the time when the degeneration of LC neurons occurs.

    We have recently proposed that the LC neurons may actually play a role in triggering plaque formation by providing small seeds of aggregated Aβ that may accumulate—for a yet unknown reason—at the terminals of their projections, in the cortex [3]. Such accumulations may be small and scarce, and may easily go undetected in the mouse brain. In fact, Liu et al. note that they cannot exclude a possible presence of intracellular Aβ in the monoaminergic neurons, in the mouse model of AD they investigate. In another mouse model of AD (i.e., the PDAPP mouse), dystrophic TH-containing nerve terminals are found in locations that contain neuritic plaques [4]. We have proposed that, rather than being affected by the plaques, these terminals may actually participate in their initiation [3]. We note that our hypothesis does not contradict at all the idea that the LC neurons die by retrograde mechanisms, as suggested by Liu et al., and others. In cell culture, we detect Aβ accumulations at the terminals of CAD cells (a LC-derived neuronal cell line [5]) long before any signs of neurodegeneration can be detected [6]. Thus, it is possible that these neurons die—in the end—due to some pathological events at the terminals of their processes. Our model, which implicates LC neurons in facilitating the initiation of plaque pathology in AD, is in line with a study by Heneka et al. [7], which showed that degeneration of neurons in the LC can indeed promote AD pathology in the APP23 mouse (another model of AD). It appears that the neurons in the brainstem may play a more important role in the pathobiology of AD then previously thought, and the paper by Liu et al. provides further support to this less investigated aspect of AD.


    . Amyloid pathology is associated with progressive monoaminergic neurodegeneration in a transgenic mouse model of Alzheimer's disease. J Neurosci. 2008 Dec 17;28(51):13805-14. PubMed.

    . Abeta oligomers induce neuronal cell cycle events in Alzheimer's disease. J Neurosci. 2008 Oct 22;28(43):10786-93. PubMed.

    . Seeding neuritic plaques from the distance: a possible role for brainstem neurons in the development of Alzheimer's disease pathology. Neurodegener Dis. 2008;5(3-4):250-3. Epub 2008 Mar 6 PubMed.

    . The PDAPP mouse model of Alzheimer's disease: locus coeruleus neuronal shrinkage. J Comp Neurol. 2005 Nov 28;492(4):469-76. PubMed.

    . Characterization of a CNS cell line, CAD, in which morphological differentiation is initiated by serum deprivation. J Neurosci. 1997 Feb 15;17(4):1217-25. PubMed.

    . Neuritic deposits of amyloid-beta peptide in a subpopulation of central nervous system-derived neuronal cells. Mol Cell Biol. 2006 Jul;26(13):4982-97. PubMed.

    . Locus ceruleus degeneration promotes Alzheimer pathogenesis in amyloid precursor protein 23 transgenic mice. J Neurosci. 2006 Feb 1;26(5):1343-54. PubMed.

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Locus coeruleus neurofibrillary degeneration in aging, mild cognitive impairment and early Alzheimer's disease. Neurobiol Aging. 2007 Mar;28(3):327-35. PubMed.
  2. . Noradrenergic mechanisms in neurodegenerative diseases: a theory. Brain Res Brain Res Rev. 2004 Apr;45(1):38-78. PubMed.
  3. . Noradrenergic changes and memory loss in aged mice. Brain Res. 1985 Dec 16;359(1-2):292-9. PubMed.
  4. . A quantitative histological study of neuronal loss from the locus coeruleus of ageing mice. Neuropathol Appl Neurobiol. 1985 Jan-Feb;11(1):55-60. PubMed.
  5. . The PDAPP mouse model of Alzheimer's disease: locus coeruleus neuronal shrinkage. J Comp Neurol. 2005 Nov 28;492(4):469-76. PubMed.
  6. . Catecholaminergic neuronal loss in locus coeruleus of aged female dtg APP/PS1 mice. J Chem Neuroanat. 2007 Nov;34(3-4):102-7. PubMed.
  7. . Early locus coeruleus degeneration and olfactory dysfunctions in Tg2576 mice. Neurobiol Aging. 2009 Feb;30(2):272-83. PubMed.

Further Reading

Primary Papers

  1. . Amyloid pathology is associated with progressive monoaminergic neurodegeneration in a transgenic mouse model of Alzheimer's disease. J Neurosci. 2008 Dec 17;28(51):13805-14. PubMed.