The yin and yang of transforming growth factor β (TGF-β) is nowhere more apparent than in Alzheimer disease. It’s been known for over a decade that expression of the cytokine is increased in AD brain (Flanders et al., 1995). Tony Wyss-Coray and colleagues showed that TGF-β overproduction by astrocytes is neuroprotective, and that TGF-β1 knockouts display increased neuronal death (see ARF related news story). That group also showed that TGF-β promotes clearance of Aβ by microglia and protects mice from Aβ deposition (Wyss-Coray et al., 2001). Conversely, the same researchers showed that high TGF-β levels in mice cause cerebrovascular amyloid angiopathy (Buckwalter et al., 2002), and Denis Vivien and colleagues tied high TGF-β to elevated APP expression and Aβ production (see ARF related news story). One difficulty in sorting out the risk/benefit ratio for TGF-β is that many cells in the nervous system and beyond respond to the cytokine, making it hard to know which players mediate various effects.

Two new studies, one just published and one presented for the first time at the Society for Neuroscience annual meeting in Atlanta, Georgia, last month, attempt to get at this question, not by knocking out the production of TGF-β, but instead by blocking its signaling in selected groups of target cells. Both take a similar approach of engineering transgenic mice to carry a dominant negative form of the TGF-β type II receptor. But in one study, the dominant negative was expressed in neurons, while in the other it was in peripheral macrophages, and that change in location turns out to make a world of difference.

In the first case, work from Tony Wyss-Coray and colleagues at Stanford University Medical School, California, shows that impeding neuronal TGF-β signaling promotes neurodegeneration and can exacerbate amyloid pathology in AD mice. That study, published in the November issue of the Journal of Clinical Investigation, contrasts with unpublished data presented at the SfN meeting that showed getting rid of TGF-β signaling in peripheral macrophages has an anti-amyloid effect. There, Terrence Town of Yale Medical School, New Haven, Connecticut, showed that damping down TGF-β action seems to unleash the Aβ-gobbling tendencies of macrophages. When taken together, the results from both groups suggest that boosting TGF-β centrally, or inhibiting it peripherally, might both be viable approaches to reducing amyloid deposition.

The study of neuronal TGF-β starts with first author Ina Tesseur and colleagues showing that the levels of the high-affinity TGF-β type II receptor subunit protein are decreased to about half of normal levels in brain tissue from people with AD. The decline is apparent early—in AD patients with mild dementia. It appears specific for AD, and was not observed in postmortem brain from patients with other neurodegenerative disorders, including Parkinson disease or progressive frontotemporal dementia. Immunostaining of brain sections revealed the receptor was mainly present on neurons, with far fewer in astrocytes. These results, along with the same group’s previous observations that TGF-β is increased in AD brain, suggest that something is amiss with this cytokine pathway in this disease.

To find the possible consequences of decreased TGF-β signaling in neurons, the researchers produced transgenic mice that expressed the TGF-β receptor type II with a deletion of the cytoplasmic serine/threonine kinase domain. This protein acts as a dominant negative inhibitor of endogenous receptor signaling, and its expression was restricted to cortical and hippocampal neurons with the neuron-specific prion promoter and tetracycline-sensitive expression.

Consistent with TGF-β’s neurotrophic actions, cultured neurons from the transgenic mice showed increased cell death, and the animals demonstrated age-dependent neuron loss in vivo, as indicated by significant decreases in MAP2, synaptophysin, and NeuN immunoreactivity in very old mice (34 months of age) compared to control animals. GFAP was elevated, indicating ongoing astrocytosis.

When they looked for AD pathology in the transgenics, the researchers found a trend toward elevation of Aβ, but the increase did not reach significance, and no amyloid deposits were observed. When they crossed the dominant negative mice with J20 mice, which express human APP with the Swedish and Indiana mutations, however, they did get enhanced Aβ accumulation. Amyloid deposition and thioflavin S deposits were increased about twofold in old crossed mice (20 months), but not in younger ones (14 months or younger). A decrease in MAP2 immunoreactivity indicated neuronal loss as well, which was not seen in TGF-βRII transgenics at the same age.

While the researchers found no evidence in the mice for enhanced APP expression or processing, or decreased Aβ degradation, studies in cultured cells pointed to an increase in Aβ production when TGF-β signaling decreased. In rat neuroblastoma cells expressing human APP, they saw an increase in Aβ secretion in response to either the dominant negative receptor, or a kinase inhibitor of TGF-β signaling. Inhibiting TGF-β action resulted in beading and retraction of neurites and rounding and detachment of cells, independent of whether the cells also expressed APP or not. These results suggest that TGF-β is necessary to maintain the health and integrity of cells, and that its reduction may promote APP processing and Aβ production.

“Collectively, these studies suggest that defects in TGF-β signaling may contribute to AD pathogenesis by promoting neurodegeneration and initiating a feedback loop in which the degenerating cell produces more Aβ, thereby enhancing amyloid deposition,” write Pritam Das and Todd Golde, Mayo Clinic, Jacksonville, Florida, in an accompanying piece. More work will be needed to find out why the TGF-β receptor is downregulated in AD brain (perhaps in response to upregulation of TGF-β itself?), how this alters Aβ production, and whether changes in processing explain the increased amyloid deposition observed in crossed mice.

Separate from its neurotrophic actions, TGF-β plays an important role in the immune response, and Town’s SfN talk focused on the interplay between the worlds of the CNS and peripheral innate immune system. Town, a researcher in Richard Flavell’s lab at Yale, hooked up the TGF-βRII dominant negative to the CD11c promoter to create mice lacking TGF-β signaling in cells of the innate immune system, including macrophages and dendritic cells (but not microglia). When he crossed these mice with Tg2576 AD mice, he saw their cortical and hippocampal plaque load and Aβ peptide content reduced by 60-80 percent at 18 months. The mice also showed a dramatic 60-90 percent reduction in cerebral vascular amyloid, and reduced astrogliosis. These changes were associated with behavioral differences, as the crossed mice showed less hyperactivity in a novel environment and in the Y maze compared to the parental Tg2576 strain.

What was the basis of lower amyloid in the mice? In brain, staining for the macrophage marker CD45 nearly doubled, and Town observed CD45- and CD11b-positive infiltrating macrophages. The macrophages contained Aβ in vivo, and they showed a higher capacity to phagocytose Aβ in vitro than control cells. From these experiments, Town concluded that getting rid of TGF-βRII signaling removed a negative regulatory influence on the macrophages, allowing them to become activated to clear Aβ. Importantly, the activation was not associated with proinflammatory cytokine production either in vivo or in vitro. The results that shutting down TGF-β signaling in the periphery promotes plaque clearance contrasts with some of Wyss-Coray’s earlier work suggesting that TGF-β production in the brain promotes plaque removal by stimulating microglial uptake (Wyss-Coray et al., 2001). Together, the studies support the idea that macrophage and microglial activation are not just two phenotypes, and show that different approaches to modulating TGF-β action may have very different outcomes. Town concluded by saying that blocking TGF-β immune signaling in the periphery may be a promising therapeutic avenue for AD.—Pat McCaffrey


  1. The findings from Tony Wyss-Coray’s group suggest that that downregulation of neuronal TGF-βRII not only seems to “sensitize” neurons to injury in the context of Alzheimer pathology, but it also seems to amplify Alzheimer pathology by promoting amyloidosis. In particular, the authors found elevated levels of neocortical and hippocampal Aβ1-42 in older AD mice crossed with mice expressing the dominant negative receptor, and in vitro, genetic or pharmacologic knockdown of TGF-β signaling resulted in both neuronal loss and enhanced release of Aβ.

    How do these findings dovetail with our work presented at the SfN meeting? Tony’s group has directly blocked TGF-β signaling in neurons by inducibly overexpressing a dominant negative TGF-βRII construct (lacking the signaling domain—very similar to what I am using), and they then crossed these mice to PDAPP AD mice to directly address the question of whether depriving neurons of the neurotrophic TGF-β signal promotes neuronal injury in the context of AD pathology. In my system I’ve driven the dominant negative TGF-βRII by the CD11c promoter (expressed by most cells of the innate immune system, including macrophages and dendritic cells) and crossed these mice to the Tg2576 AD mice to address the question of whether inhibiting TGF-β signaling on innate immune cells alters AD-like pathology. As Tony points out, TGF-β is a pleiotropic cytokine, which can have differing effects depending on concentration and cell type receiving the signal. He shows that blocking this signal on neurons is deleterious in the AD paradigm because it deprives these cells of an important neurotrophic signal. In my system, blocking TGF-β signaling in innate immune cells “revs these cells up” and causes them to clear Aβ/β-amyloid both in vivo and in vitro. Specifically, it seems that it is the peripheral macrophages that are the effectors in this response, as blockade of TGF-β in these cells causes them to enter into the brain and likely clear Aβ/β-amyloid. This response is not accompanied by proinflammatory cytokine production in vivo or in vitro, so it seems that this form of macrophage “activation” is beneficial because it promotes Aβ/β-amyloid phagocytosis in the absence of an overt inflammatory response. I guess the key message is that if we are going to target TGF-β as a therapeutic modality for AD, we have to be careful about where in the body we modulate this pathway; we’d ideally like to promote this pathway in neurons and block it in peripheral macrophages.

Make a Comment

To make a comment you must login or register.


News Citations

  1. More on TGF-β—Can It Protect against AD?
  2. Astrocytes—Part of the Solution or Part of the Problem?

Paper Citations

  1. . Altered expression of transforming growth factor-beta in Alzheimer's disease. Neurology. 1995 Aug;45(8):1561-9. PubMed.
  2. . TGF-beta1 promotes microglial amyloid-beta clearance and reduces plaque burden in transgenic mice. Nat Med. 2001 May;7(5):612-8. PubMed.
  3. . Molecular and functional dissection of TGF-beta1-induced cerebrovascular abnormalities in transgenic mice. Ann N Y Acad Sci. 2002 Nov;977:87-95. PubMed.

Further Reading

Primary Papers

  1. . Deficiency in neuronal TGF-beta signaling promotes neurodegeneration and Alzheimer's pathology. J Clin Invest. 2006 Nov;116(11):3060-9. PubMed.
  2. . Dysfunction of TGF-beta signaling in Alzheimer's disease. J Clin Invest. 2006 Nov;116(11):2855-7. PubMed.