Years ago, Tan and colleagues discovered that CD45 could inhibit microglial activation triggered by various stimuli, including Aβ (Tan et al., 2000), microbes, and crosslinking of the co-stimulatory protein CD40 (Tan et al., 2000). “We wondered whether CD45 could play a role in AD-like pathogenesis in a mouse model,” Tan told ARF.

To find out, first authors Yuyan Zhu and Huayan Hou crossed PSAPP mice (which overexpress human amyloid precursor protein and presenilin-1 with familial AD-linked mutations) with CD45-knockout mice, and analyzed their progeny at four or eight months of age. They assessed fibrillar Aβ burden by staining mouse brain sections with thioflavin S, and Aβ oligomers with Charles Glabe’s OC antibody. CD45-deficient PSAPP mice had more Aβ oligomers at four months, and greater fibrillar Aβ load at eight months, compared to PSAPP littermates. Western analysis also showed elevated levels of dimeric and oligomeric Aβ in brain extracts, both extracellular and intracellular, from CD45-/- PSAPP mice. By ELISA, the CD45-deficient group had more insoluble and soluble Aβ in the brain and, correspondingly, less soluble plasma Aβ. Together, the data suggest that PSAPP mice lacking CD45 have trouble clearing Aβ out of the brain and into the periphery.

To see whether these defects stemmed from faulty microglia, the authors stained brain sections with anti-Aβ antibodies and looked to see whether the amyloid clumps also lit up with microglial activation markers (Iba1, CD11b, or CD40), as this would indicate that the phagocytes were reaching their desired target. The immunohistochemical studies showed Iba1-positive microglia cozying up near plaques in PSAPP mice, but spread out more randomly, farther away from plaques, in CD45-deficient PSAPP animals. The latter group also produced more proinflammatory cytokines (TNF-α and IL-1β). The findings held up in vitro, where CD45-sufficient primary microglia gobbled Aβ peptides faster than did their CD45-deficient counterparts, also in confocal imaging studies, which revealed fluorescent Aβ within the cytoplasm of wild-type microglia but remaining on the surface of CD45-deficient cells. These data suggest that microglia without CD45 exist in a “runaway proinflammatory state,” that is ineffective at clearing plaques or controlling oligomeric Aβ buildup, the authors write.

These impairments eventually lead to greater problems in these mice. By immunohistochemistry and stereological analysis, neurons in the entorhinal cortex were injured and dying in eight-month-old CD45-deficient PSAPP mice. Mitochondria, too, were failing. Younger CD45-deficient transgenic mice and PSAPP mice with normal CD45 did not show these abnormalities.

“This paper shows that microglia play an important role in removal of soluble and insoluble Aβ, and that CD45 is a key molecule in these pathways,” commented Haruhiko Akiyama of Tokyo Institute of Psychiatry. Akiyama and colleagues reported years ago that microglia have diverse roles in the central nervous system—sometimes taking up soluble Aβ without inflammatory responses, other times phagocytosing insoluble, aggregated Aβ with inflammatory responses (Akiyama et al., 1999; see also Akiyama et al., 2000 for review).

Others followed with more evidence that microglia can be both Jekylls (El Khoury et al., 2007; Sinard et al., 2006) and Hydes (Fan et al., 2007) in AD. In fact, activated microglia exhibit a continuum of responses. They adopt different phenotypes, with beneficial phagocytosis and harmful inflammation on opposing ends of the spectrum (see Michelucci et al., 2009 and Perry et al., 2010 for review).

During aging, microglia seem to shift further toward the harmful phenotype. The transition accelerates in AD mouse models, where the cells initially help clear brain amyloid, but increasingly lose this capability and instead churn out more inflammatory cytokines (Hickman et al., 2008). Based on the new data, “I think CD45 plays a critical role in regulating the switch, helping to keep microglia in the phagocytic phenotype,” Tan said. Consistent with this idea, microglia express “stunningly high levels” of this phosphatase, Ben Barres noted, based on unpublished gene profiling data from his lab at Stanford University, Palo Alto, California.

At a glance, the present findings appear to fly in the face of a recent paper by Mathias Jucker’s group at the University of Tübingen in Germany. Jucker’s team surprised some researchers when they showed that removing microglia from the brains of APP transgenic mice had no effect at all on plaque formation or clearance (Grathwohl et al., 2009 and ARF related news story). However, Tan’s and Jucker’s studies “are asking fundamentally different questions,” noted Terrence Town, University of California, Los Angeles, a co-corresponding author with Tan on the current paper. Jucker ablated microglia altogether for a period of weeks using a suicide gene approach, whereas the present study “deleted a specific molecule, CD45, from birth in order to change the microglial phenotype in a specific way,” Town explained. Tan wonders whether Jucker and colleagues would have seen effects on plaque dynamics if they had depleted microglial for longer periods. “Furthermore, we do not know if the depletion had effects on oligomeric Aβ formation in their mice,” Tan noted.

All told, the new findings “add to the growing body of literature that illustrates the dynamic complexity of the role of neuroinflammation in AD and other neurodegenerative disorders,” wrote Donna Wilcock of the University of Kentucky, Lexington, in an e-mail to ARF (see full comment below). Working with Carol Colton at Duke University Medical Center in Durham, North Carolina, Wilcock and colleagues reported that deletion of the nitric oxide synthase 2 (NOS2) gene from APP transgenic mice resulted in neurodegeneration—a critical feature missing from most AD mouse models (Wilcock et al., 2008 and ARF related news story). NOS2 is an inducible enzyme that drives production of nitric oxide to help regulate immune responses. That study, together with the present data, suggests that AD transgenic mice “require a shift in the normal inflammatory response to progress disease pathology to cause neurodegeneration,” Wilcock wrote.

What about promoting CD45-mediated microglial clearance of β amyloid pharmacologically as a new therapeutic approach? Wilcock cautioned that manipulating the brain’s immune system could be “considerably difficult.... Perturbations to this system in either direction can clearly exacerbate pathology as opposed to reduce it.”

In a follow-up to the published data, Tan’s lab has begun screening for compounds that boost the CD45 pathway in microglia. The scientists are also testing whether brain-localized knockdown of CD45 in adult mice would give similar results as the current genetic knockout approach.—Esther Landhuis.

References:
Zhu Y, Hou H, Rezai-Zadeh K, Giunta B, Ruscin A, Gemma C, Jin J, Dragicevic N, Bradshaw P, Rasool S, Glabe CG, Ehrhart J, Bickford P, Mori T, Obregon D, Town T, and Tan J. CD45 deficiency drives amyloid-beta peptide oligomers and neuronal loss in AD mice. J. Neurosci. 26 Jan 2011;31(4):1355-1365. Abstract

Watts JC, Giles K, Grillo SK, Lemus A, DeArmond SJ, Prusiner SB. Bioluminescence imaging of Abeta deposition in bigenic mouse models of Alzheimer’s disease. PNAS Early Edition. 24 Jan 2011. Abstract