Β-amyloid peptides stimulate astrocytes to unleash a slew of chemicals, yet it has been hard to discern the impact of these gliotransmitters on surrounding neurons. Mook-Jung’s team focused on one—adenosine triphosphate (ATP)—and explored its effects on Aβ-mediated neuronal dysfunction in a series of in-vitro experiments. They first demonstrated that synthetic, commercially available Aβ42 peptide triggers ATP release when added to cultures of primary mouse astrocytes or U373 human astroglioma cells. However, because it is hard to collect pure astrocyte-derived ATP in reasonable amounts, Mook-Jung said, subsequent experiments tested the effect of exogenous ATP on Aβ-treated neurons.

In these assays, first author Eun Sun Jung and colleagues found that ATP protected against Aβ toxicities—namely, loss of dendritic spines, decreased hippocampal long-term potentiation, and reduced levels of synaptic molecules, such as the NMDA receptor 2A subunit and PSD-95. Furthermore, the researchers determined that ATP exerts these protective effects through P2 type-purinergic receptor signaling, because a P2 antagonist blocked the ATP-induced drop in synaptic proteins. Purinergic receptors mediate cellular functions such as vascular responses and apoptosis. P2 receptors are preferentially activated by ATP.

Whereas other studies have examined how glial factors affect neurons, “this is one of the first to look at ATP alone,” Mook-Jung told ARF. And though the research does not prove that the neuronal benefits come from astrocyte-derived ATP, “it raises the possibility that a glial source of ATP—if one could harness it—could provide protection,” said Philip Haydon, Tufts University School of Medicine, Boston, Massachusetts, in an interview with ARF. “The key thing is they show a simple molecule could have these protective effects.”

But the situation is complex. Francesco Di Virgilio and colleagues at the University of Ferrara, Italy, reported that in AD mouse models, microglia, the other major non-neuronal cell in the brain, release ATP that induces inflammation (Sanz et al., 2009; see Di Virgilio comment below). Whether the findings apply in vivo remains to be seen. Future studies could examine whether selectively blocking ATP release from astrocytes worsens disease in an AD mouse model, Haydon suggested (see full comment below).

The second paper extends a chemokine theme. Prior research, led by coauthor Cristina Limatola, uncovered a mode of neuron-glia communication whereby neurons churn out fractalkine (CX3CL1) that signals to fractalkine receptor (CX3CR1) on microglia, which then make adenosine that signals back to neurons, protecting them from glutamate-induced death (Lauro et al., 2008; Lauro et al., 2010).

In the present study, first author Maria Rosito and colleagues turn their attention toward a related chemokine, CXCL16, and its receptor CXCR6. Both appear in astrocytes, microglia, and neurons. Like fractalkine, CXCL16 is a transmembrane chemokine that is released from the membrane by metalloproteinases such as ADAM10 and ADAM17. These same enzymes drive non-amyloidogenic processing of amyloid precursor protein (APP) by cutting smack dab within its amyloid-β fragment, thus preventing formation of AD-associated peptides.

Through a series of in-vitro experiments, the researchers worked out a molecular pathway for CXCL16-mediated neuron-glia crosstalk. They spiked excess glutamate into mixed cultures of neuronal and glial cells, triggering neuronal death in a manner relevant to a variety of situations including brain injuries, stroke, and neurodegenerative disease. However, when the same cells encountered soluble CXCL16 as well, it dose-dependently protected against cell death.

CXCL16 seems to protect by specifically signaling to receptors on astrocytes, since the benefits persisted even when microglia activation in hippocampal neuronal cultures was blocked pharmacologically. Intriguingly, these scientists also linked neuroprotection to an ATP relative. By removing all extracellular adenosine or specifically blocking A3R adenosine receptors, they blocked CXCL16 protection. The team also determined that CXCL16 spurs astrocytes to secrete the chemokine CCL2 (aka monocyte chemoattractant protein-1 or MIP-1), and that neutralizing CCL2 antibodies reduces the neuroprotection.

“They’ve clearly done a nice piece of mechanistic biology,” said Terrence Town of the University of California, Los Angeles. He wonders, though, how much of the neuroprotection will pan out in vivo. Town cited recent mouse studies by Richard Ransohoff and Bruce Lamb of the Cleveland Clinic Foundation, Ohio, showing that microglial fractalkine signaling has opposing effects on the two hallmark pathologies of AD, Aβ plaques and neurofibrillary tangles. Knocking out the microglial fractalkine receptor in APPPS1 (which develops rapid Aβ pathology) and A-R1.40 transgenic mice (which develop milder pathology) reduced plaque load, compared to AD mice with normal microglia. Yet loss of the fractalkine receptor intensified tau hyperphosphorylation in a transgenic mouse model of human tauopathy (see ARF related news story). This makes it hard to determine whether microglial fractalkine signaling is ultimately harmful or beneficial. In another study, lack of the microglial fractalkine receptor prevented neurodegeneration in triple transgenic AD mice, suggesting that fractalkine signaling is not neuroprotective (ARF related news story on Fuhrmann et al., 2010).

Ransohoff found the present study “very novel and interesting,” but agreed that in-vivo studies are needed “to see whether any of this work relates to cells in their native context,” he told ARF.—Esther Landhuis.

References:
Jung ES, An K, Hong HS, Kim JH, Mook-Jung I. Astrocyte-Originated ATP Protects Abeta1-42-Induced Impairment of Synaptic Plasticity. J Neurosci. 2012 Feb 29;32(9):3081-3087. Abstract

Rosito M, Deflorio C, Limatola C, Trettel F. CXCL16 Orchestrates Adenosine A3 Receptor and MCP-1/CCL2 Activity to Protect Neurons from Excitototix Cell Death in the CNS. J Neurosci. 2012 Feb 29;32(9):3154-3163. Abstract