8 January 2009. Collagen, a major component of connective tissues and extracellular matrices, is not the first thing that comes to mind when one thinks of nerve cells. And yet, as reported in the January 4 Nature Neuroscience online, neurons in the hippocampus and cerebral cortex not only make collagen VI, but they might even use it to protect themselves from amyloid-β (Aβ) toxicity. “The finding was a complete surprise,” according to Lennart Mucke, Gladstone Institute of Neurological Disease, University of California, San Francisco, the principal author on the paper. The researchers came across the collagen connection while conducting microarray analysis for genes that are up- or downregulated in the dentate gyrus of transgenic mice expressing human amyloid precursor protein (J20 line). Though that microarray analysis is still ongoing, “the one gene that really stood out from all the others was collagen VI,” said Mucke, which prompted the researchers to publish that finding ahead of the total analysis.
As first author Jason Cheng and colleagues note, the role of various forms of collagen in peripheral organs is well established, but little is known about collagen in the brain. “The real novelty is twofold,” Mucke told ARF, “first that neurons produce collagen VI and second that they produce it to protect themselves from Aβ.” This may not be restricted to animal models, since the authors found elevated collagen VI in postmortem brain samples from AD patients. “This is a series of definitive and elegant experiments that shed some new light on Alzheimer’s pathogenesis,” said Terrence Town at Cedars-Sinai Medical Center, Los Angeles. “I think it is nice because it is one of those unanticipated findings,” he added. Town was not involved in the work.
Cheng and colleagues used RNA amplification analysis to confirm the microarray data. They found that transcripts of the collagen VI α1 subunit (Col6a1) are elevated about fourfold in dentate gyrus of hAPP mice, while the α1(VI) protein was elevated nearly 20-fold. Analysis of postmortem human samples revealed about a fourfold increase in Col6a1 mRNA compared to matched control samples. To see where the collagen is coming from, the researchers turned to cell cultures. They found that addition of Aβ42 oligomers elevated Col6a1 mRNA and α1(VI) protein in primary hippocampal and cortical neurons, whereas the amyloid peptide had no effect on collagen production in astrocytes.
What might lead to increased expression of collagen VI by neurons? The authors decided to focus on a Smad3 responsive element in the collagen gene, since Smad3 is activated by transforming growth factor β (TGFβ), a cytokine that is elevated in hAPP mouse brain and also in the AD brain (see ARF related news story). Cheng and colleagues found that blocking Smad3 prevented Aβ from boosting collagen VI expression. They also found that Aβ failed to elevate collagen VI when administered to neurons from TGFβ receptor knockout mice. Both experiments support the idea that TGFβ signaling, perhaps in response to Aβ toxicity, activates collagen VI expression. Town told ARF that this finding is in keeping with the protective role attributed to TGFβ in the brain. Town’s group has also shown that knocking down TGFβ signaling in peripheral immune cells spurs Aβ clearance (see ARF related news story). “The effects of TGFβ are exquisitely concentration- and cell-dependent, but clearly neurons have evolved to need TGFβ as a protective signal,” he said.
Mucke and colleagues’ findings also indicate that enhanced collagen VI production is protective. Primary neuronal cultures from collagen α1(VI)-deficient mice were more susceptible to Aβ toxicity, as judged by increased cell death, and this susceptibility could be rescued by treating the neurons with collagen VI fibrils. The protection seems linked to weakening Aβ oligomer toxicity, since Cheng and colleagues found that the oligomers were sequestered into large aggregates that co-localized with collagen VI in the extracellular space. “So the way we picture this is that the neurons basically coat themselves, pouring the collagen into the extracellular milieu so that it can sequester the oligomers and prevent them from attacking,” said Mucke.
It may not be so surprising that collagen (derived from the Greek word for glue) and Aβ42 peptides—which are famously sticky—might bind to each other. However, the collagen VI-Aβ interaction seems more specific than that, since the researchers found that collagen I did not protect against Aβ toxicity. Mucke said that they have not yet tested other types of collagen (and there are many). “I guess the most thought-provoking question is whether these effects will hold up in vivo,” said Town. “For the last 20 years or so it has been a thorn in the side of the AD community that we don’t get much frank neuronal loss in mouse models,” he said. “I wonder if a cross of these collagen VI-deficent animals with one of the AD mouse model lines would promote neuronal injury.” According to Mucke, such studies are underway in his laboratory. Alzforum will stick with the story.—Tom Fagan.
Cheng JS, Dubal DB, Kim DH, Legleiter J, Cheng IH, Yu G-Q, Tesseur I, Wyss-Coray T, Bonaldo P, Mucke L. Collagen VI protects neurons against Abeta toxicity. Nature Neuroscience. 2009 Jan 4. Abstract