Many people with Alzheimer’s wake up frequently at night, nap excessively during the day, and lose their sense of time. Besides sleep cycles, the circadian clock also controls astrocyte and microglia activation, and knocking out one of the core clock genes, Bmal1, hastens plaque formation and promotes neuroinflammation and synaptic damage. Now, researchers led by Erik Musiek and Brian Lananna at Washington University in St. Louis report that Bmal1 regulates the gene for YKL-40, a marker of Alzheimer’s disease. Further, deleting the YKL-40 gene, Chi3l1, alters glial inflammatory responses, promotes astrocyte and microglial phagocytosis of Aβ, and mitigates amyloid plaque formation—at least in mice. The findings, reported in the December 16 Science Translational Medicine, connect the circadian clock to glial responses in AD. Oddly, YKL-40 itself seems to have no circadian rhythm.

  • A variant in Chi3l1, the gene forYKL-40, associates with slower AD progression.
  • Deletion of Bmal1, a circadian clock gene, strongly suppresses Chi3l1 expression in mice.
  • Chi3l1 knockout improves glial phagocytosis and suppresses amyloidosis.

“We knocked out the core circadian clock to find dysregulated genes, and Chi3l1 came up,” said Musiek. “We happened to run across this important AD biomarker in this circadian screening.”

Chi3l1 encodes chitinase-3-like protein, aka YKL-40, a glycoprotein expressed primarily in astrocytes. Released into the cerebrospinal fluid in response to neuroinflammation, it is elevated in AD CSF (Craig-Schapiro et al., 2010). 

“In this elegant paper, Lananna et al. provide further evidence that YKL-40 plays a detrimental role in the pathogenesis of Alzheimer’s disease,” said Alberto Lleó, director of the Memory Unit of the Neurology Department at Hospital de la Santa Creu i Sant Pau, Barcelona, Spain. “Taken together, the study highlights the key role of astrocytes in the pathogenesis of AD and places YKL-40 as a potential therapeutic target for this condition.”

The story didn’t start with Bmal1 and the circadian clock in mice, though. The scientists first tried to figure out if, and how, YKL-40 influences how fast Alzheimer’s disease gets worse. They examined data from 778 people who had AD and were enrolled in longitudinal observational studies through the Knight Alzheimer’s Disease Research Center at Washington University. Of these, about 202 carried a genetic variant in the Chi3l1 locus, rs10399931, which has been shown to decrease CSF YKL-40 levels (Deming et al., 2016). They found that people who were homozygous for a thymine at this position had a 16 percent slower rate of AD progression, as measured by the CDR-sb, than those homozygous for cytosine (see image below).

Fast or Slow? In people with the TT genetic variant in the Chi3l1 locus rs10399931, Alzheimer’s disease progressed more slowly (pink) than in those with the CC variant (blue). [Courtesy of Erik Musiek.]

This data suggested that YKL-40 modulates AD progression. The next step was to examine how this might work. For this, the researchers turned to mice. They crossed APP/PS1-21 mice with wild-type or Chi3l1 knockout mice and analyzed brains when the offspring were 8 months old. At this age, APP/PS1-21 mice have usually have developed significant plaque pathology in the cortex and hippocampus (Radde et al., 2006). 

Staining with the antibody X34, which labels β-pleated sheet forms of insoluble Aβ, revealed that, without Chi3l1, there were 21 percent fewer fibrillar plaques, in a 17 percent smaller area in the hippocampus, though not in the cortex. However, staining with an antibody that recognizes all forms of Aβ revealed a much more pronounced reduction in plaque burden—55 percent in the hippocampus and 42 percent in the cortex. What gives? The APP/PS1-21 mice make a lot of dense fibrillar plaques, Musiek explained. It is possible that glial cells can’t remove those as well as diffuse plaques, causing this discrepancy between the two staining methods. Musiek thinks the YKL-40 knockout may have a stronger effect in mice that accumulate mostly diffuse plaques. “We’d like to go back and use a less aggressive model to see how they respond,” he said.

Are glia responsible for the reduction in plaques? Without YLK-40, hippocampal and cortical astrocytes in the APP/PS1-21 mice were less activated, as shown by fewer GFAP-positive cells around plaques. Primary astrocytes cultured from the knockouts gobbled up 13 percent more zymosan-coated latex beads than did control astrocytes from APP/PS1-21 mice, indicating the former were more phagocytic. They also swallowed 50 percent more Aβ42 peptides.

In microglia, the phagosome marker CD68 was up across all brain regions examined in YKL-40 knockout APP/PS1-21 mice relative to APP/PS1-21 controls, again indicating more voracious phagocytosis. Surprisingly, loss of Chi3l1 had a greater effect in microglia than in astrocytes, with phagocytosis of zymosan beads up by 2.5-fold and fluorescent Aβ by twofold.

“I think microglia are just better at Aβ phagocytosis; astrocytes can do it, but microglial are more efficient,” Musiek told Alzforum. “When we look at Chi3l1 KO mice, we don’t know if the increase in microglial CD68 around plaques is due to loss of Chi3l1 in the microglia themselves or in the astrocytes—which are then signaling to the microglia. Our transcriptomics data suggest that astrocytes are the primary cells that express Chi3l1, but perhaps microglia can express it under certain circumstances, and it can exert a cell-autonomous effect. We don’t know yet.”

Healthy Appetite. Microglial uptake of latex beads (left) and Aβ (right) increased when the cells were transfected with a small interfering RNA (siChi3l1) that knocked down YKL-40 production (green). A control scrambled siRNA (siScr) had no effect. Neither did siChi3l1 when cells were treated with the phagocytosis blocker cytochalasin D. [Courtesy of Erik Musiek.]

Finally, the scientists asked what molecular mechanisms might be regulating Chi3l1 expression. This is where the circadian clock comes in. Transcriptomic data revealed that Chi3l1 was downregulated in Bmal1 knockout mice brains by 89 percent, whereas deleting circadian clock proteins Per1 and Per2, which inhibit Bmal1-driven circadian regulation, increased Chi3l1 expression. This suggests that the circadian clock controls Chi3l1 transcription.

Strangely, perhaps, Chi3l1 mRNA did not rise and fall on a 24-hour cycle. The authors think this might be because the mRNA’s long half-life masks daily fluctuations in transcription.

Overall, this study suggests that disrupting the circadian clock by knocking out Bmal1 will suppress YKL-40 expression. This would drive up glial phagocytosis, reduce glial activation, and slow Aβ plaque formation and AD progression. Still, Musiek does not think that a broken clock necessarily signals better times ahead. “It really matters what cell type you’re talking about—microglia, astrocytes, or neurons,” he explained. Overall, a broken clock would likely be bad for the brain. “We have several papers coming out that look at this in more detail.”—Helen Santoro 


  1. YKL-40 has been measured in CSF in the context of neurodegenerative disease for several years and is typically considered an indicator of activated astrocytes. Here, through a comprehensive series of experiments, the Musiek lab and collaborators provide far more information on YKL-40 than we’ve had previously.

    Among the interesting findings is the fact that individuals who carry a genetic polymorphism resulting in lower CSF YKL-40 appear to have attenuated cognitive decline. Furthermore, the animal and cell experiments suggest that YKL-40 may impact plaque burden in the brain via modulation of microglial phagocytosis of amyloid. It would be interesting to see if the impact of YKL-40 on amyloid is also supported by human studies. For example, do carriers of the genetic variant of the CHI3L1 gene that causes lower CSF YKL-40 concentrations also show attenuated amyloid accumulation?

    Additionally, this study provides data showing that the astrocyte circadian clock regulates Chi3l1 transcription. It is too early to tell how this links abnormalities in human sleep and AD pathology, but the study does provide intriguing new information pointing toward a role for glia in the sleep/AD connection.  

Make a Comment

To make a comment you must login or register.


Research Models Citations

  1. APPPS1

Paper Citations

  1. . YKL-40: a novel prognostic fluid biomarker for preclinical Alzheimer's disease. Biol Psychiatry. 2010 Nov 15;68(10):903-12. PubMed.
  2. . Chitinase-3-like 1 protein (CHI3L1) locus influences cerebrospinal fluid levels of YKL-40. BMC Neurol. 2016 Nov 10;16(1):217. PubMed.
  3. . Abeta42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology. EMBO Rep. 2006 Sep;7(9):940-6. PubMed.

Further Reading

Primary Papers

  1. . Chi3l1/YKL-40 is controlled by the astrocyte circadian clock and regulates neuroinflammation and Alzheimer's disease pathogenesis. Sci Transl Med. 2020 Dec 16;12(574) PubMed.