Knowledge of the brain’s immune system has come a long way since scientists thought the brain was completely cut off from the body’s defenses. Now, using a technique called mass cytometry, researchers have characterized the diverse populations of immune cells in the mouse brain—and there are a lot of them. Scientists led by Asya Rolls, Israel Institute of Technology, Haifa, report in the July 24 Nature Neuroscience that microglia keep company with B cells, T cells, and natural killer cells. Some were expected in the normal brain, but others were a complete surprise. This study catalogues which immune cells inhabit the healthy mouse brain, where they reside, and where they originated. It provides the basis for similar future analyses in models of stress, infection, or neurodegeneration, to see how immune cells factor in to those conditions.

“I hope that using this approach, we can generate a map of the various immune cells in the brain, and see how their composition changes with age and different stages of disease,” Rolls wrote to Alzforum.

“This is one of the first characterizations of brain myeloid compartment cells by mass cytometry,” wrote Mariko Bennett, Stanford University, California, to Alzforum (see full comment below). “The value of this manuscript is not only as a standalone paper, but as the baseline by which future mass cytometry-based analyses of brain pathologies—Alzheimer’s disease included—can be compared.”

“Single-cell mass cytometry is a fantastic technology that is revealing aspects of our biology about which we have no idea,” said Burkhard Becher, University of Zurich. Previous data suggested the immune system has access to the brain, albeit limited, and this study delivers unbiased confirmation in high resolution, he said. 

Brain Immune Cells: Resident myeloid cells (microglia) make up the bulk of immune cells in the brain, but it is also inhabited by dendritic cells, memory T cells, B, and natural killer cells. [Courtesy of Korin et al., Nature.]

In mass cytometry, single cells are dissociated and their surface proteins labeled with antibodies. Mass cytometry can accommodate dozens of antibodies, each of which is labeled with a unique metal ion. After cells are labeled via their respective surface markers, they are fed in single file through a mass cytometry machine, also known as the cytometry by time of light (CyTOF) machine, made by South San Francisco-based Fluidigm. There the metal ions that were bound to the membrane are analyzed, giving each cell a unique fingerprint that represents the surface markers it expressed. This enables scientists to identify subsets of cells by the proteins showcased on their membranes.

For this study, first author Ben Korin and colleagues removed the brains of healthy mice, including the choroid plexus and meninges surrounding the brains. After dissociating the brains into single cells, the researchers labeled the cells with 44 antibodies that recognize markers typically found on cells of the innate and adaptive immune systems. They then carried out mass cytometry and categorized the immune cells in the brain preparation.

Example of the SPADE diagrams used in this mass cytometry study to represent the hierarchical organization of cells. The size of a node correlates with the number of cells; the color gradient represents expression level. This type of diagram can be generated for each molecular marker. [Courtesy of Korin et al., Nature.]

The researchers first distinguished between resident cells—the microglia—and cells that had infiltrated from the periphery. For that they used CD45, one of the few surface proteins known to be expressed more highly by infiltrating cells than those from within the brain. Among the infiltrating cells, the group found myeloid and lymphoid cells, dendritic cells, monocytes, macrophages, CD4 and CD8 T cells, B cells, and natural killer cells (see image at left). CD4 T cells previously had been found to alter normal brain activity, but few of the others have been found in brain except in pathological situations. Most infiltrating cells were hiding out in the meninges and the choroid plexus. Removing these two structures cut the percentage of infiltrating cells in the brain from 17 to less than 2.

Korin and colleagues then examined differences between immune populations in the brain and the blood. It turned out that the two are quite distinct. For instance, natural killer cells in the brain more often expressed CD62L, a marker of maturity and strong cytotoxicity. CD8 T cells there expressed CD86, a marker of unknown function. On the other hand, B cells in the blood expressed more IgM.

Single-cell mass cytometry also proved useful in parsing the heterogeneity among microglia that has long puzzled scientists. Some expressed more chemokine receptors, pointing to their sensing and trafficking function; another set expressed high levels of MHC-II, hinting at antigen presentation. Still others expressed the TCR-β immunoreceptor—not previously found on microglia—which rose with inflammation. The data imply that microglia are functionally diverse.

Lastly, the researchers found that the cell surface glycoprotein CD44 distinguishes resident from infiltrating myeloid cells. Almost all infiltrating cells expressed CD44, while almost no resident ones did. This adds to a short list of markers that distinguish the two populations, the authors wrote.

How does single-cell mass cytometry compare to single-cell genomics analysis? “With mass cytometry, you look at which proteins are actually translated,” said Bart Eggen, University of Groningen, The Netherlands. This complements gene expression profiling but gets a step closer to cellular function, he added. While this study confirms that there are many more cell types of the immune system in the CNS under normal conditions than just microglia, its novelty is that the authors identify the cell types and classify them by their cell surface markers and regional locations in the brain.

It is unclear yet what those cells are doing under physiological conditions or to what extent they participate in disease, Eggen said. However, he and Becher agreed that researchers can now do similar analyses on disease models and see how immune cells are different and how those changes affect resident cells of the CNS. “These are resources that ultimately will guide all of our future efforts to understand diseases,” said Becher.—Gwyneth Dickey Zakaib


  1. This is one of the first characterizations of brain myeloid compartment cells by CYTOF—which allows for multidimensional analysis on multiple protein markers on a single cell.

    The value of this manuscript as a resource is not only as a standalone paper, but also as the baseline against which to compare future CYTOF-based analyses of brain pathology, AD included.

    A paramount challenge of most cytometry is that it is only as valid as the antibodies used to detect the proteins of interest—and there are so many bad antibodies out there that love to stick to brain myeloid cells! I was delighted to see the thoughtful consideration of this limitation by the authors’ inclusion of corresponding RNAseq data of their major findings.

    I can't wait to dig into their data set even more! Their validation of CD44 as a marker of all "infiltrating" or "surveying" immune cells in the CNS is exciting—especially in considering what this means ... how does CD44 contribute to brain immune surveillance at baseline and in disease?

Make a Comment

To make a comment you must login or register.


No Available References

Further Reading


  1. . Microglial brain region-dependent diversity and selective regional sensitivities to aging. Nat Neurosci. 2016 Mar;19(3):504-16. Epub 2016 Jan 18 PubMed.
  2. . Mass Cytometry: Single Cells, Many Features. Cell. 2016 May 5;165(4):780-91. PubMed.

Primary Papers

  1. . High-dimensional, single-cell characterization of the brain’s immune compartment. Nat Neuro 24 July 2017