In a variation on the theme of the brain as immune-privileged, the central nervous system may have access to tailor-made immune cells. This is according to two new studies, one led by Jonathan Kipnis and the other by Marco Colonna, both at Washington University in St. Louis, and published June 3 in Science. While other organs make do with immune cells circulating in the blood, the meninges taps its own private stock of monocytes and B cells in the nearby skull and vertebrae bone marrow. By way of tiny vascular corridors in the bone, these cells travel directly to the meninges, where they reside until provoked by injury or inflammatory insult, whereupon they infiltrate the brain or spinal cord. These myeloid cells never set foot in the blood and are distinct from their counterparts circulating there. The researchers propose that these directly sourced immune cells become CNS specialists, imbued with the intricate knowledge of how to protect the brain.

  • Myeloid cells migrate from the skull and vertebral marrow directly into the meninges.
  • From there, inflammation or injury triggers their entry into the brain or spinal cord.
  • B cells that enter the meninges may be “educated” by CNS antigens.

“This is an amazing contribution to the field,” said Oleg Butovsky of Brigham and Women’s Hospital in Boston. “It will ignite new avenues of investigation to understand how these cells contribute to brain health and disease.”

The findings may also help settle debates about the relative roles in neurodegenerative processes of resident microglia versus infiltrating monocytes from the blood, Butovsky said. Perhaps the monocytes hail from this third niche—the nearby bone marrow. More work is needed to define markers of the bone-marrow-derived cells and clarify how their functions differ from those of their resident and blood-borne counterparts, he added.

“I believe that this is potentially a huge breakthrough because it changes our view of the relatively impregnable blood-brain barrier. Who needs the blood if the marrow can provide leukocytes directly?” wrote Matthias Nahrendorf of Massachusetts General Hospital in Boston. Nahrendorf previously discovered the passageways connecting the skull marrow to the meninges in people and in mice. “The implications for inflammatory disorders (and their treatment) cannot be overstated,” he wrote.

Marrow to Meninges. The bone marrow (BM) is brimming with B cells (green). The image left captures one of them crossing a channel into the dura. The B cell expresses no IgM (right), indicating it is in an early developmental stage. [Courtesy of Brioschi et al., Science, 2021.]

The bone marrow houses hematopoietic stem cells, which give rise to all blood cells in the body, including immune cells. This rich cellular nursery births myeloid cells, including monocytes, macrophages, and neutrophils. Lymphocytes, including B and T cells, undergo the early stages of their development in the bone marrow before venturing off to peripheral organs—the spleen and thymus, respectively—to become fully mature. When Nahrendorf and colleagues spied tiny vascular corridors connecting the skull marrow to the meninges, they upended the long-held dogma that the brain, like any other organ in the body, receives its supply of immune cells from the blood (Aug 2018 news). In that study, Nahrendorf focused on the travel of neutrophils along these passageways.

Do other marrow-made cells travel directly into the brain without entering the blood? Both Science papers addressed this question, focusing on different types of immune cells. Kipnis’s group zeroed in on myeloid cells, while Colonna’s group focused on the lymphoid lineage. In Kipnis’s group, co-first authors Andrea Cugurra, Tornike Mamuladze, Justin Rustenhoven, and colleagues started by asking what proportion of myeloid cells in the CNS border regions—i.e., the meninges—hail from the blood. To address this question, the researchers used parabiosis to join the circulatory systems of a wild-type mouse with a UBC-GFP mouse, which expresses the green fluorescent protein under control of the ubiquitin C promoter. All hematopoietic cells in this mouse express GFP. Two months later, they measured the proportion of different cell types in the wild-type mouse that glowed green. While about 30 percent of myeloid cells in the blood and peripheral organs of the wild-type mouse came from the parabiont, less than 10 percent of myeloid cells in the cranial or spinal dura did. This suggested that meningeal myeloid cells predominantly derived from a source other than blood.

If not from the blood, where did these meningeal myeloid cells come from? Suspecting they came from a local marrow source, the researchers removed the top bulk of the skull, aka the calvarium, from a wild-type mouse and replaced it with the calvarium from a UBC-GFP mouse. One month later, the skull graft had fused with the adjacent wild-type skull, and Cugurra and colleagues spotted vascular channels connecting the skull transplant into the recipient’s dura. Strikingly, they also spotted donor monocytes, macrophages, and neutrophils from the transplanted calvarium within the meninges, suggesting the skull marrow can supply the meninges with myeloid cells.

Bone-marrow-transplant experiments further cinched the skull as the primary source of meningeal myeloid cells. The researchers used irradiation, coupled with strategically placed lead shields, to ablate bone marrow cells either in the brain or in the rest of the body. They then infused these irradiated mice with fresh bone marrow cells from a UBC-GFP mouse, as these only reconstitute ablated sections of marrow. In WT mice that had had bone marrow wiped out everywhere bar the skull, most of the cells in the meninges did not come from the donor. However, in mice that had had their bone marrow ablated only in the skull, the green donor cells reconstituted the lost marrow there, and many myeloid cells in the connected meninges glowed green. Together, this suggested that the skull marrow, as opposed to other sources, was the key contributor of myeloid cells to the meninges.

How do these myeloid cells respond in the event of CNS damage or injury? Cugurra and colleagues used three insults, in combination with parabiosis, to investigate this question. They found that whether roused by experimental autoimmune encephalomyelitis, spinal cord injury, or optic nerve crush, marrow-derived myeloid cells—distinguished from blood cells by their lack of GFP expression—infiltrated the CNS from the meninges. Monocytes from the blood, expressing the GFP of the parabiont, also flooded the CNS under these different conditions. In all, the findings suggest that when spurred by CNS damage or neuroinflammation, meningeal monocytes from the nearby marrow infiltrate the CNS. What’s more, the researchers observed the same vascular channels connecting the vertebrae marrow to the spinal meninges as they had observed in the skull.

Secret Passageways. In the spine of a mouse with EAE, immune cells pass from vertebral bone marrow into the spinal dura via small channels, and ultimately infiltrate the spinal cord. Left: H&E staining. Right: immunofluorescence. [Courtesy of Cugurra et al., Science, 2021.]

Did the monocytes from the marrow-meningeal route differ from those that infiltrated from the blood? Using single-cell RNA sequencing to compare gene-expression profiles of both cell types in the spinal cord, the researchers found that marrow-derived monocytes expressed a more immunosuppressive phenotype, suggestive of wound healing, while their blood-derived counterparts took on a more pro-inflammatory stance.

Kipnis said further studies will need to tease out the different roles of meningeal versus blood-derived monocytes. Perhaps the meningeal monocytes are “first responders” that signal for reinforcements from the blood, he suggested. Or maybe they temper infiltrating blood cells to minimize damage or promote healing. These and many other possibilities need investigation.

Butovsky thinks that under homeostatic conditions, meningeal monocytes may work remotely, sending signals into the brain parenchyma that influence the immune environment there. Another recent study led by Kipnis raised the idea that regulatory T cells in the meninges may impose an immunosuppressive regime on the CNS (May 2021 news). 

The findings could also help explain discrepancies about the origins of myeloid cells that infiltrate the central nervous system in response to neurodegeneration. Some scientists reported that in mouse models of amyloidosis, myeloid cells surrounding plaques came from outside of the brain, presumably from the blood, by virtue of the surface markers they expressed (Dec 2014 conference news; Feb 2015 conference news). Others, including Colonna, found no such infiltration when they used parabiosis to track cells that came from the blood (May 2016 news on Wang et al., 2016). “This marrow reservoir provides the missing link,” Kipnis said. “The cells did not come from the blood or the brain, but rather from the adjacent marrow.” Butovsky, who had described infiltrating monocytes in other models of neurodegeneration including ALS, agreed, noting that meningeal monocytes delivered directly from the CNS-adjacent marrow could be the infiltrators.

Skull B Cells: Bound for Brain Boarding School
Myeloid cells were not the only ones moving from the marrow into the meninges. B lymphocytes, which make up a third of immune cells in the meninges, also took the direct route. That was the upshot of Colonna’s study, which employed similar techniques (sans the calvaria transplants) as the study led by Kipnis. The authors discovered they were investigating similar questions about a year ago, when both papers were nearly finished, and decided to submit them together.

Using parabiosis and bone marrow chimera experiments, co-first authors Simone Brioschi, Wei-Le Wang, Vincent Peng, and colleagues pegged the bone marrow in the skull as the predominant source for B cells in the meninges. Strikingly, single-cell RNA sequencing determined that these meningeal cells spanned every stage of B cell development. They included pro-B and pre-B cells, in which segments of the immunoglobulin receptor are assembled, and the immature B cells, which express the IgM receptor on the cell surface. Typically, these developmental steps are thought to occur only in the bone marrow. Immature B cells then travel in the blood to finish their maturation in the spleen. In keeping with this, Brioschi and colleagues found that the blood and spleen were almost entirely devoid of the early B cells, and consisted of immature and, to a greater extent, mature B cells.

The presence of baby B cells in the meninges suggests a unique developmental niche, in which budding B cells are “educated” by CNS antigens, Colonna said. This could potentially weed out B cells that might attack antigens specific to the brain, the researchers proposed. To investigate this, the researchers used a transgenic mouse that carries an immunoglobulin heavy chain specific for myelin oligodendrocyte glycoprotein (MOG), a CNS antigen. In these mice, far more MOG-specific B cells were detected in the tibial bone marrow than in the dural meninges, suggesting the latter weeded out CNS-reactive B cells.

The details of how nascent B cells are recruited and raised in the meninges are still unclear, although the researchers did find that fibroblasts in the dura express chemokines that attract and support their development.

Finally, Brioschi and colleagues found that the origin of B cells mingling in the meninges changed with age. Using single-cell RNA sequencing and B cell immunogloublin sequencing, the researchers compared the characteristics of dural B cells in 2- to 3-month-old mice versus 20- to 25-month-old mice. Not only were there more B cells in the older dura, an increasing number of them were clones of blood B cells. This suggested that as the mice aged, more blood B cells infiltrated the meninges. The researchers also detected a transcriptional cluster of B cells that was almost exclusively found in aged mice. Dubbed “age-associated B cells,” aka ABCs, these mature B cells expressed higher levels of IgM and, interestingly, ApoE, than did other mature B cells. The IgMs expressed by the B cells in old mice were also riddled with somatic mutations—a telltale sign that they had undergone a process called affinity maturation, which only occurs in response to a cognate antigen. Therefore, the ABCs were antigen experienced. The researchers also detected an uptick in plasma cells—i.e., terminally differentiated B cells actively pumping out antibodies—in the dura of aged compared with young mice.

“That the bone marrow supplies the dura with B cells and monocytes is incredible,” noted Arya Biragyn of the National Institutes of Health in Bethesda, Maryland. Biragyn was intrigued by what these meningeal B cells might do under neuroinflammatory conditions such as aging and neurodegenerative disease. “If brain inflammation occurs, will these cells infiltrate the parenchyma?” he wondered. Colonna’s study did not investigate B cells in the parenchyma. However, Biragyn recently reported that B cells exacerbated amyloidosis and memory loss in mice (Kim et al., 2021). Biragyn also pointed out that immune cells, including B cells, do not exist in isolation, but rather extensively cross-regulate each other. Therefore, it will be crucial to understand how immune cells from different pools interact in the context of aging and disease, he said.

To Malú Tansey of the University of Florida, the existence of reservoirs of immune cells trained to serve the needs of neurons makes perfect sense. “This is a strategy to keep the brain not in an immune-privileged state, but rather in an immune-specialized state,” she said. In cases where uneducated, peripheral cells do infiltrate the brain—a situation Tansey equated to a bull in a China shop—perhaps the skull marrow-derived immune cells run interference to limit damage, she speculated.

Going forward, it will be important to understand how aging impacts the function of immune cells in these CNS-adjacent marrow reservoirs, Tansey said. She is convinced that the aging of the immune system is a key contributor to neurodegenerative disease, noting that the preponderance of AD risk variants are expressed in immune cells. While that has spurred much research on the role of microglia in AD, perhaps the variants influence AD risk by altering the function of cells in these CNS-adjacent marrow reservoirs as well.—Jessica Shugart


  1. Both manuscripts extend the concept that the bone marrow in the skull is potentially a special place. Small channels crossing the inner shell of the skull connect bone marrow and the meninges. The data from both labs show that these channels are routes of leukocyte migration from the site of production (marrow) to the CNS. The exciting insight is that the channels may serve as a shortcut, or back door, to the CNS, and work from the Colonna and Kipnis labs implicates the channels’ importance in several inflammatory CNS conditions.

    I believe that this is potentially a huge breakthrough because it changes our view of the relatively impregnable blood-brain barrier. Who needs the blood if the marrow can provide leukocytes directly? Since we had seen that the channels exist in humans, the implications for inflammatory disorders (and their treatment) cannot be overstated.

  2. The finding that a pool of B cells and myeloid cells in the meninges derives from local skull and vertebral bone marrow, as opposed to the blood, raises the possibility that specific pools of these immune cells are uniquely primed to target CNS diseases. That these cells can enter the parenchyma during different neuroinflammatory diseases must alter the way we understand the pathophysiology of many different neurological diseases, as well the methods to treat them.

    This work raises many important questions that are critical to understand neuroinflammatory diseases. Are there functional differences between local and blood-derived immune cells? Can we target these differences to modulate the immune response to treat diseases such as multiple sclerosis and Alzheimer’s disease? How do these local skull and vertebral reservoirs change during aging? Does every tissue and organ in the body contain a local reservoir of immune cells primed for organ specific function?

  3. Together, these papers point to a previously unknown source of immune cells that are poised to augment the brain’s response to injury, inflammation, and disease.

    While skull and vertebrae are known sites of blood formation, it was not known that these sites seed the adjacent, protective covering of the brain—the meninges—with immune cells. Both papers used clever and orthogonal methods to show the origin of the meningeal cells; they likely reach the meninges via special bone channels. These channels, which seem to share some similarities to diploic veins in our skulls, suggest that the idea of local bone marrow sources for infiltration into the CNS can be extrapolated from mice to us! The skull- and vertebra-derived pools of local monocytes, neutrophils, and B cells seem to be at the ready to act when things go awry.

    These findings may raise more questions than they answer. Are infiltrating immune cells in the brain always from these sources? Are they genetically, or perhaps more importantly, functionally distinct from cells from more distant bone marrow sources? And how is all of this changed in various conditions, especially in times of craniospinal radiation for malignancy, when these cells are presumably wiped out?

    These findings could explain discrepancies between the different functions found for microglia and infiltrating cells. Depending on the experimental paradigm, one might imagine different results if one approach favors more distant sources of cells, whilst another more local sources.

    Beyond the brain and spinal cord, the work raises the possibility that there is no such thing as one bone marrow source for immune cells; instead, tissues are populated by their own local bone marrow.

  4. These two papers are each very interesting and raise many questions. I would just add that bone marrow cells called megakaryocytes were found clogging brain capillaries in an autopsy study of some cases of COVID-19 (Nauen et al., 2021).


    . Assessing Brain Capillaries in Coronavirus Disease 2019. JAMA Neurol. 2021 Jun 1;78(6):760-762. PubMed.

  5. These are two great papers that, in my opinion, challenge and change our current view on immune privilege and immune responses of the CNS. Both studies use comparable state-of-the-art methods and approaches to convincingly show the presence of monocyte and neutrophil pools in the skulls and vertebral bone marrow that closely associate with the CNS and serve as a reservoir for meningeal and CNS myeloid cells (Cugurra et al.) and provide a lymphopoietic niche in the meninges for B cells (Brioschi et al.).

    Besides their obvious importance for the pathogenesis and treatment of CNS disorders, it is intriguing to understand more deeply the relationship between peripheral and central pools of these cell types in other disorders, for example hematological diseases and cancer. Are these pools completely differentially regulated, or is there some synchrony? And could they be interchangeable in function if need be? In other words, can peripheral signals mobilize these closely CNS-associated cells to take an active part in non-CNS disorders?

    Another question that is also very intriguing to me is whether cues in the skull and vertebrae, compared to, say, the iliac bone, provide different maturation environments to each other, or if they interact.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Tiny Passageways Connect Skull Bone Marrow to the Brain
  2. As Mice Age, T Cells Traipse Around Their Meninges. Mayhem Ensues
  3. TREM2 Data Surprise at SfN Annual Meeting
  4. Microglia in Disease: Innocent Bystanders, or Agents of Destruction?
  5. Barrier Function: TREM2 Helps Microglia to Compact Amyloid Plaques

Paper Citations

  1. . TREM2-mediated early microglial response limits diffusion and toxicity of amyloid plaques. J Exp Med. 2016 May 2;213(5):667-75. Epub 2016 Apr 18 PubMed.
  2. . Therapeutic B-cell depletion reverses progression of Alzheimer's disease. Nat Commun. 2021 Apr 12;12(1):2185. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Skull and vertebral bone marrow are myeloid cell reservoirs for the meninges and CNS parenchyma. Science. 2021 Jun 3; PubMed.
  2. . Heterogeneity of meningeal B cells reveals a lymphopoietic niche at the CNS borders. Science. 2021 Jun 3; PubMed.