Bypassing the blood-brain barrier

Brain function requires tight regulation of the cerebral microenvironment, an outcome achieved through specialized brain barriers. The presence of these barriers and the observation that skin grafts transplanted into the brain are poorly rejected (1) helped establish the dogma that the central nervous system (CNS) is immune privileged; that is, immune responses are not elicited against antigens originating from the CNS. However, recent studies have identified entry and exit points for immune cells, cytokines, and brain-derived antigens or waste molecules through distinct anatomical routes in the meninges (a triple-layered membrane that surrounds the brain parenchyma). These meningeal “backdoors” might allow for blood-brain barrier (BBB) bypass, providing an interface for peripheral immune system interaction with the CNS, and could play an important role in neuroinflammation and neurodegeneration.

The BBB, a highly selective semipermeable structure segregating peripheral blood from the brain, is characterized by tight junctions between endothelial cells that line the vessels and low expression of leukocyte adhesion molecules (2). Thus, although the brain contains resident macrophages (microglia), peripheral immune cells have limited access to the brain parenchyma. A series of classical studies revealed that if an animal is peripherally immunized with an allograft (from an unmatched donor) prior to allograft transplantation into the brain, rejection occurs rapidly (1). This suggested that T cells primed in the periphery can gain access to the brain. Thus, although adaptive immune responses in the brain are weak, T cells are not blind to brain-derived antigens.

More recently, the meninges have been recognized as a critical site where immune surveillance of the CNS takes place. The meninges are a triple-layer structure enveloping the brain. Within the two inner layers, called the pia and arachnoid mater (collectively, leptomeninges), flows cerebral spinal fluid (CSF). The outermost dura layer is formed by thick connective tissue, which, unlike the brain and leptomeninges, contains vasculature largely lacking tight junctions between endothelial cells (3), allowing extensive leukocyte surveillance (4). CSF is produced by specialized epithelial cells of the choroid plexus, fills the brain ventricles, and flows toward the subarachnoid space within the leptomeninges. From the subarachnoid space, CSF enters the Virchow-Robin spaces surrounding large penetrating arteries and flows across astrocyte (a type of stromal cell) projections into the brain parenchyma, where it mixes with the interstitial fluid (ISF). From there, the macromolecule-containing ISF-CSF returns, through perivenular spaces, into the CSF circulation. This “glymphatic” system allows CSF transit through the brain, collecting waste and brain-derived antigens (5). Additionally, it provides a route for meningeal immune cell–derived factors, including cytokines, to enter the brain, where they can modulate CNS function (6).

CSF volumes must be tightly regulated to avoid hydrocephalus and associated neurological sequelae. The prevailing concept was that CSF drains into dural sinuses, vascular structures carrying venous blood from the brain, through arachnoid granulations (observed only in large animals, including humans), prior to its absorption into the jugular vein. However, whether the dural sinuses drain CSF is unclear and, regardless, such drainage would not result in efficient presentation of CNS-derived antigens to the peripheral immune system. For appropriate immune surveillance and antigen presentation, lymphatic drainage to lymph nodes is required. The brain parenchyma, unlike almost all peripheral tissues, contains no lymphatic network to facilitate waste removal. For decades, it was believed that macromolecules and immune cells are drained along the olfactory nerves, across the cribriform plate, into the nasal mucosa, and thence into draining lymph nodes. If this were the case, however, then molecules inhaled through the nose (including pathogens) would be draining into the same lymph nodes as those receiving brain-derived antigens. Such a scenario is unlikely, especially given recent work demonstrating that lymph drainage from the gut is partitioned into separate lymph nodes, endowing each with a specific function (inflammatory responses or antigen tolerance) (7), highlighting the intricate precision of antigen sampling and immune responses within tissues. Why would immune surveillance of the brain, a master organ, be shared with airborne pathogens? It seems more likely that there would be a dedicated system, and indeed a functional CNS-draining lymphatic network housed within the meninges has been characterized (3).

Once the CSF exits the brain parenchyma, it is carried across the arachnoid mater into dural spaces, where it is drained through meningeal lymphatic vessels into the deep cervical lymph nodes (dCLNs) situated within the neck, allowing interactions between brain-derived antigens and peripheral T cells. This provides a system of fluid flow that can relay messages from the periphery to the parenchyma and conversely drain parenchymal antigens to the periphery.

Thus, a unified theory of CNS cytokine passage, antigen drainage, and immune surveillance, that is, meningeal bypass, can be envisioned. A cytokine traveling within the peripheral blood is unlikely to cross the BBB, owing to the presence of endothelial cell tight junctions. The cytokine may, however, cross the fenestrated dural vasculature, which lacks tight junctions. From there, the cytokine may traverse the arachnoid mater allowing CSF entrance and subsequently, through glymphatic influx, the brain parenchyma. Although the mechanisms underlying dural-to-CSF flow are unclear, manipulations of dural immune cell populations and specific cytokines alter cognitive functions, including learning and social behaviors, through signaling directly on cytokine receptors expressed on neurons (6, 8). Additionally, CSF reaches the dura (9) and small-molecule dyes administered to the skull surface reach the meninges, and subsequently the brain parenchyma (10), collectively supporting dural-CSF-brain as a feasible route of cytokine sensing.

In addition to providing a route for molecules to reach the brain independent of the BBB, the meninges also allow the peripheral immune system to sample the brain in a distal site. An antigen that is produced in the CNS will be carried along with the perfusing CSF out of the brain into the subarachnoid space. From there, the antigen eventually drains into the meningeal lymphatic vessels (soluble drainage) or is taken up by meningeal antigen-presenting cells, including migratory dendritic cells present in the dura (4), before eventual passage to dCLNs (cellular drainage). Under homeostatic conditions, peripheral T cells, therefore, do not require access to the brain, because they can sample the entire parenchymal antigenic repertoire within the meningeal spaces. Owing to the permeable vasculature, T cells can easily access the dura to interact with local cells presenting CNS-derived antigens (3, 9). Such a scenario would allow T cells to be retained on the basis of their antigenic specificity, or make an immediate exit through nearby dural lymphatics (see the figure).

Collectively, this highlights the dura as a feasible site of CNS immune surveillance and indeed extensive dural immune cell heterogeneity, including T cells, dendritic cells, macrophages, B cells, innate lymphoid cells type 2 (ILC2s), neutrophils, and monocytes, is present under steady-state conditions to sample brain-derived products and rapidly respond to CNS insults (4). This proposed meningeal bypass allows brain-specific antigens to be sampled by immune cells, and immune cells to travel within brain borders to modulate CNS function. Thus, does the BBB really bestow immune privilege to the brain? Clearly it limits homeostatic entrance of immune cells, yet this does not confer a lack of peripheral CNS antigen recognition or a contribution from immune cell–derived cytokines. Given the postmitotic nature of parenchymal neurons, the brain is particularly sensitive to inflammatory-mediated cell death. It is therefore conceivable that in the course of its evolution, the CNS “pushed” all of its immune activity to its borders, namely to the meninges, so that neuronal function should not be impeded by infiltrating immune cells. Evolution of the meninges did not alter the close relationship between the brain and the immune system; rather, instead of the two systems interacting and communicating directly, they do so through this distinctive meningeal interface, which could be viewed as an immune organ servicing the brain.

Several recent findings have highlighted how dysfunction in these meningeal influx and efflux systems can facilitate disease progression. Serving as a sink for brain-derived waste, CSF must be effectively drained to maintain cerebral clearance. During aging and Alzheimer’s disease, clearance of brain waste, including neurotoxic soluble β-amyloid (Aβ) species, is impaired (2), as is the conduit for subsequent CSF drainage, the meningeal lymphatics (11). These impairments aggravate Alzheimer’s disease pathology, leading to increased amyloid plaque buildup and exacerbation of cognitive dysfunction. Similarly, disruption of meningeal lymphatic drainage worsens Parkinson’s disease–like pathology in a mouse model of mutant α-synuclein overexpression (12), a pathological protein found in Lewy bodies, which abnormally aggregate within neurons and precipitate cell death. It is plausible that meningeal lymphatic drainage is critical in additional neurodegenerative diseases displaying protein aggregation, including Huntington’s disease and tauopathies.

Disrupting lymphatic drainage enhances the severity of certain neurodegenerative diseases but ameliorates neuroinflammatory conditions. Meningeal lymphatic ablation attenuates the severity of inflammation in experimental autoimmune encephalomyelitis (EAE), a mouse model of multiple sclerosis, by limiting T cell trafficking and antigen drainage to dCLNs (9). It is likely that similar antigen drainage contributes to additional autoimmune disorders, and perhaps dysfunctional drainage facilitates the aggressiveness of CNS tumors and underlies the lack of appropriate immune rejection.

Recent studies have also highlighted how meningeal entrance pathways facilitate neuroinflammatory responses. During EAE, meningeal T cells traffic across inflamed leptomeningeal vasculature, allowing CNS entrance independent of passage through the BBB (13). In aged mice, T cells accumulate within the neurogenic niche and through signaling of the cytokine interferon-γ (IFN-γ) impair adult neurogenesis (14). Although the precise entrance route(s) for T cells remains unclear, it is plausible that aging meningeal dysfunction, similar to that observed during inflammatory conditions, allows T cells to enter the CNS.

The identification of vascular channels directly connecting bone marrow of the skull with the surface of the brain has recently been described (15). Although their functional homeostatic relevance is unclear, during a stroke these channels allow direct neutrophil migration from skull bone marrow, through the meninges, and into the brain parenchyma, allowing CNS penetration independent of the BBB. Clearly, meningeal routes for brain-immune interactions exist beyond traditional concepts of BBB crossing, and understanding the meninges as an immune-blood-brain interface in homeostasis and disease could provide critical targets for disease-modifying therapeutic interventions.