Inflammation in the CNS must be tightly regulated to respond efficiently to infection with neurotropic pathogens. Access of immune cells to the CNS and their positioning within the tissue are controlled by stromal cells that construct the barriers of the CNS. Although the role of the endothelium in regulating the passage of leukocytes and small molecules into the CNS has been studied extensively, the contribution of fibroblastic stromal cells as portals of entry into the CNS was only recently uncovered. We review the critical immune-stimulating role of meningeal fibroblasts in promoting recruitment and retention of lymphocytes during CNS inflammation. Activated meningeal fibroblastic stromal cells have the capacity to rapidly elaborate an immune-competent niche that sustains protective immune cells entering the CNS from the draining cervical lymph node. Such stromal cell niches can ultimately foster the establishment of tertiary lymphoid tissues during chronic neuroinflammatory conditions.
Stromal cells build the infrastructure of an organ, ensuring its specific form, nutrient and oxygen supply, innervation, and drainage. Embedded into the framework of stromal cells are the cells of a tissue’s parenchyma, which define the specific function of the organ. In secondary lymphoid organs (SLOs), fibroblastic reticular cells (FRCs), blood endothelial cells (BECs), and lymphatic endothelial cells (LECs) represent stromal cell populations that ensure structural compartmentalization and immunological function of these tissues. Aside from their role in maintaining the unique architecture of SLOs, stromal cells play a critical role in promoting immune responses. CD31+ podoplanin (PDPN)+ LECs and CD31+ PDPN− BECs regulate lymphocyte migration into and out of SLOs (1, 2), whereas specialized populations of CD31− PDPN+ FRCs produce chemokines and cytokines that recruit and maintain T and B lymphocytes within defined microenvironmental niches (3–5) and generate a conduit system that facilitates the delivery of Ag and small signaling molecules (6). Under inflammatory conditions, transcriptional changes within these immune-stimulating stromal cells support the increased migration of lymphocytes into the lymph node (LN) (7), the extraction and presentation of Ag (8), and the repositioning of cognate T and B cells to increase the efficiency of immune cell activation and differentiation (4).
Notably, stromal cells at nonlymphoid sites may also be endowed with immune-stimulating functions under homeostatic or inflammatory conditions. In the steady-state, dermal and lung-resident stromal cells produce CCL19 and CCL21 to regulate lymphocyte recruitment to the organ (9, 10). These immunological functions may be enhanced upon inflammatory stimuli, as demonstrated by the upregulated expression of lymphoid chemokines and maturation of lymphoid-like fibroblasts in the lungs following viral infection (11). Moreover, chronic peripheral inflammation, as in the case of autoimmunity or tumors, may also result in persistent lymphocyte accumulation and maturation of immune-stimulating fibroblasts in organized ectopic aggregates commonly referred to as tertiary lymphoid tissues (TLTs) (12). Certain aspects of lymphoid-like stromal cell activation were observed in the CNS following acute pathogenic infection, as well as in the context of CNS autoimmunity. In this article, we review the spatial distribution of stromal cells in the CNS, with emphasis on meningeal fibroblastic stromal cells, and highlight their immunological function and pathways of activation during acute and chronic neuroinflammatory processes.
Stromal cells: gatekeepers to the CNS parenchyma
CNS stromal cells comprise BECs and their surrounding pericytes and smooth muscle cells, meningeal fibroblasts, LECs, and epithelial cells of the choroid plexus. Importantly, the anatomically defined position and interaction of these different stromal cell types predict different properties of their barrier function. We focus on those CNS stromal cells that regulate immune cell entry into the CNS from the bloodstream and control transition of inflammatory responses from regions proximal to CNS barriers into the parenchyma (Fig. 1A).
In addition to delivering oxygen, nutrients, and metabolites throughout the organ, BECs are a principal constituent of the blood–brain barrier (BBB) and the blood–cerebrospinal fluid barrier. With the exception of capillaries that border the choroid plexus, brain BECs are unfenestrated and tightly connected via intercellular tight junction proteins and adhesion molecules that maintain the barrier function (13). The main conducting arteries of the brain stem form the external and internal carotid and vertebral arteries that bifurcate at the Circle of Willis, traverse the meningeal surface of the CNS, and branch into the parenchyma. The BBB for solutes is located uniquely at the level of capillaries (14). Beyond the capillaries, the vascular tree widens again, transiting into postcapillary venules that ultimately rejoin meningeal veins. At this location, BECs play an important role in directing lymphocytes in the inflamed CNS via the expression of P-selectin, chemoattractants (such as CXCL12), and adhesion molecules (such as ICAM1 and VCAM1; reviewed in Ref. 15). Passage of cells (16) and drainage of macromolecules (17) from the cerebrospinal fluid to the cervical LN are regulated by LECs embedded within the dura mater and expressing lymphatic vessel endothelial hyaluronan receptor 1 and the transcription factor prospero homeobox 1.
Pericytes and vascular smooth muscle cells.
Vascular smooth muscle cells (VSMCs) and pericytes constrict blood vessels, regulate vascular permeability, and secrete a basement membrane (BM) that maintains endothelial cell organization and polarity. Topologically, VSMCs can be distinguished from pericytes by their formation of a contractile layer (tunica media) around and distinct from the endothelial BM, whereas pericytes are embedded within the endothelial BM (18). Pericytes and VSMCs share overlapping and dynamic expression of mesenchymal markers, such as α-smooth muscle actin, platelet-derived growth factor receptor (PDGFR)-β, and the chondroitin sulfate proteoglycan 4 (NG2) (18). A continuous layer of VSMCs and pericytes ensheathes larger arteries and arterioles in the meninges and parenchyma (Fig. 1A). This layer of smooth muscle cells becomes discontinuous and eventually absent at the level of capillaries and postcapillary venules (19). Consistent with the requirement to maintain chemical stability in the CNS, the BBB vasculature has the highest pericyte coverage of any organ (20). Pericytes positioned at the interface of BBB endothelial cells and the glia limitans were shown to regulate vascular permeability and instruct astroglial endfeet polarization in a PDGFR-β–dependent manner (21, 22). Moreover, under neuroinflammatory conditions, pericytes may promote leukocyte migration into the CNS via upregulation of proinflammatory mediators and adhesion molecules, such as CXCL10, CXCL8, and ICAM1 (23, 24). Intriguingly, a continuum in the phenotype and morphology of arteriole VSMCs, BBB pericytes, and venular VSMCs was identified, suggesting that different pericyte subsets exist along the neurovascular tree (19). Determining which pericyte subsets participate in neuroinflammatory responses warrants further investigation.
Fibroblasts in the CNS are concentrated in the meninges and, to some extent, along submeningeal perivascular spaces. The meninges are composed of three layers (Fig. 1A). Within the outer layer, the dura is composed of distinct fibroblast subsets and collagen fibers that attach the dura to the inner surface of the skull (25). This layer is highly vascularized and contains the dural lymphatics (26). The two inner meningeal layers, the arachnoid and pia, make up the leptomeninges that cover the entire surface of the CNS. The arachnoid is composed of an upper layer of densely aligned cells bound by tight junction proteins (27) and a deeper layer made up of loosely arranged fibroblasts and collagen that form the arachnoid trabeculae that connect to the underlying pial layer (28). The pia contains a thin sheet of flattened fibroblasts, under which is a BM of collagen and elastic fibers that separates the pia from the glia limitans and neuronal tissue (Fig. 1A). The large conducting arteries and veins that traverse the meninges are embedded within this layer of pial fibroblasts, which continue to line the glia limitans alongside branching arteries that pinch into the parenchyma (29). These fibroblasts are connected by desmosomes and small nexus junctions (28) and secrete extracellular matrix (ECM) components (30, 31) and vimentin (28), and at least a subset exhibits a CD31− PDPN+ PDGFRα+ signature (31–33).
Meningeal fibroblasts form a protective cover over the CNS parenchyma from the earliest stages of CNS development (34). Mice lacking glial BM components (30, 35), or molecules involved in meningeal fibroblast development or cell adhesion and ECM remodeling (36, 37), demonstrate neuronal outgrowth and disrupted glia limitans. Postnatally, meningeal fibroblasts continue to maintain the integrity of the pial–glial barrier, in a process that requires continued communication with neighboring glial cells (38, 39). The fibroblasts lining the glia limitans of parenchymal blood vessels gradually become patchy as the perivascular space of smaller arterioles becomes narrower, and the glial and endothelial BM fuse into a single membrane at the level of capillaries (29). In postcapillary venules, fibroblasts line the glial BM, although their distribution is reported to be sparser than in precapillary arterioles (29). Although, the cross-talk between fibroblasts and their adjacent stromal and glial cells under homeostatic and inflammatory conditions remains incompletely defined, recent studies demonstrate that at least a subset of meningeal fibroblasts acquire an immune-stimulating phenotype following pathogenic or sterile neuroinflammation (31, 33). Collectively, these stromal cells critically regulate acute and chronic inflammatory processes in the CNS.
CNS fibroblastic stromal cell responses to acute infection
The anatomy of CNS barriers reflects the necessity to protect the organ’s fragile neuronal tissue and ensure homeostasis and protection from external insults. Nevertheless, numerous pathogens have evolved diverse strategies to breach CNS barriers and replicate within this tissue (40). For example, Toxoplasma gondii (41) and SIV (42) use hematopoietic cells as a Trojan horse to enter the CNS. Other viruses, such as HSV-1, gain access to the CNS by infecting peripheral nerves (43), whereas CMV takes advantage of the more permissive, fenestrated barrier of the choroid plexus and circumventricular organs (44). Certain neurotropic viruses, including West Nile virus, Dengue virus, and Japanese encephalitis virus, reach the CNS via the hematogenic route following transmission via arthropod vectors (45). The initial antiviral immune response following infection can be driven by glial cells and perivascular macrophages that rapidly produce inflammatory and chemotactic mediators to promote the recruitment and reactivation of immune cells in the CNS. In addition to inducing direct damage to infected neuronal and glial cells, pathogens may manipulate the immune system to escape immune detection or provoke overzealous immune responses that exacerbate immunopathology. Hemorrhagic fever viruses from the filovirus family, for example, strongly suppress early host IFN responses and block adaptive immunity, thereby allowing for uncontrolled viral spread (46). The ensuing proinflammatory cytokine storm promotes increased trafficking of activated myelomonocytic cells to the CNS and ultimately results in barrier breakdown and enhanced vascular permeability (40). A similar fatal loss of barrier function due to enhanced cellular trafficking of neutrophils to the infected CNS occurs during lymphocytic choriomeningitis virus (LCMV) infection (47). Therefore, it is important for several cell types to rapidly sense pathogens and respond efficiently to limit excessive pathogen spread and immunopathology.
CNS stromal cell activation and remodeling.
Bidirectional communication between parenchymal cells (such as glial cells and neurons) and CNS stromal cells is likely required to coordinate and amplify local immune responses. Indeed, mechanisms of glial and endothelial cell cross-talk were described (48), and emerging findings suggest a crucial engagement of meningeal and submeningeal endothelial cells with surrounding fibroblastic cells to engage efficient pathogen-specific immunity (33). Although quiescent under homeostatic conditions, fibroblastic cells residing along the barriers of the CNS swiftly upregulate an immune-stimulating program upon pathogen infection (33). Interestingly, although the expression of diverse pattern recognition receptors, including TLRs, by BECs in the CNS was implicated in various inflammatory settings (49–51), danger-sensing mechanisms of CNS-resident fibroblasts have not been elucidated. However, given the phenotypic and functional similarities between LN and inflammation-activated CNS fibroblastic cells (33) and the fact that LN fibroblasts can directly sense pathogens via TLRs (7, 52, 53), it is possible that these CNS stromal cells may rely on similar mechanisms to directly sense invading pathogens. Alternatively, CNS stromal cells may be activated in a bystander manner by inflammatory cytokines produced by infected cells or by infiltrating immune cells following local or systemic infection. For instance, infected microglia (54) and neurovascular pericytes (55) were shown to produce significant amounts of TNF that can activate nearby fibroblastic cells (12). Additionally, brain endothelial cells were demonstrated to respond to systemic type I IFN by upregulating adhesion molecules and chemokines that, in turn, may act on adjacent cells (56). It is likely that this first wave of activating factors may prime fibroblastic stromal cells to respond rapidly to subsequent local inflammatory cues, such as the engagement of the lymphotoxin (LT)-β receptor (LTβR) by its heterotrimeric ligand LTα1β2 expressed by transmigrating lymphocytes (31). Nevertheless, future studies are needed to further dissect the cross-talk between CNS endothelial and fibroblastic stromal cells during neuroinflammation.
Although the mechanism of pathogen or danger sensing remains unclear, several lines of evidence demonstrate the inflammation-induced activation and remodeling of perivascular fibroblasts following neurotropic infection (Fig. 1B). Indeed, infection with the parasite T. gondii (57), as well as noncytopathic (47) and cytopathic (33) viruses, was shown to induce the elaboration of an ER-TR7+ fibroblastic cell network in the meninges that is otherwise undetectable in the naive CNS. In the case of T. gondii infection, this fibroblastic cell network was decorated with CCL21, providing a highway for CD8+ T cells to optimize their search for the pathogen (57). The crucial role of CCR7 for optimal control of the parasite (58) suggests that CCL21-producing fibroblasts provide critical support for T. gondii clearance. Recently, using a model of CNS-restricted cytopathic viral infection in a reporter mouse that marks CCL19-expressing fibroblasts (59), it was demonstrated that endothelial and fibroblastic cells lining meningeal and submeningeal vessels in the brain generate substantial amounts of CCL19 and CCL21 to recruit virus-specific CD8+ T cells to the CNS, thereby ensuring successful viral clearance (33) (Fig. 1B). The swift activation of these cells followed the trajectory of the quickly replicating mouse hepatitis virus, from the olfactory neurons at the cribiform plate and spreading distally along the brain toward the spinal cord. Moreover, CCL19-expressing fibroblasts were observed to increase the expression of the adhesion molecules ICAM1 and VCAM1, which may facilitate their interaction with incoming lymphocytes. Notably, during acute mouse hepatitis virus infection, CNS stromal cells were found to be the only cellular compartment producing CCL19 and CCL21, and partial depletion of these stromal cells led to the reduced ability of the host to control viral replication in the CNS (33). Lastly, intracranial infection with the noncytopathic LCMV promotes the appearance of an ER-TR7+ reticular cell network positioned around meningeal blood vessels that stained positive for viral Ag (47). Although this latter study focused primarily on elucidating the mechanism underlying fatal seizures in LCMV-infected animals, it would be interesting to delineate the function of the reticular cell network vis-à-vis the recruitment of neutrophils and monocytes responsible for barrier breakdown and/or the regulation of antiviral CD8+ T cell chemotaxis.
Maintenance of tissue-resident memory T cells.
Following an acute immune response, pathogens will either be eradicated or will establish a state of latent, chronic infection. In the case of successful eradication of the pathogen, the immune system creates a pool of memory cells that provides enhanced protection upon pathogen re-exposure. A fraction of memory T cells, referred to as tissue-resident memory T (TRM) cells, reside within the tissue previously targeted by the pathogen and confer long-term local immune protection. TRM cells have been detected in many peripheral tissues, including the gut (60) and skin (61), as well as the brain (62). Upon intracranial infection with an attenuated strain of LCMV, a population of CD8+ TRM cells is generated that swiftly acquires antiviral reactivity and provides protection against viral re-exposure (62). Interestingly, these cells mainly reside in fibroblast-rich regions of the meninges and around the choroid plexus (62). The finding that VCAM1+ fibroblastic stromal cells in the bone marrow are critical for the maintenance of CD4+ (63) and CD8+ (64) memory T cells through the provision of IL-7 indicates that fibroblastic stromal cells create important niches for memory cell persistence. Because fibroblastic stromal cells of SLOs represent the major source for IL-7 and IL-15 (53, 65, 66), it is possible that FRC-like fibroblasts maintain a population of TRM cells in the CNS following acute infection.
Taken together, these studies demonstrate that stromal cells strategically positioned at critical sites in the CNS have the capacity to swiftly respond to pathogenic insults by acquiring immunological functions highly similar to those of stromal cells in SLOs (Fig. 1B). These phenotypic and functional adaptations allow CNS stromal cells to recruit immune cells from the periphery, positioning such immune cells at sites in need of protection, and further shaping their functionality to achieve optimal elimination of pathogens from the CNS.
CNS stromal cell responses to chronic inflammation
The potential immunopathology associated with myeloid and T cell infiltration into inflamed tissues renders these responses inherently dangerous. Importantly, fibroblastic stromal cells foster activation of immune responses, as outlined above, as well as contribute to the attenuation of acute and chronic inflammatory reactions (67–69); this may depend on the stage of disease, with TLTs evolving over time to host a variety of pro- and anti-inflammatory cell types. Changes to fibroblastic cell types in the CNS under circumstances of chronic inflammation are best exemplified in the context of the immune-mediated, demyelinating disease multiple sclerosis (MS). In MS, autoreactive lymphocytes are primed against a CNS-reactive Ag in the periphery, enter the CNS via the vasculature, and are reactivated in the perivascular spaces of postcapillary venules and meningeal vessels (70). These two sites of preferred leukocyte accumulation correspond with classical MS lesions: demyelinated, inflammatory white matter lesions disseminating from perivascular infiltrates (71) or a gradient of demyelination and neuronal loss in cortical lesions adjacent to meningeal leukocytic aggregates (72, 73). Aside from the associated neuropathology, chronic inflammatory conditions induce changes to the endothelium (74) and promote the upregulation of chemokines (75) and the elaboration of a fibroblastic network at sites of leukocyte infiltration (76) (Fig. 1C). Notably, similar lymphoid-like changes are recapitulated during experimental autoimmune encephalomyelitis (EAE) (31, 77, 78). Although not MS, this model nevertheless provides a useful tool for dissecting the mechanism of inflammation-induced stromal cell maturation in the CNS.
The first demonstration of a lymphoid-like microenvironment within the MS-affected CNS came from Prineas (79), who observed thin-walled, lymphatic capillary-like channels and plasma cells associated with fibroblastic cells in perivascular infiltrates adjacent to demyelinated MS plaques in an organization reminiscent of the LN medulla. Further studies also reported the expression of the high endothelial venule marker HECA-452 on postcapillary venules proximal to MS white matter lesions (80). Although TLTs have only been observed in the meninges, and not in perivascular white matter lesions, an ECM-rich meshwork closely interacting with infiltrating lymphocytes was noted at the core of perivascular white matter lesions in MS (76). It remains possible that perivascular fibroblastic cells in the meninges versus at postcapillary venules exhibit inherent differences in their immune-stimulating potential.
Leukocytic aggregates composed of B cells, plasma cells, T cells, macrophages, monocytes, and fibroblasts were reported in the brain meninges of individuals with progressive MS (81), as well as in individuals newly diagnosed with MS (73). Notably, their presence correlates with a thinning of the adjacent glia limitans, cortical neuropathology, and accelerated disease progression (82). Meningeal TLTs are proposed to contribute to cortical pathology by creating an immune-competent stromal cell niche in the CNS that recruits and retains activated lymphocytes. In MS, this is primarily evidenced by the accumulation of Ag-experienced, Ig class-switched B cells in close proximity to CD21+ CD35+ CXCL13-producing fibroblasts, which resemble follicular dendritic cells (FDCs). In lymphoid tissues, FDCs facilitate germinal center responses via the capture and presentation of Ag in SLOs (83). The presence of FDC-like fibroblasts is also observed at later stages of meningeal TLT formation in different EAE models (77, 84). Notably, the immune-stimulating potential of meningeal fibroblasts may not be limited to CXCL13, but may exemplify a broader scope of chemokine and cytokine expression. Indeed, a recent study in a model of adoptive-transfer EAE demonstrated the induced expression of CXCL13, CCL19, IL-6, TGF-β, and IL-23 in the meningeal stromal cell compartment upon the transfer of preprimed highly inflammatory Th17 cells into naive recipients (31). Collectively, these studies suggest that an immune-competent niche is elaborated within the inflamed CNS of individuals with MS and EAE animals (Fig. 1C), although further study is required to understand how this niche evolves at different stages of the disease process.
Observations from various animal models suggest that TLT formation in peripheral tissues recapitulates key aspects of lymphorganogenesis, including the expression of inflammatory cytokines (such as LT and TNF) and lymphoid chemokines (such as CCL19, CCL21, and CXCL13) and the presence of high endothelial venules (12). Indeed, LTβR signaling in the stromal cell compartment was shown to contribute to TLT formation in the meninges in EAE (31, 84). However, it appears that multiple molecular pathways govern neolymphogenesis in the CNS because pharmacological treatment or genetic ablation of the LTβR is not sufficient to impede the elaboration of an ECM-rich, fibroblast network in the meninges in a model of adoptive-transfer EAE (31). In this same study, a combination of IL-17 and IL-22 contributes to the earliest stages of meningeal fibroblast remodeling in vitro, and pharmacological blockade of these cytokines attenuates fibroblast remodeling in vivo. Notably, IL-22 also was demonstrated to play a role in the maturation of intestinal lymphoid follicles in the colon during Citrobacter infection downstream of the LT pathway (85), and, more recently, to contribute to the immunological maturation of stromal cells within murine salivary gland TLTs (86). These and other studies point to a role for IL-17– and IL-22–secreting cells, such as Th17 cells and group 3 innate lymphoid cells, in promoting TLT formation at various sites of chronic inflammation (87), including EAE (31, 78). Although it remains unclear whether an elaborate fibroblastic cell network underpins meningeal TLTs in MS, recent postmortem histological analyses demonstrate the preferential presence of RORγt+ lymphocytes (with variable coexpression of IL-17) at sites of meningeal TLT, but not at sites of diffuse meningeal infiltration, suggesting that Th17 cells and possibly also group 3 innate lymphoid cells partake in inducing these niches (88). Nevertheless, the TLT-inducing role of Th17 cells (or other RORγt+ cells) may not be so black and white. Indeed, it was demonstrated that leukocytes coexpressing the classical Th1 and Th17 transcription factors T-bet and RORγt accumulate within parenchymal lesions and that IFN-γ+ Th17 cells preferentially cross the human brain endothelium in in vitro coculture experiments and in EAE (89). Moreover, the adoptive transfer of preprimed Th1- and Th17-skewed lymphocytes was observed to induce a similar degree of meningeal stromal cell remodeling and TLT formation in the aforementioned adoptive-transfer EAE studies (31). Lastly, administration of the inflammatory Th1 cytokines TNF-α and IFN-γ directly into the subarachnoid space of rats preprimed with a CNS Ag is sufficient to induce meningeal TLT-like structures and associated cortical pathology (90). Collectively, these observations suggest that cues from distinct molecular pathways are integrated by meningeal stromal cells to induce their physical remodeling and immunological maturation during CNS autoimmunity. Given the importance of CNS barrier stromal cells in response to infections, it is not surprising that there are redundancies in the molecular requirements for the formation of immune-competent niches at such vulnerable sites.
Collectively, different models of pathogen infection and immune-mediated chronic neuroinflammation consistently demonstrate that CNS stromal cells acquire immune-stimulating potential to recruit immune cells from the periphery and guide them to sites of inflammation. In this article, we focused mainly on fibroblastic stromal cells, which are closely associated with endothelial stromal cells in the meningeal and submeningeal vasculature. Although fibroblastic stromal cells may display more morphological plasticity to elaborate a highway or niche for infiltrating immune cells, efficient leukocyte migration into the CNS undoubtedly requires coordinated responses among various stromal cell types. Overall, it appears that fibroblastic stromal cells are strategically positioned to respond rapidly to inflammatory stimuli, thus contributing to optimal protection against pathogens. Yet under chronic inflammatory conditions, the elaborated immune-competent niche becomes a haven for lymphocytes that influence disease progression. In the context of chronic inflammation, we can distinguish between fibroblast remodeling and maturation, which appear to be governed by distinct and sequential molecular pathways. Nevertheless, it remains unclear to what extent the magnitude, quality, and duration of local immune responses direct stromal cell remodeling versus maturation in acute pathogen infections. In addition to positioning immune cells at sites of inflammation, activated stromal cells can secrete proinflammatory cytokines and survival factors that retain activated immune cells. The finding that TRM cells accumulate at meningeal surfaces following viral infection raises the possibility that a subset of stromal cells may secrete the survival factors to maintain this pool of memory cells. Lastly, although mature stromal cell niches were suggested to propagate immune responses in the CNS, this remains to be formally shown and likely depends on the disease state. For example, it is unclear whether stromal cells play a role in the reactivation of cytopathic or autoreactive T cells in the earliest phase of neuroinflammation during which periodic new episodes of inflammation in the CNS are actively occurring, versus the case for secondary progressive MS, in which new inflammatory episodes are less common. Although advanced imaging studies demonstrate that meningeal macrophages direct T cell reactivation at the onset of CNS autoimmunity (91), it remains possible that local stromal cells contribute to the polarizing cytokine milieu that supports T cell reactivation in acute or early CNS inflammation. Indeed, it will be important to understand whether the repertoire of cytokines and survival factors produced by immune-stimulating stromal cells is skewed under different inflammatory conditions in the CNS as it is in SLOs, as well as which molecular pathways activate immune-stimulating programs following acute versus chronic CNS inflammation.
This work was supported by grants from the Swiss Multiple Sclerosis Society, Stiftung OPOS Zugunsten von Wahrnehmungsbehinderten, and the Helmut Horten Foundation and Swiss National Science Foundation Grant 146133 (all to B.L.).
The funders had no role in study design, data collection or analysis, decision to publish, or preparation of the manuscript.
Abbreviations used in this article:
blood endothelial cell
experimental autoimmune encephalomyelitis
follicular dendritic cell
fibroblastic reticular cell
lymphocytic choriomeningitis virus
lymphatic endothelial cell
platelet-derived growth factor receptor
secondary lymphoid organ
tertiary lymphoid tissue
tissue-resident memory T
vascular smooth muscle cell.
The authors have no financial conflicts of interest.