A mesothelium divides the subarachnoid space into functional compartments

Abstract

The central nervous system is lined by meninges, classically known as dura, arachnoid, and pia mater. We show the existence of a fourth meningeal layer that compartmentalizes the subarachnoid space in the mouse and human brain, designated the subarachnoid lymphatic-like membrane (SLYM). SLYM is morpho- and immunophenotypically similar to the mesothelial membrane lining of peripheral organs and body cavities, and it encases blood vessels and harbors immune cells. Functionally, the close apposition of SLYM with the endothelial lining of the meningeal venous sinus permits direct exchange of small solutes between cerebrospinal fluid and venous blood, thus representing the mouse equivalent of the arachnoid granulations. The functional characterization of SLYM provides fundamental insights into brain immune barriers and fluid transport.

Full text

BRAIN ANATOMY
A mesothelium divides the subarachnoid space into
functional compartments
Kjeld Møllgård
1
*†, Felix R. M. Beinlich
2
†, Peter Kusk
2
†, Leo M. Miyakoshi
2
†, Christine Delle
2
,
Virginia Plá
2
, Natalie L. Hauglund
2
, Tina Esmail
2
, Martin K. Rasmussen
2
, Ryszard S. Gomolka
2
,
Yuki Mori
2
, Maiken Nedergaard
3
*
The central nervous system is lined by meninges, classically known as dura, arachnoid, and pia mater.
We show the existence of a fourth meningeal layer that compartmentalizes the subarachnoid space
in the mouse and human brain, designated the subarachnoid lymphatic-like membrane (SLYM). SLYM is
morpho- and immunophenotypically similar to the mesothelial membrane lining of peripheral organs and
body cavities, and it encases blood vessels and harbors immune cells. Functionally, the close apposition
of SLYM with the endothelial lining of the meningeal venous sinus permits direct exchange of small
solutes between cerebrospinal fluid and venous blood, thus representing the mouse equivalent of the
arachnoid granulations. The functional characterization of SLYM provides fundamental insights into
brain immune barriers and fluid transport.
Emerging evidence supports the concept
that cerebrospinal fluid (CSF) acts as a
quasi-lymphatic system in the central
nervous system (1). Cardiovascular pul-
satility drives CSF inflow along periar-
terial spaces into deep brain regions (2,3),
where CSF exchange with interstitial fluid,
facilitated by glial aquaporin 4 (AQP4) water
channels (4), takes place. Fluid and solutes
from the neuropil are cleared along multiple
routes, including perivenous spaces and cra-
nial nerves, for ultimate export to the venous
circulation via meningeal and cervical lym-
phatic vessels (5,6). CSF reabsorption may
also occur at the sinuses via the arachnoid
granulations—although this has not been
described in rodents (7–10). Despite the ef-
forts dedicated to studying CSF flow along the
glymphatic-lymphatic path, it remains to be
determined how CSF is transported within
the large cavity of the subarachnoid space
(11,12). In this study, we explored how CSF
and immune cell trafficking are organized
within the subarachnoid space surrounding
the brains of mice and humans.
The meningeal membranes were first ana-
lyzedbyinvivotwo-photonmicroscopyinthe
somatosensory cortex of Prox1-EGFP
+
reporter
mice (Prox1, prospero homeobox protein 1;
EGFP, enhanced green fluorescent protein).
Prox1 is a transcription factor that determines
lymphatic fate (13,14). Second harmonic gen-
eration was used to visualize unlabeled col-
lagen fibers, while the vascular volume was
labeledbyaCascadeBlueconjugateddextran,
and astrocytes were labeled by sulforhod-
amine 101 (SR101, intraperitoneally) (15,16).
Below the parallel-oriented collagen bundles
in dura, we noted a continuous monolayer of
flattened Prox1-EGFP
+
cells intermixed with
loosely organized collagen fibers. This sub-
arachnoid lymphatic-like membrane (SLYM)
divides the subarachnoid space into an outer
superficial compartment and an inner deep
compartment lining the brain (Fig. 1A). Quan-
titative in vivo analysis of the somatosensory
cortex revealed that the thickness of SLYM
itself was 14.2 ± 0.5 mm, hence thinner than
dura (21.8 ± 1.3 mm, n= 6 mice). The dura vas-
culature is surrounded by collagen fibers,
whereas SLYM covers the subarachnoid ves-
sels. The organization and calibers of the two
sets of vasculature also exhibit distinct differ-
ences (Fig. 1, B and C).
A key question is whether SLYM constitutes
an impermeable membrane that functionally
compartmentalizes the subarachnoid space.
To test this, Prox1-EGFP
+
mice were first in-
jected with 1-mm microspheres conjugated
to a red fluorophore into the subdural outer
superficial compartment of the subarachnoid
space along with an injection of 1-mm micro-
spheres conjugated to a blue fluorophore dis-
tributed within the inner deep subarachnoid
space compartment by cisterna magna injec-
tion (Fig. 2A). In vivo two-photon microscopy
showed that the red microspheres were con-
fined to the outer superficial compartment,
whereas the blue microspheres remained
trapped in the inner deep subarachnoid space
compartment. Quantitative analysis showed
that the 1-mm microspheres did not cross
SLYM from either side. Yet, many solutes in
CSF, such as cytokines and growth factors, are
considerably smaller than 1 mmindiameter(17).
Therefore, we sought to determine whether a
small tracer could pass through SLYM. In these
experiments, tetramethylrhodamine (TMR)–
dextran (3 kDa) was administered into the
deep inner subarachnoid space via the cister-
na magna in Prox1-EGFP
+
mice. In six mice, the
small tracer did not cross the EGFP-expressing
RESEARCH
Møllgård et al., Science 379,84–88 (2023) 6 January 2023 1of5
1
Department of Cellular and Molecular Medicine, Faculty of
Health and Medical Sciences, University of Copenhagen, 2200
Copenhagen, Denmark.
2
Division of Glial Disease and
Therapeutics, Center for Translational Neuromedicine, Faculty of
Health and Medical Sciences, University of Copenhagen, 2200
Copenhagen, Denmark.
3
Division of Glial Disease and
Therapeutics, Center for Translational Neuromedicine, University
of Rochester Medical Center, Rochester, NY 14642, USA.
*Corresponding author. Email: nedergaard@urmc.rochester.edu
(M.N.); kjm@sund.ku.dk (K.M.)
†These authors contributed equally to this work.
Fig. 1. In vivo imaging depicts a fourth meningeal layer. (A) In vivo two-photon imaging of Prox1-EGFP
+
reporter mice viewed through a closed cranial window placed over the somatosensory cortex. Maximum
projection and three-dimensional (3D) views depict the spatial distribution of dura mater collagen fibers
(gray) detected by second harmonic generation. Prox1-EGFP
+
cells (green) intermixed with the irregular
sparse collagen fibers (purple) localized below dura. This subarachnoid lymphatic-like membrane is
abbreviated SLYM. Blood vessels outlined by Cascade Blue conjugated dextran (red, 10 kDa, iv) are located at
the cortical surface. (Inset) A lateral view of the 3D reconstruction with all the layers displayed individually
along the zaxis to facilitate spatial comprehension. (B) Two-photon imaging over the sensorimotor cortex
in a Prox1-EGFP reporter mouse. The vasculature was outlined by intravenous injection of TMR-dextran
(2000 kDa), and z-stacks were collected. Representative 3D reconstruction of the z-stacks. The vasculature
in dura (magenta) is embedded in collagen fibers (white). In contrast, the vasculature in the subarachnoid
space (red) is overlaid by SLYM (green). (C) Orthogonal sections through the z-stack show that the
vasculature in dura is surrounded by collagen fibers. SLYM is located beneath dura, in close apposition
with the large-caliber subarachnoid vessels.
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023

SLYM (Fig. 2B and fig. S1). Yet, in mice with
dural damage and leakage of CSF, the tracer
was observed on both sides of the EGFP
+
mem-
brane (fig. S1). Thus, SLYM divides the sub-
arachnoid space into an upper superficial and
a lower deep compartment for solutes ≥3kDa.
SLYM is therefore a barrier that limits the ex-
change of most peptides and proteins, such as
amyloid-band tau, between the upper and
lower subarachnoid space compartments.
Live brain imaging avoids fixation artifacts
(18) but cannot immunophenotypically char-
acterize the meningeal membranes. To preserve
the integrity of the meningeal membranes,
sections were next obtained from whole heads
of Prox1-EGFP
+
mice. Immunohistochemistry
revealed that Prox1-EGFP
+
cells lined the ven-
tral parts of the entire brain surface (Fig. 3A).
Immunolabeling showed that the Prox1-EGFP
+
SLYM cells were positive for another lymphatic
marker, podoplanin (PDPN) (19), but not for the
lymphatic vessel endothelial receptor 1 (LYVE1)
(20) (Fig. 3, A, lower right panels, and D). SLYM
also labeled for the cellular retinoic acid–
binding protein 2 (CRABP2) (Fig. 3, A and D),
which is restrictively expressed in dural and
arachnoid cells during early development (21).
In contrast to SLYM, lymphatic vessels in dura
were positive for all the classical lymphatic
antigens, Prox1-EGFP
+
,PDPN
+
, LYVE1
+
,and
VEGFR3
+
, but was CRABP2
−
(fig. S2). Nota-
bly, analysis of adult human cerebral cortex
depicted that above the pia mater, a CRABP2
+
/
PDPN
+
membrane was present in the entire
subarachnoid space (Fig. 3, B and C). Thus,
SLYM also surrounds the human brain. We
Møllgård et al., Science 379,84–88 (2023) 6 January 2023 2of5
Fig. 2. SLYM represents a barrier that subdivides
the subarachnoid space into two compartments.
(A) Representative image of a 3D view of maximum
projection collected after dual injections of red
microspheres (red, 1 mm) into the outer superficial
subarachnoid space (subdural) and blue micro-
spheres delivered into the inner deep subarachnoid
space by cisterna magna injection (blue, 1 mm) in a
Prox1-EGFP
+
mouse. Graphs show a comparison
of red microspheres versus blue microspheres
detected in both the outer and inner subarachnoid
space (SAS). Two-tailed unpaired ttest; outer
SAS, P< 0.01; inner SAS, P< 0.01; n= 4 mice.
(B) Representative in vivo z-stack of a Prox1-EGFP
mouse injected with a 3-kDa TMR-conjugated dextran CSF tracer delivered via the cisterna magna. Upper panels depict SLYM (green) and the perivascular distribution
of the dextran (red) as well as the two channels merged. Lower panel displays the merge of the two channels and orthogonal optical sections showing that tracer
is confined to below the membrane. Graph shows the mean tracer intensity detected below and above the membrane. Two-tailed unpaired ttest with Welch’scorrection,
P< 0.01, n= 6 mice. Significance shown as **P< 0.01. au, arbitrary units; CM, cisterna magna; d, dorsal; v, ventral.
Fig. 3. Immunophenotypic characterization
of SLYM in the mouse and human brain.
(A) Sections of Prox1-EGFP
+
mouse brain after
decalcification of the whole head counterstained
with Mayer’s hematoxylin (M-HE, purple) show
that SLYM (arrowheads) is positively immunolabeled
for CRABP2 (brown) and Prox1-EGFP
+
/PDPN
+
/
LYVE1
−
/VEGFR3
−
and encases the entire brain,
covering its dorsal and ventral portions (purple
and blue insets, respectively). (B) Adult human
brain sections immunolabeled for CRABP2 and PDPN
reveal the presence of SLYM (arrowheads) that
enwraps the subarachnoid space blood vessels
(arrow). Ependymal and pia mater cells are also
PDPN
+
(asterisks). (C) Serial sections of the same
adult human material immunolabeled for Prox1,
PDPN, CLDN11, E-CAD, and LYVE1. SLYM is indicated
by arrowheads. (D) Confocal images of SLYM
immunolabeling showing positive labeling for
PDPN and CRABP2 (both in red). No signal was
detected for LYVE1 or VEGFR3. (E) Schematic
representations of the immunophenotypical
characterization of the meningeal layers, meningeal
lymphatic vessels, and arachnoid trabeculations.
For arachnoid trabeculae, CRABP2* signifies that
the trabeculae are CRABP2
−
in the outer SAS
but CRABP2
+
in the inner SAS. For pia, PDPN*
indicates that pia is PDPN
+
in many regions of pia, but not all. VEGFR3* signifies that pia was VEGFR3
+
only in a few regions. Aq, aqueduct; BA, basilar artery;
BS, brain stem; BV, blood vessel; Cb, cerebellum; Ctx, cerebral cortex; Ep, ependyma; LV, lateral ventricle.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023

infer that the SLYM monolayer of Prox1-EGFP
+
cells organizes into a membrane rather than
vessel structures and exhibits a distinctive
set of lymphatic markers (Fig. 3E). To dis-
tinguish SLYM from the structures forming
the arachnoid mater, we used immunolabeling
against claudin-11 (CLDN-11), a main constit-
uent of the tight junctions that create the
arachnoid barrier cell layer (ABCL) (22). CLDN-
11wasdenselyexpressedinABCLaswellasin
the stromal cells of the choroid plexus, but
SLYM was CLDN-11
−
(fig. S3, A and B). Ad-
ditionally, ABCL was distinctively positive for
E-cadherin (E-Cad) (Fig. 3C), as previously re-
ported (23,24). We also compared SLYM to the
arachnoid trabeculae (25), collagen-enriched
structures that span the subarachnoid space,
finding that cells surrounding the arachnoid
trabeculae are Prox1-EGFP
−
/LYVE1
−
(fig. S3C).
Pial cells covering the cortical surface also ex-
hibited an immune-labeling profile that dif-
fered from that of SLYM (figs. S3 and S4). We
conclude that SLYM constitutes a fourth men-
ingeal layer surrounding the mouse and hu-
man brain displaying lymphatic-like features
(Prox1-EGFP
+
,PDPN
+
, LYVE1
−
,CRABP2
+
,
VEGFR3
−
,CLDN-11
−
,andE-Cad
−
)andthat
SLYM is phenotypically distinct from dura, the
arachnoid, and pia mater (Fig. 3E). Interest-
ingly, SLYM expressed PDPN, sharing a trait
with the mesothelium lining the body cavities
(26). Accordingly, we observed PDPN
+
cells
lining the kidney, as well as PDPN
+
podocytes
in the kidneys of adult C57BL/6J mice (fig.
S5A). In a human fetus, a PDPN
+
membrane
corresponding to pericardium, pleura, and peri-
toneum encases the developing heart, lungs,
and intestinal tract, respectively. PDPN
+
lym-
phatic vessels were also observed in the lungs
and intestinal tract (fig. S5, B and C). Thus,
SLYM may represent the brain mesothelium
and, as such, covers blood vessels in the sub-
arachnoid space (Fig . 1) (26). The mesothelium
is present where tissues slide against each
other and is believed to act as a boundary
lubricant to ease movement (27). Physiological
pulsations induced by the cardiovascular sys-
tem, respiration, and positional changes of the
head are constantly shifting the brain within
the cranial cavity. SLYM may, like other meso-
thelial membranes, reduce friction between
the brain and skull during such movements.
Does SLYM have additional functions? The
arachnoid villi and granulations are defined
as protrusions of the arachnoid membrane
into the lateral walls of the sinus veins and
are believed to act as passive filters that drain
Møllgård et al., Science 379,84–88 (2023) 6 January 2023 3of5
Fig. 4. SLYM forms subarachnoid villus–like
structures at the venous sinus walls in mice.
(Aand B) Schematic diagrams illustrating the
region of interest. (Cand D) Parasagittal
consecutive sections from a decalcified mouse
whole head stained for (C) the SLYM marker
Prox1-EGFP and (D) the arachnoid barrier cell
marker E-Cad. Rectangular insets on the left in
(B), (C), and (D) are shown in higher magnification
on the right. A Prox1-positive arachnoid villus–
like structure (AV) and a vein from the dorsal
venous system are in direct contact with the
transverse sinus wall (SW), which is lacking an
intervening ABCL [inset in (C) and (D)]. The ABCL
[arrows in inset in (C)] are not stained for Prox1,
in contrast to the strongly stained SLYM layer,
whereas the opposite pattern of reactivity is
depicted in the adjacent section [inset in (D)], where
ABCL is positive for E-Cad and SLYM is negative.
Arrowheads point to pia. In (C), the narrow dura
layer, indicated by small arrows, is facing the
venous endothelial layer (VEC), indicated by
slender darker arrows. (C) and (D) are the same
magnification, as are their insets. (E) Confocal
imaging of Prox1-EGFP and E-Cad shows that the
signals do not colocalize.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023

CSF from the subarachnoid space into the
venous sinus system (7–10). The arachnoid
villi and granulations are present in the brains
of humans, primates, and larger animals such
as dogs, but not in the brains of rodents
(28,29). We critically reexamined this issue to
evaluate the distribution of SLYM in relation
to the superior sagittal and transverse sinus.
Sections obtained from decalcified heads of
Prox1-EGFP
+
mice showed that Prox1-EGFP
+
SLYM cells often were in direct contact with
the venous sinus endothelial cells (Fig. 4A).
Thus, the arachnoid barrier cell layer (CLDN-
11
+
/E-Cad
+
), which normally separates dura
from the subarachnoid space, was lacking
in discrete areas allowing SLYM to directly
contact the venous sinus wall (Fig. 4B). Prox1-
EGFP
+
SLYM cells were not positive for CLDN-
11 or E-Cad, which distinguish the arachnoid
barrier cell layer (fig. S6).
Are the close appositions of SLYM and the
venous endothelial cells permeable, allowing
the exchange of small molecules between
blood and CSF? To test this, we used the prin-
ciples of bioluminescence, wherein the con-
vergence, in the same compartment, of an
enzyme with its substrate is needed to trigger
light emission. First, we delivered the lucifer-
ase enzyme from Oplophorus gracilirostris
(NanoLuc) fused to the fluorescence tag
mNeongreen (GeNL, 44 kDa) (30) into CSF
via the cisterna magna of wild-type (C57bl/6)
mice, and allowed it to circulate for 30 min
to ensure thorough distribution by the glym-
phatic system. The distribution of GeNL was
verified by mNeongreen fluorescence. Then,
the blood-brain barrier (BBB)–impermeable
substrate fluorofurimazine (FFz, 433 Da) (31)
was administered intravenously (fig. S7, A
to C) (32). After intravenous injection of FFz,
a bright bioluminescence signal catalyzed by
GeNL was detected specifically near the large
venous sinus wall (fig. S7, A and B). The bio-
luminescence signal was particularly strong
around the confluence of sinuses (fig. S7B).
The distribution of the bioluminescence sig-
nal was quantified by plotting the mean sig-
nal intensity profiles perpendicular to the
venous wall of the transverse sinus and supe-
rior sagittal sinus. The mean bioluminescence
signal profiles intersected with the fluores-
cence signal profiles of the intravascular tracer
(TMR-dextran, 70 kDa) or with shadow im-
aging of the inverted GeNL signal outlining
the vascular wall (fig. S7C). Thus, the biolu-
minescence signal was restricted to the venous
wall of the two major sinuses lacking a BBB
(33,34), consistent with the notion that FFz
is BBB-impermeable and requires the cata-
lyzation enzyme NanoLuc to generate photons
(fig. S7, A to C). In control experiments, FFz
was delivered intravenously, while the GeNL
injection into CSF was omitted. In these con-
trol experiments, no bioluminescence signal
was detected from the exposed cortex, includ-
ingfromthesinusvenouswall(fig.S7D).In
another set of control experiments, GeNL was
injected into the soft ear tissue, while FFz
was delivered intravenously. Consistent with
the notion that peripheral blood vessels are
leaky (11), light emission was clearly observed
in the region of the ear injected with Nano-
Luc but not in surrounding noninjected re-
gions of the same ear. No signal was observed
in the venous compartment, likely reflecting
that blood flow rapidly diluted the biolumines-
cence signal (fig. S7E). Together, this analysis
shows that a small molecule, FFz, can enter
the central nervous system (CNS) from the
blood and activate an enzyme, NanoLuc, pres-
ent in CSF, resulting in the generation of
photons along the wall of the venous sinus.
On the basis of the juxtaposition of SLYM
and the venous endothelium in histological
examination (Fig. 4A), the selective generation
of photons when luciferase was injected into
CSF, and the fact that the substrate was pres-
ent in the vascular compartment (fig. S7, A
to C), we propose that the apposition of the
venous endothelia and SLYM represents ro-
dent arachnoid villus–like structures, com-
parable to those in human brain.
The mesothelium surrounding peripheral
organs acts as an immune barrier (26). Does
SLYM also impede the entry of exogenous
particles into CSF? In vivo two-photon imag-
ing of Prox1-EGFP
+
mice injected intraven-
ously with rhodamine 6G (Rhod6G) to label
leukocytes (35) showed that a large number
of Rhod6G
+
myeloid cells are embedded in
SLYM(Fig.5A).ThenumberofRhod6G
+
leukocytes in dura and SLYM was directly
comparable, suggesting a prominent role of
SLYM in CNS immune responses, which sup-
ports the finding that leptomeninges are den-
sely populated with immune cells (36)(Fig.
5A). How do systemic inflammation and aging
affect the immune cell populations residing in
SLYM? Ex vivo analysis of brain sections
obtained from Prox1-EGFP
+
mice showed that,
Møllgård et al., Science 379,84–88 (2023) 6 January 2023 4of5
Fig. 5. SLYM hosts a large number of myeloid cells. (A) (Left) In vivo two-photon microscopy of Prox1-
EGFP
+
mice injected with Rhod6G (red) shows that SLYM (EGFP
+
, green) is permeated by myeloid cells
similar to dura (collagen fibers, gray). Middle panels show orthogonal sections depicting Rhod6G
+
cells in
dura and SLYM, respectively. (Right) In vivo quantification of the number of Rhod6G
+
cells present in dura
and SLYM. Values are expressed as mean ± SEM, two-tailed unpaired ttest with Welch’s correction, P=
0.5748, n= 7 mice. (B) Representative image showing the accumulation of CD45
+
cells along the pial
vessels. (C) The percentage of area covered by CD45
+
cells was significantly increased both in aged (12- to
13-month-old) mice and in response to inflammation (LPS 4 mg/kg, ip, 24 hours). Values are expressed as
mean ± SEM, two-tailed unpaired ttest with Welch’scorrection,P< 0.005, n=3mice.(D)LYVE1macrophages
were also found in the SLYM layer, being more prominent in aged and LPS-treated animals than in healthy
young Prox1-EGFP mice. (E) The mannose receptor CD206 was detected at similar levels in the young,
aged, and LPS-treated groups, suggesting that SLYM may act as a niche for border-associated mouse
macrophages. Significance shown as *P<0.05.Ctr,control.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023

in the control group, CD45
+
cells were abun-
dant, located mostly along pial vessels in the
surface of the brain (Fig. 5B). This observa-
tion, together with the significant increase
in CD45
+
in inflammation-prone conditions
[aging and lipopolysaccharide (LPS)–treated
mice, 4 mg/kg of body weight, intraperito-
neally (ip), 24 hours] (Fig. 5C), suggests that
SLYM can act as a CD45
+
recruiting and/or
proliferating site in pathological conditions.
Of note, the dose of LPS used (4 mg/kg) did
notaffecttheBBB(fig.S8).Additionalim-
mune markers showed that LYVE1
+
(Fig. 5D),
CD206
+
(Fig. 5E), and CD68
+
(fig. S9) macro-
phages can be found in SLYM, together with
dendritic cells (CD11c
+
) (fig. S9). Despite the
absence of CD3
+
and CD19
+
lymphocytes (fig.
S9), our results indicate that SLYM functions
as a niche for immunological surveillance. Thus,
in young, healthy mice, SLYM hosts CD45
+
cells, but the number and diversity of innate
immune cells rapidly expands in LPS-induced
inflammation and was also significantly al-
tered in aged mice. We conclude that SLYM
fulfills the characteristics of a mesothelium
by acting as an immune barrier that prevents
exchange of small solutes between the outer
and inner subarachnoid space compartments
and by covering blood vessels in the sub-
arachnoid space.
Discussion
The critical roles of the meningeal membranes
lining the brain have only recently been
acknowledged (5,37). It is now known that
CSF is drained by a network of lymphatic ves-
sels in the meninges and that suppression of
this drainage accelerates protein aggregation
and cognitive decline in animal models of
neurodegeneration (38–40). SLYM is Prox1
+
/
PDPN
+
/LYVE1
−
/CRABP2
+
/VEGFR3
−
/CLDN-11
−
/
E-Cad
−
and thereby distinct from the tradi-
tional meningeal membranes, including dura,
arachnoid, and pia, as well as the meningeal
lymphatic vessels and the arachnoid trabecula.
SLYM subdivides the subarachnoid space
into two compartments, suggesting that CSF
transport is more organized than currently
acknowledged. For example, SLYM covering
the vasculature in the inner subarachnoid
space will guide CSF influx along the penetrat-
ing arterioles into the brain parenchyma with-
out circulating solutes present in the outer
subarachnoid space compartment. Yet the
discovery of a fourth meningeal layer, SLYM,
has several implications beyond fluid trans-
port. The observation that SLYM is a barrier
for CSF solutes that have a molecular weight
larger than 3 kDa will require more detailed
studies but indicates a need to redefine the
concept of CNS barriers to include SLYM. The
meningeal membranes are hosts to myeloid
cells responsible for immune surveillance of
the CNS (5,37), and SLYM, owing to its close
association with the brain surfaces, is likely to
play a prominent role in this surveillance.
Herein, we showed a large increase in the num-
ber and diversity of immune cells residing in
SLYM in response to acute inflammation and
natural aging. Physical rupture of SLYM could,
by altering CSF flow patterns, explain the pro-
longed suppression of glymphatic flow after
traumatic brain injury as well as the height-
ened posttraumatic risk of developing Alz-
heimer’s disease (41,42). Rupture of SLYM
will also permit the direct passage of immune
cells from the skull bone marrow (33,43)into
the inner subarachnoid space, with direct ac-
cess to the brain surfaces, possibly explaining
the prolonged neuroinflammation after trau-
matic brain injury (44). SLYM may also be
directly involved in CNS immunity, in addi-
tiontobeinghosttomanyimmunecells.
Lymphatic-like tissues can transform quickly
in the setting of inflammation, which in the
brain may be of notable relevance for dis-
eases such as multiple sclerosis (45).
REFERENCES AND NOTES
1. J. J. Iliff et al., Sci. Transl. Med. 4, 147ra111 (2012).
2. H. Mestre et al., Nat. Commun. 9, 4878 (2018).
3. N. E. Fultz et al., Science 366, 628–631 (2019).
4. H. Mestre et al., eLife 7, e40070 (2018).
5. A. Louveau et al., Nature 523, 337–341 (2015).
6. A. Aspelund et al., J. Exp. Med. 212, 991–999
(2015).
7. L. H. Weed, J. Anat. 72, 181–215 (1938).
8. A. Key, G. Retzius, Studien in der Anatomie des
Nervensystems und des Bindegewebes (Samson & Wallin,
1876).
9. W. E. le Gros Clark, J. Anat. 55,40–48 (1920).
10. K. Ohta et al., Kurume Med. J. 49, 177–183
(2002).
11. M. K. Rasmussen, H. Mestre, M. Nedergaard,
Physiol. Rev. 102, 1025–1151 (2022).
12. A. Drieu et al., Nature 611, 585–593 (2022).
13. I. Choi et al., Blood 117, 362–365 (2011).
14. J. T. Wigle et al., EMBO J. 21, 1505–1513
(2002).
15. K. Masamoto et al., Neuroscience 212, 190–200
(2012).
16. A. Nimmerjahn, F. Kirchhoff, J. N. Kerr, F. Helmchen,
Nat. Methods 1,31–37 (2004).
17. K. Kothur, L. Wienholt, F. Brilot, R. C. Dale, Cytokine 77,
227–237 (2016).
18. H. Mestre, Y. Mori, M. Nedergaard, Trends Neurosci. 43,
458–466 (2020).
19. M. Tomooka, C. Kaji, H. Kojima, Y. Sawa, Acta Histochem.
Cytochem. 46, 171–177 (2013).
20. S. Banerji et al., J. Cell Biol. 144, 789–801 (1999).
21. M. A. Asson-Batres, O. Ahmad, W. B. Smith, Cell Tissue Res.
312,9–19 (2003).
22. C. B. Brøchner, C. B. Holst, K. Møllgård, Front. Neurosci. 9,75
(2015).
23. J. DeSisto et al., Dev. Cell 54,43–59.e4 (2020).
24. J. Derk, H. E. Jones, C. Como, B. Pawlikowski,
J. A. Siegenthaler, Front. Cell. Neurosci. 15, 703944
(2021).
25. M. M. Mortazavi et al., World Neurosurg. 111, 279–290
(2018).
26. S. E. Mutsaers, F. J. Pixley, C. M. Prêle, G. F. Hoyne, Curr. Opin.
Immunol. 64,88–109 (2020).
27. B. A. Hills, J. R. Burke, K. Thomas, Perit. Dial. Int. 18, 157–165
(1998).
28. A. Jayatilaka, Ceylon J. Med. Sci. 18,25–30 (1969).
29. D. G. Potts, V. Deonarine, J. Neurosurg. 38, 722–728
(1973).
30. K. Suzuki et al., Nat. Commun. 7, 13718 (2016).
31. Y. Su et al., Nat. Methods 17, 852–860 (2020).
32. M. P. Hall et al., ACS Chem. Biol. 7, 1848–1857
(2012).
33. J. Rustenhoven et al., Cell 184, 1000–1016.e27
(2021).
34. P. Mastorakos, D. McGavern, Sci. Immunol. 4, eaav0492
(2019).
35. H. Baatz, M. Steinbauer, A. G. Harris, F. Krombach, Int. J.
Microcirc. Clin. Exp. 15,85–91 (1995).
36. A. Merlini et al., Nat. Neurosci. 25, 887–899
(2022).
37. A. Louveau et al., J. Clin. Invest. 127, 3210–3219
(2017).
38. S. Da Mesquita, Z. Fu, J. Kipnis, Neuron 100, 375–388
(2018).
39. Z. Xu et al., Mol. Neurodegener. 10, 58 (2015).
40. L. Wang et al., Brain Pathol. 29, 176–192 (2019).
41. A. Z. Mohamed, P. Cumming, J. Götz,
F. Nasrallah; Department of Defense Alzheimer’s Disease
Neuroimaging Initiative, Eur. J. Nucl. Med. Mol. Imaging 46,
1139–1151 (2019).
42. J. D. Flatt, P. Gilsanz, C. P. Quesenberry Jr., K. B. Albers,
R. A. Whitmer, Alzheimers Dement. 14,28–34 (2018).
43. S. Brioschi et al., Science 373, eabf9277 (2021).
44. J. J. Iliff et al., J. Neurosci. 34, 16180–16193
(2014).
45. N. B. Pikor, A. Prat, A. Bar-Or, J. L. Gommerman,
Front. Immunol. 6, 657 (2016).
ACKNO WLED GME NTS
We acknowledge P. S. Froh and H. Nguye n (Department o f
Cellular and Mole cular Medicine, Fac ulty of Health and M edical
Sciences, Univer sity of Copenhag en, Denmark) f or their excelle nt
technical assis tance for the histo logy and immunoh istochemistr y
of the decalcifi ed samples. We also thank D. Xue for exp ert
graphical suppo rt, B. Sigurdsso n for analysis, and H. Hirase,
N. Cankar, and N. C. Petersen for critical reading of the manuscript.
Funding: Funding was provided by Lundbeck Foundation grant
R386-2021-165 (M.N.), Novo Nordisk Foundation grant
NNF20OC0066419 (M.N.), the Vera & Carl Johan Michaelsen’s
Legat Foundation (K.M.), National Institutes of Health grant
R01AT011439 (M.N.), National Institutes of Health grant
U19NS128613 (M.N.), US Army Research Office grant MURI
W911NF1910280 (M.N.), Human Frontier Science Program grant
RGP0036 (M.N.), the Dr. Miriam and Sheldon G. Adelson Medical
Research Foundation (M.N.), and Simons Foundation grant 811237
(M.N.). The views and conclusions contained in this article are
solely those of the authors and should not be interpreted as
representing the official policies, either expressed or implied, of
the National Institutes of Health, the Army Research Office, or the US
Government. The US Government is authorized to reproduce and
distribute reprints for Government purposes notwithstanding any
copyright notation herein.The funding agencies have taken no part on
the design of the study, data collection, analysis, interpretation, or in
writing of the manuscript. Author contributions: K.M. and M.N.
designed the study. F.R.M.B., P.K., L.M.M., C.D., V.P., M.K.R., R.S.G.,
N.L.H., T.E., and Y.M. performed the ex periments, collected the data,
and performed the analysis. K.M. and M.N. wrote the manuscript.
All authors read and approved the final version of the manuscript.
Competing interests: The authors declare that they have no
competing interests. Data and materials availability: All data are
available in the main text or the supplementary materials.
License information: Copyright © 2023 the authors, some rights
reserved; exclusive licensee American Association for the
Advancement of Science. No claim to original US government
works. https://www.science.org/about/science-licenses-journal-
article-reuse
SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.adc8810
Materials and Methods
Figs. S1 to S9
Table S1
References (46–54)
MDAR Reproducibility Checklist
Movie S1
View/request a protocol for this paper from Bio-protocol.
Submitted 6 May 2022; resubmitted 13 September 2022
Accepted 7 December 2022
10.1126/science.adc8810
Møllgård et al., Science 379,84–88 (2023) 6 January 2023 5of5
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023

Use of this article is subject to the Terms of service
Science (ISSN ) is published by the American Association for the Advancement of Science. 1200 New York Avenue NW, Washington, DC
20005. The title Science is a registered trademark of AAAS.
Copyright © 2023 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works
A mesothelium divides the subarachnoid space into functional compartments
Kjeld MøllgårdFelix R. M. BeinlichPeter KuskLeo M. MiyakoshiChristine DelleVirginia PláNatalie L. HauglundTina
EsmailMartin K. RasmussenRyszard S. GomolkaYuki MoriMaiken Nedergaard
Science, 379 (6627), • DOI: 10.1126/science.adc8810
An extra layer lines the brain
The traditional view is that the brain is surrounded by three layers, the dura, arachnoid, and pia mater. Møllgård
et al. found a fourth meningeal layer called the subarachnoid lymphatic-like membrane (SLYM). SLYM is
immunophenotypically distinct from the other meningeal layers in the human and mouse brain and represents a
tight barrier for solutes of more than 3 kilodaltons, effectively subdividing the subarachnoid space into two different
compartments. SLYM is the host for a large population of myeloid cells, the number of which increases in response to
inflammation and aging, so this layer represents an innate immune niche ideally positioned to surveil the cerebrospinal
fluid. —SMH
View the article online
https://www.science.org/doi/10.1126/science.adc8810
Permissions
https://www.science.org/help/reprints-and-permissions
Downloaded from https://www.science.org at Copenhagen University on January 05, 2023