Lymph node cortical sinus organization and
relationship to lymphocyte egress dynamics
and antigen exposure
Irina L. Grigorovaa,1,2, Mikhail Panteleevb,c, and Jason G. Cystera,2
aThe Howard Hughes Medical Institute and Department of Microbiology and Immunology, University of California, San Francisco, CA 94143;bCenter for
Theoretical Problems of Physico-Chemical Pharmacology, Russian Academy of Sciences, Moscow 125167, Russia; andcNational Research Center for
Hematology, Russian Academy of Medical Sciences, Moscow 125167, Russia
Edited by Max D. Cooper, Emory University, Atlanta, GA, and approved October 13, 2010 (received for review July 9, 2010)
Recent studies have identified cortical sinuses as sites of sphingo-
sine-1-phosphate receptor-1 (S1P1)-dependent T- and B-cell egress
from the lymph node (LN) parenchyma. However, the distribution
of cortical sinuses in the entire LN and the extent of lymph flow
within them has been unclear. Using 3D reconstruction and intra-
vital two-photon microscopy we describe the branched organiza-
tion of the cortical sinus network within the inguinal LN and show
that lymphocyte flow begins within blunt-ended sinuses. Many
cortical sinuses are situated adjacent to high endothelial venules,
and some lymphocytes access these sinuses within minutes of en-
tering a LN. However, upon entry to inflamed LNs, lymphocytes
rapidly up-regulate CD69 and are prevented from accessing cortical
sinuses. Using the LN reconstruction data and knowledge of lym-
phocyte migrationand cortical sinus entrydynamics,wedeveloped
a mathematical model of T-cell egress from LNs. The model sug-
gests that random walk encounters with lymphatic sinuses are
the major factor contributing to LN transit times. A slight discrep-
ancy between predictions of the model and the measured transit
times may be explained by lymphocytes undergoing a few rounds
of migration between the parenchyma and sinuses before depart-
ing from the LN. Because large soluble antigens gain rapid access to
cortical sinuses, such parenchyma–sinus shuttling may facilitate
antigen capture|inflammation|interstitial fluid|laminar flow
cytes get into LNs from the blood through high endothelial ven-
ules (HEVs); they then move into the T zone and B-cell follicles,
respectively, and migrate there in a stromally guided random
walk. If no antigenic stimuli have been encountered, lymphocytes
leave into the efferent lymphatics after a characteristic residence
time in murine lymph nodes (LNs) of 6–10 h for T cells and 12–
24 h for B cells (1–3). During some immune responses, egress of
naive lymphocytes from LNs is transiently blocked by IFNα/β.
This egress “shutdown” has been modeled by systemic treatment
with double-stranded mRNA mimetic polyinosine-polycytidylic
acid [poly(I:C)] that induces secretion of IFNα/β, as well as with
lymphocytic choriomeningitis virus (LCMV) infection (4). The
induced block in egress partially depends on lymphocyte-intrinsic
up-regulation of CD69. CD69 is a transmembrane protein that
negatively regulates sphingosine-1-phosphate receptor-1 (S1P1),
a receptor for sphingosine-1-phosphate (S1P) that is required for
lymphocyte egress (4, 5). However, the stage at which CD69
inhibits cell departure from the LNs has not been fully defined.
Recent studies of murine LNs identified lymphatic vascular
endothelial gene-1 (LYVE-1)+cortical sinuses as sites of T- and
B-cell exit from the LN parenchyma (6–8). Using intravital two-
photon laser scanning microscopy (TPLSM), it was shown that
intrinsic expression of S1P1on T lymphocytes regulates their ac-
cess into the cortical sinuses, whereas fluid flow within these
sinuses mediates their retention and passive transport toward the
ymphocyte recirculation between blood, lymphoid organs, and
lymph is essential for immune surveillance. T and B lympho-
medulla and efferent lymphatic (7). However, cortical sinuses
often appear packed with cells (6, 9–12) and it has been unclear
whether all of these structures exhibit fluid flow.
A systematic view of lymphocyte recirculation through LNs
under normal and inflamed conditions requires better un-
derstanding of (i) the distribution of the lymphocyte exit sites in
the LNs (LYVE-1+sinuses with flow in them) relative to entry
sites (HEVs), (ii) the time within which lymphocytes could access
and transmigrate into the exit sites after arrival into the LNs, (iii)
the fraction of cells that return back into LN parenchyma from
LYVE-1+sinuses, and (iv) the regulation of these processes by
local inflammation. In this work we have undertaken studies to
address these questions. By confocal microscopy we obtained a
3D reconstruction of HEV and LYVE-1+sinus distribution in an
entire inguinal lymph node (ILN). By TPLSM we find that mul-
tiple blunt-ended cortical sinuses show evidence of cell flow.
Large soluble antigen gains efficient access into cortical sinuses,
providing further evidence for fluid flow in these structures while
also suggesting that they may function as sites of antigen acqui-
sition by B cells. Many cortical sinuses are proximal to HEVs, and
newly entered lymphocytes have rapid access into these exit sites.
In a model of local inflammation induced by poly(I:C), naive ly-
mphocytes coming into an inflamed LN rapidly up-regulate CD69
and are blocked from accessing cortical sinuses. We developed
a quantitative model of T-cell egress from ILNs that incorporates
the experimentally measured distribution of the cortical sinuses,
known T-cell motility parameters, and sinus entry efficiency. We
time in the LN, it is necessary to propose that lymphocytes shuttle
between the parenchyma and sinuses a few times before reaching
the efferent lymphatics.
Blunt-Ended Cortical Sinuses Exhibit Flow. To gain quantitative in-
formation about the positioning and morphology of cortical
sinuses in murine LNs, we performed a 3D reconstruction of
LYVE-1+sinuses and HEVs in an entire ILN by serial sectioning;
staining for LYVE-1, CD31 (PECAM1), and CD4; and analyzing
Author contributions: I.L.G. and J.G.C. designed research; I.L.G. performed research and
modeling analysis of LN egress; I.L.G. and J.G.C. analyzed data; M.P. designed the model
of antigen diffusion against liquid flow and performed dynamic simulations; and I.L.G.,
M.P., and J.G.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Freely available online through the PNAS open access option.
1Present address: Department of Microbiology and Immunology, University of Michigan
Medical School, Ann Arbor, MI 48109.
2To whom correspondence may be addressed. E-mail: email@example.com or jason.cyster@
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| November 23, 2010
| vol. 107
| no. 47
20- to 30-μm sections and then compiled into a single file. This
sinuses that started near HEVs in interfollicular regions or at the
follicle/T-zone boundary (Fig. 1), in agreement with previous
studies(12,13).Severalconnectingsinuses that extendedfromthe
subcapsular sinus (SCS) to medullary sinuses were also identified
and usually were located adjacent to a B-cell follicle as well as
passing near HEVs (Fig. 1 and Movies S1 and S2). The 3D map
obtained of the LYVE-1+structures allowed sinus distribution
with respect to HEVs and the total sinus surface area to be mea-
sured, providing necessary parameters for egress modeling, as
described further below. As an approach to examine whether the
blunt-ended cortical sinuses showed evidence of cell flow, we de-
thesinusand not touching the wall, within thelumen and touching
the wall, or within the parenchyma near the sinus (Fig. 2 A and B).
This analysis showed that most of the cells within the lumen and
indicating they were more rounded and thus likely caught and
retained in fluid flow (Fig. 2 A and B) (7). We then asked whether
we could detect cells flowing within blunt-ended structures using
intravital TPLSM. In a series of experiments we observed a steady
lymph flow in the SCS (detected by faint autofluorescence of the
the surgery, but after this time the flow became more variable.
Studies in animal models and in humans have shown that pro-
longed anesthesia is associated with a reduction in lymph flow (14,
15). When the TPLSM analysis was performed within the first 2–3
in the direction of the medulla (Movies S3 and S4). We therefore
suggest that blunt-ended sinuses act as lymphocyte-retaining
High Molecular Weight Antigen Gains Rapid Access to Cortical Sinuses.
Because we detected slowly flowing cells in many cortical sinuses,
we speculated that lymph-borne antigens accessing medullary
sinuses may diffuse via the lymph to cortical sinuses. To test this
possibility, we injected mice with phycoerythrin-coupled hen egg
lysozyme (HEL-PE) (∼300 kDa), an antigen complex that is too
large to access the LN parenchyma via conduits (16). To facilitate
visualization of the antigen without the complication of loss
during tissue preparation, we first transferred two populations of
reporter B cells that are able to bind HEL with high affinity: wild-
type MD4 Ig-transgenic B cells that localize in follicles and
CXCR5-deficient MD4 Ig-transgenic B cells that localize to inter-
follicular and follicle-proximal T-zone regions. These are the sites
where most of the cortical sinuses are situated. Three minutes
after injection of HEL-PE into the footpad, many MD4 B cells
in the cortical sinuses, medulla, and SCS had bound the antigen,
whereas the majority of MD4 B cells in the parenchyma were
unstained (Fig. 3A). Similar labeling was observed in ILNs 20–30
min after s.c. injection of HEL-PE in the flank and the base of the
tail (Fig. 3B). The more rapid labeling after footpad compared
with flank injection may reflect greater hydrostatic pressure in the
former injection site, leading to more rapid movement of antigen
through the afferent lymph. At 40–60 min after s.c. injection of
HEL-PE, MD4 B cells in proximity to the cortical, medullary, and
SC sinuses of the draining ILN had bound the antigen (Fig. 3C).
The rate of HEL-PE appearance in LN sinuses and capture by
MD4B cells varied between animals, likely reflecting variationsin
drainage from the injection site. However, a similar sequence of
labeling in sinuses and surrounding parenchyma was seen in
multiple experiments (Fig. S1). HEL-PE gave brighter staining in
the medulla than in cortical sinus regions after both s.c. and
footpad injections (Fig. 3 and Fig. S1). This result is consistent
with access of HEL-PE to cortical sinuses by diffusion from the
medullary sinuses against the direction of the slowly flowing
through an ILN from its cortical side, obtained by reconstruction of serial
LYVE-1 antibody stained cortical, medullary, and subcapsular sinuses (SCS), as
well as macrophages in the medullary region. MR, medullary region between
the two lobes of ILN. (Scale bar, 300 μm.) Data are representative of ILN serial
section analysis from three mice (Movie S1). (B) Projection view of a 46-μm-
thick section of ILN 70 μm underneath the capsule from the cortical side,
stained to detect LYVE-1 (green) and CD31 (red) and imaged by confocal
microscopy. (Scale bar, 300 μm.) Data are representative of ILN serial section
analysis from two mice (Movie S2). Arrowheads in A and B indicate blunt-
ended sinuses. Arrow in A indicates a sinus connected to the SCS at both ends.
Projection view of LYVE-1+sinuses and HEVs in ILN. (A) 3D view
Confocal microscopy image showing a section from a blunt-ended LYVE-1+
sinus, stained for LYVE-1 (green) and CD4 (red). (Scale bar, 50 μm.) (B) Axis
ratio of CD4 T cells inside blunt-ended LYVE-1+cortical sinuses. Data are
shown for T cells inside and not touching the sinus wall (white circle),
touching the sinus wall (light gray circle), and outside of the sinuses (dark
gray circle). Data are combined from two serial reconstructions of ILNs (two
mice), and in each, four different blunt-ended structures were analyzed.
Medians are shown by horizontal lines.
Lymphocytes within blunt-ended cortical sinuses are rounded. (A)
images of peripheral LN sections, stained to detect LYVE-1 (green) and the
transferred MD4 or MD4 CXCR5−/−B cells (by HEL-A647, blue). (A) Projection
view of 6.4-μm-thick section of a popliteal LN with MD4 CXCR5−/−B cells 3
min after injection of HEL-PE into the footpad. (Scale bar, 100 μm.) (B)
Projection view of 6.4-μm-thick section of ILN with MD4 and MD4 CXCR5−/−B
cells 30 min after s.c. injection of HEL-PE into the flanks and the base of the
tail. (Scale bar, 50 μm.) (C) Combined view of three adjacent sections of ILN
(19- to 26-μm-thick projection views) with MD4 B cells, 40 min after s.c. in-
jection of HEL-PE into the flanks and base of the tail. (Scale bar, 100 μm.)
Squares indicate MD4 and MD4 CXCR5−/−B cells stained with HEL-PE. FO,
follicle; T, T zone.
Emergence of s.c. injected HEL-PE in cortical sinuses. (A–C) Confocal
| www.pnas.org/cgi/doi/10.1073/pnas.1009968107Grigorova et al.
cortical sinus lymph. In support of this hypothesis, dynamic cal-
culations of diffusion against laminar liquid flow (SI Materials and
Methods) suggest that protein antigens with molecular mass ∼300
kDa should rapidly spread into the sinuses if the liquid flow rate is
comparable to or even a few fold faster than the experimentally
measured flow rate of cells in the sinuses (7) (Fig. S2).
Previous studies have shown that the small molecular weight
from the SCS or via conduits (17, 18), although how this antigen
distributes with respect to cortical sinuses was not determined.
Two to 5 min after footpad injection of HEL-A647, popliteal LNs
showed a wave front-like labeling pattern of the MD4 B cells that
propagated through the follicles away from the capsule and the
brightly stained medulla (Fig. S3). Although the extent of MD4
B-cell staining in the cortical sinuses was variable within this time
period, they always appeared stained where the HEL-A647 wave
front from the follicles or from the medulla reached the sinuses.
Overall, the similar kinetics of HEL-A647 localization into the
follicles and cortical sinuses suggest that the cortical sinuses
themselves are not a preferential path for LN access by small anti-
follicles or spreading through sparse follicular conduits (17, 18)
By 5–10 min afterHEL-A647 footpad injection, most MD4 B cells
proximal to or within the T zone were also stained.
Inourpreviousanalysis of T-cellbehaviorin cortical sinuses, we
noted that some cells moved from the sinus back into the paren-
chyma, although this behavior was less profound in regions with
flow (7). Here we observed examples of B cells migrating from
cortical sinus regions with flow back into the parenchyma (Movies
S4 and S5). These observations suggest some amount of lympho-
cyte “shuttling” between sinuses and parenchyma, as also sug-
gested by the quantitative analysis below, and might provide
a mechanism for B-cell return into the parenchyma after an en-
counter with large soluble antigens in the sinuses.
Rapid Access of Lymphocytes into Cortical Sinuses. The close prox-
imity of cortical sinuses to HEVs (Fig. 1) led us to ask how quickly
cells entering LNs through HEVs can gain access to the sinuses.
To address this question mice were injected with lymphocytes i.v.
followed by LN section analysis to determine positioning of the
into sinuses were observed. The cells arriving in the LN at the
junction of HEVs with lymphatic sinuses appeared aligned along
the wall of the sinus (Fig. 4A). However, at 20 and 30 min after
injection, many transferred lymphocytes were detected within the
sinuses (Fig. 4B, Movie S6). These observations suggest that
S1P1, which is down-regulated on lymphocytes in the blood
(20), resensitizes sufficiently within 20 min of lymphocyte entry to
the LN parenchyma for their productive transmigration into
Lymphocyte Exclusion from Cortical Sinuses During Inflammation.
after their entry into the LN implies that a fraction of lymphocytes
maygo into theefferent lymphaticsbefore surveyingforantigen in
the parenchyma. However, antigen presentation in lymphoid
organs is often accompanied by inflammation. We therefore
attempted to assess what happens to the newly arriving cells in
locally inflamed LNs. IFNα/β induced in response to poly(I:C)
injection causes up-regulation of lymphocyte CD69 and inhibition
of the egress-promoting function of S1P1(4). Six hours after sys-
LNs were partially or fully collapsed (Fig. S4A). To test whether
sinus collapse occurs because of the displacement of CD69 up-
regulated lymphocytes from the exit structures, we asked whether
lymphocyte-intrinsic deficiency in CD69 would reverse the phe-
bone marrow (BM) chimeras showed that at 6 h after poly(I:C)
treatment, cortical sinuses in [CD69−/−↦ WT] BM chimeras were
less collapsed than in [CD69+/+↦ WT] BM chimeras (Fig. S4B).
Moreover,quantitativeanalysis ofCD69−/−and CD69+/+T andB
cells cotransferred into wild-type recipient mice showed that
CD69−/−cells were enriched in the cortical sinuses relative to the
basis of these data, we infer that CD69 up-regulation reduces
lymphocyte entry into cortical sinuses. This reduction results in a
in mice treated with FTY720 or in lymphatic sphingosine kinase-
deficient mice (8, 21).
To address how quickly CD69 is up-regulated on lymphocytes
arriving into an inflamed LN, we induced localized LN in-
flammation by footpad injection of poly(I:C) with simultaneous
control injection of PBS into the contralateral footpad (Fig. 5E).
By 6 h after injection, cells in poly(I:C) draining popliteal LNs
started up-regulating CD69 whereas in the contralateral LN (and
the other peripheral LNs) CD69 up-regulaton did not occur.
Analysis of CD69 up-regulation by lymphocytes arriving into the
inflamed popliteal LN showed that B cells and CD8 T cells started
up-regulation of CD69 by 30–60 min after their arrival (Fig. 5F).
The onset of CD69 up-regulation was a little delayed in CD4
T cells, taking place between 60 and 90 min after their entry (Fig.
5E). These data show that CD69 up-regulation by lymphocytes in
the inflamed LN is rapid compared with the median LN residence
time and thus could significantly reduce the chance of potentially
antigen-reactive lymphocytes undergoing premature exit from an
Quantitative Model of T-Cell Egress from ILN. In a currently envi-
sioned model of lymphocyte egress from LNs, naive lymphocytes
can encounter LYVE-1+sinuses while moving in the LN by
a stromally guided random walk. Upon the encounter they either
where they can become captured by flow and carried into the ef-
ferent lymphatics (6, 7). However, whether (i) distribution of the
upon contact, and (iii) lymphocyte motility within the T zone are
sufficient to explain the experimentally measured kinetics of lym-
phocyte egress from the LN is unclear. To test this, we developed
two quantitative models of T-cell egress from the ILN, using two
different approaches to model T-cell motility within the lymph
node: the simple model (SM) (22) and the conditional probability
model (CPM) (SI Materials and Methods). On the basis of these
quantitative models and the experimentally measured parameters
(i–iii), we calculated the rate of lymphocyte entry into the sinuses
images of peripheral LN sections from mice, i.v. injected 30 min earlier with
(yellow), and stained to detect LYVE-1 (green) and CD31 (blue). (A) Projection
view of 1.9-μm-thick section. (Scale bar, 50 μm.) (B) Projection view of 10-μm-
thick section (Movie S6). (Scale bar, 100 μm.) Data are representative of two
Rapid access of lymphocytes into cortical sinuses. (A and B) Confocal
Grigorova et al. PNAS
| November 23, 2010
| vol. 107
| no. 47
(Fig. 6 and Fig. S7) and compared it with an experimentally mea-
sured residence time of T cells in ILNs.
Both the SM and the CPM models of lymphocyte egress
predict that ∼30% of T cells arriving in the ILN through HEVs
encounter the exit structures within the first 30 min, and by 2 h
half of them encounter the sinuses at least once (Fig. 6 A and B).
The newly arriving lymphocytes have a higher probability of
encountering the sinuses compared with the cells residing in the
ILN for >2 h (Fig. 6 A and B, note the slope of the curve), due to
the proximity of the cortical sinuses to HEVs. Overall, the CPM
predicts an ∼20–25% faster rate of T-cell encounter with the
sinuses and egress compared with the SM (Fig. 6). Under an
assumption that the efficiency of lymphocyte transmigration into
the sinuses does not change with time following their arrival in
the ILN, the expected half-life of T cells in the ILN (before they
transmigrate into LYVE-1+sinuses) is predicted to be ∼4–5 h
(Fig. 6 C and D). This value would be only slightly delayed if the
cells were unable to transmigrate into the sinuses (for example,
due to S1P1resensitization requirements) within the first 15 or
30 min after their entry (Fig. 6 C and D).
In addition to the quantitative modeling, we performed a volu-
metric analysis of T-cell access to LYVE-1+sinuses (SI Materials
and Methods). This analysis is based on the ratio of the cumulative
the T zone (extracted from the representative ILN) and on the
experimentally measured T-cell transmigration frequency and
thetime T cells spent in contact with thesinuses. In contrast to the
quantitativemodeling,volumetric analysis does not depend on the
distribution of the entry sites and T-cell motility. The range of
half-lives for T-cell transmigration into LYVE-1+sinuses esti-
mated by volumetric analysis iseither fasterthan or comparable to
the predictions of the quantitative model (SI Materials and Meth-
ods and Fig. S8), with both of them suggesting a slightly faster rate
of T-cell access into the sinuses from the T zone than the reported
flammation. (A–D) Analysis of confocal images of peripheral LN sections
from mice that had received cotransferred CD69+/+(labeled with CFSE, blue)
and CD69−/−(labeled with CMTMR, red) T or B lymphocytes and were trea-
ted for the last 6 h with PBS or poly(I:C). (A and B) Representative sections,
stained to detect LYVE-1 (green). (Scale bars, 100 μm.) (C and D) Quantitative
analysis showing relative changes in the ratio of the CD69−/−to CD69+/+(C)
T-cell or (D) B-cell numbers in the LYVE-1+cortical sinuses compared with
the (C) T zone and (D) follicles. In each mouse at least ten 10- to 30-μm-thick
(375 × 375 μm) sections containing cortical sinuses were analyzed. (E and F)
Kinetics of CD69 up-regulation by T and B lymphocytes after their entry into
locally inflamed popliteal LNs. (E) Schematic of the experiment. Six hours
after poly(I:C) injection into the right footpad (RFP) and PBS into the left
footpad (LFP), Ly5.1 recipient mice were i.v. injected with Ly5.2 lymphocytes.
After 20 min, further entry of lymphocyte into the LNs was blocked by i.v.
injection of α4- and αL-specific antibodies. (F) Flow cytometric analysis of
CD69 up-regulation by the transferred cells in inflamed right popliteal LN
(RPLN, Lower) or control LNs (LPLN, Upper) at 1 h (blue), 1.5 h (red), 2 h
(green), and 3 h (violet) after transfer. Data shown are representative of
three experiments performed at 6 h and one experiment at 18 h.
Exclusion of CD69hilymphocytes from cortical sinuses during in-
cortical sinuses in the ILN. Calculations were performed using T lymphocyte
motility modeled as a simple random walk (SM) (A and C) or a conditional
probability random walk (CPM) (B and D) and 3D reconstruction of HEVs and
LYVE-1+sinuses of a representative ILN. For description see SI Materials and
Methods. (A and B) Fraction of T cells that have not encountered cortical
LYVE-1+sinuses at various times after cell arrival into the ILN through HEVs.
Simulations were performed with minimal distance set to 2 μm and contact
distance of 5 μm (black circles) and 3 μm (gray circles). The dashed red line
indicates the region of constant probability of encountering LYVE-1+sinuses
between ∼2 and 10 h after cell entry and higher probability of sinus contact
before 2 h after entry. τ1/2, calculated half-life for the first encounters of cells
with sinuses. (C and D) The fraction of cells that have not transmigrated into
LYVE-1+sinuses from the parenchyma if the probability of transmigration
upon contact is 0.3. The probability of transmigration is constant (black
circles), zero in the first 15 min after entry (white circles), or zero in the first
30 min after entry (gray circles). Simulations were performed with minimal
distance set to 2 μm and contact distance of 5 μm. τ1/2, calculated half-life for
cell transmigration into sinuses (Fig. S6).
Predictions of the quantitative models of lymphocyte entry into
| www.pnas.org/cgi/doi/10.1073/pnas.1009968107 Grigorova et al.
rate of CD4 and CD8 T-cell exit from LNs (3). If the kinetics of
lymphocyte exit from ILNs is tightly regulated (as would be
expected on the basis of the previous report) (3), then the dis-
crepancy between cell entry into the sinuses and egress from the
ILN may be caused by overestimation of cell entry into some cor-
tical sinuses. The frequency of T-cell transmigration into the
sinuseswas similar in variousregions ofILNs imagedby TPLSM in
four separate experiments (Fig. S5A) (7, 8). Therefore, we specu-
late that T-cell transmigration frequency does not vary a lot be-
tween various regions of the exit sites, although we cannot exclude
its gradual variation over time (timer mechanism). For a timer
mechanism to be relevant, it would have to operate over a long
timescale, because we found little effect of reducing sinus entry
efficiency for the 15–30 min after entry. However, multiple exam-
ples of lymphocyte return into parenchyma from the sinuses led us
to propose an alternative mechanism that might contribute to
lymphocyte transit time within LNs. Lymphocyte recirculation
between the sinuses and parenchyma was experimentally observed
before and in this work (Movies S4 and S5) (7, 8), especially in
restrained by macrophages close to the capsule. The experimental
assessment of the physiological extent of lymphocyte return from
the cortical sinuses back into the parenchyma may be somewhat
ambiguous due to reduction in the fluid flow occurring during the
T-cell egress suggest that to achieve an agreement between the
calculated rate of cell entry into the sinuses and the experimentally
measured half-life of T cells in the ILNs (8–9 h) (3), about half of
the T cells that transmigrate into the sinuses should return back
into the T zone.
In this study we demonstrate that blunt-ended cortical sinuses in
LNs contain many rounded lymphocytes that are moving toward
medullary sinuses. We suggest that the presence of lymph flow in
cortical sinuses promotestheroundingup and capture of cellsthat
have entered in an S1P1-dependent manner, reducing their pro-
pensity to migrate back into the parenchyma and helping ensure
they reach medullary sinuses and exit the LN. Using quantitative
information about major LN exit sites, the volume of the T zone,
and the established migration parameters of LN T cells, we esti-
mate LN transit times that are quite close to the measured values,
suggesting that these are the major parameters determining naive
lymphocyte transit time. Although the model predicts egress times
slightly faster than those that have been measured, our dynamic
analysis shows that some cells do return to the parenchyma from
sinuses, providing a likely explanation for the discrepancy.
The pathways of fluid flow into cortical sinuses have not been
well defined. Conduits have been suggested to channel lymph ar-
riving in the SCS to HEVs for return to blood circulation (12, 19).
Because many blunt-ended cortical sinuses initiate near HEVs, it
seems likely that such channeling also delivers lymph to cortical
sinuses, and some tracer studies are consistent with this in-
terpretation (23–25). In TPLSM experiments we observed simul-
taneous decreases in the amount of afferent lymph in the SCS
(suggested by collapse of the ILN capsule after a few hours of
intravital imaging) and in the cell flow rate in cortical sinuses,
parenchyma and/or conduits. Due to the detection of higher pro-
tein concentrations in efferent compared with afferent lymph, the
close proximity between HEVs and lymphatics, and their positive
staining for aquaporin, it was previously suggested that lymph may
flow from the blunt-ended cortical sinuses into the blood (12).
However, we believe the more parsimonious explanation of the
available data are that there is fluid accumulation by both HEVs
and cortical sinuses. Moreover, LN capillaries may contribute to
the formation of interstitial fluid in the LN such that there is some
local passage of fluid from the blood into the cortical sinuses (26).
A typical view of the LN immune surveillance function is that
lymphocytes enter the LN and then survey for antigen or antigenic
peptidesondendriticcells (DCs),a processthatmayrequirehours.
However,theimmunesurveillance function ofLNs might be better
considered in a hierarchical way, where the primary level of sur-
veillance is for LN inflammation and the secondary level is for
antigen or peptides displayed by DCs. With this view, a key re-
quirement for successful immune surveillance is that the inflamed
state of the LN is detected rapidly by entering lymphocytes. We
show that when T and B cells enter an inflamed LN, CD69 is in-
duced within an hour and access to cortical sinuses is strongly
poly(I:C) that occurred as a result of CD69 induction in the lym-
phocyte is similar to the emptying of sinuses that is seen within
and suggests that the presence of lymphocytes within the sinuses is
important for holding them open. This result also suggests that
tethers, to hold them open and implies that the amount of fluid
flow within the structures is influenced by their cellular content.
Recent studies identified a number of paths by which antigens
arriving in LNs can reach B cells. Whereas small antigens (<70
kDa)can gain directaccess tofolliclesvia the SCS floor orthrough
conduits, particulate and opsonized antigens can be captured and
displayed to B cells by SCS macrophages or transported to areas
near follicles by DCs (18, 27, 28). The ability of large protein
antigens to gain rapid access to cortical sinuses, most likely by dif-
fusion in the lymph, suggests a unique mode of B cell encounter
with antigen. Interestingly,although the pathwayis predicted to be
access should be orders of magnitude less efficient (Fig. S2). We
demonstrate that cognate B cells within or adjacent to cortical
sinuses rapidly acquire the large HEL-PE antigen complex. The
the parenchyma, and the suggestion from the modeling analysis
at least once, supports the possibilitythat this is a relevantpathway
of local B-cell antigen encounter. Under inflammatory conditions,
antigen exposure by entry into lymphatic sinuses will be antago-
nized by CD69, but B cells may continue to probe sinuses in an
S1P1-indpendent manner (7), allowing for continued exposure to
sinusoidal antigen. DCs might also acquire large soluble antigens,
perhaps including viral antigens (29), by sampling cortical sinus
fluids. B cells are occasionally observed migrating on the luminal
side of the SCS (8, 30), and consistent with direct exposure of such
cells to SCS lymph, some antigen-labeled cognate B cells were
detected in the SCS within minutes of HEL-PE injection. Thus, all
large soluble antigens during their period of drainage into the LN.
Materials and Methods
either the National Cancer Institute or Jackson Laboratories. C57BL/6 Tg(UBC-
GFP)30Scha/J (004353) were from Jackson Laboratories. MD4 (31), CXCR5−/−
(32), and CD69−/−(33) mice were from internal colonies. [CD69+/+↦ WT] and
[CD69−/−↦ WT] BM chimeras were generated by reconstitution of irradiated
Ly5.2 mice with bone marrow from CD69+/+and CD69−/−mice, as described
(6). Immunizations and treatments with poly(I:C) were performed as de-
scribed in SI Materials and Methods. Animals were housed in a specific
pathogen-free environment in the Laboratory Animal Research Center at
University of California (San Francisco), and all experiments conformed to
ethical principles and guidelines approved by the Institutional Animal Care
and Use Committee.
Cell Isolation and Flow Cytometry. T and B cells were isolated from spleen,
peripheral and mesenteric LNs, and purified as described (7). Purities were
typically >95%. Lymphocyte preparations before adoptive transfer and lymph
node cells postimaging were stained with various fluorochrome-conjugated
Grigorova et al. PNAS
| November 23, 2010
| vol. 107
| no. 47
antibodies from BD Pharmingen as described (7), and data were acquired on
a FACS Calibur (BD) and analyzed with FlowJo software (Treestar).
Reagents. HEL was conjugated to Alexa-Fluor 647 (Molecular Probes labeling
kit) and purified using BioSpin 6 columns (Bio-Rad Laboratories) or to phy-
coerythrin and purified as described (34).
Cell Labeling and Adoptive Transfers. Donor cells were labeled, where indi-
rhodamine (CMTMR) (Invitrogen/ Molecular Probes) or 5 μM of carboxy-
fluorescein diacetate succinimidyl ester (CFSE) (Invitrogen/Molecular Probes),
inDMEMcontaining 1%FBS for 25min at 37°C,andthenwashed byspinning
through alayerofFBS.Labeledorunlabeled cellswere adoptivelytransferred
into the tail vein of recipient mice.
Confocal Microscopy and Intravital Two-Photon Microscopy. Techniques used
were similar to those previously described (7, 30). See SI Materials and
Methods for details.
Development of the quantitative model of T-cell exit from ILNs and
volumetric analysis were performed as described in SI Materials and
Statistical Analysis. All statistical analysis was performed in GraphPad Prism
(GraphPad Software). For comparison of multiple nonparametric datasets we
used the Kruskal–Wallis test followed by Dunn’s posttest comparison be-
tween multiple groups.
ACKNOWLEDGMENTS. We thank T. Nakayama (Chiba University, Japan) for
CD69−/−mice; M. Lipp (The Max-Delbrück Center for Molecular Medicine,
Berlin) for CXCR5−/−mice; P. Beemiller, K. Suzuki, and X. Wang for technical
help; Fred Schaufele for help with confocal microscopy; H. Li for critical
feedback on the mathematical model; and T. Arnon and K. Suzuki for com-
ments on the manuscript. I.L.G. was supported by an Irvington Institute
Fellowship of the Cancer Research Institute and an Immunology program
National Institutes of Health training grant. J.G.C. is an Investigator of the
Howard Hughes Medical Institute. This work was supported in part by Na-
tional Institutes of Health Grants AI45073 and AI74847.
1. Ford WL, Simmonds SJ (1972) The tempo of lymphocyte recirculation from blood to
lymph in the rat. Cell Tissue Kinet 5:175–189.
2. Westermann J, Puskas Z, Pabst R (1988) Blood transit and recirculation kinetics of
lymphocyte subsets in normal rats. Scand J Immunol 28:203–210.
3. Tomura M, et al. (2008) Monitoring cellular movement in vivo with photoconvertible
4. Shiow LR, et al. (2006) CD69 acts downstream of interferon-alpha/beta to inhibit S1P1
and lymphocyte egress from lymphoid organs. Nature 440:540–544.
5. Bankovich AJ, Shiow LR, Cyster JG (2010) CD69 suppresses sphingosine-1-phosophate
receptor-1 function through interaction with membrane helix 4. J Biol Chem 285:
6. Pham TH, Okada T, Matloubian M, Lo CG, Cyster JG (2008) S1P1 receptor signaling
overrides retention mediated by G alpha i-coupled receptors to promote T cell egress.
7. Grigorova IL, et al. (2009) Cortical sinus probing, S1P1-dependent entry and flow-
based capture of egressing T cells. Nat Immunol 10:58–65.
8. Sinha RK, Park C, Hwang IY, Davis MD, Kehrl JH (2009) B lymphocytes exit lymph
nodes through cortical lymphatic sinusoids by a mechanism independent of
sphingosine-1-phosphate-mediated chemotaxis. Immunity 30:434–446.
9. Söderström N, Stenström A (1969) Outflow paths of cells from the lymph node
parenchyma to the efferent lymphatics—observations in thin section histology. Scand
J Haematol 6:186–196.
10. Kelly RH (1975) Functional anatomy of lymph nodes. I. The paracortical cords. Int Arch
Allergy Appl Immunol 48:836–849.
11. Bélisle C, Sainte-Marie G (1981) Tridimensional study of the deep cortex of the rat
lymph node. III. Morphology of the deep cortex units. Anat Rec 199:213–226.
12. Ohtani O, Ohtani Y (2008) Structure and function of rat lymph nodes. Arch Histol
13. He Y (1985) Scanning electron microscope studies of the rat mesenteric lymph node
with special reference to high-endothelial venules and hitherto unknown lymphatic
labyrinth. Arch Histol Jpn 48:1–15.
14. Dahan A, Mendelman A, Amsili S, Ezov N, Hoffman A (2007) The effect of general
anesthesia on the intestinal lymphatic transport of lipophilic drugs: Comparison
between anesthetized and freely moving conscious rat models. Eur J Pharm Sci 32:
15. Yamada S, Kubo M, Hayashida Y (1988) Lymph flow dynamics into the thoracic duct
of the rat. Jpn J Physiol 38:729–733.
16. Gretz JE, Norbury CC, Anderson AO, Proudfoot AE, Shaw S (2000) Lymph-borne
chemokines and other low molecular weight molecules reach high endothelial
venules via specialized conduits while a functional barrier limits access to the
lymphocyte microenvironments in lymph node cortex. J Exp Med 192:1425–1440.
17. Pape KA, Catron DM, Itano AA, Jenkins MK (2007) The humoral immune response is
initiated in lymph nodes by B cells that acquire soluble antigen directly in the follicles.
18. Roozendaal R, et al. (2009) Conduits mediate transport of low-molecular-weight
antigen to lymph node follicles. Immunity 30:264–276.
19. Lämmermann T, Sixt M (2008) The microanatomy of T-cell responses. Immunol Rev
20. Lo CG, Xu Y, Proia RL, Cyster JG (2005) Cyclical modulation of sphingosine-1-
phosphate receptor 1 surface expression during lymphocyte recirculation and
relationship to lymphoid organ transit. J Exp Med 201:291–301.
21. Pham TH, et al. (2010) Lymphatic endothelial cell sphingosine kinase activity is
required for lymphocyte egress and lymphatic patterning. J Exp Med 207:17–27.
22. Beauchemin C, Dixit NM, Perelson AS (2007) Characterizing T cell movement within
lymph nodes in the absence of antigen. J Immunol 178:5505–5512.
23. Anderson AO, Shaw S (1993) T cell adhesion to endothelium: The FRC conduit system
and other anatomic and molecular features which facilitate the adhesion cascade in
lymph node. Semin Immunol 5:271–282.
24. Gretz JE, Anderson AO, Shaw S (1997) Cords, channels, corridors and conduits: Critical
architectural elements facilitating cell interactions in the lymph node cortex. Immunol
25. Sixt M, et al. (2005) The conduit system transports soluble antigens from the afferent
lymph to resident dendritic cells in the T cell area of the lymph node. Immunity 22:
26. Bélisle C, Sainte-Marie G (1990) Blood vascular network of the rat lymph node:
Tridimensional studies by light and scanning electron microscopy. Am J Anat 189:
27. Gonzalez SF, Pitcher LA, Mempel T, Schuerpf F, Carroll MC (2009) B cell acquisition of
antigen in vivo. Curr Opin Immunol 21:251–257.
28. Phan TG, Gray EE, Cyster JG (2009) The microanatomy of B cell activation. Curr Opin
29. Gonzalez SF, et al. (2010) Capture of influenza by medullary dendritic cells via SIGN-
R1 is essential for humoral immunity in draining lymph nodes. Nat Immunol 11:
30. Phan TG, Green JA, Gray EE, Xu Y, Cyster JG (2009) Immune complex relay by
subcapsular sinus macrophages and noncognate B cells drives antibody affinity
maturation. Nat Immunol 10:786–793.
31. Goodnow CC, et al. (1988) Altered immunoglobulin expression and functional
silencing of self-reactive B lymphocytes in transgenic mice. Nature 334:676–682.
32. Förster R, et al. (1996) A putative chemokine receptor, BLR1, directs B cell migration
to defined lymphoid organs and specific anatomic compartments of the spleen. Cell
33. Murata K, et al. (2003) CD69-null mice protected from arthritis induced with anti-type
II collagen antibodies. Int Immunol 15:987–992.
34. Suzuki K, Grigorova I, Phan TG, Kelly LM, Cyster JG (2009) Visualizing B cell capture of
cognate antigen from follicular dendritic cells. J Exp Med 206:1485–1493.
| www.pnas.org/cgi/doi/10.1073/pnas.1009968107Grigorova et al.