Separation of splenic red and white pulp occurs before birth
in a LT??-independent manner
Mark F. R. Vondenhoff,*,1Guillaume E. Desanti,†,1Tom Cupedo,‡Julien Y. Bertrand,§
Ana Cumano,†Georg Kraal,* Reina E. Mebius,* and Rachel Golub†,2
†Unite du De ´veloppement des Lymphocytes, INSERM U668, Institut Pasteur, Paris, France; *Department of
Molecular Cell Biology and Immunology, VU University Medical Center, Amsterdam, The Netherlands;‡Erasmus
Medical Center, Department of Hematology, Rotterdam, The Netherlands; and§University of California, San Diego,
La Jolla, California, USA
Peyer’s patches, lymphoid tissue inducer (LTi)
cells are crucial in triggering stromal cells to re-
cruit and retain hematopoietic cells. Although LTi
cells have been observed in fetal spleen, not much
is known about fetal spleen development and the
role of LTi cells in this process. Here, we show that
LTi cells collect in a periarteriolar manner in fetal
spleen at the periphery of the white pulp anlagen.
Expression of the homeostatic chemokines can be
detected in stromal and endothelial cells, suggest-
ing that LTi cells are attracted by these chemo-
kines. As lymphotoxin (LT)?1?2can be detected
on B cells but not LTi cells in neonatal spleen,
starting at 4 days after birth, the earliest formation
of the white pulp in fetal spleen occurs in a
LT?1?2-independent manner. The postnatal devel-
opment of the splenic white pulp, involving the
influx of T cells, depends on LT?1?2expressed by
B cells. J. Leukoc. Biol. 84: 152–161; 2008.
For the formation of lymph nodes and
Key Words: lymphoid tissue inducer cells ? organogenesis ? embryo
? chemokines ? stroma ? mouse
The spleen is composed of a branching splenic artery that
eventually ends in venous sinuses. The arterial branches,
central arterioles, are surrounded by a layer of lymphoid tissue,
the white pulp. It consists of T cell areas and B cell follicles,
more or less resembling the organization found in lymph nodes
(LN). The arterial blood ends in a sinusoid system in the area
surrounding the T and B cell zones, thereby forming an ana-
tomical border between the white and red pulp, the marginal
zone. Here, sinusoid spaces formed by lining cells continuous
with the endothelium of the arterioles and reticular fibroblasts
make up a network through which the blood freely percolates
on its way to the red pulp and can be scanned for pathogens
and debris by macrophages and dendritic cells (DC). The blood
runs from the marginal zone through the red pulp cords into the
venous sinuses, enabling the spleen to exert its function as a
filter of the blood by removal of effete RBCs [1–5]. For devel-
opment and organization of lymphoid organs, the members of
the TNF superfamily are crucial. The organogenesis of Peyer’s
patches (PP) and LN is dependent on the expression of lym-
photoxin (LT)?1?2and other TNF family members [6, 7].
During fetal and neonatal life, LT?1?2is expressed by
CD45?CD4?IL-7R??CD3–lymphoid tissue inducer (LTi)
cells in PP and LN anlagen [8, 9]. LTi cells are attracted by
nonhematopoietic stromal cells that express the homeostatic
chemokines CCL19, CCL21, and CXCL13, as well as the
adhesion molecules VCAM-1, ICAM-1, and mucosal addressin
cell adhesion molecule-1 (MAdCAM-1) [10–12]. These mole-
cules are thought to be induced by the interaction of LTi and
stromal cells, which leads to LT? receptor (LT?R) and TNFR-I
triggering . Expression of these molecules favors the sub-
sequent recruitment and retention of more LTi cells as well as
other hematopoietic cells .
For the proper development of the murine neonatal lymphoid
part of the spleen, indications that LT?1?2is involved so far
only stem from data obtained after birth, which show that
LT?1?2-expressing B cells are required for the induction of
sufficient CCL21, produced by stromal cells in the T cell zone
areas . Considering the important role of LTi cells in the
organogenesis of other lymphoid organs, relatively little is
known about their role in spleen development, although their
presence has been demonstrated in the fetal spleen as early as
Embryonic Day 13.5 (E13.5) [8, 15]. Transfer of in vitro,
IL-7-activated, splenic LTi cells was shown to restore the B
and T lymphocyte (B/T) segregation in spleens of LT??/?mice
, suggesting a role for LTi cells in white pulp development.
As these experiments may not reflect the actual interactions
and cellular requirements during fetal and neonatal spleen
development, we studied the role of LTi cells in the developing
splenic white pulp in more detail. Our results show that neo-
natal, splenic LTi cells lack LT?1?2expression and that
LT?1?2, required for correct formation of the splenic white
pulp, is expressed by B cells starting approximately 4 days
after birth. Analysis of the earliest events during spleen devel-
opment demonstrates that compartmentalization of red and
1These authors contributed equally to this work.
2Correspondence: Unite du De ´veloppement des Lymphocytes, INSERM
U668, Institut Pasteur, 25, Rue du Dr Roux, 75724 Paris cedex 15, France.
Received September 28, 2007; revised March 13, 2008; accepted March 13,
152 Journal of Leukocyte Biology
Volume 84, July 2008
0741-5400/08/0084-152 © Society for Leukocyte Biology
white pulp areas starts to be regulated during embryogenesis.
By grafting fetal spleens, distinct white pulp areas containing
donor-derived LTi and B cells and red pulp areas harboring
erythrocytes can be observed. Therefore, the white pulp stroma
are already primed before E15.5 to segregate away from the red
pulp areas. Moreover, we suggest that this early phase of
stroma instruction is independent of the LT?R pathway during
murine spleen development.
MATERIALS AND METHODS
C57BL/6 mice were purchased from Harlan (Horst, The Netherlands) and from
Charles River (Saint Germain sur l’Arbresie, France), and LT??/?mice were
purchased from Charles River (Maastricht, The Netherlands). All mouse
strains were bred in the animal facilities of the VU University Medical Center
(VUMC; The Netherlands) or the Institut Pasteur (France) and kept under
routine laboratory conditions. The ethical committees of the VUMC and
Institut Pasteur approved all animal procedures.
Mice were mated overnight, and the day of vaginal-plug detection was marked
as E0.5. Pregnant females were killed at different time-points, and embryos
were harvested and frozen in OCT-embedding medium (Sakura Finetek Europe
BV, The Netherlands) or used for dissection.
Cryosections (7 ?m) were fixed in dehydrated acetone for 2 min and air-dried
for an additional 15 min. Endogenous avidin was blocked with an avidin-biotin
block (Vector Laboratories, Burlingame, CA, USA). Sections were then prein-
cubated in PBS supplemented with 5% (v/v) mouse serum for 10 min. Incu-
bation with primary antibody for 45 min was followed by a 30-min incubation
with Fluor-Alexa-labeled conjugate (Invitrogen Life Technologies, Breda, The
Netherlands) when needed. All incubations were carried out at room temper-
ature. Before embedding in polyvinyl alcohol, sections were counterstained
with Hoechst 33342 (Invitrogen Life Technologies) for 10 min. Stainings were
analyzed on a Leica TCS-SP2-AOBS confocal laser-scanning microscope
(Leica Microsystems Nederland BV, The Netherlands), and images were ob-
tained with Leica confocal software. Image processing involved contrast en-
hancement and region of interest selection, which was carried out with Jasc
Paintshop Pro 7.0. Lenses used were dry lenses: 20? (HC PL APO CS 0.7);
40? (HCX PLAN APO 0.85).
Fetal spleens grafted under the kidney capsule were dissected and incu-
bated overnight in 15% saccharose PBS at 4°C prior to cryosectioning.
Sections were analyzed on a Zeiss Axioplan 2 imaging upright microscope with
a Zeiss Axiocam Hrc camera and Zeiss Axiovision 4.2 software.
Whole-mount fetal spleen immunostaining
E15.5 fetal spleens were incubated overnight in 4% paraformaldehyde in PBS
while agitating at 4°C. Spleens were washed twice in PBS 1?, 10% FCS, 0.1%
Triton X-100 solution while agitating during 90 min and incubated overnight
with RMA4-5 (anti-CD4-FITC) antibody at 4°C under agitation. After washes,
spleens were stained with Hoechst 33342 for 90 min, washed, and then
embedded in Vectashield (Vector Laboratories) and analyzed on a Zeiss
Axioplan 2 imaging upright microscope with a Zeiss Axiocam Hrc camera and
Zeiss Axiovision 4.2 software.
Fetal spleen grafts
The fetal spleen harvested from E14.5 to E16.5 C57BL/6 (CD45.2) embryos
was placed under the kidney capsule of Rag2/?c?/?(CD45.1) mice by surgery.
The Rag2/?c?/?(CD45.1) recipient mice  were anesthetized by peritoneal
injection of 1.4 mg/g ketamine (Merial, France) and 7 ?g/g xylazine (Sigma-
Aldrich Steinheim, Germany) diluted in PBS. Two hours before the anesthesia,
some Rag2/?c?/?(CD45.1) recipient mice were injected with a total C57BL/6
(CD45.2) fetal liver cell suspension. For negative control, sham-operated
recipients were anesthetized, and their kidney capsule was opened without
grafting a fetal spleen.
For flow cytometry and immunofluorescence, the following antibodies were
used: GK1.5 (anti-CD4 [18, 19]), MECA-367 (anti-MAdCAM-1 ), MP33
(anti-CD45, BD Biosciences, Belgium), 6B2 (anti-B220 [21, 22]), 3E2 (anti-
ICAM-1 [23, 24]), and MOMA-2 (a pan macrophage marker) . All of the
antibodies were affinity-purified from hybridoma cell culture supernatants with
protein G-Sepharose (Pharmacia, Uppsala, Sweden) and biotinylated or labeled
with Alexa-Fluor 488, Alexa-Fluor 546, or Alexa-Fluor 633 (Invitrogen Life
Technologies). 429 (anti-VCAM-1, BD Biosciences), A7R34 (anti-IL-7R?,
eBioscience, San Diego, CA, USA), 11D4.1 [anti-vascular endothelial (VE)-
cadherin, BD Biosciences], RMA4-5 (anti-CD4, BD Biosciences), 30-F11
(anti-CD45, BD Biosciences), 104 (anti-CD45.2, BD Biosciences), RA3-6B2
(anti-B220, BD Biosciences), 1D3 (anti-CD19, BD Biosciences), Ter-119 (BD
Biosciences), PK136 (anti-NK1.1, BD Biosciences), RAM34 (anti-CD34, BD
Biosciences), Avas 12?1 [anti-vascular endothelial growth factor receptor
(VEGFR)-2, BD Biosciences], HL3 (anti-CD11c, BD Biosciences), 145-2C11
(anti-CD3ε, BD Biosciences), 7G6 (anti-CD21, BD Biosciences), and A20
(anti-CD45.1, BD Biosciences) anti-VE-cadherin (Alexis Corp., Switzerland) were
used as biotinylated, fluorescently labeled or unconjugated, primary antibodies.
The antibody anti-rat-IgG-tetramethylrhodamine-isothiocyanate (Chemicon Inter-
national, El Segundo, CA, USA) was used as a secondary antibody.
LN and spleens were dissected using a stereomicroscope, and single cell
suspensions were made by digestion with 0.5 mg/ml collagenase type IV
(Sigma-Aldrich) in PBS, 2% FBS, for 30 min at 37°C while constantly stirring.
For surface LT?1?2detection, cells were pretreated with 2.4G2 (anti-CD16/
32), supplemented with 5% normal mouse serum for 30 min, and subsequently
incubated with a LT?R-human IgG fusion protein  for 60 min. Anti-
human-IgG-PE (Jackson ImmunoResearch Laboratories, West Grove, PA,
USA) was used as a second-step conjugate. Splenocytes from adult LT??/?
mice were used as negative controls. Flow cytometric analysis was performed
on a FACSCalibur using CellQuest software (BD Biosciences). Cells were
negatively gated for TCR-?? to exclude all T cells. 7-Aminoactinomycin D
(Molecular Probes, Eugene, OR, USA) was used to exclude dead cells. For
each staining, negative control stainings were carried out, in which LT?R-
human IgG fusion protein was incubated together with anti-LT? mAb (BBF6),
which prevents the binding of LT?R-human IgG (Alexis Benelux, The Neth-
erlands) to cell surface LT?1?2.
For analysis of fetal spleen and fetal spleen grafts, the organs were dissected
using a stereomicroscope, and single cell suspensions were made by dissoci-
ation with a 26-gauge, 3/8-inch needle. Propidium iodide (Sigma-Aldrich) was
used to exclude dead cells. Flow cytometric analyses were performed in an
upgraded LSR (Becton Dickinson, San Jose, CA, USA) using FlowJo software
(Tree Star, Ashland, OR, USA).
RT-PCR, semi-quantitative PCR, and quantitative
Stromal cells from E15.5 fetal spleen were sorted by flow cytometry on a
MOFLO (Dako, Fort Collins, CO, USA), lysed in RLT buffer (Qiagen, Ger-
many), and frozen on dry ice. Total RNA was extracted with an RNeasy micro
kit (Qiagen), according to the manufacturer’s protocol. The cDNA was prepared
from 3 ? 103to 3 ? 106cells using random primers, SuperScript II RT, RNase
OUT in a reaction volume of 24 ?l. PCR reactions (Invitogen Life Technolo-
gies) were performed with Amplitaq Gold (Applied Biosystems, Bridgewater,
NJ, USA) in a reaction volume of 25 ?l.
The primers used were ?-actin, forward: CCGCGAGCACAGCTTCTTT,
?-actin, reverse: CTTTGCACATGCCGGAGC; Vcam-1, forward: GCTCTGG-
GAAGCTGGAACGA, Vcam-1, reverse: TTCATGAGCTGGTCACCCTTGA.
cDNA was diluted at 1/10 and 1/100, and results from the 1/100 dilution were
used to measure intensity of the band using the ImageJ 1.38 application.
Intensity was normalized using the ?-actin transcript levels (100%).
The cDNA samples used for quantitative real-time PCR were obtained by
using RT2PCR Array First Strand kit (SuperArray Bioscience Corp., Freder-
ick, MD, USA) for the RT step. The RT2Profiler PCR Array “Mouse Chemo-
Vondenhoff et al.
kines and Receptors” (SuperArray Bioscience Corp.) and the RT2Real-Time
SYBR Green/6-carboxy-X-rhodamine PCR Master Mix (SuperArray Bioscience
Corp.) were used for real-time PCR quantification. The real-time PCR was per-
formed on a 7300 Real-Time PCR system (Applied Biosystems). The products of
each reaction were checked by 2% agarose gel electrophoresis migration.
CD4?IL-7R??CD3ε–cells from 3-day-old mice intestine previously dis-
sociated by dispase (Gibco-BRL, Grand Island, NY, USA) treatment and
CD4hiLin–(Lin: Gr-1, B220, Ter119, CD11c, CD19, NK1.1, CD3ε) cells from
E15.5 fetal spleen were sorted by flow cytometry on a MOFLO (Dako), lysed in
TRIzol (Gibco-BRL), and frozen at –20°C. Total RNA was extracted according
to the TRIzol (Gibco-BRL) manufacturer’s protocol. Oligo (dT)-primed cDNA
was prepared using SuperScript II RT (Invitogen Life Technologies) in a
reaction volume of 20 ?l. PCR reactions were performed with Amplitaq Gold
(Applied Biosystems) in a reaction volume of 25 ?l. The primers used were
retinoid-related orphan receptor (ROR)?/?t, forward: GCCTCCTGAGAGCCT-
CAGG, ROR?/?t, reverse: CACCTCCTCCCGTGAAAG.
Characterization of the developing spleen
Human fetal spleen development was recently characterized,
and already at the 2nd stage of development when the first
lymphocytes start to colonize the organ, distinct areas could be
observed where the homeostatic chemokine CXCL13 was ex-
pressed. The presence of CXCL13 in arterial smooth muscle
cells as well as cells around these arterioles  suggested the
interaction of these stromal cells with LTi cells in analogy with
the developing LN and a role in the further development of the
spleen. To further define the earliest events in spleen devel-
opment, we addressed whether these distinct areas could also
be found in murine spleen. Here, fetal spleens from E18.5 were
stained for endothelial, hematopoietic, as well as stromal mark-
ers. Combined staining of tyrosine kinase with Ig and epider-
mal growth factor homology domain 2 (Tie-2/Tek), a marker for
endothelial cells , and CD45 revealed the presence of an
artery with a contracted lumen and a vein with an open lumen
in between hematopoietic cells at this time-point in develop-
ment (Fig. 1, A and D). CD45?hematopoietic cells were
remotely located in a ring-like pattern around the artery (Fig.
1B), and the venous blood vessel was immediately surrounded
by hematopoietic cells (Fig. 1C). To define the expression of
MAdCAM-1, a marker for LN stromal organizer cells  as
well as high endothelial venules , fetal spleens were
stained for MAdCAM-1 in combination with the endothelial
marker VE-cadherin . The results showed that most VE-
cadherin? endothelial cells expressed MAdCAM-1 at high
levels (Fig. 1E and Supplemental Fig. 1, A–C). Stainings of
subsequent sections showed that endothelial cells of large
vessels also expressed ICAM-1, and VCAM-1 was absent or
expressed at low levels (Fig. 1F). In addition, around the artery,
VCAM-1?cells were found. As LTi cells were reported by us
to be present at early stages in developing spleen , we
performed additional stainings for IL-7R? and CD4 in subse-
quent sections to detect these cells. Interestingly, CD4?IL-
Fig. 1. Murine fetal spleens at developmental stage E18.5. Immunofluorescence stainings (A–C and E–G) and schematic drawings (D and H). (E–G) Serial
sections. (A) Murine fetal spleen consisting of Tie-2?endothelial cells and CD45?hematopoietic cells (Tie-2 in green; CD45 in red). The outline of the fetal spleen
is indicated in white. A major vein (A and C, arrowheads) and artery (A and B, double arrowheads) can be distinguished at this time-point in development. (D)
Schematic drawing of fetal spleen showing outline and position of vein and artery. (E) High magnification image of blood vessel endothelial cells expressing
MAdCAM-1 (in green) and VE-cadherin (in red). VE-cadherin?blood vessels with the highest MAdCAM-1 expression (yellow) are marked by arrows. Vein
(arrowhead) and artery (double arrowhead) are MAdCAM-1?VE-cadherin?(orange). (F) Expression of ICAM-1 (in red) and VCAM-1 (in green) by endothelial cells
and splenic stromal cells. ICAM-1?VCAM-1?arterial endothelial cells (orange) are immediately surrounded by VCAM-1?stromal cells (double arrowhead, green).
The outline of these cells is indicated in white. Arrowhead indicates a vein. (G) LTi cells expressing IL-7R? (in green) and CD4 (in red) positioned distal to the
artery and VCAM-1?stromal cells, forming a ring-like structure (double arrowhead) that is not found around the vein (arrowhead). (H) Schematic drawing of the
outlines of arterial endothelial cells (i), VCAM-1 ? ICAM-1- stromal cells (ii), and LTi cells forming a ring-like structure (iii). Outward from this structure, blood
vessels with the highest MAdCAM-1 expression are found. Overview image of fetal spleens is shown at 20? original magnification in A. Detailed images of fetal
spleens are shown at 40? original magnification, Zoom 2, in B, C, and E–G. Data are representative of four individual mice.
154 Journal of Leukocyte Biology
Volume 84, July 2008
7R??LTi cells were located in a ring-like pattern at the
periphery of the VCAM-1?stromal cells, adjacent to MAd-
CAM-1hiVE-cadherin?endothelial cells (Fig. 1, E–G). The
distinct locations of these cell types are depicted in a sche-
matic drawing (Fig. 1H). To confirm that CD4?IL-7R??cells
were indeed LTi cells, we performed triple stainings for ROR?,
which is indispensable for LTi differentiation , IL-7R?,
and CD4. These stainings showed that all CD4?IL-7R??
contained ROR?, indicating that these cells were indeed LTi
cells (Supplemental Fig. 1, D–H and Supplemental Fig. 2).
Few ROR??IL-7R??CD4?cells were found at developmental
stage E18.5 (data not shown).
To see whether a distinct pattern of LTi cells in the prox-
imity of vessels could also be observed at earlier stages of
spleen development, serial sections of E14.5 and E16.5
spleens were analyzed. Therefore, we investigated the vascu-
lature of the fetal spleen in more depth by staining sections of
E16.5 and E14.5 fetal spleens for VE-cadherin combined with
4?,6-diamidino-2-phenylindole (DAPI) and MECA-32 staining.
The latter antibody has been described as an endothelial
marker . The combination of both endothelial markers
allowed us to focus on the artery, which can be distinguished
by expression of VE-cadherin and lack of MECA32 (VE-
cadherin?MECA-32–). At E16.5 in development, we could
identify LTi cells around this VE-cadherin?MECA-32–artery,
again organized in a ring-like pattern. In addition, we found
single LTi cells and small, poorly organized clusters in other
areas of the fetal spleen (Fig. 2, A–D). At E16.5, most of the
vessels were MAdCAM-1?(Supplementary Fig. 1, I–K), sim-
ilar to E18.5 splenic vessels. In addition, VCAM-1?stromal
cells could be distinguished that were organized around the
large VE-cadherin?MECA-32–artery (Fig. 2, B and C). Stro-
mal cells and LTi cells appeared to be intermingled, instead of
strictly separated, as seen in E18.5 fetal spleens. Although at
E14.5, LTi cells could be detected in the fetal spleen in the
vicinity of blood vessels, they were not yet organized in a
ring-like structure, as observed at E16.5 and E18.5 (Fig. 2,
E–H). Strikingly, we could only identify one major blood vessel
in the center of the spleen at E14.5. The lumen of this vessel
was large, and apart from the expression of VE-cadherin, it was
weakly stained for MECA-32 (Fig. 2F). In addition, at this
time-point in development, numerous smaller vessels ex-
pressed MAdCAM-1 and ICAM-1, but VCAM-1 expression in
the fetal spleen was limited at this developmental time-point
(Fig. 2G and Supplementary Fig. 1, L–O). VCAM-1 expression
could be observed in deeper layers of the stomach wall, indi-
Fig. 2. Murine fetal spleens at stages E14.5, E15.5, and E16.5 of development.
Immunofluorescence staining of serial sections from E16.5 (A–D) and E14.5
(E–H) spleens. (A) Double staining for MECA-32 (in green) and VE-cadherin (in
red) counterstained with DAPI (in blue) to visualize nuclei. Most of the smaller
vascular blood vessels appeared to express MECA-32 and VE-cadherin. (B)
Higher magnification image showing a MECA-32?vein that is VE-cadherinlo
(arrowhead) and a VE-cadherin?artery that lacks expression of MECA-32 (dou-
ble arrowhead). (C) ICAM-1?VCAM-1?artery (double arrowhead) surrounded by
VCAM-1?stromal cells (ICAM-1 in red; VCAM-1 in green). (D) IL-
7R??CD4?CD45loLTi cells are organized around arterial endothelial cells
(double arrowhead) (CD45 in green; IL-7R? in red; CD4 in blue). (E) E14.5
spleen (arrow) and stomach (arrowhead) containing MECA-32?VE-cadherin?
blood vessels (MECA-32 in green; VE-cadherin in red) counterstained with DAPI
(in blue). (F) At E14.5, all blood vessels (orange) are double-positive for
MECA-32 and VE-cadherin including a larger blood vessel. The outline of the
larger vessel is indicated in white. (G) At this time-point, most endothelial cells
express MAdCAM-1 (in blue) and ICAM-1 (in red), and VCAM-1 expression is virtually absent in the fetal spleen (in green). (H) In the near vicinity of the larger
vessel, CD45?hematopoietic cells are localized, including IL-7R??CD4?CD45loLTi cells (CD45 in green; IL-7R? in red; CD4 in blue). Images showing
sagital-sectioned fetal spleen (E16.5) or transverse-sectioned fetal spleen and stomach (E14.5) at 20? original magnification. Higher magnification images of fetal
spleens are shown at 40? original magnificaiton and Zoom 2 in B–D and F–H. The outline of major blood vessels is marked in white. (I) On E15.5 fetal spleen,
whole-mount immunofluorescent staining was performed to locate CD4?cells. Data are representative of four (E14.5) or five (E15.5, E16.5) individual mice.
Vondenhoff et al.
Spleen development 155
cating that it was not a result of a detection failure (Supple-
mentary Fig. 1, L and M). To see whether LTi cells indeed
preferentially distribute to certain splenic areas, whole-mount
E15.5 spleens were stained for CD4. These stainings showed
that CD4 cells were clustered and that these clusters were
distributed linearly along the length of the entire spleen (Fig.
2I), indicating that the early compartmentalization of the fetal
spleen may attract LTi cells to periarteriolar domains. In
addition, a limited number of single CD4?cells were found in
the spleen at other areas.
Distinct stromal subsets are present in the
As LTi cells are localized within special areas of the spleen, we
assumed that these cells are colocalizing with a specific stro-
mal cell subset and that various stromal cell populations could
be expected to exist in fetal spleens. We therefore analyzed
stromal cells present in E15.5 fetal spleen by FACS (Fig. 3A).
After elimination of leukocytes (CD45?) and erythrocytes
(Ter119?), the E15.5 fetal spleen stroma could be separated
into two populations, based on the expression of VE-cadherin.
As can be inferred from immunofluorescent stainings (Figs. 1
and 2), splenic stromal cells that surround arterial vessels lack
endothelial markers such as VE-cadherin. In addition, endo-
thelial cells express VE-cadherin and are CD45–, determined
by FACS analysis (Fig. 3, A and B). To confirm that VE-
cadherin?cells indeed represent endothelial cells, expression
of other endothelial markers was analyzed. Although 17% of
the CD45–Ter119–cells could be distinguished as a population
lial cells, the remaining 82% of fetal spleen CD45–Ter119–
cells was characterized as CD34–ICAM1–/loVEGFR2–VE-cad-
herin–Tie-2–stromal cells (Fig. 3, A and B). By semiquantita-
tive RT-PCR, we found VCAM-1 transcripts in both subsets
with a probable lowest quantity in the CD34–stromal subset
(Fig. 3C). As LTi cells were found to be associated with
endothelial cells in the spleen, we addressed whether splenic
CD45–cells could produce the chemokines that are known to
attract LTi cells. Quantities of CXCL12, CXCL13, CCL19, and
CCL21b transcripts were determined in CD34–stromal and
CD34?endothelial subsets of CD45–Ter119–cells (Fig. 3D).
The expression levels of CXCL12 and CXCL13 were relatively
similar in both CD45–Ter119–subsets. In contrast, levels of
CCL19 and CCL21b transcript expression were 100- and
1000-fold higher in the CD34?endothelial fraction than in the
CD34–stromal fraction, respectively.
White pulp anlagen is already primed
in fetal spleen
The expression of homeostatic chemokines by CD45–Ter119–
cells and the preferential localization of LTi cells to VE-
cadherin?endothelial cells at the periphery of the MAd-
CAM-1?stromal areas raised the question of whether these
periarteriolar areas constituted the anlagen of the white pulp.
To address whether these designated areas could mature into
splenic white pulp areas, spleens from E15.5 C57BL6 embryos
were grafted under the kidney capsule of adult Ly5.1 Rag2/
?c?/?immunodeficient mice (Fig. 4A). The use of Rag2/
?c?/?alymphoid mice was essential to avoid substantial con-
tamination by host-derived lymphocytes. B cell precursors are
present in fetal spleen at the time of transplantation, and they
are able to differentiate into mature CD19?B220?B cells in
situ. When the grafts were analyzed at 3 weeks after transplan-
tation, they showed good vascularization and had increased in
size, showing that the hematopoietic progenitors present in
E15.5 fetal spleens were able to proliferate and differentiate in
situ. Grafts performed with E14.5 spleen also showed prolif-
eration and differentiation of CD45?progenitors within the
grafted spleens (data not shown).
Analysis by immunofluorescence showed that B cells and
LTi cells colocalized in an area that did not contain Ter119?
erythrocytes, indicative of a B/LTi-containing white pulp area
and a Ter-119-containing red pulp area (Fig. 4A). In Figure
4A, only LTi cells are stained to clearly observe the structure
of the spleen (LTi clusters and vein). Development of these
areas within the grafted spleens indicated that the stroma had
been instructed to retain LTi cells to restricted areas that are
common to B cells. It appeared that the designation of the white
pulp anlagen, containing LTi cells, had already taken place at
E15.5, independently of mature, circulating lymphocytes.
In parallel, splenic grafts were analyzed by FACS to assess
the presence of donor-derived LTi, B, NK, and T cells (Fig.
4B). Depending on the graft, the proportion of hematopoietic
donor cells (CD45.2?) varied, whereas the percent of host
(CD45.1?) hematopoietic cells, containing myeloid and
CD11c?DC, was constant (approximately 10%; Fig. 4B).
Within the pool of donor-derived hematopoietic cells, clear
populations of LTi (CD4hiCD3–IL7R??) and B (CD19?B220?)
cells were seen, whereas some of these B cells expressed high
levels of CD21, potentially representing marginal Zone B cells
(Fig. 4B). Few graft-derived NK cells could be observed (Fig. 4B).
Injection of E15.5 total fetal liver cells from C57BL/6 em-
bryos during splenic transplant permits full reconstitution of
the hematopoietic compartment of the alymphoid recipients. In
these cases, the B and T cells and erythrocytes were clearly
distributed into organized white and red pulp areas. The white
pulp areas showed normal B/T segregation, indicating that T
and B cells could normally enter the white pulp areas (Fig. 4A).
LT?1?2expression during spleen development
LTi cells in developing LN are instrumental for the induction
of homeostatic chemokines in developing LN through their
expression of LT?1?2. According to the literature, the devel-
opment of the spleen requires LT?1?2-expressing cells postna-
tally [14, 35, 36], whereas we showed here that homeostatic
chemokines could already be observed before birth. Therefore, we
addressed at what time-point the earliest LT?1?2-expressing cells
could be observed using FACS analysis of postnatal spleens.
Mesenteric LN (MLN) from the same animals were included as
positive controls. Analysis of Days 0, 2, and 4 and adult wild-type
spleens revealed a gradual increase of LT?1?2on splenic B cells
(Fig. 5A). At Days 0 and 2, only 1% of all B cells expressed
LT?1?2, and this increased to 2.7% at Day 4. In adult spleens,
12.5% of total B cells expressed LT?1?2(Fig. 5B).
In contrast, few LT?1?2-expressing LTi cells could be
detected in neonatal spleens, and LTi cells from MLN
expressed high levels of LT?1?2, in accordance with our
156Journal of Leukocyte Biology
Volume 84, July 2008
earlier report (Fig. 5C) . When compensated for a spe-
cific binding (Supplementary Fig. 3), the percentage of
2.2% of all LTi cells, implicating that LT?1?2
are rare (Fig. 5D).
?LTi cells in the spleen varies between 1.5% and
LT?1?2-dependent white pulp development
As we observed few LT?1?2-expressing cells at Day 2 after
birth, we reasoned that before Day 2, splenic development,
including the formation of the white pulp anlagen, occurred
independently of LT?1?2-mediated LT?R triggering. To con-
firm this, spleens from LT??/?mice taken at Days 2, 4, and 6
after birth were analyzed. In wild-type and LT??/?mice, small
accumulations of B220?B cells could be found at 2 days after
birth around arterioles in the spleen, indicative of undisturbed
formation of white pulp areas. In addition, few CD3?T cells
were seen at this time-point in or near B cell areas in wild-type
and LT??/?animals (Fig. 6A). The effect of the absence of
Fig. 3. The stromal compartment of E15.5 fetal
spleen. (A) At E15.5, the fetal spleen (FS) stroma
were isolated as CD45–propidium iodide (PI)–
Ter119–cells, and two subsets were distinguished
based on CD34 and ICAM-1 expression. Most cells
were CD34–ICAM1–/lo, and a small subset was
formed by CD34?ICAM1?cells. FSC, Forward-
scatter; SSC, side-scatter. (B) These subsets of
stromal cells were analyzed for the surface expres-
sion of VEGFR2, VE-cadherin, and Tie-2. (C)
Semiquantitative RT-PCR analysis of the Vcam1
gene in the E15.5 fetal spleen populations (dilution
with a factor 10). Intensity was quantified to mea-
sure the difference of Vcam1 expression normal-
ized by ?-actin levels. (D) Quantitative real-time
PCR was performed to study the relative expression
of Cxcl12, Cxcl13, Ccl19, and Ccl21b chemokines
from the CD34–ICAM1–(blue histograms) and CD34?ICAM1?(orange histograms) stromal populations. The average of Gadph and hypoxanthine guanine
phosphoribosyl transferase housekeeping gene levels was used for the normalization. Results are obtained from two independent cell sorting, and each sorted
population is represented by a specific histogram.
Vondenhoff et al.
Spleen development 157
LT?1?2-expressing cells could be observed around Day 4,
when B cell areas were still small in LT??/?, and they had
clearly increased in size in wild-type mice from Days 2 to 4.
In addition, T cell areas were completely absent in LT??/?
spleens at all stages analyzed, and these were clearly
present in wild-type spleen at Day 4 and had further en-
larged at Day 6 (Fig. 6A). Despite the dramatic differences for
the B and T cell populations, we could not find a clear difference
with respect to MOMA-2?macrophages. In fact, the amount of
MOMA-2?cells that were found in close proximity to central
arterioles appeared even higher in LT??/?spleens when com-
pared with wild-type spleens at Day 4 (Fig. 6A, and data not
An additional difference was observed when splenic white
pulp stromal cell populations of wild-type mice were compared
with LT??/?mice. When staining for VE-cadherin, VCAM-1,
and ICAM-1, VE-cadherin-negative stromal cell populations
could be analyzed. VCAM-1?ICAM-1locells were located
directly around the central arteriole. At the periphery of this
white pulp area, VCAM-1?ICAM-1?cells could be seen (Sup-
plemental Fig. 4). When wild-type and LT??/?spleens were
compared, these cellular subsets could be distinguished at Day
2, but the increase in VCAM-1?ICAM-1?at the periphery of
the white pulp area seen at Days 4 and 6 in wild-type was not
observed in LT??/?mice (Fig. 6B). These results indicate a
cation where the marginal zone will develop and is in agree-
ment with earlier observations that LT?1?2/LT?R interactions
are mandatory for homeostasis and proper functioning of mar-
ginal zones in adult spleens [16, 36–39].
?-mediated increase of VCAM-1?ICAM-1?at the lo-
Fig. 4. Early priming of the fetal spleen stroma. LTi cells maintain in fetal spleen graft and colocalize with
B cells. (A) Three weeks after engraftment of CD45.2?fetal spleens under the kidney capsule of CD45.1?
Rag2/?c?/?mice, grafts were analyzed by immunohistology for LTi cells (CD4 in green), B cells (B220 in
white), and erythrocytes (Ter119 in red; FS Graft). The first quadrant represents LTi cells forming ring-like
structures (white ovals) with a vein (arrowhead) devoid of LTi cells. The second quadrant displays LTi cells
and B cells in close contact to each other. The same experiment, supplemented with injection of CD45.2?
fetal liver cells (FL), was used as a positive control for splenic architecture, and most CD4?cells (green) are
T cells (FS Graft?FL Injection). Original blue bar, 200 ?m. (B) Grafts were analyzed for their composition.
Cells were separated between graft-derived (CD45.2?) and host-derived (CD45.1?) and analyzed by flow
cytometry for the presence of LTi (CD4?CD3ε–IL7R??), B (CD19?B220?and B220?CD21?), NK
(NK1.1?CD3ε–), and T (CD3ε?) cells and DC (CD11c?). Data are representative of analysis of three
individual grafted mice.
158 Journal of Leukocyte Biology
Volume 84, July 2008
The potential role of LTi cells in white pulp ontogeny has been
a matter of debate, as these cells express the ligand for the
LT?R in developing LN, and signaling through this receptor is
indispensable for normal white pulp development [37, 40].
Although it has been shown that triggering of the LT?R by
LT?1?2-expressing B cells is crucial for white pulp develop-
ment, the possibility remained that triggering of the LT?R by
other cells, such as activated T cells or LTi cells, further
facilitated this process. Here, we show for the first time that the
majority of splenic LTi cells lacks cell surface expression of
the LT?R ligand. It has been shown that fetal and neonatal
splenic LTi cells express the LT? and LT? transcripts [9, 41].
The adult spleen environment, but not the fetal, has also been
proposed to regulate the LT? transcript expression in LTi cells
by TL1A . Our results suggest that the neonatal, splenic
environment does not lead to cell surface expression of LT?R
ligands on LTi cells. An in vitro, dose-dependent IL-7 stimu-
lation is required for an effective LT?R ligand expression by
intestinal LTi cells . We previously detected Il-7 transcript
expression by fetal spleen nonhematopoietic cells . At Day
4 after birth, the time-point that T cell areas start to emerge, a
ligand for the LT?R, most likely LT?1?2, is only expressed by
B cells. Thus, for white pulp development, B cells will give the
LT?1?2-dependent, inductive signal, similar as LTi cells do
during LN and PP formation, and this is in accordance with
earlier observations . This signal is required to attract more
B and T cells to the developing white pulp. In contrast to the
white pulp development into functional T and B cell areas, the
separation of white and red pulp areas occurs before birth and
is independent of LT?-expressing cells
In E15.5 fetal spleen, patches of LTi cells are found
throughout the spleen with a distribution reminiscent of white
pulp areas of adult spleen. Although our results indicate that
LTi cells do not directly contribute to white pulp development,
the localization of these cells, their close association with blood
vessels, and their probable interaction with VCAM-1?ICAM-
1–/lostromal cells at E18.5 are intriguing but might just be a
reflection of the chemokines and adhesion molecules that are
The Cxcl12 and Cxcl13 transcripts were similarly expressed
within the “stromal” and “endothelial” stromal cell subset. On
the contrary, the CD34?ICAM?endothelial stromal cell subset
expressed higher levels of Ccl19 and Ccl21b transcripts than
the stromal subset. Assuming that these relative transcript
expressions are revealing the protein expression levels, we
propose that the perivascular distribution of LTi cells could be
mediated by the high expression of CCL19 and CCL21 che-
mokines from the endothelial cells. Indeed, the lower expres-
sion of CCL19 and CCL21 by the stromal cells would establish
CCL19 and CCL21 concentration gradients, decreasing from
the vascular endothelial cells to the inner part of the fetal
spleen. The CXCL12 and CXCL13 chemokine expressions by
the splenic stroma probably also participate in the attraction
and retention of LTi cells into the fetal spleen. Early expres-
sion of CXCL13, before the entry of B cells, has been observed
around arterioles in developing human spleen . Thus, expres-
Fig. 5. LT?R-binding cells in neonatal spleen and MLN. LT?R-human IgG
(huIgG) binding to B cells (A and B) or LTi cells (C and D) was analyzed by
flow cytometry at indicated days after birth, and four mice per group were
Vondenhoff et al.
Spleen development 159
sion of these homeostatic chemokines most likely contributes to
the periarteriolar localization of LTi cells in the fetal spleen.
Our graft experiments show LTi and B cells colocalizing into
well-defined areas within fetal spleen, illustrating the early spe-
cific capacities of the spleen stromal cells. In the grafts, all B and
LTi cells are donor-derived and have differentiated in situ. Both
spleen stromal populations expressed the transcripts for CXCL13
and CXCL12, important for LTi cell interaction and B cell attrac-
tion and differentiation [11, 43, 44]. Moreover, we observed a
small host-derived CD11c?DC population in these grafts. It was
shown that CD11c?cells can be found in a neonatal spleen, only
a few days after birth [45, 46]. Hence, DC have migrated into the
grafted spleen in response to the chemokines produced.
In our grafts, a typical adult spleen organization was ob-
served when total fetal liver cells were injected previously. The
B/T lymphocytes that differentiated from these fetal liver cells
were attracted by the fetal spleen graft into their specific
domains, reinforcing the capacity of the white pulp anlagen to
express chemokines essential for the development of its archi-
tecture. It was shown that CCL21 could be expressed by two
different genes after an activation of a LT?1?2-dependent or
-independent pathway, whereas the two chemokines conserve
the same chemotactic activity . Most of CCL21 expression
depends on LT?1?2in the adult spleen but not in nonlymphoid
tissue . It is possible that the first chemokine expression
found in the fetal spleen at E15.5 is triggered in a similar
Finally, our data show that the fetal spleen stroma are
instructed at an early time-point to segregate lymphoid and
nonlymphoid zones, which correlate with red and white pulp areas
of the adult organ. The graft experiments and the low numbers of
B or LTi cells in the fetal spleen that express the LT?R suggest
that these first driving events are LT?R-independent.
This work was supported by fellowships from Ministe `re Fran-
c ¸ais de la Recherche et de l’Enseignement Supe ´rieur, from
Association pour la recherche sur le cancer (ARC), from ANR,
and from the European Stem Cell Research Program to
G. E. D., a Vici grant from the Netherlands Organization for
Scientific Research (918.56.612) to R. E. M., and a genomics
grant from the Netherlands Organization for Scientific Research
(050-10-120) to M. F. R. V. The authors declare that there is no
conflict of financial interest. We acknowledge A. Louise for cell
sorting and the Flow Cytometry Core Facility (Institut Pasteur,
France) as well as E. Perret for her advice and the Dynamic
Imagery Core Facility (Institut Pasteur).
1. Veerman, A. J., van Ewijk, W. (1975) White pulp compartments in the spleen
of rats and mice. A light and electron microscopic study of lymphoid and
non-lymphoid cell types in T- and B-areas. Cell Tissue Res. 156, 417–441.
2. Sasou, S., Satodate, R., Katsura, S. (1976) The marginal sinus in the
perifollicular region of the rat spleen. Cell Tissue Res. 172, 195–203.
3. Kupiec-Weglinski, J. W., Austyn, J. M., Morris, P. J. (1988) Migration
patterns of dendritic cells in the mouse. Traffic from the blood, and T
cell-dependent and -independent entry to lymphoid tissues. J. Exp. Med.
4. Schmidt, E. E., MacDonald, I. C., Groom, A. C. (1993) Comparative
aspects of splenic microcirculatory pathways in mammals: the region
bordering the white pulp. Scanning Microsc. 7, 613–628.
5. Mebius, R. E., Kraal, G. (2005) Structure and function of the spleen. Nat.
Rev. Immunol. 5, 606–616.
6. De Togni, P., Goellner, J., Ruddle, N. H., Streeter, P. R., Fick, A.,
Mariathasan, S., Smith, S. C., Carlson, R., Shornick, L. P., Strauss-
Schoenberger, J., et al. (1994) Abnormal development of peripheral lym-
phoid organs in mice deficient in lymphotoxin. Science 264, 703–707.
7. Korner, H., Cook, M., Riminton, D. S., Lemckert, F. A., Hoek, R. M.,
Ledermann, B., Kontgen, F., Fazekas de St Groth, B., Sedgwick, J. D.
(1997) Distinct roles for lymphotoxin-? and tumor necrosis factor in
organogenesis and spatial organization of lymphoid tissue. Eur. J. Immu-
nol. 27, 2600–2609.
8. Mebius, R. E., Rennert, P., Weissman, I. L. (1997) Developing lymph
nodes collect CD4?CD3- LTb? cells that can differentiate to APC, NK
cells, and follicular cells but not T or B cells. Immunity 7, 493–504.
9. Yoshida, H., Honda, K., Shinkura, R., Adachi, S., Nishikawa, S., Maki, K.,
Ikuta, K., Nishikawa, S. I. (1999) IL-7 receptor ?? CD3(–) cells in the
embryonic intestine induces the organizing center of Peyer’s patches. Int.
Immunol. 11, 643–655.
10. Honda, K., Nakano, H., Yoshida, H., Nishikawa, S., Rennert, P., Ikuta, K.,
Tamechika, M., Yamaguchi, K., Fukumoto, T., Chiba, T., Nishikawa, S. I.
(2001) Molecular basis for hematopoietic/mesenchymal interaction during
initiation of Peyer’s patch organogenesis. J. Exp. Med. 193, 621–630.
11. Finke, D., Acha-Orbea, H., Mattis, A., Lipp, M., Kraehenbuhl, J. (2002)
CD4?CD3– cells induce Peyer’s patch development: role of ?4?1 inte-
grin activation by CXCR5. Immunity 17, 363–373.
Fig. 6. Early stages of white pulp development occur in the absence of LT?.
Immunofluorescence staining of 2-, 4-, and 6-day-old spleens from C57BL/6
(B6) and LT??/?mice was done using (A) anti-B220, anti-CD3, and anti-
MOMA-2 antibodies to detect, respectively, B cells, T cells, and macrophages
and (B) anti-VE-cadherin, anti-ICAM-1, and anti-VCAM-1 antibodies to de-
tect stromal and endothelial cell subsets. Data are representative of three
individual mice per group.
160 Journal of Leukocyte Biology
Volume 84, July 2008
12. Luther, S. A., Ansel, K. M., Cyster, J. G. (2003) Overlapping roles of Download full-text
CXCL13, interleukin 7 receptor ?, and CCR7 ligands in lymph node
development. J. Exp. Med. 197, 1191–1198.
13. Mebius, R. E. (2003) Organogenesis of lymphoid tissues. Nat. Rev. Im-
munol. 3, 292–303.
14. Ngo, V. N., Cornall, R. J., Cyster, J. G. (2001) Splenic T zone development
is B cell dependent. J. Exp. Med. 194, 1649–1660.
15. Desanti, G. E., Bertrand, J. Y., Golub, R. (2007) Fetal spleen develop-
ment, the ride toward multiple functions. Functional Development and
Embryology 1, 78–90.
16. Kim, M. Y., McConnell, F. M., Gaspal, F. M., White, A., Glanville, S. H.,
Bekiaris, V., Walker, L. S., Caamano, J., Jenkinson, E., Anderson, G.,
Lane, P. J. (2007) Function of CD4?CD3– cells in relation to B- and
T-zone stroma in spleen. Blood 109, 1602–1610.
17. Colucci, F., Turner, M., Schweighoffer, E., Guy-Grand, D., Di Bartolo, V.,
Salcedo, M., Tybulewicz, V. L., Di Santo, J. P. (1999) Redundant role of
the Syk protein tyrosine kinase in mouse NK cell differentiation. J. Im-
munol. 163, 1769–1774.
18. Wilde, D. B., Marrack, P., Kappler, J., Dialynas, D. P., Fitch, F. W. (1983)
Evidence implicating L3T4 in class II MHC antigen reactivity; monoclonal
antibody GK1.5 (anti-L3T4a) blocks class II MHC antigen-specific pro-
liferation, release of lymphokines, and binding by cloned murine helper T
lymphocyte lines. J. Immunol. 131, 2178–2183.
19. Dialynas, D. P., Wilde, D. B., Marrack, P., Pierres, A., Wall, K. A.,
Havran, W., Otten, G., Loken, M. R., Pierres, M., Kappler, J., et al. (1983)
Characterization of the murine antigenic determinant, designated L3T4a,
recognized by monoclonal antibody GK1.5: expression of L3T4a by func-
tional T cell clones appears to correlate primarily with class II MHC
antigen-reactivity. Immunol. Rev. 74, 29–56.
20. Streeter, P. R., Berg, E. L., Rouse, B. T., Bargatze, R. F., Butcher, E. C.
(1988) A tissue-specific endothelial cell molecule involved in lymphocyte
homing. Nature 331, 41–46.
21. Coffman, R. L., Weissman, I. L. (1981) B220: a B cell-specific member of
the T200 glycoprotein family. Nature 289, 681–683.
22. Coffman, R. L., Weissman, I. L. (1981) A monoclonal antibody that
recognizes B cells and B cell precursors in mice. J. Exp. Med. 153,
23. Takei, F. (1985) Inhibition of mixed lymphocyte response by a rat mono-
clonal antibody to a novel murine lymphocyte activation antigen (MALA-
2). J. Immunol. 134, 1403–1407.
24. Prieto, J., Takei, F., Gendelman, R., Christenson, B., Biberfeld, P., Pa-
tarroyo, M. (1989) MALA-2, mouse homologue of human adhesion mole-
cule ICAM-1 (CD54). Eur. J. Immunol. 19, 1551–1557.
25. Kraal, G., Rep, M., Janse, M. (1987) Macrophages in T and B cell
compartments and other tissue macrophages recognized by monoclonal
antibody MOMA-2. An immunohistochemical study. Scand. J. Immunol.
26. Cupedo, T., Lund, F. E., Ngo, V. N., Randall, T. D., Jansen, W., Greuter,
M. J., de Waal-Malefyt, R., Kraal, G., Cyster, J. G., Mebius, R. E. (2004)
Initiation of cellular organization in lymph nodes is regulated by non-B
cell-derived signals and is not dependent on CXC chemokine ligand 13.
J. Immunol. 173, 4889–4896.
development of the human spleen: from primordial arterial B cell lobules to a
non-segmented organ. Histochem. Cell Biol. 128, 205–215.
28. Dumont, D. J., Yamaguchi, T. P., Conlon, R. A., Rossant, J., Breitman,
M. L. (1992) tek, a novel tyrosine kinase gene located on mouse chromo-
some 4, is expressed in endothelial cells and their presumptive precursors.
Oncogene 7, 1471–1480.
29. Cupedo, T., Vondenhoff, M. F., Heeregrave, E. J., De Weerd, A. E.,
Jansen, W., Jackson, D. G., Kraal, G., Mebius, R. E. (2004) Presumptive
lymph node organizers are differentially represented in developing mes-
enteric and peripheral nodes. J. Immunol. 173, 2968–2975.
30. Mebius, R. E., Schadee-Eestermans, I. L., Weissman, I. L. (1998) MAd-
CAM-1 dependent colonization of developing lymph nodes involves a
unique subset of CD4?CD3– hematolymphoid cells. Cell Adhes. Commun.
31. Breier, G., Breviario, F., Caveda, L., Berthier, R., Schnurch, H., Gotsch,
U., Vestweber, D., Risau, W., Dejana, E. (1996) Molecular cloning and
expression of murine vascular endothelial-cadherin in early stage devel-
opment of cardiovascular system. Blood 87, 630–641.
32. Eberl, G., Marmon, S., Sunshine, M. J., Rennert, P. D., Choi, Y., Littman,
D. R. (2004) An essential function for the nuclear receptor ROR?(t) in the
generation of fetal lymphoid tissue inducer cells. Nat. Immunol. 5, 64–73.
33. Hallmann, R., Mayer, D. N., Berg, E. L., Broermann, R., Butcher, E. C.
(1995) Novel mouse endothelial cell surface marker is suppressed during
differentiation of the blood brain barrier. Dev. Dyn. 202, 325–332.
34. Fu, Y. X., Chaplin, D. D. (1999) Development and maturation of second-
ary lymphoid tissues. Annu. Rev. Immunol. 17, 399–433.
35. Yu, P., Wang, Y., Chin, R. K., Martinez-Pomares, L., Gordon, S., Kosco-
Vibois, M. H., Cyster, J., Fu, Y. X. (2002) B cells control the migration of
a subset of dendritic cells into B cell follicles via CXC chemokine ligand
13 in a lymphotoxin-dependent fashion. J. Immunol. 168, 5117–5123.
36. Tumanov, A., Kuprash, D., Lagarkova, M., Grivennikov, S., Abe, K.,
Shakhov, A., Drutskaya, L., Stewart, C., Chervonsky, A., Nedospasov, S.
(2002) Distinct role of surface lymphotoxin expressed by B cells in the
organization of secondary lymphoid tissues. Immunity 17, 239–250.
37. Futterer, A., Mink, K., Luz, A., Kosco-Vilbois, M. H., Pfeffer, K. (1998)
The lymphotoxin ? receptor controls organogenesis and affinity maturation
in peripheral lymphoid tissues. Immunity 9, 59–70.
38. Ettinger, R., Browning, J. L., Michie, S. A., van Ewijk, W., McDevitt,
H. O. (1996) Disrupted splenic architecture, but normal lymph node
development in mice expressing a soluble lymphotoxin-? receptor-IgG1
fusion protein. Proc. Natl. Acad. Sci. USA 93, 13102–13107.
39. Mackay, F., Majeau, G. R., Lawton, P., Hochman, P. S., Browning, J. L.
(1997) Lymphotoxin but not tumor necrosis factor functions to maintain
splenic architecture and humoral responsiveness in adult mice. Eur. J.
Immunol. 27, 2033–2042.
40. Kuprash, D. V., Alimzhanov, M. B., Tumanov, A. V., Grivennikov, S. I.,
Shakhov, A. N., Drutskaya, L. N., Marino, M. W., Turetskaya, R. L.,
Anderson, A. O., Rajewsky, K., Pfeffer, K., Nedospasov, S. A. (2002)
Redundancy in tumor necrosis factor (TNF) and lymphotoxin (LT) signal-
ing in vivo: mice with inactivation of the entire TNF/LT locus versus
single-knockout mice. Mol. Cell. Biol. 22, 8626–8634.
41. Kim, M. Y., Toellner, K. M., White, A., McConnell, F. M., Gaspal, F. M.,
Parnell, S. M., Jenkinson, E., Anderson, G., Lane, P. J. (2006) Neonatal
and adult CD4? CD3– cells share similar gene expression profile, and
neonatal cells up-regulate OX40 ligand in response to TL1A (TNFSF15).
J. Immunol. 177, 3074–3081.
42. Bertrand, J. Y., Desanti, G. E., Lo-Man, R., Leclerc, C., Cumano, A.,
Golub, R. (2006) Fetal spleen stroma drives macrophage commitment.
Development 133, 3619–3628.
43. Gunn, M. D., Ngo, V. N., Ansel, K. M., Ekland, E. H., Cyster, J. G.,
Williams, L. T. (1998) A B-cell-homing chemokine made in lymphoid
follicles activates Burkitt’s lymphoma receptor-1. Nature 391, 799–803.
44. Tokoyoda, K., Egawa, T., Sugiyama, T., Choi, B. I., Nagasawa, T. (2004)
Cellular niches controlling B lymphocyte behavior within bone marrow
during development. Immunity 20, 707–718.
45. Sun, C. M., Fiette, L., Tanguy, M., Leclerc, C., Lo-Man, R. (2003) Ontogeny
and innate properties of neonatal dendritic cells. Blood 102, 585–591.
46. Dakic, A., Shao, Q. X., D’Amico, A., O’Keeffe, M., Chen, W. F., Shortman,
K., Wu, L. (2004) Development of the dendritic cell system during mouse
ontogeny. J. Immunol. 172, 1018–1027.
47. Lo, J. C., Chin, R. K., Lee, Y., Kang, H. S., Wang, Y., Weinstock, J. V.,
Banks, T., Ware, C. F., Franzoso, G., Fu, Y. X. (2003) Differential regulation
of CCL21 in lymphoid/nonlymphoid tissues for effectively attracting T cells to
peripheral tissues. J. Clin. Invest. 112, 1495–1505.
Vondenhoff et al.