The Journal of Experimental Medicine
JEM © The Rockefeller University Press
Vol. 202, No. 7, October 3, 2005 955–965
TIM-2 is expressed on B cells and in liver
and kidney and is a receptor for
Thomas T. Chen,
Suzy V. Torti,
Frances M. Brodsky,
William E. Seaman,
Jason G. Cyster,
Eréne C. Niemi,
and Michael R. Daws
Christopher D.C. Allen,
Mary C. Nakamura,
Frank M. Torti,
Veterans Administration Medical Center, San Francisco, CA 94121
Department of Medicine, Department of Microbiology and Immunology,
The G.W. Hooper Foundation and the Departments of Biopharmaceutical Sciences and Pharmaceutical Chemistry, University
of California San Francisco, CA 94143
Department of Cancer Biology, Wake Forest University Health Sciences, Wake Forest University, Winston-Salem, NC 27157
Howard Hughes Medical Institute,
T cell immunoglobulin-domain and mucin-domain (TIM) proteins constitute a receptor family
that was identified first on kidney and liver cells; recently it was also shown to be expressed
on T cells. TIM-1 and -3 receptors denote different subsets of T cells and have distinct
regulatory effects on T cell function. Ferritin is a spherical protein complex that is formed by
24 subunits of H- and L-ferritin. Ferritin stores iron atoms intracellularly, but it also
circulates. H-ferritin, but not L-ferritin, shows saturable binding to subsets of human T and B
cells, and its expression is increased in response to inflammation. We demonstrate that
mouse TIM-2 is expressed on all splenic B cells, with increased levels on germinal center B
cells. TIM-2 also is expressed in the liver, especially in bile duct epithelial cells, and in renal
tubule cells. We further demonstrate that TIM-2 is a receptor for H-ferritin, but not for
L-ferritin, and expression of TIM-2 permits the cellular uptake of H-ferritin into endosomes.
This is the first identification of a receptor for ferritin and reveals a new role for TIM-2.
T cell immunoglobulin-domain and mucin-
domain (TIM) proteins constitute a receptor
family that was identified first on kidney and
liver cells; recently it was also shown to be ex-
pressed on T cells (1–5). In humans, the TIM
receptor family seems to include only three
receptors, TIM-1, -3, and -4, whereas in the
mouse it may include as many as eight (4).
Human TIM-3 and -4 have apparent orthologs
in mice, based on sequence homology, but
human TIM-1 is almost equally homologous to
mouse TIM-1 (41%) and mouse TIM-2 (36%),
which are 66% homologous to each other.
In the mouse, the TIM gene family is
linked to a locus (
hypersensitivity and the production of Th2
cytokines (6). In accord with a role for TIM
receptors in immunity, TIM-1 is expressed
preferentially by Th2 cells, and polymorphisms
) that regulates airway
in human TIM-1 are associated with atopy,
asthma, and rheumatoid arthritis (5–10). Mouse
TIM-1 binds to TIM-4, which is expressed on
antigen-presenting cells, and ligation of TIM-1
potentiates T cell activation (7, 11, 12). In
contrast to TIM-1, TIM-3 is expressed prefer-
entially on Th1 cells. Blockade or loss of
this receptor in mice accelerates autoimmunity,
which suggests that ligation of the receptor is
Although TIM receptors are of evident im-
portance in immunity, their expression outside
of the immune system indicates that these re-
ceptors may have broader functions. Thus, in
primates and rodents, TIM-1 is expressed on
renal tubular cells; in primates, an alternatively
spliced form is expressed on liver cells, where it
has been usurped as a receptor for hepatitis A
(1–3, 16). The functions of TIM receptors on
these nonhematopoietic cells are unknown.
Ferritin is a spherical protein complex that
stores up to 4,000 iron atoms as an oxidized
mineral core (17). It is a heteropolymer that is
T.T. Chen and L. Li contributed equally to this work.
M.R. Daws’ present address is Department of Anatomy,
University of Oslo, Oslo, Norway NO317.
The online version of this article contains supplemental material.
William E. Seaman:
Abbreviations used: GC,
germinal center; IHC, immuno-
histochemistry; RT, room
temperature; TIM, T cell
and mucin domain; TREM,
triggering receptor expressed
on myeloid cells.
TIM-2 IS A B CELL RECEPTOR FOR H-FERRITIN | Chen et al.
formed by 24 subunits of H- and L-ferritin; their ratios vary
in different tissues and in response to iron, growth factors,
inflammation, or malignancy (18). Ferritin primarily is ex-
pressed intracellularly, where it regulates iron mineralization
and sequestration, and thereby buffers reactive oxygen spe-
cies. This effect of H-ferritin is essential for the antiapoptotic
effect of NF-
B, and the transcription of H-ferritin is up-
regulated by NF-
B (19). However, ferritin also circulates,
and earlier evidence suggests that H-ferritin acts as an im-
mune regulator, through binding to subsets of lymphocytes
and myeloid cells. Thus, ferritin inhibits T cell proliferation
in response to mitogens, it impairs the maturation of B cells
in vitro (20, 21), and it has immunosuppressive effects in
vivo (22). Additionally, H-ferritin, but not L-ferritin, shows
saturable binding to subsets of human T and B cells (23–26).
Despite the evidence for H-ferritin receptors on the cell sur-
face, none had been identified.
We demonstrate that TIM-2 is expressed at low levels on
all splenic B cells and is expressed at higher levels on germi-
nal center (GC) B cells. Outside the hematopoietic system,
TIM-2 is expressed in liver, especially in bile duct epithelial
cells, and in renal tubule cells. We further demonstrate that
TIM-2 serves as a selective receptor for H-ferritin, but not
for L-ferritin, and that binding of H-ferritin to TIM-2 leads
to the endocytosis of extracellular H-ferritin. The expression
of a surface receptor for H-ferritin is consistent with a role
for H-ferritin in modulating cell function—beyond its role
in storing iron—and the endocytic function of TIM-2 pro-
vides a new pathway for altering levels of H-ferritin inde-
pendent of gene expression.
TIM-2 is expressed on all splenic B cells, with high levels on
GC B cells
An expressed sequence tag for TIM-2 was isolated from the
database based on its partial homology to the Ig domain of
triggering receptor expressed on myeloid cells
(27). By expression of this cDNA, we prepared a mAb
against the extracellular domain of TIM-2. To define the
levels of TIM-2 on lymphocyte cell subsets, mice were not
immunized or were immunized with T-dependent antigens;
staining with anti–TIM-2 was assessed on subsets that were
defined by their surface phenotype. Studies from unimmu-
nized or immunized mice revealed that although TIM-2 was
expressed on follicular B cells, it was expressed at
higher levels on GC B cells (range 1.6–3.5) (Fig. 1 A). In
contrast, TIM-2 was not detected on T cells (CD4
CD8) (Fig. 1 A).
The expression of TIM-2 in splenic B cells, and its pref-
erential expression in GC B cells was confirmed by quantita-
tive RT-PCR, using SYBR Green. For these studies,
splenic B cell subsets and T cells were isolated by fluores-
cence-activated cell sorting and used to prepare total RNA.
Transcripts for TIM-2 were
B cells than in follicular B cells, whereas levels of TIM-2
transcripts in the marginal zone B cells were between the
10-fold more abundant in GC
two (Fig. 1 B). Little or no transcripts for TIM-2 were de-
tected in splenic T cells.
The preferential expression of TIM-2 on GC B cells
also was evident by immunohistochemistry (IHC). For
these studies, we raised a rabbit antiserum against a peptide
from the cytoplasmic domain of TIM-2. On fixed sections
from spleen, this antiserum demonstrated clusters of TIM-2
cells in the center of lymphoid follicles, although scat-
tered cells were seen elsewhere, including the red pulp (Fig.
2 A, middle panel). No staining was seen with control anti-
serum (Fig. 2 A, left panel); staining by the anti–TIM-2 an-
tiserum was blocked in the presence of the immunizing
peptide, which demonstrated that binding was antigen-spe-
cific (Fig. 2 A, right panel). Immunofluorescent studies
demonstrated that most, if not all, of the TIM-2
cytometric analysis of TIM-2 expression on mouse spleen cells. Spleens
were isolated 8 d after immunization with SRBCs. Each curve is normalized,
so that the peaks are of equal height. Staining of GC cells is more than
twice that of follicular B cells (Fol. B cells), although all B cells stain with
anti–TIM-2, compared with an isotype control mAb (dotted line; the control
shown is for GC B cells, but equivalent staining was seen with control
staining of follicular B cells, not depicted). TIM-2 was not detected on
CD4? or CD8? T cells. The numbers indicate geometric mean fluorescence
intensity (MFI) for TIM-2 staining of each cell subset. (B) Quantitative RT-PCR
analysis of TIM-2 expression. Total RNA was prepared from isolated subsets
of spleen cells (sorted by FACS) or from isolated tissues, and was analyzed
by quantitative RT-PCR. Expression of mRNA is shown as a ratio of TIM-2
mRNA/mRNA of two housekeeping genes, hypoxanthine-guanine phos-
phoribosyltransferase (HPRT) and GAPDH. Fol, follicular; MZ, marginal zone.
TIM-2 is expressed preferentially on GC B cells. (A) Flow
JEM VOL. 202, October 3, 2005
were clustered in the center of the follicles were B cells
(Fig. 2 B). To identify the relationship of TIM-2
GCs more clearly, mice were immunized with sheep red
cells 8 d before examination. By staining with peanut agglu-
tinin, the clusters of TIM-2
B cells were shown to be lo-
cated within GCs (Fig. 2 C). The IHC and immunofluores-
cence studies did not reveal expression of TIM-2 on all B
cells, as was detected by flow cytometry; we attribute this to
the lower level of TIM-2 expression on non-GC B cells,
which evidently is below the detection threshold of IHC.
In accord with studies by flow cytometry and quantitative
RT-PCR, T cell zones lacked TIM-2
tom row). As with IHC studies, staining with anti–TIM-2
cells (Fig. 2 B, bot-
was blocked in the presence of the immunizing peptide
TIM-2 is expressed in localized areas of the liver and kidneys
We used our rabbit antiserum to TIM-2 to demonstrate the
localized expression of TIM-2 in sections of fixed tissue from
liver and kidney. In the liver, the expression of TIM-2 was
most prominent in bile duct epithelial cells (Fig. 3, A and B).
To a lesser degree, staining of hepatocytes was seen, especially
at junctures of hepatocytes. This suggested that TIM-2 may
be expressed preferentially on bile canaliculi, which are
formed as openings between hepatocytes. Variable staining of
hepatocyte nuclei also was seen. All binding was blocked by
on spleen sections. (A) The anti–TIM-2 antiserum binds specifically to
clusters of cells in splenic follicles. Paraffin-embedded sections from
unimmunized mice were stained with an isotype-matched control rabbit
antiserum (control IgG), anti–TIM-2 antiserum (TIM-2); or anti–TIM-2
antiserum blocked with an excess of the immunizing peptide (TIM-2 ?
peptide), followed by goat anti–rabbit IgG conjugated to horseradish
peroxidase. The slides were counterstained with hematoxylin. (B) TIM-2–
positive cells are predominantly B cells. Adjacent cryostat sections from
Immunohistochemistry and immunofluoresence for TIM-2
unimmunized mice were co-stained for TIM-2 and B220 (top row) or
TIM-2 and CD3 (bottom row). TIM-2 staining appears green (left columns),
B220 and CD3 staining appear red (middle columns), and merged images
are shown in the far right columns. Bars, 100 ?m. (C) TIM-2–positive cells
are localized to GCs. Adjacent cryostat sections from the spleen of a
mouse that was immunized with SRBCs was stained for peanut agglutinin
(PNA, blue, left panel) to show the GCs, and with IgD (brown, right panel)
to show the surrounding follicular mantle zone. The section in the right
panel was stained for TIM-2 (blue) and IgD (brown).
TIM-2 IS A B CELL RECEPTOR FOR H-FERRITIN | Chen et al.
the immunizing peptide (unpublished data). Thus, like hu-
man TIM-1, mouse TIM-2 is expressed in the liver, but it is
particularly concentrated in bile duct epithelia and possibly
bile canaliculi. In the kidney, expression of TIM-2 was con-
centrated in tubular epithelial cells with the morphology of
distal tubular epithelial cells (Fig. 3, C and D). Staining was
blocked by the immunizing peptide (unpublished data).
Identification of a ligand for TIM-2
A soluble ligand for TIM-2 is released by macrophage
To identify ligands for mouse TIM-2, we cre-
ated a chimeric TIM-2 receptor, in which the extracellular
domain of TIM-2 (containing a FLAG epitope at the 5
was linked to the cytoplasmic domain of CD3
transmembrane region from CD8. This chimeric receptor
then was expressed in the T cell line, BWZ.36 (BWZ),
which contains a lacZ reporter gene under the control of
three copies of an NFAT regulatory element (28). The
cytoplasmic domain of the chimeric TIM-2 receptor
contains three immunoreceptor tyrosine-based activation
motifs. Activation of these immunoreceptor tyrosine-based
activation motifs recruits the tyrosine kinases, ZAP-70 and
Syk, and results in activation of the NFAT element and ex-
pression of lacZ. The expression of the chimeric TIM-2 re-
ceptor was confirmed by flow cytometry, using mAb’s to
FLAG or to TIM-2. Untransfected BWZ.36 cells did not
express TIM-2, as demonstrated by flow cytometry or by
PCR (unpublished data).
To screen cell lines for the expression of TIM-2 ligands,
a variety of hematopoietic and nonhematopoietic cells lines
was cocultured overnight with the BWZ.TIM-2/CD3
porter cells, followed by assessment of lacZ activation. Three
, using the
cell lines, all of macrophage lineage, stimulated the
cells: MT2, RAW264.7, and J774 (un-
published data). Furthermore, cell-free supernatants from
these macrophage cell lines also activated lacZ production by
cells, and this activation was blocked
by our mAb to TIM-2 (which does not activate TIM-2 un-
less it is cross-linked) (Fig. 4). The level of BWZ.TIM-2/
activation by cell supernatant was increased when the
macrophages were stimulated with LPS, and activation was
blocked by anti–TIM-2. Thus, macrophage cell lines pro-
duce a soluble ligand for TIM-2.
Previous studies indicated that TIM-2 binds to sem-
aphorin4A (Sema4a), and mice that are genetically defi-
cient in Sema4a have immune defects (29, 30). Therefore,
we prepared a soluble Sema4A/FcIg fusion protein, using the
same protein fragments used by Kumanogoh et al. This chi-
meric protein did not stimulate our reporter cell lines at con-
centrations up to 180 nM (unpublished data). Also, by flow
cytometry, the same Sema4a/FcIg chimeric protein did not
bind to the BWZ.TIM-2 cells or to CHO.TIM-2 cells, even
at concentrations of 40
g/ml of protein (
lished data). Thus, we have been unable to confirm the bind-
ing of Sema4A to TIM-2. Although this may require addi-
tional factors that are not represented in our detection
systems, these results indicated that the TIM-2 ligand that was
detected in macrophage supernatants was not Sema4A.
330 nM; unpub-
Molecular cloning of H-ferritin as a ligand for TIM-2.
To identify soluble ligands for TIM-2 that are produced by
macrophages, we screened a cDNA expression library that
was prepared from unstimulated MT2 macrophages. 144
pools of cDNA, each containing
clones, were used to transfect 293T cells. After 5 d in culture,
supernatants from these cells were screened for the activation
180 different cDNA
in kidney and liver. Paraffin-embedded sections from normal mice were
stained with anti–TIM-2 antiserum and counterstained with hematoxylin.
Representative sections are shown from liver (A and B) or kidney (C and D).
A and C, bars, 60 ?m; B and D, bars, 20 ?m. Staining for TIM-2 was
blocked by the immunizing peptide. In B, the arrow indicates bile duct
epithelial cells, and in D, the arrow indicates distal tubular epithelial cells;
both were stained with anti–TIM-2.
Immunohistochemistry demonstrates expression of TIM-2
RAW 264.7 (RAW), and J774 macrophage cell lines were cultured with or
without LPS (100 ng/ml) for 24 h. Conditioned medium from these cells
was used to stimulate the BWZ.TIM-2/CD3? reporter line for a period of 16 h.
For blockade of ligand action, anti–TIM-2 (?-TIM-2) mAb was used at 50
?g/ml. All wells were supplemented with PMA (10 ng/ml). Data are plotted
as lacZ production relative to that stimulated by ionomycin and normalized
to PMA alone. Bars represent mean of triplicate samples, and error bars
indicate standard error.
Release of TIM-2 ligand by macrophage cell lines. MT2,
JEM VOL. 202, October 3, 2005
stimulated the reporter cells, and the cDNAs from these pools
were pursued to isolate single colonies. The cDNA that was
isolated from both pools was identical and corresponded ex-
actly to mouse H-ferritin (GenBank M60170).
To verify that H-ferritin acts as a ligand for TIM-2, we
produced and purified recombinant H- and L-ferritin in bac-
teria as described previously (31), except that the protein was
passed over a polymyxin B column to remove endotoxin.
Recombinant H-ferritin stimulated the BWZ.TIM-2/CD3
reporter cell lines, with maximal stimulation at
(Fig. 5 A). H-ferritin did not stimulate untransfected BWZ
cells (Fig. 5 A), or BWZ cells that were transfected with
expression of these receptors at similar levels (unpublished
data). Similar results were obtained with H-ferritin produced
in eukaryotic (293T) cells (unpublished data). Furthermore,
H-ferritin, but not L-ferritin, stimulated the BWZ.TIM-2/
reporter cells (Fig. 5 B).
As an additional approach to documenting the interac-
tion between TIM-2 and H-ferritin, we expressed H-ferritin
on the surface of CHO cells by linking it to the CD8 trans-
membrane domain coupled to the CD3
main. We then prepared a soluble TIM-2–FcIg fusion pro-
tein, and examined binding by this protein to H-ferritin on
the surface of CHO cells, as assessed by flow cytometry. The
soluble TIM-2–FcIg fusion protein bound to CHO.H-fer-
ritin cells, but not to untransfected CHO cells, and binding
to the CHO.H-ferritin cells was blocked by antibody to
TIM-2 (Fig. 6). In sum, by two approaches, TIM-2 binds
H-ferritin. TIM-2 does not bind L-ferritin, and H-ferritin
does not bind TIM-1 or -3.
cells. Two supernatants strongly
, or TREM-2/CD3
TIM-2 is expressed on the cell surface and in endosomes.
To examine the expression of TIM-2 further, we transfected
BW5147 mouse T cells with the TIM-2 cDNA coupled to
enhanced GFP (EGFP) at its cytoplasmic tail. After clonal se-
lection, cells expressing TIM-2 were examined by deconvo-
lution fluorescence microscopy. Transfected cells were
mixed with untransfected cells on the same slide, as a con-
trol. In the absence of H-ferritin, TIM-2/EGFP
onstrated green fluorescence on the cell surface and in local-
ized perinuclear regions (Fig. 7). To characterize the nature
of the intracellular compartments in which TIM-2 is con-
centrated, we exposed TIM-2/EGFP
conjugated transferrin, which serves as a marker for endo-
cells to Alexa-568–
activates BWZ.TIM-2/CD3? reporter cells (black squares) but not untrans-
fected BWZ cells (open circles). The BWZ.TIM-2/CD3? reporter cells and
the untransfected BWZ cells responded equally to PMA and ionomycin
(unpublished data). (B) Recombinant H-ferritin (black squares) but not
L-ferritin (open triangles) activates BWZ.TIM-2/CD3? reporter cells.
H-ferritin but not L-ferritin ligates TIM-2. (A) H-ferritin
H-ferritin on the cell surface. CHO.H-ferritin cells were stained with
TIM-2–Fc (open curve, dashed) or with secondary antibody only (filled
curve). Addition of anti–TIM-2 monoclonal mAb (50 ?g/ml; open curve)
reduced binding to levels seen with secondary antibody alone.
TIM-2–FcIg fusion protein binds to CHO cells expressing
BW5147 T cells (which lack TIM-2) were transfected with the gene for
TIM-2, tagged at its COOH terminus with EGFP. Cells were exposed to
transferrin, and coupled to the red dye Alexa-568, as a marker for endosomes.
The figure shows the localization of TIM-2 and transferrin over the course
of 1 h. As shown in the top row, TIM-2 localizes to the cell membrane and
to localized intracellular compartments; this distribution is changed little
by exposure to transferrin. As shown in the second row, transferrin is in-
ternalized largely within 15 min, identifying endosomes. In the third row,
all cells are identified by staining of nuclei with DAPI. In the bottom row,
fluorescence by TIM-2, transferrin, and DAPI are overlayed, demonstrating
that the intracellular compartments expressing TIM-2 and transferrin
are largely overlapping. Note that the panels for 15 and 30 min include,
in addition to cells expressing TIM-2, cells that express little or no TIM-2
(arrowhead in 15-min panel). As expected, transferrin is endocytosed by
TIM-2? and TIM-2? cells. Bar, 10 ?m.
TIM-2 is expressed on the cell surface and in endosomes.
TIM-2 IS A B CELL RECEPTOR FOR H-FERRITIN | Chen et al.
somes. As expected, transferrin was internalized by TIM-2/
EGFP and TIM-2/EGFP cells (Fig. 7). In the TIM-2/
cells, transferrin and TIM-2 displayed punctate pat-
terns that became progressively overlapping (Fig. 7). Nota-
bly, transferrin localized to the perinuclear compartment
where TIM-2 is expressed prominently, which demon-
strated that at least a substantial portion of this compartment
represents membrane-derived endosomes.
TIM-2 is required for the cellular uptake of H-ferritin
We next examined the consequences of exposing the TIM-
cells to H-ferritin, tagged by conjugation to
Alexa-568. H-ferritin bound only to TIM-2/EGFP
and confirmed that H-ferritin does not bind to the cells in the
absence of TIM-2 (Fig. 8). With incubation at 37
ritin was internalized progressively by TIM-2/EGFP
Within 5 min, intracellular H-ferritin was evident in the
form of punctate structures beneath the cell surface, where it
colocalized with TIM-2; over the next 15–30 min, it colo-
calized increasingly with the perinuclear TIM-2/EGFP.
Thus, TIM-2 transports H-ferritin to punctate structures
within the cell, identified by transferrin as endosomes.
We have demonstrated that TIM-2 is expressed in localized
areas of the spleen, liver, and kidneys, and that it binds
H-ferritin, but not L-ferritin. Further, after binding of
H-ferritin to TIM-2 on the cell surface, H-ferritin is inter-
nalized into endosomes.
In the spleen, we found that TIM-2 is expressed on all B
cells, with expression at higher levels on GC B cells. Thus,
different members of the TIM family are expressed in differ-
ent subsets of immune cells; TIM-2 is on B cells, whereas
TIM-1 is inducible on Th2 cells, and TIM-3 is inducible on
Th1 cells. TIM-4 initially was identified on splenic stromal
cells, but recent studies indicated that it also is expressed on
antigen-presenting cells, where it serves as a ligand for TIM-1
(7, 12). We have not detected TIM-2 on resting T cells,
nor have we been able to induce its expression on T cells in
vitro by treatment with mitogen (ConA) or by stimulation
of T cells under conditions that induce Th1 or Th2 cells
(unpublished data). However, the induction of TIM-1 or -3
on T cells requires repeated stimulation, and we may not
have identified proper conditions to induce the expression of
TIM-2 on T cells. Regardless, although TIM-2 is 66% ho-
mologous to TIM-1, these two TIM receptors differ in their
expression on lymphocytes.
Our demonstration that TIM-2 binds to H-ferritin is the
first identification of a cell surface receptor for ferritin, al-
though the existence of ferritin receptors was postulated pre-
viously, based on saturable binding of ferritin to the surface
of several cell types, including B and T lymphocytes. Thus,
binding of H-ferritin to lymphocytes is increased in mito-
gen-stimulated cells, and purified H-ferritin forms surface
patches on T cells, after which it is endocytosed into lyso-
somes (23, 24, 32). Studies using MOLT-4 human T cells
indicated that they express a receptor for H-ferritin with an
association constant of
ceptor is increased on proliferating cells (25, 26). Biochemi-
cal evidence for an H-ferritin receptor also was found on
liver cells, including activated liver lipocytes (33, 34). In ad-
dition, specific ferritin binding was demonstrated on eryth-
roid precursor cells (35, 36), brain tissue (37, 38), and pla-
cental membranes (39).
In one report, a putative ferritin receptor was purified
from human liver (40). Although the molecular weight of
this receptor, 53 kD, is generally consistent with the pre-
dicted size of glycosylated TIM receptors, it was not identi-
fied further, and the specificity of this receptor for H-ferritin
was not shown. The liver seems to express at least two bind-
ing sites for H-ferritin; one binds L- and H-ferritin with
equal affinity, whereas the second, like TIM-2, binds only to
H-ferritin (33, 41). In healthy individuals, circulating ferritin
is predominantly L-ferritin. It was suggested that the first re-
ceptor, which binds H- and L-ferritin, may serve to regulate
levels of serum ferritin, whereas the H-ferritin receptor may
subserve independent cellular functions in response to a se-
lective increase in H-ferritin.
Ferritin is increased in inflammation, and we showed
previously that TNF-
and IL-1 activate transcription of the
H-ferritin gene in an additive manner, providing at least one
mechanism by which H-ferritin may be increased in inflam-
L/mol, and that the re-
not by cells lacking TIM-2. The experiments used the same cell line as in
Fig. 7, but here the cells were exposed to H-ferritin, identified by conjugation
to Alexa-568. As shown in the top row, the distribution of TIM-2 on the
cell surface becomes beaded within 5 min after incubation with H-ferritin;
its expression on the cell surface declines thereafter. As shown in the
second row, H-ferritin is internalized rapidly by cells expressing TIM-2. In
the third row, all cells are identified by staining of nuclei with DAPI. In the
bottom row, fluorescence by TIM-2, H-ferritin, and DAPI are overlayed.
Note that the panels for 0 and 30 min include, in addition to cells expressing
TIM-2, cells that express little or no TIM-2. Only cells expressing TIM-2
bind and internalize H-ferritin. The results are representative of four
separate experiments. Bar, 10 ?m.
H-ferritin is internalized by cells expressing TIM-2 but
JEM VOL. 202, October 3, 2005
mation (42, 43). The sources of circulating ferritin are not
defined; however, our cloning of H-ferritin from a macro-
phage line is in accord with previous studies of H-ferritin
production by macrophages, including studies that demon-
strated that LPS stimulates H-ferritin production by J774 and
RAW264.7 macrophage cells, both of which produced an
LPS-inducible ligand for TIM-2 (44, 45). Recent studies
have shown that transcription of H-ferritin is induced by
B, and the consequent interaction of H-ferritin with
iron serves to buffer the generation of reactive oxygen spe-
cies, an activity that is required for the capacity of NF-
–induced apoptosis (19). Our studies dem-
onstrate that TIM-2 permits the endocytosis of H-ferritin,
revealing that cellular levels of H-ferritin are not dependent
solely on transcription. This finding opens a new pathway
into understanding the role of ferritin in cell function.
The properties of TIM-2 that are required for endocyto-
sis of H-ferritin by TIM-2 remain to be elucidated. The cy-
toplasmic domain of TIM-2 has three tyrosine residues. The
membrane-proximal tyrosine is part of a YxxM motif,
which has been associated with endocytic localization as well
as with activation of PI3 kinase (46); however, the proximity
of this motif to the inner leaflet of the cell membrane may
render it nonfunctional. The other two tyrosines are not part
of known signaling or targeting motifs, but one motif
(E-[ED]-x-x-Y-x-x-E) is conserved through several TIM
receptors, which suggests functional significance.
By IHC of the liver, mouse TIM-2 is expressed highly in
bile ducts and, to a lesser extent, in hepatocytes. In accord with
this, we find high levels of TIM-2 transcripts in the liver (un-
published data). The expression of TIM-2 in bile duct epithelial
cells suggests that it may be involved in the transport of ferritin
into or out of bile. Ferritin excretion into bile is believed to in-
volve lysosomal exocytosis, but the mechanisms of excretion are
not well defined (47, 48). From our studies, it is possible that
this process involves TIM-2. Humans express high levels of
TIM-1 in the liver; however, in the mouse, levels of transcripts
for TIM-2 expression are 100–1,000 times higher than levels of
transcripts for TIM-1 (unpublished data). Although the localiza-
tion of TIM-1 in the human liver has not been described, these
results suggest the possibility that, in the liver, mouse TIM-2
may be a functional ortholog of human TIM-1.
No human ortholog for mouse TIM-2 has been identi-
fied, although mouse TIM-2 is only slightly less homologous
to human TIM-1 than is mouse TIM-1. Thus, human TIM-1
may share functions with mouse TIM-2 and -1, including
the capacity to bind H-ferritin. Alternatively, the capacity of
mouse TIM-2 to bind H-ferritin may have been usurped by
a different human receptor. Whatever the nature of the hu-
man ferritin receptor, the expression of TIM-2 in mice
roughly parallels that of known ferritin binding sites in hu-
mans. Thus, the expression of TIM-2 on mouse lympho-
cytes and hepatic cells corresponds with binding of H-fer-
ritin to human lymphocytes and liver cells, except that we
have not identified TIM-2 on resting T cells. However,
TIM-2 is expressed on mouse EL-4 thymoma cells (unpub-
lished data), and there may be conditions under which its
expression is induced on fresh T cells. Studies are in progress
to define the role of human TIMs in ferritin binding.
Our studies also may bear on the role of H-ferritin in ma-
lignancy. H-ferritin is increased often in malignancy, and its
expression correlates with poor prognosis. Thus, the expres-
sion of transcripts for H-ferritin by breast cancer cells is an ad-
verse prognostic indicator (49). Similarly, in ovarian cancer,
metastatic cells express higher transcripts for H-ferritin than
do nonmetastatic cells (50). In melanoma, levels of circulating
H-ferritin are elevated, and the levels of H-ferritin correlate
with levels of CD4
Further, H-ferritin was shown to activate regulatory T cells
through mechanisms that require dendritic cells (52). Studies
in rats similarly showed an up-regulation of H-ferritin during
the induction of hepatocellular carcinoma (53). Our studies
identify a receptor in mice through which H-ferritin may se-
lectively alter cell functions, including immune functions. It
will be of interest to pursue the possibility that the production
of H-ferritin by malignant cells alters the host response
through the TIM-2 ferritin receptor.
As an additional note, while this paper was under review,
Chakravarti, et al. demonstrated that transcripts for TIM-2
can be induced selectively in Th2 T cells (54). We did not
detect elevated levels of TIM-2 transcripts in freshly isolated
splenic T cells. In preliminary results we did, like Chakra-
varti, et al., find transcripts for TIM-2 in T cells after activa-
tion in vitro, but as noted in Results we have not defined
conditions that will induce detectable levels of TIM-2 on
the surface of T cells. We are testing the possibility that this
may be regulated by H-ferritin.
regulatory T cells (51).
MATERIALS AND METHODS
All animal studies were approved by the Animal Studies Subcommit-
tee of the Research & Development Committee, San Francisco Veterans Ad-
ministration Medical Center, by the UCSF Animal Care Committee.
C57BL/6 (B6), or congenic B6-Ly5.2 or B6-
obtained from The Jackson Laboratory and the National Cancer Institute, and
were maintained in a barrier facility. Where indicated, mice were immunized
i.p. with 3.5 ? 108 SRBCs (Colorado Serum Company), or with 50 ?g
4-hydroxy-3-nitrophenyl–acetyl-chicken gamma globulin (NP30-CGG; Bio-
search Technologies) precipitated in alum. Following euthanasia under ap-
proved protocols, tissues were obtained for flow cytometry and IHC at day 8
for SRBCs and day 14 for NP30-CGG, the peak of the GC response.
Gpi1a mice were
Cell lines. The BWZ.36 (BWZ) mouse T cell lymphoma cell line (BW5147
cells transfected with the cDNA for lacZ under control of four copies of the
NFAT promoter) was provided by N. Shastri (University of California Berke-
ley, Berkeley, CA) (28). The MT2 mouse macrophage cell line was provided
by M. McKichan (University of California San Francisco, San Francisco, CA).
All other cell lines are from the American Type Culture Collection. Cell lines
other than MT2 were grown in RPMI 1640 supplemented with 10% heat-
inactivated FBS, 25 ?M 2-ME, 2 mM L-glutamine, 100 U/ml penicillin, and
100 ?g/ml streptomycin (cRPMI 1640). MT2 cells were grown in IMDM
supplemented with 10% heat-inactivated FBS, 25 ?M 2-ME, 2 mM L-gluta-
mine, 100 U/ml penicillin. and 100 ?g/ml streptomycin.
Production of a monoclonal antibody and of antisera to TIM-2.
For monoclonal antibody production, rats were immunized two times in the
footpad with TIM-2–Fc fusion protein emulsified in Titermax adjuvant, fol-
TIM-2 IS A B CELL RECEPTOR FOR H-FERRITIN | Chen et al.
lowed by two immunizations with CHO cells expressing FLAG-tagged TIM-2.
Lymphocytes from the popliteal lymph nodes were fused with YB-2/0 plas-
macytoma cells, using standard techniques. Supernatants were screened for
binding to BWZ.TIM-2 cells, but not BWZ cells. Subsequent studies demon-
strated that the antibody did not bind to BWZ cells expressing TIM-1 or -3, as
documented by expression of a FLAG epitope at the 5? end (Fig. S1, available
Rabbit antisera against the cytoplasmic domain of TIM-2 were pro-
duced by ProSci Inc. by immunization with a synthetic peptide (DQVYI-
IEDTPYPEEES), corresponding to the 16 terminal residues of the full-
length TIM-2 protein. This sequence is not found in other members of the
TIM family. Antisera were affinity purified with the immunizing peptide.
Flow cytometry and cell sorting. Spleen cells were analyzed by flow
cytometry or sorted by FACS as described (55). Cells were gated with wide
forward scatter gates but narrow side scatter gates to include blast cells but
to exclude granulocytes, and then were gated further on singlet cells using
width and height parameters. Dead cells were excluded with propidium io-
dide. Follicular B cells and GC B cells were gated as depicted previously
(55). For flow cytometry, follicular B cells were defined as CD19?, GL7?,
Fas?, IgDhigh, CD4?, CD8?, and GC B cells were defined as CD19?,
GL7high, Fas?, IgDlow, CD4?, CD8?. For cell sorting, follicular B cells were
defined as CD19?, CD21med, CD23?; GC were defined as CD19?,
GL7high, Fas?, IgDlow; and marginal zone B cells were defined as CD19?,
CD21high, CD23?. The following monoclonal antibodies were used in
these studies (from BD Biosciences unless otherwise indicated): rat anti-
mouse CD4 (clone RM4-5, Alexa 405; Caltag), rat anti–mouse CD8 (clone
53–6.7, PerCP), rat anti–mouse CD19 (clone 1D3, PE-Cy7, APC, APC-
Cy7), rat anti–mouse CD21 (clone 7G6, FITC), rat anti–mouse CD23
(clone B3B4, PE), hamster anti–mouse CD95/Fas (clone Jo2, PE-Cy7), rat
anti–mouse T and B cell activation antigen (clone GL7, FITC), and rat
anti–mouse IgD (clone 11–26, PE, Southern Biotechnology Associates
Inc.). The monoclonal antibody to TIM-2 was biotinylated by using EZ-
Link (Pierce Chemical Co.) according to the manufacturer’s directions, and
binding was detected with streptavidin APC (Invitrogen). Detection of
FLAG-tagged TIM-2 used mAb M2 (biotinylated; Sigma-Aldrich). Detec-
tion of binding by the TIM-2/Fc chimeric protein was detected by using a
PE-labeled goat anti–human IgG Fc-specific secondary antibody with min-
imal cross-reactivity to other species (Jackson ImmunoResearch Laborato-
ries). Cells were analyzed on a FACSCalibur flow cytometer (Becton Dick-
inson), except that multicolor stains of fresh B cells were analyzed on a
FACSAria flow cytometer (Becton Dickinson), using FlowJo software
(Treestar). Fresh B cells were sorted on a Mo-Flo (DakoCytomation).
Quantitative PCR. Total RNA and first-strand cDNA were prepared as
described (55). For quantitative PCR, 4–8% of the cDNA was placed in a fi-
nal volume of 50 ?l containing SYBR green PCR Master Mix (Applied Bio-
systems, ABI) and primers (300 nM). Samples were analyzed on a 7300 real-
time PCR system (ABI) with the following thermal-cycler conditions: 50?C
for 2 min, 95?C for 10 min, and then 40 cycles of 95?C for 15 s followed by
60?C for 1 min. Data were analyzed by the Comparative CT Method (ABI
7700), giving a relative ratio of target gene mRNA to housekeeping gene
(hypoxanthine-guanine phosphoribosyltransferase, HPRT or GAPDH)
mRNA. Primers were designed by using PrimerExpress version 2.0 and se-
quences were as follows: TIM-2: forward, CCAACACCAGCACACACA-
GAGACCT, reverse, TGGCTTCTGTGGAGGGATTACTTCA; HPRT:
forward, AGGTTGCAAGCTTGCTGGT, reverse, TGAAGTACTCAT-
TATAGTCAAGGGCA; GAPDH: forward, GGTCTACATGTTCCAG-
TATGACTCCAC, reverse, GGGTCTCGCTCCTGGAAGAT.
Immunohistochemistry. Freshly harvested tissues were fixed by immer-
sion in neutral 10% buffered formalin (Fisher Scientific) overnight, followed
by embedding in paraffin. Tissue sections were cut at 10-?m thickness,
mounted on precleaned glass slides, and baked for 1 h at 80?C. Slides were
dewaxed with Citrisolv, equilibrated in distilled water, and subjected to an-
tigen retrieval by microwaving at low power for 5 min in 10 mM citrate
buffer pH 6.0 then cooling for 20 min. Endogenous peroxidase activity was
blocked by incubating in 0.3% hydrogen peroxide for 5 min; nonspecific
binding was blocked by incubating with 10% normal goat serum in 10 mM
Tris-HCl, pH 7.5, 150 mM NaCl (TBS) for 1 h at room temperature (RT).
Sections were incubated with affinity-purified anti–TIM-2 rabbit antiserum
diluted 1:1,000 in TBS with 0.1% Tween 20 (TBST)/5% BSA for 1 h at
RT, with or without immunizing peptide at 1 ?g/ml; washed extensively
in TBST; then incubated with goat anti–rabbit IgG conjugated to horserad-
ish peroxidase (Caltag) diluted 1:1,000 in TBST/5% BSA for 30 min at RT.
After washing in TBST, bound antibody was detected with 3,3?-diami-
nobenzidine (BioFX), counterstained with hematoxylin (Fisher Scientific),
and mounted with Permount (Fisher Scientific). Slides were photographed
under 10 and 40? objectives on an Olympus BX41 microscope equipped
with a DP12 digital camera.
For experiments examining GCs, dual-color IHC on cryostat sections
from mice that were immunized with SRBCs was performed as described
(55). Slides were observed with a 10? objective on a Leica DMLB micro-
scope equipped with an Optronics Engineering MDE1580 CCD camera.
Immunofluorescence of frozen tissue sections. Freshly harvested
spleens were embedded rapidly in OCT compound (Miles Inc.), and imme-
diately frozen in liquid nitrogen. Tissue sections were cut at 8-?m thickness,
mounted on precleaned glass slides, and fixed for 10 min in ice-cold acetone.
After rehydration in PBS, slides were treated with a biotin-blocking system
(DakoCytomation), and nonspecific binding was blocked by incubating in
TBS plus 10% normal goat serum for 1 h at RT. Sections were incubated
with affinity-purified anti–TIM-2 rabbit antiserum diluted 1:250 in TBST/
5% BSA as well as biotinylated antibodies against B220 (Caltag) or CD3 (BD
Biosciences). After extensive washing in TBST, slides were incubated with
goat anti–rabbit IgG conjugated to Alexa 488 and streptavidin conjugated to
Alexa 546 (Invitrogen) at 1 ?g/ml in TBST; mounted with Gelmount
(Biomeda); and visualized on a Zeiss Axioskop 2 FS fluorescent microscope
attached to a Hamamatsu Orca CCD camera. Digital images were acquired
with a 10? objective, contrast and brightness were adjusted automatically us-
ing OpenLab3 defaults, and the unmodified TIFF files were exported.
Vector construction. An expressed sequence tag corresponding to mu-
rine TIM-2 (AA575518) was identified in GenBank based on partial ho-
mology to the Ig domain of TREM-2, by using the tblastn algorithm. The
construct was obtained from the American Type Culture Collecrtion and
confirmed by sequencing. The full-length cDNA was expressed intact or
with the addition of a FLAG epitope, using a vector provided by L. Lanier
(University of California San Francisco, San Francisco, CA). In addition,
the extracellular domain of TIM-2 was expressed as a chimeric molecule
linked to the transmembrane domain of CD8 and the cytoplasm domain of
CD3?. This was achieved by constructing a vector containing a CD8 leader
sequence (amino acids 1–21, AAH25715) and FLAG epitope, followed by
the extracellular domain of TIM-2 (amino acids 22–231), the CD8 trans-
membrane domain (amino acids 187–215, AAH25715), and the CD3? cy-
toplasmic domain (amino acids 216–327, AAF34793) (using a vector pro-
vided by A. Weiss, University of California San Francisco, San Francisco,
California). A similar construct was used to express full-length mouse
H-ferritin on the cell surface, substituting the H-ferritin cDNA sequence
for the extracellular domain of TIM-2. These chimeric DNAs were shuttled
into the pcDNA4 vector (Stratagene), and the integrity of the constructs
was confirmed by sequencing. BWZ cells were transfected with either con-
struct by electroporation, and were selected in cRPMI 1640 containing
0.75 mg/ml Zeocin (Invitrogen).
The extracellular domain of TIM-2 was expressed in soluble form as an
Fc fusion protein. This construct was prepared in a pCDM8 vector into
which amino acids 22–230 of TIM-2 were inserted between leader se-
quence for signaling lymphocytic activation molecule (amino acids 1–24,
NP_038758.1) and the human IgG1 Fc domain (amino acids 243–473,
CAA75030) (vector provided by L. Lanier). The chimeric cDNA was shut-
JEM VOL. 202, October 3, 2005
tled into the pcDNA4 vector (Stratagene), and the integrity of the construct
was confirmed by sequencing. 293T cells in exponential growth were trans-
fected with 3 ?g plasmid DNA by using FuGENE (Roche); transfected
cells were selected in cRPMI 1640 containing 0.75 mg/ml Zeocin (Invitro-
gen). Ig fusion protein was purified from conditioned medium using pro-
tein G affinity chromatography.
Preparation and characterization of H- and L-ferritin. Mouse H-
and L-ferritin were produced in Escherichia coli, and were purified by su-
crose density gradient centrifugation as described (31). Endotoxin was re-
moved by adsorption to immobilized polymyxin B (Detoxi-GelTM, Pierce
Chemical Co.), according to the methods of the manufacturer. After ad-
sorption, endotoxin levels were below 5 EU/ml, as assessed by the chro-
mogenic end-point limulus amebocyte lysate assay (QCL-1000, Cambrex
Bio Science). Additionally, mouse H-ferritin was produced in human 293T
kidney cells by transfection with the cDNA, modified to use the signal se-
quence from CD8, followed by the sequence encoding a FLAG epitope
(DYKDDDDK), and then the H-ferritin sequence. For the expression of
H-ferritin on the cell surface, this construct was extended to include the
CD8 transmembrane domain followed by the cytoplasmic domain of
CD3?, and cells were transfected by electroporation. For fluorescence mi-
croscopy, H- and L-ferritin were labeled covalently with Alexa-568 by us-
ing a commercial kit (Invitrogen).
Reporter assay. The BWZ line, derived by Sanderson and Shastri (28)
from BW5147 T cells, contains a lacZ reporter construct regulated by four
copies of an NFAT regulatory element. BWZ cells transfected with the
TIM-2/CD3? chimeric receptor or (as a control) a TREM-2/CD3? chi-
meric receptor, were seeded in 96-well plates at 105 cells/well in cRPMI
1640 supplemented with 10 ng/ml PMA. To test for stimulation of these
cells by macrophage (or other) cell lines, the stimulator cells were added at a
ratio of 5:1 (i.e., 5 ? 105 cells/well), in triplicate. Alternatively, 100 ?l of
“conditioned” medium from confluent cells was added to triplicate wells.
For blocking experiments, anti–TIM-2 mAb was added to cells at 50 ?g/ml
at RT 15 min before ferritin. Plates were incubated for 16 h at 37?C in a 5%
CO2 humidified atmosphere. Cells were washed once in PBS, and lacZ ac-
tivity was determined by incubating the cells with 150 ?M chlorophenol-
red-?-D-galactopyranoside in PBS supplemented with 100 mM 2-mercap-
toethanol, 9 mM MgCl2, and 0.125% NP-40. After sufficient color
development, absorbance was measured at 595 nm, and was corrected for
background absorbance at 650 nm. Values were normalized by subtracting
the absorbance of wells that were treated with PMA alone. Maximum stim-
ulation was defined as the response to PMA plus 1 ?M ionomycin. For
stimulation of the reporter cell lines by purified mouse H- or L-ferritin, 5 ?
105 cells/well were stimulated in RPMI 1640 supplemented with 1% FBS.
For blocking experiments, anti–TIM-2 mAb or an isotype-matched control
mAb (2C7, rat IgG2a anti-ovalbumin) was added to cells at 10 ?g/ml at
RT 15 min before ferritin. Cells were assayed as above.
Molecular cloning of TIM-2 ligands. To screen for soluble TIM-2
ligands, a cDNA library that was prepared from MT2 macrophage cells was
transformed into bacteria, and the bacterial titer was determined by growth
on ampicillin plates. 144 pools of bacteria, each containing ?180 CFUs,
were grown overnight, and their DNA was isolated by using a Qiagen8
miniprep kit. Each miniprep was transfected into 293T to screen for soluble
TIM-2 ligands as follows: 15,000 293T cells were plated in 96-well plates
and allowed to adhere overnight. The next day, 0.5 ?g DNA was mixed
with 1.5 ?l FuGENE in 100 ?l OptiMEM medium and incubated for 30
min. Medium was aspirated from the 293T cells, and the OptiMEM/DNA
mix was added. 5 h later, 100 ?l RPMI 1640 supplemented with 20% FCS,
2 mM L-glutamine, 100 U/ml penicillin, and 100 ?g/ml streptomycin was
added to each well. After 5 d, conditioned medium was collected and cell
debris was removed by centrifugation. 100 ?l of this conditioned medium
was mixed with 105 BWZ.TIM-2-? reporter cells in the presence of 10 ng/
ml PMA, and the plates were incubated for 16 h at 37?C in a 5% CO2 hu-
midified atmosphere. Assay development was performed as detailed above.
Positive minipreps were retransformed into bacteria, and the screening was
repeated with pools containing 12 CFUs, and finally, with single colonies.
Expression of TIM-2 covalently linked to enhanced GFP. To cou-
ple EGFP to the TIM-2 cytoplasmic (C?) terminus, the full-length TIM2
coding sequence, including the endogenous signal sequence, was amplified
by PCR by using the primers 5?-CCGGGAATTCATGAATCAGAT-
TCAAGTCTTC-3? and 5?-TACCGTCGACTGGGACTCTTCT-
TCGGGGTAAGG-3?, and then cloned into the EcoRI and SalI sites of
the pEGFP-N1 expression vector (BD Clontech). Clones were verified for
integrity by sequencing and were transfected into BW5147 cells by using
FuGENE6. Stable transfectants were selected with G418 (Cellgro) at 1 mg/
ml in cRPMI 1640. Expression of the fusion construct was verified by flow
cytometry, assessing expression of GFP and, on the cell surface, TIM-2.
Deconvolution microscopy. BW5147 cells that were transfected with
TIM2-EGFP were washed twice in PBS and incubated in uptake buffer
(PBS supplemented with 10 mM Tris-HCl, 10 mM Hepes, 5 mM glucose,
1 mg/ml BSA, pH 7.4) for 30 min at 37?C. Cells were chilled on ice, and
H-ferritin (1 ?g/ml) or transferrin (50 ?g/ml), each tagged with Alexa-568
(Invitrogen), was added. After 30 min on ice, cells were washed once with
ice-cold PBS, and internalization was initiated by adding fresh uptake buffer
prewarmed to 37?C. Aliquots were removed at defined time points and
were added to excess ice-cold PBS with 0.02% sodium azide to stop inter-
nalization. Cells were pelleted, fixed in PBS/4% paraformaldehyde, and
mounted on coverslips coated with 0.1% poly-L-lysine (Sigma-Aldrich).
For some experiments, untransfected BW5147 cells were added as controls.
Images were collected by an API DeltaVision DV3 Restoration microscope
using a MicroMax 5 MHz cooled CCD camera (Roper Scientific); decon-
volution was performed using API SoftWoRx software.
Online supplemental material. Fig. S1 shows the binding of anti–TIM-2
mAb to TIM-2, but not to TIM-1 or -3. Online supplemental material is
available at http://www.jem.org/cgi/content/full/jem.20042433/DC1.
We thank S. Hayashi, J. Brown, and Y. Xu for expert technical assistance, and E. Theil
for helpful discussions.
This work was supported by the Veterans Administration and by grants R01
CA87922-01A1 (to W.E. Seaman), DK42412 (to S.V. Torti and F.M. Torti), and
GM038093 (to F.M. Brodsky) from the National Institutes of Health. T.T. Chen was
supported by an Advanced Career Development Award from the Veterans
Administration. C.D.C. Allen was supported by a Howard Hughes Medical Institute
predoctoral fellowship. M.C. Nakamura was supported by a Research Scholar Grant
RSG-01-167-01-LIB from the American Cancer Society.
The authors have no conflicting financial interests.
Submitted: 29 November 2004
Accepted: 16 August 2005
1. Feigelstock, D., P. Thompson, P. Mattoo, Y. Zhang, and G.G. Kaplan.
1998. The human homolog of HAVcr-1 codes for a hepatitis A virus
cellular receptor. J. Virol. 72:6621–6628.
2. Ichimura, T., J.V. Bonventre, V. Bailly, H. Wei, C.A. Hession, R.L.
Cate, and M. Sanicola. 1998. Kidney injury molecule-1 (KIM-1), a
putative epithelial cell adhesion molecule containing a novel immuno-
globulin domain, is up-regulated in renal cells after injury. J. Biol.
3. Ichimura, T., C.C. Hung, S.A. Yang, J.L. Stevens, and J.V. Bonventre.
2004. Kidney injury molecule-1: a tissue and urinary biomarker for
nephrotoxicant-induced renal injury. Am. J. Physiol. Renal Physiol.
4. Kuchroo, V.K., D.T. Umetsu, R.H. DeKruyff, and G.J. Freeman.
2003. The TIM gene family: emerging roles in immunity and disease.
Nat. Rev. Immunol. 3:454–462.
5. Khademi, M., Z. Illes, A.W. Gielen, M. Marta, N. Takazawa, C.
TIM-2 IS A B CELL RECEPTOR FOR H-FERRITIN | Chen et al.
Baecher-Allan, L. Brundin, J. Hannerz, C. Martin, R.A. Harris, et al.
2004. T Cell Ig- and mucin-domain-containing molecule-3 (TIM-3)
and TIM-1 molecules are differentially expressed on human Th1 and
Th2 cells and in cerebrospinal fluid-derived mononuclear cells in mul-
tiple sclerosis. J. Immunol. 172:7169–7176.
6. McIntire, J.J., S.E. Umetsu, O. Akbari, M. Potter, V.K. Kuchroo, G.S.
Barsh, G.J. Freeman, D.T. Umetsu, and R.H. DeKruyff. 2001. Identi-
fication of Tapr (an airway hyperreactivity regulatory locus) and the
linked Tim gene family. Nat. Immunol. 2:1109–1116.
7. Meyers, J.H., S. Chakravarti, D. Schlesinger, Z. Illes, H. Waldner, S.E.
Umetsu, J. Kenny, X.X. Zheng, D.T. Umetsu, R.H. Dekruyff, et al.
2005. TIM-4 is the ligand for TIM-1, and the TIM-1-TIM-4 interac-
tion regulates T cell proliferation. Nat. Immunol. 6:455–464.
8. McIntire, J.J., S.E. Umetsu, C. Macaubas, E.G. Hoyte, C. Cinnioglu,
L.L. Cavalli-Sforza, G.S. Barsh, J.F. Hallmayer, P.A. Underhill, N.J.
Risch, et al. 2003. Immunology: hepatitis A virus link to atopic disease.
9. Chae, S.C., J.H. Song, Y.C. Lee, J.W. Kim, and H.T. Chung. 2003.
The association of the exon 4 variations of Tim-1 gene with allergic
diseases in a Korean population. Biochem. Biophys. Res. Commun. 312:
10. Chae, S.C., J.H. Song, S.C. Shim, K.S. Yoon, and H.T. Chung. 2004.
The exon 4 variations of Tim-1 gene are associated with rheumatoid
arthritis in a Korean population. Biochem. Biophys. Res. Commun. 315:
11. Umetsu, S.E., W.L. Lee, J.J. McIntire, L. Downey, B. Sanjanwala, O.
Akbari, G.J. Berry, H. Nagumo, G.J. Freeman, D.T. Umetsu, and
R.H. Dekruyff. 2005. TIM-1 induces T cell activation and inhibits the
development of peripheral tolerance. Nat. Immunol. 6:447–454.
12. Shakhov, A.N., S. Rybtsov, A.V. Tumanov, S. Shulenin, M. Dean,
D.V. Kuprash, and S.A. Nedospasov. 2004. SMUCKLER/TIM4 is a
distinct member of TIM family expressed by stromal cells of secondary
lymphoid tissues and associated with lymphotoxin signaling. Eur. J. Im-
13. Monney, L., C.A. Sabatos, J.L. Gaglia, A. Ryu, H. Waldner, T. Cher-
nova, S. Manning, E.A. Greenfield, A.J. Coyle, R.A. Sobel, et al. 2002.
Th1-specific cell surface protein Tim-3 regulates macrophage activation
and severity of an autoimmune disease. Nature. 415:536–541.
14. Sabatos, C.A., S. Chakravarti, E. Cha, A. Schubart, A. Sanchez-Fueyo,
X.X. Zheng, A.J. Coyle, T.B. Strom, G.J. Freeman, and V.K.
Kuchroo. 2003. Interaction of Tim-3 and Tim-3 ligand regulates T
helper type 1 responses and induction of peripheral tolerance. Nat. Im-
15. Sanchez-Fueyo, A., J. Tian, D. Picarella, C. Domenig, X.X. Zheng,
C.A. Sabatos, N. Manlongat, O. Bender, T. Kamradt, V.K. Kuchroo,
et al. 2003. Tim-3 inhibits T helper type 1-mediated auto- and alloim-
mune responses and promotes immunological tolerance. Nat. Immunol.
16. Bailly, V., Z. Zhang, W. Meier, R. Cate, M. Sanicola, and J.V. Bonven-
tre. 2002. Shedding of kidney injury molecule-1, a putative adhesion
protein involved in renal regeneration. J. Biol. Chem. 277:39739–39748.
17. Theil, E.C. 2003. Ferritin: at the crossroads of iron and oxygen metab-
olism. J. Nutr. 133:1549S–1553S.
18. Torti, F.M., and S.V. Torti. 2002. Regulation of ferritin genes and
protein. Blood. 99:3505–3516.
19. Pham, C.G., C. Bubici, F. Zazzeroni, S. Papa, J. Jones, K. Alvarez, S.
Jayawardena, E. De Smaele, R. Cong, C. Beaumont, et al. 2004. Ferritin
heavy chain upregulation by NF-kappaB inhibits TNFalpha-induced
apoptosis by suppressing reactive oxygen species. Cell. 119:529–542.
20. Matzner, Y., C. Hershko, A. Polliack, A.M. Konijn, and G. Izak.
1979. Suppressive effect of ferritin on in vitro lymphocyte function. Br.
J. Haematol. 42:345–353.
21. Morikawa, K., F. Oseko, and S. Morikawa. 1995. A role for ferritin in
hematopoiesis and the immune system. Leuk. Lymphoma. 18:429–433.
22. Harada, T., M. Baba, I. Torii, and S. Morikawa. 1987. Ferritin selec-
tively suppresses delayed-type hypersensitivity responses at induction or
effector phase. Cell. Immunol. 109:75–88.
23. Renau-Piqueras, J., F. Miragall, and J. Cervera. 1985. Endocytosis of
cationized ferritin in human peripheral blood by resting T-lympho-
cytes. Cell Tissue Res. 240:743–746.
24. Fargion, S., A.L. Fracanzani, B. Brando, P. Arosio, S. Levi, and G.
Fiorelli. 1991. Specific binding sites for H-ferritin on human lympho-
cytes: modulation during cellular proliferation and potential implica-
tion in cell growth control. Blood. 78:1056–1061.
25. Moss, D., L.W. Powell, P. Arosio, and J.W. Halliday. 1992. Effect of
cell proliferation on H-ferritin receptor expression in human T lym-
phoid (MOLT-4) cells. J. Lab. Clin. Med. 120:239–243.
26. Moss, D., L.W. Powell, P. Arosio, and J.W. Halliday. 1992. Character-
ization of the ferritin receptors of human T lymphoid (MOLT-4) cells.
J. Lab. Clin. Med. 119:273–279.
27. Daws, M.R., L.L. Lanier, W.E. Seaman, and J.C. Ryan. 2001. Cloning
and characterization of a novel mouse myeloid DAP12-associated re-
ceptor family. Eur. J. Immunol. 31:783–791.
28. Sanderson, S., and N. Shastri. 1994. LacZ inducible, antigen/MHC-
specific T cell hybrids. Int. Immunol. 6:369–376.
29. Kumanogoh, A., S. Marukawa, K. Suzuki, N. Takegahara, C. Wa-
tanabe, E. Ch’ng, I. Ishida, H. Fujimura, S. Sakoda, K. Yoshida, and
H. Kikutani. 2002. Class IV semaphorin Sema4A enhances T-cell acti-
vation and interacts with Tim-2. Nature. 419:629–633.
30. Kumanogoh, A., T. Shikina, K. Suzuki, S. Uematsu, K. Yukawa, S.
Kashiwamura, H. Tsutsui, M. Yamamoto, H. Takamatsu, E.P. Ko-
Mitamura, et al. 2005. Nonredundant roles of Sema4A in the immune
system: defective T cell priming and Th1/Th2 regulation in Sema4A-
deficient mice. Immunity. 22:305–316.
31. Rucker, P., F.M. Torti, and S.V. Torti. 1997. Recombinant ferritin:
modulation of subunit stoichiometry in bacterial expression systems.
Protein Eng. 10:967–973.
32. Moss, D., A.R. Hibbs, D. Stenzel, L.W. Powell, and J.W. Halliday.
1994. The endocytic pathway for H-ferritin established in live MOLT-4
cells by laser scanning confocal microscopy. Br. J. Haematol. 88:746–753.
33. Moss, D., S. Fargion, A.L. Fracanzani, S. Levi, M.D. Cappellini, P.
Arosio, L.W. Powell, and J.W. Halliday. 1992. Functional roles of the
ferritin receptors of human liver, hepatoma, lymphoid and erythroid
cells. J. Inorg. Biochem. 47:219–227.
34. Ramm, G.A., R.S. Britton, R. O’Neill, and B.R. Bacon. 1994. Identi-
fication and characterization of a receptor for tissue ferritin on activated
rat lipocytes. J. Clin. Invest. 94:9–15.
35. Gelvan, D., E. Fibach, E.G. Meyron-Holtz, and A.M. Konijn. 1996.
Ferritin uptake by human erythroid precursors is a regulated iron up-
take pathway. Blood. 88:3200–3207.
36. Meyron-Holtz, E.G., B. Vaisman, Z.I. Cabantchik, E. Fibach, T.A.
Rouault, C. Hershko, and A.M. Konijn. 1999. Regulation of intracel-
lular iron metabolism in human erythroid precursors by internalized
extracellular ferritin. Blood. 94:3205–3211.
37. Hulet, S.W., S. Powers, and J.R. Connor. 1999. Distribution of trans-
ferrin and ferritin binding in normal and multiple sclerotic human
brains. J. Neurol. Sci. 165:48–55.
38. Hulet, S.W., S. Menzies, and J.R. Connor. 2002. Ferritin binding in
the developing mouse brain follows a pattern similar to myelination
and is unaffected by the jimpy mutation. Dev. Neurosci. 24:208–213.
39. Liao, Q.K., P.A. Kong, J. Gao, F.Y. Li, and Z.M. Qian. 2001. Expres-
sion of ferritin receptor in placental microvilli membrane in pregnant
women with different iron status at mid-term gestation. Eur. J. Clin.
40. Adams, P.C., L.W. Powell, and J.W. Halliday. 1988. Isolation of a hu-
man hepatic ferritin receptor. Hepatology. 8:719–721.
41. Anderson, G.J., W.P. Faulk, P. Arosio, D. Moss, L.W. Powell, and J.W.
Halliday. 1989. Identification of H- and L-ferritin subunit binding sites
on human T and B lymphoid cells. Br. J. Haematol. 73:260–264.
42. Miller, L.L., S.C. Miller, S.V. Torti, Y. Tsuji, and F.M. Torti. 1991.
Iron-independent induction of ferritin H chain by tumor necrosis fac-
tor. Proc. Natl. Acad. Sci. USA. 88:4946–4950.
43. Wei, Y., S.C. Miller, Y. Tsuji, S.V. Torti, and F.M. Torti. 1990. Inter-
leukin 1 induces ferritin heavy chain in human muscle cells. Biochem.
Biophys. Res. Commun. 169:289–296.
44. Recalcati, S., D. Taramelli, D. Conte, and G. Cairo. 1998. Nitric ox-
JEM VOL. 202, October 3, 2005
ide-mediated induction of ferritin synthesis in J774 macrophages by in-
flammatory cytokines: role of selective iron regulatory protein-2
downregulation. Blood. 91:1059–1066.
45. Kim, S., and P. Ponka. 2000. Effects of interferon-gamma and li-
popolysaccharide on macrophage iron metabolism are mediated by ni-
tric oxide-induced degradation of iron regulatory protein 2. J. Biol.
46. Wu, H., D.A. Windmiller, L. Wang, and J.M. Backer. 2003. YXXM
motifs in the PDGFbeta receptor serve dual roles as PI 3-kinase bind-
ing motifs and tyrosine-based endocytic sorting signals. J Biol Chem.
278:40425–40428 [erratum published in 278:45040].
47. Zuyderhoudt, F.M., P. Vos, G.G. Jorning, and J. Van Gool. 1985. Fer-
ritin in liver, plasma and bile of the iron-loaded rat. Biochim. Biophys.
48. Cleton, M.I., J.W. Sindram, L.H. Rademakers, F.M. Zuyderhoudt,
W.C. De Bruijn, and J.J. Marx. 1986. Ultrastructural evidence for the
presence of ferritin-iron in the biliary system of patients with iron
overload. Hepatology. 6:30–35.
49. Yang, D.C., F. Wang, R.L. Elliott, and J.F. Head. 2001. Expression of
transferrin receptor and ferritin H-chain mRNA are associated with
clinical and histopathological prognostic indicators in breast cancer.
Anticancer Res. 21:541–549.
50. Tripathi, P.K., and S.K. Chatterjee. 1996. Elevated expression of ferritin
H-chain mRNA in metastatic ovarian tumor. Cancer Invest. 14:518–526.
51. Gray, C.P., P. Arosio, and P. Hersey. 2003. Association of increased lev-
els of heavy-chain ferritin with increased CD4? CD25? regulatory
T-cell levels in patients with melanoma. Clin. Cancer Res. 9:2551–2559.
52. Gray, C.P., P. Arosio, and P. Hersey. 2002. Heavy chain ferritin acti-
vates regulatory T cells by induction of changes in dendritic cells.
53. Wu, C.G., M. Groenink, A. Bosma, P.H. Reitsma, S.J. van Deventer,
and R.A. Chamuleau. 1997. Rat ferritin-H: cDNA cloning, differen-
tial expression and localization during hepatocarcinogenesis. Carcino-
54. Chakravarti, S., C.A. Sabatos, S. Xiao, Z. Illes, E.K. Cha, R.A. Sobel,
X.X. Zheng, T.B. Strom, and V.K. Kuchroo. 2005. Tim-2 regulates T
helper type 2 responses and autoimmunity. J. Exp. Med. 202:437–444.
55. Allen, C.D., K.M. Ansel, C. Low, R. Lesley, H. Tamamura, N. Fujii,
and J.G. Cyster. 2004. Germinal center dark and light zone organiza-
tion is mediated by CXCR4 and CXCR5. Nat. Immunol. 5:943–952.