| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Light Chains Recruit T Cells Independent of Complement1
* Section of Allergy and Clinical Immunology, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT 06520;
Noda Institute for Scientific Research, Chiba-ken, Japan;
Department of Immunology, Jagiellonian University College of Medicine, Krakow, Poland;
Department of Pharmacology and Pathophysiology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands; and
¶ Department of Dermatology and Allergology, Ludwig-Maximillians-University, Munich, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
L chains alone mediated CS initiation that was C-independent, but
was dependent on mast cells. Thus, B-1 cell-mediated CS initiation
required for T cell recruitment is due to activation of C by specific
IgM pentamer, and also subunits of IgM, while L chains act via
another C-independent but mast cell-dependent
pathway. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, to mediate classical late CS (1, 2).4 Initiation of the elicitation of CS
requires local release of the vasoactive amine serotonin by mast cells
and platelets (3, 4), and also release of the vasoactive
cytokine TNF-
by mast cells (2) to induce
expression of adhesion molecules, such as ICAM-1 and VCAM-1, on the
luminal surface of local endothelium (5). Expressed
adhesion molecules likely bind integrins expressed on the surface of
circulating activated CS-effector T cells (6, 7), leading
to T cell recruitment into the tissues.
Recently, we studied early steps in the CS initiation process. We found
that C activation was required to locally generate C5a to bind
C5aRs on mast cells and platelets, causing release of serotonin and
TNF- required for T cell recruitment (1, 2, 8). We
established that B cell Abs mediated C activation for CS initiation by
demonstrating absent CS ear swelling and local elaboration of IFN-
in pan B cell-deficient µMT (1) and
JH-/-4 mice, and in xid
mice4 predominantly deficient in B-1 B cells that
make the C-activating IgM (9). Elicitation of CS was
reconstituted in sensitized xid mice by adoptive transfer of
B-1 cells.4 Because B-1 B cells produce IgM Ab,
we tested the ability of anti-trinitrophenyl (TNP) monoclonal IgM
Ab to initiate CS in TNP-chloride (TNP-Cl) sensitized B-1
cell-deficient mice.4 Two different anti-TNP
IgM mAb given i.v. to sensitized xid mice just before
challenge reconstituted elicitation of 24 h CS, i.e., IgM Ab
restored recruitment of effector T cells.4
The current study analyzes the forms of IgM that mediate CS initiation
and demonstrates involvement of not only pentameric IgM, but also
subunits of IgM. Because prior study suggested that an Ag-binding and
CS-initiating factor (10, 11, 12) of molecular mass
smaller than IgM was involved, we examined the role of IgM subunits.
Interestingly, we found that IgM subunits had CS-initiating activity,
including dimers of covalently linked H and L chains, and also
separated H and L chains. Remarkably, even anti-TNP L chain and
recombinant anti-TNP chain both mediated CS initiation.
Although prior studies suggested that CS initiation was C-dependent
(1, 2, 8), the chain-mediated effect was not
C-dependent, but was mast cell-dependent, as described recently
(13). This likely is due to direct binding to mast cells
via a postulated receptor for L chains, enabling subsequent activation
by added Ag (13).
We concluded that IgM pentameric Ab is a major B-1 cell product
mediating CS initiation via C activation to generate C5a to activate
mast cell and platelet release of vasoactive mediators, leading to T
cell recruitment. Additionally, CS initiation also can be mediated by
subunits of IgM such as H-L dimers, perhaps via C activation,
but L chains act seemingly via direct mast cell activation to
initiate CS.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hybridoma 32.17 producing anti-TNP IgM (provided by J. Fleishman, Washington University, St. Louis, MO) was grown in DMEM with 10% FCS. Culture supernatant (1L) was applied to a 35-ml TNP-BSA Sepharose 4B column preequilibrated at 4°C. Then, the column was washed with PBS (pH 7.2) until OD280 was <0.01. Bound proteins were eluted with base (3 ml of 0.2 M Na2CO3 (pH 11)) followed by PBS. OD280 of eluates was measured by spectrophotometer and IgM H chain quantitated by ELISA. IgM H chain-positive fractions were pooled and dialyzed against PBS at 4°C, and protein concentrations measured by Bradford assay (Bio-Rad, Hercules, CA) using BSA standard.
Preparation of IgM from ascites with anti-H chain affinity chromatography
Anti-TNP IgM was prepared from subclone 13.4-G-8, derived from another hybridoma 13.4 (provided by F. T. Liu, Department of Dermatology, University of California, Davis, CA) by limiting dilution to isolate higher IgM Ab-producing cells. 13.4-G-8 cells were injected i.p. into nude mice and ascites precipitated with 50% saturated NH4SO4, dialyzed thoroughly against PBS, and centrifuged at 10,000 rpm x 30' to remove particles. The enriched IgM fraction was purified via an anti-mouse IgM H chain-linked agarose affinity column (Sigma-Aldrich, St. Louis, MO) equilibrated at 4°C with PBS made with LPS-free water, and then eluted with buffer (Pierce, Rockford, IL).
Superose 6 gel filtration and ELISA quantitation of anti-TNP IgM mAb
Fast protein liquid chromatography (FPLC; Amersham Pharmacia Biotech, Piscataway, NJ) using a Superose 6 HR 10/30 gel filtration column was used for further sizing purification of TNP affinity-purified 32.17 IgM. The column was equilibrated with 0.1 M Tris-HCl (pH 8.0) at 0.2–0.5 ml/min for 3 h. TNP affinity column-purified 32.17 IgM was concentrated by 100 K molecular mass cut-off concentrators (Millipore, Bedford, MA) to 2 mg/ml, and 1 mg IgM applied to the column and proteins were eluted at a flow rate of 0.3 ml/ml using 0.1 mM Tris-HCl (pH 8.0). Fractions (0.5 ml) were collected and assayed for IgM by ELISA vs standard mouse IgM (Calbiochem, La Jolla, CA) by dilution in 0.2 M carbonate buffer (pH 9.0) and coating wells at 37°C for 2 h, followed by blocking nonspecific binding with 3% milk, and then development using goat anti-mouse irrelevant myeloma IgM followed by anti-goat IgG-HRP (Caltag Laboratories, South San Francisco, CA), measuring color development at 450 nm. Peaks with high IgM were pooled for further studies.
Reductive alkylation of IgM to obtain subunits
Anti-TNP IgM (32.17 or 13.4) (1 mg/ml in 150 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA (pH 8)) was reduced strongly with 20 mM DTT or reduced mildly with 0.25 mM DTT for 2 h at 37°C. We determined previously that mild reduction resulted in predominant H-L dimers (i.e., covalently joined H and L chains), and strong reduction resulted in separated H and L chains. Iodoacetamide (40 mM) was added to alkylate the dissociated fragments of IgM, followed by thorough dialysis against PBS. Intact IgM and reductively alkylated fragments were analyzed by SDS-PAGE with 2–15% gradient gels (8 x 7 x 0.75-cm; Daiichi Chemical, Tokyo, Japan) under nonreducing conditions unless otherwise stated. Electrophoresis was conducted at 150 V starting at 25 mA for 1.5 h. Resolved proteins were visualized by staining with Coomassie blue for 1 h and then destaining with 5% acetic acid.
Isolation of TNP-binding IgM subunits from actively TNP-Cl contact-sensitized mice
Thirty CBA/J male mice (6–8 wk old) were contact sensitized with 150 µl of 5% picryl chloride (PCl) (TNP-Cl), and were sacrificed 1 day later. Then, lymph nodes and spleens containing CS-initiating cells were removed and processed to single-cell suspensions (14). The resulting lymphoid cells were cultured in complete RPMI medium without serum for 48 h at 1.7 x 107 cells/ml at 37°C in 5% CO2 (5 x 109 cells in 300 ml). Supernatant was collected by centrifuging the cells and applied to a TNP-BSA Sepharose B column that was washed thoroughly with PBS, and bound proteins were eluted with base (3 ml of 0.2-M Na2CO3, pH 11), 2-ml fractions collected, and OD280 eluted peak dialyzed at 4°C with PBS, resulting in protein at 2 mg/ml that was reduced and alkylated as above (14), and was later injected into 4-day PCl immune xid mice at 100 µg per recipient to test for initiation of CS, or analyzed by Western blotting.
Western blotting
For analysis of in vivo-derived anti-TNP binding material, 4
ml of TNP column-purified material was dialyzed against PBS,
concentrated with 10,000 molecular mass cut-off concentrators
(Millipore), and then was analyzed by Western blotting. Proteins were
separated by 12% SDS-PAGE under reducing and nonreducing conditions
and transferred onto a nitrocellulose membrane (Bio-Rad) at 30 V
overnight at 4°C. The membrane was blocked with 5% milk for 2 h
at 25°C followed by staining with goat anti-H or anti-
chain Abs (Caltag Laboratories), followed by anti-goat IgG-alkaline
phosphatase. Specific bands were developed by using
5-bromo-4-chloro-3-indolyl phosphate one component substrate
(Kirkegaard & Perry Laboratories, Gaithersburg, MD).
Reconstitution of PCl-sensitized B-1 cell-deficient xid mice by i.v. injection of IgM or subunits
Male 5- to 6-wk-old CBA/J (The Jackson Laboratory, Bar Harbor, ME) and CBA/N-xid mice (Taconic Farms, Germantown, NY) mice were rested at least 1 wk before use and maintained in microisolators according to guidelines of the Animal Care and Use Committee of Yale University School of Medicine (New Haven, CT). Spontaneous C5-deficient DBA/2, C5-normal DBA/1, or mast cell-deficient male WBB6F1/J-Kitw/Kitw-v mice and +/+ congenic wild-type WBB6F1+/+ controls (The Jackson Laboratory) also were used.
Mice were contact sensitized with 150 µl of 5% TNP-Cl in absolute ethanol and acetone (4:1) on the shaved chest and abdomen. Four days later, CS responses were elicited by a topical application on both ears with 10 µl of 0.4% TNP-Cl in acetone and olive oil (1:1). Thickness of challenged ears was measured with a dial caliper (Ozaki, Tokyo, Japan) or micrometer (Mitutoyo, Tokyo, Japan) before, and at 2 and 24 h after challenge. Increases in ear thickness in groups of 4–6 mice were expressed as the mean ± SE x 10-2 mm.
To reconstitute defective elicitation of CS in B-1 cell-deficient
CBA/N-xid mice, we i.v. transferred anti-TNP IgM
pentamer, or its subunit H-L dimer, or fully separated mixed H and L
chains. Reconstitution of CS with anti-TNP L chains in pan B
cell-deficient µMT mice was performed similarly. Prospective
recipient xid or µMT mice were sensitized with TNP-Cl and
4 days later, after i.v. transfer of IgM or subunits or L chains,
were ear challenged with TNP-Cl. One group of sensitized µMT mice was
injected i.v. with 5 µg TNP-specific L chain 30 min before challenge.
In some experiments, component C5 was depleted by injection of
anti-C5 mAb vs an isotype-matched control (Alexion, Cheshire, CT),
just 1 h before ear challenge (1).
In vitro hapten binding and C binding by IgM pentamer and subunits
To investigate the hapten-binding capacity of IgM subunits, anti-sialyl-LeX carbohydrate IgM mAb was used so that flow cytometry assay could be used, and was prepared from FH-6 hybridoma (American Type Culture Collection (ATCC), Rockville, MD), similar to 32.17 IgM purification. Reductive alkylation described above generated H-L dimers and separated H and L chains. HL-60 human leukemia cells, which naturally highly express surface sialyl-LeX determinants, were incubated for 45 min with dilutions of anti-sialyl-LeX monoclonal IgM pentamer, H-L dimer, or separated H and L chains in 4°C PBS containing 2% FBS and 0.05% NaN3. Biotin-conjugated anti-µ Ab (Cappel, Aurora, OH), and then streptavidin-PE (BD PharMingen, San Diego, CA) were added sequentially to detect sialyl-LeX-specific IgM and subunits bound on the surface of HL-60 cells using flow cytometry.
In a second direct hapten-binding assay, ELISA plates were coated either with TNP-BSA or BSA (50 µl, 100 µg/ml) overnight at 4°C, followed by blocking with 1% BSA in PBS at 25°C for 2 h, and then 100 µl IgM or subunits were added for 1 h at 25°C. Then, biotinylated anti-µ (Cappel) followed by 1/3000 diluted HRP-conjugated streptavidin (Vector Laboratories, Burlingame, CA) were added to detect bound IgM or subunits. Tetramethylbenzidine peroxidase substrate and tetramethylbenzidine one component stop solution (Kirkegaard & Perry Laboratories) were used for color development followed by measurement at 450 nm.
For assay of C binding after attaching anti-sialyl-LeX IgM mAb to the surface of highly sialyl-LeX expressing HL-60 cells (5 x 106) were incubated with IgM or subunits at 0°C for 45 min, washed once with gelatin Veronal buffer2+ (0.1% gelatin, 0.15 mM CaCl2, 0.5 mM MgCl2), and fresh normal human serum (2.5%) diluted with gelatin Veronal buffer2+ added for 60 min at 37°C as a source of C. Human C3b deposited on the cell surface was detected by flow cytometry using 2 µg/µl biotinylated goat IgG anti-human C3b (Cappel), followed by PE-conjugated streptavidin (1/400; BD PharMingen).
Isolation of TNP-specific L chain and recombinant L chain
TNP-specific IgG1 chain-containing mAb was isolated from
culture supernatant of murine 1B7-11 hybridoma (ATCC) by affinity
purification using protein G-Sepharose (Amersham Pharmacia Biotech)
(13). Purified IgG was reduced with 200 mM 2-ME in 5 mM Tris-HCl, 150
mM NaCl, pH 8.0 (60 min, 37°C), and subsequently alkylated with 300
mM iodoacetamide (3 min on ice). The treated sample was concentrated
(Centriprep-10; Millipore) and immediately subjected to high
performance gel filtration chromatography (Superdex-200; Amersham
Pharmacia Biotech). Ig H and L chains were separated in 6 M guanidine
(pH 6.5) using a flow rate of 0.5 ml/min. Fractions containing Ig
chains were pooled and were dialyzed extensively against PBS followed
by batch treatment with protein-G Sepharose beads to further exclude
contamination with H chain. Purity of the Ig H and chain fractions
was verified by reducing and nonreducing SDS-PAGE followed by Coomassie
blue or silver staining.
Recombinant Ig L chain was produced by PCR amplification of the
cDNA fragment coding for VL-CL from TNP-specific IgG1-producing
hybridoma 1B7-11 (ATCC). The amplified product was ligated into a pGEX
vector and the recombinant product was expressed in Escherichia
coli using standard molecular biological techniques. Recombinant
GST-fusion proteins were purified using glutathione-Sepharose affinity
chromatography (13).
L chain reconstitution of CS in µMT mice
CS reactions were studied in pan B cell-deficient 6- to
12-wk-old female C57BL/6 µMT mice (n = 4), kindly
provided by Dr. J. Louis (University of Lausanne, Lausanne,
Switzerland). Experiments were performed as described previously
(15). Briefly, mice were sensitized on one ear with 2%
TNP-Cl (PCl; 20 µl of a 4:1 acetone/olive oil solution), and 7 days
later were challenged with 1% TNP-Cl (PCl; 20 µl in 1:9
acetone/olive oil) applied on the previously untreated ear. Specific
ear swelling was determined by measuring ear thickness before and
24 h after the PCl challenge using a micrometer (Oditest;
Kroeplin, Schlüchtern, Germany). Nonspecific ear swelling caused
by 1% TNP-Cl (PCl) challenge in naive animals (irritant reaction) was
subtracted from the results of the experimental groups. One group of
µMT mice received one i.v. injection of 5.0 µg TNP-specific L
chains 0.5 h before the TNP-Cl (PCl) ear challenge, and a control
group received PBS injection i.v.
Statistics
Differences in values of ear swelling for various experimental groups were examined for significance using the two-tailed students t test, and in addition by analysis of variance by two-tailed ANOVA. Data are presented as the mean response ± SE, and p < 0.05 was taken as the level of significance.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We showed previously that elicitation of CS with TNP-Cl was impaired in pan B cell-deficient µMT (1) and JH-/- mice4 as well as in CBA/N-xid mice with dominant B-1 B cell deficiency.4 Reconstitution with B-1 cells from TNP-immune donors restored the impaired CS.4 Because B-1 cells mainly produce IgM (9), we tested that B-1 cell-derived anti-hapten IgM Ab was involved and found that hybridoma-derived anti-TNP IgM alone could reconstitute CS in previously sensitized B-1 cell-deficient xid mice.4 Therefore, we concluded that reconstitution of CS with immune B-1 cells likely was due to their production of anti-hapten IgM Ab.
To test whether IgM pentamer was responsible, we further purified
anti-TNP IgM mAb 32.17 via a combination of a TNP hapten affinity
column followed by FPLC size exclusion chromatography. The FPLC peaks
"I" through "IV" (Fig. 1
a) were analyzed by
nonreducing SDS-PAGE (Fig. 1b). The FPLC excluded void
fraction of high molecular mass material eluted as peak I and contained
98% pure IgM pentamer (900 kDa; Fig. 1b, lane
B). In contrast, the starting hybridoma-derived IgM eluted from
the anti-TNP affinity column (Fig. 1b, lane
A) contained predominant IgM pentamer that just entered the gel
above the 670-kDa marker, and also smaller IgM subunit bands, possibly
representing joined H-H dimer (160 kDa), H-L2
trimer (130 kDa), and some H-L dimer. The subsequent FPLC peak II (Fig. 1a) may predominately represent H-L2
trimer (Fig. 1b, lane C), and peak III may
represent mainly H-H and trace amounts of H-L trimer and H-L dimers, as
shown in lane D of Fig. 1b.
|
c).
Thus, hybridoma-derived anti-TNP IgM consisted of 
50% pentamer,
with the rest H-L chain dimers and trimers, and possibly separate H and
L chains.
The anti-TNP IgM pentamer 37.12 (FPLC Peak I) was tested for
ability to reconstitute CS in TNP-Cl-sensitized B-1 cell-deficient
xid mice. Fig. 1d shows that TNP-Cl-sensitized
and -challenged xid mice did not elicit 2 and 24 h CS
(group B), compared with vehicle immunized mice that were challenged on
the ears similarly (group A). In contrast, 2 and 24 h CS was
reconstituted when TNP-Cl immune xid mice received i.v. 100
µg FPLC-purified anti-TNP IgM pentamer 1 day before challenge
(group C). Thus, IgM pentamer likely was at least partially responsible
for initiating elicitation of CS that led to local T cell recruitment.
Surprisingly, fully reduced and alkylated anti-TNP IgM also
mediated 2-h CS-initiating activity that reconstituted 24-h CS
responses in xid recipients (group D), equivalent to or
greater than those mediated by intact pentameric IgM (group C). In
contrast, control nonspecific mouse myeloma IgM mediated neither 2-h
swelling nor 24-h CS in PCl-sensitized and ear-challenged
xid mice (group E). Thus, the CS-initiating activity of both
intact IgM pentamer and fully reduced anti-TNP IgM was
specific.
IgM H-L chain dimers, and separated H and L chains can also initiate CS
To identify which IgM subunits mediated CS initiation, we
subjected intact IgM of hybridoma 13.4 (Fig. 2a, lane 1) to
either mild reduction (0.25 mM DTT), which produced predominant dimers
of joined H-L chains (lane 2) or to strong reduction
(20 mM DTT), and then alkylation, resulting in separate H and L chains
(lane 3) observed in nonreducing SDS-PAGE.
Predominant pentamer restored 24 h CS in TNP-Cl sensitized
xid mice at a total dose of 100 µg or perhaps even 20 µg
i.v. per mouse (Fig. 2b, groups C and D). In addition,
covalently linked H-L dimers mediated CS initiation at doses of 100 and
20 µg per mouse (groups E and F). Surprisingly, even a mixture of
separated H and L chains derived from this hybridoma 32.17 mediated CS
initiation, also at doses of 100 and 20 µg per mouse (groups G and
H). We concluded that IgM pentamer, dimers of H and L chains, and even
a mixture of separated H and L chains, all could initiate CS. Mildly
reduced IgM contained a small amount of monomeric subunit
(H2-L2) that could have
contributed to activity, but pentamer could not have contributed, since
fully reduced IgM with strong CS-initiating activity (Fig. 1d, group D, and Fig. 2b, groups G and H)
contained no pentamer.
|
CS initiation by IgM subunits was unusual, so we tested if
subunits could bind hapten, since CS initiation is hapten-specific
(1, 2, 12).4 Using another IgM
hybridoma (FH-6) of different Ag specificity, we used a sensitive flow
cytometry assay to detect binding of this anti-sialyl-Lex
monoclonal IgM and its subunits to the carbohydrate hapten sialyl-LeX
that is expressed on the surface of HL-60 cells. Fig. 3a shows the pentameric
monoclonal IgM of FH-6 bound stronger than its derived IgM subunits at
10 µg/ml. With dilution, the last detectable binding concentration of
pentamer was 0.01 µg/ml (Fig. 3a, top), while
predominant H-L dimers bound at 10 and 1.0 µg/ml, but barely at 0.1
µg/ml; i.e., 10–20% binding compared with pentamer (Fig. 3a, middle). Finally, separated IgM H and L
chains had less binding at 10 µg/ml, and did not bind at 1.0 µg/ml
(Fig. 3a, bottom); i.e., 3% compared with
pentamer.
|
b, •) compared with BSA (
).
Furthermore, H-L dimers had much less binding, requiring 100-1000
µg/ml (
) that was significant vs binding to BSA (
). In fact,
even separated H and L chains had detectable but less binding to
TNP-BSA over this range (
) that still was significant compared with
BSA (
). We concluded that using two different IgM mAb with two
different Ag specifities and two different binding assays, one using
the CS-initiating anti-TNP IgM mAb (Figs. 2b and
3b), showed that IgM subunits significantly bound hapten,
but less than pentamer. This could account in part for ability of IgM
subunits to mediate CS initiation (Fig. 2b). IgM and H-L dimers activate C, but separated H and L chains do not
CS initiation by IgM requires C activation for local generation of
C5a to recruit T cells (1, 2, 8). We tested whether
subunits of anti-sialyl-LeX FH-6 IgM that weakly bound hapten could
then activate C. Fig. 3c shows a modification of the
hapten-cell flow cytometry assay, with initial incubation of intact IgM
or its subunits with HL-60 cells, then addition of diluted fresh human
serum, and then detection of bound C3b by using biotin-conjugated
anti-human C3b Ab by adding PE streptavidin. Binding of IgM
pentamer at 10 and 1.0 µg/ml enabled strong activation of C, but 0.1
µg/ml and 0.01 µg/ml IgM activated C much less (Fig. 3c,
top). IgM H-L dimers also activated C at 10 µg/ml, less so
at 1.0 µg/ml, and just barely at 0.1 µg/ml (Fig. 3c, middle); i.e., C activation 25%
compared with IgM pentamer. In contrast, separated IgM H and L chains
that weakly bound hapten (Fig. 3, a and b) could
not activate C (Fig. 3c, bottom). We concluded
that compared with intact pentamer, subunits of IgM consisting of H-L
dimers had less but significant ability to bind hapten and activate C,
which might account for ability to mediate CS initiation. In contrast,
separated H and L chains bound hapten very weakly and could not
activate C. Thus, C activation could not be responsible for the ability
of separated H and L chains to mediate CS initiation, since they weakly
bound hapten and could not activate C.
IgM subunits from lymphoid cells of actively TNP-Cl sensitized mice also initiate elicitation of CS
Pentameric anti-TNP IgM and subunits derived from established hybridomas initiated CS. However, the smallest subunits of separated H and L chains bound Ag weakly and could not activate C, but still initiated CS. This was puzzling and raised questions about the relevance of using hybridoma-derived Ag-specific mAb compared with the physiological situation in vivo. Therefore, we attempted to obtain CS-initiating anti-TNP IgM material derived in vivo by generating culture supernatants from mixed lymph node and spleen cells of TNP-Cl contact-sensitized mice, and then partially purifying TNP-binding material on a TNP hapten affinity column. We used 1-day TNP-Cl immune cells that mediate CS initiation (16)4 that likely is due to release of IgM by a small subpopulation of sensitized B-1 cells.4 Previously, this TNP-Cl-induced partially purified anti-TNP preparation was referred to as PCl factor or PCl-F (10, 11, 12).
The purified in vivo-derived anti-TNP material was reduced, then
after SDS-PAGE was tested via Western blot probed with anti-µ,
anti-, and anti-
chain Ab under reduced and nonreduced
conditions. Anti-µ Ab probe of TNP-Cl immune anti-TNP IgM
revealed IgM H chain corresponding to 80 kDa and an additional smaller
band corresponding to 66 kDa under reducing conditions, which likely is
deglycosylated H chain in the anti-TNP material, but no pentamer or
H-L dimers could be seen (Fig. 4a, lane 2).
Furthermore, probing with anti- and anti- Ab showed that
anti-TNP material contained L chain but not L chain
(lanes 5 and 8). Surprisingly, analysis
under nonreducing conditions also revealed that in vivo-derived
anti-TNP material that bound to and eluted from TNP affinity column
also contained some free µ and chains in addition to expected IgM
pentamer (lanes 3 and 6). This suggests
that anti-TNP µ and chains exist in vivo in TNP-Cl immune
mice. In contrast, lanes 1, 4, and 7
show a positive control mouse myeloma IgM stained with
anti-µ, anti-, and anti-. Notably, when the reduced
anti-TNP material was injected i.v. at a dose of 100 µg per mouse
into previously TNP-Cl-sensitized xid mice, there was small
but significant 2-h ear swelling elicited by TNP-Cl ear challenge (Fig. 4b, group E, left), and importantly, this 2-h
response led to strong reconstitution of 24-h CS (Fig. 4b,
group E, right). We concluded that anti-TNP IgM subunits
derived from 1-day PCl immune lymphoid cells had CS-initiating
activity, like subunits of hybridoma-derived anti-TNP IgM
mAb.
|
L chain-induced 2-h responses reconstitute 24-h CS in
TNP-Cl-sensitized xid mice
TNP affinity-enriched IgM subunits resulting from strong reduction
and alkylation were purified further, and isolated L chains but not
H chains (16) were found to mediate 2-h ear swelling
responses (Fig. 5a, group D).
To test whether the 2-h activity of anti-TNP L chains could
initiate CS, we used two anti-TNP L chain preparations; native
purified monoclonal anti-TNP chain and a recombinant
anti-TNP L chain. Recombinant murine anti-TNP L chain
cloned in E. coli (13) was expressed as a GST
fusion protein. We tested if the two anti-TNP L chain preparations
could reconstitute CS in PCl-sensitized xid mice. As before,
positive control immunized CBA/J mice elicited strong 2-h initiating
and 24-h effector phases of CS on day 4 (Fig. 5a, group A),
while similarly immunized and tested xid males elicited
neither 2- nor 24-h responses (group B). In contrast, monoclonal
anti-TNP L chains, given i.v. to TNP-Cl-sensitized
xid mice 1 day before challenge on day 3 enabled elicitation
of 2-h ear swelling 1 day later that resulted in
reconstitution of 24-h CS responses (group D), likely due to
recruited intact xid-derived CS-effector T cells induced by
prior immunization with TNP-Cl.4 Of great
importance, recombinant anti-TNP L chain (5 µg) similarly
reconstituted CS in TNP-Cl-immunized and -challenged xid
mice (group E). This was not due to contaminating GST, which was
inactive (group C). We concluded that both purified polyclonal and
recombinant monoclonal anti-TNP L chains can initiate CS.
|
We have shown previously that the 2-h ear swelling response is
dependent on C activation (1, 2, 8). Therefore, we
determined whether the 2-h ear swelling activity of anti-TNP L
chains involved activation of C. We used anti-C5 mAb treatment as a
C-depleting agent (1), and found that 2-h responses
induced by anti-TNP L chain (Fig. 5b, group D) were not
inhibited (group F) compared with recipients treated with isotype
control mAb (group E). In contrast, positive control 2-h responses in
1-day TNP-Cl active immune mice were strongly inhibited by anti-C5
mAb (group C), but not by the isotype control (group B).
We used another C-affecting procedure, use of genetically C5-deficient
mice (8), and found that anti-TNP L chain-induced 2-h
ear swelling was equivalent in C5-intact DBA/1 compared with
C5-deficient DBA/2 mice (Fig. 5c, group C vs D). In
contrast, responses of 1-day TNP-Cl immune mice were reduced
significantly in C5-deficient compared with C5-intact mice
(group A vs B). We concluded that 2-h ear swelling mediated by L chain
was C-independent, corroborating the in vitro results (Fig. 3c, bottom).
Isolated L chain-dependent 2-h responses are mast
cell-dependent, and consequent CS initiation is B cell-independent
Previous studies showed that 2-h swelling responses in actively
TNP-Cl-sensitized mice and CS responses to anti-TNP L chains depend
on activation of local mast cells (13). Thus, we tested
whether L chain-induced 2-h responses were mast cell-dependent. As
expected, responses in 1-day TNP-Cl active immune mice were mast
cell-dependent (Fig. 5d, group C vs D). Furthermore,
injection of anti-TNP L chains did not result in the ability to
elicit the 2-h responses in these
WBB6F1/J-Kitw/Kitw-v mast
cell-deficient mice (groups E vs F). Therefore, direct C-independent
mast cell activation likely accounts for the 2-h ear swelling activity
induced by anti-TNP L chains. This established a pathway that
was alternative to indirect mast cell triggering in CS via the C
cascade following mast cell stimulation through the C5a receptor
(2, 17). The latter occurs in actively TNP-Cl-sensitized
mice, most likely via anti-TNP IgM pentamers derived from B-1
cells.4
To determine whether anti-TNP-specific L chains could provide
CS-initiating early mast cell activation in the absence of other B
cell-derived signals, we studied anti-TNP L chains in CS
reactions of B cell-deficient µMT mice. As previously reported
(1), TNP-Cl-sensitized pan B cell-deficient mice, that in
this case received PBS i.v. as control, were not capable of developing
immune inflammatory 24-h CS ear-swelling responses when challenged with
the sensitizing hapten (Fig. 6). In
contrast, i.v. transfer of 5.0 µg of anti-TNP L chains 30 min
before the ear challenge fully reconstituted elicitation of 24-h
CS-responses in TNP-Cl-sensitized pan B cell-deficient µMT mice.
Thus, endogenous B cell-derived Ig chains do not appear to be required
for the initiation of T cell-mediated CS responses by L chains.
Taken together, these data support the concept that unlike anti-TNP
IgM pentamers, anti-TNP L chains initiate CS responses via
direct mast cell activation. Also, in contrast to indirect mast cell
activation through C5aRs via B-1 cell-derived IgM, anti-TNP L
chains mediate CS initiation independent of other Ig chains.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
L chains, can initiate CS. Thus,
using B-1 cell-deficient xid mice, we show that
anti-TNP-specific L chains derived from either hybridoma IgM or
generated from TNP-immunized mice, or via recombinant techniques, can
reconstitute the elicitation of defective CS responses. Interestingly,
although past experiments found that CS initiation was dependent on C
generation of C5a (2), the early ear swelling reactions
and CS initiation mediated by isolated anti-TNP chains were
found to be C-independent. This perhaps is due to direct binding of the
chains to mast cells via a postulated receptor for L chains,
enabling subsequent activation by added Ag. Indeed, cross-linking of
cell surface proteins of primary cultured mast cells via added L
chains results in degranulation and lipid mediator production
(13).
We showed previously that pan B cell-deficient µMT and
JhD-/- mice had impaired CS ear swelling
responses and absent local elaboration of IFN-, the crucial cytokine
produced by recruited CS-effector T cells in these Th1 responses
(1, 2).4 This suggested that B cells
and perhaps Igs could be directly or indirectly involved in elicitation
of these classical in vivo T cell-mediated CS responses. We also showed
that predominantly B-1 cell-deficient xid mice that have no
serum IgM also exhibit poor CS responses.4
Reconstitution of the xid mice or the
JhD-/- mice with B-1 cells from TNP-immune mice
injected just before challenge restored elicitation of CS responses in
an Ag-specific manner.4 Because B-1 cells are
major producers of IgM (9), we postulated that
anti-TNP-specific IgM produced by the B-1 cells might play a role
in CS initiation. This was confirmed by reconstitution of CS in
immunized xid mice by transfer of hybridoma anti-TNP IgM
Ab.4 However, we could not rule out the
possibility that other active mediators released by B-1 cells in vivo
could also mediate this activity. We now show that highly purified
preparations of pentameric anti-TNP IgM indeed are able to
reconstitute CS in xid mice, but in addition, even IgM
subunits can induce CS initiation. However, it is unlikely that the
doses we used of 100 µg IgM or subunits reflect the actual amount
needed in vivo in actively sensitized mice, since as little as 0.7 ml
of 1-day immune sera suffices,4 and there may be
other factors produced in vivo that boost the CS-initiating effect of
the IgM Ab, resulting in less being required.
The mechanisms by which IgM could initiate CS responses are not fully
understood. Thus, restoration of CS by B-1 cell-produced IgM Ab in
immunized xid mice could be due to several mechanisms.
Pentameric IgM can activate the classical C pathway resulting in
generation of C component fragments, such as C3a and C5a that could,
especially C5a, directly activate local microvessels (17, 18) to enable T cell recruitment (5). Indeed, we
found that there is elaboration of C5a at the local site within 1
h of challenge (2), and that CS responses are absent in C5
(8) and C5a receptor-deficient mice (2).
Alternatively, C fragments like C5a could lead to vascular activation
indirectly by binding C5aRs (17, 18), thereby
degranulating mast cells (19) and platelets
(20), resulting in release of vasoactive mediators like
TNF- (2, 5) and serotonin (3) that are
required to activate microvessels for the T cell recruitment
(5). Mast cell-derived TNF- also could play a role in T
cell recruitment via the induction of chemokines (15).
Either a direct vascular activation pathway or indirect vascular
activation via mast cell and platelet mediators would lead to release
of TNF- and serotonin to increase vascular permeability and crucial
expression of endothelial adhesion molecules such as ICAM-1 and VCAM-1
(5). This would allow local luminal binding and
extravasation of Ag-specific CS-effector T cells into the tissues to
thus initiate CS responses. This would be followed by T cell activation
on APC in the extravascular tissues for production of proinflammatory
IFN- and local induction of chemokines like IP-10 at the 24-h CS
site (1, 2).4 Notably, the
anti-TNP IgM Ab mainly is circulating and is located mainly in
the intravascular space, raising the question of how the IgM
binds the Ag to elicit responses. In elicitation of CS, locally
dispersed hapten bound to self Ag may penetrate the intravascular space
to enable binding to IgM. How IgM might get into the tissues is still
an open question, but we speculate that nonspecific irritation caused
during Ag challenge may alter the permeability of local vessels,
allowing circulating IgM to leak into the tissues and thus bind Ag to
begin CS initiation, or L chain activity initiates IgM transit into the
tissues.
It has been shown previously that separation of Ig H and L chains by
reduction of interchain disulfide bonds decreased C activation by 90%
(21). In this study, we also observed that fully separated
H and L chains of IgM did not activate C (Fig. 3c). However,
mildly reduced IgM that contained H-L dimers, and thus had some of the
remaining crucial interchain disulfide bonds, exhibited 25% C
activation compared with intact IgM pentamer molecules (Fig. 3c). Another study on the importance of interchain disulfide
bonds showed that separated and noncovalently reassembled IgG-derived H
+ L chains were devoid of ability to bind the first C component, C1
(22), whereas IgG-derived covalently linked H + L dimers
were as active as the intact parent IgG. Thus, these prior studies, as
well as our current observations, stress the requirement for interchain
disulfide bonds for binding and activation of C. However, we did not
find any correlation between the in vitro C-activating ability of H-L
dimers, or separated H and L chains with ability to reconstitute CS ear
swelling responses, since all the subunits were equally active (Fig. 2b). Therefore, we propose that initiation of CS by H-L
dimers and by separated H and L chains that actually turned out to be
mediated by chains alone may thus occur via a C-independent
mechanism.
Another prior study compared the binding of C1q to pentameric IgM, or
to monomeric H2-L2 subunits
of IgM, or to intact H2-L2
of IgG Ab specific for the same hapten Ag, which was fluorescein
(23). Interestingly, these investigators observed that the
C1q-binding ability of the IgM monomer
H2-L2 subunit was similar
to that of 2-L2 IgG. In
other words, when the pentameric IgM is reduced to the monomeric
H2-L2 subunit, and binds
the fluorescein hapten, it behaved much like a bivalent IgG Ab in its
ability to bind C1q, and then activate C. It was concluded that the
superior binding of C to intact IgM pentamer relative to the monomeric
H2-L2 subunit, or to
similar H2-L2 IgG, was
related to the polymeric structure and multiple active binding sites of
IgM Fc portions for binding C1q, rather than to structural or affinity
differences between the two hapten-binding Ig isotypes, i.e., IgM or
IgG. In the current study, we observed that monomeric
H2-L2 IgM subunits had
25% C-binding capacity compared with intact IgM pentamer. In view
of the observations above concerning binding of fluorescein
(23), this is not surprising since our data suggest that
one can expect only 20% C activation by IgM
H2-L2 monomer compared with
intact IgM pentamer.
This study is the first to show that fully separated L and H chains
also can initiate CS, and that this may proceed by yet another, and
C-independent, mechanism. We propose that separated L chains, exclusive
of H chains, may act directly by binding to postulated receptors on
mast cells (11, 13, 14, 24), since L chain-mediated CS
initiation was shown in this study (Fig. 5, b and
c) to be C-independent while being mast cell-dependent (Fig. 5d). Thus, in accord with a prior study (13),
our data demonstrate mast cell dependency of CS initiation via L
chains. Cross-linking of L chains bound to these putative mast cell
receptors leads to activation of mast cells (13),
resulting in release of vasoactive mediators required for T cell
recruitment. A similar receptor was proposed previously for mast cell
sensitizing material in a TNP-specific factor called PCl-F that was
derived from culture supernatants of lymphoid cells of TNP-Cl
contact-sensitized mice. Indeed, this PCl-F was absorbed out by in
vitro incubation with purified mast cells (14, 19). In the
current study, TNP-Cl CS was used to induce similar in vivo-derived
TNP-binding material that was shown to initiate CS (Fig. 4b). Thus, mast cells may be involved in CS initiation via
two different pathways. One may be an indirect mast cell-activating
process via pentameric IgM and some subunits activating C to generate
bioactive peptides like C5a to lead to vascular activation indirectly
via stimulating release of mast cell and platelet mediators. In
contrast, the other pathway for CS initiation activates mast cells
directly via IgM-derived L chains binding to mast cell receptors via
Ag, and activating mast cell release of vasoactive mediators
independent of C.
Because anti-TNP IgM or its subunits obtained from hybridomas that also mediated CS initiation may not be representative of what is present in vivo, we produced in vivo-derived anti-TNP IgM subunits by TNP-Cl skin contact-sensitizing mice and then harvesting 1-day immune lymphoid cell supernatants. As noted, originally such material that can initiate CS was described as PCl-F (10, 11, 12, 14, 19, 24), but was poorly characterized. This PCl-F seemed to initiate CS via mast cell activation and partial degranulation (19). An important characteristic of PCl-F was ability to initiate CS even after reduction of disulfide bonds followed by alkylation (10, 11, 12). This suggested that interchain or intrachain disulfide bonds in PCl-F may not be essential for CS initiation. The current study shows that subunits of anti-TNP IgM even are released directly from in vitro cultured hybridomas, or via reduction and alkylation of intact pentamers, and can initiate CS-like PCl-F. This new finding that anti-TNP IgM subunits can reconstitute CS in immune xid mice prompted us to investigate further the relationship between the poorly characterized PCl-F and anti-TNP IgM Ab subunits that both can mediate CS initiation.
We postulated that subunits of anti-TNP IgM might have been present
in the in vivo-derived PCl-F preparation, and may have been involved in
CS initiation. Indeed, immunoblotting of hapten affinity purified in
vivo-derived PCl-F that was run under nonreducing conditions and
stained with anti-µ anti--specific Ab showed remarkably
that even unreduced native PCl-F contained determinants of free IgM H
chain and L chain (Fig. 4a, lanes 3 and
6). Therefore, PCl-F derived from supernatants of lymphoid
cells from contact-sensitized mice likely contains anti-TNP IgM
subunits produced in vivo by lymph node and spleen B cells only 1 day
following contact sensitization. One possibility, albeit very small,
that we cannot rule out, is that these anti-TNP IgM subunits were
generated via in vitro processing of the in vivo-derived IgM pentameric
material produced by 1-day immune cells. Notably, these 1-day immune
cells are able to transfer CS initiation
(16),4 as does 1-day immune sera
that likely contains B-1 cell-derived immune
IgM.4 In fact, prior studies also have shown that
PCl-F contained smaller size molecules (30–50 kDa) (10, 11) that can mediate early ear swelling that at the time were
not thought of as derived from B-1 cell IgM. This finding now suggests
that even smaller subunits of IgM contained in PCl-F may be involved in
CS initiation. Indeed, in this study, we found that anti-TNP
material obtained from 1-day immune cells contained free IgM H chains
and L chains (Fig. 4a, lanes 3 and
6). Thus, we now have shown that fully reduced IgM that was
derived from an anti-TNP hybridoma or from this in vivo-derived
PCl-F that was reduced and alkylated are both able to reconstitute CS.
This allowed elicitation of both 2- and 24-h components of CS in
sensitized xid mice that provided the recruited CS-effector
T cells (Figs. 2b and 4b). Therefore, as
suggested by immunoblot analysis, anti-TNP IgM subunits possibly
present in vivo and similar to those derived in vitro from
IgM-producing hybridomas may have the ability to initiate CS.
This is consistent, since the presence of in vivo circulating L chains has been reported previously in several instances. Thus, concentrations of circulating Ig L chains have been associated with some clinical disorders of the immune system, such as multiple sclerosis (25, 26), Sjorgen’s Syndrome (27), systemic lupus erythematosus (28), chronic lymphocytic leukemia (29), and in normal patients and in those with kidney disease (30). Reports of abnormal synthesis of Igs by hybridomas also suggest that H chains normally are not secreted by plasma cells; they accumulate and cause plasma cells to die off, while L chains alone can be synthesized and secreted by some mouse myeloma cell lines (31). We also observed the presence of H and L chains in the anti-TNP IgM hybridoma supernatants in this study. In another instance, an H+ B cell lymphoma that did not secrete IgM still was able to secrete H chains (31, 32). Thus, putting these prior observations together, our proposal that anti-TNP IgM subunits could be produced in vivo is plausible. Studies are in progress to further characterize such in vivo-derived Ig subunits in contact-sensitized mice.
The most direct evidence that separated IgM subunit chains have
CS-initiating ability was demonstration that recombinant monoclonal
anti-TNP L chain alone, like purified monoclonal anti-TNP L
chain, could transfer 2-h ear swelling activity that mediated CS
initiation, both in pan B cell-deficient µMT mice (Fig. 6) and in B-1
cell-deficient xid mice (Fig. 5). The reconstitution of CS
in µMT mice with L chains rules out the possibility that endogenous
Ig chains such as H chains could have combined in vivo with the
transferred L chain to mediate CS-initiating activity. Although
reconstitution of µMT mice with purified TNP-specific L chain
restores CS, we cannot rule out the possibility that transferred
isolated chains may associate with non-Ig molecules in vivo to
exert the CS-initiating activity. In contrast to CS initiation by
anti-TNP IgM pentamer present in 1-day immune actively sensitized
mice, the CS-initiating activity of TNP-specific L chains appears to be
mediated independent of C. Accordingly, mice treated with anti-C5
or C5-deficient mice did not have decreased 2-h CS responses mediated
by L chain compared with responses in 1-day immune mice that were
inhibited by anti-C5 and by C5 deficiency, and thus, were
C-dependent, presumably due to dependence on intact IgM (Fig. 5, b and c). Additionally and in contrast, since
anti-TNP L chain-mediated 2-h CS responses like those in 1-day
immune mice were absent in mice that lack mast cells (Fig. 5d, group F), we propose that the isolated L chains act
directly on mast cells, possibly by binding to a putative receptor,
followed likely by cross-linking with multivalent challenge TNP-self
Ag, and thus release vasoactive mediators needed for local recruitment
of T cells (13).
In summary, we show for the first time that IgM Ab and its subunits
likely derived from B-1 cells early after immunization are involved in
the initiation of T cell-mediated CS responses. Furthermore, we propose
that initiation of CS responses that leads to the local recruitment of
CS-effector T cells can occur via at least two Ag-specific pathways
(33). The first is C-dependent via IgM pentamer that
activates C to generate fragments like C5a to trigger mast cell release
of vasoactive mediators indirectly. The second is C-independent, via
smaller IgM subunits such as L chains, that directly lead to mast
cell activation for T cell recruitment. Having such different and
redundant pathways makes sense physiologically to guarantee the
mediation of such a critical and essential T cell recruitment step in
vivo for effector T cell-mediated immune responses.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Philip W. Askenase, Section of Allergy and Clinical Immunology, Department of Internal Medicine, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06520-8013. E-mail address: philip.askenase{at}yale.edu
3 Abbreviations used in this paper: CS, contact sensitivity; FPLC, fast protein liquid chromatography; TNP, trinitrophenyl; TNP-Cl, TNP chloride; PCl, picryl chloride; PCl-F, PCl factor.
4 R. F. Tsuji, M. Szczepanik, I. Kawikova, V. Paliwal, R. A. Campos, M. Akahira-Azuma, N. Baumgarth, L. A. Herzenberg, and P. W. Askenase. B cell dependent T cell responses: IgM antibodies are required to elicit contact sensitivity. Submitted for publication.
Received for publication April 9, 2002. Accepted for publication August 2, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
: a possible initiating role of B cells. J. Exp. Med. 186:1015.. J. Immunol. 162:1648.4 integrin by CD4 T cells is required for their entry into brain parenchyma. J. Exp. Med. 177:57.1
1 and 22 in models of hypersensitivity and arthritis. J. Clin. Invest. 105:721.[Medline]
and light chains in the cerebrospinal fluid of patients with multiple sclerosis and other neurological diseases. J. Neuroimmunol. 19:119.[Medline]
light chain levels in chronic progressive multiple sclerosis. Pathobiology 62:29.[Medline]
light chains in the sera of patients with Sjorgren’s syndrome. J. Immunol. 130:2663.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  M. Kneilling, R. Mailhammer, L. Hultner, T. Schonberger, K. Fuchs, M. Schaller, D. Bukala, S. Massberg, C. A. Sander, H. Braumuller, et al. Direct crosstalk between mast cell-TNF and TNFR1-expressing endothelia mediates local tissue inflammation Blood, August 20, 2009; 114(8): 1696 - 1706. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Kerfoot, M. Szczepanik, J. W. Tung, and P. W. Askenase Identification of Initiator B Cells, a Novel Subset of Activation-Induced Deaminase-Dependent B-1-Like Cells That Mediate Initiation of Contact Sensitivity J. Immunol., August 1, 2008; 181(3): 1717 - 1727. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. U. Norman, S. Hulliger, P. Colarusso, and P. Kubes Multichannel Fluorescence Spinning Disk Microscopy Reveals Early Endogenous CD4 T Cell Recruitment in Contact Sensitivity via Complement J. Immunol., January 1, 2008; 180(1): 510 - 521. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Campos, M. Szczepanik, M. Lisbonne, A. Itakura, M. Leite-de-Moraes, and P. W. Askenase Invariant NKT Cells Rapidly Activated via Immunization with Diverse Contact Antigens Collaborate In Vitro with B-1 Cells to Initiate Contact Sensitivity J. Immunol., September 15, 2006; 177(6): 3686 - 3694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Hwang, J. Yamanouchi, P. Santamaria, and P. Kubes A Critical Temporal Window for Selectin-dependent CD4+ Lymphocyte Homing and Initiation of Late-Phase Inflammation in Contact Sensitivity J. Exp. Med., May 3, 2004; 199(9): 1223 - 1234. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
