| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Microbiology and Immunology, Loyola University Chicago, Maywood, IL 60153
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
genes that are
expressed at different levels in mucosal tissues. To analyze the
factors involved in the differential expression of these
C genes, we cloned and sequenced the promoters of the
I regions that control the expression of sterile mRNA.
We found that all C genes, including C3
and C8, which are not expressed, and C4,
which is expressed at high levels, have similar nucleotide sequences in
the I region, and all contain the recognition elements
for TGF-ß in the promoter. B lymphocytes from popliteal lymph nodes
or Peyer’s patch activated in vitro could be induced by TGF-ß to
express sterile IgA transcripts of all IgA isotypes, except
C2, C3, and C8. Many
single B lymphocytes transcribed sterile mRNA of more than one IgA
isotype, which demonstrates that transcription of sterile mRNA alone
does not regulate the IgA isotype switch. The addition of IL-2 led to
the expression of transcripts of mature IgA of all isotypes, except
C2, C3, and C8. The
predominantly expressed isotype in these experiments was
C4. With the use of an IgA4-specific mAb we found that
IgA4+ plasma cells are unevenly distributed throughout the
small intestine such that many of the IgA+ plasma cells in
the duodenum-jejunum produced IgA4, whereas in the lower part of the
ileum IgA4-producing cells were almost absent. Because the microbial
flora varies throughout the intestine, we suggest that the microbial
flora creates different local environments and thus affects either
isotype switching or homing of IgA-expressing
cells. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
genes, whereas other
species have one or two C genes (1). The nucleotide
sequences of the rabbit C genes indicate that all 13
genes are functional (2). Schneiderman et al. (3) expressed the 13
C genes as chimeric IgA with a mouse VH gene
in mouse L chain-producing myeloma cells and showed that the secreted
IgAs could both fix complement and associate with secretory component
(4). Spieker-Polet et al. (5) used RNase protection analysis with
C gene-specific probes and found that in vivo, 10 or 11
C genes are expressed. The level of expression varied
for the different IgA isotypes; C4, C5,
C6, C9, C10,
C12, and C13 were expressed at high
levels, whereas C1, C2, and
C 7/11 were expressed at lower levels, and expression of
C3 and C8 could not be detected. All
C genes (excluding C3 and
C8) were expressed in gut-associated lymphoid tissues
(GALT),5 mesenteric lymph node
(MLN), and mammary gland, but surprisingly, in lung and tonsil one
C gene, C4, was expressed predominantly. It is well documented that in humans IgA1 and IgA2 are differentially expressed in various tissues. Brandtzaeg (6) found that in the respiratory tract and tonsil >80% of IgA plasma cells express IgA1. In the gut the distributions of IgA1 and IgA2 also were unequal and varied throughout the length of the intestine such that in the duodenum-jejunum 80% of the IgA plasma cells produced IgA1, whereas in the ileum the percentage of IgA1 plasma cells was 60%, and in the colon it was only 36%. Several investigators (7, 8, 9, 10) have shown for mouse and rabbit that switching to IgA occurs at high efficiency in Peyer’s patch (PP) and that the switched cells then localize to the lamina propria of the small intestine. Therefore, it is possible that the uneven distribution of IgA isotypes is due to differences in homing. However, it is also possible that switching occurs on site and is influenced by the local environment, which might be influenced by the microbial flora.
To study the conditions that lead to isotype switch and IgA secretion, several investigators studied the effect of ILs and TGF-ß in in vitro culture systems using resting B cells from various tissues. For instance, Coffman et al. (11), Lebman et al. (12), and Kim and Kagnoff (13) showed that murine splenic B cells, when stimulated to proliferate by LPS, can be induced by TGF-ß to secrete IgA and that the secretion was enhanced by IL-2. Other investigators reported similar effects of TGF-ß on murine spleen cells, but they found that IL-5 (14), IL-4 (15), or a combination of IL-4 plus IL-5 (16) enhanced IgA production. Further, Defrance et al. (17) and Kitani and Strober (18) reported that mitogen-activated human B lymphocytes, purified from tonsil or from peripheral blood, were stimulated by TGF-ß and IL-10 to secrete IgA.
Although the mechanism of isotype switch is not fully understood (for
reviews see Refs. 19 and 20), several investigators (21, 22, 23, 24) reported
that before switch recombination can occur, a sterile mRNA is produced,
and TGF-ß is required for this step. This sterile mRNA is initiated
in the I region, which is located 5' of the switch region, and a short
I exon is spliced to the unrearranged heavy chain
constant region gene. Although the function of this sterile mRNA is not
known, Wakatsuki and Strober (25) found that transfection of a mouse
cell line with an antisense I RNA led to down-regulation
of sterile transcripts and reduced IgA secretion. Further, Jung (26)
showed that I
1-/- mice, which are
defective in the I exon of IgG1, have no IgG1 in the periphery. Both
studies indicated that the I region, the sterile mRNA, or both are
required for isotype switch to occur.
Studies from several laboratories have shown that the expression of the
I exon is regulated by a highly conserved TATA-less
promoter of 150–200 nucleotides located 5' of the I
exon (27, 28). Further, in vitro expression and deletion experiments
identified TGF-ß-responsive elements in the I region
of mouse (27) and human (28, 29) C genes.
On the basis of these data, it is now accepted that in human and mouse
TGF-ß induces the expression of sterile
I-C mRNA, and that different ILs may be
required for the subsequent steps that lead to IgA switch recombination
and secretion. To explain the differences in the levels of expression
of the 13 C genes in rabbit, we identified their
I regions and searched for TGF-ß response elements in
the promoters. We also investigated in vitro which cytokines are
required for the induction of sterile and mature IgA mRNA of 11 of the
C genes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
regions
The I regions associated with 11 C
genes (all but C10 and C13) were
identified by hybridizing DNA of C-containing phage or
cosmid clones with a murine I-containing 1.4-kb
BamHI-HindIII fragment (provided by J. Stavnezer,
University of Massachusetts, Worcester, MA). The
I-containing fragments, located approximately 2 kb 5' of
the S region of each C gene, were
subcloned into M13, and the nucleotide sequences were determined using
a Sequenase kit (United States Biochemical, Cleveland, OH).
Oligonucleotide primers
Primers used for PCR amplification and their location are listed
in Table I
. The sequences for the 5'V and
5'J primers that were used to amplify C cDNA have been
published previously (30). The 5' primer used to amplify the
I genes is indicated in Fig. 1. The 3' pan
C-specific oligomers in exon 1 and exon 2 (3'
Ex1 and 3'Ex2) and the 3' oligomers
specific for the 13 C genes were used to amplify both
sterile and mature cDNA. The sequences of these primers are based on
published nucleotide sequences of the C genes (2) and
are, in most cases, sufficiently different not to lead to nonspecific
priming. Several primers, however, were checked for specificity. For
C7 and C11 the same primers were used
because the sequences of large regions of the two genes are identical.
The C4 primer weakly amplifies C2;
however, the C2 primer is specific for
C2. Because C2 is expressed only very
weakly, or in most cases not at all, the strongly expressed
C4 cDNA could be unequivocally identified. Lastly,
because the sequence of C13 is highly similar to those
of C7, C10, and C11, we
tested whether the C13 primer would also amplify DNA
from cosmid clones containing genomic DNA of these three
C genes by using a pan C-specific 5'
oligomer from exon 1 and the C13-specific 3' oligomer.
None of them was amplified.
|
|
RNA was prepared by homogenization of 1–3 g of frozen tissue in guanidinium thiocyanate followed by CsCl centrifugation and was stored in 75% ethanol at -20°C. cDNA was prepared from 3 µg of RNA with an oligo(dT) primer and reverse transcriptase (Super Script, Life Technologies, Gaithersburg, MD). For single rounds of PCR, the cDNA, equivalent to 1–2 ng of RNA was amplified with primers at a concentration of 0.2 µM. The amplification was performed with Taq polymerase (Perkin-Elmer) and 1 U/sample of 25 µl of buffer containing 15 mM MgCl2 in 30 cycles (45 s at 94°C, 1 min at 60°C, 1 min at 72°C).
Preparation of mAbs to IgA
Spleen cells of BALB/c mice immunized with Nonidet P-40 lysate
of 107 mouse myeloma cells expressing chimeric
dansyl-binding IgA4/mouse VH heavy chains and
dansyl-specific mouse L chains (3) were fused with Ag8 cells (mouse
myeloma P3 x 63-Ag8.653) using conventional methods. The
supernatant fluids of hybridomas were screened by ELISA using
dansyl-gelatin-coated plates incubated first with supernatants of the
transfected cells, followed by the hybridoma supernatant to be tested
and then with biotinylated goat anti-mouse Fc, avidin
horseradish peroxidase, and substrate, according to the manufacturer’s
instructions (ABC, Vector Laboratories, Burlingame, CA). The
specificity of each anti-IgA Ab was determined by testing it
against each of the supernatants of cells transfected with each of 12
IgA isotypes using the same ELISA. (The DNA of C5 and
C6 used by Schneiderman et al. (3) for the transfections
were of the f72,g74 allotype; all other genes were of the f71,g75
allotype.) For the remaining isotype, IgA12, for which no transfectomas
were available as well as for additional testing, we used supernatants
of rabbit hybridomas (31) secreting rabbit IgA of known isotype. For
this ELISA, plates were coated with goat anti-rabbit L chain Ab,
then incubated sequentially with: 1) supernatant containing a rabbit
IgA of known isotype, 2) the anti-IgA to be tested, and 3)
biotinylated goat anti-mouse Ig. The remaining steps of the assay
were as described above. One mAb, 102, reacted with all 13 IgA isotypes
and is designated pan IgA specific. Another mAb, 41-4, reacted with the
supernatant of IgA4-transfected cells and with the supernatants of
IgA2- and IgA8-transfected cells. Because in vivo we did not detect
expression of C8 mRNA, and expression of
C2 mRNA, was, if detectable at all, very low, this mAb
is designated IgA4 specific.
Cell cultures
IgM+ B cells were isolated from popliteal lymph
nodes (PLN) of unimmunized rabbits by simultaneous labeling with the
following three mouse mAbs: anti-rabbit IgA (mAb 102),
anti-rabbit CD 43 (L11/135) (32, 33), and anti-rabbit T cell Ab
(9AE10) (34). The secondary Ab was FITC-labeled goat anti-mouse Ig.
The unlabeled, B cell-enriched population, representing between 4 and
18% of the total number of sorted cells, was collected on a FACStar
Plus (Becton Dickinson, Mountain View, CA; made available in the FACS
Facility, Loyola University Chicago, Maywood, IL) and was 85–95%
IgM+, 96–99% CD43- 9AE10-, and
99% IgA-. A total of 1–2 x 105 cells
were incubated in 0.2 ml of RPMI with 15% FCS in 96-well round-bottom
plates for 4 days. Cells were stimulated to proliferate by adding
anti-b4 or anti-b5 
-chain allotypic Ab (25 µg/ml) and
irradiated (5000 rad) CD40L-transfected CHO cells (provided by Melanie
Spriggs, Immunex, Seattle, WA) at a concentration equal to 10% the
number of lymphocytes. In some experiments TGF-ß (R&D Systems,
Minneapolis, MN; 2 ng/ml) and recombinant human IL-2 (Genzyme
Diagnostics, Cambridge, MA; 10 ng/ml) were added.
RT-PCR of a defined number of cells
After 4 days in culture, 1–25 cells were deposited by a FACStar Plus directly into 10 µl of lysis buffer (10 mM DTT, 1% Nonidet P-40, 0.1 µM oligo(dT), and 16 U of RNA Guard (Boehringer Mannheim, Indianapolis, IN)/10 µl) on ice. (When single cells were analyzed, propidium iodide was added before sorting, and only live cells were collected.) The cDNA was prepared by adding 15 µl of first-strand buffer, 10 mM DTT, and 500 µM of each dNTP with 200 U of reverse transcriptase (SuperScript, Life Technologies) and 32 U of RNA Guard per sample. Samples were incubated for 60 min at 42°C, denatured at 96°C, and chilled on ice. For the first round of PCR, 25 µl of reaction mixture (Taq polymerase buffer, 1.5 mM MgCl2, 300 µM of each dNTP, 0.06 µM of each primer, and 2.5 U of Taq polymerase (Perkin-Elmer, Norwalk, CT)) were added. Amplification was performed for 35 cycles: 45 s at 94°C, 1 min at 60°C, and 1 min at 72°C. Samples were stored at 2°C. For the second round of PCR, 1–2 µl of the first amplification reaction was used as template, and nested or semi-nested primers were used as described in the legends to the figures and tables. Samples were visualized on polyacrylamide gels.
Immunofluorescence staining of gut sections
Sections (5 µm) of the jejunum and ileum were taken at a distance of 10 cm from the stomach and 5–10 cm from the sacculus rotundus and were stained with a pan IgA-specific or an IgA4-specific mAb. Different sections were used for each mAb. Both mAbs were counterstained with FITC-conjugated goat anti-mouse Ig.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Transcription of I-C mRNA
(designated sterile IgA mRNA) is required before IgA switch
recombination and is dependent on TGF-ß. Because a lack of functional
TGF-ß response elements in the I promoter might
explain why some rabbit C genes are not expressed, we
determined whether the I promoters of all
C genes contain the TGF-ß response elements. By
hybridizing C-containing cosmid and phage clones (2)
with a mouse I probe, we identified potential
I regions associated with 11 of the 13 C
genes. The I region of C13 could not be obtained because
the phage clone containing C13 does not have DNA 5' of
the constant region (2). We also did not succeed in identifying the I
region of C10. By nucleotide sequence analysis (Fig. 1
), we found that each of the 11 genes
had sequences highly similar to each other and to the I
promotor regions of mouse and human. Although several nucleotide
differences were observed throughout the promoter regions among the 11
sequences, these differences did not correlate with whether the
C genes were expressed. We have not identified a
transcriptional start site, but we assume that a start site is located
similarly to those in human and mouse, i.e., in the region 3' of the
TGF-ß response element and 5' of the location of the I
oligomer used to amplify I mRNA (Fig. 1). All 11
C genes contained the TGF-ß response element as well
as several other regulatory elements. Thus, we cannot attribute the
differences in the level of expression of the C genes to
missing TGF-ß response elements.
Induction of expression of sterile mRNA and IgA switch in vitro
In vitro assay for IgA switching.
Because all C genes contain a TGF-ß response element,
we expected that the expression of sterile or germline mRNA of all
isotypes could be induced by TGF-ß and that we could investigate the
requirements for switch recombination for the different IgA isotypes in
vitro. We used cells from PLN, a nonmucosal tissue, because we expected
that the B cells of PLN would not be precommitted to switch to IgA. The
cells were stimulated to proliferate by anti-L chain Ab and
CD40L-presenting cells, and TGF-ß and other cytokines were added to
induce the isotype switch. Evidence of isotype switch was obtained by
testing for I and C mRNA with RT-PCR.
Effect of TGF-ß and IL-2 on IgA switch.
We first determined the percentage of PLN B cells that can express
I- C mRNA (designated sterile IgA mRNA)
in response to TGF-ß. We incubated purified B cells, as described
above, with or without TGF-ß, and after 4 days in culture we
collected single cells by FACS and performed RT-PCR to amplify sterile
IgA mRNA. We found that most B cells from the cultures to which TGF-ß
was added expressed sterile IgA mRNA. For instance, in one
representative experiment (Fig. 2A) 10 of 13 single cells from
cultures with TGF-ß expressed sterile IgA mRNA, whereas in control
cultures without TGF-ß (Fig. 2B) we detected sterile IgA
mRNA in only two of 15 cells. The sizes of the sterile mRNAs were
slightly different, but the differences did not correlate with
different isotypes, as will be seen later (Table II).
The size differences may be due to the use of different transcriptional
start sites. We also determined the expression of VDJ-C
transcripts (designated mature IgA mRNA) and found that none of these
cells expressed functional IgA mRNA (data not shown). Similar results
were obtained in two independent experiments. We conclude that PLN
cells can be induced by TGF-ß to express sterile IgA mRNA in vitro.
|
|
A). In contrast, few, if any,
pools from control cultures without IL-2 expressed mature IgA RNA (Fig. 3B). To test whether IL-2 alone was sufficient to induce the
switch to IgA, we activated cells in the presence of IL-2 with or
without added TGF-ß. We found that seven of 10 samples from cultures
to which IL-2 and TGF-ß were added expressed IgA mRNA (Fig. 4A), whereas only two of 10
samples from cultures to which IL-2 but no TGF-ß was added expressed
IgA mRNA (Fig. 4B). Based on these results we conclude that
for switching from IgM to IgA in rabbit, as has been shown in human and
mouse, TGF-ß induces the expression of sterile IgA mRNA before the
switch to functional IgA mRNA can occur. Further, we found that upon
the addition of IL-2, cells switched and expressed mature IgA mRNA.
|
|
Isotype-specific primers.
To determine which IgA isotypes are expressed by the cultured PLN
cells, we designed antisense (3') oligomers specific for each
C gene (Table I) and tested whether these primers could
amplify by RT-PCR all individual C isotypes from cDNA of
gut and appendix tissues, which are known to express all IgA isotypes
(5). We found, in agreement with our earlier results obtained by RNase
protection (5), that all C genes except
C3 and C8 could be amplified, and that
C2 was expressed at a low level and was barely
detectable by PAGE (Fig. 5). When we used
cDNA at a 100-fold higher concentration, C3 and
C8 were amplified also, but at levels that could barely
be visualized by PAGE (data not shown). We conclude that these primers
can be used to amplify cDNA of each C gene. The
specificity of the primers is discussed in Materials and
Methods.
|
), and we found three cells that expressed C4
only, three that expressed C9 only, and one that
expressed sterile mRNA of more than one isotype (C4,
C9, and C13; Table II, PLN). The sizes of
the sterile transcripts of different cells shown in Fig. 2 were
somewhat different, but these size differences do not correlate to the
isotypes, since cells 1 and 3, expressing the somewhat smaller
transcript, as well as cells 6 and 13, which have a larger transcript,
all express the same isotype, I4. To determine whether
the expression of sterile mRNA of more than one IgA isotype was common
to many cells, we performed the same experiment with cells from PP, a
tissue in which cells switch frequently to IgA (8). Cells from PP were
purified, activated, and analyzed in the same way as described for
cells from PLN. We found that 14 of 18 B cells from cultures activated
with TGF-ß and IL-2 expressed sterile IgA transcripts. We further
analyzed seven of these 14 single cells with isotype-specific primers
and found that all seven cells expressed sterile mRNA of more than one
IgA isotype (Table II, PP) and two of these cells, 3 and 7, also
expressed mature IgA mRNA. When we tested these two cells with
isotype-specific primers (specific for C1,
C4, C5, C9,
C10, C12, or C13), we
found that both cells expressed mature IgA mRNA only of one isotype,
C4, and in addition multiple sterile mRNAs. These
results imply that even after switching has occurred, sterile mRNA can
still be expressed. It is possible that sterile transcripts can be
expressed from different alleles. This could explain the results
obtained with cell 7 (Table II, PP), in which mature and sterile
transcripts of the same isotype, C4, are expressed. The
number of sterile IgA isotypes expressed by a single cell varied
between two and five. Thus, it appears that expression of sterile mRNA
of several IgA isotypes can be induced at the same time in a single
cell. Therefore, on the basis of expression of sterile IgA mRNA, we
cannot predict the isotype to which the cell will switch.
Although no isotype appears to be expressed predominantly, we found no
sterile C2, C3, or C8 mRNA
in any of the single cells analyzed. These are the same three isotypes
in which mature mRNA was either undetectable or barely detectable
in mucosal tissues (see Fig. 5). These data support the theory that
isotype switch is preceded by the expression of sterile IgA mRNA.
We determined the nucleotide sequence of the I exons
of four genes, C4, C9,
C11, and C12 (Fig. 6). We found that the exons are spliced
to exon 1 of the constant regions. We have not determined the
transcriptional start site, but if it is positioned like that in human
and mouse (see Fig. 1), the I exons of rabbit encompass
about 200 nucleotides. The sequences of the I exons were
similar to each other but different from those in human and mouse
(23); however, the organization and size of the Ia regions in rabbit
are very similar to those in human and mouse.
|
). However, in the second experiment
(Expt 2) performed under the same conditions as Expt 1, many different
IgA isotypes were expressed (Table III). In this second experiment
C4, C5, C7/11,
C9, C12, and C13 were
expressed frequently; C1 and C10 were
expressed less frequently; and as was observed for sterile mRNA,
C2, C3, and C8 were not
amplified in any of eight pools of 25 cells. Some pools expressed as
many as eight different C mRNAs, which implies that in
this experiment one of every three cells has undergone isotype switch
(since it is unlikely that a single cell expresses more than one mature
mRNA). The difference between the results of the two experiments is
striking. One has to consider, though, that for every experiment the B
cells are obtained from a different rabbit. The total PLN population is
negatively sorted to obtain µ+ cells, but not all sorted
cells are µ+. Consequently, in every experiment different
kinds and numbers of accessory cells that can influence the isotype
switch may be present. Also in vivo, large differences in the amount
and variety of IgA expression are found in the various mucosal tissues.
The relative levels of mature IgA mRNA of the different IgA isotypes in
one experiment (no. 2, Table III) correlate well with the levels found
in vivo in mucosal tissues with an RNase protection assay (5).
Conversely, the results of the other experiment (no. 1, Table III), in
which mostly C4 was expressed, closely resemble the in
vivo findings for lung and tonsil. We conclude from these experiments
that B cells from PLN can be stimulated by TGF-ß and IL-2 in vitro to
undergo isotype switch and to express functional mRNA of several IgA
isotypes but that some as yet unknown factors regulate the choice of
the isotype.
|
We sought to determine whether the differential expression of IgA
isotypes as determined by RNase protection is reflected at the protein
level in situ. It had been shown in humans that only one of the two
isotypes, IgA1, was found in the respiratory tract and that the same
IgA isotype was found in 80% of the IgA plasma cells in the
duodenum-jejunum (6). We had found in rabbit that mostly one
C gene, C4, is expressed in the
respiratory tract (5). Therefore, we hypothesized that in rabbit IgA4
might also be more highly expressed in the jejunum than in the ileum.
To test this hypothesis, we produced an mAb that is specific for the
IgA4 isotype and another mAb that is pan specific for all IgA isotypes
(see Materials and Methods). These mAbs were used to
determine the location and the fraction of IgA+ cells in
sections of the duodenum-jejunum and the ileum (see Materials and
Methods). With the pan -specific mAb we found large numbers of
IgA+ plasma cells in both areas (Fig. 7, B and D). In
contrast, we found many IgA4+ plasma cells in the
duodenum-jejunum and very few in the ileum (Fig. 7, A and
C). We conclude that in rabbit, as in human, the IgA
isotypes are unevenly distributed in the small intestines.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our goal was to explain the difference in levels of expression of
the different IgA isotypes in rabbit. Previous studies, using RNase
protection assays, showed that the 13 rabbit C genes are
expressed at different levels in vivo. The present data from in vitro
studies with PLN cells support our earlier findings in that several of
the C genes are expressed frequently, whereas others are
expressed less frequently or not at all, and C4 is often
the predominantly expressed isotype. Significantly, for the isotypes
that are not expressed or are expressed at low levels in vivo, we did
not find sterile IgA mRNA expressed in response to TGF-ß in vitro. We
suggest that the low expression or lack of expression of some isotypes
is due to an inefficient I promoter that does not
promote transcription of sterile mRNA, a necessary step for switch
recombination to occur. However, the low expression of these isotypes
does not appear to result from the lack of TGF-ß response elements.
Instead, it could result from changes in other regulatory elements in
the I promotor, such as ATF/CRE, PU-1, and the E-box.
Further, the low expression might be influenced by positive as well as
negative regulating regions at some distance 5' of the I
promoter, as shown for mouse (27) and human (28). The influence of the
I promoter and possibly other more 5' regions of the DNA
on expression of sterile I transcripts needs to be
investigated.
Expression of mature IgA transcripts
TGF-ß, without addition of ILs, could induce B cells from the
PLN that were stimulated to proliferate by anti-L chain Ab and
CD40-L-presenting cells to express sterile IgA mRNA at a high
frequency. For example, in one experiment, 75% of the B cells had
sterile transcripts. Further, we found that many single B cells from
PLN as well as from PP simultaneously expressed sterile transcripts of
two to five different IgA isotypes. Expression of two sterile
transcripts, I1-C1 and
I
-C, in single murine cells has been
reported previously (35). In both rabbit and murine examples, the
expression of sterile transcripts of different isotypes is initiated by
the same factor, i.e., expression of I1 and
I by IL-4, albeit at different concentrations, and
expression of all rabbit I isotypes by TGF-ß. This
observation implies that although expression of sterile mRNA precedes
the actual switch, the selection of the C gene to which
the cell will switch does not occur when expression of sterile mRNA is
initiated. In some experiments expression of mature IgA transcripts was
induced in as many as 25% of the PLN B cells when IL-2 was added to
the cultures. These results imply that the addition of TGF-ß and IL-2
creates in vitro an environment that induces switching to IgA. We do
not know how IL-2 contributes to isotype switching, whether it has a
direct effect on B cells or whether it stimulates accessory cells
present in the cultures that are then activated to promote the isotype
switch. It has been reported, for instance, that CD40L-presenting cells
can promote isotype switch (36), and Schrader et al. (37) reported
evidence that dendritic cells promote IgA isotype switching. In vivo
few, if any, cells of PLN express IgA, which may imply that some cells,
like CD40-L expressing cells and/or factors such as TGF-ß, IL-2 or
ILs produced in response to IL-2, are not present or are not active in
PLN. Consistent with the low expression of IgA in vivo in PLN, we
observed that, when we fused cells of PLN from locally immunized
rabbits with the rabbit fusion partner 240E, none of approximately 50
hybridomas produced IgA (unpublished results), whereas 40–60% of
hybridomas from cells of mesenteric lymph node or PP from systemically
immunized rabbits secreted IgA.
Preferential expression of C4
In several in vitro experiments C4 was
preferentially expressed, whereas in other experiments many
C isotypes were expressed. This difference might be
explained by the presence or the absence of accessory cells in our
cultures. Various non-B cells might be important in regulating the
isotype switch. It has been reported, for instance, that dendritic
cells can skew the expression of human IgA in vitro to the IgA1 isotype
(38). Because we negatively sorted PLN B cells, we do not know how many
and what kind of cells, which might contribute to the isotype switch in
some unknown manner, were present in the in vitro cultures of the
different experiments. One can argue that a particular environment,
perhaps created by the presence or the absence of accessory cells,
supports switch recombination only to C4. We found
another situation supporting this hypothesis when we analyzed large
numbers of cells, i.e., 200,000, compared with pools of 25 cells used
for the data reported in Fig. 3 and Table III. If the cells were
cultured without IL-2, where only approximately 1% of the cells
undergo switch to IgA, we found that those switched almost exclusively
to C4.
The preference for the switch to C4 has precedence in
vivo, where we found that in tonsil and lung the predominantly
expressed C gene is C4, and occasionally
even in PP we found that C4 was the only
C gene expressed (5). We cannot exclude that
IgA4-expressing cells preferentially home to lung and tonsil. However,
we think it is more likely that the IgA switch occurs in all mucosal
lymphoid organs, but in lung and tonsil, the switch to
C4 is favored. Because we sometimes find only
C4 expressed in PP, where IgA switch actively leads to
expression of many IgA isotypes, our hypothesis that certain conditions
or environments favor the switch to C4 over all other
IgA isotypes is strengthened. One can also argue that C4
is most easily switched to, perhaps because it is the most 5' of the
C genes, whereas switch to the other isotypes requires
special or different conditions. We do not know what these conditions
might be; because we have shown that most TGF-ß-activated single
cells express sterile transcripts of more than one IgA isotype, we
conclude that the selection of a single C gene occurs
after TGF-ß induces expression of sterile transcripts. Based on our
in vitro studies, it appears that IL-2 is necessary, directly or
indirectly, for the switch to any C gene, and we
postulate that additional factors might be needed to direct the switch
to C4 or to the other IgA isotypes.
Another observation regarding C4 was that IgA4 plasma
cells were unevenly distributed throughout the small intestines. A
similar uneven distribution of IgA1 and IgA2 in humans has been
reported (6). Switching probably does not occur in the lamina propria
of the gut (7, 8). More likely, the lamina propria is populated by B
cells or plasma cells that have undergone switching in the PP or
elsewhere and have migrated into the lamina propria. The mechanism
resulting in localization of IgA4 plasma cells in one end of the small
intestines and not in the other end of the intestines, although the
total number of all IgA plasma cells appears to be similar, remains a
mystery. If homing is the reason for the uneven distribution, one would
have to postulate that receptors for IgA4-expressing cells are absent
in the lower part of the intestines. It is tempting to postulate that
the microbial flora of the gut play a role in this uneven distribution.
The number of microbes in the duodenum of the rabbit is relatively low,
whereas the ileum harbors a great variety of them (39). Microbial Ags
themselves might attract cells that express a particular IgA, or the
microbial flora may create a particular environment by activating
effector cells that attract cells expressing particular IgA isotypes.
One can also speculate that sequential IgA switching occurs in the
lower part of the small intestines, such that in the gut
IgA4-expressing cells undergo a second switch recombination to express
another IgA isotype. Sequential switching has been observed for several
Ig isotypes (40, 41, 42). C4 is the most 5' of the
C genes; therefore, a second switch to any of the 3'
C genes is possible. This scenario could explain the
absence of IgA4-expressing cells in the lower part of the gut, if one
postulates that an increasing number or variety of microbial Ags
directly or indirectly induces the sequential switch.
The results presented in this report suggest that the expression of
multiple C genes is regulated on different levels.
First, the I promotor regulates the expression of
sterile transcripts, a necessary first step for isotype switching to
occur, and individual cells can transcribe simultaneously several
I-C genes. Next, the local environment of
the different tissues regulates the expression of functional mRNA,
determining which of the transcribed C genes will be
rearranged. In addition, the data show that IgA plasma cells that
secrete a particular IgA isotype are distributed unevenly throughout
the small intestine, suggesting that IgA isotype expression is
influenced directly or indirectly by the gut flora or other Ags.
|
Footnotes |
|---|
2 The following sequences have been deposited in GenBank: accession numbers AF129765–AF129775.
3 Current address: Center of Genetics, University of Illinois, Chicago, IL 60612.
4 Address correspondence and reprint requests to Dr. K. L. Knight, Department of Microbiology and Immunology, Loyola University Chicago, 2160 South First Ave., Maywood, IL 60153. E-mail address:
5 Abbreviations used in this paper: PLN, popliteal lymph node; PP, Peyer’s patch; GALT, gut-associated lymphoid tissue; MLN, mesenteric lymph node; CD40L, CD40 ligand.
Received for publication November 9, 1998. Accepted for publication February 17, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
genes. EMBO J. 8:4041.[Medline]
genes as chimeric rabbit-mouse IgA antibodies. Proc. Natl. Acad. Sci. USA 86:7561.1 and C2 germ-line and mature mRNA transcripts in human peripheral blood B cells. J. Immunol. 153:1466.[Abstract]
1 and C2 germ-line transcripts. Int. Immunol. 3:1107. germline transcripts in the B lymphoma I.29 µ: Transforming growth factor-ß activates transcription of the unrearranged C gene. J. Immunol. 147:4374.[Abstract]
constant region gene by an ATF element and by novel transforming growth factor-ß1-responsive elements. J. Immunol. 149:2914.[Abstract]
1 and I2 germline promotor elements; proximal positive and distal negative elements may regulate the tissue specific expression of C1and C2 germline transcripts. Int. Immunol. 5:271.1 region contains a TGF-ß1 responsive enhancer and a putative recombination hotspot. Int. Immunol. 7:1191.1 and immunoglobulin heavy chain transcripts in single murine splenic B-cells. Mol. Immunol. 34:919.[Medline]
1 and Ig gene expression in murine B cells. IL-4 and the CD40 ligand-CD40 interaction provide distinct but synergistic signals. J. Immunol. 155:5637.[Abstract]
isotype switch in IL-4-treated human B lymphoblastoid cells: evidence for a sequential switch. J. Immunol. 149:1075.[Abstract]
via 4 in human B cells stimulated with IL-4 and hydrocortisone. J. Immunol. 151:4528.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  H. Spieker-Polet, P.-C. Yam, and K. L. Knight Functional Analysis of I{alpha} Promoter Regions of Multiple IgA Heavy Chain Genes J. Immunol., April 1, 2002; 168(7): 3360 - 3368. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-J. Shi, S.-R. Park, P.-H. Kim, and J. Stavnezer Roles of Ets proteins, NF-{{kappa}}B and nocodazole in regulating induction of transcription of mouse germline Ig {{alpha}} RNA by transforming growth factor-{beta}1 Int. Immunol., June 1, 2001; 13(6): 733 - 746. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. V. Volgina, M. Kingzette, S.-K. Zhai, and K. L. Knight A Single 3'{alpha} hs1,2 Enhancer in the Rabbit IgH Locus J. Immunol., December 1, 2000; 165(11): 6400 - 6405. |
