| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Institut National de la Santé et de la Recherche Médicale Unité 399, Marseille, France;
Institut de Médecine Tropicale du Service de Santé des Armées, Marseille, France;
Department of Pathology, University of Geneva, Geneva, Switzerland;
§
Institute for Animal Health, Pirbright, United Kingdom; and
¶
Medical Research Council Laboratory of Molecular Biology, Cambridge, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
20–120 nucleotide insertions
upstream of the coding region between exons 2 and 3, and/or 100–190
nucleotide deletions of exon 4. Sequence analysis revealed that these
aberrant mRNA species arose by alternative splicing due to the presence
in the NZW strain of a 794-bp sequence insertion in the second intron,
containing a cluster of short interspersed nucleotide elements. Both
the presence of sequence insertion and aberrantly spliced mRNAs were
specific to mice bearing the
Cd22a and
Cd22c alleles.
Up-regulation of CD22 expression after LPS activation appeared impaired
in Cd22a spleen cells (twice
lower than in Cd22b B cells).
Furthermore, we show that partial CD22 deficiency, i.e., heterozygous
level of CD22 expression, markedly promotes the production of IgG
anti-DNA autoantibodies in C57BL/6
(Cd22b) mice bearing the Y
chromosome-linked autoimmune acceleration gene, Yaa.
Taken together, these results suggest that a lower up-regulation of
CD22 on activated B cells (resulting from Cd22 gene
anomaly in Cd22a mice or from
CD22 heterozygosity in mutants obtained by gene targeting) is
implicated in autoantibody production, providing support for
Cd22a as a possible candidate
allele contributing to lupus susceptibility. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CD22 is a B cell-specific transmembrane glycoprotein with seven extracellular Ig-like domains, which apparently functions as an adhesion receptor to bind a series of sialic acid-dependent ligands expressed on all leukocyte classes, and as a coreceptor for the B cell Ag receptor (BCR) (reviewed in Ref. 12). Noteworthy, the cytoplasmic tail of CD22 contains three immunoreceptor tyrosine-based inhibitory motifs that are rapidly phosphorylated upon BCR cross-linking, and can bind the tyrosine phosphatase SHP-1, a negative regulator of signaling from the BCR (13, 14). A link between dysregulated CD22 expression and lupus-like autoimmune disease has been suggested by the finding that mice with a disrupted CD22 gene develop increased serum titers of autoantibodies to DNA and DNA-histone complex (15, 16).
In the mouse, CD22 has been defined by the Lyb-8 alloantigen system, Cd22a and Cd22b, encoding the CD22.1 and CD22.2 Ags, respectively (17, 18, 19). The Cd22a allele has been shown to be expressed in NZB, NZW, DBA, and PL/J strains, whereas the Cd22b allele is expressed in BALB/c, C57BL/6, CBA/N, and C3H/HeJ strains. Recently, we described a third allelic form, Cd22c, expressed in the autoimmune-prone BXSB strain and its parental strain, SB/Le (20). Noteworthy, these three alleles have striking differences in the most distal extracellular regions constituting the ligand-binding domains. Functional differences between CD22.1 and CD22.2 have been suggested based on in vitro experiments, showing remarkable enhanced BCR-evoked calcium influx in CD22.2-deficient variants of the WEHI-231 B cell line, which derive from a (BALB/c x NZB)F1 mouse (21).
Since intrinsic defects in the B cell lineage of lupus-prone mice, including the NZW strain, are likely to play a critical role in their predisposition to autoimmune disease (22, 23), it is important to define the possible contribution of the Cd22a allele in the genetic predisposition of NZW mice to autoimmune disease. In this study, we describe that in addition to the wild-type transcripts, Cd22a- and Cd22c-bearing mice also synthesize abnormally processed CD22 mRNA transcripts. The presence of aberrant transcripts in these mouse strains could be explained by insertion of intronic repetitive elements and induction of alternative splicing involving cryptic splice sites. Together with the observation of differential increase of CD22 surface expression on B cells following activation with LPS in Cd22a, and production of autoantibody in heterozygous CD22-deficient mice bearing the Yaa gene, the data strengthen the possibility of a role for CD22 in determining the susceptibility to lupus-like disease, and offer an explanation for the role of Cd22a in this respect.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
AKR/J, C3H/HeJ, C57BL/6J, CBA/J, DBA/2, MRL, NZB, NZW, and SWR mice were purchased from Harlan-France (Gannat, France). BXSB and SB mice were obtained from The Jackson Laboratory (Bar Harbor, ME). BALB/c mice were obtained from IFFA-Credo (Lyon, France). B6.Yaa males were established by repeated backcrosses (11). CD22-deficient mice were generated as described previously (15). CD22+/- animals were bred twice against B6 mice before breeding with either B6.Yaa males or B6 females in our own animal facility. The inheritance of Cd22 gene disruption was ascertained by assessing a decreased CD22 expression by peripheral blood B lymphocytes and by PCR amplification of the neomycin resistance gene using genomic DNA prepared from tail biopsies. Blood samples were collected every 2 mo from the age of 2 mo by orbital sinus puncture, and the sera were stored at -20°C until use.
RT-PCR, PCR, and DNA sequencing
First strand cDNAs were synthesized with an oligo(dT) primer and
2.5 µg of total RNA extracted from spleen, lymph node, bone marrow,
peritoneal cells, and thymocytes. RT-PCR fragments (500–600 bp)
spanning all the CD22-translated sequence were generated by using six
different sets of PCR primers, A, B, C, D, E, and F, described in Table I
. For amplification of the second intron
of the Cd22 gene, 100 ng of genomic DNA prepared from livers
of the various strains was amplified with exon 2 primer used for RT-PCR
described in set A, in combination with a primer hybridizing to exon 3
(set G). In addition to specifically amplifying abnormally processed
CD22 sequences containing the 18-bp intron 2 sequence, RT-PCR fragments
(600 bp) were generated by using a primer complementary to the 3'
flanking intron sequence of exon 2 in combination with exon 5 primer
used for RT-PCR described in set A (set H). For amplification of the
CD22-disrupted allele (800 bp), primers Neo-1
(5'-GCCCAGCGTCTTGTCATTGGCG-3') and Neo-2
(5'-GGGTAGCCAACGCTATGTCC-3') hybridizing to the neomycin resistance
gene were used. In semiquantitative RT-PCR, cDNAs were adjusted to
identical concentrations by use of primers hypoxanthine
phosphoribosyltransferase-1 (5'-GTTGGATACAGGCCAGACTTTGTTG-3') and
hypoxanthine phosphoribosyltransferase-2
(5'-GAGGGTAGGCTGGCCTATAGGCT-3'). Oligonucleotides were purchased from
Life Technologies (Rockville, MD).
|
RT-PCR and PCR products generated with primer pairs A and G, respectively, were ligated into pGEMT plasmid (Promega, Madison, WI) and cloned in the JM109 bacterial strain. Nucleotide sequences were determined by the dideoxynucleotide chain-terminating method by an ABI PRISM 310 DNA sequencer (PE Applied Biosystems, Foster City, CA), and have been deposited in EMBL Nucleotide Sequence Database. Free energy of RNA-folding forms was obtained using RNA mfold version 3.0 server by Zuker and Turner (24).
Cell preparation and activation
For analysis of CD22 expression on splenic B cells, Thy-1.2-positive cells were first depleted by magnetic separation after adsorption on anti-Thy-1.2-coated beads (Dynal A.S., Oslo, Norway). Purity of B cells was 85–95% B220+, as determined by flow cytometry. Enriched B cell suspensions were plated in 24-well plates at 2.106 cells in 2 ml medium consisting of RPMI 1640 with 10% FCS, Kanamycin, L-glutamine, 2-ME, supplemented with 10% supernatant of the X63-mIL-4 transfectant (kindly provided by Dr. Hajime Karasuyama, Tokyo, Japan) (25), and in presence of 40 µg/ml LPS (Escherichia coli 055:B5; Sigma, St. Louis, MO). CD22 expression was analyzed by flow cytometry on days 0 and 1.
mAbs and cytofluorometric analysis
The following mAbs were used: 14.8 (anti-B220), M5/114 (anti-I-A, I-E), and NIM-R6 (anti-CD22). For immunofluorescence analysis, mAbs were purified on protein G-Sepharose and coupled to biotin; staining the cells was done as described by Rolink et al. (26). PE-labeled streptavidin was purchased from Immunotech (Marseille, France). Spleen and PBMCs from 2–4-mo-old mice were first stained with NIM-R6 anti-murine CD22 (18), followed by PE-conjugated streptavidin. A total of 104 events was analyzed with a lymphocyte gate, as defined by light scatter with a FACScalibur (Becton Dickinson, Mountain View, CA), and data were processed by using CellQuest software (Becton Dickinson).
Serological assay
The presence of serum IgG anti-DNA, anti-chromatin, and anti-DNP Abs was assessed by ELISA, in which heat-denatured calf thymus DNA (Sigma), chicken chromatin, and DNP-BSA, respectively, were used as coating Ag, as previously described (27). Results are expressed in titration units (U/ml) in reference to a standard curve obtained from a pool serum of 3- to 4-mo-old MRL-lpr/lpr mice.
Genetic mapping using simple sequence length polymorphism
Genotypes were determined by PCR using selected simple length polymorphism markers purchased from Research Genetics (Huntsville, AL). DNA from B6, 129, and CD22-deficient backcross mice was extracted from tail biopsies. Amplification of the simple sequence repeat was achieved by PCR with a thermal RoboCycler (Stratagene, La Jolla, CA), as described previously (10). The relative positions of the markers were obtained from the Mouse Genome Database (MGD) mouse genetic map via the internet at http://www.informatics.jax.org (The Jackson Laboratory).
Statistical analysis
Statistical analysis was performed with the Wilcoxon two-samples test. p > 5% was considered insignificant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mouse Cd22 gene consists of 15 exons, in which exons 1
and 2 and part of exon 3 encode the 5'-untranslated region (UTR) of the
full-length cDNA (19). To analyze more closely the CD22
mRNA transcripts, RT-PCR was performed with five different CD22 primer
pairs covering exons 2–15 (see Table I
) on RNA isolated from the
spleens of 3- to 4-mo-old NZW (Cd22a) and
C57BL/6 (Cd22b) mice. RT-PCR with the CD22
primer pairs B, C, D, and E yielded the expected sizes of amplified DNA
products covering exons 4–7, 6–9, 9–13, and 10–15, respectively, in
both strains of mice (Fig. 1A). Amplification of cDNA
spanning exons 2–5 with primer pair A generated a single 604-bp DNA
product in B6 mice and a slightly smaller PCR product in NZW mice, as
expected, from the 18-bp deletion in exon 4 of the
Cd22a allele (19). However,
in addition to this fragment, a second 20–100-bp larger band was
observed in NZW mice. Estimation on ethidium bromide staining of
PCR-amplified products indicated that the amount of the aberrant CD22
mRNAs was even higher than that of the normal mRNA. Essentially
identical results were obtained with RT-PCR analysis on RNA prepared
from various lymphoid organs (spleen, lymph nodes, peritoneal cavity,
and bone marrow), but not from thymus, as shown in Fig. 1B.
Notably, these abnormal CD22 mRNA transcripts were also detected, when
CD22 primer pair F to amplify exons 1–3 was used for PCR amplification
(Fig. 1C).
|
To determine the nature of the DNA products obtained from NZW mice
after RT-PCR amplification with CD22 primer pair A, the PCR products
obtained from NZW spleen were subcloned and sequenced. Nucleotide
sequencing revealed the existence of at least nine different CD22 mRNA
species in the spleen of NZW mice (Fig. 2), in addition to the 586-bp fragment of
the expected sequence reported for the
Cd22a allele (19). A
schematic representation of the CD22 mRNA variants is depicted in Fig. 3, showing the relative location of
insertion and/or deletion sequences as well as the genomic structure of
intron 2 in the NZW strain. Variants I–V displayed larger PCR
products, 703-, 690-, 672-, 635-, and 604-bp size, which contained
structurally related sequence insertions of 117, 104, 86, 49, and 18 bp
at the junction between exons 2 and 3, respectively. The size
difference between these clones and the normal CD22 corresponded to the
unusual longer band described above. As shown in Fig. 2, the 104- and
18-bp-long sequence insertions in variants II and V, respectively,
contain an AUG codon surrounded by an A in position -3 with respect to
the A of the AUG, which might create a context suitable for initiation
of translation (28). Consequently, translation from AUG of
the 104-bp sequence insertion in variant II would create a short open
reading frame (ORF) terminating 96 bases 3' to the AUG codon of CD22.
Based on previous studies (29), this distance would
probably permit reinitiation of CD22 translation, but not with an
optimal efficiency. Insertion in variant V of the 18-bp sequence would
create an in-frame nucleotide sequence coding for 11 additional
N-terminal aa. Noteworthy, the 117-, 104-, and 86-bp sequences in
variants I–III, creating stable hairpin structures with a free energy
of -41, -45, and -37 kcal/mol, respectively, could have
negative effects on translation by blocking the migration of 40S
ribosomes (28).
|
|
). These
amplified PCR products, 535-, 533-, 486-, and 449-bp size, which
contained the above-described 104-, 49-, 86-, and 18-bp sequence
insertions, respectively, displayed either 102-, 155-, or 186-bp
deletion in exon 4 (Figs. 2 and 3). The 155-bp deletion occurring in
variant VI and IX causes a reading frameshift, which introduces a stop
codon at position 157 of exon 4. In contrast, the transcripts with 102-
and 186-bp deletions at positions 259 and 124 of exon 4 (variants VII
and VIII) are in frame and may create truncated CD22 proteins with 34-
or 62-aa deletion in Ig-like domain 1. The levels of CD22 mRNA with
exon 4 sequence deletions are probably too low to produce detectable
levels of truncated CD22 protein. All clones isolated in this study
were homologous to the mouse CD22.1 allele in exon 3 and in the partial
30- and 112-bp sequences of exons 2 and 5, respectively (data not
shown). Short interspersed nucleotide element B1 detected in intron 2 of CD22 promotes aberrant pre-mRNA splicing in NZW mice
Because sequence insertions between exons 2 and 3 corresponded
exactly to exon-intron boundaries in CD22, a genomic PCR analysis using
primer pair G complementary to sequences in exons 2 and 3 was performed
in an attempt to identify the possible origin of the additional
sequences. The PCR products generated in NZW and B6 mice were markedly
different in size, 2 and 1.2 kb, respectively. Cloning and
sequencing of these fragments confirmed the insertion of a 794-bp
sequence in the NZW gene, located 711 bp downstream from the exon
2-intron 2 junction (Fig. 4). The insert
was unrelated to known CD22 sequences, indicating that it did not arise
from duplication of adjacent regions. Database searches revealed
significant sequence homology of 354 bp, representing 44.8% of the
794-bp sequence insertion, with ID-, B1-, and B4-related families of
short interspersed nucleotide elements (Table II). A 62-bp-long sequence intervening
between the ID and B1 elements appeared to fit into the satellite class
of DNA-repeated elements. The 5' transition site from homologous to
nonhomologous intron 2 sequences between NZW and B6 strains is
localized at the border of the ID element. In addition to the sequence
insertion mentioned above, the NZW intron 2 differs from that of the B6
strain in GA and CA simple sequence repeat units, and 20 nucleotide
substitutions.
|
|
).
The 117-bp (variant I), 86-bp (variants II, III, VI, VIII), and 49-bp
(variants IV, VII) cDNA sequences derive from differential splicing of
the B1 element. The 18-bp cDNA insertion sequence (variants II, V, VI,
IX) derive from the intron 2 sequence immediately downstream from exon
2, in which a cryptic splice donor site is likely to be activated by
the three 3' splice acceptor sites of the B1 element. It should be
mentioned that the transcriptional orientation of this particular B1
repeat is opposite to that of the CD22 transcript.
|
To survey the presence of intron 2 sequence insertion in various
strains of mice, genomic DNA were amplified across intron 2. As shown
in Fig. 5A, a 2-kb PCR
product was generated not only in all
Cd22a strains tested (NZB, NZW, DBA/2, and
AKR), but also in Cd22c strains (BXSB and
SB/Le), whereas a 1.2-kb product was amplified in
Cd22b strains (BALB/c, B6, C3H/HeN, CBA/J,
SWR, and MRL). Restriction enzyme analysis on the PCR products and
additional PCR analysis using different sets of primers confirmed that
the difference was due to the same sequence insertion, as in NZW mice
(data not shown). In addition, the presence of aberrant CD22 mRNA
transcripts was similarly detected in
Cd22a and
Cd22c mice, but not in the
Cd22b strains mice tested (data not
shown).
|
), and is present in both the NZW (Cd22a)
and B6 (Cd22b) alleles (Fig. 4). Analysis
of RT-PCR products amplified from the spleen of NZW mice confirmed the
presence of two aberrant CD22 mRNA species carrying the 5' intron 2
sequence (Fig. 5B). In contrast, no PCR products were
amplified from RNA isolated from BALB/c and B6 spleens using this
primer pair, although a 600-bp DNA product was amplified with the
CD22 primer pair A amplifying exons 2–5. Lower up-regulation of CD22 surface expression by Cd22a splenic B cells compared with Cd22b splenic B cells following activation by LPS
We next sought to investigate CD22 surface expression on B cells
of Cd22a-carrying mice compared with
Cd22b mice. The rat mAb NIM-R6 apparently
recognizes a framework epitope shared by the CD22.1 and CD22.2 allelic
forms of CD22. Using this mAb, it has previously been shown that
splenic B cells from DBA/2 (CD22.1) mice express CD22 proteins on the
surface of B cells at a level comparable with those from BALB/c
(CD22.2) mice (19). A careful assessment, however, showed
that the staining intensity of the NIM-R6 mAb on B cells from NZW
(CD22.1) spleen was slightly, but significantly lower (geometric mean
42.9 ± 2.5), as compared with that of B6 (CD22.2) mice (geometric
mean 49.5 ± 1.7; p < 0.01) (Fig. 6). Although the expression levels of
CD22 showed only weak changes during the first 12 h of stimulation
with LPS and IL-4 (data not shown), splenic B cells from NZW mice
exhibited a moderate 1.7-fold increase of CD22 surface expression
(geometric mean 74.6 ± 1.8), which contrasted with a higher
up-regulation (2.7-fold) in B6 splenic B cells (geometric mean
133.4 ± 1.7; p < 0.01) (Fig. 6). Similar
differences were still observed at 48 h (geometric mean 96.1
± 5.8 in NZW mice; 140.6 ± 4 in B6 mice; p <
0.01). In both NZW and B6 cultures stimulated with LPS, however, the
extent of up-regulation of MHC class II molecules was comparable (data
not shown). To further study whether the lower levels of CD22
expression by NZW B cells are related to
Cd22a allele rather than to other genetic
systems that may also vary between NZW and B6, we performed an analysis
of the expression levels of CD22 on the B cells of DBA/2 (CD22.1) and
BALB/c (CD22.2) mice. Accordingly, unstimulated and LPS-activated B
cells from DBA/2 mice (geometric means 38.4 ± 2.3 and 60.7
± 4.2, respectively) and BALB/c mice (geometric means 49.7 ± 3.2
and 125.5 ± 5.2, respectively) were stained with NIM-R6 mAb in a
manner similar to those of NZW and B6 mice, supporting that the
difference of expression of CD22 on B cells of NZW vs B6 is due to
allelic variation in CD22.
|
Interaction between partial CD22 deficiency and Yaa gene-mediated autoimmunity-accelerating effect
To ascertain whether partial deficiency of CD22 is able to promote
autoimmunity in lupus-prone mouse strains, we explored the effect of a
2-fold difference in the CD22 expression in B6 mice bearing the
Yaa mutation. Since heterozygous
CD22+/--deficient mice do not spontaneously
produce significant titers of IgG anti-DNA autoantibodies early in
life (16), we produced B6.Yaa
CD22+/- mice and followed the development of IgG
anti-DNA. Consistent with previous analyses (15),
mature B cells in CD22+/- mice bearing the
Yaa mutation expressed one-half of the normal cell surface
density of CD22 (data not shown). At the age of 8 mo, none of the
CD22+/+ and CD22+/- male
mice without the Yaa gene developed elevated serum levels of
IgG anti-DNA autoantibodies (Fig. 7A). In contrast, increased
levels of IgG anti-DNA Abs in mice bearing the Yaa gene
were observed in 10 of 20 CD22+/- mice, and in
only 1 of 16 CD22+/+ male counterparts
(p < 0.002). Whereas none of the
CD22+/- and CD22+/+ mice
lacking the Yaa gene produced significant titers of IgG
anti-chromatin Abs (1.8 ± 1.6 and 1.4 ± 1.2 U/ml,
respectively), the CD22+/- Yaa males
produced elevated levels of IgG anti-chromatin Abs with an overall
penetrance of 50%, but that were not different from
CD22+/+ Yaa littermates (18.5 ± 14.6
and 24.3 ± 14 U/ml, respectively). Analysis of IgG anti-DNP
Abs, as an indicator of polyclonal/polyreactive IgG production,
revealed that the number of CD22+/- Yaa
mice (12/20) having elevated serum levels of IgG anti-DNP Abs was
slightly but significantly higher than CD22+/+Yaa mice (6/16; p = 0.02) (Fig. 7B).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Aberrant CD22 transcripts result from insertion of a B1 (Alu-equivalent) repetitive element in the Cd22a and Cd22c alleles
The region responsible for the occurrence of aberrant CD22 mRNAs in Cd22a and Cd22c mice was identifed in intron 2 as a genomic sequence insertion consisting of clustered short interspersed nucleotide elements (B1-, B4-, and ID-related families) not present in the mouse strains carrying the Cd22b allele. Since an inserted B1 element provides alternative splice sites, we propose that in analogy to the B2/C4 gene model (34), this element activates pre-mRNA processing of silent intron 2 and exon 4 sequences. The B1 sequence is the most highly repeated element in rodents, approximately 105 copies being dispersed throughout the mouse genome. Both B1 sequences and related human Alu sequence families are homologous to regions of the 7SL RNA, a highly conserved RNA species from which they are likely to have derived, found associated with proteins to form the signal recognition particle in the cytoplasm (35). Compelling evidence for the role played by these elements in the restructuration of the genome during evolution and rearrangement within genes stems from the observation that some of them are located at transition sites between homologous and nonhomologous sequences of duplicated genes (36, 37). Although the B1 sequences are frequently found in introns, the occurrence of aberrant pre-mRNA splicing caused by B1 elements is not a common phenomenon. The particular B1 repeat within CD22 intron 2, however, is unusual in that it does not exhibit the preferential orientation within RNA polymerase II transcription units. Furthermore, alternative splicing of the B1 element in CD22 may be contributed by use of a 5' donor splice site overlapping one of the flanking direct repeats. These sequences consist of a duplication of a short stretch of the target site of insertion, and thus do not originate from the B1 element. Identification of this genetic defect in the Cd22a and Cd22c strains of mice is significant for two reasons. First, it demonstrates the potential of the nonviral B1 family of retroposons to affect the structure and expression of genes. Second, it provides evidence for polymorphisms in the flanking region of CD22, in addition to those described in the CD22 ORF (18, 19, 20). In this regard, studies aimed at determining the form of this polymorphism in the genus Mus revealed that the intron 2 sequence insertion of CD22 was distributed across three species, Mus spretus, Mus macedonicus, and Mus spicilegus, and four different Mus musculus subspecies, but not in the Mus musculus domesticus (unpublished observations). Therefore, an actual deletion event seems to have occurred within the nascent M. m. domesticus subspecies, leading to an apparently more functional CD22.2 allele.
How may the 5'-UTRs of the aberrant CD22 transcripts influence CD22 gene expression?
Kozak (28) has examined factors in the 5'-UTR that promote efficient translation, including the observation that most eukaryotic mRNAs do not have AUG codon upstream of the translation initiation site of the major ORF. Presence of AUG codons, followed rapidly by termination codon, creates short upstream ORF that may lead to abortive translation process, resulting in reduced protein synthesis. Translation efficiency may also be influenced by differences in the secondary structures of the 5'-UTRs of distinct transcripts, such as may exist in the Cd22a transcripts; insertion into the 5'-UTR of a stem-and-loop structure may profoundly inhibit translation. Presence of aberrant transcripts, such as those resulting from the usage of cryptic splice sites in exon 4, detected among the NZW CD22 cDNAs, may also lead to synthesis of truncated, inefficient proteins. It should be noted in this respect that the actual levels of inappropriately spliced mRNAs may be underevaluated in our analysis, since these abnormal transcripts may have a lower stability.
Presence of aberrant CD22 transcripts is associated with decrease of CD22 protein membrane expression and modulation
Unstimulated Cd22a B cells express significantly lower levels of surface CD22, and most remarkably, do not up-regulate them following the LPS activation, to the same levels as Cd22b B cells. We cannot formally exclude the possibility that the observed differences in the CD22 expression on B cells might result from a putative difference in the reactivity of the NIM-R6 mAb with CD22.1 and CD22.2 allelic forms. However, the facts that the amino acid sequence in extracellular Ig-like domains 3–6 is identical between these two allelic forms of CD22 (19), and that the epitope recognized by NIM-R6 mAb is apparently located in this region (38) strongly argue against this possibility. It should be also mentioned that BXSB B cells bearing the Cd22c allele most likely share the same defect, although this cannot be proved because of a limited staining with available anti-CD22 mAbs, which do not allow a direct comparison with Cd22b B cells. Clearly, forthcoming studies are necessary to establish a direct link between the dysregulated pre-mRNA processing and CD22 expression.
Decreased expression of CD22 favors the development of autoimmunity when associated to susceptibility gene (the Yaa mutation)
The functional importance of dysregulated CD22 expression was illustrated by the finding that partial CD22 deficiency, i.e., a heterozygous level of CD22 expression, results in greatly accelerated IgG anti-DNA autoantibody production in presence of the Yaa gene. It is also significant that the CD22-disrupted allele appeared to affect IgG anti-DNP Ab titers, indicating that the increased anti-DNA Ab production may reflect some form of polyclonal activation. Since it has been shown that CD22-deficient B cells become hypersensitive to BCR cross-linking (15, 39, 40, 41), and that the Yaa gene abnormality expressed in B cells most likely promotes BCR-mediated B cell activation (42), one can speculate that the combined action of partial CD22 deficiency and Yaa mutation could trigger lupus-like autoimmune responses even in B6 mice without apparent lupus susceptibility. Our results, thus, strongly suggest that dysregulation of responsiveness to BCR signaling is of central importance in the development of lupus-like autoimmune disease. This notion is supported by the previous findings that the production of anti-DNA autoantibodies was induced in mice deficient in Lyn or SHP-1, and in mice overexpressing CD19, a BCR-associated signaling molecule, in which B cells become abnormally hyperresponsive to antigenic stimulation (43, 44, 45, 46, 47, 48, 49). In this regard, it is important to determine whether the threshold of B cell activation is lowered and whether B cells are hyperactive in mice with the Cd22 a or c allele, as compared with those with b allele. Since these features are likely to be influenced by other genetic systems that vary between mouse strains, particularly when comparing the NZW and B6 strains (9), the use of congenic recombinant mice should help address this question. In addition, the demonstration of a gene dosage effect of CD22 on the development of autoimmune responses is also consistent with other experiments performed with mice bearing a single copy of CD22, showing that an accumulation of 2-fold differences in the Lyn and SHP-1 molecules critically affects BCR signaling threshold leading to spontaneous B cell hyperactivity (50).
CD22: a candidate gene for murine lupus-like multigenic autoimmune disease
The aberrant splicing of Cd22a mRNAs represents the first description of a polymorphism functionally relevant to the mapping of a lupus-susceptibility locus on chromosome 7. This locus, designated Sle3, was initially linked with nephritis in (B6 x NZM) x NZM mice and was shown to be NZW in origin (7, 51). In addition, the presence of a second peak located in the vicinity of Cd22 at the centromeric end of Sle3 for linkage with anti-DNA autoantibody production (7) is consistent with results from a B6 x (NZW x B6.Yaa)F1 backcross progeny (10). When bred into B6 mice, Sle3 causes the production of IgG anti-DNA Abs as well as lupus nephritis, and also impacts T cell activation (9, 52). The multitude of phenotypes associated with this interval most likely represents the combined actions of distinct lupus-susceptibility genes. This would explain that Sle3 congenic homozygous autoimmune phenotype is stronger than that of CD22-deficient mice, which develop high titers of IgG anti-DNA Abs after 8 mo and do not develop renal failure (16). In supporting evidence for the presence of several lupus-susceptibility genes within this interval on chromosome 7, a locus overlapping Sle3, designated Lbw5, contributes apparently to a distinct phenotype in (NZB x NZW)F2 mice, i.e., mortality but not autoantibody production or glomerulonephritis (8).
Concluding remarks
In conclusion, the data reported in this work should prove useful as a basis for further studies that will focus on analyzing the chromosome 7 lupus-susceptibility locus. Introgression of smaller intervals in congenic strains is in progress and will allow us to test whether the sequence variations in the Cd22 gene have an effect on the development of murine lupus-like disease. Clearly, further assessment of functional capacities of the allelic forms of CD22, as a negative regulator of BCR signaling and as an adhesion receptor, and its own ligand, should help understand a potential role of the CD22 polymorphism in the development of autoimmune diseases and other diseases in which B cell function is dysregulated.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: The Scripps Research Institute, Department of Immunology, 10550 North Torrey Pines Road, La Jolla, CA 92037.
3 Current address: Leukosite Inc., 215 First Street, Cambridge, MA 02142.
4 Address correspondence and reprint requests to Dr. Luc Reininger, Institut National de la Santé et de la Recherche Médicale Unité 399, Faculté de Médecine, 27 boulevard Jean Moulin, F-13385 Marseille cedex 05, France.
5 Abbreviations used in this paper: NZB, New Zealand Black; BCR, B cell Ag receptor; NZM, New Zealand Mixed; NZW, New Zealand White; ORF, open reading frame; SLE, systemic lupus erythematosus; UTR, untranslated region.
Received for publication January 10, 2000. Accepted for publication June 23, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1 upon B cell activation. J. Exp. Med. 183:547.This article has been cited by other articles:
|
|
 |
|
  L. C. Watson, C. S. Moffatt-Blue, R. Z. McDonald, E. Kompfner, D. Ait-Azzouzene, D. Nemazee, A. N. Theofilopoulos, D. H. Kono, and A. J. Feeney Paucity of V-D-D-J Rearrangements and VH Replacement Events in Lupus Prone and Nonautoimmune TdT-/- and TdT+/+ Mice J. Immunol., July 15, 2006; 177(2): 1120 - 1128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Pisitkun, J. A. Deane, M. J. Difilippantonio, T. Tarasenko, A. B. Satterthwaite, and S. Bolland Autoreactive B Cell Responses to RNA-Related Antigens Due to TLR7 Gene Duplication Science, June 16, 2006; 312(5780): 1669 - 1672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Varki and T. Angata Siglecs--the major subfamily of I-type lectins Glycobiology, January 1, 2006; 16(1): 1R - 27R. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Nitschke, F. Lajaunias, T. Moll, L. Ho, E. Martinez-Soria, S. Kikuchi, M.-L. Santiago-Raber, C. Dix, R. M. E. Parkhouse, and S. Izui Expression of aberrant forms of CD22 on B lymphocytes in Cd22a lupus-prone mice affects ligand binding Int. Immunol., January 1, 2006; 18(1): 59 - 68. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kikuchi, H. Amano, E. Amano, L. Fossati-Jimack, M.-L. Santiago-Raber, T. Moll, A. Ida, B. L. Kotzin, and S. Izui Identification of 2 major loci linked to autoimmune hemolytic anemia in NZB mice Blood, August 15, 2005; 106(4): 1323 - 1329. [Abstract] [Full Text] [PDF]
|
|
, wild-type
CD22+/+ littermates. Ab levels are shown for peak levels
before 8 mo of age. Results are expressed in U/ml by reference to a
standard curve obtained with a serum pool of 3- to 4-mo-old
MRL-lpr/lpr mice. Lines are median values for each group
of 14–20 male mice. The broken horizontal line represents the mean + 3
SD for autoantibody levels in age-matched B6 males.