Human Molecular Genetics, 2001, Vol. 10, No. 25 2917-2931
© 2001 Oxford University Press
DNA methyltransferase 3B mutations linked to the ICF syndrome cause dysregulation of lymphogenesis genes
Human Genetics Program and Department of Biochemistry, Tulane Medical School, New Orleans, LA 70112, USA, 1Department of Microbiology and Immunology, Tulane Medical School, New Orleans, LA 70112, USA, 2Department of Pediatrics and 3Department of Human Genetics, University Medical Center St. Radboud, Nijmegen, The Netherlands, 4Institute of Human Genetics, Charitéé, Berlin, Germany, 5Ludwig-Maximilians-Universitat, Munchen, Germany, 6Wilhelm Johannsen Centre for Functional Genome Research, Department of Medical Genetics, Institute of Medical Biochemistry and Genetics, University of Copenhagen, Denmark and 7Department of Pediatric Hematology Oncology, Box 0656, MSRB I, Room A520C, University of Michigan, Ann Arbor, MI 48109, USA
Received August 17, 2001; Revised and Accepted October 8, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ICF (immunodeficiency, centromeric region instability and facial anomalies) is a recessive disease caused by mutations in the DNA methyltransferase 3B gene (DNMT3B). Patients have immunodeficiency, chromosome 1 (Chr1) and Chr16 pericentromeric anomalies in mitogen-stimulated lymphocytes, a small decrease in overall genomic 5-methylcytosine levels and much hypomethylation of Chr1 and Chr16 juxtacentromeric heterochromatin. Microarray expression analysis was done on B-cell lymphoblastoid cell lines (LCLs) from ICF patients with diverse DNMT3B mutations and on control LCLs using oligonucleotide arrays for approximately 5600 different genes, 510 of which showed a lymphoid lineage-restricted expression pattern among several different lineages tested. A set of 32 genes had consistent and significant ICF-specific changes in RNA levels. Half of these genes play a role in immune function. ICF-specific increases in immunoglobulin (Ig) heavy constant µ and
RNA and cell surface IgM and IgD and decreases in Ig
and Ig
RNA and surface IgG and IgA indicate inhibition of the later steps of lymphocyte maturation. ICF-specific increases were seen in RNA for RGS1, a B-cell specific inhibitor of G-protein signaling implicated in negative regulation of B-cell migration, and in RNA for the pro-apoptotic protein kinase C eta gene. ICF-associated decreases were observed in RNAs encoding proteins involved in activation, migration or survival of lymphoid cells, namely, transcription factor negative regulator ID3, the enhancer-binding MEF2C, the iron regulatory transferrin receptor, integrin ß7, the stress protein heme oxygenase and the lymphocyte-specific tumor necrosis factor receptor family members 7 and 17. No differences in promoter methylation were seen between ICF and normal LCLs for three ICF upregulated genes and one downregulated gene by a quantitative methylation assay [combined bisulfite restriction analysis (COBRA)]. Our data suggest that DNMT3B mutations in the ICF syndrome cause lymphogenesis-associated gene dysregulation by indirect effects on gene expression that interfere with normal lymphocyte signaling, maturation and migration. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ICF (immunodeficiency, centromeric region instability and facial anomalies) is a rare chromosome instability syndrome that is autosomal recessive. It is the only disease shown to be associated with mutations in a DNA methyltransferase gene (1,2). Chromosomal anomalies are targeted specifically to the vicinity of the centromere (pericentromeric region) of chromosome 1 (Chr1) or Chr16 in mitogen-stimulated lymphocytes and B lymphoblastoid cell lines (LCLs) from these patients (3,4). The only other types of chromosomal aberrations associated with ICF are rare abnormalities in the pericentromeric region of Chr9 (3) and telomeric associations, which have been reported so far only in two ICF LCLs (4).
Severe immunodeficiency and high frequencies of pericentromeric decondensation and rearrangements in Chr1 and Chr16 upon karyotype analysis of blood samples are diagnostic for this syndrome (3). Also, in all studied ICF tissue samples, hypomethylation has been found in satellite 2 (Sat2) DNA, the main component of centromere-adjacent (juxtacentromeric) heterochromatin, where the ICF-specific decondensation is seen (4–6). Sat2 DNA is highly methylated in normal postnatal somatic tissues (4). Various other abnormalities are often seen in ICF patients, such as, facial anomalies, intestinal problems, growth retardation and neurological dysfunction (3). Also, we recently reported that ICF LCLs are hypersensitive to the lethal effects of radiation but without deficient cell cycle checkpoint responses (7).
In order to elucidate the changes in gene expression that account for the symptoms of ICF and to understand the relationships between targeted DNA hypomethylation and gene expression, we compared RNA from LCLs of ICF patients and normal controls using an oligonucleotide microarray. Among the genes that showed significant ICF-specific decreases or increases in RNA are five Ig genes and nine genes involved in the activation or late maturation of lymphocytes. The results of our study can help explain the heretofore enigmatic nature of the immunodeficiency in ICF.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Confirming the ICF-specific phenotype of the analyzed B-cell lines
For the microarray analysis of ICF-specific gene expression, six ICF and five control polyclonal B-cell LCLs were used. The choice of B-cell lines was predicated upon the central role of B cells in abnormal Ig production in ICF and the inaccessibility of uncultured B cells from multiple patients’ blood samples because only slightly more than 30 ICF patients have been reported in the last few decades and few reach adulthood. Despite their maintenance in cell culture for extended periods of time, the LCLs exhibited highly significant ICF-specific differences in RNA levels as described below. Moreover, the ICF LCLs displayed the diagnostic karyotype of mitogen-stimulated ICF lymphocytes, namely, much higher levels of the ICF-specific chromosomal abnormalities in the pericentromeric regions of Chr1 and Chr16 than control LCLs (Table 1) and hypomethylation of Sat2 DNA in somatic cells (see below). The ICF LCLs were from five unrelated ICF patients with various mutations in DNMT3B (Table 1). The five control LCLs were from unaffected unrelated individuals (Normal A and D) or phenotypically normal parents of ICF patients (Normal B, G and C). The LCLs from the ICF patients’ parents had a similar passage history as the patients’ LCLs. All the ICF LCLs exhibited telomeric associations previously reported in two examined ICF LCLs (4).
|
We also assayed these LCLs for hypomethylation of the major DNA component of juxtacentromeric heterochromatin of Chr1, Sat2 DNA, which is characteristic of ICF. BstBI digests of the LCL DNAs were hybridized in Southern blots with a Chr1 Sat2-specific probe. All of the ICF LCLs exhibited similarly hypomethylated Sat2 DNA compared with the control LCLs, as previously demonstrated for ICF B and ICF C relative to the LCLs from their parents and to Normal A and Normal D LCLs (4). ICF B and ICF C LCLs had been checked for overall genomic 5-methylcytosine (m5C) levels by analysis of DNA digested to deoxynucleosides and shown not to be detectably m5C-deficient relative to normal LCLs (4). However, we had demonstrated that DNA from the one assayed uncultured ICF tissue, ICF brain DNA, had
7% less m5C than normal brain DNA samples (m5C to C ratios of 0.043 and 0.0455, respectively) (4).
Overview of altered gene expression in ICF LCLs
RNA from ICF and control LCLs was compared using microarrays containing oligonucleotide probes for approximately 5600 different human genes. In analyzing the data from the arrays, we looked for genes whose RNAs were significantly over- or under-represented in ICF versus control LCLs. We required that at least three of the LCLs give significant expression of that transcript (P < 0.01 comparing the perfect match to mismatch probes). There were 3881 of such probe sets, which indicates transcription of more than half of the genes on the array by at least some of the LCLs in this work. Forty-five of these displayed differences between the ICF and control LCLs at a significance level of P < 0.05 and a fold change (FC) of >2.0 or an FC > 1.5 and a P < 0.01. We also compared RNAs from five normal LCLs and three non-LCL cell lines (triplicate RNA samples from a lung carcinoma cell line, A549; a brain tumor cell line, U118; and a colorectal cancer cell line, Lovo). In an ANOVA of these samples there were 510 probe sets that were restricted in their expression to LCLs, with an FC 
1.5 and P < 0.05 for all three LCL versus non-LCL comparisons. Out of the 45 probe sets displaying ICF versus normal LCL differences, 18 were among the 510 LCL-specific probe sets. That an excess of the probe sets which show up- or downregulation in ICF are specifically expressed in LCLs relative to non-lymphoid cell lines is significant at P < 0.00001 for Fisher’s exact test (two-sided).
In Tables 2–4, we show data for 32 genes with significant ICF-specific differences in RNA levels. The remaining 13 genes were not included in the tables because their expression levels were not sufficiently high (>200 arbitrary units as the average expression for the ICF or control LCLs) or reproducible among the ICF or control LCLs (at least three of the five patients showing an FC in the same direction and the average for the ICF LCLs or control LCLs not strongly influenced by one or two outliers). These 32 genes were selected using just these parameters without regard to functionality or tissue specificity. Many of these genes had functions that could be related to the ICF phenotype and several of these showed consistent results from replicate or related probe sets. For example, the transferrin receptor gene, TFRC, which is involved in lymphocyte activation as well as the general regulation of iron metabolism, showed an average of 2.0–2.4 times less RNA signal in ICF than in normal LCLs for three different TFRC probe sets in the arrays (Table 3). The tumor necrosis factor receptor superfamily member 17 (TNFR17) gene, which is implicated in B-cell activation, and its overlapping antisense (AS) gene (8,9) gave concordant ICF-specific differences in their RNAs (Table 3; Fig. 1) consistent with the finding that LCLs which express TNFR17 RNA usually also express TNFR17 AS RNA (8). GUCY1B3 and GUCY1A3, both of which exhibited higher RNA levels in ICF than normal LCLs (Table 4; Fig. 1), encode two subunits of the same enzyme and are generally coordinately expressed (10). Moreover, the inverse relationship between ICF-associated increases in RNA levels for the Ig heavy chains IgHM and IgHD and decreases for IgHG3, IgHA1 and IgHA2 are expected for these classes of Ig genes given the genomic recombination in Ig class switch regions which deletes IgHM and IgHD exons during generation of mature heavy-chain IgG and IgA genes.
|
|
|
|
We controlled for differences among the LCLs other than their status as ICF or control cell lines. Neither differences in ploidy (Table 1) nor growth rate could explain ICF-associated increases or decreases in RNA levels and X-linked genes did not show ICF-specific differences in expression. We checked that the ICF-specific expression of genes in Tables 3 and 4 was not just a consequence of the cells overexpressing IgHM and IgHD RNA. Because Normal A had a similar pattern of Ig RNA expression to that of the ICF LCLs (Table 5), we determined whether this control cell line had RNA levels similar to those of the ICF LCLs for any of the genes in Tables 3 or 4. The only genes for which this was found to be the case were those encoding GUCY1B3 and GUCY1A3 subunits (Table 4; Fig. 1). The mean RNA signals for GUCY1B3 among the ICF LCLs, Normal A, and the other control LCLs were 1316, 813 and 162 and the analogous values for GUCY1A1 were 233, 179 and 81, respectively. Therefore, in B cells, there might be a previously unreported association of expression of soluble guanylate cyclase 1 subunit genes with the immunoglobulin status.
|
ICF-specific changes in immunoglobulin RNAs and surface immunoglobulins
The reported immune problems in most of the known ICF patients (30 patients; C.M.R.Weemaes and D.Smeets, unpublished data) are usually severe but occasionally slight. The patients may have no or low amounts of IgG, IgA and/or IgM (Table 5), and they usually have normal levels of B cells. The ICF LCLs had very low signals for IgG3 and IgA1 and IgA2 heavy-chain RNAs compared to most of the control LCLs (Tables 2 and 5). However, these ICF cell lines had high levels of IgM and IgD heavy-chain RNAs. For example, the ICF LCLs had IgD RNA signals ranging from approximately 2400 to 5500 (mean 3383) compared to signals of 400–1700 (mean 808) for the normal LCLs. No significant difference in expression of
or 
light-chain RNA was seen in the ICF versus normal LCLs. To determine the percentage of cells in each LCL expressing IgG, IgA, IgM or IgD, these and several additional polyclonal B-cell lines were analyzed by flow cytometry with fluorochrome-labeled antibodies to human Ig. Most of the cells in six of eight examined ICF LCLs were positive for cell surface IgM (sIgM+) and many of these cells were sIgD+ (Table 5 and data not shown). In contrast, all the ICF LCLs displayed negligible numbers of sIgG+ or sIgA+ cells. Also consistent with the microarray data, the normal LCLs gave very different results. Six of the seven normal LCLs had sIgG or sIgA on a large fraction (usually most) of their cells. The lack of cell-surface Ig for ICF B' and ICF K probably reflects a post-transcriptional deficiency (such as, the loss of Ig-specific chaperone-assisted folding of the heavy chain in the endoplasmic reticulum) resulting from changes during prolonged cell culture. ICF B' differed from ICF B, which was sIgM+ and sIgD+, by being passaged in a different laboratory after establishment from a single blood sample. From the highly consistent results on the levels of Ig RNA and the flow cytometry analysis at the level of protein (Table 5 plus results from the four additional LCLs mentioned above), we conclude that the B cells in the peripheral blood of ICF patients are probably mostly IgM+ (or IgD+) with extremely low percentages of IgG+ or IgA+ cells, in contrast to the cells from normal individuals which are mostly IgG+ or IgA+. The differences in the results from clinical analyses of serum Ig in ICF patients and microarray or flow cytometric analyses of ICF LCLs, notably for IgM and IgD, could be explained by defects in B-cell maturation and/or activation after V(D)J recombination in the patients.
ICF-specific changes in RNAs encoding proteins regulating lymphocyte migration, activation or homing
Nine of the genes showing ICF-specific differences in RNA levels are implicated in lymphoid cell differentiation or specific lymphoid function after V(D)J recombination and most of them are involved in signal transduction or in the control of transcription (Table 3). The genes with increased levels of RNA in ICF LCLs are the regulator of G-protein signaling gene 1 (RGS1; Fig. 1) and protein kinase C eta (PRKCH). RGS1 is a B-cell- and monocyte-specific protein that negatively regulates chemokine signaling through heterotrimeric guanine-nucleotide binding proteins (11) and controls B-cell migration within lymphoid organs in response to chemokines (12). PKC eta is present at high levels in pro-B cells and early-stage thymocytes and was shown to be pro-apoptotic in the former cells (13). The pro-B cells in ICF patients could have yet higher levels of PKC eta RNA relative to their normal counterparts than do ICF LCLs because only the ICF pro-B cells with less of an increase in this RNA might survive selection in the bone marrow.
Lower RNA levels in ICF than in control LCLs were seen for the TNFR17 and its overlapping AS gene (Table 3; Fig. 1). TNFR17, also known as B cell maturation factor, is a membrane receptor whose protein ligand (zTNF4) is important for antibody production (9). TNFR7, another TNFR superfamily member assigned roles in antibody production (14), also had significantly lower RNA levels in ICF LCLs than in normal LCLs (Table 3). RNA for B lymphoid tyrosine kinase (BLK), which might function downstream of heterotrimeric G proteins (15), was downregulated in ICF LCLs (Table 3). BLK RNA has been proposed to be involved in signal transduction in the B-cell and thymocyte lineages (16).
Two RNAs downregulated in ICF LCLs are needed for normal immunity as determined by mice knockout studies (Table 3; Fig. 1). They encode integrin ß7 , which facilitates homing to lymphoid tissue in the gut (17), and a transdominant negative protein (ID3), which inhibits the activity and expression of tissue-restricted bHLH proteins by sequestering differentiation-inducing E proteins (18). ICF-associated decreases in RNA levels were also seen for the transcription enhancer MEF2C (19,20) (Table 3; Fig. 1) and the transferrin receptor (21) (Table 3), both of which are associated with immune function. Some of the noteworthy genes that did not display ICF-specific differences in RNA levels are the B-cell antigen receptor complex components CD79a and CD79b; the lymphocyte surface protein, CD81; the B-cell surface protein CD19, a pan B-cell marker; and various members of the integrin family other than ITGB7, of the protein kinase C family other than the PKC eta gene, of the ID family other than ID3, and of the RGS family other than RGS1.
ICF-specific changes in RNA levels of other genes
Several genes encoding proteins involved in signal transduction or apoptosis, but not specifically associated with lymphogenesis, exhibited ICF-specific differences in RNA levels. GUCY1B3 and GUCY1A3 (GUCY1B1 and GUCY1A1) (22) are 0.1 Mb apart and encode the subunits of a soluble guanylate cyclase. Both RNAs were overexpressed in ICF LCLs (Table 4; Fig. 1). Their hemoprotein mediates NO signaling and is involved in many biological processes (23). Also upregulated in the ICF LCLs was the RNA for protein tyrosine phosphatase 13, which can associate with the apoptosis-inducing Fas/CD95 (Table 4). Overexpression of protein phosphatases may help maintain a resting, non-activated phenotype of circulating T cells (24). Two anti-apoptotic genes that are expressed in various tissues and were downregulated in the ICF LCLs were the cell cycle-related BTG2 (25) and the stress response gene heme oxygenase 1 (HMOX1) (26).
Several genes displaying ICF-specific differences in RNA levels are transcription factors or coactivators, e.g. MEF2C (Table 3), which is implicated in IgM secretion, and the ubiquitous transcription factor retinoic acid receptor (RAR; Table 4). RAR has been reported to inhibit activation of resting peripheral blood B cells; to induce or inhibit apoptosis depending on the concentration of retinoid ligand; and to modulate granulopoeisis (27,28). Lastly, some RNAs that were upregulated specifically in ICF LCLs are from poorly characterized human genes, namely, BBC3 (BCL-2 binding component 3), KOC1, MARK3 CNN3, or a gene whose expression is not associated with leukocytes, namely, F13A1 (Table 4).
Methylation analysis of promoter regions of four genes with ICF-specific RNA levels
ICF-linked genomic hypomethylation includes Sat2 and Sat3, D4Z4 repeats, NBL2 repeats and some X-linked genes (5,29,30). We tested the LCLs for ICF-specific hypomethylation of the promoter or 5' region of three genes that displayed ICF-linked upregulation of expression (RGS1, GUCY1B3 and F13A1; Tables 3 and 4). Because DNA hypomethylation and hypermethylation are established in different parts of the genome during embryogenesis and often during carcinogenesis (31), we also analyzed an ICF-downregulated gene (TNFR17) for possible hypermethylation of its promoter. CpG islands overlapping vertebrate promoters and 5' transcribed regions are usually constitutively unmethylated; therefore, two of the four genes for this methylation analysis were chosen partly because they did not have CpG islands in the examined region (RGS1 and TNFR17). For one of the two studied genes whose promoter region is in a CpG island (GUCY1B3), we analyzed an upstream DNA sequence with a low CpG density as well as the CpG-rich 5' region of the gene. These analyses were done by combined bisulfite restriction analysis (COBRA), a quantitative assay of C methylation in which restriction sites can be destroyed or created because of C
T conversions only at unmethylated C residues upon bisulfite modification, alkaline treatment and PCR (32). We used COBRA rather than genomic sequencing because the latter is more difficult to quantitate, especially when trying to explain 2–4-fold differences of expression. With COBRA, we can analyze methylation at individual CpG sites, as in Southern analysis, but COBRA allows more rapid examination of multiple genes and displays methylation by a positive result rather than by enzyme resistance. We analyzed two to eight CpG 5' or promoter region sites for each of the four examined genes (Table 6) and, as described below, the results obtained at different sites for a given DNA region were consistent.
|
Two of the studied genes, RGS1 and TNFR17, are preferentially expressed in mature B cells with little or no expression in T-cell lines and mitogen-activated T cells (8,12,33,34). RGS1 has not been detected in pre-B-cell lines (12). TNFR17 transcripts have been observed in pre-B-cell lines but not in lymphoid precursor nor myeloid cell lines (8). The RGS1 promoter region has a very low CpG content. The 900 bp upstream of the beginning of TNFR17 has almost 2% CpG and the predicted CpG frequency from its base composition would be
4.5% if there were no under-representation of CpGs. Therefore, it does not display the 4–5-fold under-representation of CpGs typical of bulk vertebrate DNA but it is not as rich in CpGs as a CpG island. Two TNFR17 CpG sites, one located 55 bp downstream of the transcription start site (+55; all positions given relative to the beginning of the gene; Fig. 2) and one at –128 and two RGS1 CpG sites at –83 and –2282 were amenable to COBRA (Fig. 2D and E). The –83 RGS1, –128 TNFR17 and +55 TFNR17 CpG sites were unmethylated in all LCLs and in the B-cell fraction of blood (<5% digestion; Fig. 2A and C and data not shown). They were 40–70% methylated in various normal tissues with the exception of RGS1’s site in sperm, which displayed >99% methylation. That the examined RGS1 promoter’s CpG was also hypomethylated in a non-B-cell leukocyte fraction, where this gene is not expected to be expressed, suggests that hypomethylation of this region precedes expression rather than being a consequence of it. Further upstream at the –2282 RGS1 CpG site, there was 40–90% methylation in all tested cell populations but, again, there was no correlation of methylation status with the ICF phenotype (Fig. 2).
|
Four CpG sites within the 300 bp upstream of the beginning of the widely expressed guanylate cyclase ß3 gene (GUCY1B3) were amenable to COBRA using TaqI (three sites) and RsaI (one site). This amplified promoter region was undigested (<5% digestion) and, therefore, unmethylated in all tested LCLs and tissues with two exceptions (liver and Normal D,
10–50% digestion with either of the above enzymes). This promoter region is part of a 1.7 kb CpG island extending into the beginning of the gene and has CpG and C + G contents of 4 and 57%, respectively. Further upstream, in the –380 to –650 region, which has a typical, low CpG content (1%), there was much methylation of all tested LCL and somatic tissue samples at four COBRA sites assayed with HpyCH4IV. In contrast, sperm was completely unmethylated at these sites. The levels of methylation in the somatic cell samples were variable but, again, no correlation was seen between the ICF status and the extent of methylation. COBRA assays of three CpG sites (one each for HpyCH4IV, BstUI, and TaqI) from +53 to +144 of F13A1, which encodes a subunit of the coagulation factor XIII, revealed a complete lack of methylation (<5%) in all the LCLs and tissues despite this protein’s specific association with monocytes/macrophages, megakaryocytes/platelets and placenta (35). From approximately the transcription start point to 800 bp downstream is a CpG island, which has CpG and C + G contents of 5 and 64%, respectively. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ICF is usually discovered in early childhood by the repeated occurrence of severe infections and by the finding of high frequencies of rearrangements in the pericentromeric regions of Chr1 and Chr16 in mitogen-stimulated lymphocytes (3). ICF leukocytes always display hypomethylation in Chr1 and Chr16 within the Sat2-containing heterochromatin adjacent to the centromere, i.e. the juxtacentromeric heterochromatin (1qh and 16qh), as well as in other DNA repeats (29). ICF DNA hypomethylation, which is restricted to a small percentage of the genome (7% in brain) (4), is usually the result of mutations in both alleles of DNMT3B (36). It could be due to a lack of de novo methylation of these sequences early in embryogenesis after most of the DNA methylation is lost and methylation patterns are re-established and/or to a deficiency in DNMT3B activity in postnatal cell populations. Adult cells, including ICF LCLs (37), contain DNMT3B RNA although DNMT3B expression appears to be especially high in embryonic cells (2).
We studied differential expression of genes in ICF versus normal LCLs using B-cell lines (LCLs) because the most consistent and important symptom of this syndrome involves B-cell dysfunction, namely, decreased Ig levels in serum. The six ICF LCLs that we studied retained the cellular phenotype of Sat2 hypomethylation and chromosomal abnormalities in the Sat2-rich 1qh and 16qh (Table 1). Validation of the microarray RNA analysis on LCLs was established in several ways. Among the genes showing ICF-specific over- or underexpression, lymphogenesis-associated genes were significantly over-represented compared to all LCL-expressed sequences (see Results), and these differences correlate with the immunodeficiency of ICF. For five of the genes displaying ICF-specific differences in RNA levels (TFRC, IGHM, IGHA1, GUCY1B3 and TNFR17), we had independent, closely related probe sets on the microarray that confirmed the ICF-associated changes (Tables 2–4). Lastly, we verified the ICF-specific expression from five genes (IGHG3, IGHA1, IGHA2, IGHD and IGHM) at the protein level (Table 5).
Whereas many of the genes that showed significant ICF-specific underexpression or overexpression in the LCLs can account for much of the immunodeficiency that characterizes ICF (see below), none is a good candidate to explain the chromosome instability targeted to the pericentromeric regions of Chr1 and Chr16 in mitogen-stimulated lymphocytes and in LCLs from ICF patients. Therefore, cis effects of hypomethylation at 1qh and 16qh might suffice in the context of the correct cell type (lymphocytes stimulated with mitogen in vitro or B-cell lines) (3,4,38) for predisposing to these chromosomal anomalies (4). Caveats are that the microarray that was used in this work represents only 20% of the estimated number of human genes and some of the genes that are involved in transcription control or signal transduction and display ICF-specific differences in RNA levels (Tables 3 and 4) might indirectly contribute to the ICF chromosome abnormalities.
The present work clarifies the previously unknown nature of the immunodeficiency in ICF. Sometimes, the defective response of ICF patients to antigens is seen as low in vitro stimulation indices for B or T cells upon exposure to mitogens or antigens (39–41). Infections usually affect the pulmonary or gastrointestinal tract. Serum Ig levels in blood samples are highly variable in ICF patients, but the typical finding is low serum Ig coupled with normal B- and T-cell numbers in peripheral blood (C.M.R.Weemaes and D.Smeets, unpublished data), indicating that the early stages of lymphocyte differentiation are usually not abnormal. In some patients, deficiencies are seen in only certain, variable subclasses of IgG (41; B.H.Belohradsky, unpublished data). The variability in the phenotype of ICF patients is probably partly attributable to the varied nature of DNMT3B mutations. To date, such mutant alleles in ICF patients always include at least one allele that is not a null allele and so may specify different amounts of residual DNMT3B activity (36). This is consistent with the finding that homozygous Dnmt3b knockout in mice is lethal during embryogenesis (2).
Our analysis of ICF B-cell lines for surface antigens showed that high percentages of these cells have membrane-bound IgM and/or IgD, indicating that V(D)J recombination was normal for the precursors of these cells. No surface IgG or IgA was detected in these LCLs, in contrast to the control LCLs, and none had appreciable levels of IgG or IgA RNA. Despite the unusually high percentages of cells containing surface IgM in six of the eight examined ICF LCLs, the patients from whom these were derived were moderately to severely deficient in serum IgM (Table 5 and data not shown). These findings suggest that a block at class switching is not the immediate problem in the immune system of ICF patients but rather that they have other deficiencies in lymphocyte maturation or activation.
In the microarray analysis, we found highly promising genes whose dysregulation may be largely responsible for the immunodeficiency in ICF. None of the genes exhibiting ICF-specific differences in RNA levels correspond to genes shown by microarray analysis to be dysregulated in a human colon adenocarcinoma cell line treated with the DNA demethylating (and DNA damaging) agent 5-azadeoxycytidine (42). Also, none are the same as those whose RNAs are abnormally regulated in murine fibroblasts containing a conditional knockout of Dnmt1 that causes massive demethylation (43), unlike the very limited demethylation in ICF cells. Our study revealed nine genes with ICF-specific dysregulation whose protein products are involved in lymphocyte migration, activation or survival consistent with abnormal lymphocyte maturation or function (Table 3). For example, the ICF LCL-linked overexpression of RGS1 (Table 3; Fig. 1), a B-cell-associated negative regulator of heterotrimeric G-protein signaling, probably reflects a decreased ability of B cells in ICF patients to respond to chemokines (12,44). Also, RGS1 strongly inhibits the directed migration of lymphoid cells (11,12). RGS1 expression has been postulated to lead to the retention of B cells in the germinal center of lymphoid organs (12). That significantly more RGS1 RNA was present in ICF LCLs than in the control LCLs is consistent with the hypothesis that B-cell migration within lymphoid organs is defective in ICF patients, in turn, preventing the normal final stages of maturation.
A gene that displayed ICF-specific changes in RNA levels, the integrin ß7 gene (ITGB7; Table 3; Fig. 1), has been demonstrated to be important for promoting intra-organ movement of lymphocytes (45). Whereas RGS1 is involved in negative control of lymphocyte migration and had an increased concentration of its RNA in ICF LCLs, ITGB7 promotes lymphocyte migration and, accordingly, had lower levels of RNA in ICF LCLs. Another RNA whose downregulation in ICF LCLs suggests involvement in ICF immune dysfunction encodes TNFR17, a B-cell-associated member of the tumor necrosis receptor family without a death domain (8,34) (Table 3; Fig. 1). It is the receptor for zTNF4 (also known as BLyS, BAFF, TALL-1 or THANK), which plays an important role in B-cell expansion and antibody responses (9). Functional redundancy between Tnfr17 and Taci, another zTNF4 receptor, may be responsible for the lack of a phenotype in Tnfr17 (Bcma) homozygous knockout mice (46), but such redundancy might not exist in humans, especially because there is only 56% identity between the human and murine proteins. TNFR17 has been proposed to transduce signals for cell survival and proliferation through activation of the MAP kinase pathway (47), which is an essential part of B- and T-cell development.
Two other genes that are also underexpressed in ICF LCLs and can be upregulated in stimulated B cells are transferrin receptor (21) and B-MEF-2 (apparently encoded by MEF2C) (20). MEF2C is primarily expressed in skeletal muscle, brain, and spleen, and within the lymphoid lineage, it is restricted to B cells (20,48). In B cells, MEF2C is a positive regulator of J chain expression (20,49). Because the J chain is required for optimal production of secreted pentameric IgM and dimeric IgA, decreased expression of MEF2C (Table 3; Fig. 1) could affect the ability of ICF B cells to produce secreted forms of IgM while maintaining normal expression of IgM on the surface of the B cells, consistent with the phenotype of the ICF patients in this microarray study. The product of the other gene, transferrin receptor, is critical for transport of iron into cells. Increased transferrin receptor in B cells is a marker for activation (21) and has a role in antigen trafficking in B cells (50). Thus, the decrease in transferrin receptor RNA observed in ICF LCLs in comparison to normal LCLs (Table 3) may indicate a diminished activation state of B cells in ICF patients.
A gene that functions in differentiation during embryogenesis as well as during lymphogenesis and displayed ICF-specific decreases in RNA levels is ID3. Underexpression of ID3 can help explain defective immune function without decreases in B- or T-cell numbers, as seen in the majority of ICF patients. ID3 inhibits the DNA binding of positive-acting transcription factors (especially E factors) by heterodimerizing with them. Several of these E factors play central roles in B- and T-cell maturation. Despite B and T cells being phenotypically normal and in normal numbers in Id3 homozygous knockout mice, these mice have reduced IgG1 and IgG2a levels and large decreases in production of various subclasses of serum Ig, including IgM, in response to T-cell dependent or T-cell independent antigens (51). Part of their deficient humoral immunity may be due to a defective B-cell proliferation response to activation of the B-cell receptor complex (51). Furthermore, in another Id3–/– mouse model, Id3 was shown to be necessary for thymocyte positive selection, possibly operating through T-cell receptor signaling (18). Importantly for elucidating the immunodeficiency associated with ICF, Id3+/– mice, which should have approximately half the normal amount of Id3 RNA, also had perturbed thymocyte maturation (18). This phenotype for the heterozygotes could be a result of the loss of a delicate balance between levels of Id3/ID3 and of a thymocyte-specific transcription activator E protein. ICF LCLs also had approximately half the level of ID3 RNA compared to controls (Table 3; Fig. 1).
Some of the changes in RNA levels in ICF versus control LCLs may be secondary to changes in RNA levels specified by other genes. For example, the increase in RGS1 RNA in ICF LCLs (Table 3; Fig. 1) should lead to decreased Gi and Gq signaling resulting in downregulation of MAP kinase activity in response to platelet activation factor (PAF) (12). Effects of modest changes in RGS1 protein levels might be dramatic because RGS proteins generally have low Km values (52). Negative regulation of MAP kinase activity might be responsible for the ICF-specific decreases in ID3 RNA levels because ID3 is regulated at the transcription level by MAP kinase activity (53). Also, a lack of maturity of B cells in ICF patients’ peripheral blood could account for the low levels of some of the RNAs in ICF LCLs, e.g. integrin ß7, TNFR17 and TFRC RNAs, because those proteins are associated with leukocyte or B-cell maturation (21,34,54,55).
However, the postulated immaturity of the B cells from which the LCLs were established would not explain the ICF-linked increases in RGS1 RNA. Higher levels of transcripts from RGS1 are associated with mitogen activation of tonsil B cells or induction of G-protein signaling in a B-cell line by PAF (12,33). RGS proteins have been proposed to be necessary for feedback desensitization of G-protein activated pathways (11) and so the increase in RGS1 levels in B cells from ICF patients, who appear to be deficient in the activation of B cells, indicates a premature induction of this feedback control. We propose that RGS1 and a few other of the genes showing ICF-specific differences in RNA levels (as well as genes not represented in this array) are proximal targets of the deficiency in DNMT3B activity in ICF patients. Most of the other genes with ICF-specific changes in their RNA levels may be subject to changes in transcription or post-transcriptional processing or stabilization of their mRNA secondary to proximally affected genes that control signal transduction, other types of feedback pathways or transcription.
The DNMT3B mutations in ICF patients are often missense mutations in the catalytic domain, suggesting that it is the loss of DNA methyltransferase activity and not some other function of the protein that is responsible for the syndrome. For vertebrate DNA sequences subject to transcription control by DNA methylation, this control is almost always downregulation. Furthermore, DNA methylation is often associated with heterochromatinization, which is generally repressive to transcription. Therefore, we hypothesize that it is some of the upregulated genes in ICF LCLs whose RNA levels are most directly affected by the DNMT3B mutations and that this is the result of ICF-specific DNA hypomethylation. Nevertheless, no ICF-specific differences in promoter methylation were seen in three of the genes that had higher RNA levels in ICF than in the control LCLs. COBRA analysis of promoter regions of the two B-cell-specific genes did reveal B-cell- or leukocyte-specific hypomethylation, but this occurred irrespective of whether the cells were derived from ICF patients or unaffected individuals (Fig. 2A and C). In ICF lymphoid cells, hypomethylation of regions of the genome that are normally constitutively heterochromatic (e.g. 1qh and 16qh) and consequent decreases in heterochromatinization (3,4) might affect regulation of expression of genes elsewhere in the genome. This could occur by altering the nuclear compartmentalization of a small percentage of the gene-containing euchromatin or modulating the sequestration of certain transcription regulatory proteins at specific constitutive heterochromatin regions (56,57). Evidence suggests that, during murine lymphogenesis, centromeric heterochromatin is involved in the heritable repression of TdT and several other genes through an interaction with euchromatin regions containing these genes (56,58). Ikaros, a transcription regulatory factor that binds specifically to TdT and 5 promoters and to centromeric heterochromatin, mediates this repression of TdT and 5 transcription (57,59). Similarly, centromeric heterochromatin seems to have a repressive interaction with the human ß-globin locus (60). Different heterochromatin-rich domains, such as the human Sat-rich centromeric heterochromatin and Sat2-rich1qh and 16qh, might act independently of one another (56). Further studies will be necessary to determine whether, in normal lymphocytes, association of highly methylated DNA in constitutive heterochromatin negatively controls expression of genes that exhibit upregulation of their RNA in ICF LCLs and whether this control is abrogated in ICF lymphocytes by DNA hypomethylation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Cell lines and their characterization
All LCLs were Epstein–Barr virus-transformed B-cell lines. LCLs ICF B/B', ICF C, ICF S, ICF K and ICF G have been described previously (3,6,39,40,61); the first three were from patients P1, P6 and P14, respectively, in Wijmenga et al. (36). Normal B, C and G were from phenotypically normal parents of patients who were donors for ICF B/B', C and G LCLs. Normal A and D were from phenotypically normal individuals (4). ICF B, ICF B', ICF G, ICF K, Normal A, D and C were from males whereas the other LCLs were from females. Fifty metaphases were examined for gross chromosomal aberrations in each LCL from the same cell passage used for harvesting for the microarray expression analysis. To prepare cell pellets for RNA extraction for microarray analysis, LCLs were split 1:3.3 when they were at 5 x 105 cells/ml, incubated, and then approximately 10 x 106 cells were harvested when the cultures reached a concentration of 3 x 105 cells/ml. Chromosomal anomalies and Chr1 Sat2 DNA methylation was assessed as described by Tuck-Muller et al. (4). ICF LCLs which were used just for flow cytometry were from patient B in Xu et al. (1) and his affected brother.
Microarray analysis
Total RNA was isolated by standard methods (Trizol reagent, Invitrogen, Gibco Life Technology, Carlsbad, CA; RNeasy spin column, Qiagen, Valencia, CA). Preparation of cRNA from 5 µg of total RNA, hybridization and scanning of the oligonucleotide arrays (HuGeneFL, Affymetrix, Sanat Clara, CA) were performed as described by Le Naour et al. (62) on 5 µg of total RNA reverse transcribed with an oligo(dT)24 primer containing a T7 RNA polymerase promoter site. Then, labeled cRNA was generated by in vitro transcription. Fifteen micrograms was fragmented at 94°C for 35 min in 40 mM Tris–acetate, 100 mM potassium acetate, 30 mM magnesium acetate pH 8.1 and used to prepare 300 µl of hybridization mixture [100 mM 4-morpholineethanesulfonic acid (MES) pH 6.7, 1 M NaCl, 20 mM EDTA, 0.01% Tween 20, 0.1 mg/ml herring sperm DNA, 500 µg/ml acetylated bovine serum albumin] containing a mixture of control cRNAs for hybridization efficiency and quantitation of transcript levels. The hybridization mixtures were heated to 94°C for 5 min, equilibrated at 45°C for 5 min, then clarified by centrifugation. Ten micrograms of fragmented cRNA was hybridized to the arrays at 45°C for 16 h. The arrays were washed with 6x saline/sodium phosphate/EDTA (SSPE) at 25°C and then with 100 mM MES, 0.1 M NaCl, 0.01% Tween 20 pH 6.7, at 50°C. They were stained with streptavidin-phycoerythrin, washed with 6x SSPE, stained with biotinylated anti-streptavidin IgG, stained again with streptavidin-phycoerythrin, washed with 6x SSPE, then scanned, and subjected to image analysis (GeneArray scanner and GeneChip 4.0 software, Affymetrix).
Statistical analyses
Each probe set on the microarrays typically consists of 20 oligonucleotides complementary to a specific cDNA [perfect match (PM) features], and 20 identical oligonucleotides except for a central mismatch (MM features). Probe-pairs for which the PM – MM intensities were less than –1000 on one normal LCL that was selected as a standard were excluded from the analysis, as were probe-pairs displaying saturation in the image of the standard for either the PM or MM. For a saturated PM (or MM) value on other chips, the values were imputed by considering the ratio of non-saturating PM (or MM) values compared to the standard chip. A one-sided signed-rank test on the PM – MM differences was used to test expression for the probe set. The average intensity for each probe set was computed as the mean of PM – MM differences after trimming away the 25% highest and lowest PM – MM differences (see http://dot.ped.med.umich.edu:2000/pub/ICF/index.html). The average intensities for each chip were normalized using a piece-wise linear function that makes 99 evenly spaced quantiles which have the same values as the corresponding quantiles of the standard. In order to fit linear models, the data were then log-transformed by adding 100 to each trimmed-mean, converting still negative values to 0, and then adding 1 and taking logarithms. ICF and normal LCLs were compared using two-sample t-tests on the log-transformed data. For comparison of non-LCL cell lines with LCLs, a one-way ANOVA model was fit and the resulting t-tests for comparing pairs of groups used to judge the significance of LCL versus non-LCL differences.
DNA methylation analysis (COBRA)
COBRA was done as described by Xiong and Laird (32) and Sun et al. (63). Oligonucleotides and PCR conditions are given in Table 6. Only the expected sized PCR products were obtained. Ethidium bromide-induced fluorescence in the bands representing the cleaved versus the uncleaved PCR products was quantitated by digital analysis. Internal controls for complete digestion that were included in each digest indicated that no inhibitors of digestion were present. We demonstrated that the primer-pairs, which were designed to amplify only bisulfite-treated DNA, could amplify the genomic DNA only when it was bisulfite modified or that the PCR product from the bisulfite-modified DNA was unable to be cleaved by a non-CpG restriction endonuclease whose restriction site should be destroyed by a CT conversion in the modified DNA whereas it could be cleaved by enzymes whose restriction site should be retained after bisulfite modification.
Flow cytometric analysis of LCLs of surface marker expression
LCL cells (106) were incubated with monoclonal FITC-anti human IgD, PE-anti human IgM, biotinylated-anti human IgG, biotinylated-anti human IgA and FITC-anti human CD19 (BD Pharmingen, San Diego, CA) for 40 min on ice. The cells were washed and incubated with streptavidin-CyChrome on ice for 40 min. Appropriate conjugated isotype control antibodies were used to quantify non-specific staining. Following staining, the LCLs were washed, fixed in cold 1% paraformaldehyde and analyzed by flow cytometry. Sorting for B lymphocytes from peripheral blood was done from a leukocyte fraction obtained by lysing red blood cells and staining remaining cells for 45 min with FITC-conjugated anti-human CD19 and the subset of interest sorted.
|
ACKNOWLEDGEMENTS |
|---|
We are very grateful to G.K.Hinkel for generously sharing ICF LCLs with us and to K.Uhlmann and M.Karbasiyan for unpublished data on the ICF K mutation. We are also very thankful to Alan Tucker for preparative sorting of the B-cell fraction. This research was supported in part by NIH grants CA81506 (to M.E.) and CA26803 (to S.H.) from the National Institutes of Health.
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +1 504 584 2449; Fax: +1 504 584 1763; Email: ehrlich@tulane.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Xu, G., Bestor, T.H., Bourc’his, D., Hsieh, C., Tommerup, N., Hulten, M., Qu, S., Russo, J.J. and Viegas-Péquignot, E. (1999) Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature, 402, 187–191.[Medline]
2 Okano, M., Bell, D.W., Haber, D.A. and Li, E. (1999) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell, 98, 247–257.
3 Smeets, D.F.C.M., Moog, U., Weemaes, C.M.R., Vaes-Peeters, G., Merkx, G.F.M., Niehof, J.P. and Hamers, G. (1994) ICF syndrome: a new case and review of the literature. Hum. Genet., 94, 240–246.[ISI][Medline]
4 Tuck-Muller, C.M., Narayan, A., Tsien, F., Smeets, D., Sawyer, J., Fiala, E.S., Sohn, O. and Ehrlich, M. (2000) DNA hypomethylation and unusual chromosome instability in cell lines from ICF syndrome patients. Cytogenet. Cell Genet., 89, 121–128.[ISI][Medline]
5 Jeanpierre, M., Turleau, C., Aurias, A., Prieur, M., Ledeist, F., Fischer, A. and Viegas-Pequignot, E. (1993) An embryonic-like methylation pattern of classical satellite DNA is observed in ICF syndrome. Hum. Mol. Genet., 2, 731–735.
6 Schuffenhauer, S., Bartsch, O., Stumm, M., Buchholz, T., Petropoulou, T., Kraft, S., Belohradsky, B., Meitinger, T. and Wegner, R. (1995) DNA, FISH and complementation studies in ICF syndrome; DNA hypomethylation of repetitive and single copy loci and evidence for a trans acting factor. Hum. Genet., 96, 562–571.[ISI][Medline]
7 Narayan, A., Tuck-Muller, C., Weissbecker, K., Smeets, D. and Ehrlich, M. (2000) Hypersensitivity to radiation-induced non-apoptotic and apoptotic death in cell lines from patients with the ICF chromosome instability syndrome. Mutat. Res., 456, 1–15.[ISI][Medline]
8 Laabi, Y., Gras, M.P., Brouet, J.C., Berger, R., Larsen, C.J. and Tsapis, A. (1994) The BCMA gene, preferentially expressed during B lymphoid maturation, is bidirectionally transcribed. Nucleic Acids Res., 22, 1147–1154.
9 Gross, J.A., Johnston, J., Mudri, S., Enselman, R., Dillon, S.R., Madden, K., Xu, W., Parrish-Novak, J., Foster, D., Lofton-Day, C. et al. (2000) TACI and BCMA are receptors for a TNF homologue implicated in B-cell autoimmune disease. Nature, 404, 995–999.[Medline]
10 Budworth, J., Meillerais, S., Charles, I. and Powell, K. (1999) Tissue distribution of the human soluble guanylate cyclases. Biochem. Biophys. Res. Commun., 263, 696–701.[ISI][Medline]
11 Berman, D.M. and Gilman, A.G. (1998) Mammalian RGS proteins: barbarians at the gate. J. Biol. Chem., 273, 1269–1272.
12 Moratz, C., Kang, V.H., Druey, K.M., Shi, C.S., Scheschonka, A., Murphy, P.M., Kozasa, T. and Kehrl, J.H. (2000) Regulator of G protein signaling 1 (RGS1) markedly impairs Gi signaling responses of B lymphocytes. J. Immunol., 164, 1829–1838.
13 Morrow, T.A., Muljo, S.A., Zhang, J., Hardwick, J.M. and Schlissel, M.S. (1999) Pro-B-cell-specific transcription and proapoptotic function of protein kinase Ceta. Mol. Cell. Biol., 19, 5608–5618.
14 Prasad, K.V., Ao, Z., Yoon, Y., Wu, M.X., Rizk, M., Jacquot, S. and Schlossman, S.F. (1997) CD27, a member of the tumor necrosis factor receptor family, induces apoptosis and binds to Siva, a proapoptotic protein. Proc. Natl Acad. Sci. USA, 94, 6346–6351.
15 Deehan, M.R., Klaus, G.G., Holman, M.J., Harnett, W. and Harnett, M.M. (1998) MAPkinase: a second site of G-protein regulation of B-cell activation via the antigen receptors. Immunology, 95, 169–177.[ISI][Medline]
16 Islam, K.B., Rabbani, H., Larsson, C., Sanders, R. and Smith, C.I. (1995) Molecular cloning, characterization and chromosomal localization of a human lymphoid tyrosine kinase related to murine Blk. J. Immunol., 154, 1265–1272.[Abstract]
17 Artis, D., Humphreys, N.E., Potten, C.S., Wagner, N., Muller, W., McDermott, J.R., Grencis, R.K. and Else, K.J. (2000) ß7 integrin-deficient mice: delayed leukocyte recruitment and attenuated protective immunity in the small intestine during enteric helminth infection. Eur. J. Immunol., 30, 1656–1664.[ISI][Medline]
18 Rivera, R.R., Johns, C.P., Quan, J., Johnson, R.S. and Murre, C. (2000) Thymocyte selection is regulated by the helix–loop–helix inhibitor protein, Id3. Immunity, 12, 17–26.[ISI][Medline]
19 Glynne, R., Ghandour, G., Rayner, J., Mack, D.H. and Goodnow, C.C. (2000) B-lymphocyte quiescence, tolerance and activation as viewed by global gene expression profiling on microarrays. Immunol. Rev., 176, 216–246.[ISI][Medline]
20 Rao, S., Karray, S., Gackstetter, E.R. and Koshland, M.E. (1998) Myocyte enhancer factor-related B-MEF2 is developmentally expressed in B cells and regulates the immunoglobulin J chain promoter. J. Biol. Chem., 273, 26123–26129.
21 Futran, J., Kemp, J.D., Field, E.H., Vora, A. and Ashman, R.F. (1989) Transferrin receptor synthesis is an early event in B cell activation. J. Immunol., 143, 787–792.[Abstract]
22 Zabel, U., Weeger, M., La, M. and Schmidt, H.H. (1998) Human soluble guanylate cyclase: functional expression and revised isoenzyme family. Biochem. J., 335, 51–57.
23 Lee, Y.C., Martin, E. and Murad, F. (2000) Human recombinant soluble guanylyl cyclase: expression, purification and regulation. Proc. Natl Acad. Sci. USA, 97, 10763–10768.
24 Gjorloff-Wingren, A., Saxena, M., Han, S., Wang, X., Alonso, A., Renedo, M., Oh, P., Williams, S., Schnitzer, J. and Mustelin, T. (2000) Subcellular localization of intracellular protein tyrosine phosphatases in T cells. Eur. J. Immunol., 30, 2412–2421.[ISI][Medline]
25 Guardavaccaro, D., Corrente, G., Covone, F., Micheli, L., D’Agnano, I., Starace, G., Caruso, M. and Tirone, F. (2000) Arrest of G(1)-S progression by the p53-inducible gene PC3 is Rb dependent and relies on the inhibition of cyclin D1 transcription. Mol. Cell. Biol., 20, 1797–1815.
26 Elbirt, K.K. and Bonkovsky, H.L. (1999) Heme oxygenase: recent advances in understanding its regulation and role. Proc. Assoc. Am. Physicians, 111, 438–447.
27 Lomo, J., Smeland, E.B., Ulven, S., Natarajan, V., Blomhoff, R., Gandhi, U., Dawson, M.I. and Blomhoff, H.K. (1998) RAR, not RXR, ligands inhibit cell activation and prevent apoptosis in B-lymphocytes. J. Cell Physiol., 175, 68–77.[ISI][Medline]
28 Kastner, P., Lawrence, H.J., Waltzinger, C., Ghyselinck, N.B., Chambon, P. and Chan, S. (2001) Positive and negative regulation of granulopoiesis by endogenous RAR. Blood, 97, 1314–1320.
29 Kondo, T., Comenge, Y., Bobek, M.P., Kuick, R., Lamb, B., Zhu, X., Narayan, A., Bourc’his, D., Viegas-Pequinot, E., Ehrlich, M. and Hanash, S. (2000) Whole-genome methylation scan in ICF syndrome: hypomethylation of non-satellite DNA repeats D4Z4 and NBL2. Hum. Mol. Genet., 9, 597–604.
30 Hansen, R.S., Stoger, R., Wijmenga, C., Stanek, A.M., Canfield, T.K., Luo, P., Matarazzo, M.R., D’Esposito, M., Feil, R., Gimelli, G. et al. (2000) Escape from gene silencing in ICF syndrome: evidence for advanced replication time as a major determinant. Hum. Mol. Genet., 9, 2575–2587.
31 Ehrlich, M. (2000) DNA Alterations in Cancer: Genetic and Epigenetic Changes. Biotechniques Books, Eaton Publishing, Natick.
32 Xiong, Z. and Laird, P.W. (1997) COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res., 25, 2532–2534.
33 Hong, J.X., Wilson, G.L., Fox, C.H. and Kehrl, J.H. (1993) Isolation and characterization of a novel B cell activation gene. J. Immunol., 150, 3895–3904.[Abstract]
34 Gras, M.P., Laabi, Y., Linares-Cruz, G., Blondel, M.O., Rigaut, J.P., Brouet, J.C., Leca, G., Haguenauer-Tsapis, R. and Tsapis, A. (1995) BCMAp: an integral membrane protein in the Golgi apparatus of human mature B lymphocytes. Int. Immunol., 7, 1093–1106.
35 Kida, M., Souri, M., Yamamoto, M., Saito, H. and Ichinose, A. (1999) Transcriptional regulation of cell type-specific expression of the TATA-less A subunit gene for human coagulation factor XIII. J. Biol. Chem., 274, 6138–6147.
36 Wijmenga, C., Hansen, R.S., Gimelli, G., Bjorck, E.J., Davies, E.G., Valentine, D., Belohradsky, B.H., van Dongen, J.J., Smeets, D.F., van den Heuvel, L.P. et al. (2000) Genetic variation in ICF syndrome: evidence for genetic heterogeneity. Hum. Mutat., 16, 509–517.[ISI][Medline]
37 Hansen, R.S., Wijmenga, C., Luo, P., Stanek, A.M., Canfield, T.K., Weemaes, C.M. and Gartler, S.M. (1999) The DNMT3B DNA methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proc. Natl Acad. Sci. USA, 96, 14412–14417.
38 Fasth, A., Forestier, E., Holmberg, E., Holmgren, G., Nordenson, I., Soderstrom, T. and Wahlstrom, J. (1990) Fragility of the centromeric region of chromosome 1 associated with combined immunodeficiency in siblings: a recessively inherited entity? Acta Paediatr. Scand., 79, 605–612.[ISI][Medline]
39 Carpenter, N.J., Fillpovich, A., Blaese, R.M., Carey, T.L. and Berkel, A.I. (1988) Variable immunodeficiency with abnormal condensation of the heterochromatin of chromosomes 1, 9 and 16. J. Pediatr., 112, 757–760.[ISI][Medline]
40 Turleau, C., Cabanis, M.-O., Girault, D., Ledeist, F., Mettey, R., Puissant, H., Marguerite, P. and de Grouchy, J. (1989) Multibranched chromosomes in the ICF syndrome: immunodeficiency, centromeric instability and facial anomalies. Am. J. Med. Genet., 32, 420–424.[ISI][Medline]
41 Franceschini, P., Martino, S., Ciocchini, M., Ciuti, E., Vardeu, M.P., Guala, A., Signorile, F., Camerano, P., Franceschini, D. and Tovo, P.A. (1995) Variability of clinical and immunological phenotype in immunodeficiency–centromeric instability–facial anomalies syndrome. Report of two new patients and review of the literature. Eur. J. Pediatr., 154, 840–846.[ISI][Medline]
42 Karpf, A.R., Peterson, P.W., Rawlins, J.T., Dalley, B.K., Yang, Q., Albertsen, H. and Jones, D.A. (1999) Inhibition of DNA methyltransferase stimulates the expression of signal transducer and activator of transcription 1, 2 and 3 genes in colon tumor cells. Proc. Natl Acad. Sci. USA, 96, 14007–14012.
43 Jackson-Grusby, L., Beard, C., Possemato, R., Tudor, M., Fambrough, D., Csankovszki, G., Dausman, J., Lee, P., Wilson, C., Lander, E. and Jaenisch, R. (2001) Loss of genomic methylation causes p53-dependent apoptosis and epigenetic deregulation. Nat. Genet., 27, 31–39.[ISI][Medline]
44 Reif, K. and Cyster, J.G. (2000) RGS molecule expression in murine B lymphocytes and ability to down-regulate chemotaxis to lymphoid chemokines. J. Immunol., 164, 4720–4729.
45 Lim, S.P., Leung, E. and Krissansen, G.W. (1998) The ß7 integrin gene (Itgb-7) promoter is responsive to TGF-ß1: defining control regions. Immunogenetics, 48, 184–195.[ISI][Medline]
46 Xu, S. and Lam, K.P. (2001) B-cell maturation protein, which binds the tumor necrosis factor family members BAFF and APRIL, is dispensable for humoral immune responses. Mol. Cell. Biol., 21, 4067–4074.
47 Hatzoglou, A., Roussel, J., Bourgeade, M.F., Rogier, E., Madry, C., Inoue, J., Devergne, O. and Tsapis, A. (2000) TNF receptor family member BCMA (B cell maturation) associates with TNF receptor-associated factor (TRAF) 1, TRAF2 and TRAF3 and activates NF- B, elk-1, c-Jun N-terminal kinase and p38 mitogen-activated protein kinase. J. Immunol., 165, 1322–1330.
48 Swanson, B.J., Jack, H.M. and Lyons, G.E. (1998) Characterization of myocyte enhancer factor 2 (MEF2) expression in B and T cells: MEF2C is a B cell-restricted transcription factor in lymphocytes. Mol. Immunol., 35, 445–458.[ISI][Medline]
49 Wallin, J.J., Rinkenberger, J.L., Rao, S., Gackstetter, E.R., Koshland, M.E. and Zwollo, P. (1999) B cell-specific activator protein prevents two activator factors from binding to the immunoglobulin J chain promoter until the antigen-driven stages of B cell development. J. Biol. Chem., 274, 15959–15965.
50 Cheng, P.C., Steele, C.R., Gu, L., Song, W. and Pierce, S.K. (1999) MHC class II antigen processing in B cells: accelerated intracellular targeting of antigens. J. Immunol., 162, 7171–7180.
51 Pan, L., Sato, S., Frederick, J.P., Sun, X.H. and Zhuang, Y. (1999) Impaired immune responses and B-cell proliferation in mice lacking the Id3 gene. Mol. Cell. Biol., 19, 5969–5980.
52 Wang, J., Ducret, A., Tu, Y., Kozasa, T., Aebersold, R. and Ross, E.M. (1998) RGSZ1, a Gz-selective RGS protein in brain. Structure, membrane association, regulation by G z phosphorylati