| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print December 21, 2007
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Institute of Pathology,*University of Wuerzburg, Wuerzburg, Germany; the Laboratory of Tumor Pathology and Molecular Diagnostics,Institute of Biotechnology, Bay Zoltan Foundation for Applied Research, Szeged, Hungary; and the Department of Pathology,Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama, Japan
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
genes, but also of genes that regulate proliferation and apoptosis. Examples of such genes are the CDK4 gene and the Fas ligand gene, whose promoters are bound and controlled by NFATc factors.4-9
These and other experimental data suggested that, in addition to their role in normal lymphocyte development, NFATc factors, if aberrantly expressed, might also be involved in the generation of lymphoid tumors (see also10
). Experimental evidence for this conclusion can be derived from retroviral tagging experiments in which the murine Nfatc1 locus has been identified as a common insertion site for oncogenic viruses (see also the Mouse Retrovirus Tagged Cancer Gene Database at http://RTCGD.ncifcrf.gov).11 In these studies, all retroviral insertions were found to be located close to either the two Nfatc1 promoters or polyadenylation sites,12 suggesting that retroviral insertions modulate NFATc1 expression. At similar positions, retroviral insertion sites have also been identified within the murine Nfatc3 gene, and proviral insertion resulted in a loss or decrease of NFATc3 expression. Because infections of newborn mice with the T-cell lymphomagenic retrovirus SL3-3 MLV led to an acceleration of lymphoma generation in NFATc3-deficient mice compared with wild-type mice, we concluded that NFATc3 exerts a tumor suppressor-like activity for the generation of viral-induced lymphoid tumors.13 Investigations of NFATc2-deficient mice that develop chondrocytic neoplasms led to a similar conclusion, ie, NFATc2 can act like a tumor suppressor for the generation of chondrocytic malignancies.14
However, in other cell types and under other experimental settings, NFAT factors can also exert oncogenic activities. Transformation of pre-adipocytes by a constitutively active version of NFATc1/A showed that this NFAT protein acts as an oncogene for the development of human adipocytic and other tumors,15 whereas in human classical Hodgkin’s lymphoma (cHL), NFATc1 was found to be suppressed.16 In a subset of large B-cell lymphomas, NFATc1 and c-Rel were detected to be expressed constitutively, to interact with each other, and to control synergistically the expression of CD40 ligand/CD154, thereby maintaining the survival of these aggressive lymphomas.17 In a similar way, NFATc1 and NFATc2 contribute to the constitutive expression of B-lymphocyte stimulator in large B-cell lymphomas and mantle cell lymphomas and support the survival and proliferation of these malignant B-cell lymphomas.18 NFATc1 was found to be overexpressed in human pancreatic tumors and to enhance c-myc expression, and inhibition of calcineurin/NFATc1 signaling resulted in attenuation of cell proliferation and anchorage-independent growth.19 Similar results on a constitutive activation of calcineurin have recently been reported for T-cell acute lymphoblastic leukemias.20 Induced by integrin signals, NFATc2 (and NFAT5) promotes the invasion of human breast cancer cells21 by the induction of cyclooxygenase 2, which controls prostaglandin E2 synthesis.22 The NFATc2-induced motility and invasion of breast cancer cells can be blocked by Akt/PKB signals,23 which were shown to suppress the activation of NFATcs in thymocytes and peripheral T cells.24,25 Taken together, these data indicate that alterations in calcineurin/NFATc signaling play an important role in human cancerogenesis, and because of their immunoreceptor-dependent activation, aberrant NFATc activity affects particularly the generation and maintenance of human lymphomas.
In this study, we have investigated the expression of NFATc1 (and NFATc2) in human lymphomas. By immunohistochemical staining, we found that NFATc1 expression is suppressed in human anaplastic large cell lymphomas (ALCLs) and cHLs. Both ALCL and cHL are lymphoma entities with a defective immunoreceptor signaling.26-28 The silencing of the NFATC1 locus is correlated with low levels of histone H3 acetylation, H3-K4 trimethylation, and Sp1 factor binding in its P1 promoter region, which is part of a CpG island and controlled by numerous inducible transcription factors.29 We also show here that hypermethylation of P1 promoter DNA correlates with the suppression of NFATc1 expression that, in part, can be released by inhibition of de novo DNA methylation by 5-aza-2'-deoxycytidine (5-aza-dC). On interruption of immunoreceptor signaling by cyclosporin A (CsA) in EL-4 thymoma cells that results in the suppression of NFATc1, but not NFATc2, we observed distinct changes in histone modification, whereas DNA methylation remained unaffected. These results suggest that defects in immunoreceptor signaling in ALCL and HL cells lead to a cascade of epigenetic changes that, finally, contribute to the stable silencing of the NFATC1 promoter(s) in these human lymphomas.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Formalin-fixed, paraffin-embedded samples from 3 hyperplastic tonsils, 2 lymph nodes with nonspecific hyperplasia, and 226 malignant lymphoma cases were selected from the histopathology files of the Institute of Pathology, University of Wuerzburg; the Laboratory of Tumor Pathology and Molecular Diagnostics, Institute of Biotechnology, Bay Zoltán Foundation for Applied Research Szeged; and the Department of Pathology, Okayama University. All samples from patients were used following permission and informed consent by the patients. The list of malignant lymphoma cases that were investigated is given in Table 1
. Murine EL-4 T lymphoma cells; human Jurkat leukemic cells; the human ALCL lines SR-786, SUP-M2, and KARPAS 299; and the human HL cell lines L428, L540, L1236, and KM-H2 were grown in RPMI containing 10% fetal calf serum to a density of 2 x 105 cells per milliliter and used in all biochemical assays.
|
For detection of NFATc1 expression, a mouse IgG1 monoclonal antibody was used recognizing amino acids 197 to 304 of human NFATc1 (clone 7A6; sc-7294; Santa Cruz Biotechnology, Santa Cruz, CA), common to all major NFATc1 isoforms.30,31 Immunohistochemical reactions were performed using a streptavidin-biotin-horseradish peroxidase complex method, following heat-induced antigen retrieval performed in 0.5 mol/L Tris buffer (pH 6.0) and pressure cooker. An immunoreaction was classified positive if proposed neoplastic cells revealed moderate to marked homogeneous (cytosolic) staining for NFATc1. In some cases, moderate to marked NFATc1 immunostaining occurred in a proportion of the tumor cells, and those that showed staining in at least one-half of these cells were scored to be positive. For the exact immunolocalization of NFATc1 in plasma cells with higher sensitivity, NFATc1 staining was combined with an Ig light chain antibody (Ab) cocktail, using double immunofluorescent stains and confocal laser scanning microscopy in formalin-fixed, paraffin-embedded sections of hyperplastic tonsils.
Western blots were performed by fractionation of whole protein extracts from human and murine cell lines on 10% polyacrylamide gels followed by immunodetection using polyclonal Abs raised against the N-terminal halves of NFATc1 or c2, respectively (ImmunoGlobe, Groβostheim, Germany), or against Sp1.32 Equal protein loading of gels was determined by staining of filters with Poinceau red.
DNA Methylation Assays
CpG methylation studies of the human NFATC1 promoter region were performed using the PCR-based sodium-bisulfite DNA modification procedure.33
The PCR products were directly sequenced, and semiquantitative data on methylation degree (Figure 2B)
were derived from the sequencing peaks. The sequences of the primers used in PCR amplifications for sequencing of bisulfite-modified DNA were as follows. P1 promoter, distal block of homology: forward, 5'-AACAAATAAACRCRTCCCCRAACCTCCCCAC-3' (for primary, outer amplification), and reverse, 5'-GAAYGGGTTAGAYGGGAYGTTTGAGTTTAY-3' (for primary, outer amplification); forward, 5'-AAACRCRTCCCCRAACCTCCCCACRCCRACC-3' (nested), and reverse, 5'-GAYGGGAYGTTTGAGTTTAYGYGGGTGTTYGG-3' (nested). P1 promoter, proximal block of homology: forward, 5'-CCRCRACCCTAAAACCTACRCRATAAC (for primary, outer amplification), and reverse, 5'-GTTTTTAGGYGAGYGGTTGTYGYGGYG-3' (for primary, outer amplification); forward, 5'-CRCRATAACTCCRAACCCTACCCRC-3' (nested), and reverse, 5'-GGGYGTTYGGYGATTGTTTTYGGG-3' (nested).
|
Electrophoretic Mobility Shift Assays and Chromatin Immunoprecipitation Assays
Nuclear proteins from Jurkat T and HL cell lines were prepared and used in electrophoretic mobility shift assays (EMSAs) as described previously.29 For the detection of Sp1 binding to the P1 promoter, the following (double-stranded) oligonucleotides were incubated as radioactively labeled probes together with 5 µg of nuclear protein: –825 site, (–836) 5'-GGCCTCCCCACGCCGGCCCCTGCCA-3' (–811); –585 site, (–594) 5'-CCCCCGGCCCCCGCCCCCCGCCCCT-3' (–572); –210 site, (–220) 5'-CGCGGGGAGGGGCGGGCGCTCGGCG-3' (–196); –15 site, (–30) 5'-TCCGAACTCGCCGGCGGAGTCG-3' (–9). The nucleotides in bold correspond to the Sp1 core binding sequence; the underlined nucleotides correspond to the Sp1 consensus binding site.35
For competition, a consensus Sp1 site oligonucleotide (cSp1/3) and a mutated version (mSp1/3) (sc-2502 and sc-2503, respectively) were purchased from Santa Cruz Biotechnology. In addition, a Sp1/3–210 site methylated in all CpG residues (methyl-Sp1/3; Figure 4C
) and a tandemly arranged Sp1/Sp3 site (Sp1/3tn; Figure 4B
) from the NFATC1 P2 promoter were used. In supershift EMSAs, 1 µl of polyclonal Ab raised against Sp1 or Sp3 (sc-644x; Santa Cruz Biotechnology) was added to the incubations.
|
Real-Time PCR Assays
RNAs were extracted by TRIzol from cell lines treated with 5-aza-dC (1 to 5 µmol/L) for 4 days. RNAs were reverse transcribed using the iScript cDNA synthesis kit of Bio-Rad (Hercules, CA). SYBR Green real-time PCR assays were performed with an ABI PRISM 7700 sequence detection system according to the protocol of ABgene (Hamburg, Germany). For detecting NFATc1-, BOB.1/OBF1-, PLC2-, and Syk-specific RNAs, the following primers from Qiagen (Hilden, Germany) were used: Hs_NFATc1-1-SG (QT00094157); Hs_POU2AF1_1-SG (QT0.0001540), Hs_PLCg2 (QlT00050393), and Hs_SYK_1 (QT00050043). β-Microglobulin primers were used as an internal standard.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In immunohistochemical stainings of human hyperplastic tonsils and lymph nodes using a NFATc1-specific monoclonal antibody raised against all NFATc1 isoforms (see Materials and Methods), we observed cytosolic staining of virtually all lymphoid T and B cells (with the exception of a faint or no nuclear staining in some germinal center B cells) (Supplemental Figure S2, A and B, at http://ajp. amjpathol.org). No staining was observed in endothelial cells and epithelial cells and, as shown in double stainings with monoclonal antibodies against Ig light chains, in plasma B cells (Supplemental Figure S2, C and D). When we used the same Ab for immunostaining of a large panel of human lymphoid tumors, the cytoplasm of the majority of tumor cells was also positively stained (Table 1)
. Thus, a strong cytosolic expression of NFATc1 was found in the majority of mature B-cell neoplasms: All cases of B-cell chronic lymphoid leukemia, mantle cell lymphoma, follicular lymphoma, diffuse large B-cell lymphoma, mediastinal large B-cell lymphoma, and Burkitt’s lymphoma showed consistent NFATc1 staining, and in the majority of Burkitt’s lymphoma cells, a nuclear appearance of NFATc1 was observed. In contrast to the loss of NFATc1 expression in normal plasma cells, neoplastic plasma cells of 8 of 19 plasmacytoma/myeloma cases revealed a positive NFATc1 immunostaining (Figure 1
; Supplemental Figures S3 and S4 at http://ajp.amjpathol.org; Table 1
).
|
. The accompanying small T cells showed consistently high levels of NFATc1 expression, favoring these cells as endogenous controls in all NFATc1 stainings. In primary cutaneous ALCLs, 55% of the cases investigated showed a distinct decrease in or no NFATc1 immunostaining (Table 1
; Supplemental Figure S3A). In other peripheral T-cell lymphomas, the overwhelming majority of cases exhibited a positive NFATc1 immunostaining (Table 1
; Figure 1C
; Supplemental Figure S3B).
Similar to most ALCL cases, cHL cases showed a conspicuous loss or strong decrease in NFATc1 staining in their large Hodgkin Reed-Sternberg (HRS) cells (Figure 1E)
, and in only one case (3%), a heterogeneous NFATc1 immunostaining of moderate intensity was detected in about 50% of the Hodgkin Reed-Sternberg cells. In contrast to cHLs, in nodular lymphocyte predominance Hodgkin’s lymphoma, the tumor cells showed a strong, homogeneous NFATc1 positivity in the great majority of cases (Figure 1F
; Table 1
). These data show that like a number of other signaling molecules, NFATc1 expression is suppressed in ALCLs and cHLs, ie, in human lymphoid tumors with a defective immunoreceptor signaling.26-28
NFATC1 P1 Promoter DNA Is Hypermethylated in ALCLs and Classical Hodgkin’s Lymphoma Cells
The suppression of NFATc1 in these human lymphomas could be due to several molecular mechanisms, such as due to the genetic instability of the NFATC1 locus or repressive epigenetic silencing, including DNA methylation. To show which mechanism(s) might be involved, we tested six microsatellite NFATC1 markers with respect to a loss of heterozygosity in tumors from ALCL patients. However, in DNAs from six patients, we were unable to detect any loss of heterozygosity at the NFATC1 locus. In addition, Southern blot assays using NFATC1-specific gene probes did not reveal gross alterations of the NFATC1 gene structure in DNA from several HL lines (data not shown). These negative findings prompted us to investigate DNA methylation of NFATC1 promoter region in lymphoid tumors using PCR-based sequencing of genomic DNA on sodium-bisulfite modification.33
The promoters P1 and P2 of the NFATC1 gene spanning approximately 800 and 200 bp, respectively, are conserved between mouse and human and situated within CpG methylation islands (Figure 2A)
.29
The inducible P1 promoter can be subdivided into two highly conserved regions of approximately 250 bp, which we designated as the proximal and distal block of homology.29
Each block of homology contains multiple CpG dinucleotides, is flanked by Sp1/Sp3-like binding motifs, and harbors binding sites for several inducible transcription factors, such as for CREB/Fos/ATF-2 and NFAT factors and, in the distal block of homology, for nuclear factor-
B (Figure 2A)
. Because we know that introducing mutations into the tandemly arranged NFAT-binding site within the distal block of homology abolished almost all inducible promoter activity,29
we performed DNA methylation assays using PCR primers for the amplification of P1 DNA from positions –506 to –855.
In these assays, P1 DNA methylation was investigated in primary T cells from peripheral human blood; from Jurkat T leukemia cells; from the HL cell lines L428, L540, L1236, and KM-H2; and from the ALCL cell lines SR-786, SUP-M2, and KARPAS 299. Whereas in human primary effector T cells, NFATc1 is fully expressed on T-cell activation,31
depending on the subline, NFATc1 is either inducibly29
or constitutively expressed in Jurkat T cells, ie, NFATc1 expression does not show a marked increase on induction by T+I (Figure 2C)
. In the majority of tumor cells, ie, in all three ALCL lines and in the HL lines L428, L540, and L591, NFATc1 expression is suppressed, whereas in KM-H2 (and L1236) cells, a weak expression was observed (Figure 2C)
. This expression of NFATc1 agrees well with the methylation status of P1 promoter DNA: In peripheral blood lymphocyte T cells, the P1 DNA was free of methylation, whereas in Jurkat cells (showing constitutive expression), KM-H2 cells, and L1236 cells, 17 of central CpG residues were found to be free of methylation, and several flanking CpG residues appeared to be either partially or fully methylated. In contrast, in all ALCL cell lines and in the HL lines L428 and L540, all 29 CpG residues within the distal block of homology were fully methylated (Figure 2B)
. In agreement with published data on the DNA methylation of numerous other promoters,37
these data suggest an inverse relationship between P1 promoter DNA methylation and NFATc1 expression in ALCL and HL cell lines.
To demonstrate whether in these tumor cell lines NFATC1 gene expression is suppressed by P1 DNA methylation, we treated ALCL (SUP-M2) and HL (KM-H2, L428, and L540) cells for 4 days with 1, 2.5, or 5 µmol/L 5-aza-dC. When we measured the RNA concentrations in 5-aza-dC-treated cells, we observed a distinct increase in RNA levels by real-time PCR assays. Whereas in SUP-M2 ALCLs and in KM-H2 and L428 HL cells, a 3.5- to 6-fold increase in NFATc1 RNA concentrations was detected, in L540 HL cells, a 13–25-fold increase was detected on 3 days of 5-aza-dC treatment. This is a distinctly stronger increase in RNA expression than that observed for the expression of Bob.1 (Pou2af1) and SYK genes that are known to be controlled by DNA methylation (Figure 3)
.38,39
It shows that DNA methylation contributes to suppression of the NFATC1 locus both in ALCL and HL cells.
|
The NFATC1 promoters bear several binding motifs for Sp1, a transcription factor that was described to interfere with DNA methylation.40,41
Interestingly, these putative Sp1-binding sites appear to mark the borders of hypomethylated P1 DNA segments that we identified in Jurkat cells, KM-H2 HL cells, and L1236 HL cells (Figure 2B)
. To see whether these sites are bound by Sp1 in Jurkat and HL cells, we performed EMSAs using radioactively labeled probes, including the putative Sp1 sites around the positions –825 and –585. Incubation with nuclear proteins from Jurkat cells or HL cells resulted in the appearance of several retarded bands. Although factor binding to the distal –825 site appeared to be weaker than to the proximal –585 site, supershift assays with Abs raised against Sp1 and the related factor Sp3 indicate that Sp1 as well as Sp3 bind to both sites (Figure 4A
, the shifted Sp1-Ab bands labeled by arrows in lanes 3 and 7 and the disappearance of Sp3 complex in lanes 4 and 8). This conclusion is supported by the efficient competition of factor binding using a 50 mol/L excess of an unlabeled Sp1 consensus binding site (Figure 4A
, lanes 2 and 6).
Sp1/Sp3 binding could also be demonstrated for putative Sp1 sites around the positions –210 and –15 that flank the proximal block of P1 homology. EMSAs with a consensus Sp1 binding site (Sp1/3cons) illustrate a very similar factor binding as to the –210 and –15 sites (Figure 4B)
. Competition with a 50-fold excess of an unlabeled tandemly arranged Sp1/Sp3 site from the P2 promoter (Sp1/3tn) or a Sp1/Sp3 consensus site (cSp1/3) led to suppression of Sp1 and Sp3 complex formation, whereas a mutated Sp1/Sp3 site (mSp1/3) was unable to compete (Figure 4B)
. Likewise, incubation with Abs specific for Sp1 or Sp3 resulted in the supershift of Sp1 or Sp3 complexes, respectively (Figure 4B
, arrows). Moreover, methylation of DNA probe did not inhibit Sp1/Sp3 binding to the –210 site (Figure 4C)
, because it did not impair Sp1 binding to other sites.42
It is very likely that the Sp1/3 sites around the positions –210 and –15 also flank a window of 20 de-methylated CpG dinucleotides that is protected against DNA methylation.
When we compared factor binding to Sp1 sites in nuclear protein from Jurkat cells with that from HL cells, a weaker binding was detected in EMSAs using protein extracts from L428 and L540 HL cells compared with Jurkat cells (Figure 4A
, compare lanes 11 to 16 with lanes 9 and 10). Whereas in immunoblots, somewhat less Sp1 protein was detected in HL cells than in Jurkat cells (Figure 4D)
, in ChIP assays, the P1 promoter appeared to be unbound by Sp1 in L540 HL cells but bound in Jurkat cells (Figure 4E)
. These findings suggest that the binding of Sp1 contributes to DNA methylation status and thereby to the activity of NFATC1 P1 promoter.
Epigenetic Histone Modifications Reflect the Activity of NFATC1 Promoter in Lymphoma Cells
We also questioned whether HL cells that do not express NFATc1 differ in histone modification of their NFATC1 promoter chromatin from those that do express NFATc1. To distinguish active from inactive chromatin sites, we used Abs raised against acetylated histone H3, trimethylated H3-K4, and di- and trimethylated H3-K9 in ChIP assays. Both acetylated and trimethylated H3-K4 histones are indicative for transcriptional active chromatin, whereas methylated H3-K9 is a property of inactive chromatin.43-45
As shown in Figure 5, A and B
, in L540 HL cells that do not express NFATc1, the P1 region contains much less acetylated H3 than those in KM-H2 cells and Jurkat cells that express NFATc1, and trimethylated H3-K4 histones are under-represented at the P1 promoter in L540 cells compared with KM-H2 cells. In contrast, similar low levels of H3-K9 di- and trimethylation were observed for L540, KM-H2, and Jurkat cells, although HP1 protein that binds to the methylated H3-K946
was detected at the P1 promoter in L540 but not in Jurkat cells (Figure 5B)
. Contrary to P1 chromatin, almost identical levels of acetylated and methylated H3 histones and HP1 were detected at the β-actin and DHFR promoters in L540, KM-H2, and Jurkat cells (Figure 5)
.
|
, treatment of EL-4 cells for 2 days with low doses of CsA abrogated the induction of NFATc1 expression by T+I that mimic TCR signals in T cells. In contrast, no or a very weak suppressive effect of CsA treatment was observed on NFATc2 expression. At the P1 promoter, a distinct decrease was detected in histone H3 acetylation and H3-K4 trimethylation, whereas these histone modifications remained constant at the Nfatc2 promoter chromatin. Under the same conditions, we also tested the DNA methylation status of Nfatc1 promoter DNA, but we were unable to detect any methylation of CpG residues on CsA treatment of EL-4 cells at the P1 promoter (data not shown).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
We have shown here a close correlation between the DNA methylation of 5'-CpG island in the NFATC1 P1 promoter region and silencing of P1 promoter activity. In ALCL and cHL cell lines that do not express NFATc1, all CpG dinucleotides within P1 appeared to be methylated, and treatment of these cells by 5-aza-dC led to a re-activation of NFATc1 expression. Although tumor cells are often characterized by a global hypomethylation, the 5'-CpG islands in the promoter region of many genes, in particular of tumor suppressor genes, are often hypermethylated, thereby silencing these genes in tumors.37 DNA hypermethylation and transcriptional silencing have been described for the 5'-CpG islands of RB1, VHL, CDKN2A (p14ARF and p16INK4a), MSX2, and BRCA1 genes,37,47 and the NFATC1 gene is a further addition to this list.
In striking contrast to the silencing of NFATc1 expression, the NFATC2 gene, which also bears a 5'-CpG island in its promoter region, is active in cHL cells. This leads to the question of why the CpG islands of closely related genes become either hypermethylated or remain hypomethylated in tumors. Typical examples of such tumor suppressor genes are the BRCA1 and MSX2 (MLH/MSH) genes: Although the 5'-CpG island of the BRCA1 gene is hypermethylated in breast and ovarian tumors,48 that of the BRCA2 gene is not,49 and although the 5'-CpG island of the MSX2 (MLH1) gene is hypermethylated in colon, gastric, and endometrial tumors, those of MSH2, -3, and -6 genes are nonmethylated.50 It is remarkable that CpG island hypermethylation frequently affects genes that are controlled by two different promoters, ie, by a strong one that directs the most prominent transcripts and a weaker one that directs minor transcripts. Examples are the BRCA1, APC, CDKN2A (p14ARF and p16INK4a), and several other genes (see 37 for a more detailed discussion), including the NFATC1 gene. Although it is unknown why CpG islands of genes with two promoters are predominantly methylated, the recruitment of DNA methyltransferases (DNMTs) to promoters with a high CpG content (or other structural abnormalities) might be one explanation. The expression of both DNMT1 and DNMT3b is increased in several solid and hematological tumors,51,52 and high local concentrations of CpG, histone modifications, and/or the binding of heterochromatic proteins, eg, of HP1, might lead to the predominant recruitment of DNMTs to such chromosomal domains and genes.
Despite frequent hypermethylation of tumor suppressor genes controlled by two promoters, the complete silencing of the NFATC1 gene by DNA hypermethylation in ALCL and HL tumor cells is a surprising finding. Because of the activity of the inducible promoter P1, the NFATC1 gene is strongly induced on immunoreceptor and co-receptor stimulation at several stages of T-cell development.12,53
Under optimal conditions, this results in the massive synthesis of NFATc1/
A, a relatively short NFATc1 protein that differs remarkably in its N-terminal peptide and short C terminus from NFATc2 and from NFATc1/βC, whose synthesis is directed by the constitutive NFATC1 promoter P2. Although NFATc2 and NFATc1/βC exert a pro-apoptotic activity in T cells29
and Burkitt’s lymphoma B cells,54
NFTc1/A acts in a strong anti-apoptotic manner on ectopic expression in NFATc1-deficient DT40 chicken B cells (E. Kondo, unpublished data). Together with other experimental data,15,19
this favors NFATc1/A as an oncogene for the generation of (certain) lymphoid tumors. Although the contribution of NFATc1 proteins to tumor generation has to be elucidated in detail, the hypermethylation of the NFATC1 gene from which a protein with oncogenic properties is expressed casts doubt on the general view37
that only tumor suppressor genes but no oncogenes are hypermethylated in their 5'-CpG islands.
A further finding of general interest might be the detection of a "window of hypermethylation" in the NFATC1 P1 promoter (Figure 2)
. In several lymphoma cells in which NFATc1 is expressed at moderate levels, the binding sites for inducible transcription factors within P1 form such windows that are flanked by Sp1/Sp3 binding sites. Sp1/Sp3 sites are known to affect de novo DNA methylation,40,41
and it is likely that at the NFATC1 P1 promoter they suppress DNA methylation in a directional manner. More experiments have to show whether Sp1, Sp3, or other factors associating with these two transcription factors inhibit the activity of DNMTs.
The data from this study allow a preliminary view on the molecular events that take place at the NFATC1 P1 promoter region during the silencing of the NFATC1 gene. They indicate that persistent immunoreceptor-mediated signals are necessary to keep the NFATC1 locus open in lymphoid cells. In peripheral T lymphocytes, missing TCR signals lead at first to a drop in P1-directed transcripts (transcribing exon 1), whereas transcripts directed by P2 (transcribing exon 2) are continuously synthesized.53
The complete interruption of TCR-mediated signals by CsA (Figure 6)
shows that missing TCR signals lead to de-acetylation and decrease of H3-K4 trimethylation at the P1 promoter. Subsequently, such repressive events might result in hypermethylation of promoter DNA and in complete repression of the NFATC1 locus, as it occurs in ALCLs and cHLs.
It remains to be shown why the expression of NFATc1, including the synthesis of NFATc1/A, ie, a protein with oncogenic properties, is switched off in cHL and other lymphoid tumors. In contrast, the synthesis of nuclear factor-B, which shares structural properties with NFATs,1-3
and of AP-1, the most common partner of NFATs in activated T cells,55
is constitutively switched on in Hodgkin’s lymphoma. In other human lymphomas and tumors, the expression of NFATc1 is enhanced and appears to contribute to tumorigenesis (19
; E. Kondo, unpublished data).
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Mildred Scheel Foundation for Cancer Research, the Deutsche Forschungsgemeinschaft, the Wilhelm Sander Foundation, the Alexander von Humboldt Foundation, and the Hungarian Scientific Research Fund (OTKA-T 046663 KON).
A.A., L.K., and T.S. contributed equally to this work.
Supplemental material for this article can be found on http://ajp. amjpathol.org.
Accepted for publication October 1, 2007.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
