Autoimmune-polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED) is the only systemic autoimmune disease with a monogenic background known so far revealing no association with the major histocompatibility complex region. We have recently isolated the gene defective in this syndrome and characterized several different mutations in individuals with the disorder. The novel gene, AIRE, contains a putative bipartite nuclear targeting signal predicting a nuclear location of the corresponding protein. The presence of two PHD-type zinc finger domains as well as the newly described putative DNA-binding domain, SAND, in the amino acid sequence of the APECED protein implies that it may be involved in the regulation of gene expression. Using transient expression of AIRE cDNA in mammalian cells we demonstrate here the nuclear location of the APECED protein. Immunohistochemical staining of transfected cells revealed that most of the recombinant 58 kDa APECED protein is present in the form of nuclear dots. By double immunofluorescence labelling we further show that these APECED-containing structures and the previously described PML nuclear bodies are largely non-overlapping. The AIRE protein was also visualized in multiple human tissues: a subset of the cells in thymus, in spleen and in lymph node showed nuclear staining with APECED antiserum. Immunofluorescence labelling of peripheral blood mononuclear leukocytes also revealed a nuclear body-like staining pattern in a fraction of these cells. These data from both in vitro and ex vivo systems, together with the predicted structural features of the APECED protein, suggest that this protein is most probably involved in the regulation of gene expression.
Autoimmune-polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED, PGD type I, Aps type I; OMIM 240300) is an autosomal recessive monogenic disease (1). Of the systemic autoimmune diseases described so far it is the only one inherited in a Mendelian fashion. This rare disorder is enriched in isolated populations like the Finns (incidence 1:25 000) (2), the Iranian Jews (incidence 1:6000–9000) (3) and the Sardinians (4) and sporadic disease cases are reported worldwide. The classical triad of symptoms for APECED consists of hypoparathyroidism, primary adrenocortical failure and chronic mucocutaneous candidiasis. However, the clinical phenotype of the disease consists of variable combinations of failure of several endocrine and non-endocrine organs, including the parathyroid glands, adrenal cortex, gonads, pancreatic ß-cells, thyroid gland and gastric parietal cells. Additional features include, besides the chronic candidal infections, various ectodermal dystrophies, most characteristic being dystrophy of the dental enamel and nails. The most severe complications of the disease are the oral squamous cell carcinoma observed in a small proportion of patients and fulminant autoimmune hepatitis, which may be fatal. The disease usually manifests itself in childhood, but new symptoms may develop throughout life (2).
APECED is the first autoimmune disorder that has been mapped outside the HLA region on chromosome 6. We initially assigned the disease locus by linkage analysis to a 2.6 cM region on chromosome 21q22.3 (5). Locus homogeneity of the disease was demonstrated in patients originating from different populations, including Iranian Jews and Finns (6). After constructing a physical map covering the critical chromosomal region of 800 kb (7), we were able to isolate and characterize the defective gene in this disease by the traditional positional cloning strategy (8). The same gene was simultaneously cloned independently by another group (9).
The highest steady-state mRNA levels for the novel gene, which was named AIRE for autoimmune regulator, are observed in the thymus, pancreas, testis and fetal liver. The AIRE gene is predicted to encode a protein of 545 amino acids with a deduced molecular weight of 57.7 kDa. The protein is very rich in proline and no evident charge clusters or periodicity patterns can be recognized. The predicted secondary structure consists mainly of coils, in accord with the high proline content of the protein. The amino acid sequence harbours a putative bipartite nuclear targeting signal between amino acids 113 and 133, hinting at a nuclear location of the protein (8). The AIRE gene also contains two plant homeodomain (PHD)-type zinc fingers, motifs that are found in many nuclear proteins, including transcriptional co-activators and chromatin-modulating proteins of the polycomb and trithorax groups (10). The presence of these domains and the high proline content of the protein are compatible with a nuclear location and suggest a function in the regulation of gene expression. In addition to the PHD motifs, a newly characterized domain named SAND, which is suspected to function as a DNA-binding domain, is found between amino acids 189 and 264 in the APECED protein (11). Other proteins containing the SAND domain, including Sp100/Sp140 proteins and DEAF-1 DNA-binding transcription factor, are mainly modular chromatin-associated proteins, whose function is not yet fully understood.
In order to gain further understanding of the function of the APECED protein, we investigated the subcellular localization of the polypeptide by immunostaining in transiently transfected mammalian cell lines, in relevant human tissue sections and in peripheral lymphocytes. The results confirm the nuclear location of the APECED protein and its dispersion in a subgroup of nuclear bodies.
Expression of the APECED protein
To evaluate the protein coding capacities of the AIRE cDNA, in vitro translation and western blotting experiments were performed. In vitro transcription-translation of the ARE-pGEM3(Z)+ construct revealed a single 58 kDa polypeptide, which is in conformity with the calculated molecular weight of 57.7 kDa (Fig. 1A). For in vivo detection of the corresponding APECED protein, polyclonal rabbit antibodies were raised and affinity purified against a synthetic 17 amino acid peptide analogous to amino acid residues 151–167, located in a hydrophilic region of the APECED protein. The AIRE cDNA in the SV-poly vector was transiently expressed in transfected COS-1 cells. Western blotting of the transfected cell lysates with the affinity-purified peptide antiserum revealed two polypeptides, one of 76 and the other of 58 kDa. The antibodies recognized the 76 kDa polypeptide in both untransfected and the vector-transfected COS-1 cells, indicating that the 58 kDa peptide represents the specific APECED protein (Fig. 1B). The pre-immune sera did not visualize the 58 kDa polypeptide and the signal was totally abolished by pre-incubation of the antiserum with the peptide (data not shown). The absence of a difference between the mobilities of the APECED polypeptide expressed in vitro and in vivo would suggest that no significant post-translational modification of the polypeptide has taken place, at least in the COS-1 cells.
For double immunofluorescence assays, we produced an SV-poly APECED construct with an N-terminal FLAG tag and expressed this construct in COS-1 cells. The western blots stained with M2 antibody recognizing the FLAG epitope revealed a single band of ∼60 kDa, not identified in untransfected cells (Fig. 1C). Thus, the anti-FLAG antibody detected a polypeptide with a mobility slightly slower than the untagged protein visualized with the polyclonal antibody, providing further evidence of the specificity of this antibody.
Subcellular location of the APECED protein in transfected cells
For immunohistochemical analyses both the AIRE cDNA and the corresponding N-terminally FLAG-tagged expression constructs were transfected and expressed in multiple cell lines. The subcellular location of the protein was investigated by indirect immunofluorescence in conjunction with laser confocal microscopy, using the two antibodies described above. Figure 2A shows that the APECED protein expressed in COS-1 cells gives a prominent speckled pattern of fluorescence throughout the nucleoplasm, apart from the nucleoli. Immunostaining with the polyclonal APECED antiserum and the M2 anti-FLAG antibody overlapped totally, confirming that the polyclonal antibody detects protein coded by the transfected cDNA and not any endogenous protein (Fig. 3). The speckled pattern of APECED fluorescence was observed not only in COS-1 cells, but also in transfected HeLa, NIH 3T3 (Fig. 2C and E) and CV-1 cells (data not shown), all revealing lower expression levels. The number of nuclear body-like structures in the COS-1 cell nuclei varied from ∼30 to 200 and the size of these structures varied slightly from cell to cell. In ∼25% of the transfected cells, besides the nuclear dots, some perinuclear staining was also detected. This may have been due to overexpression of the APECED protein in the transient expression systems. Furthermore, in ∼20% of the transfected COS-1, NIH 3T3 and CV-1 cells a punctate and an intermediate filament-like cytoplasmic immunofluorescense pattern was observed in addition to the nuclear bodies and the proportion of transfected HeLa cells featuring cytoplasmic staining was even higher (Fig. 2B, D and F). The intermediate filaments stained by APECED antiserum showed complete co-localization with vimentin filaments in a double labelling experiment (data not shown). Fractionation of the transfected COS-1 cells into nuclear and cytoplasmic fractions confirmed that the majority of immunoreactive APECED polypeptide resided in nuclei (Fig. 4).
The APECED protein is localized in nuclear bodies separate from PML bodies
The dot-like nuclear distribution of fluorescence observed with APECED antiserum resembles that seen with many other proteins known to form nuclear bodies (NBs), e.g. promyelotic leukaemia protein (PML) (12–14), Sp100 (15–16) and Sp140 (17) or LYSP100 (18), etc. In order to investigate the nature of the nuclear stuctures containing APECED polypeptide, we performed experiments with double immunofluorescence labelling using both antiserum against APECED protein and antibodies against either PML or Sp100, the major components of the PML NBs. The APECED protein did not show any detectable co-localization with either of these proteins (Fig. 5). Moreover, overexpression of AIRE cDNA seemed to reduce the number of the PML bodies when compared with adjacent non-transfected cells. These observations would indicate that APECED forms independent spherical nuclear structures predominantly distinct from PML NBs and suggest that high level transient expression of APECED polypeptide either displaces PML and Sp100 from the NBs or disrupts them.
Distribution of APECED protein in human tissues
To determine the distribution of APECED protein in different human tissues some immunohistochemical labellings were undertaken. In thymus strong staining was seen in the thymic corpuscles (Fig. 6a). Labelling appeared in the epithelial cell nuclei and in some corpuscles the cytoplasm of the cells was also slightly labelled. The cortical thymocytes were negative in all samples whereas the medullary thymocytes exhibited faint to moderate staining in different specimens (Fig. 6b). Scattered cells with a larger labelled nucleus were detected in the medulla and close to blood vessels and beneath the capsule (Fig. 6c). These may represent epithelial reticular cells. In the spleen a large number of lymphocytes in the red pulp were labelled with faint to moderate intensity (Fig. 6d). A small number of lymphocytes exhibited strong staining. The proliferating lymphocytes in the white pulp were not labelled. At high magnification the labelling in lymphocyte nuclei was seen close to the nuclear membrane and in dots inside the nucleus. The neutrophilic granulocytes were very strongly labelled (Fig. 6). In lymph nodes the majority of the lymphocytes exhibited faint to moderate staining and part of the cells was clearly labelled. Strongly labelled neutrophilic granulocytes could be observed (data not shown).
In smears of peripheral blood leukocytes strong staining could be seen in a large population of lymphocytes, neutrophilic granulocytes and monocytes (Fig. 6g). APECED protein was also demonstrated ex vivo by immunofluorescent microscopy after cytocentrifugation on a microscope slide in a small fraction of peripheral mononuclear leukocytes (in ∼5% of the cells analysed) (Fig. 7). The present but low expression of the AIRE gene in peripheral leukocytes was also verified by RT-PCR (data not shown).
Many nuclear factors are present either partly or entirely in distinct bodies or subnuclear compartments that produce a punctate pattern when immunostained and analysed by indirect immunofluorescence. Some of these subnuclear bodies contain factors involved in the transcription and processing of RNA. Recently, a growing number of both subnuclear bodies and their components have been shown to be intimately linked with different human diseases (19). At present, at least four different types of NBs in the interchromatin compartment have been identified on the basis of their morphological appearance and composition. Coiled bodies containing p80 coilin, fibrillarin and several splicing factors may play a role in small nuclear ribonucleoprotein (snRNP) transport or maturation, or both (20). PML bodies ate nuclear domains probably involved in transcription. They are specifically disrupted in human acute promyelotic leukaemia. Infection with certain viruses also disrupts PML bodies, which suggests that they may somehow be involved in antiviral defence (21). PML bodies are mainly composed of PML and Sp100 proteins, although some additional proteins, like heterochromatin protein 1 (22), have been suggested to form aggregates with these structures. The GATA transcription factor containing nuclear bodies comprise a third group of NBs (23) and RNA cleavage bodies a fourth one (24).
Here we describe the cellular location of the protein defective in APECED disease both in vitro in cellular systems and in tissues and cells obtained ex vivo. Two different antisera were used to detect the transiently expressed APECED protein. The polyclonal antisera was raised against an APECED peptide and the monoclonal antiserum detected the FLAG epitope. This allowed double labelling experiments in cells transfected with FLAG-tagged constructs and confirmation of the specificity of the antiserum. The N-terminal tag did not interfere with intracellular processing or distribution of the expressed protein.
The subcellular location of the APECED protein was initially analysed in transiently transfected COS-1 cells. Recombinant APECED protein was found to be predominantly localized in nuclear speckles resembling NBs. Interestingly, the number and size of the nuclear dots varied from cell to cell. These data could reflect a modulation of the involved nuclear structures during the cell cycle. Double immunofluorescence assays disclosed that APECED protein seldom overlapped with the PML or Sp100 proteins, the major components of the PML NBs. This was somewhat surprising, since the APECED protein and the Sp100 protein, both containing the SAND domain in their amino acid sequence, are speculated to have arisen from a common ancestor and therefore to have similar function (11). Moreover, further investigations are still needed to clarify whether APECED protein represents a component of known nuclear structures or whether it is a part of a novel structure.
In addition to the nuclear labelling, depending on the cell type ∼20–60% of the transfected cells showed some cytoplasmic staining that appeared either as distinct granules or as intermediate filament-like staining. These findings could be due to overexpression of the APECED protein, which may saturate the nuclear transport mechanisms. However, the observation that these cytoplasmic structures are also detected in cell lines exhibiting lower expression levels, in particular in HeLa cells, does not support this line of reasoning. The presence of APECED protein in both the nucleus and the cytoplasm was further supported by the results of cell fractionation experiments and may indicate that the protein shuttles from the nucleus to the cytoplasm. Similar results on the localization of APECED protein in transfected cells are also reported by Rinderle et al. (25).
Immunohistochemical staining of human tissue sections from thymus, spleen, liver and lymph node, selected as relevant tissues for autoimmune processes, revealed preferentially nuclear staining in some subpopulations of cells. Further, APECED protein could also be demonstrated in some peripheral blood mononuclear leukocytes by us, as well as by Rinderle et al. (25). Additional experiments are still needed to reveal the precise character of APECED-positive cells and the pattern of expression of the AIRE gene in different cells of the human immune system.
At present, the biological function of the APECED protein still remains elusive. The protein contains two PHD domains, which are defined by a characteristic arrangement of cysteine and histidine residues, C4HC3, and are potential zinc coordination motifs (10,26). This domain is repeated four times in the Drosophila homeotic protein trithorax, a positive regulator of transcription that is responsible for maintenance of the transcription of homeotic selector genes (27). It is also present in TIF1, a potential co-activator protein for nuclear hormone receptors (28). Two other proteins containing this domain, CBP and p300, mediate transcriptional activation through the cAMP-responding transcription factor CREB (29). These sequence homologies provide preliminary indications that the APECED protein may participate in transcriptional regulation by either protein-DNA or protein-protein interaction (30).
The amino acid sequence of APECED protein harbours a newly described SAND domain, whose function is not well understood (11). The proteins containing this domain have in common a conserved sequence of ∼80 amino acids. The best clue to the function of the SAND domain is supplied by the DEAF-1 DNA-binding transcription factor (31), which also contains the novel domain. Confirmation of the DNA-binding function of the SAND domain in this protein would seem to suggest that all SAND proteins are DNA-binding transcription factors. This domain appears in contiguity with a variety of others, including the bromodomain, the PHD domain and the MYND finger, which is quite surprising, since bromodomains and PHD domains are seldom found in combination with DNA-binding domains. It is noteworthy that the SAND domain is also observed in Sp100/Sp140 proteins, which are well-established components of NBs. The corresponding parallel arrangement of domains in APECED protein and Sp140 would suggest that APECED protein, though highly diverged, shares a common ancestry with the Sp100 protein group and might therefore function similarly. The SAND proteins, which also contain PHD domains, are suggested to regulate gene expression through modulation of chromatin structure (11). Based on data presented here, the APECED protein is predominantly located in the nucleus and found in cells of tissues and peripheral blood relevant to the human immune system. Future functional analyses will show whether APECED protein truly is involved in the regulation of gene expression and also how the defective structure of this protein in APECED patients results in generalized autoimmune disease disturbing the function of multiple tissues.
Materials and Methods
To obtain a specific antibody against the APECED protein, two rabbits were immunized by s.c. injections with a soluble synthetic peptide of 17 amino acids (amino acids 151–167) in Freund's complete adjuvant. The primary injection with 50 µg of the peptide was followed by three additional boosts of 100 µg after 10 days and 6 and 10 weeks, respectively. The blood was collected 1 week after the last immunization and the antibodies directed against the AIRE peptide were affinity purified using nitrocellulose-bound CnBr. The antibody titres and specificity were determined by immunofluorescence and western blotting. For double stainings, a commercially available anti-FLAG M2 monoclonal mouse antibody binding to FLAG fusion proteins (Eastman Kodak, Cambridge, UK) was used. Other antibodies used were polyclonal rabbit anti-PML and polyclonal rabbit anti-Sp100, kindly donated by Drs H. de The, A. Dejean, H. Kamei, K. Howe and E. Solomon, and monoclonal mouse anti-vimentin, a kind gift from Prof. I. Virtanen.
The expression constructs
The AIRE cDNA was earlier isolated from a human adult thymus cDNA library (8). This cDNA was cloned into EcoRI sites of the mammalian expression vectors SV-poly (32), pcDNA3(+) (Invitrogen, Leek, The Netherlands) and pGEM3(Z)+ (Promega, Madison, WI). For double immunofluorescence labelling, AIRE cDNA was tagged N-terminally with the FLAG epitope recognized by the M2 monoclonal antibody. In vitro mutagenesis, performed with a Chamelon Double-stranded, Site-directed Mutagenesis kit (Stratagene, La Jolla, CA), was used to produce this construct and the full-length AIRE cDNA in SV-poly was used as a template. Restriction enzyme selection of single mutants was performed with a selection oligonucleotide located in the polylinker region of SV-poly, which destroyed the EcoRV restriction enzyme cleaving site. Mutant clones were identified by PCR amplification and analysed by solid phase sequencing (33).
Cell culture and transfection
African green monkey kidney cell lines COS-1 (ATCC CRL 1650) and CV-1 (ATCC CCL 70), the human cervical carcinoma cell line HeLa (ATCC CCL 2) and the mouse fibroblast cell line NIH 3T3 (ATCC CRL 1658) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and penicillin and streptomycin. One day before the transfection, the cells were seeded on 3 cm cell culture dishes at 4 × 105 cells/dish. The 85% confluent cells were transfected either with 2 mg of the AIRE cDNA construct in an SV-poly vector (32) or with the corresponding 5′-end FLAG-tagged construct, using the SuperFect transfection method (Qiagen, Hilden, Germany). Analyses were performed 48 h after the transfection. An SV-poly expression vector control was used to monitor background expression of APECED protein in COS-1 cells.
COS-1 cells grown for 24 h on 100 mm plastic dishes were transfected with the AIRE cDNA in SV-poly vector. Forty-eight hours after transfection the cells were washed twice with ice-cold phosphate-buffered saline (PBS), scraped from the dishes and pelleted. The pelleted cells were washed twice with isotonic buffer [10 mM HEPES, pH 7.4, 0.25 M sucrose, 0.1 mM phenylmethyl-sulfonyl fluoride (PMSF)] and then resuspended in this buffer supplemented with 2 mM EDTA. The cells were disrupted by Dounce homogenization (25 strokes) in 1 ml of the buffer. The nuclei were isolated by low speed centrifugation (500 g for 10 min) and the cytoplasmic fraction was removed. This step was repeated and the nuclei were then suspended in 10 mM HEPES, pH 7.4, supplemented with 0.25 M sucrose, 2 mM MgCl2 and 0.1 mM PMSF and incubated for 15 min on ice. The nuclei were pelleted (500 g for 10 min) and suspended in 1x Laemmli buffer.
In vitro translation of the APECED protein
The cDNA previously isolated from a human adult thymus cDNA library (8) was subcloned into the EcoRI site of the pGEM3(Z)+ vector (Promega). The plasmid was then used as a template in the TnT Reticulocyte Lysate System (Promega). An AGA construct in pGEM7(Z)+ (Promega) (34) was used as a positive control and the pGEM3(Z)+ vector as a negative control (Fig. 1A). Western blot analysis
For western blot analysis, total protein extracts of transfected COS-1 cells were separated on an 11% SDS-polyacrylamide gel and then electroblotted onto Hybond-N nitrocellulose membranes (Amersham, Buckinghamshire, UK). After blocking with 5% defatted dried milk solution, the filters were incubated with a 1:100 dilution of polyclonal antisera against the APECED peptide for 2 h at room temperature. Subsequently the blots were incubated with alkaline phosphatase-conjugated anti-rabbit IgG (Promega) at 1:7500 dilution. Each of the incubation steps was followed by three washes for 5 min in PBS. Visualization was performed with NBT-BCIP substrate (Promega).
Immunofluorescent cell staining and confocal microscopy
The transfected cells were cultured on gelatine-coated glass coverslips 24 h before fixation with 4% paraformaldehyde (PFA) in PBS at room temperature for 15 min. The cells were washed twice with PBS and permeabilized with 0.1% Triton X-100 in PBS for 15 min at room temperature. Non-specific staining was blocked with 0.5% bovine serum albumin (BSA) in PBS before addition of the primary antibody. The coverslips were then incubated with the primary antibody in 0.5% BSA in PBS for 2 h. The polyclonal rabbit antibody against the APECED peptide was used at 1:100 dilution, the polyclonal rabbit anti-PML serum at 1:200, the rabbit Sp100 antiserum at 1:200 and the mouse anti-vimentin antibody at 1:100 dilutions. The monoclonal mouse anti-FLAG antibody M2 was used at 1:500 dilution. When double labelling was performed, the cells were incubated with the two primary antibodies simultaneously. The coverslips were subsequently incubated with rhodamine-conjugated anti-rabbit IgG and fluorescein-conjugated anti-mouse IgG (Immunotech, Marseille, France), which were used at 1:150 dilution. The immunofluorescence was analysed using a Leica DMR confocal microscope with a 63× objective.
The peripheral blood mononuclear cell fraction was separated by Lymphoprep gradient centrifugation, density 1.077 g/ml (Nycomed Pharma, Oslo, Norway). For the ex vivo immunofluorescence analysis, peripheral blood mononuclear cells were cytocentrifuged (200 r.p.m. for 4 min, Cytospin 2; Shandon, Pittsburgh, PA) on a microscope slide and air dried overnight. The cells were fixed with 4% PFA for 30 min at room temperature and permeabilized with 1% Triton X-100 for 15 min at room temperature. Non-specific staining was blocked with 5% BSA in PBS, also used with the antibodies in the following detection steps. APECED protein was detected with the antibody against the APECED peptide and subsequently visualized with FITC-conjugated sheep anti-rabbit antibody at 1:50 dilution [F(ab′)2 fragment; Boehringer Mannheim, Mannheim, Germany] and DAPI counterstaining of the nuclear DNA was performed. As a control for the specificity of the primary and secondary antibodies, a pre-immune antiserum, as well as cytochemistry void of primary antibody, was performed (data not shown). A two-colour digital image analysis was used for aquisition and display, based on an Olympus BX50 microscope and Photometrics PXL camera (Photometrics, Tucson, AZ).
Human tissues (thymus, spleen and lymph node) were obtained after surgery and stored frozen at −70°C. Human peripheral blood leukocytes (PBLs) were obtained from healthy volunteers. Frozen sections (10 µm thick) were thawed on poly-L-lysine-coated glasses and fixed with 4% PFA and 0.2% 0.1 M picric acid in phosphate buffer (pH 7.3) for 5 min. PBLs were smeared on glasses and fixed similarly. Endogenous peroxidase activity was destroyed by treating the tissues with 0.3% hydrogen peroxide for 20 min. The tissues were incubated overnight at 4°C with the APECED antibody diluted 1:25–100 in PBS containing 1% BSA and 0.3% Triton X-100. Subsequently the sections were incubated with biotinylated goat anti-rabbit IgG (diluted 1:300) and ABC complex (Vectastain Elite Kit; Vector Laboratories, Burlingame, CA) for 30 min each. Immunoreaction was visualized with nickel-intensified DAB as chromogen. The sections were then dehydrated and embedded in Entellan. Sections were examined under a Nikon FXA microscope equipped with a PCO SensiCam digital camera (PCO, Kelheim, Germany). Images were processed using CorelDraw software (Corel, Ontario, Canada) and printed with an ALPS MD-2300 printer (ALPS Electric, Cork, Ireland). Controls included presaturation of the antibody with 20–40 µg/ml APECED peptide and pre-immune serum diluted 1:50. Pre-immune serum yielded strong staining in the cytoplasm of a large number of cells in all tissues studied but no nuclear staining. The affinity-purified APECED antibody stained mainly nuclei in all samples analysed but a small number of cells with cytoplasmic staining could be observed in different tissues. The cytoplasmic staining could not be abolished with presaturation and is thus considered non-specific. Most of the nuclear staining disappeared after presaturation. Only staining considered specific is described in Results.
We thank Anu Jalanko for giving expert methodological advice. Tuula Airaksinen, Lea Puhakka, Ulla Jukarainen and Riika Salmela are appreciated for exellent technical assistance. Aarno Palotie, Meelis Kolmer and Jorma Palvimo are thanked for valuable discussions. Drs H. de The, H. Kamei, A. Dejean, K. Howe and E. Solomon kindly provided the PML and Sp100 antibodies and Prof. I. Virtanen the anti-vimentin antibody. We thank Jean Margaret Perttunen for revision of the English language. This work was financially supported by grants from the Ulla Hjelt Fond of the Foundation for Pediatric Research, the Academy of Finland (P.B.), Medical Research Fond of Tampere University Hospital (M.P.-H.), Duodecim Foundation and by resources from the MD/PhD Program of the University of Helsinki, Finland (J.K.).