Deregulated expression of p16INK4a and p53 pathway members in benign and malignant myoepithelial tumours of the salivary glands.

AIMS: Myoepithelial salivary gland tumours are uncommon and follow an unpredictable biological course. The aim was to examine their molecular background to acquire a better understanding of their clinical behaviour. METHODS AND RESULTS: Expression of protein (E2F1, p16(INK4a), p53, cyclin D1, Ki67 and Polycomb group proteins BMI-1, MEL-18 and EZH2) was investigated in 49 benign and 30 primary malignant myoepithelial tumours and five histologically benign recurrences by immunohistochemistry and the findings correlated with histopathological characteristics. Benign tumours showed a higher percentage of cells with expression of p16(INK4a) pathway members [p16(INK4a) and E2F1 (both P < 0.001), and cyclin D1, P = 0.002] compared with normal salivary gland. Furthermore, malignant tumours expressed p53 (P = 0.003) and EZH2 (P = 0.09) in a higher percentage. Recurrences displayed more p53 + tumour cells (P = 0.02) than benign primaries. Amongst the benign tumours, the clear cell type had the highest proliferation fraction (P = 0.05) and a higher percentage of EZH2 was detected in the plasmacytoid cell type (P = 0.002). CONCLUSIONS: This study is the first to demonstrate that deregulation of the p16(INK4a) senescence pathway is involved in the development of myoepithelial tumours. We propose that additional inactivation of p53 in malignant primaries and benign recurrences contributes to myoepithelial neoplastic transformation and aggressive tumour growth.

Expression of the p16(INK4a) gene product, methylation of the p16(INK4a) promoter region and expression of the polycomb-group gene BMI-1 in squamous cell lung carcinoma and premalignant endobronchial lesions.

It is generally assumed that squamous cell carcinoma develops in a stepwise manner from normal bronchial epithelium towards cancer by the accumulation of (epi)genetic alterations. Several mechanisms including mutations and homozygous deletions or hypermethylation of the p16(INK4a) promoter region can cause loss of p16 expression. Recent studies suggest overexpression of the polycomb-group gene BMI-1 might also down-regulate p16 expression. In this study, we analyzed the p16 expression in relation to the methylation status of the p16 promoter region of the p16(INK4a) gene and the expression of BMI-1 in bronchial squamous cell carcinomas (SCC) and its premalignant lesions. Nine (69%) SCC showed loss of p16 expression and 10 (77%) showed expression of BMI-1. Of four p16 positive samples two (50%) were BMI-1 positive, whereas among nine p16 negative samples, eight (89%) revealed BMI-1 staining. Four (44%) p16 negative samples were hypermethylated at the p16(INK4a) promoter region; the other p16 negative tumors that showed no hypermethylation revealed BMI-1 staining. Only two premalignant lesions showed absence of p16 expression, of which one (carcinoma in situ) was hypermethylated at the p16(INK4a) promoter region and the other (severe dysplasia) showed BMI-1 expression. In total, 11 precursor lesions (48%) revealed BMI-1 expression. In conclusion, the results of this study suggest that loss of p16 expression by promoter hypermethylation is inconsistently and occurs late in the carcinogenic process at the level of severe dysplasia. To what extent overexpression of the polycomb-group protein BMI-1 attributes to down regulating of p16 expression remains unclear.

T cell expansions in lymph nodes and peripheral blood in HIV-1-infected individuals: effect of antiretroviral therapy.

OBJECTIVE: To evaluate dynamics in CD8 T cell expansions during highly active antiretroviral therapy (HAART). DESIGN: Various T cell subsets were isolated from blood and lymph nodes and analysed for T cell receptor (TCR) diversity. METHODS: TCR complementarity determining region 3 (CDR3) spectratyping and single-strand conformation polymorphism (SSCP) analyses were performed in combination with sequencing to assess clonality of the subsets. RESULTS: Strongly skewed CDR3 patterns in total CD8 cells and the CD8 subsets CD45RO+CD27+ and CD45RO-CD27+ showed substantial dynamics in dominant CDR3 sizes, resulting in relative improvement of CDR3 size diversity in the first months of therapy. During sustained treatment, TCR diversity changed only moderately. SSCP profiles confirmed oligoclonality of TCR CDR3 perturbations. Various dominant CDR3 sizes for CD4 and CD8 T cells present in lymph nodes, but not in peripheral blood mononuclear cells, before the start of therapy emerged in peripheral blood early during therapy. CONCLUSIONS: HAART induces substantial changes in CD8 TCR diversity, eventually resulting in improvement of the repertoire. Clonal expansions observed in lymph nodes before therapy were observed in peripheral blood after therapy, suggesting that recirculation of CD4 and CD8 T cells from lymph nodes contributes to the early T cell repopulation. Decreased immune activation and possibly naive T cell regeneration subsequently decreased clonal expansions and perturbations in the CD8 TCR repertoire.

Diversity of the T-cell receptor BV repertoire in HIV-1-infected patients reflects the biphasic CD4+ T-cell repopulation kinetics during highly active antiretroviral therapy.

OBJECTIVES: Highly active antiretroviral therapy (HAART) induces a decline in viral load and a biphasic increase in peripheral blood CD4+ T-cell counts in HIV-infected patients. To evaluate the effect of HAART on T-cell receptor (TCR) diversity of repopulating naive and memory CD4+ T cells, complementarity determining region 3 (CDR3) spectratyping was performed. DESIGN: For four patients treated with HAART, CD45RO+ (memory) and CD45RA+ (naive) CD4+ T cells were isolated from peripheral blood leukocyte samples obtained 1 week before, 1-2 months after, and 9-11 months after start of treatment. METHODS: CDR3 regions were amplified by TCR-BV-specific nested PCR from CD4+ T-cell subsets. CDR3 size distributions and single-strand conformation polymorphism profiles were compared as an indication for TCR diversity. RESULTS: Increasing blood CD4+ T-cell counts during the first 2 months of treatment coincided with increased perturbation of CDR3 patterns in CD4+ T-cell subsets, suggesting an early oligoclonal repopulation. At later timepoints, CDR3 size diversity increased when T-cell counts did not substantially decrease. Memory and naive CD4+ T cells generally showed comparable levels of perturbation. CONCLUSION: Diversity of the TCR repertoire reflected biphasic T-cell repopulation during HAART, compatible with initial redistribution and later CD4+ T-cell production. Sustained elevation of T-cell counts will in principle result in restoration of TCR diversity.

Cloning of size-selected human immunoglobulin heavy-chain rearrangements from third complementarity-determining region fingerprint profiles.

Methods have been developed to rapidly visualize the size distribution of third complementarity-determining regions (CDR3) in immunoglobulin (Ig) and T-cell receptor (TCR) molecules. DNA fragments spanning the Ig or TCR CDR3 are generated by PCR using primers at fixed positions in the variable and constant segments. These fragments differ in length due to size variation of the CDR3s. Visualization of the amplification products in polyacrylamide gels as a "CDR3 fingerprint profile" is a rough measure for the complexity of the Ig and TCR antigen-binding specificities. We report an adaptation of this method for the analysis of human Ig heavy-chain genes that incorporates silver staining, which allows for the fine analysis of specific regions of the profiles. This is especially useful for the study of low-abundant transcripts.

Reconstitution of the B cell repertoire after bone marrow transplantation does not recapitulate human fetal development.

Immune reconstitution during bone marrow transplantation has been proposed to produce a fetal-type immune system. This characteristic may contribute to the relative immunodeficiency that occurs in the early post-transplant period. This review reappraises recent studies of immunoglobulin heavy chain genes produced by the recovering immune system. Comparison of these genes to those that are generated by fetal and adult B cells, demonstrates that there is no evidence to support the conclusion that adult lymphocytes in the graft reverse to a fetal stage of differentiation. In terms of lymphocyte diversity, the inadequacy of the recovering immune system is more likely to be explained by a combination of other factors - such as the delayed occurrence of somatic hypermutation and class switching, and clonal dominance.

Ig heavy chain CDR3 size diversities are similar after conventional peripheral blood and ex vivo expanded hematopoietic cell transplants.

It is largely unknown whether the immune repertoire can be reconstituted successfully after high-dose chemotherapy and transplantation using ex vivo expanded hematopoietic stem cell (HSC) grafts. It is critically important for the transplant outcome that immune repertoire reconstitution progresses after ex vivo expanded HSC graft transplants at least as efficiently as that seen after conventional HSC transplants. Previously, we showed that the T cell receptor V beta (TCRVB) third complementarity determining region (CDR3) diversification after ex vivo expanded bone marrow (BM) HSC graft transplants was similar to that seen after conventional peripheral blood stem cell transplants (PBSCTs). In the present study, the CDR3 diversity of the six immunoglobulin (Ig) heavy chain variable region gene (V(H)) families was examined in five breast cancer patients who were transplanted with ex vivo expanded BM HSCs as the only source of stem cells. For comparison, 12 healthy adults and four conventional PBSCT recipients were also studied. Using both CDR3 fingerprinting and single strand conformation polymorphism (SSCP) methodologies, it is shown that the contribution of the V(H) families to the overall repertoire among healthy adults is highly variable and not always proportional to V(H) family member size. After both ex vivo expanded HSC transplants and conventional PBSCTs, the V(H) CDR3 repertoires were limited in size diversity at 6 weeks post transplant. By 6 months, however, V(H) families displayed a repertoire diversity that was as complex as that seen in healthy adults. No difference was seen between ex vivo expanded HSC graft transplant recipients and conventional PBSCT recipients in V(H) repertoire diversity. In one patient there was a follow-up analysis 12 months after ex vivo expanded graft transplant, and the diversity of the V(H) families was maintained. In all patients, the amino acid size of the CDR3 regions fell within adult limits at all time points post transplant. These results indicate that B cell repertoire regeneration after ex vivo expanded hematopoietic cell graft transplants is similar to that seen after conventional PBSCT.

High expression of Polycomb group protein EZH2 predicts poor survival in salivary gland adenoid cystic carcinoma.

BACKGROUND: The prognosis of adenoid cystic carcinoma (ACC), a malignant salivary gland tumour, depends on clinicopathological parameters. To decipher the biological behaviour of ACC, and to identify patients at risk of developing metastases, additional markers are needed. METHODS: Expression of the cell cycle proteins p53, cyclin D1, p16(INK4a), E2F1 and Ki-67, together with the Polycomb group (PcG) proteins BMI-1, MEL-18, EZH2 and EED was investigated immunohistochemically 21 formalin-fixed, paraffin-embedded primary ACCs in relation to tumour characteristics. RESULTS: ACC revealed significantly increased expression of the cell cycle proteins compared to normal salivary tissue (n = 17). Members of the two PcG complexes displayed mutually exclusive expression in normal salivary gland tissue, with BMI-1 and MEL-18 being abundantly present. In ACC, this expression pattern was disturbed, with EZH2 and EED showing significantly increased expression levels. In univariate analysis, presence of recurrence, poor differentiation and high EZH2 levels (>25% immunopositivity) significantly correlated with unfavourable outcome. ACCs with high proliferative rate (>25% Ki-67 immunopositivity) significantly correlated with high levels of EZH2 and p16. Only the development of recurrence was an independent prognostic factor of survival in multivariate analysis. CONCLUSIONS: Expression of PcG complexes and of essential cell cycle proteins is highly deregulated in ACC. Also, EZH2 expression has prognostic relevance in this malignancy.

Polycomb-group genes as regulators of mammalian lymphopoiesis.

Polycomb proteins form DNA-binding protein complexes with gene-suppressing activity. They maintain cell identity but, also, contribute to the regulation of cell proliferation. Mice with mutated Polycomb-group genes exhibit various hematological disorders, ranging from the loss of mature B and T cells to development of lymphomas. Lymphopoiesis in humans is associated with characteristic expression patterns of Polycomb-group genes in defined lymphocyte populations. Collectively, these results indicate that Polycomb-group genes encode novel gene regulators involved in the differentiation of lymphocytes. The underlying mechanism is related, most probably, to gene silencing by chromatin modification, and might affect proliferative behavior and account for the irreversibility of lineage choice.

Analysis of clonal diversity in mouse immunoglobulin heavy chain genes selected for size of the antigen combining site.

Size-diversity of Ig and T cell receptor antigen binding (CDR3) regions can be visualized by "CDR3 fingerprinting", and provides an estimate of B- or T-cell repertoire complexity. The method does not identify clonal diversity, however, which can only be determined by random sequencing of the CDR3s. In this study we demonstrate that a combination of fingerprinting and single strand conformation polymorphism (SSCP) analysis can be used for a rapid estimation of clonal diversity within mouse Ig antigen binding regions selected for size. This application may be useful in the analysis of clonal expansion within B- and T-cell repertoires.

Ig heavy chain third complementarity determining regions (H CDR3s) after stem cell transplantation do not resemble the developing human fetal H CDR3s in size distribution and Ig gene utilization.

Previous studies have suggested that the B-cell repertoire after stem cell transplantation resembles the developing repertoire in the fetus. Fetal and adult repertoires differ strikingly at the molecular level in Ig heavy chain third complementarity determining region (H CDR3) size distribution and Ig gene utilization. Previously, the posttransplant repertoire has not been studied fully in this regard. In this study, we analyzed H CDR3s posttransplant using CDR3 fingerprinting, single-strand conformation polymorphism (SSCP), and random sequencing. Eleven adult patients who received either autologous (n = 6) or allogeneic adult sibling (n = 5) hematopoietic stem cell transplants were studied. IgM H CDR3 repertoires demonstrated limited clonal diversity within the first 6 to 10 weeks posttransplant. By 3 to 4 months, the IgM H CDR3 repertoires were as diverse as those in healthy adults. Reconstitution of the IgM diversity correlated with the expansion of the multimember VH3 family. By contrast, the contribution of the single-member VH6 family was limited in most patients up to 6 to 9 months. No evidence was seen for greater contribution of VH6 posttransplant. IgG repertoires remained clonally restricted at all times. In all patients, H CDR3 sizes fell within adult limits. Direct nucleotide sequencing of H CDR3s showed adult-type N-nucleotide insertions and Ig gene utilization. These results indicate that the emerging repertoire posttransplant does not resemble the developing fetal repertoire at the molecular level.

Human Ig heavy chain CDR3 regions in adult bone marrow pre-B cells display an adult phenotype of diversity: evidence for structural selection of DH amino acid sequences.

Ig repertoires generated at various developmental stages differ markedly in diversity. It is well documented that Ig H chain genes in human fetal liver are limited with regard to N-regional diversity and use of diversity elements. It is unclear whether these characteristics persist in pre-B cell H chain genes of adult bone marrow. Using Ig H chain CDR3 fingerprinting and sequence analysis, we analyzed the diversity of Ig H chain third complementarity determining regions (HCDR3) in adult bone marrow pre-B and mature B lymphocytes. Pre-B cell HCDR3 sequences exhibited adult characteristics with respect to HCDR3 size, distribution of N regions and usage of diversity elements. This suggested that pre-B cells in adults are distinct from fetal B cell precursors with regard to Ig H chain diversification mechanisms. At the DNA sequence level, HCDR3 diversity in mature B cells was similar to that in pre-B cells. Pre-B HCDR3s, however, frequently contained a consecutive stretch of hydrophobic amino acids, which were rare in mature B cells. We propose that highly hydrophobic pre-B HCDR3s may be negatively selected on the basis of structural limitations imposed by the antigen binding site. At the same time, usage of hydrophilic HCDR3 sequences (thought to support HCDR3 loop formation) may be promoted by positive selection.

Restricted utilization of germ-line VH3 genes and short diverse third complementarity-determining regions (CDR3) in human fetal B lymphocyte immunoglobulin heavy chain rearrangements.

Twenty-one independent immunoglobulin heavy chain VH3DJH rearrangements were cloned and sequenced from livers of human fetuses at 7, 13 and 18 weeks of gestation. The VH elements expressed were not somatically mutated. Eight out of the estimated 30 VH3 elements were utilized with a preference for five of them. One of these VH3 sequences, designated FL13-28, represented a thus-far unknown VH3 gene segment. From the six functional JH elements the JH3 and JH4 segments were utilized preferentially and from the estimated 30 D segments the DQ52 element and the Dxp family were found to rearrange frequently. D elements were utilized both in normal and inverted orientation, as single copies or in D to D fusions. Addition of N nucleotides, removal of nucleotides from the coding sequences and utilization of DIR elements (D genes with irregular recombination signals) further expanded the third complementarity-determining region (CDR3) diversity. One fourth of the fetal CDR3 regions lacked N regions. Due to utilization of DQ52, the relative absence of N regions and extensive exonuclease activity operating on the D elements, the fetal CDR3 regions were significantly shorter than those found in adult B lymphocytes.

Expression pattern of the most JH-proximal human VH gene segment (VH6) in the B cell and antibody repertoire suggests a role of VH6-encoded IgM antibodies in early ontogeny.

We have developed a mAb (JE-6) that recognizes an Id encoded by the most JH-proximal human VH gene segment (VH6) in or near germ-line configuration. This mAb was used to determine the frequency of Id JE6+ B cells in large collections of monoclonal EBV-transformed and short term B cell lines derived from fetal, neonatal, and adult lymphoid tissues. Moreover, we investigated the presence of Id JE-6+ Ig in sera from neonates and adults and determined the (auto)antigen binding properties of VH6-encoded IgM mAb. We detected a fivefold overrepresentation of VH6-expressing IgM producing B cells in fetal tissues, cord blood, and adult bone marrow relative to adult blood. In cord blood, but not in adult blood sera, germ-line VH6-encoded IgM molecules were readily detectable. IgM secreted by VH6-expressing B cell clones displayed highly conserved and virtually identical autoantigen binding properties, independent of the length and composition of the IgH chain CDR3 region and L chain isotype. Collectively, these results suggest that the VH6 gene and the antibodies it encodes play an important role in early human ontogeny.

Non-random employment of V beta 6 and J beta gene elements and conserved amino acid usage profiles in CDR3 regions of human fetal and adult TCR beta chain rearrangements.

We have studied the usage of V beta 6, D beta, and J beta elements, and the composition of the CDR3 regions of human fetal TCR beta chain rearrangements in a 17 week old fetal thymus and in fetal liver, bone marrow, spleen, and cord blood at 11 and 13 weeks of gestation. These fetal sequences were compared with TCR beta chain rearrangements obtained from post-partum thymus, adult spleen, and adult peripheral blood mononuclear cells. Both fetal and adult TCR V beta 6 rearrangements exhibited a non-random usage of V beta and J beta elements. Up to 90% of the sequences obtained at 11 weeks of gestation used J beta 2 elements, most notably J beta 2.1. In the 13 and 17 week old fetal and in the adult tissues, J beta 1 elements were used in approximately 30% of the rearrangements while, within the J beta 2 rearrangements, J beta 2.1 and J beta 2.7 were used most frequently. Both fetal and adult TCR beta chain CDR3 regions showed non-random usage of amino acids. However, the early fetal repertoire was further limited due to the relative absence of N-regions in up to 60% of the 11 and 13 week old TCR beta chain rearrangements, resulting in smaller antigen binding sites. In fetal and adult TCR beta chain rearrangements the distribution patterns of the length of N-regions and the usage profiles of J beta elements were similar in hematopoietic and peripheral organs, suggesting no apparent preference for particular TCR beta chain rearrangements.(ABSTRACT TRUNCATED AT 250 WORDS)

T cell receptor repertoire diversity and clonal expansion in human neonates.

Newborn human infants, particularly those born prematurely, are susceptible to infection with a variety of microorganisms. We questioned whether limitations in the T cell repertoire contribute to the neonatal immunocompromised state. To describe developmental changes of the T cell repertoire, cDNA segments corresponding to third complementarity regions (CDR3) of human umbilical cord blood T cell receptors (TCR) from 24-41-wk gestational age were amplified with TCR family-specific probes. The resulting amplified CDRs were visualized by fingerprinting and single strand conformation polymorphism (SSCP) analysis. At 24-wk gestation there were no limitations in TCRBV family usage, and the degree of CDR3 size heterogeneity was not different from the adult. However, earlier in gestation, CDR3s were shorter for all families and gradually increased in size until term. The extent of oligoclonal expansion observed in cord blood was greater than in adult peripheral blood (p = 0.03). T cell oligoclonal expansion was greatest at 29-33-wk gestation and declined toward term. Expansions were detectable in both CD4+ and CD8+ subpopulations. Our findings indicate that the genetic mechanisms of repertoire diversification appear intact as early as 24 wk of gestation, but repertoire diversity is limited as a result of smaller CDR3 sizes. In addition, there was a developmentally regulated progression of oligoclonally expanded T cells. These differences in the TCRBV repertoire add to the body of evidence demonstrating immaturity of the neonatal immune system. However, the role that these subtle differences are likely to play in the relative immunodeficiency of the neonate remains to be determined.

Ex vivo-expanded hematopoietic cell graft recipients exhibit T cell repertoire diversity similar to that seen after conventional stem cell transplants.

The feasibility of using ex vivo-expanded hematopoietic progenitor cells to reconstitute hematopoiesis after high-dose chemotherapy is presently being examined. Early studies have shown that myeloid and erythroid hematopoiesis can be successfully reconstituted after high-dose chemotherapy and ex vivo-expanded hematopoietic cell transplantation. The lymphoid reconstitution, however, has not been addressed previously. In this study, we examined the diversity of the T cell receptor V beta chain (TCRBV) repertoires in 5 breast cancer patients who were transplanted with ex vivo-expanded bone marrow mononuclear cells as the only source of hematopoietic graft. Using the TCRBV third complementarity determining region (CDR3) fingerprinting methodology, it is shown that CD4(+) and CD8(+) T cell subsets after ex vivo-expanded hematopoietic cell graft transplants exhibit TCRBV diversities that are similar in complexity when compared to those seen after conventional autologous peripheral blood stem cell transplants (PBSCT). No apparent difference in the extent of CDR3 diversity was found between ex vivo expanded and conventional autologous PBSCT recipients when the CD4(+) and CD8(+) subsets were further separated into CD45RA(+) "naïve" and CD45RO(+) "memory" subsets. The diversity of the CD45RA(+) naïve subsets was as complex as that of the CD45RO(+) memory subsets. These results indicate that T cell repertoire diversification is not further compromised when ex vivo-expanded hematopoietic cells are used instead of autologous peripheral blood stem cells as the only source of graft.

Expression of the human immunoglobulin heavy chain VH6 gene element by fetal B lymphocytes.

Fetal B lymphocytes in mice and humans use a limited number of the available VH gene segments. Mouse fetal B cells primarily utilize 3' VH elements, suggesting that the localization of these elements determines their rearrangement frequency. The previously reported non-random usage of human VH genes has been more difficult to explain. In this study the authors analysed the expression of the most proximal 3' human VH element (VH6) using a monoclonal antibody (JE-6). VH6 expression was assessed in various B cell differentiation stages from fetal liver, bone marrow and spleen at 12-20 weeks of gestation. The authors demonstrate that the level of VH6 expression does not exceed a stochastic usage frequency. This suggests that the localization of VH6 does not significantly promote its expression during human fetal life, and that other factors must affect the usage of VH genes during human fetal development.

Cutting edge: polycomb gene expression patterns reflect distinct B cell differentiation stages in human germinal centers.

Polycomb group (Pc-G) proteins regulate homeotic gene expression in Drosophila, mouse, and humans. Mouse Pc-G proteins are also essential for adult hematopoietic development and contribute to cell cycle regulation. We show that human Pc-G expression patterns correlate with different B cell differentiation stages and that they reflect germinal center (GC) architecture. The transition of resting mantle B cells to rapidly dividing Mib-1(Ki-67)+ follicular centroblasts coincides with loss of BMI-1 and RING1 Pc-G protein detection and appearance of ENX and EED Pc-G protein expression. By contrast, differentiation of centroblasts into centrocytes correlates with reappearance of BMI-1/RING1 and loss of ENX/EED and Mib-1 expression. The mutually exclusive expression of ENX/EED and BMI-1/RING1 reflects the differential composition of two distinct Pc-G complexes. The Pc-G expression profiles in various GC B cell differentiation stages suggest a role for Pc-G proteins in GC development.

Differential expression of human Polycomb group proteins in various tissues and cell types.

Polycomb group proteins are involved in the maintenance of cellular identity. As multimeric complexes they repress cell type-specific sets of target genes. One model predicts that the composition of Polycomb group complexes determines the specificity for their target genes. To study this hypothesis, we analyzed the expression of Polycomb group genes in various human tissues using Northern blotting and immunohistochemistry. We found that Polycomb group expression varies greatly among tissues and even among specific cell types within a particular tissue. Variations in mRNA expression ranged from expression of all analyzed Polycomb group genes in the heart and testis to no detectable Polycomb group expression at all in bone marrow. Furthermore, each Polycomb group gene was expressed in a different number of tissues. RING1 was expressed in practically all tissues, while HPH1 was expressed in only a few tissues. Also within one tissue the level of Polycomb group expression varied greatly. Cell type-specific Polycomb group expression patterns were observed in thyroid, pancreas, and kidney. Finally, in various developmental stages of fetal kidney, different Polycomb group expression patterns were observed. We conclude that Polycomb group expression can vary depending on the tissue, cell type, and development stage. Polycomb group complexes can only be composed of the Polycomb group proteins that are expressed. This implies that with cell type-specific Polycomb group expression patterns, cell type-specific Polycomb group complexes exist. The fact that there are cell type-specific Polycomb group targets and cell type-specific Polycomb group complexes fits well with the hypothesis that the composition of Polycomb group complexes may determine their target specificity. J. Cell. Biochem. Suppl. 36: 129-143, 2001.