Transdifferentiation in Lymphoid Malignancies Although B cells, T cells, histiocytes and dendritic cells all derive from a common stem cell, it has generally been believed that once lineage commitment takes place, reversion to another lineage does not occur. Recent studies in murine systems have suggested that modulation of transcription factors in vitro can lead to reprogramming of B-cells into macrophages. However, it is not known whether similar events take place in vivo, and under what circumstances they might occur. We have recently reported reprogramming of precursor B-cells and T-cells into histiocytes and Langerhans cell respectively. In both situations, patients with precursor B-cell or T-cell lymphoblastic lymphoma/leukemia (LBL/ALL) developed a tumor lacking phenotypic evidence of B-cells or T-cells, but exhibiting markers of histiocytes and Langerhans cells. The histiocytic and dendritic cell neoplasms were clonally related to their B-cell and T-cell counterparts, and demonstrated identical clonal IgH and TCR gene rearrangements. The above cases raise the question of whether one tumor lineage was transdifferentiated into the other, or if both arose from a malignant transformation of a common stem cell, as is the case in chronic myelogenous leukemia and many other myeloproliferative disorders. In the cases reported by us, the second hypothesis would presuppose the existence of a state of primitive differentiation in which IgH or T-cell receptor (TCR)gene rearrangement preceded the development of the neoplastic clone, and its subsequent differentiation towards a histiocytic or dendritic cell lineage. The existence of such a precursor is unlikely, given the characteristic germline configuration of the IgH and TCR genes in de novo histiocytic or Langerhans cell neoplasms. The possibility of lineage switching is supported by recent data showing the capacity of differentiated B cells to be reprogrammed immunophenotypically and functionally into macrophages. It is unclear, however, why reprogramming of this kind would occur in vivo. One hypothesis is that with treatment of acute leukemia, chemosensitive cells with lymphoblastic differentiation are killed, while cells of the same clone preferentially exhibiting histiocytic differentiation are more chemoresistant and survive to present later as histiocytic neoplasms. However, in at least one instance the Langerhans cell tumor arose simultaneously with precursor-T cell ALL, indicating that therapy did not play a role in driving the reprogramming of the T-cell neoplasm. In the above instances, programming occurred in a neoplastic cell with an immature, LBL phenotype. Reprogramming of a mature lymphoid malignancy seems less likely to occur. The literature contains rare reports of histiocytic or dendritic cell neoplasms in patients with mature B-cell malignancies. In a paper to be submitted shortly for publication we provide evidence that histiocytic sarcomas can arise in patients with follicular lymphoma. The histiocytic tumors are clonally related to the original B-cell neoplasm, and carry the BCL-2/JH translocation characteristic of follicular lymphoma. Epigenetic profiling of follicular lymphomas (FLs) to detect new diagnostic and therapeutic targets. In collaboration with J. Keith Killian, M.D., Ph.D., a former resident, and currently an Assistant Clinical Investigator in LP, and Paul Meltzer, M.D., Ph.D., Chief, Genetics Branch, CCR, NCI, we are performing epigenetic profiling of FLs. Methodology developed in their lab at the NCI permits the examination of the genome-wide methylation status of DNA recovered from formalin-fixed paraffin-embedded (FFPE) sections. Epigenetic DNA programming involves somatic, heritable, lineage-specific, covalent, enzymatic DNA modifications such as cytosine methylation. Epigenetic programming occurs as a natural part of cell and tissue growth and differentiation, and also in pathologic permutations thereof. Experimental evidence shows that epigenetic modifications of DNA correlate in many cases with changes in the functional activity of genes, such as transcription. Profiling of DNA methylation at CpG sites is a promising methodology for pathologic classification of tumors, including identification of patient- and tumor-specific molecular features that may be of clinical therapeutic significance. Recently, microarray genotyping platforms have been adapted for high-throughput, pan-epigenomic profiling, such that >1500 CpG targets can be simultaneous profiled in multiple clinical samples. One platform for high-throughput methylation analysis is the Illumina Bead Array platform. In a collaborative pilot study to test the performance and utility of this technology, Dr. Killian profiled DNA methylation in a set of FL and FH lymph nodes. Our study resulted in discovery of a DNA gene hypermethylation pattern characteristic of FL. Gene ontology (GO) biologic and molecular function categorization of these target genes indicates significant involvement of developmental, cell-adhesion, and morphogenesis pathways, among others. Additional analyses revealed genes characteristically undergoing hypomethylation in FL. Interestingly, gene hypomethylation in FL targets different cellular processes and growth-related pathways than gene hypermethylation; for example, stem cell markers such as CD34 and teratocarcinoma-derived growth factor 1 (TDGF1) were uniquely hypomethylated in FL. Thus, DNA methylation profiling provides molecular diagnostic markers as well as insight into the biologic pathway perturbations in lymphoma
Showing the most recent 10 out of 90 publications
