METHODS IN LABORATORY INVESTIGATION - A SELECTIVE CULTURE SYSTEM FOR GENERATING TERMINAL DEOXYNUCLEOTIDYL TRANSFERASE-POSITIVE LYMPHOID-CELLS IN-VITRO .3. STRUCTURE OF THE BONE-MARROW MICROENVIRONMENT FOR EARLY LYMPHOPOIESIS

BACKGROUND: We have previously demonstrated the feasibility of generat ing terminal deoxynucleotidyl transferase-positive (TdT+) lymphoid pre cursor cells in vitro in the nonadherent compartment of a long-term xe nogeneic culture system in which rat bone marrow (BM) cells are seeded onto established mouse BM adherent cell layers. We have also noted th at the appearance of TdT+ cells in these cultures is preceded by the f ormation of clusters of lymphoblasts in close association with the mou se BM adherent cell layer. Inasmuch as the selective generation of suc h primitive lymphoid cells is not ordinarily observed in homogeneic (i .e., mouse: mouse, rat: rat) BM cultures, the nature of the microenvir onment for the generation of committed lymphoid stem/pregenitor cells has not yet been detailed. Consequently, the aim of this study was to define the cellular components in the adherent compartment of our xeno geneic culture system that are associated with the earliest stages of lymphopoiesis in vitro. EXPERIMENTAL DESIGN: The nature of the interac tions between rat BM lymphoid precursor cells and mouse BM adherent mi croenvironmental cells was investigated by a combination of immunophen otyping and scanning and transmission electron microscopy of primary c ultures. The kinetics of formation and composition of lymphoid cluster s were also determined morphologically and phenotypically. Results wer e compared with those of other investigators who have studied lymphopo iesis in intact BM or in homogeneic cultures of pre-B cells. RESULTS: Two distinct microenvironmental regions are represented within the mou se BM adherent cell layer: (a) paucilayer (PL) regions, composed of tw o or three horizontally oriented layers of alkaline phosphatase-positi ve mouse stromal cells; and (b) multilayer (ML) regions, containing 4 to 8 layers of such stromal cells. In both regions, proliferating rat lymphoid cells, expressing the HIS24 (B220) and/or HIS50 (heat stable antigen) early B-lineage antigens, are ''sandwiched'' between adjacent layers of stromal cells and enveloped by cytoplasmic processes from i nterdigitating mouse macrophages (pseudoemperipolesis). More than 95% of the lymphoid cells are of rat origin, whereas more than 95% of the nonlymphoid cells are of mouse origin. Large clusters, containing up t o 1,000 lymphoid cells, preferentially develop in the ML regions and a re comprised primarily of TdT+ cells. Small clusters containing 5 to 5 0 lymphoid cells, preferentially develop in the PL regions and are com prised primarily of TdT- cells, that can generate TdT+ cells upon tran sfer onto fresh adherent cells layers. Formation of individual small c lusters, which outnumber large clusters by approximately 10-fold, is i nitiated by as few as 25 unfractionated rat BM cells. This process is not preceded by a lag period, and is linear with respect to time and c ell dose. Formation of large clusters requires approximately 30 times more input cells, and is linear with respect to time after a lag of 5 days. CONCLUSIONS: The number of small lymphoid clusters formed in vit ro closely approximates the frequency of lymphoid stem/progenitor cell s in the BM inoculum (3 to 5%). This suggests that, under ideal condit ions, individual clusters are clonally derived and the seeing efficien cy of the culture system approaches 100%. The results further suggest that large clusters are formed by the coalescence of numerous small cl usters within ML regions of the adherent cell layer; and that the form ation of ML regions may be supported by an underlying monolayer of mac rophages. A novel aspect of this system appears to be the frequency of pseudoemperipolesis, rather than phagocytosis, of primitive lymphoid cells by macrophages, that has also been noted in vivo. Moreover, the ML regions themselves bear a close resemblance to the recently describ ed pro-B cell-enriched, multicellular aggregate fraction of freshly ha rvested mouse BM. Hence, this system appears to structurally recreate in vitro the in vivo microenvironment for the development of pro-B cel ls. Moreover, the close physical associations between the developing l ymphoid cells, stromal cells and macrophages in tricellular complexes point to important regulatory roles for the latter cell types in early lymphopoiesis. Given the absence of differentiation beyond the TdT+ p ro-B cell stage, our culture system would appear to provide a selectiv e in vitro model for study of the earliest stages of lymphopoiesis and leukemogenesis in rat, mouse and, potentially, human BM.