Toward a layered immune system

Leonore A. Hetzenberg and Leonard A. Herzenberg Department of Genetics Stanford University School of Medicine Stanford, California 94305 The broad outlines of developmental relationships among cells in the immune system appeared to be established some years ago with the introduction of the idea that pluripotent hematopoietic stem cells emerge early in fetal life and persist as a self-replenishing population that con- tinues to divide and differentiate throughout life without changing its original potential. Recent studies of lympho- cyte development challenge this paradigm (Herzenberg and Stall, 1989; Linton et al., 1989), suggesting that there are instead several types of hematopoietic stem cells that have evolved sequentially and function at specified times during development; these create layers of progressively more advanced populations of lymphocytes and myeloid cells that collectively provide the diverse capabilities of the mammalian immune system. According to previous dogma, the self-replenishing he- matopoietic stem cells that reside initially in fetal liver and then in adult spleen and bone marrow provide a continu- ing source of myeloid and lymphoid stem cells. These my- eloid and lymphoid stem cell populations are themselves self-replenishing and collectively give rise to all erythro- cytes, myelocytes (granulocytes, macrophages, etc.), and lymphocytes found in the fetus, neonate, and adult. Thus, cells in the hematopoietic system are viewed as a single lineage that is derived from a single progenitor popula- tion, and lymphocytes are one of two major subdivisions of that lineage. The origin of various types of 6 and T lymphocytes can be incorporated into an extension of this model that in- troduces further branchpoints in the overall lineage-one that separates developmental pathways for B cells and T cells, and others that distinguish pathways for various subsets of T and B cells. Some workers refer to each of these pathways as a separate lineage; others reserve the term lineage for those pathways that appear to originate from a distinct progenitor with (at least a limited) capacity for self-regeneration and the ability to give rise to the later cells in the pathway. However, all agree that lymphocyte differentiation proceeds along pathways that at some level merit the designation lineages, because lymphocytes ori- ginate with distinct progenitors that divide and differenti- ate to populate sequential stages of the pathway. The B cell lineages proposed by Klinman and col- leagues (Linton et al., 1989) fall within this general frame- work. The existence of these lineages is predicated on evi- dence indicating that, contrary to conventional textbook wisdom, virgin B cells participating in primary antibody re- sponses are developmentally distinct from memory B cells that participate in secondary antibody responses to the same antigens. Some questions can be raised con- cerning the generality of these findings and their interpre-