PU.1, a Shared Transcriptional Regulator of Innate and Adaptive Immune Cell Fates

In the late 1980s many molecular biology laboratories were in search of tissue-specific transcription factors that regulated cell type-specific patterns of gene activity. The hope was that these key transcriptional regulatory proteins would provide molecular insight into the development and functioning of various differentiated cell types. Immunobiologists were at the forefront in this exciting endeavor, and the cloning of the cDNA for the transcription factor PU.1 was particularly notable in this historical context. At that point in time, only a handful of mammalian, cell type-specific transcription factors had been biochemically characterized and cloned as cDNAs. They included Oct-2, a lymphoid-specific transcription factor that bound the conserved octamer element in Ig gene promoters, and Pit-1, a pituitary-specific transcriptional regulator (1, 2, 3). The isolation of the cDNA for PU.1 by the Maki laboratory was particularly intriguing in that the transcription factor was found to be expressed specifically in macrophages as well as B cells and was therefore implicated in the development and/or functioning of both innate as well as adaptive immune cell types (4). Furthermore, the deduced amino acid sequence of PU.1 revealed it to be a member of the ets gene family. The founding member of this family, v-ets, encoded by the avian leukemia virus E26, had been known to cause erythroblastic and myeloblastic leukemias in chickens (5, 6). It turned out, based on a complementary research initiative of F. Moreau-Gachelin, that PU.1 was encoded by the Spi-1 gene (7). Deregulated expression of the Spi-1 gene caused by proviral insertions of the spleen focus-forming virus led to erythroleukemias in mice. Research spurred by the discovery of PU.1 would go on to establish its pivotal role in the generation of innate and adaptive immune cells. Importantly, mutations in the PU.1 gene could also be shown to cause acute myeloid leukemia in mice (8).

In a curious way, my research path initially contributed to the discovery of PU.1 and then was profoundly influenced by it. As a postdoctoral fellow with Phillip Sharp at the Massachussetts Institute of Technology, I was collaborating with members of David Baltimore’s laboratory to discover and clone transcription factors that regulated the expression of Ig genes in B lymphocytes. During this period (1985–1988) I first extended the use of the electrophoretic mobility shift DNA binding assay, EMSA, to identify and characterize nuclear transcription factors (9). I then developed a novel method for cloning genes encoding such regulatory factors by screening lambda phage cDNA expression libraries with their DNA binding sites as probes (10). Using EMSAs, the Baltimore and Sharp laboratories were able to rapidly discover the transcription factors NF-κB, Oct-2, and E2A (11, 12). Subsequently, screening of lambda phage expression libraries led to the isolation of the Oct-2 and E2A cDNAs (3, 13).

The cloning and characterization of PU.1 by R. Maki’s laboratory was stimulated by earlier work from W. Schaffner’s group, which was focused on the analysis of the prototypical transcriptional enhancer encoded by the simian virus SV40 genome. The Schaffner laboratory had managed to isolate an SV40 variant that could replicate in lymphoid cell lines (14). Molecular analysis of the variant revealed a novel transcriptional enhancer generated by duplication of a purine-rich motif (GAGGAA) termed the PU box. Using EMSAs, the Schaffner laboratory characterized a nuclear factor present in extracts of B lymphoid cell lines that bound the PU motif. Binding of this factor to the enhancer correlated with the growth properties of the SV40 variant.

In the context of discovering nuclear factors regulating the transcription of the MHC class II I-Aβ gene, the Maki laboratory used the expression screening strategy that I had developed (10). One of the MHC class II I-Aβ promoter sequences used in the screen resulted in the isolation of a cDNA encoding a protein that specifically recognized the same PU motif that had been previously described by the Schaffner laboratory (4). The gene encoding this protein was expressed in both B and macrophage cell lines and in the mouse spleen. Importantly, the protein could be shown to function as a sequence-specific transcriptional activator and its DNA-binding domain was predicted to contain a helix-turn-helix motif. This latter prediction was satisfied when R. Maki’s group, collaborating with that of K. Ely, determined the crystal structure of the binding domain in a complex with the purine-rich DNA motif (15). The structure revealed a novel architecture involving a winged helix-turn-helix motif.

The article by R. Maki and his colleagues describing PU.1 concluded with the following sentence: “The fact that it is a cell type-specific transcriptional activator suggests that it may play an important role in the differentiation or activation of macrophages and B cells.” This indeed turned out to be the case. Subsequent research by my laboratory and those of R. Maki, D. Tenen, and T. Graf revealed PU.1 to be a pivotal regulator of both innate and adaptive immune cell fates and an antagonist of erythrocyte and megakaryocyte development (16, 17, 18).

My laboratory at the Howard Hughes Medical Institute at the University of Chicago became interested in PU.1 because it was implicated in regulating the activity of Ig light chain gene enhancers based on the initial findings of M. Atchison’s group (19, 20). Given that numerous genes appeared to be regulated by PU.1 in B lineage cells as well as macrophages, we pursued a genetic analysis of PU.1 function in mice by engineering a null mutation using homologous recombination in embryonic stem cells. This work, performed collaboratively with C. Simon, established that PU.1 was required for the development of both innate (macrophages and granulocytes) and adaptive (B and T lymphocytes) immune cells (21). The Maki laboratory confirmed our findings by engineering a distinct mutation in the Spi-1 gene (22). We went on to demonstrate that PU.1 functioned in a cell-intrinsic manner at the level of multipotential lymphoid-myeloid progenitors (23). These findings had considerable developmental significance, because they provided genetic evidence for a shared precursor of the innate and adaptive cells of the immune system (24). To explore molecular functions of PU.1 in the development of the lymphoid-myeloid system, we established powerful cell culture systems. PU.1^−/− hematopoietic progenitors could be transduced with retroviral vectors and then cultured under conditions that supported the development of macrophages, neutrophils, mast cells, or B lymphoid cells. These systems enabled us to complement the PU.1 mutation in distinct lineages as well as to selectively bypass the various molecular functions of PU.1 (25, 26). D. Tenen’s group had implicated PU.1 in regulating the transcription of genes encoding cytokine receptors including M-CSFR and GM-CSFR (27, 28). We were able to confirm that PU.1 not only activates the transcription of these cytokine receptor genes but also that of IL-7Rα (26, 29). Thus, PU.1 was shown to be required for the survival and proliferation of both myeloid and lymphoid progenitors.

An important issue raised by the aforementioned experimental analyses was how PU.1 functioned to specify distinct cell fates in the immune system. Was it simply a constitutive factor or did it play an instructive role in the specification of various cell fates? Experiments from my laboratory and those of T. Graf and D. Tenen provided strong support for the latter possibility (16, 18, 25, 30). These analyses suggested fundamental molecular principles involving graded levels of PU.1 as well as its cooperative or antagonistic interplay with the GATA family of transcription factors in the specification of distinct cell fates within the hematopoietic system. A low concentration of PU.1 protein was shown to induce the B cell fate, whereas a higher concentration promoted macrophage differentiation and blocked B cell development. These experiments provided the first evidence that physiologically graded levels of a transcription factor were used to specify distinct cell fates in the immune and hematopoietic system. Recent analyses from my laboratory have revealed PU.1 to be a pivotal component of distinct gene regulatory networks that orchestrate macrophage and B cell development, thereby helping to fulfill the molecular vision with which the article by R. Maki and his colleagues concluded (31, 32).