Lineage-specific transcription factors in unexpected places

Abstract

Foxp3 is a transcription factor closely associated with the regulatory T cell (Treg) lineage in humans and mice. In the immune system, Foxp3 appears highly specific for Treg, and is not known to be expressed by other immune cell types. In this issue of the European Journal of Immunology, a paper reports that human dendritic cells (DC) transfected with ectopic Foxp3 unexpectedly acquire an immunosuppressive phenotype. Foxp3-transfected DC suppressed proliferation of naïve T cells, and biased the differentiation of CD4⁺ cells into Treg-like cells that themselves expressed Foxp3. The molecular mechanism of these effects required functional activity of the immunoregulatory enzyme indoleamine 2,3-dioxygenase (IDO). Thus, a transcription factor not native to DC nevertheless conferred elements of a regulatory phenotype following ectopic expression.

Foxp3 would seem to be a classic example of a lineage-specific transcription factor in the immune system (reviewed in ref. [1]). Foxp3 is expressed stably and at high levels by a small subset of CD4⁺ T cells with regulatory/suppressor activity [2]. Transfection of Foxp3 cDNA into naïve CD4⁺ T cells promotes differentiation into Treg-like cells [3]; whereas genetic deficiency or induced ablation of Foxp3 leads to loss of regulatory activity and autoimmunity [4]. While Foxp3 is not the only molecular determinant of the Treg lineage (and there are other, Foxp3-independent pathways that can create T cells with functional regulatory activity [5]), nevertheless Foxp3 is mechanistically important and non-dispensable transcription factor required for the differentiation of a major population of regulatory T cells (Treg).

In this issue of the European Journal of Immunology, Lipscomb, et al. [6], examines whether ectopic over-expression of Foxp3 in dendritic cells (DC) – a cell type that does not normally express Foxp3 – would alter their functional properties. As far as is known, Foxp3 has no role in DC biology; however, other reports have recently made the rather surprising observation that certain tumor cells express Foxp3, and that this expression appeared to render the tumor cells functionally immunosuppressive [7] (although the molecular mechanism of these findings remains unknown). Reasoning by analogy, Lipscomb and colleagues [6] asks whether ectopic expression of Foxp3 in DC might similarly render DC immunosuppressive.

The authors [6] found that naïve T cells exposed to Foxp3-transfected DC showed marked inhibition of proliferation, lost the ability to differentiate into pro-inflammatory Th1 cells, and upregulated the GATA-3 and Foxp3 transcription factors (associated with Th2 and Treg differentiation, respectively). Consistent with their observed upregulation of Foxp3 mRNA and protein, naïve T cells exposed to Foxp3-transfected DC displayed measurable functional suppressor activity, suggestive of an acquired Treg-like phenotype. Thus, DC that were forced to express ectopic Foxp3 became quite different to their normal counterparts, and were much more immunosuppressive – including eliciting de novo Foxp3 expression in the T cells they suppressed [6].

This effect of ectopic Foxp3 expression on DC was unexpected, and its molecular mechanism remains to be fully elucidated. However, the fact that Foxp3 had a biologic effect even in an ectopic location may represent an example of a broader theme that is emerging from studies of lineage-related transcription factors in the immune system. In the following discussion we will review recent findings in two examples, STAT3 and T-bet, which suggest that certain transcription factors appear to play a role in more than one lineage, yet in each case the effect is to bias the response toward a particular overall immunologic outcome (e.g. tolerance, Th1-type inflammation, etc.).

One example of this phenomenon is the transcription factor STAT3. STAT3 can be expressed by a variety of tumor cells, and one of the effects of STAT3 is to alter the immunologic attributes of the tumor: reducing proinflammatory cytokine production, and driving the creation of an immunosuppressive local milieu [8]. In the cells of the immune response elicited by STAT3-expressing tumors, STAT3 is also expressed by macrophages and DC, where STAT3 appears to bias these cells toward immunosuppression and tolerance [9–10]. Thus, STAT3 plays a functional role in both the tumor cells and in the immune cells that the tumor recruits. Even though these are very different cell types, the effect of STAT3 expression in both cases seems to favor the same overall outcome (immunosuppression and tolerance). Further reinforcing this concept, recent data suggest that STAT3 is also expressed in Tregs. In Tregs, STAT3 helps to maintain their immunosuppressive and regulatory properties [11]. Of course, STAT3 also does many other things as well, and tolerance induction is not an inherent function of STAT3. Rather, it is the particular cell type and the particular context that confer the particular immunological effect of STAT3. The surprise, however, is that in cells as different as tumor cells, macrophages, DC and Treg, a single transcription factor would be biologically active, and would contribute to the same overall immunologic outcome.

An analogous example occurs in the case of the transcription factor T-bet. In CD4⁺ helper T cells, T-bet promotes differentiation along the Th1 lineage [12]. In CD8⁺ T cells, T-bet expression drives differentiation into cytotoxic effector T cells (CTL) [13] – the same differentiation pathway that is promoted by help from T-bet-expressing CD4⁺ Th1 cells. Recently, T-bet has also been shown to be active in a subset of Treg [14]. Among other effects, this T-bet expression helps direct Treg to sites of Type 1 (Th1/CTL) inflammation, by promoting expression of the chemokine receptor CXCR3 (which also facilitates homing of CD8⁺ CTL to the same locales [15]). Thus, even in “opposing” lineages (effectors and suppressors) the effects of T-bet appear conceptually related, bringing different lineages together in a coordinated fashion at the same physical location.

What can we conclude from these examples? They would seem to argue against a simple one-to-one model of unique, lineage-specific transcription factors – i.e., in which Foxp3 controls only the Treg lineage, T-bet controls the Th1 lineage, and so forth. The cited examples hint at a deeper level of coordination, in which certain transcription factors seem associated with an overall functional outcome (e.g. tolerance, Th1-type inflammation, immunosuppression, etc). Under this hypothesis, when several different cell types must collaborate to achieve a complex immunologic outcome, they might all express the same transcription factor, with each cell type responding in its own, cell-specific fashion. How, exactly, such a system would evolve remains something of a puzzle – how, for example, would STAT3 expression in a tumor cell come to elicit an immune response that is itself governed by STAT3? But it is known that the signals and pathways in the immune system are finite, and all the participants at a particular site of immune response tend to encounter a similar set of local microenvironmental signals. Perhaps these stereotyped signaling pathways – and hence their downstream transcription factors – have evolved to elicit the various intracellular responses in each cell type that are appropriate to its specific role in the overall local immune reaction.

This hypothesis is speculative; and even if it is valid, it still leaves a mystery in the current study of Lipscomb et al. [6], because Foxp3 is not thought to play any normal role in DC. Why, then would ectopic expression of Foxp3 elicit any phenotypic program in DC at all (much less a teleologically “correct” immunosuppressive program)? One theoretical possibility might be that Foxp3 actually plays some previously unappreciated role in DC biology; or, alternatively (as seems more likely) it may be that DC are not so far removed developmentally from the T cell lineage that they have lost all ability to respond to T cell-specific transcription factors [16]. Thus, while Foxp3 may not normally be expressed by DC, it may still be able to activate certain downstream pathways that have meaning for DC biology.

Finally, Lipscomb et al. [6] show that IDO, an immunoregulatory enzyme that catabolizes tryptophan, appeared to participate as a downstream effector mechanism in the immunosuppressive phenotype elicited by Foxp3. IDO is an inducible gene in both myeloid and plasmacytoid DC, and can be upregulated by a variety of stimuli (reviewed in ref. [17]). Consistent with the current report, IDO expression by DC has been shown to bias CD4⁺ T cell differentiation toward a Foxp3⁺ regulatory phenotype [18–19]. The mechanism by which the transfected Foxp3 gene caused IDO induction is currently unclear. (Empty adenovirus vector itself did not induce IDO, which was an important control because IDO can be nonspecifically induced by viral infection [20]). Whatever the mechanism of IDO induction by Foxp3, once induced, IDO appeared to form an important downstream component of the immunosuppressive DC phenotype.

In conclusion, ectopic Foxp3 expression in DC creates unexpected alterations in the biology of human monocyte-derived DC. The molecular basis of this effect, and the downstream pathways that are the target of Foxp3, remain to be elucidated. But the observation itself is striking: that a transcription factor associated with tolerance should create a “tolerogenic” phenotype in a cell type that would normally (as far as we know) never express Foxp3. On a practical level, the ability to transfect human DCs with a single gene that confers an immunosuppressive phenotype may be useful as a strategy for creating tolerogenic DCs, which could be a valuable clinical preparation if successful. At a more fundamental level, the effect of ectopic Foxp3 expression in DCs may serve as a reminder that lineage-associated transcription factors in the immune system – even those that appear quite lineage-specific – may also play a biologic role in other cell lineages as well. And, at least in some examples, the very different downstream effects exerted by the same transcription factor in different cell types may serve to coordinate complex immune responses toward the same overall outcome.