A doppelgänger of T cell development

Innate lymphoid cells (ILCs) were recently identified as the innate counterpart of adaptive T-cells, after the discovery of several ILC subsets with effector functions strikingly similar to T-cell subsets. Although our understanding of ILCs is still preliminary, it is becoming clear that transcriptional programs controlling ILC terminal differentiation and effector functions closely mirror those of conventional αβ T-cells (Fig. 1, reviewed in ref. 1). Understanding how ILCs develop and diversify was initially impeded by the lack of defined intermediate progenitors, preventing the distinction between initial ILC specification, terminal differentiation and function. However, the recent identification of a common helper innate lymphoid progenitor (CHILP),^2,3 and the discovery of an earlier innate lymphoid progenitor (EILP) to both cytotoxic and helper ILCs⁴ has been a significant advance in this process. Characterization of these intermediate progenitors demonstrates that, like T-cells, ILCs develop in a stepwise manner, and suggest an appealing picture in which each step of ILC development would be a doppelgänger of a T-cell development stage (Fig. 1).

Figure 1.

Comparison between adult ILC and conventional αβ T-cell development. ILC and T-cell development is represented from a common lymphoid progenitor into terminally differentiated effectors. Transcription factors known or thought to mediate transition from one stage to the next are indicated.EILP: early innate lymphoid progenitor, NKP: natural killer progenitor, cNK: conventional natural killer, CHILP: common helper innate lymphoid progenitor, ILC: innate lymphoid cell, DP: CD4⁺CD8⁺ double positive, Tc: T cytotoxic, Th: T helper.

Supporting this parallel, transcription factors such as RUNX molecules, GATA3 and TCF-1, that are known to play critical functions at the CD4⁺CD8⁺ double positive (DP) stage of thymocyte development, have recently been found to be required in developing ILCs. Importantly, these transcription factors seem to control similar transitions in ILC and T-cell development (Fig. 1). It is unknown whether these factors have similar functions in both cell lineages, but if so, one could try and translate functions for these players from one lineage to another. For example, TCF-1 (and to a lesser extent LEF-1) was recently described to play a critical role in the development of helper versus cytotoxic lineages from DP thymocytes. This function is achieved by controlling Thpok induction in developing CD4⁺ T-cells, and repressing CD4⁺ T-cell lineage genes in collaboration with RUNX in CD8⁺ T-cells.⁵ In ILCs, TCF-1 is known to be required for the development of both helper and killer ILC progenitors.⁴ However, the developmental stage at which it is required and its function remain unknown. A tempting speculation is that TCF-1 is required at or after the EILP stage, to drive the expression of helper factors in developing helper ILCs, and works together with RUNX to silence helper genes in developing natural killer cells (NK), mirroring its function in thymocyte development. In contrast to the many similarities between conventional CD4⁺ T-cells and helper ILCs development, one apparent disparity is their dependency on distinct Zbtb family transcription factors, namely Thpok (or Zbtb7b) in T-cells, vs. PLZF (or Zbtb16) in ILCs. It is not known whether Thpok and PLZF have similar functions. Interestingly, only a mild ILC defect is observed in the PLZF knockout mouse, suggesting possible redundancy between members of the Zbtb family in helper ILC development.³

As ILC and T-cell lineages have such similar effector functions and are apparently controlled by a similar transcriptional network beyond EILP and DP stages, understanding how these cell lineages diverge into innate and adaptive arms of the immune system remains a puzzle. Early T-cell development requires many unique players that control the somatic recombination, expression, and selection of genes encoding the αβ TCR complex; these requirements lack parallels during early ILC development. Additionally, transitions from DP to mature naïve T-cells, and further to effector T-cells, are under the control of the αβ TCR. This dependency is linked to the quiescent state of T-cells in the absence of TCR-signaling and might be imposed by a transcriptional program in adaptive T-cells that remains to be better characterized.⁶ On the other hand, the earliest described ILC progenitors do not have detectable T-cell potential.⁴ This suggests that mechanisms of repression of the T-cell lineage are part of the early ILC transcriptional program. Factors controlling the generation of EILP and commitment to the ILC lineage have not yet been established. Candidates such as NFIL3, TOX, or Id2 might fulfill these functions since they are expressed highly in EILP,⁴ and are required for ILC but not early T-cell development. Such factors could play unique functions in ILC development, and control the divergence of innate versus adaptive lineages. Alternatively, as most transcription factors are used in both innate and adaptive lineages, unique combinations of factors could result in innate vs. adaptive lineage choice. Uncovering the factors and mechanisms that specify innate versus adaptive fates during early development is a fascinating and challenging new issue. The ongoing identification of early intermediate stages between lymphoid progenitors and ILCs should eventually allow these questions to be addressed.