Integrating T Cell Activation Signals to Regulate Gene Expression through Cyclosporin-Sensitive NFAT

The discovery and subsequent development of the fungal-derived cyclosporin A (CsA) in the 1970s by Jean Borel and colleagues (1) revolutionized the field of solid organ transplantation, demonstrating the power of immune system modulation in a clinical setting. Unlike other recent targeted drug development approaches, CsA was pursued because of its biological activity as a potent and specific immunosuppressive, and in particular a T cell activation inhibitor, before its mechanism of action was understood. By the early 1990s, however, it was clear that CsA bound to an intracellular receptor (immunophilin), and this drug complex could inhibit calcineurin, a Ca²⁺- and calmodulin-dependent protein phosphatase, thus disrupting a major signaling pathway initiated by TCR recognition of cognate Ag (2). The clinical import of CsA and the structurally distinct microbial product FK506 (3) developed in the following decade, along with the utility of these drugs as tools to dissect T cell activation, led to a quest to understand how these drugs, and their downstream inhibitory target calcineurin, regulated T cell activation at the level of gene expression.

Because CsA and FK506 were known to be inhibitors of early T cell effector functions (4–6), research attention turned to mechanistically integrating the affected calcium-mediated signaling pathway and downstream early gene regulation. The Il2 gene was an obvious target for this, because significant IL-2 production is T cell specific, cell activation dependent, expressed early after activation, and CsA inhibitable. Matching the characteristics of IL-2 production from activated T cells, an activation-dependent and tissue-specific enhancer of the Il2 gene could be shown to bind a protein found in the nucleus of activated T cells, appropriately termed NFAT (7). It was also clear that NFAT did not act alone, because occupancy of multiple sites on the enhancer was required for full gene activation. NFAT itself was composed of both pre-existing and inducible components, the former of which Rao and colleagues (8) termed NFATp (now recognized as NFAT1). Importantly, CsA was found to target NFAT activity (2, 9), directly linking the biological effects of this important immunosuppressive to specific transcriptional regulation.

In the first two papers of particular focus in this commentary, Rao and colleagues (8, 10) made significant leaps forward in understanding how NFAT regulated Il2 gene expression in coordination with other factors, and how CsA could influence this activity. It is notable that these characterizations preceded the cloning of the NFAT gene, unheard of in modern times, and largely relied on the ability to detect protein–DNA complexes by changes in mobility of radioactively labeled oligonucleotides (“gel shift” assays). Taking advantage of a CD4⁺ murine T cell clone that could be activated by Ab-mediated cross-linking of CD3ε, two activation-dependent nuclear protein complexes were detected bound to an oligonucleotide containing an NFAT site derived from the Il2 gene. Using combinations of mutagenesis and competition with unlabeled oligonucleotides, more detailed mapping revealed that the slower migrating protein–DNA complex could be inhibited with sequences containing an AP-1 site.

AP-1 is not a single factor but a family of dimeric transcriptional regulators that are widely expressed and involved in many diverse cellular functions, from cell division to migration to apoptosis (11). In this instance, specific Abs identified AP-1 factors Fos and Jun in the slower-migrating complex. Although Fos and Jun proteins are upregulated on T cell activation, they are not CsA sensitive. Of particular note, when hypotonic cell lysates were used instead of nuclear extracts (the former also containing cytosolic proteins), the faster-migrating protein–DNA complex was detected even in unactivated cells. This provided additional support for the finding of pre-existing NFAT in resting cells, and that one aspect of NFAT regulation involved a change in subcellular localization, specifically nuclear import (12). In an elegant series of complementation experiments, the authors then showed that the-slower migrating complex would form on mixing a nuclear extract from activated cells depleted of Fos (containing NFAT, but not AP-1) with one from cells activated in the presence of CsA (deficient in nuclear NFAT but containing AP-1).

The follow-up study aimed to explore in more depth the characteristics of NFATp, the factor detected in unstimulated cells. Using a combination of heparin-agarose (a mimic of polyanionic DNA) and specific DNA-affinity chromatography, the investigators were able to purify an ∼120-kDa putative NFAT protein from a transformed murine T cell clone. This protein isolate was a phosphoprotein susceptible to treatment with calcineurin, resulting in a phosphatase-dependent change in gel mobility. Importantly, both the protein isolate with or without calcineurin treatment would bind an NFAT sequence from the Il2 gene and would form a slower-migrating cocomplex on the DNA with recombinant Fos and Jun. This again supported a model where calcineurin did not regulate NFAT DNA binding or association with other factors per se, but instead controlled nuclear entry. Notably, Fos and Jun were unable to bind the NFAT oligonucleotide in the absence of NFAT protein, nor would this same NFAT oligonucleotide compete with Fos-Jun binding to other AP-1 sites. A series of protein mutagenesis experiments indicated that DNA binding by Fos-Jun heterodimers (or Jun homodimers) was nevertheless required to form the NFAT cocomplex. These results could be explained by the presence of an adjacent AP-1–like sequence on the NFAT oligonucleotide that, when mutated, prevented Fos-Jun binding. Together, these data suggested that the juxtaposition of an NFAT binding site and an adjacent AP-1–like site on the Il2 gene DNA was key to formation of this multiprotein transcriptional regulator complex.

The ability to at least partially isolate the NFAT protein led Hogan and colleagues (13) to clone the first NFAT gene in the third paper highlighted in this commentary. Based on the protein sequence of two tryptic peptides derived from the 120-kDa protein isolate, degenerate oligonucleotides were designed and used to PCR-amplify a gene fragment from T cell–derived cDNA. This gene fragment was then used as a probe to screen a cDNA library. As was predicted for NFAT, the identified gene was expressed in T cells, but not fibroblasts. But more direct evidence for the identity of the putative NFAT gene was a series of experiments using a 464-amino acid recombinant protein made from the isolated gene sequence. This protein bound an Il2 gene NFAT sequence, associated with Jun homodimers or Fos-Jun heterodimers in a protein–DNA complex, was required for AP-1 binding to the NFAT-containing sequence, and activated transcription in combination with AP-1.

One of the conundrums of gene regulation by a ubiquitous set of factors such as AP-1 is how, despite the lack of a cell-specific expression pattern, these proteins can mediate specific gene responses in response to environmental cues. Although the immune cell context is hardly unique in this regard, the highlighted work did reveal one part of the answer, and that is cooperativity of AP-1 with other transcriptional regulators such as NFAT, which have a more cell-specific expression pattern, can target transcriptional complexes to relevant sites and have a built-in sensor function, in this case mediated by cell activation-induced dephosphorylation controlling subcellular localization. Moreover, this transcriptional regulator complex allows integration of distinct signaling pathways, a Ca²⁺ calcineurin-dependent pathway for NFAT activation, and a MAPK-dependent pathway regulating expression and activation of AP-1.

Not surprisingly, much complexity has been revealed in the ensuing nearly three decades, including identification of four calcium-regulated NFAT proteins (and various isoforms) expressed in a wide variety of cells, and three of which (NFAT1, NFAT2, NFAT4) are expressed in T cells (14). Moreover, adjacent NFAT and AP-1 motifs allowing cooperative binding between factors with extensive protein–protein interactions (15) are not unique to the Il2 gene but are shared among other members of a transcriptional module in the context of T cell activation (16). The list of functions mediated by NFAT proteins has also grown beyond T cell activation, fittingly including recent work from Rao and colleagues (17) finding that genes encoding transcriptional regulators of the TOX-family that induce CD8 T cell exhaustion are targeted by NFAT, interestingly, even in the absence of Fos and Jun. Unquestionably, these earlier studies provided significant keys to understanding the molecular basis for the convergence of multiple cell signaling pathways in the regulation of early gene expression during T cell activation.