The Old One–Two: Solving the Mystery of Cognate Help

The pivotal findings reported by Noelle et al. (1) in 1992 resolved a puzzle that had been 40 years in the making. In the mid-20th century, seminal studies from Owen (2), Medawar and colleagues (3, 4) and Burnet (5) established the concept of an immunological self, forcing re-evaluation of the paradigms driving the field. Indeed, the instructional models that had been forwarded to explain specificity and affinity maturation in humoral responses, already under duress as the rules of molecular biology and protein chemistry advanced, lacked the means to accommodate a mechanism for self versus nonself discrimination, producing an intractable conundrum. The key, of course, was to address this problem at a cellular level. Although Ehrlich had enunciated a cell-centric approach to explaining immune responses half a century earlier (discussed in Ref. 6), his side-chain hypothesis had long been abandoned, and it also lacked a tractable means of dealing with this new but fundamental concept. Within this context, Jerne (7), Talmage (8), and Burnet (9, 10) collectively advanced a paradigm, now termed the clonal selection hypothesis, that would assume prominence during the next 50 years. By positing a clonal distribution of Ag receptor specificities, this model provided a means through which immune repertoires could be purged of unwanted specificities in a quantized manner, thus eliminating self-reactive clones but sparing beneficial specificities. By the mid-1980s, most of the potential stumbling blocks for the clonal selection hypothesis had been cleared: the concept of clonally distributed receptors had been definitively established (11) (reviewed in Ref. 12), and the genetic impasse posed by the need for vast receptor diversity had been overcome once the process of somatic recombination involving multiple gene segments had been revealed (13–15).

Despite these advances, an ironic caveat to the clonal selection hypothesis nonetheless endured: the Ag receptor was required to play both an immunogenic and tolerogenic role, prompting the question of how signals through the same receptor could drive diametrically opposite outcomes. This apparent logical contradiction was recognized early on, and various models were forwarded to deal with it. Among these, the two signal hypothesis advanced by Bretscher and Cohn (16) has proven durable. According to this model, a signal through the Ag receptor, termed signal 1, culminated in cell death unless superseded by a second, Ag receptor–independent, signal. These overarching ideas laid the conceptual groundwork for the studies that were reported in this issue’s Pillars of Immunology article.

Observations during the two decades preceding the highlighted article painted an increasingly complex picture of how such second signals were delivered. For example, classic hapten-conjugate immunization strategies showed that there were multiple ways in which signal 2 could be delivered to B cells. Thus, haptens coupled to mitogens such as bacterial LPS could directly activate hapten-specific B cells. In contrast, when haptens were coupled to protein carriers, signal 2 required and emanated from Th cells. Moreover, this T-dependent form of signal 2 was most effective when the epitope recognized by the B cell (the hapten) was physically attached to the protein portion of the Ags recognized by the T cell (the carrier). This so-called “linked recognition” was further refined with the discovery of MHC restriction of Ag recognition by T cells, yielding the realization that it reflected a cognate event: the presentation of processed carrier peptides to the T cell in the context of MHC class II molecules on the B cell. However, in this and other studies it was evident that the TCR–MHC class II interactions per se did not constitute the elusive signal 2 delivered to the B cell by activated Th cells. Within this milieu of increasing conceptual complexity, Noelle and colleagues (17–21) had already invested considerable effort in generating the tools and insights necessary to dissect the requirements and molecular basis for these interactions. These, together with a combination of cellular, biochemical, and molecular approaches, converged to reveal the basis for signal 2 in T-dependent B cell responses, thus resolving the molecular basis of cognate help.

Noelle et al. (17) had already shown that membrane preparations from activated CD4⁺ T cells could promote activation of B cells, as indicated by an uptick in RNA synthesis measured by the incorporation of radiolabeled uridine. Importantly, this was only true for activated T cell membrane preparations, suggesting that T cell activation itself was a requisite step toward acquiring the capacity for signal 2 delivery. Additionally, prior studies from others had suggested that Abs to CD40 had stimulatory effects on B cells, implicating this molecule as a candidate signal 2 target. Armed with this knowledge, and coupled with the emerging technologies that allowed generation of fusion proteins and mAbs, their search ensued for a molecule on these membranes that might be responsible. They reasoned that if CD40 was indeed the target of the elusive signal 2, then the matching ligand on activated T cell membranes should be both blocked by and detected with a fusion protein that incorporated CD40 on one end and an Ig constant domain on the other end.

Accordingly, using the RNA synthesis assay as a surrogate for receipt of signal 2 from activated T cell membranes, several mAbs and fusion proteins—each directed against various candidate T cell surface molecules—were assessed for their ability to block the interactions between activated T cell membranes and B cells that were responsible. Abs specific for several adhesion and coreceptor molecules, including LFA, ICAM, and CD4, had no apparent effect. This result tended to rule out a role for these molecules, arguing against adhesive interactions per se or some form of MHC class II reverse signaling, although these had been forwarded as signal 2 candidates in the existing literature. These also served as controls, demonstrating that simply adding Abs to surface molecules present on the activated T cell membranes would not block the transcription-promoting effects. Because Abs to a litany of known molecules could not block the activation of B cells by T cell membranes, mAbs were raised that recognized molecules on activated, but not resting, T cells. The Abs that fit these initial criteria were screened for their ability to block the activation of B cells by activated T cells. MR1, the only Ab with this characteristic to be found in multiple fusions, blocked T cell–dependent B cell activation in vitro. In parallel, the functional activities of CD40-Ig proved that it, similar to MR1, could reduce transcriptional activity to background levels in a dose-dependent manner. Importantly, the CD40-Ig fusion protein, as well as MR1, did not interfere with RNA synthesis increases associated with LPS-driven B cell activation, giving confidence that this inhibition was interfering with signals delivered by activated T cell membranes, but not other classes of molecules capable of providing a second signal. Moreover, CD40-Ig or MR1 had to be added within the first 24 h of coculture; otherwise, the effect was lost, suggesting that once the signal had been delivered, the cascade of downstream events had already been triggered. Finally, in keeping with the observation that only activated T cell membranes could engender these downstream activation measures, a fluorescently labeled CD40-Ig construct and MR1 detected their target on activated, but not resting, T cells. Importantly, CD40-Ig blocked the staining of activated T cells by MR1 and vice versa, providing a critical piece of evidence that MR1 was recognizing a ligand for CD40.

Confident that the CD40-Ig fusion protein was indeed interacting with a T cell–derived, CD40-interacting ligand, the team turned to immunoprecipitation to identify and isolate the molecule responsible. The results were clear: the fusion protein as well as MR1 precipitated a 39-kDa band from activated CD4⁺ T cell lysates. With that, the ligand for CD40, now termed CD154, was discovered and characterized.

The findings reported in this seminal paper not only provided a molecular mechanism of linked recognition and cognate T cell help, but they also set the stage for the subsequent 25 y of basic and translational research focused on these molecules and their interaction. These areas of research include the now well-established roles of CD40 and CD154 in germinal center initiation and selection that are crucial for affinity maturation and effective humoral immunity (22; reviewed in Refs. 23, 24), as well as more recently appreciated roles both within and outside of the immune system (25, 26).