In a very different biological context, the TNF family ligand ectodysplasin A (EDA) participates in the development of skin appendages, such as teeth, hair, eccrine sweat glands and numerous other glands (Fig. 2). EDA-deficiency in human and various animal species such as mice and dogs causes one form of ectodermal dysplasia (X-linked hypohidrotic ectodermal dysplasia, XLHED). There is presently no cure for ectodermal dysplasia. This disease can be particularly dangerous for young, undiagnosed patients, decreases the quality of life, can have severe impacts on self-esteem, and generates important financial burden for palliative treatments. In the past, our group has demonstrated the feasibility to cure XLHED in very young mice and dogs by protein replacement therapy (with the collaboration of Dr M. Casal at the University of Pennsylvania for experiments performed in dogs) (Fig.3). Recently, we have given full support to the initiators of Edimer, a newly founded company whose goal is to launch clinical trials in XLHED patients by the end of 2011, and with whom we are associated from a research point of view.
Fig. 2: Involvement of the TNF family ligand EDA in the development of skin appendages. Membrane-bound EDA must be released as a soluble protein in order to activate its receptor, EDAR. At the site of EDAR activation, a NF-kB-dependent pathwayinduces neutralization of placode inhibitors, thus opening the way to the development of skin-derived structures such as hair, teeth or sweat glands. Hair and teeth that form despite the absence of EDA are often morphologically abnormal, suggesting that EDA is required for development and morphology.
BAFF and APRIL
TNF family ligands signal by engagement of one or several receptors of the TNF receptor family. Some TNF ligands, in addition to their cognate receptors, interact with unconventional partners such a proteoglycans. Most TNF family ligands assemble as homotrimers and display three receptor-binding sites. Recruitment of three receptors to the trimeric ligand is the first signaling step. In the past few years, several groups including our own have accumulated evidence that the form of ligand presented to responsive cells is very important. It is now generally considered that membrane-bound and soluble TNF family ligands, including TNF itself, can trigger distinct effects. Membrane-bound ligands may be presented in a dense manner, “cross-linked” by the membrane to which they are physically attached, allowing the recruitment of more than three receptors in close proximity. Depending on the receptor, this may or may not be required for the generation of an optimal intracellular signal. As an example, TNF-R1 signals in response to soluble TNF, whereas TNF-R2 requires membrane-bound TNF. We wondered whether this concept could be extended to various forms of soluble ligands. Indeed, soluble BAFF exists either as trimers (BAFF 3-mer), or as an ordered assembly of 20 trimers (BAFF 60-mer), whereas APRIL can interact with carbohydrate side chains of proteoglycans that may cross-link soluble trimeric APRIL (Fig. 1). We reasoned that the occurrence of BAFF and APRIL as trimers and higher order multimers might be functionally relevant, should they trigger receptors differently. BAFF binds to BAFF-R, TACI and BCMA, whereas APRIL binds TACI and BCMA. We indeed found that BAFF 3-mer and BAFF 60-mer could both signal through BAFFR, but that TACI (and BCMA) only responded to BAFF 60-mer and not to BAFF 3-mer. Similar observations were made with APRIL, which needed to be cross-linked in order to activate TACI. In other experiments, we have compared wild type mice, BAFF -/- mice and mice expressing membrane-bound BAFF only. When required, these mice were also provided with recombinant forms of BAFF 3-mer or BAFF 60-mer. Membrane-bound BAFF was unable to maintain the mature B cell pool, whereas soluble BAFF could. Membrane-bound BAFF (or BAFF 60-mer) was however required to induce specific B cell markers in the mature B cell pool. In collaboration with Dr B. Huard in Geneva, we also addressed the physiological relevance of APRIL-proteoglycan interaction in human. Proteoglycan-bound APRIL improved survival of primary human plasma cells ex vivo. Also, APRIL produced in small amounts by epithelial tonsil cells in a normal situation, and in greater amounts by neutrophils upon inflammation, accumulated on neighboring proteoglycans and correlated with the presence of numerous plasma cells. Proteoglycan-bound APRIL is therefore likely to provide a survival niche for plasma cells.
Of course, several questions remain to be addressed: what are the relative contributions of membrane-bound BAFF and BAFF 60-mer in vivo? Which of the different forms of BAFF and APRIL are to be found in vivo? Does their ratio change in pathological conditions? Are they all sensitive to BAFF-blocking drugs currently tested in clinical trials?