The essential role of the FADD molecule was highlighted by generating FADD mutant null mice 1112. Indeed, FADD knockout mice were not viable. FADD null embryos died in utero at day 12.5 of development, due to underdevelopment, abdominal hemorrhage, and cardiac failure. Moreover, FADD loss of function did not result in a lymphoproliferative disorder as observed in viable Fas mutant mice 1314. These results indicated that in addition to its well known role in cell death, FADD was also implicated in survival/proliferation of some cell types.

FADD is the main signal transducing intermediate adaptor molecule of several death receptors including Fas, TNF-R1 (tumor necrosis factor receptor 1), DR3 (death receptor 3), TRAIL-R1 (TNF-related apoptosis-inducing ligand, DR4), and TRAIL-R2 (DR5) 4111215. All these receptors possess, in their intra-cytoplasmic tail, a DD homologous to the DD of FADD allowing FADD recruitment to the activated receptor. FADD can be recruited either directly to Fas and TRAIL-Rs (Figure 2A) or indirectly to TNF-R1 (Figure 2B). In the latter case, FADD is recruited via another DD-containing adaptor molecule (TRADD, TNF receptor-associated protein with DD). Next, FADD recruits DED-containing initiator pro-caspase 8 or 10 through DED/DED interactions 161718, thus forming the death-inducing signaling complex (DISC) 19. Autoprocessing of initiator pro-caspase leads to activation of effective caspases which cleave intracellular substrates, causing the apoptotic death of the cell 2021.

Control of FADD recruitment to the DISC can occur following several mechanisms depending on the cell type and the death receptor 22. The best characterized death receptor signaling inhibitors are DED-containing viral and cellular FLICE-inhibitory proteins (v-FLIPs and c-FLIPs, respectively) 2324. Inhibition of Fas-, TNF-R1-, and TRAIL-Rs-induced apoptosis by endogenous FLIPs results from binding of the c-FLIPs to the DED of FADD, thus hindering pro-caspase 8 activation (Figure 2A). Similarly, v-FLIPs inhibit apoptosis mediated by death receptors either by binding to FADD and blocking pro-caspase 8 processing, or by binding to pro-caspase 8 and inhibiting FADD interaction. Therefore, equilibrium between FADD and the expression of its inhibitors determines the outcome of the death receptor-stimulated cell, i.e. apoptosis or survival.

All the main death receptors described up to now require FADD adaptor for transmitting their apoptotic signal. Consequently, FADD is a central protein that controls multiple essential cellular processes including cellular homeostasis and elimination of pathological cells, particularly during the course of an immune response.

Formation of cytoplasmic death effector filaments (DEF) by oligomerization of DED-containing proteins, including FADD, is responsible for death receptor independent cellular apoptosis 25. Indeed, FADD over-expression by itself is known to induce cell death through DEF formation that recruits and activates pro-caspase 8. However, the existence of DEF has not been established in vivo, and increasing evidences suggest that DEF could be artefactual structures resulting from protein over-expression. As a consequence, the ability of endogenous FADD to aggregate and form DEF in normal situation should be reconsidered.

Beside being a main death adaptor molecule, FADD is also required for T cell proliferation. The first evidence of this property of FADD came from observations made in chimeric FADD knockout mice. Five-week-old chimeric FADD^-/- mice presented a lack of thymocytes compared to wild type animals, with few or no CD4⁺ CD8⁺ double positive thymocytes remaining 12. Moreover, several groups have demonstrated that FADD deficiency in peripheral T lymphocytes resulted in an inhibition of mitogen-induced T cell proliferation 122627282930. The mechanism leading to FADD-dependent T cell proliferation did not involve the early events associated with cell proliferation since expression level and functionality of the IL-2 receptor, level of IL-2 secretion, mobilization of intracellular calcium, and activation of NF-kB, p38-MAPK, and p44/42-MAPK appeared normal in FADD^-/- T lymphocytes 122930. Recent data showed that FADD^-/- T lymphocytes entered the cell cycle upon mitogenic stimulation, but died during progression through the cell cycle 30. Therefore, lack of proliferation of FADD-deficient T cells results from defective survival associated with progression through the cell cycle rather than defective activation. Up to now, the molecular pathway implicated in FADD-mediated survival of lymphocytes has not been described.

In addition to impairing survival during cell division, FADD deficiency also leads to a dysregulation of the cell cycle machinery 31. The pattern of expression of molecules implicated in both G1/S and G2/M transitions was aberrant in FADD^-/- lymphocytes, resulting in spontaneous entry and progression through the cell cycle of 10% of freshly isolated FADD^-/- T cells (as compared to less than 2% of wild type T cells) 31.

The mechanisms responsible for FADD regulation of cell cycle progression are not fully understood. However, the phosphorylation of the 194 human and 191 mouse Ser of the protein (Figure 1) has recently drawn attention. Indeed, human FADD was phosphorylated at Ser 194, by a still unidentified 70 kDa protein kinase, in cells arrested in G2/M, whereas it was unphosphorylated in G1/S 7. Generation of FADD Ser 191 mutant mice confirmed that FADD phosphorylation is involved in proliferation in vivo 8. Replacement of Ser 191 by an aspartic acid resulting in constitutive phosphorylation of the FADD protein led to abnormal development of FADD mutant mice that shared the same phenotype as FADD deficient mice 8, including few CD4⁺ CD8⁺ thymocytes and defective progression through the cell cycle 1231.

Tumor cells are constantly cycling cells. Although it does not seem to directly affect the cell cycle progression, FADD phosphorylation at Ser 194 sensitizes these cells to reagents that induce G2/M arrest such as the Taxol anticancer drug 32. In human prostate cancer cell lines, treatment with Taxol resulted in Ser 194 FADD phosphorylation and G2/M arrest 33. Moreover, etoposide or cisplatin chemotherapeutic drug-induced apoptosis of these cells was enhanced by pretreatment with Taxol, a process that was inhibited by cellular over-expression of an unphosphorylable FADD mutant 33. Therefore, tumor cells that express a Ser 194 FADD mutant that cannot be phosphorylated or are unable to phosphorylate FADD at this Ser position are expected to resist apoptosis induced by anticancer drugs that induce G2/M arrest, and to be insensitive to the synergistic effect of chemotherapy. Obviously, a lack of FADD expression will have the same consequences.

The role of the FADD adaptor in cancer was initially demonstrated by generating RAG-1 deficient transgenic mice that target expression of a FADD dominant negative mutant in lymphocytes. With age these mice developed thymic lymphoblastic lymphoma, whereas FADD^+/+ RAG-1^-/- mice did not 34. Moreover, thymic lymphoblastic lymphomas were never observed in FADD^-/- RAG-1^+/+ or FADD^-/- RAG-1^+/- mice, demonstrating that absence of FADD expression was necessary but not sufficient to induce tumor development in this model 34. These results were the first demonstration that FADD adaptor can act as a tumor suppressor in vivo.

Using a mouse model of thyroid adenoma/adenocarcinoma, we showed spontaneous disappearance of FADD protein expression during the course of tumor development 35. The so called gsp transgenic mice expressed an oncogene specifically in thyroid follicular cells (TFC), and developed thyroid hyperplasia that eventually transformed into hyperfunctioning adenomas or adenocarcinomas around the age of 8 months 36. The fact that gsp mice developed hyperfunctioning adenomas or adenocarcinomas belatedly suggested that oncogene expression conferred a predisposition but that an additional event was necessary for thyroid tumor development. We found that FADD protein was highly expressed in all non-pathological and in almost all hyperplastic thyroid glands from gsp mice. In contrast, thyroid adenoma/adenocarcinoma expressed low or no FADD protein 35. These results raised the possibility that loss of FADD protein expression could be an additional event contributing to tumorigenesis, and suggested that FADD plays a role as a tumor suppressor.

We recently showed that absence of FADD protein expression in cancer cells is also a relevant phenomenon in human malignancies 37. We looked for FADD protein expression in human acute myeloid leukemia (AML) cells. Leukemic cells of most AML patients are resistant to Fas-mediated cell death despite expressing the Fas receptor 38 and/or the FasL molecule 39. Moreover, chemotherapeutic drugs used for AML treatment can kill target cells via several mechanisms, including death receptor-induced apoptosis 40414243. We performed a retrospective study of 70 consecutive patients with de novo AML treated homogeneously, and found that leukemic cells of 2/3 of patients at diagnosis expressed low or no FADD protein 37. Moreover, in this cohort of patients, we showed that absent/low FADD protein expression in AML cells at diagnosis was a new independent prognostic factor for poor response to chemotherapy (in terms of complete remission rate, event free and overall survivals) 37. Importantly, absent/low FADD protein expression in AML cells at diagnosis was a prognostic factor even for patients classified as standard- or good-risk AML cases by cytogenetic and molecular criteria 37. As a consequence, this new prognostic factor is of clinical importance since it will allow early identification of patients with chemoresistant AML who could benefit from more intensive post-remission therapy.

The fact that absence of FADD expression was found in different types of tumor cells both in mice and humans strongly suggested that absence of FADD contributed to tumor development. Indeed, lack of FADD protein can confer numerous advantages on pathological cells, which predominantly result in tumor survival/growth gain (Figure 3).

Absence of FADD expression could confer multiple growth advantages on cancer cells (Figure 3), and is expected to contribute to disease progression. As a consequence, finding how to restore FADD protein expression in FADD-negative tumor cells represents a research field with potentially direct clinical applications. In fact, it is possible that some of the current cancer treatments act, at least partially, through this mechanism. For instance, carboplatin is a cytotoxic drug potentially effective at reestablishing functional FADD protein expression. Indeed, the human tongue carcinoma cell lines SCC-9 and SCC-25 express very low levels of FADD protein. Moreover, treatment with carboplatin enhances FADD protein expression, thus rendering cancer cells sensitive to Fas-mediated apoptosis 60. The combination of carboplatin with chemotherapeutic drugs that induce death receptor-mediated cell death may result in improved treatment.

Molecules implicated in FADD regulation of expression also represent new therapeutic targets. However, very little is known about the mechanisms leading to absent/low FADD protein expression in tumor cells. In SCC-9 and SCC-25 carcinoma cell lines, carboplatin upregulated FADD protein expression by increasing FADD mRNA 60. However, FADD mRNA was normally expressed in mouse thyroid adenoma/adenocarcinoma and in human AML cells, and lack of FADD mRNA could not account for poor FADD protein expression in these cancer cells 3537. These results suggest that several mechanisms could be implicated in loss of FADD protein, depending on the type of cell and the environmental pressure. Besides, docking FADD protein away from the death receptor, in the nucleus for example, would have the same consequences. Understanding such mechanisms is the first necessary step towards development of new anticancer drugs targeting molecules that regulate FADD protein expression.