This scientific commentary refers to ‘Neutralization of TNFSF10 ameliorates functional outcome in a murine model of Alzheimer’s disease’ by Cantarella et al. (doi: ).
Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), now known as TNFSF10, is a member of the TNF superfamily identified by Wiley and colleagues (1995). There are five known subtypes of TNFSF10 receptor, and binding of TNFSF10 to DR4 (TNFRSF10A) and DR5 (TNFRSF10B) receptors, for example, can trigger cell death through activation of caspases leading to apoptosis (Ashkenazi et al., 2008). Activation of TNFSF10 receptors has been suggested to have anti-tumour efficacy; however, variable levels of TNFSF10 receptor expression on tumour cells may limit the usefulness of this approach (Ashkenazi et al., 2008). Whereas TNFSF10 is not detectable in healthy human brain, its expression is upregulated in several neurodegenerative diseases including Alzheimer’s disease (Uberti et al., 2004). In this issue of Brain, Cantarella and colleagues report that an anti- TNFSF10 antibody can reduce brain amyloid-β load and activation of TNFSF10 apoptotic receptors, as well as improve cognition, in a triple transgenic mouse model of Alzheimer’s disease (Cantarella et al., 2014).
Much of the neuronal death in Alzheimer’s disease occurs in brain regions linked to the processing of memory (Walsh and Selkoe, 2004). The pathological hallmarks of the disease are the accumulation of neurotoxic oligomers of amyloid-β, and the formation of neurofibrillary tangles—intraneuronal aggregates of hyperphosphorylated tau protein—both of which lead to neuronal death (Walsh and Selkoe, 2004). The inflammatory process may also have a role in Alzheimer’s disease: an increase in inflammation has been linked to progression of the disease (Wyss-Coray, 2006), while polymorphisms in the TNFSF10 DR4 receptor gene (TNFRSFA) may influence Alzheimer’s disease susceptibility (Edgunlu et al., 2013). Moreover, amyloid-β deposition has been suggested to activate TNFSF10 apoptotic pathways in neurons, whereas blockade of this cascade via a TNFSF10-neutralizing monoclonal antibody appears to prevent neurotoxicity in vitro (Cantarella et al., 2003). While epidemiology research suggests that non-steroidal anti-inflammatory drugs may have therapeutic efficacy in Alzheimer’s disease, clinical trials in patients failed to show a direct effect (Heneka et al., 2011). Nevertheless, therapies targeting specific inflammatory compounds may hold promise. Here, Cantarella and colleagues demonstrate that intraperitoneal injection of an anti-TNFSF10 antibody for 6 months improves cognition in an Alzheimer’s disease animal model. They also report a reduction in inflammatory markers such as interleukin 1, inducible nitric oxide synthase—which may increase oxidative stress in neurons—and cyclooxygenase, which has previously been linked to Alzheimer’s disease pathology. Furthermore, the results suggest mechanisms of neurotoxicity that may be relevant to a variety of neurodegenerative diseases, including stroke (Cui et al., 2010).
TNFSF10 is thought to be expressed by peripheral immune cells, glial cells and neurons. However, serum levels of TNFSF10 are not consistently increased in patients with Alzheimer’s disease relative to healthy controls (Genc et al., 2009). This suggests that, if TNFSF10 plays a pathogenic role it is secreted within the CNS. One possibility is that, in response to either autocrine or paracrine signalling, neurons secrete TNFSF10 to induce apoptosis of cells affected by amyloid-β and neurofibrillary tangle pathology. Consistent with this, amyloid-β has been shown to mediate neuronal death through TNFSF10 death pathways (Cantarella et al., 2003). Furthermore, while there is clear evidence for increased numbers of astrocytes and activated microglia in the vicinity of amyloid plaques, there is less consensus on their role in pathology (Wyss-Coray, 2006). Some argue that an increase in these cells leads to further neuronal stress, which might be associated with increased activation of TNFSF10 death pathways. Others argue that astrocytes and microglia are important for clearing and degrading neurotoxic amyloid-β oligomers (Wyss-Coray, 2006).
Cantarella and colleagues now show the presence of their anti-TNFSF10 antibody in the brain of an Alzheimer’s disease mouse model, 10 days after its intraperitoneal injection. This might suggest that the antibody has the ability to penetrate the brain and to neutralize TNFSF10 pathways directly. However, there were no differences in brain levels of anti-TNFSF10 antibodies in the Alzheimer’s disease mouse model versus control mice. This could indicate that the antibodies block TNFSF10 secreted by infiltrating monocytes, which may play a role in the inflammation process in Alzheimer’s disease. Thus, evidence suggests that neutralization of TNFSF10 may block death signals mediated by TNFSF10 derived from three potential sources (Fig. 1): (i) neurons that secrete TNFSF10 under stress caused by an increase in levels of oligomeric amyloid-β; (ii) infiltrating monocytes that are attracted to amyloid plaques; or (iii) cells such as astrocytes and microglia, which secrete TNFSF10 in response to oligomeric amyloid-β and an increase in stress signals from neurons.
Possible mechanisms of action of anti-TNFSF10 antibodies in the Alzheimer’s disease mouse model brain. There are three mechanisms by which TNFSF10 (TRAIL)-specific antibodies might reduce TNFSF10 death signals in neurons: (1) blocking TNFSF10 secreted by neurons; (2) neutralizing TNFSF10 from infiltrating monocytes; and (3) blocking TNFSF10 secreted by astrocytes and microglial cells.
Immunotherapeutic approaches using antibodies to target amyloid pathology have shown promise in animal models (Weiner and Frenkel, 2006). However, the results of clinical trials have been less convincing (Farlow and Brosch, 2013). This might be due to the fact that the antibodies were given in the late stages of the disease. Early intervention might still prove to be beneficial, although this will rely on the development of effective diagnostic tools. Here, Cantarella and colleagues show that anti-TNFSF10 antibodies have therapeutic efficacy when given prior to the emergence of clinical pathology in mice. It will be interesting to see whether efficacy can still be shown after the onset of cognitive impairment. Furthermore, given that neurofibrillary tangle pathology has been strongly linked to neuronal death in Alzheimer’s disease, it will be important to assess the efficacy of anti-TNFSF10 antibodies also on the development of neurofibrillary tangles.
Cantarella and colleagues convincingly demonstrate a reduction in total amyloid-β load and suggest that it might be due to a reduction in stress signals of inflammatory markers from astrocytes and microglia. However, it is worth examining whether anti-TNFSF10 antibodies might also increase the efficacy with which these phagocytotic cells take up and degrade amyloid-β oligomers. Some anti-inflammatory modulators have been reported to increase phagocytosis of amyloid-β by these cells (Heneka et al., 2011). This might suggest that the observed reduction in activated astrocytes and microglia, thought to reflect a reduction in total brain amyloid-β, may also be related to the fact that these cells help to clear amyloid-β. Of note, this is an interesting example of using an antibody to target a specific inflammatory parameter in order to indirectly affect brain amyloid-β, rather than directly targeting amyloid-β itself.
Ultimately therefore, this study by Cantarella and colleagues suggests a new avenue for therapeutic intervention in Alzheimer’s disease in the form of blocking neuronal stress signals that might trigger neurodegeneration.
