2 Aug 2013. The struggle to develop immunotherapy for prion disease has yielded new clues into how cellular prion protein (PrPc) kills neurons. In the August 1 Nature, researchers led by Adriano Aguzzi at University Hospital Zurich, Switzerland, report that some antibodies that bind a globular domain at the C-terminus of the prion protein potently exacerbate neurodegeneration rather than stop it. That’s right, the antibodies hasten neuron death. Surprisingly, locking the N-terminal flexible domain of PrPc protected against the harmful C-terminal antibodies. Aguzzi and colleagues conclude that toxic signals from the globular domain can be transmitted through the N-terminal flexible region. While the data suggest an intriguing model for prion toxicity and potentially opens new therapeutic avenues, experts caution that the relevance to human disease is unclear.
Earlier studies in mice helped lay the groundwork for using vaccines to treat prion disease. Expression of anti-prion antibodies in animals protected them from prions injected into the brain and spleen (Heppner et al., 2001), and immunization with anti-prion antibodies prevented disease from spreading to the brain (see ARF news story). Since researchers believe prion disease results from interactions between PrPc and its pathological, misfolded variant (PrPSc), Aguzzi and colleagues sought to model this pathology by generating a series of 19 monoclonal antibodies that recognize different epitopes in the N-terminal flexible region and C-terminal globular domain of mouse PrPc (Polymenidou et al., 2008) .
In the current study, co-lead authors Tiziana Sonati, Regina Reimann, and Jeppe Falsig analyzed the antibodies' effects on cerebellar slices from wild-type or Tga20 transgenic mice. Tga20 mice overexpress brain PrPc eight-fold and succumb to prion disease. Much to the authors' surprise, “many of the antibodies were super toxic,” Aguzzi told Alzforum. “They were wiping out the (Tga20) slices in a few days.” Wild-type cells died too, albeit more slowly. The toxicity was epitope-dependent. All eight antibodies causing dramatic cell loss recognized the globular domain (GD) of prion protein.
The demonstration of neuronal death by anti-prion antibodies was in itself not new. In an earlier study, scientists at Scripps Research Institute, La Jolla, California, showed that antibody-mediated crosslinking of PrPc could kill brain cells in mice (see ARF news story). Curiously, though, the Zurich scientists saw that the mouse cerebellar slices shrank drastically no matter which form of GD antibodies they used—whole antibody, Fab fragments, even tiny pieces with a single antigen-binding site. This suggested that toxicity does not require crosslinking. The researchers began to suspect that the binding of GD antibodies induces conformational changes on PrPc that trigger neurotoxicity. However, using X-ray diffraction, solution nuclear magnetic resonance (NMR), and other biophysical techniques, they determined that one of the toxic anti-GD antibodies (called POM1) did not distort the GD region of human PrPc.
Could POM1 alter the N-terminal flexible PrPc region instead? To find out, the researchers administered it to brain slices expressing a PrPc mutant with most of the flexible tail lopped off. This mutant—but not those with deletions in other non-GD regions—allowed POM1 binding yet protected the cells against all tested neurotoxic antibodies. These data suggest that GD-mediated neurodegeneration requires the PrPc flexible tail. Bolstering that idea, the researchers rendered POM1 harmless by pre-incubating cells with a different antibody (POM2) that blocks the PrPc flexible tail.
Moving the analyses in vivo, Sonati and colleagues injected POM1 or control immunoglobulins into the cerebella of Tga20 mice. Magnetic resonance imaging (MRI) revealed brain lesions as soon as a few hours after injection. Substantial neuron loss and glial scars showed up a week later. Mice expressing PrPc without its flexible tail resisted POM1 toxicity, whereas animals expressing other PrPc mutants did not. Furthermore, mice expressing a PrPc mutant lacking the central domain between N-terminal tail and C-terminal globular domain were telling. These animals spontaneously develop neurodegeneration, but the researchers showed that a single injection of an anti-flexible tail antibody improved their survival.
All told, the data point to a model where GD antibody binding mimics the interaction of pathological PrPSc with PrPc, causing the PrPc flexible tail to misfold and transmit a toxic signal to the neuron. Aguzzi explains the principle findings in this video.
[Video credit: ©University of Zurich]
How do these findings relate to ongoing prion immunotherapy programs? Other prion scientists find the results intriguing but difficult to interpret. “Anything is toxic if you inject enough of it. The question is whether you can get a therapeutic effect with a dose that is safe,” said John Collinge, who is developing therapeutic antibodies for prion disease at the University College London, UK. In response to the 2004 Scripps study that reported anti-prion antibodies cause massive neuron loss in mice, Collinge and colleagues tested their own therapeutic PrP antibodies, along with neurotoxic antibodies from the Scripps group. The scientists reported finding no evidence of cell death with any of them (Klöhn et al., 2012). A humanized version of one antibody is currently in safety testing with a view toward clinical trials, Collinge told Alzforum.
Collinge said his and the Scripps studies used anti-prion antibodies at extremely high concentrations in the brain—nearly three orders of magnitude above the binding constant. Those were already “much higher than the concentrations that cure prion-infected cells,” Collinge said, pointing out that Aguzzi’s present study tested anti-prion antibodies at still higher concentrations. Other questions remain. For example, Collinge said, the POM15 and POM17 prion antibodies recognize the same GD epitope with similar affinity, yet POM17 is toxic while POM15 is not.
Gianluigi Forloni of the Mario Negri Institute in Milan, Italy, said the new data are “interesting and could be useful” but also wonders how well they apply to human disease. Forloni agreed with the authors that “the dramatic effects of GD ligands in vivo suggest that anti-GD autoimmunity may cause neurological conditions.” Aguzzi told Alzforum his team has received ethical review board clearance for a retrospective study to look for anti-GD autoantibodies in patients with idiopathic neurodegeneration.—Esther Landhuis
Sonati T, Reimann RR, Falsig J, Baral PK, O’Connor T, Hornemann S, Yaganoglu S, Li B, Herrmann US, Wieland B, Swayampakula M, Rahman MH, Das D, Kav N, Riek R, Liberski PP, James MN, Aguzzi A. The toxicity of antiprion antibodies is mediated by the flexible tail of the prion protein. Nature. 1 August 2013. Abstract.