05 May 2005

Mutant proteins that cause neurodegenerative disease, such as amyloid-β precursor protein, parkin, and huntingtin, are produced in many different neurons of the brain. So why do only certain subtypes of neurons degenerate in Alzheimer disease, Parkinson disease, and Huntington disease (HD)? Could it be that these proteins are less the Trojan horse but more the trebuchet, hurtling insults at specific cells from afar? A report in today’s Neuron suggests that, in the case of HD, toxicity from the outside is as much to blame as that on the inside of neurons.

In collaboration with colleagues in the US and Japan, William Yang and coworkers at the University of California used conditional expression of mutant huntingtin to test the cell-autonomous, or Trojan horse, model of HD in mice. This model suggests that toxicity occurs solely within cells and does not require pathological interaction among cells. First author Xiaofeng Gu and colleagues used the Cre/LoxP recombination system to delete transcriptional stop signals and so unblock expression of mutant huntingtin (mhtt) in selected neurons of the brain. They generated two types of animals. In one, mhtt (exon 1 of the human gene with a 103 CAG trinucleotide repeat) was expressed in all neurons, while in the second, mhtt was only expressed in pyramidal neurons of the cortex. If the cell-autonomous model held up, then the pyramidal neurons should degenerate when mhtt is expressed in those neurons alone. In fact, this is not what the authors observed.

Gu and colleagues found that in the pyramidal cell-only model, there was hardly any degeneration of those neurons, despite the fact that mhtt was robustly produced in layers II and III of the cortex. The animals had very little HD-like pathology at all. They had no locomotor deficits at any age tested (up to 10 months), had no reactive gliosis, which is common in HD models, and tests which normally detect ultrastructural changes in pyramidal neurons, such as increased uptake of osmium and the presence of abnormal, wavy apical dendrites, proved negative. In contrast, when mhtt was expressed throughout the brain, including the pyramidal neurons and the medium spiny neurons of the striatum—the two regions most badly affected in HD patients—pathological effects were rampant. Gliosis was increased about sixfold in the cortex and threefold in the striatum compared to that in wild-type animals. The transgenic animals also had significantly reduced locomotor activity once they reached about 6 months old, and it continued to decline. Thirty-five out of 37 pyramidal neurons tested had wavy apical dendrites (as compared to only eight out of 45 in the pyramidal cell-only model), and electron micrographs showed dark, heavily stained neurons in the cortices and striata.

“In summary, our neuropathological studies show that restricting mhtt expression to cortical pyramidal neurons produced very limited pathology in these neurons,” write the authors. Instead, the data indicate that some cell-cell interaction might be necessary for degeneration of these neurons.

To test this idea, Gu and colleagues examined neuronal circuitry. They found that in animals globally expressing mhtt in the brain, excitatory input to pyramidal neurons was normal. Inhibitory input from GABAergic cortical interneurons, however, was significantly affected. While the amplitude of inhibitory postsynaptic currents was similar to controls, the frequency was reduced by about 25 percent. The authors conclude that “pathological cell-cell interactions can significantly contribute to neuronal toxicity in vivo in a mouse model of HD.” They also raise the possibility that similar interactions could drive pathology in Alzheimer disease, Parkinson disease, and other diseases mediated by polyglutamine expansions. Cell-cell interactions may also be important in amyotrophic lateral sclerosis as shown recently by mouse models of the disease in which expression of mutant superoxide dismutase in glia leads to degeneration of neurons (see ARF related news story).—Tom Fagan