Firstly, Dr Cathleen Lutz (The Jackson Laboratory, USA) discussed the range of current mouse models of MND and the potential for generating new models. Lutz explored the many reasons why MND is a hard disease to model, and how the SOD1-G93A model has become the ‘workhorse’ for MND mouse modelling. However, as subgroups and new genetic mutations of MND are identified, new avenues for creating mouse models are opening up.
Models of SOD1, TDP43 and C9ORF72 were discussed, highlighting the ever-increasing array of MND mouse models which all have their own advantages and disadvantages for use.
With new CRISPR technology, new models can be created on different genetic backgrounds much faster and more efficiently than previously, which raises the question of whether we should be exploring the effect of genetic background variations on each of the models. Currently models are made using inbred strains to avoid genetic variation. However, is it more representative to create population models of the disease, in which there is genetic variation, as would be seen in samples of patients?
Dr Adam Walker (now at Macquarie University, Australia) presented data on a mouse model of TDP-43 from the University of Pennsylvania, USA. A mouse model was created which expressed hTDP-43∆NLS when doxycycline was withdrawn from the diet; dosing with doxycycline suppressed expression. Mice were characterised after withdrawal of doxycycline and recovery was also investigated when mice were fed doxycycline after a 6 week withdrawal period.
Mice were found to express hTDP-43∆NLS in the brain and spinal cord after 1 week of doxycycline withdrawal and had TDP43 pathology in the brain and spinal cord. A progressive motor phenotype was evident from 2 weeks post-withdrawal with brain and muscle atrophy by 4 weeks post-withdrawal and spinal cord motor neuron loss from 6 weeks post-withdrawal. The end stage of the disease course was around 10 weeks post-withdrawal but cause of death was unclear.
Conversely, when fed doxycycline after a 6 week withdrawal period, TDP43 pathology was cleared, reinnervation occurred, motor phenotype improved and death was prevented. This model therefore provides promise that potential therapies for MND may halt disease progression and improve symptoms, even at the late stages of disease.
Following this, Dr Qiang Zhu (University of California, USA) presented on the development of mouse models of C9ORF72. Mouse models for loss and gain of function were discussed, as the mechanism of C9ORF72 in MND is under debate.
Loss of function was modelled using a knockout of C9ORF72, in which the heterozygous knockout mouse showed no significant defects and normal survival. The homozygous knockout mice showed no signs of a motor phenotype but did show signs of mild age-dependent social changes. However, the homozygous knockout mice also developed T-cell lymphoma (C9ORF72 is a tumour suppressor).
Gain of function was modelled using an insertion of GGGGCC repeats in C9ORF72. The BAC transgenic mice showed repeat size, expression-dependent repeat-associated non-ATG (RAN) foci and RAN products accumulation, age dependency of RAN aggregate formation, hippocampal neuron loss, increased anxiety and impaired memory and learning ability in an age-dependent manner. Moreover, the efficacy of anti-sense oligonucleotides (ASO) was tested in these transgenic mice. ASO treatment significantly decreases repeat-containing RNA, RNA foci and polydipeptides without significantly reducing overall C9orf72.
These models suggest a potential role of loss and/or gain of function in C9ORF72 MND, and demonstrate the efficacy of ASO therapy.