On 2016 Aug 16, Niels Vrang commented:

We welcome discussion and critique although we do not approve of the rather broad spectrum of points raised in this comment. We do stand by our results which have been generated by several skilled technicians and scientists known in the art of neuroscience and the methods employed live up to scientific standards. Our findings are but a series of findings trying to understand the putative role of GLP-1 agonists in neurodegenerative disorders and in this particular case the findings were negative. The discussion of the paper clearly recognise that others have found positive data. We believe that the best way to advance science is by reporting both positive and negative data and we are thankful for PLoS One for making the reporting of negative findings possible.

On 2016 Aug 08, Christian Holscher commented:

This is a misleading and unscientific paper. The authors measure amyloid plaque load in a London mutation APP mouse model which does not develop amyloid plaques! In the London mutation, the amyloid stays mostly inside the cells. Fig. 5 clearly shows that the brains are virtually free of plaques, yet the authors conclude that the drug failed to reduce amyloid plaque load! They should have measured total amyloid levels using the western blot technique. The memory tasks are similarly dubious. The wild type control mice are a lot lighter than the transgenic mice (Fig 1A) and they swim a lot faster in the water maze (Fig. 3A-C), which makes the interpretation of the result questionable. The reduced latency can be explained entirely by the faster swim speed. The memory tests of the APP/PS1 mice are just as unscientific. Fig. 4 shows that the saline treated APP/PS1 mice do not show a memory deficit in both tasks when compared to controls. The drug cannot improve a non-existing memory deficit! The experiment did not work and needs to be repeated. Showing these graphs in the publication is misleading. This study is deeply flawed and the conclusions are not supported by the data shown. It should be retracted.

Prof. Christian Holscher, PhD Lancaster University, UK

On 2016 Aug 08, Christian Holscher commented:

This is a misleading and unscientific paper. The authors measure amyloid plaque load in a London mutation APP mouse model which does not develop amyloid plaques! In the London mutation, the amyloid stays mostly inside the cells. Fig. 5 clearly shows that the brains are virtually free of plaques, yet the authors conclude that the drug failed to reduce amyloid plaque load! They should have measured total amyloid levels using the western blot technique. The memory tasks are similarly dubious. The wild type control mice are a lot lighter than the transgenic mice (Fig 1A) and they swim a lot faster in the water maze (Fig. 3A-C), which makes the interpretation of the result questionable. The reduced latency can be explained entirely by the faster swim speed. The memory tests of the APP/PS1 mice are just as unscientific. Fig. 4 shows that the saline treated APP/PS1 mice do not show a memory deficit in both tasks when compared to controls. The drug cannot improve a non-existing memory deficit! The experiment did not work and needs to be repeated. Showing these graphs in the publication is misleading. This study is deeply flawed and the conclusions are not supported by the data shown. It should be retracted.

Prof. Christian Holscher, PhD Lancaster University, UK