Following the recent publication of a paper called “GABAergic inhibition in dual-transmission cholinergic and GABAergic striatal interneurons is abolished in Parkinson disease“, Erwan Bézard, director of the Neurodégénérative Diseases Institute of Bordeauxpublished a commentary, that you can find here.

Here’s our answer:

First of all, we always pay attention to discussions and criticisms and we welcome debates. It is unfortunate that not all innovation and discoveries of this work have been analysed because the discussion of the conceptual openings that this work brings is important and deserve to be discussed. As a reminder, we have demonstrated in a constrained way the release of acetylcholine and GABA by the same neurons of the striatum with the best possible techniques, ranging from recording of pairs of neurons, to the quantification of the total number of positive LHX6 and Chat cells in the lateral striatum using the iDISCO clarification technique, and intracellular qPCRs and multiple recordings of the same neuron to determine the role of intracellular chloride levels in its response. The demonstration that the GABA component is impaired in the absence of dopamine -cell attached and then re-recording of the same neuron with variable chloride levels- is a major point of interest in the context of understanding Parkinson’s pathology.

  1. 6-hydroxydopamine (6-OHDA) isn’t of course an ideal method. Moreover, the ideal experimental model of Parkinson disease does not exist and our colleague is well placed to know it. Still, Mr. Bézard, as everyone else, bases his work on this same model (Bastide et al, Sci Rep, 2017). Our aim was to show that a drug that has nothing to do with dopamine –its primary target is exclusively the chloride co-transporter- reduces the electrical and behavioural signals of Parkinson by restoring low chloride levels. The intention is not to pretend that the dopamine doesn’t have an effect, but it is a well-defined pharmaceutical entity that goes out of the usual dopamine-centred routine and that shows decisively that, as in other pathologies, the chloride levels are increased (Ben-Ari, TINS, 2017).
  2. Regarding Mr. Bézard’s argument to oversell this work, it might have been fair to point out that in parallel we are doing clinical trials based on the same molecule. We did a successful pilot trial with bumetanide on 4 patients (Damier et al, 2016) and a large phase 2 trial is being developed on 40 patients. In other words, we are conducting parallel experimental studies to test and understand the mechanisms and clinical trials to test in the patient the hypothesis: from bench to bed !

The numerous works that begin to show the liberation of 2 transmitters will become a major research and innovation topic, and this work bring an important contribution by showing that one of these 2 transmitters can keep a pathological signature that could constitute a therapeutic target.

Y. Ben-Ari

Neurochlore, BA Therapeutics et Ben-Ari Institute of Neuroarcheology  


An answer to Erwan Bézard 

The main scientific criticism of E. Bézard on our paper concerns the behavioral aspect of our study. Before answering clearly and numerically to these criticisms targeted on the 6e-f figures, we would like to say that it is a pity that the numerous other data have not also been the subject of an analysis.

As we show below, with the support of figures, all the assumptions of E. Bézard appear to be wrong. We would like to point out the, in our long life of researchers, this is the first time that we have been dealing with a reviewer who makes statistics at sight and gives results with this method.

Figure 6e :


We did this experiment as following. We measured the time for a group A of control mice and a groupe B of 6-OHDA mice to cross a roller path (roller test). Their crossing times (continuous variable) have been compared using an “unpaired 2 tailed-test” (after verifying the normality of the distribution of the crossing time for each group) in order to validate our set-up and define if the 6-OHDA mice showed a significant deficit compared to the control mice. Once this data was verified, we returned to group B one month after the first test and compared their crossing time before (new test) and after treatment with bumetanide. We cheked the normality of the distribution of the crossing time differences, and did a « paired 2-tailed t-test ».

E. Bézard tells us that a one- or two-way ANOVA test would have been adequate and would have resulted in a non-effect. This ANOVA test is not suitable because we don’t have 3 distinct groups but one control A group and two paired B groups (maybe we should have done two separated diagrams). However, we did that test. The ANOVA test is in the table below (print screen) and clearly shows significant differences (green) between the crossing times of the control (ctr) and 6-OHDA (60HDA1)mice, and between 6-OHDA (6-OHDA2) and 6-OHDA2+Bumetanide (Bum) mice. It also shows an absence of significance between 6-OHDA1 mice’s first test and the same 6-OHDA2 mice’s second test as well as between control mice and 6-OHDA2+ Bumetanide mice.

6e 1


Figure 6f :


Pole test: this is discontinuous data since we compared scores (although some statisticians think they could be considerate as continuous, see Carifio & Perla 2008). The three groups of mice are independent. We couldn’t test the 6-OHDA mice before and after treatment because the increase in weight with age prevented any comparison.

E. Bézard tells us that a non-parametric ANOVA test should have been used (for example Kruskal-Wallis) and that it would have resulted in a non-effect. Here is the requested test (print screen) that shows the significant differences (yellow) between control and 6-OHDA mice and between 6-OHDA and 6-OHDA+bumetanide mice. On the other hand, there is no difference between control and 6-OHDA+ Bumetanide mice.

6f 1.png

We also did the requested statistical analysis in python and the result is still significant:


Sanaz Eftekhari, Natalia Lozovaya and Constance Hammond with Hamed Rabiei for python statistical analysis (the code can be given on request).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s