Nodding off when reading mouse EAE papers? Is it because they are full of mushroom food? Wishful thinking or simply terrible data analysis?

Autoimmune encephalomyelitis in NOD mice is not initially a progressive multiple sclerosis model. Baker D, Nutma E, O’Shea, Cooke A, Orian JM, Amor S, Baker D .Ann Clin Trans Neurol 2019 https://doi.org/10.1002/acn3.792

Objective

Despite progress in treating relapsing multiple sclerosis (MS), effective inhibition of nonrelapsing progressive MS is an urgent, unmet, clinical need. Animal models of MS, such as experimental autoimmune encephalomyelitis (EAE), provide valuable tools to examine the mechanisms contributing to disease and may be important for developing rational therapeutic approaches for treatment of progressive MS. It has been suggested that myelin oligodendrocyte glycoprotein (MOG) peptide residues 35‐55 (MOG^35‐55)‐induced EAE in nonobese diabetic (NOD) mice resembles secondary progressive MS. The objective was to determine whether the published data merits such claims.

Methods

Induction and monitoring of EAE in NOD mice and literature review.

Results

It is evident that the NOD mouse model lacks validity as a progressive MS model as the individual course seems to be an asynchronous, relapsing‐remitting neurodegenerative disease, characterized by increasingly poor recovery from relapse. The seemingly progressive course seen in group means of clinical score is an artifact of data handling and interpretation.

Interpretation

Although MOG^35‐55‐induced EAE in NOD mice may provide some clues about approaches to block neurodegeneration associated with the inflammatory penumbra as lesions form, it should not be used to justify trials in people with nonactive, progressive MS. This adds further support to the view that drug studies in animals should universally adopt transparent raw data deposition as part of the publication process, such that claims can adequately be interrogated. This transparency is important if animal‐based science is to remain a credible part of translational research in MS.

(People east of the UK may have seen this lauch early on Sunday. My mistake)

Some people are going to wake up today and my name and the name of Dr. Love will be mud. We will be off their Christmas card lists for good and probably their grant awarding list also. Why…because the inference is that some people have issues with data analysis, which is a terrible indictment.

This is because they are claiming their work is relevant to progressive MS, when it is probably isn’t in the way it is being used. Fear not. They will probably ignore this paper and carry on regardless. It is a shame because the underpinning work is often excellent, spoilt by some dogey EAE studies.

We now have EAE experiments by numbers, where every thing is done in a prescriptive way perhaps without sufficient thought. Sadly it is often done in a way that has limited relevance to humans. Why complain?…I’m not doing myself any favours putting my head on the block. A trouble maker? However, if no one says anything, bad practise simply amplifies and it leads to non-translatable science. The neurologists will blame the animals, when they should be looking hard at the science of the perps.

Indeed, I was at a meeting when this model was touted to justify a trial in progressive MS. The agent, in my mind was unlikely to work and may have actually caused harm. This galvanised my desire to investigate

So do I sit back and say nothing? Like most people would, or say something?. What would you prefer?

You may ask, why is this important?

Well you are rattling tins and spending good money supporting this type of stuff and you and me are pinning our hopes on this.

In addition the science lemmings simply follow in the footsteps and before you know it , it is dogma

Anyway what did we do?

It all boils down to the desire to study and treat progressive MS. In this case it is said that the Non Obese Diabetic mouse gets progressive disease. The NOD mice are known that they can spontaneously get a type 1 diabetes-like condition

In the paper the EAE shows a continued worsening without relapse. Therefore, you say it looks like progressive worsening without relapse. This could be a real benefit in drug testing. This is the claim made but a host of other labs using this model

However, we had seen this profile in our own work and could see the animals were relapsing but they were not relapsing at the same time and were recovering poorly. It was not progressive disease that was independent of relapses. The impression of progressive worsening was an artefact of the data handling.

In the 1990s we did work with NODs and didn’t see this type of progressive worsening. We went back to our lab books and dug out some experiments from the pre-millenium done with myelin peptides in NODs and saw relapses. Therefore, a fish was smelt.

However, we used the optimal myelin peptides for disease induction and not the suboptimal peptide that everyone was using for their work with C57BL/6 (nasty tempered, black furred mouse used for gene manipulation work) or the NOD (white fur, big feet and gets diabetes) mice.

Therefore, we wrote round including USA, Israel, Japan asking for raw data and nothing…or positive vibes (Sometimes repeated positive vibes) followed by no action. However, we did get a response when we went to Australia and got the data. They dug out an experiment that had been done so we didn’t have to do more experiments and sent us the data. Sure enough relapsing disease occurred. It wasn’t really progressive

Was it a lab thing? Different labs do things to get differences. It is possible and we are there to be proved wrong. However , I think the answer is no.

Indeed. There is nothing like authors hanging themselves.

In this case the authors describe the model as secondary progressive EAE, but show that it is relapsing. No body would be bothered to read a citation we used, so we requested permission to reproduce the figures and put it there for everyone to see. Make you own minds up. Is is progressive disease or relapsing?

Remember this is aiming to inform on clinical trials in humans.

As you know I am an EAE fan and am being constructive, so the young scientists can read and learn. The old farts are probably a lost cause because they won’t change and couldn’t give a rats bottom in what I say. (Except when the blade is pushed between my ribs when they are reviewing something of ours). But a change in practise is needed.

Clinical trials are doing it, why not animal studies?

EAE is a severe condition, we should be striving to limit these experiments

I want there to be use made out of pre-clinical studies, it is not about making a sausage factory of papers, it is about finding treatments and benefiting humanity.

All we are saying is that there should be transparency. It is now time to produce the raw data for the animal studies. This is especially important as about 50% of papers use the wrong statistics (which is more likely to show a positive effect) to analyse their data. In the most cases this makes no difference as the results come out similar. But not always.

If it doesn’t pass the “smack you in the eye test” then it is unlikely to be meaningful anyway.

You should demand transparency of the charities that you support and say that they should require this of the animal studies they support. Then check retrospectiviely to ensure it is done and if they don’t do it, pull the funds. Simple……let people hang themselves.

If you are an editor. Time to change!.

A simple excel sheet as supplementary data… is all that it needs. (Yep in fifty years time it will be unreadable by the computer software being used at that time).

Some journals request a data availability statement but rather than rely on peoples good will, make it compulsory. I suspect if Nature and Science required this, things will start to change and some of the star science that they publish, will turn out not to be. Indeed I use a few examples when teaching about “how to read a paper”. We can see what is good, bad and can determine whether to repeat the work or not. We can refute bad science using the data that creates the science in the first place.

It could save many animals being used and allow one to question dogey data handling instantly. Surely, it is time that the s**t sticks to the science-teflon men and women that have been peddling their dogey wares for years and years. It is often the opinion leaders that pedel bad practise so it needs a grass roots change. If you believe in your science you should not be scared of transparency. It you make it all up..time to to s**t youself, let’s shame the brown pants out of science

It’s time for change!

This is open access so read it for yourself and make your mind up.