Dodgy EAE experiments

Genetic Background Can Result in a Marked or Minimal Effect of Gene Knockout (GPR55 and CB2 Receptor) in Experimental Autoimmune Encephalomyelitis Models of Multiple Sclerosis Sofia Sisay, Gareth Pryce, Samuel J. Jackson, Carolyn Tanner, Ruth A. Ross, Gregory J. Michael, David L. Selwood, Gavin Giovannoni, David BakerResearch Article | published 09 Oct 2013 | PLOS ONE10.1371/journal.pone.0076907

Endocannabinoids and some phytocannabinoids bind to CB1 and CB2 cannabinoid receptors, transient receptor potential vanilloid one (TRPV1) receptor and the orphan G protein receptor fifty-five (GPR55). Studies using C57BL/10 and C57BL/6 (Cnr2tm1Zim) CB2 cannabinoid receptor knockout mice have demonstrated an immune-augmenting effect in experimental autoimmune encephalomyelitis (EAE) models of multiple sclerosis. However, other EAE studies in Biozzi ABH mice often failed to show any treatment effect of either CB2 receptor agonism or antagonism on inhibition of T cell autoimmunity. The influence of genetic background on the induction of EAE in endocannabinoid system-related gene knockout mice was examined. It was found that C57BL/6.GPR55 knockout mice developed less severe disease, notably in female mice, following active induction with myelin oligodendrocyte glycoprotein 35-55 peptide. In contrast C57BL/6.CB2 (Cnr2Dgen) receptor knockout mice developed augmented severity of disease consistent with the genetically and pharmacologically-distinct, Cnr2tm1Zim mice. However, when the knockout gene was bred into the ABH mouse background and EAE induced with spinal cord autoantigens the immune-enhancing effect of CB2 receptor deletion was lost. Likewise CB1 receptor and transient receptor potential vanilloid one knockout mice on the ABH background demonstrated no alteration in immune-susceptibility, in terms of disease incidence and severity of EAE, in contrast to that reported in some C57BL/6 mouse studies. Furthermore the immune-modulating influence of GPR55 was marginal on the ABH mouse background. Whilst sedative doses of tetrahydrocannabinol could induce immunosuppression, this was associated with a CB1 receptor rather than a CB2 receptor-mediated effect. These data support the fact that non-psychoactive doses of medicinal cannabis have a marginal influence on the immune response in MS. Importantly, it adds a note of caution for the translational value of some transgenic/gene knockout and other studies on low-EAE susceptibility backgrounds with inconsistent disease course and susceptibility.

This study looks at the effect of gene knockdown in two different strains of mice C57BL/6 which is about the most EAE resistant strain there is, but it is the background that most gene knockouts are made on and the Biozzi ABH strain, which is the mouse of choice of Team G. 

What this says is that there can be some immune effects noted in mice that lack certain cannabinoid-related genes up in one case (CB2 knockout) and down in another case (GPR55 knockout) on the low susceptibility C57BL/6. When you have robust EAE as occurs in the ABH mouse system most of these effects disappear or become minimal and the studies in ABH help reinterpret much of the previously published studies of cannabinoid knockout C57BL/6 mice. 

So who really cares, interesting scientifically but what is the relevance to MS. We used our own data to show that you can get whatever you like in the C57BL/6 system, if you have no quality control in the system and suggests that some drug therapy studies on which you pin your future hopes, may be a bust. Indeed we could list a whole series of agents that do not do much in our hands but have been proposed to be wonder cures. 

So we published data that to my mind really belonged in the rubbish bin and had been there for a number of years until I was prodded into action and published this to make a point. This is because this is the type of rubbish that is found acceptable and published in good journals week in week out. As long as it shows the right answer in it goes. Some people really need to take their refereeing responsibilities seriously and start to question this, which is an important message. This is because next thing you know the media gets hold of it and starts to create false hopes. 

Unfortunately it is not just the bad science merchants that are the culprits. The referees made us not name names with examples, but it is not surprising that people new to the field follow these bad examples.

No wonder many studies in animals fail to translate into human treatment. This is because they would fail to translate even into another mouse strain. It is feasible therefore that many of the published studies are largely artefact. So clinical trials are planned based on data that will not stand the test of time. This is why replication is important.

Next time you read a paper where the gene knockout makes the disease worse look where the control group is. Is it low. Now compare the control group where the the gene knockout makes things better, is the control group high. They should be in the same place if there is quality control in the system, it does not take a rocket scientist to spot this, but if it proves the mechanism then it is being lapped up.  Don’t get me wrong there is some great science in C57BL/6 EAE, we need to improve standards by the EAE fraternity.

This paper shows that not all knockouts are created equal and maybe it is time to reinvestigate some studies. 

CoI This work from Team G

About the author



  • "this is the type of rubbish that is published in good journals week in and week out" sounds like inept peer review. If reputable journals publish non-reproducible experiments their credibility will be shot. Also, any data from EAE knockouts have an intrinsic skepticism when translating to human MS for obvious reasons.

    • Yes inept peer review and something that needs to be addressed,

      "But If reputable journals publish non-reproducible experiments their credibility will be shot….
      well not really!"…..Look at the Journal Nature for example

      It has been suggested that about 75% of the experimental stuff published there is not reproducible (I will post on this latter with source material I am not liabling). Given the fact until recently that there were no methods..this is hardly surprising.

      Our Nature paper was of course reproducible as is our stuff in the Journal of Lowly stuff…I am proud of that . We may not always be first but usually we are right!.

      Knockouts have skepticism for obvious reasons…please explain your stance.

  • EAE knockouts have given discrepant results (TNF knockout in EAE models have ameliorated disease in strain specific mice) but when translating to human MS the limitations of EAE become apparent. Lack of genetic diversity in strain specific mice does not reflect the "real MS". In terms of screening for potential targets, especially immunogenic, has lead to partially effective therapies.

    • The knockouts gave some of the results as with antagonists used in wild types, so it is not necessarily the gene knockout but the model…however with TNF neutralization the results reported can vary depending on context and wonder if we are comparing apples with oranges?

      Early EAE verses late MS. With rolipram a chemical believed to inhibit TNF either does nothing on inhibits early EAE but makes symptoms in chronic animals worse.. It makes MS worse too.

By MouseDoctor



Recent Posts

Recent Comments