Neuroimmunology: Updates and antibody test utilization

Hot Topic

Expires: May 24, 2024



Andrew McKeon, M.B., B.Ch., M.D.
Director, Neuroimmunology Laboratory
Professor of Laboratory Medicine and Pathology,
Division of Clinical Biochemistry
Department of Laboratory Medicine and Pathology 
at Mayo Clinic, Rochester, Minnesota



Welcome to Hot Topic by Mayo Clinic Laboratories. My name is Dr. Andrew McKeon. Today I'm going to do a presentation on updates to antibody testing in the Neuroimmunology Laboratory and a few comments about antibody test utilization. 

I'm one of the directors at the Neuroimmunology Laboratory and I also work as a neurologist in the Department of Neurology at Mayo Clinic. 


Here are some disclosures regarding patents for various autoantibodies that we've discovered at Mayo Clinic, some of which are implemented in our test profiling already and some of which are to come at a later date. 


An overview of today's talk is a review of neurological phenotype-based evaluations and why over time we have determined that this is a better approach than a more scatter gun approach to autoantibody testing that is currently available through the paraneoplastic evaluation. To review some of the coming changes to these test profiles and to review from a clinical practice perspective. When is it appropriate to order serum? When is it appropriate to order CSF, and when is it appropriate order both?


So, by way of background, autoantibody testing for autoimmune neurological diseases such as encephalopathies, movement disorders, and myelopathies has become increasingly complex over time. That is because of the number of autoantibodies that have been discovered that are biomarkers for these diseases. The complexity has also been added to by the varying disease specificity. So for some phenotypes, there are particular autoantibodies that are important. For other phenotypes, it's different antibodies again, and there is some overlap of autoantibody specificity between different phenotypes. The specimens used are also important. So for some evaluations a serum suffices for testing; for others, CSF might be the main thing to test for, plus serum is generally appropriate in those situations also. Then the assay types vary. So not all antibodies that are reported are the same in terms of their clinical utility. Growing tests volumes is also an issue, and so over time, we've had to try to streamline our processes in order to bring about the best care possible in terms of turnaround and in terms of the test utilization. And then also the process of verification of autoantibodies over time has changed. And so there are much stricter guidelines now for what kinds of autoantibodies can be offered for clinical use. And then also maintenance of quality within the laboratory. So all of these factors have made testing more complex, but also clinical care more complex. 

So broadly speaking, we're still dealing with two main groups of antibodies. These are all potential biomarkers of disease, but some of these are disease effectors such as NMDA receptor autoantibodies that can have effects on the receptors themselves, down regulating these. But others are simply just biomarkers of disease likely reflecting a T cell-mediated process. So an example of a T cell effector-related antibody will be PCA-1 or anti-Yo. But over time, the number of antibodies in each of these groups has increased. 

Neuronal nuclear/cytoplasmic Abs

Taking the last group first, these neuronal nuclear or cytoplasmic antibody biomarkers that aren't pathogenic but reflective of a paraneoplastic T-cell mediated process. As you can see, this has increased over time, so people in medical school would have learned about Hu, Ri, and Yo. But look at just the extent of autoantibodies that are now available in this space. Most of these have paraneoplastic significance. I've highlighted some of the exceptions in that regard such as GAD65 antibody and adaptor protein AP3B2. And then there are also some circumstances where patients are more immunotherapy responsive. So in general, we don't think of these disorders as being particularly immunotherapy responsive, but GFAP IgG antibody is an exception in that regard. 

Plasma membrane protein (pathogenic) Abs

Overall, these disorders have a worse neurological prognosis compared to the plasma membrane directed antibodies that I've listed here. In contrast to the last group, the paraneoplastic significance varies and I've highlighted the ones in green that have a stronger cancer association, such as for NMDA receptor antibody and ovarian teratoma AMPA receptor and GABA-B receptor antibodies. And then also to antibodies generally associated Hodgkin lymphoma. For a lot of these other antibodies that I've indicated here, there's a low predictive value for cancer overall. Often the cancer types can be diverse. And in fact, in the literature sometimes there's controversy as to whether there is actual paraneoplastic significance at all. But again, you can see between the two lists that the profile of antibodies has generally increased over time. Some of these have quite distinct neurological associations such as NMDA receptor encephalitis. And this will contrast with some of the multifocal type paraneoplastic associated antibodies such that we saw on the previous slide. So overall, because of this complexity, we think at this stage that it is not prudent to include these all in one evaluation. But more so to base it on the neurological phenotype that the patient presents with. 

Number of Abs increasing – Movement disorders

So as you can see, say for movement disorders, the number of autoantibodies here has quite dramatically increased over time as to what is relevant. Most of these are in clinical service already, but some of them are to come, yet. For example, if we took PCA-1 antibody, this is largely associated with ataxia. But something like amphiphysin antibody could be associated with ataxia, or stiff person syndrome, or a myelopathy or neuropathy. So amphiphysin antibody could end up in multiple different evaluations. It could either be pertinent to myelopathy or to movement disorders or to peripheral neuropathy. 

Neurologic phenotype-specific evaluations

The neurological phenotype specific evaluations are made up from the autoantibodies that I demonstrated to you and some others that I didn't include here. But really these are based around what is the phenotype of the patient that's presenting in front of us. What autoantibodies will be pertinent in this circumstance? This could be quite a restricted number of antibodies, say for retinopathy or for CNS demyelinating disease, we're just talking about a couple of biomarkers in serum. So for retinopathy, this could be Recoverin antibody and CRMP-5 antibody for CNS demyelinating disease. This will be aquaporin 4 MOG antibody, just in serum only. In contrast for an automated encephalopathy, both serum and CSF testing will be recommended. And there will be a vast array of analytes or antibodies that will be appropriate to test for in that circumstance. For example, LGI1 antibody will be readily detected in serum and also very specifically detected in serum. In contrast, NMDA receptor antibody and GFAP antibodies would be more sensitively and specifically detected in spinal fluid. So in order to be comprehensive in that situation, we recommend testing both serum and spinal fluid. 

When to consider?

People often ask us, when should we be considering these? Am I at risk for missing these diagnoses? And our response to that is to really think about the clinical history: How the patient is presenting? What’s on the neurological exam?  In almost all circumstances, these disorders develop subacutely. So over days to weeks. These aren't hyper-acute disorders like stroke. These aren't chronic progressive disorders like dementia. There's a rapid progression in these symptoms over time. There sometimes can be a fluctuating course, but people have significant loss of functionality over a course of weeks. Sometimes, but not always, there can be an autoimmune history in the background. So a person might have thyroid disease as a clue. Sometimes, but not always, there could be a cancer history preexisting of the patient as well, or there could be strong family histories of these disorders. 

Differential diagnosis

And then the differential diagnosis should be considered as well. So for patients presenting with the subacute onset neurological problems, we need to think about all of these other things as well. A rapidly progressive degenerative disorder like CJD, primary neurological cancer, the effects of medications or toxins, potentially infection, vitamin deficiencies, other inflammatory disorders for which there are no antibody biomarkers such as multiple sclerosis, sarcoidosis, or this rare CLIPPERS disorder where patients get a lot of inflammation in the brain stem and pons in particular. Then systemic autoimmune diseases, metabolic disorders, endocrinopathies, such as hypothyroidism or hyperthyroidism, psychiatric disorders. And then there are other tests that should be considered in these situations as well as subacute onset disorders. So cerebral spinal fluid testing is also very important in general for these disorders that I've just mentioned, but also for autoimmune neurological disease. Up to 50 percent of patients with autoimmune encephalitis are antibody negative. And thus, testing for the CSF or looking on MRI for clues of autoimmunity are important in those patients as well. So as well as sending autoantibody testing or testing for infection, once you think about testing, IgG index synthesis rate and oligoclonal bands, and kappa free light chains can be a surrogate for oligoclonal bands testing. 

Paraneoplastic evaluation?

What about the paraneoplastic evaluation? Well, this was an early workhorse diagnostic test for paraneoplastic disorders when there were limited numbers of antibodies available. These patients generally had diverse phenotypes and the antibodies therein were very classical paraneoplastic antibodies. But over time there have been more biomarkers made available, some of which have been added to the paraneoplastic evaluation, which have more limited cancer associations, have a lower specificity for neurological phenotype. And in reality, over time in terms of how the science has proceeded and how tests have been developed, there are many, many more biomarkers available. Many have neurological phenotype specificity. Many of these have limited or no cancer association. So as a result, we've spent a number of years now building out neurological phenotype-specific evaluations, and it has now gotten to the point where we have the whole neuraxis covered, and so the paraneoplastic evaluation really has become redundant.  At this stage we recommend not to order this evaluation. Instead we recommend to order a neurological phenotype-specific evaluation. And so whether that's retinopathy, demyelinating disease presentation, encephalopathy, epilepsy or dementia, myelopathy, neuropathy, or a movement disorder, we have it covered. 

What will I get with neurological phenotype-specific evaluations?

So what will I get with these neurological phenotype-specific evaluations? Well, all the clinically validated IgG biomarker antibodies for that phenotype. That will include paraneoplastic and the other autoimmune biomarkers. So it covers patients who have multifocal neurological phenotypes. So for example, if you have a patient who presents with an encephalomyelopathy and you're wondering could this patient have ANA1 antibody which has been associated with that crossover phenotype, that antibody will be found in both evaluations. So in no circumstance is it really appropriate to order two evaluations on the same specimen type. What we do recommend is that people would order both specimen types where appropriate, such as for encephalopathy. And we only make one specimen type available where appropriate for some disorders such as CNS demyelinating evaluation. Also, we should point out that these evaluations will then be updated over time as new biomarkers become available, such as the Kelch-11 antibody in seminoma in patients with brainstem encephalitis and limbic encephalitis. So that antibody will be incorporated over time into our encephalopathy and movement evaluations. But the paraneoplastic evaluation is not being updated with any of these new biomarkers and hasn't been for some time. So this evaluation is now not only nonspecific at times, but it's also insensitive. 

What should I order?

What should I order in summary? Order a neurological phenotype-specific evaluation. Order a maximum of one serum and one CSF. 

Serum versus CSF?

So as mentioned earlier, there are some evaluations that are really only pertinent in serum, and these are listed here. So retinopathy, CNS demyelinating disease, but also some of our longstanding myasthenia gravis/Lambert-Eaton syndrome tests, peripheral neuropathy, dysautonomia, GI dysmotility, and necrotizing myopathies. 

There are others where we recommend both serum and CSF to be tested. Encephalopathy, epilepsy, dementia, movement disorders, myelopathy, stiff person spectrum, and then also pediatric CNS disorders. This is in order to make sure that we capture the full range of antibodies in terms of their analytical sensitivity and specificity in the different specimen types. As mentioned earlier, you can have different antibodies that are more readily detected in different specimen types. So for example, NMDA receptor antibody detected more readily in CSF. And then in serum, LGI1 more readily detected. 

What about patients with known cancers?

So what about patients with known cancers? So an example of this would be a patient who is attending an oncologist and has small cell lung cancer and is being treated for small cell lung cancer and then develops a rapidly progressive myelopathy or neuropathy. The question that arises is, is this due to the treatment for the cancer? Or could the patient have developed a paraneoplastic neurological disorder? So in order to cater for these patients, while we are eliminating the paraneoplastic evaluation, we are also bringing out smaller cancer-specific evaluations that will contain autoantibodies specific for that cancer type. But this is really only relevant to oncology practice and not to neurological practice. 

Changes to evaluations: VGKC, N, PQ, Alpha-3 AChR

As well as the paraneoplastic evaluation, we also have some changes that we want to make to the phenotype-specific evaluations. And that really pertains to antibodies that are detected by this immunoprecipitation method: VGKC, N-type calcium channel antibody, P/Q-type calcium channel antibody, and the Alpha 3 AChR antibody. While these are very sensitive screening tests for these particular antibodies, they also have low specificity as we've published on each of these different antibodies. In contrast, there are higher specificity tests for some of these antibodies available at this point. We also have the LGI1 and CASPR2 antibodies now available that really have made VGKC complex immunoprecipitation assays redundant at this point. 

Abs with lower “general use” specificity (immunoprecipitation or ELISA detected)

There are a number of autoantibodies that are tested for by immunoprecipitation or ELISA that were initially very good in terms of their specificity for particular disease types, but over time, either have become redundant, which I will explain later on and will be obsoleted, or had been put into general use as neurology disease antibodies but were less specific outside of the initial intended phenotype and have become the subject of controversy. So there are five changes that I've listed on this slide that account for these changes. First of all, the muscle acetylcholine receptor binding antibody. This had become an antibody in general use in our paraneoplastic evaluation and some other evaluations. There is good specificity of this autoantibody for myasthenia gravis only and so we're retaining that particular antibody for that testing, but not for other phenotypes. The Alpha-3 ganglionic acetylcholine receptor autoantibody similarly has very good specificity for patients with confirmed dysautonomia and so we think that this is a reasonable test to continue to use in our dysautonomia-relevant evaluations. So this would be the autoimmune dysautonomia and GI dysmotility evaluations. The P/Q-type calcium channel antibody has good specificity in Lambert-Eaton syndrome and has moderate specificity among patients with autoimmune ataxia. Although this is detected by immunoprecipitation assay and can have some specificity issues, in ataxia this is really still the best test available for this particular P/Q-type calcium channel antibody. As a result, we are continuing to offer this antibody in the Lambert-Eaton or myasthenic-related evaluations. And we will continue to offer this in the movement disorder evaluation. Overall N-type calcium channel antibody has been found over time not to be additive in terms of specificity for cancer or for neurological phenotype similarly with striational antibody. Both of these are being eliminated from our evaluations altogether. So these tests will be considered obsoleted. 


So in summary, we recommend neurological phenotype-based testing. We recommend testing one serum and spinal fluid. The paraneoplastic evaluation will thus be obsoleted, as this is not neurological phenotype specific. Antibodies with limited specificity beyond their original use, such as the P/Q-type calcium channel antibody and Alpha-three ganglionic acetylcholine receptor antibody, are being eliminated from certain evaluations but are being maintained in the phenotype-relevant evaluations. Antibodies with limited specificity, N-type calcium channel antibody and striational, are being obsoleted altogether. For some of the phenotype evaluations, serum will continue to suffice, such as the CNS demyelinating evaluation, aquaporin 4, and MOG antibody. But for others, serum and CSF will be needed such as autoimmune encephalopathy evaluation. 

Thank you

Thank you very much for your attention. 


Contact us:

Photo credit

Shutterstock / vitstudio

MCL Education (@mmledu)

MCL Education

This post was developed by our Education and Technical Publications Team.