The evolving landscape of autoantibody testing for autoimmune neurological diseases

Autoimmune neurology is a subspecialty of neurology that focuses on the interaction between the immune system and the nervous system. These autoimmune neurological disorders occur when the immune system misfires against components of the nervous system, leading to a range of diverse neurological symptoms and conditions. The immune system’s impact on the nervous system can manifest in numerous ways, affecting the central nervous system, spinal cord, peripheral nerves, and neuromuscular junction.

The challenge with autoimmune neurological disorders is that they often mimic more common underlying etiologies, making them difficult to diagnose in a timely manner. Yet prompt treatment often is critical, with significant impact on patient morbidity and mortality. The diagnosis requires clinicians to consider the clinical presentation, neurological examination, imaging and electrodiagnostic testing, as well as laboratory investigation. The diagnostic criteria for autoimmune neurological disorders frequently include identifying underlying neural-specific antibodies in patient’s serum or cerebrospinal fluid (CSF). These antibodies are critical biomarkers that help establish the diagnosis.

The field of autoimmune neurology is ever-changing. In the last several years there have been new paraneoplastic neurological syndrome (PNS) criteria (1), calls for updates to autoimmune encephalitis diagnostic criteria (2), newly discovered antibodies and corresponding tests, and a move towards phenotype-specific antibody evaluations. This article explains the current biomarkers and assays available, and how clinical laboratories should select and interpret them to ensure the best patient care possible.

Tracking a growing list of biomarkers

Accelerated discovery of new autoantibodies and their nervous system targets has led to an expanding list of clinically relevant antibody biomarkers. These biomarkers are critical as their detection can fulfill diagnostic criteria, rule out certain diagnoses, aid in the search for malignancy and provide treatment guidance.

Dozens of new antibody assays are now offered clinically — distributed across a variety of phenotype-specific antibody evaluations or profiles. One of the largest autoimmune neurology antibody panels currently available includes measurement of thirty reportable antibodies (Mayo Clinic Laboratories; Autoimmune Movement Disorder, Evaluation). Discovery of new antibodies has significantly accelerated over the last 20 years, with more than one novel antibody being discovered per year thanks to new methodologies for antigen identification, translational research efforts to identify and characterize unique staining patterns in tissue observed on indirect immunofluorescence assays (IFA), and lastly, expansion of in vitro diagnostics companies into the field of autoimmune neurology.

In addition to new antibody discoveries, researchers have elucidated the antigen identities of previously recognized tissue IFA patterns in recent years. Immunoprecipitation-mass spectrometry (IP-MS) has been used to identify several antigen targets which were previously only known by their unique staining pattern on tissue IFA. For instance, MAP1B was confirmed as the antigen target of the Purkinje cell antibody-type 2 (PCA-2) pattern, an IgG biomarker of PNS with small-cell lung carcinoma (SCLC) (3, 4).

Similarly, DACH1 was recently identified as the antigen for the anti-neuronal antibody-type 3 (ANNA-3) pattern (5). Another antigen discovery technique with good yield has been phage display immunoprecipitation sequencing (PhIP-Seq), which uses a library of phage particles presenting peptides covering the whole human proteome or targeted subsets of the proteome. In 2019, researchers reported the identification of KLHL11 antibodies using PhIP-Seq on samples from twelve male patients who presented with similar neurological features near the time of a diagnosis of seminoma (6). Since then, KLHL11 antibodies have been offered as biomarkers for paraneoplastic autoimmune encephalitis associated with testicular cancers (7). Another recent study utilized protein microarrays for novel antigen/antibody discovery in autoimmune neurological diseases. The authors demonstrated that the protein microarrays were useful for detecting antibodies targeting intracellular antigens, but also detected antibodies that target cell surface antigens that were missed by PhIP-Seq (8).

Besides offering new antibody tests, laboratories must keep up with current literature assessing the clinical relevance and diagnostic value of historically offered antibody biomarkers. Previously described antibody biomarkers may prove to have inferior clinical utility and eventually must be retired in favor of other superior biomarkers.

For example, historically, laboratories tested for striational antibodies (StrAbs) in patients with myasthenia gravis (MG). A recent retrospective study spanning 6 years of testing concluded that, despite a statistically significant paraneoplastic association, StrAbs were neither specific nor sensitive in predicting malignancy or neurologic phenotypes (9), limiting their clinical utility.

Similarly, laboratories previously included N-type voltage-gated calcium channel (VGCC) antibodies as part of PNS antibody evaluations, but evidence from multiple studies suggested that they provide limited clinical value in the context of other superior prognostic antibody biomarkers, such as SOX1 antibodies, as predictors of malignancy (10, 11). Both N-type VGCC and Striational antibodies have been removed from the disease-specific evaluations to improve the positive predictive value of these panels (9, 11).

The discovery of clinically relevant antibody biomarkers will require laboratories to develop and implement novel tests. They will also need to implement ongoing improvements in testing methodologies and evolution of disease-specific antibody evaluations. Currently, most neural-specific autoantibody tests are brought to market as laboratory-developed tests.

Appropriate utlization of antibody tests

Test misutilization continues to be a concern in the field of autoimmune neurology. This is exacerbated by the complexity and cost of this testing. Individual antibody tests are generally not available as standalone tests, with some exceptions, due to the heterogenous clinical presentations associated with a given antibody and the inability to associate a specific clinical presentation to a specific antibody in most cases. Because of this, ordering single antibody tests has low yield in most situations.

To overcome these limitations, laboratories have grouped antibodies based on associated clinical phenotypes. While ordering single antibody tests may be more cost-effective initially and provide shorter turnaround times, this practice is less efficient given the lack of a one-phenotype-one-antibody relationship. Using a single or limited testing panel will require additional testing in most cases, making this more expensive and time inefficient.

Another risk of individual orderable antibody testing is that critical antibodies will be missed due to lack of physician awareness. This is particularly problematic in the rapidly evolving field of autoimmune neurology as several new biomarkers become available each year. Rather, expert-designed antibody profiles provide testing for multiple antibodies centered around the clinical phenotype. These so-called phenotype-specific panels often include more than 10 unique antibody targets unique to a given clinical presentation.

At the same time, these larger antibody panels can be more expensive and have longer turnaround times compared with single orderable antibody test. Furthermore, the likelihood of a false positive result increases as more tests are performed, and this is exacerbated by the inclusion of targets with lower disease-specificity, such as StrAbs mentioned earlier (12). However, the need to accurately diagnose and differentiate rapidly developing diseases that are treatable, or to predict paraneoplastic malignancy, outweighs the limitations of panel testing, as early detection improves patient outcomes (13).

Most phenotype-specific neural antibody panels contain overlapping analytes due to overlapping antibody-phenotype associations (e.g., GABA-B receptor encephalitis may present with both rapidly progressing dementia as well as ataxia). Therefore, a detailed clinical evaluation with targeted panel testing is most appropriate. However, it is common for several concurrent panels with overlapping antibodies or testing to be requested with days or weeks of each other. In most cases, this redundant testing is inappropriate. Most reference laboratories will attempt to identify these potential errors, the process for detecting this misutilization is inefficient.

A recent study investigated the ordering practice for suspected autoimmune encephalitis and paraneoplastic disorders at a tertiary referral medical center (14). This study found that there was significant and unnecessary redundant panel testing ordered in ~10% of cases. Of these, there were very few instances where an additional antibody was discovered or where clinical management was positively affected, indicating that the practice has limited clinical utility. Familiarization with components of antibody testing panels is critical to eliminate redundant testing. Send-out laboratories may consider building mechanisms to detect and review redundant ordering to ensure clinical appropriateness.

Another common issue leading to poor test utilization is ordering neural antibody evaluations in patients with a low pre-test probability. Several algorithms have been developed to establish pre-test diagnosis probability and guide test ordering. The Antibody-Prevalence-in-Epilepsy (APE)/Antibody-Prevalence-in-Epilepsy-and-Encephalopathy (APE2) score can be used to predict the likelihood of neural-specific autoantibodies in autoimmune epilepsy cases (15), or in patients with cognitive dysfunction (16). The possibility of necrotizing autoimmune myopathy (NAM) can be predicted using an immune-mediated necrotizing myopathy (IMNM) calculator (17). Institutions that have implemented these tools have reported reductions in inappropriate test ordering.

Lastly, to improve the overall diagnostic accuracy of panel testing, it is recommended that both serum and CSF samples be tested simultaneously in most cases of autoimmune encephalitis. Most antibody evaluations are available in formats for both serum and CSF, with few exceptions. The sensitivity or specificity for different antibodies vary according to the specimen type (See table 1). Since, in most cases, positivity for a specific antibody cannot be predicted based on the clinical phenotype, the approach of testing both specimens ensure optimal clinical test performance.

For some antibodies, such as those targeting N-methyl-D-aspartate receptor (NMDAR) and GFAP antibodies, testing in CSF is more sensitive and specific. For others, such as LGI1 and CASPR2 antibodies, testing serum is more sensitive (18–20). When comparing the results from paired serum and CSF samples, the clinical specificity of GFAP antibody positivity for meningoencephalomyelitis was greater than 95% in CSF compared with less than 10% when positive in serum only (21). For some antibody targets, such as GABAB receptor and NMDAR, isolated positivity in serum is associated with a higher risk of false positivity. Laboratorians should take extra precautions when interpreting discordant results from different specimen types, especially when patients have atypical clinical presentations.

Choosing the most appropriate neural antibody detection approach

There are five major analytical techniques universally applied to detection of neural antibodies. These include tissue-based indirect immunofluorescence (IIF) or immunohistochemistry (IHC), recombinant cell-based assays (CBA), immunoblotting, ELISA, and radioimmunoprecipitation assays. In some cases, there are strong supportive peer-reviewed studies to support the use of specific testing methodologies for specific neural antibodies. However, for many neural antibodies, there have not been systematic studies addressing method-specific clinical performance (See table online for recommended testing methods per antibody).

Neural antibody biomarkers can target either extracellular (cell surface) or intracellular antigens. Antibodies targeting intracellular antigens are typically thought to be indirectly involved in the disease. These antibodies often have high affinity to nonconformational epitopes. Furthermore, they target single protein antigens rather than antigen complexes. These antibodies can often be detected using methods that do not maintain the antigen in its native conformation (e.g., IIF, immunoblot or ELISA). Examples of this include antibodies against the cytoplasmic protein Yo/PCA-1 or antibodies against the nuclear/cytoplasmic protein Hu/ANNA-1. These antibodies are most commonly measured utilizing IIF assays and confirmed with line immunoblot assays (LIA) or ELISA, where the antigen is at least partially denatured.

In contrast, clinically relevant antibodies against the NDMAR are directly involved in disease pathogenesis, where they bind and modulate the receptors off the cell surface. NMDAR antibodies cannot be readily detected by methods such as LIA or ELISA that utilize denatured protein subunits of the receptor. Instead, laboratories use CBAs used where the antigen is expressed recombinantly in human cells. This approach has proven to be the most sensitive and specific method for detecting NMDAR antibodies.

Testing methodologies should be considered based on the biological and physiological nature of the antibodies as well as their clinical use, such as screening vs. confirmation tests (22). Antibodies against collapsin response-mediator protein-5 (CRMP5), also called CV2, are onconeural antibodies associated with SCLC and thymoma (23, 24). A study showed that 7.5% of seropositive samples identified by immunohistochemistry on rat brain tissue were missed by two widely used commercial LIAs, but positive on CBA (25).

Additionally, clinical laboratories should not depend solely on LIA for CRMP5 antibody testing nor use these tests as screening tests. A standard procedure with screening using tissue-based IFA followed by confirmation with a specific assay is recommended for CRMP5 antibody testing in most situations (22). In general, it is not optimal to use LIAs as the sole methodology for detection of neural antibodies for the reasons outlined above for CRMP5.

Most clinically relevant neural autoantibodies are of the IgG isotype. However, depending on the antigen, the clinically relevant antibodies can favor one or more of the IgG subclasses. Generally, anti-human pan-IgG is used as the secondary antibody in most assays, but using a subclass-specific secondary antibody can increase test specificity or provide additional clinical information in some situations (26, 27). NF155 is paranodal protein that forms a protein complex with contactin-1 and CASPR1 to ensure the integrity of myelin in the nodes and paranodes. Studies have shown that detecting the IgG4 subclass of NF155 antibodies most strongly associates with suboptimal response to intravenous immunoglobulin (IVIG) therapy. NF155 IgG4 positive patients also have unique clinical and electrodiagnostic signatures, and favorable long-term outcomes when compared with NF155 cases with isolated IgM or antibodies of other IgG subtypes (28).

Laboratory standardization and quality assurance

Most testing for antibodies related to autoimmune neurology occurs in a handful of large national reference laboratories and occasionally smaller specialty laboratories. The methods used, the composition of antibodies included in various panels, the testing algorithms followed, the provided interpretations, and the degree of consultation available all vary dramatically across laboratories. Therefore, although each of these laboratories offer an “autoimmune encephalopathy” antibody evaluation, there are significant differences between them.

Understanding the differences across laboratories has become more difficult in the last several years as the number of available antibody biomarkers increased, as well as the number of laboratories offering these antibody panels. Several studies have dealt with this problem.

To assess cross-laboratory agreement in neural antibody testing in Italian laboratories, an external quality assessment program was organized in 2018 (29). This study found partial or large interlaboratory disagreement across commercial and in-house tests for detecting antibodies against AQP4, MOG, and gangliosides. Even for laboratories using the same commercial kits, such as a commercial ELISA for myelin associated glycoprotein (MAG) antibodies, the researchers detected significant differences in results, indicating cross-laboratory difference in either performance or interpretation of the test.

Another international multicenter study of testing methodologies for MOG antibody suggested that live CBA showed excellent agreement (96%) across testing sites for high titer and negative samples, followed by fixed-CBA (90%), while ELISA showed no concordance with CBAs for detecting MOG-IgG (30). However, the agreement for borderline positive and negative samples were suboptimal across similar CBA assays (30). These studies highlight the need to improve the standardization of neural antibody testing across laboratories.

Part of the problem relates to the fact that these antibodies are rare, and given the limited number of laboratories performing the testing, there have not been standard reference material or external proficiency testing (PT) programs that readily allow laboratories to compare to their peers. To fulfill regulatory requirements, most laboratories resort to alternative assessment of performance (AAP) with internal blinded testing. Much of the autoantibody testing performed for autoimmune neurological disorders still depends on manual and subjective interpretation of indirect tissue IFA and microscopy-based CBAs. These assays are thus highly dependent on the experience of the individual interpreting the results under the microscope.

To further evolve and improve the field, there is a need to establish expert-led consensus criteria for the interpretation of tissue IFA and CBAs. This includes best practices for the use of testing algorithms, results interpretation, development of shared reference materials, movement away from subjective testing platforms where possible, and further effort to define the differences in testing methodologies. 

Jack L. Wu, PhD, MLS(ASCP)CM, is a senior developer for the Translational Research, Innovation and Test Development Office (TRITDO) and the Neuroimmunology Laboratory in the Department of Laboratory Medicine and Pathology at Mayo Clinic, Rochester, Minnesota. +Email: [email protected]

John R. Mills, PhD, DABCC, is a director of the Neuroimmunology Laboratory and an assistant professor in the Department of Laboratory Medicine and Pathology at Mayo Clinic, Rochester, Minnesota. +Email: [email protected]

References

1. Graus F, Vogrig A, Muñiz-Castrillo S, et al. Updated diagnostic criteria for paraneoplastic neurologic syndromes. Neurol Neuroimmunol Neuroinflamm 2021; doi: 10.1212/NXI.0000000000001014.
2. Orozco E, Valencia-Sanchez C, Britton J, et al. Autoimmune encephalitis criteria in clinical practice. Neurol Clin Pract 2023; doi: 10.1212/CPJ.0000000000200151.
3. Gadoth A, Kryzer TJ, Fryer J, et al. Microtubule associated protein (MAP) 1B: Antigen of PCA-2 IgG, biomarker of small-cell lung carcinoma-related paraneoplastic neurological autoimmunity (S41.007). Neurology 2017; doi: 10.1212/WNL.88.16_supplement.S41.007.
4. Gadoth A, Kryzer TJ, Fryer J, et al. Microtubule-associated protein 1B: Novel paraneoplastic biomarker. Ann Neurol 2017; doi: 10.1002/ana.24872.
5. Zekeridou A, Yang B, Lennon VA, et al. Anti-neuronal nuclear antibody 3 autoimmunity targets Dachshund homolog 1. Ann Neurol 2022; doi: 10.1002/ana.26320.
6. Mandel-Brehm C, Dubey D, Kryzer TJ, et al. Kelch-like protein 11 antibodies in seminoma-associated paraneoplastic encephalitis. N Engl J Med 2019; doi: 10.1056/NEJMoa1816721.
7. Maudes E, Landa J, Muñoz-Lopetegi A, et al. Clinical significance of Kelch-like protein 11 antibodies. Neurol Neuroimmunol Neuroinflamm 2020; doi: 10.1212/NXI.0000000000000666.
8. McKeon A, Lesnick C, Vorasoot N, et al. Utility of protein microarrays for detection of classified and novel antibodies in autoimmune neurologic disease. Neurol Neuroimmunol Neuroinflamm 2023; doi: 10.1212/NXI.0000000000200145.
9. Shelly S, Mills JR, Dubey D, et al. Clinical utility of striational antibodies in paraneoplastic and myasthenia gravis paraneoplastic panels. Neurology 2021; doi: 10.1212/WNL.0000000000012050.
10. Titulaer MJ, Klooster R, Potman M, et al. SOX antibodies in small-cell lung cancer and Lambert-Eaton myasthenic syndrome: Frequency and relation with survival. J Clin Oncol 2009; doi: 10.1200/JCO.2008.20.6169.
11. Majed M, Zekeridou A, Lennon V, et al. Testing for N-type voltage gated calcium channel antibody has limited utility in evaluating patients with suspected Lambert-Eaton myasthenic syndrome. Neurology 2022; doi: 10.1212/01.wnl.0000903272.62514.70.
12. Ebright MJ, Li S-H, Reynolds E, et al. Unintended consequences of Mayo paraneoplastic evaluations. Neurology 2018; doi: 10.1212/WNL.0000000000006577.
13. Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurol 2016; doi: 10.1016/S1474-4422(15)00401-9.
14. Fredrich SE, Vernino S, and Blackburn KM. Antibody testing for neurological autoimmune disorders: Evaluation of best practices at a tertiary referral center. Front Neurol 2021; doi: 10.3389/fneur.2021.690415.
15. Dubey D, Singh J, Britton JW, et al. Predictive models in the diagnosis and treatment of autoimmune epilepsy. Epilepsia 2017; doi: 10.1111/epi.13797.
16. Dubey D, Kothapalli N, McKeon A, et al. Predictors of neural-specific autoantibodies and immunotherapy response in patients with cognitive dysfunction. J Neuroimmunol 2018; doi: 10.1016/j.jneuroim.2018.07.009.
17. Triplett JD, Shelly S, Livne G, et al. Diagnostic modelling and therapeutic monitoring of immune-mediated necrotizing myopathy: Role of electrical myotonia. Brain Commun 2020; doi: 10.1093/braincomms/fcaa191.
18. Gresa-Arribas N, Titulaer MJ, Torrents A, et al. Antibody titres at diagnosis and during follow-up of anti-NMDA receptor encephalitis: A retrospective study. Lancet Neurol 2014; doi: 10.1016/S1474-4422(13)70282-5.
19. Gadoth A, Pittock SJ, Dubey D, et al. Expanded phenotypes and outcomes among 256 LGI1/CASPR2-IgG-positive patients. Ann Neurol 2017; doi: 10.1002/ana.24979.
20. Flanagan EP, Hinson SR, Lennon VA, et al. Glial fibrillary acidic protein immunoglobulin G as biomarker of autoimmune astrocytopathy: Analysis of 102 patients. Ann Neurol 2017; doi: 10.1002/ana.24881.
21. Flanagan EP, Dubey D, Hinson S, et al. Specificity of glial fibrillary acidic protein IgG autoantibody (GFAP-IgG) for autoimmune meningoencephalomyelitis diagnosis (S20.002). Neurology 2017; doi: 10.1212/WNL.88.16_supplement.S20.00.
22. Budhram A, Dubey D, Sechi E, et al. Neural antibody testing in patients with suspected autoimmune encephalitis. Clin Chem 2020; doi: 10.1093/clinchem/hvaa254.
23. Honnorat J, Antoine JC, Derrington E, et al. Antibodies to a subpopulation of glial cells and a 66 kDa developmental protein in patients with paraneoplastic neurological syndromes. J Neurol Neurosurg Psychiatry 1996; doi: 10.1136/jnnp.61.3.270.
24. Yu Z, Kryzer TJ, Griesmann GE, et al. CRMP-5 neuronal autoantibody: Marker of lung cancer and thymoma-related autoimmunity. Ann Neurol 2001;49:146-54.
25. Sabater L, Saiz A, Dalmau J, et al. Pitfalls in the detection of CV2 (CRMP5) antibodies. J Neuroimmunol 2016; doi: 10.1016/j.jneuroim.2015.11.009.
26. Waters P, Woodhall M, O’Connor KC, et al. MOG cell-based assay detects non-MS patients with inflammatory neurologic disease. Neurol Neuroimmunol Neuroinflamm 2015; doi: 10.1212/NXI.0000000000000089.
27. Kadoya M, Kaida K, Koike H, et al. IgG4 anti-neurofascin155 antibodies in chronic inflammatory demyelinating polyradiculoneuropathy: Clinical significance and diagnostic utility of a conventional assay. J Neuroimmunol 2016; doi: 10.1016/j.jneuroim.2016.10.013.
28. Shelly S, Klein CJ, Dyck PJB, et al. Neurofascin-155 immunoglobulin subtypes: Clinicopathologic associations and neurologic outcomes. Neurology 2021; doi: 10.1212/WNL.0000000000012932.
29. Gastaldi M, Zardini E, Scaranzin S, et al. Autoantibody diagnostics in neuroimmunology: Experience from the 2018 Italian neuroimmunology association external quality assessment program. Front Neurol 2019; doi: 10.3389/fneur.2019.01385.
30. Reindl M, Schanda K, Woodhall M, et al. International multicenter examination of MOG antibody assays. Neurol Neuroimmunol Neuroinflamm 2020; doi: 10.1212/NXI.0000000000000674.

View the full CLN July/August 2024 issue.