March 2012: Volume 38, Number 3
Diagnostic Utility for Diagnosis of Encephalitis, Paraneoplastic Syndromes, and Ataxia
By Marvin J. Fritzler, MD, PhD
Autoantibodies are key biomarkers in establishing the diagnosis and prognosis of a variety of autoimmune conditions, including several neurological conditions such as encephalitis, paraneoplastic syndromes (PNS), and ataxia. While these three conditions represent only a part of the spectrum of neurological conditions associated with autoimmunity, their diagnosis can be difficult due to clinical and serological overlap.
Recently, a number of excellent reviews of autoantibodies in neurological diseases have been published (1–10). This article describes the most salient laboratory findings for encephalitis, PNS, and ataxia. The information presented here generally refers to IgG antibodies because, for the most part, the importance of other autoantibody isotypes (IgM, IgA, IgD, IgE) or subclasses (IgG3, IgG4) is not clear.
Encephalitis. An inflammation of brain tissue, encephalitis presents with a spectrum of signs and symptoms, including fever, headache, fatigue, paralysis, visual disturbances, seizures, and loss of consciousness. Other features of limbic encephalitis include sleep and mood disturbances, memory loss, and hallucinations. Many cases of encephalitis can be traced to infectious or drug-induced etiologies, but increasingly researchers have found that autoantibodies are involved.
PNS. Patients with PNS have clinical or sub-clinical malignancies, but the symptoms of the disease are not the result of direct tumor invasion or metastases. PNS features a wide spectrum of neurological deficits, such as cerebellar ataxia, personality disorders with memory loss, multi-focal cerebral dysfunction, and more. Occasionally, PNS autoantibodies are referred to as onconeural antibodies (ONA).
Ataxia. Ataxia describes disturbances of motor coordination, often manifest as a movement disorder and loss of balance. Ataxias are associated with a variety of genetic, infectious, toxic, and metabolic conditions. Those related to autoimmune diseases fall into at least two broad categories based on the localization of the nervous system disorder: cerebellar and sensory ataxia. Patients with paraneoplastic cerebellar degeneration, paraneoplastic encephalomyelitis, multiple sclerosis, celiac disease, and cerebellar ataxia with polyendocrine autoimmunity can be characterized by the presence of circulating autoantibodies.
Detection of Autoantibodies
Among the remarkable advances in neurology in the last 2–3 decades has been identification of highly specific autoantibodies and their molecular targets. A variety of methods can detect autoantibodies to these various antigens. Traditionally, many laboratories relied on indirect immunofluorescence (IIF) using neural tissues such as hippocampus and cerebellum in which distinctive patterns of staining provided clues to the identity of the autoantibody. In some cases, the staining is so distinctive that additional testing may not be required. For others, however, the IIF patterns are not specific and laboratories must perform other specialized techniques to confirm the antibody specificity.
In some cases, the use of another tissue can help differentiate autoantibody systems that have similar staining patterns of neural tissues. Laboratories also use other techniques, such as immunoprecipitation (IP) of metabolically radiolabeled cellular proteins or IP of radiolabeled proteins produced by in vitro transcription and translation (TnT-IP) of the cognate cDNA. In addition, radioimmunoprecipitation assays (RIPA) that depend on IP of the antigen target that is tightly bound by various radiolabeled toxins, facilitates detection of some cell surface antibodies.
Because current evidence indicates that the majority of reactive epitopes are conformational, other techniques such as western blotting (WB) and line immunoassays (LIA) may be less reliable for some autoantibody systems. To improve detection, researchers have developed cell-based assays (CBA) that use tissue culture cells transfected with the appropriate cDNA. These cells over-express the cognate protein, which can be detected by conventional IIF. Researchers also have developed a number of other sophisticated techniques such as addressable laser bead immunoassays, but their use tends to be restricted to highly specialized laboratories.
Some key concepts apply to virtually all of the autoantibodies described in this review. Autoantibodies titers are not always correlated with clinical prognosis or disease course, but as a rule-of-thumb, high titer antibodies tend to be more specific for the disease. Studies also show that testing cerebral spinal fluid (CSF) adds little value to positive serum tests; however, patients with negative serum results sometimes have positive CSF results. Such findings require clinical acumen to decipher.
Cell Surface/Synaptic Targets
Autoantibodies in encephalitis, PNS, and ataxia target antigens that are cell surface receptors and proteins that have critical roles in synaptic transmission. Described here are the more prominent target antigens and their clinical associations.
AMPAR. AMPAR (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor) GluR1/2 autoantibodies are typically associated with limbic encephalitis, which is also found in other PNS, including a closely related syndrome associated with anti-leucine-rich glioma-inactivated (LGI1) antibodies described below. GluR3 antibodies are typically associated with Rasmussen’s encephalitis, which appears to be primarily T-cell mediated.
Caspr2. Caspr2 (contactin-associated protein-like 2) autoantibodies are associated with encephalitis and peripheral nerve hyperexcitability. Clinical symptoms tend to be more diverse than those associated with classical limbic encephalitis due to the frequent co-occurrence of other autoantibodies. Patients display complex disorders, manifesting as motor weakness, atrophy, fasciculation, and bulbar symptoms, all of which are attributed to motor neuron disorder. An early and accurate diagnosis is important because patients respond well to immunotherapy. Although these antibodies have also been linked to acquired neuromyotonia, many of these patients do not have Caspr2 autoantibodies.
GABABR. GABABR (γ-aminobutyric acid-B receptor) autoantibodies are associated with a syndrome of limbic encephalitis that may differ from that seen with other autoantibodies (Table 1). Approximately 50% of antibody-positive patients have tumors, primarily small cell lung carcinoma (SCLC). Researchers have suggested that GABABR autoimmunity accounts for most of the SCLC–associated limbic encephalitis. This antibody also should be suspected in patients with paraneoplastic limbic encephalitis previously attributed to GAD65 antibodies. Although autoantibodies to GABABR inhibit the cognate receptor function, unlike AMPAR autoantibodies, they do not cause receptor internalization.
Click here for Table 1
LGI1. LGI1 (leucine-rich glioma-inactivated 1) autoantibodies interact with two epilepsy-related proteins, presynaptic ADAM23 and postsynaptic ADAM22, organizing a trans-synaptic macromolecular complex that includes presynaptic potassium channels and postsynaptic AMPARs. Disruption of LGI1 function by these antibodies causes increased excitability of inhibitory neurons, explaining the clinical features of seizures and limbic encephalopathy. Considering that seizures may precede other symptoms of limbic dysfunction, an early and accurate diagnosis followed by immunotherapy has significant clinical implications. Some patients present with rapidly progressive dementia and/or myoclonic-like movements that resembles Creutzfeldt-Jakob disease.
NMDAR. NMDAR (N-methyl-D-aspartate receptors) belongs to a family of ligand-controlled transmembrane cationic channels composed of various protein subunits that include NR1 and NR2A-D. Patients with limbic encephalitis have autoantibodies that target the extracellular domain of the NR1 subunit, whereas the autoantibodies in neuropsychiatric systemic lupus erythematosus (SLE) predominantly bind the NR2 subunit of NMDAR. These autoantibodies initiate cross-linking and internalization of receptors and cause decreased current recordings in hippocampal neurons. Pathological lesions are characterized by B- and plasma-cell infiltrates, a relative paucity of T-cells, IgG deposits that are devoid of complement components, and decreased levels of NMDAR.
Autoantibodies to the NR1 subunit are a highly specific marker of anti-NMDA receptor encephalitis. This form of encephalitis is one of the most common types of autoimmune encephalitis, typically affecting young women with ovarian teratomas; however, it also occurs in women without tumors, as well as in men and children. More than 25% of women with new onset idiopathic epilepsy have this autoantibody. An extracellular epitope is located in the N-terminal domain of the NR1subunit of the NMDAR. Antibodies directed against the NR2 subunit are found in neuropsychiatric SLE, where certain dsDNA antibodies may cross-react with epitopes on the NR2A and NR2B subunits.
Laboratories typically detect anti-NMDAR autoantibodies by IIF on cryostat sections of hippocampus or cerebellum. More specific assays include CBA, which detects NMDAR antibodies with higher sensitivity.
VGCC. Autoantibodies directed against voltage-gated calcium channel (VGCC) also are known as anti-voltage dependent calcium channels or anti-LEMS (Lambert-Eaton myasthenic syndrome). VGCC is a macromolecular complex, composed of P, Q, N, L, T, and R subunits. The P/Q channels are functionally related to the release of acetylcholine neurotransmitters from the nerve endings, whereas N-type channels are key components of the autonomic conduction system. In contrast to antibody blockade of the post-synaptic acetylcholine receptor targeted by myasthenia gravis autoantibodies, VGCC antibodies bind presynaptic nerve terminals.
Up to 95% of LEMS-positive sera also contain anti-P/Q type autoantibodies, but results from conventional immunoassays correlate poorly with disease severity. The majority of LEMS patients have SCLC, and ≤ 95% of these patients also have anti-P/Q type antibodies. It is important to recognize that anti-VGCC antibodies are present in 76% of LEMS patients without detectable tumors. In addition, anti P/Q type antibodies are observed in up to 7% of SCLC without LEMS, although anti-P/Q type antibodies have not been reported in myasthenia gravis despite the strong overlap of clinical signs and symptoms. P/Q or N antibodies, however, do co-exist with anti-acetylcholine receptor antibodies in what is referred to as the overlap MG-LEMS syndrome. A few cases of anti-P/Q type antibodies in the sera of cerebellar ataxia patients have been reported.
Researchers also have reported N-type calcium channel antibodies in 30–40% of patients with LEMS and L-type calcium channel antibodies in ≤75% of patients with amyotrophic lateral sclerosis, although these observations have yet to be validated. Furthermore, other studies have reported autoantibodies against different subunits of L-type VGCC in myasthenia gravis and congenital heart block sera. RIPA using 125I-ω-conotoxin-labelled calcium channels detects anti-P/Q-type calcium channel antibodies. More recently, a CBA has become available, and an ELISA is being developed. RIPA using1255I-ω-conotoxin-labelled calcium channels to detect N-type calcium channel antibodies.
VGKC. The eight α-subunits of VGKC Shaker type are designated as Kv1.1–1.8. Researchers have described three Kv1-associated proteins (Caspr2, LGI1, Tag-1/Contactin-2) as the antigenic target structures of potassium channel antibodies. Patients with limbic encephalitis predominantly have antibodies directed against LGI1 and less commonly against Caspr2; however, patients with neuromyotonia have Caspr2 antibodies as the main VGKC antibody specificity. Anti-VGKC antibodies are biomarkers for acquired neuromyotonia, but prevalence is highly dependent on the assays used. With radioimmunoassays, <65% of neuromyotonia patients have VGKC antibodies. By comparison, IIF is much more sensitive, detecting up to 100%.
The clinical spectrum of VGKC antibody-positive patients includes acquired neuromyotonia, Morvan's syndrome, limbic encephalitis, other subacute encephalopathies, and certain manifestations of autoimmune dysautonomia. In most cases, the manifestations of these syndromes are sub-acute and not associated with a tumor malignancy. The spectrum of autoantibodies directed towards different VGKC subunits varies among neuromyotonia patients, but the most prevalent target is the Kv1.6 isotype (91%), followed by Kv1.2 (83%) and Kv1.1 (41%). Approximately 20% of neuromyotonia patients have a thymoma; SCLC is less common. Researchers also have detected VGKC antibodies in severe bowel autonomic dysmotility, idiopathic epilepsy, and myasthenia gravis. Some studies indicate that VGKC antibodies also are associated with an encephalopathy that resembles Jakob-Creutzfeldt disease, marked by rapidly progressive dementia.
Laboratories detect antibodies to VGKC components using RIPA with 125I-α-dendrotoxin-labeled native or recombinant α-subunits Kv1.1, 1.2, and 1.6, or TnT IP and CBA immunoassays.
A second class of autoantibodies involved in encephalitis, PNS, and ataxia target intracellular antigens. The genesis of these antibodies has become clearer with our understanding of how they are processed during programmed cell death (apoptosis) and, in some cases, how they are released from living cells as circulating microparticles or extracellular exosomes.
Amphiphysin. This autoantibody is a marker for PNS, but it also is associated with a wide range of clinical features, including encephalomyelitis, subacute sensory and sensorimotor neuronopathy. Patients with paraneoplastic stiff person syndrome (SPS) have these autoantibodies. In contrast, GAD65 autoantibodies suggest a diagnosis of idiopathic SPS. Amphiphysin autoantibodies are most commonly associated with SCLC and breast tumors, and patients often have other markers, such as CV2/CRMP5, Hu, and Nova-1 antibodies.
Laboratories detect amphiphysin autoantibodies by IIF. The test requires that the laboratory technologist observe more intense staining in the axons of the molecular layer than the granular layer. Other techniques for this antibody are WB and LIA. A positive IIF test, however, is not conclusive evidence for the presence of amphiphysin antibodies. Laboratories that use IIF as a screening test also should perform reflex testing with a specific assay like LIA.
CRMP5/CV2. CRMP5 (collapsing response mediator protein) autoantibodies, first referred to as CV2 antibodies, target a member of the CRMP1–5 family. CRMP5/CV2 autoantibodies that bind to an N-terminal epitope are most often observed in PNS patients; however, they are not specific for a particular PNS. The clinical presentation and findings can be quite diverse, including sub-acute sensory neuropathy, myelopathy and chorea, paraneoplastic cerebellar degeneration, and limbic encephalitis. In >50% of cases, these symptoms present 3–4 years before a tumor is clinically evident. The most common malignancies in this clinical setting are SCLC (≤75%) and thymoma (5%). Patients with a SCLC and CRMP5/CV2 antibodies have a better prognosis than those with Hu antibodies. This antibody also is found in approximately 5% of patients without tumors and in about 10% of patients with a tumor who do not have clinical features of a paraneoplastic syndrome. Other ONA antibodies, such as those binding amphiphysin (~20%), Hu antibodies (~15%), Ri (~10%) and PCA-2 (~45%) are found in up to 75% of positive sera.
HuD/ANNA type I. Before the molecular identity of anti-Hu was elucidated, these antibodies were characterized by an IIF pattern of neuronal staining referred to as anti-neuronal nuclear antibodies type I (ANNA type I). HuD antibodies are among the most prevalent ONA antibodies and are highly specific biomarkers for PNS that presents as limbic encephalitis and sensory neuropathies. In >50% of patients, these neurological features precede tumor diagnosis by 3–4 years. HuD antibodies are typically associated with SCLC, although they have been described in a variety of other tumors, including neuroblastoma, prostate carcinoma, breast carcinoma, rhabdomyosarcoma, seminoma, and adenocarcinoma. Because patients with HuD autoantibodies have a better prognosis with respect to the tumor, there is some thought that antibodies are part of the immune response to suppress the tumor. HuD antibodies are usually detected in high titers (85%), but the titers do not correlate with neurological symptoms or presence of a tumor. In a minority (2%) of positive patients, there is no detectable tumor. Patients who have anti-HuD antibodies often have other ONA antibodies, such as Ri (25%), CRMP5/CV2 (15%) and amphiphysin (10%).
Because the IIF nuclear staining pattern of anti-HuD antibodies is remarkably similar to anti-Ri/NOVA-1 antibodies and a number of other autoantibodies found in systemic autoimmune rheumatic diseases, laboratories should differentiate them by testing the serum using LIA, ELISA, or some other analyte-specific immunoassay. HuD autoantibodies can be differentiated from Ri autoantibodies by staining sections of intestine where HuD antibodies react with the neuronal nuclei of the myenteric plexus.
NOVA-1/Ri/ANNA-2. Anti-neuronal nuclearantibodies type 2 (ANNA-2)/Ri/NOVA-1 autoantibodies are directed against neuronal cell nuclei. Anti-Nova-1 antibodies are rare ONA and are a biomarker of PNS that is primarily associated with breast cancer and SCLC. They also are associated with oculo-motor disturbances, cerebellar degeneration, and brain stem encephalitis. In >50% of patients, these neurological symptoms are observed 3–4 years before a tumor is clinically evident. A higher proportion of women have this autoantibody compared to men (2:1), and a patient also can have antibodies to Hu (25%), CRMP5/ CV2 (10%), and amphiphysin (5%).
Laboratories detect these antibodies by IIF on cryopreserved sections of primate cerebellum and intestine. Nova-1 antibodies show a characteristic granular staining of virtually all neuronal nuclei, and in some cases weak cytoplasmic staining is observed. Nova-1 autoantibodies can be differentiated from HuD autoantibodies because the former do not react with neuronal nuclei of the myenteric plexus.
Ma2/Ta/PNMA2. This autoantibody’s targets include three proteins of neuronal nuclei, Ma1/Ma, Ma2/Ta and Ma3, that share similar amino acid sequences, but the dominant epitope is in Ma2. Ma2/Ta autoantibodies are diagnostic markers of PNS, and the majority (>75%) of patients have testicular carcinoma. The antibodies also are found in ≤40% of patients with cerebellar ataxia in the setting of SCLC. PNMA2 antibodies are typically found in younger patients (22–45 years) and are more common in men than women. Associated clinical features include limbic encephalitis, paraneoplastic cerebellar degeneration, cerebellar ataxia, and brainstem encephalitis. In >50%, the neurological signs and symptoms are observed before a tumor is clinically detectable. In a small minority of patients (5%), PNMA2 antibodies are found in patients without a detectable tumor.
Laboratories detect anti-PNMA2 by IIF on cryopreserved sections of primate cerebellum where the pattern of nucleolar staining is observed. Since a number of other non-tissue specific nucleolar autoantibodies are observed in normal individuals as well as other autoimmune conditions, laboratories should reflex test positive IIF staining by LIA.
APCA/Yo/CDR62 type I. The early nomenclature of anti-Purkinje cell antibodies type 1 (APCA-1) was based on the typical IIF staining pattern of cerebellar Purkinje cells. The primary target of APCA-1/Yo antibodies is CRD62, a member of a family of neuronal proteins that also includes CDR34 and CDR52. CRD62 is highly expressed in the cerebellar Purkinje cells and in cells of the brainstem and certain tumors. Antibodies to Yo/CDR62 are the second most common ONA antibodies, and they are highly sensitive diagnostic (>90%) markers of a PNS characterized by cerebellar degeneration called Yo syndrome and ataxia. In >50% of patients, the neurological symptoms are observed 3–4 years before a tumor is clinically detectable. The autoantibodies are predominantly associated with ovarian or breast carcinoma but have also been observed in carcinomas of the uterus, stomach, salivary glands, and esophagus. The vast majority of patients with anti-CDR62/Yo antibodies are women. Current evidence suggests that unlike patients with antibodies to some other ONA, patients with Yo antibodies do not have an improved prognosis with respect to outcome of the malignancy.
Laboratories detect anti-CDR62/Yo by IIF employing cryopreserved sections of mammalian cerebellum where the typical pattern of robust cytoplasmic staining of the Purkinje cells is noted. This pattern is so highly specific for anti-CDR62/Yo that some laboratories consider it sufficient to make the serological diagnosis; however, others confirm the test by LIA using a recombinant CDR62 antigen.
Important Diagnostic Aids
Identification and detection of autoantibodies in the serum and CSF of patients who present with a wide variety of neurological signs and symptoms has become a significant diagnostic and prognostic aid to clinicians. As current technologies are refined and newer diagnostic technologies are developed, enhanced diagnostic sensitivity and specificity of these assays is critically important. Future advances will require a strong interface between clinicians, laboratory scientists, and the diagnostic industry.
- Gozzard P, Maddison P. Which antibody and which cancer in which paraneoplastic syndromes? Pract Neurol 2010;10:260–70.
- Lancaster E, Martinez-Hernandez E, Dalmau J. Encephalitis and antibodies to synaptic and neuronal cell surface proteins. Neurology 2011;77:179–89.
- Graus F, Saiz A, Dalmau J. Antibodies and neuronal autoimmune disorders of the CNS. J Neurol 2010;257:509–17.
- Luqmani RA, Suppiah R, Grayson PC, Merkel PA, et al. Nomenclature and classification of vasculitis—update on the ACR/EULAR diagnosis and classification of vasculitis study (DCVAS). Clin Exp Immunol 2011;164 Suppl 1:11–3.
- Tuzun E, Dalmau J. Limbic encephalitis and variants: classification, diagnosis and treatment. Neurologist 2007;13:261–71.
- Kayser MS, Dalmau J. The emerging link between autoimmune disorders and neuropsychiatric disease. J Neuropsychiatry Clin Neurosci 2011;23:90–7.
- Braik T, Evans AT, Telfer M, McDunn S. Paraneoplastic neurological syndromes: unusual presentations of cancer. A practical review. Am J Med Sci 2010;340:301–8.
- Sadeghian H, Vernino S. Progress in the management of paraneoplastic neurological disorders. Ther Adv Neurol Disord 2010;3:43–52.
- Flanagan EP, Caselli RJ. Autoimmune encephalopathy. Semin Neurol 2011;31:144–57.
- Koike H, Tanaka F, Sobue G. Paraneoplastic neuropathy: wide-ranging clinicopathological manifestations. Curr Opin Neurol 2011;24:504–10.
Marvin J. Fritzler, MD, PhD, is a professor of medicine and clinician-scientist at the University of Calgary. Email: firstname.lastname@example.org
Disclosures: Marvin J. Fritzler has received honoraria or gifts from ImmunoConcepts Inc., BioRad, and INOVA Diagnostics, Mikrogen GmbH, Euroimmun GmbH, and Dr. Fooke Labroatorien GmbH.