Association of Paraneoplastic Neurological Disorders With Glutamic Acid Decarboxylase Antibodies.

Abstract

An increasingnumberof antineural IgGantibodies havebeen detected since the 1980s in patients with autoimmune central nervous system disease. Antibodies to intracellular antigens were initially found and antibodies to antigens on the neural surfacehavenowbeen identified. Antibodies to the intracellular enzyme glutamic acid decarboxylase (GAD) were one of the earliest antibodies discovered.1 However, not all antibodies have the same value. This means that not all are pathogenic, specific for defined syndromes, and indicative for responsiveness to immunological treatment. Antibodies to theN-methyl-D-aspartate receptor anddefinedantigens in the voltage-gated potassium channel complex, such as leucine-rich glioma inactivated protein 1 and contactinassociatedprotein2 (most likely theantibodies to theα-amino3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor, the γ-aminobutyric acidB receptor, the glycine receptor, andothers as well), are high-rank antibodies in that sense. However, other antibodies have their flaws because they lack 1 or more of these criteria. Antibodies to GAD are quite specific and are found in stiff-man syndrome, cerebellar ataxia, limbic encephalitis, and temporal lobe epilepsy.2 Specificity beyond 1 single syndrome is not uncommon. Even the high-rank antibodies are not entirely specific, as observed in theN-methylD-aspartate receptor antibodies.3,4 However, as already discussed in the Solimena et al article,1 the intracellular location ofGADmakesan invivo immunological reactionunlikely;GAD antibodies are probably not pathogenic on their own.1 The response of GAD antibodies to treatment is not as effective as, for example, antibodies to elements of the voltage-gated potassium channel complex.5 Taken together, GAD antibodies, even if they are high titer, are somehow considered secondclass antibodies. Despite these limitations, GAD antibodies have fascinating features. Neurological researchers have made observations that keepchallenging the too-nihilistic viewon thevalue of these antibodies. Glutamic acid decarboxylase antibodies are found in patients with neurological symptoms usually at high titers and can be picked up by a number of techniques, such as radioimmunoprecicitation or enzyme-linked immunosorbent assays, rodent brains, cell-based assays, and immunoblots. These antibodies are relatively frequent. Although theyaredirected to an intracellular target, theyarenot as regularly associatedwithparaneoplastic syndrome (PNS) as the classic onconeural antibodies that are also directed to intracellular antigens, such as Hu and Yo, among others.2 Glutamic acid decarboxylase antibodies have prognostic implications because they usually indicate long-term difficult-totreat courses. For example, epilepsy associated with GAD antibodies is usually resistant to antiepileptic and immunotherapeutical interventions.6 However, in stiff-man syndrome with GAD antibodies, there is a prospective placebocontrolled immunotherapy trial demonstrating a superior outcomeon intravenous immunoglobulins.7Glutamicaciddecarboxylase autoimmunity often does not stand alone but is part of polyendocrine autoimmunity and is frequently associatedwithadditional antibodies (eg, thyroidantibodies).2Glutamic acid decarboxylase is a prominent enzyme in the central nervous system. It catalyzes the conversion of the main excitatory neurotransmitter (glutamate) to the main inhibitory transmitter (γ-aminobutyric acid). Should that not suggest a potential pathogenic potency of the GAD antibodies, given the frequent hyperexcitatory characters of associated neurological syndromes? Several animal transfer studieshave been performed with material from patients with GAD antibodies.These studieshavebeen interpretedas showingpathogenic effects of the antibodies, possibly modulated by subtle differences in antigenic targeting (see the Manto et al study8 withfurther references).However,otherauthorssuggestedthat it is not theGADantibodies but accompanying other antibodies that exert thesepathogenic effects.Anexample for this position is a study by Gresa-Arribas et al9 with Francesc Graus, MD, PhD, as the senior author. It is 1 of more than 20 studies on GAD antibodies that he has authored. In this issueof JAMANeurology, Arinoet al10 presentnovel data on GAD antibodies. In their collection of 121 cases with high-titer GAD antibodies, the authors found an unexpectedly large number of patients (n = 15; 12%) with tumors classified as having a definite or possible PNS. None of them had classic onconeural antibodies. However, 8 of 15 patients with PNS had additional antibodies directed to the following targets: γ-aminobutyric acid B receptors, γ-aminobutyric acid A receptors, glycine receptors, or unknown targets on neuronal cell surfaces. This underlines the impression that GAD antibodiesmore likely come togetherwith additional antibody reactivitiescomparedwithotherantibodies.Althoughtherehave been some small-scale studies on patientswithGAD antibodies and tumors, to our knowledge, this is the first large systematic studyof this constellation.Arinoet al10 arewell known for their experience with PNS. Thismakes it possible that the large number of these syndromes in their biobank is owing to a higher likelihood that materials from patients with tumors are sent to their laboratory.However, even if one takes intoacRelated article page 874 Opinion