The clinical significance of atypical indirect immunofluorescence patterns on primate cerebellum in paraneoplastic antibody screening

Paraneoplastic neurological syndromes (PNS) are remote effects of cancer not caused by tumor growth, infiltration, metastasis or chemotherapy [1, 2]. They are predominantly the result of an autoimmune process, mediated by T-lymphocytes and/or antineuronal antibodies (AN Ab) [1, 3]. The current hypothesis is that the immune response is directed against neuronal antigens ectopically expressed by the tumor [1, 2, 4]. Often, PNS and corresponding Ab surface before a tumor is discovered [5‐9]. Therefore, it is important to meticulously search for AN Ab since they might urge clinicians to initiate oncological screening [10, 11]. Moreover, since some Ab are closely linked with a certain PNS, detection of AN Ab facilitates the diagnosis of that specific PNS.

Based upon targeted antigen location, AN Ab can be divided into two subgroups: intracellular (group I) or on the cell membrane (group II) (Fig. 1) [1, 12, 13]. Their detection is classically performed by screening assays such as indirect immunofluorescence (IIF) screening on primate cerebellum, followed by typing assays such as line blots or cell-based assays [1, 13‐16]. Intraneuronal and anti-Amphiphysin Ab (Fig. 1), the most often paraneoplastic (PN) Ab, show typical IIF patterns on primate cerebellum such as positive neuronal nuclei (anti-Hu) or coarse granular staining of cytoplasm of Purkinje cells (anti-Yo) [1, 12, 13, 17]. However, atypical patterns that are not specifically related to a PN Ab (i.e. negative on typing for intraneuronal and anti-Amphiphysin Ab), are also reported. For example, the Purkinje cell layer can show positivity without a specific pattern (coarse granular staining of Purkinje cell cytoplasm) being reported, while in e.g. the molecular layer a ‘pan-layer’ fluorescence instead of a specific pattern (e.g. dot-like fluorescence or positivity of neuronal nuclei) is often seen. The clinical significance of such patterns is not yet known. An overview of antineuronal antibodies with their associated fluorescence patterns, neoplasms and clinical features can be seen in Table 1.

In this study, we demonstrated that atypical IIF patterns on primate cerebellar slices are most often observed in neurological diseases other than PNS, including epilepsy and peripheral neuropathies. In only approximately 3.7% of cases, patients with an atypical IIF pattern were diagnosed with a PNS. Among the different disease groups, a different distribution of atypical patterns was seen (i.e. on a group level). This, however, did not result in any added diagnostic information as among different disease groups, a significant different distribution between positive and negative was absent for each and every pattern group. Indeed, within the same diagnostic group, multiple different layers were reported to be positive (Fig. 2).

Another possible cause for the clinically irrelevant results is the binding of atypical pattern generating AN Ab (i.e. anti-GAD or group II Ab) or non-AN Ab to antigens in primate cerebellum. This would lead to the positivity of certain layers, despite the absence of typical pattern-generating group I Ab (intraneuronal and anti-Amphiphysin Ab). Such an explanation has previously been suggested by Haukanes et al. [21]. In their study, some ADHD patients’ samples demonstrated a positive staining of the cytoplasm of Purkinje cells on IIF screening [21]. However, since typing assays were negative for most patients, the authors decided that the reported patterns were probably due to a non-specific binding of Ab to unknown antigens [21]. Of interest, our study observed the presence of AN Ab other than typical pattern-generating antibodies (i.e. anti-GAD and group II Ab) in 20 of the 156 ‘non-PNS’ patients. These Ab included anti-GAD, anti-GABA_B, anti-VGKC and anti-aquaporine-4 Ab and most were searched for in the majority of patients (e.g. 98/162 for anti-GAD, 87/162 for anti-VGKC and 76/162 for anti-NMDA receptor Ab). As the cumulative prevalence of those antibodies—as reported by the large scale study of Dahm et al.—in a ‘neurological disease population’ (0.5% for anti-GAD and ≤ 0.1% for the other Ab) is clearly lower than seen in our cohort, the percentage of patients with anti-GAD or group II Ab might be enriched in a cohort of patients having an atypical IIF pattern on primate cerebellar slices [22]. Therefore, it might be recommended that if the clinical presentation correlates with the presence of an AN Ab but an atypical IIF pattern and a negative line blot are produced by the laboratory, the search for Ab might be extended to cover the cell-surface (group II) Ab as well. In such case, before considering assays such as cell-based assays to identify group II Ab, the IIF screening can be extended to hippocampal slices as group II Ab have been shown to generate more typical patterns on hippocampal slices [14]. As such, the pattern generated on the hippocampal slices can provide valuable information regarding the identity of the group II Ab present, and hence suggest which cell-based assay should be performed. Moreover, the combination of both hippocampus slices and cell-based assays can provide higher sensitivities and specificities than cell-based assays alone [23].

Some limitations were encountered in this study. Firstly, our study did not include paired serum and CSF samples for most patients, as for only five patients screening was performed on both serum and CSF (only on serum, N = 154; only on CSF, N = 3). Nevertheless, it would be of interest to learn whether an atypical IIF pattern seen in serum could be confirmed in CSF. If so, it could also be worthwhile to investigate whether such paired results are related to the presence of AN Ab or to the clinical diagnosis of PNS. A second limitation was encountered when investigating possible explanations for the positivity of IIF screening assays, in light of the absence of group I Ab (excluding anti-GAD). Our secondary outcome analysis provided evidence that the presence of AN Ab other than the typical-pattern generating Ab could be put forward as a possible explanation. However, since this was a retrospective study using results gathered in daily routine clinical practice, not all samples received the same extent of AN Ab screening (i.e. not all samples were tested on the presence of every known AN Ab). A probable explanation for this could be that the consulting neurologist did not consider group II Ab as a possible cause for symptoms seen in a patient and, hence, the neurologist did not request any analysis for group II Ab (including surface Ab screening. Nevertheless, an extensive AN Ab screening for each patient with a positive yet atypical IIF pattern would further contribute to the understanding of the significance of such atypical IIF patterns. Therefore, a follow-up study where for each sample an extensive search for all AN Ab is performed (including for group II Ab) in order to generate a more complete overview of patients’ positive layers and AN Ab present, would be interesting.

A limited number of studies investigated the performance characteristics of the IIF screening on primate cerebellum. This was done in the context of the sensitivity and specificity of IIF on cerebellar slices for diagnosing PNS [4, 24, 25]. Our study, however, investigated the clinical significance of atypical IIF patterns on cerebellar slices, i.e. patterns that are not related to the presence of group I Ab (excluding anti-GAD Ab) and that are not followed by a positive line blot typing assay. Now, we show that such atypical IIF patterns are of limited clinical significance. Nevertheless, the enrichment of AN Ab (including anti-GAD, anti-VGKC and anti-GABA_B Ab) in a cohort of patients with atypical IIF patterns, calls for further investigation with respect to action required upon the presence of an atypical pattern on primate cerebellar slices. Furthermore, in 3.7% of the cases, a diagnosis of a PNS had been reached in the presence of an atypical IIF pattern without positive line blot typing, warranting caution in regard to the depiction of such IIF screening results as purely clinically irrelevant.