Early-onset immunotherapy by intravenous immunoglobulin and corticosteroids in well characterized onconeural-antibody-positive paraneoplastic neurological syndrome

Paraneoplastic neurological syndromes (PNS) are defined as immune mediated disorders, potentially affecting the whole nervous system, and are often associated with cancer 1. Some clinical patterns are highly frequent in these syndromes, such as paraneoplastic cerebellar degeneration (PCD), limbic encephalitis (LE), subacute sensory neuronopathy (SSN) and Lambert–Eaton myasthenic syndrome (LEMS) 1. In the 1980s, antibodies targeting intracellular neuronal proteins were described as biomarkers of PNS 1, advocating an underlying tumour. Indeed, more than 90% of patients with such antibodies suffered from an associated cancer. Recently, patients with a similar clinical presentation to PNS have been reported with other autoantibodies targeting cell-surface synaptic antigens. An associated cancer is less frequent in these patients. These findings have changed the concept of PNS 2, and the new PNS classification is now based on the location of the targeted antigen, instead of being defined by clinical symptoms or oncological status.

Two main groups of patients can be delineated according to the neuronal localization of the antigens targeted by the associated autoantibodies: group 1, antibodies with cytoplasmic neuronal antigens (CNA-antibodies); and group 2, antibodies with cell-surface neuronal antigens (CSNA-antibodies). Each group shows profound differences in the pathophysiology and the pathogenic role of the antibodies. Presence or absence of a tumour, prognostic and treatment responses also differ fundamentally from one group to the other 2. In group 1 a tumour is almost always present, the neurological symptoms are severe and patients generally do not improve with immunomodulatory treatment. In contrast, in group 2 a tumour is rarely present, although some differences may be observed according to the subtype of associated antibody; the neurological symptoms are severe, but improve with immunomodulatory treatment in the majority of patients.

Patients with CNA-antibodies and PNS are rare, representing fewer than 0·01% of patients with cancer 3. The main CNA-antibodies are directed against HuD 4, Ri 5, Yo 6, Ma1/2 7, CV2/CRMP5 8, Sox1 9 and Zic4 proteins 10, and are associated strongly with tumours. Anti-Hu and anti-Yo antibodies are the most frequent 11, and are related strongly to small cell lung carcinoma (SCLC) 4 and gynaecological tumours (breast and ovarian carcinomas) 6, respectively. The pathophysiological role of CNA-antibodies is still under debate. These antibodies are not considered to be directly pathogenic, acting instead as markers of a cytotoxic T cell immune response directed towards neurones. Indeed, cerebral biopsies and autopsies of patients with these autoantibodies have shown the presence of cytotoxic T cells in parenchyma, associated with a profound loss of neurones 12. Thus far, no animal models of PNS have successfully recreated the neuronal loss observed in patients, despite several attempts in murine models, including injection of anti-Hu and anti-Yo antibodies 13,14, immunization with purified Yo or HuD antigens 13,14 and injection with activated T cells (especially in the case of anti-Hu and anti-Yo antibodies) 14.

Irreversible neuronal death has been associated with the neurological symptoms presented by patients explaining the poor response to therapy. Given these therapeutic challenges, the main objective has been to stabilize the neurological symptoms. The relative rarity of these disorders, together with the heterogeneity of clinical patterns, are the main reasons explaining the scarce clinical trials in the field of PNS associated with CNA-antibodies. However, a few retrospective clinical studies 3–6,15–19 and prospective studies 20,21 (Table 1) exist and give some therapeutic guidance into the management of these conditions. The cornerstone in treating PNS associated with CNA-antibodies is curing the tumour. Tumour treatment must be the first objective to stop the immune reaction and the neuronal death. All retrospective studies suggest that rapid tumour treatment improves patient outcome. However, prospective studies are necessary to confirm these observations. The reported response to immunotherapy is generally poor, due probably to neuronal loss and death. Only two prospective trials have been conducted 20,21 in patients with anti-Hu, anti-Yo or anti-CV2/CRMP5 antibodies. One prospective open-label study compared plasma exchange plus conventional cancer chemotherapy (10 patients) to plasma exchange plus continuous oral cyclophosphamide (10 patients). Patients treated with cyclophosphamide showed sustained neurological improvement which continued after completion of the study 21. A second prospective, uncontrolled, unblinded trial suggested a possible benefit of human chorionic gonadotrophin (hCG) in the treatment of PNS 20. In the absence of large clinical trials, most therapeutic alternatives in patients with CNA-antibodies derive from observational clinical studies and case reports 3,5,15–19. The usefulness of corticoids has been suggested in patients with anti-Hu 22, anti-Yo 23 antibodies or in patients with PNS and without antibodies 24,25. Therapeutic effects of intravenous immunoglobulins (IVIg) appear to be limited, but in a very small proportion of patients with definite PNS (mainly with anti-Hu or anti-Yo antibodies) a favourable response has been described after IVIg therapy 26–28. However, this response was not reproduced in a study of 19 patients with anti-Yo antibodies and cerebellar ataxia treated with immunomodulators (including IVIg) 5. Plasma exchange is sometimes proposed instead of IVIg 29, and other authors have observed a stabilization of patients with anti-Hu and anti-Yo antibodies following treatment with cyclophosphamide in combination with corticoids and IVIg 16. Stabilization of a few patients presenting anti-Hu and anti-Yo antibodies has been observed in uncontrolled and unblinded trials with rituximab 17,30,31. Additionally, in a single-centre trial testing tacrolimus in PNS patients presenting with anti-Hu antibodies, some patients showed good tolerance and a trend towards neurological improvement 3. Given the potential role of IVIg in the management of PNS patients, we are currently conducting a Phase II multi-centre clinical trial to determine the efficacy of IVIg in the first 3 months of treatment. The trial (called Iason) is aiming to recruit 19 patients with anti-Hu, anti-Yo or anti-CV2/CRMP5 antibodies, and results are expected in 2016.

Table 1.

Main treatment studies in group 1 antibodies: antibodies with cytoplasmic neuronal antigens (CNA-antibodies)

Antibodies (number of patients)	Type of study	Treatment modalities	Results	Authors
Hu, Yo, CV2 (329)	Retrospective	Corticoids, IVIg, PE, cyclophosphamide, azathioprine	Infrequent clinical improvement	Uchuya et al., 1996 19 Keime-Guibert et al., 2000 16 Rojas et al., 2000 6 Graus et al., 2001 15 Sillevis Smitt et al., 2002 18
Yo, Hu, DNER and Ri (48)	Retrospective	Tumour treatment, immunosuppression	Few patients improved after tumour therapy	Shams'ili et al., 2003 5
Ma2 (38)	Retrospective	Tumour therapy if applicable, and corticoids or IVIg	A third of patients with favourable outcome and 50% with clinical degradation	Dalmau et al., 2004 7
Hu and Yo (9)	Retrospective	Rituximab	4/9 with clinical improvement	Shams'ili et al., 2006 17
Hu, Yo and CV2 (26)	Retrospective	Tacrolimus	Few clinical improvement	Orange et al., 2012 3
Hu, Yo, CV2 (20)	Prospective	PE and cyclophosphamide versus PE and chemotherapy	Better outcome in patients treated with cyclophosphamide	Vernino et al., 2004 21
Hu (15)	Prospective	hCG	A third of patients harboured a significant improvement	van Broekhoven et al., 2010 20

hCG = human chorionic gonadotrophin; IVIg = intravenous immunoglobulin; PE = plasma exchange.

Further prospective clinical trials are essential to compare standardized treatment approaches more effectively, as well as to confirm that the improved outcomes observed in patients treated with immunomodulation are not the result of possible biases such as subtle differences in disease severity or differences in follow-up between retrospectively identified groups. These prospective trials must be conducted in homogeneous groups of patients according to the associated circulating autoantibodies in order to create guidelines for effective treatment.

Disclosure

J. H. has no conflicts of interest to disclose.