Inflammatory demyelinating polyneuropathy: A complication of immunotherapy in malignant melanoma

Annals of Oncology 11: 1197-1200, 2000. © 2000 Kluner Academic Publishers. Primed in the Netherlands.

Clinical case Inflammatory demyelinating polyneuropathy: A complication of immunotherapy in malignant melanoma D. A. Anthoney,1 I. Bone2 & T. R. J. Evans1 [ CRC Department of Medical Oncology, Beatson Oncology Centre, Western Infirmary; Southern General Hospital, Glasgow, UK

Summary Paraneoplastic syndromes (PNS) involving the central nervous system are a rare manifestation of malignant disease. As they commonly precede the diagnosis of malignancy their acute manifestations do not often present themselves to oncologists in the first instance. It is currently believed that most, if not all, neurological PNS are autoimmune in nature. Proteins expressed ectopically on the surface of tumour cells generate an immune response which cross-reacts with the same, or similar, proteins in the nervous system resulting in damage. This can involve a single cell type of the nervous system whilst in other cases the impairment is more widespread.

Case history A 55-year-old lady presented to our institute having originally undergone wide local excision of a lentigo maligna malignant melanoma (Clark level 4; Breslow thickness 1.0 mm) from her left upper arm in July 1987. Three years later she relapsed with a left axillary lymph node metastasis which was completely excised. On referral in February 1991 she was randomised into the Adjuvant Interferon in High Risk Melanoma trial and drew the no further treatment arm. She remained disease free until July 1996 when she was investigated for iron deficiency anaemia. A stricture in her mid-ileum was detected and resection revealed malignant melanoma with metastases in omental lymph nodes. Further investigations detected lymphadenopathy anterior to the IVC and a possible liver metastasis. She was started on interferon-a (maximum dose 9 mega-units, subcutaneously, 3 x weekly), and CT after two months of treatment revealed a decrease in the abdominal lymphadenopathy and resolution of the hepatic lesion. Interferon continued until April 1997 when repeat CT showed progression of the lymphadenopathy between IVC and aorta. At this time she described symptoms of tiredness, general fatigue and tingling of her hands and feet. No focal neurological deficit could be detected and her symptoms were attributed to the interferon therapy. She represented in January 1998 with headaches, dizziness and staggering (to her left side). Neurological

2

Institute of Neurological Sciences,

The following report is of a case of chronic inflammatory demyelinating polyneuropathy (CIDP) occurring in metastatic malignant melanoma, following treatment with interferon-a. We review the current literature on this rare association and speculate on its pathogenesis, and the implications for future therapeutic strategies in melanoma targeting tumour antigens. Key words: immunotherapy, inflammatory demyelinating polyneuropathy, interferon, malignant melanoma Abbreviations: IVC - inferior vena cava; SIADH - syndrome of inappropriate antidiuretic hormone; IHC - immunohistochemical

examination revealed a positive Romberg's test but no other specific abnormality. CTand MRI of brain showed no metastases, meningeal deposits or other disease. The CSF was normal except for an elevated protein level (1.20 g/1), serum biochemistry was normal and haematology revealed a simple neutrophilia. She was empirically started on steroids (dexamethasone 4 mg three times daily) with an immediate improvement in her symptoms. Neurological review by one of the authors (I. Bone) a few days after steroids were started showed no abnormalities (Romberg's sign now negative) and the dose was gradually reduced with no significant clinical deterioration. Approximately eight weeks later further neurological symptoms developed (bilateral ptosis with demonstrable fatigueability, loss of lower limb reflexes) which resolved spontaneously over a period of weeks. An anti-acetylcholine receptor antibody test, performed at this time, was negative. In October 1998 she developed frontal headaches, loss of balance, numbness of her hands and feet and was admitted to hospital for further investigations. Clinically she displayed decreased sensation to light touch over her fingers and loss of light touch and pinprick sensation over her feet bilaterally. There was also weakness of all lower limb muscle groups (MRC grade 3) and absent reflexes. Repeat radiological and biochemical investigations were unchanged except for a rise in the CSF protein level (2.87 g/1). An auto-antibody screen

1198 Table 2. Reported cases of malignant melanoma associated CIDP.

Table I. Nerve conduction studies. Nerve\segment

Distal latency (ms)

Conduction velocity (m/s)

Amplitude (mV)

Median (elbow - APB) Ulnar (AE - ADM) Tibial (Pop - AH)

18.4(4.1) 18.3(3.9) 14.5(4.9)

36.7(49) 28.3(49) 49.0(41)

3.29(4) 2.19(6) 0.61

Abbreviations: APB - abductor policus brevis; ADM - abductor digiti minimi; AH - abductor hallicus (upper or lower limit of normal values given in parentheses). Very significant abnormalities of motor conduction are seen in the upper limbs. The conduction velocities seen in the ulnar nerve are within the demyelinating range. Significant reductions in distal latency and amplitude are also observed in the Tibial nerve. The results of the electrodiagnostic studies as shown are strongly supportive of an acquired demyelinating polyneuropathy.

(including anti-HU, anti-YO and anti-glycolipid antibodies) was negative except for an anti-GM3 IgM titre of 1 /240 (normal range < 1 /500). CT-guided biopsy of the observed para-aortic lymphadenopathy was performed and confirmed the presence of melanoma. She was transferred to the regional neurological institute for further investigation. Nerve conduction studies were performed (Table 1) and on the basis of these results, as well as the clinical features and significantly elevated CSF protein, a diagnosis of chronic inflammatory demyelinating polyneuropathy (CIDP) was made. She was started on intravenous immunoglobulin with an initial symptomatic improvement. However, her symptoms worsened after about four weeks and she was started on steroids (prednisolone 60 mg alternate days) which resulted in some improvement in her neurological function. At present she is able to mobilise with support and has increasing sensation in her lower limbs. She remains on prednisolone (30 mg alternate days) but has not required any further treatment for her melanoma.

Discussion Chronic inflammatory demyelinating polyneuropathy (CIDP) is a slowly progressive, sometimes relapsing, steroid dependent, demyelinating polyneuropathy primarily of the limbs [1]. In its non-paraneoplastic form it is slightly commoner in males and can occur at any age with up to a third of patients giving a history of antecedent viral infection. An immunological pathogenesis for CIDP is suggested by the association with various HLA antigen subgroups and observed responses to immunosuppressive therapy. The classical clinical features are of a symmetrical mixed sensorimotor neuropathy, presenting with paraesthesiae (sometimes painful), loss of vibration and jointposition senses, and limb weakness both of proximal and distal muscle. Usually, all reflexes are lost. The diagnosis of CIDP is based predominantly on a combination of clinical, electrophysiological and histological findings [2, 3].

Primary Patient Age (years) site of melanoma

Time from diagnosis melanoma to onset CIDP

Antimelanoma therapy

Reference

1

62

Unknown. Axillary lymph node biopsy +ve

CIDP predates melanoma by four months

Surgery (lymph node removal)

[5]

2

43

Unknown. Axillary lymph node biopsy +ve

CIDP predates melanoma by six months

Surgery (lymph node removal)

[5]

3

49

Pre-sternal

1 year

Surgery + radiotherapy

[5]

4

73

Not

12-15 years

Surgery, a-interferon, IL-2, melanoma vaccine

[6]

11 years

Surgery, a-interferon

Present case

reported

5

55

Upper arm

The association of CIDP with malignant disease is extremely rare. A handful of previously reported cases have suggested an association between adenocarcinoma of the pancreas, colon and bile duct and sensorimotor neuropathies bearing the major clinical features of CIDP [4]. The relationship between CIDP and malignant melanoma is equally rare with the patient presented here bringing the number of reported cases to five [5, 6] (Table 2). In two of these cases melanoma was detected incidentally during the diagnostic work-up for patients presenting with features typical of CIDP. In the other cases CIDP arose in patients with pre-existing melanoma having undergone multiple and varied therapies. In view of the small number of reported cases is the association between malignant melanoma and CIDP anything more than a rare medical curiosity? Melanocytes are of neural crest origin and it is well established that malignant melanoma tumour cells express surface glycoproteins which share immunogenic similarities with glycoproteins on the surface of cells in the central and peripheral nervous systems [7, 8]. The gangliosides expressed on the surface of melanoma cells include GM3, GD3, GM2 and GD2 the latter two being highly immunogenic in man. These gangliosides are also found expressed on Schwann cells in the peripheral nervous system. Evidence is now accumulating to suggest that antibodies raised against these glycoproteins on the surface of melanoma cells may also be responsible for initiating CIDP. Malignant melanoma is known to be a potentially immunogenic tumour and hence development of mela-

1199 noma vaccines has been one important line of research over the past twenty years. Early trials of melanoma vaccines used vaccinia virus lysates of cell cultures containing melanoma cell derived antigens [9]. Fuller et al. [10] reported two cases of patients with malignant melanoma treated by intradermal injection of vaccinia melanoma cell lysates who developed demyelinating polyneuropathy, with one fulfilling the criteria for diagnosis of CIDP while the other displayed similar clinical features with a more acute course. Monoclonal antibodies against specific cell surface antigens have also been investigated as another form of melanoma immunotherapy. A phase I trial of monoclonal anti-GD2 antibody in melanoma resulted in some patients developing senorimotor polyneuropathy, with or without SIADH [11]. Yuki et al. have shown that monoclonal antibody to GD2 reacts with cells in the posterior lobe of the pituitary gland and with the myelin sheaths of peripheral nerves [12]. In addition, of 257 patients with a range of neurological disorders (including 21 with CIDP) the serum of seven patients displayed significant titres of anti-GD2 antibody whereas none of the 50 normal controls or 18 patients with non-neurological autoimmune diseases displayed antibody [12]. As the antibodies in these patients are polyclonal the authors speculate that anti-GD2 antibody subtype specificity may be important in determining whether SIADH is part of the clinical complex of neuropathy observed in melanoma patients after vaccination. Further reinforcement of a link between CIDP and anti-glycoside antibodies is provided by the use of immunostaining in a case of CIDP arising in a patient with a 15-year history of malignant melanoma [6]. Strong and highly specific IgM reactivity to ganglioside GM2 was observed in the patients serum. Thin sections of the patients original melanoma were subjected to IHC analysis using the patients own serum and rabbit polyclonal anti-GM2 antiserum. Many of the melanoma cells reacted with an IgM component of the patients own serum and these same cells reacted specifically with the anti-GM2 antibody. The authors conclude that these results suggest that neuropathy in patients with melanoma could relate to an immune response to crossreacting antigens on the tumour and in the nerve. They label this 'molecular mimicry'. In those cases in which the neurological features of CIDP are closely associated temporally with the onset of melanoma, the molecular mimicry hypothesis would fit with the presentation of new cross-reacting antigens on the melanoma cells. However, cases in which the melanoma has been present for many years before the onset of neurological symptoms are more difficult to explain. Alteration in the character of surface glycosides on the melanoma cells is a possible explanation as to the sudden initiation of an immune reaction leading to demyelinating polyneuropathy. Yet, evidence suggests that expression of surface glycosides on melanoma cells is stable during tumour progression [8], therefore this is unlikely to be a feasible mechanism.

The mainstay of treatment for CIDP is the use of steroids. In non-steroid responsive patients other treatments used have included immunosuppressant agents, plasma exchange and intravenous immunoglobulin. More recently a number of reports have shown that dramatic improvement can be achieved in steroid unresponsive CIDP by the use of interferon-a [13, 14]. It is known that cytokines, including interferon-a, have multiple roles in the induction and suppression of immune responses and autoimmunity. It is not surprising therefore that substantial evidence exists indicating that interferon-a therapy can induce auto-immune diseases. The thyroid gland has been the predominant target of this autoimmune response in interferon-a therapy with several reports of reversible thyroid dysfunction in association with transient thyroid autoantibody formation [15,16]. Interferon-a induced autoimmunity has been recognised in other organ systems with the development of insulin-dependent diabetes mellitus, myasthenia gravis and cutaneous disorders (including lichen planus) [1719]. The early trials of interferons were associated with a number of neurological side-effects including peripheral neuropathy, although with no evidence of auto-immune reactivity [20]. Nevertheless a number of authors have reported more specific neurological syndromes arising during treatment with interferon-a [21, 22] including a patient who developed CIDP whilst on interferon-a treatment for hepatitis C. Although there was no direct immunological evidence of interferon inducing the inflammatory neuropathy in this case the authors suggest possible mechanisms for its involvement. It is interesting to speculate whether interferon-a was also responsible for triggering the onset of CIDP in the patient presented in this paper. She had had melanoma for more than 10 years prior to developing CIDP and her disease was stable at the time her neurological symptoms arose. It is unlikely, therefore, that clonal expansion of tumour cells with novel surface glycosides was responsible for initiating her inflammatory demyelination. She had received a course of interferon-a shortly before her symptoms evolved and it is plausible that this initiated an auto-immune reaction against her melanoma and in addition the development of cross-reactivity with antigens expressed on neuronal cells leading to an auto-immune CIDP. We conclude that this case highlights an important potential side-effect and morbidity of immunotherapies in the management of malignant melanoma. This has the potential to be particularly relevant in the future with the likely increase in the use of melanoma vaccines. Acknowledgements The authors would like to thank the contributions made to the management of this case and to compiling this report by Dr C. Mann and Dr H. J. Willison, Institute of Neurological Sciences.

1200 References 1. Walton J. Disorders of peripheral nerves. In Walton J (ed): Brain's Diseases of the Nervous System. Oxford/New York: Oxford University Press 1993; 604-5. 2. Ad Hoc Subcommittee of the American Academy of Neurology AIDS Task Force. Research criteria for diagnosis of chronic inflammatory demyelinating polyneuropathy. Neurology 1991; 41:617-8. 3. Albers JW, Kelly JJ. Acquired inflammatory demyelinating polyneuropathy: Clinical and electrodiagnostic features. Muscle Nerve 1989; 12: 435. 4. Antoine JC, Mosnier JF, Lapras J et al. Chronic inflammatory demyelinating polyneuropathy associated with carcinoma. J Neurol Neurosurg Psychiatr 1996; 60: 188-90. 5. Bird JS, Brown MJ, Shy ME et al. Chronic inflammatory demyelinating polyneuropathy associated with malignant melanoma. Neurology 1996; 46: 822-4. 6. Weiss MD, Luciano CA, Semino-Mora C et al. Molecular mimicry in chronic inflammatory demyelinating polyneuropathy and melanoma. Neurology 1998; 51: 1738-41. 7. Noronha AB, Harper JR, Ilyas AA et al. Myelin-associated glycoprotein shares an antigenic determinant with a glycoprotein of melanoma cells. J Neurochem 1986: 47:1558-65. 8. TsuchidaT, Saxton RE, Morton DL et al. Gangliosides of human melanoma. Cancer 1989; 63: 1166-74. 9. Hersey P. Active immunotherapy with viral lysates of micrometastases following surgical removal of high-risk melanoma. World J Surg 1992; 16: 251-60. 10. Fuller GN, Spies JM, Pollard JD et al. Demyelinating polyneuropathies trigerred by melanoma immunotherapy. Neurology 1994; 44: 2404-5. 11. Saleh MN, Khazaeh MB, Wheeler RH et al. Phase I trial of the munne monoclonal anti-GD2 antibody 14G2a in metastatic melanoma. Cancer Res 1992; 52: 4342-7. 12. Yuki N, Yamada M, Tagawa Y et al. Pathogenesis of the neurotoxicity caused by anti-GD2 antibody therapy. J Neurol Sci 1997; 149: 127-30. 13. Sabatelli M, MignognaT, Lippi G et al. Interferon-a may benefit steroid unresponsive chronic inflammatory demyelinating poly-

14. 15. 16.

17. 18. 19. 20. 21. 22.

neuropathy (Letter). J Neurol Neurosurg Psychiatr 1995; 58: 638-9. Gorson KC, Ropper AH, Clark BD et al. Treatment of chronic inflammatory demyelinating polyneuropathy with interferon-a-2a. Neurology 1998; 50: 84-7. Scalzo S, Genaro A, Boccoli G. Primary hypothyroidism associated with interleukin-2 and interferon-a-2 therapy of melanoma and renal carcinoma. Eur J Cancer 1990; 26. 1152-6. Sauter NP, Atkins MB, Mier JW et al. Transient thyrotoxicosis and persistent hypothyroidism due to acute autoimmune thyroiditis after interleukin-2 and interferon-a therapy for metastatic carcinoma: A case report. Am J Med 1992; 92: 441-4. Fabris P, Betterle C, Greggio NA et al. Insulin-dependent diabetes mellitus during alpha-interferon therapy for chronic viral hepatitis. J Hepatol 1998; 28: 514-7. Bora I, Karli N, Bakar M et al. Myasthema gravis following IFN-a-2a treatment Eur Neurol 1997; 38: 68. Dalekos GN, Hatzis J, Tsianos EV. Dermatologic disease during interferon-a therapy for chronic viral hepatits. Ann Intern Med 1998; 128: 409-10. Smedley H, Katrak M, Sikora K. Neurological effects of recombinant human interferon. BMJ 1983; 286: 262-4. Rutkove SB. An unusual axonal polyneuropathy induced by lowdose interferon-a-2a. Arch Neurol 1997; 54: 907-8. Marzo E, Tintore M, Fabregues O et al. Chronic inflammatory demyelinating polyneuropathy during treatment with interferon-a. J Neurol Neurosurg Psychiatr 1998; 65: 604.

Received 12 January 2000; accepted 19 April 2000.

Correspondence to:

Dr D. A. Anthoney CRC Department of Medical Oncology Beatson Oncology Centre Western Infirmary Dumbarton Road Glasgow G116NT UK E-mail: [email protected]