Follow-up and prognosis of the anti-LGI1 encephalitis with and without teratoma and anti-NMDAR encephalitis with teratoma
We followed the anti-LGI1 encephalitis patient with teratoma for five months; one patient in the NT group was lost to follow up after discharge, but the other eight were followed for an average of 26.1±12.0 months; and the three anti-NMDAR encephalitis patients for average 27.0±17.8 months. Teratoma was all asymptomatic in the four AE cases, which was diagnosed through routine gynecologic examination before admission in one patient and diagnosed because of aetiological screening for AE in the other three. After immunotherapy and removal of the teratoma, the teratoma associated anti-LGI1 and anti-NMDAR encephalitis patients recovered, becoming seizure-free and living independently, especially 67% of anti-NMDAR encephalitis recovered fully (mRS=0) without symptom. Similarly, immune therapy (MPT, IVIG, plasma exchange and/or immunosuppressant) was effective in 78% of the NT group; one patient died of pneumonia because of comorbid myasthenia gravis and one patient died of hepatic failure due to chronic schistosomiasis liver disease and hepatitis B, deteriorated after MPT followed by oral azathioprine. None of the patients relapsed within the follow-up period.
Discussion
Unlike anti-NMDAR encephalitis, anti-LGI1 encephalitis is a subtype of AE with heterogenous clinical manifestations, rarely comorbid with tumor[12, 14]. To our knowledge, ours is the first reported case of anti-LGI1 encephalitis with teratoma. By comparing this case with other cases of anti-LGl1 encephalitis without teratoma in our center, we have derived several possible diagnostic biomarkers of anti-LGI1 encephalitis in hopes of promoting its early diagnosis.
The mechanism of anti-LGI1 encephalitis is mediated by pathogenic auto-antibodies, which directly attack presynaptic and postsynaptic protein complexes[2]. Although limited data indicates the correlation of tumor and anti-LGI1 encephalitis, one study showed that less than 10% of patients have tumors, including thymoma, thyroid, lung and renal cell tumors[13, 14]. We think that anti-LGI1 encephalitis was associated with the teratoma in our case for several reasons: 1) IHC staining confirmed that partial or focal positive nuclear staining of LGI1 was appreciated in some tumor cells. Although by Karen Head et al’s work, LGI1 did not express in normal mice thyroid[17], we think the positive expression of LGI1 in the tumor cells of teratoma in our case is still reasonable, based on reasons below: a) considering the nature of benign germ cell tumor of struma-ovarii in this case, the neoplastic thyroidal follicular cells were different from cells in normal thyroid and may acquire the abnormal expression of LGI1 protein during the development of teratoma; b) based on the reported immunohistochemical images from the article by Karen Head et al[17], Fig3 B showed positive staining of LGI1 within the ovarian follicle-like structure, suggesting its possible expression in the germ cells. Although normal thyroid tissue derived from the endodermal layer typically showed negative LGI1 expression, the possibility of positive expression of LGI1 in some neoplastic thyroidal follicular cells within the struma-ovarii cannot be totally excluded. 2) pathological results indicated that teratoma tissue mainly consisted of thyroid gland, which is susceptible to immunological attack and easily promotes the generation of various antibodies as antigen[18] (additionally, thyroid cancer can be comorbid with anti-LGI1 encephalitis [13, 14]); 3) after removal of the teratoma, her cognitive function recovered greatly; and 4) many reports about AE or PLE associated with ovarian teratoma have been published since 1998, including Yang Y.W. et al ’s case of episodic dystonia highly consistent with characteristic FBDS described in anti-LGI1 encephalitis today[19-25].
The most representative and high-incidence subtype of AE associated with teratoma was anti-NMDAR encephalitis; NMDA receptor-expressing neurons have been described in the neural tissue within teratoma, which was pathogenetic of anti-NMDAR encephalitis [26]. Unlike NMDA receptors, which are the extracellular domain of the GluN1 subunit[27], LGI1 is not a structural component of a receptor or ion channel, but is rather secreted by neurons, perhaps explaining why the incidence of tumor is much lower in anti-LGI1 encephalitis compared with anti-NMDAR encephalitis. Nevertheless, LGI1 forms a trans-synaptic complex with presynaptic proteins and is involved in synaptic transmission of neuronal excitability[28], so anti-LGI1 encephalitis may still be comorbid with tumors such as teratoma.
Summary data from our case series indicates similar clinical manifestations as seen in many larger series. These specific manifestations are probably early diagnostic clues to anti-LGI1 encephalitis, including, 1) acute or sub-acute onset of RPD in relatively older aged population[3, 29, 30], 2) FBDS[3, 29, 31], 3) hyponatremia[3, 30] and 4) hyperintensity of bilateral hippocampus on T2 and T2-FLAIR MRI[3, 29].
More importantly, our anti-LGI1 encephalitis case with teratoma presented several specific features that might differentiate it from other subtypes of AE, but perhaps even other cases of anti-LGI1 encephalitis without teratoma, and might therefore be clues. First of all, her chronic onset of anxiety has never been reported in other subtypes of AE by our best knowledge. For anti-NMDAR encephalitis, 86% of 100 cases presented headache, fever of a non-specific viral-like illness as prodromal symptom in one case series report[32], and in another study anti-NMDAR encephalitis patients developed psychiatric symptom and short term memory loss in less than two weeks[33]. 77% of 22 anti-GABABR encephalitis patients presented isolated recurrent seizure at onset[34]. In 76 LGI1 antibody positive patients, only 2.6% presented isolated anxiety as initial symptom and the median duration time for isolated initial symptoms was 2 months[35].
Secondly, seizure is one of the most common symptoms of AE, but the clinical manifestation varies among different subtypes. The incidence of seizure in anti-NMDAR encephalitis and anti-GABABR encephalitis is the highest, followed by anti-LGI1 and anti-AMPA encephalitis. Whereas the most typical seizure type with diagnostic specificity is FBDS in anti-LGI1 encephalitis, seen in almost half of 39 confirmed cases, FBDS were not seen in our case with teratoma[3]. Anti-NMDAR encephalitis has more seizure types at onset; a study of seven cases found that 43% had GTCS, 43% had focal seizures and 14% had both simultaneously [36]. Seizure was relatively uncommon in anti-AMPA encephalitis (20% of patients), typically GTCS[37]. SE is seen more often in anti-GABA encephalitis, based on the limited literature: seen in 58% of one series of 12 anti-GABAAR encephalitis cases[38] and 27% in another series of 11 confirmed GABABR encephalitis subjects[39], compared with only 6% of 100 anti-NMDAR encephalitis patients[40] and 5% of 19 anti-LGI1 encephalitis patients[41]). Interestingly, our case with teratoma developed SE on her 8th day of admission. We think there are two interpretations, 1) SE may be more typical in anti-LGI1 encephalitis with teratoma and 2) the patient’s immunotherapy started relatively late, one month after onset of RPD, facilitating the development of SE.
Thirdly, the persistent short term memory impairment until 5 months following the acute disease stage in our patient was highly consistent with reported anti-LGI1 encephalitis. One study found 62% of 85 LGI1/CASPR2-IgG-positive patients with central involvement had residual cognitive and personality disturbances[15]. Another study revealed that 23% of 76 patients with LGI1 antibody positive had moderate or severe cognitive impairment at two years follow-up[35]. Recovery from cognitive and neuropsychological symptoms is quicker and more complete in other forms of AE. 75% of 100 confirmed patients with ani-NMDAR encephalitis showed full recovery (mRS 0, MMSE 29-30) or very mild functional deficits (mRS 1-2, MMSE 25-28) at median 17 months (range 1-194 months) follow-up[32]. This is consistent with our three anti-NMDAR encephalitis, which presented fully recovery in 67% of cases. Serial observation of patients with anti-GABABR encephalitis indicated that 35% recovered fully and 40% improved markedly, except for recurrent seizures in 50% of 20 cases comorbid with SCLC[42]. Our case suggested that the cognitive deficits in anti-LGI1 encephalitis might need longer time to recover than those in anti-NMDAR or GABABR encephalitis.
Overall, given our experience with our small anti-LGI1 encephalitis cohort, we suggest that physicians should consider the diagnosis of anti-LGI1 encephalitis, and screen for the existence of teratomas, in patients with chronic onset of mood disorders followed by RPD, SE and psychiatric symptoms. Early diagnosis is vital because immunotherapy was effective in anti-LGI1 encephalitis and if instituted early might avoid deterioration and need for ICU management. Further illumination of specific clinical biomarkers of anti-LGI1 encephalitis with teratoma will require collection of more cases.