Coeliac disease, epilepsy, and cerebral calcifications: association with TG6 autoantibodies


Dr Alexandra M Johnson, Neurology Department, Sydney Children’s Hospital, High Street, Randwick, NSW 2031, Australia. E-mail:


A 4-year-old boy presented with occipital seizures but normal initial neuroimaging and proved refractory to antiepileptic medications. On repeat neuroimaging after 1 year, he had developed bi-occipital calcification and was then found to have positive coeliac serology. He was diagnosed with coeliac disease, epilepsy, and cerebral calcifications (CEC) and became seizure free after starting the gluten-free diet. Positive antibody binding to neurons and glia was demonstrated on indirect immunofluorescence. High levels of immunoglobulin-A directed against transglutaminase isoenzyme 6 (TG6) were found in the patient’s serum. The positive response to the diet, TG6 antibodies, and neuronal antibody binding suggest that CEC might be autoimmune in nature, as in other extra-intestinal manifestations of gluten-related diseases, such as gluten ataxia.


Coeliac disease, epilepsy, and cerebral calcifications


Transglutaminase isoenzyme 6

What this paper adds

  •  An autoimmune mechanism may contribute to the pathogenesis of coeliac disease, epilepsy, and cerebral calcifications, with serum TG6 autoantibodies and neuronal antibodies found in this case.

Many neurological manifestations of gluten-related diseases have been shown to have an autoimmune basis, the best characterized being gluten ataxia. Antibodies to transglutaminase isoenzyme 6 (TG6) have previously been associated with gluten ataxia, with some evidence for a causative role.1,2

Coeliac disease with epilepsy and cerebral calcifications (CEC) is a rare neurological syndrome, characterized by medically resistant, occipital epilepsy. Previously, co-occurring folate deficiency in CEC has been thought to be contributory. We describe a patient with CEC responsive to the gluten-free diet and positive for TG6 autoantibodies. Informed consent to publish this report was obtained from the parents of the child.

Case Report

A 4-year-old boy presented with medication-resistant occipital epilepsy, with complex partial seizures every 2 to 3 days. Seizures often lasted over 30 minutes, with version, eye deviation, alteration of awareness, and amaurosis, followed by vomiting or secondary generalization. Antiepileptic drug trials included carbamazepine, sodium valproate, and levetiracetam, with only slight improvement in seizure frequency.

Initial investigations included normal cerebral computed tomography (CT) and magnetic resonance imaging (MRI). Video electroencephalography recorded seizures originating independently from bilateral occipital regions, with concordant ictal single-photon emission computed tomography (SPECT). Neuropsychological testing described average cognitive abilities overall, with attentional difficulties. Consistent with an occipital focus, there was a marked discrepancy between intact verbal skills and a weakness in visual–motor and visual–integration skills (below age-expected letter formation and word recognition/integration).

After 1 year, positron emission tomography (PET)/CT and repeat MRI (Fig. 1) demonstrated the development of bilateral calcifications over the parieto-occipital regions. Initial systems review had been unrevealing, and the patient was thriving. On specific questioning, diarrhoea had been present for several years, with watery stools and intermittent abdominal distension.

Figure 1.

 (a) Axial computed tomography scan of the patient’s head in July 2009 (age 4y 7mo) and (b) at diagnosis of coeliac disease, epilepsy, and cerebral calcifications (CEC) in September 2010 (age 5y 9mo) demonstrated development of calcification over 1 year (arrows). (c) Axial magnetic resonance imaging (Flash sequence) in September 2009 (age 4y 9mo) and (d) at diagnosis of CEC in October 2010 (age 5y 10mo) showed development of calcification bi-occipitally (arrows).

Routine coeliac antibodies were positive (endomysial immunoglobulin-A [IgA] antibody, gliadin IgA and IgG antibodies, transglutaminase 2 IgA and IgG antibodies) and endoscopy with biopsy confirmed subtotal villous atrophy. Decreased folate levels were noted (serum folate 4.2 nmol/L [normal 6.5–40.5], cerebrospinal fluid [CSF] methylenetetrahydrofolate 3.3 nmol/L [normal 40–120]). CSF pterins were not elevated and oligoclonal bands (IgG) negative. Testing serum for antibodies to different transglutaminase isoenzymes, as described in Hadjivassiliou et al.1, revealed high levels of anti-TG6 IgA (TG6 IgA 33 U/mL [normal 0–14], TG6 IgG 5.1 U/mL [normal 0–35]; commercial standards [Zedira, Darmstadt, Germany] were used to calculate units).

Indirect immunofluorescence using the patient’s serum and CSF was performed on monkey oesophagus to detect IgG and IgA endomysial antibodies, and monkey cerebellum/cerebrum to detect IgG neuronal antibodies (both INOVA Diagnostics, San Diego, CA, USA). Endomysial antibodies (IgG and IgA) were strongly positive in both serum (1:>640) and CSF (1:>40). When patient serum or CSF was applied to cerebellar and cerebral tissue, cytoplasmic immunofluorescence on neurons and glia was noted, not seen in controls with coeliac disease alone (Fig. 2).

Figure 2.

 (a) Indirect immunofluorescence using the patient’s sera (1:160) on monkey cerebellum showed strong coarse staining on the molecular layer (M), neuronal cells of the granular layer (G), white matter (W) tracts, and vessels (BV) (magnification ×200). (b) Indirect immunofluorescence using the patient’s sera (1:160) on monkey cerebrum showed non-specific diffuse staining (magnification ×200). (c, d) Comparison with serum (1:50) from a child with coeliac disease but without epilepsy showed staining of the blood vessels only in (c) the cerebellum and (d) the cerebrum (magnification ×200).

The patient was placed on a gluten-free diet and has remained seizure free for 18 months. Seizure control was achieved within 2 weeks of starting the diet and before supplementation with folate was begun. Antiepileptic drugs were withdrawn several months later. The family reported improvements in behaviour, reading, and writing though noted continued attentional difficulties requiring methylphenidate. Deterioration in seizure control was noted subsequently as the methylphenidate preparation contained gluten in the excipient. Freedom from seizures was achieved again once the medication was withdrawn.


The concept of gluten-related disease has rapidly expanded over recent decades from pure enteropathy as seen in coeliac disease, to involvement of skin in dermatitis herpetiformis and the cerebellum in gluten ataxia. Although the T-cell response in the gut mucosa is key to the development of coeliac disease, transglutaminase autoantibodies are increasingly recognized as playing a role in the extraintestinal pathophysiology of gluten-related diseases.3–5

Neurological manifestations have been reported in 10 to 22% of patients with coeliac disease and include gluten ataxia, polyneuropathy, myopathy, epilepsy, CEC, leukoencephalopathy, and headache.6 Such manifestations were initially assumed to be secondary to vitamin deficiencies due to malabsorption (e.g. folate, vitamins B12, D, and E). However, as neurological manifestations can arise without enteropathy, their occurrence may be better explained by immune mechanisms. In gluten ataxia, successful treatment by restriction of gluten7 or therapy with intravenous immunoglobulin8 is also supportive of an immune aetiology. Oligoclonal bands are present in the CSF of up to 50% of patients with gluten ataxia.7

Transglutaminases are calcium-dependent enzymes, found in the vasculature, gut, and brain. They catalyse post-translational modification of glutamine residues through isopeptide linkage, deamidation, or esterification. Transglutaminase 2 is the most widely expressed and abundant member of the transglutaminase family. It has roles in tissue repair, acute phase response, and vascular permeability. In coeliac disease, transglutaminase 2 is involved in the deamidation of gluten peptides, essential to the development of the T-cell response to gluten.3–5

Specific transglutaminase isoenzyme antibodies have been linked to particular extra-intestinal syndromes, the best characterized being transglutaminase 3 antibodies in dermatitis herpetiformis and TG6 antibodies in gluten ataxia.1 In gluten ataxia, cerebellar Purkinje cell depletion, perivascular leukocyte infiltration, and deposition of anti-TG6 IgA has been demonstrated on post-mortem tissue.1 An animal model has shown induction of ataxia upon injection of sera or isolated transglutaminase antibodies from patients with gluten ataxia or coeliac disease.2 Functionally, TG6 is able to perform the essential roles of transglutaminase in coeliac disease, with deamidation of human leukocyte antigen (HLA)-binding regions of gluten peptides. It also has the ability to complex with gluten peptides and form thioester- and isopeptide-linked complexes.5 This lends credence to a functional role for TG6 in autoimmunity in gluten sensitivity.

TG6 autoantibodies have been thought to be a more specific marker for the neurological manifestations, with the median TG6 antibody concentration being significantly higher than transglutaminase 2 antibody in patients with coeliac disease associated neurological complications.1 However, this has recently been questioned in one study. An unusually high level of patients who were positive for TG6 IgA (commercial assay) has been noted in comparison with gastrointestinal symptoms only.9 The respective commercial assay has also given many false-positive results in a blood donor collective in our hands, and we have subsequently verified gluten dependence of TG6 antibodies using a cohort with coeliac disease (D Aeschlimann, P Aeschlimann, M Hadjivassiliou, and K Kaukinen, personal communication, 2012). Methodological differences may therefore explain the differences between studies.

Our case highlights the fact that TG6 autoantibodies can also be found in CEC, something not previously reported. TG6 antibodies have been linked previously to gluten ataxia, stiff person syndrome, and gluten neuropathy. Furthermore, mutations in the gene encoding TG6 have been linked to another neurological disorder: spinocerebellar ataxia.10

Indirect immunofluorescence using sera from our patient showed antibody binding to glial and neuronal tissue, not previously shown in coeliac related epilepsy. Previous studies using coeliac disease sera have shown IgA-class antibody binding to cerebral blood vessels alone.2,11 This demonstration of serum neuronal antibodies in CEC further supports immune mediation in the pathogenesis of CEC.

CEC is a rare neurological disorder reported most commonly in Italian, Argentinian, and Spanish patients. Epilepsy onset occurs at a mean of 6 years (range 1–28y),12,13 most commonly with occipital seizures. Drug resistance is common (63% in the study by Gobbi12) with evolution towards epileptic encephalopathy. As with other extra-intestinal manifestations of gluten sensitivity, CEC is found in asymptomatic patients and those with typical coeliac disease.13 Restriction of dietary gluten has been previously shown to improve the seizures,13 and in our patient we saw a dramatic and rapid improvement.

In CEC, calcifications develop with time, with a mean of 1 year described between a normal scan and the appearance of calcifications.13 The cause of the calcification in CEC is unknown. There is thought to be no correlation between the extent of calcification and severity of the epilepsy.12 Persistence of the calcification after resolution of seizures on the gluten-free diet suggests this may be an epiphenomenon rather than being causative. Case reports exist of improvements in epilepsy on the gluten-free diet in patients with coeliac disease without calcifications, and auto-immunity has been postulated.14

Folate deficiency (serum and CSF) has been noted in many patients with CEC,15 though is not universal. Folate deficiency has been suspected to be the cause of CEC in previous cases because of their co-occurrence, and the link between folate deficiency and cerebral calcification in other situations (e.g. methotrexate treatment, primary cerebral folate deficiency). Our patient was noted to be systemically low in folate, but had a marked improvement in seizure control from the gluten-free diet alone, before supplementation with folate. Although folate deficiency may contribute to the pathogenesis of CEC, this timing implies that an alternate factor was the main stimulus for his epilepsy.

Although further studies are required, our case report suggests that CEC is an autoimmune process, similar to gluten ataxia. We found autoantibodies to TG6 in serology and IgG binding to cerebral and cerebellar neurons on immunofluorescence. This case demonstrates that autoimmunity to TG6 may be seen in other extra-intestinal manifestations of gluten-related disease affecting the central nervous system aside from gluten ataxia. The prompt response to the gluten-free diet and relapse on rechallenging with gluten also supports gluten as the trigger and driver of the neuroimmunological response in CEC.


Funding as a Neurology Fellow (to AMJ), came from Sydney Children’s Hospital Foundation. Funding for TG6 antibody (to MH) came from Coeliac UK and the Bardhan Research and Education Trust of Rotherham, UK.