Objective X linked Charcot-Marie-Tooth disease (CMTX) is a hereditary neuropathy caused by mutations in GJB1 coding for connexin-32, a gap junction protein expressed in Schwann cells, but also found in oligodendrocytes. Four patients with CMTX developing central nervous system (CNS) demyelination compatible with multiple sclerosis (MS) have been individually published. We presently sought to systematically investigate the relationship between CMTX and MS.
Methods Over 20 years, 70 consecutive patients (36 men) with GJB1 mutations were identified at our Neurogenetics Unit, Athens, Greece, and assessed for clinical features suggestive of MS. Additionally, 18 patients with CMTX without CNS symptoms and 18 matched controls underwent brain MRI to investigate incidental findings. Serum from patients with CMTX and MS was tested for CNS immunoreactivity.
Results We identified three patients with CMTX who developed clinical features suggestive of inflammatory CNS demyelination fulfilling MS diagnostic criteria. The resulting 20-year MS incidence (4.3%) differed significantly from the highest background 20-year MS incidence ever reported from Greece (p=0.00039). The search for incidental brain MRI findings identified two CMTX cases (11%) with lesions suggestive of focal demyelination compared with 0 control. Moreover, 10 cases in the CMTX cohort had hyperintensity in the splenium of the corpus callosum compared with 0 control (p=0.0002). No specific CNS-reactive humoral factors were identified in patients with CMTX and MS.
Conclusions We have demonstrated a higher than expected frequency of MS in patients with CMTX and identified incidental focal demyelinating lesions on brain MRI in patients with CMTX without CNS symptoms. This provides circumstantial evidence for GJB1 mutations acting as a possible MS risk factor.
- multiple sclerosis
- CNS demyelination
Statistics from Altmetric.com
Charcot-Marie-Tooth disease (CMT) is a collective term encompassing a group of inherited demyelinating or axonal neuropathies with a broad range of phenotypes, inheritance patterns and causative genes. It is one of the most common inherited neuromuscular disorders with a prevalence of around 1 in 2500.1 Multiple sclerosis (MS) is the most common acquired inflammatory demyelinating disease of the central nervous system (CNS) with a prevalence exceeding 1 in 1000 in most Western populations.2
X linked CMT (CMTX) constitutes around 8% of the CMT spectrum.1 It is characterised by an age-related loss of myelin fibres and inappropriately thin myelin sheaths in peripheral nerves, suggestive of chronic demyelination.3 CMTX is caused by mutations in GJB1 coding for connexin-32 (Cx32). Cx32 is a gap junction protein expressed in peripheral Schwann cells, but also found on oligodendrocytes within the CNS.3
On rare occasions, symptomatic or asymptomatic CNS involvement has been reported in patients with different forms of CMT.4 5 In the case of CMTX, subclinical CNS involvement, documented on brain MRI or evoked potentials, is not uncommon.6 7 This usually takes the form of transient confluent symmetric white matter lesions (WML) with posterior predominance on T2 MRI sequences, but can also be detected as more permanent diffusor tensor imaging abnormalities.5 Less commonly, clinical involvement ranging from extensor plantars to acute transient encephalopathy can be observed in patients with CMTX.7 8
Over the past 20 years, several reports have linked CMT with MS in individual patients. In the case of CMT1A, by far the most common form of CMT, four cases with concomitant MS have been reported.9–11 There have also been single case reports of MS in other rare forms of CMT.4 12–14 Despite the relative rarity of CMTX, there have been at least four reports to date of individual patients who developed CNS demyelinating disease compatible with a diagnosis of MS.15–18 This has raised the possibility of a true association between CMTX and MS, beyond their co-occurrence strictly by chance.
In the present study we sought to systematically investigate the relationship between CMTX and MS. We approached this by establishing the frequency of MS in a consecutive cohort of Greek patients with CMTX, and by assessing the frequency of asymptomatic WML suggestive of demyelination on brain MRI of patients with CMTX without symptoms of MS or other CNS disease.
The Greek CMTX cohort
The Neurogenetics Unit at the 1st Department of Neurology, Eginition Hospital, National and Kapodistrian University of Athens, has provided molecular diagnostic testing for suspected CMTX to patients with hereditary neuropathies from all regions of Greece, acting effectively as a national reference centre. In total, 70 patients with mutations in GJB1 have been identified since 1997 with Sanger sequencing.19 These patients have originated from 32 families harbouring 22 different GJB1 mutations.
Screening for MS in patients with CMTX
Since the earliest GJB1 mutations identified in Greek patients, we have shown an active interest in investigating possible CNS involvement in these cases.8 18–21 Over the years, out of 70 patients with CMTX, 64 were examined at the Neurogenetics Unit and directly assessed for symptoms and signs suggestive of MS. For the remaining six patients medical case notes were reviewed for evidence of CNS involvement. All patients with CMTX with clinical features suggestive of MS were further investigated with brain MRI and, where possible, with lumbar puncture for the detection of cerebrospinal fluid (CSF) oligoclonal bands (OCB).
Brain MRI in patients with CMTX and matched controls
Eighteen patients with CMTX (recently enrolled in a neuropsychological study) and 18 age and gender demographically matched healthy controls were included in the study. All participants underwent the same imaging protocol on a 3.0T Achieva TX Philips manufactured MRI scanner (Philips, Best, The Netherlands) equipped with an eight-channel head coil. The imaging protocol included a T1 weighted spin echo (SE) sequence, high-resolution T2 (3D-T2) weighted fluid attenuated inversion recovery (FLAIR) sequence, T2 weighted turbo SE sequence and T2 weighted double inversion recovery (DIR) sequence. We excluded patients fulfilling MS diagnostic criteria, patients with current CNS symptoms and patients with concomitant medical conditions that could affect cognitive function (given that these patients were also enrolled in a neuropsychological study of CMTX). Assessment of MRI images in patients and controls was undertaken by an experienced neuroradiologist (GV) and reviewed by an MS specialist (GKou). Findings were classified into four categories: (1) lesions suggestive of focal demyelination (termed MS-like T2 lesions), partly fulfilling the criteria of radiologically isolated syndrome (RIS), that is, ovoid, well-circumscribed and homogeneous focal T2 lesions measuring >3 mm22; (2) non-specific focal WML; (3) diffuse hyperintensity of the splenium of the corpus callosum (CC); and (4) diffuse white matter hyperintensity, most prominent posteriorly.
Serological investigation of patients with CMTX and MS
Sera from two patients with CMTX and MS (cases 1 and 2) were further investigated for evidence of humoral autoimmunity. More specifically, we placed the patients’ sera (dilution 1/10) on monkey brain cerebellum frozen tissue (Euroimmun, Lubeck, Germany; FA1111-1005), followed by incubation with a secondary FITC-labelled antihuman IgA GM antibody for visualisation under a fluorescence microscope in order to search for any specific antibodies binding to brain tissue antigens. In addition, we searched for antibodies against the specific antigens: MOG (according to Waters et al),23 AQP4 (as described by the manufacturer: Euroimmun; kit FA 1128-50), contactin-1, neurofascin-155 and neurofascin-186 (referral lab: Clinical Immunological Laboratory, Euroimmun; immunofluorescence with cells transfected with contactin-1, neurofascin-155 and neurofascin-186). Finally, we searched for antibodies against the gangliosides GD1a, GD1b, GD2, GD3, GM1, GM2, GM3, GM4, GQ1b, GT1a, GT1b and sulfatide using dot blot (as described by the manufacturer; GA Generic Assays, Berlin, Germany; number 5003).
Data were analysed using contingency tables, risk estimate, Χ2 (with Yates correction where appropriate), Fisher’s exact test and t-test. Statistical analysis was performed using SPSS (V.20.0).
Standard protocol registrations and patient consents
All patients gave written informed consent for molecular genetic testing and inclusion in the present study, according to the Declaration of Helsinki.
Patients with CMTX fulfilling diagnostic criteria for MS
From a total cohort of 70 consecutive patients diagnosed with CMTX over 20 years, three fulfilled diagnostic criteria for MS.24 This represents a minimum of three incident cases over a 20-year period. The highest MS incidence ever reported from Greece in an epidemiological study was in 2008.25 The authors reported a mean annual incidence of 9.137/100 000 for the period 1997–2006 in Western Greece. This translates into 18 incident cases of MS/10 000 population over a 20-year period, a figure significantly different from the three incident cases out of 70 observed in the CMTX cohort (p=0.00039, Fisher’s exact test). The relative risk can be calculated at 23.8 (95% CI 7.2 to 79.0).
The first case (table 1) was a 52-year-old man with a family history of CMT and pes cavus since childhood, but no other significant early symptoms.18 Nerve conduction studies confirmed a neuropathy with intermediate velocities. He carried a c.191G>A (p.Cys64Tyr) point mutation in GJB1. At age 46, he developed optic neuritis (ON). Brain MRI revealed multiple periventricular, subcortical, juxtacortical, callosal and brainstem T2 hyperintense lesions, some gadolinium enhancing, strongly suggestive of MS (figure 1A–E). Visual evoked potentials were prolonged ipsilaterally. CSF analysis revealed no OCB and normal IgG index. The ON remitted following corticosteroid treatment. In the following year he relapsed twice (ON; gait unsteadiness) and MRI showed further evidence of disease activity. He was successfully treated with steroids and started on prophylactic treatment with beta-interferon. Despite this, within 2 years he further relapsed twice (spinal cord sensory syndrome; vertigo and gait unsteadiness) and had further MRI evidence of disease activity (including a cervical T2 lesion compatible with demyelination on spinal cord MRI). He was again successfully treated with steroids and switched to natalizumab. He has remained relapse free since. On most recent follow-up (54 years old), he had difficulty with tandem gait, saccadic pursuit, extensor plantars, brisk upper limb tendon reflexes, reduced knee jerks, absent Achilles reflexes, mild peripheral atrophy in upper and lower limbs, reduced vibration sense peripherally in the lower limbs, pes cavus and hammer toes. Neuropsychological testing revealed deficits in verbal working memory, deficits in episodic memory (impaired encoding, but preserved consolidation of novel verbal information), marginally impaired phonemic fluency and intact executive functions.
The second case (table 1) was a 47-year-old woman with a family history of CMT, who reported poor performance in sports and difficulty running since childhood. Symptoms followed a slowly progressive course and by her early 40s she had a characteristic steppage and slightly ataxic gait. Nerve conduction studies confirmed a severe neuropathy with intermediate velocities. She carried a c.462T>G (p.Tyr154stop) point mutation in GJB1. At age 45 she developed an acute deterioration in gait, which responded to oral corticosteroids. A few months later she relapsed with a brainstem syndrome characterised by vertigo, vomiting and severe gait unsteadiness. Brain MRI revealed multiple periventricular, subcortical, juxtacortical, callosal, brainstem and cerebellar T2 lesions, some gadolinium enhancing, strongly suggestive of MS (figure 1F–J). Thoracic spinal cord MRI revealed a gadolinium-enhancing T2 lesion at T8-9 compatible with demyelination. CSF analysis revealed the presence of OCB and a raised IgG index. Acute symptoms remitted fully following intravenous steroids. Six months later she relapsed further with vertigo, unilateral facial numbness, unilateral reduction in acoustic acuity and severe gait unsteadiness requiring bilateral support, and had further evidence of disease activity on MRI. She was again treated successfully with steroids. At this point a decision was made to start natalizumab, as per rapidly progressive relapsing MS. She has remained relapse free since. On most recent follow-up (48 years old), she required unilateral assistance to walk, had saccadic pursuit, extensor plantars, reduced upper limb tendon reflexes, brisk knee jerks, absent Achilles reflexes, severe peripheral atrophy in upper and lower limbs, was quadriparetic and had reduced vibration sense peripherally in the lower limbs, pes cavus and hammer toes. Neuropsychological testing revealed impaired verbal and visuospatial working memory, deficits in episodic memory (impaired encoding, but preserved consolidation of novel verbal information) and executive deficits, that is, low processing speed and impaired cognitive flexibility.
The third case (table 1) was a 64-year-old man with a family history of CMT and difficulty in fast walking and running since childhood. Nerve conduction studies confirmed a neuropathy with intermediate velocities. He carried a c.164C>T (p.Thr55Ile) point mutation in GJB1. He reported in his 40s several episodes of limb numbness and/or weakness and gait unsteadiness with acute onset and several days to weeks’ duration that remitted spontaneously, highly suggestive of CNS demyelinating attacks. Brain MRI revealed multiple periventricular, subcortical and juxtacortical T2 lesions, strongly suggestive of MS. Brainstem auditory evoked potentials were prolonged bilaterally. CSF analysis revealed the presence of OCB and a raised IgG index. On most recent follow-up (66 years old), he had difficulty with tandem gait and walking on heels, saccadic pursuit, positive Romberg’s sign, severe upper and lower limb peripheral muscular atrophy, upper and lower limb peripheral weakness, absent tendon reflexes, equivocal plantars, bilateral dysmetria, fine postural tremor, distal lower limb loss of vibration and bilateral pes cavus with hammer toes. Neuropsychological testing was not performed.
Other notable cases
There are a further two CMTX cases that presented in this 20-year period and are worth drawing attention to as providing supporting evidence for the possibility of MS-like CNS demyelination in these patients. The first case was a 21-year-old man, one of two siblings, who presented with episodes of generalised weakness and dysarthria lasting hours to days. He had brain MRI showing transient bilateral symmetrical confluent white matter T2 lesions, often reported in patients with CMTX.7 8 Of note, he had positive OCB in his CSF, indicative of an intrathecal immune response.20 Both siblings carried the c.164C>T (p.Thr55Ile) mutation in GJB1. The second case was a 15-year-old male adolescent who had on examination asymmetrical pyramidal signs and on brain MRI periventricular T2 lesions suggestive of focal demyelination.21 It was not possible to examine the CSF for OCB in this case. He carried a c.619A>G (p.Asn175Ser) in GJB1.
Incidental brain MRI findings in patients with CMTX and controls
Table 2 summarises the incidental brain MRI findings in a cohort of 18 patients with CMTX and no evidence of CNS symptoms. Characteristic examples of these findings are shown in figure 2. Table 3 summarises the differences in incidental brain MRI findings between patients and controls. Two patients had silent lesions suggestive of focal demyelination: a 44-year-old man with a sizeable (1.02 cm max diameter) ovoid ipsilateral periventricular lesion (figure 2A), and a 61-year-old woman with a sizeable (1.15 cm max diameter) ovoid subcortical lesion (figure 2B). No controls had equivalent findings, but the difference was not significant. Non-specific WMLs (figure 2C–E) were more common in patients than controls, but again this did not reach statistical significance. In total, 10 patients displayed characteristic hyperintensity of the splenium of the CC on T2 FLAIR and DIR sequences (figure 2F–H), a finding absent from controls and significantly different (p=0.0002; table 3). Finally, diffuse white matter hyperintensity with posterior predominance (figure 2I,J) was more common in patients compared with controls, but again this was not statistically significant. Comparing MRI findings in male and female patients with CMTX, CC hyperintensity and diffuse white matter hyperintensity were more common in men than women (p=0.003 and p=0.013, respectively; Fisher’s exact test), whereas non-specific WMLs were similar (p=1.000; Fisher’s exact test) (see also table 2).
Serological investigation of patients with CMTX with MS
We did not observe any specific binding of the two patients’ sera (cases 1 and 2) on white and grey matter of monkey brain cerebellum frozen tissue. In addition, there was absence of anti-MOG and anti-AQP4 antibodies in these sera using cell-based assays as described above. Searching for antibodies to contactin-1, neurofascin-155, neurofascin-186 and 12 gangliosides, none were detected in either patient.
This study presents circumstantial evidence for an association between CMTX and MS. The evidence rests on an observed increased incidence of CNS inflammatory demyelinating disease fulfilling diagnostic criteria for MS in a Greek CMTX cohort (3/70), compared with background incidence; on supportive evidence of an intrathecal immune response in a patient with CMTX with CNS symptoms and of brain demyelinating lesions in a patient with CMTX with asymmetric CNS signs, neither of which fulfilled criteria for MS; and on the identification of incidental brain lesions suggestive of demyelination in a cohort of patients with CMTX without evidence of CNS symptoms.
One potentially significant aspect of this report lies in the putative identification of a novel risk factor predisposing to the development of MS, the most common CNS inflammatory demyelinating disease. There are few undisputed risk factors that lead to substantial increases in MS risk. These include human leucocyte antigen status, geographic latitude, low vitamin D levels, history of infectious mononucleosis, smoking and adolescent obesity. These factors, all common, roughly convey independent 1.5-fold to 3-fold increases in relative risk.26 The identification of a potential new risk factor, although rare, is, therefore, of some importance. Furthermore, although this could clearly be an overestimation, CMTX seems to increase the relative risk of MS substantially more than established risk factors. Much larger CMTX cohorts would, however, be necessary to allow a more accurate calculation of relative risk. Nevertheless, given the rarity of CMTX, GJB1 mutations are not expected to play a crucial role in the bigger picture of MS, their overall impact as a risk factor being considerably lower.
The association between CMTX and MS, if confirmed in further studies, would be important for understanding the interaction between CMTX and MS, although such findings would not be easily translated to the bigger picture of MS, given the overall rarity of CMTX. A superimposed inflammatory demyelinating neuropathy has been described in patients with CMT1A and CMTX, and is probably more frequent than expected by chance.27 This observation suggests that inherited peripheral demyelination may trigger an autoimmune reaction against peripheral myelin. It is intriguing to hypothesise an equivalent phenomenon directed against CNS myelin in patients carrying Cx32 mutations. The mechanism may include the exposure of myelin antigens to which central tolerance has not been induced, which can then be presented to autoreactive T cells and trigger an autoimmune reaction against CNS myelin.28 We did not find any evidence of a specific humoral factor directed against CNS antigens in the sera of patients presently reported. AQP4 and MOG antibodies were absent.29 Antibodies against contactin-1, neurofascin-155 and neurofascin-186, which could potentially be associated with peripheral and central demyelination, were also not identified.30
What cannot be determined from available data is whether the hypothesised immune triggering happens in the periphery and affects the CNS following activated immune cell migration or is actually instigated within the CNS itself, as a result of pre-existing CNS involvement in CMTX. It is intriguing that for all three mutations presently reported (Cys64Tyr, Thr55Ile and Tyr154stop) there is evidence from previous reports and from the present MRI screening of CNS involvement.8 19 31 Furthermore, one of the published cases of CMTX and MS carried the Arg164Trp mutation, which also has established CNS manifestations.32 33 No specific data on CNS involvement were found for the other two mutations identified in the literature in patients with CMTX and MS (Ser26Leu and Tyr211His).16 17 Pertaining to the above, it is noteworthy that patients with CMT1A, a much more common form of CMT, have only rarely been reported to develop MS.9–11 CMT1A, caused by duplications of PMP22, is generally considered not to have CNS manifestations, whether clinical or subclinical.5 34 There is, however, some recent limited evidence to the contrary.35
It is interesting that studies of gap junctions in postmortem brain samples from patients with MS have reported that Cx32 is reduced in chronic MS plaques and in the normal appearing white matter, suggesting that loss of Cx32 may be implicated in MS progression.36 Similar findings have been reported in the experimental autoimmune encephalomyelitis (EAE) model of MS.36 In fact, Cx32 knockout mice have a worse clinical and pathological outcome of EAE compared with wild-type mice, displaying a significantly higher degree of demyelination and axonal loss in lumbar spinal cord compared with wild-type mice, an observation particularly pertinent to the present study.37
Presently, we have also demonstrated that patients with CMTX without current CNS symptoms have a high frequency of incidental white matter abnormalities on brain MRI, which, at least in the case of hyperintensity in the splenium of the CC, is significantly increased, compared with controls. Regarding focal lesions suggestive of demyelination, we have been guided by studies on patients with RIS, which suggest that ovoid, well-circumscribed, focal T2 lesions measuring >3 mm have prognostic significance.22 In general, white mater abnormalities seen in patients with CMTX can be broadly classified into two groups: first, localised lesions more or less suggestive of focal demyelination; and second, diffuse lesions primarily affecting the posterior white matter and often specifically involving the splenium of the CC. Both these abnormalities have been previously noted in individual case reports of patients with CMTX.20 38 However, rather surprisingly, no systematic study of conventional brain MRI on patients with CMTX has been reported to date. Recently, a diffusion tensor imaging study demonstrated white matter abnormalities in a series of 11 patients with GJB1 mutations.5 Interestingly, these abnormalities were limited to men, a finding only partly corroborated by present conventional MRI data, showing that callosal and diffuse white matter hyperintensity was more common in men than women.
The pathogenesis of WMLs in CMTX has not been fully elucidated. It is well known that Cx32 is expressed by oligodendrocytes.3 The possibility of a dominant negative effect of mutant Cx32 on other connexins expressed by oligodendrocytes was originally postulated.3 However, early studies demonstrated that Cx32 mutants associated with CNS phenotypes were retained in the endoplasmic reticulum and Golgi apparatus of oligodendrocytes, suggesting a toxic gain of function.39 In fact, recent studies have found that under the influence of systemic inflammation transdominant effects of mutant Cx32 on other connexins can disrupt oligodendrocyte gap junctions, although oligodendrocyte dysfunction is further exacerbated by intracellularly retained mutants (eg, Thr55Ile).40 Nevertheless, other studies have recently observed that even mutations associated with a complete loss of Cx32 can produce CNS dysfunction.41 Although a specific genotype-phenotype correlation between Cx32 mutation location (extracellular, intracellular or transmembrane) and WMLs has not been observed, recent studies suggest that the ability of Cx32 mutants to form gap junction plaques and produce adequate levels of junctional coupling in oligodendrocytes may relate to CNS manifestations.42 To all of the above, the occasional triggering of immune-mediated demyelination is also speculated to contribute to WMLs, as previously discussed.30
A recent review has comprehensively documented the vast majority of CMTX cases with CNS involvement reported in the literature.43 It includes two of the four cases reporting inflammatory CNS demyelination compatible with a diagnosis of MS.15 18 We have carefully analysed all remaining 72 patients with CMTX with CNS manifestations listed by these authors, focusing on aspects such as distribution and nature of WMLs, evidence of active inflammation on MRI, presence of intrathecal antibody response, as well as nature and course of CNS symptoms. We could not identify other cases fulfilling MS diagnostic criteria. The vast majority of cases had transient CNS symptoms lasting usually a few hours, bilateral symmetric white matter hyperintensities, absence of gadolinium enhancement, no evidence of CSF oligoclonal bands and evidence of WML resolution within a few months, whenever checked.43 In fact, there is evidence that these transient WMLs do not involve demyelination, as suggested by increases rather than decreases in magnetisation to transfer ratio during the acute phase.7 The comprehensive list documented in the review by Wang and Yin43 includes a case with atypical small scattered WMLs, but no CNS symptoms, which resembles the second of our ‘other notable cases’.44
It is noteworthy that two of the present cases, as well as other two reported in the literature had, during CNS relapse, a clearly favourable response to acute treatment with high-dose intravenous steroids.16 17 This supports the inflammatory nature of disease attacks in these patients, also evidenced by the presence of gadolinium-enhancing lesions on MRI. Furthermore, three of the above patients had no evidence of inflammatory CNS disease activity (either clinically or on MRI) following relatively long-term treatment with natalizumab, suggesting that inhibiting lymphocyte migration into the CNS by blocking alpha4-integrin is an effective way of aborting CNS relapses in these patients.45 Such information is important to physicians treating patients with CMTX, who develop inflammatory CNS manifestations.
This study has important limitations, many of which relate to sample size. It would be particularly important to replicate the increased MS incidence in larger CMTX cohorts from other populations. It would also be important to screen larger numbers of patients with CMTX for incidental brain MRI findings. This should be complemented by a large age and sex-matched control group, given that previous reports suggest that up to 2.4% of normal controls may fulfil RIS criteria.22 46 A further limitation, unrelated to sample size, concerns not looking specifically for the presence of anti-Cx32 antibodies in the sera of patients with CMTX and MS. It should be noted however that previous searches for such antibodies in MS have been negative.47 Finally, based on recent studies suggesting that GJB1 promoter mutations may be a more common cause of CMTX than previously suspected, this study is limited by the absence of promoter screening.48 Notwithstanding the above limitations, it remains surprising that the association between CMTX and MS has not been previously noted in reports of larger cohorts of patients with CMTX.49 It is possible that the distinguishing factor is one of the current authors’ active interests in both CMT and inflammatory CNS demyelinating disease, a rare occurrence in our age of ever increasing subspecialising.
In conclusion, the present study presents circumstantial evidence for an association between CMTX and an inflammatory CNS demyelinating disease fulfilling diagnostic criteria for MS, suggesting that pathogenic mutations in GJB1 may act as a rare risk factor for the development of MS.
Contributors GKou conceived and designed the study, examined the patients clinically, reviewed the imaging findings, analysed the data and wrote the manuscript. MB contributed to study design, examined the patients clinically, analysed the data and revised the manuscript. GV contributed to study design, interpreted the imaging findings, analysed the data and revised the manuscript. JT contributed to the study design, performed and interpreted the immunological analysis, and revised the manuscript. CKar performed the genetic testing and revised the manuscript. EK designed and performed the imaging analysis, and revised the manuscript. DT, MA, EA, ME and CKil examined the patients clinically and revised the manuscript. DK contributed to the study design, performed and interpreted the neuropsychological testing, and revised the manuscript. CP contributed to the study design, interpreted the neuropsychological testing and revised the manuscript. MP conceived the study, contributed to the study design, examined the patients clinically and revised the manuscript. GKar conceived the study, contributed to the study design, performed the genetic testing and revised the manuscript.
Funding This work was partly supported by a grant from Teva Pharmaceuticals (grant number 12846, special account for research grants, National and Kapodistrian University of Athens).
Competing interests GK reports grants from Teva Pharmaceuticals and Genesis Pharma; personal fees from Novartis, Genesis Pharma, Sanofi-Genzyme and Teva Pharmaceuticals; non-financial support from Merck, Sanofi-Genzyme and Genesis Pharma; MB reports no disclosures; GV reports no disclosures; JT reports shares in a diagnostic laboratory (Tzartos Neurodiagnostics) in Athens; DK reports no disclosures; CK reports no disclosures; EK reports no disclosures; DT reports no disclosures; MA reports research grants from Biogen, Merck-Serono, Novartis, Teva, Bayer and Genzyme, as well as lecture-fees from Novartis, Teva, Biogen and Genzyme; EA reports research grants from Biogen, Merck-Serono, Novartis, and Sanofi-Aventis, as well as lecture-fees from Teva; M-EE reports consultation services and honoraria from Novartis, Biogen and Teva; CK reports research grants from Biogen, Novartis, Teva, and Merck-Serono; CP reports no disclosures; MP reports no disclosures; GK reports no disclosures.
Patient consent Obtained.
Ethics approval Eginition Hospital Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Anonymised data will be shared on request by any qualified investigator.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.