CANVAS-related RFC1 mutations in patients with immune-mediated neuropathy

Cerebellar ataxia, neuropathy, and vestibular areflexia syndrome (CANVAS) has recently been attributed to biallelic repeat expansions in RFC1. More recently, the disease entity has expanded to atypical phenotypes, including chronic neuropathy without cerebellar ataxia or vestibular areflexia. Very recently, RFC1 expansions were found in patients with Sjögren syndrome who had neuropathy that did not respond to immunotherapy. In this study RFC1 was examined in 240 patients with acute or chronic neuropathies, including 105 with Guillain-Barré syndrome or Miller Fisher syndrome, 76 with chronic inflammatory demyelinating polyneuropathy, and 59 with other types of chronic neuropathy. Biallelic RFC1 mutations were found in three patients with immune-mediated neuropathies, including Guillain-Barré syndrome, idiopathic sensory ataxic neuropathy, or anti-myelin-associated glycoprotein (MAG) neuropathy, who responded to immunotherapies. In addition, a patient with chronic sensory autonomic neuropathy had biallelic mutations, and subclinical changes in Schwann cells on nerve biopsy. In summary, we found CANVAS-related RFC1 mutations in patients with treatable immune-mediated neuropathy or demyelinating neuropathy.


Genetic analyses
DNA was extracted from peripheral blood using the Qiagen Kit (Qiagen, Hilden Germany).Primers used for the amplification of the short range of the repeat region in RFC1 were as described.When no normal size band was detected, the sample was subjected to repeat-primed polymerase chain reaction (PCR) for AAGGG (pathogenic), ACAGG (pathogenic), AGGGC (pathogenic), AGAGG (possibly pathogenic), AAGGC (possibly pathogenic), AAAGG (variable penetrance), AAAAG (likely non-pathogenic), AAA GGG (likely non-pathogenic), and AAGAG (likely non-pathogenic) repeat configurations 1,2,4 .Primers used for repeat-primed PCR are described in supplemental material.All patients provided written informed consent for genetic analyses.All methods in this study were performed in accordance with the Declaration of Helsinki.The study protocol was approved by the Institutional Review Boards of Kagoshima University and Kindai University.

Autoantibody measurement
Anti-ganglioside antibodies were examined in patients with GBS, MFS, CIDP, and IgM paraproteinemic neuropathy using GM1, GM2, GM3, GD1a, GD1b, GD3, GT1b, and GQ1b as reported previously 12,13 .Anti-GalNAc-GD1a antibody was additionally examined in patients with GBS and MFS, and anti-GT1a antibody was examined when anti-GQb1 antibody was detected 12,13 .Anti-MAG antibody and glycolipid sulfoglucuronyl paragloboside (SGPG) antibody, cross-reactive to MAG, were examined in patients with IgM paraproteinemic neuropathy as reported previously 5 .All methods in this study were performed in accordance with the Declaration of Helsinki.The study protocol was approved by the Institutional Review Boards of Kindai University.

RFC1 analyses
RFC1 mutations were found in three patients with immune-mediated neuropathy, such as GBS, MAG neuropathy, and idiopathic sensory ataxic neuropathy and one patient with sensory autonomic neuropathy (Fig. 1).No expansion of AGGGC, AGAGG, AAGGC, AAAGG, AAAAG, AAA GGG , or AAGAG was found in the four patients.No mutations were found in the control group.

Clinical information of patients with RFC1 mutations
Clinical and genetic information of patients with RFC1 mutations is summarized in Table 1.Detailed clinical information is provided upon request.Results of NCS are summarized in Supplementary Table 1.

Nerve biopsy findings of patient 4
Sural nerve biopsy revealed loss of myelinated and unmyelinated nerves (Fig. 2A and B), a typical finding for CANVAS-related neuropathy 1 .Amyloidosis was excluded by Congo red staining (not shown).An electron microscopic image revealed many collagen pockets (Fig. 2C).Notably, electron microscopic analysis of Schwann cells revealed cytoplasmic inclusion bodies, dense material, or accumulated membranous material (Fig. 2D-G).No apparent abnormality was found in the vascular systems (Fig. 2H).

Discussion
We found biallelic RFC1 mutations in three patients with immune-mediated neuropathy and one with nonimmune mediated neuropathy.This has been the first study to show an association between repeat expansions in RFC1 and treatable immune-mediated neuropathy.Patient 1 had typical monophasic GBS with anti-ganglioside antibodies wherein weakness was completely resolved after IVIG, but dysesthesia in the upper limbs remained 3 years after disease onset.Patient 2, who suffered from idiopathic sensory ataxic neuropathy with mild motor deficit, had a recurrent and remission clinical course with some improvement in motor and sensory conditions after repeated IVIG treatment.Patient 3 had MAG demyelinating neuropathy with stable symptoms and decreased IgM levels after rituximab treatment.All three patients tested positive for autoantibodies.Although positivity for rheumatoid factor, IgM anti-IgG, found in patient 2 might not directly cause neuropathy, a previous study demonstrated that rheumatic patients with rheumatoid factor had significantly more frequent neuropathies than those without rheumatoid factor (83% vs. 44%) 14 .Common symptoms in the patients included sensory disturbances, similar to those in Patient 4 with chronic sensory autonomic neuropathy with mild motor deficit and in previous studies on CANVAS 2 .
Pathological studies for CANVAS have been limited, with electron microscopic analyses being reported in only one study 15 .While the reported electron microscopic findings were limited to descriptions regarding Schwann cells associated with unmyelinated axons, we found several rare findings concerning myelinating Schwann cells, with a cytoplasmic inclusion body, dense materials, or accumulated membranous materials.Although Patient 4 did not have electrophysiological evidence of demyelinating neuropathy, subclinical changes in Schwann cells may have occurred.In Patient 2, demyelination was suggested to have occurred in the sural nerve on biopsy at the early stage and in the tibial nerves on NCS at the late stage.In addition, Patient 3 with MAG neuropathy had demyelination in the median nerve on NCS.These findings may indicate that patients with RFC1 mutations occasionally develop demyelination neuropathy or Schwann cell damage.
A possible mechanism by which RFC1 mutations are associated with immune-mediated neuropathy includes the vulnerability of the nerves themselves or that of Schwann cells to autoantibodies.In fact, we showed subclinical abnormal findings in Schwann cells producing myelin as mentioned above.Another possible mechanism is the abnormal function of blood-nerve barrier, given that the blood-nerve barrier may protect the nervous system from toxic materials, including autoantibodies 16 .However, we found no abnormalities around blood vessels, the site of the blood-nerve barrier, though the absence of morphological changes cannot rule out its functional change.Alternatively, abnormal immune-response or production of autoantibodies may be promoted in such patients.For example, neuronal or myelin antigens, or an expanded repeat RNA, might be abnormally Table 1.Clinical and genetic information of patients with RFC1 mutations.NCS nerve conduction study, GBS Guillain Barré syndrome, ISAN idiopathic sensory ataxic neuropathy with mild motor deficit, MAGN Myelin-associated glycoprotein neuropathy, SAN sensory autonomic neuropathy with mild motor deficit, RF rheumatoid factor, -, absent; + , mild; + + , moderate; + + + , severe; nd not done, IVMP intravenous methylprednisolone therapy (1 g/day for 3 days), IVIG intravenous immunoglobulin therapy (0.4 g/kg/day for 5 days).§ Data was available only from her medical record, § § x means # of cycles of the treatment.www.nature.com/scientificreports/exposed as immunogens 17 .Despite these fascinating mechanisms, the association between RFC1 mutations and immune-mediated neuropathy remains inconclusive, because the relatively low incidence of mutations among a large cohort of studied patients may raise the possibility that the association is coincidental.Biallelic mutations may suggest loss of function in RFC1, which is involved in DNA repair 18 .A recent report describing expansion mutations in an allele and truncation mutations in another supports this mechanism 19 .Loss of DNA repair protein function in neuropathy and cerebellar ataxia is reminiscent of Aprataxin-related disorders 20 .However, a previous study did not find any evidence suggesting loss of DNA repair function of RFC1 in fibroblasts 1 .Our electron microscopic findings suggested cytoplasmic inclusions in Schwann cells, a situation similar to that of SCA2-related neuropathy, a gain of function disease 8 .Further accumulation of evidence is needed to clarify the pathomechanism using neuronal or glial cell models or animal models.
One limitation of this study is the small number of patients with immune-mediated neuropathy who were positive for RFC1 mutations.In addition, detailed pathological studies in patients with immune-mediated neuropathy were lacking, as mentioned above.
In summary, we found CANVAS-related RFC1 mutations in patients with treatable immune-mediated neuropathy or demyelinating neuropathy.Thus, immunotherapy should not be terminated solely based on the identification of RFC1 mutations; however, repeated immunotherapy to unresponsive patients should be avoided 10,21 .Nonetheless, future large-scale studies are needed before definitive conclusions can be established.

Figure 1 .
Figure 1.Repeat-primed PCR results for RFC1.Pathological repeats of AAGGG or ACAGG were expanded in patients with Guillain-Barré syndrome, idiopathic sensory ataxic neuropathy with mild motor deficit (ISAN), MAG neuropathy, or sensory autonomic neuropathy with mild motor deficit (SAN).

Figure 2 .
Figure 2. Sural nerve biopsy findings in Patient 4. (A) A light microscopic image showing marked loss of myelinated fibers, without amyloidosis or vasculitis.(B) A lager image showing marked loss of both large and small myelinated fibers.(C) An electron microscopic image showing many collagen pockets (black arrow), reflecting unmyelinated fiber damage.(D) A Schwann cell had cytoplasmic dense materials (blue arrow).(E) Another Schwann cell had a cytoplasmic inclusion (yellow arrow).(F) The other Schwann cell had a cytoplasmic inclusion similar to the one shown in E (yellow arrow).(G) Membranous materials were seen between myelin sheaths (white arrow) or between a myelin sheath and an axon (white arrowhead).(H) Cells that constitute vessels had no abnormality without aggregation.