Autoantibody to the Rab6A/Rab6B in Autoimmune Cerebellar Ataxia Associated with Sjogren’s Syndrome

Background To report a novel autoantibody against Purkinje cell in a patient with autoimmune cerebellar ataxia (ACA) associated to Sjogren’s syndrome (SS). The Patients on one centre with cerebellar ataxia of unknown cause, who were tested positive with tissue-based indirect immunouorescence assay (TBA) on rat cerebellum sections and negative for comprehensive anti-neural autoantibodies panel, were investigated for novel autoantibody identication. Among them, one patient with comorbid ACA and SS was qualied for further exploration. His-immunoprecipitation (HIP) combined with mass spectrometric (MS) analysis was used to identify the target antigen, which was conrmed by recombinant cell based assay (CBA) and antibody neutralization experiments. 5e4 and the The s to suppress the repeated detection of the same fragment ion peaks. The relative collision energy for MS2 was set at 27% for HCD. Raw MS data was analysed using the Proteome Discoverer 2.0 software (Thermo Fisher Scientic, USA) to identify the proteins and searched against Rattus norvegicus proteins from the UniProt sequence database. The search parameters were as follows: precursor ion mass tolerance ± 15 ppm; MS/MS tolerance ± 0.1 Da; 2 missed cleavages were allowed with the enzyme of trypsin; cysteine was set as xed modication of carbamidomethylation and methionine was set as variable modications of oxidation. The protein identication was supported by at least one unique peptide. The results were ltered based on a false discovery rate (FDR) of no more than 1%.


Abstract Background
To report a novel autoantibody against Purkinje cell in a patient with autoimmune cerebellar ataxia (ACA) associated to Sjogren's syndrome (SS).

Methods
The Patients on one centre with cerebellar ataxia of unknown cause, who were tested positive with tissuebased indirect immuno uorescence assay (TBA) on rat cerebellum sections and negative for comprehensive anti-neural autoantibodies panel, were investigated for novel autoantibody identi cation. Among them, one patient with comorbid ACA and SS was quali ed for further exploration. Hisimmunoprecipitation (HIP) combined with mass spectrometric (MS) analysis was used to identify the target antigen, which was con rmed by recombinant cell based assay (CBA) and antibody neutralization experiments.

Results
TBA of the patient's serum and cerebrospinal uid (CSF) for autoantibody testing revealed binding of IgG antibody, mainly IgG1, to Purkinje cell and granular layer of rat cerebellum. Rab6A was identi ed as the autoantigen by MS and Western blot, subsequently veri ed by CBA with HEK293 cells expressing human Rab6A/Rab6B. Furthermore, recombinant human Rab6A/Rab6B protein to neutralize the autoantibodies' tissue reaction was performed by a parallel con rmed approach.

Conclusion
Autoantibody against Rab6A/Rab6B may be a novel biomarker in diagnosis of ACA, especially in patients with comorbid ACA and SS. The role of the antibody in mechanism of ACA warrants further study.

Background
The disorder of suspected immune-mediated cerebellar ataxia without the identi cation of a well-known trigger or pathogenic neuronal antibody is refered to as autoimmune cerebellar ataxia (ACA) [1] . Some ACA patients have coexisting non-neurological autoimmune diseases, such as Sjogren's syndrome (SS), thyroid autoimmune diseases, and vitiligo, which suggest autoimmune tendency and served as an indication in diagnosis of ACA [1,2] . SS is an autoimmune chronic lymphocytic in ammatory disease involving the exocrine glands (ocular or salivary gland) in the setting of antinuclear antibodies, particularly to Ro/SSA and La/SSB [3] . It can occur alone as primary SS or in conjunction with other connective tissue diseases (secondary SS). Primary SS primarily affects exocrine glands, but may have extra-glandular manifestations, including the neurologic system, with the prevalence of 8-49% [4] .
Cerebellar ataxia is one of the described neurological manifestations [5] .
Antibody-mediated dysfunction is one of possible aetiologies of cerebellar ataxia related to primary SS. In 2001, Owada et al detected an antibody that reacted with a protein of 34 kDa from the extract of spinal cord, dorsal root ganglion or cerebellar cortex in a primary SS patient with motor weakness and cerebellar ataxia [6] . Here, we report the identi cation of a novel neural autoantibody against Rab6A/ Rab6B protein in a patient with comorbid ACA and SS.

Samples
The samples (sera) are from the patients of cerebellar ataxia of unknown cause who were registered to the program of encephalitis and paraneoplastic syndrome (PNS) of Peking Union Medical College Hospital (PUMCH) from July 2018 to November 2019. This study was approved by the Ethics Committee of PUMCH (JS-891 and JS-2184), and informed consent was obtained from each patient. As shown in Fig. 1, sera of enrolled patients were rstly tested by tissue-based indirect immuno uorescence assay (TBA) on rat cerebellum sections, positive sera were further screened for well-established anti-neural autoantibodies using recombinant protein, either immunoblot for intracellular or cell-based assay (CBA) for extracellular autoantibodies (including antibodies target to aquaporin 4, NMDA-R, CASPR2, AMPA-R, LGI1, GABAb-R, GAD-65, ITPR1, ZIC4, PKCγ, AP3B2, PCA-2, CARP VII, Homer-3, NCDN, CV2/CRMP5, PNMA2, Ri, Yo, Hu and amphiphysin). When a TBA-positive serum was negative for the above autoantibody screenings, the sample was investigated for new antibody identi cation. Among them, a 43year-old female was involved with clinically suspected primary SS combined with unexplained cerebellar ataxias. The patient experienced extensive tests, yet no infectious, metabolic nor genetic cause was revealing. However, oligoclonal bands of the patient's CSF was positive and detection of anti-SSA and anti-Ro antibodies strongly indicated immune-mediated pathogenesis. TBA test revealed strong IgG1 reactivity with cerebellar granular cell layers and Purkinje cell layers, but not with a broad panel of recombinant expressed anti-neural autoantibodies. Therefore, the patient (serum) was selected for novel antibody identi cation. Additionally, 10 age-and sex-matched SS patients without neurological manifestation were enrolled to identify. ZF-0308, ZSBio, China) and Alexa Fluor 555 labelled goat anti-rabbit IgG antibody (Cat. Ab150078, Abcam, UK) at room temperature for 30 min. The slides were then washed again, embedded in glycerol and observed by DMi8 microscope (Leica, Germany). For evaluation of IgG subclasses, patient serum was tested on rat cerebellum sections as described above, with the following modi cations applied: unconjugated sheep anti-human IgG antibodies speci c for IgG subclasses 1 to 4 (Nodics-Mubio, Netherlands, 1:100) were substituted for the FITC labelled goat anti-human IgG antibody, and AF568 labelled donkey anti-sheep IgG (Invitrogen; absorbed against human IgG, 1:200) was used to detect the subclass speci c antibodies (Fig. 3).

SDS-PAGE and Mass spectrometry (MS)
20 µl eluted proteins were electrophoresed on 12% SDS-PAGE gels. After visualizing with Coomassie Brilliant Blue G-250, the divisible protein bands were cut and destained with 25 mM ammonium bicarbonate/50% Acetonitrile. Proteins in the gel particles were extracted by PAGE gel protein extraction kit (Sangon Biotech, China). Dried extracts were resuspended in lysis buffer and subjected to online reverse phase nano LC-MS/MS analysis with 50% of sample loading using an Easy nLC 1000 (Thermo Fisher Scienti c,USA), coupled to a Q-Exactive Plus mass spectrometer (Thermo Fisher Scienti c,USA). Peptide samples were concentrated on a 2 cm trap column (100 µm diameter) and separated on a 12 cm capillary column (75 µm diameter), both packed in-house with 1.9 µm C18 reverse-phase fused silica (Michrom Bioresources, Inc., Auburn, CA). The samples were separated at a ow rate of 0.6 µL/min with a 71 min linear gradient from 5-30% mobile phase B (phase A: 0.1% formic acid in water, phase B: 0.1% FA in ACN), followed by a quick ramp from 30% mobile phase B to 95% mobile phase B within 1 min, where samples were held for 6 min before a quick ramp down; then, the column was re-equilibrated. Eluted peptides were analysed with a Q-Exactive Plus mass spectrometer (Thermo Fisher Scienti c, USA). The MS survey scan was analysed over a mass range of 300-1400 Da with a resolution of 70000 at m/z 200. The isolation width was 3 m/z for precursor ion selection. The automatic gain control (AGC) was set to 3e6, and the maximum injection time (MIT) was 60 ms. The MS2 was analysed using data-dependent mode searching for the 20 most intense ions fragmented in the HCD. For each scan with a resolution of 17500 at m/z 200, the AGC was set at 5e4 and the MIT was 80 ms. The dynamic exclusion was set at 18 s to suppress the repeated detection of the same fragment ion peaks. The relative collision energy for MS2 was set at 27% for HCD. Raw MS data was analysed using the Proteome Discoverer 2.0 software (Thermo Fisher Scienti c, USA) to identify the proteins and searched against Rattus norvegicus proteins from the UniProt sequence database. The search parameters were as follows: precursor ion mass tolerance ± 15 ppm; MS/MS tolerance ± 0.1 Da; 2 missed cleavages were allowed with the enzyme of trypsin; cysteine was set as xed modi cation of carbamidomethylation and methionine was set as variable modi cations of oxidation. The protein identi cation was supported by at least one unique peptide. The results were ltered based on a false discovery rate (FDR) of no more than 1%.

Recombinant Expression of Antigens in HEK293 Cells
For the preparation of substrates for the cell-based assay, HEK293 cells (CRL-1573

Clinical data
During July 2018 to November 2019, we collected sera from patients with unexplained cerebellar ataxia for TBA on rat cerebellum sections and by extensive anti-neural autoantibodies screening. Five of them were positive in TBA but negative in all established autoantibody screenings, and in the serum of one patient we identi ed the novel autoantibody (Fig. 1).
The patient is a 43-year-old woman with a medical history of elevated fasting glucose levels presented with progressive dizziness, unsteady gait, nausea and vomiting for two months. Several days before the onset of symptoms, she got injection of the yearly in uenza virus vaccine. She had feelings of dry eyes and dry mouth recently. Neurological examinations revealed intention tremor, dysdiadochokinesia, and abnormal heel-to-shin test bilaterally, with failure of heel-to-toe walking, and a wide-based, staggering gait. No nystagmus, weakness, or Babinski sign was noticed. Brain magnetic resonance imaging (MRI) showed abnormalities and atrophy in the bilateral cerebellum (Fig. 2). CSF analysis displayed in ammatory features (elevated cell count, protein levels and immunoglobulin index, and positive oligoclonal bands). Other clinical examination showed that unstimulated whole saliva ow rates and Schirmer's test were decreased, and ocular staining score was increased. Whole body (without head) PET/CT revealed no notable abnormalities. The serum was positive for anti-SSA and anti-Ro antibodies AG, Lüebeck, Germany) were also negative in either serum or CSF. Therefore, this patient was investigated for novel antibody. In the TBA, The patient's serum and CSF displayed a ne-granular cytoplasmic IgG staining in the Purkinje cell layer and granular cell layer of the rat cerebella (Fig. 3). The IgG subclass repertoire of the antibody was analysed in the patient and revealed mainly IgG1 antibody (Fig. 4).
The patient was nally diagnosed as ACA associated with SS and received intravenous immunoglobulin, steroids, and cyclophosphamide (CTX) treatment. Symptoms of dizziness, nausea and vomiting were resolved, while ataxia persisted. Then the patient received tocilizumab 400 mg every month. As a result, the symptoms of ataxia improved gradually and uorescence pattern on rat cerebella was reduced obviously (Fig. 3).
Identi cation of target autoantigen HIP of rat cerebellum and the index patient's serum revealed a protein with apparent molecular mass of 21 kDa by SDS-PAGE. The protein was absent in HIP of therapy treatment serum or control serum (Fig. 5A). The precipitated protein was identi ed as Rattus norvegicus (Rat) Rab6A (UNIPROT acc. # Q9WVB1 ) by MALDI-TOF analysis (Fig. 5B). Western blot analysis of the HIP exhibited a speci c reaction at 21 kDa using patient's serum and anti-Rab6A antibody, respectively (Fig. 5C). Rab6A protein belongs to the Rab6 family, in human this family consists of 4 different isoforms: Rab6A, Rab6A', Rab6B and Rab6C. Rab6A' is generated by alternative splicing of the RAB6A gene and differs from Rab6A by only three amino acids. Both proteins are ubiquitously expressed and are together the most abundant Golgiassociated Rab proteins. Rab6B is encoded by a separate gene RAB6B and is preferentially expressed in brain, especially in microglial cells, pericytes and Purkinje cells [7] . RAB6C is a primate-speci c retrogene transcribed in a limited number of human tissues. Since Rab6B protein expressed in a brain speci c manner and show 91% identity with Rab6A protein [7] , we doubted whether the patient's serum also reactive with Rab6B. Indeed, both Rab6A and Rab6B antibody produced a similar uorescence pattern on rat cerebella comparable to that generated by the patient's serum (Fig. 6).
As a proof of correct antigen identi cation, the patient's sample was then tested using recombinant HEK293 cells which expressed either human Rab6A or Rab6B protein (Fig. 7). Serum and CSF of patient reacted with the cells expressing Rab6A or Rab6B but not with mock transfection protein, control serum reacted with none of them.
The congruence of the autoantibody target Rab6A /Rab6B was further demonstrated by the proof of neutralization of antibody binding to brain tissue: the reaction of the patient's autoantibody on tissue, especially in the Purkinje cell layer, could be abolished by pre-incubation with HEK293 lysate containing Rab6A or Rab6B (Fig. 8). Antibody binding was unaffected when a comparable fraction from mocktransfected HEK293 cells was used.
Speci city of anti-Rab6a/ Rab6b autoantibody Sera from 10 SS patients without neurological manifestation were analysed by CBA in parallel to the sample of the index patient. None of these control sera produced a similar immuno uorescence pattern to that of the index sera on the recombinant Rab6a/ Rab6b substrate (Supplementary material). It suggested that anti-Rab6a/ Rab6b autoantibody was speci c to disease of ACA, but not SS.

Discussion
We identi ed an novel autoantibody against Rab6A/Rab6B in a patient with comorbid ACA and SS who met the diagnostic criteria for primary ACA recently recommended by Hadjivassiliou M. et al [1] . The patient's serum reacted with Purkinje cell and granular layer of rat cerebellum but not with hippocampal Whether anti-Rab6A/Rab6B antibody is pathogenic is unknown. There is some indirect evidence for its potential pathogenic role. First, the antibody mainly response to Purkinje cell, which is known to be a cell type expressed exclusively in the cerebellum, and the patient presented cerebellar ataxia manifestation. Second, the antibody belonged to the IgG1 subclass, a strong complement activator, suggesting that it may act on Purkinje cell via complement-dependent mechanism, which is well-established feature in other autoantibody-associated disorders [8,9] . Third, CSF analysis displayed in ammatory features and immunosuppressive therapy was followed by clinical stabilization and ataxia improvement. Testing of serum sample taken after treatment with intravenous immunoglobulin, steroids, CTX and tocilizumab showed a decline of serum titer. On the other hand, Rab6A and Rab6B are intracellular antigens located in membranes encompassing the Golgi or endoplasmic reticulum (ER). Rab6A is restricted to the Golgi apparatus, whereas Rab6B is distributed over Golgi and ER membranes. It's believed that nuclear or cytoplasmic antigens are not accessible to immune attack in situ. As antibodies targeting intracellular antigens (e.g. Hu, Yo Ri, Ta) are not pathogenic, it is therefore possible that Rab6A/ Rab6B antibody has no such impact [9] .
Rab6A and Rab6B belong to GTPases of the Rab family, which are important regulators of intracellular transport and membrane tra cking in eukaryotic cells. There are about 70 members of Rab family in human, Rab6 proteins are the most abundant Golgi associated protein and they play an important role in retrograde Golgi-endoplasmic reticulum and intra-Golgi transport [10] . Although the pathogenicity of Rab6A/B hasn't been well understood, some other Rab proteins have been reported to associated with a multitude of inherited genetic disorders and acquired diseases ranging from peripheral neuropathy [11,12] , neurodegeneration [13] , to immunode ciency [14] and cancer [15] . A spastic ataxia related mutation has been identi ed in the gene of VPS13 protein, a Rab6 effector [16] , mutation in RAB3GAP2, gene of Rab3 GTPase-activating protein, was reported in hereditary spastic paraplegias (HSPs) [17] . It's interesting that mutations in gene of ITPR1, another Purkinje cell antigen, are also reported in HSPs [18] . Notably, the activity of some Rabs are regulated by calcium e ux [19] , as most of identi ed antigens that targeted by Purkinje cell antibodies are contributed in maintaining intracellular calcium homeostasis [20] , there may be potential functional relationship between them.
Comprehensive evaluation of anti-Rab6A/Rab6B antibodies in disease remains challenging. By far, few publications have been reported about Rab6 in neurological disease or multi-system autoimmune diseases. Case report here may bring some hint for understanding of Rab6A/Rab6B role in disease, however we need to collect more cases and to make further investigations to explore the clinical signi cance of Rab6A/Rab6B, as well as to differentiate Rab6A from Rab6B in their contribution to the disease.

Conclusions
In summary, we identi ed anti-Rab6A/B antibody as a new autoantibody against Purkinje cell in a patient with ACA and responsive to immunotherapy. Our nding expands the spectrum of diagnostic anti-neuronal antibodies of autoimmune cerebellar ataxia. The role of this novel antibody in mechanism of ACA warrants further study.

Consent for publication
Written informed consents were obtained from all patients.

Availability of data and materials
The datasets during and/or analysed during the current study available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests. Screening chart of patient samples followed in our study for new antigen identi cation. We enrolled patients with clinically suspected immune-mediated cerebellar ataxias who attended our hospital during July 2018 to December 2019. All individuals gave written informed consent for participation in the study. We reviewed their medical records and the patients didn't have a known trigger or any pathogenic neuronal antibodies. Patients' sera were tested by TBA on rat cerebellum sections and screened for anti-AQP4, thyroid peroxidase, thyroglobulin, NMDA-R, CASPR2, AMPA-R, LGI1, GABAb-R, GAD-65, CV2/CRMP5, PNMA2, Ri, Yo, Hu, amphiphysin, ITPR1, ZIC4, PKCγ, AP3B2, PCA-2, CARP VII, Homer-3, and NCDN antibodies. After the screening, 5 patients with a distinct pattern on positive TBA and negative detection of the above antibodies were further investigated for novel antibodies. Additionally, 10 age-and sex-matched SS patients were enrolled to test the speci city. Immuno uorescence staining of rat cerebellar tissues. Cryosections of rat cerebellum were incubated with patient serum (before therapy and after therapy, 1:100) and CSF (before therapy, undiluted) or control serum (1:100) in the rst step and with FITC-labeled goat anti-human IgG in the second step (green). A ne IgG staining of the Purkinje cell layer and granular cell layer was obtained. Upper row: X200 magni cation; lower row: X400 magni cation.

Figure 8
Neutralization of antibody reaction on tissue. Neutralization of immuno uorescence reaction on neuronal tissues. Serum was pre-incubated with extracts of HEK293 cells transfected with empty control vector or with the plasmid harboring human RAB6A or RAB6B cDNA. The extract containing Rab6A or Rab6B greatly reduced the immune reaction of the serum on rat cerebellum tissue. X200 magni cation.