Symmetric dimethylation of poly-GR correlates with disease duration in C9orf72 FTLD and ALS and reduces poly-GR phase separation and toxicity

Funding LG is funded by the Leonard Wolfson Centre for Experimental Neurology. SB acknowledges a longterm fellowship from EMBO. ADG is supported by the NIH (grant number NS097263). TL is supported by an Alzheimer’s Research UK Senior Fellowship and the Leonard Wolfson Centre for Experimental Neurology. AI receives funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (648716 C9ND), Alzheimer’s Research UK, the Motor Neurone Disease Association and the UK Dementia Research Institute which receives its funding from DRI Ltd, funded by the UK Medical Research Council, Alzheimer’s Society and Alzheimer’s Research UK. The Queen Square Brain Bank for Neurological Disorders is supported by the Reta Lila Weston Institute for Neurological Studies and the Medical Research Council. The MRC London Neurodegenerative Diseases Brain Bank receives funding from the UK Medical Research Council (MR/L016397/1). The MRC London Neurodegenerative Diseases Brain Bank and Manchester Brain Bank are part of the Brains for Dementia Research programme, jointly funded by Alzheimer’s Research UK and Alzheimer’s Society.


Antibody generation
Rabbit monoclonal ADMA-GR and SDMA-GR antibodies were generated by UCB Pharma. Two separate rabbits were immunised with peptide N-acetyl-GRGRGRGR-NH2 modified with either asymmetric dimethyl arginine (ADMA-GR) or symmetric dimethyl arginine (SMDA-GR) (Peptide Protein Research Ltd) respectively. Immunisations were also performed with ADMA-PR and SDMA-PR peptides but no specific antibodies were generated. Immunisations were performed using the peptides conjugated to KLH, OVA and BSA in 3 successive sub-cutaneous administrations with the first dose being given in complete Freund's adjuvant and subsequent shots provided in incomplete Freund's adjuvant with two-week gaps between immunisations. Termination occurred 14 days after the final boost with single cell suspensions of spleen, blood, lymph node and bone marrow prepared and frozen in 10 % dimethyl sulfoxide (DMSO) in fetal calf serum (FCS) at -80 °C.
B cell culture screening was performed using a method previously reported [20]. Briefly, the presence of peptide-binding antibodies in B cell culture supernatants was determined using a homogeneous fluorescence-based binding assay using Superavidin beads (Bangs Laboratories) coated with either biotinylated ADMA-GR or SDMA-GR. Binding was revealed with a goat anti-rabbit IgG Fcγ-specific Cy-5 conjugate (Jackson ImmunoResearch). Plates were read on a Mirrorball fluorescence cytometer (TTP Labtech).
Following primary screening, positive supernatants were consolidated onto 96-well bar-coded master plates then screened for binding against ADMA-GR, SDMA-GR, unmodified-GR peptide, as well as unmodified N-acetyl-PRPRPRPR-NH2 peptide (PR), and PR peptide modified with either asymmetric dimethyl arginine (ADMA-PR) or symmetric dimethyl arginine (SMDA-PR) (Peptide Protein Research Ltd). In addition, an irrelevant peptide from an unrelated protein was used as a negative control. Binding assays involved an ELISA using a streptavidin capture step. The ELISA assay involved coating 384well Maxisorp ELISA plates (Thermo Fisher Scientific) with Streptavidin at 2 µg/ml in a carbonate coating buffer (dH2O, + 0.16 % Na2CO3, + 0.3 % NaHCO3) before biotinylated peptide was added at 2 µg/ml. Plates were blocked with 1 % bovine serum albumin (BSA) in phosphate buffered saline (PBS) and then incubated with 10 µl/well of B cell culture supernatant. Plates were then incubated with a secondary HRP-conjugated goat anti-rabbit IgG Fc-specific antibody (1:5000; Jackson ImmunoResearch) and visualisation of binding was revealed using 3,3',5,5'-Tetramethylbenzidine substrate (TMB; Millipore). The optical density was measured at 630 nM using Synergy 2 microplate reader (BioTek).
B cell supernatants demonstrating specificity to the ADMA-GR or SDMA-GR peptides were selected for variable region recovery and cloning as previously described [8], utilising a fluorescent foci technique using beads coated with biotinylated peptide to identify and isolate antigen-specific B cells from positive culture wells. Specific antibody variable region genes were recovered from single B cells by reverse transcription (RT)-PCR using heavy and light chain variable region-specific primers. PCR primers contained restriction sites at the 3' and 5' ends allowing cloning of the variable region into a rabbit IgG1 (VH) or rabbit kappa (VL) mammalian expression vector. Heavy and light chain constructs were cotransfected into HEK-293 cells using 293fectin transfection reagent (Invitrogen). After 7 days of recombinant antibody expression, the supernatants were harvested and antibodies were re-screened for selectivity using the specificity assays described above. IgG were purified on an AKta system (GE healthcare) using affinity chromatography (protein A) followed by size exclusion (SE)-HPLC to produce a product >98% pure monomer species.

Cases
Tissue from C9orf72 cases used in this study was obtained from the Queen Square Brain Bank (QSBB) for Neurological Disorders, UCL Queen Square Institute of Neurology, University College London

Double-label immunofluorescence
7 µm thick paraffin-embedded frontal cortex sections were stained using a rat anti-poly(GR) antibody (5H9,1:25, [15,14]) in combination with either rabbit anti-ADMA-GR (UCB, 1:2000) or SDMA-GR (UCB, 1:50). After appropriate pre-treatment, tissue sections were incubated with the primary antibodies for 1 hour at room temperature, washed in TBS-T and incubated with the secondary antibodies chicken antirat Alexa Fluor 488 (Thermo Fisher Scientific, 1:1000) and donkey anti-rabbit Alexa Fluor 568 (Thermo Fisher Scientific, 1:1000) for 1 hour at room temperature. Tissue sections were washed and mounted using Vectashield anti-fade mounting medium containing DAPI (Vector Laboratories) for nuclear counterstaining. Sections were viewed with a Leica DM5500B fluorescence microscope using the 40x or 63x objectives and Leica Application Suite X software with 3D deconvolution post-processing.

Immunohistochemistry image acquisition and quantification
Frontal cortex sections stained with p62, ADMA-GR and SDMA-GR antibodies were scanned at 40x magnification using a Leica Slide Scanner SCN400 and viewed on Digital Image Hub (Leica Biosystems). Ten regions of interest were selected per case and the number of p62, ADMA-GR or SDMA-GR inclusions in each image were manually quantified to give the total number of inclusions per case. For normalisation to p62, the percentage of p62 inclusions containing ADMA-GR or SDMA-GR was calculated by dividing the total number of ADMA-GR or SDMA-GR inclusions by the total number of p62 inclusions.

Correlation analysis
All statistical analysis was performed using GraphPad Prism. All correlations were performed using Spearman rank correlation test. In all instances, a p-value of less than 0.05 was considered statistically significant.

Phase separation assay
Peptides were chemically synthesized by Pepscan (Lelystad, Netherlands), dissolved in milliQ water at 1 mM and stored at -20 °C. For droplet formation, peptides were diluted to the indicated concentrations in 100 mM K2HPO4/KH2PO4 buffer at pH7 containing 30 % polyethylene glycol (PEG; Sigma-Aldrich).
OD600 of 40 µl samples was measured in clear-bottom 384 well plates (Thermo Fisher Scientific) using a SPARK Multimode microplate reader (Tecan Life Sciences). Data was analyzed using Microsoft Excel and GraphPad Prism. For imaging on plastic, droplets were incubated in plastic Cell Counter slides (Bio-Rad) and the chambers were sealed using nail varnish to prevent evaporation (described in detail in [4]). For imaging on glass, droplets were incubated in glass chambers made of microscope cover slips (Thermo Fisher Scientific), microscope slides (Thermo Fischer Scientific) and silicone spacers (Grace Bio-labs), as used previously [5]. Samples were imaged at room temperature on a Zeiss LSM 780 Meta NLO confocal microscope equipped with a 20x long-range objective. Pictures were processed with FIJI software.

Cytotoxicity assay
Primary mouse cortical neurons were dissociated into single cell suspensions from E16.5 C57BL/6 mice (Jackson Laboratory) cortices using a papain dissociation system (Worthington Biochemical Corporation). Neurons were seeded onto poly-L-lysine coated plates (0.1 % w/v) and grown in Neurobasal media (Gibco) supplemented with B-27 serum-free supplement (Gibco), GlutaMAX, and Penicillin-Streptomycin (Gibco) in a humidified incubator at 37 °C, with 5 % CO2. Three days after seeding a half-media change was performed and peptides were added at indicated concentration for 24h. Cytotoxicity in primary neuron cultures was measured by lactose dehydrogenase (LDH) release assays (Promega, CytoTox 96® Non-Radioactive Cytotoxicity Assay), according to manufacturer's instructions. LDH readout was measured using a SPARK Multimode microplate reader (Tecan Life Sciences). Data was analyzed using Microsoft Excel and GraphPad Prism. Subsequently, the cells were fixed in 4 % formaldehyde in PBS, and stained according standard protocols. Antibodies used were rabbit NeuN antibody (ABN78, EMD Millipore) and AlexaFluor 488 secondary antibody (Life Technologies). Samples were imaged on a Zeiss LSM 780 Meta NLO confocal microscope. Pictures were processed with FIJI software.

Poly-GR peptide uptake experiment
We labelled poly-GR20 peptides with an Alexa-488 labelling kit (Thermo-Fisher Scientific) as described before [4]. U2OS (ATCC) cells were cultured in DMEM medium (Thermo-Fisher Scientific) containing 10% FBS (Invitrogen) at 37°C and 5% CO2, and handled according to standard procedures. Cells were seeded on glass cover slips and allowed to adhere for 24 hours. Labelled peptides were added to the cell medium at a concentration of 1 µM and incubated for 1 hour. Cells were washed before fixing with 4% PFA. Cells were imaged on a Zeiss LSM 780 Meta NLO confocal microscope. Pictures were processed and analysed with FIJI software.

Supplementary discussion
This study describes the characterisation of two novel antibodies designed to detect dimethylated poly-GR in C9orf72 patient post-mortem tissue and is the first to demonstrate that poly-GR can be detected both in asymmetric and symmetric dimethylated forms. The ADMA-GR and SDMA-GR specific antibodies were found to label a proportion of DPR protein neuronal cytoplasmic inclusions in the frontal cortex of both C9orf72-FTLD and C9orf72-ALS cases. Symmetrically dimethylated poly-GR showed a consistently positive correlation with disease duration and age at death, indicating that this posttranslational modification may have a significant effect on disease outcome.
Previous studies have suggested that this DPR protein may be post-translationally modified by the addition of methyl groups to arginine residues due to co-localisation of poly-GR with PRMT enzymes and immunoreactivity to an antibody which recognises proteins that are asymmetrically dimethylated.
However, the antibodies used in the previously published studies were not specific for the poly-GR protein, making it difficult to determine whether the poly-GR protein itself was methylated, or whether it associated with other proteins that can undergo arginine methylation [3,7,19]. In this study, ELISA analysis demonstrated that novel antibodies designed to detect either symmetrically or asymmetrically modified poly-GR were specific for their respective peptide antigens, and did not show reactivity to unmethylated forms of poly-GR or to the oppositely dimethylated poly-GR antigen. The ADMA-GR antibody showed some reactivity to ADMA-PR, and the SDMA-GR antibody showed some reactivity to SDMA-PR. However, given that the antibodies had a higher affinity for the methylated poly-GR proteins, and poly-PR is rarely detected by immunohistochemistry in human post-mortem brain [9,24], it was determined that these antibodies were unlikely to identify ADMA-PR and SDMA-PR in post-mortem tissue.
Immunohistochemical assessment of the ADMA-GR and SMDA-GR antibodies showed nuclear and cytoplasmic staining in the frontal cortex of both healthy control and C9orf72 cases. This is most likely the detection of dimethylated epitopes of other GR-containing proteins that often undergo methylation at sites with a consensus glycine/arginine rich motif, and is consistent with other studies that have investigated ADMA in human brain tissue [3,19]. The detection of inclusions, however, was specific to the C9orf72 cases, although we cannot rule out that inclusions also contain non-DPR methylated GRcontaining proteins. The number of ADMA-GR positive inclusions was consistently higher than the number of SDMA-GR positive inclusions in all C9orf72 cases, consistent with a report showing a high degree of co-localisation between GR inclusions and an ADMA-specific antibody [19]. This is unsurprising given that asymmetric dimethylation of arginine residues is known to be the most prevalent form of methylation in physiological systems, with the monomethylated and symmetrically dimethylated forms thought to occur at levels of about 20 % to 50 % that of the asymmetrically dimethylated form [1].

ADMA-GR and SDMA-GR containing inclusions were detected in both C9orf72-FTLD and C9orf72-ALS
cases. This indicates that arginine methylation of poly-GR is not specific to a clinical phenotype and cannot be used to pathologically differentiate between C9orf72 ALS and FTLD cases, as is the case with FTLD-FUS and ALS-FUS, where methylated FUS inclusions are found in the ALS-FUS cases [10].
Double immunofluorescence demonstrated that both ADMA-GR and SDMA-GR co-localise with poly-GR, indicating that these antibodies are detecting DPR protein inclusions in the C9orf72 cases. In agreement with the previous published studies on methylation of poly-GR inclusions, co-localisation was not observed for all poly-GR inclusions, suggesting that some inclusions may contain unmethylated or mono-methylated forms of poly-GR [3,19]. The lack of methylation in some poly-GR inclusions may reflect the dynamic reversibility of arginine methylation. Although arginine methylation was initially thought to be a permanent post-translational modification, a number of studies have emerged supporting the reversible nature of methylarginine and have identified putative arginine demethylases [2,6,[21][22][23]. The poly-GR inclusions detected that were unmethylated may therefore reflect poly-GR proteins yet to undergo arginine methylation or that have been demethylated. The ability to be reversibly methylated could have important physical or functional consequences for the poly-GR proteins. This has been demonstrated for the FUS protein where arginine methylation alters the proteins ability to phase separate and changes its interaction with its nuclear importer [11,16].
Our observation of a significant positive correlation between SDMA-GR inclusions in the frontal cortex and both disease duration age at death indicates a potentially protective nature of SDMA-GR inclusions, as patients that have higher numbers of SDMA-GR inclusions have a longer disease duration and later age at death. Importantly, the association between SDMA-GR and disease duration or age at death remained when the DPR protein inclusion burden of each case was considered by normalising the number of SDMA-GR positive inclusions to the number of p62 positive. This indicates that these associations are specific for SDMA-GR and are not related to total DPR protein pathology in the frontal cortex. and age at death could be hypothesised to be due to the symmetric dimethylation making poly-GR more biophysically inert. Indeed, we observed that both ADMA-and SDMA modification of poly-GR reduced its ability to phase separate through decreasing interaction strength between poly-GR molecules. As arginine promotes phase separation by both electrostatic and pi-pi interactions in this paradigm [5], methylation likely affects one or both of these interactions. Consistent with the possibility that phase separation of poly-GR contributes to its toxicity [13], these modifications also reduced poly-GR toxicity in primary neuronal cultures.
Further studies are needed to explore why only SDMA-GR and not ADMA-GR associates with longer disease duration and age at death. There is little published research into the differences in biological roles and functional consequences of ADMA and SDMA modifications, however it could be speculated we described previously [12]. We made use of a cancer cell line that is more resistant to GR-toxicity, and short incubation times to allow us to disentangle uptake from cellular toxicity (e.g. influx due to loss of membrane integrity).   Data are Spearman's correlation coefficient r (95 % confidence interval (CI)) and p value. Significance level was set at p < 0.05 (two-sided).