Novel Insights into the Multiple Sclerosis Risk Gene ANKRD55

Previously, we identified rs6859219, a single nucleotide polymorphism (SNP) located in intron 7 of ANKRD55, as a genetic risk factor for multiple sclerosis (MS) (1); we subsequently established this association with genome-wide significance in a replication study including a total of 13,849 subjects (2). Recently, a proxy of rs6859219 (r² = 0.9, D′ = 1), rs71624119, also was associated with MS in a comprehensive fine-mapping of autoimmune disease–related genomic regions (3). Other studies linked ANKRD55 to several autoimmune diseases, including rheumatoid arthritis (4–6), celiac disease (7), Crohn’s disease (8), and type 2 diabetes (9). An SNP near ANKRD55 also was associated with changes in N-glycosylation of IgG (10), which is altered in rheumatoid arthritis. Interestingly, defective N-glycosylation was also implicated in MS (11). The absence of strong linkage disequilibrium (LD) between SNPs in ANKRD55 and other MS candidate genes in the area [i.e., the IL31RA-IL6ST cluster (1, 12)] is suggestive of an independent role for this gene in MS susceptibility.

ANKRD55 and its expressed proteins are poorly characterized. Ankyrin repeats consist of 33–34 residue motifs conformed in two α-helices separated by loops, and they function almost exclusively to mediate protein–protein interactions (13). As one of the most abundant motifs in nature, ankyrin repeats are present primarily in eukaryotic proteins with very diverse functions, including transcription factors (TFs), cytoskeletal proteins and cell cycle regulators, among others (13).

The ANKRD55 locus, located on chromosome 5q11.2, is complex. The Ensembl 75 release [(14) http://www.ensembl.org] of February 2014 includes six alternatively spliced transcript variants: four are protein coding, and the remaining two processed transcripts without protein products. However, only Ensembl transcript 001 (corresponding to the full-length form) is regarded as the consensus coding sequence, with the remainder not being consistently annotated across genomic databases because of the lack of experimental evidence for the transcripts and the resulting proteins coded for. Information about cells and tissues that express ANKRD55 protein, its subcellular localization, and its function is relatively scarce. The Human Protein Atlas (http://www.proteinatlas.org) documents ANKRD55 expression in many tissues of diverse organ systems, as well as in all myeloid cell lines tested.

Orthologs of ANKRD55 exist in other primate and nonprimate vertebrate species. For mice, Ensembl annotates four transcript variants, named Ankrd55 201 through 204, with each of these giving rise to a different protein isoform. The mouse protein encoded by transcript 201, which is also the only protein annotated by UniProt, is 84% identical to the human protein encoded by transcript 001. In this article, we use the Ensembl nomenclature to refer to the transcripts.

The present study was conducted to shed light on the expression of ANKRD55 in the immune and nervous systems, to investigate the possible relationship of rs6859219 with ANKRD55 expression regulation, and to explore the relationship of ANKRD55 with MS. In the first instance, we aimed to characterize the human ANKRD55 transcript variants expressed in immune cells and compare the expression profile of immune cells with a human neuronal cell line. Human ANKRD55 protein expression and subcellular localization were also assessed. Once the most abundant RNA transcripts were identified, we correlated the expression of each of these transcripts with rs6859219 genotypes in PBMCs and subpopulations. Moreover, to gain insight into the possible role of ANKRD55 in the normal and inflamed CNS, we examined ANKRD55 protein expression in a murine model of experimental autoimmune encephalomyelitis (EAE), as well as in neuron and microglia cultures under inflammatory conditions.

Human blood samples were collected after written informed consent was obtained, and the study was approved by the local ethics committees from Bilbao (Comité Ético de Investigación Clínica de Euskadi) and Barcelona (Comité Ético de Investigación Clínica Hospital Universitari Vall D’Hebron). Mice experiments were performed in strict accordance with European Union and governmental regulations (Decret 53/2013 BOE no. 34 and Comunidad de Madrid ES280790000184). The Ethics Committee on Animal Experimentation of the Instituto Cajal, Consejo Superior de Investigaciones Científicas, approved all procedures (protocol number 2013/03 CEEA-IC). Measures to improve welfare assistance and clinical status, as well as end point criteria, were established to minimize suffering and ensure animal welfare. Briefly, wet food pellets were placed on the bed-cage when the mice began to develop clinical signs to facilitate access to food and hydration.

PBMCs were isolated from whole blood by a Ficoll-Hypaque density gradient (GE Healthcare, Chalfont St. Giles, U.K.), as indicated by the manufacturer, frozen in the presence of 10% DMSO, and stored in liquid nitrogen until use. Jurkat and U937 cells were kindly provided by Prof. Carlos Matute (University of the Basque Country, Leioa, Spain) and maintained in RPMI 1640 medium supplemented with 10% FBSi and 2 mM l-glutamine (all from Sigma-Aldrich, Madrid, Spain). SH-SY5Y cells were kindly provided by Prof. Pablo Villoslada (Institute of Biomedical Research August Pi Sunyer, Barcelona, Spain) and maintained in DMEM/F12 GlutaMAX supplement medium (Thermo Fisher Scientific, Waltham, MA), complemented with 10% FBSi (Sigma-Aldrich).

The murine microglial cell line (BV-2) was obtained from Interlab Cell Line Collection San Martino-Istituto Scientifico Tumori-Istituto Nazionale per la Ricerca sul Cancro (Genova, Italy). The cells were grown in DMEM supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Thermo Fisher Scientific) and maintained under standard cell culture conditions at 37°C and 5% CO₂. BV2 cells were plated at a density of 1 × 10⁶/well in 6-well culture plates for Western blot analysis and to a density of 2 × 10⁴/well in 24-well culture plates for immunocytochemistry analysis. One hour before experiments, cells were subjected to restricted conditions (fresh serum-free DMEM) before being treated with LPS (50 ng/ml; Sigma-Aldrich) and IFN-γ (100 U/ml; PeproTech, London, U.K.) for 24 h. Primary mix glial cultures were prepared as follows. In brief, after decapitation, forebrains of newborn (1 d postnatal) C57BL/6 mice were dissociated mechanically, filtered through a 150-μm nylon mesh, and resuspended in DMEM supplemented with 10% FBS, 10% horse serum, 4 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. The cells were cultured on poly-d-lysine–coated (15 μg/ml) 75-cm² flasks (Falcon, LePont de Claix, France), and the medium was replenished 7 d after initial seeding. Upon reaching confluence (12–16 d), the flasks were shaken at 230 rpm at 37°C for 3 h to remove loosely adherent microglial cells. The supernatant was plated in six-well culture plates (1 × 10⁶ cells/well) for Western blot analysis and in 24-well culture plates (5 × 10⁴/well) for immunocytochemistry analysis in DMEM supplemented with 10% FBS, 10% horse serum, 4 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cells were grown in a humidified environment containing 5% CO₂ and held at a constant temperature of 37°C. One hour prior to LPS exposure, medium was removed from microglial cultures and replaced by fresh serum-free DMEM. Then, cells were stimulated with 50 ng/ml LPS for 24 h.

Hippocampal neuronal cultures were prepared from mice embryos (embryonic day 17–18), as previously described (15). After the dissection, cortices were stripped of meningeal tissue and dissociated in PBS containing trypsin without EDTA (Thermo Fisher Scientific) and deoxyribonuclease I (1 mg/ml; Roche Diagnostics, Manheim, Germany) at 37°C for 15 min. A single-cell suspension was prepared by triturating tissue, the pellet was resuspended in DMEM supplemented with heat-inactivated horse-serum (10% v/v), and cells were plated on poly-l-lysine–coated (15 μg/ml; Sigma-Aldrich) 6-well culture plates (1 × 10⁶/well) for Western blot analysis and on 24-well culture plates (2 × 10⁴/well) for immunocytochemistry for 3 h. Medium was changed to remove nonadherent cells, and new Neurobasal medium was added supplemented with penicillin/streptomycin (1% v/v), GlutaMAX (2 mM; all from Thermo Fisher Scientific), and 2% B-27 supplement (Thermo Fisher Scientific). After three d in vitro, nonneuronal cell division was halted by exposure to 5 ng/ml cytosine-d-arabinofuranoside (Sigma-Aldrich). Cells were grown in a humidified environment containing 5% CO₂ and held at a constant temperature of 37°C for 6 d. One hour prior to LPS exposure, medium was removed from neuronal cultures and replaced by fresh Neurobasal medium, and cells were stimulated with 100 ng/ml LPS for 24 h.

Specific pairs of primers were designed for the detection of five of the six transcript variants described in Ensembl by conventional RT-PCR (Table I). Transcript variant 008, which encodes a processed transcript that is not translated into protein and has only been reported in cerebral cortex with low confidence [(16), http://www.ebi.ac.uk/gxa], could not be studied because its sequence overlaps completely with transcript 001 (Fig. 1). For the detection of transcript 001 and/or 002, four redundant pairs of primers were used (Table I, pairs 1, 2, 6, and 7). For sequencing of transcript 005, new primers were designed to cover the whole open reading frame in two overlapping fragments (Table I, pairs 8–10). Primer 005.2R_alt was used in samples in which amplification with 005.2R was too weak or absent. For the quantitative PCR (qPCR) experiments, specific primers for transcripts 005, 001, and 007 were designed (Table I, pairs 11–13).

Clinical, demographic, and genotype data for the patients and controls included in the gene expression study are provided in Table III. Patients were diagnosed with clinically definite MS according to the criteria of Poser et al. (17). Two PBMC sample collections were used: the first collection consisted of PBMCs from 43 MS patients with different clinical courses and stages of activity (15 relapsing-remitting MS [RRMS] patients in relapse, 13 RRMS patients in remission, and 15 primary progressive MS [PPMS] patients) and 14 healthy controls recruited at the Vall d’Hebron Hospital in Barcelona. The second collection contained PBMCs from 53 healthy controls provided by the Basque BioBank for Research (www.biobancovasco.org).

CD4⁺, CD8⁺, CD14⁺, CD19⁺, and CD56⁺ subpopulations were isolated from PBMCs of 23 healthy donors using CD4⁺ (Th cells; cat. no. 130-045-101), CD8⁺ (T cytotoxic; cat. no. 130-045-201), CD14⁺ (monocytes; cat. no. 130-050-201), CD19⁺ (B cells, cat. no. 130-050-301) and CD56⁺ (NK cells; cat. no. 130-050-401) MACS MicroBeads, following the manufacturer’s instructions (Miltenyi Biotec, Madrid, Spain). PBMC samples were split into two parts. CD8⁺ cells were purified by positive selection from one part, and CD56⁺ cells were retrieved from the resulting CD8⁻ fraction. CD19⁺ cells were retrieved from the second part by positive selection, CD14⁺ cells were obtained from the CD19⁻ fraction, and CD4⁺ cells were subsequently obtained from the CD14⁻ fraction. The purity of each subpopulation was verified by flow cytometry using PE mouse anti-human CD4⁺ (130-098-167), PE-Vio 770 mouse anti-human CD8⁺ (cat. no. 130-098-060), FITC mouse anti-human CD14⁺ (cat. no. 130-080-701), VioBlue mouse anti-human CD19⁺ (cat. no. 130-098-606), and PE-Vio 770 mouse anti-human CD56⁺ (cat. no. 130-098-132) Abs, as well as PE mouse IgG2a (cat. no. 130-098-849), PE-Vio 770 mouse IgG2a (cat. no. 130-098-564), FITC mouse IgG2a (cat. no. 130-098-877), VioBlue mouse IgG1 (cat. no. 130-099-756), and PE-Vio 770 mouse IgG1 (cat. no. 130-098-563) isotype controls (all from Miltenyi Biotec).

Total RNA from Jurkat, U937, and SH-SY5Y cell lines was obtained using TRI RNA Isolation Reagent (Sigma-Aldrich), and total RNA from human PBMCs was obtained with TRI or a NucleoSpin RNA II kit (MACHEREY-NAGEL), following the manufacturers’ protocols, and treated with RNase-free DNase I (Invitrogen). RNA concentration was determined using a NanoDrop 2000 spectrophotometer, and integrity was verified by running the samples in an agarose gel. RNA concentration was normalized prior to reverse transcription. Total RNA (1000 or 100 ng) was reverse transcribed with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems).

Optimal amplification conditions of the primer pairs were tested on cDNA from each of the cell types and finally set to initial denaturation at 95°C for 5 min, followed by 40 cycles of denaturation at 95°C (35 s), annealing at 54°C (35 s), and extension at 72°C (45 s), and a final extension step at 72°C for 10 min. For primer pairs 8 and 9 (Table I), which yielded longer products, denaturation and annealing times were 45 s, and extension time was 75 s. When using cDNA from PBMCs, cells from at least two to four individuals were tested in each experiment. PCRs were carried out with 5, 25, and 50 ng template cDNA/reaction. Each experiment (i.e., each PCR in the same conditions) was performed at least in duplicate. Reactions were run in a Veriti 96-Well Fast Thermal Cycler (Applied Biosystems), and PCR products were subjected to agarose gel electrophoresis. PCR products of expected or unexpected molecular sizes consistently found in more than one reaction were verified by sequencing.

qPCR was performed with 5 ng cDNA/reaction using Fast SYBR Green Master Mix and primers for different ANKRD55 transcripts specified above (IDT). GAPDH and ACTB (predesigned primers from IDT and QIAGEN) were used as internal controls for normalization, and data were analyzed using the 2^−ΔCt method (18). Standard curves with serial dilutions of cDNA were performed to test the efficiency of the primers, ensuring that all of the primer pairs had amplification efficiencies within the accepted range (90–110%) and that the efficiency of the primers for the internal controls and the target transcripts differed by no more than 10% from each other (18). Reactions were run in triplicate, and no–reverse transcription and no-template controls were included.

Digital droplet PCR (ddPCR) assays were performed for absolute quantification of human ANKRD55 transcript copy numbers (19) in CD4⁺ T cells using isoforms 001, 005, and 007 and GAPDH qPCR primers. cDNA concentration and primers used were the same as for qPCR. Each 20-μl mix reaction containing 1× ddPCR EvaGreen Supermix (Bio-Rad, Hercules, CA), cDNA, and primers was pipetted into a droplet generator cartridge (Bio-Rad). Then, 70 μl Droplet Generation Oil for EvaGreen (Bio-Rad) was loaded into each oil well. The cartridge was placed into a QX200 Droplet Generator (Bio-Rad) to generate droplets and then transferred to a 96-well polypropylene plate (Eppendorf, Hamburg, Germany). The plate was heat-sealed with foil using a PX1 PCR Plate Sealer and placed in a thermal cycler C1000 Touch (both from Bio-Rad). Thermal cycling conditions were 1 cycle at 95°C for 5 min, 40 cycles of 95°C for 30 s, 60°C for 1 min, and 1 cycle at 4°C for 5 min, 90°C for 5 min, with 2°C/s ramp rate. After PCR, the 96-well plate was loaded into the QX200 Droplet Reader (Bio-Rad) and the fluorescence intensity of individual droplets was measured. Data analysis was performed with QuantaSoft droplet reader software (Bio-Rad).

Female C57BL/6 mice were purchased from Harlan (Barcelona, Spain) and housed in the animal facilities at Instituto Cajal under controlled 12-h light/dark cycle, temperature (20 ± 2°C), and relative humidity (40–50%). Mice had free access to standard food and water. Induction of EAE was performed at 6–8 wk of age by s.c. immunization with 300 μg myelin oligodendrocyte glycoprotein (MOG)_35–55 peptide (Sección de Síntesis de Péptidos, Centro de Biología Molecular Severo Ochoa, Consejo Superior de Investigaciones Científica) and 200 μg Mycobacterium tuberculosis (H37Ra Difco) in a 1:1 mix with IFA (CFA; Sigma-Aldrich). On the same day and 2 d later, mice were injected i.p. with 200 ng pertussis toxin (Sigma-Aldrich) in 0.1 ml PBS. Control animals (CFA; n = 7) were obtained by inoculation with the same emulsion without MOG and without injections of pertussis toxin. All animals were sacrificed at day 17 postimmunization (when all animals of the EAE group [n = 9] were symptomatic) for ELISA and Western blot analysis. The mice were examined daily for clinical signs of EAE, and disease scores were measured as follows: 0, no disease; 1, limp tail; 2, limp tail and hind limb weakness; 3, hind limb paralysis; 4, hind limb and front limb paralysis; and 5, moribund and death.

Mouse primary cultures (hippocampal neurons and mouse microglia), BV2 cell line cultures, and lumbar spinal cord, brain, and spleen from CFA and EAE animals were lysed in TBS (pH 7.6) containing 10% glycerol, 1% Nonidet P-40, 1 mM EDTA, 1 mM EGTA, and complete protease inhibitors mixture (Roche Diagnostics). The samples were sonicated and centrifuged for 15 min at 13,500 rpm, and supernatants were mixed with 5× Laemmli sample buffer and boiled for 5 min. Equal amounts of protein (30 μg) were resolved on 10% SDS-PAGE and electroblotted at 90 V for 70 min at 4°C to nitrocellulose (Amersham Biosciences). The membranes were blocked for 1 h at room temperature (RT) in 5% (w/v) BSA (Thermo Fisher Scientific) in TBS with 0.1% Tween 20 (TBSTw). Then, they were incubated overnight at 4°C with the ANKRD55 primary Ab (1:500; Sigma-Aldrich) in 5% BSA-TBSTw, washed extensively with 5% BSA-TBSTw solution, and incubated with HRP-conjugated anti-rabbit IgG (1:10,000) secondary Ab (Jackson ImmunoResearch, West Grove, PA) for 1 h at RT. Finally, the blots were rinsed, and the peroxidase reaction was developed by ECL (Amersham Biosciences). The blots were stripped in 62.5 mM Tris-HCl (pH 6.8) containing 2% SDS and 0.7% 2-ME and reprobed with monoclonal anti–α-tubulin (1:10,000; Sigma-Aldrich).

SH-SY5Y cells were fractionated into different subcellular compartments following the protocol of Holden and Horton (20). Nuclei from human CD4⁺ cells and Jurkat, SH-SY5Y, and U937 cell lines were isolated by sucrose gradient centrifugation (21). For immunoblotting, 8.5 μg protein of each subcellular fraction was electrophoresed through a 10% SDS-PAGE gel and transferred onto polyvinylidene difluoride membranes (Merck Millipore). The resulting blot was incubated with anti-ANKRD55 (1:1000; Aviva Systems Biology) at 4°C overnight, in the presence or absence of a 200-fold molar excess of ANKRD55 blocking peptide (2 μg/ml), whose sequence is part of both Ensembl isoforms 001 and 005 and is the original immunogen used to generate the polyclonal Ab (sequence: LSQESRTEPTRPPPSQSSRPQKKERRFNVLNQIFCKNKKEEQRAHQKDPS; Aviva Systems Biology) and then with HRP-conjugated anti-rabbit secondary IgG Ab (1:3000; Cell Signaling Technology) for 1 h at RT. anti-GAPDH (1:1000; Merck Millipore), anti-histone H3 (1:1000; Cell Signaling Technology), and anti-GRP94 (1:1000; Enzo Life Sciences) Abs were used to validate fraction purity.

Jurkat, U937, and SH-SY5Y cells were seeded on coverslips coated with poly-d-lysine (Sigma-Aldrich). U937 cells were differentiated with 25 ng/ml PMA for 40 h and stimulated with 5 μg/ml LPS (both from Sigma-Aldrich) for 1, 4, and 24 h. Subsequently, cells were fixed in 4% paraformaldehyde-PBS for 20 min, incubated in 0.2% Triton X-100 in PBS for 30 min, blocked in 3% BSA (all from Sigma-Aldrich) in 0.2% Triton X-100 in PBS for 30 min, and stained with anti-ANKRD55 (1:500; Aviva Systems Biology) for 1 h at RT. Staining with Alexa Fluor 488–conjugated anti-rabbit Ab (1:500; Invitrogen) was performed for 1 h at RT, followed by staining with DAPI (1:1000). To confirm specificity of ANKRD55 staining, a blocking peptide for the ANKRD55 Ab (Aviva Systems Biology) was used (see preceding paragraph). Images were obtained using a Leica TCS CW SP8 STED Super-Resolution microscope with a 63× immersion objective and excitation wavelengths of 408 and 488 nm. For SH-SY5Y cells, we used a zoom factor ranging from 0.75 to 1.5×.

Mouse hippocampal neurons, mouse microglia, and BV2 microglial cells were seeded on poly-d-lysine–coated coverslips. After 24 h of treatment, cells were fixed in 4% paraformaldehyde for 20 min, blocked, and incubated for 3 d with rabbit anti-ANKRD55 Ab (1:100; Sigma-Aldrich), mouse anti-MAP2 Ab (1:200; Calbiochem, Darmstadt, Germany) for visualizing neurons, and mouse anti-CD11b (1:200; Serotec, Oxford, U.K.) and Alexa Fluor 594–conjugated anti-phalloidin (1:50, Thermo Fisher Scientific) to see microglia. After washing, cells were incubated with Alexa Fluor 594–conjugated goat anti-mouse Ab and Alexa Fluor 488–conjugated goat anti-rabbit Ab (Thermo Fisher Scientific) and counterstained with DAPI.

ANKRD55 levels in cell lysates from primary mouse microglial cell cultures, as well as ANKRD55 content in lumbar spinal cord and brain from intact and EAE animals, were measured with a mouse-specific solid-phase sandwich ELISA kit, using an Ab for ANKRD55, following the manufacturer’s recommendations (Cusabio Biotech, Hubei, People's Republic of China). The minimum detectable dose of mouse ANKRD55 was <7.8 pg/ml, and the intra and inter coefficient of variation was <8 and <10%, respectively.

Mononuclear cells from the brains and spinal cord of EAE mice at the peak of the disease were obtained as described (22), with some modifications. Briefly, the brains and spinal cord were cut into pieces, mechanically triturated, passed through a 100-μm cell strainer, and diluted with isotonic Percoll (stock isotonic percoll; GE Healthcare) to a final concentration of 30%. This suspension was slowly layered on the top of 70% SIP and centrifuged for 30 min at 500 × g at 18°C. The debris and myelin on the top of the tube was removed, and 2 ml the 70–30% interphase was collected, washed, and resuspended in FACS buffer (PBS + 0.1% BSA) for cell counting and Ab staining. Cells were then incubated with anti-CD16/CD32 (Affymetrix, Santa Clara, CA) to block nonspecific Fc binding for 10 min at 4°C, followed by incubation with Abs against CD45 (0.125 μg/ml), CD4 (0.125 μg/ml), and CD11b (0.125 μg/ml) (all from BD Pharmingen, Erembodegem, Belgium) for 30 min. Then, the cells were suspended in permeabilization/fixation buffer (Affymetrix) and incubated with anti-ANKRD55 (8 μg/ml; Santa Cruz Biotechnology, Dallas, TX) for 30 min. A secondary PE-conjugated Ab (Santa Cruz Biotechnology) was used to detect ANKRD55. Debris and duplets were excluded from the analysis, and ≥10,000 events were acquired in each experiment on a FACSAria flow cytometer and analyzed using FACSDiva analysis software (both from BD Biosciences, San Diego, CA). The gating was performed with CD45 Ab and the side scattered light. We analyzed CD4⁺ T cells, monocytes (CD45^+high CD11b⁺), and microglia (CD45^+low CD11b⁺). The geometric mean parameter was used to analyze the fluorescence intensity of ANKRD55.

Analysis of the gene expression data was performed with GraphPad v.6. The Mann–Whitney test (for comparison of two groups), paired t test (for comparison of CD4⁺ and the other subpopulations), or Kruskal–Wallis test (for comparison of three groups) was applied. For correlation analysis among the three transcripts, the Pearson correlation coefficient was used. Statistical comparisons for ddPCR analyses were performed using the Mann–Whitney U test. For mouse experiments, all data are expressed as mean ± SEM, with the exception of the clinical score data, which is presented as mean ± SD. An unpaired two-tailed Student t test or nonparametric Mann–Whitney U test was used to determine the statistical significance in all cases. The level of significance was set at p ≤ 0.05. with p ≤ 0.05 as significant, p ≤ 0.01 as very significant, and p ≤ 0.001 as highly significant.

Because ANKRD55 constitutes a risk gene for different autoimmune diseases affecting diverse organs, it is likely to exert an important function in the immune system; therefore, ANKRD55 gene expression was assessed in human PBMCs. In addition, ANKRD55 isoform expression was assessed in the monocyte-like cell line U937 and the lymphoblastoid cell line Jurkat; SH-SY5Y, a human neuroblastoma cell line, was included to test for possible differential expression patterns of ANKRD55 between immune and nonimmune cells, because of the reported expression of ANKRD55 in human brain (16). A summary of the results obtained using specific pairs of primers (Table I) by conventional RT-PCR is presented in Table II. No amplification of transcripts 002 and 006 was found in any of the studied cell types by means of the primer pairs 2 and 6 and pair 4, respectively. Transcript 001, the canonical full-length form, was not detectable by conventional RT-PCR in any of the cell types using primer pairs 1, 2, or 6, but it was detected with primer pair 7 (Fig. 1, Table I). These primers, spanning exons 3–6 (Fig. 1), were not completely specific to variant 001 and could, in principle, also amplify 002 and 006. However, as shown below, the presence of isoform 001 was unequivocally confirmed with qPCR primer pair 12; it is selective for the exon 7–8 junction in isoform 001 (Fig. 1, Table I), which is missing in 006 and 002. With primer pair 7, a band of the expected size was detected and verified by sequencing in SH-SY5Y cells, and the presence of this transcript was confirmed in an independent experiment using 50 ng of template cDNA (Table II). With 50 ng of template cDNA, a band of the expected size was also detected in U937 and PBMCs (Table II), but it was not observed in any other experiment, suggesting that, in these cell types, the 001 transcript might be expressed at low levels at the limit of sensitivity for conventional RT-PCR. Despite all of these efforts, no amplification of the full-length transcript was detected in Jurkat cells in any of our experiments.

In contrast, the 005 transcript was readily detected in PBMCs using 5 ng cDNA/reaction (Table II). A band of the expected size was consistently observed in various independent experiments using cDNA from MS patients and healthy donors, and sequencing confirmed that it corresponded to the 005 transcript. To further verify that the sequence agreed with that available in Ensembl, two additional pairs of primers were designed to cover the entire open reading frame of the transcript in two overlapping fragments (Fig. 1, Table I). The transcript could also be detected in PBMCs using these primer pairs, and sequencing confirmed its identity as isoform 005. This isoform was also detected in Jurkat and U937 cells, although, with 5 ng cDNA/reaction, the band was much fainter than that observed in PBMCs, and it could not be detected in all of the experiments. No amplification of the 005 transcript was detected in SH-SY5Y cells under any of our experimental conditions (Table II).

Transcript 007 was detected and confirmed by sequencing in PBMCs using 5 ng cDNA/reaction, but no amplification of this transcript was observed in Jurkat, U937, or SH-SY5Y cells in any conditions (Table II).

Next, we screened the coding region of the 005 transcript in MS patients (Table III) for possible nonsynonymous mutations. We resequenced the full coding region of 005 in 12 MS patients selected according to their genotypes for rs6859219 using the primer pairs 8, 9, and 10 (Fig. 1, Table I). Two homozygotes for the protective allele (AA) and five heterozygotes (AC) and five homozygotes for the risk allele (CC) were included. A summary of the results is given in Table IV. No previously undescribed coding mutation was found in any of the patients. A nonsynonymous coding SNP, rs321776, which changes the amino acid at position 57 of isoform 001 from valine to methionine and is classified as benign by SIFT and PolyPhen function prediction tools, was found as heterozygote genotype in one patient who was also heterozygote for rs6859219 and in homozygosis in one patient who was homozygote for the protective (A) allele of rs6859219. Additionally, two homozygotes for rs6859219 were heterozygote carriers of the minor allele of a synonymous coding SNP, rs321775, and one homozygote for rs6859219 was also homozygote for the minor allele of rs321775. In addition, one heterozygote for rs6859219 was a heterozygote carrier of a rare SNP in the 5′ untranslated region, rs77365409, and another heterozygote for rs6859219 was a carrier of a rare synonymous variant, rs60779428. Data from the 2011 genome-wide association study (GWAS) (23) was available to us via the International Multiple Sclerosis Genetic Consortium. To evaluate the potential relevance of the SNPs identified by sequencing, we considered the association results of these SNPs or their proxies in that study. The missense SNP, rs321776, was genotyped in the GWAS, and for the other three, SNPs that were in at least moderate LD were included (for rs321775, r² = 1, D′ = 1; for rs77365409, r² = 0.271, D′ = 1; for rs60779428, r² = 0.428, D′ = 0.734). Because none of these SNPs showed nominal significance (p < 0.05) in the GWAS, which contained 9,772 cases and 17,376 controls, it is very unlikely that they are of relevance; therefore, they were not considered further.

To assess whether ANKRD55 expression was influenced by genotypes of rs6859219 and whether ANKRD55 expression was altered in MS, we performed qPCR on PBMCs of 15 RRMS patients in relapse, 13 RRMS patients in remission, 15 PPMS patients, and 66 healthy controls (14 controls from Barcelona and 52 controls from Bilbao) (Table III). We surveyed the specific expression of each of the transcripts that were detected in the RT-PCR experiments (005, 001, and 007) by using specific pairs of primers (Table I, pairs 11, 12 and 13). No significant differences were found between patients and controls (Fig. 2), RRMS and PPMS patients, or patients in remission and in relapse (data not shown). Separating all individuals by genotypes for rs6859219, we observed a significant association between this SNP and expression of all three transcripts: homozygotes for the risk allele (CC) showed higher expression of ANKRD55 005, 001, and 007 than did carriers of the protective allele (Fig. 2, right panels). The allelic dosage effect was consistent with the genetic risk model for the disease (CC individuals presented a higher odds ratio than CT heterozygotes). Separate analysis of patients and controls did not reach statistical significance, but the same trend was observed in both groups (data not shown). To identify the main cellular sources of ANKRD55 transcripts, PBMCs from 23 healthy controls were separated to enrich for CD4⁺, CD8⁺, CD14⁺, CD19⁺, and CD56⁺ cells and analyzed by qPCR. This showed that the three ANKRD55 transcripts were uniquely and highly expressed in CD4⁺ T cells but were undetectable or minimally expressed in the other subpopulations (Fig. 3). Individual transcript levels were highly correlated, implying that rs6859219 coincides with a cis-expression quantitative trait locus regulating expression levels of the three transcripts in CD4⁺ T cells. Risk CC homozygotes produced significantly higher levels of isoforms 001 and 005 in CD4⁺ T cells. Absolute quantification of transcript copy numbers by ddPCR showed 007 to be the most expressed transcript with 33.6 copies/μl, followed by 001 (16.95 copies/μl) and 005 (5.8 copies/μl) in risk homozygotes (CC genotype, averaged over four healthy controls). Homozygotes for the protective allele produced 5.4-fold fewer copies of 007, 4.6-fold fewer copies of 001, and 8.9-fold fewer copies of 005 (AA, averaged over three healthy controls).

We next assessed the protein expression and subcellular distribution of ANKRD55 in Jurkat, U937, and SH-SY5Y cell lines by immunofluorescence microscopy, nuclear enrichment through sucrose gradient ultracentrifugation, and biochemical fractionation to obtain cytosolic, membranous organelles, and nuclear fractions. In PMA/LPS-treated U937 cells (data not shown) and PMA-treated U937 cells, as well as in SH-SY5Y cells, the cytosol, as well as the nucleus to a greater extent, demonstrated a strong signal for ANKRD55 reactivity (Supplemental Fig. 1). The nuclear signal, but not the cytoplasmic one, largely disappeared, as assessed by microscopy, in the presence of a specific ANKRD55-blocking peptide coinciding with the immunogen used to raise the Ab. Subcellular ANKRD55 distribution was more difficult to interpret in undifferentiated U937 cells (data not shown) and in Jurkat cells (Supplemental Fig. 1) because of the relatively large nucleus. However, the presence of a nuclear ANKRD55 signal was confirmed in all cell types studied. To validate the expression patterns of ANKRD55 observed using confocal microscopy, the subcellular localization of isoforms 001 (UniProt identifier Q3KP44-1) and 005 (UniProt identifier Q3KP44-2) of ANKRD55 in primary CD4⁺ T cells and Jurkat, SH-SY5Y, and U937 cells was assessed in highly enriched nuclear fractions (Fig. 4A). The data confirmed the location of protein isoforms 001 and 005 in the nucleus of U937 and Jurkat cells, as well as primary CD4⁺ T cells, and revealed a high abundance of isoform 001 in the nucleus of SH-SY5Y cells. In SH-SY5Y cells, protein isoform 005 was weakly detectable in cytosol and membrane organelle fractions and was virtually absent from the nuclei obtained through sucrose ultracentrifugation or biochemical fractionation (Fig. 4). Instead, a smaller protein (30 kDa), not seen in any of the other cell types analyzed, was detectable in the nucleus but not in the cytosol and membrane organelles. Thus, this analysis revealed the predominantly nuclear location of ANKRD55 005 and 001.

To further explore the expression of ANKRD55 in the CNS and to assess expression changes in response to inflammation, we undertook experiments in mice. Immunocytochemistry of mouse hippocampal neuronal cell cultures using an Ab that recognizes the mouse isoform 201 showed that ANKRD55 is located primarily in the nucleus (Fig. 5A). Western blot confirmed the presence of a band of 68 kDa (the molecular size corresponding to mouse isoform 201) in this cell type (Fig. 5C). After exposure to LPS (100 ng/ml) for 24 h, the intensity of this band increased significantly (p ≤ 0.001), as revealed by quantification of the Western blot signal (Fig. 5B, 5C). Microglial cells showed strong constitutive expression of ANKRD55, as assessed by immunocytochemistry, Western blot, and ELISA (Fig. 6). Significant upregulation of ANKRD55 (68-kDa band) was observed 24 h after LPS (50 ng/ml) stimulation, as assessed by ELISA (Fig. 6B, p < 0.05, versus control) and by Western blot (Fig. 6C, 6D, p < 0.01, versus control). Immunocytochemistry data indicated that this upregulation occurred principally in the nucleus (Fig. 6A). Similar results were obtained with the BV2 mouse microglial cell line, because combined LPS (50 ng/ml) and IFN-γ (100 U/ml) activation induced a significant increase in protein expression, again primarily in the nucleus (Supplemental Fig. 2, p < 0.01, versus control). Collectively, these data indicate that ANKRD55 is located predominantly in the nucleus of mouse hippocampal neurons and microglia and that inflammatory stimuli induce its overexpression, thus supporting a putative-specific role for this protein in immune-inflammatory–related processes.

We first confirmed ANKRD55 protein expression in the spinal cord of intact animals by mass spectrometry (Supplemental Fig. 3). To determine whether ANKRD55 expression was associated with autoimmunity of the CNS, we immunized C57BL/6 mice with MOG_33–55 to induce the EAE model of MS (Fig. 7A). The brain, spinal cord, and spleen were removed from EAE mice at day 17 postimmunization (peak of the disease occurs between 15–20 d postinfection), when all animals of the group were symptomatic. Interestingly, EAE mice overexpressed ANKRD55 in spinal cord, as assessed by ELISA (Fig. 7B, p < 0.001, versus control) and confirmed by Western blot (Fig. 7C, 7D, p < 0.05, versus control). Brain samples from EAE mice also showed a significant increase in ANKRD55, as measured by a specific mouse ELISA (Fig. 7E, p < 0.05, versus control). Spleen samples were also positive for constitutive ANKRD55 expression; again, its production was increased in mice with EAE (Fig. 7F, 7G, p < 0.05 versus control). In an attempt to determine whether immune cells were contributing to the presence of ANKRD55 in the CNS of EAE mice, we performed new EAE experiments to isolate CNS-infiltrating mononuclear cells at the peak of the disease (17 d postinfection). Flow cytometric analysis of CNS infiltrates indicated (Fig. 8A) that a percentage of CD4⁺ T cells and monocytes was positive for ANKRD55 in EAE mice (Fig. 8B). The intensity of fluorescence was higher in CD4⁺ T cells compared with monocytes. With regard to microglia (intrinsic brain cells), a small percentage of cells appeared positive for ANKRD55, although the intensity of fluorescence was comparable to that of CD4⁺ T cells (Fig. 8B, 8C). Therefore, although intrinsic CNS cells may contribute to the presence of ANKRD55 (as determined by ELISA and Western blot), a portion of ANKRD55 detected in the CNS derives from the immune system. In accordance with human data, CD4⁺ T cells seem to be the cellular population in which ANKRD55 has the strongest expression. We performed a sorting analysis of PBMCs from blood pooled from two EAE mice at the peak of disease; the expression of ANKRD55 was higher in CD4⁺ T cells compared with CD4⁻ T cells, in agreement with human data (6-fold increase).

The first identified MS-associated SNP in ANKRD55, rs6859219 (1, 2), is located in a genomic region with a high recombination rate; it has only three proxies with r² > 0.5 and nine proxies with r² > 0.2. It is interesting to note that all of the SNPs at this locus that are reported to be associated with MS, as well as many of those associated with other autoimmune diseases, correspond to rs6859219 itself or its closest proxies. The most strongly associated SNP at the ANKRD55 locus of the latest comprehensive association study of MS (3) appears to be a perfect proxy of rs6859219. Sequencing of the coding region of transcript 005 in MS patients did not reveal previously undescribed deleterious mutations. Additionally, examination of data from the 2011 International Multiple Sclerosis Genetics Consortium GWAS (23) allowed us to delineate in more detail the association signals in the ANKRD55 region. rs6859219 was not genotyped in this GWAS, but 133 SNPs in a 470-kbp area around ANKRD55 were included. Although none of these SNPs reached genome-wide significance, three had p values < 0.01. Of these, the most significant one, rs10065637 (p = 0.0002), was in perfect LD with rs6859219 (D′ = 1, r² = 1); the other two were in moderate LD with it: rs10040327 (p = 0.004, D′ = 1, r² = 0.57) and rs13328207 (p = 0.001, D′ = 0.759, r² = 0.256). The clustering of association signals in the same narrow LD area suggests that variation in this area is presumably of crucial relevance for the regulation of ANKRD55; our gene-expression data, showing an association between rs6859219 genotype and ANKRD55 expression, argue in favor of this.

The qPCR and, especially, ddPCR results from this study indicate that 001 is the most highly expressed protein-coding transcript in human CD4⁺ T cells, whereas the noncoding transcript variant 007 is by far the most abundant transcript overall. By immunofluorescence microscopy, using an Ab that recognizes 001 and 005 isoforms, ANKRD55 immunoreactivity was detected in the nucleus of SH-SY5Y, Jurkat, and U937 cell lines in the resting state and in PMA-stimulated U937 cells. ANKRD55 reactivity was also present in the cytosol of all of the studied cell types, although, unlike the nuclear signal, the cytosolic one did not disappear upon addition of an ANKRD55-specific blocking peptide. Immunoblotting of sucrose gradient–purified nuclei confirmed the presence of 005 and 001 in the nucleus of primary CD4⁺ T cells and in those of the Jurkat and U937 cell lines, as well as of 001 in the nucleus of SH-SY5Y cells. In SH-SY5Y cells, protein isoform 005 was weakly detectable in cytosol and membrane organelle fractions, but it was virtually absent from the nuclei obtained through biochemical fractionation or ultracentrifugation. Instead, a smaller immunoreactive protein (30 kDa), which is detectable in the nucleus but not in the cytosol and is not seen in any of the other cell types analyzed, may correspond to a lineage-specific immunoreactive truncation product of a larger nuclear isoform because an ANKRD55 splice variant coding for a protein of this size has not been reported. Although the 005 transcript could not be detected in SH-SY5Y and 001 was not detected in Jurkat cells using RT-PCR, both protein forms were observed by immunoblotting in both cell lines. Discrepancies between mRNA and protein levels are well documented. Only ∼40% of the variation in protein concentration is explained by knowing mRNA abundances (24). mRNAs are generally less stable than proteins (average half-life of 2.6–7 h versus 46 h, respectively) and exhibit a larger dynamic range of concentrations (25). Biologically, steady-state levels of proteins are also highly influenced by their functions, with matching stoichiometries between interacting proteins in the same physical complex (26). Although such considerations may go some way toward understanding this discrepancy, our blots show that the protein isoform 005 is only borderline detectable in SH-SY5Y cells. On the basis of these data, even if transcriptional and posttranslational regulation of ANKRD55 differs between immune and nervous cells, both protein isoforms are likely to be essential in both cell types, as well as to survive for some time following termination of their mRNA transcription.

Interestingly, the noncoding transcript variant 007 was easily detected in PBMCs and CD4⁺ T cells, but not in immortalized cell lines, by conventional RT-PCR. ANKRD55 007 is 853 bp long and contains seven exons, some of which partially or completely overlap with the exons of the protein-coding transcripts, whereas others are made up of intronic sequences (Fig. 1). According to these features, it can be categorized as a long noncoding RNA (lncRNA) (27). lncRNAs are RNAs > 200 nt that do not contain an open reading frame; although their biological meaning and function still need to be elucidated, ample evidence, including specific subcellular localization, tissue specificity, and association with disease, indicates that they may be essential regulators of cell function (27–29). The importance of lncRNAs in MS is highlighted by recent studies showing that various SNPs associated with MS in GWAS are located in genomic regions that code for lncRNAs (30) or regulate lncRNA expression in cis in human monocytes (31). Abundant expression of ANKRD55 007 in CD4⁺ T cells may be an indication of the unique cell-specific functional role for this noncoding RNA; it is an intriguing finding that merits further investigation. We were not able to detect expression of isoforms 002 and 006 under any conditions tested. It is of relevance to note that, of the six Ensembl ANKRD55 transcripts, isoforms 002 and 006 were assigned the lowest GENCODE Transcript Support Level of 5 (i.e., evidence that a transcript annotation is de facto expressed in humans based on transcript sequences from the International Nucleotide Sequence Database Collaboration), which is given to transcripts that are not supported at all by an mRNA or an expressed sequence tag. There also is no literature to support these isoforms. Thus, we cannot exclude that the failure to detect these isoforms in our study is simply due to their nonexistence or scarcity.

Importantly, we found that the risk (C) allele of rs6859219 is associated with higher expression of ANKRD55 mRNA in PBMCS and CD4⁺ T cells, qualifying this SNP as a novel cis-expression quantitative trait locus for ANKRD55. The genotype seems to increase the overall expression of the gene but not the splicing or differential expression of transcripts, because the effect was observed for all three transcripts assessed by qPCR (005, 007, and 001). Given that higher levels of ANKRD55 were observed in healthy carriers of the risk allele, and not only in patients, it is conceivable that the high level of ANKRD55 increases susceptibility to pathologic inflammation, rather than being a consequence of it. Therefore, further exploration of a possible proinflammatory effect of ANKRD55 is indicated.

According to the Variant Effect Predictor tool of ENSEMBL, rs6859219 overlaps with a transcription enhancer region in several cell types, including CD14⁺ monocytes and lymphoblastoid cell lines (gm12878). It has only three proxies with r² > 0.5, and all of these overlap with regulatory regions and TF binding sites (Supplemental Table I). Of these, rs6859219 achieves the highest RegulomeDB regulatory score (i.e., 2b) via supporting data for TF binding + any transcriptional motif + DNase footprint + DNase peak, compatible with an intronic enhancer region. Based on chromatin mark datasets from the ENCODE and National Institutes of Health Epigenomics Projects, rs10065637, the perfect proxy of rs6859219 (separated by 271 bp), was found to occur very close to a H3K4me3 chromatin mark (distance of only 5 bp) that is specific to CD4⁺ regulatory T cells and CD4⁺ memory primary cells (height/distance score of 0.80, implying a high cell-type specificity given a height/distance score cut-off of 0.53) (32). In a study on rheumatoid arthritis, rs7731626, which is located in another ANKRD55 intron 6.1 kb from rs6859219, was found to regulate the expression of ANKRD55 in CD4⁺ T cells but not in CD14⁺CD16⁻ cells (5) (no ANKRD55 transcript-specific information). rs7731626 occurs in only partial disequilibrium with rs6859219, and neither the former SNP nor its proxies have emerged as risk SNPs for MS. Thus, rs7731626 may mark a distinct regulatory region, of lesser relevance to MS, although, similar to rs6859219, it is active in CD4⁺ T cells (RegulomeDB regulatory score of 2a, Supplemental Table I). CD4⁺ T cells are considered a crucial natural source of MS risk gene transcripts (23) and are of major relevance in the immunopathogenesis of MS.

In this study, we also characterized the expression of mouse ANKRD55 protein in primary CNS cells and in EAE mice. In particular, we detected the mouse isoform 201 (which shows 84% homology with the human full-length protein 001) by Western blot, and the presence of ANKRD55 in the CNS was confirmed using a specific mouse ELISA. Our results indicate that neurons and primarily microglial cells are producers of ANKRD55 201 and that the expression of this protein is increased under inflammatory conditions. The BV2 microglial cell line showed the same profile of response to inflammation as the primary cultures. The production of ANKRD55 201 by hippocampal neurons and its localization in the nucleus are interesting findings that require further investigation, particularly in relation to its upregulation under inflammatory stimuli. Further studies are ongoing to characterize in detail the subcellular distribution of ANKRD55 in CNS cells to shed some light on its unknown function. We also have preliminary data showing ANKRD55 expression in primary rat astrocyte and oligodendrocyte cultures (data not shown), supporting the relevance of continuing the study of its role in the different CNS cellular types. Importantly, ANKRD55 protein was overproduced in the brain, spinal cord, and spleen of EAE mice compared with healthy animals. The flow cytometry data of CNS infiltrates indicate that immune cells contribute to ANKRD55 in the CNS of EAE mice, particularly CD4⁺ T cells, consistent with the data observed in humans. Even if the mechanisms behind this upregulation remain to be elucidated, this observation provides proof of concept that ANKRD55 is affected by an autoimmune condition of the CNS, supporting future studies on the role of this protein in MS.

In summary, this study contributes toward understanding ANKRD55 in the context of MS and the immune and nervous systems. Our human data confirm the existence of transcript variants 005 and 007 and establish 005 and 001 as the main protein-coding transcripts in immune cells. Correlation of rs6859219 with expression of ANKRD55 in CD4⁺ T cells implies a functional link with MS, and higher expression of the gene in carriers of the risk allele points to a potential proinflammatory role for ANKRD55. Our results in human and mouse cells also provide evidence that ANKRD55 proteins are predominantly nuclear. Pending further investigation, this may hint toward a potential function as a constituent or modulator of transcription complexes via ankyrin repeat–driven protein–protein interactions, demonstrated for nuclear ankyrin repeat domain proteins belonging to the NF-κB and Notch transcription pathways (33). Moreover, we demonstrate that ANKRD55 protein expression is upregulated in response to inflammatory stimuli in murine neurons and microglia and that it is overexpressed in the CNS and spleen of the EAE mouse model of MS.

The authors have no financial conflicts of interest.