Brain Human Monoclonal Autoantibody from Sydenham Chorea Targets Dopaminergic Neurons in Transgenic Mice and Signals Dopamine D2 Receptor: This information is current as Implications in Human Disease of September 26, 2021. Carol J. Cox, Meenakshi Sharma, James F. Leckman, Jonathan Zuccolo, Amir Zuccolo, Abraham Kovoor, Susan E. Swedo and Madeleine W. Cunningham J Immunol 2013; 191:5524-5541; Prepublished online 1 Downloaded from November 2013; doi: 10.4049/jimmunol.1102592 http://www.jimmunol.org/content/191/11/5524 http://www.jimmunol.org/

Supplementary http://www.jimmunol.org/content/suppl/2013/11/01/jimmunol.110259 Material 2.DC1 References This article cites 84 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/191/11/5524.full#ref-list-1 by guest on September 26, 2021 Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Brain Human Monoclonal Autoantibody from Sydenham Chorea Targets Dopaminergic Neurons in Transgenic Mice and Signals Dopamine D2 Receptor: Implications in Human Disease

Carol J. Cox,* Meenakshi Sharma,† James F. Leckman,‡,x,{ Jonathan Zuccolo,* Amir Zuccolo,* Abraham Kovoor,† Susan E. Swedo,x,‖ and Madeleine W. Cunningham*

How autoantibodies target the brain and lead to disease in disorders such as Sydenham chorea (SC) is not known. SC is charac- terized by autoantibodies against the brain and is the main neurologic manifestation of streptococcal-induced rheumatic fever.

Previously, our novel SC-derived mAb 24.3.1 was found to recognize streptococcal and brain Ags. To investigate in vivo targets Downloaded from of human mAb 24.3.1, VH/VL genes were expressed in B cells of transgenic (Tg) mice as functional chimeric human VH 24.3.1– mouse C-region IgG1a autoantibody. Chimeric human–mouse IgG1a autoantibody colocalized with tyrosine hydroxylase in the a basal ganglia within dopaminergic neurons in vivo in VH 24.3.1 Tg mice. Both human mAb 24.3.1 and IgG1 in Tg sera were found to react with human dopamine D2 receptor (D2R). Reactivity of chorea-derived mAb 24.3.1 or SC IgG with D2R was confirmed by dose-dependent inhibitory signaling of D2R as a potential consequence of targeting dopaminergic neurons, reaction with surface-

exposed FLAG epitope-tagged D2R, and blocking of Ab reactivity by an extracellular D2R peptide. IgG from SC and a related http://www.jimmunol.org/ subset of streptococcal-associated behavioral disorders called “pediatric autoimmune neuropsychiatric disorder associated with streptococci” (PANDAS) with small choreiform movements reacted in ELISA with D2R. Reaction with FLAG-tagged D2R distinguished SC from PANDAS, whereas sera from both SC and PANDAS induced inhibitory signaling of D2R on transfected cells comparably to dopamine. In this study, we define a mechanism by which the brain may be altered by Ab in movement and behavioral disorders. The Journal of Immunology, 2013, 191: 5524–5541.

ydenham chorea (SC) is well established as the major neu- streptococcal pharyngitis (3, 4), tics and obsessive-compulsive rologic sequela of Streptococcus pyogenes–induced rheumatic disorder (OCD) appear to be associated with streptococcal and S fever (1) and is characterized by involuntary movements other types of infections (5–9), as well as with autoantibodies by guest on September 26, 2021 and neuropsychiatric disturbances, including obsessive-compulsive against neuronal Ags (10–14). symptoms and hyperactivity (2). Although SC is associated with Movement, behavioral, and neuropsychiatric disorders affect millions of children worldwide. Studies suggesting that infection and autoantibodies might contribute to the pathogenesis of some *Department of Microbiology and Immunology, University of Oklahoma Health Sciences Center, Biomedical Research Center, Oklahoma City, OK 73104; groups of movement and behavioral disorders began with studies †Department of Biomedical and Pharmacological Sciences, College of Pharmacy, of SC and its relationship to streptococcal infection and auto- University of Rhode Island, Kingston, RI 02881; ‡Yale Child Study Center, Yale University School of Medicine, New Haven, CT 06519; xDepartment of Pediatrics, antibodies against the brain in streptococcal-induced rheumatic Yale University School of Medicine, New Haven, CT 06519; {Department of Psy- fever (15–19). However, actualization of this hypothesis was not ‖ chiatry, Yale University School of Medicine, New Haven, CT 06519; and Pediatrics recognized until recently when it was discovered that SC was and Developmental Neuroscience Branch, National Institute of Mental Health, De- partment of Health and Human Services, Bethesda, MD 20892 treatable by plasmaphoresis, which correlates with removal of Received for publication September 9, 2011. Accepted for publication September 26, autoantibodies against the brain (20, 21). Other childhood dis- 2013. eases, including some groups of OCD or tics, have also been re- This work was supported by Grants HL35280 and HL56267 from the National Heart, lated to streptococcal infections, with subsequent development Lung, and Blood Institute, a supplement from the National Institute of Mental Health, of autoantibodies (10, 13, 20, 22–25). These disorders may be and a grant from the Oklahoma Center for Advancement of Science and Technology. M.W.C. is the recipient of a National Institutes of Health Merit Award. J.F.L. receives identified as pediatric autoimmune neuropsychiatric disorder as- grant support from Grifols and the National Institutes of Mental Health Intramural sociated with streptococci (PANDAS) (22) or pediatric acute- program for a randomized clinical trial of i.v. Ig treatment for children with pediatric onset neuropsychiatric syndrome in the presence of other types autoimmune neuropsychiatric disorder associated with streptococci being conducted at the National Institutes of Health Clinical Center under the leadership of S.E.S. of infections (23, 26). Address correspondence and reprint requests to Dr. Madeleine W. Cunningham, The basal ganglia has been implicated as a target of poststrep- Department of Microbiology and Immunology, University of Oklahoma Health Sci- tococcal immune responses (25, 27). Streptococcal-specific Abs ences Center, Biomedical Research Center, Room 217, 975 NE 10th Street, Okla- in SC were shown to react with neurons in human basal ganglia, homa City, OK 73104. E-mail address: [email protected] and neuronal-specific Ab titers were associated with both severity The online version of this article contains supplemental material. and duration of choreic episodes (15, 16). Husby et al. (16) found Abbreviations used in this article: ADHD, attention deficit hyperactivity disorder; that chorea patient sera reacted strongly with cytoplasmic, but not BBB, blood–brain barrier; CaM kinase II, calcium-calmodulin dependent protein kinase II; D1R, dopamine D1 receptor; D2R, dopamine D2 receptor; GlcNAc, N- nuclear, Ags in human caudate and subthalamic nuclei, as well as acetyl-b-D-glucosamine; HC, H chain; LC, L chain; OCD, obsessive-compulsive cerebral cortex neurons. The antibrain reactivity correlated with disorder; PANDAS, pediatric autoimmune neuropsychiatric disorder associated with streptococci; SC, Sydenham chorea; Tg, transgenic; TH, tyrosine hydroxylase; both severity and duration of choreic symptoms. Immunomodu- TRITC, tetramethyl rhodamine isothiocyanate; TS, Tourette’s syndrome. latory therapies, such as plasma exchange, provide strong evi- www.jimmunol.org/cgi/doi/10.4049/jimmunol.1102592 The Journal of Immunology 5525 dence that SC is caused by a pathogenic Ab response (20). Other behavioral abnormalities. All human subjects gave informed consent, and groups of neuropsychiatric and movement disorders, including studies were approved by human subjects protection committees at the PANDAS, may also be associated with an autoantibody response University of Oklahoma Health Sciences Center, Yale University, and the National Institute of Mental Health. Eight of ten SC, 27/27 PANDAS, 5/5 against the brain (13, 28) and abnormal brain magnetic resonance ADHD, and 18/19 control samples were presented in a different manner imaging (29). Improvement was also seen in these cases after by Brimberg et al. (35), who did not include comparison with Tourette’s plasmaphoresis (21, 30). syndrome (TS)/OCD sera. The hypothesis that SC and its pathogenesis may develop as Construction of Tg vectors a result of signaling autoantibodies directed against neuronal cells in the brain was delineated by studies using SC-derived human Rearranged Ig HC and LC V genes of human mAb 24.3.1 were cloned previously, and the VH and VL sequences are published (11). Tg constructs mAbs. SC mAb cross-reactivity was first associated with the group were designed to produce a chimeric Ab (human V-gene/mouse C-region A streptococcal carbohydrate epitope N-acetyl-b2D-glucosamine Ab) when expressed in Tg mice. (GlcNAc) and neuronal surface Ag lysoganglioside GM1 and a he- 24.3.1 HC Tg construct. A 2.8-kb BamHI-EcoRI fragment of plasmid lical intracellular brain protein tubulin (10, 11). The cross-reactivity pG11-LHCuM (provided by Dr. Brett Aplin, University of Melbourne, between GlcNAc and a helical peptide structures has been studied Melbourne, Victoria, Australia) was excised and religated. This contains Ig extensively (31–33). Chorea-derived human anti-streptococcal promoter, leader, and enhancer regions, as well as a unique SalI restriction site for cloning a V-gene insert (36). The mAb 24.3.1 rearranged VH gene, mAb 24.3.1 also induced increased calcium-calmodulin depen- previously cloned into a PCR vector, was amplified using the following dent protein kinase II (CaM kinase II) activity in a human neu- primers: 24HCTg59,592CGCGTCGACTTGCTTTTCTGTTTCCAAGC- ronal cell line (10), and intrathecal passive transfer of mAb 24.3.1 AGGGGTCAATTCAGGTGCAGCTGGTG-39 and 24HCTg39,592CGC- induced elevated tyrosine hydroxylase (TH) in rat brains (12). GTCGACAGATCTAGCACACAGACCCTATTACTCACATACCTGCA- Downloaded from GAGACAGTGAGGAGACGGTGAC-39. These primers contain consensus mAb 24.3.1 exhibited stronger avidity for neuronal Ags, which splice acceptor or donor sequences plus SalI restriction sites (bold type) may explain its ability to signal neuronal cells, but the extracel- for cloning. The 456-bp amplified PCR product was digested with SalI lular targets responsible for the signaling were not defined. SC (Promega) and cloned into the corresponding SalI site in the modified Tg mAb specificities were similar to the neuronal-specific IgG in sera vector, creating a 3.2-kb plasmid. Verification of the presence of the 24.3.1 or cerebrospinal fluid from active chorea (10). Ab-mediated cell VH insert was done by diagnostic restriction enzyme digestion, as well as by sequencing of PCR amplification products. A mouse HC construct signaling leading to excess dopamine release was suggested to be containing the IgG1 constant regions and membrane exons was kindly http://www.jimmunol.org/ involved in the immunopathogenesis of SC, as well as other CNS provided by Dr. Chris Goodnow (John Curtin School of Medical Research, disorders, but the functional in vivo targets explaining mecha- The Australian National University, Canberra ACT, Australia). Experi- nisms in disease were not identified. ments with this vector were described previously (37); parts of this vector are the same as pG11-LHCuM. A 9.5-kb region of the IgG1 plasmid, To gain further insight into in vivo functional Ab targets that may flanked by HindIII restriction sites, contains the Sg1 region, the constant be involved in the mechanisms of SC and related CNS disorders, and hinge exons, and the membrane exons. Primers were designed to amplify we created transgenic (Tg) mice expressing chorea-derived human this region and the upstream regions, based on the published genomic sequence (GenBank). Three regions (including the cloned 24.3.1 V gene mAb 24.3.1 H chain (HC) and L chain (LC) V-region (VH and VL) H genes as part of a chimeric (human V-gene/mouse C-region) IgG1a site) were PCR amplified and tandemly ligated, as previously described (38), to create the 24HCTg Tg construct. Ab. Mice Tg for mAb 24.3.1 V genes were validated by charac- by guest on September 26, 2021 24.3.1 LC Tg construct. The LC Tg vector, pSV-LkCkΔJ4J5, generously teristic cross-reactive anti-neuronal Ab specificities in serum and provided by Dr. Brett Aplin (University of Melbourne, Melbourne, VIC, of mAbs produced from lymphocytes of Tg mice. Chimeric 24.3.1 Australia), is an 11.4-kb plasmid containing a 6.4-kb region with the Tg Ab penetrated the brain and dopaminergic neurons in the basal LC promoter and leader sequences, the SalI restriction site for cloning ganglia of Tg mice, most likely in the substantia nigra or ventral VL gene, and the mouse k C region. Amplification primers, TgLK-F5 (59- CGGTCTGTGAAAAACCCCT-39) and TgLK-R4 (592TCCCCCTGAA- tegmental area. However, not all neurons targeted were dopami- CCTGAAACATA-39), which spanned this region, were designed based nergic, and a significant portion of them were located in the cortex. on known plasmid and insert DNA sequences; this 6.5-kb fragment was Furthermore, human mAb 24.3.1 was similar to sera from SC and amplified by PCR and subsequently cloned into the PCR cloning vector, other movement and neuropsychiatric disorders, which had ele- pCR2.1 TOPO-TA (Invitrogen), creating the LKTA vector. Verification of vated IgG to human dopamine D2 receptor (D2R). SC-derived the presence and orientation of insert was done by diagnostic restriction enzyme digestion, as well as by sequencing of PCR amplification products. mAb 24.3.1 and SC IgG reacted with FLAG-tagged D2R, whereas The 24.3.1 VL gene was amplified using primers 24LCTg59 (59-CGCG- PANDAS IgG did not. However, human chorea mAb 24.3.1, sera TCGACTTGCTTTTCTGTTTCCAAGCAGGAGCCATCGGGGAAATT- from the human SC mAb B cell donor, and PANDAS sera all in- AGTGATGACG-39) and 24LC39 (592CGCGTCGACAGATCTAGC- duced inhibitory signaling of D2R, which may be the consequence ACACAGACCCTATTACTCACATACGTTTGATCTCCAGCTTGGTCC- C-39), containing consensus splice acceptor or donor sequences and SalI of Ab targeting of dopaminergic neurons. Our novel findings restriction sites (bold type), which yield a 456-bp product. The VL gene suggest that signaling autoantibodies may play a role in chorea amplification product was purified and digested with SalI (Promega). The and other movement and behavioral disorders. digested DNA was then ligated into the corresponding SalI site of the LKTA vector, creating the 24LC-LKTA Tg vector. Correct orientation of the 24.3.1 LC V-gene insert was verified by DNA sequencing using pri- Materials and Methods mers flanking the insert, including TgLK-For24 (59-ATGGGCATCAAA- Human blood samples ATGGAGTCA-39) and 24LCRev1 (59-AGGCTGCTGATGGTGACAGT- 39). Expression of the chimeric human V-gene/mouse C-region Ab from Blood samples from SC and PANDAS patients and healthy controls were the 24.3.1 transgenes was tested by transient transfection into mouse mye- obtained from the National Institute of Mental Health (S.E.S.), the Child loma cells, as described below. Study Center, Yale University (J.F.L.), and the University of Oklahoma Health Sciences Center (M.W.C.). SC was diagnosed using Jones Criteria. Generation of Tg mice PANDAS was diagnosed using the published National Institute of Mental Health criteria (22) and included patients displaying fine piano playing Purified 24.3.1 HC and LC Tg constructs (13.8-kb and 6.4-kb BamHI DNA movements of the fingers and toes in addition to obsessive-compulsive fragments) were used for pronuclear microinjection into C57BL/6 mouse symptoms (24). Attention deficit hyperactivity disorder (ADHD) was di- embryos at Xenogen Biosciences/Caliper Life Sciences (Cranbury, NJ). agnosed on the basis of a structured psychiatric interview (Kiddie- Genomic DNA used to create the HC Tg vector was derived from a BALB/C Schedule for Affective Disorders and Schizophrenia or Diagnostic Inter- mouse (IgG1a allotype); having the transgenes in C57BL/6 mice (IgG1b view for Children and Adolescents) and the Connors Parent and Teacher allotype) allows for detection of the HC transgene by ELISA and FACS, Rating Scales (34). Healthy controls had no evidence of neurologic or and most importantly, in vivo in the brain. The 24.3.1 chimeric transgenes 5526 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS were coinjected, and litters consisting of a total of 65 mice were born. Invitrogen. Commercial universal forward and reverse primers for M13 Screening for incorporation of transgene(s) was done by tail DNA PCR (Promega) were also used. using mAb 24.3.1 V-gene–specific primers. Nine founder mice were generated that were either HC (VH) only or double-integration positive Tg DNA sequencing (V +V ). Southern hybridization was performed to confirm correct ori- H L Sequencing was performed by the University of Oklahoma Health Sciences entation of the transgene(s) in the genome and to determine copy number. Center Laboratory for Genomics and Bioinformatics. Nucleotide align- Sequencing of PCR products confirmed that the inserted VH and VL genes in founder animals were correct and unmutated. Sera from founders were col- ments were performed using the basic local alignment search tool analysis lected and screened by ELISA to verify expression of the chimeric IgG1a Ab. program (National Center for Biotechnology Information), IMGT/V-QUEST Founders were maintained at Xenogen and backcrossed to C57BL/6 mice. Tg (International Immunogenetics Information System), and WUR MUSCLE mice were shipped to the University of Oklahoma Health Sciences Center multiple alignment analysis program (http://www.bioinformatics.nl/tools/ muscle.html). for experimental studies and housed at the University of Oklahoma Health Sciences Center Animal Facility. Negative littermates or C57BL/6 wild-type Transient transfections mice (Taconic) were used as experimental controls. All experiments on animals were performed in accordance with relevant institutional (National Sp2/0 mouse myeloma cells were maintained in culture in IMDM (Life Institutes of Health and University of Oklahoma Health Sciences Center) Technologies) containing 10 IU/ml penicillin/streptomycin (Life Tech- guidelines and regulations as set by the University of Oklahoma Health nologies), 10 mg/ml gentamicin (Sigma), and 10% (v/v) heat-inactivated Sciences Center Institutional Animal Care and Use Committee. FCS (Hyclone). Cells were fed 18 h prior to transfection. A total of 10–15 mg each DNA species to be transfected and 1.5 mg the plasmid were DNA isolation and purification linearized separately by restriction enzyme digestion and purified using Small-scale plasmid DNA isolation was performed using a Wizard Plus Amicon Micropure-EZ centrifugal filter devices (Millipore). Electro- Minipreps DNA Purification system (Promega). Endofree Plasmid Maxi poration was performed using parameters previously published for Sp2/0 3 7 Downloaded from kit (QIAGEN) and PureYield Plasmid Midiprep and Maxiprep Systems cells (39). Each linearized DNA sample was added to 1.6 10 Sp2/0 (Promega) were used for medium- and large-scale plasmid DNA prepa- cells, and samples were incubated on ice for 10 min. The cells were then m ration for cloning, cell transfections, and transgenesis microinjection. electroporated at 960 F, 2 kV/cm using Gene Pulser Xcell (Bio-Rad). After an additional 10-min incubation on ice, each set of cells was Genomic DNA from mouse tail tissue samples was purified using the Easy- transferred to a plastic media dish containing 3 ml prewarmed IMDM with DNA isolation kit (Invitrogen). All kits were used following the manu- facturers’ instructions. DNA constructs used for microinjection were fur- 20% FCS. Cells were cultured at 37˚C overnight, after which the super- ther purified by dialysis against TE buffer (10 mM Tris, 0.25 mM EDTA natant was collected for ELISA, and transfectants were collected and stored in TRI Reagent (Molecular Research Center) at 280˚C for subse- [pH 7.5]) using Millipore V-series mixed cellulose ester microdialysis http://www.jimmunol.org/ membranes (0.025 mm) following the transgene preparation protocol quent RNA extraction and RT-PCR. supplied by Xenogen. Concentration and purity of DNA samples were Southern hybridization determined by spectrophotometric analysis at 260/280 nm using a Synergy HT Multidetection microplate reader (Bio-Tek) with KC4 software. Southern blotting was performed to determine orientation and copy number of transgenes. Purified genomic DNA from Tg mice and control C57BL/6 PCR wild-type mice was provided by Xenogen. Hybridization was done using Stratagene’s Illuminator Chemiluminescent Detection System, following PCR amplification of DNA was performed using 0.1–1 ng plasmid template the Southern blot procedure described in the accompanying manual. DNA or 50–400 ng genomic template DNA and 0.5 mM59 and 39 oli- Probes were purified PCR products of the variable regions from the 24.3.1 gonucleotide primers, using Platinum Taq PCR Supermix, High Fidelity HC or LC (24HCTg59–39 and 24LCTg59–39). (Invitrogen). Amplification was performed using a Gene-Amp PCR System by guest on September 26, 2021 2700 (Applied Biosystems). Cycling parameters were dependent on the RNA isolation and RT-PCR primers and template being used, but followed this general profile: 94˚C for 1 min, followed by 30–45 cycles of 94˚C for 15–30 s, 50–65˚C for Total RNA was isolated from whole spleen or tissue culture cell lines using 15–30 s, 72˚C for 1 min/kb, and then 10 min at 72˚C. PCR amplification TRI Reagent (Molecular Research Center, Cincinnati, OH) following the products were separated by electrophoresis on 1% (w/v) Tris HCl–borate– protocol provided. Contaminating DNA was removed by treatment with EDTA ethidium bromide–containing agarose gels and visualized under UV DNase I (1 U/mg RNA) (Invitrogen), per the manufacturer’s protocol. RT- light, as per standard molecular biological procedures. PCR was performed using the Superscript III One-step RT-PCR System with Platinum Taq DNA Polymerase (Invitrogen), using either gene- DNA cloning specific primers or Ig-Primer Sets (human or mouse) from Novagen. Reactions were performed on a Gene-Amp PCR System 2700 (Applied Molecular weight markers, restriction endonucleases, and DNA-modifying Biosystems) using the following cycling profile: 45˚C for 30 min (first- enzymes (Promega, Invitrogen, New England BioLabs) were used ac- strand cDNA synthesis reaction), denaturation at 94˚C for 2 min, followed cording to the manufacturers’ instructions. Insert DNA and/or plasmids for by 50 cycles of 94˚C for 15 s, 50˚C for 30 s, 68˚C for 1 min, and a 5-min cloning were digested with appropriate restriction enzymes and size sep- extension at 68˚C. Results of amplification were assessed by visualizing arated by agarose gel electrophoresis. Digested DNA samples were sepa- an aliquot of the RT-PCR products (10 ml each) on an ethidium bromide– rated on 0.8–1.25% modified TAE (40 mM Tris-acetate, 0.1 mM Na2 stained 1% Tris HCl–borate–EDTA agarose gel. The remainder of the EDTA [pH 8]) agarose gels, and DNA bands were excised and purified amplification products were purified by column filtration or gel extraction, using the Montage DNA Gel Extraction kit (Millipore) or the QIAquick as described above. Purified RT-PCR products were either cloned into Gel Extraction Kit (QIAGEN). DNA samples were concentrated using plasmid vectors or submitted for sequencing. Microcon YM-50 centrifugal filter units (Millipore) if necessary. PCR am- plification products (for cloning or direct sequencing) were purified by the Treatments to open blood–brain barrier above-mentioned gel-extraction methods or by Montage PCR filter units (Millipore). Small DNA products (,1 kb), which were cloned into the To open the blood–brain barrier (BBB) and stimulate Tg B cells, Tg and pCR2.1-TOPO TA vector (Invitrogen), were transformed into One Shot non-Tg littermates were immunized with streptococcal cell wall and LPS, TOP10 Chemically Competent Escherichia coli cells (Invitrogen), following following the method of Kowal et al. (40). Serum was collected before and the provided protocols. Inserts cloned into Tg or other vectors were ligated after immunization and analyzed by capture ELISA using anti-mouse using T4 DNA ligase (Invitrogen), following the recommended insert/vector IgG1a. Animals were sacrificed at the end of the experiments, brains DNA ratios. Larger DNA inserts and DNA fragments cloned into large Tg were collected and placed in formalin for histopathological analysis and vector constructs with a propensity for recombination were transformed into immunohistochemical staining, and serum was collected for ELISA. Spleens One Shot OmniMax 2 T1 Phage–resistant cells (Invitrogen) or SURE cells were collected from Tg mice and used for RNA extraction/RT-PCR, FACS (Stratagene), following the suppliers’ protocols. Growth and propagation of analysis, and hybridoma production. Peripheral lymph nodes and bone transformed cells and selection of positive transformants were performed marrow were harvested for FACS analysis. following standard molecular biological protocols. ELISAs Oligonucleotide primers Ninety-six–well polyvinyl, Immunolon-4 microtiter plates (Dynatech Lab- Primers were designed according to known sequence, using the Primer 3 oratories) were coated with 10 mg/ml Ag overnight at 4˚C. Plates were (v.0.4.0) primer design program (SourceForge) and were synthesized by reacted with sera or hybridoma supernatant overnight at 4˚C. Ab binding The Journal of Immunology 5527 was detected using alkaline phosphatase conjugated to specific sec- with FBS stain buffer after incubations; sequential incubations were per- ondary Ab or to streptavidin. Plates were developed with 1 mg/ml p- formed for cell samples that were stained with more than one fluoro- nitrophenylphosphate-104 substrate (Sigma) at 50 ml/well. OD was quanti- chrome. All incubations were conducted on ice in the dark. Wash steps fied at 405 nm in an Opsys MR microplate reader (Dynex Technologies) or were carried out by adding 1 ml FBS stain buffer to each tube, followed by a Synergy HT Multidetection microplate reader (Bio-Tek) using KC4 soft- centrifugation at 1000 rpm for 5 min at 4˚C in a Beckman Coulter Allegra ware. Capture ELISAs were performed using purified mouse anti-mouse 25R centrifuge. Cells were labeled with Abs coupled to biotin (anti-mouse IgG1a (Igh-4a) mAb (BD Pharmingen) or goat anti-mouse IgG1 (Sigma) IgG1a, anti-mouse IgG1b) or one of the following fluorochromes: PerCP- as the coating Ab. Biotin-conjugated mAbs [mouse anti-mouse IgG1a, mouse Cy5.5 (anti-mouse B220), R-PE (anti-mouse CD23), or allophycocyanin anti-mouse IgG1b (Igh-4b), rat anti-mouse Igk (BD Pharmingen)] diluted (anti-mouse CD21). Biotin-conjugated Abs were also incubated with 1/5000 in 1% BSA-PBS (or 1/500 for anti-Igk)] were used as secondary Ab streptavidin-FITC. All labeled Abs were purchased from Pharmingen. where appropriate. Biotinylated conjugates were further incubated with Isotype-control Abs (Biotin Mouse IgG2a k, Biotin Mouse IgM k, PerCP- ExtrAvidin-Alkaline Phosphatase (Sigma) at the same dilution. Ag panel Cy5.5 Rat IgG2a k, PE-Rat IgG2a k, allophycocyanin Rat IgG2b k) were ELISA testing included ds- and ssDNA (Invitrogen), human dopamine used at the same concentrations as the Abs of interest. After labeling, cells receptor D2L (D2R long isoform; PerkinElmer) (41), and b-tubulin (MP were resuspended in 500 ml13 fixative containing 1.4% formaldehyde Biomedical), as well as rabbit skeletal tropomyosin, mouse laminin, and (R&D Systems). Labeled cells were stored in the dark at 4˚C overnight and BSA (all from Sigma). Also tested in the ELISA were pepsin cleavage then analyzed on a FACSCalibur automated four-color benchtop flow products of streptococcal M5 protein (PepM5) and GlcNAc-BSA, which cytometer (BD Biosciences) at the University of Oklahoma Health Sci- were prepared as previously described (42, 43). Lysoganglioside (Sigma) ences Center Flow and Image Cytometry Laboratory. Dead and irrelevant was also tested and was coated at 20 mg/ml on microtiter plates, as pre- cell populations were excluded by setting gates on the basis of forward and viously described (10). Biotin-conjugated mAb mouse anti-mouse IgG1a side scatter profiles. Stained cells were analyzed initially to estimate the was used as secondary Ab in the Ag panel ELISA. Alkaline phosphatase– percentage of cells expressing B220, a marker present on cells committed conjugated anti-human IgM (Sigma) was used for detection of human mAb to the B lineage. B cell subpopulations were identified by anti–B220-

24.3.1. For testing of human sera for the presence of D2R Abs, plates were PerCP-Cy5.5 in combination with fluorescently labeled Abs against the Downloaded from coated with human dopamine receptor D2L (PerkinElmer) at 10 mg/ml. other cell surface markers, anti-CD23 and anti-CD21, to characterize the Alkaline phosphatase–conjugated goat anti-human IgG (g-chain specific) cells with regard to maturation level and location (i.e., marginal zone or F(ab9)2 fragment (Sigma) was used for detection. To determine the D2R follicular). Computer analysis was done using Summit v 4.3 software Ab titer, sera were diluted 1/100 in 1% BSA in PBS buffer and subse- (Dako Colorado). quently diluted 2-fold, and ELISA was performed as described above. For the D2R Ab titers, tests were performed separately three times, and the Immunohistochemistry titration was done in duplicate in each assay. Titers were determined as the highest dilution with OD value of 0.10 at 2 h. Controls used included Colocalization studies were performed using paraffin-embedded Tg mouse http://www.jimmunol.org/ brain tissue sections. Brain sections were tested to determine whether Tg Ab serum alone, secondary Ab conjugate alone, and 1% BSA alone. IgG1a expressed in vivo by B cells deposited in the brains of Tg mice. Competitive-inhibition ELISA with D2R peptide Formalin-fixed mouse brains were cut, mounted on slides, and stained following previously described protocols (11, 44). Briefly, tissue was fixed Competitive-inhibition ELISA was performed in duplicate, as previously overnight in 10% buffered formalin, sectioned (5-mM thick), and mounted described (32, 44). Concentration of mAbs was 100 ng/ml. Peptide in- onto Microprobe probe-On Plus slides. Mounted tissue was baked for 60 hibitors were prepared as 1-mg/ml solutions in PBS (pH 7.2) and serially min and deparaffinized, followed by rehydration using graded ethanol. diluted from 500 to 1 mg/ml. The diluted inhibitors were mixed with an Heat-mediated Ag retrieval was performed using a rice cooker and sodium equal volume of each mAb and incubated at 37˚C for 1 h, followed by an citrate buffer (10 mM sodium citrate, 0.05% Tween 20 [pH 6]). Slides were overnight incubation at 4˚C. The mAb-inhibitor mixture was added to 96- heated for 20 min and allowed to cool to room temperature. Slides were by guest on September 26, 2021 well microtiter plates coated with human dopamine receptor D2L (Perkin- then washed twice in PBS (pH 7.4) and blocked with a protein blocker for Elmer) at 10 mg/ml in bicarbonate buffer. Samples were allowed to in- 15 min at room temperature, followed by two additional washes in PBS. cubate overnight at 4˚C. Ab binding was detected using alkaline Colocalization by double immunostaining used Abs conjugated to FITC phosphatase–conjugated anti-human IgM (Sigma). The remainder of the and tetramethyl rhodamine isothiocyanate (TRITC). For localization of a a assay was performed as described above. Maximal (100%) reactivity was chimeric IgG1 Ab, biotin-conjugated mouse anti-mouse IgG1 was used determined by mAb–PBS incubation without inhibitors. Mouse mono- (5 mg/ml; BD Pharmingen) in combination with FITC-conjugated strep- clonal anti-human D2R Ab clone 1B11 was purchased from Sigma and tavidin (1:20; Invitrogen). For immunostaining of dopaminergic neurons, was detected using alkaline phosphatase–conjugated goat anti-mouse IgG rabbit polyclonal Ab to TH (Sigma) was used as the primary Ab (1:100), (Fab specific; Sigma). Mouse mAb 1B11 was produced after immunization with TRITC-conjugated sheep anti-rabbit polyclonal Ab as the secondary of mice with human D2R residues 1–110. Ab (1:100; Sigma). For immunostaining of D2R, rabbit anti-dopamine D2 receptor polyclonal Ab (Abcam) was used as the primary Ab (1:35), with Human D2R peptide sequences TRITC-conjugated sheep anti-rabbit polyclonal Ab mentioned above used as the secondary Ab (1:100). Tissue staining by primary Abs was done by D2R peptides were synthesized by ImmunoKontact (AMS Biotechnology, overnight incubation at 4˚C, and fluorescently labeled secondary Abs were Lake Forest, CA) and purified by HPLC. The sequences and designations incubated on tissues for 1 h at room temperature in a humidified chamber are as follows: DRD2 E1.1, MDPLNLSWYDDDLERQNWSR and DRD2 in the dark on the following day. Washes were done in PBS (pH 7.4). For E1.2, SRPFNGSDGKADRPHYNYYA. Control peptide was a previously positive control for anti-mouse IgG1a staining, mouse brain tissue was described human cardiac myosin peptide: FTRLKEALEKSEARRKE- incubated with IgG1a Ab (mouse mAb to neuron-specific b-III tubulin LEERMVSL (45, 46). [Abcam]) (1:500) for 2 h, followed by washes and incubation with biotin- labeled anti-mouse IgG1a and streptavidin-FITC, as previously described. Flow cytometry Mouse liver tissue was used as negative control for TH Ab staining. Prior to viewing, slides were treated with Prolong Gold Anti-fade reagent with Lymphocytes from spleen, bone marrow, and lymph nodes were fluo- DAPI (Invitrogen), per the manufacturer’s recommended protocol, to re- rescently labeled and analyzed by FACS. Spleens, lymph nodes, and bone duce photobleaching and stain nuclei. Slides were viewed with an Olym- marrow from Tg mice and non-Tg littermates were harvested, and single- pus BX41 microscope equipped with DAPI, FITC, and TRITC filters, and cell suspensions were prepared. Spleens were pushed through autoclaved micrographs were imaged with a SPOT InSight 2 FireWire Camera. screens to homogenize tissue. Isolation of spleen cells and cells from Images were merged using SPOT Software V. 4.6 (Diagnostic Instru- peripheral lymph nodes was performed using Histopaque-1083 Hybri-Max ments). Colocalization studies were performed on paraffin-embedded (Sigma). Bone marrow was isolated from mouse femurs as follows: heads normal mouse brain tissue sections using SC-derived human mAbs 24.3.1, of femurs were cut off and discarded, and bones were placed in 0.6-ml 37.2.1,and31.1.1(100ng/ml).DetectionofhumanmAbswasbyFITC- centrifuge tubes, each with a hole in the end made by an 18-gauge needle. conjugated goat anti-human IgM (1:50; Jackson ImmunoResearch). Incu- Each tube was placed in a 1.5-ml centrifuge tube and pulse centrifuged for bation with media only (IMDM supplemented with 10% FCS) was per- 30 s. RBCs in the collected bone marrow were removed by treatment with formed as a control. the Mouse Erythrocyte Lysing Kit (R&D Systems), following the manu- facturer’s protocol. Cells from spleen, bone marrow, and lymph nodes Generation of chimeric Ab–secreting hybridomas were resuspended at 1.0 3 106 cells/ml in 500 mlFBSstainbuffer (Pharmingen) and incubated for 30 min with the appropriate biotinylated Spleens from Tg mice were harvested and mashed through screens using or directly fluorescein-labeled Ab for FACS analysis. Cells were washed Histopaque-1083 Hybri-Max (Sigma). Cells were fused with K6H6/B5 5528 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS hybridoma fusion partner, a nonsecreting human 3 mouse heterohybridoma chemiluminescence was generated using SuperSignal West Femto Chemi- line (ATCC CRL-1823; American Type Culture Collection), using poly- luminescent Substrate (Pierce) and measured using a GloMax Microplate ethylene glycol 1000, as previously described (32). Hybridomas were se- Luminometer (Promega). To normalize the signal for cell number, cells lected using hypoxanthine, aminopterin, and thymidine–supplemented were incubated with ethidium bromide (1 mM in PBS, 5 min), and fluo- media and allowed to proliferate in TH medium, after which cell culture rescence intensity was measured using a SpectraMax Microplate Reader supernatants were tested for Ig production and specific Ag binding. Sub- (Molecular Devices; excitation, 530 nm; emission, 590 nm). Data are sequently, hybridomas were cloned by limiting dilution at ,1 cell/well and reported as the ratio of luminescence/ethidium bromide fluorescence then subcloned twice to produce B cell hybridoma clones. Clones were intensity. maintained in DMEM with 10% FCS under standard cell culture conditions. Statistical analysis Screening of hybridomas For ELISA and cAMP assays, the effect of multiple treatment groups was Hybridoma supernatants were screened by capture ELISA for isotype evaluated by one-way ANOVA using GraphPad Prism 5 software. Dif- (IgG1, IgK) and allotype (IgG1a, IgG1b) of Ab and by Ag panel ELISA for ference of the means was evaluated by the Tukey test. For human sera antigenic specificity. V-gene cDNA sequence analysis was performed on ELISA data (Fig. 7), analysis was performed using Mann–Whitney and each hybridoma clone. Total RNA was extracted from 5 3 106 hybrid- Kruskal–Wallis tests. oma cells using TRI Reagent (Molecular Research Center). RT-PCR was performed using mAb 24.3.1 V-gene specific primers, Mouse Ig-Primers (Novagen), and primers specific to g and k constant regions. The resulting Results RT-PCR products were purified and sequenced. Tg mice cAMP assays To further investigate in vivo targets of human mAb 24.3.1

Assays were done using A9 L-hD2 S.C.18 (ATCC CRL-10225), a fibroblast from SC, Tg mice expressing HC and LC V-region genes of anti- Downloaded from cell line expressing human D2R. The untransfected A9 cell line (ATCC streptococcal/anti-brain human mAb 24.3.1 were generated by CCL-1.4) was used as a control. Cells were grown and isolated for the assay pronuclear microinjection of HC and LC Tg vector constructs into following the published procedure of Tang et al. (47). Briefly, cells were plated in 12-well culture plates and grown to confluence (∼1 3 106 cells/ C57BL/6 mouse embryos. The human VL and VH genes were cloned well) in IMDM supplemented with 10% FBS (Life Technologies) at 37˚C. into Tg vectors containing upstream regulatory (promoter, leader, Prior to assays, cells were washed three times with prewarmed DMEM and enhancer) and constant regions of murine Igk (Fig. 1A) and supplemented with 0.1% ascorbic acid and 20 mM Tris (pH 7.4) and

murine IgG1 allotype a (Fig. 1B). The HC construct also contained http://www.jimmunol.org/ preincubated in 400 ml the same medium for 10 min at 37˚C. Incubations were started by adding 100 ml the above-mentioned prewarmed DMEM the murine IgG1 membrane exons. Tg constructs were coinjected containing test drugs (i.e., dopamine, 10 mg/ml [Sigma]), human serum into C57BL/6 mouse embryos by Xenogen Biosciences-Caliper (1:50), or hybridoma cell supernatant. Incubations were done for 20 min at Life Sciences (Cranbury, NJ). Litters were screened with mAb 37˚C. Overnight treatment with pertussis toxin (100 ng/ml; Sigma) was 24.3.1 V-gene–specific primers for incorporation of the gene(s). done as a control. cAMP assay was performed using a cAMP Direct Im- The chimeric V transgene IgG1a allotype allowed detection of munoassay Kit (BioVision), following the manufacturer’s protocol, and H a standard curve from 3.12 to 200 pmol was run on each plate. Concen- expression in the C57BL/6 background (b allotype) by anti-allotypic tration of cAMP in samples was calculated from the standard curve using Ab. Tg mice produced were primarily single Tg (containing VH GraphPad Prism 5 or KC4 software. only).

Expression of transgene(s) in mice resulted in production of a by guest on September 26, 2021 Ab-binding experiments using FLAG epitope-tagged D2R a transfectants chimeric (human V-gene/mouse IgG1 C-region) Ab. The Tg24.3.1- IgG1a Ab was expressed in serum at 1–4 mg/ml, with the highest HEK293T cells were cultured in 96-well plates in DMEM with 10% FBS levels observed when the double Tg (V +V ) was expressed (Fig. (Sigma), 50 U/ml penicillin, and 50 mg/ml streptomycin. Transient ex- H L pression of D2R was achieved by transfecting cells with a plasmid vector 2). Expression levels were monitored at monthly intervals. Fig. 2 (pcDNA/Zeo; Invitrogen) containing a FLAG-tagged D2L receptor cDNA shows levels in representative 6-mo-old Tg mice. construct using Lipofectamine LTX (Invitrogen) transfection reagent. FACS analysis determined B cell surface expression of Tg24.3.1- Control cells were transfected with the empty vector. Forty-eight hours IgG1a chimeric Ab. Splenic B cells from Tg mice expressed similar posttransfection, cells were fixed using 4% w/v formaldehyde (15 min at 4˚ levels of IgG1a and endogenous IgG1b (data not shown), indi- C) in PBS. Cells were blocked for 45 min with 3% BSA in PBS before incubation with the indicated Abs (1 h, 20˚C) in the blocking buffer. The cating no allelic exclusion. In bone marrow, some Tg mice had a M2 anti–FLAG-tagged mouse mAb (Sigma) was used at a dilution of B cells that expressed more than twice as many chimeric IgG1 1:250 as a positive control. The affinity-purified human mAbs were used at receptors as IgG1b receptors. Previous work suggests that Tg mice a concentration ∼3 mg/ml. Human sera from PANDAS and SC patients and expressing Ig genes have reductions in relative B cell numbers in normal healthy controls were used at 1:100 dilution. Cells were washed three times with PBS at 4˚C and then incubated (1 h, 20˚C) with the ap- spleen, bone marrow, and peripheral blood and accelerated mat- propriate HRP-conjugated anti-mouse or anti-human secondary Abs (Jack- uration and more rapid emigration of B cells due to earlier, more son ImmunoResearch; 1:5000 dilution in blocking buffer). HRP-catalyzed synchronous expression of the Tg HC (48).

FIGURE 1. Schematic representation of DNA constructs used to generate 24.3.1 Tg mice. (A) LC construct. The Vk construct contains the hu- man mAb 24.3.1 SalI-flanked VJ region in addi- tion to Ig promoter (Pk) and leader (Lk) se- quences and the mouse k C region (Ck). (B)HC construct. Human mAb 24.3.1 VDJ region flanked by SalI restriction sites was cloned into a mouse IgG1a Tg construct that was described previously (37). Restriction sites: B, BamHI; C, ClaI; N, NotI. S, Switch region; M, membrane region. The Journal of Immunology 5529

FIGURE 2. Expression of chimeric IgG1a in sera of Tg mice. Capture a ELISA using anti-mouse IgG1 confirmed the expression of 24.3.1 VH chi- meric Ab in sera of 6-mo-old 24.3.1 Tg mice.

Chimeric IgG1a Ab expressed in Tg mice demonstrated cross-reactive anti-streptococcal/anti-brain Ab specificities in

Tg sera and Tg mAbs similar to human mAb 24.3.1 Downloaded from To determine whether binding specificities of chimeric Tg24.3.1- IgG1a Ab matched those recognized by human mAb 24.3.1, sera from Tg mice were reacted with a panel of streptococcal and brain a Ags. Single VH 24.3.1 and double VH+VL 24.3.1 Tg serum IgG1 Ab retained specificity to brain Ags lysoganglioside and tubulin,

as well as streptococcal group A carbohydrate epitope GlcNAc http://www.jimmunol.org/ and streptococcal wall membrane Ag (Fig. 3A, 3B). These sera did not react with ssDNA, dsDNA, or BSA, indicating the spec- ificity of the Tg Abs. Fig. 3A shows double Tg serum reactiv- ity, and single VH Tg sera are shown in Fig. 3B. Furthermore, mAbs were produced from splenic B cells of VH 24.3.1 Tg mice. Thirteen Tg B cell hybridoma clones producing strong anti- a lysoganglioside binding by the Tg IgG1 Ab were selected. VH transgene expression in the hybridomas was verified by RT-PCR and subsequent cDNA sequencing through the V(D)J junction. by guest on September 26, 2021 a All thirteen Tg-derived VH 24.3.1 mAbs (IgG1 ) were similar and retained specificities to tubulin, lysoganglioside, and GlcNAc; three representative clones are shown in Fig. 3C. The specificities found in Tg mouse sera or Tg hybridomas expressing the 24.3.1 VH transgene were very similar to the human chorea-derived mAb 24.3.1. Twelve mAbs, including three mAbs shown, did not react with ssDNA or dsDNA. One 24.3.1 Tg-derived mAb (66-4.8) was unique in that it strongly recognized dsDNA (OD = 0.6). None of the Tg-derived mAbs reacted with BSA. These data further con- firm the validity of the VH 24.3.1 Tg Ab in the Tg mice. FIGURE 3. Chimeric IgG1a Tg Ab expressed in 24.3.1 Tg mouse sera Nucleotide sequence analysis of Tg B cell hybridoma–derived and Tg-derived mAbs react with streptococcal and neuronal Ags. Double

VL genes showed similarities in LC VL gene usage, with some (VH+VL) Tg mouse sera (A) and single (VH) Tg sera (B) tested against clones preferentially using Vk 21E and Vk 1 (bb1) LC genes pre- a panel of Ags by ELISA using anti-mouse IgG1a retained cross-reactive viously described in mouse anti–peptide-DWEYSVWLSN/dsDNA Ab specificities to tubulin, lysoganglioside, and streptococcal Ags. (C)Ag a cross-reactive autoantibodies associated with lupus and neurologic panel reactivity of 24.3.1 Tg mAbs (IgG1 ) in ELISA. mAbs produced symptoms (49, 50). Five Abs used the Vk 21E subgroup (21.12) from Tg mouse splenic B cells retained cross-reactive specificities to tu- LC gene, and two used the Vk 1 (bb1) LC gene (Table I). Tg- bulin, lysoganglioside, and streptococcal Ags. The p value was obtained by the Tukey test using one-way ANOVA. ns, Not significant. derived mAbs encoded by the 21.12 LC gene (66-1.23, 66-2.5, 66- 2.11, 66-4.8, and 66-4.12) had .95% homology with mouse mAbs (16-9, 19-43, and 37g) that reacted with the DWEYS peptide that Tg24.3.1-IgG1a could be detected in serum, expressed by [associated with neuropsychiatric lupus (50)] and with multiple B cells, and deposited in brain. Therefore, the Tg24.3.1-IgG1a Ab autoantigens, including dsDNA, cardiolipin, and fibronectin. expressed in B cells and released from plasma cells in serum VL LCs pairing with our Tg VH24.3.1 have often been repeated penetrated the brain and was taken up by neuronal cells in vivo. in other reported autoantibody specificities, especially systemic lu- pus erythematosus (49). In vivo chimeric Tg Ab localized in dopaminergic neurons in To summarize, our chimeric Tg24.3.1-IgG1a Ab was expressed basal ganglia of Tg VH24.3.1 mice in B cells and sera of Tg mice. We expected that our Ab would To determine where chimeric Tg 24.3.1-IgG1a Ab (IgG1a) was originate from the sera and penetrate the brain. Therefore, we localized in the brain in vivo, we investigated tissue sections of a a conducted experiments to determine whether the Tg24.3.1-IgG1 brains from Tg VH24.3.1 mice. Anti-IgG1 allotypic Ab detected Ab could target brain tissues. The studies described below show Tg24.3.1-IgG1a within neuronal cell bodies in the basal ganglia, 5530 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS

Table I. VL gene usage in Tg mAbs derived from 24.3.1 VH Tg mouse (Fig. 4J, 4K). In Fig. 4L, DAPI-stained brain denotes the basal ganglia region where colocalized fluorescence (Fig. 4A–F) was Hybridoma Vk Jk observed. IgG1a-positive neuronal cells were also observed in the 66-1.9 1 (bb1) 2 (86%)a cortex of Tg mouse brain and did not stain with anti-TH Ab (Fig. 66-1.23 21.12b 2 (98%) 4N, 4O). Therefore, our data suggest that our Tg autoantibody 66-1.24 1 (bb1) 2 (92%) localized in vivo in both dopaminergic and nondopaminergic b 66-2.5 21.12 2 (97%) neurons in the basal ganglia within the substantia nigra/ventral 66-2.11 21.12b 2 (94%) 66-4.8 21.12b 2 (95%) tegmental area and cortex, respectively. The presence of D2R on 66-4.12 21.12b 4 (91%) other types of neurons could explain our results (53–55). DAPI- stained brain (Fig. 4M) indicates the region where IgG1a–FITC Hybridoma cell lines are listed with the most closely homologous reported germ- a line variable and joining segments. staining in the cortex was observed. No IgG1 staining was seen in aPercentage homology to germline sequence is noted in parentheses. Sequencing the hippocampus (Fig. 4Q, 4R). Fig. 4P shows DAPI staining of was performed on at least three replications and with different primer combinations. bVk21E subgroup. hippocampus. Our Tg24.3.1-IgG1a Ab colocalized in dopaminergic neurons in the substantia nigra or ventral tegmental area, which are part of the most likely the substantia nigra or ventral tegmental area (Fig. basal ganglia. Dopaminergic neurons originating in the substantia 4A). Ab to TH was used to identify dopaminergic neurons of nigra/ventral tegmental area and terminating in the dorsal striatum Tg mouse brain (Fig. 4B) because TH converts tyrosine to dopa- play a critical role in basal ganglia functions (56–59). The sub- mine and is a marker for dopaminergic neurons (51, 52). Chimeric stantia nigra and the adjacent ventral tegmental area, both intrinsic Downloaded from Tg24.3.1-IgG1a Ab colocalized in TH-rich dopaminergic neurons nuclei of the basal ganglia, contain dopaminergic neurons that in vivo (Fig. 4A–C). Fig. 4A–F show Tg 24.3.1-IgG1a Ab within project into the striatum. Nigrostriatal axon terminals release do- dopaminergic neurons in brains of Tg VH24.3.1 mice; enlarged pamine into the striatum. D2Rs are expressed on the dopaminergic images are also shown (Fig. 4D–F). Similar results were observed neurons in the substantia nigra and ventral tegmental area, as well a in five of seven Tg VH24.3.1 mice, whereas no IgG1 was detected in as on neurons in the striatum (55, 58, 60, 61). However, dopa- brains of non-Tg mice (Fig. 4G). Fig. 4H shows TH Ab staining in minergic neuronal cell bodies are only seen in the substantia nigra http://www.jimmunol.org/ non-Tg brain, and Fig. 4I shows the merged image for Fig. 4G and and ventral tegmental area, and the axons of these cells project Fig. 4H. Secondary Ab controls were negative in all experiments into the striatum. All dopamine receptor subtypes are expressed in

a FIGURE 4. Chimeric Tg24.3.1 VH IgG1 Ab–tar- by guest on September 26, 2021 geted dopaminergic neurons in Tg mouse brain in vivo. Colocalization of Tg 24.3.1 IgG1a (anti-IgG1a Ab, green) and TH (anti-TH Ab, red) by immunofluorescent staining of mouse brain tissue. TH is a marker for dopaminergic neurons. (A–I) For each row, left panel shows IgG1a (FITC labeled), center panel shows TH Ab (TRITC labeled), and right panel is merged image (FITC-TRITC). Brain sections (basal ganglia) of Tg mouse (original magnification 320), showing FITC- labeled anti-mouse IgG1a (A), TRITC-labeled TH Ab (B), and merged image (C). (D–F) Enlarged images of (A–C). Brain sections (basal ganglia) of non-Tg litter- mate tested with anti-mouse IgG1a (G), anti-TH Ab (H), and both (merged image; I). Conjugated secondary Ab controls (secondary Ab, no primary Ab): FITC- conjugated streptavidin (1:20; J) and TRITC-conju- gated sheep anti-rabbit Ab (1:100; K). (L)Tgmouse brain stained with DAPI. Rectangle denotes basal gan- glia region of colocalization of images shown in (A–I) (original magnification 34). Cortex tissue section of Tg mouse brain: DAPI staining (original magnifi- cation 310) of cortex tissue section (M). (N) 24.3.1 IgG1a-positive (FITC-labeled) cells shown in rectangle in (M). (O) Enlarged image of cortex tissue section. Hippocampus tissue section of Tg mouse brain: DAPI staining (original magnification 310) of hippocampus tissue section (P). (Q) No staining for 24.3.1 IgG1a observed in the hippocampus. (R) Enlarged image of hippocampus. The Journal of Immunology 5531 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 5. Tg24.3.1-IgG1a and human SC mAb 24.3.1 reacted and colocalized with D2R. (A) Reactivity of 24.3.1 Tg mAbs and Tg mouse serum (IgG1a) with human D2R in the ELISA using anti-mouse IgG1a. Tg mAbs reacted similarly to human mAb 24.3.1 with D2R. Other human chorea-derived mAbs 37.2.1 and 31.1.1 reacted less with D2R than mAb 24.3.1. The p value was obtained by the Tukey test using one-way ANOVA. (B) Chimeric a a a Tg24.3.1 VH IgG1 Ab colocalized with anti-D2R in Tg mouse brain in vivo. Colocalization of Tg 24.3.1 IgG1 (anti-IgG1 Ab) and rabbit polyclonal anti- D2R Ab by immunofluorescent staining of mouse brain tissue. Top row, Brain sections of Tg mouse, showing FITC-labeled anti-mouse IgG1a (left panel), TRITC-labeled anti-D2R Ab (center panel), and merged image (right panel) (original magnification 320). Middle row, Enlarged images from top row. Bottom row, Secondary Ab controls: FITC-conjugated streptavidin (1:20; left panel) and TRITC-conjugated sheep anti-rabbit Ab (1:100; right panel) were negative. (C) Human SC-derived mAb 24.3.1 (IgM) colocalized with anti-D2R in normal mouse brain tissue sections. (Figure legend continues) 5532 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS the striatum, but D1Rs and D2Rs are most abundant (58). D2Rs Transient expression of FLAG-tagged D2R is evidenced by the are also expressed by striatal interneurons (54). D2Rs have pre- enhanced binding of the anti-FLAG Ab specifically to the cells synaptic and postsynaptic localizations and functions (54, 55, 61). transiently transfected with cDNA for the D2R construct (Fig. The basal ganglia location of the Tg24.3.1-IgG1a–targeted dopa- 6A). Enhanced binding of human chorea-derived mAb 24.3.1 to minergic TH+ neurons was determined based on the presence cells transiently expressing the D2R construct is observed relative of TH as a marker for dopaminergic neurons in the substantia to untransfected cells that do not express D2R. Chorea-derived nigra/ventral tegmental area of the basal ganglia. Dopaminergic mAbs 31.1.1 and 37.2.1 exhibited significantly enhanced bind- (TH+) neuronal cell bodies are absent from the striatum. ing to the FLAG-tagged D2R–expressing cells but were less re- Additional colocalization studies were performed on normal active compared with mAb 24.3.1 (Fig. 6B). Thus, all three SC mouse brain tissue to determine whether SC-derived human mAb mAbs bind significantly to D2R (Fig. 6B). 24.3.1 (IgM) reacted with dopaminergic neurons. As shown in Sup- To determine whether IgG in SC and PANDAS sera reacted plemental Fig. 1, mAb 24.3.1 binds to TH-rich dopaminergic neurons with extracellular D2R surface exposed with a FLAG tag, human as expected, because dopaminergic neurons were targeted in the Tg sera from SC and PANDAS patients and normal control sera were mice (Fig. 4). The right panels in Supplemental Fig. 1 are merged tested in the FLAG assay. We found that all SC sera tested reacted images. Human mAbs 37.2.1 and 31.1.1 were observed to bind consistently with extracellular D2R in the FLAG epitope-tagged dopaminergic neurons (Supplemental Fig. 1). Media controls and transfected cell line, whereas all PANDAS sera were negative secondary Ab controls were negative in all experiments. Thus, (Fig. 6C). Our data correlate well with other studies that found SC both the chorea-derived mAb 24.3.1 and the Tg expressed 24.3.1 IgG positive and PANDAS IgG negative by FACS analysis of

Ab bind to dopaminergic neurons (Fig. 4, Supplemental Fig. 1). HEK293 cell lines expressing D2R (64). Table II compares results Downloaded from

a from the FLAG-tagged D2R assay with other D2R assays, in- Tg serum Ab and IgG1 Tg-derived mAbs reacted with human cluding ELISA and signaling through D2R in the cAMP assay D2R (cAMP results described below indicate that both SC and To investigate the possibility that the Tg-derived Abs, as well as the PANDAS sera signal through D2R, whereas only the SC IgG binds chorea-derived human mAb 24.3.1, might bind to human D2R, to the FLAG-tagged D2R).

a receptor present on dopaminergic neurons (54, 55), human mAb To investigate the reactivity of human SC mAb 24.3.1 with D2R, http://www.jimmunol.org/ 24.3.1, Tg sera, and Tg-derived mAbs were tested for reactivity mAb 24.3.1 was reacted in ELISA with several peptides that we with human D2R. Single- and double-Tg mouse serum and Tg synthesized from the D2R extracellular region; the most reactive hybridoma mAbs (IgG1a) exhibited strong reactivity (OD . 2.0) peptide was chosen for a competitive-inhibition ELISA. Chorea with human D2R (Fig. 5A). All 13 Tg-derived mAbs exhibited mAbs 24.3.1, 31.1.1, and 37.2.1 all reacted most strongly with D2R reactivity with D2R; three of those tested are shown in Fig. 5A. peptide E1.1 (OD 1.0) and were tested further to determine whether Strong reactivity with D2R was seen for chorea-derived human binding to whole D2R could be inhibited in vitro by blocking with mAb 24.3.1, from which the VH and VL genes were derived and D2R E1.1 in competitive-inhibition ELISA. D2R peptide E1.1 was expressed in the Tg mice (Fig. 5A). Tg sera, as well as Tg-derived a strong inhibitor of SC mAb 24.3.1, 31.1.1, and 37.2.1 binding to mAbs, reacted with D2R similar to mAb 24.3.1. Reactivity of D2R and inhibited D2R recognition by chorea mAbs in a dose- by guest on September 26, 2021 chorea-derived mAbs 37.2.1 and 31.1.1 with D2R was consider- dependent manner (Fig. 6D, right panel, Fig. 6E). A commer- ably lower than seen with mAb 24.3.1. cial anti-human D2R mAb, specific for the region included in the To determine whether our chimeric Tg24.3.1-IgG1a Ab colo- D2R E1.1 peptide (mAb 1B11; Sigma), was compared with mAb calized with D2R, we treated the Tg brain sections with rabbit 24.3.1 in the competitive-inhibition ELISA (Fig. 6D, left panel). A polyclonal anti-D2R Ab and anti-IgG1a Ab to detect the Tg24.3.1 previously characterized peptide from human cardiac myosin (45, Ab. As shown in Fig. 5B, chimeric Tg24.3.1–IgG1a Ab colocalized 46) did not strongly inhibit binding of either anti-D2R mAb 1B11 with anti-D2R (merged images shown in right panels)inTgmouse or chorea mAb 24.3.1 to whole D2R in ELISA (Fig. 6D, control brain. Secondary Ab controls were negative. peptide). In Fig. 6D, the y-axis scale is different for the two Abs To compare binding of Tg24.3.1-IgG1a Ab after expression in shown; the condition “no peptide” shows an OD of 2 with the Tg mice with direct binding of human mAb 24.3.1 to brain tissues, commercial D2R E1.1 Ab and an OD ∼1 with the SC mAb 24.3.1. normal mouse brain tissue sections were reacted with SC-derived Fig. 6D shows that the two Abs differ significantly in the inhibi- human mAb 24.3.1 and rabbit polyclonal Ab to D2R. As shown in tion curves. A higher concentration of inhibitor peptide D2R E1.1 Fig. 5C, human mAb 24.3.1 colocalized with anti-D2R similar to (∼62 mg/ml) was required to reach ∼50% inhibition of cross- Tg 24.3.1-IgG1a Ab (merged images, right panels). Media con- reactive SC mAb 24.3.1, whereas ,1 mg/ml peptide produced trols and secondary Ab controls were negative. ∼50% blocking of the commercial positive-control D2R E1.1 peptide Ab. Neither Ab was significantly inhibited with 500 mg/ml Chorea-derived mAb 24.3.1 strongly reacted with FLAG of a “control peptide”: 25% inhibition was seen for the SC mAb epitope-tagged D2R transfectants on the cell surface and D2R 24.3.1, and 12% inhibition was observed for the commercial extracellular peptide sequence positive-control anti-D2R E1.1 peptide Ab. Inhibition of both To further confirm dopamine receptor binding of our human mAbs was significantly greater with the D2R E1.1 peptide (Fig. chorea-derived mAb 24.3.1, we examined the binding of the Ab 6D). In addition, extracellular D2R peptide E1.2 did not exhibit to HEK293T cells transiently expressing a D2R construct that con- strong inhibition, and it behaved similarly when reacted with the tained a FLAG epitope tag at the extracellular N terminus (62, 63). commercial anti-D2R Ab and SC mAb 24.3.1 (Fig. 6D). Two

Colocalization of human SC mAb 24.3.1 (anti-human IgM Ab) and rabbit polyclonal anti-D2R Ab in normal mouse brain tissue sections. Top row, Brain tissue sections of normal mouse (original magnification 310), incubated in vitro with human SC mAb 24.3.1, showing FITC-labeled anti-human IgM (left panel), TRITC-labeled rabbit polyclonal D2R Ab (center panel), and merged image (right panel). Second row, Enlarged images from top row. Media controls (IMDM supplemented with 10% FBS) and secondary Ab controls (FITC-conjugated anti-human IgM, 1:50; TRITC-conjugated sheep anti-rabbit Ab, 1:100) were negative. The concentration of mAbs tested (24.3.1, 37.2.1, 31.1.1) was 100 ng/ml. ns, Not significant. The Journal of Immunology 5533 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 6. Human SC mAb 24.3.1 reacted with FLAG epitope-tagged D2R at the cell surface and with a D2R peptide. (A) Quantification of the relative levels of cell surface binding of anti-FLAG M2 Ab to HEK293T cells transfected with cDNA for the FLAG-tagged D2R construct or HEK293T cells transfected with an empty vector (untransfected cells). White bars represent HEK293T cells treated with control media (no Ab), and black bars represent HEK293T cells treated with anti-FLAG M2 Ab (mean 6 SD, n = 5). (B) Quantification of the relative levels of cell surface binding of human chorea- derived SC mAbs to HEK293T cells transfected with cDNA for the FLAG-tagged D2R construct (black bars) or with empty vector (untransfected cells, white bars) (mean 6 SD, n = 5). All of the chorea-derived SC mAbs showed significant binding to D2R transfectants compared with cells transfected with empty vector. (C) Human SC sera reacted with FLAG epitope-tagged D2R at the cell surface. Quantification of the relative levels of cell surface binding of human PANDAS sera (P), SC sera (SC), and sera from normal healthy controls (N) to HEK293T cells transfected with cDNA for the FLAG-tagged D2R construct (black bars) or with empty vector (untransfected cells, gray bars) (mean 6 SD, n = 4). Four SC sera tested (SC1–4) (Figure legend continues) 5534 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS

Table II. Summary of immunoreactivity of human SC and PANDAS sera

Human Sera ID No. D2R ELISA IgG Titer cAMP D2 FLAG SC SC1 8,000 Positive Positive (p , 0.001) (p , 0.001) SC2 16,000 Positive Positive (p , 0.001) (p , 0.05) SC3 16,000 Positive Positive (p , 0.001) (p , 0.05) SC4 32,000 Positive Positive (p , 0.0001) (p , 0.001) SC5 16,000 Positive Not tested (p , 0.001) SC6 (mAb donor) 32,000 Positive Not tested (p , 0.0001) SC7 16,000 Positive Not tested (p , 0.0001) SC8 16,000 Positive Not tested (p , 0.001) PANDAS P1 5,120 Positive Negative Downloaded from (p , 0.01) P2 8,000 Positive Negative (p , 0.001) P3 8,000 positive Negative (p , 0.001) P4 128,000 positive Negative (p , 0.0001) http://www.jimmunol.org/ P5 8,000 Positive Not tested (p , 0.001) P6 16,000 Positive Not tested (p , 0.001) P7 32,000 Positive Not tested (p , 0.001) Normals N1 16,000 Negative Negative N2 2,000 Negative Negative N3 8,000 Negative Negative

N4 2,000 Negative Negative by guest on September 26, 2021 N5 16,000 Negative Not tested N6 2,000 Negative Not tested N7 4,000 Negative Not tested other SC mAbs, 31.1.1 and 37.2.1, produced from the same in- IgG in SC and PANDAS with fine choreiform movements dividual, also showed similar peptide reactivity and inhibition reacted with D2R patterns (Fig. 6E). D2R peptide E1.2 did not show significant To determine whether autoantibodies against D2R were elevated in inhibition compared with E1.1. There was a significant difference sera from humans with SC and related disorders, we used ELISA to between the inhibition observed for peptide E1.1 versus E1.2 investigate a small group of well-characterized human sera from (mAb 24.3.1 and mAb 31.1.1, p , 0.001; mAb 37.2.1, p , 0.01) SC, PANDAS, TS, and OCD. Our PANDAS (22) group represents (Fig. 6D, 6E). Tg mAbs also reacted with D2R peptides in the children who had acute-onset OCD and/or tics, as well as fine direct ELISA and showed stronger specificity to D2R E1.1 peptide choreiform piano-playing movements (fingers and toes), following compared with D2R E1.2 peptide (Fig. 6F). Although all three SC a streptococcal infection, which may indicate a type of PANDAS. mAbs reacted similarly with the D2R peptide, they differed with Human sera were titrated in the ELISA against the D2R mem- regard to their signaling and binding properties of D2R and did not brane Ag. As shown in Fig. 7A, IgG reactivity with D2R was react with D2R with the same intensity in the ELISA. significantly elevated in SC and in a PANDAS group with fine

showed significant (p , 0.001 or p , 0.05) binding to D2R transfectants compared with cells transfected with empty vector. Four PANDAS sera tested (P1– 4) were negative and did not show significant binding. Normal control sera tested were negative. *, **, ***, #, and ## in (A)–(C) denote p values shown below the graphs. (D) D2R peptide E1.1 significantly inhibited binding of human SC mAb 24.3.1 to D2R in a dose-dependent manner. Left panel, Dose- dependent inhibition of anti-D2R mAb 1B11 binding to D2R by extracellular D2R peptide E1.1 (p , 0.001). Right panel, Significant (p , 0.001) inhibition of chorea-derived mAb 24.3.1 binding to D2R by D2R peptide E1.1. Control peptide and extracellular D2R peptide E1.2 did not strongly inhibit binding and behaved similarly when reacted with the commercial mouse anti-D2R mAb and our chorea-derived human mAb 24.3.1. p , 0.001, D2R E1.1 versus E1.2. Media control (negative) not shown. (E) Dose-response inhibition of two other human SC mAbs 31.1.1 and 37.2.1 by D2R peptide E1.1. Dose- dependent binding inhibition of human SC mAbs 31.1.1 and 37.2.1 to D2R by extracellular D2R peptide E1.1 (p , 0.001). D2R peptide E1.2 did not show significant inhibition compared with E1.1 (mAb 31.1.1, p , 0.001; mAb 37.2.1, p , 0.01). D2R peptide E1.2 did not strongly inhibit. Extracellular D2R peptide E1.2 did not strongly inhibit. Media control (negative) not shown. mAb concentration used was 100 ng/ml. (F)SCTgVH24.3.1 mAbs reacted with D2R peptides in the ELISA. Tg mAbs expressing human mAb 24.3.1 HC (VH) were tested in the direct ELISA for reactivity with D2R peptides E1.1 and E1.2. Tg mAbs showed specificity for peptide D2R E1.1. Media control (negative) not shown. p , 0.001, D2R E1.1 versus D2R E1.2. The Journal of Immunology 5535 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 7. IgG in SC and PANDAS reacted with D2R. (A) IgG from SC and PANDAS sera reacted with human D2R. Human sera from SC, PANDAS, TS/OCD, and ADHD were titrated for reactivity with human D2R. A strong reaction of SC (p = 0.0005) and PANDAS (p = 0.0291) sera was observed with D2R in the ELISA. (B and C) IgG from SC and PANDAS reacted with D2R compared with the different response against D1R. Reactivity of human sera from SC, PANDAS, TS/OCD, and ADHD with dopamine receptors D1R and D2R in the ELISA. Ab titers are shown in (B) and OD values are shown in (C) for comparison of titers versus OD at 405 nm (1:500 dilution). Horizontal line indicates mean. The p values were obtained using the Mann–Whitney and Kruskal–Wallis tests. choreiform movements but not in TS, OCD, ADHD, or normal 7C used the same data as shown in Fig. 7B, but it shows OD age-matched subjects. The dot plot shows normal controls with an values at 1:500 serum dilution rather than titers for comparison. anti-D2R mean titer of 6,000, whereas the mean titer in SC was 17,600 (∼3-fold increase). In PANDAS, the mean titer was 13,449 Chorea-derived human mAb 24.3.1 signaled human D2R and (.2-fold the normal mean). Our ELISA data suggest that auto- inhibited adenylate cyclase activity in D2R transfectants antibodies against D2R were associated more significantly with To study Ab activation of the D2 inhibitory receptor, D2R-expressing movement disorders, such as SC or PANDAS with fine choreiform transfectants were reacted with human mAbs, and changes in movements, and not with TS, OCD, or ADHD. Fig. 7B shows IgG adenylate cyclase were measured as cAMP levels in cell lysates. reactivity of SC and PANDAS IgG with D2R compared with Chorea-derived human mAb 24.3.1, which reacted with the human dopamine D1 receptor (D1R). When anti-D2R was compared with D2R membrane Ag in the ELISA, signaled D2R and inhibited anti-D1R in SC and PANDAS, the p values were significantly cAMP in the D2R-expressing transfectants, comparable to acti- different (p = 0.001 and p , 0.0001, respectively; Fig. 7B). Fig. vation of D2R by dopamine. Fig. 8A shows D2R transfectants after 5536 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS

FIGURE 8. Human mAb 24.3.1 signaled human D2R in transfected D2R cells by inhibiting adenylate cyclase activity comparable to dopamine. Inhibition of adenylate cyclase by receptor activation in human D2R-transfected cell lines (A9-hD2R) was quantified by cAMP levels measured in cell lysates. Incubation of Downloaded from the D2R cell line with chorea-derived mAbs 24.3.1, 31.1.1, and 37.2.1. Results are shown from cAMP di- rect-competition immunoassay. (A) D2R-transfected cells treated with mAb 24.3.1 showed a 53% decrease in cAMP, which was comparable to dopamine-treated cells (p , 0.001). Cells treated with pertussis toxin are http://www.jimmunol.org/ also shown; this treatment was used as an experimental control, because D2R-mediated activation and inhibi- tory signaling with concomitant decreases in cAMP can be blocked by preincubation of cells with pertussis toxin. (B) Unresponsive control (nontransfected) A9 cells treated with dopamine and mAbs. (C) Dose-re- sponse curve showed cAMP levels in D2R-transfected cells following treatment with mAb 24.3.1 at 1, 5, 10, 25, and 50 ng and unresponsive control (nontrans- fected) A9 cells after same treatment. Assays were by guest on September 26, 2021 performed in triplicate. The p value was obtained by the Tukey test using one-way ANOVA. The following concentrations of mAbs were tested: 24.3.1, 90 ng/ml; 31.1.1, 100 ng/ml; and 37.2.1, 100 ng/ml. ns, not sig- nificant.

treatment with dopamine and human chorea-derived mAbs 24.3.1, 37.2.1 did not activate D2R (0% for mAb 37.2.1) because cAMP 31.1.1, and 37.2.1. Transfectants treated with 1 mM dopamine was not significantly reduced. As expected, a receptor-mediated as a positive control showed 53% inhibition of cAMP compared decrease in cAMP levels was not inhibited in cells preincubated with untreated transfectants, which is consistent with results of with pertussis toxin. D2Rs are coupled to G proteins of the previous studies that showed ∼50% inhibition (47). Human chorea- Gi/o subfamily, which are pertussin toxin sensitive; consequently, derived mAb 24.3.1–treated D2R-transfected cells showed a reduc- pertussin toxin treatment blocked D2R-mediated decreases in tion in cAMP that was comparable to that of dopamine stimulation. cAMP (47, 65). Control nontransfected A9 cells showed no sig- Although D2R transfectants incubated with chorea-derived mAb nificant differences in cAMP levels in untreated and treated cells 31.1.1 also showed a small reduction in cAMP (∼18%), it was (Fig. 8B). Thus, human chorea-derived mAb 24.3.1 showed a significantly less than with mAb 24.3.1, and chorea-derived mAb significant activation of D2R compared with the same isotype and The Journal of Immunology 5537 concentrations of other chorea-derived mAbs. Fig. 8C shows a dose-response curve of D2R transfectants treated with 1, 5, 10, 25, or 50 ng of mAb 24.3.1, whereas nontransfected A9 cells were unresponsive after the same treatment. The data show that mAb 24.3.1 inhibitory signaling through D2R is dose dependent. mAbs derived from Tg24.3.1-IgG1a mice were tested for D2R signaling, but no reduction in cAMP was seen (data not shown). Lack of D2R signaling by Tg mAbs may be due to mouse LC pairing with a our human/mouse Tg VH 24.3.1–IgG1 , which may affect the epitope recognition or avidity required for signaling. Because Tg 24.3.1-IgG1a Ab localized in dopaminergic neurons in vivo, and human mAbs and Tg mAbs all recognized D2R, the data strongly suggest that D2R on neurons may be one of the Ab targets in vivo, as was shown in our colocalization studies on Tg mouse brain tissue (Fig. 5B). Inhibitory signaling of adenylate cyclase activity in D2R transfectants by human SC sera

In support of the hypothesis that SC mAb 24.3.1 and Tg 24.3.1- Downloaded from IgG1a Ab came from chorea donor sera with D2R signaling, acute serum from the mAb 24.3.1 human B cell donor was shown to have an elevated serum anti-D2R IgG titer of 32,000 by ELISA, which is ∼5-fold greater than the normal mean (6,000), and the acute serum showed strong signaling inhibition of D2R compa-

rable to dopamine (Fig. 9A). Assay results, shown in Fig. 9A, http://www.jimmunol.org/ illustrate the inhibition as a decrease in cAMP levels in transfected cell lysates of human D2R transfectants by the mAb 24.3.1 donor serum but not in control A9 cells (Fig. 9B). Results are shown after treatment with dopamine, anti-D2R serum, normal serum, and mAb 24.3.1. As shown in Fig. 9A, transfectants treated with mAb donor sera (SC6) showed significantly decreased cAMP levels (40% decrease) relative to untreated cells, and this was comparable to the inhibitory response of mAb-treated transfected cells (43%). Transfectants treated with another SC sera (SC7), by guest on September 26, 2021 with a serum anti-D2R IgG titer of 16,000, also showed a sim- ilar decreased (∼40%) cAMP level. Additionally, cAMP levels were significantly decreased (p , 0.001) in transfectants treated with five other SC sera with anti-D2R titers of 8,000 or 16,000. Dopamine-treated D2R transfectants showed a 48% reduction in cAMP levels, whereas levels in transfected cells treated with normal serum (anti-D2R titers of 2,000, 8,000, or 16,000) were not significantly inhibited (Fig. 9A). Control nontransfected A9 cells (Fig. 9B) showed no significant differences in cAMP levels between treated and untreated cells. Our results suggest that the high level of anti-D2R Abs detected in ELISA (32,000) in the human SC mAb 24.3.1 donor sera and other human SC sera sig- naled the human D2R, with a concomitant reduction in adenylate cyclase (cAMP) activity. Normal sera with an anti-D2R titer of FIGURE 9. D2R transfectants treated with human SC sera with elevated 2,000, 8,000, or 16,000 did not signal D2R. Our results illustrate anti-D2R Ab show decreased cAMP levels. Inhibition of adenylate cyclase how the human chorea donor sera elicited the same inhibitory by receptor stimulation in human D2R-transfected cell lines (A9-hD2R) signaling response as did mAb 24.3.1 in D2R-transfected cells, was quantified by cAMP levels measured in cell lysates. Incubation with whereas normal sera with comparable anti-D2R ELISA titers did human SC mAb 24.3.1 Ab donor serum, other SC sera, and normal healthy not. control sera. Results are shown from cAMP direct-competition immuno- assay. (A) D2R-transfected cells treated with SC Ab donor serum (SC6; Inhibitory signaling of adenylate cyclase activity in D2R anti-D2R titer 32,000) and SC serum 7 (SC7; anti-D2R titer 16,000) transfectants by human PANDAS sera showed a 40% decrease in cAMP levels comparable to dopamine (48%)- Using the D2R-transfected cell line, human PANDAS sera with and mAb 24.3.1–treated cells (43% decrease); p , 0.0001. cAMP levels , elevated anti-D2R Ab ELISA titers (P7: 32,000, P6: 16,000, P2: were also significantly decreased (p 0.001) in transfectants treated with five other SC sera (SC1, SC2, SC3, SC5, SC8) with anti-D2R titers of 8,000) were examined for signaling of D2R and adenylate cyclase, 8,000 and 16,000. cAMP level was not reduced significantly in trans- as measured by decreased cAMP. Fig. 10 shows inhibitory signal fectants treated with normal serum (anti-D2R titers 2,000, 8,000, or in cAMP levels in transfected cell lysates of D2R transfectants 16,000). (B) Control (nontransfected) A9 cells were unresponsive. Assays (Fig. 10A) but not in control (nontransfected) A9 cells (Fig. 10B). were performed in triplicate. The p values were obtained by the Tukey test Importantly, the cAMP response of D2R transfectants to PANDAS using one-way ANOVA. Concentration of mAb used was 24.3.1–90 ng/ml. sera P7 (32,000 D2R titer) was comparable to their response to ns, Not significant. 5538 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS

difference in cAMP levels. Our results suggest that sera from SC, as well as from other streptococcal-associated neuropsychiatric sequelae, such as PANDAS with fine piano playing choreiform movements, signal the human D2R.

Discussion Little is known about the targets of autoantibodies in the brain in movement or behavioral disorders, and the effects of autoantibody targeting dopaminergic neurons or their dopamine receptors is virtually unknown. Studies suggest that anti-streptococcal Abs target the brain in human disease or their animal models and may lead to altered behaviors or movements (10, 13, 15, 16, 35, 64, 66– 68). Two recent studies (35, 64) suggested that Abs in SC rec- ognize the D2R in humans, and studies (35, 66, 67) in rats and mice linked behavioral changes with streptococcal exposure and D2R. Our studies and those of other investigators used different methodologies to detect autoantibodies against D2R (64). Table II summarizes our human sera studies in three immunoassays, as

shown in Figs. 6, 7, 9, and 10. The SC-derived human mAb Downloaded from studies are found in Figs. 5–8 and Supplemental Fig. 1, and the Tg Ab studies are found in Figs. 2–6. Collectively, the studies suggest that D2R is the target of autoantibodies produced in SC and PANDAS. Abs against D2R measured by ELISA recognize both extracellular and intracellular epitopes, whereas functional Ab

signaling or extracellular binding assays would recognize only http://www.jimmunol.org/ extracellular domains of the receptor. Our peptide, which inhibited mAb 24.3.1, was an extracellular epitope at the exposed N ter- minus of D2R. To test the hypothesis that autoantibodies from movement and behavioral disorders target neurons and possibly D2R in the brain, we developed a novel Tg mouse expressing an SC-derived brain autoantibody. The Tg autoantibodies exhibited cross-reactive anti- neuronal specificities, in particular D2R reactivity, and targeted dopaminergic neurons in vivo. We show in our study that a human by guest on September 26, 2021 SC-derived Ab, with defined specificity and with a defined VH- region gene linked to a mouse IgG1a C-region gene, targeted neurons in the Tg mouse brain that was most likely the substantia nigra/ventral tegmental area of the basal ganglia. Only the sub- stantia nigra and ventral tegmental area contain dopaminergic neuronal cell bodies that project long axonal extensions into the striatum. Dopaminergic neuronal cell bodies are not observed in the striatum. The Tg24.3.1-IgG1a construct was found within the cytoplasm of dopaminergic neurons in vivo, sparing the nucleus. Tg serum Ab, as well as mAbs produced from Tg VH24.3.1 FIGURE 10. D2R-transfected cells treated with PANDAS sera showed B cells, retained cross-reactive Ab specificities to tubulin, lyso- decreased cAMP levels compared with transfectants treated with normal ganglioside, and streptococcal Ags, even when the human VL was control sera. (A) D2R-transfected cells were treated with PANDAS sera replaced with a mouse VL chain. During these studies, a new and sera from normal controls. PANDAS sera with elevated human anti- functional reactivity of human mAb 24.3.1 was found that sig- D2R Ab ELISA titers were assayed for inhibitory signaling [anti-D2R naled D2R and that may be the consequence of Tg24.3.1 IgG1a titers: P7: 32,000; P6: 16,000; P2: 8,000] and showed significantly de- creased cAMP levels (46%, 39%, 23%) comparable to decreased cAMP targeting dopaminergic or other types of neurons expressing D2R. a levels in response to the neurotransmitter dopamine (44%) and chorea Tg serum Ab and IgG1 hybridoma mAbs demonstrated the same mAb 24.3.1 (48%) treatments. D2R transfectants treated with normal sera strong reactivity with D2R as did human SC-derived mAb 24.3.1. of various ELISA anti-D2R titers (N2: 2000, N7: 4000, N3: 8000) did not Our data support the hypothesis that, in human disease, Abs show significant reductions in cAMP. (B) Control (nontransfected) A9 cells against D2R in SC or PANDAS with fine choreiform movements were unresponsive. Assays were performed in triplicate. The p value was may target dopaminergic neurons in the basal ganglia and cortex obtained by Tukey test using one-way ANOVA. Concentration of mAb and alter brain function to contribute to movement and behavioral ∼ 24.3.1 used was 90 ng/ml. PANDAS sera and normal controls: n =7 disorders. Fine choreiform movements were reported in 95% of (three of each are shown in the figure). ns, Not significant. the first 50 cases of PANDAS (22). We have suspected for some time that mAb 24.3.1 signaled dopamine (44%), as well as chorea-derived mAb 24.3.1 (48%) a functional receptor on the surface of neuronal cells. Our current (Fig. 10A). cAMP was not inhibited or decreased in the D2R- results from our Tg mouse studies reveal that chimeric Tg24.3.1 transfected cells treated with normal sera of various anti-D2R Ab deposited in neurons in Tg brain tissues and colocalized with ELISA titers (Normal N2: 2000; Normal N7: 4000; Normal N3: D2R. Our new data link the signaling of D2R with autoantibodies 8000). Nontransfected A9 cells (Fig. 10B) showed no significant in SC and other diseases for which treatment with D2R blocker The Journal of Immunology 5539 drugs are effective (68–72). Our studies published in Brimberg protein–mediated signaling (53). D1Rs typically stimulate an in- et al. (35) demonstrated reactivity of SC and PANDAS IgG in crease in cAMP, whereas, in contrast, D2 is inhibitory, with a de- ELISA, whereas studies by Dale et al. (64) demonstrated anti- crease in cAMP, as shown for mAb 24.3.1 and chorea and PANDAS D2R Ab in SC using cell-based flow cytometry assays. Neither sera. The roles of D2Rs are generally more complex than D1Rs of these studies investigated anti-D2R Ab function/signaling or because they are expressed as both pre- and postsynaptic receptors localization in brain tissues. This study extends these findings to (53). If D2R is an in vivo Ag, we would expect TH+ neurons, as demonstrate the anti-D2R Ab targeting of neurons in basal ganglia well as postsynaptic TH2 negative neurons expressing D2R in the and cortex. cortex, to be targeted by Tg24.3.1-IgG1a Ab. Future studies will Previous studies (10, 12, 13) showed that SC and PANDAS IgG, be required to determine whether postsynaptic or other types of as well as chorea-derived mAb 24.3.1, activated CaM kinase II in neurons are involved in the reaction with and uptake of mAb a human neuronal cell line, leading to TH activation and subse- 24.3.1. Although D2R was the focus of our current study, it is quent dopamine release. TH is the rate-limiting enzyme in do- quite possible that other neuronal cell surface extracellular Ags pamine synthesis and was increased in the brain of Lewis rats after and receptors may be involved in the signaling of neuronal cells intrathecal infusion of mAb 24.3.1 (12). Dopamine is the main by Abs in chorea, PANDAS, and other related movement and catecholamine neurotransmitter in the CNS and is synthesized neuropsychiatric conditions. from tyrosine via TH and stored in vesicles in axon terminals. In this study, we did not establish a Tg mouse line or test be- Dopaminergic signaling is a balance between dopamine release havior in mice expressing Tg 24.3.1-IgG1a. The continual loss of and reuptake by the presynaptic terminal. The five dopamine re- viable litters from the Tg 24.3.1 mothers could have been due to ceptor types are divided into two families: D1-like (stimulatory) the presence of Tg24.3.1-IgG1a in serum of mothers where Ab Downloaded from and D2-like (inhibitory), both of which act through G protein– may have affected development of the neonatal mice. Evidence coupled protein receptors (73, 74). Dopamine plays a pivotal role suggests that maternal autoantibodies may affect development of in modulating the functioning of the basal ganglia brain circuitry the brain (81). (56, 59, 73), and abnormalities in the functioning of this neuro- Our previous work demonstrated cross-reactivity of human transmitter have been implicated in disorders like SC (75). mAb 24.3.1 with lysoganglioside GM1 (10) and tubulin (11). a Tg24.3.1-IgG1 mice provided ample evidence to show that Lysoganglioside is associated with lipid rafts in the membrane, http://www.jimmunol.org/ human-derived V-gene 24.3.1 retained cross-reactive specificities whereas tubulin is intracellular and is the most abundant protein in to streptococcal and neuronal Ags, as expressed in vivo as a chi- the brain. Tubulin was reported to be associated with postsynaptic meric human V-gene/mouse C-region IgG1a Ab molecule. Human density (82). Our previous studies also showed that human mAb chorea-derived mAb 24.3.1 is an IgM (10); however, we devel- 24.3.1 activates CaM kinase in a human neuronal SKNSH cell oped a chimeric human/mouse Tg24.3.1-IgG1a construct that is lines (10). CaM kinase activation was reported in association with more likely than IgM to cross into the brain. In humans with heterodimers of dopamine receptors containing both D1 and D2 chorea, it is well-established that their cerebrospinal fluid and sera (83). A recent study of SC reports that the ratios of Abs against contain IgG anti-neuronal signaling autoantibodies against the Ag both D1R and D2R correlated directly with the Universidade specificities reported for our Tg24.3.1-IgG1a construct, as well as Federal de Minas Gerais Sydenham’s Chorea Rating Scale for by guest on September 26, 2021 our human mAb 24.3.1 (10). Thus, it is noteworthy that human symptoms of SC (84). In this study, we compared the reactivity of mAb 24.3.1 and Tg mAbs possessed the same neural Ag specif- SC and PANDAS sera against D1R and D2R. Although anti-D1R icities (Figs. 3, 5) and peptide epitopes (Fig. 6) as did those found was much less reactive than was anti-D2R IgG, it was significantly for IgG in chorea (Fig. 7). In addition, the signaling properties of elevated in the sera from SC and PANDAS and will be investi- sera can be abrogated by removal of IgG (35). gated further in future studies. Our Tg24.3.1-IgG1a construct was expressed in mouse sera and Cross-reactivity of anti-neuronal autoantibodies alone may not penetrated the brain targeting dopaminergic neurons (Fig. 4). Our lead to disease, as suggested by the results with normal sera and focus was on Tg Ab expression and targeting of neuronal cells. To other mAbs, but when the avidity of the Ab causes signaling of open the BBB, we used LPS and streptococcal cell walls in the a receptor it may be a risk factor in the development of movement mice to promote Ab penetration of the brain. Requirements for or neuropsychiatric symptoms. The similarity in Ag specificity, BBB penetration were established previously by Kowal et al. (40) but differences in signaling, of D2R by the human chorea-derived in the pathogenesis of anti-DNA Abs and in neurologic lupus (50, mAbs 24.3.1, 37.2.1, and 31.1.1 illustrates this principle that cross- 76). Tg24.3.1-IgG1a mice not receiving the BBB stimulants did reactivity by itself may not necessarily lead to signaling of a re- not have Tg24.3.1-IgG1a Ab in the brain. ceptor. Thus, our studies of mAb 24.3.1 in comparison with the two Most importantly, in vivo, Tg 24.3.1 Ab (IgG1a) colocalized other chorea-derived human mAbs show their differences in sig- with TH Ab, a marker for dopaminergic neurons (51). Thus, our naling, avidity (10), and perhaps recognition of different epitopes study identified dopaminergic neurons as the in vivo targets of in the CaM kinase assay (10), as well as in the D2R-signaling Tg24.3.1-IgG1a in mouse brain. Colocalization studies gave the assay (Fig. 8). appearance that in vivo Tg24.3.1-IgG1a was internalized within The molecular basis for the cross-reactivity of mAb 24.3.1 with TH+ neuronal cells in cross-sections of the Tg mouse basal gan- D2R and two other neuronal Ags may be due, in part, to Abs against glia. Internalization might be expected, because previous studies the group A streptococcal carbohydrate Ag GlcNAc and its cross- showed that Abs to neuronal receptors can be internalized (77), reactivity with lysoganglioside (10) and tubulin (11). We know and it is well established that dopamine receptors undergo recy- from extensive studies (31) that Abs against a helical or specific cling (78–80). Additionally, our current work confirms colocali- peptide sequences containing aromatic and nonpolar residues are zation of deposited chimeric Tg24.3.1-IgG1a with dopamine cross-reactive with GlcNAc. It is notable that the peptide most receptor D2R. Our study strongly supports the in vitro recognition strongly recognized (Fig. 6) by mAb 24.3.1 has many nonpolar of D2R by both Tg24.3.1-IgG1a Ab and chorea-derived human and aromatic amino acids in sequence. The previous study by mAb 24.3.1. Dopamine is critically involved in numerous physi- Shikhman et al. (31) used alanine substitutions to prove that the ological processes, and dopamine receptor functions are associ- presence of these nonpolar and aromatic residues contributed to ated with the regulation of the second messenger cAMP through G the cross-reactivity between dissimilar molecules. 5540 BRAIN AUTOANTIBODY TARGETS DOPAMINERGIC NEURONS

We do not completely understand the basis of the discrepancy compulsive symptoms: a prospective blinded cohort study. Pediatrics 121: 1188– 1197. in lack of binding of PANDAS sera to FLAG-tagged extracellular 7. Murphy, T. K., R. Kurlan, and J. Leckman. 2010. The immunobiology of D2R expressed in transfected HEK293T cells (extracellular N- Tourette’s disorder, pediatric autoimmune neuropsychiatric disorders associated with Streptococcus, and related disorders: a way forward. J. Child Adolesc. terminal octapetide DYKDDDDK [FLAG]-tagged D2L isoform) Psychopharmacol. 20: 317–331. (63) compared with PANDAS sera signaling through D2R ex- 8. Gause, C., C. Morris, S. Vernekar, C. Pardo-Villamizar, M. A. Grados, and pressed on live mouse fibroblast A9 L-hD2 S.C.18 cells (A9 L cell H. S. Singer. 2009. Antineuronal antibodies in OCD: comparisons in children line hD2 subclone #18, ATCC CRL-10225) (Fig. 10A). The dif- with OCD-only, OCD+chronic tics and OCD+PANDAS. J. Neuroimmunol. 214: 118–124. ferences could be explained if PANDAS and SC Abs bind dif- 9. Murphy, T. K., E. A. Storch, A. B. Lewin, P. J. Edge, and W. K. Goodman. 2012. ferent extracellular epitopes on D2R. It is conceivable that the Clinical factors associated with pediatric autoimmune neuropsychiatric disorders FLAG tag may interfere with the binding of PANDAS Abs in associated with streptococcal infections. J. Pediatr. 160: 314–319 10. Kirvan, C. A., S. E. Swedo, J. S. Heuser, and M. W. Cunningham. 2003. Mimicry HEK293T cell experiments. Alternatively, differences in post- and autoantibody-mediated neuronal cell signaling in Sydenham chorea. Nat. translational modifications of D2R in different cell lines also could Med. 9: 914–920. result in interference in PANDAS Abs binding one cell line but 11. Kirvan, C. A., C. J. Cox, S. E. Swedo, and M. W. Cunningham. 2007. Tubulin is a neuronal target of autoantibodies in Sydenham’s chorea. J. Immunol. 178: not the other. 7412–7421. In summary, based on several pieces of evidence, our findings 12. Kirvan, C. A., S. E. Swedo, D. Kurahara, and M. W. Cunningham. 2006. Streptococcal mimicry and antibody-mediated cell signaling in the pathogenesis support recognition of D2R by mAb 24.3.1. We demonstrated mAb of Sydenham’s chorea. Autoimmunity 39: 21–29. 24.3.1 binding to a FLAG-tagged dopamine D2R, as well as in- 13. Kirvan, C. A., S. E. Swedo, L. A. Snider, and M. W. Cunningham. 2006. duction of cAMP signaling in vitro in D2R-transfected cell lines. Antibody-mediated neuronal cell signaling in behavior and movement disorders. J. Neuroimmunol. 179: 173–179. Although mAb 24.3.1 identified a peptide epitope of D2R, we 14. Brilot, F., V. Merheb, A. Ding, T. Murphy, and R. C. Dale. 2011. Antibody Downloaded from cannot be certain of the in vivo target. Our Tg24.3.1-IgG1a Ab binding to neuronal surface in Sydenham chorea, but not in PANDAS or Tourette mouse model translated to human SC by evidence of in vivo ex- syndrome. Neurology 76: 1508–1513. 15. Bronze, M. S., and J. B. Dale. 1993. Epitopes of streptococcal M proteins that pression of chorea-derived human VH Ab that targeted dopami- evoke antibodies that cross-react with human brain. J. Immunol. 151: 2820–2828. nergic neurons. In addition, human sera IgG from SC and PANDAS 16. Husby, G., I. van de Rijn, J. B. Zabriskie, Z. H. Abdin, and R. C. Williams, Jr. induced inhibitory signaling of the D2R in transfectants. Reac- 1976. Antibodies reacting with cytoplasm of subthalamic and caudate nuclei neurons in chorea and acute rheumatic fever. J. Exp. Med. 144: 1094–1110. tivity with D2R may be present in SC and PANDAS that displays 17. Zabriskie, J. B. 1967. Mimetic relationships between group A streptococci and http://www.jimmunol.org/ small choreiform piano playing movements (22). Finally, our evi- mammalian tissues. Adv. Immunol. 7: 147–188. 18. Zabriskie, J. B., K. C. Hsu, and B. C. Seegal. 1970. Heart-reactive antibody dence supports the hypothesis that in brain disorders, such as associated with rheumatic fever: characterization and diagnostic significance. SC and PANDAS/pediatric acute-onset neuropsychiatric syn- Clin. Exp. Immunol. 7: 147–159. drome, Ab-mediated D2R signaling on dopaminergic neurons could 19. Zabriskie, J. B. 1985. Rheumatic fever: the interplay between host, genetics, and microbe. Lewis A. Conner memorial lecture. Circulation 71: 1077–1086. contribute to alteration of the central dopamine pathways and de- 20. Swedo, S. E. 1994. Sydenham’s chorea. A model for childhood autoimmune velopment of movement and neuropsychiatric symptoms. neuropsychiatric disorders. JAMA 272: 1788–1791. 21. Perlmutter, S. J., S. F. Leitman, M. A. Garvey, S. Hamburger, E. Feldman, H. L. Leonard, and S. E. Swedo. 1999. Therapeutic plasma exchange and in- Acknowledgments travenous immunoglobulin for obsessive-compulsive disorder and tic disorders We thank Dr. Chris Goodnow for the IgG1a HC vector and Dr. Brett Aplin in childhood. Lancet 354: 1153–1158. by guest on September 26, 2021 22. Swedo, S. E., H. L. Leonard, M. Garvey, B. Mittleman, A. J. Allen, for the Igk LC vector. We also thank Janet Baker, Adita Blanco, and Kathy S. Perlmutter, L. Lougee, S. Dow, J. Zamkoff, and B. K. Dubbert. 1998. Pediatric Alvarez for expert technical assistance. We acknowledge Dr. David Dyer autoimmune neuropsychiatric disorders associated with streptococcal infections: and Dr. Allison Gillaspy and the University of Oklahoma Health Sciences clinical description of the first 50 cases. Am. J. Psychiatry 155: 264–271. Center Laboratory for Genomics and Bioinformatics for DNA sequencing, 23. Swedo, S. E., H. L. Leonard, B. B. Mittleman, A. J. Allen, J. L. Rapoport, S. P. Dow, M. E. Kanter, F. Chapman, and J. Zabriskie. 1997. Identification of as well as Jim Henthorn and the University of Oklahoma Flow and Image children with pediatric autoimmune neuropsychiatric disorders associated with Cytometry Laboratory for FACS analysis. We thank the dedicated parents streptococcal infections by a marker associated with rheumatic fever. Am. J. of children with movement and behavioral disorders for support. We also Psychiatry 154: 110–112. thank Dr. Shu Man Fu (University of Virginia College of Medicine, Char- 24. Swedo, S. E., J. L. Rapoport, D. L. Cheslow, H. L. Leonard, E. M. Ayoub, lottesville, VA) for critical review of the original manuscript and Dafna D. M. Hosier, and E. R. Wald. 1989. High prevalence of obsessive-compulsive symptoms in patients with Sydenham’s chorea. Am. J. Psychiatry 146: 246–249. Lotan and Dr. Daphna Joel (Tel Aviv University, Tel Aviv, Israel) for 25. Swedo, S. E., H. L. Leonard, M. B. Schapiro, B. J. Casey, G. B. Mannheim, helpful discussions. M. C. Lenane, and D. C. Rettew. 1993. Sydenham’s chorea: physical and psy- chological symptoms of St Vitus dance. Pediatrics 91: 706–713. 26. Swedo, S. E., and P. J. Grant. 2005. Annotation: PANDAS: a model for human Disclosures autoimmune disease. J. Child Psychol. Psychiatry 46: 227–234. M.W.C. serves as Chief Scientific Officer of Moleculera Labs, which pro- 27. Giedd, J. N., J. L. Rapoport, M. J. Kruesi, C. Parker, M. B. Schapiro, A. J. Allen, H. L. Leonard, D. Kaysen, D. P. Dickstein, W. L. Marsh, et al. 1995. Sydenham’s vides testing of anti-neuronal Abs in children with neuropsychiatric and chorea: magnetic resonance imaging of the basal ganglia. Neurology 45: 2199– movement disorders. M.W.C. and C.J.C. declare a financial interest in Mol- 2202. eculera Labs. J.F.L. serves on the medical advisory board of Moleculera 28. Singer, H. S., C. R. Loiselle, O. Lee, K. Minzer, S. Swedo, and F. H. Grus. 2004. Labs. Anti-basal ganglia antibodies in PANDAS. Mov. Disord. 19: 406–415. 29. Giedd, J. N., J. L. Rapoport, M. A. Garvey, S. Perlmutter, and S. E. Swedo. 2000. MRI assessment of children with obsessive-compulsive disorder or tics associ- ated with streptococcal infection. Am. J. Psychiatry 157: 281–283. References 30. Allen, A. J., H. L. Leonard, and S. E. Swedo. 1995. Case study: a new infection- 1. Stollerman, G. H. 2001. Rheumatic fever in the 21st century. Clin. Infect. Dis. triggered, autoimmune subtype of pediatric OCD and Tourette’s syndrome. J. 33: 806–814. Am. Acad. Child Adolesc. Psychiatry 34: 307–311. 2. Marques-Dias, M. J., M. T. Mercadante, D. Tucker, and P. Lombroso. 1997. 31. Shikhman, A. R., N. S. Greenspan, and M. W. Cunningham. 1994. Cytokeratin Sydenham’s chorea. Psychiatr. Clin. North Am. 20: 809–820. peptide SFGSGFGGGY mimics N-acetyl-beta-D-glucosamine in reaction with 3. Taranta, A., and G. H. Stollerman. 1956. The relationship of Sydenham’s chorea antibodies and lectins, and induces in vivo anti-carbohydrate antibody response. to infection with group A streptococci. Am. J. Med. 20: 170–175. J. Immunol. 153: 5593–5606. 4. Taranta, A. 1959. Relation of isolated recurrences of Sydenham’s chorea to 32. Shikhman, A. R., and M. W. Cunningham. 1994. Immunological mimicry be- preceding streptococcal infections. N. Engl. J. Med. 260: 1204–1210. tween N-acetyl-beta-D-glucosamine and cytokeratin peptides. Evidence for 5. Kurlan, R., and E. L. Kaplan. 2004. The pediatric autoimmune neuropsychiatric a microbially driven anti-keratin antibody response. J. Immunol. 152: 4375– disorders associated with streptococcal infection (PANDAS) etiology for tics and 4387. obsessive-compulsive symptoms: hypothesis or entity? Practical considerations 33. Shikhman, A. R., N. S. Greenspan, and M. W. Cunningham. 1993. A subset of for the clinician. Pediatrics 113: 883–886. mouse monoclonal antibodies cross-reactive with cytoskeletal proteins and 6. Kurlan, R., D. Johnson, and E. L. Kaplan; Tourette Syndrome Study Group. group A streptococcal M proteins recognizes N-acetyl-beta-D-glucosamine. J. 2008. Streptococcal infection and exacerbations of childhood tics and obsessive- Immunol. 151: 3902–3913. The Journal of Immunology 5541

34. Conners, C. K., K. C. Wells, J. D. Parker, G. Sitarenios, J. M. Diamond, and 58. Surmeier, D. J., L. Carrillo-Reid, and J. Bargas. 2011. Dopaminergic modulation J. W. Powell. 1997. A new self-report scale for assessment of adolescent psy- of striatal neurons, circuits, and assemblies. Neuroscience 198: 3–18. chopathology: factor structure, reliability, validity, and diagnostic sensitivity. J. 59. Graybiel, A. M. 2000. The basal ganglia. Curr. Biol. 10: R509–R511. Abnorm. Child Psychol. 25: 487–497. 60. Kawaguchi, Y., C. J. Wilson, S. J. Augood, and P. C. Emson. 1995. Striatal 35. Brimberg, L., I. Benhar, A. Mascaro-Blanco, K. Alvarez, D. Lotan, C. Winter, interneurones: chemical, physiological and morphological characterization. J. Klein, A. E. Moses, F. E. Somnier, J. F. Leckman, et al. 2012. Behavioral, Trends Neurosci. 18: 527–535. pharmacological, and immunological abnormalities after streptococcal exposure: 61. Gerfen, C. R., K. A. Keefe, and E. B. Gauda. 1995. D1 and D2 dopamine re- a novel rat model of Sydenham chorea and related neuropsychiatric disorders. ceptor function in the striatum: coactivation of D1- and D2-dopamine receptors Neuropsychopharmacology 37: 2076–2087. on separate populations of neurons results in potentiated immediate early gene 36. Aplin, B. D., C. L. Keech, A. L. de Kauwe, T. P. Gordon, D. Cavill, and response in D1-containing neurons. J. Neurosci. 15: 8167–8176. J. McCluskey. 2003. Tolerance through indifference: autoreactive B cells to the 62. Celver, J., M. Sharma, and A. Kovoor. 2010. RGS9-2 mediates specific inhibition nuclear antigen La show no evidence of tolerance in a transgenic model. J. of agonist-induced internalization of D2-dopamine receptors. J. Neurochem. Immunol. 171: 5890–5900. 114: 739–749. 37. Pogue, S. L., and C. C. Goodnow. 2000. Gene dose-dependent maturation and 63. Celver, J., M. Sharma, and A. Kovoor. 2012. D(2)-Dopamine receptors target receptor editing of B cells expressing immunoglobulin (Ig)G1 or IgM/IgG1 tail regulator of G protein signaling 9-2 to detergent-resistant membrane fractions. J. antigen receptors. J. Exp. Med. 191: 1031–1044. Neurochem. 120: 56–69. 38. Kim-Saijo, M., T. Akamizu, K. Ikuta, Y. Iida, K. Ohmori, K. Matsubara, 64. Dale, R. C., V. Merheb, S. Pillai, D. Wang, L. Cantrill, T. K. Murphy, H. Ben- Y. Matsuda, M. Suzuki, F. Matsuda, and K. Nakao. 2003. Generation of Pazi, S. Varadkar, T. D. Aumann, M. K. Horne, et al. 2012. Antibodies to surface a transgenic animal model of hyperthyroid Graves’ disease. Eur. J. Immunol. 33: dopamine-2 receptor in autoimmune movement and psychiatric disorders. Brain 2531–2538. 135: 3453–3468. 39. Li, J., Y. Wang, Q. X. Li, Y. M. Wang, J. J. Xu, and Z. W. Dong. 1998. Cloning of 65. Fukunaga, K., and N. Shioda. 2012. Novel dopamine D2 receptor signaling 3H11 mAb variable region gene and expression of 3H11 human-mouse chimeric through proteins interacting with the third cytoplasmic loop. Mol. Neurobiol. 45: light Chain. World J. Gastroenterol. 4: 41–44. 144–152. 40. Kowal, C., L. A. DeGiorgio, T. Nakaoka, H. Hetherington, P. T. Huerta, 66. Yaddanapudi, K., M. Hornig, R. Serge, J. De Miranda, A. Baghban, G. Villar, B. Diamond, and B. T. Volpe. 2004. Cognition and immunity; antibody impairs and W. I. Lipkin. 2010. Passive transfer of Streptococcus-induced antibodies memory. Immunity 21: 179–188. reproduces behavioral disturbances in a mouse model of pediatric autoimmune Downloaded from 41. Dal Toso, R., B. Sommer, M. Ewert, A. Herb, D. B. Pritchett, A. Bach, neuropsychiatric disorders associated with streptococcal infection. Mol. Psy- B. D. Shivers, and P. H. Seeburg. 1989. The dopamine D2 receptor: two mo- chiatry 15: 712–726. lecular forms generated by alternative splicing. EMBO J. 8: 4025–4034. 67. Hoffman, K. L., M. Hornig, K. Yaddanapudi, O. Jabado, and W. I. Lipkin. 2004. 42. Krisher, K., and M. W. Cunningham. 1985. Myosin: a link between streptococci A murine model for neuropsychiatric disorders associated with group A beta- and heart. Science 227: 413–415. hemolytic streptococcal infection. J. Neurosci. 24: 1780–1791. 43. Swerlick, R. A., M. W. Cunningham, and N. K. Hall. 1986. Monoclonal anti- 68. Dale, R. C. 2005. Post-streptococcal autoimmune disorders of the central ner- bodies cross-reactive with group A streptococci and normal and psoriatic human vous system. Dev. Med. Child Neurol. 47: 785–791.

skin. J. Invest. Dermatol. 87: 367–371. 69. Shannon, K. M., and G. M. Fenichel. 1990. Pimozide treatment of Sydenham’s http://www.jimmunol.org/ 44. Galvin, J. E., M. E. Hemric, K. Ward, and M. W. Cunningham. 2000. Cytotoxic chorea. Neurology 40: 186. mAb from rheumatic carditis recognizes heart valves and laminin. J. Clin. Invest. 70. Demiroren, K., H. Yavuz, L. Cam, B. Oran, S. Karaaslan, and S. Demiroren. 106: 217–224. 2007. Sydenham’s chorea: a clinical follow-up of 65 patients. J. Child Neurol. 45. Cunningham, M. W., H. C. Meissner, J. S. Heuser, B. A. Pietra, D. K. Kurahara, 22: 550–554. and D. Y. Leung. 1999. Anti-human cardiac myosin autoantibodies in Kawasaki 71. Moretti, G., M. Pasquini, G. Mandarelli, L. Tarsitani, and M. Biondi. 2008. What syndrome. J. Immunol. 163: 1060–1065. every psychiatrist should know about PANDAS: a review. Clin. Pract. Epidemol 46. Li, Y., J. S. Heuser, S. D. Kosanke, M. Hemric, and M. W. Cunningham. 2004. Ment. Health 4: 13. Cryptic epitope identified in rat and human cardiac myosin S2 region induces 72. Sokol, M. S. 2000. Infection-triggered anorexia nervosa in children: clinical myocarditis in the Lewis rat. J. Immunol. 172: 3225–3234. description of four cases. J. Child Adolesc. Psychopharmacol. 10: 133–145. 47. Tang, L., R. D. Todd, A. Heller, and K. L. O’Malley. 1994. Pharmacological and 73. Pivonello, R., D. Ferone, G. Lombardi, A. Colao, S. W. Lamberts, and functional characterization of D2, D3 and D4 dopamine receptors in fibroblast L. J. Hofland. 2007. Novel insights in dopamine receptor physiology. Eur. J.

and dopaminergic cell lines. J. Pharmacol. Exp. Ther. 268: 495–502. Endocrinol. 156(Suppl. 1): S13–S21. by guest on September 26, 2021 48. Goodnow, C. C. 1992. Transgenic mice and analysis of B-cell tolerance. Annu. 74. Pollack, A. 2004. Coactivation of D1 and D2 dopamine receptors: in marriage, Rev. Immunol. 10: 489–518. a case of his, hers, and theirs. Sci. STKE 2004: pe50. 49. Putterman, C., B. Deocharan, and B. Diamond. 2000. Molecular analysis of the au- 75. Harris, K., and H. S. Singer. 2006. Tic disorders: neural circuits, neurochemistry, toantibody response in peptide-induced autoimmunity. J. Immunol. 164: 2542–2549. and neuroimmunology. J. Child Neurol. 21: 678–689. 50. DeGiorgio, L. A., K. N. Konstantinov, S. C. Lee, J. A. Hardin, B. T. Volpe, and 76. Kowal, C., L. A. Degiorgio, J. Y. Lee, M. A. Edgar, P. T. Huerta, B. T. Volpe, and B. Diamond. 2001. A subset of lupus anti-DNA antibodies cross-reacts with the B. Diamond. 2006. Human lupus autoantibodies against NMDA receptors me- NR2 glutamate receptor in systemic lupus erythematosus. Nat. Med. 7: 1189–1193. diate cognitive impairment. Proc. Natl. Acad. Sci. USA 103: 19854–19859. 51. Lo Bianco, C., J. L. Ridet, B. L. Schneider, N. Deglon, and P. Aebischer. 2002. 77. Hughes, E. G., X. Peng, A. J. Gleichman, M. Lai, L. Zhou, R. Tsou, alpha -Synucleinopathy and selective dopaminergic neuron loss in a rat T. D. Parsons, D. R. Lynch, J. Dalmau, and R. J. Balice-Gordon. 2010. Cellular lentiviral-based model of Parkinson’s disease. Proc. Natl. Acad. Sci. USA 99: and synaptic mechanisms of anti-NMDA receptor encephalitis. J. Neurosci. 30: 10813–10818. 5866–5875. 52. Pickel, V. M., T. H. Joh, and D. J. Reis. 1975. Ultrastructural localization of 78. Wolstencroft, E. C., G. Simic, N. thi Man, I. Holt, T. Lam, P. R. Buckland, and tyrosine hydroxylase in noradrenergic neurons of brain. Proc. Natl. Acad. Sci. G. E. Morris. 2007. Endosomal location of dopamine receptors in neuronal cell USA 72: 659–663. cytoplasm. J. Mol. Histol. 38: 333–340. 53. Beaulieu, J. M., and R. R. Gainetdinov. 2011. The physiology, signaling, and 79. Parton, R. G., and A. A. Richards. 2003. Lipid rafts and caveolae as portals for pharmacology of dopamine receptors. Pharmacol. Rev. 63: 182–217. endocytosis: new insights and common mechanisms. Traffic 4: 724–738. 54. De Mei, C., M. Ramos, C. Iitaka, and E. Borrelli. 2009. Getting specialized: 80. Sudhof, T. C. 2004. The synaptic vesicle cycle. Annu. Rev. Neurosci. 27: 509– presynaptic and postsynaptic dopamine D2 receptors. Curr. Opin. Pharmacol. 9: 547. 53–58. 81. Lee, J. Y., P. T. Huerta, J. Zhang, C. Kowal, E. Bertini, B. T. Volpe, and 55. Anzalone, A., J. E. Lizardi-Ortiz, M. Ramos, C. De Mei, F. W. Hopf, B. Diamond. 2009. Neurotoxic autoantibodies mediate congenital cortical im- C. Iaccarino, B. Halbout, J. Jacobsen, C. Kinoshita, M. Welter, et al. 2012. Dual pairment of offspring in maternal lupus. Nat. Med. 15: 91–96. control of dopamine synthesis and release by presynaptic and postsynaptic do- 82. Gardoni, F., E. Marcello, and M. Di Luca. 2009. Postsynaptic density-membrane pamine D2 receptors. J. Neurosci. 32: 9023–9034. associated guanylate kinase proteins (PSD-MAGUKs) and their role in CNS 56. Gerfen, C. R., T. M. Engber, L. C. Mahan, Z. Susel, T. N. Chase, F. J. Monsma, disorders. Neuroscience 158: 324–333. Jr., and D. R. Sibley. 1990. D1 and D2 dopamine receptor-regulated gene ex- 83. Hasbi, A., B. F. O’Dowd, and S. R. George. 2011. Dopamine D1-D2 receptor pression of striatonigral and striatopallidal neurons. Science 250: 1429–1432. heteromer signaling pathway in the brain: emerging physiological relevance. 57. Sano, H., Y. Yasoshima, N. Matsushita, T. Kaneko, K. Kohno, I. Pastan, and Mol. Brain 4: 26. K. Kobayashi. 2003. Conditional ablation of striatal neuronal types containing 84. Ben-Pazi, H., J. A. Stoner, and M. W. Cunningham. 2013. Dopamine receptor dopamine D2 receptor disturbs coordination of basal ganglia function. J. Neu- autoantibodies correlate with symptoms in Sydenham’s Chorea. PLoS ONE 8: rosci. 23: 9078–9088. e73516.