Hets 2021

‫סקירות‬ ‫הרפואה • כרך ‪ • 150‬חוב' ‪ • 5‬מאי ‪2011‬‬ ‫האם לטפל בטרשת נפוצה טבה?‬ ‫נורמאליים (‪ ,)NAWM‬דהיינו ללא נגעים‪ ,‬חל אובדן של סיבי חומר‬ ‫לבן‪ .‬בנוסף לפגיעה בחומר הלבן‪ ,‬שהיא פגיעה במעטפת המיאלין סביב‬ ‫כיום מקובל להתחיל בטיפול אימונומודולטורי מונע מייד לאחר‬ ‫האקסונים‪ ,‬קיימת פגיעה גם בחומר האפור‪ ,‬כלומר בנוירונים עצמם‪.‬‬ ‫שאובחנה טרשת נפוצה עם מהלך התקפי לסירוגין [‪ .]24‬מטרת‬ ‫‪ Calabrese‬וחב' [‪ ]10‬הראו‪ ,‬כי מספר הנגעים הקורטיקליים ונפחם‬ ‫הטיפול היא להפחית את תדירות ההתקפים החדים וחומרתם‪,‬‬ ‫(חומר אפור) שאינם מערבים חומר לבן‪ ,‬קטן יותר בחולי טרשת נפוצה‬ ‫ולדחות את התקדמות המוגבלות הנוירולוגית שאליה חשוף החולה‪,‬‬ ‫טבה בהשוואה לחולי טרשת נפוצה עם מהלך התקפי פעיל (‪.)RRMS‬‬ ‫ומכאן הדחיפות בתחילת הטיפול בשלב מוקדם ככל האפשר של‬ ‫במעקב שנערך כעבור שנה‪ ,‬מצאו החוקרים כי התרחשה עלייה‬ ‫משמעותית במספר הנגעים ונפחם רק בקבוצת החולים עם מהלך‬ ‫המחלה [‪.]25‬‬ ‫בעקבות זאת עולה השאלה‪ :‬האם יש צורך להתחיל בטיפול גם‬ ‫התקפי פעיל‪ ,‬ולא בקבוצת החולים עם מהלך טב‪.‬‬ ‫בחולי טרשת נפוצה טבה‪ ,‬שכן טרשת נפוצה טבה היא מחלה שבה‬ ‫במחקר זה הודגם הקשר בין מספר הנגעים הקורטיקליים למצב‬ ‫התקדמות הפגיעה הנוירולוגית נמוכה‪ ,‬ועל כן ייתכן שטיפול אינו‬ ‫הקליני המועדף של חולי טרשת נפוצה טבה [‪ .]10‬בנוסף הודגם‪ ,‬כי גם‬ ‫הכרחי כלל או לפחות שתקופת תחילת הטיפול יכולה להידחות‬ ‫בשנים רבות [‪ .]26‬למעשה‪ ,‬אין תשובה מדעית לשאלה זו‪ ,‬וחלק‬ ‫לפגיעה בחומר האפור יש משמעות בהגדרת טרשת נפוצה טבה‪.‬‬ ‫מחולי טרשת נפוצה טבה בוודאי מקבלים טיפול מונע‪.‬‬ ‫ה‪ .‬הגדרת פעילות המחלה ביחס למיקום הנגעים‬ ‫רופאים אינם מחויבים אתית לטפל אם על פי השקפת דעתם‬ ‫‪ Pittock‬וחב' [‪ ]19‬הראו‪ ,‬כי הממצאים בתהודה מגנטית בחולי טרשת‬ ‫המקצועית אין בטיפול כדי להביא תועלת למטופל [‪ .]27‬ככל שנדע‬ ‫נפוצה טבה משתנים‪ ,‬וקיים טווח רחב של נפח נגעים אפשרי ‪ -‬החל‬ ‫מהם הגורמים הפרוגנוסטיים של טרשת נפוצה טבה קבועה ושל‬ ‫מנפח קטן מאוד וכלה בנפח גדול מאוד‪ .‬דהיינו‪ ,‬הממצאים בתהודה‬ ‫טרשת נפוצה טבה מתחלפת המתקדמת למהלך מחלה פעיל‪ ,‬כך‬ ‫מגנטית אינם במיתאם למדד ה–‪ ,EDSS‬הנמוך בחולי טרשת נפוצה‬ ‫טבה‪ Ceccarelli .‬וחב' [‪ ]20‬הראו‪ ,‬כי ההבדל בין חולי טרשת נפוצה‬ ‫ניטיב להבין מתי דרושה התערבות טיפולית‪.‬‬ ‫טבה לחולי טרשת נפוצה עם מהלך התקפי לסירוגין‪ ,‬מתבטא יותר‬ ‫בפיזור הנגעים ובמיקומם ופחות בהיקף השינויים המבניים במוח‬ ‫לסיכום‬ ‫(מספר הנגעים ונפחם)‪ .‬מסקנתם הייתה‪ ,‬כי המצב הקליני המועדף‬ ‫של חולי טרשת נפוצה טבה נובע ממעורבות נמוכה באונה הקדמית‪,‬‬ ‫הגדרת טרשת נפוצה טבה מחייבת זמן מעקב ארוך‪ .‬לכן‪ ,‬נודעת‬ ‫הן מבחינת הנגעים בחומר הלבן והן מבחינת החומר הלבן שנראה‬ ‫חשיבות למדדים נוספים היכולים להעיד על צורת מחלה זו‪ .‬הערכה‬ ‫נורמאלי (‪.]20[ )NAWM‬‬ ‫קוגניטיבית ומדידות מתקדמות‬ ‫התקדמות המחלה ממצב טב למהלך מחלה‬ ‫בתהודה מגנטית של המוח מספקים ˆ טרשת נפוצה היא מחלה הגורמת‬ ‫מתקדם‪ -‬טרשת נפוצה טבה מתחלפת‬ ‫מידע רב ערך שיכול לעזור בהגדרה לנכות בגיל צעיר‪ ,‬אך בחלק‬ ‫ומעקב אחר חולי טרשת נפוצה מחולי טרשת נפוצה מהלך‬ ‫כפי שתואר לעיל‪ ,‬טרשת נפוצה מוגדרת כמחלה עם מהלך טב באופן‬ ‫טבה‪ .‬לנוכח חשיבות התסמינים המחלה טב (‪ ,)Benign‬והיא מוגדרת‬ ‫רטרוספקטיבי בלבד‪ ,‬לאחר משך מחלה ארוך‪ .‬אולם רבים מהחולים‬ ‫שאינם מוטוריים‪ ,‬ראוי לשקול כטבה באופן רטרוספקטיבי‪,‬‬ ‫שסווגו כבעלי הפנוטיפ הטב‪ ,‬עלולים לפתח לאחר זמן מחלה ארוך‬ ‫להעריך מחדש את ההגדרה של לאחר עשר שנות מחלה‪.‬‬ ‫מחלה מתקדמת וחמורה יותר‪ .‬דהיינו‪ ,‬בחלק מהחולים‪ ,‬מהלך של‬ ‫טרשת נפוצה טבה‪.‬‬ ‫טרשת נפוצה טבה הוא מצב המוגבל בזמן‪ .‬זיהוי גורמים פרוגנוסטיים‬ ‫אנו מציעים לחלק את קבוצת ˆ המהלך הטב בחולי טרשת‬ ‫דמוגרפיים או קליניים‪ ,‬יסייעו להבחין בין חולי טרשת נפוצה טבה‬ ‫הטרשת נפוצה טבה לתתי קבוצות נפוצה יכול להישאר כך או‬ ‫שיעברו למהלך מחלה מתקדם (להלן טרשת נפוצה טבה מתחלפת)‬ ‫על פי הקריטריונים הבאים‪ )1( :‬להפוך פעיל גם לאחר עשור‪.‬‬ ‫לבין אלו שיישארו במצב של טרשת נפוצה טבה לשנים ארוכות (להלן‬ ‫‪ )2( ;EDSS‬תפקוד קוגניטיבי; (‪)3‬‬ ‫היקף הפגיעה ומיקומה בחומר‬ ‫ˆ‬ ‫תהודה מגנטית של המוח‪ .‬קבוצת‬ ‫טרשת נפוצה טבה קבועה)‪.‬‬ ‫החולים עם הסתמנות קלינית של הלבן והאפור של המוח‪ ,‬כפי‬ ‫‪ Sayao‬וחב' [‪ ]21‬עקבו אחר קבוצת חולי טרשת נפוצה טבה‬ ‫כל שלושת המרכיבים שהוזכרו שמודגם בתהודה מגנטית‬ ‫שהוגדרו כבעלי משך מחלה ארוך של עשור‪ ,‬עם פגיעה נוירולוגית‬ ‫לעיל‪ ,‬שונה מקבוצת חולים המציגה (‪ ,)MRI‬ותפקודם הקוגניטיבי‬ ‫על פי מדד ‪ EDSS‬השווה ל–‪ 3.0‬או נמוך ממנו‪ ,‬ובדקו את התקדמות‬ ‫רק חלק מהמרכיבים‪ ,‬והדבר בעל של חולי טרשת נפוצה עם‬ ‫המחלה ‪ 20‬שנה לאחר התפרצותה‪ 52.1% .‬מהחולים המשיכו להיות‬ ‫משמעות פרוגנוסטית וטיפולית‪ .‬מהלך טב‪ ,‬מהווים גורמים‬ ‫מוגדרים כחולים בטרשת נפוצה טבה‪ ,‬בעוד שהשאר עברו למצב‬ ‫כך‪ ,‬בחולה עם ‪ EDSS‬נמוך‪ ,‬פגם פרוגנוסטיים משמעותיים בכל‬ ‫מוגבלות מתקדם יותר מבחינת היקף הנכות הנוירולוגית‪ .‬יתרה‬ ‫קוגניטיבי וממצאים בתהודה הקשור להתקדמות המחלה‪.‬‬ ‫מכך‪ 21.3% ,‬מחולים שאובחנו כחולים בטרשת נפוצה טבה בעשור‬ ‫מגנטית של מחלה מפושטת‪,‬‬ ‫הראשון למחלתם‪ ,‬הפכו מוגבלים בצורה חמורה ביותר ‪ -‬עם ניקוד‬ ‫‪ EDSS‬הגבוה מ–‪ ,6.0‬אשר מעיד על נכות משמעותית וירידה בתפקוד‬ ‫סבירותו להישאר חולה בטרשת נפוצה טבה לאחר משך מחלה ארוך‪,‬‬ ‫[‪ .]21‬בנוסף‪ ,‬נמצא מיתאם בין הנכות הנוירולוגית על פי ה–‪EDSS‬‬ ‫שונה מאשר חולה עם ‪ EDSS‬נמוך‪ ,‬ללא פגם קוגניטיבי וללא ממצאים•‬ ‫ובין הישארות במצב קבוע של טרשת נפוצה טבה [‪.]22‬‬ ‫בתהודה מגנטית‪.‬‬ ‫‪ Glad‬וחב' [‪ ]23‬עקבו אחר קבוצת חולי טרשת נפוצה טבה‪,‬‬ ‫במחקר נמצא מיתאם בין הישארות החולים בקבוצת טרשת נפוצה‬ ‫מחבר מכותב‪ :‬גלעד וינדר‬ ‫טבה קבועה לבין המדדים הפרוגנוסטיים הבאים‪ :‬מהלך מחלה‬ ‫המרכז לטרשת נפוצה‬ ‫התחלתי של התקפים לסירוגין (‪ ;)RRMS‬מין ‪ -‬נשים; גיל צעיר‬ ‫מרכז רפואי שיבא‬ ‫בפרוץ המחלה; וממוצע קצב התקפים בשנה הנמוך מ–‪ 0.2‬לאורך‬ ‫תל השומר‪ ,‬רמת גן‬ ‫שנות המחלה [‪.]23‬‬ ‫דוא\"ל‪464 [email protected] :‬‬ ‫‪49‬‬

2011 ‫ • מאי‬5 '‫ • חוב‬150 ‫הרפואה • כרך‬ ‫סקירות‬ 1. Zifman E & Amital H, Neuropsychological and MRI 15. Mesaros S, Rocca MA, ‫ביבליוגרפיה‬ Prediction of Neurological measures predict short-term Riccitelli G & al, Corpus Diseases by Using evolution in benign multiple Callosum Damage and multiple sclerosis at Autoantibodies: Wishful sclerosis. Neurology, Cognitive Dysfunction in 20 years. Neurology, Thinking Come True. 2009; 73: 494-495. Benign MS. Human Brain 2007;68:496-500. IMAJ, 2008;10:29-31. Mapping, 2009. 9. Portaccio E, stromillo 22. Pittock SJ, Does benign 2. Noseworthy JH, Lucchinetti ML, Goretti B & al, 16. Hines M, Chiu L, McAdams multiple sclerosis today C, Rodriguez & al, Multiple Neuropsychological and MRI LA & al, Cognition and imply benign multiple sclerosis. N Engl J Med, measures predict short term the corpus callosum: sclerosis tomorrow? 2000; 343:938-952. evolution in benign multiple Verbal fluency, visuospatial Implications for treatment. sclerosis. Neurology, ability, and language Neurology, 2007;68:480-481. 3. Hawkins SA & McDonnell 1996; 46:907-911. liberalization related to GV, Benign multiple midsagittal surface area of 23. SB Glad, HI Nyland, JH sclerosis? Clinical course, 10. Calabrese M, Filippi M, callosal subregions. Behav Aarseth & al, Long-term long term follow up, and Rovaris M & al, Evidence Neurosci, 1992; 106:3-14. follow-up of benign assessment of prognostic for relative cortical multiple sclerosis in factors. J Neurol Neurosurg sparing in benign multiple 17. Filippi M, Rocca M, Arnold Hordaland County, Western Psychiatry, 1999; 67:148-152. sclerosis: a longitudinal D & al, EFNS guidelines on Norway. Multiple Sclerosis, magnetic resonance the use of neuroimaging 2009; 15: 942-950. 4. Ramsaransing GS & De imaging study. Multiple in the management of Keyser J, Benign course in Sclerosis, 2009; 15:36-41. multiple sclerosis. Neurol, 24. Thompson AJ, Benign multiple sclerosis: a review. 2006; 13: 313-25. multiple sclerosis Acta Neurol Scand, 2006; 11. Kurtzke FJ, Rating - EDITORIAL 113:359-369. neurologic impairment 18. Yulin GE, Meng L, Grossman COMMENTARY. J Neurol in multiple sclerosis. RI & al, Applications of Neurosurg Psychiatry, 5. Poser S, Wikström J & Neurology, 1983;33:1444. Diffusion Tensor MR 1999;67:138. Bauer HJ, Clinical data Imaging in Multiple and the identification of 12. Amato MP, Zipoli V, goretti Sclerosis. Annals of the New 25. Rudick RA, Disease- special forms of multiple B & al, benign multiple York Academy of Sciences, modifying drugs for sclerosis in 1271 cases sclerosis: cognitive, 2005; 1064: 202-219. relapsing- remitting studied with a standardized psychological and social multiple sclerosis documentation system. J aspects in a clinical cohort. J 19. Pittock SK, Noseworthy and future directions Neurol Sci, 1979; 40:159-68. Neurol, 2006; 253:1054-1059. JH & Rodriguez M, MRI for multiple sclerosis findings in benign multiple therapeutics. Arch Neurol, 6. Hutchinson M, Disability 13. Rao SM, Leo GJ, Ellington L sclerosis are variable. J 1999; 56: 1079 - 1084. due to multiple sclerosis: a & al, Cognitive dysfunction Neurol, 2007; 254:539-541. community-based study of in multiple sclerosis. II. 26. Ramsaransing G, Maurits N, an Irish county. Ir Med J, Impacton employment 20. Ceccarelli I, Rocca MA, Zwanikken C & De Keyser 1986; 79:48-50. and social functioning. Pagani E & al, The J, Early prediction of a Neurology, 1991; 41:692-696. topographical distribution benign course of multiple 7. Ramsaransing G, Maurits of tissue injury in benign sclerosis on clinical N, Zwanikken C & al, Early 14. Rovaris M, Riccitelli G, MS: A 3T multiparametric grounds:a systematic prediction of a benign Judica E & al, Cognitive MRI study. NeuroImage, review. Multiple Sclerosis, course of multiple sclerosis impairment and structural 2008 ;39:1499-1509. 2001; 7: 345-347. on clinical grounds: a brain damage in benign systemic review. Multiple multiple sclerosis. 21. Ana-Luiza Sayao, Virginia 27. Kivity S, Borow M & sclerosis, 2001; 7:345-347. Neurology, 2008; Devonshire & Helen Shoenfeld Y, Hippocrates' 71:1521-1526. Tremlett, Longitudinal Oath is Challenged. 8. Portaccio E, Stromillo follow-up of “benign” IMAJ, 2009; 9: 581-584. ML, Goretti B & el, ‫כרוניקה‬ ?‫האם רכיבה על אופניים קלים מקצרת את זמן הנסיעה‬ ‫ פעמים באופניים‬30 ‫ ק\"ג) או‬9.5( ‫ פעמים באופניים קלים‬26 ‫ רכיבה באופניים לעבודה נפוצה‬,‫בערים צפופות בכלי רכב‬ ‫ שניות‬32 ‫ ההבדל בזמן הנסיעה הממוצע היה‬.)‫ ק\"ג‬13.5( ‫כבדים‬ ‫ האם משקל האופניים משפיע על המאמץ ומשך‬.‫יותר ויותר‬ .)P = 0.72( ?‫הנסיעה‬ ‫ כי אופניים קלים אינם משפיעים על זמן הנסי־‬,‫המחבר סיכם‬ ,‫ שנערך לפי כל כללי המחקר המקובלים‬,‫מחקר אישי‬ ‫ וממליץ לרוכבים להפחית ממשקלם במקום לקנות אופניים‬,‫עה‬ )BMJ ‫בוצע בידי יועץ הרדמה וטיפול נמרץ בשם ג'רמי גרובס‬ ‫ וזמן‬,‫ המשתתף היחיד במחקר היה הוא עצמו‬.2010;341:1296( ...‫יקרים וקלים‬ ‫ נמדד כשהוא רוכב‬,‫ ק\"מ‬43.5 ‫ שמרחקו‬,‫הנסיעה לעבודה וחזרה‬ ‫איתן ישראלי‬ 465 50

Incomplete response to colchicine in M694V homozygote FMF patients Autoimmunity reviews | 2012 ‫ דר' חגית יונת‬:‫מנחה‬ '‫מנהלת מחלקה פנימית א‬ ‫אחראית גנטיקה של מבוגרים במכון הגנטי‬ [email protected] ‫נעמה שחטר‬ ‫אונ' תל אביב‬ ‫השתתפה כסטודנטית בפרויקט ח״ץ‬ 2010-2012 ‫בין השנים‬ [email protected] 51

Autoimmunity Reviews 12 (2012) 72–76 Contents lists available at SciVerse ScienceDirect Autoimmunity Reviews journal homepage: www.elsevier.com/locate/autrev Review Incomplete response to colchicine in M694V homozygote FMF patients Merav Lidar a,b,f, Hagith Yonath c,d,f, Naama Shechter c, Fabienne Sikron e, Siegal Sadetzki e,f, Pnina langevitz a,b,f, Avi Livneh a,b,f, Elon Pras c,f,⁎ a Department of Medicine F, Sheba Medical Center, Israel b The Heller Institute of Medical Sciences, Sheba Medical Center, Tel Hashomer, Israel c The Danek Gartner Institute of Human Genetics, Sheba Medical Center, Tel Hashomer, Israel d Department of Medicine A, Sheba Medical Center, Tel Hashomer, Israel e Cancer and Radiation Epidemiology Unit, Sheba Medical Center, Gertner Institute for Epidemiology and Health Policy Research, Tel Hashomer, Israel f Sackler Faculty of Medicine, Tel Aviv University, Israel article info abstract Available online 2 August 2012 Background: Previous studies have shown that with prophylactic colchicine 65% of the patients suffering from Familial Mediterranean fever (FMF) will show a complete response, 30% a partial response and about 5% will Keywords: show minimum or no response. These studies were performed before the isolation of the disease gene. FMF Genotyping enables us to study the response rates according to specific mutations. We have witnessed a Mutation large number of M694V homozygotes who do not respond well to colchicine despite being treated with max- Colchicine imal sustained doses. M69V homozygotes Aim: To assess the response rates to colchicine in M694V homozygote FMF patients in comparison to other None responders prevalent genotypes. Methods: We conducted a telephonic survey which included 112 FMF patients: 40 M694V homozygotes, and 2 comparison groups of 41 M694V/V726A compound heterozygotes and 31 V726A homozygotes. The ques- tionnaire included demographic, social and clinical features, colchicine dose, response rates and reported side effects. Results: M694 homozygotes showed a more severe disease, and were treated with higher doses of colchicine (av- erage dose 1.98± 0.56 compared to 1.47± 0.58, p =0.0001 and 1.13± 0.41, p b 0.001 in the M694V/V726A com- pound heterozygotes and the V726A homozygotes, respectively); Colchicine related side effects were noted in 40% of the M694V homozygotes. The average rate of attacks in treated M694V homozygotes (0.70±1.06) was higher compared to the two other groups (0.14± 0.26, p =0.002 and 0.08± 0.20, p=0.0009, respectively) and only 25% of them reported no attacks in the last year. None of the patients who took part in this study had amyloidosis. Side effects limiting the dose of colchicine were noted in 40% of the M694V homozygotes. Conclusions: Despite receiving higher doses of colchicine the prevalence of complete responders among M694V homozygotes is much lower than previously appreciated. The results highlight the need for additional treat- ment modalities for these patients. © 2012 Elsevier B.V. All rights reserved. Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 2. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Take-home messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ⁎ Corresponding author at: Institute of Human Genetics, Sheba Medical Center Israel. Tel.: +972 3 5302998; fax: +972 3 5302914. E-mail address: [email protected] (E. Pras). 1568-9972/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.autrev.2012.07.025 Downloaded for Anonymous User (n/a) at Sheba Medical Center from ClinicalKey.com by Elsevier on October 19, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved. 52

M. Lidar et al. / Autoimmunity Reviews 12 (2012) 72–76 73 1. Introduction included 41 M694V/V726A compound heterozygotes and 31 V726A ho- mozygotes. The patients for each genotype were randomly chosen from Familial Mediterranean fever (FMF) is an autoinflammatory disorder a list of more than 1000 patients genotyped in the last 10 years for characterized by recurrent attacks of fever accompanied by sterile peri- MEFV mutations. Clinical information was obtained through a tele- tonitis, arthritis, pleuritis, and a typical inflammatory skin rash termed phone interview by the aid of a structured questioner that included de- erysipelas-like erythema (ELE). The development of renal amyloidosis tails on the ethnic origin, the age of disease onset, the age of diagnosis, type AA is the most devastating manifestation of the disease and in the the number of attacks before treatment, the disease symptoms, the past was a major cause of morbidity and mortality in these patients presence of amyloidosis, the dose of colchicine, side effects from treat- [1]. The disease is caused by mutations in the MEFV gene, which is com- ment, response to treatment, years of formal education, type of profes- posed of 10 exons and encodes a 781 amino acid protein [2]. To date sion (academic or non-academic), employment status, reception of more than 50 disease-associated mutations have been identified, most social benefits and divorce rates. of which are extremely rare (Infevers data base http://fmf.igh.cnrs.fr/ infevers). Very high FMF carrier rates have been described among the Demographic and clinical characteristics of the M694V homozygote Mediterranean and Middle Eastern population ranging from 1:5 in group were compared with each of the comparison groups using the X2 North African Jews, Arabs and Turks, to 1:3 among Iraqi Jews and Arme- test for discrete variables and the t‐test for continuous variables. nians [3–6]. Most patients have mutations in exon 10 which is the lon- gest exon in this gene, located in the C-terminal end of the protein The means of colchicine dose and frequency of attacks before and encoding the B30.2 domain. The 2 most common exon 10 mutations after treatment were calculated and compared between the M694V ho- are M694V which in Israel is found mainly among North African Jews mozygote group and each of the two comparison groups using t-test. but also in Iraqi Jews and in Arabs, and V726A which is prominent in Ashkenazi Jews but can also be found in Iraqi Jews, Arabs and Druze All of the statistical analyses were done with Statistical Analysis [7]. M694V in the homozygous state predisposes to a severe disease System software version 9.1 (SAS Institute, Inc., Cary, NC). Statistical compared to other mutations [8]. The vast majority of FMF patients are significance was set at p b 0.05, using two-tailed tests. treated with prophylactic colchicine which was introduced in the mid 70s of the last century [9]. Under this treatment 60–65% of the patients 3. Results were reported to achieve a complete remission, 30–35% a partial remis- sion and 5–10% were defined as non-responders [10]. These studies No significant differences were found in the age or sex distribution were performed before patients could be stratified according to their between the M694V homozygotes group and both comparison groups mutations. A recent study from Turkey on pediatric M694V homozy- (the M694V/V726A compound heterozygote group and the V726A ho- gotes has reported complete response and non‐response in 36% and mozygote group). While 67% of the M694V homozygotes were of 18% of the patients, respectively, response rates way below the expected North African origin, 40% of the V726A homozygotes were of Ashkenazi [11]. and/or Iraqi, reflecting the frequency of these two mutations in these populations (pb 0.001). Significant differences between the M694V ho- In this study we assessed the response rates to colchicine among mozygotes and each of the two comparison groups were detected in the adult M694V homozygotes and compared them to two groups of pa- age of onset, the age at diagnosis, the frequency of attacks prior to treat- tients, M694V/V726A compound heterozygotes and V726A homozy- ment and the appearance of arthritis. M694V homozygotes also gotes. We show that despite receiving high doses of colchicine, reported a much higher frequency of exertional leg pain, a complaint response rates in M694V homozygotes are far less favorable than pre- that appears between the attacks and shows poor response to treat- viously appreciated. ment. No differences were detected in the frequency of the abdominal and pleural attacks between the groups (Table 1). Interestingly, none 2. Materials and methods of the V726A homozygote patients suffered from attacks manifested only by fever. M694V homozygotes were treated with higher doses of Patients were recruited at the FMF clinic at the Sheba Medical Cen- colchicine (average dose close to 2 mg, Table 2). Forty percent of them ter, Israel. The study was approved by the institutional review board reported side effects, compared to 17% and 23% in the M694V/V726A and participants gave informed consent. compound heterozygotes and the V726A homozygotes, respectively. In none of the patients did the side effects which consisted solely of gas- Overall 112 patients previously tested for MEFV mutations were in- trointestinal complaints, result in complete stop of colchicine, rather the cluded; 40 homozygotes for M694V and 2 comparison groups that dose was lowered to the maximal tolerable amount. Table 1 Clinical characteristics of the patients by mutation. M694V homozygotes M694V/V726A V726A homozygotes P12a (M694V homozygotes P13b (M694V homozygotes compound heterozygotes vs. M694V/V726A heterozygotes) vs. V726A homozygotes) Total 40 41 31 0.94 0.09 Age 40.0 ± 12.7 39.8 + 14.6 45.3 + 12.8 Range 23–67 24–79 26–66 0.92 0.09 Females 52.5 53.7 32.3 b 0.0001 b 0.0001 Age at onset 5.65 ± 5.8 15.59 ± 12.9 24.42 ± 11.7 b 0.0001 b 0.0001 Age at diagnosis 14.15 ± 10.9 23.22 ± 13.0 33.71 ± 12.0 No. of attacks per month 2.6 ± 2.0 1.5 ± 1.3 1.3 ± 1.63 0.0025 0.005 0.17–8.0 0.17–4.5 0.083–8.0 before treatment 11 (27.5%) 11(26.8%) 0 (0.0%) 0.94 b 0.001 Fever only 35 (87.5%) 37 (90.2%) 29 (93.6)% 0.73 0.45 Abdominal attacks 19 (47.5%) 22 (53.7%) 9 (29.0)% 0.60 0.11 Pleural attacks 32 (80.0%) 17 (41.5%) 4 (12.9%) b 0.0004 Arthritis 7 (17.5%) 2 (4.9%) 2 (6.4%) 0.09 b 0.0001 ELE 35 (89.7%) 19 (47.5%) 15 (48.4%) b 0.0001 0.28 Exertional leg pain 0 0 0 Amyloidosis b 0.0001 a p12: the p value when comparing the first and second groups (homozygote—M694V with compound heterozygote. b p13: the p value when comparing the first and third groups (homozygote—M694V with homozygote—V726A. Downloaded for Anonymous User (n/a) at Sheba Medical Center from ClinicalKey.com by Elsevier on October 19, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved. 53

74 M. Lidar et al. / Autoimmunity Reviews 12 (2012) 72–76 Table 2 M694V/V726A V726A homozygotes P12 (M694V homozygotes P13 (M694V homozygotes The effects of colchicine by study group/mutation. compound heterozygotes vs. M694V/V726A heterozygotes) vs. V726A homozygotes) 31 M694V homozygotes 41 1.13 ± 0.41 0.0001 p b 0.001 1.47 ± 0.58 7 (22.6%) 0.05 0.10 Total 40 7 (17.1%) 0.08 ± 0.2 0.002 0.0009 Dose of colchicine 1.98 ± 0.56 0.14 ± 0.26 Side effects from treatment 16 (40.0%) 0–1.0 Average number of attacks 0.70 ± 1.06 0–1.25 per month on treatment 0–4.33 M694V homozygotes were much less responsive to treatment com- Other therapeutic options have been tried, with varying success pared to the other two groups. M694V homozygotes suffered from an rates only in individual cases or in small, non-randomized trials. None- average attack rate of 0.70± 1.06 per month compared to 0.14± 0.26 theless, patients suffering frequent attacks on a maximal tolerated oral (p= 0.002) and 0.08 ±0.20 (p= 0.0009) in the M694V/V726A com- dose of colchicine, may be offered a weekly intravenous supplement pound heterozygotes and the V726A homozygotes, respectively or the addition of a TNF alpha blocking agent [15]. Oral thalidomide (Table 2). Only 25% of the M694V homozygotes reported no attacks in and interferon-alpha inhibitors are in rare use nowadays, due to disqui- the last year compared to 51% and 74% in the M694V/V726A compound eting side effects and limited utility [15]. Given that IL-1 has emerged as heterozygotes and the V726A homozygotes, respectively (Table 3). Four the major cytokine mediating FMF associated inflammation, IL-1 M694V homozygotes reported no change in symptoms or worsening of blocking agents are the current therapeutic focus in severe cases of symptoms on colchicine compared to only one such patient in the FMF. Anecdotal reports suggest prompt response in the majority of pa- M694V/V726A compound heterozygote group and another in the tients, for a variety of indications including resistant FMF in the pres- V726A homozygote group. These patients can be defined as non- ence of secondary amyloidosis and colchicine-intolerance in renal responders. It is important to emphasize that none of the patients failure or transplanted patients [16]. who participated in this study suffered from amyloidosis, reflecting the fact that even in poor responders or non-responders, colchicine Presently, we re-assessed the response to colchicine in FMF patients, still has an important role in preventing this devastating complication. showing that an inadequate response rate (more than one attack every two months) is not uniformly distributed among the different FMF geno- No differences between the groups were detected in divorce rate, types. M694V homozygotes exhibited a hefty, 30% inadequate response the number of years of formal education, or the number of those who rate, which is 5–6 fold higher than reported in V726A homozygotes and hold an academic profession. M694V homozygotes disclosed signifi- M694V/V726A compound heterozygotes. It is well appreciated that the cantly higher rates of unemployment compared with the V726A ho- M694V mutation, more so in its homozygous form, is associated with a mozygotes and were more likely to receive social benefits compared severe clinical phenotype. Indeed, soon after the advent of the FMF with the compound heterozygote group (Table 4). gene we, and others, have shown that patients with the M694V/M694V genotype had an earlier age of disease onset, a higher frequency of joint 4. Discussion and skin involvement and required higher doses of colchicine to control the disease compared to the other 3 prevalent genotypes [8,17,18]. Also, Colchicine, an alkaloid used in the treatment of gout since the first the contribution of M694V homozygosity to the development of renal century, has constituted the cornerstone of treatment for FMF in the amyloidosis, independent from disease severity, was recognized [19]. past 40 years, drastically changing the natural history of the disease. The rate of amyloidosis in M694V homozygotes who received no colchi- Aside from its success at attenuating attack frequency and severity cine before the age of 20 years, was found to be an astonishing 61% [20]. in the majority of FMF patients, colchicine is effective at preventing and arresting the development of renal amyloidosis [12,13]. Over the years, we have developed two sets of severity scores for FMF. The earlier of the two, also known as the \"Pras severity score\" Colchicine is given in a daily dose varying between 1 to 2 mg/d, [21], included criteria such as age of disease onset, number of attacks depending on the patient's clinical response. Only a small proportion per month before initiation of colchicine treatment, presence of ar- of patients benefit from escalation of the dose above 2 mg/d while the thritis, erysipelas-like erythema, amyloidosis and daily colchicine majority attains no additional clinical benefit. Moreover, with increasing dose needed to control attacks. The latter, \"Mor severity scale\"[22], doses, the rate of side effects, notably gastrointestinal cramps, abdomi- omitted the number of monthly attacks in order to avoid recall bias nal pain and diarrhea, which frequently hinder compliance even at a and added in its place criteria such as the number of attack sites in- lower dose, increases, as does potential drug toxicity [14]. Over the volved during a single attack and over the course of the disease. years colchicine has emerged as a wonder drug for FMF. It is cheap, con- Whereas in the Pras score, a colchicine dose of 1 mg/d received a sin- venient, safe and effective and for most FMF patients there is no practical gle severity point, increasing stepwise to 4 points in patients who re- alternative. quired 2 mg/d or more in order to control attacks, the Mor score included among its severity criteria only the need for ≥2 mg/d of Table 3 Distribution of attacks among patients. M694V homozygotes M694V/V726A V726A homozygotes P12 (M694V homozygotes P13 (M694V homozygotes compound heterozygotes vs. M694V/V726A heterozygotes) vs. V726A homozygotes) Total (n) 40 41 31 0.004 b 0.0001 Patients without attacks for the 10 (25.0%) 21 (51.2 %) 23 (74.2%) last year 17 (42.5%) 18 (43.9%) 7 (22.6%) Patients with attacks up to once 12 (30.0%) 2 (4.9%) 1 (3.2%) every 2 months Patients with more than one attack 1 (2.5%) 0 (0.0%) 0 (0.0%) every two months Unknown Downloaded for Anonymous User (n/a) at Sheba Medical Center from ClinicalKey.com by Elsevier on October 19, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved. 54

M. Lidar et al. / Autoimmunity Reviews 12 (2012) 72–76 75 Table 4 Social and demographic details. M694V homozygotes M694V/V726A compound heterozygotes V726A homozygotes P12 (M694V homozygotes P13 (M694V homozygotes vs. M694V/V726A heterozygotes) vs. V726A homozygotes) Total 40 41 31 0.79 0.46 Years of formal 14.0 ± 2.44 13.8 ± 3.87 14.5 ± 3.56 0.82 0.15 education 18 (45.0%) 20 (48.8%) 20 (64.5)% Academic profession 0.35 0.02 14 (34.2%) 10 (24.3%) 3 (9.7%) 0.04 0.06 or students 8 (20.0%) 2 (4.9%) 1 (3.2%) 0.35 0.68 Unemployed 4 (12.9%) 1 (3.3%) 3 (7.4%) Social benefits Divorced colchicine to control attacks. Evidently, the underlying assumption in 5. Conclusions both cases was that colchicine dose can be increased to the point of clinical response without constraint rendering response to treatment, We show that FMF patients who are M694V homozygotes, demon- a seemingly redundant criterion. However, presently we show that a strated a less favorable response to prophylactic colchicine treatment, significant proportion of M694V homozygotes cannot tolerate the manifested by the need for higher daily doses and resulting in an in- colchicine dose needed to control their attacks and exhibit incom- creased rate of side effects. This group of poor and non-responder FMF plete clinical response on their maximal tolerated dose. Notably, patients shall be the prime candidates for new alternative therapies, M694V homozygotes experienced more side effects from colchicine such as IL-1 inhibition, in the near future. compared to M694V/V726A compound heterozygotes and V726A ho- mozygotes, indisputably due to consumption of higher colchicine Take-home messages doses. This is evident in the significantly lower response rates in the M694V homozygotes than in the two other control groups. Further- • FMF is caused by mutations in the MEFV gene, which is composed of more, a complete response was noted in 51% of the M694V/V726A 10 exons. compound heterozygotes and in close to 72% of the V726A homozy- gotes reflecting the fact that the M694V/V726A and V726A homozy- • The 2 most common mutations are V726A and M694V which pre- gotes are intermediate and mild genotypes, respectively. It may be disposes to a severe disease. inferred that incomplete clinical response in M694V homozygotes prompts physicians to escalate colchicine dosage. However, when • The vast majority of FMF patients are treated with prophylactic col- side effects develop before full clinical response is attained; colchicine chicine. dose is subsequently lowered ultimately maintaining patients on a suboptimal dose. The narrow therapeutic/toxic ratio of colchicine be- • M694V homozygote patients demonstrate a less favorable response comes a major predicament in the chronic treatment of M694V ho- to colchicine. mozygotes, analogous to the situation encountered when applying colchicine in acute gout. Notwithstanding, we also re-confirm that • New alternative therapies are needed for poor and non-responder about 90% of the M694V homozygotes do respond, albeit incomplete- FMF patients. ly, to colchicine and most importantly, do not develop amyloidosis. References The decreased response rate of M694V homozygotes to colchicine was suggested previously, for example in a Turkish study conducted [1] Sohar E, Gafni J, Pras M, Heller H. Familial Mediterranean fever. A survey of 470 among 222 pediatric FMF patients [11]. On the other hand, while we cases and review of the literature. Am J Med 1967;43:227-53. have previously shown a 2-fold concentration of colchicine in mononu- clear blood cells in colchicine-responsive as opposed to non-responsive [2] The International FMF Consortium. Ancient missense mutations in a new member patients, these clinical phenotypes did not correlate with M694V homo- of the RoRet gene family are likely to cause familial Mediterranean fever. Cell zygosity or with any other particular genotype [23]. This may be 1997;90:797-807. explained by the fact that while homozygosity to M694V clearly plays a role in incomplete response to colchicine, an additional genetic factor, [3] Kogan A, Shinar Y, Lidar M, Revivo A, Langevitz P, Padeh S, et al. Common MEFV namely the P glycoprotein efflux pump (MDR-1), is implicated in col- mutations among Jewish ethnic groups in Israel: high frequency of carrier and chicine response or lack thereof. This pump is responsible for the trans- phenotype III states and absence of a perceptible biological advantage for the car- port of colchicine into the polymorphonuclear cells and is encoded by rier state. Am J Med Genet 2001;102:272-6. the ABCB-1 gene [24]. Two recent studies have found an association be- tween a polymorphism in MDR-1 and the response to colchicine in FMF [4] Ozer FL, Kaplaman E, Zileli S. Familial Mediterranean fever in Turkey. A report of patients [25,26]. twenty cases. Am J Med 1971;50:336-9. Presently, we did not find differences in the level of education, the [5] Barakat MH, Karnik AM, Majeed HW. el-Sobki NI, Fenech FF. Familial Mediterra- percentage of patients with an academic profession, unemployment nean fever (recurrent hereditary polyserositis) in Arabs: a study of 175 patients rates or divorce rates between M694V homozygotes and the control and review of the literature. Q J Med 1986;233:837-47. groups. We did, however, find higher rates of patients receiving social benefits among this group. Homozygosity for M694V is common among [6] Rogers DB, Shohat M, Petersen GM, Bickal J, Congleton J, Schwabe AD, et al. Famil- North African Jewish decedents, a group that socio-economically fell ial Mediterranean fever in Armenians: autosomal recessive inheritance with high below the average in Israel. In recent years however, the gap is rapidly gene frequency. Am J Med Genet 1989;34:168-72. closing and this is manifested by the non-significant differences in the pa- rameters mentioned above. Thus, our results suggest that social factors [7] Zaks N, Shinar Y, Padeh S, Lidar M, Mor A, Tokov I, et al. Analysis of the three most play a small role, if any, in the determination of disease severity and re- common MEFV mutations in 412 patients with familial Mediterranean fever. Isr sponse to colchicine. Med Assoc J 2003;5:585-8. [8] Shinar Y, Livneh A, Langevitz P, Zaks N, Aksentijevich I, Koziol DE, et al. Genotype– phenotype assessment of common genotypes among patients with familial Med- iterranean fever. J Rheumatol 2000;27:1703-7. [9] Zemer D, Revach M, Pras M, Modan B, Schor S, Sohar E, et al. A controlled trial of colchicine in preventing attacks of familial Mediterranean fever. N Engl J Med 1974;291:932-4. [10] Zemer D, Pras M, Sohar E, Gafni J. Continuous colchicine therapy in familial Med- iterranean fever. Harefuah 1976;90:440-2. [11] Soylemezoglu O, Arga M, Fidan K, Gonen S, Emeksiz HC, Hasanoglu E, et al. Unresponsiveness to colchicine therapy in patients with familial Mediterranean fever homozygous for the M694V mutation. J Rheumatol 2010;37:182-9. [12] Zemer D, Pras M, Sohar E, Modan M, Cabili S, Gafni J. Colchicine in the prevention and treatment of the amyloidosis of familial Mediterranean fever. N Engl J Med 1986;314:1001-5. [13] Livneh A, Zemer D, Langevitz P, Laor A, Sohar E, Pras M. Colchicine treatment of AA amyloidosis of familial Mediterranean fever. An analysis of factors affecting outcome. Arthritis Rheum 1994;37:1804-11. Downloaded for Anonymous User (n/a) at Sheba Medical Center from ClinicalKey.com by Elsevier on October 19, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved. 55

76 M. Lidar et al. / Autoimmunity Reviews 12 (2012) 72–76 [14] Livneh A, Langevitz P, Zemer D, Padeh S, Migdal A, Sohar E, et al. The changing [20] Mimouni A, Magal N, Stoffman N, Shohat T, Minasian A, Krasnov M, et al. Familial face of familial Mediterranean fever. Semin Arthritis Rheum 1996;26:612-27. Mediterranean fever: effects of genotype and ethnicity on inflammatory attacks and amyloidosis. Pediatrics 2000;105:E70. [15] Lidar M, Scherrmann JM, Shinar Y, Chetrit A, Niel E, Gershoni-Baruch R, et al. Colchi- cine nonresponsiveness in familial Mediterranean fever: clinical, genetic, pharmaco- [21] Pras E, Livneh A, Balow Jr JE, Kastner DL, Pras M, Langevitz P. Clinical differences be- kinetic, and socioeconomic characterization. Semin Arthritis Rheum 2004;33: tween North African and Iraqi Jews with familial Mediterranean fever. Am J Med 273-82. Genet 1998;75:216-9. [16] Meinzer U, Quartier P, Alexandra JF, Hentgen V, Retornaz F, Kone-Paut I. Interleu- [22] Mor A, Shinar Y, Zaks N, Langevitz P, Chetrit A, Shtrasburg S, et al. Evaluation of dis- kin-1 targeting drugs in familial Mediterranean fever: a case series and a review ease severity in familial Mediterranean fever. Semin Arthritis Rheum 2005;35:57-64. of the literature. Semin Arthritis Rheum 2011;41:265-71. [23] Lidar M, Scherrmann JM, Shinar Y, Chetrit A, Niel E, Gershoni-Baruch R, et al. Colchi- [17] Majeed HA, El-Shanti H, Al-Khateeb MS, Rabaiha ZA. Genotype/phenotype corre- cine nonresponsiveness in familial Mediterranean fever: clinical, genetic, pharmaco- lations in Arab patients with familial Mediterranean fever. Semin Arthritis Rheum kinetic, and socioeconomic characterization. Semin Arthritis Rheum 2004;33: 2002;31:371-6. 273-82. [18] Samuels J, Aksentijevich I, Torosyan Y, Centola M, Deng Z, Sood R, et al. Familial [24] Niel E, Scherrmann JM. Colchicine today. Joint Bone Spine 2006;73:672-8. Mediterranean fever at the millennium. Clinical spectrum, ancient mutations, [25] Tufan A, Babaoglu MO, Akdogan A, Yasar U, Calguneri M, Kalyoncu U, et al. Asso- and a survey of 100 American referrals to the National Institutes of Health. Med- icine (Baltimore) 1998;77:268-97. ciation of drug transporter gene ABCB1 (MDR1) 3435C to T polymorphism with colchicine response in familial Mediterranean fever. J Rheumatol 2007;34:1540-4. [19] Gershoni-Baruch R, Brik R, Zacks N, Shinawi M, Lidar M, Livneh A, et al. The con- [26] Ozen F, Silan C, Uludag A, Candan F, Silan F, Ozdemir S, et al. Association between tribution of genotypes at the MEFV and SAA1 loci to amyloidosis and disease se- ABCB1 (MDR1) gene 3435 C>T polymorphism and colchicine unresponsiveness verity in patients with familial Mediterranean fever. Arthritis Rheum 2003;48: of FMF patients. Ren Fail 2011;33:899-903. 1149-55. Double allogenic mesenchymal stem cells transplantations could not enhance therapeutic effect compared with single transplantation in systemic lupus erythematosus. The clinical trial of allogenic mesenchymal stem cells (MSCs) transplantation for refractory SLE patients has shown significant safety and ef- ficacy profiles. However, the optimum frequency of the MSCs transplantation (MSCT) is unknown. In this concern, Wang D, et al. (Clin Dev Immunol 2012;2012:273291) observed whether double transplantations of MSCs is superior to sin- gle transplantation. Fifty-eight refractory SLE patients were enrolled in this study, in which 30 were randomly given single MSCT, and the other 28 were given double MSCT. Patients were followed up for rates of survival, disease remission, and relapse, as well as transplanta- tion-related adverse events. SLE disease activity index (SLEDAI) and serologic features were evaluated. These results showed that no remark- able differences between single and double allogenic MSCT were found in terms of disease remission and relapse, amelioration of disease activity, and serum indexes in an SLE clinical trial with more than one year followup. This study demonstrated that single MSCs transplan- tation at the dose of one million MSCs per kilogram of body weight was sufficient to induce disease remission for refractory SLE patients. Downloaded for Anonymous User (n/a) at Sheba Medical Center from ClinicalKey.com by Elsevier on October 19, 2021. For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved. 56

Pupillometer-based objective chromatic perimetry in normal eyes and patients with retinal photoreceptor dystrophies Investigative Ophthalmology & Visual Science | 2013 I have being taking part in the Arrow project since 2012. It ‫ פרופ' יגאל רוטנשטרייך‬:‫מנחה‬ was a great pleasure to supervise young MD students at early stages of their career. I believe that research is a vital part of ,‫ראש מרפאה לאלקטרופיזיולוגיה‬ being a good physician. In the research projects the students ‫מכון גולדשלגר לחקר העין‬ became familiar with the basic strategies designing a clinical trial, defining study aims, patient selection and statistical [email protected] plan. They learned the importance of data collection accuracy and organizing the data as well as experience in ophthalmology, the study of which is very limited during the regular clinical studies. I believe these are important tools that that will help them in their daily medical work. As the famous phrase: \"I have learned from all my teachers, but I've learned much more from my student\" . The young students contributed immensely to the success of the projects with fresh new ideas and high motivation to learn.   The ultra-fast accelerating changes in medicine in the 21st century,  present unprecedented challenges to practicing MDs. I believe that medical research enables physicians to remain updated and to better evaluate the new changes and adapt to them. A good analogy would be baking a cake versus reading a book of recipes.  As a surgeon and clinician researcher, I know how hard it is to find the time and energy to apply this in our busy schedule. Pursuing medical research enables the MDs to apply their research-based insights when treating patients, not strictly adhering to the protocols, with more personalized medicine approaches. The Arrow Project enables the students to experience the world of research clinicians. The aim for me was to teach MD students how to employ research tools for better, personalized patient care. ‫מוחמד מחאגנה‬ ‫אונ' תל אביב‬ ‫השתתף כסטודנט בפרויקט ח\"ץ‬ 2012-2014 ‫בין השנים‬ [email protected] 57

Retina Pupillometer-Based Objective Chromatic Perimetry in Normal Eyes and Patients With Retinal Photoreceptor Dystrophies Alon Skaat,1 Ifat Sher,1 Andrew Kolker,1,2 Sivan Elyasiv,1 Elkana Rosenfeld,1,3 Mohamad Mhajna,1,3 Shlomo Melamed,1 Michael Belkin,1,3 and Ygal Rotenstreich1,3 1The Maurice and Gabriela Goldschleger Eye Institute, Sheba Medical Center, Tel Aviv University, Tel Hashomer, Israel 2Department of Ophthalmology, The George Washington University, Washington, DC 3The Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel Correspondence: Ygal Rotenstreich, PURPOSE. To evaluate a novel objective perimetry using multifocal chromatic pupil light reflex Goldschleger Eye Research Institute, in normal participants and patients with photoreceptor dysfunction, and to relate this new Sheba Medical Center, 52621 Tel technique with subjective dark-adapted chromatic Goldmann perimetry. Hashomer, Israel; Ygal.rotenstreich@ sheba.health.gov.il. METHODS. Thirty-two eyes of 17 retinitis pigmentosa (RP) or cone–rod dystrophy patients and 20 eyes of 12 healthy individuals were tested. A computerized infrared video pupillometer AS and IS contributed equally to the was used to record changes in pupil diameter in response to short- and long-wavelength work presented here and should stimuli (peak 485 and 640 nm, respectively; light intensity 40 cd/m2) at 13 different points of therefore be regarded as equivalent the 308 visual field (VF), under background illumination of 2.7 cd/m2. The pupillary response authors. (PR) of patients was compared with PR obtained from normal control participants. In 11 patients, the pupillary responses were also compared with their findings on dark-adapted Submitted: October 12, 2012 chromatic Goldmann perimetry. Accepted: March 5, 2013 RESULTS. Significantly reduced pupillary responses were obtained in RP patients in response to Citation: Skaat A, Sher I, Kolker A, et the short-wavelength stimulus in nearly all perimetric locations (P < 0.03). By contrast, in al. Pupillometer-based objective response to the long-wavelength stimulus, RP patients demonstrated significantly reduced PR chromatic perimetry in normal eyes mostly in peripheral locations (P � 0.02). In a cone–rod dystrophy patient, the PR to both and patients with retinal long- and short-wavelength stimuli was significantly lower in the scotoma area identified by photoreceptor dystrophies. Invest the dark-adapted chromatic Goldmann perimetry. In all patients that were tested by the Ophthalmol Vis Sci. 2013;54;2761– chromatic Goldmann, minimal PR was recorded in areas that were nondetected in the 2770. DOI:10.1167/iovs.12-11127 chromatic Goldmann perimetry. CONCLUSIONS. This study demonstrates the potential feasibility of using pupillometer-based chromatic perimetry for objectively assessing VF defects and retinal function in patients with retinal dystrophies. (ClinicalTrials.gov number, NCT01021982.) Keywords: visual field, perimetry, retinitis pigmentosa, pupils, retinal dystrophy E valuation of visual field (VF) defects is important for clinical examinations varies considerably, further emphasizing the need diagnosis and monitoring of various ophthalmologic diseas- for new technological advances that allow earlier and objective es. VF is assessed mainly by subjective perimetry techniques, detection of VF defects and their progression with higher levels including standard automated perimetry and short-wavelength of certainty than are currently available.9 automated perimetry.1 Two insurmountable limitations of these methods are the need for patient cooperation and the The pupillary light reflex is an objective indicator of retinal subjectivity of patients’ responses.2 Therefore, testing of young and optic nerve functions.10–13 Several studies used a pupil- children, the elderly, and individuals with compromised lometer with achromatic stimulus for objective determination communication is almost certain to yield unreliable results. of the visual field.9,12–23 However, a comparison between visual Moreover, patients’ responses can be affected by their levels of and papillary sensitivity revealed that they are not sufficiently fatigue, wakefulness, and attentiveness during the long correlated to be of clinical use.9,14–23 A different method of procedure. Hence, constant monitoring and instruction of multifocal pupillographic perimetry (TrueField Analyzer; Tek- participants by suitably qualified personnel are needed to tronix, Beaverton, OR) uses white and colored stimuli analyzing obtain reliable results. Furthermore, test–retest variability, both eyes simultaneously.24–26 Although this technology is particularly in peripheral locations and in regions of VF deficits, promising, it cannot differentiate between the rod and cone makes it difficult to determine whether the VF is worsening systems. over the course of serial examinations.1,3–5 Frequent examina- tions are needed and misdiagnosis of early stages is common.6–8 Kardon et al.,27,28 using full-field stimuli, developed a Unfortunately, in routine clinical practice the frequency of VF protocol for assessing the contribution of rods, cones, and melanopsin ganglion cells to the pupillary response (PR). These studies provided evidence that the PR to different wavelengths, Copyright 2013 The Association for Research in Vision and Ophthalmology, Inc. 2761 www.iovs.org j ISSN: 1552-5783 Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 58

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2762 stimulus intensities, and stimulus durations reflects activation Light Stimuli of different outer and inner retinal cells. It was suggested that the transient PR to a low-intensity, short-wavelength stimulus Light stimuli were presented using a ganzfeld dome apparatus reflects rod activity, that the transient PR to a long-wavelength (multifocal dark-adaptometer; Roland Consult Stasche & Finger stimulus is predominantly driven by cones, and that a sustained GmbH) placed 330 mm from the patient’s eye, and controlled PR to a continuous high-intensity short-wavelength stimulus is with a stimulus generator and custom software. The untested derived primarily from the direct intrinsic activation of eye was occluded. Stimuli were presented from the center, and melanopsin-containing retinal ganglion cells (mRGCs).27–29 participants were asked to fixate on a red light-emitting diode These and similar protocols were successfully used to assess fixation light presented from 13 different locations in the VF the function of outer and inner retinal cells in patients with (central, superior, inferior, temporal, and nasal fields at angles retinitis pigmentosa (RP) and patients with RPE65 muta- of 108, 208, and 308). Wavelengths of the light stimuli selected tions.30,31 However, because these methods use a wide light for this study were 640 6 5 nm for red light (long wavelength) source that stimulates the entire retina, they are not applicable and 480 6 5 nm for blue light (short wavelength). A light for multifocal testing to identify VF defects. intensity of 40 cd/m2 was chosen after preliminary calibrations that enabled us to identify the minimal stimulus intensity that In a previous study we demonstrated that a modified yielded a substantial PR in peripheral VF locations in five Goldmann dark-adapted chromatic perimeter can be used to normal participants. Each stimulus was presented using identify cone or rod VF defects.32 Here, we tested the stimulus size V (64 mm2) on a background luminance of 2.7 possibility of using a novel chromatic perimetry technique in cd/m2. Stimulus duration was 1000 ms and the interstimulus which the retina was stimulated in a multifocal pattern by interval was 10 seconds. using a narrow (64 mm2) light beam at different wavelengths. Use of this narrow beam resulted in generation of an objective Pupil Measurement VF test. The PR in this modified system (dark-adaptometer; Roland Consult Stasche & Finger GmbH, Brandenburg, Pupil diameters were recorded in real time by a computerized Germany) was automatically recorded at various VF locations. infrared pupillometer (Roland Consult Stasche & Finger We compared between normal participants and patients with GmbH), which consisted of a monitor with viewing optics RP or cone–rod dysfunction. We also associated the chromatic for presentation of a light stimulus to the subject. Pupil pupillometer-based perimetry findings of patients with their tracking was performed by an infrared high-resolution camera electroretinography (ERG), and dark-adapted chromatic Gold- inside the dark-adaptometer that recorded the PR at a sampling mann perimetry recordings. rate of 34 Hz. The software (Roland Consult Stasche & Finger GmbH) searched for the pupil in every image. A correction METHODS factor was used to get the diameter in millimeters and pupil diameters were measured with an accuracy of 0.1 mm (Roland Participants Consult Stasche & Finger GmbH). The study was conducted according to the tenets of the The subject’s eye was inclined at 158 to the center, at the Declaration of Helsinki, received approval from the Sheba position where the stimulus was presented (Fig. 1). The Medical Center Institutional Review Board Committee, and was subject had an uninterrupted VF in excess of 308 in all registered at www.clinicaltrials.gov (registration no. meridians. A recordable PR was obtained in both eyes of all NCT01021982). Informed written consent was obtained from patients except for two, both of whom had difficulty in fixating all participants. Twenty eyes of 12 normal healthy age-matched on most fixation locations in one eye. (P ¼ 0.067 compared with patients) volunteers (six males, six females; mean 6 SD age: 38 6 14.4 years; range: 25–65 years) The subjects were requested to blink several times before were included in the study. Four participants could not have the start of the recording and refrain from blinking during the both eyes tested. Inclusion criteria were normal eye examina- recording. Real-time video imaging of the eye was carefully tion, best-corrected visual acuity (BCVA) of 20/20, normal monitored by the examiner during the test. Tests in which color vision test (Roth-28-hue test), no history of past or the subject blinked were excluded and the subject was present ocular disease, no use of any topical or systemic retested. medications that could adversely influence efferent pupil movements, and normal 24-2 Swedish Interactive Threshold Analysis of Pupillary Responses Algorithm (SITA), developed for the Humphrey standard perimeter (Humphrey Field Analyser II, SITA 24-2; Carl Zeiss Percentage pupil contraction at each time point was deter- Meditec, Inc., Jena, Germany). mined by the formula: % pupil contraction ¼ 100 3 [The difference between the highest initial diameter at the The study group (eight males and nine females; mean 6 SD beginning of the stimulus and the lowest diameter in response age: 48.8 6 15.5 years; range: 27–72 years) comprised 30 eyes to that stimulus]/[The highest initial pupil diameter], as of 16 patients with RP and two eyes of a patient with cone–rod described by Kardon et al.27 Previous studies demonstrated dystrophy. Inclusion criteria for RP patients were typical that contraction of the pupil is a true PR when the initial abnormal fundus appearance and a previously recorded ERG pupillary contraction (time at which the maximum accelera- that was abnormal under scotopic or photopic conditions or tion occurs) falls within a definite time window (200–450 ms both (in compliance with the protocol of the International after stimulus onset).13,27 Accordingly, we recorded pupillary Society for Clinical Electrophysiology of Vision, which specifies contraction only when the initial pupillary contraction was the absence or diminution of b-wave amplitude below the fifth within this time window. Figure 1 shows an example of a pupil percentile with prolonged implicit times compared with tracing from a normal subject. All calculations were done by an normal participants).33 Exclusion criteria were a concurrent independent experienced masked technician. The test dura- ocular disease and any other condition affecting the PR. Data tion was approximately 5 minutes for each eye. In preliminary recorded for all patients included sex, diagnosis or genetic studies we repeated each measurement twice in normal defect if known, and ERG responses. participants and found no significant difference between repeated measurements (P > 0.05, n ¼ 14). Furthermore, in all perimetry locations the PR did not significantly differ Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 59

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2763 FIGURE 1. (A) The computerized chromatic multifocal pupillometer. (B) An infrared video image of a pupil while being recorded with the pupil tracking system. (C) Thirteen different locations in the visual field were tested: central, superior, inferior, temporal, and nasal fields at angles of 108, 208, and 308. (D) An example of pupil recordings from the chromatic multifocal pupillometer. The x-axis presents the time scale in milliseconds and the y-axis records the pupil diameter in millimeters. between the left and right eye (P > 0.16, n ¼ 7) in the healthy controls for all perimetry locations using a one-way ANOVA. participant. Agreement between the chromatic pupillometer recordings and the dark-adapted chromatic Goldmann (that yields a yes/no Chromatic Dark-Adapted Visual Field result) was assessed using two-sample t-test and the Mann– Whitney nonparametric test. A value of P < 0.05 was Eleven patients were tested for kinetic VF by dark-adapted considered statistically significant. chromatic Goldmann perimetry.32 Briefly, a Goldmann perim- eter (940-ST; Haag-Streit AG, Liebefeld, Switzerland) was used RESULTS to map patient’s conventional and two-color dark-adapted VFs. Patients were dark adapted for 30 minutes prior to testing. The All participants easily tolerated the protocol without any setting used for stimuli were V3c for the long-wavelength discomfort. PR to the short-wavelength stimulus significantly stimulus and 2 log units lower in luminance (V3c) for the exceeded the PR to the long-wavelength stimulus at all short-wavelength stimulus. perimetry locations (P < 0.01). Thus, the mean percentage pupil contraction in the normal participants in response to the Statistics short-wavelength stimulus was (mean 6 SE) 28.6 6 0.38%, and only 14.7 6 0.41% in response to the long-wavelength stimulus Statistical analysis was performed using a commercial software (Figs. 2A, 2B). program (SAS for Windows, version 9.2; SAS Institute, Inc., Cary, NC; or SPSS for Windows, version 20.0; SPSS, Inc., When both eyes were included in the analysis (repeated- Chicago, IL). For two-eye analysis, comparison between measures ANOVA), no significant differences in mean PR to the patients and healthy controls for all perimetry locations was long-wavelength stimulus were observed between RP patients performed using a one-way ANOVA with repeated measures and normal participants in the majority of locations (P > 0.05), (eye side).34 Since for some of the participants only one eye except nasal at 208, temporal at 308, and superior at 308 (P � was examined, the mixed model was applied to address this 0.02, Fig. 2A). Similar results were obtained when the pupillary issue. For single-eye analysis, we compared between the responses of one eye of patients and normal participants were pupillary recordings of the right eye of patients and healthy compared. Thus, the mean PR to the long-wavelength stimulus Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 60

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2764 FIGURE 2. Percentage change in pupil diameter in both eyes of RP FIGURE 3. Percentage change in pupil diameter in the right eye of RP patients and normal participants in response to both long-wavelength patients and normal participants in response to both long-wavelength (A) and short-wavelength stimuli (B). Data are presented as the mean (A) and short-wavelength stimuli (B). Data are presented as the mean 6 SE of 20 eyes of normal participants and 30 eyes of RP patients. 6 SE of 10 eyes of normal participants and 16 eyes of RP patients. Visual field locations are marked as Central (Ctr), Nasal (N), Temporal Visual field locations are marked as described in Figure 2. (T), Inferior (I), and Superior (S) at varying angles (108, 208, 308). VF testing, the individual reports on three patients, two with in RP patients was significantly lower compared with normal RP and one with cone–rod dystrophy, are presented here. participants in only two locations (nasal 208 and temporal 308, P � 0.04, Fig. 3A). RP Patient #1 In contrast, the mean PR to the short-wavelength stimulus A 68-year-old white male with isolated RP had a BCVA of 20/30 exhibited by the RP patients was significantly lower than the in both eyes. Fundus examination demonstrated pigment PR of control participants in all locations (P < 0.03), except epithelial atrophy, arteriolar narrowing, and bone spicules temporal 108, when both eyes were analyzed (Fig. 2B). When and clumps in the midperiphery of both eyes (the right eye is only one eye was included in the analysis, the mean PR to the shown in Fig. 4A). This patient had significant ERG and VF loss short-wavelength stimulus exhibited by the RP patients was (Table and Fig. 4B). The PR of the right eye of this patient to significantly lower than the PR of control participants in a both the long- and the short-wavelength stimuli in a majority of majority of locations (P � 0.011) except temporal 108 and VF locations, was lower by over 3-fold compared with the superior at 108 and 208 (Fig. 3B). The lowest responses by RP mean PR of the right eye in the normal group (Fig. 4C). In most patients were consistently recorded in peripheral locations of the tested perimetry locations there was an agreement (208 and 308). between the pupillometer-based perimetry and the chromatic Goldmann perimetry. Thus, the pupillometer-based perimetry To validate the novel chromatic pupillometer-based perim- showed that the PR to the long-wavelength stimulus was etry technique we compared the chromatic pupillometer highest at temporal 208 and superior 308 (Fig. 4C). These recordings of 11 patients with their chromatic dark-adapted locations corresponded to the areas where the long-wave- Goldmann findings. In a majority of locations the chromatic length stimulus was detectable by the chromatic Goldmann pupillometer recordings of PR to short-wavelength stimulus perimetry (Fig. 4B). In areas where the long-wavelength were in agreement with the chromatic Goldmann findings both stimulus was not detected by the chromatic Goldmann (108 when single-eye and two-eye analyses were performed (P < superior, and all inferior and nasal locations), the PR response 0.05; Supplementary Table S4). By contrast, in a majority of was lower than 15% of mean normal values. Similarly, the locations there was no significant correlation between the PR highest pupillary responses to the short-wavelength stimulus to long-wavelength stimulus and the chromatic Goldmann were recorded at 308 superior and at 108 and 208 temporal, in recordings (Supplementary Table S3). In all patients tested, agreement with the areas where the short-wavelength stimulus minimal pupillary responses were recorded in areas that were was detected by chromatic Goldmann (Figs. 4B, 4C). The nondetected in the dark-adapted chromatic Goldmann (Sup- lowest PR to the short-wavelength stimulus (<10% of normal plementary Tables S1–S4). To illustrate the pattern of recorded PR values compared with the results of chromatic Goldmann Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 61

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2765 FIGURE 4. A 68-year-old white male with isolated RP. (A) Color photograph of the right eye shows an abnormal fundus with retinal pigment epithelial atrophy, arteriolar narrowing, and bone spicules. (B) Chromatic dark-adapted Goldmann perimetry of the right eye. (C) Comparison of PR of the patient’s right eye in response to both short- and long-wavelength stimuli, as a percentage of mean normal values. Visual field locations are marked as described in Figure 2. mean values) was obtained in locations that were not detected the chromatic Goldmann perimetry. Thus, the PR to the long- by the chromatic Goldmann (center and all nasal locations). wavelength stimulus (Fig. 5C) was 75% or higher of the mean Lower correspondence was observed in areas that were on the response of the right eye in normal participants at the center, isopters of the chromatic Goldmann VF (temporal 108 in the at 108 and 208 in all locations, and at 308 nasal. These locations long-wavelength and superior 208 for both wavelength stimuli). corresponded to areas where the long-wavelength stimulus was detectable on chromatic Goldmann perimetry (Fig. 5B). By RP Patient #2 contrast, in areas where the long-wavelength stimulus was A 37-year-old white male with autosomal dominant RP had a nondetectable by the chromatic Goldmann or close to the BCVA of 20/20 in both eyes. Fundus examination demonstrated isopter (308 temporal, inferior, and superior), minimal pupil- pigment epithelial atrophy and bone spicules in his right eye lary responses were recorded (54%, 60%, and 28% of mean (Fig. 5A). ERG responses were abnormal (Table) and chromatic normal values, respectively). Similar agreement was observed Goldmann perimetry demonstrated characteristic constriction in the PR to the short-wavelength stimulus. In areas where the (Fig. 5B). In most of the locations tested there was an short-wavelength stimulus was detectable by chromatic Gold- agreement between the pupillometer-based perimetry and mann perimetry (at the center, all nasal locations, inferior 108 and 208, and superior 108 and 208), the recorded PR was Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 62

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2766 TABLE. Demographic, Genetics, Visual Acuity, and ERG Findings for Patients Included in the Study Sex Age Origin Age at Genetics Visual Acuity ERG-Photopic,* lV ERG-Scotopic,† lV Diagnosis RE LE RE LE RE LE 1 M 68 Austria/ Spain 53 Sporadic 20/30 20/30 11 14 89 18 24 2 M 37 Lithuania 30 AD 20/25 ND 18 14 49 57 46 20 17 3 M 58 Iran 57 AR 20/60 20/60 11 151 162 7 90 56 4 M 40 Poland 5 AR 20/60 20/60 47 5 ND 5 F 69 Turkey/Greece 43 AR 20/100 ND 19 21 05 80 86 6 F 34 Iraq 10 AR 20/20 20/30 ND 0 00 10 13 14 7 M 72 Poland 53 X-linked RP 0.5/36 1/18 2 45 0 00 8 F 60 Greece 50 AR 20/40 HM 31 0 84 7 00 9 F 64 Russia 20 AD 20/800 20/800 0 0 0 10 12 157 185 10 F 31 Russia 29 AD 20/40 20/20 13 21 219 207 27 11 M 27 Iraq/ Morocco 2 Sporadic 20/50 20/100 5 12 M 55 Iraq 11 AD 20/63 20/40 0 13 F 32 Iraq 2 AR 20/32 20/25 3 14 F 33 Uzbekistan 2 AR 20/25 20/20 0 15 F 64 Morocco 30 AR 20/100 20/50 7 16 M 44 Turkey 39 AR 20/800 20/800 19 17 M 42 Syria/Egypt 19 AR cone rod 20/100 20/100 30 AD, autosomal dominant; AR, autosomal recessive; RE, right eye; LE, left eye; ND, not determined. * Mean normal values: 144 lV. † Mean normal values: 507 lV. maximal (>66% of mean normal values). The lowest pupillary Goldmann (all nasal locations), minimal PR was recorded responses were recorded in areas where the short-wavelength (<37% of mean normal values). Lower correlation was stimulus was not detectable on chromatic Goldmann perimetry observed in 108 superior that is on the isopter and at 308 (temporal 208 and 308, superior 308, and inferior 308). Lower temporal for the short-wavelength stimulus. correspondence was observed in nasal 308 for the short- wavelength stimulus, which was on the isopter of the DISCUSSION chromatic Goldmann VF. In this study we successfully used the PR to multifocal Cone–Rod Dystrophy Patient #17 chromatic stimulus as an objective mean to perform perimetry. RP patients exhibited significantly reduced mean PR to the A 42-year-old male with autosomal recessive cone–rod short-wavelength stimulus at most locations, whereas their dystrophy had a BCVA of 20/100 at both eyes. Fundus mean responses to the long-wavelength stimulus were similar examination demonstrated macular pigment epithelial atrophy to those of normal participants at central locations but were (the right eye is shown in Fig. 6A) and chromatic Goldmann significantly reduced at peripheral locations. Studies from the perimetry revealed a dense central scotoma and temporal groups of Kardon27–29 and Stieger30 demonstrated that parafoveal shifting of fixation (Fig. 6B). ERG responses were transient PR to a low-intensity, short-wavelength stimulus abnormally reduced, especially under conditions of light reflects rod activity and that transient PR to a long-wavelength adaptation (Table). In his responses to both the short- and stimulus is predominantly driven by cones. In retinitis the long-wavelength stimuli, this patient demonstrated an pigmentosa patients, the loss of rod function exceeds the agreement in most perimetry locations between the pupil- reduction of cone function and VF loss typically begins with lometer-based perimetry and the chromatic Goldmann perim- peripheral VF constriction.35,36 These findings suggest that the etry. Thus, the PR to the long-wavelength stimulus (Fig. 6C) PR to short-wavelength stimulus measured by our chromatic was highest (>58% of mean normal values) in areas where the pupillometer may be mediated by rods, whereas the PR to long-wavelength stimuli were detectable by chromatic Gold- long-wavelength stimulus measured here may be mediated by mann perimetry (center, 108 and 308 temporal, 108 inferior, cones. Future studies will be aimed at determining a clinically and 208 superior). The PR to the long-wavelength stimulus was applicable protocol for assessing these cell contributions to the minimal (<34% of mean normal values) in areas that were PR measured by our chromatic pupillometer. This will enable nondetectable on the chromatic Goldmann (208 and 308 both the objective detection of affected areas and the nasal). Lower correlation was observed in 108 nasal and 208 identification of the damaged photoreceptor cells underlying temporal. The 308 superior and inferior locations in this the defect in these locations. patient were not measurable, probably because of his difficulty in using a parafoveal fixation on peripheral inferior When both eyes were analyzed, the only measurement in and posterior targets. An agreement between the two which the PR to short-wavelength stimulus in RP patients was perimetry tests was also demonstrated in response to the not significantly lower than that in normal participants was at short-wavelength stimulus (Fig. 6C). In areas where the short- the temporal 108 location. This might be explained by recent wavelength stimulus was detectable by the chromatic Gold- optical coherence tomographic findings in RP patients mann (center, 108 and 208 temporal, 108 and 208 inferior, and showing increased outer macular thickness in the nasal 308superior), the patient exhibited maximal PR (>48% of quadrant, which corresponds to the temporal 108 location.37 mean normal values), whereas in areas where the short- wavelength stimulus was undetectable in the chromatic In cone–rod dystrophy, the deficit in the cones far exceeds that in rods. In the cone–rod dystrophy patient described here, the decline in PR to both the long- and the short-wavelength Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 63

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2767 FIGURE 5. A 37-year-old white male with autosomal dominant RP. (A) Color photograph of the right eye shows an abnormal fundus with retinal pigment epithelial atrophy and bone spicules. (B) Chromatic Goldmann perimetry of the right eye. (C) Comparison of PR of the patient’s right eye in response to both short- and long-wavelength stimuli, as a percentage of mean normal values. Visual field locations are marked as described in Figure 2. stimuli was similar in the scotoma area, unlike our RP patients, threshold-subjective kinetic test that yields a binary qualitative in whom the decline in response to the short-wavelength recording (yes/no result). By contrast the chromatic pupillom- stimulus was more pronounced, suggesting that the new eter is a suprathreshold, objective quantitative test. Hence, it is perimetry method may assist in diagnosis of diseases affecting likely that the pupillometer can detect reduced retinal function different retinal cells. to suprathreshold stimuli in a numeric manner. Since the response to the long-wavelength stimulus was less affected in To validate the new chromatic perimetry technique we RP patients, this difference between the two systems was more decided to compare the novel pupillometer method with an profound. In some cases the lower agreement could be established perimetry technique. We chose the chromatic dark- explained by PR measurements in areas corresponding to the adapted Goldmann because it uses multifocal chromatic short- chromatic Goldmann VF isopters. and long-wavelength stimuli for perimetry determination and monitoring of patients with retinal dystrophies.32,38 We Our findings that similar results were obtained using two- observed a good agreement between the chromatic pupillom- eye and single-eye analyses provide further evidence for the eter-based perimetry and the malfunctioning areas identified validity of the new perimetry method. Minor differences by the dark-adapted chromatic Goldmann perimetry. Further- between the results of the two analyses could be explained by more, a correlation was observed between the two methods in the smaller sample size using a single-eye analysis. Although a majority of locations in response to the short-wavelength the results presented here are highly promising for use of the stimulus. Our findings that the PR to a long-wavelength multifocal chromatic PR as an objective parameter of retinal stimulus did not correlate in a majority of locations with the function in disease conditions, further research is needed to chromatic dark-adapted Goldmann, could be explained by the improve and refine this technique. The prototype instrument differences between the two methods: the Goldmann is a used in this study was constructed for proof of concept and Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 64

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2768 FIGURE 6. A 42-year-old male with autosomal recessive cone–rod dystrophy. (A) Color photograph of the right eye shows an abnormal fundus with macular pigment epithelial atrophy. (B) Chromatic Goldmann perimetry of the right eye. (C) Comparison of PR of the patient’s right eye in response to both short- and long-wavelength stimuli, as a percentage of mean normal values. Visual field locations are marked as described in Figure 2. uses only a single central stimulus. Thus, some minimal mRGCs. However, since transient PR was recorded, the cooperation on the part of the participants was still needed contribution of rods probably exceeded that of mRGCs. for fixation on targets. In some cases, the need for a patient to Furthermore, the pathology of RP patients is caused primarily fixate on a location correlating with a scotoma could explain a by degeneration of the photoreceptors, and the defect in reduced correspondence between the pupillometry-based mRGCs is less significant,35 suggesting that the difference in PR perimetry and the Goldmann perimetry findings. For example, to the short-wavelength stimulus between the RP patients and in the cone–rod dystrophy patient, who had difficulty fixating the normal participants was largely due to rod degeneration. In at central locations because of a central scotoma, we found an attempt to discriminate between responses of mRGCs and relatively less correspondence between the Goldmann and the the rods, we are currently testing the PR to short-wavelength pupillometry-based test results, specifically in more peripheral stimuli at different intensities. Once a protocol for differential locations. We believe that this reduced correspondence may be cell-type contribution to PR is established, this chromatic due to the central location of the stimulus and the limited pupillometer will enable determination of the functionality of ability of this patient to fixate on peripheral VF locations using inner and outer retinal cells at different locations in the retina. parafoveal fixation. This limitation is likely to be overcome in a Areas with nonfunctional photoreceptors and some functional future design of the instrument, where participants will be mRGCs may be more suitable candidates for cell-based or asked to look forward and stimuli will be individually genetic therapies. In future studies we will also examine the introduced at different VF locations. A second limitation of correlation between age and severity of VF loss in a larger the current study was the use of the short-wavelength stimulus group of patients as well as determine the appropriate at 40 cd/m2. Based on the findings of Kardon et al.,27 this light parameters for testing the PR of dichromats. In addition, use intensity can exert a pupillary response both in rods and in of a smaller spot size will be investigated, with the aim of Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 65

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2769 achieving better perimetric resolution. Taken together, this 14. Aoyama T. Pupillographic perimetry. The application to novel chromatic pupillometer may enable sensitive and clinical cases (author’s transl) [in Japanese]. Nihon Ganka objective characterization of VF defects and may be used to Gakkai Zasshi. 1977;81:1527–1538. diagnose and monitor patients with photoreceptor dystrophies in the upcoming clinical trials. 15. Reuther R, Krastel H, Alexandridis E. Disturbances of the pupil reflex associated with homonymous hemianopic paracentral Acknowledgments scotoma (author’s transl) [in German]. Arch Psychiatr Nervenkr. 1981;229:259–266. The authors thank Esther L. Shabtai (MSc, from the Statistics Unit of Tel Aviv University, Israel, who received financial compensation 16. Trimarchi F, Bianchi PE, Gelmi C, Franchini F, Baldini M. for the statistical analysis) for the statistical analysis and Shirley Diagnostic value of objective campimetry. J Neurol. 1981;225: Smith for scientific English editing (who received financial 167–173. compensation for the work). 17. Alexandridis E, Krastel H. Pupillographic perimetry using the Supported in part by a Claire and Amade Martier fund research ‘‘octopus.’’ Fortschr Ophthalmol. 1989;86:692–694. grant (YR) and in part by a grant from the Israeli Ministry of Immigrant Absorption (IS). Ygal Rotenstreich has submitted a 18. Wilhelm H. The pupil and retrogeniculate visual pathway: patent application through Sheba Medical Center (patent applica- overview. Ophthalmologe. 1996;93:319–324. tion #PCT/L2010/00624). 19. Yoshitomi T, Matsui T, Tanakadate A, Ishikawa S. Comparison Disclosure: A. Skaat, None; I. Sher, None; A. Kolker, None; S. of threshold visual perimetry and objective pupil perimetry in Elyasiv, None; E. Rosenfeld, None; M. Mhajna, None; S. clinical patients. J Neuroophthalmol. 1999;19:89–99. Melamed, None; M. Belkin, None; Y. Rotenstreich, P 20. Kardon RH, Kirkali PA, Thompson HS. Automated pupil References perimetry. Pupil field mapping in patients and normal subjects. Ophthalmology. 1991;98:485–495; discussion 495– 1. Sharma P, Sample PA, Zangwill LM, Schuman JS. Diagnostic 496. tools for glaucoma detection and management. Surv Oph- thalmol. 2008;53(suppl 1);S17–S32. 21. Schmid R, Luedtke H, Wilhelm BJ, Wilhelm H. Pupil campimetry in patients with visual field loss. Eur J Neurol. 2. Turalba AV, Grosskreutz C. A review of current technology 2005;12:602–608. used in evaluating visual function in glaucoma. Semin Ophthalmol. 2010;25:309–316. 22. Skorkovska K, Wilhelm H, Ludtke H, Wilhelm B. How sensitive is pupil campimetry in hemifield loss? Graefes Arch Clin Exp 3. Young WO, Stewart WC, Hunt H, Crosswell H. Static threshold Ophthalmol. 2009;247:947–953. variability in the peripheral visual field in normal subjects. Graefes Arch Clin Exp Ophthalmol. 1990;228:454–457. 23. Jensen W. A description of a method for objective perimetry [in German]. Albrecht Von Graefes Arch Klin Exp Ophthal- 4. Heijl A, Lindgren A, Lindgren G. Test-retest variability in glauco- mol. 1976;201:183–191. matous visual fields. Am J Ophthalmol. 1989;108:130–135. 24. Rosli Y, Bedford SM, James AC, Maddess T. Photopic and 5. Blumenthal EZ, Sample PA, Berry CC, et al. Evaluating several scotopic multifocal pupillographic responses in age-related sources of variability for standard and SWAP visual fields in macular degeneration. Vision Res. 2012;69:42–48. glaucoma patients, suspects, and normals. Ophthalmology. 2003;110:1895–1902. 25. Carle CF, James AC, Kolic M, Loh YW, Maddess T. High- resolution multifocal pupillographic objective perimetry in 6. Chauhan BC, Garway-Heath DF, Goni FJ, et al. Practical glaucoma. Invest Ophthalmol Vis Sci. 2011;52:604–610. recommendations for measuring rates of visual field change in glaucoma. Br J Ophthalmol. 2008;92:569–573. 26. Maddess T, Ho YL, Wong SS, et al. Multifocal pupillographic perimetry with white and colored stimuli. J Glaucoma. 2011; 7. Collaborative Normal-Tension Glaucoma Study Group. Com- 20:336–343. parison of glaucomatous progression between untreated patients with normal-tension glaucoma and patients with 27. Kardon R, Anderson SC, Damarjian TG, Grace EM, Stone E, therapeutically reduced intraocular pressures. Am J Ophthal- Kawasaki A. Chromatic pupil responses: preferential activa- mol. 1998;126:487–497. tion of the melanopsin-mediated versus outer photoreceptor- mediated pupil light reflex. Ophthalmology. 2009;116:1564– 8. Advanced Glaucoma Intervention Study. 2. Visual field test 1573. scoring and reliability. Ophthalmology. 1994;101:1445–1455. 28. Kardon R, Anderson SC, Damarjian TG, Grace EM, Stone E, 9. Cibis GW, Campos EC, Aulhorn E. Pupillary hemiakinesia in Kawasaki A. Chromatic pupillometry in patients with retinitis suprageniculate lesions. Arch Ophthalmol. 1975;93:1322– pigmentosa. Ophthalmology. 2011;118:376–381. 1327. 29. Park JC, Moura AL, Raza AS, Rhee DW, Kardon RH, Hood DC. 10. Grozdanic SD, Matic M, Sakaguchi DS, Kardon RH. Evaluation Toward a clinical protocol for assessing rod, cone, and of retinal status using chromatic pupil light reflex activity in melanopsin contributions to the human pupil response. healthy and diseased canine eyes. Invest Ophthalmol Vis Sci. Invest Ophthalmol Vis Sci. 2011;52:6624–6635. 2007;48:5178–5183. 30. Lorenz B, Strohmayr E, Zahn S, et al. Chromatic pupillometry 11. Skorkovska K, Ludtke H, Wilhelm H, Wilhelm B. Pupil dissects function of the three different light-sensitive retinal campimetry in patients with retinitis pigmentosa and func- cell populations in RPE65 deficiency. Invest Ophthalmol Vis tional visual field loss. Graefes Arch Clin Exp Ophthalmol. Sci. 2012;53:5641–5652. 2009;247:847–853. 31. Kawasaki A, Crippa SV, Kardon R, Leon L, Hamel C. 12. Asakawa K, Shoji N, Ishikawa H, Shimizu K. New approach for Characterization of pupil responses to blue and red light the glaucoma detection with pupil perimetry. Clin Ophthal- stimuli in autosomal dominant retinitis pigmentosa due to mol. 2010;4:617–623. NR2E3 mutation. Invest Ophthalmol Vis Sci. 2012;53:5562– 5569. 13. Hong S, Narkiewicz J, Kardon RH. Comparison of pupil perimetry and visual perimetry in normal eyes: decibel 32. Rotenstreich Y, Fishman GA, Lindeman M, Alexander KR. The sensitivity and variability. Invest Ophthalmol Vis Sci. 2001; application of chromatic dark-adapted kinetic perimetry to 42:957–965. retinal diseases. Ophthalmology. 2004;111:1222–1227. 33. Marmor MF, Fulton AB, Holder GE, et al. ISCEV standard for full-field clinical electroretinography (2008 update). Doc Ophthalmol. 2009;118:69–77. Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 66

Pupillometer-Based Objective Chromatic Perimetry IOVS j April 2013 j Vol. 54 j No. 4 j 2770 34. Armstrong RA. Statistical guidelines for the analysis of data 37. Garcia-Martin E, Pinilla I, Sancho E, et al. Optical obtained from one or both eyes. Ophthalmic Physiol Opt. coherence tomography in retinitis pigmentosa: reproduc- 2013;33:7–14. ibility and capacity to detect macular and retinal nerve fiber layer thickness alterations. Retina. 2012;32:1581– 35. Hartong DT, Berson EL, Dryja TP. Retinitis pigmentosa. Lancet. 1591. 2006;368:1795–1809. 38. Rotenstreich Y, Belkin M, Sadetzki S, et al. A randomized 36. Grover S, Fishman GA, Brown J Jr. Patterns of visual field crossover trial of 9-cis b-carotene rich powder in retinitis progression in patients with retinitis pigmentosa. Ophthal- pigmentosa patients. JAMA Ophthalmol. In press. mology. 1998;105:1069–1075. Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933467/ on 05/13/2015 Terms of Use: 67



‫‪16/6-idiotype expressing antibodies induce brain inflammation‬‬ ‫‪and cognitive impairment in mice: the mosaic of central nervous‬‬ ‫‪system involvement in lupus‬‬ ‫‪BMC Medicine | 2013‬‬ ‫חברי ח\"ץ היקרים‪,‬‬ ‫מנחה‪ :‬פרופ' יהודה שינפלד‬ ‫כמי שמלווה את מחקרכם במשך שנים ורואה אח\"כ את תוצאות הקריירה‬ ‫המרכז למחלות אוטואימוניות‬ ‫שלכם‪ ,‬אני מודע לתרומה העצומה והשפעת המחקר עליכם‪.‬‬ ‫נשיא אוניברסיטת אריאל‬ ‫אומר בקצרה‪ :‬רופא חוקר הוא רופא יותר טוב כי הוא תמיד שואל למה‪.‬‬ ‫‪[email protected]‬‬ ‫מאחל לכם הצלחה גדולה הן בחלק הקליני והן בחלק המחקרי‪.‬‬ ‫אשמח לפגוש אתכם בעתיד‪.‬‬ ‫בהצלחה‪,‬‬ ‫שלכם פרופ' יהודה שינפלד‬ ‫פרויקט ח״ץ חשף אותי לעבודה מחקרית בעיקר מחקר בסיסי אך לא רק‪.‬‬ ‫מורן לנדאו רבי‬ ‫עבדתי עם צוות מתחומי הביולוגיה וכן מתחומי הרפואה השונים‪ .‬הצטרפתי‬ ‫אונ' תל אביב‬ ‫לישיבות מחקר וכנסים‪ .‬יצרתי קשרים שעודם קיימים‪.‬‬ ‫השתתפה כסטודנטית בפרויקט ח״ץ‬ ‫זכיתי להיות תלמידתו של פרופ׳ שיינפלד ועוד רבים וטובים מהמכון‬ ‫בין השנים ‪2010-2011‬‬ ‫האוטואימוני‪.‬‬ ‫‪[email protected]‬‬ ‫בעיני מדובר בחוויה מעשירה ומלמדת רבות‪.‬‬ ‫‪69‬‬

Kivity et al. BMC Medicine 2013, 11:90 http://www.biomedcentral.com/1741-7015/11/90 RESEARCH ARTICLE Open Access 16/6-idiotype expressing antibodies induce brain inflammation and cognitive impairment in mice: the mosaic of central nervous system involvement in lupus Shaye Kivity1,2*†, Aviva Katzav1,3†, Maria Teresa Arango1,4,5, Moran Landau-Rabi1, Yaron Zafrir1, Nancy Agmon-Levin1, Miri Blank1, Juan-Manuel Anaya4, Edna Mozes6, Joab Chapman1,3 and Yehuda Shoenfeld1,7 Abstract Background: The 16/6-idiotype (16/6-Id) of the human anti-DNA antibody was found to induce experimental lupus in naïve mice, manifested by production of autoantibodies, leukopenia and elevated inflammatory markers, as well as kidney and brain involvement. We assessed behavior and brain pathology of naive mice injected intra- cerebra-ventricularly (ICV) with the 16/6-Id antibody. Methods: C3H female mice were injected ICV to the right hemisphere with the human 16/6-Id antibody or commercial human IgG antibodies (control). The mice were tested for depression by the forced swimming test (FST), locomotor and explorative activity by the staircase test, and cognitive functions were examined by the novel object recognition and Y-maze tests. Brain slices were stained for inflammatory processes. Results: 16/6-Id injected mice were cognitively impaired as shown by significant differences in the preference for a new object in the novel object recognition test compared to controls (P = 0.012). Similarly, the preference for spatial novelty in the Y-maze test was significantly higher in the control group compared to the 16/6-Id-injected mice (42% vs. 9%, respectively, P = 0.065). Depression–like behavior and locomotor activity were not significantly different between the16/6-Id-injected and the control mice. Immunohistochemistry analysis revealed an increase in astrocytes and microglial activation in the hippocampus and amygdala, in the 16/6-Id injected group compared to the control. Conclusions: Passive transfer of 16/6-Id antibodies directly into mice brain resulted in cognitive impairments and histological evidence for brain inflammation. These findings shed additional light on the diverse mosaic pathophysiology of neuropsychiatric lupus. Keywords: Systemic lupus erythematosus, 16/6 idiotype, Anti-DNA, Neuropsychiatric lupus, Cognitive impairment Background proposed a standard nomenclature of case definitions, Neuropsychiatric systemic lupus erythematosus (NPSLE) reporting standards and diagnostic testing recommen- refers to a complex set of syndromes involving the central dations for the 19 neuropsychiatric Systemic lupus nervous system (CNS) in up to 56% of lupus patients [1-5]. erythematosus (SLE) syndromes [6]. While some of the Due to the varied diagnostic criteria applied to define focal manifestations (for example, stroke) can be explained NPSLE, the American College of Rheumatology has by vasculitic or thrombotic lesions, the pathogenicity of more diffuse manifestations of NPSLE (for example, cogni- * Correspondence: [email protected] tive impairment, depression and psychosis) remains rela- †Equal contributors tively obscure. Nevertheless, studies have demonstrated the 1The Zabludovicz Center for Autoimmune Diseases, Sheba Medical Center, importance of various factors involved in the development Tel-Hashomer, 2 Derech Sheba St., Ramat Gan 52621, Israel of diffuse neuropsychiatric manifestations, such as the 2Rheumatic Disease Unit, Sheba Medical Center, Tel-Hashomer, 2 Derech presence of autoantibodies, inflammatory mediators (for Sheba St., Ramat Gan 52621, Israel Full list of author information is available at the end of the article © 2013 Kivity et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 70

Kivity et al. BMC Medicine 2013, 11:90 Page 2 of 9 http://www.biomedcentral.com/1741-7015/11/90 example, cytokines, matrix metalloproteinases), neuropep- in the animal facility at Sheba Medical Center. The mice tides and endocrine factors [7-10]. Other factors, such as were raised under standard conditions, 23 ± 1°C, 12- medications and primary neurologic and psychiatric disor- hour light cycle (6:30 am to 6:30 pm) with ad libitum ders, may play a major role as well. access to food and water. The Sheba Medical Center Animal Welfare Committee approved all procedures. More than 20 brain specific and non-specific autoanti- bodies have been proposed to be involved in the Monoclonal 16/6-Id expressing antibodies mechanism of NPLSE [11], including anti-neuronal [12], The human monoclonal anti–DNA antibodies were anti-ribosomal–P [13,14], anti-phospholipid [15] anti- produced by a hybridoma derived from fusion of the bodies, as well as anti NR2/anti-DNA antibodies that GM4672 lymphoblastoid cell line and peripheral blood cross react with N-methyl-D-aspartate (NMDA) recep- or splenic lymphocytes obtained from three lupus pa- tors [3,16]. During the last two decades, anti-DNA tients. The human mAb that bears the 16/6-Id (IgG1/k) idiotypes were characterized, and found to play an im- has been characterized previously [33]. The mAb was se- portant role in systemic lupus erythematosus and NPSLE creted by hybridoma cells that were grown in culture [17]. The 16/6 idiotype (Id) antibody is a human anti- and were purified by using a protein G-sepharose col- single-stranded-DNA (anti-ssDNA) monoclonal antibody umn (Pharmacia, Fine Chemicals, Uppsala, Sweden). (mAb) originated from a patient with cold agglutinin dis- ease [18]. The 16/6-Id was found to be polyspecific [19], The injection process is based on a detailed protocol cross reacting with cytoskeletal proteins (vimentin), reported by Shoenfeld et al. [34]. Mice were anesthe- platelets, lymphocyte membranes, pathogens such as tized by intra-peritoneal (i.p.) injection of ketamine Klebsiela polysaccharides and Mycobacterium tuberculosis (100 mg/kg) and xylazine (20 mg/kg). The skull was glycoproteins, brain glycolipids and tumor cells [20-22]. carefully exposed, and a small hole was drilled with a The presence of 16/6-Id was detected in 30% of lupus 25-gauge needle above the right lateral ventricle (2 patients, and their levels were found to correlate with mm lateral to the midline and 2.5 mm posterior to the disease activity [23,24]. Elevated titers of 16/6-Id were bregma). A 27-gauge needle attached to a Hamilton syr- also detected in NPSLE patients [25]. Deposits of 16/6-Id inge was inserted at this point to a depth of 2 mm, where were found in the skin, kidney and brain tissue [21,26,27], preliminary tests had confirmed accurate ICV placement and were found to bind human cortical brain tissue by injection of dye. Then 1 μl of anti-DNA 16/6-Id mAb sections ex vivo. The presence of circulating 16/6-Id or control IgG was slowly infused, the needle was with- was detected in patients with other autoimmuine diseases drawn and the skin over the scalp was sutured. All anti- as well (for example, polymyositis, systemic sclerosis) body solutions used contained 6 mg protein/ml. Each [28,29]. Immunization of naïve mice with the human anti- mouse received only a single injection. DNA 16/6-Id mAb was shown to induce experimental lupus manifested both serologically and clinically. A wide Experimental design profile of mice autoantibodies (for example, mouse 16/6- Twenty-one CH3 mice were injected ICV to the right Id, and antibodies against dsDNA, ssDNA, Ro, La, RNP, hemisphere, 11 with human 16/6-Id antibodies and 10 Sm, histones, cardiolipin and phosphatydilserine), were with commercial human IgG antibodies (control). The detected, as well as leukopenia, elevated erythrocyte sedi- forced swimming test (FST) was performed at Days 16 mentation rate (ESR), proteinuria and the deposition of and 23 after antibody injection, the staircase test at Days immunoglobulins in the kidney mesangium [30-32]. In 14 and 26, the novel object recognition at Days 19 and addition, recent-preliminary data showed histological brain 20 and the Y-maze test at Day 21. At Day 24, under changes in mice with experimental SLE induced by active anesthesia, a systemic perfusion was performed, and the immunization with the 16/6-Id (A. Marom and E. Mozes, brains were collected. Immunofluorescence staining was unpublished results). Therefore, we hypothesized that the performed to detect markers of inflammation or neur- 16/6-Ids have a pathogenic role in neuropsychiatric lupus. onal degeneration (see below). In the present study we investigated the effect of 16/6-Id on behavioral and cognitive functions, as well as on the Cognitive and behavioral tests brain pathology of naïve mice injected intra-cerebra Forced swimming test -ventricularily (ICV) with the 16/6-Id. This test is based on Porsolt et al.‘s description [35]. Mice were placed in individual glass beakers (height 39 Methods cm, diameter 21.7 cm) with water 15 cm deep at 25°C. On the first day, mice were placed in the cylinder for a Mice, antibody injection and experimental design pretest session of 15 minutes, and later were removed from the cylinder, and then returned to their home Mice cages. Twenty-four hours later (Day 2), the mice were Three-month-old, female C3H mice were obtained from Harlan Laboratories, Jerusalem, Israel, and were housed 71

Kivity et al. BMC Medicine 2013, 11:90 Page 3 of 9 http://www.biomedcentral.com/1741-7015/11/90 re-exposed to the swimming condition in a similar envir- during training was replaced by a novel object. All ob- onment, and then subjected to a test session for six mi- jects were balanced in terms of physical complexity and nutes. The behavioral measure scored was the duration were emotionally neutral. The box and the objects were (in seconds) of immobility, defined as the absence of thoroughly cleaned by 70% alcohol before each session escape-oriented behaviors, such as swimming, jumping, to avoid possible instinctive odorant cues. A preference rearing, sniffing or diving, recorded during the six- index, a ratio of the amount of time spent exploring any minute test. A depression-like behavior was considered one of the two items (old and new in the retention ses- as an increased immobility time. sion) over the total time spent exploring both objects, was used to measure recognition memory. Individual ani- Staircase test mals demonstrating insufficient task performance were Locomotor and explorative activity was evaluated by the excluded from later specific statistical analyses for the fol- staircase test, as described previously by Katzav et al. lowing reasons: (1) non-exploration, which was defined as [15]. This test analyzes locomotor and exploratory activ- no objection interaction or (2) technical malfunctions dur- ities (stair-climbing) and anxiety (rearing). The staircase ing data collection. maze consisted of a polyvinyl chloride enclosure with five identical steps, 2.5 × 10 × 7.5 cm. The inner height Y maze test of the walls was constant (12.5 cm) along the whole The Y maze test was used to assess spatial memory. It length of the staircase. The box was placed in a room was comprised of three arms, built of black Perspex. with constant lighting and isolated from external noise. Each arm was 8 × 30 × 15 cm at an angle of 120° from Each mouse was tested individually. The animal was the others. One arm was randomly selected as the start placed on the floor of the staircase with its back to the arm. Each mouse was placed twice in the start arm. On staircase. The number of stairs climbed and the number the first trial, lasting for five minutes, one of the other of rears were recorded during a three-minute period. two arms was randomly chosen to be blocked whereas Climbing was defined as each stair on which the mouse on the second trial, lasting for two minutes, both arms placed all four paws; rearing was defined as each in- were open. The two trials were separated by a two- stance the mouse rose on hind legs (to sniff the air), ei- minute interval, during which the mouse was returned ther on the stair or against the wall. The number of to its home cage. The time spent in each of the arms stairs descended was not taken into account. Before each was measured. Between each trial and between each test, the animal was removed and the box cleaned with a mouse, the maze was cleaned with a 70% alcohol solu- diluted alcohol solution to eliminate smells. tion and dried. Discrimination of spatial novelty was assessed by a preference index [37]: time in the new Novel object recognition test arm - time old arm/time in the new arm + time in the This is a visual recognition memory test based on a old arm, assessing spatial memory. The mouse is method described by Tordera et al. [36]. The apparatus, expected to recognize the old arm as old and spend an open field box (50 × 50 × 20 cm), was constructed more time in the new arm. from plywood painted white. Three phases (habituation, training and retention) were conducted on two separate Immunofluorescence staining test days. Before training, mice were individually habitu- Brain perfusion and fixation ated by allowing them to explore the box for one hour. The mice were anesthetized by an i.p. injection of keta- No data were collected at this phase. During training mine (100 mg/kg) and xylazine (20 mg/kg) and sacrificed sessions, two identical objects were placed into the box by transcardiac perfusion with phosphate buffered saline in the northwest and southeast corners (approximately (PBS) followed by perfusion with 4% paraformaldehyde 5 cm from the walls), 20 cm away from each other (sym- (PFA, Sigma-Aldrich Israel Ltd., Rehovot Israel) in phos- metrically) and then the individual animal was allowed phate buffer (PO4, pH 7.4). After perfusion, the brain was to explore for five minutes. Exploration of an object was quickly removed and fixed overnight in 4% PFA (in PO4, defined as directing the nose to the object at a distance pH 7.4) at 4°C. On the following day, the brain was of ≤1 cm and/or touching it with the nose; turning cryoprotected by immersion in 30% sucrose in 0.1M PO4 around or sitting near the object was not considered as (pH 7.4) for 24 to 48 hours at 4°C before brain cutting. exploratory behavior. The time spent in exploring each object was recorded. The animals were returned to their Brain cutting and preservation home cages immediately after training. During the reten- Frozen coronal sections (30 to 50 μm) were cut on a tion test, the animals were placed back into the same sliding microtome (Leica Microsystems GmbH, Wetzlar, box after a four-hour interval, and allowed to explore Germany), collected serially and kept in a cryoprotectant freely for five minutes. One of the familiar objects used at −20°C until staining. Staining was performed as 72

Kivity et al. BMC Medicine 2013, 11:90 Page 4 of 9 http://www.biomedcentral.com/1741-7015/11/90 follows. Six mice (three IgG control and three 16/6 Id) spent near the new object vs. old object, P = 0.655). This were used for immunohistochemistry. Brain sections suggests a specific visual recognition memory impairment were stained free-floating, incubated with the first anti- in the 16/6-Id mice. Similarly, cognitive performance in bodies overnight at 4°C. The slices were then washed in the Y-maze test is presented as a preference index for new PBS + 0.1% Triton X-100, and incubated at room (additional percent time spent in the novel arm) in both temperature for one hour with the corresponding fluor- groups (Figure 2). The control IgG mice spent 46% add- escent chromogens-conjugated secondary antibody. Sec- itional time in the new lane while the mice injected with tions were stained for specific antigens with antibodies 16/6-Id spent 9% additional time in the new lane (P = against activated microglia (anti-Iba1, pAb, Abcam, 0.015 by t-test). Cambridge, UK) and astrocytes (anti-GFAP mAb, Dako, Carpinteria, CA, USA). Counter staining was performed In the forced swimming test there was no significant with Hoechst (Sigma-Aldrich Israel Ltd., Rehovot Israel). difference between 16/6-Id injected and control mice in depression-like behavior at Days 16 and 24 after injec- Statistical analysis tion. Average immobility times of the control mice vs. Results are expressed as the mean ± SEM. The differ- 16/6-Id injected mice were 117.6 ± 65.9 vs. 160 ± 72.8 ences in mean for average immobility time in the FST, (P = 0.159 by t-test) and 182.5 ± 45.4 vs. 205.7 ± 42.7 the staircase test parameters (number of rearing and sec (P = 0.238 by t-test) on Days 16 and 24, respectively. stair-climbing events), novel object recognition and Y- maze tests were evaluated by T-test. Significant results In the staircase test, there was no significant difference were determined as P <0.05. between the average rearing and stair-climbing counts, among mice from control-IgG vs. 16/6-Id (23.7 ± 2.6 vs. Results 21.8 ± 2.5 rearings, and 24.5 ± 2.3 vs. 16.5 ± 4.4 stair- climbing events, respectively, P >0.016). The results also Cognitive and behavioral performance did not change from Day 14 to 26. The results of cognitive performance in the novel object recognition test are presented as the proportion time Brain pathology spent near objects (new and old) in both groups (Figure 1). Brain sections were stained for activated microglia and There was a significant preference for attention to the astrocytes (as markers for inflammation). The 16/6-Id new object in the control group (64% time spent near the injected mice demonstrated increased microglial activa- new object compared to 36% time spent near the old ob- tion (Iba-1 staining), at the hippocampus (CA1, CA3, ject, P = 0.012), while no difference in the preference was dentate gyrus, stratum radiatum) as well as the amyg- seen in the mice injected with 16/6-Id (56% vs. 44% time dala, compared to IgG control (Figure 3). The difference Figure 1 16/6-Id injected mice displayed impaired performance Figure 2 16/6-Id injected mice displayed impaired spatial in the novel object recognition test. Results are presented as the memory in the Y-maze test. Results are presented as the proportion of time spent near the old and new objects by the 16/6- proportion of time (mean ± SEM) spent in the new arm introduced Id (gray bars) and IgG control (black bars) injected mice. The control by the 16/6-Id (gray bars) and IgG control (black bars) injected mice. mice (IgG) significantly preferred the new object (64% vs. 36% for In the figure it is shown that the control group (IgG injected) spent the proportion time near the new vs. old objects respectively; P = more time in the new lane as compared to the 16/6 injected group. 0.01), while the 16/6-Id injected mice had no significant preference They have recognized the old lane as known and preferred to either objects (56% vs. 44% new vs. old; P = 0.5). Results exploring the new lane, which means that their spatial memory is presented as mean ± SEM. * Statistically significant (P <0.05). conserved. There was a significant difference in additional time spent in the new lane between the 16/6 and IgG group (0.46 vs. 0.09, P = 0.02 respectively). * Statistically significant (P <0.05). 73

Kivity et al. BMC Medicine 2013, 11:90 Page 5 of 9 http://www.biomedcentral.com/1741-7015/11/90 Figure 3 Increased brain inflammation (activated microglia) in 16/6-Id mice in the hippocampal regions (CA1, CA3). Staining of activated microglia (green, white arrows) was more prominent in the 16/6-Id injected mice brains (A, C) compared to control mice brains (B, D) in the hippocampal regions CA1 (A, B) and CA3 (C, D). Hoechst nucleus staining – blue, GFAP staining – red. Magnification ×40. in microglial activation staining was not seen in the chemical roles such as maintenance of extracellular ion neucortex and piriform cortex, between 16/6-Id and balance. However, in special situations, astrocytes may control-IgG mice. Increased staining for astrocytes increase in number as an inflammatory reaction aimed (GFAP staining) was also noted in the CA3 hippocampal for scaring and repairing CNS tissue. Microglia serve as region in the 16/6-Id injected mice compared to controls scavengers and are activated in an inflammatory reac- (Figure 4). tion. The presence of more astrocytes (gliosis) or the ac- tivation of microglia in brain tissue can implicate an Discussion inflammatory state. Our hypothesis regarding the patho- In the present study we have observed that passive genesis of 16/6-Id antibodies induced-brain impairment transfer of 16/6-Id antibodies directly to mice brains includes several mechanisms: 1) Neuronal degeneration resulted in a selective cognitive impairment, expressed may be caused by direct or indirect injury to hippocam- as visual recognition and spatial memory deficits. De- pal area. For example, recently Berry et al. demonstrated pressive behavior (FST) and locomotor activity (staircase that anti-ATP synthase autoantibodies, purified from test) were not altered in the 16/6-Id injected mice, when Alzheimer’s disease patients, can lead to cognitive im- compared to the control group. Our findings suggest pairment and hippocampal neuron apoptosis in naïve that 16/6-Id antibodies may have a role in the pathogen- mice [38]. Other neurotoxic autoantibodies, such as anti- esis of cognitive impairment observed in some patients phospholipid and anti-ribosomal P antibodies, were shown with SLE [8]. to penetrate living cells and cause functional cellular injury and apoptosis by inhibiting protein synthesis [39,40]. 2) Immunostaining of brain sections from both groups Neuronal function modification. 16/6-Id antibodies may revealed increased presence of activated microglia and recognize and bind antigens expressed on neurons of the astrocytes, in the hippocampal region of the 16/6-Id hippocampus and may affect brain cells by alter signaling, injected mice, compared to the controls. The hippocam- cell function and neurotransmitter pathways [41]. 3) Brain pus has an important function in memory processing, inflammation. Injection of 16/6-Id antibodies may lead to therefore, its damage by an inflammatory processes may brain inflammation involving activation of microglia and affect cognitive performance in the 16/6-Id injected astrocytes, and the production of pro-inflammatory mice. Astrocytes in steady state conditions are mainly cytokines. This inflammatory response can disrupt the responsible for biochemical support and several other 74

Kivity et al. BMC Medicine 2013, 11:90 Page 6 of 9 http://www.biomedcentral.com/1741-7015/11/90 Figure 4 Increased brain inflammation (astrocyes) in 16/6-Id mice in the hippocampal region (CA3). Staining of astrocytes (red) in the hippocampal CA3 region was more prominent in the 16/6-Id injected mice brains (A) compared to control mice injected with commercial IgG (B). Hoechst nucleus staining - blue. Magnification ×40. blood–brain barrier, facilitating entry into the brain by in- experiments, the BBB was breached temporarily by in- flammatory factors, including circulating cells of the im- jection with LPS to imitate an infection [48], while mune system, cytokines, immune-complex mediated small others used noradrenalin to imitate a stressful condition; vessel inflammation, and complement components. The both conditions were implicated in triggering disease inflammatory reaction may induce cognitive changes ob- flare-ups in SLE and NPSLE patients. The studies of served in the injected mice. Diamond et al. added another layer to the current un- derstandings regarding the role of different auto-antibodies We have extensively studied the pathogenesis of differ- in the pathogenesis of NPSLE. Another technique to by- ent autoantibodies and their influence on the brain. In- pass the BBB was used by us in several experiments. In the jection of anti-ribosomal-P antibodies ICV to naïve mice ICV technique, antibodies were injected directly into the resulted in depressive-like behavior in these mice lateral ventricle in the mouse brain, allowing antibody dis- [42,43]. In another study, we found that injection of persal throughout the brain tissue. In our previous studies, antiphospholipid syndrome patients with antibodies in- an experimental NPSLE was induced by passive transfer of duced memory deficits and hyperactivity [15,44]. This anti-ribosomal-P antibodies directly to mice brains [43]. suggests that a certain antibody is linked with each spe- The intra-cerebra-ventricularly injected mice exhibited a cific disease manifestation. The presence of numerous depression-like behavior, not associated with motor or cog- autoantibodies, at least 174 in SLE and 20 in NPSLE, nitive deficits, and was significantly attenuated by prolong which might have a role in the mechanism of the disease treatment with an anti-depressant (fluoxetine), but not were reported during the past years [11,45]. This may with anti-psychotic drug (haloperidol). Interestingly, the explain the diversity of 19 neuropsychiatric manifesta- anti-ribosomal-P antibody specifically stained neurons which tions which can be demonstrated in more than 50% of are related to limbic and olfactory brain areas: the hippo- SLE patients [46]. We propose a hypothesis, that in campus, cingulate cortex and the primary olfactory piriform NPSLE patients different manifestations are the result of cortex [43]. The depressed mice also exhibited a decreased an interplay among various auto-antibodies and genetic smell threshold capability [42], as well as olfactory and lim- and environmental factors. For this process to occur, bic imaging alterations, when manganese-enhanced-mag- auto-antibodies produced in the body must be able to netic resonance imaging (MRI) was performed [49]. cross the blood–brain barrier (BBB). It is presumed that the BBB can become transiently “unlocked” following an Another issue of this puzzle was stressed almost two de- inflammatory insult, an immune complex damage or ex- cades ago when the importance of the idiotypic network posure to infectious endotoxins (for example, lipopoly- in the induction of various autoimmune diseases was ac- saccharide, LPS), allowing antibody penetration. In knowledged [22,50]. One proposed mechanism of action addition, different auto-antibodies may attach to differ- of the 16 /6-Id is via the idiotypic network, in which injec- ent epitopes, expressed unevenly in different brain areas tion of human anti-DNA 16/6-Id mAbs induces the or neuronal networks. In the studies of Diamond et al., generation of anti-Id, and anti-anti-Id, and so on. The pro- anti-DNA antibodies which can cross-react with the duction of 16/6-Id was found to be induced also by several NR2 - anti-NMDA receptor were found in the sera, CSF infectious agents (for example, Klebsiella pneumonia and brains of SLE patients [16,47]. These antibodies [51,52] and Mycobacterium tuberculosis [53]); this could were shown to alter brain cell function and to mediate point to the role of infections in initiating the disease in a apoptotic death in vivo and in vitro [16,47]. In their genetically susceptible individual [54]. 75

Kivity et al. BMC Medicine 2013, 11:90 Page 7 of 9 http://www.biomedcentral.com/1741-7015/11/90 The finding, that 16/6-Id antibodies were detected in design and coordination of the study. EM and JC participated in the design other autoimmune diseases, such as PM/DM and sclero- and coordination of the study and helped to draft the manuscript. YS derma without them expressing central nervous symp- conceived of the study, participated in its design and coordination, and toms is interesting. Perhaps, in some diseases (for helped to draft the manuscript. All authors read and approved the final example, SLE) a variety of systemic factors enable the al- manuscript. tering of BBB permeability. These factors may include other circulating antibodies, inflammatory elements, as Acknowledgments well as vasogenic agents, growth factors and free radicals. We would like to thank Itzik Sehayek for his substantial contribution to this This phenomenon is not unusual in the autoimmunity research. field, for instance, anti-Ro antibodies are associated with myositis or sub-acute skin manifestations in some SLE pa- Author details tients and not in Sjogren patients. 1The Zabludovicz Center for Autoimmune Diseases, Sheba Medical Center, Tel-Hashomer, 2 Derech Sheba St., Ramat Gan 52621, Israel. 2Rheumatic The current finding, that the 16/6-Id is related to Disease Unit, Sheba Medical Center, Tel-Hashomer, 2 Derech Sheba St., spatial novelty and visual recognition memory impair- Ramat Gan 52621, Israel. 3Department of Neurology, Sagol Neuroscience ments in mice, may attest for immune-mediated damage Center, Sheba Medical Center, Tel-Hashomer, 2 Derech Sheba St., Ramat Gan to brain areas relevant for these functions. There is a 52621, Israel. 4Center for Autoimmune Diseases Research (CREA), School of wide agreement that spatial long-term memory and ob- Medicine and Health Sciences, Universidad del Rosario, Kr 24 N°63C-69, ject recognition is dependent on the functioning of the Bogotá, DC 111221, Colombia. 5Doctoral Program in Biomedical Sciences, hippocampal region [55]. Taken together, these concepts Universidad del Rosario, Kr 24 N°63C-69, Bogotá, DC 111221, Colombia. may promote the idea for a treatment for NPSLE via 6Department of Immunology, The Weizmann Institute of Science, 234 Herzl blocking or inhibiting the 16/6-Id. This can be done per- Street, Rehovot 76100, Israel. 7Incumbent of the Laura Schwarz-Kip Chair for haps by treatment with intravenous gamma-globulin, Research of Autoimmune Diseases, Sackler Faculty of Medicine, Tel-Aviv which harbors anti-idiotypic antibodies itself, and has University, Haim Lebanon St., Ramat Aviv 69978, Israel. shown some efficacy in the treatment of NPSLE patients [56]. Other therapeutic means may involve the utilization Received: 28 September 2012 Accepted: 13 December 2012 of inhibitory peptides based on the complementarity de- Published: 4 April 2013 termining region of anti-DNA antibodies. Indeed, such a peptide was shown to be effective in animal models and in References a limited number of lupus patients [57-59]. 1. Ainiala H, Loukkola J, Peltola J, Korpela M, Hietaharju A: The prevalence of Conclusions neuropsychiatric syndromes in systemic lupus erythematosus. Neurology Passive transfer of anti-DNA 16/6-idiotype directly to 2001, 57:496–500. mice brains resulted in cognitive impairment, supported 2. Brey RL, Holliday SL, Saklad AR, Navarrete MG, Hermosillo-Romo D, by cognitive testing impairments, and changes in brain Stallworth CL, Valdez CR, Escalante A, del Rincon I, Gronseth G, Rhine CB, histological analysis. Therefore, the 16/6-idiotype may Padilla P, McGlasson D: Neuropsychiatric syndromes in lupus: prevalence have a role in cognitive decline, as well as other neuro- using standardized definitions. Neurology 2002, 58:1214–1220. psychiatric manifestations, which are found in lupus 3. Lapteva L, Nowak M, Yarboro CH, Takada K, Roebuck-Spencer T, Weickert T, patients. Bleiberg J, Rosenstein D, Pao M, Patronas N, Steele S, Manzano M, van der Veen JW, Lipsky PE, Marenco S, Wesley R, Volpe B, Diamond B, Illei GG: Anti- Abbreviations N-methyl-D-aspartate receptor antibodies, cognitive dysfunction, and anti-ssDNA: Anti-single-stranded-DNA; BBB: Blood–brain barrier; CNS: Central depression in systemic lupus erythematosus. Arthritis Rheum 2006, nervous system; FST: Forced swimming test; ICV: Intra-cerebra-ventricularly; i. 54:2505–2514. p: Intra-peritoneal; LPS: Lipopolysaccharide; mAb: Monoclonal antibody; 4. Greenwood DL, Gitlits VM, Alderuccio F, Sentry JW, Toh BH: Autoantibodies MRI: Magnetic resonance imaging; NMDA: N-methyl-D-aspartate; in neuropsychiatric lupus. Autoimmunity 2002, 35:79–86. NPSLE: Neuropsychiatric systemic lupus erythematosus; PBS: Phosphate 5. Borchers AT, Aoki CA, Naguwa SM, Keen CL, Shoenfeld Y, Gershwin ME: buffered saline; PFA: Paraformaldehyde; SLE: Systemic lupus erythematosus; Neuropsychiatric features of systemic lupus erythematosus. Autoimmun 16/6-Id: 16/6-idiotype. Rev 2005, 4:329–344. 6. The American College of Rheumatology nomenclature and case Competing interests definitions for neuropsychiatric lupus syndromes. Arthritis Rheum 1999, The authors declare that they have no competing interests. 42:599–608. 7. Rekvig OP, Putterman C, Casu C, Gao HX, Ghirardello A, Mortensen ES, Authors’ contributions Tincani A, Doria A: Autoantibodies in lupus: culprits or passive SK participated in the immunohistochemical and behavioral studies and bystanders? Autoimmun Rev 2012, 11:596–603. drafted the manuscript. AK participated in the immunohistochemical and 8. Kozora E, Hanly JG, Lapteva L, Filley CM: Cognitive dysfunction in systemic behavioral studies and helped to draft the manuscript. MTA participated in lupus erythematosus: past, present, and future. Arthritis Rheum 2008, the immunohistochemical studies and performed the statistical analysis. MLR 58:3286–3298. and YZ participated in the behavioral studies. NAL participated in the 9. Kapadia M, Sakic B: Autoimmune and inflammatory mechanisms of CNS behavioral studies and helped to draft the manuscript. MB participated in damage. Prog Neurobiol 2011, 95:301–333. the design and helped to draft the manuscript. JMA participated in the 10. Oikonomou KG, Zachou K, Dalekos GN: Alpha-actinin: a multidisciplinary protein with important role in B-cell driven autoimmunity. Autoimmun Rev 2011, 10:389–396. 11. Zandman-Goddard G, Chapman J, Shoenfeld Y: Autoantibodies involved in neuropsychiatric SLE and antiphospholipid syndrome. Semin Arthritis Rheum 2007, 36:297–315. 12. Weiner SM, Klein R, Berg PA: A longitudinal study of autoantibodies against central nervous system tissue and gangliosides in connective tissue diseases. Rheumatol Int 2000, 19:83–88. 13. Bonfa E, Golombek SJ, Kaufman LD, Skelly S, Weissbach H, Brot N, Elkon KB: Association between lupus psychosis and anti-ribosomal P protein antibodies. N Engl J Med 1987, 317:265–271. 76

Kivity et al. BMC Medicine 2013, 11:90 Page 8 of 9 http://www.biomedcentral.com/1741-7015/11/90 14. Toubi E, Shoenfeld Y: Clinical and biological aspects of anti-P-ribosomal 35. Porsolt RD, Bertin A, Jalfre M: Behavioral despair in mice: a primary protein autoantibodies. Autoimmun Rev 2007, 6:119–125. screening test for antidepressants. Arch Int Pharmacodyn Ther 1977, 229:327–336. 15. Katzav A, Pick CG, Korczyn AD, Oest E, Blank M, Shoenfeld Y, Chapman J: Hyperactivity in a mouse model of the antiphospholipid syndrome. 36. Tordera RM, Totterdell S, Wojcik SM, Brose N, Elizalde N, Lasheras B, Del Rio Lupus 2001, 10:496–499. J: Enhanced anxiety, depressive-like behaviour and impaired recognition memory in mice with reduced expression of the vesicular glutamate 16. DeGiorgio LA, Konstantinov KN, Lee SC, Hardin JA, Volpe BT, Diamond B: A transporter 1 (VGLUT1). Eur J Neurosci 2007, 25:281–290. subset of lupus anti-DNA antibodies cross-reacts with the NR2 glutamate receptor in systemic lupus erythematosus. Nat Med 2001, 37. Dix SL, Aggleton JP: Extending the spontaneous preference test of 7:1189–1193. recognition: evidence of object-location and object-context recognition. Behav Brain Res 1999, 99:191–200. 17. Shoenfeld Y, Isenberg DA, Rauch J, Madaio MP, Stollar BD, Schwartz RS: Idiotypic cross-reactions of monoclonal human lupus autoantibodies. 38. Berry A, Vacirca D, Capoccia S, Bellisario V, Malorni W, Ortona E, Cirulli F: J Exp Med 1983, 158:718–730. Anti-ATP synthase autoantibodies induce neuronal death by apoptosis and impair cognitive performance in C57BL/6 mice. J Alzheimers Dis 2013, 18. Shoenfeld Y, Hsu-Lin SC, Gabriels JE, Silberstein LE, Furie BC, Furie B, Stollar 33:317–321. BD, Schwartz RS: Production of autoantibodies by human-human hybridomas. J Clin Invest 1982, 70:205–208. 39. Sun KH, Tang SJ, Lin ML, Wang YS, Sun GH, Liu WT: Monoclonal antibodies against human ribosomal P proteins penetrate into living cells and 19. Shoenfeld Y, Rauch J, Massicotte H, Datta SK, Andre-Schwartz J, Stollar BD, cause apoptosis of Jurkat T cells in culture. Rheumatology (Oxford) 2001, Schwartz RS: Polyspecificity of monoclonal lupus autoantibodies produced 40:750–756. by human-human hybridomas. N Engl J Med 1983, 308:414–420. 40. Koscec M, Koren E, Wolfson-Reichlin M, Fugate RD, Trieu E, Targoff IN, 20. Andre-Schwartz J, Datta SK, Shoenfeld Y, Isenberg DA, Stollar BD, Schwartz Reichlin M: Autoantibodies to ribosomal P proteins penetrate into live RS: Binding of cytoskeletal proteins by monoclonal anti-DNA lupus hepatocytes and cause cellular dysfunction in culture. J Immunol 1997, autoantibodies. Clin Immunol Immunopathol 1984, 31:261–271. 159:2033–2041. 21. Amital-Teplizki H, Avinoach I, Coates AR, Kooperman O, Blank M, Shoenfeld 41. Matus S, Burgos PV, Bravo-Zehnder M, Kraft R, Porras OH, Farias P, Barros LF, Y: Binding of monoclonal anti-DNA and anti-TB glycolipids to brain Torrealba F, Massardo L, Jacobelli S, Gonzalez A: Antiribosomal-P tissue. Autoimmunity 1989, 4:277–287. autoantibodies from psychiatric lupus target a novel neuronal surface protein causing calcium influx and apoptosis. J Exp Med 2007, 22. Blank M, Shoenfeld Y: The story of the 16/6 idiotype and systemic lupus 204:3221–3234. erythematosus. Isr Med Assoc J 2008, 10:37–39. 42. Katzav A, Ben-Ziv T, Chapman J, Blank M, Reichlin M, Shoenfeld Y: Anti-P 23. Isenberg DA, Shoenfeld Y, Madaio MP, Rauch J, Reichlin M, Stollar BD, ribosomal antibodies induce defect in smell capability in a model of Schwartz RS: Anti-DNA antibody idiotypes in systemic lupus CNS-SLE (depression). J Autoimmun 2008, 31:393–398. erythematosus. Lancet 1984, 2:417–422. 43. Katzav A, Solodeev I, Brodsky O, Chapman J, Pick CG, Blank M, Zhang W, 24. Galeazzi M, Bellisai F, Sebastiani GD, Morozzi G, Marcolongo R, Houssiau F, Reichlin M, Shoenfeld Y: Induction of autoimmune depression in mice by Cervera R, Levy Y, George J, Sherer Y, Shoenfeld Y: Association of 16/6 and anti-ribosomal P antibodies via the limbic system. Arthritis Rheum 2007, SA1 anti-DNA idiotypes with anticardiolipin antibodies and clinical 56:938–948. manifestations in a large cohort of SLE patients. European Concerted Action on the Immunogenetics of SLE. Clin Exp Rheumatol 1998, 44. Menachem A, Chapman J, Katzav A: Hyperactivity induced by 16:717–720. antiphospholipid syndrome serum. Ann N Y Acad Sci 2009, 1173:422–426. 25. Amital-Teplizki H, Bearman JE, Miele PW Jr, Sukenik S, Shoenfeld Y: A multidimensional autoantibody analysis specifying systemic lupus 45. Sherer Y, Gorstein A, Fritzler MJ, Shoenfeld Y: Autoantibody explosion in erythematosus patients with neuropsychiatric symptomatology. Isr J Med systemic lupus erythematosus: more than 100 different antibodies found Sci 1992, 28:422–427. in SLE patients. Semin Arthritis Rheum 2004, 34:501–537. 26. Isenberg DA, Collins C: Detection of cross-reactive anti-DNA antibody 46. Unterman A, Nolte JE, Boaz M, Abady M, Shoenfeld Y, Zandman-Goddard G: idiotypes on renal tissue-bound immunoglobulins from lupus patients. Neuropsychiatric syndromes in systemic lupus erythematosus: a meta- J Clin Invest 1985, 76:287–294. analysis. Semin Arthritis Rheum 2010, 41:1–11. 27. Isenberg D, Dudeney C, Wojnaruska F, Bhogal BS, Rauch J, Schattner A, 47. Faust TW, Chang EH, Kowal C, Berlin R, Gazaryan IG, Bertini E, Zhang J, Naparstek Y, Duggan D: Detection of cross reactive anti-DNA antibody Sanchez-Guerrero J, Fragoso-Loyo HE, Volpe BT, Diamond B, Huerta PT: idiotypes on tissue-bound immunoglobulins from skin biopsies of lupus Neurotoxic lupus autoantibodies alter brain function through two patients. J Immunol 1985, 135:261–264. distinct mechanisms. Proc Natl Acad Sci U S A 2010, 107:18569–18574. 28. Konikoff F, Isenberg DA, Kooperman O, Kennedy RC, Rauch J, Theodor E, 48. Kowal C, DeGiorgio LA, Nakaoka T, Hetherington H, Huerta PT, Diamond B, Shoenfeld Y: Common lupus anti-DNA antibody idiotypes in chronic liver Volpe BT: Cognition and immunity; antibody impairs memory. Immunity diseases. Clin Immunol Immunopathol 1987, 43:265–272. 2004, 21:179–188. 29. Shoenfeld Y, Livne A, Blank M, Argov S, Krup M, Fleishmakher E, Sukenik S, 49. Kivity S, Tsarfaty G, Agmon-Levin N, Blank M, Manor D, Konen E, Chapman J, Teplizki H: A human monoclonal anti-DNA antibody derived from a Reichlin M, Wasson C, Shoenfeld Y, Kushnir T: Abnormal olfactory function patient with polymyositis having the common lupus 16/6 idiotype. demonstrated by manganese-enhanced MRI in mice with experimental Immunol Lett 1988, 19:77–83. neuropsychiatric lupus. Ann N Y Acad Sci 2010, 1193:70–77. 30. Mendlovic S, Brocke S, Shoenfeld Y, Ben-Bassat M, Meshorer A, Bakimer R, 50. Shoenfeld Y, Ben-Yehuda O, Messinger Y, Bentwitch Z, Rauch J, Isenberg DI, Mozes E: Induction of a systemic lupus erythematosus-like disease in Gadoth N: Autoimmune diseases other than lupus share common anti- mice by a common human anti-DNA idiotype. Proc Natl Acad Sci U S A DNA idiotypes. Immunol Lett 1988, 17:285–291. 1988, 85:2260–2264. 51. Isenberg DA, Shoenfeld Y: An analysis of autoimmunity through studies 31. Shoenfeld Y, Mozes E: Pathogenic idiotypes of autoantibodies in of DNA antibody idiotypes. Autoimmunity 1988, 1:67–75. autoimmunity: lessons from new experimental models of SLE. FASEB J 1990, 4:2646–2651. 52. El-Roiey A, Gross WL, Luedemann J, Isenberg DA, Shoenfeld Y: Preferential secretion of a common anti-DNA idiotype (16/6 Id) and anti- 32. Shoenfeld Y: Idiotypic induction of autoimmunity: a new aspect of the polynucleotide antibodies by normal mononuclear cells following idiotypic network. FASEB J 1994, 8:1296–1301. stimulation with Klebsiella pneumoniae. Immunol Lett 1986, 12:313–319. 33. Waisman A, Shoenfeld Y, Blank M, Ruiz PJ, Mozes E: The pathogenic 53. Shoenfeld Y, Vilner Y, Coates AR, Rauch J, Lavie G, Shaul D, Pinkhas J: human monoclonal anti-DNA that induces experimental systemic lupus Monoclonal anti-tuberculosis antibodies react with DNA, and erythematosus in mice is encoded by a VH4 gene segment. Int Immunol monoclonal anti-DNA autoantibodies react with Mycobacterium 1995, 7:689–696. tuberculosis. Clin Exp Immunol 1986, 66:255–261. 34. Shoenfeld Y, Nahum A, Korczyn AD, Dano M, Rabinowitz R, Beilin O, Pick 54. Kivity S, Agmon-Levin N, Blank M, Shoenfeld Y: Infections and CG, Leider-Trejo L, Kalashnikova L, Blank M, Chapman J: Neuronal-binding autoimmunity–friends or foes? Trends Immunol 2009, 30:409–414. antibodies from patients with antiphospholipid syndrome induce cognitive deficits following intrathecal passive transfer. Lupus 2003, 55. Broadbent NJ, Squire LR, Clark RE: Spatial memory, recognition memory, 12:436–442. and the hippocampus. Proc Natl Acad Sci U S A 2004, 101:14515–14520. 77

Kivity et al. BMC Medicine 2013, 11:90 Page 9 of 9 http://www.biomedcentral.com/1741-7015/11/90 56. Milstone AM, Meyers K, Elia J: Treatment of acute neuropsychiatric lupus with intravenous immunoglobulin (IVIG): a case report and review of the literature. Clin Rheumatol 2005, 24:394–397. 57. Mozes E, Sharabi A: A novel tolerogenic peptide, hCDR1, for the specific treatment of systemic lupus erythematosus. Autoimmun Rev 2010, 10:22–26. 58. Lapter S, Marom A, Meshorer A, Elmann A, Sharabi A, Vadai E, Neufeld A, Sztainberg Y, Gil S, Getselter D, Chen A, Mozes E: Amelioration of brain pathology and behavioral dysfunction in mice with lupus following treatment with a tolerogenic peptide. Arthritis Rheum 2009, 60:3744–3754. 59. Sthoeger ZM, Sharabi A, Molad Y, Asher I, Zinger H, Dayan M, Mozes E: Treatment of lupus patients with a tolerogenic peptide, hCDR1 (Edratide): immunomodulation of gene expression. J Autoimmun 2009, 33:77–82. doi:10.1186/1741-7015-11-90 Cite this article as: Kivity et al.: 16/6-idiotype expressing antibodies induce brain inflammation and cognitive impairment in mice: the mosaic of central nervous system involvement in lupus. BMC Medicine 2013 11:90. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit

‫‪Superior temporal gyrus thickness correlates‬‬ ‫‪with cognitive performance in multiple sclerosis‬‬ ‫‪Brain Structure & Function | 2013‬‬ ‫מנחה‪ :‬פרופ' יואב צ'פמן‬ ‫מנחה‪ :‬פרופ' ענת אחירון‬ ‫מנהל מחלקה נוירולוגית‬ ‫מייסדת פרויקט ח\"ץ‪ ,‬מנהלת המרכז‬ ‫לטרשת נפוצה ואחראית הקתדרה‬ ‫‪[email protected]‬‬ ‫למחלות אוטואימוניות אוניברסיטת ת\"א‬ ‫‪[email protected]‬‬ ‫התמזל מזלי והתקבלתי לפרוייקט ח”ץ לפני ‪ 14‬שנים‪ .‬במידה רבה‪ ,‬רבים‬ ‫אסף אחירון‬ ‫מהדברים שלמדתי‪ ,‬סיגלתי והכרתי בתקופה שבה לקחתי חלק בפרוייקט‪,‬‬ ‫אונ' תל אביב‬ ‫בהרצאות האינטימיות‪ ,‬ובהדרכות האישיות‪ ,‬מלווים אותי עד היום‪.‬‬ ‫השתתף כסטודנט בפרויקט ח״ץ‬ ‫תמציתה של הרפואה המצטיינת בעיני היא מה שלמדתי להכיר‬ ‫בין השנים ‪2007-2009‬‬ ‫בפרוייקט הזה ‪ -‬היכולת להעריך מידע מבוסס מחקר‪ ,‬לבצע מחקר‬ ‫רלוונטי‪ ,‬עדכני ואחראי‪ ,‬וכמובן‪ ,‬רפואה מבוססת רופאים מעולים‪.‬‬ ‫‪[email protected]‬‬ ‫מה שניתן לי בפרוייקט הזה אני אוחז כל יום בשתי ידיים וממשיך בבניית‬ ‫הקריירה שלי‪.‬‬ ‫היכולות המאתגרות של תחילת דרכי לוו על ידי רופאים מצויינים שרצו‬ ‫לעזור וידעו איך‪ .‬הרופא‪-‬חוקר שאני היום מאפשר לי להעביר את כל מה‬ ‫שהתחיל באותה תוכנית ומאז רק צמח‪ ,‬ולראות את התוכנית מנקודת מבטם‬ ‫של החוקרים המנוסים‪ ,‬ענפי הקריירה‪ .‬גם כאן קיים סיפוק לא קטן‪.‬‬ ‫‪79‬‬

Brain Struct Funct (2013) 218:943–950 DOI 10.1007/s00429-012-0440-3 ORIGINAL ARTICLE Superior temporal gyrus thickness correlates with cognitive performance in multiple sclerosis Asaf Achiron • Joab Chapman • Sigal Tal • Eran Bercovich • Hararai Gil • Anat Achiron Received: 3 March 2012 / Accepted: 19 June 2012 / Published online: 12 July 2012 Ó Springer-Verlag 2012 Abstract Decreased cortical thickness that signifies gray analysis demonstrated that only the LH superior temporal matter pathology and its impact on cognitive performance gyrus thickness was associated with cognitive performance is a research field with growing interest in relapsing– and its thickness correlated with motor skills (r = 0.65, remitting multiple sclerosis (RRMS) and needs to be fur- p = 0.003), attention (r = 0.45, p = 0.042), and informa- ther elucidated. Using high-field 3.0 T MRI, three-dimen- tion processing speed (r = 0.50, p = 0.025). Our findings sional T1-FSPGR (voxel size 1 9 1 9 1 mm) cortical show that restricted cortical thinning occurs in RRMS thickness was measured in 82 regions in the left hemi- patients with mild disease and that LH superior temporal sphere (LH) and right hemisphere (RH) in 20 RRMS gyrus atrophy is associated with cognitive dysfunction. patients with low disease activity and in 20 age-matched healthy subjects that in parallel underwent comprehensive Keywords Multiple sclerosis Á Cortical thickness Á MRI Á cognitive evaluation. The correlation between local cortical Cognitive performance Á Superior temporal gyrus atrophy and cognitive performance was examined. We identified seven regions with cortical tissue loss that dif- Introduction fered between RRMS and age-matched healthy controls. These regions were mainly located in the frontal and Relapsing–remitting multiple sclerosis (RRMS) is classi- temporal lobes, specifically within the gyrus rectus, inferior cally regarded as a white matter disease, although neuronal frontal sulcus, orbital gyrus, parahippocampal gyrus, and injury and gray matter atrophy have been extensively superior temporal gyrus, with preferential left asymmetry. described in RRMS patients, even in the early stage of the Increased cortical thickness was identified in two visual disease (Chard et al. 2004; Tiberio et al. 2005; Zivadinov sensory regions, the LH inferior occipital gyrus, and the and Minagar 2009). As white matter pathology has shown RH cuneus, implicating adaptive plasticity. Correlation only moderate correlation with cognitive abilities (Geurts and Barkhof 2008), our interest was directed at gray matter A. Achiron (&) Á J. Chapman loss and its relation to cognitive impairment in RRMS Department of Neurology, Multiple Sclerosis Center, patients, taking into account that oligodendrocytes consti- Sheba Medical Center, Sackler School of Medicine, tute about 75 % of the glial cells in the cortex (Pelvig et al. Tel-Aviv University, Tel-Hashomer, Israel 2008). Gray matter atrophy has been studied in RRMS e-mail: [email protected] patients using novel methods of imaging to better under- stand its mechanisms, pathogenesis, and clinical relevance. S. Tal Advanced non-conventional MRI techniques such as MTR, Department of Radiology, Sheba Medical Center, MRS, and DTI provide better detection of diffuse injury in Sackler School of Medicine, Tel-Aviv University, the cortical gray matter and of gray matter demyelinating Tel-Hashomer, Israel lesions (Inglese et al. 2004; Ge et al. 2001; Bozzali et al. 2002). Moreover, the findings of gray matter loss were E. Bercovich Á H. Gil Á A. Achiron validated by immunohistochemistry methods showing Multiple Sclerosis Center, Sheba Medical Center, Sackler School of Medicine, Tel-Aviv University, Tel-Hashomer, Israel 123 80

944 Brain Struct Funct (2013) 218:943–950 cortical demyelination with gray-matter lesions discon- used a computerized query to select patients with mild nected from foci of white-matter demyelination as well as RRMS according to the following inclusion criteria: gray-matter inflammation that occurs in the absence of (a) relapsing–remitting disease course; (b) age range demyelination (Vercellino et al. 2005; Rudick and Trapp between 18 and 55 years; (c) neurological disability by the 2009). Cortical gray matter loss was reported to appear in expanded disability status scale (EDSS) score (Kurtzke the early clinical stage of RRMS, even as early as during 2008) \\3.0; (d) no corticosteroid treatment within 4 weeks the first clinical event (Calabrese et al. 2007). Several preceding brain MRI or cognitive assessments; (e) not studies have focused on global cortical gray matter loss within the period of acute relapse; (f) no major depression (Amato et al. 2004, 2007), as well as on specific cortical or anxiety as assessed by the related Hamilton question- gray matter loss (Tekok-Kilic et al. 2007; Morgen et al. naires for depression and anxiety; (g) cognitive assessment 2006), and evaluated the relation between gray-matter performed using the MindstreamsÒ Computerized Cogni- atrophy and cognitive impairment. These studies demon- tive Battery (MCCB) (Achiron et al. 2007); (h) brain MRI strated a widespread pattern of cortical thinning in RRMS examination performed using 3D high-resolution fast patients with cognitive dysfunction and suggested cortical spoiled gradient-echo (FSPGR) MS protocol; and (i) brain thinning as a marker of a more aggressive disease. MRI and cognitive assessments performed within up to 6 months interval. Data for normal healthy subjects were We hypothesized that cortical gray matter damage obtained from the hospital radiology MRI unit database, affecting specific brain areas could be related to a deficit in which includes MRI data of healthy volunteers recruited a related cognitive domain and that the assessment of the from the local community and hospital staff. None of the correlation between cortical thickness and cognitive per- healthy volunteers suffered from any neurological disease formance may provide new insights into the mechanism or took central nervous system active medications. All underlying the development of cognitive impairment in healthy subjects gave written informed consent prior to RRMS patients as well as enable us to better understand MRI examination and were interviewed and examined to plasticity mechanisms that prevent gray-matter loss. exclude neurological or cognitive abnormalities. For each Accordingly, we were interested to assess RRMS patients MS patient an age-matched healthy subject was randomly with mild disease as in these patients we hypothesized that selected. The study was approved by the Sheba IRB compensatory mechanisms enhancing neural plasticity committee. Computerized software anonymization (http:// operate to reduce disease activity and even result in dicom.online.fr) ensured that personal identities were kept increased cortical volume in specific regions. To test these confidential. hypotheses, we applied high-resolution MRI protocol and an advanced image analysis technique to interrogate MRI acquisition protocol cortical thinning in a group of RRMS patients with mild disease evident by low neurological disability, in compar- Brain MRI was performed on 3.0-T MR scanner (Signa, ison with age-matched healthy subjects. We also investi- GE) using high-resolution eight-channel head coil for both gated the structural correlates of the thinned regional brain patients and controls. Imaging sequence of three-dimen- areas by integrating cognitive performance measures sional fast spoiled gradient-echo (fSPGR) with isotropic derived from a comprehensive cognitive battery. Our voxel size of 1 9 1 9 1 mm, TE = 2 ms, TR = 6 ms, findings indicate a specific gray matter pattern loss and TI = 450 ms, 146 contiguous sagittal slices with field of suggest the occurrence of brain plasticity in RRMS with view (FOV) of 256 9 256 mm, matrix 256 9 256 mm, mild disease course. and flip angle of 20, were acquired. Methods T2 lesion volume measurements were performed using MSAnalyze software that segments structures in 3D and Settings and study design applies a multispectral technique based on bayesian clas- sification of brain tissue using active contour methods Data were obtained through the Sheba Multiple Sclerosis (Achiron et al. 2002). Center Computerized Database Registry. The Multiple Sclerosis Center Database Registry was implemented in Cortical thickness measurements 1995 and includes patients’ demographic and clinical data archived by an advanced computerized electronic record- MR images were imported to BrainVoyagerQX image keeping software system. Patients’ records are updated by analysis software (http://www.brainvoyager.com) for seg- the Center’s neurologists prospectively, during each patient mentation, reconstruction, inflation, and cortical thickness visit into the center. For the purpose of the current study we measurements. Cortical thickness measurements were per- formed using the Laplace method algorithm, implemented 123 81

Brain Struct Funct (2013) 218:943–950 945 in BrainVoyagerQX. This algorithm provides precise and gyri. The MRI data was uploaded to the software and reliable results and was validated in a phantom study (Ha- exported to an EXCEL file by an experienced technician idar and Soul 2006; Lee et al. 2006). Regions in the left who was blinded to the clinical and cognitive data. Results hemisphere (LH) and right hemisphere (RH) were measured were further inspected by an experienced neuroradiologist. using the software brain atlas. Specifications of the methods for cortical thickness measurements were previously Cognitive assessments described in detail (Geuze et al. 2008). In brief, the software first performs a pre-analysis stage of correction of in- Cognitive assessment was performed using the MCCB homogeneity in scanner fields to increase the contrast (http://www.neurotrax.com). The MCCB measures cog- between gray and white matter. Then, each MRI scan is nitive domains including executive function, attention, normalized to a standard spatial Talairach space providing a information processing speed, memory, motor skills, common coordinate reference. Next, isolation of the cere- verbal function, and visual spatial performance. MCCB bral cortex is achieved by removing the skull, cerebellum, index scores are computed to summarize performance in spinal cord, and basal ganglia and filling of the ventricles. each cognitive domain. A global cognitive score (GCS) is This stage is followed by segmentation of the gray/white computed as the average of the index scores. GCS and matter border and the gray/pia matter border. Cortical domain index scores were normalized according to strat- thickness is calculated by solving Laplace’s equation for the ifications of age and education using standard z score potential between the inner-gray/white matter and outer- formulae to a standardized scale (mean 100, SD 15). The gray/pia matter surfaces. After acquisition of cortical MCCB was validated in healthy and in mild cognitive thickness measurements for each subject based on the MRI impairment subjects as well as in MS patients (Achiron coronal sections, BrainVoyagerQX reconstructs a three- et al. 2007). The battery consists of interactive software dimensional presentation of the cortex (Fig. 1a–c). Next, producing accuracy and response time data and is rou- cortex-based alignment is performed to align macro-ana- tinely performed by MS patients followed at our center as tomical (gyri and sulci) structures of each subject to the a tool for profiling cognitive function and detection of group’s average brain. This step is performed by aligning cognitive impairment. Patients performed practice ses- the reconstructed cortices using curvature information sions prior to individual tests and were instructed to reflecting the gyral and sulcal folding patterns. Accord- respond as quickly as possible in the relevant cognitive ingly, a mean measurement of cortical thickness for each domains. The test’s duration is about 50 min and it was cortical surface point is established in three-dimensional conducted under the supervision of an experienced space by high-resolution surface-averaging algorithms. psychologist. Cortical surface point is defined by the distance between the gray/white and pial surfaces. The software generates highly Statistical analysis accurate models of these surfaces and then measures the distance between them. The distance between the two sur- Statistical analysis was performed using the SASÒ software faces is the thickness of the cortical gray matter. This cor- version 9.1 (SAS Institute, Cary, NC, USA). Analyses tical matching approach substantially improves statistical included descriptive statistics for demographic, clinical, analysis across subjects by enabling anatomical integration radiological and cognitive data. T test was performed to and reducing anatomical variability. Not only the resulting test for significant differences in cortical thickness mea- statistical map is projected on the surface rendering of a surements within various brain regions between MS cortical reconstruction, but also individual anatomical patients and healthy subjects. Correlations between the information (as provided by labeled cortical voxels of gyral differing brain regions with cognitive tasks scores were and sulcal patterns) is actively used in the statistical analysis performed by the Spearman rank correlation coefficient. of single subject and group data, with the scope of Data are presented as mean ± SE and a p value \\0.05 was enhancing sensitivity and improving the spatial correspon- considered significant. dence across brains. Cortical thickness measurements of 41 regions in each hemisphere were exported for statistical Results analysis. The software automatically assigns a neuroana- tomical label to each location on a cortical surface model. Subjects The technique is based on probabilistic information esti- mated from a manually labeled training. This procedure Twenty RRMS patients with mild disease, 11 females, 9 incorporates both geometric information derived from the males, mean age 32.2 ± 2.5 years were included in this cortical model and neuroanatomical data from the training study. Mean disease duration was 5.7 ± 1.2 years (range set. The result is a complete labeling of cortical sulci and 123 82

946 Brain Struct Funct (2013) 218:943–950 Fig. 1 Cortical thickness measurements. a Segmentation of the gray/white matter border and of the gray/pia matter. b Acquisition of cortical thickness measurements for one subject. c Thickness measurements in three- dimensional reconstruction of the cortex 1–22 years), neurological disability assessed by the EDSS Cortical thickness and cognitive performance was 1.4 ± 0.2, GCS was 95.5 ± 2.1 and T2 lesion volume was 5.9 ± 1.4 ml. Nine RRMS patients (45 %) were Correlation analysis performed between the cortical treated with immunomodulatory drugs for a mean treat- regions that significantly differed in thickness between ment period of 2.6 years. The mean period interval RRMS patients and healthy subjects and cognitive tasks between brain imaging and cognitive assessment was scores disclosed that only the LH superior temporal gyrus 1.6 ? 0.8 months. Demographic, clinical, cognitive and correlated with GCS (r = 0.51, p = 0.019), motor skills MRI data for patients included in the study are shown in (r = 0.65, p = 0.003), attention (r = 0.45, p = 0.042), Table 1. The control group included 20 age-matched and information processing speed (r = 0.50, p = 0.025). healthy subjects, 12 females, 8 males, with a mean age of 33.9 ± 2.4 years. Discussion Regional cortical thickness measurements In the current study we compared cortical gray matter thickness in RRMS patients with age-matched healthy Prominent cortical thinning in RRMS patients as compared subjects and demonstrated gray matter atrophy even in MS with healthy subjects was localized to the gyrus rectus, patients with relatively mild disability. The image analysis inferior frontal sulcus, orbital gyrus, parahippocampal software we used enabled voxel by voxel calculation that gyrus, and superior temporal gyrus. It is noteworthy that established a mean measurement of cortical thickness for these differences were evident only in the LH (Table 2; each cortical surface point reducing anatomical variability. Fig. 2a). Two areas, the LH Inferior occipital gyrus and the The data therefore are absolute and represent decrease or RH Cuneus, both related to the visual pathways, were increase in absolute terms and not findings relative to found to be thicker in MS patients as compared to healthy global normalization. The findings support previous studies subjects (Fig. 2b). 123 83

Brain Struct Funct (2013) 218:943–950 947 Table 1 Descriptive data for study patients RRMS (N = 20) areas measured (41 in each hemisphere), suggesting that Variable the gray matter damage is relatively restricted. It is of interest that all these seven cortical regions are coupled Gender 11 (55) with primary sensory input as follows: (1) the LH gyrus Females [n (%)] 9 (45) rectus associated with social behavior and social cognition Males [n (%)] in tasks involving facial recognition, altruism and attribu- 32.2 ± 2.5 tion of intentions and perception of anger in others (Brunet Age (years) 30 et al. 2000; Goldberg et al. 2006; Moll et al. 2002); (2) the Mean ± SE LH inferior frontal sulcus and (3) the LH orbital gyrus Median 5.7 ± 1.2 related to olfactory, taste, somatic sensory association 4 cortex and the visual association cortex (Price 2007); (4) Disease duration (years) the LH parahippocampal gyrus related to memory encod- Mean ± SE 1.4 ± 0.2 ing and retrieval (Maguire and Mummery 1999); (5) the Median 2.0 LH superior temporal gyrus associated with auditory and 9 (45) speech comprehension; (6) the LH inferior occipital gyrus EDSS and (7) the RH cuneus both connected with vision. Sensory Mean ± SE 5.9 ± 1.4 deficits are typical of MS patients at onset and along the Median 4 disease course and our findings indicate that involvement of the sensory-related regions is reflected by cortical thin- IMD treated [n (%)] 95.5 ± 2.1 (101) ning. Additional unexpected finding was the lateralization T2-lesion volume (ml) 94.3 ± 3.2 (96) of cortical atrophy, as six of the seven cortical regions 95.4 ± 2.3 (100) found to be thinner in RRMS patients were located in the Mean ± SE 96.3 ± 3.5 (100) LH. This could support previously reported preferential Median 100.3 ± 3.3 (98) left-sided involvement in MS with a cluster of reduced gray Cognitive performance, mean ± SE (median) 96.5 ± 3.1 (102) matter present on the left, extending from pre-rolandic Global cognitive score 92.3 ± 3.4 (97) cortex to the superior temporal area (Prinster et al. 2006). Information processing speed 94.4 ± 3.2 (101) Moreover, lateralized functional changes have been Attention described in MS in relation to cognitive dysfunction Verbal function including left fronto-temporal metabolic and cerebral blood Visual spatial perception flow reductions indicating a predominant left temporo- Executive function parietal involvement with a preferential correlation to Memory deficits in verbal fluency and verbal memory (Pozzilli et al. Motor skills 1991, 1992). In accordance, alterations of the P600 com- ponent of cortical event-related potentials in the left frontal RRMS relapsing–remitting multiple sclerosis, EDSS expanded dis- and temporo-parietal areas in MS patients with memory ability status scale, IMD immunomodulatory drugs disturbances were reported (Sfagos et al. 2003). These findings suggest that in the early stage of the disease a gray Table 2 Cortical regions that differ between RRMS patients and matter loss occurs in the LH, and this preferential left healthy subjects atrophy that correlates with cognitive dysfunction (Pan et al. 2001) may explain specific deficits such as decreased Cortical region RRMS Control p performance for the paced auditory serial addition test, patients subjects which activates preferentially left fronto-parieto-temporal (N = 20) (N = 20) areas (Audoin et al. 2005). LH gyrus rectus 3.3 ± 0.07 3.5 ± 0.1 0.037 A striking result in our study was that two cortical LH inferior frontal sulcus 3.1 ± 0.07 3.4 ± 0.08 0.034 regions, the LH Inferior occipital gyrus and the RH cuneus LH orbital gyrus 3.4 ± 0.09 3.8 ± 0.08 0.015 were found to be thicker in RRMS patients compared with LH parahippocampal gyrus 3.2 ± 0.1 3.6 ± 0.1 0.0059 normal subjects. These results imply that cortical reorga- LH superior temporal Gyrus 3.0 ± 0.08 3.3 ± 0.07 0.0085 nization and increased recruitment of cortical networks LH inferior occipital Gyrus 3.1 ± 0.07 2.8 ± 0.07 0.018 within adjacent brain areas may operate as an adaptive RH cuneus 2.7 ± 0.06 2.5 ± 0.07 0.015 mechanism to compensate for decreased cognitive perfor- mance (Staffen et al. 2002). Myelination is known to be Thickness measured in millimeter enhanced by increasing the electrical activity of neigh- boring axons, clearly linking neuronal electrical activity to similarly showing gray matter loss in patients with mild disability (Sailer et al. 2003; Calabrese et al. 2010). Our study further extends these observations by demonstrating the pattern of anatomic decrease of cortical thickness in specific brain regions. We found that cortical thickness was significantly decreased only in seven out of 82 cortical 123 84

948 Brain Struct Funct (2013) 218:943–950 Fig. 2 Comparison of cortical thickness between RRMS patients and healthy subjects. a Significantly thinner cortical areas in RRMS patients are demonstrated in the gyrus rectus (yellow), inferior frontal sulcus (green), orbital gyrus (red), parahippocampal gyrus (gray), and superior temporal gyrus (purple). Right LH outer surface, left LH inner surface. b Significantly thicker cortical areas in RRMS patients are demonstrated in the LH inferior occipital gyrus (yellow) and the RH cuneus (red). Right LH outer surface, left RH inner surface myelinogenesis (Demerens et al. 1996). This functional cognitive domains and demonstrated that the decrease in plasticity could lead to synaptic proliferation and to an the LH superior temporal gyrus thickness is a marker of increase in cortical thickness in the recruited areas, and as cognitive impairment specifically related to attention and was demonstrated in our patients, especially in areas rela- information processing speed. As the superior temporal ted to the visual pathways. Functional plasticity was gyrus is known to be associated with auditory and speech demonstrated in animal models (Anderson et al. 2002), comprehension, our findings suggest that the structural after training (Draganski et al. 2004), and in stroke patients integrity of this region affects the capacity of MS patients where an increase in the activation response in the ventral to appropriately and in-time comprehend spoken sentences postcentral gyrus relative to controls was also associated and similarly indicate impaired ability to attentively per- with an increase in cortical thickness in the same area form this task over time, even if no significant neurological (Schaechter et al. 2006). It is of importance to note that in disability is demonstrated. RRMS patients, early in the disease process, increase in the thickness of specific cortical areas could reflect plasticity- The limitations of our study are related to the fact that related processes associated with anatomical changes. In statistical significances were calculated without adjustment this line, a recent study reported of increased brain acti- to multiple comparisons due to the small sample size. vation in MS patients during response inhibition by fMRI, However, only the brain areas found to show significant suggesting that MS does not affect the ability of the brain differences between MS patients and healthy subjects were to compensate or reorganize with increasing demands subjected to correlation analysis with cognitive parameters, (Smith et al. 2009), and similarly, in another fMRI study thus reducing the number of variables and the effect of (Rocca et al. 2003), it was shown that in MS patients with multiple testing. multiple T2-weighted lesions, cortical reorganization occurs over a rather distributed sensorimotor network. Taken together, our findings, though preliminary, pro- vide new insight into the cortical pathology in RRMS Finally, our study assessed whether the extent of local patients with mild disease, revealing asymmetric (L [ R) cortical atrophy was associated with the extent of cognitive focal cortical thinning mainly involving primary sensory performance using in-depth comprehensive cognitive regions associated with a compensating increased thickness assessment. Of the cortical regions that differed in thick- of visual regions. ness between RRMS patients and healthy subjects, only the LH superior temporal gyrus correlated with GCS, motor References skills, attention, and information processing speed. Though decrease in the gray matter volume of the LH superior Achiron A, Gicquel S, Miron S, Faibel M (2002) Brain MRI lesion temporal gyrus was previously reported in MS patients load quantification in multiple sclerosis: a comparison between (Prinster et al. 2006), our study is the first that evaluated automated multispectral and semi-automated thresholding com- the association between cortical thickness and specific puter-assisted techniques. Magn Reson Imaging 20:713–720 123 85

Brain Struct Funct (2013) 218:943–950 949 Achiron A, Doniger GM, Harel Y, Appleboim-Gavish N, Lavie M, reconstruction algorithms using MRI phantom. Neuroimage Simon ES (2007) Prolonged response times characterize cognitive 31:572–584 performance in multiple sclerosis. Eur J Neurol 14:1102–1108 Maguire EA, Mummery CJ (1999) Differential modulation of a common memory retrieval network revealed by positron emis- Amato MP, Bartolozzi ML, Zipoli V, Portaccio E, Mortilla M, Guidi sion tomography. Hippocampus 9:54–61 L, Siracusa G, Sorbi S, Federico A, De Stefano N (2004) Moll J, de Oliveira-Souza R, Eslinger PJ, Bramati IE, Moura˜o- Neocortical volume decrease in relapsing–remitting MS patients Miranda J, Andreiuolo PA, Pessoa L (2002) The neural with mild cognitive impairment. Neurology 63:89–93 correlates of moral sensitivity: a functional magnetic resonance imaging investigation of basic and moral emotions. J Neurosci Amato MP, Portaccio E, Goretti B, Zipoli V, Battaglini M, Bartolozzi 22:2730–2736 ML, Stromillo ML, Guidi L, Siracusa G, Sorbi S, Federico A, De Morgen K, Sammer G, Courtney SM, Wolters T, Melchior H, Blecker Stefano N (2007) Association of neocortical volume changes CR, Oschmann P, Kaps M, Vaitl D (2006) Evidence for a direct with cognitive deterioration in relapsing–remitting multiple association between cortical atrophy and cognitive impairment sclerosis. Arch Neurol 64:1157–1161 in relapsing–remitting MS. Neuroimage 30:891–898 Pan JW, Krupp LB, Elkins LE, Coyle PK (2001) Cognitive Anderson BJ, Eckburg PB, Relucio KI (2002) Alterations in the dysfunction lateralizes with NAA in multiple sclerosis. Appl thickness of motor cortical subregions after motor-skill learning Neuropsychol 8:155–160 and exercise. Learn Mem 9:1–9 Pelvig DP, Pakkenberg H, Stark AK, Pakkenberg B (2008) Neocor- tical glial cell numbers in human brains. Neurobiol Aging Audoin B, Ibarrola D (2005) Functional MRI study of PASAT in 29:1754–1762 normal subjects. MAGMA 18:96–102 Pozzilli C, Passafiume D, Bernardi S, Pantano P, Incoccia C, Bastianello S, Bozzao L, Lenzi GL, Fieschi C (1991) SPECT, Bozzali M, Cercignani M, Sormani MP, Comi G, Filippi M (2002) MRI and cognitive functions in multiple sclerosis. J Neurol Quantification of brain gray matter damage in different MS Neurosurg Psychiatry 54:110–115 phenotypes by use of diffusion tensor MR imaging. AJNR Am J Pozzilli C, Fieschi C, Perani D, Paulesu E, Comi G, Bastianello S, Neuroradiol 23:985–988 Bernardi S, Bettinardi V, Bozzao L, Canal N (1992) Relationship between corpus callosum atrophy and cerebral metabolic asym- Brunet E, Sarfati Y, Hardy-Bayle MC, Decety J (2000) A PET metries in multiple sclerosis. J Neurol Sci 112:51–57 investigation of the attribution of intentions with a nonverbal Price JL (2007) Definition of the orbital cortex in relation to specific task. Neuroimage 11:157–166 connections with limbic and visceral structures and other cortical regions. Ann N Y Acad Sci 1121:54–71 Calabrese M, Atzori M, Bernardi V, Morra A, Romualdi C, Rinaldi L, Prinster A, Quarantelli M, Orefice G, Lanzillo R, Brunetti A, Mollica McAuliffe MJ, Barachino L, Perini P, Fischl B, Battistin L, C, Salvatore E, Morra VB, Coppola G, Vacca G, Alfano B, Gallo P (2007) Cortical atrophy is relevant in multiple sclerosis Salvatore M (2006) Grey matter loss in relapsing–remitting at clinical onset. J Neurol 254:1212–1220 multiple sclerosis: a voxel-based morphometry study. Neuroim- age 29:859–867 Calabrese M, Rinaldi F, Mattisi I, Grossi P, Favaretto A, Atzori M, Rocca MA, Pagani E, Ghezzi A, Falini A, Zaffaroni M, Colombo B, Bernardi V, Barachino L, Romualdi C, Rinaldi L, Perini P, Gallo Scotti G, Comi G, Filippi M (2003) Functional cortical changes P (2010) Widespread cortical thinning characterizes patients in patients with multiple sclerosis and nonspecific findings on with MS with mild cognitive impairment. Neurology 7:321–328 conventional magnetic resonance imaging scans of the brain. Neuroimage 19:826–836 Chard DT, Griffin CM, Rashid W, Davies GR, Altmann DR, Kapoor Rudick RA, Trapp BD (2009) Gray-matter injury in multiple R, Barker GJ, Thompson AJ, Miller DH (2004) Progressive grey sclerosis. N Engl J Med 361:1505–1506 matter atrophy in clinically early relapsing–remitting multiple Sailer M, Fischl B, Salat D, Tempelmann C, Scho¨nfeld MA, Busa E, sclerosis. Mult Scler 10:387–391 Bodammer N, Heinze HJ, Dale A (2003) Focal thinning of the cerebral cortex in multiple sclerosis. Brain 126:1734–1744 Demerens C, Stankoff B, Logak M, Anglade P, Allinquant B, Schaechter JD, Moore CI, Connell BD, Rosen BR, Dijkhuizen RM Couraud F, Zalc B, Lubetzki C (1996) Induction of myelination (2006) Structural and functional plasticity in the somatosensory in the central nervous system by electrical activity. Proc Natl cortex of chronic stroke patients. Brain 129:2722–2733 Acad Sci USA 93:9887–9892 Sfagos C, Papageorgiou CC, Kosma KK, Kodopadelis E, Uzunoglu NK, Vassilopoulos D, Rabavilas AD (2003) Working memory Draganski B, Gaser C, Busch V, Schuierer G, Bogdahn U, May A deficits in multiple sclerosis: a controlled study with auditory (2004) Neuroplasticity: changes in grey matter induced by P600 correlates. J Neurol Neurosurg Psychiatry 74:1231–1235 training. Nature 427:311–312 Smith AM, Walker LA, Freedman MS, DeMeulemeester C, Hogan MJ, Cameron I (2009) fMRI investigation of disinhibition in Ge Y, Grossman RI, Udupa JK, Babb JS, Kolson DL, McGowan JC cognitively impaired patients with multiple sclerosis. J Neurol (2001) Magnetization transfer ratio histogram analysis of gray Sci 281:58–63 matter in relapsing–remitting multiple sclerosis. AJNR Am J Staffen W, Mair A, Zauner H, Unterrainer J, Niederhofer H, Neuroradiol 22:470–475 Kutzelnigg A, Ritter S, Golaszewski S, Iglseder B, Ladurner G (2002) Cognitive function and fMRI in patients with multiple Geurts JJ, Barkhof F (2008) Grey matter pathology in multiple sclerosis: evidence for compensatory cortical activation during sclerosis. Lancet Neurol 7:841–851 an attention task. Brain 125:1275–1282 Tekok-Kilic A, Benedict RH, Weinstock-Guttman B, Dwyer MG, Geuze E, Westenberg HG, Heinecke A, de Kloet CS, Goebel R, Carone D, Srinivasaraghavan B, Yella V, Abdelrahman N, Vermetten E (2008) Thinner prefrontal cortex in veterans with Munschauer F, Bakshi R, Zivadinov R (2007) Independent posttraumatic stress disorder. Neuroimage 41:675–681 contributions of cortical gray matter atrophy and ventricle Goldberg II, Harel M, Malach R (2006) When the brain loses its self: prefrontal inactivation during sensorimotor processing. Neuron 50:329–339 Haidar H, Soul JS (2006) Measurement of cortical thickness in 3D brain MRI data: validation of the Laplacian method. J Neuroim- aging 16:146–153 Inglese M, Ge Y, Filippi M, Falini A, Grossman RI, Gonen O (2004) Indirect evidence for early widespread gray matter involvement in relapsing–remitting multiple sclerosis. Neuroimage 21:1825– 1829 Kurtzke JF (2008) Historical and clinical perspectives of the expanded disability status scale. Neuroepidemiology 31:1–9 Lee JK, Lee JM, Kim JS, Kim IY, Evans AC, Kim SI (2006) A novel quantitative cross-validation of different cortical surface 123 86

950 Brain Struct Funct (2013) 218:943–950 enlargement for predicting neuropsychological impairment in Vercellino M, Plano F, Votta B, Mutani R, Giordana MT, Cavalla P multiple sclerosis. Neuroimage 36:1294–1300 (2005) Grey matter pathology in multiple sclerosis. J Neuropa- Tiberio M, Chard DT, Altmann DR, Davies G, Griffin CM, Rashid W, thol Exp Neurol 64:1101–1107 Sastre-Garriga J, Thompson AJ, Miller DH (2005) Gray and white matter volume changes in early RRMS: a 2-year longi- Zivadinov R, Minagar A (2009) Evidence for gray matter pathology tudinal study. Neurology 64:1001–1007 in multiple sclerosis: a neuroimaging approach. J Neurol Sci 282:1–4 123 87



Thrombin regulation of synaptic plasticity: implications for physiology and pathology Experimental Neurology | 2013 ‫ פרופ' יואב צ'פמן‬:‫מנחה‬ ‫מנהל מחלקה נוירולוגית‬ [email protected] ‫היוש‬-‫דר' זאב יצקזון‬ ‫אונ' תל אביב‬ ‫השתתף כסטודנט בפרויקט ח״ץ‬ 2011-2013 ‫בין השנים‬ ‫ובהמשך הצטרף כמנחה‬ [email protected] 89

Experimental Neurology 247 (2013) 595–604 Contents lists available at SciVerse ScienceDirect Experimental Neurology journal homepage: www.elsevier.com/locate/yexnr Thrombin regulation of synaptic plasticity: Implications for physiology and pathology Nicola Maggio a,b,⁎, Zeev Itsekson b, Dan Dominissini c,d, Ilan Blatt b,e, Ninette Amariglio c, Gideon Rechavi c,d, David Tanne b,e, Joab Chapman b,e a Talpiot Medical Leadership Program, The Chaim Sheba Medical Center, 52621 Tel HaShomer, Israel b Department of Neurology, The J. Sagol Neuroscience Center, The Chaim Sheba Medical Center, 52621 Tel HaShomer, Israel c Cancer Research Center, The Sheba Medical Center, 52621 Tel HaShomer, Israel d Sackler School of Medicine, Tel Aviv University, 69978 Tel Aviv, Israel e Department of Neurology, Sackler School of Medicine, Tel Aviv University, 69978 Tel Aviv, Israel article info abstract Article history: Thrombin, a serine protease involved in the coagulation cascade has been recently shown to affect neuronal Received 16 November 2012 function following blood–brain barrier breakdown. Several lines of evidence have shown that thrombin may Revised 24 January 2013 exist in the brain parenchyma under normal physiological conditions, yet its role in normal brain functions Accepted 18 February 2013 and synaptic transmission has not been established. In an attempt to shed light on the physiological functions Available online 27 February 2013 of thrombin and Protease Activated Receptor 1 (PAR1) in the brain, we studied the effects of thrombin and a PAR1 agonist on long term potentiation (LTP) in mice hippocampal slices. Surprisingly, different concentra- Keywords: tions of thrombin affect LTP through different molecular routes converging on PAR1. High thrombin concen- Thrombin trations induced an NMDA dependent, slow onset LTP, whereas low concentrations of thrombin promoted a PAR1 VGCCs, mGluR-5 dependent LTP through activated Protein C (aPC). Remarkably, aPC facilitated LTP by activating LTP PAR1 through an Endothelial Protein C Receptor (EPCR)-mediated mechanism which involves intracellular cal- Synaptic plasticity cium stores. These findings reveal a novel mechanism by which PAR1 may regulate the threshold for synaptic Hippocampus plasticity in the hippocampus and provide additional insights into the role of this receptor in normal and path- Extracellular proteases ological conditions. © 2013 Elsevier Inc. All rights reserved. Introduction in the brain under physiological conditions as well as its roles in nor- mal brain functions and synaptic transmission have not yet been Cerebrovascular events induced by either ischemia or hemorrhage completely clarified. lead to blood–brain barrier (BBB) breakdown and exposure of the brain to blood constituents (Yang and Rosenberg, 2011). Among others, PAR1 belongs to a family of seven transmembrane domains, thrombin, a serine protease involved in the coagulation cascade, has G protein-coupled receptors whose activation requires the cleavage of been shown to contribute to the stroke pathology following these condi- a peptide bond at the N-terminal extracellular side which binds the sec- tions (Chen et al., 2012; Wang et al., 2012a, 2012b). High concentrations ond extracellular loop of the same receptor thus activating it (Sokolova of thrombin in the brain also perturb normal physiology by saturating and Reiser, 2008). In the brain PAR1 is expressed both in neurons and as- synaptic plasticity and inducing seizures (Isaeva et al., 2012; Maggio trocytes (Junge et al., 2004; Luo et al., 2007). Its activation has been et al., 2008, 2013). These effects depend on the activation of the throm- shown to modulate synaptic transmission and plasticity (Lee et al., bin receptor, the Protease Activated Receptor 1 (PAR1) and consequent 2007), yet the specific contribution of the astrocytic vs. the neuronal potentiation of NMDA receptor functions (Gingrich et al., 2000; Maggio receptor remains under investigation. In addition, while in peripheral et al., 2008). Previous studies, however, have shown that thrombin may organs PAR1 has been shown to be activated by a pool of proteases, exist in the brain parenchyma in normal physiological conditions e.g. activated Protein C (aPC), leading to different outcomes (Mosnier (Turgeon et al., 2000). Indeed, the mRNAs for both the thrombin et al., 2007), the role of other PAR1 agonists in the brain has not been precursors prothrombin and factor Xa, the enzyme converting pro- fully investigated. thrombin into thrombin, have been detected in several areas of the forebrain (Dihanich et al., 1991; Shikamoto and Morita, 1999). Nev- In an attempt to shed light on the physiological vs. pathological ertheless, both the mechanisms leading to the thrombin production functions of thrombin and PAR1 in the brain, we studied the effects of different concentrations of thrombin and PAR1 agonist (PAR1-AP) ⁎ Corresponding author at: Department of Neurology, The Chaim Sheba Medical Center, on long term potentiation (LTP) using hippocampal slices. Surprisingly, 52621 Tel HaShomer, Israel. Fax: +972 35304752. we found that diverse thrombin concentrations differently regulate the threshold for synaptic plasticity in the hippocampus. These data provide E-mail address: [email protected] (N. Maggio). additional insights into the role of this receptor in normal and patholog- ical conditions. 0014-4886/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.expneurol.2013.02.011 90

596 N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 Methods Endothelial Protein C Receptor (EPCR) (P20, 1:100; Santa Cruz Bio- tech, CA, USA). Hippocampal sections (25 μm) were blocked in 10% Drugs normal serum in 0.01 M PBS/0.25% Triton for 1 h at room tempera- ture (RT). After 48 h incubation at 4 °C with the primary antibody The following drugs were prepared from frozen stocks: thrombin (NeuN, PAR-1, and EPCR with 2% normal serum), sections were exposed (Sigma Aldrich, Rehovot, Israel and Enzyme Research Laboratories, to the appropriate secondary antibody (1:500, DyLight flourophores— Swansea, England); PAR1 agonist (PAR-1AP, SFLLRN, Sigma Aldrich, 594; 488; 633, Thermo-Scientific, Rockford, IL, USA) for 1 h and finally Rehovot, Israel); PAR1 antagonist (SCH79797, Tocris Bioscience, mounted and coverslipped with Flouromount (Southern Biotechnology). Bristol, United Kingdom); plasmin (Sigma Aldrich, Rehovot, Israel); Slides were imaged with a Zeiss LSM 510 confocal microscope and data activated Protein C (aPC, Sigma Aldrich, Rehovot, Israel); inactivated were acquired and analyzed using a computer assisted image analysis Protein C (iPC, Enzyme Research Laboratories, Swansea, England); system. APV (Sigma Aldrich, Rehovot, Israel); nifedipine (Sigma Aldrich, Rehovot, Israel); MPEP (Tocris Bioscience, Bristol, United Kingdom); Intracerebroventricular (ICV) injections thapsigargin (Alomone Labs, Jerusalem, Israel); cyclopiazonic acid (CPA, Alomone Labs, Jerusalem, Israel); α-NAPAP (Sigma Aldrich, Mice were anesthetized with an intraperitoneal (IP) injection of ke- Rehovot, Israel). As the specific activity of the α-thrombin from var- tamine (100 mg/kg) and xylazine (20 mg/kg). The skull was carefully ious vendors ranged between 2700 and 3200 NIH U/mg by compar- exposed, and a small hole was drilled above the right lateral ventricle ison to Lot K of the NIH standard, we estimated the concentration of (2 mm lateral to the midline and 2.5 mm posterior to the bregma). A active α-thrombin that corresponds to 1 U/ml activity, by calculating a rat monoclonal antibody blocking EPCR (R252 clone, Sigma Aldrich, conversion factor using pure α-thrombin (3200 U/mg), as previously Rehovot, Israel) was injected at a concentration of 30 μg/ml (2 mm described (Gingrich et al., 2000). In this lot, the manufacturer reported depth) using a 27-gauge needle attached to a Hamilton syringe. Follow- this protein to be >95% α-thrombin as determined by gel electrophore- ing the slow infusion of 1 μl of antibody solution, the needle was with- sis, hence a solution with 1 U/ml α-thrombin should be 9 nM by molec- drawn, and the skin over the scalp was sutured. Control mice were ular weight for thrombin of 36.7 kDa. For the sake of simplicity, we used injected with a vehicle solution. Hippocampal slices were cut 36 h fol- a conversion factor of 1 U/ml = 10 nM α-thrombin throughout the lowing the procedure. text to estimate the concentration of active α-thrombin (henceforth re- ferred to as thrombin) from various vendors. Results Electrophysiology Thrombin induces slow onset LTP in CA1 Animal handling was approved by the Institutional Animal Care and Thrombin concentrations rise following BBB breakdown (Chen et al., Use Committee, which adheres to the national law, and NIH rules. Briefly, 2010, 2012). Exposure of mice hippocampal slices to high concentration 4–5 months old male C57BL/6 mice were rapidly decapitated and 350 μm of thrombin (1 U/ml thrombin; [Thrombin]high) for 12 min produced a coronal dorsal hippocampal slices were used. Slices were incubated for gradual increase in population EPSP recorded in the stratum radiatum 1.5 h in a humidified, carbogenated (5% CO2 and 95% O2) gas atmosphere of region CA1 of the hippocampus. This gradual change was specific to at 33 ± 1 °C and were perfused with artificial CSF [containing (in mM) the EPSP, because no parallel change was noticed in the presynaptic 124 NaCl, 2 KCl, 26 NaHCO3, 1.24 KH2PO4, 2.5 CaCl2, 2 MgSO4, and 10 glu- volley produced in response to the stimulation (Fig. 1A). The in- cose, pH 7.4] in a standard interface chamber. Recordings were made crease in EPSP slope rose gradually over 40 min of recording, reaching with a glass pipette containing 0.75 M NaCl (4 MΩ) placed in the stratum a plateau level, which was 73% above control (n = 11 slices; 1.73 ± radiatum CA1. Stimulation was evoked using a Master 8 pulse stimulator 0.78; P b 0.01), similar to that induced by tetanic stimulation (HFS) of (A.M.P.I., Jerusalem, Israel) and was delivered through two sets of bipolar the alternate pathway prior to drug application. Application of 2, 3, 5 nichrome electrodes placed on either side of the recording electrode such and 10 U/ml thrombin produced the same effect (data not shown). A that two independent stimulation channels were used for each slice. The comparable, slow-rising, persistent increase in population EPSP was use of two parallel pathways allowed comparison of the effects of drug seen after bath application of 1 μM SFLLRN, a PAR1 receptor agonist application in the same slice (Maggio and Segal, 2007a, 2007b). LTP was ([PAR1-AP]high; n = 11 slices; 1.62 ± 0.071; P b 0.01; Fig. 1B). To induced by high-frequency stimulation consisting of 100 pulses at twice verify that the effect of thrombin was mediated through activation the test intensity, delivered at a frequency of 100 Hz (HFS; 100 Hz, 1 s). of a genuine PAR1 receptor, the selective PAR1 antagonist SCH79797 Before applying the tetanic stimulation, baseline values were recorded (SCH) was used. SCH had no effect on established tetanic LTP (Fig. 1C) at a frequency of 0.033 Hz. Responses were digitized at 5 kHz and stored (n = 11; 1.84 ± 0.079), yet the response to thrombin, tested in the on a computer. Off-line analysis and data acquisition were performed non tetanized pathway, was completely abolished (0.99 ± 0.066). In using Spike 2 software (CED, Cambridge, England). All numerical data order to address whether thrombin-induced slow onset LTP shared a are expressed as mean ± SEM, and EPSP slope changes after tetanic stim- downstream mechanism with the conventional LTP, thrombin was ap- ulation were calculated with respect to baseline. There were no systemat- plied in the presence of the NMDA receptor antagonist 2(R)-APV ic differences in the magnitudes of the baseline responses in the different (APV, 50 μM). Under these conditions, thrombin was unable to produce conditions. All values reported refer to 30 min before tetanic stimulation. LTP (0.97 ± 0.062 at 25 min following thrombin application; n = 11 Unless otherwise indicated, statistical evaluations were performed by ap- slices) (Fig. 1D). Likely, the effect of thrombin was abolished in presence plying Student's t test for paired and unpaired data, as the case may be of its specific inhibitor, α-NAPAP (Suppl. Fig. 1A). These experiments (Origin 8.0). P values of b0.05 were considered a significant difference be- confirm the results we previously obtained in rats (Maggio et al., tween means. 2008), namely that the slow onset LTP induced by [Thrombin]high in- volves NMDA receptors. Immunohistochemistry The effects of thrombin and PAR-1AP are concentration dependent The following antibodies were used for immunodetection: mouse antibodies raised against Neuronal nuclear antigen (NeuN) (1:100; Thrombin produces a variety of effects in the brain being detri- Millipore, Billerica, MA, USA); rabbit antibodies to Protease Activated mental at high concentrations and protective at low concentrations Receptor-1 (PAR1) (1:50; Abcam, Cambridge, UK); goat antibodies to (Hua et al., 2009; Striggow et al., 2000). In order to test whether the 91

N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 597 Fig. 1. [Thrombin]high induces slow onset, NMDA-dependent LTP in hippocampal slices. (A, B) Short application (12 min) of thrombin (1 U/ml, [Thrombin]high; A) as well as of PAR1-AP (1 μM, [PAR1-AP]high; B) produces a gradual increase in population EPSP in stratum radiatum CA1 without affecting population synaptic volleys (traces on top). (C) Thrombin induced slow onset LTP is blocked by the PAR1 antagonist SCH79797 (1 μM). (D) The NMDA antagonist APV (50 μM) totally blocks thrombin induced slow onset LTP. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. Fig. 2. Thrombin and PAR1-AP effects on LTP are concentration dependent. (A) Short application (15 min) of thrombin (100 mU/ml, [Thrombin]low) does not induce a slow onset LTP. As well, it does not affect the level of a tetanus induced LTP evoked at the second pathway 20 min after drug removal. (B) Short application (15 min) of PAR1-AP (100 nM, [PAR1-AP]low) does not induce a slow onset LTP, however it enhances the height of a tetanus induced LTP evoked at the second pathway 20 min after drug removal. (C) The PAR1 antagonist SCH 79797 blocks the [PAR1-AP]low-mediated enhancement of LTP. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. 92

598 N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 effects of thrombin on LTP are concentration dependent, we exposed and 2.21 ± 0.065 at 75 min of recordings for the first and the second the hippocampal slices to 100 mU/ml thrombin ([Thrombin]low) for pathway respectively; P b 0.001; n = 12 slices; Fig. 3D). Application 15 min. In these experiments, thrombin was bath applied following of 100 nM aPC led to the same result (Fig. 3E). PAR1 activation was in the delivery of HFS at the first control pathway. At this concentration, charge of this effect as aPC in presence of SCH was not able to induce thrombin did not affect the established LTP, yet it failed to induce a slow an enhanced LTP (1.83 ± 0.083 and 1.84 ± 0.076 at 75 min of re- onset potentiation of EPSP at the second pathway (1.01 ± 0.068 at cordings for the first and the second pathway respectively; P = 0.54; 50 min of recordings; Fig. 2A). Furthermore, HFS at the second pathway, n = 12 slices; Fig. 3F). These experiments indicate that aPC mimics the 20 min after drug removal, evoked an LTP of similar level to the one in the effects of [PAR1-AP]low and causes enhanced LTP in stratum radiatum first, control pathway (n = 12 slices; 1.66 ± 0.059 vs. 1.64 ± 0.063, re- of CA1. spectively; P = 0.43; Fig. 2A). Thrombin concentrations ranging from 500 mU/ml to 1 mU/ml led to similar results (data not shown). Interest- The Endothelial Protein C Receptor (EPCR) is expressed in the hippocampus, ingly, however, exposure of slices to a lower concentration of PAR1-AP colocalizes with PAR1 and mediates the effects of aPC resulted in a novel, different phenomenon. Following the induction of LTP by HFS at the first pathway, the slices were exposed to 100 nM aPC is known to activate PAR1 upon binding its own receptor EPCR PAR1-AP ([PAR1-AP]low) for 15 min. [PAR1-AP]low did not enhance syn- (Rezaie, 2005; Riewald and Ruf, 2005). Recently, studies addressing a aptic transmission per se, yet 20 min after drug withdrawal an enhanced possible neuroprotective role of aPC following glutamate toxicity LTP could be evoked by HFS at the second pathway (n = 12 slices; have shown that EPCR is expressed in cortical and hippocampal neu- 2.23 ± 0.086; P b 0.001 over LTP at the first pathway; Fig. 2B). Similar rons in cell cultures (Gorbacheva et al., 2009). Besides, to our knowl- data were obtained using concentrations of PAR1-AP ranging from edge, the exact location of this receptor in the hippocampus has not 200 nM to 10 nM (data not shown). In order to verify that the genuine been studied so far. As a first attempt to test whether EPCR may be in- activation of PAR1 is required to produce an enhancement of LTP, volved in the aPC induced enhancement of LTP, we qualitatively [PAR1-AP]low was bath applied in presence of SCH. Remarkably, in this mapped its expression in the hippocampus using immunofluores- condition HFS at the second pathway generated a normal level of LTP cence. EPCR was found to be widely distributed in the whole hippo- comparable to the one obtained in the control pathway (n = 12 slices; campus (Fig. 4A). In CA1 it was localized both in neurons (Fig. 4B) 1.79 ± 0.77 vs. 1.75 ± 0.084, respectively; P = 0.39; Fig. 2C). These ex- and astrocytes (Fig. 4C). EPCR has been reported to activate PAR1 periments show that exposure to [PAR1-AP]low enhances LTP through a through a physical interaction between the two receptors (Riewald PAR1 mediated mechanism. This effect is specific to [PAR1-AP]low and is and Ruf, 2005), and indeed a confocal colocalization study found not shared by [Thrombin]low. that EPCR and PAR1 were highly co-expressed in the same CA1 cells (Fig. 4D). A second step towards the understanding of the role of Activated Protein C (aPC) and not plasmin mimics the effects of [PAR1-AP]low EPCR in aPC-mediated phenomena consists of testing the ability of on LTP aPC to enhance LTP in presence of EPCR blockers. Taking into account that a pharmacological approach to reliably block these receptors has PAR1-AP consists of a short amino-acid sequence and its mecha- not been developed as yet, we tried to tackle this issue by ICV injec- nism of PAR1 activation (i.e. direct interaction with its ligand bind- tions of EPCR blocking antibodies (R-252) and consequent use of hip- ing site) differs from that of endogenous proteases. In this respect, pocampal slices from these animals. In these experiments, slices [PAR1-AP]low-mediated enhancement of LTP could be due to a pecu- were cut 36 h after recovery from the ICV injection procedures. EPCR liar property of the drug which at specific concentrations may act as blockade did not influence the ability of the slice to undergo normal a general agonist, thus causing this effect. In order to exclude this LTP. Applying HFS to the first pathway resulted in a full blown LTP possibility, we tested whether other endogenous PAR1 agonists could (1.68 ± 0.064; n = 13 slices; Fig. 5A). Remarkably, however, HFS de- mimic the effects of [PAR1-AP]low on LTP. livered at the second pathway in presence of 100 nM aPC evoked a po- tentiation similar to the one in the control pathway (1.70 ± 0.066; Plasmin has been shown to activate PAR1 both in peripheral tissues P = 0.49; n = 13 slices; Fig. 5A), therefore inhibiting the aPC's effect. and in the brain where it potentiates NMDA receptors currents Differently, EPCR block did not impair the effects of PAR1-AP. In this set- (Mannaioni et al., 2008). In hippocampal slices, application of 1 U/ml ting, [PAR1-AP]low still enhanced LTP compared to control conditions plasmin ([Plasmin]high) for 15 min induced a gradual increase in popu- (1.70 ± 0.065 and 2.18 ± 0.074 at 75 min of recordings for the lation EPSP at stratum radiatum CA1 over 40 min of recording. This po- first and the second pathway respectively; P b 0.001; n = 13 slices; tentiation reached a plateau level (1.75 ± 0.059; n = 11 slices; Fig. 3A) Fig. 5B). Likely, [PAR1-AP]high produced a slow onset LTP in EPCR similar to the one HFS induced in the alternate path prior to drug appli- blockade condition. These experiments indicate that EPCR mediates cation (1.81 ± 0.053; P = 0.46; Fig. 3A). Like [Thrombin]low, exposure the effects of aPC acting most likely upstream to PAR1. to 100 mU/ml plasmin ([Plasmin]low) for 15 min neither generated a slow onset LTP nor enhanced LTP evoked by HFS at a second pathway aPC induces NMDA independent, VGCC and mGluR-5 dependent LTP following drug removal (1.77 ± 0.073; n = 11 slices; Fig. 3B). Finally, as in the case of [Thrombin]high, the [Plasmin]high-mediated slow onset In order to understand the mechanisms leading to aPC mediated LTP was due to PAR1 activation as it was completely blocked in pres- LTP, slices were treated with aPC in presence of the NMDA blocker ence of SCH (1.02 ± 0.076; n = 11 slices; Fig. 3C). These experiments APV. 50 μM APV were bath applied right after HFS on the first pathway. show that plasmin and thrombin share similar outcomes on LTP and 15 min later 100 nM aPC was added to the medium. Interestingly, do not mimic the effects of [PAR1-AP]low. under NMDA blocking condition aPC was still able to evoke an en- hanced LTP when HFS was delivered at the second path (1.83 ± 0.076 In peripheral tissues as well as in the blood coagulation cascade, it and 2.21 ± 0.074; at 75 min of recordings for the first and the second has been shown that activation of PAR1 by aPC results in different, pathway respectively; P b 0.001; n = 12 slices; Fig. 6A). Considerably, somewhat opposite effects than those of thrombin induced PAR1 ac- APV was able to block LTP in the first, control pathway, but did not im- tivation (Feistritzer and Riewald, 2005; Ludeman et al., 2005; Riewald pair the aPC enhanced LTP (0.99 ± 0.071 and 2.22 ± 0.078; at 75 min and Ruf, 2005). Hence, we tested whether aPC could mediate the en- of recordings for the first and the second pathway respectively; hancement of LTP induced by [PAR1-AP]low. Following the induction P b 0.001; n = 9 slices; Fig. 6B). We then tested whether two addi- of LTP in the first pathway, 1 μM aPC was bath applied for 15 min tional key players in synaptic plasticity such as VGCCs or mGluR-5 (Fig. 3D). Surprisingly, aPC did not increase basal synaptic transmis- would have a role in mediating aPC induced LTP. Surprisingly both sion, however an enhanced LTP was evoked when HFS was delivered at the second pathway after the washout of the drug (1.81 ± 0.074 93

N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 599 Fig. 3. aPC shares the effects of [PAR1-AP]low on LTP. (A) Short application (15 min) of plasmin (1 U/ml, [Plasmin]high) induces a gradual increase in population EPSP in stratum radiatum CA1 without affecting population synaptic volleys (traces on top). (B) Short application (15 min) of plasmin (100 mU/ml) neither induces a slow onset LTP nor affects the level of a tetanus induced LTP evoked at the second pathway 20 min after drug removal. (C) The PAR1 antagonist SCH 79797 blocks the [Plasmin]high-slow onset LTP. (D, E) Short application (15 min) of aPC (1 μM D; 100 nM E) does not induce a slow onset LTP, however it enhances the level of tetanus induced LTP evoked at the second pathway 20 min following drug removal. (F) The PAR1 antagonist SCH 79797 blocks the aPC mediated enhancement of LTP. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. VGCC and mGluR-5 antagonists blocked the ability of aPC to induce en- decrease in the established synaptic potentiation was observed follow- hanced LTP. In one experiment, bath applying aPC in presence of 20 μM ing the exposure of the slice to this drug (1.66 ± 0.076 at 80 min com- nifedipine led to a normal potentiation following HFS delivery to a sec- pared to 1.87 ± 0.083 right after HFS; P b 0.001; Fig. 7A). We further ond path (1.82 ± 0.081 and 1.78 ± 0.082; at 75 min of recordings for explored these observations using cyclopiazonic acid (CPA), a blocker the first and the second pathway respectively; P = 0.34; n = 12 slices; of endoplasmic reticulum Ca2+ ATPases. Bath application of 20 μM Fig. 6C). Similarly, a normal level of potentiation was evoked following CPA blocked the potentiation facilitated by aPC (1.87 ± 0.086 and HFS at the second pathway when aPC was bath applied in pres- 1.52 ± 0.082 at the onset following HFS for the first and the second ence of 10 μM of the mGluR-5 antagonist MPEP (1.79 ± 0.085 pathway, respectively; n = 14 slices, P b 0.001; Fig. 7B). In addition, and 1.84 ± 0.082 at 75 min of recordings for the first and second similar to thapsigargin, CPA impaired the already established LTP at pathway respectively; P = 0.31; n = 12 slices; Fig. 6D). These ex- the control pathway. The notable advantage of CPA is the possibility periments show that aPC induces an NMDA-independent, VGCC to wash it out from the medium. This allows the refill of intracellular and mGluR-5-dependent LTP. calcium stores and the recovery of their function. Remarkably, deliver- ing HFS in presence of CPA and aPC resulted in a potentiation of 55% Calcium stores are involved in aPC mediated LTP over baseline (Fig. 7C). However, HFS applied in presence of aPC follow- ing CPA removal facilitated LTP (1.47 ± 0.057 and 2.19 ± 0.082 at aPC has been shown to modulate calcium release from intracellu- 150 min of recording for the first and the second pathway respectively; lar stores through an EPCR–PAR1 dependent mechanism (Domotor P b 0.001; Fig. 7C). These experiments indicate that calcium stores are et al., 2003). In order to test whether aPC mediated LTP is dependent involved in aPC mediated LTP and confirm the role of these organelles on calcium stores, we bath applied aPC with thapsigargin, an inhibitor both in the establishment and maintenance of LTP in normal conditions of endoplasmatic reticulum Ca2+ ATPases (Maggio and Segal, 2007a, (Maggio and Segal, 2007a, 2007b). 2007b). Remarkably, aPC in presence of 1 μM thapsigargin was not able to facilitate LTP; HFS at the second pathway evoked a lower Thrombin enhances LTP by promoting the formation of aPC LTP than that of the control path (1.87 ± 0.083 and 1.53 ± 0.086 at the onset following HFS for the first and the second pathway, respec- In the coagulation cascade, thrombin contributes to the generation tively; n = 14 slices, P b 0.001; Fig. 7A). In addition, thapsigargin of aPC (Griffin et al., 2007). This initiates a feedback loop that ulti- even affected the LTP evoked at the first, control pathway as a slow mately results in the attenuation of the thrombin signaling (Griffin 94

600 N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 Fig. 4. The aPC receptor, EPCR, is expressed on neurons, on astrocytes and colocalizes with PAR1. (A) EPCR is widely expressed throughout the hippocampal fields (NeuN, red, EPCR, green; 10×) and is localized in neurons (B, NeuN, red, EPCR, green; 40×) and astrocytes (C, GFAP, red, EPCR, green; 40×). (D) PAR1 (red) and EPCR (green) colocalize in CA1 (40×). The white arrow in A indicates the area of CA1 where the pictures in B, C, and D were taken. et al., 2007). In this respect, we assumed that [Thrombin]low could as activation in the perfusing solution. In these experiments, HFS to the sec- well promote the formation of aPC in hippocampal slices. Indeed, if this ond channel in presence of [Thrombin]low and iPC led to facilitation of might be the case, [Thrombin]low could facilitate LTP through an aPC me- LTP compared to control conditions (1.77 ± 0.085 and 2.28 ± 0.065 diated mechanism. In order to test this hypothesis, we exposed slices to at 20 min following delivery of HFS at the first and the second pathway inactivated Protein C (iPC), a compound that can be converted to aPC in respectively; P b 0.001; n = 14; Fig. 8C). Interestingly, [Thrombin]low in presence of thrombin. In the first set of experiments, [Thrombin]low did presence of α-NAPAP was unable to produce this result (Suppl. Fig. 1B). not show any notable effect on LTP: application of this drug following This effect is peculiar to [Thrombin]low and not shared by [Thrombin]high. HFS at the control pathway did not result in a facilitation of LTP when a Notably, perfusing [Thrombin]high in presence of APV neither evoked an second HFS was delivered at the second pathway (1.76 ± 0.075 and enhanced LTP (Suppl. Fig. 2) nor facilitated LTP following simultaneous 1.77 ± 0.086 at 20 min following delivery of HFS at the first and the sec- perfusion with iPC. These experiments demonstrate that [Thrombin]low ond pathway respectively; P = 0.62; n = 14; Fig. 8A). Then, in a differ- is likely to facilitate LTP by promoting aPC formation in hippocampal ent set of slices, we bath applied 1 μM iPC following HFS at the first slices. pathway. In these conditions, HFS delivered at the second pathway also produced a potentiation of similar levels to the one in the control path Discussion (1.75 ± 0.083 and 1.74 ± 0.076 at 20 min following delivery of HFS at the first and the second pathway respectively; P = 0.66; n = 14; In this study, we addressed the role of thrombin in modulating LTP. Fig. 8B). Strikingly, however exposure of slices to [Thrombin]low and Interestingly, we observed that different concentrations of thrombin led iPC led to a different outcome. Here different solutions of [Thrombin]low to diverse outcomes (Fig. 9). [Thrombin]high induced an NMDA depen- and iPC were simultaneously perfused in the slice to avoid possible iPC dent, slow onset LTP while [Thrombin]low promoted a VGCC, mGluR-5 95

N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 601 Fig. 5. EPCR blockade prevents the aPC-mediated, but not the [PAR1-AP]low-mediated enhancement of LTP. In slices from anti-EPCR ICV-injected animals application of aPC does not enhance LTP (A) Contrarily, in this setting both [PAR1-AP]high (B) and [PAR1-AP]low (C) are able to induce a slow onset LTP or to enhance the level of a tetanus induced LTP evoked at the second pathway, respectively. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. dependent LTP through the activation of aPC (Fig. 9). aPC facilitated LTP The highest concentrations of thrombin in the brain occur during by activating PAR1 through an EPCR-mediated mechanism and involve- brain hemorrhages (Gong et al., 2008; Hua et al., 2007, 2009) howev- ment of intracellular calcium stores. er high levels of thrombin have been reported following ischemic Fig. 6. aPC induces an NMDA-independent, VGCC and mGluR5-dependent LTP. (A) The NMDA antagonist APV (50 μM) does not block the aPC-induced LTP. (B) APV blocks LTP in the control pathway, but it does not affect aPC-induced LTP. (C) The voltage gated calcium channels blocker nifedipine (20 μM) inhibits the aPC-induced LTP. (D) The mGluR5 blocker MPEP (10 μM) hampers the aPC-induced LTP. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Up- ward arrows indicate the time of HFS. 96

602 N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 Fig. 7. Calcium stores are involved in aPC induced LTP. (A, B) Both the irreversible (thapsigargin, 1 μM; A) and the reversible (cyclopiazonic acid, CPA, 20 μM; B) SERCA pump in- hibitors prevent the aPC-induced LTP. (C) HFS in presence of aPC after CPA washout restores aPC-induced LTP. Averaged EPSP are plotted versus time. Representative traces at in- dicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. stroke (Chen et al., 2010, 2012). In these settings, BBB opening and rup- tightly regulated by multiple processes involving endogenous prote- ture of brain endothelial cells activate the coagulation cascade leading to ases (Siller-Matula et al., 2011). Definitely, a similar system might intracerebral production of thrombin (Chodobski et al., 2012). Interest- exist in the brain in order to balance thrombin concentrations at the ingly, neurological deficits in acute ischemic stroke animal models may synapse during physiological events. Therefore, depending on the avail- be attenuated following thrombin blockade (Chen et al., 2010, 2012; ability of this system, different concentrations of thrombin might arise Karabiyikoglu et al., 2004). Our present data as well as our previous re- at the synapse following a determined synaptic event and hence differ- port (Maggio et al., 2008) show that [Thrombin]high evokes an NMDA ently influence LTP (Komai et al., 2000; Maggio et al., 2008). dependent LTP which saturates the ability of a neuronal network to un- dergo further NMDA dependent potentiation. This result might explain In this study, we also showed the unique role of aPC in regulating the inability to recover from a cerebrovascular trauma: the saturation LTP. Interestingly, the effects of aPC are concentration independent, of synaptic connectivity by thrombin-activated mechanisms does not therefore the presence of aPC at the synaptic cleft might shift by itself allow the brain to use LTP-like plastic processes for either acquisition the threshold of synaptic plasticity towards novel LTP facilitation in a of new “memories” or adaptation to new motor plans after the insult previously potentiated network. While the aPC receptor EPCR seems (Maggio et al., 2008). Definitely, additional experiments are required to be widely expressed in the hippocampus, it is not entirely clear to explore the duration of thrombin action in the brain as well as the whether in the brain the production of aPC exclusively depends on ways to overcome thrombin-related cognitive deficits. In this respect, the thrombin-mediated conversion of iPC or it may be locally synthe- it could be interesting to follow on whether patients discharged on sized de novo. In the blood stream, PC seems to circulate in its inactive, dabigatran, a new thrombin direct inhibitor, would have better long zymogen form and the switch into its active state occurs following its in- term cognitive outcomes following stroke than other patients released teraction with thrombin (Griffin et al., 2007). Clarifying this issue in the on alternative anticoagulants. brain might lead to important findings for further understanding of the mechanisms regulating synaptic plasticity. However, as the case may There are suggestions that thrombin may be produced during nor- be, this study's, our previous (Maggio et al., 2008) as well as other evi- mal synaptic transmission (Turgeon et al., 2000), yet its exact amount dences (Dihanich et al., 1991; Gingrich et al., 2000; Shikamoto and at the synaptic cleft under physiological conditions is unknown. In the Morita, 1999) point towards a role of clotting factors in the regulation blood stream, thrombin has a short half life and its concentration is of synaptic plasticity. Lately, aPC has received attention due to its Fig. 8. [Thrombin]low enhances LTP through aPC. (A, B) [Thrombin]low (A) as well as inactivated Protein C (iPC; B) neither induce a slow onset LTP nor affect the level of a tetanus induced LTP evoked at the second pathway. (C) [Thrombin]low in presence of iPC enhances the level of a tetanus induced LTP evoked at the second pathway. Averaged EPSP are plotted versus time. Representative traces at indicated times (a, b) are shown on top of each section. Upward arrows indicate the time of HFS. 97

N. Maggio et al. / Experimental Neurology 247 (2013) 595–604 603 and provides additional insights on the role of this receptor in phys- iological vs. pathological conditions. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.expneurol.2013.02.011. Conflict of interest The authors declare no competing financial interests. Fig. 9. Diverse thrombin concentrations differently regulate LTP. [Thrombin]high in- Acknowledgments duces an NMDA dependent, slow onset LTP while [Thrombin]low promotes a VGCC, mGluR dependent LTP through the activation of aPC. aPC facilitates LTP by activating The authors wish to acknowledge Dr. Eduard Korkotian for helping with the imaging of the histological sections and Dr. Menahem Segal PAR-1 through an EPCR-mediated mechanism and involvement of intracellular calci- for commenting on the early drafts of the manuscript. NM is the recip- ient of a Talpiot Fellowship at the Sheba Medical Center. The funding um stores. agency did not have a role in the study design; collection, analysis, and interpretation of data; in the writing of the report; and in the deci- sion to submit the paper for publication. cytoprotective and anticoagulant properties which may be of benefit References in stroke therapy (Griffin et al., 2012; Wang et al., 2012a, 2012b; Zlokovic and Griffin, 2011). Our data further promote these positive Chen, B., Cheng, Q., Yang, K., Lyden, P.D., 2010. Thrombin mediates severe neurovascular properties of aPC. Precisely, the evidence that in the brain aPC en- injury during ischemia. Stroke 41, 2348–2352. hances LTP through alternative pathways requiring VGCCs and mGLUR-5 may result in better cognitive outcomes in cases where Chen, B., Friedman, B., Whitney, M.A., Winkle, J.A., Lei, I.F., Olson, E.S., Cheng, Q., Pereira, B., NMDA receptors are saturated by thrombin. Definitely, more data Zhao, L., Tsien, R.Y., Lyden, P.D., 2012. Thrombin activity associated with neuronal on the role of aPC in stroke are needed to confirm this hypothesis. damage during acute focal ischemia. J. Neurosci. 32, 7622–7631. The evidence that both thrombin and aPC may exert opposite ef- Chodobski, A., Zink, B.J., Szmydynger-Chodobska, J., 2012. Blood–brain barrier patho- fects acting on the same receptor, i.e. PAR1, is puzzling indeed. In en- physiology in traumatic brain injury. Transl. Stroke Res. 2, 492–516. dothelial cells, it has been shown that PAR1 activation either by thrombin or aPC may lead to different, opposite outcomes due to Dihanich, M., Kaser, M., Reinhard, E., Cunningham, D., Monard, D., 1991. Prothrombin the stimulation of Gq/G12/13 protein cascade in the former case and mRNA is expressed by cells of the nervous system. Neuron 6, 575–581. Gi in the latter (Riewald and Ruf, 2005). There might be a similar sit- uation in the brain where PAR1 has been shown to interact with Domotor, E., Benzakour, O., Griffin, J.H., Yule, D., Fukudome, K., Zlokovic, B.V., 2003. Ac- multiple G proteins (McCoy et al., 2012). An additional hypothesis tivated protein C alters cytosolic calcium flux in human brain endothelium via to explain these results may take into account the efficacy of differ- binding to endothelial protein C receptor and activation of protease activated ent concentrations of PAR1-AP in producing diverse effects. Specifi- receptor-1. Blood 101, 4797–4801. cally, it might be that two types of PAR1 bearing different affinities for the PAR1-AP ligand might exist at the synapse. A low affinity PAR1 Feistritzer, C., Riewald, M., 2005. Endothelial barrier protection by activated protein C could directly be activated by thrombin and mediate the [Thrombin]high through PAR1-dependent sphingosine 1-phosphate receptor-1 crossactivation. Blood induced slow onset LTP while a high affinity receptor, not promptly 105, 3178–3184. cleaved by thrombin, might be in charge of the aPC–EPCR effects. These two different receptors could be linked to different G proteins Gingrich, M.B., Junge, C.E., Lyuboslavsky, P., Traynelis, S.F., 2000. Potentiation of NMDA and be expressed in diverse cell types (astrocytes vs. neurons). Addi- receptor function by the serine protease thrombin. J. Neurosci. 20, 4582–4595. tional pharmacological and molecular studies are needed in order to validate this hypothesis. Gong, Y., Xi, G., Hu, H., Gu, Y., Huang, F., Keep, R.F., Hua, Y., 2008. Increase in brain thrombin activity after experimental intracerebral hemorrhage. Acta Neurochir. In the present study, we have not clearly addressed whether the Suppl. 105, 47–50. effects of thrombin and aPC are mediated by the neuronal and/or as- trocytic PAR1. EPCR localization by itself does not add any further hint Gorbacheva, L., Davidova, O., Sokolova, E., Ishiwata, S., Pinelis, V., Strukova, S., Reiser, G., to this quest as it seems that both EPCR and PAR1 are colocalized in 2009. Endothelial protein C receptor is expressed in rat cortical and hippocampal the same cells in the whole hippocampus. Nevertheless, the cellular neurons and is necessary for protective effect of activated protein C at glutamate localization of PAR1 signaling seems to be a fundamental issue in excitotoxicity. J. Neurochem. 111, 967–975. order to fully understand the contribution of this receptor to synaptic plasticity. On the one hand, PAR1 knock out animals might not be an Griffin, J.H., Fernandez, J.A., Gale, A.J., Mosnier, L.O., 2007. Activated protein C. J. Thromb. appropriate tool to utilize in solving this issue as signals from both Haemost. 5 (Suppl. 1), 73–80. cell types are being deleted in these animals. On the other hand, the use of molecular techniques that will downregulate PAR1 expression Griffin, J.H., Zlokovic, B.V., Mosnier, L.O., 2012. Protein C anticoagulant and cytoprotective in specific cell types is more promising and might lead to new funda- pathways. Int. J. Hematol. 95, 333–345. mental insights into the role of this receptor in synaptic plasticity. Hua, Y., Keep, R.F., Hoff, J.T., Xi, G., 2007. Brain injury after intracerebral hemorrhage: In summary, we have shown that thrombin may differently regulate the role of thrombin and iron. Stroke 38, 759–762. synaptic plasticity based on its concentration levels at the synaptic cleft. We believe this represents a novel mechanism by which PAR1 may Hua, Y., Keep, R.F., Gu, Y., Xi, G., 2009. Thrombin and brain recovery after intracerebral regulate the threshold for synaptic plasticity in the hippocampus hemorrhage. Stroke 40, S88–S89. Isaeva, E., Hernan, A., Isaev, D., Holmes, G.L., 2012. Thrombin facilitates seizures through activation of persistent sodium current. Ann. Neurol. 72, 192–198. Junge, C.E., Lee, C.J., Hubbard, K.B., Zhang, Z., Olson, J.J., Hepler, J.R., Brat, D.J., Traynelis, S.F., 2004. Protease-activated receptor-1 in human brain: localization and functional ex- pression in astrocytes. Exp. Neurol. 188, 94–103. Karabiyikoglu, M., Hua, Y., Keep, R.F., Ennis, S.R., Xi, G., 2004. Intracerebral hirudin in- jection attenuates ischemic damage and neurologic deficits without altering local cerebral blood flow. J. Cereb. Blood Flow Metab. 24, 159–166. Komai, S., Matsuyama, T., Matsumoto, K., Kato, K., Kobayashi, M., Imamura, K., Yoshida, S., Ugawa, S., Shiosaka, S., 2000. Neuropsin regulates an early phase of schaffer- collateral long term potentiation in the murine hippocampus. Eur. J. Neurosci. 12 (4), 1479–1486. Lee, C.J., Mannaioni, G., Yuan, H., Woo, D.H., Gingrich, M.B., Traynelis, S.F., 2007. Astro- cytic control of synaptic NMDA receptors. J. Physiol. 581, 1057–1081. Ludeman, M.J., Kataoka, H., Srinivasan, Y., Esmon, N.L., Esmon, C.T., Coughlin, S.R., 2005. PAR1 cleavage and signaling in response to activated protein C and thrombin. J. Biol. Chem. 280, 13122–13128. Luo, W., Wang, Y., Reiser, G., 2007. Protease-activated receptors in the brain: receptor expression, activation, and functions in neurodegeneration and neuroprotection. Brain Res. Rev. 56, 331–345. Maggio, N., Segal, M., 2007a. Striking variations in corticosteroid modulation of long- term potentiation along the septotemporal axis of the hippocampus. J. Neurosci. 27, 5757–5765. Maggio, N., Segal, M., 2007b. Unique regulation of long term potentiation in the rat ventral hippocampus. Hippocampus 17, 10–25. 98