Volume 5, Number 4, July 2018 Neurology.org/NN · Academy Officers RalphL.Sacco,MD,MS,FAAN,President JamesC.Stevens,MD,FAAN,PresidentElect AnnH.Tilton,MD,FAAN,VicePresident CarlayneE.Jackson,

Volume 5, Number 4, July 2018Neurology.org/NN

A peer-reviewed clinical and translational neurology open access journal

ARTICLE

Anti-NMDA receptor encephalitis and nonencephalitic HSV-1 infection e458

ARTICLE

Monitoring B-cell repopulation aft er depletion therapy in neurologic patients e463

ARTICLE

Neopterin is associated with hippocampal subfi eld volumes and cognition in HIV e467

ARTICLE

Mononuclear cell transcriptome changes associated with dimethyl fumarate in MS e470

Academy OfficersRalph L. Sacco, MD, MS, FAAN, PresidentJames C. Stevens, MD, FAAN, President ElectAnn H. Tilton, MD, FAAN, Vice PresidentCarlayne E. Jackson, MD, FAAN, SecretaryJanis M. Miyasaki, MD, MEd, FRCPC, FAAN, TreasurerTerrence L. Cascino, MD, FAAN, Past President

Executive Office, American Academy of NeurologyCatherine M. Rydell, CAE, Executive Director/CEO201 Chicago AveMinneapolis, MN 55415Tel: 612-928-6100

Editorial OfficePatricia K. Baskin, MS, Executive EditorKathleen M. Pieper, Senior Managing Editor, NeurologySharon L. Quimby, Managing Editor, Neurology® Clinical PracticeMorgan S. Sorenson, Managing Editor, Neurology® Neuroimmunology& Neuroinflammation

Lee Ann Kleffman, Managing Editor, Neurology® GeneticsCynthia S. Abair, MA, Senior Graphics EditorAndrea R. Rahkola, Production Editor, NeurologyRobert J. Witherow, Senior Editorial AssociateKaren Skaja, Senior Editorial AssociateKaitlyn Aman Ramm, Editorial AssistantKristen Swendsrud, Editorial AssistantAndrea Willgohs, Editorial Assistant

PublisherWolters KluwerBaltimore, MD

Publishing StaffKim Jansen, Executive PublisherJessica Heise, Production Team Leader, Neurology JournalsMegen Miller, Production EditorStacy Drossner, Production Associate

Neurology® is a registered trademark of the American Academy of Neurology(registration valid in the United States).

Neurology® Neuroimmunology & Neuroinflammation (eISSN 2332-7812) is anopen access journal published online for the American Academy of Neurology,201 Chicago Ave, Minneapolis, MN 55415, by Wolters Kluwer Health, Inc. at14700 Citicorp Drive, Bldg 3, Hagerstown, MD 21742. Business offices arelocated at Two Commerce Square, 2001 Market St., Philadelphia, PA 19103.Production offices are located at 351 West Camden Street, Baltimore, MD21201-2436. © 2018 American Academy of Neurology.

Neurology® Neuroimmunology & Neuroinflammation is an official journal of theAmerican Academy of Neurology. Journal website: Neurology.org/NN, AANwebsite: AAN.com

Copyright and Permission Information: Please go to the journal website(www.neurology.org/nn) and click the Permissions tab for the relevant article.Alternatively, send an email to [email protected]. Generalinformation about permissions can be found here: https://www.lww.com/journal-permission.

Disclaimer: Opinions expressed by the authors and advertisers are notnecessarily those of the American Academy of Neurology, its affiliates, or ofthe Publisher. The American Academy of Neurology, its affiliates, and thePublisher disclaim any liability to any party for the accuracy, completeness,efficacy, or availability of the material contained in this publication (includingdrug dosages) or for any damages arising out of the use or non-use of any of thematerial contained in this publication.

Advertising Sales Representatives: Wolters Kluwer, 333 Seventh Avenue, NewYork, NY 10001. Contacts: Eileen Henry, tel: 732-778-2261, fax: 973-215-2485,[email protected] and in Europe: Avia Potashnik, WoltersKluwer, tel: +44 207 981 0722; +44 7919 397 933 or e-mail: [email protected].

Careers & Events: Monique McLaughlin, Wolters Kluwer, Two CommerceSquare, 2001 Market Street, Philadelphia, PA 19103, tel: (215)-521- 8468, fax:215-521-8801; [email protected].

Reprints: Meredith Edelman, Commercial Reprint Sales, Wolters Kluwer,Two Commerce Square, 2001 Market Street, Philadelphia, PA 19103, tel:215-356-2721; [email protected]; [email protected].

Special projects: US & Canada: Alan Moore, Wolters Kluwer, Two CommerceSquare, 2001 Market Street, Philadelphia, PA 19103, tel: 215-521-8638,[email protected]. International: Andrew Wible, Senior Manager,Rights, Licensing, and Partnerships, Wolters Kluwer, [email protected].

http://Neurology.org/NN

http://AAN.com

https://www.lww.com/journal-permission

https://www.lww.com/journal-permission

mailto:[email protected]


mailto:reprintsolutions@wolterskluwer<?show $132#?>.com

mailto:reprintsolutions@wolterskluwer<?show $132#?>.com




A peer-reviewed clinical and translational neurology open access journal Neurology.org/NN

Neurology® Neuroimmunology & Neuroinflammation

EditorJosep O. Dalmau, MD, PhD

Deputy EditorScott S. Zamvil, MD, PhD, FAAN

Associate EditorsMichael Heneka, MDDennis Kolson, MD, PhDFriedmann Paul, MDLan Zhou, MD, PhD

Editorial BoardRohit Aggarwal, MD, MSSergio E. Baranzini, PhDSusanne Benseler, MDJeffrey L. Bennett, MD, PhD, FAANSeverine Boillee, PhDMatthijs C. Brouwer, MD, PhDWolfgang Bruck, MDDavid B. Clifford, MD, FAANRussell C. Dale, MBChB, MRCP, PhDOscar H. Del Brutto, MD, FAANThomas G. Forsthuber, MDKazuo Fujihara, MDJeffrey M. Gelfand, MD, MAS, FAANJoseph El Khoury, MDMarkus P. Kummer, PhD

Jun Li, MD, PhDLin Mei, MD, PhDAndrew McKeon, MDChristopher Power, MDRichard M. Ransohoff, MDChristine Stadelmann, MDIsrael Steiner, MDLawrence Steinman, MDJesper Tegner, PhD, MScSilvia N. Tenembaum, MDMaarten J. Titulaer, MD, PhDAri Waisman, PhDHugh J. Willison, MBBS, PhDGregory F. Wu, MD, PhD

Neurology® Journals

Editor-in-ChiefRobert A. Gross, MD, PhD, FAAN

Deputy EditorBradford B. Worrall, MD, MSc, FAAN

Section Editors

BiostatisticsRichard J. Kryscio, PhDChristopher A. Beck, PhDSue Leurgans, PhD

Classification of Evidence EvaluationsGary S. Gronseth, MD, FAAN

PodcastsStacey L. Clardy, MD, PhDJeffrey B. Ratliff, MD, Deputy Podcast Editor

OmbudsmanDavid S. Knopman, MD, FAAN

Scientific Integrity AdvisorRobert B. Daroff, MD, FAAN

Classification of Evidence Review TeamMelissa J. Armstrong, MDRichard L. Barbano, MD, PhD, FAANRichard M. Dubinsky, MD, MPH, FAANJeffrey J. Fletcher, MD, MScGary M. Franklin, MD, MPH, FAANDavid S. Gloss II, MD, MPH&TMJohn J. Halperin, MD, FAANJason Lazarou, MSc, MDSteven R. Messe, MD, FAANPushpa Narayanaswami, MBBS, DM, FAANAlex Rae-Grant, MD

Vision Neurology® Neuroimmunology &Neuroinflammation will be the premier peer-reviewed journal for experts in the fields ofneuroimmunology and neuroinflammation.

Mission Neurology: Neuroimmunology &Neuroinflammation will provide neurologistswith peer-reviewed articles, editorials, andreviews to enhance patient care, education,and clinical & translational research.

EditorialInquiries

Tel: 612-928-6400Toll-free: 800-957-3182 (US)Fax: [email protected]

StayConnected

facebook.com/NeurologyNN

twitter.com/greenjournal

youtube.com/user/NeurologyJournal


TABLE OF CONTENTS Volume 5, Number 4, July 2018 Neurology.org/NN

Editor’s Corner

e472 Body surface area and B-cell repopulationJ. Dalmau

Open Access

Editorial

e471 Infection-triggered autoimmunity: The case of herpessimplex virus type 1 and anti-NMDAR antibodiesR.C. Dale and M. Nosadini

Open Access Companion article, p. e458

Articles

e458 Anti-NMDA receptor encephalitis andnonencephalitic HSV-1 infectionA. Salovin, J. Glanzman, K. Roslin, T. Armangue, D.R. Lynch, andJ.A. Panzer

Open Access Editorial, p. e471

e463 Monitoring B-cell repopulation after depletiontherapy in neurologic patientsE. Ellwardt, L. Ellwardt, S. Bittner, and F. Zipp

Open Access

e467 Neopterin is associated with hippocampal subfieldvolumes and cognition in HIVD.A. Fleischman, K. Arfanakis, S. Leurgans, S.M. Keating, M. Lamar,D.A. Bennett, O.M. Adeyemi, and L.L. Barnes

Open Access

e465 Expanded autologous regulatory T-lymphocyteinfusions in ALS: A phase I, first-in-humanstudyJ.R. Thonhoff, D.R. Beers, W. Zhao, M. Pleitez, E.P. Simpson,J.D. Berry, M.E. Cudkowicz, and S.H. Appel

Open Access Class of Evidence

e466 Molecular signature of Epstein-Barr virus infection inMS brain lesionsM.A. Moreno, N. Or-Geva, B.T. Aftab, R. Khanna, E. Croze,L. Steinman, and M.H. Han

Open Access

e470 Mononuclear cell transcriptome changes associatedwith dimethyl fumarate in MSA.R. Gafson, K. Kim, M.T. Cencioni, W. van Hecke, R. Nicholas,S.E. Baranzini, and P.M. Matthews

Open Access

Clinical/Scientific Notes

e459 Fecal microbiota transplantation associated with10 years of stability in a patient with SPMSS. Makkawi, C. Camara-Lemarroy, and L. Metz

Open Access

e460 Obinutuzumab, a potent anti–B-cell agent, forrituximab-unresponsive IgM anti-MAGneuropathyG. Rakocevic, U. Martinez-Outschoorn, and M.C. Dalakas

Open Access Class of Evidence

e464 Granulomatous myositis induced by anti–PD-1monoclonal antibodiesN. Uchio, K. Taira, C. Ikenaga, A. Unuma, M. Kadoya, A. Kubota,T. Toda, and J. Shimizu

Open Access

e468 Mortality in neuromyelitis optica is stronglyassociated with African ancestryM.A. Mealy, R.A. Kessler, Z. Rimler, A. Reid, L. Totonis, G. Cutter,I. Kister, and M. Levy

Open Access

http://neurology.org/nn

e469 Unusual neurologic presentation of aseptic abscessessyndromeP. Nicolas, O. Guerrier, A. Benoit, F. Durand-Dubief, G. Raverot,S. Debarbieux, C. Delteil, A. Vasiljevic, E. Jouanneau, F. Cotton,R. Marignier, and S. Vukusic

Open Access

Correction

e473 High-dose cyclophosphamidewithout stem cell rescuein immune-mediated necrotizing myopathies Cover image

Muscle biopsy showing PD-L1 overexpression on granuloma cells andnon-necrotic fibers surrounding granuloma. See “Granulomatousmyositis induced by anti–PD-1 monoclonal antibodies.”See page e464

Editor’s Corner Volume 5, Number 4, July 2018Neurology.org/NN

Josep Dalmau, MD, PhD, Editor, Neurology® Neuroimmunology & Neuroinflammation

Body surface area and B-cellrepopulationNeurol Neuroimmunol Neuroinflamm 2018;5:e472. doi:10.1212/NXI.0000000000000472

Many studies in neuroimmunology are focused on biomarkers of treatment response and diseaseoutcome. These studies are clinically important considering the increasing armamentarium ofdrugs and paraclinical tests available and represent the main theme that unifies several of thearticles included in the current issue of Neurology® Neuroimmunology & Neuroinflammation (N2).Dr. Ellwardt et al.1 investigated the factors that influence B-cell repopulation after B-cell depletiontherapy in patients with neuromyelitis optica spectrum disorders andMS. The cohort consisted of45 patients at their first treatment, 42 received rituximab and 3 ocrelizumab, and most weretreated with body surface area (BSA)-adapted dosage of medication. B-cell repopulation wasdefined as the first detection of CD19+ cells above 1% of total CD45+ lymphocytes aftertreatment. Multiple factors were investigated including BSA, age, sex, CSF parameters, pre-treatment therapy, and absolute lymphocyte and leukocyte counts during treatment. The studyshows that the only factor with significant influence on the CD19+ B-cell recovery was the BSA.Patients with a larger BSA had a higher probability to reach 1% CD19 cells sooner than patientswith a smaller BSA after B-cell depletion therapy. The authors suggest that this finding is due toa systematic underestimation of rituximab dosage when the Dubois equation is used for the BSAcalculation, particularly in patients with large height and weight. Calculating the BSA with theMosteller equation partially overcame this effect. In addition, the authors suggest that the use ofthe arbitrary dose of 375 mg/m2 is not sufficient and should be increased, especially for patientswith a high BSA. The authors acknowledge that some effects may have been overlooked becauseof the limited number of analyzed events, and that future investigations should consider includingnovel biomarkers (e.g., CD27+ B cells among others), and the presence of gene polymorphisms(e.g., FCGR3A) that have been suggested to predict efficacy of B-cell targeted therapies.

There is currently no reliable prediction marker for early treatment response to any diseasemodifying treatment for relapsing remitting MS (RRMS). Therefore, to determine individualpharmacodynamic responses that can distinguish patients who will respond to dimethyl fumarate(DMF), Dr. Gafson et al.2 examined the short-term changes in gene expression in peripheral bloodmononuclear cells (PBMCs) of patients with RRMS treated with DMF. Blood samples wereobtained from 24 patients at baseline, 6 weeks, and 15 months after initiating treatment. UsingRNAseq, the authors identified a robust short-term transcriptomic response to DMF in PBMCsthat was associated with activation of the Nrf2 and inhibition of the NFκB pathways in responders.In addition, these patients showed stabilization of gene expression between 6 weeks and 15months.By contrast, no early transcriptional changes were identified in nonresponders, who also showedgreater expression of proinflammatory pathway genes as compared with healthy controls. Findingsfrom this study confirm previously reported modulating effects of DMF on genes related toantioxidant,3 anti-inflammatory,4 and NFκB pathways.5 The authors acknowledge the small samplesize and the use of only 3 time points in sample testing, indicating the need of further confirmatorywork. Yet, these preliminary findings highlight the sensitivity of RNA-Seq whole transcriptomic

From the ICREA-IDIBAPS, Hospital Clınic, University of Barcelona, Spain; and the Department of Neurology, University of Pennsylvania, Philadelphia.

Funding information and disclosures are provided at the end of the article. Full disclosure form information provided by the authors is available with the full text of this article atNeurology.org/NN.

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloadingand sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

MORE ONLINE

Editor SummaryNPub.org/N2/edsum

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. e1

http://ascopubs.org/doi/full/10.1212/NXI.0000000000000472

http://nn.neurology.org/lookup/doi/10.1212/NXI.0000000000000472

http://creativecommons.org/licenses/by-nc-nd/4.0/

http://NPub.org/N2/edsum

assays of PBMC, which although currently are expensive, havethe potential to be streamlined andmore accessible in the future.

Several studies have shown that neopterin, a pteridine producedin macrophages and dendritic cells, is an early marker of HIV-associated cellular immune activation and provides a sensitivemeasure of treatment efficacy and HIV disease course,6,7 pre-dicting HIV-related mortality.8 The levels of neopterin are alsoassociated with compromise of the blood-brain barrier, general-ized brain atrophy, and HIV-associated cognitive impairment.9,10

Based on these findings, Fleischman et al.11 investigated in 66HIV-infected, virally suppressed patients, the effects of the im-mune response and inflammatory cascade on MRI-derived re-gional brain volumes and cognition. Cross-sectional associationswere examined between serum immune markers of activation(neopterin) and inflammation (interleukin [IL]-1β, IL-6, tumornecrosis factor alpha, andC-reactive protein) and several regionalbrain volumes and cognition. None of the inflammatory markerswas associated with any regional brain volume, but higher neo-pterin levels were associated with lower presubiculum, cornuammonis (CA)1, and CA4/dentate volumes, and deficits ofcognition and semantic and working memory. Despite the limi-tations of the study noted by the authors (small number ofpatients with limited age range, no measurement of CSF neo-pterin, lack of adjustment to premorbid intelligence, and cross-sectional study), the findings suggest that serum neopterin levelsand 3-TMRI neuroimaging (both clinically available and easy toobtain)may be useful in identifyingHIV-infected adults at risk ofdeveloping further neuronal injury and cognitive impairment.

In addition to these studies, the July issue of N2 containsother interesting articles that I hope will catch your attention.

Study fundingNo targeted funding reported.

DisclosureJ. Dalmau is the editor of Neurology: Neuroimmunology &Neuroinflammation; is on the editorial board of Neurology

UpToDate; holds patents for and receives royalties fromMa2 autoantibody test, NMDA receptor autoantibody test,GABA(B) receptor autoantibody test, GABA(A) receptorautoantibody test, DPPX autoantibody test, and IgLON5autoantibody test; and receives research support fromEuroimmun, NIH, Fundacio CELLEX, Instituto Carlos III(CIBERER, and Fondo de Investigaciones Sanitarias). Fulldisclosure form information provided by the authors isavailable with the full text of this article at Neuro-logy.org/NN.

Received May 8, 2018. Accepted in final form May 8, 2018.

References1. Ellwardt E, Ellwardt L, Bittner S, Zipp F. Monitoring B-cell repopulation after de-

pletion therapy in neurological patients. Neurol Neuroimmunol Neuroinflamm 2018;5:e463. doi: 10.1212/NXI.0000000000000463.

2. Gafson AR, Kim K, Cencioni MT, et al. Mononuclear cell transcriptomechanges associated with dimethyl fumarate in multiple sclerosis. NeurolNeuroimmunol Neuroinflamm 2018;5:e470. doi: 10.1212/NXI.0000000000000470.

3. Zhao G, Liu Y, Fang J, Chen Y, Li H, Gao K. Dimethyl fumarate inhibits the ex-pression and function of hypoxia-inducible factor-1alpha (HIF-1alpha). BiochemBiophys Res Commun 2014;448:303–307.

4. Tahvili S, Zandieh B, Amirghofran Z. The effect of dimethyl fumarate on geneexpression and the level of cytokines related to different T helper cell subsets inperipheral blood mononuclear cells of patients with psoriasis. Int J Dermatol 2015;54:e254–e260.

5. Hund AC, Lockmann A, Schon MP. Mutually enhancing anti-inflammatory activitiesof dimethyl fumarate and NF-kappaB inhibitors–implications for dose-sparing com-bination therapies. Exp Dermatol 2016;25:124–130.

6. Mildvan D, Spritzler J, Grossberg SE, et al. Serum neopterin, an immune activationmarker, independently predicts disease progression in advanced HIV-1 infection. ClinInfect Dis 2005;40:853–858.

7. Wirleitner B, Schroecksnadel K, Winkler C, Fuchs D. Neopterin in HIV-1 infection.Mol Immunol 2005;42:183–194.

8. Stein DS, Lyles RH, Graham NM, et al. Predicting clinical progression or death insubjects with early-stage human immunodeficiency virus (HIV) infection: a compar-ative analysis of quantification of HIV RNA, soluble tumor necrosis factor type IIreceptors, neopterin, and beta2-microglobulin. Multicenter AIDS Cohort Study.J Infect Dis 1997;176:1161–1167.

9. Anesten B, Yilmaz A, Hagberg L, et al. Blood-brain barrier integrity, intrathecalimmunoactivation, and neuronal injury in HIV. Neurol Neuroimmunol Neuro-inflamm 2016;3:e300. doi: 10.1212/NXI.0000000000000300.

10. Fuchs D, Moller AA, Reibnegger G, Stockle E, Werner ER, Wachter H. Decreasedserum tryptophan in patients with HIV-1 infection correlates with increased serumneopterin and with neurologic/psychiatric symptoms. J Acquir Immune Defic Syndr1990;3:873–876.

11. Fleischman DA, Arfanakis K, Leurgans SE, et al. Neopterin is associated with hip-pocampal subfield volumes and cognition in HIV. Neurol Neuroimmunol Neuro-inflamm 2018;5:e467. doi: 10.1212/NXI.0000000000000467.

e2 Neurology: Neuroimmunology & Neuroinflammation | Volume 5, Number 4 | July 2018 Neurology.org/NN









EDITORIAL OPEN ACCESS

Infection-triggered autoimmunityThe case of herpes simplex virus type 1 and anti-NMDAR antibodies

Russell C. Dale, MRCP, and Margherita Nosadini, MD

Neurol Neuroimmunol Neuroinflamm 2018;5:e471. doi:10.1212/NXI.0000000000000471

Correspondence

Prof. Dale

[email protected]

The recent paradigm that autoimmune encephalitis is associated with autoantibodies thatbind to conformational epitopes of cell surface antigens involved in neurotransmission hastransformed the field of autoimmune neurology and created new biomarkers for the cli-nician to define treatable acquired neurologic conditions. In parallel, a growing arma-mentarium of immune therapeutic agents is becoming available. But fundamental questionsremain, particularly “why does this happen?” There is cohort-derived evidence suggestingage and ethnic vulnerability and emerging evidence of HLA associations.1,2 Anti–NMDAreceptor (NMDAR) encephalitis, which is defined by a characteristic clinical syndrome andthe presence of CSF autoantibodies to the NR1 subunit of the NMDAR, is the mostcommon autoimmune encephalitis. In anti-NMDAR encephalitis, many female adults havean ovarian teratoma that contains NMDAR-expressing neural tissue, providing a hypo-thetical example of “cross-reactive” autoimmunity between the NMDARs in the tumor andin the brain. However, large cohorts have shown that ovarian teratomas are very uncommonin pediatric patients,3 so what is the “trigger” of the loss of immune tolerance in youngchildren? Some of these children have preceding herpes simplex type 1 (HSV1) enceph-alitis, a destructive highly inflammatory encephalitis that predominantly affects temporallobes and surrounding structures. After recovery, approximately 20% of children withHSV1 encephalitis deteriorate with a “second phase” of encephalitis, this time with clinicalfeatures typical of anti-NMDAR encephalitis, including florid movement disorders andencephalopathy, associated with the emergence of anti-NMDAR antibodies.4–6 This clearexample of HSV1 encephalitis inducing secondary anti-NMDAR encephalitis yielded thehypothesis that the inflammatory destruction of neural tissue with release of neural antigensinto the surrounding brain, circulation and then lymphatic system, resulted in a reactivationof autoreactive lymphocytes against NMDAR antigens and production of pathogenic anti-NMDAR antibodies.

However a new article, in this month’s Neurology® Neuroimmunology and Neuroinflammation,strengthens the evidence for the alternative pathophysiologic mechanism between HSV1 andanti-NMDAR autoimmunity: namely that HSV1 infection increases the likelihood of de-veloping anti-NMDAR encephalitis, possibly through “molecular mimicry,” without the needfor brain tissue destruction (i.e., without HSV1 encephalitis).7 Salovin et al. performed testingfor previous HSV1 infection (immunoglobulin G [IgG] serology) in stored serum samplesfrom children with anti-NMDAR encephalitis (with typical clinical features and positive serumanti-NMDAR antibodies) with no clinical or radiologic evidence of HSV1 encephalitis. Therewere 2 cohorts, the first from Philadelphia and the second from Barcelona (combined mean age9 years, range 1–17). They found statistically significant elevation of HSV1 IgG seropositivity inpatients with anti-NMDAR encephalitis compared with controls in the Philadelphia cohort andthe combined cohorts (49% vs 21% in the combined cohorts, p = 0.007). It is important thatthey used hospital controls who were children with other neuroinflammatory disorders, whichreduced the likelihood that the findings were due to nonspecific inflammatory associations andmakes it more likely to be specific to anti-NMDAR encephalitis. Second, they ensured that

From the Neuroimmunology Group (R.C.D.), Kids Neuroscience Centre, Kids Research at the Children’s Hospital at Westmead, University of Sydney, Australia; and the PaediatricNeurology and Neurophysiology Unit (M.N.), Department of Women’s and Children’s Health, University Hospital of Padua, Italy.



RELATED ARTICLE

Anti-NMDA receptorencephalitis andnonencephalitic HSV-1infection

Page e458

Copyright © 2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology. 1

http://dx.doi.org/10.1212/NXI.0000000000000471




patients did not have IV immunoglobulin (IVIG) before se-rologic sampling, as IVIG may induce false-positive infectiousserologic findings, given the polyclonal multidonor origins.Third, the patients were age matched with controls—this isvery important, particularly with an extremely common in-fection such as HSV1, which has increasing seropositivity withincreasing childhood age. As expected, the Barcelona controlgroup, who were older, had higher HSV1 seropositivity.Fourth, the authors also measured IgG against 2 other verycommon viral infections in childhood, namely cytomegalo-virus and Epstein-Barr virus, and found no difference betweenanti-NMDAR encephalitis patients and controls.

There were some weaknesses in this approach, partly relatedto the retrospective nature of the study. First, it would havebeen of interest to better understand the relationship of thetiming of infection and production of anti-NMDAR anti-bodies, which could have been performed by measuringHSV1 IgM antibodies, and convalescent HSV1 IgG sam-pling. Similarly, there is no reporting of the time betweenonset of anti-NMDAR encephalitis and serum sampling or ofthe type and timing of preceding clinical HSV1 infection.Second, and acknowledged by the authors, there was nocontrolling for race or socioeconomic factors, which mayinfluence HSV1 infection, as the presence of other childrenin the family, and housing overcrowding can influence theprevalence of common childhood infectious disease. Third,it is still plausible that HSV1 could induce a subclinical HSV1encephalitis in some patients, which could be missed clini-cally or radiologically—proving HSV1 PCR negativity inCSF would have added to this study and strengthened thefindings further. Indeed, the detection of “asymptomatic”positive CSF PCR for HSV1 in anti-NMDAR encephalitispatients has been reported and discussed in a recent review.6

Moreover, the number of patients in the Salovin study whofulfilled Graus et al8 clinical criteria for anti-NMDARencephalitis is unclear. It is also unclear whether NMDARantibodies were present in CSF (or only serum). Finally,although cytomegalovirus and Epstein-Barr virus testinghelped understand the specificity of the HSV1 findings,testing for other infections, particularly Mycoplasma pneu-moniae would have added to the study, as Mycoplasma hasbeen proposed to trigger anti-NMDAR encephalitis,9 andother types of herpes viruses beside HSV1 have beenreported in anti-NMDAR encephalitis.6

Regardless of these limitations, the article by Salovin et al. isimportant and redresses the balance in the debate about in-duction of anti-NMDAR autoimmunity and supports thehypothesis that preceding viral infection can result in loss ofimmune tolerance in predisposed individuals.

Author contributionsR.C. Dale: no duplicate publication; drafting/revising themanuscript; accepts responsibility for conduct of researchand will give final approval; acquisition of data; and principalinvestigator/guarantor. M. Nosadini: no duplicate publication;drafting/revising the manuscript; analysis or interpretation ofdata; and accepts responsibility for conduct of research and willgive final approval.


DisclosureR.C. Dale served on the scientific advisory board of NationalBlood Authority; received speaker honoraria from Biogen andBristol-Myers Squibb; served on the editorial board ofMSARD,Neurology: Neuroimmunology & Neuroinflammation, and Euro-pean Journal of Paediatric Neurology; and received researchsupport from the NHMRC and Multiple Sclerosis ResearchAustralia. M. Nosadini reports no disclosures. Full disclosureform information provided by the authors is available with thefull text of this article at Neurology.org/NN.

References1. JonesHF,MohammadSS,ReedPW, et al. Anti-N-methyl-d-aspartate receptor encephalitis in

M�aori and Pacific Island children inNewZealand.DevMedChildNeurol 2017;59:719–724.2. Mueller SH, Farber A, Pruss H, et al. Genetic predisposition in anti-LGI1 and anti-

NMDA receptor encephalitis. Ann Neurol 2018;83:863–869.3. Titulaer MJ, McCracken L, Gabilondo I, et al. Treatment and prognostic factors for

long-term outcome in patients with anti-NMDA receptor encephalitis: an observa-tional cohort study. Lancet Neurol 2013;12:157–165.

4. Armangue T, Leypoldt F, Malaga I, et al. Herpes simplex virus encephalitis is a triggerof brain autoimmunity. Ann Neurol 2014;75:317–323.

5. Mohammad SS, Sinclair K, Pillai S, et al. Herpes simplex encephalitis relapse with choreais associated with autoantibodies to N-methyl-D-aspartate receptor or dopamine-2 re-ceptor. Mov Disord 2014;29:117–122.

6. Nosadini M, Mohammad SS, Corazza F, et al. Herpes simplex virus-induced anti-N-methyl-D-aspartate receptor encephalitis: a systematic literature review with analysisof 43 cases. Dev Med Child Neurol 2017;59:796–805.

7. Salovin A, Glanzman J, Roslin K, Armangue T, Lynch DR, Panzer JA. Anti-NMDAreceptor encephalitis and nonencephalitic HSV-1 infection. Neurol NeuroimmunolNeuroinflamm 2018;90:e458. doi: 10.1212/NXI.0000000000000458.

8. Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmuneencephalitis. Lancet Neurol 2016;15:391–404.

9. Florance NR, Davis RL, Lam C, et al. Anti-N-methyl-D-aspartate receptor (NMDAR)encephalitis in children and adolescents. Ann Neurol 2009;66:11–18.

2 Neurology: Neuroimmunology & Neuroinflammation | Volume 5, Number 4 | July 2018 Neurology.org/NN




ARTICLE OPEN ACCESS

Anti-NMDA receptor encephalitis andnonencephalitic HSV-1 infectionAmy Salovin, MS, Jason Glanzman, Kylie Roslin, Thais Armangue, MD, PhD, David R. Lynch, MD, PhD,

and Jessica A. Panzer, MD, PhD


Correspondence

Dr. Lynch

[email protected]

AbstractObjectiveTo determine whether there is an association between nonencephalitic herpes simplex virus 1(HSV-1) infection and anti-NMDA receptor encephalitis (anti-NMDARE).

MethodsAntibody testing was performed using samples from 2 cohorts in a case-control observationalstudy. The cohort “Philadelphia” included 16 serum samples of pediatric anti-NMDARE casesand 42 age-matched controls with other neuroinflammatory disorders studied at the Children’sHospital of Philadelphia and University of Pennsylvania. The cohort “Barcelona” contained 23anti-NMDARE patient samples and 26 age-matched participants with other neuro-inflammatory disorders studied at IDIBAPS-Hospital Clinic, University of Barcelona. Thepresence of HSV-1 IgG antibodies was examined by ELISA. As an additional control, IgGantibodies to cytomegalovirus (CMV) and Epstein-Barr virus viral capsid antigen (EBV-VCA)were determined.

ResultsIn each cohort, more participants with anti-NMDARE than controls had anti-HSV-1 IgGantibodies. In the Philadelphia cohort (58 participants), 44% of anti-NMDARE cases hadantibodies to HSV-1 compared with 14% controls (OR 4.67, 95% CI 1.3–17.3, p = 0.031). Inthe Barcelona cohort (49 participants), 52% of participants with anti-NMDARE had antibodiesto HSV-1 compared with 31% of controls (OR 2.45, 95% CI 0.7–7.9, p = 0.155). Overall, 49%of anti-NMDARE cases have antibodies to HSV-1 in these 2 combined cohorts compared with21% of controls (Mantel-Haenszel OR 3.21, 95% CI 1.3–7.7, p = 0.007).

ConclusionPast HSV-1 infection was found in significantly more anti-NMDARE cases than controls. Thissuggests a meaningful association between nonencephalitic HSV-1 infection and developmentof anti-NMDARE.

From the Division of Neurology (A.S., J.G., K.R., D.R.L., J.A.P.), Department of Pediatrics, Children’s Hospital of Philadelphia, PA; Clinical and Experimental Neuroimmunology Program(T.A.), August Pi i Sunyer Biomedical Research Institute (IDIBAPS), Hospital Clınic, University of Barcelona; Pediatric Neuroimmunology Unit (T.A.), Department of Neurology, Sant Joande Deu Childrens Hospital, University of Barcelona, Spain; and Department of Neurology (D.R.L., J.A.P.), Perelman School of Medicine, University of Pennsylvania, PA.


The Article Processing Charge was funded by the authors.







Anti-NMDA receptor encephalitis (anti-NMDARE) is a re-cently described autoimmune encephalitis associated withpsychosis, seizures, dyskinesias, hypoventilation, and auto-nomic instability, attributed to antibodies against the GluN1subunit of the NMDA receptor (NMDAR).1 This disorder isnow the most commonly identified nonviral cause of en-cephalitis.2 In approximately 50% of teenage girls and adultwomen with the disorder, it is triggered by an NMDAR-expressing ovarian teratoma.3 By contrast, greater than 90% ofyoung girls or males with the disorder have no identifiedtrigger, although it may be associated with viral infection.4

Approximately 10%–25% of patients with herpes simplexvirus 1 (HSV-1) encephalitis have an immune-mediatedrelapse associated with GluN1-specific antibodies and anti-NMDARE symptoms.4,5 The association of nonencephaliticHSV-1 infection with anti-NMDARE has not been examined.To investigate the possibility that nonencephalitic HSV-1infection could trigger the production of anti-NMDAR anti-bodies, this study examined HSV-1 seropositivity in 2 par-ticipant cohorts, with and without anti-NMDAR encephalitis,to ascertain if an association exists between nonencephaliticHSV-1 infection and anti-NMDARE.

MethodsParticipant materialCases were identified by neurologist diagnosis followingreview of clinical information (table) combined with results ofanti-NMDAR antibody testing from serum sent to neuro-immune repositories at the Hospital of the University ofPennsylvania, the Children’s Hospital of Philadelphia, orUniversity of Barcelona. Sera from the Children’s Hospital ofPhiladelphia (CHOP) and University of Pennsylvania(UPenn) repositories were combined into 1 cohort (Phila-delphia). Sera from the University of Barcelona were analyzedas a separate cohort (Barcelona). We identified case sera fromchildren aged 0–11 years deposited in the Philadelphiarepositories between 2008 and 2011 and from childrenaged 0–17 years submitted between 2008 and 2017 in theUniversity of Barcelona repository. Children with a knowntumor or with IV immunoglobulin treatment before samplecollection were excluded. We identified 39 combined casesfulfilling these criteria, 16 from the Philadelphia cohort and 23from the University of Barcelona. The mean age of thecombined cohorts was 9.0 ± 4.7 years (range 1–17 years). ThePhiladelphia cohort contained children with a mean age of6.8 ± 3.1 years, and the Barcelona cohort contained childrenwith a mean age of 10.6 ± 5.0. To limit sampling bias, 68age-matched controls with other potential neuroimmunedisease were identified from the same repositories. From

Philadelphia, 43 controls were identified, and 1 was sero-positive for HSV2 IgG and was excluded. Included controlshad a mean age of 7.9 ± 2.4 years. From the University ofBarcelona, 26 controls were identified with a mean age of10.5 ± 5.2 years. No participants had knownHSV encephalitis.

Standard protocol approvals, registrations,and patient consentsSamples were collected in accordance with the guidelines ofInstitutional Review Boards of the CHOP, UPenn, andUniversity of Barcelona. Informed consent was obtained fromeach participant.

ELISA analysisThe following commercially available solid-phase ELISA kitswere used according to the manufacturer’s instructions:HSV-1 IgG (20-HSGHU-E01, ALPCO and ab108737,Abcam), HSV2 IgG (20-H2GHU-E01, ALPCO andab108739, Abcam), CMV IgG (GD84, Genesis Diagnostics/ALPCO and ab108724, Abcam), and EBV-VCA (ab108730,Abcam). All HSV-1 IgG-positive samples were screened forHSV2 IgG. Results were determined by comparison with pos-itive, negative, and cutoff controls. Sera with values greater than10% of the cutoff control were considered positive. Sera wasassayed and analyzed with masked case/control designation.

Statistical analysisStatistical analyses were performed using STATA/GraphPad.Missing variables were excluded from analysis. The Fisherexact test was used to compare categorical variables, witha significance level of 0.05. Continuous variables were analyzedusing the Wilcoxon rank-sum test. Data from cohort combi-nation are presented as a Mantel-Haenszel–adjusted OR.

ResultsHigher prevalence of HSV-1 seropositivity inanti-NMDARE cases

Philadelphia cohortA significantly higher number of anti-NMDARE cases hadIgG antibodies against HSV-1 (44%; 7/16) compared withage-matched controls (14%; 6/42), leading to an OR of 4.66(95% CI 1.3–17.3, p = 0.031) (figure). The mean age of anti-HSV-1 IgG-positive and -negative participants was not sig-nificantly different.

Barcelona cohortCompared with age-matched controls, a higher number ofanti-NMDARE cases had IgG antibodies against HSV-1.Fifty-two percent (12/23) of cases hadHSV-1 IgG antibodies,

GlossaryCHOP = Children’s Hospital of Philadelphia; CMV = cytomegalovirus; EBV = Epstein-Barr virus;HSV = herpes simplex virus1; NMDAR = NMDA receptor; NMDARE = NMDA receptor encephalitis; UPenn = University of Pennsylvania.



whereas 31% (8/26) of controls were HSV-1 IgG positive(OR 2.46, 95%CI 0.8–7.9, p = 0.155) (figure). Although therewas no statistical difference, the trend in this cohort matchesthat of the Philadelphia cohort.

OverallCombination of the 2 cohorts reveals that an overall 19 (49%)of 39 anti-NMDARE cases were seropositive for HSV-1 IgG(figure), significantly higher than age-matched controls(14/68 [21%], Mantel-Haenszel–adjusted OR 3.21, 95% CI1.3–7.7, p = 0.007). The mean age of HSV-1 IgG-positive and-negative participants was not significantly different.

To control for the possibility that nonspecific antibody ele-vation in case sera could yield false positives on ELISA, wealso tested for immunoreactivity to cytomegalovirus (CMV)and Epstein-Barr virus (EBV), infections without knownanti-NMDARE association. CMV and EBV-VCA IgG sero-positivity did not differ between anti-NMDARE cases andcontrols in either cohort, although the number of participantspositive for CMV and EBV-VCA IgG antibodies was higher inthe Barcelona cohort than that in the Philadelphia cohort(Philadelphia—CMV: 5/15 [33%], EBV: 9/16 [30%];Barcelona—CMV: 14/18 [78%], EBV: 18/22 [82%])(figure; Philadelphia—CMV: OR 1.2, 95% CI 0.32–4.05,p = 0.823, EBV: OR 1.0, 95% CI 0.3–3.2, p = 0.999;Barcelona—CMV: OR 1.6, 95% CI 0.39–6.24, p = 0.532,EBV: OR 1.8, 95% CI 0.4–7.0, p = 0.505).

DiscussionIn this study, anti-NMDARE cases had a higher prevalence ofHSV-1 IgG antibodies than age-matched controls in twoseparate cohorts. Antibodies to unrelated viruses (CMV andEBV-VCA) were not elevated in anti-NMDARE participants,suggesting that this finding is selective for HSV-1. HSV-1 IgGseroprevalence is greater than 60% by mid-adulthood.6

Therefore, we evaluated the association between HSV-1 in-fection and anti-NMDARE in children, in which reportedHSV-1 IgG seroprevalence is only 26% at age 7 years.7 Therate of positive HSV-1 IgG antibodies detected in ourparticipants with anti-NMDARE is elevated compared withthis population statistic and the age-matched control groupdescribed here. In this retrospective study, with limiteddemographic information available, we were unable to controlfor race or socioeconomic status. However, HSV-1 seropos-itivity in anti-NMDARE participants was greater than thatreported for any demographic group in this age range.7 Pro-spective studies may better control for demographic factors toconfirm this finding.

The development of anti-NMDAR immunoreactivity fol-lowingHSV encephalitis has been hypothesized to result fromHSV-mediated brain injury that exposes NMDARs to theimmune system, triggering an immune response.4 However,none of our anti-NMDARE participants with HSV antibodieshad clinically evident HSV encephalitis. The present data

Table Descriptive statistics for Philadelphia and Barcelona cohorts

Philadelphia Barcelona

Control Index Total p Control Index Total p

Sex

Female (N) 17 8 25 17 13 30

Female (%) 43.6 50.0 45.5 0.772 65.4 56.5 61.2 0.407

Unknown (N) 1 0 1 4 6 10

Age

Mean ± SD, y 7.9 ± 2.5 6.8 ± 3.1 7.6 ± 2.7 0.242 10.4 ± 5.1 10.6 ± 5.0 10.5 ± 5.0 0.888

Median, y (Q1, Q3) 8 (7, 10) 7.5 (3.8, 9) 8 (6, 10) 12.5 (5.5, 15) 12 (6, 15) 12 (5, 15)

Clinical measures; patients presenting with N (%)

Pleocytosis 8 (14.5) 7 (43.8) 15 (27.3) 0.163 5 (19.2) 9 (39.1) 14 (28.6) 0.080

Unknown/unavailable cell count 17 (30.9) 5 (31.3) 22 (40.0) 7 (26.9) 8 (34.8) 15 (30.6)

Abnormal MRI 8 (14.5) 4 (25.0) 12 (21.8) 0.441 3 (11.5) 2 (8.7) 5 (10.2) 0.999

Unknown/unavailable MRI 24 (61.5) 4 (25.0) 28 (50.9) 15 (30.8) 13 (34.8) 28 (32.7)

Seizures 16 (41.0) 12 (75.0) 28 (50.9) 0.037 4 (15.4) 12 (52.2) 16 (32.7) 0.013

Encephalopathy 25 (64.1) 15 (93.8) 40 (72.7) 0.043 20 (76.9) 17 (73.9) 37 (75.5) 0.999

Movement disorder 21 (53.8) 14 (87.5) 35 (63.6) 0.029 16 (61.5) 13 (56.5) 29 (59.2) 0.777

Data are presented as number of participants and frequency (%). p values are reported from the Fisher exact test, except age, which is reported from theWilcoxon rank-sum test.

Neurology.org/NN Neurology: Neuroimmunology & Neuroinflammation | Volume 5, Number 4 | July 2018 3


suggest an alternate disease mechanism, such as molecularmimicry, in which an epitope expressed by HSV is structurallysimilar to a portion of the NMDAR. Shared epitopes between

HSV-1 and disease specific autoantigens have been implicatedin neuroimmune conditions including myasthenia gravis.8

Alternatively, HSV infection may alter NMDAR expression,even outside the CNS, in a way that makes the receptor moreimmunogenic. For example, HSV-1 has been reported to“disarm” the unfolded protein response.9 Increased expres-sion of misfolded proteins may result in recognition byautoreactive immune cells.10 Finally, HSV-1 could modulatethe immune system in a manner that allows inappropriaterecognition of native NMDARs.

The present data suggest that nonencephalitic HSV-1 infectionmay trigger the subsequent development of anti-NMDARE.Prospective studies are needed to determine the timing of anti-NMDARE onset after nonencephalitic HSV-1 infection, alongwith mechanistic studies evaluating the mechanism by whichsuch events are involved in disease pathogenesis.

Author contributionsAmy Salovin: study concept and design, acquisition of data,analysis and interpretation of data, and manuscript prepara-tion. Jason Glanzman, Kylie Roslin, and Thais Armangue:acquisition of data and analysis and interpretation of data.David R. Lynch and Jessica A. Panzer: study concept anddesign, analysis and interpretation of data, study supervision,and critical revision of the manuscript for intellectual content.

AcknowledgmentThe authors thank Dr. Josep Dalmau (University Pennsylva-nia and University Barcelona) for his critical review of themanuscript and the Children’s Hospital of PhiladelphiaBiostatistics and Data Management Core for statisticalconsultation.

Study fundingStudy funded by the NIH (R21NS088148-01-02).

DisclosureA. Salovin, J. Glanzman, and K. Roslin report no disclosures.T. Armangue received research support from the MutuaMadrileña Foundation. D.R. Lynch is an associate editor ofthe Journal of Neurogenetics; holds a patent for creation of testfor anti-NMDA receptor encephalitis; consulted for BracketIncorporated; and received research support from EdisonPharmaceutical, Shire Pharmaceutical, Reata Pharmaceutical,Horizon Pharma, and the NIH. J.A. Panzer is deceased (nodisclosures are included for this author). Full disclosure forminformation provided by the authors is available with the fulltext of this article at Neurology.org/NN.

Received December 27, 2017. Accepted in final form February 26, 2018.

References1. Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor encephalitis: case

series and anlysis of the effects of antibodies. Lancet Neurol 2008;7:1091–1098.2. Gable MS, SheriffH, Dalmau J, Tilley DH, Glaser CA. The frequency of autoimmune

N-methyl-D-aspartate receptor encephalitis surpasses that of individual viral etiologiesin young individuals enrolled in the California encephalitis project. Clin Infect Dis2012;54:899–904.

Figure Seropositivity for IgG against HSV-1, CMV, and EBV

Samples from anti-NMDARE cases or controls from 2 cohorts were tested viaELISA for immunoreactivity to viral antigens. Samples were tested for IgGantibodies to HSV-1 and, as a control, IgG antibodies to CMV and EBV-VCA.(A) In 1 cohort from the neuroimmune repositories at the Children’s Hospitalof Philadelphia and theUniversity of Pennsylvania, 44%of anti-NMDARE caseswere seropositive for HSV-1 IgG compared with 14% of age-matched controls(p = 0.031, Fisher exact test). There was no difference in IgG positivity betweencases and controls for either CMV (33% cases and 30% controls) or EBV-VCA(29%cases and 29% controls). Data are presented aspercent positive sampleswith error bars indicating 95% CIs. (B) In a second cohort from the Universityof Barcelona, anti-NMDARE cases were more likely to be HSV-1 IgG positive(52% of anti-NMDARE cases compared with 31% of age-matched controls,p = 0.155, Fisher exact test). There was no difference in IgG positivity betweencases and controls for either CMV (78% cases and 69% controls) or EBV-VCA(82% cases and 72% controls). (C) Overall, when the cohorts are combinedtogether, anti-NMDARE cases were significantly more likely to be HSV-1 IgGseropositive (49% of anti-NMDARE cases compared with 21% of age-matchedcontrols, p = 0.007). There was no difference in IgG positivity between casesand controls for either CMV (58% cases and 45% controls) or EBV-VCA (71%cases and 63% controls). *p < 0.05; **p < 0.01. CMV = cytomegalovirus; EBV =Epstein-Barr virus; HSV = herpes simplex virus; NMDARE = NMDA receptorencephalitis; VCA = viral capsid antigen.




3. Titulaer MJ, McCracken L, Gabilondo I, et al. Treatment and prognostic factors forlong-term outcome in patients with anti-NMDA receptor encephalitis: an observa-tional cohort study. Lancet Neurol 2013;12:157–165.

4. Armangue T, Leypoldt F, Malaga I, et al. Herpes simplex virus encephalitis is a triggerof brain autoimmunity. Ann Neurol 2014;75:317–323.

5. Mohammad SS, Sinclair K, Pillai S, et al. Herpes simplex encephalitis relapse withchorea is associated with autoantibodies to N-Methyl-D-aspartate receptor ordopamine-2 receptor. Mov Disord 2014;29:117–122.

6. Xu F, Sternberg MR, Kottiri BJ, et al. Trends in herpes simplex virus type 1 and type 2seroprevalence in the United States. JAMA 2006;296:964–973.

7. Xu F, Lee F, Morrow R. Seroprevalence of herpes simplex virus type 1 in children inthe United States. J Pediatr 2007;151:374–377.

8. Schwimmbeck PL, Dyrberg T, Drachman DB, Oldstone MBA. Molecular mimicryand myasthenia gravis. An autoantigenic site of the acetylcholine receptor alpha-subunit that has biologic activity and reacts immunochemically with herpes simplexvirus. J Clin Invest 1989;84:1174–1180.

9. Burnett HF, Audas TE, Liang G, Lu RR. Herpes simplex virus-1 disarms the unfoldedprotein response in the early stages of infection. Cell Stress Chaperones 2012;17:473–483.

10. Todd DJ, Lee AH, Glimcher LH. The endoplasmic reticulum stress response inimmunity and autoimmunity. Nat Rev Immunol 2008;8:663–674.



ARTICLE OPEN ACCESS

Monitoring B-cell repopulation after depletiontherapy in neurologic patientsErik Ellwardt, MD, Lea Ellwardt, PhD, Stefan Bittner, MD, and Frauke Zipp, MD


Correspondence

Dr. E. Ellwardt

[email protected]

AbstractObjectiveTo determine the factors that influence B-cell repopulation after B-cell depletion therapy inneurologic patients and derive recommendations for monitoring and dosing of patients.

MethodsIn this study, we determined the association of body surface area (BSA; calculated by bodyweight and height with the Dubois formula), sex, pretreatment therapy, age, CSF data, andwhite blood cell counts with the risk and timing of B-cell repopulation, defined as 1% CD19+

cells (of total lymphocytes), following 87 B cell–depleting anti-CD20 treatment cycles of 45neurologic patients (28 women; mean age ± SD, 44.5 ± 15.0 years).

ResultsPatients with a larger BSA had a higher probability to reach 1% CD19+ cells than those witha smaller BSA (p < 0.05) following B-cell depletion therapy, although those patients hadreceived BSA-adapted doses of rituximab (375 mg/m2). Sex, pretreatment, age, CSF data, orabsolute lymphocyte and leukocyte counts during treatment did not significantly influenceCD19+ B-cell recovery in the fully adjusted models. Intraindividual B-cell recovery in patientswith several treatment cycles did not consistently change over time.

ConclusionsB-cell repopulation after depletion therapy displays both high inter- and intra-individual vari-ance. Our data indicate that a larger BSA is associated with faster repopulation of B cells, evenwhen treatment is adapted to the BSA. A reason is the routinely used Dubois formula,underestimating a large BSA. In these patients, there is a need for a higher therapy dose. BecauseB-cell count–dependent therapy regimes are considered to reduce adverse events, B-cellmonitoring will stay highly relevant. Patients’ BSA should thus be determined using theMosteller formula, and close monitoring should be done to avoid resurgent B cells and diseaseactivity.

From the Focus Program Translational Neurosciences (FTN) and Immunology (FZI) (E.E., S.B., F.Z.), Rhine Main Neuroscience Network (rmn2), Department of Neurology, UniversityMedical Center of the Johannes Gutenberg University Mainz; and Institute of Sociology and Social Psychology (L.E.), University of Cologne, Germany.









Studies that use depleting antibodies such as ocrelizumab andrituximab have proven the clinical efficacy of targeting B cellsin inflammation such as MS1,2 or neuromyelitis optica spec-trum disorder (NMOSD).3 The reappearance of B cells in theperipheral blood correlates with increased disease activity4

and is monitored by fluorescence-activated cell sorting(FACS) analysis. Although the efficacy of B-cell targetingtherapies is closely linked to sufficient depletion, a consider-able heterogeneity concerning monitoring and retreatmentprotocols complicates their use in clinical practice. Thereappearance of B cells can be defined when CD19+ cellsreach 1% of lymphocyte counts,5 which then leads to the nexttreatment cycle. However, some protocols use fixed timeintervals (e.g., every 6 months in clinical trials for ocrelizumab)but also 2% or absolute lymphocyte counts such as 10 per μLfor CD19+ cell monitoring.4,6 Our clinical experience is thatonce patients reach 1%, they rapidly surpass 2% within days. Asan alternative approach, monitoring of CD27+ memory B cellshas been suggested as a sensitive marker after rituximabtreatment.7,8 Still, there is a high interindividual variance be-tween patients who require personalized monitoring foreach patient. Inconsistency exists not only concerning B-cellmonitoring parameters but also dose regimens. Whereassome patients receive body surface area (BSA)-adapted doses(usually 375 mg/m2), some receive fixed doses.9

To identify factors influencing repopulation of B cells, weanalyzed CD19+ cell counts of neurologic patients (aged17–76 years) treated at least once with rituximab or ocreli-zumab and examined age, sex, medical pretreatment, oligo-clonal bands and cell count in the CSF before the firsttreatment, as well as concomitant leukocyte and lymphocytecounts and the BSA.

MethodsStandard protocol approvals, registrations,and patient consentsWe identified patients in the Neurology Department at theUniversity Medical Center Mainz, Germany, from 2007 untilMarch 2017 who had received at least 1 infusion withrituximab or ocrelizumab. Patient data were acquired retro-spectively and anonymized according to §36/37 of theRhineland-Palatinate state hospital law. Patients of the Uni-versity Medical Center Mainz provided their written consentthat data and remaining material can be used for research.

Data acquisitionFor those patients, we collected all available CD19+ cellcounts obtained by flow cytometric (FACS) analysis during

treatment in our clinic. FACS data were usually acquired 3 and6 months after the last treatment and monthly thereafter. Inindividual cases, close monitoring was done (e.g., if CD19+

cells were rising already after 3 months). In this analysis, wealso collected personal characteristics (e.g., age, sex, and BSA)and received information on medical pretreatments and di-agnosis from medical letters and reports. The BSA had beencalculated for all patients using the Dubois formula�0:007184 × height½cm�0:725 × body weight½kg�0:425� at eachtreatment. To compare the accuracy of the BSA calculation,

we also used the Mosteller formulaffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiheight½cm� × body weight½kg�

3;600

q.

The data included a record of the dates for all measurementsand treatments.

Definition of B-cell repopulationB-cell repopulation was defined as the first detection ofCD19+ cells above 1% of total CD45+ lymphocytes afterCD19+ cell depletion as measured by FACS. The number ofdays between the last rituximab infusion and the detection ofrepopulation was calculated. This repopulation of B cells isclassified as “event” throughout the article.

Statistical methodsThe data were characterized by a panel structure with re-peated observations per patient. Specifically, there were 45patients, resulting in a total of 373 patient observations duringthe observation period. Within the entire observation period,each patient could experience the critical event of repopula-tion, i.e., CD19+ cells above 1% with respective treatment,multiple times. Ultimately, there were 37 patients, who to-gether experienced the event 87 times.

The analytical procedure was 2-fold. First, controlling for thenumber of treatment cycles (1st, 2nd, 3rd, and so forth), weassessed the influence of various risk factors, such as age, sex,and BSA, on the likelihood of the critical event. For thispurpose, we performed Cox proportional hazards regression.Hazard ratios (HRs) below 1 indicated decreased risk of thecritical event depending on a factor, whereas the HR above 1indicated increased risk.

Second, we were interested in the time until the (re)oc-currence of the event, i.e., when CD19+ cells were mea-sured above 1%. Therefore, the model only included thoseobservations in which the critical event had occurred (n =87). We regressed the number of days since the previoustreatment on age, sex, previous therapy, and BSA in a linearrandom effects regression model, while controlling for thenumber of treatment cycles. We assessed the unadjustedand adjusted effects of all risk factors in both procedures

GlossaryBSA = body surface area; FACS = fluorescence-activated cell sorting; HR = hazard ratio; NMO = neuromyelitis optica;NMOSD = neuromyelitis optica spectrum disorder; SE = standard error.



and performed sensitivity analyses on missing values forthe BSA.

ResultsCohort characteristicsThe cohort consisted of 45 patients (62.2% female) agedbetween 17 and 76 years at their first treatment cycle (mean ±SD: 44.5 ± 15.0 years) (table 1). The disease spectrum in-cluded mainly patients with NMOSDs and MS (figure 1A).Forty-two patients received rituximab; 3 had received ocreli-zumab in clinical studies and were monitored thereafter. Themajority (71.1%) of patients received the BSA-adapteddosage of 375 mg/m2, whereas 28.9% received fixed doses(10 patients: 300 mg and 3 patients: 1,000 mg). Of note,37.8% of patients were treatment naive without previousimmune therapy treatment. Among previous therapies, aza-thioprine, cyclophosphamide, and interferon-beta were mostfrequent. The washout period was adapted according to thegeneral guidelines, and lymphocyte counts were normal be-fore starting the new therapy. Only patients with no otherconcomitant immunosuppressive therapy have been included.For 37 patients, we recorded at least 1 event (when CD19+

cells surpassed 1% of lymphocytes after CD19+ cell depletion);in total, 87 treatment cycles could be used for data analysis inour cohort. Averaged CD19+ cells first reached 1% after253 days or 8.3 months (mean ± SD: 253.37 ± 81.67) afterlast treatment. Earliest repopulation occurred after 108days (3.6 months), and latest repopulation after 554 days(18.2 months).

BSA-dependent B-cell recovery after treatmentwith 375 mg/m2

A total of 305 observations had complete information withoutmissing BSA values. This constituted data from 29 patientswho had collectively experienced a total of 79 events. TheBSA-adapted dosage of 375 mg/m2 resulted in a greater dosefor patients with a larger BSA (patient dosages ranged from574 to 975mg). Underlying diagnosis did not predict the timeuntil B-cell recovery. However, patients with a larger BSAwere at significantly higher risk of experiencing repopulationthan those with a smaller BSA (HR 1.015 per cm2, CI1.003–1.028, p < 0.05, fully adjusted Cox proportional haz-ards regression model; figure 1B, table e-1, links.lww.com/NXI/A45: model 7 and table e-2), although they received anadapted dose of 375 mg/m2. The fully adjusted randomeffects model revealed that with every 10 cm2 increase in theBSA, the time until CD19 cell repopulation was reduced by 14days (B = −1.389, standard error [SE] = 0.554, p < 0.05; tablee-3: model 7). The Kaplan-Maier curves of the 1st and 4thquartiles of the BSA underline this observation (HR 30.5, CI11.43–81.44, p < 0.001, Log-rank [Mantel-Cox] test; figure1C) by showing a faster B-cell repopulation for patients witha larger BSA. This phenomenon was also reflected in longertreatment intervals in patients with a small BSA and shortertreatment intervals in patients with a large BSA (figure 2A).

Several different equations for predicting the BSA frommeasurements of height and weight have been derived (e.g.,Dubois, Haycok, or Mosteller). The correlation between allthe formulas is generally high. The Dubois equation ismost frequently used, although specific recommendations forcalculating the BSA are not included in neurologic medical

Table 1 Baseline cohort characteristics

Patient characteristicsPatients(N = 45)

Rituximab, n 42

Ocrelizumab, n 3

Age at the first dose (y ± SD) 44.5 ± 15.0

Female, n (%) 28 (62.2)

No previous disease-modifying therapy, n (%) 17 (37.8)

Previous disease-modifying therapy, n (%) 28 (62.2)

Azathioprine 11 (24.4)

Interferon beta-1a/1b 10 (22.2)

Cyclophosphamide 8 (17.8)

Fingolimod 7 (15.6)

Natalizumab 6 (13.3)

Mycophenolate mofetil 5 (11.1)

Dimethyl fumarate 4 (8.9)

Glatiramer acetate 3 (6.7)

Mitoxantrone 3 (6.7)

Cyclosporine 2 (4.4)

Infliximab 2 (4.4)

Methotrexate 1 (2.2)

BSA-adapted dose 375 mg/m2, n (%) 32 (71.1)

Average dose (mg) 633.3 ±235.5

Fixed dose, n (%) 13 (28.9)

300 mg 10 (22.2)

1,000 mg 3 (6.7)

No. of events (CD19+ cells first surpass 1% afterdepletion) of all observations

87 of 373

Time when CD19+ cells first surpass 1%, d 251.2 ± 83.6

25th percentile 204.25

Median 245

75th percentile 302

CD19+ cells before the first treatment (%) (n = 18) 11.3 ± 5.8

Abbreviation: BSA = body surface area.Patients who received either rituximab or ocrelizumab in our neurologicclinic.Data are presented as mean ± SD, unless otherwise indicated.


http://links.lww.com/NXI/A45



guidelines. However, it has been previously suggested thatDubois might lead to lower values than Mosteller in patientswith large height and weight.10 To address this issue, wecompared resulting BSA-adjusted dosages calculated usingthe Dubois or Mosteller equation to determine the BSA(figure 2B). The resulting dosage differs for patients witha high BSA (4th quartile) also in our cohort (871.9 ± 10.22 vs896 ± 11.67 mg, p < 0.001, paired t test). For patients witha low BSA, there was no significant difference.

No influence of age, sex, pretreatment, CSF, orwhite blood cell counts on B-cell recoveryThe age of patients did not show a conclusive relationshipwith repopulation of CD19+ B cells (figure 3A, tables e-1 ande-2, links.lww.com/NXI/A45). However, although patientswho received any immune therapy before B-cell depletion(HR 1.444, CI 0.918–2.272, p > 0.05, Cox proportionalhazards regression models; figure 3B, table e-1: model 7, andtable e-2: model 6) and male patients (HR 1.453, CI0.809–2.607, p > 0.05, Cox proportional hazards regressionmodels; figure 3C) tended to repopulate earlier, a significanteffect might not have been achieved because of the limited

event numbers in the fully adjusted models (see also randomeffects model, tables e-3 and e-4). Leukocyte and lymphocytecounts did not show any correlation with B-cell repopulationbehavior. We divided patients into 2 groups: (1) patients whoexperienced on at least 1 occasion leukopenia (leukocytes <3,500/μL) or lymphopenia (lymphocytes < 1,000/μL) and(2) patients who never had a reduced leukocyte or lympho-cyte count throughout the B-cell depletion period. We foundno differences between patients who had experienced leuko-penia or lymphopenia and the remaining patients regardingdays until B-cell repopulation (258.2 ± 13.98 days [1], 251.3 ±11.15 days [2], p = 0.73, unpaired 2-sided students t test;figure 3D). Neither the presence of oligoclonal bands nor theabsolute cell count in the CSF after diagnosis but before thefirst treatment showed any correlation with B-cell recovery(figure 3E). The presence of aquaporin-4 antibodies ofpatients with NMOSD also did not show any correlation, andB-cell counts before treatment did not correlate with B-cellrepletion kinetics (B = −3.688, SE = 2.513, p = 0.142, tablee-5). In addition, absolute cell counts for T lymphocytes andnatural killer cells did not change over time after CD20+ celldepletion (figure 3F).

Figure 1 Faster repopulation of CD19+ cells in patients with a high BSA after treatment with 375 mg/m2 rituximab

(A) Disease spectrum of patients whoreceived rituximab or ocrelizumab.Other inflammatory neurologic diseases(OIND) included neurosarcoidosis,rheumatoid meningoencephalomyelitis,and neurolupus disease. (B) The timeuntil CD19+ cells are first detected above1% negatively correlates with the BSA inpatients who already received a BSA-adapted dosage of 375 mg/m2 (n = 79treatment cycles from 29 patients wereincluded, *p < 0.05, fully adjusted Coxproportionalhazards regressionmodel).(C) Survival analysis for the 1st and 4thquartiles of the BSA for surpassing 1%CD19+ cells with significant differences(***p < 0.001, log-rank [Mantel-Cox]test).




Results were not caused by outliers with a smaller or largerBSA but appeared to be robust against values from extremeends (i.e., lower and upper 5% quantiles). The unadjustedeffects of previous therapy and sex disappeared in the modelsadjusted for the BSA. This change likely was based on a re-duction in the sample size rather than adjustment. Additionaltests revealed that patients with missing values on the BSA rana lower risk of the event than patients with nonmissing values(table e-2, links.lww.com/NXI/A45). The limited statisticalpower and model fit do not yet permit conclusive statementson the remaining risk factors.

Intraindividual B-cell repopulation aftermultiple rituximab treatment cyclesSeveral patients received more than 1 rituximab treatmentcycle. In total, we could identify 18 patients who had sufficientdata to compare B-cell repopulation events over time in thesame individual. As depicted in figure 4A, there was no clearpattern observed in the B-cell recovery after the first and lastdocumented rituximab treatment (248.9 ± 13.5 days after thefirst rituximab treatment, 256.5 ± 15.6 days after the last

rituximab treatment, p = 0.71, unpaired 2-sided Student ttest). For 2 patients, B-cell recovery (surpassing 1% of CD19+

cells) could be analyzed for up to 8 treatment cycles (see 1example in figure 4B).

DiscussionThe efficacy of B-cell depletion, especially in progressive MS,has recently drawn interest toward the role of B lymphocytes inpathology and their relevance as therapeutic targets, as well astheir response to treatment with disease-modifying drugs suchas alemtuzumab and dimethyl fumarate. In our study, factorsthat influence the repopulation of B cells in neurologic patientsafter B-cell depletion were analyzed. Whereas obvious param-eters such as age, CSF characteristics, and leukocyte counts werenot involved in the interindividual variability of B-cell recovery,we identified the risk of a larger BSA for an earlier repopulationof B cells and thus potential disease activity. A systematic un-derestimation of rituximab dosage calculation based on thewidely used Dubois equation in patients with a high BSA wasidentified as an explanation. Calculating the BSA with the

Figure 2 Repopulation of CD19+ cells in low and high BSA patients and calculation of the BSA

(A) Comparison of CD19+ cell recovery in2 patients with high and low BSA overseveral treatment cycles. (B) The BSA(m2) calculated using the Dubois for-mula significantly underestimates dos-ages for our patients with a high BSA(4th quartile, 871.9 ± 10.22 vs 896 ±11.67 mg, p < 0.001, paired t test).




Mosteller equation partially overcomes this effect. However, theuse of the arbitrary dose of 375 mg/m2 is not sufficient butshould be increased especially for patients with a high BSA.

It is important to not miss the time point of B-cell repopu-lation, as this is most likely linked to resurging disease activity,

at least in patients with MS and neuromyelitis optica(NMO).11,12 Because of the large intraindividual variation ofB-cell repopulation, which has also been described for pedi-atric patients with NMO,12 close monitoring of B cells isrecommended with special attendance to high BSA patients tonot miss the early repopulators—the lowest interval was 108

Figure 3 Age, pretreatment, sex, and leukopenia do not influence CD19+ cell repopulation

(A) Age, (B) pretreatment, (C) sex (n = 87 treatment cycles from37patients, HR calculated from fully adjusted Coxproportional hazards regressionmodel, tablee-2, links.lww.com/NXI/A45), and (D) absolute lymphocyte or leukocyte counts (p = 0.73, unpaired 2-sided Student t test) or the presence of oligoclonal bands(OCBs) in the CSF (E, p = 0.41, unpaired 2-sided Student t test) do not influence the time until CD19+ cells reach 1% of lymphocytes. The box-and-whisker plotindicates the median value (center line), the 25th–75th percentiles (box), and the 10th–90th percentiles (whiskers). (F) Absolute counts of T lymphocytes andnatural killer cells did not significantly change in our cohort (mean ± SEM, one-way ANOVA with the Bonferroni multiple comparison test, p = 0.56) over timeafter CD20 depletion.




days (3.6 months) in our cohort. We propose to monitorthose patients initially after 2 and 4 months and biweekly toevery 4 weeks thereafter. Patient-adjusted treatment intervalsdecrease potential side effects (e.g., infusion reactions), de-crease patient efforts, and save costs as B cells on averagerepopulate after 8.3 months instead of 6months (according tothe treatment regimen with fixed doses).

Other factors such as age, sex, underlying diagnosis, andpretreatment with other immunosuppressive drugs did notshow a significant influence on B-cell repopulation. Smalleffects might have been overlooked because of the limitednumber of analyzed events. In male patients and patients withpretreatment, B cells tended to reappear sooner.We could notconfirm the association between leukopenia and prolongedrecovery of B cells as suggested before.13 The presence orabsence of oligoclonal bands in the CSF also did not show anycorrelation. Concomitant leukopenia and/or lymphopeniawere not associated with an impaired B-cell repopulationbehavior in our cohort. Also, the repeated application of rit-uximab in our patients did not lead to a habituation effect ofB-cell recovery. Besides, the overall efficacy and safety of thetreatment was good with no severe adverse events or clinicalrelapses in all patients during the study, in line with existingclinical data.

Randomized controlled trials assessing the efficacy and safetyof different dosage protocols for neurologic disorders aremostly lacking. In a study with patients with NMO, the du-ration of depletion was dose dependent as indicated bya repopulation time of 184 days with 1,000 mg vs 99 days with100 mg.6 The widely used BSA-adapted dosage of 375 mg/m2

was first introduced in early phase I lymphoma trials14,15

failing to demonstrate a clear relationship between the doseintensity and tumor response. The original regimen of ritux-imab (375 mg/m2 per week for 4 weeks) in adult neurologywas therefore mainly based on empirical concerns. Data from

rheumatoid arthritis studies favor a high-dose rituximab pro-tocol using 2 cycles of 1,000 mg, while optimal treatmentparadigms have not yet been defined.16,17 It should be kept inmind that necessary dosages of B cell–depleting antibodieswill most likely depend on the concentration and effect in thetarget organ itself to target tissue-resident B cells. Lymphnode biopsies performed in lymphoma patients show in-complete depletion18; moreover, the functional properties ofthe remaining B cells may change.19 However, in patients withMS, rituximab was shown to deplete B cells from the CSF andsupposedly also from brain tissue,20,21 while this direct CNSeffect is probably limited based on the fact that rituximabconcentration in the CSF reaches 2% of serum values.22

Future investigations using B cell–depleting therapies in neu-roimmunologic diseases should not only address the optimaldosing protocol based on the clinical efficacy and safety but alsoassess the use of novel biomarkers beyond measuring CD19+

B cells in the peripheral blood. Potentially interesting B-cellsubtypes arememory and effector B cells (e.g., CD27+ cells, butalso CD27−IgM−IgD− memory cells, late-stage lineage plas-mablasts, or increasingly acknowledged cytokine-producingB cells). Furthermore, gene polymorphisms in the FCGR3Agene encoding the FcγRIIIa18 have been suggested to predictthe efficacy of B cell–targeted therapies while these findingshave not yet been transferred to neuroimmunologic patients.

Author contributionsErik Ellwardt: designed the study, performed data analysis,and wrote the manuscript. Lea Ellwardt: performed statisticalanalysis. Stefan Bittner: designed the study and wrote themanuscript. Frauke Zipp: designed the study, evaluated data,and wrote the manuscript.

AcknowledgmentThe authors thank Cheryl Ernest for proofreading and editingthe manuscript.

Figure 4 Intraindividual repopulation of CD19+ cells following multiple treatment cycles

(A) Comparison of time until CD19+ cells were first detected above 1% after repetitive rituximab cycles in the same patients does not reveal any significantdifferences over time. (B) CD19+ cell repopulation of one of the patients from (A) (red line) in 8 rituximab treatment cycles.



Study fundingThis work was supported by the German Research Council(DFG, CRC-TR-128).

DisclosureE. Ellwardt and L. Ellwardt report no disclosures. S. Bittnerhas received consultation funds and travel compensation fromBiogen Idec, Sanofi-Genzyme, Roche, Novartis, and MerckSerono. F. Zipp received travel funding from Teva, Novartis,Merck Serono, Bayer, Biogen Idec, Ono Pharma, Genzyme,Sanofi-Aventis, and Octapharma; received research supportfrom Teva, Novartis, Merck Serono, Bayer, DFG, BMBFKKNMS, PCORI PRAG-MS, and Progressive MS Alliance;received consultation funds from Biogen Idec Germany, OnoPharma, Genzyme, Sanofi-Aventis, and Octapharma; andconsulted for Teva, Merck Serono, Novartis, Bayer Health-care, Biogen Idec Germany, Ono Pharma, Genzyme, Sanofi-Aventis, and Octapharma. Full disclosure form informationprovided by the authors is available with the full text of thisarticle at Neurology.org/NN.

Received December 22, 2017. Accepted in final form March 22, 2018.

References1. Hauser SL, Bar-Or A, Comi G, et al. Ocrelizumab versus interferon Beta-1a in re-

lapsing multiple sclerosis. New Engl J Med 2017;376:221–234.2. Hauser SL, Waubant E, Arnold DL, et al. B-cell depletion with rituximab in

relapsing–remitting multiple sclerosis. New Engl J Med 2008;358:676–688.3. Damato V, Evoli A, Iorio R. Efficacy and safety of rituximab therapy in neuromyelitis

optica spectrum disorders: a systematic review andmeta-analysis. JAMANeurol 2016;73:1342–1348.

4. Pellkofer HL, Krumbholz M, Berthele A, et al. Long-term follow-up of patients withneuromyelitis optica after repeated therapy with rituximab. Neurology 2011;76:1310–1315.

5. Yang CS, Yang L, Li T, et al. Responsiveness to reduced dosage of rituximab inChinese patients with neuromyelitis optica. Neurology 2013;81:710–713.

6. Greenberg BM, Graves D, Remington G, et al. Rituximab dosing and monitoringstrategies in neuromyelitis optica patients: creating strategies for therapeutic success.Mult Scler 2012;18:1022–1026.

7. Kim SH, KimW, Li XF, Jung IJ, Kim HJ. Repeated treatment with rituximab based onthe assessment of peripheral circulating memory B cells in patients with relapsingneuromyelitis optica over 2 years. Arch Neurol 2011;68:1412–1420.

8. Kim SH, Huh SY, Lee SJ, Joung A, Kim HJ. A 5-year follow-up of rituximab treatment inpatients with neuromyelitis optica spectrum disorder. JAMANeurol 2013;70:1110–1117.

9. Jacob A, Weinshenker BG, Violich I, et al. Treatment of neuromyelitis optica withrituximab: retrospective analysis of 25 patients. Arch Neurol 2008;65:1443–1448.

10. Fancher KM, Sacco AJ, Gwin RC, Gormley LK, Mitchell CB. Comparison of twodifferent formulas for body surface area in adults at extremes of height and weight.J Oncol Pharm Pract 2016;22:690–695.

11. Stuve O, Leussink VI, Frohlich R, et al. Long-term B-lymphocyte depletion withrituximab in patients with relapsing-remitting multiple sclerosis. Arch Neurol 2009;66:259–261.

12. Nosadini M, Alper G, Riney CJ, et al. Rituximab monitoring and redosing in pediatricneuromyelitis optica spectrum disorder. Neurol Neuroimmunol Neuroinflamm 2016;3:e188.

13. Radaelli M, Moiola L, Sangalli F, et al. Neuromyelitis optica spectrum disorders: long-term safety and efficacy of rituximab in Caucasian patients. Mult Scler 2016;22:511–519.

14. Maloney DG, Grillo-Lopez AJ, White CA, et al. IDEC-C2B8 (Rituximab) anti-CD20monoclonal antibody therapy in patients with relapsed low-grade non-Hodgkin’slymphoma. Blood 1997;90:2188–2195.

15. Maloney DG, Liles TM, Czerwinski DK, et al. Phase I clinical trial using escalatingsingle-dose infusion of chimeric anti-CD20 monoclonal antibody (IDEC-C2B8) inpatients with recurrent B-cell lymphoma. Blood 1994;84:2457–2466.

16. Buch MH, Smolen JS, Betteridge N, et al. Updated consensus statement on the use ofrituximab in patients with rheumatoid arthritis. Ann Rheum Dis 2011;70:909–920.

17. Bredemeier M, de Oliveira FK, Rocha CM. Low- versus high-dose rituximab forrheumatoid arthritis: a systematic review and meta-analysis. Arthritis Care Res 2014;66:228–235.

18. Ternant D, Cartron G, Henin E, TodM, Girard P, Paintaud G. Model-based design ofrituximab dosage optimization in follicular non-Hodgkin’s lymphoma. Br J ClinPharmacol 2012;73:597–605.

19. Kamburova EG, Koenen HJ, Borgman KJ, Ten Berge I, Joosten I, Hilbrands LB. Asingle dose of rituximab does not deplete B cells in secondary lymphoid organs butalters phenotype and function. Am J Transplant 2013;13:1503–1511.

20. Cross AH, Stark JL, Lauber J, Ramsbottom MJ, Lyons JA. Rituximab reduces B cellsand T cells in cerebrospinal fluid of multiple sclerosis patients. J Neuroimmunol 2006;180:63–70.

21. Martin Mdel P, Cravens PD, Winger R, et al. Depletion of B lymphocytes fromcerebral perivascular spaces by rituximab. Arch Neurol 2009;66:1016–1020.

22. Komori M, Lin YC, Cortese I, et al. Insufficient disease inhibition by intrathecalrituximab in progressive multiple sclerosis. Ann Clin translational Neurol 2016;3:166–179.




ARTICLE OPEN ACCESS

Neopterin is associated with hippocampalsubfield volumes and cognition in HIVDebra A. Fleischman, PhD, Konstantinos Arfanakis, PhD, Sue Leurgans, PhD, Sheila M. Keating, PhD,

Melissa Lamar, PhD, David A. Bennett, MD, Oluwatoyin M. Adeyemi, MD, and Lisa L. Barnes, PhD


Correspondence

Dr. Fleischman

[email protected]

AbstractObjectiveHIV infection sets off an immediate immune response and inflammatory cascade that can leadto neuronal injury and cognitive impairment, but the relationship between immune markers,regional brain volumes, and cognition remains understudied in HIV-infected adults.

MethodsCross-sectional associations were examined between serum immune markers of activation(neopterin) and inflammation (interleukin [IL]-1β, IL-6, tumor necrosis factor alpha, andC-reactive protein) with regional brain volumes (cortical, subcortical, total gray matter,hippocampus, and subfields) and cognition in 66 HIV-infected, virally suppressed, adults whounderwent 3.0-T MRI as part of the Research Core of the Rush Center of Excellence onDisparities in HIV and Aging. Immune markers were assayed from frozen plasma, values wereentered into linear regression models as predictors of regional brain volumes, and interactiveeffects of immune response and regional brain volumes on cognition were examined.

ResultsNo inflammatory marker was associated with any regional brain volume. Higher neopterin levelwas associated with lower total hippocampal, presubiculum, and cornu ammonis (CA) subfieldvolumes. Higher neopterin level and lower total hippocampal volume were independentlyassociated with lower episodic memory, and neopterin level fully mediated the effect ofhippocampal atrophy on episodic memory. Higher neopterin levels were associated with lowerpresubiculum, CA1, and CA4/dentate volumes and lower semantic memory, working memory,and global cognition.

ConclusionImmune activation in response to HIV infection, measured by neopterin, has a deleterious andtargeted effect on regional brain structure, which can be visualized with clinically available MRImeasures of hippocampus and its subfields, and this effect is associated with lower cognitivefunction.

From the Rush Alzheimer’sDiseaseCenter (D.A.F., K.A., S.L., M.L., D.A.B., L.L.B.), Rush University Medical Center; the Department of Neurological Sciences (D.A.F., S.L., M.L., D.A.B., L.L.B.),the Department of Behavioral Sciences (D.A.F., M.L., L.L.B.), the Department of Preventive Medicine (S.L.), the Department of Diagnostic Radiology and Nuclear Medicine (K.A.), RushUniversity Medical Center; Ruth M. Rothstein CORE Center (O.M.A.); the Biomedical Engineering (K.A.), Illinois Institute of Technology, Chicago; the Blood Systems Research Institute(S.M.K.), San Francisco, CA; and the University of California at San Francisco (S.M.K.), Laboratory Medicine.









It has long been known that HIV infection produces an almostimmediate immune activation and inflammatory cascade thatcan persist in the face of effective highly active antiretroviraltherapy (HAART) treatment1 and can lead to neuronal in-jury2 and cognitive impairment.3 The earliest phase of theHIV immune response occurs when interferon-stimulated(IFN-γ) T-lymphocytes activate monocytes, macrophages,and dendritic cells to produce pteridines, a class of chemicalsthat stimulate the inflammatory cascade through the release ofcytokines (e.g., interleukin [IL]-1β, IL-6, and tumor necrosisfactor alpha [TNFα]) and acute phase reactants (e.g.,C-reactive protein [CRP]).

Neopterin, a pteridine produced in monocyte-derived mac-rophages4 and dendritic cells,5 is a well-established earlymarker of HIV-associated cellular immune activation.6 Theneopterin level correlates with measures of viral load andCD4 count, is a sensitive measure of treatment efficacy andHIV disease course,7–9 and predicts HIV-related mortality.10

Neopterin level is associated with compromise of blood-brain barrier integrity and generalized brain atrophy in HIV11

and HIV-associated cognitive impairment.2 Thus, neopterinlevel may signal activation of HIV-related immune mecha-nisms that can lead to neuronal injury, perhaps in regionsmost susceptible to metabolic disturbance, neurologic dis-ease and alterations in cognitive function, but no study of thefiner associations of the neopterin level with MRI-derivedregional brain volumes and cognition in HIV has beenreported.

Here, we examined relationships between immune activa-tion measured by serum neopterin, inflammation measuredby serum markers (IL-1β, IL-6, TNF-α, and CRP) knownto be present in the HIV immune response, even withsuccessful HAART treatment, and associated with brainabnormalities and cognitive impairment and decline,12–143T-MRI-derived regional brain structure, and cognition in HIV-infected adults. Given the evidence that HIV is associatedwith increased levels of immune activation and chronic in-flammation, we tested the hypothesis that higher level ofimmune activation, measured by neopterin, and higher in-flammation, measured by these 4 inflammatory markers,would be specifically associated with volume loss in thehippocampus, a brain region known to be highly susceptibleto metabolic stress and neurologic disease15 and HIV in-fection.16 Furthermore, we hypothesized that this volumereduction would be associated with lower episodic memory,which depends critically on hippocampal integrity.

MethodsParticipantsThe study was conducted among participants in the ResearchCore of the Rush Center of Excellence on Disparities in HIVand Aging (CEDHA), a longitudinal study of older adultswith HIV infection. HIV+ participants (n = 175) wererecruited from an urban infectious disease clinic. Eligibilityrequirements included documented HIV infection, AfricanAmerican, or white, age 50 and older, CD4+ ≥ 200 cells/mm3

on HAART or CD4+ ≥ 500 HAART naive, and viral loadsranging from undetectable up to 50,000 copies. All partic-ipants underwent annual structured clinical evaluations thatincluded an assessment of risk factors for cognitive decline,a battery of cognitive function tests, and a blood draw.

HIV+ participants who were willing to undergo MRI wereapproached if they were nondemented and able to give in-formed consent. Participants were excluded if there was anycontraindication for MRI, history of head injury with loss ofconsciousness over 30 minutes, known large lesions, MS orhydrocephalus, seizure disorder, schizophrenia, current use ofneuroleptic, narcotic medication, or anti-inflammatory med-ications. Seventy-eight of the 175 participants recruited intothe research core were ineligible based on the above criteria.Of those 78 remaining, 11 participants were eligible but couldnot be contacted, 11 participants were reluctant to scan, 5 hadunsuccessful scans, and 4 did not have blood available forassay, making the final sample 66 participants.

Standard protocol approvals, registrations,and patient consentsThe study was approved by the Rush Institutional ReviewBoard, and all participants signed study consent.

Cognitive testingEach participant underwent cognitive function testing thatincluded 18 tests of episodic memory, semantic memory,working memory, perceptual speed, and visuospatial ability.17

See supplementary materials for details, links.lww.com/NXI/A47.

Serum immune activation andinflammatory markersNeopterin was measured in undiluted plasma by ELISA(Alpco, Salem, NH; standard curve measuring 0.5–63 ng/mLwith 3% intraplate and 7% interplate coefficient of variation[CV]%). ELISA results were acquired using MolecularDevices E-Max plate reader with Softmax Pro v5.4. MILLI-PLEXMAPHuman High Sensitivity T Cell Panel was used to

GlossaryCA = cornu ammonis; CEDHA = Center of Excellence on Disparities in HIV and Aging; CRP = C-reactive protein; CV =coefficient of variation; FDR = false discovery rate; HAART = highly active antiretroviral therapy; IL = interleukin; INF =interferon; TNFα = tumor necrosis factor alpha.





measure IL-1β, IL-6, and TNF-α in undiluted plasma (Milli-pore, Billerica, MA; standard curve measuring 0.18–2,000 pg/mL with 7% intra- and inter-plate CV). MILLIPLEX MAPCardiovascular Disease Panel 3 was used to measure CRP indiluted plasma (1: 40,000) (standard curve measuring0.01–50 ng/mL with 1% intraplate CV% and 8% interplateCV%). See supplementary materials for details, links.lww.com/NXI/A47.

Brain image acquisition and processingAll MRI data were collected on a 3-T Philips MRI scanner(Best, Netherlands) with a 32-channel head coil. See supple-mentarymaterials for further details (links.lww.com/NXI/A47)on MRI acquisition and postprocessing.18

Statistical approachMRI data were transferred to SAS (SAS Institute, Inc., Cary,NC, 2012) for further data manipulation, linkage with de-mographic and serum immune marker data, and statisticalanalyses. Mean, SD, skewness coefficients, Q-Q plots, andhistograms were obtained for age, education, CD4, serummarkers, brain, and cognitive measures. Because the dis-tributions of neopterin and CRP were markedly skewed,logarithmic transformations (base 10) were used in all anal-yses. Because the other 3 markers were below the detectionlimit in many participants, we used Spearman correlations indescriptive statistics, and we used dichotomized values(detected/not detected) in analytic models. First, we exam-ined the correlations between demographic and diseasecharacteristics with immune markers and with brain variables;then, we used regression models (with brain measures asoutcomes) to examine the association of each of the immunemarkers with whole-brain measures (total cortical, total sub-cortical, total gray, and total hippocampus) with terms tocorrect for age, education, race, sex, and CD4 level. Based onthe findings, we then examined the associations of neopterinwith hippocampal subfield volumes, with false discovery rate(FDR) corrections for multiple comparisons. Finally, becauseof known associations of hippocampus volume with memory,we first considered episodic memory as the outcome ofa multiple regression model with terms for neopterin andsubfield volume and their interaction. In secondary models,we replaced episodic memory with semantic memory, work-ing memory, perceptual speed, and visuospatial ability. Allregression coefficients are reported per SD of the outcome;for continuous covariates, the regression coefficients are basedon standardized covariates.

Data availabilityAnonymized data not published within this article will beshared by request from any qualified investigator.

ResultsDescriptive information for demographic, disease status, im-mune markers, brain volume, and cognitive measures is givenin table 1.

First, we examined correlations between the demographic,disease, and immune marker variables No immune marker ordisease variable was associated with race or age (ps > 0.05),and females had higher CRP level (female: mean [log10] 4.17,SD = 0.55; male: mean = 3.81, SD = 0.52, t = 2.33, p = 0.023).

Persons with higher neopterin level had lower CD4 counts(r = −0.369, p < 0.01), but there was no association with nadirCD4 or disease duration. No inflammatory marker was as-sociated with CD4 or any other disease variable. The in-flammatory markers were significantly interrelated, but nonewas related to neopterin. Correlations between the immunemarker variables are shown in supplementary materials, tablee-1, links.lww.com/NXI/A47.

Next, we examined correlations between the demographicand disease severity variables with regional brain volumes. Allselected regions had higher volumes with lower age andhigher CD4 count (ps < 0.05). There were no significantassociations with sex or race.

In a set of linear regression models, corrected for de-mographics (age, education, sex, and race) and CD4, we nextexamined associations between immune markers and totalcortical, total subcortical, total gray, and total hippocampalvolume. No immune marker was associated with total cor-tical, total subcortical, or total gray matter volume. Personswith higher neopterin levels had lower total hippocampalvolumes (estimate = 0.398, SE = 0.128, p = 0.003). Estimates,SEs, and p values are shown in table 2. The full model forneopterin and total hippocampal volume is shown in sup-plementary materials, table e-2, links.lww.com/NXI/A47.Because neopterin was the only immune marker associatedwith brain volume, no further analyses were performed withinflammatory markers. The remaining analyses concentratedon associations between neopterin, regional hippocampalvolumes, and cognition.

Next, a finer examination of the associations of neopterin withhippocampal volume was undertaken using 7 subfield vol-umes: presubiculum, subiculum, cornu ammonis (CA)1,CA2/3, CA4/dentate, fimbria, and hippocampal fissure.Volumes of presubiculum, subiculum, and fimbria were higherwith lower age (ps < 0.05). There were no associations withsex or race. Persons with higher CD4 count had higher vol-ume in all fields except the hippocampal fissure (ps < 0.05).

In a series of linear regression models, corrected for de-mographic variables and CD4, persons with higher neopterinlevel had lower volumes of presubiculum, and the CA (CA1,CA2/3, and CA4/dentate) fields, after correcting probabilitylevels for multiple comparisons using FDR. Estimates, SEs,and p values for the regression models are shown in table 3.

As a final step, we examined the associations of neopterin andregional brain volumes first with episodic memory in separatelinear regression models controlled for demographic and








disease variables, and then, in secondary analyses, with globalcognition, semantic memory, working memory, perceptualspeed, and visuospatial ability. Persons with lower neopterinlevels (estimate = −0.347, SE = 0.126, p < 0.01) and higherhippocampal volumes (estimate = 0.257, SE = 0.121, p < 0.05)had better episodic memory. When neopterin and hippo-campal volumes were entered in a single model, the associ-ation between neopterin and episodic memory was slightly

attenuated, but not eliminated (estimate = −0.284, SE =0.135, p = 0.039), but the association between hippocampalvolume and episodic memory was eliminated (p = 0.22),suggesting that neopterin level fully mediated the associationbetween hippocampal volume and episodic memory. Theinteraction of neopterin and hippocampal volume was notsignificant. Full models are shown in supplementary mate-rials, table e-3, links.lww.com/NXI/A47.

Table 1 Demographic characteristics, HIV disease status, immune, brain, and cognitive measures (n = 66)

Measure Mean SD Minimum Maximum

Age at scan 58.5 5.3 50.6 71.3

Years education 13.5 3.2 8.0 24.0

Male, n (%) 48 (73%)

Black, n (%) 45 (68%)

CD4 638.5 313.5 228.0 1,872.0

Nadir CD4a 276.4 191.2 6.0 756.0

Disease duration (y) 16.0 7.1 1.0 32.0

Neopterinb 0.42 0.17 0.14 0.88

IL-1β, n (%) 35 (53%)

IL-6, n (%) 44 (67%)

TNF-α, n (%) 52 (79%)

CRPc 3.91 0.58 2.65 5.09

Cortical gray volumed 262.99 21.58 211.03 308.28

Subcortical gray volume 36.09 3.82 28.60 44.64

Total gray volume 358.38 30.61 287.20 414.26

Hippocampal volume 5.14 0.71 3.61 6.87

Presubiculum 0.60 0.09 0.42 0.81

Subiculum 0.80 0.11 0.58 1.10

CA1 0.42 0.05 0.29 0.53

CA2/3 1.74 0.17 0.87 1.63

CA4/dentate 0.70 0.10 0.50 0.92

Fimbria 0.07 0.03 0.01 0.12

Hippocampal fissure 0.06 0.03 0.02 0.16

Global cognitionc 0.15 0.55 −0.97 1.22

Episodic memory 0.25 0.67 −1.26 1.59

Semantic memory 0.07 0.79 −1.65 1.56

Working memory 0.16 0.78 −1.43 1.85

Perceptual speed 0.09 0.82 −1.68 1.85

Visuospatial ability 0.04 0.80 −1.38 1.57

Abbreviations: CA = cornu ammonis; CRP = C-reactive protein; IL = interleukin; TNF-α = tumor necrosis factor alpha.a n = 53.b Log10-transform of units.c All brain volumes are tenths of percents of intracranial volume.d All cognitive scores converted to Z-scores (mean = 0).




Interactions between the neopterin level and hippocampalsubfield volumes were statistically significant for other cognitivedomains. The standardized interaction term of the neopterinlevel with CA1 for semantic memory (estimate = 0.321, SE =0.114, p < 0.01), CA4/dentate for semantic memory (estimate= 0.29, SE = 0.14, p = 0.04), presubiculum for global cognition(estimate = 0.215, SE = 0.106, p = 0.047), and presubiculum forworking memory (estimate = 0.262, SE = 0.121, p = 0.034),were all statistically significant. Figure 1 illustrates the associa-tion of hippocampal subfield volumes to cognition for the 2extreme quintiles (10th and 90th) of the neopterin level. Theobservation that the slopes for low levels of neopterin arenegative and the slopes for high levels of neopterin are positivereflects the fact that all the interaction terms are positive.

DiscussionIn this study, 66 HIV-infected adults who had higher serumneopterin levels, a measure of cellular immune activation, had

lower total hippocampal volume and lower presubiculum andCA subfield volumes. Episodic memory was better for personswith lower neopterin levels and for persons with higher hip-pocampal volumes. In mediation analyses, neopterin levelfully accounted for the effect of total hippocampal volume onepisodic memory. In hippocampal subfield analyses, personswith higher neopterin levels had lower CA1, CA4/dentate,and presubiculum volumes and lower semantic memory,working memory, and global cognition.

Neopterin production is the complex result of direct stimu-lation from IFN-γ, the only cytokine that directly induces it,19

and numerous indirect effects and interactions betweenmultiple cytokines and thus reflects the entirety of immunenetwork response on monocytes, macrophages, and dendriticcells.20Neopterin is often elevated before antibody serocon-version or the onset of any clinical symptoms21 and hencemay be an earlier and more sensitive indicator of HIV-associated immune response than any single, or any set of,inflammatory markers.6,20 In this study, we measured 4 in-flammatory markers that are known to be present in HIVinfection and associated with brain abnormalities and cogni-tive impairment and decline, IL-1β, IL-6, TNF-α, and CRP,but none was associated with brain volume. Only neopterin,a marker for immune activation, was associated with CD4,a marker of disease severity, and with regional brain volumes,specifically the hippocampus and its subfields.

No brain region is as susceptible to noxious insult as thehippocampus.15 The pyramidal neurons of the CA subfields,particularly CA1,22 are well known to be vulnerable toanoxia-ischemia,23 metabolic stress (e.g., hypoglycemia24),microvascular pathology,25 and neurodegenerative diseasesuch as Alzheimer disease and vascular dementia.26 HIV viralload is high in the hippocampus27 and its subfields, withpyramidal neurons in CA3 possibly being more vulnerable tothe virus than CA1,16 and reduced hippocampal volumeshave been demonstrated in HIV-infected adults.28 Theresults of this study suggest that hippocampal atrophy thatoccurs in HIV is associated with immune activation, asmeasured by neopterin.

Table 2 Associations of immune markers and brain volumes (estimate, SE)a

Cortical gray Subcortical gray Total gray Hippocampus

Neopterin −0.223 (0.127) −0.080 (0.131) −0.172 (0.130) −0.398 (0.128)b

IL-1β 0.049 (0.236) 0.230 (0.236) 0.093 (0.239) 0.072 (0.252)

IL-6 0.142 (0.241) 0.240 (0.241) 0.195 (0.244) 0.101 (0.256)

TNF-α −0.111 (0.279) −0.049 (0.281) −0.089 (0.284) −0.156 (0.297)

CRP −0.073 (0.116) 0.051 (0.117) −0.041 (0.118) 0.022 (0.123)

Abbreviations: CRP = C-reactive protein; IL = interleukin; TNF-α = tumor necrosis factor alpha.a Regression coefficients fromseparatemodels for brain volumeswith a term for an immunemarker,models controlled for age, sex, education, race, andCD4status. IL1-β, IL-6, and TNF-α were included as binary variables (detected/not detected), and the coefficient is scaled to the SD of the brain volume. Forneopterin and CRP, the coefficients are standardized regression coefficients.b p < 0.01.

Table 3 Associations of neopterin and hippocampalsubfield volumes (estimate, SE)a

Subfield Estimate SE Corrected p valueb

Presubiculum −0.31 0.12 0.015

Subiculum −0.30 0.13 0.061

CA1 −0.46 0.14 0.002

CA2/3 −0.46 0.14 0.002

CA4/dentate −0.46 0.13 0.003

Fimbria −0.19 0.11 0.135

Fissure −0.18 1.28 1.089

Abbreviation: CA = cornuammonis.a Estimated from separate models of standardized hippocampal subfieldoutcome with a term for standardized neopterin, controlled for age, sex,education, race, and CD4 status. Thus, all coefficients in this table areexpressed as standardized regression coefficients.b FDR (false discovery rate) corrected for multiple comparisons.Bold values are statistically significant.



Hippocampal integrity is critical for normal memory func-tion,29 and in this study, hippocampal volume was, indeed,positively associated with episodic memory. Persons withhigher neopterin levels had worse episodic memory, and, inmediation analyses, the neopterin level fully accounted for theeffect of hippocampal atrophy on episodic memory. Theseresults suggest that immune activation, as measured by neo-pterin, has a clear and robust deleterious effect on episodicmemory in HIV-infected adults.

In hippocampal subfield analyses, persons with higher neo-pterin levels had lower volumes in 2 CA subfields (CA1 andCA/4dentate) and lower semantic memory. Episodic andsemantic memory are considered long-term, declarativememory processes known to be strongly dependent on hip-pocampal circuitry,30 particularly the CA subfields, with CA4/dentate known to be a critical region for encoding in-formation, and CA1 known to be a critical region for retrieval

of that information.31 The results of this study suggest thatimmune activation in HIV-infected adults, measured byneopterin, may signal early volume reductions in hippocampalsubfields that are responsible for normal encoding and re-trieval of declarative information.

Although it has long been known that the hippocampus iscritical for the encoding and retrieval of long-term declarativememories, more recently, the region has been shown to playa central role in other cognitive processes such as workingmemory32 or the ability to maintain and manipulate multipleitems simultaneously in short-term memory. The hippo-campus is recruited in complex working memory span tasksthat involve processing relations between item features,33 andthe subiculum subfield has been shown to be preferentiallyactivated in response to discriminating highly associated fea-tures from multiple viewpoints in a scene discriminationtask.34 In this study, persons with higher levels of neopterin

Figure Association of hippocampal subfield volumes to cognition by neopterin level, volumes, and cognition adjusted forage, education, race, sex, and CD4

Colored lines show estimated cognition vs volume for extreme percentiles of neopterin (90th percentile in blue and 10th percentile in red). Solid points aredata for persons in extreme quintiles (blue, 5th quintile; red, 1st quintile). Data for persons in the middle quintiles of neopterin are shown in open circles.



had lower volumes of the presubiculum, 1 of 4 componentregions of the subiculum, and lower working memory andglobal cognition. This finding is consistent with the notionthat immune activation, measured by neopterin, may flagHIV-associated neuronal injury in a vulnerable subfield andconsequent risk for global cognition and working memoryimpairments. Overall, the results of this study suggest thatneopterin level may serve as a proxy for ongoing risk of HIV-associated brain injury and neurocognitive disorder, but lon-gitudinal studies are needed.

There are multiple mechanisms potentially linking neopterinbiology with neuronal damage, neurodegeneration, and con-sequent cognitive impairment in HIV infection. One candi-date mechanism is the disruption by chronic immuneactivation of the kynurenine pathway of tryptophan oxidativemetabolism, which can induce cell damage through dysregu-lation of NMDA receptor function or generation of reactiveoxygen species.35 Higher neopterin level is associated withlower serum tryptophan, brain atrophy, and dementia in HIVinfection,11 and neurotoxic glutamate is known to be elevatedin the CSF and plasma of patients with HIV.36CA1 hippo-campal cells have a high expression of NMDA receptors(which are critical for learning and memory) and are thusvulnerable to glutamate-mediated excitotoxicity.37 Indeed, theintegrity of this pathway has been implicated in other medicalconditions that involve chronic immune activation, brain in-jury, and cognitive impairment including Alzheimer disease,38

metabolic syndrome, and vascular cognitive impairment.39

Thus, it is feasible that chronic immune activation in HIVinfection could be leading to neuronal loss and diminishedcognitive function through an effect on the kynureninepathway, but the biological mechanisms underlying the as-sociation between chronic immune activation, brain integrity,and cognitive function in HIV infection are complex and re-main understudied.

This study had important strengths and limitations. First, theimmune markers for this study were selected to target im-mune activation and inflammation, but only neopterin wasassociated with regional brain neuronal injury and cognitiveimpairment. Considering that the study participants were fullyvirally suppressed, low levels of plasma inflammatory markersmay not have the dynamic range to identify associations withmarkers of HIV-induced pathology. Low concentrations forthese cytokines are not uncommon in well-controlled HIVinfection, and it is known that there is variable detectability ofIL-6, TNF-α, and IL-1β in HIV-infected individuals. The ro-bust associations of neopterin with regional brain volumesand cognition shown here support previous studies demon-strating that subclinical CNS damage and consequent cogni-tive impairment do occur in HIV even when the virus is wellcontrolled.2 Second, the number of subjects was small, andthe lack of associations between regional brain volumes andthe chosen inflammatory markers may be due to inadequatepower, especially for the effects of the interaction of neopterinwith brain volumes on cognition. However, the inflammatory

markers were highly intercorrelated, and not related to neo-pterin, and associations with neopterin were robust, sug-gesting that the neopterin level may, indeed, be a window onan upstream immune response that is potentially damaging tothe integrity of brain volume and cognition. Third, we usedserum neopterin rather than CSF neopterin; however,a number of studies have shown that serum and CSF neo-pterin are correlated21 and that serum neopterin parallels HIVRNA levels.40 Fourth, the age range of our sample was limited,and the findings cannot be generalized to much younger orolder HIV-infected adults. Fifth, the analyses did not adjustfor premorbid general intelligence, which could influenceindividual level of cognitive reserve available to participantsafter contracting HIV. Sixth, the study is cross-sectional andtherefore cannot demonstrate a causative relationship be-tween neopterin level and brain volume and cognition. Fi-nally, although our sample is small and cross-sectional, it wasexceptionally well characterized within the parent CEDHAstudy, which is a longitudinal prospective study of HIV, andwe will be able to perform longitudinal analyses in the future.Limitations notwithstanding, the results of this study suggestthat the serum neopterin level and 3T neuroimaging, both ofwhich are clinically available and relatively easy to obtain, maybe useful in identifying those HIV-infected adults who are atrisk of developing further neuronal injury and cognitive im-pairment. This information could prove invaluable for earlyimplementation of targeted pharmacologic and behavioralinterventions.

Author contributionsStudy concept and design: D.A. Fleischman and K. Arfanakis.Analysis or interpretation of the data: D.A. Fleischman,K. Arfanakis, S. Leurgans, S.M. Keating, M. Lamar, D.A.Bennett, and L.L. Barnes. Drafting of the manuscript: D.A.Fleischman, K. Arfanakis, S. Leurgans, S.M. Keating,M. Lamar, D.A. Bennett, and L.L. Barnes. Statistical analysis:S. Leurgans. Obtained funding: D.A. Fleischman, K. Arfanakis,D.A. Bennett, O.M. Adeyemi, and L.L. Barnes.

AcknowledgmentThe authors thank the staff of the Rush Center of Excellenceon Disparities in HIV and Aging, the research staff of the RuthM. Rothstein Core Center, and the staff of the RushAlzheimer’s Center, particularly Woojeong Bang, MS, forstatistical analysis and Niranjini Rajendran, MS, for imagepostprocessing.

Study fundingThis study was supported by the National Institute of Mi-nority Health and Health Disparities grant P20MD6886;National Institute of Aging grant P30AG010161; and IllinoisDepartment of Public Health.

DisclosureD.A. Fleischman received research support from the NationalInstitute ofMinorityHealth andHealthDisparities. K. Arfanakisserved on the editorial board of Brain Imaging and Behavior and



received research support from the NIA and NINDS. S. Leur-gans serves as an associate editor of statistics for Neurology andreceived research support from the NIH. S.M. Keating receivedresearch support from the NIH/NHLBI, NIH/NIAID, USFDA, amfAR Institute for HIV Cure. M. Lamar received re-search support from the NIA, NINDS, and NHLBI. D.A.Bennett served on the scientific advisory board of VigorousMinds, Takeda Pharmaceuticals, AbbVie, Indian Institute ofScience, MCSA, Columbia ADRC, CCNA, CIMA-Q, WFU,Emory, MODEL-AD, REGARDS, HRS, MIDUS, and NationalAdvisory Council on Aging; serves on the editorial board ofNeurology, Current Alzheimer Research, and Neuroepidemiology;and received research support from the NIH. O.M. Adeyemiconsulted for Gilead and received research support from Abb-Vie. L.L. Barnes serves on the editorial board of the Journal ofAging and Health and received research support from the NIA.Full disclosure form information provided by the authors isavailable with the full text of this article at Neurology.org/NN.

Received October 16, 2017. Accepted in final form April 4, 2018.

References1. Heaton RK, Clifford DB, Franklin DR Jr, et al; CHARTER Group. HIV-associated

neurocognitive disorders persist in the era of potent antiretroviral therapy: CHAR-TER Study. Neurology 2010;75:2087–2096.

2. Anesten B, Yilmaz A, Hagberg L, et al. Blood-brain barrier integrity, intrathecalimmunoactivation, and neuronal injury in HIV. Neurol Neuroimmunol Neuro-inflamm 2016;3:e300.

3. Ances BM, Ellis RJ. Dementia and neurocognitive disorders due to HIV-1 infection.Semin Neurol 2007;27:86–92.

4. Huber C, Batchelor JR, Fuchs D, et al. Immune response-associated production ofneopterin:release from macrophages primary under control of interferon-gamma.J Exp Med 1984;160:310–316.

5. Wirleitner B, Reider D, Ebner S, et al. Monocyte-derived dendritic cells releaseneopterin. J Leukoc Biol 2002;72:1148–1153.

6. Wachter H, Fuchs D, Hausen A, Reibnegger G, Werner ER. Neopterin as a marker foractivation of cellular immunity: immunologic basis and clinical application. In: SpiegelHE, editor. Advances in Clinical Chemistry, vol 27. San Diego, CA: Academic Press;1989:82–141.

7. FuchsD, SpiraTJ,HausenA, et al. Neopterin as a predictivemarker for disease progressionin human immunodeficiency virus type 1 infection. Clin Chem 1989;35:1746–1749.

8. Mildvan D, Spritzler J, Grossberg SE, et al. Serum neopterin, an immune activationmarker, independently predicts disease progression in advanced HIV-1 infection. ClinInfect Dis 2005;40:853–858.

9. Wirleitner B, Schroecksnadel I, Winkler C, Fuchs D. Neopterin in HIV-1 infection.Mol Immunol 2005;42:183–194.

10. Stein DS, Lyles RH, Graham NM, et al. Predicting clinical progression or death insubjects with early-stage human immunodeficiency virus (HIV) infection: a compar-ative analysis of quantification of HIV RNA, soluble turmor necrosis factor type IIreceptors, neopterin, and B2-microglobulin. Multicenter AIDS Cohort Study. J InfDis 1997;176:1161–1167.

11. Fuchs D, Moller AA, Reibnegger G, Stockie E, Werner ER, Wachter H. Decreasedserum tryptophan in patients with HIV-1 infection correlates with increased serumneopterin and with neurologic/psychiatric symptoms. J Acquir Immune Defic Syndr1990;3:873–876.

12. Bridges SL, Jenq G, Moran M, Kuffner T, Whitworth WC, McNicholl J. Single-nucleotide polymorphisms in tumor necrosis factor receptor genes: definition of novelhaplotypes and racial/ethnic differences. Arthritis Rheum 2002;46:2045–2050.

13. Khera A, McGuire DK, Murphy SA, et al. Race and gender differences in C-reactiveprotein levels. J Am Coll Cardiol 2005;46:464–469.

14. Cox DE, Hoffman SC, DiMercurio BS, et al. Cytokine polymorphic analyses indicateethnic differences in the allelic distribution of interleukin-2 and interleukin-6.Transplantation 2001;72:720–726.

15. Bartsch T. The hippocampus in neurological disease. In: Bartsch T, editor. TheClinical Neurobiology of the Hippocampus: An Integrative View. Oxford, UnitedKingdom: Oxford University Press; 2012:200–205.

16. Torres-Munoz J, Stockton P, Tacoronte N, Roberts B, Maronpot RR, Petito CK.Detection of HIV-1 gene sequences in hippocampal neurons isolated from post-mortem AIDS brains by laser capture microdissection. J Neuropathol Exp Neurol2001;60:885–892.

17. Krueger KR, Adeyemi O, Leurgans S, et al. Association of cognitive activity andneurocognitive function in blacks and whites with HIV. AIDS 2017;31:437–441.

18. Malone IB, Leung KK, Clegg S, et al. Accurate automatic estimation of total intracranialvolume: a nuisance variable with less nuisance. Neuroimage 2015;104:366–372.

19. Werner-Felmayer G, Werner ER, Fuchs D. Hausen A, Reibnegger G, Wachter H.Neopterin formation and tryptophan degradation by a human myelomonocytic cellline (THP-1). Cancer Res 1990;50:2863–2867.

20. Sucher R, Schroecksnadel K, Weiss G, Margreiter R, Fuchs D, Brandacher G. Neo-pterin, a prognostic marker in human malignancies. Cancer Lett 2010;287:13–22.

21. Sonnerborg AB, von Stedingk LV, Hansson LO, Strannegard OO. Elevated neopterinand beta 2-microglobulin levels in blood and cerebrospinal fluid occur early in HVI-1infection. AIDS 1989;3:277–283.

22. Nikonenko AG, Radenovic L, Andjus PR, Skibo GG. Structural features of ischemicdamage in the hippocampus. Anat Rec (Hoboken) 2009;292:1914–1921.

23. Schmidt-Kastner R, Freund TF. Selective vulnerability of the hippocampus in brainischemia. Neuroscience 1991;40:599–636.

24. Auer RN, Wieloch T, Olsson Y, Siesjo BK. The distribution of hypoglycemic braindamage. Acta Neuropathol 1984;64:177–191.

25. Kril J, Patel S, Harding A, Halliday G. Patients with vascular dementia due to mi-crovascular pathology have significant hippocampal neuronal loss. J Neurol Neuro-surg Psychiatry 2002;72:747–751.

26. Zarow C, Vinters HV, Ellis WG, et al. Correlates of hippocampal neuron number inAlzheimer’s disease and ischemic vascular dementia. Ann Neurol 2005;57:896–903.

27. Fujimura RK, Goodkin K, Petito CK, et al. HIV-1 proviral DNA load across neuro-anatomic regions of individuals with evidence for HIV-1 associated dementia. J AcquirImmune Defic Syndr Hum Retrovirol 1997;16:146–152.

28. Kallianpur KJ, Shikuma C, Kirk GR, et al. Peripheral blood HIV DNA is associatedwith atrophy of cerebellar and subcortical gray matter. Neurology 2013;80:1792–1799.

29. Bartsch T, Schonfeld R, Muller FJ, et al. Focal lesions of human hippocampal CA1neurons in transient global amnesia impair placememory. Science 2010;328:1412–1415.

30. FleischmanDA,Wilson RS, Gabrieli JD, Schneider JA, Bienias JL, Bennett DA. Implicitmemory and Alzheimer’s disease neuropathology. Brain 2005;128:2006–2015.

31. Eldridge LL, Engel SA, Zeineh MM, Bookheimer SY, Knowlton BJ. A dissociation ofencoding and retrieval processes in the human hippocampus. J Neurosci 2005;25:3280–3286.

32. Faraco CC, Linsworth N, Langley J, et al. Complex span tasks and hippocampalrecruitment during working memory. Neuroimage 2011;55:773–787.

33. Wilson TW, Proskovec AL, Heinrichs-Graham E, et al. Aberrant neuronal dynamicsduring working memory operations in the aging HIV-infected brain. Sci Rep 2017;7:41568.

34. Hodgetts CJ, Voets NL, Thomas AG, Clare S, Lawrence AD, Graham KS. Ultra-high-field fMRI reveals a role for the subiculum in scene perceptual discrimination.J Neurosci 2017;37:3150–3159.

35. Stone TW, Darlington L. Pharmacology of the kynurenine pathway. Int Congress Ser2007;1304:298–304.

36. Farrarese C, Aliprandi A, Tremolizzo L, et al. Increased glutatmate in CSF and plasmaof patients with HIV dementia. Neurology 2001;57:671–675.

37. Wang X, Ranu P, Chen X-W, Limpeanchob N, Kuman KN, Michaelis EK. Highintrinsic oxidative stress may underlie selective vulnerability of the hippocampal CA1region. Mol Brain Res 2005;140:120–126.

38. Pawlek GE, Bien B, Pawlek D. Kynurenine and its metabolites in Alzheimer’s diseasepatients. Adv Med Sci 2010;55:204–211.

39. Oxenkrug GF. Interferon-gamma-inducible kynurenines/pteridines inflammationcascade: implication for aging and aging-associated psychiatric medical disorders.J Neural Transm 2011;118:75–85.

40. Hagberg L, Cinque P, Gisslen M, et al. Cerebrospinal fluid neopterin: an informativebiomarker of central nervous system immune activation in HIV-1 infection. AIDS ResTher 2010;7:15.




ARTICLE OPEN ACCESS CLASS OF EVIDENCE

Expanded autologous regulatory T-lymphocyteinfusions in ALSA phase I, first-in-human study

Jason R. Thonhoff, MD, PhD,* David R. Beers, PhD,* Weihua Zhao, MD, PhD, Milvia Pleitez, MD,

Ericka P. Simpson, MD, James D. Berry, MD, Merit E. Cudkowicz, MD, and Stanley H. Appel, MD


Correspondence

Dr. Appel

[email protected]

AbstractObjectiveTo determine whether autologous infusions of expanded regulatory T lymphoctyes (Tregs)into patients with amyotrophic lateral sclerosis (ALS) are safe and tolerable during early andlater stages of disease.

MethodsThree patients with ALS, with no family history of ALS, were selected based on their differingsites of disease onset and rates of progression. Patients underwent leukapheresis, and Tregswere subsequently isolated and expanded ex vivo. Tregs (1 × 106 cells/kg) were administeredIV at early stages (4 doses over 2 months) and later stages (4 doses over 4 months) of disease.Concomitant interleukin-2 (2 × 105 IU/m2/injection) was administered subcutaneously 3times weekly over the entire study period. Patients were closely monitored for adverse effectsand changes in disease progression rates. Treg numbers and suppressive function were assayedduring and following each round of Treg infusions.

ResultsInfusions of Tregs were safe and well tolerated in all patients. Treg numbers and suppressivefunction increased after each infusion. The infusions slowed progression rates during early andlater stages of disease. Spearman correlation analyses showed that increased Treg suppressivefunction correlated with slowing of disease progression per the Appel ALS scale for eachpatient: patient 1: ρ (rho) = −0.60, p = 0.003; patient 2: ρ = −0.71, p = 0.0026; and patient 3: ρ =−0.54, p = 0.016. Measures of maximal inspiratory pressure also stabilized, particularly in 2patients, during Treg infusions.

ConclusionsThese results demonstrate the safety and potential benefit of expanded autologous Treginfusions, warranting further clinical trials in patients with ALS. The correlation between Tregsuppressive function and disease progression underscores the significance of using Treg sup-pressive function as an indicator of clinical status.

Classification of evidenceThis study provides Class IV evidence. This is a phase I trial with no controls.

MORE ONLINE

Class of EvidenceCriteria for ratingtherapeutic and diagnosticstudies

NPub.org/coe

*These authors contributed equally to the manuscript.

From the Houston Methodist Neurological Institute (J.R.T., D.R.B., W.Z., M.P., E.P.S., S.H.A.), Houston Methodist Hospital Research Institute, Stanley H. Appel Department ofNeurology, Houston, TX; and Neurological Clinical Research Institute (J.D.B., M.E.C.), Massachusetts General Hospital, Boston, MA.


The Article Processing Charge was funded by Department of Neurology, Houston Methodist Hospital.





http://www.NPub.org/coe



CD4+CD25+FOXP3+ regulatory T lymphocytes (Tregs) area subpopulation of T lymphocytes that are immunosuppres-sive and maintain tolerance to self-antigens, with their dys-function playing a pivotal role in the development ofautoimmune disorders.1–4 In amyotrophic lateral sclerosis(ALS) mice, infusions of Tregs slow disease progression andprolong survival, and Tregs suppress the proliferation of re-sponder T lymphocytes and the activation of microglia.5,6 Inpatients with ALS, the expression of the Treg master tran-scription factor FOXP3 is reduced in rapidly progressingpatients,7 with subsequent impairment of Treg suppressivefunctions; FOXP3 expression and Treg suppressive functionscorrelate with the extent and rapidity of disease progression.8

When expanded ex vivo in the presence of interleukin (IL)-2and rapamycin, Treg suppressive function is restored.8,9

These data suggest that infusions of expanded autologousTregs will improve Treg suppressive function in vivo andslow rates of disease progression in patients with ALS. Totest these hypotheses, a first-in-human phase 1 study wasinitiated to determine whether infusions of expanded autol-ogous Tregs into patients with ALS were safe and tolerableduring early and later stages of disease. IL-2 was administeredconcomitantly in the patients in an effort to stabilize andpossibly enhance the suppressive functions of the infusedTregs. Furthermore, the relationship between ex vivo Tregsuppressive functions and patients’ clinical statuses wasexplored.

MethodsPrimary research questionAre infusions of expanded autologous Tregs safe and tolerablein patients with ALS during both early and later stages ofdisease? This study is a phase 1 trial with no controls andprovides Class IV evidence that infusions of expanded au-tologous Tregs are safe and tolerable during early and laterstages of disease.

Standard protocol approvals, registrations,and patient consentsApproval from the Food and Drug Administration and In-stitutional Review Board at Houston Methodist Hospital wasobtained before study initiation. Written informed consentwas obtained before enrollment. The study was registered onclinicaltrials.gov (NCT03241784).

Study design and patient selection criteriaThis study was conducted at the Houston MethodistNeurological Institute. Patients with no family history of

ALS, and with differing sites of symptom onset and rates ofdisease progression, were recruited from Houston Meth-odist Hospital’s Muscular Dystrophy Association (MDA)/Amyotrophic Lateral Sclerosis Association (ALSA) ALSclinic. Three patients with arm, bulbar, and leg-onset ALS,respectively, were enrolled in the trial (table). Patients wererecruited, treated, and followed up between January 2016and February 2018.

The patients underwent a total of 8 infusions of expandedautologous Tregs with concomitant subcutaneous IL-2injections. Four Treg infusions were administered every 2weeks at an early stage of the disease, followed by 4 Treginfusions administered every 4 weeks at a later stage. EachTreg dose was 1 × 106 cells/kg. The Treg dose was empiricallydetermined but was selected within the range of what hasbeen shown to be safe and tolerable in patients with type 1diabetes.4 IL-2 was administered subcutaneously 3 timesweekly at a dose of 2 × 105 IU/m2/injection beginning the dayafter the first Treg infusion and continued throughout thestudy period.

Infusions of expanded autologous TregsLeukapheresis was performed 1 month before the first Treginfusion. Tregs were isolated and expanded ex vivo in theGood Manufacturing Practice–compliant facility at M.D.Anderson Cancer Center according to a previously describedprotocol.8,9 Each Treg infusion was administered IV througha peripheral line, and patients were closely monitored for anyinfusion-related adverse responses for 4 hours after theinfusion.

Clinical evaluationsThe revised ALS Functional Rating Scale (ALSFRS-R), AppelALS Rating Scale (AALS),10 andmaximal inspiratory pressure(MIP) measurements were performed immediately beforeeach Treg infusion, every 2 weeks during each round ofinfusions, and monthly after each round. Forced vital capacity(FVC) was monitored at each evaluation as a component ofthe AALS. Patients were asked about adverse events at eachencounter.

Assessing Treg percentage and suppressivefunction in the peripheral bloodPeripheral blood was drawn 1 month before the first Treginfusion, immediately before each infusion, the day aftereach infusion, every 2 weeks during each round of infusions,and monthly after each round. The percentage of CD4+CD25+

FOXP3+ Tregs within the total CD4+ population was assessedby flow cytometry.8 Treg suppressive function on the

GlossaryAALS = Appel Amyotrophic Lateral Sclerosis Rating Scale; ALS = amyotrophic lateral sclerosis; ALSA = Amyotrophic LateralSclerosis Association;ALSFRS-R = Revised Amyotrophic Lateral Sclerosis Functional Rating Scale; FVC = forced vital capacity;IL = interleukin; MDA = Muscular Dystrophy Association; MIP = maximal inspiratory pressure.


http://clinicaltrials.gov


proliferation of autologous responder T lymphocytes wasassessed by [3H]-thymidine incorporation.8

Statistical analysisCorrelation between changes in the AALS and Treg sup-pressive function was determined by Spearman correlationusing GraphPad Prism 7 software and depicted by Spearmanrho (ρ). Two-tailed p values <0.05 were considered statisti-cally significant. Data collected between the first and fifth Treginfusions were compared with values from the day of the firstinfusion. Data collected after the fifth Treg infusion werecompared with the values from the day of the fifth infusion.

Data availabilityIndividual deidentified patient data not published within thearticle including clinical evaluations and Treg percentage andsuppressive function results will be shared by request fromany qualified investigator.

ResultsSafetyNo infusion-related adverse events or clinically significantchanges in safety laboratories or electrocardiogram findingswere observed. All patients noted dramatic increases in thefrequency, intensity, and distribution of fasciculations duringeach round of infusions. Fasciculations were noted withina few minutes to a few days after each Treg infusion andlasting from days to more than 1 month after the completionof each round of infusions. Patient 1 experienced increasedmuscle cramps in his legs from weeks 2–6, 2 falls on weeks 17

and 23, and an episode of pharyngitis on week 10. Patient 2underwent placement of a percutaneous endoscopic gas-trostomy tube on week 9. He developed aspiration pneu-monia on week 19, and his IL-2 injections were temporarilysuspended until week 23. His progressive dysphagia andepisode of aspiration pneumonia were likely due to hisbulbar ALS. Patient 2 dropped out of the study on week 50because of his progressive disease and was placed in hospicecare. He died on week 51 because of respiratory failuresecondary to ALS. Patient 3 developed 2 suspected gastro-intestinal infections and an upper respiratory infection be-tween weeks 24 and 29. She reported mild dyspnea onexertion beginning on week 48.

Treg percentage and suppressive functionincreased during infusionsIn all patients, Treg percentage (figure 1, A–C) and sup-pressive function (figure 1, D–F) increased during the firstround of infusions, declined between each round of infusions,and increased again during the second round.

EnhancedTreg suppressive function correlatedwith slowing of functional declineIn all patients, the rate of decline of the ALSFRS-R and AALSslowed for 2 months during the first round of infusions, ac-celerated between each round of infusions, and slowed againover 4 months during the second round (figure 2, A–C).Spearman correlation showed that increased Treg suppressivefunction correlated with slowing of disease progression perthe AALS for each patient (figure 2, D–F; ρ = −0.60, p = 0.003in patient 1; ρ = −0.71, p = 0.0026 in patient 2; and ρ = −0.54,p = 0.016 in patient 3). The larger the increase in Treg

Table Patient characteristics

Patient No. 1 2 3

Age (yr) 47 46 56

Sex Male Male Female

Initial weight (kg) 92 77 79

Site of symptom onset Arm Bulbar Leg

Time from symptom onset to diagnosis (mo) 7 18 12

Time from symptom onset to the 1st Treg infusion (mo) 14 24 38

Riluzole use at study entry Yes Yes Yes

Noninvasive ventilation use at study entry No Yes No

ALSFRS-R just before the 1st Treg infusion 44 36 41

AALS just before the 1st Treg infusion 50 65 68

FVC (L) just before the 1st Treg infusion 5.25 2.84 2.24

FVC (% predicted) just before the 1st Treg infusion 92 56 77

MIP (cm H2O) just before the 1st Treg infusion 120 50 100

Abbreviations: AALS = Appel Amyotrophic Lateral Sclerosis Rating Scale; ALSFRS-R = revised Amyotrophic Lateral Sclerosis Functional Rating Scale; FVC =forced vital capacity; MIP = maximal inspiratory pressure.



suppressive function, the smaller the decline in the AALS atthe next clinical evaluation.

MIPs stabilized during infusionsIn all patients, the FVC remained relatively unchanged duringthe Treg infusions and between each round (figure 3, A–C).In patients 1 and 3, the MIPs were stable during the firstround of infusions and deteriorated between each round. TheMIPs again stabilized during the second round. Patient 2 wastreated with noninvasive ventilation before enrollment andcontinued treatment throughout the study. In patient 2, MIPswere low, but remained relatively stable during each round ofinfusions and between each round (figure 3, D–F).

DiscussionThree patients with ALS were infused with autologous ex-panded Tregs with concomitant subcutaneous IL-2 injec-tions at early and later stages of disease. Tregs were alsoinfused at a later stage of disease to determine whether theinfusions remained safe and the beneficial effects on disease

progression could be extended by increasing the dosing in-terval. Treg infusions were safe and well tolerated regardlessof the burden of disease. Treg suppressive function corre-lated with changes in the AALS; the greater the improve-ment in Treg suppressive function, the slower the rate ofclinical progression. This correlation supports the valueof Treg suppressive function as a meaningful indicator ofclinical status. In addition, Treg infusions did not adverselyaffect respiratory function and appeared to stabilize the de-cline in MIPs in the 2 patients who were not being treatedwith noninvasive ventilation.

In a previous pilot study, low-dose IL-2 administered sub-cutaneously for 1 year in 5 patients with ALS was safe andtolerable, but did not appear to alter the clinical course orincrease endogenous Treg numbers, likely related to impairedendogenous Treg responsiveness to IL-2 (unpublishedresults). In the present study, subcutaneous injections of low-dose IL-2 were administered to stabilize the infused expandedTregs. During infusions in all patients, several data pointswere observed in which IL-2 could have enhanced the pro-liferation and function of infused Tregs. However, IL-2 was

Figure 1 Treg percentage and suppressive function increased during each round of Treg infusions

Arrows and vertical dotted lines represent Treg infusions. The 1st Treg infusion was administered on week 0 and then every 2 weeks for a total of 4infusions. The 5th Treg infusion was administered in each patient on weeks 48, 27, and 33, respectively, and then every 4 weeks for a total of 4 infusions.The percentage of CD4+CD25+FOXP3+ Tregs within the total CD4+ cell population is shown for patient 1 (A), patient 2 (B), and patient 3 (C). Tregpercentages are shown at baseline (weeks 4.6, 3.0, and 4.9 in each patient, respectively), the days of the 1st and 5th Treg infusions, the day after each Treginfusion, every 2 weeks during each round of infusions, and 1month after each round. The data point collected the day after the 4th Treg infusion (week 6)in patient 3 was not determined because of a flow staining error. Treg suppressive function is shown on the same time points as the Treg percentages forpatient 1 (D), patient 2 (E), and patient 3 (F).



not of critical value in the interim between each round whenTreg percentage and suppressive function and clinical statusdeteriorated.

Common to all patients was the perceived increase in fas-ciculations. The rapid onset of fasciculations suggests a pe-ripheral action in the lower motor neuron possibly mediated by

Figure 2 Disease progression slowed during each round of Treg infusions and correlated with increased Treg suppressivefunction

Arrows and vertical dotted lines represent Treg infusions. Clinical progression is depicted by the ALSFRS-R (white points) and AALS (black points) stages of thedisease for patient 1 (A.a), patient 2 (B.a), and patient 3 (C.a). Clinical progression lines during each round of Treg infusions are enlarged in side panels for theearly (1) and later (2) stages of disease. Correlation between changes in the AALS and Treg suppressive function is shown for patient 1 (A.b), patient 2 (B.b), andpatient 3 (C.b). Lines represent the best fit as determined by linear regression analysis. Data were analyzed by Spearman correlation, and p values <0.05 wereconsidered statistically significant. AALS = Appel Amyotrophic Lateral Sclerosis Rating Scale; ALSFRS-R = revised Amyotrophic Lateral Sclerosis FunctionalRating Scale.



immune modulation of ectopic axonal firing. In the earlier IL-2–alone pilot study, increased fasciculations were not observed,suggesting that the infused Tregs directly or indirectly causedthe fasciculations in this study. In addition, common to allpatients was the occurrence of infections during the study:pharyngitis in patient 1, aspiration pneumonia in patient 2, andgastrointestinal and upper respiratory infections in patient 3. Apotential increased risk of infections with Treg and IL-2treatment is concerning and requires further study in a largernumber of patients with ALS.

Although this study lacked blinding and placebo controls,slowing of disease progression was observed during the initialinfusions at an early stage of disease and the subsequent 4monthly infusions at a later stage of disease. Administering theALSFRS-R to the patients during the initial infusions every 2weeks enhances the likelihood of a placebo effect. However,we have not observed stabilization of the AALS in previousstudies of subcutaneous IL-2 injections (unpublished) orinfusions of allogeneic hematopoietic stem cells.11 Further-more, the increased Treg suppressive function, which corre-lated with the clinical state, and the observed stabilization ofMIPs were not likely to have been influenced by a placeboeffect. Increased clinical progression rates were observed

between each round of infusions, but it was not clear whetherthe progression was related to the cessation of Treg infusionsor would have occurred spontaneously. More pertinent is theobservation that subsequent Treg infusions were beneficial atlater stages of disease with increased disease burden and rateof progression. Circulating functional Tregs may slow diseaseprogression by suppressing peripheral proinflammatorymonocytes/macrophages and responder T lymphocytes, aswell as entering the CNS and suppressing activated microglia.Defining peripheral and central actions of Tregs merits furtherinvestigation.

The results from this study support the need for a phase 2,randomized, placebo-controlled trial over a longer period totest the clinical efficacy, safety, and tolerability of differentdoses of Tregs in a larger number of patients with ALS. Thegoal of future studies is to determine whether optimized dosesof Tregs infused at regular intervals prolong slowed pro-gression and minimize the more rapid progression associatedwith cessation of Treg infusions.

Author contributionsJason R. Thonhoff: study concept and design; acquisition,analysis and interpretation of data; and drafting of the

Figure 3 Maximal inspiratory pressures stabilized during Treg infusions

Arrows and vertical dotted lines represent Treg infusions. FVCmeasurements are represented as%predicted values for patient 1 (A), patient 2 (B), and patient3 (C). Measurements are shown at baseline (weeks 4.6, 3.0, and 4.9 in each patient, respectively), immediately before each Treg infusion, every 2weeks duringeach round of infusions, and 1 month after each round. MIP measurements are shown in cm H2O for patient 1 (D), patient 2 (E), and patient 3 (F). MIPs areshown at the same time points as FVC measurements. The MIP values were erroneously not determined for patient 1 immediately before the 5th Treginfusion, and for patient 3, 1 month after the second round of infusions. The solid gray line connects the points between each round of infusions (A-F). FVC =forced vital capacity; MIP = maximal inspiratory pressure.



manuscript. David R. Beers: study concept and design; ac-quisition, analysis and interpretation of data; and critical re-vision of the manuscript. Weihua Zhao: acquisition, analysisand interpretation of data and critical revision of the manu-script. Milvia Pleitez: study concept and design and acquisi-tion, analysis, and interpretation of data. Ericka P. Simpson,James D. Berry, and Merit E. Cudkowicz: study concept anddesign and critical revision of the manuscript. Stanley H.Appel: study concept and design, study supervision, acquisi-tion of data, analysis and interpretation of data, and criticalrevision of the manuscript.

AcknowledgmentThe authors dedicate this article to their friend and colleagueJenny S. Henkel, who made considerable contributions to thedevelopment of this project through her research. They aregrateful to Dan Neal of the University of Florida for hisassistance in the statistical analyses. They also appreciate theassistance of Luis Lay, Della Brown, Shixiang Wen, DouglasCasey, JinghongWang, Joanna Espinosa, and Yuanfang Liu ofthe Houston Methodist Neurological Institute and AnnetteDeMattos and Liz Simpson of the Neurological ClinicalResearch Institute, Massachusetts General Hospital. They arealso grateful to the Stem Cell Transplantation and CellularTherapy facility at M.D. Anderson Cancer Center formanufacturing the Tregs.

Study fundingALS Finding a Cure® and the ALSA.

DisclosureJ.R. Thonhoff is supported by the HoustonMethodist ClinicianScientist Recruitment and Retention Program. D.R. Beers,W. Zhao, and M. Pleitez report no disclosures. E.P. Simpsonreceives publishing royalties from McGraw-Hill and wasa speaker for Alexion Pharmaceuticals. J.D. Berry served on thescientific advisory board of MT Pharma, Denali Therapeutics,and Orion Corporation; received travel funding from ALSA,MDA, and ALS ONE; is a consultant to Denali Therapeutics;has received fellowship support fromVoyager Therapeutics andresearch funding from Cytokinetics, Neuraltus, Amylyx Phar-maceuticals, Brainstorm Cell Therapeutics, NIH/NINDS,

MDA, ALS Finding a Cure, ALSA, and ALS ONE; holdsa patent formicroRNAs in neurodegenerative disorders; and hisspouse is on the editorial board ofNeuropsychopharmacology. M.E. Cudkowicz serves on the scientific advisory board of Cyto-kinetics, SAB, Lilly and DSMB; is on the editorial board ofNeurotherapeutics and JAMA Neurology; holds a patent formetabolomics in ALS; receives publishing royalties fromUpToDate; serves as a consultant to Biogen Idec, ImmunityPharma, Biogen, AveXis, Revalesio, Mitsubishi Tanabe Pharma,Avanir, MDA, Cytokinetics, Lilly, Karyopharm, and Orion; andreceived research support from the NINDS, ALSA, and ALSFinding a Cure Foundation. S.H. Appel serves as a scientificconsultant to Mitsubishi Tanabe Pharma, Neuraltus, and UCBBiopharma; received speaker honoraria from and served on thespeaker’s bureau of Avanir; received research support from theALSA, ALS Finding a Cure, and Lee Rizzuto Foundation; andserved as an expert consultant in ALS case. Full disclosure forminformation provided by the authors is available with the full textof this article at Neurology.org/NN.

Received February 6, 2018. Accepted in final form April 11, 2018.

References1. Sakaguchi S. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells in

immunological tolerance to self and non-self. Nat Immunol 2005;6:345–352.2. Viglietta V, Baecher-Allan C, Weiner HL, Hafler DA. Loss of functional suppression

by CD4+CD25+ regulatory T cells in patients with multiple sclerosis. J Exp Med2004;199:971–979.

3. Dejaco C, Duftner C, Grubeck-Loebenstein B, Schirmer M. Imbalance of regulatoryT cells in human autoimmune diseases. Immunology 2006;117:289–300.

4. Bluestone JA, Buckner JH, Fitch M, et al. Type 1 diabetes immunotherapy usingpolyclonal regulatory T cells. Sci Transl Med 2015;7:315ra189.

5. Beers D, Henkel JS, Zhao W, et al. Endogenous regulatory T lymphocytes ameliorateamyotrophic lateral sclerosis in mice and correlate with disease progression in subjectswith amyotrophic lateral sclerosis. Brain 2011;134:1293–1314.

6. Zhao W, Beers DR, Liao B, Henkel JS, Appel SH. Regulatory T lymphocytes fromALS mice suppress microglia and effector T lymphocytes through different cytokine-mediated mechanisms. Neurobiol Dis 2012;48:418–428.

7. Henkel JS, Beers DR, Wen S, et al. Regulatory T-lymphocytes mediate amyotrophiclateral sclerosis progression and survival. EMBO Mol Med 2013;5:64–79.

8. Beers DR, Zhao W, Wang J, et al. ALS patients’ regulatory T lymphocytes are dys-functional, and correlate with disease progression rate and severity. J Clin InvestInsight 2017;2:e89530.

9. Alsuliman A, Appel SH, Beers DR, et al. A robust, good manufacturing practice-compliant, clinical-scale procedure to generate regulatory T cells from patients withamyotrophic lateral sclerosis for adoptive cell therapy. Cytotherapy 2016;18:1312–1324.

10. Haverkamp LJ, Appel V, Appel SH. Natural history of amyotrophic lateral sclerosis ina database population: validation of a scoring system and a model for survival pre-diction. Brain 1995;118:707–719.

11. Appel SH, Engelhardt JI, Henkel JS, et al. Hematopoietic stem cell transplantation inpatients with sporadic amyotrophic lateral sclerosis. Neurology 2008;71:1326–1334.




ARTICLE OPEN ACCESS

Molecular signature of Epstein-Barr virusinfection in MS brain lesionsMonica A. Moreno, PhD, Noga Or-Geva, PhD, Blake T. Aftab, PhD, Rajiv Khanna, PhD, Ed Croze, PhD,

Lawrence Steinman, MD, and May H. Han, MD


Correspondence

Dr. Han

[email protected]

AbstractObjectiveWe sought to confirm the presence and frequency of B cells and Epstein-Barr virus(EBV) (latent and lytic phase) antigens in archived MS and non-MS brain tissue byimmunohistochemistry.

MethodsWe quantified the type and location of B-cell subsets within active and chronic MS brain lesionsin relation to viral antigen expression. The presence of EBV-infected cells was further confirmedby in situ hybridization to detect the EBV RNA transcript, EBV-encoded RNA-1 (EBER-1).

ResultsWe report the presence of EBV latent membrane protein 1 (LMP-1) in 93% of MS and 78% ofcontrol brains, with a greater percentage of MS brains containing CD138+ plasma cells andLMP-1–rich populations. Notably, 78% of chronic MS lesions and 33.3% of non-MS brainscontained parenchymal CD138+ plasma cells. EBV early lytic protein, EBV immediate-earlylytic gene (BZLF1), was also observed in 46% of MS, primarily in association with chroniclesions and 44% of non-MS brain tissue. Furthermore, 85% of MS brains revealed frequentEBER-positive cells, whereas non-MS brains seldom contained EBER-positive cells. EBV in-fection was detectable, by immunohistochemistry and by in situ hybridization, in both MS andnon-MS brains, although latent virus was more prevalent in MS brains, while lytic virus wasrestricted to chronic MS lesions.

ConclusionsTogether, our observations suggest an uncharacterized link between the EBV virus life cycle andMS pathogenesis.

From the Department of Neurology and Neurological Sciences (M.A.M., N.O., L.S., M.H.H.), Stanford University School of Medicine, Multiple Sclerosis Center; InterdepartmentalProgram in Immunology (M.A.M., N.O., L.S., M.H.H.), Stanford; Atara Biotherapeutics (B.T.A., E.C.), San Francisco, CA; and Queensland Institute of Medical Research (R.K.), Brisbane,Queensland, Australia.


The Article Processing Charge was funded by Atara Therapeutics.







MS is a chronic, autoimmune-mediated disorder of the CNS,associated with neurodegeneration and progressive neuro-logic disability.1,2 Some studies report Epstein-Barr virus(EBV) in MS,3–7 whereas others find no association.8,9

Contradictory results have been attributed to differences inmethodology and tissue preservation. To date, unequivocaldemonstration of EBV infection in MS lesions is lacking.

EBV is a double-stranded DNA ɣ-herpes virus that entersB cells via the tonsillar lymphoid tissue.6 Within B cells, EBVswitches between latent and lytic forms via modulation ofEBV nuclear genes, the latent membrane proteins (LMPs) 1,2A, and 2B,10 and the viral immediate-early protein, BZLF1(ZEBRA).11,12 The risk of developing MS increases after EBVinfection,13,14 and EBV infection has been shown to be moreprevalent in patients with MS.15,16

Previous reports show a high frequency of CNS infiltratingB cells positive for EBV RNA transcript, EBV-encoded RNA(EBER), by in situ hybridization and for the EBV LMP-1protein.3 However, subsequent studies, using identical MSbrain tissue, failed to replicate these findings.17,18 Here, wedemonstrate EBV infection in both MS and control brains,using archived MS and healthy brain samples. We reporthigher numbers of parenchymal plasma cells and LMP-1+

cells in chronic plaques (CPs) and chronic active plaques(CAPs) compared with controls. We observed BZLF1 pro-tein expression in both MS and non-MS brains. However, itsexpression was restricted to CPs in MS brains.

MethodsHuman brain samplesWe included archived, autopsy brain samples and selectedpreserved biopsy brain specimens from a panel of 17 MS(mean age 59.3 years [range 29–98 years]; 7 men, 9 women,and 1 without sex specification) and 9 control brains (withoutneurologic disease) (mean age 72.8 years [range 49–89years]; 4 men and 5 women) in this study. We did not observeany differences in staining patterns between biopsy andpostmortem samples. All samples were processed and sub-jected to an identical staining protocol. Demographics in-cluding age and sex are shown in table. Samples were obtainedfrom the Stanford Pathology Department (NeuropathologyDivision), the Maritime Brain Tissue Bank, Dalhousie Uni-versity, and the University ofWashington Alzheimer’s DiseaseResearch Center, the Adult Changes in Thought Study, andMorris K Udall Center of Excellence for Parkinson’s Disease

Research. ForMS tissue, areas analyzed included the temporallobe cortex, occipital lobe cortex, frontal lobe cortex, temporallobe cortex, and parietal lobe. Human MS brain tissue re-search was performed according to Institutional ReviewBoard guidelines approved by Stanford Human SubjectsResearch.

Standard protocol approvals, registrations,and patient consentsArchived, deidentified autopsy and biopsy brain samples wereobtained and used according to the institutional guidelines.

ImmunohistochemistryImmunohistochemistry was performed on formalin-fixed,paraffin-embedded (FFPE) tissue in 4-μm-thick tissue sec-tions. Briefly, slides were deparaffinized in xylene (3 × 1minute), xylene: 100% ethanol (at a ratio of 1:1 for 1 minute),100% ethanol (2 × 1 minute), 95% ethanol (1 minute), 70%ethanol (1 minute), and 50% ethanol (1 minute). Slides werethen rinsed in cold water and washed in phosphate-bufferedsaline (PBS). For characterization of cellular infiltration andinflammation, tissue was stained with hematoxylin and eosin(H&E). For heat-induced antigen retrieval, slides were im-mersed in 10 nM sodium citrate buffer (pH 6) and heated ina microwave to 98°C for 20 minutes. Slides were thenquenched in 0.3% hydrogen peroxide (H2O2) for 20 minutesto quench endogenous peroxidase activity and rinsed in PBS.To block nonspecific binding of antigens to the tissue, slideswere immersed in 2% normal horse serum (NHS) for 10minutes at room temperature (RT). Primary antibody wasdiluted in NHS and incubated overnight at 4°C. The next day,slides were washed in PBS and incubated with secondaryantibody (biotinylated, affinity-purified anti-immunoglobulin;1:125 dilution in NHS) for 30 minutes at RT. Samples werenext washed in PBS and incubated with Elite ABC for 45minutes at RT. After this incubation, slides were washed onceagain and stained with freshly made 3, 39diaminobenzidine for1 minute. Samples were then counterstained with hematox-ylin for 1 minute, rinsed in water, and then coverslipped withPermount. Immunohistochemistry was performed on thesame sections using antibodies against myelin basic protein(MBP) (abcam, ab7349, clone 12; 1:100), CD3 (Dako,M7254, clone F7.2.38; 1:25), CD68 (Dako, M0876, clonePG-M1; dilution 1:50), CD20 (abcam, Ab9475, clone L26;dilution 1:25), CD138 (Sigma, 138M-14, clone B-A38; 1:50),LMP-1 (SC-71023, clone 3H2104,ab,c; dilution 1:100),LMP-1 (SC-57721, clone CS1/2/3/4; 1:200), and EBVZEBRA (BZLF1) (SC-53904, dilution 1:200). Positive controlsincluded tonsils with infectious mononucleosis and diffuse

GlossaryCAP = chronic active plaque; CP = chronic plaque; DLBCL = diffuse large B-cell lymphoma; EBER = EBV-encoded RNA;EBNA = Epstein-Barr nuclear antigen; EBV = Epstein-Barr virus; FFPE = formalin-fixed, paraffin-embedded; LMP = latencymembrane protein; MBP = myelin basic protein; NHS = normal horse serum; PBS = phosphate-buffered saline; PV =perivascular space; RT = room temperature.



large B-cell lymphoma. Negative controls included normaltonsils.

In situ hybridizationIn situ hybridization (ISH) was performed on the same FFPEblocks, which were used for immunohistochemistry. TheEBER-1 dinitrophenyl (DNP) probe was used to detect theexpression of EBER-1 system (ISH iView kit; VentanaMedical Systems Inc., Tucson, AZ, Cat# 760-097). EBER-ISHwas performed using an automated Ventana BenchMark XT

system (Ventana Inc.) in accordance with the manufacturer’sprotocol. Briefly, FFPE sections were treated with EZ Prepbuffer (Ventana Inc.) to remove paraffin, rehydrated, and thendigested with ISH protease 1 (Ventana Inc., Cat# 780-4147).EBER 1 DNP probe was then administered and allowed tohybridize, followed by stringency washes as per the manu-facturer’s instructions using SSC buffer (Ventana Inc., Cat #950-110). Slides were counterstained with Red CounterstainII (Ventana Inc., Cat# 780-2218). Serial sections of all sam-ples were also stained with oligo-T probes to ascertain RNA

Table Characteristics of MS and healthy control brain samples

Patient Age/sex MS type Lesion type Region CD138 LMP-1 BZLF1PV/P(CD138+ cells)

MS-1 Unknown/M Tumefactive MS CAP PL +++ +++ − PV

MS-2 Unknown/unknown Unknown CAP Unknown + ++ − P

MS-3 30/F Unknown CAP FLa + +++ − PV, P

MS-4 62/F Tumefactive MS CAP FLa +++ +++ − P

MS-5 42/F Unknown CAP PLa − ++ − PV

MS-6 39/M Tumefactive MS CAP TLCa +++ +++ − PV, P

MS-7 29/M Unknown CAP OLCa − − − NA

MS-8 66/F Unknown CP TLC +++ ++ + PV, P

MS-9 47/M SPMS CP OLC +++ +++ + PV, P

MS-10 50/M SPMS CP OLC +++ +++ ++ PV, P

MS-11 98/M Unknown CP OLC ++ +++ − PV, P

MS-12 45/F Unknown Unknown TLCa − − − NA

MS-13 88/F Unknown CP FL ++ +++ − PV, P

MS-14 60/F Unknown CP FL + + + PV, P

MS-15 84/M PPMS CP FL +++ +++ + PV, P

MS-16 86/F PPMS CP OLC +++ +++ ++ PV

MS-17 64/F PPMS CP OLC +++ +++ + PV

Cntl-1 66/F NA NA TLC + + ++ PV, P

Cntl-2 Unknown/unknown NA NA Unknown ++ +++ +++ PV, P

Cntl-3 49/M NA NA OLC + − − P

Cntl-4 84/F NA NA OLC +++ − − PV

Cntl-5 55/M NA NA OLC ++ ++ +++ PV

Cntl-6 63/F NA NA OLC − ++ − PV

Cntl-7 92/M NA NA OLC − + − PV

Cntl-8 89/F NA NA OLC ++ + + PV

Cntl-9 71/F NA NA FL +++ + − PV

Abbreviations: CAP = chronic active plaque; CP = chronic plaque; CNTL = control; PV = perivascular; P = parenchyma; NA = not applicable; OLC = occipital lobecortex; TLC = temporal lobe cortex; FL = frontal lobe; PL = parietal lobe; SPMS = secondary progressive MS; PPMS = primary progressive MS; LMP-1 = latentmembrane protein 1; BZLF1 = EBV immediate-early lytic gene; syndecan-1 (CD138) = a plasma cell marker.Treatment status of MS cases is unknown.Semiquantification is expressed as “−” (no cells/mm2), “+” (<5 cells/mm2), “++” (5–10 cells/mm2), and “+++” (>10 cells/mm2).a Biopsy samples. All other patients shown are autopsy samples.



preservation in each sample. EBV-infected tissue from a tonsilclassified as having infectious mononucleosis was used asa positive control for EBER-1. Tonsil tissue obtained froma healthy individual was used as a negative control.

Light microscopy andsemiquantitative analysisAll images were acquired using a Zeiss Axio Imager M2 mi-croscope (Carl Zeiss Microscopy), viewed using either 10- or63-fold magnification. For the purposes of blinding, the tissuesamples had only their autopsy or biopsy identifier number.Semiquantification of CD138+, LMP-1+, and BZLF1+ cellnumbers was performed by manually counting the number ofpositive cells having a clearly visible nucleus. The number ofcells in a given tissue section was determined by counting cellspresent in 3 × 3-cm tissue sections from autopsy samples (MSsamples, n = 11; controls samples, n = 9). Biopsy samples (MSsamples, n = 6; control samples, n = 0) were used to de-termine the number of positive cells present in 2 × 2-cm tissuesections. Results are expressed as follows: no cells/mm2 (−),<5 cells/mm2 (+), 5–10 cells/mm2 (++), and >10 cells/mm2

(+++). The data sets generated during and/or analyzedduring the current study are included in this published articleand are also available from the corresponding author onreasonable request.

ResultsCharacterization of MS and non-MSbrain tissueIn MS, there are multiple, focal areas of inflammation-drivendemyelination in the CNS called plaques or lesions.2 They areobserved as areas of loss of myelin, characterized by di-minished staining with basic protein (MBP) both within andin the surrounding region of the plaque. CAPs are charac-terized by “a demyelinated area with sharply defined marginsand recent areas of inflammatory demyelinating areas at theedges.” CAP often contains lymphocytes and macrophage/microglia within perivascular cuffs and the brain parenchyma.19,20

CPs contain “areas of demyelination with well-demarcated bor-ders and abundant astrogliosis, but few or no inflammatorycells.”2,21–23 Inflammation, though variable, is present in almostall types of MS plaques observed in the MS brain.

H&E staining was performed to confirm the presence of in-flammatory cell infiltration in active (figure 1A) and to a lesserdegree in chronic MS lesions (figure 1B). Inflammatoryinfiltrates, largely composed of lymphocytes (figure 1G) andmacrophages (figure 1, J, and K), were present to a variabledegree in all MS tissue samples studied. Staining with anti-MBP confirmed the loss of myelin, which is observed as a lossof brown staining in both active and chronic MS plaques(figure 1, D and E). Staining with CD3 and CD68 shows anincreased accumulation of CD3+ T lymphocytes and CD68+

macrophages/microglia in and around perivascular cuffs inMS lesions (both active and chronic) and control brains

(figure 1, G–L). However, the degree of inflammatory cellinfiltration decreases with disease duration as seen in CPs(figure 1, H and K).

Immunohistochemical detection of B cells andEBV latent and lytic proteinsTo determine the presence and frequency of EBV infection inMS (n = 17) and non-MS brains (n = 9), we performedimmunohistochemistry to determine the expression of thepan B-cell marker (CD20) and plasma cell marker (CD138),in addition to EBV markers LMP-1 and BZLF1 (figure 2).CD20+ B lymphocytes and CD138+ plasma cells were ob-served in all MS (figure 2, A–E) and most non-MS controlbrain samples (figure 2, C, F, and M), although in non-MStissue, these cells were often confined to the vasculature(figure 2, C and F). However, scarce numbers of CD138+

plasma cells (64.7% of MS/33.3% of control samples) weredetected outside the vasculature, most appearing to havemigrated into the parenchyma (figure 2D and table). We nextperformed immunohistochemistry to detect antibodies rec-ognizing EBV latent or lytic proteins. To detect EBV in itslatent form, we used LMP-1, which is an important latentlyexpressed viral protein encoded by EBV. LMP-1 functions, inpart, by ensuring efficient maturation of naive B cells intolong-lived memory B cells. Its expression coincides with EBVgrowth programs. LMP-1 staining was observed in both CAPs(figure 2G) and in CPs (figure 2H) and in control brain tissueobtained from healthy individuals (figure 2, I and N). To testfor the presence of lytic EBV infection, we measured thepresence of a well-documented EBV lytic protein, BZLF1. Weobserved BZLF1+ cells in both MS and control brain samples(table and figure 2, K and L).

The frequency of BZLF1+ staining was enriched in a CP (7 of9 samples; 78%) compared with control tissue (4 of 9 sam-ples; 44%). Notably, we observed no detectable BZLF1+ cellsin CAPs present in MS tissue (table and figure 2, J and O). Asa control, we used immunohistochemistry to detect CD20,CD138, LMP-1, and BZLF1 in tissue sections obtained froma tonsil classified as having infectious mononucleosis (positivecontrol) compared with a disease-free tonsil derived froma healthy individual (negative control). As expected, wedetected the presence of CD20+ B cells and CD138+ plasmacells in tissue derived from both the infectious mononucleosisand disease-free tonsil (figure 3, A–D). We detected robustexpression of EBV LMP-1 and diffuse BZLF1 staining indiseased tissue (figure 3, E and G), while detecting no LMP-1or BZLF1 in disease-free tonsil tissue (figure 3, F and H).

Detection of EBER transcripts in B cells andplasma cells infiltrating the MS brainEBERs (noncoding small RNAs) are expressed in all knownforms of EBV latency and serve as the “gold standard” fordetecting latent EBV infection.24,25 In situ hybridization forEBER-1 transcripts was performed on 7 MS and 4 non-MSbrain samples. Sporadic populations of EBER+ cells wereobserved in 6 of 7 MS brain samples (figure 4, B and D),



whereas only a single EBER+ cell was detected in 2 of 4control samples (figure 4F). RNA preservation was confirmedin all MS and non-MS tissue samples using oligo-T probes

(figure 4, A, C, and E). Robust EBER transcript expressionwas observed in a positive control containing a tonsil withinfectious mononucleosis (figure 4N). We performed

Figure 1 Histopathologic features of a chronic active and a chronic plaque in the MS brain

Representative hematoxylin & eosin (H&E) staining (A–C) and immunohistochemistry (D–L) of active MS (A, D, G, J), chronic MS (B, E, H K), and healthy control(C, F, I, L) brain samples. Perivascular (PV) inflammation (A and B), demyelination, as indicated by the loss ofMBP staining (black arrows) (D and E), presence ofinflammatory cells, CD3+ T lymphocytes (red arrowheads) were prominent within activeMS lesions (G) and present to a lesser extent in chronicMS lesions (H).Numerous macrophages/microglia (black arrowheads) were observed in and around PV cuffs in active MS (J) and were also observed chronic MS lesions,although to a lesser extent (K). Healthy controls, without neurologic disease, showed little or noCD3+ immunoreactivity (I) and positive CD68+ immunostainingin the parenchyma resembling resident microglia (black arrowheads) (L). Normal tonsils were used as a positive control and show CD3+ (M) and CD68+ (N)immunostaining. CAP = chronic active plaque, CP = chronic plaque. Scale bars = 50 μm.



immunohistochemistry for LMP-1 and in situ hybridizationfor EBER+ cells on serial sections of MS brain samples (figure4, H, I, K, and L). RNA preservation was confirmed in adja-cent sections (Figure 4, G and J). We confirmed colocalized(figure 4, H–I) or proximal (figure 4, K–L) LMP-1 immu-noreactivity and EBER+ signal (figure 4, H, I, K, and L).

Quantification of B cells and EBV latent andlytic proteinsMS and control samples were categorized by the number ofimmunohistologically detectable CD138+, LMP-1+, andBZLF-1+ cells/mm2. Results are grouped accordingly andpresented as follows: no cells/mm2 (−), <5 cells/mm2 (+),5–10 cells/mm2 (++), and >10 cells/mm2 (+++) (table andfigure 5). In samples containing less than 10 cells/mm2 (−, +,and ++) of either CD138 or LMP-1, we did not observedetectable increases in patients with MS compared withcontrols (figure 5, A and B). Although results did not reach

statistical significance, MS samples had an increased fre-quency of samples containing greater than 10 cells/mm2

(+++) of either CD138 or LMP-1 compared with controls(figure 5, A and B), suggesting that control of an EBV in-fection may be diminished in patients with MS, resulting in anincreased presence of EBV-infected plasma cells vs that seenin healthy controls. Conversely, the percentage of samplescontaining greater than 10 cells/mm2 (+++) of the lytic phasemarker, BZLF1, showed a greater frequency in controls vs MSsamples (figure 5C), although this finding was not statisticallysignificant. When comparing the general presence or absenceof EBV in MS and controls samples, there was no observabledifference in the number of patients expressing CD138, LMP-1, or BZLF1-positive cells (figure 5, D–F) (n = 17 patientswith MS and n = 9 controls). CD138+ plasma cell numbersfound restricted within perivascular spaces (PV) were greaterin controls vs both CAP and CP samples (figure 5G). Nota-bly, 33% of control samples contained parenchymal CD138+

Figure 2 Immunohistological detection of EBV latent and early lytic proteins in MS and control brains

CD20+ B lymphocytes (A–C) and CD138+ plasma cells (black arrowheads) (D–F) in the parenchyma and vasculature in a chronic active plaque (CAP) (A andD), ina chronic plaque (CP) (B and E), and in the vasculature of a healthy control brain sample (C and F). Percentage of CAP, CP, and control brain samples expressingCD138 protein detectable by immunohistochemistry are shown (M). We observed latent membrane protein-1 (LMP-1) expression (red arrowheads) in CAP,CP, and control brain samples (G–I). Percentage of CAP, CP, and control brain samples expressing LMP-1 protein detectable by immunohistochemistry areshown (N). Cells expressing LMP-1 (red arrowheads) were found in the vasculature of CAPs (G) and control brains (I), and within the parenchyma in a CP (H).The expression of the viral immediate-early protein BZLF1 was not observed in a CAP (J) and was observed in and around the vasculature in tissue from a CP(red arrowheads) (K). BZLF1 was also observed to a lesser extent in healthy controls (red arrowheads) (L). Percentage of CAP, CP, and control brain samplesexpressing BZLF1 protein detectable by immunohistochemistry are shown (O). Scale bar = 50 μm. Pictures are representative of analysis from17MS samplesand 9 healthy controls. EBV = Epstein-Barr virus.



plasma cells compared with 57% and 78% in CAPs and CPs,respectively (figure 5H).

DiscussionEBV infection is estimated to affect upward of 95% of adultsworldwide.26 EBV infection causes infectious mononucleosis,

an acute infection with multiorgan involvement. Chronic EBVinfection is associated with various types of malignancy, in-cluding Burkitt lymphoma and certain head and neckcancers.26,27 The notion that EBV may be related to thepathogenesis of MS stemmed from early studies, whichshowed elevated EBV antibody titers in the sera of MS vscontrol patients.28 Whether MS might result from an immu-nopathologic response toward an active EBV infection

Figure 3 Immunohistochemical analyses of positive and negative Epstein-Barr virus (EBV) control tissues usingimmunostaining

Pictures are representative of resultsfrom 2 different experiments using EBV-positive (tonsil with infectious mono-nucleosis) and EBV-negative (normaltonsil) control samples. Immunohisto-chemistry revealed the presence ofCD20+ B lymphocytes (A–B) and CD138+

plasma cells (C–D) in our positive andnegative control tissue (A–D). LMP-1+

(brown staining) and BZLF1+ (brownstaining emphasized with black arrows)cells are shown in a tonsil from a patientwith infectious mononucleosis (E, G). NoLMP-1+ or BZLF1+ cells were observed inour negative control tissue (normal ton-sil) (F, H). Scale bar = 50 μm.



Figure 4 Detection of EBER+ cells in MS and control brains by in situ hybridization

In situ hybridization for EBER detects EBER+ cells (blue-black nuclei, black arrows) in 4 representative MS brains (B, D, H, and K) and 1 representative controlbrain (F). In situ hybridization for EBER and immunohistochemistry for LMP-1 show EBER+ and LMP-1+ cells in the same region or nearby regions in 2representative MS brains (G–L). Tonsil tissue from a patient with infectious mononucleosis (M–O) shows colocalization of EBER and LMP-1. RNA preservationin samples is corroborated by in situ hybridization for oligo dT in serial sections from the sameMS and control samples (A, C, E, G, J, andM). Scale bar = 50 μm.



brought into the CNS by immigrating B cells has been de-bated, and investigation has generated contradicting results.Although some studies report EBV in MS,3–7 others find noassociation.8,9 Initial studies aimed to show that the presenceof EBV in the MS brain by in situ hybridization yieldednegative results.29 More recent studies using immunohisto-chemistry, in situ hybridization, and reverse transcriptase PCRtechniques in post-mortemMS brains report EBV infection in21 of 21 brains analyzed.3 A similar study, using the samecohort of MS brain samples, detected EBV infection in 2 of 12

MS cases analyzed by similar methodologies.17 These con-tradictory results stem from studies using basically similartechnologies. Although technologies may be similar, meth-odological differences may lead to differences in the sensitivityof the assays. Especially, in the case of postmortem autopsybrain specimens, it becomes crucial to assure proper pro-cessing and preservation. Autopsy brain samples with sub-optimal tissue preservation may lead to altered assaysensitivity. In addition, there is a wide spectrum of pathologicfeatures of MS lesions, depending on the type and severity of

Figure 5 Increased frequency of parenchymal CD138- and LMP-1–positive cells in MS

Formalin-fixed paraffin embedded brain tissue from MS and control brains without neurologic disease were cut into 4-μm sections. Hematoxylin and eosin(H&E) and immunohistochemistry were performed using antibodies against latent membrane protein 1 (LMP-1), Epstein-Barr virus (EBV) immediate-earlylytic gene (BZLF1), and Syndecan-1 (CD138), a plasma cell marker. For each MS and control sample, the number of CD138+, LMP-1+, and BZLF1+ cells witha visible nucleus was counted manually to allow semiquantitative analysis and categorization of these markers. Results are semiquantitative and expressedas percentage of patients expressing as no cells/mm2, <5 cells/mm2, 5–10 cells/mm2, and >10 cells/mm2 (A–C). Semiquantitative analysis of CD138 (D) and EBVantigen-positive cells (E,F) inMSandhealthy control samples (D-F). CD138+ cells inMS and control sampleswere characterized by their location in perivascularregions or in the parenchyma (G and H), revealing an increased frequency of parenchymal CD138+ cells in CAPs and CPs vs controls (H). The number of cellswas counted from samples (MS: n = 11 autopsy samples andn =6 biopsy samples; controls samples: n = 9 autopsy samples n = 0biopsy samples). The numberof cells was counted on 3 × 3-cm autopsy sections and on 2 × 2-cm sections for biopsy samples.



the disease and the stage of the lesions.23,30 In addition, falsepositives may be introduced by the use of commercial in situhybridization detection kits in combination with nonspecificEBV antibodies.23 Alternatively, EBV LMP-1 staining, thoughoften strong, may also be focal and weak, thereby contributingto false-negative results.31 All of these factors may greatlycontribute to variations in assay sensitivity.

Here, we used a well-characterized MS brain tissue bank andan array of specific antibodies and reagents, which are ac-cepted as reagents able to reliably detect LMP-1 and BZLF1by immunohistochemistry3,17,23,32 in combination with in situhybridization methods considered the “gold standard” ap-proach for detecting EBERs and latent EBV infection.24,25 Wealso include oligo-T probes to ascertain RNA preservation ineach sample. B lymphocyte and EBV antigen expression in theCNS were detected by both immunohistochemistry and insitu hybridization in the MS brain, although expression wasnot unique to the MS brain. CD138+ plasma cells were oftenobserved outside the vasculature (in the brain parenchyma) inboth chronic and chronic active MS plaque and mainly re-stricted to PV spaces in control samples, suggesting that theMS brain is more permissible to plasma cell extravasation intothe parenchyma. The newly discovered lymphatic vessels inthe CNS serve as highways, which can carry immune cells toand from the healthy brain.33 B cells have been shown to enterall parts of the normal human brain, including the paren-chyma, though in low numbers.34,35 As MS progresses,CD138+ plasma cells accumulate and persist within the CNS,even in the absence of observed inflammation.22,36 It remainsunclear whether the extravasation of plasma cells into theparenchyma is dictated by EBV infection or whether MSpathogenesis involves extravasation of EBV-infected plasmacells deep into the tissue, potentially perpetuating the in-flammatory response in MS.

In both MS and control samples, LMP-1 immunostaining waspredominately observed in the membrane of cells with somestaining of the endosomal compartment (figure 2, G–I). Inour positive and negative control tissue, LMP-1 revealedpredominantly a membrane staining pattern (figure 3E).Differences in these staining patterns are likely due to differ-ences in B cells in the brain vs peripheral tissues (controltonsils). LMP-1 is a lipid raft–associated protein, which canaccumulate in the lipid-rich rafts on the surface of mostB cells.37 Endosomes are similar to the composition of lipid-rich rafts on the surface of B cells and have been shown torecruit LMP-1 in a CD63-dependent manner. On accumula-tion within endosomal vesicles, LMP-1 can escape degrada-tion and persist within host B cells.38

Of interest, the lytic form of EBV, as indicated by immuno-histologic protein expression of BZLF1, was readily detectablein CPs but not in CAPs. Previous reports depicting BZLF1staining patterns have been variable.12,39–43 Here, we showBZLF1 staining in the MS brain as diffuse cytoplasmicstaining along with a darker punctate-like staining (figure 2K).

Our observations are in line with recent studies showingBZLF1 staining in theMS brain.41 This staining pattern differsfrom what we observe in tonsils from a patient with infectiousmononucleosis (figure 3G).

EBV lytic proteins, such as BZLF1, can suppress the pro-duction of IL-2 and IL-6.44 In addition to its role in B-cellmaturation,45 IL-6 is known to play an important role inboth neurogenesis and oligodendroglia genesis during healthand following injury.46 In mice with astrocytes, whichoverexpress IL-6, enhanced revascularization resulted inmore rapid healing after traumatic brain injury.47 EBV-mediated suppression of pro- and anti-inflammatory cyto-kines may greatly influence the influx of inflammatory cells,which can serve to clear damaged cells or demyelinateddebris in an attempt to repair tissue. Whether lytic EBV canmodulate the MS brain microenvironment or whether theMS brain microenvironment (i.e., CPs or CAPs) can mod-ulate the EBV life cycle (i.e., latent or lytic) warrants furtherinvestigation.

Previous studies attempting to demonstrate EBV in the MSbrain observed the presence of EBV-infected B cells inlymphoid-like B-cell follicles.3 These initial studies werechallenged by the inability of different groups to consistentlyidentify EBV-infected B cells in cortical structures such asmeningeal follicles with germinal centers. We were also un-able to identify any lymphoid-like structures in the tissue weexamined, leaving open the question of whether such struc-tures exist and whether meningeal B-cell follicles represent animportant site for accumulation of EBV-infected B cells in theMS brain.

In summary, we observed that EBV infection is present inboth MS and control brains, although EBV-positive cells weremore prevalent and more densely populated in the MS brain.The expression of BZLF1+ cells did not differ between MSand control brains. However, we did not observe BZLF1 ex-pression in CAPs, suggesting that the MS brain may be sen-sitive to changes in the EBV virus life cycle. Of interest, 85% ofMS brains revealed frequent EBER-positive cells, whereasnon-MS brains contained seldom EBER-positive cells. Theseresults corroborate some of the controversial results reportedpreviously for EBER expression in patients with MS.3 Takentogether, our results that are derived from a well-characterizedMS brain tissue bank support previous studies demonstratingthe presence of EBV in the MS brain. Further studies in-vestigating the EBV replication cycle and the role of EBV-infected B cells present in meninges and follicular structuresin MS brains as it relates to disease pathology and plaqueformation are warranted.

Author contributionsM.A. Moreno, M.H. Han, L. Steinman, B.T. Aftab, andR. Khanna formulated the original problem and provided di-rection and guidance. M.A. Moreno performed, collected, andanalyzed data for IHC and wrote the manuscript. N. Or-Geva



collected and analyzed data for ISH and IHC.N.Or-Geva,M.H.Han, and L. Steinman reviewed and edited the manuscript.

AcknowledgmentBrain samples are provided by the Maritime Brain TissueBank, Department of Anatomy and Neurobiology, Faculty ofMedicine, Dalhousie University, and the University ofWashington Alzheimer’s Disease Research Center (AG05136),the Adult Changes in Thought Study (AG006781), andMorris K Udall Center of Excellence for Parkinson’s DiseaseResearch (NS062684).

Study fundingFunding was provided by Atara Biotherapeutics.

DisclosureM.A.Moreno consulted for and received research support fromAtara Biotherapeutics and received research support from theNIH. N. Or-Geva holds a patent for Use of anti-third party cellsfor treatment of a variety of diseases including cancer andautoimmune disease or as protectors of genetically-engineeredcell therapy. B.T. Aftab has been employed by, received re-search support from, and holds stock/stock options in AtaraBiotherapeutics. R. Khanna is on the scientific advisory board ofand is a consultant of Atara Biotherapeutics; is the editor-in-chief of Clinical & Translational Immunology; has patents orpatents pending for T cell epitopes and immunotherapy forvirus-associated diseases including malignancies and autoim-mune diseases; received research support from Atara Bio-therapeutics and National Health and Medical ResearchCouncil; and holds stock options in and received licensingpayments from Atara Biotherapeutics. E. Croze has consultedfor Atara Biotherapeutics, Iris Bay is his consulting company.L. Steinman served on the scientific advisory board of Novartis,Receptos, Atreca, Tolerion, Teva, and AbbVie; received travelfunding and/or speaker honoraria fromCelgene and AbbVie; ison the editorial board of MS Journal and Proceedings NationalAcademy of Sciences; has a patent pending on Cytokines andtype 2 interferons; has multiple patents on Antigen specifictolerance; received research support from Atara Bio-therapeutics, Celgene, and Biogen; holds stock options andboard membership in Tolerion; and is a member of the boardof directors of BioAtla. M. Han served on the advisory com-mittee of Novartis; received travel funding and/or speakerhonoraria from the University of California, San Diego, andCMSC; consulted for Sanofi Genzyme; and received researchsupport from the Guthy Jackson Charitable Foundation forNMO Research and Leducq Foundation. Full disclosure forminformation provided by the authors is available with the fulltext of this article at Neurology.org/NN.

Received December 21, 2017. Accepted in final form April 25, 2018.

References1. Fitzner D, Simons M. Chronic progressive multiple sclerosis—pathogenesis of neu-

rodegeneration and therapeutic strategies. Curr Neuropharmacol 2010;8:305–315.2. Popescu BFG, Pirko I. Pathology of multiple sclerosis : where do we stand ? Con-

tinuum (Minneap Minn) 2013;19:901–921.

3. Serafini B, Rosicarelli B, Franciotta D, et al. Dysregulated Epstein-Barr virus infectionin the multiple sclerosis brain. J Exp Med 2007;204:2899–2912.

4. Ascherio A, Munger KL. Epstein-barr virus infection and multiple sclerosis: a review.J Neuroimmune Pharmacol 2010;5:271–277.

5. Ascherio A, Munger KL. Environmental risk factors for multiple sclerosis. Part I: therole of infection. Ann Neurol 2007;61:288–299.

6. Nemerow GR, Mold C, Schwend VK, Tollefson V, Cooper NR. Identification ofgp350 as the viral glycoprotein mediating attachment of Epstein-Barr virus (EBV) tothe EBV/C3d receptor of B cells: sequence homology of gp350 and C3 complementfragment C3d. J Virol 1987;61:1416–1420.

7. Veroni C, Serafini B, Rosicarelli B, Fagnani C, Aloisi F. Transcriptional profile andEpstein-Barr virus infection status of laser-cut immune infiltrates from the brain ofpatients with progressive multiple sclerosis. J Neuroinflamm 2018;15:18.

8. Opsahl ML, Kennedy PGE. An attempt to investigate the presence of Epstein Barrvirus in multiple sclerosis and normal control brain tissue. J Neurol 2007;254:425–430.

9. Sargsyan SA, Shearer AJ, Ritchie AM, et al. Absence of Epstein-Barr virus in the brainand CSF of patients with multiple sclerosis. Neurology 2010;74:1127–1135.

10. Price AM, Luftig MA. Dynamic Epstein-barr virus gene expression on the path to B-cell transformation. Adv Virus Res 2014;88:279–313.

11. Takada K, Shimizu N, Sakuma S, Ono Y. Trans activation of the latent Epstein-Barrvirus (EBV) genome after transfection of the EBV DNA fragment. J Virol 1986;57:1016–1022.

12. Rooney CM, Rowe DT, Ragot T, Farrell PJ. The spliced BZLF1 gene of Epstein-Barrvirus (EBV) transactivates an early EBV promoter and induces the virus productivecycle. J Virol 1989;63:3109–3016.

13. Thacker EL, Mirzaei F, Ascherio A. Infectious mononucleosis and risk for multiplesclerosis: a meta-analysis. Ann Neurol 2006;59:499–503.

14. Nielsen TR, Rostgaard K, Nielsen NM, et al. Multiple sclerosis after infectiousmononucleosis. Arch Neurol 2007;64:72–75.

15. Pohl D, Krone B, Rostasy K, et al. High seroprevalence of Epstein-Barr virus inchildren with multiple sclerosis. Neurology 2006;67:2063–2065.

16. Alotaibi S, Kennedy J, Tellier R, Stephens D, Banwell B. Epstein-Barr virus in pediatricmultiple sclerosis. JAMA;291:1875–1879, 2004.

17. Willis SN, Stadelmann C, Rodig SJ, et al. Epstein-Barr virus infection is not a char-acteristic feature of multiple sclerosis brain. Brain 2009;132:3318–3328.

18. Peferoen LAN, Lamers F, Lodder LNR, et al. Epstein Barr virus is not a characteristicfeature in the central nervous system in established multiple sclerosis. Brain 2010;133.

19. Han MH, Lundgren DH, Jaiswal S, et al. Janus-like opposing roles of CD47 inautoimmune brain inflammation in humans and mice. J Exp Med 2012;209:1325–1334.

20. Han MH, Hwang S, Roy DB, et al. Proteomic analysis of active multiple sclerosislesions reveals therapeutic targets. Nature 2008;451:1076–1081.

21. Kutzelnigg A, Lucchinetti CF, Stadelmann C, et al. Cortical demyelination and diffusewhite matter injury in multiple sclerosis. Brain 2005;128:2705–2712.

22. Frischer JM, Bramow S, Dal-Bianco A, et al. The relation between inflammation andneurodegeneration in multiple sclerosis brains. Brain 2009;132:1175–1189.

23. Lassmann H, Niedobitek G, Aloisi F, Middeldorp JM. Epstein-Barr virus in themultiple sclerosis brain: a controversial issue-report on a focused workshop held in theCentre for Brain Research of the Medical University of Vienna, Austria. Brain 2011;134:2772–2786.

24. Khan G, Coates PJ, Kangro HO, Slavin G. Epstein Barr virus (EBV) encoded smallRNAs: targets for detection by in situ hybridisation with oligonucleotide probes. J ClinPathol 1992;45:616–620.

25. Gulley ML, Tang W. Laboratory assays for Epstein-Barr virus-related disease. J MolDiagn 2008;10:279–292.

26. Cohen JI, Primary immunodeficiencies associated with EBV disease. Curr TopMicrobiol Immunol 2015;390:241–265.

27. Hjalgrim H, Askling J, Sørensen P, et al. Risk of Hodgkin’s disease and other cancersafter infectious mononucleosis. J Natl Cancer Inst 2000;92:1522–1528.

28. Sumaya CV, Myers LW, Ellison GW. Epstein-Barr virus antibodies in multiple scle-rosis. Arch Neurol 1980;37:94–96.

29. Hilton DA, Love S, Fletcher A, Pringle JH. Absence of Epstein-Barr virus RNA inmultiple sclerosis as assessed by in situ hybridisation. J Neurol Neurosurg Psychiatry1994;57:975–976.

30. Lucchinetti C, Bruck W, Parisi J, et al. Heterogeneity of multiple sclerosis lesions:implications for the pathogenesis of demyelination. Ann Neurol 2000;47:707–717.

31. Gulley ML, Glaser SL, Craig FE. Guidelines for interpreting EBER in situ hybrid-ization and LMP1 immunohistochemical tests for detecting Epstein-Barr virus inHodgkin lymphoma. Am J Clin Pathol 2002;117:259–267.

32. Niedobitek G, Agathanggelou A, RoweM, et al. Heterogeneous expression of Epstein-Barr virus latent proteins in endemic Burkitt’s lymphoma. Blood 2016;86:659–665.

33. Louveau A, Smirnov I, Keyes TJ, et al. Structural and functional features of centralnervous system lymphatic vessels. Nature 2015;523:337–341.

34. Blauth K, Owens GP, Bennett JL. The ins and outs of B cells in multiple sclerosis.Front Immunol 2015;6:1–7.

35. Anthony IC, Crawford DH, Bell JE. B lymphocytes in the normal brain: contrasts withHIV-associated lymphoid infiltrates and lymphomas. Brain 2003;126:1058–1067.

36. Moreno M, Bannerman P, Ma J, et al. Conditional ablation of astroglial CCL2 sup-presses CNS accumulation of M1 macrophages and preserves axons in mice withMOG peptide EAE. J Neurosci 2014;34:8175–8185.

37. Muntasell A, Berger AC, Roche PA. T cell-induced secretion of MHC class II-peptidecomplexes on B cell exosomes. EMBO J 2007;26:4263–4272.




38. Verweij FJ, van Eijndhoven MAJ, Hopmans ES, et al. LMP1 association with CD63 inendosomes and secretion via exosomes limits constitutive NF-κB activation. EMBO J2011;30:2115–2129.

39. Angelini DF, Serafini B, Piras E, et al. Increased CD8+ T cell response to Epstein-Barrvirus lytic antigens in the active phase of multiple sclerosis. Plos Pathog 2013;9.

40. SunCC, Thorley-LawsonDA. Plasma cell-specific transcription factor XBP-1s binds to andtransactivates the Epstein-Barr virus BZLF1 promoter. J Virol 2007;81:13566–13577.

41. Hassani A, Corboy JR, Al-Salam S, Khan G, Epstein-Barr virus is present in the brainof most cases of multiple sclerosis and may engage more than just B cells. PLoS One2018;13:e0192109.

42. Frangou P, Buettner M, Niedobitek G. Epstein-Barr virus (EBV) infection in epi-thelial cells in vivo: rare detection of EBV replication in tongue mucosa but not insalivary glands. J Infect Dis 2005;191:238–242.

43. Calattini S, Sereti I, Scheinberg P, Kimura H, Childs RW, Cohen JI. Detection of EBVgenomes in plasmablasts/plasma cells and non-B cells in the blood of most patientswith EBV lymphoproliferative disorders by using Immuno-FISH. Blood 2010;116:4546–4559.

44. Jochum S, Moosmann A, Lang S, Hammerschmidt W, Zeidler R. The EBV immu-noevasins vIL-10 and BNLF2a protect newly infected B cells from immune recog-nition and elimination. Plos Pathog 2012;8:e1002704.

45. Hunter CA, Jones SA. IL-6 as a keystone cytokine in health and disease. Nat Immunol2015;16:448–457.

46. Erta M, Quintana A, Hidalgo J. Interleukin-6, a major cytokine in the central nervoussystem. Int J Biol Sci 2012;8:1254–1266.

47. PenkowaM, Giralt M, LagoN, et al. Astrocyte-targeted expression of IL-6 protects theCNS against a focal brain injury. Exp. Neurol 2003:130–148.



ARTICLE OPEN ACCESS

Mononuclear cell transcriptome changesassociated with dimethyl fumarate in MSArie R. Gafson, MBBS, MA, PhD, MRCP, Kicheol Kim, PhD, Maria T. Cencioni, PhD, Wim van Hecke, PhD,

Richard Nicholas, FRCP, Sergio E. Baranzini, PhD, and Paul M. Matthews, DPhil, FRCP, FMedSci


Correspondence

Prof. Matthews

[email protected]

AbstractObjectiveTo identify short-term changes in gene expression in peripheral blood mononuclear cells(PBMCs) associated with treatment response to dimethyl fumarate (DMF, Tecfidera) inpatients with relapsing-remitting MS (RRMS).

MethodsBlood samples were collected from 24 patients with RRMS (median Expanded Disability StatusScale score, 2.0; range 1–7) at baseline, 6 weeks, and 15 months after the initiation of treatmentwith DMF (BG-12; Tecfidera). Seven healthy controls were also recruited, and blood sampleswere collected over the same time intervals. PBMCs were extracted from blood samples andsequenced using next-generation RNA sequencing. Treatment responders were defined usingthe composite outcome measure “no evidence of disease activity” (NEDA-4). Time-course andcross-sectional differential expression analyses were performed to identify transcriptomicmarkers of treatment response.

ResultsTreatment responders (NEDA-4 positive, 8/24) over the 15-month period had 478 differ-entially expressed genes (DEGs) 6 weeks after the start of treatment. These were enriched fornuclear factor (erythroid-derived 2)-like 2 (Nrf2) and inhibition of nuclear factor κB (NFκB)pathway transcripts. For patients who showed signs of disease activity, there were no DEGs at 6weeks relative to their (untreated) baseline. Contrasting transcriptomes expressed at 6 weekswith those at 15 months of treatment, 0 and 1,264 DEGs were found in the responder andnonresponder groups, respectively. Transcripts in the nonresponder group (NEDA-4 negative,18/24) were enriched for T-cell signaling genes.

ConclusionShort-term PBMC transcriptome changes reflecting activation of the Nrf2 and inhibition ofNFκB pathways distinguish patients who subsequently show a medium-term treatment re-sponse with DMF. Relative stabilization of gene expression patterns may accompany treatment-associated suppression of disease activity.

From the Division of Brain Sciences (A.R.G., M.T.C., R.N.), Department of Medicine, Imperial College London; the Department of Neurology (K.K.), Weill Institute for Neurosciences,University of California, San Francisco; icometrix (W.v.H.), Begaultlaan, Leuven, Belgium; the Department of Neurology (S.E.B.), Weill Institute for Neurosciences, Institute for HumanGenetics and Graduate Program in Bioinformatics, University of California, San Francisco; and Division of Brain Sciences (P.M.M.), Department of Medicine, the Centre forNeurotechnology and the UK Dementia Research Institute, Imperial College London.



This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY), which permits unrestricted use, distribution, and reproduction in anymedium, provided the original work is properly cited.





http://creativecommons.org/licenses/by/4.0/

Relapsing-remitting MS (RRMS) is an autoimmune diseaseaffecting the CNS. Many disease-modifying treatments(DMTs) are available, but none have efficacy in all patients, allare expensive, and all are associated with possible adverseevents.1 Stratifying patients to the best tolerated and mostefficacious treatment either before or soon after commencingtreatment would enhance relative benefits and reduce harm.2,3

Currently, themost common approach to ensuring that patientsare receiving efficacious medication involves vigilant diseasemonitoring using clinical measures and serial MRI. The mostrecent composite “no evidence of disease activity” (NEDA-4)outcome measure combines 4 indices of disease activity:relapses, Expanded Disability Status Scale (EDSS) progression,new MRI activity (Gd+ lesions or new/newly enlarging T2lesions), and relative brain volume loss.4,5 While sensitive todisease activity, a limitation of this approach is the time ofobservation required to meaningfully assess changes in thesecomponent measures. There is a need for shorter term pre-dictive markers of clinical outcomes.

Dimethyl fumarate (DMF) (BG-12; Tecfidera) is a first-linetherapy of moderate efficacy approved for use in RRMS. How-ever, a recent post hoc analysis revealed that only 26% of patientsin the original DEFINE/CONFIRM trials were without clinicalevidence of disease activity or new or enlarging T2-hyperintenselesions on MRI at 2 years.6 This suggests that there may beresponder and nonresponder populations. However, currently,there is no early response stratification marker to distinguishthem. The hypothesized mechanism of action involving modu-lation of nuclear factor (erythroid-derived 2)-like 2 (Nrf2)7,8 andnuclear factor κB (NFκB)9 in immune cells suggested thatpharmacodynamic effects on the transcriptome of peripheralblood mononuclear cells (PBMCs) could be used to predicttreatment response.10 To test this, we used next-generationRNAsequencing to identify short-term changes in gene expression at6 weeks after treatment initiation that are associated withmedium-term treatment response defined by the compositeoutcome measure NEDA-4 at 15 months after treatment.

MethodsStandard protocol approvals, registrations,and patient consentsOur research study was reviewed and approved by the NHSResearch Ethics Committee of LondonCamden and Islington

(14/LO/1896). All patients provided written informedconsent.

The study cohort included 24 patients with RRMS (medianEDSS score, 2; range 1–7) recruited from the Imperial Col-lege Healthcare NHS Trust. Inclusion criteria were diagnosisof RRMS by McDonald criteria,11 age between 18 and 65years, intent to commence DMF, and otherwise treatment-free (other DMTs or steroids) for at least 3 months beforesample collection. Exclusion criteria were known or suspectedintolerance or contraindication to MRI. Seven age- and sex-matched healthy volunteers who were not receiving anyprescribed or over the counter medicines were recruited bylocal advertising.

Patients and healthy volunteers attended for 3 study visits. Forpatients, these were at baseline, before the onset of treatment,6 weeks after the initiation of treatment with DMF, and after15 months of DMF treatment. In the volunteer cohort, therewere also 3 study visits over the same time intervals, but nodrug was taken. The EDSS was conducted by a single, trainedphysician (A.R.G.).

Sample collectionNonfasting venous blood samples were collected at eachstudy visit in EDTA tubes. PBMCs were extracted from freshwhole blood within 1 hour of sample collection using a Ficollgradient (Histopaque-1077; Sigma Life Science, St. Louis,MO). Buffy coat containing PBMCs was aspirated usingsterile Pasteur pipettes and washed with sterile phosphate-buffered saline (PBS). Two wash cycles were performed,followed by resuspension of cell pellet in 10 mL PBS. Cellswere then counted using Trypan blue, and aliquots of 5–10million cells were taken for RNA extraction. RNA extractionwas performed on fresh pellet directly after PBMC extractionusing the Qiagen RNeasy kit as per the manufacturer’sguidelines (Qiagen, Hilden, Germany).

RNA-Seq protocolRNAwasmaintained at −80°C before sequencing. The qualityof the RNA prepared was confirmed, and all samples weresequenced at the same time. Quality was assessed as the RNAconcentration measured using a Nanodrop and Qubit 2.0Fluorometer (Life Technologies, Carlsbad, CA), and RNAintegrity was tested using TapeStation (Agilent Technologies,Palo Alto, CA). All samples showed a 260-nm/280-nm

GlossaryAR-BVL = annualized rate of brain volume loss; cDNA = complementary DNA;CDP = confirmed disability progression;DE =differentially expressed;DEG = differentially expressed gene;DMF = dimethyl fumarate;DMT = disease-modifying treatment;EDSS = Expanded Disability Status Scale; IPA = Ingenuity Pathway Analysis; LFC = log-fold change; LOOCV = leave-one-outcross-validation; MCS = Mental Component Score; MSFC = MS Functional Composite; NEDA = no evidence of diseaseactivity; NFκB = nuclear factor κB;Nrf2 = nuclear factor (erythroid-derived 2)-like 2; PBMC = peripheral blood mononuclearcell; PCS = Physical Summary Score; PBS = phosphate-buffered saline; RRMS = relapsing-remitting MS; SF-36 = 36-itemShort-Form.



fluorometer intensity ratio of >2 and an RNA integritynumber of >9.

Library preparation was performed using the Illumina NEB-Next Ultra RNA Library Preparation Kit following the manu-facturer’s recommendations (NEB, Ipswich, MA). MessengerRNA was enriched with Oligo d(T) beads and then denaturedfor 15 minutes at 94°C. Complementary DNA (cDNA) wasthen synthesized, end paired, and adenylated at 39 ends. Then,a universal adapter was ligated to cDNA fragments along withthe index sequence, and the library was further enriched withlimited cycle PCR. The size of the resulting sequencing librarieswas measured using Agilent TapeStation and quantified usingQubit 2.0 Fluorometer and quantitative PCR (Applied Bio-systems, Carlsbad, CA). RNA sequencing was performed on anIllumina HiSeq platform (San Diego, CA) (2 × 150 base pairs;paired-end configuration). The quality of the raw sequencedata (Fastq files) was assessed by Phred scoring. The Phred Qscore of 30 was >96% for all samples.

Raw sequences were aligned to the human reference genomeGRCh38.p10 using Dynamic Read Analysis for Genomicssoftware. Gene hit counts were calculated from the outputBAM files using HTSeq-count, a python library that countsaligned reads overlapping exons for each gene.12,13 Only readsmapping unambiguously to a single gene are counted usingthe software, and reads possibly mapping to more than 1 geneare discarded.

MRI scansThe patients with RRMS underwent MRI at the ImperialCollege Clinical Imaging Facility at 6 weeks and 15 monthsafter the start of treatment (Siemens Verio 3T; 32-channelhead coil; T1-and T2-weighted structural scans). Scans wereanalyzed using MSmetrix, a scanner-independent and clini-cally approved software developed by Icometrix to extractwhole brain atrophy, lesion volume changes, and the numberof new lesions between 2 timepoints.14,15 The longitudinalapproach taken by MSmetrix incorporates both spatial andtemporal information for accurate and consistent lesion seg-mentation based on Markov Random Field modeling anddifference imaging across the 2 timepoints.16

Clinical outcomesClinical outcomes of patients were assessed by a single,trained physician (A.R.G.) before the onset of treatment, 6weeks after the initiation of treatment with DMF, and after15 months of DMF treatment (A.R.G.). At each clinicalvisit, detailed patient histories were taken including dataconcerning any new clinical relapses over the study period,a full EDSS assessment, MS Functional Composite (MSFC)scoring,17 and 36-item Short-Form (SF-36)18 quality-of-lifequestionnaire.

The NEDA-4 outcome measure was used to designatepatients as responders or nonresponders at 15 months.NEDA-4 was defined as no evidence of relapses, active MRI

lesions (both new or enlarged T2 lesions), 6-month con-firmed disability progression (CDP) (defined as an increase inthe EDSS score of 1.5 points from a baseline score of 0, of 1.0point from a baseline score of 1.0 or more, or 0.5 points frombaseline greater 5.0), or a mean annualized rate of brain vol-ume loss (AR-BVL) of more than 0.4%.5 Secondary outcomemeasures included MSFC Z-score comparisons at baselineand 15 months. The SF-36 was scored at baseline and 15months as a Physical Summary Score (PCS) and a MentalComponent Score (MCS).

Statistical analysesTreatment responders and nonresponders as classified byNEDA-4 criteria were studied using both time-course andcross-sectional analyses. Independent differential expressionanalyses were performed on count data derived from HT-Sequsing DESeq219 for the responder and nonresponder groups.DESeq2, which performs differential expression analysis byfirst performing a regularized logarithm transformation, fol-lowed by detection and correction of dispersion estimates thatare too low through modeling using average expressionstrength over all samples, was used for this. An assumptionmade by this software is that genes of similar average ex-pression have similar dispersion. Where counts are low ordispersion is high for a specific gene, DESeq2 shrinks the log-fold change (LFC) toward zero.20 The software provides anLFC value across conditions and an adjusted p value correctedfor multiple comparisons using the Benjamini-Hochberg20

method.

Time-course differential expression contrasts were per-formed between baseline and 6-week samples and thenbetween 6-week and 15-month samples, using a pairedapproach that controlled for intraindividual variations.Cross-sectional contrasts between patients and controlswere performed, controlling for age as a covariate. Adjustedp value for significance (Padj) was set at Padj < 0.05. Foldchange cutoffs were thresholded at log2-fold changeof ±0.3.

Downstream pathway analysis was performed using IngenuityPathway Analysis (IPA) software (Qiagen). Genes of interestimputed into IPA had Padj < 0.05 and log2-fold change of±0.3. Enrichment analysis was performed using g:Profilerisolated to KEGG pathways (p < 0.0005).

MSFC Z-score and SF-36 comparisons at baseline and 15months were calculated using a paired Student t test. Per-mutation analysis was performed using DESeq2 (100-fold)with random selection of patients with RRMS (n = 8). Thesignificance of differences in distributions derived from leave-one-out cross-validation (LOOCV) was determined usinga Student t test (p < 0.05).

Data availabilityAnonymized data will be shared by request from any qualifiedinvestigator.



ResultsBaseline demographics and clinical information of patientsand approximately age- and sex-matched healthy controls areshown in table 1. Eight patients had been on a previous DMT(although none received drug in the 3 months precedingsample collection), and 16 were treatment naive. Twopatients had thalassemia trait, 2 had psoriasis, and 2 had au-toimmune thyroid disorders. Table e-1, links.lww.com/NXI/A50 lists concurrent medications taken by patients withRRMS.

NEDA-4 was achieved by 8/24 patients (33%) over the 15-month period after initiating treatment with DMF. An AR-BVL greater than −0.4% (range, −0.44% to −2.19%) wasfound for 12 patients (50%). Enlarging or new lesions oc-curred in 9 patients (38% and 4 of these had an AR-BVL<−0.4%). Three patients experienced relapses, and 6-monthCDP occurred in 4 patients (2/4 of whom also experiencedrelapses).

The median change in the MSFC score from baseline to 15months for the whole cohort was +0.21 (range, −0.27 to1.33) (p < 0.005). The median change in the SF-36 PCS was+4.4 (range, −39.4 to 51.25), and the median change in theMCS was +2.9 (range, −24.9 to 39.0), but these changeswere not statistically different (p = 0.24 and p = 0.1,respectively).

Short-term pharmacodynamic effects of DMFWe first tested for differentially expressed gene (DEG) be-tween the healthy controls and all the patients with MS beforethe start of DMF, at baseline. Five hundred twenty-two geneswere differentially expressed (DE) between patients andcontrols (Padj < 0.05). Of these, 254 were downregulated inpatients and 268 were upregulated. There was enrichment ofKEGG pathways “B-cell activation involved in the immuneresponse” and “TNF signaling pathway” (p < 0.001).

We assessed the pharmacodynamic effects of DMF inpatients, independently testing for those in the clinical re-sponder and nonresponder groups. In the responder group,there were 478 DEGs 6 weeks after the start of treatment withDMF relative to baseline (padj < 0.05). These differencesshowed enrichment of transcripts related to the Nrf2 pathway(p < 0.0005) (figure 1) and increased expression of thoseassociated with inhibiting NFκB responses (overlap p <0.0005) (figure 2). In the nonresponder group, no consistentDEGs were identified 6 weeks after the start of treatmentrelative to baseline (table 2).

We confirmed the significance of this difference in responderand nonresponder groups by testing for effects of outliervalues using LOOCV. The median numbers of DEGs aftertreatment were 404 and 0 in the responder and nonrespondergroups, respectively (p < 0.0005). We also assessed RNA-Seqdata from untreated healthy controls (n = 7). Comparisonfrom baseline to the end of a 6-week period without anyintervention showed only 7 DEGs (padj < 0.05) (table 2).

Treatment response is associatedwith a stablepattern of gene expressionBetween 6 weeks and 15 months, 0 and 1,264 DEGs weredetected in the responder and nonresponder groups, re-spectively (table 2). We further confirmed a difference be-tween the 2 groups using a 100-fold permutation analysis inrandomly selected combinations of 8 patients with RRMS.The median number of DEGs in this analysis was 702 (range,31–3,230). In healthy controls (n = 7), who were not givenany intervention and who were followed up over the sametime period, there were 180 DEGs (table 2).

The large number of DEGs found between 6 weeks and 15months in the nonresponder group prompted us to test forresponse heterogeneity within this group. We first tested forindividual outliers. Based on PCA of the 16 nonresponders at15 months, responses in 2 patients appeared as outliers andwere therefore removed from further analysis (figure e-1,links.lww.com/NXI/A48). After the removal of these 2 out-liers, 2 distinct nonresponder groups (arbitrarily called groupsA and B) were identified in a subsequent round of PCA(figure e-2). We then independently assessed DEGs between6 weeks and 15 months in these groups: 560 DEGs werefound for group A and 648 for group B (117 [11%] of theseDEGs overlapped between the 2 groups). We tested for thesignificance of these differences relative to the stable expres-sion pattern in the responder group using LOOCV. Themedian numbers of DEGs with LOOCVwere 270 and 497 forgroups A and B, respectively, which both are different to theequivalent analysis for the responder group, in which themedian number of DEGs was 0 (p = 0.03 and p = 0.004,respectively).

In group A, the most enriched canonical pathways were in-volved with Th1 and Th2 activation and T-cell receptor sig-naling (p < 0.0001). The DEGs showed enrichment for the

Table 1 Patient and healthy volunteer demographic data

Patients withMS (n = 24)

Healthycontrols (n = 7)

% Women 50% 43%

Mean age (y) 39.6 ± 13.5 36.7 ± 9.7

Average disease duration fromdiagnosis (y)

4 ± 4 NA

Average disease duration fromfirst symptom (y)

7 ± 6 NA

EDSS (median, range) 2 (1–7) NA

Treatment-naive patients 16 NA

Current smoker 2 2

Abbreviations: EDSS = Expanded Disability Status Scale; NA = not applicable.Values quoted as mean ± SD unless otherwise indicated.






KEGG pathway “T-cell receptor signaling” (p < 0.0001). Acomparison of patients with healthy controls at 15 monthsalso demonstrated enrichment of the T-cell receptor signalingpathway, suggesting that failure to respond well to treatmentis associated with T-cell dysregulation. There were noenriched canonical or KEGG pathways in group B.

DMF treatment is associated with short-termrelative normalization of gene expressionin respondersAfter controlling for sex, 668 DEGs were found in samplesfrom patients in the responder group relative to the healthycontrols at baseline (padj < 0.05). However, 6 weeks after thestart of treatment, only 3 genes were DE between thesepatients and the healthy controls (table 3). At 15 months,there were 85 DEGs between these patients and healthycontrols, although only 14 genes (2%) overlapped with theDEGs found at baseline (figure 3A).

Four hundred seventy-eight DEGs were found in samplesfrom patients in the nonresponder group relative to thehealthy controls at baseline (padj < 0.05) (table 3). Ninety-eight (21%) overlapped with those identified from the re-sponder group. At 6 weeks after the start of treatment, 18genes were DE between the nonresponder patients andhealthy controls, 8 of which (44%) also had been identifiedbaseline (figure 3B). At 15 months, 391 DEGs were found innonresponder group A and 340 in nonresponder group B(table 3).

DiscussionThere is currently no reliable early treatment response pre-diction marker for any RRMS DMT. Here, we explored thehypothesis that short-term, individual pharmacodynamicresponses can distinguish patients who will respond to DMF.

Figure 1 (A–D) Nuclear factor (erythroid-derived 2)-like 2–related transcripts are increased 6 weeks after treatment inresponders but not in nonresponders or healthy controls

Boxplots represent variance-stabilized transformed counts for transcripts (A) FOSL1, (B) ATF4, (C) MAFG, and (D) MGST1 at baseline and 6 weeks inresponders, nonresponders, and healthy controls. ATF4 = activating transcription factor 4; FOSL1 = fos-related antigen 1; MAFG = transcription factor MafG;MGST1 = microsomal glutathione S-transferase 1.



To do this, we tested whether gene expression changes at 6weeks are associated with the medium-term clinical responseto DMF. Using RNAseq, we observed that a robust short-term transcriptomic response to DMF in PBMCs was asso-ciated with activation of the Nrf2 and inhibition of the NFκBpathways in treatment responders. In addition to a robust

short-term pharmacodynamic response to DMF in treatmentresponders, we observed stabilization of gene expression be-tween 6 weeks to 15 months. By contrast, no early tran-scriptional changes were observed after starting DMF innonresponders. We also found greater expression of proin-flammatory pathway genes in nonresponders than in healthycontrols.

A number of previous studies also have investigated thepharmacodynamic effects of DMF on gene expression. Thesedescribed modulation of genes related to antioxidantpathways,21–24 anti-inflammatory pathways,25,26 andNFκB27,28 that may be related to therapeutic effects. Ourresults confirmed changes in expression in all 3 of thesepathways in the subset of patients in whom the drug sup-presses apparent disease activity. We also demonstrated rel-ative stabilization of gene expression over the medium term in

Figure 2 (A–D) NFκB-related transcripts are increased 6 weeks after treatment in responders but not in nonresponders orhealthy controls

Boxplots represent variance-stabilized transformed counts for transcripts (A) CD83, (B) ICAM1, (C) NFκBIA, and (D) NFκBIE at baseline and 6 weeks inresponders, nonresponders, and healthy controls. CD83 = cluster of differentiation 83; ICAM1 = intercellular adhesion molecule 1; NFκB = nuclear factor κB;NFκBIA = nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha; NFκBIE = nuclear factor of kappa light polypeptide gene enhancerin B-cells inhibitor, epsilon.

Table 2 Number of differentially expressed genebetween baseline and 6weeks (short term) and 6weeks and 15 months (medium term) inresponders, nonresponders, and controls

Responders Nonresponders Controls

Baseline vs 6 wk 478 0 7

6 wk vs 15 mo 0 1,264 180



treatment responders. By contrast, the nonrespondersshowed substantial numbers of DEGs over the same timeperiod. In a subset of these nonresponders (group A), wefound increased expression of immune activation pathways.

This study identified DEGs between patients with RRMS andhealthy controls at baseline, before treatment. Although someprevious studies also reported DEGs between thesegroups,29,30 our use of RNA-Seq allowed a wider range ofdiscriminatory transcripts to be identified.31 However, PBMCexpression differences between treatment-responsive RRMSpatients and healthy controls did not persist after the initiationof treatment with DMF. We also provided data suggestingadditional pharmacodynamic effects from stabilization of geneexpression relative to treatment nonresponders. We speculatethat the latter reflects enhanced immune homeostasis and

suggest that the transcriptome differences relative to healthycontrols and their dynamics may be markers of pharmaco-dynamic response in MS. Supportive evidence for the latterhypothesis comes from observations of dynamic tran-scriptome expression with respiratory syncytial virus infectionof an otherwise healthy control monitored with repeatedtranscriptomics over 14 months.32

Overall, our treatment response frequency in this pragmatic,real-world study group was similar to that reported before inlarger populations. We found that 33% of patients achievedNEDA-4 after 15 months of treatment. This is consistent witha recent post hoc analysis reporting on the medium-termNEDA outcomes of the phase III trials (DEFINE/CONFIRM) of DMF in RRMS.6 We also observed an im-provement in the overall MSFC score at 15 months, as alsoreported from the initial DEFINE/CONFIRM studies.33 Al-though others have been able to show an improvement inQOL with DMF as measured by SF-36,34 we were unable toreplicate this finding.

We acknowledge limitations of our work. The small samplesize and using only 3 timepoints limit the confidence in ourselection of transcriptomic response markers and estimationof their effect size. Further confirmatory work is needed.However, we attempted to reduce the effect of these factors byincreasing the rigor of our statistical analyses (e.g., performingLOOCV and permutation tests). The use of a control group atmatching timepoints also allowed us to compare our findingsin the patients with those in a matched, healthy population

Table 3 Number of differentially expressed gene in cross-sectional analysis between patients withrelapsing-remitting MS and controls at baseline,6 weeks, and 15 months

Respondersvs control

Nonrespondersvs control

Baseline 668 478

6 wk 3 18

15 mo 85 391 (group A)

340 (group B)

Figure 3 (A–B) Dimethyl fumarate treatment is associated with a relative normalization of gene expression in respondersbut not in nonresponders

Venn diagrams represent the number of differentially expressed gene in responder (A) and nonresponder (B) groups compared with controls at baseline, 6weeks, and 15 months.



while controlling for potential effects of time. We also mini-mized the effect of artefacts arising from “batch” effects bysequencing all samples at the same time and in the samesequencing facility.

To enhance the power to discriminate responders and thenonresponders, we used the NEDA-4 criteria, which rely onthe assessment of the apparent rate of brain atrophy. In doingso over this short time frame (compared with that used forclinical decision making), we could not formally take intoaccount the possibility of “pseudoatrophy.” We attempted tominimize the potential effect of this confound by “rebaselin-ing” our patients after the initial 6 weeks on medication whengreatest artefactual atrophy might take place.35 To the extentthat we may have misclassified the true treatment response,this approach will underestimate the efficacy of DMF. How-ever, the most likely effect of this will be to reduce the sensitivityto detection of transcriptomic outcomes discriminating treat-ment responders, rather than to generate false-positive signals.Finally, although we performed intragroup predictive testingwith leave-one-out cross-validation, without an independentreplication cohort, we were unable to test the predictive powerof our findings formally.

We have provided evidence that DMF can alter PBMC tran-scriptome profiles of patients withMS even over the short term.The changes that we found support current hypotheses formechanisms of action via activation of Nrf2 and suppression ofNFκB pathways. In addition, we discovered evidence thata treatment response to DMF is associated with enhancedimmune homeostasis that “normalizes” gene expression in thePBMC fraction. Validation and extension of these results mayhave implications for patient stratification for best use of DMFin other inflammatory conditions.36,37 Our work highlights thesensitivity of RNA-Seq transcriptomic pharmacodynamicmeasures of drug response. Although the RNA-Seq wholetranscriptome assays of PBMCs now are relatively costly, RNA-Seq and related methods have the potential to be streamlinedand provided at modest cost to become a future clinical labo-ratory assay if their value is demonstrated.

Author contributionsA.R. Gafson: acquisition of data and analysis and in-terpretation of data. K. Kim: analysis and interpretation ofdata. M.T. Cencioni: acquisition of data. R. Nicholas: studyconcept and design and acquisition of data. S.E. Baranzini:analysis and interpretation of data. P.M. Matthews: studyconcept and design, analysis and interpretation of data, andstudy supervision. W. van Hecke: acquisition of data, imageanalysis and contribution as "in kind" funding for imagemanagement.

AcknowledgmentThe authors gratefully acknowledge support from theImperial College Healthcare Trust Biomedical ResearchCentre. The authors gratefully acknowledge generous fundingfrom the Wellcome Trust, the Medical Research Council,

Imperial College Biomedical Research Centre, and the UKDementia Research Institute.

Study fundingStudy funded by the Wellcome Trust and Imperial CollegeBiomedical Research Centre.

DisclosureA.R. Gafson reports no disclosures. K. Kim received researchsupport from the NIH/NINDS. M.T. Cencioni reports nodisclosures. W. van Hecke is the founder/CEO of and holdsstock or stock options in icometrix. R. Nichols served on thescientific advisory board of Biogen, Roche, and Genzyme;received travel funding and/or speaker honoraria from Bio-gen; and received research support from the Multiple Scle-rosis Society. S.E. Baranzini served on the scientific advisoryboard of Novartis, EMD Serono, and Sanofi-Aventis; receivedgifts from Novartis; received travel and speaker honorariumfrom Novartis; served on the editorial board of MultipleSclerosis Journal, Neurology, and Neurology: Neuroimmunology& Neuroinflammation; has a patent pending for a gene ex-pression signature that could identify patients at high risk ofdeveloping MS; consulted for Novartis, EMD Serono, andTeva Neuroscience; and received research support from theNIH/NINDS, Department of Defense, and NMSS. P.M.Matthews served on the scientific advisory boards of Ipsen,Roche, Biogen, Adelphi Communications MS, Orbimed, andVUMC; received travel funding and/or speaker honorariafrom Biogen and Novartis; served on the editorial board ofNature Rev Neurology and J Neuroimaging; received publishingroyalties from MIT Press and Dana Press; was employed byGlaxoSmithKline; served on the speakers bureau of AdelphiCommunications; received research support from Glax-oSmithKline, Biogen, MRC, EU Marie Curie, NIHR, MSSociety of GB andNI, and Progressive MS Alliance; and holdsstock and stock options in GlaxoSmithKline. Full disclosureform information provided by the authors is available with thefull text of this article at Neurology.org/NN.

Received March 19, 2018. Accepted in final form April 23, 2018.

References1. Comi G, Radaelli M, Soelberg Sorensen P. Evolving concepts in the treatment of

relapsing multiple sclerosis. Lancet 2017;389:1347–1356.2. Gafson A, Craner MJ, Matthews PM. Personalised medicine for multiple sclerosis

care. Mult Scler 2017;23:362–369.3. Matthews PM, Edison P, Geraghty OC, Johnson MR. The emerging agenda of

stratified medicine in neurology. Nat Rev Neurol 2014;10:15–26.4. Giovannoni G, Turner B, Gnanapavan S, Offiah C, Schmierer K, Marta M. Is it time to

target no evident disease activity (NEDA) in multiple sclerosis? Mult Scler RelatDisord 2015;4:329–333.

5. Kappos L, De Stefano N, Freedman MS, et al. Inclusion of brain volume loss ina revised measure of “no evidence of disease activity” (NEDA-4) in relapsing-remitting multiple sclerosis. Mult Scler 2016;22:1297–1305.

6. Havrdova E, Giovannoni G, Gold R, et al. Effect of delayed-release dimethyl fumarateon no evidence of disease activity in relapsing-remitting multiple sclerosis: integratedanalysis of the phase III DEFINE and CONFIRM studies. Eur J Neurol 2017;24:726–733.

7. Linker RA, Lee DH, Ryan S, et al. Fumaric acid esters exert neuroprotective effects inneuroinflammation via activation of the Nrf2 antioxidant pathway. Brain 2011;134:678–692.

8. Scannevin RH, Chollate S, Jung MY, et al. Fumarates promote cytoprotection ofcentral nervous system cells against oxidative stress via the nuclear factor (erythroid-derived 2)-like 2 pathway. J Pharmacol Exp Ther 2012;341:274–284.




9. Gerdes S, Shakery K, Mrowietz U. Dimethylfumarate inhibits nuclear binding ofnuclear factor kappaB but not of nuclear factor of activated T cells and CCAAT/enhancer binding protein beta in activated human T cells. Br J Dermatol 2007;156:838–842.

10. Dubey D, Kieseier BC, Hartung HP, et al. Dimethyl fumarate in relapsing-remittingmultiple sclerosis: rationale, mechanisms of action, pharmacokinetics, efficacy andsafety. Expert Rev neurotherapeutics 2015;15:339–346.

11. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis:2010 revisions to the McDonald criteria. Ann Neurol 2011;69:292–302.

12. Anders S, Pyl PT, Huber W. HTSeq: a Python framework to work with high-throughput sequencing data. Bioinformatics 2015;31:166–169.

13. Fonseca NA, Marioni J, Brazma A. RNA-Seq gene profiling: a systematic empiricalcomparison. PLoS One 2014;9:e107026.

14. Jain S, Sima DM, Ribbens A, et al. Automatic segmentation and volumetry of multiplesclerosis brain lesions from MR images. Neuroimage Clin 2015;8:367–375.

15. Smeets D, Ribbens A, Sima DM, et al. Reliable measurements of brain atrophy inindividual patients with multiple sclerosis. Brain Behav 2016;6:e00518.

16. Jain S, Ribbens A, Sima DM, et al. Two time point MS lesion segmentation in brainMRI: an expectation-maximization framework. Front Neurosci 2016;10:576.

17. Fischer JS, Rudick RA, Cutter GR, et al. TheMultiple Sclerosis Functional CompositeMeasure (MSFC): An integrated approach to MS clinical outcome assessment. MultScler 1999;5:244.

18. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I.Conceptual framework and item selection. Med Care 1992;30:473–483.

19. LoveMI, HuberW, Anders S. Moderated estimation of fold change and dispersion forRNA-seq data with DESeq2. Genome Biol 2014;15:550.

20. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and pow-erful approach to multiple testing. Stat Soc Ser B Methodol 1995;57:289–300.

21. Zhao G, Liu Y, Fang J, Chen Y, Li H, Gao K. Dimethyl fumarate inhibits the ex-pression and function of hypoxia-inducible factor-1alpha (HIF-1alpha). BiochemBiophys Res Commun 2014;448:303–307.

22. Ha CM, Park S, Choi YK, et al. Activation of Nrf2 by dimethyl fumarate improvesvascular calcification. Vascul Pharmacol 2014;63:29–36.

23. Lin SX, Lisi L, Dello Russo C, et al. The anti-inflammatory effects of dimethylfumarate in astrocytes involve glutathione and haem oxygenase-1. ASN Neuro 2011;3:e00055.

24. Han R, Xiao J, Zhai H, Hao J. Dimethyl fumarate attenuates experimental autoim-mune neuritis through the nuclear factor erythroid-derived 2-related factor 2/hemoxygenase-1 pathway by altering the balance of M1/M2 macrophages.J Neuroinflamm 2016;13:97.

25. Tahvili S, Zandieh B, Amirghofran Z. The effect of dimethyl fumarate on gene ex-pression and the level of cytokines related to different T helper cell subsets in pe-ripheral blood mononuclear cells of patients with psoriasis. Int J Dermatol 2015;54:e254–e260.

26. Wilms H, Sievers J, Rickert U, Rostami-Yazdi M, Mrowietz U, Lucius R. Dime-thylfumarate inhibits microglial and astrocytic inflammation by suppressing thesynthesis of nitric oxide, IL-1beta, TNF-alpha and IL-6 in an in-vitro model of braininflammation. J neuroinflammation 2010;7:30.

27. Hund AC, Lockmann A, Schon MP. Mutually enhancing anti-inflammatory activitiesof dimethyl fumarate and NF-kappaB inhibitors–implications for dose-sparing com-bination therapies. Exp Dermatol 2016;25:124–130.

28. Loewe R, Pillinger M, de Martin R, et al. Dimethylfumarate inhibits tumor-necrosis-factor-induced CD62E expression in an NF-kappa B-dependent manner. J InvestDermatol 2001;117:1363–1368.

29. Ratzer R, Sondergaard HB, Christensen JR, et al. Gene expression analysis ofrelapsing-remitting, primary progressive and secondary progressive multiple sclerosis.Mult Scler 2013;19:1841–1848.

30. Achiron A, Gurevich M, Friedman N, Kaminski N, Mandel M. Blood transcriptionalsignatures of multiple sclerosis: unique gene expression of disease activity. AnnNeurol 2004;55:410–417.

31. Ramanathan M, Weinstock-Guttman B, Nguyen LT, et al. In vivo gene expressionrevealed by cDNA arrays: the pattern in relapsing-remitting multiple sclerosis patientscompared with normal subjects. J Neuroimmunol 2001;116:213–219.

32. Chen R, Mias GI, Li-Pook-Than J, et al. Personal omics profiling reveals dynamicmolecular and medical phenotypes. Cell 2012;148:1293–1307.

33. Giovannoni G, Gold R, Kappos L, et al. Delayed-release dimethyl fumarate anddisability assessed by the multiple sclerosis functional composite: integrated analysisof DEFINE and CONFIRM. Mult Scler J Exp Transl Clin 2016;2:2055217316634111.

34. Kappos L, Gold R, Arnold DL, et al. Quality of life outcomes with BG-12 (dimethylfumarate) in patients with relapsing-remitting multiple sclerosis: the DEFINE study.Mult Scler 2014;20:243–252.

35. Zivadinov R, Reder AT, Filippi M, et al. Mechanisms of action of disease-modifyingagents and brain volume changes in multiple sclerosis. Neurology 2008;71:136–144.

36. Al-Jaderi Z, Maghazachi AA. Utilization of dimethyl fumarate and related moleculesfor treatment of multiple sclerosis, cancer, and other diseases. Front Immunol 2016;7:278.

37. Xie X, Zhao Y, Ma CY, et al. Dimethyl fumarate induces necroptosis in colon cancercells through GSH depletion/ROS increase/MAPKs activation pathway. Br J Phar-macol 2015;172:3929–3943.



CLINICAL/SCIENTIFIC NOTES OPEN ACCESS

Fecal microbiota transplantation associated with10 years of stability in a patient with SPMSSeraj Makkawi, MBBS, FRCPC, ABPN, MSCS, CRND, Carlos Camara-Lemarroy, MD, and Luanne Metz, MD,

FRCPC


Correspondence

Dr. Makkawi

[email protected]

Several studies link the gut microbiome and MS immunopathogenesis. Fecal microbiotatransplantation (FMT), the process of transferring fecal microbiota from a healthy donor toa patient, successfully treats recurrent Clostridium difficile enterocolitis and may benefit auto-immune diseases. The potential short-term efficacy of this procedure has been reported in MS.Here, we describe a patient with secondary progressive MS (SPMS) who achieved diseasestability for over 10 years following FMT.

Case reportA 61-year-old woman with MS was followed up in the Calgary MS Clinic since 1988 (age 33years). She had 7 relapses between 1998 and 2001, and her MRI showed numerous periven-tricular, juxtacortical, brainstem, and corpus callosum lesions confirming relapsing remittingMS (RRMS). CSF was not obtained. She started glatiramer acetate in 2001 and has remainedon it since. She has been relapse-free and had stable brain MRI (no new lesions) since.However, between 2001 and 2005, her balance, ambulation, lower limb power, bladder func-tion, and fatigue gradually worsened. Her Expanded Disability Status Scale (EDSS) increasedfrom 2.0 to 3.0. Because of progressive symptom worsening, and stable MRI, this suggesteda diagnosis of SPMS. During 2005 and 2006, she had several bouts of C difficile enterocolitisfollowing clindamycin treatment of a gingival infection. Her EDSS score increased to 6.0. As shewas resistant to multiple courses of metronidazole and vancomycin, she received a singleFMT in 2006 from her partner via rectal enema using a standard protocol.1 Her EDSS scoreimmediately stabilized without other treatment or lifestyle changes. Over the next 10 years, herFunctional System scores minimally improved, as did Modified Multiple Sclerosis FunctionalComposite scores after 2012 when this measure was routine in our clinic (figure).

DiscussionThe gut microbiome consists of billions of bacteria, fungi, and viruses, which participate inmaintaining gut homeostasis. Recently, there has been explosion of interest in the gut micro-biota and their impact on normal host immune function and in the development of autoimmunediseases such as MS.

Evidence suggesting a role for the gut microbiome in MS originated from preclinical inves-tigations using the murine model of experimental autoimmune encephalomyelitis (EAE).2

Mice bred under germ-free conditions are resistant to EAE induction, and after induction, theyhave fewer activated pathogenic lymphocytes and increased T-regulatory cells. These changes,and resistance to EAE, can be reversed after recolonization with fecal microbiota from mice

From the Department of Clinical Neurosciences (S.M., C.C.-L., L.M.), Cumming School of Medicine, University of Calgary, Alberta, Canada; and King Saud Bin Abdulaziz University forHealth Sciences (S.M.), Jeddah, Saudi Arabia.









bred under standard conditions. Together, this evidencebuilds a case for the gut microbiome as a neuroimmunemodulator in EAE.3

Recent studies have shown that gut microbiome compositionsin patients with RRMS differ from those in controls. Somedifferences have been associated with relapse risk andperipheral immunologic derangements.4 However, a clear andconsistent “MS microbiome phenotype” has not been de-scribed, and amyriad of different species have been implicated.

The potential role of the gut microbiome in MS pathogenesishas led to proposals for microbiome modification as a thera-peutic strategy. Although dietary interventions and the use ofprobiotics have not been effective, FMT presents an idealmethod of microbiome modification. Decades of safe usemake it an attractive treatment option for autoimmune dis-eases, including MS.5

In a previous report, three people with MS had significantneurologic symptom improvement after FMT for constipation.6

To the best of our knowledge, this case is the first reportsuggesting the potential long-term benefit of FMT on MSdisease progression. An inflammatory response, triggered byrecurrent C difficile infections, likely led to rapid neurologicworsening in 2006 that stabilized after correction of gutdysbiosis by FMT. Her symptoms increased gradually,although her EDSS did not increase continuously before2006, so it is possible that she did not have SPMS. Herneurologic deficits may have been more amenable to reversalthan would be expected if they had been longstanding. Wehypothesize that her transplanted gut microbiome failed todrive progression and possibly permitting limited endogenousrepair. However, as we do not know the composition of hermicrobiome before she acquired C difficile, or after the FMT,this remains speculative. As no interventions have yet beenshown to stimulate repair, the time course of recovery in MSremains unknown. Control of ongoing inflammation is,however, believed to be important.

The prospect of a novel treatment of MS based on a soundtheoretical basis is encouraging. Clinical trials are required to

Figure EDSS, FS scores, and Modified MSFC scores

(A) EDSS and FS scores. (B) Modified MSFC scores. All scores were recorded in the patient chart at the time of each visit. The first clinic visit after fecalmicrobiota transplantation was in 2007. Oral SDMT replaces the PASAT. EDSS = Expanded Disability Status Scale; FS = Functional System; MSFC = MultipleSclerosis functional composite; SDMT = Symbol Digit Modalities Test; PASAT = Paced Auditory Serial Addition Test.



guide the future use of FMT by evaluating its effectiveness,safety profile, and mechanism of action. Indeed, a phase-2 trialis ongoing.7 Challenges will include donor and patient se-lection, dosing, and selection of outcomes.

Author contributionsS.M. and C.C.-L. prepared the initial draft of the manuscript.S.M., C.C.-L., and L.M. contributed to the revision of themanuscript. S.M., C.C.-L., and L.M. agreed to the publicationof this version of the manuscript.


DisclosureS. Makkawi and C. Camara-Lemarroy report no disclosures.L. Metz served on the scientific advisory board of DSMB andRick Hansen Institute; has a patent pending for Combinationof minocycline plus hydroxychloroquine in demyelinatingdisease; consulted for EMD Serono; and received research

support from the Multiple Sclerosis Society of Canada. Fulldisclosure form information provided by the authors isavailable with the full text of this article at Neurology.org/NN.

Received January 3, 2018. Accepted in final form February 28, 2018.

References1. Louie T, Adams PC. Nature’s therapy for recurrent Clostridium difficile diarrhea. Can J

Gastroenterol Hepatol 2008;22:455–456.2. Berer K, Mues M, Koutrolos M, et al. Commensal microbiota and myelin autoantigen

cooperate to trigger autoimmune demyelination. Nature 2011;479:538–541.3. Adamczyk-Sowa M, Medrek A, Madej P, Michlicka W, Dobrakowski P. Does the gut

microbiota influence immunity and inflammation in multiple sclerosis pathophysi-ology? J Immunol Res 2017;2017:7904821.

4. Probstel AK, Baranzini SE. The role of the gut microbiome in multiple sclerosis riskand progression: towards characterization of the “MS microbiome”. Neuro-therapeutics 2018;15:126–134.

5. Xu MQ, Cao HL, Wang WQ, et al. Fecal microbiota transplantation broadening itsapplication beyond intestinal disorders. World J Gastroenterol 2015;21:102–111.

6. Borody TJ, Leis SM, Campbell J, Torres M, Nowak A. Fecal microbiota trans-plantation (FMT) in multiple sclerosis (MS) [abstract]. Am J Gastroenterol 2011;106:S352.

7. Kremenchutzky M. Fecal microbial transplantation in relapsing multiple sclerosispatients. In: ClinicalTrials.gov [online]. Available at: clinicaltrials.gov/ct2/show/NCT03183869. Accessed February 2, 2018.



http://ClinicalTrials.gov

https://clinicaltrials.gov/ct2/show/NCT03183869

https://clinicaltrials.gov/ct2/show/NCT03183869


CLINICAL/SCIENTIFIC NOTES OPEN ACCESS CLASS OF EVIDENCE

Obinutuzumab, a potent anti–B-cell agent,for rituximab-unresponsive IgManti-MAG neuropathyGoran Rakocevic, MD, FAAN, Ubaldo Martinez-Outschoorn, MD, and Marinos C. Dalakas, MD, FAAN


Correspondence

Dr. Dalakas

[email protected]

Anti-MAG demyelinating neuropathy is difficult to treat. All immunotherapies have failedexcept for rituximab, a chimeric B-cell–depleting monoclonal antibody against CD20, thathelps up to 40% of patients based on 2 controlled and several uncontrolled series.1–3 Becausethe majority of these patients are left disabled, stronger anti–B-cell agents might be promising.

We describe clinical response and autoantibody changes after treatment with obinutuzumab(Gazyva), a new generation of humanized anti-CD20 monoclonal antibodies, in 2 patients withanti-MAG neuropathy who continued to worsen despite multiple courses of rituximab. Obi-nutuzumab, approved for chronic lymphocytic leukemia (CLL), exerts greater peripheral andlymphoid B-cell depletion4 and might be more effective in rituximab-refractory patients.

Classification of evidenceThis is a single observational study without controls and provides Class IV evidence thatobinutuzumab is safe to use in patients with IgM anti-MAG demyelinating neuropathy.

Patients and treatmentsPatient 1A 71-year-old man developed feet paresthesias that progressed in 4 years to bilateral foot drop.Workup revealed distal demyelinating neuropathy, a benign IgMκ monoclonal gammopathy,elevated IgM levels, and high-titer anti-MAG antibodies (table). The gammopathy was benignincluding normal bone marrow biopsy. He received 3 monthly courses of IVIG without ben-efits. Rituximab, 2 g, was ineffective without affecting the IgM level or anti-MAG titers while hisweakness continued to worsen. Obinutuzumab was then administered in 6 cycles over 6months, as per the CLL protocol, as follows: day 1: 100 mg; day 2: 900 mg; days 8 and 18:1,000 mg each; and 1,000 mg thereafter monthly for 5 months.

Patient 2A 65-year-old man, developed distal leg numbness and paresthesias 13 years ago followingsuccessful therapy for colorectal cancer. The neuropathic symptoms gradually worsened withsensory ataxia and muscle weakness. Workup revealed a demyelinating neuropathy, an IgMκgammopathy, normal bone marrow biopsy, and high-titer anti-MAG antibodies (table). Hissymptoms transiently improved with oral corticosteroids and IVIG. Over the following 7 years,he received 5 courses of rituximab, 2 g every year. His gait and stamina improved after the first 2treatments, but there was no further benefit. He gradually progressed with more weakness,requiring MAFOs and canes for ambulation, and prominent hand tremors. The IgMκ spike and

MORE ONLINE

Class of EvidenceCriteria for ratingtherapeutic and diagnosticstudies

NPub.org/coe

From the Department of Neurology (G.R., M.C.D.), Department of Hematology (U.M.-O.), Thomas Jefferson University, Philadelphia, PA; and Neuroimmunology Unit (M.C.D.),Department of Pathophysiology, Faculty of Medicine, National and Kapodistrian University of Athens, Greece.







http://NPub.org/coe



high anti-MAG antibody titers persisted. Because of severedisease worsening and continuing disability not respondinganymore to rituximab, he was treated with obinutuzumab,administered for 6 months as described above.

ResultsThere was no clinical improvement or worsening in thepatients’ neuropathic symptoms 6 and 12 months aftertreatment with obinutuzumab. In patient 1, the neurologicdeficits remained unchanged several months after therapy.Patient 2, 1 year after therapy, showed signs of progressionin pace consistent with his pretreatment course; no accel-erated worsening related to obinutuzumab was observed.Both patients tolerated the treatment well. Except fortransient mild thrombocytopenia, there were no compli-cations during the administration or the follow-up period.

Despite no clinical benefits, however, the IgM levels normal-ized and remained normal up to a year after obinutuzumab inboth patients (table). Of interest, the anti-MAG antibody titers,6 months after treatments, were also normalized and remainedlow up to 12 months; the IgMκ spike, however, remainedunchanged without discernible differences in the light chain(table). In patient 2, 1 year after obinutuzumab, the anti-MAGtiters started to rise, reaching now >70,000 units.

DiscussionThe clinical success of first-generation glycoengineered type-I,anti–CD20-mediated, B-cell–depleting, monoclonal antibodiesin autoimmune neurologic and rheumatological disorders hasprovided the rationale for using more potent next-generationanti-CD20 agents. For example, ocrelizumab and ofatumumabseemmore effective than rituximab in progressive and relapsingMS.5,6 Obinutuzumab, a third-generation, glycoengineeredtype-II, humanized anti-CD20 monoclonal antibody approvedfor CLL, has increased binding affinity to the Fc receptor onB cells and enhanced complement and antibody-dependentcytotoxicity resulting in extensive B-cell lysis of peripheralB cells, including some within the lymphoid tissues; because it

also affects IL-6 production, it is expected to cause more sus-tained depletion of memory B cells and affect antibody pro-duction. These effects prompted us to evaluate its efficacy inpatients with rituximab-refractory anti-MAG–mediated neu-ropathy.3 Obinutuzumab, administered for 6 months, was safebut did not improve the patients’ symptomatology even up toa year of follow-up. In contrast to rituximab, however, it nor-malized the IgM level and anti-MAG antibody titers (table).This observation suggests an effect beyond B-cell depletion;B cells play a key role in antigen presentation, complementactivation, and cytokine production, such as IL-1, IL-6, and IL-10, that affect immunoregulatory B and T cells and antibodyproduction by plasma cells.7 These preliminary results, even ina limited number of 2 patients, suggest that the IgM anti-MAGantibodies, despite being pathogenic,8 do not seem to correlatewith clinical response. Whether this is related to our patients’advanced disease and severe axonal degeneration or to in-effectiveness of obinutuzumab is unclear. The good toleranceof the drug, however, the more profound induction of B-celldepletion, and effect on antibodies, as demonstrated withnormalization of IgM and anti-MAG titers, suggest that obi-nutuzumab might still be considered as an early treatment ofthis difficult-to-treat neuropathy.

Author contributionsDr. Rakocevic and Dr. Martinez: study concept and design,acquisition of data, analysis and interpretation, and criticalrevision of the manuscript for important intellectual content.Dr. Dalakas: study concept and design, analysis and in-terpretation, critical revision of the manuscript for importantintellectual content, and study supervision.


DisclosureM.Dalakas served on the scientific advisory board of Novartis,Baxalta, and Octapharma; received travel funding and/orspeaker honoraria from Merck/Serono, Octapharma, andPfizer AG; served on the editorial board of/as an editor ofNeurology, BMC Neurology, Acta Myologica, Acta Neurologica

Table IgM levels andanti-MAGantibody titers before and after treatmentwith obinutuzumab in 2 patientswith anti-MAGneuropathy

PatientsIgM levels(normal 40–230 mg/dL)

IgM monoclonalspike

Anti-MAG titers by EIA(normal ≤ 1:1600 units)

Patient 1

Before obinutuzumab 524 mg/dL Present >1:102,400

After obinutuzumab 229 mg/dL Present <1:1,600 (normalized)

Patient 2

Before obinutuzumab 420 mg/dL Present >1:102,400

After obinutuzumab 173 mg/dL Present <1:1,600 (normalized)



Scandinavica, and Therapeutic Advances in Neurology; con-sulted for Therapath, Baxter, Octapharma, CSL, and theDysimmune Diseases Foundation; received institutionalsupport to Thomas Jefferson University and University ofAthens fromMerck Serono, Genzyme, Novartis, the Guillain-Barre Syndrome/CIDP Foundation, Dysimmune DiseasesFoundation, CSL, Biogen, and Newfactor; G. Rakocevicreports no disclosures. U.Martinez-Outschoorn served on theeditorial board of the American Journal of Pathology; receivedresearch support from Otsuka Pharmaceuticals and the NIH/NCI. Full disclosure form information provided by theauthors is available with the full text of this article at Neu-rology.org/NN.

Received January 8, 2018. Accepted in final form March 5, 2018.

References1. Dalakas MC, Rakocevic G, SalajeghehM, et al. Placebo-controlled trial of rituximab in

IgM anti-myelin-associated glycoprotein antibody demyelinating neuropathy. AnnNeurol 2009;65:286–293.

2. Ferfoglia R, Guimarães-Costa R, Viala K, et al. Long-term efficacy of rituximab in IgManti-myelin-associated glycoprotein neuropathy: RIMAG follow-up study. J PeripherNerv Syst 2016;21:10–14.

3. Dalakas MC. Rituximab an anti-MAG neuropathy: more evidence for efficacy andmore predictive factors. J Neurol Sci 2017;377:224–226.

4. Dalakas MC. B cells as therapeutic targets in autoimmune neurological disorders. NatClin Pract Neurol 2008;4:557–567.

5. Montalban X, Hauser SL, Kappos L, et al. Ocrelizumab versus placebo in primaryprogressive multiple sclerosis. N Engl J Med 2017;376:209–220.

6. Bar-Or A, Grove R, Austin D, et al. The MIRROR Study: a randomized, double-blind,placebo-controlled, parallel-group, dose-ranging study to investigate the safety andMRI efficacy of subcutaneous ofatumumab in subjects with relapsing-remittingmultiple sclerosis. Neurology 2014;82:S23.006.

7. Li R, Rezk A, Healy LM, et al. Cytokine-defined B cell responses as therapeutic targetsin multiple sclerosis. Front Immunol 2015;6:626.

8. Latov N. Pathogenesis and therapy of neuropathies associated with monoclonalgammopathies. Ann Neurol 1995;37:S32–S42.






Granulomatous myositis induced by anti–PD-1monoclonal antibodiesNaohiro Uchio, MD, Kenichiro Taira, MD, Chiseko Ikenaga, MD, Atsushi Unuma, MD, Masato Kadoya, MD,

Akatsuki Kubota, MD, PhD, Tatsushi Toda, MD, PhD, and Jun Shimizu, MD, PhD


Correspondence

Dr. Shimizu

[email protected]

With the expanding use of immune checkpoint blockers typified by anti–programmed death-1(PD-1) and anti–cytotoxic T-lymphocyte-associated protein 4 monoclonal antibodies (Abs)for antitumor therapy, the number of patients showing immune-related adverse events (irAEs)is increasing. Skeletal muscle is one of the target tissues of irAEs and several features ofmyopathy as irAEs have been reported: myasthenia gravis (MG) overlap, cardiac involvement,necrotizing myopathy, and inflammatory myopathy with T-cell and B-cell infiltration.1–4

However, the immunopathogenesis of muscle destruction remains unclear. Here, we report 2cases of granulomatous myositis after anti–PD-1 therapy.

Case reportsCase 1A 79-year-old woman with stage IV lung adenocarcinoma developed nonfatigable mild prox-imal limb weakness (Medical Research Council [MRC] grade 4) 14 days after the fourth cycleof nivolumab. She presented no skin rash and showed no ocular, bulbar, truncal, or respiratorysymptoms. She had well-controlled concurrent breast cancer. Laboratory test results showed anelevated serum creatine kinase (CK) level of 1,638 IU/L. Myositis-specific autoantibodies(MSAs), anti–acetylcholine receptor (AChR) Ab, and anti–titin Ab were negative. Anti–striated muscle Abs were not measured. Electrocardiography findings were normal. EMGshowed spontaneous activity. Chest CT showed no lesions except the cancers.

Biopsy of the biceps brachii muscle showed patchy mononuclear cell infiltrates and gran-uloma formation in muscle fascicles (figure 1A). In the granulomas, CD11c+ M1 andCD163+ M2 macrophages were abundant (figure e-1, links.lww.com/NXI/A46), andCD11c+ M1 macrophages frequently invaded non-necrotic fibers and formed granuloma-tous collection inside the basal lamina of muscle fibers (figure 1, B–D). CD8+ T cellsinvading non-necrotic fibers were also observed (figure 1E). PD-1+ cells were scattered ingranulomas (figure 1F), and programmed death-ligand 1 (PD-L1) was upregulated on thenon-necrotic fibers around granulomas (figure 1G). Electron microscopy confirmed themacrophage invasion of non-necrotic fibers (figure e-2, links.lww.com/NXI/A46). Nivo-lumab was discontinued, and oral prednisolone 50 mg was started. Her symptoms improveddramatically with complete resolution, and prednisolone was tapered off within 3 weeks.During 20 months of follow-up, she showed no neurologic recurrence and received con-ventional chemotherapy.

Case 2A 70-year-old man who received pembrolizumab (anti–PD-1 Ab) and axitinib (tyrosine kinaseinhibitor) for treatment of renal cell carcinoma developed left ptosis, diplopia, and weaknesswith myalgia in his neck and left shoulder (MRC grade 4) without bulbar, respiratory, or cardiac

From the Department of Neurology (N.U., K.T., C.I., A.U., A.K., T.T., J.S.), Graduate School of Medicine, The University of Tokyo, Japan; and Department of Neurology and Anti-agingMedicine (M.K.), National Defense Medical College, Saitama, Japan.











symptoms 2 days after the second cycle of pembrolizumab.He showed no fluctuation of symptoms or fatigability. Lab-oratory test results showed an elevated serum CK level of1,832 IU/L, positivity for anti-AChR Ab (0.8 nmol/L), andnegativity for MSAs. Anti–striated muscle Abs were notmeasured. EMG of the right arm and repetitive nerve stimu-lation (RNS) in the left deltoid muscle showed no abnor-malities. Electrocardiography and chest CT findings werenormal.

Biopsy of the left deltoid muscle showed mononuclear cellinfiltration and giant-cell granulomas (figure 1H). CD11c+

M1 macrophages frequently invaded non-necrotic fibers(figure 1, I and J). Small granulomas replacing muscle fibersalways contained both CD11c+ M1 and CD163+ M2

macrophages (figure 1, K and L). Pembrolizumab was dis-continued and IV immunoglobulin, oral prednisolone30 mg, and pyridostigmine were started. His symptomsimproved markedly with normalization of serum CK within1 month. Prednisolone was tapered to 7.5 mg over the next 5months. During 6 months of follow-up, he showed noneurologic or oncological recurrence.

DiscussionIn this study, we reported cases of granulomatous myositisas an irAE. Although patients with necrotizing myopathywith a fatal outcome were reported previously,2 our 2patients showed a favorable outcome of immunosuppres-sive therapy.

Figure 1 Histopathologic features of case 1 (A–G) and case 2 (H–L)

(A–G) Case 1. (A) Hematoxylin and eosin (H&E) staining showing granulomas in endomysium. (B–D) Serial sections of a high-power view of the square area in(A) with (B) H&E staining and immunohistochemical analysis of (C) CD11 and (D) collagen IV CD11c+M1macrophages invading non-necrotic muscle fibers andforming granulomatous collection inside collagen IV-positive basal lamina. (E) CD8+ T cells infiltrating endomysium and invading non-necrotic fiber (arrows).(F) ScatteredPD-1+ cells in granuloma and (G) PD-L1 overexpression ongranuloma cells andnon-necrotic fibers surrounding granuloma. (H–L) Case 2. (H) H&Estaining of endomysial granuloma with multinucleated giant cell (arrow). Serial sections with (I) H&E staining and (J) immunohistochemical analysis of CD11showing CD11c+ macrophages attacking non-necrotic fiber. Serial sections immunostained for (K) CD11c and (L) CD163 showing macrophages forminggranulomatous collection within muscle fibers consisting of CD11c+ M1 and CD163+ M2 macrophages (arrows). Bars: 50 μm for A, E, F, G, H, K, and L, and10 μm for B, C, D, I, and J. PD-1 = programmed death-1; PD-L1 = programmed death-ligand 1.



One patient (case 2) with low-titer anti-AChR Ab may beoverlapped withMG. Low titers of anti-AChRAbwith normalRNS were reported in patients with MG or myopathy–MGoverlap as irAEs.2,4

The pathomechanisms of myopathy as an irAE are still unclear.Activated T cells, which target an antigen shared by tumors andskeletal muscles, have been suggested to have a role in irAEs.1

Our study showed not only T cells attacking non-necroticfibers but also M1 macrophages attacking non-necrotic fibersand forming granulomas, indicating that macrophages alongwith T cells are involved in the immunopathogenesis. Bothproinflammatory M1 and anti-inflammatory M2 macrophagesare involved in granuloma formation.5 Recent studies suggestthat PD-1 inhibition directly affects macrophages and PD-1depletion induces M1 polarization.6,7 In our cases, PD-1blockade may have promoted M1 polarization, leading to theM1 macrophage invasion of non-necrotic fibers with granu-loma formation.

Of interest, PD-L1 was overexpressed on non-necrotic fibersaround granulomas. PD-1/PD-L1 pathway activation inhibitsimmune-mediated attacks by T cells. Therefore, PD-L1upregulation on non-necrotic fibers may be a protective re-action. To explore measures to prevent myopathy as an irAE,further studies on the immunopathogenesis are necessary.

Author contributionsDr. Uchio: study design and concept, acquisition of data,analysis and interpretation, and drafting of the manuscript.Drs. Taira, Ikenaga, Unuma, and Kadoya: acquisition of data.Dr. Kubota: critical revision of the manuscript. Dr. Toda:study supervision. Dr. Shimizu: study design and concept,acquisition of data, analysis and interpretation, and criticalrevision of the manuscript.

AcknowledgmentThe authors thank Drs. Ryo Usui, Keiichi Hokkoku (TeikyoUniversity School of Medicine), Tatsuhiko Naito, MasaoOsaki, and Yoshikazu Uesaka (Toranomon Hospital) for helpin the acquisition of clinical data of the patients.


DisclosureN. Uchio, K. Taira, C. Ikenaga, A. Unuma, M. Kadoya, A.Kubota, and T. Toda report no disclosures. J. Shimizu issupported by a Health and Labour Sciences Research Granton Rare and Intractable Diseases (Validation of Evidence-based Diagnosis and Guidelines, and Impact on QOL inPatients with Neuroimmunological Diseases) from the Min-istry of Health, Labour and Welfare of Japan and a Grant-in-Aid for Scientific Research (KAKENHI; 15K09347) fromJSPS. Full disclosure form information provided by theauthors is available with the full text of this article at Neu-rology.org/NN.

Received January 26, 2018. Accepted in final form March 29, 2018.

References1. Johnson DB, Balko JM, Compton ML, et al. Fulminant myocarditis with combination

immune checkpoint blockade. N Engl J Med 2016;375:1749–1755.2. Liewluck T, Kao JC, Mauermann ML. PD-1 inhibitor-associated myopathies:

emerging immune-mediated myopathies. J Immunother 2018;41:208–211.3. Gandiga PC, Wang AR, Gonzalez-Rivera T, Sreih AG. Pembrolizumab-associated

inflammatory myopathy. Rheumatology (Oxford) 2018;57:397–398.4. Suzuki S, Ishikawa N, Konoeda F, et al. Nivolumab-related myasthenia gravis with

myositis and myocarditis in Japan. Neurology 2017;89:1–8.5. Hilhorst M, Shirai T, Berry G, Goronzy JJ, Weyand CM. T cell—macrophage

interactions and granuloma formation in vasculitis. Front Immunol 2014;5:1–14.6. Gordon SR, Maute RL, Ben W, et al. PD-1 expression by tumour-associated mac-

rophages inhibits phagocytosis and tumour immunity. Nature 2017;545:495–499.7. Chen W, Wang J, Jia L, Liu J, Tian Y. Attenuation of the programmed cell death-1

pathway increases the M1 polarization of macrophages induced by zymosan. CellDeath Dis 2016;7:e2115.






Mortality in neuromyelitis optica is stronglyassociated with African ancestryMaureen A. Mealy, PhD(c), RN,* Remi A. Kessler, BA,* Zoe Rimler, BS, Allyson Reid, BA, Lauren Totonis, BS,

Gary Cutter, PhD, Ilya Kister, MD, and Michael Levy, MD, PhD


Correspondence

Dr. Levy

[email protected]

BackgroundNeuromyelitis optica spectrum disorder (NMOSD) is a severe autoimmune disease of the opticnerve, spinal cord, and, less frequently, brain.1 Severity and degree of recovery from relapses arethe factors that determine long-term visual and motor disability and mortality.1 Mortality ratesin NMOSD worldwide range from 9% to 32%, depending on age, relapse rate, and recoveryfrom attacks.2,3 In examining mortality data from 2 large, ethnically diverse NMOSDCenters intheMid-Atlantic United States, we observed a striking race distribution: most deceased patientswere of African ancestry. In this analysis, we focus on race as a risk factor for mortality inNMOSD.

MethodsThis is a retrospective study of all patients with NMOSD seen at 2 large US-based clinics: JohnsHopkins Hospital (Baltimore, MD) and New York University (New York, NY). NMOSD wasdefined by the 2015 International Panel of NMO Diagnosis.4 Race was patient reported,whereas all other clinical and demographic factors, including the cause of death, were confirmedby site investigators. Patients not seen in the previous 12 months were called to verify livingstatus. Time to diagnosis, frequency of clinic/hospital visits, and treatment regimen for relapsesacted as surrogates of health care access. Institutional review boards from both institutionsapproved this study.

ResultsA total of 427 NMOSD patients were included in this analysis, 328 from Johns HopkinsHospital and 99 from New York University. In total, 30 patients died during follow-up (table),with an annual mortality rate of 0.68 deaths per 100 patient-years. The mean disease duration attime of death was 6.9 years. Patients of African ancestry constituted 41% of our clinic pop-ulation, but they comprised 90% of the deceased NMOSD patients, with average age at death of52.3 years. The other 3 deceased patients included an Asian woman aged 85 years, a Caucasianman aged 69 years, and a Caucasian woman aged 65 years. The overall mortality rate in our totalcohort was 7.0%, and among those of African ancestry was 15.4% (p < 0.0001).

Patients in each race group were similar regarding age, sex, aquaporin-4 serostatus, time todiagnosis, acute treatment care, and access to our clinics (table). Although more deceasedpatients were untreated at final follow-up (22% vs. 4%), there was no difference in treatmentrates among the races.

*Shared first authorship.

From the Department of Neurology (M.A.M., R.A.K., L.T., M.L.), Johns Hopkins University, Baltimore, MD; the Department of Neurology (Z.R., A.R., I.K.), New York University; and theUniversity of Alabama (G.C.), Birmingham, AL.








Table Demographic and clinical characteristics of cohorts

Characteristic

Deceased cohort only Total NMOSD cohort

Africanancestry

Caucasianancestry Other

Africanancestry

Caucasianancestry

Hispanicancestry

Asianancestry Total

No. of patients 27 2 1 175 198 35 19 427

Female sex (%) 21 (78) 1 (50) 1(100)

156 (89) 163 (82) 29 (83) 19 (100) 367(86)

Proximity to clinic (%),within mid-Atlantic

27 (100) 1 (50) 1(100)

164 (94) 147 (74) 32 (91) 16 (84) 359(84)

AQP4 seropositive, (%) 23 (85) 1 (50) 1(100)

155 (89) 144 (73) 25 (71) 16 (84) 340(80)

Age at disease onset in y,mean (SD)

43.1 (15.0) 60.2 (3.1) 78.0 37.1 (15.8) 42.6 (17.1) 35.4 (19.5) 46.0 (16.0) 39.9(16.7)

Duration of disease in y,mean (SD)

7.0 (5.0) 6.6 (0.1) 7.0 10.6 (8.1) 10.2 (7.3) 10.0 (6.9) 11.2 (7.9) 10.4(7.7)

Delay in diagnosis 4.0 (4.7) 0.1 (0) 1.0 4.0 (6.0) 4.3 (7.0) 3.6 (4.6) 3.7 (4.4) 4.1(6.3)

Preventive medicationat final follow-up (%)

Azathioprine 4 (15) 0 (0) 0 (0) 12 (7) 15 (8) 6 (17) 1 (5) 34 (8)

Rituximab 9 (33) 1 (50) 0 (0) 94 (54) 117 (59) 19 (54) 11 (58) 241(56)

Mycophenolate mofetil 6 (22) 1 (50) 1(100)

52 (30) 54 (27) 5 (14) 5 (26) 116(27)

Other 2 (7) 0 (0) 0 (0) 9 (5) 8 (4) 3 (9) 1 (5) 21 (5)

None 6 (22) 0 (0) 0 (0) 8 (4) 4 (2) 2 (6) 1 (5) 15 (4)

Received PLEX for the lastrelapse, (%)

22 (81) 2 (100) 0 (0) N/Aa

NMOSD relapse within 1 ybefore death (%)

22 (81) 1 (50) 0 (0) N/A

High cervical/brainstem (%) 20 (74) 1(50) 0 (0)

Age at death in y, mean (SD) 50.1 (14.8) 66.8 (3.0) 85.0 N/A

Cause of death (%) N/A

Complications of NMOSD 19 (70) 2 (100) 1(100)

Cancer 4 (15) 0 (0) 0 (0)

Liver failure 1 (4) 0 (0) 0 (0)

Renal failure 1 (4) 0 (0) 0 (0)

Advanced AIDS 1 (4) 0 (0) 0 (0)

Suicide 1 (4) 0 (0) 0 (0)

Comorbidities, excludingcause of death (%)

N/A

High blood pressure 8 (30) 1 (50) 1(100)

Diabetes 7 (26) 1 (50) 1(100)

Other autoimmune disease 4 (15) 0 (0) 1(100)

Mood disorder 4 (15) 0 (0) 0 (0)

Continued



In 22 of 30 deceased patients (73%), cause of death wasrelated to NMOSD (table). Most deaths, 70%, were precededby a relapse in the brainstem and/or upper cervical spinal cordwithin the previous 12months despite preventivemedicationsin 80% of patients at the time of the fatal relapse.

DiscussionOur study involved a very large patient sample—427 patientsfrom 2 large specialized NMO centers. The overall mortalityrate was 7.0% (30 of 427 patients). This rate is slightly lowerthan in contemporary studies (9%–13%)3,5 and considerablyimproved from older landmark studies (22%–32%).2,6,7

The decrease in mortality over time is likely due to earlierdiagnosis, use of plasmapheresis for acute relapses, and pre-ventive immunotherapies, which have been shown to decreaserelapse rates in observational studies.8 It is important that thedefinition of NMOSD has changed over the past 2 decades,allowing for milder cases to be diagnosed. Thus, the decreasein mortality could also be in part a technical artifact(“Will Rogers effect”).

The most striking finding of this study is the observation thatnearly all the deceased patients in our combined cohort wereof African ancestry. Patients of African ancestry make up 41%of the total cohort, but account for 90% of the mortality. Thisis unlikely due to chance (p < 0.0001, Fisher test) or todifferences in delay in diagnosis, clinic access, or treatment(table). This is also unlikely to be due to differences in deathrates among races: according the CDC, mortality rate amongthose of African race was 0.8% and 0.7% among Caucasians(2009–2014). One other study has identified African ancestryas a strong predictor of mortality in NMOSD. In Brazil, wherethe estimated mortality rate among all patients with NMOSDis 23%, the mortality rate among Afro-Brazilians is a stagger-ing 58%.9 Two European studies did not implicate Africanrace as a risk factor for mortality, but the proportion of Africanpatients in their cohorts was much lower than in the EasternUS and Brazilian NMOSD cohorts.2,10,11 Our results have

important implications for management of patients of Africanancestry with NMOSD. Further studies, especially pro-spective studies assessing factors that affect the severity ofrelapses, may shed light on the high risk of death amongpatients of African ancestry with NMOSD.

Author contributionsM. Levy, R.A. Kessler, M.A. Mealy, G. Cutter, and I. Kistercontributed to the design and conceptualization of the study,analysis and interpretation of the data, and drafting of themanuscript.

Study fundingThis study was funded by the NIH, K08 NS078555(M. Levy).

DisclosureM.A. Mealy received speaker honoraria from the Consortiumof Multiple Sclerosis Centers and research support from theNIH/NCATS. R.A. Kessler, Z. Rimler, A. Reid, and L. Totonisreport no disclosures. G. Cutter served on the scientific advi-sory board of AMO Pharmaceuticals, Apotek, Gilead Phar-maceuticals, Horizon Pharmaceuticals, Modigenetech/Prolor,Merck, Merck/Pfizer, Opko Biologics, Sanofi-Aventis, ReataPharmaceuticals, Receptos/Celgene, Teva, NHLBI, andNICHD; received speaker honoraria from the Consortium ofMultiple Sclerosis Centers and, Teva; served on the editorialboard ofMultiple Sclerosis, JASN, and Alzheimer’s & Dementia:Translational Research & Clinical Interventions; is president ofand holds stock in Pythagoras; consulted for Argenx BVBA,Atara Biotherapeutics, Consortium of Multiple SclerosisCenters, Genzyme, Genentech, Innate Therapeutics, JanssenPharmaceuticals, Klein-Buendel Incorporated, MedImmune,Medday, Nivalis, Novartis, Opexa Therapeutics, Roche, SavaraInc, Somahlution, Teva, Transparency Life Sciences, and TGTherapeutics; participates in the NARCOMS MS PatientRegistry funded by the CMSC; received research supportfrom the NIH/NINDS, NIH/NIAID, UAB Center for AIDSResearch, NIH/NHLBI, NIH/NICHD, NIH/NIA, US De-partment of Defense, NIH/MIAMS Children’s Hospital

Table Demographic and clinical characteristics of cohorts (continued)

Characteristic

Deceased cohort only Total NMOSD cohort

Africanancestry

Caucasianancestry Other

Africanancestry

Caucasianancestry

Hispanicancestry

Asianancestry Total

Asthma 3 (11) 0 (0) 0 (0)

Renal insufficiency 3 (11) 0 (0) 0 (0)

Thyroid disease 3 (11) 0 (0) 0 (0)

Hyperlipidemia 2 (7) 0 (0) 0 (0)

Cancer 1 (3) 1 (50) 0 (0)

Abbreviations: AQP4 = aquaporin-4; NMOSD = neuromyelitis optica spectrum disorder; PLEX = plasma exchange.a Previously published data in NMOSD reported that approximately 20%–78% of patients are escalated to plasma exchange.12



(Boston), Consortium of MS Centers, and MyastheniaGravis Foundation of America; and reviewed statistical datafor Galderma Litigation. I. Kister received served on thescientific advisory board of Biogen IDECMS Franchise DataGeneration and Genentech; consulted for Biogen Idec; andreceived research support from Biogen Idec, Serono,Novartis, Genzyme, Guthy-Jackson Charitable Foundation,and National Multiple Sclerosis Society. M. Levy served onthe scientific advisory board of Asterias, Chugai, and Alexion;is on the editorial board of Multiple Sclerosis and RelatedDisorders; holds patents for Aquaporin-4 sequence that elicitspathogenic T cell response in animal model of neuromyelitisoptica and Use of peptide for diagnostic and therapeuticdevelopments; consulted for Guidepoint Global, GersonLehrman Group, and Cowen Group; and received researchsupport from ViroPharma/Shire, Acorda, AcoPharma,Sanofi, Genzyme, Alnylan, Alexion, Terumo BCT, andNINDS. Full disclosure form information provided by theauthors is available with the full text of this article atNeurology.org/NN.

Received August 24, 2017. Accepted in final form April 9, 2018.

References1. Mealy MA,Wingerchuk DM, Greenberg BM, LevyM. Epidemiology of neuromyelitis

optica in the United States: a multicenter analysis. Arch Neurol 2012;69:1176–1180.2. Kitley J, Leite MI, Nakashima I, et al. Prognostic factors and disease course in

aquaporin-4 antibody-positive patients with neuromyelitis optica spectrum disorderfrom the United Kingdom and Japan. Brain 2012;135:1834–1849.

3. Collongues N, Marignier R, Jacob A, et al. Characterization of neuromyelitis opticaand neuromyelitis optica spectrum disorder patients with a late onset. Mult Scler J2014;20:1086–1094.

4. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnosticcriteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177–189.

5. Cabre P, Gonzalez-Quevedo AG, Bonnan M, et al. Relapsing neuromyelitis optica:long term history and clinical predictors of death. J Neurol Neurosurg Psychiatry2008;80:1162–1164.

6. Wingerchuk DM,HogancampWF, O’Brien PC,Weinshenker BG. The clinical courseof neuromyelitis optica (Devic’s syndrome). Neurology 1999;53:1107–1114.

7. Wingerchuk DM, Weinshenker BG. Neuromyelitis optica: clinical predictors of a re-lapsing course and survival. Neurology 2003;60:848–853.

8. Kessler RA, Mealy MA, Levy M. Treatment of neuromyelitis optica spectrum dis-order: acute, preventive, and symptomatic. Curr Treat Options Neurol 2016;18:2.

9. Papais-Alvarenga RM, Carellos SC, Alvarenga MP, Holander C, Bichara RP, ThulerLC. Clinical course of optic neuritis in patients with relapsing neuromyelitis optica.Arch Ophthalmol 2008;126:12–16.

10. Chan KH, Lee R, Lee JCY, et al. Central nervous system inflammatory demyelinatingdisorders among Hong Kong Chinese. J Neuroimmunol 2013;262:100–105.

11. Jarius S, Ruprecht K, Wildemann B, et al. Contrasting disease patterns in seropositiveand seronegative neuromyelitis optica: a multicentre study of 175 patients. J Neuro-inflammation 2012;9:1–18.

12. Abboud H, Petrak A, Mealy M, Sasidharan S, Siddique L, Levy M. Treatment of acuterelapses in neuromyelitis optica: steroids alone versus steroids plus plasma exchange.Mult Scler 2016;22:185–192.





Unusual neurologic presentation of asepticabscesses syndromePhilippe Nicolas, MD,* Olivier Guerrier, MD,* Amandine Benoit, MD, Francoise Durand-Dubief, MD,

Gerald Raverot, MD, PhD, Sebastien Debarbieux, MD, Clemence Delteil, MD, Alexandre Vasiljevic, MD,

Emmanuel Jouanneau, MD, PhD, Francois Cotton, MD, PhD, Romain Marignier, MD, PhD,

and Sandra Vukusic, MD, PhD


Correspondence

Philippe Nicolas

[email protected]

Clinical caseA 25-year-old woman of North African origin, with no personal or family history, successivelydeveloped noninfectious abscess-like formations involving multiple organs including the CNS.

First, she developed large pustular and ulcerative skin lesions evocative of pyoderma gan-grenosum (PG) (figure, A and B) with hyperthermia up to 38.5°C. Skin biopsies revealeda sterile nonfollicular neutrophilic infiltration of the dermis (figure, C and D), supporting thisdiagnosis. Blood samples showed an inflammatory reaction, but all infectious explorations werenegative. A chest-abdomen-pelvis CT scan revealed numerous nodular splenic formations(figure, E). Corticosteroids (1 mg/kg/d) were initiated for the treatment of PG, and the patientrapidly improved with healing of the cutaneous lesions and regression of the splenic abscesses(figure, F).

Seven months later, she presented with meningeal syndrome with hyperthermia. Brain MRIwas normal. CSF analysis revealed meningitis with 800 leukocytes (90% neutrophils) andhypoglycorrhachia at 1.5 mmol/L, but which was sterile with a negative universal PCR. Thediagnosis of aseptic meningitis was considered. Nonetheless, the patient was treated withamoxicillin, cefotaxime, and dexamethasone for 10 days with a spectacular improvement after24 hours.

Six months later, she presented with progressively worsening headache associated with spatio-temporal disorientation and hyperthermia. Brain MRI revealed hydrocephalus secondary toa voluminous gadolinium-enhancing hypothalamic-pituitary multiloculated lesion (figure, G).The endocrinologic workup showed global pituitary insufficiency requiring supplementation byhydrocortisone, levothyroxine, and sex hormones. CSF analysis revealed pleocytosis with 90leukocytes/mm3 (75% lymphocytes), an increased IgG index (0.77), and oligoclonal bands.Direct microbiologic analyses and cultures were negative in blood and CSF, including myco-bacteria. The QuantiFERON®-TB test was negative. A transsphenoidal biopsy of the pituitarylesion highlighted multiple microabscesses infiltrating and destroying the normal anteriorpituitary tissue. Microabscesses were characterized by a central area of necrotic debris andneutrophils surrounded by a chronic inflammatory infiltrate of plasma cells and macrophages;there was no granuloma (figure, I–K). Cultures were sterile, and the universal and mycobacteria

*These authors contributed equally to the manuscript.

From the Service de Neurologie, sclerose en plaques (P.N., O.G., A.B., F.D.-D., R.M., S.V.), pathologies de la myeline et neuro-inflammation, Hopital Neurologique Pierre Wertheimer,Hospices Civils de Lyon, Bron; Centre des Neurosciences de Lyon (A.V., R.M.), INSERM 1028 et CNRS UMR5292, Equipe FLUID et Observatoire Français de la Sclerose en Plaques;Universite de Lyon (A.V., R.M.); INSERM U1052 (G.R., E.J.), CNRS UMR5286, Cancer Research Center of Lyon; Federation d’Endocrinologie (G.R., E.J.), Groupement Hospitalier Est,Hospices Civils de Lyon, Bron; Service de Radiologie (F.C.), Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Universite Claude Bernard Lyon1, Universite de Lyon, Pierre Benite;CREATIS (F.D.-D., F.C.), UMR 5220 CNRS & U1044 Inserm, Universite Claude Bernard Lyon1, Universite de Lyon, Villeurbanne; Service d’anatomie et cytologie pathologiques (C.D.),Centre Hospitalier de la Timone, France; Centre de Pathologie et Neuropathologie Est (A.V.), Groupement Hospitalier Est, Hospices Civils de Lyon, Bron; Service de dermatologie (S.D.),Centre Hospitalier Lyon Sud, 69495 Pierre Benite; and Service de Neurochirurgie B (E.J.), Hopital Neurologique Pierre Wertheimer, Hospices Civils de Lyon, Bron, France.









PCR were negative. Remarkably, the patient spontaneouslyimproved with the disappearance of disorientation andheadaches and vanishing of the multiloculated lesion (figure,

H), without any specific treatment except substitutive hor-monal replacement; hydrocortisone is not believed to haveanti-inflammatory properties at this dose.

Figure Skin, spleen, and hypothalamic-pituitary lesions

(A–D) Skin lesions: pyoderma gangrenosum (PG)-like ulceration of the elbow (A) and large pustules on the ankle surrounded by an inflammatory basis (B).Histopathologic aspect (courtesy of Dr. B. Balme Centre Hospitalier Lyon Sud) showing a dermal neutrophilic abscess (red circle). HES (hematoxylin, eosin,saffron) staining. Original magnification ×4 (C) and original magnification ×10 (D). (E and F) Spleen lesions: contrast-enhanced abdominal CT showingsplenomegaly related to numerous splenic abscesses (red arrow) (E). Disappearance of the splenic abscesses after steroid treatment (F). (G–K) Hypothalamic-pituitary lesion. (G) Brain MRI: sequential contrast-enhanced sagittal T1-weighted images showing a voluminous (31 mm/17 mm/13 mm) gadolinium-enhancing hypothalamic-pituitary multiloculated lesion (red arrow). (H) Spontaneous vanishing of the hypothalamic-pituitary lesion. (I–K) Pituitaryhistopathology showing microabscesses (I; black arrow) composed of numerous neutrophils (J; arrowhead) and necrotic debris (K). Original magnification×100 (I) and ×400 (J and K). Special histochemical stains (Gram, Grocott, PAS [Periodic acid-Schiff], and Ziehl) did not detect any bacteria, fungi, or myco-bacteria. Universal and mycobacteria PCR were negative. Standard, parasite, and fungal cultures were sterile.



The initially considered diagnoses were ruled out. First, mi-crobiologic investigations, including mycobacteria, were notsuggestive of a septic origin. Second, clinical examination andbiological, radiologic, and neuropathologic features were notsupportive of sarcoidosis—no granuloma—Behcet disease,and histiocytosis. Finally, there was no evidence of malig-nancy, including lymphoma to date, nor immunodeficiency.

DiscussionIn the light of the remarkable history of multiple abscess-likelesions, involving the skin, spleen, meninges and CNS, theextensive microbiologic workup that remained negative, thehistopathologic features, and the corticosensitivity, we finallysettled on the diagnosis of aseptic abscesses (AA) syndrome.We did not perform genetic explorations, and therefore, wecannot exclude the possibility of an associated monogenicautoinflammatory disorder or immunodeficiency syndrome.

AA is a rare autoinflammatory disease mainly affecting youngadults. It is characterized by recurrent attacks of fever anddeep round aseptic lesions potentially involving all organs.1 Aneutrophilic dermatosis, such as PG, may be associated.1,2

Extensive microbiologic workups remain negative, and anti-biotics fail to cure patients but relapses can be prevented bycorticosteroids and immunosuppressive drugs such as TNF-αantagonists or IL-1 receptor antagonists.1,3 There is no spe-cific feature to distinguish a septic from an aseptic abscess onMRI, although it would be helpful in establishing the di-agnosis and eventually preventing the need for biopsies.

The pathophysiology of AA remains unknown but seems tobe a multifactorial combination of neutrophil dysfunction,inflammation, and genetic predisposition.2,4,5 AA shares somefeatures with other entities such as pyogenic arthritis, pyo-derma gangrenosum, and acne syndrome—a monogenicautoinflammatory disorder characterized by PG, acne, andpyogenic sterile arthritis—Behçet disease, and neutrophilicdermatosis such as PG or Sweet syndrome. Neurologic in-volvement in those neutrophilic diseases has already beendescribed in BD6 and in SS.7 These entities widely overlapwith AA, sharing features including a neutrophilic infiltrate,a high incidence of aseptic splenic abscesses, and cortico-sensitivity. The clinical and radiologic presentations were notsupportive of neuro-Behcet and we decided to retain the di-agnosis of AA because the diagnostic criteria for neuro-Sweet7

were not fulfilled, although there is a continuum between PG,SS, and AA.

In conclusion, our case emphasizes the possibility of CNSinvolvement of AA syndrome and highlights the need of anearly diagnosis to rapidly start an appropriate treatment.

Author contributionsData collection: P. Nicolas, O. Guerrier, A. Benoit, G. Raverot,S. Debarbieux, and F. Cotton. Data analysis and interpretation:

P. Nicolas, O.Guerrier, A. Benoit, F. Durand-Dubief,G. Raverot, S. Debarbieux, C. Delteil, A. Vasiljevic,E. Jouanneau, F. Cotton, R. Marignier, and S. Vukusic. Lit-erature search: P. Nicolas, O.Guerrier, and A. Benoit.Figures: P. Nicolas, G. Raverot, S. Debarbieux, C. Delteil,A. Vasiljevic, and F. Cotton. Drafting the article: P. Nicolas,O.Guerrier, R. Marignier, and S. Vukusic. Critical revision ofthe manuscript: A. Benoit, F. Durand-Dubief, G. Raverot,S. Debarbieux, C. Delteil, A. Vasiljevic, E. Jouanneau, andF. Cotton. Final approval of the version to be published:P. Nicolas, O. Guerrier, A. Benoit, F. Durand-Dubief,G. Raverot, S. Debarbieux, C. Delteil, A. Vasiljevic, E. Jouan-neau, F. Cotton, R. Marignier, and S. Vukusic. P. Nicolas andO. Guerrier have contributed equally to the manuscript.


DisclosureP. Nicolas, O. Guerrier, and A. Benoit report no disclosures.F. Durand-Dubief served on the scientific advisory board ofRoche and Merck Serono and received travel funding and/orspeaker honoraria from Biogen, Merck Serono, Sanofi-Aventis, Novartis Pharma, and Roche. G. Raerot, S. Debar-bieux, C. Delteil, A. Vasiljevic, E. Jouanneau, and F. Cottonreport no disclosures. R. Marignier served on the scientificadvisory board of MedImmune and received speaker hono-raria and travel funding fromBiogen Idec, Genzyme, Novartis,Merck Serono, Roche, Sanofi-Aventis, and Teva. S. Vukusicserved on the scientific advisory board of Biogen Idec, Cel-gene, Geneuro, Genzyme, Merck Serono, Novartis Pharma,Roche, Sanofi-Aventis, and Teva; received travel funding and/or speaker honoraria from Biogen Idec, Genzyme, MerckSerono, Novartis, Roche, Sanofi-Aventis, and Teva Pharma;consulted for Biogen Idec, Celgene, Geneuro, Genzyme,Merck Serono, Novartis, Roche, Sanofi-Aventis, and Teva;and received research support from Biogen Idec, Genzyme,Medday, Merck Serono, Novartis, Roche, Sanofi-Aventis,and Teva. Full disclosure form information provided by theauthors is available with the full text of this article atNeurology.org/NN.

Received February 12, 2018. Accepted in final form April 23, 2018.

References1. Andre MF, Piette JC, Kemeny JL, et al. Aseptic abscesses: a study of 30 patients

with or without inflammatory bowel disease and review of the literature. Medicine(Baltimore) 2007;86:145–161.

2. Wallach D, Vignon-Pennamen MD. From acute febrile neutrophilic dermatosis to neu-trophilic disease: forty years of clinical research. J AmAcadDermatol 2006;55:1066–1071.

3. Bardy A, Guettrot-Imbert G, Aumaıtre O, Andre MF. Efficacy of Il-1β blockade inrefractory aseptic abscesses syndrome. Mod Rheumatol 2014;24:217–219.

4. Andre MF, Aumaıtre O, Grateau G, et al. Longest form of CCTGmicrosatellite repeatin the promoter of the CD2BP1/PSTPIP1 gene is associated with aseptic abscessesand with Crohn disease in French patients. Dig Dis Sci 2010;55:1681–1688.

5. Braswell SF, Kostopoulos TC, Ortega-Loayza AG. Pathophysiology of pyodermagangrenosum (PG): an updated review. J Am Acad Dermatol 2015;73:691–698.

6. Hisanaga K. Neuro-neutrophilic disease: neuro-Behçet disease and neuro-Sweetdisease. Intern Med 2007;46:153–154.

7. Hisanaga K, Iwasaki Y, Itoyama Y; Neuro-Sweet Disease Study Group. Neuro-Sweetdisease: clinical manifestations and criteria for diagnosis. Neurology 2005;24:1756–1761.




CORRECTION

High-dose cyclophosphamide without stem cell rescue in immune-mediated necrotizing myopathiesNeurol Neuroimmunol Neuroinflamm 2018;5:e473. doi:10.1212/NXI.0000000000000473

In the Article “High-dose cyclophosphamide without stem cell rescue in immune-mediatednecrotizing myopathies” by C.A. Mecoli et al.,1 the Acknowledgments section should includethe following: “CAM was supported by National Institutes of Health/National Institute ofArthritis and Musculoskeletal and Skin Diseases grant T32AR048522.” The authors regret theomission.

Reference1. Mecoli CA, Lahouti AH, Brodsky RA, Mammen AL, Christopher-Stine L. High-dose cyclophosphamide without stem cell rescue in

immune-mediated necrotizing myopathies. Neurol Neuroimmunol Neuroinflamm 2017;4:e381.

Copyright © 2018 American Academy of Neurology 1


Documents

Volume 5, Number 4, July 2018 Neurology.org/NN · Academy Officers RalphL.Sacco,MD,MS,FAAN,President JamesC.Stevens,MD,FAAN,PresidentElect AnnH.Tilton,MD,FAAN,VicePresident CarlayneE.Jackson,