ARTICLE OPEN ACCESS

Altered fovea in AQP4-IgGndashseropositiveneuromyelitis optica spectrum disordersSeyedamirhosein Motamedi MSc Frederike C Oertel MD Sunil K Yadav PhD Ella M Kadas PhD

Margit Weise MSc Joachim Havla MD Marius Ringelstein MD Orhan Aktas MD Philipp Albrecht MD

Klemens Ruprecht MD Judith Bellmann-Strobl MD Hanna G Zimmermann PhD Friedemann Paul MD and

Alexander U Brandt MD

Neurol Neuroimmunol Neuroinflamm 20207e805 doi101212NXI0000000000000805

Correspondence

Dr Brandt

alexanderbrandtcharitede

AbstractObjectiveTo investigate disease-specific foveal shape changes in patients with neuromyelitis opticaspectrum disorders (NMOSDs) using foveal morphometry

MethodsThis cross-sectional study included macular spectral domain optical coherence tomographyscans of 52 eyes from 28 patients with aquaporin-4 immunoglobulin G (AQP4-IgG)-sero-positive NMOSD 116 eyes from 60 patients with MS and 123 eyes from 62 healthy controls(HCs) retrospectively and an independent confirmatory cohort comprised 3333 patientswith NMOSDMS The fovea was characterized using 3D foveal morphometry We includedperipapillary retinal nerve fiber layer (pRNFL) thickness and combined macular ganglion celland inner plexiform layer (GCIPL) volume to account for optic neuritis (ON)-related neu-roaxonal damage

ResultsGroup comparison showed significant differences compared with HC in the majority of fovealshape parameters in NMOSD but not MS Pit flat disk area average pit flat disk diameter innerrim volume and major slope disk length as selected parameters showed differences betweenNMOSD and MS (p value = 0017 0002 0005 and 0033 respectively) This effect wasindependent of ON Area under the curve was between 07 and 08 (receiver operating char-acteristic curve) for discriminating between NMOSD and MS Pit flat disk area and average pitflat disk diameter changes independent of ON were confirmed in an independent cohort

ConclusionsFoveal morphometry reveals a wider and flatter fovea in NMOSD in comparison to MS andHC Comparison to MS and accounting for ON suggest this effect to be at least in partindependent of ON This supports a primary retinopathy in AQP4-IgGndashseropositive NMOSD

From the Experimental and Clinical Research Center (SM FCO JB-S HGZ FP AUB) Max-Delbruck Center for Molecular Medicine and Charite - Universitatsmedizin Berlincorporate member of Freie Universitat Berlin Humboldt-Universitat zu Berlin and Berlin Institute of Health NeuroCure Clinical Research Center (SM FCO SKY EMK JB-SHGZ FP AUB) Charite ndash Universitatsmedizin Berlin corporate member of Freie Universitat Berlin Humboldt-Universitat zu Berlin and Berlin Institute of Health GermanyDivision of Neuroinflammation and Glial Biology (FCO) University of California San Francisco Nocturne GmbH (SKY EMK) Berlin Department of Neurology (MW MR OAPA) Medical Faculty Heinrich Heine University Dusseldorf Institute of Clinical Neuroimmunology (JH) LMU Hospital Ludwig-Maximilians University Munich Department ofNeurology (MR) Center for Neurology and Neuropsychiatry LVR-Klinikum Dusseldorf Department of Neurology (KR FP) Charite - Universitatsmedizin Berlin corporate memberof Freie Universitat Berlin Humboldt-Universitat zu Berlin and Berlin Institute of Health Germany and Department of Neurology (AUB) University of California Irvine

Go to NeurologyorgNN for full disclosures Funding information is provided at the end of the article

The Article Processing Charge was funded by the authors

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 40 (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

Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1

Aquaporin-4 immunoglobulin G (AQP4-IgG)-seropositiveneuromyelitis optica spectrum disorder (NMOSD) is an in-flammatory astrocytopathy defined by pathogenic serum im-munoglobulin G antibodies against aquaporin-41ndash3

Optic neuritis (ON) is a hallmark of NMOSD and leads tosevere neuroaxonal damage in optic nerve and retina associ-ated with oftentimes severe vision loss4ndash8 Retinal opticalcoherence tomography (OCT) can be used to measure thisdamage9ndash12 Peripapillary retinal nerve fiber layer (pRNFL)and combined macular ganglion cell and inner plexiform layer(GCIPL) typically become thinner whereas inner nuclearlayer (INL) becomes thicker as a result of ON613ndash15

Recently a foveal thickness (FT) reduction has been reportedin eyes never experiencing an ON in patients with AQP4-IgGndashseropositive NMOSD1617 suggesting either subclinicaloptic nerve inflammation or primary retinal astrocytopathy inNMOSD8 This change in FT appeared to be driven bya change in foveal shape with a normally V-shaped foveaappearing more widened and U-shaped with flattened disk ineyes of patients with AQP4-IgGndashseropositive NMOSD17

Because FT is a weak measure for foveal shape we developeda 3D foveal morphometry method which we previously de-scribed and validated in detail18 Here we use this approach toinvestigate the foveal shape in patients with AQP4-IgGndashseropositive NMOSD We compare findings againstmeasurements in patients with MS which also presents withON and against healthy controls (HCs) Our goal was toinvestigate whether foveal changes are characteristic to AQP4-IgGndashseropositive NMOSD and not simply caused by ON

MethodsStudy populationIn this analysis we retrospectively included data from anongoing observational cohort study in patients with NMOSDat the NeuroCure Clinical Research Center at CharitemdashUniversitatsmedizin Berlin Germany acquired from August2013 to November 2016 Inclusion criteria were a minimum ageof 18 years and fulfilling the diagnostic criteria for AQP4-IgGndashseropositive NMOSD according to the 2015 InternationalConsensus Diagnostic Criteria7 AQP4-IgGndashseropositivitywas tested using a cell-based assay (Euroimmun LubeckGermany) Exclusion criteria were any other neurologic orophthalmologic disorder (eg glaucoma diabetes and

refractive error gt6 diopters) which can affect the retina19

Eyes with an episode of ON within the last 6 months beforethe OCT examinations were excluded Of 46 patientsenclosed in the study we included 28 patients with NMOSDin the analysis after applying the inclusion and exclusion cri-teria (table 1) We additionally included 60 patients withrelapsing-remitting MS according to the 2010 revisedMcDonald criteria20 from 2 cohort studies about MS andclinically isolated syndrome and 62 HCs both groups age andsex matched to the NMOSD cohort in this study (table 1)Data from 17 patients with AQP4-IgGndashseropositive NMOSD(61) were already included in a previous study by Oertelet al17 High-contrast visual acuity was measured using EarlyTreatment in Diabetes Retinopathy Study charts at a 4-mdistance with an Optec 6500 P system (Stereo Optical Chi-cago IL) with best correction and under photopic conditions

A confirmatory cohort consisting of macular OCTs from 58eyes of 33 patients with AQP4-IgGndashseropositive NMOSD(eyes with a history of ON [ON+] 27 33 women age 492 plusmn154 years) and 62 eyes of 33 patients with MS (ON+ 12 32women age 497 plusmn 147 years) from longitudinal prospectiveobservational cohort studies at the Department of NeurologyUniversitatsklinikum Dusseldorf at Heinrich Heine UniversityDusseldorf Germany was included in this study following thesame inclusion and exclusion criteria MS and NMOSD groupswere well matched in this cohort for age (p = 0812) and sex (p= 1) but not for the proportion of eyes with ON (p = 0001)The NMOSD group is well matched to the Berlin cohort forage (p = 0113) sex (p = 0214) and ON+ (p = 0507)

Ethics statementThe study was approved by the local ethics committee atCharitemdashUniversitatsmedizin Berlin (EA113109 EA116312 and EA118210) The confirmatory OCT data were col-lected under approval from the local ethics committee atHeinrich Heine University Dusseldorf (4389R) The study wasconducted according to the Declaration of Helsinki in its cur-rently applicable version and the applicable German and Eu-ropean laws All the participants gave written informed consent

Optical coherence tomographyAll retinal OCT images (exploratory and confirmatory co-hort) were taken using Spectralis spectral domain OCTdevices from Heidelberg Engineering (Heidelberg Ger-many) with activated eye tracker and automatic real-time(ART) averaging The pRNFL thickness was calculated usingstandard ring scans around the optic nerve head (12deg single B

GlossaryAQP4-IgG = aquaporin-4 immunoglobulin GART = automatic real timeAUC = area under the curveB = estimate FT = fovealthickness GCIPL = combined macular ganglion cell and inner plexiform layer HC = healthy control ILM = inner limitingmembrane INL = inner nuclear layerMOG = myelin oligodendrocyte glycoproteinNMOSD = neuromyelitis optica spectrumdisorder OCT = optical coherence tomography ON = optic neuritis ONminus = eyes without a history of ON ON+ = eyes witha history of ON pRNFL = peripapillary retinal nerve fiber layerROC = receiver operating characteristic SE = standard error of B

2 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

scan with 1536 A scans 16 le ART le100) The volume of theGCIPL and INL was calculated in a 6-mm diameter aroundthe fovea and FT was measured in a 1-mm-diameter areaaround the fovea based on macular volume scans (25deg times 30deg61 B scans with 768 A scans per each B scan ART = 15)Intraretinal layer segmentation was performed and correctedif needed using Heidelberg Eye Explorer (HEYEX version19100) by an experienced grader All OCT scans werequality controlled according to the OSCAR-IB criteria1921

and OCT data are reported in accordance with the AdvisedProtocol for OCT Study Terminology and Elements(APOSTEL) recommendations22 Four eyes from the ex-ploratory cohort were excluded from the study because ofinadequate OCT scan quality

Foveal morphometryAll macular volume OCTs were analyzed using a 3D fovealmorphometry described previously in detail18 In brief firstthe 3D macular scan is flattened based on the segmentation ofthe Bruch membrane as the reference plane and then theinner limiting membrane (ILM) surface is smoothed andreconstructed radially using a cubic Bezier polynomial modelBased on the reconstructed ILM surface several parametersare defined to describe the foveal and parafoveal shape Threesurfaces are defined in this foveal shape analysis method rimdisk which connects the points on the surface with the max-imum height (rim points) slope disk which connects the

points with the maximum slopes in the parafoveal area and pitflat disk which characterizes the flatness of the foveal pit Eachsurface is described by 4 parametersmdashthe length in thedominant direction (major axis) major length the length inthe second dominant direction (perpendicular to the majoraxis) minor length the area and the average diameter Inaddition the distance between the fovea (theminimum point)and the center of rim disk average pit depth the distancebetween the fovea and the reference plane central fovealthickness the average height of the rim points average rimheight the volume between the ILM surface and the referenceplane rim volume the volume between the ILM surface andrim disk pit volume the volume between the ILM surface andthe reference plane within 1-mm-diameter cylinder centeredat the fovea inner rim volume and the average slope at themaximum slope points average maximum pit slope are mea-sured by this method to characterize the 3D foveal shapeFigure 1 gives an overview of the method and the definedparameters Test-retest reliability was excellent in all fovealmorphometry parameters with intraclass correlation coef-ficients gt09 in all cases (table e-1 linkslwwcomNXIA270)

Statistical analysisSex and ON differences between groups were tested using χ2

tests and age differences were tested using 2-sample Wil-coxon tests Performance measurements were based on thearea under the curve (AUC) for receiver operating

Table 1 Demographic description of NMOSD MS and HC cohorts

NMOSD MS HC

No of patients (N) 28 60 62

No of eyes (N) 52 116 123

Sex (female) (N [])a 26 (93) 55 (92) 56 (90)

Age (y) (mean plusmn SD)b 436 plusmn 115 390 plusmn 109 417 plusmn 135

Patients with a history of ON (N []) 16 (57) 29 (48) mdash

Eyes with a history of ON (N [])c 20 (38) 32 (28) mdash

Number of ONs per eye (median [range]) 15 (1ndash8) 1 (1ndash2) mdash

VA for ON eyes (logMAR) (mean plusmn SD)d 036 plusmn 069 minus009 plusmn 010

EDSS score (median [range]) 3 (0ndash65) 2 (0ndash45) mdash

Disease duration (y) (mean plusmn SD) 68 plusmn 48 79 plusmn 84 mdash

pRNFL (μm) (mean plusmn SD) 812 plusmn 221 908 plusmn 153 976 plusmn 88

GCIPL (mm3) (mean plusmn SD) 169 plusmn 029 183 plusmn 024 194 plusmn 014

INL (mm3) (mean plusmn SD) 094 plusmn 009 096 plusmn 006 095 plusmn 006

FT (μm) (mean plusmn SD) 2629 plusmn 149 2721 plusmn 203 2723 plusmn 231

Abbreviations EDSS = ExpandedDisability Status Scale FT = foveal thickness GCIPL = combinedmacular ganglion cell and inner plexiform layer volume HC =healthy controls INL = inner nuclear layer volume logMAR = logarithm of the minimum angle of resolution NMOSD = neuromyelitis optica spectrumdisorders ON = optic neuritis pRNFL = peripapillary retinal nerve fiber layer thickness VA = high-contrast visual acuitya Sex match p value = 1b Age match p value HC vs MS = 0382 HC vs NMOSD = 0437 MS vs NMOSD = 0056c ON match p value = 0199d VA measurements for 25 (78) ON eyes of patients with MS and 17 (85) ON eyes of patients with NMOSD were available

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 3

characteristic (ROC) curves All linear regression analyseswere performed using linear mixed-effect models includingintereye within-patient correlations age and sex as randomeffects The marginal and conditional coefficients of de-termination of the linear models were calculated with pseudoR-squared Stepwise logistic regression analysis for model se-lection was performed by the Akaike InformationCriteria withboth backward and forward modes of stepwise search basedon generalized linear models In this exploratory study wecorrected p values for multiple testing using the Benjamini-Hochberg procedure In addition the identified parameterdifferences were tested in a second independent cohortobtained with the same scanning protocol at a different centerOne-sided p values were reported for the confirmatory cohortnot corrected for multiple testing Sample size for the confir-matory cohort was calculated for 1-sided 2-sample t-test with90 power and a significance level of 005 To adjust thisestimate for eye-based statistics we added 60 sample size toaccount for intereye within-patient effects and additionalcovariates All statistical analysis were performed in R version35023 with packages stats lme4 lmerTest MuMIn ROCRggplot2 plotROC pwr multcomp and ggpubr packages Thep values less than 005 were considered significant

Data availabilityAll data are available on reasonable request from the corre-sponding author

ResultsFoveal shape changes in NMOSD and MSFirst we analyzed foveal shape in patients with NMOSD andMS and compared results with measurements in HCs Fovealshape was altered in patients with NMOSD but only mildlyaffected in patients with MS both in comparison to HCs (table2) Foveal parameters stratified by history of ON are includedin supplemental data (table e-2 linkslwwcomNXIA271)

In contrast both patients with MS and NMOSD showedneuroaxonal damage typically occurring after ON pRNFL andGCIPL were lower in patients with NMOSD in comparison toHCs (pRNFL standard error of B [SE] = minus177 [30] μmp lt 0001 GCIPL B [SE] = minus027 [004] mm3 p lt 0001) butalso in patients with MS in comparison to HCs (pRNFL B[SE] = minus70 (20) μm p lt 0001 GCIPL B [SE] = minus012 [003]mm3 p = 0001)

Parameter selectionNext we selected parameters with the highest potential to beabnormal in NMOSD We therefore analyzed parameterperformance in discriminating between eyes from patientswith NMOSD and MS regardless of ON status (table 3 andfigure 2A) This was followed by a linear regression analysisagainst diagnosis history of ON and their interaction effect toderive effect sizes and group differences accounting for ON

Figure 1 Three-dimensional foveal shape analysis method overview

(A) ILM surface smoothing and radial reconstruction using the cubic Bezier polynomial (B) Rim height average pit depth and central foveal thickness (C) Rimdisk (blue) slope disk (red) and pit flat disk (green) and major and minor axes on each surface (D) Rim volume pit volume and inner rim volume APD =average pit depth CFT = central foveal thickness ILM = inner limiting membrane major = major axis minor = minor axis

4 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

Table 3 shows the results of this selection process ordered byAUC The best parameter selected from the ROC analysis waspit flat disk area (AUC= 0798 figure 2A) To derive a final setof relevant parameters we computed a stepwise logistic re-gression model to predict NMOSD vs MS including only theparameters with AUC ge07 This selected 4 parameters pitflat disk area average pit flat disk diameter inner rim volumeand major slope disk length (figure 2 DndashG)

Association with ON and neuroaxonal damageA crucial question is whether these parameters react to ON-related damage or are indeed at least partially independent To

further investigate this we repeated the AUC analysis but thistime separately for the eyes without a history of ON (ONminus) andthe ON+ (table e-2 linkslwwcomNXIA271) Indeed fovealshape was altered also in the ONminus Here the best foveal shapeparameter to distinguish ONminus from patients with NMOSD andpatients with MS was minor pit flat disk length (AUC = 0804figure 2B) In ON+ the best-performing parameter to discrim-inate between patients with NMOSD and patients with MS wasmajor pit flat disk length (AUC = 0817 figure 2C) Of noteNMOSD ONminus also showed signs of mild neuroaxonal damagecomparedwithHC (pRNFL B [SE] = minus57 [24] μm p= 0017GCIPL B [SE] = minus012 [004] mm3 p = 0001)

Table 2 Foveal shape analysis parameters results (mean plusmn SD) and linear regression analysis for NMOSD and MS vs HC

HC (mean plusmnSD)

MS (mean plusmnSD)

NMOSD (mean plusmnSD)

HC vs NMOSD HC vs MS

B (SE) p B (SE) p

Average pit depth (mm) 0117 plusmn 0021 0111 plusmn 0018 0101 plusmn 0026 minus0017(0005)

lt0001 minus0006 (0003) 0078

Central foveal thickness (mm) 0231 plusmn 0015 0229 plusmn 0017 0228 plusmn 0017 minus0004(0004)

0330 minus0002 (0003) 0535

Average rim height (mm) 0348 plusmn 0014 0340 plusmn 0016 0328 plusmn 0018 minus0021(0003)

lt0001 minus0009 (0002) lt0001

Average rim disk diameter (mm) 2184 plusmn 0115 2152 plusmn 0110 2132 plusmn 0130 minus0056(0027)

0037 minus0034 (0020) 0082

Rim disk area (mm2) 3717 plusmn 0387 3606 plusmn 0371 3545 plusmn 0436 minus0184(0090)

0041 minus0116 (0066) 0079

Major rim disk length (mm) 0630 plusmn 0066 0613 plusmn 0063 0600 plusmn 0075 minus0032(0015)

0039 minus0018 (0011) 0112

Minor rim disk length (mm) 0619 plusmn 0065 0599 plusmn 0062 0590 plusmn 0072 minus0030(0015)

0042 minus0021 (0011) 0055

Average slope disk diameter (mm) 0663 plusmn 0119 0653 plusmn 0152 0771 plusmn 0136 0104 (0028) lt0001 minus0012 (0025) 0626

Slope disk area (mm2) 0361 plusmn 0131 0358 plusmn 0180 0486 plusmn 0164 0120 (0032) lt0001 minus0005 (0028) 0858

Major slope disk length (mm) 0068 plusmn 0025 0068 plusmn 0034 0090 plusmn 0029 0021 (0006) lt0001 minus48eminus4 (0005) 0930

Minor slope disk length (mm) 0053 plusmn 0019 0052 plusmn 0026 0072 plusmn 0026 0019 (0005) lt0001 minus0001 (0004) 0772

Average pit flat disk diameter(mm)

0215 plusmn 0030 0211 plusmn 0039 0257 plusmn 0052 0042 (0008) lt0001 minus0005 (0006) 0440

Pit flat disk area (mm2) 0037 plusmn 0010 0036 plusmn 0015 0054 plusmn 0025 0017 (0003) lt0001 minus0001 (0002) 0691

Major pit flat disk length (mm) 00067 plusmn00018

00065 plusmn00027

00098 plusmn 00048 00032(00007)

lt0001 minus00001(00004)

0725

Minor pit flat disk length (mm) 00058 plusmn00016

00056 plusmn00023

00083 plusmn 00036 00026(00005)

lt0001 minus00002(00004)

0628

Rim volume (mm3) 1045 plusmn 0153 0983 plusmn 0133 0910 plusmn 0170 minus0141(0036)

lt0001 minus0061 (0025) 0013

Inner rim volume (mm3) 0104 plusmn 0018 0103 plusmn 0019 0088 plusmn 0015 minus0017(0004)

lt0001 minus0001 (0003) 0753

Pit volume (mm3) 0252 plusmn 0043 0246 plusmn 0051 0259 plusmn 0044 0005 (0010) 0606 minus0007 (0008) 0402

Average maximum pit slope(degrees)

1216 plusmn 338 1110 plusmn 241 986 plusmn 311 minus242 (074) 0001 minus103 (052) 0047

Abbreviations B = estimate HC = healthy controls MS = patients with MS NMOSD = patients with neuromyelitis optica spectrum disorders SE = standarderror of BSignificant p values are marked in bold

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 5

Foveal changes may also be driven not by ON per se but bythe amount of neuroaxonal damage after ON Here MS maybe a problematic control group because the amount of ON-related retinal damage is typically lesser than in NMOSD1024

To investigate whether differences in neuroaxonal damage areindeed able to explain the observed foveal differences werepeated the linear regression model analyses for the pre-viously selected 4 parameters but this time corrected addi-tionally for GCIPL and INL (table 4) which can be

considered sensitive parameters for ON severity Here it wasconfirmed that the observed group differences are unlikely tobe explained by ON severity differences alone

See supplemental data for more results on regression analysiscorrected for the neuroaxonal damage (table e-3 linkslwwcomNXIA272) Figure 2 HndashI shows rim volume and av-erage pit depth as example foveal shape parameters signifi-cantly dependent onON status but not on diagnosis Figure 2

Table 3 AUC and linear regression analysis results for NMOSD vs MS sorted in ascending order of AUC

NMOSD vs MS(AUC)

Linear regression with interaction effects of diagnosis and ON history

MS vs NMOSD ON2 vs ON+ NMOSD times ON

R2Marg R2

CondB (SE) p B (SE) p B (SE) p

Pit flat disk area (mm2) 0798 0011 (0005) 0021 0001 (0002) 0544 0015 (0004) 0001 0176 0843

Average pit flat disk diameter(mm)

0796 0033 (0011) 0002 0004 (0005) 0423 0029 (0009) 0002 0200 0876

Minor pit flat disk length (mm) 0796 00018(00007)

0011 00002(00003)

0645 00020(00007)

0007 0172 0827

Major pit flat disk length (mm) 0790 00019(00009)

0038 00003(00004)

0456 00031(00008)

lt0001 0177 0855

Inner rim volume (mm3) 0755 minus0012(0004)

0006 minus0002(0001)

0170 minus0007(0003)

0006 0144 0940

Minor slope disk length (mm) 0744 0017 (0006) 0013 0001 (0002) 0554 0008 (0004) 0051 0111 0941

Slope disk area (mm2) 0739 0102 (0042) 0022 0010 (0011) 0371 0053 (0022) 0022 0096 0958

Average slope disk diameter(mm)

0739 0094 (0035) 0010 0006 (0009) 0509 0050 (0019) 0010 0116 0956

Major slope disk length (mm) 0733 0017 (0008) 0040 0002 (0002) 0265 0010 (0004) 0019 0083 0963

Average rim height (mm) 0664 minus0008(0003)

0022 minus0007(0002)

0001 minus0009(0004)

0022 0107 0882

Rim volume (mm3) 0627 minus0053(0032)

0124 minus0061(0016)

lt0001 minus0044(0032)

0164 0081 0878

Average pit depth (mm) 0602 minus0007(0005)

0147 minus0007(0002)

lt0001 minus0009(0004)

0025 0069 0931

Pit volume (mm3) 0594 0012 (0011) 0369 minus0006(0004)

0239 0002 (0008) 0845 0013 0917

Average maximum pit slope(degrees)

0593 minus074 (059) 0210 minus073 (026) 0008 minus124 (051) 0019 0073 0908

Major rim disk length (mm) 0556 minus0010(0015)

0671 minus0021(0006)

0002 minus0005(0013)

0697 0025 0902

Average rim disk diameter(mm)

0550 minus0014(0026)

0657 minus0039(0012)

0002 minus0010(0023)

0657 0027 0891

Rim disk area (mm2) 0549 minus0043(0088)

0628 minus0128(0039)

0002 minus0038(0077)

0628 0027 0892

Minor rim disk length (mm) 0545 minus0004(0015)

0761 minus0022(0007)

0002 minus0008(0013)

0756 0029 0878

Central foveal thickness (mm) 0543 minus0001(0004)

0891 15eminus4

(0001)0891 minus0001

(0002)0891 0002 0946

Abbreviations AUC = area under the curve B = estimate MS = patients with MS NMOSD = patients with neuromyelitis optica spectrum disorders NMOSD timesON = interaction effect of diagnosis andON ON = optic neuritis ONminus = eyes without a history of ON ON+ = eyes with a history of ON SE = standard error of BR2Cond = conditional R-squared R2

Marg = marginal R-squaredSignificant p values and AUC ge07 are marked in bold

6 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

Figure 2 ROC curves exploratory data analysis for selected parameters and sample fovea

ROC curves for best-performing foveal shape and standard OCT parameters discriminating between (A) NMOSD vs MS (B) NMOSD ONminus vs MS ONminus and (C)NMOSDON + vsMSON+ Box and dot plots for (D) pit flat disk area (E) average pit flat disk diameter (F) inner rim volume and (G)major slope disk length theselected 4 foveal shape parameters (H) Rim volume and (I) average pit depth as example foveal shape parameters affected by ON but not diagnosis Asample central (foveal) B scan of (J) MS ONminus and (K) NMOSD ONminus chosen from the median of the selected pit flat disk parameters in each groupdemonstrating the difference in foveal pit (pit flat disk) between NMOSD and MS in eyes without a history of ON AUC = area under the curve FT = fovealthickness pRNFL = peripapillary retinal nerve fiber layer thickness INL = inner nuclear layer volume HC = healthy controls MS = patients with MS NMOSD =patients with neuromyelitis optica spectrum disorders ON = optic neuritis ONminus = eyes without a history of ON ON+ = eyes with a history of ON ROC =receiver operating characteristic

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 7

JndashK shows sample central B scans (crossing the fovea) ofONminus from patients with NMOSD and MS chosen from themedian of the selected pit flat disk parameters in each group

Parameter confirmationFinally we tested whether the parameters identified in the ex-ploratory analysis could be confirmed in an independent cohortof patients with NMOSD and MS measured with the samedevice and protocol at an independent center Based on dif-ferences in the selected parameters we determined the mini-mum sample size for a confirmatory cohort as n = 38 29 35and 59 eyes per group based on measurements for pit flat diskarea average pit flat disk diameter inner rim volume and majorslope disk length respectively In this confirmatory cohort pitflat disk area and average pit flat disk diameter were confirmedto be significantly different inNMOSD in comparison toMS (B[SE] = 0007 [0004] mm2 p = 0035 and B [SE] = 0018[0010]mm p = 0039 respectively) neither dependent onON(p = 0254 and 0184) nor on NMOSD-specific ON (p = 0293and 0382) Differences in inner rim volume were not significantin the confirmatory cohort (diagnosis p = 0125 ON p =0080 NMOSD-specific ON p = 0056) Major slope disklength only showed a significant association with NMOSD-specific ON (diagnosis p = 0155 ON history p = 0370NMOSD-specific ON B [SE] = 0012 [0007] mm p = 0046)

DiscussionUsing a novel foveal morphometry approach we here showthat foveal shape is altered in patients with AQP4-IgGndashseropositive NMOSD Our results further support thatthese changes cannot be explained by neuroaxonal damageresulting from ON alone

Foveal morphometry described a flatter and wider fovea inAQP4-IgGndashseropositive NMOSD both in comparison to MSand HC (figure 2 JndashK) This is characterized by increased pitflat disk area increased average pit flat disk diameter reducedinner rim volume and increased major slope disk length Al-though neuroaxonal damage from ON altered the foveal shapeas well we observed robust changes in these parameters also ineyes never experiencing anON and when correcting for ON orneuroaxonal damage in the statistical models in all eyes

The foveal shape changes in AQP4-IgGndashseropositive NMOSDreported in our study are supported by previous studies whichinvestigated thickness or volume changes as indirect evidencefor foveal shape changes Jeong et al16 and Oertel et al17

showed a significant reduction in FT in eyes of patients withAQP4-IgGndashseropositive NMOSD independent of ON incomparison to HCs

Table 4 Linear regression results for the selected foveal shape analysis parameters corrected for GCIPL and INL

Linear regression with interaction effects of diagnosis and ON history corrected for GCIPL

MS vs NMOSD ON2 vs ON+ GCIPL (mm3) NMOSD times ON

R2Marg R2

CondB (SE) p B (SE) p B (SE) p B (SE) p

Pit flat disk area (mm2) 0010(0005)

0031 minus32eminus4

(0002)0886 minus0016

(0006)0010 0012

(0004)0010 0212 0858

Average pit flat disk diameter(mm)

0030(0011)

0006 minus83eminus5

(0005)0986 minus0038

(0013)0006 0023

(0009)0019 0230 0889

Inner rim volume (mm3) minus0011(0004)

0013 minus77eminus4

(0001)0554 0010

(0004)0013 minus0006

(0003)0038 0159 0945

Major slope disk length (mm) 0017(0008)

0051 0002 (0002) 0360 55eminus5

(0006)0992 0010

(0004)0033 0082 0962

Linear regression with interaction effects of diagnosis and ON history corrected for INL

MS vs NMOSD ON2 vs ON+ INL (mm3) NMOSD times ON

R2Marg R2

CondB (SE) p B (SE) p B (SE) p B (SE) p

Pit flat disk area (mm2) 0012(0005)

0017 0001(0002)

0771 0041(0024)

0145 0013(0004)

0014 0193 0851

Average pit flat disk diameter(mm)

0035(0011)

0005 0003(0005)

0482 0075(0054)

0200 0027(0010)

0013 0208 0882

Inner rim volume (mm3) minus0012(0004)

0009 minus0002(0001)

0212 0003(0018)

0860 minus0008(0003)

0009 0143 0939

Major slope disk length (mm) 0016(0008)

0061 0002(0002)

0312 minus0014(0028)

0623 0010(0004)

0022 0083 0962

Abbreviations B = estimate GCIPL = combined macular ganglion cell and inner plexiform layer volume INL = inner nuclear layer volume MS = patients withMS NMOSD = patients with neuromyelitis optica spectrum disorders NMOSD times ON = interaction effect of diagnosis and ON ON = optic neuritis ONminus = eyeswithout a history of ON ON+ = eyes with a history of ON SE = standard error of B R2

Cond = conditional R-squared R2Marg = marginal R-squared

Significant p values are marked in bold

8 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

Apathophysiologic explanation for the observed changes could bethe presence of a primary retinopathy in AQP4-IgGndashseropositiveNMOSD mediated by AQP4-IgG The principal glial cell of theretina is the Muller cell expresses AQP4 and is enriched aroundthe fovea25 Muller cell bodies reside in the INL but processstretch through the whole thickness of the retina linking retinalneurons and photoreceptors with blood vessels Importantly an-imal studies have shown complement-independent AQP4 loss inMuller cells in rats induced by AQP4-IgG which is in line with anAQP4-IgGndashmediated primary retinopathy in NMOSD2627

AQP4-IgGndashmediated primary retinal astrocytopathy has beenalso suggested in human by AQP4-IgGndashseropositive NMOSDautopsy cases28 AQP4 is expressed in Muller cell end feetmdashanalogous to astrocytic end feetmdashat the blood-retina barrier29

For AQP4-IgG circulating in serum to reach its antigen the blood-retina barrier presumably needs to be disrupted Blood-retina or-brain barrier disruptions are typically associated with an acuteinflammatory event and then conceptionally linked to an acuteattack involving complement30 It is unclear whether or to whichextent blood-retinabrain barrier disruptions occur in NMOSDand other diseases that do not lead to full attack cascades A recentanalysis of the NMOSD momentum trial data31 revealed thatelevated glial fibrillary acidic protein levels in serum were associ-ated with an increased attack risk independently suggesting thatthere is indeed subclinical astrocyte damage outside attacks (An-nual European Committee for Treatment and Research in Mul-tiple Sclerosis [ECTRIMS] 2019 P1609)32 A recent study couldfurther show that in rats blood-brain barrier breakdown is notnecessary for NMOSD pathology but that NMOSD-like diseasecan be caused by AQP4-IgG circulating in CSF33 We recentlyreported progressive GCIPL loss without ON in a longitudinalstudy investigating an overlapping cohort34 Further evidence fora primary retinopathy comes from Tian et al35 reporting innerretinal layer thinning independent from ON Significant changesin vascularization of the fovea were also shown in patients withAQP4-IgGndashseropositive NMOSD in comparison to HCs usingOCT angiography36ndash38

Alternatively foveal changes could be caused by subclinicalON Occasionally studies have reported neuroaxonal damagealso in eyes without prior ON in AQP4-IgGndashseropositiveNMOSD which could be interpreted as evidence as such39

However earlier studies had cohort heterogeneity due toincomplete antibody characterization or inclusion of patientswith antibody-negative NMOSD Furthermore ON inNMOSD often occurs near the chiasm and neuroaxonaldamage can be caused by chiasmal crossover from an affectedeye to the fellow eye40 which might not be clinically apparentIn this study we found evidence of neuroaxonal damage alsoin eyes reported to have never experienced ON which is inagreement with our recent findings in a multicenter study34 Itis possible that our data set also included fellow eyes that wereaffected from cross-chiasmal affects during a contralateral ONThe number of patients with NMOSD only experiencingtransverse myelitis and no ON was too small which is why werefrained from analyzing eyes of these patients separately Inconsequence we cannot fully determine to which extent the

observed changes may be caused or affected by covert ON andneuroaxonal damage of the optic nerve

Themain limitationof our studywas the low sample size forAQP4-IgGndashseropositiveNMOSD especially for patients without a historyof ON which is unfortunately common in studies investigatingNMOSD Another limitation of this study is that the diagnosticvalue is unclear as we only used scans from 1 OCT device usinga single scanning protocol It is unclear how scans from differentOCT devices and scanning protocols can be compared

Foveal morphometry may potentially be useful for differentialdiagnosis of AQP4-IgGndashseropositive NMOSD TypicallypRNFL and GCIPL as well as other OCT parameters associ-ated with neuroaxonal damage are mostly nonspecific to theunderlying ON etiology In contrast many foveal morphom-etry parameters showed significant differences betweenpatients with NMOSD and MS in this study Parameter se-lection resulted in 4 promising parameters describing fovealdifferences pit flat disk area average pit flat disk diametermajor slope disk length and inner rim volume Only the first 2parameters could be confirmed in an independent cohort Thereason may be the different frequency of ON in the confir-matory cohort between patients with MS and NMOSD whichexemplifies the need for additional confirmation especially ineyes that are inconspicuous in regard to neuroaxonal damagefrom ON Future work should further investigate this bycomparing patients with myelin oligodendrocyte glycoprotein(MOG)-IgGndashseropositive disease against patients withMOGAQP4-IgG double-negative NMOSD as well as clarify effectsof scan protocols and foveal variability in healthy persons

Study fundingSupported by the Einstein Foundation Berlin (Einstein JuniorScholarship to SM) the German Federal Ministry of Eco-nomic Affairs and Energy (BMWI EXIST 03EFEBE079 toAUB and EMK) German Research Foundation (DFGExc257 to FP and AUB) German Federal Ministry of Educa-tion and Research (BMBF Neu2 ADVISIMS to FP andAUB as well as part of the ldquoGerman Competence NetworkMultiple Sclerosisrdquo (KKNMS) project NationNMO01GI1602B toOA) andNovartis (research grant to HGZ)

DisclosureEM Kadas SK Yadav AU Brandt S Motamedi and FPaul are named as coinventors on the patent application forthe foveal shape analysis method used by this manuscript(ldquoMethod for estimating shape parameters of the fovea byoptical coherence tomographyrdquo International PublicationNumber ldquoWO2019016319 A1rdquo) EM Kadas SK Yadav FPaul and AU Brandt are cofounders and hold shares intechnology start-up Nocturne GmbH which has commercialinterest in OCT applications in neurology EM Kadas andSK Yadav are now employees of Nocturne GmbH HGZimmermann received a research grant from Novartis Allother authors report no relevant disclosures Go to Neurol-ogyorgNN for full disclosures

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 9

Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationJanuary 30 2020 Accepted in final form April 16 2020

References1 Jarius S Paul F Franciotta D et al Mechanisms of disease aquaporin-4 antibodies in

neuromyelitis optica Nat Clin Pract Neurol 20084202ndash2142 Paul F Jarius S Aktas O et al Antibody to aquaporin 4 in the diagnosis of neuro-

myelitis optica PLoS Med 20074e1333 Zekeridou A Lennon VA Aquaporin-4 autoimmunity Neurol Neuroimmunol

Neuroinflamm 20152e110 doi 101212NXI00000000000001104 Jarius S Wildemann B Paul F Neuromyelitis optica clinical features immunopatho-

genesis and treatment neuromyelitis optica Clin Exp Immunol 2014176149ndash1645 Schmidt F Zimmermann H Mikolajczak J et al Severe structural and functional

visual system damage leads to profound loss of vision-related quality of life inpatients with neuromyelitis optica spectrum disorders Mult Scler Relat Disord20171145ndash50

6 Schneider E Zimmermann H Oberwahrenbrock T et al Optical coherence to-mography reveals distinct patterns of retinal damage in neuromyelitis optica andmultiple sclerosis PLoS One 20138e66151

7 Wingerchuk DM Banwell B Bennett JL et al International consensus diagnosticcriteria for neuromyelitis optica spectrum disorders Neurology 201585177ndash189

8 Yamamura T Nakashima I Foveal thinning in neuromyelitis optica a sign of retinalastrocytopathy Neurol Neuroimmunol Neuroinflamm 20174e347 doi 101212NXI0000000000000347

9 Oertel FC Zimmermann H Paul F Brandt AU Optical coherence tomography inneuromyelitis optica spectrum disorders potential advantages for individualizedmonitoring of progression and therapy EPMA J 2018921ndash33

10 Bennett JL de Seze J Lana-Peixoto M et al Neuromyelitis optica and multiplesclerosis seeing differences through optical coherence tomography Mult SclerHoundmills Basingstoke Engl 201521678ndash688

11 Oertel FC Zimmermann H Mikolajczak J et al Contribution of blood vessels toretinal nerve fiber layer thickness in NMOSD Neurol Neuroimmunol Neuroinflamm20174e338 doi 101212NXI0000000000000338

12 Oberwahrenbrock T Traber GL Lukas S et al Multicenter reliability of semi-automatic retinal layer segmentation using OCT Neurol Neuroimmunol Neuro-inflamm 20185e449 doi 101212NXI0000000000000449

13 Kaufhold F ZimmermannH Schneider E et al Optic neuritis is associated with innernuclear layer thickening and microcystic macular edema independently of multiplesclerosis PLoS One 20138e71145

14 Syc SB Saidha S Newsome SD et al Optical coherence tomography segmentationreveals ganglion cell layer pathology after optic neuritis Brain 2012135521ndash533

15 Pache F Zimmermann H Mikolajczak J et al MOG-IgG in NMO and relateddisorders a multicenter study of 50 patients Part 4 afferent visual system damageafter optic neuritis in MOG-IgG-seropositive versus AQP4-IgG-seropositive patientsJ Neuroinflammation 201613282

16 Jeong IH KimHJ KimNH Jeong KS Park CY Subclinical primary retinal pathologyin neuromyelitis optica spectrum disorder J Neurol 20162631343ndash1348

17 Oertel FC Kuchling J Zimmermann H et al Microstructural visual system changes inAQP4-antibodyndashseropositive NMOSD Neurol Neuroimmunol Neuroinflamm 20174e334 doi 101212NXI0000000000000334

18 Yadav SK Motamedi S Oberwahrenbrock T et al CuBe parametric modeling of 3Dfoveal shape using cubic Bezier Biomed Opt Express 201784181

19 Tewarie P Balk L Costello F et al TheOSCAR-IB consensus criteria for retinal OCTquality assessment PLoS One 20127e34823

20 Polman CH Reingold SC Banwell B et al Diagnostic criteria for multiple sclerosis2010 revisions to the McDonald criteria Ann Neurol 201169292ndash302

21 Schippling S Balk LJ Costello F et al Quality control for retinal OCT in multiplesclerosis validation of the OSCAR-IB criteria Mult Scler Houndmills BasingstokeEngl 201521163ndash170

22 Cruz-Herranz A Balk LJ Oberwahrenbrock T et al The APOSTEL recom-mendations for reporting quantitative optical coherence tomography studies Neu-rology 2016862303ndash2309

23 R Core Team R A Language and Environment for Statistical Computing [online]Vienna R Foundation for Statistical Computing 2018 Available at wwwR-projectorgAccessed June 13 2019

24 Ratchford JN Quigg ME Conger A et al Optical coherence tomography helps differ-entiate neuromyelitis optica and MS optic neuropathies Neurology 200973302ndash308

25 Bringmann A Pannicke T Grosche J et al Muller cells in the healthy and diseasedretina Prog Retin Eye Res 200625397ndash424

26 Felix CM Levin MH Verkman AS Complement-independent retinal pathologyproduced by intravitreal injection of neuromyelitis optica immunoglobulin GJ Neuroinflammation 201613275

27 Zeka B Hastermann M Kaufmann N et al Aquaporin 4-specific T cells and NMO-IgG cause primary retinal damage in experimental NMOSD Acta NeuropatholCommun 2016482

28 Hokari M Yokoseki A Arakawa M et al Clinicopathological features in anteriorvisual pathway in neuromyelitis optica Ann Neurol 201679605ndash624

29 Goodyear MJ Crewther SG Junghans BM A role for aquaporin-4 in fluid regulationin the inner retina Vis Neurosci 200926159ndash165

30 Takeshita Y Obermeier B Cotleur AC et al Effects of neuromyelitis optica-IgG at theblood-brain barrier in vitro Neurol Neuroimmunol Neuroinflamm 20174e311 doi101212NXI0000000000000311

31 Cree BAC Bennett JL Kim HJ et al Inebilizumab for the treatment of neuromyelitisoptica spectrum disorder (N-MOmentum) a double-blind randomised placebo-controlled phase 23 trial Lancet Lond Engl 20193941352ndash1363

32 ECTRIMS 2019mdashlate breaking news abstracts Mult Scler J 201925890ndash938

Appendix Authors

Name Location Contribution

SeyedamirhoseinMotamedi MSc

CharitemdashUniversitatsmedizinBerlin Germany

Collected the data conductedthe statistical analysiscontributed to development ofthe foveal shape analysismethod and drafted themanuscript for intellectualcontent

Frederike COertel MD

CharitemdashUniversitatsmedizinBerlin Germany

Collected the data performedOCT quality control andsegmentation contributed todata interpretation andrevised the manuscript forintellectual content

Sunil K YadavPhD

CharitemdashUniversitatsmedizinBerlin Germany

Developed the foveal shapeanalysis method and revisedthe manuscript forintellectual content

EllaM Kadas PhD CharitemdashUniversitatsmedizinBerlin Germany

Contributed to developmentof the foveal shape analysismethod and revised themanuscript for intellectualcontent

Margit WeiseMSc

Heinrich Heine UniversityDusseldorf Germany

Major role in acquisition ofthe confirmatory data andrevised the manuscript forintellectual content

Joachim HavlaMD

Ludwig-MaximiliansUniversity Munich Germany

Contributed to datainterpretation and revised themanuscript for intellectualcontent

MariusRingelstein MD

Heinrich Heine UniversityDusseldorf Germany

Major role in acquisition ofthe confirmatory data andrevised the manuscript forintellectual content

Orhan Aktas MD Heinrich Heine UniversityDusseldorf Germany

Major role in acquisition ofthe confirmatory data andrevised the manuscript forintellectual content

Philipp AlbrechtMD

Heinrich Heine UniversityDusseldorf Germany

Major role in acquisition ofthe confirmatory data andrevised the manuscript forintellectual content

KlemensRuprecht MD

CharitemdashUniversitatsmedizinBerlin Germany

Contributed to studymanagement and revised themanuscript for intellectualcontent

Judith Bellmann-Strobl MD

CharitemdashUniversitatsmedizinBerlin Germany

Contributed to studymanagement and revised themanuscript for intellectualcontent

Hanna GZimmermannPhD

CharitemdashUniversitatsmedizinBerlin Germany

Contributed to datainterpretation and revised themanuscript for intellectualcontent

Friedemann PaulMD

CharitemdashUniversitatsmedizinBerlin Germany

Contributed to studymanagement contributed todata interpretation andrevised the manuscript forintellectual content

Alexander UBrandt MD

CharitemdashUniversitatsmedizinBerlin Germany

Designed and conceptualizedthe study supervised thestatistical analysis anddrafted the manuscript forintellectual content

10 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 NeurologyorgNN

33 Hillebrand S Schanda K Nigritinou M et al Circulating AQP4-specific auto-antibodies alone can induce neuromyelitis optica spectrum disorder in the rat ActaNeuropathol (Berl) 2019137467ndash485

34 Oertel FC Havla J Roca-Fernandez A et al Retinal ganglion cell loss in neuro-myelitis optica a longitudinal study J Neurol Neurosurg Psychiatry 2018891259ndash1265

35 Tian DC Su L Fan M et al Bidirectional degeneration in the visual pathway inneuromyelitis optica spectrum disorder (NMOSD) Mult Scler J 2018241585ndash1593

36 Huang Y Zhou L ZhangBao J et al Peripapillary and parafoveal vascular networkassessment by optical coherence tomography angiography in aquaporin-4 antibody-positive neuromyelitis optica spectrum disorders Br J Ophthalmol 2019103789ndash796

37 Kwapong WR Peng C He Z Zhuang X Shen M Lu F Altered macular microvas-culature in neuromyelitis optica spectrum disorders Am J Ophthalmol 201819247ndash55

38 Green AJ Cree BAC Distinctive retinal nerve fibre layer and vascular changes inneuromyelitis optica following optic neuritis J Neurol Neurosurg Psychiatry 2009801002ndash1005

39 Ringelstein M Harmel J Zimmermann H et al Longitudinal optic neuritis-unrelatedvisual evoked potential changes in NMO spectrum disorders Neurology 202094e407ndashe418

40 Ramanathan S Prelog K Barnes EH et al Radiological differentiation of optic neuritiswith myelin oligodendrocyte glycoprotein antibodies aquaporin-4 antibodies andmultiple sclerosis Mult Scler Houndmills Basingstoke Engl 201622470ndash482

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 5 | September 2020 11

DOI 101212NXI000000000000080520207 Neurol Neuroimmunol Neuroinflamm

Seyedamirhosein Motamedi Frederike C Oertel Sunil K Yadav et al seropositive neuromyelitis optica spectrum disordersminusAltered fovea in AQP4-IgG

This information is current as of June 23 2020

ServicesUpdated Information amp

httpnnneurologyorgcontent75e805fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent75e805fullhtmlref-list-1

This article cites 39 articles 9 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectionretinaRetina

httpnnneurologyorgcgicollectiondevics_syndromeDevics syndromefollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm