Гиперотражающая сетчатка связана с дегенерацией сетчатки после оптического неврита при расстройствах спектра нейромиелита и множественном склерозе.

INTRODUCTION

Multiple sclerosis (MS) and neuromyelitis optica spectrum disorders (NMOSD) are autoimmune inflammatory diseases of the central nervous system (CNS) characterized by recurrent inflammation, demyelination, and neurodegeneration [ 1 , 2 ]. MS is one of the most common neuroinflammatory diseases affecting more than 2.8 million people globally [ 3 ]. The majority of people with MS (pwMS) exhibit a relapsing disease course, which is defined by recurrent attacks [ 1 ]. If a patient has a first attack but does not fulfill the diagnostic criteria for MS, it is considered a clinically isolated syndrome (CIS) [ 4 ]. Up to 70% of people with CIS (pwCIS) will subsequently be diagnosed with MS. [ 5 ] In contrast to MS, NMOSD is a rare disease with a prevalence of 4–13/100,000 people [ 6 ]. Most people with NMOSD (pwNMOSD) have antibodies against aquaporin‐4 (AQP4‐IgG), an astrocytic water channel, suggesting an astrocytopathy as part of the disease pathology [ 7 ]. Both diseases share several clinical characteristics including the common manifestation of optic neuritis (ON). ON describes an inflammation of the optic nerves that subsequently leads to demyelination, neurodegeneration, and severe vision loss [ 8 ].

Microglia, the resident immune cells of the CNS, are key players in MS and NMOSD pathology. During neuroinflammation, microglia recruit lymphocytes, trigger T‐cell activation, and release inflammatory cytokines [ 9 ]. Yet, microglia are also involved in maintaining CNS homeostasis and removing apoptotic cells [ 9 ]. In MS, widespread activation of microglia in the CNS including on the rim of chronic active lesions is observed [ 10 ]. Thus, they are often considered drivers of chronic disease progression in MS. [ 10 ] In NMOSD, microglial activation was also shown in early pathology, but it remains unclear to which extent they fuel ongoing disease activity [ 11 ]. In murine NMOSD models, microglia seem to be necessary for NMOSD pathogenesis [ 12 ]. Despite the distinct pathology of MS and NMOSD, microglia might thus be promising therapeutic targets. Nevertheless, quantifying microglia and microglial activation in vivo remains challenging. Imaging modalities such as 18‐kDa translocator protein (TSPO) positron emission tomography (PET) have been proposed for the analysis of microglia. Yet, TSPO‐PET is still not widely available and poses an additional radiation burden [ 13 ]. Non‐invasive imaging techniques for the quantification of microglia and microglial activation are therefore highly warranted.

A potential alternative for measuring microglial activity might be retinal optical coherence tomography (OCT). OCT is currently used in clinical practice to non‐invasively measure retinal structural parameters such as peripapillary retinal nerve fiber layer (pRNFL), combined ganglion cell and inner plexiform layer (GCIPL), or inner nuclear layer (INL) — particularly for the diagnosis of acute and past ON [ 14 , 15 , 16 ]. Several studies have validated pRNFL and GCIPL as markers of retinal neurodegeneration as well as mirroring brain atrophy and predicting future disease activity [ 17 , 18 ]. INL was suggested as a marker of inflammatory disease activity and therapy response [ 15 ]. To quantify microglia and microglial activity, the novel OCT parameter hyperreflective retinal foci (HRF) have become of increasing interest. HRF are present in healthy people in various degrees but elevated numbers of HRF have been described in several ophthalmologic and neurological diseases including diabetic retinopathy [ 19 ], Fabry disease [ 20 ], and age‐related macular degeneration (AMD) [ 21 ] as well as recently in MS. [ 22 ] Recent pathology data of HRF in AMD indicate HRF in the outer retina to mainly consist of melanolipofuscin‐laden mononuclear phagocytes and differentiated microglia [ 23 ]. In a mice model of diabetic retinopathy, an increased microglial activity was also detected in the area of HRF [ 24 ]. Serological evidence of elevated proinflammatory cytokines in the cerebrospinal fluid of pwMS in relation to HRF numbers in the inner retina indicates them also as a marker of activated microglia in MS. [ 21 , 25 ] Although promising, the definitive etiology of HRF has not been proven.

Previous research on HRF in CNS inflammatory diseases focused solely on pwMS without a clinical history of ON. To expand the clinical usability of HRF, we here 1) compared the frequency of HRF between eyes of pwCIS, pwMS, and pwNMOSD with previous ON (ON + eyes) and without a history of ON (ON − eyes) and 2) investigated the association of HRF with markers of retinal neurodegeneration.

METHODS

Study population

Data for this retrospective analysis were derived from three completed cohort studies between 2011 and 2021 at the Neuroscience Clinical Research Center (NCRC) at Charité—Universitätsmedizin Berlin, Germany. Adult subjects with CIS or early MS that experienced an ON as a first clinical manifestation were included from the CIS‐cohort (EA1/182/10). Inclusion criteria consisted of diagnosis of CIS within six months of study inclusion or diagnosis of MS within two years according to the 2017 McDonald criteria [ 4 ]. Exclusion criteria were a diagnosis of secondary progressive MS and any ophthalmologic disease that might interfere with OCT. Subjects with a minimum age of 18 years, a diagnosis of AQP4‐IgG‐positive NMOSD according to the 2015 International Consensus diagnostic criteria for NMOSD and a medical history of ON were included from the NMO‐cohort study (EA1/041/14). AQP4‐IgG seropositivity was tested by fixed cell‐based assay (Euroimmun, Lübeck, Germany). Healthy controls (HC) were enrolled from the VIMS study (EA1/163/12) and were free of any ophthalmologic or neurologic disease. HCs were age‐ and sex‐matched to enrolled subjects with NMOSD. Collected characteristics included sex, age, expanded disability severity scale (EDSS) score, disease duration, number of previous ON, and disease‐modifying therapies (DMT). In accordance with Wiendl et al. [ 26 ] low‐efficacy DMT consisted of interferon, glatiramer acetate, dimethyl fumarates, azathioprine, and mycophenolate mofetil. High‐efficacy treatment consisted of rituximab, tocilizumab, and belimumab.

Optical coherence tomography

All retinal OCT examinations were performed using Heidelberg Engineering Spectralis spectral domain OCT (Heidelberg Engineering, Heidelberg, Germany) with automatic real‐time (ART) image averaging without pupillary dilatation and were quality assessed according to OSCAR‐IB criteria [ 27 , 28 ]. The pRNFL thickness was calculated with activated eye tracker using 3.4‐mm ring scans around the optic nerve head (12°, 1536 A‐scans 16 ≤ ART ≤100). The GCIPL and INL thickness was measured in a 6‐mm diameter around the fovea based on a macular volume scan (25° × 30°, 61 B‐scans with 768 A‐scans per B‐scan, ART = 15). Segmentation of pRNFL and GCIPL was performed semi‐automatically using Heidelberg Eye Explorer (Eye Explorer 1.9.10.0 with viewing module 6.0.9.0; Heidelberg Engineering) and corrected, if necessary, by an experienced rater (CL). Reported results adhered to the APOSTEL 2.0 recommendations [ 29 , 30 ]. OCT images of study participants were blinded regarding ON status but not for different disease groups.

Hyperreflective retinal foci

Counting of HRF were performed using the central B‐scan crossing the fovea in a 3000 μm diameter around the center by an experienced rater (IV). As previously described, HRF were defined as small (< 30 μm), punctiform elements with a reflectivity similar to the RNFL but without any back shadowing [ 31 ]. HRF were quantified in the GCIPL and INL separately. HRF quantification was confirmed for all scans by a second independent rater (PK). For analyses of HRF, we used the open‐source software OCT‐marker (OCT‐marker, Kay Gawlik, Charité—Universitätsmedizin Berlin, Germany) that enables image modification (i.e., gamma correction, contrast, brightness) to improve visibility of HRF and reduction of background noise (Figure 1 ). The interrater reliability between the two independent raters (IV, PK) was assessed using the intraclass correlation coefficient (ICC). The ICC for quantification of HRF was 0.8 and therefore classified as good. In cases of quantification differences between the two raters, an expert consensus between the two raters was reached during the secondary review.

Statistical methods

All statistical analysis and graphical visualizations were performed using R (version 4.2.2). p ‐values <0.05 were defined as statistically significant. p ‐values <0.07 were defined as statistical trend. Categorical variables are described in absolute numbers and relative percentages, ordinal variables are reported as median and interquartile range (IQR), and continuous variables are presented as mean and standard deviation (SD). Total HRF consisted of the sum of HRF in the GCIPL and INL. Linear mixed models were employed for all analyses regarding pRNFL, GCIPL, and INL thickness and HRF numbers to account for multiple observations in the same patient (left and right eye). First, differences between the disease entities were examined. Second, a comparison of eyes with and without ON within the respective disease group was performed. β and standard errors were reported. Linear mixed models were not applicable when analyzing a potential association between HRF and retinal structural parameters distinctly in ON − eyes and ON + eyes of pwCIS and pwMS as these patients only had a single unilateral ON. In these cases, spearman correlations were applied, and correlation coefficients ( r ) were reported. Due to the exploratory nature of this study, the uncorrected results are being reported. A preliminary adjustment using false discovery rate correction revealed only an impact on associations between HRF numbers and retinal structural parameters in ON + eyes (Tables S1–S3 ).

Ethics approval and consent to participate

All cohort studies were approved by the local ethics committee at Charité—Universitätsmedizin Berlin (EA1/182/10, EA1/041/14 EA1/163/12). The studies were conducted in accordance with the Declaration of Helsinki in its currently applicable version and the applicable European and German laws. All participants provided written informed consent.

RESULTS

Study population

In total, 227 eyes of 123 people, which included 35 (28.5%) HC, 35 (28.5%) pwNMOSD, 32 (26.0%) pwMS, and 21 (17.0%) pwCIS (Table 1 ). Nineteen eyes were excluded due to inadequate macular OCT image quality. Study population characteristics are summarized in Table 1 . PwCIS and pwMS were significantly younger than HC and NMOSD (pwCIS: HC p < 0.001, pwNMOSD p < 0.001; pwMS: HC p < 0.001, pwNMOSD p < 0.001). Further, pwNMOSD were significantly more often female than pwCIS ( p = 0.049) and pwMS. No differences were seen regarding age and sex between pwCIS and pwMS as well as between pwNMOSD and HC.

HRF are a retinal feature in NMOSD , CIS, and MS

Total HRF were not associated with age ( β = −0.02 ± 0.05 [n.s.]) or sex in the whole cohort ( β = −0.66 ± 1.47 [n.s.]) as well as in all subgroups (data not shown). Longer disease duration was associated with less occurrence of total HRF in pwNMOSD ( β = −0.07 ± 0.02 [ p < 0.001]) but not in pwsCIS ( β = 0.81 ± 0.48 [n.s.]) or pwMS ( β = −0.03 ± 0.19 [n.s.]). Yet, a longer time interval from the last ON showed a trend of less total HRF ( β = −0.05 ± 0.03 [ p = 0.064]). When comparing the influence of DMT in pwCIS and pwMS on HRF frequency, patients taking DMT had fewer HRF in GCIPL, even though this was not a significant difference ( β = −1.47 ± 1.06 [n.s.]).

Total HRF counts were higher in pwCIS, pwMS, and pwNMOSD compared to HC (pwNMOSD β = −5.69 ± 1.39 [ p < 0.001], pwCIS β = −8.83 ± 1.09 [ p < 0.001], pwMS β = −5.44 ± 1.07 [ p < 0.001], Figure 2 ). These differences persisted when comparing HRF in GCIPL (pwNMOSD β = −2.55 ± 0.87 [ p = 0.005], pwCIS β = −5.01 ± 0.74 [ p < 0.001], pwMS β = −3.07 ± 0.76 [ p < 0.001]) and INL separately (pwNMOSD β = −3.15 ± 0.68 [ p < 0.001], pwCIS β = −3.80 ± 0.49 [ p < 0.001], pwMS β = −2.36 ± 0.43 [ p < 0.001]).

PwCIS had higher HRF counts in GCIPL ( β = −2.06 ± 1.02 [ p = 0.048]) and INL than pwMS ( β = −1.43 ± 0.57 [ p = 0.015]). PwCIS also had higher HRF counts in GCIPL ( β = −2.58 ± 1.17 [ p = 0.032]) but not in INL compared to pwNMOSD ( β = −0.67 ± 0.89 [n.s.]). No significant differences in HRF count were seen between pwMS and pwNMOSD.

HRF frequency is associated with ON history in pwMS but not in pwCIS or pwNMOSD

ON + eyes of pwMS had a lower number of HRF in the GCIPL but not in the INL (Figure 3 ) than ON − eyes (HRF GCIPL: β = −1.71 ± 0.78 [ p = 0.036]; HRF INL: β = −0.45 ± 0.42 [ p = n.s.]). In pwCIS and pwNMOSD, HRF number did not differ between ON + eyes and ON − eyes in the GCIPL (pwCIS: β = −1.47 ± 1.42 [n.s.]; pwNMOSD: β = −2.05 ± 1.33 [n.s.]) nor in the INL (pwCIS: β = 0.28 ± 0.61 [n.s.]; pwNMOSD: β = 0.41 ± 0.87 [n.s.]; Table 2 ). When comparing ON + eyes between different etiologies, pwCIS had more HRF in the GCIPL but not in INL than pwNMOSD (HRF GCIPL: β = −3.28 ± 1.52 [ p = 0.038]; HRF INL: β = −0.86 ± 1.30 [n.s.]). HRF frequencies did not differ significantly in ON + eyes between other groups. Across diagnoses, a greater time interval from last ON showed a trend of less total HRF and HRF in GCIPL (Total HRF β = −0.05 ± 0.03 [ p = 0.064]; HRF GCIPL β = −0.04 ± 0.02 [ p = 0.056]). Adjusting for GCIPL thickness did not influence this trend (data not shown).

In ON − eyes, HRF in the GCIPL and INL were more frequently detected in patients' eyes than in HC eyes (HRF GCIPL: pwNMOSD β = −3.47 ± 0.90 [ p < 0.001], pwCIS β = −5.69 ± 0.86 [ p < 0.001], pwMS β = −3.77 ± 0.74 [ p < 0.001]; HRF INL: pwNMOSD β = −2.92 ± 0.58 [ p < 0.001], pwCIS β = −3.67 ± 0.54 [ p < 0.001], pwMS β = −2.13 ± 0.47 [ p < 0.001]). Further, ON − eyes of pwCIS had significantly higher numbers of HRF in INL but not in GCIPL compared to ON − eyes of pwMS (HRF GCIPL: t = −1.66 [ p = 0.105]; HRF INL: t = −2.26 [ p = 0.029]). A pairwise comparison of ON − eyes between diseases did not yield any significant associations.

Correlation between HRF and retinal structural parameters

Across subjects with preexisting demyelinating disease, ON + eyes had a thinner pRNFL and GCIPL compared to ON − eyes (pRNFL: pwMS β = −12.40 ± 2.95 [ p < 0.001], pwCIS: β = −6.88 ± 2.20 [ p = 0.005], pwNMOSD: β = −26.21 ± 4.55 [ p < 0.001]; GCIPL: pwMS β = −8.89 ± 1.47 [ p < 0.001], pwCIS: β = −6.61 ± 1.24 [ p < 0.001], pwNMOSD: β = −8.07 ± 1.90 [ p < 0.001]).

In pwMS and pwNMOSD, Total HRF were positively associated to pRNFL (pwMS: β = 0.11 ± 0.04 [ p = 0.012]; pwNMOSD β = 0.16 ± 0.05 [ p = 0.004]) and GCIPL thickness (pwMS: β = 0.17 ± 0.07 [ p = 0.026]; pwNMOSD β = 0.35 ± 0.13 [ p = 0.010]) but not in pwCIS (pRNFL: β = −0.01 ± 0.09 [n.s.]; GCIPL: β = 0.06 ± 0.16 (n.s.)). Furthermore, in pwMS and pwNMOSD number of HRF in GCIPL were positively associated to GCIPL thickness (pwMS: β = 0.15 ± 0.06 [ p = 0.015]; pwNMOSD: β = 0.22 ± 0.08 [ p = 0.006]). These associations persisted in ON + eyes of pwMS and pwNMOSD but not in ON − eyes (ON + eyes: pwMS r = 0.38 [ p = 0.040]; pwNMOSD β = 0.26 ± 0.11 [ p = 0.031]; ON − eyes: pwMS r = 0.17 [n.s.]; pwNMOSD β = 0.10 ± 0.14 [n.s.]). There was no significant association for a number of Total HRF or HRF in INL with INL thickness (Figure 3 ).

DISCUSSION

This study shows higher numbers of HRF in the GCIPL and INL in pwCIS, pwMS, and pwNMOSD than in HC with pwCIS presenting with the highest HRF numbers. Further, in pwMS and pwNMOSD, HRF in GCIPL were positively associated with GCIPL thickness in ON + eyes. This confirms and extends prior cross‐sectional and longitudinal data comparing HRF frequencies in pwMS and HC [ 22 , 32 ].

Microglia play an active role during acute inflammation and phagocytosis of damaged tissue during the post‐acute stage. We and others showed in animal models of neuroinflammation such as experimental autoimmune encephalomyelitis (EAE), that the number of retinal microglia increases specifically during these phases and decreases thereafter [ 33 ]. A prior experimental study also suggested that HRF might be a surrogate marker of localized blood‐retina barrier (BRB) disruption by activated microglia [ 24 ]. This is in line with our results showing that the time after ON is inversely correlated with the total HRF count.

On the contrary, our data also reveal a positive correlation between neuroaxonal layer thickness (pRNFL, GCIPL) and HRF count in pwMS and pwNMOSD. Retinal neurodegeneration after ON is particularly characterized by apoptosis of retinal ganglion cells (RGC) in the GCIPL [ 34 ]. Histologic studies in EAE mice also showed that the early retrograde degeneration of RGC is mediated by microglia and enabled by the breakdown of the BRB [ 35 , 36 ]. Yet, beyond the acute phase of ON when RGC loss already happened, BRB disruption and inflammation potentially decrease. As our cohort consisted predominantly of patients with ON in the chronic stage, a relevant GCIPL atrophy had already occurred. A histopathological study of post‐mortem retinal tissues of pwMS revealed an increased presence of retinal microglia in GCIPL and INL [ 37 ]. Further, retinal microglia were seen gathered around degenerated ganglion cells, particularly in the GCIPL [ 37 ]. While longitudinal studies will be necessary to fully understand the HRF dynamics, one could speculate that the formation of HRF is additionally dependent on the presence of RGCs as a substrate. This would also explain the higher numbers of HRF in ON − eyes compared to ON + eyes. Additionally, the dependence of HRF on RGC could also account for the missing association between HRF and INL thickness. INL thickening might reflect inflammatory activity in pwMS and is potentially caused by dynamic fluid shifts and Müller cell dysfunction [ 38 ] which are likely not primarily due to retinal microglia. However, due to the still not fully comprehended nature of HRF, all interpretations of the current results remain speculative. Further histopathological and in vivo examinations are therefore warranted.

CIS and early MS are considered to have similar pathologic mechanisms [ 1 ]. Both stages of the disease are characterized by active demyelinating lesions in the central nervous system with lymphocyte and microglia infiltrations [ 39 ]. After the 2017 iterations of the McDonald criteria, the amount of patients diagnosed with MS after a first demyelinating attack increased substantially [ 40 ]. As the diagnosis of CIS and MS comes closer, previous studies failed to find significant differences in immunological profiles between both diseases [ 41 ]. Therefore, we do not believe that differences in the number of HRF between CIS and MS are the result of alternating pathophysiologic mechanisms. Further, we ruled out an influence of age, sex, or disease duration as possible causes for differences in HRF numbers in pwCIS and pwMS. A potential explanation for the elevated number of HRF in pwCIS might be that the occurrence is influenced by immunomodulating medications. Even though none of the currently available disease‐modifying drugs for MS are directly targeting microglia, the indirect effects of these treatments are likely [ 42 ]. Additionally, the use of DMT was shown to slow down retinal neurodegeneration of INL and GCIPL in pwMS [ 43 ]. In our cohort, pwMS were only treated with low‐efficacy treatment due to the availability of treatments at the time of study inclusion and the short disease duration. Treated patients had numerically but not statistically fewer numbers of HRF compared to treatment‐naïve patients. Future prospective data will be necessary to confirm this effect. Further, the more pronounced retinal neurodegeneration in pwMS compared to pwCIS could be an additional confounder on HRF formation.

Strengths of the current study include the structured analysis of HRF using a software that was specifically developed for OCT evaluations, the exclusion of other concomitant eye diseases, and the correction of retinal segmentation by experienced raters. However, several limitations need to be discussed: We only performed a cross‐sectional analysis of OCT images. As HRF can change over time, longitudinal analysis might offer novel insights into the origin of HRF in neuroinflammatory diseases by evaluating possible changes in HRF numbers during relapses. Larger prospective trials are warranted to further assess the value of HRF as a prediction marker for future relapses. Also, OCT images of the acute phase of ON were not available. Experimental studies in EAE mice showed especially in the early stage of acute ON a pronounced increase in microglial activity [ 33 ]. Analysis of HRF in patients with acute ON might be an opportunity to reproduce these experimental results in humans. In addition, the comparability between pwNMOSD with pwCIS and pwMS is limited as pwNMOSD had a notably longer disease duration. Regarding pwMS, we only included patients with a relapsing–remitting disease course. As microglial activation is believed to be a key driver of clinical progression [ 42 ], future studies should evaluate HRF as a potential marker for progression in pwMS. Despite using a software specialized in the analysis of OCT images, an automated quantification of HRF would increase comparability and reduce the risk of human‐made errors. So far, no validated algorithm for automated segmentation of HRF is available even though some studies have attempted to establish one in diabetic retinopathy [ 44 , 45 ]. Lastly, due to the structure of our data, blinding of the OCT images could only be done for ON status. We tried to reduce a potential rater bias by adhering to our protocol and ensuring quality control through analyzing by two separate raters. Still, in patients with severe ON or multiple previous ipsilateral ON, the raters could have detected a potential ON due to severe RNFL or GCIPL atrophy.

In conclusion, we demonstrated a higher number of HRF in INL and GCIPL in demyelinating diseases compared to HC. Due to the strong correlation of HRF to GCIPL thickness, HRF formation seems to be affected by RGC count. More advanced histopathologic studies are required to confirm that HRF represents activated microglia. Future clinical studies should also focus on establishing automated quantification software for HRF and subsequently evaluate the potential predictive utility of HRF, for example, for reconstitution of visual acuity after ON or disease activity independent of attacks.

AUTHOR CONTRIBUTIONS

Philipp Klyscz: Conceptualization; methodology; writing – original draft; writing – review and editing; formal analysis; investigation; data curation; validation; visualization. Ifat Vigiser: Investigation; methodology; writing – review and editing. Gilberto Solorza Buenrostro: Software; writing – review and editing; data curation. Seyedamirhosein Motamedi: Writing – review and editing; software; data curation. Carla Johanna Leutloff: Writing – review and editing; data curation. Patrick Schindler: Writing – review and editing; investigation. Tanja Schmitz‐Hübsch: Funding acquisition; writing – review and editing; resources. Friedemann Paul: Resources; writing – review and editing; funding acquisition. Hanna Gwendolyn Zimmermann: Conceptualization; writing – review and editing; methodology. Frederike Cosima Oertel: Conceptualization; supervision; visualization; writing – review and editing; project administration; resources; funding acquisition.

CONFLICT OF INTEREST STATEMENT

PKL, IV, GSB, SM, and CJL report no conflicts of interest. PS received travel support by UCB, received speaker‘s honoraria by Roche and Alexion, and served on an advisory board by Alexion. TSH reports speaker's honoraria from Honoraria AbbVie, Bayer, Roche and research grants bms/Celgene, and Roche. FP has received honoraria and research support from Alexion, Bayer, Biogen, Chugai, MerckSerono, Novartis, Genyzme, MedImmune, Shire, and Teva Pharmaceuticals, and serves on scientific advisory boards for Alexion, MedImmune, Novartis, and UCB. He has received funding from Deutsche Forschungsgemeinschaft (DFG Exc 257), Bundesministerium für Bildung und Forschung (Competence Network Multiple Sclerosis), Guthy‐Jackson Charitable Foundation, EU Framework Program 7, and National Multiple Sclerosis Society of the USA. He serves on the steering committee of the N‐Momentum study with inebilizumab (Horizon Therapeutics) and the OCTiMS Study (Novartis). He is an associate editor with Neurology, Neuroimmunology, and Neuroinflammation and academic editor with PloS One. HGZ reports grants and speaking honoraria from Novartis and personal fees from Bayer Healthcare, both unrelated to this project. FCO reports past research funding by the American Academy of Neurology, the National Multiple Sclerosis Society (US), and the German Association of Neurology (DGN). FCO reports current research support by the Hertie Foundation for Excellence in Clinical Neurosciences, by Novartis AG, by the DFG‐TWAS program—both unrelated to this project. She also reports speaker honoraria by UCB.

Supporting information

Data S1.

ACKNOWLEDGEMENTS

The authors would like to thank the administrative office of the Neuroscience Clinical Research Center (NCRC) and the REDCap team. Further, the authors thank all persons who participated in the study.

DATA AVAILABILITY STATEMENT

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

REFERENCES

1 Filippi M , Bar‐Or A , Piehl F , et al. Multiple sclerosis . Nat Rev Dis Primers . 2018 ; 4 : 43 . 30410033 10.1038/s41572-018-0041-4

2 Jarius S , Paul F , Weinshenker BG , Levy M , Kim HJ , Wildemann B . Neuromyelitis optica . Nat Rev Dis Primers . 2020 ; 6 : 85 . 33093467 10.1038/s41572-020-0214-9

3 Walton C , King R , Rechtman L , et al. Rising prevalence of multiple sclerosis worldwide: insights from the atlas of ms, third edition . Mult Scler . 2020 ; 26 : 1816 ‐ 1821 . 33174475 10.1177/1352458520970841 PMC7720355

4 Thompson AJ , Banwell BL , Barkhof F , et al. Diagnosis of multiple sclerosis: 2017 revisions of the mcdonald criteria . Lancet Neurol . 2018 ; 17 : 162 ‐ 173 . 29275977 10.1016/S1474-4422(17)30470-2

5 Allen CM , Mowry E , Tintore M , Evangelou N . Prognostication and contemporary management of clinically isolated syndrome . J Neurol Neurosurg Psychiatry . 2020 ; 92 : 391 ‐ 397 . 10.1136/jnnp-2020-323087 33361410

6 Flanagan EP , Cabre P , Weinshenker BG , et al. Epidemiology of aquaporin‐4 autoimmunity and neuromyelitis optica spectrum . Ann Neurol . 2016 ; 79 : 775 ‐ 783 . 26891082 10.1002/ana.24617 PMC4988933

7 Lennon VA , Wingerchuk DM , Kryzer TJ , et al. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis . Lancet . 2004 ; 364 : 2106 ‐ 2112 . 15589308 10.1016/S0140-6736(04)17551-X

8 Bennett JL , Costello F , Chen JJ , et al. Optic neuritis and autoimmune optic neuropathies: advances in diagnosis and treatment . Lancet Neurol . 2023 ; 22 : 89 ‐ 100 . 36155661 10.1016/S1474-4422(22)00187-9

9 Lloyd AF , Davies CL , Holloway RK , et al. Central nervous system regeneration is driven by microglia necroptosis and repopulation . Nat Neurosci . 2019 ; 22 : 1046 ‐ 1052 . 31182869 10.1038/s41593-019-0418-z PMC6597360

10 Absinta M , Maric D , Gharagozloo M , et al. A lymphocyte‐microglia‐astrocyte axis in chronic active multiple sclerosis . Nature . 2021 ; 597 : 709 ‐ 714 . 34497421 10.1038/s41586-021-03892-7 PMC8719282

11 Lucchinetti CF , Guo Y , Popescu BF , Fujihara K , Itoyama Y , Misu T . The pathology of an autoimmune astrocytopathy: lessons learned from neuromyelitis optica . Brain Pathol . 2014 ; 24 : 83 ‐ 97 . 24345222 10.1111/bpa.12099 PMC3905574

12 Chen T , Bosco DB , Ying Y , Tian DS , Wu LJ . The emerging role of microglia in neuromyelitis optica . Front Immunol . 2021 ; 12 : 616301 . 33679755 10.3389/fimmu.2021.616301 PMC7933531

13 Airas L , Rissanen E , Rinne JO . Imaging neuroinflammation in multiple sclerosis using tspo‐pet . Clin Translat Imaging . 2015 ; 3 : 461 ‐ 473 . 10.1007/s40336-015-0147-6 PMC4887541 27331049

14 Oertel FC , Specovius S , Zimmermann HG , et al. Retinal optical coherence tomography in neuromyelitis optica . Neurol Neuroimmunol Neuroinflamm . 2021 ; 8 : e1068 . 34526385 10.1212/NXI.0000000000001068 PMC8448522

15 Petzold A , Balcer LJ , Calabresi PA , et al. Retinal layer segmentation in multiple sclerosis: a systematic review and meta‐analysis . Lancet Neurol . 2017 ; 16 : 797 ‐ 812 . 28920886 10.1016/S1474-4422(17)30278-8

16 Petzold A , Fraser CL , Abegg M , et al. Diagnosis and classification of optic neuritis . Lancet Neurol . 2022 ; 21 : 1120 ‐ 1134 . 36179757 10.1016/S1474-4422(22)00200-9

17 Zimmermann HG , Knier B , Oberwahrenbrock T , et al. Association of retinal ganglion cell layer thickness with future disease activity in patients with clinically isolated syndrome . JAMA Neurol . 2018 ; 75 : 1071 ‐ 1079 . 29710121 10.1001/jamaneurol.2018.1011 PMC6143115

18 Saidha S , Al‐Louzi O , Ratchford JN , et al. Optical coherence tomography reflects brain atrophy in multiple sclerosis: a four‐year study . Ann Neurol . 2015 ; 78 : 801 ‐ 813 . 26190464 10.1002/ana.24487 PMC4703093

19 Bolz M , Schmidt‐Erfurth U , Deak G , Mylonas G , Kriechbaum K , Scholda C . Optical coherence tomographic hyperreflective foci: a morphologic sign of lipid extravasation in diabetic macular edema . Ophthalmology . 2009 ; 116 : 914 ‐ 920 . 19410950 10.1016/j.ophtha.2008.12.039

20 Atiskova Y , Rassuli R , Koehn AF , et al. Retinal hyperreflective foci in fabry disease . Orphanet J Rare Dis . 2019 ; 14 : 296 . 31878969 10.1186/s13023-019-1267-2 PMC6933914

21 Coscas G , De Benedetto U , Coscas F , et al. Hyperreflective dots: a new spectral‐domain optical coherence tomography entity for follow‐up and prognosis in exudative age‐related macular degeneration . Ophthalmologica . 2013 ; 229 : 32 ‐ 37 . 23006969 10.1159/000342159

22 Pilotto E , Miante S , Torresin T , et al. Hyperreflective foci in the retina of active relapse‐onset multiple sclerosis . Ophthalmology . 2020 ; 127 : 1774 ‐ 1776 . 32359844 10.1016/j.ophtha.2020.03.024

23 Augustin S , Lam M , Lavalette S , et al. Melanophages give rise to hyperreflective foci in amd, a disease‐progression marker . J Neuroinflammation . 2023 ; 20 : 28 . 36755326 10.1186/s12974-023-02699-9 PMC9906876

24 Mat Nor MN , Guo CX , Green CR , Squirrell D , Acosta ML . Hyper‐reflective dots in optical coherence tomography imaging and inflammation markers in diabetic retinopathy . J Anat . 2023 ; 243 : 697 ‐ 705 . 37222261 10.1111/joa.13889 PMC10485584

25 Puthenparampil M , Torresin T , Franciotta S , et al. Hyper‐reflecting foci in multiple sclerosis retina associate with macrophage/microglia‐derived cytokines in cerebrospinal fluid . Front Immunol . 2022 ; 13 : 852183 . 35664007 10.3389/fimmu.2022.852183 PMC9160385

26 Wiendl H , Gold R , Berger T , et al. Multiple sclerosis therapy consensus group (mstcg): position statement on disease‐modifying therapies for multiple sclerosis (white paper) . Ther Adv Neurol Disord . 2021 ; 14 : 17562864211039648 . 34422112 10.1177/17562864211039648 PMC8377320

27 Tewarie P , Balk L , Costello F , et al. The oscar‐ib consensus criteria for retinal oct quality assessment . PLoS One . 2012 ; 7 : e34823 . 22536333 10.1371/journal.pone.0034823 PMC3334941

28 Schippling S , Balk LJ , Costello F , et al. Quality control for retinal oct in multiple sclerosis: validation of the oscar‐ib criteria . Mult Scler . 2015 ; 21 : 163 ‐ 170 . 24948688 10.1177/1352458514538110

29 Cruz‐Herranz A , Balk LJ , Oberwahrenbrock T , et al. The apostel recommendations for reporting quantitative optical coherence tomography studies . Neurology . 2016 ; 86 : 2303 ‐ 2309 . 27225223 10.1212/WNL.0000000000002774 PMC4909557

30 Aytulun A , Cruz‐Herranz A , Aktas O , et al. Apostel 2.0 recommendations for reporting quantitative optical coherence tomography studies . Neurology . 2021 ; 97 : 68 ‐ 79 . 33910937 10.1212/WNL.0000000000012125 PMC8279566

31 Kodjikian L , Bellocq D , Bandello F , et al. First‐line treatment algorithm and guidelines in center‐involving diabetic macular edema . Eur J Ophthalmol . 2019 ; 29 : 573 ‐ 584 . 31238719 10.1177/1120672119857511

32 Pengo M , Miante S , Franciotta S , et al. Retinal hyperreflecting foci associate with cortical pathology in multiple sclerosis . Neurol Neuroimmunol Neuroinflamm . 2022 ; 9 : e1180 . 35606113 10.1212/NXI.0000000000001180 PMC9128002

33 Cruz‐Herranz A , Oertel FC , Kim K , et al. Distinctive waves of innate immune response in the retina in experimental autoimmune encephalomyelitis . JCI Insight . 2021 ; 6 : e149228 . 34100385 10.1172/jci.insight.149228 PMC8262300

34 Horstmann L , Schmid H , Heinen AP , Kurschus FC , Dick HB , Joachim SC . Inflammatory demyelination induces glia alterations and ganglion cell loss in the retina of an experimental autoimmune encephalomyelitis model . J Neuroinflammation . 2013 ; 10 : 120 . 24090415 10.1186/1742-2094-10-120 PMC3851328

35 Manogaran P , Samardzija M , Schad AN , et al. Retinal pathology in experimental optic neuritis is characterized by retrograde degeneration and gliosis . Acta Neuropathol Commun . 2019 ; 7 : 116 . 31315675 10.1186/s40478-019-0768-5 PMC6637505

36 Fairless R , Williams SK , Hoffmann DB , et al. Preclinical retinal neurodegeneration in a model of multiple sclerosis . J Neurosci . 2012 ; 32 : 5585 ‐ 5597 . 22514320 10.1523/JNEUROSCI.5705-11.2012 PMC6703474

37 Green AJ , McQuaid S , Hauser SL , Allen IV , Lyness R . Ocular pathology in multiple sclerosis: retinal atrophy and inflammation irrespective of disease duration . Brain . 2010 ; 133 : 1591 ‐ 1601 . 20410146 10.1093/brain/awq080 PMC2877904

38 Balk LJ , Coric D , Knier B , et al. Retinal inner nuclear layer volume reflects inflammatory disease activity in multiple sclerosis; a longitudinal oct study . Mult Scler J Exp Transl Clin . 2019 ; 5 : 2055217319871582 . 31523449 10.1177/2055217319871582 PMC6728683

39 Frischer JM , Weigand SD , Guo Y , et al. Clinical and pathological insights into the dynamic nature of the white matter multiple sclerosis plaque . Ann Neurol . 2015 ; 78 : 710 ‐ 721 . 26239536 10.1002/ana.24497 PMC4623970

40 van der Vuurst de Vries RM , Mescheriakova JY , Wong YYM , et al. Application of the 2017 revised mcdonald criteria for multiple sclerosis to patients with a typical clinically isolated syndrome . JAMA Neurol . 2018 ; 75 : 1392 ‐ 1398 . 30083703 10.1001/jamaneurol.2018.2160 PMC6248116

41 Camara‐Lemarroy CR , Silva C , Metz LM , et al. Multimodal peripheral fluid biomarker analysis in clinically isolated syndrome and early multiple sclerosis . Mult Scler Relat Disord . 2021 ; 50 : 102809 . 33581614 10.1016/j.msard.2021.102809

42 Healy LM , Stratton JA , Kuhlmann T , Antel J . The role of glial cells in multiple sclerosis disease progression . Nat Rev Neurol . 2022 ; 18 : 237 ‐ 248 . 35190704 10.1038/s41582-022-00624-x

43 Button J , Al‐Louzi O , Lang A , et al. Disease‐modifying therapies modulate retinal atrophy in multiple sclerosis: a retrospective study . Neurology . 2017 ; 88 : 525 ‐ 532 . 28077493 10.1212/WNL.0000000000003582 PMC5304463

44 Yu C , Xie S , Niu S , et al. Hyper‐reflective foci segmentation in sd‐oct retinal images with diabetic retinopathy using deep convolutional neural networks . Med Phys . 2019 ; 46 : 4502 ‐ 4519 . 31315159 10.1002/mp.13728

45 Okuwobi IP , Ji Z , Fan W , Yuan S , Bekalo L , Chen Q . Automated quantification of hyperreflective foci in sd‐oct with diabetic retinopathy . IEEE J Biomed Health Inform . 2020 ; 24 : 1125 ‐ 1136 . 31329137 10.1109/JBHI.2019.2929842