Разоблачение смешения воспаления-дегенерации в ранних мембранах: предварительные исследования из ферритина, нейрогранина, ТРЕМ2 и клеточного слоя сетчатки.

Introduction

The pathophysiology of Multiple Sclerosis (MS) involves a complex interplay between immune-mediated inflammation and neurodegeneration, leading to a cascade of pathological events. Understanding these processes, particularly in the early stages of the disease, is crucial for developing effective diagnostic and therapeutic strategies [ 1 , 2 ].

Recent advances in biomarker research have provided new insights into the molecular underpinnings of MS [ 3 , 4 ]. Among the promising candidates are ferritin, neurogranin, Triggering Receptor Expressed on Myeloid cells 2-TREM2, and neurofilaments light chain (NfL) [ 5 – 9 ]. These biomarkers offer a window into the intricate relationship between inflammation and neurodegeneration in MS.

Ferritin, a protein primarily involved in iron storage and homeostasis, has garnered attention in MS research due to its potential role in neuroinflammation and oxidative stress [ 10 – 12 ]. Elevated levels of ferritin in the cerebrospinal fluid (CSF) have been linked to various neurological disorders, including MS [ 6 , 13 , 14 ]. The presence of ferritin in the central nervous system (CNS) may reflect iron-mediated oxidative damage and microglial activation, both of which contribute to neurodegeneration [ 13 , 14 ].

Neurogranin, a postsynaptic protein crucial for synaptic plasticity and long-term potentiation, has been proposed as a marker of synaptic dysfunction and neurodegeneration [ 15 , 16 ]. While extensively studied in Alzheimer's disease, its role in MS remains less understood [ 15 , 16 ]. Investigating neurogranin levels in MS patients could provide valuable insights into synaptic integrity and cognitive function, shedding light on the synaptic alterations that occur in the disease.

TREM2, an innate immune receptor expressed on microglia and other myeloid cells, plays a pivotal role in regulating neuroinflammation and phagocytosis [ 17 , 18 ]. Human genetics and preclinical studies have tried to identify key contributions of TREM2 to several neurodegenerative conditions and its role in remyelination. [ 17 , 19 , 20 ]. Its potential as a biomarker for microglial activation and neuroinflammation in MS warrants further exploration, as it may offer clues to the immune mechanisms driving disease progression.

Neurofilaments, particularly NfL, have emerged as robust biomarkers of axonal damage in MS. Elevated levels of NfL in the CSF and serum are indicative of ongoing neuroaxonal injury and have been associated with disease activity and progression [ 9 ]. The measurement of NfL levels provides a dynamic marker of neurodegeneration, reflecting the extent of axonal damage in real time [ 9 ].

The optical coherence tomography (OCT) has become a valuable tool for assessing neurodegeneration in MS, being the retina an extension of the CNS, mirrors pathological changes occurring in the brain and spinal cord [ 21 – 23 ]. The retinal ganglion cell layer (GCL) thinning is provided as useful marker of disease disability and brain atrophy with a good sensitivity in monitoring longitudinal changes [ 24 ]. Despite the growing interest in these biomarkers, their interrelationships and collective role at the time of MS diagnosis have not been fully investigated.

This pilot study aims to investigate the levels of ferritin, neurogranin, TREM2, and neurofilaments in the CSF at the time of MS diagnosis. Furthermore, we seek to explore the correlations between these biomarkers and their associations with clinical, cognitive, and OCT parameters.

Methods

Study setting

We conducted a pilot cross-sectional study at the MS Center of Foggia, Italy from June 1, 2023, to March 31, 2024.

Participants

Inclusion criteria were: age 18–55 years; (2) a confirmed Relapsing Multiple Sclerosis (RMS) diagnosis based on the revised McDonald criteria [ 25 ], (3) symptom onset within 24 months from diagnosis confirmation; (4) high disease activity in the year before diagnosis; (5) not being previously exposed to disease-modifying therapies (DMTs).

Additionally, we included patients a comparison group of patients age/sex matched with a 2:1 ratio with RMS defined as symptomatic controls (SCs) according to consensus definition [ 26 ] (Table e 1 ).

Table 1 Baseline characteristics of the enrolled cohort SCs ( n  = 13) RMS ( n  = 26) p value* Female n (%) 9 (69.2) 18 (69.2) ns Age at sampling (year), (mean ± SD) 35.1 ± 9.5 34.8 ± 9.8 ns CSF parameters  Link index 0.3 ± 0.2 1.1 ± 0.4  < .05 Oligoclonal bands  Absence 13 (100) 2 (7.7%)  < .05  Presence 0 24 (92.3)  < .05  N. of oligoclonal bands (mean ± SD) / 5.5 ± 2.3 / RMS patients features  N of relapses in the year before diagnosis (mean ± SD) / 2.6 ± 0.5 /  EDSS at the time of diagnosis confirmation (median, IQR) / 3.0 (2.5–3.0) /  Patients with brain MRI lesions on T1 Gad + weighted sequences n (%) / 26 (100) /  SDMT Z-score (mean ± SD) / −0.6 ± 1.1 /  CVLT-II Z-score (mean ± SD) / −0.4 ± 0.9 /  BVMT Z-score (mean ± SD) / −0.7 ± 1.2 /  pRNFL(μm) / 118.9 ± 13.2 /  mRNFL(μm) / 21.3 ± 4.8 /  GCL (μm) / 43.3 ± 4.4 / CSF cerebrospinal fluid, OCB oligoclonal bands, N number, EDSS Expanded Disability Status Scale, MRI magnetic resonance imaging, OCT optical coherence tomography, pRNFL peripapillary retinal nerve fiber layer thickness, mRNFL macular RNFL thickness, GCL ganglionar cells layer thickness, SDMT Symbol Digit Modalities Test, CVLT-II California Verbal Learning Test-II, BVMT , Brief Visuospatial Memory Test, SCs symptomatic controls, IQR interquartile range, SD standard deviation *Chi-square tests and Fisher-Exact tests were used to compare categorical variables. Quantitative variables were compared with Mann Whitney-test

Baseline characteristics of the enrolled cohort

CSF cerebrospinal fluid, OCB oligoclonal bands, N number, EDSS Expanded Disability Status Scale, MRI magnetic resonance imaging, OCT optical coherence tomography, pRNFL peripapillary retinal nerve fiber layer thickness, mRNFL macular RNFL thickness, GCL ganglionar cells layer thickness, SDMT Symbol Digit Modalities Test, CVLT-II California Verbal Learning Test-II, BVMT , Brief Visuospatial Memory Test, SCs symptomatic controls, IQR interquartile range, SD standard deviation

*Chi-square tests and Fisher-Exact tests were used to compare categorical variables. Quantitative variables were compared with Mann Whitney-test

We excluded patients who had received immunosuppressive or immunomodulatory medications for other conditions, or who had been exposed to steroids within 30 days prior to the initial blood sample collection. High disease activity was defined as having at least two clinical relapses within the last 12 months and the presence of gadolinium-enhancing lesions on T1-weighted magnetic resonance imaging (MRI) sequences, in accordance with Italian Medicines Agency rules. MRI scans obtained within 6 months of confirmed diagnosis were designated as baseline MRIs. All imaging was performed using a consistent 3-Tesla MRI scanner (Siemens, Magnetom Skyra model).

Procedures and covariates definitions

Clinical and cognitive evaluations, and OCT acquisition

In RMS cohort, disability was assessed using Expanded Disability Status Scale (EDSS) by a Neurostatus certified neurologist. Cognitive function evaluation was conducted by neuropsychologist using Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS). This standardized battery consists of three specific tests: the Symbol Digit Modalities Test (SDMT), the California Verbal Learning Test-II (CVLT-II), and the Brief Visuospatial Memory Test (BVMT).

The resulting scores were converted to Z-scores based on established normative values.

The OCT scans were performed using a Spectralis device from Heidelberg Engineering (Heidelberg, Germany) in a darkened room without pupillary dilation. For our analysis, we utilized the average peripapillary retinal nerve fiber layer (pRNFL) thickness (μm), the macular RNFL thickness (mRNFL) (μm), and ganglionar cell layer (GCL) thickness (μm). A detailed description of the OCT scanning protocol has been provided according to previously validated reports [ 27 , 28 ].

To avoid confounding effects, we excluded eyes with a history of ON ( n  = 3) from our analysis due to the significant localized damage associated with this condition. For patients with unilateral ON, only the unaffected eye was included in the study. For all other participants, we calculated the mean values from both eyes for our analysis.

OCT examinations were conducted for all participants within 2 weeks from the BICAMS and EDSS assessment.

Biological assessment

Lumbar puncture and peripheral blood samples withdrawal were routinely performed as part of the diagnostic workflow after informed consent acquisition.

The CSF parameters included were white blood count, total protein, and CSF/serum albumin ratio. IgG oligoclonal bands (OCBs) were quantified in serum and CSF using an Immage 800 nephelometer (Beckman Culter, Nyon, Switzerland). They were analyzed and quantified by isoelectric focusing and immunoblotting. The Link index was also calculated (ratio of CSF IgG to CSF albumin, divided by the ratio of serum IgG to serum albumin). This ratio-of-ratios, when exceeding 0.7 (or the laboratory-specific threshold), was indicative of intrathecal IgG synthesis [ 29 ].

Peripheral blood samples were collected in Vacutainer tubes (BDInc, Franklin Lakes, NJ) without anticoagulant and centrifuged at 3000 rpm for 7 min to separate the serum fraction from the blood clot. CSF samples were collected into polypropylene tubes and centrifuged at 1000 rpm for 10 min to remove the cellular fraction.

Cell-free CSF and serum were stored at −80 °C until further analysis.

NfL analysis

NfL were quantified using ELLA system. Cell-free CSF and serum samples prepared with the protocol recommended on the analysis specification technical data sheet and not containing visible particulate matter did not require additional centrifugation before use. A 1:2 dilution into the specified cartridge inlet was used for the ELLA analysis.

The concentration of CSF NfL and serum NfL was measured with a Simple Plex Human NF-L Cartridge®(SPCKB-PS-002448-Bio-Techne, Minneapolis, MN, USA) using antibodies to NfL.

Neurogranin, sTREM2, and Ferritin analysis

A Multiplex Assay (HNS2MAG-95K, Millipore) using antibodies to Neurogranin, TREM2, and Ferritin was used to analyze the cell-free CSF of RMS and SCs. The measurement was performed by Luminex 200, following the procedure of protocol.

Outcomes

Our specific objectives were to investigate the levels of ferritin, neurogranin, and TREM2 in our homogeneous cohort of RMS patients at the time of diagnosis to verify their potential as disease biomarkers; to explore associations in the RMS cohort among these three investigated biomarkers with serum and CSF NfL, previously validated biomarkers for MS; to examine the associations between all these biomarkers and clinical, cognitive, and OCT parameters in RMS patients.

Statistical analysis

Normality distribution was verified using the Shapiro–Wilk test. A comparison between the two groups was conducted. For normally distributed data, an independent t test was employed to compare means, while for non-normally distributed data, the Mann–Whitney U test was utilized.

Biological markers are expressed as mean ± standardized error of the mean.

Statistically significance was settled to 0.05. To evaluate the diagnostic ability of the classifier, a Receiver-Operating Characteristic (ROC) curve was generated, and the Area Under the Curve (AUC) was calculated to quantify the overall performance. The ROC curve plots the True-Positive Rate against the False-Positive Rate at various threshold settings. Finally, a correlation analysis was performed to examine the relationships between variables in different cells. Partial correlation analysis adjusted for age and sex and expressed with Spearman's rank correlation coefficient was used to measure the strength and direction of associations between investigated biomarkers and RMS patients features.

Statistical analyses were performed using Python version 3.11.6. The following packages were utilized for specific analyses: Pandas 2.1.1 and NumPy 1.24.4 for descriptive statistics; Scikit-learn 1.3.1 for ROC analysis; SciPy 1.11.3 for correlation analysis; Matplotlib 3.8.0 and Seaborn 0.13.0 for data visualization.

Results

A total cohort of 39 subjects (69.2% women), with a mean age of 34.9 ± 9.7 years, was analyzed.

Demographic, clinical, neuropsychological, and OCT characteristics are shown in Table  1 .

Figure  1 reports the box plot comparing the levels of neurogranin, ferritin, and TREM-2 in patients with MS compared to OND. Neurogranin (741.5 ± 78.5 vs 403.7 ± 159.3 pg/mL, p  = 0.04) had higher levels in RMS patients when compared to those with SCs. Ferritin and TREM-2 did not show differences in MS when compared to SCs (17,858.6 ± 1240.6 vs 15,030.9 ± 1694.2 pg/mL, p  = 0.200 and 3020.3 ± 306 vs 2336.4 ± 237.7 pg/mL, p  = 0.119, respectively).

Box plot of neurogranin, ferritin, and TREM-2 levels in RMS and SCs (pg/mL). Data are expressed as mean ± SEM. RMS, relapsing–remitting MS; SCs symptomatic controls; TREM-2, Triggering Receptor Expressed on Myeloid cells 2. SEM, standard error of the mean

The ROC curve indicated that neurogranin could be considered as a disease biomarker: AUC 0.733 ( p  = 0.01, 95% CI 0.448–0.972). The ROC curve for TREM-2 has an approaching significance with AUC 0.699 ( P  = 0.051, 95% CI 0.433–0.928) (Fig.  2 ).

Logistic regression analyses with ROC curve output of patients with RMS plotted against the SCs group. The AUC, with 95% CI, is given for each parameter. The surface expression of each parameter for patients with RMS ( n  = 26) is combined as true positives and plotted against SCs as true negatives ( n  = 13). The diagonal dividing the ROC space represents the random event. A logistic regression analysis with combined parameter results has been performed for “all parameters”, parameters with AUC > 0.70, and AUC > 0.75. a Neurogranin; b TREM-2; c ferritin. ROC, receiver-operating characteristic; AUC, area under the curve; RMS, relapsing–remitting MS; SCs symptomatic controls; TREM-2, Triggering Receptor Expressed on Myeloid cells 2

The mean level of serum NfL in RMS patients was 12.4 ± 8.1 (pg/mL) and CSF NfL was 450.2 ± 451.2 (pg/mL).

The correlation analysis among the three investigated biomarkers and NFL revealed no significant correlation (Fig.  3 ).

Heatmap of partial correlations between MS biomarkers. The heatmap displays correlations between five biomarkers: Neurogranin, TREM-2, Ferritin, serum NfL, and CSF NfL. The biomarkers are listed on both axes, creating a symmetrical matrix. Each cell represents the correlation coefficient between two biomarkers, ranging from 0 to 1. The color intensity corresponds to the correlation strength: darker red indicates stronger positive correlations, while lighter shades represent weaker correlations. The diagonal shows perfect correlation (1.00) of each biomarker with itself. Only the upper triangle of the matrix is filled, as the lower triangle would be a mirror image. Correlation values are displayed in each cell, rounded to two decimal places. Note that only significant correlations ( p  < 0.05) are shown; blank cells indicate non-significant correlations

Strong positive correlation was found between ferritin and both neurogranin ( r  = 0.690, p  < 0.01) and TREM-2 ( r  = 0.570, p  < 0.05). A positive correlation was found also between CSF NfL and serum NfL ( r  = 0.858, p  < 0.001) ( Fig.  3 ).

The association among biomarkers and clinical, neuropsychological, and OCT parameters are shown in eTable2.

The most relevant finding was the inverse correlation between Ferritin and GCL thickness and the same trend between neurogranin and GCL thickness ( r  = 0.509, p  = 0.004 and r  = 0.375, p  = 0.034, respectively).

Additionally, both Ferritin and TREM-2 had a positive correlation with number of relapses in the year before diagnosis ( r  = 0.547, p  = 0.013 and r  = 0.552, p  = 0.012, respectively). TREM-2 had a positive association with baseline EDSS ( r  = 0.500, p  = 0.015).

Ancillary, SDMT (raw score) was negatively correlated with CSF number of OCB ( r  = −0.56, p  = 0.012) (data not showed).

Discussion

In this pilot study, neurogranin might emerge as a potential biomarker for MS, particularly in patients diagnosed within 24 months from clinical onset. The study hints at a possible relationship between ferritin, neurogranin, and TREM-2, which could suggest complex interactions between inflammation, oxidative stress, and neuronal damage in MS. While ferritin may not definitively serve as a specific marker for MS, its potential correlation with neurogranin might indicate the importance of considering multiple biomarkers to better understand the disease's processes and progression.

Furthermore, the potential inverse association of ferritin and neurogranin with GCL warrants further investigation. However, the specific link between ferritin levels and GCL damage has not been directly examined in the context of MS, underscoring the necessity for additional research to elucidate these potential relationships.

Iron metabolism in retinal ganglion cells involves intricate processes of iron uptake, storage, and regulation [ 30 , 31 ]. Iron overload could lead to retinal diseases, emphasizing the importance of understanding the regulatory mechanisms for potential targeted therapies.

Studies have shown that iron could exacerbate the activated state of microglia, contributing to inflammation and further neuronal injury [ 32 ]. Additionally, the interplay between iron accumulation, mitochondrial dysfunction, and inflammation is critical in the pathogenesis of neurodegenerative disorders, including MS [ 33 ].

Moreover, the iron uptake by microglia occurs through specific pathways, and alterations in iron-related proteins have been observed in various neurodegenerative conditions [ 34 ]. This suggests that iron dysregulation may play a significant role in the pathological processes involving microglia and neuronal injury [ 35 ]. Overall, the interplay between iron, microglial activation, and neuronal injury is a critical area of research in understanding neurodegenerative diseases.

Further research should also investigate the interplay between iron metabolism and synaptic damage in MS.

Additionally, the observed association between the number of relapses in the year preceding diagnosis and the levels of both Ferritin and TREM-2 suggests their possible link with high disease activity.

Ancillary, the negative association between the SDMT, the most sensitive test for cognitive impairment in MS, and the number of OCBs in the CSF, considered a marker of inflammation, underscores in this homogeneous cohort of patients with early MS how an interplay between inflammation and neurodegeneration from diagnosis [ 36 – 38 ].

Longitudinal studies with larger sample size are needed to validate these findings and explore the potential of these biomarkers in predicting disease progression and treatment response.

The first limit of the study is its preliminary nature. The effect size (Cohen's d) of 0.77 for CSF neurogranin levels between MS patients and OND, demonstrates a moderate-to-large effect. The current power of 0.58, while below the conventional 0.8 threshold, is reasonable for a pilot study. To achieve 80% power, future studies should aim for at least 27 subjects per group. This pilot provides valuable insights for hypothesis generation and sample size estimation for more extensive studies. Undoubtedly, the findings are preliminary and should be interpreted with caution. The 2:1 case–control ratio, while not ideal, is justified given the exploratory nature of this research. Future studies should strive for a more balanced design and larger sample sizes to confirm these preliminary findings and explore other potential biomarkers in MS. By elucidating these relationships, we hope to gain preliminary insights into the complex interplay between inflammation and neurodegeneration in MS. This knowledge could potentially lead to the identification of novel biomarker combinations for early diagnosis, prognosis, and monitoring of therapeutic responses in MS.

This pilot study represents a crucial step toward unraveling the complex "inflammation-degeneration tangle" in early MS (Fig.  4 ), potentially paving the way for more comprehensive investigations and novel therapeutic approaches.

Interconnected network of biomarkers and pathophysiological processes in MS. This diagram illustrates the complex relationships between investigated biomarkers (including CSF and serum neurofilaments, neurogranin, ferritin, and TREM2), physiological processes (such as iron metabolism and oxidative stress), and potential therapeutic targets in MS. Key elements include the central role of biomarkers, their correlations, and their connections to ganglion degeneration, targeted therapies, and neuroprotection. The diagram emphasizes the multifaceted nature of MS pathology and the importance of diverse biomarkers in understanding and potentially treating the disease

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 382 KB)

Supplementary file2 (PDF 160 KB)

Author contribution

Aurora Zanghì: conceptualization, data curation, formal analysis, and writing-draft. Annamaria Greco: conceptualization, data curation, and biological analysis; Ermete Giancipoli: conceptualization, data curation, and methodology. Hayrettin Tumani: conceptualization, data curation, and methodology; Carlo Avolio: conceptualization and data curation. Emanuele D’Amico: conceptualization, data curation, visualization, supervision, and writing, original draft .

Funding

The researchers were independent from funders and sponsors. All researchers could access all the data.

Data availability

Anonymised data will be shared upon request from any qualified investigator for the sole purpose of replicating procedures and results presented in the report provided that data transfer is in agreement with EU legislation on the general data protection regulation.

Disclosure

This study was not sponsored.

Declarations

Conflict of interest

Aurora Zanghì has nothing to disclose related to the submitted manuscript. Annamaria Greco has nothing to disclose related to the submitted manuscript. Ermete Giancipoli has nothing to disclose related to the submitted manuscript. Hayrettin Tumani has nothing to disclose related to the submitted manuscript. Carlo Avolio has nothing to disclose related to the submitted manuscript. Emanuele D’Amico has nothing to disclose related to the submitted manuscript.

Research ethics and patient consent

The study protocol was approved by the local ethics committee (Comitato Etico Foggia, CE/109/2024/PO). All patients signed a written informed consent to participate in the study. The current report does not contain any individual or identifying information. The study was conducted in accordance with the ethical principles of the Declaration of Helsinki.

References

Associated Data

Supplementary Materials

Supplementary file1 (PDF 382 KB)

Supplementary file2 (PDF 160 KB)

Data Availability Statement

Anonymised data will be shared upon request from any qualified investigator for the sole purpose of replicating procedures and results presented in the report provided that data transfer is in agreement with EU legislation on the general data protection regulation.

Disclosure

This study was not sponsored.