Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 https://doi.org/10.1186/s12974-022-02635-3 RESEARCH Open Access Peripheral myeloid‑derived suppressor cells are good biomarkers of the efficacy of fingolimod in multiple sclerosis Celia Camacho‑Toledano1, Isabel Machín‑Díaz1, Leticia Calahorra1, María Cabañas‑Cotillas1, David Otaegui2, Tamara Castillo‑Triviño2,3, Luisa María Villar4, Lucienne Costa‑Frossard4,5, Manuel Comabella6, Luciana Midaglia6, José Manuel García‑Domínguez7, Jennifer García‑Arocha1, María Cristina Ortega1 and Diego Clemente1* Abstract Background: The increasing number of treatments that are now available to manage patients with multiple sclerosis (MS) highlights the need to develop biomarkers that can be used within the framework of individualized medicine. Fingolimod is a disease-modifying treatment that belongs to the sphingosine-1-phosphate receptor modulators. In addition to inhibiting T cell egress from lymph nodes, fingolimod promotes the immunosuppressive activity of myeloid-derived suppressor cells (MDSCs), whose monocytic subset (M-MDSCs) can be used as a biomarker of dis‑ ease severity, as well as the degree of demyelination and extent of axonal damage in the experimental autoimmune encephalomyelitis (EAE) model of MS. In the present study, we have assessed whether the abundance of circulating M-MDSCs may represent a useful biomarker of fingolimod efficacy in EAE and in the clinical context of MS patients. Methods: Treatment with vehicle or fingolimod was orally administered to EAE mice for 14 days in an individualized manner, starting the day when each mouse began to develop clinical signs. Peripheral blood from EAE mice was col‑ lected previous to treatment and human peripheral blood mononuclear cells (PBMCs) were collected from fingoli‑ mod to treat MS patients’ peripheral blood. In both cases, M-MDSCs abundance was analyzed by flow cytometry and its relationship with the future clinical affectation of each individual animal or patient was assessed. Results: Fingolimod-treated animals presented a milder EAE course with less demyelination and axonal damage, although a few animals did not respond well to treatment and they invariably had fewer M-MDSCs prior to initiat‑ ing the treatment. Remarkably, M-MDSC abundance was also found to be an important and specific parameter to distinguish EAE mice prone to better fingolimod efficacy. Finally, in a translational effort, M-MDSCs were quantified in MS patients at baseline and correlated with different clinical parameters after 12 months of fingolimod treatment. M-MDSCs at baseline were highly representative of a good therapeutic response to fingolimod, i.e., patients who met at least two of the criteria used to define non-evidence of disease activity-3 (NEDA-3) 12 months after treatment. Conclusion: Our data indicate that M-MDSCs might be a useful predictive biomarker of the response of MS patients to fingolimod. Keywords: EAE, MDSCs, S1PR, Responder and non-responder, Personalized medicine Background *Correspondence: dclemente@sescam.jccm.es Multiple sclerosis (MS) is a chronic inflammatory demy- 1 Neuroimmuno‑Repair Group, Hospital Nacional de Parapléjicos-SESCAM, elinating disease of the central nervous system (CNS) that Finca La Peraleda s/n, 45071 Toledo, Spain Full list of author information is available at the end of the article affects about 2.8 million people worldwide [1]. However, © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 2 of 22 the development of MS does not follow a unique pattern most positive response to fingolimod in MS patients but rather it has an unpredictable course. Around 85% of has been addressed [14–16]. However, there is little patients develop the relapsing–remitting (RR-MS) form data regarding the innate immune cells to predict the of the disease, which is essentially characterized by epi- responsiveness of MS patients to fingolimod, attributed sodes of neurological dysfunction (relapses) followed by to ­C D56bright NK cells [17], with no data regarding the partial or full recovery [2]. Our current understanding of myeloid cell compartment. the influence of the immune system on MS pathology has Myeloid-derived suppressor cells (MDSCs) are a heter- led to the approval of 17 disease-modifying treatments ogeneous population generated in the bone marrow dur- (DMTs) for RR-MS, providing neurologists and patients ing myelopoiesis that contribute to the innate immune a wide range of therapeutic options [3]. Clinicians must response. Under pathological conditions, blocking their select treatment strategies for each individual based on differentiation in conjunction with exposure to inflam- a personalized risk and benefit assessment [4]. How- matory mediators activates these cells, provoking their ever, while the therapeutic options are constantly evolv- acquisition of immunosuppressive functions [18, 19]. In ing, escalation is still a common strategy followed in the the experimental autoimmune encephalomyelitis (EAE) decision-making process for MS treatment [5]. It consists mouse model of MS, splenic MDSCs are composed in using a first-line low risk, moderate efficacy therapy, of two subpopulations: polymorphonuclear-MDSCs and then monitoring the effectiveness of the treatment. (PMN-MDSCs, ­CD11b+Ly-6CintLy-6G+) and monocytic- When the DMT fails in controlling the disease activity, MDSCs (M-MDSCs, ­CD11b+Ly-6ChiLy-6G−, which are it is switched to a more efficacious one. This approach is phenotypically indistinguishable from Ly-6Chi inflam- quite conservative and it is most applicable to patients matory monocytes in blood, i.e., Ly-6Chi cells [20]). The with a lower risk of developing an aggressive form of abundance of M-MDSCs has repeatedly been related to MS [6]. Mounting evidence indicates that early imple- disease severity in the EAE model [21–24]. Indeed, in sev- mentation of efficacious treatment may prevent chronic eral immune-mediated animal models other than EAE, immune activation and better preserve long-term neural it appears that fingolimod or other S1PR agonists may integrity [7]. Indeed, the most effective drugs should be be implicated in the infiltration capacity, the expansion indicated in patients with very active MS associated with and the immunosuppressive activity of MDSCs through an earlier development of disability. Among other DMTs, their antagonism of S1PR1, 3 and 5 [25–28]. Hence, those designed to act as sphingosine-1-phosphate recep- there appears to be a clear relationship between the tor (S1PR) modulators are a large family of drugs that has therapeutic response to fingolimod and the abundance the ability to be used for this purpose. of MDSCs. However, the utility of these cells as poten- Gilenya™, is an S1PR antagonist that was the first oral Fingolimod (FTY720), known commercially as tial biomarkers of the response to fingolimod in inflam- matory conditions (including EAE and MS) remains DMT approved for the treatment of RR-MS [8]. It is a totally unexplored. Circulating PMN-MDSCs (identified structural analogue of sphingosine that upon phospho- as ­CD11b+CD33+HLA-DR−/lowCD14−CD15+LOX1+ rylation acts as a modulator of S1PR1, 3, 4 and 5, with cells) and M-MDSCs (­ CD11b+CD33+HLA-DR−/ low + − higher affinity for S1PR1. Fingolimod is a potent inhibi- CD14 CD15 cells) [29], yet MDSCs have been little tor of lymphocyte egression from lymphoid tissues as a studied in MS and there is conflicting data as to the cell result of the internalization and degradation of S1PR1 type most strongly involved or altered during the clinical [9], although additional biological effects on other CNS course of MS [30–32]. There is also little information on or immune cells have also been described [10]. The the possible effect of DMTs on MDSCs, most of which heterogeneity of MS makes it difficult and costly to has been limited to glatiramer acetate (GA), which has determine the effectiveness of any specific treatment, been reported to induce a negligible increase in MDSCs including fingolimod. Hence, it is important to iden- [32]. Notably, there is no data about MDSCs as a predic- tify biomarkers that provide information about disease tive biomarker for MS treatments. progression and treatment efficacy for each individual In the current work, we present novel data in the con- to avoid undesirable side effects, and to reduce the text of EAE and MS patients indicating that M-MDSCs cost burden to Health Systems. In this sense, a num- could serve as a biomarker for responsiveness to fin- ber of biomarkers have been described to monitor the golimod treatment and its efficacy in EAE. Patients who effects of exposure to a drug and its pharmacodynam- exhibited a good therapeutic response to fingolimod at ics, as well as to identify the likelihood of toxicity as 12 months of treatment had a greater level of circulat- an adverse effect of fingolimod treatment [11–13]. In ing M-MDSCs at baseline relative to those considered recent years, the role of cells of the adaptive immune to be non-responders. Our observations in EAE and MS response as predictive biomarkers to anticipate the might serve to identify new tools that in turn, could help Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 3 of 22 modify the prescribing criteria of neurologists’ and ulti- Council Directive 2010/63/EU, 90/219/EEC, Regula- mately, patients’ quality of life. tion No. 1946/2003), and with the Spanish National and Regional Guidelines for Animal Experimentation (RD Methods 53/2013 and 178/2004, Ley 32/2007 and 9/2003, Decreto Induction of EAE 320/2010). Chronic progressive EAE was induced by subcutaneous immunization of female 6- to 8-week-old C57BL/6 mice Evaluation of the clinical signs (Janvier Labs, France) with 200 µL of a Myelin Oligo- The clinical signs were established through observational dendrocyte Glycoprotein solution (200 µg, M ­ OG35–55 evaluation of the functional behavior of each mouse, peptide: Genscript, New Jersey, USA) emulsified in com- allowing different parameters of the EAE clinical course plete Freund’s Adjuvant (CFA) that contained 4 mg/mL to be defined: (i) the day of onset, the 1st day mice had of Mycobacterium tuberculosis (BD Biosciences, Franklin a clinical score ≥ 0.5; (ii) clinical score at onset; (iii) the Lakes, New Jersey, USA). Immunized mice were admin- maximum clinical score; (iv) the day when animals istered Pertussis toxin intravenously (iv) through the tail reached the maximum clinical score or “peak” defined vein (250 ng/mouse: Sigma-Aldrich, St. Louis, MO, USA) as the 1st day a clinical score ≥ 1.5 was repeated; (v) the in a final volume of 100 µL on the day of immunization severity index, the maximum clinical score/the days and 48 h later. EAE was scored clinically on a daily basis elapsed from onset to peak [24]; (vi) the accumulated by three observers in a double-blind manner as follows: clinical score from onset to peak and the total accumu- 0, no detectable signs of EAE; 0.5, half limp tail; 1, whole lated clinical score, considered as the sum of the individ- limp tail; 1.5, hind limb inhibition (unsteady gait and/ ual clinical scores from the day of onset until the peak or or poor hind leg grip); 2, hind limb weakness; 2.5, bilat- the end of the treatment, respectively; (vii) the percent- eral partial hind limb paralysis or unilateral full hind age recovery [(the maximum clinical score − the residual limb paralysis; 3, complete bilateral hind limb paralysis score at the end of treatment) × 100/maximum clinical (paraplegia); 3.5, partial forelimb paralysis; 4, tetraplegia; score]; (viii) the recovery index [(the maximum clinical 4.5, moribund; and 5, death. In accordance with ethi- score at peak − the residual score at the end of treat- cal regulations, humane end-point criteria were applied ment)/days elapsed from the peak to the recovery]; and when an animal reached a clinical score ≥ 4, when clinical (ix) the residual score, the clinical score at the end of the score ≥ 3 was reached for more than 48 h, or when self- follow-up. mutilation was evident, persistent retention of urine, 35% weight loss and signs of stress or pain for more than 48 h, Flow cytometry analysis of peripheral blood even if the EAE score was < 3. The generation of sounds, and splenocytes lordokyphosis or stereotypic behavior were considered Blood was obtained from the retro-orbital venous sinus as signs of stress. In the present study, no animals pre- of isoflurane-anesthetized mice at disease onset and sented signs of pain or stress, and none reached a clinical collected in 2% EDTA tubes, and the erythrocytes were score > 3.5. then lysed in a 15-mL tube with 2.5 mL ACK lysis buffer: The individualized treatment and follow-up was car- 8.29 g/L ­NH4Cl; 1 g/L ­KHCO3; 1 mM EDTA in distilled ried out as follows. The 1st day clinical signs of EAE ­H2O at pH 7.4 (Panreac). Subsequently, the lysis reaction were detected (“onset”, therapeutic window during which was stopped by adding wash buffer (sterile 1× Phosphate the clinical score debuts between 0.5 and 1.5), each ani- Buffered Saline—PBS) and the blood cells were recovered mal was administered fingolimod orally (N = 18; 3 mg/ by centrifugation at 210 g for 5 min at room temperature kg in distilled water: EAE-FTY720, kindly provided by (RT). The Fc cell receptors were then blocked for 10 min Novartis Pharma AG, Switzerland [33–35]) or the vehicle at 4 °C with an anti-CD16/CD32 antibody (10 μg/mL: alone (N = 19; distilled water: EAE-Veh) on a daily basis BD Biosciences 553142) and the cells were resuspended for 14 days. All mice were killed on the 15th day post- in 25 μL of staining buffer: PBS supplemented with 10% onset (dpo) and four different variables were analyzed in heat-inactivated fetal bovine serum (FBS: Linus); 25 mM all the EAE mice: (a) the clinical signs; (b) the peripheral HEPES buffer (Gibco); and 2% Penicillin/Streptomycin blood content; (c) splenic content; and (d) histopatholog- (P/S: Gibco). The cells were further labeled for 20 min ical parameters. at 4 °C in the dark with the following fluorochrome- All the animal manipulations were approved by the conjugated monoclonal antibodies diluted in 25 μL of institutional ethical committee (Comité Ético de Experi- staining buffer (Additional file 1: Table S1): anti-Ly- mentación Animal del Hospital Nacional de Parapléjicos), 6C-FITC (AL-21 clone), anti-Ly-6G-PE (1A8 clone), and all the experiments were performed in compliance anti-CD11b-PerCP-Cy5.5 (M1/70 clone: all from BD with the European guidelines for animal research (EC Biosciences); anti-MHC-II-PE-Cy7 (M5/114.15.2 clone), Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 4 of 22 anti-CD11c-APC (N418 clone), anti-F4/80-eFluor450 dehydrated in a graded ethanol series and xylol solu- (BM8 clone) to stain myeloid cells (eBioscience-Thermo tions, and they were mounted with Entellan (Merck) for Fisher Scientific); and anti-CD3e-PB (500A2 clone), preservation. anti-CD4-PE (RM4-5 clone), anti-CD8a-FITC (53–6.7 To analyze axonal damage, the non-phosphorylated clone: all from BD Bioscience), anti-CD25-BV421 form of the neurofilament protein (SMI-32) was stained (PC61.5 clone) and anti-CD69-APC (H1.2F3 clone) for by immunohistochemistry. After several rinses in PB, lymphoid cell staining (Invitrogen). The blood cells were the sections were treated with 10% methanol in PB for then rinsed with staining buffer, recovered by centrifuga- 20 min and then incubated for 1 h at RT in incubation tion at 210 g for 5 min, fixed for 10 min at RT with 4% buffer: 5% normal donkey serum (NDS: Merck) and 0.2% paraformaldehyde (PFA: Sigma-Aldrich) in 0.1 M Phos- Triton X-100 (Merck) diluted in PBS. Immunostaining phate Buffer (pH 7.4, PB), resuspended in PBS and finally was performed by incubating the sections overnight at analyzed in a FACS Canto II cytometer (BD Biosciences) 4 °C with the SMI-32 primary antibody (Biolegend: Addi- at the Flow Cytometry Core Facility of the Hospital tional file 1: Table S1) in incubation buffer. The sections Nacional de Parapléjicos. The data obtained were ana- were then incubated for 1 h at RT with the correspond- lyzed using the FlowJo 10.6.2 software (Tree Star Inc.). ing fluorescent secondary antibody diluted in staining Fresh spleens were collected from euthanized mice at buffer (1:1000: Invitrogen), the cell nuclei were stained the end of fingolimod/vehicle administration and the tis- with Hoechst (10 µg/mL: Sigma-Aldrich) and the sam- sue was processed mechanically to obtain a single cell ples were mounted in Fluoromont-G (Southern Biotech). suspension that was passed through a 40-µm filter (BD Omission of the primary antibody as a control staining Biosciences), before it was washed in a 50-mL tube with yielded no signal. cold supplemented RPMI medium: RPMI 1640 (Gibco) with 2 mM l-Glutamine (l-Glu: ThermoFisher), 10% FBS Image acquisition and analysis and 1% P/S. Erythrocytes were lysed with 1 mL of ACK To quantify the degree of demyelination, sections were lysis buffer and the reaction was stopped by adding PBS. selected randomly from the sub-cohort group of 18 EAE After that, splenocytes were recovered by centrifugation mice (at least three sections per mouse, each separated at 210 g for 5 min and immunostained with myeloid/ by 340 µm). EC stained spinal cord sections were ana- lymphoid markers following the same protocol as for the lyzed on a stereological Olympus BX61 microscope to peripheral blood samples. quantify demyelination, using a DP71 camera (Olym- pus) and VisionPharm software for anatomical mapping. Tissue extraction and histopathological analysis Superimages were acquired at a ×10 magnification using After removing their spleen, all the animals were per- the mosaic tool and analyzed with ImageJ, expressing the fused transcardially with 2% PFA, and their spinal cords results as the total area of white matter demyelination were dissected out and post-fixed for 4 h at RT in the and relative to the whole white matter area. same fixative. After immersion in 30% (w/v) sucrose in To quantify axonal damage, ×20 tile scan images from Nussloch) were thaw-mounted onto Superfrost® Plus PB for 12 h, coronal cryostat sections (20 µm thick: Leica, the spinal cord of each animal were obtained on an IX83 inverted microscope (Olympus) equipped with CellSense slides. A randomly selected group of 18 EAE mice from Dimensions software for large-area imaging. The area the main cohort previously used for the spleen analy- of axon damage relative to the whole spinal cord or the sis was studied, obtaining histopathological parameters total demyelinated area was analyzed with an ad hoc from the spinal cord: the area and percentage of demyeli- plug-in designed by the Microscopy and Image Analysis nated white matter, and the axonal damage associated to Core Facility at the Hospital Nacional de Parapléjicos. white matter lesions. After selecting the demyelinated area, a threshold for Demyelinated regions were quantified in eriochrome immunofluorescence was established and SMI-32 immu- cyanine (EC) stained spinal cord tissue. Sections of the nostaining was measured, representing the result as the spinal cord were air-dried at RT for 12 h followed by 2 h percentage of axonal damage within the inflammatory at 37 °C, placed in fresh acetone for 5 min and air-dried lesion. again for 30 min before they were stained for 30 min with 0.2% EC solution. The sections were then differentiated MS patient cohort for M‑MDSC blood analysis in 5% alum and borax-ferricyanide for 10 min each, and 31 RR-MS patients fulfilling the 2010-revised McDon- after rinsing with abundant tap water, correct differen- ald diagnostic criteria [36] who initiated fingolimod tiation was assessed under the microscope to ensure treatment were recruited at the Vall d’Hebron Hospi- myelinated areas were stained blue and demyelinated tal (Barcelona), the Ramón y Cajal Hospital (Madrid) or areas appeared yellowish-white. Finally, the slides were the Donostia Hospital (San Sebastián). The study was Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 5 of 22 approved by the corresponding local ethics committees, at RT in the dark, the cells were washed with PBS and and informed written consent was obtained from all the recovered at 390 g for 5 min. PBMCs were then resus- participants in accordance with the Helsinki declaration. pended in 25 μL of staining buffer and incubated for None of the MS patients in this study received corticos- 10 min at 4 °C with Beriglobin (50 μg/mL: CSL Behring) teroids in the last 30 days and they were treated with to block the Fc receptors. Subsequently, 25 µL of the fingolimod (0.5 mg/day) based on current guidelines. antibody mix diluted in staining buffer was added to the In patients who switched treatment from a prior DMT, cells for 30 min at 4 °C in the dark. The antibody panel a minimum washout period of 24 h was established for (Additional file 1: Table S1) to stain M-MDSCs included those who had been receiving interferon-beta (IFN-β) anti-CD15-FITC (HI98 clone), anti-CD14-PerCP- or GA, of 2 months for natalizumab, and of 1 month for CyTM5.5 (Mφ29 clone), anti-CD11b-PE-Cy7 (ICRF44 azathioprine. All the patients completed 12 months of clone), anti-CD33-APC (WM53 clone) and anti-HLA- fingolimod treatment (for demographic and clinical char- DR-BV421 (G46-6 clone: BD Biosciences). After stain- acteristics see Additional file 2: Table S2). ing, the PBMCs were rinsed with staining buffer at 390 g Neurological evaluation was carried out at baseline, 6 for 5 min at RT, resuspended in 0.1% PFA diluted in PBS and 12 months after initiation of the fingolimod therapy and stored at 4 °C in the dark until the next day. The with additional evaluation in case of relapses. Relapses samples were then analyzed in a FACS Canto II cytom- were defined as a worsening of neurological impairment eter (BD Biosciences) with the corresponding controls or the appearance of new symptoms attributable to MS in and 100.000 events were recorded in the mononuclear the absence of fever or infection, lasting at least 24 h and cell (MNC) gate. Compensation controls were obtained preceded by at least 1 month of stability [37]. A standard- by incubating each single antibody with the anti-mouse ized brain magnetic resonance imaging (MRI) protocol Ig,k/Negative Control Compensation Particle Set (Com- (axial T2, axial T2 FLAIR and T1) was implemented to pBeads, BD Biosciences). Fluorescence Minus One con- reliably detect T2 lesions at baseline and new or enlarg- trols were set-up in pooled samples to avoid any variation ing lesions at 12 months after initiation of the fingoli- between the experimental conditions and the data were mod therapy. The transverse T1-weighted sequence was analyzed with FlowJo v.10.6.2 software (FlowJo, LLC-BD systematically repeated after gadolinium (Gd) injection Biosciences). (0.2 mmol/kg) whenever T2 lesions were demonstrated. Both MRI scans were performed using the same 1.5-T Statistical analysis equipment with a slice thickness varying from 2 to 5 mm. The data were analyzed with SigmaPlot version 11.0 (Sys- Disability was assessed by a trained neurologists using tat Software Inc.). Age, and gender balance were com- the Kurtzke Expanded Disability Status Scale (EDSS) [38] pared using a Fisher exact test. A Student’s t-test was prior to the initiation of fingolimod therapy and then at used to compare pairs of the different groups of mice/ 6 and 12 months, and whole blood was collected from patients with parametric distributions or a Mann–Whit- these patients at the same time points during follow-up. ney U test was used for non-parametric distributions. A Spearman test was carried out for the correlation analy- Flow cytometry analysis of peripheral blood cells ses. A ROC curve analysis was performed to determine from fingolimod‑treated patients the discriminative ability of the different clinical parame- Fresh anti-coagulated EDTA blood was obtained by veni- ters scored and to determine the threshold of M-MDSCs, puncture the same day of the treatment, before the first with the highest sensitivity and specificity to separate dose of fingolimod and after 12 months of treatment. responder from non-responder mice. In addition, treat- Human peripheral blood mononuclear cells (PBMCs) ment efficacy was also determined by a ROC curve anal- were isolated by Ficoll density gradient centrifugation ysis. Finally, a Cox regression analysis was performed (GE-171440-02, Merck), collected from the interphase, to determine the probability of reaching the population washed with isolation buffer (2.23 g/L d-glucose, 2.2 g/L median of the clinical parameters presented by the vehi- sodium citrate, 0.8 g/L citric acid, 0.5% Bovine Serum cles as a function of different categorical and non-cate- Albumin in PBS) and further centrifuged at 350 g for gorical variables. The minimum statistical significance 10 min at RT. The cell pellet was resuspended in FBS, was set at p < 0.05: *p < 0.05, **p < 0.01, ***p < 0.001. counted and aliquoted 1:1 in FBS with 20% dimethylsul- foxide (DMSO: Sigma-Aldrich). The samples were then Results stored long-term in liquid nitrogen at − 170 °C until use. Individualized fingolimod treatment accelerates disease Freshly thawed PBMCs (1 × ­106) were washed twice recovery in EAE with PBS and stained with Zombie NIR Dye (Biolegend) Fingolimod has been seen previously to improve the following the manufacturer’s instructions. After 15 min clinical course of EAE [39, 40], although in all these Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 6 of 22 earlier studies fingolimod treatment was carried out on 20, 21, 41, 42]. Interestingly, high levels of splenic Ly-6Chi randomly assigned groups of mice with diverse clinical cells at the peak of EAE have been directly associated scores, without taking into account the disease severity of with a milder disease course [24, 43]. Hence, to explore each individual animal. In order to better approximate to whether M-MDSCs could be used a predictive tool for the clinical situation of MS patients, we used an individu- fingolimod-induced reduction in the severity of the clini- alized treatment protocol that started at the specific point cal course of EAE, we first investigated the correlation of disease onset for each animal. Differences between the between the abundance of Ly-6Chi cells in the periph- experimental groups were observed from 1 dpo, which eral blood at disease onset and different short- and long- became constantly significant from 4 dpo onwards. In term clinical parameters of the clinical course of EAE. general terms, the individualized fingolimod treatment As in the case of EAE-Veh, the higher proportion of Ly- significantly reduced the maximum clinical score of each 6Chi cells relative to the total blood content was directly animal, enhancing their recovery at a higher and faster related to a milder EAE disease course: lower maximum rate (the percentage recovery and recovery index, respec- score at peak, lower total accumulated and residual clini- tively). In addition, a decrease in their residual score and cal scores, and stronger relative disease recovery in EAE- reduced total accumulated score was observed relative to FTY720 mice (Fig. 2A–D). These data pointed to the the EAE-Veh animals (Fig. 1A, Table 1). The amelioration quantification of Ly-6Chi cells as an interesting parameter in the clinical signs observed in mice administered fin- in the search for biomarkers that link disease severity golimod was paralleled by a reduction in demyelination: reduction after fingolimod treatment. demyelinated area- EAE-Vehicle: 59,796.97 ± 22,444.54, To check whether the initial randomized distribution EAE-FTY720: 22,168.25 ± 15,461.25, p = 0.043); % of the of EAE mice influenced the long-term effects of fingoli- demyelinated area within the white matter (EAE-Veh: mod, we compared the clinical and immunological base- 6.26 ± 2.30, EAE-FTY720; 2.14 ± 1.44, p = 0.042: Fig. 1B, line conditions between the two experimental groups. C, F); and % of axonal damage within the demyelinated The day of onset and the clinical score at that point did area (EAE-Veh: 0.98 ± 0.22, EAE-FTY720: 0.43 ± 0.12, not differ between these groups: day at onset (EAE-Veh p = 0.045: Fig. 1D, E, G). Moreover, the individualized 13 dpi Inter Quartile Range—IQR 12–14), EAE-FTY720: fingolimod treatment also dampened the abundance of 13 dpi IQR 12–13, p = 0.789); and clinical score at onset splenic ­CD3+, ­CD8+ and C ­ D3+CD25+ T lymphocytes (EAE-Veh: 0.75 IQR 0.50–0.75), EAE-FTY720: 0.75 IQR (Table 1). 0.5–1.0, p = 0.763). Importantly, no major differences In summary, the individualized follow-up in each were reported in the abundance of Ly-6Chi cells between mouse from the point of disease onset appeared to be a the animals treated with vehicle or fingolimod, neither good strategy to explore the differential efficacy of fin- relative to the total number of blood cells or in the mye- golimod treatment in EAE. loid subset (Fig. 3A, B). Moreover, the Ly-6Chi cells from both animal groups showed a similar mean fluorescence The abundance of Ly‑6Chicells in peripheral intensity (MFI) in terms of their activation or the differ- blood is related to a milder long‑term EAE course entiation markers expressed prior to vehicle or fingoli- in fingolimod‑administered mice mod administration (Fig. 3C–E). Therefore, these data To establish the M-MDSC values that might predict a clearly indicated that the content and characteristics of future response to fingolimod in MS, we explored the the circulating Ly-6Chi cells was not biased between the relationship between their abundance in the blood and two groups of animals prior to treatment. the future short and long-term clinical course in fin- We also looked for differences in the abundance of golimod-administered EAE mice. Highly immunosup- circulating T cells, the main targets for both fingolimod pressive M-MDSCs in the spleen and spinal cord of EAE and MDSCs, prior to vehicle/fingolimod administration. mice, as well as circulating M-MDSCs, are phenotypi- No differences were detected in ­CD3+ or ­CD8+ T lym- cally indistinguishable from ­CD11b+Ly-6ChiLy-6G−/low phocytes, although there was a slightly higher percent- inflammatory monocytes, also called Ly-6Chi cells [18, age of ­CD4+ T lymphocytes in animals that then received (See figure on next page.) Fig. 1 Individualized fingolimod treatment accelerates disease recovery and decrease the histopathological damage in EAE mice. A Individualized fingolimod treatment from disease onset significantly reduced the clinical course and severity of EAE in mice (EAE-Veh N = 19; EAE-FTY720 N = 18). B, C Panoramic images of the spinal cord of one EAE-Veh (B) and one EAE-FTY720 (C) mouse stained with EC (myelin, blue). Dashed lines define the demyelinated area. D, E Detailed view of SMI-32 staining (axonal damage, red) in infiltrated areas (delimitated by dashed lines) of the spinal cord of EAE-Veh (D) and EAE-FTY720 (E) mice (nuclei are stained with Hoechst). F, G Quantification of the relative demyelination within the white matter or the axonal damage within the infiltrated area in the histological sub-cohort of EAE mice (EAE-Veh N = 9, EAE-FTY720 N = 9). Scale bar: B, C = 250 µm; D, E = 50 µm Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 7 of 22 Fig. 1 (See legend on previous page.) Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 8 of 22 Table 1 Comparison of the clinical parameters and splenic T cell populations in EAE mice following individualized treatment with fingolimod or vehicle Vehicle FTY720 p value Clinical parameters Maximum score 3 (2.25–3) 2.38 (1.18–2.75) < 0.05 dpi peak 15 (14–17) 15 (14–15.25) ns Severity index 0.75 (0.50–1.00) 0.625 (0.33–0.75) ns Peak accumulated score 6.50 (5.63–9.00) 5.69 (3.68–6.81) ns Total accumulated score 37.25 (23.00–42.38) 18.88 (12.59–27.16) < 0.01 % Recovery 25.00 (12.50–33.30) 72.73 (53.41–80.24) < 0.01 Recovery index 0.093 (0.04–0.17) 0.16 (0.10–0.21) < 0.05 Residual score 2.25 (1.00–2.63) 0.50 (0.50–0.88) < 0.01 Spleen % ­CD3+/Total blood cells 15.15 ± 0.97 11.06 ± 0.93 < 0.01 % ­CD4+/Total blood cells 9.11 ± 0.64 7.75 ± 0.63 ns % ­CD8+/Total blood cells 4.98 ± 0.36 2.79 ± 0.28 < 0.001 % ­CD3+CD25+/Total blood cells 1.32 ± 0.08 1.09 ± 0.07 < 0.05 % ­CD3+CD69+/Total blood cells 1.77 ± 0.12 1.86 ± 0.10 ns The data are expressed as the median (IQR) for the clinical parameters and the mean ± SEM for the splenic parameters. The results of applying a Student’s t-test or a non-parametric Mann–Whitney test are shown: dpi day post-immunization; IQR interquartile range; ns non-significant fingolimod (Table 2). To check whether the difference in which in turn showed no difference between them ­CD4+ T lymphocytes might reflect an imbalance in the (Table 3). Indeed, similarities between EAE-Vehicle mice immunological environment, we compared the Ly-6Chi/T and EAE-FTY720 non-responders were also observed in lymphocyte ratio between the two experimental groups. the splenic immunological status at the end of the follow- No differences were detected in the relative abundance of up. Interestingly, only responder mice had a significantly Ly-6Chi cells and C­ D3+, ­CD4+ or C­ D8+ T lymphocytes smaller proportion of splenic ­CD3+, ­CD4+, ­CD8+ and between mice administered fingolimod or the vehicle ­CD3+CD25+T lymphocytes (Table 3). alone, downplaying the significance of the differences at When the level of circulating Ly-6Chi cells in the blood baseline after considering the inflammatory microenvi- prior to fingolimod administration was analyzed, EAE- ronment as a whole. These results clearly showed that the FTY720 non-responder mice had a remarkable lower therapeutic effect of fingolimod observed is not related to abundance of these cells within the whole blood cells a clinical or immunological bias associated with the indi- than EAE-FTY720 responders prior to treatment ini- vidualized treatment of EAE. tiation (Fig. 4D, E). Importantly, Ly-6Chi cell levels had excellent discriminatory power to assess the risk of Circulating Ly‑6Chi cells represent an excellent biomarker being responder or non-responder based on the median to predict the responsiveness to fingolimod in EAE mice value of the maximum peak clinical score, the percent- Although fingolimod was therapeutically effective in age recovery of the clinical score and the residual clini- most of the EAE-FTY720 mice (responder mice), 3 of the cal score (AUC 0.956, 95% C.I.—confidence interval 18 mice showed a clinical course indistinguishable from 0.853–1.058; p < 0.05). In all cases, we established 2.75% the control EAE-Veh mice, i.e., non-responders (Fig. 4A– of Ly-6Chi cells within the whole leukocyte popula- C). To quantify this effect, we considered non-responder tion at disease onset as the cut-off value indicative of a mice as those with a similar or worse clinical level than future therapeutic responsiveness in EAE mice (100% the median of the EAE-Veh group. Thus, EAE-FTY720 specificity, 95% C.I. 29.24–100%; 86.67% sensitivity, 95% non-responders were defined as those with a maximum C.I. 59.54–98.34%; Fig. 5A, B). By contrast, the mere clinical score ≥ 3, with a clinical recovery score ≤ 25%, abundance of this cell type was not able to discriminate a recovery of the clinical score between the peak and between responder and non-responder EAE-FTY720 the end of the follow-up ≤ 0.5 and a residual clinical mice based on the median accumulated clinical score at score ≥ 2.25. Following this distribution, EAE-FTY720 the end of the experiment (AUC 0.875, 95% CI 0.703– responder mice exhibited a significantly milder disease 1.047; p = 0.092) B (AUC 0.667, 95% CI 0.305–1.028; course than EAE-Veh or EAE-FTY720 non-responders, p = 0.374; Fig. 5C–F). Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 9 of 22 Fig. 2 Ly-6Chi cell abundance at EAE onset is related to milder short- and long-term disease courses. A–D The amount of circulating Ly-6Chi cells at disease onset is inversely correlated with the peak score (A), the total accumulated clinical score (B) and the residual score (C), and it is directly correlated with the % recovery (D) in both EAE-Veh and EAE-FTY720 mice In summary, our data clearly indicated that circu- Circulating Ly‑6Chi cells constitute a good biomarker lating Ly-6Chi cell content at the onset of EAE is an of fingolimod treatment efficacy in EAE extraordinary tool to determine future therapeutic As in clinical practice, examination of EAE responder responsiveness of EAE mice to fingolimod treatment in mice yielded high variability in the efficacy demonstrated terms of the maximum affectation, the extent of recov- by fingolimod treatment in each individual mouse. In ery and the residual score achieved by each individual fact, our data revealed that the higher the number of animal. circulating Ly-6Chi cells prior to treatment, the greater (See figure on next page.) Fig. 3 Ly-6Chi cells are homogeneously represented in EAE-Vehicle and EAE-FTY720 mice before treatment. A Representative plots of Ly-6Chi cells from EAE mice that will be treated with vehicle (left panel) or fingolimod (right panel). B The abundance of Ly-6Chi cells relative to the total cells or to the myeloid component at disease onset was similar between both experimental study groups prior to the beginning of treatment. C Representative histograms from the main phenotypic markers of circulating Ly-6Chi cells at disease onset. D, E The MFI of the different markers of Ly-6Chi cells did not differ prior to onset of treatment Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 10 of 22 Fig. 3 (See legend on previous page.) Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 11 of 22 Table 2 Immunological cells in the blood of EAE mice at disease the difference in the maximum score achieved by EAE- onset FTY720 mice relative to the median score of the EAE- Vehicle FTY720 p value Veh mice (Fig. 6A). Hence, Ly-6Chi cells may be a putative biomarker not only of responsiveness to fingolimod treat- % ­CD3+/Total blood cells 12.94 ± 1.32 16.77 ± 1.63 ns ment, but also to the individual degree of drug efficacy. % ­CD4+/Total blood cells 8.09 ± 0.77 10.91 ± 1.03 < 0.05 To assess this phenomenon, we classified the effect of % CD8 /Total blood cells 3.32 ± 0.30 4.42 ± 0.51 ns fingolimod in each EAE mouse as high or low based on % Ly-6Chi/Total blood cells:% ­CD3+ 0.47 ± 0.13 0.39 ± 0.10 ns the median of the EAE-FTY720 group as a whole. Fin- % Ly-6Chi/Total blood cells:% ­CD4+ 0.72 ± 0.19 0.58 ± 0.15 ns golimod treatment was considered to be high efficacy % Ly-6Chi/Total blood cells:% ­CD8+ 1.83 ± 0.55 1.68 ± 0.50 ns when an EAE-FTY720 mouse had a maximum clinical % Ly-6Chi/Myeloid cells:% ­CD3+ 0.58 ± 0.16 0.50 ± 0.13 ns score < 2.5, a total accumulated clinical score < 19, clini- % Ly-6Chi/Myeloid cells:% ­CD4+ 0.89 ± 0.23 0.75 ± 0.18 ns cal recovery > 72.73%, a recovery index > 0.16 or a resid- % Ly-6Chi/Myeloid cells: % ­CD8+ 2.25 ± 0.65 2.15 ± 0.62 ns ual clinical score < 1.0. The abundance of Ly-6Chi cells The data are expressed as the mean ± SEM. The result of the Student’s t-test or a prior to fingolimod treatment could not discriminate non-parametric Mann–Whitney test is shown: ns not significant between animals prone to present high/low treatment Fig. 4 Circulating levels of Ly-6Chi cells are higher in fingolimod responder EAE mice. A–C Graphic representation of the clinical course of one representative EAE-Veh (A Ly-6Chicells/total blood cells = 2.26% or Ly-6Chi cells/myeloid cells = 3.15%), one non-responder (B Ly-6Chicells/ total blood cells = 2.32%; Ly-6Chicells/myeloid cells = 3.20%) and one responder EAE-FTY720 mouse (C Ly-6Chicells/total blood cells = 4.81% or Ly-6Chicells/myeloid cells = 6.02%). EAE-Veh animals developed a severe clinical course undistinguishable from non-responder EAE-FTY720 mice. D, E Responder EAE-FTY720 mice (N = 3) had a much higher abundance of circulating Ly-6Chicells among the total blood cells (D) or the myeloid component (E) than non-responder EAE-FTY720 mice (N = 15) prior to initiating treatment Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 12 of 22 Table 3 Comparison of the clinical parameters and splenic T cell populations in EAE-Vehicle, and EAE-FTY720 responder and non- responder mice EAE-Vehicle EAE-FTY720 p value Non-responder Responder Clinical parameters Maximum score 3 3 2.25* < 0.01 (2.25–3) (3–3) (1.25–2.63) dpi peak 15 15 15 ns (14–17) (13–15) (14–16) Severity index 0.75 0.75 0.63 ns (0.50–1) (0.60–1) (0.33–0.69) Peak accumulated score 6.5 7 4.5* < 0.05 (5.63–9) (6.50–9.75) (3–6.63) Total accumulated score 37.25 41.38 18* < 0.001 (23–45.38) (34.38–45.50) (10.90–24.75) % Recovery 25 8.33 75* < 0.001 (12.5–33.30) (0–25) (62.50–80.95) Recovery index 0.09 0.04 0.17* < 0.01 (0.04–0.17) (0–0.09) (0.13–0.22) Residual score 2.25 2.75 0.5* < 0.001 (1–2.63) (2.25–3) (0.5–0.75) Spleen % ­CD3+/Total blood cells 15.15 ± 0.97 15.68 ± 3.27 10.13 ± 0.77* < 0.01 % ­CD4+/Total blood cells 9.11 ± 0.64 10.51 ± 2.17 7.20 ± 0.56* < 0.05 % ­CD8+/Total blood cells 4.98 ± 0.36 4.47 ± 0.91 2.46 ± 0.21* < 0.001 % CD3 ­CD25+/Total blood cells 1.32 ± 0.08 1.33 ± 0.25 1.04 ± 0.07* < 0.05 % CD3 ­CD69+/Total blood cells 1.77 ± 0.12 2.24 ± 0.31 1.78 ± 0.10 ns The data are expressed as the median (IQR). One-way ANOVA or a non-parametric ANOVA on ranks tests are shown: dpi day post-immunization; ns not significant; IQR interquartile range *Represents significantly different groups efficacy based on the median of the maximum clinical The responsiveness and efficacy to fingolimod treatment score (AUC 0.765, 95% CI 0.524–1.007; p = 0.058), the in EAE is exclusively influenced by the level of circulating % recovery of the clinical score (AUC 0.700, 95% 0.450– Ly‑6Chi cells 0.950; p = 0.155), or the recovery index (AUC 0.741, 95% The existence of differences in the responsiveness and CI 0.491–0.990; p = 0.085). Interestingly, Ly-6Chi cell efficacy to fingolimod could be the result of the sum of level was seen to be an important parameter to distin- different clinical and immunological factors other than guish EAE mice in which fingolimod acts with high effi- Ly-6Chi cell levels. To test this hypothesis, we performed cacy based on the median value of the total accumulated a Cox proportional hazard regression analysis to deter- clinical score (Fig. 6B) and of the residual score at the mine the likelihood of EAE mice achieving a maximum end of fingolimod administration (Fig. 6D). In fact, hav- clinical score ≥ 3 (median of the EAE-Veh mice, i.e., para- ing more than 5.15% (100% specificity, 99% C.I. 55.50– plegia) or a residual score < 2 as a function of the follow- 100%; 66.67% sensitivity, 99% C.I. 21.91–95.84%; Fig. 6C) ing variables found at disease onset: dpi; clinical score; or 4.59% (100% specificity, 95% C.I. 26.59–100%; 57.14% the abundance of Ly-6Chi cells; C ­ D3+, ­CD4+, ­CD8+ T hi sensitivity, 95% C.I. 22.34–87.33%; Fig. 6E) of circulating cells; the ratios between Ly-6C cells/lymphocyte sub- Ly-6Chi cells at disease onset were valid cut-off points in sets, and some other myeloid cell subsets, such as neu- EAE mice to distinguish those prone to experience high trophils ­(CD11b+Ly-6CintLy-6Ghi) or dendritic cells efficacy of fingolimod treatment in terms of the total ­(CD11b+CD11c+); and the experimental animal group, accumulated clinical score or residual score, respectively. i.e., treated or not with fingolimod. Importantly, Ly-6Chi In summary, the proportion of peripheral blood Ly-6Chi cell abundance at disease onset and the absence of fin- cells at disease onset appeared to be a good biomarker of golimod treatment were the only two variables associated fingolimod efficacy, whereby the higher the level of these with a lower risk of reaching paraplegia or completing cells the less severe the disease after fingolimod treatment. the treatment with a residual clinical score < 2 (p < 0.05 Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 13 of 22 Fig. 5 Ly-6Chi cell content is an excellent biomarker to predict responsiveness to fingolimod in EAE mice. A ROC curve analysis showing that Ly-6Chi cell content presented the highest discriminatory ability for responder and non-responder EAE mice classified using the median of maximum clinical score, the % recovery or the residual score. B Cut-off value for the Ly-6Chicells/total blood cells to predict good responsiveness to fingolimod. C–F The level of Ly-6Chi cells could not discriminate between responder and non-responder EAE-FTY720 mice based on the median of the total accumulated clinical score (C, D) or the recovery index (E, F). AUC​ area under the curve (N = 18 mice) Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 14 of 22 Fig. 6 Ly-6Chi cells are a good biomarker for the efficacy of fingolimod treatment in EAE. A The higher the Ly-6Chicell content at onset, the greater the difference in the maximum score reached by EAE-FTY720 mice compared to the median of the EAE-Veh group. B, D ROC curve analysis showing that Ly-6Chi cells had a good capacity to discriminate high efficacy as opposed to low efficacy EAE mice through the median accumulated clinical score (B) and the residual score (D). C, E Cut-off values for the Ly-6Chi cells/total blood cells to predict those EAE mice that will present a higher total accumulated or a lower residual score than the median of EAE-FTY720 mice at the end of the experimental evaluation Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 15 of 22 in both cases). A high abundance of Ly-6Chi cells at dis- 12 months, the baseline M-MDSC level was correlated ease onset was a clear protective factor (HR 0.0949, 95% to the different criteria considered to define non-evi- C.I. 0.0106–0.851) while the absence of fingolimod treat- dence of disease activity (NEDA-3), i.e., no new relapses; ment was a prominent risk factor for reaching paraple- absence of confirmed disability progression (CDP), gia (HR 11.195, 95% C.I. 1.212–103.397). On the other understood as 1 point increase in the EDSS for patients hand, a high abundance of Ly-6Chi cells at disease onset with a basal EDSS ≤ 5 (1.5 points when baseline EDSS increased the likelihood of completing the treatment with was 0) or of 0.5 points in patients with basal EDSS > 5; little clinical sequelae (HR 1.894, 95% C.I. 1.238–2.898), and no increase in brain MRI activity [a composite whereas not being treated with fingolimod prevented measure of any new T2 and/or gadolinium-enhancing such an outcome (HR 0.149, 95% C.I. 0.0358–0.623). All lesions (Gd+ lesions)] relative to the baseline MRI [14]. these data clearly indicate that the abundance of Ly-6Chi After 12 months of treatment, 74.19% of patients had cells is a good and specific biomarker for fingolimod not experienced relapses, with a significantly lower treatment efficacy in the EAE model. annual relapse rate (ARR) than the year before start- ing fingolimod treatment (1.10 ± 0.18 vs 0.32 ± 0.11; M‑MDSC abundance is independent of the fingolimod p < 0.001). Despite the fact that 77.42% of patients main- to treat MS patients’ characteristics at baseline tained or decreased their EDSS after treatment, no signif- To establish a parallel with the observations in the EAE icant differences were observed in the EDSS score before mouse model, we carried out a translational study in and after fingolimod treatment [3.0 (2.0–4.0) vs 2.5 (2.0– 31 MS patients (demographic data are shown in Addi- 4.0); p = 0.603]. Finally, the proportion of patients with no tional file 2: Table S2) to assess whether circulating new T2 lesions after 12 months of fingolimod treatment M-MDSC abundance at baseline may be related to the was 61.29%, whereas 87.10% of patients did not present future response of MS patients to fingolimod. We first new ­Gd+ lesions after 12 months of follow-up. However, assessed whether the M-MDSC content of the mono- M-MDSC abundance at baseline was not related to the nuclear cell fraction (M-MDSCs/MNCs) at baseline was future clinical course at 12 months after treatment (Addi- independent of the patients’ age (r = 0.0649, p = 0.726), tional file 3: Table S3). ­ d+ enhanc- EDSS (r = − 0.102, p = 0.584), or number of G To address the ability of M-MDSCs to discriminate ing lesions (r = 0.199, p = 0.279). Switching treatment between MS patients who will respond distinctly to fin- from natalizumab (Ntz) to fingolimod has been associ- golimod, “responder” MS patients were considered as ated with a risk of MS reactivation [44–46]. In order to those who met all the NEDA-3 criteria (R-MSNEDA-3), avoid misinterpretation of the M-MDSC abundance due whereas “non-responder” MS patients were those who to the presence of patients with a very active immune met two or less NEDA-3 criteria at the end of the fol- response, we discriminated patients who switched treat- low-up period (NR-MSNEDA-3). Accordingly, 13 of the ment from Ntz (Ntz-patients, N = 11) from those who 31 patients (41.94%) were considered as R-MSNEDA-3. were naïve or who had previously received another first- No relationship was found between the abundance of line treatment (Non-Ntz patients, N = 20). Interestingly, M-MDSCs and the different clinical parameters found the level of M-MDSCs at baseline was similar between after 12 months follow-up in R-MSNEDA-3 (Additional Non-Ntz and Ntz patients (Non-Ntz 1.23 ± 0.30, Ntz file 3: Table S3) and hence, the percentage of M-MDSCs 1.22 ± 0.66; p = 0.397). In fact, the M-MDSCs in both at baseline was not significantly different between the Ntz and Non-Ntz patients remained independent of the R-MSNEDA-3 and NR-MSNEDA-3 subgroups (R-MSNEDA-3 baseline parameters: age (Non-Ntz r = − 0.125, p = 0.594; 1.54 ± 0.56, NR-MSNEDA-3 1.01 ± 0.32; p = 0.447). Ntz r = 0.397, p = 0.210); EDSS (Non-Ntz r = − 0.144, In a second approximation, we restricted the concept p = 0.538; Ntz r = 0.0961, p = 0.755), number of ­ G d+ of “responder” MS patients to those the so-called clini- enhancing lesions (Non-Ntz r = 0.179, p = 0.444; Ntz cal response (R-MSCR), i.e., MS patients with no new r = 0.100, p = 0.755). Therefore, our data indicated that relapses and no CDP after 12 months of treatment [47, M-MDSC abundance at baseline was very clearly inde- 48]. In this sense, 20 of the 31 patients (64.52%) were pendent of demographics at and switching from Ntz considered as R-MSCR. Again, no relationship was found treatment. between the abundance of M-MDSCs and the different clinical parameters found after 12 months follow-up in M‑MDSC abundance at baseline is greater in fingolimod R-MSCR (Additional file 3: Table S3). Moreover, the per- responder patients centage of M-MDSCs at baseline was not significantly To establish whether the abundance of M-MDSCs different between the R-MSCR and NR-MSCR subgroups might be related to the future response to fingolimod at (R-MSCR 1.44 ± 0.39, NR-MSCR 0.85 ± 0.45; p = 0.094). Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 16 of 22 In a next step, we restricted the concept of “responder” are also strongly represented in R-MS patients, irrespec- MS patients (R-MS) to those showing any good therapeu- tive of their demographic and baseline parameters. Fur- tic response (GTR), i.e., MS patients who meet 2 or more thermore, we identified that more circulating M-MDSCs of the NEDA-3 criteria after 12 months of treatment [14, are present at baseline in R-MS patients with no signs of 15]. Accordingly, “non-responder” MS patients (NR-MS) CDP after a 12-month follow-up. were those who met one or none NEDA-3 criteria. In The growing list of DMTs available for the treatment of this sense, 25 out of 31 patients (80.61%) were classified MS [49] makes the identification of predictive biomarkers as R-MS. Again, no relationship was established between for treatment response ever more necessary to help clini- the abundance of M-MDSCs and the clinical parame- cians adopt optimal treatment management strategies for ters after a 12-month follow-up within the R-MS group MS patients [5]. In this context, promising results have (Additional file 3: Table S3). Importantly, R-MS patients been obtained from preclinical experimental procedures, did have a significantly higher percentage of M-MDSCs although the translation of these findings to the clinic is than NR-MS (R-MS 1.48 ± 0.35, NR-MS 0.16 ± 0.06; often disappointing [50, 51]. This could be due to diver- p < 0.05: Fig. 7A, B). As such, the abundance of M-MDSCs gences between EAE and MS, as well as the more simpli- did appear to be a useful tool to identify MS patients with fied interpretation of preclinical observations [52, 53]. To a suboptimal therapeutic response to fingolimod. facilitate the translation of new potential biomarkers dis- Finally, we assessed whether the level of M-MDSCs covered in a preclinical setting into the clinical environ- differed between R-MS and NR-MS for each of the clini- ment, the approximation of the EAE treatment strategy cal criteria used in the definition of NEDA-3. M-MDSCs to the clinical reality of MS should be addressed. A large were unable to distinguish R-MS and NR-MS patients number of studies aimed at evaluating the efficacy of new based on the absence of new relapses or new MRI activ- treatments in EAE administer the drugs to animals pro- ity: relapses—R-MS 1.26 ± 0.35, NR-MS 1.14 ± 0.60 phylactically [54–56]. Conversely, a large amount of pre- (p = 0.701); MRI activity—R-MS 1.50 ± 0.45, NR-MS clinical therapeutic data was obtained by treating at the 0.80 ± 0.29 (p = 0.441: Fig. 7C, D). Importantly, the level same time point all EAE animals once they had reached of M-MDSCs was higher in MS patients with no evidence a certain clinical score, or when only 50–60% of the ani- of an increase in disability (R-MS 1.43 ± 0.34, NR-MS mals had developed clinical signs of EAE [57–59], ignor- 0.17 ± 0.07; p < 0.05: Fig. 7E). ing the inter-individual biological variability in disease In summary, these data displayed a clear parallel with severity found among immunized mice [24]. This type of the preclinical results reported previously in the EAE approach is easier to follow as all animals are treated uni- mice and they showed the ability of M-MDSC abundance formly, monitoring is unified and the experimental pro- to discriminate a good response to fingolimod treatment cedure is completed simultaneously. However, these two in MS patients. Moreover, M-MDSCs were enriched in experimental approximations are far from close to real patients whose disability remained stable or diminished clinical practice. In order to more closely resemble the after 12 months of fingolimod treatment. clinical situation, we performed individualized adminis- tration and clinical assessment of each EAE mouse, and Discussion thus, vehicle or fingolimod treatment commenced at the The present study demonstrates the potential of circulat- time that each mouse began to develop clinical signs of ing Ly-6Chi cells/M-MDSCs as a biomarker of response EAE. Our data regarding the amelioration of the clinical to fingolimod in both EAE mice and MS patients, respec- course, the decrease in T cells and the histopathological tively. From a preclinical point of view, our results show preservation after fingolimod treatment demonstrated that peripheral Ly-6Chi cell content at disease onset is that this kind of experimental approach is as therapeuti- the only immunological variable that allows us to dis- cally valid as other preclinical designs [60, 61]. However, tinguish the presence of a highly effective therapeutic it also adds some valuable information about the inter- response to fingolimod in the EAE model. M-MDSCs individual heterogeneity regarding the efficacy of the (See figure on next page.) Fig. 7 M-MDSCs are more abundant at baseline in patients with a good therapeutic response to fingolimod. A Flow cytometry strategy of analysis for M-MDSCs in human PBMC samples. In both “responder” and “non-responder” MS patients, immature myeloid cells were gated from mononuclear subpopulation as C ­ D33+HLA-DR−/low populations. ­CD33+HLA-DR−/low cells were assessed for CD14 and CD15 expression to identify the M-MDSC subset defined as ­CD14+CD15−. B M-MDSCs were more abundant in R-MS patients (those who meet two or more criteria for NEDA-3: good therapeutic response—GTR) than NR-MS patients. C, D M-MDSCs were similar between R-MS and NR-MS patients according the absence of new relapses (C), or new T2 lesions or ­Gd+ enhancing lesions in MRI after 12 months of treatment (D). E M-MDSCs were clearly higher in patients with stable disability 12 months after fingolimod treatment. R-MS “Responder”-MS patients; NR-MS “non-responder”-MS patients. B R-MS N = 25, NR-MS N = 6; in C R-MS N = 23, NR-MS N = 8; in D R-MS N = 19, NR-MS N = 12; in E R-MS N = 26, NR-MS N = 5 Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 17 of 22 Fig. 7 (See legend on previous page.) Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 18 of 22 molecule tested, something closer to the reality of patient through different experimental strategies increased the management. number and activity of splenic and CNS infiltrated Ly- Our data indicate that the level of Ly-6Chi cells is a good 6Chi cells [32, 43, 69–72]. The seemingly controversial and specific biomarker of clinical efficacy and respon- inverse relationship between the abundance of Ly-6Chi siveness to fingolimod in EAE mice, i.e., it is useful to dis- cells (i.e., inflammatory monocytes/M-MDSCs) and clin- tinguish responder mice, and those that will have a less ical improvement in EAE has been explained by recent severe clinical development and fewer sequelae, following experimental observations. It is currently not possible fingolimod treatment. Ly-6Chi cells are able to distinguish to discriminate circulating Ly-6Chi inflammatory mono- fingolimod-responsive mice on the basis of maximum cytes from M-MDSCs in mice [18, 20, 21, 41, 42]. The clinical score, residual score, and percent recovery, but vast majority of studies explored Ly-6Chi function around fail on the basis of cumulative score or recovery rate. This the peak of EAE disease, when the greatest tissue dam- would indicate that the level of Ly-6Chi cells is adequate age has already occurred and Ly-6Chi cells had acquired to predict the effect of fingolimod associated with attenu- immunosuppressive functions, leading to a decrease in ation of early CNS changes, i.e., reduced influx of inflam- T cell proliferation and Th1/Th17 differentiation [41]. matory monocytes and T cells. However, prediction of However, the deleterious role attributed to Ly-6Chi cells later phenomena more related to clinical recovery, where in EAE results from their study during the pro-inflamma- remyelination and neuroprotection are more relevant, tory phase of this animal model. In fact, an increase in may be confounded by the direct effects of fingolimod on these cells leads to an earlier onset and a worse disease CNS resident cells, such as oligodendrocytes, astrocytes severity [73], whereas their elimination resulted in milder or even neurons [62]. disease courses [74]. The presence of experimental bias resulting from the To interpret these studies, it is important to consider randomized distribution of the groups can be ruled out the true functional consequences of Ly-6Chi cells as they since the timings, the level of clinical affectation and the switch from behaving as antigen presenting cells at EAE immunoregulatory balance between M-MDSCs/T cells onset to becoming T cell suppressors at the peak of the were similar between both experimental groups prior to EAE clinical course [41]. Importantly, some studies iden- treatment. Importantly, our case study is one of the few tified peripheral blood Ly-6Chi cells as precursors of to demonstrate that EAE is very useful to test the respon- M-MDSCs in certain tissues. At the peak of EAE, about siveness and efficacy of a given DMT. Most drugs that 50% of the anti-inflammatory Arg-I+ monocyte-derived are successful in EAE fail to translate adequately into the myeloid cells in the CNS are derived from i­NOS+ pro- clinic. This is due, among other reasons, from not tak- inflammatory cells that colonize this organ at disease ing into account the bimodal distribution of behavioral onset [75]. Moreover, almost 98% of the Arg-I+ anti- responses to drugs [63]. The nomenclature of “respond- inflammatory phagocytes infiltrating the CNS at the peak ers” and “non-responders” has been commonly used of the clinical course are thought to be derived from Ly- in MS patients for some time [14, 64, 65]. However, it 6ChiCCR2+ monocytes that invade the inflamed tissue is anecdotic in the study of new treatments in the EAE from the onset of the disease [76]. Interestingly, it was model [66, 67]. Our individualized study represents shown that all the Arg-I+ cells present in the spinal cord another exception since it mirrors the natural variabil- at the peak of EAE have a phenotype of M-MDSCs and ity that exists among animals in response to fingolimod suppressor activity [77]. Along similar lines, the crucial treatment, similar to that witnessed in MS patients. Data role of pro-inflammatory Ly-6Chi cells in effective repair from this kind of experimental design will help identify at later stages of chronic inflammatory pathologies was so-called “disease classifying biomarkers”, which distin- reported recently [78]. All these data suggest that the vast guish the MS endophenotypes that can yield measurable majority of immunoregulatory phagocytic cells present in predictions in terms of prognosis and response to a cor- the CNS at the peak of EAE would be derived from cir- responding treatment, a phenomenon that is crucial to culating inflammatory monocytes that share their phe- improve future MS trials [68]. notype with M-MDSCs. Therefore, circulating Ly-6Chi To our knowledge, this is the first study that describes monocytes appear to be the only appropriate cell subset the potential of Ly-6Chi cells as biomarkers of response to be analyzed in order to validate M-MDSCs as a predic- to DMTs in EAE. These findings are in agreement with tive tool for the response to fingolimod in EAE. our earlier observations demonstrating a clear relation- The promising preclinical results mentioned above and ship between the abundance of splenic Ly-6Chi cells and the need to find biomarkers that predict the response of the severity of the clinical course of EAE mice [23, 24]. the plethora of medications already approved to treat MS, In addition, it was reported that clinical improvements would make M-MDSCs promising candidates to be used induced in EAE or other experimental models of MS in clinical studies. To our knowledge, this is the first work Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 19 of 22 that interrogates the role of M-MDSCs as predictive bio- Abbreviations AUC​: Area under the curve; CDP: Confirmed disability progression; CFA: markers of fingolimod treatment in MS. The only work Complete Freund’s adjuvant; CI: Confidence interval; CNS: Central nervous that has explored the impact of DMTs on M-MDSCs system; DMSO: Dimethylsulfoxide; DMTs: Disease-modifying treatments; dpo: indicated a non-significant increase in the number of Day post-onset; EAE: Experimental autoimmune encephalomyelitis; EAE-FTY: Fingolimod administered mice; EAE-Veh: Vehicle administered EAE mice; EC: total M-MDSCs in RR-MS patients treated with GA Eriochrome cyanine; EDSS: Expanded Disability Status Scale; FBS: Fetal bovine [32]. In fact, the scarcity of data about M-MDSCs and serum; FTY720: Fingolimod; GA: Glatiramer acetate; Gd: Gadolinium; GTR​ MS are somehow contradictory, and encourages a more : Good therapeutic response; HZ: Hazard ratio; IFN-β: Interferon-beta; IQR: Interquartile range; iv: Intravenously; l-Glu: l-Glutamine; MDSCs: Myeloid- in depth analysis on larger patient cohorts [31, 32]. To derived suppressor cells; M-MDSCs: Monocytic-myeloid-derived suppressor date, the analyses of immune cells as predictive bio- cells; MNCs: Mononuclear cells; MOG: Myelin oligodendrocyte glycoprotein; markers of the response to fingolimod centered on cells MRI: Magnetic resonance imaging; MS: Multiple sclerosis; NDS: Normal donkey serum; NEDA-3: Non-evidence of disease activity-3; NR-MS: Non-responder MS involved in adaptive immunity [14, 15],a s described patients who did not meet 2 or more of the NEDA-3 criteria after 12 months of for Ntz [79] or IFN-β [80]. By contrast, differences in treatment; NR-MSNEDA-3: Non-responder MS patients who did not meet all the cells of the innate immune response between those MS NEDA-3 criteria after 12 months of treatment; Ntz: Natalizumab; P/S: Penicillin/ streptomycin; PBMCs: Peripheral blood mononuclear cells; PBS: Sterile 1× patients who respond efficiently to fingolimod or not has phosphate buffered saline; PFA: Paraformaldehyde; PMN-MDSCs: Polymor‑ been restricted to regulatory ­CD56bright NK cells, with phonuclear-myeloid-derived suppressor cells; R-MS: Responder MS patients a higher abundance in R-MS patients than in NR-MS who MS patients who met 2 or more of the NEDA-3 criteria after 12 months of treatment; R-MSCR: Responder MS patients with no new relapses and no patients [17]. However, the implication of myeloid cells CDP after 12 months of treatment; NR-MSCR: Non-responder MS patients as biomarkers of response to fingolimod has yet to be who either relapses and/or CDP after 12 months of treatment; R-MSNEDA-3: analyzed. In this sense, this study shows for the first time Responder MS patients who met all the NEDA-3 criteria after 12 months of treatment; RR-MS: Relapsing–remitting multiple sclerosis; RT: Room tempera‑ that the abundance of M-MDSCs at baseline is higher in ture; S1PR: Sphingosine-1-phosphate receptor; SMI-32: Non-phosphorylated MS patients that respond well to fingolimod. Moreover, form of the neurofilament protein. the abundance of M-MDSCs in R-MS patients is mainly associated to the absence of CDP. Our results are consist- Supplementary Information ent with our previous data regarding the higher abun- The online version contains supplementary material available at https://​doi.​ dance of circulating M-MDSCs in RR-MS patients at org/​10.​1186/​s12974-​022-​02635-3. relapse relative to the lower EDSS at baseline and after a Additional file 1: Table S1. List of antibodies used for flow cytometry and 1-year follow-up [81]. Therefore, EDSS seems to be the histology. most relevant clinical parameter related to the enrich- Additional file 2: Table S2. Demographic data from the fingolimod ment of immunoregulatory M-MDSCs. In this regard, it patients included in the study. has been shown that the immunosuppressive activity of Additional file 3: Table S3. Correlations between M-MDSCs and clinical M-MDSCs is impaired in patients with progressive MS parameters after 12 months of fingolimod treatment. [31]. Moreover, the level of splenic M-MDSCs in EAE is related to increased myelin and neuronal preservation Acknowledgements [24, 81] and the secretome of M-MDSCs has the abil- The authors would like to thank Dr Virginia Vila-del Sol and Ángela Marquina ity to promote proliferation, survival and differentiation Rodríguez at the Flow Cytometry Core Facility of the Hospital Nacional de Parapléjicos for their assistance with the phenotypic analysis of the different of oligodendrocyte precursor cells [82]. Taken together, cell populations, and Dr José Ángel Rodríguez-Alfaro and Dr Javier Mazarío at our data reinforce the idea that M-MDSCs could be a the Microscopy Core Facility of the Hospital Nacional de Parapléjicos for their good tool to predict the improvement of clinical param- assistance with the confocal imaging and histological quantifications. Fin‑ golimod was kindly supplied by Novartis Pharma AG under a signed Material eters related to neuroprotection and/or neuroreparative Transfer Agreement. enhancement, which requires further experimental study. Author contributions CC-T was a major contributor in writing the manuscript and carried out most Conclusions of the experimental procedures. IM-D and JG-A contributed to the process‑ ing and analysis of the blood samples, both from the EAE mouse model and Our data strongly suggest that the percentage of circulat- fingolimod-treated patients. LC contributed to the processing and analysis ing M-MDSCs seems to be a good parameter to identify of the PBMCs isolated from fingolimod-treated patients. MC performed the patients with a higher probability of responding subopti- histopathological analysis in mice. DO, TC-T, LMV, LC-F, MC and LM isolated baseline (prior to the treatment) PBMCs samples from their own fingolimod mally to fingolimod treatment. This observation paves the patients, and performed the clinical evaluations. MCO contributed to the way for the existence of good biomarkers for monitoring preclinical experimental design and contributed to the editing of the manu‑ or predicting effective response to this or other DMTs script. JMG-D contributed to the clinical experimental design. DC designed the experiments, contributed to the data evaluation and analysis, and was a in the same family (ponesimod, siponimod, ozanimod), major contributor in writing and editing the manuscript. All authors read and which will facilitate their more accurately informed rec- approved the final manuscript. ommendation for patient management in the near future. Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 20 of 22 Funding 5. Callegari I, Derfuss T, Galli E. Update on treatment in multiple sclerosis. La This work was supported by the Instituto de Salud Carlos III (PI18/00357, Presse Médicale. 2021;50(2):104068. https://​doi.​org/​10.​1016/j.​lpm.​2021.​ RD16-0015/0019, PI21/00302, all co-funded by the European Union), the 104068. Fundación Merck Salud (FMS_2020_MS), Esclerosis Múltiple España (REEM- 6. Chalmer TA, Kalincik T, Laursen B, Sorensen PS, Magyari M, Sellebjerg F, EME-S5 and REEM-EME_2018), ADEMTO, ATORDEM and AELEM. CC-T holds et al. Treatment escalation leads to fewer relapses compared with switch‑ a predoctoral fellowship from the Instituto de Salud Carlos III (FI19/00132, ing to another moderately effective therapy. J Neurol. 2019;266:306–15. co-funded by the European Union). LC and JG-A were hired under PI18/00357 https://​doi.​org/​10.​1007/​s00415-​018-​9126-y. and RD16/0015/0019, respectively. DC, MCO and IM-D were hired by SESCAM. 7. Merkel B, Butzkueven H, Traboulsee AL, Havrdova E, Kalincik T. Timing of high-efficacy therapy in relapsing-remitting multiple sclerosis: a system‑ Availability of data and materials atic review. Autoimmun Rev. 2017;16(6):658–65. https://​doi.​org/​10.​1016/j.​ The data sets used and/or analyzed during the current study are available autrev.​2017.​04.​010. from the corresponding author on reasonable request. 8. Mehling M, Kappos L, Derfuss T. Fingolimod for multiple sclerosis: mecha‑ nism of action, clinical outcomes, and future directions. Curr Neurol Neurosci Rep. 2011;11:492–7. https://​doi.​org/​10.​1007/​s11910-​011-​0216-9. Declarations 9. Bascuñana P, Möhle L, Brackhan M, Pahnke J. Fingolimod as a treat‑ ment in neurologic disorders beyond multiple sclerosis. Drugs R D. Ethics approval and consent to participate 2020;20:197–207. https://​doi.​org/​10.​1007/​s40268-​020-​00316-1. All the animal manipulations were approved by the institutional ethical 10. Antel J. Mechanisms of action of fingolimod in multiple sclerosis. Clin Exp committee (Comité Ético de Experimentación Animal del Hospital Nacional Neuroimmunol. 2014;5(1):49–54. https://​doi.​org/​10.​1111/​cen3.​12079. de Parapléjicos), and all the experiments were performed in compliance with 11. Sáenz-Cuesta M, Alberro A, Muñoz-Culla M, Osorio-Querejeta I, the European guidelines for animal research (EC Council Directive 2010/63/EU, Fernandez-Mercado M, Lopetegui I, et al. The first dose of fingolimod 90/219/EEC, Regulation No. 1946/2003), and with the Spanish National and affects circulating extracellular vesicles in multiple sclerosis patients. Int J Regional Guidelines for Animal Experimentation (RD 53/2013 and 178/2004, Mol Sci. 2018;19(8):2448. https://​doi.​org/​10.​3390/​ijms1​90824​48. Ley 32/2007 and 9/2003, Decreto 320/2010). The study with human samples 12. Kuhle J, Disanto G, Dobson R, Adiutori R, Bianchi L, Topping J, et al. was approved by the corresponding local ethics committees of the Ramón y Conversion from clinically isolated syndrome to multiple sclerosis: a large Cajal hospital, Valle d’Hebron Hospital and Donostia Hospital, and informed multicantre study. Mult Scler. 2015;21(8):1013–24. https://​doi.​org/​10.​ written consent was obtained from all the participants in accordance with the 1177/​13524​58514​568827. Helsinki declaration. 13. Arvin AM, Wolinsky JS, Kappos L, Morris MI, Reder AT, Tornatore C, et al. Varicella-zoster virus infections in patients treated with fingolimod: risk Consent for publication assessment and consensus recommendations for management. JAMA Not applicable. Neurol. 2015;72(1):31–9. https://​doi.​org/​10.​1001/​jaman​eurol.​2014.​3065. 14. Teniente-Serra A, Hervás JV, Quirant-Sánchez B, Mansilla MJ, Grau-López Competing interests L, Ramo-Tello C, et al. Baseline differences in minor lymphocyte sub‑ The authors declare no competing financial interests. populations may predict response to fingolimod in relapsing-remitting multiple sclerosis patients. CNS Neurosci Ther. 2016;22(7):584–92. https://​ Author details doi.​org/​10.​1111/​cns.​12548. 1 Neuroimmuno‑Repair Group, Hospital Nacional de Parapléjicos-SESCAM, 15. Quirant-Sánchez B, Hervás-García JV, Teniente-Serra A, Brieva L, Moral-Tor‑ Finca La Peraleda s/n, 45071 Toledo, Spain. 2 Multiple Sclerosis Unit, Biodon‑ res E, Cano A, et al. Predicting therapeutic response to fingolimod treat‑ ostia Health Institute, 20014 Donostia‑San Sebastián, Spain. 3 Neurology ment in multiple sclerosis patients. CNS Neurosci Ther. 2018;24(12):1175– Department, Hospital Universitario Donostia, San Sebastián, Spain. 4 Immu‑ 84. https://​doi.​org/​10.​1111/​cns.​12851. nology Department, Hospital Universitario Ramón y Cajal, Instituto Ramón 16. Quirant-Sánchez B, Presas-Rodriguez S, Mansilla MJ, Teniente-Serra A, y Cajal de Investigación Sanitaria (IRYCIS), Madrid, Spain. 5 Multiple Sclerosis Hervás-García JV, Brieva L, et al. Th1Th17CM lymphocyte subpopulation Unit, Neurology, Ramón y Cajal University Hospital, Madrid, Spain. 6 Neurol‑ as a predictive biomarker of disease activity in multiple sclerosis patients ogy‑Neuroimmunology Service, Centre d’Esclerosi Múltiple de Catalunya under dimethyl fumarate or fingolimod treatment. Mediat Inflamm. 2019. (Cemcat), Institut de Recerca Vall d’Hebron, Hospital Universitari Vall d’Hebron, https://​doi.​org/​10.​1155/​2019/​81478​03. Universitat Autònoma de Barcelona, Barcelona, Spain. 7 Multiple Sclerosis Unit, 17. Moreno-Torres I, González-García C, Marconi M, García-Grande A, Department of Neurology, Hospital General Universitario Gregorio Marañón, Rodríguez-Esparragoza L, Elvira V, et al. Immunophenotype and transcrip‑ Madrid, Spain. tome profile of patients with multiple sclerosis treated with fingolimod: setting up a model for prediction of response in a 2-year translational Received: 24 August 2022 Accepted: 30 October 2022 study. Front Immunol. 2018. https://​doi.​org/​10.​3389/​fimmu.​2018.​01693. 18. Bronte V, Brandau S, Chen SH, Colombo MP, Frey AB, Greten TF, et al. Recommendations for myeloid-derived suppressor cell nomenclature and characterization standards. Nat Commun. 2016;7(1):12150. https://​ doi.​org/​10.​1038/​ncomm​s12150. References 19. Millrud CR, Bergenfelz C, Leandersson K. On the origin of myeloid-derived 1. Walton C, King R, Rechtman L, Kaye W, Leray E, Marrie RA, et al. Rising suppressor cells. Oncotarget. 2017;8(2):3649–65. https://​doi.​org/​10.​ prevalence of multiple sclerosis worldwide: insights from the Atlas of MS, 18632/​oncot​arget.​12278. third edition. Mult Scler J. 2020;26(14):1816–21. https://​doi.​org/​10.​1177/​ 20. Veglia F, Sanseviero E, Gabrilovich DI. Myeloid-derived suppressor 13524​58520​970841. cells in the era of increasing myeloid cell diversity. Nat Rev Immunol. 2. Hauser SL, Cree BAC. Treatment of multiple sclerosis: a review. Am J Med. 2021;21:485–98. https://​doi.​org/​10.​1038/​s41577-​020-​00490-y. 2020;133(12):1380–90. https://​doi.​org/​10.​1016/j.​amjmed.​2020.​05.​049. 21. Zhu B, Bando Y, Xiao S, Yang K, Anderson AC, Kuchroo VK, et al. C ­ D11b+ 3. Liu Z, Liao Q, Wen H, Zhang Y. Disease modifying therapies in relapsing- Ly-6Chi suppressive monocytes in experimental autoimmune encephalo‑ remitting multiple sclerosis: a systematic review and network meta-anal‑ myelitis. J Immunol. 2007;179(8):5228–37. https://​doi.​org/​10.​4049/​jimmu​ ysis. Autoimmun Rev. 2021;20(6): 102826. https://​doi.​org/​10.​1016/j.​autrev.​ nol.​179.8.​5228. 2021.​102826. 22. Mishra MK, Wang J, Silva C, MacK M, Yong VW. Kinetics of proinflamma‑ 4. Rae-Grant A, Day GS, Marrie RA, Rabinstein A, Cree BAC, Gronseth GS, tory monocytes in a model of multiple sclerosis and its perturbation by et al. Practice guideline recommendations summary: disease-modifying laquinimod. Am J Pathol. 2012;181(2):642–51. https://​doi.​org/​10.​1016/j.​ therapies for adults with multiple sclerosis. Neurology. 2018;90(17):777– ajpath.​2012.​05.​011. 88. https://​doi.​org/​10.​1212/​WNL.​00000​00000​005347. 23. Moliné-Velázquez V, Ortega MC, Vila del Sol V, Melero-Jerez C, de Castro F, Clemente D. The synthetic retinoid Am 80 delays recovery in a model of multiple sclerosis by modulating myeloid-derived suppressor cell fate Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 21 of 22 and viability. Neurobiol Dis. 2014;67:149–64. https://​doi.​org/​10.​1016/j.​ 41. Zhu B, Kennedy JK, Wang Y, Sandoval-Garcia C, Cao L, Xiao S, et al. nbd.​2014.​03.​017. Plasticity of Ly-6Chi myeloid cells in T cell regulation. J Immunol. 24. Melero-Jerez C, Alonso-Gómez A, Moñivas E, Lebrón-Galán R, Machín- 2011;187(5):2418–32. https://​doi.​org/​10.​4049/​jimmu​nol.​11004​03. Díaz I, de Castro F, et al. The proportion of myeloid-derived suppressor 42. Damuzzo V, Pinton L, Desantis G, Solito S, Marigo I, Bronte V, et al. cells in the spleen is related to the severity of the clinical course and Complexity and challenges in defining myeloid-derived suppressor cells. tissue damage extent in a murine model of multiple sclerosis. Neurobiol Cytom B. 2015;88(2):77–91. https://​doi.​org/​10.​1002/​cyto.b.​21206. Dis. 2020;140: 104869. https://​doi.​org/​10.​1016/j.​nbd.​2020.​104869. 43. Alabanza LM, Esmon NL, Esmon CT, Bynoe MS. Inhibition of endogenous 25. Liu G, Bi Y, Wang R, Yang H, Zhang Y, Wang X, et al. Targeting S1P1 activated protein C attenuates experimental autoimmune encepha‑ receptor protects against murine immunological hepatic injury through lomyelitis by inducing myeloid-derived suppressor cells. J Immunol. myeloid-derived suppressor cells. J Immunol. 2014;192(7):3068–79. 2013;191(7):3764–77. https://​doi.​org/​10.​4049/​jimmu​nol.​12025​56. https://​doi.​org/​10.​4049/​jimmu​nol.​13011​93. 44. Sempere AP, Martín-Medina P, Berenguer-Ruiz L, Pérez-Carmona N, 26. Li Y, Zhou T, Wang Y, Ning C, Lv Z, Han G, et al. The protumorigenic poten‑ Sanchez-Perez R, Polache-Vengud J, et al. Switching from natalizumab to tial of FTY720 by promoting extramedullary hematopoiesis and MDSC fingolimod: an observational study. Acta Neurol Scand. 2013. https://​doi.​ accumulation. Oncogene. 2017;36:3760–71. https://​doi.​org/​10.​1038/​onc.​ org/​10.​1111/​ane.​12082. 2017.2. 45. Cohen M, Maillart E, Tourbah A, De Sèze J, Vukusic S, Brassat D, et al. 27. Asakura T, Ishii M, Namkoong H, Suzuki S, Kagawa S, Yagi K, et al. Switching from natalizumab to fingolimod in multiple sclerosis. JAMA Sphingosine 1-phosphate receptor modulator ONO-4641 stimulates Neurol. 2014;71(4):436–41. https://​doi.​org/​10.​1001/​jaman​eurol.​2013.​ ­CD11b+Gr-1+ cell expansion and inhibits lymphocyte infiltration in the 6240. lungs to ameliorate murine pulmonary emphysema. Mucosal Immunol. 46. Butzkueven H, Giacomini PS, Cohan S, Ziemssen T, Sienkiewicz D, Zhang 2018;11:1606–20. https://​doi.​org/​10.​1038/​s41385-​018-​0077-5. Y, et al. Safety of fingolimod in patients with multiple sclerosis switched 28. Le Huu D, Matsushita T, Jin G, Hamaguchi Y, Hasegawa M, Takehara from natalizumab: results from TRANSITION-A 2-year, multicenter, obser‑ K, et al. FTY720 ameliorates murine sclerodermatous chronic graft- vational, cohort study. Brain Sci. 2022;12(2):215. https://​doi.​org/​10.​3390/​ versus-host disease by promoting expansion of splenic regulatory brain​sci12​020215. cells and inhibiting immune cell infiltration into skin. Arthritis Rheum. 47. Río J, Nos C, Tintoré M, Téllez N, Galán I, Pelayo R, et al. Defining the 2013;65(6):1624–35. https://​doi.​org/​10.​1002/​art.​37933. response to interferon-β in relapsing-remitting multiple sclerosis patients. 29. Veglia F, Perego M, Gabrilovich D. Myeloid-derived suppressor cells com‑ Ann Neurol. 2006;59:344–52. https://​doi.​org/​10.​1002/​ana.​20740. ing of age review-article. Nat Immunol. 2018;19:108–19. https://​doi.​org/​ 48. Gasperini C, Prosperini L, Tintoré M, Sormani MP, Filippi M, Rio J, et al. 10.​1038/​s41590-​017-​0022-x. Unraveling treatment response in multiple sclerosis: a clinical and MRI 30. Ioannou M, Alissafi T, Lazaridis I, Deraos G, Matsoukas J, Gravanis A, et al. challenge. Neurology. 2019;92(4):180–92. https://​doi.​org/​10.​1212/​WNL.​ Crucial role of granulocytic myeloid-derived suppressor cells in the 00000​00000​006810. regulation of central nervous system autoimmune disease. J Immunol. 49. Wiendl H, Gold R, Berger T, Derfuss T, Linker R, Mäurer M, et al. Multiple 2012;192(3):1334. https://​doi.​org/​10.​4049/​jimmu​nol.​13900​73. Sclerosis Therapy Consensus Group (MSTCG): position statement on 31. Iacobaeus E, Douagi I, Jitschin R, Marcusson-Ståhl M, Andrén AT, Gavin C, disease-modifying therapies for multiple sclerosis (white paper). Ther Adv et al. Phenotypic and functional alterations of myeloid-derived suppres‑ Neurol Disord. 2021. https://​doi.​org/​10.​1177/​17562​86421​10396​48. sor cells during the disease course of multiple sclerosis. Immunol Cell 50. Segal BM, Constantinescu CS, Raychaudhuri A, Kim L, Fidelus-Gort R, Biol. 2018;96(8):820–30. https://​doi.​org/​10.​1111/​imcb.​12042. Kasper LH. Repeated subcutaneous injections of IL12/23 p40 neutralis‑ 32. Cantoni C, Cignarella F, Ghezzi L, Mikesell B, Bollman B, Berrien-Elliott MM, ing antibody, ustekinumab, in patients with relapsing-remitting multiple et al. Mir-223 regulates the number and function of myeloid-derived sclerosis: a phase II, double-blind, placebo-controlled, randomised, dose- suppressor cells in multiple sclerosis and experimental autoimmune ranging study. Lancet Neurol. 2008;7(9):796–804. https://​doi.​org/​10.​1016/​ encephalomyelitis. Acta Neuropathol. 2017;133:61–77. https://​doi.​org/​10.​ S1474-​4422(08)​70173-X. 1007/​s00401-​016-​1621-6. 51. Hartung HP, Kieseier BC. Atacicept: targeting B cells in multiple sclerosis. 33. Webb M, Tham CS, Lin FF, Lariosa-Willingham K, Yu N, Hale J, et al. Sphin‑ Ther Adv Neurol Disord. 2010;3(4):205–16. https://​doi.​org/​10.​1177/​17562​ gosine 1-phosphate receptor agonists attenuate relapsing-remitting 85610​371146. experimental autoimmune encephalitis in SJL mice. J Neuroimmunol. 52. Steinman L, Zamvil SS. How to successfully apply animal studies in 2004;153(1–2):108–21. https://​doi.​org/​10.​1016/j.​jneur​oim.​2004.​04.​015. experimental allergic encephalomyelitis to research on multiple sclerosis. 34. Colombo E, Di Dario M, Capitolo E, Chaabane L, Newcombe J, Martino G, Ann Neurol. 2006;60(1):12–21. https://​doi.​org/​10.​1002/​ana.​20913. et al. Fingolimod may support neuroprotection via blockade of astrocyte 53. T’Hart BA, Jagessar SA, Kap YS, Haanstra KG, Philippens IHCHM, Serguera nitric oxide. Ann Neurol. 2014. https://​doi.​org/​10.​1002/​ana.​24217. C, et al. Improvement of preclinical animal models for autoimmune- 35. Ntranos A, Hall O, Robinson DP, Grishkan IV, Schott JT, Tosi DM, et al. mediated disorders via reverse translation of failed therapies. Drug Discov FTY720 impairs CD8 T-cell function independently of the sphingosine- Today. 2014;19(9):1394–401. https://​doi.​org/​10.​1016/j.​drudis.​2014.​03.​023. 1-phosphate pathway. J Neuroimmunol. 2014;76(3):325–37. https://​doi.​ 54. Rossi S, Lo Giudice T, De Chiara V, Musella A, Studer V, Motta C, et al. Oral org/​10.​1002/​ana.​24217. fingolimod rescues the functional deficits of synapses in experimental 36. Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, et al. autoimmune encephalomyelitis. Br J Pharmacol. 2012;165(4):861–9. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald https://​doi.​org/​10.​1111/j.​1476-​5381.​2011.​01579.x. criteria. Ann Neurol. 2011;69(2):292–302. https://​doi.​org/​10.​1002/​ana.​ 55. Tsai H-C, Huang Y, Garris CS, Moreno MA, Griffin CW, Han MH. Effects of 22366. sphingosine-1-phosphate receptor 1 phosphorylation in response to 37. Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. FTY720 during neuroinflammation. JCI Insight. 2016;1(9): e86462. https://​ New diagnostic criteria for multiple sclerosis: guidelines for research doi.​org/​10.​1172/​jci.​insig​ht.​86462. protocols. Ann Neurol. 1983. https://​doi.​org/​10.​1002/​ana.​41013​0302. 56. Kyran EL, Robinson C, Kocovski P, Li Z, Dang PT, Hale MW, et al. Multiple 38. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an pathological mechanisms contribute to hippocampal damage in the expanded disability status scale (EDSS). Neurology. 1983;13(3):227–31. experimental autoimmune encephalomyelitis model of multiple sclero‑ https://​doi.​org/​10.​1002/​ana.​41013​0302. sis. NeuroReport. 2018;29(1):19–24. https://​doi.​org/​10.​1097/​WNR.​00000​ 39. Seki N, Maeda Y, Kataoka H, Sugahara K, Sugita T, Chiba K. Fingolimod 00000​000920. (FTY720) ameliorates experimental autoimmune encephalomyelitis (EAE): 57. Mangano K, Nicoletti A, Patti F, Donia M, Malaguarnera L, Signorelli S, et al. II. FTY720 decreases infiltration of Th17 and Th1 cells into the central Variable effects of cyclophosphamide in rodent models of experimen‑ nervous system in EAE. Inflamm Regen. 2010;30(6):542–8. https://​doi.​org/​ tal allergic encephalomyelitis. Clin Exp Immunol. 2010;159(2):159–68. 10.​2492/​infla​mmreg​en.​30.​542. https://​doi.​org/​10.​1111/j.​1365-​2249.​2009.​04050.x. 40. Hou H, Cao R, Miao J, Sun Y, Liu X, Song X, et al. Fingolimod ameliorates 58. Wilmes AT, Reinehr S, Kühn S, Pedreiturria X, Petrikowski L, Faissner S, the development of experimental autoimmune encephalomyelitis by et al. Laquinimod protects the optic nerve and retina in an experimental inhibiting Akt-mTOR axis in mice. Int Immunopharmacol. 2016;30:171–8. https://​doi.​org/​10.​1016/j.​intimp.​2015.​11.​024. Camacho‑Toledano et al. Journal of Neuroinflammation (2022) 19:277 Page 22 of 22 autoimmune encephalomyelitis model. J Neuroinflamm. 2018;15(1):183. 76. Locatelli G, Theodorou D, Kendirli A, Jordão MJC, Staszewski O, Phul‑ https://​doi.​org/​10.​1186/​s12974-​018-​1208-3. phagar K, et al. Mononuclear phagocytes locally specify and adapt their 59. Yang T, Zha Z, Yang X, Kang YZ, Wang X, Tong Y, et al. Neuroprotective phenotype in a multiple sclerosis model. Nat Neurosci. 2018;21(9):1196– effects of fingolimod supplement on the retina and optic nerve in the 208. https://​doi.​org/​10.​1038/​s41593-​018-​0212-3. mouse model of experimental autoimmune encephalomyelitis. Front 77. Moliné-Velázquez V, Cuervo H, Vila-Del Sol V, Ortega MC, Clemente D, Neurosci. 2021;15: 663541. https://​doi.​org/​10.​3389/​fnins.​2021.​663541. De Castro F. Myeloid-derived suppressor cells limit the inflammation by 60. Benson CA, Wong G, Tenorio G, Baker GB, Kerr BJ. The MAO inhibitor promoting T lymphocyte apoptosis in the spinal cord of a murine model phenelzine can improve functional outcomes in mice with established of multiple sclerosis. Brain Pathol. 2011;21(6):678–91. https://​doi.​org/​10.​ clinical signs in experimental autoimmune encephalomyelitis (EAE). 1111/j.​1750-​3639.​2011.​00495.x. Behav Brain Res. 2013;252:302–11. https://​doi.​org/​10.​1016/j.​bbr.​2013.​06.​ 78. Rizzo G, Di Maggio R, Benedetti A, Morroni J, Bouche M, Lozanoska- 019. Ochser B. Splenic L­ y6Chi monocytes are critical players in dystrophic 61. Martín-Álvarez R, Paúl-Fernández N, Palomo V, Gil C, Martínez A, Mengod muscle injury and repair. JCI Insight. 2020;5(2): e130807. https://​doi.​org/​ G. A preliminary investigation of phoshodiesterase 7 inhibitor VP3.15 10.​1172/​jci.​insig​ht.​130807. as therapeutic agent for the treatment of experimental autoimmune 79. Signoriello E, Lanzillo R, Brescia Morra V, Di Iorio G, Fratta M, Carotenuto A, encephalomyelitis mice. J Chem Neuroanat. 2017;80:27–36. https://​doi.​ et al. Lymphocytosis as a response biomarker of natalizumab therapeutic org/​10.​1016/j.​jchem​neu.​2016.​12.​001. efficacy in multiple sclerosis. Mult Scler. 2016;22(7):921–5. https://​doi.​org/​ 62. Yazdi A, Ghasemi-Kasman M, Javan M. Possible regenerative effects of 10.​1177/​13524​58515​604381. fingolimod (FTY720) in multiple sclerosis disease: an overview on remy‑ 80. Alenda R, Costa-Frossard L, Alvarez-Lafuente R, Espejo C, Rodríguez- elination process. J Neurosci Res. 2020;98(3):524–36. https://​doi.​org/​10.​ Martín E, de la Maza SS, et al. Blood lymphocyte subsets identify optimal 1002/​jnr.​24509. responders to IFN-beta in MS. J Neurol. 2018;265(1):24–31. https://​doi.​ 63. Jm P, Sz I. Importance of the non-responder in deciphering animal behav‑ org/​10.​1007/​s00415-​017-​8625-6. ior of experimental autoimmune encephalomyelitis. J Mult Scler. 2016;3: 81. Ortega MC, Lebrón-Galán R, Machín-Díaz I, Naughton M, Pérez-Molina I, e108. https://​doi.​org/​10.​4172/​2376-​0389.​1000e​108. García-Arocha J, et al. Myeloid-derived suppressor cells are relevant fac‑ 64. Rudick RA, Lee JC, Simon J, Ransohoff RM, Fisher E. Defining inter‑ tors to predict the severity of multiple sclerosis. bioRxiv. 2022. http://​biorx​ feron β response status in multiple sclerosis patients. Ann Neurol. iv.​org/​conte​nt/​early/​2022/​04/​20/​2022.​04.​20.​488896.​abstr​act. 2004;56(4):548–55. https://​doi.​org/​10.​1002/​ana.​20224. 82. Melero-Jerez C, Fernández-Gómez B, Lebrón-Galán R, Ortega MC, 65. Río J, Comabella M, Montalban X. Predicting responders to therapies for Sánchez-de Lara I, Ojalvo AC, et al. Myeloid-derived suppressor cells sup‑ multiple sclerosis. Nat Rev Neurol. 2009;5(10):553–60. https://​doi.​org/​10.​ port remyelination in a murine model of multiple sclerosis by promoting 1038/​nrneu​rol.​2009.​139. oligodendrocyte precursor cell survival, proliferation, and differentiation. 66. Ian S, Zagon LAH. Low dose naltrexone treatment of established Glia. 2021;69(4):905–24. https://​doi.​org/​10.​1002/​glia.​23936. relapsing-remitting experimental autoimmune encephalomyelitis. J Mult Scler. 2015;2:136. https://​doi.​org/​10.​4172/​2376-​0389.​10001​36. 67. Hammer LA, Zagon IS, McLaughlin PJ. Improved clinical behavior of Publisher’s Note established relapsing-remitting experimental autoimmune encephalo‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑ myelitis following treatment with endogenous opioids: implications for lished maps and institutional affiliations. the treatment of multiple sclerosis. Brain Res Bull. 2015;112:42–51. https://​ doi.​org/​10.​1016/j.​brain​resbu​ll.​2015.​01.​009. 68. Manouchehri N, Zhang Y, Salter A, Hussain RZ, Hartung HP, Hemmer B, et al. Clinical trials in multiple sclerosis: potential future trial designs. Ther Adv Neurol Disord. 2019;12:1756286419847095. https://​doi.​org/​10.​1177/​ 17562​86419​847095. 69. Mecha M, Feliú A, Machín I, Cordero C, Carrillo-Salinas F, Mestre L, et al. 2-AG limits Theiler’s virus induced acute neuroinflammation by modulat‑ ing microglia and promoting MDSCs. Glia. 2018;66(7):1447–63. https://​ doi.​org/​10.​1002/​glia.​23317. 70. Melero-Jerez C, Suardíaz M, Lebrón-Galán R, Marín-Bañasco C, Oliver- Martos B, Machín-Díaz I, et al. The presence and suppressive activity of myeloid-derived suppressor cells are potentiated after interferon-β treat‑ ment in a murine model of multiple sclerosis. Neurobiol Dis. 2019;127:13– 31. https://​doi.​org/​10.​1016/j.​nbd.​2019.​02.​014. 71. Wang Z, Zheng G, Li G, Wang M, Ma Z, Li H, et al. Methylprednisolone alleviates multiple sclerosis by expanding myeloid-derived suppressor cells via glucocorticoid receptor β and S100A8/9 up-regulation. J Cell Mol Med. 2020;24(23):13703–14. https://​doi.​org/​10.​1111/​jcmm.​15928. 72. Chung J-S, Tamura K, Akiyoshi H, Cruz PD, Ariizumi K. The DC-HIL/syn‑ decan-4 pathway regulates autoimmune responses through myeloid- derived suppressor cells. J Immunol. 2014;192(6):2576–84. https://​doi.​ Ready to submit your research ? Choose BMC and benefit from: org/​10.​4049/​jimmu​nol.​13018​57. 73. King IL, Dickendesher TL, Segal BM. Circulating Ly-6C+ myeloid precur‑ • fast, convenient online submission sors migrate to the CNS and play a pathogenic role during autoimmune demyelinating disease. Blood. 2009;113(14):3190–7. https://​doi.​org/​10.​ • thorough peer review by experienced researchers in your field 1182/​blood-​2008-​07-​168575. • rapid publication on acceptance 74. Mildner A, MacK M, Schmidt H, Brück W, Djukic M, Zabel MD, et al. • support for research data, including large and complex data types ­CCR2+Ly-6Chi monocytes are crucial for the effector phase of autoim‑ munity in the central nervous system. Brain. 2009;132(Pt 9):2487–500. • gold Open Access which fosters wider collaboration and increased citations https://​doi.​org/​10.​1093/​brain/​awp144. • maximum visibility for your research: over 100M website views per year 75. Giles DA, Washnock-Schmid JM, Duncker PC, Dahlawi S, Ponath G, Pitt D, et al. Myeloid cell plasticity in the evolution of central nervous system At BMC, research is always in progress. autoimmunity. Ann Neurol. 2018;83(1):131–41. https://​doi.​org/​10.​1002/​ ana.​25128. Learn more biomedcentral.com/submissions