Multiple Sclerosis Resource Centre

Welcome to the Multiple Sclerosis Resource Centre. This website is intended for international healthcare professionals with an interest in Multiple Sclerosis. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Progressive multiple sclerosis cerebrospinal fluid induces inflammatory demyelination, axonal loss, and astrogliosis in mice

Experimental Neurology, pages 620 - 632


Multiple sclerosis (MS) is an autoimmune disease characterized by inflammatory demyelination and neurodegeneration throughout the CNS, which lead over time to a condition of irreversible functional decline known as progressive MS. Currently, there are no satisfactory treatments for this condition because the mechanisms that underlie disease progression are not well understood. This is partly due to the lack of a specific animal model that represents progressive MS. We investigated the effects of intracerebroventricular injections of cerebrospinal fluid (CSF) derived from untreated primary progressive (PPMS), secondary progressive (SPMS), and relapsing/remitting (RRMS) MS patients into mice. We found discrete inflammatory demyelinating lesions containing macrophages, B cell and T cell infiltrates in the brains of animals injected with CSF from patients with progressive MS. These lesions were rarely found in animals injected with RRMS-CSF and never in those treated with control-CSF. Animals that developed brain lesions also presented extensive inflammation in their spinal cord. However, discrete spinal cord lesions were rare and only seen in animals injected with PPMS-CSF. Axonal loss and astrogliosis were seen within the lesions following the initial demyelination. In addition, Th17 cell activity was enhanced in the CNS and in lymph nodes of progressive MS-CSF injected animals compared to controls. Furthermore, CSF derived from MS patients who were clinically stable following therapy had greatly diminished capacity to induce CNS lesions in mice. Finally, we provided evidence suggesting that differential expression of pro-inflammatory cytokines present in the progressive MS CSF might be involved in the observed mouse pathology. Our data suggests that the agent(s) responsible for the demyelination and neurodegeneration characteristic of progressive MS is present in patient CSF and is amenable to further characterization in experimental models of the disease.



  • A novel mouse model for progressive multiple sclerosis is proposed.
  • Progressive MS patient CSF caused inflammatory demyelination in the murine CNS.
  • Th17 cells were involved in the observed CNS lesion pathology.
  • CSF-induced brain lesions are drastically reduced by clinical treatment.
  • MS patient cytokines mediate murine lesion formation.

Abbreviations: CSF - cerebrospinal fluid, CTRL - control, EDSS - expanded disability status scale, ITMTX - intrathecal methotrexate, MS - multiple sclerosis, PPMS - primary progressive multiple sclerosis, SPMS - secondary multiple sclerosis progressive, RRMS - relapsing/remitting multiple sclerosis.

Keywords: Progressive multiple sclerosis, Intracerebroventricular injection, Cerebrospinal fluid, Demyelination lesion pathology, Inflammatory cytokines.


MS is an autoimmune disease of unknown origin, characterized by demyelination and axonal loss throughout the CNS (Gironi et al, 2000 and Compston and Coles, 2008). MS often starts out as a clinically isolated syndrome followed by a series of alternating periods of remission and exacerbation, referred to as RRMS (Hafler, 2004 and Sadiq, 2005). While patients typically return to near normal neurologic function at the end of each episode, over time, failure of the CNS to remyelinate MS lesions ( Franklin, 2002 ) and regenerate axons (Trapp et al, 1998 and Kornek et al, 2000) can lead to an irreversible progression of clinical disability (SPMS) (Ferguson et al, 1997 and Lovas et al, 2000). In addition, 10%–15% of MS patients will have clinical progression from onset without remissions (PPMS) ( Miller and Leary, 2007 ).

At present, the therapeutic options for patients with progressive MS are limited, and no regenerative treatment exists for this condition. In addition, although many animal models have been successfully employed to reproduce specific features of the histopathology and neurobiology of multiple sclerosis, there is no single model that fully captures the entire complexity of progressive MS and its heterogeneity ( Gold et al., 2006 ). Moreover, tissue specimens from patients with progressive MS are not generally available, and post-mortem pathology poorly represents the dynamic biological events related to ongoing disease pathogenesis. In contrast, CSF is readily obtainable and can be studied throughout the course of the disease. These proprieties in conjunction with the fact that molecules secreted by resident and infiltrating cells of the CNS drain into the CSF make it a useful tool to monitor CNS biology and disease activity for progressive MS (Awad et al, 2010, Harris et al, 2013, and Stangel et al, 2013). Therefore, in vivo studies using CSF obtained from progressive MS patients are likely to yield important and new insights into the mechanisms of progressive disease.

We recently reported on a study that combined the use of CSF derived from progressive MS patients and cultured human neural progenitors cells, in order to understand the mechanisms of stem cell-driven CNS repair ( Cristofanilli et al., 2013 ). Here, we established a patient-specific model of investigation based on serial injections of CSF derived from MS patients into the mouse brain. Our data show for the first time that characteristic MS lesion pathology can be induced in mouse brain using patient CSF. We hope that this model, which represents a step forward in understanding the cellular and molecular mechanisms underlying disease progression in MS, can help design more effective treatments for this condition.


Patient selection and CSF collection

CSF was collected with IRB approval and informed consent from 28 patients (20 with clinically definite MS ( McDonald et al., 2001 ) and 8 non-MS controls) ( Table 1 ) seen at the International Multiple Sclerosis Management Practice, the clinical affiliate of the Tisch MS Research Center of New York. Of the 20 MS patients, 8 were secondary progressive, 6 primary progressive, and 6 relapsing remitting. Eight non-MS control (CTRL) CSF samples were obtained for diagnostic purpose from untreated patients with other neurological diseases, including inflammatory (human T lymphotropic virus type-I associated myelopathy and transverse myelitis) and non-inflammatory diseases (spinal cord injury, spinal stenosis, and stroke). All of the MS patients in this study had active disease (see below for definition), and none of them received any immunomodulatory treatment for at least 6 months prior to CSF collection (untreated samples). A second CSF collection was performed in 4 patients who appeared to have stabilized as assessed by the expanded disability status scale (EDSS) after treatment with intrathecal methotrexate (ITMTX) (3 patients) or natalizumab (1 patient). CSF was obtained with sterile techniques either by standard lumbar puncture or by access port aspiration of implanted pumps. CSF samples were processed immediately and kept on ice. Samples were centrifuged at 200 × g for 15 min to remove cells. All samples were confirmed to be free of red blood cell contamination. Aliquots of CSF were stored at − 80 °C until use.

Table 1 Patient demographics.

# of patients (CSF) 6 8 6 8
Age at sample collection, mean (SD), years 50.3 (12.0) 48.1 (9.3) 48.0 (12.4) 45.5 (6.46)
EDSS at sample collection, mean (SD) 6.6 (0.98) 7.5 (0.86) 0.88 (0.74) NA
Disease duration at sample collection, mean (SD), years 13.83 (9.8) 21.7 (7.8) 2.50 (3.7) NA
CSF total protein, mean (SD), μg/ml 653.2 (237.3) 657.6 (242.6) 786.2 (162.1) 499.7 (236.3)
CSF total albumin, mean (SD), μg/ml 399.4 (203.9) 294.7 (95.85) 380.0 (120.0) 302.9 (141.2)
CSF cell count, mean (SD), /ml 2541 (2979) 3652 (7844) 4669 (4919) 2795 (3987)

CSF = cerebrospinal fluid. EDSS = expanded disability status scale. PP = primary progressive. SP = secondary progressive. RR = relapsing, remitting. CTRL = controls.

Clinical assessment of MS patients

All patients in the study had a complete neurological examination at the onset of the study. In addition, routine brain MRI scans were performed on all study subjects 1–2 weeks before CSF collection. Active disease was defined by the presence of any one of the following criteria in the 6 months preceding CSF sample collection: (1) one or more relapses documented by a neurologist's examination; (2) change in 0.5 point or greater in the EDSS score; and (3) change in MRI, specifically a change in the number or size of lesions or the presence of gadolinium-enhancing lesions.

Cannula implantation

Mice (C57/BL6, males) were purchased from The Jackson Laboratory (Bar Harbor, ME). At 8 weeks of age, animals were anesthetized with a solution of ketamine (100 mg/kg) and xylazine (10 mg/kg). They were positioned in a Kopf Small Animal Stereotaxic Instrument such that their heads were stable and immobile. The skull was exposed and a small hole was drilled at the following stereotactic coordinates in reference to the bregma: − 1 mm (anteroposterior axis), 0 mm (medial–lateral axis). A custom made cannula (26 gauge, 5 mm pedestal, cut 3.5 mm below the pedestal; Plastics One, Roanoke, VA) was inserted for its entire length at this position to reach the dorsal third ventricle. The cannula was glued in place using Loctite 454 Instant Adhesive and then with dental cement. Two to three sutures were used to hold the skin around the incision together. A dummy (Plastics One), made to fit the length and gauge of the cannula, was screwed into the guide to keep the passage unobstructed. All animal experiments were approved by the IACUC committee of St Luke's Roosevelt Hospital Center of New York and conformed to NIH guidelines.

CSF injections

Mice were allowed to recover for one week before injections to ensure proper healing of the wound and stability of the cannula. Prior to the injection, the dummy was unscrewed from the guide. An injection needle (Plastics One) was attached to polyethylene tubing (Becton Dickinson, Sparks, MD) and rinsed first with 100% ethanol and then with sterile saline. 22 μl of sterile CSF was loaded into the tubing using a Hamilton Syringe. CSF or saline was injected over the course of 10 min. The needle was kept in place for an additional 5 min to avoid liquid backflow. Injections were given twice a week, for 1, 2, or 4 weeks. Mice were injected daily with BrdU (10 mg/mL) 5 μl/g of body weight from the day of the first CSF or saline injection until sacrifice.

Rota-Rod testing

An accelerating speed Rota-Rod (Panlab/Harvard Apparatus, Holliston, MA, USA) was used to assess locomotive impairment in the mice. Mice were trained twice a day for a week prior to cannula implantation, and then a baseline was measured prior to the first CSF injection for each mouse. Time until fall was measured as the average of three trials the day after each twice weekly injection for four weeks and then normalized against individual baselines.

Mouse sample collection

CSF collection

Samples were collected from all the mice injected with either saline, PPMS-, SPMS-, or CTRL-CSF that survived 8 injections ( Table 2 A). Anesthetized mice were positioned in a Kopf Small Animal Stereotaxic Instrument so that their heads were stable. The skin and muscle covering the cisterna magna were cut and pulled to the side to reveal the membrane of the cisterna magna. A fire-pulled borosilicate glass capillary tube (O.D. 1.5 mm, 0.75 mm, Sutter Instrument Company, Novato, CA) was inserted approximately 1 mm into the membrane and CSF was collected by capillary suction (5–8 μl per sample). Samples were transferred to an Eppendorf tube, flash frozen in liquid nitrogen, and stored at − 80 °C until use.

Table 2 Injection cohort.

  PP SP RR CTRL Saline
A: CSF injections over 4 weeks (twice a week)
# of patients CSF injected 6 8 6 8 NA
# of mice injected 41 45 18 44 16
# of mice surviving 8 injections 38 38 15 37 13
# of surviving mice sacrificed for RNA/protein study 13 14 0 14 7
# of surviving mice sacrificed for pathology 25 24 15 23 6
B: Time course
# of patients CSF injected 2 2      
# of mice injected 20 20      
# of mice sacrificed/dead early due to surgical complication 1 0      
# of mice sacrificed for pathology after 2 injections (1 week) 10 10      
# of mice sacrificed for pathology after 4 injections (2 weeks) 9 10      
C: Treated-CSF injections
# of patients CSF injected 2 2      
# patients treated with ITMTX 1 2      
# patients treated with natalizumab 1 0      
# of mice sacrificed for pathology 18 18      
Protein and RNA extractions

Brain and spinal cord were removed and cut in half longitudinally. One half was placed in RIPA buffer (Cell Signaling Technologies, Danvers, MA) with Halt Protease and Phosphatase Inhibitor Single-Use Cocktail (Thermo Scientific, Rockford, IL) and then sonicated for protein extraction. The other half was processed for RNA extraction using QIAzol Lysis Reagent and RNeasy Lipid Tissue Midi Kit (Qiagen, Venlo, Netherlands).

Animal perfusion

Animals were intracardially perfused first with saline and then with 4% paraformaldehyde (PFA) in 0.1 M PBS, pH 7.4. Brains and spinal cords were post-fixed in formaldehyde for at least 24 h, then processed and embedded in a paraffin cassette. Tissues were sliced at 5 μm with a microtome and mounted onto Unifrost Plus Microscope Slides (VWR, Chester, PA). Slides were dried overnight before staining.

Tissue staining


Microscope slides were baked for 30 min at 37 °C to dry and deparaffinized with a xylene–alcohol scale. Antigen retrieval was performed using antigen unmasking solution from Vector Laboratories (Burlingame, CA). Slides were washed and blocked in 10% normal goat serum (NGS) with 0.2% Triton X-100 for 1 h. Following blocking, tissue was incubated with primary antibodies or the appropriate control (negative control, isotype control, antibody preabsorbed with the relative immunogenic peptide) for 2 h at room temperature. Antibodies used included mouse anti-myelin basic protein (MBP, 1:1000), mouse anti-neurofilament SMI312 (NF, 1:1000), mouse anti-tubulin-β-III (all from Covance, Denver, PA), rabbit anti-glial fibrillary acidic protein (GFAP, 1:500) (Dakocytomation, Glostrup, Denmark), rabbit anti-Iba1 (1:400) (Wako Chemicals, Richmond, VA), rabbit anti-CD3 (1:100) (Abcam, Cambridge, MA), rabbit anti-amyloid precursor protein (APP, 1:100) (Life Technologies, Carlsbad, CA), mouse anti BrdU (1:200) (Developmental Studies Hybridoma Bank, Iowa City, IA). After PBS washes, slides were incubated for 2 h at RT in the dark with the appropriate AlexaFluor conjugated secondary antibodies (Life Technologies) diluted in blocking solution. Slides were washed for 5 min in 4′,6-diamidino-2-phenylindole (DAPI) diluted in PBS (0.5 μg/ml), followed by two PBS washes. Tissue was sealed under a cover slip using Fluoromount (Sigma-Aldrich, St. Louis, MO). Samples were analyzed using a Zeiss LSM 510 Meta Confocal microscope running Zen imaging software at 10 ×, 20 ×, 40 ×, and 63 × magnification (Thornwood, NY).


After incubation with primary antibody (rabbit anti CD 138 at 1:400, and rabbit anti CD19 at 1:400 dilution, Abbiotec, San Diego, CA, and rabbit anti GFAP at 1:500), slides were incubated with horseradish peroxidase-conjugated secondary antibodies (1 h at RT). After developing with DAB (Vector Laboratories, Burlingame, CA) and counterstaining with hematoxylin (Sigma-Aldrich, St. Louis, MO), slides were sealed with Cytoseal Mounting Media.

Luxol fast blue (LFB) and hematoxylin and eosin (H&E) stainings

Microscope slides were baked for 30 min at 37 °C to dry and deparaffinized with a xylene–alcohol scale. Slides were immersed in 2 changes of 95% ethanol for 1 min each, and then placed in LFB solution (0.1% Solvent Blue 38 (Sigma) in 95% Ethanol, + 5 ml 10% glacial acetic acid, filtered) at 58 °C for at least 3 h. Following staining, slides were immersed in 95% alcohol and then distilled water to remove excess stain. Differentiation was accomplished by submerging the slides in 0.05% lithium carbonate for 10–20 s, followed by two changes (1 min each) in 70% ethanol. These steps were repeated if further differentiation was required. Slides were washed in distilled water and then stained for 5 min in Cresyl Violet Solution (0.25% Cresyl violet acetate (Sigma) + 5 drops 10% glacial acetic acid per 300 ml solution, filtered). Slides were then washed twice with 95% ethanol (1 min each) and 100% ethanol (2 min each), cleared in 3 changes of xylene, dried, and then mounted using Cytoseal (Thermo Scientific, Portsmouth, NH). For H&E staining, slides were baked and deparaffinized as above, and then standard procedures were used to stain slides with hematoxylin and eosin in order to visualize generalized histological features.

Quantitative analysis of immunostained images

Brains were cut into 300 sections at 5 μm/section, and 13 sections (1 out of every 25) were stained per mouse. Excluding 3 sections spanning the injection site, a total of ten images per animal spanning 3 mm of the corpus callosum were analyzed. Image J software (Image Processing and Analysis in Java) was used to calculate total cell count for BrdU+and DAPI+cells and to measure mean intensity for GFAP, Iba1, MBP, and NF stains.

For each lesion, the total volume was determined as follows: in every section the lesion was found, the affected cross sectional area was measured. Subsequently, these areas were averaged and the lesion volume (in mm3) was calculated using the following formula:


stripin: si1.gif

For each mouse, the corpus callosum lesioned volume was determined by summing data from all lesions.

Quantitative real time PCR (q-RT-PCR)

q-RT-PCR was performed on the Applied Biosystems 7900 HT Fast Real-Time PCR System using individual TaqMan probes or customized TaqMan low density array and Taqman Gene Expression Master Mix (Applied Biosystems, Foster City, CA). For array PCR, samples were loaded and centrifuged twice at room temperature for 1 min at 1200 rpm. Fold change in RNA levels was calculated using the ΔΔ Ct method (relative quantitation), with the saline group used as the calibrator, and 18S rRNA expression as the internal control. Data were represented on a heat map generated with R (The R Project for Statistical Computing).

Luminex assay

For mouse CSF analysis, samples were pooled together from mice injected with the same patient CSF and mixed with Bio-Plex Pro™ Mouse Cytokine Th17 Panel A 8-Plex (Bio-Rad Laboratories, Hercules, CA). Samples were pooled together by equal volume and were further diluted 1:4 with the manufacturer sample diluent to be analyzed in duplicate. For human CSF analysis, individual samples were mixed undiluted with Bio-Plex Pro™ Human Cytokine 17-Plex Assay (Bio-Rad) and analyzed in duplicate. CSF samples that were found to be blood-contaminated were excluded from the analysis. Both types of CSF samples were run on a BioPlex 200 system (Bio-Rad) according to the manufacturer's protocol. Briefly, 50 μl of each CSF supernatant and various concentrations of each cytokine standard (Bio-Rad) were added to 50 μl of antibody-conjugated beads (Bio-Rad) in a 96-well filter plate (Millipore, Billerica, MA). After a 30 min incubation, the plate was washed and 25 μl of a biotinylated antibody solution (Bio-Rad) was added to each well, followed by another 30 min incubation. The plate was then washed and 50 μl of streptavidin-conjugated phycoerythrin (PE; Bio-Rad) was added to each well and incubated for 10 min. Following a final wash, the contents of each well were resuspended in 125 μl of assay buffer (Bio-Rad) and analyzed using a Bio-Plex Array Reader (Bio-Rad). The cytokine concentrations were calculated by reference to a standard curve for each cytokine derived using various concentrations of the cytokine standards (0.2, 0.78, 3.13, 12.5, 50, 200, 800 and 3200 pg/ml) assayed in the same manner as the CSF samples. The lower detection limit for each cytokine was: 0.19 pg/ml for IL-1β; 0.06 pg/ml for IL-2; 0.02 pg/ml for IL-4; 0.22 pg/ml for IL-5; 0.17 pg/ml for IL-6, IL-7, and TNF-α; 0.13 pg/ml for IL-8, IL-17, and MIP-1β; 0.23 pg/ml for IL-10; 0.27 pg/ml for IL-12p70; 0.09 pg/ml for IL-13; 0.06 pg/ml for G-CSF; 0.31 pg/ml for GM-CSF; 0.24 pg/ml for IFN-γ; and 0.12 pg/ml for MCP-1.

CSF stimulation of cultured T cells and FACS analysis

Cervical, axillary, and inguinal lymph nodes were harvested from CSF injected animals after 8 injections. Samples from the same animal were pooled together and single cell suspensions were prepared using a 70 μM cell strainer. CD4 + cells were isolated using Dynabeads® FlowComp™ Mouse CD4 kit (Life Technologies), plated at 4 × 105 cells/well in 96-well culture plates and cultured in RPMI containing 10% FBS, 1% minimum essential medium, 0.1% beta-mercaptoethanol with (stimulated) or without (unstimulated) 10% CSF derived from the same sample that was injected into the animal. CSF was replenished every other day. After 2 weeks cells were treated with brefeldin A (GolgiPlug; BD Bioscience, San Jose, California), harvested, pre-incubated for 15 min with staining buffer containing BD Fc Block (1 μg/106cells in 100 μl), and surface-stained with anti-CD4-V450 (BD Bioscience). After being fixed and permeabilized with BD Cytofix/Cytoperm™ Fixation/Permeabilization Solution Kit (BD Bioscience) cells were stained with IFN-γ-APC and anti-IL17A-PE (both from BD Bioscience). Isotype-matched IgG were used as a negative control. The stained cells were analyzed using a FACS Aria from BD Bioscience.

Statistical analysis

SPSS was used for statistical analysis. Data were presented as mean ± standard error of the mean (SEM), and one or two-way ANOVA with post-hoc analysis (Tukey HSD or Bonferroni Test) or student t-test was used to assess the significance of the data. P-values < 0.05 were considered statistically significant.


Progressive MS-CSF induces inflammatory-demyelinating lesions in the mouse brain

CSF samples were collected as described in Methods and their cell count, total protein content, and albumin levels were determined. No significant differences were found among groups ( Table 1 ). To investigate the hypothesis that spinal fluid obtained from MS patients could induce MS-like pathology in the mouse CNS, we injected the collected CSF or saline into the third ventricles of mice ( Table 2 ). After four weeks of twice weekly injections, 16 out of 25 animals injected with PPMS-CSF and 16 out of 24 injected with SPMS-CSF developed discrete demyelinating lesions in the corpus callosum ( Table 3 , Fig. 1 and Supplemental Figs. 1 A–D), and at one or more other distinct locations in the brain, such as the cingulum, the dorsal fornix, the fasciculus retroflexus, the lateral ventricles, the primary somatosensory cortex, and the stria medullaris ( Table 4 and Supplemental Fig. 2 ). Only 2 mice out of 15 injected with RRMS-CSF developed similar lesions ( Tables 3 A and 4 ). No lesions were observed when mice were injected with either control-CSF ( Supplemental Fig. 1 E) or saline (data not shown). Regression analysis ( Supplemental Fig. 1 F) showed a correlation between the percentage of lesioned animals and patient EDSS score (R2 = 0.3445). No correlation was found between percentage of lesioned animals and patient age (R2 = 0.0013, data not shown). Inter-patient variation accounted for 8 of the 17 mice injected with progressive MS-CSF that did not develop brain lesions. Specifically, CSF from patients 1PP and 4SP did not cause lesions in any of the mice into which it was injected ( Table 3 A). Excluding these two outlying patients, 78% (32 out of 41) of mice injected with untreated progressive MS-CSF developed brain lesions. Therefore, the remaining 22% of mice did not develop lesions because of intra-sample variation.

Table 3 Lesioned animals among surviving mice.

  Patient CSF code Diagnosis/treatment (if any) Lesioned animals/total surviving animals SC lesioned animals/total surviving animals
A: CSF injections over 4 weeks (twice a week) 1PP Primary progressive MS 0/4 0/4
2PP Primary progressive MS 4/5 1/5
3PP Primary progressive MS 2/3 0/3
4PP Primary progressive MS 3/4 1/4
5PP Primary progressive MS 3/4 0/4
6PP Primary progressive MS 4/5 1/5
Total for primary progressive MS 16/25 3/25
1SP Secondary progressive MS 2/2 0/2
2SP Secondary progressive MS 1/2 0/2
3SP Secondary progressive MS 3/4 0/4
4SP Secondary progressive MS 0/4 0/3
5SP Secondary progressive MS 4/4 0/5
6SP Secondary progressive MS 2/3 0/3
7SP Secondary progressive MS 1/2 0/2
8SP Secondary progressive MS 3/3 0/3
Total for secondary progressive MS 16/24 0/24
1RR Relapsing remitting MS 0/3 0/3
2RR Relapsing remitting MS 1/2 0/2
3RR Relapsing remitting MS 0/2 0/2
4RR Relapsing remitting MS 0/3 0/3
5RR Relapsing remitting MS 1/3 0/3
6RR Relapsing remitting MS 0/2 0/2
Total for relapsing remitting MS 2/15 0/11
B: Time course 2 injections (1 week) 2PP Primary progressive MS 0/5 0/5
6PP Primary progressive MS 0/5 0/5
Total for primary progressive MS 0/10 0/10
5SP Secondary progressive MS 0/5 0/5
8SP Secondary progressive MS 0/5 0/5
Total for secondary progressive MS 0/10 0/10
C: Time course 4 injections (2 weeks) 2PP Primary progressive MS 3/4 0/5
6PP Primary progressive MS 4/5 0/5
Total for primary progressive MS 7/9 0/9
5SP Secondary progressive MS 4/5 0/5
8SP Secondary progressive MS 4/5 0/5
Total for secondary progressive MS 8/10 0/10
D: Treated-CSF injections 2PP Primary progressive MS/ITMTX 0/9 0/9
6PP Primary progressive MS/Natalizumab 2/9 0/9
5SP Secondary progressive MS/ITMTX 1/9 0/9
8SP Secondary progressive MS/ITMTX 0/9 0/9
Total for treated progressive MS 3/36 0/36

Fig. 1 Inflammatory demyelinated lesion in the corpus callosum of a mouse injected for 4 weeks with MS-CSF and relative quantification analysis in all mice. (A) Luxol fast blue staining of a brain section of a mouse injected with MS-CSF. Black arrow indicates demyelinated lesion. (a) Luxol fast blue staining of a brain section for the mouse in A, showing the needle (cannula) track. (B) Tile scan reconstruction of immunostained brain section (C–F merged signals) pinpointing lesion localization (white arrow). (C) MBP immunostaining. (D) Iba1+macrophages. (E) BrdU+proliferative cells (white) over DAPI (blue). (F) C–E merged signals. (G) Higher magnification of lesioned area in F showing Iba1+macrophages-engulfed myelin debris (signal for BrdU+cells omitted). (H) H&E staining showing inflammation and foamy macrophages (black arrowheads) in an adjacent lesioned area (solid black lines). (I–K) Quantitative representation of the mean intensity of (I) Iba1 and (K) MBP immunostainings and total number of (J) BrdU+cells in the corpus callosum of injected mice. Values are normalized to saline group. N = number of animals. Significant P values obtained by one-way ANOVA are reported for each protein [post-hoc analysis:Tukey HSD Test(BrdU and Iba1: P < 0.005 for PP and SP vs. CTRL or RR; MBP: P < 0.005 for PP and SP vs. CTRL or RR and RR vs. CTRL)]. PP = primary progressive; SP = secondary progressive; RR = relapse-remitting; CTRL = control.

Table 4 Brain lesion distribution among lesioned animals.

Patient CSF code Mouse # CC CC lesioned volume (× 10− 3 mm3) CG DF FR LV PSC SM
2PP 1 X 8.75 X         X
2 X 4.65   X        
3 X 7.13 X     X   X
4 X 6.32     X X    
3PP 1 X 3.124 X          
2 X 4.23 X     X    
4PP 1 X 7.77     X X    
2 X 6.18       X    
3 X 5.22       X    
5PP 1 X 4.45   X   X    
2 X 7.66         X  
3 X 8.89   X        
6PP 1 X 9.11 X     X X  
2 X 7.56   X     X  
3 X 10.65   X        
4 X 8.22 X     X    
Average for PPMS (± SDM) = 6.87 (± 2.1) × 10− 3 mm3
1SP 1 X 5.51 X          
2 X 6.18     X      
2SP 1 X 5.34   X   X    
3SP 1 X 7.92            
2 X 8.13   X   X    
3 X 7.54       X    
5SP 1 X 11.33 X         X
2 X 8.68 X         X
3 X 8.52   X        
4 X 9.64       X X  
6SP 1 X 5.21 X          
2 X 3.74       X    
7SP 1 X 2.89 X     X    
8SP 1 X 10.38 X          
2 X 6.25       X    
3 X 10.75   X     X  
Average for SPMS (± SDM) = 7.38 (± 2.5) × 10− 3 mm3
2RR 1 X 4.38 X     X    
5RR 1 X 2.65 X     X    

Average for RRMS (± SDM) = 3.51 (± 1.22) × 10− 3 mm3.

CC = corpus callosum; CG = cingulum; DF = dorsal fornix; FR = fasciculus retroflexus; LV = lateral ventricles; SM = stria medullaris.

Serial section analysis revealed that lesions in the corpus callosum were localized 600 to 1000 μm (on the anteroposterior axis) away from the injection site ( Fig 1 A) with an average volume of 6.87(± 2.1) × 10− 3 mm3and 7.38(± 2.5) × 10− 3 mm3for PPMS and SPMS groups, respectively ( Table 4 ). Due to their size and frequency we chose to further characterize specifically these lesions. Immunostaining revealed the presence of BrdU+proliferative cells ( Figs. 1 E and F), Iba1+macrophages ( Figs. 1 D and F) with engulfed myelin debris ( Fig. 1 G) inside the lesions. Quantitative analysis of the corpus callosum reveals a significant upregulation of Iba1 signal intensity ( Fig. 1 I) as well as the number of BrdU+cells ( Fig. 1 J) in both progressive MS groups compared to RRMS, control and saline groups. The opposite significant trend was found for MBP intensity ( Fig. 1 K). No significant differences were found between PPMS and SPMS for all the above proteins. Within each lesion the vast majority (i.e. 90%) of BrdU+cells were proliferating Iba1+macrophages (data not shown).

Progressive MS-CSF induces axonal loss and astrogliosis in the mouse brain

In progressive MS, demyelinated CNS lesions are often characterized by the loss of axons and the presence of a glial scar ( Trapp et al., 1998 ). To determine if these key pathological features were also present in our CSF-induced lesions, we stained consecutive brain sections containing a demyelinated area (e.g. Fig. 1 ) with markers for axonal integrity and astrogliosis ( Fig. 2 ). Neurofilament staining ( Fig. 2 B) revealed a complete loss of axonal projections within the demyelinated area ( Fig. 2 A). In addition, degenerating axons, visualized by APP staining were present around the lesion edges ( Fig. 2 C) but not inside, confirming the complete axonal loss at the lesion's core. Furthermore, an elaborate net of GFAP+astrocytes resembling a glial scar filled completely the demyelinated area ( Figs. 2 D–E), a pre-requisite for sclerotic scar development seen in MS brain pathology (Smith and Sommer, 1992 and Williams et al, 2007). Quantification of the immunostained corpus callosum showed a significant increase of GFAP ( Fig. 2 G) and a significant decrease of neurofilament signal intensities ( Fig. 2 H) in both progressive MS groups compared to RRMS, control and saline (set as a calibrator). No significant differences were found between PPMS and SPMS.


Fig. 2 Axonal loss, astrogliosis, and immune cell infiltration in MS-CSF induced brain lesion. Axonal loss, degenerative axons, and astrogliosis visualized by immunostaining with (B) neurofilament, (C) APP, and (D) GFAP, respectively. Immunohistochemistical staining for GFAP (brown) in (E) the lesioned area (solid white lines) and in (F) NAWM of the same mouse corpus callosum shown in A–D. Quantitative representation of (G) GFAP and (H) neurofilament immunofluorescent signal intensities in the corpus callosum of injected mice. Values are normalized to saline group. N = number of animals. Significant P values obtained by one-way ANOVA are reported for both proteins [post-hoc analysis:Tukey HSD Test(GFAP: P < 0.005 for PP and SP vs. CTRL or RRMS; NF: P < 0.005 for PP vs. CTRL or RRMS)]. (I) CD3 staining revealing the presence of T cells inside the lesion with scale bar representing the same scale in A–D. (J) CD138+plasma cells (black arrows) and (K) CD19+B cells (black arrows) in the lesioned area (solid white lines). (A) = MBP immunostaining.

T and B cells are involved in CSF-induced lesion pathology

To investigate whether infiltration of peripheral immune mediators was involved in the lesion pathology, we looked for the presence of T ( Fig. 2 I) and B lymphocytes ( Figs. 2 J–K) in lesioned brains. Numerous CD3+T-cells ( Fig. 2 I) were found throughout the lesion area ( Fig. 3 P) but were virtually absent in the normal appearing white matter (data not shown). Similar data were obtained when looking for CD138+antibody-producing plasma B-cells ( Fig. 2 J) and, to a lesser extent, for CD19+immature B-cells ( Fig. 2 K).


Fig. 3 Spinal cord inflammatory demyelinating lesion and locomotor evaluation after 4 weeks of CSF injections. (A) Tile scan reconstruction of immunostained spinal cord section pinpointing lesion localization. (B) MBP immunostaining. (C) Iba1+macrophages. (E) BrdU+proliferative cells (white) over DAPI (blue). (E) B–D merged signals. (F) Axonal loss, (G) astrogliosis, and (H) T cell infiltration visualized by immunostaining for neurofilament, GFAP, and CD3 respectively. (I) H&E staining. (K) Locomotor performance comparison between mice treated with CTRL-CSF, PPMS-CSF, or SPMS-CSF.

MS-CSF only rarely induces typical demyelinating lesions in mouse spinal cord despite extensive inflammation

The vast majority of animals that developed brain lesions also presented with extensive meningeal inflammation (BrdU+/Iba1+cells) spanning every segment of the spinal cord and sporadically with foci of subpial demyelination ( Supplemental Fig. 3 A). However, only a small percentage of those mice developed discrete demyelinated, inflammatory lesions in their spinal cord ( Fig. 3 and Table 3 A). These lesions were much smaller than the ones found in the brain but exhibited similar characteristics such as demyelination ( Fig. 3 B), BrdU+proliferative cells ( Fig. 3 D), macrophages ( Fig. 3 C), axonal loss ( Fig. 3 F), astrogliosis ( Fig. 3 G), and CD3+T-cell infiltration ( Fig. 3 H). However, neither CD138+plasma cells nor CD19+B-cells were found in this type of lesion (data not shown). Interestingly, discrete spinal cord lesions were observed only in mice injected with PPMS-CSF, and no spinal cord inflammation was detected in animals injected with RRMS-CSF (data not shown). Locomotor performance, evaluated by Rota-Rod testing, did not vary among treated groups ( Fig. 3 K). These data are consistent with the fact that, in the vast majority of treated animals, the spinal cord was spared from severe injury.

Demyelination precedes axonal loss and glial scar formation

To further characterize the lesion genesis, we performed a time course experiment, in which animals were injected biweekly with progressive MS-CSF and sacrificed 3 days after the second or fourth injection ( Table 2 B). For this experiment, we selected patient CSF based on the percentage (80% or higher) of mice that had previously developed lesions at 4 weeks post treatment. After 1 week of treatment (2 injections), no lesions were detected in either the brain or spinal cord of treated animals ( Table 3 B). After 2 weeks (4 injections), lesions were seen with demyelination ( Table 3 C and Figs. 4 A, B, and K), macrophages ( Fig. 4 C), BrdU+proliferative cells ( Fig. 4 D), and T-cells ( Fig. 4 J). However, axonal loss within these lesions was incomplete ( Fig. 4 F), as documented by the presence of numerous APP+degenerating axons ( Fig. 4 G) and by the level of NF intensity in the corpus callosum which was significantly higher than it at 4 weeks ( Fig. 4 N). In addition, GFAP+astrocytes were found surrounding the lesion but not inside ( Figs. 4 H–I), suggesting that glial scar formation had not been initiated as confirmed by lower signal intensity for this protein compared to 4 weeks ( Fig. 4 N). Intriguingly, CD138+plasma cells ( Fig. 4 L) and CD19+B-cells ( Fig. 4 M) were also found mainly outside the area of demyelination.


Fig. 4 Inflammatory demyelinated lesion in the corpus callosum of a mouse injected for 2 weeks with progressive MS-CSF. (A) Tile scan reconstruction of immunostained brain section (B–D merged signals) pinpointing lesion localization. (B) MBP immunostaining. (C) Iba1+macrophages. (D) BrdU+proliferative cells (white) over DAPI (blue). (E) H&E staining showing inflammation and foamy macrophages (black arrowheads) in the lesioned area (solid black lines) in the same mouse. (F) Axonal loss, (G) degenerative axons, and (H) astrogliosis, visualized by immunostaining for neurofilament, APP, and GFAP, respectively. (I) Immunohistochemical staining for GFAP (brown) showing reactive astrocytes surrounding the lesioned area. (J) CD3 + T cells (red) infiltrate the lesion. Scale bar is also representative of the magnification of images in panels B, C, D, F, G, and H. (K) Luxol fast blue staining. (L) CD138+plasma cells (black arrows) and (M) CD19+B cells (black arrows) near the lesioned area (solid white lines). (N) Quantification of the expression of BrdU, GFAP, Iba1, MPB, and NF in the corpus callosum of 2 week mice (number of animals examined = 19; PPMS and SPMS groups combined) normalized to mice injected with the same patient CSF for 4 weeks (number of animals examined = 17; PPMS and SPMS groups combined). Asterisks denote a statistically significant difference between the two time points as determined by student t-tests [*P < 0.05; **P < 0.01].

To determine whether treatment with CSF, derived from progressive MS patients or control individuals, affects gene expression in the mouse CNS, RNA was extracted from the brain and spinal cord of a subset of mice that received 8 CSF injections ( Table 2 AB). After collection, individual RNA samples were pooled together for each treated group and analyzed in triplicate (technical replicates) by q-RT-PCR using a custom TaqMan low density array ( Fig. 5 A) containing pre-hybridized probes to detect the RNA expression of 47 genes. Compared to saline, CSF groups showed a similar transcriptional profiling in the brain, whereas in the spinal cord, gene expression profiling of mice injected with PPMS-CSF was unique. In both brain and spinal cord, genes that presented the largest variation among groups (± 1 fold compared to saline) were associated with the immune system and included: T-cell (CD3, CD4, CD28), B-cell (CD19), and macrophage (CD11c, Csf2) receptors; cytokines (IL-17, IL-4); chemokines (CXCL13); transforming growth factor beta-1 (Tgfb-1); and interferon-gamma (Ifng). For most of these genes, their expression changes were greater in the progressive MS-CSF injected groups compared to control. To perform statistical analysis for the above selected genes, we ran individual monoplex q-PCR assays for each gene. For this experiment, samples derived from mice injected with the same patient CSF were pooled together before the assay such that each sample represented an individual patient for a total of 6 PP, 8 SP, and 8 control samples (biological replicates). For both brains and spinal cords, similar expression trends to the ones reported in Fig. 5 A were observed for all the genes examined. However, in brain samples, due to high intra-group variations, no statistical significance was achieved (data not shown). Interestingly, in the spinal cords significant upregulation of T cell specific genes CD3 and CD4, and the B cell chemoattractant CXCL13 was reported for PPMS compared to SPMS and CTRL ( Fig. 5 B).


Fig. 5 RNA level comparisons after 4 weeks of CSF injections. (A) Heat map representation of gene expression in brains and spinal cords of animals treated with PPMS-CSF, SPMS-CSF, or CTRL-CSF. Values are the average of 3 technical replicates normalized to saline (calibrator) and represented on a logarithmic scale. (B) Gene expression comparison in spinal cord of CSF-injected animals for the subset of genes shown in (A) that displayed statistically significant variation between groups. Significant P values obtained by one-way ANOVA are reported for each gene [post-hoc analysis:Tukey HSD Test(CD3, CD4, CXCL13: P < 0.005 for PPMS vs. SPMS or CTRL)].

Progressive MS-CSF treatments enhanced Th17 cell activity in the mouse CNS and in its periphery

Based on our immuno-pathology and RNA data suggesting T cell involvement in the above described mouse pathology, we investigated the CNS and peripheral expression of Th1 and Th17 cells, which are believed to be key modulators in MS. In the CNS, the expression of cytokines associated with the Th1/Th17 pathway was analyzed by Luminex-ELISA in brains, spinal cords, and mouse CSF ( Figs. 6 A–C). As described above for the individual monoplex assay, samples derived from mice injected with the same patient CSF were pooled together before the assay. No significant differences between CSF-injected groups were found in either brain ( Fig. 6 A) or spinal cord ( Fig. 6 B). However, when the mouse CSF was analyzed, a significant increase in the expression of IL-6 and IL-17, two Th17 associated proteins, and of IL-12p40, an antagonist of Th1 activation (Kato et al, 1996, Rothe et al, 1997, and Kalinski et al, 2001) was observed in animals injected with CSF derived from patients with progressive MS compared to controls ( Fig. 6 C) or saline (data not shown). In addition, although higher expression of IL-10, which is secreted by several types of T cells including Th1, was detected in the progressive MS-CSF treated animals, the expression of Th1 associated proteins IL-12p70 and interferon-gamma (IFN-γ) was found to be below the assay detection limit. These data suggest that, in our animal model of progressive MS, Th17 rather than Th1 cells were active in the mouse CNS. To investigate if activation of the Th17 pathway was also present in the peripheral immune system, we analyzed the ratio between Th17 and Th1 cells by immunostaining and FACS analysis of cultured lymph node cells derived from CSF injected animals and stimulated in vitro with the same patient CSF injected into the animal. Th17 cells were found to be significantly more abundant in samples derived from mice injected with PPMS-CSF compared to with control CSF or saline with or without in vitro stimulus ( Fig. 6 D). A similar trend was also reported for the SPMS samples after in vitro CSF-stimulation although these differences were not statistically significant. A post-hoc analysis for each individual patient's CSF revealed that animals that developed CNS lesions had a higher Th17 cell count, with or without stimulus, compared to those that did not (data not shown). In addition, the Th17/Th1 ratio was increased in both progressive MS-CSF treated groups after in vitro stimulation ( Fig. 7 E) whereas control and saline groups remained unaltered.


Fig. 6 Effect of CSF injections on Th1–Th17 immune responses in the mouse CNS and in its periphery. (A–C) Protein level comparisons in the CNS. (A) brain, (B) spinal cord, and (C) mouse CSF. Black horizontal bars represent control values (set as calibrator). Error bars represent SEM. **P = 0.008, one-way ANOVA [post-hoc analysis:Tukey HSD Test(IL-6, IL-17: P < 0.001 for PP or SP vs. control; IL-10, IL-12p40: P < 0.005 for PP or SP vs. control)]. BDL = below detection limit. N = number of biological replicates examined. (D–E) Th1–Th17 peripheral immune responses. (D) T cell subtype comparisons between CSF injected animals with (++) or without (−−) in vitro stimulus. (E) Th17/Th1 cell ratio comparisons between in vitro stimulated and unstimulated samples. Error bars represent SEM. *P = 0.03, two-way ANOVA [post-hoc analysis:Bonferroni Test(Th17−−: P < 0.001 for PP vs. saline; P < 0.005 for PP vs. SP and CTRL. Th17++: P < 0.001 for PP vs. CTRL and saline)]. N = number of mice examined. SP = secondary progressive; PP = primary progressive; CTRL = control.


Fig. 7 Protein level comparisons of pro and anti-inflammatory cytokines present in the injected human CSF cohort. (A) Comparisons between CSF samples derived from untreated patients. Error bars represent SEM. Significant P values obtained by one-way ANOVA are reported for each cytokine [post-hoc analysis:Tukey HSD Test(IL-17: P < 0.005 for PPMS, SPMS, or RRMS vs. CTRL; IL-6: P < 0.005 for PPMS vs. CTRL or RRMS; IL-8 and MIP-1β: P < 0.005 for PPMS vs. CTRL)]. The number of biological replicates examined was: 6 for PPMS and RRMS; 8 for SPMS and CTRL. (B) Comparisons between CSF samples derived from progressive MS patients before and after treatment. Significant P values obtained by student t-test are reported for each cytokine. The number of biological replicates examined was 4 (2 PPMS and 2 SPMS combined post-assay). PPMS = primary progressive MS; SPMS = secondary progressive MS; CTRL = control.

CSF-induced brain lesions are drastically reduced by clinical treatment and correlate with patient CSF levels of pro-inflammatory cytokines

To determine whether long term treatment of MS patients had an effect on the above described lesion pathology, we collected CSF from treated patients whose untreated CSF was used previously in our study. Specifically, we obtained CSF from four patients whose CSF, while untreated, induced lesions in the vast majority of injected mice and was used in the time course experiment described above. Three of these patients were treated for a year with ITMTX, while the last one received natalizumab for 18 months prior to CSF collection ( Table 2 C). None of the mice developed spinal cord lesions ( Table 3 D). Only 1 out of 27 animals injected with ITMTX treated CSF and 2 out of 9 animals injected with natalizumab treated CSF showed distinct brain lesions ( Table 3 D). Furthermore, the number of BrdU+cells in the corpus callosum, as well as the signal intensity for GFAP, Iba1, MBP and NF was found to revert to the control/saline levels reported in the untreated study (data not shown).

To investigate which factor(s) in the untreated patient's CSF might have been involved in the mechanism behind lesion formation, we compared the expression level of 17 human cytokines associated with either pro- or anti-inflammatory immune responses in our CSF cohort of untreated and treated patients by Luminex-ELISA. The expression of 10 cytokines (IL-1β, IL-2, IL-4, IL-5, IL-10, IL-12p70, G-CSF, GM-CSF, IFN-γ, and TNF-α) was found below the assay detection limit in the entire cohort (data not shown). In the untreated cohort ( Fig. 7 A), of the 7 detectable cytokines, IL-17 was significantly higher in the control group compared to each MS subtype with PPMS being the lowest. The opposite trend was found for IL-6, IL-8, MCP-1 (CCL2) and MIP-1β (CCL4) with significant differences found between PPMS vs. CTRL or RRMS for IL-6 and PPMS vs. CTRL for IL-8 and MIP-1β expression. IL-7 and IL-13 showed higher expression in the SPMS, but differences among groups were not significant. In the treated cohort, data from PPMS and SPMS samples were grouped together before comparisons to increase statistical power and then compared with their matched untreated samples. Interestingly, the CSF levels of IL-6, IL-8, MCP-1, and MIP-1β were significantly reduced after patient's treatment to those of controls (CTRL, shown in Fig. 7 A), whereas CSF levels for IL-17, IL-7, and IL-13 were unchanged ( Fig. 7 B).


In the last two decades, the gradual introduction of disease-modifying agents has significantly changed the outcome of RRMS patients ( Hauser et al., 2013 ). By contrast, the mechanisms of disease progression are poorly defined, and the treatment options for progressive forms of MS, characterized by the steady accumulation of irreversible disability, are relatively inadequate ( Hawker et al., 2009 ). These shortcomings are perhaps because there are no good clinical or experimental paradigms to predict if and when a transition from RR to SPMS will occur or if a patient will develop PPMS after an initial clinically isolated syndrome ( Rovaris et al., 2006 ). Indeed, even serial conventional MRI brain scans which are widely accepted as surrogate markers of disease activity for RRMS are not sensitive enough to monitor the progressive phase of MS ( Barkhof et al., 2009 ), and more specific in vivo markers of neurodegeneration are needed for a reliable assessment of putative new therapeutic options. In addition, the fact that no single available animal model can replicate every aspect of progressive MS contributes to our incomplete understanding of the immunological and pathological mechanisms and to the lack of effective treatments for this condition. In this study, we investigated whether an experimental model representative of progressive MS could result from using human CSF as a pathology initiating agent. We showed that spinal fluid derived from patients with progressive forms of MS injected intracerebroventricularly into mice results in the development of discrete lesions of inflammatory demyelination. Time course analysis revealed that demyelination preceded axonal loss and was associated with macrophage proliferation and T-cell infiltration and only later did complete axonal loss, astrogliosis and B-cell/plasma cells infiltration occurred. Lesions were rarely observed when animals were injected with RRMS-CSF and did not occur with inflammatory or non-inflammatory control-CSF or saline injections. The difference in induced lesion frequency between progressive MS-CSF and RRMS-CSF treated animals is likely a reflection of the known clinical and pathological differences between the two forms of the disease. RRMS is characterized by disease exacerbations and remissions, and thus, the CSF represents a pool of pro-inflammatory and anti-inflammatory or regulatory healing factors whereas, in progressive disease, the CSF is a product of the relentless disease worsening. Characterization of the cellular basis of this lesion formation showed that the Th17 but not the Th1 pathway was stimulated by CSF treatments in the mouse CNS and lymph nodes. These data are consistent with the emerging literature that shows that pathogenic Th17 cells play a major role in MS disease pathogenesis (Matusevicius et al, 1999, Kebir et al, 2007, Han et al, 2008, and Nylander and Hafler, 2012). Finally, to better understand the mechanism underlying the induction of mouse pathology by the progressive MS-CSF, we investigated whether patient treatment might affect lesion formation. We found that injections of CSF from treatment responders (with ITMTX or natalizumab) resulted in a markedly reduced number of lesioned animals and in a drastic attenuation of the cellular components (macrophages, reactive astrocytes, immune cells) underlying lesion pathology. These data together suggest that the factors involved in lesion formation are favorably altered with treatment. Cytokine profiling of the injected treated and untreated CSF suggested that at least some of these factors are likely to be pro-inflammatory molecules differentially expressed in our CSF cohort. In particular, CCL2, CCL4, IL-6, and IL-8, which were found to be higher in both types of progressive MS-CSF compared to RRMS and control-CSF, and then reduced to control level after treatment, are likely to be involved in the lesion pathology. These findings are in agreement with what has been reported by others. For instance, CSF levels of CCL4 have been found to be higher in PPMS (but not in RRMS) vs. non-inflammatory neurological diseases ( Matsushita et al., 2013 ), and strong staining of this cytokine has been shown in macrophages within MS plaques ( Simpson et al., 1998 ). Similar reports have been found for the CSF expression of CCL2, IL-6 and IL-8 (Navikas et al, 1996, Stelmasiak et al, 2000, Ishizu et al, 2005, and Edwards et al, 2013). Interestingly, the fact that the highest levels of these cytokines were found in PPMS-CSF, together with higher gene expression of the pro-inflammatory chemokine CXCL13 in the PPMS-injected mouse spinal cord, could potentially explain the known propensity for spinal cord involvement in PPMS. This notion was also supported by finding spinal cord lesions only in mice injected with PPMS-CSF. As a possible mechanism of lesion formation, we propose that the synergistic effects of the aforementioned human cytokines induce inflammatory pathways such as the Th17 one, which ultimately cause demyelinating pathology in mice. Based on this hypothesis, the higher level of human IL-17 found in control CSF, which did not cause lesions, can be explained by the inability of this human cytokine to cross react with the mouse or the lack of other synergistic factors compared to progressive MS-CSF necessary to induce lesion pathology.

Attempts to transfer the pathophysiology of MS into animals by the use of patient CSF (cells and/or supernatant) have been made in the past. The most prominent studies, which relied on the use of a single injection of RRMS-CSF (collected during disease exacerbation) into the cisterna magna of immunocompromised (SCID) mice, resulted in a series of controversial reports that raised doubts about the reproducibility of the data (Saeki et al, 1992, Hao et al, 1994, Sakoda et al, 1994, Jones et al, 1995, and Fujimura et al, 1997). Our study establishes for the first time that acellular CSF derived from untreated progressive MS patients can cause typical MS-like pathological lesions in the mouse CNS.

In our disease model, no discernible clinical locomotor or other impediments were observed in mice. However, our data clearly shows the development of definite pathological lesions seen in the white matter of their brains at locations distinct from the injection sites of CSF. These findings are novel and, if validated, will provide a biological model that may help to better understand the mechanisms underlying lesion formation and the pathophysiology associated with MS disease progression. Furthermore, this unique model may be employed to discover novel biomarkers for CNS inflammation, demyelination, axonal degeneration as well as disease activity and progression. In addition, this model could be used to evaluate MS patient treatment outcomes in mice and to guide customized therapeutic regimens, which could ultimately lead to the development of specific therapies targeting PPMS and SPMS.

The following are the supplementary data related to this article.

Download file

Supplementary Fig. 1 Demyelinating lesions appear only in the corpus callosum of mice injected with MS-CSF. (A–D) Demyelination visualized by the absence of Luxol fast blue staining (arrows) in brain sections of mice injected with MS-CSF. (E) Brain section of a mouse injected with control-CSF. (F) Regression analysis between patient EDSS score (x-axis) and percentage of lesioned animals (y-axis). R = correlation coefficient.

Download file

Supplementary Fig. 2 Other inflammatory demyelinated areas in the brain of a mouse injected for 4 weeks with MS-CSF. MBP, Iba1, BrdU, and DAPI staining are represented in green, red, white, and blue respectively. Black arrows indicate lesioned areas. CC = corpus callosum.

Download file

Supplementary Fig. 3 Spinal cord inflammation after 4 weeks of progressive MS-CSF injections. Meningeal inflammation visualized by (A) immunostaining for Iba1 (red) and BrdU (white) and by (B) H&E spanning the cervical, thoracic, and lumbar segments of the spinal cord of a mouse injected with progressive MS-CSF. Arrows indicate foci of subpial demyelination. (C) H&E staining of control spinal cord.


This work was supported by The Emerald Foundation (grant to MC) and by The Tisch MS Research Center of New York (private funds). We would like to thank all the clinicians and nurses of the International Multiple Sclerosis Management Practice for patient CSF collection and all of the patients that donated their CSF for this study. We also thank Xinhe Liu, from our center, for assisting with the animal surgical procedures and Jonathan Davila, Postdoctoral Fellow at Stanford University, for his help with the generation of the heat map used in Fig. 5 .


  • Awad et al., 2010 A. Awad, B. Hemmer, H.P. Hartung, B. Kieseier, J.L. Bennett, O. Stuve. Analyses of cerebrospinal fluid in the diagnosis and monitoring of multiple sclerosis. J. Neuroimmunol.. 2010;219:1-7
  • Barkhof et al., 2009 F. Barkhof, P.A. Calabresi, D.H. Miller, S.C. Reingold. Imaging outcomes for neuroprotection and repair in multiple sclerosis trials. Nat. Rev. Neurol.. 2009;5:256-266 Crossref
  • Compston and Coles, 2008 A. Compston, A. Coles. Multiple sclerosis. Lancet. 2008;372:1502-1517 Crossref
  • Cristofanilli et al., 2013 M. Cristofanilli, B. Cymring, A. Lu, H. Rosenthal, S.A. Sadiq. Cerebrospinal fluid derived from progressive multiple sclerosis patients promotes neuronal and oligodendroglial differentiation of human neural precursor cells in vitro. Neuroscience. 2013;250C:614-621 Crossref
  • Edwards et al., 2013 K.R. Edwards, J. Goyal, T. Plavina, J. Czerkowicz, S. Goelz, A. Ranger, D. Cadavid, J.L. Browning. Feasibility of the use of combinatorial chemokine arrays to study blood and CSF in multiple sclerosis. PLoS One. 2013;8:e81007 Crossref
  • Ferguson et al., 1997 B. Ferguson, M.K. Matyszak, M.M. Esiri, V.H. Perry. Axonal damage in acute multiple sclerosis lesions. Brain. 1997;120(Pt 3):393-399 Crossref
  • Franklin, 2002 R.J. Franklin. Why does remyelination fail in multiple sclerosis?. Nat. Rev. Neurosci.. 2002;3:705-714 Crossref
  • Fujimura et al., 1997 H. Fujimura, Y. Nakatsuji, S. Sakoda, K. Toyooka, Y. Okuda, H. Yoshikawa, M. Kaido, Y. Saeki, T. Mima, T. Kishimoto, T. Yanagihara. Demyelination in severe combined immunodeficient mice by intracisternal injection of cerebrospinal fluid cells from patients with multiple sclerosis: neuropathological investigation. Acta Neuropathol.. 1997;93:567-578 Crossref
  • Gironi et al., 2000 M. Gironi, A. Bergami, E. Brambilla, F. Ruffini, R. Furlan, G. Comi, G. Martino. Immunological markers in multiple sclerosis. Neurol Sci. 2000;21:S871-875
  • Gold et al., 2006 R. Gold, C. Linington, H. Lassmann. Understanding pathogenesis and therapy of multiple sclerosis via animal models: 70 years of merits and culprits in experimental autoimmune encephalomyelitis research. Brain. 2006;129:1953-1971 Crossref
  • Hafler, 2004 D.A. Hafler. Multiple sclerosis. J. Clin. Invest.. 2004;113:788-794
  • Han et al., 2008 M.H. Han, S.I. Hwang, D.B. Roy, D.H. Lundgren, J.V. Price, S.S. Ousman, G.H. Fernald, B. Gerlitz, W.H. Robinson, S.E. Baranzini, B.W. Grinnell, C.S. Raine, R.A. Sobel, D.K. Han, L. Steinman. Proteomic analysis of active multiple sclerosis lesions reveals therapeutic targets. Nature. 2008;451:1076-1081 Crossref
  • Hao et al., 1994 Q. Hao, T. Saida, M. Nishimura, K. Ozawa, K. Saida. Failure to transfer multiple sclerosis into severe combined immunodeficiency mice by mononuclear cells from CSF of patients. Neurology. 1994;44:163-165
  • Harris et al., 2013 V.K. Harris, N. Donelan, Q.J. Yan, K. Clark, A. Touray, M. Rammal, S.A. Sadiq. Cerebrospinal fluid fetuin-A is a biomarker of active multiple sclerosis. Mult. Scler.. 2013;19(11):1462-1472 10.1177/1352458513477923 Crossref
  • Hauser et al., 2013 S.L. Hauser, J.R. Chan, J.R. Oksenberg. Multiple sclerosis: prospects and promise. Ann. Neurol.. 2013;74(3):317-327 10.1002/ana.24009 (Review)
  • Hawker et al., 2009 K. Hawker, P. O'Connor, M.S. Freedman, P.A. Calabresi, J. Antel, J. Simon, S. Hauser, E. Waubant, T. Vollmer, H. Panitch, J. Zhang, P. Chin, C.H. Smith. Rituximab in patients with primary progressive multiple sclerosis: results of a randomized double-blind placebo-controlled multicenter trial. Ann. Neurol.. 2009;66:460-471 Crossref
  • Ishizu et al., 2005 T. Ishizu, M. Osoegawa, F.J. Mei, H. Kikuchi, M. Tanaka, Y. Takakura, M. Minohara, H. Murai, F. Mihara, T. Taniwaki, J. Kira. Intrathecal activation of the IL-17/IL-8 axis in opticospinal multiple sclerosis. Brain. 2005;128:988-1002 Crossref
  • Jones et al., 1995 R.E. Jones, Y. Chou, A. Young, M. Mass, A. Vandenbark, H. Offner, D. Bourdette. T cells with encephalitogenic potential from multiple sclerosis patients and Lewis rats fail to induce disease in SCID mice following intracisternal injection. J. Neuroimmunol.. 1995;56:119-126
  • Kalinski et al., 2001 P. Kalinski, P.L. Vieira, J.H. Schuitemaker, E.C. de Jong, M.L. Kapsenberg. Prostaglandin E(2) is a selective inducer of interleukin-12 p40 (IL-12p40) production and an inhibitor of bioactive IL-12p70 heterodimer. Blood. 2001;97:3466-3469 Crossref
  • Kato et al., 1996 K. Kato, O. Shimozato, K. Hoshi, H. Wakimoto, H. Hamada, H. Yagita, K. Okumura. Local production of the p40 subunit of interleukin 12 suppresses T-helper 1-mediated immune responses and prevents allogeneic myoblast rejection. Proc. Natl. Acad. Sci. U. S. A.. 1996;93:9085-9089 Crossref
  • Kebir et al., 2007 H. Kebir, K. Kreymborg, I. Ifergan, A. Dodelet-Devillers, R. Cayrol, M. Bernard, F. Giuliani, N. Arbour, B. Becher, A. Prat. Human TH17 lymphocytes promote blood–brain barrier disruption and central nervous system inflammation. Nat. Med.. 2007;13:1173-1175 Crossref
  • Kornek et al., 2000 B. Kornek, M.K. Storch, R. Weissert, E. Wallstroem, A. Stefferl, T. Olsson, C. Linington, M. Schmidbauer, H. Lassmann. Multiple sclerosis and chronic autoimmune encephalomyelitis: a comparative quantitative study of axonal injury in active, inactive, and remyelinated lesions. Am. J. Pathol.. 2000;157:267-276 Crossref
  • Lovas et al., 2000 G. Lovas, N. Szilagyi, K. Majtenyi, M. Palkovits, S. Komoly. Axonal changes in chronic demyelinated cervical spinal cord plaques. Brain. 2000;123(Pt 2):308-317 Crossref
  • Matsushita et al., 2013 T. Matsushita, T. Tateishi, N. Isobe, T. Yonekawa, R. Yamasaki, D. Matsuse, H. Murai, J. Kira. Characteristic cerebrospinal fluid cytokine/chemokine profiles in neuromyelitis optica, relapsing remitting or primary progressive multiple sclerosis. PLoS One. 2013;8:e61835 Crossref
  • Matusevicius et al., 1999 D. Matusevicius, P. Kivisakk, B. He, N. Kostulas, V. Ozenci, S. Fredrikson, H. Link. Interleukin-17 mRNA expression in blood and CSF mononuclear cells is augmented in multiple sclerosis. Mult. Scler.. 1999;5:101-104
  • McDonald et al., 2001 W.I. McDonald, A. Compston, G. Edan, D. Goodkin, H.P. Hartung, F.D. Lublin, H.F. McFarland, D.W. Paty, C.H. Polman, S.C. Reingold, M. Sandberg-Wollheim, W. Sibley, A. Thompson, S. van den Noort, B.Y. Weinshenker, J.S. Wolinsky. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann. Neurol.. 2001;50:121-127 Crossref
  • Miller and Leary, 2007 D.H. Miller, S.M. Leary. Primary-progressive multiple sclerosis. Lancet Neurol.. 2007;6:903-912 Crossref
  • Navikas et al., 1996 V. Navikas, D. Matusevicius, M. Soderstrom, S. Fredrikson, P. Kivisakk, A. Ljungdahl, B. Hojeberg, H. Link. Increased interleukin-6 mRNA expression in blood and cerebrospinal fluid mononuclear cells in multiple sclerosis. J. Neuroimmunol.. 1996;64:63-69 Crossref
  • Nylander and Hafler, 2012 A. Nylander, D.A. Hafler. Multiple sclerosis. J. Clin. Invest.. 2012;122:1180-1188 Crossref
  • Rothe et al., 1997 H. Rothe, R.M. O'Hara Jr., S. Martin, H. Kolb. Suppression of cyclophosphamide induced diabetes development and pancreatic Th1 reactivity in NOD mice treated with the interleukin (IL)-12 antagonist IL-12(p40)2. Diabetologia. 1997;40:641-646 Crossref
  • Rovaris et al., 2006 M. Rovaris, C. Confavreux, R. Furlan, L. Kappos, G. Comi, M. Filippi. Secondary progressive multiple sclerosis: current knowledge and future challenges. Lancet Neurol.. 2006;5:343-354 Crossref
  • Sadiq, 2005 S. Sadiq. L.P. Rowland (Ed.) Multiple sclerosis 11th edition in: Merritt's Textbook of Neurology. 134 (, 2005) 941-962
  • Saeki et al., 1992 Y. Saeki, T. Mima, S. Sakoda, H. Fujimura, N. Arita, T. Nomura, T. Kishimoto. Transfer of multiple sclerosis into severe combined immunodeficiency mice by mononuclear cells from cerebrospinal fluid of the patients. Proc. Natl. Acad. Sci. U. S. A.. 1992;89:6157-6161 Crossref
  • Sakoda et al., 1994 S. Sakoda, Y. Nakatsuji, H. Fujimura. Transfer of multiple sclerosis into SCID mice. Nihon Rinsho. 1994;52:2946-2951
  • Simpson et al., 1998 J.E. Simpson, J. Newcombe, M.L. Cuzner, M.N. Woodroofe. Expression of monocyte chemoattractant protein-1 and other beta-chemokines by resident glia and inflammatory cells in multiple sclerosis lesions. J. Neuroimmunol.. 1998;84:238-249 Crossref
  • Smith and Sommer, 1992 M.E. Smith, M.A. Sommer. Association between cell-mediated demyelination and astrocyte stimulation. Prog. Brain Res.. 1992;94:411-422 Crossref
  • Stangel et al., 2013 M. Stangel, S. Fredrikson, E. Meinl, A. Petzold, O. Stuve, H. Tumani. The utility of cerebrospinal fluid analysis in patients with multiple sclerosis. Nat. Rev. Neurol.. 2013;9:267-276 Crossref
  • Stelmasiak et al., 2000 Z. Stelmasiak, M. Koziol-Montewka, B. Dobosz, K. Rejdak, H. Bartosik-Psujek, K. Mitosek-Szewczyk, E. Belniak-Legiec. Interleukin-6 concentration in serum and cerebrospinal fluid in multiple sclerosis patients. Med. Sci. Monit.. 2000;6:1104-1108
  • Trapp et al., 1998 B.D. Trapp, J. Peterson, R.M. Ransohoff, R. Rudick, S. Mork, L. Bo. Axonal transection in the lesions of multiple sclerosis. N. Engl. J. Med.. 1998;338:278-285 Crossref
  • Williams et al., 2007 A. Williams, G. Piaton, C. Lubetzki. Astrocytes—friends or foes in multiple sclerosis?. Glia. 2007;55:1300-1312 Crossref


Tisch MS Research Center of New York, New York, NY 10019, USA

lowast Corresponding author at: Tisch MS Research Center of New York, 521 West 57th St. 4th Fl. New York, NY 10019, USA. Fax: + 1 212 265 8194.

Conflict of interest: The authors declare no competing financial interests.