OUP user menu

Combined therapy with methylprednisolone and erythropoietin in a model of multiple sclerosis

Ricarda Diem , Muriel B. Sättler , Doron Merkler , Iris Demmer , Katharina Maier , Christine Stadelmann , Hannelore Ehrenreich , Mathias Bähr
DOI: http://dx.doi.org/10.1093/brain/awh365 375-385 First published online: 15 December 2004

Summary

Neurodegenerative processes determine the clinical disease course of multiple sclerosis, an inflammatory autoimmune CNS disease that frequently manifests with acute optic neuritis. None of the established multiple sclerosis therapies has been shown to clearly reduce neurodegeneration. In a rat model of experimental autoimmune encephalomyelitis, we recently demonstrated increased neuronal apoptosis under methylprednisolone therapy, although CNS inflammation was effectively controlled. In the present study, we combined steroid treatment with application of erythropoietin to target inflammatory as well as neurodegenerative aspects. After immunization with myelin oligodendrocyte glycoprotein (MOG), animals were randomly assigned to six treatment groups receiving different combinations of erythropoietin and methylprednisolone, or respective monotherapies. After MOG-induced experimental autoimmune encephalomyelitis became clinically manifest, optic neuritis was monitored by recording visual evoked potentials. The function of retinal ganglion cells, the neurons that form the axons of the optic nerve, was measured by electroretinograms. Functional and histo pathological data of retinal ganglion cells and optic nerves revealed that neuron and axon protection was most effective when erythropoietin treatment that was started at immunization was combined with high-dose methylprednisolone therapy given from days 1 to 3 of MOG-induced experimental autoimmune encephalomyelitis. In contrast, isolated neuronal or axonal protection without clinical benefit was achieved under monotherapy with erythropoietin or methylprednisolone, respectively.

  • methylprednisolone
  • erythropoietin
  • experimental autoimmune encephalomyelitis
  • neuroprotection
  • visual evoked potentials
  • AMPA = α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
  • BDNF = brain-derived neurotrophic factor
  • CFA = complete Freund's adjuvant
  • EAE = experimental autoimmune encephalomyelitis
  • Epo = erythropoietin
  • ERG = electroretinogram
  • FG = fluorogold
  • IGF = insulin-like growth factor
  • MAPK = mitogen-activated protein kinase
  • MBP = myelin basic protein
  • MOG = myelin oligodendrocyte glycoprotein
  • MPred = methylprednisolone
  • RGC = retinal ganglion cell
  • VEP = visual evoked potential

Introduction

Multiple sclerosis is an autoimmune CNS disease that has long been thought to be primarily characterized by inflammation and demyelination. In the last few years, histopathological and MRI studies (Losseff et al., 1996; Trapp et al., 1998; Peterson et al., 2001) as well as data from animal models (Smith et al., 2000; Meyer et al., 2001) have reintroduced the presence of axonal and neuronal degeneration in multiple sclerosis. This neurodegenerative aspect of multiple sclerosis has the strongest impact on the development of permanent neurological deficits in patients (Trapp et al., 1999).

Experimental autoimmune encephalomyelitis (EAE) is the principal model of multiple sclerosis (Wekerle et al., 1994). By the use of different agents and modes of immunization and of different animal strains, EAE models mimicking the whole histopathological spectrum of the human disease could be established (Storch et al., 1998; Kornek et al., 2000). Previously, we demonstrated that EAE induced by immunization of female Brown Norway rats with recombinant myelin oligodendrocyte glycoprotein (MOG) strongly reflects the neurodegenerative aspects of multiple sclerosis. In this model, severe optic neuritis that occurs in 80–90% of the animals leads to acute axonal degeneration of the optic nerve and consecutive apoptosis of retinal ganglion cells (RGCs), the neurons that form its axons (Meyer et al., 2001; Hobom et al., 2004).

Clinically established strategies for the treatment of multiple sclerosis mainly target the autoimmune response by using anti-inflammatory, immunomodulatory and immunosuppressive agents. Although these substances have been proved to be beneficial in terms of modifying the clinical disease course (Noseworthy et al., 1999), for none of them have clear neuroprotective properties been demonstrated. In many studies, the concept of achieving neuroprotection as a secondary phenomenon resulting from the treatment of inflammation and autoimmunity was not confirmed (Hickman et al., 2003; Parry et al., 2003). Whereas common pharmacological multiple sclerosis therapies are well characterized with respect to the autoimmune reaction, little is known about direct effects on neuronal survival. For methylprednisolone (MPred), the standard therapy of acute multiple sclerosis relapses (Brusaferri and Candelise, 2000), we showed negative effects on RGC survival during MOG-induced optic neuritis, although inflammatory infiltration of the optic nerve was reduced (Diem et al., 2003). This pro-apoptotic action of the steroid was caused by inhibition of an endogenous neuroprotective pathway involving the phosphorylation of mitogen-activated protein kinases (MAPKs).

Erythropoietin (Epo), the main regulator of erythropoiesis in mammals (Jelkmann, 1992), has shown neuroprotective properties in models of brain injury, such as experimental ischaemia, trauma and epilepsy (Bernaudin et al., 1999; Brines et al., 2000). Beneficial effects on the clinical disease course were demonstrated in EAE induced by myelin basic protein (MBP) in rats (Brines et al., 2000). Compared with the effects of MPred during autoimmune optic neuritis (Diem et al., 2003), we recently observed that Epo acts in an antagonistic way by up-regulating active MAPKs in RGCs without influencing inflammatory infiltrates, demyelination or axonal damage of the optic nerve (Sättler et al., 2004).

In the present study we tested the hypothesis raised from our previous data that a combination of MPred and Epo might act synergistically during MOG-induced optic neuritis and thereby protect RGC bodies as well as their axons. By comparing different treatment protocols, we show that early application of a neuroprotective agent such as Epo together with steroid therapy during the acute stage of the disease is an effective strategy to prevent axonal and neuronal degeneration in inflammatory autoimmune CNS diseases.

Material and methods

Rats

Female Brown Norway rats 8–10 weeks of age were used in all experiments. They were obtained from Charles River (Sulzfeld, Germany) and kept under environmentally controlled conditions without the presence of pathogens.

All experiments that involved animal use were performed in compliance with the relevant laws and institutional guidelines. These experiments have been approved by the local authorities of Braunschweig, Germany.

Immunogen

Recombinant rat MOG, corresponding to the N-terminal sequence of rat MOG (amino acids 1–125), was expressed in Escherichia coli and purified to homogeneity by chelate chromatography (Weissert et al., 1998). The purified protein in 6 M urea was then dialysed against 0.01 M sodium acetate, pH 3, to obtain a preparation that was stored at −20°C.

Induction and evaluation of EAE

The rats were anaesthetized by inhalation of diethylether and injected intradermally at the base of the tail with a total volume of 100 µl inoculum, containing 50 µg MOG in saline emulsified (1 : 1) with complete Freund's adjuvant (CFA) (Sigma, St Louis, MO, USA) containing 200 µg heat-inactivated Mycobacterium tuberculosis (strain H 37 RA; Difco Laboratories, Detroit, MI, USA). Rats were scored for clinical signs of EAE and weighed daily. The signs were scored as follows: grade 0.5, distal paresis of the tail; grade 1, complete tail paralysis; grade 1.5, paresis of the tail and mild hind leg paresis; grade 2.0, unilateral severe hind leg paresis; grade 2.5, bilateral severe hind limb paresis; grade 3.0, complete bilateral hind limb paralysis; grade 3.5, complete bilateral hind limb paralysis and paresis of one front limb; grade 4, complete paralysis (tetraplegia), moribund state, or death. Rats were followed until day 8 of the disease (end of study). Statistical significance was assessed using Bonferroni corrected one-way analysis of variance (ANOVA) followed by Duncan's test.

Retrograde labelling of RGCs

Two weeks before immunization, adult Brown Norway rats were anaesthetized with intraperitoneal chloral hydrate (0.42 mg/kg body weight), the skin was incised mediosagittally, and holes were drilled into the skull above each superior colliculus (6.8 mm dorsal and 2 mm lateral from bregma). We injected stereotactically 2 µl of the fluorescent dye fluorogold (FG; 5% in normal saline) (Fluorochrome, Englewood, CO, USA) into both superior colliculi.

Electrophysiological recordings

The rats were anaesthetized by intraperitoneal injection of 10% ketamine (0.65 ml/kg; Atarost, Twistringen, Germany) together with 2% xylazine (0.35 ml/kg; Albrecht, Aulendorf, Germany) and mounted on a stereotaxic device. During the experiment, body temperature was maintained between 35° and 37°C with a heating pad, and the electrocardiogram was monitored continuously on an oscilloscope. For recording of visual evoked potentials (VEPs) from the primary visual cortex, two gold screw electrodes with a tip diameter of 1 mm were placed 3–4 mm lateral to the interhemispheric fissure and 1 mm frontal to the lambda fissure. Reference electrodes were placed 1 mm lateral to the midline and 1 mm before bregma. The electroretinogram (ERG) was recorded with chlorinated silver ball electrodes as described (Meyer et al., 2001). Visual stimuli were presented on a 17-inch monitor (Acer View 76i) positioned 20 cm in front of the eye. The display was centred on a position ∼40° medially from the pupil axis. Light flashes of 20 μs duration were used at a rate of 1 Hz, and bar stimulation consisted of vertical gratings of variable spatial frequency, alternating in phase with a temporal frequency of 1.8 Hz at 66% Michelson contrast (constant mean luminance, 15 cd/m2). Signals were amplified 10 000-fold and bandpass-filtered between 0.1 and 100 Hz, and 128 events were averaged to improve the signal-to-noise ratio. Amplitudes of pattern ERG and pattern VEP were determined as described earlier (Meyer et al., 2001). Assessment of visual acuity was also described earlier (Meyer et al., 2001). VEP and ERG measurements were performed at clinical onset of the disease. To monitor the disease course and to investigate therapeutic effects of the different treatment protocols, a second measurement of VEPs and ERGs was done on day 8 of the disease.

Treatment of animals

Animals were randomly assigned to the different treatment groups (n = 6 animals or n = 12 eyes for each group). Four eyes in each treatment group were used for western blot analysis and did not undergo electrophysiological assessment. Groups one (‘early Epo’) and two (‘late Epo’) received daily intraperitoneal injections of 5000 units recombinant human Epo (Janssen Cilag, Neuss, Germany)/kg bodyweight in 1 ml of 0.9% NaCl from the day of immunization onwards or started at disease onset, respectively. In addition to systemic application of Epo (given from immunization or disease manifestation onwards), group three (‘early Epo + MPred’) and four (‘late Epo + MPred’) were treated with intraperitoneal injections of MPred (20 mg/kg; Urbason®; Hoechst Marion Roussel, Frankfurt/Main, Germany) on days 1–3 of the disease. Group five (‘MPred’) received MPred alone from day 1 to day 3 of MOG-EAE. Group six (‘vehicle’) was treated with daily intraperitoneal injections of 1 ml 0.9% NaCl from the day of immunization onwards. Pre-experiments with application of vehicle over different treatment periods (from immunization or disease manifestation on) showed no differences concerning clinical outcome, electrophysiological or histopathological data.

Quantification of RGC density

At the end of the second recording session, the rats received an overdose of chloral hydrate and were perfused via the aorta with 4% paraformaldehyde in phosphate-buffered saline. The brain, the optic nerves and both eyes were removed, and the retinas were dissected and flat-mounted on glass slides. They were examined by fluorescence microscopy (Axiophot 2; Zeiss, Göttingen, Germany) using an UV filter (365/397 nm), and RGC densities were determined by counting labelled cells in three areas (62 500 µm2) per retinal quadrant at eccentricities of 1/6, 3/6 and 5/6 of the retinal radius. Cell counts were performed by two independent investigators following a blind protocol. Statistical significance was assessed using Bonferroni-corrected one-way ANOVA followed by Duncan's test.

Histopathology

Animals were perfused 8 days after disease onset and postfixed overnight in 4% paraformaldehyde. Optic nerves were removed and embedded in paraffin. Histological evaluation was performed on 4 µm thick slices stained with haematoxylin and eosin, Luxol-fast blue, and Bielschowky's silver impregnation to assess inflammation, demyelination and axonal pathology. Axonal densities were determined in vertical sections of the optic nerves stained by Bielschowsky's silver impregnation. Overview photographs (200× magnification) and high-magnification photographs (1000× magnification) were made with a CCD camera (Color View II; Soft imaging System®). The number of axons in each optic nerve was counted in at least 14 standardized microscopic fields of 2500 µm2 (Brück et al., 1997). Mean axon density was calculated for each optic nerve. The surface area of the optic nerve was measured using the analySIS® Docu software (Soft imaging System). Demyelinated areas were determined as a percentage of the whole optic nerve cross-section. The investigators who performed neuropathological examinations were blinded to the electrophysiological results of the study. Statistical significance was assessed using Bonferroni-corrected one-way ANOVA followed by Duncan's test.

Western blotting

In each treatment group, retinal protein lysates were prepared at day 3 of manifest EAE, 6 h after the last dose of vehicle, MPred and/or Epo was given. The western blot analysis was performed as described elsewhere (Diem et al., 2001). After incubation with the primary antibody against phospho-p44/42 MAPK [Thr180/Tyr182; New England Biolabs, Schwalbach, Germany; 1:200 in 1% skim milk in 0.1% Tween 20 in phosphate-buffered saline (PBS-T)], membranes were washed in PBS-T and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies against rabbit IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:3000 in PBS-T). Labelled proteins were detected using the ECL-plus reagent (Amersham, Arlington Heights, IL, USA).

p44/42 MAPK protein levels were detected using a primary antibody (sc-93-G; Santa Cruz Biotechnology) diluted 1 : 500 in 1% skim milk in PBS-T, and an HRP-conjugated secondary antibody against goat IgG (Santa Cruz Biotechnology; 1 : 3000 in PBS-T).

Results

Clinical disease course

Rats were randomly assigned into six treatment groups containing six animals each. The day of disease manifestation did not significantly differ between the groups. Vehicle-treated animals developed symptoms at day 15.9 ± 2.0 after immunization. In the early Epo treatment group (daily application of Epo from the day of immunization until day 8 of the disease; 5000 units/kg), the first neurological symptoms occurred at day 15.6 ± 0.8. Disease onset was at day 17.0 ± 2.6 in the late Epo group (treatment with Epo given from disease manifestation on until day 8 of EAE; 5000 U/kg). The early Epo + MPred group (20 mg/kg MPred given from day 1 to day 3 of EAE in addition to the early Epo treatment protocol) developed first signs of EAE at day 17.1 ± 1.9. In the late Epo + MPred group (20 mg/kg MPred given from day 1 to day 3 of EAE in addition to the late Epo treatment protocol), disease onset occurred 17.3 ± 0.8 days after immunization. Animals treated with MPred as a monotherapy from day 1 to day 3 of MOG-EAE (20 mg/kg) showed the first neurological symptoms at day 15.6 ± 1.5 after immunization.

In addition, the severity of symptoms at disease manifestation was similar in the different groups (vehicle group, mean clinical score 1.1 ± 0.4; early Epo, 1.0 ± 0.2; late Epo, 0.8 ± 0.1; early Epo + MPred, 0.8 ± 0.3; late Epo + MPred, 1.3 ± 0.3; MPred, 0.8 ± 0.3). Whereas the clinical score remained stable until day 8 of MOG-EAE in vehicle-treated animals (1.0 ± 0.2), in the group treated with MPred as a monotherapy (0.9 ± 0.3), as well as in late Epo-treated animals (0.6 ± 0.2), significant improvement was observed in both groups that received combination therapies. Mean clinical score in the early Epo + MPred group dropped to 0.1 ± 0.1 (P < 0.008). Rats treated according to the late Epo + MPred protocol performed normally at day 8 of MOG-EAE (mean clinical score, 0). Animals that received Epo from the day of immunization onwards showed a tendency to a better clinical outcome, which was, however, not statistically significant (mean clinical score of 0.3 ± 0.3 at day 8 of MOG-EAE).

Functional assessment of optic nerves and retinal ganglion cells

To diagnose optic neuritis in vivo and to monitor the function of RGCs, the neurons that form the axons of the optic nerve, we performed VEP and ERG measurements in response to flash and pattern stimulation. Flash VEP experiments were performed to test the axonal signalling of the optic nerve whereas pattern VEP recordings were used to estimate the animal's visual acuity. ERG measurements in response to flash stimulation indicate the intact function of the whole retina. In contrast, pattern ERG is a specific electrophysiological marker for RGCs. Recently, we have shown that healthy sham-immunized rats have visual acuity values of 1.31 ± 0.16 cycles/° determined by pattern VEP recordings and 1.10 ± 0.13 cycles/° in the pattern ERG measurements (Meyer et al., 2001). The electrophysiological results of the different treatment groups in our present study are summarized in Table 1. VEP and ERG recordings in each animal were performed at day 1 and day 8 of clinically manifest EAE. Compared with all other treatment protocols, only application of Epo from the day of immunization onwards combined with high-dose MPred therapy (early Epo + MPred) led to partial recovery of VEP responses to pattern stimulation and thereby to regaining of visual acuity on day 8 of MOG-EAE (Table 1; Fig. 1A–F). Also the electrophysiological responses to flash stimulation were improved in this group: at day 8 of the disease, eight out of eight tested eyes responded to flash light. Compared with pattern VEP responses, electrophysiological responses to flash stimulation showed a certain variability during clinical disease course that seemed to be partly independent of therapeutic intervention. Under vehicle treatment, three out of eight tested eyes recovered their response to flash light, whereas spontaneous recovery in response to pattern stimulation could not be observed (Table 1).

Fig. 1

Combined therapy of Epo started on the day of immunization and MPred given from day 1 to day 3 of MOG-EAE leads to recovery of VEP responses in rats with severe optic neuritis. Representative examples of VEP recordings from different treatment groups at day 1 (A, C, E) and day 8 (B, D, F) of MOG-EAE are given. In each group, recordings from days 1 and 8 were taken from the same individual animal and the same eye, respectively. Flash light stimuli are indicated by arrows. (A, B) No VEP responses to flash light stimulation were obtained from an animal of the vehicle-treated group on days 1 (d1 EAE veh) and 8 (d8 EAE veh) of MOG-EAE. (C, D) After treatment with MPred, no recovery of VEPs was seen at day 8 of MOG-EAE (d8 EAE after mpred) when compared with recordings from the same animal on day 1 of the disease (d1 EAE before mpred). (E, F) Under treatment with Epo from the day of immunization onwards, no potentials in response to flash light stimulation were produced on the day of clinically manifest EAE (d1 EAE early epo before mpred). After combination with MPred, clear potentials following flash light stimulation and in response to large-pattern bar stimulation (three alternating bars) were recorded (d8 EAE early epo + mpred). The stimulating pattern is indicated on top of the respective VEP recording sequence.

View this table:
Table 1

Results of VEP and ERG recordings (obtained on days 1 and 8 of MOG-EAE) of the different treatment groups

VEP day 1 (flash)VEP day 8 (flash)VEP day 1 (pattern)VEP day 8 (pattern)ERG day 1 (flash)ERG day 8 (flash)ERG day 1 (pattern)ERG day 8 (pattern)
Vehicle0/83/88/88/8
Early Epo3/81/88/88/80.35 ± 0.06 cyc/°0.30 ± 0.05 cyc/°
Late Epo4/84/87/88/8
Early Epo + MPred4/88/80.16 ± 0.03 cyc/°8/88/80.37 ± 0.04 cyc/°0.29 ± 0.06 cyc/°
Late Epo + MPred0/84/88/87/8
MPred0/82/83/86/8
Control (CFA-immunized)8/8 (day 18 p.i.)1.31 ± 0.16 cyc/° (day 18 p.i.)8/8 (day 18 p.i.)1.10 ± 0.13 cyc/° (day 18 p.i.)
  • Each group contained eight tested eyes. The number of eyes with detectable potentials to flash stimulation is given as the number out of eight tested eyes. The responses to pattern VEP and ERG are given as visual acuity values calculated from VEP or ERG amplitudes and the spatial frequency of the pattern stimulation. Control rats immunized with CFA were measured on day 18 after immunization (p.i.).

In the flash ERG measurements, most of the animals (42 out of 48 tested eyes at day 1 or 45 out of 48 tested eyes at day 8 of EAE) showed intact function of the whole retina (Table 1). Recordings of ERGs in response to pattern stimulation revealed significantly better RGC function at days 1 and 8 of MOG-EAE in both groups that received Epo (alone or in combination with MPred) from the day of immunization on (early Epo, early Epo + MPred) (Table 1; Fig. 2E, F) when compared with all other treatment groups. Both of these early Epo-treated groups showed a decrease in visual acuity determined by pattern ERG from day 1 to day 8 of the disease, independent of cotreatment with MPred (Table 1). In none of the other treatment groups were ERG potentials in response to pattern stimulation recorded at disease onset (Table 1; Fig. 2A–D). Spontaneous recovery of pattern ERG responses under vehicle treatment was also not observed (Table 1; Fig. 2A, B).

Fig. 2

Improved RGC function under treatment with Epo started on the day of immunization remains stable following MPred pulse therapy. Representative examples of pattern ERG recordings from different treatment groups on day 1 (A, C, E) and day 8 (B, D, F) of MOG-EAE are given. In each group, recordings from days 1 and 8 were taken from the same individual animal and the same eye, respectively. The stimulating pattern is indicated on bottom of the left panel. (A, B) No ERG responses to pattern bar stimulation (three alternating bars) were recorded in an animal from the vehicle-treated group on days 1 (d1 EAE veh) and 8 (d8 EAE veh) of MOG-EAE. (C, D) Treatment with MPred did not lead to recovery of ERG potentials in response to pattern bar stimulation (3 alternating bars) at day 8 of MOG-EAE (d8 EAE after mpred) when compared with recordings from day 1 of the disease (d1 EAE before mpred). (E, F) Under daily treatment with Epo given from the day of immunization onwards, clear ERG potentials in response to three alternating bars were obtained on day 1 of MOG-EAE (d1 EAE early epo before mpred). After additional application of MPred from day 1 to day 3 of the disease, stable ERG responses to pattern bar stimulation were recorded on day 8 of the disease in the same animal (d8 EAE early epo + mpred).

Histopathological evaluation of optic nerves

To correlate the results from VEP measurements with histopathological data of the optic nerves, we analysed axonal density, demyelination and inflammation. Representative optic nerve sections of the different treatment groups stained with Bielschowsky's silver impregnation, Luxol–fast blue and haematoxylin–eosin are given in Fig. 3. Mean optic nerve surface areas of the six treatment groups did not differ significantly (vehicle, 0.27 ± 0.07 mm2; early Epo, 0.25 ± 0.02 mm2; late Epo, 0.28 ± 0.01 mm2; early Epo + MPred, 0.28 ± 0.07 mm2; late Epo + MPred, 0.27 ± 0.04 mm2; MPred, 0.26 ± 0.06 mm2). In accordance with the functional results, the animal group which was treated using the early Epo + MPred protocol had significantly higher axon counts (32.732 ± 7585 axons/mm2; mean ± SEM; P < 0.008) (Fig. 4A) when compared with vehicle-treated animals (10.539 ± 3113 axons/mm2). Also, the extent of myelin damage determined as the percentage of the demyelinated area with respect to the whole optic nerve cross-section differed significantly when compared with vehicle treatment (early Epo + MPred, 6.1 ± 2.92% versus vehicle 66.1 ± 15.28%; mean ± SEM; P < 0.001) (Fig. 4B). Inflammatory infiltrates were markedly reduced in all three animal groups which received MPred (as monotherapy or cotreatment), whereas Epo alone had no influence on the extent of inflammatory infiltration.

Fig. 3

Histopathological analysis of optic nerves obtained on day 8 of MOG-EAE after different treatment regimes. Differences in surface areas of the optic nerves presented result from individual selection but do not reflect general differences between the treatment groups. (AC) Vehicle treatment. (DF) Early Epo (daily application of Epo started on the day of immunization). (GI) Early Epo + MPred (combined application of Epo started on the day of immunization and MPred given from day 1 to day 3 of MOG-EAE). (JL) Treatment with MPred from day 1 to day 3 of MOG-EAE. (A, D, G, J) Haematoxylin and eosin staining shows extensive cellular infiltration (purple spots) in the vehicle- (A) and in the early Epo (D)-treated optic nerve, whereas cellular infiltrates were reduced after treatment with MPred alone (J) or in combination with Epo (G). (B, E, H, K) Representative optic nerves stained with Luxol–fast blue reveal pronounced myelin preservation (blue) in the early Epo + MPred-treated optic nerve (H). After vehicle treatment (B), the whole optic nerve cross-section appears demyelinated (purple). The optic nerve treated with Epo as monotherapy from the day of immunization on (E) looks almost completely demyelinated. The blue area at the left side (low-magnification image) indicates a small part containing preserved myelin. Arrowheads indicate macrophages filled with myelin degradation products showing ongoing demyelination (high-magnification image). After treatment with MPred (K), large parts of the optic nerve show intact myelin (blue). (C, F, I, L) Bielschowsky silver impregnation shows marked axonal preservation in animals treated with MPred alone (L) or in combination with Epo (I). Remaining axons are indicated by green arrows; blue arrowheads indicate infiltrating macrophages. Scale bars: 100 or 20 µm for haematoxylin and eosin or Luxol–fast blue staining; 20 µm for Bielschowsky silver impregnation.

Fig. 4

Combined treatment with Epo started at immunization and MPred given from day 1 to day 3 of MOG-EAE reduces axonal damage and demyelination of the optic nerve. (A) Data are mean ± SEM of Bielschowsky-stained optic nerve axons/mm2 at day 8 of MOG-EAE. veh, vehicle treatment; mpred, treatment with MPred from day 1 to day 3 of MOG-EAE; epo early + mpred, combined application of Epo starting on the day of immunization and MPred given from day 1 to day 3 of MOG-EAE; epo late + mpred, combined treatment with Epo starting at disease onset and MPred given from day 1 to day 3 of MOG-EAE; epo early, daily application of Epo starting on the day of immunization; epo late, daily application of Epo starting at disease onset. *Statistically significant compared with vehicle treatment (P < 0.008; Bonferroni-corrected one-way ANOVA followed by Duncan's test). (B) The extent of demyelination is given as the percentage of the whole optic nerve cross-section stained with Luxol–fast blue. Data are mean ± SEM obtained at day 8 of MOG-EAE. **Statistically significant compared with vehicle treatment (P < 0.001; Bonferroni-corrected one-way ANOVA followed by Duncan's test).

An increased number of remaining axons was also detected in the group which received MPred as a monotherapy (25.856 ± 6516 axons/mm2; P < 0.008 when compared with vehicle) (Fig. 4A). In this group, a trend towards an increased amount of remaining myelin was observed (41.4 ± 13.02 versus 66.1 ± 15.2% in vehicle-treated animals) (Fig. 4B), which was not statistically significant. In the group that was treated according to the late Epo + MPred protocol, both axonal density (13.426 ± 5302 versus 10.539 ± 3113 axons/mm2 in vehicle-treated animals) (Fig. 4A) and demyelination (33.9 ± 14.33 versus 66.1 ± 15.2% for vehicle) (Fig. 4B) showed a trend towards a better outcome, which was not statistically significant.

Histopathological results in both treatment groups that received Epo as monotherapy (early Epo, late Epo) did not differ from those of the vehicle-treated animals (Fig. 4A, B). The numbers of remaining axons/mm2 in early Epo- and late Epo-treated animals were in the range of 6348 ± 1033 and 7600 ± 909, respectively. The extent of demyelination was also similar to that of the vehicle-treated group. The mean percentage of the demyelinated area was 67.28 ± 8.47% in rats treated with Epo from the day of immunization onwards, and 48.5 ± 14.6% when Epo therapy was started at disease onset (vehicle, 66.1 ± 15.2%).

Retinal ganglion cell counts

To compare the functional data from RGCs obtained by pattern ERG recordings with their survival rates, we counted FG-positive RGCs at day 8 of MOG-EAE in the different treatment groups. In control retinas of healthy CFA-immunized rats, mean RGC density was 2730 ± 145 cells/mm2 (mean ± SEM; n = 9) (Diem et al., 2003). Previously, we demonstrated that in our MOG-EAE model RGCs degenerate during severe optic neuritis (Meyer et al., 2001) and that this apoptotic RGC death was augmented under high-dose MPred therapy (Diem et al., 2003). Furthermore, we observed that Epo as a monotherapy started on the day of immunization protects RGCs from secondary cell death following autoimmune inflammation of the optic nerve (Sättler et al., 2004). In the present study comparing different treatment periods and combination protocols, Epo therapy started on the day of immunization was superior to all other therapeutic regimens. At day 8 of MOG-EAE, RGC counts in the early Epo-treated animal group were in the range of 1499 ± 41 cells/mm2 (mean ± SEM; n = 8; P < 0.008 when compared with vehicle treatment) (Fig. 5C, E). RGC counts in vehicle-treated animals dropped to 775 ± 112 until day 8 of MOG-EAE (n = 8) (Fig. 5A, E). In contrast to the early Epo treatment protocol, Epo given as a monotherapy started at disease onset (late Epo) had no significant effect on RGC survival when compared with vehicle treatment (960 ± 60 versus 775 ± 112 RGCs/mm2; n = 8 each) (Fig. 5E). In agreement with our previous results, MPred given from day 1 to day 3 of MOG-EAE decreased the number of surviving RGCs to 441 ± 61/mm2 (n = 8; P < 0.008 when compared with vehicle treatment) (Fig. 5B, E). This effect was completely abolished under combined treatment with Epo if Epo therapy was started at disease onset (late Epo + MPred, 760 ± 85 RGCs/mm2; n = 8; P < 0.008 when compared with MPred alone) (Fig. 5E). If MPred therapy was combined with application of Epo following the early Epo treatment protocol, the survival-promoting effect of the cytokine predominated. Under this combination therapy, RGC survival at day 8 of MOG-EAE was still promoted to 1211 ± 43 cells/mm2 (n = 8; P < 0.008 when compared with MPred alone or when compared with vehicle-treated animals) (Fig. 5D, E).

Fig. 5

Daily application of Epo alone starting at immunization or in combination with MPred given from day 1 to day 3 of MOG-EAE promotes RGC survival during optic neuritis. (A) Representative whole-mount area at three-sixths of the retinal radius from a vehicle-treated animal at day 8 of MOG-EAE. RGCs are labelled with FG. (B) When compared with vehicle treatment, the number of FG-positive RGCs is reduced after MPred monotherapy. (C) Epo treatment started at immunization increases the number of surviving RGCs counted at day 8 of MOG-EAE. (D) Under combined treatment with Epo starting on the day of immunization and MPred given from day 1 to day 3 of MOG-EAE, RGC counts are still increased when compared with vehicle treatment. Scale bar, 100 µm. (E) Data are mean ± SEM of retrogradely labelled RGCs/mm2. veh, vehicle treatment; mpred, treatment with MPred from day 1 to day 3 of MOG-EAE; epo early + mpred, combined application of Epo started at the day of immunization and MPred given from day 1 to day 3 of MOG-EAE; epo late + mpred, combined treatment with Epo started at disease onset and MPred given from day 1 to day 3 of MOG-EAE; epo early, daily application of Epo started at the day of immunization; epo late, daily application of Epo started at disease onset. *Statistically significant when compared with vehicle treatment (P < 0.008); **statistically significant when compared with MPred monotherapy (P < 0.008; Bonferroni-corrected one-way ANOVA followed by Duncan's test).

Signal transduction in retinal ganglion cells

Recently, we observed that MPred or Epo can regulate intracellular concentrations of phosphorylated MAPKs in RGCs in an antagonistic way (Diem et al., 2003; Sättler et al., 2004). The functional relevance of this pathway for RGC survival in our MOG-EAE model has been shown by experiments using a pharmacological inhibitor of MAPK kinase, the upstream kinase, which in turn phosphorylates and thereby activates MAPKs (Diem et al., 2003). To investigate whether the extent of MAPK phosphorylation under the different treatment regimens in the present study correlates with RGC survival, we performed western blot analysis on phosphorylated and unphosphorylated forms of p44/42-MAPKs (data not shown). These experiments revealed that Epo-induced phosphorylation of MAPKs does not depend on the different treatment intervals used in this study. In both animal groups which received Epo as a monotherapy (early Epo, late Epo), protein concentrations of phospho-p44/42-MAPKs were increased when compared with vehicle treatment, whereas levels of the unphosphorylated forms of both proteins were unchanged. In accordance with our previous results, high-dosage MPred therapy led to downregulation of phospho-p44/42-MAPKs in RGCs. This effect was abolished under combined application of MPred and Epo if Epo therapy was administered according to the late Epo protocol. If Epo as a combination therapy was started on the day of immunization, protein levels of phosphorylated MAPKs were still higher than those under vehicle treatment, indicating that Epo given early has a stronger influence than MPred on the regulation of this neuroprotective pathway.

Discussion

Recently, we demonstrated that MPred has direct pro-apoptotic effects on RGCs during MOG-induced optic neuritis by inhibiting MAPK phosphorylation in these neurons, although inflammatory infiltrates of the optic nerve were reduced (Diem et al., 2003). In contrast, Epo was observed to promote phosphorylation of MAPKs in RGCs without exerting any positive effects on histopathological changes in the optic nerve (Sättler et al., 2004). Under the hypothesis that protective influences might be complementary to each other if both substances are combined, we tested different treatment protocols and combination therapies of Epo and MPred in our present study.

Application of Epo was performed systemically due to its ability to cross the blood–brain barrier under physiological and pathophysiological conditions (Brines et al., 2000). As revealed by functional, electrophysiological data as well as histopathological results from retinas and optic nerves, we show that combination of Epo started on the day of immunization and MPred given from day 1 to day 3 of MOG-EAE was most effective in protecting RGCs as well as their axons. Compared with this protocol, Epo given as monotherapy from the day of immunization onwards was superior with respect to pure neuronal cell body protection but failed to achieve any axonal rescue. Functionally, this was reflected by improved results from pattern ERG measurements without any improvement in VEP data. From the rat model of surgical axotomy of the optic nerve, it is known that apoptotic cell death of 80–90% of RGCs occurs following optic nerve transection (Isenmann et al., 1997; Diem et al., 2001). In this model, it has also been demonstrated that neuroprotective therapies delay apoptotic cell death of RGCs rather than resulting in permanent rescue or even promoting neuronal regeneration (Kermer et al., 1999). From these observations, it can be concluded that, also during autoimmune optic neuritis, classical neuroprotective therapies alone exert ‘cosmetic’ effects on RGC survival that must remain transient as long as axon degeneration continues. Accordingly, from the very few successfully tested neuroprotective substances in EAE models, stable effects on neuronal cell body survival were produced only by those that influenced axonal pathology as well. It has been shown that antagonists of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate type of glutamate receptor rescued motoneurons in the lumbar spine during MBP-induced EAE (Smith et al., 2000) and protected spinal cord axons from EAE-associated degeneration in an adoptive transfer model in rodents (Pitt et al., 2000). Neuron- as well as axon-protective effects in EAE might also be expected from substances with combine anti-inflammatory and neuroprotective properties. Minocycline, for instance, has been shown to reduce inflammation in EAE (Brundula et al., 2002) and to slow down disease progression in a model of amyotrophic lateral sclerosis (Kriz et al., 2002). At least parts of the neuroprotective effects of this tetracycline derivate seem to be mediated through inhibition of glutamate excitotoxicity (Darman et al., 2004).

We demonstrate in our present study that Epo does not have any significant effect on neuronal survival when monotherapy is started at disease onset of EAE. This is explained by observations from our previous work showing that apoptotic cell death of RGCs starts around 1 week before clinical manifestation of EAE. At disease onset, rats have lost 54% of their RGCs (Hobom et al., 2004), indicating that effective neuroprotective therapies should cover subclinical periods of autoimmune CNS inflammation. Transferring this to the human disease where the starting point of autoimmunity cannot be determined, it might be useful to extend neuroprotective treatment for a longer period than that of acute neurological deterioration. In support of this notion, recent MRI studies in patients with relapsing–remitting or primary progressive multiple sclerosis revealed the existence of neurodegenerative processes independent of inflammation and clinically manifest disease activity (De Stefano et al., 2003a, b). However, late, short-duration application of Epo in our present study was beneficial when treatment was combined with high-dose MPred therapy. Under these conditions, Epo completely abolished negative steroid effects on neuronal survival by antagonistic regulation of phospho-MAPKs, as shown by western blot analysis. These data reveal that Epo treatment started simultaneously with steroid pulse therapy is effective in protecting neurons from acute MPred-induced augmentation of cell death but fails to significantly influence the underlying, advanced EAE-associated processes of neurodegeneration.

Comparing our results from VEP measurements with histopathological data on the optic nerves, good functional outcome was obtained only in the animal group that was treated according to the early Epo + MPred protocol, although beneficial effects of MPred on axon counts were also seen if MPred was given as monotherapy. This can be explained by the severe reduction of RGCs after application of MPred alone. It is known that RGC loss can occur before alterations of VEP potentials are detectable. But after RGC death has reached a certain threshold, it is not only ERG potentials that are affected (Hobom et al., 2004), severe functional deficits of VEP responses can also be the consequence (Hood and Greenstein, 2003). From a functional point of view, these data again argue for the necessity of therapeutically targeting both neuronal cell bodies and axons to avoid effects that do not result in any clinically relevant benefit. Comparing histopathological changes in the optic nerves between our different steroid-treated rat groups, protection from axonal damage and demyelination in animals treated according to the early Epo + MPred protocol was better than under MPred therapy alone. In contrast, no additional effects on histopathological changes of the optic nerves were seen when Epo treatment was started simultaneously with MPred. In a study on MBP-EAE in Lewis rats, it was shown that Epo treatment started at day 3 after immunization decreased CNS tissue levels of tumour necrosis factor-α and interleukin-6 (Agnello et al., 2002). For both of these proinflammatory cytokines, relevant influences on the development of demyelination and axonal damage have been described in different EAE models (Okuda et al., 1999; Probert et al., 2000). Although Epo alone did not improve histopathological changes of the optic nerves in our study, it is conceivable that its suppressing effect on the production of proinflammatory cytokines early during development of EAE synergistically enhances MPred-induced tissue protection.

In a recent study, it has been demonstrated that simultaneous treatment with an anti-inflammatory antibody, an AMPA/kainate receptor antagonist and the N-terminal tripeptide of insulin-like growth factor (IGF) reverses EAE in mice (Kanwar et al., 2004). The concept of ‘benign autoimmunity’ might help to explain the effectiveness of combining anti-inflammatory treatment with application of neurotrophin-like substances, such as IGF and Epo, as shown in our EAE model. According to this hypothesis, autoimmune inflammatory conditions can promote neuronal survival via increased secretion of neurotrophic factors by cells of the immune system. High levels of brain-derived neurotrophic factor (BDNF) and other glial cell line-derived neurotrophins were detected in T cells in the spinal cord from animals with MBP-EAE, which in turn led to increased neuronal survival of spinal motoneurons after additional axotomy in this model (Hammarberg et al., 2000). Production of BDNF upon antigen stimulation was shown in a study on T-cell lines specific for myelin autoantigens, such as MBP and MOG (Kerschensteiner et al., 1999). In this study, immune-cell-derived BDNF was demonstrated to support the survival of sensory neurons in vitro. BDNF immunoreactivity was also detected in inflammatory cells in lesional areas within the brain of multiple sclerosis patients (Kerschensteiner et al., 1999). Neurotrophic factors in turn activate a variety of protective intracellular neuronal pathways, such as MAPK phosphorylation, which has been described in many studies (Yamada et al., 2001; Barnabe-Heider et al., 2003). With this background, delivery of Epo as an exogenous neurotrophin-like substance might compensate for the lack of endogenous neurotrophic factor support resulting from anti-inflammatory treatment of EAE or multiple sclerosis.

In summary, we present evidence for beneficial effects of combined anti-inflammatory and neuroprotective treatment in a rat model of multiple sclerosis that especially reflects neurodegenerative aspects of the disease. Significant differences in functional outcome resulting from alternative protocols point up the importance of exactly selecting treatment intervals for the therapeutic agents and thoroughly considering their interactions concerning intracellular signal transduction. Compared with the Epo-induced effects in this study, a similar activation of intracellular signalling steps or neuronal cell body protection might also be achieved by classical neurotrophins. However, limitations concerning their application mode make Epo and its recently developed carbamylated derivate (Leist et al., 2004) more promising candidates for testing in the human disease as well.

Acknowledgments

This work was supported by the Medical Faculty of the University of Göttingen, Germany (junior research group; R. D., D. M., C. S.), Biogen Idec and the Gemeinnützige Hertie Stiftung. Erythropoietin was kindly provided by Janssen Cilag, Neuss, Germany. We thank Inna Boger for expert technical assistance. We also thank F. Staub for critically reading this manuscript.

References

View Abstract