Tumor necrosis factor α and its receptors in relapsing-remitting multiple sclerosis

Journal of Neurological Sciences 152 (1997) 51–61

Tumor necrosis factor a and its receptors in relapsing-remittingmultiple sclerosis

a , b a c d bGianvito Martino , Antonella Consiglio , Diego M. Franciotta , Angelo Corti , Massimo Filippi ,a a bKoen Vandenbroeck , Francesca L. Sciacca , Giancarlo Comi ,

a , b ,*Luigi M.E. GrimaldiaNeuroimmunology Unit, Dept. of Biotechnology (DIBIT), San Raffaele Scientific Institute, Via Olgettina 58, 20132 Milano, Italy

bDept. of Neurology, University of Milano, San Raffaele Scientific Institute, Via Olgettina 60, 20132 Milano, ItalycLaboratory of Neuroimmunology, Dept. of Neurology, ‘C. Mondino’ Neurological Institute, University of Pavia, Via Palestro 7, 27100 Pavia, Italy

dImmunobiotechnology Unit, Dept. of Biotechnology (DIBIT), San Raffaele Scientific Institute, Via Olgettina 58, 20132 Milano, Italy

Received 26 November 1996; received in revised form 17 April 1997; accepted 28 April 1997

Abstract

In the attempt to further characterize the extent and timing of tumor necrosis factor (TNF)a-system activation during multiple sclerosis(MS), we performed a cross-sectional and a longitudinal study in a total of 73 relapsing-remitting MS patients. We assessed serum levelsof soluble TNFa, soluble TNFa receptor 1 (R1) and soluble TNFa receptor 2 (R2) in 65 relapsing-remitting MS patients in differentphases of disease. TNFa, R1 and R2 serum levels measured in MS patients did not differ from those measured in healthy individuals anddid not correlate with (a) clinical relapses, (b) presence of gadolinium-enhancing brain-magnetic resonance imaging (MRI) lesions, and (c)bioactivity of TNFa. We also measured in 8 additional relapsing-remitting MS patients peripheral blood mononuclear cells (PBMC)mRNA levels of TNFa, R1 and R2 every 15 days for one year. In 4 of these patients we also measured levels of soluble TNFa, R1 andR2 every 15 days for 5 months across a clinical exacerbation. PBMC TNFa, R1 and R2 mRNA levels and serum levels of soluble R1 andR2, but not TNFa, fluctuated concordantly (P,0.05) and peaked a mean of 6 weeks before clinical and MRI evidence of disease activity.Moreover, we found a significant positive correlation between cumulative TNFa and R2 mRNA levels (measured during the follow-upperiod in the 8 MS patients studied serially) and the number of clinical attacks recorded in these patients during the study. Our data showthat serum levels of soluble TNFa, R1, and R2 in MS patients do not differ from those of healthy individuals. However, although withinnormal values, the transcription and production rate of all these molecules fluctuate concordantly in the peripheral blood during the courseof the disease (with the exception of soluble TNFa) and their maximal elevation significantly precedes the occurrence of clinicalexacerbations. It is not clear whether soluble TNFa escapes recognition by commonly used assays or is simply not released in its solubleform in MS patients. In any case, measurement of TNFa mRNA levels and R1 and R2 mRNA and protein levels appears to be a betterindicator of disease fluctuations during the course of MS than assessments of soluble TNFa protein. 1997 Elsevier Science B.V.

Keywords: Multiple sclerosis; TNFa; TNFa receptor 1; TNFa receptor 2

1. Introduction expressed transmembrane receptors, R1 (55 kD) and R2(75 kD) (Beutler, 1994; Tracey, 1994) members of the

The activities of the proinflammatory cytokine, tumor nerve growth factor receptor /TNF receptor (NGFR/necrosis factor (TNF)a, are mediated by two ubiquitously TNFR) superfamily. Both TNFa receptors are coded by

separate non polymorphic genes sharing less than 25%* sequence similarity. They also occur in a soluble formCorresponding author. Tel.: 139 2 26432791; fax 139 2 2434855;

e-mail: [email protected] resulting from proteolysis (Porteu and Nathan, 1990) of

0022-510X/97/$17.00 1997 Elsevier Science B.V. All rights reserved.PII S0022-510X( 97 )00142-1

52 G. Martino et al. / Journal of Neurological Sciences 152 (1997) 51 –61

the corresponding transmembrane receptor (Nophar et al., measurements of TNFa and/or R1 and R2 mRNA levels1990). The soluble TNFa receptors have been shown to are a better indicator of disease activity occurring duringmodulate TNFa activity in vitro by inhibiting receptor MS than their protein counterparts.binding and by stabilizing the bioactive trimeric structureof TNFa (Nophar et al., 1990; Aderka et al., 1992; Banneret al., 1993).

2. Materials and methodsIn a complex interplay with several other pro-inflamma-tory cytokines, TNFa exerts a direct effector function in

2.1. Cross-sectional study patientsexperimental and human demyelination (Selmaj and Raine,1988; Hofman et al., 1989; Ruddle et al., 1990; Selmaj et

Levels of soluble TNFa, R1 and R2 were measured inal., 1991a,b) and contributes to the cellular activation thatserum samples obtained from 65 patients (41 women, 24precedes the occurrence of central nervous system (CNS)men; mean age, 33.3 years; range, 19–55 years) affectedinflammatory and demyelinating lesions in patients withby clinically definite relapsing-remitting MS (mean diseasemultiple sclerosis (MS) (Beck et al., 1988; Rieckmann etduration, 6.9 years; range, 0.5–25 years) according to theal., 1995). TNFa mRNA levels have been reported toclinical classification of Poser et al. (1983). At the time ofincrease in peripheral blood mononuclear cells (PBMC)sample collection, all MS patients underwent clinicalfrom patients with MS from 2 to 6 weeks prior to theexamination and were evaluated by the Expanded Disabili-appearance of clinical manifestations of the disease (Riec-ty Status Scale (EDSS) (Kurtzke, 1983) (mean EDSS atkmann et al., 1995). However, the measurement of solubleentry, 2.5). We subdivided relapsing-remitting MS patientsTNFa levels has yielded controversial results (Beck et al.,into those experiencing a clinically active (34 patients) or1988; Franciotta et al., 1989; Sharief and Hentges, 1991;stable (31 patients) phase of the disease. Venous blood wasRudick and Ransohoff, 1992; Chofflon et al., 1992; Dore-obtained from all active patients prior to any treatment. WeDuffy et al., 1995) thus questioning the clinical relevancedid not include in the study patients who had receivedof these observations.immunosuppressive drugs during the preceding 3 months.Elevations of soluble TNFa R1 levels have been

reported during phases of disease activity in serum andcerebrospinal fluid samples from MS patients (Matsuda et 2.2. Longitudinal study patientsal., 1994). Attempts to confirm this additional finding havebeen, again, controversial (Rieckmann et al., 1994a; Har- PBMC mRNA levels of TNFa, R1, and R2 weretung et al., 1995; Spuler et al., 1996). measured every 15 days for one year in 8 patients (6

The levels in humans of R2 are approximately 4-5 times women, 2 men; mean age, 25 years; range, 21–37 years)that of R1, and R2 has an higher affinity for TNFa than affected by clinically definite relapsing-remitting MSR1. Thus, it is likely that R2 plays a more relevant role (mean disease duration, 6.3 years; range 2–10 years).than R1 in modulating TNFa bioactivity in vivo. However, Patients were clinically evaluated using the EDSS at theto our knowledge, R2 has never been evaluated in MS time of all sample collections (mean EDSS at entry, 2.5).patients. All patients but one experienced at least one clinical attack

The goal of this study was to determine the dynamics of during the study (1 patient, 4 attacks; 2 patients, 3 attacks;the TNFa system in MS patients by measuring in 73 3 patients, 2 attacks; 1 patient, 1 attack; 1 patient, norelapsing-remitting MS patients TNFa, R1 and R2 protein attacks) and were treated during the attack with a 2 to 3as well as mRNA levels. Disease activity was evaluated at week course of steroids. To evaluate the concordancethe same time using clinical assessment and brain magnetic between mRNA and soluble protein levels of TNFa, R1,resonance imaging (MRI). TNFa, R1, and R2 mRNAs and R2 we also obtained, at the time of PBMC collection,were detected in PBMC-derived total mRNA from MS serum samples from 4 of these MS patients for 3 to 5patients by reverse transcriptase-polymerase chain reaction months across a relapse.(RT-PCR). An immunoenzymatic assay was used tomeasure serum total TNFa protein (free plus solublereceptor-bound TNFa) levels and a cytolytic bioassay was 2.3. Brain-MRI studiesused to measure its bioactivity. The soluble forms of R1and R2 were measured by enzyme-linked immunoadsor- Brain-MRI was performed with a 1.5 Tesla machinebent assays (ELISA). We found that protein and mRNA (Siemens SP63, Erlangen, Germany) at the time of samplelevels of TNFa, R1, and R2 did not differ between MS collection in the 39 MS patients studied cross-sectionallypatients and controls. However, their values fluctuate (19 clinically active and 20 stable) and monthly in allconcordantly in the peripheral blood during the course of patients followed longitudinally. Five mm contiguous axialMS and peak (either at mRNA or protein levels) prior to T -weighted (SE 600/17; field of view, 230 mm, matrix1

the occurrence of clinical exacerbations. Interestingly, size, 1983256; 2 excitations) slices throughout the whole

G. Martino et al. / Journal of Neurological Sciences 152 (1997) 51 –61 53

brain were obtained 5 to 7 min after the injection of preliminary experiments on cDNA from the U937 cell line.gadolinium-diethylenetriamine pentaacetic acid (0.1 mmol / The number of cycles was increased from 10 to 40, in 5kg intravenously over 1–2 min). New or persisting enhanc- cycles increments; 25 cycles were selected as conditionsing lesions were counted in each scan by one of us (M.F.), yielding measurements belonging to the exponential part ofunaware of the patient clinical status or immunological the amplification curve. Amplified PCR products wereresults. visualized in ethidium bromide-stained 1% agarose gels

run in TAE buffer, blotted onto nylon membranes, and32hybridized with specific P-labeled (Ready Prime kit,

2.4. Measurements of soluble TNFa, R1, and R2: Amersham, Buckinghamshire, UK) probes representing theproteins amplified region. Probes were obtained by elution of RT-

PCR products from the U937 cell line representing specificSerum TNFa levels were measured using a commercial- sequences of TNFa, R1 and R2. Autoradiograms of the

ly available ELISA (Medgenix Diagnostics, Fleurus, Bel- probed TNFa, R1 and R2 PCR products were quantifiedgium) that detects both free and soluble receptor-bound by densitometry (Computing Densitometer, MolecularTNFa. The detection limit of the assay was 10 pg/ml. Dynamics; Image Quant Software, version 3.3). The rawSoluble R1 was measured by an ELISA (detection limit of data was normalized against the intensity of the b-actinthe assay was 0.3 ng/ml) developed in our laboratory band. For inter-subject comparisons, the normalized inten-(Corti et al., 1994a). Soluble R2 was measured (detection sities were further corrected using the normalized inten-limit of the assay was 0.15 ng/ml) using a commercial sities of the U937 cell-derived positive controls. All cDNAELISA (Bender, Vienna, Austria). samples from each patient were run in the same PCR

round. To minimize experimental errors, all cDNA sam-ples from a given patients were amplified simultaneously

2.5. Measurement of serum cytolytic activity of TNFa and a positive control (i.e. U937 cell cDNA) was includedin every PCR experiments.

Serum TNFa bioactivity was measured by a cytolytic The following primers were used: TNFa (product: 490bioassay using mouse WEHI 164 clone13 (Espevik and bp), antisense 59-GAC CTC TCT CTA ATC AGC CC-39,Nissen-Meyer, 1986). The calibration curve of TNFa sense 59-CAG ACC CGT CCA GAT GAA AC-39, R1cytotoxicity was prepared by testing, in parallel, eight (product: 659 bp), antisense 59-GGG GGG ATC CTCTNFa standard solutions (Woc Ltd.,Vaduz, Switzerland) at ATG TGG TGC CTG AGT CCT C-39; sense 59-ATAconcentrations ranging from 0.274 to 200 pg/ml (specific GAA GCT TGG CAT GGG CCT CTC CAC CGT G-39;

7activity 3.2310 U/mg). TNFa cytolytic activity was R2 (product: 430 bp), antisense 59-AGT GCG TTA TCTcalculated by interpolation from the calibration curve; TGC CCA GG-39, sense 59-TGG AAA GCA TCA CCTresults were expressed as equivalents of standard TNFa CAG GC-39, b-actin (product: 658 bp); antisense 59-CTAcytotoxicity. The detection limit of the assay was 10 GAA GCA TTT GCG GTG GAC GAT G-39, sensepg/ml. 59-ACG GGG TCA CCC ACA CTG TGC-39.

2.6. Measurements of TNFa, R1, and R2: mRNA 2.7. Statistical analyses

PBMC were obtained by Ficoll Hypaque density cen- Protein levels of soluble TNFa, R1 and R2 measured intrifugation of EDTA-treated blood. Total RNA was iso- the cross-sectional study were compared with clinicallated from PBMC using guanidine-isothiocyanate solution variables using the Mann Whitney U Test for unpairedand purified by ultracentrifugation on a cesium chloride data. TNFa, R1 and R2 protein and mRNA levels mea-gradient (Maniatis et al., 1989). A T-primed first strand kit sured in patients followed longitudinally were comparedwas used for the reverse transcription of total RNA into by regression analysis. mRNA levels of TNFa, R1 and R2cDNA (Ready-to-go kit, Pharmacia, Upsala, Sweden). measured during the follow-up period in the 8 MS patientsPCR amplification of cDNA sequences specific for TNFa, studied longitudinally were summed up and their meanR1 and R2 was performed using 1 mM cytokine primers, values were correlated with the number of clinical attacks200 mM of each dNTP, 50 mM KCl, 10 mM Tris-HCl (pH suffered during the follow-up period by regression analy-8.3), 1.5 mM MgCl . After a hot start at 95ºC for 5 min, 1 sis. The P values for regression analysis were derived from2

U Taq Polymerase was added and 25 PCR cycles were a t-test to test the hypothesis: H : regression line slope ±0

carried out. Each PCR cycle consisted of 1 minute of 0. Days elapsed (before and after) between a TNFa, R1 oramplification at 95ºC, 1 minute of primer annealing at R2 peak and a clinical attack or a peak in the number of60ºC, and 1 minute of product dissociation at 72ºC. The brain-MRI lesions were compared using the Mann Whitneynumber of PCR amplification cycles was selected based on U test for unpaired data.

54G

.M

artinoet

al./

Journalof

Neurological

Sciences152

(1997)51

–61

Table 1Clinical characteristics and serum levels of soluble TNFa, R1 and R2 in patients with MS

a bStatus of No of Sex Age Disease duration (y) EDSS TNFa (pg /ml) TNFa receptors (ng/ml)MS patients patients F/M (y)

Antigenic Cytolytic R1 R2

Clinically active 34 23/11 30.461.1 (19–45) 4.160.5(0.5–10) 2.660.2(1–7.5) 16.661.2 (10–30.2) 36.265.0 (15–103) 1.260.1 (0.3–2.3) 3.960.2 (2.2–8.6)Clinically stable 31 18/13 36.761.5 (25–55) 10.061.2(0.5–25) 2.460.3(0–7) 16.661.2 (10.3–26.2) 37.264.1 (19.4–66.9) 1.360.1 (0.4–2) 4.160.3 (2.4–9.4)

c dBrain-MRI positive 24 14/10 32.561.5 (19–47) 6.061.0(0.5–17) 2.460.3(1–7.5) 15.161.4 (10.3–29.3) 20,30 1.060.1 (0.3–2.3) 4.160.2 (2.3–7.5)c dBrain-MRI negative 15 10/5 33.961.8 (25–46) 10.361.9(0.5–21) 2.460.4(0–5) 18.361.3 (10–26.2) 20, 20, 20 1.060.1 (0.4–1.7) 4.760.5 (2.5–9.4)

a Values are means 6 standard errors. Values in parenthesis are ranges.b TNFa and R1 mean values were calculated only in serum samples showing measurable levels. Antigenic form of TNFa represents free as well as receptor-bound soluble TNFa.c Patients showing or not showing brain-MRI enhancing lesions.d Single measurable values.

G.

Martino

etal.

/Journal

ofN

eurologicalSciences

152(1997)

51–61

55

Fig. 1. Regression analyses comparing TNFa, R1 and R2 mRNA and protein levels obtained in MS patients followed longitudinally. Panels: (A) TNFa mRNA vs. R1 serum levels (P50.0007), (B)TNFa mRNA vs. R1 mRNA (P50.0001), (C) TNFa mRNA vs. R2 serum levels (P50.041), (D) TNFa mRNA vs. R2 mRNA (P50.0001), (E) R1 mRNA vs. R2 mRNA (P50.027), (F) R1 mRNAvs. R2 serum levels (P50.035), (G) R1 serum levels vs. R2 serum levels (P50.0001), (H) R2 mRNA vs. R1 serum levels (P50.0001), and (I) R2 mRNA vs. R2 serum levels (P50.0007).Comparisons which did not reach a statistical significance (TNFa serum levels vs. R1 serum levels: P50.11; TNFa serum levels vs. R2 serum levels: P50.33; TNFa serum levels vs. TNFa mRNA:P50.28; TNFa serum levels vs. R1 mRNA: P50.38; TNFa serum levels vs. R2 mRNA: P50.21; R1 serum levels vs. R1 mRNA: P 0.125) were omitted from the graph. Values in the x- and y-axisrepresent the measure units for each variable while curves within the graphs represent 95% confidence interval of the slopes.


Fig. 2. Serial measurements of soluble TNFa (filled circles), R1 (filled diamonds) and R2 (filled squares) protein levels in 4 MS patients studied every 15days across an acute clinical attack. Follow up periods ranged between 3 and 5 months. Each symbol on x-axis represents a single measurement while onthe y-axis are represented the serum levels of TNFa (pg /ml), R1 (ng/ml) and R2 (ng/ml). Arrows represent clinical attacks and black boxes the period ofsteroid treatment. The two TNFa receptors seem to fluctuate concordantly in all 4 patients over time while the TNFa does not.

3. Results without gadolinium-enhancing brain-MRI lesions. No cor-relation was found between levels of soluble TNFa or

3.1. Cross-sectional study serum cytolytic activity of TNFa and levels of R1 and R2.

Results are summarized in Table 1. Soluble TNFa

protein was measurable (.10 pg/ml) in serum samples 3.2. Longitudinal studyfrom 22 of 34 (65%) patients with active and 17 of 31(55%) patients with stable MS. Bioactive TNFa was In the 4 MS patients experiencing a clinical attack inmeasurable (.10 pg/ml) in 17 of 34 (50%) patients with which levels of soluble TNFa, R1 and R2 were measuredactive and 13 of 31 (42%) patients with stable MS. The every 15 days for 3 to 5 months across an acute attack, wefinding that mean levels of cytolytic TNFa were higher found that protein levels of soluble TNFa (range, 0–13.9than those of soluble TNFa probably indicates that sample pg/ml) did not correlate with protein levels of R1 (range,cytolytic activity was not solely due to TNFa activity. No 1.1–5.2 ng/ml) or R2 (range, 1.8–6.2 ng/ml) (Figs. 1 andcorrelation could be found between levels of soluble TNFa 2).and serum cytolytic activity of TNFa (R50.2; P50.35). mRNA of TNFa, R1 and R2 were concomitantly

R1 was measurable in all but three patients; R2 was detected in 103 of 169 (61%) PBMC extracts during themeasurable in all patients. R1 and R2 levels measured in longitudinal study in the 8 MS patients. Fluctuations ofMS patients were always among the normal reference arbitrary units of TNFa (range, 0–4.0), R1 (range, 0–6.1),values (6S.D.) of our lab (R255.262.1; R151.360.3). and R2 (range, 0–5.2) mRNA were recorded in all

Serum levels of soluble TNFa, R1, R2 and serum patients. Regression analysis of fluctuations of TNFa

cytolytic activity of TNFa were similar in clinically active mRNA levels showed a clear association with fluctuationsvs. stable MS patients and in patients with vs. patients of R1 (Fig. 1B) and R2 (Fig. 1D) mRNA levels. TNFa

Table 2Days elapsed before or after a peak of PBMC TNFa, R1 and R2 mRNA levels and the corresponding clinical attack or a peak of brain-MRI enhancinglesions

Variables Clinical Attacks Brain-MRI enhancing lesionsa b †Days before (range) Days after (range) P value Days before (range) Days after (range) P value

TNFa 44.4611.8 (0-117) 93.5612.5 (22–198) 0.011 47.7613.4 (0–161) 78.8611.5 (20–128) 0.062R1 55.8619.6 (0–280) 89.4614.1 (22–198) 0.038 46.1611.1 (0–114) 98.4627.5 (28–299) 0.085R2 46.7611.8 (0–118) 56.268.5 (23–105) 0.293 50.8610.8 (0–123) 55.3611.2 (20–136) 0.800a Values are means 6 standard errors.b P values were obtained using the Mann Whitney U test.


elevations within sustained curves (i.e. peaks) were consid-ered for statistical analysis as marker of maximal activa-tion of the TNFa system (Fig. 4).

The 8 relapsing-remitting MS patients experienced 17clinical attacks during the study (1 patient, 4 attacks; 2patients, 3 attacks; 3 patients, 2 attacks; 1 patient, 1 attack;1 patient, no attacks) (mean annual exacerbation rate: 2.1).Variations of TNFa, R1 and R2 mRNA levels wereanalyzed in the 15 attacks in which we could demonstratethe beginning and the end of a sustained increase overbasal level. A peak of TNFa, R1 and R2 coincided withthe beginning of a clinical attack in 5, 6, and 4 occasions,respectively. TNFa and R1 peaked simultaneously at thebeginning of 5 clinical attacks while TNFa and R2 orTNFa, R1, and R2 peaked together in 3 attacks.

The elevated frequency of the TNFa, R1 and R2 peaks(mean 3.2, 3.0, 3.7 peaks /year /patient, respectively) madedifficult to interpret the temporal relationship betweenTNFa system-specific events and clinical variables. Topartially overcome this ties, we measured the time elapsedbetween a TNFa, R1, or R2 peak and the ensuing clinicalattack and compared it with time elapsed from a clinicalattack and the ensuing TNFa, R1 or R2 peak. Since thetime elapsed between a TNFa or R1 peak and thefollowing clinical attack was significantly shorter than thatbetween a clinical attack and the ensuing TNFa or R1peak (P,0.05) (Table 2), we favor a non random associa-tion between these events and clinical exacerbations. R2mRNA peaks did not correlate with time elapsed before orafter a clinical attack (Table 2).

During the study we performed 80 gadolinium-enhancedbrain-MRI scans. One patient had 13 monthly scans, 3patients had 12 monthly scans, 2 patients had 11 monthlyscans, and 2 patients had 5 monthly scans. A total of 462enhancing lesions were found in 66 scans (343 lesionswere new and 119 were persisting from the previous scan);no enhancing lesions were detected in the remaining scans.

Fig. 3. Regression analysis correlating the cumulative TNFa, R1 and R2We plotted the number of active lesions over time andmRNA mean level measurements for each one of the 8 MS patientsgenerated for each patient curves reflecting disease activitystudied longitudinally during the follow-up period (1 open circle, 1

patient). Values in the y-axis represent the mean values of the TNFa, R1 as detected by MRI. We adopted for statistical analysis theand R2 for each patient and are expressed as arbitrary units. Curves maximal number of enhancing lesions (i.e. MRI peaks) aswithin the graphs represent 95% confidence intervals of the slopes. Mean marker of CNS inflammation. A total of 20 MRI peaksTNFa (R50.78; P50.02) and R2 (R50.72; P50.04) mRNA cumulative

were recorded (mean MRI peaks per patient 2.5, rangelevels correlate significantly with the number of clinical attacks while1–4). An MRI peak was contemporary to 4 of 16 (25%)mean R1 mRNA cumulative levels do not (R50.49; P50.21).clinical attacks; in the remaining attacks MRI was notperformed. We then analyzed the temporal relationshipbetween MRI peaks and TNFa, R1 and R2 mRNA peaks

mRNA levels significantly correlated also with soluble R1 (Fig. 4). Although values did not reach statistical signifi-(Fig. 1A) and R2 (Fig. 1C) serum levels. Mean values of cance, the time elapsed was shorter between an MRI peakcumulative mRNA levels of TNFa and R2, but not of R1, and the preceding TNFa (P50.06) and R1 (P50.08)were found to correlate significantly with the number of mRNA peak than between an MRI peak and the followingclinical attacks recorded during the follow-up period in the TNFa and R1 mRNA peak (Table 2), again suggesting a8 MS patients studied longitudinally (Fig. 3). non random occurrence of the two events. R2 mRNA

We plotted levels of TNFa, R1 and R2 mRNA over time peaks did not to correlate with time elapsed before or afterfor each patient. Maximal TNFa, R1 and R2 mRNA an MRI peak (Table 2).


Fig. 4. (A) Serial measurements of the number of gadolinium (gd)-enhancing brain-MRI lesions in a representative MS patient followed every 15 days forone year. (B) TNFa (filled circles), R1 (filled squares) and R2 (filled diamonds) mRNA levels, measured in PBMC from the same patient of panel A, andexpressed in arbitrary units which represent the ratio between densitometric mRNA value of the variable measured (TNFa, R1, R2) and the densitometricvalue of a house keeping gene (b-actin). In both panels each symbol on the x-axis represents a single measurement during the follow-up period. Arrowsrepresent clinical attacks and the black boxes the period of steroid treatment.

provided discordant results (Beck et al., 1988; Franciotta et4. Discussional., 1989; Sharief and Hentges, 1991; Rudick and Ran-sohoff, 1992; Chofflon et al., 1992; Dore-Duffy et al.,Following the report by Selmaj and Raine (1988)1995). Discrepancies may have arisen from the use ofshowing a potential demyelinating effect for TNFa, its roledifferent assays, each able to detect only some of thein the pathogenesis of MS has been the object of intensevarious antigenic forms of TNFa (Grimaldi et al., 1992;investigations. We now know that TNFa contributes toCorti et al., 1994b) or from the existence of two distinctCNS inflammation during phases of disease activity notTNFa receptors (R1 and R2), whose soluble and mem-only by directly damaging myelin forming cells, but alsobrane-bound forms can complex with soluble TNFa in theby fostering CNS-based immune activation (class II ex-circulation or at the site of TNFa production (Nophar etpression, stimulation of interferon-g synthesis, etc.) and byal., 1990; Aderka et al., 1992; Banner et al., 1993; Beutler,participating to the activation processes occurring in the1994; Tracey, 1994) thus hiding the TNFa epitopesperipheral lymphoid organs, well before CNS-specificrecognized by the anti-TNFa antibodies used in theeffector cells cross a permissive blood-brain barrier (Mar-specific immunoenzymatic assays.tin et al., 1992; Olsson, 1995). Accordingly, fluctuation of

So far, studies on TNFa receptors levels in MS haveTNFa levels in peripheral blood from MS patients couldfocused only on soluble R1 (Matsuda et al., 1994; Riec-represent a peripheral marker of immunological changeskmann et al., 1994a; Hartung et al., 1995; Spuler et al.,leading to demyelination in patients with MS. However, in1996), but, again, with discordant results. Matsuda et al.vitro PBMC stimulation with mitogens or in vivo direct(1994) found increased levels of soluble R1 in serum asmeasurements of soluble TNFa levels in MS patients have


well as in cerebrospinal fluid samples from ‘active’ MS dramatically affect the yield of PCR products. Even whenpatients compared to ‘stable’ patients. Rieckmann et al. these parameters are controlled, as in our experiments,(1994a) found that soluble R1 was upregulated during there is sometimes a tube-to-tube variation (i.e. eventsphases of disease stability. Hartung et al. (1995) reported occurring during the first few cycles or small temperaturean overproduction of soluble R1 during phases of variances across thermal cycler block) that preclude accur-disease activity. Finally, Spuler et al. (1996) found no ate quantitation (Gilliland et al., 1990). The use ofsignificant differences of R1 levels in serum and CSF automated cyclers using fluorescinated probes to monitorobtained serially from 10 MS patients during a one year the development of the polymerase reaction in its exponen-follow-up. tial phase might improve the validity of PCR-based

We measured soluble TNFa by an ELISA able to detect cytokine quantitations.‘free’ as well as ‘receptor-bound’ soluble TNFa in our Our longitudinal study provides at least two possiblecross-sectional study of 65 MS patients and failed to show explanations for the apparent paradox of an increase insignificant differences between MS patients and healthy TNFa (and its receptors) gene transcription correlatingindividuals, or among MS patients with different clinical with phases of disease activity and the lack of corre-and MRI evidence of disease activity. Moreover, we did sponding increase at protein level.not find any correlation between levels of soluble R1 and First, maximal transcription over basal levels of TNFa

R2 and clinical or MRI evidence of disease activity, thus and its receptors was observed well before clinical or MRIfailing to obtain preliminary support at protein level in evidence of disease activity. TNFa and TNFa receptors, infavor of an active role of soluble TNFa in the peripheral fact, fluctuated concordantly during the follow up period inactivation of the immunological process able to ignite and all patients studied, reaching the highest levels approxi-sustain CNS demyelination in these patients. mately 40 days before clinical or gadolinium-enhanced

To avoid the possible pitfalls deriving from the use of brain-MRI evidence of disease activity. Accordingly,ELISA for soluble TNFa protein measurements, we sought measurements at the time of a clinical attack might missindirect evidence of TNFa production by measuring its the time of maximal production of the soluble form ofmRNA in PBMC from MS patients followed longitudinal- these molecules.ly. TNFa mRNA, in fact, have been reported to rise 4 Second, the longitudinal study shows that mRNA ofweeks before a clinical attack and to correlate with disease TNFa is transcribed and mRNA and soluble proteins of itsactivity assessed by MRI (Rieckmann et al., 1995, 1994b). two receptors are concordantly elevated. However, noWe found that TNFa gene transcription actually occurs, significant elevation in TNFa secretion can be found in thefluctuates, and peaks before the occurrence of clinical or blood of MS patients. This finding might reflect: (a) theMRI evidence of disease activity, thus confirming previous ‘consumption’ of soluble TNFa around the site of pro-reports. Since TNFa activity is finely tuned by the duction /action (CNS demyelinating plaques? lymphonodalpresence of its receptors, we also evaluated R1 and R2 PBMC?) and its accumulation within the circulation onlygene transcription. We found that the two receptor genes when overproduced or underconsumed; (b) the retention ofare transcribed, during the course of MS, and their the translated TNFa into the membrane of TNFa-secretingtranscription rate fluctuates concordantly with that of cells (Grell et al., 1995). Considering the membrane andTNFa, thus suggesting an active role of the two TNFa soluble forms of TNFa have different biological effectsreceptors in the regulation of TNFa activity in MS. A (Grell et al., 1995), the measurement of the former mightfurther indication that the TNFa system is associated with turn to be more informative on the role of TNFa in theMS activity arises from the positive correlation found pathogenesis of MS, thus providing a simple explanationbetween cumulative TNFa and R2 mRNA levels, mea- for the discrepancies between mRNA and soluble proteinsured in the MS patients followed longitudinally, and the measurements.number of clinical attacks recorded during the follow-up We conclude that serum levels of soluble TNFa can notperiod in each of these patients. be used to assess activation of the TNFa system in MS.

The use of RT-PCR to quantitate the mRNA values, Although the role played by this pleiotropic cytokine in thehowever, suggest caution in the interpretation of the data, peripheral activation of immune responses during MSespecially when inter-patients values are compared. The remains to be elucidated, the concordant values of PBMCmain constrain of this technique in obtaining quantitative TNFa mRNA levels and soluble R1 and R2 protein serumdata is related to the amplification process. Because levels indicate that serial measurements of soluble TNFa

amplification is an exponential process (at least initially) receptors in serum of MS might represent a reliable markersmall differences in any of the variables that control the of activation of the TNFa system and of MS diseasereaction rate (concentration of polymerase, dNTPs, Mg, activity in general. Although, our data suggest that activa-DNA, and primers; annealing, extension and denaturing tion of the TNFa system during MS heralds clinicaltemperatures; cycle length and number; rate of ‘primer- exacerbations, the predictive value of this associationdimer’ formation; presence of contaminating DNA) will should be assessed in an ‘ad hoc’ study.


Scheurich, P., 1995. The transmembrane form of tumor necrosis factorAcknowledgementsis the prime activating ligand of the 80 kDa tumor necrosis factorreceptor. Cell 83, 793–802.

This work was supported by grants from Associazione Grimaldi, L.M.E., Martino, G., Franciotta, D.M., Brustia, R., Castagna,Italiana Sclerosi Multipla (AISM), and Consiglio A., Pristera, R., Lazzarin, A., 1992. Undetectable tumor necrosisNazionale delle Ricerche (CNR) ‘Progetto Citochine’. We factor-alpha in spinal fluid from HIV-1-infected patients. Ann. Neurol.

31, 687–689.are indebted to R. Bergamaschi, V. Cosi (Department ofHartung, H.P., Reiners, K., Archelos, J.J., Michels, M., Seeldrayers, P.,Neurology, University of Pavia), Bruno Colombo, M.

Heidenreich, F., Pflughaupt, K.W., Toyka, K.V., 1995. CirculatingRodegher, S. Mammi, V. Martinelli, A. Poggi, M. Rovarisadhesion molecules and tumor necrosis factor receptor in multiple

and the staff of the San Raffaele Multiple Sclerosis Center sclerosis: correlation with magnetic resonance imaging. Ann. Neurol.(Department of Neurology, University of Milano) for 38, 186–193.

Hofman, F.M., Hinton, D.R., Johnson, K., Merrill, J.E., 1989. Tumorproviding patients and contributing to this investigations.necrosis factor identified in multiple sclerosis brain. J. Exp. Med. 170,We thank G. Scotti, A. Campi, and C. Pereira (Department607–612.of Neuroradiology, University of Milano) for helping in

Kurtzke, J.F., 1983. Rating neurological impairment in multiple sclerosis:MRI data collection and Barbara Colombo for technical an expanded disability scale (EDSS). Neurology 33, 1444–1452.help. K. Vandenbroeck is a recipient of a fellowship from Maniatis, T., Fritsch, E.F., J. Sambrook, 1989. Molecular Cloning, A

Laboratory Manual. Cold Spring Harbor Laboratory, New York, 2ndthe Commission of the European Communities.edition, pp. 7.1–7.87.

Martin, R., McFarland, H.F., McFarlin, D.E., 1992. Immunologicalaspects of demyelinating diseases. Annu. Rev. Immunol. 10, 153–187.

Matsuda, M., Tsukada, N., Miyagi, K., Yanagisawa, N., 1994. IncreasedReferenceslevels of soluble tumor necrosis factor receptor in patients withmultiple sclerosis and HTLV-1-associated myelopathy. J. Neuroim-

Aderka, D., Engelmann, H., Maor, Y., Brakebush, C., Wallach, D., 1992. munol. 52, 333–340.Stabilization of the bioactivity of tumor necrosis factor by its soluble Nophar, Y., Kemper, O., Brakebusch, C., Englemann, H., Zwang, R.,receptors. J. Exp. Med. 175, 323–329. Aderka, D., Holtmann, H., Wallach, D., 1990. Soluble forms of tumor

Banner, D.W., D’Arcy, A., Janes, W., Gentz, R., Schoenfeld, H.J., Broger, necrosis factor receptors (TNF-Rs). The cDNA for the type I TNF-R,C., Loetscher, H., Lesslauer, W., 1993. Crystal structure of the soluble cloned using amino acid sequence data of its soluble form, encodeshuman 55 kd TNF receptor-human TNF beta complex: implications both the cell surface and a soluble form of the receptor. EMBO J. 9,for TNF receptor activation. Cell 73, 431–445. 3269–3278.

Beck, J., Rondot, P., Catinot, L., Falcoff, E., Kirchner, H., Wietzerbin, J., Olsson, T., 1995. Cytokine-producing cells in experimental autoimmune1988. Increased production of interferon gamma and tumor necrosis encephalomyelitis and multiple sclerosis. Neurology 45 (suppl 6),factor precedes clinical manifestation in multiple sclerosis: do cyto- S11–S15.kines trigger off exacerbations?. Acta Neurol. Scand. 78, 318–323. Porteu, F., Nathan, C., 1990. Shedding of tumor necrosis factor receptors

Beutler, B., 1994. Tumor necrosis factor: the molecules and their by activated human neutrophils. J. Exp. Med. 172, 599–607.emerging role in medicine. Raven Press, New York. Poser, C.M., Paty, D.W., Scheinberg, L., McDonald, I., Davis, F.A.,

Chofflon, M., Juillard, C., Juillard, P., Gauthier, G., Grau, G.E., 1992. Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberberg, D.H., Tourtellot-Tumor necrosis factor-alpha synthesis by cerebrospinal fluid-derived T te, W.W., 1983. New diagnostic criteria for MS: guidelines for researchcell clones from patients with multiple sclerosis. Eur. Cytokine Netw. protocols. Ann. Neurol. 13, 227–231.3, 523–531. Ruddle, N.H., Bergman, C.M., McGrath, K.M., Lingenheld, E.G., Grun-

Corti, A., Bagnasco, L., Cassani, G., 1994. Identification of an epitope of net, M.L., Padula, S.J., Clark, R.B., 1990. An antibody to lymphotoxintumor necrosis factor (TNF)-receptor type 1 (p55) recognized by and tumor necrosis factor prevents transfer of experimental allergicTNF-a-antagonist monoclonal antibody. Lymph. Cytok. Res. 13, 183– encephalomyelitis. J. Exp. Med. 172, 1193–1200.190. Rieckmann, P., Martin, S., Weichselbraun, I., Albrecht, M., Kitze, B.,

Corti, A., Poiesi, C., Merli, S., Cassani, G., 1994. Tumor necrosis factor Weber, T., Tumani, H., Broocks, A., Luer, W., Helwig, A., Poser, S.,(TNF) a quantification by ELISA and bioassay: effects of TNFa- 1994. Serial analysis of circulating adhesion molecules and TNFsoluble TNF receptor (p55) complex dissociation during assay incuba- receptor in serum from patients with multiple sclerosis: sICAM-1 is antions. J. Immunol. Meth. 177, 191–198. indicator for relapse. Neurology 44, 2367–2372.

Dore-Duffy, P., Newman, W., Balabanov, R., Lisak, R.P., Mainolfi, E., Rieckmann, P., Albrecht, M., Kitze, B., Weber, T., Tumani, H., Broocks,Rothlein, R., Peterson, M., 1995. Circulating, soluble adhesion pro- A., Luer, W., Poser, S., 1994. Cytokine mRNA levels in mononuclearteins in cerebrospinal fluid and serum of patients with multiple blood cells from patients with multiple sclerosis. Neurology 44,sclerosis: correlation with clinical activity. Ann. Neurol. 37, 55–62. 1523–1526.

Espevik, T., Nissen-Meyer, J., 1986. A highly sensitive cell line, WEHI Rieckmann, P., Albrecht, M., Kitze, B., Weber, T., Tumani, H., Broocks,164 clone 13, for measuring cytotoxic factor / tumor necrosis factor A., Luer, W., Helwig, A., Poser, S., 1995. Tumor necrosis factor-alphafrom human monocytes. J. Immunol. Meth. 95, 99–105. messenger RNA expression in patients with relapsing-remitting multi-

Franciotta, D.M., Grimaldi, L.M.E., Martino, G., Piccolo, G., Bergamas- ple sclerosis is associated with disease activity. Ann. Neurol. 37,chi, R., Citterio, A., Melzi D’Eril, G.V., 1989. Tumor necrosis factor in 82–88.serum and cerebrospinal fluid of patients with multiple sclerosis. Ann. Rudick, R.A., Ransohoff, R.M., 1992. Cytokine secretion by multipleNeurol. 26, 787–789. sclerosis monocytes. Relationship with disease activity . Arch. Neurol.

Gilliland, G., Perrin, S., Bunn, H.F., 1990. Competitive PCR for 49, 265–270.quantitation of mRNA. In: Innis, M.A., Gelfand, D.H., Sninsky, J.J., Selmaj, K.W., Raine, C.S., 1988. Tumor necrosis factor mediates myelinWhite, T.J. (Eds.), PCR protocols. A guide to methods and applica- and oligodendrocyte damage in vitro. Ann. Neurol. 23, 339–346.tions. Academic Press, San Diego, pp. 60–69. Selmaj, K.W., Raine, C.S., Cannella, B., Brosnan, C.F., 1991. Identifica-

Grell, M., Douni, E., Wajant, H., Lohden, M., Clauss, M., Maxeiner, B., tion of lymphotoxin and tumor necrosis factor in multiple sclerosisGeorgopoulos, S., Lesslauer, W., Kollias, G., Pfizenmaier, K., lesions. J. Clin. Inv. 87, 949–954.


Selmaj, K.W., Raine, C.S., Cross, A.H., 1991. Anti-tumor necrosis factor Spuler, S., Yousry, T., Scheller, A., Voltz, R., Holler, E., Hartmann, M.,therapy abrogates autoimmune demyelination. Ann. Neurol. 30, 694– Wick, M., Hohlfeld, R., 1996. Multiple sclerosis: prospective analysis700. of TNF-a and 55kDa TNF receptor in CSF and serum in correlation

Sharief, M.K., Hentges, R., 1991. Association between tumor necrosis with clinical and MRI activity. J. Neuroimmunol. 66, 57–64.factor-alpha and disease progression in patients with multiple sclerosis. Tracey, K., 1994. Tumor necrosis factor-alpha. In: Thomson, A. (Ed.),New Engl. J. Med. 325, 467–472. The Cytokine Handbook. Academic Press, London, pp. 289–304.

Tumor necrosis factor α and its receptors in relapsing-remitting multiple sclerosis

Documents

Transcript of Tumor necrosis factor α and its receptors in relapsing-remitting multiple sclerosis