Immunogenicity of Interferon-b in MultipleSclerosis Patients: Influence of Preparation,

Dosage, Dose Frequency, andRoute of Administration

Christian Ross, MD,* Katja Maria Clemmesen, MD,* Morten Svenson, PhD,* Per Soelberg Sørensen, MD,†Nils Koch-Henriksen, MD,‡ Gunhild Lange Skovgaard, MD,* Klaus Bendtzen, MD,*

and the Danish Multiple Sclerosis Study Group§

A total of 754 consecutive patients with relapsing-remitting multiple sclerosis were investigated for interferon-b(IFNb) antibodies by protein-G affinity chromatography and antiviral neutralization bioassay during 24 months on 6MIU (22 mg) of subcutaneous IFNb-1a once weekly (n 5 143) or three times weekly (n 5 160), 6 MIU (30 mg) ofintramuscular IFNb-1a once weekly (n 5 140), or 8 MIU every other day of IFNb-1b (n 5 311). The proportion ofbinding antibodies was higher in those receiving IFNb-1b compared with 6 MIU of IFNb-1a three times weekly (97 vs89% at 12 months), and fewer became positive if 6 MIU of IFNb-1a was administered once weekly (58 vs 89%). Fewerpatients on intramuscular than subcutaneous IFNb-1a became positive (33 vs 58%). The binding and neutralizingcapacities were higher in the IFNb-1b group than in the IFNb-1a groups; these differences, however, were not significantafter 12 months. The number of positive patients varied considerably and depended on the amount of IFN added to thebioassay; adding 10 LU/ml or more masked antibody detection. Antibodies induced by either preparation neutralizedboth IFNb species but not IFNa. In conclusion, IFNb-induced antibodies are frequently found in multiple sclerosispatients, and IFNb-1b is more immunogenic than IFNb-1a. The immunogenicity of IFNb-1a increases with the fre-quency of administration and if it is given subcutaneously.

Ross C, Clemmesen KM, Svenson M, Sørensen PS, Koch-Henriksen N, Skovgaard GL, Bendtzen K,Danish Multiple Sclerosis Study Group. Immunogenicity of interferon-b

in multiple sclerosis patients influence of preparation, dosage, dose frequency,and route of administration. Ann Neurol 2000;48:706–712

Interferon (IFN)-b has been shown to reduce the ex-acerbation rate, severity of clinical relapses, progressionof permanent neurological deficits, and lesion load onmagnetic resonance imaging (MRI) in patients suffer-ing from relapsing-remitting and secondary progressivemultiple sclerosis (MS).1–6

A potential problem with prolonged clinical use ofrecombinant human cytokines, including IFNb, is an-tibody induction.7 For example, neutralizing antibodiesreduce the effect of IFNa in viral hepatitis and hema-tological cancers, in which recurrence of infection ormalignant cells in the blood can be accurately mea-sured.8–10 Reduced bioavailability of IFNb has alsobeen demonstrated in MS patients with neutralizingantibodies,11,12 and antibody-positive patients have

presented with relapse rates and MRI results similar tothose of patients receiving placebo.13–16

The reported frequencies of neutralizing anti-IFNbantibodies in MS have varied from 7 to 42%, althoughbinding antibodies have been found in up to 78% ofIFNb-treated patients.14,16,17 These numbers not onlydepend on the IFNb preparations, dosage, dose fre-quency, and route of administration but also on theantibody detection methods, emphasizing the need forstandardized IFN antibody measurements. Unfortu-nately, attempts to achieve this have failed, and to-gether with difficulties in assessing the impact of theseantibodies in individual MS patients, this has led tocontroversies with regard to the clinical importance ofIFNb-induced antibodies.

From the *Laboratory for Clinical Interferon Research, Institute forInflammation Research, and †Department of Neurology, Rigshospi-talet University Hospital, Copenhagen, and ‡Department of Neu-rology, Aalborg Sygehus, Ålborg, Denmark.

§Members of the Danish Multiple Sclerosis Study Group are listedin the Appendix.

Received Feb 10, 2000, and in revised form May 23. Accepted forpublication May 24, 2000.

Address correspondence to Prof Bendtzen, Institute for Inflamma-tion Research IIR 7521, Rigshospitalet University Hospital, 9 Bleg-damsvej, DK-2100 Copenhagen, Denmark.

706 Copyright © 2000 by the American Neurological Association

In this study, we have used optimized assays to mea-sure antibody binding and neutralization of IFNb. Toassess the influence of IFNb preparation, dosage, andfrequency and route of administration, these tests wereperformed in a central laboratory to analyze prospec-tively sampled sera from all IFNb-treated MS patientsin Denmark.

Materials and MethodsPatients and ControlsFrom 1996 to 1998, 754 patients began treatment withIFNb in Denmark. As part of a national protocol, bloodsamples from all patients were collected at 0, 3, 6, 12, 18,and 24 months of therapy. Patients were treated with either6 MIU (22 mg) of IFNb-1a (Rebif) subcutaneously (SC)once weekly (n 5 143), 6 MIU (22 mg) of IFNb-1a (Rebif)SC three times weekly (n 5 160) (because this treatment wasnot approved until May 1998, antibody formation in thesepatients was followed for 12 months only), 6 MIU (30 mg)of IFNb-1a (Avonex) intramuscularly (IM) once weekly(n 5 140), or 8 MIU of IFNb-1b (Betaferon) SC everyother day (n 5 311). The sex ratio, mean age, and meanEstimated Disability Status Scale score by treatment areshown in the Table. Blood was collected by venipuncture 48hours after injection of IFNb, and the sera were isolated andstored at 220°C until assay.

Patients from the Prevention of Relapses and Disability byInterferon Beta-1a Subcutaneously in Multiple Sclerosis(PRISMS) study were included as dosage controls.4 Thesepatients received 6 MIU (22 mg) or 12 MIU (44 mg) ofIFNb-1a (Rebif) or placebo SC three times weekly. Normalhuman sera (n 5 2,450) were obtained from healthy Danishblood donors, and 74 different batches of human IgG (En-dobulin) were provided by Immuno (Heidelberg, Germany).

InterferonsHuman recombinant IFNb-1a was donated by Serono(Copenhagen, Denmark). Escherichia coli–derived recombi-nant IFNb-1b was purchased from PeproTech (Rocky Hill,CA). International IFN standards were donated by the Na-tional Institutes of Health (Bethesda, MD).

125I-Labeling of IFNbIFNb-1a and IFNb-1b were radiolabeled by the chloramineT method and purified as detailed previously.18

Affinity Chromatography Assay for BindingAnti-IFNb AntibodySera were diluted in phosphate-buffered saline (PBS), pH7.2, and 25% fetal calf serum (FCS) and incubated for 2hours at 22°C and overnight at 4°C in 100 ml of PBS con-taining 3,000 cpm of 125I-IFNb.

Binding of IgG to 125I-IFNb was investigated by affinitychromatography assay (ACA) using columns with 500 ml ofprotein-G Sepharose CL-4B (Pharmacia, Uppsala, Sweden).Unbound 125I-IFNb was collected at 4°C with PBS and0.5% FCS as running buffer. 125I-IFNb bound to IgG waseluted with 100 nM of glycine-HCl, pH 2.4. Nonspecificbinding was assessed in the presence of 200 ng of unlabeledIFNb. Radioactivity was determined in a gamma counter(1272 CliniGamma; LKB, Wallac, OY, Finland).

Binding titers were calculated as the reciprocal value of theminimum serum dilution causing significant binding of3,000 cpm per 100 ml of 125I-IFNb-1a, which was greaterthan the mean plus 3 SD of the 125I-IFNb binding capacityof sera from 754 untreated MS patients.18 The binding ca-pacity was calculated as the percentage bound of added tracerusing 25% serum in the screening assay. Three sera withlow, medium, and high IFNb binding capacities and a rabbitantibody to human IFNb were used as inter- and intra-assayvariation controls.18

Dissociation AssaySera were evaluated by ACA and diluted in PBS to yield50% maximum binding. The protein concentration was keptconstant by varying the amounts of added FCS. Immunecomplexes were formed by incubating the diluted sera with3,000 cpm of 125I-IFNb for 24 hours. Dissociation of thesecomplexes was measured 0, 5, 24, and 48 hours after addi-tion of greater than 200 times excess unlabeled IFNb.Bound and free tracer were separated by protein-G chroma-tography, and the times to 50% dissociation were assessed.

IFN Neutralization BioassayAn antiviral neutralization bioassay (ANB) was carried outessentially as described.18 Briefly, MC-5 cells, a subclone ofA549 cells, were seeded in microtrays at a concentration of10,000 cells per well and incubated at 37°C in a 5% CO2

atmosphere. The IFNb preparations were added at concen-trations of 3, 10, and 100 laboratory units (LU)/ml, corre-sponding to 0.9, 3, and 30 IU/ml (1 LU is the amount ofIFN inducing 50% protection against challenge virus). After

Table. Patient Characteristics

Rebif(once weekly)

Rebif(three times weekly)

Avonex(once weekly)

Betaferon(every other day)

n 143 160 140 311Sex ratio (female:male) 1.98:1 2.02:1 2.26:1 1.96:1Age (mean yr) 37.1 36.4 38.4 38.4Age (range yr) 18–55 14–62 14–67 16–64Estimated Disability Status Scale score (mean) 3.0 2.8 2.5 2.9Estimated Disability Status Scale score (range) 0–5.5 0–6.0 0–6.0 0–6.0

Ross et al: Anti-IFNb Antibodies in MS Patients 707

24 hours, encephalomyocarditis virus was added at the low-est concentration that by itself caused 100% cell death. Afteranother 24 hours, the antiviral effect of IFN was measuredusing nitroblue tetrazolium assay. To eliminate interferenceby endogenous serum toxicity, individual sera were alwaystested at dilutions of 20, 100, 1,000, and 2,000 times andevaluated in ANB at the lowest dilution showing no toxicity.To avoid false-positive and false-negative results, separate ex-periments to control for endogenous antiviral activity andserum toxicity were included in each analysis.

EthicsThe study was approved by the Central Scientific EthicsCommittee of Denmark and carried out according to theHelsinki II Declaration.

StatisticsFisher’s exact test was used to compare the numbers ofantibody-positive patients, and the Mann Whitney U testwas used for comparisons of binding and neutralizing capac-ities. A probability value greater than 0.05 was considered tobe nonsignificant.

ResultsWe first investigated the frequency of anti-IFNb anti-bodies in healthy individuals, in untreated MS patients,and in normal human IgG preparations. Specific anti-bodies were demonstrated in ACA in only 2 of 2,450normal blood donors (0.08%) and at a low level in 1of 754 MS patients (0.13%) before therapy. In com-parison, 24 of 74 (32%) different batches of pooledhuman IgG were positive.

Number of MS Patients Developing BindingAnti-IFNb AntibodiesThe ACA for anti-IFNb antibodies exhibited inter-and intra-assay variations below 5%. The binding wassaturable and specific for IFNb, but an excess of bothIFNb-1a and IFNb-1b could displace the binding inall cases tested.

As shown in Figure 1, most patients treated withIFNb every other day or three times weekly serocon-verted within 3 months, with the highest number inIFNb-1b–treated patients. In contrast, less than 10%of patients receiving IFNb-1a once weekly becameantibody-positive after 3 months. After 12 months,58% of patients treated with 6 MIU (22 mg) ofIFNb-1a SC once weekly had developed antibodiescompared with 89% of those treated with 6 MIU (22mg) of IFNb-1a SC three times weekly. A significantlylower proportion (33%) developed antibodies when 6MIU (30 mg) of IFNb-1a was administered IM onceweekly. The highest frequency of antibodies was seenin patients treated with 8 MIU of IFNb-1b SC everyother day (97%). Similar findings were seen after 24months of therapy. Six of the antibody-positive pa-tients on IFNb-1b and 2 of the antibody-positive pa-

tients on IFNb-1a tested negative after 24 months stillon treatment. Only one of these had high-capacitybinding antibodies after 3 to 12 months of therapy.

Patients included from the PRISMS study4 served asdosage controls: 0, 6 MIU (22 mg), or 12 MIU (44mg) of IFNb-1a SC three times weekly. After 12months of therapy, more of the Danish patients treatedwith 6 MIU three times weekly became positive (89%)compared with patients from the PRISMS studytreated either with a similar dose (82%) or with ahigher dose of 12 MIU three times weekly (74%; p ,0.05). None of the placebo-treated patients developedanti-IFNb antibodies.

A significantly higher proportion of female (64%)than male (47%) patients became positive after 12months of treatment with IFNb-1a SC once weekly(p , 0.02). This was essentially unchanged at 18 and24 months.

Fig 1. Number of multiple sclerosis (MS) patients with inter-feron (IFN)-b–induced antibodies. Sera were positive in affin-ity chromatography assay (ACA) if the IgG bound greater thanthe mean plus 3 SD (' .18%) of the 125I-IFNb bindingcapacity of sera from 754 untreated MS patients. ● 8 MIU ofIFNb-1b subcutaneously (SC) every other day (total of 28MIU per week; n 5 311). h 6 MIU of IFNb-1a SC threetimes weekly (total of 18 MIU per week; n 5 160). { 6MIU of IFNb-1a SC once weekly (total of 6 MIU per week;n 5 143). ‚ 6 MIU of IFNb-1a intramuscularly onceweekly (total of 6 MIU per week; n 5 140). Significant dif-ferences (Fisher’s exact test): Group ● versus Groups { and ‚,respectively, at all time points: p , 0.0001; Group ● versusGroup h: p , 0.0001 at 3 and 6 months and p , 0.02 at12 months; Group h versus Group {: p , 0.0001 at 3 and6 months and p , 0.005 at 12 months; Group h versusGroup ‚: p , 0.0001 at all time points; Group { versusGroup ‚: p , 0.01 at 6 months and p , 0.001 at 12 and18 months.

708 Annals of Neurology Vol 48 No 5 November 2000

Binding Capacities of Sera fromAntibody-Positive PatientsAs shown in Figure 2, sera from patients treated withIFNb-1b had higher binding capacities than thosefrom the other groups. In addition, IFNb-1a inducedantibodies with higher binding capacities if adminis-tered three times weekly rather than once weekly and ifgiven SC rather than IM.

Interestingly, sera from the PRISMS dosage controlgroup receiving 12 MIU of IFNb-1a three timesweekly (total of 36 MIU per week) had significantlylower binding capacities compared with sera from pa-tients on 6 MIU of IFNb-1a three times weekly. Morethan 10% of these sera had measurable levels of IFNb,suggesting the presence of circulating IFNb up to 48hours after subcutaneous administration of high dosesof the cytokine.

Number of Antibody-Positive Patients withNeutralizing AntibodiesThe study started as a comparison of patients onIFNb-1b every other day and IFNb-1a once weekly.All patients had sera investigated for both IFN bindingand neutralization. In patients belonging to the othergroups, only sera positive for binding antibodies inACA were tested by ANB. As shown in Figure 3, morethan 70% of patients on 8 MIU of IFNb-1b every

other day were positive after 6 months compared withless than 25% of patients in the group receiving 6MIU of IFNb-1a once weekly. This difference dimin-ished after 24 months of therapy.

The number of patients with neutralizing antibodiesvaried considerably depending on the selected sensitiv-ity of the bioassay. Thus, as shown in Figure 4, theaddition of increasing amounts of IFNb in ANB low-ered the proportion of antibody-positive results whentesting the same subpopulation of patients.

Neutralizing Capacities of Sera fromAntibody-Positive PatientsAs shown in Figure 5, sera from patients in theIFNb-1b group (8 MIU every other day SC) had thehighest neutralizing capacity. There were no significantdifferences after 12 months of subcutaneous IFNbtherapies. In contrast, sera from patients on intramus-cular IFNb-1a reached a plateau, and the neutralizingcapacities were significantly lower in the groups onsubcutaneous therapies after 12 and 18 months.

It is noteworthy that the Grossberg and Kawademethod for calculating neutralizing titers19 revealed thehighest titers in sera from 4 patients treated withIFNb-1a.

Fig 2. Binding capacities of sera from antibody-positive pa-tients. The patients are the same as those shown inFigure 1. The tracer background ranged from 7 to 12%.Results are shown as mean values 6 SD of the raw data.18

Significant differences (Mann Whitney U test): Group ● ver-sus Groups { and ‚, respectively: p , 0.001 at 3 to 18months and p , 0.01 at 24 months; Group ● versus Grouph: p , 0.05 at 3 and 6 months; Group h versus Group {:p , 0.02 at 3, 6, and 12 months; Group h versus Group‚: p , 0.02 at 3 months, and p , 0.01 at 6 and 12months; Group { versus Group ‚: p , 0.01 at 6 and 18months.

Fig 3. Number of antibody-positive patients with interferon(IFN)-b neutralizing antibodies. An antiviral neutralizationbioassay was carried out with 3 LU/ml of IFNb added. Serawere considered positive if they neutralized greater than themean plus 3 SD of the neutralizing capacity of sera from 754untreated multiple sclerosis patients. Patients included in thisinvestigation were treated as follows: ● 8 MIU of IFNb-1bsubcutaneously (SC) every other day (n 5 311); { 6 MIU ofIFNb-1a SC once weekly (n 5 143). Significant differences(Fisher’s exact test): p , 0.0001 at 3, 6, 12, and 18 monthsand p , 0.001 at 24 months.

Ross et al: Anti-IFNb Antibodies in MS Patients 709

Specificity of the IFNb AntibodiesAntibodies induced by IFNb-1a or IFNb-1b alwaysbound and neutralized both IFNb species.

Dissociation Time of Preformed Immune ComplexesSera from 5 patients on prolonged IFNb therapyshowed lower capacities to bind IFNb even thoughtheir neutralizing capacities increased at the same time.To evaluate whether this was due to affinity matura-tion, we tested the kinetics of dissociation of immunecomplexes formed by antibodies in these sera and ex-ogenous IFNb. Figure 6 shows the outcome of one ofthree investigations giving essentially similar results.The decline in binding capacity and the increased neu-tralizing capacity after 12 months of therapy were ac-companied by a substantial increase in dissociationtime of the immune complexes.

DiscussionIFNa and IFNb have been used therapeutically inmany diseases, and IFNb is now the drug of choice inthe treatment of relapsing-remitting MS.1,2,4–6,20 Thedevelopment of anti-IFNb antibodies during therapyhas been reported in variable numbers of MS pa-tients.2,13,15,16,21,22 Some antibodies have been foundto bind but not to neutralize IFNb, and others neu-tralize IFNb in ex vivo neutralization assays. Reducedbioavailability of IFNb and a positive correlation be-

tween neutralizing antibodies and reduced clinical effi-cacy have also been reported.2,11,12 It is still unclear,however, whether different IFNb preparations vary intheir capacity to induce antibodies and whether theseantibodies neutralize the effect of continued therapy. Itis also unknown whether “rescue therapy” with another

Fig 4. Influence of bioassay sensitivity on the number of pa-tients with measurable neutralizing antibodies. Data from 40patients treated with 8 MIU of interferon (IFN)-b-1b subcu-taneously (SC) every other day. Bioassays were carried outwith the addition of IFNb-1b at concentrations of 3 LU/ml(● 5 high sensitivity), 10 LU/ml (J 5 medium sensitivity),or 100 LU/ml (C 5 low sensitivity). (See Fig 3 for criterionfor positivity.)

Fig 5. Neutralizing capacities of sera from antibody-positivepatients. The patients are the same as those shown inFigure 1. Sera from antibody-positive patients were tested inan antiviral neutralization bioassay at 5% (vol/vol), andinterferon-b was added at 3 LU/ml. Data are shown as meanvalues 6 SEM: Group ● versus Group {: p , 0.001 at 3and 6 months; Group ● versus Group h: p , 0.01 at 3months; Group h versus Group {: p , 0.01 at 3 and 6months; Group h versus Group ‚: p , 0.01 at 6, 12, and18 months; Group { versus Group ‚: p , 0.01 at 6 and12 months.

Fig 6. Binding and neutralizing capacity versus dissociationtime. Results from 1 of 3 patients on interferon (IFN)-b-1btherapy. J–J 5 binding capacity. C 5 neutralizing capacityusing 1 LU/ml of IFNb. ● 5 neutralizing capacity using 10LU/ml of IFNb. ✘ 5 dissociation time.

710 Annals of Neurology Vol 48 No 5 November 2000

IFNb preparation may overcome the problem of invivo neutralizing antibodies as has been shown in thecase of IFNa therapy.23,24

For clinical assessment, it is important to optimizetests for both binding and neutralizing antibodies toreflect the in vivo situation as closely as possible. Forexample, ANB, in which IFN is added in quantitiesout of proportion to those needed for virus inactiva-tion, would tend to underestimate clinically relevantanti-IFN activities. Conversely, human blood samplesoften contain substances that are toxic to cells used inANB. If appropriate measures are not taken, bioassaysof such samples would erroneously report the presenceof anti-IFN activity.

The study presented here is unique in that all IFNb-treated patients in an entire country had antibodiesmeasured prospectively for 12 to 24 months in a singlespecialized laboratory. The results show that althoughanti-IFNb antibodies in MS patients before IFNbtherapy are as rare as in the normal population (0.1%),most of the patients develop measurable binding andneutralizing antibodies within a few months of therapy.Indeed, almost all patients treated with IFNb-1b de-veloped binding antibodies, and up to 80% developedneutralizing antibodies. The positive sera from theIFNb-1b group also contained the highest levels of an-tibodies compared with those from the IFNb-1agroups. More frequent administration induced a higherrate of seroconversion as exemplified by 6 MIU (22mg) of IFNb-1a SC once weekly versus 6 MIU (22mg) of IFNb-1a SC three times weekly. Dosage controlsera from the PRISMS study4 confirmed that more fre-quent administration of IFNb-1a resulted in an in-creased proportion of positive cases, but the antibodytiters in these cases remained unaltered regardless ofIFNb-1a dosage. Sera from patients on the highest andmost frequent dosage of IFNb-1a (12 MIU three timesweekly) had significantly lower binding capacities thanthose from patients on 6 MIU of IFNb-1a three timesweekly. This, however, was most likely caused by thepresence in these sera of IFNb-1a not yet cleared fromthe previous administration. IM administration in-duced antibodies less frequently than administration,which is possibly due to faster clearance from the in-jected tissue.

The number of MS patients with neutralizing anti-bodies was heavily influenced by the chosen level ofsensitivity of ANB, particularly with regard to theamounts of IFN applied. Ten LU/ml has been recom-mended as standard, where 1 LU is the amount of IFNinducing 50% protection against challenge virus. Ourdose-response curves showed a plateau of maximumprotection at IFN levels between 2.5 and 3 LU/ml,however. With the use of 10 LU/ml of IFN, two thirdsmight be neutralized by antibodies without being no-ticed, and with the use of 5% serum, this would cor-

respond to a blood neutralizing capacity of up to 150LU/ml of IFN. Thus, it is highly likely that the use ofneutralization assays with the addition of 100 LU/mlof IFN explains why some investigators report very lowfrequencies of neutralizing antibodies in MS patientstreated with IFNb.15

Both the quantity and quality of anti-IFN antibodiesmay affect neutralization of IFN in vivo and thus theclinical importance of these antibodies. A factor thatmight be important in this regard is epitope matura-tion, (ie, the accumulation of antibodies with higheraffinities during the course of therapy, possibly at theexpense of low-affinity antibodies). This would explainour observation of increased neutralization capacityalong with decreased binding capacity in sera of pa-tients on prolonged therapy and the fact that these seraformed immune complexes with exogenous IFNb thatdissociated much less readily than complexes in seradrawn earlier during therapy.

This investigation was not intended to study theclinical effects of anti-IFN antibodies and whetherthese antibodies jeopardize further therapy with IFN.This may be clarified from a later clinical assessment ofthis large cohort of Danish MS patients. Nevertheless,reduced levels of biological markers of in vivo IFN ac-tivity have already been reported in MS patients withIFN neutralizing antibodies, and some of these patientssuffer relapse rates and MRI levels equal to placebo lev-els.1,2,4,5,20 Furthermore, because antibodies inducedby both IFNb-1a and IFNb-1b cross-reacted in allcases investigated in our study, a shift from one IFNbpreparation to the other is unlikely to benefit patientsin whom antibodies have already been induced.

We conclude that even though sera may be screenedfor anti-IFNb antibodies in binding assays, ANB is re-quired for a more accurate understanding of these an-tibodies. This is best carried out in specialized labora-tories and should include quantitation using serialsample dilutions along with controls for toxicity andendogenous IFN activity. Assay sensitivity should al-ways be demonstrated. Finally, the use in ANB of 10LU/ml of IFN or more is most likely too high toscreen for clinically relevant anti-IFN antibodies.

AppendixThe Danish Multiple Sclerosis Study Group includes P.S.Sørensen, Rigshospitalet University Hospital, Copenhagen(Chairman); S. Helweg-Larsen, Rigshospitalet UniversityHospital, Copenhagen; J. Frederiksen, Glostrup UniversityHospital, Glostrup; K. Jensen, Hillerød Hospital, Hillerød;A. Heltberg, Roskilde Hospital, Roskilde; A. Nordenbo,Holbæk Hospital, Holbæk; F. Boesen, Slagelse Hospital, Sla-gelse; S. Deth, Næstved Hospital, Næstved; O. Kristensen,Odense University Hospital, Odense; R. Kirkebye, Sønder-borg Hospital, Sønderborg; I. Zeeberg, Vejle Hospital, Vejle;E. Stenager, Esbjerg Hospital, Esbjerg; H.J. Hansen, Århus

Ross et al: Anti-IFNb Antibodies in MS Patients 711

University Hospital, Århus; N.B. Jensen, Viborg Hospital,Viborg; J. Tørring, Holstebro Hospital, Holstebro; andN. Koch-Henriksen, Aalborg Hospital, Ålborg (Secretary).

This work was supported by the Danish Multiple Sclerosis Society,the IMK Almene Fond, and the Danish Biotechnology Program.

Drs Allan R. Thomsen (Institute of Microbiology and Immunology,University of Copenhagen), Morten B. Hansen (Blood Bank Rigs-hospitalet, Copenhagen), and Rudolph Schosser (Immuno GmbH,Heidelberg, Germany) kindly provided encephalomyocarditis vi-rus, sera from normal individuals, and human IgG preparations.Ares-Serono kindly provided sera from the PRISMS study. MereteTjalve, Anne Ambjørnsen, and Ole Christensen provided excellenttechnical assistance.

References1. The IFNB Multiple Sclerosis Study Group. Interferon b-1b is

effective in relapsing-remitting multiple sclerosis. I. Clinical re-sults of a multicenter, randomized, double-blind, placebo-controlled trial. Neurology 1993;43:655–661

2. The IFNB Multiple Sclerosis Study Group and The Universityof British Columbia MS/MRI Analysis Group. Interferon b-1bin the treatment of multiple sclerosis: final outcome of the ran-domized controlled trial. Neurology 1995;45:1277–1285

3. Jacobs LD, Cookfair DL, Rudick RA, et al. Intramuscular in-terferon beta-1a for disease progression in relapsing multiplesclerosis. Ann Neurol 1996;39:285–294

4. PRISMS (Prevention of Relapses and Disability by Interferonb-1a Subcutaneously in Multiple Sclerosis) Study Group. Ran-domised double-blind placebo-controlled study of interferonb-1a in relapsing/remitting multiple sclerosis. Lancet 1998;352:1498–1504

5. European Study Group on interferon b-1b in secondary pro-gressive MS. Placebo-controlled multicentre randomised trial ofinterferon b-1b in treatment of secondary progressive multiplesclerosis. Lancet 1998;352:1491–1497

6. Arnason BG. Immunologic therapy of multiple sclerosis. AnnuRev Med 1999;50:291–302

7. Bendtzen K, Hansen MB, Ross C, et al. High-avidity autoan-tibodies to cytokines. Immunol Today 1998;19:209–211

8. von Wussow P, Freund M, Block B, et al. Clinical significanceof anti-IFN-a antibody titres during interferon therapy. Lancet1987;2:635–636

9. Steis RG, Smith JW, Urba WJ, et al. Resistance to recombinantinterferon a-2a in hairy-cell leukemia associated with neutralizinganti-interferon antibodies. N Engl J Med 1988;318:1409–1413

10. Milella M, Antonelli G, Santantonio T, et al. Neutralizing an-

tibodies to recombinant a-interferon and response to therapy inchronic hepatitis C virus infection. Liver 1993;13:146–150

11. Rudick RA, Simonian NA, Alam JA, et al. Incidence and sig-nificance of neutralizing antibodies to interferon beta-1a inmultiple sclerosis. Neurology 1998;50:1266–1272

12. Deisenhammer F, Reindl M, Harvey J, et al. Bioavailability ofinterferon beta 1b in MS patients with and without neutralizingantibodies. Neurology 1999;52:1239–1243

13. Antonelli G. Development of neutralizing and binding antibod-ies to interferon (IFN) in patients undergoing IFN therapy. An-tiviral Res 1994;24:235–244

14. The IFNB Multiple Sclerosis Study Group and the Universityof British Columbia MS/MRI Analysis Group. Neutralizing an-tibodies during treatment of multiple sclerosis with interferonb-1b: experience during the first three years. Neurology 1996;47:889–894

15. Abdul-Ahad A, Shah S, Galazka A. Incidence of antibodies tointerferon-b in patients treated with recombinant humaninterferon-b1a from mammalian cells. Cytokines Cell MolTher 1997;3:27–32

16. Kivisakk P, Alm GV, Tian WZ, et al. Neutralising and bindinganti-interferon-b-1b (IFN-b-1b) antibodies during IFN-b-1btreatment of multiple sclerosis. Mult Scler 1997;3:184–190

17. Abdul-Ahad A, Shah S, Galazka A. Treatment of multiple scle-rosis with interferon b-1b. Neurology 1997;49:641–642

18. Ross C, Svenson M, Hansen MB, et al. High avidity IFN-neutralizing antibodies in pharmaceutically prepared humanIgG. J Clin Invest 1995;95:1974–1978

19. Grossberg SE, Kawade Y. The expression of potency of neutral-izing antibodies for interferons and other cytokines. Biotherapy1997;10:93–98

20. Rudick RA, Fisher E, Lee JC, et al. Use of the brain parenchy-mal fraction to measure whole brain atrophy in relapsing-remitting MS. Neurology 1999;53:1698–1704

21. Antonelli G. In vivo development of antibody to interferons: anupdate to 1996. J Interferon Cytokine Res 1997;17(Suppl 1):S39–S46

22. Antonelli G, Bagnato F, Pozzilli C, et al. Development of neu-tralizing antibodies in patients with relapsing-remitting multiplesclerosis treated with IFN-b1a. J Interferon Cytokine Res 1998;18:345–350

23. Khan OA, Dhib-Jalbut SS. Neutralizing antibodies to inter-feron beta-1a and interferon beta-1b in MS patients are cross-reactive. Neurology 1998;51:1698–1702

24. Antonelli G, Simeoni E, Bagnato F, et al. Further study on thespecificity and incidence of neutralizing antibodies to interferon(IFN) in relapsing remitting multiple sclerosis patients treatedwith IFN beta-1a or IFN beta-1b. J Neurol Sci 1999;168:131–136

712 Annals of Neurology Vol 48 No 5 November 2000