Journal of Neurology

, Volume 258, Issue 5, pp 904–907

Neutralizing antibodies in interferon beta treated patients with multiple sclerosis: knowing what to do now

Commentary to: 10.1007/s00415-010-5844-5 “One-year evaluation of factors affecting the biological activity of interferon beta in multiple sclerosis patients” by S. Malucchi et al.
  • Til Menge
  • Hans-Peter Hartung
  • Bernd C. Kieseier
Short Commentary

DOI: 10.1007/s00415-011-5941-0

Cite this article as:
Menge, T., Hartung, HP. & Kieseier, B.C. J Neurol (2011) 258: 904. doi:10.1007/s00415-011-5941-0

Longterm treatment of parenterally administered interferon beta (IFNb) is one of the mainstays of multiple sclerosis (MS) therapy [1]. The first trials were conducted in the early 1980s when human fibroblast derived IFNb was administered intrathecally in an open-label fashion [2]. In ten patients treated monthly for 6 months a significant reduction of the annualized relapse rate was noted compared to the relapse rate at baseline. This effect was not observed in ten control patients with MS that, for ethical reasons, did not undergo repeat sham lumbar punctures. This prompted a larger multicentre trial in which again human IFNb purified from cultured fibroblasts was administered intrathecally [3]. After the advent of recombinant techniques to produce larger amounts of IFNb, the first pivotal trial of IFNb-1b administered subcutaneously every other day in 372 patients with relapsing-remitting MS was published in 1993 [4], and the drug was approved by the FDA for use in MS the same year. Since then two additional IFNb-1a products were evaluated in pivotal trials that yielded positive results and have been marketed subsequently [5, 6]. Treatment with IFNb was also shown to delay disease progression and halt disease activity in patients with clinically isolated syndromes [7, 8, 9]. This observation along with neuropathological and MR imaging studies emphasized the importance of early diagnosis and treatment initiation of patients [10]. Overall the effect size of the IFNb preparations is comparable and moderate [1] and the effects of these drugs have been proven to be sustained for more than 15 years [11]. Additionally the long-term safety profile of IFNb treatment is excellent [12, 13].

However, with a so called biological, i.e. a proteinaceous therapeutic compound, the issue of IFNb immunogenicity is inherent and has remained a matter of debate [14]. Antibodies against IFNb can be detected in up to 47% of patients undergoing treatment with IFNb (Table 1) [14]. These antibodies are either purely binding, and then termed binding antibodies (BAbs) or inhibit IFNb activity and then termed neutralizing antibodies (NAbs) (Fig. 1). The proportion of NAbs is variable and differs between available IFNb products (table 1) [14]. Once the bioactivity of injected IFNb is blocked by persistent and high-titer NAbs the beneficial clinically detectable treatment effects of IFNb seem to vanish, albeit with a latency of some one to four years. A diminution of the downregulatory effect on MR disease activity becomes apparent earlier, i.e. 6-18 months after development of high titer Nabs. [15]. Immediate negative effects have not been sufficiently demonstrated. It is of note that most analyses were done in a post-hoc fashion on data from trials originally designed and powered to show treatment effects of IFNb rather than detrimental effects of NAbs [16]. Thus, there is ongoing controversy regarding the clinical impact of NAbs on the management of MS patients [17]. Current European guidelines recommend switching therapy in patients with persistently high-titer NAbs with suspected clinical or MRI activity [16, 18]. Importantly, NAbs may also affect endogenous IFNb as cross-reactivity between antibodies against IFNb-1a or IFNb-1b and endogenous, human fibroblast derived IFNb was shown only recently by us [19].
Table 1

Prevalences of NAbs in prospective IFNb treatment trials



NAbs positive patients (%)


IFNb-1b s.c.

 1.6 MU qod







 8.0  MU qod













 8.0  MU qod (MxA assay)



[28, 29, 38]

IFNb-1a s.c.

 22  μg

  q wk







[6, 31]




 44  μg

  q wk







[6, 31]













IFNb-1a i.m.

 30  μg q wk






[36, 37]







  Assay not known




 60  μg q wk



[36, 37]

NAbs status was determined by CPE unless stated otherwise. NAbs were considered positive if titers >20  TRU/mL were measured

s.c. subcutaneously, i.m. intramuscularly, qod every other day, q wk once weekly, t.i.w. three times weekly

Fig. 1

IFNb-induced signal transduction and possible interferences of anti-IFNb antibodies. Binding of IFNb to its receptor (IFN α/β receptor) assembles an active signalling complex resulting in signal transduction through phosphorylation of tyrosine kinases (e.g, Tyk-2, Jak-1). Phosphorylated tyrosine kinases may activate certain transcription factors, that lead to IFNb-mediated gene transcription. Gene translation to proteins, e.g. neopterin, myxovirus resistance protein A (MxA), results in IFNb-mediated cellular function. BAbs may not interfere with the interaction of IFNb with its receptor, leaving the downstream pathways unaltered (middle panel). In contrast, NAbs hinder the engagement of IFNb with its receptor, thus inhibiting receptor activation. Intracellular downstream responses will not be turned on and IFNb response proteins not be secreted (right panel). Therefore, the assessment of the MxA response either by measuring the MxA protein or its mRNA gene transcript is a sensitive and validated marker for bioactivity of IFNb in humans

Antibodies to IFNb are measured in a two-step procedure. First, BAbs are detected by standard antibody assays. In a second step the neutralizing capacity of these antibodies is demonstrated, for which the in vitro cytopathic effect (CPE) assay can be regarded as the gold standard [20]. Recently it was demonstrated that measuring the mRNA response of the IFNb-induced antiviral MxA protein ex vivo by quantitative PCR yields a good correlate of the bioavailability of IFNb in a given subject [16, 20]. However, since there are inter-individually different response curves of the MxA mRNA expression, a small proportion of patients may be classified false-positively [21, 22, 23]. To minimize this potential error it would be ideal to compare individual MxA mRNA responses on IFNb therapy with the respective responses at treatment initiation of the same individual [24].

In this issue of the Journal S. Malucchi et al. have set out to evaluate BAbs and NAbs in 167 patients with MS after 1 year of treatment with IFNb. The proportions of patients treated with either IFNb product were roughly identical. Twenty-two percent of patients showed a pathological, i.e. negative, MxA mRNA response, indicative of reduced bioavailability. Twenty-two of those 37 patients were both positive for BAbs by capture ELISA and for NAbs by CPE; therefore, in these patients the presence of NAbs as detected by CPE resulted in a reduced in vivo bioactivity of IFNb. Eleven of all 167 patients, i.e. 6.6%, had NAbs at high-titers (>100 TRU/mL), ten of these completely abrogated the MxA response. There were also 9.6% samples with a borderline diminished MxA response. Notably, in three of the 37 MxA-negative patients, i.e. in 1.8% of the entire cohort studied, neither BAbs nor NAbs were detected. The reasons for this may be treatment incompliance, presence of soluble IFNb receptors that capture the injected IFNb [23] or a false-negative result as discussed above.

This study is very helpful and interesting for several reasons and lets us participate in challenging the NAbs issue by one of the leading NAbs research groups in Europe: It demonstrates a good, albeit incomplete, link between the presence of high-titer NAbs (as measured by CPE) and the loss of bioactivity in vivo, and it identifies a small (~ 2%), yet substantial number of patients with loss of IFNb bioactivity not explainable by the presence of NAbs. It is worthwhile to assess these patients in greater detail, as suggested by the authors, given the availability of alternative treatment options, although these may have potential side effects not seen with IFNb. Thus, before switching a patient to glatiramer acetate or escalating the patient to, e.g., natalizumab, a false-negative MxA result and noncompliance should be ruled out. Lastly, this study reproduces independently in a well-defined cohort of patients the frequency of NAbs-positivity and high-titer NAbs-positivity after exactly one year of treatment reported earlier. It should be rewarding to follow up those 6.6% of patients with a reduced, not yet negative MxA responses in order to quantify the in proportion of patients that will eventually completely loose their IFNb bioactivity, i.e. become NAbs positive.

Emergence, frequency and origin of NAbs remain challenging issues while to date consensus seems to have been achieved regarding their immediate impact on IFNb bioactivity and their delayed impact on disease activity. The paper by S. Malucchi et al. does not attempt to settle the matter — a task too vast for one group anyway. But the careful and detailed expert analysis of the antibody status helps to further our understanding of this complicated picture of treatment induced humoral immunity.


Supported in part by the German Ministry for Education and Research (BMBF), “German Competence Network Multiple Sclerosis (KKNMS)”.

Conflict of interest

TM has received honoraria and travel support from BayerHealthcare, BiogenIdec, and MerckSerono. HPH and BCK received honoraria for consulting and speaking, with approval by the rector of HHU, from BayerHealthcare, BiogenIdec, MerckSerono, Novartis, Teva Sanofi Aventis. Their department, with approval by the rector of HHU, received unrestricted grants for research on neutralizing antibodies from these companies.

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • Til Menge
    • 1
  • Hans-Peter Hartung
    • 1
  • Bernd C. Kieseier
    • 1
  1. 1.Department of Neurology, Medical FacultyHeinrich-Heine-UniversityDüsseldorfGermany

Personalised recommendations